Understanding and Optimizing the Local Catalyst Environment in CO₂ Reduction Electrodes - CaltechTHESIS

Understanding and Optimizing the Local Catalyst Environment in CO₂ Reduction Electrodes - CaltechTHESIS
CaltechTHESIS
A Caltech Library Service
About
Browse
Deposit an Item
Instructions for Students
Understanding and Optimizing the Local Catalyst Environment in CO₂ Reduction Electrodes
Citation
Welch, Alexandra Justine
(2022)
Understanding and Optimizing the Local Catalyst Environment in CO₂ Reduction Electrodes.
Dissertation (Ph.D.), California Institute of Technology.
doi:10.7907/4s78-cq55.
Abstract
Understanding and managing the local microenvironments in carbon dioxide reduction catalysts is crucial for optimizing device performance. In particular a locally high pH can increase catalyst selectivity and activity, as well as indicate which part of the catalyst is most active. In this thesis we begin by studying how nanoporous catalysts can induce this locally high pH in an aqueous system. We observe an increase in both Faradaic efficiency and partial current density for carbon monoxide in the nanoporous system relative to a planar metal film. We then show that this same nanoporous architecture can be used for improved device performance in a gas diffusion electrode configuration. We also perform copper underpotential deposition and secondary ion mass spectroscopy to show that almost half of the catalyst is not in contact with the electrolyte in this configuration. Then we use confocal fluorescent microscopy to image the local pH in a gas diffusion electrode to determine which parts of the electrode are most active. Through a combination of experiment and simulations we find that the catalyst within thin cracks of the microporous layer is most active for carbon dioxide reduction. While the study of local pH and wetting is the main focus of this thesis, we also explore how light can be used to improve selectivity and activity. In particular we study gold nanoparticles on p-type gallium nitride and copper nanoparticles on p-type nickel oxide. Finally, this thesis also explores how carbon dioxide conversion can actually be deployed. We discuss opportunities for combining carbon dioxide capture and conversion, as well as evaluate different pathways for renewable methane generation.
This thesis gives in depth analysis of electrochemical carbon dioxide reduction catalysts as well as putting this research into the larger context of how such devices can be deployed. We hope that by combining systems level thinking and specific device studies better carbon dioxide conversion systems can be realized.
Item Type:
Thesis (Dissertation (Ph.D.))
Subject Keywords:
carbon dioxide reduction, solar fuels, confocal microscopy, nanoporous
Degree Grantor:
California Institute of Technology
Division:
Engineering and Applied Science
Major Option:
Applied Physics
Thesis Availability:
Public (worldwide access)
Research Advisor(s):
Atwater, Harry Albert
Thesis Committee:
Greer, Julia R. (chair)
Goddard, William A., III
See, Kimberly
Atwater, Harry Albert
Defense Date:
1 October 2021
Non-Caltech Author Email:
ajwelch1031 (AT) gmail.com
Record Number:
CaltechTHESIS:11062021-151820825
Persistent URL:
DOI:
10.7907/4s78-cq55
Related URLs:
URL
URL Type
Description
DOI
Article adapted for ch. 2
DOI
Article adapted for ch. 4
DOI
Article adapted for ch. 5
DOI
Article adapted for ch. 6
DOI
Article adapted for ch. 7
DOI
Article adapted for ch. 8
ORCID:
Author
ORCID
Welch, Alexandra Justine
0000-0003-2132-9617
Default Usage Policy:
No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:
14420
Collection:
CaltechTHESIS
Deposited By:
Alexandra Welch
Deposited On:
21 Dec 2021 22:15
Last Modified:
08 Nov 2023 00:12
Thesis Files
PDF
- Final Version
See Usage Policy.
110MB
Repository Staff Only:
item control page
CaltechTHESIS is powered by
EPrints 3.3
which is developed by the
School of Electronics and Computer Science
at the University of Southampton.
More information and software credits
Understanding and
optimizing the local
catalyst environment in
CO2 reduction electrodes

Thesis by

Alex J. Welch

In Partial Fulfillment of the Requirements for
the degree of
Doctor of Philosophy

CALIFORNIA INSTITUTE OF TECHNOLOGY
Pasadena, California

2022
(Defended October 1st, 2021)

ii

ã 2021
Alex J. Welch
ORCID: 0000-0003-2132-9617

iii

ACKNOWLEDGEMENTS
Over the past five years I have grown so much, and this was made possible by an
extraordinary group of people: colleagues and mentors who pushed me to explore new scientific
territory and believed in my ability to tackle any challenges, friends and family who supported me
no matter what and were always down to hang out and laugh. This experience would not have been
the same without all of you and I am truly thankful.
First, I would like to thank my advisor, Harry Atwater. I remember trying to decide if I
should go to Caltech or stay at Stanford for graduate school, and all of my professors at Stanford
kept telling me that I should go to Caltech because of what an amazing mentor Harry is. They
could not have been more right. I am constantly inspired by Harry’s passion and excitement for
science, and no matter how busy Harry is, he always found time to meet with me when I asked.
Harry has taught me to always be curious and that no project is impossible.
Next, I would like to thank Giulia Tagliabue and Joe Duchene. I came into graduate school
having not taken a chemistry class since high school, and therefore had a lot of questions about
what was going on. Both of them took so much time to train me on all of the equipment, explain
concepts to me, and answer all of my many questions. I have learned so much from both of them.
Particularly, Giulia taught by example to always comes into lab with a smile and to not get
discouraged when an experiment doesn’t work. Joe taught me how to really think about the
purpose of every experiment and to design the most effective studies. After they both became
professors, I put a post-it note on my desk saying “What would Giulia/Joe do?” and it helps me get
into the right mindset to get through challenges.
I would also like to thank many other members of the Atwater team and other researchers in
Jorgensen. I would not have made it nearly as far on all of my projects without all of the
brainstorming that I did with people like Aidan, Eowyn, Annette, Lucy, Ian, Ibadillah, CX, Sophia,
Matthias, Nick, Artur, and Kate. They were always around the lab and happy to take a look at any
experimental problem I was having or help me think of how to design an experiment to probe a
particular question. Aidan helped me solve many of the cell design problems I was having with the
confocal microscope and it has been great seeing him carry on the work with nanoporous gold. I
have had so much fun working with Annette, and it is exciting to see her take over the pH imaging
experiments. Ibadillah was a pleasure to work with on techno-economic analysis project. Ian
taught me so much about electrochemistry, and I really appreciate how he always puts down what
he is doing to answer any question that I have. Eowyn is an amazing friend and encouraged me

iv
when experiments were not working. I also want to thank Dagny, Joeson, and Magel for working
on the e-beam with me. In particular I would like to thank Joeson for teaching so much about how
the e-beam operates, working tirelessly to fix the e-beam after the flood, and for all of our wine
nights. There are many more people in the A-team to thank than I have time to here, but you should
know that without the community that the A-team provided during my time, Caltech would have
been much less fun.
I also want acknowledge my mentors that helped get me to Caltech. First, I want to thank
Pat Burchat and Kam Moler for getting me excited about physics and for all of their advice through
the years. I also want to thank Mark Brongersma for his help in deciding what classes I should take.
I especially want to thank Yi Cui for letting me work in his lab starting freshman fall. Po-Chun was
very patient in teaching me how to do all the experiments and Yi always had amazing new ideas for
us to try. In my senior year I worked with Jen Dionne on plasmonic upconverters with Guru. Jen’s
enthusiasm for science is something that I strive for and Guru helped me to believe in my abilities
as a scientist.
Now, while all of the scientific mentorship and help was crucial for this thesis, I could not
have done it without the support of my friends and family. My first-year roommates Sophia,
Sophie, and Sarah helped me find my home in LA. I was lucky enough to get to live with Sophie
for the next 3 years; she was always there for me no matter what was going on. I also want to give
a shout out to some other Caltech friends from my year: Haley, Cora, Camilla, and Areum. The
group that was my anchor through all of Caltech and makes up my LA family is Taco Tuesday.
Every other Tuesday, even during the pandemic, we ate tacos, drank margaritas, and laughed.
Megan, Kendall, Eowyn, Spencer, Alexander, Alex C., Alex S., Jason, Lauren, Nathan, John, Amy,
Sydney, Elyse, Cullen, Sophie, Taylor, Claire, and Jamie: I feel like I could pick up the phone and
call any of you right now and you would drop everything to help me. Your friendship means the
world to me. I would like say thank you to the skiing crew, those who have not already been listed
are Jackie, Eric, Madison, Anna, Clayton, Tarek, and Brian; thank you for all of the adventures.
Two friends that I would like to say an extra thank you to are Megan and Kristine, my scuba
buddies; whether we are at the cellar or on an adventure there is no one I would rather be with.
Kristine, it was so much fun to get to live with you during the pandemic and camp every weekend.
Megan, Mondays are now my favorite day of the week because of you. You bring so much
happiness to everyone you interact with.
My family has made me who I am today. My dad is always there to share my love for
science and building tables. My mom has taught me compassion by always being there and caring

for others. John and Tom, you always make me laugh and there is no one who will know me quite
like you two. Grandma ,thanks for always being so proud of me. All of you have given me so
much support and love throughout my life. I look forward to many more card games, hikes, and
general good times. Thank you for always believing me and helping me become my best self.
And last, and most importantly, thank you Jamie. Thank you for making me smile every
day and being my partner in all things. Thank you for letting me epoxy tables in our bathroom and
for always being excited about my work. Thank you for everything.

vi

ABSTRACT
Understanding and managing the local microenvironments in carbon dioxide
reduction catalysts is crucial for optimizing device performance. In particular a locally
high pH can increase catalyst selectivity and activity, as well as indicate which part of the
catalyst is most active. In this thesis we begin by studying how nanoporous catalysts can
induce this locally high pH in an aqueous system. We observe an increase in both
Faradaic efficiency and partial current density for carbon monoxide in the nanoporous
system relative to a planar metal film. We then show that this same nanoporous
architecture can be used for improved device performance in a gas diffusion electrode
configuration. We also perform copper underpotential deposition and secondary ion
mass spectroscopy to show that almost half of the catalyst is not in contact with the
electrolyte in this configuration. Then we use confocal fluorescent microscopy to image
the local pH in a gas diffusion electrode to determine which parts of the electrode are
most active. Through a combination of experiment and simulations we find that the
catalyst within thin cracks of the microporous layer is most active for carbon dioxide
reduction. While the study of local pH and wetting is the main focus of this thesis, we
also explore how light can be used to improve selectivity and activity. In particular we
study gold nanoparticles on p-type gallium nitride and copper nanoparticles on p-type
nickel oxide. Finally, this thesis also explores how carbon dioxide conversion can
actually be deployed. We discuss opportunities for combining carbon dioxide capture
and conversion, as well as evaluate different pathways for renewable methane
generation.
This thesis gives in depth analysis of electrochemical carbon dioxide reduction
catalysts as well as putting this research into the larger context of how such devices can
be deployed. We hope that by combining systems level thinking and specific device
studies better carbon dioxide conversion systems can be realized.

vii

PUBLISHED CONTENT AND CONTRIBUTIONS
Operando Local pH Measurement within Gas Diffusion Electrodes Performing
Electrochemical Carbon Dioxide Reduction
Alex J. Welch, Aidan Q. Fenwick, Annette Böhme, Hsiang-Yun Chen, Ian Sullivan, Xueqian Li,
Joseph S. Duchene, Chengxiang Xiang, and Harry A. Atwater, Jour. Phys. Chem. C. 2021
DOI: 10.1021/acs.jpcc.1c06265
Contributions: conception and design, fabrication and data collection, data analysis,
manuscript writing, and revision
Comparative Techno-Economic Analysis of Renewable Generation of Methane Using
Sunlight, Water, and Carbon Dioxide
Alex J. Welch, Ibadillah A Digdaya, Ron Kent, Paul Ghougassian, Harry A. Atwater, and
Chengxiang Xiang ACS Energy Letters 2021 6 (4), 1540-1549.
DOI: 10.1021/acsenergylett.1c00174
Contributions: conception, model design, data analysis, manuscript writing, and revision.
Bicarbonate or Carbonate Processes for Coupling Carbon Dioxide Capture and
Electrochemical Conversion
Alex J. Welch, Emily Dunn, Joseph S. DuChene, and Harry A. Atwater, ACS Energy
Letters 2020 5 (3), 940-945. DOI: 10.1021/acsenergylett.0c00234
Contributions: conception, data collection and analysis, manuscript writing, and revision.
Optical Excitation of a Nanoparticle Cu/p-NiO Photocathode Improves Reaction Selectivity
for CO2 Reduction in Aqueous Electrolytes
Joseph S. DuChene, Giulia Tagliabue, Alex J. Welch, Xueqian Li, Wen-Hui Cheng, Harry A.
Atwater, Nano Letters 2020 20 (4), 2348-2358. DOI: 10.1021/acs.nanolett.9b04895
Contributions: device fabrication, characterization, revision.
Nanoporous Gold as a Highly Selective and Active Carbon Dioxide Reduction Catalyst
Alex J. Welch, Joseph S. DuChene, Giulia Tagliabue, Artur Davoyan, Wen-Hui Cheng,
and Harry A. Atwater, ACS Applied Energy Materials 2019 2 (1), 164-170.
DOI: 10.1021/acsaem.8b01570
Contributions: fabrication and data collection, data analysis, manuscript writing, and revision
Quantifying the Roles of Surface Plasmon Excitation and Hot Carrier Transport in Plasmonic
Devices
Giulia Tagliabue, Adam S. Jermym, Ravishankar Sundararaman, Alex J. Welch, Joseph S.
Duchene, Ragip Pala, Artur R. Davoyan, Prineha Narang, and Harry A. Atwater, Nature
Communications 2018 9, 3394.
DOI: 10.1038/s41467-018-05968-x
Contributions: fabrication and characterization, revision

viii
Hot Hole Collection and Photoelectrochemical CO2 Reduction with Plasmonic Au/p-GaN
Photocathodes
Joseph S. DuChene, Giulia Tagliabue, Alex J. Welch, Wen-Hui Cheng, and Harry A. Atwater,
Nano Letters 2018 18 (4), 2545-2550. DOI: 10.1021/acs.nanolett.8b00241
Contributions: fabrication and characterization, revision

ix

TABLE OF CONTENTS

Acknowledgements………………………………………………………………………..iii
Abstract ……………………………………………………………………........................vi
Published Content and Contributions……………………………………...........................vii
Table of Contents…………………………………………………………………………..ix
List of Figures…………..………………………………………………………………….xi
List of Tables……………………………………………………………………………xxiii
Chapter I: Introduction to CO2 reduction……………………………………………………1
1.1 Role of solar fuels in mitigating climate change……………………………………1
1.2 Fundamentals of CO2 reduction…………………………………………….………2
1.3 Catalyst composition………………………………………..……..……..………....4
1.4 Nanostructuring and surface facets of catalyst……………………………….....…...5
1.5 Improved CO2 reduction via plasmonic metal nanoparticles…………………….….6
1.6 Local pH effects…………………………...………………………………..……….7
1.6.1 Electrolyte concentration and relevant species…………………………....………7
1.6.2 Dependence of HER and CO2 reduction on pH………………………………….10
1.6.3 Techniques for the measurement of local pH…………………………………….10
1.7 Device components and CO2 delivery method……………………………………..11
1.8 Thesis outline………………………………………………………………...……15
Chapter II: Nanoporous gold as a highly selective and active carbon
dioxide reduction catalyst…………………………………………………..……….…..…21
2.1 Introduction…………………………………………………………….….……....21
2.2 Fabrication and characterization of catalyst…………………………..…….……...22
2.3 Electrochemical characterization……………………………………..…………....27
2.4 Influence of local pH on performance……………………………..………………31
2.5 Catalyst Stability…………………………………………………..………………33
2.6 Conclusion……………………………………………………..………………….36
Chapter III: Nanoporous gold GDE………………………………….……………………39
3.1 Introduction………………………………………………….…………………….39
3.2 Fabrication and characterization of nanoporous gold………………………..……..41
3.3 Electrochemical characterization of nanoporous gold………………………..…….45
3.4 Copper underpotential deposition characterization………………………….……..47
3.5 Secondary ion mass spectroscopy characterization……………………….……….50
3.6 Conclusion………………………………………………………………………...52
Chapter IV: Operando local pH measurement within gas diffusion
electrodes performing electrochemical carbon dioxide reduction…………………………55
4.1 Introduction………………………………………………………………………..55
4.2 Experimental set up and characterization……………………………………….....58
4.3 Results and Discussion…………………………………………………………….64
4.4 Conclusion……………………………………………………………...…………75
Chapter V: Hot hole collection and photoelectrochemical CO2
reduction with plasmonic Au/p-GaN photocathodes………………………………………80

5.1 Introduction……………………………………………………………….……….80
5.2 Fabrication and characterization of Au/p-GaN……………………………….……83
5.3 Electrochemical studies of Au/p-GaN…………………………….…………….....85
5.4 Fabrication, characterization, and electrochemical studies of Au/p-NiO………......90
5.5 Selectivity of Au/p-GaN in light and dark conditions………………….…………..92
5.6 Conclusion..…………………………………………………………………...…..93
Chapter VI: Optical excitation of a nanoparticle Cu/p-NiO photocathode
improves reaction selectivity for CO2 reduction in aqueous electrolytes…………….……..98
6.1 Introduction………………………………………………………………...……...98
6.2 Fabrication and characterization of electrode……………………………………..101
6.3 Photoelectrochemical studies………………………………………..…………...103
6.4 Discussion of mechanism……………………………………………...…………110
6.5 Conclusion……………………………………………………..………………...112
Chapter VII: Bicarbonate or carbonate processes for coupling carbon
Dioxide capture and electrochemical conversion………………………………..……….119
7.1 Introduction…………………………………………………………….………...119
7.2 CO2 capture and conversion steps……………………………..………………….121
7.3 Direct (bi)carbonate conversion………………………………………………….123
7.4 Conclusion……………………………………………………………………….126
Chapter VIII: Comparative techno-economic analysis of renewable
Generation of methane using sunlight, water, and CO2………………………..………….129
8.1 Introduction………………………………………………………………………129
8.2 Carbon dioxide capture…………………………………………………………...133
8.3 Water generation…………………………………………………...…………….135
8.4 Hydrogen generation……………………………………………...……………...137
8.5 Methane generation………………………………………………...…………….140
8.6 Conclusion……………………………………………………...……………,,,...145
Chapter IX: Conclusions…………………………………………………..……………..150
Appendix A……………………………………………………………..………………..155
Appendix B………………………………………………………………..…………..…167

xi

LIST OF FIGURES

1.1 Synopsis of various steps involved in capture CO2 and transforming it into valuable
chemicals……………………………………………………………………………………2
1.2 Adsorption energy of CO* vs the adsorption energy of H*. Marks in red are metals that
primarily produce H2, teal marks produce products with more than two carbons, blue marks
produce mainly CO, and yellow marks produce predominantly HCOOH. Reproduced from
Bragger et al. Copyright 2017, John Wiley and Sons………………………………..………....4
1.3 Concentration of carbon dioxide, hydrogen, hydroxyl, bicarbonate, carbonate, and potassium
ions as a function of bulk pH in a potassium (bi)carbonate electrolyte at 25°C and a pressure of
1atm. Reproduced from Singh, et al. Copyright 2015, Royal Society of Chemistry ………......9
1.4 Schematic of a CO2 reduction electrochemical cell. Electrons flow to the working electrode,
cathode, to reduce CO2 to hydrocarbons (HC), and electrons are removed from H2O to create
O2 at the counter electrode, anode. The dotted regions indicate boundary layers (BL). Arrows
indicate which way different species migrate through the BL. The reference electrode is used to
measure the voltage being applied to the working electrode and the current is measured between
the cathode and anode ………………...…………………………………………………....12
1.5 Different electrochemical CO2 reduction reaction schemes are shown in (a) and (b). (a)
shows a schematic of aqueous-phase CO2 reduction where CO2 is first dissolved in an aqueous
electrolyte then reduced at catalyst surface. (b) shows a vapor fed CO2 reduction system where
the CO2 is delivered in gas phase to catalyst on a gas diffusion electrode; the electrolyte can either
be aqueous or polymer. (c) shows a plot of Faradaic efficiency vs. partial current density for
different reported devices. Open shapes refer to aqueous-phase CO2 reduction and filled shapes
represent vapor-fed CO2 reduction devices. The numbers on the plot refer to references from
the original paper. Reproduced from Higgins et al. Copyright 2019, American Chemical
Society……….…………………………………………………………………………......13
1.6 (a) shows a schematic of gas diffusion CO2 reduction electrode. The orange lines are carbon
fibers, the orange circles are carbon black, and the black circles are catalyst. (b)-(d) show
different states that the catalyst can be in; (b) is flooded, (c) is wetted, and (d) is dry. This was
reproduced and edited from Weng et al. Copyright 2018, Royal Society of Chemistry…..........14
2.1 Helium FIB images of (a) top-down view of a nanoporous Au (np-Au) film that was etched
at room temperature (RT) and (b) top-down view of a np-Au film that was etched at low
temperature (LT). (c) SEM cross-section image of a RT-etched np-Au film. All scale bars
represent 100 nm. (d) Electrochemical surface area enhancement as a function of film thickness
for RT-etched np-Au films as determined by Cu underpotential deposition experiments……..22
2.2 XRD spectra of (a) RT np-Au (~800 nm thick) on a glass substrate, (b) planar Au film on
glass, and (c) Au foil after flame annealing. (d) Zoom-in of (200) peak where planar Au and npAu have been increased by 20x. From the XRD it is evident that the planar Au film is highly

xii
oriented in the (111) orientation. The data also shows that the full width half max of the np-Au
is much larger than that of the Au foil, indicating that the np-Au has smaller grains………….24
2.3 Bright field transmission electron microscopy (TEM) images of RT np-Au film. The red
arrows denote grain boundaries. All scale bars represent 5 nm…………………….…………25
2.4 Transmission electron microscopy (TEM) analysis of RT np-Au films. (a) Bright-field TEM
of a particular region of the np-Au film along with (b) the corresponding selected area electron
diffraction (SAED) pattern. (c1-c7) Dark-field TEM images from the np-Au film obtained from
the particular spots numbered in panel b. (d) Bright-field TEM image of np-Au film along with
dark-field TEM images numbered 1-3. All scale bars in TEM images represent 20 nm……….26
2.5 Transmission electron microscopy (TEM) analysis of RT np-Au films. (a-c) Bright-field
TEM of a particular region of the np-Au film along with (d-f) a corresponding dark-field TEM
image obtained from that particular region of the np-Au film. All scale bars in all TEM images
represent 20 nm…………………………………………………………………………….26
2.6 Electrochemical performance of Au cathodes. Faradaic efficiency (FE) for CO (filled bars)
and H2 (open bars) as a function of applied potential (E) with (a) 661 ± 10 nm thick room
temperature etched (RT) np-Au film, (b) 664 ± 5 nm thick low temperature etched (LT) np-Au
film, (c) planar Au film, and (d) commercial Au foil. Partial current density (J) for CO (filled
circles) and H2 (open circles) as a function of applied potential for (e) 661 ± 10 nm thick RT npAu film, (f) 664 ± 5 nm thick LT np-Au film, (g) planar Au film, and (h) commercial Au foil.
Each data point represents the average FE for CO or H2 obtained over 2-3 h of continuous
electrolysis at the indicated potential with iR compensation. The partial current densities also
represent the average value observed over the same time period. All data was obtained from the
same electrode along the potential sweep……………………………………………………28
2.7 (a) Cyclic voltammetry of LT np-Au (grey curve) and RT np-Au (dark blue curve) in 0.5 M
H2SO4 obtained at a scan rate of 50 mV s-1. From these data we can determine that the LT
sample has ~3x greater electrochemical surface area as compared to the RT sample. (b) shows
a histogram of pore widths measured on these two samples at three different locations on the
sample. Representative SEM cross-section images of (c) RT np-Au and (d) LT np-Au films. The
scale bars on both images correspond to 100 nm. These images correspond to the actual
electrodes used for electrochemical tests involving different electrolyte concentrations………29
2.8 (a) SEM cross-section image of ~150 nm-thick RT np-Au film. The scale bar represents 100
nm. (b) Faradaic efficiency as a function of applied potential (E) for 150 nm-thick RT np-Au
film. (c) Partial current density for CO (JCO) from a 150 nm-thick and ~800 nm-thick RT np-Au
sample. Considering the 4x smaller surface area of the thinner film, the relative JCO between the
two films is unexpected. We hypothesize that this is due to mass transport limitations……….30
2.9 Influence of electrolyte concentration on CO2 reduction selectivity with Au cathodes. (a-c)
The Faradaic efficiency for CO as a function of applied potential (E) obtained at two different
electrolyte concentrations (both saturated with CO2) for (a) 809 ± 15 nm thick room
temperature etched (RT) np-Au film, (b) 821 ± 22 nm thick low temperature etched (LT) npAu film, and (c) 50 nm thick planar Au film. (d-f) The corresponding average current density (J)
obtained at the applied potential (E) observed at two different electrolyte concentrations for (d)
RT np-Au, (e) LT np-Au, and (f) planar Au film. (g-i) Predicted solution pH at the surface of

xiii
the electrode for (g) RT np-Au, (h) LT np-Au, and (i) planar Au film. A planar electrode
geometry is assumed for the simulations………………………...………………………….31
2.10 Extended electrochemical stability data for Au cathodes. The Faradaic efficiency for CO
(filled circles) and H2 (open circles) was measured every 15 min via gas chromatography over
the course of 24 h at an applied potential of E = –0.5 VRHE with iR compensation for (a) room
temperature (RT)-etched np-Au, (b) planar Au film, and (c) Au foil…………….……...…….34
2.11 Extended electrochemical stability data for a RT np-Au film (~800 nm thick). The Faradaic
efficiency for CO (filled points) and H2 (open points) was measured every 15 min via gas
chromatography over the course of 110 h at an applied potential of E = –0.5 VRHE with iR
compensation………………………..…………………………………………………...35
2.12 (a,b) SEM images of a ~800 nm thick RT np-Au film (a) before and (b) after testing for
110 h at −0.5 V vs. RHE. (c,d) SEM images of a planar Au film (c) before and (d) after testing
for 24 h at −0.5 V vs RHE. (e,f) SEM images of a Au foil (e) before and (f) after testing for 24
h at −0.5 V vs RHE. There is no visible difference between any of the planar samples before
and after testing. In the np-Au sample there is some minor coarsening of the ligaments, but no
significant changes to the film morphology are observed…………………….………..…….35
2.13 XRD spectra of Au films before and after 24 h of testing for (a) ~800 nm thick RT np-Au
film, (b) planar Au film, and (c) Au foil. The peak at 68° in the RT np-Au film and the planar
Au film is due to the Si substrate. In Figure 2.2, the XRD patterns were collected from films
supported on a glass substrate to avoid the peak from the Si substrate. Negligible differences
were observed between Au peaks obtained from on Si vs. glass substrates………….………..36
3.1 A schematic of the nanoporous gold gas diffusion electrode used in this study in a vapor
CO2 fed device. The bottom image shows the electrode structure. The support consists of a
gas diffusion layer composed of carbon fibers, on top of which is coated carbon black and PTFE
which makes up the microporous layer. The nanoporous gold is coated on top of this. The top
three panels show the different configurations that the nanoporous gold can be in during
operation: flooded, wetted, and dry………………………………….……………………...40
3.2 shows SEM images of the electrode at different phases in the fabrication process. (a)-(c)
shows images of the bare carbon paper, Sigracet 38BC. (d)-(f) show images of the gold silver
alloy on the carbon paper. (g)-(i) show images of the nanoporous gold morphology from a 35%
Au alloy that forms after the nitric acid etch…………………………………………………42
3.3 SEM characterization of nanoporous gold (np-Au) electrodes with a varying gold atomic
percent (%Au) of 15%Au (a-c), 25%Au (d-f), 35%Au (g-i), and 45%Au (j-l)……… ……………..42
3.4 SEM characterization of planar gold (pl-Au) and nanoporous gold (np-Au) electrodes with
varying gold atomic percent (%Au) (a-d). CO2R performance of the pl-Au and np-Au electrodes
is shown via Faradaic efficiencies (e-h) and partial current densities (i-l). For all plots, CO is
denoted in pink and H2 by blue. Each data point is the average of three distinct
electrodes…………………………………………………………..………………………43
3.5 SEM characterization of nanoporous gold (np-Au) electrodes with a varying gold atomic
percent (%Au) on silicon. 20%Au (a-c), 25%Au (d-f), and 30%Au (g-i)…………… …………..44

xiv
3.6 cross sectional SEM of a 300 nm thick 35% gold nanoporous gold electrode on carbon
paper. From this image we can see how uneven the carbon paper substrate is and that the
nanoporous gold has a homogenous pore structure through the entire thickness…….………44
3.7 Contact angle measurements of nanoporous gold electrodes with varying atomic gold
percentages. 0 atomic percent Au indicates that there is no catalyst layer and 100 atomic percent
Au indicates that there was a solid gold film deposited. All samples were deposited on Sigracet
38BC unless otherwise noted………………………………………………….……………46
3.8 (a) relative surface area enhancements of a 300 nm thick Au electrode and 300 nm thick 35
%Au np-Au normalized to the Au on Si sample as calculated by Cu UPD. (b), Surface area
enhancement of 100 nm, 300 nm, and 900 nm thick 35%Au electrodes. The dashed line shows
a linear fit with an R2 = 1.00……………………………………………..……………..……47
3.9 (a) relative surface area of three identical 300 nm thick, 35%Au np-Au electrodes under
different CO2 flow rates normalized to the surface area of the same electrode in an aqueous CO2
fed system. (b) Faradaic efficiencies of the electrodes at different CO2 flow rates. (c) Partial
current densities of the electrodes at different CO2 flow rates. All experiments for b and c were
carried out at -0.92 VRHE…………………………………………………………………….49
3.10 Secondary ion mass spectroscopy data summary. (a) shows normalized Cu counts relative
to normalized depth into the np-Au electrode with 0 as the surface of the np-Au electrode in
contact with the electrolyte and 1 as the np-Au electrode in contact with the microporous layer.
(b) is the ratio of the of the vapor CO2 fed counts to the aqueous CO2 fed counts in (a) and
represent the portion of the catalyst that is in contact with the electrolyte as a function of
depth………………………………………………………………………………………50
3.11 Secondary ion mass spectroscopy raw data for aqueous CO2 fed system and vapor CO2 fed
system. (a) shows the gold counts, (b) the copper counts, and (c) the carbon counts………….51
4.1 Overview of a Cu gas diffusion electrode (GDE) for CO2 reduction studies. (a) Crosssectional diagram of the custom electrochemical cell designed to enable in situ confocal
fluorescent microscopy experiments. (b) Schematic of a typical Cu GDE, not to scale. (c), (d)
Scanning electron microscope (SEM) images of a Cu GDE. (e), (f) SEM images of an uncoated
microporous layer. (g), (h) SEM images of the gas diffusion layer……………………………57
4.2 (a) shows a schematic of the electrochemical cell used for imaging the pH via confocal
fluorescent microscopy. The bottom plate is the gas chamber and the top plate holds the
electrolyte. This setup has no membrane and the electrolyte is constantly being flowed across
the active catalyst layer. (b) shows a top-down photo of the electrochemical cell without the
microscope objective. (c) shows a photo of the entire experimental setup with the objective in
the cell, the electrolyte bath, and the pump to circulate the electrolyte through the
electrochemical cell………………………………………………………………………....59
4.3 Characterization of the pH-sensitive DHPDS fluorescent dye. (a) shows the absorbance of
DHPDS for different pH solutions. vertical black lines denote the two different excitation
wavelengths (lex1 = 458 nm) and (lex2 = 458 nm) used for the study. (b) shows the ratio of

xv
fluorescence emission from a 458 nm and 488 nm excitation wavelength as a function of solution
pH. After acquiring all of the data, we fit the pH data to the function, y = –a/(1+exp(–b*(x –
c)))+d. We found the coefficients to be a = –33.72, b = 1.413, c = 8.083, and d = 5.571 for
95% confidence bounds. We therefore have an error of 0.3 pH units. (c) shows the current (J)
vs applied electrode potential (E) for a CO2 reduction electrode with (dashed line) and without
(solid line) DHPDS dye in the electrolyte……………………………………………………60
4.4 Characterization of how the pH-sensitive DHPDS fluorescent dye affects the activity and
selectivity of the copper GDE. (a) shows the current density (J) vs time before the dye is added
(0 to 34 minutes) and after the dye is added to the electrolyte at 35 minutes. The partial current
density for HER increases but the CO2 reduction partial current densities remain stable. (b)
shows the Faradaic efficiency for gas products vs time. The dye is added to the electrolyte at 35
minutes. The Faradaic efficiency for HER increases but the CO2 reduction Faradaic efficiency
remains stable. (c) shows the Faradaic efficiency for liquid products before and after the dye was
added. The Faradaic efficiency for the CO2 reduction reactions remain similar before and after
the addition of the dye, albeit with slight increase in ethanol (orange) and decrease in formic acid
(blue) …………………………………………………………………………...……….....60
4.5 Electrocatalytic characterization of a GDE composed of carbon paper coated with 300 nm
of Cu on top of the microporous layer. (a) Faradaic Efficiency and (b) partial current density, J,
for each product as a function of electrode potential, E. The figure legend applies to both panels
(a) and (b)………….. …………………………………………………………………..…..61
4.6 Stability of the Cu GDE working electrode potential (Ewe) over time. (a) shows four different
electrochemical tests where the current density is set to -3.4 mA/cm2. From this we can see that
there are only very small changes in potential of the working electrode between tests. (b) shows
electrochemical tests with varying current density. We note that the potential of the working
electrode is very stable after the first 5 minutes, indicating that electrode is stable throughout the
run……………………………………………………………….……………...…………62
4.7 Operando mapping of solution pH in three dimensions over a Cu GDE. Maps are obtained
at the same location on the electrode at different heights above the electrode surface and at
different current densities. From top to bottom, each row of maps corresponds to 27 µm above,
0 µm (at the surface), 15 µm below, and 30 µm below the electrocatalyst surface. From left to
right, each column of maps was obtained at 0 mA/cm2 (no reaction under open circuit
conditions), –1.6 mA/cm2, –3.4 mA/cm2, –14.7 mA/cm2, and –28.0 mA/cm2. The pH color
scale and the scale bar in the bottom right-hand corner apply to all images…………………63
4.8 pH maps at three different locations along the electrode surface. Position 1 is off to the side
of the electrolyte inlet, position 2 is near the electrolyte inlet, and position 3 is near the electrolyte
outlet. The first row shows pH maps all taken at –3.4 mA/cm2 and the second row shows pH
maps all taken at –7.0 mA/cm2. We observe the hot spots for all 3 positions at -3.4 mA/cm2
and we do not observe the hot spots at –7.0 mA/cm2………………………………..………64
4.9: Influence of electrolyte flow on the spatial resolution of pH maps. (a) and (b) are pH maps
stitched together, taken at 9 µm above the surface of the electrode and at a current density of –
14.7 mA/cm2. In (a) the left half of the image has no Cu while the right half has Cu catalyst.
The electrolyte is flowing from left to right across the surface of the electrode. In (a) the left half

xvi
of the image is coated with Cu while the right half has no Cu catalyst. The electrolyte is flowing
from left to right across the surface. (c) shows a schematic that is not to scale of what the pH
gradient looks like in both x-y and x-z planes. The area within the orange circle in the x-y plane
indicates where the Cu catalyst is located on the GDE. The position where image (a) was taken
is indicated in panel (c) by the red square and the red line labeled ‘a’. The position where image
(b) was taken is indicated in panel (c) by the red square and the red line labeled
‘b’…………………………………………………………………………………………..65
4.10 Influence of physical confinement on CO2 reduction performance. (a) Low-magnification
SEM image of a Cu gas diffusion electrode; (b) High-magnification SEM image of a Cu gas
diffusion electrode with an overlay of the Cu signal obtained from an EDS map, red shading
indicates Cu covered regions. (c) Measured pH as a function of trench width. The orange data
points denote the average trench width. The error bars in the abscissa axis indicate the variation
in trench width with the smallest and largest end points denoting the thinnest and the widest
point along the trench. The error bars in the ordinate axis represent the standard deviation of
pH values within the trench. (d) and (e) pH maps obtained from two representative trenches
with different widths located at different regions on the electrode at a distance of 8 µm below
the electrode surface while operating at a current density of –3.4 mA/cm2……………....…..66
4.11: SEM and EDS maps of two locations on a GDE with 300 nm Cu. The SEM images at the
left show the location for all EDS maps in the corresponding row. From the EDS maps we can
see that at the bottom of the cracks, there is Cu while less carbon and PTFE (Fluorine signal) is
present……………………………………………………………………………………..68
4.12 SEM and EDS maps of three locations on a GDE with 300 nm Cu. Red shading denotes
carbon, green shading denotes Cu, and blue indicates fluorine. From the EDS maps we can see
that there is Cu deposited on the side walls of the trenches……………..……………………68
4.13 pH maps at three different locations along the electrode surface. All measurements were
taken at –3.4 mA/cm2. The image in the first row is taken at the surface of the electrode and
the second row is at 8 µm below the surface (0 µm). We observe that for all cases the pH within
the narrow trench is higher than it is at the surface of the GDE……………………………...69
4.14 shows a COMSOL simulation of the velocity of the electrolyte in the electrochemical cell
with the objective submerged in the electrolyte. (a) shows the velocity in the x-z plane with the
inlet on the left and outlet on the right. (b) shows the velocity in the y-z plane with the electrolyte
flowing into the page. From these simulations it is clear that the velocity underneath the
objective is small……………………………………………………………………………69
4.15 Simulations of local pH within and around trenches of various dimensions in the GDE. (a)
Schematic of the model used for simulations indicating regions of CO2 flux (white), current
density (orange), concentration boundary conditions (green), and electrolyte flow (blue). pH
map in the x-z plane at a uniform current density of –3.4 mA/cm2 for a trench with a depth of
50 µm deep and a width of (b) 20µm and (d) 5 µm. (c) pH map in the x-z plane for a trench [50
deep x 5µm wide] with an average current density of –3.4 mA/cm2 where the current density in
the trench is twice as high as the current density on the surface. The CO2 flux is constant through
all surfaces and boundary conditions are kept the same for all simulations. The pH scale bar
applies to all pH profiles (b)-(d)……………………… …………………………………….73

xvii
4.16 shows COMSOL simulations of the pH profile in a trench that is 5µm wide and 50 µm
deep at different current densities. The current density is constant over all surfaces. (a) shows
the pH profile at –3.4 mA/cm2, (b) shows the pH profile at –7.0 mA/cm2, and (c) shows the
pH profile at –14.7 mA/cm2. From these simulations we observe that the pH near the catalyst
layer increases as the current density increases………………………………………..……..74
5.1 Hot carrier collection across an interfacial Schottky barrier at metal/semiconductor
heterojunctions. (a) Qualitative energy band diagram of a plasmonic metal (e.g. Au) in physical
contact with an n-type semiconductor (e.g. TiO2), depicting the conduction band edge (ECB),
valence band edge (EVB), band gap (EG), Fermi level (EF), and the interfacial Schottky barrier
(ΦB). Plasmon excitation creates hot electrons (red) and hot holes (blue) above and below the
EF of Au, respectively, with a distribution of energies governed by the metal band structure and
the incident photon energy (hv = 2.4 eV). Only those hot electrons with sufficient energies
above ΦB (indicated by dashed line) can surmount the interfacial barrier and populate available
CB levels of the n-type semiconductor support. (b) Qualitative energy band diagram of a
plasmonic metal (e.g. Au) in physical contact with a p-type semiconductor (e.g. p-GaN),
depicting ECB, EVB, EG, EF, and ΦB. Plasmon excitation creates hot electrons (red) and hot holes
(blue) above and below the EF of Au, respectively. Only those hot holes with sufficient energies
below ΦB (indicated by dashed line) can surmount the interfacial barrier and populate available
VB levels of the p-type semiconductor support………………………………………..…….80
5.2 Optical properties of p-type GaN (p-GaN) substrate. (a) Absorption of p-GaN substrate,
demonstrating strong absorption in the UV region with no significant features across the visible
regime. Thus, any visible-light features observed from the Au/p-GaN system can be attributed
to the surface plasmon resonance of the Au nanoparticles. (b) Tauc plot of p-GaN indicates an
optical band gap of EG = 3.35 eV, consistent with the expected EG of 3.4 eV for
GaN…………………………………………………………………………….………….83
5.3 Plasmonic Au/p-GaN photocathode device structure. (a) Schematic of Au/p-GaN
photocathode on sapphire substrate. (b) SEM image with corresponding size-distribution
histogram of Au nanoparticles (mean diameter, d = 8.2 ± 1.6 nm) on p-GaN substrate. (c)
Absorption spectra of plasmonic Au/p-GaN photocathode (red curve) compared to bare pGaN substrate (blue curve). The plasmonic device shows a prominent surface plasmon
resonance feature due to the Au nanoparticles at ca. 570 nm. Inset shows a digital photograph
of the colorless p-GaN substrate and the purple Au/p-GaN device. (d) Solid-state currentvoltage (I-Eappl) behavior from Au/p-GaN heterostructures. Ohmic contact to the p-GaN
substrate was achieved through deposition of a thin-film Au/Ni alloy (top panel). In contrast, a
metal-semiconductor Schottky diode was obtained across the Au/p-GaN heterojunction
(bottom panel). Fitting of these data yields a Schottky barrier height of ΦB = 1.1 eV across the
Au/p-GaN interface………………………………………………………………………..83
5.4 Mott-Schottky plot of electrochemical impedance data obtained from bare p-GaN
photocathodes obtained at 2 kHz. The negative slope confirms the p-type character of the
GaN substrates used herein. From a linear fit of the data we obtain a carrier concentration of
ca. 1 x 1019 cm-3, similar to the acceptor doping level of NA = 3-7 x 1018 cm-3 specified by the
manufacturer, and a flat-band potential (Efb) of ca. 2.0 VRHE (V vs. RHE)….………………..85

xviii
5.5 Photoelectrochemical characterization of plasmonic Au/p-GaN photocathodes. (a) Linear
sweep voltammetry of plasmonic Au/p-GaN (red) and bare p-GaN (blue) photocathodes at
10 mV s-1 under periodic (0.5 Hz), visible-light irradiation (λ > 495 nm) at an incident power
of I0 = 600 mW cm-2. (b) Chronoamperometry of plasmonic Au/p-GaN (red) and bare pGaN (blue) photocathodes under periodic (0.5 Hz), visible-light irradiation (λ > 495 nm) while
poised at a fixed applied potential of −0.4 VRHE. (c) Chronopotentiometry of the open-circuit
voltage (Voc) from plasmonic Au/p-GaN photocathodes under visible-light irradiation. (d)
Incident photon-to-charge conversion efficiency (IPCE) of plasmonic Au/p-GaN (red) and
bare p-GaN (blue) photocathodes immersed in 50 mM K2CO3 electrolyte while held at a fixed
potential of −0.4 VRHE. The absorption spectra of each device are plotted with the IPCE
spectra to aid comparison between photoelectrochemical performance & light absorption…86
5.6 (a) Cyclic voltammograms of plasmonic Au/p-GaN and bare p-GaN photocathodes under
dark conditions (black and grey curves) and visible-light (λ > 495 nm) irradiation at I0 = 500 mW
cm-2 (red and blue curves). While the plasmonic Au/p-GaN device exhibits an obvious light
response (red curve), no difference in current was observed for the bare p-GaN device (blue
curve). (b) Close-up view of the cyclic voltammograms from bare p-GaN photocathode under
dark (grey) and visible-light (λ > 495 nm) irradiation (blue) at I0 = 500 mW cm-2. No difference
could be observed between dark (grey) and light (blue) conditions, as these two curves lay
directly on top of one another, confirming that the p-GaN support does not respond to visible
light. This observation is consistent with the large band gap of p-GaN (see Figure 5.2).
Therefore, all visible-light responses observed from plasmonic Au/p-GaN photocathodes can
be unambiguously assigned to hot-hole injection from Au to the valence band of p-GaN upon
plasmon excitation………………………………………………………………………….86
5.7 Photocurrent (Jph = Jlight – Jdark) response obtained from plasmonic Au/p-GaN
photocathodes showing a linear trend with respect to incident light power (I0)……… …….87
5.8 Chronopotentiometry of the open-circuit voltage (Voc) from bare p-GaN photocathodes
under UV-light irradiation. The positive shift in Voc upon UV light exposure confirms the ptype character of the GaN substrates used herein…………………………………..……...88
5.9 (a) X-ray diffraction pattern from 20 nm-thick NiO films on FTO glass showing the
characteristic (200) peak of NiO. (b) X-ray photoelectron spectroscopy spectrum of the Ni 2p
region, showing the characteristic binding energies of NiO. (c) Tauc plot of the NiO film
exhibiting a band gap of ca. 3.7 eV. (d) Mott-Schottky plot obtained from 20 nm-thick NiO
films on FTO glass substrate, which shows a negative slope indicative of p-type conductivity.
From these data, the flat-band potential (Efb) is estimated to be ca. 0.75 VRHE (Volts vs. RHE)
with an acceptor concentration of ca. 1 x 1019 cm-3. All these data are consistent with previous
literature reports of p-type NiO thin films.2 (e) Scanning electron microscopy image of Au
nanoparticles uniformly decorated on the p-NiO surface with corresponding size-distribution
histogram of the Au nanoparticles, with an average Au diameter of 10 ± 1 nm. (f) Absorbance
of the bare p-NiO photocathode (grey) and the plasmonic Au/p-NiO photocathode (black).
A prominent surface plasmon resonance feature due to the Au nanoparticles is observed
around 560 nm. Inset shows a digital image of the FTO glass substrate, p-NiO/FTO substrate,
and Au/p-NiO/FTO substrate, from left to right. A faint purple color is observed from the
Au/p-NiO device……………………………………………..…………………………..90

xix
5.10 Photoelectrochemistry of plasmonic Au/p-NiO photocathodes. (a) Solid-state currentvoltage (I-Eappl) behavior from Au/p-NiO films exhibiting Ohmic behavior, consistent with
previous literature for Au/NiO contacts. (b) Chronoamperometry from Au/p-NiO
photocathodes (black) under visible-light excitation (λ > 495 nm) at 500 mW cm-2 while poised
at −0.4 VRHE. A prompt, reproducible plasmonic photocurrent is clearly observed, indicating
hot-hole collection by the p-NiO support upon plasmon excitation. For comparison, the bare
p-NiO film (grey) exhibits no observable photocurrent. (c) Photoelectrochemical action
spectrum obtained from Au/p-NiO device (black points), showing a clear relationship with
the surface plasmon resonance of the Au nanoparticles (black curve)………………………91
5.11 Time-course of gas evolution from bare p-GaN photocathode under dark electrolysis
conditions at −1.8 VRHE in CO2-saturated 50 mM K2CO3 electrolyte……………..………..92
5.12 Photoelectrochemical CO2 reduction with plasmonic Au/p-GaN photocathodes. (a)
Time-course of gas evolution from plasmonic Au/p-GaN photocathode during controlled
potential electrolysis under dark conditions. (b) Time-course of gas evolution from plasmonic
Au/p-GaN photocathode during controlled potential electrolysis under plasmon excitation (λ
> 495 nm). All electrolysis experiments were performed at −1.8 VRHE in CO2-saturated 50 mM
K2CO3 electrolyte without sacrificial reagents……………………………..……………….92
6.1 Plasmonic Cu/p-NiO photocathode device structure. (a) Schematic of Cu/p-NiO
photocathode on fluorine-doped tin oxide (FTO) glass showing the approximate dimensions
of the Cu nanoparticles (~8 nm in diameter) and the p-NiO layer (~60 nm thick) on the FTO
glass substrate. (b) Quantitative energy level diagram showing the relative positions of the pNiO valence band (EVB) and conduction band (ECB) relative to the Cu Fermi level (EF). The
difference in energy between the p-NiO valence band and the Cu Fermi level is expected to
allow the formation of an interfacial Schottky barrier (ΦB) to hot hole injection at the Cu/pNiO interface of around 1 eV. Photoexcitation of Cu nanoparticles with photon energy (hv)
below the band gap (EG) of the p-NiO support generates hot electrons and hot holes on the
Cu surface. The p-NiO support facilitates charge separation across the metal-semiconductor
interface by allowing the collection of hot holes from the metal while also confining the hot
electrons on the Cu surface to drive CO2 reduction (inset)…………………………………99
6.2 Materials characterization of p-type NiO films. (a) X-ray diffraction pattern from NiO film
on FTO glass showing the characteristic (200) peak of NiO. All other peaks can be attributed
to the underlying FTO substrate. (b) X-ray photoelectron spectroscopy high-resolution scan
of the Ni 2p region, showing the characteristic binding energies and satellite features of NiO.
(c) Mott-Schottky plot obtained from NiO films on FTO glass substrate, which shows a
negative slope indicative of p-type conductivity. From a linear fit of these data, the flat-band
potential (Efb) is estimated to be ca. 0.75 VRHE (Volts vs. RHE) with an acceptor concentration
of ca. 1 x 1019 cm-3. (d) Tauc plot of the NiO film showing a band gap of around 3.7 eV. All
these data indicate material properties consistent with previous literature reports of p-type
NiO thin films...................................................................................................................................100
6.3 Materials characterization of the plasmonic Cu/p-NiO photocathode. (a) SEM image with
corresponding size-distribution histogram of Cu nanoparticles (mean diameter, d = 8 ± 2 nm)
on a 60 nm thick p-NiO film supported on FTO glass. (b) X-ray photoelectron spectroscopy
high-resolution scan of the Cu 2p region from as-synthesized Cu/p-NiO photocathodes. (c)
Absorption spectra of the plasmonic Cu/p-NiO photocathode before (yellow curve) and after

xx
(red curve) electrochemical reduction via three successive cyclic voltammetry scans. The
spectrum of the bare p-NiO film (blue curve) is also shown for comparison. (d) Cyclic
voltammograms from plasmonic Cu/p-NiO photocathode (yellow to red curves) and bare pNiO films (blue curve) at a scan rate of 50 mV s-1. Black arrows indicate the scan direction.
The reduction of Cu oxides into metallic Cu is evidenced by the progressively smaller cathodic
wave around 0.7 VRHE (yellow curve) that eventually disappears after the third successive scan
(red curve). A representative voltammogram from bare p-NiO films (blue curve) is shown for
reference………………………………………………………………..………………..102
6.4 Cyclic voltammetry of bare p-NiO cathodes under dark (black curve) and visible light
(orange curve) showing that bare p-NiO film exhibits no measurable light response across the
potential sweep. This observation is consistent with the large band gap of p-NiO. Therefore,
all visible-light responses observed from plasmonic Cu/p-NiO photocathodes can be
unambiguously assigned to hot-hole injection from Cu to the valence band of p-NiO upon
optical excitation of the Cu nanoparticles…………………………..…………………….103
6.5 Photoelectrochemical characterization of plasmonic Cu/p-NiO photocathodes. (a) Linear
sweep voltammetry J(E) of plasmonic Cu/p-NiO photocathode at a scan rate of 20 mV s-1
under dark conditions (dotted black curve) and under visible-light irradiation (λ = 560 ± 40
nm) (solid red curve). (b) Chronoamperometry J(t) of the photocurrent (Jph = Jlight – Jdark)
obtained from plasmonic Cu/p-NiO (red) and bare p-NiO (blue) photocathodes under
periodic, visible-light irradiation (λ = 560 ± 40 nm) while potentiostatically poised at an applied
potential of E = −0.2 VRHE. (c) Power-dependence of the photocurrent Jph(I0) obtained from
the plasmonic Cu/p-NiO photocathode. (d) Chronopotentiometry V(t) of the open-circuit
voltage (Voc) obtained from the plasmonic Cu/p-NiO photocathode (red curve) and the bare
p-NiO cathode (blue curve) under visible-light irradiation (λ = 560 ± 40
nm)……………………………………………………………………………………....105
6.6 CO2 reduction with bare p-NiO cathodes under dark conditions. (a-b) Faradaic efficiency
for H2 (squares), CO (circles), and HCOO− (triangles) with (c-d) corresponding partial current
density J for the hydrogen evolution reaction (JHER, squares), carbon monoxide (JCO, circles), and
formate (JHCOO-, triangles). Electrolysis was performed under dark conditions in a CO2-saturated
50 mM K2CO3 electrolyte. The device was held potentiostatically at each applied potential for
2 hours while the gas products were sampled every 15 minutes and analyzed by gas
chromatography. Liquid products were collected and analyzed by HPLC at the end of each run.
Each data point represents the average of three independent trials and the error bar indicates
the standard deviation……………………………………………………………………..106
6.7 Comparison of the partial current densities (J) obtained from bare p-NiO (open data
points) and Cu/p-NiO cathodes (filled data points) under dark conditions. (a) Partial current
density for the hydrogen evolution reaction (JHER) from p-NiO (open squares) relative to
Cu/p-NiO (filled squares). (b) Partial current density for carbon monoxide (JCO) from p-NiO
(open circles) relative to Cu/p-NiO (filled circles). (c) Partial current density for formate
(JHCOO-) from p-NiO (open triangles) relative to Cu/p-NiO (filled triangles). Each data point
represents the average of three independent trials and the error bar indicates the standard
deviation. We observed a significant increase in the J for CO2 reduction products CO and
HCOO− with the addition of Cu nanoparticles, while almost no change in the amount of H2

xxi
that was evolved. We therefore attribute the significant amount of H2 that is evolved from
the plasmonic Cu/p-NiO device to the activity of the underlying p-NiO film, which almost
exclusively produces H2 under CO2 reduction conditions…………………………………107
6.8 Distribution of CO2 reduction products obtained from plasmonic Cu/p-NiO
photocathodes as a function of the applied electrochemical potential (E). Faradaic efficiency
(a–c) and associated partial current density (d–f) for the production of (a,d) hydrogen (H2)
(squares), (b,e) carbon monoxide (CO) (circles), and (c,f) formate (HCOO−) (triangles) during
controlled potential electrolysis under dark conditions (blue symbols) and under visible-light
irradiation (yellow symbols). Plasmon excitation was performed with λ = 560 ± 40 nm at an
incident power of 160 mW cm-2. Data points and error bars represent the average value and
standard deviation, respectively, obtained from three independent trials…………….……108
7.1 Schematic representation of the steps for various prototype systems designed to capture
CO2 and/or convert it to either concentrated and pressurized CO2 or to a value-added
product. The blue arrows represent prototype processes that capture and convert CO2, grey
arrows represent prototype processes that only focus on CO2 conversion, and the pink arrow
represents the process that we propose. The numbers in brackets correspond to references
for the various processes [5-13,15,18]……………………………………………………120
7.2 Synopsis of the various steps involved in capturing CO2 and transforming it into valuable
chemicals……………………………………………………………………………...…121
7.3 Schematic representation of processes and energy requirements for various proposed
schemes that capture CO2 and transform it to value-added products. The bottom rows show
the commercial syngas synthesis process, with either a feed stock of coal or of natural gas. The
energy needed to produce 1 mol of CO while related to the cost needed to operate the plant
does not encompass the full picture of expenses such as materials, maintenance, and
labor………………………………………………………………………...……………124
7.4 Schematic representation of processes and energy requirements for various proposed
schemes that capture CO2 and transform into formic acid………………………………….125
8.1 Schematic of various pathways to capture CO2, generate H2, and generate CH4 from sunlight,
H2O, and sunlight……………………………………………………………………….....130
8.2 Schematic representation of various technology pathways for sustainable generation of
methane from sunlight, water, and carbon dioxide…………………….………………….132
8.3 The cost breakdown of the green methane from thermochemical and biochemical
processes. The feedstock of the thermochemical and biochemical process assumed water
from utility, CO2 from direct air capture, and H2 from LTE with an electricity price of
$49/MWh………………………………………………………..………………………141
8.4 Cost of methane from (a) photoelectrochemical (PEC) and (b) electrochemical methanation
processes as a function of key performance metrics in those technologies. (a) The cost of PEC
methanation as a function of the STF conversion efficiency and the cost per area of PEC
material. (b) The cost of electrochemical methanation as a function of the energy efficiency of

xxii
the device and the operating current density. The green region in both plots shows where the
cost is equal to or less than thermochemical or biochemical methanation. For all systems
compared the cost of CO2 is assumed to be $278/ton and the electricity price is $49/MWh…142
8.5 Cost of the (a) thermochemical and (b) biochemical methanation processes as a function of
H2 cost and CO2 cost. (c) A side-by-side comparison between the biochemical methanation
process and the thermochemical methanation process………………………………….…144
9.1 shows an SEM and a schematic of a nanoporous gold catalyst. On the left is a cross sectional
SEM of a 25% gold nanoporous gold sample. On the right is a schematic of the nanoporous
gold during electroreduction conditions. The red curve represents the pH in solution as a
function of distance from the electrode surface. The solid part of the red curve has been
calculated based on experimental conditions and the dashed portion of the red line is an assumed
pH inside of the nanoporous gold…………………………………………………………150
9.2 shows a schematic of the confocal fluorescent microscopy experimental set up. On the left
we see the water immersion objective scanning the surface of a CO2 reduction gas diffusion
electrode. On the right is an example of a pH map; for more details see Chapter 4………..151
9.3 shows a schematic of the illuminated gold nanoparticles on p-type GaN. The circle on the
right shows the band diagram of the p-GaN and the Au nanoparticles, showing the energy that
the hot holes need to be injected into the semiconductor………………………….………152
9.4 shows a schematic of illuminated copper nanoparticles on p-type NiO…………………153

xxiii

LIST OF TABLES

1.1 List of relevant electrochemical reactions and their corresponding
equilibrium potentials……………………………………………………..…3
1.2 List of techniques for measuring the local pH. The table also shows relevant
metrics to compare techniques, such as temporal resolution, spatial resolution,
and pH range………...……………………………………………….…….11
4.1 Model parameters………………………………………………………71
8.1: Summary of cost, TRLs, and demonstrated scale of different technological
pathways for renewable generation of methane. The cost of methane generation
in the thermochemical and biochemical routes assumed water from utility, CO2
from
direct
air
capture,
and
H2
from
LTE
as
the
feedstock…………………..……………………………………………...132

Chapter 1

INTRODUTION TO CO2 REDUCTION
1.1

Role of solar fuels in the mitigation of climate change
Climate change is perhaps the largest threat facing our planet today, due to the catastrophic

consequences, and the global coordinated effort it will take to mitigate the problem.1 For every additional 0.5°C
increase in global temperature there will be a discernible increase in heat waves, heavy precipitation, sea level
rise, extreme weather events, and drought. In order to mitigate the worst of these effects we must stay below a
1.5°C change in global temperature. Models have shown that this will only be possible if go to negative
greenhouse gas emissions by 2050.2 To achieve this goal, renewable energy needs to become a dominant source
of energy globally.3 Currently, the two largest forms of renewable energy are solar photovoltaics and wind, due
to the many options for geographical deployment and the large availability of both resources.4 However, despite
the abundance of these resources, solar and wind energy both experience daily, seasonal, and geographic
variation, thus suggesting the need for efficient and inexpensive energy storage to fully replace carbon-based
fuels.5 Batteries, while useful for grid power management and short term storage, are unlikely to be able to
provide necessary seasonal storage due to prohibitive costs and charge leakage. Other storage technologies exist
– such as pumped hydro, compressed air, and fly wheels – but they also have limitations including scalability,
versatility, and storage time.6,7 In addition to their inability to handle seasonal storage, these technologies are not
nearly as energy dense as chemical fuels, and are therefore unlikely to be able to power applications such as
intercontinental flights, long-haul shipping, or orbital rockets.8
The transformation of renewable energy into chemical bonds can solve both of these challenges because
it provides an energy dense and long-term storage solution. In particular the transformation of carbon dioxide
(CO2) serves to store renewable energy as chemical fuels, which are carbon neutral. This is important because
there are many industries that are difficult to decarbonize,3† such as the production of cement,9,10 steel,11 plastic,12
ammonia,13 and aluminum,14,15 which account for 12% of total carbon emissions. By removing CO2 from the
atmosphere we can offset the emissions from these industries. Transforming CO2 into valuable chemicals allows
us to close the carbon cycle and stop the emission of greenhouse gases providing a long term, energy dense
storage solution that also removes CO2 from the atmosphere.

† Decarbonization means to remove the emission of greenhouse gases from a process.

The following steps are most commonly followed in order for CO2 to be captured and transformed: (i) a
CO2 source, (ii) a capture medium, (iii) a process to release CO2 from the capture medium, (iv) CO2 compression
into a concentrated gas stream, and (v) conversion of CO2 into fuels, chemicals, and/or materials (e.g.,
hydrocarbons).16 In this thesis we focus on the conversion portion (step 5) of this process because this requires
the most energy and is least understood. In the final chapters we perform technoeconomic analysis (TEA) on
the whole system to see what how a CO2 capture/conversion system could be deployed.
ACS Energy Letters

Viewpoint

Figure
Synopsis
varioussteps
stepsinvolved
involvedinincapturing
capture CO
and transforming
transformingititinto
intovaluable
valuablechemicals.
chemicals16
Figure 2. 1.1:
Synopsis
of theofvarious
CO22 and
1.2

Fundamentals of electrochemical CO2 reduction
The cathodic reaction in electrochemical CO2 reduction (CO2R) follows this form:
𝑥𝐶𝑂! + 𝑛𝐻" + 𝑛𝑒 # → 𝐶$ 𝐻% 𝑂& + 𝑚𝐻! 𝑂

(1.1)

The anodic reaction is the oxygen evolution reaction (OER) and is as follows:
2𝐻! 𝑂 → 𝑂! + 4𝐻" + 4𝑒 #

(1.2)

While other anodic reactions are possible, water is the only viable source of electrons and protons if CO2R is to
be scaled.17
Through this section CO2R reaction will be discussed in terms of thermodynamics. It will also be
discussed the cause of the challenges of doing CO2R instead of the hydrogen evolution reaction (HER). The
following chapters will address how the catalyst and local catalyst environments can influence which reaction
Figure 3. Schematic
of processes and energy
requirements
for various proposed
that capture
CO2 andhowever
transformthe
it to
pathway
occurs.representation
The first electrochemical
studies
of CO2 reduction
began inschemes
the early
19th century

value-added products. The bottom rows show the commercial syngas synthesis process, with a feed stock of either coal or natural gas. The
energy needed to produce 1 mole of CO while related to the cost needed to operate the plant does not encompass the full picture of
expenses such as materials, maintenance, and labor.

captures CO2 from a geothermal power plant, then pressurizes
and heats the CO2 in the presence of H2 generated via water
electrolysis to create methanol and water (Carbon Recycling
International). This process requires 1.53 MJ/mol CO2

CO2, respectively (Supporting Information, section 1E).
Finally, an approach for carbon sequestration uses a process
that injects CO2 into cement during production, allowing
cement producers to use less binder and offset their CO2

first more systematic study was done by Hori is 1985.18,19 He explored a variety of different metals, CO2
pressures, etc. and documented how they effected the onset potential and Faradaic efficiency towards different
products. Table 1.1 shows common CO2R reactions with their equilibrium potentials and names of the
products, similar such tables have been shown in many previous CO2R reviews.18,20–24 The CO2R standard
potentials are calculated via the Gibbs free energy equation from data provided by NIST.25 For gas phase
products gas-phase thermochemistry data is used while for aqueous products Henry’s law data is used. All
equilibrium potentials are reported vs the reversible hydrogen electrode (RHE) which is independent of pH.
CO2 is always considered a gas and H2O a liquid in Table 1.1.
CO2 Reduction Reactions

E0 (V vs RHE)

Name of Product

→ %!""%(*+)

-0.12

Formic acid

!"2

+ 2% + + 2& −

→ !"(-) + %2 "

-0.10

Carbon monoxide

!"2

+ 6% + + 6& −

→ !%3 "%(*+) + %2 "

0.03

Methanol, MeOH

!"2 + 8% + + 8& − → !%4 - + 2%2 "

0.17

Methane

!"2 + 4% + + 4& − → !(2) + 2%2 "

0.21

Graphite

!"2 + 2% + 2&

→ !""% 2(2)

-0.47

Oxalic acid

→ !%3 !HO 78 +3%2 "

0.06

Acetaldehyde

→ !2 %4 - + 4%2 "

0.08

Ethylene

→ !2 %5 "% *+ + 3%2 "

0.09

Ethanol

→ !%3 !""% *+ + 2%2 "

0.11

Acetic acid

2!"2 + 2% + 2&

2!"2 + 10% + 10&
2!"2 + 12% + 12&
2!"2

+ 12% + + 12& −

2!"2

+ 8% + + 8& −

2!":

+ 14% ; + 14& <

→ !: %= > + 4%: "

0.14

Ethane

→ !: %@ !%" AB + 5%: "

0.09

Propionaldehyde

→ !C HD OH 78 + 5%: "

0.10

Propanol, PrOH

3C": + 16% + 16&

3!": + 18% + 18&

Other important reactions
2% ; + 2& < → %:

2%: " → ": + 4% + 4&

E0 (V vs RHE)

Name of Reaction

0.0

Hydrogen evolution reaction (HER)

1.23

Oxygen evolution reaction (OER)

Table 1.1: list of relevant electrochemical reactions and their corresponding equilibrium potentials.
From Table 1.1 we are able to see that CO2R is not only difficult because CO2 is the most stable form
of carbon under environmental conditions, but the variety of reaction pathways and small differences in
equilibrium potentials make selectively producing one product challenging. First, the close spacing between
many of the equilibrium potentials indicates that merely tuning the applied bias will not be able to easily select
for one product. Second, many of the reactions require multiple protons and electrons, and it is difficult to find
a single catalyst that can sustain all of the intermediates for a single pathway. In order for a reactant to be
reduced, it must bind to the electrode surface so that electrons can be transferred to it. In a reaction pathway

with multiple electron transfers each intermediate must also stay bound to the electrode surface, so that all of
the electrons can be transferred. If an intermediate does not stay bound to the surface it will float into the
electrolyte and the final desired product will not be made. Therefore, the more electrons required for a particular
reaction, the more difficult it is to find a catalyst that selectively and actively performs the reaction. Third, the
potential of the hydrogen evolution reaction (HER) is right in the middle of the CO2R reaction potentials. In
an aqueous environment, water has a molarity of 55.5M, whereas the maximum CO2 concentration is 33mM.26
It is a challenge to selectively to do CO2R as opposed to HER when there is so much water available, which also
can be reduced at a similar potential. In this thesis we explore how through a combination of catalyst engineering
and understanding of the local catalyst environment we can overcome these challenges to make highly selective
and active CO2R devices.
1.3

Catalyst compositions
A number of considerations are necessary when choosing a catalyst, the first of which is what products

the metal is capable of making. This is largely defined by the adsorption energy of three important intermediates
H*, COOH*, CO*.27‡ An intermediate is any chemical that exists as a reactant is reduced to the final product.
As discussed previously, for a reactant to be reduced it must adsorb to the electrode surface, allowing for
electrons to transfer. In order for multiple electrons to be transferred each subsequent intermediate must stay
adsorbed to the electrode surface.

