conjugated organic polymers as photocathode materials in...
TRANSCRIPT
Conjugated Organic Polymers as Photocathode
Materials in Organic Photoelectrochemical Cells
by
Patrick Fortin
B.Sc., University of Alberta, 2013
Thesis Submitted in Partial Fulfillment of the
Requirements for the Degree of
Doctor of Philosophy
in the
Department of Chemistry
Faculty of Science
© Patrick Fortin 2019
SIMON FRASER UNIVERSITY
Spring 2019
Copyright in this work rests with the author. Please ensure that any reproduction or re-use is done in accordance with the relevant national copyright legislation.
ii
Approval
Name:
Degree:
Title:
Examining Committee:
Date Defended/Approved:
Patrick Fortin
Doctor of Philosophy (Chemistry)
Conjugated Organic Polymers as Photocathode Materials in Organic Photoelectrochemical Cells
Chair: Corina Andreoiu Associate Professor
Steven Holdcroft Senior Supervisor Professor
Byron D. Gates Supervisor Associate Professor
Loren Kaake Supervisor Assistant Professor
Vance E. Williams Internal Examiner Associate Professor
Timothy L. Kelly External Examiner Associate Professor Department of Chemistry University of Saskatchewan
March 29, 2019
iii
Abstract
The work presented in this thesis aims at studying the photoelectrochemical
properties of conjugated organic polymers, with a particular interest in their application as
photocathode materials to perform the hydrogen evolution reaction. Whereas inorganic
semiconductors have been the primary focus in the area of solar water splitting, organic
semiconductors have only recently emerged as a new class of photoelectrode material.
Furthermore, the research emphasis on organic semiconductor systems used for water
splitting applications has largely been focused on improving performance through device
engineering. Reports concerning the interfacial thermodynamic and electronic processes
are less frequent. Given the difference in photophysical properties compared to inorganic
semiconductors and the change in chemical environment compared to solid-state organic
electronics, the characterization and elucidation of interfacial processes at the organic
semiconductor-electrolyte remains an area of need.
The ability of uncatalyzed P3HT photocathodes to perform the hydrogen evolution
reaction in acidic aqueous media is first investigated. Whereas previous reports in the
literature have attributed the aqueous photoactivity of P3HT to the hydrogen evolution
reaction, this study reveals that residual dissolved oxygen is largely responsible for the
observed photocurrents and that, despite favorable thermodynamics, hydrogen is not
evolved at P3HT photocathodes in the absence of a catalyst.
Following the initial investigation of P3HT photocathode performance, strategies
to improve photocurrent densities at organic photocathodes are explored. In these studies,
nanostructured photocathodes are prepared from P3HT:PCBM nanoparticles. The
optoelectronic and morphological properties, as well as the photoelectrochemical
performance of the nanostructured photocathodes are compared to planar P3HT:PCBM
photocathodes. To achieve hydrogen production, platinum nanoparticle catalysts are
deposited onto the organic layer photoelectrochemically, where the increased surface
area of the nanostructured electrodes leads to enhanced catalyst loadings and increased
photocurrent densities compared to the planar photocathodes.
Finally, the influence of PCBM on interfacial energy alignment of the redox couple
at the semiconductor-electrolyte interface is investigated in a non-aqueous electrolyte
using a benzoquinone redox couple. Photoelectrochemical measurements show that the
presence of PCBM at the semiconductor-electrolyte interface leads to the formation of an
interfacial dipole layer and decreases the built-in potential developed at the
semiconductor-electrolyte interface.
iv
Keywords: conjugated organic polymers, photoelectrochemistry, water splitting,
hydrogen evolution, nanoparticles, P3HT, P3HT:PCBM
v
Dedication
To Kristine, Shadow and Belle
vi
Acknowledgements
I would like to thank Prof. Steven Holdcroft for the opportunity to be a part of his
research group. His guidance has been instrumental in my growth as a scientist during
my time at SFU. I would also like to thank my supervisory committee members, Prof. Loren
Kaake and Prof. Byron Gates, for their insightful comments and constructive feedback
over the course of my project.
I would like to thank all the present and former members of the Holdcroft group
that I have worked with over the years. Special thanks need to be given to Dr. Graeme
Suppes for all the knowledge and technical expertise he has shared with me, as well as
Dr. Pankaj Chowdhury and Mr. Subash Rajasekar for all the great technical discussions
we have had.
I would like to thank the SFU Department of Chemistry and 4D Labs staff who have
helped me along the way: everyone at the electronics and machine shop for helping to
keep my experiments running, Mr. Bruce Harwood for fabricating the glass
photoelectrochemical cell, and Dr. Xin Zhang for the extensive training and wealth of
knowledge regarding electron microscopy.
vii
Table of Contents
Approval .......................................................................................................................... ii
Abstract .......................................................................................................................... iii
Dedication ....................................................................................................................... v
Acknowledgements ........................................................................................................ vi
Table of Contents .......................................................................................................... vii
List of Tables ................................................................................................................... x
List of Figures................................................................................................................. xi
List of Acronyms, Symbols and Abbreviations .............................................................. xvi
Chapter 1. Introduction .............................................................................................. 1
1.1. Hydrogen as an Alternative Fuel ............................................................................ 1
1.2. Photoelectrolysis of Water ..................................................................................... 4
1.2.1. Fundamentals of Semiconductor Photoelectrochemistry ............................... 4
1.2.2. The Hydrogen Evolution and Oxygen Evolution Reactions ............................ 9
1.2.3. Device Architecture and Design Principles .................................................. 10
Semiconductor Photocatalyst Systems .................................................................. 11
Photoelectrochemical Cells .................................................................................... 13
Strategies to Increase Performance ....................................................................... 18
1.3. Conjugated Organic Polymers ............................................................................. 21
1.3.1. Overview ..................................................................................................... 21
1.3.2. Electronic Processes in Organic Semiconductors ........................................ 23
1.3.3. History ......................................................................................................... 25
1.3.4. P3HT ........................................................................................................... 28
1.4. Organic Photoelectrochemical Cells .................................................................... 31
1.5. Photoelectrode Performance and Experimental Methods .................................... 34
1.5.1. Benchmark and Diagnostic Efficiencies ....................................................... 34
Solar-to-Hydrogen Conversion Efficiency ............................................................... 34
Applied Bias Photon-to-Current Efficiency (ABPE) ................................................. 35
Incident Photon-to-Current Efficiency (IPCE) ......................................................... 36
Absorbed Photon-to-Current Efficiency (APCE) ..................................................... 37
1.5.2. Experimental Methods ................................................................................. 38
Photoelectrochemical Cell ...................................................................................... 38
Photocurrent .......................................................................................................... 39
Faradaic Efficiency ................................................................................................. 40
Built-In Potential ..................................................................................................... 40
Ultraviolet-Visible Spectroscopy ............................................................................. 41
Photoluminescence Spectroscopy ......................................................................... 42
Grazing Incidence Wide Angle X-Ray Scattering ................................................... 42
Dynamic Light Scattering ....................................................................................... 44
1.6. Thesis Scope ....................................................................................................... 45
viii
Chapter 2. On the Origin of Photocurrent at Uncatalyzed P3HT Films in Aqueous Acidic Media ...................................................................................................... 47
2.1. Introduction .......................................................................................................... 47
2.2. Experimental ....................................................................................................... 48
2.2.1. Materials ...................................................................................................... 48
2.2.2. Film Preparation .......................................................................................... 49
2.2.3. Photoelectrochemistry ................................................................................. 49
2.2.4. Hydrogen Gas Detection ............................................................................. 50
2.2.5. Dissolved Oxygen Detection ........................................................................ 50
2.3. Results ................................................................................................................ 50
2.3.1. P3HT photocurrent vs. O2 concentration ...................................................... 50
2.4. Conclusion........................................................................................................... 53
Chapter 3. Hydrogen Evolution at Conjugated Polymer Nanoparticle Electrodes 55
3.1. Introduction .......................................................................................................... 55
3.2. Experimental ....................................................................................................... 58
3.2.1. Materials ...................................................................................................... 58
3.2.2. P3HT:PCBM Nanoparticle Films .................................................................. 58
3.2.3. Deposition of Pt ........................................................................................... 59
3.2.4. Photoelectrochemistry ................................................................................. 59
3.2.5. Gas Chromatography .................................................................................. 60
3.2.6. Electron Microscopy .................................................................................... 60
3.2.7. Spectroscopy ............................................................................................... 60
3.2.8. Grazing Incidence Wide Angle X-ray Scattering .......................................... 61
3.2.9. Dynamic Light Scattering ............................................................................. 61
3.3. Results and Discussion ....................................................................................... 61
3.3.1. Nanoparticle Synthesis ................................................................................ 61
3.3.2. Characterization of Nanoparticles ................................................................ 63
3.3.3. Photoelectrochemistry of Nanoparticle-based Photocathodes ..................... 68
3.3.4. Conclusion ................................................................................................... 71
Chapter 4. Energy Level Alignment and Interfacial Dipole Layer Formation at the P3HT:PCBM-Electrolyte Interface in Organic Photoelectrochemical Cells ... 74
4.1. Introduction .......................................................................................................... 74
4.2. Experimental ....................................................................................................... 76
4.2.1. Materials ...................................................................................................... 76
4.2.2. P3HT:PCBM Films ...................................................................................... 76
4.2.3. Photoelectrochemistry ................................................................................. 76
4.3. Results and Discussion ....................................................................................... 77
4.3.1. Open Circuit Potential & Onset Potential for Photocurrent ........................... 77
4.3.2. Role of a PEDOT:PSS Charge Selective Layer ........................................... 80
4.4. Conclusion........................................................................................................... 84
Chapter 5. Conclusions and Future Directions ...................................................... 86
ix
5.1. Conclusions ......................................................................................................... 86
5.2. Future Directions ................................................................................................. 87
5.2.1. Rational Polymer Design ............................................................................. 87
Conjugated Organic Polymers with High Dielectric Constants ................................ 88
Band-Edge Engineering ......................................................................................... 90
5.2.2. Device Engineering ..................................................................................... 93
Polymer Orientation ............................................................................................... 93
Nanostructured Photoelectrodes ............................................................................ 95
References ................................................................................................................... 98
x
List of Tables
Table 3-1 Solubility of P3HT & PCBM in various solvents36 and solvent properties. .............................................................................................. 62
xi
List of Figures
Figure 1.1 Relative Fermi level (EF) positions of a) intrinsic, b) p-type and c) n-type semiconductors. Also shown are the valence band (VB) and conduction band (CB) levels..................................................................... 6
Figure 1.2 The relative energy levels at the semiconductor-electrolyte interface before (left) and after (right) equilibration for a) p-type and b) n-type semiconductors. ECB, EVB, EF and Eredox represent the positions of the conduction band, valence band, Fermi level and redox potential, respectively. EF, equil represents the equilibrium level attained between EF and Eredox after contact. Vbi and W represent the built-in potential and the depletion layer width, respectively. .............................................. 7
Figure 1.3 Schematic illustration of the semiconductor-electrolyte interface and the Helmholtz layer. The Gouy layer has been omitted for simplicity as it can be ignored for the applications discussed in this thesis. ............. 8
Figure 1.4 Diagram showing photon absorption, charge migration and subsequent redox reactions at a) one-step and b) two-step photocatalyst systems. ........................................................................... 13
Figure 1.5 Photoelectrochemical water splitting configurations for a) p-type photocathodes, b) n-type photoanodes and c) tandem photoelectrode systems. ........................................................................ 14
Figure 1.6 a) Schematic depiction of nanoparticle, nanorod and branched nanorod electrodes. b-d) Cross-sectional view of the nanoparticle, nanorod and branched nanorod electrodes. e-g) Top-down view of the nanoparticle, nanorod and branched nanorod electrodes. The insets of c) and d) represent magnified SEM images of their respective electrodes.60 Reprinted with permission from Cho, I. S.; Chen, Z.; Forman, A. J.; Kim, D. R.; Rao, P. M.; Jaramillo, T. F.; Zheng, X. Nano Lett. 2011, 11, 4978. Copyright 2011 American Chemical Society. .................................................................................. 20
Figure 1.7 Orbital configurations of the ground state (S0), the first excited singlet state (S1) and the first exctied triplet state (T1), where the arrows represent the electron spin. .................................................................... 24
Figure 1.8 Jablonski diagram indicating the electronic and vibrational levels of singlet and triplet states, as well as radiative and non-radiative transitions. Excitation processes are represented by straight, solid arrows; radiative recombination processes are represented by straight, dashed arrows; and non-radiative transitions are represented by undulated arrows. .......................................................... 25
Figure 1.9 Chemical structures of a few select conjugated organic polymers .......... 27
Figure 1.10 Chemical structure of two D-A polymers. Donor units are colored red and acceptor units are colored blue. ....................................................... 28
Figure 1.11 A representation of the three possible coupling configurations of 3-hexylthipohene monomers: head-to-head, head-to-tail, and tail-to-tail. ‘Hex’ is used to represent the hexyl functional group (-C6H13). ......... 29
xii
Figure 1.12 Diagram depicting the bulk heterojunction morphology achieved when a P3HT:PCBM film is cast on an ITO substrate. ........................... 30
Figure 1.13 Energy level diagram of a P3HT photocathode on ITO and H+/H2 redox couple in solution. ......................................................................... 32
Figure 1.14 Organic photoelectrochemical cell with a device architechture of ITO/PEDOT:PSS/P3HT:PCBM/MoS3.114 Reprinted from Queyriaux, N.; Kaeffer, N.; Morozan, A.; Chavarot-Kerlidou, M.; Artero, V. J. Photochem. Photobiol. C Photochem. Rev. 2015, 25, 1, Copyright 2015, with permission from Elevier. ........................................................ 33
Figure 1.15 Diagram of the glass cell used for photoelectrochemical measurements. ...................................................................................... 38
Figure 1.16 Example of a typical electorolysis trace. The three important regions are labelled, where i. represents the dark region, ii. the photocurrent transient region, and iii. the steady-state photocurrent region................. 39
Figure 1.17 a) An idealized example of a typical OCP measurement, where regions i and ii represent the OCP measured under dark and illuminated conditions, respectively. b) A linear sweep voltammogram under chopped illumination, a vertical dashed line has been drawn from the onset of photocurrent to highlight Vonset..................................... 41
Figure 1.18 Diagram showing the scattering geometry of a typical GIWAXS measurement. ........................................................................................ 43
Figure 1.19 Diagram of an idealized P3HT unit cell of where the (100) and (010) orientations are shown in the lamellar and π-stacking directions, respectively. ........................................................................................... 44
Figure 1.20 Outline of this thesis organized in a flow chart. ....................................... 46
Figure 2.1 Electrolysis of P3HT-coated ITO electrodes at -0.24 V vs. SCE (0 V vs. RHE) performed in 0.1 M H2SO4 (black), 0.1 M NaCl (dotted), 0.1 M NaOH (dashed), and 0.5 M Na2SO3 oxygen scavenger (dot-dashed). Light intensity: 100 mW cm-2 (300-700 nm), under chopped illumination cycles of 30 s dark / 1 min light. .................. 51
Figure 2.2 Gas chromatographic analysis of H2 gas in the photoelectrochemistry cell after 202.5 mC of charge over 60 s (dot-dashed) and after passing 32.5 mC of electrolytic charge over 10 s (dashed) at a Pt foil (i.e., conventional electrochemical HER); 6 h photoelectrolysis at a P3HT-ITO electrode (dotted) after passing 56.6 mC of cathodic charge. The solid black signal represents a sample of air (i.e, blank signal). Signals are offset for clarity. ....................................................... 52
Figure 2.3 Photoelectrolysis of P3HT coated ITO electrodes in 0.1 M H2SO4 with different dissolved oxygen concentrations. Oxygen deficient (<0.01 ppm) conditions (red), ambient (7.80 ppm) conditions (green) and oxygen saturated (> 20.0 ppm) conditions (blue). Light intensity: 100 mW cm-2 (300-700 nm). The electrodes were biased at -0.24 V vs. SCE (0 V vs. RHE). Chopped illumination cycles consist of 30 s dark / 1 min light. ............................................................ 53
Figure 3.1 General schemes showing the formation of organic polymer nanoparticles by (a) the mini-emulsion method and (b) the precipitation method. .............................................................................. 57
xiii
Figure 3.2 Size distribution of P3HT:PCBM nanoparticle diameter determined by (a) dynamic light scattering (DLS) and (b) transmission electron spectroscopy (TEM). For DLS measurements, 1 drop of nanoparticle dispersion was diluted in 2 mL of IPA. TEM samples were prepared by drop casting a diluted nanoparticle dispersion in IPA on a lacey carbon TEM grid. .................................................................................... 64
Figure 3.3 Normalized UV-vis spectra of a P3HT:PCBM nanoparticle film (solid black), a solution cast P3HT:PCBM bulk heterjunction film (solid red), a dispersion of P3HT:PCBM nanoparticles in IPA (dashed black) and a P3HT:PCBM solution in o-DCB (dashed red). nanoparticle and solution cast films were ~90 nm thick. .................................................... 65
Figure 3.4 (a) GIWAXS data obtained for P3HT:PCBM nanoparticle (black) and solution-cast (red) films. The angle of incidence was 0.16°. Inset: Lamellar stacking of P3HT chains in an edge-on fashion along the (100) direction. 2D GIWAXS plots of the P3HT:PCBM nanoparticle and solution-cast electrodes are shown in (b) and (c), respectively. ....... 66
Figure 3.5 Steady-state fluorescence emission spectra of a solution-cast P3HT film (dashed), a solution cast P3HT:PCBM film (red) and a P3HT:PCBM nanoparticle film (black), all on ITO. An excitation wavelength of 515 nm was used. ........................................................... 67
Figure 3.6 SEM images of a P3HT:PCBM nanoparticles on ITO electrodes. Images (a) and (b) depict the nanoparticle films at magnification values of 60k and 250k, respectively. ..................................................... 67
Figure 3.7 Platinum deposition by electrolysis under chopped illumination for P3HT:PCBM nanoparticle (black) and solution-cast (red) films. The potential at the photoelectrodes was held at -0.24 V vs. SCE (0 V vs. RHE). ........................................................................................ 68
Figure 3.8 SEM images, collected using the backscattered electron detector, of Pt deposited after a deposition period of 100 s on (a) a P3HT:PCBM nanoparticle electrode and (b) a solution-cast P3HT:PCBM thin film electrode. ............................................................................................... 69
Figure 3.9 Linear sweep voltammetry (a) and electrolysis (b) under chopped illumination comparing P3HT:PCBM nanoparticle (black) and solution-cast (red) photocathodes in 0.1 M H2SO4. Experiments were performed both with (solid) and without (dashed) Pt nanoparticle catalysts. Linear sweep voltammetry of Pt deposited on bare ITO is shown in (c). Linear sweep voltammograms were measured at a scan rate of 5 mV/s and electrolysis measurements were performed at -0.24 V vs. SCE (0 V vs. RHE). .......................................................... 70
Figure 3.10 Electrolysis under chopped illumination comparing P3HT:PCBM solution-cast photocathodes prepared using Pt deposition periods of 100 s (solid) and 250 s (dashed). Measurements were performed at - 0.24 V vs. SCE (0 V vs. RHE). ............................................................. 71
Figure 4.1 a) OCP vs. light intensity plot for P3HT (circles), 10 wt% PCBM (squares), 25 wt% PCBM (diamonds) and 50 wt% PCBM (triangles) and b) linear sweep voltammetry results for pristine P3HT (black), 10 wt% PCBM (red), 25 wt% PCBM (green) and 50 wt% PCBM (blue). Device architecture for all devices is ITO/P3HT:PCBM.
xiv
Measurements were carried out in 0.1 M Bu4N PF6 in ACN with 1 mM BZQ. Potentials measured vs. Ag/Ag+ reference (0.01 M AgNO3). For OCP measurements, the points plotted, and error bars represent the average OCP values measured at each light intensity and their standard deviation, respectively. For LSV measurements, the Y-axis is offset for clarity, vertical dashed lines are drawn from the onset potential of their respective measurements. ........................................... 78
Figure 4.2 Energy level diagram of a) P3HT and b) P3HT:PCBM (50 wt% PCBM) photocathodes on ITO, in contact with the BZQ redox couple. ............................................................................................................... 79
Figure 4.3 Figure 3. a) OCP vs. light intensity plot for P3HT (circles), 10 wt% PCBM (squares), 25 wt% PCBM (diamonds) and 50 wt% PCBM (triangles) in P3HT:PCBM and b) linear sweep voltammetry results for pristine P3HT (black), 10 wt% PCBM (red), 25 wt% PCBM (green) and 50 wt% PCBM (blue). Device architecture for all devices is ITO/PEDOT:PSS/P3HT:PCBM. Measurements were carried out in 0.1 M Bu4N PF6 in ACN with 1 mM BZQ. Potentials measured vs. Ag/Ag+ reference (0.01 M AgNO3). For OCP measurements, the points plotted, and error bars represent the average OCP values measured at each light intensity and their standard deviation, respectively. For LSV measurements, the Y-axis is offset for clarity, vertical dashed lines are drawn from the onset potential of their respective measurements. ..................................................................... 81
Figure 4.4 Photoelectrolysis using pristine P3HT (black), 10 wt% PCBM (red), 25 wt% PCBM (green) and 50 wt% PCBM (blue) in P3HT:PCBM films. A bias of -0.8 V (vs. Ag/Ag+) was applied for all electrolysis measurements. All electrodes have an active layer thickness of 220 nm and a device architecture of ITO/PEDOT:PSS/P3HT:PCBM. Measurements were carried out in 0.1 M Bu4N PF6 in ACN with 1 mM BZQ. ...................................................................................................... 83
Figure 4.5 Photoelectrolysis at pristine P3HT (black), 10 wt% PCBM (red), and 25 wt% PCBM (green). A bias of -0.8 V (vs. Ag/Ag+) was applied for all electrolysis measurements. All electrodes have a device architecture of ITO/PEDOT:PSS/P3HT:PCBM. Measurements were carried out in 0.1 M Bu4N PF6 in ACN with 1 mM BZQ. .......................... 84
Figure 5.1 Chemical structures of MDMO-PPV, PBD and their oligo(ethylene glycol)-bearing analogues. ..................................................................... 89
Figure 5.2 Chemical structures of the polymers used to study the effect of different conjugated units223 and fluorinated backbone substituents225 on molecular energy levels. .................................................................... 92
Figure 5.3 Chemical structures of PCDTBT and APFO-Green 1 ............................. 93
Figure 5.4 Schematic illustration of the a) rubbing and b) friction transfer processes used to create thin polymer films with controlled polymer chain orientation.236 Reprinted from Brinkmann, M.; Hartmann, L.; Biniek, L.; Tremel, K.; Kayunkid, N. Macromol. Rapid Commun. 2014, 35, 9, Copyright 2013, with permission from WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. .............................................................. 95
xv
Figure 5.5 Scematic illustration of the nanoimpronting method used to create nanostructured polymer films. ................................................................ 96
Figure 5.6 Schematic illustration of the formation of polymer nanostructures using the templating method.242 Copyright 2011, reproduced with permission from the Royal Society of Chemistry. ................................... 97
xvi
List of Acronyms, Symbols and Abbreviations
A Absorbance
ABPE Applied bias photon-to-current efficiency
APCE Absorbed photon-to-current efficiency
B-NR Branched nanorod
BHJ Bulk heterojunction
CDL Depletion layer capacitance
CH Helmholtz capacitance
Ctot Total capacitance
CE Counter electrode
COP Conjugated organic polymer
CB Conduction band
D Diffusion coefficient
D-A Donor-acceptor
DLS Dynamic light scattering
EF Fermi level
Eredox Redox potential
GC Gas chromatography
GIWAXS Grazing incidence wide angle x-ray scattering
HH Head-to-tail
HER Hydrogen evolution reaction
HT Head-to-tail
IPCE Incident photon-to-current efficiency
ITO Indium-doped tin oxide
HOMO Highest occupied molecular orbital
I Intensity of transmitted light
Io Intensity of incident light
jph Photocurrent density
jsc Short circuit current density
kB Boltzmann constant
LSV Linear sweep voltammetry
LUMO Lowest unoccupied molecular orbital
NA Density of acceptor dopants
xvii
NP Nanoparticle
NR Nanorod
OCP Open circuit potential
OER Oxygen evolution reaction
OFET Organic field effect transistor
OPEC Organic photoelectrochemical cell
OPV Organic photovoltaic
ORR Oxygen reduction reaction
P Power density
P3HT Poly(3-hexylthiophene)
PCBM [6,6]phenyl-C61-butyric acid methyl ester
PCE Power conversion efficiency
PL Photoluminescence
q Elementary charge
RHE Reversible hydrogen electrode
S0 Singlet ground state
S1 First excited singlet state
SCE Saturated calomel electrode
SEM Scanning electron microscopy
SHE Standard hydrogen electrode
STH Solar-to-hydrogen efficiency
T Temperature
TEM Transmission electron microscopy
TT Tail-to-tail
T1 First excited triplet state
Vbi Built-in potential
Vph Photovoltage
VOC Open circuit voltage
Vonset Onset potential
VB Valence band
W Depletion layer width
WE Working electrode
XRD X-ray diffraction
εo Permittivity of vacuum
xviii
εr Relative dielectric constant
η Viscosity
ηe-/h+ Absorptance
ηF Faradaic efficiency
ηinterface Interfacial charge transfer efficiency
ηtransport Charge transport efficiency
λ Wavelength
1
Chapter 1. Introduction
The introduction of this thesis will serve to highlight the importance of renewable
energy in the world today, with an emphasis on solar fuels research in general, and
gradually narrowing the scope to discuss the potential role of organic
photoelectrochemical cells (OPECs). Here, fundamental concepts of semiconductor
physics and photoelectrochemistry will be introduced, along with key concepts and
characteristics pertaining to conjugated organic polymers that are central to the research
undertaken. Various photoelectrochemical technologies and design principles will be
described, along with important characterization techniques and their underlying
concepts. Finally, a thorough literature review and historical overview of organic
photoelectrochemical cells is provided to demonstrate the rapid progress in the field,
establish the current state of the field and provide insight into where future research in the
field may be headed.
1.1. Hydrogen as an Alternative Fuel
A rapidly increasing global population and urbanization of developing countries
has resulted in rising global energy demand, predicted to grow from 18.5 TW in 20171 to
30 TW in 2050.2 A serious concern associated with the increasing energy demand is the
increase in CO2 emissions from the consumption of carbon-based energy sources. The
implementation of renewable energy technologies is, therefore, crucial to limiting the rise
in CO2 emissions, and consequently their contributions to increasing global temperatures.
Current renewable energy sources include wind, hydroelectric, tidal, geothermal, biomass,
nuclear and solar. Of these sources, however, only solar energy has the potential to meet
the global energy demand, as it is estimated that 120,000 TW of power reaches the earth’s
surface every year from the sun.
Although the prospect of meeting the energy demands of the entire planet using
direct solar energy is appealing, the intermittent nature of sunlight and the limited capacity
of the electrical grid infrastructure bring about the need for large-scale energy storage
solutions. While various energy storage technologies exist (e.g., mechanical,
2
electrochemical, superconducting magnetic, and cryogenic), chemical energy storage
offers the advantage of high energy density and facile transportation for on demand use.3
Hydrogen, specifically, is an attractive fuel as it can be produced directly through the
photoelectrolysis of water, where solar energy is used to drive redox reactions at
semiconductor photoelectrodes immersed in aqueous solutions, resulting in the evolution
of hydrogen and oxygen gas, as shown in Equation 1.
2H2O (l) + hν → 2H2 (g) + O2 (g) (1)
The hydrogen produced can then be stored, transported and used on demand in
fuel cells or combustion engines, while emitting only water vapour as a by-product.
Considering the above, it is understandable that the use of abundant renewable resources
(i.e., water and sunlight) to produce a fuel with no carbon emissions upon consumption is
often referred to as the “Holy Grail” of renewable energy.
Although hydrogen represents an attractive alternative to fossil fuels, the lack of
economically viable methods to produce clean and sustainable hydrogen at large-scales
represents a major challenge towards widespread adoption. Currently, 80-85% of the
hydrogen produced globally is derived via steam methane reforming, while the remainder
is produced via coal gasification and the electrolysis of water.4 The major drawback of
hydrogen production from hydrocarbon feedstocks is the production of CO2 as a by-
product. While downstream CO2 sequestration and/or utilization technologies may provide
short term relief regarding carbon emissions, ultimately, carbon-free hydrogen production
is required to sustain any sort of “hydrogen economy”. Although CO2 is not a direct by-
product of water electrolysis, it is an energy intensive process due to the large operating
voltages (~1.9 V) required to operate electrolyzers at their optimal current densities
(~1 A/cm2). Additionally, the electrode materials employed are generally expensive and
not appropriate for the large-scale production of hydrogen. However, efforts are being
made in the area of water electrolysis to develop new materials and technologies in hopes
of lowering operating voltages and material costs.5
Although great progress towards the ultimate goal of solar hydrogen production
has been made over the years, a number of challenges remain to be addressed.
