ABSTRACT
Title of dissertation: FUNCTIONAL IMAGING OF PHOTOVOLTAIC
MATERIALS AT THE NANOSCALE
Elizabeth M. Tennyson, Doctor of Philosophy, 2018
Dissertation directed by: Professor Marina S. Leite
Department of Materials Science
and Engineering, Institute for Research
in Electronics and Applied Physics
The ideal photovoltaic technology for global deployment must exhibit two key
attributes: (i) high power-conversion efficiency, enabling a solar panel with a large
power output per area and, (ii) low-cost/W, due to either being derived from earth-
abundant materials and/or ease of fabrication. For the past two decades, extensive
efforts have been made to boost the efficiency of some of the most promising high per-
formance and low-cost photovoltaic materials, such as CdTe, Cu(In,Ga)Se2 (CIGS),
and hybrid organic-inorganic perovskites, to achieve higher efficiency devices. How-
ever, improvement in the overall performance is still limited by the open-circuit
voltage (Voc). All of the solar cell materials listed above are composed of grains
and grain boundaries on the order of micro- and nanometers, respectively, and their
nanoscale interfaces can cause electrical charge carriers to become trapped and re-
combine non-radiatively, reducing the Voc. Therefore, in this thesis, I implement
high spatial resolution functional imaging techniques to resolve the local voltage
variations in the thin-film polycrystalline and hybrid perovskites materials for pho-
tovoltaic applications.
First, I spectrally and spatially resolve the local photovoltage of CIGS solar
cells through confocal optical microscopy to build a qualitative voltage tomography.
From these photovoltage results, I discover variations in the electrical response of
>20% that are also on the same length scale as the grains composing the CIGS ma-
terial. Therefore, by enhancing the spatial resolution beyond the diffraction limit,
the electronic properties of individual grains and the interfaces between the grains
can be fully resolved. For this, I implement Kelvin probe force microscopy (KPFM),
and demonstrate a universal method to directly map the Voc of any photovoltaic ma-
terial with nanoscale spatial resolution. Next, we extend this ability of KPFM to
rapidly image (16 sec/map) the real-time dynamics of perovskite solar cells, which
are notorious for their slow and unstable electrical output. Through fast-KPFM
imaging, we discover regions within a single grain that show a residual Voc response
which pervades for ≈9 min, likely caused by a slow ion migration process. Finally,
to understand how different perovskite compositions influence the behavior of the
nanoscale electrical response, I utilize KPFM to realize both irreversible and re-
versible Voc signals. Compiling all these results discussed above, throughout my
Ph.D. I have yielded the following contributions: (i) evidence that the photovoltage
of polycrystalline solar cell materials varies at the same length scale as the grains
composing them, (ii) a nanoscale imaging platform to directly map the Voc with
unprecedented spatial resolution, and (iii) a technique to map the real-time voltage
response of many perovskite compositions, ultimately indicating that the elements
constituting the perovskite cation and halide positions are both directly related to
their reversible vs. irreversible electrical nature. From these contributions, I foresee
the functional imaging methods developed in this thesis to be widely implemented
as a diagnostic tool for the rational design of photovoltaics with enhanced electrical
performance and lower cost.
FUNCTIONAL IMAGING OF PHOTOVOLTAIC
MATERIALS AT THE NANOSCALE
by
Elizabeth M. Tennyson
Dissertation submitted to the Faculty of the Graduate School of the
University of Maryland, College Park, in partial fulfillment
of the requirements for the degree of
Doctorate of Philosophy
2018
Advisory Committee:
Professor Marina S. Leite, Chair/Advisor
Professor Eric D. Wachsman
Professor Richard L. Greene
Professor John Cumings
Professor Gottlieb S. Oehrlein
©c Copyright by
Elizabeth M. Tennyson
2018

Acknowledgments
I would like to dedicate this Ph.D. thesis to everyone who has helped and
supported me along this graduate school journey. To start, I would like express
my gratitude for my adviser Prof. Marina S. Leite. Your guidance has made me a
better scientist, writer, presenter, critical thinker, and researcher. You’ve taught me
countless useful research skills and tips, and because of your encouragement I know
that I will do well after graduate school. Although it wasn’t always easy, thank you
for the motivating and believing in me. I’d also like to thank the members of my
Ph.D. Thesis Committee: Profs. John Cumings, Richard Greene, Gottlieb Oehrlein,
and Eric Wachsman for their perspective and ideas.
To past and present Leite Lab members/friends: Chen Gong, Mariama Dias,
John Howard, Zack Benson, Allen Chang, Eric Bailey, William Gunnarsson (i.e.
Brady), Bernardo Neves, Alex Yulaev, Sabrina Curtis, Griffin Godbey, Garrett
Wessler, and Zinab Jadidi. I want to express particular emphasis on Chen, who
joined the Leite Lab at the same time as I did. You and I helped pioneer the lab
and I am very grateful to have you by my side for that, you are/were a brillant and
excellent group member. Also, Mariama Dias, thank you! You are an incredible
mentor and I am so glad to have you in my life.
To members/friends/collaborators of the Munday Lab: Joseph Garrett, Lisa
Krayer, Yunlu Xu, and Joseph Murray. In particular, I owe a lot of my AFM
knowledge and skills to Joe G. Thank you for being patient, a great teacher, and
a kind person. I’d also like to give a special thanks Lisa for being an excellent
resource and fun person to discuss research and life, with. Lastly, I’d like to thank
Prof. Jeremy Munday for letting me to use his lab’s instruments, tools, and optics
ii
while Chen, Prof. Leite, and I were still building up the READ Lab.
To the other non-MSE and non-IREAP collaborators that either helped with
obtaining a result or provided solar cell samples for the research performed in
this Thesis. The folks at Naval Research Lab: Jesse Frantz, Jason Myers, Ro-
bel Bekele, and Jasbinger Sanghera. University of Nebraska: Jinsong Huang and
Miao Hu. From UMD Aerospace department: Suok-Min Na. From ORNL/CNMS:
Liam Collins, Alison Pawlicki, Anton Ievlev, and Stephen Jesse. From HZB: Anto-
nio Abate. From the University of Cambridge: Bart Roose, Mojtaba Abdi Jalebi,
Kangyu Ji, and Sam Stranks.
There are also many staff members at the University of Maryland who deserve
my gratitude for helping me with research challenges, health care logistics, pay
roll, etc. From the MSE Department: Kathleen Hart, Nina Morris (Kay), Mike
McNicholas, and Patricia Lorenzana. From SAC: Karen Gaskell. From IREAP:
Nolan Ballew, Jay Pyle, Nancy Boone, Dottie Brosius (special shout out for hosting
me at Thanksgiving 2016!), Ed Condon, Judi Gorski, Kathryn Tracey Metzler,
Taylor Prendergast, and Bryan Quinn. The Nanocenter: Valery Ray, Mark Lecates,
Jon Hummel, John Abrahams (rock on), and Tom Loughran (who also always made
a special effort to check up on my family).
To UMD-MSE Alumni who proved to me that graduation was possible and
provided me with some excellent guidance, advice, and exclusive resources: Travis
Dietz, Romaine Isaacs, Josh Taillon, Wonseok Hwang, Sean Fackler, Norvik Voska-
nian, Dominik Metzler, Jasper Drisko, Rick Suchoski, Alex Kozen, Greg Hitz, Elliot
Bartis, Thomas Hays, Adam Jolley, Matthew Murray, Eric Rosenthal, Alex Pearse,
and Gary Paradee.
I graciously acknowledge all of the funding agencies and organizations who
have awarded me with monetary support so that I could accomplish and complete
this dissertation: the James A. Clark School of Engineering, the UMD Dean’s Fel-
lowship, the UMD Jacob K. Goldhaber travel grant, the 2015 and 2017 UMD Grad-
uate School Summer Research Fellowship, the 2016-2017 UMD Graduate Deans
Dissertation Fellowship, the 2017-2018 Hulka Energy Research Fellowship, and the
iii
2017-2018 All-S.T.A.R. Fellowship. Other monetary awards include: MRS Gradu-
ate Student Silver Award, OSA Student Poster Competition, IREAP best speaker
award, UMD Researchfest 2015, IEEE 40th PVSC best poster.
To the 2016 Joint US-Africa Materials Institute for giving me the opportunity
of a lifetime. In particular, the organizers: Profs. Sossina Haile, Peter Green, and
Simon Billinge.
It is also extremely important that I thank the all the professors from the
Physics department at my alma mater, University of Wisconsin - La Crosse. There
are a particular few that inspired and supported my decision to attend graduate
school at Maryland: Profs. Sudha, Seth King, and Shauna Sallmen. The incredible
atmosphere that the professors brought to the UWL Physics department is what
motivated me to continue on with my education.
I thank all of my dear friends who have either shared my woes, listened to
my rants, or helped distract me from my work when I needed a break (some of
these brave individuals even held the honor of performing all three of these tasks).
Some include: Manatee(Travis Dietz, Katie Frey, Michael Van Order, Andrew Knoll,
and Alyson Becker) our Sunday night dinners, will always be cherished by me, and
may Zengo forever rest in peace. Some amazing ladies back in the midwest: Kelse
Brellenthin (Triplett? It depends when this thesis gets published), Ceejae Lynch,
and Smerg, for listening and being my life-long friends, if there is one thing that is a
constant in this life it will be our friendship. Kwembur Morks (a.k.a. Morko) for the
advice, support, and for sharing a perspective I cannot get anywhere else. Andrew
Prudhom, who was the only person who understood my La Crosse to graduate school
transition. Katie and Dave Shahin, for the laughs, the games, the competitions, and
the ride in the Tesla. Long Nyugen for being my first official office mate, continuing
friend, and introducing me to many of your great friends such as Vikram Um and
Adria Schwarber who just hijacked your shout out. Tristan McAloon and Sammi
for pushing me to stay my first year when I almost did not. Also: Bethany and
Patrick Stanley, Teddy Wilson, Pinghsan Luan, Jiaqi Dai, and Andrew Dunkman.
To Travis C. Dietz, PhD, who was with me for every step along the way in
iv
my graduate school adventure. You made the process infinitely better, it is just a
shame we won’t be able to quantify it. From all-nighters spent studying for exams,
to all-nighters spent writing papers/dissertations, and everything in between (Why
crystal melt?). Thank you for all the love and laughter you have shared with me
these past five years. I look forward to seeing what life brings with you post-graduate
school. This is our last chance to figure out the mysterious green Jello conundrum.
Let’s see how it goes.
I know that I will not be capable of completely expressing my gratitude for
my family, but I can try. To Joe: I wish you were here, and to my other siblings:
Jason, Dan, Linsey, and Scott, thank you for never asking when I might graduate or
what I am doing with my life, because I never knew, and to my siblings’ significant
others: Marcia, Tiffany, and Chrisshon for their kind words of motivation. To my
nieces: Ariana, Aubrey, McKenzy, and Ivy — this is for you, I want you to know
that you can live whatever life you want and don’t have to settle for anything. Re-
member to always embrace these wise words from Beyoncé: “Okay ladies now let’s
get in-formation”. To my nephews: Gabriel, Oliver, and Quentin — Work hard
to reach your goals in life, whatever they may end up being, I already know you
boys all have your unique abilities. To Dennis and Mike, thanks for expressing your
interest in my research, or at least pretending to. To Nicki: for checking in and
always remembering to wish me well whenever there was an achievement or event
(big or small) related to graduate school. I know you always enjoyed when I sent
you my papers. Lastly, to my incredible parents: to say “I wouldn’t be here without
them” goes beyond just the physical and literal meaning, their undeniable mental
encouragement and belief in me was unparalleled. I am astounded by all their love
and support. I mean, they read every paper, that’s amazing! Thanks Dad for always
double checking my results by comparing them to your own sophisticated research.
Elizabeth M. Tennyson
February 2018
College Park, MD
v
Table of Contents v
List of Figures viii
List of Tables x
1 Introduction to Photovoltaics 1
1.1 The need for photovoltaics . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Pursuing high-performance and low-cost PV . . . . . . . . . . . . . . 2
1.3 Fundamentals of solar cells . . . . . . . . . . . . . . . . . . . . . . . . 3
2 Research Objectives 7
3 Scanning Probe Microscopy of Energy Harvesting Systems 9
3.1 Introduction to scanning probe microscopy . . . . . . . . . . . . . . . 10
3.2 AFM for energy harvesting systems . . . . . . . . . . . . . . . . . . . 12
3.2.1 Hybrid perovskites and thin-film materials for PV: opportu-
nities and current limitations . . . . . . . . . . . . . . . . . . 13
3.2.2 Probing electrical properties . . . . . . . . . . . . . . . . . . . 20
3.2.3 Imaging local chemical properties . . . . . . . . . . . . . . . . 29
3.2.4 Optical properties at the nanoscale . . . . . . . . . . . . . . . 33
3.3 Probing dynamic processes in perovskite solar cells . . . . . . . . . . 38
3.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4 Photovoltage Tomography in Polycrystalline Solar Cells 45
4.1 The need for tomography in thin film photovoltaic devices . . . . . . 46
4.2 Experimental methods . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.2.1 CIGS solar cell fabrication . . . . . . . . . . . . . . . . . . . . 48
4.2.2 Device macroscopic characterization . . . . . . . . . . . . . . . 49
4.2.3 Raman spectroscopy mapping . . . . . . . . . . . . . . . . . . 50
4.2.4 Photovoltage imaging . . . . . . . . . . . . . . . . . . . . . . . 50
4.3 Correlating chemical and electrical maps in CIGS . . . . . . . . . . . 51
4.4 Spectrally dependent photovoltage imaging results and discussion . . 54
4.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
vi
5 Nano-imaging of open-circuit voltage in photovoltaics 67
5.1 Introduction to local voltage mapping in thin-film solar cells . . . . . 68
5.2 Experimental methods . . . . . . . . . . . . . . . . . . . . . . . . . . 71
5.2.1 Solar cell fabrication . . . . . . . . . . . . . . . . . . . . . . . 71
5.2.2 KPFM measurements . . . . . . . . . . . . . . . . . . . . . . . 72
5.2.3 Macroscopic characterization of devices . . . . . . . . . . . . . 72
5.2.4 EQE measurements of GaAs . . . . . . . . . . . . . . . . . . . 74
5.2.5 Voc image analysis . . . . . . . . . . . . . . . . . . . . . . . . . 75
5.3 Direct imaging of Voc at the nanoscale: results and discussion . . . . . 75
5.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
6 Real-time nanoscale open-circuit voltage dynamics of perovskite
solar cells 89
6.1 Introduction to nano-imaging in perovskite solar cells . . . . . . . . . 90
6.2 Experimental methods . . . . . . . . . . . . . . . . . . . . . . . . . . 92
6.2.1 Perovskite solar cell fabrication . . . . . . . . . . . . . . . . . 92
6.2.2 Illuminated Kelvin-probe force microscopy Voc derivation . . . 93
6.2.3 KPFM implementation . . . . . . . . . . . . . . . . . . . . . . 97
6.2.4 Image analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
6.3 Fast-KPFM results and discussion . . . . . . . . . . . . . . . . . . . . 99
6.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
7 Perovskite composition-dependent nanoscale voltage response 111
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
7.2 Experimental methods . . . . . . . . . . . . . . . . . . . . . . . . . . 114
7.2.1 ToF-SIMS characterization . . . . . . . . . . . . . . . . . . . . 114
7.2.2 Raman spectroscopy and microscopy . . . . . . . . . . . . . . 114
7.2.3 Solar cell synthesis . . . . . . . . . . . . . . . . . . . . . . . . 114
7.2.4 Macroscopic measurements . . . . . . . . . . . . . . . . . . . . 116
7.2.5 Kelvin probe force microscopy experiments . . . . . . . . . . . 116
7.3 Composition analysis of perovskites . . . . . . . . . . . . . . . . . . . 117
7.4 Probing perovskites’ electrical response at the nanoscale . . . . . . . 119
7.4.1 Irreversible nanoscale electrical response in MAPbBr3 . . . . . 121
7.4.2 Fully reversible and stable Cs-incorporated triple cation per-
ovskites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
7.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
8 Perspective on functional imaging of PV materials 133
8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
8.2 Electrons as a source of excitation . . . . . . . . . . . . . . . . . . . . 134
8.3 Photons as a source of excitation . . . . . . . . . . . . . . . . . . . . 139
8.4 Probing the degradation of perovskite solar cells . . . . . . . . . . . . 144
8.5 Correlative microscopy and the need for big data analytics . . . . . . 146
8.6 Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
8.7 Future work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
vii
A Products of this research 151
A.1 Awards and honors . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
A.2 Publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
A.3 Presentations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
B Additional publications 155
C University of Maryland service 157
C.1 Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
Bibliography 159
viii
List of Figures
1.1 NREL 2018 efficiency chart . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Macroscopic dark and light J-V curves . . . . . . . . . . . . . . . . . 4
1.3 Radiative and non-radiative charge carrier recombination . . . . . . . 6
3.1 Principle of AFM operation . . . . . . . . . . . . . . . . . . . . . . . 11
3.2 SEM images of selected polycrystalline materials for photovoltaics . . 14
3.3 Nanoimaging electrical properties by AFM . . . . . . . . . . . . . . . 21
3.4 Photovoltaic performance at the nanoscale . . . . . . . . . . . . . . . 28
3.5 Nanoimaging chemical properties in solar cell materials by AFM . . . 31
3.6 Nanoimaging optical properties in solar cell materials by AFM . . . . 34
3.7 Imaging electrical, chemical, and optical dynamics in perovskites . . . 40
4.1 CIGS polycrystalline solar cells . . . . . . . . . . . . . . . . . . . . . 49
4.2 Scanning photovoltage experimental setup . . . . . . . . . . . . . . . 51
4.3 CIGS electron backscatter diffraction pattern . . . . . . . . . . . . . . 52
4.4 Chemical composition and local electrical response of CIGS solar cells 53
4.5 Raman microscopy on CIGS: additional region . . . . . . . . . . . . . 54
4.6 Laser ON/OFF photovoltage signal verification . . . . . . . . . . . . 55
4.7 Additional photovoltage maps of CIGS solar cells . . . . . . . . . . . 55
4.8 Spectrally dependent voltage maps on CIGS solar cell . . . . . . . . . 57
4.9 Spectrally dependent voltage histograms . . . . . . . . . . . . . . . . 58
4.10 CIGS optical properties . . . . . . . . . . . . . . . . . . . . . . . . . 59
4.11 SEM images of as-is vs. polished CIGS grains . . . . . . . . . . . . . 59
4.12 Electron, optical, scanning probe microscopy CIGS grains . . . . . . . 60
4.13 Voltage distribution of a polished CIGS sample . . . . . . . . . . . . 61
4.14 Mapping voltage with high spatial resolution in CIGS solar cells . . . 63
4.15 Characterization of the mesoscale behavior of CIGS solar cells . . . . 64
5.1 Light J-V curves for all solar cell devices . . . . . . . . . . . . . . . . 73
5.2 Dark J-V for GaAs solar cell . . . . . . . . . . . . . . . . . . . . . . 73
5.3 Angle dependence of external quantum efficiency for GaAs device . . 74
5.4 Energy diagram of KPFM measurements . . . . . . . . . . . . . . . . 76
5.5 Macroscopic characterization of GaAs . . . . . . . . . . . . . . . . . . 79
5.6 Voc vs. photon flux . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
5.7 Comparison of macro- and nanoscale Voc measurements . . . . . . . . 80
5.8 KPFM performed on four solar cell devices . . . . . . . . . . . . . . . 82
ix
5.9 Directional independent KPFM imaging on Si . . . . . . . . . . . . . 83
5.10 Influence of ∆H on voltage response . . . . . . . . . . . . . . . . . . . 85
5.11 Quantifying KPFM spatial resolution . . . . . . . . . . . . . . . . . . 86
6.1 Real-time measurements of voltage dynamics in perovskites . . . . . . 93
6.2 Macroscopic characterization of perovskite solar cell device . . . . . . 94
6.3 SEM images of perovskite thin film . . . . . . . . . . . . . . . . . . . 94
6.4 Monitoring all relevant signals during FM-KPFM measurements . . . 96
6.5 Environmental conditions during in situ imaging of perovskites . . . . 97
6.6 AFM topography images . . . . . . . . . . . . . . . . . . . . . . . . . 98
6.7 Imaging nanoscale variations in the Voc of perovskite solar cells . . . . 100
6.8 Dynamics of perovskite solar cells at the nanoscale . . . . . . . . . . 102
6.9 Voltage profile for fast H-KPFM illumination scans . . . . . . . . . . 104
6.10 Spatial and temporal variation of the residual voltage within perovskites106
6.11 Real-time nanoscale Voc dynamics of perovskite photovoltaics . . . . . 108
7.1 Chemical composition analysis of perovskites by ToF-SIMS . . . . . . 118
7.2 Raman microscopy on perovskite thin films . . . . . . . . . . . . . . . 119
7.3 Macroscopic output and short-term stability of perovskite devices . . 120
7.4 Imaging the Voc response in perovskites at the nanoscale . . . . . . . 122
7.5 Relative humidity during KPFM measurements . . . . . . . . . . . . 122
7.6 Nanoscale irreversible voltage response in MAPbBr3 perovskites . . . 124
7.7 Dynamic, reversible electrical response of MAPbBr3 . . . . . . . . . . 126
7.8 Dynamic, reversible electrical response of MAPbBr3: region 2 . . . . 127
7.9 Nanoscale reversible voltage responses in Cs-mixed perovskites . . . . 128
7.10 Dynamic electrical response of Cs-mixed perovskite . . . . . . . . . . 129
7.11 Analysis of perovskites nanoscale electrical stability . . . . . . . . . . 131
8.1 Electrons as an excitation source to nano-image PV functionality . . 135
8.2 Photons as a source of excitation to nano-image PV functionality . . 141
8.3 Imaging the nanoscale chemical distribution of a perovskite thin-film 149
x
List of Tables
3.1 Theoretical vs Experimental Solar Cell Parameters . . . . . . . . . . 16
8.1 Nanoscale functional imaging methods performed on PV materials . . 137
xi
Chapter 1: Introduction to Photovoltaics
1.1 The need for photovoltaics
To swiftly and effectively replace of fossil energy-based technologies we re-
quire a high-efficiency, low-cost, and ubiquitous photovoltaic (PV) technology. For-
tunately, renewable energy is expected to become a significant percentage of the
global energy sources as the total amount of power consumed globally continues
to increase each year. According to the National Renewable Energy Laboratory
(NREL) Renewable Electricity Futures Study, 80% of the power generated in the
United States could be from renewable sources by 2050. [1] For this to occur PV
electricity generation must increase beyond 10% in the US, [2] as only 1-2% of our
electricity is currently produced from solar power today. [3] In general, single crys-
talline PV materials generally outperform their polycrystalline counterparts, [4] yet
the extremely high cost associated with monocrystalline epitaxial fabrication meth-
ods does not outweigh the benefits of efficient PV devices, hindering implementation.
Non-epitaxial compounds, such as hybrid perovskites and thin-film polycrystalline
materials are very promising alternatives, but improvement in device stability and
power-conversion efficiency (η) are still required for successful deployment, as will
be discussed in detail throughout this dissertation.
1
1.2 Pursuing high-performance and low-cost PV
Polycrystalline materials such as CuInxGa(1−x)Se2 (CIGS), CdTe, or hybrid
organic-inorganic perovskites are a potential alternative for low-cost PV, but im-
provement η is required to ensure low-cost/W, see Figure 1.1. The efficiency chart [5]
displays each PV technology’s record performance as a function of time. CIGS,
CdTe, and perovskites all produce η >22%, however Silicon, which represents ∼90%
of the solar cell market, boasts η >25%. Extensive efforts have been made to in-
crease the short circuit current density (Jsc) of thin-film polycrystalline solar cells.
Yet, the performance of the best solar cells is still limited by the open-circuit volt-
age (Voc), the maximum voltage available from a PV device, which is related to the
radiative recombination (photon emission) within the material. The Voc of both well-
established and emerging polycrystalline materials is considerably below theoretical
predictions. [6–13]
2
Figure 1.1: NREL 2018 efficiency chart. Accessed online from ref [5].
The key pathway to achieve higher efficiency is through minimizing the non-
radiative recombination processes within the semiconductor. [14–23] Therefore, I
suggest investigating the Voc, as it is related to the amount of carrier recombination
in the material and it plays a key role in the overall performance of solar cells.
1.3 Fundamentals of solar cells
Solar cells are fundamentally governed by the diode equation, where the Voc
depends on the dark current (I0) and on the short-circuit current (Isc) as:
[ ] [ ]
nkT Isc
Voc = ln + 1 ≈
nkT Isc
ln (1.1)
q I0 q I0
n is the diode ideality factor, k is Boltzmann’s constant, T is the temperature, q is
3
the electronic charge, and I0 depends on recombination events within the material.
Figure 1.2 shows the most common macroscopic characterization technique for PV
devices: a dark and light current-voltage (J-V ) curve. At zero bias in the dark, a
solar cell does not generate any current and will act as a diode, only allowing charge
carriers to flow in a specific direction. When excited by photons, the light-absorbing
PV material utilizes the energy from the light to produce charge carriers. Ideally, the
generated electrons and holes are efficiently separated, swept toward and collected
by the electrical contacts, with minimal losses. This process introduces a light-
generated current, indicated as Jsc in Figure 1.2. The solar cell power-conversion
efficiency is calculated by maximum power point (Mpp) divided by the incident
power.
Figure 1.2: Dark and light J-V curves. Black curve: Dark J-V, the curve matches
that of a standard diode. Blue curve: Light J-V, here the solar cell produces a light-
generated current, therefore, performing work at specific external voltage biases. Jsc:
short-circuit current density, Voc: open-circuit voltage, Mpp: maximum power point.
For an ideal PV device, the Voc is only limited by radiative recombination pro-
cesses and is equal to the material bandgap term (Eg ) in volts minus the thermody-
q
namic losses: Voc ≈ (Eg )−0.4. [6,24–27] This equation was first derived by Shockleyq
4
and Queisser in 1961, and is so-called the Shockley-Queisser limit. [6] Generation
and recombination are electronic excitation and relaxation processes that increase
and reduce the number of free charge carriers, respectively. The most relevant type
of generation in PV is the absorption of a photon. For every generation process
there is an equivalent recombination process, which is typically the loss of a carrier
through the decay of an electron to a lower energy state. When the energy associ-
ated with this decay is released by the emission of a photon that has an equivalent
energy to the material bandgap (Eg), the event is termed radiative recombination,
see Figure 1.3(a). This thermodynamically necessary process is dominant in direct
Eg materials, such as GaAs, perovskites, CIGS, and CdTe, and is not necessar-
ily harmful to the device. [24] Nevertheless, in non-radiative recombination, Figure
1.3(b), electron-hole pairs recombine at defect states, inhibiting the collection of a
charge carrier, generating heat, and unavoidably degrading the performance. There
are other types of non-radiative recombination events, such as Auger, but here I am
highlighting only the most relevant processes for multi-crystalline solar cell materi-
als.
5
Figure 1.3: Radiative and non-radiative charge carrier recombination. Pos-
sible recombination events that may occur after a incident photon is absorbed by
a solar cell material. (a) Radiative recombination: an electron directly recombines
with a hole and emits a photon the same energy as the semicondutor’s bandgap.
(b) Trap-assisted non-radiative recombination: an electron relaxes to a defect energy
state and emits a photon with an energy much lower than the bandgap.
Polycrystalline light-absorbers are composed of micron-sized grains, and thus,
consist of many interfaces or defects, all of which could potentially be sites of trap-
assisted recombination. Therefore, while macroscopic light J-V measurements are
useful for determining the average device performance, a nanoscale imaging spec-
troscopy method to spatially resolve the local optical and electrical properties to
access charge carrier recombination processes within polycrystalline materials for
PV is still missing and is critical for understanding device deficiencies. Throughout
this dissertation I advance our understanding of materials for PV by resolving the
charge generation and collection with nanoscale resolution. For this, I have imple-
mented novel nanospectroscopic imaging metrologies, such as scanning photovoltage
and Kelvin probe force microscopies.
6
Chapter 2: Research Objectives
In this dissertation I accomplish the following research objectives by developing
and implementing functional imaging techniques to map the local electronic charge
carrier processes (i.e. generation, recombination, and collection) within promising,
next-generation PV materials.
(i) Photovoltage tomography in polycrystalline solar cells. By selecting
the wavelength of the local source of illumination, I build a photovoltage tomography
of the PV material, based on its absorption coefficient and the wavelength-dependent
SPCM. Our results using scanning confocal microscopy (with a 100× objective lens)
show spatial variations in the generation of carriers. The light penetration depth
is directly related to the experimentally determined material absorption coefficient
(α = (4πk )) which establishes a tomography of photovoltage in three dimensions.
λ
(ii) Nano-imaging of open-circuit voltage in photovoltaics. Using
Kelvin probe force microscopy (KPFM) I locally probe the open-circuit voltage
(Voc) with nanoscale resolution of solar cells through a series of dark and illumi-
nated KPFM scans. Since KPFM is a variant of scanning probe microscopy, both
the morphology and electrical properties of the solar cell material are simultaneously
acquired. KPFM measures the work function difference between the cantilever tip
7
and sample surface. Upon solar cell illumination, the work function shifts, and the
local Voc is calculated from this change. Here, I reveal new insights about how the
grain boundaries within polycrystalline devices affect the overall performance.
(iii) Real-time nanoscale open-circuit voltage dynamics of perovskite
solar cells. A variant of KPFM is utilized to rapidly image a specimen with
temporal resolution on the order of seconds, more than two orders of magnitude
faster than conventional frequency modulated KPFM. By using a scanning probe
technique with such high spatial and temporal resolution we realize how the material
changes electrically with time. This is especially interesting for perovskite solar
cells, as they are known for their electrical instability when exposed to air and/or
moisture. Here, I will image the dynamics of perovskite devices in real-time.
(iv) Imaging fully reversible and stable perovskites at the nanoscale.
With the wide variety of perovskite materials available for PV applications, it is
crucial to determine which composition will have the highest electrical stability and
moisture tolerance. I leverage my expertise in KFPM and probe multiple perovskite
thin-films to measure both irreversible and reversible voltage responses. I further use
Fast-KPFM, described in Objective (iii), to track the irreversible electrical signal.
8
Chapter 3: Scanning Probe Microscopy of Energy Harvesting Sys-
tems
Our scientific understanding of the nanoscale world is continuously growing
ever since atomic force microscopy (AFM) has enabled us to “see” materials at this
length scale. Beyond morphology, functional imaging is becoming standard practice
as new AFM-based techniques are continuously extending its capabilities. Resolving
material properties with high spatial accuracy is now extremely critical, as future
next-generation energy harvesting systems are comprised of complex and intricate
nanoscale features. Here, we provide a literature review of recent research discover-
ies that implemented AFM methods to measure and determine how the electrical,
chemical, and/or optical properties influence the overall device behavior. We dedi-
cate a portion of this chapter to perovskite solar cells, which are of primary interest
to photovoltaic research, and highlight the remarkable progress made towards un-
derstanding and controlling their instabilities. We conclude with a summary and
outlook anticipating the most pressing materials related challenges associated with
solar cells, and how they will likely be overcome in the near future by nanoimaging
through AFM. This chapter is adapted from E.M. Tennyson et. al, ACS Energy
Letters, 2 2761-2777 (2017) Front Cover [28]
9
3.1 Introduction to scanning probe microscopy
Imagine gently gliding along the surface of an object with the tip of your index
finger — what types of information can you obtain? Most likely, you will feel the texture
(surface roughness), temperature, shape, hardness, etc., of the item. All this information is
acquired in less than one second by neurons firing from your fingertip to your brain. Atomic
force microscopy (AFM) is analogous to this situation; except, replace your fingertip by
a nanoscale probe and the neuron reactions with the microscope detection system. AFM
is, thus, an extremely powerful tool that probes the morphological, electrical, magnetic,
mechanical, chemical, and optical properties of the sample under examination. Since its
first implementation in the 1980’s, [29] AFM has revolutionized our understanding of a
variety of nanoscale PV systems, [30,31], as will be presented in this Chapter.
In most AFM-based setups there is a probe with a radius ranging from 1-100 nm
that is positioned very close to or in contact with the surface of a sample, illustrated in
Figure 3.1. The spatial resolution of AFM imaging methods is, then, primarily limited
by the size of the probe. The interaction between the AFM tip and sample is detected
by a photodiode that collects laser light deflected off the back of the cantilever. The
photodiode is calibrated with specified x-y coordinates and continuously keeps track of
the laser position. Therefore, any minute movement of the probe, whether caused by
a small change in sample height or electronic carrier concentration, is detected. The
feedback loop system adjusts the cantilever accordingly, depending on the measurement
mode of interest. There are many forces and interactions, such as electrostatic, capillary,
or chemical potential, that can excite the tip and/or sample causing the deflection laser’s
x-y coordinates to change. Beyond the surface roughness, [32, 33] properties such as the
10
local electrical, chemical, and optical spatial variations are imaged, which will be discussed
in detail later.
photodiode AFM controller + 
feedback
deflection laser
control and 
analysis
Figure 3.1: Principle of AFM operation. Illustration of an AFM probe scanning
the surface of a textured, non-uniform sample. A laser is deflected from the back of
the probe onto a photodiode and any change in the tip-sample interaction is translated
as the laser’s coordinates on the photodiode. As a consequence, the feedback system
within the AFM adjusts itself accordingly. Out of scale for clarity.
Here, we introduce the remarkable progress made in determining the properties and
performance of materials for energy harvesting by AFM through functional and correlative
nanoscale imaging. The micro- and nanostructure of heterogeneous photovoltaic (PV) and
materials is well known to influence their overall performance. [34, 35] Therefore, under-
standing the electrical and chemical processes that harm device operation is crucial for
future renewable energy generation, conversion, and storage systems. We choose to focus
on AFM methods because they encompasses key desirable features such as: the ability
to perform measurements in operando conditions with 3D nanoscale spatial resolution, is
non-destructive, and requires very little sample preparation. Further, AFM is undoubt-
edly a versatile tool to image the functionality of energy harvesting materials, and its
full potential has yet to be realized, as new techniques will shed light on faster, dynamic
processes. We target selected AFM-based methods that are specifically implemented to
11
investigate the nanoscale properties of energy conversion systems that translate into device
performance, such as the electrical and optical response of solar cells (through measure-
ments of photovoltage, photon-induced current, and charge carrier lifetime), and chemical
composition. In the realm of photovoltaics, we provide an in-depth discussion of each
imaging method that has been utilized to probe relevant properties of next-generation
solar cells. In particular, we focus on performance mapping and perovskite dynamics as
it is at the forefront of interest for scientists and engineers working on PV. These novel
scanning methods have the potential to elucidate key open questions about the physical
behavior of perovskite devices.
3.2 AFM for energy harvesting systems
The most promising emerging PV materials consist of micro-structured semicon-
ductors with boundaries and interfaces that can lead to undesirable non-radiative re-
combination of charge carriers, reducing the power-conversion efficiency (η) of the device
significantly below its theoretical limit. [34,35] The applications of these heterogeneous ma-
terials, including perovskites, CdTe, Cu(In,Ga)Se2 (CIGS), Cu2ZnSnS4 (CZTS), quantum
dots, organics, dye-sensitized, etc., are aimed toward lower-cost and/or higher-efficiency
solar markets. Often, the devices are fabricated as thin-films (≤3 µm thickness), which
can be deposited on a variety of surfaces with distinct mechanical properties, including
flexible substrates. Despite the potential low cost ($/kW) of these devices, it is currently
not well understood how their structure affects charge carrier generation, recombination,
and collection at the nanoscale. Thus, imaging the key physical and chemical properties
of these materials is crucial to their further development. In this section of the Thesis
I describe how AFM methods have been realized to probe the electrical, chemical, and
12
optical properties of non-epitaxial materials for PV.
3.2.1 Hybrid perovskites and thin-film materials for PV: opportunities
and current limitations
The absorbing layer of halide hybrid perovskites and of most polycrystalline mate-
rials for PV (CdTe, CIGS, CZTS, and polycrystalline GaAs) is composed of grains in the
microscale range, as shown in Figure 3.2, with well-defined facets, interfaces, and bound-
aries. The macroscopic performance of the devices strongly depends on these mesoscale
constructs, where the behavior of the average density of atoms corresponding to length
scales of 5-50 nm (and to a volume commonly comprised of thousands of atoms) is more
relevant than the contribution of the individual atoms themselves. [36] Thus, it is imper-
ative to resolve the performance of the device at similar, relevant length scales. While
conventional electrical characterization methods are extremely useful, they do not provide
information about how each grain and interface contributes to/affects the electrical and
optical responses of the PV devices. Scanning probe microscopy (SPM) and scanning
(transmission) electron microscopies (STEM and SEM) have been successfully applied to
characterize materials for solar cells with high spatial resolution. [37–39] Recently, mi-
croscopic techniques have been combined to investigate the structural, optical, chemical,
and electrical properties of a variety of relevant materials. [40,41] However, the nanoscale
functional imaging of critical mesoscale phenomena in PV materials and in operando de-
vices is still missing. The performance of PV devices is defined by its main figures-of-merit:
open-circuit voltage (Voc), short-circuit current density (Jsc), fill factor (FF), and η. Thus,
the ability to map these quantities with nanoscale spatial resolution will likely unravel the
intricate physical and chemical processes that define device functionality at the mesoscale,
13
elucidating the origin of the modest performance of some technologies, and of the light-
induced chemical reactions that take place in perovskites. Below we summarize the major
advantages and challenges of the selected non-uniform solar cell materials.
(a) Perovskite (b) CdTe (c) CIGS (d) CZTS (e) Polycrystalline GaAs
3 μm
Figure 3.2: SEM images of selected polycrystalline materials for photo-
voltaics: (a) hybrid perovskites, reprinted from ref [42], (b) CdTe, (c) CIGS, (d)
CZTS, reprinted from ref [43], and (e) polycrystalline GaAs, reprinted from ref [44].
In all cases, the absorbing layer is composed of micron sized grains or smaller, where
their optical and electrical responses vary in the same length scale.
3.2.1.1 Hybrid organic-inorganic perovskites
Since 2009, the research community has witnessed the rise of hybrid halide per-
ovskites as a competitive PV technology. [7,45,46] The perovskite material has the chemical
formula ABX3, where A is an organic cation (most commonly used is methyl ammonium,
MACH3NH3), B is an inorganic cation (typically Pb or Sn) [47], and X is a halide (typ-
ically iodine, but chloride and bromide have also been used). The interchangeability of
elements within the compound enables bandgap energy (Eg) tuning from 1.6 eV to 3.2
eV, very appropriate for multi-junction designs. At present, perovskites hold a record
power-conversion efficiency similar to CdTe and CIGS, at η = 22.7 ± 0.8%. [35] However,
stabilizing the high performance has proven to be a challenge and is one of the main ongo-
ing focuses of research in the field. [48] Scientists have observed PV characteristics change
in a matter of seconds [39,49] and the primary processes that cause this instability point
towards ion migration. [50] Most perovskites are composed by sub-micron scale grains
14
(Figure 3.2(a)), where spatial variations in their chemical composition resulting from the
spin coating deposition process during the device fabrication are possibly responsible for
the spatial electrical variations reported in the literature, [39,51] as well as ion migration.
Therefore, imaging electronic charge separation and collection at the numerous interfaces
could elucidate the pathways for carrier transport within the material. [52] Full compre-
hension and control of the time-dependent physical and chemical processes responsible for
the commonly observed degradation of perovskites will likely lead to stable, high-efficiency
devices. Additionally, by probing the local electrical response of new lead-free perovskite
alternatives [53] the community will ultimately identify which materials are stable under
PV operation conditions and, thus, suitable for device development, which is currently
unknown.
15
16
V
material E (eV) V (V) V (V) oc,meas J 2 2
Jsc,meas
g oc,theory oc,meas V sc,theory (mA/cm ) Jsc,meas (mA/cm )oc,theory J η (%)sc,theory
perovskite 1.58 1.325 1.105 0.86 27.2 25.0 0.92 22.7 ± 0.8
CdTe 1.45 1.168 0.887 0.76 30.5 30.25 0.99 22.1 ± 0.5
CIGS 1.15 0.89 0.741 0.83 42.1 87.8 0.90 22.6 ± 0.5
CZTS 1.13 0.884 0.513 0.58 43.3 35.2 0.81 12.6 ± 0.3
poly-GaAs 1.42 1.160 0.994 0.89 31.7 23.2 0.73 18.4 ± 0.5
Table 3.1: Theoretical vs Experimental Solar Cell Parameters
3.2.1.2 CdTe
CdTe/CdS heterojunction solar cells are a mature technology, available on today’s
PV market. Thus, improving devices’ performance for a well-established technology can
significantly impact the associated industry. At present, the power-conversion efficiency
record of CdTe solar cells is η = 22.1 ± 0.5% and typical module performance is η =
18.6 ± 0.6%. [4] While the Jsc of world record devices has practically achieved its maxi-
mum, there is a 25% gap between the theoretical and the measured Voc of polycrystalline
CdTe solar cells, see Table 3.1. It has been recently shown that single crystalline CdTe
devices (boundaries free) can achieve Voc as high as 1.096 V, [54] corresponding to ∼20%
voltage improvement over the best values ever obtained for polycrystalline CdTe solar
cells (limited at ∼0.9 V, as shown in Table 3.1). This result strongly indicates that the
grain boundaries do indeed negatively impact the recombination rate of charge carriers.
However, fabricating monocrystalline CdTe thin films is an extremely slow and expensive
process. Therefore, identifying and controlling where carrier recombination events take
place within polycrystalline CdTe will facilitate the design of devices with suppressed non-
radiative recombination, higher Voc and, thus, performance. It has been recently suggested
that not every CdTe GB results in voltage reduction; [37, 55] therefore, the response of
individual grains (with ∼1 µm in diameter, as shown in Figure 3.2(b)) and its interfaces
must be resolved, which requires high spatial resolution electrical measurements.
3.2.1.3 Cu(In(1−x)Gax)Se2 (CIGS)
The extensive characterization of CIGS/CdS heterojunction devices strongly indi-
cates that this material can become a marketable PV technology. Currently, CIGS is the
17
second highest performing thin-film solar cell at η = 22.6± 0.5%. [4] The CIGS Eg can be
tuned from 1.02 to 1.68 eV as a function of Ga/In content (useful for multi-junction ap-
proaches), has high optical absorption, is compatible with flexible substrates, and has the
potential for low-cost/W installation. [56] Nevertheless, the Voc and the Jsc of these devices
can still be improved, as shown in Table 3.1. Note that, recently, the Jsc of small cells (0.51
cm2) reached >41 mA/cm2 (uncertified), [57] approaching the theoretical value. However,
its Voc is still >20% below the theoretical prediction. The p-doped CIGS absorbing layer
is composed of grains with approximately 1 µm in diameter, see Figure 3.2(c). Voltage
variations of >20% have been observed, shown in the next Chapter, despite the uniform
chemical composition of the grains. [40] Therefore, enhanced power conversion efficiencies
will require mapping how the different grains and interfaces composing the CIGS layer
affect charge carrier generation, recombination, and collection. In particular, the possible
correlation between the physical behavior of in operando CIGS devices and its structural
properties (resulting from a distribution of grain orientations) is still unclear. [56]
3.2.1.4 Cu(Zn,Sn)(S,Se)2 (CZTS)
CZTS belongs to a class of material similar to CIGS (see Figure 3.2(d) for grain
morphology). The main difference, as the chemical formula suggests, is that the elemental
compound does not contain possibly scarce elements like In or Ga, and is purposefully
restricted to using earth-abundant elements. Here, the Zn and Sn atoms replace the group
III elements (In and Ga), producing a structure with charge neutral valence state. [58,59]
Although extensive efforts to increase the performance of CZTS solar cells, its Voc is 50%
below the theoretical prediction for this material (Table 3.1). Moreover, improvements in
the Jsc are also still necessary. As a result of its modest Voc, the record device efficiency is
18
η = 12.6±0.3%. [4,60] This lower performance is primarily associated with the fabrication
of films with high void density or impure phases within the Cu2ZnSnS4, such as SnxSy, [61]
despite multiple attempts to optimize the thin film deposition parameters. Determining
how these variations in chemical composition affect the local electrical response of CZTS
thin films could aid in the fabrication of higher efficiency devices.
3.2.1.5 Polycrystalline GaAs
Single crystalline GaAs offers near-ideal optical and electrical properties. As a re-
sult, today’s world record single junction devices are made of monocrystalline GaAs, with
power-conversion efficiency of η = 28.8 ± 0.9% under AM1.5 global illumination. [4] More-
over, this semiconductor compound, together with InP and other III-V alloys, has been
successfully implemented in high-efficiency multi-junction solar cells for concentrator sys-
tems and space applications. [62–64] However, the epitaxial fabrication methods required
for manufacturing monocrystalline III-V semiconductor modules have prohibitive costs
for terrestrial applications. Polycrystalline GaAs is, thus, an attractive alternative to
monocrystalline compounds [44] as they can be fabricated both n- and p-type by close-
spaced vapor transport (CSVT), see Figure 3.2(e). [65, 66] The record cell efficiency for
multicrystalline GaAs is η = 18.4 ± 0.5%, [4] which is significantly below its single crys-
talline counterpart and likely due to either the structural defects in the GaAs polycrystals
or surface recombination within the distinct facets. By spatially resolving the Jsc and Voc
of the GaAs mesoscale constructs we can measure the local response of each crystal facet
independently. For instance, it is well known that the GaAs substrate orientation affects
the density of defects. [67,68] Yet, an analysis of the contribution of the individual GaAs
polycrystals to charge carrier recombination and collection and, therefore, to the overall
19
device performance is critical and still missing.
From Table 3.1, it is evident that there is an opportunity for material scientists, elec-
trical engineers, physicists, and chemists to advance the state-of-knowledge of perovskites
and polycrystalline materials for PV. As the mesoscale constructs strongly influence the
efficiency of charge carrier generation, recombination, and collection events, it is imper-
ative to resolve the relevant performance properties of these materials/devices at similar
length scales.
3.2.2 Probing electrical properties
Understanding the nanoscale electrical characteristics of solar cells is the next step
towards realizing devices with controlled response and, thus, enhanced performance. In
AFM based techniques, the electrical signal is the main diagnostic tool used to indicate
whether the local voltage and/or current of a PV device is affected by the materials mor-
phology and how its spatial variation compares with the average, macroscopic device re-
sponse. Nanoscale electrical characterization can help answer critical, still not fully solved
questions such as: “How do the aggregates in perovskite thin films affect overall device
morphology and performance?”, [69] “Does the lifetime of charge carriers vary as a func-
tion of incident wavelength in organic PV?”, [70] and “What is the specific contribution of
grain boundaries in CIGS and CdTe solar cells′ to the local electrical signal?”. [37,71] we
review the most relevant AFM methods that have made contributions towards revealing
the effect of the light-absorbing microstructure on the electronic response.
20
Electrical properties imaging
perovskite polymer CdTe
(a) KPFM (c) trEFM (d) cAFM
3 -23
topography 0 V bias -2 V bias
1 µm
-124 90 106 176 -33 57
nm mV mV 0 -26
6 -0.25
topography 0 V bias -2 V bias
-81 67 -30 61 -22 64
nm mV mV Distance / µm 0 -1.55
(b) EFM (e) SCM
topography EFM: -2 V bias EFM: -4 V bias Differential map
400 nm 400 nm
0 60 0 4 0 3
Height (nm) (dU/dz)/dV (a.u.) (dU/dz)/dV (a.u.)
CIGS solar cell
Organic material
Figure 3.3: Nanoimaging electrical properties by A(aF) KimM Y.C. et a(l.a, A)dv. ETngo. Mpat.o 20g16r DaOIp: 10h.1y002/aenmd.201502104 
(b) Araki, K. Jap. J. of Appl. Phys. 2016 DOI: 10.7567/JJAP.55.070305 
KPFM of two perovskite solar cells, with and without (tc)h Coex, Pi.An. ect aol., rJ. pPhyos. rChaemt.i Loettn. 201o5 fDO5I:10..71021m/acos.jplc%lett.5b01360 
of PbI as function of voltage bias. Illumination condition(d)s Tu:tej5a, 0M.0 et anl., mJ. Phyis.n Chcemi.d Leett.n 20t16 l1i0g.10h21t/acas.jpncledtt.6b023992
(e) Koji, M. et al., Meas. Sci. Technol. 2014 DOI: 10.1088/0957-0233/25/4/044020
photon flux = 0.3 kW/m2. Adapted with permission from ref [69]. ©c 2015 WILEY-
VCH Verlag GmbH & Co. KGaA, Weinheim. (b) AFM and corresponding EFM
images of a MDMO-PPV:PCBM (1:4) polymer blend thin film acquired at different
applied biases. Far right: differential map of the two biased EFM maps. Adapted
from ref [72]. Copyright 2016 The Japanese Journal of Applied Physics. (c) Top:
trEFM map overlaid with topography in a polymer half-processed solar cell in a 1.5
× 1.5 µm2 region. Bottom: Profile of topography and relative charging rate acquired
for two different photon excitations. Adapted from ref [70]. (d) cAFM maps of the
local current signal for both non-passivated and passivated CdTe solar cells in a 10
× 10 µm2 area. Adapted from ref [71]. (e) SCM on the cross-section of a CIGS solar
cell defines the location of the p-n junction and qualitatively shows the density of
dopants in each respective layer. Adapted with permission from ref [73]. ©c 2014 IOP
Publishing Ltd.
A variant of AFM that measures local voltage with nanoscale spatial resolution is
called Kelvin probe force microscopy (KPFM), which we have recently demonstrated to
map the open-circuit voltage in solar cells [37,39] (a detailed discussion is presented in the
next section and in Chapter 5). KPFM measures the work function difference between a
conductive probe and the sample surface, termed the contact potential difference (VCPD).
KPFM was theorized by Lord Kelvin in 1898 [74] and first implemented a couple of decades
ago to characterize the local electrical signal of metal and semiconductor surfaces. [75,76]
21
With PbI2 No PbI2
Height / nm
CdCl2 treatment as-deposited
pA nA
nA nA
Normalized Charging Rate Normalized
Charging Rate
The work function (W ) is the energy level difference between the vacuum and Fermi level
(EF ). The probe, typically Si coated with a thin conductive metal layer, is assumed to have
a constant W in non-contact scanning mode. In equilibrium (no light), the solar cell’s W
at the surface stays constant; however, in the case of non-uniform, heterogenous materials
the work function may vary locally depending on grain orientation, [77] EF pinning, [78,79]
charge trapping, [80] and chemical composition, among other factors. When the sample
is illuminated, it generates charge carriers and is, thus, in a non-equilibrium state.
KPFM is frequently used as a diagnostic tool to image the voltage response before
and after specific deposition steps, relevant for optimizing the fabrication parameters of the
films. In Figure 3.3(a), Seidel’s group compared how the electrical signal of (FAPbI3)0.85
(MAPbBr3)0.15 (FA = formamidinium, MA = methylammonium) perovskite compounds
with and without an additional (5.7 mol%) PbI2 reacts locally when an external voltage
bias (inducing ion migration) is applied, by KPFM. [69] It was found that the VCPD of
the sample without the extra treatment step presents a dramatic change as a function of
bias, particularly at the grain boundaries, while the VCPD of the device with excess PbI2
remains constant at both the grain interiors and grain boundaries. These observations
were attributed to ionic migration suppression due to the incorporation of excess PbI2,
which in turn improves the overall electrical properties of the solar cell.
A derivative of KPFM, called either intensity-modulated or illuminated-assisted
KPFM, measures the decay of the surface photovoltage (SPV), which is the subtraction
between an illuminated- and a dark- KPFM scan in the same location. Here, the VCPD
depends on the frequency of the laser pulse, and a transient SPV signal resolved on the
order of sub-milliseconds is observed. [81, 82] The decay rate of the SPV is related to the
minority charge carrier lifetime, as shown in refs [83, 84]. Briefly, when the illumination
22
source is in OFF mode, the PV voltage decays exponentially and is proportional to e−1/τ ,
where τ is the minority carrier lifetime. These pioneer works yield insights that link the
local photovoltage, lifetime, and topography, especially useful for correlating the role of
the microstructure to a specific type of recombination mechanism.
Electrostatic force microscopy (EFM), of the same family as KPFM, also measures
the electrical signal of the tip-sample interaction while applying a voltage to the tip in
non-contact mode. The deflection of the AFM probe is linked to the electrostatic force
(FES) between the sample surface and tip. [85, 86] That is, the oscillation path (z -axis)
of the cantilever changes depending on the electrical charge distribution located at the
sample surface. The electrostatic force is described as:
1 dC
FES = V
2 (3.1)
2 dz
where z is the distance between the tip and sample, C is the tip-to-sample capacitance,
and V is a voltage term that encompasses all voltages flowing within the AFM experimen-
tal setup. The amount of tip deflection is related to the charge carrier density; therefore,
by utilizing EFM one can map electrical properties such as capacitance, surface voltages,
or electrostatic potentials. [87] The capability of this technique is exemplified in Fig-
ure 3.3(b), where the surface charge of a poly[2-methoxy-5-(3A,7A-dimethyloctyloxy)-1,4-
phenylenevinylene] (MDMO-PPV):phenyl-C61-butyric-acid-methyl ester (PCBM) poly-
mer blend thin film is mapped as a function of external voltage. [72] Here, Takuya’s group
was able to determine how the electrostatic properties of an organic PV system influence
device performance. Both the permanent (addition of two EFM maps set at different
voltage bias values) and field-induced (differential of two biased EFM maps) charge com-
ponents were acquired. While the permanent portion of the EFM signal did not affect
23
performance, the differential map shows sharp, dark contrasts, which could be related to
charge trapping.
Time-resolved EFM (trEFM), an extension of traditional EFM, measures the charg-
ing rate by tracking the resonance frequency shift that occurs between the probe and the
sample due to charge accumulation. [88,89] Under illumination or external voltage bias a
solar cell generates electrons and holes at a specific rate. A faster charging rate implies
an increase in charge buildup under the probe, shifting the cantilever frequency further
away from its equilibrium resonance peak, dependent on both the incident photon energy
and quantum efficiency. Each of these quantities are known for varying spatially in or-
ganic heterogenous PV materials. Figure 3.3(c) shows a 3D topography map overlaid with
trEFM data on a MDMO-PPV:PCBM (wt/wt 1:4) organic blend cast from toluene. [70]
The black curve denotes topography, while the green and red ones are the normalized
charging rates under 532 nm (above bandgap) and 850 nm (below bandgap) illumination,
respectively. The trEFM line scans present similar spatial trends and, thus, are indepen-
dent of the incident wavelength. This result suggests that the internal quantum efficiency
of the polymer material is equal across a wide range of photon energies. Ginger’s group
observed that segregated regions of PCBM have lower charging rates than if the film were
to have a completely uniform blend.
One way to measure the current of a PV device at the nanoscale is to probe the
sample using conductive- or photoconductive-AFM (cAFM or pcAFM). Here, a conductive
AFM probe in contact with the sample surface acts as a local electrode that collects the
solar cells dark (cAFM) and/or light-generated current (pcAFM). This mature method
has been widely used to image a variety of materials for solar cells, such as perovskites,
[51, 90–92] thin-film polcrystalline, [71, 93, 94] and others. [95–99] One pressing question
24
that cAFM can help answer is: What is the specific role of the CdCl2 treatment on
the electrical properties of CdTe solar cells? It has been ubiquitously demonstrated, at
a macroscopic level, that the CdCl2 improves PV performance by passivating the CdTe
grains and interfaces; [100] yet, resolving how it affects the local electrical signal is still not
well understood. Thus, cAFM has been applied to elucidate how this extra fabrication
step enhances CdTe performance — see in 3.3(d). [71] Here, Rockett’s group imaged two
CdTe solar cells, with and without CdCl2 treatment, under multiple external voltages
using a Pt-coated Si probe. Both the top-left and bottom-right maps were obtained at a
bias of +6 V while the top-right and bottom-left were acquired at -6 V. Under positive
bias the p-n junction of the CdTe solar cell experiences reverse breakdown voltage due to
the geometry of the measurement. In this case, the current contrast between the grains
and grain boundaries in Figure 3.3(d) is accentuated for the CdCl2 treated sample, caused
by injection of electrons into the grain boundaries from the probe. This indicates that
the grain boundaries are less p-type than the grain interiors in the treated sample. While
applying a negative voltage, the sample undergoes forward bias and the current flow in
this regime is due to electron injection from the p-n junction into the CdTe layer. The
magnitude of the current under negative applied bias is limited by the electron collection of
the AFM tip and/or by the recombination of charge carriers at the back Schottky contact.
While this technique is a great approach to directly map the current generated by a solar
cell with nanoscale spatial resolution, much care should be taken when acquiring images,
as the thin conductive coating of the probe can be easily removed, producing artifacts in
the electrical signal.
Scanning capacitance microscopy (SCM) maps the differential capacitance (δC/δV),
enabling dopant concentration imaging in microelectronic systems such as solar cells.
25
Briefly, the capacitance of a semiconductor can be tuned by applying an alternating elec-
tric field to a conductive-coated tip, and is subsequently measured by tracking the third
harmonic frequency of the cantilever. [101] The type of doping (n or p) can be deduced
based on the sign of δC/δV, while the magnitude of the capacitance is based on carrier
concentration. [102] Further, there has been some research effort to quantitatively link
δC/δV to charge carrier concentration in different types of semiconductors materials, such
as Si, GaN, and others. [103] Daisnke’s group imaged a cross-sectional SCM map of a CIGS
solar cell, Figure 3.3(e), showing the qualitative carrier concentration at the CdS/CIGS
n-p junction. [73] Both the accumulation and depletion regions were directly mapped, im-
portant for determining where the electrons and holes are generated throughout the PV
films.
3.2.2.1 Attributing electrical properties to performance
While standard electrical property mapping is an extremely informative way to gain
fundamental insights about PV materials and devices, quantifying and relating nanoscale
charge carrier phenomena back to device performance is a critical next-step that will
lead to the rational design of solar cells with high power-conversion efficiencies. There
has been novel research that links the acquired electrical output to a relevant solar cell
parameter, namely photovoltage and photocurrent. We recently modified illuminated-
KPFM to quantify the local open-circuit voltage (Voc), [37] one of the most important
figures-of-merit that define PV performance, discussed in depth in Chapter 5. For this,
the AFM probe must be grounded with respect to the bottom electrode of the solar cell
and the experimental setup calibrated. In this way, when both a dark and an illuminated
KPFM map are acquired on the same region of a sample, and subsequently subtracted,
26
the remaining value is the quasi-Fermi level (∆µ). As seen in Equation 3.2, ∆µ is directly
proportional to the Voc of the PV device:
∆µ = Voc × q (3.2)
where, q is the elementary charge. Figure 3.4(a) shows a Voc map containing two grain
boundaries in a CdTe solar cell with distinct and contrasting electrical behavior. The
defect highlighted on the left locally reduces the Voc by 60 mV while the grain boundary on
the right acts similar to the grain interiors, maintaining a voltage response comparable to
its surrounding grains. Here, we determined the quantitative contribution of CdTe grains
and grain boundaries to the local Voc. We also revealed local voltage variations in CIGS
solar cells as high as 250 mV, demonstrating the large influence that the microstructure can
have on the recombination of charge carriers, consistent with other KPFM measurements
on CIGS. [104] Overall it was demonstrated that this powerful nanoimaging approach is
a universal tool to directly map the local Voc of mono-, polycrystalline, and perovskite
solar cell devices, and works in ambient environment. [37, 39] For more information see
Chapters 5 and 6.
27
Performance imaging
(a) Topography Dark KPFM (b) Current tomography
1.3 10
1.29
500nm -10
0 2 μm
Voc Light KPFM Experimental schematic
 0.58 
1.14
  