‡ The * notation denotes that these chemicals are in the adsorbate form.

Figure 1.2: Adsorption energy of CO* vs the adsorption energy of H*. Marks in red are metals that primarily produce H2, teal
marks produce products with more than two carbons, blue marks produce mainly CO and yellow marks produce predominantly
HCOOH. Reproduced from Bragger et al. Copyright 2017, John Wiley and Sons27
The first adsorption energy to consider is for H* because this defines the metals ability to perform the

1/1

hydrogen evolution reaction (HER). Most transition metals have a high adsorption energy for H* (red marks
in Fig. 1.2) making them very active for HER and therefore not suitable CO2 reduction catalysts.19,28 Late
transition and p-block metals have much positive binding energy for H* making them much worse HER
catalysts29 (Pd, Cu, Ag, Au, Zn, Cd, Hg, In, Tl, Sn, Pb).
The next consideration is the binding energy for CO*. Metals that bind CO* very strongly (red marks
in Fig. 1.2) are poor CO2R catalysts because their active sites are quickly poisoned.30 We say the sites are poisoned
because the CO occupies the location where the CO2 needs to bind to in order to be reduced. On the other
hand, metals that bind CO* weakly (blue and yellow marks in Fig. 1.2) are unable to make higher order carbon
products because the CO is released quickly from the surface. Metals that produce HCOOH as opposed to CO
have a very similar adsorption energy of COOH* and H*, while metals that produce CO have a larger difference
in the adsorption energy of COOH* and H*. The only metal known to date that can efficiently produce multicarbon products is Cu (teal mark in Fig 1.2), and this is likely due to the fact that it is has a medium adsorption
energy for CO* and a positive adsorption energy for H*.
1.4

Nanostructuring and surface facets of catalyst

After choosing what material to use the next important consideration is how to structure the catalyst.
The first consideration is how to structure the catalyst so that the most favorable surface facet is facing the
electrolyte. Different surface facets of metals adsorb intermediates differently due to their different electronic
structure and number of dangling bonds. Therefore, being able to create devices with specific surface faceting
can improve device performance. An important value for understanding and predicting the activity of a
particular surface facet is the free energy (DG). When intermediates are absorbed onto the surface this changes
the DG of that surface. The ideal catalyst would have no change in DG change throughout a reaction.
In the case of CO2 reduction to CO, the Au (111) surface shows a change in DG double that of the Au
(211) surface.31 This implies that if an electrode is made with only Au (211) it will be highly active for CO. In
addition, for Au there is a linear relation between the generalized coordination number of a surface and the
overpotential required for CO2R, with less undercoordinated surfaces requiring lower overpotentials.32 For Cu
the story becomes more complicated due to the number of possible reaction pathways. However, there are still
striking differences in products between different facets. With Cu (100) being least active for HER and most
active for C2H4 production, and Cu (111) being most active for CH4 production.33 The ability to structure a
catalyst and control the exposed surface facets allows you optimize for different products and make the most
efficient devices
Another advantage of nanostructuring the catalyst is that we can increase the activity per gram of material
used. In general, undercoordinated sites are more active CO2R, due to their perturbed electronic structure,
which changes DG.34,35 Therefore nanoparticles become an exciting structure for the catalyst because of the high
density of undercoordinated sites. In the case of Au31,32,36 and Cu35,37–39 it has shown that generally as particles
become smaller CO2R increases.40

Nanoparticles are not the only way to create a high density of

undercoordinated sites, a variety of structures have been explored from nanopores, nanospikes, nanorods,
nanofoam, etc. The advantages of these contiguous nanostructured films are that they do not require a
conductive support and have more active sites relative to the top-down area as compared to a planar film. The
increase in electrochemically active surface area relative to the top-down area allows for increased device
performance. Chapter 2 further explores how nanostructuring gold can improve its selectivity and activity.
1.5

Improved CO2 reduction via plasmonic metal nanoparticle
Light is an additional tool that can be used to enhance the selectively of a catalyst, in particular, plasmonic

metal nanoparticles for CO2 reduction. The given enhancement depends on the specifics of particular system
and it requires detailed experiments to determine which mechanism is dominating, as multiple may occur at

once. These mechanisms include (1) resonant photon scattering, (2) resonant energy transfer, (3) hot electron
transfer, and (4) local photo thermal heating.41–46
To understand how light can change the selectivity of a plasmonic metal nanoparticle it is first important
to understand the photophysics of how upon resonant optical excitation of plasmonic metal nanoparticle
electrons are excited above the Fermi level energy of the metal and then decay back. Before the optical
excitation, all electrons are distributed according to the Fermi-Dirac distribution,47 in which nearly all electrons
are restricted to states below Fermi energy and a single temperature (TF) describes system. Immediately (t = 1100fs) after optical excitation electrons are promoted above the Fermi level into an electron gas. At this point,
these electrons are referred to as ‘hot’ because their electronic temperature has not equilibrated with the lattice.41
These hot electrons when they are promoted above the Fermi level leave behind hot holes, deep vacancies in
the electronic structure. These hot electrons and hot holes then decay (t = 100 fs -1ps) to a Fermi-Dirac like
distribution with a higher electronic temperature than the lattice.48–50 The hot carriers then continue to decay
until they relax to the original electron distribution and the heat is transferred throughout the lattice. The goal
of plasmonic photocatalysis is to harness the energy of these carriers before they relax back to the ground state
and to use this harvested energy for improved activity and selectivity of the catalyst.
The time scale at which these hot carriers decay is much faster than that of the time scale of chemical
reactions, picoseconds vs microseconds, as indicated by the photophysics. It is therefore useful to create a
plasmonic metal semiconductor junction to separate hot electrons and hot holes, thereby prolonging their
lifetime. Upon optical excitation the plasmonic metal hot electrons are lifted above the Fermi energy. A fraction
of the hot carriers moves toward the metal/semiconductor interface. Hot electrons with energy above the
valence band of the semiconductor or hot holes with energy below the valence band have a probability of being
transferred into the semiconductor (there is still a chance that they be reflected.)47,50 In the case of CO2 reduction
we usually use a p-type semiconductor so that it can harvest the hot holes out of the plasmonic metal. This
leaves behind hot electrons that can then be transferred into the adsorbate. Due to the unique energy of these
electrons, it can affect the selectivity of the reaction.41,50–52 In this thesis we explore how plasmonic Au
nanoparticles on p-type GaN53 and plasmonic Cu nanoparticles on p-type nickel oxide54 can enhance the CO2
reduction reaction.
1.6

Local pH effects
Local pH, the measure of the proton concentration, is a critical parameter to device performance, and

yet there currently exists minimal experimental understanding. While significant research has been put into the

understanding of the catalyst,18,21,55 there has been less effort to understand the local pH despite its ability to alter
reaction selectivity and activity. In this section we will review how pH is influenced by and influences CO2
concentration, activity of HER and CO2R, and selectivity of CO2R. It will also be discussed what techniques
can be used to measure the local pH.
1.6.1

Electrolyte concentration and relevant species
There are two different pH values to consider in CO2 reduction the pH at the catalyst surface and the

bulk pH. Before any electrochemical reactions happen the pH locally and in the bulk are the same and defined
by the electrolyte concentration and CO2 saturation. Figure 1.3 shows the concentration of various species in a
KHCO3/K2CO3 electrolyte as a function of pH. The values for these concentrations can be found by looking
at the rate and equilibrium constants for the following chemical reactions relating CO2, HCO3-, CO32-, H2O, H+,
and OH-. The first species that we consider is CO2 because its concentration determines whether the electrode
will be limited by its own activity or mass transport limitations. The amount of CO2 dissolved in the electrolyte
on pressure, temperature, and salinity of the electrolyte. The equilibrium between the gas and liquid phase of
CO2 is given by Henry’s constant (K0):
(!"#

𝐾' = 3)

!"#

(1.3)
*''

ln(K ' ) = 93.4517 < + = − 60.2409 + 23.3585 ln <*''= + S 30.023517 − 0.023656 <*''= +

0.0047036 <*''= 4

(1.4)

where 𝑐,-# is the concentration of dissolved CO2, 𝑓,-# is the gas phase fugacity of CO2, T is temperature in
Kelvin, and S is salinity defined by the amount of salt in grams in 1kg of water (given in parts per thousand).
This shows that the concentration of CO2 decreases with increasing salt concentration.
The dissolved CO2 dissociates at pH greater that 5 to bicarbonate and carbonate according to (1.5) and
1.6). The forward rate of reaction of (1.5) and (1.6) are 3.71*10-2 s-1 and 59.44 s-1 respectively.
𝐶𝑂!(/0) + 𝐻! 𝑂 ⇌ 𝐻" + 𝐻𝐶𝑂2# ,
𝐻𝐶𝑂2# ⇌ 𝐻" + 𝐶𝑂2!# ,

𝑝𝐾 = 6.37

𝑝𝐾 = 10.25

(1.5)
(1.6)

The final reaction of note is for the dissociation of water. The forward rate of reaction of (1.7) is 2.4*10-5
mol/(L*s).
𝐻! 𝑂 ⇌ 𝐻" + 𝑂𝐻# ,

𝑝𝐾/ = 14

(1.7)

Figure 1.3: Concentration of carbon dioxide, hydrogen, hydroxyl, bicarbonate, carbonate, and potassium ions as a function of bulk
pH in a potassium (bi)carbonate electrolyte at 25°C and a pressure of 1atm. Reproduced from Singh, et al. Copyright 2015,
Royal Society of Chemistry.26
Using equations (1.3)-(1.7), Fig. 1.3 shows that to maximize the amount of dissolved CO2 in an aqueous
environment the pH must be kept below 9.2. However, ideally the reaction would occur as close to this pH as
possible so that the number of protons can be reduced to suppress HER. We can also see that while the
concentration of dissolved CO2 is maximized below pH 9.2, the concentration never goes above 33mM, showing
that in aqueous environments the reaction will be mass transport limited at current densities above roughly 20
mA/cm2. To overcome this limitation devices have been designed that deliver CO2 in the gas phase, which is
discussed further is section 1.7. This initial look at CO2, HCO3-, CO32-, H2O, H+, and OH- in bulk electrolyte is
useful, however once the current density is non-zero at the electrode surface H+, OH-, and CO2 begin to be
consumed and/or created at the electrode surface. This creates a boundary layer between the electrode surface
and the bulk pH. The size of this boundary layer depends on the buffering capability of the electrolyte and the
magnitude of the current density. Typical boundary layers for aqueous CO2 reduction tend to be between 10-

10
100 µm.56 This means that the pH at the electrode surface can be quite different than that of the bulk. It is
critical that we know what the pH is at the surface of the electrode because that we ultimately affect the reaction.
Chapter 3 focuses exclusively on measuring the local pH of a CO2 reduction electrode.
1.6.2

Dependence of HER and CO2 reduction on pH
Both the HER and CO2R require protons to complete the reaction (see Table 1.1). However, there are

different mass transport requirements relative to protons for both HER and CO2R. This has been explored by
Hall, et al. on a Au rotating disk electrode.57 In the experiment a polished polycrystalline is rotated at various
speeds, and the rotation speed is proportional to the convective flow of reagents to the electrode surface.58 As
the Au electrode rotation speed is increased from 625 rpm to 3500 rpm, the current density for CO2R is
unchanged but the HER current density increases by ~22%.57 This shows that CO2R remains relatively
unaffected by limited protons, whereas limited protons significantly suppresses HER. From this it becomes
clear that by controlling the pH near the surface of the electrode, HER is suppressed. However, it is important
to keep in mind that if the pH is too high that the concentration of CO2 will decrease (Fig 2.2) and this will limit
the current density for CO2R. That is why it is critical that change in pH only occur locally so that the CO2 can
still be there in high concentration, which is only possible in a non-equilibrium state.
Another important consideration is that while the pH does not affect the total current density for CO2R,
it does impact which reactions occur. There have been numerous reports59–64 of how the pH impacts what
products are made, as will be discussed in later chapters of this thesis. One example shows that as the pH at the
surface of the electrode increases the Faradaic efficiency (FE) for HER decreases.59 They show this in an
aqueous system on a Ag foam electrode using in situ Raman spectroscopy. We also see that on microporous
Au electrodes that the FE for CO goes up and HER decreases as the pore depth increases.57 This suggests that
as the pores’ depth increase, it becomes more difficult for protons to diffuse in, thus allowing for increased pH
at the same current density and therefore creating more favorable conditions for CO2R.
1.6.3

Techniques for the measurement of local pH

11

Table 1.2: List of techniques for measuring the local pH. The table also shows relevant metrics to compare techniques, such as
temporal resolution, spatial resolution, and pH range.
There are many techniques that can be used to measure the local pH (Table 1.2) each with their own
advantages and disadvantages.65 In this thesis (Chapter 4) we focus on using fluorescent confocal microscopy.
This technique was chosen because it is able to provide spatial resolution in x, y and z. It can also be performed
in situ with only small perturbations to the electrochemical reaction allowing it to be a platform to study many
different catalysts. Finally, it can be used over a wide range of pH values by selecting the appropriate dye.
Chapter 3 provides greater detail of the specifics of the device and the pH maps of CO2 reduction electrodes.
1.7

Device components and CO2 delivery method
Another consideration for device performance is the design of the device itself. There are many

considerations to take into account when designing the device such as distance between electrodes, CO2 delivery
method, membrane, reference electrode, light, etc. While there are generic guidelines for device configuration,
much depends on the purpose of the device (e.g., whether the device uses a photoanode or needs to be
compatible with a confocal microscope). This section goes through the device components, why different
configurations are chosen, and implications of different device design choice. Fig. 1.4 shows a schematic of an
electrochemical cell. Electrons flow to the working electrode, cathode, to reduce CO2 to hydrocarbons. The

12
voltage of the working electrode is measured by a reference electrode, usually a Ag/AgCl, which has a stable and
well known electrode potential. It is important that the reference electrode is as close as possible to the working
to minimize a solution resistance between the two. The reference electrode is needed because it is not possible
to measure the current and voltage between the working and counter electrodes at the same time; introducing a
reference electrode solves this challenge.
A membrane is placed between the cathode and anode to prevent any reduced liquid products from
being oxidized as well as to stop aqueous O2 from poisoning the working electrode. There are several types of
membranes used, cation exchange membrane (CEM), anion exchange membranes (AEM), and bi-polar
membranes (BPM). CEMs allows for cations to flow freely through and AEMs allow for anions to move freely
through. BPMs have a CEM and AEM layer, which causes water to dissociate and protons to be on the CEM
side and hydroxyls to be on the AEM side. This enables solutions of different pH to be maintained on either
side of membrane, however it increases the cell resistance. At the counter electrode, anode, water is oxidized to
oxygen, equation (1.2). For all of the experiments performed in this thesis a platinum (Pt) foil or mesh was used
for the anode due to its stability and low overpotential. However, Pt is expensive, so in a practical device a
nickel (Ni) anode would be a more appropriate choice.
e-

← HCO3H2O
HC

CO32- →
← H+

CO32- →
←H

OH- →

BL

HCO3- →

OH- →

2H+
½ O2

Counter Electrode

H2O

Membrane

CO2
H+

Reference Electrode

Working Electrode

K+

BL

BL

BL

Figure 1.4: Schematic of a CO2 reduction electrochemical cell. Electrons flow to the working electrode, cathode, to reduce CO2 to
hydrocarbons (HC), and electrons are removed from H2O to create O2 at the counter electrode, anode. The dotted regions indicate

13
boundary layers (BL). Arrows indicate which way different species migrate through the BL. The reference electrode is used to
measure the voltage being applied to the working electrode and the current is measured between the cathode and anode.
When combining these various components there are several important factors to consider —

Figure 1. Different electrochemical CO R reactor schemes. (a) Aqueous-phase CO
Energy Letters
Perspective
employing an (b) aqueou
and of
then
reduced
at a catalyst
surface.
CO R repeatability.
minimization of cell resistance, uniformity
voltage
across
electrodes,
andVapor-fed
experimental

Minimization of cell resistance can be accomplished by placing the anode and cathode as close together as

Moving Membrane
toward choice
practical
possible without running into mass transport limitations.
can alsoreactor
impact thedecell resistance,
signs that operate using CO2 delivered
with thinner membranes are generally more conductive. It is also important for the voltage across the cathode
to the cathode in
the vapor phase can
26
to be uniform, which can be achieved via a parallel
plate
configuration.
This
is notperformance
always possible, e.g., when
help to overcome these
66
using a photocathode, causing the anode and cathode
to be offset. challenges.
The need for a parallel plate configuration
and solubility

and to have reproducibility between experiments, has led to the use of compression cells.26 This allows the
along with acid/base buffer (CO2/HCO3−/CO3−2) equilibria
anode, cathode, and reference electrode to be placed in the exact same position
each time and in the optimal
lead to intrinsic challenges toward achieving high conversion
20
rates and
energy
practical
reactordue
configuration. Despite the cost, the best material
to make
thefficiencies.
cell out of is Moving
polyethertoward
ether ketone
(PEEK)
designs that operate using CO2 delivered to the cathode in the
to its hardness and chemical stability. The hardness is crucial so that plastic fittings can be screwed into it without
vapor phase (Figure 1b,c) can help to overcome these
performance
andcell
solubility
Such gas-diffusion
damaging the threads, as well as to prevent
the risk of the
cracking challenges.
under compression.
The chemical
electrodes (GDEs) can achieve this by employing a porous
stability also means that the cell can be thoroughly
in nitric
aciddiffusion
and/or aqua
regia
reducing
the risk
catalystcleaned
layer along
with
media
to thus
facilitate
reactant
transport
and
GDEs
beenappear
used obvious,
in otherthey
of contamination between experiments. While
many of
thedistribution.
considerations
listed have
here may
electrochemical energy-conversion devices such as fuel cells
are critical in being able to obtain consistentand
results
and to be able
to reproduce
results in
electrolyzers,
where
the architecture
hasliterature.
been optimized
for
high
current
density
and
low
transport
losses.
However, the
sEnergy
Perspective
Letters
Perspective
ergy
Letters
Perspective
adaptation to CO2R will require further advancement,
as
different operating strategies and understandings are needed to

address p
to avoid
Furtherm
form a
ideally an
species
electroly
A rece
various e
a review
years on
Perspect
GDE de
context
discussio
opportun
translate
purpose
review o
developm
perspecti
research

re 1. Different electrochemical CO2R reactor schemes. (a) Aqueous-phase CO2R, where CO2 is first solubilized in an aqueous electrolyte
then reduced at a catalyst surface. Vapor-fed CO2R employing an (b) aqueous or (c) polymer electrolyte.

address product selectivity considerations, which is important
Moving toward practical reactor de- carbon monoxide,
to avoid andtheformate.
need (b)
for Energy
costly efficiencies
downstreamversusseparations.
partial current densit
address
product
selectivity
considerations,
which
is important
phase CO reduction where CO is first
dissolved
in
an
aqueous
electrolyte
then
reduced
at
catalyst
surface.
(b)
shows
vapor
address
product
selectivity
considerations,
which
is
important
address
product
selectivity
considerations,
which
is
important
arebe
shown
in fed
solid
Performances
obtained
for vapor-fed
CO R lectrodes
ard
practical
reactor
deFurthermore,
the
actual
electrolyte
can
either
aqueous
to symbols,
to avoid the
need
for
costly
downstream
separations.
delivered
igns
that
operate
using
CO
Moving
toward
practical
reactor
deving
toward
practical
reactor
deavoid
costly
downstream
separations.
to to
avoid
the the
needneed
forinfor
costly
downstream
separations.
reactors
are
shown
hollow
symbols.
All
energy
efficiencies
were
Furthermore,Furthermore,
the
actual electrolyte
can
either
becanaqueous
to aqueous
perate
using
COusing
Furthermore,
actual
electrolyte
can
either
be aqueous
2 delivered
gns
that
operate
using
the the
actual
electrolyte
either
be
to to
ns
that
operate
COCO
2 delivered
2 delivered
formefficiency
a catalyst/liquid
electrolyte
interface
(Figure
1b), are
or the reve
form a catalyst/liquid
electrolyte
(Figure
or(Figure
=interface
,1b),
where
E 1b),
and
mula:energy
catalyst/liquid
electrolyte
interface
1b),
dethe
in cathode
the
vapor
phase
can
formform
a catalyst/liquid
electrolyte
interface
(Figure
or Eor
ocome
the
cathode
in
the
vapor
phase
ohe
the
vapor
phase
cancanion-conducting
cathode
in in
the
vapor
phase
can
ideally
polymer
that
can
transport
charged
ideally
an ion-conducting
charged
ideally
an ion-conducting
thatthat
can can
transport
charged
CO
R product,
and
V ispolymer
thepolymer
uncompensated
cell
these
performance
overcome
these
performance
an Hion-conducting
polymer
thattransport
canvoltage.
transport
charged
pelp
to to
overcome
these
performance
orideally
OH
form
species (e.g.,species
Hspecies
OH
)catalyst/polymerand
catalyst/polymer(e.g.,
or or
OH
) aand
formform
a catalyst/polymer(e.g.,
H) and

ctrochemical
CO
schemes.
(a) Aqueous-phase
COCO
solubilized
insolubilized
an aqueous
2R reactorCO
2R, where CO
2 is first
eDifferent
1. Different
electrochemical
schemes.
(a)CO
Aqueous-phase
solubilized
an aqueous
electrolyte
electrochemical
R reactor
schemes.
(a) Aqueous-phase
CO2CO
is first
inelectrolyte
aninaqueous
electrolyte
2R reactor
2R, where
2 is first
2CO
2R, where
employing
an
(b)
aqueous
or
(c)
polymer
electrolyte.
catalyst
surface.
Vapor-fed
CO
21
employing
an
(b)
aqueous
or
(c)
polymer
electrolyte.
hen
reduced
a catalyst
surface.
Vapor-fed
R employing
an
(b)
aqueous
or
(c)
polymer
electrolyte.
reduced
at aatFigure
catalyst
surface.
Vapor-fed
CO2CO
2RCO
1.5:
Different
electrochemical
reduction
reaction
chemes
are
shown
in
(a)
and
(b).
(a)
shows
schematic
of aqueousFigure
2.
State-of-the-art
performance
of
vapor-fed
CO
devices.
(a) Faradaic efficie

(Eanode
− Ecathode
) × FE
Vcell

ysolubility
challenges.
help
to overcome
these performance
nd
solubility
challenges.
challenges.

2 +

+ cell

21 21

anode

cathode

electrolyte interface
(Figure
1c). (Figure
+ 1c).1c). −
electrolyte
interface
(Figure
electrolyte
interface
22

21

318

14
CO2 reduction system where the CO2 is delivered in gas phase to catalyst on a gas diffusion electrode; the electrolyte can either be
Paperdensity for different reported devices. Open shapes refer
aqueous or polymer. (c) shows a plot of Faradaic efficiency vs. partial current

to aqueous-phase CO2 reduction and filled shapes represent vapor-fed CO2 reduction devices. The numbers on the plot refer to
references from the original paper. Reproduced from Higgins et al. Copyright 2019, American Chemical Society.67
The final device consideration which is critical to device performance is the choice of CO2 delivery
Published on 04 June 2018. Downloaded by California Institute of Technology on 11/1/2018 9:17:54 PM.

method. There are two different options for CO2 delivery, either the CO2 can be dissolved into the electrolyte,
or the CO2 can be delivered to the catalyst in the gas phase using a gas diffusion electrode. When scientists first
began studying CO2 reduction aqueous CO2 systems were predominantly used (Fig. 1.5a).18,19 This system while
complex is simpler to understand relative to a gas fed system. This is due to the fact that when the CO2 is
aqueous the whole system fits into the mold a typical electrochemical reaction, where the reactant is dissolved
in the electrolyte, planar metal catalysts can be used, and the anode and cathode are fully wetted and electrically
Fig. 1 Schematic of a gas diffusion electrode.

connected via the electrolyte. Despite the advantages of simplicity, it is not possible to run CO2 reduction at
of CO
42 mM. This
30% (33mM).
higher than26,67
thatIn
of
2 issolubility
high current densities in aqueous CO2 fed systems, due to the
low
of level
CO2is inonlywater

Paper

dissolved CO2 (33 mM) and cannot account for the order of
increase
CO2R current
density observed
experiaddition the pH of the electrolyte must be kept near neutral magnitude
to maximize
COin2 solubility,
meaning
that there
is
mentally. (2) Recent experimental and theoretical work have
26
an abundance of protons available creating favorable conditions
for HER.
demonstrated
the importance of water and hydrated cations on
the elementary processes involved in CO2R.17,18 Therefore, we
In order to overcome these challenges, researchers have
developed
where
be
2 canwith
propose
that it isporous
necessaryelectrodes
for the catalyst
to beCO
covered
electrolyte
in
order
to
be
active.
This
means
that
although
CO
2 is
fed to the catalyst in gas phase, Fig 1.5b.67 This has allowed for
devices to be built that not only had improved
supplied to the GDE from the gas phase, the reactant at the
site is still dissolved CO2.
current density, but also improved selectivities. We see in Figcatalyst
1.5c
that
the maximum current density observed
The performance of a GDE greatly depends on the local
environment
within thewe
CL see
and that
the balance
for an aqueous fed device was around 20 mA/cm2, but for vapor
fed devices
manybetween
devicestransport
reach
phenomena and reaction kinetics. Based on the capillary pressure,
CL pore-size
and their
wettability,
pores can be
hundreds of mA/cm2. We also see an increased Faradaic efficiency
fordistribution
higherView
order
products
likethe
ethylene
Article Online
either flooded (Fig. 2a) or dry (Fig. 2c). The partially wetted CL case
PCCP
depicted in Fig. 2b occurs when
there is a mixture of flooded and

a)

b)

c)

d)

Fig. 1 Schematic of a gas diffusion electrode.

of CO2 is 42 mM. This level is only 30% higher than that of
dissolved CO2 (33 mM) and cannot account for the order of
magnitude increase in CO2R current density observed experimentally. (2) Recent experimental and theoretical work have
demonstrated the importance of water and hydrated cations on
the elementary processes involved in CO2R.17,18 Therefore, we

Fig. 2 Schematic of pore conditions in the catalyst layer. (a) Flooded
pore: pore volume filled with electrolyte. (b) Wetted pore: a thin layer of
electrolyte covers the pore walls. (c) Dry pore: catalyst inactive due to lack
of an ionic pathway.

dry pores. Flooded pores completely eliminate gas channels within
the CL, resulting in high mass-transport resistances for gaseous
reactants. Dry pores will be inactive due to the lack of aqueous
16974 The
| Phys.
Chem.
Chem. Phys.,
2018, 20, 16973--16984
electrolyte and an ionic pathway.
film
of electrolyte
in the
wetted pores needs to be thin to minimize CO2 transport
resistance to the catalyst, but thick enough to maintain good

dry pores. Flo
the CL, resu
reactants. Dr
electrolyte an
wetted pore
resistance to
ionic conduc
defined as t
pressure, w
gas-phase p
hydrophilic
increases, ot
large hydrop
small hydrop
needs to adj
the CL pores
Because t
systems invo
processes, it
of a particul
CL without
chemistry an
GDE perform
most part, t
et al. have fo
had no effec
behaviour w
shown that
current den
Other param
methods, et
detailed sur
Endrodi et a
designs of G
efforts on th
modeling, D

15
Figure 1.6: (a) shows a schematic of gas diffusion CO2 reduction electrode. The orange lines are carbon fibers, the orange circles are
carbon black, and the black circles are catalyst. (b)-(d) show different states that the catalyst can be in; (b) is flooded, (c) is wetted,
and (d) is dry. This was reproduced and edited from Weng et al. Copyright 2018, Royal Society of Chemistry.68
These improvements are due to the increased flux of CO2 to the catalyst which allows for higher current
densities. In addition, the electrolyte can have a much higher pH since the CO2 is being delivered in gas phase,
which suppresses HER due to the lack of protons. A typical gas diffusion electrode (GDE) is composed of
carbon fiber and carbon black support with catalyst coated on top (Fig. 1.6a). On the backside of the carbon
paper is a gas chamber of pure CO2; the CO2 diffuses through the network of carbon fibers towards the
electrolyte. Coated on the carbon fibers is the microporous layer which is typically composed of carbon black
and polytetrafluoroethylene (PTFE). The carbon black provides conductivity and the PTFE is hydrophobic to
prevent flooding. On top of the microporous layer the catalyst is coated and in contact with the electrolyte.
Despite the advantages of this new device configuration, it adds new complexity to the system. In the
aqueous system all of the catalyst was always in contact with the electrolyte, if the catalyst is not in contact with
the electrolyte no reaction can occur since there are no protons for the reaction and the cathode is not electrically
connected to the anode. In a vapor fed device there are now three states that the catalyst can be in flooded,
wetted, and dry (Fig 1.6b-d). When the catalyst is dry no reaction can occur because it not contacted with the
electrolyte. If the catalyst is flooded, the high current densities cannot be achieved due to the low solubility of
CO2 in water which causes mass transport limitations. To create a stable wetted condition the pressure of the
CO2 and electrolyte must be carefully balanced. Catalyst coatings are also often employed to prevent flooding.
This device also raises new questions about the physics of the device: (1) how thick is the electrolyte layer, (2)
what is the concentration of CO2, H+, and OH- in this layer, and (3) what percentage of the catalyst is in this
wetted condition.

This thesis seeks to experimentally answers these questions (Chapter 4) and use this

knowledge to create an optimized GDE structure (Chapter 3).
1.8

Thesis Outline
This thesis provides deeper understanding to how the local catalyst environment can affect the CO2R

reaction. Chapter 2 focuses on how nanostructured catalysts can induce local pH gradients for improved CO2
reduction selectivity. Chapter 3 combines nanostructured catalysts and gas diffusion electrodes for improved
device performance. Chapter 4 demonstrates a confocal fluorescent microscopy system to measure the local
pH in COR gas diffusion electrodes. Chapter 5 and 6 shows how light can be used to affect the CO2R reaction,

16
specifically looking at plasmon-mediated electron transfer in metal/semiconductor heterojunctions. Then in
Chapter 6 and 7 the focus shifts to how various CO2 capture and conversion systems to provide insight on what
the optimal systems are to be deployed commercially and provides techno-economic analysis. Finally, Chapter
8 summarizes the main outcomes and contributions of this thesis. Additionally, it provides perspective on future
experiments for local pH measurements.

BIBLIOGRAPHY CHAPTER 1
1.
2.

3.

4.
5.
6.
7.

Masson-Delmotte, V.; Zhai, P.; Pörtner, H.-O.; Roberts, D.; Skea, J.; Shukla, P. R.; Pirani,
A. IPCC Report: Global Warming of 1.5°C; 2018.
Le Quéré, C.; Andrew, R. M.; Friedlingstein, P.; Sitch, S.; Pongratz, J.; Manning, A. C.;
Korsbakken, J. I.; Peters, G. P.; Canadell, J. G.; Jackson, R. B.; Boden, T. A.; Tans, P. P.;
Andrews, O. D.; Arora, V. K.; Bakker, D. C. E.; Barbero, L.; Becker, M.; Betts, R. A.;
Bopp, L.; Chevallier, F.; Chini, L. P.; Ciais, P.; Cosca, C. E.; Cross, J.; Currie, K.; Gasser,
T.; Harris, I.; Hauck, J.; Haverd, V.; Houghton, R. A.; Hunt, C. W.; Hurtt, G.; Ilyina, T.;
Jain, A. K.; Kato, E.; Kautz, M.; Keeling, R. F.; Klein Goldewijk, K.; Körtzinger, A.;
Landschützer, P.; Lefèvre, N.; Lenton, A.; Lienert, S.; Lima, I.; Lombardozzi, D.; Metzl,
N.; Millero, F.; Monteiro, P. M. S.; Munro, D. R.; Nabel, J. E. M. S.; Nakaoka, S.; Nojiri,
Y.; Padín, X. A.; Peregon, A.; Pfeil, B.; Pierrot, D.; Poulter, B.; Rehder, G.; Reimer, J.;
Rödenbeck, C.; Schwinger, J.; Séférian, R.; Skjelvan, I.; Stocker, B. D.; Tian, H.; Tilbrook,
B.; van der Laan-Luijkx, I. T.; van der Werf, G. R.; van Heuven, S.; Viovy, N.; Vuichard,
N.; Walker, A. P.; Watson, A. J.; Wiltshire, A. J.; Zaehle, S.; Zhu, D. Global Carbon Budget
2018. Earth Syst. Sci. Data Discuss. 2018.
Davis, S. J.; Lewis, N. S.; Shaner, M.; Aggarwal, S.; Arent, D.; Azevedo, I. L.; Benson, S.
M.; Bradley, T.; Brouwer, J.; Chiang, Y. M.; Clack, C. T. M.; Cohen, A.; Doig, S.; Edmonds,
J.; Fennell, P.; Field, C. B.; Hannegan, B.; Hodge, B. M.; Hoffert, M. I.; Ingersoll, E.;
Jaramillo, P.; Lackner, K. S.; Mach, K. J.; Mastrandrea, M.; Ogden, J.; Peterson, P. F.;
Sanchez, D. L.; Sperling, D.; Stagner, J.; Trancik, J. E.; Yang, C. J.; Caldeira, K. Net-Zero
Emissions Energy Systems. Science (80-. ). 2018, 360 (6396).
Lewis, N. S. Toward Cost Effective Solar. Science (80-. ). 2007, 798 (1), 798–802.
Dowling, J. A.; Rinaldi, K. Z.; Ruggles, T. H.; Davis, S. J.; Yuan, M.; Tong, F.; Lewis, N.
S.; Caldeira, K. Role of Long-Duration Energy Storage in Variable Renewable Electricity
Systems. Joule 2020, 4 (9), 1907–1928.
Yang, Z.; Zhang, J.; Kintner-meyer, M. C. W.; Lu, X.; Choi, D.; Lemmon, J. P.
Electrochemical Energy Storage for Green Grid: Status and Challenges. Chem. Rev. 2011,
111 (5), 3577–3613.
U.S. Department of Energy. Grid Energy Storage Report; 2013.

17

8.
9.
10.
11.
12.
13.

14.
15.
16.
17.
18.
19.
20.

21.

Administration, U. S. E. I. Few Transportation Fuels Surpass the Energy Densities of
Gasoline and Diesel. Today In Energy 2013, 9991.
What is cement? https://www.worldcoal.org/coal/uses-coal/coal-cement (accessed Feb
10, 2020).
Andrew, R. M. Global CO2 Emissions from Cement Production, 1928-2018. Earth Syst.
Sci. Data 2019, 11 (4), 1675–1710.
World Steel Association. World Steel Figures. 2020 World steel Fig. 2020.
Geyer, R.; Jambeck, J. R.; Law, K. L. Production, Use, and Fate of All Plastics Ever Made.
Sci. Adv. 2017, 3 (7), 25–29.
Rissman, J.; Bataille, C.; Masanet, E.; Aden, N.; Morrow, W. R.; Zhou, N.; Elliott, N.; Dell,
R.; Heeren, N.; Huckestein, B.; Cresko, J.; Miller, S. A.; Roy, J.; Fennell, P.; Cremmins, B.;
Koch Blank, T.; Hone, D.; Williams, E. D.; de la Rue du Can, S.; Sisson, B.; Williams, M.;
Katzenberger, J.; Burtraw, D.; Sethi, G.; Ping, H.; Danielson, D.; Lu, H.; Lorber, T.;
Dinkel, J.; Helseth, J. Technologies and Policies to Decarbonize Global Industry: Review
and Assessment of Mitigation Drivers through 2070. Appl. Energy 2020, 266 (November
2019), 114848.
Ore, I.; Pigments, I. O.; Rock, P.; Crystal, Q.; Earths, R.; Ash, S. Mineral Commodity
Summaries 2021; 2021.
Nitopi, S. A. Understanding the Factors That Govern Activity and Selectivity of the
Electrochemical Carbon Dioxide Reduction Reaction on Copper Catalysts, Stanford,
2019.
Welch, A. J.; Dunn, E.; Duchene, J. S.; Atwater, H. A. Bicarbonate or Carbonate Processes
for Coupling Carbon Dioxide Capture and Electrochemical Conversion. ACS Energy Lett.
2020, 5 (3), 940–945.
Dau, H.; Limberg, C.; Reier, T.; Risch, M.; Roggan, S.; Strasser, P. The Mechanism of
Water Oxidation: From Electrolysis via Homogeneous to Biological Catalysis.
ChemCatChem 2010, 2 (7), 724–761. https://doi.org/10.1002/cctc.201000126.
Hori, Y.; Vayenas, C. G.; White, R. E.; Gamboa-Aldeco, M. E. Electrochemical CO2
Reduction on Metal Electrodes. In Modern Aspects of Electrochemistry; Springer: New York,
2008; pp 89–189. https://doi.org/https://doi.org/10.1007/978-0-387-49489-0_3.
Hori, Yoshio Kikuchi, Katsuhei Suzuki, S. Production of CO and CH4 in the
Electrochemical Reduction of CO2 at Metal Electrodes in Aqueous Hydrogencarbonate
Solution. Chem. Lett. 1985, 1695–1698.
Appel, A. M.; Bercaw, J. E.; Bocarsly, A. B.; Dobbek, H.; Dubois, D. L.; Dupuis, M.;
Ferry, J. G.; Fujita, E.; Hille, R.; Kenis, P. J. A.; Kerfeld, C. A.; Morris, R. H.; Peden, C.
H. F.; Portis, A. R.; Ragsdale, S. W.; Rauchfuss, T. B.; Reek, J. N. H.; Seefeldt, L. C.;
Thauer, R. K.; Waldrop, G. L. Frontiers, Opportunities, and Challenges in Biochemical
and Chemical Catalysis of CO2 Fixation. Chem. Rev. 2013, 113 (8), 6621–6658.
White, J. L.; Baruch, M. F.; Pander, J. E.; Hu, Y.; Fortmeyer, I. C.; Park, J. E.; Zhang, T.;
Liao, K.; Gu, J.; Yan, Y.; Shaw, T. W.; Abelev, E.; Bocarsly, A. B. Light-Driven
Heterogeneous Reduction of Carbon Dioxide: Photocatalysts and Photoelectrodes. Chem.

18

22.
23.
24.
25.
26.

27.
28.
29.
30.
31.
32.
33.
34.
35.

Rev. 2015, 115 (23), 12888–12935.
Gattrell, M.; Gupta, N.; Co, A. A Review of the Aqueous Electrochemical Reduction of
CO2 to Hydrocarbons at Copper. J. Electroanal. Chem. 2006, 594 (1), 1–19.
Qiao, J.; Liu, Y.; Hong, F.; Zhang, J. A Review of Catalysts for the Electroreduction of Carbon
Dioxide to Produce Low-Carbon Fuels; 2014; Vol. 43. https://doi.org/10.1039/c3cs60323g.
Kortlever, R.; Shen, J.; Schouten, K. J. P.; Calle-Vallejo, F.; Koper, M. T. M. Catalysts and
Reaction Pathways for the Electrochemical Reduction of Carbon Dioxide. J. Phys. Chem.
Lett. 2015, 6 (20), 4073–4082.
Linstrom, P. J.; Mallard, W. G. NIST Chemistry WebBook, NIST Standard Reference
Database Number 69. National Institute of Standards and Technology. 2014, p 20899.
Singh, M. R.; Clark, E. L.; Bell, A. T. Effects of Electrolyte, Catalyst, and Membrane
Composition and Operating Conditions on the Performance of Solar-Driven
Electrochemical Reduction of Carbon Dioxide. Phys. Chem. Chem. Phys. 2015, 17 (29),
18924–18936.
Bagger, A.; Ju, W.; Varela, A. S.; Strasser, P.; Rossmeisl, J. Electrochemical CO2 Reduction:
A Classification Problem. ChemPhysChem 2017, 18 (22), 3266–3273.
Noda, H.; Ikeda, S.; Oda, Y.; Imai, K.; Maeda, M.; Ito, K. Electrochemical Reduction of
Carbon Dioxide at Various Metal Electrodes in Aqueous Potassium Hydrogen Carbonate
Solution. Chem. Soc. Japan 1990, 63 (9), 2459–2462.
Nørskov, J. K.; Bligaard, T.; Logadottir, A.; Kitchin, J. R.; Chen, J. G.; Pandelov, S.;
Stimming, U. Trends in the Exchange Current for Hydrogen Evolution. J. Electrochem. Soc.
2005, 152 (3), J23.
Kuhl, K. P.; Hatsukade, T.; Cave, E. R.; Abram, D. N.; Kibsgaard, J.; Jaramillo, T. F.
Electrocatalytic Conversion of Carbon Dioxide to Methane and Methanol on Transition
Metal Surfaces. J. Am. Chem. Soc. 2014, 136 (40), 14107–14113.
Zhu, W.; Michalsky, R.; Metin, Ö.; Lv, H.; Guo, S.; Wright, C. J.; Sun, X.; Peterson, A. A.;
Sun, S. Monodisperse Au Nanoparticles for Selective Electrocatalytic Reduction of CO2
to CO. J. Am. Chem. Soc. 2013, 135 (45), 16833–16836.
Zhang, W.; He, J.; Liu, S.; Niu, W.; Liu, P.; Zhao, Y.; Pang, F.; Xi, W.; Chen, M.; Pang, S.
S.; Ding, Y. Atomic Origins of High Electrochemical CO2 Reduction Efficiency on
Nanoporous Gold. Nanoscale 2018, 10 (18), 8372–8376.
De Gregorio, G. L.; Burdyny, T.; Loiudice, A.; Iyengar, P.; Smith, W. A.; Buonsanti, R.
Facet-Dependent Selectivity of Cu Catalysts in Electrochemical CO2 Reduction at
Commercially Viable Current Densities. ACS Catal. 2020, 10 (9), 4854–4862.
Welch, A. J.; Duchene, J. S.; Tagliabue, G.; Davoyan, A.; Cheng, W. H.; Atwater, H. A.
Nanoporous Gold as a Highly Selective and Active Carbon Dioxide Reduction Catalyst.
ACS Appl. Energy Mater. 2019, 2 (1), 164–170. https://doi.org/10.1021/acsaem.8b01570.
Nitopi, S.; Bertheussen, E.; Scott, S. B.; Liu, X.; Engstfeld, A. K.; Horch, S.; Seger, B.;
Stephens, I. E. L.; Chan, K.; Hahn, C.; Nørskov, J. K.; Jaramillo, T. F.; Chorkendorff, I.
Progress and Perspectives of Electrochemical CO2 Reduction on Copper in Aqueous
Electrolyte. Chem. Rev. 2019, 119 (12), 7610–7672.

19

36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.

Mariano, R. G.; McKelvey, K.; White, H. S.; Kanan, M. W. Selective Increase in CO2
Electroreduction Activity at Grain-Boundary Surface Terminations. Science (80-. ). 2017,
358 (6367), 1187–1192.
Manthiram, K.; Beberwyck, B. J.; Alivisatos, A. P. Enhanced Electrochemical
Methanation of Carbon Dioxide with a Dispersible Nanoscale Copper Catalyst. J. Am.
Chem. Soc. 2014, 136 (38), 13319–13325.
Loiudice, A.; Lobaccaro, P.; Kamali, E. A.; Thao, T.; Huang, B. H.; Ager, J. W.; Buonsanti,
R. Tailoring Copper Nanocrystals towards C2 Products in Electrochemical CO2
Reduction. Angew. Chemie - Int. Ed. 2016, 55 (19), 5789–5792.
Reske, R.; Mistry, H.; Behafarid, F.; Roldan Cuenya, B.; Strasser, P. Particle Size Effects
in the Catalytic Electroreduction of CO2 on Cu Nanoparticles. J. Am. Chem. Soc. 2014, 136
(19), 6978–6986.
Hammer, B.; Morikawa, Y.; Nørskov, J. K. CO Chemisorption at Metal Surfaces and
Overlayers.
Phys.
Rev.
Lett.
1996,
76
(12),
2141–2144.
Linic, S.; Christopher, P.; Ingram, D. B. Plasmonic-Metal Nanostructures for Efficient
Conversion of Solar to Chemical Energy. Nat. Mater. 2011, 10 (12), 911–921.
Atwater, H. A.; Polman, A. Plasmonics for Improved Photovoltaic Devices. Nat. Mater.
2010, 9 (3), 205–213.
Warren, S. C.; Thimsen, E. Plasmonic Solar Water Splitting. Energy Environ. Sci. 2012, 5
(1), 5133–5146.
Hou, W.; Cronin, S. B. A Review of Surface Plasmon Resonance-Enhanced
Photocatalysis. Adv. Funct. Mater. 2013, 23 (13), 1612–1619.
Linic, S.; Aslam, U.; Boerigter, C.; Morabito, M. Photochemical Transformations on
Plasmonic Metal Nanoparticles. 2015, 14 (June).
Linic, S.; Christopher, P.; Xin, H.; Marimuthu, A. Catalytic and Photocatalytic
Transformations on Metal Nanoparticles with Targeted Geometric and Plasmonic
Properties. Acc. Chem. Res. 2013, 46 (8), 1890–1899.
Ashcroft, N. W.; Mermin, N. D. Solid State Physics; Hartcourt College: New York, 1976.
Link, S.; El-Sayed, M. A. Optical Properties and Ultrafast Dynamics of Metallic
Nanocrystals. Annu. Rev. Phys. Chem. 2003, 54, 331–366.
Hartland, G. V. Optical Studies of Dynamics in Noble Metal Nanostructures. Chem. Rev.
2011, 111 (6), 3858–3887.
Brongersma, M. L.; Halas, N. J.; Nordlander, P. Plasmon-Induced Hot Carrier Science
and Technology. Nat. Nanotechnol. 2015, 10 (1), 25–34.
Clavero, C. Plasmon-Induced Hot-Electron Generation at Nanoparticle/Metal-Oxide
Interfaces for Photovoltaic and Photocatalytic Devices. Nat. Photonics 2014, 8 (2), 95–103.
Christopher, P.; Moskovits, M. Hot Charge Carrier Transmission from Plasmonic
Nanostructures. Annu. Rev. Phys. Chem. 2017, 68 (March), 379–398.
Duchene, J. S.; Tagliabue, G.; Welch, A. J.; Cheng, W. H.; Atwater, H. A. Hot Hole
Collection and Photoelectrochemical CO2 Reduction with Plasmonic Au/p-GaN

20

54.
55.
56.

57.
58.
59.
60.
61.

62.
63.
64.
65.
66.
67.
68.

Photocathodes. Nano Lett. 2018, 18 (4), 2545–2550.
Duchene, J. S.; Tagliabue, G.; Welch, A. J.; Li, X.; Cheng, W. H.; Atwater, H. A. Optical
Excitation of a Nanoparticle Cu/p-NiO Photocathode Improves Reaction Selectivity for
CO2 Reduction in Aqueous Electrolytes. Nano Lett. 2020, 20 (4), 2348–2358.
De Luna, P.; Hahn, C.; Higgins, D.; Jaffer, S. A.; Jaramillo, T. F.; Sargent, E. H. What
Would It Take for Renewably Powered Electrosynthesis to Displace Petrochemical
Processes? Science (80-. ). 2019, 364 (6438).
Singh, M. R.; Clark, E. L.; Bell, A. T. Effects of Electrolyte, Catalyst, and Membrane
Composition and Operating Conditions on the Performance of Solar-Driven
Electrochemical Reduction of Carbon Dioxide. Phys. Chem. Chem. Phys. 2015, 17, 18924–
18936.
Hall, A. S.; Yoon, Y.; Wuttig, A.; Surendranath, Y. Mesostructure-Induced Selectivity in
CO2 Reduction Catalysis. J. Am. Chem. Soc. 2015, 137 (47), 14834–14837.
Bard, A. J.; Faulkner, L. R. Fundamentals and Applications of Needle Trap Devices; 2002; Vol. 2.
Zhang, Z.; Melo, L.; Jansonius, R. P.; Habibzadeh, F.; Grant, E. R.; Berlinguette, C. P. PH
Matters When Reducing CO2in an Electrochemical Flow Cell. ACS Energy Lett. 2020, 5
(10), 3101–3107.
Lu, X.; Zhu, C.; Wu, Z.; Xuan, J.; Francisco, J. S.; Wang, H. In Situ Observation of the
PH Gradient near the Gas Diffusion Electrode of CO2 Reduction in Alkaline Electrolyte.
J. Am. Chem. Soc. 2020, 142 (36), 15438–15444.
Wang, L.; Nitopi, S. A.; Bertheussen, E.; Orazov, M.; Morales-Guio, C. G.; Liu, X.;
Higgins, D. C.; Chan, K.; Nørskov, J. K.; Hahn, C.; Jaramillo, T. F. Electrochemical
Carbon Monoxide Reduction on Polycrystalline Copper: Effects of Potential, Pressure,
and PH on Selectivity toward Multicarbon and Oxygenated Products. ACS Catal. 2018, 8
(8), 7445–7454.
Yang, K.; Kas, R.; Smith, W. A. In Situ Infrared Spectroscopy Reveals Persistent Alkalinity
near Electrode Surfaces during CO2 Electroreduction. J. Am. Chem. Soc. 2019, 141 (40).
Nesbitt, N.; Smith, W. Water Activity Regulates CO2 Reduction in Gas-Diffusion
Electrodes. ChemRxiv 2021, No. January. https://doi.org/10.26434/chemrxiv.13571141.
Nesbitt, N.; Smith, W. A. Operando Topography and Mechanical Property Mappinig of
CO2 Reduction Gas-Diffusion Electrodes Operating at High Currrent Densities. J.
Electrochem. Soc. 2021.
Monteiro, M. C. O.; Koper, M. T. M. Measuring Local PH in Electrochemistry. Curr. Opin.
Electrochem. 2021, 25, 100649.
Corson, E. R.; Creel, E. B.; Kim, Y.; Urban, J. J.; Kostecki, R.; McCloskey, B. D. A
Temperature-Controlled Photoelectrochemical Cell for Quantitative Product Analysis.
Rev. Sci. Instrum. 2018, 89 (5).
Higgins, D.; Hahn, C.; Xiang, C.; Jaramillo, T. F.; Weber, A. Z. Gas-Diffusion Electrodes
for Carbon Dioxide Reduction: A New Paradigm. ACS Energy Lett. 2019, 4, 317–324.
Weng, L.-C.; Bell, A. T.; Weber, A. Z. Modeling Gas-Diffusion Electrodes for CO2

21

Reduction. Phys. Chem. Chem. Phys. 2018, No. 20, 16973–16984.

22
Chapter 2

NANOPOROUS GOLD AS A HIGHLY SELECTIVE AND ACTIVE CARBON
DIOXIDE REDUCTION CATALYST
2.1

Introduction
The ability to reduce CO2 into useful chemicals or fuels will not only enable clean technology, but it will

also close the carbon cycle by recycling CO2 and preventing its further addition to the atmosphere.1 The CO2
reduction products can either be liquid fuels like ethanol or gaseous products like syngas (H2 and CO), which
are feedstocks for thermocatalytic transformations via the Fischer−Tropsch process.2–5 To date, CO2 reduction
is not a widespread technology because of low energy efficiency associated with high overpotentials, a lack of
electrocatalytic stability, and poor selectivity for the CO2 reduction reaction (CO2RR) over the H2 evolution
reaction (HER), which results in low partial-current densities for the product of interest.6
Various approaches have been explored to improve the activity and selectivity of Au-based
electrocatalysts for the CO2RR, from controlling nanocrystal size to tailoring of the exposed crystal facets, and
even surface functionalization with molecular coatings.7–12 Recently, nanoporous catalytic architectures have
shown promise for electrochemical CO2 reduction due to their large internal surface area and prevalence of
stepped sites and grain boundaries inherent in their complex structure of highly-curved metal ligaments.13–19 This
propensity for under-coordinated atomic sites has been suggested to play a pivotal role in improving the
selectivity of CO2 reduction in nanoporous silver (np-Ag) cathodes by stabilizing CO2− intermediates involved
in the electrochemical conversion of CO2 to CO.13 Similar mechanisms have been invoked to explain the
electrocatalytic performance of nanoporous gold (np-Au) films.18,19 While the highly-irregular surface atomic
structure of np-Au is well known, relatively less attention has been devoted to exploring how molecular transport
into and out of such a tightly-confined catalytic system may also affect the selectivity for CO2 reduction within
the porous network. Mesoporous Au films with controlled pore sizes around 200 nm in diameter have previously
been shown to exhibit increased selectivity for CO2 reduction with increasing film thickness from 0.5 μm up to
2.7 μm.20 The improved selectivity is attributed to the formation of a pH gradient within the porous network as
protons are consumed during electrolysis faster than they can be replenished by the supporting electrolyte; this
effect is increased with increasing thickness of the mesoporous metal cathode. Although this study only adjusted
the overall film thickness, these observations strongly suggest that fine-tuning the metal porosity by controlling

23
the pore size could offer a simple route to further improving the selectivity of porous cathodes for
electrochemical CO2 reduction in aqueous electrolytes.
Here, we use np-Au films with pore diameters on the order of tens of nanometers to explore the
influence of metal porosity on the selectivity for CO2 reduction in aqueous electrolytes. Due to their small pore
diameters, the porous network of metal ligaments is able to sustain pH gradients within np-Au films that are
half as thick (~800 nm) as those previously reported in mesoporous Au films (~2 μm).20 This effect becomes
more prominent upon further decreasing the pore diameter from ~30 nm down to ~10 nm, as evidenced by
electrochemical studies. We find that np-Au films are highly selective for the conversion of CO2 to CO with
high Faradaic efficiency (FE ~99%) at modest overpotentials (η = 0.40 V), while at the same time delivering
large partial current densities for CO (JCO = 6.2 mA cm-2). Finally, we demonstrate that these np-Au films exhibit
excellent electrochemical durability and maintain Faradaic efficiency of ~80% for CO production over 4.5 days
of continuous electrolysis at an applied potential of E = –0.5 V vs the reversible hydrogen electrode (RHE).
2.2

Fabrication and Characterization of Catalyst

Figure 2.1: Helium FIB images of (a) top-down view of a nanoporous Au (np-Au) film that was etched at room temperature
(RT) and (b) top-down view of a np-Au film that was etched at low temperature (LT). (c) SEM cross-section image of a RT-etched

24
np-Au film. All scale bars represent 100 nm. (d) Electrochemical surface area enhancement as a function of film thickness for RTetched np-Au films as determined by Cu underpotential deposition (UPD) experiments.
The np-Au films were fabricated by electron-beam deposition of Ag and Au alloys onto clean silicon
(Si) substrates (Si [p-type, 0-10 W cm, (100) orientation, 620 ± 25 µm thick, University Wafers]), followed by
selectively etching Ag from the Ag/Au alloy with nitric acid (see Methods). Briefly, An AMOD dual electron
beam deposition system (System 02520, Angstrom Engineering) was used to fabricate all samples. First, Si
substrates were cleaned by sonicating sequentially in acetone, isopropanol, and deionized water for five minutes
each. The samples were then stored in deionized water until they were dried with N2 prior to being placed in
the electron beam deposition chamber. The materials for electron beam deposition were ordered from
Plasmaterials. The Au target was 99.99% pure with 3-6 mm random size pieces, Ag was 99.99% pure with 3-6
mm random size pieces, and Ti was 99.995% pure in 0.25” diameter pellets. First, 2 nm of Ti was deposited at
a rate of 1 Å/s, then 50 nm of Au was deposited at a rate of 2 Å/s. Next, Au and Ag were co-deposited at a
rate of 2Å/s and 6Å/s, respectively, to create a 25% Au and 75% Ag alloy. Over the course of the deposition
the partial pressure of the chamber would rise from ~10-7 torr to ~10-6 torr and the temperature would rise from
20 °C to 60 °C for a 1μm thick sample. If the samples were then placed in a beaker of 70 wt.% HNO3 for 10
min to etch out the silver. After this time had elapsed, they were rinsed with deionized water 10 times before
being dried with N2. The room temperature np-Au morphology was obtained by etching the Au/Ag alloy films
at room temperature (~22 °C) (denoted RT np-Au), while the low temperature samples were etched at in a
freezer (-20°C) (denoted LT np-Au). Secondary ion mass spectrometry (SIMS) indicates that approximately 1.3
at.% of residual Ag remains in the structure after etching, consistent with prior reports.21
Figure 2.1 shows helium focused ion beam (He FIB) images of np-Au samples that were etched at room
temperature (Figure 2.1a) and at low temperature (Figure 2.1b), displaying an average ligament thickness of 28
± 8 nm and 10 ± 2 nm, respectively. Chemical etching at low temperatures restricts the surface mobility of the
Au atoms during etching and ensures that the ligament diameter is decreased.22 A scanning electron microscope
(SEM) cross-section image of the ~800 nm thick RT etched np-Au film shows that the entirety of the film is
porous down to the planar Au base layer (Figure 2.1c). As shown in this cross-sectional image, we routinely
observed that the fully etched np-Au films were approximately 20% thinner (~800 nm) than the initial thickness
of the Au/Ag alloy (~1 µm). To characterize the electrochemical surface area of the np-Au films, we performed
Cu underpotential deposition (UPD) experiments. A Copper(II) sulfate (≥99%, Sigma Aldrich) (CuSO4)
solution (0.1 M) in 0.5 M H2SO4 was prepared and used as the deposition bath for all Cu UPD experiments.
The solution was bubbled with N2 (Research grade from Airgas) for 30 min to remove dissolved O2 from the

25
solution prior to starting the experiment. The working electrode was a planar Au film or a np-Au film of various
thicknesses with a Pt mesh counter electrode and a Ag/AgCl reference electrode. The potential of the working
electrode was swept from 450 mV to 50 mV vs. Ag/AgCl at a scan rate of 5 mV s-1. A total of three to five
electrodes were measured at each thickness ranging from 100 nm to 1.6 μm and the electrochemical surface area
enhancement was obtained by taking the average surface area of the np-Au film relative to that of a planar Au
film of known geometrical surface area. The maximum roughness factor of ~57 for the thickest films was
obtained using this method. It is also important to note that the surface area increases linearly with film thickness,
indicating that the entire surface area of the np-Au film is electrochemically accessible (Figure 2.1d).

Figure 2.2: XRD spectra of (a) RT np-Au (~800 nm thick) on a glass substrate, (b) planar Au film on glass, and (c) Au foil
after flame annealing. (d) Zoom-in of (200) peak where planar Au and np-Au have been increased by 20x. From the XRD it is
evident that the planar Au film is highly oriented in the (111) orientation. The data also shows that the full width half max of the
np-Au is much larger than that of the Au foil, indicating that the np-Au has smaller grains.
To estimate the average grain size of the np-Au films, we performed X-ray diffraction (XRD) on a RTetched sample of ~800 nm thickness (Figure 2.2a). XRD spectra were taken using an X’PERT-PRO MRD Serial
# DY3178 made by PANalytic. The scan went from 30° to 120°. The voltage was 45 kV, current 40 mA and

26
the beam attenuator was Ni 0.125 mm automatic. For comparison, we also examined a 50 nm-thick planar Au
base layer, and a commercial Au foil (Alfa Aesar 99.9975%) (Figures 2.2b,c). These data show that the average
full-width half maximum of the diffraction peaks from np-Au are larger than the Au foil (Figures 2.1d).
According to the Scherrer equation,23 the np-Au film and Au foil have an average grain size of 20 ± 4 nm and
77 ± 23 nm, respectively. These calculations assume a shape factor of unity and do not take into account the
possibility of microstrain.24 We also performed transmission electron microscopy (TEM) to directly visualize the
distribution of grain boundaries within individual ligaments of the np-Au film. Consistent with prior reports,25–
27

we observed many grain boundaries along the surface of the curved Au ligaments (Figure 2.3). Dark-field

TEM images were also collected to estimate the average grain size within the np-Au film (Figure 2.4 and Figure
2.5). The average grain size that we observed in TEM (17 ± 4 nm) is very similar with the estimate obtained
through analysis of the XRD data. It is known that grain boundaries and under-coordinated sites in Au
electrocatalysts serve as the active sites for CO2 reduction,28,29 suggesting that these np-Au films should exhibit
substantial activity for CO2 reduction.

Figure 2.3: Bright field transmission electron microscopy (TEM) images of RT np-Au film. The red arrows denote grain
boundaries. All scale bars represent 5 nm.

27

Figure 2.4: Transmission electron microscopy (TEM) analysis of RT np-Au films. (a) Bright-field TEM of a particular region
of the np-Au film along with (b) the corresponding selected area electron diffraction (SAED) pattern. (c1-c7) Dark-field TEM
images from the np-Au film obtained from the particular spots numbered in panel b. (d) Bright-field TEM image of np-Au
film along with dark-field TEM images numbered 1-3. All scale bars in all TEM images represent 20 nm.

28
Figure 2.5: Transmission electron microscopy (TEM) analysis of RT np-Au films. (a-c) Bright-field TEM of a particular region
of the np-Au film along with (d-f) a corresponding dark-field TEM image obtained from that particular region of the np-Au film.
All scale bars in all TEM images represent 20 nm.
2.3

Electrochemical characterization
A two-compartment electrochemical cell made of polyether ether ketone (PEEK) was used to perform

CO2RR experiments on the np-Au films. A volume of 2 mL of 50 mM K2CO3 was added to each compartment,
which were separated by an anion exchange membrane (AGC, Selemion AMV). The np-Au film served as the
working electrode with a Ag/AgCl leakless reference electrode and a Pt foil counter electrode. The Pt foil
(99.99% , 0.05 mm thick, Alfa Aesar) was soaked in 10 wt.% HNO3 for 1 h and then flame annealed to remove
contaminants before each experiment. The flame annealing process entails holding a flame to the foil until it
glows red then rinsing the foil in water and drying. This process is repeated twice. The same procedure was
also applied to the Au foil before testing. CO2 saturated 50 mM K2CO3 (pH 6.8) was prepared by bubbling CO2
(Research grade from Airgas) into the electrolyte for 30 min prior to experiments. It is known that higher
supporting electrolyte concentrations provide higher current densities during electrolysis,30–32 but we found this
concentration sufficient to enable reliable evaluation of our electrodes. Each electrolyte compartment was
bubbled with CO2 at a rate of 5 SCCM through a fine glass dispersion frit to maximize the speed of delivery of
CO2 into solution. The outflowing gas was sent through a flow meter to check that the flow of CO2 in and out
of the cell was the same, ensuring that it was thoroughly sealed against gas leaks. The outflowing gas was sent
through a vapor trap to remove all water from the air before it was fed into a (SRI-8610) gas chromatograph.
All experiments were performed at room temperature with a Biologic VSP-300 potentiostat. All
potentials were converted to the reversible hydrogen electrode (RHE) scale using the equation: E vs. RHE = E
vs. Ag/AgCl + 0.197 V + 0.059 V pH-1 × solution pH. Before each experiment, potentiostatic electrochemical
impedance spectroscopy (PEIS) was performed to determine the solution resistance of the cell, which was
typically between 30 – 60 W. The applied electrochemical potential was then compensated by 85% using iR
compensation of the potentiostat. The electrochemical cell was dismounted and rinsed multiple times after each
experiment and then stored in 10 wt.% HNO3. Before using the cell for the next experiment, it was sonicated
for 10 min in water at least 4 times.
To analyze the chemical products, the electrode was held at the desired potential for at least 2 h allowing
for the completion of eight gas chromatography measurements. The gas chromatograph (SRI-8610) used a
Hayesep D column and a Molsieve 5A column with N2 as the carrier gas. The gaseous products were detected

29
using a thermal conductivity detector (TCD) for CO detection and a flame ionization detector (FID) equipped
with a methanizer for H2 detection. Quantitative analysis of gaseous products was based on calibration with
several gas standards over many orders of magnitude in concentration. To measure liquid products, the
electrolyte on the anode and cathode were sampled at the end of the run and tested with high performance liquid
chromatography (HPLC). However, no liquid products were ever observed. Between different potential
experiments all of the electrolyte was removed and then the cell was rinsed three times with water before new
electrolyte was add and bubbled with CO2.

