INFLUENCE OF LIGAND EXCHANGE ON COPPER REDOX SHUTTLES IN DYE-SENSITIZED SOLAR 
CELLS 

By 

Eric James Firestone 

A DISSERTATION 

Submitted to 
Michigan State University 
in partial fulfillment of the requirements   
for the degree of 

Chemistry – Doctor of Philosophy 

2024 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ABSTRACT 

Dye-sensi(cid:415)zed  solar  cells  (DSSCs)  are  recognized  as  a  promising,  eco-friendly  alterna(cid:415)ve  to 

tradi(cid:415)onal photovoltaics, characterized by their unique light-harves(cid:415)ng capabili(cid:415)es and poten(cid:415)al 

for enhanced eﬃciency and stability in renewable energy applica(cid:415)ons. In this study, the focus is 

on  the  development  of  copper  redox  complexes,  which  exhibit  varying  responses  to  ligand 

exchange upon the introduc(cid:415)on of 4-tertbutylpyridine (TBP), a cri(cid:415)cal factor inﬂuencing electron 

transfer processes and the overall performance of DSSCs. Three copper complexes, copper(II/I) 

N,N(cid:3387)-Dibenzyl-N,N(cid:3387)-bis(6-methylpyridin-2-ylmethyl)ethylenediamine  triﬂate,  [Cu(dbmed)]OTf1/2, 

copper(II/I) 2,6-bis[1,1-bis(2-pyridyl)ethyl]pyridine triﬂate, [Cu(PY5)]OTf1/2 and copper(II/I) 6,6(cid:3387)-

bis(1,1-di(pyridine-2-yl)ethyl)-2,2(cid:3387)-bipyridine bistriﬂimide, [Cu(bpyPY4)]TFSI1/2 were inves(cid:415)gated, 

with  synthe(cid:415)c  and  electrochemical  methodologies,  including  UV-Vis  spectroscopy,  NMR,  and 

cyclic  voltammetry,  being  u(cid:415)lized  to  examine  the  redox  behavior  and  ligand  exchange 

phenomena. The performance of DSSC devices with each redox mediator was measured to be 

4.32%, 2.01%, and 1.23% respec(cid:415)vely. Methodology developed in this study, which involves using 

redox poten(cid:415)al as an indicator to predict ligand exchange events, represents an expansion and 

reﬁnement  of  exis(cid:415)ng  concepts  found  in  the  literature.  By  building  on  prior  research,  this 

approach not only deepens the understanding of  copper complex behaviors in DSSCs but also 

oﬀers a more nuanced perspec(cid:415)ve for enhancing the design and eﬃciency of these innova(cid:415)ve 

solar cells. 

 
 
 
 
Copyright by  
ERIC JAMES FIRESTONE 
2024  

 
 
  
  
  
  
  
  
  
  
 
 
 
 
 
 
 
 
 
 
 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
 
 
 
 
 
 
 
 
ACKNOWLEDGMENTS 

Firstly I would like to thank my mentor, Thomas Hamann, for all the years he spent working to 

help  me  be(cid:425)er  myself  as  a  scien(cid:415)st  and  a  chemist.  Without  his  assistance  I would not be  the 

scien(cid:415)st I am today. I would also like to thank the rest of my commi(cid:425)ee, Dr. James McCusker, Dr. 

Aaron  Odom,  Dr.  Mitch  Smith,  and  Dr.  Thomas  O’Halloran,  for  their  assistance  in  helping  me 

understand mmy chemistry over the years. I want my fellow Hamannites, for all of the assistance 

both with the science and non-scien(cid:415)ﬁc issues that have arisin over the years. I especially want 

to thank those that were on the DSSC project,  Yujue, Aus(cid:415)n, Samhita, Abubaker, and Michael, 

whose constant discussions and advice greatly improved my understanding of my work. For those 

that were under my care these last three years thank you for being kind to my sugges(cid:415)ons and 

instruc(cid:415)on. I am certain I made mistakes, but you both were kind and pa(cid:415)ent through them all. I 

would also like to thank my friends both from before and during my (cid:415)me here at MSU, Josh, Karl, 

Mike,  Jackie,  Rachel,  Julie,  Adam,  Connor,  Nick,  Ana,  Courtney,  Spencer,  Shane,  Hailey,  Lizzie, 

Dominic,  Cash,  Ben,  Jack,  Ka(cid:415)e,  Allison,  Sophia,  Maggie,  Arzoo,  Noel,  Courtney,  Rob,  Tom, 

Veronica, each and every one of you will hold a special place in my heart, without you all I don’t 

think I would have been able to sane across these last few years. Finally I would like to thank my 

family, without whose support I would never be able to make this far in my pursuit of scien(cid:415)ﬁc 

knowledge. From the bo(cid:425)om of my heart, thank you. 

iv 

 
 
 
 
TABLE OF CONTENTS 

Chapter 1 – Introduc(cid:415)on ................................................................................................................ 1 
1.1 Mo(cid:415)va(cid:415)ons in solar energy ................................................................................................... 1 
1.2 History of DSSCs .................................................................................................................... 3 
1.3 Redox Shu(cid:425)les in DSSCs ........................................................................................................ 5 

Chapter 2 – Inves(cid:415)ga(cid:415)on of [Cu(dbmed)](OTf)1/2 ....................................................................... 12 
2.1 Introduc(cid:415)on ......................................................................................................................... 12 
2.2 Experimental Details ........................................................................................................... 13 
2.3 Results and Discussion ........................................................................................................ 18 
2.4 Conclusion ........................................................................................................................... 25 

Chapter 3 – Inves(cid:415)ga(cid:415)on of [Cu(bpyPY4)](TFSI)1/2 ..................................................................... 26 
3.1 Introduc(cid:415)on ......................................................................................................................... 26 
3.2 Experimental Details ........................................................................................................... 27 
3.3 Results and Discussion ........................................................................................................ 32 
3.4 Conclusions .......................................................................................................................... 47 

Chapter 4 – Inves(cid:415)ga(cid:415)on of [Cu(PY5)]2+/+ ................................................................................... 49 
4.1 Introduc(cid:415)on ......................................................................................................................... 49 
4.2 Experimental Details ........................................................................................................... 51 
4.3 Results and Discussion ........................................................................................................ 55 
4.4 Conclusion ........................................................................................................................... 81 

Chapter 5 – Developing a model to predict ligand exchange in DSSCs ...................................... 83 
5.1 Introduc(cid:415)on ......................................................................................................................... 83 
5.2 Experimental Details ........................................................................................................... 86 
5.3 Results and Discussion ........................................................................................................ 89 
5.4 Conclusions .......................................................................................................................... 98 

WORKS CITED ............................................................................................................................... 99 

APPENDIX A: CHAPTER 2 SUPPLEMENTARY DATA ..................................................................... 110 

APPENDIX B: CHAPTER 3 SUPPLEMENTARY DATA ..................................................................... 114 

APPENDIX C: CHAPTER 4 SUPPLEMENTARY DATA ..................................................................... 116 

v 

 
 
 
 
Chapter 1 – Introduc(cid:415)on 

1.1 Mo(cid:415)va(cid:415)ons in solar energy 

Energy is a cornerstone of modern society, and its consump(cid:415)on has profound implica(cid:415)ons on our 

environment, most notably in the form of increasing CO₂ emissions. Data from the past decade, 

speciﬁcally  between  2010  and  2022,  highlights  a  rise  in  CO₂  emissions  by  1.0%  annually, 

amoun(cid:415)ng to roughly 9.9 ± 0.5 GtC yr⁻¹. This has resulted in an atmospheric CO2 concentra(cid:415)on 

51% higher than pre-industrial levels.1 An alarming 82% of these emissions can be traced back to 

fossil fuel consump(cid:415)on.2 This rapid surge in energy use, par(cid:415)cularly from fossil fuels, has resulted 

in energy-related CO₂ emissions surging to record highs, further underscoring the urgent need for 

carbon-neutral energy sources. 

If  current  trends  in  emission  mi(cid:415)ga(cid:415)ons  and  renewable  energy  adop(cid:415)on  persist,  projec(cid:415)ons 

suggest a 3°C global temperature rise from pre-industrial levels by 2100.3 This poten(cid:415)al escala(cid:415)on 

intensiﬁes  the  impera(cid:415)ve  for  a  shi(cid:332)  toward  carbon-neutral  energy  solu(cid:415)ons.  Solar  energy, 

boas(cid:415)ng an overwhelming theore(cid:415)cal poten(cid:415)al of approximately 173,000 TW, dwarfs the current 

global demand of 2.9 TW and outpaces poten(cid:415)al outputs from other renewable sources like wind 

and hydro.4,5 Although renewable energy's contribu(cid:415)on to global power reached a commendable 

12.8%  in  2022,5  broader  adop(cid:415)on  is  constrained  by  factors  such  as  genera(cid:415)on  costs  and 

fabrica(cid:415)on complexity. 

1 

 
 
Figure 1.1: The 1.5 AM solar spectrum which represents the average amount of light that 
reaches the surface of the Earth across the con(cid:415)nental United States.6 

Solar  cells  present  a  promising  avenue  in  this  regard.  Their  rising  prevalence,  combined  with 

declining costs due to the SunShot Ini(cid:415)a(cid:415)ve, signiﬁes the growing feasibility of solar energy. The 

SunShot  Ini(cid:415)a(cid:415)ve  is  a  DOE  program  with  the  goal  to  reduce  the  cost  of  solar  energy  by  75%, 

mee(cid:415)ng  the  2020  goal by  2017.7  While  the  reduc(cid:415)on  in fabrica(cid:415)on  costs  of  tradi(cid:415)onal  silicon 

solar cells has played a role in this price drop, further reduc(cid:415)ons face challenges chieﬂy due to 

stringent fabrica(cid:415)on requirements. The quest for more cost-eﬀec(cid:415)ve solar solu(cid:415)ons necessitates 

the explora(cid:415)on of alterna(cid:415)ve photovoltaics with simpler fabrica(cid:415)on processes, reduced energy 

input, and the use of readily available materials. 

2 

 
 
 
Dye-Sensi(cid:415)zed  Solar  Cells  (DSSCs)  emerge  as  a  frontrunner  in  this  category.  Their  poten(cid:415)al 

aﬀordability stems from their straigh(cid:414)orward fabrica(cid:415)on and the u(cid:415)liza(cid:415)on of earth-abundant 

materials.  However,  to  truly  posi(cid:415)on  DSSCs  as  a  game-changer  in  the  renewable  energy 

landscape,  research  must  intensify  around  enhancing  their  eﬃciency  and  trimming  material 

costs, thereby ensuring their economic viability and widespread adop(cid:415)on.8 

1.2 History of DSSCs 

The legacy of DSSCs is based on the development of the theory of photoelectrochemistry, which 

can be traced back to Edmond Becquerel's groundbreaking work in 1839.9 Since the discovery of 

photoelectrochemistry, it has been used to develop methods of genera(cid:415)ng clean energy.  

Semiconductors are materials with speciﬁc bandgaps that can absorb photons of suitable energy. 

When they absorb these photons, they generate electron-hole pairs that can separate within the 

semiconductor and par(cid:415)cipate in redox reac(cid:415)ons on diﬀerent surfaces. U(cid:415)lizing this idea, in 1873 

and  1887  respec(cid:415)vely,  Vogel  and  Moser  independently  explored  "sensi(cid:415)za(cid:415)on"  using  silver 

halides  mixed  with  visible-light-absorbing  dyes.10  The  theory  was  further  elucidated  almost  a 

century later, in 1969, by Gerischer, who detailed how excited state dissolved dyes could inject 

electrons  into  ZnO  semiconductor  crystals.11  Further  advancements  were  made  with  the 

anchoring of ruthenium and quinone sensi(cid:415)zers by two researchers, Weber and Matsumura in 

1979 and 1980, respec(cid:415)vely.12,13 

3 

 
 
Figure 1.2: The composi(cid:415)on of a DSSC.  

Grӓtzel  and  his  team  exponen(cid:415)ally  enhanced  current  genera(cid:415)on  using  polycrystalline  anatase 

TiO2  ﬁlms  and  the  iodide/triodide  redox  shu(cid:425)le.14  By  1993,  they  achieved  a  10%  power 

conversion  eﬃciency  (PCE),  thanks  to  reﬁned  TiO2  electrodes  and  expanded  light-harves(cid:415)ng 

sensi(cid:415)zers.15 However, with minimal improvements beyond 10% PCE the progress plateaued un(cid:415)l 

2010  when  outer-sphere  redox shu(cid:425)les became prevalent  in  DSSCs.16  The  subsequent  decade 

saw organic sensi(cid:415)zers paired with outer-sphere redox shu(cid:425)les, pushing PCEs to 15%.17–20  

In the ensuing sec(cid:415)ons, the intricate opera(cid:415)ons of the dye-sensi(cid:415)zed solar cell will be delved into, 

followed by a discussion on the factors limi(cid:415)ng the PCE of these devices. 

DSSCs are a unique type of photovoltaic that distributes charge separa(cid:415)on and collec(cid:415)on over 

mul(cid:415)ple components, in contrast to conven(cid:415)onal crystalline semiconductor photovoltaics. In a 

4 

 
 
 
DSSC,  a  dye  or  sensi(cid:415)zer  is  anchored  to  semiconductor  nanopar(cid:415)cles,  usually  ac(cid:415)ng  as  the 

primary light absorber. When this dye absorbs a photon, an electron is excited to a state where it 

can be injected into the semiconductor's conduc(cid:415)on band. The role of the semiconductor, o(cid:332)en 

materials like TiO2, is to transport these electrons to an FTO substrate (ﬂuorine-doped (cid:415)n oxide). 

However, for the dye to con(cid:415)nue absorbing light, it needs another electron. This is where a redox 

shu(cid:425)le, dissolved in the electrolyte, comes into play. It diﬀuses to the dye's surface, dona(cid:415)ng an 

electron and ensuring the dye's readiness for subsequent photon absorp(cid:415)ons. The oxidized redox 

shu(cid:425)le then moves to the counter electrode to accept an electron, comple(cid:415)ng the circuit. 

The eﬃciency of a DSSC in conver(cid:415)ng sunlight to electrical power depends on several factors. The 

excited dye must have suﬃciently nega(cid:415)ve poten(cid:415)al to inject its electron into the semiconductor. 

The  injec(cid:415)on  must  be  faster  than  excited  dye decays  or recombines  with  the  semiconductor's 

electrons.  The  rate  of  dye  regenera(cid:415)on  should  outpace  its  recombina(cid:415)on  rate.  Furthermore, 

electrons in the semiconductor need to be collected quickly to prevent their recombina(cid:415)on with 

the oxidized redox shu(cid:425)les. The interplay of these processes largely determines the DSSC's power 

conversion eﬃciency. 

1.3 Redox Shu(cid:425)les in DSSCs 

DSSCs draw their essence from the ability to regenerate con(cid:415)nually. Central to this regenera(cid:415)ve 

ability is the redox couple. Predominantly found in a liquid electrolyte form, the redox couple—

both in its reduced and oxidized states—acts as the linchpin, bridging conduc(cid:415)vity between the 

photoanode and the counter electrode. 

For the dynamic to work eﬃciently, the reduced form of the redox couple necessitates a rela(cid:415)ve 

nega(cid:415)ve  reduc(cid:415)on  poten(cid:415)al  compared  to  the  dye  in  its  ground  state.  Once  the  dye  molecule 

5 

 
 
dispatches an electron to the semiconductor, it undergoes oxida(cid:415)on. This is where the reduced 

redox couple, ac(cid:415)ng as an electron donor, plays its part. It regenerates the oxidized dye, priming 

it for the next photon absorp(cid:415)on. The pace at which this regenera(cid:415)on occurs is paramount—it 

contributes  to  the  photocurrent's  magnitude.  Once  it  donates  its  electron,  the  reduced  redox 

species  becomes  oxidized,  primarily  found  near  the  photoanode.  Through  liquid  electrolyte 

diﬀusion, this oxidized form dri(cid:332)s to the counter electrode to reclaim an electron from a catalyst 

surface, comple(cid:415)ng the cycle. 

Not  only  do  redox  couples  play  a  vital  kine(cid:415)c  role,  but  their  poten(cid:415)al  also  shapes  the  DSSC's 

photovoltage output. The open-circuit voltage (VOC) is deﬁned by the energy diﬀerence between 

the Fermi level of the semiconductor and the solu(cid:415)on poten(cid:415)al of the redox shu(cid:425)le. The Fermi 

level of a semiconductor is deﬁned by the equa(cid:415)on  

𝐸(cid:3007) = 𝐸(cid:3030) −

(cid:3038)(cid:3251)(cid:3021)
(cid:3032)

𝑙𝑛

(cid:3041)(cid:3252)
(cid:3015)(cid:3252)

(1.1) 

where EC is the energy of the conduc(cid:415)on band minimum, kB is the Boltzmann constant, T is the 

temperature in Kelvin, e is the elementary charge of an electron, nC is the density of conduc(cid:415)on 

band  electrons,  and  NC  is  the  density  of  states  in  the  conduc(cid:415)on  band.21  The  Fermi  level  is 

inﬂuenced  by  both  the  inﬂux  of  electrons  into  the  conduc(cid:415)on  band  and  the  rate  at  which 

electrons  exit  the  conduc(cid:415)on  band,  whether  through  recombina(cid:415)on  or  charge  separa(cid:415)on.  By 

deploying diverse redox couples, one can skillfully modulate VOC, providing avenues for enhanced 

cell performance. 

6 

 
 
 
 
 
The  equilibrium  poten(cid:415)al  of  the  liquid  electrolyte  at  the  counter  electrode  sets  the  cathode's 

energy level. A redox couple with a more posi(cid:415)ve redox poten(cid:415)al increases the cell's VOC, with a 

limit of the dye’s ground state poten(cid:415)al. In order to eﬃciently regenerate the dye, some energy, 

ac(cid:415)ng  as  a  driving  force  for  electron  transfer,  is  inevitably  lost.  If  the  electron  acceptor  (the 

oxidized form of the redox couple) in the electrolyte has a more posi(cid:415)ve poten(cid:415)al than the Fermi 

Level, it can lead to electron recombina(cid:415)on. According to Marcus theory, faster electron transfer 

rates occur with larger driving forces in the normal region. However, redox couples that transfer 

electrons  rapidly  and  have  a  higher  poten(cid:415)al  o(cid:332)en  face  electron  recombina(cid:415)on,  reducing  the 

electron ﬂow under equilibrium. Thus, adjus(cid:415)ng the redox proper(cid:415)es of the couple is crucial for 

enhancing solar cell performance. 

Triiodide/iodide redox couple's pioneering endeavors set a robust founda(cid:415)on for DSSC eﬃciency. 

These couples, due to their superior electron transfer kine(cid:415)cs, became instrumental in ensuring 

high photocurrent outputs. Studies, notably by Boschloo, delved deep into their mechanisms.22 

Here, the iodide ion would regenerate the dye, subsequently genera(cid:415)ng the diiodide radical. This 

radical  undergoes  rapid  dispropor(cid:415)ona(cid:415)on,  yielding  triiodide  and  iodide.  The  more  posi(cid:415)ve 

reduc(cid:415)on  poten(cid:415)al  of  triiodide  results  in  a  reduc(cid:415)on  of  electron  recombina(cid:415)on,  thereby 

increasing the cell's current output. But this comes at a price: energy loss. 