Photoelectrochemical water splitting represents a complex technology where devices
must be able to efficiently absorb sunlight, possess appropriate thermodynamics for the
3
desired hydrogen and oxygen evolution reactions to take place, generate sufficient
photovoltage to drive these reactions, and exhibit sufficiently fast kinetics at the
semiconductor-electrolyte interface. Additionally, these devices must exhibit long-term
stability under operating conditions and must not be prohibitively expensive. These
requirements determine whether a material can be considered for use in
photoelectrochemical water splitting technologies and are generally divided into three
broad categories: efficiency, stability and cost. The baseline targets for the economical
production of solar hydrogen established by the US Department of Energy are 10% solar-
to-hydrogen efficiency, an operating lifetime of 5 years and a cost of $2-4 per kilogram of
H2 produced. Although it is not uncommon to find materials that exceed one, or even two,
of the three requirements, a material that meets all three requirements has yet to be
discovered. For example, many metal oxide semiconductors represent inexpensive
materials that produce stable photocurrents for extended periods of time, however they
suffer from poor efficiencies.6,7 On the opposite end of the spectrum, multijunction
photoelectrodes based on III-V compound semiconductors are capable of reaching solar-
to-hydrogen efficiencies over 10%, but suffer from long-term stability issues and use
expensive materials.8,9
In addition to the scientific challenges, the future of solar driven hydrogen
production is further complicated by technical obstacles such as storage, infrastructure,
and, depending on the type of reactor used, the separation of hydrogen and oxygen gases
from the potentially explosive gas mixture that is created. Storing the generated hydrogen
is an issue because, although the gravimetric energy density of hydrogen is very high, its
volumetric energy density is rather low. To circumvent the volumetric energy density
issues, various storage technologies are being explored. These technologies range from
simple solutions such as high pressure storage containers10 and liquid cryo-storage,11 to
more sophisticated approaches like high surface area metal organic frameworks that are
capable of storing hydrogen through physisorption.12 Additionally, the lack of infrastructure
is often cited as a major technical and economical barrier to the widespread adoption of
hydrogen as an energy source. Currently, there is little infrastructure in place to support a
hydrogen economy and would require significant capital investment to construct the large-
scale hydrogen production, storage, long-distance transmission, local distribution and
dispensing systems needed. Although the seemingly insurmountable task of creating the
4
infrastructure for a new primary energy source from the ground up looms large, several
reports and proposals exist that argue its viability.13,14
1.2. Photoelectrolysis of Water
1.2.1. Fundamentals of Semiconductor Photoelectrochemistry
Semiconductors represent the photoabsorbing component of
photoelectrochemical devices and are responsible for the generation of charge carriers
that will ultimately be used in the electrochemical redox reactions that take place at the
electrolyte interface. The understanding of the complex interactions of semiconductors
with incoming solar radiation and the liquid electrolyte is crucial to furthering the area of
photoelectrolysis. This section outlines the fundamental properties of semiconductors, as
well as the important processes that occur when semiconductors interact with light and at
the semiconductor-electrolyte interface.
Semiconductors are a class of materials whose conductive properties can be
manipulated so as to behave as either insulators or conductors under certain conditions.
The conductivity of a semiconductor is best described by the band model. As
semiconductors are made up of many covalently bonded atoms, the energy levels become
so dense that broad energy bands are formed as opposed to the discrete energy levels
observed for single atoms or molecular systems.15 When an electron resides in the lower
energy band, referred to as the valence band, no conduction may occur and the
semiconductor acts as an insulator,* however, when an electron gains enough energy to
populate the higher energy band, known as the conduction band, it is now free to
participate in conduction. The excitation of an electron to the conduction band leaves
behind a vacancy in the valence band, providing an opportunity for other valence band
electrons to move and occupy the newly formed vacancy. For simplicity, this vacancy is
often thought of as a particle with a positive charge that moves through the valence band,
known as a hole. The minimum energy needed to promote an electron from the valence
band to the conduction band is referred to as the band gap energy.
* This assumes that the valence band is full.
5
Exposing a semiconductor to solar radiation is a common way to excite a
semiconductor electron from the valence band to the conduction band, where
electromagnetic radiation with energy equal to or greater than the semiconductor band
gap will be absorbed. The formation of photogenerated electron-hole pairs, often referred
to as excitons, is the fundamental process of solar energy production and conversion.
These photoexcited electrons, however, exist in a meta-stable state where they will
eventually recombine with a hole. This charge carrier recombination represents a major
loss mechanism in photoelectrochemical applications.16
Also central to semiconductor photoelectrochemistry is the concept of the Fermi
level, which represents a hypothetical energy level where, at absolute zero, all energy
states below the Fermi level are occupied and all energy states above the Fermi level are
unoccupied. At finite temperatures it is convenient to describe the Fermi level as an energy
level that, at thermodynamic equilibrium, has a 50% chance of being occupied by an
electron at a given temperature. For intrinsic (i.e., undoped) semiconductors, the Fermi-
level lies in the middle of the band gap. For extrinsic (i.e., doped) semiconductors, charge
carrier densities can be manipulated by introducing either donor or acceptor atoms. Take,
for example, a defect-free crystal lattice made up of silicon atoms, a group IV element with
four valence electrons. Each silicon atom will share one valence electron with each of its
four neighbouring silicon atoms, who also share one of their electrons, forming a
covalently bonded network where all available valence electrons are participating in
bonding, i.e., no free electrons are available to participate in conduction. By replacing one
silicon atom with a group V element (e.g., phosphorous) that contains five valence
electrons, the group V atom will share one valence electron with each of its four
neighbouring silicon atoms and one electron remains “free” to participate in conduction.
As a result, the number of negative charge carriers (electrons), is greater than the number
of positive charge carriers (holes) and is referred to as n-type doping. The opposite is true
if a silicon atom is replaced by a group III element (e.g., boron) that contains three valence
electrons. The lack of a shared electron to one of the surrounding silicon atoms will result
in a vacancy (hole) that is free to participate in conduction. In this case, holes will
outnumber electrons and is referred to as p-type doping. A major consequence of doping
is the shift in Fermi level within the semiconductor, as n-type doping results in the Fermi
level shifting upwards towards the bottom of the conduction band and p-type doping will
shift the Fermi level downwards towards the top of the valence band. As the relative ratio
6
of positive and negative charge carriers can be controlled through doping, the concept of
majority and minority carriers can be applied. In p-type materials, holes act as majority
carriers while electrons act as minority carriers; the opposite is true for n-type materials,
i.e., electrons are the majority carriers and holes are the minority carriers.
Figure 1.1 Relative Fermi level (EF) positions of a) intrinsic, b) p-type and c) n-type semiconductors. Also shown are the valence band (VB) and conduction band (CB) levels.
The semiconductor-electrolyte interface plays an important role in semiconductor
photoelectrochemistry and the ability of a semiconductor to perform a
photoelectrochemical reaction is highly dependent on both the interfacial thermodynamics
and kinetics. The relative positions of the semiconductor Fermi level and the
electrochemical potential of the redox couple in solution are crucial to device performance,
as are the electron transfer kinetics across the semiconductor-electrolyte interface. For a
semiconductor in contact with electrolyte, the electrochemical potential of the redox couple
must lie between the valence and conduction band edges (i.e., within the band gap) of the
semiconductor and above (below) the Fermi level for p-type (n-type) semiconductors.
The difference in energy between the semiconductor Fermi level and
electrochemical potential of the redox couple leads to spontaneous charge transfer at the
interface, where majority carriers will be transferred from the semiconductor to the
electrolyte until an equilibrium is reached. As a result, there will be a region that extends
from the semiconductor surface into the bulk that is depleted of majority carriers, aptly
referred to as the depletion layer. Consequently, an internal electric field that extends the
width of the depletion layer is developed. The band bending diagrams depicting these
phenomena are shown below in Figure 1.2.
The depletion layer plays an important role in semiconductor
photoelectrochemistry as any photogenerated charges that originate within the depletion
layer can be effectively separated by the internal electric field, which will influence the
7
overall performance of the photoelectrochemical device. The depletion layer is determined
by several semiconductor properties and can be calculated by Equation 2, where W is the
Figure 1.2 The relative energy levels at the semiconductor-electrolyte interface before (left) and after (right) equilibration for a) p-type and b) n-type semiconductors. ECB, EVB, EF and Eredox represent the positions of the conduction band, valence band, Fermi level and redox potential, respectively. EF, equil represents the equilibrium level attained between EF and Eredox after contact. Vbi and W represent the built-in potential and the depletion layer width, respectively.
depletion layer width, εr is the relative dielectric constant of the semiconductor, εo is the
permittivity of vacuum, Vbi is the built-in potential, q is the elementary charge and NA is the
doping density.
W = √2εrεoVbi
qNA (2)
The transfer of majority carriers across the semiconductor-electrolyte interface
during the equilibration process results in a net charge at the semiconductor surface. For
p-type semiconductors, the transfer of positively charged majority carriers (i.e., holes)
across the interface leads to a negatively charged surface. To counterbalance this surface
charge, a redistribution of ions occurs within the electrolyte near the semiconductor
surface, forming two oppositely charged layers at the semiconductor-electrolyte interface.
The resulting formation is termed the electrochemical double layers and is analogous to
an electrical capacitor consisting of two oppositely charged plates separated by some
distance. As such, a potential drop exists across the electrochemical double layer. The
8
electrolyte component of the double layer consists of a compact layer of solvated ions at
the immediate semiconductor-electrolyte boundary, known as the Helmholtz layer, and a
diffuse layer that extends into the bulk of the electrolyte, known as the Gouy layer. When
the concentration of the electrolyte is sufficiently high (≥0.01 M), the contribution from the
diffuse layer can be neglected. As a result, the potential drop in solution will be confined
to the Helmholtz layer.
Figure 1.3 Schematic illustration of the semiconductor-electrolyte interface and the Helmholtz layer. The Gouy layer has been omitted for simplicity as it can be ignored for the applications discussed in this thesis.
The formation of the electrochemical double layer becomes important when
considering the application of an external bias potential to the photoelectrochemical cell,
which is common in solar water splitting applications. When an external bias is applied,
the potential difference will be distributed over both the depletion layer within the
semiconductor and the Helmholtz layer in solution. The total capacitance can then be
modeled as two capacitors in series:
1
𝐶tot=
1
𝐶DL+
1
𝐶H (3)
where Ctot is the total capacitance, CDL is the depletion layer capacitance and CH is the
Helmholtz capacitance. Given that the charge, Q, in both layers are equal and the
relationship between capacitance, charge and potential is given as:
𝐶 =𝑄
𝑉 (4)
9
the potential distribution across the both the depletion (ΔVDL) and the Helmholtz (ΔVH)
layers can be described by:
∆𝑉DL
∆𝑉H=
𝐶H
𝐶DL (5)
Because of the inverse relationship between capacitance and the distance
between two oppositely charged plates in the classical capacitor model, or the thickness
of each layer for the electrochemical double layer, the Helmholtz capacitance will be much
larger than the depletion layer capacitance. The typical width of a Helmholtz layer is
~2-5 Å, whereas depletion layer widths typically range from 5-500 nm.2 As CH>>CDL, any
applied potential will fall only across the depletion layer of the semiconductor. The
application of a bias potential to a semiconductor working electrode will, therefore,
increase both the magnitude of band bending and depletion layer width within the
semiconductor. This is an important property of the semiconductor-electrolyte interface
that becomes useful for the characterization of photoelectrochemical cells, as will be
further discussed in Section 1.5.2.
1.2.2. The Hydrogen Evolution and Oxygen Evolution Reactions
Solar-driven water splitting takes advantage of the photogenerated charges
produced within the semiconductor layer to perform complementary redox reactions at the
electrolyte interface, where the reduction and oxidation reactions are responsible for the
evolution of hydrogen and oxygen gases, respectively. Under acidic conditions, the
reduction and oxidation reactions can be written as
4 H+ + 4e− → 2 H2 𝐸° = 0.00 V vs. SHE
2 H2O → O2 + 4H+ + 4e− 𝐸° = 1.23 V vs. SHE
2 H2O → 2H2 + O2 𝐸cell° = −1.23 V vs. SHE
In theory, the thermodynamic driving force required for water splitting to occur is
1.23 V. This suggests that a single semiconductor material that possesses a band gap
>1.23 eV and whose conduction band and valence band energies straddle the redox
potentials of the hydrogen evolution reaction (HER) and oxygen evolution reaction (OER),
respectively, can be used to successfully drive both the HER and OER at the
semiconductor-electrolyte interface. In practice, however, electron transfer at the
10
semiconductor-electrolyte interface produces losses due to the concentration and kinetic
overpotentials required to perform the HER and OER. As a result, the energy needed for
the successful photoelectrolysis of water is often cited as 1.6-2.4 V.17
The large overpotentials required for photoelectrolysis effectively limits the
materials that are capable of “overall water splitting” to semiconductors with large band
gaps. Because of their large band gaps, most of these materials require photons with
sub-400 nm wavelengths (i.e., ultraviolet radiation) to generate photoexcited electrons.
The dependence on ultraviolet radiation reduces the maximum achievable efficiencies for
these materials, as it represents only a fraction of the solar spectrum. Even if all UV light
up to 400 nm was utilized without any losses, the resulting solar-to-hydrogen efficiency
would only be 2%.18 As photons with 400-800 nm wavelengths (i.e., visible light) represent
nearly half of the incoming solar radiation, the efficient use of photons in the visible region
is required for highly efficient photoelectrolysis. For example, if all photons up to 600 nm
were utilized, an efficiency of 16% could be realized. Further increasing the wavelength of
utilized photons to 800 nm would result in an efficiency of 32%.18
The limitations brought about by the overpotentials required for the
photoelectrolysis of water have motivated researchers to develop photoelectrochemical
water splitting systems with configurations that take advantage of semiconductor
properties targeted towards either the HER or OER, multijunction configurations that use
semiconductors with complementary band gaps to maximize absorption across the solar
spectrum and the implementation of cocatalysts to improve charge transfer kinetics. The
following section will explore the development of device architectures and design
principles commonly used in photoelectrolysis.
1.2.3. Device Architecture and Design Principles
As the field of solar water splitting matures, device architectures and design
principles are continually improving to push performance higher. Nevertheless, the
majority of water splitting systems can be classified into two major categories, namely
photocatalytic and photoelectrochemical systems.
11
Semiconductor Photocatalyst Systems
Photocatalytic systems consist of semiconductor particles suspended in an
electrolyte and are capable of performing photocatalytic redox reactions in the absence of
any externally applied bias. The simple design allows for relatively inexpensive and facile
scale-up of reactor technologies.19 Two strategies exist for achieving photocatalytic water
splitting using heterogenous semiconductor catalysts; the first method is a one-step
system, where a single photoabsorber possessing a conduction band edge more negative
than the hydrogen evolution reaction (HER) redox potential and valence band edge more
positive than the oxygen evolution reaction (OER) redox potential is employed to
simultaneously generate hydrogen and oxygen gas. An efficient one-step photocatalytic
system must satisfy the following requirements: (i) have a low band gap for the efficient
use of visible light, (ii) be resistant to corrosion under photoirradiation and (iii) possess
suitable band edge energetics for both hydrogen and oxygen evolution.
While metal oxides are generally stable against photocorrosion, it is difficult to
develop a metal oxide photocatalyst that has both sufficiently negative conduction band
to perform the HER and a sufficiently narrow band gap to ensure good utilization of the
solar spectrum due to the highly positive valence band formed by the O 2p orbital.20
Alternatively, non-oxide materials, such as sulfide and nitride semiconductors, possess
appropriate band levels for overall water splitting using visible light, however, these
materials are generally plagued by stability issues as the self-oxidation process competes
with oxygen evolution. This has been shown for CdS photocatalysts, where the S2- anions
are preferentially oxidized over water molecules in the absence of a sacrificial electron
donor.21 One class of semiconductors that has been shown to be an effective one-step
water splitting system are metal oxynitrides. When (Ga1-xZnx)(N1-xOx) is modified with a
hydrogen evolution cocatalyst, Rh2-yCryO3, stoichiometric amounts of H2 and O2 are
evolved, and an apparent quantum yield of 5.9% at 420-440 nm has been achieved.22
As it is difficult to find a material that meets all the above requirements, a more
commonly applied strategy is the two-step, or z-scheme, water splitting system. The two-
step method allows for the use of two separate semiconductor materials, each optimized
for either the water reduction or oxidation processes and couples them with a redox shuttle
that will react with the photogenerated holes produced at the HER photocatalyst and
photogenerated electrons produced at the OER photocatalyst. The two-step system also
12
allows for the product gases to be generated separately, which is not possible for a one-
step system. Although the two-step system has significant advantages over the one-step
system, it requires twice as many photons to achieve water splitting and faces challenges
with regards to the backwards reaction of the redox shuttle at the photocatalyst materials.23
The two-step water splitting scheme was first conceptualized in 1979 by Allen J.
Bard,24 but the successful stoichiometric evolution of H2 and O2 in a two-step
photocatalytic system was not demonstrated until 2001.25 This was accomplished using a
Pt-loaded anatase TiO2 photocatalyst to perform the HER, bare rutile TiO2 photocatalyst
to perform the OER and an iodate/iodide (IO3-/I-) redox mediator. Because of the large
bandgap of TiO2, this system is, however, only capable of operating under UV irradiation.
A visible light driven, two-step photocatalyst system was developed later that same year,
employing a Pt-loaded SrTiO3 photocatalyst doped with Cr and Ta as the hydrogen
evolution photocatalyst, a Pt-loaded WO3 oxygen evolution photocatalyst, and an
iodate/iodide (IO3-/I-) redox mediator.26 Although the quantum efficiency for this system
was rather low (0.1% at 420.7 nm), it represented an important milestone in visible light
driven photocatalytic water splitting. More recently, La- and Rh-codoped SrTiO3
(SrTiO3:La,Rh) and Mo-doped BiVO4 (BiVO4:Mo) photocatalysts were embedded onto a
gold layer and modified with a Ru species co-catalyst.27 The underlying gold layer allows
for direct charge transfer between the SrTiO3:La,Rh hydrogen evolution photocatalyst and
the BiVO4:Mo oxygen evolution photocatalyst, effectively removing the need for a redox
mediator in solution and, hence, any competing backwards reactions caused by the redox
mediator. Under visible light irradiation, the Ru-modified SrTiO3:La,Rh/Au/BiVO4:Mo sheet
was able to split pure water, evolving stoichiometric amounts of H2 and O2 with a solar-to-
hydrogen energy conversion efficiency of 1.1% and apparent quantum yield of 33% at
419 nm.
Although photocatalytic systems represent a viable water splitting system, the
remainder of this thesis will focus on photoelectrochemical cells. Comprehensive reviews
of heterogenous photocatalysts can be found elsewhere.18,28,29
13
Figure 1.4 Diagram showing photon absorption, charge migration and subsequent redox reactions at a) one-step and b) two-step photocatalyst systems.
Photoelectrochemical Cells
Photoelectrochemical systems consist of either a two- or three-electrode setup,
where the electrodes are connected to a potentiostat, allowing for the control over
variables such as applied potential or current. The minimum requirements for a
photoelectrochemical cell are a semiconductor working electrode, where photogenerated
charges are used to perform a redox reaction at the semiconductor-electrolyte interface,
and a counter electrode, at which the complementary redox reaction occurs.
Photoelectrochemical systems can be further broken down into photoanodes, where a
n-type semiconductor is used as the working electrode to perform photooxidation
reactions; photocathodes, where a p-type semiconductor is used as the working electrode
to perform photoreduction reactions; and tandem cells, where a n-type photoanode and
p-type photocathode are simultaneously employed to perform the complementary
photooxidation and photoreduction reactions without any external bias.
While single photoelectrode systems allow for the use of low band gap materials
with properties tailored for either the reduction or oxidation of water, they generally require
an external bias to generate a potential large enough to split water. The use of tandem
photoelectrochemical cells offers the advantage of exploring various combinations of
semiconductors with complementary absorption and stability characteristics, leading to
systems capable of achieving higher photovoltages and overall water splitting in the
absence of an external bias.17
14
Since the first report of photoelectrochemical water splitting, which employed a
TiO2 photoanode and a Pt counter electrode,30 a multitude of semiconductor materials
have been developed and used as photoelectrodes for solar water splitting applications.
A few of the most promising materials are highlighted below.
Figure 1.5 Photoelectrochemical water splitting configurations for a) p-type photocathodes, b) n-type photoanodes and c) tandem photoelectrode systems.
Photoanode Materials
TiO2-based photoanodes have been extensively studied since the inaugural work
of Fujishima and Honda in 1972,31 as TiO2 is an earth-abundant, non-toxic material that
exhibits good stability. Due to its large band gap (3.2 eV for anatase and 3.0 eV for rutile
TiO2), however, the maximum theoretical solar-to-hydrogen (STH) efficiency for these
materials is very low (1.3% for anatase and 2.2% for rutile TiO2).32 As a result, various
metal oxide semiconductors have been investigated over the years in an attempt to
increase the utilization of visible light (i.e., decrease the band gap of the semiconductor)
while maintaining the desirable properties associated with TiO2, such as good stability,
abundance and low-toxicity. A promising metal oxide that has emerged is α-Fe2O3
(hematite).
Hematite is a highly abundant material that exhibits good stability and low toxicity.
Additionally, it has a favorable band gap, reported between 1.9 and 2.2 eV. The reduction
in band gap, compared to TiO2, increases the maximum theoretical STH efficiency of
hematite to 12.9% for a band gap value of 2.2 eV.32 Despite these promising properties,
hematite also possess significant drawbacks, including: short diffusion lengths (2-4 nm)
and short charge carrier lifetimes (on the order of picoseconds), leading to large
recombination rates; relatively low absorption coefficient (~103 cm-1), resulting in the need
for relatively thick films (400-500 nm) to ensure efficient photon collection; and poor water
15
oxidation kinetics.33 To overcome these limitations, strategies including nanostructuring
and cocatalyst modification have been applied to hematite photoanodes.† For example,
photoanodes consisting of hematite nanosheets modified with Ag and cobalt phosphate
nanoparticle cocatalysts have reportedly achieved a photocurrent density of 4.7 mA/cm2
at 1.23 V vs. RHE and a STH efficiency of 0.55%.34
In addition to metal oxides, n-type group IV semiconductors (e.g., n-Si), group III-V
compound semiconductors (e.g., n-GaAs and n-InP) and group II-VI compound
semiconductors (e.g., n-CdTe) have been the subject of investigation for
photoelectrochemical water splitting applications. In general, these materials possess
narrow band gaps (e.g., 1.1 eV for Si, 1.42 eV for GaAs and 1.5 eV for CdTe) and high
charge carrier mobilities, allowing them to achieve large current densities and high STH
efficiencies. Unfortunately, these materials are also very susceptible to photocorrosion or
photopassivation, making them highly unstable under operating conditions.35,36 Efforts in
this area have, therefore, been focused on improving the stability of these materials. The
use of a thin TiO2 layer grown by atomic layer deposition has been shown to prevent the
corrosion of Si, GaAs and GaP photoanodes.37 When modified with a NiO cocatalyst, Si
photoanodes exhibited photocurrent densities >30 mA/cm2 at 1.17 V vs. RHE for over 100
hours of continuous oxygen evolution, while the analogous GaAs and GaP photoanodes
exhibited operational lifetimes of over 25 hours. In a separate study, CdTe photoanodes
that had a TiO2 protecting layer decorated with NiO cocatalyst exhibited a steady
photocurrent density of 21.5 mA/cm2 at 2.07 V over four days of operation.38
Photocathode Materials
Cu2O is an attractive material for use as a photocathode as it is abundant, has low
toxicity and possesses a band gap of 2 eV, which translates to a theoretical maximum
STH efficiency of 18%.39 In contrast to metal oxides used for photoanodes, Cu2O suffers
from poor stability and is susceptible to photocorrosion. As the redox potentials of the
reduction and oxidation of monovalent copper oxide lie within the band gap, parasitic
self-reduction and self-oxidation pathways to Cu and CuO, respectively, are kinetically
and thermodynamically possible.40 The incorporation of protective layers have been
shown to improve stability, for example, an aluminium-doped zinc oxide (AZO) layer and
† An in-depth discussion regarding cocatalyst modification and nanostructuring is presented in the proceeding sections: Cocatalysts and Nanostructures
16
TiO2 layer were sequentially deposited on top of a Cu2O photocathode by atomic layer
deposition, followed by the photoelectrochemical deposition of a MoS2+x cocatalyst.41 The
n-type AZO creates a p-n junction with the p-type Cu2O, promoting efficient transfer of
photogenerated electrons from Cu2O into the AZO layer and the TiO2 layer protects the
buried Cu2O-AZO junction. When used in conjunction with a MoS2+x cocatalyst, the Cu2O-
based photocathode exhibited stable photocurrents of 5.7 mA/cm2 at 0 V vs. RHE over a
continuous operating period of 10 hours.
More recently, I-III-VI2 chalcopyrites, a class of p-type semiconductors where I =
Cu, Ag; III = Ga, In; VI = S, Se, have emerged as promising photocathode materials. These
materials possess large minority-carrier diffusion lengths (up to a few micrometers),
leading to enhanced charge separation; high absorption coefficients (~105 cm-1), allowing
for efficient photon collection using thin films; and tunable band gap energies (1.0-2.4 eV),
achieved by adjusting the alloy composition.42 Photocurrent densities as high as
16 mA/cm2 at 0 V vs. RHE were achieved using a Cu(In,Ga)(Se,S)2 photocathode
passivated with a ZnS layer, and further modified with a Pt hydrogen evolution
cocatalyst.43 The stability of these materials has also been demonstrated using a CuGaSe2
photocathode modified with a thin CdS layer and Pt cocatalyst.44 Here, stable
photocurrents of ~3.5 mA/cm2 at 0 V vs. RHE were measured under constant operation
for more than 10 days.
Finally, p-type group IV semiconductors (e.g., p-Si) and group III-V compound
semiconductors (e.g., p-GaN, p-InP and p-GaAs) have been studied as photocathodes for
the production. Similar to their n-type analogues used as photoanodes, these materials
exhibit large photocurrents and STH efficiencies, but suffer from poor stability under
operating conditions. Examples of these materials used for photoelectrochemical
hydrogen evolution have been published for decades, but more recently, the incorporation
of passivation layers and nanostructured architectures have been reported as strategies
to improve device stability and performance.45–49
Tandem Cell Configurations
Tandem photoelectrochemical cells incorporate both a n-type photoanode and
p-type photocathode to create a two-absorber system capable of unassisted, solar-driven
water splitting. Semiconductors with complementary spectral absorption profiles are used
to ensure efficient utilization of incoming solar radiation. For example, a wider band gap
17
semiconductor would be used as the front photoelectrode, absorbing the high energy
photons while the lower energy photons would be transmitted to, and subsequently
absorbed by, the narrow band gap semiconductor used to form the back photoelectrode.
By matching the band edge positions of the photoanode and photocathode materials, the
combined photovoltages produced at each photoelectrode is sufficient to overcome the
thermodynamic threshold to perform the hydrogen and oxygen evolution reactions and
their associated overpotentials, removing the need for any external bias.
The combination of efficient photon utilization and large achievable photovoltages
allows for tandem cells to be highly efficient. Calculations done by the Lewis group have
shown that the theoretical maximum achievable STH efficiency for a tandem
photoelectrochemical cell with an optimal band gap combination is 29.7%.50 This model
assumes optimal band gaps of 1.60 eV and 0.95 eV for the front and back absorbers,
respectively, a fill factor of 0.85, a solution resistance of 5 Ω/cm2 and the use of low
overpotential Pt and RuO2 coctalysts. Using the same assumptions, theoretical STH
efficiencies greater than 25% can be achieved by pairing semiconductor materials with
band gaps between 1.6-1.8 eV with Si, which has a band gap of 1.1 eV.
Practically, however, measured STH efficiencies have been much lower. The
current benchmark for unassisted solar water system using a tandem structure is a 14%
solar-to-hydrogen efficiency produced by a III-V photovoltaic tandem absorber.51 Here, the
front layer consists of GaInP (Eg=1.78 eV) with a modified AlInP overlayer and Rh
cocatalyst, while the back layer consists of a GaInAs (Eg=1.26 eV) photoabsorber, where
the photogenerated minority carriers are shuttled to a separate RuO2 electrode via
external circuit, where they are used to perform the oxygen evolution reaction. Although
this configuration exhibits high performance, the materials and preparation methods are
prohibitive to large scale applicability. Ultimately, cheap and abundant materials will be
necessary to meet the technical requirements for large-scale hydrogen production.
In an effort to develop tandem photoelectrochemical cells from earth abundant
materials, Jang et al. combined a hematite photoanode modified with a NiFeOx oxygen
evolution cocatalyst and a Si photocathode modified with a TiO2/Pt overlayer in a tandem
configuration.52 This configuration was able to achieve unassisted solar water splitting with
a STH efficiency of 0.91%, however the authors noted that the photocurrents achieved at
18
the back Si photocathode were reduced five-fold (compared to photocurrents measured
as a standalone photocathode) due to the initial absorption of hematite.
To rectify the issue of parasitic photon absorption by the front absorber in tandem
cells, the photoelectrodes can be placed in a parallel (i.e., side-by-side) configuration. Ding
et al. developed a tandem photoelectrochemical cell based on a Ni-modifed Si solar cell
as the photocathode and a Mo-doped BiVO4 photoanode decorated with a FeOOH oxygen
evolution cocatalyst, where a higher STH efficiency was achieved when the
photoelectrodes were employed in a side-by-side configuration (2.21%) over a stacked
configuration (0.77%).53 Although the stacked photoelectrode configuration allows for
more efficient utilization of incoming solar irradiation and wireless configurations that are
compatible with convenient assembly, the increased efficiencies of the side-to-side
configuration cannot be overlooked.
Strategies to Increase Performance
As evidenced by the previous sections, the efficient production of hydrogen and
oxygen rarely occurs at bare semiconductor electrodes and several strategies have been
developed to improve the efficiency of these photoelectrochemical systems. This section
will cover two general strategies to improve water splitting systems: (i) using HER/OER
cocatalysts to promote the desired water splitting reactions and (ii) creating
nanostructured photoelectrodes to take advantage of specific size effects at the
nanoscale.
Cocatalysts
The generation of intermediate species during the hydrogen and oxygen evolution
reactions, coupled with the multielectron nature of these reactions leads to kinetic barriers
that must be overcome in addition to the 1.23 V thermodynamic requirement. In most
cases, the formation of intermediate species at bare semiconductor surfaces presents a
large energy barrier, and thus large overpotentials are required to drive the HER/OER. By
loading a cocatalyst onto the semiconductor surface, the overpotentials required to
perform the HER/OER can be reduced drastically.