Figure 2.6: Electrochemical performance of Au cathodes. Faradaic efficiency (FE) for CO (filled bars) and H2 (open bars) as a
function of applied potential (E) with (a) 661 ± 10 nm thick room temperature etched (RT) np-Au film, (b) 664 ± 5 nm thick
low temperature etched (LT) np-Au film, (c) planar Au film, and (d) commercial Au foil. Partial current density (J) for CO (filled
circles) and H2 (open circles) as a function of applied potential for (e) 661 ± 10 nm thick RT np-Au film, (f) 664 ± 5 nm thick
LT np-Au film, (g) planar Au film, and (h) commercial Au foil. Each data point represents the average FE for CO or H2 obtained
over 2-3 h of continuous electrolysis at the indicated potential with iR compensation. The partial current densities also represent the
average value observed over the same time period. All data was obtained from the same electrode along the potential sweep.
Figure 2.6a shows the Faradaic efficiency (FE) of a RT np-Au film (661 ± 10 nm thick) for both CO
(filled blue bars) or H2 (white bars) as a function of the applied potential (E) from −0.3 VRHE to −1.1 VRHE (V
vs. RHE). Each data point shown in Figure 2.6 represents the average FE for CO or H2 obtained over 2-3 h of
continuous electrolysis at the indicated potential with iR compensation. All data was obtained from the same

30
electrode along the potential sweep. The RT np-Au film exhibits a maximum FE for CO of 90% at E = −0.5
VRHE with the remainder of the current producing H2. We note that no liquid products were detected for any of
the Au electrodes studied. Notably, the LT np-Au film (664 ± 5 nm thick) obtains a maximum FE for CO (filled
grey bars) of 99% at E = −0.5 VRHE and maintains at least 80% FE for CO from −0.3 VRHE to −0.7 VRHE before
the HER (white bars) begins to account for a larger portion of the products at more negative applied potentials
(Figure 2.6b). To examine the influence of the np-Au morphology on CO2 reduction selectivity, we tested the
activity of a 50 nm-thick planar Au film, which is the base Au layer of the np-Au electrodes. As shown in Figure
2.6c, the planar Au film primarily produces H2 (white bars) across the entire potential window; the FE for CO
production (green bars) only reaches ~40% at −0.5 VRHE. We also evaluated the activity of a commercial Au foil
(99.9975%, 0.1 mm thick, Alfa Aesar) to confirm that our experimental conditions and cell configuration are
capable of adequately reproducing commonly observed activity trends for Au films.33 As shown in Figure 2.6d,
the Au foil obtained a maximum FE for CO of 92% at −0.5 VRHE (filled red bars), consistent with prior reports.33

Figure 2.7: (a) Cyclic voltammetry of LT np-Au (grey curve) and RT np-Au (dark blue curve) in 0.5 M H2SO4 obtained at a
scan rate of 50 mV s-1. From these data we can determine that the LT sample has ~3x greater electrochemical surface area as
compared to the RT sample. (b) shows a histogram of pore widths measured on these two samples at three different locations on the
sample. Representative SEM cross-section images of (c) RT np-Au and (d) LT np-Au films. The scale bars on both images

31
correspond to 100 nm. These images correspond to the actual electrodes used for electrochemical tests involving different electrolyte
concentrations.
A significant advantage of the np-Au morphology over the planar Au electrodes is illustrated by the high
partial current density for CO (JCO) relative to H2 (Figures 2.6e-h). At an applied potential of −0.7 VRHE, the LT
np-Au film exhibits a peak JCO of 8.1 mA cm-2 (Figure 2.6f), which is four times higher than the Au foil (Figure
2.6h) and eight times higher than the planar Au film (Figure 2.6g). At the optimum applied potential for CO
production (−0.5 VRHE), the LT np-Au film displays JCO of nearly 6.2 mA cm-2 while the RT np-Au film JCO is
around 4.5 mA cm-2. Interestingly, the LT np-Au film shows only a slight increase in JCO as compared to the RT
np-Au despite the ~3 times increase in surface area between the LT and RT np-Au films (Figure 2.7a). This
lower than expected JCO from LT np-Au films likely arises due to mass transport limitations, whereby the
geometry of the electrochemical cell does not allow for sufficient delivery of CO2 throughout the porous
electrode to keep up with the electrochemically active surface area of the entire film. This hypothesis is confirmed
by comparing the FE and JCO for RT-etched samples that were ~800 nm and ~150 nm thick (Figure 2.8).
Despite the drastic difference in electrochemical surface area as determined by Cu UPD (Figure 2.1d), the thicker
film only showed a ~30% increase in JCO under our experimental conditions (Figure 2.8c). Note that the linear
relationship between surface area of the np-Au film and alloy thickness (Figure 2.1d) implies that the entire
network is accessible to the electrolyte, suggesting that a large fraction of dissolved CO2 does not penetrate the
entire depth of the film at the current densities studied. These results suggest that to better use the full
electrocatalytic surface area of np-Au for CO2 reduction requires that the geometry of the cell be modified to
flow the CO2 directly through the np-Au film so that CO2 is efficiently delivered to the catalyst, as opposed to
simply flowing the CO2 past the electrode surface. Indeed, it has recently been shown that such a tactic is highly
beneficial for improving the rate of electrochemical CO2 reduction.34,35

Figure 2.8: (a) SEM cross-section image of ~150 nm-thick RT np-Au film. The scale bar represents 100 nm. (b) Faradaic
efficiency as a function of applied potential (E) for 150 nm-thick RT np-Au film. (c) Partial current density for CO (JCO) from a

32
150 nm-thick and ~800 nm-thick RT np-Au sample. Considering the 4x smaller surface area of the thinner film, the relative JCO
between the two films is unexpected. We hypothesize that this is due to mass transport limitations.
Both the RT and LT np-Au films exhibit superior FE for CO (Figures 2.6a,b) relative to the planar Au
film or Au foil (Figures 2.6c,d) across the entire potential window studied. It has previously been shown that the
residual Ag in the np-Au film is not the source of the high FE for CO.18 We therefore attribute such significant
improvements in catalytic selectivity to the prevalence of grain boundaries that exist within the np-Au structure
relative to the planar Au film. Another factor that likely contributes to such marked improvements in selectivity
is the ability of the np-Au film to support locally-alkaline pH conditions within the porous network as protons
are consumed during electrolysis. Such an effect has previously been observed in mesoporous Au electrodes,
which serves to suppress the rate of HER while the rate of CO2 reduction is relatively unaffected.20
2.4

Influence of local pH on performance

Figure 2.9: Influence of electrolyte concentration on CO2 reduction selectivity with Au cathodes. (a-c) The Faradaic efficiency for
CO as a function of applied potential (E) obtained at two different electrolyte concentrations (both saturated with CO2) for (a) 809
± 15 nm thick room temperature etched (RT) np-Au film, (b) 821 ± 22 nm thick low temperature etched (LT) np-Au film, and
(c) 50 nm thick planar Au film. (d-f) The corresponding average current density (J) obtained at the applied potential (E) observed

33
at two different electrolyte concentrations for (d) RT np-Au, (e) LT np-Au, and (f) planar Au film. (g-i) Predicted solution pH at
the surface of the electrode for (g) RT np-Au, (h) LT np-Au, and (i) planar Au film. A planar electrode geometry is assumed for
the simulations.
To examine the influence of local pH gradients within the nanoporous network on the selectivity of npAu films, we examined the electrochemical activity of a RT and a LT np-Au film (Figures 2.9d, e) at two different
electrolyte concentrations (50 mM K2CO3 and 200 mM K2CO3 – both fully saturated with CO2). Increasing the
electrolyte concentration increases the buffering capacity of the solution,20,32 which reduces any swings in local
pH that are anticipated to form within the pores of the np-Au films as protons are consumed during electrolysis.
It was therefore anticipated that the np-Au films would exhibit reduced FE for CO in 200 mM K2CO3 electrolyte
if an increased solution pH within the porous network was responsible for the high selectivity observed on the
np-Au films. As shown in Figure 2.9a and b, the selectivity on both RT np-Au and LT np-Au is essentially
unchanged at low applied potentials (−0.3 VRHE), but as the current density increases with increased applied bias
(Figures 2.9d,e), any pH gradient that may form within the np-Au film in the 50 mM K2CO3 electrolyte (Figures
2.9a,b circles) is diminished by the improved buffering capacity of the 200 mM K2CO3 electrolyte (Figures 2.9a,b
diamonds). In contrast, no change in FE for CO is observed on a planar Au film at any applied potential (Figure
2.9c), confirming that the change in selectivity observed on the np-Au electrodes is not simply a consequence
of the increased electrolyte concentration (Figures 2.9c,f). These results strongly suggest that a pH gradient is
forming within the porous network of the Au electrocatalyst and serves an important role in achieving high
selectivity of CO2 reduction in aqueous electrolytes.
Interestingly, the LT np-Au film shows a larger reduction in FE for CO (Figure 2.9a) than the RT npAu film (Figure 2.9b). Analysis of the pore sizes between these two samples reveals that the LT np-Au film has
pores with an average diameter of 10 ± 2 nm while the RT np-Au sample has pores with an average diameter
of 28 ± 8 nm (Figure 2.7b). We therefore attribute the improved selectivity of the LT np-Au film relative to the
RT np-Au film to the smaller pore size of the former (~10 nm) relative to the latter (~30 nm), which more
effectively supports a high local pH within the porous network that improves the selectivity for CO2RR relative
to HER (Figures 2.6a,b). Previous work has shown that increasing the thickness of a mesoporous Au catalyst to
~2 µm helps to achieve a similar effect within ~200 nm pores.20 Our observations suggest that further reducing
the pore volume below 10 nm may enable realizing such an effect within even thinner nanoporous metal films
than those studied here.
To further explore whether the pH gradient is developed within the porous Au network or occurs largely
in the boundary layer, we simulated the pH profile as a function of distance away from the electrode surface

34
using the model previously reported by Gupta, et al.30 Briefly, the model assumes a planar electrode geometry,
which is a valid assumption for calculating the pH at the outer surface of the electrode because the flux of
reactants and products must be the same for either a porous or planar electrocatalyst at this location. The
assumption of a planar electrode is clearly incapable of accounting for changes in the transport of reactants and
products into and out of the porous film itself, and we therefore interpret any experimental deviations from the
model to originate from changes occurring within the porous network of metal ligaments. The inputs into the
model are the electrolyte concentration, the total current density, and the Faradaic efficiency for CO and H2. A
70 µm-thick boundary layer was assumed based on the experimental flow rate of CO2 of 5 SCCM through the
catholyte.31 As shown in Figures 2.9g-i, these calculations predict very little change in local pH at the electrode
surface between the two electrolyte concentrations, albeit small deviations from the bulk electrolyte are predicted
for the RT np-Au film and the planar Au electrode (Figures 2.9g,i). While significant reductions in FE for CO
were observed on both the RT and LT np-Au films (Figures 2.9a and 2.9b, respectively), no change in FE was
observed experimentally on the planar Au film (Figure 2.9c). This obvious contradiction between the results of
experiment with those from the model indicates that the local pH changes must be occurring within the porous
network itself. Otherwise, we would have observed a similar reduction in FE for CO with the planar electrode
at the higher electrolyte concentration. We note that these experimental observations are consistent with a
previous report on mesoporous Au films,20 yet are achieved with much thinner films. Taken together, these
results indicate that the pore diameter of porous metal electrocatalysts is a critical parameter for optimizing their
selectivity, and suggest that control over the pore size on the nanometer length scale may offer further
improvements in electrochemical selectivity.
2.5

Catalyst Stability

35

Figure 2.10: Extended electrochemical stability data for Au cathodes. The Faradaic efficiency for CO (filled circles) and H2 (open
circles) was measured every 15 min via gas chromatography over the course of 24 h at an applied potential of E = –0.5 VRHE with
iR compensation for (a) room temperature (RT)-etched np-Au, (b) planar Au film, and (c) Au foil.
We further evaluated the electrocatalytic stability of these Au films for the CO2RR at an applied potential
of E = –0.5 VRHE (with iR compensation). Significantly, we observed that the np-Au film maintained a high FE
for CO (nearly 90%) over the course of 24 h of continuous electrocatalytic testing (Figure 2.10a). In stark
contrast, the Au foil and planar Au films exhibit drastic reductions in FE for CO over just one day of testing at
the same applied potential (Figures 2.10b,c). Continued testing of a different RT np-Au film for 4.5 days (110
h) showed continued catalytic stability (Figure 2.11). Comparison of the SEM images before and after testing
show no significant changes in morphology except that the np-Au ligaments appear to coarsen slightly (Figure
2.12). Analysis of these films by XRD indicates no significant changes in peak width before and after testing,
but all films showed a decrease in the overall signal magnitude from diffraction peaks associated with high-index

36
reflections (Figure 2.13). We note that the activity of the Au foil can be recovered if the flame anneal treatment
is repeated, but such a process is undesirable as it hinders long-term continual operation under CO2RR
conditions. These observations serve to highlight the benefit of using the nanoporous metal structure to perform
CO2 reduction: the prevalence of grain boundaries offer numerous active sites on the metal ligaments while the
porous network is able to support a locally-alkaline pH within the film that helps improve electrocatalytic
selectivity for the CO2RR over the HER.

Figure 2.11: Extended electrochemical stability data for a RT np-Au film (~800 nm thick). The Faradaic efficiency for CO
(filled points) and H2 (open points) was measured every 15 min via gas chromatography over the course of 110 h at an applied
potential of E = –0.5 VRHE with iR compensation.

Figure 2.12: (a,b) SEM images of a ~800 nm thick RT np-Au film (a) before and (b) after testing for 110 h at −0.5 V vs.
RHE. (c,d) SEM images of a planar Au film (c) before and (d) after testing for 24 h at −0.5 V vs RHE. (e,f) SEM images of
a Au foil (e) before and (f) after testing for 24 h at −0.5 V vs RHE. There is no visible difference between any of the planar

37
samples before and after testing. In the np-Au sample there is some minor coarsening of the ligaments, but no significant changes to
the film morphology are observed.

Figure 2.13: XRD spectra of Au films before and after 24 h of testing for (a) ~800 nm thick RT np-Au film, (b) planar Au
film, and (c) Au foil. The peak at 68° in the RT np-Au film and the planar Au film is due to the Si substrate. In Figure 2.2, the
XRD patterns were collected from films supported on a glass substrate to avoid the peak from the Si substrate. Negligible differences
were observed between Au peaks obtained from on Si vs. glass substrates.
2.6

Conclusions
In conclusion, we have demonstrated that np-Au films constitute a promising electrocatalytic

architecture for CO2 reduction to yield CO in aqueous electrolytes. The np-Au films exhibit a maximum Faradaic
efficiency for CO of 99% at −0.5 VRHE while operating at a partial current density for CO in excess of 6 mA cm2

. We attribute the catalytic performance of np-Au to its high electrochemical surface area possessing a large

number of grain boundaries and its ability to support a local depletion of protons within the porous network.
Significantly, these np-Au films maintain a Faradaic efficiency of greater than 80% over the course of 110 h of
continuous electrolysis at −0.5 VRHE, while the activity and selectivity of both planar Au films and Au foils
diminishes significantly over much shorter periods of operation (~4 h). These studies highlight the benefits of
nanoporous metal cathodes for CO2 reduction and indicate that the pore size is an important parameter to
control for improving selectivity in these promising electrocatalytic architectures.

38

BIBLIOGRAPHY CHAPTER 2
1. Jones, J.-P.; Parkash, G. K. S.; Olah, G. A. Electrochemical CO2 Reduction: Recent Advances and
Current Trends. Isr. J. Chem. 2014, 54, 1451–1466.
2. White, J. L.; Baruch, M. F.; Pander, J. E.; Hu, Y.; Fortmeyer, I. C.; Park, J. E.; Zhang, T.; Liao, K.;
Gu, J.; Yan, Y.; Shawn, T. W.; Abelev, E.; Boarsly, A. B. Light-Driven Heterogeneous Reduction of
Carbon Dioxide: Photocatalysts and Photoelectrodes. Chem. Rev. 2015, 115, 12888–12935.
3. Whipple, D.; Kenis, P. Prospects of CO2 Utilization via Direct Heterogeneous Electrochemical
Reduction. J. Phys. Chem. Lett. 2010, 1, 3451–3458.
4. Kuhl, K. P.; Hatsukade, T.; Cave, E. R.; Abram, D. N.; Kibsgaard, J.; Jaramillo, T. F. Electrocatalytic
Conversion of Carbon Dioxide to Methane and Methanol on Transition Metal Surfaces. J. Am.
Chem. Soc. 2014, 136, 14107–14113.
5. Kuhl, K. P.; Cave, E. R.; Abram, D. N.; Jaramillo, T. F. New Insights into the Electrochemical
Reduction of Carbon Dioxide on Metallic Copper Surfaces. Energy Environ. Sci. 2012, 5, 7050–7059.
6. Hori, Y. Electrochemical CO2 Reduction on Metal Electrodes. In Modern Aspects of Electrochemistry;
Vayenas, C. G., White, R. E., Gamboa-Aldeco, M. E., Eds.; Springer New York: New York, NY,
2008; pp 89–189.
7. Cao, Z.; Kim, D.; Hong, D.; Yu, Y.; Xu, J.; Lin, S.; Wen, X.; Nichols, E. M.; Jeong, K.; Reimer, J. A.;
Yang, P.; Chang, C. J. A Molecular Surface Functionalization Approach to Tuning Nanoparticle
Electrocatalysts for Carbon Dioxide Reduction. J. Am. Chem. Soc. 2016, 138, 8120–8125.
8. Fang, Y.; Flake, J. C. Electrochemical Reduction of CO2 at Functionalized Au Electrodes. J. Am.
Chem. Soc. 2017, 139, 3399–3405.
9. Zhu, W.; Michalsky, R.; Metin, Ö.; Lv, H.; Guo, S.; Wright, C. J.; Sun, X.; Peterson, A. A.; Sun, S.
Monodisperse Au Nanoparticles for Selective Electrocatalytic Reduction of CO2 to CO. J. Am.
Chem. Soc. 2013, 135, 16833–16836.
10. Zhu, W.; Zhang, Y.-J.; Zhang, H.; Lv, H.; Li, Q.; Michalsky, R.; Peterson, A. A.; Sun, S. Active and
Selective Conversion of CO2 to CO on Ultrathin Au Nanowires. J. Am. Chem. Soc. 2014, 136,
16132–16135.
11. Chen, Y.; Li, C. W.; Kanan, M. W. Aqueous CO2 Reduction at Very Low Overpotential on OxideDerived Au Nanoparticles. J. Am. Chem. Soc. 2012, 134, 19969–19972.
12. Liu, M.; Pang, Y.; Zhang, B.; De Luna, P.; Voznyy, O.; Xu, J.; Zheng, X.; Dinh, C. T.; Fan, F.; Cao,
C.; de Arquer, F. P. G.; Safaei, T. S.; Mepham, A.; Klinkova, A.; Kumacheva, E.; Filleter, T.; Sinton,
D.; Kelley, S. O.; Sargent, E. H.Enhanced Electrocatalytic CO2 Reduction via Field-Induced Reagent
Concentration. Nature 2016, 537, 382.
13. Lu, Q.; Rosen, J.; Zhou, Y.; Hutchings, G. S.; Kimmel, Y. C.; Chen, J. G.; Jiao, F. A Selective and
Efficient Electrocatalyst for Carbon Dioxide Reduction. Nat. Commun. 2014, 5, 3242.
14. Dutta, A.; Morstein, C. E.; Rahaman, M.; Cedeño López, A.; Broekmann, P. Beyond Copper in CO2
Electrolysis: Effective Hydrocarbon Production on Silver-Nanofoam Catalysts. ACS Catal. 2018,
8357–8368.
15. Sen, S.; Liu, D.; Palmore, G. T. R. Electrochemical Reduction of CO2 at Copper Nanofoams. ACS
Catal. 2014, 4, 3091–3095.
16. Ma, M.; Trześniewski, B. J.; Xie, J.; Smith, W. A. Selective and Efficient Reduction of Carbon
Dioxide to Carbon Monoxide on Oxide-Derived Nanostructured Silver Electrocatalysts. Angew.
Chem. Int. Ed. 2016, 55, 9748–9752.
17. Zhang, Y.; Luc, W.; Hutchings, G. S.; Jiao, F. Photoelectrochemical Carbon Dioxide Reduction
Using a Nanoporous Ag Cathode. ACS Appl. Mater. Interfaces 2016, 8, 24652–24658.

39
18. Zhang, W.; He, J.; Liu, S.; Niu, W.; Liu, P.; Zhao, Y.; Pang, F.; Xi, W.; Chen, M.; Zhang, W.; et al.
Atomic Origins of High Electrochemical CO2 Reduction Efficiency on Nanoporous Gold.
Nanoscale 2018, 10, 8372–8376.
19. Chen, C.; Zhang, B.; Zhong, J.; Cheng, Z. Selective Electrochemical CO2 Reduction over Highly
Porous Gold Films. J. Mater. Chem. A 2017, 5, 21955–21964.
20. Hall, A. S.; Yoon, Y.; Wuttig, A.; Surendranath, Y. Mesostructure-Induced Selectivity in CO2
Reduction Catalysis. J. Am. Chem. Soc. 2015, 137, 14834–14837.
21. Biener, J.; Biener, M. M.; Madix, R. J.; Friend, C. M. Nanoporous Gold: Understanding the Origin of
the Reactivity of a 21st Century Catalyst Made by Pre-Columbian Technology. ACS Catal. 2015, 5,
6263–6270.
22. Qian, L. H.; Chen, M. W. Ultrafine Nanoporous Gold by Low-Temperature Dealloying and Kinetics
of Nanopore Formation. Appl. Phys. Lett. 2007, 91, 083105.
23. Patterson, A. L. The Scherrer Formula for X-Ray Particle Size Determination. Phys. Rev. 1939, 56,
978–982.
24. Jiang, H. G.; Rühle, M.; Lavernia, E. J. On the Applicability of the X-Ray Diffraction Line Profile
Analysis in Extracting Grain Size and Microstrain in Nanocrystalline Materials. J. Mater. Res. 1999,
14, 549–559.
25. Hodge, A. M.; Biener, J.; Hsiung, L. L.; Wang, Y. M.; Hamza, A. V.; Satcher, J. H. Monolithic
Nanocrystalline Au Fabricated by the Compaction of Nanoscale Foam. J. Mater. Res. 2005, 20, 554–
557.
26. Mathur, A.; Erlebacher, J. Size Dependence of Effective Young’s Modulus of Nanoporous Gold.
Appl. Phys. Lett. 2007, 90, 061910.
27. Petegem, S. V.; Brandstetter, S.; Maass, R.; Hodge, A. M.; El-Dasher, B. S.; Biener, J.; Schmitt, B.;
Borca, C.; Swygenhoven, H. V. On the Microstructure of Nanoporous Gold: An X-Ray Diffraction
Study. Nano Lett. 2009, 9, 1158–1163.
28. Mariano, R. G.; McKelvey, K.; White, H. S.; Kanan, M. W. Selective Increase in CO2
Electroreduction Activity at Grain-Boundary Surface Terminations. Science 2017, 358, 1187.
29. Feng, X.; Jiang, K.; Fan, S.; Kanan, M. W. Grain-Boundary-Dependent CO2 Electroreduction
Activity. J. Am. Chem. Soc. 2015, 137, 4606–4609.
30. Gupta, N.; Gattrell, M.; MacDougall, B. Calculation for the Cathode Surface Concentrations in the
Electrochemical Reduction of CO2 in KHCO3 Solutions. J. Appl. Electrochem. 2006, 36, 161–172.
31. Clark, E. L.; Resasco, J.; Landers, A.; Lin, J.; Chung, L.-T.; Walton, A.; Hahn, C.; Jaramillo, T. F.;
Bell, A. T. Standards and Protocols for Data Acquisition and Reporting for Studies of the
Electrochemical Reduction of Carbon Dioxide. ACS Catal. 2018, 8, 6560–6570.
32. Singh, M. R.; Clark, E. L.; Bell, A. T. Effects of Electrolyte, Catalyst, and Membrane Composition
and Operating Conditions on the Performance of Solar-Driven Electrochemical Reduction of
Carbon Dioxide. Phys. Chem. Chem. Phys. 2015, 17, 18924–18936.
33. Cave, E. R.; Montoya, J. H.; Kuhl, K. P.; Abram, D. N.; Hatsukade, T.; Shi, C.; Hahn, C.; Nørskov,
J. K.; Jaramillo, T. F. Electrochemical CO2 Reduction on Au Surfaces: Mechanistic Aspects
Regarding the Formation of Major and Minor Products. Phys. Chem. Chem. Phys. 2017, 19, 15856–
15863.
34. Dinh, C. T.; Burdyny, T.; Kibria, M. G.; Seifitokaldani, A.; Gabardo, C. M.; García de Arquer, F. P.;
Kiani, A.; Edwards, J. P.; De Luna, P.; Bushuyev, O. S.; Zou, C.; Quintero-Bermudez, R.; Pang, Y.;
Sinton, D.; Sargent, E. H. CO2 Electroreduction to Ethylene via Hydroxide-Mediated Copper
Catalysis at an Abrupt Interface. Science 2018, 360, 783.
35. Hoang, T. T. H.; Ma, S.; Gold, J. I.; Kenis, P. J. A.; Gewirth, A. A. Nanoporous Copper Films by
Additive-Controlled Electrodeposition: CO2 Reduction Catalysis. ACS Catal. 2017, 7, 3313–3321.

40
Chapter 3

OPERANDO DETERMINATION OF CATALYTICALLY-ACTIVE REGIONS IN
NANOPOROUS GOLD GAS DIFFUSION ELECTRODES FOR HIGHLY
SELECTIVE CARBON DIOXIDE REDUCTION
3.1

Introduction
The electrochemical reduction of carbon dioxide (CO2R) to value-added commodities represents a

promising means to store renewable electricity and create a closed carbon cycle for industrial chemicals.1–3 CO2R
can produce a wide variety of products based on catalyst choice,4–6 structuring,4,7 and treatment.2,8 Of these
products, carbon monoxide (CO) is especially interesting due to its ubiquitous role as a precursor in cornerstone
chemical processes such as Fischer-Tropsch reactions,9–11 hydroformylation of alkenes to aldehydes,12 methanol
production,10,13 and metal refinement.14 In addition, CO2R to CO is a two-electron process which translates to
lower energy inputs as compared to other multi-electron products.10,15 CO production has been demonstrated
with high selectivities, with Faradaic efficiencies (FE) for CO reaching >90% in aqueous-based electrolyzers.7,8
While current densities between -10 mA/cm2 and -20 mA/cm2 can be achieved through electrode nanostructuring,16–20 further improvements in traditional aqueous systems are limited due to low CO2 solubility and
long diffusion lengths.21,22 In order for electrochemical CO2R to be economically viable, current densities of
greater than -100 mA/cm2 are required.23,24
Gas diffusion electrodes (GDEs) present an alternative electrode design capable of addressing mass
transport and CO2 solubility issues.25 In a GDE, a blend of CO2 gas and water vapor is flowed across a porous
catalyst layer in contact with a liquid electrolyte.21,25 High current densities have been achieved by GDEs tuned
for high performance in a range of different configurations such as alkaline electrolyzers and membrane
electrode assemblies.26–29

Scheme 1
Flooded

41

Wetted

Dry

Electrolyte
Nanoporous Au
Microporous Layer
Gas Diffusion Layer
Figure 3.1: A schematic of the nanoporous gold gas diffusion electrode used in this study in a vapor CO2 fed device. The bottom
image shows the electrode structure. The support consists of a gas diffusion layer composed of carbon fibers, on top of which is coated
carbon black and PTFE which makes up the microporous layer. The nanoporous gold is coated on top of this. The top three panels
show the different configurations that the nanoporous gold can be in during operation - flooded, wetted, and dry.
The large disparity in current density between an aqueous electrolyzer and a GDE can be attributed to
overcoming the mass transport limits in an aqueous system. In an aqueous electrolyzer, CO2 must first dissolve
into the electrolyte then diffuse to the electrode surface. This process is mass transported limited by the rate at
which CO2 can dissolve into the electrolyte and this precludes the total current density achievable to a regime
usually bounded by a performance of less than -20 mA/cm2. The catalyst in an aqueous electrolyzer can be
considered to be fully flooded. The term flooded means that the catalyst is fully submerged in the electrolyte.
An idealized GDE, on the other hand, advantageously forms a triple phase boundary in which a meniscus of
water wraps the catalyst, and channels of gas flow through the catalyst layer (Figure 3.1). This reduces the
diffusion length and substantially increases the CO2 transport rate to the catalyst surface, thus enabling much
higher current densities in the range of -100 of mA/cm2 to -1000 of mA/cm2 in ultra high preforming devices.
We refer to this GDE state as wetted. There is also the potential that portions of a catalyst in a GDE are dry

42
due to excess gas pressure thus rendering them inactive (Figure 3.1). While in an ideal scenario all of the catalyst
in the GDE would be wetted, it likely that a combination of flooded, wetted, and dry conditions exist
simultaneously. The interaction between the catalyst and the electrolyte has been probed computationally;21,25
however, there is little experimental understanding of this layer. Here we seek to understand the relationship
between device performance and catalyst wetting.
It has been previously demonstrated that nanoporous gold (np-Au) is a promising CO2 to CO catalyst.7
In aqueous CO2 fed systems, np-Au has shown improved selectivity and catalytic activity relative to planar Au
(pl-Au) due to the high density of under-coordinated sites,30 prevalence of grain boundaries,30,31 and high surface
area.32,33 Furthermore, pH gradients are built up within the porous structure and this increased pH suppresses
the parasitic hydrogen evolution reaction (HER);34 however, these experiments were only demonstrated in the
aqueous CO2 fed configuration. Here, np-Au GDEs are used as a model catalyst system for a vapor CO2 fed
CO2R device. We demonstrate that np-Au GDEs achieve >95% selectivity for CO at partial current densities
for CO production (JCO) greater than -150 mA/cm2. One of the outstanding questions in the field of GDEs is
understanding what portion of the catalyst layer is meaningfully contributing to active catalysis. We take
advantage of the highly porous nature of our electrode to determine via, scanning electron microscopy (SEM),
copper under potential deposition (Cu UPD), secondary ion mass spectroscopy (SIMS), and electrochemical
product analysis to determine which fraction of the catalyst is flooded/wetted, and dry. We find that only 56%
of the available catalyst is active under vapor CO2 fed conditions and that the bottom 75% of the catalyst layer
exhibits the largest difference in wetting between an aqueous CO2 fed vs vapor CO2 fed system. These
investigations yield improved methods of in situ catalyst characterization which can facilitate the optimization
and adoption of CO2R electrolyzers on an industrially-relevant scale.
3.2

Fabrication and characterization of nanoporous gold
The np-Au GDEs used in our studies were fabricated by electron beam co-deposition of Au and Ag

targets onto the microporous side of a carbon paper substrate (Sigracet 38BC) to produce a AgxAu1-x alloy of
uniform distribution. The relative atomic percent of Au relative to Ag is tuned by varying the deposition rates
of each metal. The np-Au electrode was produced by soaking the AgxAu1-x alloy in concentrated nitric acid (70%
weight/volume) for 15 minutes at room temperature. Concentrated nitric acid dealloys the Ag to produce the
desired np-Au morphology. (See Figure 5.2 for SEM images of the fabrication process).

43

Bare Sigracet 38BC

Au/Ag alloy

300 nm

300 nm

300 nm

500 nm

500 nm

300 nm

Post electrochemistry

After HNO3 etch

500 nm

500 nm

500 µm

500 µm

500 µm

500 µm

Figure 3.2: shows SEM images of the electrode at different phases in the fabrication process. (a)-(c) shows images of the bare carbon
paper, Sigracet 38BC. (d)-(f) show images of the gold silver alloy on the carbon paper. (g)-(i) show images of the nanoporous gold
morphology from a 35% Au alloy that forms after the nitric acid etch.
15 atomic % Au

25 atomic % Au

150 nm

150 nm

600 nm

3 µm

150 nm

150 nm

600 nm

600 nm

600 nm

3 µm

45 atomic % Au

35 atomic % Au

3 µm

3 µm

44
Figure 3.3: SEM characterization of nanoporous gold (np-Au) electrodes with a varying gold atomic percent (%Au) of 15%Au (ac), 25%Au (d-f), 35%Au (g-i), and 45%Au (j-l).
Figure 1
Planar Au

a)

b)

Faradaic
Efficiency (%)

25%Au np-Au

c)

600 nm

600 nm
e) 100

15%Au np-Au

d)

600 nm

35%Au np-Au

600 nm

f)

g)

h)

j)

k)

l)

80
60
40
20

J (mA/cm2)

i)

-40
-80
-120
-160
-1

CO
H2
-0.8 -0.6 -0.4
E (V vs RHE)

CO
H2
-1

-0.8 -0.6 -0.4
E (V vs RHE)

CO
H2
-1

-0.8 -0.6 -0.4
E (V vs RHE)

CO
H2
-1

-0.8 -0.6 -0.4
E (V vs RHE)

Figure 3.4: SEM characterization of planar gold (pl-Au) and nanoporous gold (np-Au) electrodes with varying gold atomic percent
(%Au) (a-d). CO2R performance of the pl-Au and np-Au electrodes is shown via Faradaic efficiencies (e-h) and partial current
densities (i-l). For all plots, CO is denoted in pink and H2 by blue. Each data point is the average of three distinct electrodes.
Representative, top-down SEM images for a range of 300 nm thick np-Au samples of varying gold
atomic percent (%Au) from 15%Au to 35%Au and a pl-Au control are shown in Figure 3.3 and Figure 3.4a-d. Clear
morphological distinctions can be observed between the samples of different %Au. The electrodes consist of a
three dimensional network of gold ligaments, pores, and cracks. The cracks in the np-Au morphology are due
to the volume contraction (10-30%) that occurs from the removal of Ag.35 The crack sizes notably increase at
%Au below 25%Au and yield a discontinuous film at 15%Au, through which the underlying carbon paper substrate
is visible (Figure 3.3a-c). The size of the cracks at 35%Au are greatly diminished as a near continuous network of
ligaments is achieved (Figure 3.3g-i). However, the uneven nature of the underlying carbon paper substrate
makes this visualization difficult. To better understand the morphology changes, np-Au was deposited on Si
substrates as shown in Figure 3.5. A nanoporous morphology is no longer observed above 45%Au due to the

45
lack of continuous channels of Ag present in the base alloy (Figure 3.3j-l). At these atomic ratios, Au forms a
protective barrier that prevents the nitric acid from penetrating deeper into the alloy to remove the residual
Ag.35 The coarseness of the nanoporous ligaments and pores are modestly reduced as the %Au is increased.
Cross sectional SEM demonstrates that the nanoporous ligaments are consistently distributed throughout the
entire volume of the film after dealloying (Figure 3.6). Figure 3.2 shows SEM images of the bare carbon paper
substrate, the AgxAu1-x base alloy before the nitric acid etch, and after the etch.
25 atomic % Au

20 atomic % Au

200 nm

200 nm

10 µm

200 nm

1 µm

1 µm

30 atomic % Au

1 µm

10 µm

10 µm

Figure 3.5: SEM characterization of nanoporous gold (np-Au) electrodes with a varying gold atomic percent (%Au) on silicon.
20%Au (a-c), 25%Au (d-f), and 30%Au (g-i).

1 µm

46
Figure 3.6: cross sectional SEM of a 300 nm thick 35% gold nanoporous gold electrode on carbon paper. From this image we
can see how uneven the carbon paper substrate is and that the nanoporous gold has a homogenous pore structure through the entire
thickness.
3.3

Electrochemical characterization of nanoporous gold
CO2R performance for each electrode was evaluated in a two-compartment flow cell. A Selemion anion

exchange membrane was used to separate the Pt mesh anode from the np-Au cathode.

A 1M KHCO3

electrolyte saturated with CO2 electrolyte was independently recirculated through the anode and cathode
chambers. Despite the fact that the increased alkalinity of KOH electrolytes has been shown to improve CO2R
selectivity,28,29,36 KOH is considered to be a sacrificial medium because the hydroxide anions are converted into
bicarbonate and carbonate upon contact with any unreacted CO2 gas.37 The carbonate salts are then known to
precipitate out of solution and potentially clog the pores of the anion exchange membrane,38 which necessitates
that the KOH electrolyte be continuously replaced throughout the electrolysis.39,40 The KHCO3 electrolyte
thereby offers a sustainable alternative, albeit often at the expense of CO2R selectivity. The CO2 gas was
delivered through a serpentine channel located behind the GDE at a constant flow rate of 50 standard cubic
centimeters per minute (SCCM) unless otherwise noted. A leakless Ag/AgCl electrode was used as the reference
electrode. A gas chromatograph was used to quantify the concentration of product gasses in the effluent stream
and a potentiostat was used to control the applied electrochemical potential. All electrode potentials (E) are
reported relative to the reversible hydrogen electrode (RHE) scale (VRHE).
Figure 3.4e-h shows the FE of a pl-Au and np-Au GDEs (300 nm thick) for both CO and H2 and Figure
3e shows the corresponding partial current densities (Figure 1i-l) as a function of applied potential between the
range of -0.42 VRHE to -0.92 VRHE in a vapor CO2 fed configuration. Each measurement represents the averaged
result of three distinct electrodes. The pl-Au GDE has a maximum current density (J) of -106 mA/cm2 at -0.92
VRHE and a maximum FE for CO of 82% at -0.82 VRHE. After benchmarking the performance of the pl-Au film
as a reference, we then tested a series of np-Au electrodes to evaluate the performance over a range of varying
%Au. Both the 15%Au and 25%Au electrodes demonstrate a notable improvement on the pl-Au electrode in terms
of current density (J) and FE across the entire potential range. At 15%Au and 25%Au the performance of the
GDEs is relatively similar with the 25%Au electrode offering a slightly higher performance. The 25%Au np-Au
has a JCO that is 1.48X that of the planar Au film at -0.92 VRHE and 2.92x improvement over the pl-Au with a
peak FE for CO of 93% at -0.42 VRHE. The 35%Au electrode significantly outperforms the pl-Au across the
entire potential range with a minimum enhancement in JCO of over 2x. At -0.92 VRHE the JCO is -164 mA/cm2

47
with an FE for CO of 88%. At lower potentials the FE for CO remains above 90% with a peak of 95% at -0.42
VRHE. Additionally, the JCO of the 35%Au electrodes at -0.82 VRHE is higher at -130 mA/cm2 than the -0.92 VRHE
data points for all other systems. We hypothesize that the enhanced performance of the 35%Au electrode is due
to the near continuous ligament coverage across the electrode surface that may improve electrolyte wetting,
decrease the potential for flooding and increase the residency of CO2 in the catalyst layer by reducing the
distribution of cracks that would otherwise allow the gas to circumvent the catalyst regions.

Figure 3.7: Contact angle measurements of nanoporous gold electrodes with varying atomic gold percentages. 0 atomic percent Au
indicates that there is no catalyst layer and 100 atomic percent Au indicates that there was a solid gold film deposited. All samples
were deposited on Sigracet 38BC unless otherwise noted.
We sought to better understand the difference in performance between the np-Au catalysts of varying
%Au by understanding how the electrolyte interacts with each system. Catalyst wetting is a critical component
of device performance and we therefore carried out contact angle analysis of each of the electrodes to understand
the macroscopic wetting properties. Figure 3.7 shows the contact angle of a variety of samples including the
base carbon paper, 300 nm thick Au on Si substrate, pl-Au on carbon paper, and np-Au on carbon paper. The
carbon paper itself is hydrophobic since it contains a fluorinated microporous layer and thus exhibits a contact
angle of 159° whereas the 300 nm thick pl-Au GDE has a contact angle of 106°. The 15%Au GDE sample has
a similar hydrophobicity to the pl-Au reference with a contact angle of 108°. Interestingly, the 35%Au was found
to be the most hydrophilic with a contact angle of 59°. We attribute the higher hydrophobicity of lower %Au
samples to the large cracks in the catalyst layer that penetrate through to the hydrophobic microporous layer,
which allows the water droplet to come into contact with the fluorinated microporous layer. The 35%Au has a
more complete metal coverage, which minimizes contact with the microporous layer. Therefore, the hydrophilic

48
network of pores is able to wick water into its structure and lower the contact angle. While this provides
correlation between wetting and device performance this does not explain the improvement in CO2R
performance as catalyst hydrophobicity has been demonstrated in literature to improve CO2R this trend was not
observed in our system. To better understand our system, we must therefore consider wetting on the nanoscale
as distinct from the bulk properties highlighted by contact angle analysis.
3.4

Copper underpotential deposition characterization
Figure 2

a)
npAu-CP

Au-CP
Au-Si

b)

Figure 3.8: (a) relative surface area enhancements of a 300 nm thick Au electrode and 300 nm thick 35 %Au np-Au normalized
to the Au on Si sample as calculated by Cu UPD. (b), Surface area enhancement of 100 nm, 300 nm, and 900 nm thick 35%Au
electrodes. The dashed line shows a linear fit with an R2 = 1.00.
Cu UPD allows us to further probe wetting on the nanoscale by evaluating where electrolyte is
present. The Cu UPD process electroplates a monolayer of Cu onto any Au surface exposed to the electrolyte
solution. It is then possible to integrate the cyclic voltammogram to determine how much charge was passed,
which is proportional to the surface area.7,41,42 We are therefore able to determine the surface area of the catalyst
that is in contact with the electrolyte. We first measured the surface area of 300 nm of Au deposited onto a
silicon wafer (Au-Si) as a baseline and normalized all further Cu UPD measurements to this value (Figure

49
3.8a). A 300 nm thick film of Au deposited onto carbon paper (pl-Au) exhibited a 15x area enhancement
compared to the Au-Si. This enhancement indicated that the irregular carbon paper base substrate increases the
rugosity and subsequently the surface area of the pl-Au electrode. In comparison a 300 nm thick 35%Au np-Au
electrode on carbon paper has a 78x increase in surface area compared to the Au-Si and a 5x increase in surface
area compared to the pl-Au electrode. This significant increase in surface area demonstrated that the
nanotexturing method greatly enhanced the available surface area of a np-Au electrode relative to a pl-Au
electrode. While cross-sectional SEM demonstrated a cross-sectionally consistent np-Au morphology (Figure
S4), we sought to verify this interpretation with Cu UPD experiments. To achieve this, three distinct, 35%Au
electrodes with a catalyst layer thickness of 100 nm, 300 nm, and 900 nm were fabricated and intentionally
flooded by soaking them in an aqueous CO2 fed electrochemical cell to ensure that the entire surface area of the
catalyst layer was electrochemically accessible (Figure 3.8b). We found that the surface area as measured by Cu
UPD increased linearly with the thickness of the catalyst. The linear increase in surface area with film thickness
confirmed that our electrodes are indeed cross sectionally consistent and that the Cu2+ ions in solution penetrate
into the entire depth of the catalyst layer in the aqueous CO2 fed configuration. If this were not the case, we
would have observed that the surface area enhancement measured by Cu UPD would taper off with increasing
np-Au film thickness. This effect would result from the Cu2+ ions inability to diffuse throughout the extend of
the np-Au network.

50

Figure 3

Surface Area Relative to
Aqueous CO2 Fed Electrode

a)

b)

0.8
0.6
0.4
0.2

Faradaic
Efficiency (%)

100
80
60
40

CO
H2

20

c)

J (mA/cm2)

-40
-80
-120

CO
H2

-160

20

40

Flow Rate of CO2 (SCCM)

Figure 3.9: (a) relative surface area of three identical 300 nm thick, 35%Au np-Au electrodes under different CO2 flow rates
normalized to the surface area of the same electrode in an aqueous CO2 fed system. (b) Faradaic efficiencies of the electrodes at different
CO2 flow rates. (c) Partial current densities of the electrodes at different CO2 flow rates. All experiments for (b) and (c) were carried
out at -0.92 VRHE.
Interestingly, we observed a 5x increase in surface area for the 35%Au sample compared to the pl-Au,
with only a 2–3.5x increase in J. To understand the discrepancy between surface area and J we carried out a set
of Cu UPD experiments under vapor CO2 fed conditions (Figure 3.9a). We hypothesized that the surface area
calculated by Cu UPD from an electrode in an aqueous CO2 fed system represents the true surface area of the
of the np-Au catalyst layer as the entirety of the Au is in contact with the electrolyte, i.e. flooded. The surface
area of a fully flooded np-Au electrode is normalized to one and any deviation from this value by a vapor CO2
fed electrode can be attributed to a percentage of the catalyst layer being inaccessible to the electrolyte. At a
flow rate of 50 SCCM the measured surface area of a 35%Au electrode decreased to 57% of its maximum value,
indicating that only ~½ the electrode is in contact with the electrolyte. As the flow rate was reduced to 10
SCCM and 2 SCCM the accessible surface area increased to 66% and 87% of the total available surface area

51
respectively. We attribute this effect to an increase of flooded pores in the np-Au catalyst layer due to the
pressure drop as a result of the reduced flow rate. To evaluate this, we performed a series of CO2R experiments
on 300 nm thick, 35%Au np-Au electrodes at 50 SCCM, 10 SCCM and 2 SCCM at -0.92 VRHE. We found that
the current density decreased from -190 mA/cm2 for the 50 SCCM sample to -107 mA/cm2 for the 10 SCCM
sample and further to -91 mA/cm2 for the 2 SCCM sample (Figure 3c). This drop in current density indicates
that the catalyst layer is flooding as the flow rate of CO2 is reduced and that the flooded sections of the electrode
have regions more similar to the mass transport limited, aqueous CO2 fed system rather than the vapor CO2 fed
system. To further verify this, we tracked the FE for H2 of each electrode and found it to increase from 12%
for the 50 SCCM sample to 34% for the 10 SCCM sample and to 48% for the 2 SCCM sample (Figure
3.9b). This increase in HER indicates that the catalyst cannot carry out CO2R and instead, increasingly engages
in parasitic HER in the flooded pores. This result indicates that the catalyst layer exists in as a combination of
three distinct states (Figure 3.1). State one is the fully flooded state that aqueous CO2 fed systems exists in. This
state of operation heavily suppresses current density due the mass transport limitations caused by the longer
diffusion lengths that CO2 must travel through the electrolyte to reach the catalyst surface. State two is the vapor
CO2 fed, GDE configuration in which thin layers of electrolyte wet the catalyst surface and gaseous pores
effectively deliver CO2 throughout the catalyst layer. In this state, the much smaller diffusion length allows high
current densities to be achieved. It should be noted that these two states are indistinguishable by Cu UPD but
that an increase in flooding is apparent as current density is reduced and HER is increased. In state three, the
Figure 4

catalyst is completely dry and unavailable for catalysis.
3.5

Secondary ion mass spectroscopy characterization

b)

Aqueous CO2 fed electrode

0.8

Vapor CO2 fed electrode

0.6
0.4
0.2

0.2

0.4

0.6

0.8

Normalized Depth into np-Au

Fraction of Catalyst in
Contact with Electrolyte

a)

Normalized Cu Counts

Mostly flooded

Transition region

2/3 dry and 1/3 in
contact with electrolyte

0.8
0.6
0.4
0.2

0.2

0.4

0.6

0.8

Normalized Depth into np-Au

Figure 3.10: Secondary ion mass spectroscopy data summary. (a) shows normalized Cu counts relative to normalized depth into
the np-Au electrode with 0 as the surface of the np-Au electrode in contact with the electrolyte and 1 as the np-Au electrode in contact

52
with the microporous layer. (b) is the ratio of the of the vapor CO2 fed counts to the aqueous CO2 fed counts in (a) and represent
the portion of the catalyst that is in contact with the electrolyte as a function of depth.
a)

b)

c)

Figure 3.11: Secondary ion mass spectroscopy raw data for aqueous CO2 fed system and vapor CO2 fed system. (a) shows the gold
counts, (b) the copper counts, and (c) the carbon counts.
In order to verify that these Cu UPD results indicate that a certain fraction of the catalyst is unwetted,
we electroplated Cu onto a flooded and operando electrode. This should yield an electrode with copper metal
plated onto the regions that were in contact with electrolyte and an absence of copper where the electrode is
dry. Therefore, the presence of Cu is a proxy for wetting and will be referred to as such from now on. Secondary
Ion Mass Spectroscopy (SIMS) was carried out on an aqueous CO2 fed electrode and a vapor CO2 fed electrode
to elucidate where the Cu was spatially deposited (Figure 3.10a and Figure 3.11). We make the assumption that
the np-Au electrodes of each sample are near identical and then compare the Cu/Au ratio of each distinct
electrode. We find that the total integrated Cu/Au count of the vapor CO2 fed electrode is 57% that of the
aqueous CO2 fed electrode (Figure 3.10a). This result is in close agreement with our Cu UPD finding that the
vapor CO2 fed electrode has 56% the surface area of the aqueous CO2 fed one. Next, we took the ratio of these
results to determine the percentage of the vapor CO2 fed electrode that is in contact with the electrolyte as a
function of depth (Figure 3.10b). We define 0 as the surface of the np-Au electrode in contact with the
electrolyte and 1 as the np-Au electrode in contact with the microporous layer. If the vapor fed/aqueous fed
ratio is equal to 1, the wetting of the electrodes are equivalent whereas fractional deviations indicate that the
aqueous CO2 fed electrode has more gold in contact with the electrolyte. We find that the vapor fed/aqueous
fed depth profile can be separated into three distinct regions. In region 1 (the top 10% of the catalyst) both
electrodes have a comparable wetting which indicates that the vapor CO2 fed electrode is likely highly flooded
this region. Region 2 is a transition zone where the wetting in the catalyst layer of the vapor CO2 fed is

53
increasingly reduced compared to the aqueous CO2 fed electrode. Region 3 consists of the bottom 75% of the
catalyst layer and here the vapor CO2 fed electrode has 2/3 less wetting than the aqueous CO2 fed electrode.
This finding shows that a significant fraction of the catalyst layer of the vapor CO2 fed electrode is not in contact
with the electrolyte and therefore does not contribute to CO2R. These previously unknown results allow us to
determine the percentage of the catalyst that contributes towards CO2R. We anticipate that these results will
allow for improved GDE designs that are capable of harnessing all of the available electrode surface area to
drive electrocatalytic CO2R.
3.6

Conclusion
In this work we provide a new method for how to evaluate the active regions of GDEs for CO2R.

Through a combination of Cu UPD and SIMS we were able to show that under operating conditions for a vapor
CO2 fed device that the top 10% of the nanoporous gold layer is fully flooded, while the rest of the catalyst layer
exists as combination of the flooded/wetted and dry. Strikingly, 43% of the catalyst is not in contact with the
electrolyte at all. In addition to providing this new insight into the active regions of the catalyst we also
demonstrate a highly active CO2R device. The np-Au GDEs are capable of attaining an FE for CO of up to
95% and a JCO of -168 mA/cm2. We hope that our analysis will enable our colleagues in the field to better
understand and maximize the amount of catalyst that contributes towards CO2R.

BIBLIOGRAPHY CHAPTER 3
1. Garg, S.; Li, M.; Weber, A. Z.; Ge, L.; Li, L.; Rudolph, V.; Wang, G.; Rufford, T. E. Advances and
Challenges in Electrochemical CO2 Reduction Processes: An Engineering and Design Perspective
Looking beyond New Catalyst Materials. J. Mater. Chem. A 2020, 8 (4), 1511–1544.
2. Schreier, M.; Héroguel, F.; Steier, L.; Ahmad, S.; Luterbacher, J. S.; Mayer, M. T.; Luo, J.; Grätzel, M.
Solar Conversion of CO2 to CO Using Earth-Abundant Electrocatalysts Prepared by Atomic Layer
Modification of CuO. Nat. Energy 2017, 2 (7).
3. Lin, S.; Diercks, C. S.; Zhang, Y. B.; Kornienko, N.; Nichols, E. M.; Zhao, Y.; Paris, A. R.; Kim, D.;
Yang, P.; Yaghi, O. M.; Chang, C. J. Covalent Organic Frameworks Comprising Cobalt Porphyrins for
Catalytic CO2 Reduction in Water. Science (80-. ). 2015, 349 (6253), 1208–1213.
4. Kuhl, K. P.; Cave, E. R.; Abram, D. N.; Jaramillo, T. F. New Insights into the Electrochemical
Reduction of Carbon Dioxide on Metallic Copper Surfaces. Energy Environ. Sci. 2012, 5 (5), 7050–7059.
5. Kuhl, K. P.; Hatsukade, T.; Cave, E. R.; Abram, D. N.; Kibsgaard, J.; Jaramillo, T. F. Electrocatalytic
Conversion of Carbon Dioxide to Methane and Methanol on Transition Metal Surfaces. J. Am. Chem.
Soc. 2014, 136 (40), 14107–14113.
6. Hatsukade, T.; Kuhl, K. P.; Cave, E. R.; Abram, D. N.; Jaramillo, T. F. Insights into the Electrocatalytic
Reduction of CO2 on Metallic Silver Surfaces. Phys. Chem. Chem. Phys. 2014, 16 (27), 13814–13819.

54
7. Welch, A. J.; Duchene, J. S.; Tagliabue, G.; Davoyan, A.; Cheng, W. H.; Atwater, H. A. Nanoporous
Gold as a Highly Selective and Active Carbon Dioxide Reduction Catalyst. ACS Appl. Energy Mater.
2019, 2 (1), 164–170.
8. Rosen, B. A.; Salehi-khojin, A.; Thorson, M. R.; Zhu, W.; Whipple, D. T.; Kenis, P. J. A.; Masel, R. I.
Ionic Liquid – Mediated Selective. Science (80-. ). 2011, 334 (November), 643–645.
9. Yang, J.; Ma, W.; Chen, D.; Holmen, A.; Davis, B. H. Fischer-Tropsch Synthesis: A Review of the
Effect of CO Conversion on Methane Selectivity. Appl. Catal. A Gen. 2014, 470, 250–260.
10. Verma, S.; Kim, B.; Jhong, H. R. M.; Ma, S.; Kenis, P. J. A. A Gross-Margin Model for Defining
Technoeconomic Benchmarks in the Electroreduction of CO2. ChemSusChem 2016, 9 (15), 1972–1979.
11. Hernández, S.; Farkhondehfal, M. A.; Sastre, F.; Makkee, M.; Saracco, G.; Russo, N. Syngas Production
from Electrochemical Reduction of CO2: Current Status and Prospective Implementation. Green Chem.
2017, 19 (10), 2326–2346.
12. Pruett, R. L.; Smith, J. A. A Low-Pressure System for Producing Normal Aldehydes by
Hydroformylation of α Olefins. J. Org. Chem. 1969, 34 (2), 327–330.
13. Jhong, H. R. M.; Ma, S.; Kenis, P. J. Electrochemical Conversion of CO2 to Useful Chemicals: Current
Status, Remaining Challenges, and Future Opportunities. Curr. Opin. Chem. Eng. 2013, 2 (2), 191–199.
14. Bloemacher, D. Carbonyl Iron Powders: Its Production and New Developments. Met. Powder Rep. 1990,
45 (2), 117–119.
15. Bushuyev, O. S.; De Luna, P.; Dinh, C. T.; Tao, L.; Saur, G.; van de Lagemaat, J.; Kelley, S. O.; Sargent,
E. H. What Should We Make with CO2 and How Can We Make It? Joule 2018, 2 (5), 825–832.
16. Nitopi, S.; Bertheussen, E.; Scott, S. B.; Liu, X.; Engstfeld, A. K.; Horch, S.; Seger, B.; Stephens, I. E. L.;
Chan, K.; Hahn, C.; Nørskov, J. K.; Jaramillo, T. F.; Chorkendorff, I. Progress and Perspectives of
Electrochemical CO2 Reduction on Copper in Aqueous Electrolyte. Chem. Rev. 2019, 119 (12), 7610–
7672.
17. Lu, Q.; Rosen, J.; Zhou, Y.; Hutchings, G. S.; Kimmel, Y. C.; Chen, J. G.; Jiao, F. A Selective and
Efficient Electrocatalyst for Carbon Dioxide Reduction. Nat. Commun. 2014, 5, 1–6.
18. Asadi, M.; Kim, K.; Liu, C.; Addepalli, A. V.; Abbasi, P.; Yasaei, P.; Phillips, P.; Behranginia, A.; Cerrato,
J. M.; Haasch, R.; Zapol, P.; Kumar, B.; Klie, R. F.; Abiade, J.; Curtiss, L. A.; Salehi-Khojin, A.
Nanostructured Transition Metal Dichalcogenide Electrocatalysts for CO2 Reduction in Ionic Liquid.
Science (80-. ). 2016, 353 (6298), 467–470.
19. Gao, D.; Zhou, H.; Wang, J.; Miao, S.; Yang, F.; Wang, G.; Wang, J.; Bao, X. Size-Dependent
Electrocatalytic Reduction of CO2 over Pd Nanoparticles. J. Am. Chem. Soc. 2015, 137 (13), 4288–4291.
20. Chen, Y.; Li, C. W.; Kanan, M. W. Aqueous CO2 Reduction at Very Low Overpotential on OxideDerived Au Nanoparticles. J. Am. Chem. Soc. 2012, 134 (49), 19969–19972.
21. Weng, L. C.; Bell, A. T.; Weber, A. Z. Modeling Gas-Diffusion Electrodes for CO2 Reduction. Phys.
Chem. Chem. Phys. 2018, 20 (25), 16973–16984.
22. Ma, M.; Clark, E. L.; Therkildsen, K. T.; Dalsgaard, S.; Chorkendorff, I.; Seger, B. Insights into the
Carbon Balance for CO2 Electroreduction on Cu Using Gas Diffusion Electrode Reactor Designs.
Energy Environ. Sci. 2020, 13 (3), 977–985.
23. Singh, M. R.; Goodpaster, J. D.; Weber, A. Z.; Head-Gordon, M.; Bell, A. T. Mechanistic Insights into
Electrochemical Reduction of CO2 over Ag Using Density Functional Theory and Transport Models.
Proc. Natl. Acad. Sci. U. S. A. 2017, 114 (42), E8812–E8821.
24. Weekes, D. M.; Salvatore, D. A.; Reyes, A.; Huang, A.; Berlinguette, C. P. Electrolytic CO2 Reduction in
a Flow Cell. Acc. Chem. Res. 2018, 51 (4), 910–918.
25. Higgins, D.; Hahn, C.; Xiang, C.; Jaramillo, T. F.; Weber, A. Z. Gas-Diffusion Electrodes for Carbon
Dioxide Reduction: A New Paradigm. ACS Energy Lett. 2019, 4 (1), 317–324.
26. García de Arquer, F. P.; Dinh, C. T.; Ozden, A.; Wicks, J.; McCallum, C.; Kirmani, A. R.; Nam, D. H.;
Gabardo, C.; Seifitokaldani, A.; Wang, X.; Li, Y. C.; Li, F.; Edwards, J.; Richter, L. J.; Thorpe, S. J.;

55
Sinton, D.; Sargent, E. H. CO2 Electrolysis to Multicarbon Products at Activities Greater than 1 A
Cm−2. Science (80-. ). 2020, 367 (6478), 661–666.
27. Grigioni, I.; Sagar, L. K.; Li, Y. C.; Lee, G.; Yan, Y.; Bertens, K.; Miao, R. K.; Wang, X.; Abed, J.; Won,
D. H.; Garciá De Arquer, F. P.; Ip, A. H.; Sinton, D.; Sargent, E. H. CO2 Electroreduction to Formate
at a Partial Current Density of 930 MA Cm-2 with InP Colloidal Quantum Dot Derived Catalysts. ACS
Energy Lett. 2021, 6 (1), 79–84.
28. Dinh, C. T.; García De Arquer, F. P.; Sinton, D.; Sargent, E. H. High Rate, Selective, and Stable
Electroreduction of CO2 to CO in Basic and Neutral Media. ACS Energy Lett. 2018, 3 (11), 2835–2840.
29. Verma, S.; Hamasaki, Y.; Kim, C.; Huang, W.; Lu, S.; Jhong, H. R. M.; Gewirth, A. A.; Fujigaya, T.;
Nakashima, N.; Kenis, P. J. A. Insights into the Low Overpotential Electroreduction of CO2 to CO on a
Supported Gold Catalyst in an Alkaline Flow Electrolyzer. ACS Energy Lett. 2018, 3 (1), 193–198.
30. Zhang, W.; He, J.; Liu, S.; Niu, W.; Liu, P.; Zhao, Y.; Pang, F.; Xi, W.; Chen, M.; Pang, S. S.; Ding, Y.
Atomic Origins of High Electrochemical CO2 Reduction Efficiency on Nanoporous Gold. Nanoscale
2018, 10 (18), 8372–8376.
31. Chen, C.; Zhang, B.; Zhong, J.; Cheng, Z. Selective Electrochemical CO2 Reduction over Highly Porous
Gold Films. J. Mater. Chem. A 2017, 5 (41), 21955–21964.
32. Wang, R.; Wang, C.; Cai, W. Bin; Ding, Y. Ultralow-Platinum-Loading High-Performance Nanoporous
Electrocatalysts with Nanoengineered Surface Structures. Adv. Mater. 2010, 22 (16), 1845–1848.
33. Sukeri, A.; Saravia, L. P. H.; Bertotti, M. A Facile Electrochemical Approach to Fabricate a Nanoporous
Gold Film Electrode and Its Electrocatalytic Activity towards Dissolved Oxygen Reduction. Phys. Chem.
Chem. Phys. 2015, 17 (43), 28510–28514.
34. Hall, A. S.; Yoon, Y.; Wuttig, A.; Surendranath, Y. Mesostructure-Induced Selectivity in CO2 Reduction
Catalysis. J. Am. Chem. Soc. 2015, 137 (47), 14834–14837.
35. Biener, J.; Biener, M. M.; Madix, R. J.; Friend, C. M. Nanoporous Gold: Understanding the Origin of the
Reactivity of a 21st Century Catalyst Made by Pre-Columbian Technology. ACS Catal. 2015, 5 (11),
6263–6270.
36. Verma, S.; Lu, X.; Ma, S.; Masel, R. I.; Kenis, P. J. A. The Effect of Electrolyte Composition on the
Electroreduction of CO2 to CO on Ag Based Gas Diffusion Electrodes. Phys. Chem. Chem. Phys. 2016, 18
(10), 7075–7084.
37. Winter, M.; Brodd, R. J. What Are Batteries, Fuel Cells, and Supercapacitors? Chem. Rev. 2004, 104 (10),
4245–4269.
38. Gülzow, E.; Schulze, M. Long-Term Operation of AFC Electrodes with CO2 Containing Gases. J. Power
Sources 2004, 127 (1–2), 243–251. https://doi.org/10.1016/j.jpowsour.2003.09.020.
39. Gülzow, E. Alkaline Fuel Cells: A Critical View. J. Power Sources 1996, 61 (1–2), 99–104.
40. Gouérec, P.; Poletto, L.; Denizot, J.; Sanchez-Cortezon, E.; Miners, J. H. The Evolution of the
Performance of Alkaline Fuel Cells with Circulating Electrolyte. J. Power Sources 2004, 129 (2), 193–204.
41. Zhang, B. A.; Ozel, T.; Elias, J. S.; Costentin, C.; Nocera, D. G. Interplay of Homogeneous Reactions,
Mass Transport, and Kinetics in Determining Selectivity of the Reduction of CO2 on Gold Electrodes.
ACS Cent. Sci. 2019, 5 (6), 1097–1105.
42. Ross, M. B.; Dinh, C. T.; Li, Y.; Kim, D.; De Luna, P.; Sargent, E. H.; Yang, P. Tunable Cu Enrichment
Enables Designer Syngas Electrosynthesis from CO2. J. Am. Chem. Soc. 2017, 139 (27), 9359–9363.

56
Chapter 4

OPERANDO LOCAL pH MEASUREMENT WITHIN GAS DIFFUSION
ELECTRODES PERFORMING ELECTROCHEMICAL CARBON DIOXIDE
REDUCTION
4.1

Introduction
While the cost of renewable electricity has declined markedly, selective, energy-efficient synthesis of

storable chemical fuels is necessary to enable widespread adoption of sustainable energy. One approach is to
transform solar energy into chemical fuels and fuel precursors via artificial photosynthesis. Recently, significant
advances have been made in the design of gas diffusion electrodes (GDEs) for electrochemical carbon dioxide
(CO2) reduction at high current densities. While promising, GDEs have not yet achieved their full potential for
product selectivity and energy efficiency due to the complexity of the electrocatalytic reactions involved in
making fuels from CO2 reduction.
Many parameters may influence the selectivity and activity of the CO2 reduction reaction, the most
obvious of which are the catalyst1–3 and applied potential.4,5 Aside from these two critical parameters, GDE
system configurations (flow through vs. flow by),6 local electrolyte viscosity,7 concentration and identity of
cations in the electrolyte,8 salt deposition on the GDE,9 membrane structure and composition,10 bicarbonate and
carbonate formation in the electrolyte,11 hydrophobicity of thee GDE,12 and other factors can have a significant
influence on device performance.
Of particular interest in this work is the local pH established at the electrode surface during fuel synthesis.
Electrochemical solar fuel-forming reactions create hydroxide ions (OH–) at the catalyst surface during the
reaction when in alkaline electrolyte which alters the local pH near the cathode,13,14 thus strongly impacting both
product selectivity and activity.15–18 Therefore, it is important to distinguish the local pH near the electrode
surface from the pH in the bulk electrolyte. Although challenging to determine experimentally, the local pH
near the electrode surface should be measured under operating conditions to provide the necessary insight
required to further improve the activity, selectivity, and stability of these fuel-forming devices.
In this study, we focus on understanding the pH in GDEs because this device architecture has increased
the performance of CO2 reduction electrodes by an order of magnitude due to their ability to deliver CO2 in the
gas phase, thereby overcoming the mass transport limitations encountered in more traditional electrocatalytic
devices.19 Not only does this architecture allow for higher current densities, but also improved product selectivity

57
in CO2 reduction.19–21 In CO2 reduction devices where the CO2 is dissolved directly into the electrolyte, the
maximum current density is less than –30 mA/cm2 due to low CO2 solubility in aqueous electrolytes (around
[34 mM] at maximum).22 The main challenge of the GDE architecture is in developing the appropriate device
structure to maintain what is referred to as a “wetted” condition. Here, a thin layer of electrolyte coats the catalyst
to provide water molecules while simultaneously allowing for rapid dissolution of CO2 through the electrolyte
to avoid mass-transport limitations. If the water layer coating the catalyst is too thick, the catalyst becomes
flooded and its operation is more similar to a bulk aqueous electrolyte CO2 reduction device.23 Alternatively, if
there is no electrolyte, the catalyst has no access to water molecules and no reaction can occur. To achieve this
wetted condition the GDE is composed of a gas diffusion layer, microporous layer, and catalyst layer (Figure
4.1). The microporous layer is perhaps the most critical because the concentration of polytetrafluoroethylene
(PTFE) allows the wetting to be tuned.12,23 Tailored GDE architectures have demonstrated current densities
greater than 1 A/cm2 for multi-carbon products.24
It is vital to understand the local pH within GDEs due to the high CO2 flow rates and high current
densities at which these devices operate. Interestingly, these two characteristics have opposing effects on the
local pH near the electrode surface. High current densities result in the creation of multiple hydroxide ions per
unit time, thus rapidly increasing the pH, while any unreacted CO2 will acidify the electrolyte via reaction with
OH– to form HCO3– and H+ ions. If the pH is increased locally, the activity of the hydrogen evolution reaction
(HER) decreases substantially, while the CO2 reduction reaction becomes relatively more favorable.17,18 While
both reactions consume water molecules, the rate of H2 evolution has been shown to be far more dependent on
the local pH than the rate of CO2 reduction.17,25 In addition to suppressing HER, the local pH also influences
which CO2 reduction pathways are most energetically favorable.16,18,26 Alkaline conditions in particular promote
the formation of multi-carbon products (C2+) such as ethanol, propanol, acetate, etc.16,25,26 Theoretical models
have been developed to simulate the local pH near operating CO2 reduction GDEs,23,27,28 however it is difficult
to accurately represent the complex electrochemical environment created by the triple phase boundary at the
catalyst surface. We therefore seek to directly measure the local pH near an operating GDE and experimentally
validate the results of these prior simulations.
There are various techniques that can be used to probe the local pH, such as fluorescence
microscopy,15,29–34 scanning electrochemical cell microscopy (SECM),35–37 surface enhanced Raman spectroscopy
(SERS),38 and surface-enhanced infrared absorption spectroscopy (SEIRAS).39 Previously, SERS has been used
to measure the local pH in a CO2 reduction GDE with a 1 M KOH electrolyte (pH 14). It was found that the
local pH was near 7 in the absence of any current flow, and as the current density increased to 100 mA/cm2, the
local pH increased to 10. It is interesting to note that even with an electrolyte with a bulk pH of 14, the electrode

58
surface remained in the range pH range of 7-10 for a wide range of current densities. However, this
measurement did not provide any resolution along the plane of the electrode surface. SECM studies confirm
that the local pH increases during device operation. 35–37 While SECM allows for better spatial resolution than
confocal fluorescent microscopy, it is unable to probe the pH within the trenches of the GDE, Figure 4.1e, f.
Our study builds on previous work by mapping the pH both on the surface and within the heterogeneous
reaction environments encountered in GDEs. This experimental approach therefore allows us to correlate the
width of trenches in GDEs to the local pH.
Here, we use confocal microscopy and a pH-sensitive two-color fluorescent dye to probe the operando
local pH of a copper (Cu) GDE under CO2 reduction conditions with micron-scale resolution in all three spatial
dimensions [within the plane of the electrode (x, y) and perpendicular to its surface (z)]. This approach offers
new insight into how CO2 reduction affects the local electrolyte pH near the Cu catalyst. Interestingly, our study
indicates that at low overpotentials, the pH varies widely across the electrode surface. Specifically, in narrow
trenches throughout the electrode, the pH is significantly elevated compared to the surroundings. Our findings
highlight the electrocatalytic heterogeneity in GDEs and strongly suggest that these regions of locally-high pH
are the most active parts of the electrode for CO2 reduction.