0.47
Figure 3.4: Photovoltaic performance at the nanoscale. (a) Illuminated-KPFM
on a CdTe solar cell results in direct imaging of the local Voc by subtracting an
illuminated scan by a dark one. Adapted with permission from ref [37]. ©c 2015
WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. (b) A variant of cAFM,
(a) Tennyson, E.M. et al. Adv. Energyc Monatderu. 2c0t1i5n dgoi: t1o0.m10o02g/areanpmh.2i0c15A70F12M3 , biased near Voc. A tough diamond-coated probe was
(b) Luria, J. et al. Nature Energy 2016u dsoei:1d0.1t0o38/entecnehrgya.2w01a6y.150the top surface layer after each current map of a CdTe solar
cell. Adapted by permission from Macmillan Publishers Ltd: Nature Energy [93],
copyright 2016.
AFM-based techniques are often limited to imaging properties at the sample surface;
however, a method recently developed, conducting tomographic AFM, maps the photocur-
rent in 3D, demonstrated in Figure 3.4(b). [93] In pcAFM mode, a diamond-coated probe
with a high toughness value, iteratively etches away the top few nanometers of the surface
to reveal CdTe grains and their respective photo-generated current. While this technique
is destructive in the scanned region, charge carrier transport pathways previously hidden,
are measured. Using this method, Huey’s group identified inter- and intra-grain planar
defects attributed to local wurtzite and zincblende structure variations, consistent with
stacking faults imaged with transmission electron microscopy (TEM) of the same sample.
Further, by scanning the region biased, grain boundaries with n-type characteristics were
resolved.
28
V µm
pA
V
3.2.3 Imaging local chemical properties
Very recently, researchers have extended AFM to map the chemical information of
materials for PV at the nanoscale. [105] Now, more than ever, identifying the elemental
distribution with nanoscale spatial resolution is relevant for heterogenous, next-generation
PV technologies to improve the state-of-knowledge of crucial unanswered questions, such
as: “Where do the Br3, Cl3, or I3 ions segregate in mixed perovskite materials upon
illumination?”, [106,107] “Is the non-uniform electrical response of perovskites related to
variations in chemical composition?”, [39,51] “Is it possible to chemically resolve where the
different species in a block co-polymer thin film preferentially accumulates/segregates?”,
[108] “What is the primary substance limiting the electrical conductivity in polymer solar
cells?”, [109] and “Are the majority of grain boundaries in CZTS thin-film solar cells of
Cu-rich or Cu-poor composition?”. [59]
Tip-enhanced Raman spectroscopy (TERS) is an effective method to image the
chemical composition of materials. [110–112] The experimental setup for TERS requires
a high-intensity polarized light focused on the apex of a metal-coated AFM probe (e.g.
Ag or Au) held ∼20 nm above the sample surface. The interaction between the inci-
dent far-field photons and the tip induces a large electromagnetic near-field enhancement,
capable of exciting Raman vibration modes. The amplification effect is attributed to local-
ized surface plasmons that are generated by the incident photon energy, which is usually
near the resonant peak of the metal coating on the AFM cantilever. [110, 113, 114] The
strong excited optical field is coupled into a spectrometer, where each pixel in a TERS
map corresponds to a spectrum. TERS is particularly useful when analyzing systems
with inhomogeneous chemical composition distribution, such as perovskite and organic
29
solar cells, as the agglomeration or separation of elemental compounds can deteriorate
the electrical performance of the devices. Figure 3.5(a) shows the topography (top) and
corresponding TERS image (bottom) of a PCBM (acceptor) and poly-(3-hexylthiophene)
(P3HT, donor) polymer solar cell. [109] Meixner′s group attributes the bright red feature
in the center of the chemical map to a PCBM aggregate that formed on the surface of
the polymer due to an unoptimized annealing step. PCBM was identified by both the
Raman and photoluminescence (PL) spectra in this region, and the spatial distribution
of the chemical trace was only discernable in the TERS imaging. Confocal microscopy
was attempted in the same location and no change in the Raman signal was detected.
This PCBM aggregate is responsible for the limited electron transfer observed in these
polymeric solar cells, resolved by TERS measurements.
30
Chemical property imaging
Polymer Polymer
(a) TERS (b) PiFM 1492 cm-1 1733 cm-1
125 nm
1.79 2.86 0.79 1.76
0 600 Wavenumber (cm-1) PiFM signal (a.u.) PiFM signal (a.u.)
nm
(c) PTIR 0.7 250
AFM PTIR
410 543
kcounts/0.5s 0 0
Polymer – catalyst energy harvester
Figure 3.5: Nanoimaging chemical properties in solar cell materials by
AFM. (a) Topography (top) and near-field TERS map (bottom) of a polymer
material, wavenumber range from 283 to 2926 cm−1 was integrated. Region III
indicates the agglomeration of PCBM molecules. Adapted with permissions from
ref [109]. ©c 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. (b) PiFM
(a) Wang, X. et al., Adv. Func. Mater. 2010 DOI: 10.1002/adfm.200901930
images acquired at two distinct wavenumber excitations show(bi)n Nogwakt, hD. eet atl.,w Scoi. Adpv. h20a16s DeOsI: 10o.1f126/sciadv.1501571
a co-block polymer material. Reprinted/adapted from [107]. ©c(c) TGhhoshe, S.A et aul.,t Nhato Mratser, 20s1o5 DmOI:e 10.1038/NMAT4220
rights reserved; exclusive licensee American Association for the Advancement of Sci-
ence. Distributed under a Creative Commons Attribution NonCommercial License
4.0 (CC BY-NC) http://creativecommons.org/licenses/by-nc/4.0/. (c) Topography
and PTIR absorption map acquired at 3054 cm−1 of a PDPB nanofiber. Correspond-
ing PTIR absorption spectrum of the C-H stretching modes both before and after
catalysis. Adapted by permission from Macmillan Publishers Ltd: Nature Materi-
als [115], copyright 2015.
Photo-induced force microscopy (PiFM) is an emerging approach that was developed
in 2010. [116] Here, the near-field Raman signal is detected in noncontact mode. Thus,
changes in the time averaged force gradient of the tip-sample interaction are measured,
rather than a spectrum. [117] During the experiment one (tunable) or two lasers (of
different wavelengths) are focused onto the sample; they modulate at frequencies that
are selected based on the first and/or second resonance of the cantilever. The external
excitation generates a unique response in the force curve that is captured by a lock-in
amplifier. [118] In general, the experimental setup is compatible with bottom-illumination
with a high numerical aperture objective lens. The spatial resolution of the PiFM image
is slightly better than the simultaneously acquired topography map because the PiFM is
31
PiFM signal (a.u.)
Absorption (a.u.)
µm
both excited and detected in the near-field regime. The signal-to-noise ratio of PiFM is
substantially larger than the one of TERS because the latter is detected in the far-field
regime. These advantages extend chemical imaging spatial resolution, which can be used
to reveal nanoscale domains within composites, or to determine the specific segregation
regions in mixed perovskites. [106, 119] As a proof of concept and to demonstrate the
power of this technique, Park′s group imaged a block co-polymer material, poly(styrene-b-
methyl methacrylate) (PS-b-PMMA), with very well defined Raman molecular signatures
and identified each chemical compound with <10 nm resolution, see Figure 3.5(b). [108]
A point-by-point PiFM response spectrum is shown on the left, executed by sweeping
a quantum cascade laser at the given wavenumber and recording the magnitude of the
cantilevers deflection. On the right side of Figure 3.5(b), two PiFM images in the same
location are acquired by consecutively exciting the sample at wavenumbers 1492 cm−1
(for PS), and 1733 cm−1 (for PMMA), where both polymers are clearly resolved. Regions
of PS appear brighter in the map using the 1492 cm−1 excitation because its near-field
Raman signal is stimulating the cantilever at its resonance frequency, detected by lock-in
amplification. Note that it is necessary to know the Raman spectrum of the sample under
investigation before performing PiFM to ensure appropriate selection of the excitation
wavenumber. PiFM is, thus, a very promising AFM-based tool to probe materials for PV
ranging from perovskites to CIGS, where local variations in chemical composition might
affect carrier recombination and collection.
Photothermal induced resonance (PTIR) microscopy, or AFM-IR, combines precise
IR spectroscopy with the spatial resolution of an AFM. [120,121] Here, an external, tunable
IR laser is pulsed while the AFM cantilever is in contact mode. This photon excitation
induces local heating of the sample faster than the acquisition of the AFM feedback
32
and proportional to the thermal expansion coefficient of the material. As each chemical
species has a unique thermal expansion coefficient, chemical identification is attainable at
the nanoscale. [122,123] In most cases, devices or thin films under investigation are placed
on top of a glass prism, where the incident light is reflected up to the sample by total
internal reflection. [124] This method has been applied to resolve ferro-elastic domains
[125] and the distribution of Cl as a function of wavelength and annealing time, [107]
in perovskites. In Figure 3.5(c), both topography and PTIR maps are presented on a
UV- and visible-light photocatalytic energy harvesting material, poly(diphenylbutadiyne)
(PDPB). [115] Here, the conductive polymer was fabricated as a nanofiber, because the
photocatalytic reaction of the one-dimensional design was more efficient than its bulk
counterpart. One major challenge with current light harvesters is their long-term stability,
thus, Remita′s group investigated PDPB, a promising material that does not require any
precious metal nanoparticle co-catalyst. Nanoscale PTIR measurements demonstrated the
polymer reusability, as any composition degradation would be visualized in the absorption
map. Further, the spectra on the right side of Figure 3.5(c) show the reproducible C-
H stretching vibration modes at 3054 cm−1 both before and after undergoing catalysis.
The overlap in the spectra indicates that there was no change in chemical composition,
indicative of a stable, photocatalytic material. This chemical mapping technique could
also extend to metallic nanostructures for energy harvesting, [126] not possible with TERS
or PiFM, as many metals (e.g. Ag and Au) do not have a clear Raman response.
3.2.4 Optical properties at the nanoscale
Combining the information gained from a traditional optical microscope with the
spatial resolution of an AFM yields a highly impactful imaging approach that can reveal
33
optical phenomena (i.e. reflection, absorption, and transmission), and electrical properties
of materials for energy harvesting. Acquiring an optical response at the nanoscale requires
a small antenna (or probe) that either excites and/or collects the near-field optical signal,
entitled near-field scanning optical microscopy (NSOM, or SNOM). Here, the optical infor-
mation obtained is beyond the diffraction limit of light, restricted instead to the diameter
of the NSOM probe. [127] Through NSOM measurements scientists are beginning to an-
swer questions such as: “Does the migration of MA+ and I− ions in perovskites impact the
index of refraction of the material?”, [128] “How can one efficiently couple far- to near-field
light in order to image radiative recombination in nanoscale structures of optoelectronic
devices?”, [129] “What is the spectral dependence on the local photocurrent enhancement
in CdTe and CZTS grain boundaries?”, [130, 131] and “What is the correlation between
NSOM mapping (optical properties) v3
the facets of GaAs polycrystals and their carrier radiative emission rate?”. [44,66]
InP NW a-Si solar cell
(a) Campanile Confocal (b) (c)
3 2
0 0
(d) pristine positive poling negative poling
(apertureless) perovskite
Figure 3.6: Nanoimaging optical properties in solar cell materials by AFM.
Clockwise: (a) PL maps acquired with a near-field campanile probe and cCodnTfeocal ob-
jective, respectively, of an InP nanowire in a 2 × 2 µm2 region. From [129]. Reprinted
with permission from AAAS. (b) Illustration of NSOM in collection mode utilizing
an aluminum-coated glass fiber tip to compare the near-fi(ea) lBado, Wo. etpal.t, Siccienacel201r2eDOsI: 10.1126/scien(b) Paetzold, U.W. et al., Nano Lett 201o4 DnOI:a 10n.10c21e
ce.12o27f977/nl503249n
nanopatterned region vs. a flat region on an a-Si device. [132(c]) Le(itce, )M.SE. etffal.,e ACcS tNainvo 2e014EDOQI: 10E.1021m/nn5a05p2585
(d) Zhang, Y et al., ACS Nano 2016 DOI: 10.1021/acsnano.6b03104
measured at 850 nm (top) and average spectrally-dependent EQE graph (bottom)
calculated from NSOM-LBIC images of a CdTe solar cell. Adapted from ref [130].
(d) Left: schematic of s-NSOM. Right: s-NSOM measurements on a perovskite mi-
crowire under different poling conditions, used to measure ion migration; scale bar
is 500 nm. In the s-NSOM amplitude maps: red = high signal, blue = low signal.
Adapted from ref [128].
34
kCs
In near-field optics the distance between the illumination/collection source (the
NSOM probe) and the sample surface is significantly less than the wavelength of light
(∼10 nm). The theoretical description of the near-field effect was introduced in the 1920’s
[133] but was realized more than half a century later. [134, 135] The major experimental
innovations that led to the successful implementation of NSOM are: (i) depositing a thin
metal coating over the NSOM probe to help guide light into the nanoscale aperture, and
(ii) including a feedback loop to continuously monitor/adjust the tip-sample distance.
NSOM can be performed in multiple configurations, [136] ranging from excitation-only to
reflection+collection. Below, we review the two most useful NSOM modes for solar cells:
(i) collection and/or excitation and (ii) apertureless, where there is a discussion about
how these methods can be extended to determine relevant information about PV device
performance.
In NSOM-PL, also detailed in Chapter 8, the probe acts as both a collection and
an excitation source. For this, incident light passes through the small hole in the NSOM
probe and the subsequent photon emission caused by the radiative recombination of charge
carriers in the solar cell is collected back through the same aperture. One example of high-
efficiency, bidirectional far-to-near field coupling is presented in Figure 3.6(a), [129] where
Weber-Bargioni’s group employed a new probe design shaped similar to a campanile bell
tower, to promote a more effective waveguide for far-to-near field coupling in a wide spec-
tral range. In Figure 3.6(a), NSOM and confocal PL maps of an InP nanowire (NW) are
compared; each pixel represents the amplitude of the emission signal at 802 nm. It is
immediately apparent that the image acquired with the campanile probe has much higher
spatial resolution, elucidating the spatial distribution of PL within a single NW. The
non-uniform emission is associated to trap-state modifications of the localized excitons,
35
an important discovery if InP NWs are to be implemented into light-harvesting devices.
Since this study, there have been more NSOM-PL investigations [137] of the optoelectron-
ics properties of solar cell materials including perovskites [138] and InGaN/GaN, [139]
each presenting novel insights about the local distribution of radiative and non-radiative
recombination mechanisms.
NSOM using an aperture in the collection-only configuration illuminates a large
region of the sample and gathers the evanescent fields of photons propagating through
the surface of the device. The NSOM probe is maintained at a distance of ∼20 nm above
the sample, close enough to allow tunneling of the guided photons. In the example shown
in Figure 3.6(b), [132] Paetzold et al. have used NSOM in collection mode to resolve
the near-field enhancement induced by patterned nanostructures within a solar cell as a
function of both wavelength and light polarization. [132] Their results complement the
macroscopic measurements of external quantum efficiency (EQE) improvement due to the
addition of metallic nanostructures. [140]
One application of NSOM in excitation-only mode is light beam induced current
(LBIC) microscopy, where the photogenerated current can be spatially resolved at the
nanoscale. [141] A spot size equivalent to the diameter of the aperture illuminates a solar
cell and locally generates a flow of charge carriers. The photocurrent signal is typically
detected by using the macroscopic contacts of the device. Resolving the light-generated
current at the nanoscale is relevant when the feature size that composes a solar cell material
is of a similar dimension, e.g. grain boundaries. [131] In Figure 3.6(c) spectrally-dependent
NSOM-LBIC is applied to a CdTe solar cell. [130] Nanoscale spatial variations of the EQE
(calculated from photocurrent measurements) at grain boundaries (GB) vs. grain cores
(GC) were measured. Overall, Leite et al. found that GBs generate more photocurrent,
36
hypothesized to be a local reduction in bandgap. One advantage of this optical technique
over pcAFM is that it is performed in non-contact mode. Thus, contact artifacts from the
tip are reduced. However, the signal-to-noise ratio is substantially lower in NSOM-LBIC,
and more sensitive current detectors are required.
Apertureless NSOM or scattering-NSOM (s-NSOM) follows similar operating prin-
ciples as collection-only mode, i.e. excitation uses far-field light and the optical response
from the sample is collected in the near-field. [142] However, here, the probe is not hol-
low, but instead a sharp, metallic tip (most cases an AFM probe) kept <20 nm from the
illumination spot. The apex enhances and disperses the near-field interaction induced by
the incident light occurring at the sample surface, the signal is then acquired with far-
field detectors. To further increase the near-field enhancement effect, it is suggested to
use an AFM probe that is coated with a thin layer of Au. [143] Because in s-NSOM the
tip no longer wields an aperture, the spatial resolution is equivalent to standard AFM.
As previously mentioned, ion migration in perovskite materials is hypothesized to in-
duce instabilities and degradation in the overall electrical performance. In Figure 3.6(d),
Bao’s group implemented apertureless NSOM to nanoimage, in situ, how the electric-field
poling direction of a perovskite device influences MA+ and I− ion migration. [128] A mi-
crowire is used to isolate and confine the direction of the ions. Near-field optical maps
were obtained under mid-IR (11.7 µm) illumination to map the index or refraction and
free-charge carrier distribution. On the left of Figure 3.6(d) an illustration of the exper-
imental setup is shown. The top portion of the right three images displays a schematic
of the observed phenomenon occurring in the microwire under pristine (no applied volt-
age), positive poling (+10 V), and negative poling (-10 V) conditions. The topography,
near-field optical amplitude, and phase images under these three conditions are presented
37
for comparison, respectively. According to the topography maps, the perovskite material
expands in the poling direction, which is also in the same orientation as the migration of
the MA+ ions. The authors calculate that the material expansion is 4.4%; the measured
shrinking-swelling effect in the perovskite structure could be one source of device degra-
dation. Further, the near-field amplitude (here, representing the local carrier density)
response is uniform under pristine conditions, but during poling the distribution of charge
carrier changes: MA+ ions accumulate near the negative side and I− shifts toward the
positive electrode, possibly generating a temporary p-i-n junction. In the near-field phase
maps (related to the perovskites refractive index), the negative electrode has a much larger
signal, indicating a decrease in absorption due to the swelling of the perovskite. Although
very informative, this microscopy method suffers from high background noise due to the
far-field excitation. Therefore, double lock-in or other background-suppression methods
should be implemented to extract the near-field signal. [136]
3.3 Probing dynamic processes in perovskite solar cells
Due to the high interest in perovskites for energy harvesting, this Chapter section
is dedicated to cutting-edge research that investigates the electrical, chemical, and/or
optical properties of these intriguing materials with nanoscale spatial resolution. With
promising, single-junction power conversion efficiencies >22% [4] and a new record tandem
perovskite/Si solar cell of 23.9%, [144] the prospect for commercializing this PV technology
is approaching fruition. Moreover, there has been a recent article by Nazeeruddin’s group
announcing a one-year stable perovskite solar cell, [145] indicating that our ability to
optimize and control these complex materials is increasing. Yet, there are still many
challenges that lie ahead and we are continuing to discover answers to important scientific
38
questions such as: “Can we capture the real-time effects of light illumination on the local
Voc of perovskites?”, [39] “How do the radiative recombination distributions of charge
carriers in MAPbI3 vs. MAPbBr3 perovskite compounds differ?”, [138] “Are Cs-containing
triple-cation perovskites more stable than conventional ones?”, [146, 147] “What is the
role of humidity on the perovskite local electrical response?”, [50,148] and “What are the
perovskite material options that are stable upon illumination?”. [149,150]
Perovskite solar cells experience performance instabilities caused by transient pro-
cesses upon material illumination, [151] and exposure to humidity and oxygen. [152, 153]
Imaging these dynamical events at relevant length scales have become an important next-
step towards identifying stable materials and devices. Rapid temporal mapping will likely
elucidate the transport of charge carriers after excitation, spatially identifying where trap-
states are located. In Figure 3.7(a), our group measured the real-time, photo-induced
voltage response of a MAPbI3 perovskite solar cell before, during, and after illumination
with fast-KPFM. [39] More details will be discussed in Chapter 6, but briefly, each KPFM
map (left of image) is acquired in 16 seconds (two orders of magnitude faster than con-
ventional methods), therefore, it is possible to capture how the electrical signal changes
as a function of time. Immediately after the light is turned OFF the photo-generated
voltage ceases. However, the charge distribution is very different from the initial dark
scan, as seen by comparing both the KPFM map and the voltage histograms (right side
of panel). In fact, the perovskite’s electrical performance remains out of equilibrium for
nine minutes, until finally stabilizing and returning to its initial dark conditions. While
perovskite performance instability on the time scale of several minutes has been observed
macroscopically, [49] by imaging with nanoscale spatial resolution, two distinct regions
with contrasting voltage decay rates were found to coexist within a single MAPbI3 grain,
39
resulting from ion migration. [39] The next step is to identify the relationship between
this dynamic electrical response with local chemical composition variations. This point
will be addressed in more detail inPtehreovfsuktiutere dwynoarkmpicosr _ti vo2n of this thesis.
(a) KPFM Histogram (b) Topography as-prepared
1.2
95oC for 5 min 95oC for 5 min 
+ 110oC for 10 min
0
(c) Topography NSOM PL Topography NSOM PL
MAPbBr3 MAPbI3
0 53 nm 0.6 26.6 0 465 nm 0.2 2.6
Figure 3.7: Imaging electrical, chemical, and optical dynamics in perovskite
(a) Garrett, J.L.s, eot al. aNarno Lcette. 2l0l1s7 dobi: 1y0.102A1/aFcs.nManole.tt.7(ba00)289S equence of fast KPFM on a perovskite solar cell under
(b) Chae, J. et adl., aNarnko Leatt n20d15 DiOlIl: 1u0.m102i1n/acas.ntaenodlettc.5ob0n35d56itions; each map takes 16 seconds. Right graph shows volt-
(c) Vrucinic, M. et al., Adv. Science 2015 doi: 10.1002/advs.201500136
age histograms for 42 fast-KPFM maps. [39] (b) Topography and three PTIR maps
measured at different annealing temperatures of a thin-film perovskite, where the
blue regions of the map refer to high Cl concentration; scale bar is 1 µm. Adapted
from [107]. (c) NSOM topography and simultaneous PL images of two perovskite
compositions, MAPbBr3 and MAPbI3, scale bar = 4 µm. Reproduced with permis-
sion. [138] Copyright the authors, published under CC-BY 4.0 license 2015.
Mapping the local elemental distribution in perovskites is of major interest for the
PV community to help determine where ions are segregating and/or migrating. For ex-
ample, in Figure 3.7(b), a MAPbI3−xClx sample is annealed in situ and the AFM (top
left) and PTIR signals are acquired during each thermal step. By identifying the PTIR
proportion between high and low incident photon energies, local changes in the perovskite
bandgap are distinguished. It is well understood that the addition of Cl increases the
bandgap energy of the MAPbI3, therefore, regions of high Cl concentration are discerned
40
PTIR Intensity Ratio
(blue in the PTIR maps, red indicates Cl-poor areas). Local increases in bandgap in the
as-prepared sample as high as 0.3 eV were measured. However, upon further annealing the
bandgap continuously decreases, indicating that Cl content is reduced with temperature,
likely caused by MACl sublimation. Additional spectrally-dependent PTIR imaging with
high spatial resolution on the wide variety of perovskite materials would be extremely in-
formative, particularly to determine the role of absorption of different elements causing the
instabilities of solar cell efficiency. Recently, a novel AFM-based chemical imaging method
called peak-force infrared microscopy mapped the distribution of methylammonium in a
MAPbBr3 nanocrystal. [154]
To image the radiative recombination centers in perovskite materials, Deschler’s
group implemented NSOM-PL on both MAPbI3 and MAPbBr3 thin films. Figure 3.7(c)
displays the simultaneously acquired topography and PL maps for each sample. The
sample containing Br3 has a more homogenous morphology, while the I3 film has needle-
like features with 25% of the scanned region comprised of voids. The PL signal shows
localized hot spots (two orders of magnitude increase) in the Br3 sample that are not
correlated with the topography. However, the I3 containing thin film displays some PL
variation, but the brighter regions are not nearly as dramatic as in the Br3 one. Their
results, in combination with time-resolved PL measurements on the same samples, revealed
that regions of high radiative recombination for both thin films have slower charge carrier
decay rates. The authors concluded that the areas with higher PL emission are decoupled
from the low PL regions through either a slow diffusion process or energy barrier that
prevents the charge carriers from moving laterally in the perovskite layer. It is also
hypothesized that higher luminescent sites are regions of enhanced structural order and
better crystalline quality. AFM is a vital tool set for advancing the state-of-knowledge
41
of the dynamic responses of perovskites when exposed to ambient environment, where
measurements prior to full device development are extremely important for the screening
and classification of stable alternatives.
3.4 Summary
In this introductory chapter, we presented selected AFM methods that led to re-
cent discoveries related to the electrical, chemical, and optical responses of materials and
devices for the most promising polycrystalline PV materials. Revealing the fundamen-
tal properties of these systems is crucial to increase our understanding of the underlying
mechanisms that define device performance at the nanoscale. Ultimately, we envision that
high spatial resolution investigations will lead to the redesign of materials with improved
performance, as well as devices.
In this summary section we provide an outlook of how AFM can be further imple-
mented to address pressing problems in materials for photovoltaics and energy storage.
First, we speculate how fast AFM methods could be used to help elucidate the role of the
ambient environment onto the degradation of perovskites. Then, we discuss the need for
advancing NSOM-based methods which, to date, are not widely implemented. Measure-
ments using NSOM are particularly relevant to tackle open questions in polycrystalline
PV.
Nanoscale mapping of performance through KPFM and cAFM can potentially de-
convolute the role of humidity, oxygen, illumination, bias, and temperature on perovskites
degradation, including their effect on both grains and grain boundaries. These measure-
ments are likely to redefine the design of solar cells using conventional and emerging
perovskites, considering their contact interfaces. We anticipate the nanoimaging of per-
42
ovskites through “fast” AFM techniques, such as heterodyne-KPFM [34, 39, 155], KPFM
G-mode extension, [156] or 2ω D-KPFM [157] to elucidate their dynamic electrical behav-
ior at relevant length scales, critical to identify the driving forces behind its instabilities.
In particular, we expect researchers to correlate the changes in the optoelectrical proper-
ties (photovoltage, photocurrent, and PL) with possible chemical composition variations
through correlative microscopy (KPFM + cAFM + PiFM) on the very same grains. AFM
experiments where one can elucidate the effects of ambient environment on the local elec-
trical and optical properties will probably focus on correlative microscopy, in order to
define how each parameter affects charge carrier generation, recombination, and collec-
tion, all critical for device performance. We also expect NSOM-PL, KPFM, and TERS
to help us identify which perovskites compositions are stable under illumination. While
extensive effort has focused on determining the ideal synthesis parameters for optimized
device performance, substantial work lies ahead to identify what triggers ion migration
and/or material segregation upon sample illumination and exposure to ambient condi-
tions. This class of material has ∼100 options, [149] thus, measurements on thin films will
accelerate a high throughput and effective material selection, prior to device processing
development.
Concerning polycrystalline materials for PV, direct measurements of light-matter
interactions through underutilized NSOM-based techniques are critical for tackling rel-
evant open questions. For local optical measurements, better NSOM probes are being
designed and engineered to optimize the far-to-near field light coupling, as the current
transmission loss is still 3-6 orders of magnitude. [158,159] More efficient probes will allow
for mapping on a broader variety of low-emission materials, including CIGS and CZTS.
NSOM-PL will further impact polycrystalline GaAs fabrication, [160] as the dependency
43
of radiative recombination on different types grain facets could be resolved. We also antic-
ipate KPFM measurements combined with scanning electron microscopy to become more
common, [161] where one can correlate, in situ, the electrical response of grains and grain
boundaries with their crystalline orientations. To probe the p-n junction in inorganic solar
cells, such as CdTe and CIGS, KPFM and cAFM should be applied in cross-section after
cleaning the sample in an oxygen-free environment, to prevent surface recombination due
to the material’s exposure to air. For that, it’s best practice to operate the AFM inside a
glove box, where the samples can also be cleaved.
In the past three decades AFM has become the standard characterization tool to
image the morphology of a specimen. In the years following its conception, scientists
have modified the instrumentation to image beyond its typical measurement mode, and
unleashed an array of new research opportunities to map other material properties. We
have no doubt that novel adaptations will continue to develop, as speculated above. It’s
all at the tips of our fingers!
44
Chapter 4: Photovoltage Tomography in Polycrystalline Solar Cells
To date, the performance of all polycrystalline photovoltaics is limited by
their Voc, an indicator of charge carrier recombination within the semiconductor
layer. Thus, the successful implementation of high efficiency and low-cost solar cells
require the control and suppression of non-radiative recombination centers within
the material. In this Chapter, we spectrally and spatially resolve the photovoltage
of polycrystalline thin-film CIGS solar cells. Micro-Raman and EDX maps obtained
on the same grains showed that the chemical composition of the CIGS layer is very
uniform. Surprisingly, we observed concurrent spatial variations in the photovoltage
generated across the device, strongly indicating that structural properties are likely
responsible for the non-uniform mesoscale behavior reported here. We build a to-
mography of the photovoltage response at 1-sun global illumination, mimicking the
operation conditions of solar cells. Further, we spatially resolve the voltage within
the CIGS grains, where we found variations >20%. Our functional characterization
could be implemented to identify where non-radiative recombination preferentially
takes place, enabling the implementation of non-uniform materials for future devices
with higher Voc. This chapter is adapted from E.M. Tennyson et. al, ACS Energy
Letters, 1 899905 (2016) [40]
45
4.1 The need for tomography in thin film photovoltaic de-
vices
This chapter investigates thin film Cu(In,Ga)Se2 (CIGS) solar cells, as they are a
practical alternative to commercially available Si photovoltaic (PV) devices due to their
low cost and high efficiency. [162, 163] Further, these technologies have the advantage of
being fabricated on top of either a rigid or flexible substrate, extending solar cell devices
towards a variety of applications. To date, the record CIGS device has an efficiency of η
= 22.6% and open-circuit voltage of Voc = 741 mV, which was fabricated using the well-
established multi-stage co-evaporation technique. [164] However, the sequential deposition
steps substantially increase the overall production cost of CIGS, preventing its wide com-
mercialization. [165,166] An alternative fabrication technique is the sputter deposition of
CIGS through a single quaternary target. [167–169] This method can be implemented in
large-scale applications and also avoids a necessary but toxic post-deposition selenization
step. [170] Despite the advantages of the quaternary target method, the typical Voc of the
PV devices is ∼500 mV, well below the Shockley-Queisser theoretical limit (∼813 mV
for CIGS with Eg = 1.213 eV). [6, 7] Thus, as discussed in Chapters 1 and 3, the Voc is
currently restricting the performance of CIGS solar cells. Because the CIGS absorbing
layer is inhomogeneous and composed of micro-scale grains, it has been speculated that
the electrical response of the material could be limited either by composition variations
within the grains or by the different interfaces throughout the thin film polycrystalline
layer. [56, 171]
In order to probe the local structural and electrical characteristics of materials for
PV, scanning probe and electron microscopies have been widely implemented. [56,172,173]
46
The local electrical response of these inhomogeneous materials have been determined
by laser-beam [95, 174–176] and electron-beam induced current, (EBIC) [55, 177–179]
conductive-atomic force microscopy, [180–183] Kelvin probe force microscopy, (KPFM)
[37, 180, 184] scanning probe microwave capacitance, [185] near-field scanning optical mi-
croscopy, [130–132,186,187] cathodoluminescence, [188,189] and photoluminescence. [190–
195] Further, spectroscopic methods, including electron energy loss spectroscopy (EELS)
[196] and X-ray microscopy, [197, 198] have been implemented to reveal how chemical
composition variations within the active layers affect the electrical response of the de-
vices. Recently, micro-Raman mapping also became a tool to link material composition to
optoelectronic properties. [11, 199] Nevertheless, a high spatial resolution (<1 µm) spec-
troscopic imaging platform to resolve possible photovoltage variations of polycrystalline
solar cells that does not require special sample preparation or an inert environment is still
missing.
Here, we spectrally and spatially resolved the photovoltage of polycrystalline thin
film CIGS solar cells from a quaternary target, which is related to the charge carrier recom-
bination events within the absorbing layer of the device. Micro-Raman maps obtained on
the same grains revealed that the chemical composition of the CIGS layer is very uniform,
in agreement with our energy-dispersive X-ray spectroscopy (EDX) results. Surprisingly,
we observed concurrent spatial variations in the photovoltage generated across the device,
strongly indicating that structural properties are likely responsible for the non-uniform
mesoscale behavior reported here. We built a qualitative tomography of the photovoltage
through a sequence of spectrally dependent voltage scans at low photon injection level,
mimicking realistic operation conditions of photovoltaic devices. The mesoscale electri-
cal response of the material varied at the same length scale as the grains, and different
47
sites (interfaces and groups of grains) showed distinct photovoltage, depending on wave-
length (λ), where voltage variations >20% were measured. These results demonstrate that
the of CIGS solar cells is limited by the types of interfaces existing between the grains
composing the active layer of the devices. Further, we provide a universal and direct
method to map voltage with high spatial resolution without the use of a scanning probe
or vacuum environment that can be applied to other optoelectronic devices, such as LEDs,
photodetectors, etc.
4.2 Experimental methods
4.2.1 CIGS solar cell fabrication
The CIGS solar cell samples were fabricated via sputtering from a quaternary tar-
get. [167] The substrate for all samples was soda lime glass coated with 700 nm of Mo.
coated with 2.0 µm thick p-type CIGS films deposited by RF magnetron sputtering. Then,
a chemical bath deposition of 50 nm thick n-type CdS layer was performed, and a 500
nm transparent conductive top contact with ZnO and Al:ZnO was sputter deposited to
enhance current collection. Lastly, Ni/Al grids were evaporated to serve as the top con-
tact. The samples were scribed by hand to obtain individual cells, each with an area of
approximately 0.54 cm2. Figure 4.1(a) shows a schematic of the CIGS devices. The inset
of Figure 4.1(b) is a photograph of six electrically isolated CIGS devices on a single 1 ×
1 in2 glass substrate.
48
Figure 4.1: CIGS polycrystalline solar cells fabricated by sputter deposition
from a quaternary target. (a) Schematic of CIGS solar cell, out of scale for clarity.
(b) Representative light J-V curve for device illuminated with AM1.5G at 1-sun,
yielding Voc = 490 mV and η = 7.77%. The inset shows a photograph of six solar
cells electrically isolated on a single 1 × 1 in2 glass substrate. (c) EQE measurement
of the same device shown in (b).
4.2.2 Device macroscopic characterization
Light J-V curves were acquired using a Xenon lamp solar simulator, with 2.0 × 2.0
in2 (50.8 × 50.8 mm2) output, and an Air Mass 1.5G filter (see Figure 4.1(b) for light J-V
curve). The output power of the lamp was set to 1.0-sun (100 mW/cm2). The light source
was calibrated with a 20 × 20 mm2 reference monocrystalline Si solar cell with a fused
silica window. For the EQE measurement Figure 4.1(c), a xenon arc lamp was directed
into a monochromator. Two silicon diodes were initially used: one as a reference and the
other for calibration. A beam splitter guided half of the light to the fixed reference diode;
the remaining light was directed through another optical path and collected at the position
of either the calibration diode or the solar cell sample. Independent measurements using
the silicon calibration diode were performed before each EQE measurement was acquired.
The ratio of the spectral responses between the solar cell and the calibration diode was
computed every 10 nm. The EQE for CIGS was determined between 390 nm and 1100
nm. The low EQE for wavelengths <500 nm results from the expected light absorption
49
by the ZnO (with Eg ∼3.3 eV) layer and the CdS n-layer (with Eg = 2.4 eV) [200].
4.2.3 Raman spectroscopy mapping
A 532 nm laser beam was focused onto the sample using a 100×microscope objective
lens (0.75 NA), the scattered light was collected and collimated by the same objective.
The scattered signal was then collected by a spectrometer and detected with a liquid
nitrogen-cooled charge-coupled-device (CCD). The sample was mounted on an x-y piezo
scan stage for the map acquisition. Thus, a Raman spectrum was obtained for each pixel
in an image. By applying selective wavenumber filters we produced Raman maps for the
relevant vibrations of CIGS.
4.2.4 Photovoltage imaging
Incident light was coupled in to a 25 µm multimode fiber attached to the optical
microscope and transmitted through a 100× objective lens, NA = 0.75, focusing a ∼1
µm diameter spot onto the surface of the solar cell. The incident photon flux was 1.9 ×
1023 photons/m2/s for the photovoltage maps shown in Figures 4.4, 4.6, and 4.13, and
>1025 photons/m2/s for Figures 4.8, 4.9, and 4.14. The charge carriers generated by the
controlled illumination were collected through their corresponding contacts, one connected
to ground, and the other connected to a lock-in amplifier, used to read the electrical output
directly from the device at a high signal-to-noise ratio. In all cases, the top contact was<25
µm away from the scanned region. Here, the lock-in amplifier inherently subtracted the
dark current signal from the illuminated scans [174]; thus, the acquired voltage maps refer
to an effective open-circuit voltage. A piezo scanner was used to acquire the photovoltage
maps for each position of the solar cell, with high spatial resolution. Figure 4.2 displays
50
a schematic of the experimental setup described here. Because the acquisition of each
voltage map took ∼1 hr, we corrected the drift after each scan to ensure the validity of
our wavelength dependent analysis, where the center of the scanned area was recalibrated
with the spatial resolution of the piezo scanner (<100 nm). The average between two
neighboring points was taken for all maps shown in this Chapter.
Figure 4.2: Scanning photovoltage experimental setup. Incident light is cou-
pled in to a 25 µm multimode fiber attached to the optical microscope. The light is
transmitted through a 100× objective, NA = 0.75, focusing a ∼1 µm diameter spot
onto the surface of the solar cell. The charge carriers are collected by their corre-
sponding macroscopic contacts, one connected to ground, and the other connected
to a lock-in amplifier, which reads the electrical output directly from the device at a
high signal-to-noise ratio. A piezo scanner is used to map the voltage signal for each
position of the device, resulting in a map with lateral spatial resolution <500 nm,
defined by the beam spot size.
4.3 Correlating chemical and electrical maps in CIGS
We utilize standard characterization methods to obtain the macroscopic photovoltaic
performance of the CIGS solar cells investigated here as described in the Experimental
Methods section and in ref [167]. The CIGS layer is composed of grains with ∼1 µm in
size, preferentially orientated along the 〈112〉 direction (see Figure 4.3 for a representative
SEM and electron back scatter diffraction (EBSD) map).
51
Figure 4.3: CIGS electron backscatter diffraction pattern. (a) SEM image of
CIGS grains. (b) EBSD map along the normal direction of the CIGS surface showing
preferential grain orientation along 〈112〉 direction, as indicated by the color scale on
the right.
We investigate the relationship between the chemical composition and the electrical
properties of the CIGS devices by micro-Raman imaging and EDX in conjunction with
scanning photovoltage microscopy on the same grains. For the scanning photovoltage
measurements a 100× objective lens is used to focus the local source of excitation, see
Experimental Methods.
Remarkable variations in the photovoltage were measured with high spatial reso-
lution, as shown in Figure 4.4. Note that the average voltage signal is small (<2 mV)
because of the confined volume of excitation during the measurements (spot area equal
0.95 µm2 compared to the total area of the device, equal to 0.54 cm2), causing a voltage
drop to occur between the dark and light regions of the cell. [201, 202] Yet, the electrical
variation in the scanned area is 0.5 mV, considerably higher than the noise (<10 µV). The
representative Raman spectrum with both the strong Cu2−xSe and the chalcopyrite peaks
is presented in Figure 4.4(c) (Raman shifts at 295 cm−1 and 172 cm−1 highlighted in gray
and brown, respectively), [203,204] where the inset refers to the optical micrograph of the
region mapped. The micro-Raman images corresponding to the Cu2−xSe and chalcopyrite
peaks were acquired on the same location as the voltage map and exhibit uniform surface
52
chemical composition (despite surface segregation of Cu2−xSe); see Figure 4.4(d,e) (see
Figure 4.5 for additional Raman maps). The EDX maps of Cu, Ga, In and Se presented
in Figure 4.4(f)-(i) show very uniform elemental distribution in all cases (within the spa-
tial resolution of the SEM probe used). Our chemical composition measurements strongly
indicate that the spatial variation observed in the CIGS electrical response is likely due to
the distribution of grain orientation or grain boundary types, which could result in areas
of high rates of charge carrier recombination. [55,205]
Figure 4.4: Chemical composition and local electrical response of CIGS so-
lar cells. (a) SEM image of CIGS. (b) Photovoltage map showing spatial variation
>15%; illumination conditions: 100× objective, λ = 575 ± 3 nm and incident power
= 298 ± 3 µW. (c) Representative Raman spectrum. Inset: optical micrograph indi-
cating the region where EDX, photovoltage, and Raman measurements were acquired
(white box). Raman maps of (d) Cu2−xSe peak (grey filter in (c)) and (e) chalcopy-
rite peak (brown filter in (c)). Illumination conditions for Raman: λ = 532 nm and
incident power = 100 ± 5 µW. EDX maps showing (f) Cu, (g) Ga, (h) In, and (i) Se
uniform distribution.
53
Figure 4.5: Raman microscopy on CIGS: additional region.(a,e) Optical mi-
crographs of CIGS solar cell. The white box refers to the 10 × 10 µm2 region where
micro-Raman imaging measurements were performed. (b,f) Chalcopyrite peak with
Raman shift = 172 cm−1 and (c,g) Cu2−xSe peak with Raman shift = 295 cm
−1,
where each pixel in the images is a spectrum within the brown and gray highlighted
region in (d) and (h), respectively. (d,h) Single Raman spectra extracted from the
center pixel of its corresponding Raman map.
4.4 Spectrally dependent photovoltage imaging results and
discussion
We performed a test measurement to verify the photovoltage signal. For this, the
incident laser light was switched ON and OFF, see Figure 4.6. Furthermore, to prove
that the voltage variation in these photovoltage measurements were representative and
reproducible we acquired maps on a variety of CIGS solar cells as shown in Figure 4.7.
Here, we imaged three similarly fabricated devices that all show voltage variations at the
same length scale as the composing CIGS grains.
54
Figure 4.6: Laser ON/OFF photovoltage signal verification. (a) Optical mi-
crograph of CIGS solar cell; the white square indicates the region where the voltage
scan was performed. (b) Scanning photovoltage map of a typical CIGS solar cell
showing spatial variation in voltage. Illumination conditions: 600 nm laser; incident
power = 1.44 µW; objective lens: 100×, NA = 0.75.
Figure 4.7: Additional photovoltage mapping on CIGS solar cells.(a)-(c) Op-
tical micrographs of three different CIGS solar cells where the white square indicates
the corresponding scanned photovoltage map below. The bright white region in each
micrograph is the top electrical contact; here the black dashed line is a guide for the
eye to emphasize the interface between the metal and the semiconductor layers. To
ensure effective charge carrier collection we scanned regions <25 µm away from the
top macroscopic contact. (d)-(f) Photovoltage maps acquired for solar cells 1, 2 and
3, respectively, with spatial variation on the same length scale as the CIGS grains.
55
To understand the contribution of the photovoltage generated within the CIGS
grains to the overall performance of the solar cell as a function of light penetration depth
we spectrally resolve the electrical response of the CIGS sample while mimicking the
illumination conditions of a working device. See Figure 4.8(a) and (c) for the 1-sun spec-
trum at AM1.5G illumination [206] and the EQE, respectively, where the colored dots
represent the wavelengths at which the microscopic photovoltage measurements were per-
formed. Figure 4.8(b) shows an optical micrograph of a selected region of the CIGS solar
cell. We acquire the photovoltage maps using the same number of incident photons in all
cases (Figure 4.8(d)-(u)). The electrical signal continually increases with the wavelength,
reaching a maximum value at 950 nm (Figure 4.8(r)), approaching the bandgap of the fab-
ricated CIGS layer (equal to 1.15 eV, as determined by photoluminescence). As expected,
beyond the semiconductor bandgap (corresponding to wavelength λ >1000 nm) the elec-
trical signal drops dramatically (Figure 4.8(t) and (u)), indicating that the very modest
photogenerated voltage is likely due to trap states within the polycrystalline layer [131].
Nevertheless, spatial variations are still observed and are well above the noise level (20
nV) of the photovoltage measurements introduced here. The in depth electrical response
of the CIGS layer is obtained without cross-sectioning (destroying) the sample. Although
the surface of the sample is being probed, the electrical response of the subsurface grains
and interfaces was acquired due to the absorption coefficient of the material. By compar-
ing the spatial variation in photovoltage of individual scans (e.g. Figure 4.8(g) and (n)),
one notices that the behavior changes as a function of wavelength, demonstrating that the
photovoltage signal is resulting from the electrical response within the material (and not
its surface only). Thus, through our imaging method we acquire a tomography of the elec-
trical characteristics of the absorbing material where each scan corresponds to a snapshot
56
of the mesoscale behavior of the grains composing the CIGS layer. The word tomography
is used here to emphasize that the device is imaged in ’sections’ by illuminating it with
different wavelengths, which, by definition, corresponds to a distinct penetration depth
within the CIGS layer.
Figure 4.8: Spectrally dependent voltage maps on CIGS solar cell. (a)
1-sun AM1.5G spectrum as a function of wavelength. (b) Optical micrograph of
CIGS solar cell acquired with 100× objective, NA = 0.75, spot size ∼1 µm. The
white box highlights the 7 × 7 µm2 area where the photovoltage measurements were
performed. (c) EQE measurements. The colored dots in (a) and (c) represent the
wavelengths where the microscopic measurements were performed. (d)-(u) Spectrally
dependent scanning photovoltage microscopy scans. For each scan: the label indicates
the wavelength, and the color scale bar is in µV.
The spatial distribution of the voltage for each scan is used to quantify the occur-
rence of preferential charge carrier generation/collection sites as a function of wavelength
57
and, thus, light penetration depth within the CIGS layer. As shown by the histograms
in Figure 4.9, the voltage average value and distribution for each map vary as a function
of wavelength. While the absorption coefficient is larger for shorter wavelengths (Figure
4.10), a smaller average value of the voltage is obtained. This behavior suggests that car-
riers generated near the top surface are more likely to recombine non-radiatively, leading
to a smaller voltage. This observation is corroborated by the EQE data (Figure 4.1(c)),
which also shows a decrease in carrier collection. For longer wavelengths, the absorption
occurs deeper within the CIGS layer and leads to increased photovoltage until the bandgap
is reached, where the voltage again decreases due to lack of photon absorption. The pre-
cise cause of the broadening of the voltage distribution with wavelength is unknown and
is currently being investigated.
Figure 4.9: Spectrally dependent voltage histograms. 3D plot of voltage his-
tograms (normalized) as a function of wavelength for CIGS solar cell, obtained from
scanning photovoltage microscopy (shown in Figure 4.8).
58
Figure 4.10: CIGS optical properties. (a) Index of refraction n and extinction
coefficient k as a function of wavelength, as measured by ellipsometry on the polished
CIGS sample. (b) Absorption coefficient, α, as a function of wavelength.
The possible contribution of surface roughness and consequent light reflection and
scattering to the voltage signal was excluded by a detailed comparison between as-is
and polished grains, where spatial variations in photovoltage to occur irrespective of the
morphology of the grains. The CIGS sample was chemically polished with aqua regia using
a method described in ref [207]. The sample was immersed in a 1:3 solution of HCl:HNO3
at room temperature for nine seconds, followed by a rinse in DI water. Figure 4.11 shows
a cross-section SEM of both as-is and polished grains while Figure 4.12 shows a top-view
SEM, optical micrographs, atomic force microscopy scans, and line profiles.
Figure 4.11: SEM images of as-is vs. polished CIGS grains. Cross-section
SEM images of (a,c) as-is and (b,d) polished CIGS samples.
59
Figure 4.12: Electron, optical, scanning probe microscopy of as-is and pol-
ished CIGS grains. SEM, optical and AFM images and line scans of (a-d) as-is
and (e-h) polished CIGS grains post-sputter fabrication.
In Figure 4.13 the voltage distribution maps for both the as-is and polished samples
are shown along with their macroscopic light J-V curves. The non-uniformity of the
voltage signal in both as-is and polished CIGS samples demonstrates that different charge
carrier dynamics are taking place at the distinct grains. Note that there is no correlation
between the spatial variation in the voltage signal and the optical micrographs (reflected
light). Light J-V curves of Figure 4.13(c) as-is and Figure 4.13(f) polished solar cells were
tested under AM1.5 global illumination (1-sun, 100 mW/cm2). Here, the low voltage
from the polished sample results primarily from parasitic shunt resistance (Rsh) that
reduces the amount of current flowing through the solar cell, which most likely results
from surface recombination within the CIGS layer after the chemical polishing. Note that
the series resistance (Rs) in both samples is similar, which is not related to the voltage
of the device. Although the polished sample has surface roughness of ∼100 nm (Figure
4.12), we were unable to acquire reliable information about the grains crystallographic
orientation by EBSD due to the ZnO 500 nm layer (Figure 4.1(a)). Thus, the precise
60
correlation between the spatial variations in Voc and the structural properties of the CIGS
requires the implementation of a selective chemical etching for removing the oxide layer.
Nevertheless, this etching would necessarily be destructive and the exposure of the n-CdS
would favor surface recombination, which directly affects the Voc of the device.
Figure 4.13: Voltage distribution of a polished CIGS sample. Optical image
of (a) as-is and (d) polished CIGS grains imaged under the same illumination con-
ditions. Scanning photovoltage images for (b) as-is and (e) polished grains, using
100× objective, NA = 0.75. In both cases, the scans were acquired <25 µm from the
electrical contacts, and were illuminated with 600 ± 3 nm laser source at a power =
1.44 µW.
To further evaluate the spatial inhomogeneity of charge carrier dynamics within
the CIGS layer we analyze the voltage enhancement for the set of maps shown in Figure
4.8. Figure 4.14(a) shows a photovoltage scan acquired at 750 nm overlaid on the SEM
image for the same 7 × 7 µm2 region. Despite the fact that the scanning photovoltage
measurements are limited by diffraction, notable spatial variations were consistently ob-
served. Figure 4.14(b) displays the voltage enhancement as a function of wavelength for
the positions (pixels) indicated by the colored circles in Figure 4.8(a). Here, we define
the enhancement as the voltage of a pixel, Vx,y, divided by the average value of the entire
region, 〈V〉, (52.1 µV in this case). Note that both Vx,y and 〈V〉 change for each incident
61
wavelength (due to the CIGS absorption coefficient, see Figure 4.10).
Surprisingly, some regions of the sample showed a modest voltage improvement of
10% for short wavelengths (400-500 nm), and a remarkable enhancement of 40% close
to CIGS bandgap. The low electrical signal (black and cyan circles in Figure 4.14(b))
approaches the average value (green solid line) as the energy of the incident source of
excitation nears the bandgap of CIGS. Interestingly, certain regions of the CIGS layer
(red points in Figure 4.14(b)) present local voltage that varies 10% below and above the
average response of the device, for wavelengths <550 nm and >600 nm, respectively. Near
the material bandgap the voltage enhancement reaches 20%. By spectrally and spatially
resolving the photovoltage generated by the CIGS solar cell we obtain a qualitative tomog-
raphy of the local electrical characteristics of the device and probe their limiting factor:
the Voc (see Figure 4.14(c)). Here, the maps are displayed on the same voltage scale to
emphasize the spatial variations as a function of the incident wavelength, which are likely
due to the structural properties of the CIGS layer, instead of composition inhomogeneities
between different grains. The longer wavelength voltage maps encompass the near-surface
strong light absorption, and the qualitative tomography displays the influence of the each
wavelength’s light absorption on the photovoltage.
62
Figure 4.14: Mapping voltage with high spatial resolution in polycrystalline
solar cells. (a) Voltage map overlaid with SEM image, acquired at λ = 750 ± 3
nm and incident power of 59 ± 1 nW. (b) Voltage enhancement as a function of
wavelength for selected regions of the map shown in (a). The green line corresponds
to the average voltage value of each map, and the error bars refer to the standard
deviation of each λ-dependent scan. (c) Qualitative tomography of photovoltage
showing spatial variation as a function of λ. For each measurement: 100× objective
lens (NA = 0.75).
Through a sequence of spectrally dependent photocurrent scans we determine the
∆Voc map for the polished CIGS device under 1-sun illumination (Figure 4.15). This map
is obtained by adding the contribution of each photocurrent scan to the overall short-
circuit current map Isc,total of the device, weighted by the number of photons incident into
the sample at AM1.5G illumination, as:
∑
Jsc,total = Jsc,λn (4.1)
λn
∫ (λn+1+λ)/2
J measuredsc,λn = Isc,λ × (φ1−sun × λdλ)/(Pinc × λn) (4.2)n
(λ+λn−1)/2
where Jsc,λn is the short-circuit current density for each wavelength λn where a scan was
acquired, Imeasuredsc,λ is the short-circuit current measured (a map), φn 1−sun is the irradiance
corresponding to 1-sun AM1.5G illumination (from Figure 4.8(a)), and Pinc is the incident
63
power for each measurement (wavelength dependent). Then, the Voc map is determined by
using the diode equation, where the ideality factor and the dark saturation current density
(Jdark) were obtained from macroscopic measurements. The spatially resolved Voc map
combines the contribution of the surface and the bulk properties of the grains, weighted
by the sun spectrum, mimicking the performance of a solar cell under realistic operation
conditions.
Figure 4.15: Characterization of the mesoscale behavior of CIGS solar cells.
Voc map for CIGS solar cell, with sub-micron spatial resolution, obtained from a
sequence of spectrally dependent scans.
4.5 Conclusion
Here, we have implemented a universal, non-destructive imaging method to map
the spatial variation of the electrical response of optoelectronic devices in situ that en-
abled a tomography of the generated photovoltage within the solar cell absorbing layer.
The spatial resolution of the diffraction-limited method described in this chapter could
be improved by using near-field scanning optical microscopy probes as the source of illu-
mination, [131] which can achieve resolution <100 nm. Nevertheless, the interpretation
of the data in these cases often requires the consideration of near-field effects, and possi-
bly the need for 3-dimensional simulations to help us understand the primary near-field
64
light-matter interactions taking place. This functional characterization reveals how the
electrical response of CIGS devices from a quaternary target changes spatially, considering
light penetration depth. Further, the sub-micron spatial resolution of the photovoltage
scans presented here elucidates a pressing open question in the field of photovoltaics: does
the Voc of polycrystalline CIGS devices from a quaternary target vary spatially? Yes.
We demonstrated that the photovoltage of CIGS solar cells from a quaternary target
varies locally in >20%, a behavior mapped here for the first time. We used micron-
Raman imaging and EDX to determine the chemical composition of the grains. While the
distribution of Cu, In, Ga and Se was found to be uniform at the micron-scale, the Voc of
these devices is still considerably lower than expected. Scanning photovoltage microscopy
was implemented as an in situ functional characterization tool to resolve the local electrical
response of CIGS, where remarkable variations in voltage were measured. A sequence of
wavelength dependent voltage scans were used to build a tomography of the photovoltage
distribution within the CIGS layer; where some regions of the material outperform the
average, macroscopic behavior, potentially affecting the performance of the device. Finally,
we presented a micron-scale resolved Voc map under broadband illumination, revealing
the spatial distribution of voltage for the CIGS solar cell under 1-sun global illumination
operation conditions. This functional imaging framework provides a novel tool to map the
electrical response of inhomogeneous materials for photovoltaics with micron-scale length
grains, such as perovskites, CdTe, and CZTS. Further, this non-destructive method can
be expanded to other optoelectronic devices, such as photodetectors and LEDs.
65
66
Chapter 5: Nano-imaging of open-circuit voltage in photovoltaics
For most photovoltaic devices, the current density is near its maximum value;
however, achieving a large open-circuit voltage has proven difficult for nearly all
photovoltaic technologies. Understanding the carrier recombination processes on the
nanoscale that lead to this voltage reduction is critical, yet no technique exists to
microscopically map recombination and the resulting open-circuit voltage. Here we
present a novel metrology to image the open-circuit voltage with spatial resolution
<100 nm, more than five orders of magnitude better than previous methods, using
a variant of illuminated Kelvin Probe Force Microscopy. We apply this technique
to several materials (GaAs, Si, CdTe, and CIGS), and find that the open-circuit
voltage varies locally by more than 200 mV, suggesting that spatial variation of
non-radiative recombination strongly affects the overall device performance. This
technique enables new insights into the loss mechanisms that hinder solar cells and
provides a new platform to image device performance with nanoscale resolution.
This chapter is adapted from E.M. Tennyson et. al, Advanced Energy Materials, 5
1501142 (2015) Front Cover [37]
67
5.1 Introduction to local voltage mapping in thin-film solar
cells
For most photovoltaic devices, the record short-circuit current density (Jsc) is near
its theoretical limit; however, achieving a large open-circuit voltage (Voc) has proven dif-
ficult for nearly all photovoltaic technologies. [7, 208] Understanding the carrier recombi-
nation processes, caused by structural or compositional variations at the nanoscale, that
lead to this voltage reduction is critical, yet no technique exists to quantitatively map re-
combination and the resulting Voc at the nanoscale. Here we present a novel metrology to
image the Voc with spatial resolution <100 nm, more than five orders of magnitude better
than previous methods, using a variant of illuminated Kelvin Probe Force Microscopy. We
apply this technique to several materials (GaAs, Si, CdTe, and CIGS), and find that the
Voc varies locally by more than 200 mV, suggesting that spatial variation of non-radiative
recombination strongly affects the overall device performance. This technique enables new
insights into the loss mechanisms that hinder solar cells and provides a new platform to
image device performance with nanoscale resolution.
The Voc of a solar cell is a measurement of the maximum voltage generated by the
device under illumination, which is proportional to the quasi Fermi level splitting of the
semiconductor p-n junction. This generated voltage and the overall electrical behavior of
a device strongly depend on the non-radiative recombination rate of the charge carriers
within a material, as described in Chapter 1, which is affected by the defects inherently
present within the semiconductor. Despite all the effort in developing higher perfor-
mance thin-film polycrystalline solar cells, such as CdTe, CuInxGa(1−x)Se2 (CIGS), and
Cu2ZnSnS4 (CZTS), the difference between the theoretically predicted and the best exper-
68
imentally achieved Voc is still considerably large, as shown in Chapter 3. For Si, extensive
research has been dedicated to design and implement nanostructured light-trapping ar-
chitectures to boost light absorption [209–212] however, there are very few experiments
showing how the Voc is affected. Thus, for any micron- and nano-scale structured PV
device, assessing variations in Voc with nanoscale resolution and spatially resolving where
recombination occurs within the material can potentially change the pathway for designing
higher performance devices.
Imaging methods based on atomic force microscopy (AFM) techniques have been ex-
tensively used to characterize the structural and electrical properties of PV materials and
full devices. [172,182,213–222] In particular, Kelvin probe force microscopy (KPFM) has
been implemented to probe the electrical characteristics of a variety of PV materials and
devices, ranging from organic materials [223–226] and oxides [227] to III-V semiconduc-
tors for multijunction designs [228–230] and polycrystalline thin-films. [184, 219,231–236]
The local optoelectronic properties and changes in material composition have also been
mapped using near-field scanning optical microscopy (NSOM) probes as local sources of
excitation. [130,131,186,237–240] Recently, photoluminescence has emerged as a promising
tool to map charge recombination [129,241,242] and carriers diffusion [194] with high spa-
tial resolution. At low-temperature (70 K), photoluminescence imaging with sub-micron
resolution has been implemented to map a 10 meV quasi-Fermi level splitting in CIGS
solar cells, where variations in the intensity signal were attributed to changes in the ma-
terial composition. [190] Nevertheless, none of these imaging techniques provide a direct
measurement of Voc within the material at operating conditions. A straightforward, uni-
versal, and accurate method to measure the Voc (and hence non-radiative recombination
processes) with high spatial resolution in PV materials is still missing.
69
Here we present a new imaging technique based on illuminated-KPFM to map the
Voc of optoelectronic devices with nanoscale resolution <100 nm. We map the contact po-
tential difference of a half or fully processed solar cells in the dark and under illumination
to determine the quasi Fermi level splitting of the p-n junction. We spatially resolve the
Voc of different PV materials, including III-V direct bandgap semiconductors, monocrys-
talline Si, and polycrystalline thin-films, demonstrating the universality of our platform
for imaging devices performance. While single crystalline GaAs and Si solar cells show
very uniform Voc maps, CdTe and CIGS present dramatic variations in Voc, larger than 0.2
V, which have never been resolved by other microscopy methods. We find that the local
variations in the Voc are due to the fact that different grain orientations can act as distinct
centers for recombination within the material. Although KPFM measurements under illu-
mination (named surface photovoltage - SPV) have been realized in a variety of solar cells
technology [84, 243–253], here a direct correlation between KPFM measurements (light-
minus dark-KPFM) and the Voc of photovoltaic devices, through the measurement of the
quasi-Fermi level splitting is demonstrated for the first time. This nanoscale metrology is
non-destructive and can be implemented in ambient conditions, allowing for the diagnosis
of how the different processing steps can affect the recombination within the material and,
therefore, the ultimate performance of an optoelectronic device. For non-uniform semi-
conductor materials, mapping local variations in Voc is extremely useful to identify which
types of interfaces can suppress non-radiative recombination processes.
70
5.2 Experimental methods
5.2.1 Solar cell fabrication
The GaAs solar cell was obtained from M-Comm. The cells were grown on an n-type
GaAs wafer (300 µm thick, 5 × 1018 cm−3), and the active regions were cladded by AlGaAs
(Eg = 1.80 eV, 3 × 1018 cm−3) passivation and back surface field layers. The GaAs solar
cell used for system calibration was extremely small (1 × 3 mm2). As a consequence,
the ideality factor of the device is particularly high, due to edge surface recombination
effects. [254] The monocrystalline Si solar cell is a commercially available device, formed
by n- and p-doped Si layers, with pyramidal texturing to enhance light trapping. The
commercially available CdTe solar cell measured was formed by the following layers (from
top to bottom): 4.0 mm of glass substrate, 550 nm of a bilayer transparent conductive
oxide, 50 nm of n-type CdS, and 3.5 µm of p-type CdTe, which was submitted to the
standard CdCl2 treatment. Due to the geometry of the device, the illuminated-KPFM
measurements were performed on the exposed p-doped CdTe grains with the original back
contact. The CIGS solar cell sample was fabricated by sputtering from a quaternary
target, which consistently produces polycrystalline, highly oriented grains with controlled
stoichiometry, and a bandgap of 1.1 eV. [167] Soda lime glass coated with 500 nm of Mo
was used as the substrate. P -doped CIGS films with 2.0 µm in thickness were deposited
by RF magnetron sputtering. A 50 nm thick n-type CdS layer was deposited by chemical
bath deposition. For all samples, no anti-reflection coating was added.
71
5.2.2 KPFM measurements
It has been shown that AM-KPFM is susceptible to internal AC-signal couplings
which can lead to a topography-dependent offset in the applied voltage. [255] Thus, to en-
sure no topography pick-up we measure the cantilever response to electrostatic excitation
in both the in- and out-of-phase channels across the entire resonance. The response is that
of a simple harmonic oscillator, without any sign of AC coupling. Because the coupling
scales with frequency, a cantilever with a low resonance frequency is chosen. Furthermore,
the coupling is independent of illumination, so it does not affect the results pertaining to
the change in surface potential caused by incident light intensity. The cantilever spatially
averages the tip-sample interaction due to the long-range nature of the electrostatic inter-
action, [256] reducing the spatial voltage variation signal, [257] importantly, this does not
change the voltage shift due to illumination.
5.2.3 Macroscopic characterization of devices
Dark and light J-V curves were acquired using a full spectrum solar simulator,
with 2.0 × 2.0 in2 (50.8 × 50.8 mm2) collimated output, and an Air Mass 1.5G filter.
The output power of the xenon lamp was independently controlled, ranging from 0.1 to
1.0-sun (100 mW/cm2). The light source was calibrated using a 20 × 20 mm2 reference
monocrystalline Si solar cell with a fused silica window. Light J-V measurements for all
the devices measured here are presented in Figure 5.1. A dark J-V measurement of the
GaAs solar cell was used to determine the dark current of the device in Figure 5.2.
72
Figure 5.1: Light J-V curves for four solar cell devices. (a) GaAs (Voc = 0.83
V), (b) monocrystalline Si (Voc = 0.59 V), (c) CdTe (Voc = 0.56 V), and (d) CIGS
(Voc = 0.50 V). Inset: active layers of each device (out of scale for clarity). For all
samples, no anti-reflection coating was added.
Figure 5.2: Dark J-V for GaAs solar cell. (a) Semilog dark J-V and (b) linear
fit used to determine Jdark.
73
Figure 5.3: Angle dependence of external quantum efficiency for the GaAs
solar cell. EQE measurements at normal incidence (black) and 77.5o (red) using
polarized light, in the same geometry of the illuminated-KPFM measurements.
5.2.4 EQE measurements of GaAs
A xenon arc lamp was directed into a monochromator. Two identical silicon diodes
were used to regulate the spectral response of the EQE measurements; one as a reference
and the other for calibration. A beam splitter guides half of the output light to the fixed
reference diode, the other half was directed through an optical path and collected at the
position of the calibration diode or solar cell sample. The ratio of the spectral responses
between the reference and calibration diode was computed in 10 nm increments. The EQE
of the solar cell was measured from 390 nm and 900 nm. Independent calibrations were
performed before each EQE measurement. A polarizer was used to match the polarization
of the incident light with the illuminated-KPFM measurements. Figure 5.3 shows the EQE
measurement taken on GaAs solar cell used to calculated the calibration constant.
74
5.2.5 Voc image analysis
The SPV images were corrected for drift using a correlation function which computes
the similarity between two topographical scans as they shift in time. The function outputs
a value that corresponds to the magnitude in which the image is shifted in x and y
directions. Once the topography correction was complete the potential maps of the dark-
and illuminated-KPFM scans were subtracted.
5.3 Direct imaging of Voc at the nanoscale: results and dis-
cussion
We implement illuminated-KPFM as a new tool to directly map the Voc of photo-
voltaic materials with nanoscale resolution. In KPFM, the contact potential difference
(CPD) is proportional to the work function difference between the tip and the surface
of the material under investigation (qVCPD = Wtip −Ws). For a solar cell, as shown in
Figure 5.4(a) and (b), the CPD signal depends on the illumination conditions as:
qV d dCPD = Wtip −W ds (in the dark), (5.1)
and
qV il = W ilCPD tip −W ils = qV dCPD + ∆µ (under illumination) (5.2)
where q is the electron charge, the superscripted d and il refer to dark and illuminated
conditions, respectively, and ∆µ is the quasi Fermi level splitting. The Voc of the device
is:
∆µ
Voc = = V
il d
q CPD
− VCPD (5.3)
75
(a) dark-KPFM qV dCPD (b) illuminated-KPFM qV dCPD (c) KPFM setup Wd il function tip Wtip
E Wd E Wil generator vac Wd p vac Wil pd n nW Wil cantilever m m lock-in holder 
E -  
laser	
  	