While noble metal catalysts are generally thought of as the best performing
catalysts, e.g., IrO2 and RuO2 as oxygen evolution catalysts and Pt as a hydrogen
evolution catalyst, their high cost and low abundance make them unsuitable for large scale
19
water splitting applications. Similar to the semiconductor candidates for solar-driven water
splitting, the cocatalysts employed should be inexpensive and readily available. For this
reason, non-precious metal oxygen and hydrogen evolution catalysts have been the
subject of many investigations. For example, the Jaramillo group has compared the
overpotentials needed to achieve a current density of 10 mA/cm2, the estimated current
density needed to reach 10% STH efficiency, for twenty-six oxygen evolution catalysts
and eighteen hydrogen evolution catalysts.54 From this benchmarking study, a few
promising non-precious OER catalysts were identified. While twelve non-precious oxygen
evolution catalysts were able to achieve current densities of 10 mA/cm2 at overpotentials
between 0.35-0.45 V (compared to 0.29 V for Ru), only six showed promising long-term
stability where the overpotential required to maintain the current density of 10 mA/cm2
remained constant, or even decreased slightly in some cases, after twenty-four hours. The
six promising candidates were Co/P, CoFe, NiCo, NiFe, NiMoFe and NiZn.
In the same study, two hydrogen evolution catalysts, NiMo and NiMoCo, were
shown to exhibit comparable overpotentials to polished polycrystalline Pt and platinized
Pt electrodes. The overpotentials associated with both Pt electrodes is 0.04 V, while NiMo
and NiMoCo exhibit overpotentials of 0.045 and 0.05 V, respectively. CoMo also exhibited
a relatively low overpotential of 0.1 V. Moreover, the overpotentials required remained
stable over twenty-four hours for all three of these catalysts (i.e., NiMo, NiMoCo and
CoMo).
The cocatalysts identified here represent only a small subset of materials that have
been developed as oxygen and hydrogen evolution catalysts. Many other OER cocatalysts
exist, such as perovskites, spinels, metal hydroxide and oxyhydroxides, metal
chalcogenides and metal pnictides, and have been summarized in a recent review by
Suen et al.55 Similarly, other non-precious HER cocatalysts such as, metal sulfides,
selenides, carbides, nitrides, and phosphides are reviewed by Zou and Zang.56
Nanostructures
As the area of nanoscience has experienced rapid advancement, nanostructuring
of existing semiconductor materials has become a popular method of improving water
splitting devices and offers distinct advantages over bulk semiconductor properties. These
advantages include high aspect ratios, leading to shorter path lengths that photogenerated
charge carriers must travel before reaching the semiconductor-electrolyte interface;
20
improved light distribution due to the scattering of light at nanoscale features; tunability of
energy levels due to quantum confinement effects; and enhanced charge transfer across
the semiconductor-electrolyte interface due to increased surface area. A multitude of
nanostructured photoelectrodes exist, including nanoparticles,57 nanorods,58 nanotubes,59
and branched nanorods.60 Although the performance of a material for water splitting
applications will ultimately depend on its intrinsic properties, nanostructuring has been
beneficial to the development of improved photoelectrode materials.
Figure 1.6 a) Schematic depiction of nanoparticle, nanorod and branched nanorod electrodes. b-d) Cross-sectional view of the nanoparticle, nanorod and branched nanorod electrodes. e-g) Top-down view of the nanoparticle, nanorod and branched nanorod electrodes. The insets of c) and d) represent magnified SEM images of their respective electrodes.60 Reprinted with permission from Cho, I. S.; Chen, Z.; Forman, A. J.; Kim, D. R.; Rao, P. M.; Jaramillo, T. F.; Zheng, X. Nano Lett. 2011, 11, 4978. Copyright 2011 American Chemical Society.
The ability of nanoparticle (NP), nanorod (NR) and branched nanorod (B-NR) TiO2
photoanodes to perform the oxygen evolution reaction was compared by Cho et al, where
it was shown that the branched nanorods exhibited the highest performance.60 The
photocurrent densities measured for the NP, NR and B-NR photoanodes were 0.01, 0.31
21
and 0.83 mA/cm2, respectively, at 0.8 V vs. RHE. Additionally, the incident photon to
current efficiencies of the NP, NR and B-NR photoanodes were 6% at 370 nm, 42% at
380 nm and, 67% at 380 nm respectively. The improved efficiency at the branched
nanorods is attributed to the improved charge carrier separation and collection due to the
small diameters of the branches, leading to a reduced pathlength for photogenerated
charges to reach the semiconductor-electrolyte interface. Moreover, the authors report a
4-fold increase in surface area for the branched nanorods over bare nanorods, leading to
enhanced hole transfer at the semiconductor-electrolyte interface.
Branched nanorods offer better charge transport properties over nanoparticles, as
it has been reported that electron mobilities in nanoparticle films are hindered by trap
states that exist at grain boundaries between particles.61 Additionally, branched nanorods
exhibit improved absorption over nanoparticles as the absorption properties of
nanoparticle films are often limited by reflection losses at the surface.62 Although the
charge transport and absorption properties of nanorods and branched nanorods are
similar, the increased surface area of branched nanorods leads to increased charge
transfer at the semiconductor-electrolyte interface. For these reasons, branched nanorods
represent a promising nanostructure geometry that can be applied to other
semiconductors that exhibit better water splitting capabilities than TiO2.
1.3. Conjugated Organic Polymers
1.3.1. Overview
Conjugated organic polymers (COPs) are macromolecules made up entirely of
main-group elements, such as hydrogen, carbon, nitrogen, oxygen and sulfur.
Conjugation in organic molecules refers to the overlap of adjacent atomic p-orbitals,
forming molecular π-orbitals. This orbital overlap leads to the delocalization of π-electrons
within the molecule. In COPs, conjugation is brought about by alternating single and
double bonds that make up the polymer backbone. If heteroatoms are present, the lone
pair electrons located within the pz orbital of the heteroatom will also contribute to the
conjugation of the polymer backbone. As a result, COPs exhibit semiconducting
properties, lending them to a larger classification of materials referred to as organic
semiconductors.
22
Organic semiconductors look to combine the optoelectronic properties of
traditional inorganic semiconducting materials with the desirable properties associated
with organic molecules, namely mechanical flexibility; synthetic tunability, allowing the
optoelectronic properties of a material to be tailored through synthetic organic chemistry;
and solubility, allowing these materials to be compatible with large scale, roll-to-roll,
solution-processing techniques (e.g., slot die coating, spray coating, ink-jet printing). Along
with COPs, two other main subcategories of organic semiconductors exist: (i) amorphous
molecular films, formed by the deposition of organic small molecules and (ii) molecular
crystals, where conjugated molecules form highly ordered structures, held together
through van-der-Waals interactions. For this thesis, only COPs will be considered further.
Amorphous molecular films and molecular crystals have been reviewed elsewhere.63,64
Although COPs are semiconducting materials, important distinctions regarding
their semiconducting properties exist when compared to their inorganic counterparts. The
first consideration is the nature of conductivity in inorganic semiconductors and COPs.
Intrinsic conductivity at room temperature is attainable for certain inorganic
semiconductors due to free charge carrier generation by thermal excitation. This is due,
in part, to the low band gap energies of inorganic semiconductors (e.g., 0.67 eV for Ge,
1.1 eV for Si and 1.4 eV for GaAs). Additionally, the relative dielectric constant, εr, of
inorganic semiconductors is large (e.g., 16.0 for Ge, 11.7 for Si and 13.1 for GaAs),
resulting in minimal coulombic attraction between photogenerated electron-hole pairs,
allowing charge carriers to conduct freely within the valence and conduction bands upon
excitation. In contrast, εr values of COPs are typically between 2 and 4, leading to large
coulombic attraction between photogenerated electron-hole pairs. As a result, COPs
exhibit exciton diffusion lengths on the order of 10–20 nm, resulting in high recombination
rates. Further discussion on the short diffusion lengths and high recombination rates of
COPs is undertaken in section 1.3.4., along with commonly employed strategies to
overcome these limitations.
As mentioned previously, COPs are characterized by their extended molecular
π-orbitals and delocalized π-electrons. As the number of overlapping π-orbitals increases
with increasing monomer units added to the polymer chain, the density of discrete
molecular orbitals increases until a continuous band is formed. In the ground state, the
lower band is comprised of occupied π-orbitals, while the upper band is made up of
unoccupied π*-orbitals. In organic semiconductors, the highest occupied molecular orbital
23
(HOMO) and lowest unoccupied molecular orbital (LUMO) are analogous to the valence
band maximum and conduction band minimum in inorganic semiconductors, respectively.
Additionally, the overlapping π-orbitals are instrumental in charge transport
through polymer films. High charge carrier mobilities are observed along the polymer
backbone, where π-orbital overlap is strong, and moderate charge carrier mobilities are
observed in the π-π stacking direction for cofacially aligned polymer chains, where
moderate π-orbital overlap exists. The dependence of charge carrier mobility on orbital
overlap results in significantly enhanced mobilities in polymer films that form ordered films.
More recently however, indacenodithiophene-based polymers have been shown
to exhibit high charge carrier mobilities, outperforming highly ordered, semicrystalline
polymers, despite having a nearly amorphous microstructure.65 It should be noted that
although these polymers are often cited as being amorphous, there exists enough small
range order to produce detectable diffraction peaks by synchrotron XRD experiments. To
explain this deviation from the longstanding view that high charge carrier mobilities are
only achievable for polymers with highly crystalline microstructures, Noriega et al. have
proposed a model where the small crystalline domains that exist within the amorphous
polymer matrix are linked together by a common polymer chain extending from one
ordered domain, through the amorphous region, and into the second ordered domain.66
These bridging polymer chains allow charges to be transported between ordered polymer
domains throughout the film. Studies involving these indacenodithiophene polymers have
led researchers to postulate that an increased tolerance to disorder may represent a more
significant factor in designing high mobility polymers than increased crystallinity. Indeed,
further studies have shown that indacenodithiophene polymers display a low degree of
energetic disorder, owing to their planar, torsion-free backbone configuration.67
1.3.2. Electronic Processes in Organic Semiconductors
The following section will provide a brief overview of the ground/excited states and
the vibronic structure of organic semiconductors. The ground state of a molecule refers to
an electron configuration with the lowest possible electronic energy. For conjugated
organic polymers (COPs), this corresponds to a configuration in which electrons occupy
only the molecular orbitals up to and including the HOMO level, leaving the LUMO
unoccupied. When an electron is promoted from the HOMO to the LUMO, the molecule
24
transitions from the ground state to an excited state configuration and is referred to as an
electronic transition.
In addition to electronic energies, vibrational energies must be considered as
molecules are not stationary, but oscillate around their equilibrium position.68 Each
electronic state will, therefore, have multiple vibrational energies associated with it. As a
result, transitions from the ground state to any vibrational level in the excited state can
occur, where the resulting absorption spectrum represents a superposition of the
transitions to different excited state vibrational levels. The simultaneous change in
vibrational and electronic energy levels is referred to as a vibronic transition.
Finally, to obtain a complete description of the polymer’s state, the spin of the
electrons must be taken into consideration, where electron spins may only have a value
of either +½ or -½. If the spins of the unpaired electrons located in the π- and π*-orbitals
of the excited state are antiparallel (i.e., add up to a total spin of zero), the molecule is
said to be in the singlet state. Conversely, if the excited state electron spins are parallel
(i.e., add up to a total spin of one), the molecule is said to be in the triplet state.
Figure 1.7 Orbital configurations of the ground state (S0), the first excited singlet state (S1) and the first exctied triplet state (T1), where the arrows represent the electron spin.
In addition to optical excitations of an electron from the ground to an excited state,
the process of recombination represents an important phenomenon in organic electronics.
During recombination, an electron residing in a π*-orbital of an excited state molecule will
relax down to the ground state, emitting a photon with energy equal to that of the transition.
Kasha’s rule states that the emitting level of a given multiplicity will be the lowest level of
that multiplicity.69 Simply put, for any radiative relaxation process, the transition must
25
originate from the lowest vibrational level for a given excited state. As a result, any electron
excited to a higher vibrational level within the excited state must first lose its vibrational
energy by vibrational relaxation, returning to the zeroth vibrational level of the excited state
before emission can occur. The emissive transition may, however, take place to any
vibrational level in the ground state. Because of Kasha’s rule, the emission spectra of a
given molecule will therefore be red-shifted relative to its absorption spectra, referred to
as a Stokes shift.
Figure 1.8 Jablonski diagram indicating the electronic and vibrational levels of singlet and triplet states, as well as radiative and non-radiative transitions. Excitation processes are represented by straight, solid arrows; radiative recombination processes are represented by straight, dashed arrows; and non-radiative transitions are represented by undulated arrows.
1.3.3. History
The earliest reported conjugated organic polymer appeared in the literature in
1862, where polyaniline was presumably formed by the electrochemical oxidation of
aniline.70 Despite this early discovery, it would take nearly 100 years for the area of
conducting and semiconducting polymers to gain significant traction, beginning with the
26
synthesis of polyacetylene in 1958.71 The development of conjugated polymers continued
into the 1960s, most notably with the synthesis of polypyrrole in 1968.72 Although the
advent of electroactive plastics was exciting, the usefulness of polyaniline, polyacetylene
and polypyrrole remained limited as they were synthesized as insoluble powders. A
tremendous breakthrough occurred in 1974 when researchers were able to successfully
synthesize polyacetylene directly into thin films.73 Just a few years later, a discovery with
even greater significance was made when the collaborative efforts of Heeger, MacDiarmid
and Shirakawa showed that through the doping of polyacetylene films with Cl2, Br2, I2 and
AsF5, their conductivity could be enhanced by a factor of 107.74 In fact, Heeger,
MacDiarmid and Shirakawa were awarded the 2000 Nobel Prize in Chemistry for their
work on electrically conductive plastics. The area of conjugated organic polymers saw an
influx of researchers following the work on chemically doped polyacetylene films, leading
to the discovery of numerous polymers and their applications in organic electronics such
as organic field effect transistors, organic light emitting diodes and organic photovoltaics.
Of the polymers discovered, polyphenylene, polythiophene and polyphenylene
vinylene were especially well studied, however, much like the unsubstituted conjugated
polymers that came before them, these polymers were largely insoluble. In an effort to
synthesize soluble conjugated polymers, it was found that the introduction of alkyl or
alkoxy sidechains to the conjugated backbone afforded solubility to the polymers. Early
examples of soluble conjugated organic polymers are poly([2-methoxy-5-(3,7-
dimethyloctyloxy)]-1,4-phenylene-vinylene) (MDMO-PPV) and poly(3-hexylthiophene)
(P3HT). When combined with a fullerene electron acceptor, these polymers showed
reasonable performance when employed as active layers in photovoltaic devices, e.g., up
to 3% and 6.5% power conversion efficiencies (PCE) for MDMO-PPV75 and P3HT76
respectively. Despite the reasonable PCEs achieved with these polymers, MDMO-PPV
has been plagued by poor morphological control, leading to increased recombination
rates, and P3HT is limited by its rather large band gap of 2 eV and high lying HOMO level,
resulting in poor utilization of the solar spectrum and a small open circuit voltage (VOC),
respectively.77
27
Figure 1.9 Chemical structures of a few select conjugated organic polymers
Synthetic design strategies have led to a class of conjugated organic polymers
known as donor-acceptor (D-A) polymers, where the conjugated backbone is made up of
alternating electron donating and electron accepting units.78 In general, D-A polymers
exhibit smaller band gap energies, owing to the hybridization of the energy levels of the
donor and acceptor moieties. Additionally, the donor-acceptor configuration leads to the
HOMO being located mainly on the donor moiety, while the LUMO is located mainly on
the acceptor moiety. As a result, a partial positive charge resides on the donor unit and a
partial negative charge resides on the acceptor unit upon absorption of a photon, this aids
in stabilizing the photogenerated exciton. While ensuring that a polymer has a sufficiently
low band gap will ensure efficient absorption of the solar spectrum, the HOMO position
will also influence the VOC, where a lower lying HOMO will lead to a higher VOC.‡ Therefore,
a balance between lowering the band gap while maintaining a sufficiently high VOC is
required. Many donor and acceptor building blocks have been synthesized and their
effects on the optoelectronic properties of polymer systems investigated;78–80 and
advances in donor-acceptor polymers have led to single-junction organic solar cells with
PCEs above 10%.81–84
‡ This assumes that the energetics of the acceptor (fullerene or non-fullerene) remain unchanged.
28
Figure 1.10 Chemical structure of two D-A polymers. Donor units are colored red and acceptor units are colored blue.
Although these D-A polymers offer significant advantages over the second-
generation conjugated polymers, like P3HT, they are synthetically challenging to produce,
making them prohibitively expensive to employ in exploratory areas of research, such as
organic photoelectrochemical cells. The research undertaken in this thesis, therefore,
focuses on P3HT, where advances in synthetic methodology, reasonable optoelectronic
properties and extensive literature reports make it an ideal candidate for a proof-of-
concept material in organic photoelectrochemical cells.
1.3.4. P3HT
Poly(3-hexylthiophene) (P3HT) was first developed in 198585 to address the
solubility issues of unsubstituted polythiophene, itself developed only a few years earlier
in 1980.86,87 While the conjugated thiophene backbone is maintained, the hexyl chains
afford solubility to the polymer, allowing for solution processing of polymer films. In
addition to solubility, the incorporation of alkane side chains has significant influence over
the structural, electronic and optical properties of the polymer.88 Due to the asymmetry of
the 3-hexylthiophene monomer, three different monomer coupling configurations exist:
head-to-tail, head-to-head, and tail-to-tail. The term regioregularity is used to describe the
ratio of head-to-tail couplings within the polymer chain. High regioregularity is desired for
producing films with a high degree of order,89 leading to higher charge carrier mobilities
29
compared to regiorandom polymer configurations.90 When regioregularity is sufficiently
high, backbone planarization and aggregation of P3HT chains in the solid state leads to
the formation of crystalline domains. P3HT films are, therefore, characterized as being
semi-crystalline; as such, the degree of crystallinity within P3HT films can measured using
x-ray scattering techniques.91
Figure 1.11 A representation of the three possible coupling configurations of 3-hexylthipohene monomers: head-to-head, head-to-tail, and tail-to-tail. ‘Hex’ is used to represent the hexyl functional group (-C6H13).
P3HT has a band gap of ~2 eV, providing adequate absorption of the solar
spectrum, with P3HT films exhibiting an absorption maximum situated at 515 nm.
Additionally, the large absorption coefficient of P3HT leads to shallow absorption depths,
allowing for efficient collection of solar radiation at thin films, on the order of 100 nm.92 The
small relative dielectric constant of P3HT (εr = 3.5), however, results in exciton diffusion
lengths on the order of 10–20 nm and increased recombination rates. To overcome the
challenges associated with P3HT’s small εr value, electron accepting materials, such as
fullerenes (e.g., [6,6]phenyl-C61-butyric acid methyl ester (PCBM)), or n-type organic
semiconductors, are routinely employed to enhance exciton separation.
The implementation of a bilayer configuration in organic photovoltaics to improve
charge separation, and consequently device efficiency, was first reported by Tang, where
a perylene electron acceptor was deposited on top of a copper pthalocyanine donor,
leading to power conversion efficiencies of 1%.93 Later, the discovery of photoinduced
30
electron transfer from conjugated organic polymers to fullerene (C60) would lead to
fullerenes becoming the electron accepting material of choice in organic photovoltaics.94
Subsequent research in the area of organic photovoltaics focused on designing
polymer/fullerene bilayers to enhance charge separation at the donor-acceptor interface,
however, charge separation in this bilayer configuration is limited to photogenerated
excitons formed within 10-20 nm of the donor-acceptor interface. Any excitons generated
further than one diffusion length away from the interface will recombine, leading to high
recombination rates in the bulk of the semiconductor layer. By increasing the interfacial
area between electron donating and electron accepting domains, the pathlength from the
exciton generation site to the donor-acceptor interface is minimized. This is generally
accomplished by employing a bulk heterojunction (BHJ) morphology, where phase
separation between the donor and acceptor materials is controlled, resulting in an
interpenetrating donor-acceptor blend.95 If the exciton can reach the donor-acceptor
interface before recombination occurs, effective electron transfer from donor to the
acceptor material can occur, while the hole remains in the donor phase. The separated
charges are now free to percolate through the donor and acceptor materials and be
collected at their respective electrodes. A prevailing hypothesis in the area of organic
photovoltaics is that the formation of highly ordered active layer morphologies consisting
of alternating donor and acceptor domains that are twice the width of the exciton diffusion
length should lead to efficient charge separation and excellent charge transport. This has
spurred research efforts to fabricate nanostructured active layers for use in organic solar
cells.96 To date, however, no reports of nanostructured active layers outperforming the
more traditional BHJ approach to device fabrication exist.
Figure 1.12 Diagram depicting the bulk heterojunction morphology achieved when a P3HT:PCBM film is cast on an ITO substrate.
In general, the operation of organic photovoltaic cells and organic
photoelectrochemical cells can be broken down into three major steps: (i) photon
31
absorption, where a photon with energy equal to or greater than the band gap of the
organic semiconductor is absorbed, resulting in the excitation of a ground state electron
in the HOMO to the LUMO, leaving behind a hole. (ii) exciton migration, where the
photogenerated electron-hole pair migrates through the system. During the exciton
migration step, if the electron-hole pair can reach the donor-acceptor interface it can be
separated into mobile charge carriers. If the electron-hole pair is not able to reach the
donor-acceptor interface within the excited state lifetime of the material, radiative
recombination can occur, resulting in the emission of a photon. (iii) Charge collection,
where any separated mobile charge carriers that have made their way through the system
can be extracted at their respective contacts.
1.4. Organic Photoelectrochemical Cells
The first literature example where researchers took advantage of organic
semiconductor properties for solar fuel generation was the demonstration of the aqueous
photoactivity of polyacetylene films in 1981.97 Only a few years later, in 1985, poly(p-
phenylene) was investigated as a heterogenous photocatalyst for the photoreduction of
water to hydrogen gas, in the presence of a sacrificial electron donor.98 The area of organic
photocatalysts for water splitting remains an active area of research, where various
materials such as graphitic carbon nitride polymers,99,100 poly(azomethine) networks,101
pyrene-based conjugated microporous polymers,102 and carbazole-phenylene-based
polymers,103 to highlight a few, are being studied. Although interesting and important to
the development of organic semiconductor electrochemistry, further discussion of organic
heterogeneous photocatalysts will be forgone for the sake of conciseness. The focus of
this section will, therefore, be directed towards organic photoelectrochemical cells
(OPECs) exclusively.
Until recently, the area of OPECs has remained relatively dormant since the
seminal work on polyacetylene photoelectrodes. The only other example of an organic
photoelectrochemical cell for solar fuels generation surfaced in 1996, where El-Rashiedy
established the aqueous photoactivity of P3HT thin film electrodes.104 Since then, P3HT,
aptly referred to as the fruit fly of conjugated organic polymers, has been the photocathode
material of choice for the budding area of OPECs. Appropriate thermodynamics (i.e., the
redox potential of the hydrogen evolution reaction (HER) lies above P3HT’s Fermi level,
but below its LUMO level) and abundant literature studies have made P3HT an obvious
32
choice as a proof-of-concept material from which researchers can better understand the
fundamental principles that govern water splitting processes at organic photoelectrodes.
Figure 1.13 Energy level diagram of a P3HT photocathode on ITO and H+/H2 redox couple in solution.
A renewed interest in OPEC research began in 2012 when Lanzarini et al.
demonstrated the production of stable photocurrent densities at P3HT:PCBM
photocathodes in aqueous NaCl electrolyte.105 Since then, publications in the area of
OPECs have been released regularly, and their numbers steadily increasing. In only a few
years, organic photoelectrode performance for water splitting reactions has increased
dramatically, from current densities of sub-μA/cm2 to milliamps/cm2. These improvements
have come largely from the incorporation of hydrogen evolution catalysts, such as Pt106,107
and MoS3,108 as well as incorporating interfacial charge selective layers. Interfacial layers
are routinely employed in organic electronic devices and serve multiple purposes including
tuning the energy level alignment at the electrode-active layer interface,109 facilitating
ohmic contact; defining the polarity of the electrodes, where either contact can serve as
the anode or cathode depending on the materials employed, leading to device
architectures with increased stability110 and better compatibility with roll-to-roll processing
techniques;111 improving charge selectivity by preventing the flow of unwanted charge
carriers across the interfacial layer (i.e., holes to cathode and electrons to anode) that
would result in recombination and lower device efficiency;112 and controlling the
morphology, where vertical stratification of the donor and acceptor materials of the BHJ
can be controlled through surface energy considerations, leading to enhanced charge
collection at both the top and bottom electrodes.113
33
Figure 1.14 Organic photoelectrochemical cell with a device architechture of ITO/PEDOT:PSS/P3HT:PCBM/MoS3.114 Reprinted from Queyriaux, N.; Kaeffer, N.; Morozan, A.; Chavarot-Kerlidou, M.; Artero, V. J. Photochem. Photobiol. C Photochem. Rev. 2015, 25, 1, Copyright 2015, with permission from Elsevier.
In the context of OPECs, numerous materials have been investigated as hole
transport layers, such as graphene derivatives (e.g., graphene oxide and reduced
graphene oxide),115 amorphous WO3,116 CuI,117 MoO3,118 polyaniline,119 2D MoS2 flakes120
and 3-D MoO3 nanolamellar scaffolds.121 In addition to hole transport layers, electron
transport layers have been used to improve charge transfer between the organic layer and
HER catalyst,122 as well as protect the underlying photoactive organic layer from
degradation.123 These advances in device engineering have led to OPECs with
photocurrent densities as high as 8 mA/cm2 at 0 V vs. RHE, through the incorporation of
a CuI hole transport layer, P3HT:PCBM BHJ active layer, nanostructured TiO2 electron
transport layer and Pt HER catalyst.117 Although these photocurrents are outperformed by
those of p-type silicon photocathodes (27 mA/cm2 at 0 V vs. RHE for silicon nanowire
arrays)124 and copper-based chalcogenides (>20 mA/cm2 at 0 V vs. RHE for CdS and Pt
modified CIGS photocathodes),31 they are comparable to top performing metal-oxide
photocathodes (7.6 mA/cm2 at 0 V vs. RHE for Cu2O with an Al:Zn/TiO2 layer),125 and are
approaching those of group III-V compound materials (11 mA/cm2 at 0 V vs. RHE for
GaInP2 with a graded TiO2/MoOx/MoSx layer).126
Although large photocurrents are impressive on the laboratory scale, ultimately a
technology will have to be compatible with large-scale, cost-effective manufacturing
methods to be commercially relevant. The advancements in device engineering have
34
indeed allowed for all solution processed OPECs to be assembled. The first example of
an all solution processed photocathode employed a device architecture of
ITO/MoOx/P3HT:PCBM/MoS3 and produced photocurrent densities of 2 mA/cm2 at
0 V vs. RHE for one hour.126 Recently, an all solution processed photocathode with a
device architecture of FTO/CuI/P3HT:PCBM/TiO2/Pt/PEI was produced, achieving
photocurrent densities as high as 5.25 mA/cm2 at 0 V vs. RHE.127 This is approaching
champion all-solution processed photocathodes based on inorganic devices capable of
achieving 8 mA/cm2 at 0 V vs. RHE using Bi-modified CuInS2 (CIS(Bi)) in a device
architecture of Mo/CIS(Bi)/CdS/TiO2/Pt photocathodes.128
Demonstrations of OPECs have expanded beyond the hydrogen evolution
reaction (HER) to include the oxygen reduction reaction (ORR), for both dissolved oxygen
detection129 and as an intermediate step in the production of H2O2;130 organic photoanodes
for water oxidation;131,132 and hybrid inorganic/organic tandem cells for unassisted, overall
water splitting.133,134 Additionally, other research areas related to OPECs are emerging,
such as modified fullerenes for both HER135 and ORR,136 as well as the fabrication of
photoelectrodes with hybrid organic/inorganic bulk heterojunctions for use in
photoelectrochemical energy conversion and storage, e.g., P3HT nanofibers with CdS
nanoparticles137 and P3HT with CdSe nanocrystals.138 However, in addition to device
engineering, a fundamental understanding of the energetics and electronic processes at
the organic-electrolyte interface is necessary for future progress, but reports in these
areas are less common.and the few publications that exist focus on studies performed in
the absence of PCBM, i.e., simplified P3HT-electrolyte interfaces.139–143
1.5. Photoelectrode Performance and Experimental Methods
1.5.1. Benchmark and Diagnostic Efficiencies
Solar-to-Hydrogen Conversion Efficiency
Overall solar-to-hydrogen (STH) conversion efficiency is the most important
parameter used to characterize PECs and is used to compare the performance of various
materials against one another, analogous to power conversion efficiency in photovoltaics.
As the most significant and only acceptable global comparison standard, STH efficiency
is regarded as the benchmark efficiency in solar water splitting.144
35
STH efficiency describes the overall efficiency of a photoelectrochemical water
splitting device exposed to broadband solar irradiation under zero-bias conditions, i.e., the
working electrode (WE) and counter electrode (CE) are operated under short circuited
conditions. Additionally, for true STH measurements, the WE and CE must be immersed
in identical electrolytes with the same pH due to the Nernstian bias of 59 mV per pH unit
that would arise from the chemical bias between the two solutions were they to be
contained in separate compartments.145 The electrolyte must also be free from any
sacrificial donor or acceptor species that would otherwise provide additional reaction
pathways. Incident radiation should first pass through an AM1.5G filter so as to simulate
true sunlight reaching the terrestrial surface. The AM1.5G filter provides incident light that
is a good representation of the solar spectrum on a clear day, around noon, for a tilted PV
unit situated at the mid latitudes.146
With the above definitions and conditions in mind, STH efficiency can be
expressed as:
STH = [|jsc|×1.23 V × ηF
Ptotal]
AM1.5G (6)
Where jsc is the short circuit photocurrent density normalized to the electrode area
(in mA/cm2), 1.23 V represents the thermodynamic water splitting potential, ηF is the
faradaic efficiency for hydrogen evolution and Ptotal is the power density of the incident
illumination (in mW/cm2).
In addition to the benchmark STH efficiency measurements, multiple diagnostic
efficiencies exist that provide valuable information about the materials being investigated.