Figure 4.1: Overview of a Cu gas diffusion electrode (GDE) for CO2 reduction studies. (a) Cross-sectional diagram of the custom
electrochemical cell designed to enable in situ confocal fluorescent microscopy experiments. (b) Schematic of a typical Cu GDE, not to

59
scale. (c), (d) Scanning electron microscope (SEM) images of a Cu GDE. (e), (f) SEM images of an uncoated microporous layer.
(g), (h) SEM images of the gas diffusion layer.
A Cu-based GDE was investigated owing to the unique ability of Cu to produce C2+ products (e.g.
ethanol).3,40 Figure 4.1 shows a schematic of the GDE and experimental setup used, as well as scanning electron
microscopy (SEM), and optical bright-field microscopy images of different layers of the device. The GDE used
here and in many other devices19,25,41,42 is composed of three layers: (1) a gas diffusion layer, (2) a microporous
layer, and (3) a catalyst layer (Figure 4.1b). The CO2 first diffuses through the gas diffusion layer, composed of
carbon fibers (Figure 4.1 g, h), and then through the microporous layer, which is made of carbon black coated
in hydrophobic PTFE to regulate local electrolyte availability (Figure 4.1e, f). After diffusion through the
microporous layer, the CO2 comes into contact with the electrolyte at the catalyst layer (Figure 4.1c, d) where it
can then undergo reduction to yield a variety of chemical products.
Some CO2 reacts at the catalyst surface into products such as CO, HCOOH, or CH4. The remaining
unreacted portion of the CO2 then passes into the electrolyte and increases its acidity.43 While several reports
quantified the one pass CO2 utilization efficiency,44 the vast majority of the CO2 reduction experiments did not
seek to optimize the utilization of CO2. The competition between these two processes – CO2 acidification and
hydroxide ion generation – can be investigated via measurement of the local pH at the catalyst-electrolyte
interface.
4.2

Experimental set up and characterization
We used confocal fluorescent microscopy to measure the local pH due to its high spatial resolution

relative to the dimensions of interest in the system. Figure 4.1a shows a schematic of the experimental set up
and Figure 4.2 shows a more detailed schematic, as well as photos of the cell. The technical resolution of this
system is 280 nm in x-y and 560 nm in z. However due to noise from the electrolyte pump and diffusion, the
resolution is on the order of a micron under our conditions. A laser is used to excite a ratiometric two-color pH
sensitive fluorescent dye, 6,8-dihydroxy-1,3-pyrenedisulfonic acid (DHPDS) in the electrolyte.45 This approach
ensures that the pH is independent of the concentration of DHPDS in the solution.15 Figure 4.3 shows the
absorbance vs. wavelength of DHPDS in different standard solutions of known pH. At the most acidic pH of
4.6, the peak absorbance is centered at ~400 nm, at pH 8.5 the peak absorbance is at ~455 nm, and at pH 11.7
the peak absorbance shifts toward ~480 nm. We focused our studies on near-neutral to basic pH conditions
based on the results of previous work,22,25 which reported that these conditions are most favorable for CO2

60
reduction. We therefore sequentially excite the dye line by line with a 458 nm (lex1) and 488 nm (lex2) laser in
order to achieve the resolution over the widest range of relevant pH values.
DHPDS was calibrated by collecting the fluorescence from two different wavelength excitations (458
nm and 488 nm) for a range of pH solutions. The solutions were made by mixing KOH, bicarbonate, water
and standard buffer solutions. The pH of these solutions was measured by a Denver Instruments Ultra Basic
pH meter and results were confirmed with color changing pH strips. The pH meter was calibrated with buffer
solutions at pH 4, pH 7, and pH 10 before every set of measurements. The solution was then placed under the
microscope and the water immersion objective was immersed in it. The objective was then focused on a point
in the middle of the solution. Three images were then taken here with sequential line by line excitation of 458
nm and 488 nm. The fluorescence intensity signal was gathered from 515 – 700 nm. The ratio of the
fluorescence from these two excitations was then plotted vs. pH to generate Figure 4.3b. The pH was measured
again after the image was taken to make sure the pH reading is accurate. After acquiring all of the data, we fit
the pH data to the function, y = –a / (1 + exp(–b * (x – c))) + d. We found that coefficients to be a = –33.72,
b = 1.413, c = 8.083, and d = 5.571 for a 95% confidence bounds. We therefore have an error of 0.3 pH units.

Figure 4.2: (a) shows a schematic of the electrochemical cell used for imaging the pH via confocal fluorescent microscopy. The bottom
plate is the gas chamber and the top plate holds the electrolyte. This setup has no membrane and the electrolyte is constantly being

61
flowed across the active catalyst layer. (b) shows a top-down photo of the electrochemical cell without the microscope objective. (c) shows
a photo of the entire experimental setup with the objective in the cell, the electrolyte bath, and the pump to circulate the electrolyte
through the electrochemical cell.

Figure 4.3: Characterization of the pH-sensitive DHPDS fluorescent dye. (a) shows the absorbance of DHPDS for different
pH solutions. vertical black lines denote the two different excitation wavelengths (lex1 = 458 nm) and (lex2 = 458 nm) used for
the study. (b) shows the ratio of fluorescence emission from a 458 nm and 488 nm excitation wavelength as a function of solution
pH. After acquiring all of the data, we fit the pH data to the function, y = –a/(1+exp(–b*(x – c)))+d. We found the
coefficients to be a = –33.72, b = 1.413, c = 8.083, and d = 5.571 for 95% confidence bounds. We therefore have an error of
0.3 pH units. (c) shows the current (J) vs applied electrode potential (E) for a CO2 reduction electrode with (dashed line) and
without (solid line) DHPDS dye in the electrolyte.

Figure 4.4: Characterization of how the pH-sensitive DHPDS fluorescent dye affects the activity and selectivity of the copper
GDE. (a) shows the current density (J) vs time before the dye is added (0 to 34 minutes) and after the dye is added to the
electrolyte at 35 minutes. The partial current density for HER increases but the CO2 reduction partial current densities remain

62
stable. (b) shows the Faradaic efficiency for gas products vs time. The dye is added to the electrolyte at 35 minutes. The Faradaic
efficiency for HER increases but the CO2 reduction Faradaic efficiency remains stable. (c) shows the Faradaic efficiency for liquid
products before and after the dye was added. The Faradaic efficiency for the CO2 reduction reactions remain similar before and
after the addition of the dye, albeit with slight increase in ethanol (orange) and decrease in formic acid (blue).
A CO2-saturated solution of 100 mM KHCO3 with a bulk pH of 6.8 was used as the electrolyte in our
experiments, ensuring that the bulk pH will be at the lower sensitivity limit of the DHPDS. The DHPDS dye
is electrochemically stable under CO2 reduction reaction conditions (Figure 4.3c and Figure 4.4). In Figure 4.3c,
the current-voltage characteristics of the electrode are nearly identical with or without the ratiometric dye. Upon
addition of the dye, HER activity slightly increased but left the CO2 reduction reaction activity remains largely
unchanged (Figure 4.4). Based on these control experiments, the DHPDS dye is relatively inert with regards to
GDE operating conditions.

Figure 4.5: Electrocatalytic characterization of a GDE composed of carbon paper coated with 300 nm of Cu on top of the
microporous layer. (a) Faradaic Efficiency and (b) partial current density, J, for each product as a function of electrode potential, E.
The figure legend applies to both panels (a) and (b).
We first characterized the electrochemical performance of our Cu-based GDE prior to imaging the local
solution pH. An AMOD dual electron beam deposition system (System 02520, Angstrom Engineering) was
used to fabricate all samples. 300 nm of Cu was deposited onto the microporous layer of the gas diffusion
electrode (GDE) at a rate of 2 Å/s. The substrate holder was rotating for all depositions. Over the course of
the deposition the partial pressure of the chamber would rise from ~10-7 torr to ~10-6 torr and the temperature
would rise from 20 °C to 30 °C for a 300 nm thick sample. For samples that were used for confocal fluorescent
microscopy, an aluminum foil shadow mask with 3 mm diameter holes was used. For electrodes used for

63
product detection, no shadow mask was used. After deposition, the samples were first spray coated with a
solution of carbon black and Nafion. The solution is one-part DI water and one-part isopropyl alcohol with 2.5
mg of carbon black per ml of solution and 50 ml of 5 weight % Nafion per ml of solution. For both coatings
the solution was sprayed from a distance of 3 inches for 1 second per square inch of electrode. After the samples
were coated, they were dried under vacuum overnight.
Chronoamperometry experiments were performed across a range of applied potentials and the products
were measured via gas chromatography and high-performance liquid chromatography (Figure 4.5). All applied
potentials (E) are reported vs. the Reversible Hydrogen Electrode (E vs. RHE). As shown in Figure 4.5a, at
potentials more positive than –1.0 V vs. RHE, the Cu GDE produced primarily H2 with some C1 and C2
products. Consistent with prior observations from Cu GDEs, H2 is the dominant product at low overpotentials,
while higher overpotentials favor CO2 reduction.4,46 Cu requires higher overpotentials to perform the C-C
coupling reactions necessary to synthesize C2 products.21 At –1.0 V vs. RHE, we begin to observe many CO2
reduction products, with the largest fraction consisting of C2 products, ethylene and ethanol. Higher
overpotentials were not evaluated because the limited pH sensitivity range of the DHPDS dye is not suited for
higher current densities. We therefore restricted our electrocatalytic characterization to those conditions that
could be directly examined with confocal microscopy.

Figure 4.6: Stability of the Cu GDE working electrode potential (Ewe) over time. (a) shows four different electrochemical tests where
the current density is set to -3.4 mA/cm2. From this we can see that there are only very small changes in potential of the working
electrode between tests. (b) shows electrochemical tests with varying current density. We note that the potential of the working electrode
is very stable after the first 5 minutes indicating that electrode is stable throughout the run.
For the confocal microscopy experiments, an electrochemical compression cell similar to the one used
for the electrocatalytic characterization (Figure 4.1a and Figure 4.2) was employed. We note that the cell

64
membrane was removed to accommodate the short focal length (1.7 mm) of the objective so that it could be
positioned close to the cathode. Additionally, the cell was rotated 90° to accommodate the geometry of the
confocal microscope.

The choice and design of the electrochemical cell is further discussed in the

Supplementary Information. The DHPDS dye (100 µM) was dissolved in the electrolyte and the electrolyte was
replaced between every experiment to ensure that the initial conditions were standardized to keep the flux of
hydroxide ions constant between experiments. The electrode was stable between experiments with minimal
changes in the potential of the working electrode after 5 minutes (Figure 4.6). For each current density that was
tested, a series of images were taken 3 µm apart in the z-direction (perpendicular to the electrode surface). Figure
4.7 shows two-dimensional (x,y) maps of solution pH obtained from a single location on the cathode surface at
varying distances from the electrode surface (within a column) and at different current densities (along a row).
The color scale in each map from blue to yellow denotes the local solution pH from pH = 7 to 10.

Figure 4.7: Operando mapping of solution pH in three dimensions over a Cu GDE. Maps are obtained at the same location on
the electrode at different heights above the electrode surface and at different current densities. From top to bottom, each row of maps
corresponds to 27 µm above, 0 µm (at the surface), 15 µm below, and 30 µm below the electrocatalyst surface. From left to right,
each column of maps were obtained at 0 mA/cm2 (no reaction under open circuit conditions), –1.6 mA/cm2, –3.4 mA/cm2, –
14.7 mA/cm2, and –28.0 mA/cm2. The pH color scale and the scale bar in the bottom right-hand corner apply to all images.

65

4.3

Results and discussion
In the first column of Figure 4.7, at 0 mA/cm2, the solution pH is uniform throughout the z-direction.

The second row of pH maps at 0 µm defines the surface of the electrode; as the electrode is not flat, the highest
point of the electrode in the image area was chosen as the 0 µm height. Black regions in the pH map indicate
areas where no fluorescence was observed and therefore no electrolyte was present. The bottom row of pH
maps in Fig. 4.7 shows the solution pH within a trench in the microporous layer. As the current density increases
from left to right along a row, the local solution pH increases both within the trench of the microporous layer
and above the electrode surface. It is particularly interesting to note that the pH is not completely uniform over
the electrode surface, which can be most easily seen in the 0 µm and –15 µm height maps at a current density
of –3.4 mA/cm2. We are only able to observe this inhomogeneity at low current density where the pH gradient
built up is not large. As the current density is further increased and all catalyst particles become electrochemically
active we are no longer able to disentangle the pH gradient creation from individual locations along the catalyst.
This effect was repeatedly observed at multiple locations across the electrode surface, as shown in Figure 4.8.
We observe much smaller local variations at –1.6 mA/cm2 because the applied bias is smaller.

Figure 4.8: pH maps at three different locations along the electrode surface. Position 1 is off to the side of the electrolyte inlet,
position 2 is near the electrolyte inlet, and position 3 is near the electrolyte outlet. The first row shows pH maps all taken at –3.4
mA/cm2 and the second row shows pH maps all taken at –7.0 mA/cm2. We observe the hot spots for all 3 positions at -3.4
mA/cm2 and we do not observe the hot spots at –7.0 mA/cm2.

66
We can use pH as a proxy for the total current density, as each electron catalyzing either the HER or
the CO2 reduction reaction corresponds to the creation of one hydroxide ion in the electrolyte. Hence, higher
pH regions are indictive of higher activity. As the current density was further increased, the fluorescent signal
from the dye eventually saturates. To confirm that only electrochemically-active areas of the electrode were
responsible for locally increasing the solution pH, a map was obtained 9 µm above the electrode surface over a
region that was only partially covered with Cu (Figure 4.9). The electrolyte flowed left to right across the
electrode and the current density was set at –14.7 mA/cm2. As shown in Figure 4.9a, the left side of map has a
pH ~7, which was obtained above a region of the catalyst without Cu, while the right side of the map has a pH
~ 9, which was taken from above a region coated in Cu. It is clear from this map that regions of locally-high
pH only occur near regions of the electrode where hydroxide ions are being created via electrocatalysis. In
addition, we measured the pH under the same conditions as the back edge of the electrolyte (Figure 4.9b).

Figure 4.9: Influence of electrolyte flow on the spatial resolution of pH maps. (a) and (b) are pH maps stitched together, taken at
9 µm above the surface of the electrode and at a current density of –14.7 mA/cm2. In (a) the left half of the image has no Cu
while the right half has Cu catalyst. The electrolyte is flowing from left to right across the surface of the electrode. In (a) the left half
of the image is coated with Cu while the right half has no Cu catalyst. The electrolyte is flowing from left to right across the surface.
(c) shows a schematic that is not to scale of what the pH gradient looks like in both x-y and x-z planes. The area within the

67
orange circle in the x-y plane indicates where the Cu catalyst is located on the GDE. The position where image (a) was taken is
indicated in panel (c) by the red square and the red line labeled ‘a’. The position where image (b) was taken is indicated in panel
(c) by the red square and the red line labeled ‘b’.
When the average pH at the surface of the electrode is below pH 9.5 (J < –14.7 mA/cm2, or applied
potentials less negative than –0.7 V vs. RHE) the electrode mostly produces C1 products and H2. In contrast,
many C2 products were observed when the solution pH at the surface of the electrode was above pH 10 (J > –
28.0 mA/cm2, or at –1.0 V vs RHE). Potentials greater than –0.9 V vs RHE are required to produce these
higher current densities and C2 products, but the local pH also plays a role in suppressing the HER and
promoting CO2 reduction.17,39 The activity of CO2 reduction is independent, whereas the HER activity is greatly
dependent on the hydroxide ion concentration.17 For CO reduction on Cu, locally high pH conditions reduce
the free energy required for important steps along the reaction pathway to yield C2 products such as ethanol.16
Areas of locally high pH may also reduce the free energy for CO2 reduction pathways. Thus, our results indicate
that a highly-alkaline local pH increases the selectivity towards C2 products while decreasing the selectivity
towards the HER.

68
Figure 4.10: Influence of physical confinement on CO2 reduction performance. (a) Low-magnification SEM image of a Cu gas
diffusion electrode; (b) High-magnification SEM image of a Cu gas diffusion electrode with an overlay of the Cu signal obtained
from an EDS map, red shading indicates Cu covered regions. (c) Measured pH as a function of trench width. The orange data
points denote the average trench width. The error bars in the abscissa axis indicate the variation in trench width with the smallest
and largest end points denoting the thinnest and the widest point along the trench. The error bars in the ordinate axis represent the
standard deviation of pH values within the trench. (d) and (e) pH maps obtained from two representative trenches with different
widths located at different regions on the electrode at a distance of 8 µm below the electrode surface while operating at a current density
of –3.4 mA/cm2.
Finally, we explore the local pH within trenches in the microporous layer, as shown in Figure 4.10. The
trenches are randomly distributed throughout the electrode and have an average of width of 18.8 µm ± 8 µm
(Figure 4.10a). Figure 4.10b shows a higher-magnification SEM image of a crack, with an overlay of an energy
dispersive spectroscopy (EDS) map indicating regions that contain Cu (red shading). Figures 4.11 and 4.12
show more EDS maps of Cu, F, and C from different trenches. From this data we found that Cu is not only
coated on top of the microporous layer, but also at the bottom and on the sides of the trenches, suggesting that
CO2 reduction can be performed within these confined regions of the electrode. These trenches within the
microporous layer offer an interesting opportunity for studying the influence of physical confinement on the
CO2 reduction process in a GDE device. Accordingly, we employed our pH mapping techniques to these regions
of the electrode to see how the reduced dimensions of the device influence the local solution pH near the active
Cu electrocatalyst. At a current density of –3.4 mA/cm2, we found that the solution pH within a narrow trench
(~3.2 µm wide) was pH 9.5. We emphasize that this local pH was much higher than the pH of 7.3 within a
comparatively wider trench (~16.2 µm wide). Interestingly, even at this relatively low current density (–3.4
mA/cm2), the more confined electrochemically-active region produces a higher local pH than more open
regions, which serves to suppress HER without substantially impeding CO2 reduction.47,48 We also note that the
pH within a narrow trench is higher than the surrounding surface of the electrode (Figure 4.13). We proceeded
to map to elucidate the trend of local pH vs. trench width, as shown in Figure 4.10c. While we find that the
electrolyte flow is very small between the objective and the catalyst in our COMSOL flow model (Figure 4.14),
we only measured the pH within trenches that were perpendicular to the electrolyte flow to ensure that the flow
dynamics are as comparable as possible. We observed that as the trench width decreased, the local solution pH
within the trench increased. This observation is consistent with prior electrochemical studies,17,47 and has

69
important implications for the design of more active GDEs capable of performing selective CO2 reduction at
lower overpotentials with improved energy efficiency.

Figure 4.11: SEM and EDS maps of two locations on a GDE with 300 nm Cu. The SEM images at the left show the
location for all EDS maps in the corresponding row. From the EDS maps we can see that at the bottom of the cracks, there is Cu
while less carbon and PTFE (Fluorine signal) is present.

Figure 4.12: SEM and EDS maps of three locations on a GDE with 300 nm Cu. Red shading denotes carbon, green shading
denotes Cu, and blue indicates fluorine. From the EDS maps we can see that there is Cu deposited on the side walls of the trenches.

70

Figure 4.13: pH maps at three different locations along the electrode surface. All measurements were taken at –3.4 mA/cm2.
The image in the first row is taken at the surface of the electrode and the second row is at 8 µm below the surface (0 µm). We
observe that for all cases the pH within the narrow trench is higher than it is at the surface of the GDE.

Figure 4.14: shows a COMSOL simulation of the velocity of the electrolyte in the electrochemical cell with the objective submerged
in the electrolyte. (a) shows the velocity in the x-z plane with the inlet on the left and outlet on the right. (b) shows the velocity in the
y-z plane with the electrolyte flowing into the page. From these simulations it is clear that the velocity underneath the objective is
small.

71
We additionally confirm the experimentally-measured pH within the trenches for various widths and
current densities by simulating the solution pH using the charge transfer and bulk reactions in a two-dimensional
COMSOL model (Figure 4.15).49 Simulations were performed with a stationary COMSOL Multiphysics 5.5
model with a combination of the Laminar Flow Module and the Transport of Diluted Species Module. To
estimate the appropriate flow velocity close to the electrode surface, a three-dimensional COMSOL model with
the Laminar Flow Module simulating the electrolyte flow in the electrochemical cell and around the immersion
objective used for experiments was set up, see Fig. 4.14. The velocity at the inlet of the cell was experimentally
determined to be 1.3 mm/s. As expected, the flow velocity underneath the objective is found to be much lower
than around it. The average flow velocity 30 µm above the electrode surface was determined to be 0.14 µm/s.
This value was used as an input for the two-dimensional model of pH around a trench.
The geometry of the two-dimensional model of the pH is depicted in Fig 4.15d. At the inlet (left), we
assume a flow velocity of 0.14 µm/s, at the outlet (right) a zero-pressure condition is applied. On the electrode
surface as well as on the trench walls and at the bottom of the trench we apply no-slip boundary conditions.
Both convection and diffusion are taken into consideration for the transport properties. The inputs here are the
velocity field calculated with the laminar flow module and the diffusion coefficients of all species (see Table 4.1).
We assume a boundary layer thickness of 60 µm where we apply concentration boundary conditions as expected
for a CO2-saturated 100 mM KHCO3 electrolyte in equilibrium (see Table 4.1). The same concentrations are
used as inflow concentrations at the inlet.
On the surface of the electrode as well as at the bottom and the side wall of the trench we assume a
µ345

CO2 flux of 1 (3# ∙7. This value was determined experimentally by measuring the difference in the CO2 flow
rate with a flow meter before and after passing the GDE in the electrochemical cell. Note that this method only
provides

an

estimate

as

the

CO2

flux

though

the

GDE

is

not

homogeneous.

Also on the electrode surface, the trench bottom and walls we assume that there is catalytic activity. The chargetransfer reactions considered on these walls are
CO2 (aq) + H2O + 2e- à CO + 2OH2H+ + 2e- à H2
2H2O + 2e- à H2 + 2OHThe source terms are determined with
𝑅8 = −∑

𝑠9 ∙ 𝐽 ∙ 𝐹𝐸8
𝑛9 ∙ 𝐹

72
with 𝑠9 the stoichiometric coefficient of equation j, J the current density, 𝐹𝐸8 the Faradaic efficiency of
species n determined experimentally (see Fig 4.5, the Faradaic efficiency for CO is assumed to be the sum of the
Faradaic Efficiency of all carbon products), 𝑛9 the number of transferred electrons, and F Faraday’s constant.
Furthermore, bulk carbonate reactions are assumed to take place in the whole electrolyte body:
CO2(aq)+H2O à H++HCO3HCO3- à H++CO32CO2(aq)+OH- à HCO3HCO3-+OH- à H2O+CO32H2O à H++OHThe source terms are calculated with
𝑅8 = O 𝑠9 ∙ (𝑘9 Q 𝑐8 − 𝑘#9 Q 𝑐8 )

7$ ;'

7$ :'

where 𝑠9 is the stoichiometric coefficient of reaction j, 𝑘9 the forward rate constant, 𝑘#9 the reverse

reaction rate constant, and 𝑘#9 = =$ with the equilibrium constant 𝐾9 (see Table 4.1). 𝑐8 is the concentration of

species n.
Table 4.1: Model parameters
Diffusion Coefficients (Weng et al, Phys. Chem. Chem. Phys., 2018, 20, 16973-16984)
Species
Value
K+
H+
OHCO2
HCO3CO32Equilibrium Concentrations
Species

𝑐𝑚'
𝑐𝑚
9.311 ∙ 10%&
𝑐𝑚
5.293 ∙ 10%&
𝑐𝑚
1.91 ∙ 10%&
𝑐𝑚'
%&
1.185 ∙ 10
𝑐𝑚'
%&
0.91 ∙ 10
1.957 ∙ 10%&

Value

K+

𝑚𝑜𝑙
100 (

H+

1.58 ∙ 10%)

OHCO2

𝑚𝑜𝑙
𝑚(
𝑚𝑜𝑙
6.31 ∙ 10%& (
𝑚𝑜𝑙
37.13 (

73
𝑚𝑜𝑙
𝑚(

HCO3-

99.94

CO32-

3.02 ∙ 10%'

𝑚𝑜𝑙
𝑚(

Forward Rate Constants (Schulz et al, Mar. Chem., 2006,100,53-65)
k1

3.71 ∙ 10%'

k2

59.44

k3

2.23 ∙ 10(

k4
kw

𝑚𝑜𝑙 ∙ 𝑠
6 ∙ 10
𝑚𝑜𝑙 ∙ 𝑠
𝑚𝑜𝑙
1.4 ∙ 10%(
𝐿∙𝑠

Equilibrium Constants (Schulz et al, Mar. Chem., 2006,100,53-65)
K1
K2
K3
K4
Kw

𝑚𝑜𝑙
𝑚𝑜𝑙
%./.('
10
𝐾.
𝐾0
𝐾'
𝐾0
𝑚𝑜𝑙'
10%.) '
10%+.(-

For the model (Figure 4.15a), we assumed a constant flux of CO2 of 1 µmol cm-2 s-1 through the planar
electrode surface, the trench walls and trench bottom (see SI for details). We used the experimental current
density of –3.4 mA/cm2 at the same boundaries because SEM and EDS images of our samples show that copper
coats the surface, trench side walls and trench bottom (see Figures 4.11 and 4.12). We assume a boundary layer
thickness of 60 µm where a concentration boundary condition matching the concentration of CO2-saturated
100 mM KHCO3 electrolyte was applied. The electrolyte flow from left to right was also taken into consideration
in the model. The appropriate flow velocity is determined with a three-dimensional COMSOL model of the
flow in the electrochemical cell (Figure 4.14).

74

Figure 4.15: Simulations of local pH within and around trenches of various dimensions in the GDE. (a) Schematic of the model
used for simulations indicating regions of CO2 flux (white), current density (orange), concentration boundary conditions (green), and
electrolyte flow (blue). pH map in the x-z plane at a uniform current density of –3.4 mA/cm2 for a trench with a depth of 50 µm
deep and a width of (b) 20µm and (d) 5 µm. (c) pH map in the x-z plane for a trench [50 deep x 5µm wide] with an average
current density of –3.4 mA/cm2 where the current density in the trench is twice as high as the current density on the surface. The
CO2 flux is constant through all surfaces and boundary conditions are kept the same for all simulations. The pH scale bar applies
to all pH profiles (b)-(d).
As we expected, the pH was increased close to the electrode surface and inside the trench due to charge
transfer reactions locally creating hydroxide ions. The pH decreased further away from the electrode surface
due to convection and diffusion within the bulk electrolyte. There was a dip in the pH profile above the trench
because more CO2 comes through the electrode at this point and acidifies the electrolyte close to the trench.
Additionally, there is increased CO2 flux here because CO2 is able to diffuse not only through the bottom of the
trench but also through the sidewalls. This feature was also observed experimentally, as shown in the pH map
in Figure 4.7 for –3.4 mA/cm2 at –15 µm. We note that the pH gradient is nearly symmetric above the crack
because the electrolyte velocity is low. At higher current densities we observed that pH increases in the trench
and on the surface of the electrode as expected (Figure 4.16).

75

Figure 4.16: shows COMSOL simulations of the pH profile in a trench that is 5µm wide and 50 µm deep at different current
densities. The current density is constant over all surfaces. (a) shows the pH profile at –3.4 mA/cm2, (b) shows the pH profile at
–7.0 mA/cm2, and (c) shows the pH profile at –14.7 mA/cm2. From these simulations we observe that the pH near the catalyst
layer increases as the current density increases.

Comparing Figure 4.17b and Figure 4.17d it is clear that the pH is considerably lower in the wider
trenches than the narrower trenches which is in agreement with experimental results (Fig 4.10c). In Figure 4.17c,
we simulated a trench with the same dimensions as the trench shown in Figure 4.17d. However, we modeled a
non-uniform catalyst activity where the activity in the trench is twice as high as it is on the electrode surface,
while maintaining the same average current density over the whole electrode. We observed that in this case, the
pH within the trench is higher than the pH above the surface of the electrode which we did not observe in
experiment (see Figure 4.13). This discrepancy leads to the conclusion that the experimentally observed higher

76
pH in narrow trenches cannot only be explained by the confinement of the trench, but must also be due to
increased catalytic activity within the trench.
4.4

Conclusion
We have employed confocal fluorescent microscopy to elucidate how the operando local pH changes with

current density as a function of distance above and below the surface of a Cu-based GDE. It is clear from the
experimentally obtained pH maps that there are non-uniform hotspots of locally high pH across the catalyst
even at relatively low overpotentials. Through experimental results confirmed by simulations, we show that the
catalyst within narrow trenches is more active than catalyst at the surface of the electrode. We also observed
that the pH was higher in narrow trenches as opposed to wider trenches, and we confirmed this result with
COMSOL simulations. Further work must be done to understand why catalyst in narrow trenches performs
better than catalyst in wider trenches, and this will be the subject of ongoing studies. Nevertheless, the ability
to locally image the solution pH in three dimensions (x, y, and z) with micron spatial resolution is an important
tool for understanding and identifying which part of the catalyst is most productive under real operating
conditions. Our results have therefore demonstrated that the overpotential required to perform selective CO2
reduction can be reduced within narrow trenches. We anticipate that this knowledge will help inform the design
and construction of more efficient CO2 reduction devices.

BIBLIOGRAPHY CHAPTER 4
1.
2.
3.
4.
5.
6.
7.

Hori, Yoshio Kikuchi, Katsuhei Suzuki, S. Production of CO and CH4 in the Electrochemical Reduction
of CO2 at Metal Electrodes in Aqueous Hydrogencarbonate Solution. Chem. Lett. 1985, 1695–1698.
Kortlever, R.; Shen, J.; Schouten, K. J. P.; Calle-Vallejo, F.; Koper, M. T. M. Catalysts and Reaction
Pathways for the Electrochemical Reduction of Carbon Dioxide. J. Phys. Chem. Lett. 2015, 6 (20), 4073–
4082.
Bagger, A.; Ju, W.; Varela, A. S.; Strasser, P.; Rossmeisl, J. Electrochemical CO2 Reduction: A
Classification Problem. ChemPhysChem 2017, 18 (22), 3266–3273.
Kuhl, K.; Jaramillo, T. F. New Insights into the Electrochemical Reduction of Carbon Dioxide on
Metallic Copper Surfaces. 2012, 7050–7059.
Cave, E. R.; Montoya, J. H.; Kuhl, K. P.; Abram, D. N.; Hatsukade, T.; Shi, C.; Hahn, C.; Nørskov, J. K.;
Jaramillo, T. F. Electrochemical CO2 Reduction on Au Surfaces: Mechanistic Aspects Regarding the
Formation of Major and Minor Products. Phys. Chem. Chem. Phys. 2017, 19 (24), 15856–15863.
Endrődi, B.; Bencsik, G.; Darvas, F.; Jones, R.; Rajeshwar, K.; Janáky, C. Continuous-Flow
Electroreduction of Carbon Dioxide. Prog. Energy Combust. Sci. 2017, 62, 133–154.
Bohlen, B.; Wastl, D.; Radomski, J.; Sieber, V.; Vieira, L. Electrochemical CO2 Reduction to Formate on
Indium Catalysts Prepared by Electrodeposition in Deep Eutectic Solvents. Electrochem. commun. 2020,

77

8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.

110 (October 2019), 106597.
Singh, M. R.; Kwon, Y.; Lum, Y.; Ager, J. W.; Bell, A. T. Hydrolysis of Electrolyte Cations Enhances the
Electrochemical Reduction of CO2 over Ag and Cu. J. Am. Chem. Soc. 2016, 138 (39), 13006–13012.
Jeanty, P.; Scherer, C.; Magori, E.; Wiesner-Fleischer, K.; Hinrichsen, O.; Fleischer, M. Upscaling and
Continuous Operation of Electrochemical CO2 to CO Conversion in Aqueous Solutions on Silver Gas
Diffusion Electrodes. J. CO2 Util. 2018, 24 (January), 454–462.
Weekes, D. M.; Salvatore, D. A.; Reyes, A.; Huang, A.; Berlinguette, C. P. Electrolytic CO2 Reduction in
a Flow Cell. 2018.
Rabinowitz, J. A.; Kanan, M. W. The Future of Low-Temperature Carbon Dioxide Electrolysis Depends
on Solving One Basic Problem. Nat. Commun. 2020, 11 (1), 10–12.
Kim, B.; Hillman, F.; Ariyoshi, M.; Fujikawa, S.; Kenis, P. J. A. Effects of Composition of the
Microporous Layer and the Substrate on Performance in the Electrochemical Reduction of CO2 to CO.
J. Power Sources 2016, 312, 192–198.
Davis, S. J.; Lewis, N. S.; Shaner, M.; Aggarwal, S.; Arent, D.; Azevedo, I. L.; Benson, S. M.; Bradley, T.;
Brouwer, J.; Chiang, Y.; et al. Net-Zero Emissions Energy Systems. 2018, 9793.
Birdja, Y. Y.; Pérez-Gallent, E.; Figueiredo, M. C.; Göttle, A. J.; Calle-Vallejo, F.; Koper, M. T. M.
Advances and Challenges in Understanding the Electrocatalytic Conversion of Carbon Dioxide to Fuels.
Nat. Energy 2019, 4 (9), 732–745.
Leenheer, A. J.; Atwater, H. A. Imaging Water-Splitting Electrocatalysts with PH-Sensing Confocal
Fluorescence Microscopy. J. Electrochem. Soc. 2012, 159 (9), H752–H757.
Liu, X.; Schlexer, P.; Xiao, J.; Ji, Y.; Wang, L.; Sandberg, R. B.; Tang, M.; Brown, K. S.; Peng, H.; Ringe,
S.; et al. pH Effects on the Electrochemical Reduction of CO2 towards C2 Products on Stepped Copper.
Nat. Commun. 2019, 10 (1).
Hall, A. S.; Yoon, Y.; Wuttig, A.; Surendranath, Y. Mesostructure-Induced Selectivity in CO2 Reduction
Catalysis. J. Am. Chem. Soc. 2015, 137 (47), 14834–14837.
Zhang, Z.; Melo, L.; Jansonius, R. P.; Habibzadeh, F.; Grant, E. R.; Berlinguette, C. P. PH Matters When
Reducing CO2 in an Electrochemical Flow Cell. ACS Energy Lett. 2020, 5 (10), 3101–3107.
Higgins, Drew Hahn, Christopher Chengxiang, Xiang Jaramillo, Thomas F. Weber, A. Z. Gas-Diffusion
Electrodes for Carbon Dioxide Reduction: A New Paradigm. ACS Energy Lett. 2019, 4, 317–324.
Inaba, M.; Jensen, A. W.; Sievers, G. W.; Escudero-Escribano, M.; Zana, A.; Arenz, M. Benchmarking
High Surface Area Electrocatalysts in a Gas Diffusion Electrode: Measurement of Oxygen Reduction
Activities under Realistic Conditions. Energy Environ. Sci. 2018, 11 (4), 988–994.
Nitopi, S.; Bertheussen, E.; Scott, S. B.; Liu, X.; Engstfeld, A. K.; Horch, S.; Seger, B.; Stephens, I. E. L.;
Chan, K.; Hahn, C.; et al. Progress and Perspectives of Electrochemical CO2 Reduction on Copper in
Aqueous Electrolyte. Chem. Rev. 2019, 119 (12), 7610–7672.
Singh, M. R.; Clark, E. L.; Bell, A. T. Effects of Electrolyte, Catalyst, and Membrane Composition and
Operating Conditions on the Performance of Solar-Driven Electrochemical Reduction of Carbon
Dioxide. Phys. Chem. Chem. Phys. 2015, 17, 18924–18936.
Weng, L.-C.; Bell, A. T.; Weber, A. Z. Modeling Gas-Diffusion Electrodes for CO2 Reduction. Phys.
Chem. Chem. Phys. 2018, No. 20, 16973–16984.
García de Arquer, F. P.; Dinh, C. T.; Ozden, A.; Wicks, J.; McCallum, C.; Kirmani, A. R.; Nam, D. H.;
Gabardo, C.; Seifitokaldani, A.; Wang, X.; et al. CO2 Electrolysis to Multicarbon Products at Activities
Greater than 1 A Cm−2. Science (80-. ). 2020, 367 (6478), 661–666.
Dinh, C.; Burdyny, T.; Kibria, G.; Seifitokaldani, A.; Gabardo, C. M.; Arquer, F. P. G. De; Kiani, A.;
Edwards, J. P.; Luna, P. De; Bushuyev, O. S.; et al. CO2 Electroreduction to Ethylene via HydroxideMediated Copper Catalysis at an Abrupt Interface. Science (80-. ). 2018, 360, 783–787.
Wang, L.; Nitopi, S. A.; Bertheussen, E.; Orazov, M.; Morales-Guio, C. G.; Liu, X.; Higgins, D. C.; Chan,
K.; Nørskov, J. K.; Hahn, C.; Jaramillo, T. F. Electrochemical Carbon Monoxide Reduction on

78

27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.

Polycrystalline Copper: Effects of Potential, Pressure, and PH on Selectivity toward Multicarbon and
Oxygenated Products. ACS Catal. 2018, 8 (8), 7445–7454.
Nesbitt, N.; Smith, W. Water Activity Regulates CO2 Reduction in Gas-Diffusion Electrodes. ChemRxiv
2021, No. January.
Suter, S.; Haussener, S. Optimizing Mesostructured Silver Catalysts for Selective Carbon Dioxide
Conversion into Fuels. Energy Environ. Sci. 2019, 12 (5), 1668–1678.
Monteiro, M. C. O.; Koper, M. T. M. Measuring Local PH in Electrochemistry. Curr. Opin. Electrochem.
2021, 25, 100649.
Rudd, N. C.; Cannan, S.; Bitziou, E.; Ciani, I.; Whitworth, A.; Unwin, P. R. Fluorescence Confocal Laser
Scanning Microscopy as a Probe of PH Gradients in Electrode Reactions and Surface Activity Nicola.
Anal. Chem. 2005, 77, 6205–6217.
Bouffier, L.; Doneux, T. Coupling Electrochemistry with in Situ Fluorescence (Confocal) Microscopy.
Curr. Opin. Electrochem. 2017, 6 (1), 31–37.
Pande, N.; Chandrasekar, S. K.; Lohse, D.; Mul, G.; Wood, J. A.; Mei, B. T.; Krug, D. Electrochemically
Induced PH Change: Time-Resolved Confocal Fluorescence Microscopy Measurements and
Comparison with Numerical Model. J. Phys. Chem. Lett. 2020, 11 (17), 7042–7048.
Vitt, J. E.; Engstrom, R. C. Imaging of Oxygen Evolution and Oxide Formation Using Quinine
Fluorescence. Anal. Chem. 1997, 69 (6), 1070–1076.
Bowyer, W. J.; Xie, J.; Engstrom, R. C. Fluorescence Imaging of the Heterogeneous Reduction of
Oxygen. Anal. Chem. 1996, 68 (13), 2005–2009.
Nesbitt, N.; Smith, W. A. Operando Topography and Mechanical Property Mappinig of CO2 Reduction
Gas-Diffusion Electrodes Operating at High Currrent Densities. J. Electrochem. Soc. 2021.
Botz, A.; Clausmeyer, J.; Öhl, D.; Tarnev, T.; Franzen, D.; Turek, T.; Schuhmann, W. Local Activities of
Hydroxide and Water Determine the Operation of Silver-Based Oxygen Depolarized Cathodes. Angew.
Chemie - Int. Ed. 2018, 57 (38), 12285–12289.
Dieckhöfer, S.; Öhl, D.; Junqueira, J. R. C.; Quast, T.; Turek, T.; Schuhmann, W. Probing the Local
Reaction Environment During High Turnover Carbon Dioxide Reduction with Ag-Based Gas Diffusion
Electrodes. Chem. - A Eur. J. 2021, 27 (19), 5906–5912.
Lu, X.; Zhu, C.; Wu, Z.; Xuan, J.; Francisco, J. S.; Wang, H. In Situ Observation of the PH Gradient near
the Gas Diffusion Electrode of CO2 Reduction in Alkaline Electrolyte. J. Am. Chem. Soc. 2020, 142 (36),
15438–15444.
Yang, K.; Kas, R.; Smith, W. A. In Situ Infrared Spectroscopy Reveals Persistent Alkalinity near Electrode
Surfaces during CO2 Electroreduction. J. Am. Chem. Soc. 2019, 141 (40).
Nitopi, S.; Bertheussen, E.; Scott, S. B.; Liu, X.; Engstfeld, A. K.; Horch, S.; Seger, B.; Stephens, I. E. L.;
Chan, K.; Hahn, C.; et al. Progress and Perspectives of Electrochemical CO2 Reduction on Copper in
Aqueous Electrolyte. 2019.
Wang, Y.; Shen, H.; Livi, K. J. T.; Raciti, D.; Zong, H.; Gregg, J.; Onadeko, M.; Wan, Y.; Watson, A.;
Wang, C. Copper Nanocubes for CO2 Reduction in Gas Diffusion Electrodes. Nano Lett. 2019.
Nguyen, T. N.; Dinh, C. T. Gas Diffusion Electrode Design for Electrochemical Carbon Dioxide
Reduction. Chem. Soc. Rev. 2020, 49 (21), 7488–7504.
Ripatti, D. S.; Veltman, T. R.; Kanan, M. W. Carbon Monoxide Gas Diffusion Electrolysis That Produces
Concentrated C2 Products with High Single-Pass Conversion. Joule 2019, 3 (1), 240–256.
Welch, A. J.; Dunn, E.; Duchene, J. S.; Atwater, H. A. Bicarbonate or Carbonate Processes for Coupling
Carbon Dioxide Capture and Electrochemical Conversion. ACS Energy Lett. 2020, 5 (3), 940–945.
Hakonen, A.; Hulth, S. A High-Performance Fluorosensor for PH Measurements between 6 and 9.
Talanta 2010, 80 (5), 1964–1969.
Hori, Y. Electrochemical CO 2 Reduction on Metal Electrodes. Mod. Asp. Electrochem. 2008, No. 42, 89–
189.

79
47.
48.
49.

Welch, A. J.; Duchene, J. S.; Tagliabue, G.; Davoyan, A.; Cheng, W. H.; Atwater, H. A. Nanoporous
Gold as a Highly Selective and Active Carbon Dioxide Reduction Catalyst. ACS Appl. Energy Mater. 2019,
2 (1), 164–170.
Hall, A. S.; Yoon, Y.; Wuttig, A.; Surendranath, Y. Mesostructure-Induced Selectivity in CO2 Reduction
Catalysis. J. Am. Chem. Soc. 2015, 137 (47), 14834–14837.
Schulz, K. G.; Riebesell, U.; Rost, B.; Thoms, S.; Zeebe, R. E. Determination of the Rate Constants for
the Carbon Dioxide to Bicarbonate Inter-Conversion in PH-Buffered Seawater Systems. Mar. Chem.
2006, 100 (1–2), 53–65.

80
Chapter 5

HOT HOLE COLLECTION AND PHOTOELECTROCHEMICAL CO2 REDUCTON
WITH PLASMONC Au/p-GaN PHOTOCATHODES
5.1

Introduction
Harvesting non-equilibrium hot carriers from plasmonic-metal nanostructures offers unique

opportunities for driving photochemical reactions at the nanoscale. Despite numerous examples of hot electrondriven processes, the realization of plasmonic systems capable of harvesting hot holes from metal nanostructures
has eluded the nascent field of plasmonic photocatalysis. Here, we fabricate gold/p-type gallium nitride (Au/pGaN) Schottky junctions tailored for photoelectrochemical studies of plasmon-induced hot-hole capture and
conversion. Despite the presence of an interfacial Schottky barrier to hot-hole injection of more than 1 eV across
the Au/p-GaN heterojunction, plasmonic Au/p-GaN photocathodes exhibit photoelectrochemical properties
consistent with the injection of hot holes from Au nanoparticles into p-GaN upon plasmon excitation. The
photocurrent action spectrum of the plasmonic photocathodes faithfully follows the surface plasmon resonance
absorption spectrum of the Au nanoparticles and open-circuit voltage studies demonstrate a sustained
photovoltage during plasmon excitation. Comparison with Ohmic Au/p-NiO heterojunctions confirms that the
vast majority of hot holes generated via interband transitions in Au are sufficiently hot to inject above the 1.1
eV interfacial Schottky barrier at the Au/p-GaN heterojunction. We further investigated plasmon-driven
photoelectrochemical CO2 reduction with the Au/p-GaN photocathodes, and observed improved selectivity
for CO production over H2 evolution in aqueous electrolytes. Taken together, our results offer experimental
validation of photoexcited hot holes more than 1 eV below the Au Fermi level and demonstrate a
photoelectrochemical platform for harvesting hot carriers to drive solar-to-fuel energy conversion.
The generation of non-equilibrium “hot” electron-hole pairs via surface plasmon decay within metal
nanostructures holds great promise for initiating and controlling chemical reactions at the nanoscale.1-6 However
the capture and conversion of photoexcited hot carriers presents challenges, given their very short mean-free
paths (lmfp ~2–20 nm) and excited-state lifetimes (t ~fs–ps).7-15 Hot carrier collection schemes typically involve
the formation of an interfacial Schottky barrier (ΦB) between plasmonic metals (e.g. Au) and wide band gap
semiconductors (e.g. n-type TiO2) to quickly capture hot electrons in a plasmonic photosensitization strategy
similar to that employed in dye-sensitized solar cells (Figure 5.1a). Although numerous optoelectronic systems
have been devised to harness plasmonic hot electrons for sub-band gap photodetection16-21 and plasmon-driven

81
photocatalysis,22-32 little is known about hot holes derived from surface plasmon decay. Recent theoretical
calculations have predicted an asymmetry in the energy distributions between hot electrons and hot holes relative
to the metal Fermi level (EF) in common plasmonic metals like Au and Cu.8-12 Due to the high density of
electronic d-band states, photoexcitation above the interband threshold of the metal (d-band to sp-band
transition) can generate hot holes that are much “hotter” (further away from the Fermi level) than hot electrons
(Figure 5.1b).8-12 In Au nanostructures, an imbalance in hot carrier distributions would be expected to occur for
photon energies hν > 1.8 eV.8 Resonant optical excitation of the dipole plasmon mode in spherical Au
nanoparticles (hv ~2.4 eV) should therefore preferentially produce hot holes within the Au d-band that reside
far below the Au Fermi level.8-12,33 This substantial asymmetry between the energy distributions of hot carriers
implies a greater collection efficiency of hot holes relative to hot electrons for a comparable Schottky barrier
height. The strong oxidizing power of these hot d-band holes also offers the potential for driving various
oxidation reactions if they could be transferred to an appropriate catalyst. Indeed, photo-oxidation of adsorbed
citrate molecules in the plasmon-driven synthesis of colloidal Ag and Au nanoprisms is known to proceed more
efficiently via “hot” d-band holes as compared to “warm” sp-band holes.34-39 Strategies that can efficiently and
selectively harvest hot holes from metal nanostructures would therefore offer significant benefits for plasmonic
photochemistry.40-42

Figure 5.1: Hot carrier collection across an interfacial Schottky barrier at metal/semiconductor heterojunctions. (a) Qualitative
energy band diagram of a plasmonic metal (e.g. Au) in physical contact with an n-type semiconductor (e.g. TiO2), depicting the
conduction band edge (ECB), valence band edge (EVB), band gap (EG), Fermi level (EF), and the interfacial Schottky barrier

82
(ΦB). Plasmon excitation creates hot electrons (red) and hot holes (blue) above and below the EF of Au, respectively, with a
distribution of energies governed by the metal band structure and the incident photon energy (hv = 2.4 eV). Only those hot
electrons with sufficient energies above ΦB (indicated by dashed line) can surmount the interfacial barrier and populate available
CB levels of the n-type semiconductor support. (b) Qualitative energy band diagram of a plasmonic metal (e.g. Au) in physical
contact with a p-type semiconductor (e.g. p-GaN), depicting ECB, EVB, EG, EF, and ΦB. Plasmon excitation creates hot electrons
(red) and hot holes (blue) above and below the EF of Au, respectively. Only those hot holes with sufficient energies below ΦB
(indicated by dashed line) can surmount the interfacial barrier and populate available VB levels of the p-type semiconductor
support.
To date, however, nearly all studies of hot carrier collection have focused on the capture and
conversion of hot electrons with n-type semiconductors (Figure 5.1a). Comparatively fewer studies have
examined the sensitization of wide band gap p-type semiconductors, in which the plasmonic metal injects hot
holes directly into the valence band of an adjoining p-type semiconductor support (Figure 5.1b). Such a
plasmonic photosensitization scheme would enable adsorbed molecules to harvest hot electrons directly from
metal nanostructures while obviating the need for sacrificial reagents that are often used to facilitate charge
separation. Despite much promise, harvesting hot holes from metal nanostructures is more challenging than
hot electron collection, given the relatively short mean-free path (lmfp ~ 5-10 nm) of hot holes 1-2 eV below
the EF compared to hot electrons 1-2 eV above the EF (lmfp ~20 nm).8-11 There are also far fewer wide band
gap p-type semiconductors available for sensitization than n-type semiconductors. Although p-type nickel
oxide (p-NiO) has been shown to serve as a carrier-selective contact for hot-hole collection from photoexcited
Au nanoparticles,43,44 the absence of a suitable Schottky barrier at the Au-NiO interface permits the collection
of holes from the metal Fermi level (EF) down to the photon energy (EF – hν). Such a system is therefore
incapable of selectively probing the population of “hot” holes generated deep below the metal Fermi level
upon plasmon excitation. As a result, the energy distribution and associated prevalence of hot holes in metal
nanostructures upon photoexcitation is not well understood.
Here, we employ p-type gallium nitride (p-GaN) as a semiconductor support for photoelectrochemical
studies of hot-hole collection from Au nanoparticles. As a wide band gap (Eg ~3.4 eV) semiconductor that
exhibits p-type conductivity, p-GaN is an ideal semiconductor support for investigating hot-hole collection
above the interband threshold of a plasmonic nanoantenna. Importantly, the sizable Schottky barrier (ΦB > 1
eV) established across the Au/p-GaN interface provides a suitable platform for probing the existence of very
hot holes deep below the metal Fermi level upon photoexcitation of Au nanoparticles. Our
photoelectrochemical studies show that plasmonic Au/p-GaN photocathodes indeed support cathodic

83
photocurrents consistent with the collection of plasmon-induced holes that are more than 1 eV below the Au
Fermi level and the measured action spectrum faithfully follows the surface plasmon resonance of the Au
nanoparticles. Significantly, open-circuit voltage measurements demonstrate a sustained plasmonic
photovoltage across the metal-semiconductor heterojunction whose sign is consistent with the injection of
hot holes into the p-GaN support. We further used these Au/p-GaN photocathodes for plasmon-driven CO2
reduction, demonstrating the utility of these plasmonic photocathodes for artificial photosynthesis. These
results demonstrate the feasibility of harvesting hot holes from plasmonic-metal nanostructures and open a
route for the design of plasmonic photocathodes that can capture visible light to drive photochemistry.
5.2

Fabrication and characterization of Au/p-GaN
Plasmonic Au/p-GaN photocathodes were constructed via evaporation of Au thin-films onto

commercial p-type GaN/sapphire substrates (c-axis 0001 orientation) (5 μm thick GaN) (Pam-Xiamen).
Immediately before Au deposition, the p-GaN substrates were first pre-treated with dilute NH4OH solution
(0.02% v/v%) for 30 s to remove native oxide, followed by 30 s of copious washing in Nanopure water. It
was empirically found that such surface treatments were critical for achieving good device performance. The
p-GaN/sapphire substrate was then blown dry with N2 gas and loaded into the vacuum chamber. A 1.5 nmthick film of Au was then deposited onto the p-GaN surface using electron-beam physical vapor deposition
at a base pressure of ca. 1 x 10-7 torr and a deposition rate of 1.0 Å s-1. The Au/p-GaN films were then
annealed in ambient air at 300 °C for 1 h to ensure coalescence of the discontinuous Au thin-film into Au
nanoparticles and achieve good adhesion with the underlying p-GaN surface. The bare GaN-on-sapphire
substrate is optically transparent and displays a direct optical band gap of ca. 3.35 eV (Figure 5.2), consistent
with the 3.4 eV band gap expected for GaN.45-48 The wide band gap of p-GaN ensures that any visible-light
response observed at photon energies below 3.4 eV from the Au/p-GaN device can be attributed to hot-hole
injection from the Au nanoparticles. Scanning electron microscopy (SEM) shows Au nanoparticles of
diameter, d = 8.2 ± 1.6 nm uniformly distributed across the p-type GaN surface after briefly annealing in
ambient air at 300 °C for 1 h (Figure 5.3a,b). The absence of stabilizing surfactants and molecular linkers is
advantageous for establishing direct physical contact at the Au/p-GaN interface while also exposing a clean
Au surface for catalysis. It is further noted that no adhesion layer is required to construct the metalsemiconductor heterojunction, excluding any possible contribution from interfacial metal layers on the device
operation. The Au/p-GaN substrate adopts a purple color and exhibits a prominent absorption peak in the
visible region at ca. 570 nm, attributable to the surface plasmon resonance of Au nanoparticles (Figure 5.3c,

84
red curve). The fringes present in both absorption spectra are due to Fabry-Pérot interferences within the
high-index GaN layer.

Figure 5.2: Optical properties of p-type GaN (p-GaN) substrate. (a) Absorption of p-GaN substrate, demonstrating strong
absorption in the UV region with no significant features across the visible regime. Thus, any visible-light features observed from the
Au/p-GaN system can be attributed to the surface plasmon resonance of the Au nanoparticles. (b) Tauc plot of p-GaN indicates
an optical band gap of EG = 3.35 eV, consistent with the expected EG of 3.4 eV for GaN.

Figure 5.3: Plasmonic Au/p-GaN photocathode device structure. (a) Schematic of Au/p-GaN photocathode on sapphire
substrate. (b) SEM image with corresponding size-distribution histogram of Au nanoparticles (mean diameter, d = 8.2 ± 1.6
nm) on p-GaN substrate. (c) Absorption spectra of plasmonic Au/p-GaN photocathode (red curve) compared to bare p-GaN
substrate (blue curve). The plasmonic device shows a prominent surface plasmon resonance feature due to the Au nanoparticles at
ca. 570 nm. Inset shows a digital photograph of the colorless p-GaN substrate and the purple Au/p-GaN device. (d) Solid-state

85
current-voltage (I-Eappl) behavior from Au/p-GaN heterostructures. Ohmic contact to the p-GaN substrate was achieved through
deposition of a thin-film Au/Ni alloy (top panel). In contrast, a metal-semiconductor Schottky diode was obtained across the
Au/p-GaN heterojunction (bottom panel). Fitting of these data yields a Schottky barrier height of ΦB = 1.1 eV across the
Au/p-GaN interface.
Solid-state electrical measurements were conducted to verify that a Schottky barrier (ΦB) was
established across the Au/p-GaN interface. Ohmic contacts to the p-GaN substrate were fabricated
according to previous literature protocols49,50 via co-deposition of a 10 nm-thick Ni/Au (50/50 atomic %)
alloy followed by annealing in ambient air for 1 h at 500 °C. A metal-semiconductor Schottky junction was
then constructed from 100 nm-thick Au contacts. Electrical measurements were conducted under an optical
microscope using piezoelectric microprobes (Imina Technologies, miBots™) to electrically address the
contact pads on the p-GaN substrate. As shown in Figure5.3d (top panel), Ohmic behavior was observed
when both probes were electrically connected to the Ni/Au alloy contacts. In contrast, rectifying device
characteristics were observed when one of the microprobes was moved onto the Au contact pad (Figure 5.3d,
bottom panel). Fitting of these data to the diode equation yields a Schottky barrier height of ΦB = 1.1 eV
across the Au/p-GaN heterojunction, similar to the 1.2 eV barrier previously observed for Au/p-GaN
contacts.51 The Au/p-GaN photocathode therefore provides a suitable photoelectrochemical platform to
probe the production of hot holes with energies in excess of 1 eV in Au nanoparticles.
5.3

Electrochemical studies of Au/p-GaN
Photoelectrochemical studies were performed using a potentiostat in a three-electrode configuration

with the Au/p-GaN photocathode as working electrode, a platinum wire mesh counter electrode, and a
saturated calomel electrode (SCE) as the reference electrode. The electrolyte (50 mM K2CO3) was sparged
with N2 gas for 30 min prior to experiments, which were performed under a N2 blanket. All electrochemical
potentials are reported with respect to the reversible hydrogen electrode (RHE). Mott-Schottky analysis of
electrochemical impedance data obtained at 2 kHz confirms the p-type conductivity of the bare GaN films
(Figure 5.4). From a linear fit of the data we obtained a carrier concentration of ca. 1 x 1019 cm-3, similar to
the acceptor doping level of NA = 3-7 x 1018 cm-3 specified by the manufacturer. The flat-band potential (Efb)
is ca. 2.0 VRHE (V vs. RHE), consistent with prior reports.45-47 We estimate the width of the depletion region
(Wd) within the p-GaN substrate to be ca. 11–20 nm, implying that hot-hole tunneling through the barrier can
be neglected. Any hot holes collected by the underlying p-GaN support upon plasmon excitation must

86
therefore possess sufficient energy to surmount the ~1 eV Schottky barrier at the metal-semiconductor
interface.

Figure 5.4: Mott-Schottky plot of electrochemical impedance data obtained from bare p-GaN photocathodes obtained at 2 kHz.
The negative slope confirms the p-type character of the GaN substrates used herein. From a linear fit of the data we obtain a
carrier concentration of ca. 1 x 1019 cm-3, similar to the acceptor doping level of NA = 3-7 x 1018 cm-3 specified by the
manufacturer, and a flat-band potential (Efb) of ca. 2.0 VRHE (V vs. RHE).
The potential-current (E-J) behavior of the plasmonic photocathode was assessed via linear sweep
voltammetry under periodic (0.5 Hz), visible-light excitation (λ > 495 nm) at a scan rate of 10 mV s-1 to
simultaneously monitor the dark current (J) and the photocurrent (Jph). As shown in Figure 5.5a (red curve),
the Au/p-GaN device displayed cathodic photocurrents (Jph) along the potential sweep consistent with the
collection of hot holes across the metal-semiconductor heterojunction. The Au/p-GaN photocathode
exhibits an onset potential (Eon) of ca. 0.4 VRHE, and attained a Jph of ca. 1.8 μA cm-2 at −0.8 VRHE. Cyclic
voltammograms of bare p-GaN and Au/p-GaN photocathodes under dark and light conditions are shown in
Figure 5.6. The bare p-GaN photocathode exhibits no discernable Jph response under visible-light excitation
across the entire potential sweep (Figure 5.5a, blue curve and Figure 5.6). Chronoamperometry J(t)
experiments were then performed with the plasmonic photocathode poised at −0.4 VRHE. As shown in Figure
5.5b, the plasmonic Au/p-GaN device exhibits a prompt and reproducible cathodic Jph of ca. 1.3 μA cm-2
under periodic, visible-light illumination (λ > 495 nm). For comparison, no visible light response was observed
from the bare p-GaN support under otherwise identical experimental conditions (Figure 5.5b, blue curve). As
expected for a hot carrier-driven process,52 the plasmonic photocathode exhibits a linear Jph response with
respect to incident light power (Figure 5.7).

87

Figure 5.5: Photoelectrochemical characterization of plasmonic Au/p-GaN photocathodes. (a) Linear sweep voltammetry of
plasmonic Au/p-GaN (red) and bare p-GaN (blue) photocathodes at 10 mV s-1 under periodic (0.5 Hz), visible-light
irradiation (λ > 495 nm) at an incident power of I0 = 600 mW cm-2. (b) Chronoamperometry of plasmonic Au/p-GaN (red)
and bare p-GaN (blue) photocathodes under periodic (0.5 Hz), visible-light irradiation (λ > 495 nm) while poised at a fixed
applied potential of −0.4 VRHE. (c) Chronopotentiometry of the open-circuit voltage (Voc) from plasmonic Au/p-GaN
photocathodes under visible-light irradiation. (d) Incident photon-to-charge conversion efficiency (IPCE) of plasmonic Au/p-GaN
(red) and bare p-GaN (blue) photocathodes immersed in 50 mM K2CO3 electrolyte while held at a fixed potential of −0.4
VRHE. The absorption spectra of each device are also plotted with the IPCE spectra to aid comparison between photoelectrochemical
performance and light absorption.

88
Figure 5.6: Cyclic voltammetry of photocathodes. (a) Cyclic voltammograms of plasmonic Au/p-GaN and bare p-GaN
photocathodes under dark conditions (black and grey curves) and visible-light (λ > 495 nm) irradiation at I0 = 500 mW cm-2 (red
and blue curves). While the plasmonic Au/p-GaN device exhibits an obvious light response (red curve), no difference in current was
observed for the bare p-GaN device (blue curve). (b) Close-up view of the cyclic voltammograms from bare p-GaN photocathode under
dark (grey) and visible-light (λ > 495 nm) irradiation (blue) at I0 = 500 mW cm-2. No difference could be observed between dark
(grey) and light (blue) conditions, as these two curves lay directly on top of one another, confirming that the p-GaN support does not
respond to visible light. This observation is consistent with the large band gap of p-GaN (see Figure 5.2). Therefore, all visible-light
responses observed from plasmonic Au/p-GaN photocathodes can be unambiguously assigned to hot-hole injection from Au to the
valence band of p-GaN upon plasmon excitation.

Figure 5.7: Photocurrent (Jph = Jlight – Jdark) response obtained from plasmonic Au/p-GaN photocathodes showing a linear
trend with respect to incident light power (I0).
Plasmon-driven charge separation across the Au/p-GaN interface was examined by monitoring the
open-circuit voltage (Voc) of the device under visible-light excitation (λ > 495 nm). For a p-type semiconductor
in contact with an aqueous electrolyte, Fermi level equilibration between the semiconductor (EF) and the
redox species in the electrolyte (EF,Redox = EF) establishes a depletion region within the semiconductor that
results in downward band-bending near the surface.47 Upon optical excitation, the electric field across the
space-charge layer draws photogenerated electrons to the semiconductor-liquid junction where they are
scavenged by redox species in the electrolyte (e.g. protons), leaving behind charge vacancies (i.e. holes) to
accumulate within the bulk of the semiconductor. This accumulation of holes causes a downward shift of
their quasi-Fermi level (EF,p) with respect to the dark equilibrium level (EF), driving the Voc of the p-type

89
photocathode to more positive values.47 Indeed, band gap excitation of bare p-GaN photocathodes with UV
light induces a positive shift in the Voc of the device (Figure 5.8). As there are no free electrons photogenerated
directly within the p-GaN conduction band upon illumination with sub-band gap visible light, any changes in
the Voc of the Au/p-GaN photocathode can be attributed to the injection of hot holes into the p-GaN valence
band. Thus, monitoring changes in the Voc of the Au/p-GaN photocathode upon plasmon excitation enables
the detection of hot hole injection across the metal-semiconductor heterojunction in the absence of an applied
electrical bias, Eappl. The Au/p-GaN photocathode was allowed to equilibrate with the electrolyte for several
hours in the dark until a steady Voc baseline was obtained (Voc,dark = 0.62 VRHE). Plasmon excitation of the Au
nanoparticles promptly drives the injection of hot holes into the p-GaN support (Figure 5.5c), as evidenced
by the positive shift of Voc upon the incidence of visible light (Voc,light = 0.64 VRHE). A plasmonic photovoltage
(Vph = Voc,light − Voc,dark) of ca. 20 mV was eventually established within the photocathode and sustained over
30 s before the light was turned off. The hot holes present in the p-GaN valence band must then recombine
across the Au/p-GaN interface to re-establish the equilibrium Voc obtained under dark conditions. The
observation of a plasmonic photovoltage demonstrates the ability to maintain charge separation across the
Schottky barrier at the Au/p-GaN interface under steady-state operation.