   ∆H	
   amplifier 
c	
   Ec	
  
ε feedback ε F	
  F	
   Δµ V	
   for DC 
3kΩ	
   solar	
  cell	
   voltage 
E tip metal tip 
bottom contact 
metal v	
   E + v	
   x,y	
  piezo	
  stage	
  
p-type p-type 
n-type n-type 
Figure 5.4: Energy diagram of KPFM measurements. Band profile of n-p
junction (a) in equilibrium (dark) and (b) under illumination. Evac, Ec, Ev, and F
refer to the energy of the vacuum level, the conduction band, the valence band, and
d(il) d(il)
the Fermi level, respectively. Wn and Wp are the work function for the dark
(illuminated) n- and p-type layers. EFn, and EFp refer to the Fermi level splitting ∆µ
upon illumination. (c) Cross-sectional illustration of illuminated-KPFM setup, used
to map open-circuit voltage Voc of photovoltaic materials with nanoscale resolution.
The metrology is implemented in ambient environment and only requires the bottom
contact of the device. For each scan, the probe passes over the surface twice: in the
first one topography is acquired, and in the second one the potential difference is
recorded at a height ∆H from the first pass. The illumination source is a 660 nm
continuum laser (represented by the red arrow). The surface photovoltage (SPV),
difference between the illuminated- and dark-KPFM signals, is proportional to the
Voc of the solar cell.
Note that this technique does not require the preceding knowledge of the work function of
the sample or the tip, because ∆µ is determined by measuring the difference between these
two quantities under distinct illumination conditions. Thus, the nanoscale maps of the Voc
of partially processed devices are determined from a pair of dark- and illuminated-KPFM
images.
Amplitude-modulated (AM) KPFM is implemented using metal-coated Pt probes
with mechanical resonance ω2π = 86.7 kHz in ambient conditions - see schematic setup
in Figure 5.4(c). For each scan, the probe passes over the surface twice. On the first
pass, a feedback loop controls the sample height in order to maintain constant cantilever
oscillation amplitude; and topography is acquired in non-contact mode. On the second
pass, the cantilever is held at a constant height (∆H) relative to the path it traced on
the first pass − see Figure 5.4(c). The tip oscillates slightly below resonance during
76
the topographic scan, to maintain stable tapping mode. [258] The amplitude of the tip
oscillation is ∼60 nm (though this value varies slightly from scan to scan, depending on
the optical lever sensitivity of the probe).
The Voc of a photovoltaic device is given by:
( ) ( )
nkT Iil nkT Iil
Voc = ln + 1 =∼ ln (5.4)
q Idark q Idark
where n is the ideality factor, k is Boltzmann’s constant, T is the temperature, and Iil and
Idark are the light generated and the dark saturation current, respectively. For a partially
illuminated solar cell the Voc can be written as:
( ) ( )
partial nkT q ∗ φ ∗ EQEo τ ∗Ail ∗ EQEλ,ΘVoc (φ) = ln + ln = V fulloc (φ) + β (5.5)q Jdark Acell ∗ EQEo
where Iil = q∗φEQE(λ,Θ)∗Ail, φ is the number of photons per area per time from the light
source, EQEo and EQEλ,Θ are the external quantum efficiencies at normal incidence and
at an incident angle Θ for illumination at wavelength λ, Jdark is the dark current density,
τ is the transmission fraction of the optical setup, Ail is the area of the solar cell being
illuminated, Acell is the total area of the solar cell, and β is a calibration constant (<0),
given by:
( )
nkT τAilEQEλ,Θ
β = ln (5.6)
q AcellEQEo
77
Note that for illumination of the full device, as is the case for a solar simulator measure-
ment, β = 0. With KPFM, the solar cells are only partially illuminated, and the light
generated current is very small compared to the situation of full cell illumination. To
determine the expected Voc for full illumination using KPFM:
V fulloc (φ) = V
partial
oc (φ) + β (5.7)
where V partialoc (φ) is the illuminated- minus dark-KPFM voltage signal, i.e. the surface
photovoltage (SPV) of the cell. Alternatively, if V fulloc (φ) is determined by an independent
method, e.g. from solar simulator measurements, for the same photon flux as the KPFM
measurements, β can be experimentally determined as:
β ∼= V partial fulloc (φ)− Voc (φ). (5.8)
For our calibration of β, φ corresponds to 1-sun illumination (see Figure 5.6).
Figure 5.5 shows the macroscopic characterization of a GaAs solar cell, without
an anti-reflection coating, used to calibrate and test this new metrology; a schematic of
the solar cell is presented in Figure 5.5(a). Figure 5.5(b) shows the external quantum
efficiency (EQE) measurement for the device under normal incidence illumination. The
dark and light J-V macroscopic measurements were performed using a conventional solar
simulator, where we adjust the light intensity to mimic different numbers of suns. The
electrical characterization of the device shows good diode behavior, with Voc = 0.83 ± 0.01
V under 1-sun illumination (see Figure 5.5(c)). From the intensity-dependent light J-V
curves we infer the Voc dependence with illumination (see Figure 5.5(d)), where the photon
78
Figure 5.5: Macroscopic characterization of GaAs. (a) Schematic of the solar
cell. (b) External quantum efficiency measurement. (c) Light J-V curves as a func-
tion of illumination (incident power). At AM 1.5G 1-sun illumination, Voc = 0.83±
0.01 V. (d) Open-circuit voltage Voc as a function of incident power (photon flux).
Figure 5.6: Voc vs. photon flux. Voc as a function of photon flux (φ) obtained
from solar simulator (black) and nanoscale KPFM measurements (green) for a GaAs
solar cell. (a) Voc of a fully (solar simulator) and partially (KPFM) illuminated
device. (b) Voc after correction for calibration factor β = −0.73 V, to compare with
fully illuminated devices. Note that both measurements were performed at the same
injection level, allowing for the direct comparison of the voltage dependence on the
incident light. The ideality factor, n, equals 1.9 ± 0.2.
79
Figure 5.7: Comparison of macro- and nanoscale Voc measurements. (a) AFM
topography image (peak to valley = 5 nm). (b)-(i) KPFM images (VCPD) for different
incident powers (light source: 660 nm laser). (j) Voc determined by macroscopic J-
V measurements (black) and by nanoscale KPFM (green) as a function of incident
power from the light sources. For the KPFM measurements, the Voc was obtained by
averaging over the scanned area, and the error bars refer to the standard deviation.
Inset shows uniform Voc map with nanoscale resolution.
flux corresponds to the AM1.5 global illumination, from 0.1 to 1.0-sun (100 mW/cm2).
Figure 5.7 shows a sequence of intensity-dependent KPFM measurements. The
GaAs solar cell is ideal to test this method because it is a very uniform sample (single
crystal) with minimum height variation (roughness is 5 nm, see Figure 5.7(a)). The CPD
(voltage signal) of the device increased with the incident laser power (photon flux) - see
Figure 5.7(b)-(i) - indicating that more carriers are generated upon increasing illumination,
as expected. In all cases, the voltage was very uniform because the epitaxial GaAs solar
cell has an extremely low density of recombination centers. [24] The uniform colored scans
shown in Figure 5.7(b)-(i) confirm the constant CPD signal of the cell, expected for a GaAs
monocrystalline device with extremely low density of defects. Therefore, the uniform color
of the scans indicates that the CPD does not vary within the spatial resolution of the
AFM probe (30 nm in diameter). Figure 5.7(j) shows the macroscopic Voc (from solar
80
simulator measurements) and the Voc signal determined by KPFM, both as a function of
the photon flux. For the KPFM measurements, the Voc is the averaged signal obtained
from the scans shown in Figures 5.7(b)-(i). The calibration constant is β = -0.73 V (see
Figure 5.6). The KPFM measurements are in excellent agreement with the solar simulator
ones: the ideality factor obtained from KPFM is n = 1.9 ± 0.2, which is within 10% (see
Experimental Methods). According to the KPFM measurements, the average Voc of the
solar cell at 1-sun illumination is Voc = 0.83 ± 0.02 V (spatially resolved in the inset
of Figure 5.7(j)). These results demonstrate that by illuminated-KPFM the macroscopic
value of the Voc for a device can be determined, and local variations in Voc can be measured
with nanoscale resolution. This can be applied to any PV material.
The illuminated-KPFM metrology is extended to a variety of solar cell devices,
allowing for direct imaging of local variations in Voc (i.e. ∆Voc) with nanoscale resolution.
The ∆Voc maps do not require system calibration; i.e. β is not needed. Figure 5.8 shows
the SEM, AFM, dark- and illuminated-KPFM measurements for all samples investigated
in this work: GaAs, monocrystalline Si, CdTe and CIGS, with morphologies ranging from
epitaxial to granular surfaces. GaAs presents a very uniform Voc (Figure 5.8(q)), with
spatial variations within the noise level of the KPFM measurements (∼10 mV).
The Voc map for Si (Figure 5.8(r)) is also very uniform, as expected for a monocrys-
talline material. Yet, we investigate the influence of shadowing effects on the Voc values
for the textured Si sample − see Figure 5.9. Briefly, we find that to compensate the shad-
owing effects of the textured facets we simply need to measure illuminated-KPFM under
two illumination conditions.
For CdTe (Figure 5.8(s)), different grain interfaces displayed unique Voc responses,
indicated by the dashed areas. This result suggests that the grain boundaries act as dis-
81
GaAs Si CdTe CIGS 
(a)	
   (b)	
   (c)	
   (d)	
  