Despite  their prowess,  the  intricacies  of  triiodide/iodide  electron  transfer  mechanisms remain 

complex, promp(cid:415)ng researchers to explore alternate redox couples. Outer-sphere one electron 

transi(cid:415)on  metal  redox  couples  oﬀer  a  more  straigh(cid:414)orward  electron  transfer  mechanism, 

sidestepping  the  complexi(cid:415)es  seen  in  iodide/triiodide  couples.  An  example  is  the  cobalt(III/II) 

tri(2,2’-bipyridine) couple, which has gained immense trac(cid:415)on.23 Their poten(cid:415)al is exemplary, but 

7 

 
 
challenges persist, par(cid:415)cularly in low electron transfer rates which leads to high driving force to 

regenerate the dye. 

In summary, redox couples not only act as the heartbeat of DSSCs but also oﬀer a rich tapestry of 

variables  for  researchers  to  manipulate.  Balancing  energy  loss  with  eﬃcient  regenera(cid:415)on  and 

electron  transfer  remains  the  ever-present  challenge,  poin(cid:415)ng  towards  a  promising  avenue  of 

future explora(cid:415)on. 

Recent advancements in solar cell technology have spotlighted the promise of copper(II/I) redox 

couples.  Following  several  years  of  research,  the  eﬃciency  of  solar  cells  featuring  copper(II/I) 

complexes has soared, posi(cid:415)oning itself among the elite. Leading to the current highest eﬃciency 

DSSC devices to u(cid:415)lize copper(II/I) redox complexes.20 Remarkably, cells employing these couples 

consistently demonstrate an open-circuit voltage (VOC) exceeding 1 V, a notable leap from cells 

using cobalt complexes, which exhibit less posi(cid:415)ve redox poten(cid:415)als. 

Historically copper complexes were studied due to their role in electron transfer processes, both 

biologically  and  inorganically.  Earlier  studies  gravitated  towards  understanding  copper's  role 

within metalloproteins.24 These proteins, o(cid:332)en sizable and challenging to study, necessitated the 

explora(cid:415)on of small molecule analogs. This eventually propelled studies into the redox chemistry 

of copper(II) coordina(cid:415)on complexes. In this context, the so-called "blue-copper" Type I protein 

garnered signiﬁcant a(cid:425)en(cid:415)on due to its dis(cid:415)nc(cid:415)vely posi(cid:415)ve reduc(cid:415)on poten(cid:415)als, which hinted 

at the protein's u(cid:415)lity in natural electron transport systems.25 Intriguingly, these proper(cid:415)es also 

align perfectly with the requisites for solar energy conversion, paving the way for their adop(cid:415)on 

in this realm. 

8 

 
 
The  ini(cid:415)a(cid:415)on  of  this  paradigm  shi(cid:332)  can  be  a(cid:425)ributed  to  Ha(cid:425)ori  in  2005,  who  introduced  the 

poten(cid:415)al of copper complexes in DSSCs.26 The study delineated three diverse redox couples, all 

with unique ligand environments, and their compa(cid:415)bility with the N719 dye. Of par(cid:415)cular interest 

was the [Cu(SP)(mmt)]0/- complex, which mimicked blue-copper proteins. This seminal work laid 

the groundwork for subsequent researchers to unravel the intricacies of copper redox couples, 

ul(cid:415)mately  culmina(cid:415)ng  in  Bai's  breakthrough  in  2011  that  showcased  a  marked  eﬃciency  leap 

from  1.4%  to  7.0%,  by    pairing  a  copper  redox  couple  with  an  organic  dye  bound  to  the 

semiconductor.27 

As researchers delved deeper into the interplay between dyes and copper redox couples, more 

nuanced  ﬁndings  emerged.  Colombo's  explora(cid:415)on  with  bulky  asymmetric  phenanthroline 

copper(II/I)  complexes,  for 

instance,  highlighted  the 

importance  of  careful  electrolyte 

composi(cid:415)on.28 By 2016, a ﬂurry of research showcased high-eﬃciency DSSCs using copper redox 

couples,  with  Cong's  and  Freitag's  works  standing  out.  Both  explored  diﬀerent  ligands  and 

prepara(cid:415)on  methods,  yet  achieved  eﬃciencies  hovering  around  9%.29,30  Saygili's  2016  study 

further  propelled  the  ﬁeld,  achieving  over  10%  eﬃciency  with  various  copper  bipyridyl  redox 

couples.31 

The pace of advancement in this domain did not abate, with Hu's 2018 study introducing a stable 

tetradentate copper(II/I) redox couple, which achieved an impressive eﬃciency of 9.2%.32 As of 

the current research landscape, the copper redox mediator copper(II/I) 4,4’,6,6’-tetramethyl-2,2’-

bipyridine, [Cu(tmpy)2]2+/+, and the organic dye 3-{6-{4-[bis(2',4'-dihexyloxybiphenyl-4-yl)amino-

]phenyl}-4,4-dihexyl-cyclopenta-[2,1-b:3,4-b']dithiphene-2-yl}-2-cyanoacrylic 

acid, 

Y123, 

combina(cid:415)on remains at the pinnacle of DSSC performance.20 However, the quest for perfec(cid:415)on 

9 

 
 
con(cid:415)nues, as researchers strive to grasp the intricacies of electron transfer processes within these 

couples, all in a bid to op(cid:415)mize dye regenera(cid:415)on and electron recombina(cid:415)on kine(cid:415)cs. 

Base addi(cid:415)ves in DSSCs, such as pyridines and imidazoles, play a cri(cid:415)cal role in enhancing redox 

mediator performance.33,34 Lewis bases such as 4-tert-butylpyridine (TBP) are used as addi(cid:415)ves 

in DSSCs to increase the performance of the DSSC devices.35–37 This increase is a(cid:425)ributed to a shi(cid:332) 

in  the  (cid:415)tania  conduc(cid:415)on  band  edge  to  a  more  nega(cid:415)ve  poten(cid:415)al  and  a  reduc(cid:415)on  in 

recombina(cid:415)on  from  the  (cid:415)tania  surface  by  adsorbing  to  the  (cid:415)tania.33,34  These  addi(cid:415)ves  can 

undergo an interac(cid:415)on with copper complexes upon oxida(cid:415)on, leading to the forma(cid:415)on of new 

complexes  with  altered  proper(cid:415)es.38  It  is  currently  being  inves(cid:415)gated  as  to  the  extent  of  the 

interac(cid:415)on between the base addi(cid:415)ves and the copper complexes but there are two prevailing 

theories. The ﬁrst is that a ligand subs(cid:415)tu(cid:415)on can  occur, where, upon oxida(cid:415)on  of the  copper 

complex, the base addi(cid:415)ves will displace the ligand bound the copper complex forming a copper-

base complex.38 The second hypothesis is that, upon oxida(cid:415)on, the base  will coordinate to an 

open coordina(cid:415)on site on the copper center forming a copper-ligand-base complex.39 Through 

either path the presence of these addi(cid:415)ves can modify the redox poten(cid:415)al, stability, and kine(cid:415)cs 

of copper-based redox mediators. 

The interac(cid:415)on between base addi(cid:415)ves and copper redox mediators can signiﬁcantly impact DSSC 

performance.  Ligand  subs(cid:415)tu(cid:415)on  reac(cid:415)ons  can  inﬂuence  the  energe(cid:415)cs  of  electron  transfer 

processes,  aﬀec(cid:415)ng  charge 

injec(cid:415)on,  recombina(cid:415)on,  and  overall  cell  eﬃciency.33,34,38 

Understanding these eﬀects is essen(cid:415)al for op(cid:415)mizing DSSC design and opera(cid:415)on. 

Previous  research  has  demonstrated  that  ligand  subs(cid:415)tu(cid:415)on  reac(cid:415)ons  can  either  enhance  or 

hinder  DSSC  performance,  depending  on  the  speciﬁc  copper  complex  and  base  addi(cid:415)ve 

10 

 
 
involved.38 Studying the factors that govern these outcomes provides valuable insights into the 

underlying mechanisms and informs the ra(cid:415)onal design of more eﬃcient redox mediators. 

Developing  a  predic(cid:415)ve  model  for  ligand  subs(cid:415)tu(cid:415)on  reac(cid:415)ons  is  essen(cid:415)al  for  advancing  the 

design and op(cid:415)miza(cid:415)on of copper-based redox mediators in DSSCs. Such a model would enable 

researchers  to  predict  the  outcomes  of  diﬀerent  ligand  subs(cid:415)tu(cid:415)on  scenarios  and  guide  the 

selec(cid:415)on of base addi(cid:415)ves and copper complexes to achieve desired performance enhancements. 

In this thesis, a comprehensive explora(cid:415)on of copper redox mediators in dye-sensi(cid:415)zed solar cells 

(DSSCs)  is  undertaken,  focusing  on  their  interac(cid:415)ons  with  base  addi(cid:415)ves  and  the  consequent 

implica(cid:415)ons for DSSC performance. The intricate interplay between copper complexes and base 

addi(cid:415)ves is thoroughly examined, with the goal of contribu(cid:415)ng to the fundamental understanding 

of ligand subs(cid:415)tu(cid:415)on processes and developing a predic(cid:415)ve model to aid in the ra(cid:415)onal design of 

eﬃcient DSSCs. 

11 

 
 
 
Chapter 2 – Inves(cid:415)ga(cid:415)on of [Cu(dbmed)](OTf)1/2 

2.1 Introduc(cid:415)on 

The  demand  for  carbon-neutral  energy  genera(cid:415)on  has  accelerated  the  development  of 

renewable  technologies,  par(cid:415)cularly  in  the  domain  of  solar  energy.  Within  this  realm,  dye-

sensi(cid:415)zed solar cells (DSSCs) have been iden(cid:415)ﬁed for their cost eﬃciency and photon-to-current 

conversion capabili(cid:415)es across diverse ligh(cid:415)ng condi(cid:415)ons.8 

A standard DSSC consists of a chromophore anchored to a semiconduc(cid:415)ng mesoporous metal 

oxide surface. Upon photoexcita(cid:415)on, an electron is injected into the semiconductor's conduc(cid:415)on 

band. The oxidized chromophore is then regenerated by a redox shu(cid:425)le present in the solu(cid:415)on. 

This electron, a(cid:332)er traversing an external circuit, is eventually received by the counter electrode. 

Here, the oxidized redox shu(cid:425)le is reduced, comple(cid:415)ng the current circuit. 

Historically,  the  iodide/triiodide  (I−/I3−)  redox  shu(cid:425)le  system  was  predominantly  used  but 

presented  challenges  such  as  corrosiveness,  non-tunability,  and  compe(cid:415)(cid:415)ve  absorp(cid:415)on  in  the 

visible spectrum. Addi(cid:415)onally, the two-electron redox process of the I−/I3− system restricted the 

theore(cid:415)cal maximum photovoltage of DSSC devices.22 Co(III/II)-based redox shu(cid:425)les presented 

their own set of limita(cid:415)ons, primarily due to the signiﬁcant inner-sphere reorganiza(cid:415)on energy 

during the electron transfer process, which aﬀected dye regenera(cid:415)on kine(cid:415)cs.23 

Copper-based redox shu(cid:425)les have been explored as an alterna(cid:415)ve due to their reduced inner-

sphere  reorganiza(cid:415)on  energies  for  electron  transfer.  The  geometrical  adaptability  of  copper 

complexes, inﬂuenced by their oxida(cid:415)on state and ligand environment, is of signiﬁcant interest. 

While  many  DSSCs  with  copper  have  primarily  u(cid:415)lized  bidentate  ligands  such  as  bipyridine  or 

phenanthroline, there has been a shi(cid:332) towards tetradentate Cu(II/I) redox shu(cid:425)les, resul(cid:415)ng in 

12 

 
 
improved stability and reduced photovoltage losses.40,41 Innova(cid:415)ons by Freitag and associates, 

where a tetradentate Cu(II/I) redox shu(cid:425)le was integrated into DSSCs, have exhibited enhanced 

stability and reduced photovoltage losses.42 The rigidity of these tetradentate ligands can further 

lower reorganiza(cid:415)on energies, op(cid:415)mizing DSSC performance metrics. 

However,  DSSC  addi(cid:415)ves  like  TBP  and  N-methylbenzimidazole  (NMBI),  known  to  enhance  cell 

performance, have demonstrated interac(cid:415)ons with copper-based redox shu(cid:425)les. Speciﬁcally, TBP 

has been observed to coordinate with Cu(II) species, leading to the displacement of bidentate 

ligands.38 This interac(cid:415)on impacts the redox poten(cid:415)al, electron transfer kine(cid:415)cs, and achievable 

photovoltage.  Given  these  observa(cid:415)ons,  there's  an  ongoing  inves(cid:415)ga(cid:415)on  into  the  poten(cid:415)al  of 

tetradentate ligands to provide stability in the face of these challenges. 

In this chapter, the focus will be on the interac(cid:415)ons and stability of copper-based redox mediators 

in DSSCs, par(cid:415)cularly in the presence of base addi(cid:415)ves, and the poten(cid:415)al advantages oﬀered by 

increased ligand den(cid:415)city. 

2.2 Experimental Details 

All  NMR  spectra  were  taken  on  an  Agilent  DirectDrive2  500  MHz  spectrometer  at  room 

temperature and referenced to residual solvent signals. All NMR spectra were evaluated using 

the  MestReNova  so(cid:332)ware  package  features.  Cyclic  voltammograms  were  obtained  using 

µAutolabIII poten(cid:415)ostat using BASi glassy carbon electrode, a pla(cid:415)num mesh counter electrode, 

and a fabricated 0.01 M AgNO3, 0.1 M TBAPF6 in acetonitrile Ag/AgNO3 reference electrode. All 

measurements  were  internally  referenced  to  ferrocene/ferrocenium  couple  via  addi(cid:415)on  of 

ferrocene  to  solu(cid:415)on  a(cid:332)er  measurements  or  run 

in  a  parallel  solu(cid:415)on  of  the  same 

solvent/electrolyte.  UV-Vis  spectra  were  taken  using  a  PerkinElmer  Lambda  35  UV-Vis 

13 

 
 
spectrometer using a 1 cm path length quartz cuve(cid:425)e at 480 nm/min. Elemental analysis data 

were  obtained  via  Midwest  Microlab.  For  single-crystal  X-ray  diﬀrac(cid:415)on,  single  crystals  were 

mounted on a nylon loop with paratone oil using a Bruker APEX-II CCD diﬀractometer. Crystals 

were maintained at T ¼ 173(2) K during data collec(cid:415)on. Using Olex2, the structures were solved 

with  the  ShelXS  structure  solu(cid:415)on  program  using  the  Direct  Methods  solu(cid:415)on  method. 

Photoelectrochemical  measurements  were  performed  with  a  poten(cid:415)ostat  (Autolab  PGSTAT 

128N) in combina(cid:415)on with a xenon arc lamp. An AM 1.5 solar ﬁlter was used to simulate sunlight 

at 100 mW cm-2, and the light intensity was calibrated with a cer(cid:415)ﬁed reference cell system (Oriel 

Reference Solar Cell & Meter). A black mask with an open area of 0.07 cm-2 was applied on top of 

the cell ac(cid:415)ve area. A monochromator (Horiba Jobin Yvon MicroHR) a(cid:425)ached to the 450 W xenon 

arc light source was used for monochroma(cid:415)c light for IPCE measurements. The photon ﬂux of the 

light  incident  on  the  samples  was  measured  with  a  laser  power  meter  (Nova  II  Ophir).  IPCE 

measurements were made at 20 nm intervals between 400 and 700 nm at short circuit current. 

TEC 15 FTO was cut into 1.5 cm by 2 cm pieces which were sonicated in soapy DI water for 15 

minutes,  followed  by  manual  scrubbing  of  the  FTO  with  Kimwipes.  The  FTO  pieces  were  then 

sonicated in DI water for 10 minutes, rinsed with DI water, and sonicated in isopropanol for 15 

minutes. The FTO pieces were dried under a stream of air then immersed in an aqueous 40 mM 

solu(cid:415)on  TiCl4  solu(cid:415)on  for  30  minutes  at  70  °C.  The  water  used  for  the  TiCl4  treatment  was 

preheated to 70 °C prior to adding 2 M TiCl4 to the water. The 40 mM solu(cid:415)on was immediately 

poured onto the samples and placed in a 70 °C oven for the 30-minute deposi(cid:415)on. The FTO pieces 

were  immediately  rinsed  with  18  MΩ  water  followed  by  isopropanol  and  were  annealed  by 

hea(cid:415)ng from room temperature to 500 °C, holding at 500 °C for 15 minutes. A 0.36 cm2 area was 

14 

 
 
doctor bladed with commercial 30 nm TiO2 nanopar(cid:415)cle paste (DSL 30NRD). The transparent ﬁlms 

were placed in a 100 °C oven for 30 minutes. The samples were annealed in an oven that was 

ramped to 325 °C for 5 minutes, 375 °C for 5 minutes, 450 °C for 5 minutes, and 500 °C for 15 

minutes. The 30 nm nanopar(cid:415)cle ﬁlm thickness was 8.2 μm. A(cid:332)er cooling to room temperature, 

a second TiCl4 treatment was performed as described above. When the anodes had  cooled to 

room temperature, they were soaked in a dye solu(cid:415)on of 0.1 mM Y123 in 1:1 acetonitrile:tert-

butyl alcohol for 18 hours. A(cid:332)er soaking, the anodes were rinsed with acetonitrile and were dried 

gently under a stream of nitrogen. 

The PEDOT counter electrodes were prepared by electrodeposi(cid:415)on in a solu(cid:415)on of 0.01 M EDOT 

and 0.1 M LiClO4 in 0.1 M SDS in 18 MΩ water. A constant current of 8.3 mA for 250 seconds was 

applied to a 54 cm2 piece of TEC 15 FTO with predrilled holes using an equal size piece of FTO as 

the counter electrode. The PEDOT electrodes were then washed with DI water and acetonitrile 

before being dried under a gentle stream of nitrogen and cut into 1.5 cm by 1.0 cm pieces. The 

working and counter electrodes were sandwiched together with 25 μm surlyn ﬁlms by placing 

them on a 140 °C hotplate and applying pressure. The cells were then ﬁlled in a nitrogen ﬁlled 

glove box with electrolyte through one of the two predrilled holes and were sealed with 25 μm 

surlyn backed by a glass coverslip and applied heat to seal with a soldering iron. The electrolyte 

consisted of 0.25 M Cu(I), 0.065 M Cu(II), 0.1 M Li(Counter-ion), in acetonitrile with either no 4-

tert-butylpyridine or 0.5 M 4-tert-butylpyridine added. All batches were made with at least 10 

cells per batch. Contact to the TiO2 electrode was made by soldering a thin layer of indium wire 

onto the FTO. 