Figure 5.8: Chronopotentiometry of the open-circuit voltage (Voc) from bare p-GaN photocathodes under UV-light irradiation.
The positive shift in Voc upon UV light exposure confirms the p-type character of the GaN substrates used herein.
The photoelectrochemical action spectrum, Jph(λ), was assessed via incident photon-to-current
conversion efficiency (IPCE) measurements with the photocathode poised at −0.4 VRHE in 50 mM K2CO3
electrolyte. Figure 5.5d shows that the IPCE of the plasmonic photocathode (red curve) faithfully follows the
surface plasmon resonance of the Au nanoparticles (light red curve) across the visible regime. For comparison,
no Jph response is observed from the bare p-GaN device under otherwise identical experimental conditions

90
(Figure 5.5d, blue curve). Taken together, these photoelectrochemical data confirm that the visible-light
response observed from the Au/p-GaN device is derived from hot-hole injection upon plasmon excitation
of the Au nanoparticles. Given the considerable Schottky barrier height at the Au/p-GaN interface (ΦB = 1.1
eV), such a result is encouraging for the design of optoelectronic devices that operate via the collection of hot
holes from metal nanostructures.
5.4

Fabrication, characterization, and electrochemical studies of Au/p-NiO
The influence of interfacial barrier height on hot-hole collection from Au nanoparticles was then

investigated through the construction and evaluation of a plasmonic photocathode composed of Au
nanoparticles on p-type nickel oxide (p-NiO) films. Plasmonic Au/p-NiO photocathodes were constructed
via evaporation of Au thin-films onto p-type NiO (p-NiO) substrates. The p-NiO films were synthesized on
fluorine-doped tin oxide (FTO) glass substrates using electron-beam physical vapor deposition. Ni metal was
deposited at a rate of 0.25 Å s-1 under flowing O2 gas at 6 sccm. After deposition of a 20 nm-thick NiO film
on the FTO substrate, a 2 nm-thick film of Au was then deposited onto the p-NiO surface using electronbeam physical vapor deposition at a base pressure of ca. 1 x 10-7 torr and a deposition rate of 1.0 Å s-1. The
Au/p-NiO films were then annealed in ambient air at 300 °C for 1 h to ensure coalescence of the
discontinuous Au thin-film into Au nanoparticles and achieve good adhesion with the underlying p-NiO
surface.
As a wide band gap semiconductor that natively exhibits p-type conductivity and forms an Ohmic
contact with Au,44 plasmonic Au/p-NiO photocathodes offer a complementary photoelectrochemical system
for assessing the collection of hot holes from Au nanoparticles in the absence of an interfacial barrier (Figure
S5.9a-d). Au nanoparticles of ca. 10 ± 1 nm in diameter were uniformly decorated onto the p-NiO surface in
the same way as for the Au/p-GaN device (Figure 5.9e). The Au/p-NiO photocathodes exhibit a surface
plasmon resonance feature at ca. 560 nm with nearly identical overall magnitude as the Au/p-GaN system
(Figure 5.9f). Solid-state I-V measurements confirm Ohmic contact between Au nanoparticles and the p-NiO
film (Figure 5.10a). Chronoamperometry of Au/p-NiO photocathodes yielded photocurrents of ca. 2.7 μA
cm-2 at −0.4 VRHE under periodic, visible-light irradiation (λ > 495 nm), while no Jph response was obtained
from the bare p-NiO substrate (Figure 5.10b). The incident photon-to-charge conversion efficiency (IPCE)
Jph(λ) of Au/p-NiO qualitatively follows the surface plasmon resonance of the Au nanoparticles, indicating
that all photocurrent is attributable to hot-hole injection from Au nanoparticles to the p-NiO support (Figure
5.10c). The striking similarity in photoelectrochemical performance between Au/p-NiO (ca. 2.7 μA cm-2 at
−0.4 VRHE) and Au/p-GaN (ca. 1.3 μA cm-2 at −0.4 VRHE) indicates that the vast majority of hot holes

91
generated in the Au/p-GaN device are sufficiently hot to inject above the 1.1 eV interfacial Schottky barrier
at the Au/p-GaN heterojunction. This experimental result is consistent with prior theoretical predictions of
the hot-hole distribution generated on Au nanoparticles upon photoexcitation above the interband threshold
of the metal.8-11 Considering the vast difference in barrier height between these two material systems (~1 eV),
such a result is promising for the eventual exploitation of plasmon-derived hot holes for solar photochemistry.

Figure 5.9: Characterization of p-type NiO films and plasmonic Au/p-NiO films. (a) X-ray diffraction pattern from 20 nmthick NiO films on FTO glass showing the characteristic (200) peak of NiO. (b) X-ray photoelectron spectroscopy spectrum of
the Ni 2p region, showing the characteristic binding energies of NiO. (c) Tauc plot of the NiO film exhibiting a band gap of ca.
3.7 eV. (d) Mott-Schottky plot obtained from 20 nm-thick NiO films on FTO glass substrate, which shows a negative slope
indicative of p-type conductivity. From these data, the flat-band potential (Efb) is estimated to be ca. 0.75 VRHE (Volts vs. RHE)
with an acceptor concentration of ca. 1 x 1019 cm-3. All these data are consistent with previous literature reports of p-type NiO
thin films.2 (e) Scanning electron microscopy image of Au nanoparticles uniformly decorated on the p-NiO surface with
corresponding size-distribution histogram of the Au nanoparticles, with an average Au diameter of 10 ± 1 nm. (f) Absorbance
of the bare p-NiO photocathode (grey) and the plasmonic Au/p-NiO photocathode (black). A prominent surface plasmon
resonance feature due to the Au nanoparticles is observed around 560 nm. Inset shows a digital image of the FTO glass substrate,
p-NiO/FTO substrate, and Au/p-NiO/FTO substrate, from left to right. A faint purple color is observed from the Au/pNiO device.

92

Figure 5.10: Photoelectrochemistry of plasmonic Au/p-NiO photocathodes. (a) Solid-state current-voltage (I-Eappl) behavior
from Au/p-NiO films exhibiting Ohmic behavior, consistent with previous literature for Au/NiO contacts. (b)
Chronoamperometry from Au/p-NiO photocathodes (black) under visible-light excitation (λ > 495 nm) at 500 mW cm-2 while
poised at −0.4 VRHE. A prompt, reproducible plasmonic photocurrent is clearly observed, indicating hot-hole collection by the
p-NiO support upon plasmon excitation. For comparison, the bare p-NiO film (grey) exhibits no observable photocurrent. (c)
Photoelectrochemical action spectrum obtained from Au/p-NiO device (black points), showing a clear relationship with the surface
plasmon resonance of the Au nanoparticles (black curve).
5.5

Selectivity of Au/p-GaN in light and dark conditions
To that end, Au/p-GaN photocathodes were employed for plasmon-driven carbon dioxide (CO2)

reduction in CO2-saturated 50 mM K2CO3 electrolyte in a three-electrode configuration with Pt gauze as the
counter electrode and a saturated calomel electrode as the reference electrode. No sacrificial reagents were
used in the reaction. The photocathode was potentiostatically poised at −1.8 VRHE while the reactor headspace
gas was periodically sampled and analyzed via gas chromatography over a 48 h period. It has previously been
reported that p-GaN nanowires, which predominantly expose m-planes, produce carbon monoxide (CO),
hydrogen (H2), and methane (CH4) in the absence of a metal co-catalyst.53 Under our experimental conditions,
we find that the bare p-GaN substrate (c-plane) produces primarily H2 and CO, with trace amounts of CH4
under dark electrolysis (Figure 5.11). The bare p-GaN surface exhibits a substantial preference for proton
reduction over CO2 reduction, evolving H2 at a rate that is ca. 3.5 times greater than CO. In contrast, the
addition of Au nanoparticles substantially alters the selectivity of the device for CO2 reduction. Under dark
electrolysis, the Au/p-GaN device exhibits improved selectivity for CO2 reduction to CO with a significant
reduction in H2 evolution relative to the bare p-GaN surface (Figure 5.12a). From these data, the Au/p-GaN
device produces CO-to-H2 at a ratio of ca. 4:1 under dark conditions. Plasmon excitation of the Au
nanoparticles with visible light (λ > 495 nm) was found to increase the CO evolution rate by 20%, from ca. 4
nmol h-1 in the dark to ca. 5 nmol h-1 in the light, while exerting little influence on the rate of H2 evolution
(Figure 5.12b). Plasmonic Au/p-GaN photocathodes therefore exhibit improved selectivity for

93
photoelectrochemical CO2 reduction in aqueous electrolytes, producing CO-to-H2 at a ratio of 5:1 upon
plasmon excitation (Figure 5.12b). The photoelectrochemical stability of these plasmonic photocathodes is
evidenced by the continuous evolution of gaseous products over 96 h of electrochemical operation. Overall,
these results are consistent with recent experimental observations of improved selectivity for plasmon-driven
CO2 reduction with plasmonic-metal nanostructures25,30 and demonstrate the utility of plasmonic
photocathodes for artificial photosynthesis.

Figure 5.11: Time-course of gas evolution from bare p-GaN photocathode under dark electrolysis conditions at −1.8 VRHE in
CO2-saturated 50 mM K2CO3 electrolyte.

94
Figure 5.12: Photoelectrochemical CO2 reduction with plasmonic Au/p-GaN photocathodes. (a) Time-course of gas evolution
from plasmonic Au/p-GaN photocathode during controlled potential electrolysis under dark conditions. (b) Time-course of gas
evolution from plasmonic Au/p-GaN photocathode during controlled potential electrolysis under plasmon excitation (λ > 495
nm). All electrolysis experiments were performed at −1.8 VRHE in CO2-saturated 50 mM K2CO3 electrolyte without sacrificial
reagents.
5.6

Conclusion
In summary, we have demonstrated the collection of plasmon-induced hot holes from Au

nanoparticles via construction of an interfacial Schottky barrier with p-type GaN. Although the photocurrents
reported here are relatively modest (~μA cm-2), substantial enhancements in device performance may be
achievable by implementing this photosensitization scheme for p-GaN nanowire arrays to dramatically
increase the available surface area for light collection and catalysis.54-57 Further improvements could also be
realized through roughening the metal-semiconductor interface to relieve the momentum matching
restrictions for hot hole transmission across planar metal-semiconductor Schottky junctions investigated here.
More broadly, the realization of a plasmon-driven photocathode capable of collecting hot holes at least 1.1
eV below the Au Fermi level holds tantalizing prospects for plasmonic photochemistry, given the considerable
oxidizing power of such carriers. These results support previous observations of plasmon-driven water
oxidation with Au nanoparticles,58-60 and suggest room for further improvements if the hot holes could be
efficiently transferred to an appropriate catalyst. Further studies are needed to fully elucidate the factors
governing plasmonic hot hole collection in p-type semiconductor systems along with their associated carrier
dynamics. Though at an early stage, our demonstration of a plasmon-driven artificial photosynthetic system
for CO2 conversion offers promise for the eventual exploitation of both hot electrons and hot holes in fuelforming photochemical reactions.

BIBLIOGRAPHY CHAPTER 5
1. Linic, S.; Christopher, P.; Ingram, D. B. Plasmonic-metal nanostructures for efficient conversion of solar
to chemical energy. Nat. Mater. 2011, 10, 911-921.
2. Clavero, C. Plasmon-induced hot-electron generation at nanoparticle/metal-oxide interfaces for
photovoltaic and photocatalytic devices. Nat. Photon. 2014, 8, 95-103.
3. Brongersma, M. L.; Halas, N. J.; Nordlander, P. Plasmon-induced hot carrier science and technology. Nat.
Nanotechnol. 2015, 10, 25-34.
4. Christopher, P.; Moskovits, M. Hot Charge Carrier Transmission from Plasmonic Nanostructures. Annu.
Rev. Phys. Chem. 2017, 68, 379-398.

95
5. Linic, S.; Aslam, U.; Boerigter, C.; Morabito, M. Photochemical transformations on plasmonic metal
nanoparticles. Nat. Mater. 2015, 14, 567-576.
6. Hartland, G. V.; Besteiro, L. V.; Johns, P.; Govorov, A. O. What’s so Hot about Electrons in Metal
Nanoparticles? ACS Energy Lett. 2017, 2, 1641-1653.
7. Manjavacas, A.; Liu, J. G.; Kulkarni, V.; Nordlander, P. Plasmon-induced hot carriers in metallic
nanoparticles. ACS Nano 2014, 8, 7630-7638.
8. Brown, A. M.; Sundararaman, R.; Narang, P.; Goddard III, W. A.; Atwater, H. A. Nonradiative Plasmon
Decay and Hot Carrier Dynamics: Effects of Phonons, Surfaces, and Geometry. ACS Nano 2016, 10, 957966.
9. Sundararaman, R.; Narang, P.; Jermyn, A. S.; Goddard III, W. A.; Atwater, H. A. Theoretical predictions
for hot-carrier generation from surface plasmon decay. Nat. Commun. 2014, 5, 5788.
10. Bernardi, M.; Mustafa, J.; Neaton, J. B.; Louie, S. G. Theory and computation of hot carriers generated by
surface plasmon polaritons in noble metals. Nat. Commun. 2015, 6, 7044.
11. Govorov, A. O.; Zhang, H.; Gun’ko, Y. K. Theory of Photoinjection of Hot Plasmonic Carriers from
Metal Nanostructures into Semiconductors and Surface Molecules. J. Phys. Chem. C 2013, 117, 1661616631.
12. Liu, J G.; Zhang, H.; Link, S.; Nordlander, P. Relaxation of Plasmon-Induced Hot Carriers. ACS Photon.
2017, DOI: 10.1021/acsphotonics.7b00881.
13. Harutyunyan, H.; Martinson, B. F. A.; Rosenmann, D.; Khorashad, L. K.; Besteiro, L. V.; Govorov, A.
O.; Wiederrecht, G. P. Anomalous ultrafast dynamics of hot plasmonic electrons in nanostructures with
hot spots. Nat. Nanotechnol. 2015, 10, 770-774.
14. Hartland, G. V. Optical Studies of Dynamics in Noble Metal Nanostructures. Chem. Rev. 2011, 111, 38583887.
15. Brown, A. M.; Sundararaman, R.; Narang, P.; Schwartzberg, A. M.; Goddard III, W. A.; Atwater, H. A.
Experimental and Ab Initio Ultrafast Carrier Dynamics in Plasmonic Nanoparticles. Phys. Rev. B 2017, 118,
087401.
16. Knight, M. W.; Sobhani, H.; Nordlander, P.; Halas, N. J. Photodetection with active optical antenna. Science
2011, 332, 702-704.
17. Zheng, B. Y.; Zhao, H.; Manjavacas, A.; McClain, M.; Nordlander, P.; Halas, N. J. Distinguishing between
plasmon-induced and photoexcited carriers in a device geometry. Nat. Commun. 2015, 6, 7797.
18. Li, W.; Valentine, J. Metamaterial Perfect Absorber Based Hot Electron Photodetection. Nano Lett. 2014,
14, 3510-3514.
19. Goykhman, I.; Desiatov, B.; Khurgin, J.; Shappir, J.; Levy, U. Locally Oxidized Silicon Surface-Plasmon
Schottky Detector for Telecom Regime. Nano Lett. 2011 11, 2219-2224.
20. Chalabi, H.; Schoen, D.; Brongersma, M. L. Hot-Electron Photodetection with a Plasmonic Nanostripe
Antenn. Nano Lett. 2014, 14, 1374-1380.
21. Li, W.; Coppens, Z. J.; Besteiro, L. V.; Wang, W.; Govorov, A. O.; Valentine, J. Circularly polarized light
detection with hot electrons in chiral plasmonic metamaterials. Nat. Commun. 2015, 6, 8379.
22. Mubeen, S.; Hernandez-Sosa, G.; Moses, D.; Lee, J.; Moskovits, M. Plasmonic photosensitization of a
wide band gap semiconductor: converting plasmons to charge carriers. Nano Lett. 2011, 11, 5548-5552.
23. Mubeen, S.; Lee, J.; Singh, N.; Kramer, S.; Stucky, G. D.; Moskovits, M. An autonomous photosynthetic
device in which all charge carriers derive from surface plasmons. Nat. Nanotechnol. 2013, 8, 247-251.
24. Mubeen, S.; Lee, J.; Liu, D.; Stucky, G. D.; Moskovits, M. Panchromatic photoproduction of H2 with
surface plasmons. Nano Lett. 2015, 15, 2132-2136.
25. Robatjazi, H.; Zhao, H.; Swearer, D. F.; Hogan, N. J.; Zhou, L.; Alabastri, A.; McClain, M. J.; Nordlander,
P.; Halas, N. J. Plasmon-induced selective carbon dioxide conversion on earth-abundant aluminumcuprous oxide antenna-reactor nanoparticle. Nat. Commun. 2017, 8, 27.

96
26. Swearer, D. F.; Zhao, H.; Zhou, L.; Zhang, C.; Robatjazi, H.; Martirez, J. M. P.; Krauter, C. M.; Yazdi, S.;
McClain, M. J.; Ringe, E.; Carter, E. A.; Nordlander, P.; Halas, N. J. Plasmonic Photocatalysis of Nitrous
Oxide into N2 and O2 using Aluminum-Iridium Antenna-Reactor Nanoparticles. Proc. Natl. Acad. Sci. U.
S. A. 2016, 113, 8916-8920.
27. Mukherjee, S.; Zhou, L.; Goodman, A. M.; Large, N.; Ayala-Orozco, C.; Zhang, Y.; Nordlander, P.; Halas,
N. J. Hot-Electron-Induced Dissociation of H2 on Gold Nanoparticles Supported on SiO2. J. Am. Chem.
Soc. 2014, 136, 64-67.
28. Zhou, L.; Zhang, C.; McClain, M. J.; Manjavacas, A.; Krauter, C. M.; Shu, T.; Berg, F.; Everitt, H. O.;
Carter, E. A.; Nordlander, P.; Halas, N. J. Aluminum Nanocrystals as a Plasmonic Photocatalyst for
Hydrogen Dissociation. Nano Lett. 2016, 16, 1478-1484.
29. Zhang, C.; Zho, H.; Zhou, L.; Schlather, A. E.; Dong, L.; McClain, M. J.; Swearer, D. F.; Nordlander, P.;
Halas, N. J. Al–Pd Nanodisk Heterodimers as Antenna–Reactor Photocatalysts. Nano Lett. 2016, 16, 66776682.
30. Zhang, X.; Li, X.; Zhang, D.; Su, N. Q.; Yang, W.; Everitt, H. O.; Liu, J. Product selectivity in plasmonic
photocatalysis for carbon dioxide hydrogenation. Nat. Commun. 2017, 8, 14542.
31. Aslam, U.; Chavez, S.; Linic, S. Controlling energy flow in multimetallic nanostructures for plasmonic
catalysis. Nat. Nanotechnol. 2017, 12, 1000-1005.
32. Zhong, Y.; Ueno, K.; Mori, Y.; Shi, X.; Oshikiri, T.; Murakoshi, K.; Inoue, H.; Misawa, H. Plasmonassisted Water Splitting Using Two Sides of the Same SrTiO3 Single-crystal Substrate: Conversion of
Visible Light to Chemical Energy. Angew. Chem. Int. Ed. 2014, 126, 10350-10354.
33. Sá, J.; Tagliabue, G.; Friedli, P.; Szlachetko, J.; Rittmann-Frank, M. H.; Santomauro, F. G.; Milne, C. J.;
Sigg, H. Direct observation of charge separation on Au localized surface plasmons. Energy Environ. Sci.
2013, 6, 3584-3588.
34. Thrall, E. S.; Steinberg, A. P.; Wu, X.; Brus, L. E. J. Phys. Chem. C 2013, 117, 26238-26247.
35. Langille, M. R.; Personick, M. L.; Mirkin, C. A. Plasmon-mediated syntheses of metallic nanostructures.
Angew. Chem. Int. Ed. 2013, 52, 13910-13940.
36. Wu, X.; Redmond, P. L.; Liu, H.; Chen, Y.; Steigerwald, M.; Brus, L. Photovoltage mechanism for room
light conversion of citrate stabilized silver nanocrystal seeds to large nanoprisms. J. Am. Chem. Soc. 2008,
130, 9500-9506.
37. Schlather, A. E.; Manjavacas, A.; Lauchner, A.; Marangoni, V. S.; DeSantis, C. J.; Nordlander, P.; Halas,
N. J. Hot Hole Photoelectrochemistry on Au@SiO2@Au Nanoparticles. J. Phys. Chem. Lett. 2017, 8, 20602067.
38. Wu, X.; Thrall, E. S.; Liu, H.; Steigerwald, M.; Brus, L. Plasmon Induced Photovoltage and Charge
Separation in Citrate-Stabilized Gold Nanoparticles. J. Phys. Chem. C 2010, 114, 12896-12899.
39. Zhai, Y.; DuChene, J. S.; Wang, Y.-C.; Qiu, J.; Johnston-Peck, A. C.; You, B.; Guo, W.; DiCaccio, B.;
Qian, K.; Zhao, E. W.; Ooi, F.; Hu, D.; Su, D.; Stach, E. A.; Zhu, Z.; Wei, D. W. Polyvinylpyrrolidoneinduced anisotropic growth of gold nanoprisms in plasmon-driven synthesis. Nat. Mater. 2016, 15, 889895.
40. Zhao, J.; Nguyen, S. C.; Ye, R.; Ye, B.; Weller, H.; Somorjai, G. A.; Alivisatos, A. P.; Toste, F. D. A
Comparison of Photocatalytic Activities of Gold Nanoparticles Following Plasmonic and Interband
Excitation and a Strategy for Harnessing Interband Hot Carriers for Solution Phase Photocatalysis. ACS
Cent. Sci. 2017, 3, 482-488.
41. Kim, Y.; Torres, D. D.; Jain, P. K. Activation Energies of Plasmonic Catalysts. Nano Lett. 2016, 16, 33993407.
42. Moskovits, M. The case for plasmon-derived hot carrier devices. Nat. Nanotechnol. 2015, 10, 6-8.

97
43. Nakamura, K.; Oshikiri, T.; Ueno, K.; Wang, Y.; Kamata, Y.; Kotake, Y.; Misawa, H. Properties of
plasmon-induced photoelectric conversion on a TiO2/NiO p-n junction with au nanoparticles. J. Phys.
Chem. Lett. 2016, 7, 1004-1009.
44. Robatjazi, H.; Bahauddin, S. M.; Doiron, C.; Thomann, I. Direct Plasmon-Driven Photoelectrocatalysis.
Nano Lett. 2015, 15, 6155-6161.
45. Beach, J. D.; Collins, R. T.; Turner, J. A. J. Electrochem. Soc. 2003, 7, A899-A904.
46. Kibria, M. G.; Mi, Z. Artificial photosynthesis using metal/nonmetal-nitride semiconductors: current
status, prospects, and challenges. J. Mater. Chem. A 2016, 4, 2801-2820.
47. Kamimura, J.; Bogdanoff, P.; Ramsteiner, M.; Corfdir, P.; Feix, F.; Geelhaar, L.; Riechert, H.
Photoelectrochemical Properties of (In,Ga)N Nanowires for Water Splitting Investigated by in Situ
Electrochemical Mass Spectroscopy. Nano Lett. 2017, 17, 1529-1537.
48. Kibria, M. G.; Chowdhury, F. A.; Zhao, S.; Al Oltaibi, B.; Trudeau, M. L.; Guo, H.; Mi, Z. Visible lightdriven efficient overall water splitting using p-type metal-nitride nanowire arrays. Nat. Commun. 2015, 6,
6797.
49. Ho, J.-K.; Jong, C.-S.; Chiu, C. C.; Huang, C.-N.; Shih, K.-K. Low-resistance ohmic contacts to p-type
GaN achieved by the oxidation of Ni/Au films. J. Appl. Phys. 1999, 86, 4491-4497.
50. Jang, H. W.; Kim, S. Y.; Lee, J.-L. Mechanism for Ohmic contact formation of oxidized NiÕAu on ptype GaN. J. Appl. Phys. 2003, 94, 1748-1752.
51. Wu, C. I.; Kahn, A. Investigation of the chemistry and electronic properties of metal/gallium nitride
interfaces. J. Vac. Sci. Technol. B 1998, 16, 2218-2223.
52. Christopher, P.; Xin, H.; Linic, S. Visible-light-enhanced catalytic oxidation reactions on plasmonic silver
nanostructures. Nat. Chem. 2011, 3, 467-472.
53. Al Otaibi, B.; Fan, S.; Wang, D.; Ye, J.; Mi, Z. Wafer-Level Artificial Photosynthesis for CO2 Reduction
into CH4 and CO Using GaN Nanowires. ACS Catal. 2015, 5, 5342-5348.
54. Boettcher, S. W.; Spurgeon, J. M.; Putnam, M. C.; Warren, E. L.; Turner-Evans, D. B.; Kelzenberg, M. D.;
Maiolo, J. R.; Atwater, H. A.; Lewis, N. S. Photoelectrochemical Hydrogen Evolution Using Si Microwire
Arrays. Science 2010, 327, 185-187.
55. Kelzenberg, M. D.; Boettcher, S. W.; Petykiewicz, J. A.; Turner-Evans, D. B.; Putnam, M. C.; Warren, E.
L.; Spurgeon, J. M.; Briggs, R. M.; Lewis, N. S.; Atwater, H. A. Enhanced absorption and carrier collection
in Si wire arrays for photovoltaic applications. Nat. Mater. 2010, 9, 239-244.
56. Wang, D.; Pierre, A.; Kibria, M. G.; Cui, K.; Han, X.; Bevan, K. H.; Guo, H.; Paradis, S.; Hakima, A.-R.;
Mi, Z. Wafer-level photocatalytic water splitting on GaN nanowire arrays grown by molecular beam
epitaxy. Nano Lett. 2011, 11, 2353-2357.
57. Al Otaibi, B.; Nguyen, H. P. T.; Zhao, S.; Kibria, M. G.; Fan, S.; Mi, Z. Highly Stable Photoelectrochemical
Water Splitting and Hydrogen Generation Using a Double-Band InGaN/GaN Core/Shell Nanowire
Photoanode. Nano Lett. 2013, 13, 4356-4361.
58. Shi, X.; Ueno, K.; Takabayashi, N.; Misawa, H. Plasmon-enhanced photocurrent generation and water
oxidation with a gold nanoisland-loaded titanium dioxide photoelectrode. J. Phys. Chem. C 2012, 117, 24942499.
59. Nishijima, Y.; Ueno, K.; Kotake, Y.; Murakoshi, K.; Inoue, H.; Misawa, H. Near-infrared plasmon-assisted
water oxidation. J. Phys. Chem. Lett. 2012, 3, 1248-1252.
60. Wang, S.; Gao, Y.; Miao, S.; Liu, T.; Mu, L.; Li, R.; Fan, F.; Li, C. Positioning the Water Oxidation Reaction
Sites in Plasmonic Photocatalysts. J. Am. Chem. Soc. 20160 139, 11771-11778.

98
Chapter 6

OPTICAL EXCITATION OF A NANOPARTICLE Cu/p-NiO PHOTOCATHODE
IMPROVES REACTION SELECTIVITY FOR CO2 REDUCTION IN AQUEOUS
ELECTROLYTES
6.1

Introduction
Artificial photosynthesis seeks to mimic the catalytic machinery of natural photosynthetic systems with

inorganic materials capable of converting carbon dioxide (CO2), water (H2O), and sunlight into useful chemicals
(e.g. ethanol, ethylene, etc.).1-6 Unfortunately, the realization of such a process is currently hindered by catalytic
challenges associated with selective conversion of CO2 into desired products without the proliferation of
unwanted side reactions.1-6 The complexity of the reaction pathway, which involve multiple proton-coupled
electron transfer steps, requires a process for preferentially activating specific chemical intermediates to reliably
and selectively produce a single product of interest.1-6 The ongoing search for selectivity has inspired numerous
strategies to improve the preferential conversion of CO2 into desired products, including nanostructuring of the
electrocatalyst,7-9 elemental alloying,10,11 engineering of the exposed catalytic surface facets12-15 and grain
boundaries,16-18 manipulating the local solution pH,19-21 judicious choice of chemical additives to the electrolyte
itself,22,23 or the use of ionic liquids to limit the availability of protons.24,25
Despite numerous examples of improved catalyst selectivity via the aforementioned approaches, to date,
the use of light as a tool for guiding the selectivity of CO2 reduction has received considerably less attention.26-33
Given that the most commonly used metals for electrocatalytic CO2 reduction, namely Ag, Au, and Cu, all
support surface plasmon excitations, nanostructured metal catalysts offer new opportunities for exploiting their
unique optical properties to shape the selectivity of chemical reactions.26,34-39 In particular, the plasmon-driven
production of energetic “hot” carriers on metal nanostructures has shown great promise for photocatalysis,34-39
but the prompt decay (t ~ 1 ps) of hot carriers into phonon modes of the metal nanocrystal requires a strategy
for quickly separating hot electron-hole pairs on an ultrafast timescale.38,39 To that end, numerous studies have
established the benefits of forming an interfacial Schottky barrier between a plasmonic metal and a wide band
gap n-type semiconductor (e.g. Au/TiO2) for separating hot carriers across the metal-semiconductor
heterojunction.40-49 Providing a channel for collecting hot electrons within the conduction band of the n-type
semiconductor support effectively limits recombination processes and extends the lifetime of the chargeseparated state to allow photochemistry to proceed.47-49 Yet to promote plasmon-driven CO2 reduction directly
on the metal surface requires quickly extracting hot holes from below the metal Fermi level with a wide band

99
gap p-type semiconductor so that hot electrons can accumulate on the metal and initiate reactions with adsorbed
molecules. The ability to quickly collect hot holes from metal nanostructures via charge transfer to the support
also obviates the need for sacrificial reagents commonly used in plasmonic photocatalysis. Indeed, we have
recently demonstrated the utility of interfacing plasmonic Au nanoparticles with p-type GaN to enable
photoelectrochemical CO2 reduction with plasmonic Au/p-GaN photocathodes.28 Unfortunately, the limited
number of p-type semiconductors suitable for such studies has hindered the development of plasmonic devices
capable of harvesting hot holes from metal nanostructures for applications in photocatalysis or photodetection.
Our ability to manipulate and control hot carriers is currently restricted by insufficient knowledge of plasmoninduced hot holes; to date, relatively few experimental studies have been reported.28,50-58
Here, we employ p-type nickel oxide (p-NiO) as a wide band gap semiconductor support to harvest hot
holes from photoexcited Cu nanoparticles and enable photoelectrochemical CO2 reduction with plasmonic
Cu/p-NiO photocathodes (Figure 6.1a). Nickel oxide is commonly used as a hole transport material in a variety
of photovoltaic and photoelectrochemical devices due to its excellent chemical stability, high optical
transparency, and suitable p-type character.59-64 Furthermore, because p-NiO films can be deposited by a variety
of low-cost methods, p-NiO may offer a more scalable option than p-type GaN as a candidate wide band gap
p-type semiconductor to facilitate charge separation. In plasmonic devices, p-NiO has previously been used to
collect photogenerated hot holes from Au nanoparticles, where an Ohmic contact is reportedly formed at the
Au/p-NiO interface.65-67 As a support for Cu nanoparticles, the valence band position of p-NiO (ca. −5.4 eV vs.
vacuum)60,63 relative to the Cu Fermi level (ca. −4.5 eV vs. vacuum)68 is anticipated to establish a modest Schottky
barrier (ΦB) at the Cu/p-NiO interface that facilitates charge separation by selectively collecting plasmoninduced hot holes from the metal (Figure 6.1b). The large band gap of p-NiO (~3.7 eV)63,64 ensures that any
visible light incident upon the Cu/p-NiO device is incapable of directly exciting charge carriers within the pNiO film, and it therefore serves solely to collect hot holes from the Cu nanoparticles. Furthermore, the
conduction band edge of p-NiO (ca. −1.7 eV vs. vacuum)60 relative to the Cu Fermi level provides rectification
across the Cu/p-NiO interface by presenting a sizable energy barrier (~ 3 eV) to hot electron transfer at the
metal-semiconductor heterojunction (Figure 6.1b). This plasmonic Cu/p-NiO device structure thereby limits
recombination processes by providing a pathway for hot hole collection within the underlying p-NiO film while
allowing for the accumulation of hot electrons on the Cu nanoparticles to drive CO2 reduction.

100

Figure 6.1: Plasmonic Cu/p-NiO photocathode device structure. (a) Schematic of Cu/p-NiO photocathode on fluorine-doped
tin oxide (FTO) glass showing the approximate dimensions of the Cu nanoparticles (~8 nm in diameter) and the p-NiO layer
(~60 nm thick) on the FTO glass substrate. (b) Quantitative energy level diagram showing the relative positions of the p-NiO
valence band (EVB) and conduction band (ECB) relative to the Cu Fermi level (EF). The difference in energy between the p-NiO
valence band and the Cu Fermi level is expected to allow the formation of an interfacial Schottky barrier (ΦB) to hot hole
injection at the Cu/p-NiO interface of around 1 eV. Photoexcitation of Cu nanoparticles with photon energy (hv) below the
band gap (EG) of the p-NiO support generates hot electrons and hot holes on the Cu surface. The p-NiO support facilitates
charge separation across the metal-semiconductor interface by allowing the collection of hot holes from the metal while also
confining the hot electrons on the Cu surface to drive CO2 reduction (inset).
Photoelectrochemical studies of plasmonic Cu/p-NiO photocathodes confirm that visible-light
excitation of Cu nanoparticles induces hot hole injection to the p-NiO valence band along with hot electron
transfer to adsorbed molecules in the supporting electrolyte. The incidence of visible light was found to exert
a significant influence over the selectivity of Cu nanoparticles for CO2 reduction. Specifically, we observed
that optical excitation of the Cu nanoparticles preferentially promoted the production of both carbon
monoxide (CO) and formate (HCOO−) while simultaneously limiting the evolution of hydrogen (H2) in
aqueous electrolytes. These results suggest that optical excitation of the metal alters the electrochemical

101
reaction mechanism occurring on the Cu surface, with implications for the design of plasmonic photocatalysts
that exhibit improved selectivity for CO2 reduction. Overall, our studies demonstrate the utility of p-type
semiconductors for the development of plasmonic photocathodes capable of artificial photosynthesis and
open new possibilities for manipulating and controlling photochemistry at the nanoscale with plasmonic-metal
nanostructures.
6.2

Fabrication and characterization of electrode

Figure 6.2: Materials characterization of p-type NiO films. (a) X-ray diffraction pattern from NiO film
on FTO glass showing the characteristic (200) peak of NiO. All other peaks can be attributed to the
underlying FTO substrate. (b) X-ray photoelectron spectroscopy high-resolution scan of the Ni 2p region,
showing the characteristic binding energies and satellite features of NiO. (c) Mott-Schottky plot obtained
from NiO films on FTO glass substrate, which shows a negative slope indicative of p-type conductivity.
From a linear fit of these data, the flat-band potential (Efb) is estimated to be ca. 0.75 VRHE (Volts vs. RHE)
with an acceptor concentration of ca. 1 x 1019 cm-3. (d) Tauc plot of the NiO film showing a band gap of
around 3.7 eV. All these data indicate material properties consistent with previous literature reports of ptype NiO thin films.
Plasmonic Cu/p-NiO photocathodes were constructed via electron beam physical vapor deposition.
The p-type NiO (p-NiO) films were first synthesized on fluorine-doped tin oxide (FTO) glass substrates by

102
depositing Ni metal at a rate of 0.25 Å s-1 under flowing O2 gas at 6 sccm. After deposition of a 60 nm-thick
NiO film on the FTO substrate, the film was annealed in ambient air at 300 °C for 1 h to ensure complete
conversion to the desired p-NiO phase, (see Figure 6.2). Mott-Schottky analysis of the p-NiO film confirms
that they exhibit p-type conductivity with a flat-band potential (Efb) of around 0.75 VRHE (V vs. RHE) and a
carrier density of around 2 x 1019 cm-3 (Figure 6.2c). We note that these material properties of the assynthesized p-NiO films are consistent with previous reports.61,63 After the heat treatment, 3 nm of Cu was
then deposited onto the p-NiO surface using electron-beam physical vapor deposition at a base pressure of
ca. 1 x 10-7 torr and a deposition rate of 1.0 Å s-1. No interfacial adhesion layer was used at the Cu/p-NiO
heterojunction. Free from stabilizing surfactants required in colloidal nanoparticle synthesis, our approach
ensures direct physical contact at the Cu/p-NiO interface while also exposing a clean Cu surface for catalysis.
Scanning electron microscopy (SEM) imaging of the Cu/p-NiO device shows Cu nanoparticles distributed
across the p-type NiO surface with a mean diameter d of 8 ± 2 nm (Figure 6.3a). Analysis of the Cu oxidation
state by X-ray photoelectron spectroscopy (XPS) indicates that the as-deposited Cu nanoparticles likely
oxidize to a mix of both Cu(I) and Cu(II) oxidation states69 upon exposure to ambient air (Figure 6.3b). This
interpretation is further supported by the optical properties of the Cu/p-NiO films, which appear dark grey
in color and display a broad peak in the visible region spanning from around 600 nm to 800 nm (Figure 6.3c,
yellow curve). This optical response is similar to that previously observed in CuO nanoparticles.70,71 In
contrast, the bare p-NiO films are nearly transparent across the visible spectrum (Figure 6.3c, blue curve) and
exhibit a wide band gap EG of around 3.7 eV (Figure 6.2d).

103

Figure 6.3: Materials characterization of the plasmonic Cu/p-NiO photocathode. (a) SEM image with corresponding sizedistribution histogram of Cu nanoparticles (mean diameter, d = 8 ± 2 nm) on a 60 nm thick p-NiO film supported on FTO
glass. (b) X-ray photoelectron spectroscopy high-resolution scan of the Cu 2p region from as-synthesized Cu/p-NiO photocathodes.
(c) Absorption spectra of the plasmonic Cu/p-NiO photocathode before (yellow curve) and after (red curve) electrochemical
reduction via three successive cyclic voltammetry scans. The spectrum of the bare p-NiO film (blue curve) is also shown for
comparison. (d) Cyclic voltammograms from plasmonic Cu/p-NiO photocathode (yellow to red curves) and bare p-NiO films
(blue curve) at a scan rate of 50 mV s-1. Black arrows indicate the scan direction. The reduction of Cu oxides into metallic Cu
is evidenced by the progressively smaller cathodic wave around 0.7 VRHE (yellow curve) that eventually disappears after the third
successive scan (red curve). A representative voltammogram from bare p-NiO films (blue curve) is shown for reference.
6.3

Photoelectrochemical studies
Photoelectrochemical studies were performed in a three-electrode configuration with the Cu/p-NiO

photocathode as the working electrode, a platinum wire mesh counter electrode, and a saturated calomel
electrode (SCE) as the reference electrode. The aqueous electrolyte (50 mM K2CO3) was sparged with CO2
gas for 30 min prior to all electrochemical experiments, which were performed under a CO2 blanket to prevent
the ingress of atmospheric O2 into the supporting electrolyte. All electrochemical potentials are reported with
respect to the reversible hydrogen electrode (RHE). As shown in Figure 6.3d (yellow curve), cyclic
voltammetry of the Cu/p-NiO device indicates that surface oxides formed on the Cu nanoparticles upon

104
exposure to ambient air are successfully reduced into metallic Cu(0) at applied potentials more negative than
0.6 VRHE. The absence of such features from bare p-NiO films (Figure 6.3d, blue curve) confirms that these
cathodic and anodic waves are attributable to the redox features of the Cu nanoparticles. Subsequent cyclic
voltammetry scans across the potential window from 0.8 VRHE to 0.2 VRHE indicate that any residual
cupric/cuprous oxides are fully converted into metallic Cu, as evidenced by the progressively smaller reduction
wave around 0.6 VRHE that eventually disappears upon the third successive scan (Figure 6.2d, red curve). This
electrochemical observation is consistent with recent in operando spectroscopic evidence of the
electrocatalytically active phase of Cu-based cathodes.72 We also note that a change in the optical absorption
was observed for these Cu/p-NiO films immediately after cyclic voltammetry was performed. The freshly
cycled Cu/p-NiO photocathodes exhibit a new spectral feature located around 630 nm that we attribute to
the surface plasmon resonance of metallic Cu nanoparticles (Figure 6.3c, red curve). Collectively, these results
strongly suggest that the oxide formed on the Cu nanoparticles upon exposure to ambient air is successfully
reduced back into the metallic state under CO2 reduction conditions.

Figure 6.4: Cyclic voltammetry of bare p-NiO cathodes under dark (black curve) and visible light (orange
curve) showing that bare p-NiO film exhibits no measurable light response across the potential sweep.
This observation is consistent with the large band gap of p-NiO. Therefore, all visible-light responses
observed from plasmonic Cu/p-NiO photocathodes can be unambiguously assigned to hot-hole injection
from Cu to the valence band of p-NiO upon optical excitation of the Cu nanoparticles.

105
The current-potential (J-E) behavior of the plasmonic Cu/p-NiO photocathode was assessed via
linear sweep voltammetry at a scan rate of 20 mV s-1 under both dark conditions and visible-light excitation
(λ = 560 ± 40 nm FWHM) with a high-power LED (I0 = 160 mW cm-2). As shown in Figure 6.5a (dotted
black curve), the Cu/p-NiO device displayed a cathodic current (J) along the potential (E) sweep from 0 VRHE
to −1.0 VRHE. The incidence of visible light (Figure 6.5a, solid red curve) imparts increased cathodic
photocurrent (Jph) relative to that observed in the dark and reduces the potential required for the onset of the
Faradaic current from around −0.4 VRHE in the dark to around −0.3 VRHE in the light (see inset of Figure 6.5a).
For comparison, the bare p-NiO photocathode exhibits no change in current density under visible-light
excitation (Figure 6.4). Chronoamperometry J(t) experiments demonstrate that the plasmonic Cu/p-NiO
device exhibits a prompt and reproducible cathodic photocurrent Jph under periodic, visible-light illumination
(λ = 560 ± 40 nm) while held potentiostatically at −0.2 VRHE (Figure 6.5b, red curve). For comparison, no
visible light response was observed from bare p-NiO supports under otherwise identical experimental
conditions (Figure 6.5b, blue curve). The plasmonic Cu/p-NiO photocathode displays a linear Jph response
with respect to the incident light power and reaches a maximum Jph of around 5 μA cm-2 (Figure 6.5c).
Chronopotentiometry Voc(t) experiments were then performed to examine plasmon-driven charge separation
across the metal-semiconductor heterojunction.28 Hot hole injection from photo-excited Cu nanoparticles
into the p-NiO film under open-circuit conditions leads to the accumulation of holes within the valence band
of p-NiO, causing a shift in the Voc of the Cu/p-NiO photocathode to more positive potentials relative to
the equilibrium Voc observed under dark conditions. Therefore, plasmon-induced hot hole transfer across the
metal-semiconductor heterojunction can be observed by monitoring the influence of light on the Voc of the
device. Indeed, the plasmonic Cu/p-NiO device exhibits an increase in Voc upon exposure to visible-light
irradiation and eventually establishes a plasmonic photovoltage Vph (Vph = Voc,light – Voc,dark) of around 15 mV
(Figure 6.5d, red curve). No Vph response was observed from bare p-NiO films (Figure 6.5d, blue curve).
Taken together, these data are consistent with plasmon-induced hot hole injection into the p-NiO valence
band along with photoelectrochemical reduction of molecular species in the supporting electrolyte.

106

Figure 6.5: Photoelectrochemical characterization of plasmonic Cu/p-NiO photocathodes. (a) Linear sweep voltammetry J(E)
of plasmonic Cu/p-NiO photocathode at a scan rate of 20 mV s-1 under dark conditions (dotted black curve) and under visiblelight irradiation (λ = 560 ± 40 nm) (solid red curve). (b) Chronoamperometry J(t) of the photocurrent (Jph = Jlight – Jdark)
obtained from plasmonic Cu/p-NiO (red) and bare p-NiO (blue) photocathodes under periodic, visible-light irradiation (λ =
560 ± 40 nm) while potentiostatically poised at an applied potential of E = −0.2 VRHE. (c) Power-dependence of the
photocurrent Jph(I0) obtained from the plasmonic Cu/p-NiO photocathode. (d) Chronopotentiometry V(t) of the open-circuit
voltage (Voc) obtained from the plasmonic Cu/p-NiO photocathode (red curve) and the bare p-NiO cathode (blue curve) under
visible-light irradiation (λ = 560 ± 40 nm).
The influence of plasmon excitation on the selectivity of plasmonic Cu/p-NiO photocathodes for the
CO2 reduction reaction (CO2RR) was then studied in a two-compartment compression cell specifically designed
to enable photoelectrochemical operation.73 The CO2 reduction reaction (CO2RR) was conducted in a threeelectrode configuration with Cu/p-NiO or bare p-NiO cathode as the working electrode, Pt wire gauze as the
counter electrode, and a saturated calomel electrode (SCE) as the reference electrode. All photoelectrochemical
experiments were conducted within a custom-built, airtight cell equipped with a quartz window.

The

photoelectrochemical experiments were performed in K2CO3 electrolyte (pH 7) that was fully saturated with
CO2 by vigorous bubbling of the cathode and anode compartments for 1 h before commencing with the

107
experiment. We note that photoelectrochemical collection of hot holes from the Cu nanoparticles via the
underlying p-NiO film obviates the need for sacrificial reagents commonly used in plasmonic photocatalysis.
The photocathode was potentiostatically poised at a given potential from −0.7 VRHE to −0.9 VRHE for 2 h to
collect multiple data points. The potential window available for plasmon-driven CO2 reduction studies is
restricted by the limited stability of the oxide support during electrochemical operation at applied potentials
more negative than −0.9 VRHE. Visible-light irradiation (λ = 560 ± 40 nm FWHM) with a high-power LED was
incident on the sample through the quartz window at an incident power of I0 = 160 mW cm-2 (measured at the
quartz window). The photo(electro)catalysis experiments were performed under continuous flow conditions in
a flow-cell device with periodic sampling of the reactor headspace every 15 min over the course of the 2 h
reaction with a gas chromatograph. The gas chromatograph (SRI-8610) was equipped with a Hayesep D column
and a Molsieve 5A column using N2 as carrier gas. The gaseous products were detected using a thermal
conductivity detector (TCD) and flame ionization detector (FID) equipped with a methanizer. Quantitative
analysis of gaseous products was based on calibration with several gas standards over many orders of magnitude
in concentration. Liquid products were collected from the cathode and anode compartments at the end of the
2 h electrolysis experiment and analyzed by high-pressure liquid chromatography (Thermo Fischer Dionex
UltiMate 3000). All catalytic experiments at a given potential were repeated in triplicate.

Figure 6.6: CO2 reduction with bare p-NiO cathodes under dark conditions. (a-b) Faradaic efficiency for H2 (squares), CO
(circles), and HCOO− (triangles) with (c-d) corresponding partial current density J for the hydrogen evolution reaction (JHER, squares),

108
carbon monoxide (JCO, circles), and formate (JHCOO-, triangles). Electrolysis was performed under dark conditions in a CO2-saturated
50 mM K2CO3 electrolyte. The device was held potentiostatically at each applied potential for 2 h while the gas products were
sampled every 15 min and analyzed by gas chromatography. Liquid products were collected and analyzed by HPLC at the end of
each run. Each data point represents the average of three independent trials and the error bar indicates the standard deviation.

Figure 6.7: Comparison of the partial current densities (J) obtained from bare p-NiO (open data points) and Cu/p-NiO cathodes
(filled data points) under dark conditions. (a) Partial current density for the hydrogen evolution reaction (JHER) from p-NiO (open
squares) relative to Cu/p-NiO (filled squares). (b) Partial current density for carbon monoxide (JCO) from p-NiO (open circles)
relative to Cu/p-NiO (filled circles). (c) Partial current density for formate (JHCOO-) from p-NiO (open triangles) relative to Cu/pNiO (filled triangles). Each data point represents the average of three independent trials and the error bar indicates the standard
deviation. We observed a significant increase in the partial current densities for CO2 reduction products CO and HCOO− with the
addition of Cu nanoparticles, while almost no change in the amount of H2 that was evolved. We therefore attribute the significant
amount of H2 that is evolved from the plasmonic Cu/p-NiO device to the activity of the underlying p-NiO film, which almost
exclusively produces H2 under CO2 reduction conditions.
As shown in Figure 6.8, the observed product distributions obtained under dark electrocatalysis are
dependent on the applied electrochemical potential (E). The reported Faradaic efficiency for each chemical
product represents the average value obtained from three independent trials and the error bars indicate the
standard deviation. At an applied potential of E = −0.7 VRHE, the Cu/p-NiO photocathode evolves primarily
hydrogen (H2) along with carbon monoxide (CO) and formate (HCOO−) as minor products. At more negative
applied potentials (E = −0.9 VRHE), CO and HCOO− begin to comprise a more significant fraction of the total
Faradaic efficiency (~40%) under dark conditions (Figure 6.8b-c). We emphasize that our cyclic voltammetry
results (Figure 6.5d), together with the observed changes in the optical properties of the device (Figure 6.5c),
indicate that the oxidation state of the Cu nanoparticles under these applied potentials is metallic Cu(0). This

109
conclusion is also supported by recent in operando spectroscopic evidence.72 Although a significant fraction of the
products evolved from the Cu/p-NiO device consist of H2 under dark conditions (Figure 6.8a, blue points), this
product is largely attributable to the activity of the underlying p-NiO film that remains exposed to the electrolyte.
Indeed, the bare p-NiO substrate produces almost exclusively H2 with ~98% Faradaic efficiency across the
entire potential window from −0.7 VRHE to −0.9 VRHE (Figure 6.6). The product distribution observed for the
bare p-NiO control sample under CO2RR conditions is consistent with a prior study of NiO-based cathodes.62
We therefore assign nearly all CO2RR products observed from the plasmonic Cu/p-NiO photocathodes to the
catalytic activity of the Cu nanoparticles (Figure 6.7).

Figure 6.8: Distribution of CO2 reduction products obtained from plasmonic Cu/p-NiO photocathodes as a function of the
applied electrochemical potential (E). Faradaic efficiency (a–c) and associated partial current density (d–f) for the production of
(a,d) hydrogen (H2) (squares), (b,e) carbon monoxide (CO) (circles), and (c,f) formate (HCOO−) (triangles) during controlled
potential electrolysis under dark conditions (blue symbols) and under visible-light irradiation (yellow symbols). Plasmon excitation
was performed with λ = 560 ± 40 nm at an incident power of 160 mW cm-2. Data points and error bars represent the average
value and standard deviation, respectively, obtained from three independent trials.
As shown in Figure 6.8, optical excitation of the plasmonic Cu/p-NiO photocathodes with 565 ± 40
nm light from a high-power LED (I0 = 160 mW cm-2) induces a marked change in the distribution of chemical
products compared to that observed during dark electrocatalysis. Specifically, we observed a reduction in the
Faradaic efficiency for H2 evolution (Figure 6.8a, squares) concomitantly with an increase in the Faradaic

110
efficiency for both carbon monoxide (Figure 6.8b, circles) and formate (Figure 6.8c, triangles) at all applied
potentials. The biggest change in selectively was observed at −0.7 VRHE, where the Faradaic efficiency for H2
falls from nearly 94% in the dark to around 58% in the light (Figure 6.8a). This substantial reduction in HER
activity was accompanied by a sizable improvement in the selectivity for CO2 reduction: the Faradaic efficiency
for both CO and HCOO− increased by three times relative to that observed in the dark and begin to account
for nearly 50% of the total Faradaic efficiency from the device. The partial current densities associated with
the H2 evolution reaction (JHER), the production of CO (JCO), and the production of formate (JHCOO−) are shown
in Figure 6.8d-f, respectively. At an applied potential of E = −0.7 VRHE, we observed little change in JHER
between dark and light conditions (Figure 6.8d), but a notable increase in both JCO and JHCOO− was observed
(Figure 6.8e-f). As we moved to more negative applied potentials, the proportion of JCO and JHCOO− continued
to increase along with a sizable reduction in JHER relative to that observed during dark electrocatalysis. At the
most negative potential studied (E = −0.9 VRHE), the JHER was reduced by nearly 33% from around 1.5 mA
cm-2 in the dark to around 1 mA cm-2 in the light, while both JCO and JHCOO− are nearly three times greater than
they were in the dark. Overall, these results indicate that optical excitation of the Cu nanoparticles increases
their selectivity for the CO2RR relative to the HER at all applied potentials. We exclude the possibility that
such changes in CO2RR selectivity arise solely due to plasmonic heating of the electrocatalytic surface, as it
has previously been shown that increased electrolyte temperatures promote H2 evolution while reducing the
selectivity for CO2 reduction.73 Such heating-induced trends in electrocatalytic selectivity are clearly opposite
to those observed here. We note that these results are interesting in light of previous observations of plasmonenhanced selectivity involving gas-phase photocatalysis conducted at elevated temperatures, in which the
conversion of CO2 and H2 to carbon monoxide (CO) or methane (CH4) was enhanced with optical excitation
of the plasmonic photocatalyst.29,33
6.4

Discussion of mechanisms
There are several possible mechanisms by which plasmon excitation of Cu nanoparticles may alter

the distribution of CO2 reduction products obtained from the plasmonic Cu/p-NiO photocathode. Here, we
consider three distinct plasmon-induced processes that could account for our observed photoelectrochemical
results. Photo-induced hot hole injection into the p-NiO valence band leads to increased electron density on
the Cu nanoparticles, which could potentially influence the reaction mechanism in a variety of ways. Hot
electrons may be selectively injected into available molecular orbitals of adsorbed species at the metalelectrolyte interface. If such a process were to occur preferentially on adsorbed CO2, hot electrons would
selectively activate CO2 to aid in formation of the CO2− species on the Cu surface. The generation of hot

111
electrons on the Cu nanoparticles via plasmon excitation may therefore help initiate the catalytic cascade on
the Cu surface by activating adsorbed CO2 molecules.
Alternatively, it is conceivable that plasmon excitation of the Cu nanoparticles serves to reduce the
evolution of H2 from the Cu surface through a process known as desorption induced by electronic transitions
(DIET).36,37 In this mechanism, preferential hot-electron transfer to adsorbed H2 or H2-evolving species (i.e.
protons, hydroxide ions, water) may destabilize surface bound molecules by populating anti-bonding orbitals
of the adsorbate and causing the molecule to dissociate on the Cu surface. Indeed, it has previously been
reported that plasmon-induced hot electrons can initiate photo-dissociation of H2 molecules.74-76 As an
additional consequence of molecular H2 dissociation, surface-bound hydrogen atoms would be available to
protonate nearby CO2− molecules and facilitate CO2 reduction. Such a process may be responsible for the
suppression of H2 observed upon optical excitation of the plasmonic Cu/p-NiO photocathode.
Finally, it is possible that plasmon-induced hot hole injection to the p-NiO support modifies the
molecular interactions with the Cu surface by altering the electronic structure of the Cu nanoparticles. It is
well known that the electronic structure of the metal d-bands plays the dominant mechanism in determining
molecular adsorption at a metal surface.77,78 Although a distribution of hot holes spanning the sp-band down
to the d-bands are created within the Cu nanoparticles upon visble-light excitation, direct transitions (d-band
to sp-band) are the dominate mechanism for hot-hole generation when irradiated above the interband
threshold of Cu (~1.6 – 1.8 eV).79,80 Thus, optical excitation of the Cu nanoparticles with 560 nm light (hv =
2.2 eV) preferentially excites hot holes within the metal d-bands77,78 that can then transfer to the underlying pNiO film. Injection of hot holes into the p-NiO valence band thereby alters the occupation of states below
the Cu Fermi level, which could thereby modify the molecular surface interactions by tuning the extent of
hybridization between the metal d-bands and the frontier orbitals of the adsorbate. This change in electronic
structure of the metal via plasmon-induced hot hole transfer to the p-NiO support offers an alternative
pathway towards shaping the selectivity of Cu nanoparticles.
Although the production of hydrogen (H2), carbon monoxide (CO), and formate (HCOO−) are all
thought to involve two proton-coupled electron-transfer steps, these three products originate from different
reactive intermediates formed on the Cu surface under reaction conditions.1,3 It therefore seems unlikely that
hot electron transfer is occurring preferentially to a short-lived CO2RR intermediate formed on the Cu surface
in operando, since the Faradaic efficiencies for both CO and HCOO− were observed to increase with light
excitation. Instead, we suspect that plasmon-induced hot electrons on the Cu nanoparticles likely play a key
role in improving the selectivity for the CO2RR by preferentially activating CO2 to form the CO2− anion.
Furthermore, if a fraction of the electrochemically-derived H2 molecules were photo-dissociated on the Cu

112
surface by hot electrons via DIET, the surface-bound hydrogen atoms would be readily available for
protonation of activated CO2− molecules. We also hypothesize that the injection of hot holes into the p-NiO
film changes the intrinsic binding affinity of the metal surface for reactant molecules by altering the d-band
structure of the Cu nanoparticles. In tandem, these processes could synergistically shape the selectivity of the
plasmonic Cu/p-NiO photocathode in favor of the CO2RR relative to the HER. As several reduction
reactions are occurring simultaneously on the plasmonic photocathode, and both the Cu nanoparticles and
the underlying p-NiO film are exposed to the electrolyte, advanced in operando spectroscopic studies are needed
to conclusively distinguish between these possible reaction mechanisms. At present, it remains unclear if
plasmon-induced hot electrons are transferred directly to adsorbed molecules on timescales commensurate
with electron-electron scattering processes (t ~ 10–100 fs) or if charge transfer occurs after establishing an
excited-state Fermi-Dirac distribution at an elevated electronic temperature (t > 1 ps). Nevertheless, these
initial photoelectrochemical observations indicate that optical excitation of the Cu nanoparticles alters the
selectivity of CO2 reduction relative to traditional electrochemical reduction performed under dark conditions.
6.5

Conclusion
In summary, we have demonstrated the benefits of using p-type NiO as a wide band gap support for

harvesting hot holes from Cu nanoparticles to allow the accumulation of hot electrons on the metal surface
and drive CO2 reduction with plasmonic Cu/p-NiO photocathodes. The collection of hot holes from the Cu
nanoparticles via injection to the p-NiO support also removes the requirements for sacrificial reagents
commonly employed in plasmon-induced photochemical reactions.

The Cu-wide band gap p-type

semiconductor Schottky junction design therefore represents a path forward for the realization of plasmondriven photocathodes capable of harnessing surface plasmon excitations to steer the selectivity of Cu surfaces
for photoelectrochemical CO2 reduction in aqueous media. We observed that plasmon excitation of the Cu
nanoparticles modulates the chemical selectivity for CO2 reduction products, increasing CO evolution and
HCOO− production while simultaneously suppressing H2 evolution. Several possible reaction mechanisms
are proposed to account for the observed influence of light on the selectivity of photoelectrochemical CO2
reduction. Although a conclusive assignment of the reaction mechanism requires in operando spectroscopy to
observe the molecular details of the reaction occurring on the plasmonic photocathode, we speculate that
plasmon-induced hot electrons likely play a key role in altering the selectivity of the reaction. Overall, our
photoelectrochemical results illustrate a promising strategy towards optically manipulating the catalytic
selectivity of Cu surfaces for CO2 conversion.

113

BIBLIOGRAPHY CHAPTER 6
Montoya, J. H.; Seitz, L. C.; Chakthranont, P.; Vojvodic, A.; Jaramillo, T. F.; Nørskov, J. K. Materials for
Solar Fuels and Chemicals. Nat. Mater. 2017, 16, 70−81.
2. Nitopi, S.; Bertheussen, E.; Scott, S. B.; Liu, X.; Engstfeld, A. K.; Horch, S.; Seger, B.; Stephens, I. E. L.;
Chan, K.; Hahn, C.; Nørskov, J. K.; Jaramillo, T. F.; Chorkendorff, I. Progress and Perspectives of
Electrochemical CO2 Reduction on Copper in Aqueous Electrolyte. Chem. Rev. 2019, 119, 7610−7672.
3. Birdja,Y.Y.;Peŕez-Gallent,E.;Figueiredo,M.C.;Göttle,A.J.; Calle-Vallejo, F.; Koper, M. T. M. Advances
and Challenges in Understanding the Electrocatalytic Conversion of Carbon Dioxide to Fuels. Nat.
Energy 2019, 4, 732−745.
4. Todorova, T. K.; Schreiber, M. W.; Fontecave, M. Mechanistic Understanding of CO2 Reduction
Reaction (CO2RR) Toward Multicarbon Products by Heterogeneous Copper-Based Catalysts. ACS
Catal. 2020, 10, 1754−1768.
5. Xu, S.; Carter, E. A. Theoretical Insights into Heterogeneous (Photo)electrochemical CO2 Reduction.
Chem. Rev. 2019, 119, 6631−6669.
6. White, J. L.; Baruch, M. F.; Pander, J. E., III; Hu, Y.; Fortmeyer, I. C.; Park, J. E.; Zhang, T.; Liao, K.;
Gu, J.; Yan, Y.; Shaw, T. W.; Ebelev, E.; Bocarsly, A. B. Light-Driven Heterogeneous Reduction of
Carbon Dioxide: Photocatalysts and Photoelectrodes. Chem. Rev. 2015, 115, 12888−12935.
7. Liu, M.; Pang, Y.; Zhang, B.; De Luna, P.; Voznyy, O.; Xu, J.; Zheng, X.; Dinh, C.-T.; Fan, F.; Cao, C.;
García de Arquer, F. P.; Safaei, T. S.; Mepham, A.; Klinkova, A.; Kumacheva, E.; Filleter, T.; Sinton, D.;
Kelley, S. O.; Sargent, E. H. Enhanced Electrocatalytic CO2 Reduction via Field-Induced Reagent
Concentration. Nature 2016, 537, 382−386.
8. Saberi Safaei, T.; Mepham, A.; Zheng, X.; Pang, Y.; Dinh, C.-T.; Liu, M.; Sinton, D.; Kelley, S. O.;
Sargent, E. H. High-Density Nanosharp Microstructures Enable Efficient CO2 Electroreduction. Nano
Lett. 2016, 16, 7224−7228.
9. De Luna, P.; Quintero-Bermudez, R.; Dinh, C.-T.; Ross, M. B.; Bushuyev, O. S.; Todorovic, P.; Regier,
T.; Kelly, S. O.; Yang, P.; Sargent, E. H. Catalyst Electro-Redeposition Controls Morphology and
Oxidation State for Selective Carbon Dioxide Reduction. Nat. Catal. 2018, 1, 103−110.
10. Kim, D.; Resasco, J.; Yu, Y.; Asiri, A. M.; Yang, P. Synergistic Geometric and Electronic Effects for
Electrochemical Reduction of Carbon Dioxide Using Gold-Copper Bimetallic Nanoparticles. Nat.
Commun. 2014, 5, 4948.
11. Kim, D.; Xie, C.; Becknell, N.; Yu, Y.; Karamad, M.; Chan, K.; Crumin, E. J.; Nørskov, J. K.; Yang, P.
Electrochemical Activation of CO2 through Atomic Ordering Transformations of AuCu Nanoparticles. J. Am. Chem. Soc. 2017, 139, 8329−8336.
12. Hahn, C.; Hatsukade, T.; Kim, Y.-G.; Vailionis, A.; Baricuatro, J. H.; Higgins, D. C.; Nitopi, S. A.;
Soriaga, M. P.; Jaramillo, T. F. Engineering Cu Surfaces for the Electrocatalytic Conversion of CO2:
Controlling Selectivity Toward Oxygenates and Hydrocarbons. Proc. Natl. Acad. Sci. U. S. A. 2017, 114,
5918−5923.
13. Lu, Q.; Rosen, J.; Zhou, Y.; Hutchings, G. S.; Kimmel, Y. C.; Chen, J. G.; Jiao, F. A Selective and
Efficient Electrocatalyst for Carbon Dioxide Reduction. Nat. Commun. 2014, 5, 3242.
14. Zhu, W.; Zhang, Y.-J.; Zhang, H.; Lv, H.; Li, Q.; Michalsky, R.; Peterson, A. A.; Sun, S. Active and
Selective Conversion of CO2 to CO on Ultrathin Au Nanowires. J. Am. Chem. Soc. 2014, 136, 16132−
16135.
1.

114
15. Zhu, W.; Michalsky, R.; Metin, Ö .; Lv, H.; Guo, S.; Wright, C. J.; Sun, X.; Peterson, A. A.; Sun, S.