5 µm 5 µm 5 µm 5 µm 
1.0 10 10 42005  (e)	
   1300 400 µm(e)	
   (f)	
   (g)	
   (h)	
  
0.8 nm 8 µm 4 nm 0 1 2 3 4 5
2
400
4 0
0.6 6 31
2 200
0
0.4 40 20 100
1
-200
10 10 -2 10 -1 100.2 200 nm 2 1-4 2 µm 520 µ0mnm  -400 2 µm 
150
0 0 -2 0 0 
01.0 10
0 2
.0 0 5
0.0 0.2 0.4 0.6 0.8 1.0 0 2 4 6 8 10 0 1 2 3 4 5
(i)	
   µ8 µm 1m.0V7 (j)	
   8 µm m8 0.5V3 (k)	
  
µm
8 m 1m.2V9 (l)	
  
µm
0mV .2V6 V mV 4 2000.8 8 0 1 2 3 4 5
520
41.8050 480
1.28
3
0 480 480
1.26
0.6 1.600 5030
0.95 1.24 250
480 1.22
6 0.4 6 0.90 24460 6 460 6 446 460 1001.20 mV1
0.85 460
1.18
0.2 0.280 1200 nm 0.75 2 µm 440 2 µm 1.16 2 µm 
440 440 1.145440 440 150
0 20.
4 1.
00 4 10
0 5 4
0.0 0.2 0.4 0.6 0.8 1.0 0 2 4 4 6 8 10 0 1 2 3 4 5
(m)	
   µm (n)	
   µm (o)	
   µm µm (p)	
  
0.8 V mV4820 44 V
200
20 420 420
1.05 520 3 1.28
1.600 3 1.260.6 500
2 2 0.95 2 2 1.24 250
04.9000 44800 410.202 400
0.4 4 2 4 mV
1.20
0.85 460
1.18
0.2800.2 1200 nm 0.72 2 µm 0.4344 0 2 µm 1.141 .162 µm 0.07 
0 0 0.75 0 0 5 1.14
0.0 0 0
0 2 4 0 20.0 06.2 0.40 8 0.6 20.8 102 15.00..02405 m V0 62 04 8 6 2 8 102 15000..40007 m 0V 61 2 8 3 4 102 5500..04007 m V 6 8 10  250.00 mV
µm
µm (q)	
   µm µm(r)	
   µm(s)	
   µm µm(t)	
  
0.11 
200 nm 2 µm 2 µm 2 µm 
 -17-00.0.01 7m V - 17-00.0.00 2m V  -17-00.0.00 6m V  -17-00.0.00 7m V
Figure 5.8: KPFM performed on four solar cell devices.(a)-(d) SEM images
of GaAs, monocrystalline Si, CdTe and CIGS solar cells. (e)-(h) AFM topography
images of the same samples. (i)-(l) Dark-KPFM images. (m)-(p) Illuminated-KPFM
images using 660 nm laser source with power output = 1 mW, which corresponds
to 1-sun illumination (100 mW/cm2). For Si sample, laser output = 10 mW. (q)-(t)
Open-circuit voltage variation (∆Voc) maps with nanoscale resolution. Dashed areas
indicate grains and interfaces with local Voc variations only resolved by illuminated-
KPFM.
82
µm
µm
µm µm
ΔVoc Light-KPFM Dark-KPFM AFM SEM 
µm
µm µmµm
µm
µm µm µm
µm
Topography	
  (nm)	
   Voltage	
  (V)	
   Voltage	
  (V)	
  
Figure 5.9: Directional independent KPFM imaging on Si. (a) Schematic
showing two directions of illumination for KPFM measurements on Si solar cells. (b)
AFM topography image. ∆Voc maps obtained from two distinct illuminated-KPFM
measurements (power out of the light source = 10 and 50 mW) for sample positioned
at (c) [100] and (d) [010] directions with respect to the AFM probe. (50 mW-10 mW)
voltage signal is equal to 0.152 ± 0.016 V and 0.125 ± 0.012 V, respectively. Image
(d) is rotated 90 degrees counter-clockwise to facilitate comparison with (c).
tinct recombination centers. The CIGS grains present remarkable variations in Voc (>200
mV), independent of the material topography, see dashed regions in Figure 5.8(t). Despite
the fact that the different facets of a crystal can present distinct work functions, [259] the
difference is taken into account because both the light- and dark-KPFM voltage signals are
measured for the same facets of the crystal. The analysis of the role of the grain interfaces
on the performance of the device is beyond the scope of this paper; however, these mea-
surements demonstrate that nanoscale analysis of Voc in non-uniform materials can resolve
and quantify the recombination processes that currently limit the performance of polycrys-
talline solar cells. [208] Controlled growth of grains with crystal orientation corresponding
to the high values of ∆Voc (∼0.15 V) could potentially boost the overall Voc of the device.
Despite the fact that CIGS grains grow preferentially along the 〈112〉 orientation, [167]
a distribution of grains is usually observed, and to the best of our knowledge there is no
83
systematic study correlating different grain orientations with device performance. Note
that while the voltage signal increases with illumination for GaAs, it decreases for Si and
CdS/CIGS because the scanned layers are p- and n-doped, respectively. For the p-type
CdTe surface, the voltage signal decreases with illumination because the work function of
CdTe (∼5.9 eV) is larger than the work function of the Pt-coated probe (∼5.1 eV) used
in the measurements, modifying the band diagram configuration presented in Figure 5.4.
The possible influence of major experimental artifacts was carefully analyzed, and no
significant contribution to our measurements was found. A detailed study of the influence
of ∆H (and therefore topography) on the SPV signal was performed and is presented in
Figure 5.10. In order to minimize the effect of topography on the SPV signal, ∆H was
kept constant for all dark and illuminated measurements presented here. Unless indicated
otherwise, the probe passed at ∆H = 20 nm above the topographic scan on the second
pass.
84
Figure 5.10: Influence of ∆H on voltage response. (a)-(c) AFM images of GaAs
solar cell and (d)-(f) corresponding KPFM showing topography influence on voltage
signal as a function of the difference in sample-probe distance between AFM and
KPFM passes (∆H). Illuminated-KPFM using incident power = 10 mW. At ∆H =
-60 nm the topography signal clearly contributes to the voltage scan even for the
case of a smooth sample. At ∆H = 80 nm the probe is too far from the surface and
the variation in the voltage signal is negligible because this situation is similar to a
macroscopic Kelvin probe measurement. (g) Average surface photovoltage (SPV) as
a function of ∆H.
We resolve local variations in Voc with spatial resolution <100 nm. Figure 5.11(a)
shows a ∆Voc map for the CdTe solar cell, showing a few interfaces that have reduced Voc
and, most likely, act as non-radiative recombination centers. The full-width-half-maximum
(FWHM) of the line profiles shown in Figures 5.11(b)-(d) for the lateral variation in ∆Voc is
consistently below 100 nm. This set of KPFM measurements was acquired using Pt probes
with 30 nm in diameter, demonstrating that we can accurately measure and resolve the Voc
85
Figure 5.11: Quantifying KPFM spatial resolution. (a) ∆Voc map for CdTe
solar cell. (b)-(d) Line scans showing consistent spatial resolution <100 nm. The
full-width-half-maximum that defines the lateral resolution of the method is indicated
in each line profile.
with truly nanoscale resolution. Higher spatial resolution can be achieved by using smaller
probes; such as high aspect ratio tips with gold nanoparticles on standard Si tips [260] or
Si probes with 2 nm in diameter. [261]
For polycrystalline PV materials, the direct correlation between the structural prop-
erties of different grains and the Voc maps can be used to identify the types of grains which
should be favored during the thin-film deposition or growth. Further, nanostructured
PV can tremendously benefit from this new technique because the identification of non-
radiative recombination centers can guide the design of the next generation nano-enabled
devices. It is well known that the Voc of a PV device can strongly depend on surface
passivation, and therefore spatial variations in the surface properties can lead to local
variations in the Voc, which can be accurately measured with our technique. Further, a
86
recombination center, which corresponds to a region with very small diffusion length, can
be accurately measured by our imaging method. Although the carrier minority diffusion
length of inorganic PV materials is greater than 100 nm, it is an average of the distance
a carrier can move from point of generation until it recombines. Illuminated-KPFM does
not require the preceding knowledge of the sample nor the probe work function because
the difference between the work function of the material surface and the probe under
illumination and in the dark suffices to determine local variations in Voc.
5.4 Conclusion
In this chapter, we presented a new method to spatially resolve and image the Voc
in partially and fully processed optoelectronic devices with nanoscale resolution, <100
nm. The Voc is a key figure of merit to determine the performance of a device; it is an
indirect measurement of the recombination processes that take place within the material.
The illuminated-KPFM technique is demonstrated in a variety of photovoltaic materials,
ranging from epitaxial GaAs to polycrystalline CIGS, where we can resolve Voc variation
larger than 200 mV at the nanoscale. This metrology, introduced here for the first time,
can be applied to any optoelectronic device, including photovoltaics, LEDs, and photode-
tectors, does not require full device processing, is non-destructive and works in ambient
environment. Moreover, it can be used to determine if specific device processing steps are
beneficial or harmful for the ultimate performance of a functional semiconductor material.
87
88
Chapter 6: Real-time nanoscale open-circuit voltage dynamics of
perovskite solar cells
Hybrid organic-inorganic perovskites are an emerging material with great po-
tential for high-performance and low-cost photovoltaics. However, for perovskites to
become a competitive and reliable solar cell technology, their instability and spatial
variation must be understood and controlled. While the macroscopic characteriza-
tion of the devices as a function of time is very informative, a nanoscale identification
of their real-time local optoelectronic response is still missing. Here, we implement
a 4D imaging method through illuminated heterodyne KPFM to spatially (<50 nm)
and temporally (16 seconds/scan) resolve the voltage of perovskite solar cells in a
low relative humidity environment. Local Voc images show nanoscale sites with volt-
age variation >300 mV under 1-sun illumination. Surprisingly, regions of voltage
that relax in seconds and after several minutes consistently coexist. Time-dependent
changes of the local Voc are likely due to intragrain ion migration and are reversible
at low injection level. These results show, for the first time, the real-time transient
behavior of the Voc in perovskite solar cells at the nanoscale. This chapter is adapted
from J.L. Garrett, E.M. Tennyson et. al, Nano Letters, 17 2554-2560 (2017) [39]
a1er	16s	 VOC	(V)	
+0.20	
+0.10	
a1er	557s	
-0.10	
400	nm	 -0.20	
89
!me	
6.1 Introduction to nano-imaging in perovskite solar cells
Recent improvements in power conversion efficiency of perovskite solar cells have
brought this class of materials to the same performance level as thin-film CdTe and CIGS
photovoltaics, demonstrating its promise as a large-scale technology. [262–265] To date,
the world record single-junction perovskite device has a power conversion efficiency of η
= 22.7%. [35] Moreover, perovskites are ideal for composing dual-junction designs with a
Si or a CIGS bottom subcell. [266, 267] The remarkable performance of perovskite solar
cells is primarily attributed to their high absorption combined with extremely long carrier
diffusion lengths, [268] including one literature report of >175 µm. [269] Despite the rapid
progress of the field, there is still a lack of understanding of why and how the material
is changing/degrading when exposed to light and humidity. [50, 270] Thus, there is a
pressing need in the scientific community to elucidate what causes the perovskite active
layer to change when the device is operating [271] (exposed to light). Macroscopic electrical
measurements, such as light J-V curves as a function of time, are extremely useful to track
the device instabilities; [272,273] however, one cannot infer how much variation exists from
grain to grain and within one grain in the perovskite layer, which is needed to determine
how the local changes affect the overall device performance.
Microscopy measurements provide a powerful framework for correlating the mor-
phology of the active layers of the solar cells with fundamental physical processes and
device performance. [37, 40, 130, 131] In particular, scanning microscopy techniques are
ideal for investigating heterogeneity at the nanoscale, [51, 90, 138] such as the spatial do-
mains on MAPbI3. [51] Recently, micron-scale photoluminescence has been applied to
resolve charge carrier recombination and ion migration in perovskite grains. [41, 138, 274]
90
Through piezo force microscopy, the ferroelectric domains of the perovskites have been
observed. [275,276] By time-dependent Kelvin probe measurements it has been suggested
that the hyterisis behavior commonly observed in perovskite devices could be related to
trapped charges and ion migration across the light absorbing layer. [277] Kelvin probe force
microscopy (KPFM) [278,279] has been used to map the electrostatic potential across mul-
tiple layers of active devices, [151, 280] to measure work function changes during device
poling, [281] and to probe the effect of grain boundaries on MAPbI3. [282] Further, KPFM
has been applied to investigate the possible origins of the hysteresis commonly observed in
light J-V measurements. [69] Despite these advances, traditional high-resolution variants
of atomic force microscopy (AFM) operate too slowly to spatially analyze dynamic events,
such as structural changes, light induced chemical reactions, and ion migration.
Here we image, in real-time, the dynamics of perovskite solar cells with nanoscale
spatial resolution (<50 nm) by mapping local changes in the open-circuit voltage (Voc)
through high-speed illuminated-KPFM (16 seconds/scan), see Figure 6.1(a). For that, we
implement a novel KPFM method that preserves the spatial sensitivity while increasing the
scan speed by >100 times when compared to conventional frequency-modulated-KPFM,
where we map nanoscale voltage domains that vary by >300 mV under 1-sun illumina-
tion. The voltage within a perovskite grain changes substantially (both spatially and
temporally) post light exposure, where regions out of and in equilibrium co-exist, and
are attributed here to intragrain ion migration. We determine the spatial distribution of
the local photogenerated voltage, how it varies as a function of time, and we quantify its
dynamic behavior as the perovskite solar cell reaches its steady state: over 9 minutes. A
time-dependent residual Voc that lasts for minutes is found, resulting from migration, even
under dark conditions. Our results demonstrate, for the first time, the transient electrical
91
behavior of perovskite solar cells with nanoscale spatial resolution.
6.2 Experimental methods
6.2.1 Perovskite solar cell fabrication
The solar cells are fabricated by the sequential spin-coating deposition of the per-
ovskite layer (CH3NH3PbI3) and the [6,6]-phenyl C61-butyric acid methyl ester (PCBM)
layer. The perovskite layer is fabricated from lead (II) iodide (PbI2) and methyl ammo-
nium iodide (MAI), each deposited on a glass/ITO/PEDOT substrate by spin coating at
6000 rpm for 35 seconds. After deposition, the sample is thermally annealed at 100◦C in
a nitrogen dimethylformamide environment, as described elsewhere. [283] A 20 nm layer
of PCBM is then deposited on top of the perovskite via spin coating at 6000 rpm for 35
seconds, and the device is annealed at 100◦C for 60 minutes. This annealing process is
implemented to passivate the perovskite and minimize photocurrent hysteresis. A 20 nm
layer of C60 is then thermally evaporated to further aid the passivation of the active layer
of the device. The macroscopic characterization of the devices (light J-V and EQE) is
performed under nitrogen environment to prevent material degradation. The band dia-
gram for the device is presented in Figure 6.1(b), where the approximate values of the
energy gaps were obtained from refs [264,284–288]. Here, the PCBM layer functions as the
electron transport material (ETM), while the PEDOT layer is the hole transport material
(HTM), where the holes are preferentially collected through the indium tin oxide (ITO)
electrical contact. The perovskite absorbing layer with nominal thickness of 300 nm is
sandwiched by the PEDOT and the PCBM (see Figure 6.2). The perovskite films are
composed of grains ≈1 µm in size and with a low density of pinholes, with Voc = 1.02 ±
92
(a) (b)
Evac
-1.8
e-
-3.9 -4.1
KPFM 
circuit -4.7 ≅ -5.0
+ -5.2h -5.4
h+ -5.8
ITO PEDOT CH3NH3PbI3 PCBM
0 0.1 0.2 0.3 (HTM) absorber (ETM) probe
Voltage (V)
Figure 6.1: Real-time nanoscale measurements of voltage dynamics in per-
ovskite solar cells. (a) Schematic of hybrid perovskite solar cells measured by
illuminated-KPFM, out of scale for clarity. The local voltage is overlaid with topog-
raphy, and the green arrow indicates the illumination direction. (b) Energy diagram
of perovskite device and AFM probe illustrating the charge separation mechanism
within the hybrid solar cell under illumination. Evac refers to the vacuum energy
level, and all energy values are in units of eV. The ITO layer is grounded with respect
to the KPFM probe; thus, the subtraction of an illuminated- from a dark-KPKM scan
provides a map of the open-circuit voltage of the device. ETM electron transport
material; HTM hole transport material.
0.02 V (see Figures 6.2 and 6.3 for macroscopic device characterization and SEM images).
6.2.2 Illuminated Kelvin-probe force microscopy Voc derivation
Kelvin probe force microscopy is a scanning probe method based on atomic force
microscopy (AFM) that measures surface potential at nanometer length-scales and utilizes
AFM to maintain small tip-sample separation, as reviewed in depth in Chapter 3. An AC
voltage is applied to a conductive AFM probe, which in turn causes the probe to oscillate,
generating a signal proportional to the difference between the voltage of the probe and
the surface. A feedback loop then applies a DC voltage (contact potential difference,
VCPD) to the probe in order to minimize the KPFM signal, which originates from the
cantilevers oscillation. The contact potential difference is defined to be VCPD = Vprobe-
Vsample. We followed the calibration procedure described in Ref [155] to determine the
93
Figure 6.2: Macroscopic characterization of perovskite solar cell device. (a)
Cross-section schematic of perovskite solar cell (out of scale for clarity). (b) Light
J-V performed under 1-sun AM1.5G illumination in reverse bias. Voc = 1.02 V,
J 2sc = 17.72 mA/cm , FF = 71.56%, η = 12.91%. Inset: device Voc variation as a
function of time, where t = 0 refers to the first light J-V measurement acquired, just
after device fabrication. (c) External quantum efficiency (EQE) measurement for the
same device.
Figure 6.3: SEM images of perovskite thin film. (a,b) Scanning electron mi-
croscopy images of PCBM layer.
94
KPFM setpoint, avoiding topographical imprint on the KPFM data. [255] Here, the in-
plane spatial variation of VCPD is equal to the spatial variation of the surface potential,
because Vprobe does not depend on probe position.
Two KPFM techniques are used throughout the results presented here: heterodyne
(H-) KPFM and frequency modulation (FM-) KPFM. H-KPFM is a newer technique,
first developed to increase the sensitivity of KPFM in vacuum, while retaining the spa-
tial resolution of FM-KPFM, [155] but it has recently been shown to operate quickly in
air, reaching scan speeds of several frames a minute. [289] Here it is used in single pass
mode, to map dynamic processes. It acquires each (128)2 pixel image in 16 seconds and
achieves a per-pixel signal standard deviation of about 30 mV. The sequential fast H-
KPFM measurements are acquired by alternating top-bottom and bottom-top scans, and
no correlation between the scan direction and the voltage signal is detected. FM-KFPM
is used to scan rougher regions of the sample, which are observed to disturb the H-KPFM
signal. It is configured in the side-band implementation with a modulation frequency of
800 Hz, [278] in the two-pass lift mode; topography is measured on the first pass and the
probe is lifted 15 nm above to record the surface potential. The KPFM loop error signal,
and 2f-KPFM signal, which is proportional to VAC [155] are also collected, to verify proper
functioning of the KPFM loop (see Figure 6.4). Lift mode helps to maintain stability of
the FM-KPFM signal on rough topography. Each FM-KPFM image takes about 30 min-
utes to record, and contains (256)2 pixels. The standard deviation per-pixel due to noise
of FM-KPFM is ≈20 mV, whether estimated by sensitivity measurements or by sampling
a uniform region of the potential.
The Voc map is obtained using the following analysis, as discussed in ref [37] and
the previous chapter:
95
Figure 6.4: Monitoring all relevant signals during FM-KPFM measure-
ments. (a) AFM topography shows a homogenous texture of grains. (b) Kelvin
probe voltage (VCPD) detected by a feedback loop at a frequency fmod = 800 Hz
from the first resonance. (c) Error signal in the feedback loop showing extremely
small variations (∼2 mV), which demonstrates the reliability of the method. (d)
In FM-KPFM, a second signal at twice the modulation frequency (2fmod) is also
recorded and is proportional to the tip-sample capacitance. The scan shows how the
sensitivity of the KPFM signal varies spatially.
( ) ( )
V partial
nkT Isc nkT Jsc ×Ail
oc = ln + 1 = ln + 1q Idark ( ) q Jd(ark ×Acell ) (6.1)
nkT Jsc nkT Ail Jdark
= ln + ln +
q Jdark q Acell Jsc
( )
V full = V partial − nkT Ail Jdarkln + = V partial ′oc oc − β (φ) (6.2)q Acell q(EQE)φ oc
( )
′ nkT Ail Jdark
β (φ) = ln + (6.3)
q Acell q(EQE)φ
Here, the superscripts partial and full refer to the illumination conditions, n (=
1.34) is determined from the slope of Figure 6.7(f), Jsc and Jdark are the short-circuit and
′
the dark current densities, respectively, and β (φ) depends on the illumination conditions
(photon flux, φ) of the measurements. At 1-sun (φ= φ(1-sun)), the short-circuit current
term is given by: Jsc = q(EQE)φ(1− sun), which is determined by a light J-V measure-
96
Figure 6.5: Controlling the environmental conditions during in situ imag-
ing of device performance. Average value of voltage (VCPD) for a sequence of
alternating dark- and illuminated-KPFM scans as a function of time (red). The rela-
tive humidity (light blue) and the temperature (navy) of the environment inside the
AFM chamber were both precisely controlled and monitored while performing the
KPFM measurements.
′
ment (see Figure 6.2). When the device is under full illumination: β = 0, as Ail = Acell
and Jdark << Jsc.
6.2.3 KPFM implementation
For the measurements presented here, a Cypher AFM (Asylum Research) is used,
with HQ:CSC37/Pt-B probes (Pt-coated Si-doped probe with <30 nm in radius and 0.3
N/m force constant, µmasch). Dry air is pumped into the microscope to maintain a low
relative humidity (<15%) and constant temperature (see Figure 6.5). The topography
of the PCBM layer is observed to be stable during the time scale of our measurements
(Figure 6.6). The ITO layer (bottom contact) is grounded with respect to the conductive
probe for the KPFM measurements. A supercontinuum laser (WhiteLase, Fianium) with
variable power density is used for the illuminated measurements at 500 nm wavelength,
and neutral density filters are used for the smallest illumination intensities (<30 nW). FM-
KPFM was used to acquire Figures 6.1, 6.7, and 6.8(b), while H-KPFM is implemented
for Figures 6.8(c)-(h), 6.10, and 6.11.
97
Figure 6.6: AFM topography images. (a)-(f) The morphology of the PCBM
layer does not change within the time interval of the fast H-KPFM measurements
(corresponding to Figure 6.8(c)-(h)).
We implement two types of KPFM to spatially resolve nanoscale changes of the
voltage in the perovskite solar cell. KPFM measures the contact potential difference
voltage signal (VCPD), which is proportional to the difference between the work function
of the conductive probe and the surface of the sample being scanned. To map the local
voltage response in large areas of the device, frequency-modulated KPFM [290–292] can
be used. To increase the scan speed of AFM techniques, common approaches include the
use of polymeric cantilevers [293] or to reduce the size of the cantilevers while imaging
the sample in an aqueous solution. [294] Although these methods are extremely useful to
probe the topography of a variety of materials, they cannot be implemented to accurately
measure voltage. Thus, to investigate the perovskites electrical properties in real-time
upon illumination, we realize a variant of heterodyne-KPFM, which enables fast scanning
(16 seconds/scan) while maintaining high spatial resolution. [155] Briefly, the frequency
at which the voltage is detected is separated from the frequencies of both the topography
98
feedback loop and the applied voltage [155, 289] (see Figure 6.4 for maps of all relevant
signals).
6.2.4 Image analysis
The KPFM images are analyzed using Igor Pro software. We used a linear drift
correction software code, [295] based on the correlation of topography scans for the calcu-
lations of Voc.
6.3 Fast-KPFM results and discussion
In order to better understand the voltage changes that occur under illumination, we
perform light intensity dependent in situ measurements of the photogenerated voltage on
a large, 16 µm2, area of the perovskites using KPFM. Figure 6.7(b)-(e) shows a sequence
of illuminated-KPFM images from 2 nW to 54 µW (corresponding to low injection level).
As the power of the incident light increases, so does the magnitude of the average VCPD
(despite the local variations in voltage, as discussed later in this manuscript). Figure 6.7(f)
shows the expected linear increase of the averaged change in VCPD under illumination, as
a function of light intensity.
As derived in the Experimental Methods section, KPFM has been recently demon-
strated as a novel method to map the local Voc of half and fully processed photovoltaic
devices with nanoscale spatial resolution, where Voc is the voltage across a solar cell at
open-circuit condition. In the illuminated- and dark-KPFM measurements presented here,
the voltage on the conductive probe is referenced to the grounded ITO bottom contact.
As a result, when the device is illuminated the splitting of the quasi-Fermi level (propor-
99
(a) dark (b) 0.01-suns (c) 0.13-suns (d) 1-sun (e) 18-suns 
0.50
0.25
1 μm 0.0
 0.30 a.u.
0.4 (f) (g) 0.74
0.3 150 nm
0.2
0.1
0.0
0.42
10-2 10-1 100 10 102
Incident power (µW)
Figure 6.7: Imaging nanoscale variations in the Voc of perovskite solar cells.
(a) Dark-KPFM and (b-e) intensity dependent illuminated-KPFM scans measured on
a perovskite device showing local variations in voltage. (f) Scan-averaged illuminated
minus dark voltage (V il -V darkCPD CPD) as a function of incident power (light intensity).
The error bars refer to the standard deviation of the averaged values, the green line
is the best fit, and the green region represents two deviations from the mean. (g)
Absolute Voc map with nanoscale spatial resolution at 1-sun illumination, overlaid
with topography.
tional to the Voc of the solar cell) is directly measured by subtracting an illuminated- from
a dark-KPFM scan. Here, we illuminate a well-defined area of the solar cell at normal
incidence and, thus, the local Voc can be expressed as: [37]
V (x, y) = [V il (x, y)− V dark − ′oc CPD CPD(x, y)] β , (6.4)
where (x,y) are the spatial coordinates, il and dark refer to illuminated and dark condi-
tions, respectively, and
( )
′ nkT Ail Jdark
β = ln + , (6.5)
q Acell Jsc
at 1-sun illumination, where n is the ideality factor (obtained from the slope of the curve
shown in Figure 6.7(f)), k is Boltzmann’s constant, T is the temperature, and Acell = 1.0
cm2, Ail = 2.0 × 10−3 cm2 are the total cell and illuminated areas, respectively, and Jdark
100
V il −V darkCPD CPD (V )
0.01-suns
0.13-suns
1-sun
18-suns
VCPD (V)
VOC (V)
′
and Jsc are the dark current and short-circuit current densities. The coefficient β (=
-0.22 ± 0.03 V) is a calibration factor that depends on the illumination conditions and on
′
the area of the device being probed. Note that when the whole cell is illuminated: β =0,
as Ail = Acell and Jsc >> Jdark. Figure 6.7(g) shows a representative Voc map for the
perovskite solar cell under 1-sun local illumination, which exhibits spatial variations >300
mV. These remarkable changes in photovoltage demonstrate that the materials electrical
characteristics vary locally from grain to grain and within a single grain. The average value
of the Voc measured by KPFM is smaller than the initial value obtained by macroscopic
light J-V measurements immediately after fabrication (Figure 6.2), which primarily results
from the Voc decay of the sample at a rate of ∼5% per day (see the inset of Figure 6.2).
Figure 6.8(a,b) show representative topography and KPFM scans in a 16 µm2 region
of the perovskite device in the dark (at equilibrium). The VCPD varies spatially with a
standard deviation of 80 mV (substantially greater than the noise) and is independent of
the morphology of the sample. These spatial variations could be due to inhomogeneities
either within the perovskite or within the PCBM layer, caused by the spin coating de-
position process used to fabricate the sample. During all measurements presented here,
the temperature was kept at 28-29 oC and the relative humidity <15% inside the AFM
enclosure, see Figure 6.5 for stability of the environment inside the AFM chamber during
dark and illuminated measurements. Figure 6.8(c)-(h) show the in situ changes in the
voltage before, during and after illumination, obtained by fast-KPFM (16 seconds/scan),
with 54 µW of incident power at wavelength of 500 nm. According to the KPFM initial
dark scan, areas of both high and low surface potential co-exist (Figure 6.8(c)), with an
average VCPD of 0.2 V. Upon illumination, Figure 6.8(d), the surface potential increases
substantially, to 0.4 V on average, as a result of the photogenerated voltage. No change
101
 e g at l o V ) b( (a) Topogr ayphhpya r g o p o T ) a( (b()c V) oinltaitgiael  dark (d) illuminated (i) 
h 
200 nm 
 mµ 1 1 µm (e) tdark= 16 s (f) tdark= 98 s 
 
-200 200  
 3. 0-  2. 0-  1. 0-  0. 0  200002-  100001 -   00  - 100001  -200002 0.0 -0.1 -0.2 -0.3 
 ) V(  V H e)eigimghnh(t  tt( nh(nmgim)e ) 
g 
H VK (V) K dark 
(a) Topography (b) Voltage 
(g) tdark= 263 s 
f 
(h) tdark= 557 s  
 