Synthesis of N,N(cid:3387)-Dibenzyl-N,N(cid:3387)-bis(6-methylpyridin-2-ylmethyl)ethylenediamine (dbmed): 

15 

 
 
2-(Chloromethyl)-6-methylpyridine 

hydrochloride 

(0.5300g, 

3.0 

mmol) 

and 

benzyltriethylammonium chloride (0.0121 g, 0.05 mmol) were added to 20 mL dichloromethane 

and s(cid:415)rred un(cid:415)l fully dissolved.  N,N’-Dibenzylethylenediamine (0.35 mL, 1.48 mmol) was then 

added  to  the  solu(cid:415)on,  a(cid:332)er  which  a  cloudy  white  solu(cid:415)on  was  formed.  Meanwhile  sodium 

hydroxide  (20.1807  g,  500 mmol)  was  added  to  20  mL  of  water  which was  then  added to  the 

dichloromethane solu(cid:415)on and s(cid:415)rred vigorously. A(cid:332)er 2 hours the organic solu(cid:415)on became clear 

and the aqueous solu(cid:415)on turned white. The solu(cid:415)on was then allowed to reﬂux overnight. A(cid:332)er 

reﬂuxing addi(cid:415)onal water was added to the solu(cid:415)on to allow the layers to fully separate then the 

organic layer was then separated from the aqueous layer. The aqueous layer was then extracted 

with  dichloromethane  two  (cid:415)mes  and  all  the  organic  layers  were  combined,  and  dried  with 

magnesium sulfate. Then the solvent was removed under low vacuum. The crude solid was then 

recrystallized  in  acetonitrile  and  dried  to  obtain  the  product  (0.4270  g,  60.1%  yield).  1H-NMR 

(CDCl3 500 MHz) δ 2.50 (s, 6H), 2.67 (s, 4H), 3.56 (s, 4H), 3.67 (s, 4H), 6.96 (d, 2H), 7.25 (m, 12H), 

7.45 (t, 2H). Matching the data of the compound made by Hu et al.40 

Synthesis of [Cu(dbmed)]OTf: 

Copper(I)  tetrakisacetonitrile  triﬂate  (0.1123  g,  0.30  mmol)  was  dissolved  in  minimal  dry 

acetonitrile.  Meanwhile  dbmed  (0.1487  g,  0.33  mmol)  was  dissolved 

in  minimal  dry 

dichloromethane. The dbmed solu(cid:415)on was added dropwise to the copper solu(cid:415)on, turning the 

solu(cid:415)on a bright yellow color. Reac(cid:415)on was s(cid:415)rred overnight then precipitated with dry ether and 

ﬁltered. The solvent was then removed via vacuum and the Cu(I)dbmed product was collected 

(0.1876 g, 94.9% yield). 1H-NMR (CDCl3 500 MHz) δ 2.67 (s, 2H), 2.72 (s, 6H), 2.82 (s, 2H), 3.57 (s, 

16 

 
 
2H), 3.74 (s, 2H), 3.85 (s, 2H), 4.16 (s, 2H), 7.20 (m, 4H), 7.29 (m, 8H), 7.43 (d, 2H), 7.83 (t, 2H). 

Matching the data of the compound made by Hu et al.40 

Synthesis of Cu(II)dbmed: 

Copper(II)  triﬂate  (0.808  g,  2.23  mmol)  was  dissolved  in  minimal  dry  acetonitrile.  Meanwhile, 

dbmed (0.1030 g, 2.29 mmol) was dissolved in minimal dry dichloromethane. The dbmed solu(cid:415)on 

was  added  dropwise  to  the  copper  solu(cid:415)on,  turning  a  deep  blue  color.  Reac(cid:415)on  was  s(cid:415)rred 

overnight then precipitated with dry ether and ﬁltered. The solvent was then removed via vacuum 

and  the  Cu(II)dbmed  product  was  collected  (0.1596  g,  87.9%  yield).  Matching  the  data  of  the 

compound made by Hu et al.40 

17 

 
 
2.3 Results and Discussion 

Scheme 2.1: Syntheses of the [Cu(dbmed)]OTf1/2 complexes with the Cu(I) species (le(cid:332)) and 
Cu(II) species (right). 

The  ligand,  N,N(cid:3387)-Dibenzyl-N,N(cid:3387)-bis(6-methylpyridin-2-ylmethyl)ethylenediamine  (dbmed),  and 

the copper ligand complexes, [Cu(dbmed)]OTf and [Cu(dbmed)]OTf2 were synthesized following 

a previously reported protocol.32 A crystal was grown of the copper(I) form, the copper center is 

coordinated  to  the  two  amine  and  two  pyridyl  nitrogen’s,  resul(cid:415)ng  in  a  tetrahedral  geometry 

which  is  common  for  copper(I)  complexes.  This  crystal  matched  the  crystal  shown  in  the 

literature.40 A(cid:425)empts at growing a crystal of the copper(II) form did not work out, forming an oil 

instead  of  crystals.  However,  it  has  been  previously  shown  that  the  copper(II)  form  is  in  the 

18 

 
 
 
 
distorted tetrahedral geometry due to strain caused by the methyl groups on the 6 posi(cid:415)on of 

the pyridine.40 It is also noted that the copper(II) complex is air stable, but the copper(I) is not, it 

oxidizes over the course of a few hours in solu(cid:415)on or a few days in the solid state. The dbmed 

ligand was characterized via 1H-NMR and both copper complexes were characterized via 1H-NMR, 

and elemental analysis, matching previous reports, shown in Figures 2.1A, 2.2A, and 2.3A. 

The eﬀect of TBP on the [Cu(dbmed)]2+/+ redox couple was evaluated in the presence of diﬀerent 

concentra(cid:415)ons of the Lewis base. TBP was added in excess of at least 15 equivalents rela(cid:415)ve to 

[Cu(dbmed)](OTf)2 in solu(cid:415)on to match the condi(cid:415)ons used in the devices.  

e
c
n
a
b
r
o
s
b
A

0.6

0.5

0.4

0.3

0.2

0.1

0

450

550

650

750

850
Wavelength / nm

950

1050

Figure 2.3: UV-Vis spectra of 2.14 mM [Cu(dbmed)]OTf2, pale green, with increasing 
concentra(cid:415)ons of TBP added, from 0 mM to 53.5 mM with the darker green being higher 
concentra(cid:415)on of TBP added, compared to a spectra of 2.14 mM [Cu(TBP)4]OTf2 in acetonitrile, 
purple. 

The  eﬀect  of  TBP  was  evaluated  using  UV−vis  spectroscopy,  containing  2  mM  of 

[Cu(dbmed)](OTf)2, where spectra were recorded as a func(cid:415)on of added TBP. As TBP was added 

to a solu(cid:415)on of [Cu(dbmed)]OTf2 and monitored using UV-Vis spectroscopy. It can be seen that 

the absorbance of the peak at ~650 nm decreases in intensity and blueshi(cid:332)s slightly. However 

19 

 
 
 
even  at  25  equivalents  of  TBP  added  to  the  solu(cid:415)on  the  spectra  does  not  match  that  of 

[Cu(TBP)4]OTf2.  

A
μ

/

t

n
e
r
r
u
C

30

20

10

0

-10

-20

-30

-1

-0.8

-0.6

-0.2

-0.4

0
Potential Applied vs Ferrocene / V

0.2

0.4

0.6

0.8

1

Figure 2.4: Cyclic voltammogram of 1.6 mM [Cu(dbmed)]OTf2 in acetonitrile containing 0.1 M 
LiOTf using a glassy carbon working electrode and a pla(cid:415)num mesh counter electrode and a 
scan rate of 100 mV/s. Increasing concentra(cid:415)ons of TBP were added from 0 mM to 38 mM, the 
darker green the higher the TBP concentra(cid:415)on. The results were compared with the cyclic 
voltammogram of 2.0 mM free dbmed ligand in red. 

As a follow up experiment, the addi(cid:415)on of TBP to a solu(cid:415)on of [Cu(dbmed)]OTf2 was monitored 

via cyclic voltammetry. As it can be seen in Figure 2.2, the cathodic peak of [Cu(dbmed)]OTf2, at  

-0.050  V  vs  ferrocene  decreases  in  intensity  and  a  second  cathodic  wave  around  -0.250  V  vs 

ferrocene grows in as TBP is added to the solu(cid:415)on. While the cathodic peak is undergoing a dras(cid:415)c 

change,  the  anodic  peak  does  not  experience  signiﬁcant  change  nor  shi(cid:332)  in  poten(cid:415)al. 

Addi(cid:415)onally, a peak appears to be growing in around 0.600 V vs ferrocene which corresponds to 

the redox poten(cid:415)al of the free ligand. This observa(cid:415)on suggests that dbmed is displaced by TBP 

upon oxida(cid:415)on, forming [Cu(TBP)4]OTf2. However, upon reduc(cid:415)on, the process reverses: TBP is 

displaced  by  the  dbmed  ligand,  reforming  [Cu(dbmed)]OTf.  This  demonstrates  that  the  ligand 

20 

 
 
 
 
 
subs(cid:415)tu(cid:415)on is fully reversible. This indicates that that as TBP is added to the solu(cid:415)on some of the 

previously bound dbmed ligand is being released from copper and is able to be oxidized.  

Figure 2.5: 1H-NMR of 9.85 mM [Cu(dbmed)]OTf2, black, in deuterated acetonitrile with a 
known amount of dichloromethane added. Various concentra(cid:415)ons of TBP were added, 18 mM, 
in grey, 50 mM, in purple, and 100 mM, in green. The spectra were compared to a spectra of 
free dbmed ligand, red. 

To determine the extent of the displacement of dbmed unambiguously and quan(cid:415)ta(cid:415)vely by TBP, 

proton NMR spectra were taken of [Cu(dbmed)]OTf2 in deuterated acetonitrile without and with 

the  addi(cid:415)on  of TBP.  Because  [Cu(dbmed)]2+  is  paramagne(cid:415)c,  the  proton  signals  of  the  dbmed 

ligands coordinated to the copper(II) center show very diﬀerent chemical shi(cid:332)s than free dbmed 

ligands  in  the  solu(cid:415)on.  The  sample  was  measured  with  up  to  100  equivalents  of  TBP  added, 

however  as  the  concentra(cid:415)on  of  TBP  was  added  to  the  solu(cid:415)on  there  were  issues  with  the 

21 

 
 
shimming due to the high intensity of the methyl group on the TBP, as a result the data shown is 

limited to 50 equivalents of TBP added and lower. As TBP is added to the solu(cid:415)on, four new peaks 

appear  in  the  NMR  spectra  between  2.4  and  3.8  ppm.  These  peaks  match  with  the  peaks  of 

unbound dbmed ligand. By u(cid:415)lizing an internal standard of a known amount of DCM, the amount 

of free ligand in solu(cid:415)on can be quan(cid:415)ta(cid:415)vely determined.  Using this method, it was found that 

the [Cu(dbmed)](OTf)2 had not fully undergone ligand subs(cid:415)tu(cid:415)on by the (cid:415)me 50 equivalents of 

TBP was added to the solu(cid:415)on. 

2
-

m
c
A
m

/

t
n
e
r
r
u
C

10

9

8

7

6

5

4

3

2

1

0

-1

-0.8

-0.6

-0.4

-0.2

0

Voltage / V

Figure 2.6: J−V curves for the [Cu(dbmed)]2+/+ DSSC devices, with TBP, dark green, and without 
TBP light green. 

Table 2.1: A summary of DSSC device performance metrics for those devices. 

η (%) 

Jsc (mA*cm-
2) 

Voc (V) 

FF 

Cu(dmbpy) 

0.85(±0.35)  2.19(±0.77)  0.58(±0.02) 

0.66(±0.05) 

Cu(dmbpy) 
with TBP 

4.32(±0.64)  7.19(±1.24)  0.89(±0.02) 

0.67(±0.05) 

22 

 
 
 
 
 
 
 
DSSC devices were fabricated  using [Cu(dbmed)]2+/+ as the redox shu(cid:425)le in conjunc(cid:415)on with  a 

poly(3,4-ethylenedioxythiophene)  (PEDOT)  counter  electrode,  and  the  commercially  available 

dye Y123 was used as the light-harves(cid:415)ng component. The devices were constructed as described 

previously. The devices with TBP added to the solu(cid:415)on yielded a respectable performance with a 

JSC of 7.19(±1.24) mAcm-2 and an overall PCE of 4.32(±0.64) as shown in Table 2.1. However, in 

the  absence  of  TBP  the  performance  was  signiﬁcantly  worse,  with  a  JSC  of  only  2.19(±0.77)    

mAcm-2 and an overall PCE of 0.85(±0.35). This indicates that TBP is s(cid:415)ll performing a pivotal role 

in increasing the current of the devices.  

Table 2.2: Comparison of the solu(cid:415)on poten(cid:415)al and the predicted Nerns(cid:415)an poten(cid:415)al of the 
electrolyte solu(cid:415)on used in each DSSC device set. 

E1/2 of 
[Cu(dbmed)]OTf1/2 
/ V 
-0.005 

Nernst 
equa(cid:415)on 
/ V 
-0.023 

Solu(cid:415)on 
poten(cid:415)al 
/V 
-0.053 

Ef Level / 
V 

-0.633 

-0.005 

-0.023 

-0.260 

-1.150 

dbmed 
dbmed with 
TBP 

23 

 
 
 
10

1

0.1

)
s
/
1
(
e
m

i
t
e
f
i
L

0.01

0

[Cu(dbmed)] without TBP

[Cu(dbmed)] with TBP

0.2

0.4

0.6

0.8

1

1.2

Voltage (V)

Figure 2.7: A plot of the electron life(cid:415)me vs poten(cid:415)al plot for DSSC devices using 
[Cu(dbmed)]OTf1/2 (green), with (pale color) and without (dark color) the addi(cid:415)on of TBP. 

By inves(cid:415)ga(cid:415)ng the solu(cid:415)on poten(cid:415)al of electrolyte composi(cid:415)ons, it is observed that the solu(cid:415)on 

poten(cid:415)al becomes about 200 mV more nega(cid:415)ve with the addi(cid:415)on of TBP to the electrolyte. Even 

though there was a nega(cid:415)ve shi(cid:332) in solu(cid:415)on poten(cid:415)al, there was a 300 mV increase in the open 

circuit voltage, indica(cid:415)ng that the fermi level of the semiconductor increased by around 500 mV 

with  the  introduc(cid:415)on  of  TBP  to  the  devices.  This  is  a  further  indicator  of  ligand  subs(cid:415)tu(cid:415)on 

occurring  and  forming  a  more  nega(cid:415)ve  redox  couple  in  solu(cid:415)on.  Furthermore,  the  electron 

life(cid:415)me  also  signiﬁcantly  increases  with  the  addi(cid:415)on  of  TBP,  to  the  extent  that  it  is  hard  to 

compare the results since there are no overlapping measured poten(cid:415)als that the (cid:415)mescale of the 

experiment.  This  dras(cid:415)c  increase  in  poten(cid:415)al  and  electron  life(cid:415)me  can  be  a(cid:425)ributed  to  the 

addi(cid:415)on of TBP, more speciﬁcally both the proper(cid:415)es of Lewis bases in DSSC electrolyte (shi(cid:332)ing 

24 

 
 
 
 
the conduc(cid:415)on band poten(cid:415)al and reducing recombina(cid:415)on) and ligand exchange occurring and 

forming a [Cu(TBP)4)]OTf1/2 complex which has very poor electron kine(cid:415)cs.33,34 

2.4 Conclusion 

In this study, the [Cu(dbmed)]OTf2 complex was found to be suscep(cid:415)ble to ligand subs(cid:415)tu(cid:415)on in 

the  presence  of  TBP,  without  measurable  full  ligand  subs(cid:415)tu(cid:415)on  occurring  when  up  to  50 

equivalents of TBP were added to the solu(cid:415)on. Interes(cid:415)ngly, this ligand subs(cid:415)tu(cid:415)on was observed 

to be reversible upon reduc(cid:415)on of the complex. Despite the occurrence of ligand exchange, the 

addi(cid:415)on  of  TBP  notably  enhanced  the  eﬃciency  of  the  DSSC  devices.  Allowing  for  dbmed  to 

con(cid:415)nue to be a viable ligand for copper redox shu(cid:425)les in DSSCs in the future. The underlying 

cause of this ligand subs(cid:415)tu(cid:415)on was determined to be the steric strain imposed by the methyl 

groups  on  the  pyridines  of  the  complex,  which  eﬀec(cid:415)vely  outcompetes  the  thermodynamic 

stability usually conferred by the increased den(cid:415)city of the ligand. 

25 

 
 
 
 
Chapter 3 – Inves(cid:415)ga(cid:415)on of [Cu(bpyPY4)](TFSI)1/2 

3.1 Introduc(cid:415)on 

The  necessity  for  carbon-neutral  energy  genera(cid:415)on  has  driven  the  advancement  of  novel 

technologies that exploit renewable sources, notably wind and solar energy.43,44 By harnessing 

sunlight for solar-to-electrical energy conversion, photovoltaic cells oﬀer a promising solu(cid:415)on.45–

47  A  special  subset  of  these,  dye-sensi(cid:415)zed  solar  cells  (DSSCs),  are  recognized  for  their  cost-

eﬀec(cid:415)ve  nature  and  their  impressive  photon-to-current  conversion  eﬃciencies  under  diverse 

ligh(cid:415)ng condi(cid:415)ons.20,48–52 A typical DSSC consists of a chromophore anchored to a semiconduc(cid:415)ng 

mesoporous metal oxide surface. When subjected to photoexcita(cid:415)on, an electron is injected into 

the  semiconductor's  conduc(cid:415)on  band.  Subsequently,  the  oxidized  chromophore  (or  dye)  is 

regenerated by a redox shu(cid:425)le present in the solu(cid:415)on. This injected electron, having traversed an 

external  circuit,  reaches  the  counter  electrode.  There,  the  oxidized  redox  shu(cid:425)le  undergoes 

reduc(cid:415)on, comple(cid:415)ng the circuit.20,47,53 

However, the tradi(cid:415)onally employed iodide/triiodide (I−/I3−) redox shu(cid:425)le system exhibits several 

drawbacks, including corrosiveness, lack of tunability, and compe(cid:415)(cid:415)ve absorp(cid:415)on in the visible 

spectrum by the triiodide ion.54–57 Notably, the I−/I3− redox couple's two-electron redox process 

o(cid:332)en leads to reduced theore(cid:415)cal maximum photovoltage output from DSSC devices, especially 

when compared to one-electron redox processes.57,58 Co(III/II)-based redox shu(cid:425)les, on the other 

hand, are accompanied by a signiﬁcant inner-sphere reorganiza(cid:415)on energy during the electron 

transfer process.59,60 This energy hinders dye regenera(cid:415)on kine(cid:415)cs due to its limited availability 

for electron transfer. Copper-based redox shu(cid:425)les, however, have demonstrated parity with the 

classical  I−/I3−  and  Co(III/II)-based  redox  shu(cid:425)les.  Their  success  is  o(cid:332)en  a(cid:425)ributed  to  the 

26 

 
 
minimized  inner-sphere  electron  transfer  reorganiza(cid:415)on  energies.  The  ability  of  copper 

complexes to adopt various geometries based on their metal oxida(cid:415)on state and ligand se(cid:427)ng is 

noteworthy.61–63 The majority of copper redox shu(cid:425)les employed in DSSCs thus far have relied on 

bidentate bipyridine or phenanthroline-based ligands.64 An approach which u(cid:415)lizes bulky groups 

adjacent to nitrogen donors to diminish structural shi(cid:332)s between Cu(II) and Cu(I) redox states. 

Recent  reports  also  suggest  that  rigid  tetradentate  ligands  can  further  op(cid:415)mize  Cu(II/I)  redox 

shu(cid:425)les  by  promo(cid:415)ng  even  lower  inner-sphere  reorganiza(cid:415)on  energies,  thus  resul(cid:415)ng  in 

heightened short-circuit current densi(cid:415)es and reduced VOC losses.65 

Lewis  bases,  such  as  TBP  and  N-methylbenzimidazole  (NMBI),  are  frequently  incorporated 

addi(cid:415)ves in DSSC devices.35–37 These compounds have been shown to improve the overall  cell 

performance.  Such  enhancements  are  o(cid:332)en  associated  with  the  ability  of  these  bases  to 

nega(cid:415)vely  shi(cid:332)  the  (cid:415)tania  conduc(cid:415)on  band  edge  and  suppress  the  interfacial  charge 

recombina(cid:415)on rates through surface adsorp(cid:415)on.36,66–68 Notably, when most copper-based redox 

shu(cid:425)les  are  inves(cid:415)gated,  TBP  has  been  observed  to  coordinate  with  various  Cu(II)  species, 

poten(cid:415)ally  displacing  polydentate  ligands.38,39,41,58,69–72  Such  interac(cid:415)ons  signiﬁcantly  inﬂuence 

the  redox  poten(cid:415)al  by  genera(cid:415)ng  mul(cid:415)ple  redox  species,  which  in  turn  aﬀects  the  maximum 

achievable photovoltage and electron transfer kine(cid:415)cs. The focal point of this research revolves 

around exploring the interac(cid:415)ons between a copper complex with a hexadentate ligand and TBP, 

with a primary goal of understanding if ligand subs(cid:415)tu(cid:415)on occurs upon oxida(cid:415)on. 