These include the applied bias photon-to-current efficiency (ABPE); the incident photon-
to-current efficiency (IPCE), also referred to as the external quantum efficiency (EQE);
and the absorbed photon-to-current efficiency (APCE), also referred to as the internal
quantum efficiency (IQE). These three major diagnostic efficiencies are discussed below.
Applied Bias Photon-to-Current Efficiency (ABPE)
During the operation of PECs, an external voltage bias may be applied to
compensate for any voltage deficiency or to enhance photogenerated charge carrier
separation by increasing the magnitude of band bending. The application of an external
36
bias, therefore, requires a separate definition from that of the STH efficiency. This is
accomplished by introducing an applied bias term, Vb, to the STH equation:
ABPE = [|jsc|×(1.23 V− |Vb|) × ηF
Ptotal]
AM1.5G (7)
Incident Photon-to-Current Efficiency (IPCE)
IPCE can be summed up concisely as “the number of electrons out versus the
number of photons in”. Practically, photocurrent at the semiconductor working electrode
is measured under monochromatic irradiation and this is compared against the number of
incident photons for the same wavelength of light, calculated using a calibrated photodiode
detector. The photocurrent and number of incident photons can be converted to electron
transfer rate and incident photon flux, respectively. Thus, IPCE is the ratio of the rate of
electron transfer at the semiconductor-electrolyte interface to the incident photon flux,
represented as a function of wavelength.
IPCE (λ) =electrons/cm2/s
photons/cm2/s =
|jph|
Pmono×
1240
λ (8)
Where jph is the photocurrent density (in mA/cm2), Pmono is the calibrated power
density of the incident monochromatic irradiation, λ is the wavelength of the
monochromatic incident irradiation and 1240 represents the multiplication of Planck’s
constant, h, and the speed of light, c.
IPCE represents an important diagnostic as it considers the efficiencies of three
fundamental processes involved during photoelectrochemical water splitting. The first
being the fraction of excitons (i.e., electron-hole pairs) generated per incident photon flux,
ηe−/h+, the second is charge transport of the photogenerated charges to the
semiconductor-electrolyte interface, ηtransport, and the third is interfacial charge transfer,
ηinterface.
IPCE ∝ ηe−/h+ ηtransport ηinterface (9)
Theoretically, integration of IPCE over the entire solar spectrum can be used to
estimate STH if collected under zero-bias conditions, however, this assumes 100%
37
Faradaic efficiency, which may not always be the case if other by products are formed or
corrosion occurs.
Absorbed Photon-to-Current Efficiency (APCE)
Both IPCE and STH implicitly include incident photons that are lost due to
reflectance or transmission. APCE is often used to subtract those lost photons and
measure efficiency based only on photons which are successfully absorbed. This is
accomplished by introducing absorptance, ηe−/h+, defined as the fraction of excitons
generated per incident photon flux, i.e., the first fundamental process described above.
Therefore:
APCE ∝ IPCE
ηe−/h+
∝ ηtransport ηinterface (10)
If a material exhibits high ηe−/h+, but a low IPCE, it can therefore be inferred that
the limiting factor will be either poor charge transport efficiency or sluggish charge transfer
kinetics at the semiconductor-electrolyte interface (or both). By re-running IPCE
measurements with a redox couple possessing better interfacial kinetics, the two limiting
factors can be decoupled as only ηinterface should be drastically affected by the nature of
the redox couple.
Absorptance can be determined by performing UV-Vis measurements to obtain
the absorbance, where absorbance is proportional to the thickness of the semiconductor
film and the concentration of the absorbing chromophore. From Beer’s law:
A = − log (𝐼
𝐼𝑜) (11)
Where A is the absorbance, Io is the incident light intensity and I is the intensity of
the transmitted light. Absorptance is expressed as:
ηe−/h+ = 𝐼𝑜−𝐼
𝐼𝑜 = 1 −
𝐼
𝐼𝑜 = 1 − 10−A (12)
APCE can therefore be expressed as:
38
APCE (λ) = |jph|
Pmono×
1240
λ×
1
1−10−A (13)
1.5.2. Experimental Methods
Many of the parameters discussed in previous sections, such as photocurrent,
faradaic efficiency and built-in potential represent some of the most significant parameters
pertaining to the characterization of PECs. The following section will provide an overview
of the experimental methods used to measure these parameters.
Photoelectrochemical Cell
Photoelectrochemical measurements require a specialized cell that allows for
multiple considerations, such as allowing for irradiation of the semiconductor working
electrode, the ability to maintain an inert atmosphere, sampling of the headspace for
analysis by gas chromatography, and the cell must be outfitted to accommodate the use
of multiple electrodes. The cylindrical body of the photoelectrochemical cell, with flat
windows on either end, allows incident irradiation to pass through the cell; the multiple
ports allow for the use of an inert gas line and various electrodes; and a septum allows for
sampling of the headspace. Typically, photoelectrochemical cells are fabricated from
quartz, to allow high energy UV irradiation to reach the working electrode, however, for
this work the photoelectrochemical cell was fabricated in house and made of glass, which
is opaque to UV light. A diagram of the photoelectrochemical cell used in this work is
presented below in Figure 1.15.
Figure 1.15 Diagram of the glass cell used for photoelectrochemical measurements.
39
Photocurrent
Arguably the most important parameter in solar water splitting, photocurrent is a
direct measure of device performance and appears in all the efficiency calculations
mentioned above. Photocurrent is measured by potentiostatic coulometry, also referred to
as electrolysis, under chopped illumination. Electrolysis measurements are performed
using a conventional three-electrode setup, where the working electrode is held at
constant potential and current is measured as a function of time. In photoelectrochemical
cells, a light source is used to illuminate the photoelectrode intermittently, resulting in three
distinct regions. The first region is the dark-current, which represents any current
measured while the photoelectrode is not being illuminated. As the redox processes
occurring at the semiconductor-electrolyte interface are driven by photogenerated
charges, dark current in photoelectrochemical cells is generally negligible. The second
region of importance is the steady-state photocurrent, which is the current measured while
the photoelectrode is under illumination. Finally, photocurrent transients appear during the
initial illumination due to charging processes that occur at the semiconductor-electrolyte
interface.147 The value of photocurrent reported is the difference between the steady-state
photocurrent and the dark current. A typical electrolysis plot is shown in Figure 1.16, with
the regions of interest highlighted.
Figure 1.16 Example of a typical electorolysis trace. The three important regions are labelled, where i. represents the dark region, ii. the photocurrent transient region, and iii. the steady-state photocurrent region.
40
Faradaic Efficiency
Although photocurrent is largely considered the best metric to determine overall
performance of PECs, bulk electrolysis is incapable of differentiating photocurrents due to
desirable water splitting reactions and other parasitic currents, such as the formation of
by-products or corrosion. Differentiation of the origin of photocurrent is crucial when
determining overall STH efficiency as the inclusion of any parasitic photocurrents will
artificially inflate the STH efficiency. The Faradaic efficiency is the fraction of photocurrent
that is used for water splitting reactions versus the total measured photocurrent. To
determine the Faradaic efficiency of a PEC, electrolysis and gas chromatography are
combined, where the theoretical amount of hydrogen gas produced at the cathode is
calculated from the amount of charge passed during electrolysis and the real amount of
hydrogen gas present after electrolysis is measured by gas chromatography. The ratio of
the measured hydrogen to the theoretical amount of hydrogen produced yields the
Faradaic efficiency. Faradaic efficiencies can be maximized, and often approach unity, by
limiting the number of competing charge transfer pathways. This can be accomplished
through purging of the electrolyte with inert gas to remove any dissolved gases and using
high purity electrolytes.
Built-In Potential
The built-in potential (Vbi) is one of several factors that determines the depletion
layer width at the semiconductor surface and will consequently influence charge
separation within the semiconductor. Both open circuit potential (OCP) and linear sweep
voltammetry (LSV) can be used to measure Vbi. OCP measurements are performed using
a conventional three-electrode setup under zero-current conditions (i.e., at open circuit).
The potential is measured as a function of time, beginning under dark conditions and then
under illumination. The difference between the OCP measured under dark and illuminated
conditions is known as the photovoltage (Vph) of the system. At sufficiently high light
intensity, all pre-existing band bending at the semiconductor surface is removed and Vph
is equal to Vbi.
Vbi can also be estimated by LSV, where LSV measurements are also carried out
using a three-electrode setup and current density is measured as a function of potential
under chopped illumination. For p-type semiconductor photoelectrodes, potentials are
swept from more anodic (i.e., positive) potentials to more cathodic (i.e., negative)
41
potentials. The potential corresponding to the onset of photocurrent (Vonset) represents the
point at which minority carriers are transferred from the semiconductor to the electrolyte,
i.e., the onset of band bending. The difference between Vonset at the illuminated
semiconductor and the reduction potential (Eredox) of the redox couple will therefore be
equal to Vbi when kinetic overpotentials are minimized. Eredox can be found experimentally
by performing cyclic voltammetry measurements, employing a working electrode with low
polarization overpotentials, such as Pt.
Figure 1.17 a) An idealized example of a typical OCP measurement, where regions i and ii represent the OCP measured under dark and illuminated conditions, respectively. b) A linear sweep voltammogram under chopped illumination, a vertical dashed line has been drawn from the onset of photocurrent to highlight Vonset.
Ultraviolet-Visible Spectroscopy
Ultraviolet-Visible (UV-vis) spectroscopy is a form of absorption spectroscopy,
where the sample is irradiated with light in the ultraviolet-visible region of the solar
spectrum. UV-vis spectroscopy measures the absorbance of the sample as a function of
wavelength, where absorbance is determined by measuring the ratio of the intensities of
the incident and transmitted light at a given wavelength. Where A is the absorbance, Io is
the intensity of incident light and I is the intensity of the transmitted light.
A = log𝐼o
𝐼 (14)
In addition to measuring how strongly a material absorbs light of a certain
wavelength, UV-vis can provide information regarding the optoelectronic properties of
conjugated organic polymers, such as the optical band gap, and insight into polymer
42
aggregation and film morphology. The optical band gap of COPs can be determined from
the wavelength at which the onset of absorption occurs by converting the onset
wavelength into energy. Although UV-vis can be used as a simple method to estimate the
band gap of a material, it should be noted that the optical band gap is not a precise
measurement of the difference between the HOMO and LUMO levels of organic
semiconductors, due to poorly screened electron-electron interactions.148 Backbone
planarization and formation of semi-crystalline domains can be observed in the P3HT UV-
vis spectra through the appearance of the S1 ← S0 0-0 transition, manifesting itself as a
vibronic shoulder at lower energy wavelengths.149 The planarization of the P3HT backbone
is associated with aggregation, where the relative peak ratio of the 0-0 and 0-1 transitions
can provide information on whether H- and J-type aggregation is dominant, as proposed
by Spano.150,151
Photoluminescence Spectroscopy
Photoluminescence (PL) spectroscopy is a form of emission spectroscopy, where
a sample is excited using a light source (typically using monochromatic light corresponding
to the maximum absorption wavelength of the sample) and the intensity of the emitted
radiation is measured as a function of wavelength. PL spectroscopy can provide important
insights into the radiative recombination rates within COPs as successful electron transfer
from the donor to acceptor phase results in the quenching of photoluminescence intensity.
PL peak ratios can also be used in accordance with the H-aggregate model to provide
information regarding polymer chain aggregation.
Grazing Incidence Wide Angle X-Ray Scattering
Grazing incidence wide angle x-ray scattering (GIWAXS) is a subset of x-ray
scattering techniques, where grazing incidence refers to the operation of x-ray scattering
in the reflection scattering geometry, as opposed to the traditional transmission scattering
geometry. In GIWAXS measurements, the incident x-ray beam approaches the sample at
a very shallow angle, that is less than the critical angle of the sample material, to ensure
all external reflection (e.g., P3HT has a critical angle of 0.178° for x-rays with a wavelength
of 1.55 Å) and scattering is measured as a function of exit angle (yz plane) and in-plane
angle (xy plane). The size of the angles measured is dependent on the sample-to-detector
distance, where wide angles (i.e. GIWAXS) are measured when the detector is situated
10-50 cm from the sample and small angles (i.e., GISAXS) at 130-300 cm.152
43
Consideration of Bragg’s Law dictates that larger scattering angles allow for the resolution
of smaller feature sizes.153 Therefore, GIWAXS is typically employed for the
characterization of P3HT thin film morphology, due to its ability to probe the atomic and
molecular distances within the crystalline domains.
Figure 1.18 Diagram showing the scattering geometry of a typical GIWAXS measurement.
Specifically, GIWAXS can be used to quantify the distribution of polymer chain
orientation within P3HT thin films. Within the crystalline domains, P3HT polymer chains
form a unit cell comprised of two main orientations: (i) Lamellar stacking, where polymer
chains stack in the direction of the hexyl sidechains. This is taken as the (100) direction.
(ii) π-stacking, where the planarized, conjugated backbones of adjacent polymer chains
align. This is considered the (010) direction. Any alignment of the polymer chains in the
direction of conjugation, is generally too disordered and therefore no observable (001)
peaks appear in P3HT scattering profiles.
44
Figure 1.19 Diagram of an idealized P3HT unit cell of where the (100) and (010) orientations are shown in the lamellar and π-stacking directions, respectively.
When cast as a thin film, the P3HT polymer chains adopt two major orientations
with respect to the substrate, the edge-on orientation, where the lamellar stacking direction
is normal to the substrate and the face-on orientation, where the π-stacking direction is
normal to the substrate. Depending on the application, different orientations will be favored
as charge transport within P3HT films has been shown to be much higher in the π-stacking
direction compared to the lamellar stacking direction.154 In organic field-effect transistors,
for example, an edge-on orientation with in-plane π-stacking is desired, whereas for
devices with a vertical architecture, such as organic photovoltaics (OPVs) and organic
photoelectrochemical cells (OPECs), the face-on orientation with π-stacking normal to the
substrate is preferred. GIWAXS provides quantitative information on the orientation of the
polymer chains in the solid state and is therefore an important measurement when
characterizing the morphology of polymeric thin films.
Dynamic Light Scattering
Dynamic light scattering (DLS) is a common method used to determine the size of
nanoparticles, where the Stokes-Einstein relationship is used to relate particle motion and
particle size. DLS functions by passing a monochromatic laser through a diluted
nanoparticle solution, where light is scattered and collected at a detector situated at 90
degrees to the incident laser beam. As the nanoparticles move due to Brownian motion,
the scattering pattern at the detector also changes. An auto-correlation function tracks the
fluctuation of the scattering intensity at the detector and provides insight into the diffusion
45
coefficient of the nanoparticles. The hydrodynamic diameter of the nanoparticles can then
be calculated using the Stokes-Einstein relationship:155
dh =kBT
3πηD (15)
where dh is the hydrodynamic diameter, kB is the Boltzmann constant, T is the temperature,
η is the viscosity of the solvent and D is the diffusion coefficient.
1.6. Thesis Scope
The following chapters will explore the photoelectrochemical properties of P3HT,
with a particular interest in its application as a photocathode material to perform the
hydrogen evolution reaction. Although solar water splitting technologies represent a
potential source of abundant renewable energy in the future, the objective of this thesis is
not necessarily to engineer the highest performing photocathodes from conjugated
organic polymers, but rather to gain a better understanding of the thermodynamic and
electronic processes that take place at the organic semiconductor-electrolyte interface.
Chapter two explores the origin of photocurrent at single layer P3HT films
deposited on ITO-coated glass in acidic aqueous solutions. Where it was previously
thought that the sub-μA/cm2 photocurrents observed at the illuminated P3HT-electrolyte
interface was due to the reduction of protons to hydrogen gas, quantitative evidence is
provided that suggests any photocurrents measured at uncatalyzed P3HT photocathodes
are largely due to the reduction of dissolved oxygen. Despite favorable thermodynamics,
this work demonstrates the need for hydrogen evolution catalysts to be employed in
concert with the P3HT photoabsorbing layer to successfully achieve hydrogen evolution.
In chapter three, P3HT:PCBM nanoparticles are synthesized and used to fabricate
nanostructured organic photocathodes. The morphology and optoelectronic properties of
the nanostructured photocathodes are compared with conventional, solution-cast thin
films of P3HT:PCBM, where control of the nanoscale morphology within each nanoparticle
leads to enhanced P3HT:PCBM phase segregation. The photo-assisted deposition of
platinum nanoparticles as hydrogen evolution reaction catalysts onto the P3HT:PCBM
photocathodes facilitates the photoreduction of protons to hydrogen gas and increased
hydrogen production is demonstrated at the nanostructured electrodes.
46
The role of PCBM at the semiconductor-electrolyte interface is discussed in
chapter four. Specifically, the influence of PCBM on the energy level alignment of the
redox couple in solution and the formation of an interfacial dipole layer are explored in
detail. It is found that the presence of PCBM results in a decreased built-in potential and
depletion layer thickness in P3HT photocathodes. This was shown to influence
photocathode performance in thinner photoelectrodes, where efficient charge separation
was achieved in the absence of a PCBM acceptor phase by tuning the semiconductor
thickness to coincide with depletion layer width.
Final remarks, followed by recommendations for future directions in the field of
organic photoelectrochemical cells mark the end of this thesis. The need for rational
polymer design and organic nanostructures highlights the future directions section.
Figure 1.20 Outline of this thesis organized in a flow chart.
47
Chapter 2. On the Origin of Photocurrent at Uncatalyzed P3HT Films in Aqueous Acidic Media
Sections of this chapter have been reproduced in part with permission from:
Suppes, G. M.; Fortin, P. J.; Holdcroft, S. J. Electrochem. Soc. 2015, 162, H55. Copyright
2015, The Electrochemical Society. The work presented in this chapter are the
contributions made to the paper by P. Fortin.
2.1. Introduction
At the onset of this project, very few reports concerning the use of conjugated
organic polymers as photocathodes for the hydrogen evolution reaction existed. At the
time, the goal of the project was to determine whether it was possible to evolve hydrogen
at uncatalyzed P3HT photocathodes. Combing through the literature, it seemed as though
it was possible. Examples of conjugated organic polymers used in sacrificial water splitting
systems can be found as early as 1985 wherein it was reported that irradiation of a
suspension of poly(p-phenylene) in the presence of a diethylamine hole scavenger led to
the photoreduction of protons to hydrogen gas.98 It was subsequently reported, using a
photoelectrochemical system, that P3HT exhibited photocathodic activity in aqueous
solutions and that photocurrent could be increased by incorporation of electron acceptors,
such as fullerene.104 The photocurrents measured at these organic photocathodes were
interpreted as being commensurate with the hydrogen evolution reaction. Additionally,
Lanzarini et al. demonstrated that P3HT:PCBM films on ITO generates stable
photocurrent in the sub-µA/cm2 range in aqueous NaCl,105 concluding that the
photocurrent was due to proton reduction at the P3HT:PCBM electrode and oxidation of
Cl- at the counter electrode.
However, careful examination of the literature reports concerning
photoelectrochemical hydrogen evolution at uncatalyzed P3HT surfaces revealed that, in
fact, no quantitative analysis (e.g., gas chromatography) had been performed to confirm
the presence of hydrogen gas. Given this revelation, along with the advancements in the
field that indicate an apparent need for catalysts to promote hydrogen evolution at
48
irradiated conjugated polymer films, the origin of photocurrent at single layer P3HT films
in aqueous solution was re-examined.
One possible reaction pathway competing with the hydrogen evolution reaction is
the one electron reduction of oxygen to the superoxide radical anion, O2•-. In aqueous
acidic media, the superoxide radical anion undergoes the following disproportionation
reaction to produce hydrogen peroxide (H2O2) and molecular oxygen (O2):156
𝑂2 + 𝑒− → 𝑂2•−
(16)
O2•− + H+ → HO2
• (17)
HO2• + O2
•− + H+ → H2O2 + O2 (18)
where the oxidation state of oxygen changes from -½ in the superoxide radical anion to -
1 in hydrogen peroxide and 0 in molecular oxygen.
In this chapter, the propensity of uncatalyzed P3HT photocathodes to perform the
hydrogen evolution reaction (HER) and oxygen reduction reaction (ORR) is investigated.
Headspace analysis, by gas chromatography, is performed after photoelectrolysis to gain
insight into whether hydrogen is created at irradiated P3HT photocathodes. Additionally,
the relationship between photocurrent density and dissolved oxygen concentration is
examined to determine additional reduction pathways that may occur at irradiated P3HT
photocathodes.
2.2. Experimental
2.2.1. Materials
Regioregular poly(3-hexylthiophene) used for the P3HT-O2 photoelectrochemical
studies was synthesized according to a literature procedure.157 The number-average
molecular weight of the polymer was 35,600 kDa with a polydispersity index (PDI) of 1.66
as determined by gel permeation chromatography using tetrahydrofuran (Sigma-Aldrich,
HPLC grade) eluant, three Waters µStyragel columns at 40°C, a Waters 1515 isocratic
HPLC pump, a Waters 2414 differential refractometer, and a Waters 2487 dual UV
absorbance detector (λ = 254 nm). Polystyrene standards (Waters) were used for
calibration. The degree of head-to-tail regioregularity was determined to be 94% by 1H
49
NMR spectroscopic analysis, according to standard literature procedure.158,159
Chlorobenzene (Sigma-Aldrich), H2SO4 (Anachmia), NaOH (BDH chemicals), NaCl
(Fisher), Na2SO3 (Fisher) were all used as received. All aqueous solutions were prepared
using water obtained from a Milli-Q water purification system (18 MΩ, EMD Millipore).
Indium tin oxide (ITO) coated glass slides (15 Ω/□) (Colorado Concept Coatings LLC) were
cleaned by sequential sonication in each of the following solutions: isopropyl alcohol,
acetone, H2O: H2O2: NH4OH (5:1:1) solution and distilled water.
2.2.2. Film Preparation
Films for P3HT-O2 photoelectrochemistry were prepared from a 25 mg/mL
chlorobenzene solution that had been stirred overnight at 65˚C. The solution was spin-
cast onto chemically cleaned ITO in a glove box (H2O < 0.1 ppm, O2 < 1 ppm) at 1000 rpm
for 1 minute, affording polymer films ~100 nm thick. Film thicknesses were determined
using an Alpha-Step IQ profilometer (KLA-Tencor). The edges of each electrode were
covered with chemically-resistant and electrically insulating epoxy purchased from
Epoxies Inc. (20-3004 HV) to prevent exposure of ITO to the electrolyte.
2.2.3. Photoelectrochemistry
The light source was a 200 W Xe/Hg lamp (Uhsio America, inc.), used in
combination with a 300-700 nm band pass filter (FSQ-KG3, Newport Corp.) and neutral
density filters (Thorlabs Inc.), to achieve 100 mW cm-2 irradiation, as measured using a
broadband power meter (841-PE, Newport Corporation) equipped with an Ophir thermal
detector head (3A-P-SH-V1). The cell configuration was designed to allow irradiation of
the polymer-electrolyte interface through the electrolyte. A water filter was placed in front
of the electrochemical cell to remove excess heat. Electrochemical measurements were
performed using a Pine Bipotentiostat (AFC-BP1). Photoelectrochemical measurements
were performed in a 3-electrode configuration using a saturated calomel electrode (SCE)
(+0.24 V vs. RHE) and a Pt wire as the reference and counter electrodes, respectively.
Solutions were sparged with argon (Praxair, purity 99.999 %) for 1.5 to 2 hours in a glass
PEC cell and sealed under a flow of argon. Pt was photoelectrochemically-deposited onto
P3HT-coated ITO electrodes from 0.1 mM K2PtCl6 in 1 M H2SO4 at -0.1 V vs. SCE
(0.14 V vs. RHE) for 60 s under 100 mW cm-2 irradiation.
50
2.2.4. Hydrogen Gas Detection
The headspace in the electrochemical cell was sampled using a 5 mL syringe,
fitted with a gas tight valve (Series A-2, VICI Precision Sampling), and analyzed using an
Agilent Technologies 6890N GC system equipped with a thermal conductivity detector. A
2.13 m Agilent J&W GC packed column in stainless steel tubing was used (inner diameter
of 2 mm, HayeSep N packing material, 60/80 mesh size). Argon was used as a carrier gas
at a flow rate of 30 mL/min1 under 46.2 psi.
2.2.5. Dissolved Oxygen Detection
An oxygen probe (Orion 850 polarographic probe) was used to measure dissolved
molecular oxygen in solution. Unless specifically stated, solutions were sparged with
argon (Praxair, purity 99.999 %) for 1.5 to 2 h prior to use and sealed under a flow of
argon, resulting in a dissolved oxygen concentration <0.01 ppm as measured by the
dissolved oxygen probe. For other experiments, the dissolved oxygen content of the
electrolyte was controlled either by using solutions exposed to ambient air, referred to as
“ambient air” solutions, having a dissolved O2 concentration of 7.80 ppm, or by bubbling
the solutions with oxygen for ten minutes, resulting in oxygen saturated solutions having
a dissolved oxygen concentration >20 ppm.
2.3. Results
2.3.1. P3HT photocurrent vs. O2 concentration
Sub-µA/cm2 photocathodic current is observed under potentiostatic conditions
when a P3HT-coated ITO electrode is immersed into 0.1 M H2SO4 solution purged with
argon, as illustrated in Figure 2.1. In the dark, no cathodic current is observed. As shown
in Figure 2.1, the initial photocurrent observed upon irradiation decays to a pseudo-steady
state. Upon removing the light source and re-illuminating the electrode, the initial current
spike decreased in magnitude and eventually disappeared, rendering a steady state
current of 200-300 nA/cm2. A similar photocurrent profile is observed in 0.1 M NaCl and
0.1 M NaOH, but the magnitudes of the photocurrents are smaller. It was considered that
the photocurrent was due to the reduction of protons, leading to the hydrogen evolution
51
reaction (HER), because the conduction band electrons of P3HT are sufficiently reductive
to reduce protons in solution.
Figure 2.1 Electrolysis of P3HT-coated ITO electrodes at -0.24 V vs. SCE (0 V vs. RHE) performed in 0.1 M H2SO4 (black), 0.1 M NaCl (dotted), 0.1 M NaOH (dashed), and 0.5 M Na2SO3 oxygen scavenger (dot-dashed). Light intensity: 100 mW cm-2 (300-700 nm), under chopped illumination cycles of 30 s dark / 1 min light.
To examine whether the photocurrent was indeed due to HER, the headspace of
the electrolysis cell was sampled after 6 h of photoelectrolysis, after passing ~56 mC
at -0.24 V vs. SCE (0 V vs. RHE) of cathode charge—equivalent to 0.29 μmol of H2 gas
(assuming 100% current-to-gas efficiency). However, the quantity of H2 gas measured by
GC, if at all produced, was below the detection limit of the instrument. In a series of control
experiments, hydrogen was evolved at a Pt foil working electrode by application
of -0.4 V vs. SCE (-0.16 V vs. RHE) for 10 s, passing 32.5 mC of charge, equivalent to
0.17 μmol of H2; and for 60 s, passing 202.5 mC of charge, equivalent to 1.05 μmol of H2.
Electrolytically-produced H2 was readily detectable by GC, as shown in Figure 2.2,
providing compelling evidence that the photocurrent observed at P3HT-coated ITO
electrodes is not commensurate with hydrogen evolution.
52
Figure 2.2 Gas chromatographic analysis of H2 gas in the photoelectrochemistry cell after 202.5 mC of charge over 60 s (dot-dashed) and after passing 32.5 mC of electrolytic charge over 10 s (dashed) at a Pt foil (i.e., conventional electrochemical HER); 6 h photoelectrolysis at a P3HT-ITO electrode (dotted) after passing 56.6 mC of cathodic charge. The solid black signal represents a sample of air (i.e, blank signal). Signals are offset for clarity.
The influence of trace oxygen was investigated as the source for photocathodic
current, as oxygen has been shown to act as an electron acceptor in solution.156,160,161
After purging the solutions with inert gas, the concentration measured by the oxygen probe
was below the limits of detection, <0.1 ppm. However, the addition of 0.5 M sodium sulfite,
a known oxygen scavenger,162 reduced the observed photocurrents shown in Figure 2.1
to non-detectable values. This confirms that trace oxygen remaining after purging with
argon is likely responsible for the photocurrent. Moreover, potentiostatic photoelectrolysis
(Figure 2.3) measurements in 0.1 M H2SO4 with increasing dissolved oxygen contents
(<0.01, 7.80, and >20.0 ppm) led to increasing photocurrent and no dark current.
Potentiostatic measurements revealed that, in the presence of a high concentration of
dissolved oxygen, the magnitude of the transient current spike observed immediately
following illumination returns upon repeated on-off cycles. In contrast, the cathodic current
spike fades after the first on-off cycles in the absence of a replenishing amount of oxygen.
The origin of this is uncertain but may involve the well-known P3HT-O2 charge transfer
complex that forms upon their contact and which serves as an electron acceptor,163,164 and
which has been observed for P3HT films in contact with water.165
53
Figure 2.3 Photoelectrolysis of P3HT coated ITO electrodes in 0.1 M H2SO4 with different dissolved oxygen concentrations. Oxygen deficient (<0.01 ppm) conditions (red), ambient (7.80 ppm) conditions (green) and oxygen saturated (> 20.0 ppm) conditions (blue). Light intensity: 100 mW cm-2 (300-700 nm). The electrodes were biased at -0.24 V vs. SCE (0 V vs. RHE). Chopped illumination cycles consist of 30 s dark / 1 min light.
That P3HT is capable of photoelectrochemically-reducing oxygen is not
unexpected since other PEC reactions are also possible if the appropriate redox couple is
present and provided the redox energy level lies below the energy of the conduction band
of P3HT. The overriding feature of these experiments is that the measured photocurrent
at the P3HT-coated photocathodes in pure aqueous or acidic solution appears largely to
be due to the reduction of dissolved oxygen in solution, even when the electrolyte is
flushed with argon for extended periods of time. Going to greater lengths to exclude
oxygen, such as the addition of oxygen scavengers, reduces the photocurrent to negligible
values. Thus, despite favourable energy level positions of the P3HT LUMO in relation to
the H+/H2 redox couple, and the observation of pH dependence on photocurrent, the HER
is not the predominant mechanism.