e 
d 
ill 
1 µm K
Height (nm) V  (V) 
dark 
200 100  0 -100 -200 0.0 -0.1 -0.2 -0.3 c 
200 100  0 -100 -200 0.0 0.0 -0.1 -0.2 -0.3 0.3 0.0 0.2 0.4 0.6 
Height (nm) 1 µm VVVK (V
0.0 0.6 
CCPPDD ( 
)V( V) ) V  (V) VCPD (V) CPD
Figure 6.8: Dynamics of perovskite solar cells at the nanoscale. (a) Topogra-
phy and (b) dark-KPFM measurements on a representative area of a perovskite pho-
(a) Topography (b) Vtooltvagoelt aic device. A sequence of fast-KPFM measurements were acquired on the region
highlighted by the white dashed square in (a,b): (c) dark-KPFM, (d) illuminated-
KPFM, (e-h) dark-KPFM scans as a function of time. Note that after exposure to
light (scan (d)), it takes the material ≈ nine minutes to return to equilibrium (scan
(h)). (i) Histograms of voltage distribution as a function of time for 42 KPFM scans
(dark and illuminated) showing reversible dynamics upon a sequence of dark scans.
Illumination conditions: 500 nm laser light at 54 µW. Relative humidity <15%.
in topography is observed during the KPFM scans (see Figure 6.6), indicating that these
perovskite devices are morphologically stable during this time frame, under the illumina-
tion conditions and the environment used in this experiment. Surprisingly, after turning
the light back OFF, a transient decay of the voltage is observed, see Figure 6.8(e)-(h).
Over time, the voltage returns towards its initial profile, where it takes ≈9 min for the
material to achieve its original surface potential spatial distribution (Figure 6.8(h)).
Through the voltage histograms displayed in Figure 6.8 we quantify this complex
transient behavior of the perovskite’s voltage. Each pixel in a 128 × 128 pixel map is
binned according to its voltage value to produce the histograms. Note how the spatial
102
Counts 
distribution of the voltage changes with time until the device reaches equilibrium again. In
particular, the histograms reveal that the initial dark voltage distribution (blue) is quite
broad, with a peak around 0.2 V, and under illumination this distribution increases to
0.4 V (green curves). However, after turning the laser OFF, the potential immediately
becomes a single, fairly narrow Gaussian (light grey curve). As time progresses, the
VCPD broadens again, primarily by extending to higher voltages (dark grey curves). This
behavior is also observed after exposing the perovskite to different illumination intensities
(ranging from 2.0 nW to 4.2 mW, see Figure 6.9). These results demonstrate that after
exposing the perovskite to light, its voltage changes as a function of time even under dark
conditions.
Another striking observation about these perovskite solar cells is that upon removing
the illumination, the Voc does not immediately return to zero. Instead, a non-zero average
voltage persists for several minutes, with a sign is opposite of the Voc under illumination.
This behavior is consistent with trap filling and subsequent ion migration, which have
been observed to take minutes to equilibrate on the macroscale. [49, 296] To quantify the
time-dependent voltage response within the perovskite solar cell, we analyze the spatial
variation of this residual Voc obtained by subtracting the equilibrium dark-KPFM image,
Figure 6.8(h), from the post-illumination dark-KPFM scans, see for example Figure 6.8(e)-
(g). Figure 6.10 displays the residual Voc images for the perovskite solar cell for a series
of snapshots. The dark measurement immediately after turning the laser OFF shows
a transient electrical behavior (Figure 6.10(a)), most likely resulting from a reversible
electronic process within the material. The representative scans shown in Figure 6.10(a)-
(e) have two very distinct regions, with initial Voc = - 0.2 V (red) and Voc = 0 V (blue),
revealing a strong bimodal voltage distribution on the electrical response of the device,
103
Figure 6.9: Voltage profile for fast H-KPFM illumination scans. Histogram
of VCPD for sequential H-KFPM scans showing signal variation as a function of
illumination, using 500 nm (green curves) and under dark conditions (grey/black
curves). Each histogram corresponds to one scan, acquired in 16 seconds in the
same 1.0 µm2 area. The incident illumination power is listed on the right. Lines
representing dark (light) scans become darker (lighter) the longer the light is OFF
(ON), as shown in grey and green, respectively.
104
not resolved by conventional macroscopic electrical measurements or by standard AFM-
based methods. As the system approaches equilibrium, the Voc in the initially red region
approaches zero, see voltage scans in Figure 6.10(b)-(e), while the blue regions remain
unchanged. Figure 6.10(f) shows the averaged Voc as a function of time for the entire
scanned area shown in (a)-(e) (black curve) and the voltage for the regions highlighted
by the white squares (red and blue curves). Each point on the graph corresponds to the
averaged value (either within the white squares or for the entire scanned region). Here, the
variance is calculated by considering each pixel in each box as an individual measurement
with 30 mV standard deviation, and it is plotted at three standard deviations. The
blue curve is constant under dark conditions (laser OFF), indicating that either some
regions of the PCBM layer do not suffer any post-illumination driven electrical changes
or that these regions return to equilibrium in less than 16 seconds. Nevertheless, the red
curve shows that an out-of-equilibrium state persists for several minutes. These time-
dependent measurements reveal that the voltage equilibration has a spatial dependence
at the nanoscale that could be responsible for the light-dependent instability observed in
perovskites. [297]
Because no net current flows out of the device at open-circuit conditions, an addi-
tional current term must balance the diode dark current (Idark) to enable a residual voltage
to exist without illumination. We explain this behavior using a simple diode model with
an additional time-dependent current term, Iion(t), which could arise as a result of ion
migration and/or a change in the density of trap states:
( )
qV
I(V ) = IL − Idark enkT − 1 + Iion(t), (6.6)
where IL is the light-generated current of the solar cell. In the dark (IL = 0), and at
105
(a) (b) 5 s (c)164 (d)32 (e) 41
30 
0 
-30 200
0 
2 2 2 2 2
-0.1 1 1 1 1 1
-0.2 200
0.1 1.0 (f) (g)
0.8 
region
0 0.6 
scanned
-0.1 0.4 
region 0.2 
-0.2 a b c d e 0.0 
16      32           64 128 256  512 0   200  400
t - toff (s) t - toff (s) 
Figure 6.10: Spatial and temporal variation of the residual photogenerated
voltage within perovskite solar cells. Topography and Voc scans acquired after
the solar cell is illuminated and brought back to dark conditions after (a) 16 s, (b) 65
s, (c) 164 s, (d) 328 s and (e) 541 s. Note that immediately after illumination, two
distinct regions are observed, one with Voc equal zero (blue) and another with Voc =
-0.2 V (red). (f) Averaged Voc as a function of time for the entire scanned area shown
in (a-e) (black), and for regions 1 and 2 (highlighted by the white squares). The error
bars refer to three standard deviations from the averaged values. The vertical grey
dashed lines correspond to the time interval where the scans (a-e) are performed. (g)
Ratio of current due to ion migration (Iion) and dark current (Idark) as a function
of time. Black line is an exponential fit to the experimental data resulting in a time
constant of 185 ± 27 s.
106
Average Voc (V) Voc (V) Height (nm) 
| Iion/Idark | 
open-circuit conditions, we have
( )
qVoc(t)
Iion(t) = Idark e nkT − 1 , (6.7)
where Voc(t) is the time-dependent residual open-circuit voltage. Post-illumination, Iion(t)
opposes the original dark current and tends toward zero on the time scale of minutes.
Figure 6.10(g), shows the decay of Iion(t) as the device reaches equilibrium, resulting in a
time constant of 185 ± 27 s.
As one of the primary limiting factors of perovskite photovoltaics is the material
instability when exposed to light, we also measure the time-dependent local changes in
Voc under illumination, as presented in Figure 6.11. Here, we determine ∆Voc by the
difference between two illuminated KPFM maps, where each resulting scan corresponds
to an instant of time subtracted from the last illuminated one in Figure 6.8(i), in light
green. The sequence of ∆Voc maps show the local spatial variation in the photogenerated
voltage. As previously observed, [296] when the perovskite material is illuminated, the
trap states due to vacancies are filled by electrons. As a result, the ions move away
from these sites (due to electrostatic repulsion). We track this ion migration process,
in real-time, by spatially resolving changes in the local voltage (∆Voc) that take place
during material illumination. We hypothesize that ions migrate away from the PCBM
layer (electron transport layer) due to the built-in electric field. Regions with ∆Voc > 0
and <0 co-exist, strongly indicating that there is an accumulation of ions and vacancies
within the illuminated area of the cell, which varies in time. At low injection level,
this migration process is reversible and does not lead to material degradation. Yet, it
changes the electrical behavior of the device locally, which ultimately affects the overall
performance of the solar cells.
107
time 
(a) After 16s (b) After 32s (c) After 48s (d) After 64s (e) After 80s 
500 nm 
-100 -50 0 +50 +100 
ΔVOC (mV) 
Figure 6.11: Real-time nanoscale Voc dynamics of perovskite photovoltaics.
Sequence of ∆Voc maps as a function of time after (a) 16 s, (b) 32 s, (c) 48 s, (d)
64 s, and (e) 80 s of illumination. Each scan is subtracted from an illuminated one
after 96 s (from 6.8i). The blue and red dashed areas highlight regions with transient
voltage response. The scans size is 1 × 1 µm2.
We attribute the dynamic behavior of the local electrical response of the perovskite
solar cell mapped here to intragrain ion migration: [41, 282] charge accumulation takes
place when the material is illuminated, and charge migration when the solar cell is no
longer in operation (dark conditions). This photo-induced ion migration is primarily
driven by the movement of to regions with depleted ions, as observed in prior photolu-
minescence experiments. [41] We demonstrate that this phenomenon is fully reversible at
low illumination injection level (Figure 6.7) and does not result in permanent material
degradation. Further, the spatial variations in the local electrical response are not accom-
panied by any changes in morphology (see Figure 6.6 for sequence of topography maps
during material illumination), but the nanoscale time-dependent variations are most likely
responsible for macroscopic instabilities observed in these perovskite solar cells. The diffu-
sion of ions under dark conditions leads to a time-dependent local residual Voc(t), mapped
here for the first time. Our results indicate that free carriers are probably responsible for
the observed transient voltage. Once the material is under dark conditions again, some
ions continuously move until the perovskite layer achieves an equilibrium state, which
108
takes place in several minutes for the material probed in this work (as shown in Figure
6.8).
6.4 Conclusion
The real-time functional imaging method reported here can be combined with
other scanning probe microscopies and big data analysis [298] to probe additional rel-
evant phenomena that currently describes the unique illumination dependent behavior
of perovskites, such as structural transformation, [299] ferroelectric domains, [275, 300]
poling, [296] and piezoelectric behavior. [275] For instance, to map structural and chemi-
cal composition changes, fast-KPFM could be combined with tip-enhanced Raman spec-
troscopy. Likewise, to investigate the possible effects of long lived traps [301] on the devices
local electrical response, our novel imaging method could be combined with pump-probe
EFM/KPFM techniques, [302, 303] which measures voltage at microsecond timescales.
Further, spectrally dependent fast-KPFM could help elucidating the mechanism respon-
sible for light-induced self-poling in perovskites.
In summary, we imaged and quantified the real-time voltage dynamics of perovskite
solar cells at the nanoscale (<50 nm) under illumination and post-illumination conditions
by fast KPFM (16 seconds/scan) in a low relative humidity environment. 1-sun local Voc
maps of the perovskites revealed absolute variations >300 mV between and within grains.
Under illumination, the photogenerated voltage of the perovskite varied spatially within
one grain. After illumination, it took the material ≈9 min to reach its equilibrium state.
This transient electrical behavior, most likely resulting from reversible ion migration within
the perovskite layer, cannot be revealed by conventional macroscopic electrical measure-
ments. The real-time nanoscale imaging experiments presented here demonstrate that the
109
Voc of perovskite photovoltaics varies locally, and that light-induced reversible ion migra-
tion could be responsible for the time-dependent transient behavior of most perovskite
photovoltaics. Substantial effort lies ahead to acquire a full picture of the mechanism(s)
responsible for the light-induced physical and chemical processes frequently observed in
this class of materials, and their effects on the maximum attainable Voc. Nevertheless, our
nanoscale time-dependent measurement represents a new imaging method to diagnose the
local electrical response of perovskites, ranging from lead-free CsBrIxBr3−x alternatives to
mixed cation and halide compositions for optimized dual-junction solar cells. Thus, map-
ping the spatial and time dynamics of perovskite materials will likely impact the design
of next-generation stable perovskite solar cells with reliable voltage response.
110
Chapter 7: Perovskite composition-dependent nanoscale voltage re-
sponse
Perovskite solar cells that incorporate small concentrations of Cs in their A-
site demonstrate a boost in both stability as well as overall device performance.
Yet, how the Cs cation influences the electrical properties at the nanoscale of the
photovoltaic material is still unknown. In this chapter, we determine how the com-
position affects the short- and long-term electrical stabilities of three perovskites
under 1-sun illumination, MAPbBr3, MAPbI3, and Cs-mixed, with <50 nm spatial
resolution. An anomalous irreversible electrical signature on MAPbBr3 is mapped
at the mesoscale, which results in entire grains with no contribution to the Voc.
These measurements verify the necessity of high spatial resolution mapping. Con-
versely, Cs-mixed perovskites exhibit fully reversible voltage, ultimately proving
that the desired electrical output of this composition persists even at the nanoscale.
Further, this material presents no spatial variation in Voc as a function of illumi-
nation, as ion motion is restricted. Finally, we reveal that the performance of the
Cs-mixed thin film remains stable as a function of four weeks. These results show
that the nanoscale electrical behavior of the perovskites is intimately connected to
their chemical composition and macroscopic response.
111
7.1 Introduction
The remarkable progress and research effort that has happened to improve per-
ovskite solar cells in the past decade is unprecedented. [153, 304] This PV material has
now achieved power-conversion efficiencies (η) for both single-junction [35, 305] and tan-
dem devices containing perovskites [144, 306] >22%. These advancements demonstrate
researchers ripening comprehension of perovskites, and thus, the fabrication processes for
higher quality materials, such as electronic charge carrier transport layers and film inter-
faces, are being refined. [307–309] While such advances are necessary to develop commer-
cial PV devices, short- and long-term stability under realistic atmosphere environments
and perovskites inability to pass accelerated ageing tests remain challenging. [152]
The heterogeneity and dynamics of the electrical behavior in perovskites begin at
the nanoscale, [28, 34, 39, 151] and locally revealing the primary performance parame-
ters with high spatial resolution can potentially be correlated to the beneficial, parasitic,
and/or transient response and the overall device performance. For instance, under il-
lumination, perovskites have local variations in Voc, [37, 39, 90] photocurrent, [51] and
photoluminescence signal [41,310,311] as a function of both space and time. The primary
culprit of the heterogeneity is still under debate, but the strongest hypothesis thus far sug-
gests halide ion migration. [49,106,312–314] Recently, triple-cation perovskites containing
methylammonium (MA), formamidinium (FA), and Cs have been presented as a promising
alternative, due to experimental evidence that Cs incorporation improves thermal, [315]
moisture, [316, 317] and photoelectrical stabilities. [318] Simultaneously, the devices can
maintain its output η, suitable for the future, large-scale deployment of this PV material.
It has been lately observed that the addition of Cs affects the structural properties of the
112
absorbing layer by decreasing its lattice constant, while simultaneously preventing ion mi-
gration. [319] Nevertheless, it is currently unknown how the electrical signal of the grains
constituting these perovskites changes upon illumination and if this composition suffers
from ion motion at the nanoscale.
Here we elucidate how the presence of Cs-incorporated triple cations lead to a fully
reversible nanoscale voltage response in perovskite solar cells for both short- and long-term
stabilities. By implementing Kelvin probe force microscopy (KPFM), one can directly im-
age spatial variations in Voc under 1-sun illumination of three perovskites with different
compositions. The sample that incorporates Cs yields a uniform and steady Voc upon light
treatment, irrespective of grains and grain boundaries. The MAPbBr3 composition shows
an anomalous, non-reversible light-induced electrical behavior. For this material, illumina-
tion under 1-sun results in >10% of the material with “zero-voltage” regions ranging from
single grains to clusters, while no morphological variations are observed. Further, through
spatially and temporally resolving the light-induced changes in the MAPbBr3 sample, we
demonstrate the effect of ion motion in these solar cells Voc, which is responsible for the
observed dynamics. These results show that the Cs-mixed perovskite is the only stable
composition measured here at both the nano- and macroscale. The Voc measurements
presented can be applied to other perovskites, with alternative chemical compositions to
track material degradation, and can be used in different atmospheres.
113
7.2 Experimental methods
7.2.1 ToF-SIMS characterization
Using a focused Ga+ ion beam and scanning electron microscope dual-beam sys-
tem, depth profiles are acquired, presented in Figure 7.1. Both the positive and negative
secondary ions were generated by ionizing the species from the sample using a 20 keV
Ga+ ion beam with current at ∼1.1 nA. The ToF-SIMS signals were recorded by a mass
spectrometer. During the measurements, the vacuum level was kept below 10−4 Torr at
all times. Three distinct locations were measured for each perovskite sample for statistical
purposes. Areas of 20 × 20 µm2 were sputtered while 10 × 10 µm2 central regions were
analyzed to avoid any edge effect.
7.2.2 Raman spectroscopy and microscopy
Raman spectra and maps for the MAPbBr3 sample were acquired with a 650 nm
laser excitation source and spectrometer, using a 100× objective lens in a confocal optical
microscopy setup. The Cs-mixed and MAPbI3 samples were measured with a NT-MDT
atomic force microscope, using an excitation source of 532 nm laser. Each pixel in the
Raman map represents the integrated area under the peak, highlighted in Figure 7.2. All
Raman maps are 5 × 5 µm2 and were performed after the KPFM maps were measured.
7.2.3 Solar cell synthesis
Fluorine doped tin oxide coated glass slides (Sigma-Aldrich, 7Ω/) were cleaned
by sonication in 2% Hellmanex solution for 15 minutes. After rinsing with deionized
114
water and ethanol, the substrates were again sonicated with isopropanol for 15 minutes
and rinsed with acetone. The substrates were treated with UV-ozone for 5 minutes, and
a 50 nm thick TiO2 electron selective layer was deposited by spray pyrolysis at 450
◦C
from a precursor solution of titanium diisopropoxide bis(acetylacetonate) (0.6 ml) and
acetylacetone (0.4 ml) in anhydrous ethanol (9 mL). After the spraying, the substrates
were left at 450◦C for 30 min and left to cool down to room temperature. Mesoporous
TiO2 films were spin-coated (4000 rpm, 20 s) from a 1:10 solution of TiO2 paste (30 NR-D,
Dyesol) in anhydrous ethanol. The films were annealed on a programmable hotplate (2000
W, Harry Gestigkeit GmbH) using a 45 minute ramp to 450◦C followed by a dwell time of
30 minutes. Perovskite films were deposited from a precursor solution; MAPbBr3: MABr
(1.2 M) and PbBr2 (1.2 M) in anhydrous DMF:DMSO 4:1 (v:v); MAPbI3: MAI (1.2 M)
and PbBr2 (1.2 M) in DMSO; Cs0.06(MA0.17FA0.83)0..94Pb(I0.83Br0.17)3: FAI (1 M), PbI2
(1.1 M), MABr (0.2 M), PbBr2 (0.2 M) and CsI (0.075 M) in anhydrous DMF:DMSO
4:1 (v:v). The perovskite solution was spin-coated in a two-step program at 1000 and
6000 rpm for 10 and 20 s respectively. During the second step, 100 µl of chlorobenzene
was poured onto the spinning substrate five seconds prior the end of the program. The
substrates were then annealed at 100◦C for one hour in a N2 glove box. Subsequently,
the substrates were cooled down for a few minutes and a PTAA (EM Index) solution (10
mg/ml in chlorobenzene) doped with bis(trifluoromethylsulfonyl)imide lithium salt (Li-
TFSI, Aldrich) and 4-tert-butylpyridine (TBP, Aldrich) was spun at 4000 rpm for 20 s,
as according to the literature. [146, 320] For the solar cells, 80 nm of Au was thermally
evaporated under high vacuum for the top contact of the devices.
115
7.2.4 Macroscopic measurements
For the power conversion efficiency measurements, a solar simulator from ABET
Technologies (Model 11016 Sun 2000) with a xenon arc lamp was used, and the solar cell
response was recorded using a Metrohm PGSTAT302N Autolab. The intensity of the solar
simulator was calibrated to 100 mW/cm2 using a Si reference cell from ReRa Solutions
(KG5 filtered). J-V curves were measured in ambient environment at room temperature,
and under reverse bias (from high to low voltages) conditions, at a scan rate of 10 mV/s.
The active area of the cells was 0.09 cm2 and cells were measured two days after their
preparation.
7.2.5 Kelvin probe force microscopy experiments
The KPFM measurements were performed using an atomic force microscope (AFM).
A conductive Pt-coated Si probe is used with a 35-nm tip to measure the work function
difference and topography of each sample simultaneously. Due to the geometry of the
perovskite half-devices, in all cases the surface of the 50-nm thick PTAA hole transport
layer was scanned. To neglect the influence of moisture during the KPFM measurements,
piping connected to a dry air source was fed into the AFM chamber, which maintained a
humidity level within the microscope to <10% relative humidity, see Figure 7.5. The sam-
ple’s were illuminated with a power-tunable 532 nm laser (1-sun), directed into the AFM.
A small glass prism was placed underneath the perovskite thin film, and reflected light up
toward the sample, producing a spot with 0.5 mm in diameter. Heterodyne-KPFM mode
was implemented during the experiment, which measures the contact potential difference
(VCPD) of the tip-sample at the second resonance of the cantilever; see refs [39,155] for de-
116
tails. The line-noise of 50 mV was quantified by examining voltage line profiles. Between
experiments, the perovskite samples were stored in a low-humidity desiccant chamber,
with the average humidity level between <10%.
7.3 Composition analysis of perovskites
Perovskites are, in most cases, an organic-inorganic hybrid material blend with a
chemical formula of ABX3 (A = monovalent cation, B = inorganic cation, and X = halide
anion). The constituents of this structural form have been researchers primary tuning knob
used to alter the optoelectronic properties of perovskites as well as improve the thermal,
structural, and photo stability. [119,150,321,322] The most common compounds for each
lattice position are: A-site: methylammonium (MA = CH3NH
+
3 ), formamidinium (FA =
HC(NH ) 2)+2 ( ) and inorganic Cs
+; B−site: Pb2+, which has the best performance in this
location (although very recent studies show that a partial substitution of Co2+ or Bi3+
could boost output power) [321,323]; and X3 site: I
−, Br−, and/or Cl−.
We investigate three perovskite thin-films with different compositions. First, the
conventional MAPbI3 (with bandgap = 1.58 eV) is selected because its electronic proper-
ties are well documented, and therefore, it is an ideal candidate for systematic testing of
the chemical composition contribution to the nanoscale electrical response. Another pure
halide, MAPbBr3, is analyzed; this perovskite has a higher bandgap (2.2 eV) and, thus,
larger Voc and limited photocurrent, and it could be utilized in tandem PV devices. The
overall electrical performance (macroscopic) of pure-halide perovskites has been observed
to degrade with exposure to either light and/or moisture, especially when the sole A-site
compound is MA. [153, 324, 325] Cation engineering is implemented by adding a small
amount of an inorganic component, Cs+, to a mixture of MA and FA, and also use a
117
(a) PTAA – 50 nm (b) PTAA – 50 nm (c) PTAA – 50 nm 
MAPbBr – 400 nm Cs-mixed– 400 nm MAPbI3 – 400 nm3 
Mesoporous TiO2 – 150 nm Mesoporous TiO2 – 150 nm Mesoporous TiO2 – 150 nm
Compact TiO2 – 50 nm Compact TiO2 – 50 nm Compact TiO2 – 50 nm
FTO (500 nm)/Glass FTO (500 nm)/Glass FTO (500 nm)/Glass
(d) PVSK TiO2 FTO/Glass (e) PVSK TiO2 FTO/Glass (f) PVSK TiO2 FTO/Glass
0.5 Br-1 0.3Ti O
0.3 TiBr-2
0.4 O
Ti Cs 0.2 O 
0.3 0.2 I
Br-1 Si Si
0.2 0.1
Si 0.1 Br-2
0.1 I
0.0 0.0 0.0
0 100 200 300 400 500 600 0 100 200 300 400 500 600 0 100 200 300 400 500 600 700
Time (s) Time (s) Time (s)
Figure 7.1: Chemical composition analysis of perovskites by ToF-SIMS.
Schematic and real-color photographs of perovskite samples: (a) MAPbBr3, (b) Cs-
mixed, and (c) MAPbI3. Samples size = 1.2 × 1.5 cm2. (d)-(f), Representative depth
profiles of ToF-SIMS measurements acquired for each sample.
mixture of halide atoms. This approach has been observed to improve the stability in
ambient environments substantially. [146, 318] The third sample investigated here has a
mixed A-site and X-site perovskite material: Cs0.06(MA0.17FA0.83)0.94Pb(I0.83Br0.17)3 (i.e.
Cs-mixed, also with bandgap ≈ 1.58 eV). [146, 326] See the Methods section for detailed
information about the perovskite thin film/device fabrication steps. Half-devices are fab-
ricated, which are composed of PTAA/perovskite/TiO2/FTO/glass, to perform nanoscale
electrical measurements through KPFM, as will be presented later in this chapter.
To confirm the existence of each species within all perovskite compositions, time-of-
flight secondary ion mass spectroscopy (ToF-SIMS) are performed and the depth profiles
of the relevant elements are collected. Figure 7.1 displays a schematic of the perovskite
samples and their corresponding thicknesses, plus the respective real color photographs
(Figure 7.1(a)-(c)), which are an indicator of their bandgaps. The elemental depth profiles
obtained by ToF-SIMS (Figure 7.1(d)-(f)) validate the presence of the constituent elements
in the perovskite films with minimal diffusion throughout the other thin-film layers, see
the experimental methods.
118
Counts
Raman data
MAPbBr3 Cs-mixed MAPbI3
250 650 150
(a) (b) (c)
200 600 125
150
550 100
100
500 75
50
50
0 250 500 750 1000 1250 1500 1750 4500 250 500 750 1000 1250 0 250 500 750 1000 1250 1500 1750
Wavenumbe -1 -1    r  (cm ) Wavenumber (cm ) Wavenumber (cm
-1)
(d) Raman  65-100 cm-1 (e) Raman  251-340 cm-1 (f) Raman 108-196 cm-1
   
      
      
   
1 μm 1 μm 1 μm
0 90 3E6 1.3E7 4.7E6 5E6
Counts (a.u.) Count s (a.u.) Coun ts (a.u.)
@UMD @ORNL @UMD/ORNL
Br-PFbi-gBurr setr7e.2tc:hRinagm, aasnymmmiecrtroicscopyMoAnYBpEe: rMovAs tkoirtseiotnhin filmsP.b-(Ia s)t-r(ect)c,hRinagm, asnymspmeec-tric
tra and (d)-(f), their corresponding Raman maps. Each pixel in the 5 × 5 µm2.
Raman map is the integration of the Raman peak highlighted by gray in the spec-
trum above, where the range in cm−1 is indicated within the maps. Raman peak
identification (a)-(c): Br3-Pb-Br3 stretching, asymmetric; MA torsion, and Pb-I3
stretching symmetric, respectively. [327,328]
As desired, Br (two isotopes shown as Br-1 and Br-2 in Figure 7.1), Cs, and I are
all confined to the perovskite layer of the samples. Further, the chemical composition
is uniform throughout each layer. This uniform distribution in the composition is in
agreement with our Raman microscopy chemical imaging, see Figure 7.2.
7.4 Probing perovskites’ electrical response at the nanoscale
Each perovskite thin-film investigated here also has a full device counterpart. Light
J-V curves and η vs. time plots are presented in Figure 7.3, detailing the average per-
formance parameters of the investigated films (see Methods for parameter specification).
From the macroscopic measurements, it is clear that the Cs-mixed sample has the highest
η at 18.6% while concurrently having excellent short-term performance stability (Figure
119
Counts (a.u.)
 
 
 