3.2 Experimental Details 

All  NMR  spectra  were  taken  on  an  Agilent  DirectDrive2  500  MHz  spectrometer  at  room 

temperature and referenced to residual solvent signals. All NMR spectra were evaluated using 

27 

 
 
the  MestReNova  so(cid:332)ware  package  features.  Cyclic  voltammograms  were  obtained  using 

µAutolabIII poten(cid:415)ostat using BASi glassy carbon electrode, a pla(cid:415)num mesh counter electrode, 

and a fabricated 0.01 M AgNO3, 0.1 M TBAPF6 in acetonitrile Ag/AgNO3 reference electrode. All 

measurements  were  internally  referenced  to  ferrocene/ferrocenium  couple  via  addi(cid:415)on  of 

ferrocene  to  solu(cid:415)on  a(cid:332)er  measurements  or  run 

in  a  parallel  solu(cid:415)on  of  the  same 

solvent/electrolyte.  UV-Vis  spectra  were  taken  using  a  PerkinElmer  Lambda  35  UV-Vis 

spectrometer using a 1 cm path length quartz cuve(cid:425)e at 480 nm/min. Elemental Analysis data 

were  obtained  via  Midwest  Microlab.  For  single-crystal  X-ray  diﬀrac(cid:415)on,  single  crystals  were 

mounted on a nylon loop with paratone oil using a Bruker APEX-II CCD diﬀractometer. Crystals 

were maintained at T ¼ 173(2) K during data collec(cid:415)on. Using Olex2, the structures were solved 

with  the  ShelXS  structure  solu(cid:415)on  program  using  the  Direct  Methods  solu(cid:415)on  method. 

Photoelectrochemical  measurements  were  performed  with  a  poten(cid:415)ostat  (Autolab  PGSTAT 

128N) in combina(cid:415)on with a xenon arc lamp. An AM 1.5 solar ﬁlter was used to simulate sunlight 

at 100 mW cm-2, and the light intensity was calibrated with a cer(cid:415)ﬁed reference cell system (Oriel 

Reference Solar Cell & Meter). A black mask with an open area of 0.07 cm-2 was applied on top of 

the cell ac(cid:415)ve area. A monochromator (Horiba Jobin Yvon MicroHR) a(cid:425)ached to the 450 W xenon 

arc light source was used for monochroma(cid:415)c light for IPCE measurements. The photon ﬂux of the 

light  incident  on  the  samples  was  measured  with  a  laser  power  meter  (Nova  II  Ophir).  IPCE 

measurements were made at 20 nm intervals between 400 and 700 nm at short circuit current. 

TEC 15 FTO was cut into 1.5 cm by 2 cm pieces which were sonicated in soapy DI water for 15 

minutes,  followed  by  manual  scrubbing  of  the  FTO  with  Kimwipes.  The  FTO  pieces  were  then 

sonicated in DI water for 10 minutes, rinsed with acetone, and sonicated in isopropanol for 10 

28 

 
 
minutes. The FTO pieces were dried in room temperature air then immersed in an aqueous 40 

mM TiCl4 solu(cid:415)on for 60 minutes at 70 °C. The water for the TiCl4 treatment was preheated to 70 

°C prior to adding 2 M TiCl4 to the water. The 40 mM solu(cid:415)on was immediately poured onto the 

samples  and  placed  in  a  70  °C  oven  for  the  60-minute  deposi(cid:415)on.  The  FTO  pieces  were 

immediately rinsed with  18 MΩ  water followed by isopropanol and were  annealed by hea(cid:415)ng 

from room temperature to 500 °C, holding at 500 °C for 30 minutes. A 0.36 cm2 area was doctor 

bladed with commercial 30 nm TiO2 nanopar(cid:415)cle paste (DSL 30NRD). The transparent ﬁlms were 

le(cid:332) to rest for 10 minutes and were then placed in a 125 °C oven for 30 minutes. The samples 

were annealed in an oven that was ramped to 325 °C for 5 minutes, 375 °C for 5 minutes, 450 °C 

for  5  minutes,  and  500  °C  for  15  minutes.  The  30  nm  nanopar(cid:415)cle  ﬁlm  thickness  was 

approximately 8 μm. A(cid:332)er cooling to room temperature, a second TiCl4 treatment was performed 

as described above. When the anodes had cooled to 80 °C, they were soaked in a dye solu(cid:415)on of 

0.1 mM Y123 in 1 : 1 acetonitrile : tert-butyl alcohol for 18 hours. A(cid:332)er soaking, the anodes were 

rinsed with acetonitrile and were dried gently under a stream of nitrogen. 

The PEDOT counter electrodes were prepared by electropolymeriza(cid:415)on in a solu(cid:415)on of 0.01 M 

EDOT and 0.1 M LiClO4 in 0.1 M SDS in 18 MΩ water. A constant current of 8.3 mA for 250 seconds 

was applied to a 54 cm2 piece of TEC 8 FTO with predrilled holes using an equal size piece of FTO 

as the counter electrode. The PEDOT electrodes were then washed with DI water and acetonitrile 

before being dried under a gentle stream of nitrogen and cut into 1.5 cm by 1.0 cm pieces. The 

working and counter electrodes were sandwiched together with 25 μm surlyn ﬁlms by placing 

them on a 140 °C hotplate and applying pressure. The cells were then ﬁlled in a nitrogen ﬁlled 

glove box with electrolyte through one of the two predrilled holes and were sealed with 25 μm 

29 

 
 
surlyn backed by a glass coverslip and applied heat to seal with a soldering iron. The electrolyte 

consisted of 0.10 M Cu(I), 0.05 M Cu(II), 0.1 M Li(Counter-ion), and 0.5 M 4-tert-butylpyridine in 

acetonitrile. All batches were made with at least 10 cells per batch. Contact to the TiO2 electrode 

was made by soldering a thin layer of indium wire onto the FTO. 

Prepara(cid:415)on  of 

copper(I) 

6,6(cid:3387)-bis(1,1-di(pyridine-2-yl)ethyl)-2,2(cid:3387)-bipyridine  Bistriﬂimide 

([Cu(bpyPY4)]TFSI: 

Copper(I)  tetrakisacetonitrile  bistriﬂimide  (0.1647  g,  0.43  mmol)  was  dissolved  in  minimal  dry 

acetonitrile. Meanwhile 6,6(cid:3387)-bis(1,1-di(pyridin-2-yl)ethyl)-2,2(cid:3387)-bipyridine (bpyPY4) (0.2545 g, 0.49 

mmol) was dissolved in minimal dry dichloromethane. The bpyPY4 solu(cid:415)on was added dropwise 

to the copper solu(cid:415)on, turning the solu(cid:415)on a dark red color. Reac(cid:415)on was s(cid:415)rred overnight then 

precipitated  with  dry  ether  and  ﬁltered.  The  solvent  was  then  removed  via  vacuum  and  the 

Cu(I)bpyPY4 product was collected (0.2829 g, 76.1% yield).  1H-NMR (CDCl3 500 MHz) δ 2.25 (s, 

6H),  7.219  (s,  4H),  7.76  (s,  4H),  8.10  (d,  2H),  8.17  (t,  2H),  8.35  (d,  2H)  Elem.  Anal.  calc.  for 

C36H28N7F7O4S2Cu: C, 50.03; H, 3.27; N, 11.34. Found: C, 50.05; H, 3.73; N, 11.08. 

Copper(II)  bistriﬂimide  (0.4215  g,  1.16  mmol)  was  dissolved  in  minimal  dry  acetonitrile. 

Meanwhile BPYPY4 (0.6942 g, 1.33 mmol) was dissolved in minimal dry dichloromethane. The 

BPYPY4  solu(cid:415)on  was  added dropwise  to  the  copper  solu(cid:415)on,  turning  the  solu(cid:415)on  a pale  blue 

color. Reac(cid:415)on was s(cid:415)rred overnight then precipitated with dry ether and ﬁltered. The solvent 

was  then  removed  via  vacuum  and  the  Cu(II)bpyPY4  product  was  collected  (0.7912  g,  63.9% 

yield). Elem. Anal. calc. for C38H28N8F12O8S4Cu + ether: C, 41.40; H, 3.14; N, 9.20. Found: C, 41.48; 

H, 3.10; N, 9.13. 

30 

 
 
All  the  electrochemical  measurements  were  performed  by  using  μAutoLabII  poten(cid:415)ostat  or 

Autolab  PGSTAT128N  in  a  home-made  electrochemical  cell  under  N2  atmosphere.  Copper 

complexes were dissolved in dry acetonitrile, the speciﬁc concentra(cid:415)ons will be speciﬁed in the 

result  and  discussion  sec(cid:415)on.  Also dissolved  in  the  solu(cid:415)on  was  0.1  M  suppor(cid:415)ng  electrolyte, 

which will be speciﬁed in the main text. 

For the cyclic voltammetry measurements, glassy carbon disk was used as the working electrode. 

The reference electrode was a home-made 0.1 M Ag/AgNO3 reference electrode. A salt bridge 

was used to prevent the leakage of the silver containing electrolyte in the reference electrode. 

The CV of ferrocene was measured in the same solvent before and a(cid:332)er measuring the sample. 

The  redox  poten(cid:415)al of  ferrocene/ferrocenium  was  used  to  correct  the  poten(cid:415)al  of  the  home-

made reference electrode. The counter electrode for CV measurement was a pla(cid:415)num mesh. Scan 

rate was varied and will be speciﬁed in the main text. For the (cid:415)tra(cid:415)on measurement, diﬀerent 

concentra(cid:415)ons of TBP were added into the solu(cid:415)on containing copper complexes and suppor(cid:415)ng 

electrolyte. The solu(cid:415)on was s(cid:415)rred and then stabilized before measuring.  

The absorp(cid:415)on spectra were measured with Lambda 35 (PerkinElmer) spectrometer. Solu(cid:415)ons 

containing copper complexes were prepared in the glovebox under an inert atmosphere with dry 

acetonitrile.  Screw-cap  quartz  cuve(cid:425)e  with  1  cm  path  length  was  used  for  all  the  absorp(cid:415)on 

measurements in solu(cid:415)on. Blank acetonitrile was used as the reference for the absorp(cid:415)on. For 

the  (cid:415)tra(cid:415)on  measurement,  a  stock  solu(cid:415)on  containing  a  constant  concentra(cid:415)on  of  copper 

complex was prepared in the glove box. Each sample contains 4 mL of the stock solu(cid:415)on and a 

stoichiometric amount of TBP. The absorp(cid:415)on spectra of diﬀerent samples were measured. 

31 

 
 
Proton  NMR  of  copper  complexes  were  measured  using  a  J-Young  NMR  tube  under  an  inert 

atmosphere  with  Agilent  DDR2  500  MHz  NMR  spectrometer  at  room  temperature.  For  the 

(cid:415)tra(cid:415)on measurement, 0.5 mL solu(cid:415)on containing a constant concentra(cid:415)on of copper complex 

was prepared in the glove box with deuterated acetonitrile. Dichloromethane was used as a non-

coordina(cid:415)ng reference proton source to calibrate the proton peak area. A(cid:332)er running NMR of the 

pure copper complex, the tube was brought back into the glove box and another aliquot of TBP 

was added into the same sample. For each run, the parameters, number of scans, temperature, 

decay (cid:415)me, and scan range, used for the NMR measurement were kept the same. 

3.3 Results and Discussion 

The  synthesis  of  the  ligand,  bpyPY4,  has  been  previously  reported,  wherein  1,1-  bis(2-

pyridyl)ethane underwent deprotona(cid:415)on with n-BuLi before the addi(cid:415)on of the electrophile, 6,6(cid:3387)-

dibromo-2,2(cid:3387)-bipyridine. The metala(cid:415)on was done u(cid:415)lizing two diﬀerent counter ions, triﬂate and 

bistriﬂimide,  in  an  a(cid:425)empt  to  overcome  solubility  limita(cid:415)ons.  The  copper(I)  triﬂate  version 

exhibits  a  solubility  of  less  than  0.1  M.  In  contrast,  the  solubility  of  the  copper(I)  bistriﬂimide 

version exceeds 1.25 M, leading to the use of bistriﬂimide versions for all studies on this copper 

complex system. As depicted in Scheme 3.1, the metala(cid:415)on is carried out using an equimolar ra(cid:415)o 

of  the  appropriate  copper  salt,  either  [Cu(acetonitrile)4](TFSI)  or  Cu(TFSI)2  where  TFSI  = 

Bistriﬂimide,  with  bpyPY4  in  acetonitrile  to  aﬀord  [Cu(bpyPY4)](TFSI)  or  [Cu(bpyPY4)](TFSI)2, 

respec(cid:415)vely. The complexes were puriﬁed  by slow diﬀusion of diethyl ether into concentrated 

acetonitrile solu(cid:415)ons, and  their characteriza(cid:415)on was accomplished  with  1H-NMR spectroscopy 

and elemental analysis. 

32 

 
 
Scheme 3.1: Syntheses of the [Cu(bpyPY4)]TFSI1/2 complexes and crystal structures of the 
ca(cid:415)ons of the Cu(I) species (le(cid:332)) and Cu(II) species (right). Depicted ellipsoids are at the 50% 
probability level. The anions and hydrogens were omi(cid:425)ed for clarity. 

Single  crystal  X-ray  diﬀrac(cid:415)on  ascertained  the  solid-state  structures  of  the  copper  complexes, 

with details documented in Tables 3.1A and 3.2A. It was observed that the hexadentate ligand 

formed  a  ﬁve-coordinate  geometry  around  the  copper(II)  center,  binding  ﬁve  out  of  the  six 

pyridine  donors  of  the  bpyPY4.  Alterna(cid:415)vely,  the  copper(I)  center  revealed  a  four-coordinate 

geometry with two pyridine donors from bpyPY4 remaining unbound. Geometric indices, τ5 and 

τ4, are used for ﬁve-coordinate and four-coordinate complexes, respec(cid:415)vely, to determine how 

much distor(cid:415)on there is between ideal geometries.73,74 The τ value in each case will have a range 

from 0 to 1. For ﬁve-coordinate complexes, a τ5 value of 0 represents an ideal square pyramidal 

geometry,  and  a  value  of  1  represents  an  ideal  trigonal  bipyramidal  geometry.  For  the  ﬁve-

coordinate  copper(II)  complex,  the  τ5  value  was  0.61,  indica(cid:415)ve  of  a  distorted  trigonal 

bipyramidal geometry. Within this conﬁgura(cid:415)on, the bipyridine fragment of the bpyPY4 ligand 

33 

 
 
 
posi(cid:415)ons itself both axially and equatorially around the copper(II) ion. Notably, the copper(II) ion 

exhibited a displacement distance of 0.225 Å from the plane deﬁned by the equatorial donors, 

poin(cid:415)ng  towards  the  axial  pyridine  donor.  As  a  result,  the  angles  involving  the  axial  pyridine 

nitrogen donor, the copper(II) ion, and the equatorial donors slightly exceed 90°. Moreover, the 

bipyridine fragment showcases a near-planar orienta(cid:415)on, with its pyridine rings forming a torsion 

angle of 5.0(4)°. 

In the case of the copper(I) complex, a four-coordinate complex, an ideal square planar geometry  

is represented by a τ4 value of 0, and an ideal tetrahedral geometry is represented by a τ4 value 

of  1.74  The  copper(I)  counterpart  showcased  a  τ4  value  of  0.63,  deno(cid:415)ng  a  highly  distorted 

tetrahedral geometry. In terms of coordina(cid:415)on, the bpyPY4 ligand coordinates the copper(I) ion 

in a manner where the bipyridine unit occupies two of the coordina(cid:415)on sites, and the remaining 

two sites have a pyridine from each of the 1,1-bis(2-pyridyl)ethane fragments bound. The average 

bond distance between the copper-bipyridine nitrogen stands at 2.082(3) Å, aligning with data 

from other tetrahedral copper(I) complexes with bipyridine ligands.75,76 Yet, the copper-pyridine 

nitrogen bond is slightly shorter, averaging 2.015(2) Å. The bipyridyl segment, in this instance, 

demonstrates  a  pronounced  distor(cid:415)on,  marked  by  a  torsion  angle  of  -18.0(4)°  between  its 

pyridine rings.  

The  1H-NMR  spectrum  of  [Cu(bpyPY4)]TFSI,  recorded  in  deuterated  acetonitrile  at  room 

temperature, integrates to 20 protons. Two broad peaks of equal intensity at δ = 7.17 and 7.73 

ppm, integra(cid:415)ng only to four protons each, combine for eight protons in total. Sharp resonances 

at δ = 810, 8.17, and 8.35 ppm, comprising two doublets and a triplet further downﬁeld in the 

aroma(cid:415)c spectrum, account for six protons, two protons per peak, and align with the bipyridine 

34 

 
 
unit. A sharp singlet at 2.25 ppm accounts for the remaining six protons, represen(cid:415)ng the two 

methyl  peaks.  However,  the bpyPY4  framework has 28  protons.  Intriguingly,  while  the  bpyPY4 

framework  boasts  28  aroma(cid:415)c  protons,  the  spectrum  under-represents  this  number  by  eight 

protons. Insights from the solid-state structure of the copper(I) complex, depicted in Scheme 3.1, 

reveal two non-coordina(cid:415)ng pyridines, which could account for the eight proton discrepancy. The 

observable  broad  peaks  at  7.17  ppm  and  7.73  ppm  suggest  a  rapid  exchange  on  the  NMR 

(cid:415)mescale between non-coordinated and coordinated pyridine donors.  

Figure 3.1: 1H-NMR spectrum (500 MHz, CD3CN) of [Cu(bpyPY4)](TFSI). 

To  delve  deeper  into  this  phenomenon,  variable-temperature  1H-NMR  spectroscopy  was 

executed, with spectra obtained from −40 to 25 °C consolidated in Figure 3.2. This allows for an 

interroga(cid:415)on  of  the  solu(cid:415)on's  dynamic  coordina(cid:415)on  behavior.  As  the  temperature  decreased, 

35 

 
 
 
new broad peaks emerged by 0°C, yielding eight pronounced peaks by −30°C. Integra(cid:415)ng to 16 

protons,  these  peaks  originate  from  the  four  terminal  pyridines,  illustra(cid:415)ng  two  dis(cid:415)nc(cid:415)ve 

chemical  environments.  The  presump(cid:415)on  is  that  two  of  these  ﬂuctua(cid:415)ng  pyridines,  one  from 

each  dipyridylethane  “arm”,  remain  uncoordinated,  each  possessing  four  dis(cid:415)nct  aroma(cid:415)c 

protons.  In  contrast,  the  other  two  are  coordinated,  accoun(cid:415)ng  for  the  remaining  four  peaks. 

Notably, chemical shi(cid:332)s for non-coordinated pyridines diverge slightly from the free ligand, but 

such a downﬁeld shi(cid:332) has been documented in prior studies for dissociated heterocycles within 

a coordinated ligand matrix.77,78 

Figure 3.2: Variable temperature 1H-NMR spectra of [Cu(bpyPY4)]TFSI in CD3CN. 