Monodisperse Au Nanoparticles for Selective Electrocatalytic Reduction of CO2 to CO. J. Am. Chem.
Soc. 2013, 135, 16833−16836.
16. Mariano, R. G.; McKelvey, K.; White, H. S.; Kanan, M. W. Selective Increase in CO2 Electroreduction
Activity at Grain- Boundary Surface Terminations. Science 2017, 358, 1187−1192.
17. Feng, X.; Jiang, K.; Fan, S.; Kanan, M. W. A Direct Grain- Boundary-Activity Correlation for CO
Electroreduction on Cu Nanoparticles. ACS Cent. Sci. 2016, 2, 169−174.
18. Feng, X.; Jiang, K.; Fan, S.; Kanan, M. W. Grain-Boundary- Dependent CO2 Electroreduction Activity.
J. Am. Chem. Soc. 2015, 137, 4606−4609.
19. Hall, A. S.; Yoon, Y.; Wuttig, A.; Surendranath, Y. Mesostructure-Induced Selectivity in CO2 Reduction
Catalysis. J. Am. Chem. Soc. 2015, 137, 14834−14837.
20. Yoon, Y.; Hall, A. S.; Surendranath, Y. Tuning of Silver Catalyst Mesostructure Promotes Selective
Carbon Dioxide Conversion into Fuels. Angew. Chem., Int. Ed. 2016, 55, 15282−15286.
21. Welch, A. J.; DuChene, J. S.; Tagliabue, G.; Davoyan, A.; Cheng, W.-H.; Atwater, H. A. Nanoporous
Gold as a Highly Selective and Active Carbon Dioxide Reduction Catalyst. ACS Appl. Energy Lett. 2019,
2, 164−170.
22. Cao, Z.; Kim, D.; Hong, D.; Yu, Y.; Xu, J.; Lin, S.; Wen, X.; Nichols, E. M.; Jeong, K.; Reimer, J. A.;
Yang, P.; Chang, C. J. A Molecular Surface Functionalization Approach to Tuning Nano- particle
Electrocatalysts for Carbon Dioxide Reduction. J. Am. Chem. Soc. 2016, 138, 8120−8125.
23. Han, Z.; Kortlever, R.; Chen, H.-Y.; Peters, J. C.; Agapie, T. CO2 Reduction Selective for C≥2 Products
on Polycrystalline Copper with N-Substituted Pyridinium Additives. ACS Cent. Sci. 2017, 3, 853−859.
24. Rosen, B. A.; Salehi-Khojin, A.; Thorson, M. R.; Zhu, W.; Whipple, D. T.; Kenis, P. J. A.; Masel, R. I.
Ionic Liquid-Mediated Selective Conversion of CO2 to CO at Low Overpotentials. Science 2011, 334,
643−644.
25. Asadi, M.; Ki, K.; Liu, C.; Addepalli, A. V.; Abbasi, P.; Yasaei, P.; Phillips, P.; Behranginia, A.; Cerrato, J.
M.; Haasch, R.; Zapol, P.; Kumar, B.; Klie, R. F.; Abiade, J.; Curtiss, L. A.; Salehi-Khojin, A.
Nanostructured Transition Metal Dichalcogenide Electrocatalysts for CO2 Reduction in Ionic Liquid.
Science 2016, 353, 467−470.
26. Yu, S.; Wilson, A. J.; Kumari, G.; Zhang, X.; Jain, P. K. Opportunities and Challenges of Solar-EnergyDriven Carbon Dioxide to Fuel Conversion with Plasmonic Catalysts. ACS Energy Lett. 2017, 2,
2058−2070.
27. Creel, E. B.; Corson, E. R.; Eichhorn, J.; Kostecki, R.; Urban, J. J.; McCloskey, B. D. Directing
Selectivity of Electrochemical Carbon Dioxide Reduction Using Plasmonics. ACS Energy Lett. 2019, 4,
1098−1105.
28. DuChene, J. S.; Tagliabue, G.; Welch, A. J.; Cheng, W.-H.; Atwater, H. A. Hot Hole Collection and
Photoelectrochemical CO2 Reduction with Plasmonic Au/p-GaN Photocathodes. Nano Lett. 2018, 18,
2545−2550.
29. Zhang, X.; Li, X.; Zhang, D.; Su, N. Q.; Yang, W.; Everitt, H. O.; Liu, J. Product Selectivity in
Plasmonic Photocatalysis for Carbon Dioxide Hydrogenation. Nat. Commun. 2017, 8, 14542.
30. Yu, S.; Jain, P. K. Plasmonic Photosynthesis of C1-C3 Hydrocarbons from Carbon Dioxide Assisted by
an Ionic Liquid. Nat. Commun. 2019, 10, 2022.
31. Yu, S.; Wilson, A. J.; Heo, J.; Jain, P. K. Plasmonic Control of Multi-Electron Transfer and C-C
Coupling in Visible-Light-Driven CO2 Reduction on Au Nanoparticles. Nano Lett. 2018, 18, 2189−
2194.
32. Yu, S.; Jain, P. K. Selective Branching of Plasmonic Photosynthesis into Hydrocarbon Production and
Hydrogen Generation. ACS Energy Lett. 2019, 4, 2295−2300.

115
33. Robatjazi, H.; Zhao, H.; Swearer, D. F.; Hogan, N. J.; Zhou, L.; Alabastri, A.; McClain, M. J.;

Nordlander, P.; Halas, N. J. Plasmon- Induced Selective Carbon Dioxide Conversion on EarthAbundant Aluminum-Cuprous Oxide Antenna-Reactor Nanoparticles. Nat. Commun. 2017, 8, 27.
34. Linic, S.; Christopher, P.; Ingram, D. B. Plasmonic-Metal Nanostructures for Efficient Conversion of
Solar to Chemical Energy. Nat. Mater. 2011, 10, 911−921.
35. Brongersma, M. L.; Halas, N. J.; Nordlander, P. Plasmon- Induced Hot Carrier Science and Technology.
Nat. Nanotechnol. 2015, 10, 25−34.
36. Christopher, P.; Moskovits, M. Hot Charge Carrier Trans- mission from Plasmonic Nanostructures.
Annu. Rev. Phys. Chem. 2017, 68, 379−398.
37. Linic, S.; Aslam, U.; Boerigter, C.; Morabito, M. Photochemical Transformations on Plasmonic Metal
Nanoparticles. Nat. Mater. 2015, 14, 567−576.
38. Hartland, G. V.; Besteiro, L. V.; Johns, P.; Govorov, A. O. What’s so Hot about Electrons in Metal
Nanoparticles? ACS Energy Lett. 2017, 2, 1641−1653.
39. Zhang, Y.; He, S.; Guo, W.; Huang, J.; Mulcahy, J. R.; Wei, W. D. Surface-Plasmon-Driven Hot
Electron Photochemistry. Chem. Rev. 2018, 118, 2927−2954.
40. Knight, M. W.; Sobhani, H.; Nordlander, P.; Halas, N. J. Photodetection with Active Optical Antennas.
Science 2011, 332, 702−704.
41. Zheng, B. Y.; Zhao, H.; Manjavacas, A.; McClain, M.; Nordlander, P.; Halas, N. J. Distinguishing
Between Plasmon-Induced and Photoexcited Carriers in a Device Geometry. Nat. Commun. 2015, 6,
7797.
42. Li, W.; Valentine, J. Metamaterial Perfect Absorber Based Hot Electron Photodetection. Nano Lett.
2014, 14, 3510−3514.
43. Tagliabue, G.; Jermyn, A. S.; Sundararaman, R.; Welch, A. J.; DuChene, J. S.; Pala, R.; Davoyan, A. R.;
Narang, P.; Atwater, H. A. Quantifying the Role of Surface Plasmon Excitation and Hot Carrier
Transport in Plasmonic Devices. Nat. Commun. 2018, 9, 3394.
44. Mubeen, S.; Lee, J.; Singh, N.; Kramer, S.; Stucky, G. D.; Moskovits, M. An Autonomous
Photosynthetic Device in Which All Charge Carriers Derive from Surface Plasmons. Nat. Nanotechnol.
2013, 8, 247−251.
45. Mubeen, S.; Lee, J.; Liu, D.; Stucky, G. D.; Moskovits, M. Panchromatic Photoproduction of H2 with
Surface Plasmons. Nano Lett. 2015, 15, 2132−2136.
46. Tian, Y.; Tatsuma, T. Mechanisms and Applications of Plasmon-Induced Charge Separation at TiO2
Films Loaded with Gold Nanoparticles. J. Am. Chem. Soc. 2005, 127, 7632−7637.
47. DuChene, J. S.; Sweeny, B. C.; Johnston-Peck, A. C.; Su, D.; Stach, E. A.; Wei, W. D. Prolonged Hot
Electron Dynamics in Plasmonic-Metal/Semiconductor Heterostructures with Implications for Solar
Photocatalysis. Angew. Chem., Int. Ed. 2014, 53, 7887−7891.
48. Pu, Y.-C.; Wang, G.; Chang, K.-D.; Ling, Y.; Lin, Y.-K.; Fitzmorris, B. C.; Liu, C.-M.; Lu, X.; Tong, Y.;
Zhang, J. Z.; Hsu, Y.-J.; Li, Y. Au Nanostructure-Decorated TiO2 Nanowires Exhibiting Photoactivity
Across Entire UV-visible Region for Photoelectrochemical Water Splitting. Nano Lett. 2013, 13,
3817−3823.
49. Qian, K.; Sweeny, B. C.; Johnston-Peck, A. C.; Niu, W.; Graham, J. O.; DuChene, D. J.; Qiu, J.; Wang,
Y.-C.; Engelhard, M. E.; Su, D.; Stach, E. A.; Wei, W. D. Surface Plasmon-Driven Water Reduction:
Gold Nanoparticle Size Matters. J. Am. Chem. Soc. 2014, 136, 9842−9845.
50. Schlather, A. E.; Manjavacas, A.; Lauchner, A.; Marangoni, V. S.; DeSantis, C. J.; Nordlander, P.; Halas,
N. J. Hot Hole Photoelectrochemistry on Au@SiO2@Au Nanoparticles. J. Phys. Chem. Lett. 2017, 8,
2060−2067.
51. Zhao, J.; Nguyen, S. C.; Ye, R.; Ye, B.; Weller, H.; Somorjai, G. A.; Alivisatos, A. P.; Toste, F. D. A
Comparison of Photocatalytic Activities of Gold Nanoparticles Following Plasmonic and Interband

116
Excitation and a Strategy for Harnessing Interband Hot Carriers for Solution Phase Photocatalysis. ACS
Cent. Sci. 2017,3,482−488.
52. Al-Zubeidi, A.; Hoener, B. S.; Collins, S. S. E.; Wang, W. Kirchner, S. R.; Jebeli, S. A. H.; Joplin, A.;
Chang, W.-S.; Link, S.; Landes, C. F. Hot Holes Assist Plasmonic Nanoelectrode Dissolution. Nano Lett.
2019, 19, 1301−1306.
53. Yu, W.; Wijesekara, K. D.; Xi, X.; Willets, K. A. Quantifying Wavelength-Dependent Plasmonic Hot
Carrier Energy Distributions at Metal/Semiconductor Interfaces. ACS Nano 2019, 13, 3629−3637.
54. Pensa, E.; Garguilo, J.; Lauir, A.; Schlucker, S.; Cortes, A.; Maier, S. A. Spectral Screening of the Energy
of Hot Holes over a Particle Plasmon Resonance. Nano Lett. 2019, 19, 1867−1874.
55. Matsui, T.; Li, Y.; Hsu, M.-H. M.; Merckling, C.; Oulton, R. F.; Cohen, L. F.; Maier, S. A. Highly Stable
Plasmon Induced Hot Hole Transfer into Silicon via a SrTiO3 Passivation Interface. Adv. Funct. Mater.
2018, 28, 1705829.
56. Güsken, N. A.; Lauri, A.; Li, Y.; Matsui, T.; Doiron, B.; Bower, R.; Regoutz, A.; Mihai, A.; Petrov, P. K.;
Oulton, R. F.; Cohen, L. F.; Maier, S. A. TiO2-x - Enhanced IR Hot Carrier Based Photodetection in
Metal Thin Film-Si Junctions. ACS Photonics 2019, 6, 953−960.
57. Tanzid, M.; Ahmadivand, A.; Zhang, R.; Cerjan, B.; Sobhani, A.; Yazdi, S.; Nordlander, P.; Halas, N. J.
Combining Plasmonic Hot Carrier Generation with Free Carrier Absorption for High-Perform- ance
Near-Infrared Silicon-Based Photodetection. ACS Photonics 2018, 5, 3472−3477.
58. Cai, Y.-Y.; Collings, S. S. E.; Gallagher, M. J.; Bhattacharjee, U.; Zhang, R.; Chow, T. H.; Ahmadivand,
A.; Ostovar, B.; Al-Zubeidi, A.; Wang, J.; Nordlander, P.; Landes, C. F.; Link, S. Single-Particle
Emission Spectroscopy Resolves d-Hole Relaxation in Copper Nanocubes. ACS Energy Lett. 2019, 4,
2458−2465.
59. He, J.; Lindström, E.; Hagfeldt, A.; Lindquist, S.-E. Dye- Sensitized Nanostructured p-Type Nickel
Oxide Film as a Photo- cathode for a Solar Cell. J. Phys. Chem. B 1999, 103, 8940−8943.
60. Irwin, M. D.; Buchholz, D. B.; Hains, A. W.; Chang, R. P. H.; Marks, T. J. p-Type Semiconducting
Nickel Oxide as an Efficiency- Enhancing Anode Interfacial Layer in Polymer Bulk-Heterojunction
Solar Cells. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 2783−2787.
61. Seo, S.; Park, I. J.; Kim, M.; Lee, S.; B, C.; Jung, H. S.; Park, N.- G.; Kim, J. Y.; Shin, H. An Ultra-Thin,
Un-Doped NiO Hole Transporting Layer of Highly Efficient (16.4%) Organic-Inorganic Hybrid
Perovskite Solar Cells. Nanoscale 2016, 8, 11403.
62. Kamata, R.; Kumagai, H.; Yamazaki, Y.; Sahara, G.; Ishitani, O. Photoelectrochemical CO2 Reduction
Using a Ru(II)-Re(I) Supra- molecular Photocatalyst Connected to a Vinyl Polymer on a NiO Electrode.
ACS Appl. Mater. Interfaces 2019, 11, 5632−5641.
63. Thimsen, E.; Martinson, A. B. F.; Elam, J. W.; Pellin, M. J. Energy Levels, Electronic Properties, and
Rectification in Ultrathin p-NiO Films Synthesized by Atomic Layer Deposition. J. Phys. Chem. C 2012,
116, 16830−16840.
64. Sai Gautam, G.; Senftle, T. P.; Alidoust, N.; Carter, E. A. Novel Solar Cell Materials: Insights from FirstPrinciples. J. Phys. Chem. C 2018, 122, 27107−27126.
65. Robatjazi, H.; Bahauddin, S. M.; Doiron, C.; Thomann, I. Direct Plasmon-Driven Photoelectrocatalysis.
Nano Lett. 2015, 15, 6155−6161.
66. Nakamura, K.; Oshikiri, T.; Ueno, K.; Wang, Y.; Kamta, Y.; Kotake, Y.; Misawa, H. Plasmon-Enhanced
Photocurrent Generation and Water Oxidation with a Gold Nanoisland-Loaded Titanium Dioxide
Photoelectrode. J. Phys. Chem. Lett. 2016, 7, 1004−1009.
67. Kao, K.-C.; Kuroiwa, Y.; Nishi, H.; Tatsuma, T. Hydrogen Evolution from Water Based on PlasmonInduced Charge Separation at a TiO2/Au/NiO/Pt System. Phys. Chem. Chem. Phys. 2017, 19,
31429−31435.
68. Anderson, P. A. The Work Function of Copper. Phys. Rev. 1949, 76, 388−390.

117
69. Dabera, G. D. M. R.; Walker, M.; Sanchez, A. M.; Pereira, H. J.; Beanland, R.; Hatton, R. A. Retarding

Oxidation of Copper Nanoparticles Without Electrical IIsolation and the Size Dependence of Work
Function. Nat. Commun. 2017, 8, 1894.
70. Mendez-Medrarno, M. G.; Kowalska, E.; Lehoux, A.; Herriisan, A.; Othani, B.; Bahena, D.; Briois, V.;
Colbeau-Justin, C.; Rodríguez- Loṕez, J. L.; Remita, H. Surface Modification of TiO2 with Ag
Nanoparticles and CuO Nanoclusters for Application in Photo- catalysis. J. Phys. Chem. C 2016, 120,
5143−5154.
71. Yu, J.; Zhuang, S.; Xu, X.; Zhu, W.; Feng, B.; Hu, J. Photogenerated Electron Reservoir in Hetero-p-n
CuO-ZnO Nano-composite Device for Visible-Light-Driven Photocatalytic Reduction of Aqueous
Cr(VI). J. Mater. Chem. A 2015, 3, 1199−1207.
72. Scott, S. B.; Hogg, T. V.; Landers, A. T.; Maagaard, T.; Bertheussen, E.; Lin, J. C.; Davis, R. C.; Beeman,
J. W.; Higgins, D.; Drisdell, W. S.; Hahn, C.; Mehta, A.; Seger, B.; Jaramillo, T. F.; Chorkendorff, I.
Absence of Oxidized Phases in Cu under CO Reduction Conditions. ACS Energy Lett. 2019, 4, 803−804.
73. Corson, E. R.; Creel, E. B.; Kim, Y.; Urban, J. J.; Kostecki, R.; McCloskey, B. D. A TemperatureControlled Photoelectrochemical Cell for Quantitative Product Analysis. Rev. Sci. Instrum. 2018, 89,
055112.
74. Mukherjee, S.; Zhou, L.; Goodman, A. M.; Large, N.; Ayala- Orozco, C.; Zhang, Y.; Nordlander, P.;
Halas, N. J. Hot-Electron- Induced Dissociation of H2 on Gold Nanoparticles Supported on SiO2. J.
Am. Chem. Soc. 2014, 136, 64−67.
75. Zhou, L.; Zhang, C.; McClain, M. J.; Manjavacas, A.; Krauter, C. M.; Shu, T.; Berg, F.; Everitt, H. O.;
Carter, E. A.; Nordlander, P.; Halas, N. J. Aluminum Nanocrystals as a Plasmonic Photocatalyst for
Hydrogen Dissociation. Nano Lett. 2016, 16, 1478−1484.
76. Zhang, C.; Zhao, H.; Zhou, L.; Schlather, A. E.; Dong, L.; McClain, M. J.; Swearer, D. F.; Nordlander,
P.; Halas, N. J. Al-Pd Nanodisk Heterodimers as Antenna-Reactor Photocatalysts. Nano Lett. 2016, 16,
6677−6682.
77. Nørskov, J. K.; Abild-Pedersen, F.; Studt, F.; Bligaard, T. Density Functional Theory in Surface
Chemistry and Catalysis. Proc. Natl. Acad. Sci. U. S. A. 2011, 108, 937−943.
78. Hammer, B.; Nørskov, J. K. Theoretical Surface Science and Catalysis Calculations and Concepts. Adv.
Catal. 2000, 45, 71−129.
79. Sundararaman, R.; Narang, P.; Jermyn, A. S.; Goddard, W. A., III; Atwater, H. A. Theoretical
Predictions for Hot-Carrier Generation from Surface Plasmon Decay. Nat. Commun. 2014, 5, 5788.
80. Brown, A. M.; Sundararaman, R.; Narang, P.; Goddard, W. A., III; Atwater, H. A. Nonradiative
Plasmon Decay and Hot Carrier Dynamics: Effects of Phonons, Surfaces, and Geometry. ACS Nano
2016, 10, 957−966.
81. Wuttig, A.; Yaguchi, M.; Motobayashi, K.; Osawa, M.; Surendranath, Y. Inhibited Proton Transfer
Enhances Au-Catalyzed CO2-to-Fuels Selectivity. Proc. Natl. Acad. Sci. U. S. A. 2016, 113,
E4585−E4593.
82. Christopher, P.; Xin, H.; Marimuthu, A.; Linic, S. Singular Characteristics and Unique Chemical Bond
Activation Mechanisms of Photocatalytic Reactions on Plasmonic Nanostructures. Nat. Mater. 2012, 11,
1044−1050.
83. Boerigter, C.; Campana, R.; Morabito, M.; Linic, S. Evidence and Implications of Direct Charge
Excitation as the Dominant Mechanism in Plasmon-Mediated Photocatalysis. Nat. Commun. 2016, 7,
10545.
84. Boerigter, C.; Aslam, U.; Linic, S. Mechanism of Charge Transfer from Plasmonic Nanostructures to
Chemically Attached Materials. ACS Nano 2016, 10, 6108−6115.

118
85. Rao, V. G.; Aslam, U.; Linic, S. Chemical Requirement for Extracting Energetic Charge Carriers from

Plasmonic Metal Nano- particles to Perform Electron-Transfer Reactions. J. Am. Chem. Soc. 2019, 141,
643−647.
86. Aslam, U.; Rao, V. G.; Chavez, S.; Linic, S. Catalytic Conversion of Solar to Chemical Energy on
Plasmonic Metal Nanostructures. Nat. Catal. 2018, 1, 656−665.
87. Wu, K.; Chen, J.; McBride, J. R.; Lian, T. Efficient Hot-Electron Transfer by a Plasmon-Induced
Interfacial Charge-Transfer Transition. Science 2015, 349, 632−635.
88. Tan, S.; Argondizzo, A.; Ren, J.; Liu, L.; Zhao, J.; Petek, H. Plasmonic Coupling at a
Metal/Semiconductor Interface. Nat. Photonics 2017, 11, 806−812.
89. Kumar, P. V.; Rossi, T. P.; Marti-Dafcik, D.; Reichmuth, D.; Kuisma, M.; Erhart, P.; Puska, M. J.;
Norris, D. J. Plasmon-Induced Direct Hot-Carrier Transfer at Metal-Acceptor Interfaces. ACS Nano
2019, 13, 3188−3195.
90. Zhang, Y.; Nelson, T.; Tretiak, S.; Guo, H.; Schatz, G. C. Plasmonic Hot-Carrier-Mediated Tunable
Photochemical Reactions. ACS Nano 2018, 12, 8415−8422.

119
Chapter 7

BICARBONATE OR CARBONATE PROCESSES FOR COUPLING CARBON
DIOXIDE CAPTURE AND ELECTROCHEMICAL CONVERSION
7.1

Introduction
Designing a scalable system to capture CO2 from the air and convert it into valuable chemicals, fuels,

and materials could be transformational for mitigating climate change.1–3 Climate models predict that negative
greenhouse gas emissions will be required by the year 2050 in order to stay below a 2 °C change in global
temperature.4 The processes of CO2 capture, CO2 conversion, and finally product separation all require
significant energy inputs; devising a system that simultaneously minimizes the energy required for all steps is an
important challenge. To date, a variety of prototype or pilot-level CO2 capture and/or conversion systems have
been designed and built targeting the individual objectives of either capture or conversion. One approach has
focused on CO2 removal from the atmosphere and storage of pure pressurized CO2.5,6 Other efforts have
concentrated on CO2 conversion processes, such as electrochemical reduction7–10 or fermentation.10,11 Only a
few concepts or analyses have been developed for complete end-to-end processes that perform both CO2
capture and transformation.12,13

120

cement
mineral

al
ic
em
ch
tro
ec
El

Release CO2
from capture
material

Capture
CO2

CO2
source

[18]

[13]
[12]

[15]

Transform
HCO3 /CO32-

CO2
[7-9]

CO2
[11]

[10]

Transform
CO2

CH3OH
C2H5OH

[6]
[5]

Compress CO2

CO, HCOOH
CO2
H2O

CH4
C2H5OH,C2H3O2C6H13OH, C4H9OH

Figure 7.1: Schematic representation of the steps for various prototype systems designed to capture CO2 and/or convert it to either
concentrated and pressurized CO2 or to a value-added product. The blue arrows represent prototype processes that capture and convert
CO2, grey arrows represent prototype processes that only focus on CO2 conversion, and the pink arrow represents the process that we
propose. The numbers in brackets correspond to references for the various processes [5-13,15,18].
Here we explore an approach for the design of a CO2 capture and conversion system - (i) formation of
bicarbonate or carbonate ((bi)carbonate) through dissolution of CO2 into basic solution, followed by (ii)
electrochemical reduction to syngas or formate, and finally (iii) transformation into useful chemicals, fuels, and
materials. Unlike traditional electrochemical systems, in which gaseous CO2 is the primary chemical feedstock
that is converted into products, our analysis focuses on the transformation of CO2 carried by (bi)carbonate
solutions (Figure 7.1). This approach offers several advantages for coupling CO2 capture with conversion. First,
capturing CO2 from the atmosphere does not require an energy-intensive heating step to recover gaseous CO2
from a capture medium for later conversion. Second, by transforming the CO2 carried by (bi)carbonate ions,
the process avoids the energy-intensive compression of gaseous CO2 and allows for significantly higher
conversion efficiency per mol of captured CO2. We outline the energy requirements for various steps in this

121
(bi)carbonate feedstock route and compare it to systems with conventional CO2 capture and conversion
processes.
7.2

CO2 capture and conversion steps

CO2 Source
Environment
- Air
- Seawater

Capture CO2
Absorption
- e.g. amines
Adsorption
- e.g. zeolites

Point Source
- Petrochemical Industry
Membranes
- e.g. ethylene
- e.g. fibers
- Iron and Steel
- Cement production
- Natural gas sweetening Biological
- Fossil fuels or Minerals - e.g. algae
- e.g. coal
Other
- e.g. mineralization

Release from Compress CO2
capture material
Heat
Vacuum cycle

~15 PSIA for
fermentation
~15-50 PSIA for
electrochemical
processes
~250-1450 PSIA
for thermochemical
processes

Transform CO2
Electrochemical
Fermentation
Photo-electrochemical
Photocatalysis
Thermocatalytic
Thermolysis
Industrial Processes

Products
CxHyOz
CxH2x
CxH2x+1
CxH2x+2
Diesel
Syngas
Ethylene
Butanol
Hexanol
Propanol
Ethanol
Plastics

Figure 7.2: Synopsis of the various steps involved in capturing CO2 and transforming it into valuable chemicals.
There are many possible configurations for CO2 capture and conversion systems, but five steps are
integral to every conventional design (Figure 7.2): (i) a CO2 source, (ii) a capture medium, (iii) a process to release
CO2 from the capture medium, (iv) CO2 compression into a concentrated gas stream, and (v) conversion of CO2
into fuels, chemicals, and/or materials (e.g. hydrocarbons). The design of an energy-efficient integrated system
for capture and conversion requires careful consideration of the CO2 source. Whether it is a point source with
a mixed-process gas stream (i.e. flue gas from a coal-fired power plant) or a relatively dilute source (i.e.
atmosphere) will dictate the CO2 concentration in the feedstock and therefore the type of capture media that is
most appropriate to use. There are a variety of options for CO2 capture media, ranging from polymer
membranes to organic and inorganic liquid adsorbers.14 After collection by the capture media, the release of
CO2 is usually accomplished by heating the capture media itself. The collected CO2 must then be compressed
to a suitable pressure for flow and injection into a conversion system. As will be shown below, these two
processes related to CO2 capture and compression potentially constitute a major energy input that must be
considered when designing a complete system for mitigating global atmospheric CO2.
To better facilitate an end-to-end comparison of complete systems, we analyze CO2 capture and
conversion systems that have been realized to date at a pilot-plant level. We divide them into three categories:

122
systems that (i) perform direct air capture, (ii) use a concentrated CO2 source for conversion, and (iii) only focus
on the conversion of CO2. The steps for each pilot-plant and prototype are shown in Figure 7.1, and a more
detailed schematic is given in the Appendix A: Figures A1-A2. We compare the pilot processes using the metric
of energy required to remove/transform one mole of CO2 from the atmosphere because this metric directly
affects the operational expenses of the plant. However, it should be noted that this metric does not account for
other expenses (labor, maintenance, materials, etc.) needed to operate the plant, or the construction capital
expense and amortization period. Due to these considerations, we have therefore chosen to not compare
fermentation processes to electrochemical or thermocatalytic processes. Fermentation is an interesting CO2
conversion technique that shows significant promise because of its high selectivity for products like ethanol and
other multiple carbon products.10–12 Fermenters also operate near room temperature, making the energy
requirements per mole of CO2 transformed low (Appendix A:1G-H).10,11 However, fermentation processes
require labor to operate and maintain the system, reaction times can be long, and fermenters are typically
operated as a batch reactor rather than in a continuous chemical process. These considerations make it difficult
to directly compare electrochemical and fermentation processes using only energy required per mole of CO2
captured/transformed as a metric. In this perspective, we are proposing an electrochemical process and will
therefore focus our attention on comparing it to similar systems.
Pilot-plants that remove CO2 from the atmosphere and produce concentrated and pressurized CO2
streams have been designed using a capture medium of potassium hydroxide (Carbon Engineering)5 or
polyamines (ClimeWorks).6 The potassium hydroxide process requires ~0.28 MJ per mole of CO2 (kJ/mol CO2)
removed from the atmosphere (Appendix A:1A)5 while the polyamine process requires 0.58 MJ/mol CO2.6 For
both of these atmospheric CO2 capture processes, the enthalpy required to release the CO2 from the capture
medium and subsequent pressurization of the gaseous CO2 dominates the total energy requirements of the
system.
Other pilot-scale processes have been developed that capture CO2 from concentrated sources. For
instance, CO2 emitted from a geothermal power plant is dissolved into a potassium hydroxide solution, which
is then pumped underground to react with naturally occurring basalt rock (CarbFix).15 By taking advantage of
location-specific geology for CO2 transformation, the system only requires 0.13 MJ/mol CO2 (Appendix A:1B).16
Another process captures CO2 from a geothermal power plant, then pressurizes and heats the CO2 in the
presence of H2 generated via water electrolysis to create methanol and water (Carbon Recycling International).
This process requires 1.53 MJ/mol CO2 (Appendix A:1C), but it also produces a chemical fuel in the form of
methanol.13 As expected, this process requires more energy than the other systems described because it
incorporates a CO2 conversion process.

123
Pilot-scale CO2 conversion systems have also received extensive attention.7–11,17 A pilot-scale CO2
electrolyzer was shown to create CO and O2 using an energy input of 0.81 MJ/mol CO2 (Appendix A:1D).9
Other pilot-scale CO2 electrolyzers have been demonstrated to make CO7 or formic acid8, requiring 0.61 MJ/mol
CO2 or 0.75 MJ/mol CO2, respectively (Appendix A:1E). Finally, an approach for carbon sequestration uses a
process that injects CO2 into cement during production, allowing cement producers to use less binder and offset
their CO2 emissions in the process; using this method, 25 lbs of CO2 can be stored per cubic yard of cement
(Appendix A:1H).17 Notably, none of the these transformation processes account for the energy required to
capture the CO2 from the atmosphere or a concentrated point source. As mentioned above, procuring a
pressurized CO2 source in suitable purity does not represent a trivial amount of energy.
In order to construct an alternative process with a modest energy consumption alternative, we seek to
evaluate the energy intensive steps of each process. Energy consumption varies greatly depending on the source
of CO2 and the desired products. While the process of CO2 conversion requires significantly greater energy than
CO2 capture, conversion schemes discussed above have yet to report their conversion efficiency. Based on our
calculations, all reported devices have CO2 conversion efficiencies less than 40% (Appendix A:3A-H) with the
majority having conversion efficiencies less than 10%.18 For many electrolysis processes the energy required per
mole of CO2 would at minimum double unless the unused CO2 is recycled through the system. To decrease the
overall energy requirements for CO2 capture and conversion, it is interesting to consider a system that converts
nearly 100% of the captured CO2.
7.3

Direct (bi)carbonate conversion

CO2 Source

CO2 Capture Media

Atmospheric
CO2

Basic Solution
(e.g. KOH)
Circulating Fans
0.01 MJ/mol CO2

CO2/HCO3−/CO32−
Processing

CO32−
HCO3−

Electrolyzer
Gas-Diffusion
Electrode
Bipolar Membrane
0.7 MJ/mol CO2

Total Energy
Requirements

CO + H2
0.7 MJ/mol CO
Fermenter
C6H13OH
C4H9OH
C2H5OH
C2H3O2-

O2

Coal

Coal Pyrolysis
400°C

Char

Gasify Char
800 – 1800°C

CO + H2
3-7.5 MJ/mol CO

Contaminants

Natural
Gas

Steam Methane
Reform

CO, H2

Syngas
Processing

CO + H2
Reverse Water
Gas Shift

0.8 MJ/mol CO

CuZnOxAlOx
CH3OH
Fischer-Tropsch
Diesel
CxH2x
CxH2x+1
CxH2x+2

H2O, CH4
Trace levels of H2O + CH4

Figure 7.3: Schematic representation of processes and energy requirements for various proposed schemes that capture CO2 and

124
transform it to value-added products. The bottom rows show the commercial syngas synthesis process, with either a feed stock of coal
or of natural gas. The energy needed to produce 1 mol of CO while related to the cost needed to operate the plant does not encompass
the full picture of expenses such as materials, maintenance, and labor.
Electrochemically reducing CO2 carried by (bi)carbonate instead of pressurized CO2 bypasses energy
intensive processes necessary for concentration and compression of gaseous CO2 (see Figure 7.3 and Figure 7.4
for schematics of the overall process designs). This change in the carbon-bearing feedstock molecule allows us
to avoid energy-intensive processes, such as heating or vacuum cycling of a sorbent material to release the
captured CO2, along with the subsequent compression of the captured CO2 into a sufficient pressure that can
adequately supply a CO2 conversion system, such as a gas-diffusion electrolyzer. Furthermore, transforming the
CO2 carried by a (bi)carbonate ion allows for high conversion efficiency per mol of captured CO2, unlike many
other electrolysis devices operating with gaseous CO2 as the carbon input (Appendix A:3A-H).
An example schematic system could combine (bi)carbonate formation with syngas production (Figure
7.3), which can then be flexibly converted to valuable chemicals, fuels, and materials. Assuming the first step of
CO2 capture uses an already demonstrated air contactor design5, ~0.01 MJ/mol CO2 would be required
(Appendix A:2A), depending on the alkalinity of the solution. The solution of (bi)carbonate could then be fed
into an electrolyzer where it is converted to syngas (CO and H2). Recent reports of gas-diffusion electrodes
using a silver catalyst at the cathode and a bipolar membrane have shown that CO2 can be generated locally near
the catalyst from a (bi)carbonate electrolyte, which can then be converted to syngas.19,20 Importantly, no CO2
was observed in the output stream of chemical products,19 thus eliminating the need to separate the syngas from
the unconverted CO2 stream. These reported prototype devices19,20 required between ~0.7 MJ/mol CO2 and
achieved a desirable syngas ratio of between 3:1 and 2:1 (H2:CO) (Appendix A:2B and Appendix A:2C). In all
of the electrolysis processes discussed above, the oxygen evolution reaction (OER) occurs at the counter
electrode. It should also be noted that the base does not need to be replenished as long as the rate of CO2
capture is matched to the rate of CO production. To create a plant that could capture and transform 1 metric
ton of CO2 per day would require a capture system of approximately 16.6 m2, an electrode area of 0.4 m2, and
an array of solar panels covering an area of 4,571 m2 (Appendix A:2D). Additional space for solar energy storage
and syngas processing would also be required, but would be negligible compared to the area needed for the
photovoltaic system generating the electricity. If this system were to be used to capture all CO2 emitted globally
daily, approximately 100 million tons of CO2, the system would need to be around 460,000 km2. This is slightly
larger than the area of California (~424,000 km2), and only around 0.09% of the area of the earth.
Notably, the total energy requirement of the (bi)carbonate conversion system is 0.7 MJ/mol CO to make
syngas, while commercial processes to make syngas from coal require 3-7.5 MJ/mol CO (Appendix A:2E).21 To

125
make syngas from natural gas requires 0.8 MJ/mol CO (Appendix A:2F-G).22 It should be noted, however, that
the energies listed for these conventional industrial processes do not account for the energy required to extract
the coal or natural gas feedstocks. The (bi)carbonate process we have outlined compares favorably on an
energetic basis to current commercial processes, and unlike typical gasification systems, it also produces a syngas
stream without any contaminants. As is well known, syngas is a versatile feedstock for further generation of a
variety of useful chemicals and fuels. For example, butanol, hexanol, acetate, and ethanol can be synthesized
from syngas by using a fermenter.10,12 Another option is to synthesize methanol by heating and pressurizing
syngas in the presence of a CuZnOxAlOx catalyst.23 A third option would be to use Fischer-Tropsch processes
to make longer-chain hydrocarbons, such as diesel fuel, which can be synthesized at 240 °C and 25 bar in the
presence of a cobalt catalyst.24
CO2 Source

CO2 Capture Media

Atmospheric
CO2

Basic Solution
(e.g. NaOH)
Circulating Fans
0.01 MJ/mol CO2

CO2/HCO3−
Processing

HCO3−

Electrolyzer
Pd-Pt Catalyst
0.4 MJ/mol CO2

Total Energy
Requirements

HCOOH
0.8 MJ/mol HCOOH

Chlor-Alkali
Processing
0.4 MJ/mol NaOH

Fermentation
Biomass
Catalysis
CH3OH

O2

NaOH

Formic Acid
Processing

Formic Acid
Fuel Cell
Electrical
Power
Hydrogen
Carrier
H2

Brine

Cl2

Figure 7.4: Schematic representation of processes and energy requirements for various proposed schemes that capture CO2 and
transform into formic acid.
Syngas is not the only chemical product that can be electrochemically synthesized from CO2-bearing
bicarbonate solutions. Several studies have shown that bicarbonate solutions can also be electrochemically
reduced to formate at low overpotentials (Appendix A:2H).25–28 The global market for formic acid is relatively
small (currently ~570 million USD) but expected to grow significantly in the coming years because of the
increased use of formic acid in the rubber industry.29 The energy required to electrochemically synthesize
formate from bicarbonate via atmospheric CO2 capture is 0.8 MJ/mol CO2 (Figure 7.4). Not only is formic acid
a useful chemical in and of itself, but formic acid can also be transformed into valuable chemicals such as
methanol by either molecular catalysts,30–33 thermocatalysis,34–36 or fermentation processes.37 Although the heat

126
of combustion is modest (0.25 MJ/mol), formic acid could conceivably be used to generate power either directly
in a formic acid fuel cell38–41 or as a hydrogen carrier42–45 for a hydrogen fuel cell.
7.4

Conclusion
This system-level design approach for CO2 capture and conversion highlights the favorable characteristics

of processes that use a (bi)carbonate solution as the carbon-bearing feedstock. Instead of reducing CO2 directly
from a pressurized gas stream, CO2 molecules contained in (bi)carbonate solutions are electrochemically reduced
to form value-added chemicals, thus removing energy intensive steps, and with a potential for a near-unity
conversion rate of the captured CO2 molecules. While we have focused on syngas production in our analyses,
the main advantage of the proposed process involves changing the starting feedstock from gaseous CO2 to
aqueous CO2 carrying (bi)carbonate solutions, whose advantage is enhanced when coupled with a direct-air
capture system to remove atmospheric CO2.

BIBLIOGRAPHY CHAPTER 7
1. Figueres, C.; et al. Emissions are still rising: ramp up the cuts. Nature 2018, 564, 27−30.
2. Huntingford, C.; Mercado, L. M. High chance that current atmospheric greenhouse concentrations
commit to warmings greater than 1.5°C over land. Sci. Rep. 2016, 6, 1−7.
3. Jackson, R. B. Global energy growth is outpacing decarbonization. Environ. Res. Lett. 2018, 13, 120401.
4. Le Quere, C. et al. Global carbon budget 2018; 2018.
5. Keith, D. W.; Angelo, D. St; Holmes, G.; Heidel, K. A process for capturing CO2 from the atmosphere.
Joule 2018, 2, 1573−1594.
6. Wurzbacher, J. A.; Gebald, C.; Piatkowski, N.; Steinfeld, A. Concurrent separation of CO2 and H2O from
air by a temperature- vacuum swing adsorption/desorption cycle. Environ. Sci. Technol. 2012, 46,
9191−9198.
7. Dioxide Materials. Dioxide Materials: Electrolyzer to convert carbon dioxide to carbon monoxide.
8. Dioxide Materials. Dioxide Materials: Electrolyzer to transform carbon dioxide into formic acid.
9. Flanders, N.; Kuhl, K.; Cave, E. Opus 12 − Recycling carbon dioxide back into fuels and chemicals; 2016.
10. Haas, T.; Krause, R.; Weber, R.; Demler, M.; Schmid, G. Technical photosynthesis involving CO2
electrolysis and fermentation. Nat. Catal. 2018, 1, 32−39.
11. Electrochaea. Applications of Electrochaea’s BioCat biomethanation technology; 2018,
12. LanzaTech. Technical background on the LanzaTech Process; 2013, http://www.arpaesummit.com/paperclip/exhibitor_docs/ 14AE/LanzaTech_Inc._131.pdf.
13. Kauw, M.; Benders, R. M. J.; Visser, C. Green methanol from hydrogen and carbon dioxide using
geothermal energy and/or hydropower in Iceland or excess renewable electricity in Germany. Energy 2015,
90, 208−217.
14. Li, B.; Duan, Y.; Luebke, D.; Morreale, B. Advances in CO2 capture technology: A patent review. Appl.
Energy 2013, 102, 1439− 1447.

127
15. Matter, J. M.; et al. Rapid carbon mineralization for permanent disposal of anthropogenic carbon dioxide
emissions. Science (Washington, DC, U. S.) 2016, 352, 1312.
16. Ragnheidardottir, E.; Sigurdardottir, H.; Kristjansdottir, H.; Harvey, W. Opportunities and challenges for
CarbFix: An evaluation of capacities and costs for the pilot scale mineralization sequestration project at
Hellisheidi, Iceland and beyond. Int. J. Greenhouse Gas Control 2011, 5, 1065−1072.
17. Monkman, S.; MacDonal, M. Ready mixed technology case study CO2 utilization in concrete mix design
optimization; 2016, http://info. carboncure.com/hubfs/Downloads/
CarbonCure%20Ready%20Mixed%20Technology%20Case%20Study. pdf?hsCtaTracking=3acd2ed8631d-4960-a516- eb14990ee747%7Cf4c138bf-0589-4751-a2c8-576a17dd4453.
18. Weng, L. C.; Bell, A. T.; Weber, A. Z. Towards membrane- electrode assembly systems for CO2 reduction:
A modeling study. Energy Environ. Sci. 2019, 12, 1950−1968.
19. Li, Y. C.; et al. CO2 electroreduction from carbonate electrolyte. ACS Energy Lett. 2019, 4, 1427−1431.
20. Li, T.; et al. Electrolytic conversion of bicarbonate into CO in a flow cell. Joule 2019, 3, 1487−1497.
21. Energy Technology Systems Analysis Programme. Syngas production from coal; 2010,
22. Baltrusaitis, J.; Luyben, W. L. Methane conversion to syngas for gas-to-liquids (GTL): is sustainable CO2
reuse via dry methane reforming (DMR) cost competitive with SMR and ATR processes? ACS Sustainable
Chem. Eng. 2015, 3, 2100−2111.
23. Tountas, A. A. Towards solar methanol: past, present, and future. Adv. Sci. 2019, 6, 1970048.
24. Samavati, M.; Santarelli, M.; Martin, A.; Nemanova, V. Production of synthetic Fischer−Tropsch diesel
from renewables: thermoeconomic and environmental analysis. Energy Fuels 2018, 32, 1744−1753.
25. Chatterjee, D.; Jaiswal, N.; Banerjee, P. Electrochemical conversion of bicarbonate to formate mediated by
4−
the complex Ru III (edta) (edta = ethylenediaminetetraacetate). Eur. J. Inorg. Chem. 2014, 2014,
5856−5859.
26. Stalder, C. J.; Chao, S.; Wrighton, M. S. Electrochemical reduction of aqueous bicarbonate to formate with
high current efficiency near the thermodynamic potential at chemically derivatized electrodes. J. Am. Chem.
Soc. 1984, 106, 3673−3675.
27. Kortlever, R.; Peters, I.; Koper, S.; Koper, M. T. M. Electrochemical CO2 Reduction to Formic Acid at
Low Over- potential and with High Faradaic Efficiency on Carbon-Supported Bimetallic Pd-Pt
Nanoparticles. ACS Catal. 2015, 5, 3916−3923.
28. Kortlever, R.; Balemans, C.; Kwon, Y.; Koper, M. T. M. Electrochemical CO2 reduction to formic acid on
a Pd-based formic acid oxidation catalyst. Catal. Today 2015, 244, 58−62.
29. Sawant, A. Global Formic Acid Market; Market Research Future, 2018.
30. De, S.; Gevers, L.; Emwas, A. H.; Gascon, J. Conversion of Formic Acid into Methanol Using a
Bipyridine-Functionalized Molecular Heterogeneous Catalyst. ACS Sustainable Chem. Eng. 2019, 7,
3933−3939.
31. Miller, A. J. M.; Heinekey, D. M.; Mayer, J. M.; Goldberg, K. I. Catalytic disproportionation of formic acid
to generate methanol. Angew. Chem., Int. Ed. 2013, 52, 3981−3984.
32. Savourey, S.; et al. Efficient disproportionation of formic acid to methanol using molecular ruthenium
catalysts. Angew. Chem., Int. Ed. 2014, 53, 10466−10470.
33. Sasayama, A. F.; Moore, C. E.; Kubiak, C. P. Electronic effects on the catalytic disproportionation of
formic acid to methanol by [Cp*IrIII(R-bpy)Cl]Cl complexes. Dalt. Trans. 2016, 45, 2436− 2439.
34. Sordakis, K.; et al. Aqueous phase homogeneous formic acid disproportionation into methanol. Green
Chem. 2017, 19, 2371−2378.
35. Tsurusaki, A.; et al. Investigation of Hydrogenation of Formic Acid to Methanol using H2 or Formic Acid
as a Hydrogen Source. ACS Catal. 2017, 7, 1123−1131.

128
36. Sordakis, K.; et al. Carbon Dioxide to Methanol: The Aqueous Catalytic Way at Room Temperature. Chem.
Eur. J. 2016, 22, 15605−15608.
37. Gleizer, S.; et al. Conversion of escherichia coli to generate all biomass carbon from CO2. Cell 2019, 179,
1255−1263.e12.
38. Rice, C.; Ha, S.; Masel, R. I.; Wieckowski, A. Catalysts for direct formic acid fuel cells. J. Power Sources 2003,
115, 229−235.
39. Yu, X.; Pickup, P. G. Recent advances in direct formic acid fuel cells (DFAFC). J. Power Sources 2008, 182,
124−132.
40. Rice, C.; et al. Direct formic acid fuel cells. J. Power Sources 2002, 111, 83−89.
41. Ji, X.; et al. Nanocrystalline intermetallics on mesoporous carbon for direct formic acid fuel cell anodes.
Nat. Chem. 2010, 2, 286−293.
42. Muller, K.; Brooks, K.; Autrey, T. Hydrogen storage in formic acid: a comparison of process options.
Energy Fuels 2017, 31, 12603− 12611.
43. Bavykina, A. V.; Goesten, M. G.; Kapteijn, F.; Makkee, M.; Gascon, J. Efficient production of hydrogen
from formic acid using a covalent triazine framework supported molecular catalyst. ChemSusChem 2015, 8,
809−812.
44. Joo,́ F. Breakthroughs in Hydrogen Storage — Formic Acid as a Sustainable Storage Material for
Hydrogen. ChemSusChem 2008, 1, 805−808.
45. Bi, Q.; et al. An aqueous rechargeable formate-based hydrogen battery driven by heterogeneous Pd
catalysis. Angew. Chem., Int. Ed. 2014, 53, 13583.

129
Chapter 8

COMPARATIVE TECHNO-ECONOMIC ANALYSIS OF RENEWABLE
GENERATION OF METHANE USING SUNLIGHT, WATER, AND CARBON
DIOXIDE
8.1

Introduction
Thirty-one percent of the primary energy consumed in the United States comes from the burning of

natural gas, 70-90% of which is comprised of methane (CH4).1 Natural gas is recovered from onshore and
offshore natural gas and oil wells, and from coal beds. Currently, the United States has enough supply of dry
natural gas to sustain current consumption for 92 years.2 Meanwhile, California consumes 2.14 MMcf (43.2
million ton) of natural gas per year,2 over a quarter of which is used to generate electric power3 and which
provides approximately 40% of electrical energy in the state.4 Since an extensive nationwide storage and
distribution network already exists for natural gas, the development of renewable methane could enable rapid
and widespread distribution of zero-carbon energy services. Thus for California to meet its renewable portfolio
standard, e.g., 60% renewable energy for electricity generation by 2030,5 and to conserve a limited resource, it is
imperative to assess how to develop and deploy technologies for renewable generation of CH4 in the next few
decades.
While an increasing number of power to gas (PtG) projects for CH4 generation or H2 generation are
being planned globally,6 the largest source of renewable CH4 currently being produced in the United States
comes from anaerobic digesters that convert cow manure into natural gas. There are currently over 250 such
systems in operation with more under construction.7 In addition to generating CH4 renewably, these anaerobic
digesters also prevent the release of CH4 – one of the most potent greenhouse gases – into the atmosphere.
However, co-location of dairy farms and anaerobic digesters alone will not yield enough renewable methane to
replace the current energy demand met by natural gas.8 For example, the residential natural gas demand in

130
California is ~25,000 ton/day,4 while an average dairy farm can only produce ~5 ton/day of natural gas from
anaerobic digesters.7,9 If all potential biogas in California was realized it could power 180,000 homes or 435,000
vehicles, which represents approximately 1.2% of all homes or 3% of all registered vehicles in the state.8 While
these anaerobic digesters co-located with dairy farms may seem to have small production capacity, they are
among the largest sources of renewable CH4 generation in the world.6 Therefore, it is important to evaluate
other more readily scalable technology pathways for renewable generation of CH4.

Figure 8.1: Schematic of various pathways to capture CO2, generate H2, and generate CH4 from sunlight, H2O, and sunlight.
Here we outline multiple technology routes for renewable generation of CH4 from sunlight, water (H2O),
and carbon dioxide (CO2) (Figure 8.1 and Figure 8.2). We evaluate the technology readiness level (TRL), the
demonstrated scale of these candidate technologies, the cost for CH4 generation, as well as the cost required to
provide the necessary feedstocks – H2O, CO2, and H2. We investigate and compare four main CO2 methanation
pathways: thermochemical (via the Sabatier reaction), biochemical, photo-electrochemical, and electrochemical
(Figure 8.2). By applying a standard discounted cash flow method to each technology, we assess the current
status, future opportunities and compare different technology pathways side-by-side.

131

Figure 8.2: Schematic representation of various technology pathways for sustainable generation of methane from sunlight, water,
and carbon dioxide.

The detailed assumptions for the TEA of technologies evaluated in this study are included in the
Supporting Information and the database files used for arriving at detailed cost values are also included in the
Supporting Information. The baseline CH4 production was assumed to be at a scale of 30 kton CH4 per year
(1,500 Mcf per year), or 81 tons CH4 per day (4 Mcf per day). We calculate the CO2 capture, H2 production and
water generation rates needed to match this production rate, e.g., H2 production rate of 40 ton/day, CO2 capture
rate of 245 ton/day and water generation rate of 365 ton/day. No carbon credits were accounted for in this

132
study. The electricity price for all current systems was assumed to be $49/MWh based on the 2018 data from
solar utility PV in California.10 The TRL is evaluated from 1 to 9, where TRL 1-2 corresponds to the observation
of basic principles in the academic development, TRL 3-4 to proof-of-concept development at lab scale, TRL
5-6 to process development and system integration from lab to pre-pilot scale, TRL 7-8 to optimization and pre-

CH4 Generation

H2 Generation

Water
Source

CO2
Capture

commercialization scale and TRL 8-9 to commercial operation at scale.11
Method

Current Cost $/ton

TRL

Point Source

60-7081

8-9

Current demonstrated
Scale
~20,000 ton/day19

Air

277 (115 in future)

7-8

~1 ton/day14

Ocean Water
Utility
Desalination
Membrane
Condensation
Low Temperature
Electrolysis (LTE)
Photoelectrochemical
(PEC)
High Temperature
Electrolysis (HTE)
Solar Thermo-chemical
(STCH)
Electrochemical
Photoelectrochemical

416 (118 in future)
<0.000332
1.6235
150
339

3-4
8-9
8-9
8-9
5-6

~1 kg/day31
~1.6 million ton/day33
~0.19 million ton/day35
~120 ton/day41
~6 kg/day46

3518

8-9

~2.6 ton/day6,51

5294

5-6

<1 kg/day58

3956

7-8

~1 ton/day56

3706
(2,500-10,000 with feedstock)
(1,500-17,000 with feedstock)
199 without feedstock,
2,797 with feedstock
189 without feedstock,
2,830 with feedstock

5-6
3-4
3-4

<1 kg/day61
<1 kg/day72
<1 kg/day75

7-8

~5 ton/day69

7-8

~4 ton/day65

Thermochemical
Biochemical

Table 8.1: Summary of cost, TRLs, and demonstrated scale of different technological pathways for renewable generation of methane.
The cost of methane generation in the thermochemical and biochemical routes assumed water from utility, CO2 from direct air capture,
and H2 from LTE as the feedstock.
The technological pathways considered in this study also included the traditional PtG routes, in which
H2 is produced via electrolysis followed by CH4 synthesis.12 In addition, while many reports 13–16 have focused

133
on individual components of a renewable methane system, this work used the state-of-the-art performance
metrics from specific technologies, including four advanced water-splitting technologies, CO2 capture from air
and oceanwater, (photo)electrochemical CO2 reduction into CH4 and thermo-chemical and biochemical
methanation from H2 and CO2, so that different technological pathways can be directly compared. Table 8.1
summarized the cost estimate from this study, TRLs and demonstrated scale for each technology, derived from
literature reports of renewable generation of CH4.
8.2

Carbon dioxide capture
Carbon dioxide can be captured from point sources of emission, atmospheric air, or oceanwater. The

location of capture often dictates where CH4 generation can occur. CO2 capture from point sources, such as oil
refineries and the cement industry, is primarily based on chemical absorption and desorption of flue gas with
amine solutions.17,18 Point source CO2 capture has been demonstrated at a rate of 20,000 ton/day by the Century
NG plant in Texas in 2010.19 For a typical coal fired power plant, a CO2 capture rate of ~5,000 ton/day provides
a CO2 emission reduction of ~10%.20 Point source CO2 capture utilizes relatively mature technology and has
been validated by large industrial scale demonstrations at multiple locations, at an average TRL of 8-9. However,
point source CO2 capture is not compatible for a negative CO2 emission future. A typical carbon capture system
for a coal fired power plant reduces the plant energy efficiency, consuming 16% of the generated energy from
the plant.21 Since point source capture has been deployed at large scale in multiple geographical locations, the
estimated cost has converged to a narrow cost range of ~$60-70/ton of CO2 from reported literature.22,23
Carbon dioxide can also be captured from the environment in a dilute form from either atmospheric air
or from ocean water. Direct air capture (DAC) has been demonstrated at an early commercial scale.14,24 The
operating principle of DAC from Carbon Engineering includes two sequential loops. In the first loop, CO2 is
captured from the atmosphere using capturing solvents, such as aqueous alkaline solutions, to form aqueous
carbonate solutions. The second loop precipitates the carbonate using Ca2+, regenerates the alkaline solution,
and releases CO2 by calcination.14 The largest system that is currently built based on this technology is capable

134
of capturing roughly 1 ton CO2/day, which corresponds to a TRL of 7-8.14 Assuming 803 kWh/ton CO2 to
power calciner, compressor, pumps, etc. with the largest single line item being the air contactor, our TEA model
predicts a cost of $277/ton of CO2 at a plant capacity of 240 tons/day (Appendix B: Table B2). This is in good
agreement with reported values.14 Assuming an energy input of 555 kWh/ton CO2, , and cost of electricity of
$10/MWh, we further estimate that the future cost for DAC can be reduced to $115/ton, upon scaling to a
plant capacity of 2,400 ton/day (Appendix B: Table B3).
Capturing CO2 from oceanwater is an attractive alternative to DAC because the concentration of CO2
is 140 times higher in oceanwater than it is in the air.16 The operating principle for oceanwater capture is to shift
the CO2/bicarbonate equilibrium toward dissolved CO2 by acidifying oceanwater, achieved via a process which
lowers pH of oceanwater, such as electrodialysis. The acidified stream is then passed through a liquid–gas
membrane contactor, which captures the gaseous CO2 from the dissolved CO2 in the aqueous stream. However,
oceanwater intake, pre-treatment, and pumping from an offshore site to an onshore capture plant accounts for
a major portion of the cost of capturing CO2 from oceanwater.16 By either co-locating, with a desalination or
electric power plant, or creating an onshore floating system, this cost can be significantly reduced.25,26
Electrochemical systems for the extraction of CO2 from oceanwater have been reported previously.16,27–30
Assuming current density of 100 mA/cm2, voltage of 1.2V, we estimate the cost to be $416/ton of CO2 for a
floating CO2 capture from ocean water system via electrodialysis with shallow intake and a plant capacity of 240
ton/day (Appendix B: Table B4). In the future, by assuming an increased system scale with throughput of 2,700
ton/day, electrodialyzer current density of 1 A/cm2, and an electrodialyzer voltage of 1.6V, the cost of the system
can be reduced to ~ $118/ton CO2 (Appendix B: Table B5). Our calculated current cost of $402/ton of CO2
for a floating ocean capture system is similar to the value reported by Eisaman et. al.16 for a system that is colocated with a water desalination plant. Differences in assumptions made for electricity price, electrodialysis
performance, pre-treatment processes, and other economic assumptions account for the differences. While this
technology appears to be promising, it is in its nascent phase, corresponding to a TRL of 5-6 for oceanwater

135
CO2 capture given that the largest system realized to date operates with a throughput of ~1 kg/day in a lab
environment.31
Encouragingly, the costs for CO2 capture from either air or oceanwater via the most compelling
processes may be able to reach <$100/ton in the future. As CO2 capture from dilute sources reaches this cost
range, large scale CO2 utilization or storage will not be limited by the physical location of point sources of CO2
emission. This will mark the transition to an infrastructure that can effectively offset CO2 emissions from
sources that are very difficult to address with a point-of-emission capture approach, such as consumer appliances
and vehicles.
8.3

Water generation
Regardless of the methanation technology used, water is required as a feedstock for the hydrogen content

in CH4. A small total amount of water is needed and its cost is low, as compared to other steps in the process;
thus the choice of water generation method is largely dependent on the system location. Water can either be
purchased at a utility scale, from a desalination plant, captured from the air with a membrane, or condensed out
of the air by engineering desired thermal properties of the material.
Utility scale water is the cheapest option due to its scale and government subsidy. The price of ground
water depends on the pumping depth, energy source, cost of energy, and the amount of water available. Prices
in 2010, according to the Organization for Economic Cooperation and Development (OECD), range from
$0.000195/ton in California to $0.00023/ton in Arizona.32 For our analysis, we used the water cost in California.
One of the largest water suppliers is the Los Angeles Department of Water and Power (LADWP), which
supplied 632 billion liters of water in 2014.33 Another water generation option is desalination, in which ocean
water is processed with a pretreatment filter to remove large particles, and is then forced under high pressure
through a membrane to perform reverse osmosis. The fresh water is then treated for drinking and the brine is
discharged to the ocean.34 Large scale desalination has only been demonstrated at a tenth the scale of utility
water sources. For example, in 2015 San Diego deployed a desalination plant that produces 68.9 billion liters of

136
fresh water per year at a cost of $1.62/ton.35 While more expensive, water supply from desalination is preferable
in locations near the ocean with minimal rainfall. Both of these technologies are assigned a TRL of 8-9 due to
their large scale and multiple plant locations.
Alternative routes for water generation involve extraction from atmospheric air. While these methods
are more expensive and have lower TRLs, they remove location restrictions for CH4 production systems. One
way to extract water from the air is via a membrane or mesh, which provides a surface upon which water vapor
in the atmosphere can condense. These droplets are harvested into a collection area under the influence of
gravity drop.36 This has been shown in numerous studies37–39 and demonstrated on a larger scale at 12 ton/day.40
We assign a TRL of 8-9 for the membrane water capturing technology.41 The cost of water obtained from
membrane capture is estimated to be $60 /ton at a plant capacity of 365 ton/day (Appendix B: Table B6).
Another way to remove water from the air is by using radiatively cooled surfaces to condense water out of the
air.42–45 In these systems, materials are designed to maximize infrared emission properties and allow the surface
to be cooler than ambient temperature. This change in temperature between surface and ambient promotes the
condensation of atmospheric water on the surface which can then be harvested.36 Currently, condensation via
radiative cooling is the most expensive option among those we considered, due to the high capital expense for
purchase of the materials to capture water at a relevant scale. The largest demonstration to date is by OPUR
(International Organization for the Utilization of Dew), which has shown a system that can generate 0.006
ton/day, which we assign a TRL of 5-6.45,46 We estimate the cost of water using radiative cooling to be $339/ton
at a scale of 365 ton/day (Appendix B: Table B7).
Despite the significant cost differences between various water generation strategies, the cost of water
remains lower than other costs of a renewable methanation system. For low temperature electrolysis if water
produced via membrane air capture is used rather than utility water the cost of hydrogen increases from $3.48/kg
to $4.53/kg. It is notable that no matter where a methane system is deployed, near a utility water source, ocean,
or desert that generating water is likely not to be a limiting factor.

137
8.4

Hydrogen generation
CO2 and water are the main raw feedstocks for direct methanation pathways (electrochemical and

photoelectrochemical); however, indirect methanation pathways (thermochemical and bio-methanation) (Figure
8.2) rely on reaction between CO2 and H2. It is therefore appropriate to assess the technology options for
renewable H2 generation from water. The renewable generation of H2 is also the most critical step in all the PtG
studies.6,12,47–49 We focus on four H2 generation technologies, including low temperature electrolysis (LTE),
photoelectrochemical (PEC), high temperature electrolysis (HTE), and solar thermo-chemical (STCH). We used
H2A analyses guidelines and applied the DOE financial and operational assumptions that are adjusted to our
case studies to calculate the cost of hydrogen generated using these technologies.50 All technologies are evaluated
at a design capacity of 40-45 ton H2/day. This average production will therefore maintain the 81 tons CH4 per
day (4 Mcf CH4 per day) production required by our baseline assumptions.
The largest scale demonstrated for LTE thus far is 10 MW, corresponding to ~2.6 ton/day generation
rate.51 MW scale LTE systems were also deployed world-wide, at an average TRL of 8-9.52 In LTE, H2 is
produced at the cathode and O2 at the anode electrochemically under a voltage bias. The cathode and anode are
separated by a membrane separator.53 Three types of LTE systems have been developed including alkaline water
electrolysis, proton exchange membrane based water electrolysis and hydroxide exchange membrane based
water electrolysis.54 State of the art proton exchange membrane (PEM) cells operate at ~2 A cm-2 and ~2 V with
a stack level efficiency of 55 kWh/kg of H2.54 We assume on operating voltage of 1.9 V/cell, current density of
2 A/cm2, stack life of 7 years.15,50 From this we estimate the current cost of a PEM system at a design capacity
of 40 ton/day is $3,518/ton (Appendix B: Table B8). The values that we calculated are similar to those calculated
previously, but significantly lower than those calculated by the DOE case study.15 The discrepancy was largely
due to the higher cost of electricity assumed by the DOE of $87/MWh and the higher after-tax real IRR of
8%.50,53 A sensitivity analysis (Appendix B: Figure B1) for the impact of electricity cost, energy efficiency, capital

138
expenditures and after-tax real IRR on the cost of hydrogen indicates that presentably the electricity cost was
the largest levers among those variables.
High temperature electrolysis (HTE) is another H2 production method using electricity. The operating
principle for HTE is very similar to LTE. High temperature electrolysis cells include a cathode for water
reduction, an anode for oxygen generation and a solid ceramic material, which is used as the electrolyte to
selectively conduct oxygen ions (O2-) at ~700°–800°C.55 It has been demonstrated at a scale of 2.6 MW,
corresponding to ~1 ton H2/day.56 Assuming an energy usage of 51 kWh/kg H2,52 we estimate the cost of HTE
system to be $3956 /ton H2 (Appendix B: Table B9), which was similarly less than the value calculated by the
DOE, $4660/ton.50 Both LTE and HTE use electricity for H2 generation, as a result, the electricity price has a
large influence on the cost of H2 from both technologies. For instance, at a scale of 40 ton H2/day, LTE system
will produce H2 at $2410/ton with an electricity price of $30/MWh, whereas at an electricity price of $60/MWh,
the cost of H2 is $4160/ton.
Photoelectrochemical (PEC) and solar thermochemical processes (STCH) produce H2 from sunlight
and water. Photoelectrochemical water-splitting cells integrate multiple functional materials and couple water
oxidation and hydrogen evolution reactions to produce molecular hydrogen and oxygen. Key PEC processes
include light absorption, photo-generated carrier transport, electrocatalysis, ionic transport and product
separation.57 Photoelectrochemical devices operate at much lower current densities, typically in the range of ~10100 mA cm-2, relative to LTE or HTE, since the production rate is matched to the solar flux. State of the art
PEC devices have exhibited a solar to hydrogen conversion efficiency of 19.3%.58 Currently PEC devices have
only been demonstrated at a laboratory scale <1 kg/day H2, giving it a TRL of 5-6.13 Assuming a solar to hydrogen
(STH) efficiency of 10%, photovoltaic (PV) efficiency of 19.1%, cost per unit area of $161/m2 (which includes
the cost of PV cells, catalyst, membrane, chassis, water processing, gas processing, power electronics, and control
system),13 the current cost of H2 from PEC is estimated to be $5294/ton (Appendix B: Table B10). However,
the projected price drops of photovoltaic materials, dramatic improvements in membrane costs, and increases

139
in solar to fuel efficiency are projected to lead to a significantly reduced cost for H2. For example, with an STH
efficiency of 20%, PV efficiency of 25%, cost per unit area of $119/m2 (which includes the cost of PV cells,
catalyst, membrane, chassis, water processing, gas processing, power electronics, and control system), we
estimate the cost of PEC H2 can reach $1775/ton in future (Appendix B: Figure B2, Appendix B: Table B11).
The estimated value is lower than values calculated by Shaner, et al.,13 due to a higher assumed solar capacity
factor of 28.4% (for California) in our case study, as opposed to 20%, and more up to date value of $0.37/W59
for the cost/Watt of photovoltaic panels (See Supporting Information).
Solar thermochemical (STCH) cycles use the heat from the sunlight to produce hydrogen and oxygen
from water. STCH uses two-step redox active metal oxide thermochemical cycles to produce H2 and O2
sequentially in two different chemical reactions.60 STCH has been demonstrated at ~1 kg/day, giving it a TRL
of 5-6.61 Using an economic model adapted from the U.S. DOE H2A analysis,50,62 assuming an STH efficiency
of 20%, plant capacity factor of 90%, we estimated H2 cost from STCH to be $3706/ton for a system with a
design capacity of 45 ton H2/day (Appendix B: Table B12).
H2 generation is the largest cost driver for indirect renewable CH4 generation and is expected to play a
critical role in a broader setting in future energy systems. However, it is important to realize that presently the
largest demonstrated H2 generation project even with the highest TRL technology, e.g., low temperature
electrolysis, is limited at < 3 ton of H2 per day. Note that a single digester co-located with a dairy farm produces
on average ~5 ton of CH4 per day, which translates to ~2.5 ton of H2 per day required based on the CO2
methanation reaction. Hence, it is not surprising to note that large electrolysis projects for renewable generation
of H2, such as a 156 ton H2/day system in France,63 have been planned in the near future.6 However, converting
H2 to CH4 has its own advantages. CH4 has ~3.5 times higher storage capacity that H2, and H2 is significantly
more difficult to store since it is corrosive and leads to embrittlement of container materials.64 Considering this
last point, many costly modifications and component replacements would be needed in the legacy gas piping,
storage, and distribution infrastructure to make it compatible with H2 distribution rather than methane

140
distribution. These infrastructure utilization considerations represent a strong argument in favor of renewable
methane as an alternative to H2, as a gas energy carrier for widespread distribution.
8.5

Methane generation
Having surveyed pathways to generate the raw feedstocks needed for renewable methane synthesis, we

now analyze and compare different methanation pathways. We separate these into two main categories: i)
indirect CO2 to CH4 conversion via thermochemical and bio-methanation methods, and ii) direct CO2 to
methane conversion via electrochemical and PEC methods.
The two indirect methane conversion methods that we focus on are thermochemical methanation via
the Sabatier reaction and biochemical methanation. Biochemical methanation has been demonstrated at 5MW
scale, or ~4.3 ton/day from Electrochaea.65 We assign a TRL of 7-8 for this technology. The basic technology
relies on anaerobic microorganisms called methanogenic archaea that are able under certain conditions to
produce CH4 from H2 and CO2 with high selectivity.66 These organisms exist naturally in the environment and
have been selectively evolved for higher selectivity in these reactors. In this process, the archaea are heated up
to 60 oC and pressurized to ~10 bar and then fed CO2 and H2.65 The organisms can then self-sustain the heat
and highly selectively convert the CO2 and H2 to CH4. Assuming a 99% efficient biological methanation reactor65,
biochemically produced CH4 is estimated to have a cost of $189/ton without the feedstock cost, and $2830/ton
(Appendix B: Table B13), assuming that the H2 is generated via LTE and the CO2 is captured from air.67
We focus on the Sabatier reaction as a method for thermochemical methanation. Similar to the
biomethanation route, CO2 is reacted with H2; however instead of using a micro-organism the reactor is heated
to ~350 °C in the presence of a catalyst (i.e. Ni). The CO2 and H2 then react exothermically to produce CH4.68
The largest Sabatier reactor built thus far is a 6MW reactor by Audi at their Audi e-gas facility in germany.69 We
estimate a TRL of 7-8 for this technology. Assuming a conversion efficiency of 93%, we calculate the cost of
thermochemical methanation to be $193/ton CH4 without the feedstock cost, and $2791/ton CH4 (Appendix
B: Table B14), using H2 generated via LTE and CO2 captured from air.