 
25 18
Cs-mixed
20 MAPbI3 15
MAPbBr3
15 12
Cs-mixed
9 MAPbI
10 3
MAPbBr
6 3
5
(a) 3 (b)
0
0.0 0.3 0.6  0.9 1.2 1.5 25 50 75 100 125 150
Voltage (V) time (s)
Figure 7.3: Macroscopic performance and short-term stability of perovskite
solar cells. (a) Light J-V curves for all the perovskite solar cells investigated here
(color-coded). For Cs-mixed, MAPbI3, and MAPbBr3: Voc = 1.11, 1.02, and 1.36 V,
J 2sc = 22.1, 18.1, and 5.5 mA/cm , FF = 76, 85, and 76%, and η = 18.6, 15.7, and
5.7%, respectively. (b) η vs. time for each perovskite device. The Cs-mixed sam-
ple shows the highest performance and stability while the MAPbI3 device degrades
rapidly with time.
7.3(b)). The MAPbBr3 solar cell has the lowest efficiency at 5.7% despite having the
largest Voc (due to its wide bandgap) and also maintains an unchanged electrical output
with time, see Figure 7.3(b). Yet the MAPbI3 device, which initially has a promising η =
15.7%, rapidly falls to <13% in two minutes, signifying that this perovskite composition is
the least electrically stable. Furthermore, despite the larger fill factor (FF), the MAPbI3
light J-V curve contains an undesirable bump between 0.8-0.9 V, which is attributed to
charge build-up close to the electrical contacts. [49] For these reasons, we have chosen
to focus these high-resolution mapping experiments on the MAPbBr3 and Cs-mixed per-
ovskite compositions, to understand how their nanoscale voltage behavior influences the
materials electrical stability.
The nanoscale, real-time Voc response of the perovskites is imaged by Kelvin-probe
force microscopy (KPFM) under 1-sun illumination conditions and extremely low humid-
ity. A schematic of the KPFM experimental setup is presented in Figure 7.4. The samples
120
Current dens ity (mA/cm2)
η (%)
are illuminated from the bottom using a 532-nm laser beam incident on a glass prism,
which reflects an equivalent photon flux of 1-sun up towards the sample. Briefly, KPFM
measures the work function difference between the atomic force microscope (AFM) probe
and the sample surface. [28] Upon illumination, the perovskite absorbs photons, which
generate charge carriers, introducing quasi-Fermi level splitting (directly proportional to
the Voc) [37] and thus, changing the work function of the material. By subtracting an
illuminated KPFM scan by a dark one performed in the same location, and maintaining
a common electrical ground between the probe and the perovskite samples back electrode
(FTO in this case), we image Voc with spatial resolution <50 nm. [37, 39] The relative
humidity in the AFM chamber is held <8% (Figure 7.5) by continuously flowing dry air
into the microscope through a piping system. See the experimental methods section for
more details on experimental parameters. The nanoscale voltage response of perovskite
samples before, during, and after illumination are mapped, termed here Dark(1), Light,
and Dark(2), respectively.
7.4.1 Irreversible nanoscale electrical response in MAPbBr3
Figure 7.6 displays measurements of the nanoscale voltage response of the MAPbBr3
perovskite under 1-sun illumination, where an irreversible response is observed upon the
material’s first exposure to light. During the entirety of our experiments, no topography
changes are observed (Figure 7.6(a)-(c)), indicating that any variations in the electrical
signal are due to non-morphological properties of the perovskites. It also demonstrates that
during the scans the AFM probe is kept intact (and all changes in tip-sample interactions
are due to changes in the electrical output of the perovskites). The MAPbBr3 perovskite
presents extraordinary, irreversible light-induced electrical behavior at the nanoscale, with
121
piezo and cantilever
AFM controller 
AFM probe and feedback
perovskite 
half-device
incident light
Figure 7.4: Imaging the Voc response in perovskites at the nanoscale.
Schematic of the illuminated-KPFM experimental setup, where the perovskite sam-
ples are back illuminated, and the local voltage signal is measured by an AFM probe
using Kelvin-probe force microscopy. Because the perovskite back contact is grounded
with respect to the tip, the data acquired under illumination is proportional to the
Voc of the solar cell.
Humidity data (for week 2)
25
20
15
10
5
0 2 4 6 8 10 12 14 16
Time (hrs)
Figure 7.5: Relative humidity during KPFM measurements. The humidity
level inside the AFM chamber was monitored every 30 seconds during the exper-
iments. A constant dry air flow was pumped into the chamber to maintain low
moisture content.
122
Humidity (%)
no morphological changes. Before this sample is exposed to light, the voltage is nearly
uniform, see Dark(1) image in Figure 7.6(d), with an average voltage equal to 1.04 V.
Upon illumination (Figure 7.6(e)), the an inhomogeneous photovoltage response which
differentiates itself from the distribution of the initial dark scan. After sample illumination
(Figure 7.6(f)), a new electrical and reversible state emerges, termed as Dark(2). There
are two distinct regions within the MAPbBr3 voltage maps in Figure 7.6(d)-(f): (i) the
background, where the light-induced voltage shows the largest increase (very light green,
voltage = 1.84 V), and (ii) unique perovskite grain and grains clusters which have become
inactive photovoltage sites, viewed as the medium-green color in Figure 7.6(e) and (f)
(voltage = 1.49 V in both cases). Note that after the initial illumination of the MAPbBr3
sample (shown in Figure 7.6(e)) the voltage responses, both Light and Dark, become
reversible.
To quantify the actual effect of the high and low voltage regions on the Voc of
′
MAPbBr3, we subtract a Dark(x) image from a Light-KPFM scan, and obtain Voc maps,
′
as indicated in Figure 7.6. Here Voc is defined as an effective Voc because it results from the
partial illumination of the perovskite (laser is focused and, thus, the excitation of carriers
is localized) using a single wavelength, and considers the dark current contribution of the
′
entire sample area. Thus, as expected, Voc differs from the macroscopic Voc, where the
volume of carriers excited is much larger. Two contrasting areas are immediately appar-
′ ′
ent when comparing the Voc(1) and Voc(2) maps (Figure 7.6(g) and (h), respectively), as
′
locally, the Voc(2) value drops to zero (dark red regions). Because the electrical response
does not change post-illumination, the photovoltage generation is now inactive at these
sites, harming the local and, thus, macroscopic performance of devices. These parasitic
regions cover approximately 12% of the scanned area shown here. Surprisingly, despite the
123
(a) AFM (b) 75(c)
1 μm 1 μm 1 μm
0
(d) Dark(1) (e) Light (f) Dark 1.9(2)
1 μm 1 μm 1 μm
0.7
Light – Dark(1) Light – Dark(2)
(g) (h) 1
1 μm 1 μm 0
𝑉𝑉′𝑜𝑜𝑜𝑜(1) = 0.727 V 𝑉𝑉
′
𝑜𝑜𝑜𝑜(2) = 0.757 V
Figure 7.6: Nanoscale irreversible voltage response in MAPbBr3 per-
ovskites. (a)-(c) AFM topography maps of the MAPbBr3 sample and their respec-
tive (d)-(f) voltage images under sequential dark, illuminated, and dark conditions.
Here, despite consistent topography, an irreversible electrical process occurs within
the MAPbBr3 perovskite film. (g,h) Spatial variation in open-circuit voltage under
′ ′
local illumination where: Voc(1) = Light minus Dark(1), and Voc(2) = Light minus 4
Dark(2). For all measurements relative humidity <8%, and for all light-KPFM scans
the illumination conditions are a 532-nm laser under 1-sun incident photon flux.
124
height (nm) voltage (V) 𝑉𝑉′𝑜𝑜𝑜𝑜 (V)
′ ′
zero-voltage grains in the Voc(2) map, the average Voc value of the entire scanned region
〈 ′( Voc(x)〉), quoted underneath Figure 7.6(g) and (h)) increases by 30 mV after illumination.
Note that a macroscopic measurement would not capture such irreversible behavior and
could inadvertently lead to the conclusion that there is an overall enhancement in per-
formance, when this is not the full story. We hypothesize that this irreversible electrical
process is caused by Br− ion motion within the perovskite film. [329–331] Because Br is
a negatively charged ion, it tends to move towards the electron transport material during
operation conditions (meaning, during illumination). The movement of these ions away
from the surface (i.e. PTAA, the hole transport layer) is likely the reason for the observed
heterogeneous depletion of voltage at the tip-sample junction.
Through fast-KPFM we determine the immediate changes that take place in the
perovskites electrical response upon illumination in two representative regions within the
scanned area of Figure 7.6, see Figures 7.7 and 7.8. Briefly, in fast-KPFM, the real-
time voltage is measured, where each map is acquired in 16 seconds. [155] As the laser
is turned ON/OFF, the transient and asymmetric voltage response of this perovskite is
captured, as displayed in the maps presented in Figure 7.7 and in the corresponding voltage
histograms. Once the sample is illuminated under 1-sun (Figure 7.6(c)) the electrical
signal changes substantially, corresponding to the photo-generated voltage. Here, the
‘slow’ voltage increase that takes place in ∼2-3 minutes (red-color histograms) results
from ion motion within the perovskite. As the material is illuminated the electrons fill in
the existing trap states.12 Consequently, the Br− anions now have a preferential motion
direction — away from these filled states. Upon turning the laser OFF a residual positive
voltage response is meaasured for ∼10 min (see gradual change in signal from Figure 7.7(f)
to (i) and red to purple histograms in Figure 7.7(j)). Here, the dynamic voltage signal
125
30 (a) AFM (b) Dark (c) Dark  Light (j)
10 min
Light ON
0 200 nm
(d) Light (e) Light for 3 min (f) Light  Dark
Light OFF
2 min
OFF
3 min ON
(g) Dark (h) Dark for 2 min (i) Dark for 10 min
OFF
0.8 1.0 1.2 1.4 1.6 1.8 2.0
voltage (V)
0.8 1.8
voltage (V)
Figure 7.7: Dynamic, reversible electrical response after the first illumi-
nation of MAPbBr3. (a) Topography map. (b)-(i) Fast Dark- and Light-KPFM
scans (16 sec/scan) shown in chronological order, demonstrating the slow reversibility
of the voltage response in 10 minutes post illumination. During scan (c) the 532 nm
laser at 1-sun is turned ON, scan direction bottom to top. During scan (f) the illumi-
nation source is turned OFF, scan direction from top to bottom. (j) Corresponding
voltage histograms for the entire fast KPFM series.
at open-circuit conditions results from the now reversible electron and Br- motion. Yet,
there is a hysteresis in the nanoscale voltage, where it takes more than 2× longer for
the voltage to equilibrate under dark conditions. Note that this time-dependent voltage
behavior could not be resolved by conventional KPFM methods, as they would not capture
the short-time voltage changes.
7.4.2 Fully reversible and stable Cs-incorporated triple cation perovskites
The macroscopic electrical performance and stability of solar cells, shown in Figure
7.3, can be traced down to their nanoscale constructs. Therefore, revealing the relevant
local voltage responses will help answer the open-question of: “Why does Cs-incorporation
significantly improve overall device efficiency and stability?” Thus, we dedicate this sec-
tion to the effect of cation engineering on the material at short- and long-term time scales.
126
height (nm)
counts (a.u.)
time
30 (a) AFM (b) Dark (c) Dark  Light (j)
200 nm Light ON0
(d) (e) Light for 5 min (f) Light  Dark
Light OFF
OFF
ON
(g) (h) Dark for 3 min (i) Dark for 7 min
OFF
0.8 1.0 1.2 1.4 1.6 1.8 2.0
voltage (V)
0.8 1.8
voltage (V)
Figure 7.8: Dynamic electrical response after illumination of MAPbBr3
region 2. (a) Topography map. (b)-(i) Fast Dark- and Light- KPFM scans (16
sec/scan) acquired, shown in chronological order, demonstrating the slow reversibil-
ity of the voltage response in ∼10 minutes post illumination. During scan (c) the
532 nm laser at 1-sun is turned ON, scan direction bottom to top. During scan
(f), the illumination source is turned OFF, scan direction from top to bottom. (j)
Corresponding voltage histograms for the entire fast KPFM series.
In Figure 7.9 the morphology (which is constant as a function of time and illumination,
Figure 7.9(a)-(c)) and electrical signal of Cs-mixed composition at the nanoscale is shown.
We measure a very uniform voltage distribution (Figure 7.9(d)-(f)) under both dark and
light conditions, irrespective of the interfaces between the grains. This is desirable for
large-scale solar cell deployment, and a distinct difference from the behavior of MAPbBr3.
The Cs-mixed sample also presents fully reversible voltage response (under the experi-
′
mental conditions tested here) determined by extracting the difference between the Voc(1)
′
and Voc(2) maps (= 9 mV), see Figure 7.9(g)-(h), which is well within the line-noise level
of our KPFM measurements. The ion motion suppression in the Cs-mixed perovskites is
further confirmed by applying fast-KPFM, as shown in Figure 7.10. As a region consist-
ing of several grains undergoes illumination cycling, the nanoscale voltage response is fully
reversible, and it immediately changes upon turning the laser ON/OFF. It is important
127
height (nm)
counts (a.u.)
time
(a) AFM (b) (c) 75
1 μm 1 μm 1 μm
0
(d) Dark(1) (m) Light (f) Dark
2.3
(2)
1 μm 1 μm 1 μm
1.5
Light – Dark(1) Light – Dark(2)
(g) (h) 0.4
1 μm 1 μm
0
𝑉𝑉𝑜𝑜𝑜𝑜(1) = 0.275 V 𝑉𝑉𝑜𝑜𝑜𝑜(2) = 0.266 V
Figure 7.9: Nanoscale reversible voltage responses in Cs-mixed perovskites.
(a)-(c), Topography and (d)-(f), KPFM maps of the Cs-mixed sample showing fully
′
reversible voltage and no significant spatial variations. (g,h) Voc maps, where the
difference in the average value between the two images is within the noise level of the
KPFM measurements. For light-KPFM scans, the illumination conditions = 532-nm
laser at 1-sun incident photon flux. All measurements held relative humidity <10%. 7
′
to reiterate that, as expected, the absolute value of the Voc for the Cs-mixed sample is
lower than the MAPbBr3 one due to its smaller bandgap energy, see Figure 7.3. Yet, the
differences in the perovskites electrical behavior can only be attributed to their composi-
tions, as the experimental parameters, grain morphology, and thin film stacking layers are
all similar. Through our scanning probe microscopy experiments we elucidate that the
presence of Cs prevents the degradation of the perovskites at the nanoscale, ultimately
leading to the stable macroscopic response observed in this system. [317,318]
Figure 7.11 shows the stability of all samples investigated here after four weeks of
KPFM experiments. As described in other reports, [146, 319, 326] by incorporating Cs
into the A-site position (which has a smaller radius than both MA and FA), [332] halide
128
height (nm) voltage (V) 𝑉𝑉′𝑜𝑜𝑜𝑜 (V)
16 s/scan 
1 x 1 μm2
30 (a) AFM (b) Dark KPFM (c) Dark; 1 min later (d) Dark  Light (e) Light
light ON
0 200 nm
(e) Light; 3 min later (f) Light  Dark (g) Dark (h) Dark; 4 min later (i) Dark  Light
1.5
light OFF
light ON
(j) Light (k) Light; 16 s later (l) Light  Dark (m) Dark (n) Dark; 7 min later 
light OFF
0.9
(o) Dark  Light (p) Light (q) Light; 6 min later (r) Light  Dark (s) Dark; 8 min later 
light ON
light OFF
Figure 7.10: Dynamic electrical response after illumination of Cs-mixed
perovskite. (a) Topography map. (b)-(s), Fast Dark- and Light- KPFM scans (16
sec/scan) acquired, shown in chronological order, demonstrating the reversible and
uniform voltage response of the Cs-mixed perovskite half-device in a 1× 1 µm2 region.
During scans (d), (i), and (o), the 532 nm laser at 1-sun is turned ON, and for scans
(f), (l), and (r) the illumination source is turned OFF. The scan direction is indicated
by the white arrows within the voltage map. The entire dataset is not shown as this
would comprise of >100+ images, therefore, we have selected time snapshots which
either (1) show a change in illumination state or (2) are representative of the voltage
distribution during a given illumination state.
129
height (nm)
voltage (V)
migration is inhibited because both MA and FA are compressed in these cases. The
histograms in Figure 7.11 display the time-dependence of the Voc of all the perovskite
compositions obtained by the KPFM maps, where the error bars correspond to the line-
noise level of the measurements. It is clear that the MAPbI3 perovskite decays week-after-
week, signifying material degradation. Meanwhile, the MAPbBr3 and Cs-mixed samples
remain stable throughout the entirety of the experiments, although MAPbBr3 does show
higher volatility. This phenomenon could be related to the measured irreversible electrical
signal and zero-voltage sites in the MAPbBr3, which in this case, are likely caused by
light-induced, mobile organic cation and/or bromine ions. [329, 330] In comparison with
the other pure halide options, the Cs-mixed perovskite has superior electrical behavior.
The significance of spatially resolving the Voc of inhomogeneous materials, such as
perovskites, lies in revealing changes in the optoelectronic behavior of grains and within
grains related to charge carrier recombination that dominates the overall device stability
and performance. Ultimately, elucidating the driving forces for material degradation upon
exposure to light (in this case), moisture, oxygen, temperature and bias will advance our
understanding of what triggers the dynamic response commonly observed in perovskites.
To date, microscopic measurements of perovskites through scanning probe and electron
microscopy-based methods have shown significant spatial variations in the photocurrent
and photovoltage generated by MA and FA perovskites, and are likely related to physical
and chemical processes that take place once the material is exposed to ambient environ-
ment or operation conditions. Thus, identifying the response of novel perovskites, such as
(RbCsMAFA)PbI3 and Bismuth-halide double perovskites, [333] to the parameters above-
mentioned is critical to elucidate which chemical compositions can be implemented as solar
cell modules.
130
1.2 MAPbI3 MAPbBr3 Cs-mixed
1.0
0.8
0.6
0.4
0.2
0.0 W W W W W Wee ee ee e
We e We We e e e ee ek e 1 k 2 k 4 k 1 k 2 k 4 k 1 k 2 k 4
Figure 7.11: Analysis of perovskites nanoscale electrical stability. Histograms
of Voc, extracted from the nanoscale KPFM experiments, normalized as a function
of the first week for each perovskite sample. The superior long-term voltage stability
of the Cs-mixed composition is evident by its steady Voc signal over four weeks.
7.5 Conclusion
We determined the influence of chemical composition on the nanoscale Voc stability
for three types of perovskites: MAPbI3, MAPbBr3, and Cs-mixed. By performing KPFM
imaging under 1-sun illumination conditions we discovered an irreversible photovoltage
response within the MAPbBr3 perovskite, demonstrating the need for high spatial res-
olution mapping. Regions of zero-photovoltage coexist with active areas upon material
illumination. The MAPbI3 conventional perovskite presented the lowest stability among
the three compositions at multiple time scales. Meanwhile, we measured a fully reversible
and uniform voltage response in Cs-mixed perovskites, which is linked to the suppression
of ion motion. The differences between the three samples’ Voc behavior are due to the ex-
istence and suppression of ion migration: as Cs is incorporated into the perovskite lattice,
the movement of ions is inhibited due to strain. As a result, the nanoscale Voc is homo-
geneous at the length scale of individual grains. Yet, for the MAPbBr3 perovskite, the
131
𝑉𝑉𝑜𝑜𝑜𝑜 (norm)
motion of Br− ions causes a hysteresis on the Voc response as a function of illumination.
This nanoscale Voc imaging/analysis can be expanded to probe the real-time response
of emerging perovskites, including options for multi-junction solar cells [306] and light-
emitting devices. For instance, by implementing fast-KPFM one can identify the role
of different atmospheres, such as N2, O2, and H2O, on the dynamic electrical properties
of these materials. Moreover, the KPFM measurements can be combined with other
AFM-based methods, such as photoconductive-AFM, to provide a complete picture of the
electrical (voltage and current) characteristics of the perovskites. As the search for reliable
perovskites accelerates, defining the chemical compositions that lead to stable devices is
key for the future deployment of high-performance and low-cost optoelectronics based on
this class of material. Because they are composed by nanoscale constructs, resolving their
individual contribution to the overall material behavior through microscopic methods will
enable the understanding of the relevant processes responsible for the dynamic response
often observed in perovskites.
132
Chapter 8: Perspective on functional imaging of PV materials
Further global adoption of photovoltaic energy conversion technologies is con-
tingent on sustained progress towards widespread grid-parity. For that, solar cell
materials composed of micro-scale grains and nanoscale boundaries show the highest
potential due to their large theoretical efficiency and low-cost fabrication methods.
Here we outline the current challenges facing hybrid perovskites and prevalent thin
film polycrystalline materials for photovoltaics. We offer an perspective on how
mesoscale functional imaging can enable a complete understanding of the physi-
cal and chemical processes restricting their performance, completing the materials
science structurepropertiesprocessingperformance tetrahedron. We provide a dis-
cussion about how novel imaging methods based on electron and scanning probe
microscopies can be realized to provide quantitative information about the relevant
parameters (figures-of-merit) that define solar cell performance. Finally, we offer our
vision for the upcoming years, wherein correlative functional microscopy will lead to
a complete narrative of the electrical, optical, structural, and chemical properties of
these materials, including their surface and bulk properties. This chapter is adapted
from E.M. Tennyson et. al, ACS Energy Letters, 2 1825 (2017) [34]
133
8.1 Introduction
In this chapter, we provide insight about how cutting-edge functional nano-imaging
methods can be implemented to advance our understanding of the fundamental physical
and chemical processes governing these non-uniform materials, including the use of both
electrons and photons as the excitation source for charge carriers. Second, we discuss the
need for real-time microscopic methods to probe the degradation mechanisms of perovskite
solar cells, still not well understood. Finally, we briefly analyze the relevance of big data
to effectively correlate the information acquired by a set of microscopic methods.
We suggest the implementation of high spatial resolution microscopy methods be-
yond their standard operation mode as a functional imaging platform to directly resolve
the role of the grains and interfaces on solar cell performance. To emulate PV device
operation conditions, both electrons and photons can be employed as the carrier exci-
tation source. Therefore, we discuss the primary advantages and limitations of selected
microscopy approaches and how they can be modified to offer quantitative information
about device figures-of-merit at the nanoscale.
8.2 Electrons as a source of excitation
Electron microscopy has proven to be an indispensable tool to acquire structural,
electrical, and chemical information of PV materials with nanoscale and atomic spatial res-
olution. SEM-based measurements are typically performed in a high vacuum atmosphere,
minimizing sample contamination. The interaction of the electron beam (e-beam) with
the material can generate secondary and back-scattered electrons, X-rays, and cathodo-
luminescence (CL), portrayed in Figure 8.1(a). A commonly used electrical measurement
134
in a SEM is termed electron-beam-induced current (EBIC), where the e-beam acts as a
local source of excitation of carriers. As the e-beam rasters across the sample, it pro-
duces a map of the generated current with spatial resolution primarily determined by the
beam size. [196] This characterization method is frequently performed on cross-sections
to determine the relative changes in charge carrier collection across p-n junctions and
other important interfaces. These measurements are especially important for perovskite
devices, as they do not have a conventional, well-defined p-n junction. However, the use of
electrons as the carrier excitation source is prone to severe image artifacts, mainly caused
by: (i) e-beam penetration path and depth, and (ii) damage to the sample.
(a) e- beam (b) EBIC e
- beam parabolic (c) CL
e- beam source mirror
top 
contact current flow conduction band spectrometer
secondary e- back- +scattered e- h CL
X-rays CL e- solar cell
Position (nm)
pre-amplifier
PV material valence band Wavelength (nm)
Figure 8.1: Electrons as a source of excitation for high-resolution functional
imaging of solar cells. (a) Illustration of the main radiation signals generated
from an e-beam upon interaction with a material. (b) Schematic of e-beam induced
current (EBIC), where the generated current is spatially resolved. (c) Illustration
of cathodoluminescence (CL) resulting from the radiative recombination of charge
carriers.
Cathodoluminescence (CL) is the physical phenomenon defined by the emission of
photons (electromagnetic radiation) within the UV-IR range, occurring when electrons re-
lax to their fundamental state, resulting from a material excitation by an e-beam [334,335]
- see Figure 8.1(c). Thus, information about where radiative recombination takes place
within a PV material can be obtained by CL. The spatial resolution is, in principle, defined
by the size of the e-beam in the SEM. CL imaging has been implemented in the investi-
gation of a variety of solar cell materials. [336–339] The successful acquisition of reliable
135
Current
CL
data heavily depends on the instrumentation used for light detection. A parabolic mirror,
inserted between the specimen and the pole piece, focuses the photons emitted from the
solar cell into a well-defined point on the detection system. Then, the collected radiation
is spectrally analyzed/resolved for every e-beam position using a high sensitivity CCD
detector and a two-dimensional emission map is recorded. In this manner, the spectrally
and spatially resolved optical response is acquired. CL has recently been implemented
to probe the optical response of the fully processed device. As an example, the CdCl2-
treatment during CdTe fabrication has been shown to reduce non-radiative recombination
in all types of GBs, as the overall intensity of the CL spectrum increases after this pas-
sivation step. [55] One unique feature of CL is the fact that it does not require device
processing. In other words, electrical contacts are not required for mapping the radiative
recombination rate of carriers within a thin film. Nevertheless, analogous to EBIC, CL
suffers from the similar limitations (see Table 8.1), due to the excitation source.
136
137
imaging technique applications limitations
Electons as a source of excitation
EBIC maps of generated current E-beam induced artifacts due to sample damage,
strong dependence on accelrating voltage, pri-
marily a surface signal.
CL radiative recombination of charge carriers Resolution is limited to e-beam volume interac-
tion and excess carrier diffusion.
Photons as a source of excitation
pcAFM photocurrent, relative nanoscale J-V Removal of conductive coating on probe due to
characterization sample contact, slow scan speeds, and effect of
contact size.
illuminated-KPFM quantitative, nanoscale Voc Topographic artifacts due to surface roughness.
NSOM LBIC photocurrent beyond diffraction limit Decouple the evanescent field around probe aper-
ture and near-field sample-probe iteractions.
NSOM PL radiative recombination of charge carriers
beyond diffraction limit
Tr-PL through NSOM minority carrier lifetime with subwave-
length spatial resolution
Table 8.1: Selected nanoscale functional imaging methods performed on PV materials along with their primary applications and limitations
Though the imaging methods described above have been extensively used to char-
acterize non-uniform solar cell materials, a combinatorial SEM-based functional imaging
platform is necessary to describe the full picture of the fundamental physics of charge
carriers. Thus, both EBIC and CL are beginning to be used in conjunction with other
SEM-based imaging capabilities to gain information about additional material properties
(e.g. chemical or structural). X-rays produced by the e-beam (Figure 8.1(a)) are used
for compositional mapping, by employing energy dispersive X-ray spectroscopy (EDX).
[340] Alternative chemical imaging techniques include electron energy loss spectroscopy
(EELS), [341] and time-of-flight secondary-ion mass spectroscopy (ToF-SIMS). [41] Addi-
tionally, the orientation of the grains and the types of interfaces present in a sample can
be mapped by electron backscattering diffraction (EBSD). Recently, EBIC, transmission
electron microscopy with selected area diffraction (TEM-SAD), and atom probe tomog-
raphy (APT) were united to investigate the CdSexTe1−x alloy that forms upon annealing
of CdTe devices. [38] Here, the zinc blende phase showed superior photo-response of when
compared to wurzite. Although very informative, EBIC measurements do not provide
quantitative information regarding current collection, nor has the generated current been
compared to macroscopic light J-V measurements.
We foresee correlative SEM and TEM approaches becoming widespread over the
coming years, as the dynamics of the charge carrier processes responsible for high (low)
PV performance is strongly related to the structural and chemical characteristics of the
grains composing the thin-film devices. To date, most of the work using SEM techniques
has focused on capturing the relative variations in current; however, measurements of Jsc,
a key figure-of-merit for solar cells, are still missing. While a photon generates only one
electron-hole pair, an electron can generate up to thousands of electron-hole pairs. The
138
volume in which the carrier excitation event occurs depends on the e-beam current and
voltage, the material density, band gap energy, its backscatter coefficient, atomic weight,
atomic number, and the quantum efficiency for generating an electron-hole pair. [342]
Thus, the volume of excitation, which is defined by the trajectory of the electrons within
the solid, must be calculated for a quantitative comparison between EBIC and Jsc. For
that, the role of the contact size must also be taken into consideration. [179] Although
the contribution of each individual atom is not as relevant as the mesoscale behavior of
the material, if atomic resolution becomes needed, we expect TEM-based EBIC to be
implemented. In this case, besides all considerations mentioned above, one must also
determine the possible overestimated contribution of surface recombination due to the
very large surface area-to-volume ratio of the samples (as a result of the lamella sample
size).
8.3 Photons as a source of excitation
The nanoscale functional imaging of PV using photons as the excitation source
requires researchers to use atomic force microscopy (AFM) or near-field scanning optical
microscopy (NSOM). Both platforms offer spatial resolution beyond the diffraction limit,
<100 nm, enabling the elucidation of the local electrical and optical properties of solar
cells. In this section, we summarize the operation principles of selected AFM- and NSOM-
based methods that provide invaluable insights about the material optoelectronic response
and focus on how they can provide quantitative information about the figures-of-merit of
the solar cells, with nanoscale spatial resolution.
Scanning probe microscopy methods have demonstrated exceptional versatility when
mapping the electrical response of PV with nanometer spatial resolution, as reviewed in
139
Chapter 3. [37,71] A common AFM mode used to determine the electrical response of solar
cells is photoconductive (pc) AFM. Here, a conductive metal probe acts as the top local
electrical contact of the PV device, as illustrated in Figure 8.2(a). Upon device illumi-
nation by an external light source, the probe directly collects the generated photocurrent
while recording the topography of the sample in contact mode. The potential beneficial
role of grain boundaries in CdTe, CIGS, and CZTS devices as current collectors has been
broadly investigated through pcAFM. [94, 192, 343] However, it is still unknown how the
grain boundaries affect charge carrier recombination, which requires mapping Voc, as will
be discussed later. pcAFM has been applied to perovskites [344] and recently integrated
with voltage biasing to produce spatially resolved J-V curves of the devices. [51] In these
measurements, individual maps of J-V parameters can be obtained, extremely valuable to
estimate relative changes in all figures-of-merit, with nanoscale spatial resolution. In order
to translate this information into macroscopic device parameters, one needs to take into
consideration (1) the extremely small volume of excitation when light focusing is used, (2)
the effect of the probe size as a current collector and charge carrier diffusion length, (3) the
photon flux in single wavelength measurements, (4) spatial variations in the absorption
coefficient of the material under investigation (which will necessarily affect charge carrier
generation), (5) and the possible optical losses of the experimental setup.
140
(a) pcAFM (c) NSOM
current flow
pre-amplifier (ii) hν long pass filter optical 
fiber
electronic Position (nm)h+
control 
e-
bottom contact
beam splitter spectrometer
(i) hν
Position (nm) excitation 
light
to control 
(b) KPFM and analysis
dark objective 
lens
qVCPD qVCPD
Position (nm)
W d W
il
p NSOM probe
E W p W e-vac n n LBIC 
Ec
Wm Wtip Wm W PL emission
flow
tip pre-amplifier
E +F h∆𝜇𝜇 e-
metal tip Position (nm)
solar cell
Ev + bottom contactp-type h
n-type x,y piezo stage
Figure 8.2: Photons as a source of excitation to nano-image PV functional-
ity through scanning probe microscopy. (a) In photoconductive-AFM (pcAFM)
a conductive probe acts as a local electrical contact. An external light source is used,
where the illumination direction, (i) or (ii), depends on device configuration. Both
surface topography and current are measured simultaneously. (b) Kelvin probe force
microscopy (KPFM) measures the work function (W ) difference (or contact potential
difference, VCPD) between the probe and the sample surface when grounded with re-
spect to one another. In the dark, d, the solar cell is in equilibrium and all the Fermi
levels are equal. Under illumination (il) there is a splitting of the quasi-Fermi level
(∆µ). The Voc is, thus, determined by subtracting dark- from light-KPFM scans.
Evac, Ec, EFn, EFp, and Ev denote the vacuum, conduction, the quasi-Fermi level
for e− and h+, and valence band energies, respectively. Adapted from [37]. (c) Near-
field scanning optical microscopy (NSOM) schematic showing two possible types of
measurements with spatial resolution beyond the diffraction limit: photoluminescence
(PL) and laser-beam-induced current (LBIC) microscopy.
While extensive efforts have been made to increase the Jsc of thin-film polycrys-
talline solar cells, higher efficiencies are still limited by the Voc (see Chapter 3). The local
variations of Voc within these materials can be determined by a much less explored and
equally powerful AFM-based method: Kelvin probe force microscopy (KPFM). [37, 39]
As previously explained, the work function difference between the sample surface and the
probe is measured, as illustrated in Figure 8.2(b). By subtracting an illuminated-KPFM
scan from a dark one, the quasi -Fermi level splitting (∆µ), and therefore, the local Voc gen-
141
Voc Topography Current Topography
erated by the material is obtained. This universal approach only requires a half-processed
device and can be performed in ambient conditions. With proper system calibration, the
quantitative Voc obtained from focused illumination is converted into a 1-sun Voc map un-
der AM1.5 global illumination. We anticipate Voc mapping through KPFM to be combined
with pcAFM (on the very same grains and boundaries) to quantify where charge carrier
recombination and collection take place, respectively. Combined, these measurements en-
able the deconvolution of both processes, critical to identify which mechanism is limiting
the performance of the device under investigation. Further, by implementing a correl-
ative microscopy approach, where the SPM methods are combined with ex situ EBSD,
one can elucidate which interfaces and grain orientations in polycrystalline materials are
responsible for the modest Voc of the technologies discussed earlier.
Functional imaging through NSOM has been applied less frequently to solar cells
when compared to AFM-based methods. Here, evanescent waves are used to break the far-
field diffraction limit. By keeping the distance between the probe and the sample surface
considerably smaller than the wavelength of light (at least two orders of magnitude), it
is possible to map the solar cells optical response with true nanoscale spatial resolution.
In NSOM, a probe or an optical fiber with a hollow center (approximately 60-200 nm in
diameter) is used as the excitation and/or collection source, as depicted in Figure 8.2(c).
One electrical measurement option using NSOM is laser-beam induced current (LBIC).
In this configuration, the photons transmitted through the probe induce photocurrent in
a highly localized region of the solar cell. This method has been applied to identify the
diffusion of sulfur in CdTe, [237] to spatially resolve the external quantum efficiency (EQE)
of CdTe solar cells, determining the contribution of grains and boundaries to the overall
signal of the device, [130,131,345] and the photo-response of grain and GBs in CIGS. [238]
142
For the quantitative comparison between the LBIC signal and the macroscopic Jsc of a
device, one must accurately determine the contribution of the dark saturation current of
the entire sample (whole device area), and extricate possible topographic artifacts due to
near-field light coupling. One attractive application of NSOM for perovskites is to realize
spectrally dependent LBIC scans to spatially monitor the existence of trap states within
the material, where ion migration seems to be a major contributor in their local electrical
response. [39]
A powerful alternative measurement is to use the NSOM probe for both the excita-
tion of carriers and the collection of emitted photons (photoluminescence, PL), resulting
from radiative recombination. The collection of the PL signal strongly depends on light-
probe coupling, and considerable effort has focused on designing optimized probes for
capturing the emitted photons. For instance, campanile tips have been identified as an
excellent geometry to enhance the coupling efficiency between far and near fields, allowing
for PL signal imaging of nanoscale features in InP nanowires for PV applications. [129]
PL microscopy with nanoscale spatial resolution can be realized to probe the impact of
GaAs polycrystals’ facets on carrier recombination. Note that device processing is not
required here. Different than InP, GaAs is a material that is not self-passivated. Thus,
one could use PL through NSOM to identify the effect of distinct surface passivation
processes (e.g. Na2S and Si3N4) to the recombination rate within the different facets com-
posing the GaAs polycrystals. Although unconventional, NSOM can also be integrated
into a SEM. In this dual probe approach the NSOM tip collects the near-field CL emission
caused by the e-beam to image the energy transport mechanisms without being limited
by photon recycling or excess carrier diffusion. [346] We foresee that combining LBIC and
PL signals with simulations of the electrodynamics at the near-field regime could enable a
143
tomography of charge carrier generation, recombination, and collection, resolving the role
of the mesoscale constructs within polycrystalline solar cells to their macroscopic device
performance.
Time-resolved PL (tr-PL) is an extremely powerful method to quantify the minority
carriers lifetime within a material. Here, the carriers that recombine radiatively (PL
signal) are detected by a time-correlating photon counting system. tr-PL spectroscopy
has been widely implemented to resolve the effect of passivation methods on the lifetime
of polycrystalline materials for PV, [347, 348] as well as the role of precursor solution
concentration (and band gap) in halide perovskites. [349, 350] When combined with an
NSOM microscope, it provides maps of the temporal dynamics of the PL signal, with
sub-wavelength spatial resolution. tr-PL through confocal optical microscopy [351, 352]
or NSOM [353] could be implemented to investigate the spatial distribution of minority
carrier lifetimes in perovskite grains and its photo-instability. A primary advantage of
tr-PL microscopy is that it enables measurements in a film, not requiring any device
processing.
8.4 Probing the degradation of perovskite solar cells
The commonly observed degradation of perovskite solar cells is the main barrier for
this emerging material to become a reliable PV technology, as previously mentioned in
Chapters 6 and 7. Thus, measuring and controlling the physical and chemical processes
responsible for the transient electrical behavior of perovskites is critical for its widespread
deployment. [50] we foresee pcAFM and illuminated-KPFM being implemented to track
material degradation. For instance, MAPbX3 (X = Br or I) is unstable under illumina-
tion, and can suffer from structural phase transition at low temperatures (<60 oC), and
144
often degrades upon exposure to moisture. [146] Thus, SPM under controlled illumination,
humidity, and temperature conditions will help elucidate the driving forces behind the ma-
terial degradation commonly observed in perovskites. For example, spectrally dependent
KPFM will likely reveal the possible presence of trap states within the material, and how
their density varies from grain to grain. The reversibility of the electrical response of
perovskites can also be probed by tuning the humidity (low, high, low) upon sequential
SPM measurements. Here, pcAFM and KPFM could resolve how both Jsc and Voc change
as the perovskite is exposed to moisture. The advantage of using microscopic methods to
probe perovskite solar cells lies in quantifying spatial variations within the samples, which
could be associated with chemical composition deviations from grain to grain.
The dynamic electrical behavior of the perovskites has been measured both in plane-
view and in cross-section by pcAFM [90] and KPFM. [39, 151] However, there is still a
pressing need for scientists to discover the driving forces responsible for the devices tran-
sient behavior when the material is exposed to light and humidity. This dynamic response
likely requires the further development of time-dependent functional microscopy methods
that will provide 4-dimensional (4D) information, space and time, at relevant length scales.
Mimicking realistic operation conditions (including the environment) of perovskite devices
during operando measurements is necessary to identify why and how the material is de-
grading. Preferably, these 4D measurements should be combined with chemical imaging
to establish a correlation between device behavior and the changes in chemical composi-
tion. The combination of illuminated-KPFM (as described earlier) with Raman imaging
and other AFM-based chemical mapping is a promising route to elucidate the effect of
photo-induced chemical reactions in device degradation. As recently demonstrated, [39]
fast-KPFM scans (16 second/frame) that allow real-time measurements of local Voc have
145
revealed (1) substantial ion migration during perovskites illumination, and (2) a residual
post-illumination voltage resulting from a time dependent ion migration current never
measured before. These fast scans can be extended to lead-free perovskite films, [354] e.g.
MASnI3 [355] or Cs2BiAgCl6, [53] as well as new compounds for multi-junction PV, such
as Rb-FA0.75MA0.15Cs0.1PbI2Br, [356] to determine their dynamic response and, thus,
material stability when exposed to light.
8.5 Correlative microscopy and the need for big data ana-
lytics
We foresee that the use of combined functional imaging microscopies will produce
large amounts of data that must be effectively analyzed for an understanding of the mate-
rial behaviors under investigation. An emerging investigation pathway is big data, which
encompasses the analysis of all data to identify critical patterns and associations (e.g.
correlation between grain boundaries and low Voc). Big data analytics involves evaluating
the detected signals with approaches from multivariate statistics and machine learning to
identify meaningful relationships between material properties, obtained from a set of mi-
croscopy measurements in this case. [357] Ultimately, the main advantage of the available
paradigms [298] is to gain insight into complex phenomena while minimizing computational
time. For example, using principal component analysis (PCA), [358] one can determine
which dimensions of the data account for the most variance in an imaging dataset. Note
that in this situation, each physical quantity is assigned as one data dimension. This
method has already proven useful when combined with band excitation magnetic force
microscopy; PCA provided the same qualitative insight as the simple harmonic oscillator
146
model while requiring ∼ 104 less computation time. [358] Machine learning approaches
will likely help scientists determine which defects and grain boundaries in polycrystalline
PV materials contribute most to performance deficiencies. For instance, a combination of
NSOM LBIC and ex situ EBSD data acquired at the same region of a sample could train
an algorithm to predict the electrical signal based solely on the angle between two adjacent
grain boundaries. Data mining may also identify degradation mechanisms in perovskite
solar cells. Namely, factor analysis can be applied to a combination of illuminated-KPFM
and tip-enhanced Raman spectroscopy under a controlled atmosphere to establish how the
electrical and chemical responses vary as a function of relative humidity. Further, time-
dependent photocurrent and photovoltage measurements of emerging lead-free perovskite
alternatives could provide a roadmap for stable materials. For the reasons outlined, we an-
ticipate the use of big data analytics on information produced by combinatorial functional
microscopy.
8.6 Outlook
In perovskites, device stability is currently the main barrier towards implementing
a reliable technology, followed by the pressing need to identify non-toxic lead-free, stable
options. The full potential of thin-film polycrystalline materials for PV is primarily lim-
ited by their Voc, which is strongly related to non-radiative recombination events within
the material. As the grains and boundaries forming the devices are acknowledged to affect
their electrical response it is imperative that we identify which types of grains and inter-
faces contribute to charge carrier non-radiative recombination. In order to determine the
local electrical properties of the mesoscale constituents of both perovskites and inorganic
thin-film PV, researchers must implement microscopy techniques beyond their standard
147
operation modes, quantifying the figures-of-merit (and, hence, the device performance) at
relevant length scales. Further, the large amount of generated data must be analyzed in
an effective manner using approaches based on big data analytics, leading to quantitative
information about the devices under investigation.
We expect the realization of correlative microscopy and the functional imaging of
materials/devices at the mesoscale to finally elucidate some critical open questions in the
field, such as: What are the light-induced chemical reactions responsible for perovskites’
degradation under 1-sun illumination? What is the role of humidity in perovskite’s per-
formance? Which types of interfaces in CdTe, CIGS, and CZTS are responsible for the
low carrier radiative recombination rate that limits their Voc? How do the different facets
composing GaAs polycrystals affect charge carrier generation, recombination, and collec-
tion? Is the diode ideality factor for all grain boundaries the same in a polycrystalline
PV device? To answer these questions, among many others, it is essential to deter-
mine the physical behavior of the mesoscale constructs when the devices are in operando.
Functional imaging methods will likely act as the central “covalent bond” between the
materials science structure-properties-processing-performance tetrahedron cornerstones,
providing unprecedented correlative insights into materials for PV at the mesoscale. In
turn, these sophisticated measurements can potentially allow the rational design of higher
performance PV devices based on low-cost materials/methods.
8.7 Future work
An exciting next direction for this research project will be to implement correlative
microscopy, as described above, to investigate how the local chemical distribution within
a perovskite material effects the dynamic nanoscale voltage response, which we previously
148
measured and quantified in Chapters 6 and 7. For this, I recently performed photother-
mal infrared resonance (PTIR) microscopy (also known as AFM-IR discussed in Chapter
3, Section 3.2.3), on a perovskite thin-film sample that contains an additional solution,
termed additive, see Figure 8.3. The inclusion of the additive shows an increase in the
macroscopic Voc, however, how it influences the electrical properties at the nanoscale is
still unknown.
Figure 8.3: Imaging the nanoscale chemical distribution of a perovskite
thin-film. AFM topography overlaid with the chemical map that was simultaneously
measured, here, the sample was excited by wavenumber = 1609 cm−1.
The preliminary measurements reveal ∼1 × 1 µm2 regions of distinct chemical
variations induced by the presence of the additive; this was proven by acquiring AFM-IR
on a sample without the additional solution (not shown here). A clear next step will be
to apply the illuminated-KPFM technique developed in this thesis on the same sample to
correlate the chemical variations to the nanoscale Voc.
149
150
Appendix A: Products of this research
A.1 Awards and honors
19. MRS Graduate Student Award (Silver) 10/2017
18. IEEE 44th Photovoltaic Specialist Conference Best Student Paper Finalist 05/2017
17. OSA Student Poster Cometition - honorable mention 04/2017
16. UMD All-S.T.A.R. fellowship 04/2017
15. UMD graduate student summer research fellowship 2017 04/2017
14. ARPA-e student program at the energy innovation summit, Washington, D.C. 01/2017
13. OSA Student Leadership Conference, Rochester, NY - UMD representative 10/2016
12. Graduate Student Deans Dissertation Fellowship 04/2016
11. Selected for Joint US-Africa Materials Institute (JUAMI) 2016 04/2016
10. Hulka Energy Research Fellowship 02/2016
9. MRS Graduate Student Exchange at NC - Chapel Hill Seminar, invited talk 02/2016
8. Institute of Research in Electronics and Applied Physics (IREAP) - best speaker 12/2015
7. UMD ResearchFest 2015 - overall winner 06/2015
6. UMD graduate student summer research fellowship 2015 04/2015
5. OSA student chapter poster competition - 1st place poster 04/2015
4. NSF graduate research fellowship program honorable mention 04/2015
3. UMD Jacob K. Goldhaber travel grant 01/2015
2. IEEE 40th Photovoltaic Specialist Conference (PVSC) - best poster 06/2014
1. UMD Dean’s Fellowship 08/2013 – 05/2017
A.2 Publications
8. J. M. Howard, E. M. Tennyson, S. Barik, E. Waks, M. Toney, A. F. Nogueira,
B. R. A. Neves, M. S. Leite, Humidity-dependent Photoluminescence Hysteresis in
Cs-containing Perovskite Thin Films – in progress
7. E. M. Tennyson, B. Roose, J. L. Garrett, C. Gong, J. N. Munday, A. Abate,
M. S. Leite, Caesium-incorporated triple cation perovskites deliver fully reversible and
stable nanoscale voltage response – in progress
6. E. M. Tennyson, C. Gong, M. S. Leite, Imaging Energy Harvesting and Storage
Systems at the Nanoscale, ACS Energy Letters, 2, 2761-2777 (2017) – invited review,
front cover
5. E. M. Tennyson, J. M. Howard, M. S. Leite, Mesoscale Functional Imaging
of Materials for Photovoltaics, ACS Energy Letters, 2, 1825-1834 (2017) – invited
perspective
4. J.L. Garrett, E. M. Tennyson, J.N. Munday, M.S. Leite, Real-time Nanoscale
open-circuit voltage dynamics of perovskite solar cells, Nano Letters, 17, 2554-2560,
(2017)
3. E. M. Tennyson, J.A. Frantz, J.M. Howard, W.B. Gunnarsson, J.D. Myers, R.Y.
Bekele, S. Na, M.S. Leite, Photovoltage tomography in polycrystalline solar cells, ACS
Energy Letters, 1, 899-905 (2016)
151
2. E. M. Tennyson, J.L. Garrett, J.A. Frantz, J.D. Myers, R.Y. Bekele, J.S. Sanghera, J.N.Munday,
M.S. Leite, Nanoimaging of open-circuit voltage in photovoltaic devices, Advanced Energy Mat-
erials, 5, 1501142 (2015) – front cover
1. E. M. Tennyson, M.S. Leite, Mapping the performance of solar cells with nanoscale resolution,
SPIE Newsroom doi:10.1117/2.1201509.006097 (2015)
Conference Proceedings:
5. E. M. Tennyson, M.S. Leite, 44th IEEE PVSC Proceedings: Nano-imaging of Performance
in Photovoltaics (06/2017)
4. E. M. Tennyson, J. L. Garrett, J. N. Munday, M. S. Leite, 43rd IEEE PVSC Proceedings:
Mapping Voc in Polycrystalline Solar Cells with Nanoscale Spatial Resolution (06/2016)
3. E. M. Tennyson, J.L. Garrett, C. Gong, J.A. Frantz, J.D. Myers, R.Y. Bekele, J.S. Sanghera,
J.N. Munday, M.S. Leite, SPIE 9561, Thin Films for Solar and Energy Technology VII Pro-
ceedings: A novel method for mapping open-circuit voltage in solar cells with nanoscale resolu-
tion (10/2015)
2. E. M. Tennyson, J.A. Frantz, J.D. Myers, R.Y. Bekele, J.S. Sanghera, S. Na, M.S. Leite,
42nd IEEE PVSC Proceedings: Imaging EQE in CIGS solar cells with high spatial resolution
(06/2015)
1. E. M. Tennyson, J.L. Garrett, C. Gong, J.A. Frantz, J.D. Myers, R.Y. Bekele, J.S. Sanghera,
J.N. Munday, M.S. Leite, 40th IEEE PVSC Proceedings: Assessing local voltage in CIGS solar
cells by nanoscale resolved Kelvin probe force microscopy and sub-mircon photoluminescence
(06/2014)
A.3 Presentations
29. E. M. Tennyson, M. S. Leite, Functional Imaging of Photovoltaics at the Nanoscale: from
Perovskites to Polycrystalline Materials. Fall MRS 2017 (11/2017) talk and MRS Graduate
Student Silver Award winner
28. E. M. Tennyson, J. L. Garrett, J. N. Munday, M. S. Leite, Real-Time Mapping of Perovskite
Intragrain Instabilities at the Nanoscale. Fall MRS 2017 (11/2017) talk
27. E. M. Tennyson, M.S. Leite, Imaging photovoltaic functionality of polycrystalline and
perovskite solar cells at the nanoscale. ACS 254th National Meeting (08/2017) talk and poster
26. E. M. Tennyson, M.S. Leite, Nano-imaging of Performance in Photovoltaics. 44th IEEE
PVSC (06/2017) best student award finalist talk
25. E. M. Tennyson, M.S. Leite, Imaging the mesoscale electrical response of polycrystalline
and perovskite solar cells. UMD ResearchFest 2017 (06/2017) poster
24. E. M. Tennyson, M.S. Leite, Mapping the performance of polycrystalline and perovskite solar
cells with nanoscale spatial resolution. OSA poster competition (04/2017) poster
23. E. M. Tennyson, M.S. Leite, Open-circuit voltage mapping of next-generation solar cells.
IREAP Graduate Student Seminar (04/2017) talk
22. E. M. Tennyson, M.S. Leite, High spatial resolution performance mapping in thin-film solar
cells. IREAP Graduate Student Seminar (11/2016) talk
21. E. M. Tennyson, M.S. Leite, Mapping the local voltage response of polycrystalline solar cells
with nanoscale resolution. Frontier in Optics Conference (10/2016) poster
20. E. M. Tennyson, M.S. Leite, Spectrally dependent photovoltage tomography map for charge
carrier recombination study in CIGS solar cells. IREAP Graduate Student Seminar (04/2016)
talk
19. E. M. Tennyson, M.S. Leite, Imaging the electrical properties of thin-film solar cells with
high spatial resolution. UMD Nanoday (04/2016) poster
18. E. M. Tennyson, M.S. Leite, Spectrally dependent photovoltage tomography map for charge
carrier recombination study in CIGS solar cells. IREAP Graduate Student Seminar (04/2016)
talk
152
17. E. M. Tennyson, M.S. Leite, Kelvin probe force microscopy: mapping the local open-circuit
voltage of solar cells with nanoscale resolution. NC Chapel Hill Seminar (03/2016) invited talk
16. E. M. Tennyson, J.L. Garrett, J.N. Munday, M.S. Leite Resolving local voltage variations in
opto-electronic devices with Kelvin probe force microscopy. APS Baltimore (03/2016) talk
15. E. M. Tennyson, M.S. Leite High resolution mapping of the local performance parameters in
solar cells. Center for Nanophysics and Advanced Materials Seminar (02/2016) talk
14. E. M. Tennyson, M.S. Leite Resolving electrical performance variations of polycrystalline
solar cells at the nanoscale. IREAP Graduate Student Seminar (10/2015) talk
13. E. M. Tennyson, J.L. Garrett, J.A. Frantz, J.D. Myers, R.Y. Bekele, J.S. Sanghera, J.N.
Munday, M.S. Leite, Mapping efficiency in solar cells with nanoscale resolution. SPIE Optics
and Photonics (08/2015) talk
12. E. M. Tennyson, J.L. Garrett, J.A. Frantz, J.D. Myers, R.Y. Bekele, J.S. Sanghera, J.N.
Munday, M.S. Leite High resolution mapping of the open-circuit voltage of solar cells using
Kelvin probe force microscopy. IEEE Photonics: Women in Photonics event (06/2015) invited
talk
11. E. M. Tennyson, J.L. Garrett, J.A. Frantz, J.D. Myers, R.Y. Bekele, J.S. Sanghera, J.N.
Munday, M.S. Leite Mapping the open-circuit voltage of solar cells with nanoscale resolution.
UMD ResearchFest 2015 (06/2015) poster and talk
10. E. M. Tennyson, J.A. Frantz, J.D. Myers, R.Y. Bekele, J.S. Sanghera, S.M. Na, A. Flatau,
M.S. Leite, Imaging EQE in CIGS solar cells with high spatial resolution. 42nd IEEE PVSC
(06/2015) talk
9. E. M. Tennyson, J.L. Garrett, J.A. Frantz, J.D. Myers, R.Y. Bekele, J.S. Sanghera, J.N.
Munday, M.S. Leite Mapping the open-circuit voltage of solar cells with nanoscale resolution.
Annual Science Fair and Graduate Student Poster Award Dinner for NCS-OSA and DC-N.
Virginia IEEE photonics societies (05/2015) poster
8. E. M. Tennyson, M.S. Leite, High resolution mapping of solar cell functionality. Electrical
and Computer Engineering Graduate Student Seminar (05/2015) talk
7. E. M. Tennyson, M.S. Leite, High resolution mapping of solar cell functionality. Institute of
Research in Electronics and Applied Physics Graduate Student Seminar (05/2015) talk
6. E. M. Tennyson, J.L. Garrett, J.A. Frantz, J.D. Myers, R.Y. Bekele, J.S. Sanghera, J.N.
Munday, M.S. Leite, Imaging open-circuit voltage and external quantum efficiency in solar cel-
ls with nanoscale resolution. Materials Research Society meeting (04/2015) talk
5. E.M. Tennyson, J.L. Garrett, J.A. Frantz, J.D. Myers, R.Y. Bekele, J.S. Sanghera, J.N.
Munday, M.S. Leite Quantitative imaging of open-circuit voltage in opto-electronic devices
with nanoscale resolution. SPIE Sensing Technology + Applications conference (04/2015) talk
4. E. M. Tennyson, J.L. Garrett, J.A. Frantz, J.D. Myers, R.Y. Bekele, J.S. Sanghera, J.N.
Munday, M.S. Leite, Mapping the open-circuit voltage of solar cells with nanoscale resolution.
OSA poster competition (04/2015) poster
3. E. M. Tennyson, J.L. Garrett, J.A. Frantz, J.D. Myers, R.Y. Bekele, J.S. Sanghera, J.N.
Munday, M.S. Leite, Kelvin probe force microscopy: Imaging open-circuit voltage in optoelec-
tronic devices. American Physical Society March Meeting (03/2015) talk
2. E. M. Tennyson, Characterization of solar cells with nano-scale resolution. IREAP Graduate
Student Seminar (10/2014) talk
1. E. M. Tennyson, J.L. Garrett, C. Gong, J.A. Frantz, J.D. Myers, R.Y. Bekele, J.S. Sanghera,
J.N. Munday, M.S. Leite, Assessing local voltage in CIGS solar cells by nanoscale resolved
Kelvin probe force microscopy and sub-micron photoluminescence. 40th IEEE PVSC (06/2014)
best poster
153
154
Appendix B: Additional publications
2. Y. Xu, E. M. Tennyson, J. Kim, S. Barik, J. Murray, E. Waks, M. S. Leite, and J. N.
Munday, Active control of photon recycling for tunable optoelectronic device, Advanced Optical
Materials, (2018) DOI: 10.1002/adom.201701323
1. L. J. Krayer, E. M. Tennyson, M. S. Leite, J. N. Munday, Near-IR imaging based on hot
carrier generation in nanometer-scale optical coatings, ACS Photonics, (2018)
DOI: 10.1021/acsphotonics.7b01021 – front cover
155
156
Appendix C: University of Maryland service
C.1 Contributions
10. Vice President of OSA Student Chapter at UMD July 2017 - May 2018
9. Restricted research sub-committee member Aug. 2016 - Feb. 2017
8. Volunteer for Women in Engineering DREAM conference Years 2015, 2016, and 2017
7. Short talk (10 min) at Materials Science and Engineering alumni event April 2015
6. Presented at MSE open house for prospective PhD students March 2015
5. Participant in Women in Engineering day Years 2014, 2015, 2016, and 2017
4. Volunteer for Maryland Day Years 2014, 2015, 2016 and 2017
3. Socialized with prospective PhD students at open house dinners
Years 2014, 2015, 2015, and 2017
2. Organized UMD ResearchFest June 2014
1. President of Materials Science Graduate Society Spring 2014
157