In  anhydrous  acetonitrile,  the  UV-visible  spectra  of  the  copper  complexes  reveal  dis(cid:415)nct 

absorp(cid:415)on characteris(cid:415)cs. The [Cu(bpyPY4)]TFSI complex shows absorp(cid:415)on peaks of moderate 

intensity spanning from 249 to 324 nm, with molar absorp(cid:415)vi(cid:415)es ranging from 51,500 to 33,400 

M−1cm−1. The copper(I) complex also displays lower intensity shoulders at 329, 391, and 450 nm, 

characterized by molar absorp(cid:415)vi(cid:415)es of 10000, 6600, and 4100 M−1cm−1, respec(cid:415)vely. A subtle 

shoulder  peak  at  538  nm  is  observed,  signifying  a  metal-to-ligand  charge  transfer  with  an 

ex(cid:415)nc(cid:415)on  coeﬃcient  of  2500  M−1cm−1.  The  copper(II)  complex  demonstrates  d−d  transi(cid:415)on 

bands at 616 and 845 nm, each with ex(cid:415)nc(cid:415)on coeﬃcients of 270 and 194 M−1cm−1, respec(cid:415)vely. 

36 

 
 
 
These  characteris(cid:415)cs  are  emblema(cid:415)c  of  a  structure  predisposed  to  a  trigonal  bipyramidal 

geometry in solu(cid:415)on. 

Figure 3.3 UV−visible spectra of the[Cu(bpyPY4)]TFSI complex (yellow) and the 
[Cu(bpyPY4)]TFSI2 complex (blue) in anhydrous CH3CN. 

The cyclic voltammogram of the copper(II) complex revealed a copper(II/I) redox couple at −0.454 

V versus the ferrocenium/ferrocene couple as shown in Figure 3.4, with a peak spli(cid:427)ng value of 

93  mV,  indica(cid:415)ng  that  it  is  quasi-reversible.  This  redox  poten(cid:415)al  is  found  to  be  surprisingly 

nega(cid:415)ve rela(cid:415)ve to most previously reported copper(II/I) redox couples. 79–84 

37 

 
 
 
A
µ
/

t
n
e
r
r
u
C

15.00

10.00

5.00

0.00

-5.00

-10.00

-15.00

-1.2

-1

-0.8

-0.6

-0.4
Potential vs Ferrocene / V

-0.2

0

0.2

0.4

Figure 3.4: A cyclic voltammogram of 2.00 mM [Cu(bpyPY4)](TFSI)2 in anhydrous CH3CN with 0.1 
M TBAPF6 as a suppor(cid:415)ng electrolyte, measured at a scan rate of 0.1 V/s on a glassy carbon 
working electrode. 

The interac(cid:415)on of TBP with the [Cu(bpyPY4)]2+/+ redox couple was probed via cyclic voltammetry 

at  varying  concentra(cid:415)ons  of  the  Lewis  base,  with  TBP  present  in  an  excess  of  at  least  10 

equivalents to align with device condi(cid:415)ons delineated later. The addi(cid:415)on of TBP to the copper 

complex shows no signiﬁcant inﬂuence on either the cathodic or anodic wave of the copper(II/I) 

redox couple, indica(cid:415)ng there is no reac(cid:415)on between the [Cu(bpyPY4)]2+ complex and TBP. This 

observa(cid:415)on  was  further  substan(cid:415)ated  by  UV-Vis  spectroscopic  analysis,  showing  that,  despite 

increasing TBP concentra(cid:415)ons, the UV-visible absorp(cid:415)on spectra remained consistent a(cid:332)er being 

corrected for dilu(cid:415)on. This reinforces the ﬁnding that TBP remains unbound to the metal center 

of the redox shu(cid:425)le. If there was subs(cid:415)tu(cid:415)on occurring the peak at 845 nm would show a rapid 

decrease in intensity due to a lack of contribu(cid:415)on from the [Cu(TBP)4]2+ complex. Further insight 

was sought using 1H-NMR spectroscopy. Titra(cid:415)on of the [Cu(bpyPY4)]2+ sample with TBP unveiled 

38 

 
 
 
 
 
no  signals  corresponding  to  free  bpyPY4,  sugges(cid:415)ng  an  absence  of  ligand  subs(cid:415)tu(cid:415)on.  This 

behavior is a(cid:425)ributed to the transient, non-coordinated pyridines within the hexadentate ligand, 

which likely occlude any open coordina(cid:415)on sites on the metal, precluding the detrimental binding 

of external Lewis bases.  

Figure 3.5: A) Cyclic Voltammogram of 2 mM [Cu(bpyPY4)](TFSI)2 with 0.1 M TBAPF6 in 
anhydrous CH3CN with increasing concentra(cid:415)on of TBP at a glassy carbon working electrode; 
scan rate = 100 mVs-1. B) UV-Vis absorp(cid:415)on spectra of 2 mM [Cu(bpyPY4)](TFSI)2 in anhydrous 
CH3CN with increasing equivalents of TBP. A spectra of 2 mM [Cu(TBP)4]TFSI is shown as a 
control. C) 1H-NMR spectra of [Cu(bpyPY4)](TFSI)2 in CH3CN with increasing equivalents of TBP. 
Unbound bpyPY4 and TBP are shown as controls. 

Stopped-ﬂow  spectroscopy  was  u(cid:415)lized  to  measure  the  cross-exchange  electron  transfer  rate 

constant, k12, between [Cu(bpyPY4)]2+ and decamethylferrocene (Fe(Cp*)2).  

39 

 
 
 
 
Scheme 2.2: Reac(cid:415)on followed in stopped-ﬂow experiments. 

Decamethylferrocene  was  chosen  for  the  cross-exchange  to  op(cid:415)mize  the  driving  force  and 

because it has a well-known self-exchange rate constant for electron transfer. Due to the large 

poten(cid:415)al diﬀerence between [Cu(bpyPY4)]2+/+ and [Fe(Cp*)2]+/0, it was assumed that the reac(cid:415)on 

went to comple(cid:415)on, with no signiﬁcant back reac(cid:415)on. Figure 3.6A shows a ﬁt of the absorbance 

at 450 nm vs (cid:415)me plot, which represents the growth of the [Cu(bpyPY4)]+ species in solu(cid:415)on due 

to the reduc(cid:415)on of [Cu(bpyPY4)]2+ by Fe(Cp*)2, using the following equa(cid:415)on:  

𝐴 = 𝐴(cid:2998) + (𝐴(cid:2868) − 𝐴(cid:2998))𝑒(cid:2879)(cid:3038)(cid:3290)(cid:3277)(cid:3294)(cid:3047) (3.1) 

For all reac(cid:415)ons, Fe(Cp*)2 was held in excess to maintain pseudo-ﬁrst order condi(cid:415)ons, allowing 

kobs to be represented by:  

𝑘𝑜𝑏𝑠 = 𝑘12[𝐹𝑒(𝐶𝑝∗ )2 ] (3.2) 

Figure 3.6: A) An absorbance vs (cid:415)me plot at 450 nm, showing the increase of [Cu(bpyPY4)]TFSI 
species (black dot) and ﬁ(cid:427)ng (red line) for the reduc(cid:415)on of [Cu(bpyPY4)]TFSI2 by 
decamethylferrocene (Fe(Cp*)2). B) The pseudo-ﬁrst order rate constant, kobs, vs the 
concentra(cid:415)on of [Fe(Cp* )2] for the reac(cid:415)ons of [Cu(bpyPY4)]TFSI2 and [Fe(Cp* )2] in CH3CN 
containing 0.1 M LiTFSI. 

40 

 
 
 
 
 
 
A linear ﬁt of the previously ﬁ(cid:425)ed kobs values vs concentra(cid:415)on of Fe(Cp*)2 provided the value of 

the cross-exchange rate constant, k12, from the slope of the line (Figure 3.6B). The ini(cid:415)al reac(cid:415)on 

mixtures and the pseudo-ﬁrst order observed rate constants can be found in Table 3.1.  

Table 3.3: The ini(cid:415)al reac(cid:415)on mixture and the observed rate constant, kobs, for the cross 
exchange between [Cu(bpyPY4)]TFSI2 and [Fe(Cp* )2]. 

Using the experimentally determined cross-exchange rate constant from Equa(cid:415)on 3.2, k12, and 

the previously determined self-exchange constant for [Fe(Cp*)2]+/0 , k11, the Marcus cross-rela(cid:415)on  

𝑘12 = (𝑘11𝑘22𝐾12𝑓12)1/2𝑊12 (3.3) 

was used to calculate the self-exchange rate constant for [Cu(bpyPY4)]2+/+ , k22, where K12 is the 

equilibrium constant, f12 is a nonlinear correc(cid:415)on term, and W12 is an electrosta(cid:415)c work term for 

bringing  the  reactants  into  contact.  The  nonlinear  correc(cid:415)on  term  and  the  work  term  were 

calculated  (see  below  for  details  on  the  calcula(cid:415)on)  and  determined  to  be  2.60  and  0.99, 

respec(cid:415)vely  (Table  3.3).  However,  it  should  be  noted  that  there  will  be  some  error  in  the 

calcula(cid:415)on  of  the  work  term  since  the  Debye-Huckel  model  is  not  expected  to  have  accurate 

results when the solu(cid:415)on has a high ionic strength, as was the case for this experiment with 0.1 

M suppor(cid:415)ng electrolyte. The equilibrium constant, K12, was determined using Equa(cid:415)on 3.4:  

𝑛𝐹𝛥𝐸 = 𝑅𝑇𝑙𝑛𝐾12 (3.4) 

where n is the number of electrons transferred, F is Faraday’s constant, ΔE is the diﬀerence in 

formal poten(cid:415)al between the oxidant and reductant in solu(cid:415)on, R is the ideal gas constant, and 

41 

 
 
 
 
 
T is the temperature in Kelvin. The redox poten(cid:415)al for [Cu(bpyPY4)]2+/+ was found to be -0.43 V 

vs ferrocene and the redox poten(cid:415)al of [Fe(Cp*)2]+/0 has been previously determined to be -0.52 

V2 vs ferrocene85, using cyclic voltammetry which gave a calculated K12 of 18.4. The self-exchange 

rate  constant  for  [Fe(Cp*)2]+/0  was  previously  iden(cid:415)ﬁed  to  be  3.8  ×  107  M-1s-1  using  NMR 

techniques.86  Using  these  values,  the  self-exchange  rate  for  the  [Cu(bpyPY4)]2+/+  couple  was 

determined to be 8.78 M-1s-1. The measurement was run mul(cid:415)ple (cid:415)mes to conﬁrm reproducibility 

and all samples were within error (see Table 3.2).  

Table 3.4: The self-exchange rate of [Cu(bpyPY4)]TFSI1/2 complex for each trial. 

The  work  required  to  move  the  reactant  complexes  to  a  distance,  r,  for  the  electron  transfer 

reac(cid:415)on was calculated using the Equa(cid:415)on 3.5.  

𝑊(cid:2869)(cid:2870) = exp [

(cid:2879)(cid:3050)(cid:3117)(cid:3118)(cid:2878)(cid:3050)(cid:3118)(cid:3117)(cid:2879)(cid:3050)(cid:3117)(cid:3117)(cid:2879)(cid:3050)(cid:3118)(cid:3118)
(cid:2870)(cid:3019)

] (3.5) 

𝑤(cid:3036)(cid:3037)(𝑟) =  

(cid:3053)(cid:3284)(cid:3053)(cid:3285)(cid:3044)(cid:3118)(cid:3015)(cid:3250)
(cid:2872)(cid:3095)(cid:3084)(cid:3116)(cid:3084)(cid:3045)((cid:2869)(cid:2878)(cid:3081)(cid:3045))

 (3.6) 

Equa(cid:415)on  3.6  is  used  to  determine  the  work  associated  with  the  forward  and  reverse  cross-

exchange reac(cid:415)on, w12 and w21 respec(cid:415)vely, and the self-exchange reac(cid:415)ons, w11 and w22. In this 

equa(cid:415)on zi and zj are the charges of the interac(cid:415)ng complexes, q is the charge of an electron, NA 

is Avogadro’s constant, ε0 is the permi(cid:427)vity of free space, ε is the sta(cid:415)c dielectric of the medium, 

𝛽 = (

(cid:2870)(cid:3044)(cid:3118)(cid:3015)(cid:3250)(cid:3010)
(cid:2869)(cid:2868)(cid:2868)(cid:2868) (cid:3116)(cid:3084)(cid:3038)(cid:3251)(cid:3021)

)(cid:2869)/(cid:2870), I is the ionic strength of the solu(cid:415)on, and kB is the Boltzmann’s constant. The 

calcula(cid:415)on has the following assump(cid:415)ons: the work is assumed to be primarily Coulombic, the 

42 

 
 
 
 
distance,  r,  is  assumed  to  be  the  center-to-center  distance  between  the  complexes,  and  the 

reactants are assumed to be spherical. The non-linear correc(cid:415)on term, f12, was calculated using 

the equa(cid:415)on  

𝑙𝑛𝑓(cid:2869)(cid:2870) =

(cid:2869)

(cid:2872)

(cid:4672)(cid:3039)(cid:3041) (cid:3117)(cid:3118)(cid:2878)
(cid:3286)(cid:3117)(cid:3117)(cid:3286)(cid:3118)(cid:3118)
(cid:3275)(cid:3118)

(cid:2922)(cid:2924)(cid:4672)

(cid:3118)

(cid:4673)

(cid:3298)(cid:3117)(cid:3118)(cid:3127)(cid:3298)(cid:3118)(cid:3117)
(cid:3267)(cid:3269)
(cid:3298)(cid:3117)(cid:3117)(cid:3126)(cid:3298)(cid:3118)(cid:3118)
(cid:3267)(cid:3269)

(cid:4673)(cid:2878)

 (3.7) 

where K12 is the equilibrium constant, w12 and w21 are the work associated with the forward and 

reverse  cross-exchange  reac(cid:415)on  respec(cid:415)vely,  R  is  the  ideal  gas  constant,  T  is  temperature  in 

Kelvin, and Z is the frequency factor, which is assumed to be 1013 M-1s-1 due to the larger inner-

sphere contribu(cid:415)ons to the total reorganiza(cid:415)on energy.  

The  total  reorganiza(cid:415)on  energy  of  the  self-exchange  reac(cid:415)on  was  calculated  following  the 

equa(cid:415)on  

𝑘(cid:2869)(cid:2869) = 𝐾(cid:2869)(cid:2870)𝑍𝛤𝑒(cid:2879)((cid:3090)(cid:3117)(cid:3117) (cid:2872)(cid:3038)(cid:3251)(cid:3021))

⁄

 (3.8) 

where k11 is the previously calculated self-exchange rate, K12 is the equilibrium constant, Z is the 

frequency factor, same as in Equa(cid:415)on 3.7, Γ is a correc(cid:415)on for nuclear tunneling, which is assumed 

to be ~1, λ11 is the total reorganiza(cid:415)on energy of the self-exchange reac(cid:415)on, kB is the Boltzmann 

constant, and T is the temperature of solu(cid:415)on. The resul(cid:415)ng λ11 was found to be 3.15 eV.  

The outer-sphere reorganiza(cid:415)on energy for two spherical reactants can be determined using  

𝜆(cid:2869)(cid:2869),(cid:3042) =

((cid:3057)(cid:3053)(cid:3044))(cid:3118)
(cid:2872)(cid:3095)(cid:3084)(cid:3116)

(cid:4672)

(cid:2869)
(cid:2870)(cid:3028)(cid:3117)

+

(cid:2869)
(cid:2870)(cid:3028)(cid:3118)

−

(cid:2869)
(cid:3045)(cid:3117)(cid:3117)

(cid:4673) (cid:3436)

(cid:2869)
(cid:3005)(cid:3290)(cid:3291),(cid:3294)(cid:3290)(cid:3287)

−

(cid:2869)
(cid:3005)(cid:3294),(cid:3294)(cid:3290)(cid:3287)

(cid:3440) (3.9) 

where λ11,o is the outer-sphere reorganiza(cid:415)on energy, Δz is the change in charge of the complex, 

q is the charge of an electron, ε0 is the permi(cid:427)vity of free space, a1 and a2 are the atomic radii of 

the  reactants,  r11  is  the  center−center  distance  between  the  reactants,  Dop,sol  is  the  op(cid:415)cal 

dielectric  constant  of  the  medium,  which  is equal  to  the  square of  the  refrac(cid:415)ve  index  of  the 

43 

 
 
medium, and Ds,sol is the sta(cid:415)c dielectric constant of the medium. The outer-sphere reorganiza(cid:415)on 

energy  for  the  self-exchange  reac(cid:415)on  was  calculated  to  be  0.99  eV.  The  inner-sphere 

reorganiza(cid:415)on energy was determined using 𝜆11,𝑖 = 𝜆11 − 𝜆11,𝑂 yielding a result of 2.16 eV for the 

inner-sphere reorganiza(cid:415)on energy of the self-exchange reac(cid:415)on. 

Table 3.5: Summary of the kine(cid:415)c data used to calculate the self-exchange rate constant of the 
[Cu(bpyPY4)]TFSI1/2 complex by monitoring the reac(cid:415)on of [Cu(bpyPY4)]TFSI2 and [Fe(Cp* )2] in 
CH3CN containing 0.1 M LiTFSI at 25°C. 

DSSC devices were fabricated with  [Cu(bpyPY4)]TFSI1/2 as the redox mediator, where poly(3,4-

ethylenedioxythiophene) (PEDOT) was u(cid:415)lized at the counter electrode and the commercial dye 

Y123 was used as the light-harvester. The electrolyte consisted of 0.10 M [Cu(bpyPY4)]TFSI, 0.05 

M [Cu(bpyPY4)]TFSI2, 0.1 M LiTFSI, and 0.5 M 4-tert-butylpyridine, 10 equivalents vs the Cu(II) in 

the electrolyte, in acetonitrile. The constructed devices gave respectable performance with a JSC 

of 4.98(±0.01) mAcm-2 and an overall power conversion eﬃciency of 1.23% for the devices with 

TBP and 4.73(±0.36) mAcm-2 and an overall power conversion eﬃciency of 0.21% for the devices 

without TBP, as shown in Table 3.6. 

44 

 
 
 
 
2
-

m
c
A
m
/
t
n
e
r
r
u
C

6

5

4

3

2

1

0

-0.5

-0.4

-0.3
-0.2
Voltage / V

-0.1

0

Figure 3.7: J−V curves for the [Cu(bpyPY4)]TFSI1/2 based DSSC devices. The devices represented 
by the black curve contains TBP while the gray curve is without TBP. 

Table 3.6: Summary of device performance, parameters represent the average of three devices. 

η (%) 

Jsc (mA*cm-2)  Voc (V) 

FF 

Cu(PY6) 

0.21(±0.07)  4.73(±0.36)  0.14(±0.01) 0.32(±0.05) 

Cu(PY6) with TBP 1.23(±0.08)  4.98(±0.01)  0.44(±0.01) 0.56(±0.02) 

Both sets of devices were aﬀected by mass transport issues which limited their current. Looking 

at  the  incident  photon-to-current  conversion  eﬃciency  (IPCE)  spectra,  which  measures  the 

current  at  various  wavelengths  to  determine  how  much  of  the  photons  absorbed  at  each 

wavelength are converted to current. The IPCE in Figure 3.8 shows that as TBP is added there is a 

slight increase in the photon-to-current conversion between 450 and 650 nm. This increase could 

be caused by decreased recombina(cid:415)on to the redox shu(cid:425)le as TBP is being added to the solu(cid:415)on. 