141

Figure 8.3: The cost breakdown of the green methane from thermochemical and biochemical processes. The feedstock of the
thermochemical and biochemical process assumed water from utility, CO2 from direct air capture, and H2 from LTE with an
electricity price of $49/MWh.

Figure 8.3 shows cost breakdown for the indirect methanation methods, assuming H2 generated via
PEM electrolysis, CO2 captured from the atmosphere and H2O from a utility source. The cost for CO2 capture
and the methanation process (both thermochemical and biochemical) are small, and it is clear that the cost of
renewable H2 generation is the largest cost component for the indirect methanation pathways. As shown in
Figure 8.3 and Appendix B: Figure B1, the cost of electricity remains the largest cost and largest lever for
renewable H2 generation via LTE, which is consistent with recent DOE reports.62

142

b.

$2,800 /ton CH4

Cost of Methane ($/ton)

$2,800 /ton CH4

Cost of Methane ($/ton)

a.

Figure
8.4: Cost
methane from (a)
photoelectrochemical
(PEC) and (b) electrochemical
methanation
processes
as athe
function
Figure
XX:ofParameter
sweep
of direct methanation
methods.
(a) shows
how
costof

of PEC methanation depends on STF and the cost per area of PEC material. (b) shows
how the cost of electrochemical methanation depends on the energy efficiency of the
the cost
per areand
of PEC
material.density.
(b) The costThe
of electrochemical
methanation
as a plots
function shows
of the energy
efficiencythe
of the
device is
and
device
current
green region
in both
where
price
equal to or less than thermochemical or biochemical methanation. For all systems
the operating current density. The green region in both plots shows where the cost is equal to or less than thermochemical or biochemical
compared the cost of CO2 is assumed to bye $278/ton and the electricity price is
$49/MWh.
methanation.
For all systems compared the cost of CO2 is assumed to be $278/ton and the electricity price is $49/MWh.
key performance metrics in those technologies. (a) The cost of PEC methanation as a function of the STF conversion efficiency and

For the direct methane conversion methods, we first consider an electrochemical system powered by
grid electricity at a high-capacity factor, similar to LTE H2 electrolysis. The main differences between an H2
electrolysis and a CO2 electrochemical system comes from the consideration of charge transfer in electrocatalysis,
which requires 8 electrons per CH4 molecule from CO2 as opposed to the 2 electrons needed for generation of
an H2 molecule. Currently, the multi-electron and proton reaction still faces significant challenges in selectivity,
activity and durability.70,71 One of the highest performing electrolysis systems developed, exhibited a Faradaic
Efficiency (FE) of 85% for CH4 generation with an overpotential of 2.8 V at ~25 mA/cm2.72 Electrochemical
CO2 conversion devices have also exhibited high operating current densities up to 700 mA/cm2 in other
reduction products such as CO and ethylene.73,74 The cost for direct electrochemical methanation process as a
function of the operating current density and the energy efficiency of cell was illustrated in Figure 8.4a. A range
of operating current density from 10 mA/cm2 to 5 A/cm2 and a range of Faradaic efficiency from 50 to 100%

143
were considered for the direct electrochemical CH4 generation. Note that the overall cell efficiency is a
combination of the Faradaic efficiency of the reaction and the operating cell voltage. At an operating current
density of 100 mA/cm2, and an energy efficiency of 15%, the cost of CH4 was estimated to be $10,700/ton CH4
(Appendix B: Table B15), assuming $278/ton CO2 captured from the atmosphere, and a plant size of 81
ton/day. If we assume a more optimistic device performance of 5 A/cm2, and an energy efficiency of 56%, the
cost of CH4 was estimated to be $2,420/ton CH4 (Appendix B: Table B15), assuming $278/ton CO2 captured
from the atmosphere, and a plant size of 81 ton/day. Based on the demonstrated current density and energy
efficiency in the literature, the electrochemical methanation is not competitive with indirect methanation
techniques. It would require very significant advancement in materials and device development to be cost
competitive to the indirect methanation processes.
Photoelectrochemical (PEC) methanation operates using the similar mechanisms as PEC H2 generation
except that the electrons are reducing CO2 instead of H2O. Both PEC H2 and PEC methanation use water as
the proton source. Photoelectrochemical (PEC) methanation also faces similar challenges as the electrochemical
methanation, specifically the fact that producing methane requires 8 electrons, and since PEC devices are limited
by the solar flux, this limits the rate of methanation. A PEC methanation device has been realized with a solar
to fuel efficiency (STF) of 0.1%.75 While PEC methanation has significant challenges in the activity and
selectivity, other PEC CO2 reduction devices, such as CO2 reduction to CO or formate, have reached STF
conversion efficiency of >10%.76 The cost for direct PEC methanation pathway as a function of the cost per
area and the STF conversion efficiency was illustrated in Figure 8.4b. A range of STF conversion efficiency from
4 to 18% was considered for direct PEC CH4 generation. At an STF of 4% for the PEC methanation device,
and a $200/m2 constructed in a PEC type 3 configuration,13 we estimate the cost of CH4 to be to be $16,930/ton
(Appendix B: Table B16). If we assume more optimistic device metrics of an STF of 18% for the PEC
methanation device, and a $10/m2 constructed in a PEC type 3 configuration,13 we estimate the cost of CH4 to
be to be $1,500/ton (Appendix B: Table B16). As shown in Figure 8.3b, for PEC methanation to be competitive

144
with indirect methanation, the cost of the system per square meter must be significantly reduced as well as
improvements in efficiency.
a.

Thermochemical
Methanation
$2,800 /ton CH4

Cost of CH4
($/ton)

b.

Biochemical
Methanation
$2,800 /ton CH4

Cost of CH4
($/ton)

c.

Biochemical Thermochemical
Temperature (°C)
63
350
Pressure (ba)
10
10
Capacity Factor
98%
98%
10 MW Reactor
480
15
footprint (m2)
Contaminant
high
H2S sensitive
Tolerance
tolerance
Largest similar plant
0.25
21.6
(million kg/day)

Figure 8.5: Cost of the (a) thermochemical and (b) biochemical methanation processes as a function of H2 cost and CO2 cost. (c)
A side-by-side comparison between the biochemical methanation process and the thermochemical methanation process.

Figures 8.5a-b show how the cost of thermochemical and biochemical methanation is affected by the
cost of the feedstocks. The cost of both approaches is very close, due to the similarity of their CapEx’s and that
the CapEx makes up 75% of the methanation cost. The CapEx’s are similar because much of the equipment
required for both systems is the same, items such as compressors, reactor, pumps, piping, etc.

The

heterogeneous catalysts in the thermochemical methanation and the anaerobic microorganisms in the
biochemical methanation were not the main cost driver for methanation. Other factors to consider are listed in
Figurer 8.5c. One advantage of biochemical methanation is its high tolerance for contaminants, whereas catalysts
for the Sabatier reaction are highly sensitive to H2S.48,77 This advantage is most important when the CO2 is being
captured from point sources, however CO2 captured from the air or oceanwater environment is likely to result
in a very pure CO2 stream, making this difference less important. Another advantage of biochemical
methanation is the lower operating temperature, making it possibly more suitable for small scale reactors.65,78
However, thermochemical methanation may be favorable when considering scaling of methanation to a large
capacity. First, the required areal footprint of the reactor is significantly smaller.49,67,79 Second, examination of
scales of similar processes via bio- vs. thermochemical processes such as Fischer Tropsch via thermochemical

145
or bio ethanol for biochemical methanation, there is nearly a 100x difference in scale at which these processes
have been demonstrated, suggesting it may be significantly easier to scale up thermochemical methanation than
biochemical methanation.
8.6

Conclusion
In summary, thermochemical or biochemical methane generation using CO2 captured from point

sources and H2 produced from low temperature electrolysis powered by renewables turned out to be the most
cost competitive pathway in the short term. The cost of renewable H2 is found to be the dominant cost
component of renewable methane synthesized by indirect methanation. We also found that the cost of CO2
from dilute sources (air or oceanwater) is likely to be competitive with CO2 from point sources as the technology
advances, and the cost of CO2 will not be a cost driver for CH4 generation. The largest demonstrated scale for
direct CO2 capture from air (~1 ton of CO2 per day), the renewable H2 generation via low temperature
electrolysis (~2.6 ton of H2 per day), the thermo-chemical methanation (~5 ton of CH4 per day) and the
biochemical methanation (~4 ton of CH4 per day) are all very small, and are dwarfed by a single anaerobic
digester co-located with a dairy farm. As cost of the renewable electricity continued to decrease, at an electricity
price of $10/MWh, we estimated that an overall optimistic cost of $983/ton of CH4 in the future, which is then
cost competitive to the market CH4 price in certain regions of the world,80 can be achieved. In the long term,
significant improvements of key performance metrics in electrochemical and photoelectrochemical methanation
can provide unique alternatives to the short-term pathway winners with more energy resilience and ultimately
achieve CH4 production cost of < $1000/ton.

BIBLIOGRAPHY CHAPTER 8
1.
2.

U.S. Energy Information Administration. August 2020, Monthly Energy Review. United States Energy
Information Administration Monthly Review 0035, (2020).
United States Energy Information Administration. EIA Annual Energy Outlook 2020. Annual Energy
Outlook (2020).

146
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.

20.
21.
22.
23.
24.
25.
26.

U.S. Energy Information Administration. Natural Gas Consumption by End Use. (2020).
United States Energy Information Administration. California State Energy Profile. U.S. Energy Information
Administration (2020).
California Renewables Portfolio Standard Program. California Legislative Information (2018).
Thema, M., Bauer, F. & Sterner, M. Power-to-Gas: Electrolysis and methanation status review. Renew.
Sustain. Energy Rev. 2019, 112, 775–787.
AgSTAR. Livestock Anaerobic Digester Database. (2020).
American Biogas Council. Biogas State Profile: California. (2015).
GMI. Successful Applications of Anaerobic Digestion From Across the World. Global Methane Initiative (GMI) (2013).
National Renewable Energy Laboratory. 2018 Annual Technology Baseline. (2018).
NASA.
Technology
readiness
levels
introduction.
(2004).
Available
at:
(Accessed: 24th January 2021)
Peters, R., Baltruweit, M., Grube, T., Samsun, R. C. & Stolten, D. A techno economic analysis of the
power to gas route. J. CO2 Util. 2019, 34, 616–634.
Shaner, M. R., Atwater, H. A., Lewis, N. S. & McFarland, E. W. A comparative technoeconomic analysis
of renewable hydrogen production using solar energy. Energy Environ. Sci. 2016, 9, 2354–2371.
Keith, D. W., Angelo, D. St., Holmes, G. & Heidel, K. A process for capturing CO2 from the atmosphere.
Joule 2018, 2, 1573–1594.
Glenk, G. & Reichelstein, S. Economics of converting renewable power to hydrogen. Nat. Energy 2019,
4, 216–222.
Eisaman, M. D. et al. Indirect ocean capture of atmospheric CO2: Part II. Understanding the cost of
negative emissions. Int. J. Greenh. Gas Control 2018, 70, 254–261.
Sang Sefidi, V. & Luis, P. Advanced Amino Acid-Based Technologies for CO2 Capture: A Review. Ind.
Eng. Chem. Res. 2019, 58, 20181–20194.
Hu, X. et al. Toward Solvent Development for Industrial CO2 Capture by Optimizing the Catalyst-Amine
Formulation for Lower Energy Consumption in the Solvent Regeneration Process. Energy and Fuels
(2019). doi:10.1021/acs.energyfuels.9b02874
Carbon Capture and Sequestration Technologies @ MIT. Century Plant Fact Sheet: Commercial EOR
using
Anthropogenic
Carbon
Dioxide.
(2019).
Available
at:
the,projects in the Permian Basin. (Accessed: 30th October 2020)
Evans, S. Around the world in 22 carbon capture projects. Carbon Brief (2014). Available at:
(Accessed:
20th
October 2020)
Hammond, G. P. & Spargo, J. The prospects for coal-fired power plants with carbon capture and storage:
A UK perspective. Energy Convers. Manag. 2014, 86, 476–489.
Chou, V., Iyengar, A. K. S., Shah, V. & Woods, M. Cost and performance baseline for fossil energy plants: Report
Number DOE/NETL-2007/1281. (2015).
Fout, T. et al. Cost and Performance Baseline for Fossil Energy Plants Volume 1a: Bituminous Coal (PC) and Natural
Gas to Electricity Revision 3. National Energy Technology Laboratory (NETL) (2015).
Wurzbacher, J. A., Gebald, C., Piatkowski, N. & Steinfeld, A. Concurrent separation of CO2 and H2O
from air by a temperature-vacuum swing adsorption/desorption cycle. Environ. Sci. Technol. 2012, 46,
9191–9198.
Eisaman, M. D. Negative Emissions Technologies: The Tradeoffs of Air-Capture Economics. Joule 2020,
4, 516–520.
Patterson, B. D. et al. Renewable CO2 recycling and synthetic fuel production in a marine environment.
Proc. Natl. Acad. Sci. U. S. A. 2019, 116, 12212–12219.

147
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.

Eisaman, M. D. et al. CO2 extraction from seawater using bipolar membrane electrodialysis. Energy
Environ. Sci. 2012, 5, 7346–7352.
de Lannoy, C. F. et al. Indirect ocean capture of atmospheric CO2: Part I. Prototype of a negative
emissions technology. Int. J. Greenh. Gas Control 2018, 70, 243–253.
Willauer, H. D., Dimascio, F., Hardy, D. R. & Williams, F. W. Feasibility of CO2 extraction from seawater
and simultaneous hydrogen gas generation using a novel and robust electrolytic cation exchange module
based on continuous electrodeionization technology. Ind. Eng. Chem. Res. 2014, 53, 12192–12200.
Willauer, H. D., DiMascio, F., Hardy, D. R. & Williams, F. W. Development of an Electrolytic Cation
Exchange Module for the Simultaneous Extraction of Carbon Dioxide and Hydrogen Gas from Natural
Seawater. Energy and Fuels 2017, 31, 1723–1730.
Digdaya, I. A. et al. A direct coupled electrochemical system for capture and conversion of CO2 from
oceanwater. Nat. Commun. 2020, 11, 1–10.
Wichelns, D. Agricultural Water Pricing : United States. (2010).
LADWP. Briefing Book 2018-19. (2019).
Water Corporation. Water Forever, towards climate resilience. (2009).
San Diego County Water Authority. Seawater Desalination. (2015). doi:10.1002/9780470276280.ch3
Jarimi, H., Powell, R. & Riffat, S. Review of sustainable methods for atmospheric water harvesting. Int. J.
Low-Carbon Technol. 2020, 15, 253–276.
Kim, Hyunho, Yang, S. et al. Water harvesting from air with metal-organic frameworks powered by
natural sunlight. Science 2017, 434, 430–434.
Schemenauer, R. S. & Cereceda, P. A proposed standard fog collector for use in high-elevation regions.
(1387).
Klemm, O. et al. Fog as a fresh-water resource: Overview and perspectives. Ambio 2012, 41, 221–234.
Drupps
Electric.
The
Value
of
Water.
(2020).
Available
at:
Drupps Electric. An Adaptable Solution to Water Security. (2020). Available at:
Maestre-Valero, J. F., Ragab, R., Martínez-Alvarez, V. & Baille, A. Estimation of dew yield from radiative
condensers by means of an energy balance model. J. Hydrol. 2012, 460–461, 103–109.
Muselli, M., Beysens, D. & Milimouk, I. A comparative study of two large radiative dew water condensers.
J. Arid Environ. 2006, 64, 54–76.
Khalil, B. et al. A review: dew water collection from radiative passive collectors to recent developments
of active collectors. Sustain. Water Resour. Manag. 2016, 2, 71–86.
Zhou, M. et al. Accelerating vapor condensation with daytime radiative cooling. arxiv 6 (2019).
doi:10.1117/12.2525125
International Organization for Dew Utilization. Condensers for sale. (2019). Available at:
Kauw, M., Benders, R. M. J. & Visser, C. Green methanol from hydrogen and carbon dioxide using
geothermal energy and/or hydropower in Iceland or excess renewable electricity in Germany. Energy
2015, 90, 208–217.
Becker, W. L., Penev, M. & Braun, R. J. Production of synthetic natural gas from carbon dioxide and
renewably generated hydrogen: A techno-economic analysis of a power-to-gas strategy. J. Energy Resour.
Technol. 2019, 141.
Rönsch, S. et al. Review on methanation - From fundamentals to current projects. Fuel 2016, 166, 276–
296.
NREL. H2A: Hydrogen Analysis Production Models. (2018).
FuelCellsWorks. World’s Largest Hydrogen Plant Opens in Fukushima. (2020).

148
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.

James, B. D., Colella, W. G. & Moton, J. M. Techno-Economic Analysis of Hydrogen Production Pathways. Stategic
Analysis (2013).
James, B., Colella, W. & Moton, J. PEM Electrolysis H2A Production Case Study Documentation. (2013).
Shiva Kumar, S. & Himabindu, V. Hydrogen production by PEM water electrolysis – A review. Mater.
Sci. Energy Technol. 2019, 2, 442–454.
Elder, R., Cumming, D. & Mogensen, M. B. High Temperature Electrolysis. Carbon Dioxide Utilisation: Closing
the Carbon Cycle: First Edition (Elsevier B.V., 2015). doi:10.1016/B978-0-444-62746-9.00011-6
Lonis, F., Tola, V., Cascetta, M., Arena, S. & Cau, G. Performance evaluation of an integrated energy
system for the production and use of renewable methanol via water electrolysis and carbon dioxide
hydrogenation. in AIP Conference Proceedings 2019, 2191.
Joy, J., Mathew, J. & George, S. C. Nanomaterials for photoelectrochemical water splitting – review. Int.
J. Hydrogen Energy 2018, 43, 4804–4817.
Cheng, W. H. et al. Monolithic Photoelectrochemical Device for Direct Water Splitting with 19%
Efficiency. ACS Energy Lett. 2018, 3, 1795–1800.
Woodhouse, M., Smith, B., Ramdas, A. & Robert Margolis. Crystalline Silicon Photovoltaic Module
Manufacturing Costs and Sustainable Pricing: 1H 2018 Benchmark and Cost Reduction Roadmap. National Renewable
Energy Laboratory (2020).
Steinfeld, A. Solar thermochemical production of hydrogen - A review. Sol. Energy 2005, 78, 603–615.
Mcdaniel, A. et al. High Efficiency Solar Thermochemical Reactor for Hydrogen Production. (2017).
Peterson, D. & Miller, E. Hydrogen and Fuel Cells Program Record - Hydrogen Production Cost from Solid Oxide
Electrolysis. Hydrogen and Fuel Cells Program Record - Hydrogen Production Cost from Solid Oxide Electrolysis (2016).
Business Portal Norwegen. Nel ASA:Framework agreement for six hydrogen plants in France. 2019,
(2017).
Amos, W. A. Cost of Storing and Transporting Hydrogen. NREL (1998).
Electrochaea. Applications of Electrochaea’s BioCat biomethanation technology. (2018).
Thauer, R. K., Kaster, A. K., Seedorf, H., Buckel, W. & Hedderich, R. Methanogenic archaea:
Ecologically relevant differences in energy conservation. Nat. Rev. Microbiol. 2008, 6, 579–591.
Electrochaea. Power-to-Gas via Biological Catalysis (P2G-Biocat). (2017).
Veselovskaya, J. V., Parunin, P. D. & Okunev, A. G. Catalytic process for methane production from
atmospheric carbon dioxide utilizing renewable energy. Catal. Today 2017, 298, 117–123.
Strategieplattform Power to Gas. Audi e-gas project. (2014).
Luna, P. De, Hahn, C., Higgins, D. & Jaffer, S. A. What would it take for renewably powered
electrosynthesis to displace petrochemical processes? 2019, 3506, 1–9.
Nitopi, S. et al. Progress and Perspectives of Electrochemical CO2 Reduction on Copper in Aqueous
Electrolyte. Chem. Rev. 2019, 119, 7610–7672.
Qiu, Y. L. et al. Copper Electrode Fabricated via Pulse Electrodeposition: Toward High Methane
Selectivity and Activity for CO2 Electroreduction. ACS Catal. 2017, 7, 6302–6310.
García de Arquer, F. P. et al. CO2 electrolysis to multicarbon products at activities greater than 1 A cm−2.
Science 2020, 367, 661–666.
De Luna, P. et al. What would it take for renewably powered electrosynthesis to displace petrochemical
processes? Science 2019, 364.
Fu, Q. et al. Hybrid solar-to-methane conversion system with a Faradaic efficiency of up to 96%. Nano
Energy 2018, 53, 232–239.
Cheng, W. H. et al. CO2 Reduction to CO with 19% Efficiency in a Solar-Driven Gas Diffusion Electrode
Flow Cell under Outdoor Solar Illumination. ACS Energy Lett. 2020, 470–476.
Spath, P. et al. Biomass to Hydrogen Production Detailed Design and Economics Utilizing the Battele Columbus
Laboratory Indirectly-Heated Gasifier. (2005).
Ghaib, K. & Ben-Fares, F. Z. Power-to-Methane: A state-of-the-art review. Renew. Sustain. Energy Rev.

149

79.
80.
81.

2018, 81, 433–446.
Martin, M. R., Fornero, J. J., Stark, R., Mets, L. & Angenent, L. T. A single-culture bioprocess of
methanothermobacter thermautotrophicus to upgrade digester biogas by CO2-to-CH4 conversion with
H2. Archaea 2013, (2013).
US Energy Information Administration. United Statets Nattural Gas Industrial Price. (2020).
Matsumiya, N., Teramoto, M., Kitada, S. & Matsuyama, H. Evaluation of energy consumption for
separation of CO2 in flue gas by hollow fiber facilitated transport membrane module with permeation of
amine solution. Sep. Purif. Technol. 2005, 46, 26–32.

150
Chapter 9

CONCLUSIONS
9.1

Conclusion
Throughout this thesis we have explored a variety of different ways to control and enhance the CO2

reduction reaction, as well as analyzed systems that synergies between different CO2 capture and conversion
technologies.

Figure 9.1: shows an SEM and a schematic of a nanoporous gold catalyst. On the left is a cross sectional SEM of a 25% gold
nanoporous gold sample. On the right is a schematic of the nanoporous gold during electroreduction conditions. The red curve
represents the pH in solution as a function of distance from the electrode surface. The solid part of the red curve has been calculated
based on experimental conditions and the dashed portion of the red line is an assumed pH inside of the nanoporous gold.
In Chapter 2 we discussed how nanoporous gold (np-Au) films, with pore sizes ranging from 10 nm to
30 nm, represent promising electrocatalytic architectures for the CO2 reduction reaction due to their large
electrochemically active surface area, relative abundance of grain boundaries, and ability to support pH gradients
inside the nanoporous network. Electrochemical studies show that np-Au films support partial current densities
for the conversion of CO2 to CO in excess of 6 mA cm-2 at a Faradaic efficiency of ~99% in aqueous electrolytes
(50 mM K2CO3 saturated with CO2). Moreover, np-Au films are able to maintain Faradaic efficiency greater than
80% for CO production over prolonged periods of continuous operation (110 h). Electrocatalytic experiments
at different electrolyte concentrations demonstrate that the pore diameter of nanoporous cathodes represents a
critical parameter for creating and controlling local pH gradients inside the porous network of metal ligaments
(Figure 9.1). These results demonstrate the merits of nanoporous metal films for the CO2 reduction reaction

151
and offer an interesting architecture for highly selective electrocatalysis capable of sustaining high catalytic
currents over prolonged periods.
Electrolyte
Catalyst

Mostly dry
CO2 CO2 CO2

Mostly
flooded

Figure 9.2: shows an SEM and a schematic of a nanoporous gold gas diffusion electrode system with arrows indicating which
regions of the catalyst are mostly dry or mostly flooded
Then in Chapter 3 we built on this catalyst by transforming it into a gas diffusion electrode. This unique
catalyst structure allowed us to achieve Faradaic efficiencies for CO greater than 95% and a maximum partial
current density for CO of -168 mA cm-2. In addition, through a combination of secondary ion mass
spectroscopy and copper underpotential deposition we showed that approximately half of the catalyst is not in
contact with the electrolyte during operation and that the majority of this dry region exists in the bottom half of
the catalyst.

Figure 9.3: shows a schematic of the confocal fluorescent microscopy experimental set up. On the left we see the water immersion
objective scanning the surface of a CO2 reduction gas diffusion electrode. On the right is an example of a pH map, for more details
see Chapter 4.
After exploring how local pH can impact catalyst performance in Chapter 2, we dive more deeply into
trying to understand the pH in Chapter 4. Here, confocal fluorescent microscopy was used to map the electrolyte
pH near a copper gas diffusion electrode during CO2 reduction with micron spatial resolution in three
dimensions (Figure 9.3). We observed that the local pH increased from pH 6.8 to greater than pH 10 as the

and ‡Joint Center for Artificial Photosynthesis, California Institute of Technology,

152

current density was increased from 0 to 28 mA/cm2 in a 100 mM KHCO3 electrolyte. Variations in the pH
across the surface indicate areas of locally increased activity. Within deep trenches of the active layer, the local

carriers from plasmonicnities for driving photoumerous examples of hot
smonic systems capable of
s has eluded the nascent
ricate gold/p-type gallium
for photoelectrochemical
d conversion. Despite the
ot-hole injection
of more
pH increases
as trench width decreases. Computational models confirm these experimental results and also
on, plasmonic Au/p-GaN
showed that
catalyst
withinofnarrow
trenches
is more
than that found
operties consistent
with
thefound
injection
hot holes
from
Au active
nanoparticles
into atp-the surface of the electrode.
This study of
suggests
that the overpotential
required
to perform
selective
2 reduction can be reduced by
nt action spectrum
the plasmonic
photocathodes
faithfully
follows
the CO
surface
he Au nanoparticles
open-circuit
voltage
studies
demonstratelayer.
a sustained
increasing theand
density
of narrow trench
regions
in the microporous
arison with Ohmic Au/p-NiO heterojunctions confirms that the vast majority of
Au are sufficiently hot to inject above the 1.1 eV interfacial Schottky barrier at the
Figure 9.4: shows
a schematic of the illuminated
nanoparticleswith
on p-type
The circle on the right shows the band
ated plasmon-driven
photoelectrochemical
CO2gold
reduction
the GaN.
Au/p-GaN
ty for CO diagram
production
over and
H2 the
evolution
in aqueous
of the p-GaN
Au nanoparticles,
showing electrolytes.
the energy that theTaken
hot holestogether,
need to be injected into the semiconductor.
oexcited hot holes more than 1 eV below the Au Fermi level and demonstrate a
ot carriers to drive solar-to-fuel energy conversion.
In Chapter 5 and 6 we then explored harvesting non-equilibrium hot carriers from plasmonic-metal
s, plasmonic photocathode, CO2 reduction, Schottky barrier, hot holes
nanostructures offers unique opportunities for driving photochemical reactions at the nanoscale. Despite

numerous examples of hot electron-driven processes, the realization of plasmonic systems capable of harvesting

8−12
frommetal
the Fermi
level) than
hot electrons
(Figure
Au
lectron−hole
hot holes from
nanostructures
has eluded
the nascent
field of1b).
plasmonicInphotocatalysis.
In Chapter 5
al nanostrucnanostructures,
imbalance
in hotnitride
carrier(Au/p-GaN)
distributions
would junctions tailored for
(Figure 9.4),
we fabricate an
gold/p-type
gallium
Schottky
ling chemical
be expected to occur for photon energies hν > 1.8 eV.8
photoelectrochemical studies of plasmon-induced hot-hole capture and conversion. Despite the presence of an
capture and
Resonant optical excitation of the dipole plasmon mode in
interfacial Schottky barrier to hot-hole injection of more than 1 eV across the Au/p-GaN heterojunction,
s challenges,
spherical Au nanoparticles (hv ∼ 2.4 eV) should therefore
−20 nm) and
plasmonic Au/p-GaN
photocathodes
photoelectrochemical
properties
preferentially
produce exhibit
hot holes
within the Au
d-bandconsistent
that with the injection of
8−12,33
ier collection
reside
far below the
Au Fermi
This The
substantial
hot holes from
Au nanoparticles
into p-GaN
upon level.
plasmon excitation.
photocurrent action spectrum of
an interfacial
asymmetry
between
thefollows
energy
distributions
hot carriers
the plasmonic
photocathodes
faithfully
the surface
plasmon of
resonance
absorption spectrum of the Au
ls (e.g., Au)
implies a greater collection efficiency of hot holes relative to hot
and open-circuit voltage studies demonstrate a sustained photovoltage during plasmon excitation.
pe TiO2) nanoparticles
to
electrons for a comparable Schottky barrier height (Figure 1).
osensitization
Comparison with Ohmic Au/p-NiO heterojunctions confirms that the vast majority of hot holes generated via
The strong oxidizing power of these hot d-band holes also
ed solar cells
interband transitions
in Au
are sufficiently
hot to inject
above oxidation
the 1.1 eV interfacial
Schottky
offers the
potential
for driving
various
reactions
if barrier at the Au/psystems have
they could
transferred
to plasmon-driven
an appropriate
catalyst. Indeed,CO2 reduction with the
GaN heterojunction.
Webe
further
investigated
photoelectrochemical
for sub-band
adsorbed
citrate
molecules
in production
the plasmonAu/p-GaNphoto-oxidation
photocathodes, andofobserved
improved
selectivity
for CO
over H2 evolution in aqueous
photocatalydriven
synthesis
of
colloidal
Ag
and
Au
nanoprisms
is
known
to
from surface
proceed more efficiently via “hot” d-band holes as compared to
ave predicted
“warm” sp-band holes.34−39 Strategies that can efficiently and
hot electrons
selectively harvest hot holes from metal nanostructures would
) in common
of the high

153
electrolytes. Taken together, our results offer experimental validation of photoexcited hot holes more than 1
Letter for harvesting hot carriers to
eV below the Au Fermi level and demonstrate a photoelectrochemical platform

drive solar-to-fuel energy conversion. Then in Chapter 6 (Figure 9.5) we report the light-induced modification

Nanoparticle
Cu/p-NiO
Photocathode
of catalytic selectivity
for photoelectrochemical
CO reduction in aqueous media using copper nanoparticles
onto2
p-type
nickel oxide photocathodes.
Optical excitation of Cu nanoparticles generates hot electrons
tivity dispersed
for CO
Reduction
in Aqueous

available for driving CO2 reduction on the Cu surface while charge separation is accomplished by hot hole
injection from the Cu nanoparticles into the underlying p-NiO support.

, Alex J. Welch, Xueqian Li, Wen-Hui Cheng,

Photoelectrochemical studies

demonstrate that optical excitation of plasmonic Cu/p-NiO photocathodes imparts increased selectivity for CO2
reduction over hydrogen evolution in aqueous electrolytes. Specifically, we observed that plasmon-driven CO2

theOnline
production of carbon monoxide and formate, while simultaneously reducing the
ett.9b04895 reduction increased
Read
evolution of hydrogen. Our results demonstrate an optical route towards steering the selectivity of artificial
photosynthetic systems with plasmon-drivensı photocathodes for photoelectrochemical CO2 reduction in
Article Recommendations

aqueous media.

* Supporting Information

dification of catalytic selectivity for
ueous media using copper (Cu)
de (p-NiO) photocathodes. Optical
electrons available for driving CO2
aration is accomplished by hot-hole
the underlying p-NiO support.
at optical excitation of plasmonic
selectivity for CO2 reduction over
ecifically, we observed that plasmonn of carbon monoxide and formate,
hydrogen. Our results demonstrate
of artificial photosynthetic systems
oelectrochemical CO2 reduction in

Figure 9.5: shows a schematic of illuminated copper nanoparticles on p-type NiO.

ctrochemistry, hot holes, plasmonic photocathode, CO2 reduction
After exploring the specific science of how to convert CO2 into chemical products, in Chapters 7 and 8
we turned our focus to how to design CO2 capture and conversion systems that work together. In typical CO2

commonly used metals for electrocatalytic CO2 reduction,
capture andnamely,
conversionAg,
systems
five all
components:
a CO2 source,
(ii) a medium to capture the CO2,
Au, there
and areCu,
support (i)surface
plasmon
lytic machinery
excitations,
metal catalysts offer new oppor(iii) the release
of the COnanostructured
2 from the capture material, (iv) compression of the released CO2, and (v) conversion
ganic materials
tunities for exploiting their unique optical properties to shape
ter (H2O), of
and
the CO2. In Chapter 7, we evaluate the energy requirements
current prototypes and pilot scale plants for
the
the selectivity of chemical reactions.26,34−39 Inofparticular,
1−6
ylene, etc.).
CO captureplasmon-driven
and conversion.production
We point out
an alternative
pathway
which
uses bicarbonate or carbonate as a
of energetic
“hot”
carriers
on metal
ess is currently2
nanostructures
shown
promise
for photocatalychemical feedstock
rather thanhas
gaseous
CO2.great
The use
of bicarbonate
or carbonate removes the need to release
with selective
34−39
but
the
prompt
decay
(t
ps)
of
hot
carriers
into
sis,
without CO
the2 from the capture medium or compress it while also offering opportunities
for higher conversion
phonon modes of the metal nanocrystal requires a strategy for
The complexity
quickly separating hot electron−hole pairs on an ultrafast time
ultiple protonscale.38,39 To that end, numerous studies have established the
a process for
benefits of forming an interfacial Schottky barrier between a
termediates to
plasmonic metal and a wide band gap n-type semiconductor
t of interest.1−6
(e.g., Au/TiO ) for separating hot carriers across the metal−

154
efficiencies. Bicarbonate and carbonate can be electrochemically converted into syngas or formate and
subsequently into valuable chemicals and fuels using existing industrial processes. We suggest that moving to a
bicarbonate or carbonate feedstock will reduce the energy required to capture and convert atmospheric CO2.
Then in Chapter 8 we give a technical and economic perspective for the generation of renewable CH4 from
sunlight, H2O, and CO2. We evaluate the technology readiness level (TRL), the demonstrated scale of these
candidate technologies, the cost for CH4 generation, as well as the cost required to provide the necessary
feedstocks – H2O, CO2, and H2. We investigate and compare four main CO2 methanation pathways:
thermochemical (via the Sabatier reaction), biochemical, photo-electrochemical, and electrochemical. A unified
technoeconomic framework allows for side-by-side comparison of different pathways for sustainable CH4
generation. Based on the state-of-the-art materials and processes, we find that the indirect methanation
processes: thermochemical or biochemical are significantly more economic than that of the direct methanation
processes: electrochemical or photoelectrochemical routes. We also provide quantitative key metrics for
electrochemical or photoelectrochemical processes to become cost competitive with the indirect methanation
processes.
Through the process of evaluating different carbon capture and conversion systems (Chapter 7 and 8)
we have shown that the conversion process is a critical parameter. Since CO2 conversion is so critical and the
least understood part of the system, we have tried to bring further understanding to how to optimize these
devices. We have shown that illumination can be used to suppress hydrogen evolution reaction as well as
enhance CO2 reduction reaction rates. We have also shown through the nanoporous gold catalyst and the
confocal fluorescent microscopy experiments that by increasing the local pH the activity and selectivity of the
catalyst are increased. Finally, we have shown that by nanostructuring a catalyst to increase its surface area and
undercoordinated sites the activity and selectivity can be enhanced. We hope that this thesis will provide critical
insight for the design of future electrochemical CO2 reduction devices.
9.2

Outlook
This thesis begins to probe some fundamental questions facing the CO2 reduction community;

particularly, where is the reaction actually occurring in the catalyst. Chapter 3 uses copper underpotential
deposition (Cu UPD) and secondary ion mass spectroscopy (SIMS) to probe this, and chapter 4 dives even
deeper by imaging the local pH during operation using confocal fluorescent microscopy. The open questions
that remain from Chapter 3 is what the ‘wetted’ layer actually looks like. Is it a thin layer of water coating the

155
catalyst as was proposed, or does it look more like a channel of electrolyte going through the pores. SIMS and
Cu UPD are not able to provide the necessary insight to answer this question, therefore I propose that future
work should be done using either Auger spectroscopy or cryogenic scanning electron microscopy (CSEM).
Auger is a promising technique because of its very high resolution which could allow us to see precisely where
the copper atoms are. (Copper is a proxy for the electrolyte because copper can only plated where the catalyst
is in contact with the electrolyte.) CSEM is also intriguing because the electrode could be frozen during CO2
reduction and then the water could be directly imaged. The more we understand about the interaction between
the water and the catalyst the better electrodes we will be able to design, and in addition this experimental
knowledge will also allow us to create more accurate models of the system.
The next phase of experiments for the confocal fluorescent microscopy will be to find dyes that are not
sensitive to pH, but are sensitive to CO2 reduction products or hydrogen. This will allow us to create maps
showing which part of the catalyst is producing what product. Then we can match which catalyst site under
what conditions is producing what product which would be very exciting. This type of data can also help to
improve models and allow us to create electrodes that are more selective for a single product.
While my work has begun answering these fundamental questions about catalyst environment and
activity, there is still much that is left to be explored. I am excited to see how future researchers continue to
probe these questions and build on this work.

156

Appendix A:

CO 2 Capture Companies

CO2 Source
Carbon
Engineering

Atmosphere

ClimeWorks

Atmosphere

Carbon Recycling
International

Geothermal
power plant

Capture Media
Air

Basic solution
beneath fans

Air

CO2

Calcium Carbonate

Geothermal
power plant

CO2

Heat adsorber

Adsorber

Compress to 50 bar
and heat to 498 K

Siemens and
Evonik

CO2

O2

Electrolyzer

Sn Gas Diffusion Electrode
CO2

Ag Gas
Diffusion
Electrode

CO2, CO, H2O
O2
CO2

Anaerobic
Microorganism

H2
CO2

Inject CO2 into Cement

O2

0.6 MJ/mol CO2

0.6 MJ/mol CO2

0.1 MJ/mol CO2

CO
CO

O2
HCOOH
O2

0.8 MJ/mol CO2
0.6 MJ/mol CO2
0.7 MJ/mol CO2

C6H13OH, C4H9OH

1.0 MJ/mol CO2

C2H5OH,C2H3O2-

7.2 MJ/mol CO2

CH4

Bio Catalyst

0.3 MJ/mol CO2

H2O Electrolysis

Ag Gas Diffusion Electrode

Electrochaea

CarbonCure

O2

CH3OH

Down flowing well to basaltic rock

Basic solution
CO2

Dioxide
Materials

Compressed CO2

CO2, H2O

Adsorber

Opus 12
CO2 Conversion Companies

CO32-

H2

Carb Fix

Total Energy
Requirements

Processing

0.5 MJ/mol CO2

H2O Electrolysis
Cement

53.9-60.6 MJ/mol CO2

Figure A1: Schematic representation of various pilot plants that capture CO2 and/or transform it to either concentrated and
pressurized CO2 or to a value-added product. The companies in the light blue region capture CO2 from the atmosphere. The
companies in the mid-blue region capture CO2 from concentrated sources. The companies in the dark blue region focus only on
transforming already captured CO2. The energies reported for these companies does not include the energy required to capture the
CO2. The energy reported for the electrolyzer systems assume 100% conversion efficiency of CO2 which is not realistic, so there would
be additional energy costs for a recirculating system.

157

cement
mineral

al
ic
em
ch
tro
ec
El

Release CO2
from capture
material

Capture
CO2

CO2
source

Compress CO2

CO2

CO2

Transform
HCO3-/CO32-

Transform
CO2

CH3OH
C2H5OH
CO, HCOOH
CO2
H2O

CH4
C2H5OH,C2H3O2C6H13OH, C4H9OH

Figure A2: Schematic representation of the steps for various prototype systems designed to capture CO2 and/or convert it to either
concentrated and pressurized CO2 or to a value-added product. The blue arrows represent prototype processes that capture and convert
CO2, grey arrows represent prototype processes that only focus on CO2 conversion, and the pink arrow represents the process that we
propose.

158
Section I: Energy Input Calculations for CO2 Capture and Conversion Processes
A: Carbon Engineering
Carbon Engineering uses fans to contact air with a basic solution, thus dissolving CO2 and transforming it into
carbonate.1 The carbonate reacts with Ca(OH)2 to form CaCO3, which is then heated to release pure CO2 for
compression and storage. Carbon Engineering reports that their process for capturing CO2 from the air requires
8.81 GJ per 1.3-1.5 metric tons of pure and compressed CO2 produced. Below we convert this value to kJ/mol
CO2 so that the value can be readily compared to other processes reported.
𝐸𝑛𝑒𝑟𝑔𝑦 =

8.81 𝐺𝐽
1 𝑡 𝑝𝑢𝑟𝑒 𝐶𝑂!
1𝑘𝑔
44.01 𝑔
1000000 𝑘𝐽
1.3 − 1.5 𝑡 𝑝𝑢𝑟𝑒 𝐶𝑂!
1000𝑘𝑔
1000𝑔 1 𝑚𝑜𝑙 𝐶𝑂!
1 𝐺𝐽
259 𝑡𝑜 298 𝑘𝐽
𝑚𝑜𝑙 𝐶𝑂!

We would like to note that Carbon Engineering has now reported that they have also developed a process
to make liquid fuels with their captured CO2. They estimate the cost will be $1/L once the process is scaled
up.2 However, they do not report any details on how they do this, what specific fuel they synthesize, or the
energy requirements of the process, so we will not discuss this process further.
B: CarbFix
Ragnheidardottier, E., et al.3 reports that the power requirement for the CarbFix pilot plant is 200 kW and it
stores 2099 metric tons of CO2 annually. Below we use these values to convert to kJ/mol CO2 so that the
value can be readily compared to other processes reported.
𝐸𝑛𝑒𝑟𝑔𝑦 =

200 𝑘𝐽 3600 𝑠𝑒𝑐 24 ℎ𝑟𝑠 365 𝑑𝑎𝑦𝑠
1 𝑦𝑟
1 𝑡 𝐶𝑂!
1𝑘𝑔
1 ℎ𝑟
1 𝑑𝑎𝑦
1 𝑦𝑟
2099 𝑡𝐶𝑂! 1000𝑘𝑔 1000𝑔
44.01 𝑔
132 𝑘𝐽
1 𝑚𝑜𝑙 𝐶𝑂! 𝑚𝑜𝑙 𝐶𝑂!

C: Carbon Recycling International
Kauw, M. et al.4 report that the Carbon Recycling International 5M plant consumes 47.9 MJ/kg methanol.
Below we convert this value to kJ/mol CO2 so that the value can be readily compared to other processes
reported.
𝐸𝑛𝑒𝑟𝑔𝑦 =

47.9 𝑀𝐽
1000 𝑘𝐽
1 𝑘𝑔
32.04 𝑔
1 𝑚𝑜𝑙 𝑚𝑒𝑡ℎ𝑎𝑛𝑜𝑙
𝑘𝑔 𝑚𝑒𝑡ℎ𝑎𝑛𝑜𝑙
1 𝑀𝐽
1000 𝑔 1 𝑚𝑜𝑙𝑒 𝑚𝑒𝑡ℎ𝑎𝑛𝑜𝑙
1 𝑚𝑜𝑙𝑒 𝐶𝑂!
1535 𝑘𝐽
𝑚𝑜𝑙 𝐶𝑂!

D: Opus12
Opus 12 reports that their pilot device uses 5kW and produces 15kg of CO per day, and the input is concentrated
CO2.5 We assume for these devices that the conversion percentage of CO2 is 100%. Below we convert these
values to kJ/mol CO2 so that the value can be readily compared to other processes reported.

159

𝐸𝑛𝑒𝑟𝑔𝑦 =

1 𝑑𝑎𝑦
24 ℎ𝑟𝑠 3600 𝑠𝑒𝑐
5 𝑘𝐽
28.01 𝑔
1 𝑚𝑜𝑙 𝐶𝑂
807 𝑘𝐽
15000𝑔 𝐶𝑂 1 𝑑𝑎𝑦
1 ℎ𝑟
1 𝑠𝑒𝑐 1 𝑚𝑜𝑙 𝐶𝑂 1 𝑚𝑜𝑙𝑒 𝐶𝑂! 𝑚𝑜𝑙 𝐶𝑂!

E: Dioxide Materials
Dioxide Materials provides information on the voltage at which to run their devices, the Faradaic efficiency, and
the stability of the devices.6,7 The electrolyzer that takes CO2 to CO runs at 500mA/cm2 at 3V full cell potential
with a FE of 95%. The cathode is made of a Ag catalyst and the membrane is made of Sustanion, leading to a
device stability of over 3000 hours.7 The electrolyzer that takes CO2 to formic acid runs at 160mA/cm2 at 3.5V
full cell potential with a FE of 90%. The cathode is made of an Sn nanoparticle catalyst, the membrane is made
of Sustanion, and the counter is IrO2.6 We assume for these devices that the conversion percentage of CO2 is
100%. We take this information and calculate the energy in kJ/mol CO2.
For CO2 to CO electrolyzer
𝑈𝑧𝐹 (3 𝑉)(2 𝑚𝑜𝑙𝑒 𝑒𝑙𝑒𝑐𝑡𝑟𝑜𝑛𝑠)(96485 C mol#* )
1 𝑘𝐽
609 𝑘𝐽
𝐸𝑛𝑒𝑟𝑔𝑦 =
𝐹𝐸
. 95
1000 𝐽 𝑚𝑜𝑙 𝐶𝑂!
For CO2 to Formic Acid electrolyzer
𝐸𝑛𝑒𝑟𝑔𝑦 =

𝑈𝑧𝐹 (3.5 𝑉)(2 𝑚𝑜𝑙𝑒 𝑒𝑙𝑒𝑐𝑡𝑟𝑜𝑛𝑠)(96485 C mol#* )
1 𝑘𝐽
750 𝑘𝐽
𝐹𝐸
. 90
1000 𝐽 𝑚𝑜𝑙 𝐶𝑂!

F: Cement production
World Coal Association reports that it requires 200kg of coal to produce 1 metric ton of cement.8 The higher
heating value of coal is approximately 28.8-32.4 MJ/kg of coal according to the European Nuclear Society.9
Carbon Cure can store 25 lbs of CO2 per cubic yard of cement.10 From these values we were able to calculate
the amount of energy required per mol of CO2 stored in the cement.
𝐸𝑛𝑒𝑟𝑔𝑦 ≅

200 𝑘𝑔 𝑐𝑜𝑎𝑙
28.8 𝑡𝑜 32.4 𝑀𝐽 1000 𝑘𝐽 1 𝑡𝑜𝑛𝑛𝑒
1 𝑘𝑔
3.15 𝑔 𝑐𝑒𝑚𝑒𝑛𝑡
1 𝑡𝑜𝑛𝑛𝑒 𝑐𝑒𝑚𝑒𝑛𝑡
1 𝑘𝑔 𝑐𝑜𝑎𝑙
1 𝑀𝐽
1000 𝑘𝑔 1000 𝑔
1 𝑐𝑚2
764555𝑐𝑚
1 𝑦𝑑
1𝑙𝑏
44.01 𝑔 𝐶𝑂!
53,837 𝑡𝑜 60,567 𝑘𝐽
1 𝑦𝑑 2
25 𝑙𝑏 𝐶𝑂! 453.6 𝑔
1 𝑚𝑜𝑙 𝐶𝑂!
𝑚𝑜𝑙 𝐶𝑂!

G: Siemens and Evonik
Haas, T., et al.11 report a system designed in collaboration between Siemens in Evonik.11 This system uses a CO2
electrolyzer to make syngas which is then fermented in the presence of bacteria. Depending on the type of
bacteria, operating conditions, and ratio of CO to H2, the fermenters either produce a 1:1 mixture of hexanol
and butanol or a combination of acetate and ethanol. For the hexanol and butanol they report that this process
requires 22kWh/kg of alcohol and then an additional 0.6 kWh/kg hexanol to separate the alcohols. For the
acetate and ethanol process they apply 4.94V over the whole CO2 electrolyzer. The fermenter is put at 36°C.11
We assume that the fermenter has a specific heat capacity of water to calculate the energy required to heat it.
We then add the energy for the electrolyzer to the energy of the fermenter to obtain a final energy, this energy

160
does not include the energy required to separate the products. In a practical system there would also be energy
costs for pumps, feeding the bacteria, capturing the CO2, etc. which we have not taken into account here.
For CO2 to hexanol and butanol
𝐸𝑛𝑒𝑟𝑔𝑦 =

22.6 𝑘𝑊ℎ 3600 𝑠𝑒𝑐 1 𝑘𝑔 𝑎𝑙𝑐𝑜ℎ𝑜𝑙 . 5 (74.12 + 102.17) 𝑔
7,171 𝑘𝐽
𝑘𝑔 𝑎𝑙𝑐𝑜ℎ𝑜𝑙
1 ℎ𝑟
1000 𝑔
1 𝑚𝑜𝑙 𝑎𝑙𝑐𝑜ℎ𝑜𝑙
𝑚𝑜𝑙 𝑎𝑙𝑐𝑜ℎ𝑜𝑙

For CO2 to acetate and ethanol
𝑈𝑧𝐹 (4.94 𝑉)(2 𝑚𝑜𝑙𝑒 𝑒𝑙𝑒𝑐𝑡𝑟𝑜𝑛𝑠)(96485 C mol#* )
1 𝑘𝐽
953 𝑘𝐽
𝐹𝐸
1000 𝐽
𝑚𝑜𝑙
1 𝑚𝑜𝑙𝑒
18.02 𝑔
1 𝑘𝑔
4.181 𝑘𝐽
𝑄 𝑤𝑎𝑡𝑒𝑟 = 𝑚𝑐𝛥𝑇 = 32 𝐿 𝐻! 𝑂 ×
4×
×3
4 (309 − 293𝐾)
22.4𝐿
1 𝑚𝑜𝑙𝑒 𝐻! 𝑂 1000 𝑔
𝑘𝑔 𝐾
0.11 𝑘𝐽
𝑚𝑜𝑙 𝐻! 𝑂
𝐸𝑛𝑒𝑟𝑔𝑦 =

H: Electrochaea12
For a BioCat 10 reaction the reported installed power is 220kW, the input of CO2 is 500 Nm3/h, the input of
H2 is 2000 Nm3/h. The output is 500 Nm3/h CH4.12
𝐸𝑛𝑒𝑟𝑔𝑦 =

220 𝑘𝐽 3600𝑠
1ℎ
1 𝑁𝑚2
22.4𝐿
36 𝑘𝐽
1𝑠
1ℎ
500 𝑁𝑚 𝐶𝑂! 1000𝐿 1 𝑚𝑜𝑙 𝐶𝑂!
𝑚𝑜𝑙 𝐶𝑂!

We assume use of an NEL H2O electrolyzer to produce the H2.4
𝐸𝑛𝑒𝑟𝑔𝑦 =

52 𝑘𝑊ℎ 3600𝑠 . 002 𝑘𝑔 𝐻!
374 𝑘𝐽
1 𝑘𝑔 𝐻!
1ℎ
1 𝑚𝑜𝑙 𝐻!
𝑚𝑜𝑙 𝐻!

Section 2: Energy Input Calculations for Bicarbonate or Carbonate Feedstock Systems
A: Energy required to dissolve CO2 in solutions of various alkalinity
Keith, D., et al.13 report an air contactor that requires 61 kWh/tCO2 for the fan which is 70% efficient and 21
kWh/tCO2 for the fan which is 82% efficient.13 In the air contactor described in this paper they have a basic
solution comprised of 2M K+, 0.45M CO32- , 1.1M OH-.13 We have calculated the energy in kJ/mol CO2
assuming this solution concentration and for a solution comprised of 2.24M K+, 2.24M HCO3- , 10-4.78M OH-.
Energy to dissolve CO2 in solution to make carbonate
Solution: 2M K+, 0.45M CO32- , 1.1M OH61 𝑘𝑊ℎ + 21 𝑘𝑊ℎ 3600 𝑠𝑒𝑐
1 𝑡𝐶𝑂!
1 𝑘𝑔
44.01 𝑔 𝐶𝑂!
13 𝑘𝐽
1 𝑡𝐶𝑂!
1 ℎ𝑟
1000 𝑘𝑔 1000 𝑔
1 𝑚𝑜𝑙
𝑚𝑜𝑙 𝐶𝑂!

161
Energy to dissolve CO2 in solution to make bicarbonate
Solution: 2.24M K+, 2.24M HCO3- , 10-4.78M OH61𝑘𝑊ℎ 21𝑘𝑊ℎ
0.45𝑀 𝐶𝑂2#!
65 𝑘𝐽
1 𝑡 𝐶𝑂! 1 𝑡 𝐶𝑂! 2.24 𝑀 𝐻𝐶𝑂2# 𝑚𝑜𝑙 𝐶𝑂!
65 𝑘𝑊ℎ 3600 𝑠𝑒𝑐 1 𝑡 𝐶𝑂!
1 𝑘𝑔
44.01 𝑔 𝐶𝑂!
10 𝑘𝐽
1 𝑡 𝐶𝑂!
1 ℎ𝑟
1000 𝑘𝑔 1000 𝑔
1 𝑚𝑜𝑙
𝑚𝑜𝑙 𝐶𝑂!
B: Bipolar membrane with Ag catalyst cathode and Ni foam anode
Sargent, E., et al.14 reports a carbonate electrolyzer that uses a Ag catalyst cathode and a Ni foam anode separated
by a bipolar membrane. The catholyte is a carbonate solution and the anolyte is a potassium hydroxide solution.
The device is tested at various potentials; for our calculations we choose to look at when the applied a full cell
potential of 3.8V, which gives the appropriate ratio of CO to H2.14
𝑈𝑧𝐹 ( 3.8𝑉)(2 𝑚𝑜𝑙𝑒 𝑒𝑙𝑒𝑐𝑡𝑟𝑜𝑛𝑠)(96485.33212 C mol#* )
1 𝑘𝐽
733 𝑘𝐽
𝐸𝑛𝑒𝑟𝑔𝑦 =
𝐹𝐸
1000 𝐽 𝑚𝑜𝑙 𝐶𝑂!
This energy does not take into account the energy of the pumps or other practical considerations that would be
needed for a functional system.
C: Bipolar membrane with Ag catalyst cathode and Ni anode
Li, T., et al.15 reports an electrolyzer that uses a Ag catalyst at the cathode and a Ni anode to transform a solution
of bicarbonate into syngas. When 3.5V is applied over the full cell the appropriate ratio of H2 and CO are
produced. Below we calculate the energy this corresponds to.15
𝑈𝑧𝐹 (3.5 𝑉)(2 𝑚𝑜𝑙𝑒 𝑒𝑙𝑒𝑐𝑡𝑟𝑜𝑛𝑠)(96485 C mol#* )
1 𝑘𝐽
675 𝑘𝐽
𝐸𝑛𝑒𝑟𝑔𝑦 =
𝐹𝐸
1000 𝐽 𝑚𝑜𝑙 𝐶𝑂!
This energy does not take into account the energy of the pumps or other practical considerations that would be
needed for a functional system.
D:
Area
required
of
electrolyzer,
capture
system,
and
solar
cells
We calculate the area required for a plant able to process 1 metric ton of CO2 per day (t-CO2/day) and how
much would area would be needed to convert the 37.1 gigatons of CO2 emitted globally annually.16 We also
assume that the plant is operates continuously for 24 hours, all power needs are generated on site via solar
energy, and that there is sufficient battery storage to power the capture and conversion system continuously.
We assume that the capture system is designed in the same way as Carbon Engineering’s air contactor. They
report that their air contactor processes 22 t-CO2/m2/year and that the plant they describe in their report can
capture 1 t-CO2/day.13 For the electrolyzer we assume that the metrics from the electrolyzer reported by Sargent,
et al. – 200 mA/cm2 at 3.8 V.17 For the solar system we assume a 350 W solar panels with an area of 2.03 m2.18
The average capacity factor in the united states is approximately 25%.19
Calculate area of capture system to capture 1 t-CO2 per day

162

𝑇𝑜𝑡𝑎𝑙 𝐴𝑟𝑒𝑎 =

1 𝑡 − 𝐶𝑂! 365 𝑑𝑎𝑦 𝑚! ∗ 𝑦𝑒𝑎𝑟
= 16.6 𝑚!
𝑑𝑎𝑦
1 𝑦𝑒𝑎𝑟
22 𝑡 − 𝐶𝑂!

Calculate area of electrolyzer to convert 1 t-CO2 per day
𝑇𝑜𝑡𝑎𝑙 𝐴𝑟𝑒𝑎 =

0.2 𝐶
10000 𝑐𝑚!
1 𝑚𝑜𝑙 𝐶𝑂!
60 𝑠𝑒𝑐 60 𝑚𝑖𝑛 24 ℎ𝑟
𝑐𝑚 ∗ 𝑠𝑒𝑐
1𝑚
3 𝑚𝑜𝑙 𝑠𝑦𝑛𝑔𝑎𝑠
1𝑚𝑖𝑛
1 ℎ𝑟
1 𝑑𝑎𝑦
1 𝑒𝑙𝑒𝑐𝑡𝑟𝑜𝑛 1 𝐶𝑂! 𝑚𝑜𝑙𝑒𝑐𝑢𝑙𝑒 𝑐𝑜𝑛𝑣𝑒𝑟𝑡𝑒𝑑
1 𝑚𝑜𝑙𝑒 𝐶𝑂!
1.6 ∗ 10#*> 𝐶
2 𝑒𝑙𝑒𝑐𝑡𝑟𝑜𝑛
6.022 ∗ 10!2 𝐶𝑂! 𝑚𝑜𝑙𝑒𝑐𝑢𝑙𝑒
.04401 𝑘𝑔
1 𝑡𝑜𝑛
2.36 𝑡𝑜𝑛 𝐶𝑂!
1 𝑡𝑜𝑛 𝐶𝑂!
1 𝑚𝑜𝑙 𝐶𝑂! 1000 𝑘𝑔
𝑚 ∗ 𝑑𝑎𝑦
0.38 𝑚! ∗ 𝑑𝑎𝑦

Calculate area of photovoltaic system needed to power plant that processes 1 t-CO2 per day
𝑃𝑜𝑤𝑒𝑟 𝑑𝑒𝑛𝑠𝑖𝑡𝑦 =
𝑇𝑜𝑡𝑎𝑙 𝐴𝑟𝑒𝑎 =

350 𝑊
1.03 𝑘𝑊ℎ
0.25
24
ℎ𝑟
2.03 𝑚!
1 𝑚!

746 𝑘𝐽
1 𝑚𝑜𝑙 𝐶𝑂! 1000 𝑘𝑔 0.0002778 𝑘𝑊ℎ
1 𝑚!
𝑘𝑊ℎ
𝑚𝑜𝑙 𝐶𝑂! . 04401 𝑘𝑔
1 𝑡𝑜𝑛
1 𝑘𝐽
1.03 𝑑𝑎𝑦 𝑠𝑜𝑙𝑎𝑟 𝑝𝑎𝑛𝑒𝑙
4571 𝑚! 0.005 𝑘𝑚!
𝑡𝐶𝑂! /𝑑𝑎𝑦 𝑡𝐶𝑂! /𝑑𝑎𝑦

Calculate area of whole system needed to process CO2 emitted globally daily
𝑇𝑜𝑡𝑎𝑙 𝑎𝑟𝑒𝑎 =

37.1 ∗ 10> 𝑡𝑜𝑛𝑠 𝐶𝑂!
1 𝑦𝑟
4571 𝑚!
1 𝑘𝑚!
1𝑦𝑟
365 𝑑𝑎𝑦𝑠 𝑡𝐶𝑂! /𝑑𝑎𝑦 (1000 𝑚)!
= 464,614 𝑘𝑚𝟐 ~ 460,000 𝑘𝑚𝟐

E: Energy required to make syngas from coal
Energy Technology Systems Analysis Programme reports that the process of synthesizing syngas from coal is
roughly 74.7% efficient.20 Syngas has roughly half of the energy density of natural gas. From this we can estimate
the energy required to make syngas.
𝐸 = 0.5 ×

33.4 − 82.7 𝑀𝐽
1𝑚2
22.4𝐿
3 𝑚𝑜𝑙 𝑠𝑦𝑛𝑔𝑎𝑠 100 1.5 − 3.7 𝑀𝐽
1 𝑚 𝑛𝑎𝑡𝑢𝑟𝑎𝑙 𝑔𝑎𝑠 1000𝐿 1 𝑚𝑜𝑙 𝑠𝑦𝑛𝑔𝑎𝑠
1 𝑚𝑜𝑙 𝐶𝑂
74.7
𝑚𝑜𝑙 𝐶𝑂

F: Energy required to make syngas from natural gas using steam reforming and the reverse water gas
shift with CO2 capture
For the process of steam reforming natural gas and then performing the reverse water gas shift to obtain the
proper ratio of CO to H2 there are many steps involved. Baltrusaitis, J. and W. Luyben21 report that the energy
for the preheater is 10.47 MW, vaporizer is 202.9 MW, steam methane reforming is 536.6 MW, the furnace for

163
the reverse water gas shift is 121.2 MW, heat exchanger two is 219.8 MW, heat exchanger is 46 MW, the steam
methane reforming cooler is 177.2 MW, the reboiler is 75.2 MW, the feed effluent heat exchanger is 53.2 MW,
the condenser is 30.2 MW, the compressor is 9.83 MW, the intercoolers are 10.76 MW, and the stripper cooler
is 40.86 MW. There are also some additional smaller energies that total 400MW. This system produces 15,013
kmol/h H2, 7500 kmol/h CO, 271 kmol/h CH4, and 94 kmol/h H2O. From this we can calculate the energy
needed to make the syngas.
𝑇𝑜𝑡𝑎𝑙 𝑝𝑜𝑤𝑒𝑟 = 10.47𝑀𝑊 + 202.9𝑀𝑊 + 536.6𝑀𝑊 + 121.2𝑀𝑊 + 46𝑀𝑊 + 177.2𝑀𝑊 + 75.2𝑀𝑊
+ 53.2 𝑀𝑊 + 30.2 𝑀𝑊 + 9.83 𝑀𝑊 + 10.76 𝑀𝑊 + 40.86𝑀𝑊 + 400𝑀𝑊
= 1,314.4 𝑀𝑊
𝐸=

1714.4 𝑀𝐽
60𝑠
60 𝑚𝑖𝑛
1ℎ𝑟
1𝑘𝑚𝑜𝑙
0.8 𝑀𝐽
1𝑠
1𝑚𝑖𝑛
1ℎ𝑟
7500 𝑘𝑚𝑜𝑙 𝐶𝑂 1000𝑚𝑜𝑙 𝑚𝑜𝑙 𝐶𝑂

G: Energy required to make syngas from natural gas using steam reforming and the reverse water gas
shift
For the process of steam reforming natural gas and then performing the reverse water gas shift to obtain the
proper ratio of CO to H2 there are many steps involved. Baltrusaitis, J. and W. Luyben21 report that the energy
for the preheater is 10.47 MW, vaporizer is 202.9 MW, steam methane reforming is 536.6 MW, the furnace for
the reverse water gas shift is 121.2 MW, heat exchanger two is 219.8 MW, heat exchanger 3 is 46 MW,
compressor is 9.83MW, and the steam methane reforming cooler is 177.2 MW here are also some additional
smaller energies that total 400MW. This system produces 15,013 kmol/h H2, 7500 kmol/h CO, 272 kmol/h
CH4, and 75 kmol/h H2O. From this we can calculate the energy needed to make the syngas.
𝑇𝑜𝑡𝑎𝑙 𝑝𝑜𝑤𝑒𝑟 = 10.47𝑀𝑊 + 202.9𝑀𝑊 + 536.6𝑀𝑊 + 121.2𝑀𝑊 + 46𝑀𝑊 + 177.2𝑀𝑊
+ 9.83 𝑀𝑊 + 400𝑀𝑊 = 1,504.2 𝑀𝑊
𝐸=

1314.4 𝑀𝐽
60𝑠
60 𝑚𝑖𝑛
1ℎ𝑟
1𝑘𝑚𝑜𝑙
0.7 𝑀𝐽
1𝑠
1𝑚𝑖𝑛
1ℎ𝑟
7500 𝑘𝑚𝑜𝑙 𝐶𝑂 1000𝑚𝑜𝑙 𝑚𝑜𝑙 𝐶𝑂

H: Pd-Pt cathode with Pt anode to reduce bicarbonate to formate
Kortlever, R., et al. reports an electrochemical cell that reduces bicarbonate to formate at low overpotentials
using a Pd-Pt cathode and Pt anode. A potential of -0.4V vs the reversible hydrogen electrode (RHE) is applied
to the cathode, which corresponds to ~90% faradaic efficiency for formate.22 We assume a full cell potential of
2V, since the cell has some resistance and the anode is performing OER which has a standard potential of 1.23V.
We have calculated the energy used by the cell below, the calculation does not include the energy that would be
required for pumps or other practical considerations.
𝐸𝑛𝑒𝑟𝑔𝑦 =

𝑈𝑧𝐹 (2 𝑉)(2 𝑚𝑜𝑙𝑒 𝑒𝑙𝑒𝑐𝑡𝑟𝑜𝑛𝑠)(96485 C mol#* )
1 𝑘𝐽
429 𝑘𝐽
𝐹𝐸
. 90
1000 𝐽 𝑚𝑜𝑙 𝐶𝑂!