2.4. Conclusion
Despite the conduction band electrons of P3HT possessing sufficient redox
potential to reduce protons, no hydrogen gas was detected at irradiated, single layer,
pristine P3HT films. Additionally, when the concentration of dissolved oxygen was
54
controlled, it was observed that photocurrent density was proportional to dissolved oxygen
concentration. Together, these results suggest that the origin of photocurrent at
uncatalyzed P3HT photocathodes is largely due to the reduction of residual O2 in solution.
In the presence of an oxygen scavenger, photocurrents are reduced to negligible amounts.
This work provides quantitative evidence that a catalyst layer must be used in
conjunction with the organic photoabsorbing layer to promote photoelectrochemical
hydrogen generation. This work provided the foundation for further research carried out in
the Holdcroft lab, where platinum nanoparticle catalysts were deposited
photoelectrochemically onto P3HT films and hydrogen gas was successfully evolved by
photoelectrolysis.107
55
Chapter 3. Hydrogen Evolution at Conjugated Polymer Nanoparticle Electrodes
Sections of this chapter have been reproduced in part with permission from:
Fortin, P. J.; Rajasekar, S.; Chowdhury, P.; Holdcroft, S. Can. J. Chem. 2018, 96, 148.
Copyright 2018, NRC Research Press. The work presented in this chapter are the
contributions made to the paper by P. Fortin.
3.1. Introduction
With the establishment that hydrogen evolution does not occur at uncatalyzed
P3HT photocathodes (Section 2.4) and a method to incorporate a hydrogen evolution
catalyst onto the photocathodes has been developed, strategies to further improve
photocurrent densities at organic photocathodes were explored.
A strategy often adopted in inorganic photoelectrode materials chemistry for
increasing photocurrent density is to increase the surface area of the electrode by way of
forming a nanostructured electrode surface, as demonstrated with a range of
nanostructured semiconductor photoelectrodes, each of which offer distinct
improvements. For instance, high aspect ratio nanowire electrodes are able to overcome
charge transport constraints of planar electrodes by minimizing the exciton diffusion path
length to the semiconductor-electrolyte interface, so that photogenerated holes and
electrons can be efficiently separated prior to their recombination.166 Nanoholes have been
integrated into 20 µm-thin Si electrodes and decorated with Pt nanoparticle catalysts.167
The high photocurrent densities achieved using these structures (23 mA/cm2 at
0 V vs. RHE) is attributed to the antireflective properties of the nanoholes, which increases
light absorption in the thin electrodes, as well as increasing the charge transfer kinetics
due to suppression of surface recombination. Recently, nonprecious-metal catalysts were
deposited onto pyramidally, nanotextured silicon substrates,168 and the enhanced
interfacial electrode area used to increase the amount of catalytic sites accessible for the
hydrogen evolution reaction, yielding photocurrents of 12 mA/cm2.
56
Organic semiconducting polymers have the distinct advantage over inorganic
counterparts in that they can be cast from simply-prepared polymer solutions. However,
when cast, they typically form smooth, uniform films leading to an electrode surface area
that is similar to the geometric area of the substrate onto which the film is cast. Drawing
from strategies used in the PEC studies of inorganic films it ought to be possible to
increase the surface area of the polymer film by adopting a nanoparticle approach to film
formation.
Nanoparticles of P3HT and P3HT:PCBM have indeed been the subject of
extensive characterization in recent years as a means to control the morphology of
photoactive layers in organic photovoltaic light harvesting systems in order to offset the
relatively short diffusion length of photogenerated excitons in organic materials.169–174
Nanoparticulate systems also allow deposition of photoactive components from water and
alcohols rather than from traditional chlorinated solvents, as demonstrated for organic
photovoltaic (OPV)175–177 and organic field effect transistor (OFET)178 devices. Organic
polymer and polymer:fullerene nanoparticle dispersions may be prepared by one of two
methods: (i) a mini-emulsion method, wherein a polymer or polymer:fullerene blend is
dissolved in a solvent and subsequently added to a surfactant-containing non-solvent,
where the solvent and non-solvent are immiscible and (ii) a surfactant-free precipitation
method, in which a polymer or polymer:fullerene blend is dissolved in a solvent and directly
injected into a non-solvent, for which the solvent and non-solvent are miscible. Both
methods of nanoparticle formation are illustrated in Figure 3.1.
Although surfactant-stabilized nanoparticles of P3HT:PCBM have been shown to
be smaller than the surfactant-free nanoparticles,179 they form core-shell structures, with
a PCBM-rich core, which may inhibit effective photogenerated charge separation and
collection, whereas the surfactant-free precipitation method forms nanoparticles that
possess a relatively uniform, blended morphology.180 With the surfactant-stabilized
nanoparticle approach, trace amounts of surfactant may remain incorporated in the active
material, potentially interfering with the formation of the bulk-heterojunction, and thus
impinging upon device performance and device lifetime.177 The surfactant-free approach
to nanoparticles mitigates these potential issues. In this method, the polymer of
57
Figure 3.1 General schemes showing the formation of organic polymer nanoparticles by (a) the mini-emulsion method and (b) the precipitation method.
choice is dissolved in a “good” solvent and the solution injected, under control, into a “poor”
solvent (the non-solvent) so that upon contact with the non-solvent the polymer chains
collapse to form spherical-like nanoparticles.
The concentration of the original polymer solution affects the size of the
nanoparticles formed. As the concentration is increased, the number of polymer chains
per unit volume is increased, leading to larger nanoparticles. In considering the formation
of bulk heterojunction films of donor and acceptor molecules, nanoparticles with the
smallest possible diameters are assumed to be preferred for photovoltaic and
photoelectrochemical applications in order to minimize the donor and acceptor domain
sizes, thereby reducing the exciton diffusion length required to reach the heterojunction
interface. Two key variables essential to controlling nanoparticle preparation are: (i) the
nature of the solvent and non-solvent, which must be miscible to ensure formation of a
58
homogenous colloidal solution of the polymer; (ii) the concentration of the initial polymer
solution, which needs to be low enough so as to form nanoparticles, yet high enough to
ensure the colloidal solution formed provides films sufficiently thick so as to be examinable
as photocathodes when cast on a substrate. Herein, we report the vacuum-free
preparation of high surface area organic photoelectrodes using surfactant-free
P3HT:PCBM nanoparticles and the subsequent investigation of their morphological and
optoelectronic properties. The influence of these properties on their performance as
organic photocathodes, when coupled with a Pt hydrogen evolution catalyst, is
investigated and compared to photoelectrodes prepared from solution-cast P3HT:PCBM
films.
3.2. Experimental
3.2.1. Materials
P3HT was purchased from Rieke Metals (MW = 50-70 kg/mol, regioregularity: 91-
94%, <0.01% metal impurities). PC61BM (American Dye Sources, Inc.), K2PtCl6 (Johnson
Matthey) and all solvents were all used as-received. All aqueous solutions were prepared
using DI water obtained from a Milli-Q water purification system(18MΩ, EMD Millipore).
Indium tin oxide (ITO) coated glass slides (15 Ω/□) (Colorado Concept Coatings LLC) were
cleaned by sequential sonication for 10 minutes in each of the following solutions:
dichloromethane, deionized water and isopropyl alcohol. Chemically-cleaned ITO slides
were subjected to a plasma treatment using 25% v/v O2 in Ar for 10 minutes (Fischione
Instruments Model 1020 plasma cleaner).
3.2.2. P3HT:PCBM Nanoparticle Films
Solutions of P3HT:PCBM was prepared by first dissolving 10 mg of P3HT in 1 mL
CHCl3 and 10 mg of PCBM in 1 mL CHCl3 , stirring each overnight at 50°C. The solutions
of P3HT and PCBM were mixed together and 1.5 mL of this solution was added by syringe
to 4 mL of isopropyl alcohol (IPA) under sonication at room temperature for 1 minute
(Branson 1510 ultrasonic cleaner). A dispersion of nanoparticles of P3HT:PCBM was
obtained, having a final concentration of 2.7 mg/mL. Next, 100 µL aliquots of the
nanoparticle dispersion were spin-cast onto ITO-coated glass slides at 2000 rpm multiple
times to achieve the desired thickness. A film thickness of ~90 nm required 10 spin-coats.
59
3.2.3. Deposition of Pt
Platinum nanoparticles were photoelectrochemically-deposited onto films of
P3HT:PCBM nanoparticles from 0.1 mM K2PtCl6 in 1 M H2SO4 by application of an
electrode potential of −0.1 V vs. SCE under chopped (100 mW/cm2) illumination using
5 second dark and 15 second light intervals for a total irradiation time of 60 seconds. The
amount of Pt deposited on the polymer films was calculated using Equation 19:
mPt = Q
nF∗ MWPt (19)
where mPt is the mass of Pt deposited via photoelectrochemical deposition, Q is the charge
passed from the working electrode to the electrolyte and is measured by taking the area
above the Pt deposition curves, F is Faraday’s constant, MWPt it the molecular weight of
Pt and n, the number of electrons, is four.
3.2.4. Photoelectrochemistry
A 200 W Xe/Hg lamp (Uhsio America, inc.) was used in combination with a
300-700 nm band pass filter (FSQ-KG3, Newport Corp.) and neutral density filters
(Thorlabs Inc.), to achieve 100 mW/cm2 irradiation, as measured using a broadband
power meter (841-PE, Newport Corporation) equipped with an Ophir thermal detector
head (3A-P-SHV1). A glass photoelectrochemical cell with two large flat windows and five
accessory ports was used and the cell configuration was designed to allow irradiation of
the polymer–electrolyte interface through the electrolyte. A water filter was placed in front
of the electrochemical cell to remove excess heat. Electrochemical measurements were
performed using a Pine Bipotentiostat (AFC-BP1). PEC measurements were performed
in a 3-electrode configuration using a saturated calomel electrode (SCE)
(+0.24 V vs RHE) and a Pt wire as the reference and counter electrodes, respectively.
Solutions were purged with nitrogen (Praxair, purity 99.999%) for 1 hour in a glass PEC
cell and a positive pressure N2 blanket maintained in the overhead volume of the cell. For
H2 evolution measurements, the cell was sealed under N2 to avoid the loss of any H2 in
the head space above the working electrode.
60
3.2.5. Gas Chromatography
The headspace in the electrochemical cell was sampled using a 5 mL syringe,
fitted with a gastight valve (Series A-2, VICI Precision Sampling), and analyzed using an
Agilent Technologies 6890N GC system equipped with a thermal conductivity detector. A
2.13 m Agilent J&W GC packed column in stainless steel tubing was used (inner diameter
2 mm, HayeSep N packing material, 60/80 mesh size). Argon (99.999%) was used as a
carrier gas at a flow rate of 30 mL/min under 46.2 psi.
3.2.6. Electron Microscopy
Scanning electron microscopy (SEM) was carried out using a Nova NanoSEM
electron microscope at an accelerating voltage of 5 kV and a working distance of 5 mm.
Films were prepared on ITO as described above and fixed to an aluminum stub using
carbon tape (SPI supplies). A secondary electron (SE) detector was used to obtain images
of the fine nanostructured features of the nanoparticle electrodes and a back scattered
electron (BSE) detector was used to visualize the Pt nanoparticle catalyst on the
nanostructured electrodes. BSE images are contrasted by atomic number, where the
bright phases correspond to the species with a higher atomic number. This is especially
useful in this work as it allows to distinguish between nanoparticle catalysts from
nanoparticle support. Bright field transmission electron microscopy (TEM) images were
acquired with a FEI Tecnai Osiris scanning transmission electron microscope (FEI
Company, Hillsboro, OR, USA) operated at 200 kV. TEM samples were prepared by
dropping an IPA solution of the nanoparticle dispersion on a TEM grid having an ultrathin
carbon film on a lacey carbon support (Prod #01824, Ted Pella Inc).
3.2.7. Spectroscopy
UV Absorption spectra were recorded on a Cary 300 Bio UV-visible
Spectrophotometer (Agilent Technologies). All solution samples were diluted to an
appropriate concentration and measurements were performed in quartz cuvettes. All thin
film samples were ~90 nm thick and prepared on ITO-coated glass substrates. Absorption
intensities of each spectrum were normalized to their respective λmax values. Steady-state
fluorescence emission spectra were performed using a PTI Quantamaster
spectrofluorometer with a xenon short arc lamp (Ushio Inc.) as the excitation source. All
61
thin film samples were mounted 22.5° normal to the incident light. All spectra were
recorded using an excitation wavelength of 515 nm, coinciding with their λmax values as
determined by UV-Vis.
3.2.8. Grazing Incidence Wide Angle X-ray Scattering
GIWAXS measurements were performed on a SAXSLAB Ganesha 300XL
instrument using a photon wavelength of 1.54 Å (Cu K-α). Measurements were taken
under vacuum at an angle of incidence of 0.16° and the scattered intensity was detected
on a Pilatus3 R 300K detector. Data analysis was performed using the saxsgui version
2.15.02 software package.
3.2.9. Dynamic Light Scattering
DLS measurements were performed at room temperature using a Zetasizer Nano
ZS (Malvern Instrument). 1 drop of the nanoparticle dispersion (2.7 mg/mL) was diluted in
2 mL of the appropriate solvent and measurements were performed in plastic cuvettes
with a 1 cm path length.
3.3. Results and Discussion
3.3.1. Nanoparticle Synthesis
P3HT:PCBM nanoparticles were synthesized via surfactant-free precipitation
method, as described above. Various combinations of solvent (in which polymer:fullerene
solutions were prepared) and non-solvent (in which the polymer:fullerene solutions were
added) were investigated in order to study the synergistic effect on both the size of
nanoparticles and nature of aggregation of the polymer nanoparticles. Solvents/non-
solvents were chosen by considering the solubility data (reported by Machui et al.)181 of
both P3HT and PCBM (Table 3-1).
While poor solubility in the non-solvent ensures that the blend of P3HT and PCBM
will co-precipitate upon the addition of the P3HT:PCBM solution to the non-solvent, the
physical properties of both the solvent and non-solvent are important to the formation of
nanoparticle films. Employing a solvent which possesses a lower boiling point and higher
62
vapour pressure than the non-solvent ensures that the good solvent of the solution
evaporates prior to evaporation of the non-solvent during the spin coating process. If the
non-solvent evaporates prior to the evaporation of the good solvent, the nanoparticles are
at risk of being re-dissolved by residual solvent during spin casting, which may result in
P3HT:PCBM films more akin to those cast from homogenous solutions.
Table 3-1 Solubility of P3HT & PCBM in various solvents181 and solvent properties
P3HT solubility (mg/mL)
PCBM solubility (mg/mL)
Boiling Point (°C)
Vapor Pressure (mmHg)
Solvents
o-dichlorobenzene 15 42 180 1
chloroform 14 29 61 158
tetrahydrofuran
1 2 65 145
Non-solvents
1-butanol < 0.1 < 0.1 118 5
2-propanol < 0.1 < 0.1 83 33
methanol < 0.1 < 0.1 65 92
ethanol < 0.1 < 0.1 78 43
dimethylsulfoxide < 0.1 0.2 189 1
water < 0.1 < 0.1 100 17
Based on the above criteria, of those listed only CHCl3 and THF were considered
for further use, as o-DCB is not removed prior to the non-solvent during the spin coating
process due to its higher boiling point and lower vapor pressure. The physical properties
of the non-solvent are equally as important to consider, ensuring good film formation
during the spin coating process. Although many non-solvents meet the required criteria of
low P3HT and PCBM solubility, high boiling point and low vapor pressure, those with very
high boiling points and very low volatility are not completely removed during the spin
casting process, leading to excessively long drying times. For this reason, DMSO, BuOH
and water were rejected as non-solvents and not considered further.
In the surfactant-free precipitation method employed, the stability of the
nanoparticles in the non-solvent is also important as there is no surfactant present to
mitigate nanoparticle aggregation. It was found that non-solvents possessing more
63
amphiphilic character were able to act as both non-solvent and surfactant, leading to
stable nanoparticle dispersions. This was apparent in the series of alcohols where
nanoparticles prepared in MeOH aggregated within a few hours, whereas nanoparticles
prepared in BuOH were stable over several days. As the alkane chain length of the alcohol
is increased, the stability increases due to their increasing amphiphilic nature; however,
boiling point and vapor pressure rapidly increase and decrease, respectively, with higher
carbon homologues, rendering alcohols with aliphatic chains longer than propanol
impractical for the formation of nanoparticle films.
3.3.2. Characterization of Nanoparticles
Dynamic light scattering (DLS) and transmission electron microscopy (TEM) were
used to determine the size of nanoparticles formed. The DLS data shown in Figure 3.2a
implies that a distribution of nanoparticle sizes exists in solution with a z-average diameter
of 140 nm. TEM images in Figure 3.2b indicate smaller nanoparticle sizes than those
obtained by DLS, but a distribution of sizes is observed nonetheless, with larger
nanoparticles having diameters of ~100 nm and smaller nanoparticles having diameters
of ~50 nm. DLS measures the hydrodynamic radius, but the measurements rely upon
Rayleigh scattering, where intensity of Rayleigh scattering scales with the 6th power of
particle diameter. Therefore the presence of larger aggregates skews the average size of
the nanoparticles to larger than actual values.178
As shown in Figure 3.3, UV-Vis absorption spectra of the nanoparticle dispersions
exhibit an absorption maximum situated at 515 nm, with a vibronic shoulder at 600 nm.
The red shift in the nanoparticle absorption spectra, with respect to the absorption of P3HT
dissolved in solution, as well as the existence of a vibronic shoulder are indicative of inter-
chain interactions due to π-π stacking. These features suggests that semi-crystalline
domains of P3HT are formed during formation of nanoparticles.26 The absorption spectra
of the nanoparticle dispersions possess similar absorption spectra to nanoparticle films
prepared therefrom, in contrast to films cast from P3HT polymer solutions where
significant interchain interactions are absent in solution but evolve during film casting.
Comparing the absorption spectra of nanoparticle and polymer solution-cast films, similar
absorption maxima are observed, however, the solution-cast film exhibits a weaker
vibronic shoulder intensity. Considering the H-aggregate model,150,151 it may be concluded
64
Figure 3.2 Size distribution of P3HT:PCBM nanoparticle diameter determined by (a) dynamic light scattering (DLS) and (b) transmission electron spectroscopy (TEM). For DLS measurements, 1 drop of nanoparticle dispersion was diluted in 2 mL of IPA. TEM samples were prepared by drop casting a diluted nanoparticle dispersion in IPA on a lacey carbon TEM grid.
that the ratio of aggregated to non-aggregated polymer domains is larger for the
nanoparticle film than solution-cast films. However, limitations on over-interpreting the
relevance of the H-aggregate model from UV-Vis spectroscopy have recently been
raised.183 These limitations arise due to the sensitivity of UV-Vis spectra to reflections at
polymer boundary layers and the influence of film thickness and refractive index on
reflectance. The latter is especially relevant when comparing nanoparticle-cast and
solution-cast films, as the surface morphologies are expected to differ significantly.
Further information on the comparative aggregation within the nanoparticle and
solution-cast films was obtained by grazing incidence wide-angle x-ray scattering
(GIWAXS). GIWAXS plots are shown in Figure 3.4. The scattering peak situated at
q= 0.4 Å-1 (d-spacing of 15.7 Å) arises from the lamellar stacking of P3HT crystallites,
considered as the (100) plane in P3HT films. The relative intensity of the scattering peak
at q= 0.4 Å-1 is proportional to the degree of ordering within the P3HT nanodomains of the
P3HT:PCBM films.184 The similar relative scattering intensity of the nanoparticle and thin
film samples suggests a similar extent of ordering between the two. The 2D plot shown
for the nanoparticle electrodes, Figure 3.4b, shows polymer chains with nearly isotropic
orientation as characterized by the broad ditribution of intensity for all peaks, whereas the
2D plot of the
65
Figure 3.3 Normalized UV-vis spectra of a P3HT:PCBM nanoparticle film (solid black), a solution cast P3HT:PCBM bulk heterjunction film (solid red), a dispersion of P3HT:PCBM nanoparticles in IPA (dashed black) and a P3HT:PCBM solution in o-DCB (dashed red). nanoparticle and solution cast films were ~90 nm thick.
solution-cast electrode, Figure 3.4c, shows only the (100) peak along qy with a faint trace
of the (200) peak, also along qy. These results suggest that the polymer chains within the
nanoparticles do not align themselves in any particular orientation with respect to the
substrate, while the polymer chains of the solution-cast film are preferentially aligned in
an edge-on orientation.
Due to high excitonic binding energies of conjugated polymers, the exciton
diffusion lengths are on the order of 10-20 nm.185 To compensate for the short exciton
diffusion lengths, electron acceptors, such as PCBM, have been introduced to promote
efficient charge separation. If the exciton recombines prior to reaching the donor-acceptor
interface, radiative emission may occur in the form of fluorescence. Ultafast photoinduced
electron transfer from a conjugated polymer to fullerene in bulk heterojunction films were
first reported by Sariciftci et al.94 Photinduced electron transfer manifests itself in
quenching of the photoluminescence intensity of the conjugated polymer film. Improving
the bulk heterojunction, by increasing the interfacial surface area between the donor and
acceptor phases, leads to enhanced quenching of the photoluminescence. Figure 3.5
shows the steady-state fluorescence spectra for both the nanoparticle and solution cast
P3HT:PCBM films, as well as a pristine P3HT film for reference. A clear increase in
quenching of the fluorescence is observed for both the nanoparticle and
66
Figure 3.4 (a) GIWAXS data obtained for P3HT:PCBM nanoparticle (black) and solution-cast (red) films. The angle of incidence was 0.16°. Inset: Lamellar stacking of P3HT chains in an edge-on fashion along the (100) direction. 2D GIWAXS plots of the P3HT:PCBM nanoparticle and solution-cast electrodes are shown in (b) and (c), respectively.
solution-cast films due to the presence of PCBM. A possible advantage which
semiconducting nanoparticles offer is greater control of the nanoscale morphology by
controlling the extent of phase segregation within each nanoparticle rather than over the
entire active layer.175 The enhanced quenching observed for the P3HT:PCBM
nanoparticle films, with respect to the solution-cast film may be due to enhanced control
of the donor-acceptor phase segregation on the nanoscale.
SEM images of the P3HT:PCBM nanoparticles cast on ITO electrodes are shown
in Figure 3.6, from which the diameter of the nanoparticles is estimated to be between 50
and 100 nm, coinciding with the TEM results described above. Under lower magnification,
Figure 3.6a, a homogeneous distribution of P3HT:PCBM nanoparticles is observed across
the electrode, free from the formation of large aggregates. Under higher magnifications,
Figure 3.6b, layers of partially coalesced spherical nanoparticles are observed, confirming
the formation of nanostructured electrodes. Partial coalescence may be caused by trace
CHCl3 remaining in the nanoparticle dispersion. Although the CHCl3 solvent evaporates
67
Figure 3.5 Steady-state fluorescence emission spectra of a solution-cast P3HT film (dashed), a solution cast P3HT:PCBM film (red) and a P3HT:PCBM nanoparticle film (black), all on ITO. An excitation wavelength of 515 nm was used.
prior to the alcohol during the spin casting process, its presence during each subsequent
coating may be sufficient to promote plasticization and partial coalescence of the
nanoparticles. Partial coalescence of the nanoparticles may, in fact, be beneficial to
photocathode activity, as previous reports have shown that coalescence of P3HT:ICBA
(indene-C60 bisadduct) nanoparticles, by thermal annealing, increases the performance of
organic photovoltaic devices due to the increase in the short circuit current density brought
about by the connectivity of nanoparticles.182
Figure 3.6 SEM images of a P3HT:PCBM nanoparticles on ITO electrodes. Images (a) and (b) depict the nanoparticle films at magnification values of 60k and 250k, respectively.
68
3.3.3. Photoelectrochemistry of Nanoparticle-based Photocathodes
Photocathodes were prepared by depositing both P3HT:PCBM nanoparticles and
solution-cast P3HT:PCBM films onto ITO, followed by the photoelectrochemical
deposition of platinum nanoparticles by potentiostatic coulometry (Figure 3.7). By taking
advantage of the photoreductive activity of P3HT and the position of the redox potential
associated with the PtCl62- anion (0.51 V vs. SCE or 0.75 V vs. RHE), which is positive of
both P3HT and PCBM LUMO energy levels, it is possible to reduce PtCl62- to metallic Pt
on the electrode surface by irradiation of the P3HT:PCBM photocathode under a bias
of -0.1 V vs. SCE (0.14 V vs. RHE). The photocurrents observed are attributed to the
reduction of PtCl62- to Pt0.
Figure 3.7 Platinum deposition by electrolysis under chopped illumination for P3HT:PCBM nanoparticle (black) and solution-cast (red) films. The potential at the photoelectrodes was held at -0.24 V vs. SCE (0 V vs. RHE).
SEM images of the electrodes obtained using a back scattered electron detector
illustrate that Pt nanoparticle catalysts (bright spots) are deposited in uniform size and
distribution on the P3HT:PCBM nanoparticle and solution-cast films (Figure 3.8).
Linear sweep voltammetry of the photoelectrodes in aqueous solution, Figure 3.9a,
confirm the role of Pt catalyst on the nature of the electrochemical reactions under
irradiation as there is no appreciable photocurrent generated in their absence. After
deposition of Pt electrocatalysts, however, a cathodic current is observed having an onset
potential at 0 V vs. SCE (0.24 V vs. RHE), which is slightly more negative than the
reported flat-band potential of P3HT 0.1 V vs. SCE (0.34 V vs. RHE) in similar
solutions.186 It is not uncommon for the onset potential to appear up to a few hundred
millivolts more negative than the flat-band potential due to recombination in the space
69
Figure 3.8 SEM images, collected using the backscattered electron detector, of Pt deposited after a deposition period of 100 s on (a) a P3HT:PCBM nanoparticle electrode and (b) a solution-cast P3HT:PCBM thin film electrode.
charge layer near flat band potential, charge trapping at surface defects and/or poor
charge transfer kinetics which leads to charge accumulation at the surface.187 Negligible
dark currents are observed for both the P3HT:PCBM nanoparticle and P3HT:PCBM
solution-cast photocathodes, however, the nanoparticle photoelectrodes exhibit a small
dark current at potentials more negative than -0.3 V vs. SCE (-0.06 V vs. RHE) which may
be due to electrocatalytic evolution of H2 at Pt which has been inadvertently deposited
onto exposed ITO sites during the Pt deposition process. Voltammograms of Pt deposited
on blank ITO were recorded (Figure 3.9c) and show the onset of H2 at the ITO/Pt
electrodes correlates well with the onset of dark current at the ITO/Pt electrodes.
To minimize the influence of H2 directly evolved at ITO/Pt sites during the photo-
assisted hydrogen evolution experiments, all subsequent photo-electrolysis experiments
were performed at sufficiently positive potentials (-240 mV vs. SCE or 0 V vs. RHE) where
the dark current was negligible. Potentiostatic coulometry data of the P3HT:PCBM
nanoparticle and of the solution-cast films, both with and without Pt catalyst, are presented
in Figure 3.9b. The experiments were performed under chopped illumination, where
significant currents are only observed during the illumination of the photoelectrodes and
return to negligible current in the dark, confirming that photogenerated charges within the
polymer films are responsible for the redox reactions which occur at the interfaces of the
working and counter electrodes within the photoelectrochemical cell.
70
Photocurrents obtained during electrolysis are increased from sub-µA values at
uncatalyzed photoelectrodes to 35 µA/cm2 and 7 µA/cm2 for the Pt-deposited,
P3HT:PCBM nanoparticle and the Pt-deposited solution-cast photoelectrodes,
respectively. To confirm hydrogen evolution at the nanoparticle photocathodes,
electrolysis was performed for 1 hour and the headspace was positively identified by gas
chromatography as containing hydrogen. The rate of hydrogen generation was calculated
to be 0.1 µmol/h.
Figure 3.9 Linear sweep voltammetry (a) and electrolysis (b) under chopped illumination comparing P3HT:PCBM nanoparticle (black) and solution-cast (red) photocathodes in 0.1 M H2SO4. Experiments were performed both with (solid) and without (dashed) Pt nanoparticle catalysts. Linear sweep voltammetry of Pt deposited on bare ITO is shown in (c). Linear sweep voltammograms were measured at a scan rate of 5 mV/s and electrolysis measurements were performed at -0.24 V vs. SCE (0 V vs. RHE).
Although larger photocurrents are observed at the nanoparticle photoelectrodes
compared to their solution-cast film counterparts, it is noted that the increased surface
71
area of the nanoparticle electrodes facilitates a larger quantity of Pt electrocatalyst to be
deposited when subjected to identical Pt deposition times under identical conditions. The
mass of Pt deposited was calculated using the total charge passed during
photoelectrochemical deposition, the details of which are provided in the experimental
details, which is a reasonable assumption as the solution had been thoroughly purged
with N2 to remove any dissolved oxygen. For the voltammetric and coulometric curves
provided, the loading of Pt catalyst deposited on the nanoparticle and solution-cast
electrodes films were 1.11 and 0.21 µg/cm2 respectively. To obtain comparable catalyst
loadings to the nanoparticle electrodes, longer Pt deposition periods were used for the
planar films. The longer deposition periods, however, lead to decreased photocathode
performance. The electrolysis results of solution-cast electrodes for which Pt has been
deposited for periods of 100 and 250 seconds are shown in Figure 3.10. These results
show that it is detrimental to extend Pt deposition for long periods of time. As deposition
time is extended, the size of the catalyst particles increases,188 resulting in a decreased
catalytic surface area and thus decreased catalytic activity.139
Figure 3.10 Electrolysis under chopped illumination comparing P3HT:PCBM solution-cast photocathodes prepared using Pt deposition periods of 100 s (solid) and 250 s (dashed). Measurements were performed at - 0.24 V vs. SCE (0 V vs. RHE).