The IPCE spectra can also be integrated to determine what the theore(cid:415)cal current of the device, 

resul(cid:415)ng with an integrated current of 7.87 mAcm-2 for the devices with TBP and 7.77 mAcm-2 for 

45 

 
 
 
 
 
 
 
the devices without TBP, we can see that roughly 3 mAcm-2 were lost due to the mass transport 

limita(cid:415)ons. 

%
/
E
C
P
I

60

50

40

30

20

10

0

9.00

8.00

7.00

6.00

5.00

4.00

3.00

2.00

1.00

0.00

2
-

m
c
A
m
/
t
n
e
r
r
u
C

400

450

500

550

600

650

700

Wavelength / nm

Figure 3.8: IPCE curves for the [Cu(bpyPY4)]TFSI1/2 based DSSC devices. The devices represented 
by the black points and curve contains TBP while the gray points and curve is without TBP. For 
the IPCE, the points represent IPCE responses while the solid lines represent the integrated 
photocurrent. 

The  eﬀect  of  TBP  on  the  fermi  level  of  the  devices  was  inves(cid:415)gated.  By  knowing  the  solu(cid:415)on 

poten(cid:415)al and the VOC of the devices the fermi level could be calculated, which can be determined 

by adjus(cid:415)ng the open circuit voltage (VOC) by the solu(cid:415)on poten(cid:415)al. For both devices with and 

without TBP the solu(cid:415)on poten(cid:415)al was roughly the same, at -0.443 V vs ferrocene for the devices 

without  TBP  and  -0.466  V  vs  ferrocene  for  the  devices  with  TBP.  However,  there  was  a  stark 

diﬀerence in the VOC between both sets of devices, devices without TBP having a VOC of only 0.14 

V vs solu(cid:415)on poten(cid:415)al where devices with TBP had a VOC of 0.44 V vs solu(cid:415)on poten(cid:415)al.  This 

indicates that the addi(cid:415)on of TBP caused the fermi level to undergo a nega(cid:415)ve shi(cid:332) of roughly 

300  mV.  Although  the  [Cu(bpyPY4)]TFSI1/2  complex  does  not  undergo  ligand  subs(cid:415)tu(cid:415)on  the 

46 

 
 
 
 
 
 
 
addi(cid:415)on of TBP to the electrolyte s(cid:415)ll dras(cid:415)cally increases the electron life(cid:415)me, by almost 1000 

fold. This indicates that the addi(cid:415)on of TBP is s(cid:415)ll able to reduce recombina(cid:415)on in the system, 

although  not  as  signiﬁcantly  as  the  [Cu(dbmed)]OTf1/2 system.  This  shi(cid:332)  in  poten(cid:415)al  could  be 

a(cid:425)ributed to the addi(cid:415)on of TBP in solu(cid:415)on, which can shi(cid:332) the poten(cid:415)al of the conduc(cid:415)on band 

of  the  (cid:415)tania  to  more  nega(cid:415)ve  poten(cid:415)als  and  reducing  recombina(cid:415)on  by  adsorbing  to  the 

(cid:415)tania.33,34  

10

1

0.1

)
s
/
1
(
e
m

i
t
e
f
i
L

0.01

0

[Cu(dbmed)] without TBP

[Cu(bpyPY4)] without TBP

[Cu(dbmed)] with TBP

[Cu(bpyPY4)] with TBP

0.2

0.4

0.6

0.8

1

1.2

Voltage (V)

Figure 3.9: A plot of the electron life(cid:415)me vs poten(cid:415)al plot for DSSC devices using 
[Cu(dbmed)]OTf1/2 (green), [Cu(bpyPY4)]TFSI1/2 (red), with (pale color) and without (dark color) 
the addi(cid:415)on of TBP. 

3.4 Conclusions 

This  study  has  demonstrated  the  complex  rela(cid:415)onship  between  ligand  design  and  metal 

coordina(cid:415)on  in  copper  complexes  using  the  bpyPY4  scaﬀold.  The  crystallographic  analysis 

revealed that the bpyPY4 ligand can adapt to both ﬁve-coordinate and four-coordinate structures 

47 

 
 
 
 
 
around  copper(II)  and  copper(I)  centers.  These  structures  showed  signiﬁcant  devia(cid:415)ons  from 

ideal geometries, indica(cid:415)ng the inﬂuence of the bpyPY4 scaﬀold on how the ligand and metal 

interact. 

The  behavior  of  these  copper-ligand  interac(cid:415)ons  was  further  explored  using  1H-NMR  studies, 

which showed a rapid interchange between pyridine donors that are coordinated and those that 

are  not.  The  results  from  variable-temperature  NMR  spectroscopy  highlighted  the  unstable 

nature of the copper-pyridine bonds, providing a detailed view of how these bonds behave in 

solu(cid:415)on. 

The UV-visible spectroscopic analysis helped improve understanding on the electronic transi(cid:415)ons 

in  these  complexes,  revealing  speciﬁc  absorp(cid:415)on  pa(cid:425)erns  that  correspond  to  their  structural 

preferences in solu(cid:415)on. The consistent behavior of the [Cu(bpyPY4)]2+/+ redox couple, even with 

added  TBP,  suggests  a  strong  metal-ligand  bond,  which  is  promising  for  applica(cid:415)ons  where 

interac(cid:415)ons with Lewis bases are expected. 

The high theore(cid:415)cal current and the nega(cid:415)ve redox poten(cid:415)al make this redox shu(cid:425)le a candidate 

for u(cid:415)liza(cid:415)on with narrow band gap dyes for high photocurrent DSSC devices.  Future studies will 

focus  on  u(cid:415)lizing  this  redox  complex  with  near  IR  absorbing  dyes  and  developing  new  ligand 

designs to vary the energe(cid:415)cs and stability of the copper species. 

Overall, this comprehensive study not only deepens our understanding of copper complexes with 

hexadentate ligands but also lays the groundwork for designing new ligands that can control and 

alter the shapes and reac(cid:415)vi(cid:415)es of metal centers. The poten(cid:415)al uses of these systems in areas 

like catalysis and electronic devices are signiﬁcant, highligh(cid:415)ng the value of detailed studies like 

this one. 

48 

 
 
Chapter 4 – Inves(cid:415)ga(cid:415)on of [Cu(PY5)]2+/+ 

4.1 Introduc(cid:415)on 

Following  Grätzel's  seminal  1993  report  on  dye-sensi(cid:415)zed  solar  cells  (DSSCs)  achieving  10% 

eﬃciency,  advancements  were  hindered  by  the  reliance  on  the  I3-/I-  redox  shu(cid:425)le.15  The  high 

overpoten(cid:415)al required for eﬃcient dye regenera(cid:415)on was a major limita(cid:415)on. Recent progress with 

outer  sphere  redox  shu(cid:425)les  oﬀers  a  solu(cid:415)on  to  mi(cid:415)gate  this  limita(cid:415)on  and  enhance  DSSC 

performance.23  Conversely,  Co(III/II)  based  redox  shu(cid:425)les 

introduce  an 

inner-sphere 

reorganiza(cid:415)on  energy  challenge  due  to  electron  transfer  transi(cid:415)ons  between  d7  Co(II)  and  d6 

Co(III)  states.  This  aﬀects  dye  regenera(cid:415)on  kine(cid:415)cs  by  limi(cid:415)ng  the  available  electron  transfer 

driving force.23 

The  robust  performance  of  copper-based  redox  shu(cid:425)les  in  DSSCs  has  posi(cid:415)oned  them  as  a 

compelling  avenue  for  overcoming  conven(cid:415)onal  limita(cid:415)ons.  Their  success  stems  from  their 

remarkable ability to a(cid:425)ain eﬃciencies that surpass previous benchmarks. With an excep(cid:415)onal 

15.2% eﬃciency under standard solar irradiance and an astonishing 34.5% eﬃciency under indoor 

ﬂuorescent  ligh(cid:415)ng  at  1000  lux  intensity,  copper-based  redox  shu(cid:425)les  have  ﬁrmly  established 

their  creden(cid:415)als.20,87  This  achievement  can  be  a(cid:425)ributed  to  their  intrinsic  advantage  of  lower 

inner-sphere  electron  transfer  reorganiza(cid:415)on  energies  compared  to  tradi(cid:415)onal  op(cid:415)ons. 

Furthermore, their versa(cid:415)le coordina(cid:415)on geometries, witnessed in copper(II) d9 complexes which 

frequently adopt six-coordinate, octahedral or tetragonal, ﬁve-coordinate, square pyramidal or 

trigonal bipyramidal, or four-coordinate, tetrahedral or square planar geometries, and copper(I) 

d10  complexes  which  predominantly  adopt  four-coordinate,  tetrahedral  or  square  planar 

geometries, underscore their adaptability.60–63 Employing ligands based on bidentate bipyridine 

49 

 
 
or phenanthroline mo(cid:415)fs, augmented with strategically posi(cid:415)oned bulky groups, these shu(cid:425)les 

can reduce the reorganiza(cid:415)on energy needed for the transi(cid:415)on between Cu(II) and Cu(I) redox 

states, facilita(cid:415)ng more eﬃcient electron transfer processes within the DSSC system.64 

Recent observa(cid:415)ons have illuminated a cri(cid:415)cal facet of copper-based redox shu(cid:425)le performance 

within DSSCs. Notably, prevalent Lewis base addi(cid:415)ves, including 4-tert-butylpyridine (TBP), have 

been  revealed  to  coordinate  with  Cu(II)  species.38  This  coordina(cid:415)on  has  poten(cid:415)ally  adverse 

impacts on copper redox mediator electrochemistry and, consequently, overall device eﬃcacy. 

Studies into this phenomenon elucidate that TBP-induced coordina(cid:415)on or ligand subs(cid:415)tu(cid:415)on—

o(cid:332)en involving bipyridyl and phenanthroline ligands—leads to intricate electrochemical changes 

within the DSSC devices. When the ligand subs(cid:415)tu(cid:415)on occurs and [Cu(TBP)4]2+ is present in the 

devices it is accompanied by a substan(cid:415)al nega(cid:415)ve shi(cid:332) in formal redox poten(cid:415)als, by hundreds 

of  millivolts.38  This  phenomenon  limits  the  a(cid:425)ainable  photovoltage,  exer(cid:415)ng  a  signiﬁcant 

constraint on device performance. 

Recent strides in copper-based redox shu(cid:425)le design have been underscored by the eﬀorts of Sun 

and colleagues, who introduced copper based redox shu(cid:425)les incorpora(cid:415)ng pentadentate ligands 

within  DSSC  devices.41  The  pentadentate  Cu(II)  complex  was  shown  to  have  a  resistance  to 

subs(cid:415)tu(cid:415)on,  even  when  exposed  to  commonly  employed  Lewis  base  addi(cid:415)ves  such  as  4-tert-

butylpyridine (TBP). This resistance to ligand subs(cid:415)tu(cid:415)on is a(cid:425)ributed to two factors. First, the 

increased den(cid:415)city of the ligand translates to a heightened overall stability of the metal complex, 

owing to the chela(cid:415)ng eﬀect. Second, they speciﬁcally designed the coordina(cid:415)on sphere's steric 

constraints  to  shield  the  copper  complexes—par(cid:415)cularly  their  oxidized  forms—from  TBP 

50 

 
 
coordina(cid:415)on. Building upon this concept, a copper-based redox shu(cid:425)le harnessing the potency 

of a pentadentate ligand, 2,6-bis[1,1-bis(2-pyridyl)ethyl]pyridine (PY5), is presented.  

4.2 Experimental Details 

All  NMR  spectra  were  taken  on  an  Agilent  DirectDrive2  500  MHz  spectrometer  at  room 

temperature and referenced to residual solvent signals. All NMR spectra were evaluated using 

the  MestReNova  so(cid:332)ware  package  features.  Cyclic  voltammograms  were  obtained  using 

µAutolabIII poten(cid:415)ostat using BASi glassy carbon electrode, a pla(cid:415)num mesh counter electrode, 

and a fabricated 0.01 M AgNO3, 0.1 M TBAPF6 in acetonitrile Ag/AgNO3 reference electrode. All 

measurements  were  internally  referenced  to  ferrocene/ferrocenium  couple  via  addi(cid:415)on  of 

ferrocene  to  solu(cid:415)on  a(cid:332)er  measurements  or  run 

in  a  parallel  solu(cid:415)on  of  the  same 

solvent/electrolyte.  UV-Vis  spectra  were  taken  using  a  PerkinElmer  Lambda  35  UV-Vis 

spectrometer using a 1 cm path length quartz cuve(cid:425)e at 480 nm/min. Elemental Analysis data 

were obtained via Midwest Microlab. Con(cid:415)nuous-wave (Cw) EPR measurements were performed 

using a Bruker E-680X spectrometer opera(cid:415)ng at X-band frequencies and equipped with an SHQ-

E cavity. EPR samples comprising 0.01 M Cu(II) complexes in acetonitrile or dichloromethane. The 

liquid  solu(cid:415)ons  were  studied  in  a  quartz  ﬂat  cell.  An  Oxford  ESR-900  cryostat  and  an  ITC-503 

temperature  controller  were  used  for  measurements  at  298  K.  Cw-EPR  simula(cid:415)ons  were 

performed with EasySpin 5.2.28, running in MATLAB 2020a.  We simulated the solu(cid:415)on spectra 

using  the  “chili”  module  from  EasySpin.  Our  simula(cid:415)ons  varied  the  coeﬃcients  for  copper 

hyperﬁne coupling, isotropic diﬀusion  correla(cid:415)on  (cid:415)me,  and intrinsic linewidth. The results are 

shown as dashed curves oﬀset from the spectra (solid lines).For single-crystal X-ray diﬀrac(cid:415)on, 

single  crystals  were  mounted  on  a  nylon  loop  with  paratone  oil  using  a  Bruker  APEX-II  CCD 

51 

 
 
diﬀractometer. Crystals were maintained at T ¼ 173(2) K during data collec(cid:415)on. Using Olex2, the 

structures  were  solved  with  the  ShelXS  structure  solu(cid:415)on  program  using  the  Direct  Methods 

solu(cid:415)on  method.  Photoelectrochemical  measurements  were  performed  with  a  poten(cid:415)ostat 

(Autolab PGSTAT 128N) in combina(cid:415)on with a xenon arc lamp. An AM 1.5 solar ﬁlter was used to 

simulate sunlight at 100 mW cm-2, and the light intensity was calibrated with a cer(cid:415)ﬁed reference 

cell system (Oriel Reference Solar Cell & Meter). A black mask with an open area of 0.07 cm-2 was 

applied on top of the cell ac(cid:415)ve area. A monochromator (Horiba Jobin Yvon MicroHR) a(cid:425)ached 

to the 450 W xenon arc light source was used for monochroma(cid:415)c light for IPCE measurements. 

The  photon  ﬂux of  the  light  incident  on  the  samples  was  measured  with  a  laser  power  meter 

(Nova II Ophir). IPCE measurements were made at 20 nm intervals between 400 and 700 nm at 

short circuit current. 

TEC 15 FTO was cut into 1.5 cm by 2 cm pieces which were sonicated in soapy DI water for 15 

minutes,  followed  by  manual  scrubbing  of  the  FTO  with  Kimwipes.  The  FTO  pieces  were  then 

sonicated in DI water for 10 minutes, rinsed with acetone, and sonicated in isopropanol for 10 

minutes. The FTO pieces were dried in room temperature air then immersed in an aqueous 40 

mM TiCl4 solu(cid:415)on for 60 minutes at 70 °C. The water the for the TiCl4 treatment was preheated 

to 70 °C prior to adding 2 M TiCl4 to the water. The 40 mM solu(cid:415)on was immediately poured onto 

the  samples  and  placed  in  a  70  °C  oven  for  the  60-minute  deposi(cid:415)on.  The  FTO  pieces  were 

immediately rinsed with  18 MΩ  water followed by isopropanol and were  annealed by hea(cid:415)ng 

from room temperature to 500 °C, holding at 500 °C for 30 minutes. A 0.36 cm2 area was doctor 

bladed with commercial 30 nm TiO2 nanopar(cid:415)cle paste (DSL 30NRD). The transparent ﬁlms were 

le(cid:332) to rest for 10 minutes and were then placed in a 125 °C oven for 30 minutes. The samples 

52 

 
 
were annealed in an oven that was ramped to 325 °C for 5 minutes, 375 °C for 5 minutes, 450 °C 

for 5 minutes, and 500 °C for 15 minutes. The 30 nm nanopar(cid:415)cle ﬁlm thickness was ~8 μm. A(cid:332)er 

cooling to room temperature, a second TiCl4 treatment was performed as described above. When 

the  anodes  had  cooled  to  80  °C,  they  were  soaked  in  a  dye  solu(cid:415)on  of  0.1  mM  Y123  in  1:1 

acetonitrile  :  tert-butyl  alcohol  for  18  hours.  A(cid:332)er  soaking,  the  anodes  were  rinsed  with 

acetonitrile and were dried gently under a stream of nitrogen. 

The PEDOT counter electrodes were prepared by electropolymeriza(cid:415)on in a solu(cid:415)on of 0.01 M 

EDOT and 0.1 M LiClO4 in 0.1 M SDS in 18 MΩ water. A constant current of 8.3 mA for 250 seconds 

was applied to a 54 cm2 piece of TEC 8 FTO with predrilled holes using an equal size piece of FTO 

as the counter electrode. The PEDOT electrodes were then washed with DI water and acetonitrile 

before being dried under a gentle stream of nitrogen and cut into 1.5 cm by 1.0 cm pieces. The 

working and counter electrodes were sandwiched together with 25 μm surlyn ﬁlms by placing 

them on a 140 °C hotplate and applying pressure. The cells were then ﬁlled in a nitrogen ﬁlled 

glove box with electrolyte through one of the two predrilled holes and were sealed with 25 μm 

surlyn backed by a glass coverslip and applied heat to seal with a soldering iron. The electrolyte 

consisted of 0.10 M Cu(I), 0.05 M Cu(II), 0.1 M Li(Counter-ion), and 0.5 M 4-tert-butylpyridine in 

acetonitrile. All batches were made with at least 5 cells per batch. Contact to the TiO2 electrode 

was made by soldering a thin layer of indium wire onto the FTO. 

Acetonitrile,    deuterated    acetonitrile,    dichloromethane,    methanol,  ethanol,  diethyl-ether, 

deionized water, 1,1-bis(2-pyridyl)ethane, 2,6-diﬂuoropyridine, 2.5 M n-butyl lithium in hexanes, 

tetrakisacetonitrile copper(I) triﬂate,  copper(II)  triﬂate ,  silver  nitrate,  tetrabutylammonium  

53 

 
 
hexaﬂuorophosphate, lithium triﬂate, lithium bistriﬂimide, silver bistriﬂimide, copper (I) chloride, 

copper (I) bistriﬂimide, and isopropyl alcohol were purchased from Sigma-Aldrich. 

Star(cid:415)ng  material  1,1-Bis(2-pyridyl)ethane,  and  ligand  PY5    were  synthesized  according  to 

published procedures.88  

[Cu(PY5)](OTf)  

A mixture of PY5 (74.5 mg, 0.168 mmol) and [Cu(ACN)4](OTf) (57.5 mg, 0.153 mmol) in anhydrous 

acetonitrile was s(cid:415)rred for 30 minutes at room temperature. The solu(cid:415)on was precipitated with 

anhydrous diethyl ether, forming a yellow solid, and the solid was collected. The solid was dried 

under vacuum. (97.8 mg 97.4% yield) 1H NMR (500 MHz, acetonitrile-d3): δ = 8.53 (d, 4H); 8.04 

(t, 1H); 7.92 (d, 2H); 7.77 (t, 4H); 7.37 (d, 4H); 7.26 (t, 4H); 2.20 (s, 6H); 2.18 (1.5 H). Elem. Anal. 