H: Production of Sodium Hydroxide
The Office of Energy Efficiency and Renewable Energy of the U.S. Department of Energy report that when
a zero-gap membrane chlor-alkali cell with an oxygen-depolarized cathode is used for the chlor-alkali process,
sodium hydroxide can be produced for 2,500 kWh/ton.23

164

𝐸𝑛𝑒𝑟𝑔𝑦 =

2,500 𝑘𝑊ℎ 3600 𝑘𝐽
1 𝑡𝑜𝑛
1 𝑘𝑔
39.997 𝑔 𝑁𝑎𝑃𝐻
360 𝑘𝐽
1 𝑡𝑜𝑛 𝑁𝑎𝑂𝐻 1 𝑘𝑊ℎ 1000 𝑘𝑔 1000 𝑔
1 𝑚𝑜𝑙 𝑁𝑎𝑂𝐻
𝑚𝑜𝑙 𝑁𝑎𝑂𝐻

Section 3: CO2 Electrolyzer Conversion Efficiencies
For all calculations in this section we assume that the CO2 is at standard temperature and pressure.
A: Gewirth, A., et al CO2 to electrolyzer24
𝑀𝑜𝑙𝑠 𝑜𝑓 𝐶𝑂! 𝑝𝑒𝑟 𝑠𝑒𝑐𝑜𝑛𝑑 = 3

7 𝑚𝑙 𝐶𝑂! 1 𝑚𝑖𝑛 22,400 𝑚𝑙 𝐶𝑂!
5.2 ∗ 10#@ 𝑚𝑜𝑙 𝐶𝑂!
43
43
4=
1 𝑚𝑖𝑛
60𝑠
1 𝑚𝑜𝑙 𝐶𝑂!
1𝑠

. 17 𝐶 1 𝑚𝑜𝑙 𝑒 − 1 𝑚𝑜𝑙 𝑒𝑡ℎ𝑦𝑙𝑒𝑛𝑒
1.5 ∗ 10#*' 𝑚𝑜𝑙 𝐶𝑂
𝑀𝑜𝑙𝑠 𝑜𝑓 𝑒𝑡ℎ𝑦𝑙𝑒𝑛𝑒 𝑝𝑒𝑟 𝑠𝑒𝑐𝑜𝑛𝑑 = 3
43
43
4=
1𝑠
96485 𝐶
12 𝑚𝑜𝑙 𝑒 −
1𝑠
. 09 𝐶 1 𝑚𝑜𝑙 𝑒 − 1 𝑚𝑜𝑙 𝑒𝑡ℎ𝑎𝑛𝑜𝑙
7.8 ∗ 10#** 𝑚𝑜𝑙 𝐶𝑂
𝑀𝑜𝑙𝑠 𝑜𝑓 𝑒𝑡ℎ𝑎𝑛𝑜𝑙 𝑝𝑒𝑟 𝑠𝑒𝑐𝑜𝑛𝑑 = 3
43
43
4=
1𝑠
96485 𝐶
12 𝑚𝑜𝑙 𝑒 −
1𝑠
𝐶𝑜𝑛𝑣𝑒𝑟𝑠𝑖𝑜𝑛 𝑃𝑒𝑟𝑐𝑒𝑛𝑡𝑎𝑔𝑒 =

2.2 ∗ 10#*' 𝑚𝑜𝑙 𝐶2
1𝑠
= 0.004%
1𝑠
5.2 ∗ 10#@ 𝑚𝑜𝑙 𝐶𝑂!

B: Siemens and Evonik CO2 electrolyzer11
90 𝑚𝑙 𝐶𝑂! 1 𝑚𝑖𝑛 22,400 𝑚𝑙 𝐶𝑂!
6.7 ∗ 10#A 𝑚𝑜𝑙 𝐶𝑂!
𝑀𝑜𝑙𝑠 𝑜𝑓 𝐶𝑂! 𝑝𝑒𝑟 𝑠𝑒𝑐𝑜𝑛𝑑 = 3
43
43
4=
1 𝑚𝑖𝑛
60𝑠
1 𝑚𝑜𝑙 𝐶𝑂!
1𝑠
. 3 𝐶 1 𝑚𝑜𝑙 𝑒 − 1 𝑚𝑜𝑙 𝐶𝑂
1.1 ∗ 10#@ 𝑚𝑜𝑙 𝐶𝑂
43
43
4 ∗ .7 =
1𝑠
96485 𝐶 2 𝑚𝑜𝑙 𝑒 −
1𝑠
1.1 ∗ 10#@ 𝑚𝑜𝑙 𝐶𝑂
1𝑠
𝐶𝑜𝑛𝑣𝑒𝑟𝑠𝑖𝑜𝑛 𝑃𝑒𝑟𝑐𝑒𝑛𝑡𝑎𝑔𝑒 =
= 1.6%
1𝑠
6.7 ∗ 10#A 𝑚𝑜𝑙 𝐶𝑂!

𝑀𝑜𝑙𝑠 𝑜𝑓 𝐹𝑜𝑟𝑚𝑎𝑡𝑒 𝑝𝑒𝑟 𝑠𝑒𝑐𝑜𝑛𝑑 = 3

C: Sargent, E., et al CO2 electrolyzer25
𝑀𝑜𝑙𝑠 𝑜𝑓 𝐶𝑂! 𝑝𝑒𝑟 𝑠𝑒𝑐𝑜𝑛𝑑 = 3

30 𝑚𝑙 𝐶𝑂! 1 𝑚𝑖𝑛 22,400 𝑚𝑙 𝐶𝑂!
2.2 ∗ 10#A 𝑚𝑜𝑙 𝐶𝑂!
43
43
4=
1 𝑚𝑖𝑛
60𝑠
1 𝑚𝑜𝑙 𝐶𝑂!
1𝑠

. 75 𝐶 1 𝑚𝑜𝑙 𝑒 − 1 𝑚𝑜𝑙 𝑎𝑐𝑒𝑡𝑖𝑐 𝑎𝑐𝑖𝑑
43
43
4 ∗ .05
1𝑠
96485 𝐶
8 𝑚𝑜𝑙 𝑒 −
4.8 ∗ 10#B 𝑚𝑜𝑙 𝑎𝑐𝑒𝑡𝑖𝑐 𝑎𝑐𝑖𝑑
1𝑠

𝑀𝑜𝑙𝑠 𝑜𝑓 𝐴𝑐𝑒𝑡𝑖𝑐 𝑎𝑐𝑖𝑑 𝑝𝑒𝑟 𝑠𝑒𝑐𝑜𝑛𝑑 = 3

165
. 75 𝐶 1 𝑚𝑜𝑙 𝑒 − 1 𝑚𝑜𝑙 𝑒𝑡ℎ𝑦𝑙𝑒𝑛𝑒
43
43
4 ∗ .65
1𝑠
96485 𝐶
12 𝑚𝑜𝑙 𝑒 −
4.2 ∗ 10#C 𝑚𝑜𝑙 𝐶𝑂
1𝑠

𝑀𝑜𝑙𝑠 𝑜𝑓 𝑒𝑡ℎ𝑦𝑙𝑒𝑛𝑒 𝑝𝑒𝑟 𝑠𝑒𝑐𝑜𝑛𝑑 = 3

𝑀𝑜𝑙𝑠 𝑜𝑓 𝑒𝑡ℎ𝑎𝑛𝑜𝑙 𝑝𝑒𝑟 𝑠𝑒𝑐𝑜𝑛𝑑 = 3

. 75 𝐶 1 𝑚𝑜𝑙 𝑒 − 1 𝑚𝑜𝑙 𝑒𝑡ℎ𝑎𝑛𝑜𝑙
6.4 ∗ 10#B 𝑚𝑜𝑙 𝐶𝑂
43
43
4 ∗ .1 =
1𝑠
96485 𝐶
12 𝑚𝑜𝑙 𝑒 −
1𝑠

5.3 ∗ 10#C 𝑚𝑜𝑙 𝐶2
1𝑠
𝐶𝑜𝑛𝑣𝑒𝑟𝑠𝑖𝑜𝑛 𝑃𝑒𝑟𝑐𝑒𝑛𝑡𝑎𝑔𝑒 =
= 2.4%
1𝑠
2.2 ∗ 10#A 𝑚𝑜𝑙 𝐶𝑂!
D: Kenis, P., et al CO2 electrolyzer26
7 𝑚𝑙 𝐶𝑂! 1 𝑚𝑖𝑛 22,400 𝑚𝑙 𝐶𝑂!
5.2 ∗ 10#@ 𝑚𝑜𝑙 𝐶𝑂!
𝑀𝑜𝑙𝑠 𝑜𝑓 𝐶𝑂! 𝑝𝑒𝑟 𝑠𝑒𝑐𝑜𝑛𝑑 = 3
43
43
4=
1 𝑚𝑖𝑛
60𝑠
1 𝑚𝑜𝑙 𝐶𝑂!
1𝑠
. 275 𝐶 1 𝑚𝑜𝑙 𝑒 − 1 𝑚𝑜𝑙 𝐶𝑂
1.4 ∗ 10#@ 𝑚𝑜𝑙 𝐶𝑂
𝑀𝑜𝑙𝑠 𝑜𝑓 𝐶𝑂 𝑝𝑒𝑟 𝑠𝑒𝑐𝑜𝑛𝑑 = 3
43
43
4 ∗=
1𝑠
96485 𝐶 2 𝑚𝑜𝑙 𝑒 −
1𝑠
5.3 ∗ 10#C 𝑚𝑜𝑙 𝐶𝑂
1𝑠
𝐶𝑜𝑛𝑣𝑒𝑟𝑠𝑖𝑜𝑛 𝑃𝑒𝑟𝑐𝑒𝑛𝑡𝑎𝑔𝑒 =
= 27.3%
#A
1𝑠
2.2 ∗ 10 𝑚𝑜𝑙 𝐶𝑂!
E: Kenis, P., et al CO2 electrolyzer27
7 𝑚𝑙 𝐶𝑂! 1 𝑚𝑖𝑛 22,400 𝑚𝑙 𝐶𝑂!
5.2 ∗ 10#@ 𝑚𝑜𝑙 𝐶𝑂!
𝑀𝑜𝑙𝑠 𝑜𝑓 𝐶𝑂! 𝑝𝑒𝑟 𝑠𝑒𝑐𝑜𝑛𝑑 = 3
43
43
4=
1 𝑚𝑖𝑛
60𝑠
1 𝑚𝑜𝑙 𝐶𝑂!
1𝑠

𝑀𝑜𝑙𝑠 𝑜𝑓 𝐶𝑂 𝑝𝑒𝑟 𝑠𝑒𝑐𝑜𝑛𝑑 = 3

. 375 𝐶 1 𝑚𝑜𝑙 𝑒 − 1 𝑚𝑜𝑙 𝐶𝑂
1.8 ∗ 10#@ 𝑚𝑜𝑙 𝐶𝑂
43
43
4=
1𝑠
96485 𝐶 2 𝑚𝑜𝑙 𝑒 −
1𝑠

𝐶𝑜𝑛𝑣𝑒𝑟𝑠𝑖𝑜𝑛 𝑃𝑒𝑟𝑐𝑒𝑛𝑡𝑎𝑔𝑒 =

5.3 ∗ 10#C 𝑚𝑜𝑙 𝐶𝑂
1𝑠
= 34.8%
#A
1𝑠
2.2 ∗ 10 𝑚𝑜𝑙 𝐶𝑂!

F: Xuan, J., et al CO2 electrolyzer28
𝑀𝑜𝑙𝑠 𝑜𝑓 𝐶𝑂! 𝑝𝑒𝑟 𝑠𝑒𝑐𝑜𝑛𝑑 = 3

50 𝑚𝑙 𝐶𝑂! 1 𝑚𝑖𝑛 22,400 𝑚𝑙 𝐶𝑂!
3.7 ∗ 10#A 𝑚𝑜𝑙 𝐶𝑂!
43
43
4=
1 𝑚𝑖𝑛
60𝑠
1 𝑚𝑜𝑙 𝐶𝑂!
1𝑠

. 350 𝐶 1 𝑚𝑜𝑙 𝑒 − 1 𝑚𝑜𝑙 𝐻𝐶𝑂𝑂#
1.8 ∗ 10#@ 𝑚𝑜𝑙 𝐻𝐶𝑂𝑂#
𝑀𝑜𝑙𝑠 𝑜𝑓 𝑓𝑜𝑟𝑚𝑎𝑡𝑒 𝑝𝑒𝑟 𝑠𝑒𝑐𝑜𝑛𝑑 = 3
43
43
4=
1𝑠
96485 𝐶
2 𝑚𝑜𝑙 𝑒 −
1𝑠

166
1.8 ∗ 10#@ 𝑚𝑜𝑙 𝐻𝐶𝑂𝑂#
1𝑠
𝐶𝑜𝑛𝑣𝑒𝑟𝑠𝑖𝑜𝑛 𝑃𝑒𝑟𝑐𝑒𝑛𝑡𝑎𝑔𝑒 =
= 4.9%
1𝑠
3.7 ∗ 10#A 𝑚𝑜𝑙 𝐶𝑂!
G: Lv, J. J., et al CO2 electrolyzer29
10 𝑚𝑙 𝐶𝑂! 1 𝑚𝑖𝑛 22,400 𝑚𝑙 𝐶𝑂!
7.4 ∗ 10#@ 𝑚𝑜𝑙 𝐶𝑂!
𝑀𝑜𝑙𝑠 𝑜𝑓 𝐶𝑂! 𝑝𝑒𝑟 𝑠𝑒𝑐𝑜𝑛𝑑 = 3
43
43
4=
1 𝑚𝑖𝑛
60𝑠
1 𝑚𝑜𝑙 𝐶𝑂!
1𝑠
. 255 𝐶 1 𝑚𝑜𝑙 𝑒 − 1 𝑚𝑜𝑙 𝐶! 𝐻D
2.2 ∗ 10#C 𝑚𝑜𝑙 𝐶! 𝐻D
𝑀𝑜𝑙𝑠 𝑜𝑓 𝑒𝑡ℎ𝑦𝑙𝑒𝑛𝑒 𝑝𝑒𝑟 𝑠𝑒𝑐𝑜𝑛𝑑 = 3
43
43
4=
1𝑠
96485 𝐶
12 𝑚𝑜𝑙 𝑒 −
1𝑠
2.2 ∗ 10#C 𝑚𝑜𝑙 𝐶! 𝐻D
1𝑠
𝐶𝑜𝑛𝑣𝑒𝑟𝑠𝑖𝑜𝑛 𝑃𝑒𝑟𝑐𝑒𝑛𝑡𝑎𝑔𝑒 =
= 2.9%
1𝑠
7.4 ∗ 10#@ 𝑚𝑜𝑙 𝐶𝑂!
H: Verma, S., et al CO2 electrolyzer30
𝑀𝑜𝑙𝑠 𝑜𝑓 𝐶𝑂! 𝑝𝑒𝑟 𝑠𝑒𝑐𝑜𝑛𝑑 = 3

10 𝑚𝑙 𝐶𝑂! 1 𝑚𝑖𝑛 22,400 𝑚𝑙 𝐶𝑂!
1.3 ∗ 10#A 𝑚𝑜𝑙 𝐶𝑂!
43
43
4=
1 𝑚𝑖𝑛
60𝑠
1 𝑚𝑜𝑙 𝐶𝑂!
1𝑠

𝑀𝑜𝑙𝑠 𝑜𝑓 𝐶𝑂 𝑝𝑒𝑟 𝑠𝑒𝑐𝑜𝑛𝑑 = 3

. 255 𝐶 1 𝑚𝑜𝑙 𝑒 − 1 𝑚𝑜𝑙 𝐶𝑂
8.3 ∗ 10#C 𝑚𝑜𝑙 𝐶𝑂
43
43
4=
1𝑠
96485 𝐶 2 𝑚𝑜𝑙 𝑒 −
1𝑠

𝐶𝑜𝑛𝑣𝑒𝑟𝑠𝑖𝑜𝑛 𝑃𝑒𝑟𝑐𝑒𝑛𝑡𝑎𝑔𝑒 =

2.2 ∗ 10#C 𝑚𝑜𝑙 𝐶𝑂
1𝑠
= 6.6%
#@
1𝑠
7.4 ∗ 10 𝑚𝑜𝑙 𝐶𝑂!

BIBLIOGRAPHY: APPENDIX A
1.
2.
3.
4.
5.
6.
7.

Singh, M. R., Clark, E. L. & Bell, A. T. Effects of electrolyte, catalyst, and membrane composition and
operating conditions on the performance of solar-driven electrochemical reduction of carbon dioxide.
Phys. Chem. Chem. Phys. 2015, 17, 18924–18936.
Carbon Engineering. Air to fuels. (2019). Available at: https://carbonengineering.com/about-a2f/.
(Accessed: 18th October 2019)
Ragnheidardottir, E., Sigurdardottir, H., Kristjansdottir, H. & Harvey, W. Opportunities and challenges
for CarbFix: An evaluation of capacities and costs for the pilot scale mineralization sequestration project
at Hellisheidi, Iceland and beyond. Int. J. Greenh. Gas Control 2011, 5, 1065–1072.
Benders, M. J., Visser, C. & Kauw, M. Green methanol from hydrogen and carbon dioxide using
geothermal energy and/or hydropower in Iceland or excess renewable electricity in Germany. Energy
2015, 1–10.
Flanders, N., Kuhl, K. & Cave, E. Opus 12 – Recycling carbon dioxide back into fuels and chemicals. (2016).
Dioxide Materials. Dioxide Materials: Electrolyzer to transform carbon dioxide into formic acid.
Available at: https://dioxidematerials.com/technology/formic-acid/. (Accessed: 10th February 2020)
Dioxide Materials. Dioxide Materials: Electrolyzer to convert carbon dioxide to carbon monoxide.
Available at: https://dioxidematerials.com/technology/co2-electrolysis/. (Accessed: 10th February

167

8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.

2020)
What is cement? World Coal Association 1–4 (2019). Available at: https://www.worldcoal.org/coal/usescoal/coal-cement. (Accessed: 10th February 2020)
European
Nuclear
Society.
Fuel
comparison.
(2014).
Available
at:
2020)
Monkman, S. & MacDonal, M. Ready mixed technology case study CO2 utilization in concrete mix design
optimization. (2016).
Haas, T., Krause, R., Weber, R., Demler, M. & Schmid, G. Technical photosynthesis involving CO2
electrolysis and fermentation. Nat. Catal. 2018, 1, 32–39.
Electrochaea. Applications of Electrochaea’s BioCat biomethanation technology. (2018).
Keith, D. W., Angelo, D. St., Holmes, G. & Heidel, K. A process for capturing CO2 from the atmosphere.
Joule 2018, 2, 1573–1594.
Li, Y. C. et al. CO2 electroreduction from carbonate electrolyte. ACS Energy Lett. 2019, 4, 1427–1431.
Li, T. et al. Electrolytic conversion of bicarbonate into CO in a flow cell. Joule 2019, 3, 1487–1497.
Le Quere, C. et al. Global carbon budget 2018. (2018).
Li, Y. C. et al. CO2 electroreduction from carbonate electrolyte. ACS Energy Lett. 2019, 4, 1427–1431.
Trina Solar. The Tallmax framed 144 half cell module. (2019).
US Energy Information Administration. Electric Power Monthly with Data for September 2012. (2012).
Energy Technology Systems Analysis Programme. Syngas production from coal. (2010).
Baltrusaitis, J. & Luyben, W. L. Methane conversion to syngas for gas-to-liquids (GTL): is sustainable
CO2 reuse via dry methane reforming (DMR) cost competitive with SMR and ATR processes? ACS
Sustain. Chem. Eng. 2015, 3, 2100–2111.
Kortlever, R., Balemans, C., Kwon, Y. & Koper, M. T. M. Electrochemical CO2 reduction to formic acid
on a Pd-based formic acid oxidation catalyst. Catal. Today 2015, 244, 58–62.
U.S. Department of Energy. Advanced Chlor-Alkali Technology (CPS #1797). Industrial Technologies Program
(2004).
Hoang, T. T. H. et al. Nanoporous copper-silver alloys by additive-controlled electrodeposition for the
selective electroreduction of CO2 to ethylene and ethanol. J. Am. Chem. Soc. 2018, 140, 5791–5797.
Dinh, C. et al. CO2 electroreduction to ethylene via hydroxide-mediated copper catalysis at an abrupt
interface. Science 2018, 360, 783–787.
Kim, B., Hillman, F., Ariyoshi, M., Fujikawa, S. & Kenis, P. J. A. Effects of composition of the
microporous layer and the substrate on performance in the electrochemical reduction of CO2 to CO. J.
Power Sources 2016, 312, 192–198.
Ma, S. et al. Carbon nanotube containing Ag catalyst layers for efficient and selective reduction of carbon
dioxide. J. Mater. Chem. A 2016, 4, 8573–8578.
Lu, X., Leung, D. Y. C., Wang, H. & Xuan, J. A high performance dual electrolyte microfluidic reactor
for the utilization of CO2. Appl. Energy 2017, 194, 549–559.
Lv, J. J. et al. A highly porous copper electrocatalyst for carbon dioxide reduction. Adv. Mater. 2018, 30,
1–8.
Verma, S. et al. Insights into the low overpotential electroreduction of CO2 to CO on a supported gold
catalyst in an alkaline flow electrolyzer. ACS Energy Lett. 2017, 3, 193–198.

168

Appendix B:
Analytical basis
Relevant process assumption and cost information for direct air capture (DAC),1 low-temperature electrolysis
(LTE),2,3 high-temperature electrolysis (HTE),3,4 solar thermochemical hydrogen (STCH), and thermochemical
methanation5 were solicited from available reports. Various sources were adopted to develop process and cost
estimates for ocean capture,6–10 photoelectrochemical (PEC) hydrogen generation,11,12 grid electrolysis
methanation,3,4 PEC methanation,11,12 and biochemical methanation.13,14 The plant design capacity was assumed
to be 240 ton/day for CO2 capture, 45 ton/day for H2 generation, 81 ton/day for CH4 production.
In some cases where cost information were already available in a certain plant capacity, the costs of similar items
of different sizes were approximated to our specific design capacity using exponential scaling factor, as follows:
𝑆G I
𝐶E = 𝐶F 3 4
(B.1)
𝑆H
where CD is the cost at design capacity, CB is the cost at known baseline capacity, SD is the design capacity, SB is
the known baseline capacity, and N is the scaling factor exponent that varies from 0.1 to unity. Unless otherwise
specified, an average value of 0.6 was as used for N.
All equipment capital costs reported in the past were adjusted to 2019 dollars by multiplying the reported base
cost from an earlier year by the ratio of a cost index (I) in 2019 to a base cost index (Ibase) that corresponds to
the year of which the cost information was obtained:10
𝐶!'*> = 𝐶J

𝐼!'*>
𝐼J

(B.2)

where C2019 is the equipment purchase cost in 2019, CR is the equipment purchase cost in reference year, I2019 is
the Chemical Engineering Plant Cost Index (CEPCI) in 2019 and IR is CEPCI in reference year.
The calculated results of the levelized product costs were adjusted to 2020 dollars using inflation rate of 1.9%:
𝐿𝐶𝑃!'!' = 𝐿𝐶𝑃!'*> (1 + 1.9%)!'!'#!'*>

(B.3)

where LCP2020 and LCP2019 are the levelized cost of product in 2020 dollars and in 2019 dollars, respectively.

169

Table B1: The default H2A hydrogen production economic model and its financial values were applied to all
systems studied.
Financial Parameters
Length of Construction Period (years)
Year of analysis
% of Capital Spent in 1st Year of Construction
Start-up Time (years)
Plant life (years)
Analysis period (years)
Depreciation Schedule Length (years)
Depreciation Type
% Debt Financing
% Equity Financing
Interest rate on debt (%)
Debt period (years)
% of Fixed Operating Costs During Start-up (%)
% of Revenues During Start-up (%)
% of Variable Operating Costs During Start-up (%)
Decommissioning costs (% of depreciable capital investment)
Salvage value (% of total capital investment)
Inflation rate (%)
State Taxes (%)
Federal Taxes (%)
Total Tax Rate (%)
Working capital (% of yearly change in operating costs)
After-tax real IRR
After-tax nominal IRR

Assumptions
2020
100%
40
40
20
MACRS
60%
40%
3.70%
Constant debt
75%
50%
75%
10%
10%
1.9%
6.0%
21.0%
25.74%
15%
6%
8.01%

170
Table B2: Assumptions for calculating the cost of direct air capture of CO2 currently. For further information
see excel sheet titled “Direct Air Capture of CO2 (current)”.
Process Assumptions
Design capacity
Average production
Capacity factor
Baseline capacity
Scale ratio
Fan energy
Fluid pumping energy
Fluid pumping energy
Power produced from slaking heat
Energy consumption
ASU power usage
Compressor power usage
Calciner
Water consumption
Water price
Industrial electricity
Natural gas price

244444.44
220000
0.9
2684931.507
0.09
61
21
27
-77
32
238
132
369
4.7
0.69307362
0.049
3.5

kg CO2/day
kg CO2/day
kg CO2/day
kWh/t-CO2
kWh/t-CO2
kWh/t-CO2
kWh/t-CO2
kWh/t-CO2
kWh/t-CO2
kWh/t-CO2
kWh/t-CO2
t-water/t-CO2
$/t-water
$/kWh
$/GJ

Baseline system cost
Air contactor
Pellet reactor
Calciner-slaker
Air separation unit
CO2 compressor
Steam turbine
Power plant
Fines filter
Other equipment
Buildings
Transformer
Total

Baseline installed cost in startup
year dollars
$135,867,821
$91,490,678
$52,110,425
$45,209,958
$20,463,455
$7,971,229
$38,904,359
$20,939,349
$115,285,393
$2,974,339
$22,129,085

Capital costs
Depreciable capital costs
Direct capital cost
Indirect capital cost
Site preparation (2% direct capital cost)
Engineering and design (10% direct capital cost)
Project contingency (15% direct capital cost)
Upfront permitting cost (legal and contractors fees) (7.5% direct capital
cost)
Total capital cost

Value in startup year dollars
$134,129,266
$2,682,585
$13,412,927
$20,119,390
$10,059,695
$180,403,863

Fixed operating costs
Labor cost ($50/FTE) ($/year) (11 FTE)
G&A ($/year) (20% labor cost)
Property taxes and insurance ($/year) (2% total capital cost)
Production maintenance and repairs ($/year) (2.9% direct capital cost)
Total fixed operating costs ($/year)

Value in startup year
$1,282,289
$256,458
$3,608,077
$3,889,749
$9,036,573

Variable operating costs
Value in startup year
Energy utilities costs
Non energy utilities costs
Total variable operating costs ($/year)

$17,823,995
$3,032,682
$20,856,677

Replacements
Value in startup year
Unplanned replacement capital cost (0.5% of total direct capital
costs/year)

$670,646

Installation
cost factor
1.86
1.70
1.77
1.43
1.16
1.12
1.07
1.76
1.06
2.68
1.06

Combined
plant scaling
0.09
0.09
0.15
0.24
0.20
0.19
0.30
0.24
0.24
0.43
0.33

Installed cost
$22,984,912
$14,167,891
$13,591,553
$15,339,052
$4,777,228
$1,667,188
$12,564,408
$8,728,853
$29,039,678
$3,445,575
$7,822,927
$134,129,266

171
Table B3: Assumptions for calculating the cost of direct air capture of CO2 in the future. For further
information see excel sheet titled “Direct Air Capture of CO2 (future)”.
Process Assumptions
Design capacity
Average production
Capacity factor
Baseline capacity
Scale ratio
Scaling factor exponent
Electricity equivalent gas input
Electricity input
Water consumption
Water price
Industrial electricity
Natural gas price

2684931.51
2550684.932
0.95
2684931.507
0.6
478.3333333
77
4.7
0.69307362
0.01
3.5

kg CO2/day
kg CO2/day
kg CO2/day

kWh/t-CO2
kWh/t-CO2
t-water/t-CO2
$/t-water
$/kWh
$/GJ

Baseline system cost
Air contactor
Pellet reactor
Calciner-slaker
Air separation unit
CO2 compressor
Steam turbine
Power plant
Fines filter
Other equipment
Buildings
Transformer
Total

Uninstalled cost in startup year
dollars
$157,996,906
$112,786,948
$75,667,193
$55,560,659
$18,440,904
$6,900,467
$31,765,944
$29,505,446
$91,609,652
$6,900,467
$19,868,587

Capital costs
Depreciable capital costs
Direct capital cost
Indirect capital cost
Site preparation (2% direct capital cost)
Engineering and design (10% direct capital cost)
Project contingency (15% direct capital cost)
Upfront permitting cost (legal and contractors fees) (7.5% direct capital cost)
Total capital cost

Value in startup year dollars
$607,003,173
$12,140,063
$60,700,317
$91,050,476
$45,525,238
$816,419,268

Fixed operating costs
Labor cost ($50/FTE) ($/year) (46.34 FTE)
G&A ($/year) (20% labor cost)
Property taxes and insurance ($/year) (2% total capital cost)
Production maintenance and repairs ($/year) (2.9% direct capital cost)
Total fixed operating costs ($/year)

Value in startup year dollars
$5,402,081
$1,080,416
$16,328,385
$17,603,092
$40,413,974

Variable operating costs
Energy utilities costs
Non energy utilities costs
Total variable operating costs ($/year)

Value in startup year dollars
$17,823,995
$3,032,682
$20,856,677

Replacements
Unplanned replacement capital cost (0.5% of total direct capital costs/year)

Value in startup year dollars
$670,646

Installation cost
factor
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00

Installed cost
$157,996,906
$112,786,948
$75,667,193
$55,560,659
$18,440,904
$6,900,467
$31,765,944
$29,505,446
$91,609,652
$6,900,467
$19,868,587
$607,003,173

172

Table B4: Assumptions for calculating the cost of ocean CO2 capture currently. For further information see
excel sheet titled “Ocean CO2 Capture (current)”.
Process Assumptions
Design capacity
Average production rate
Capacity factor
Extraction efficiency
Electrodialysis acidified stream target pH
Oceanwater target pH
Current density
Voltage
Electrodialysis energy at oceanwater target pH
CO2 stripping energy
Intake energy (0.5 bar, 5 m intake)
Pre-treatment energy (3 bar ultrafiltration, 7.6 bar nanofiltration, 0.5 bar
electrodialyzer)

246575.34
221917.8082
0.9
0.9
0.40
100
1.2
0.97736927
0.072994345
0.206827333

kg CO2/day
kg CO2/day

0.04804125

kWh/kg CO2

Pump energy (0.5 bar membrane contactor)

0.206827333

kWh/kg CO2

Industrial electricity

0.049

mA/cm2
kWh/kg CO2
kWh/kg CO2
kWh/kg CO2

$/kWh

Baseline system cost
Electrodialyzer
Pre-treatment
Gas stripping
Intake
Screening
Pumping
Total

Uninstalled cost in startup year
dollars
$15,133,812
$6,657,344
$21,691,109
$6,717,271
$22,973,169
$3,747,976

Capital costs
Depreciable capital costs
Direct capital cost
Indirect capital cost
Site preparation (2% direct capital cost)
Engineering and design (10% direct capital cost)
Project contingency (15% direct capital cost)
Upfront permitting cost (legal and contractors fees) (7.5% direct capital cost)
Total capital cost

Value in startup year dollars
$92,304,818
$1,846,096
$9,230,482
$13,845,723
$6,922,861
$124,149,981

Fixed operating costs
Labor cost ($50/FTE) ($/year) (34.95 FTE)
G&A ($/year) (20% labor cost)
Property taxes and insurance ($/year) (2% total capital cost)
Production maintenance and repairs ($/year) (2.9% direct capital cost)
Total fixed operating costs ($/year)

Value in startup year dollars
$4,074,146
$814,829
$2,483,000
$2,769,145
$10,141,120

Variable operating costs
Energy utilities costs
Non energy utilities costs
Total variable operating costs ($/year)

Value in startup year
$6,001,364
$0
$6,001,364

Replacements
Value in startup year dollars
Unplanned replacement capital cost (0.5% of total direct capital costs/year)
Replacement costs (15% of depreciable capital cost/5year)
Specified replacement cost (electrodialyzer/5year)
Specified replacement cost (60% pre-treatment/5year)
Specified replacement cost (gas stripping/10year)

$461,524.09
$18,622,497.08
$18,160,574.35
$4,793,288.03
$26,029,331.20

Installation cost
factor
1.20
1.20
1.20
1.20
1.20
1.20

Installed cost
$18,160,574
$7,988,813
$26,029,331
$8,060,725
$27,567,802
$4,497,572
$92,304,818

173

Table B5: Assumptions for calculating the cost of ocean CO2 capture in the future. For further information
see excel sheet titled “Ocean CO2 Capture (future)”.
Process Assumptions
Design capacity
Average production rate
Capacity factor
Extraction efficiency
Electrodialysis acidified stream target pH
Oceanwater target pH
Current density
Voltage
Electrodialysis energy at oceanwater target pH
CO2 stripping energy
Intake energy (0.5 bar, 5 m intake)
Pre-treatment energy (3 bar ultrafiltration, 7.6 bar nanofiltration, 0.5 bar
electrodialyzer)

2739726.03
2602739.726
0.95
0.98
0.40
1000
1.6
1.196778698
0.072994345
0.189943469

kg CO2/day
kg CO2/day

0.044119516

kWh/kg CO2

Pump energy (0.5 bar membrane contactor)

0.189943469

kWh/kg CO2

Industrial electricity

0.01

mA/cm2
kWh/kg CO2
kWh/kg CO2
kWh/kg CO2

$/kWh

Baseline system cost
Electrodialyzer
Pre-treatment
Gas stripping
Intake
Screening
Pumping
Total

Uninstalled cost in startup year
dollars
$7,721,333
$25,033,436
$73,347,986
$68,543,583
$234,420,087
$38,244,657

Capital costs
Depreciable capital costs
Direct capital cost
Indirect capital cost
Site preparation (2% direct capital cost)
Engineering and design (10% direct capital cost)
Project contingency (15% direct capital cost)
Upfront permitting cost (legal and contractors fees) (7.5% direct capital cost)
Total capital cost

Value in startup year dollars
$447,311,082
$8,946,222
$44,731,108
$67,096,662
$33,548,331
$601,633,406

Fixed operating costs
Labor cost ($50/FTE) ($/year) (126.79 FTE)
G&A ($/year) (20% labor cost)
Property taxes and insurance ($/year) (2% total capital cost)

Value in startup year dollars
$14,780,298
$2,956,060
$12,032,668

Production maintenance and repairs ($/year) (2.9% direct capital cost)

$13,419,332

Total fixed operating costs ($/year)

$43,188,359

Variable operating costs
Energy utilities costs
Non energy utilities costs
Total variable operating costs ($/year)

Value in startup year
$16,090,905
$0
$16,090,905

Replacements
Value in startup year dollars
Unplanned replacement capital cost (0.5% of total direct capital costs/year)

$2,236,555

Replacement costs (15% of depreciable capital cost/5year)
Specified replacement cost (electrodialyzer/5year)
Specified replacement cost (60% pre-treatment/5year)
Specified replacement cost (gas stripping/10year)

$60,163,341
$7,721,333
$25,033,436
$73,347,986

Installation cost
factor
1.00
1.00
1.00
1.00
1.00
1.00

Installed cost
$7,721,333
$25,033,436
$73,347,986
$68,543,583
$234,420,087
$38,244,657
$447,311,082

174

Table B6: Assumptions for calculating the cost of membrane water capture. For further information see excel
sheet titled “Membrane Water Capture”.
Process Assumptions
Design capacity
Average production rate
Capacity factor

365000
357700
98%

kg H2O/day
kg H2O/day

System cost
Equipment cost
Total

Uninstalled cost in startup year
dollars
$52,581,424

Capital costs
Value in startup year dollars
Depreciable capital costs
Direct capital cost
Indirect capital cost
Site preparation (2% direct capital cost)
Engineering and design (10% direct capital cost)
Project contingency (15% direct capital cost)
Upfront permitting cost (legal and contractors fees) (15% direct capital cost)
Total capital cost

$52,581,424
$761,944
$3,809,721
$5,714,581
$5,714,581
$68,582,251

Fixed operating costs
Labor cost ($50/FTE) ($/year) (3 FTE)
G&A ($/year) (20% labor cost)
Property taxes and insurance ($/year) (2% total capital cost)
Production maintenance and repairs ($/year) (3% direct capital cost)
Total fixed operating costs ($/year)

Value in startup year dollars
$362,980
$67,339
$1,371,645
$1,577,443
$3,379,407

Variable operating costs
Total variable operating costs ($/year)

Value in startup year dollars
$321,067

Replacements
Value in startup year dollars
Unplanned replacement capital cost (1% of total direct capital costs/year)

$525,814

Installation cost
factor
1.00

Installed cost
$52,581,424
$52,581,424

175

Table B7: Assumptions for calculating the cost of condensation water capture. For further information see
excel sheet titled “Condensation Water Capture”.
Process Assumptions
Design capacity
Average production rate
Capacity factor

365000
357700
98%

kg H2O/day
kg H2O/day

System cost
OPUR structure
Pumps
Total

Uninstalled cost in startup year
dollars
$19,993,088
$214,000

Capital costs
Value in startup year dollars
Depreciable capital costs
Direct capital cost
Indirect capital cost
Site preparation (2% direct capital cost)
Engineering and design (10% direct capital cost)
Project contingency (15% direct capital cost)
Upfront permitting cost (legal and contractors fees) (15% direct capital cost)
Total capital cost

$327,803,864
$4,750,123
$23,750,615
$35,625,923
$35,625,923
$427,556,448

Fixed operating costs
Labor cost ($50/FTE) ($/year) (3 FTE)
G&A ($/year) (20% labor cost)
Property taxes and insurance ($/year) (2% total capital cost)
Production maintenance and repairs ($/year) (3% direct capital cost)
Total fixed operating costs ($/year)

Value in startup year dollars
$362,980
$67,339
$8,551,129
$9,834,116
$18,815,564

Variable operating costs
Value in startup year dollars
Total variable operating costs ($/year)

$0

Replacements
Value in startup year dollars
Unplanned replacement capital cost (1% of total direct capital costs/year)

$3,278,039

Installation cost
factor
1.00
1.00

Installed cost
$19,993,088
$214,000
$1,273,210,134

176

Table B8: Assumptions for calculating the cost of low temperature electrolysis for H2 production. For further
information see excel sheet titled “Low Temperature Electrolysis H2”.
Process Assumptions
Design capacity
Current Density
Voltage
Baseline design capacity
Degradation Rate
Cell/stack
Stack Life
Hours per stack life
Degradation Rate
Stack oversize due to degradation
Peak production rate
Capacity factor
Average production
Total Active Area
Total Active Area (with degradation)
Total System Electrical Usage
Stack Electrical Usage
BoP Electrical Usage
Total System Input Power (Peak)
Stack Input Power (peak)
Process Water Flow Rate
Total System Cost
Stack System Cost
Mechanical BoP
Electrical BoP
Industrial electricity
Processed water

Baseline system cost
Stack capital cost
Mechanical BoP
Electrical BoP
Total

40000
1.9
50000
1.5
150
59480.4
89.2206
0.13
45200
0.97
43844
2212
2499
55.5
50.4
5.1
104.525
94.92
3.78
460
1.3
76.00
82
0.049
0.00263368

Uninstalled cost in startup year
dollars
$38,650,945
$4,086,980
$9,260,237

Capital costs
Depreciable capital costs
Direct capital cost
Indirect capital cost
Site preparation (2% direct capital cost)
Engineering and design (10% direct capital cost)
Project contingency (15% direct capital cost)
Upfront permitting cost (legal and contractors fees) (15% direct capital cost)
Non depreciable capital costs
Cost of land (5 acre, $50,000/acre) (2016 dollars)
Total capital cost

$57,747,504
$1,154,950
$5,774,750
$8,662,126
$8,662,126
$245,459
$82,246,914

Fixed operating costs
Labor cost ($50/FTE) ($/year) (8.75 FTE)
G&A ($/year) (20% labor cost)
Property taxes and insurance ($/year) (2% total capital cost)
Production maintenance and repairs ($/year) (3% direct capital cost)
Total fixed operating costs ($/year)

$1,019,641
$203,928
$1,644,938
$1,732,425
$4,600,932

Variable operating costs
Energy utilities costs
Non energy utilities costs
Total variable operating costs ($/year)

$43,520,322
$176,671
$43,696,993

Replacements
Unplanned replacement capital cost (0.5% of total direct capital costs/year)
Replacement costs (15% of depreciable capital cost/7year)

$410,007
$8,662,126

kg H2/day
A/cm2
V/cell
kg H2/day
mV/1000 hrs
years
hrs/life
V/life
kg H2/day
kg H2/day
m2
m2
kWh/kg H2
kWh/kg H2
kWh/kg H2
MW
MW
gal/kg H2
$/kW
$/cm2
$/(kg H2/day)
$/kW
$/kWh
$(2016)/gal

Installation cost
factor
1.12
1.00
1.12

Installed cost
$43,289,058
$4,086,980
$10,371,465
$57,747,504

177

Table B9: Assumptions for calculating the cost of high temperature electrolysis for H2 production. For further
information see excel sheet titled “High Temperature Electrolysis H2”.
Process Assumptions
Design capacity
Capacity factor
Average production
Total System Electrical Usage
Electrical usage
Heat usage
Heat usage
Process Water Flow Rate
Total System Cost
Stack Cost (% of uninst. SOEC Sys Cost )
BoP Cost (% of uninst. SOEC Sys Cost )
Industrial natural gas
Industrial electricity
Processed water

40000
0.8244
32976
50.9
36.8
14.1
0.05613
2.384702106
820
0.35
0.65
3.733674814
0.049
0.00263368

kg H2/day
kg H2/day
kWh/kg H2
kWh/kg H2
kWh/kg H2
mmBtu/kg H2
gal/kg H2
$/kW

$/mmbtu
$/kWh
$(2016)/gal

Baseline system cost
Stack capital cost
BoP
Total

Uninstalled cost in startup year
dollars
$38,650,945
$4,086,980

Capital costs
Value in startup year dollars
Depreciable capital costs
Direct capital cost
Indirect capital cost
Site preparation (2% direct capital cost)
Engineering and design (10% direct capital cost)
Project contingency (15% direct capital cost)
Upfront permitting cost (legal and contractors fees) (15% direct capital cost)
Non depreciable capital costs
Cost of land (1 acre, $50,000/acre) (2016 dollars)
Total capital cost

$67,016,069
$1,340,321
$6,701,607
$10,052,410
$10,052,410
$49,092
$95,211,910

Fixed operating costs
Labor cost ($50/FTE) ($/year) 13.12 FTE)
G&A ($/year) (20% labor cost)
Property taxes and insurance ($/year) (2% total capital cost)
Production maintenance and repairs ($/year) (3% direct capital cost)
Total fixed operating costs ($/year)

Value in startup year dollars
$1,529,461
$305,892
$1,904,238
$2,010,482
$5,750,074

Variable operating costs
Energy utilities costs
Non energy utilities costs
Total variable operating costs ($/year)

Value in startup year
$25,633,493
$83,829
$25,717,323

Replacements
Value in startup year dollars
Unplanned replacement capital cost (1% of total direct capital costs/year)
Specified replacement costs (27.3% of stack cost/year)
Specified replacement costs (100% of BoP cost/15 year)

$951,628
$6,403,385
$43,560,445

Installation cost
factor
1.12
1.12

Installed cost
$23,455,624
$43,560,445
$67,016,069

178

Table B10: Assumptions for calculating the cost of photoelectrochemical H2 production currently. For further
information see excel sheet titled “Photoelectrochemical H2 (current)”.
Process Assumptions
Design capacity
Average production rate
Location
PV efficiency
STH efficiency
Solar power
Solar capacity factor
PV module
Solar capture area for stack only
Solar capture area for BoP (10% of stack)
Total solar capture area
PV cells
Catalyst
Membrane
Chassis
Water processing
Gas processing
Power electronics and control system
Process Water Flow Rate
Processed water

40000
40000

kg H2/day
kg H2/day

California
19.1%
10%
1000
28.4%
0.37
1914218
191421.8
2105639.8
70.67
30
37.74869216
0.76
6.599227784
7.908190728
3.78
0.00263368

W/m2
$/W
m2
m2
m2
$/m2
$/m2
$/m2
$/m2
$/m2
$/m2
$/m2
gal/kg H2
$(2016)/gal

Baseline system cost
PV capital cost
Stack capital cost
Mechanical BoP
Electrical BoP
Total

Uninstalled cost in startup year
dollars
$148,805,565
$159,499,461
$15,487,230
$16,651,801

Capital costs
Value in startup year dollars
Depreciable capital costs
Direct capital cost
Indirect capital cost
Site preparation (2% direct capital cost)
Engineering and design (10% direct capital cost)
Project contingency (15% direct capital cost)
Upfront permitting cost (legal and contractors fees) (15% direct capital cost)
Non depreciable capital costs
Cost of land (520 acre, $50,000/acre) (2016 dollars)
Total capital cost

$406,223,878
$8,124,478
$40,622,388
$60,933,582
$60,933,582
$26,015,706
$602,853,613

Fixed operating costs
Labor cost ($50/FTE) ($/year) (20.67 FTE)
G&A ($/year) (20% labor cost)
Property taxes and insurance ($/year) (2% total capital cost)
Production maintenance and repairs ($/year) (3% direct capital cost)
Total fixed operating costs ($/year)

Value in startup year dollars
$2,409,559
$481,912
$12,057,072
$12,186,716
$27,135,260

Variable operating costs
Value in startup year
Energy utilities costs
Non energy utilities costs
Total variable operating costs ($/year)

$0
$145,348
$145,348

Replacements
Value in startup year dollars
Unplanned replacement capital cost (0.5% of total direct capital costs/year)
Replacement costs (15% of depreciable capital cost/7year)
Specified replacement cost (PV stack cost/20 year)

$2,031,119
$86,525,686
$193,447,234

Installation cost
factor
1.30
1.12
1.00
1.12

Installed cost
$193,447,234
$178,639,396
$15,487,230
$18,650,017
$406,223,878

179
Table B11: Assumptions for calculating the cost of photoelectrochemical H2 production in the future. For
further information see excel sheet titled “Photoelectrochemical H2 (future)”.
Process Assumptions
Design capacity
Average production rate
Location
PV efficiency
STH efficiency
Solar power
Solar capacity factor
PV module
Solar capture area for stack only
Solar capture area for BoP (10% of stack)
Total solar capture area
PV cells
Catalyst
Membrane
Chassis
Water processing
Gas processing
Power electronics and control system
Process Water Flow Rate
Processed water

40000
40000

kg H2/day
kg H2/day

California
19.1%
10.0%
1000
28.40%
0.24
957109
95710.9
1052819.9
70.67
30
37.74869216
0.76
6.599227784
7.908190728
3.78
0.00263368

W/m2
$/W
m2
m2
m2
$/m2
$/m2
$/m2
$/m2
$/m2
$/m2
$/m2
gal/kg H2
$(2016)/gal

Baseline system cost
PV capital cost
Stack capital cost
Mechanical BoP
Electrical BoP
Total

Uninstalled cost in startup year
dollars
$148,805,565
$159,499,461
$15,487,230
$16,651,801

Capital costs
Value in startup year dollars
Depreciable capital costs
Direct capital cost
Indirect capital cost
Site preparation (2% direct capital cost)
Engineering and design (10% direct capital cost)
Project contingency (15% direct capital cost)
Upfront permitting cost (legal and contractors fees) (15% direct capital cost)
Non depreciable capital costs
Cost of land (520 acre, $50,000/acre) (2016 dollars)
Total capital cost

$406,223,878
$8,124,478
$40,622,388
$60,933,582
$60,933,582
$26,015,706
$602,853,613

Fixed operating costs
Labor cost ($50/FTE) ($/year) (20.67 FTE)
G&A ($/year) (20% labor cost)
Property taxes and insurance ($/year) (2% total capital cost)

Value in startup year dollars
$2,409,559
$481,912
$12,057,072

Production maintenance and repairs ($/year) (3% direct capital cost)

$12,186,716

Total fixed operating costs ($/year)

$27,135,260

Variable operating costs
Value in startup year
Energy utilities costs
Non energy utilities costs
Total variable operating costs ($/year)

$0
$145,348
$145,348

Replacements
Value in startup year dollars
Unplanned replacement capital cost (0.5% of total direct capital costs/year)
Replacement costs (15% of depreciable capital cost/7year)
Specified replacement cost (PV stack cost/20 year)

$2,031,119
$86,525,686
$193,447,234

Installation cost
factor
1.30
1.12
1.00
1.12

Installed cost
$193,447,234
$178,639,396
$15,487,230
$18,650,017
$406,223,878

180
Table B12: Assumptions for calculating the cost of solar thermochemical H2. For further information see excel
sheet titled “Solar Thermochemical H2”.
Process Assumptions
Plant design capacity
Baseline design capacity
Plant capacity factor
Average production rate
STH efficiency
Process Water Flow Rate
Scale ratio
Scaling factor exponent
Lower limit for scaling capacity
Upper limit for scaling capacity
Processed water price

45000
100000
90.00%
40500
20%
2.378
0.45
0.78
20000
200000
0.00263368

kg H2/day
kg H2/day
kg H2/day
gal/kg H2

kg H2/day
kg H2/day
$(2016)/gal

System cost
Baseline uninstalled cost in
reference year dollars
ZrO2
Compression System
Solar Reactors
Vacuum Pumps
Water Pumps
Turbine
Heat Exchangers
Heliostats
Secondary Concentrators
Towers
Ferrite
Total

$50,604
$36,614,848
$27,852,508
$4,850,000
$97,785
$693,223
$378,081
$193,334,841
$738,859
$72,990,255
$50,999,508
$388,600,513

Capital costs
Value in startup year dollars
Depreciable capital costs
Direct capital cost
Indirect capital cost
Site preparation (2% direct capital cost)
Engineering and design (17.8% solar subsystem cost)
Project contingency (16.8% solar and 18% chemical system cost)
Upfront permitting cost (legal and contractors fees) (7.5% direct capital cost)
Non depreciable capital costs
Cost of land (374.44 acre, $50,000/acre) (2016 dollars)
Total capital cost

$294,872,293
$7,016,402
$36,094,585
$60,714,276
$26,311,508
$1,842,960
$426,852,024

Fixed operating costs
Labor cost ($50/FTE) ($/year) (239.07 FTE)
G&A ($/year) (20% labor cost)
Property taxes and insurance ($/year) (2% total capital cost)
Production maintenance and repairs ($/year) (0.5% solar + 6% nonsolar)
Total fixed operating costs ($/year)

Value in startup year dollars
$4,554,227
$910,845
$8,537,040
$8,318,137
$22,320,251

Variable operating costs
Energy utilities costs
Non energy utilities costs
Environmental surcharges
Total variable operating costs ($/year)

$0
$92,581
$1,208,144
$1,300,725

Replacements
Value in startup year dollars
Unplanned replacement capital cost (0.5% of total direct capital costs/year)

$2,125,045

Specified replacement costs (ZrO2 and Ferrite/5 year)

$32,580,141

Scaled uninstalled
cost in reference
year dollars
$27,145
$19,640,973
$14,940,670
$2,601,642
$52,454
$371,860
$202,811
$103,708,866
$396,340
$39,153,505
$27,357,206

Scaled uninstalled
cost in startup year
dollars
$32,296
$23,367,567
$17,775,449
$3,095,266
$62,406
$442,415
$241,291
$123,386,145
$471,539
$46,582,324
$32,547,846

Installation
cost factor
3.17
3.3
3.3
2.15
3.17

Scaled installed
cost
$32,296
$23,367,567
$56,348,173
$10,214,379
$205,940
$951,191
$764,893
$123,386,145
$471,539
$46,582,324
$32,547,846
$294,872,293

181
Table B13: Assumptions for calculating the cost of biochemical methanation. For further information see excel
sheet titled “Biochemical Methanation”.
Process Assumptions
Design capacity
Average production
Capacity factor
H2 input rate
CO2 input rate
H2 price (LTE H2)
CO2 price (DAC CO2)

85714
84000
98%
42857
235714
3.50
0.28

kg SNG/day
kg SNG/day
kg/day
kg/day
$/kg
$/kg

System cost
Compressor
Methanation
Piping installation
Total

Uninstalled cost in reference year
dollars
$5,299,039
$15,799,014

Capital costs
Value in startup year dollars
Depreciable capital costs
Direct capital cost
Indirect capital cost
Site preparation (2% direct capital cost)
Engineering and design (10% direct capital cost)
Project contingency (15% direct capital cost)
Upfront permitting cost (legal and contractors fees) (15% direct capital cost)
Total capital cost

$28,960,870
$506,353
$2,531,766
$3,797,650
$3,797,650
$39,594,290

Fixed operating costs
Labor cost ($50/FTE) ($/year) (12 FTE)
G&A ($/year) (20% labor cost)
Property taxes and insurance ($/year) (2% total capital cost)
Production maintenance and repairs ($/year) (3% direct capital cost)
Total fixed operating costs ($/year)

Value in startup year dollars
$1,449,927
$268,985
$791,886
$868,826
$3,379,624

Variable operating costs
Value in startup year dollars
Energy utilities costs
H2 cost
CO2 cost
Total variable operating costs ($/year)

$0
$53,596,721
$23,448,298
$53,596,721

Replacements
Value in startup year dollars
Unplanned replacement capital cost (1% of total direct capital costs/year)

$289,609

Uninstalled cost in
startup year dollars
$6,061,570
$18,072,489

Installation
cost factor
1.00
1.00

Installed cost
$6,061,570
$18,072,489
$4,826,812
$28,960,870

182

Table B14: Assumptions for calculating the cost of thermochemical methanation. For further information see
excel sheet titled “Thermochemical Methanation”.
Process Assumptions
Design capacity
Average production
Capacity factor
H2 input rate
CO2 input rate
H2 price (LTE H2)
CO2 price (DAC CO2)

81100
79478
98%
40000
218200
3.50
0.28

kg SNG/day
kg SNG/day
kg/day
kg/day
$/kg
$/kg

System cost
Heat exchangers
Reactors
Catalyst (Ni supported)
Compressors
MDEA unit
Membrane unit
Organic Rankine Cycle Unit
Pumps
Cooling tower
Piping installation
Total

Uninstalled cost in reference year
dollars
1957000
2919000
318000
5126000
1049000
263000
1100000
$214,000
$555,000

Capital costs
Values in startup year dollars
Depreciable capital costs
Direct capital cost
Indirect capital cost
Site preparation (2% direct capital cost)
Engineering and design (10% direct capital cost)
Project contingency (15% direct capital cost)
Upfront permitting cost (legal and contractors fees) (15% direct capital cost)
Total capital cost

$28,233,144
$564,663
$2,823,314
$4,234,972
$4,234,972
$40,091,064

Fixed operating costs
Labor cost ($50/FTE) ($/year) (12 FTE)
G&A ($/year) (20% labor cost)
Property taxes and insurance ($/year) (2% total capital cost)
Production maintenance and repairs ($/year) (3% direct capital cost)
Total fixed operating costs ($/year)

Values in startup year dollars
$1,398,861
$279,772
$801,821
$846,994
$3,327,448

Variable operating costs
Values in startup year
Energy utilities costs
H2 cost
CO2 cost
Total variable operating costs ($/year)

$0
$50,023,606
$21,706,018
$71,729,624

Replacements
Values in startup year dollars
Unplanned replacement capital cost (1% of total direct capital costs/year)

$282,331

Uninstalled cost in
startup year dollars
2701033.858
4028777.635
438900.7496
7074859.252
1447820.397
362990.2426
1518210.14
$295,361
$766,006

Installation cost
factor
2.47
2.47
2.47
1.00
1.00

Installed cost
2701033.858
4028777.635
438900.7496
7074859.252
3576116.381
896585.8991
3749979.046
$295,361
$766,006
$4,705,524
$28,233,144

183

Table B15: Assumptions for calculating the cost of electrochemical methanation. For further information see
excel sheet titled “Low Temperature electrochemical CH4”.
Process Assumptions
Design capacity
Current Density
Voltage
Cell/stack
Degradation Rate
Stack Life
Degradation Rate
Stack oversize due to degradation
Peak production rate
Capacity factor
Average production
Total Active Area
Total Active Area (with degradation)
Total System Electrical Usage
Stack Electrical Usage
BoP Electrical Usage
Total System Input Power (Peak)
Stack Input Power (peak)
Utilization
Process Water Flow Rate
Process CO2 Flow Rate
Total System Cost
Stack System Cost
Mechanical BoP
Electrical BoP
Industrial electricity
Processed water
CO2 (DAC)

81000
0.5
150
0.5
30.66
0.045
84645
0.97
82105.65
10642
11121
65.61
63.06
2.55
231.41
222.41
100%
1.89
2.75
460
1.3
38.00
82
0.049
0.00263368
0.28

kg CH4/day
A/cm2
V/cell
mV/1000 hrs
years
V/life
kg CH4/day
kg CH4/day
m2
m2
kWh/kg CH4
kWh/kg CH4
kWh/kg CH4
MW
MW
gal/kg H4
kg/kg H4
$/kW
$/cm2
$/(kg CH4/day)
$/kW
$/kWh
$(2016)/gal
$/kg CH4

System cost
Uninstalled cost in startup year dollars
Stack capital cost
Mechanical BoP
Electrical BoP
Total

$172,003,665
$3,826,797
$21,698,172

Capital costs
Depreciable capital costs
Direct capital cost
Indirect capital cost
Site preparation (2% direct capital cost)
Engineering and design (10% direct capital cost)
Project contingency (15% direct capital cost)
Upfront permitting cost (legal and contractors fees) (15% direct capital cost)
Non depreciable capital costs
Cost of land (4.4 acre, $50,000/acre) (2016 dollars)
Total capital cost

$220,772,854
$4,415,457
$22,077,285
$33,115,928
$33,115,928
$313,497,452
$247,296
$313,744,748

Fixed operating costs
Labor cost ($50/FTE) ($/year) (8.81 FTE)
G&A ($/year) (20% labor cost)
Property taxes and insurance ($/year) (2% total capital cost)
Production maintenance and repairs ($/year) (3% direct capital cost)
Total fixed operating costs ($/year)

$1,027,269
$205,454
$6,274,895
$6,623,186
$14,130,803

Variable operating costs
Energy utilities costs
CO2 cost
H2O cost
Total variable operating costs ($/year)

$96,348,776
$22,919,497
$165,424
$119,433,698

Replacements
Unplanned replacement capital cost (0.5% of total direct capital costs/year)
Replacement costs (15% of depreciable capital cost/7year + unplanned cost)

$1,567,487
$33,115,928

Installation cost factor
1.12
1.00
1.12

Installed cost
$192,644,104
$3,826,797
$24,301,952
$220,772,854

184

Table B16: Assumptions for calculating the cost of photoelectrochemical methanation. For further information
see excel sheet titled “Photoelectrochemical CH4”.
Process Assumptions
Design capacity
Average production rate
Location
PV efficiency
STH efficiency
Solar power
Solar capacity factor
PV module
Solar capture area for stack only
Solar capture area for BoP (10% of stack)
Total solar capture area
PV cells
Catalyst
Membrane
Chassis
Water processing
Gas processing
Power electronics and control system
Utilization
Process CO2 Flow Rate
Process Water Flow Rate
Processed water
CO2 (DAC)

81000
81000

kg CH4/day
kg CH4/day

California
19.1%
6.0%
1000
28.40%
0.37
6460484
646048.4
7106532.4
70.67
30
37.75
0.23
1.71
0.46
100%
2.75
3.78
0.00263368
0.28

W/m2
$/W
m2
m2
m2
$/m2
$/m2
$/m2
$/m2
$/m2
$/m2
$/m2
kg CO2/kg CH4
gal/kg CH4
$(2016)/gal
$(2016)/gal

System cost
PV capital cost
Stack capital cost
Mechanical BoP
Electrical BoP
Total

Uninstalled cost in startup year dollars

Installation cost factor

$502,218,645
$538,310,535
$13,790,861
$3,238,603

1.30
1.12
1.00
1.12

Capital costs
Depreciable capital costs
Direct capital cost
Indirect capital cost
Site preparation (2% direct capital cost)
Engineering and design (10% direct capital cost)
Project contingency (15% direct capital cost)
Upfront permitting cost (legal and contractors fees) (15% direct capital cost)
Non depreciable capital costs
Cost of land (520 acre, $50,000/acre) (2016 dollars)
Total capital cost

$1,273,210,134
$25,464,203
$127,321,013
$190,981,520
$190,981,520
$87,802,984
$1,895,761,374

Fixed operating costs
Labor cost ($50/FTE) ($/year) (20.67 FTE)
G&A ($/year) (20% labor cost)
Property taxes and insurance ($/year) (2% total capital cost)
Production maintenance and repairs ($/year) (3% direct capital cost)
Total fixed operating costs ($/year)

$12,803,772
$2,560,754
$37,915,227
$38,196,304
$91,476,057

Variable operating costs
Energy utilities costs
CO2 cost
H2O cost
Total variable operating costs ($/year)

$0
$22,610,859
$147,164
$22,758,023

Replacements
Unplanned replacement capital cost (0.5% of total direct capital costs/year)
Replacement costs (15% of depreciable capital cost/7year)
Specified replacement cost (PV stack cost/20 year)

$6,366,051
$271,193,758
$652,884,238

Installed cost
$652,884,238
$602,907,799
$13,790,861
$3,627,235
$1,273,210,134

185

Figure B1: Sensitivity analysis of low temperature electrolysis processes.

186

Solar H2 generation cost breakdown
$5,294

Axis Title

Cost ($/ton-H2)

$5,000

$4,000

27
864

855
205
10
169

$3,000
946

$2,000

316
160

291

836

294

906

215
158
157
325

current

future

$1,000

$0

$1,775
156
10
85
84

Figure B2: Cost breakdown for current and future cost of photoelectrochemical water splitting.

Decommissioning
O&M
Property taxes and insurance
Labor, G&A
Feedstock
Other capital related costs
Stack and BoP replacements
PV replacements
Mechanical BoP
Stack
PV

187

a.

Electrochemical

b.

Photoelectrochemical

Figure B3: Sensitivity analysis of (a) electrochemical and (b) photoelectrochemical methanation processes. The base case for the
methanation processes assumes CO2 captured from the atmosphere and H2 generated via low temperature electrolysis.

188

a.
c.

Thermochemical

b.

Biochemical

Figure B4: Sensitivity analysis of (a) thermochemical and (b) biochemical methanation processes. The base case for the methanation
processes assumes CO2 captured from the atmosphere and H2 generated via low temperature electrolysis.

Temp
Press
Capa
10 M
footp
Cont
Toler
Large
(milli

189

BIBLIOGRAPHY: APPENDIX B
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.

Keith, D. W., Angelo, D. St., Holmes, G. & Heidel, K. A process for capturing CO2 from the atmosphere.
Joule 2018, 2, 1573–1594.
James, B. D., Colella, W. G. & Moton, J. M. Techno-Economic Analysis of Hydrogen Production Pathways. Stategic
Analysis (2013).
NREL. H2A: Hydrogen Analysis Production Models. (2018).
Peterson, D. & Miller, E. Hydrogen and Fuel Cells Program Record - Hydrogen Production Cost from Solid Oxide
Electrolysis. Hydrogen and Fuel Cells Program Record - Hydrogen Production Cost from Solid Oxide Electrolysis (2016).
Becker, W. L., Penev, M. & Braun, R. J. Production of synthetic natural gas from carbon dioxide and
renewably generated hydrogen: A techno-economic analysis of a power-to-gas strategy. J. Energy Resour.
Technol. 2019, 141.
de Lannoy, C. F. et al. Indirect ocean capture of atmospheric CO2: Part I. Prototype of a negative
emissions technology. Int. J. Greenh. Gas Control 2018, 70, 243–253.
Eisaman, M. D. et al. Indirect ocean capture of atmospheric CO2: Part II. Understanding the cost of
negative emissions. Int. J. Greenh. Gas Control 2018, 70, 254–261.
Matsumiya, N., Teramoto, M., Kitada, S. & Matsuyama, H. Evaluation of energy consumption for
separation of CO2 in flue gas by hollow fiber facilitated transport membrane module with permeation of
amine solution. Sep. Purif. Technol. 2005, 46, 26–32.
Voutchkov, N. Desaliniation Engineering Planning and Design. (McGraw-Hill Companies, Inc, 2013).
Seider, W. D., Seader, J. D., Lewin, D. R. & Widagdo, S. Product and Process Design. (John Wiley & Sons,
2009).
Shaner, M. R., Atwater, H. A., Lewis, N. S. & McFarland, E. W. A comparative technoeconomic analysis
of renewable hydrogen production using solar energy. Energy Environ. Sci. 2016, 9, 2354–2371.
James, B. D., Baum, G. N., Perez, J. & Baum, K. N. Discovery of viable methanotrophic bacteria in permafrost
sediments of northeast Siberia. (2009).
Electrochaea. Power-to-Gas via Biological Catalysis (P2G-Biocat). (2017).
Electrochaea. Applications of Electrochaea’s BioCat biomethanation technology. (2018).