3.3.4. Conclusion
P3HT:PCBM nanoparticles were prepared via simple precipitation method and
used to fabricate nanostructured organic photocathodes. The nanoscale morphology,
polymer crystallinity and photoelectrochemical performance of the nanostructured
72
P3HT:PCBM electrodes were compared with planar P3HT:PCBM electrodes to better
understand the effects of nanostructuring on organic photocathodes.
It was postulated that control of the nanoscale morphology within individual
nanoparticles may lead to increased donor-acceptor interfacial area, and thus enhanced
charge separation compared to the bulk phase separation of donor-acceptor domains that
occurs in planar electrodes. Steady-state fluorescence spectra were collected for both the
nanostructured and planar P3HT:PCBM as a qualitative measure of photogenerated
charge separation. Although the nanostructured electrodes exhibit a slightly lower
emission signal compared to the planar electrodes, suggesting greater charge separation,
the presence of PCBM is of much greater importance than whether the particular donor-
acceptor morphology is controlled over a 90 nm thin film or within nanoparticles with
diameters of 50-100 nm.
The effect of nanoparticle formation on polymer aggregation and crystallinity was
studied through UV-Vis and GIWAXS measurements. The similarities between the
absorption spectra of P3HT:PCBM nanoparticles suspended in solution and P3HT:PCBM
nanoparticles cast as films suggests that polymer aggregation and crystallinity is
determined during nanoparticle formation. The red-shift, relative to the P3HT:PCBM
solution, and appearance of a vibronic shoulder at longer wavelengths are indicative that
semi-crystalline polymer domains are present within the nanoparticles immediately after
their synthesis and remain largely unchanged through the spin coating process. GIWAXS
measurements were performed for quantitative comparison of the polymer crystallinities
between the nanostructured and planar P3HT:PCBM electrodes. The GIWAXS
measurements revealed similar polymer crystallite sizes for the nanostructured and planar
electrodes, however, the orientation of the polymer chains with respect to the ITO
substrate differed. The polymer chains within the nanostructured electrodes do not adopt
a preferred orientation, whereas the polymer chains in the planar electrode are
preferentially aligned in an edge-on orientation.
Finally, a platinum hydrogen evolution catalyst was deposited onto both the
nanostructured and planar P3HT:PCBM photocathodes and their photoelectrochemical
performances compared in acidic aqueous solution. The nanostructured electrodes
exhibited a five-fold increase in photocurrent density during photoelectrolysis
measurements over the planar electrodes. While the fluorescence, UV-Vis and GIWAXS
73
measurements suggest that morphological differences between the nanostructured and
planar electrodes exist, the most significant factor in photoelectrochemical performance
appears to be surface area. The larger surface area of the nanoparticle electrodes allows
for greater amounts of platinum to be deposited during the catalyst deposition process,
consequently leading to enhanced photoelectrochemical performance. Pt-loading studies
at planar electrodes show that increasing the Pt deposition period past a certain point will
begin to have adverse effects on photoelectrochemical performance. The growth of
catalyst particles with increasing deposition time, results in decreased catalytic surface
area.
For photoelectrochemical reactions plagued by slow electron transfer kinetics at
the polymer-electrolyte interface, such as the HER, the performance of the
photoelectrodes will ultimately be determined by the amount of catalytic surface area
available. There are, however, photoelectrochemical reactions which have been shown to
occur directly at the polymer-electrolyte interface in the absence of an electrocatalyst,
such as the reduction of O2189 and organic redox couples, such as anthraquinone-2,7-
disulfonate (AQDS).190 The reduction of O2 and AQDS are of interest in the application of
dissolved oxygen sensors and redox flow batteries, respectively. Further study of the
photoelectrochemical reduction of various redox couples at nanostructured conjugated
polymer electrodes is warranted.
74
Chapter 4. Energy Level Alignment and Interfacial Dipole Layer Formation at the P3HT:PCBM-Electrolyte Interface in Organic Photoelectrochemical Cells
4.1. Introduction
It is well documented that in the field of organic photovoltaics, the formation of a
bulk heterojunction between an electron donating material (e.g. conjugated organic
polymer) and an electron accepting material (e.g., fullerene) is necessary to achieve high
power conversion efficiencies.191 Although the influence of PCBM at the organic-metal
interface in solid state organic electronics has been the subject of many studies, there
have been no studies concerning the influence of PCBM at the organic-electrolyte
interface in organic photoelectrochemical cells.
Previous reports, in the area of organic electronics, have shown that the interfacial
dipole layer formed at organic-metal interfaces alters the energy level alignment between
the Fermi level of the organic layer and the metal work function,192–195 and studies specific
to P3HT:PCBM bulk heterojunction systems have shown that the presence of PCBM at
the interface between the organic layer and the metal cathode leads to band unpinning
and a subsequent decrease in the magnitude of band bending in the organic layer at the
metal cathode contact.196,197 At the organic-metal interface, the wave function of the metal
extends into the organic layer, where overlap with the organic molecular states causes a
shift and broadening of the molecular levels, creating an induced density of interface
states within the band gap. Within the density of states there exists a charge neutrality
level - a level at which filling all states up to that level with electrons results in a surface
without a net charge. As a result, when the Fermi level of the organic layer lies below the
charge neutrality level, the surface will be positively charged. The spontaneous electron
transfer process which occurs between the organic and metal layers is determined by the
relative positions of the charge neutrality level and the metal work function. When the work
function of the metal lies above the charge neutrality level of the organic layer, electrons
are transferred from the metal to the interface states, creating an interfacial dipole layer
and alignment of the metal work function with the charge neutrality level of the organic
75
layer (i.e., Fermi level pinning). Guerrero et al. found that the band-bending and strength
of the dipole layer at the P3HT:PCBM-metal cathode interface are highly dependent on
the number and nature of the fullerene molecules in contact with the metal, indicating that
PCBM molecules are largely responsible for controlling the density of the interface
states.197 Simply put, when a large number of PCBM molecules are present at the
interface, a large number of induced interfacial states exist. This is consistent with reports
that the extent to which the metal work function aligns with the charge neutrality level is
dependent on the density of interface states in the vicinity of the charge neutrality
level.195,198,199 As the analogy of the semiconductor-metal Schottky barrier is routinely
applied to semiconductor-electrolyte interfaces,15,200–202 applying the above findings to the
P3HT:PCBM photoelectrochemical system may prove useful in understanding the role of
PCBM in organic photoelectrodes.
Any energy level alignment that occurs at the semiconductor-electrolyte interface
will affect the magnitude of band-bending, i.e., the built-in potential (Vbi), and the depletion
layer width. Equation 2 shows the dependence of the depletion layer width, W, on the
built-in potential, Vbi, the relative dielectric constant of the organic layer, εr, the permittivity
of vacuum, εo, the elementary charge, q, and the doping density, NA.
The depletion layer plays an important role in photoelectrochemistry as any
photoinduced charges generated within this region may be separated by the built-in
electric field. It is therefore expected that the influence of PCBM on the energy level
alignment (i.e., Vbi) at the organic semiconductor interface will have a direct impact on
exciton separation within the semiconductor photoelectrodes and subsequent
photoelectrochemical performance. This work aims to elucidate the role of PCBM at the
organic-electrolyte interface of organic photoelectrochemical cells. More specifically, the
influence of PCBM on interfacial energy alignment is investigated through
photoelectrochemical techniques and the relationship between the built-in potential,
depletion layer width and photocathode performance is discussed.
76
4.2. Experimental
4.2.1. Materials
P3HT was purchased from Rieke Metals (MW = 50-70 kg/mol, regioregularity: 91-
94%, < 0.01% metal impurities). PCBM (American Dye Sources, Inc.), and all solvents
were all used as-received. Indium tin oxide (ITO) coated glass slides (15 Ω/□) (Colorado
Concept Coatings LLC) were cleaned by sequential sonication for 10 minutes in each of
the following solutions: dichloromethane, deionized water and isopropyl alcohol.
Chemically-cleaned ITO slides were subjected to a plasma treatment in air for 10 minutes
(Harrick Plasma PDC-32G).
4.2.2. P3HT:PCBM Films
The appropriate ratios of P3HT and PCBM were weighed out depending on the
sample being prepared and dissolved by stirring in chloroform overnight at 60 °C. Thin
film electrodes were prepared by spincoating 0.4 µL aliquots of the corresponding solution
(filtered through 0.45 µm PTFE filter) onto plasma-cleaned ITO (Harrick Plasma PDC-
32G, 10 min with air plasma) and spun at 2000 rpm. The resulting P3HT and P3HT:PCBM
films were ~220 nm thick, as determined by profilometry (Alpha-Step IQ profilometer from
KLA-Tencor). For samples prepared with a PEDOT:PSS charge selective layer, a
PEDOT:PSS solution was filtered through a 0.45 µm Acrodisk filter and spin-cast onto
plasma cleaned ITO (Harrick Plasma PDC-32G, 10 min with air plasma) at 2000 rpm. The
resulting PEDOT:PSS films were placed in an oven at 120 °C for 10 mins to remove
moisture. The appropriate P3HT or P3HT:PCBM solution was spin-cast on top of the
PEDOT:PSS layer. All electrodes were coated with an epoxy coating (Epoxies Inc.) to
prevent exposure of ITO to electrolyte at the electrode edges.
4.2.3. Photoelectrochemistry
A 200 W Xe/Hg lamp (Uhsio America, inc.) was used in combination with a
300-700 nm band pass filter (FSQ-KG3, Newport Corp.) and neutral density filters
(Thorlabs Inc.), to achieve 100 mW/cm2 irradiation, as measured using a broadband
power meter (841-PE, Newport Corporation) equipped with an Ophir thermal detector
head (3A-P-SHV1). A glass photoelectrochemical cell with two large flat windows and five
77
accessory ports was used and the cell configuration was designed to allow irradiation of
the polymer–electrolyte interface through the electrolyte. A water filter was placed in front
of the electrochemical cell to remove heat. Photoelectrochemical measurements were
performed using a Pine Bipotentiostat (AFC-BP1). Non-aqueous photoelectrochemical
measurements were performed in a 3-electrode configuration using a silver/silver ion
reference electrode (0.01 M AgNO3 in 0.1 M Bu4N PF6) and a Pt wire as the reference and
counter electrodes, respectively. Solutions were purged with nitrogen (Praxair, purity
99.999%) for 1 hour in a glass PEC cell and a positive pressure N2 blanket maintained in
the overhead volume of the cell.
4.3. Results and Discussion
4.3.1. Open Circuit Potential & Onset Potential for Photocurrent
Open circuit potential and linear sweep voltammetry measurements provide useful
information on the photovoltage (Vph) and onset potential (Vonset) of a semiconductor,
respectively. Through careful consideration of experimental conditions, Vph and Vonset can
be used to extract Vbi, thereby providing quantitative information about energy level
alignment and depletion layer width. The major conditions for meaningful open circuit
potential measurements is that the illumination must be sufficiently intense to completely
remove pre-existing band bending at the surface and that the recombination rates of the
photogenerated charges in the semiconductor must not be so fast to nullify the effect of
illumination.203 A common method to determine whether the light intensity is sufficiently
intense to flatten the bands is to plot photovoltage vs. light intensity, as illustrated in Figure
4.1a. At higher intensities the photovoltage becomes saturated and when the above
conditions are met, Vph approximates Vbi. It should be noted here that fast recombination
rates associated with organic semiconductors results in the incomplete removal of band
bending at the semiconductor-electrolyte interface, consequently leading to observed
photovoltage values that are smaller than the built-in potential. Although quantitative
values of Vbi cannot be extracted from the photovoltage measurements performed, the
results remain significant as a clear trend of decreasing photovoltage is observed as
PCBM content is increased.
Through linear sweep voltammetry, Vbi can be estimated from the difference
between the onset potential (Vonset) at the organic photocathode under illumination and the
78
thermodynamic reduction potential of the redox couple, when kinetic overpotentials are
minimized. Kinetic overpotentials can be minimized either by introducing catalyst layers to
the semiconductor surface or by choosing appropriate redox couples that possess facile
electrochemical redox kinetics at the semiconductor electrode of interest. In this work,
benzoquinone (BZQ), a reversible redox couple possessing a low redox overpotential at
carbon-based electrodes, is used.
Photovoltage measurements were performed on P3HT:PCBM photocathodes with
varying PCBM content and the results plotted in Figure 4.1a. A general trend of decreasing
photovoltage is observed with increasing PCBM content. At higher PCBM ratios a negative
photovoltage is measured, signaling a reversal in photocathode behavior from p-type to
n-type behavior.
Figure 4.1 a) OCP vs. light intensity plot for P3HT (circles), 10 wt% PCBM (squares), 25 wt% PCBM (diamonds) and 50 wt% PCBM (triangles) and b) linear sweep voltammetry results for pristine P3HT (black), 10 wt% PCBM (red), 25 wt% PCBM (green) and 50 wt% PCBM (blue). Device architecture for all devices is ITO/P3HT:PCBM. Measurements were carried out in 0.1 M Bu4N PF6 in ACN with 1 mM BZQ. Potentials measured vs. Ag/Ag+ reference (0.01 M AgNO3). For OCP measurements, the points plotted, and error bars represent the average OCP values measured at each light intensity and their standard deviation, respectively. For LSV measurements, the Y-axis is offset for clarity, vertical dashed lines are drawn from the onset potential of their respective measurements.
The trend of decreasing photovoltage with increasing PCBM content is explained
by considering, firstly, that the extent of energy level alignment experienced by the BZQ
79
redox potential is dictated by the density of interface states at the organic layer, and
secondly, that the density of interface states increases as PCBM content is increased.
Therefore, at high PCBM content, the BZQ redox potential will align with the charge
neutrality level, resulting in a deviation from ideal Mott-Schottky behaviour, and thereby
reducing the magnitude of band bending. As photovoltage is proportional to band bending,
Vph (and subsequently Vbi) will be reduced at high PCBM concentration. An energy level
diagram depicting the effect of energy level alignment on Vbi is shown in Figure 4.2.
Figure 4.2 Energy level diagram of a) P3HT and b) P3HT:PCBM (50 wt% PCBM) photocathodes on ITO, in contact with the BZQ redox couple.
The reversal of the photoresponse from n-type to p-type for films possessing high
PCBM content is attributed to a non-Ohmic contact between the ITO and organic film
interfaces. This non-Ohmic contact gives rise to a Schottky barrier and results in band
bending at both the ITO-organic interface, as well as the organic-electrolyte interface. In
the case where Schottky barriers are formed at both sides of a semiconductor layer, the
dominant space charge layer will determine the overall photoresponse of the
photoelectrode.204,205 For pristine P3HT films, the Schottky barrier at the organic-
electrolyte interface is approximated by the difference between the Fermi level of P3HT
and the BZQ redox potential. Similarly, the ITO-organic Schottky barrier is approximated
by the difference between the P3HT Fermi level and the ITO work function. For pristine
P3HT films, the larger Schottky barrier at the organic-electrolyte interface dominates,
sweeping holes (majority carriers) towards the back contact and electrons (minority
carriers) towards the electrolyte, resulting in a p-type photovoltage. For P3HT:PCBM films
with high PCBM content, the Schottky barrier at the organic-electrolyte interface is
80
diminished as the BZQ redox potential is pinned to the charge neutrality level, so much so
that the ITO-organic barrier dominates the photoresponse and an n-type photovoltage is
observed.
Linear sweep voltammograms of P3HT:PCBM photocathodes with varying PCBM
concentrations are shown in Figure 4.1b; vertical lines have been overlaid on the figure
for interpretation of the onset potential at each photocathode. Comparing the
voltammograms, it is apparent that increasing the PCBM content of the photocathode
results in a negative (cathodic) shift in the onset potential for P3HT:PCBM films on ITO in
the presence of BZQ redox couple under chopped illumination. Onset potentials shift from
-0.36 V for P3HT photocathodes to -0.39 V, -0.46 V and -0.51 V for P3HT:PCBM
photocathodes containing 10, 25 and 50 wt% PCBM, respectively. All onset potential
values are compared against an Ag/Ag+ non-aqueous reference electrode.
The combined results of the photovoltage and onset potential measurements show
there is a negative shift in the flat-band potential as PCBM content in the P3HT:PCBM
films is increased, which is consistent with previous reports of P3HT:PCBM films in
solid-state organic electronic devices.197 The addition of PCBM and introduction of
interface states gives rise to an interfacial dipole layer between the organic layer and the
electrolyte, resulting in the alignment of the BZQ redox couple and the charge neutrality
level. The resulting deviation from the Mott-Schottky limit at the semiconductor-electrolyte
interface produces a measurable decrease in Vbi.
4.3.2. Role of a PEDOT:PSS Charge Selective Layer
The incorporation of interfacial materials that promote an Ohmic contact and high
charge selectivity of holes and electrons at their corresponding electrode contacts have
been shown to be important for optimal device performance in organic electronics.206
PEDOT:PSS is commonly employed as a hole transport layer in organic electronics due
to its high conductivity and optical transparency,207 as well as its high work function, which
facilitates Fermi level pinning of the anode to the positive integer charge-transfer state
(EICT+) of the organic active layer.208,209 Through the incorporation of PEDOT:PSS as a
hole transport layer between the ITO and the organic layer, Ohmic contact between the
ITO and organic layer is achieved, eliminating the Schottky barrier previously created at
the ITO-organic interface. This is supported by the photovoltage measurements using
81
PEDOT:PSS as a hole transport layer, shown in Figure 4.3a, as a p-type photoresponse
is exhibited across all concentrations of PCBM, i.e., no reversal of photoresponse to n-type
is observed. Although the incorporation of a hole transport layer will eliminate any
competing space charge layers at the ITO-organic interface, the formation of the interfacial
dipole layer at the organic-electrolyte interface still occurs when PCBM is present in the
P3HT:PCBM layer. As a result, the photovoltage measured by open circuit potential is
reduced from 0.26 V to 0.16 V and the onset potential shifts cathodically from -0.27 V
to -0.34 V (vs. Ag/Ag+) as PCBM content is increased from 0 wt% through 50 wt%.
Figure 4.3 Figure 3. a) OCP vs. light intensity plot for P3HT (circles), 10 wt% PCBM (squares), 25 wt% PCBM (diamonds) and 50 wt% PCBM (triangles) in P3HT:PCBM and b) linear sweep voltammetry results for pristine P3HT (black), 10 wt% PCBM (red), 25 wt% PCBM (green) and 50 wt% PCBM (blue). Device architecture for all devices is ITO/PEDOT:PSS/P3HT:PCBM. Measurements were carried out in 0.1 M Bu4N PF6 in ACN with 1 mM BZQ. Potentials measured vs. Ag/Ag+ reference (0.01 M AgNO3). For OCP measurements, the points plotted, and error bars represent the average OCP values measured at each light intensity and their standard deviation, respectively. For LSV measurements, the Y-axis is offset for clarity, vertical dashed lines are drawn from the onset potential of their respective measurements.
With the establishment that PCBM exerts a measurable effect on the interfacial
thermodynamics of P3HT:PCBM photoelectrochemical cells, an investigation concerning
the effect of PCBM on the magnitude of photocurrent generated during photoelectrolysis
was warranted. The consequence of any shift in Vbi caused by the presence of PCBM will
82
manifest itself in a change in the width of the space charge depletion layer, as shown by
Equation 2.
The depletion layer plays an important role in photoelectrochemistry as any
photoinduced charges generated within this region may be separated by the built-in
electric field. If the depletion layer is extended over the entire width of the organic layer it
could eliminate the need for an electron acceptor and formation of a bulk heterojunction
structure. However, the combination of a small relative dielectric constant (typically 2 to 4)
and dopant concentrations on the order of 1016 reported for organic semiconductors
generally lead to depletion layer widths on the order of ~100 nm.196,197,210 As the film
thickness of organic semiconductor photoelectrodes is typically >200 nm, a significant
portion of the active layer does not experience the effects of the built-in electric field of the
depletion layer. Additionally, as the extent of energy alignment between the BZQ redox
couple and the charge neutrality level is increased with increasing PCBM content, the
decrease in Vbi will lead to a further reduction in depletion layer width.
Using Equation 2, the depletion layer widths of the organic photocathodes can be
estimated for the extreme cases where, (i) no energy level pinning occurs (i.e., for pristine
P3HT films), and (ii) where the redox potential of the BZQ couple is fully pinned at the
charge neutrality level (i.e., for P3HT:PCBM films with high PCBM content). In both cases,
we assume a relative dielectric constant, εr, and dopant density, NA, of 3.5 and 5x1016
cm-3, respectively.196,211 The built-in potential, Vbi, at a semiconductor-electrolyte interface
is defined as the difference between the semiconductor Fermi level and the
electrochemical potential of the redox couple in solution. For P3HT films, Vbi is
approximated as the difference between the P3HT HOMO (5.0 eV) and the BZQ redox
potential (4.0 eV). As P3HT is a p-type semiconductor, its Fermi level will reside close to
the HOMO level, allowing the use of the more well-defined HOMO level as a reasonable
approximation. For P3HT:PCBM films with high PCBM content, the BZQ redox potential
is pinned at the charge neutrality level, and as result, Vbi is approximated as the difference
between the P3HT HOMO and the charge neutrality level (4.7 eV).197 Note that using the
P3HT HOMO level to approximate Vbi will result in slightly larger values than if the P3HT
Fermi level were used due to the small energy offset between the HOMO and Fermi levels.
Nevertheless, this method represents a valid approximation of the upper-limit of depletion
layer width that is useful in predicting photocathode behaviour within comparative studies.
83
The depletion layer widths calculated for P3HT and P3HT:PCBM (50 wt%) films are 88 nm
and 48 nm, respectively.
In thick films, where the depletion layer extends only partially through the organic
layer, it is reasonable to assume that the bulk heterojunction structure will remain an
important factor in the overall device performance, as a large portion of charge generation
and separation will occur in the field-free region of the film. The photoelectrolysis results
of P3HT:PCBM films with different PCBM concentrations, shown in Figure 4.4, corroborate
this hypothesis, as the photocurrents measured at 220 nm thick P3HT:PCBM
photocathodes with a PEDOT:PSS hole transport layer increase with increasing PCBM
content, despite the decrease in Vbi and depletion layer width.
Figure 4.4 Photoelectrolysis using pristine P3HT (black), 10 wt% PCBM (red), 25 wt% PCBM (green) and 50 wt% PCBM (blue) in P3HT:PCBM films. A bias of -0.8 V (vs. Ag/Ag+) was applied for all electrolysis measurements. All electrodes have an active layer thickness of 220 nm and a device architecture of ITO/PEDOT:PSS/P3HT:PCBM. Measurements were carried out in 0.1 M Bu4N PF6 in ACN with 1 mM BZQ.
To probe the effect of the relative thickness the depletion layer extending into the
organic layer, the thickness of the P3HT:PCBM photocathodes was reduced to ~90 nm
so that the depletion layer extends the entire width of the organic film. As a result, the
majority of photogenerated charges should be efficiently separated, regardless of where
they are absorbed, thus photocurrent should be proportional to Vbi, which has been shown
to decrease with increasing PCBM concentration. Figure 4.5 shows electrolysis results for
90 nm thick organic layers, where the measured photocurrent densities are highest for
84
pristine P3HT films and decrease as PCBM content is increased. These results show that
for thin organic semiconductor films, where the depletion layer width extends through the
majority of the active layer, Vbi plays a dominant role in the generation of photocurrent.
These results are significant as they show that efficient charge separation is achieved in
organic photoelectrodes in the absence of an electron acceptor if the width of the depletion
layer and width of the film thickness are carefully considered.
Figure 4.5 Photoelectrolysis at pristine P3HT (black), 10 wt% PCBM (red), 25 wt% PCBM (green), and 50 wt% PCBM (blue). A bias of -0.8 V (vs. Ag/Ag+) was applied for all electrolysis measurements. All electrodes have a device architecture of ITO/PEDOT:PSS/P3HT:PCBM. Measurements were carried out in 0.1 M Bu4N PF6 in ACN with 1 mM BZQ.
4.4. Conclusion
Photoelectrochemical techniques were carried out using a reversible redox couple
with low kinetic overpotential, BZQ, where a decrease in photovoltage from 0.26 V to
0.16 V and cathodic shift in onset potential from -0.27 V to -0.34 V (vs. Ag/Ag+) is observed
as PCBM content of the P3HT:PCBM photocathodes is increased from 0 through 50 wt%.
The photovoltage and onset potential measurements reveal that the built-in potential of
the photocathodes decreases upon addition of PCBM. The observed decrease in Vbi
appears to be due to interfacial states introduced by PCBM at the semiconductor-
electrolyte interface, forming an interfacial dipole layer that aligns the BZQ redox potential
with the charge neutrality level of the organic layer.
85
As Vbi is decreased, the depletion layer width within the organic active layer, which
is important for effective charge separation, is also decreased. This has been shown to
influence photocathode performance in thinner photoelectrodes, where efficient charge
separation was achieved in the absence of a PCBM acceptor phase by tuning the
semiconductor thickness to coincide with depletion layer width. For thicker photocathodes,
where the field-free region of the semiconductor layer is significantly larger than the
depletion layer width, the bulk heterojunction donor-acceptor morphology remains crucial
for the separation of photogenerated charges.
The importance of Ohmic contact at the ITO-organic interface is evident when the
photovoltage measurements at P3HT:PCBM photoelectrodes with and without a hole
transport layer are compared. It was discovered that the photoresponse of the electrodes
changed from p-type to n-type at high PCBM content when no hole transport layer was
present. As the redox potential of the electrolyte becomes pinned to the charge neutrality
level of the organic layer at high PCBM content, the Schottky barrier at the organic-
electrolyte interface is reduced. Consequently, the Schottky barrier present at the ITO-
organic interface becomes dominant, causing minority carriers (i.e., electrons) to now be
swept towards the ITO-organic interface and majority carriers (i.e., holes) towards the
organic-electrolyte interface. The incorporation of a hole transport layer between the ITO
and organic layers promotes Ohmic contact, resulting in a p-type photoresponse across
all PCBM concentrations.
86
Chapter 5. Conclusions and Future Directions
5.1. Conclusions
Although conjugated organic polymers have been the subject of extensive
investigation over the last few decades, they represent an emerging class of materials for
photoelectrochemical applications. As most reports concerning the study of organic
photoelectrochemical cells have focused on improving performance through device
engineering, the fundamental processes that occur at the semiconductor-electrolyte
interface have remained an underreported field of research. Given the difference in
chemical environment compared to their solid-state counterparts, the elucidation of
underlying thermodynamic and electronic processes at organic photoelectrodes
warranted further investigation. The emphasis of the research conducted in this thesis
was, therefore, placed on gaining a better understanding of the thermodynamic and
electronic processes that occur at the semiconductor-electrolyte interface.
In chapter two, an initial investigation into whether P3HT photocathodes were
capable of performing the hydrogen evolution reaction in the absence of catalyst revealed
that despite appropriate thermodynamics, no hydrogen evolution occurs at the P3HT
photocathodes. It was shown quantitatively that any photocurrents measured during
photoelectrolysis was largely due to the presence of trace amounts of dissolved oxygen.
In addition to investigations concerning fundamental interfacial processes, the
advantages of nanostructured electrodes were explored in chapter three, where
P3HT:PCBM nanoparticles were synthesized and used to prepare nanostructured
photocathodes. When coupled with a photoelectrochemically deposited platinum catalyst,
the nanostructured P3HT:PCBM photoelectrodes show enhanced hydrogen evolution
compared to their planar counterparts. The improved performance is largely attributed to
the greater surface area of the nanostructured electrodes, allowing for increased catalyst
loadings.
Although fullerenes have been shown to be an important component in solid-state
organic electronics, little is known about their influence on the semiconductor-electrolyte
87
interface in photoelectrochemical applications. As such, the influence of fullerenes,
specifically PCBM, at the semiconductor-electrolyte interface was the subject of
investigation in chapter four. It was found that the presence of PCBM at the
semiconductor-electrolyte interface influences the energy level alignment of the redox
couple in solution, leading to the formation of an interfacial dipole layer. As the redox
potential of the electrolyte species aligns with the interfacial density of states brought
about by PCBM molecules at the photocathode surface, the built-in potential and depletion
layer width of the organic semiconductor are diminished. A major consequence of
decreasing depletion layer width is a subsequent decrease in photogenerated charge
separation. When the depletion layer width of the P3HT active layer is controlled to
coincide with the total polymer film thickness, efficient charge separation is possible in the
absence of a PCBM acceptor phase, and the presence of PCBM is, in fact, detrimental to
performance of photocathodes with thin active layers.
Through fundamental studies, it is often possible to gain insights that can be used
as a guideline for the selection of materials or the development of novel materials with
targeted properties for specific applications. Considering the results discussed in chapters
two and four, it should be possible to engineer fullerene-free organic photocathodes
capable of performing the hydrogen evolution reaction through the combination of a hole
transport layer, ensuring Ohmic contact between the conductive substrate and the organic
semiconductor layer; an organic semiconductor with appropriate HOMO and LUMO
energetics; and a hydrogen evolution catalyst. Additionally, further exploration of organic
nanostructures as a strategy to achieve higher photocurrent densities is warranted, based
on the results from chapter three.
5.2. Future Directions
Based on the findings of this thesis, two general avenues are recommended for
the furtherment of organic photoelectrochemical cells; the first being rational polymer
design and the second being further engineering of current polymer systems.
5.2.1. Rational Polymer Design
A significant advantage of conjugated organic polymers is their synthetic tunability,
which can be exploited to tune the optoelectronic properties of materials. This section will
88
focus on potential synthetic strategies to improve the charge extraction efficiencies of
organic photocathodes. The dielectric constant and built-in potential are of particular
interest, as these properties have been shown to influence the depletion layer width of
semiconductor photoelectrodes significantly. Through the implementation of high
dielectric constant materials and/or maximizing the built-in potential through band-edge
engineering, it may be possible to enhance charge carrier extraction within organic
photocathodes.