Calc. for C30H25CuF3N6O3S C, 54.42; H, 3.98; N, 12.28. Found: C, 54.79; H, 3.91; N, 12.05. TOF-MS-

ES+ m/z calc. for [Cu(PY5)] C29H25CuN5 506.14; Found, 506.1407. 

[Cu(PY5)](TFSI)  

Silver  Bistriﬂimide  (47.0  mg,  0.121  mmol)  and  copper  chloride  (12.0  mg,  0.121  mmol)  was 

dissolved in minimal anhydrous ACN and was s(cid:415)rred for 30 minutes at room temperature. A(cid:332)er 

mixing, the solid silver chloride was removed from the solu(cid:415)on, and the solu(cid:415)on was then added 

to PY5 (52.5 mg, 0.118 mmol) and s(cid:415)rred overnight. The solu(cid:415)on was precipitated with anhydrous 

diethyl  ether,  forming  a  yellow  solid,  and  the  solid  was  collected.  The  solid  was  dried  under 

vacuum. (90.2 mg 96.9% yield). 1H NMR (500 MHz, acetonitrile-d3): δ = 8.53 (d, 4H); 8.04 (t, 1H); 

7.92 (d, 2H); 7.77 (t, 4H); 7.37 (d, 4H); 7.26 (t, 4H); 2.20 (s, 6H); 2.18 (1.5 H). Elem. Anal. Calc. for 

C31H25CuF6N6O4S2 C, 54.42; H, 3.98; N, 12.28. Found: C, 54.79; H, 3.91; N, 12.05. TOF-MS-ES+ m/z 

calc. for [Cu(PY5)], 506.14; Found, 506.14. 

54 

 
 
[Cu(PY5)](OTf)2 

 A mixture of PY5 (0.2924 g, 0.66 mmol) and Cu(OTf)2 (0.2210 g, 0.61 mmol) in anhydrous DCM 

was s(cid:415)rred for 30 minutes at room temperature. The solu(cid:415)on was precipitated with anhydrous 

diethyl ether, forming a blue solid, and the solid was collected. The solid was dried under vacuum. 

(0.4776 mg 92.5% yield) Elem. Anal. Calc. for C31H25CuF6N6O6S2 C, 46.24; H, 3.13; N, 7.89. Found: 

C, 45.93; H, 3.32; N, 8.38.  

[Cu(PY5)](TFSI)2  

A mixture of PY5 (47.8 mg, 0.108 mmol) and Cu(TFSI)2 (67.2 mg, 0.108 mmol) in anhydrous DCM 

was s(cid:415)rred for 30 minutes at room temperature. The solu(cid:415)on was precipitated with anhydrous 

diethyl ether, forming a blue solid, and the solid was collected. The solid was dried under vacuum. 

Any  further  puriﬁca(cid:415)on  was  done  via  recrystalliza(cid:415)on  from  ACN  with  diﬀused  ether.(46.3  mg 

40.1% yield) Elem. Anal. Calc. for C35H28CuF12N8O8S4 C, 37.93; H, 2.55; N, 10.11. Found: C, 37.38; 

H, 2.31; N, 9.41.  

4.3 Results and Discussion 

The synthesis of the ligand, PY5, has been previously documented, involving the deprotona(cid:415)on 

of  1,1-bis(2-pyridyl)ethane  by  n-BuLi  followed  by  the  addi(cid:415)on  of  2,6-diﬂuoropyridine  as  the 

electrophile.89 The copper complexes were synthesized by reac(cid:415)ng equimolar ra(cid:415)os of the PY5 

ligand  with  copper  precursors  in  diﬀerent  solvents,  leading  to  the  forma(cid:415)on  of  [Cu(PY5)]OTf, 

[Cu(PY5)]OTf2, [Cu(PY5)]TFSI, and [Cu(PY5)]TFSI2. The copper(I) bistriﬂimide was synthesized by 

combining equimolar amounts of silver bistriﬂimide and  copper chloride. The complexes were 

puriﬁed via recrystalliza(cid:415)on  from acetonitrile for copper(I)  and dichloromethane for copper(II) 

solu(cid:415)ons and characterized using 1H-NMR and elemental analysis. 

55 

 
 
Scheme 4.3: Syntheses of the [Cu(PY5)]2+/+ complexes, where is the counter ion used for each 
batch, either triﬂate (OTf) or bistriﬂimide (TFSI). 

Single crystal X-ray diﬀrac(cid:415)on revealed the solid-state structures of the complexes. A geometric 

index, τ4, can be used for four-coordinate complexes to determine how much distor(cid:415)on there is 

from  ideal  tetrahedral  or  square  planer  geometries.74  Where  a  τ4  of  1  represents  an  ideal 

tetrahedral  geometry  and  0  represents  an  ideal  square  planer  geometry.  The  [Cu(PY5)]OTf 

complex exhibited a highly distorted tetrahedral geometry with a calculated τ4 value of 0.65. The 

constrained nature of the ligand led to a signiﬁcant por(cid:415)on of the copper center being solvent-

exposed. The choice of solvent during synthesis played a role in preven(cid:415)ng a dispropor(cid:415)ona(cid:415)on 

reac(cid:415)on in which the copper(I) form dispropor(cid:415)onated to copper metal and a copper(II) complex 

when the complex was synthesized in dichloromethane. Interes(cid:415)ngly when an NMR was taken of 

the dispropor(cid:415)onated copper(II) product, it did not match the synthe(cid:415)c [Cu(PY5)]OTf2 product. 

56 

 
 
 
Acetonitrile was used as the solvent of choice when synthesizing the copper(I) form since it did 

not cause the copper(I) form to undergo dispropor(cid:415)ona(cid:415)on. 

Figure 4.1: Crystal structures of the ca(cid:415)ons of the [Cu(PY5)]OTf species (A), and [Cu(PY5)]OTf2 
species (B), and the [Cu(PY5)]TFSI2 species which was grown in dichloromethane, a non-
coordina(cid:415)ng solvent (C). Depicted ellipsoids are at the 50% probability level. The non-
interac(cid:415)ng anions and hydrogens were omi(cid:425)ed for clarity. 

The  solid-state  structure  of  [Cu(PY5)]OTf2  showed  pseudo-octahedral  geometry,  with  the  PY5 

ligand coordinated at ﬁve sites and the acetonitrile, the solvent used for synthesis, coordinated 

at  the  sixth  site.  The  average  bond  length  for  the  PY5  ligand  is  2.081  Å,  and  the  bond  length 

between the copper and the nitrogen on the acetonitrile is 2.369 Å, a signiﬁcantly longer distance. 

57 

 
 
 
To inves(cid:415)gate changes in a system with a vacant axial site, the crystals were regrown using an 

innocent  solvent,  dichloromethane.  This  process  revealed  an  interac(cid:415)on  between  one  of  the 

oxygens  on  the  triﬂuoromethanesulfonate  counter-ion  and  the  copper  center  in  the  axial 

posi(cid:415)on.  The  bond  length  was  measured  at  2.453  Å.  This  copper-counter-ion  bond  length 

increases  further  when  the  bistriﬂimide  counter-ion  is  studied,  yielding  a copper-oxygen  bond 

length of 2.728 Å. Challenges arose in purifying acetonitrile-bound copper(II) complexes due to 

the lability of acetonitrile, leading to a mixture of copper complexes with and without acetonitrile 

bound.  To  ensure  accurate  measurements,  the  copper(II) 

form  was  synthesized 

in 

dichloromethane. 

To  assess  the  structural  nuances  of  the  copper  complex  with  diﬀerent  counter  ions,  1H-NMR 

analysis was employed. The 1H-NMR spectra for both [Cu(PY5)]OTf and [Cu(PY5)]TFSI, shown in 

Figures  4.1A  and  4.4A  respec(cid:415)vely,  were  recorded  in  deuterated  acetonitrile  at  room 

temperature.  Both  spectra  display  an  integra(cid:415)on  for  25  protons.  Four  peaks  represen(cid:415)ng  the 

pyridine “arms” are observed: one at approximately 8.5 ppm and three others between 7.35 and 

7.95 ppm, with each integra(cid:415)ng to four protons. Peaks at around 8.00 ppm, integrated to one 

proton, and 7.25 ppm, integrated to two protons, correspond to the central pyridine. The peak at 

approximately 2.20 ppm, integra(cid:415)ng to six protons, is a(cid:425)ributed to the ligand's methyl groups. 

The  copper(II)  complex  is  a  paramagne(cid:415)c  compound  making  it  hard  to  pull  any  quan(cid:415)ﬁable 

informa(cid:415)on from the spectra, however, it should be noted that the NMR of both of the OTf and 

TFSI counter ion, irrespec(cid:415)ve of it they are of the copper(I) or copper(II) complexes, yield similar 

spectra. 

58 

 
 
Figure 4.3: A) Cyclic voltammograms of 2 mM [Cu(PY5)]OTf2, blue, [Cu(PY5)]TFSI2, red, in 
anhydrous CH3CN. B) [Cu(PY5)]OTf2 in dichloromethane, blue, with 5 equivalents of acetonitrile 
added to the solu(cid:415)on. All samples contained 0.1 M TBAPF6 as a suppor(cid:415)ng electrolyte, 
measured at a scan rate of 0.1 V/s on a glassy carbon working electrode. 

To  further  understand  how  the  electrochemical  behavior  of  the  complex  was  aﬀected  by  the 

diﬀerent  counter  ions,  cyclic  voltammetry  was  conducted.  In  the  [Cu(PY5)]OTf2  complex,  two 

peaks were observed in the cyclic voltammogram. One peak appeared at -0.372 V vs ferrocene, 

while the other was observed at -0.662 V vs ferrocene. When the complex was measured in the 

absence of acetonitrile, using dichloromethane as the solvent, the peak at -0.662 V went away, 

and  when  acetonitrile  was  (cid:415)trated  into  the  solu(cid:415)on,  the  peak  at  -0.622  V  returned.  Which 

59 

 
 
 
indicates that the peak at -0.662 V was the acetonitrile bound complex, and the peak at -0.372 V 

was  either  triﬂate  bound  or  a  5-coordinate  [Cu(PY5)]OTf2  complex.  When  the  [Cu(PY5)]TFSI2 

complex  was  inves(cid:415)gated  in  anhydrous  acetonitrile  it  exhibited  a  single  peak  at  -0.382  V  vs 

ferrocene,  indica(cid:415)ng  that  the  acetonitrile  was  not  axially  bound  in  the  TFSI  version  of  the 

complex. This could be due to the increased bulk of the TFSI counter-ion when compared to the 

OTf counter-ion, the TFSI may be physically blocking the acetonitrile from binding with the copper 

center.  The  minimal  change  in  redox  poten(cid:415)al  of  the  peak  around  -0.375  V  vs  ferrocene  with 

various  counter  ions  indicates  that  the  redox  couple  being  measured  does  not  require  the 

presence of acetonitrile, and it is indiﬀerent to the counter-ion present, even though the triﬂate 

counter-ion has much stronger interac(cid:415)ons with the copper center. This suggests that the peak 

at around -0.375 mV corresponds to a 5-coordinate species where the axial posi(cid:415)on is vacant. 

Figure 4.2: The UV-Vis spectra of the copper(I), A, and copper(II), B, complexes in anhydrous 
acetonitrile. The triﬂate complexes are in blue while the bistriﬂimide complexes are in red. 

Shi(cid:332)ing our focus to how the diﬀerent counter ions aﬀect the op(cid:415)cal proper(cid:415)es of the copper 

complexes, we examined their UV-visible absorp(cid:415)on spectra. The UV-Vis spectra of the copper(I) 

complexes  in  acetonitrile  displayed  peaks  below  450  nm  which  were  a(cid:425)ributed  to  π-π* 

absorp(cid:415)ons from pyridine units. Metal to ligand charge transfer bands were observed between 

60 

 
 
 
450-500  nm  in  copper(I)  complexes,  with  no  signiﬁcant  diﬀerence  between  the  triﬂate  and 

bistriﬂimide variants. 

The [Cu(PY5)]OTf2 complex exhibits two d-d transi(cid:415)ons, the ﬁrst at 597 nm, with an ex(cid:415)nc(cid:415)on 

coeﬃcient of 79.8 M-1s-1, and the second at 909 nm, with an ex(cid:415)nc(cid:415)on coeﬃcient of 13.2 M-1s-1. 

The [Cu(PY5)]TFSI2 complex only exhibits one peak at 598 nm, with an ex(cid:415)nc(cid:415)on coeﬃcient of 

53.0  M-1s-1,  although  the  tailing  on  the  peak  could  indicate  the  overlap  of  two  peaks.  The 

observa(cid:415)on  of  two  d-d  transi(cid:415)ons  is  indica(cid:415)ve  of  either  a  Jahn-Teller  distorted  octahedral 

complex, or a square pyramidal complex. However, the large diﬀerence in wavelengths between 

the  two  transi(cid:415)ons  indicates  that  this  is  a  square  pyramidal  geometry.  The  diﬀerence  in 

absorbance of spectra when the counter-ion is changed is most likely due the lack of acetonitrile 

bound to the axial posi(cid:415)on in the TFSI version of the complex. 

61 

 
 
Figure 4.4: Simulated, red, and experimental, black, EPR spectra of 2 mM [Cu(PY5)]OTf2 in 
either anhydrous acetonitrile, A, or dichloromethane, B. Spectra C shows the diﬀerence 
between the dichloromethane spectra, red, and the acetonitrile spectra, black, when corrected 
for solvent aﬀects. 

Electron  Paramagne(cid:415)c  Resonance  spectra  were  taken  in  an  a(cid:425)empt  to  elucidate  the  solu(cid:415)on 

geometry of the copper(II) form of the redox couple. Two spectra were taken of the [Cu(PY5)]OTf2 

complex,  one  of  the  complex  synthesized  and  measured  in  acetonitrile  and  the  other  in 

dichloromethane. The EPR showed that in both systems the complex is axially elongated which is 

indica(cid:415)ve  of  the  complex  having  an  elongated  octahedral,  square  pyramidal,  or  square  planer 

geometry.  The  two  systems  have  similar  g-value  but  the  hyperﬁne  coupling  is  quite  diﬀerent 

62 

 
 
 
between the two, indica(cid:415)ng that the complexes are similar in geometry but not iden(cid:415)cal. When 

the spectra were adjusted to account for the eﬀects of using two diﬀerent solvents, it becomes 

easy to see that the species in solu(cid:415)on are not iden(cid:415)cal. However, it remains unclear what the 

complex’s exact makeup in solu(cid:415)on is. 

Table 4.1: EPR Parameters of [Cu(PY5)]OTf2 complexes with liquid solu(cid:415)ons in acetonitrile (ACN) 
or dichloromethane (DCM) at room temperature. All parameters were simulated using MATLAB. 

g parallel 

g perpendicular 

hyperﬁne parallel (G) 

hyperﬁne perpendicular (G) 

In ACN 

2.28 

In DCM 

2.30 

2.04 

2.04 

437 

372 

42.6 

39.8 

To  inves(cid:415)gate  poten(cid:415)al  interac(cid:415)ons  between  the  triﬂate  counterion  and  the  copper  center  in 

solu(cid:415)on, a variable temperature  19F-NMR experiment was conducted. The experiment involved 

cooling  the  solu(cid:415)on  from  room  temperature  to  -40°C.  Upon  lowering  the  temperature,  the 

ﬂuorine peak corresponding to the triﬂate counterion exhibited a decrease in intensity coupled 

with an increase in sharpness. This observa(cid:415)on suggests that the counter-ion does not undergo 

exchange with the axial site on the copper center, implying that the counter-ion doesn't interact 

with the copper center in solu(cid:415)on. Such an exchange would typically result in broadening of the 

peak  as  the  temperature  decreases.  Notably,  this  trend  remained  consistent  irrespec(cid:415)ve  of 

whether acetonitrile or dichloromethane was used as the solvent. When this evidence is compiled 

with the previous results of the UV-Vis, cyclic voltammetry, and EPR experiments it becomes clear 

that the solu(cid:415)on geometry of the [Cu(PY5)]2+/+ complexes a square pyramidal where the axial site 

is vacant. 

63 

 
 
 
 
2100000

2000000

1900000

1800000

1700000

1600000

1500000

1400000

1300000

1200000

1100000

1000000

900000

800000

700000

600000

500000

400000

300000

200000

100000

0

-100000

-200000

10

0

-10

-20

-30

-40

-50

-60

-70

-80

-90

-100
f1 (ppm )

-110

-120

-130

-140

-150

-160

-170

-180

-190

-200

-210

Figure 4.5: 19F-NMR spectra of [Cu(PY5)]OTf2 in anhydrous deuterated acetonitrile at various 
temperatures, 25°C-red, 0°C-green, -20°C-blue, and -40°C-purple. 

64 

 
 
 
150000

100000

50000

0

-50000

-100000

10

0

-10

-20

-30

-40

-50

-60

-70

-80

-90

-100
f1 (ppm )

-110

-120

-130

-140

-150

-160

-170

-180

-190

-200

-210

Figure 48.6: 19F-NMR spectra of [Cu(PY5)]OTf2 in deuterated dichloromethane at various 
temperatures, 25°C-red, 0°C-green, -20°C-blue, and -40°C-purple. 

In order to understand how the dynamic nature of the interac(cid:415)on with the open axial site the 

self- exchange kine(cid:415)cs of both systems were inves(cid:415)gated. First the [Cu(PY5)]OTf complex was 

inves(cid:415)gated u(cid:415)lizing stopped-ﬂow spectroscopy. Stopped-ﬂow spectroscopy was u(cid:415)lized to 

measure the cross-exchange electron transfer rate constant, k12, between [Cu(PY5)]OTf2 and 

octamethylferrocene (FcMe8).  

[𝐶𝑢(𝑃𝑌5)](cid:2870)(cid:2878) + 𝐹𝑐𝑀𝑒8 

     (cid:3038)(cid:3117)(cid:3118)      
(cid:4653)⎯⎯⎯(cid:4654)   [𝐶𝑢(𝑃𝑌5)](cid:2878) + 𝐹𝑐𝑀𝑒8(cid:2878)          (4.1) 

65 

 
 
 
Figure 4.7: An absorbance vs (cid:415)me plot at 450 nm, showing the increase of [Cu(PY5)]OTf2 species 
(black dot) and ﬁ(cid:427)ng (red line) for the reduc(cid:415)on of [Cu(PY5)]OTf2 by octamethylferrocene 
(FcMe8). 

Octamethylferrocene  was  chosen  for  the  cross-exchange  to  op(cid:415)mize  the  driving  force  and 

because it has a well-known self-exchange rate constant for electron transfer.90 Due to the large 

poten(cid:415)al diﬀerence between [Cu(PY5)]2+/+ and [FcMe8]+/0, it was assumed that the reac(cid:415)on went 

to comple(cid:415)on, with no signiﬁcant back reac(cid:415)on. Figure 4.7 shows a ﬁt of the absorbance at 450 

nm vs (cid:415)me plot, which represents the growth of the [Cu(PY5)]+ species in solu(cid:415)on  due to the 

reduc(cid:415)on of [Cu(PY5)]2+ by FcMe8, using the following equa(cid:415)on: 

𝐴 =   𝐴(cid:2998) + (𝐴(cid:3016) − 𝐴(cid:2998))𝑒(cid:2879)(cid:3038)(cid:3290)(cid:3277)(cid:3294)(cid:3047) 

          (4.2) 

66 

 
 
 
For all reac(cid:415)ons, FcMe8 was held in excess to maintain pseudo-ﬁrst order condi(cid:415)ons, allowing 

kobs to be represented by:  

𝑘(cid:3042)(cid:3029)(cid:3046) =   𝑘(cid:2869)(cid:2870)[𝐹𝑐𝑀𝑒8]                   (4.3) 

A linear ﬁt of the ﬁ(cid:425)ed kobs values vs concentra(cid:415)on of FcMe8 provided the value of the cross-

exchange  rate  constant,  k12,  from  the  slope  of  the  line.  The  ini(cid:415)al  reac(cid:415)on  mixtures  and  the 

pseudo-ﬁrst order observed rate constants can be found in Table 4.2.  