Bibliography
[1] National Renewable Energy Laboratories. Renewable Electricity Futures
Study. http://www.nrel.gov/analysis/re_futures, 2016. [Online; ac-
cessed 20-Oct-2016].
[2] D. J. Norris and E. S. Aydil. Getting Moore from solar cells. Science,
338(6107):625–626, 2012.
[3] Solar Energy Industries Assosciaton. Solar Industry Data. http://www.seia.
org/Research-Resources/Solar-Industry-Data, 2016. [Online; accessed
20-Oct-2016].
[4] M. A. Green, Y. Hishikawa, W. Warta, E. D. Dunlop, D. H. Levi, J. Hohl-
Ebinger, and A. W. Y. Ho-Baillie. Solar cell efficiency tables (version 50).
Progress in Photovoltaics: Research and Applications, 25(7):668–676, 2017.
[5] National Renewable Energy Laboratories. NREL Best Research-Cell Efficien-
cies. https://www.nrel.gov/pv/assets/images/efficiency-chart.png,
2018. [Online; accessed 04-Feb-2018].
[6] W. Shockley and H.J. Queisser. Detailed balance limit of p-n junction solar
cells. Journal of Applied Physics, 32:510, 1961.
[7] P. K. Nayak, J. Bisquert, and D. Cahen. Assessing possibilities and limits for
solar cells. Advanced Materials, 23(25):2870–2876, 2011.
[8] H. Xin, J. K. Katahara, I. L. Braly, and H. W. Hillhouse. 8% effi-
cient Cu2ZnSn(S,Se)4 solar cells from redox equilibrated simple precursors
in DMSO. Advanced Energy Materials, 4(11):1301823, 2014.
[9] W. Ki and H. W. Hillhouse. Earth-abundant element photovoltaics directly
from soluble precursors with high yield using a non-toxic solvent. Advanced
Energy Materials, 1(5):732–735, 2011.
[10] Y.-T. Hsieh, Q. Han, C. Jiang, T.-B. Song, H. Chen, L. Meng, H. Zhou,
and Y. Yang. Efficiency enhancement of Cu2ZnSn(S,Se)4 solar cells via alkali
metals doping. Advanced Energy Materials, 6(7):1502386, 2016.
[11] W.-C. Hsu, H. Zhou, S. Luo, T.-B. Song, Y.-T. Hsieh, H.-S. Duan, S. Ye,
W. Yang, C.-J. Hsu, C. Jiang, B. Bob, and Y. Yang. Spatial element
159
distribution control in a fully solution-processed nanocrystals-based 8.6%
Cu2ZnSn(S,Se)4 device. ACS Nano, 8(9):9164–9172, 2014.
[12] V. Steinmann, R. E. Brandt, and T. Buonassisi. Photovoltaics: Non-cubic
solar cell materials. Nature Photonics, 9(6):355–357, 2015.
[13] Y. Zhou, L. Wang, S. Chen, S. Qin, X. Liu, J. Chen, D.-J. Xue, M. Luo,
Y. Cao, Y. Cheng, E. H. Sargent, and J. Tang. Thin-film Sb2Se3 photovoltaics
with oriented one-dimensional ribbons and benign grain boundaries. Nature
Photonics, 9(6):409–415, 2015.
[14] J. H. Yun, K. H. Kim, M. S. Kim, B. T. Ahn, S. J. Ahn, J. C. Lee, and K. H.
Yoon. Fabrication of CIGS solar cells with a Na-doped Mo layer on a Na-free
substrate. Thin Solid Films, 515(15):5876–5879, 2007.
[15] W. Jaegermann, A. Klein, and T. Mayer. Interface engineering of inorganic
thin-film solar cells - materials-science challenges for advanced physical con-
cepts. Advanced Materials, 21(42):4196–4206, 2009.
[16] A. Romeo, M. Terheggen, D. Abou-Ras, D. L. Bätzner, F.-J. Haug, M. Kälin,
D. Rudmann, and A. N. Tiwari. Development of thin-film Cu(In,Ga)Se2 and
CdTe solar cells. Progress in Photovoltaics: Research and Applications, 12:93–
111, 2004.
[17] M. Kaelin, D. Rudmann, F. Kurdesau, H. Zogg, T. Meyer, and A. N. Tiwari.
Low-cost CIGS solar cells by paste coating and selenization. Thin Solid Films,
480:486–490, 2005.
[18] T. Nakada, Y. Hirabayashi, T. Tokado, D. Ohmori, and T. Mise. Novel device
structure for Cu(In,Ga)Se2 thin film solar cells using transparent conducting
oxide back and front contacts. Solar Energy, 77(6):739–747, 2004.
[19] N. Kohara, S. Nishiwaki, Y. Hashimoto, T. Negami, and T. Wada. Electrical
properties of the Cu(In,Ga)Se2/MoSe2/Mo structure. Solar Energy Materials
and Solar Cells, 67(1-4):209–215, 2001.
[20] K. Orgassa, H. W. Schock, and J. H. Werner. Alternative back contact ma-
terials for thin film Cu(In,Ga)Se2 solar cells. Thin Solid Films, 431:387–391,
2003.
[21] K. D. Dobson, I. Visoly-Fisher, G. Hodes, and D. Cahen. Stability of
CdTe/CdS thin-film solar cells. Solar Energy Materials and Solar Cells,
62(3):295–325, 2000.
[22] W. Monch. Barrier heights of real Schottky contacts explained by metal-
induced gap states and lateral inhomogeneities. Journal of Vacuum Science
and Technology B, 17(4):1867–1876, 1999.
160
[23] Jenny Nelson. The Physics of Solar Cells (Properties of Semiconductor Ma-
terials). Inperial College Press, 2003.
[24] O. D. Miller, E. Yablonovitch, and S. R. Kurtz. Strong internal and exter-
nal luminescence as solar cells approach the Shockley-Queisser limit. IEEE
Journal of Photovoltaics, 2:303–311, 2012.
[25] R. T. Ross. Some thermodynamics of photochemical systems. Journal of
Chemical Physics, 46:45904593, 1967.
[26] M. A. Green. Radiative efficiency of state-of-the-art photovoltaic cells.
Progress in Photovoltaics: Research and Applications, 20:472476, 2012.
[27] L. C. Hirst and N. J. Ekins-Daukes. Fundamental losses in solar cells. Progress
in Photovoltaics: Research and Applications, 19:286–293, 2011.
[28] E. M Tennyson, C. Gong, and M. S. Leite. Imaging energy harvesting and
storage systems at the nanoscale. ACS Energy Letters, 2(12):27612777, 2017.
[29] G. Binnig, H. Rohrer, C. Gerber, and E. Weibel. Surface studies by scanning
tunneling microscopy. Physical Review Letters, 49(1):57–61, 1982.
[30] G. Binnig, C. F. Quate, and C. Gerber. Atomic force microscope. Physical
Review Letters, 56(9):930–933, 1986.
[31] F. J. Giessibl. Advances in atomic force microscopy. Reviews of Modern
Physics, 75(3):949–983, 2003.
[32] R. Garcia and E. T. Herruzo. The emergence of multifrequency force mi-
croscopy. Nature Nanotechnology, 7(4):217–226, 2012.
[33] M. Z. Baykara, T. C. Schwendemann, E. I. Altman, and U. D. Schwarz. Three-
dimensional atomic force microscopy – taking surface imaging to the next level.
Advanced Materials, 22(26-27):2838–2853, 2010.
[34] E. M. Tennyson, J. M. Howard, and M. S. Leite. Mesoscale functional imaging
of materials for photovoltaics. ACS Energy Letters, 2:1825–1834, 2017.
[35] M. A. Green, Y. Hishikawa, E. D. Dunlop, D. H. Levi, J. Hohl-Ebinger, and
A. W.Y. Ho-Baillie. Solar cell efficiency tables (version 51). Progress in Pho-
tovoltaics: Research and Applications, 26(1):3–12, 2018.
[36] Department of Energy: From quanta to the continuum: Opportunities for
mesoscale science. (2012).
[37] E. M. Tennyson, J. L. Garrett, J. A. Frantz, J. D. Myers, R. Y. Bekele,
Jasbinder S. Sanghera, J. N. Munday, and M. S. Leite. Nanoimaging of
open-circuit voltage in photovoltaic devices. Advanced Energy Materials,
5(23):1501142, 2015.
161
[38] J. D. Poplawsky, W. Guo, N. Paudel, A. Ng, K. More, D. Leonard, and Y. Yan.
Structural and compositional dependence of the CdTexSe1−x alloy layer pho-
toactivity in CdTe-based solar cells. Nature Communications, 7:12537, 2016.
[39] J. L. Garrett, E. M. Tennyson, M. Hu, J. Huang, J. N. Munday, and M. S.
Leite. Real-time nanoscale open-circuit voltage dynamics of perovskite solar
cells. Nano Letters, 17(4):2554–2560, 2017.
[40] E. M. Tennyson, J. A. Frantz, J. M. Howard, Gunnarsson W. B., J. D. Myers,
R. Y. Bekele, S. M. Na, and M. S. Leite. Tomography of charge carrier
collection in polycrystalline solar cells. ACS Energy Letters, 1:899–905, 2016.
[41] D. W. de Quilettes, W. Zhang, V. M. Burlakov, D. J. Graham, T. Leijtens,
A. Osherov, V. Bulovic, H. J. Snaith, D. S. Ginger, and S. D. Stranks. Photo-
induced halide redistribution in organic-inorganic perovskite films. Nature
Communications, 7, 2016.
[42] Y. Xie, F. Shao, Y. Wang, T. Xu, D. Wang, and F. Huang. Enhanced per-
formance of perovskite CH3NH3PbI3 solar cell by using CH3NH3I as additive
in sequential deposition. ACS Applied Materials & Interfaces, 7(23):12937–
12942, 2015.
[43] D.-H. Son, D.-H. Kim, S.-N. Park, K.-J. Yang, D. Nam, H. Cheong, and J.-K.
Kang. Growth and device characteristics of CZTSSe thin-film solar cells with
8.03% efficiency. Chemistry of Materials, 27(15):5180–5188, 2015.
[44] A. L. Greenaway, M. C. Sharps, J. W. Boucher, L. E. Strange, M. G. Kast,
S. Aloni, and S. W. Boettcher. Selective area epitaxy of GaAs microstructures
by close-spaced vapor transport for solar energy conversion applications. ACS
Energy Letters, 1(2):402–408, 2016.
[45] A. Kojima, K. Teshima, Y. Shirai, and T. Miyasaka. Organometal halide
perovskites as visible-light sensitizers for photovoltaic cells. Journal of the
American Chemical Society, 131(17):6050–6051, 2009.
[46] J. Seo, J. H. Noh, and S. I. Seok. Rational strategies for efficient perovskite
solar cells. Accounts of Chemical Research, 49(3):562–572, 2016.
[47] T.-B. Song, T. Yokoyama, S. Aramaki, and M. G. Kanatzidis. Perfor-
mance enhancement of lead-free tin-based perovskite solar cells with reducing
atmosphere-assisted dispersible additive. ACS Energy Letters, 2(4):897–903,
2017.
[48] J. Bisquert, Y. Qi, T. Ma, and Y. Yan. Advances and obstacles on perovskite
solar cell research from material properties to photovoltaic function. ACS
Energy Letters, 2(2):520–523, 2017.
162
[49] W. Tress, N. Marinova, T. Moehl, S. M. Zakeeruddin, Mohammad Khaja
Nazeeruddin, and M. Grätzel. Understanding the rate-dependent J-V hys-
teresis, slow time component, and aging in CH3NH3PbI3 perovskite solar cells:
the role of a compensated electric field. Energy & Environmental Science,
8(3):995–1004, 2015.
[50] J. Berry, T. Buonassisi, D. A. Egger, G. Hodes, L. Kronik, Y.-L. Loo,
I. Lubomirsky, S. R. Marder, Y. Mastai, J. S. Miller, D. B. Mitzi, Y. Paz,
A. M. Rappe, I. Riess, B. Rybtchinski, O. Stafsudd, V. Stevanovic, M. F.
Toney, D. Zitoun, A. Kahn, D. Ginley, and D. Cahen. Hybrid organic-
inorganic perovskites (HOIPs): Opportunities and challenges. Advanced Ma-
terials, 27:5102–5112, 2015.
[51] Y. Kutes, Y. Zhou, J. L. Bosse, J. Steffes, N. P. Padture, and B. D. Huey.
Mapping the photoresponse of CH3NH3PbI3 hybrid perovskite thin films at
the nanoscale. Nano Letters, 16(6):3434–3441, 2016.
[52] E. Edri, S. Kirmayer, S. Mukhopadhyay, K. Gartsman, G. Hodes, and D. Ca-
hen. Elucidating the charge carrier separation and working mechanism of
CH3NH3PbI3−xClx perovskite solar cells. Nature Communications, 5, 2014.
[53] F. Giustino and H. J. Snaith. Toward lead-free perovskite solar cells. ACS
Energy Letters, 1(6):1233–1240, 2016.
[54] Y. Zhao, M. Boccard, S. Liu, J. Becker, X.-H. Zhao, C. M. Campbell,
E. Suarez, M. B. Lassise, Z. Holman, and Y.-H. Zhang. Monocrystalline CdTe
solar cells with open-circuit voltage over 1V and efficiency of 17%. Nature
Energy, 1:16067, 2016.
[55] J. Moseley, W. K. Metzger, H. R. Moutinho, N. Paudel, H. L. Guthrey, Y. Yan,
R. K. Ahrenkiel, and M. M. Al-Jassim. Recombination by grain-boundary type
in CdTe. Journal of Applied Physics, 118(2):025702, 2015.
[56] D. Abou-Ras, S. S. Schmidt, N. Schäfer, J. Kavalakkatt, T. Rissom, T. Unold,
R. Mainz, A. Weber, T. Kirchartz, E. Simsek Sanli, P. A. van Aken, Q. M.
Ramasse, H.-J. Kleebe, D. Azulay, I. Balberg, O. Millo, O. Cojocaru-Mirédin,
D. Barragan-Yani, K. Albe, J. Haarstrich, and C. Ronning. Compositional
and electrical properties of line and planar defects in Cu(In,Ga)Se2 thin films
for solar cells. physica status solidi (RRL) Rapid Research Letters, pages
1–13, 2016.
[57] R. Kamada, T. Yagioka, S. Adachi, A. Handa, K. F. Tai, T. Kato, and H. Sugi-
moto. New world record Cu(In,Ga)(Se,S)2 thin film solar cell efficiency beyond
22%. In 2016 IEEE 43rd Photovoltaic Specialists Conference (PVSC), pages
1287–1291.
163
[58] A. Walsh, S. Chen, S.-H. Wei, and X.-G. Gong. Kesterite thin-film solar cells:
Advances in materials modelling of Cu2ZnSnS4. Advanced Energy Materials,
2(4):400–409, 2012.
[59] A. Polizzotti, I. L. Repins, R. Noufi, S.-H. Wei, and D. B. Mitzi. The state and
future prospects of kesterite photovoltaics. Energy & Environmental Science,
6(11):3171–3182, 2013.
[60] W. Wang, M. T. Winkler, O. Gunawan, T. Gokmen, T. K. Todorov, Y. Zhu,
and D. B. Mitzi. Device characteristics of CZTSSe thin-film solar cells with
12.6% efficiency. Advanced Energy Materials, 4(7):1301465, 2014.
[61] C.-Y. Su, C. Yen Chiu, and J.-M. Ting. Cu2ZnSnS4 absorption layers with
controlled phase purity. Scientific Reports, 5:9291, 2015.
[62] M. S. Leite, R. L. Woo, W. D. Hong, D. C. Law, and H. A. Atwater. Wide-
band-gap InAlAs solar cell for an alternative multijunction approach. Applied
Physics Letters, 98(9):093502, 2011.
[63] H. Cotal, C. Fetzer, J. Boisvert, G. Kinsey, R. King, P. Hebert, H. Yoon,
and N. Karam. III-V multijunction solar cells for concentrating photovoltaics.
Energy & Environmental Science, 2(2):174–192, 2009.
[64] M. S. Leite, R. L. Woo, J. N. Munday, W. D. Hong, Sh. Mesropian, D. C.
Law, and H. A. Atwater. Towards an optimized all lattice-matched In-
AlAs/InGaAsP/InGaAs multijunction solar cell with efficiency >50%. Applied
Physics Letters, 102(3):033901, 2013.
[65] A. J. Ritenour, R. C. Cramer, S. Levinrad, and S. W. Boettcher. Efficient
n-GaAs photoelectrodes grown by close-spaced vapor transport from a solid
source. ACS Applied Materials & Interfaces, 4(1):69–73, 2011.
[66] A. J. Ritenour, J. W. Boucher, R. DeLancey, A. L. Greenaway, S. Aloni, and
S. W. Boettcher. Doping and electronic properties of GaAs grown by close-
spaced vapor transport from powder sources for scalable III-V photovoltaics.
Energy & Environmental Science, 8(1):278–285, 2015.
[67] K. Kamon, M. Shimazu, K. Kimura, M. Mihara, and M. Ishii. Orientation
dependence of GaAs growth in low-pressure OMVPE. Journal of Crystal
Growth, 84(1):126–132, 1987.
[68] S. D. Hersee, E. Barbier, R. Blondeau, and R. Blondeau. A study of the
orientation dependence of Ga(Al)As growth by MOVPE. Journal of Crystal
Growth, 77(13):310–320, 1986.
[69] Y. C. Kim, N. J. Jeon, J. H. Noh, W. S. Yang, J. Seo, J. S. Yun, A. Ho-Baillie,
S. Huang, M. A. Green, J. Seidel, T. K. Ahn, and S. I. Seok. Beneficial effects
of PbI2 incorporated in organo-lead halide perovskite solar cells. Advanced
Energy Materials, 6(4):1502104, 2016.
164
[70] P. A. Cox, M. S. Glaz, J. S. Harrison, S. R. Peurifoy, D. C. Coffey, and D. S.
Ginger. Imaging charge transfer state excitations in polymer/fullerene solar
cells with time-resolved electrostatic force microscopy. The Journal of Physical
Chemistry Letters, 6(15):2852–2858, 2015.
[71] M. Tuteja, A. B. Mei, V. Palekis, A. Hall, S. MacLaren, C. S. Ferekides,
and A. A. Rockett. CdCl2 treatment-induced enhanced conductivity in CdTe
solar cells observed using conductive atomic force microscopy. The Journal of
Physical Chemistry Letters, 7(24):4962–4967, 2016.
[72] A. Kento, I. Yutaka, A. Yoshio, and M. Takuya. Fine structures of organic pho-
tovoltaic thin films probed by frequency-shift electrostatic force microscopy.
Japanese Journal of Applied Physics, 55(7):070305, 2016.
[73] M. Koji, F. Takaya, I. Hiroshi, and F. Daisuke. Characterization of carrier
concentration in cigs solar cells by scanning capacitance microscopy. Measure-
ment Science and Technology, 25(4):044020, 2014.
[74] Lord Kelvin. On electric machines founded on induction and convection. Philo-
sophical Magazine, 46(82):66–73, 1898.
[75] Nonnenmacher M., M.P. OBoyle, and H. K. Ickramasinghe. Kelvin probe force
microscopy. Applied Physics Letters, 58:2921–2923, 1991.
[76] W. Melitz, J. Shen, A.C. Kummel, and S Lee. Kelvin probe force microscopy
and its application. Surface Science Reports, 66:1–27, 2011.
[77] S. Sadewasser, T. Glatzel, S. Schuler, S. Nishiwaki, R. Kaigawa, and M. Ch
Lux-Steiner. Kelvin probe force microscopy for the nano scale characterization
of chalcopyrite solar cell materials and devices. Thin Solid Films, 431:257–261,
2003.
[78] W. Melitz, J. Shen, S. Lee, J. S. Lee, A. C. Kummel, R. Droopad, and E. T.
Yu. Scanning tunneling spectroscopy and Kelvin probe force microscopy in-
vestigation of fermi energy level pinning mechanism on InAs and InGaAs clean
surfaces. Journal of Applied Physics, 108(2):023711, 2010.
[79] S. Yogev, R. Matsubara, M. Nakamura, U. Zschieschang, H. Klauk, and
Y. Rosenwaks. Fermi level pinning by gap states in organic semiconductors.
Physical Review Letters, 110(3):036803, 2013.
[80] Y. Zhang, D. Ziegler, and M. Salmeron. Charge trapping states at the SiO2-
Oligothiophene monolayer interface in field effect transistors studied by Kelvin
probe force microscopy. ACS Nano, 7(9):8258–8265, 2013.
[81] P. A. Fernández Garrillo, L . Borowik, F. Caffy, R. Demadrille, and B. Grévin.
Photo-carrier multi-dynamical imaging at the nanometer scale in organic and
inorganic solar cells. ACS Applied Materials and Interfaces, 8(45):31460–
31468, 2016.
165
[82] H. Lepage, N. Chevalier, D. Mariolle, and O. Renault. Measuring the life-
time of silicon nanocrystal solar cell photo-carriers by using Kelvin probe
force microscopy and x-ray photoelectron spectroscopy. Nanotechnology,
25(26):265703, 2014.
[83] M. Takihara, T. Takahashi, and T. Ujihara. Minority carrier lifetime in poly-
crystalline silicon solar cells studied by photoassisted Kelvin probe force mi-
croscopy. Applied Physics Letters, 93(2):021902, 2008.
[84] G. Shao, M. S. Glaz, F. Ma, H. Ju, and D. S. Ginger. Intensity-modulated
scanning Kelvin probe microscopy for probing recombination in organic pho-
tovoltaics. ACS Nano, 8(10):10799–10807, 2014.
[85] P. Girard. Electrostatic force microscopy: Principles and some applications
to semiconductors. Nanotechnology, 12(4):485–490, 2001.
[86] J. W. Hong, K. H. Noh, Sang-il Park, S. I. Kwun, and Z. G. Khim. Surface
charge density and evolution of domain structure in triglycine sulfate deter-
mined by electrostatic-force microscopy. Physical Review B, 58(8):5078–5084,
1998.
[87] V. K. Sangwan, D. Jariwala, I. S. Kim, K.-S. Chen, T. J. Marks, L. J. Lauhon,
and M. C. Hersam. Gate-tunable memristive phenomena mediated by grain
boundaries in single-layer MoS2. Nature Nanotechnology, 10(5):403–406, 2015.
[88] D. C. Coffey and D. S. Ginger. Time-resolved electrostatic force microscopy
of polymer solar cells. Nature Materials, 5:735–740, 2006.
[89] M. Jaquith, E. M. Muller, and J. A. Marohn. Time-resolved electric force
microscopy of charge trapping in polycrystalline pentacene. The Journal of
Physical Chemistry B, 111(27):7711–7714, 2007.
[90] S. Y. Leblebici, L. Leppert, Y. Li, S. E. Reyes-Lillo, S. Wickenburg, E. Wong,
J. Lee, M. Melli, D. Ziegler, D. K. Angell, D. F. Ogletree, P.D Ashby, F. M.
Toma, J. B. Neaton, I. D. Sharp, and A. Weber-Bargioni. Facet-dependent
photovoltaic efficiency variations in single grains of hybrid halide perovskite.
Nature Energy, 1:16093, 2016.
[91] J.-J. Li, J.-Y. Ma, Q.-Q. Ge, J.-S. Hu, D. Wang, and L.-J. Wan. Microscopic
investigation of grain boundaries in organolead halide perovskite solar cells.
ACS Applied Materials & Interfaces, 7(51):28518–28523, 2015.
[92] N. Bich Phuong, K. Gee Yeong, J. William, K. Byeong Jo, and
J. Hyun Suk. Trapping charges at grain boundaries and degradation of
CH3NH3Pb(I1−xBrx)3 perovskite solar cells. Nanotechnology, 28(31):315402,
2017.
166
[93] J. Luria, Y. Kutes, A. Moore, L. Zhang, E. A. Stach, and B. D. Huey. Charge
transport in CdTe solar cells revealed by conductive tomographic atomic force
microscopy. Nature Energy, 1(16150):16150, 2016.
[94] N. Muhunthan, O. P. Singh, V. Toutam, and V. N. Singh. Electrical char-
acterization of grain boundaries of CZTS thin films using conductive atomic
force microscopy techniques. Materials Research Bulletin, 70:373–378, 2015.
[95] T. A. Bull, L. S. C. Pingree, S. A. Jenekhe, D. S. Ginger, and C. K. Luscombe.
The role of mesoscopic PCBM crystallites in solvent vapor annealed copolymer
solar cells. ACS Nano, 3(3):627–636, 2009.
[96] X.-D. Dang, A. B. Tamayo, J. Seo, C. V. Hoven, B. Walker, and T.-Q. Nguyen.
Nanostructure and optoelectronic characterization of small molecule bulk het-
erojunction solar cells by photoconductive atomic force microscopy. Advanced
Functional Materials, 20(19):3314–3321, 2010.
[97] W. C. Tsoi, P. G. Nicholson, J. S. Kim, D. Roy, T. L. Burnett, C. E. Mur-
phy, J. Nelson, D. D. C. Bradley, J.-S. Kim, and F. A. Castro. Surface
and subsurface morphology of operating nanowire:fullerene solar cells revealed
by photoconductive-AFM. Energy & Environmental Science, 4(9):3646–3651,
2011.
[98] M. Guide, X.-D. Dang, and T.-Q. Nguyen. Nanoscale characterization of
tetrabenzoporphyrin and fullerene-based solar cells by photoconductive atomic
force microscopy. Advanced Materials, 23(20):2313–2319, 2011.
[99] L. Letertre, O. Douhret, M. Bougouma, T. Doneux, R. Lazzaroni, C. Buess-
Herman, and P. Leclre. Nanoscale study of MoSe2/Poly(3-Hexylthiophene)
bulk heterojunctions for hybrid photovoltaic applications. Solar Energy Ma-
terials and Solar Cells, 145(Part 2):116–125, 2016.
[100] B. E. McCandless, I. Youm, and R. W. Birkmire. Optimization of vapor
post-deposition processing for evaporated CdS/CdTe solar cells. Progress in
Photovoltaics: Research and Applications, 7(1):21–30, 1999.
[101] K. Kobayashi, H. Yamada, and K. Matsushige. Dopant profiling on semicon-
ducting sample by scanning capacitance force microscopy. Applied Physics
Letters, 81(14):2629–2631, 2002.
[102] I. Visoly-Fisher, S. R. Cohen, and D. Cahen. Direct evidence for grain-
boundary depletion in polycrystalline CdTe from nanoscale-resolved measure-
ments. Applied Physics Letters, 82(4):556–558, 2003.
[103] F. Takaya, M. Koji, I. Hiroshi, and F. Daisuke. Analytical procedure for ex-
perimental quantification of carrier concentration in semiconductor devices by
using electric scanning probe microscopy. Measurement Science and Technol-
ogy, 25(4):044021, 2014.
167
[104] N. Nicoara, T. Lepetit, L. Arzel, S. Harel, N. Barreau, and S. Sadewasser.
Effect of the KF post-deposition treatment on grain boundary properties in
Cu(In, Ga)Se2 thin films. Scientific Reports, 7:41361, 2017.
[105] A. Dazzi and C. B. Prater. AFM-IR: Technology and applications in nanoscale
infrared spectroscopy and chemical imaging. Chemical Reviews, 117(7):5146–
5173, 2017.
[106] S. J. Yoon, M. Kuno, and P. V. Kamat. Shift happens. how halide ion defects
influence photoinduced segregation in mixed halide perovskites. ACS Energy
Letters, 2:1507–1514, 2017.
[107] J. Chae, Q. Dong, J. Huang, and A. Centrone. Chloride incorporation process
in CH3NH3PbI3−xClx perovskites via nanoscale bandgap maps. Nano Letters,
15(12):8114–8121, 2015.
[108] D. Nowak, W. Morrison, H. K. Wickramasinghe, J. Jahng, E. Potma, L. Wan,
R. Ruiz, T. R. Albrecht, K. Schmidt, J. Frommer, D. P. Sanders, and S. Park.
Nanoscale chemical imaging by photoinduced force microscopy. Science Ad-
vances, 2(3):e1501571, 2016.
[109] X. Wang, D. Zhang, K. Braun, H.-J. Egelhaaf, C. J. Brabec, and A. J. Meixner.
High-resolution spectroscopic mapping of the chemical contrast from nanome-
ter domains in P3HT:PCBM organic blend films for solar-cell applications.
Advanced Functional Materials, 20(3):492–499, 2010.
[110] T. Schmid, L. Opilik, C. Blum, and R. Zenobi. Nanoscale chemical imag-
ing using tip-enhanced raman spectroscopy: A critical review. Angewandte
Chemie International Edition, 52(23):5940–5954, 2013.
[111] S. Zaleski, A. J. Wilson, M. Mattei, X. Chen, G. Goubert, M. F. Cardinal,
K. A. Willets, and R. P. Van Duyne. Investigating nanoscale electrochemistry
with surface- and tip-enhanced raman spectroscopy. Accounts of Chemical
Research, 49(9):2023–2030, 2016.
[112] N. Jiang, D. Kurouski, E. A. Pozzi, N. Chiang, M. C. Hersam, and R. P.
Van Duyne. Tip-enhanced raman spectroscopy: From concepts to practical
applications. Chemical Physics Letters, 659:16–24, 2016.
[113] G. Nicholson Patrick and A. Castro Fernando. Organic photovoltaics: Prin-
ciples and techniques for nanometre scale characterization. Nanotechnology,
21(49):492001, 2010.
[114] B.-S. Yeo, J. Stadler, T. Schmid, R. Zenobi, and W. Zhang. Tip-enhanced
raman spectroscopy - its status, challenges and future directions. Chemical
Physics Letters, 472(1):1–13, 2009.
168
[115] S. Ghosh, N. A. Kouamé, L. Ramos, S. Remita, A. Dazzi, A. Deniset-Besseau,
P. Beaunier, F. Goubard, P.-H. Aubert, and H. Remita. Conducting poly-
mer nanostructures for photocatalysis under visible light. Nature Materials,
14(5):505–511, 2015.
[116] I. Rajapaksa, K. Uenal, and H. K. Wickramasinghe. Image force microscopy
of molecular resonance: A microscope principle. Applied Physics Letters,
97(7):073121, 2010.
[117] I. Rajapaksa and H. K. Wickramasinghe. Raman spectroscopy and microscopy
based on mechanical force detection. Applied Physics Letters, 99(16):161103,
2011.
[118] J. Jahng, D. A. Fishman, S. Park, D. B. Nowak, W. A. Morrison, H. K.
Wickramasinghe, and E. O. Potma. Linear and nonlinear optical spectroscopy
at the nanoscale with photoinduced force microscopy. Accounts of Chemical
Research, 48(10):2671–2679, 2015.
[119] D. J. Slotcavage, H. I. Karunadasa, and M. D. McGehee. Light-induced phase
segregation in halide-perovskite absorbers. ACS Energy Letters, 1(6):1199–
1205, 2016.
[120] A. Dazzi, R. Prazeres, F. Glotin, and J. M. Ortega. Local infrared microspec-
troscopy with subwavelength spatial resolution with an atomic force micro-
scope tip used as a photothermal sensor. Optics Letters, 30(18):2388–2390,
2005.
[121] Cho. H., J. R. Felts, M.-F. Yu, L. A. Bergman, A. F. Vakakis, and W. P. King.
Improved atomic force microscope infrared spectroscopy for rapid nanometer-
scale chemical identification. Nanotechnology, 24(44):444007, 2013.
[122] A. Dazzi, F. Glotin, and R. Carminati. Theory of infrared nanospectroscopy by
photothermal induced resonance. Journal of Applied Physics, 107(12):124519,
2010.
[123] J. R. Felts, K. Kjoller, M. Lo, C. B. Prater, and W. P. King. Nanometer-scale
infrared spectroscopy of heterogeneous polymer nanostructures fabricated by
tip-based nanofabrication. ACS Nano, 6(9):8015–8021, 2012.
[124] B. Lahiri, G. Holland, and A. Centrone. Chemical imaging: Chemical imaging
beyond the diffraction limit: Experimental validation of the PTIR technique.
Small, 9(3):439–445, 2013.
[125] E. Strelcov, Q. Dong, T. Li, J. Chae, Y. Shao, Y. Deng, A. Gruverman,
J. Huang, and A. Centrone. CH3NH3PbI3 perovskites: Ferroelasticity re-
vealed. Science Advances, 3(4):e1602165, 2017.
169
[126] C. Gong, M. R. S. Dias, G. C. Wessler, J. A. Taillon, L. G. Salamanca-Riba,
and M. S. Leite. Fully alloyed noble metal nanoparticles via physical deposition
for plasmonics. Advanced Optical Materials, 5, 2017.
[127] B. Hecht, B. Sick, U. P. Wild, V. Deckert, R. Zenobi, O. J. F. Martin, and
D. W. Pohl. Scanning near-field optical microscopy with aperture probes: Fun-
damentals and applications. The Journal of Chemical Physics, 112(18):7761–
7774, 2000.
[128] Y. Zhang, Y. Wang, Z.-Q. Xu, J. Liu, J. Song, Y. Xue, Z. Wang, J. Zheng,
L. Jiang, C. Zheng, F. Huang, B. Sun, Y.-B. Cheng, and Q. Bao. Reversible
structural swell-shrink and recoverable optical properties in hybrid inorganic-
organic perovskite. ACS Nano, 10(7):7031–7038, 2016.
[129] W. Bao, M. Melli, N. Caselli, F. Riboli, D. S. Wiersma, M. Staffaroni, H. Choo,
D. F. Ogletree, S. Aloni, J. Bokor, S. Cabrini, F. Intonti, M. B. Salmeron,
E. Yablonovitch, P. J. Schuck, and A. Weber-Bargioni. Mapping local charge
recombination heterogeneity by multidimensional nanospectroscopic imaging.
Science, 338(6112):1317–1321, 2012.
[130] M. S. Leite, M. Abashin, H. J. Lezec, A. G. Gianfrancesco, A. A. Talin, and
N. B. Zhitenev. Nanoscale imaging of photocurrent and efficiency in CdTe
solar cells. ACS Nano, 8(11):11883–11890, 2014.
[131] M. S. Leite, Ma. Abashin, H. J. Lezec, A. G. Gianfrancesco, A. Alec Talin,
and N. B. Zhitenev. Mapping the local photoelectronic properties of poly-
crystalline solar cells through high resolution laser-beam-induced current mi-
croscopy. IEEE Journal of Photovoltaics, 4(1):311–316, 2014.
[132] U. W. Paetzold, S. Lehnen, K. Bittkau, U. Rau, and R. Carius. Nanoscale
observation of waveguide modes enhancing the efficiency of solar cells. Nano
Letters, 14(11):6599–6605, 2014.
[133] E. H. Synge. XXXVIII. a suggested method for extending microscopic reso-
lution into the ultra-microscopic region. The London, Edinburgh, and Dublin
Philosophical Magazine and Journal of Science, 6(35):356–362, 1928.
[134] U. Dürig, D. W. Pohl, and F. Rohner. Near-field optical-scanning microscopy.
Journal of Applied Physics, 59(10):3318–3327, 1986.
[135] A. Lewis, M. Isaacson, A. Harootunian, and A. Muray. Development of a 500
Ȧ spatial resolution light microscope. Ultramicroscopy, 13(3):227–231, 1984.
[136] J.-H. Kim and K.-B. Song. Recent progress of nano-technology with NSOM.
Micron, 38(4):409–426, 2007.
170
[137] D. F. Ogletree, P. J. Schuck, A. F. Weber-Bargioni, N. J. Borys, S. Aloni,
W. Bao, S. Barja, J. Lee, M. Melli, K. Munechika, S. Whitelam, and S. Wick-
enburg. Revealing optical properties of reduced-dimensionality materials at
relevant length scales. Advanced Materials, 27(38):5693–5719, 2015.
[138] M. Vrućnić, C. Matthiesen, A. Sadhanala, G. Divitini, S. Cacovich, S. E.
Dutton, C. Ducati, M. Atatre, H. Snaith, R. H. Friend, H. Sirringhaus, and
F. Deschler. Local versus long-range diffusion effects of photoexcited states on
radiative recombination in organic-inorganic lead halide perovskites. Advanced
Science, 2(9):1500136, 2015.
[139] K. Okamoto, A. Scherer, and Y. Kawakami. Near-field scanning optical mi-
croscopic transient lens for carrier dynamics study in InGaN/GaN. Applied
Physics Letters, 87(16):161104, 2005.
[140] K. Nakayama, K. Tanabe, and H. A Atwater. Plasmonic nanoparticle
enhanced light absorption in GaAs solar cells. Applied Physics Letters,
93(12):121904, 2008.
[141] J. Das Anshuman, R. Shivanna, and K. S. Narayan. Photoconductive NSOM
for mapping optoelectronic phases in nanostructures. Nanophotonics, 3(1-
2):19–31, 2014.
[142] J. M. Atkin, S. Berweger, A. C. Jones, and M. B. Raschke. Nano-optical
imaging and spectroscopy of order, phases, and domains in complex solids.
Advances in Physics, 61(6):745–842, 2012.
[143] A. Bek, R. Vogelgesang, and K. Kern. Apertureless scanning near field op-
tical microscope with sub-10nm resolution. Review of Scientific Instruments,
77(4):043703, 2006.
[144] Imec. https://www.imec-int.com/en/articles/imec-reports, 2017.
[145] G. Grancini, C. Roldn-Carmona, I. Zimmermann, E. Mosconi, X. Lee, D. Mar-
tineau, S. Narbey, F. Oswald, F. De Angelis, M. Graetzel, and Moham-
mad Khaja Nazeeruddin. One-year stable perovskite solar cells by 2d/3d
interface engineering. Nature Communications, 8:15684, 2017.
[146] M. Saliba, Taisuke M., J.-Y. Seo, K. Domanski, J.-P. Correa-Baena, M. K.
Nazeeruddin, S. M. Zakeeruddin, W. Tress, A. Abate, A. Hagfeldt, and
M. Grätzel. Cesium-containing triple cation perovskite solar cells: Improved
stability, reproducibility and high efficiency. Energy & Environmental Science,
9(6):1989–1997, 2016.
[147] R. E. Beal, D. J. Slotcavage, T. Leijtens, A. R. Bowring, R. A Belisle, W. H
Nguyen, G. F. Burkhard, E. T. Hoke, and M. D. McGehee. Cesium lead
halide perovskites with improved stability for tandem solar cells. The Journal
of Physical Chemistry Letters, 7(5):746–751, 2016.
171
[148] D. Bryant, N. Aristidou, S. Pont, I. Sanchez-Molina, T. Chotchunangatchaval,
S. Wheeler, J. R. Durrant, and S. A. Haque. Light and oxygen induced degra-
dation limits the operational stability of methylammonium lead triiodide per-
ovskite solar cells. Energy & Environmental Science, 9(5):1655–1660, 2016.
[149] D. Yang, J. Lv, X. Zhao, Q. Xu, Y. Fu, Y. Zhan, A. Zunger, and
L. Zhang. Functionality-directed screening of Pb-free hybrid organic-inorganic
perovskites with desired intrinsic photovoltaic functionalities. Chemistry of
Materials, 29(2):524–538, 2017.
[150] L. Li, N. Liu, Z. Xu, Q. Chen, X. Wang, and H. Zhou. Precise composition
tailoring of mixed-cation hybrid perovskites for efficient solar cells by mixture
design methods. ACS Nano, 11(9):88048813, 2017.
[151] V. W. Bergmann, S. A. L. Weber, F. Javier Ramos, M. K. Nazeeruddin,
M. Grätzel, D. Li, A. L. Domanski, I. Lieberwirth, S. Ahmad, and R. Berger.
Real-space observation of unbalanced charge distribution inside a perovskite-
sensitized solar cell. Nature Communications, 5:6001, 2014.
[152] M. I. Asghar, J. Zhang, H. Wang, and P. D. Lund. Device stability of per-
ovskite solar cells - a review. Renewable and Sustainable Energy Reviews,
77:131–146, 2017.
[153] N.-G. Park, M. Grätzel, T. Miyasaka, K. Zhu, and K. Emery. Towards stable
and commercially available perovskite solar cells. Nature Energy, 1:16152,
2016.
[154] L. Wang, H. Wang, M. Wagner, Y. Yan, D. S. Jakob, and X. G. Xu. Nanoscale
simultaneous chemical and mechanical imaging via peak force infrared mi-
croscopy. Science Advances, 3(6):1–11, 2017.
[155] J. L. Garrett and J. N. Munday. Fast, high-resolution surface potential mea-
surements in air with heterodyne Kelvin probe rorce microscopy. Nanotech-
nology, 27(24):245705, 2016.
[156] L. Collins, M. Ahmadi, T. Wu, B. Hu, S. V. Kalinin, and S. Jesse. Breaking the
time barrier in Kelvin probe force microscopy: Fast free force reconstruction
using the G-mode platform. ACS Nano, 11(9):87178729, 2017.
[157] Y. Miyahara and Peter G. Force-gradient sensitive Kelvin probe force mi-
croscopy by dissipative electrostatic force modulation. Applied Physics Letters,
110(16):163103, 2017.
[158] J. B. Kim and W. S. Chang. Design and fabrication of a tip-on-aperture probe
for resolution enhancement of optical patterning. Microelectronic Engineering,
151:24–29, 2016.
172
[159] F. Huth, A. Chuvilin, M. Schnell, I. Amenabar, R. Krutokhvostov, S. Lopatin,
and R. Hillenbrand. Resonant antenna probes for tip-enhanced infrared near-
field microscopy. Nano Letters, 13(3):1065–1072, 2013.
[160] A. L. Greenaway, J. W. Boucher, S. Z. Oener, C. J. Funch, and S. W.
Boettcher. Low-cost approaches to III-V semiconductor growth for photo-
voltaic applications. ACS Energy Letters, pages 2270–2282, 2017.
[161] H. R. Moutinho, R. G. Dhere, C. S. Jiang, Y. F. Yan, D. S. Albin, and M. M.
Al-Jassim. Investigation of potential and electric field profiles in cross sections
of CdTe/CdS solar cells using scanning Kelvin probe microscopy. Journal of
Applied Physics, 108(7):074503, 2010.
[162] M. A. Green. Commercial progress and challenges for photovoltaics. Nature
Energy, 1:15015, 2016.
[163] A. Chirilã, P. Reinhard, F. Pianezzi, P. Bloesch, A. R. Uhl, C. Fella, L. Kranz,
D. Keller, C. Gretener, H. Hagendorfer, D. Jaeger, R. Erni, S. Nishiwaki,
S. Buecheler, and A. N. Tiwari. Potassium-induced surface modification
of Cu(In,Ga)Se2 thin films for high-efficiency solar cells. Nature Materials,
12(12):1107–1111, 2013.
[164] P. Jackson, R. Wuerz, D. Hariskos, E. Lotter, W.m Witte, and M. Powalla.
Effects of heavy alkali elements in Cu(In,Ga)Se2 solar cells with efficiencies up
to 22.6%. physica status solidi (RRL) Rapid Research Letters, 10(8):583–586,
2016.
[165] C.-H. Liu, C.-H. Chen, S.-Y. Chen, Y.-T. Yen, W.-C. Kuo, Y.-K. Liao, J.-
Y. Juang, H.-C. Kuo, C.-H. Lai, L.-J. Chen, and Y.-L. Chueh. Large scale
single-crystal Cu(In,Ga)Se2 nanotip arrays for high efficiency solar cell. Nano
Letters, 11(10):4443–4448, 2011.
[166] T. B. Harvey, I. Mori, C. J. Stolle, T. D. Bogart, D. P. Ostrowski, M. S.
Glaz, J. Du, D. R. Pernik, V. A. Akhavan, H. Kesrouani, D. A. Vanden Bout,
and B. A. Korgel. Copper indium gallium selenide photovoltaic devices made
using multistep selenization of nanocrystal films. ACS Applied Materials and
Interfaces, 5(18):9134–9140, 2013.
[167] J. A. Frantz, R. Y. Bekele, V. Q. Nguyena, J. S. Sanghera, A. Bruce, S. V.
Frolov, M. Cyrus, and I. D. Aggarwal. Cu(In,Ga)Se2 thin films and devices
sputtered from a single target without additional selenization. Thin Solid
Films, (519):7763–7765, 2011.
[168] J. H. Shi, Z. Q. Li, D. W. Zhang, Q. Q. Liu, Z. Sun, and S. M. Huang.
Fabrication of Cu(In,Ga)Se2 thin films by sputtering from a single quaternary
chalcogenide target. Progress in Photovoltaics: Research and Applications,
19(2):160–164, 2011.
173
[169] L. Ouyang, D. Zhuang, M. Zhao, N. Zhang, X. Li, L. Guo, R. Sun, and
M. Cao. Cu(In,Ga)Se2 solar cell with 16.7% active-area efficiency achieved
by sputtering from a quaternary target. physica status solidi (a), pages 1–5,
2015.
[170] N. G. Dhere. Scale-up issues of CIGS thin film PV modules. Solar Energy
Materials and Solar Cells, 95(1):277–280, 2011.
[171] N. E. Gorji. Degradation sources of CdTe thin film PV: CdCl2 residue and
shunting pinholes. Applied Physics A, 116(3):1347–1352, 2014.
[172] K. Durose, S. E. Asher, W. Jaegermann, D. Levi, B. E. McCandless, W. Met-
zger, H. Moutinho, P. D. Paulson, C. L. Perkins, J. R. Sites, G. Teeter, and
M. Terheggen. Physical characterization of thin-film solar cells. Progress in
Photovoltaics, 12(2-3):177–217, 2004.
[173] T. Kirchartz and U. Rau. Advanced Characterization Techniques for Thin
Film Solar Cells, pages 1–547. Wiley-VCH, 2011.
[174] D. P. Ostrowski, M. S. Glaz, B. W. Goodfellow, V. A. Akhavan, M. G. Pan-
thani, B. A. Korgel, and D. A. V. Bout. Mapping spatial heterogeneity in
Cu(In1−xGax)Se2 nanocrystal-based photovoltaics with scanning photocurrent
and fluorescence microscopy. Small, 6(24):2832–2836, 2010.
[175] S. L. Howell, S. Padalkar, K.-H. Yoon, Q. Li, D. D. Koleske, J. J. Wierer,
G. T. Wang, and L. J. Lauhon. Spatial mapping of efficiency of GaN/InGaN
nanowire array solar cells using scanning photocurrent microscopy. Nano Let-
ters, 13:51235128, 2013.
[176] D. Josell, R. Debnath, J. Y. Ha, J. Guyer, M. A. Sahiner, C. J. Reehil, W. A.
Manners, and N. V. Nguyen. Windowless CdSe/CdTe solar cells with differ-
entiated back contacts: JV, EQE, and photocurrent mapping. ACS Applied
Materials and Interfaces, 6(18):15972–15979, 2014.
[177] M. Nichterwitz, D. Abou-Ras, K. Sakurai, J. Bundesmann, T. Unold,
R. Scheer, and H. W. Schock. Influence of grain boundaries on current collec-
tion in Cu(In,Ga)Se2 thin-film solar cells. Thin Solid Films, 517(7):2554–2557,
2009.
[178] P. O. Grabitz, U. Rau, B. Wille, G. Bilger, and J. H. Werner. Spatial in-
homogeneities in Cu(In,Ga)Se2 solar cells analyzed by electron beam induced
voltage measurements. In Photovoltaic Energy Conversion, Conference Record
of the 2006 IEEE 4th World Conference on, volume 1, pages 424–427, 2006.
[179] H. P. Yoon, P. M. Haney, D. Ruzmetov, H. Xu, M. S. Leite, B. H. Hamadani,
A. A. Talin, and N. B. Zhitenev. Local electrical characterization of cadmium
telluride solar cells using low-energy electron beam. Solar Energy Materials
and Solar Cells, 117:499–504, 2013.
174
[180] J.-H. Kim, S.-Y. Kim, C.-S. Jiang, K. Ramanathan, and M. M. Al-
Jassim. Direct imaging of enhanced current collection on grain boundaries
of Cu(In,Ga)Se2 solar cells. Applied Physics Letters, 104(6):063902, 2014.
[181] D. Azulay, O. Millo, I. Balberg, H.-W. Schock, I. Visoly-Fisher, and D. Ca-
hen. Current routes in polycrystalline CuInSe2 and Cu(In,Ga)Se2 films. Solar
Energy Materials and Solar Cells, 91(1):85–90, 2007.
[182] D.C. Coffey, O.G. Reid, D.B. Rodovsky, G.P Bartholomew, and D. S. Ginger.
Mapping local photocurrents in polymer/fullerene solar cells with photocon-
ductive atomic force microscopy. Nano Letters, 7:738–744, 2007.
[183] Y. Kutes, B. A. Aguirre, J. L. Bosse, J. L. Cruz-Campa, D. Zubia, and B. D.
Huey. Mapping photovoltaic performance with nanoscale resolution. Progress
in Photovoltaics: Research and Applications, 24(3):315–325, 2015.
[184] T. Dittrich, A. Gonzales, T. Rada, T. Rissom, E. Zillner, S. Sade-
wasser, and M. Lux-Steiner. Comparative study of Cu(In,Ga)Se2/CdS and
Cu(In,Ga)Se2/In2S3 systems by surface photovoltage techniques. Thin Solid
Films, 535:357–361, 2013.
[185] M. Tuteja, P. Koirala, S. MacLaren, R. Collins, and A. Rockett. Direct obser-
vation of electrical properties of grain boundaries in sputter-deposited CdTe
using scan-probe microwave reflectivity based capacitance measurements. Ap-
plied Physics Letters, 107(14):142106, 2015.
[186] C. R. McNeill, H. Frohne, J. L. Holdsworth, J. E. Furst, B. V. King, and P. C.
Dastoor. Direct photocurrent mapping of organic solar cells using a near-field
scanning optical microscope. Nano Letters, 4(2):219–223, 2004.