Conjugated Organic Polymers with High Dielectric Constants
The implementation of materials with high dielectric constants has the potential to
produce single junction polymer photoelectrodes, capable of overcoming the large binding
energies associated with traditional conjugated organic polymers while avoiding the
drawbacks of the incorporation of fullerene acceptors discussed in Chapter 4, namely
interfacial dipole formation and energy level alignment leading to decreased built-in
potential and depletion layer width. In general, increasing the relative dielectric constant
of conjugated organic polymers can be achieved through the introduction of polar or
polarizable substituents to the polymer structure.212 One such strategy that has been used
to increase the relative dielectric constant of conjugated organic polymers is the
integration of polar oligo(ethylene glycol) side chains into the polymer structure.213
Initial studies of oligo(ethylene glycol) side chain substitution were performed using
2,5-substituted poly(p-phenylene vinylene) (PPV).214 A relative dielectric constant as high
as 5.5 was achieved when tri(ethylene glycol) (TEG) sidechains were introduced at both
the 2 and 5 positions along the polymer backbone of PPV, yielding poly(2,5-bis-
(triethoxymethoxy)-1,4-phenylene vinylene) (diPEO-PPV). Poly[2-methoxy-5-(3′,7′-
dimethyloctyloxy)-1,4-phenylenevinylene] (MDMO-PPV) was used as a reference
material, where a relative dielectric constant of 3 was obtained. The hole mobility values
of diPEO-PPV and MDMO-PPV were found to be similar, in the range of 10-4 cm2/V s. The
comparable hole mobilities of diPEO-PPV and MDMO-PPV is not unexpected as the
conjugated backbone of the polymer, which is responsible for charge transport through
the film, remains largely unaffected by sidechain substitution. Despite the increased
relative dielectric constant, the performance of photovoltaic devices prepared from diPEO-
PPV blended with a PCBM acceptor were very poor. The poor photovoltaic performance
89
was attributed to poor active layer morphology, specifically the large amount of polymer
aggregation and unfavorable phase separation that occurs at high ethylene glycol content.
Chang et al. were able to show an increase in power conversion efficiency from
6.2% to 7% in OPV devices by introducing TEG functionalized diketopyrrolopyrrole
(TEGDPP) into a triple component copolymer made up of thienylbenzodithiophene
(BDTT), diketopyrrolopyrrole (DPP) and TEGDPP units.215 The TEGDPP/DPP ratio was
gradually increased and it was found that 10% TEGDPP content exhibited optimal
photovoltaic performance when blended with a PC71BM acceptor. The enhancement in
photovoltaic performance is attributed to the ability of the TEG sidechains to promote
structural ordering (i.e., enhanced π-π stacking) within the polymer domains, while precise
control over the amount of TEG incorporated into the polymer allows for control over phase
separation and hence donor-acceptor interfacial area can be maximized. Corroborating
previous reports, it was also found that large ratios of TEG sidechains led to highly
aggregated polymer chains and large phase separation that was detrimental to
photovoltaic performance.
Figure 5.1 Chemical structures of MDMO-PPV, PBD and their oligo(ethylene glycol)-bearing analogues.
Although the aforementioned studies illustrate that uncontrolled aggregation and
phase separation due to high ethylene glycol content along the polymer backbone is
detrimental to photovoltaic performance when using polymer:fullerene blends, these
issues should not be present if fullerene-free photoelectrodes can be successfully
employed in organic photoelectrochemical cells. The combined benefits of greater
90
structural ordering and increased dielectric constant by way of sidechain modification may
be an effective strategy to enhance charge separation within organic photocathodes;
especially when considering the possibility that the importance of fullerenes may be
reduced when the built-in electric field at the organic-electrolyte interface is sufficiently
strong (see Chapter 4).
Additional synthetic strategies that incorporate the use of polar/polarizable
substituents include backbone fluorination216 and the use of heavier group 16 atoms in
sulfur containing polymers (e.g., selenium or tellurium).217,218
Band-Edge Engineering
Predictive models have previously been employed to illustrate the relationship
between the power conversion efficiency of organic photovoltaics (OPVs), the bandgap of
the conjugated organic polymer, and the LUMO level of the conjugated polymer.219
Maximizing the energy difference between the polymer donor HOMO and the fullerene
acceptor LUMO will maximize the open circuit potential (VOC) in OPVs, however a balance
between maximizing VOC while maintaining a low enough bandgap to ensure adequate
absorption of the solar spectrum is required. Additionally, the LUMO levels of donor
polymers employed in OPVs are constrained by the fullerene acceptor LUMO level (4.3 eV
for PCBM), as the polymer LUMO must lie ~0.3 eV above the fullerene LUMO to ensure
efficient charge separation at the donor-acceptor interface.220 By following the design rules
developed to ensure optimal energetics in OPVs, it should be possible to develop
materials with optimal energetics to carry out the hydrogen evolution reaction (HER) in
organic photoelectrochemical cells. While the design rules should remain largely
unchanged, one notable difference is that the LUMO level of the polymer donor is now
limited by the redox potential of the HER (4.5 eV) rather than the fullerene LUMO. In
organic photoelectrochemical cells, the energy difference between the polymer HOMO
and the redox potential of the electrolyte species is known as the built-in potential (Vbi) of
the system and represents the maximum achievable driving force for photoelectrolysis.221
Maximizing Vbi is therefore desirable for optimal photoectrochemical performance,
however, similar to OPVs, a balance between maximizing Vbi and ensuring low bandgap
is required.
Subsequent work on controlling molecular energy levels through rational design
has successfully been carried out to produce high performance conjugated organic
91
polymers.222 In general, two synthetic methods have been adopted to tune the energetics
of conjugated polymers: i) through the modification of the backbone structure, achieved
via the copolymerization of various molecular building blocks and ii) through changing the
backbone substituents along the polymer chain.
Hou et. al. incorporated different conjugated units into the backbone structure of
benzodithiophene (BDT)-based polymers and studied the effect of the various units on the
molecular energy levels and bandgap of the synthesized polymers.223 It was found that
the bandgaps of the polymers could be tuned to fall between 1 and 2 eV, depending on
the conjugated unit employed. The way in which the bandgap was lowered, however, was
not uniform across all conjugated units. Polymers containing a thienopyrazine (TPZ) unit,
for example, lowered the optical bandgap value to 1.05 eV, from 2.13 eV for the BDT
homopolymer, by simultaneously lowering the LUMO and raising the HOMO level.
Although the low bandgap is desirable for greater spectral absorption, raising the HOMO
level, of course, reduces VOC which is undesirable for photovoltaic performance. When
benzothiadiazole (BT) units were introduced into the BDT polymers, the optical bandgap
was lowered to 1.70 eV by only depressing the LUMO level, leaving the HOMO
unchanged. By lowering only the LUMO level, VOC is preserved while more efficient
absorption of the solar spectrum is achieved by the lower bandgap.
Alternatively, the use of electron-withdrawing groups as substituents on the
conjugated backbone have been used to effectively lower the HOMO and LUMO levels
with minimal effect on the optical and morphological properties of the polymer.224,225 A
series of BDT-based polymers containing thienothiophene (TT) units and additional
thiophenes as a pi-bridge were synthesized with an increasing number of fluorine atoms
attached to the BDT and/or TT units.225 By attaching two fluorine atoms on the BDT unit
and one fluorine atom on the TT unit (PBT-3F), it was possible to lower the HOMO and
LUMO levels to 5.20 eV and 3.30 eV, respectively, compared to 4.90 eV and 3.10 eV for
the polymer containing no fluorine substituents (PBT-0F). As a result, the VOC values
increased from 0.56 V to 0.78 V and the average power conversion efficiencies from 4.4%
to 8.3% for the fluorine-free and fluorine containing polymers, respectively.
92
Figure 5.2 Chemical structures of the polymers used to study the effect of different conjugated units223 and fluorinated backbone substituents225 on molecular energy levels.
Backbone fluorination represents an interesting strategy to maximize the built-in
potential of organic photoelectrochemical cells, as not only does the incorporation of
fluorine as a backbone substituent lower the HOMO level but an additional increase in the
relative dielectric constant of the conjugated polymer should be observed. Unfortunately,
the dielectric constants of the materials mentioned here were not measured, however the
synergistic effect of lowering the molecular energy levels and increasing the dielectric
constant may prove to be beneficial for photoelectrochemical applications.
Going forward, the simplest route to employing low bandgap polymers with deep
HOMO levels as photocathode materials is to use commercially available
polymers, such as poly[N-9′-heptadecanyl-2,7-carbazole-alt-5,5-(4′,7′-di-2-thienyl-2′,1′,3′-
benzothiadiazole)] (PCDTBT).226 PCDTBT has HOMO and LUMO levels situated at 5.5 eV
and 3.6 eV, respectively. The next progression in employing materials with optimal
energetics would be to synthesize conjugated polymers that have previously been studied
but are not available commercially. APFO-Green 1 is an example of a polyfluorene-based
material boasting HOMO and LUMO levels of 5.8 eV and 4.1 eV, respectively.227 The
fluorinated BDT-based polymers synthesized in the Hou group represent interesting
materials for use in organic photoelectrochemical cells as they not only possess adequate
93
energetics, but the fluorinated backbone should increase the relative dielectric constant of
the materials.224,225
Figure 5.3 Chemical structures of PCDTBT and APFO-Green 1
Finally, an entirely new design space can be explored to synthesize conjugated
organic polymers that possess optimal energetics for fullerene-free photoelectrochemical
applications, where the conjugated polymer LUMO levels are no longer constrained by the
fullerene LUMO. Additionally, if the built-in electric field formed within the depletion layer
is sufficiently strong to promote exciton dissociation, the LUMO offset of 0.3 eV previously
needed to ensure efficient charge dissociation at the donor-acceptor interface is no longer
required. As a result, the utilization of conjugated polymers with LUMO levels approaching
the 4.5 eV redox potential of the hydrogen evolution reaction should be possible. The
resulting bandgaps of materials that possess both low-lying HOMO and LUMO levels
could theoretically be comparable to those of inorganic materials (e.g., 1.1 for Si, 1.4 for
GaAs and 1.5 eV for CdTe).
5.2.2. Device Engineering
Polymer Orientation
Polymer orientation has been shown to influence charge carrier mobility in P3HT-
based organic electronics.91,228,229 Charge carrier mobility is an important parameter in
organic electronics as it influences charge extraction and recombination dynamics in
OPVs230 and allows for fast device operation of OFETs.231 It would, therefore, be
reasonable to expect studies dedicated to charge carrier mobility effects in the context of
94
organic photoelectrochemical cells (OPECs) will lead to the elucidation of additional
underlying processes for optimal performance of organic photoelectrochemical devices.
As mentioned previously, charge carrier mobility within P3HT films has been
shown to be much higher in the π-stacking direction compared to the lamellar stacking
direction.154 The device geometry of OPECs dictates that controlling the polymer
orientation to maximize face-on orientation of the polymer chains would be beneficial, as
charges are collected in the direction normal to the substrate. A few simple methods to
control polymer orientation have been demonstrated, namely by varying the solvent
volatility,228 and by controlling the spin-coating speed.232 The edge-on orientation is
thought to be the thermodynamically favored orientation of P3HT films, therefore
controlling the rate solvent evaporation has a direct correlation to the orientation of
polymer chains during solidification. Spin-coating P3HT from a less volatile solvent, such
as 1,2-dichlorobenzene, will result in a preferentially edge-on orientation of the polymer
chains, whereas spin-coating from a more volatile solvent, such as chloroform, will result
in the polymer chains preferentially aligning in the face-on orientation with respect to the
substrate. As solvent evaporation rate has been shown to be proportional to spin speed,233
faster spin speeds will also result in enhanced face-on orientation.
Although casting from a more volatile solvent at higher speeds promotes face-on
orientation of the polymer chains, a subsequent decrease in crystallinity within the polymer
layer is also observed. Shioya et al. have used polarized multiple-angle incidence
resolution spectrometry to show that as the face-on component increases in a P3HT thin
film, the crystallinity decreases in a linear fashion.234 There are, however, techniques
which use mechanical shear forces to achieve a face-on oriented film with high
crystallinity, such as rubbing and friction transfer. Rubbing is a simple technique which
uses a cylinder covered by a velvet or microfiber cloth being rolled along the polymer-
coated substrate at constant speed and pressure, and friction transfer utilizes a pneumatic
piston to press a polymer pellet against a heated, moving substrate to deposit an oriented
polymer film. Kajiya et al. have reported an 8-fold increase in the out-of-plane hole mobility
in face-on oriented P3HT films after rubbing.235
95
Figure 5.4 Schematic illustration of the a) rubbing and b) friction transfer processes used to create thin polymer films with controlled polymer chain orientation.236 Reprinted from Brinkmann, M.; Hartmann, L.; Biniek, L.; Tremel, K.; Kayunkid, N. Macromol. Rapid Commun. 2014, 35, 9, Copyright 2013, with permission from WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.
Nanostructured Photoelectrodes
Nanostructured organic photoelectrodes have the potential to increase OPEC
performance due to their large surface areas and narrow feature widths that coincide with
the short exciton diffusion lengths of organic polymers. The use of nanostructured organic
photocathodes to perform the hydrogen evolution reaction was introduced in Chapter 3,
and future directions in this area will be discussed below.
The strategy of employing nanostructures to control the morphology of conjugated
organic polymers has indeed been the subject of previous studies, although these have
focused on solid-state organic electronics.96 Nevertheless, the techniques developed for
the fabrication of nanostructured solid-state organic electronics may be used to produce
potentially useful nanostructures for use as photocathodes in organic
photoelectrochemical cells. Given their desirable properties of large surface area and
narrow feature widths, nanowires and nanorods represent the most promising
nanostructures for use as organic photoelectrodes. Nanoimprinting lithography and
templating represent two methods that have been developed to produce organic
nanostructures.
Nanoimprinting lithography utilizes a prefabricated, nanostructured mold to
introduce nanostructures into already-cast, planar polymer films by imprinting the mold
onto the polymer film under pressure. Although the initial fabrication of the nanostructured
96
mold is time-consuming and expensive, it can be reused to replicate the desired
nanostructure for multiple samples. Nanoarrays of polymethylmethacrylate (PMMA) with
10 nm diameters have been fabricated via nanoimprinting lithography,237 illustrating the
potential of this technology to produce conjugated polymer nanostructures tailored to
overcome their short exciton diffusion lengths. Organic photovoltaic devices were
fabricated using a double nanoimprinting method the create 2D dot patterned
P3HT:PCBM active layers with feature sizes varying from 200-25 nm wide and 55-80 nm
in height.238 It was shown that charge transport and separation is facilitated by the smaller
feature sizes, as evidenced by the increasing short circuit current densities and
photoluminescence quenching as feature size is decreased. Additionally, fullerene-free
polymer solar cells based on P3HT:PTCDI-C13 were fabricated using nanoimprinted P3HT
nanotrenches with feature widths of 50 nm and a depth of 100 nm.239 The authors were
able to enhance the short circuit densities by a factor of 2.5 using the nanoimprinted
devices compared to planar junction devices. From these select publications, it would
seem feasible to fabricate conjugated organic polymer nanostructures with an aspect ratio
between two and three, that also possess feature widths matching the exciton diffusion
lengths of the polymer.
Figure 5.5 Schematic illustration of the nanoimpronting method used to create nanostructured polymer films.
In general, the templating method consists of filling a nanoporous scaffold with
polymer, whereby the scaffold is subsequently removed through selective etching and
freestanding polymer nanostructures remain. This method has been used to produce
P3HT nanorods oriented perpendicularly on ITO with diameters of 50 nm and heights of
150 nm.240 Here, the authors place an anodized aluminum oxide (AAO) template on a
planar P3HT film, and the nanoholes of the template are filled by P3HT via capillary action
after heating the substrate above the melting point of P3HT, under vacuum. Enhanced
photoluminescence quenching over planar P3HT films was observed for the nanorods,
owing to the enhanced exciton separation as the nanofeatures match the exciton diffusion
97
length of P3HT. Similarly, power conversion efficiencies were enhanced by a factor of
seven when combined with a C60 accepting phase to form organic photovoltaic devices.
This enhancement is largely credited to the increased short circuit current densities. In
addition to capillary action, polymer nanotubes can be created through
electropolymerization, where the nanostructured scaffold is immersed in an electrolyte
containing the desired monomer and subsequent electropolymerization is carried
out.241,242
Figure 5.6 Schematic illustration of the formation of polymer nanostructures using the templating method.242 Copyright 2011, reproduced with permission from the Royal Society of Chemistry.
98
References
(1) World Energy Outlook 2018; International Energy Agency, 2018.
(2) Krol, R. van de; Grätzel, M. Photoelectrochemical Hydrogen Production; Springer:
New York, 2012.
(3) Aneke, M.; Wang, M. Appl. Energy 2016, 179, 350.
(4) Simpson, A. P.; Lutz, A. E. Int. J. Hydrogen Energy 2007, 32, 4811.
(5) Zeng, K.; Zhang, D. Prog. Energy Combust. Sci. 2010, 36, 307.
(6) Miller, E. L.; Paluselli, D.; Marsen, B.; Rocheleau, R. E. Sol. Energy Mater. Sol.
Cells 2005, 88, 131.
(7) Van De Krol, R.; Liang, Y.; Schoonman, J. J. Mater. Chem. 2008, 18, 2311.
(8) Khaselev, O.; Turner, J. A. Science 1998, 280, 425.
(9) Cheng, W. H.; Richter, M. H.; May, M. M.; Ohlmann, J.; Lackner, D.; Dimroth, F.;
Hannappel, T.; Atwater, H. A.; Lewerenz, H.-J. ACS Energy Lett. 2018, 3, 1795.
(10) Zheng, J.; Liu, X.; Xu, P.; Liu, P.; Yongzhi, Z.; Yang, J. Int. J. Hydrogen Energy
2012, 37, 1048.
(11) Aceves, S. M.; Espinosa-Loza, F.; Ledesma-Orozco, E.; Ross, T. O.; Weisberg, A.
H.; Brunner, T. C.; Kircher, O. Int. J. Hydrogen Energy 2010, 35, 1219.
(12) Suh, M. P.; Park, H. J.; Prasad, T. K.; Lim, D.-W. Chem. Rev. 2012, 112, 782.
(13) Ogden, J. M. Annu. Rev. Energy Environ. 1999, 24, 227.
(14) James, B. D.; Baum, G. N.; Perez, J.; Baum, K. N. Technoeconomic Analysis of
Photoelectrochemical (PEC) Hydrogen Production; Directed Technologies, Inc.:
Arlington, VA, 2009.
(15) Krishnan Rajeshwar. Encycl. Electrochem. 2007, 1.
99
(16) Fountaine, K. T.; Lewerenz, H. J.; Atwater, H. A. Nat. Commun. 2016, 7, 1.
(17) Walter, M. G.; Warren, E. L.; McKone, J. R.; Boettcher, S. W.; Mi, Q.; Santori, E.
A.; Lewis, N. S. Chem. Rev. 2010, 110, 6446.
(18) Abe, R. J. Photochem. Photobiol. C Photochem. Rev. 2010, 11, 179.
(19) Goto, Y.; Hisatomi, T.; Wang, Q.; Higashi, T.; Ishikiriyama, K.; Maeda, T.; Sakata,
Y.; Okunaka, S.; Tokudome, H.; Katayama, M.; Akiyama, S.; Nishiyama, H.; Inoue,
Y.; Takewaki, T.; Setoyama, T.; Minegishi, T.; Takata, T.; Yamada, T.; Domen, K.
Joule 2018, 2, 509.
(20) Scaife, D. E. Sol. Energy 1980, 25, 41.
(21) Hsieh, Y. H.; Huang, C. P. Colloids and Surfaces 1991, 53, 275.
(22) Maeda, K.; Teramura, K.; Domen, K. J. Catal. 2008, 254, 198.
(23) Minggu, L. J.; Wan Daud, W. R.; Kassim, M. B. Int. J. Hydrogen Energy 2010, 35,
5233.
(24) Bard, A. J. J. Photochem. 1979, 10, 59.
(25) Abe, R.; Sayama, K.; Domen, K.; Arakawa, H. Chem. Phys. Lett. 2001, 344, 339.
(26) Sayama, K.; Mukasa, K.; Abe, R.; Abe, Y.; Arakawa, H. Chem. Commun. 2001,
2416.
(27) Wang, Q.; Hisatomi, T.; Jia, Q.; Tokudome, H.; Zhong, M.; Wang, C.; Pan, Z.;
Takata, T.; Nakabayashi, M.; Shibata, N.; Li, Y.; Sharp, I. D.; Kudo, A.; Yamada, T.;
Domen, K. Nat. Mater. 2016, 15, 611.
(28) Kudo, A.; Miseki, Y. Chem. Soc. Rev. 2009, 38, 253.
(29) Osterloh, F. E. Chem. Mater. 2008, 20, 35.
(30) Fujishima, A.; Honda, K. Nature 1972, 238, 37.
(31) Kumagai, H.; Minegishi, T.; Sato, N.; Yamada, T.; Kubota, J.; Domen, K. J. Mater.
100
Chem. A 2015, 3, 8300.
(32) Jiang, C.; Moniz, S. J. A.; Wang, A.; Zhang, T.; Tang, J. Chem. Soc. Rev. 2017, 46,
4645.
(33) Sivula, K.; Le Formal, F.; Grätzel, M. ChemSusChem 2011, 4, 432.
(34) Peerakiatkhajohn, P.; Yun, J. H.; Chen, H.; Lyu, M.; Butburee, T.; Wang, L. Adv.
Mater. 2016, 28, 6405.
(35) Nakato, Y.; Ohnishi, T.; Tsubomura, H. Chem. Lett. 1975, 4, 883.
(36) Yamamoto, A.; Yano, S. J. Electrochem. Soc. 1975, 122, 260.
(37) S. Hu; M. R. Shaner; J. A. Beardslee; M. Lichterman; B. S. Brunschwig; N. S. Lewis.
Science 2014, 344, 1005.
(38) Lichterman, M. F.; Carim, A. I.; McDowell, M. T.; Hu, S.; Gray, H. B.; Brunschwig,
B. S.; Lewis, N. S. Energy Environ. Sci. 2014, 7, 3334.
(39) Moniz, S. J. A.; Shevlin, S. A.; Martin, D. J.; Guo, Z. X.; Tang, J. Energy Environ.
Sci. 2015, 8, 731.
(40) Toe, C. Y.; Zheng, Z.; Wu, H.; Scott, J.; Amal, R.; Ng, Y. H. Angew. Chemie - Int.
Ed. 2018, 57, 13613.
(41) Morales-Guio, C. G.; Tilley, S. D.; Vrubel, H.; Grätzel, M.; Hu, X. Nat. Commun.
2014, 5, 3059.
(42) Kaneshiro, J.; Gaillard, N.; Rocheleau, R.; Miller, E. Sol. Energy Mater. Sol. Cells
2010, 94, 12.
(43) Chae, S. Y.; Park, S. J.; Han, S. G.; Jung, H.; Kim, C.-W.; Jeong, C.; Joo, O.-S.;
Min, B. K.; Hwang, Y. J. J. Am. Chem. Soc. 2016, 138, 15673.
(44) Moriya, M.; Minegishi, T.; Kumagai, H.; Katayama, M.; Kubota, J.; Domen, K. J.
Am. Chem. Soc. 2013, 135, 3733.
(45) Kamimura, J.; Bogdanoff, P.; Lähnemann, J.; Hauswald, C.; Geelhaar, L.; Fiechter,
101
S.; Riechert, H. J. Am. Chem. Soc. 2013, 135, 10242.
(46) Wang, D.; Pierre, A.; Kibria, M. G.; Cui, K.; Han, X.; Bevan, K. H.; Guo, H.; Paradis,
S.; Hakima, A. R.; Mi, Z. Nano Lett. 2011, 11, 2353.
(47) Lee, M. H.; Takei, K.; Zhang, J.; Kapadia, R.; Zheng, M.; Chen, Y. Z.; Nah, J.;
Matthews, T. S.; Chueh, Y. L.; Ager, J. W.; Javey, A. Angew. Chemie - Int. Ed.
2012, 51, 10760.
(48) Oh, I.; Kye, J.; Hwang, S. Nano Lett. 2012, 12, 298.
(49) Gao, L.; Cui, Y.; Wang, J.; Cavalli, A.; Standing, A.; Vu, T. T. T.; Verheijen, M. A.;
Haverkort, J. E. M.; Bakkers, E. P. A. M.; Notten, P. H. L. Nano Lett. 2014, 14, 3715.
(50) Hu, S.; Xiang, C.; Haussener, S.; Berger, A. D.; Lewis, N. S. Energy Environ. Sci.
2013, 6, 2984.
(51) May, M. M.; Lewerenz, H.-J.; Lackner, D.; Dimroth, F.; Hannappel, T. Nat.
Commun. 2015, 6, 8286.
(52) Jang, J. W.; Du, C.; Ye, Y.; Lin, Y.; Yao, X.; Thorne, J.; Liu, E.; McMahon, G.; Zhu,
J.; Javey, A.; Guo, J.; Wang, D. Nat. Commun. 2015, 6, 7447.
(53) Ding, C.; Qin, W.; Wang, N.; Liu, G.; Wang, Z.; Yan, P.; Shi, J.; Li, C. Phys. Chem.
Chem. Phys. 2014, 16, 15608.
(54) McCrory, C. C. L.; Jung, S.; Ferrer, I. M.; Chatman, S. M.; Peters, J. C.; Jaramillo,
T. F. J. Am. Chem. Soc. 2015, 137, 4347.
(55) Suen, N. T.; Hung, S. F.; Quan, Q.; Zhang, N.; Xu, Y. J.; Chen, H. M. Chem. Soc.
Rev. 2017, 46, 337.
(56) Zou, X.; Zhang, Y. Chem. Soc. Rev. 2015, 44, 5148.
(57) Wahl, A.; Augustynski, J. J. Phys. Chem. B 1998, 102, 7820.
(58) Feng, X.; Zhu, K.; Frank, A. J.; Grimes, C. A.; Mallouk, T. E. Angew. Chemie - Int.
Ed. 2012, 51, 2727.
102
(59) Shankar, K.; Basham, J. I.; Allam, N. K.; Varghese, O. K.; Mor, G. K.; Feng, X.;
Paulose, M.; Seabold, J. A.; Choi, K.-S.; Grimes, C. A. J. Phys. Chem. C 2009, 113,
6327.
(60) Cho, I. S.; Chen, Z.; Forman, A. J.; Kim, D. R.; Rao, P. M.; Jaramillo, T. F.; Zheng,
X. Nano Lett. 2011, 11, 4978.
(61) Björkstén, U.; Moser, J.; Grätzel, M. Chem. Mater. 1994, 6, 858.
(62) Kavan, L.; Grätzel, M. Electrochim. Acta 1995, 40, 64.
(63) Schwoerer, M.; Wolf, H. C. Organic Molecular Solids; Wiley-VCH Verlag GmbH &
Co. KGaA: Weinheim, 2007.
(64) Shirota, Y. J. Mater. Chem. 2005, 15, 75.
(65) Zhang, W.; Smith, J.; Watkins, S. E.; Gysel, R.; McGehee, M.; Salleo, A.;
Kirkpatrick, J.; Ashraf, S.; Anthopoulos, T.; Heeney, M.; McCulloch, I. J. Am. Chem.
Soc. 2010, 132, 11437.
(66) Noriega, R.; Rivnay, J.; Vandewal, K.; Koch, F. P. V; Stingelin, N.; Smith, P.; Toney,
M. F.; Salleo, A. Nat. Mater. 2013, 12, 1038.
(67) Venkateshvaran, D.; Nikolka, M.; Sadhanala, A.; Lemaur, V.; Zelazny, M.; Kepa,
M.; Hurhangee, M.; Kronemeijer, A. J.; Pecunia, V.; Nasrallah, I.; Romanov, I.;
Broch, K.; McCulloch, I.; Emin, D.; Olivier, Y.; Cornil, J.; Beljonne, D.; Sirringhaus,
H. Nature 2014, 515, 384.
(68) Köhler, A.; Bässler, H. Electronic Processes in Organic Semiconductors; Wiley-
VCH Verlag GmbH & Co. KGaA: Weinheim, 2015.
(69) Kasha, M. Discuss. Faraday Soc. 1950, 9, 14.
(70) Letheby, H. Chem. Soc. 1862, 15, 161.
(71) Natta, G.; Mazzanti, G.; Corradini, P. Atti Accad. Naz. Lincei Cl. Sci. Fis. Mat. Nat.
Rend. 1958, 25, 3.
(72) Dall’Olio, A.; Dascola, Y.; Varacco, V.; Bocchi, V. C. R. Acad. Sci. Ser. C. 1968,
103
267, 433.
(73) Ito, T.; Shirakawa, H.; Ikeda, S. J. Polym. Sci. Part A Polym. Chem. 1996, 34, 2533.
(74) Chiang, C. K.; Fincher, C. B.; Park, Y. W.; Heeger, A. J.; Shirakawa, H.; Louis, E.
J.; Gau, S. C.; MacDiarmid, A. G. Phys. Rev. Lett 1977, 39, 1098.
(75) Wienk, M. M.; Kroon, J. M.; Verhees, W. J. H.; Knol, J.; Hummelen, J. C.; Van Hal,
P. A.; Janssen, R. A. J. Angew. Chemie - Int. Ed. 2003, 42, 3371.
(76) Zhao, G. J.; He, Y. J.; Li, Y. Adv. Mater. 2010, 22, 4355.
(77) Li, G.; Zhu, R.; Yang, Y. Nat. Photonics 2012, 6, 153.
(78) Kularatne, R. S.; Magurudeniya, H. D.; Sista, P.; Biewer, M. C.; Stefan, M. C. J.
Polym. Sci. Part A Polym. Chem. 2013, 51, 743.
(79) Marszalek, T.; Li, M.; Pisula, W. Chem. Commun 2016, 52, 10938.
(80) Müllen, K.; Pisula, W. J. Am. Chem. Soc. 2015, 137, 9503.
(81) Zhao, J.; Li, Y.; Yang, G.; Jiang, K.; Lin, H.; Ade, H.; Ma, W.; Yan, H. Nat. Energy
2016, 1, 15027.
(82) Liu, Y.; Zhao, J.; Li, Z.; Mu, C.; Ma, W.; Hu, H.; Jiang, K.; Lin, H. Nat. Commun.
2014, 5, 5293.
(83) He, Z.; Xiao, B.; Liu, F.; Wu, H.; Yang, Y.; Xiao, S.; Wang, C.; Russell, T. P.; Cao,
Y. Nat. Photonics 2015, 9, 174.
(84) Chen, J.; Cui, C.; Li, Y.; Zhou, L.; Ou, Q. Adv. Funct. Mater. 2015, 27, 1035.
(85) Elsenbaumer, R. L.; Jen, K. Y.; Oboodi, R. Synth. Met. 1986, 15, 169.
(86) Yamamoto, T.; Sanechika, K.; Yamamoto, A. J. Polym. Sci. Polym. Lett. Ed. 1980,
18, 9.
(87) Lin, J. W.-P.; Dudek, L. P. J. Polym. Sci. Polym. Chem. Ed. 1980, 18, 2869.
(88) Friedel, B.; McNeill, C. R.; Greenham, N. C. Chem. Mater. 2010, 22, 3389.
104
(89) Kim, Y.; Cook, S.; Tuladhar, S. M.; Choulis, S. A.; Nelson, J.; Durrant, J. R.; Bradley,
D. D. C.; Giles, M.; McCulloch, I.; Ha, C.-S.; Ree, M. Nat. Mater. 2006, 5, 197.
(90) Jiang, X.; Patil, R.; Harima, Y.; Ohshita, J.; Kunai, A. J. Phys. Chem. B 2005, 109,
221.
(91) Sirringhaus, H.; Brown, P. J.; Friend, R. H.; Nielsen, M. M.; Bechgaard, K.; Spiering,
A. J. H. Nature 1999, 401, 685.
(92) Ohkita, H.; Ito, S. Polymer 2011, 52, 4397.
(93) Tang, C. W. Appl. Phys. Lett. 1986, 48, 183.
(94) Sariciftci, N. S.; Smilowitz, L.; Heeger, a. J.; Wudl, F. Science 1992, 258, 1474.
(95) Yu, G.; Gao, J.; Hummelen, J. C.; Wudl, F.; Heeger, A. J. Science 1995, 270, 1789.
(96) Chen, J. T.; Hsu, C. S. Polym. Chem. 2011, 2, 2707.
(97) Shirakawa, H.; Ikeda, S.; Aizawa, M.; Yoshitake, J.; Suzuki, S. Synth. Met. 1981, 4,
43.
(98) Yanagida, S.; Kabumoto, A.; Mizumoto, K.; Pac, C.; Yoshino, K. J. Chem. Soc.
Chem. Commun. 1985, 8, 474.
(99) Zhao, Z.; Sun, Y.; Dong, F. Nanoscale 2015, 7, 15.
(100) Wang, X.; Maeda, K.; Thomas, A.; Takanabe, K.; Xin, G.; Carlsson, J. M.; Domen,
K.; Antonietti, M. Nat. Mater. 2009, 8, 76.
(101) Schwab, M. G.; Hamburger, M.; Feng, X.; Shu, J.; Spiess, H. W.; Wang, X.;
Antonietti, M.; Müllen, K. Chem. Commun. 2010, 46, 8932.
(102) Sprick, R. S.; Jiang, J.-X.; Bonillo, B.; Ren, S.; Ratvijitvech, T.; Guiglion, P.;
Zwijnenburg, M. a.; Adams, D. J.; Cooper, A. I. J. Am. Chem. Soc. 2015, 137, 3265.
(103) Woods, D. J.; Sprick, R. S.; Smith, C. L.; Cowan, A. J.; Cooper, A. I. Adv. Energy
Mater. 2017, 7, 1700479.
105
(104) El-Rashiedy, O. A.; Holdcroft, S. J. Phys. Chem. 1996, 100, 5481.
(105) Lanzarini, E.; Antognazza, M. R.; Biso, M.; Ansaldo, A.; Laudato, L.; Bruno, P.;
Metrangolo, P.; Resnati, G.; Ricci, D.; Lanzani, G. J. Phys. Chem. C 2012, 116,
10944.
(106) Gustafson, M. P.; Clark, N.; Winther-Jensen, B.; MacFarlane, D. R. Electrochim.
Acta 2014, 140, 309.
(107) Suppes, G. M.; Fortin, P. J.; Holdcroft, S. J. Electrochem. Soc. 2015, 162, H551.
(108) Bourgeteau, T.; Tondelier, D.; Geffroy, B.; Brisse, R.; Laberty-Robert, C.;
Campidelli, S.; de Bettignies, R.; Artero, V.; Palacin, S.; Jousselme, B. Energy
Environ. Sci. 2013, 6, 2706.
(109) Mihailetchi, V. D.; Blom, P. W. M.; Hummelen, J. C.; Rispens, M. T. J. Appl. Phys.
2003, 94, 6849.
(110) White, M. S.; Olson, D. C.; Shaheen, S. E.; Kopidakis, N.; Ginley, D. S. Appl. Phys.
Lett. 2006, 89, 143517.
(111) Hau, S. K.; Yip, H.-L.; Leong, K.; Jen, A. K.-Y. Org. Electron. 2009, 10, 719.
(112) Peumans, P.; Forrest, S. R. Appl. Phys. Lett. 2001, 79, 126.
(113) Germack, D. S.; Chan, C. K.; Joseph Kline, R.; Fischer, D. A.; Gundlach, D. J.;
Toney, M. F.; Richter, L. J.; Delongchamp, D. M. Macromolecules 2010, 43, 3828.
(114) Queyriaux, N.; Kaeffer, N.; Morozan, A.; Chavarot-Kerlidou, M.; Artero, V. J.
Photochem. Photobiol. C Photochem. Rev. 2015, 25, 90.
(115) Bellani, S.; Najafi, L.; Martín-García, B.; Ansaldo, A.; Del Rio Castillo, A. E.; Prato,
M.; Moreels, I.; Bonaccorso, F. J. Phys. Chem. C 2017, 121, 21887.
(116) Mezzetti, A.; Fumagalli, F.; Alfano, A.; Iadicicco, D.; Antognazza, M. R.; di Fonzo,
F. Faraday Discuss. 2017, 198, 433.
(117) Rojas, H. C.; Bellani, S.; Fumagalli, F.; Tullii, G.; Leonardi, S.; Mayer, M. T.;
Schreier, M.; Gra, M.; Lanzani, G.; Fonzo, F. Di; Antognazza, M. R. Energy Environ.
106
Sci. 2016, 9, 3710.
(118) Rojas, H. C.; Marra, G.; Meda, L.; Lanzani, G.; Mayer, M. T.; Antognazza, M. R.;
Fonzo, F. Di. J. Mater. Chem. A 2016, 4, 2178.
(119) Belarb, E.; Blas-ferrando, V. M.; Haro, M.; Maghraoui-meherzi, H. Chem. Eng. Sci.
2016, 154, 143.
(120) Bellani, S.; Najafi, L.; Capasso, A.; Del Rio Castillo, A. E.; Antognazza, M. R.;
Bonaccorso, F. J. Mater. Chem. A 2017, 5, 4384.
(121) Ghadirzadeh, A.; Fumagalli, F.; Mezzetti, A.; Bellani, S.; Meda, L.; Antognazza, M.
R.; Di Fonzo, F. ChemPhotoChem 2018, 283.
(122) Bourgeteau, T.; Tondelier, D.; Geffroy, B.; Brisse, R.; Cornut, R.; Artero, V.;
Jousselme, B. ACS Appl. Mater. Interfaces 2015, 7, 16395.
(123) Haro, M.; Solis, C.; Molina, G.; Otero, L. A.; Bisquert, J.; Gimenez, S.; Guerrero, A.
J. Phys. Chem. C 2015, 119, 6488.
(124) Huang, S.; Zhang, H.; Wu, Z.; Kong, D.; Lin, D.; Fan, Y.; Yang, X.; Zhong, Z.;
Huang, S.; Jiang, Z.; Cheng, C. ACS Appl. Mater. Interfaces 2014, 6, 12111.
(125) Paracchino, A.; Laporte, V.; Sivula, K.; Grätzel, M.; Thimsen, E. Nat. Mater. 2011,
10, 456.
(126) Gu, J.; Aguiar, J. A.; Ferrere, S.; Steirer, K. X.; Yan, Y.; Xiao, C.; Young, J. L.; Al-
Jassim, M.; Neale, N. R.; Turner, J. A. Nat. Energy 2017, 2, 16192.
(127) Rojas, H. C.; Bellani, S.; Sarduy, E. A.; Fumagalli, F.; Mayer, M. T.; Schreier, M.;
Grätzel, M.; Di Fonzo, F.; Antognazza, M. R. ACS Omega 2017, 2, 3424.
(128) Guijarro, N.; Prévot, M. S.; Yu, X.; Jeanbourquin, X. A.; Bornoz, P.; Bourée, W.;
Johnson, M.; Le Formal, F.; Sivula, K. Adv. Energy Mater. 2016, 6, 1501949.
(129) Bellani, S.; Ghadirzadeh, A.; Meda, L.; Savoini, A.; Tacca, A.; Marra, G.; Meira, R.;
Morgado, J.; Di Fonzo, F.; Antognazza, M. R. Adv. Funct. Mater. 2015, 25, 4531.
(130) Jakešová, M.; Apaydin, D. H.; Sytnyk, M.; Oppelt, K.; Heiss, W.; Sariciftci, N. S.;
107
Głowacki, E. D. Adv. Funct. Mater. 2016, 26, 5248.
(131) Park, S.-Y.; Kim, M.; Jung, J.; Heo, J.; Hong, E. M.; Choi, S. M.; Lee, J.-Y.; Cho,
S.; Hong, K.; Lim, D. C. J. Power Sources 2017, 341, 411.
(132) Bornoz, P.; Prévot, M. S.; Yu, X.; Guijarro, N.; Sivula, K. J. Am. Chem. Soc. 2015,
137, 15338.
(133) Shao, D.; Cheng, Y.; He, J.; Feng, D.; Zheng, L.; Zheng, L.; Zhang, X.; Xu, J.; Wang,
W.; Wang, W.; Lu, F.; Dong, H.; Li, L.; Liu, H.; Zheng, R.; Liu, H. ACS Catal. 2017,
7, 5308.
(134) Shao, D.; Zheng, L.; Feng, D.; He, J.; Zhang, R.; Liu, H.; Zhang, X.; Lu, Z.; Wang,
W.; Wang, W.; Lu, F.; Dong, H.; Cheng, Y.; Liiu, H.; Zheng, R. J. Mater. Chem. A
2018, 6, 4032.
(135) Chen, Y.; Chen, H.; Tian, H. Chem. Commun. 2015, 51, 11508.
(136) María Girón, R.; Marco-Martínez, J.; Bellani, S.; Insuasty, A.; Comas Rojas, H.;
Tullii, G.; Antognazza, M. R.; Filippone, S.; Martín, N. J. Mater. Chem. A 2016, 4,
14284.
(137) Varga, A.; Endrödi, B.; Hornok, V.; Visy, C.; Janáky, C. J. Phys. Chem. C 2015,
119, 28020.
(138) Lai, L.-H.; Gomulya, W.; Berghuis, M.; Protesescu, L.; Detz, R. J.; Reek, J. N. H.;
Kovalenko, M. V.; Loi, M. A. ACS Appl. Mater. Interfaces 2015, 7, 19083.
(139) Lin, Z.; Ji, L.; Zhang, X. Mater. Lett. 2009, 63, 2115.
(140) Mosconi, E.; Salvatori, P.; Saba, M. I.; Mattoni, A.; Bellani, S.; Bruni, F.; Santiago
Gonzalez, B.; Antognazza, M. R.; Brovelli, S.; Lanzani, G.; Li, H.; Brédas, J. L.; De
Angelis, F. ACS Energy Lett. 2016, 1, 454.
(141) Bellani, S.; Porro, M.; Caddeo, C.; Saba, M. I.; Miranda, P. B.; Mattoni, A.; Lanzani,
G.; Antognazza, M. R. J. Mater. Chem. B 2015, 3, 6429.
(142) Guerrero, A.; Haro, M.; Bellani, S.; Antognazza, M. R.; Meda, L.; Gimenez, S.;
108
Bisquert, J. Energy Environ. Sci. 2014, 7, 3666.
(143) Tullii, G.; Desii, A.; Bossio, C.; Bellani, S.; Colombo, M.; Martino, N.; Antognazza,
M. R.; Lanzani, G. Org. Electron. 2017, 46, 88.
(144) Chen, Z.; Dinh, H. N.; Miller, E. Photoelectrochemical Water Splitting Standards,
Experimental Methods, and Protocols; Springer: New York, 2013.
(145) Nozik, A. J. Nature 1975, 257, 383.
(146) Bak, T.; Nowotny, J.; Rekas, M.; Sorrell, C. . Int. J. Hydrogen Energy 2002, 27, 991.
(147) Gasiorowski, J.; Kollender, J. P.; Hingerl, K.; Sariciftci, N. S.; Mardare, A. I.; Hassel,
A. W. Phys. Chem. Chem. Phys. 2014, 16, 3739.
(148) Deibel, C.; MacK, D.; Gorenflot, J.; Schöll, A.; Krause, S.; Reinert, F.; Rauh, D.;
Dyakonov, V. Phys. Rev. B 2010, 81, 085202.
(149) Brown, P. J.; Thomas, D. S.; Köhler, A.; Wilson, J. S.; Kim, J. S.; Ramsdale, C. M.;
Sirringhaus, H.; Friend, R. H. Phys. Rev. B 2003, 67, 064203.
(150) Spano, F. C.; Silva, C. Annu. Rev. Phys. Chem. 2014, 65, 477.
(151) Spano, F. C. Acc. Chem. Res. 2010, 43, 429.
(152) Hexemer, A.; Müller-Buschbaum, P. IUCrJ 2015, 2, 106.
(153) Bragg, W. L. Proc. Cambridge Philos. Soc. 1913, 17, 43.
(154) O’Connor, B.; Kline, R. J.; Conrad, B. R.; Richter, L. J.; Gundlach, D.; Toney, M. F.;
DeLongchamp, D. M. Adv. Funct. Mater. 2011, 21, 3697.
(155) Li, Z.; Wang, Y.; Shen, J.; Liu, W.; Sun, X. Opt. Lasers Eng. 2014, 56, 94.
(156) Chin, D. H.; Chiericato, G.; Nanni, E. J.; Sawyer, D. T. J. Am. Chem. Soc. 1982,
104, 1296.
(157) Loewe, R. S.; Khersonsky, S. M.; McCullough, R. D. Adv. Mater. 1999, 11, 250.
(158) Chen, T.; Wu, X.; Rieke, R. J. Am. Chem. Soc. 1995, 117, 233.
109
(159) Mao, H.; Xu, B.; Holdcroft, S. Macromolecules 1993, 26, 1163.
(160) Costentin, C.; Evans, D. H.; Robert, M.; Savéant, J. M.; Singh, P. S. J. Am. Chem.
Soc. 2005, 127, 12490.
(161) Pozo-gonzalo, C.; Torriero, A. A. J.; Forsyth, M.; Macfarlane, D. R.; Howlett, P. C.
J. Phys. Chem. Lett. 2013, 4, 4.
(162) Choudhury, B.; Weedon, A.; Bolton, J. Langmuir 1998, 14, 6199.
(163) Abdou, M. S. A.; Holdcroft, S. Macromolecules 1993, 26, 2954.
(164) Abdou, M. S. a; Orfino, F. P.; Son, Y.; Holdcroft, S. J. Am. Chem. Soc. 1997, 119,
4518.
(165) Bellani, S.; Fazzi, D.; Bruno, P.; Giussani, E.; Canesi, E. V.; Lanzani, G.;
Antognazza, M. R. J. Phys. Chem. C 2014, 118, 6291.
(166) Foley, J. M.; Price, M. J.; Feldblyum, J. I.; Maldonado, S. Energy Environ. Sci. 2012,
5, 5203.
(167) Jung, J.-Y.; Park, M.-J.; Li, X.; Kim, J.-H.; Wehrspohn, R. B.; Lee, J.-H. J. Mater.
Chem. A 2015, 3, 9456.
(168) Chen, F.; Zhu, Q.; Wang, Y.; Cui, W.; Su, X.; Li, Y. ACS 2016, 8, 31025.
(169) Clafton, S. N.; Huang, D. M.; Massey, W. R.; Kee, T. W. J. Phys. Chem. B 2013,
117, 4626.
(170) Holmes, N. P.; Burke, K. B.; Sista, P.; Barr, M.; Magurudeniya, H. D.; Stefan, M. C.;
Kilcoyne, a. L. D.; Zhou, X.; Dastoor, P. C.; Belcher, W. J. Sol. Energy Mater. Sol.
Cells 2013, 117, 437.
(171) Labastide, J. A.; Baghgar, M.; Dujovne, I.; Venkatraman, B. H.; Ramsdell, D. C.;
Venkataraman, D.; Barnes, M. D. J. Phys. Chem. Lett. 2011, 2, 2089.
(172) Pedersen, E. B. L.; Pedersen, M. C.; Simonsen, S. B.; Brandt, R. G.; Böttiger, a.
P. L.; Andersen, T. R.; Jiang, W.; Xie, Z. Y.; Krebs, F. C.; Arleth, L.; Andreasen, J.
W. J. Mater. Chem. A 2015, 3, 17022.
110
(173) Nagarjuna, G.; Baghgar, M.; Labastide, J. A.; Algaier, D. D.; Barnes, M. D.;
Venkataraman, D.; Amherst, M.; Street, N. P.; States, U. ACS Nano 2012, 12,
10750.
(174) Roy, P.; Jha, A.; Dasgupta, J. Nanoscale 2016, 8, 2768.
(175) Ulum, S.; Holmes, N.; Darwis, D.; Burke, K.; David Kilcoyne, a. L.; Zhou, X.;
Belcher, W.; Dastoor, P. Sol. Energy Mater. Sol. Cells 2013, 110, 43.
(176) Darwis, D.; Holmes, N.; Elkington, D.; Kilcoyne, A. L. D.; Bryant, G.; Zhou, X.;
Dastoor, P.; Belcher, W. Sol. Energy Mater. Sol. Cells 2014, 121, 99.
(177) Gärtner, S.; Christmann, M.; Sankaran, S.; Röhm, H.; Prinz, E.; Penth, F.; Pütz, A.;
Türeli, A. E.; Penth, B.; Baumstümmler, B.; Colsmann, A. Adv. Funct. Mater. 2014,
26, 6653.
(178) Millstone, J. E.; Kavulak, D. F. J.; Woo, C. H.; Holcombe, T. W.; Westling, E. J.;
Briseno, A. L.; Toney, M. F.; Fréchet, J. M. J. Langmuir 2010, 26, 13056.
(179) Darwis, D.; Holmes, N.; Elkington, D.; David Kilcoyne, a. L.; Bryant, G.; Zhou, X.;
Dastoor, P.; Belcher, W. Sol. Energy Mater. Sol. Cells 2014, 121, 99.
(180) Schwarz, K. N.; Farley, S. B.; Smith, T. A.; Ghiggino, K. P. Nanoscale 2015, 7,
19899.
(181) Machui, F.; Langner, S.; Zhu, X.; Abbott, S.; Brabec, C. J. Sol. Energy Mater. Sol.
Cells 2012, 100, 138.
(182) Gärtner, S.; Christmann, M.; Sankaran, S.; Röhm, H.; Prinz, E.-M.; Penth, F.; Pütz,
A.; Türeli, A. E.; Penth, B.; Baumstümmler, B.; Colsmann, A. Adv. Mater. 2014, 26,
6653.
(183) Ehrenreich, P.; Birk, S. T.; Zimmermann, E.; Hu, H.; Kim, K.; Weickert, J.; Pfadle,
T.; Schmidt-mende, L. Sci. Rep. 2016, 6, 32434.
(184) Erb, B. T.; Zhokhavets, U.; Gobsch, G.; Raleva, S.; Stühn, B.; Schilinsky, P.;
Waldauf, C.; Brabec, C. J. Adv. Funct. Mater. 2005, 15, 1193.
111
(185) Lunt, R. R.; Giebink, N. C.; Belak, A. A.; Benziger, J. B.; Forrest, S. R.; Lunt, R. R.;
Giebink, N. C.; Belak, A. A.; Benziger, J. B.; Forrest, S. R. J. Appl. Phys. 2009, 105,
053711.
(186) Suppes, G.; Ballard, E.; Holdcroft, S. Polym. Chem. 2013, 4, 5345.
(187) van de Krol, R.; Grätzel, M. Photoelectrochemical Hydrogen Production; Springer:
New York, 2012; Vol. 102.
(188) Brülle, T.; Denisenko, A.; Sternschulte, H.; Stimming, U. Phys. Chem. Chem. Phys.
2011, 13, 12883.
(189) Bellani, S.; Ghadirzadeh, A.; Meda, L.; Savoini, A.; Tacca, A.; Marra, G.; Meira, R.;
Morgado, J.; Fonzo, F. Di; Antognazza, M. R. Adv. Funct. Mater. 2015, 25, 4531.
(190) Chowdhury, P.; Fortin, P.; Suppes, G.; Holdcroft, S. Macromol. Chem. Phys. 2016,
217, 1119.
(191) Scharber, M. C.; Sariciftci, N. S. Prog. Polym. Sci. 2013, 38, 1929.
(192) Vázquez, H.; Flores, F.; Oszwaldowski, R.; Ortega, J.; Pérez, R.; Kahn, A. Appl.
Surf. Sci. 2004, 234, 107.
(193) Crispin, X.; Geskin, V.; Crispin, A.; Cornil, J.; Lazzaroni, R.; Salaneck, W. R.;
Brédas, J.-L. J. Am. Chem. Soc. 2002, 124, 8131.
(194) Ishii, H.; Sugiyama, K.; Ito, E.; Seki, K. Adv. Mater. 1999, 11, 605.
(195) Hwang, J.; Wan, A.; Kahn, A. Mater. Sci. Eng. R Reports 2009, 64, 1.
(196) Bisquert, J.; Garcia-Belmonte, G.; Munar, A.; Sessolo, M.; Soriano, A.; Bolink, H. J.
Chem. Phys. Lett. 2008, 465, 57.
(197) Guerrero, A.; Marchesi, L. F.; Boix, P. P.; Ruiz-Raga, S.; Ripolles-Sanchis, T.;
Garcia-Belmonte, G.; Bisquert, J. ACS Nano 2012, 6, 3453.
(198) Flores, F.; Ortega, J.; Vázquez, H. Phys. Chem. Chem. Phys. 2009, 11, 8658.
(199) Vázquez, H.; Flores, F.; Kahn, A. Org. Electron. physics, Mater. Appl. 2007, 8, 241.
112
(200) Horowitz, G.; Allongue, P.; Cachet, H. J. Electrochem. Soc. 1984, 131, 2563.
(201) Kar, S.; Rajeshwar, K.; Singh, P.; DuBow, J. Sol. Energy 1979, 23, 129.
(202) Kumar, A.; Wilisch, W. C. A.; Lewis, N. S. Crit. Rev. Solid State Mater. Sci. 1993,
18, 327.
(203) Chen, Z.; Dinh, H. N.; Miller, E. Photoelectrochemical water splitting: standards,
experimental methods, and protocols; Springer: New York, 2013.
(204) Siripala, W. In Trends in Solar Energy Research; Hough, T., Ed.; Nova Science
Publishers, Inc.: New York, 2006; pp. 1–20.
(205) Siripala, W.; Perera, L. D. R. D.; De Silva, K. T. L.; Jayanetti, J. K. D. S.;
Dharmadasa, I. M. Sol. Energy Mater. Sol. Cells 1996, 44, 251.
(206) Yip, H.-L.; Jen, A. K.-Y. Energy Environ. Sci. 2012, 5, 5994.
(207) Groenendaal, L.; Jonas, F.; Freitag, D.; Pielartzik, H.; Reynolds, J. R. Adv. Mater.
2000, 12, 481.
(208) Xu, Z.; Chen, L.-M.; Chen, M.-H.; Li, G.; Yang, Y. Appl. Phys. Lett. 2009, 95,
013301.
(209) Sehati, P.; Braun, S.; Lindell, L.; Liu, X.; Andersson, L. M.; Fahlman, M. IEEE J.
Sel. Top. Quantum Electron. 2010, 16, 1718.
(210) Dibb, G. F. A.; Muth, M.-A.; Kirchartz, T.; Engmann, S.; Hoppe, H.; Gobsch, G.;
Thelakkat, M.; Blouin, N.; Tierney, S.; Carrasco-Orozco, M.; Durrant, J. R.; Nelson,
J. Sci. Rep. 2013, 3, 3335.
(211) Arnab, S. M.; Kabir, M. Z. J. Appl. Phys. 2014, 115, 034504.
(212) Brebels, J.; Manca, J. V.; Lutsen, L.; Vanderzande, D.; Maes, W. J. Mater. Chem.
A 2017, 5, 24037.
(213) Torabi, S.; Jahani, F.; Van Severen, I.; Kanimozhi, C.; Patil, S.; Havenith, R. W. A.;
Chiechi, R. C.; Lutsen, L.; Vanderzande, D. J. M.; Cleij, T. J.; Hummelen, J. C.;
Koster, L. J. A. Adv. Funct. Mater. 2015, 25, 150.
113
(214) Breselge, M.; Van Severen, I.; Lutsen, L.; Adriaensens, P.; Manca, J.;
Vanderzande, D.; Cleij, T. Thin Solid Films 2006, 511–512, 328.
(215) Chang, W. H.; Gao, J.; Dou, L.; Chen, C. C.; Liu, Y.; Yang, Y. Adv. Energy Mater.
2014, 4, 1300864.
(216) Lu, Y.; Xiao, Z.; Yuan, Y.; Wu, H.; An, Z.; Hou, Y.; Gao, C.; Huang, J. J. Mater.
Chem. C 2013, 1, 630.
(217) Patra, A.; Bendikov, M. J. Mater. Chem. 2010, 20, 422.
(218) Carrera, E. I.; Seferos, D. S. Macromolecules 2015, 48, 297.
(219) Scharber, M. C.; Muehlbacher, D.; Koppe, M.; Denk, P.; Waldauf, C.; Heeger, A.
J.; Brabec, C. J. Adv. Mater. 2006, 18, 789.
(220) Brédas, J. L.; Beljonne, D.; Coropceanu, V.; Cornil, J. Chem. Rev. 2004, 104, 4971.
(221) Gerischer, H. Pure Appl. Chem. 1980, 52, 2649.
(222) Li, Y. Acc. Chem. Res. 2012, 45, 723.
(223) Hou, J.; Park, M.; Zhang, S.; Yao, Y.; Chen, L.; Li, J.; Yang, Y. Macromolecules
2008, 41, 6012.
(224) Liu, D.; Zhao, W.; Zhang, S.; Ye, L.; Zheng, Z.; Cui, Y.; Chen, Y.; Hou, J.
Macromolecules 2015, 48, 5172.
(225) Zhang, M.; Guo, X.; Zhang, S.; Hou, J. Adv. Mater. 2014, 26, 1118.
(226) Blouin, N.; Michaud, A.; Leclerc, M. Adv. Mater. 2007, 19, 2295.
(227) Andersson, L. M.; Inganäs, O. Appl. Phys. Lett. 2006, 88, 082103.
(228) Chang, J.; Sun, B.; Breiby, D. W.; Nielsen, M. M.; Giles, M.; Mcculloch, I.;
Sirringhaus, H. Chem. Mater. 2004, 16, 4772.
(229) Salleo, A. Mater. Today 2007, 10, 38.
(230) Ebenhoch, B.; Thomson, S. A. J.; Genevičius, K.; Juška, G.; Samuel, I. D. W. Org.
114
Electron. 2015, 22, 62.
(231) Rolin, C.; Kang, E.; Lee, J. H.; Borghs, G.; Heremans, P.; Genoe, J. Nat. Commun.
2017, 8, 14975.
(232) Delongchamp, D. M.; Vogel, B. M.; Jung, Y.; Gurau, M. C.; Richter, C. A.; Kirillov,
O. A.; Obrzut, J.; Fischer, D. A.; Sambasivan, S.; Richter, L. J.; Lin, E. K. Chem.
Mater. 2005, 17, 5610.
(233) Meyerhofer, D. J. Appl. Phys. 1978, 49, 3993.
(234) Shioya, N.; Shimoaka, T.; Eda, K.; Hasegawa, T. Macromolecules 2017, 50, 5090.
(235) Kajiya, D.; Ozawa, S.; Koganezawa, T.; Saitow, K. J. Phys. Chem. C 2015, 119,
7987.
(236) Brinkmann, M.; Hartmann, L.; Biniek, L.; Tremel, K.; Kayunkid, N. Macromol. Rapid
Commun. 2014, 35, 9.
(237) Chou, S. Y.; Krauss, P. R.; Zhang, W.; Guo, L.; Zhuang, L. J. Vac. Sci. Technol. B
1997, 15, 2897.
(238) He, X.; Gao, F.; Tu, G.; Hasko, D. G.; Hüttner, S.; Greenham, N. C.; Steiner, U.;
Friend, R. H.; Huck, W. T. S. Adv. Funct. Mater. 2011, 21, 139.
(239) Cheyns, D.; Vasseur, K.; Rolin, C.; Genoe, J.; Poortmans, J.; Heremans, P.
Nanotechnology 2008, 19, 424016.
(240) Kim, J. S.; Park, Y.; Lee, D. Y.; Lee, J. H.; Park, J. H.; Kim, J. K.; Cho, K. Adv.
Funct. Mater. 2010, 20, 540.
(241) Cao, J.; Sun, J.; Shi, G.; Chen, H.; Zhang, Q.; Wang, D.; Wang, M. Mater. Chem.
Phys. 2003, 82, 44.
(242) Kowalski, D.; Tighineanu, A.; Schmuki, P. J. Mater. Chem. 2011, 21, 17909.