[187] L. Gütay, R. Pomraenke, C. Lienau, and G. H. Bauer. Subwavelength inho-
mogeneities in Cu(In,Ga)Se2 thin films revealed by near-field scanning optical
microscopy. physica status solidi (a), 206(5):1005–1008, 2009.
[188] U. Rau, K. Taretto, and S. Siebentritt. Grain boundaries in Cu(In,Ga)(Se,S)2
thin-film solar cells. Applied Physics A, 96(1):221–234, 2009.
[189] D. Abou-Ras, C. T. Koch, V. Kstner, P. A. van Aken, U. Jahn, M. A. Contr-
eras, R. Caballero, C. A. Kaufmann, R. Scheer, T. Unold, and H. W. Schock.
Grain-boundary types in chalcopyrite-type thin films and their correlations
with film texture and electrical properties. Thin Solid Films, 517(7):2545–
2549, 2009.
[190] K. Bothe, G. H. Bauer, and T. Unold. Spatially resolved photoluminescence
measurements on Cu(In,Ga)Se2 thin films. Thin Solid Films, 403404(0):453–
456, 2002.
175
[191] L. Gütay and G. H. Bauer. Spectrally resolved photoluminescence studies on
Cu(In,Ga)Se2 solar cells with lateral submicron resolution. Thin Solid Films,
515(15):6212–6216, 2007.
[192] L. Lombez, M. Soro, A. Delamarre, N. Naghavi, N. Barreau, D. Lincot, and J.-
F. Guillemoles. Revisiting the interpretation of biased luminescence: Effects
on Cu(In,Ga)Se2 photovoltaic heterostructures. Journal of Applied Physics,
116(064504):6, 2014.
[193] G. H. Bauer, R. Brggemann, S. Tardon, S. Vignoli, and R. Kniese. Quasi-fermi
level splitting and identification of recombination losses from room tempera-
ture luminescence in Cu(In1−xGax)Se2 thin films versus optical band gap. Thin
Solid Films, 480481:410–414, 2005.
[194] K. Alberi, B. Fluegel, H. Moutinho, R. G. Dhere, J. V. Li, and A. Mascarenhas.
Measuring long-range carrier diffusion across multiple grains in polycrystalline
semiconductors by photoluminescence imaging. Nature Communications, 4:1–
7, 2013.
[195] O. Neumann, R. Brüggemann, D. Hariskos, W. Witte, and G. H. Bauer.
Photoluminescence studies of polycrystalline Cu(In,Ga)Se2: Lateral inho-
mogeneities beyond abbe’s diffraction limit. Journal of Applied Physics,
118(18):185311, 2015.
[196] C. Li, Y. Wu, J. Poplawsky, T. J. Pennycook, N. Paudel, W. Yin, S. J. Haigh,
M. P. Oxley, A. R. Lupini, M. Al-Jassim, S. J. Pennycook, and Y. Yan. Grain-
boundary-enhanced carrier collection in CdTe solar cells. Physical Review
Letters, 112(15):156103, 2014.
[197] R. Chakraborty, J. Serdy, B. West, M. Stuckelberger, B. Lai, J. Maser, M. I.
Bertoni, M. L. Culpepper, and T. Buonassisi. Development of an in-situ tem-
perature stage for synchrotron x-ray spectromicroscopy. Review of Scientific
Instruments, 86(11):1–6, 2015.
[198] R. Mainz, E. Simsek Sanli, H. Stange, D. Azulay, S. Brunken, D. Greiner,
S. Hajaj, M. D. Heinemann, C. A. Kaufmann, M. Klaus, Q. M. Ramasse,
H. Rodriguez-Alvarez, A. Weber, I. Balberg, O. Millo, P. A. van Aken, and
D. Abou-Ras. Annihilation of structural defects in chalcogenide absorber films
for high-efficiency solar cells. Energy and Environmental Science, pages 1–10,
2016.
[199] C. Xue, D. Papadimitriou, Y. S. Raptis, W. Richter, N. Esser, S. Sieben-
tritt, and M. Ch. Lux-Steiner. Micro-raman study of orientation effects of
CuxSe-crystallites on cu-rich CuGaSe2 thin films. Journal of Applied Physics,
96(4):1963–1966, 2004.
[200] A. Luque and S. Hegedus. Handbook of Photovoltaic Science and Engineering.
John Wiley and Sons, Ltd, West Sussex, 2011.
176
[201] A. Fantoni, M. Vieira, J. Cruz, R. Schwarz, and R. Martins. A two-dimensional
numerical simulation of a non-uniformly illuminated amorphous silicon solar
cell. Journal of Physics D: Applied Physics, 29(12):3154, 1996.
[202] A. Cuevas and S. Lpez-Romero. The combined effect of non-uniform illumi-
nation and series resistance on the open-circuit voltage of solar cells. Solar
Cells, 11(2):163–173, 1984.
[203] S. Nomura, S. Ouchi, and S. Endo. Raman spectra of ordered vacancy
compounds in the Cu-In-Se system. Japanese Journal of Applied Physics,
36(8B):1075–1077, 1997.
[204] W. Witte, R. Kniese, and M. Powalla. Raman investigations of Cu(In,Ga)Se2
thin films with various copper contents. Thin Solid Films, 517(2):867–869,
2008.
[205] M. Gloeckler, J. R. Sites, and W. K. Metzger. Grain-boundary recombination
in Cu(In,Ga)Se2 solar cells. Journal of Applied Physics, 98(11):113704, 2005.
[206] National Renewable Energy Laboratories. Standard am1.5 global sun spec-
trum (2012), 2012.
[207] A. Slobodskyy, T. Slobodskyy, T. Ulyanenkova, S. Doyle, M. Powalla,
T. Baumbach, and U. Lemmer. In-depth analysis of the CuIn1−xGaxSe2 film
for solar cells, structural and optical characterization. Applied Physics Letters,
97(251911), 2010.
[208] M. A. Green, K. Emery, Y. Hishikawa, W. Warta, and E. D. Dunlop. Solar
cell efficiency tables (version 46). Progress in Photovoltaics: Research and
Applications, 23(7):805–812, 2015.
[209] K.R. Catchpole and A. Polman. Plasmonic solar cells. Optics Express,
16(26):21793–21800, 2008.
[210] P. Campbell and M. A. Green. Light trapping properties of pyramidally tex-
tured surfaces. Journal of Applied Physics, 62(1):243–249, 1987.
[211] J. N. Munday and H. A. Atwater. Large integrated absorption enhancement
in plasmonic solar cells by combining metallic gratings and antireflection coat-
ings. Nano Letters, 11(6):2195–2201, 2010.
[212] M. L. Brongersma, Y. Cui, and S. Fan. Light management for photovoltaics
using high-index nanostructures. Nater Materials, 13(5):451–460, 2014.
[213] H. R. Moutinho, R. G. Dhere, C. S. Jiang, M. M. Al-Jassim, and L. L. Kazmer-
ski. Electrical properties of CdTe/CdS solar cells investigated with conductive
atomic force microscopy. Thin Solid Films, 514(1-2):150–155, 2006.
177
[214] I. Visoly-Fisher, S. R. Cohen, K. Gartsman, A. Ruzin, and D. Cahen. Under-
standing the beneficial role of grain boundaries in polycrystalline solar cells
from single-grain-boundary scanning probe microscopy. Advanced Functional
Materials, 16(5):649–660, 2006.
[215] H. Sakaguchi, F. Iwata, A. Hirai, A. Sasaki, and T. Nagamura. Nanometer-
scale photoelectric property of organic thin films investigated by a photo-
conductive atomic force microscope. Japanese Journal of Applied Physics,
38(65):3908–3911, 1999.
[216] C. Ballif, H. R. Moutinho, and M. M. Al-Jassim. Cross-sectional electro-
static force microscopy of thin-film solar cells. Journal of Applied Physics,
89(2):1418–1424, 2001. Ballif, C Moutinho, HR Al-Jassim, MM.
[217] M. J. Romero, C. S. Jiang, J. Abushama, H. R. Moutinho, M. M. Al-Jassim,
and R. Noufi. Electroluminescence mapping of CuGaSe2 solar cells by atomic
force microscopy. Applied Physics Letters, 89:143120, 2006.
[218] H. R. Moutinho, R. G. Dhere, C. S. Jiang, T. Gessert, A. Duda, M. Young,
W. K. Metzger, and M. M. Al-Jassim. Role of Cu on the electrical properties
of CdTe/CdS solar cells: A cross-sectional conductive atomic force microscopy
study. Journal of Vacuum Science & Technology B, 25(2):361–367, 2007.
[219] W. Melitz, J. Shen, A.C. Kummel, and S Lee. Kelvin probe force microscopy
and its application. Surface Science Reports, 66:1–27, 2011.
[220] P. Sutter, E. Sutter, and T. R. Ohno. High-resolution mapping of nonuniform
carrier transport at contacts to polycrystalline CdTe/CdS solar cells. Applied
Physics Letters, 84(12):2100–2102, 2004.
[221] P. Eyben, F. Seidel, T. Hantschel, A. Schulze, A. Lorenz, A. U. De Castro,
D. Van Gestel, J. John, J. Horzel, and W. Vandervorst. Development and
optimization of scanning spreading resistance microscopy for measuring the
two-dimensional carrier profile in solar cell structures. physica status solidi
(a), 208(3):596–599, 2011.
[222] O. G. Reid, G. E. Rayermann, D. C. Coffey, and D. S. Ginger. Imaging
local trap formation in conjugated polymer solar cells: A comparison of time-
resolved electrostatic force microscopy and scanning Kelvin probe imaging.
The Journal of Physical Chemistry C, 114(48):20672–20677, 2010.
[223] E. J. Spadafora, R. Demadrille, B. Ratier, and B. Grévin. Imaging the car-
rier photogeneration in nanoscale phase segregated organic heterojunctions by
Kelvin probe force microscopy. Nano Letters, 10(9):3337–3342, 2010.
[224] H. Hoppe, T. Glatzel, M. Niggemann, A. Hinsch, M. Ch Lux-Steiner, and
N. S. Sariciftci. Kelvin probe force microscopy study on conjugated poly-
mer/fullerene bulk heterojunction organic solar cells. Nano Letters, 5(2):269–
274, 2005.
178
[225] S. Watanabe, Y. Fukuchi, M. Fukasawa, T. Sassa, A. Kimoto, Y. Tajima,
M. Uchiyama, T. Yamashita, M. Matsumoto, and T. Aoyama. In situ KPFM
imaging of local photovoltaic characteristics of structured organic photovoltaic
devices. ACS Applied Materials and Interfaces, 6(3):1481–1487, 2014.
[226] V. Palermo, G. Ridolfi, A. M Talarico, L. Favaretto, G. Barbarella, N. Ca-
maioni, and P. Samor. A Kelvin probe force microscopy study of the photo-
generation of surface charges in all-thiophene photovoltaic blends. Advanced
Functional Materials, 17(3):472–478, 2007.
[227] A. Soudi, C.-H. Hsu, and Y. Gu. Diameter-dependent surface photovoltage
and surface state density in single semiconductor nanowires. Nano Letters,
12(10):5111–5116, 2012.
[228] C.-S. Jiang, D. J. Friedman, J. F. Geisz, H. R. Moutinho, M. J. Romero, and
M. M. Al-Jassim. Distribution of built-in electrical potential in GaInP2/GaAs
tandem-junction solar cells. Applied Physics Letters, 83(8):1572–1574, 2003.
[229] C.-S. Jiang, H. R. Moutinho, D. J. Friedman, J. F. Geisz, and M. M. Al-
Jassim. Measurement of built-in electrical potential in IIIV solar cells by
scanning Kelvin probe microscopy. Journal of Applied Physics, 93(12):10035–
10040, 2003.
[230] A. Schwarzman, E. Grunbaum, E. Strassburg, E. Lepkifker, A. Boag,
Y. Rosenwaks, T. Glatzel, Z. Barkay, M. Mazzer, and K. Barnham. Nanoscale
potential distribution across multiquantum well structures: Kelvin probe force
microscopy and secondary electron imaging. Journal of Applied Physics,
98(8):–, 2005.
[231] R. Baier, C. Leendertz, M. C. Lux-Steiner, and S. Sadewasser. Toward quan-
titative Kelvin probe force microscopy of nanoscale potential distributions.
Physical Review B, 85(16), 2012.
[232] H. R. Moutinho, R. G. Dhere, C. S. Jiang, Y. F. Yan, D. S. Albin, and M. M.
Al-Jassim. Investigation of potential and electric field profiles in cross sections
of CdTe/CdS solar cells using scanning Kelvin probe microscopy. Journal of
Applied Physics, 108(7):074503, 2010.
[233] M. Hafemeister, S. Siebentritt, J. Albert, M. C. Lux-Steiner, and S. Sade-
wasser. Large neutral barrier at grain boundaries in chalcopyrite thin films.
Physical Review Letters, 104(19), 2010.
[234] Masaki Takihara, Takashi Minemoto, Youichi Wakisaka, and Takuji Taka-
hashi. An investigation of band profile around the grain boundary of
Cu(In,Ga)Se2 solar cell material by scanning probe microscopy. Progress in
Photovoltaics: Research and Applications, 21(4):595–599, 2013.
179
[235] S. S. Schmidt, D. Abou-Ras, S. Sadewasser, W. Yin, C. Feng, and Y. Yan.
Electrostatic potentials at Cu(In,Ga)Se2 grain boundaries: Experiment and
simulations. Physical Review Letters, 109(9):095506, 2012. PRL.
[236] H. Mönig, Y. Smith, R. Caballero, C. A. Kaufmann, I. Lauermann, M. Ch
Lux-Steiner, and S. Sadewasser. Direct evidence for a reduced density of
deep level defects at grain boundaries of CuGaSe2 thin films. Physical Review
Letters, 105(11):116802, 2010. PRL.
[237] M. K. Herndon, A. Gupta, V. Kaydanov, and R. T. Collins. Evidence for
grain-boundary-assisted diffusion of sulfur in polycrystalline CdS/CdTe het-
erojunctions. Applied Physics Letters, 75(22):3503–3505, 1999.
[238] A. A. McDaniel, J. W. P. Hsu, and A. M. Gabor. Near-field scanning opti-
cal microscopy studies of Cu(In,Ga)Se2 solar cells. Applied Physics Letters,
70(26):3555–3557, 1997. McDaniel, AA Hsu, JWP Gabor, AM.
[239] M. J. Romero, K. Alberi, I. T. Martin, K. M. Jones, D. L. Young, Y. Yan,
C. Teplin, M. M. Al-Jassim, P. Stradins, and H. M. Branz. Nanoscale mea-
surements of local junction breakdown in epitaxial film silicon solar cells. Appl.
Phys. Lett., 97(9):092107, 2010.
[240] K. Bittkau and T. Beckers. Near-field study of light scattering at rough
interfaces of a-si:h/mu c-si:h tandem solar cells. Physica Status Solidi a-
Applications and Materials Science, 207(3):661–666, 2010.
[241] W. Bao, M. Staffaroni, J. Bokor, M. B. Salmeron, E. Yablonovitch, S. Cabrini,
A. Weber-Bargioni, and P.J. Schuck. Plasmonic near-field probes: a com-
parison of the campanile geometry with other sharp tips. Optics Express,
21(7):8166–8176, 2013.
[242] P. J. Schuck, A. Weber-Bargioni, P. D. Ashby, D. F. Ogletree,
A. Schwartzberg, and S. Cabrini. Life beyond diffraction: Opening new
routes to materials characterization with next-generation optical near-field
approaches. Advanced Functional Materials, 23(20):2539–2553, 2013.
[243] R. D. Berger, Weber A.L, and Stefan A. L. Electrical characterization of
organic solar cell materials based on scanning force microscopy. European
Polymer Journal, 49(8):1907–1915, 2013.
[244] A. Castaldini, D. Cavalcoli, A. Cavallini, and M. Rossi. Scanning Kelvin probe
and surface photovoltage analysis of multicrystalline silicon. Materials Science
and Engineering: B, 91-92(0):234–238, 2002.
[245] Q. Chen, H. Zhou, T.-B. Song, S. Luo, Z. Hong, H.-S. Duan, L. Dou, Y. Liu,
and Y. Yang. Controllable self-induced passivation of hybrid lead iodide per-
ovskites toward high performance solar cells. Nano Letters, 14(7):4158–4163,
2014.
180
[246] M. Chiesa, L. Bürgi, J.-S. Kim, R. Shikler, R. H. Friend, and H. Sirring-
haus. Correlation between surface photovoltage and blend morphology in
polyfluorene-based photodiodes. Nano Letters, 5(4):559–563, 2005.
[247] J. Heo and S. Won. Scanning probe study on the photovoltaic characteristics
of a si solar cell by using Kelvin force microscopy and photoconductive atomic
force microscopy. Thin Solid Films, 546(0):353–357, 2013.
[248] C. S. Jiang, H. R. Moutinho, R. G. Dhere, and M. M. Al-Jassim. The
nanometer-resolution local electrical potential and resistance mapping of CdTe
thin films. IEEE Journal of Photovoltaics, 3(4):1383–1388, 2013.
[249] C.-S. Jiang, R. Noufi, K. Ramanathan, H. R. Moutinho, and M. M. Al-Jassim.
Electrical modification in Cu(In,Ga)Se2 thin films by chemical bath deposition
process of cds films. Journal of Applied Physics, 97(5):053701, 2005.
[250] M. Salvador, S. M. Vorpahl, H. Xin, W. Williamson, G. Shao, D.s U. Karatay,
H. W. Hillhouse, and D. S. Ginger. Nanoscale surface potential variation
correlates with local S/Se ratio in solution-processed CZTSSe solar cells. Nano
Letters, 14(12):6926–6930, 2014.
[251] C. Yang, Y. Pyekh, and S. Danyluk. Surface potential imaging of PV cells
with a Kelvin probe. Solar Energy Materials and Solar Cells, 102(0):167–172,
2012.
[252] C. Loppacher, U. Zerweck, S. Teich, E. Beyreuther, T. Otto, S. Grafström,
and L. M. Eng. FM demodulated Kelvin probe force microscopy for surface
photovoltage tracking. Nanotechnology, 16(3):S1, 2005.
[253] W. Bergbauer, T. Lutz, W. Frammelsberger, and G. Benstetter. Kelvin probe
force microscopy - an appropriate tool for the electrical characterisation of
LED heterostructures. Microelectronics Reliability, 46(911):1736–1740, 2006.
[254] M. S. Carpenter, M. R. Melloch, M. S. Lundstrom, and S. P. Tobin. Effects
of Na2S and (NH4)2S edge passivation treatments on the dark current-voltage
characteristics of GaAs p-n diodes. Applied Physics Letters, 52(25):2157–2159,
1988.
[255] S. Barbet, M. Popoff, H. Diesinger, D. Deresmes, D. Theron, and T. Melin.
Cross-talk artefacts in Kelvin probe force microscopy imaging: A comprehen-
sive study. Journal of Applied Physics, 115(144313):1–11, 2014.
[256] H. O. Jacobs, P. Leuchtmann, O. J. Homan, and A. Stemmer. Resolution
and contrast in Kelvin probe force microscopy. Journal of Applied Physics,
84(3):1168–1173, 1998.
[257] U. Zerweck, C. Loppacher, T. Otto, S. Grafström, and L. M. Eng. Accuracy
and resolution limits of Kelvin probe force microscopy. Physical Review B,
71(12):125424, 2005.
181
[258] R. Garca and A. San Paulo. Attractive and repulsive tip-sample interac-
tion regimes in tapping-mode atomic force microscopy. Physical Review B,
60(7):4961–4967, 1999.
[259] S. Sadewasser and T. Glatzel. Kelvin Probe Force Microscopy: Measuring and
Compensating Electrostatic Forces. Springer, 2011.
[260] S. Hormeño, M. Penedo, C. V. Manzano, and M. Luna. Gold nanoparticle
coated silicon tips for Kelvin probe force microscopy in air. Nanotechnology,
24(39):395701, 2013.
[261] M. Cohen, R. Shavit, and Z. Zalevsky. Observing optical plasmons on a single
nanometer scale. Scientific Reports, 4, 2014.
[262] M. A. Green, A. Ho-Baillie, and H. J. Snaith. The emergence of perovskite
solar cells. Nature Photonics, 8(7):506–514, 2014.
[263] T. M. Brenner, D. A. Egger, L. Kronik, G. Hodes, and D. Cahen. Hybrid
organic-inorganic perovskites: low-cost semiconductors with intriguing charge-
transport properties. Nature Reviews Materials, 1:15007, 2016.
[264] P. Gao, M. Grätzel, and M. K. Nazeeruddin. Organohalide lead perovskites for
photovoltaic applications. Energy & Environmental Science, 7(8):2448–2463,
2014.
[265] M. Sessolo and H. J. Bolink. Perovskite solar cells join the major league.
Science, 350(6263):917–917, 2015.
[266] J. P. Mailoa, C. D. Bailie, E. C. Johlin, E. T. Hoke, A. J. Akey, W. H.
Nguyen, M. D. McGehee, and T. Buonassisi. A 2-terminal perovskite/silicon
multijunction solar cell enabled by a silicon tunnel junction. Applied Physics
Letters, 106(12):121105, 2015.
[267] C. D. Bailie, M. G. Christoforo, J. P. Mailoa, A. R. Bowring, E. L. Unger,
W. H. Nguyen, J. Burschka, N. Pellet, J. Z. Lee, M. Grätzel, R. Noufi,
T. Buonassisi, A. Salleo, and M. D. McGehee. Semi-transparent perovskite so-
lar cells for tandems with silicon and CIGS. Energy & Environmental Science,
8(3):956–963, 2015.
[268] D. Shi, V. Adinolfi, R. Comin, M. Yuan, E. Alarousu, A. Buin, Y. Chen,
S. Hoogland, A. Rothenberger, K. Katsiev, Y. Losovyj, X. Zhang, P. A. Dow-
ben, O. F. Mohammed, E. H. Sargent, and O. M. Bakr. Low trap-state density
and long carrier diffusion in organolead trihalide perovskite single crystals.
Science, 347(6221):519–522, 2015.
[269] Q. Dong, Y. Fang, Y. Shao, P. Mulligan, J. Qiu, L. Cao, and J. Huang.
Electron-hole diffusion lengths > 175 µm in solution-grown CH3NH3PbI3 sin-
gle crystals. Science, 347(6225):967–970, 2015.
182
[270] G. Niu, X. Guo, and L. Wang. Review of recent progress in chemical stability
of perovskite solar cells. Journal of Materials Chemistry A, 3(17):8970–8980,
2015.
[271] J. You, L. Meng, T.-B. Song, T.-F. Guo, Y. M. Yang, W.-H. Chang, Z. Hong,
H. Chen, H. Zhou, and Q. Chen. Improved air stability of perovskite solar cells
via solution-processed metal oxide transport layers. Nature nanotechnology,
11(1):75–81, 2016.
[272] H. J. Snaith, A. Abate, J. M. Ball, G. E. Eperon, T. Leijtens, N. K. Noel,
S. D. Stranks, J. T.-W. Wang, K. Wojciechowski, and W. Zhang. Anomalous
hysteresis in perovskite solar cells. The journal of physical chemistry letters,
5(9):1511–1515, 2014.
[273] W. Qiu, U. W. Paetzold, R. Gehlhaar, V. Smirnov, H.-G. Boyen, J. G. Tait,
B. Conings, W. Zhang, C. B. Nielsen, I. McCulloch, L. Froyen, P. Heremans,
and D. Cheyns. An electron beam evaporated TiO2 layer for high efficiency
planar perovskite solar cells on flexible polyethylene terephthalate substrates.
Journal of Materials Chemistry A, 2015.
[274] D. W. de Quilettes, S. M. Vorpahl, S. D. Stranks, H. Nagaoka, G. E. Eperon,
M. E. Ziffer, H. J. Snaith, and D. S. Ginger. Impact of microstructure on local
carrier lifetime in perovskite solar cells. Science, 348(6235):683–686, 2015.
[275] Y. Kutes, L. Ye, Y. Zhou, S. Pang, B. D. Huey, and N. P. Padture. Di-
rect observation of ferroelectric domains in solution-processed CH3NH3PbI3
perovskite thin films. The Journal of Physical Chemistry Letters, 5(19):3335–
3339, 2014.
[276] J. Beilsten-Edmands, G. E. Eperon, R. D. Johnson, H. J. Snaith, and
P. G. Radaelli. Non-ferroelectric nature of the conductance hysteresis in
CH3NH3PbI3 perovskite-based photovoltaic devices. Applied Physics Letters,
106(17):173502, 2015.
[277] V. W. Bergmann, Yu. Guo, H. Tanaka, I. M. Hermes, D. Li, A. Klasen,
S. A. Bretschneider, E. Nakamura, R.r Berger, and S.A.L. Weber. Local
time-dependent charging in a perovskite solar cell. ACS applied materials &
interfaces, 8(30):19402–19409, 2016.
[278] U. Zerweck, C. Loppacher, T. Otto, S. Grafström, and L. M Eng. Accuracy
and resolution limits of Kelvin probe force microscopy. Physical Review B,
71(12):125424, 2005.
[279] J. M. R. Weaver and D. W. Abraham. High resolution atomic force mi-
croscopy potentiometry. Journal of Vacuum Science & Technology B: Micro-
electronics and Nanometer Structures Processing, Measurement, and Phenom-
ena, 9(3):1559–1561, 1991.
183
[280] C.-S. Jiang, M. Yang, Y. Zhou, B. To, S. U. Nanayakkara, J. M. Luther,
W. Zhou, J. J. Berry, J. van de Lagemaat, N. P. Padture, K. Zhu, and M. M.
Al-Jassim. Carrier separation and transport in perovskite solar cells studied
by nanometre-scale profiling of electrical potential. Nature Communications,
6:8397, 2015.
[281] Z. Xiao, Y. Yuan, Y. Shao, Q. Wang, Q. Dong, C. Bi, P. Sharma, A. Gru-
verman, and J. Huang. Giant switchable photovoltaic effect in organometal
trihalide perovskite devices. Nature Materials, 14(2):193–198, 2015.
[282] R. Sheng, A. W. Y. Ho-Baillie, S. Huang, M. Keevers, X. Hao, L. Jiang,
Y.-B. Cheng, and M. A. Green. Four-terminal tandem solar cells using
CH3NH3PbBr3 by spectrum splitting. The journal of physical chemistry let-
ters, 6(19):3931–3934, 2015.
[283] Z. Xiao, C. Bi, Y. Shao, Q. Dong, Q. Wang, Y. Yuan, C. Wang, Y. Gao,
and J. Huang. Efficient, high yield perovskite photovoltaic devices grown
by interdiffusion of solution-processed precursor stacking layers. Energy &
Environmental Science, 7(8):2619–2623, 2014.
[284] H.-B. Kim, H. Choi, J. Jeong, S. Kim, B. Walker, S. Song, and J. Y. Kim.
Mixed solvents for the optimization of morphology in solution-processed,
inverted-type perovskite/fullerene hybrid solar cells. Nanoscale, 6(12):6679–
6683, 2014.
[285] W. Shenghao, S. Takeaki, K. Ryusuke, and A. Katsuhiro. Energy level align-
ment of C60 / Ca interface with bathocuproine as an interlayer studied by
ultraviolet photoelectron spectroscopy. Japanese Journal of Applied Physics,
51(10S):10NE32, 2012.
[286] I. Eiji and S. Toshiki. Relationship between work function of hole collection
electrode and temperature dependence of open-circuit voltage in multilayered
organic solar cells. Japanese Journal of Applied Physics, 51(2S):02BK14, 2012.
[287] J.-Y. Jeng, Y.-F. Chiang, M.-H. Lee, S.-R. Peng, T.-F. Guo, P. Chen, and
T.-C. Wen. CH3NH3PbI3 perovskite/fullerene planar-heterojunction hybrid
solar cells. Advanced Materials, 25(27):3727–3732, 2013.
[288] M. Hu, C. Bi, Y. Yuan, Z. Xiao, Q. Dong, Y. Shao, and J. Huang. Distinct
exciton dissociation behavior of organolead trihalide perovskite and excitonic
semiconductors studied in the same system. Small, 11(18):2164–2169, 2015.
[289] Y. Sugawara, L. Kou, Z. Ma, T. Kamijo, Y. Naitoh, and Y. Jun Li. High
potential sensitivity in heterodyne amplitude-modulation Kelvin probe force
microscopy. Applied Physics Letters, 100(22):223104, 2012.
[290] S. Jesse, N. Balke, E. Eliseev, A. Tselev, N. J. Dudney, A. N. Morozovska,
and S. V. Kalinin. Direct mapping of ionic transport in a Si anode on the
184
nanoscale: time domain electrochemical strain spectroscopy study. ACS Nano,
5(12):9682–9695, 2011.
[291] A. Kumar, T. M. Arruda, A. Tselev, I. N. Ivanov, J. S. Lawton, T. A. Za-
wodzinski, O. Butyaev, S. Zayats, S.n Jesse, and S. V. Kalinin. Nanometer-
scale mapping of irreversible electrochemical nucleation processes on solid li-
ion electrolytes. Scientific Reports, 3:1621, 2013.
[292] N. Balke, S. Jesse, A. N. Morozovska, E. Eliseev, D. W. Chung, Y. Kim,
L. Adamczyk, R. E. Garcia, N. Dudney, and S. V. Kalinin. Nanoscale mapping
of ion diffusion in a lithium-ion battery cathode. Nature Nanotechnology,
5(10):749–754, 2010.
[293] J. D. Adams, B. W. Erickson, J. Grossenbacher, J. Brugger, A. Nievergelt, and
G. E. Fantner. Harnessing the damping properties of materials for high-speed
atomic force microscopy. Nature Nanotechnology, 11(2):147–151, 2016.
[294] T. Ando, N. Kodera, E. Takai, D. Maruyama, K. Saito, and A. Toda. A
high-speed atomic force microscope for studying biological macromolecules.
Proceedings of the National Academy of Sciences, 98(22):12468–12472, 2001.
[295] Milan Sonka, Vaclav Hlavac, and Roger Boyle. Image Processing, Analysis,
and Machine Vision, 4th Edition. 2015.
[296] Y. Deng, Z. Xiao, and J. Huang. Light-induced self-poling effect on
organometal trihalide perovskite solar cells for increased device efficiency and
stability. Advanced Energy Materials, 5(20), 2015.
[297] J. Yang, B. D. Siempelkamp, D. Liu, and T. L. Kelly. Investigation of
CH3NH3PbI3 degradation rates and mechanisms in controlled humidity en-
vironments using in situ techniques. ACS Nano, 9(2):1955–1963, 2015.
[298] E. Strelcov, A. Belianinov, Y.-H. Hsieh, S. Jesse, A. P. Baddorf, Y.-H. Chu,
and S. V. Kalinin. Deep data analysis of conductive phenomena on complex
oxide interfaces: Physics from data mining. ACS Nano, 8(6):6449–6457, 2014.
[299] R. Gottesman, E. Haltzi, L. Gouda, S. Tirosh, Y. Bouhadana, A. Zaban,
E. Mosconi, and F. De Angelis. Extremely slow photoconductivity response
of CH3NH3PbI3 perovskites suggesting structural changes under working con-
ditions. The journal of physical chemistry letters, 5(15):2662–2669, 2014.
[300] J. M. Frost, K. T. Butler, F. Brivio, C. H. Hendon, M. Van Schilfgaarde, and
A. Walsh. Atomistic origins of high-performance in hybrid halide perovskite
solar cells. Nano Letters, 14(5):2584–2590, 2014.
[301] T. Leijtens, S. D. Stranks, G. E. Eperon, R. Lindblad, E. M. J. Johans-
son, I. J. McPherson, H. Rensmo, J. M. Ball, M. M. Lee, and H. J. Snaith.
Electronic properties of meso-superstructured and planar organometal halide
185
perovskite films: Charge trapping, photodoping, and carrier mobility. ACS
Nano, 8(7):7147–7155, 2014.
[302] Durmus U. Karatay, Jeffrey S. Harrison, Micah S. Glaz, Rajiv Girid-
haragopal, and David S. Ginger. Fast time-resolved electrostatic force mi-
croscopy: Achieving sub-cycle time resolution. Review of Scientific Instru-
ments, 87(5):053702, 2016.
[303] J. Murawski, T. Graupner, P. Milde, R. Raupach, U. Zerweck-Trogisch, and
L. M. Eng. Pump-probe Kelvin-probe force microscopy: Principle of operation
and resolution limits. Journal of Applied Physics, 118(15):154302, 2015.
[304] J. P. Correa-Baena, A. Abate, M. Saliba, W. Tress, T. J. Jacobsson,
M. Gratzel, and A. Hagfeldt. The rapid evolution of highly efficient perovskite
solar cells. Energy & Environmental Science, 10(3):710–727, 2017.
[305] W. S. Yang, B.-W. Park, E. H. Jung, N. J. Jeon, Y. C. Kim, D. U. Lee,
S. S. Shin, J. Seo, E. K. Kim, J. H. Noh, and S. I. Seok. Iodide management
in formamidinium-lead-halide-based perovskite layers for efficient solar cells.
Science, 356(6345):1376–1379, 2017.
[306] Y.-L. Wu, D. Yan, J. Peng, T. Duong, Y. Wan, S. P. Phang, H. Shen, N. Wu,
C. Barugkin, X. Fu, S. Surve, D. Grant, D. Walter, T. P. White, K. R. Catch-
pole, and K. J. Weber. Monolithic perovskite/silicon-homojunction tandem
solar cell with over 22% efficiency. Energy & Environmental Science, 10:2472–
2479, 2017.
[307] J. Peng, Y. Wu, W. Ye, D. A. Jacobs, H. Shen, X. Fu, Y. Wan, T. Duong,
N. Wu, C. Barugkin, H. T. Nguyen, D. Zhong, J. Li, T. Lu, Y. Liu, M. N.
Lockrey, K. J. Weber, K. R. Catchpole, and T. P. White. Interface passivation
using ultrathin polymer-fullerene films for high-efficiency perovskite solar cells
with negligible hysteresis. Energy & Environmental Science, 10(8):1792–1800,
2017.
[308] Z. Wang, Q. Lin, F. P. Chmiel, N. Sakai, L. M. Herz, and H. J. Snaith. Efficient
ambient-air-stable solar cells with 2D3D heterostructured butylammonium-
caesium-formamidinium lead halide perovskites. Nature Energy, 6:17135,
2017.
[309] M. He, B. Li, X. Cui, B. Jiang, Y. He, Y. Chen, D. ONeil, P. Szymanski, M. A.
Ei-Sayed, J. Huang, and Z. Lin. Meniscus-assisted solution printing of large-
grained perovskite films for high-efficiency solar cells. Nature Communications,
8:16045, 2017.
[310] R. Brenes, D. Guo, A. Osherov, N. K. Noel, C. Eames, E. M. Hutter, S. K.
Pathak, F. Niroui, R. H. Friend, M. S. Islam, H. J. Snaith, V. Bulović, T. J.
Savenije, and S. D. Stranks. Metal halide perovskite polycrystalline films
exhibiting properties of single crystals. Joule, 1(1):155–167, 2017.
186
[311] A. Ummadisingu, L. Steier, J.-Y. Seo, T. Matsui, A. Abate, W. Tress, and
M. Grätzel. The effect of illumination on the formation of metal halide per-
ovskite films. Nature, 545(7653):208–212, 2017.
[312] C. Eames, J. M. Frost, P. R. F. Barnes, B. C. ORegan, A. Walsh, and M. S.
Islam. Ionic transport in hybrid lead iodide perovskite solar cells. Nature
Communications, 6:7497, 2015.
[313] S. J. Yoon, S. Draguta, J. S. Manser, O. Sharia, W. F. Schneider, M. Kuno,
and P. V. Kamat. Tracking iodide and bromide ion segregation in mixed halide
lead perovskites during photoirradiation. ACS Energy Letters, 1(1):290–296,
2016.
[314] Z. Li, C. Xiao, Y. Yang, S. P. Harvey, D. H. Kim, J. A. Christians, M. Yang,
P. Schulz, S. U. Nanayakkara, C.-S. Jiang, J. M. Luther, J. J. Berry, M. C.
Beard, M. M. Al-Jassim, and K. Zhu. Extrinsic ion migration in perovskite
solar cells. Energy & Environmental Science, 10(5):1234–1242, 2017.
[315] O. A. Syzgantseva, M. Saliba, M. Grätzel, and U. Rothlisberger. Stabilization
of the perovskite phase of formamidinium lead triiodide by methylammonium,
Cs, and/or Rb doping. The Journal of Physical Chemistry Letters, 8(6):1191–
1196, 2017.
[316] Q. Ma, S. Huang, S. Chen, M. Zhang, C. F. J. Lau, M. N. Lockrey, H. K.
Mulmudi, Y. Shan, J. Yao, J. Zheng, X. Deng, K. Catchpole, M. A. Green,
and A. W. Y. Ho-Baillie. The effect of stoichiometry on the stability of inor-
ganic cesium lead mixed-halide perovskites solar cells. The Journal of Physical
Chemistry C, 121(36):19642–19649, 2017.
[317] Y. Hu, M. F. Aygüler, M. L. Petrus, T. Bein, and P. Docampo. Impact
of rubidium and cesium cations on the moisture stability of multiple-cation
mixed-halide perovskites. ACS Energy Letters, 2(10):2212–2218, 2017.
[318] J.-W. Lee, D.-H. Kim, H.-S. Kim, S.-W. Seo, S. M. Cho, and N.-G. Park.
Formamidinium and cesium hybridization for photo- and moisture-stable per-
ovskite solar cell. Advanced Energy Materials, 5(20):1501310, 2015.
[319] T. Matsui, J.-Y. Seo, M. Saliba, S. M. Zakeeruddin, and M. Grätzel. Room-
temperature formation of highly crystalline multication perovskites for effi-
cient, low-cost solar cells. Advanced Materials, 29(15):1606258, 2017.
[320] M. Saliba, T. Matsui, K. Domanski, J.-Y. Seo, A. Ummadisingu, S. M Za-
keeruddin, J.-P. Correa-Baena, W. R. Tress, A. Abate, and A. Hagfeldt. Incor-
poration of rubidium cations into perovskite solar cells improves photovoltaic
performance. Science, 354(6309):206–209, 2016.
[321] M. T. Klug, A. Osherov, A. A. Haghighirad, S. D. Stranks, P. R. Brown, S. Bai,
J. T. W. Wang, X. Dang, V. Bulovic, H. J. Snaith, and A. M. Belcher. Tailoring
187
metal halide perovskites through metal substitution: influence on photovoltaic
and material properties. Energy & Environmental Science, 10(1):236–246,
2017.
[322] T. J. Jacobsson, J.-P. Correa-Baena, M. Pazoki, M. Saliba, K. Schenk,
M. Grätzel, and A. Hagfeldt. Exploration of the compositional space for
mixed lead halogen perovskites for high efficiency solar cells. Energy & Envi-
ronmental Science, 9(5):1706–1724, 2016.
[323] Y. Hu, F. Bai, X. Liu, Q. Ji, X. Miao, T. Qiu, and S. Zhang. Bismuth
incorporation stabilized α-CsPbI3 for fully inorganic perovskite solar cells.
ACS Energy Letters, pages 2219–2227, 2017.
[324] S. Yang, W. Fu, Z. Zhang, H. Chen, and C.-Z. Li. Recent advances in per-
ovskite solar cells: efficiency, stability and lead-free perovskite. Journal of
Materials Chemistry A, 5(23):11462–11482, 2017.
[325] T. A. Berhe, W.-N. Su, C.-H. Chen, C.-J. Pan, J.-H. Cheng, H.-M. Chen,
M.-C. Tsai, L.-Y. Chen, A. A. Dubale, and B.-J. Hwang. Organometal halide
perovskite solar cells: degradation and stability. Energy & Environmental
Science, 9(2):323–356, 2016.
[326] D. P. McMeekin, G. Sadoughi, W. Rehman, G. E. Eperon, M. Saliba, M. T.
Hörantner, A. Haghighirad, N. Sakai, L. Korte, B. Rech, M. B. Johnston,
L. M. Herz, and H. J. Snaith. A mixed-cation lead mixed-halide perovskite
absorber for tandem solar cells. Science, 351(6269):151–155, 2016.
[327] R. G. Niemann, A. G. Kontos, D. Palles, E. I. Kamitsos, A. Kaltzoglou,
F. Brivio, P. Falaras, and P. J. Cameron. Halogen effects on ordering and
bonding of CH3NH
+
3 in CH3NH3PbX3 (X = Cl, Br, I) hybrid perovskites:
A vibrational spectroscopic study. The Journal of Physical Chemistry C,
120(5):2509–2519, 2016.
[328] M. Ledinský, P. Löper, B. Niesen, J. Holovský, S.-J. Moon, J.-H. Yum,
S. De Wolf, A. Fejfar, and C. Ballif. Raman spectroscopy of organic-inorganic
halide perovskites. The Journal of Physical Chemistry Letters, 6(3):401–406,
2015.
[329] J. M. Azpiroz, E. Mosconi, J. Bisquert, and F. De Angelis. Defect migration
in methylammonium lead iodide and its role in perovskite solar cell operation.
Energy & Environmental Science, 8(7):2118–2127, 2015.
[330] S. Ghosh, S. K. Pal, K. J. Karki, and T. Pullerits. Ion migration heals trapping
centers in CH3NH3PbBr3 perovskite. ACS Energy Letters, pages 2133–2139,
2017.
[331] A. M. Tirmzi, R. P. Dwyer, T. Hanrath, and J. A. Marohn. Coupled slow and
fast charge dynamics in cesium lead bromide perovskite. ACS Energy Letters,
2(2):488–496, 2017.
188
[332] J.-P. Correa-Baena, M. Saliba, T Buonassisi, M. Grätzel, A. Abate, W. Tress,
and A. Hagfeldt. Promises and challenges of perovskite solar cells. Science,
358(6364):739–744, 2017.
[333] A. H. Slavney, L. Leppert, D. Bartesaghi, A. Gold-Parker, M. F. Toney, T. J.
Savenije, J. B. Neaton, and H. I. Karunadasa. Defect-induced band-edge re-
construction of a bismuth-halide double perovskite for visible-light absorption.
Journal of the American Chemical Society, 139(14):5015–5018, 2017.
[334] T. Coenen, S. V. den Hoedt, and A. Polman. A new cathodoluminescence
system for nanoscale optics, materials science, and geology. Microscopy Today,
24(03):12–19, 2016.
[335] D. Abou-Ras, T. Kirchartz, and U. Rau. Advanced Characterization Tech-
niques for Thin Film Solar Cells. Wiley-VCH Verlag GmbH Co, 2011.
[336] M. Müller, S. Ribbe, T. Hempel, F. Bertram, J. Christen, W. Witte, S. Pae-
tel, and M. Powalla. Investigation of vertical compositional gradients in
Cu(In,Ga)Se2 by highly spatially and spectrally resolved cathodoluminescence
microscopy. Thin Solid Films, 535(0):270–274, 2013.
[337] V. Consonni, N. Baier, O. Robach, C. Cayron, F. Donatini, and G. Feuillet.
Local band bending and grain-to-grain interaction induced strain nonunifor-
mity in polycrystalline cdte films. Physical Review B, 89(3):035310, 2014.
[338] C. G. Bischak, E. M. Sanehira, J. T. Precht, J. M. Luther, and N. S. Ginsberg.
Heterogeneous charge carrier dynamics in organic-inorganic hybrid materials:
Nanoscale lateral and depth-dependent variation of recombination rates in
methylammonium lead halide perovskite thin films. Nano Letters, 15(7):4799–
4807, 2015.
[339] M. I. Dar, G. Jacopin, M. Hezam, N. Arora, S. M. Zakeeruddin, B. De-
veaud, M. K. Nazeeruddin, and M. Grätzel. Asymmetric cathodoluminescence
emission in CH3NH3PbI3−xBrx perovskite single crystals. ACS Photonics,
3(6):947–952, 2016.
[340] H. Guthrey, S. Johnston, D. N. Weiss, S. Grover, K. Jones, A. Blosse, and
M. Al-Jassim. Three-dimensional minority-carrier collection channels at shunt
locations in silicon solar cells. Solar Energy, 135:163–168, 2016.
[341] A. Belghachi. Theoretical Calculation of the Efficiency Limit for Solar Cells.
InTech, 2015.
[342] J. Sempau, E. Acosta, J. Baro, J. M. Fernndez-Varea, and F. Salvat. An algo-
rithm for monte carlo simulation of coupled electron-photon transport. Nuclear
Instruments and Methods in Physics Research Section B: Beam Interactions
with Materials and Atoms, 132(3):377–390, 1997.
189
[343] Y. Kutes, B. A. Aguirre, J. L. Bosse, J. L. Cruz-Campa, D. Zubia, and B. D.
Huey. Mapping photovoltaic performance with nanoscale resolution. Progress
in Photovoltaics: Research and Applications, 24(3):315–325, 2016.
[344] G. E. Eperon, D. Moerman, and D. S. Ginger. Anticorrelation between local
photoluminescence and photocurrent suggests variability in contact to active
layer in perovskite solar cells. ACS Nano, 10(11):10258–10266, 2016.
[345] S. Smith, P. Zhang, T. Gessert, and A. Mascarenhas. Near-field optical beam-
induced currents in CdTe/CdS solar cells: Direct measurement of enhanced
photoresponse at grain boundaries. Applied Physics Letters, 85(17):3854–3856,
2004.
[346] N. M. Haegel. Integrating electron and near-field optics: Dual vision for the
nanoworld. Nanophotonics, 3(1-2):75–89, 2014.
[347] S. Shirakata and T. Nakada. Time-resolved photoluminescence in
Cu(In,Ga)Se2 thin films and solar cells. Thin Solid Films, 515(15):6151–6154,
2007.
[348] Teodor K. Todorov, Jiang Tang, Santanu Bag, Oki Gunawan, Tayfun Gokmen,
Yu Zhu, and David B. Mitzi. Beyond 11cu2znsn(s,se)4 solar cells. Advanced
Energy Materials, 3(1):34–38, 2013.
[349] V. D′Innocenzo, S. R. A. Kandada, M. De Bastiani, M. Gandini, and
A. Petrozza. Tuning the light emission properties by band gap engineering
in hybrid lead halide perovskite. Journal of the American Chemical Society,
136(51):17730–17733, 2014.
[350] D. W. de Quilettes, S. Koch, S. Burke, R. K. Paranji, A. J. Shropshire, M. E.
Ziffer, and D. S. Ginger. Photoluminescence lifetimes exceeding 8 µs and quan-
tum yields exceeding 30% in hybrid perovskite thin films by ligand passivation.
ACS Energy Letters, 1(2):438–444, 2016.
[351] E. S. Barnard, E. T. Hoke, S. T. Connor, J. R. Groves, T. Kuykendall, Z. Yan,
E. C. Samulon, E. D. Bourret-Courchesne, S. Aloni, P. J. Schuck, C. H. Peters,
and B. E. Hardin. Probing carrier lifetimes in photovoltaic materials using
subsurface two-photon microscopy. Scientific Reports, 3:2098, 2013.
[352] W. Chen, X. Wen, M. Latzel, M. Heilmann, J. Yang, X. Dai, S. Huang,
S. Shrestha, R. Patterson, S. Christiansen, and G. Conibeer. Nanoscale char-
acterization of carrier dynamic and surface passivation in InGaN/GaN mul-
tiple quantum wells on GaN nanorods. ACS Applied Materials & Interfaces,
8(46):31887–31893, 2016.
[353] A. Richter, G. Behme, M. Sptitz, Ch Lienau, T. Elsaesser, M. Ramsteiner,
R. Ntzel, and K. H. Ploog. Real-space transfer and trapping of carriers into
single GaAs quantum wires studied by near-field optical spectroscopy. Physical
Review Letters, 79(11):2145–2148, 1997.
190
[354] S. Chakraborty, W. Xie, N. Mathews, M. Sherburne, R. Ahuja, M. Asta,
and S. G. Mhaisalkar. Rational design: A high-throughput computational
screening and experimental validation methodology for lead-free and emergent
hybrid perovskites. ACS Energy Letters, 2(4):837–845, 2017.
[355] Z. Shi, J. Guo, Y. Chen, Q. Li, Y. Pan, H. Zhang, Y. Xia, and W. Huang.
Lead-free organic-inorganic hybrid perovskites for photovoltaic applications:
Recent advances and perspectives. Advanced Materials, 29(16):1605005, 2017.
[356] T. Duong, Y. Wu, H. Shen, J. Peng, X. Fu, D. Jacobs, E.-C. Wang, T. C.
Kho, K. C. Fong, M. Stocks, E. Franklin, A. Blakers, N. Zin, K. McIntosh,
W. Li, Y.-B. Cheng, T. P. White, K. Weber, and K. Catchpole. Rubidium
multication perovskite with optimized bandgap for perovskite-silicon tandem
with over 26% efficiency. Advanced Energy Materials, page 1700228, 2017.
[357] S. V. Kalinin, B. G. Sumpter, and R. K. Archibald. Big-deep-smart data in
imaging for guiding materials design. Nature Materials, 14(10):973–980, 2015.
[358] S. Jesse and S. V. Kalinin. Principal component and spatial correlation analy-
sis of spectroscopic-imaging data in scanningprobe microscopy. Nanotechnol-
ogy, 20(8):085714, 2009.
191