Table 4.2: The ini(cid:415)al reac(cid:415)on mixture and the observed rate constant, kobs, for the cross 
exchange between [Cu(PY5)]OTf2 and FcMe8. 

[Cu(PY5)(OTf)2] / M 

[FcMe8] / M 

2.74E-05 

2.58E-04 

3.10E-04 

3.61E-04 

4.13E-04 

4.65E-04 

Using the experimentally determined cross-exchange rate constant from Equa(cid:415)on 4.3, k12, and 

the previously determined self-exchange constant for [FcMe8]+/0, k11, the Marcus cross-rela(cid:415)on 

𝑘(cid:2869)(cid:2870) = (𝑘(cid:2869)(cid:2869)𝑘(cid:2870)(cid:2870)𝐾(cid:2869)(cid:2870)𝑓(cid:2869)(cid:2870))(cid:2869)/(cid:2870)𝑊(cid:2869)(cid:2870)       (4.4) 
was  used  to  calculate  the  self-exchange  rate  constant  for  [Cu(PY5)]2+/+,  k22.  Where  K12  is  the 

equilibrium constant, f12 is a nonlinear correc(cid:415)on term, and W12 is an electrosta(cid:415)c work term for 

bringing  the  reactants  into  contact.  The  nonlinear  correc(cid:415)on  term  and  the  work  term  were 

calculated  (see  below  for  details  on  the  calcula(cid:415)on)  and  determined  to  be  2.60  and  0.99, 

respec(cid:415)vely as seen in Table 4.3. However, it should be noted that there will be some error in the 

calcula(cid:415)on  of  the  work  term  since  the  Debye-Huckel  model  is  not  expected  to  have  accurate 

67 

 
 
 
 
 
 
 
results when the solu(cid:415)on has a high ionic strength, as was the case for this experiment (0.1 M 

suppor(cid:415)ng electrolyte). The equilibrium constant, K12, can be determined using Equa(cid:415)on 4.5:  

𝑛𝐹𝛥𝐸 = 𝑅𝑇𝑙𝑛𝐾(cid:2869)(cid:2870)                         (4.5) 

where n is the number of electrons transferred,  F is Faraday’s constant, ΔE is the diﬀerence in 

formal poten(cid:415)al between the oxidant and reductant in solu(cid:415)on, R is the ideal gas constant, and 

T is the temperature in Kelvin. The redox poten(cid:415)al for [Cu(PY5)]2+/+  was found to be -0.372 V vs 

ferrocene91 and the redox poten(cid:415)al of [FcMe8]+/0 has been previously determined to be -0.377 V 

vs ferrocene, using cyclic voltammetry which gave a calculated K12 of 0.62. The self-exchange rate 

constant for [FcMe8]+/0 was previously iden(cid:415)ﬁed to be 2.01 × 107 M-1s-1  using NMR techniques.92 

Using these values, the self-exchange rate for the [Cu(PY5)]OTf1/2 couple was determined to be 

88.1 (±7.3) M-1s-1.  

The  work  required  to  move  the  reactant  complexes  to  a  distance,  r,  for  the  electron  transfer 

reac(cid:415)on was calculated using the Equa(cid:415)on 4.6.  

𝑊(cid:2869)(cid:2870) =  exp [−

(cid:3050)(cid:3117)(cid:3118)(cid:2878)(cid:3050)(cid:3118)(cid:3117)(cid:2879)(cid:3050)(cid:3117)(cid:3117)(cid:2879)(cid:3050)(cid:3118)(cid:3118)
(cid:2870)(cid:3019)

] 

   (4.6) 

𝑤(cid:3036)(cid:3037)(𝑟) =

(cid:3053)(cid:3284)(cid:3053)(cid:3285)(cid:3044)(cid:3118)(cid:3015)(cid:3250)
(cid:2872)(cid:3095)(cid:3084)(cid:3116)(cid:3084)(cid:3045)((cid:2869)(cid:2878)(cid:3081)(cid:3045))  

(4.7) 

Equa(cid:415)on  4.7  was  used  to  determine  the  work  associated  with  the  forward  and  reverse  cross-

exchange reac(cid:415)on, w12 and w21 respec(cid:415)vely, and the self-exchange reac(cid:415)ons, w11 and w22. In this 

equa(cid:415)on zi and zj are the charges of the interac(cid:415)ng complexes, q is the charge of an electron, NA 

is Avogadro’s constant, ε0 is the permi(cid:427)vity of free space, ε is the sta(cid:415)c dielectric of the medium, 

𝛽 = (cid:4672)

(cid:2870)(cid:3044)(cid:3118)(cid:3015)(cid:3250)(cid:3010)
(cid:2869)(cid:2868)(cid:2868)(cid:2868)(cid:3084)(cid:3116)(cid:3084)(cid:3038)(cid:3251)(cid:3021)

(cid:2869)/(cid:2870)
(cid:4673)

, I is the ionic strength of the solu(cid:415)on and kB is the Boltzmann’s constant. The 

calcula(cid:415)on has the following assump(cid:415)ons: the work is assumed to be primarily Coulombic, the 

68 

 
 
 
 
 
 
distance,  r,  is  assumed  to  be  the  center-to-center  distance  between  the  complexes,  and  the 

reactants are assumed to be spherical.  

The non-linear correc(cid:415)on term, f12, was calculated using the equa(cid:415)on 

𝑙𝑛𝑓(cid:2869)(cid:2870) =

(cid:2869)

(cid:2872)

(cid:4672)(cid:3039)(cid:3041)(cid:3012)(cid:3117)(cid:3118)(cid:2878)
(cid:3286)(cid:3117)(cid:3117)(cid:3286)(cid:3118)(cid:3118)
(cid:3275)(cid:3118)

(cid:2922)(cid:2924)(cid:4672)

(cid:3118)

(cid:4673)

(cid:3298)(cid:3117)(cid:3118)(cid:3127)(cid:3298)(cid:3118)(cid:3117)
(cid:3267)(cid:3269)
(cid:3298)(cid:3117)(cid:3117)(cid:3126)(cid:3298)(cid:3118)(cid:3118)
(cid:3267)(cid:3269)

(cid:4673)(cid:2878)

(4.8) 

where K12 is the equilibrium constant, w12 and w21 are the work associated with the forward and 

reverse  cross-exchange  reac(cid:415)on  respec(cid:415)vely,  R  is  the  ideal  gas  constant,  T  is  temperature  in 

Kelvin, and Z is the frequency factor, which is assumed to be 1013 M-1s-1 due to the larger inner-

sphere contribu(cid:415)ons to the total reorganiza(cid:415)on energy.  

Table 4.3: Summary of the kine(cid:415)c data used to calculate the self-exchange rate constant of the 
[Cu(PY5)]OTf1/2 complex by monitoring the reac(cid:415)on of [Cu(PY5)]OTf2 and FcMe8 in acetonitrile 
containing 0.1 M LiOTf at 25°C. 

Kine(cid:415)c Parameter  Values 

K12 

0.62 

k12 / M-1s-1 

4.53E+04 

k22 / M-1s-1 

2.01E+07 

f12 

W12 

k11 / M-1s-1 

λse / eV 

λi / eV 

λo / eV 

1 

2.6 

88.1 

2.57 

1.55 

1.02 

The self-exchange of fast exchanging redox complexes can be found by inves(cid:415)ga(cid:415)ng the methyl 

peaks  in  the  1H-NMR  spectra  of  the  pure  diamagne(cid:415)c  [Cu(PY5)]TFSI  complex,  the  pure 

69 

 
 
  
 
 
 
paramagne(cid:415)c  [Cu(PY5)]TFSI2  complex,  and  mixtures  of  the  diamagne(cid:415)c  and  paramagne(cid:415)c 

species.  All  peaks  were  ﬁ(cid:425)ed  to  a  Lorentzian/Gaussian  lineshape.  The  self-exchange  rate 

constant, kse,  was  calculated from  

𝑘(cid:3046)(cid:3032) =  

(cid:2872)(cid:3095)(cid:3076)(cid:3279)(cid:3076)(cid:3291)((cid:3057)(cid:3092))(cid:3118)
(cid:3435)(cid:3024)(cid:3279)(cid:3291)(cid:2879)(cid:3076)(cid:3291)(cid:3024)(cid:3291)(cid:2879)(cid:3076)(cid:3279)(cid:3024)(cid:3279)(cid:3439)(cid:3004)

(4.9) 

where Wdp is the line width (full width at half-maximum) of the mixed species methyl resonance 

peak, Wp and Wd are the line widths of the paramagne(cid:415)c and diamagne(cid:415)c peaks, respec(cid:415)vely, Xp 

and Xd are the mole frac(cid:415)ons of the paramagne(cid:415)c and diamagne(cid:415)c species, respec(cid:415)vely, C is the 

total concentra(cid:415)on of the exchanging species, and Δν is the observed frequency shi(cid:332) rela(cid:415)ve to 

the posi(cid:415)on of the resonance for the diamagne(cid:415)c species. The diamagne(cid:415)c and paramagne(cid:415)c 

line widths and the frequency shi(cid:332) are listed in Table 4.4. This analysis produced a kse of 389 M-

1s-1 for [Cu(PY5)]TFSI1/2 in deuterated acetonitrile at 22°C.  

Table 4.4: List of parameters used to calculate the self-exchange of [Cu(PY5)]TFSI1/2. 

Wp 
Wd 
Xd 
Xp 
ΔV 
Wdp 
Kex 

102.27 
11.3 
0.404249 
0.595751 
22.562 
204.66 
389.0535 

The total reorganiza(cid:415)on energy for each of the self-exchange reac(cid:415)ons was calculated following 

the equa(cid:415)on 

𝑘(cid:2869)(cid:2869) = 𝐾(cid:2869)(cid:2870)𝑍𝛤𝑒(cid:2879)((cid:3090)(cid:3117)(cid:3117) (cid:2872)(cid:3038)(cid:3251)(cid:3021))

⁄

(4.10) 

where k11 is the previously calculated self-exchange rate, K12 is the equilibrium constant, Z is the 

frequency factor, which is assumed to be 1013 M-1s-1 due to the larger inner-sphere contribu(cid:415)ons 

70 

 
 
  
 
 
 
 
to the total reorganiza(cid:415)on energy, Γ is a correc(cid:415)on for nuclear tunneling, which is assumed to be 

~1,  λ11  is  the  total  reorganiza(cid:415)on  energy  of  the  self-exchange  reac(cid:415)on,  kB  is  the  Boltzmann 

constant, and T is the temperature of solu(cid:415)on in Kelvin. The resul(cid:415)ng λ11 was found to be 2.57 eV 

for the OTf complex and 2.41 for the TFSI complex.  

The outer-sphere reorganiza(cid:415)on energy for two spherical reactants can be determined using 

𝜆(cid:2869)(cid:2869),(cid:3042) =

((cid:3057)(cid:3053)(cid:3044))(cid:3118)
(cid:2872)(cid:3095)(cid:3084)(cid:3116)

(cid:4672)

(cid:2869)
(cid:2870)(cid:3028)(cid:3117)

+

(cid:2869)
(cid:2870)(cid:3028)(cid:3118)

−

(cid:2869)
(cid:3045)(cid:3117)(cid:3117)

(cid:4673) (cid:3436)

(cid:2869)
(cid:3005)(cid:3290)(cid:3291),(cid:3294)(cid:3290)(cid:3287)

−

(cid:2869)
(cid:3005)(cid:3294),(cid:3294)(cid:3290)(cid:3287)

(cid:3440) 

(4.11) 

where λ11,o is the outer-sphere reorganiza(cid:415)on energy, Δz is the change in charge of the complex, 

q is the charge of an electron, ε0 is the permi(cid:427)vity of free space, a1 and a2 are the atomic radii of 

the  reactants,  r11  is  the  center−center  distance  between  the  reactants,  Dop,sol  is  the  op(cid:415)cal 

dielectric  constant  of  the  medium,  which  is equal  to  the  square of  the  refrac(cid:415)ve  index  of  the 

medium, and Ds,sol is the sta(cid:415)c dielectric constant of the medium. The outer-sphere reorganiza(cid:415)on 

energy for the self-exchange reac(cid:415)on was calculated to be 1.02 eV for the OTf complex and 1.01 

eV for the TFSI complex. The inner-sphere reorganiza(cid:415)on energy was determined using 𝜆(cid:2869)(cid:2869),(cid:3036) =

𝜆(cid:2869)(cid:2869) − 𝜆(cid:2869)(cid:2869),(cid:3016) yielding a result of 1.55 eV for the OTf complex and 1.40 eV for the TFSI complex for 

the inner-sphere reorganiza(cid:415)on energy of the self-exchange reac(cid:415)on.  

71 

 
 
 
Figure 4.8: A) Cyclic Voltammogram of 2 mM [Cu(PY5)](OTf)2 with 0.1 M TBAPF6 in anhydrous 
acetonitrile with increasing concentra(cid:415)on of 4-tert-butylpyridine (TBP) at a glassy carbon 
working electrode; scan rate = 100 mVs-1. B) UV-Vis absorp(cid:415)on spectra of 2 mM [Cu(PY5)](OTf)2 
in anhydrous acetonitrile with increasing equivalents of TBP. C) 1H-NMR spectra of 
[Cu(PY5)](OTf)2 in anhydrous acetonitrile with increasing equivalents of TBP. Unbound PY5 and 
TBP are shown as controls. 

Lewis bases such as TBP are used as addi(cid:415)ves in DSSCs to increase the performance of the DSSC 

devices.35–37  Several  copper  based  redox  shu(cid:425)les  have  been  shown  to  undergo  a  ligand 

subs(cid:415)tu(cid:415)on with the TBP, even polydentate ligands.38,39 This can have a signiﬁcant impact on the 

72 

 
 
 
redox  poten(cid:415)al  and  kine(cid:415)cs  of  the  system  as  mul(cid:415)ple  redox  systems  are present  at  any given 

(cid:415)me. Looking at the cyclic voltammetry data, Upon the addi(cid:415)on of TBP to the solu(cid:415)on containing 

[Cu(PY5)]OTf2, both of the redox waves go away and a new wave grows in at -0.512 V vs ferrocene, 

indica(cid:415)ng that the [Cu(PY5)]OTf2 complex is reac(cid:415)ng with TBP. The new peak con(cid:415)nues to grow 

in as TBP is added un(cid:415)l about 10 equivalents rela(cid:415)ve to the amount of copper(II) in the solu(cid:415)on, 

indica(cid:415)ng that whatever the change that occurred has fully stabilized by 10 equivalents added. 

As  such  the  amount  of  TBP  used  in  the  devices  in  this  study  was  10  equivalents,  in  order  to 

minimize mixed species in solu(cid:415)on.  

When the addi(cid:415)on of TBP is monitored via UV-Vis spectroscopy, there is a no(cid:415)ceable blue shi(cid:332) of 

23 nm for the peak at ~600 nm, which also comes with an almost 50% increase in absorbance, 

and a blue shi(cid:332) of 58 nm for the peak at ~900 nm, with minimal change in absorbance. Once 10 

equivalents  of  TBP  is  added,  rela(cid:415)ve  to  the  copper(II)  in  solu(cid:415)on,  the  reac(cid:415)on  seems  to  have 

reached  equilibrium  and  the  peaks  no  longer  shi(cid:332).  This  agrees  with  the  previous  cyclic 

voltammetry study indica(cid:415)ng that there is a reac(cid:415)on occurring which has reached equilibrium by 

10 equivalents of TBP added to the solu(cid:415)on.  Although it should be noted that when the reac(cid:415)on 

has stabilized the resul(cid:415)ng spectra does not match the spectra of [Cu(TBP)4]OTf2, meaning that 

whatever is formed it is not the ligand subs(cid:415)tuted product. 

In  an  a(cid:425)empt  to  understand  what  the  interac(cid:415)on  of  the  [Cu(PY5)]OTf2  complex  and  TBP  was 

producing,  the  reac(cid:415)on  was  monitored  via  1H-NMR.  Upon  the  addi(cid:415)on  of  TBP  to  the 

[Cu(PY5)]OTf2 solu(cid:415)on a change in the spectra is seen  as soon as a single equivalent of TBP is 

added.  However,  a(cid:332)er  that  ini(cid:415)al  change  in  the  spectra,  the only  change  in  the  spectra  is  the 

broad peaks that seem to correspond to TBP. The broadness of the peaks seems to indicate that 

73 

 
 
there is a rapid exchange on the NMR (cid:415)mescale of the TBP and the copper center. However, the 

spectra show no signs of unbound ligand in the solu(cid:415)on as TBP added, sugges(cid:415)ng the PY5 ligand 

remains bound to the copper center. This indicates that the interac(cid:415)on being seen likely stems 

from TBP displacing the labile axial acetonitrile, resul(cid:415)ng in a [Cu(PY5)(TBP)]2+ copper complex, 

rather than the displacement of the ligand. It is also worth no(cid:415)ng that it has been posited that 

the  interac(cid:415)on  of  TBP  with  copper  complexes  may  not  necessarily  be  undergoing  a  full 

subs(cid:415)tu(cid:415)on, instead it may be just a coordina(cid:415)on of the TBP to the copper center causing the 

change in proper(cid:415)es, similar to what is being seen with [Cu(PY5)]OTf2.39,93 

CuPY5(OTf) without TBP
CuPY5(OTf) with TBP
CuPY5(TFSI) without TBP
CuPY5(TFSI) with TBP

2
-

m
c
A
m
/
t
n
e
r
r
u
C

10
9
8
7
6
5
4
3
2
1
0

-0.5

-0.4

-0.3

-0.2

-0.1

0

Voltage / V

Figure 4.9: J−V curves for the [Cu(PY5)]OTf1/2, blue, and [Cu(PY5)]TFSI1/2, red, based DSSC 
devices. The devices represented by the pale colored curves contain TBP while the dark colored 
curves are without TBP. 

74 

 
 
 
 
 
Table 4.5: Summary of device performance of devices comparing [Cu(PY5)]OTf and 
[Cu(PY5)]TFSI, parameters represent the average of ﬁve devices. 

η / % 

JSC / mAcm-2 

VOC / V 

FF 

CuPY5(OTf) 

1.30(±0.25) 

8.22(±1.16) 

0.26(±0.02) 

0.60(±0.01) 

CuPY5(OTf) with 
TBP 

1.40(±0.30) 

6.22(±1.97) 

0.35(±0.05) 

0.66(±0.04) 

CuPY5(TFSI) 

1.01(±0.35) 

6.57(±2.18) 

0.26(±0.01) 

0.59(±0.02) 

CuPY5(TFSI) with 
TBP 

1.33(±0.09) 

5.84(±0.49) 

0.36(±0.03) 

0.64(±0.01) 

DSSC devices were fabricated with the commercial dye, Y123, being used as the light-harvester, 

and where poly(3,4-ethylenedioxythiophene) (PEDOT) was u(cid:415)lized at the counter electrode. The 

devices u(cid:415)lized either [Cu(PY5)]OTf1/2 or [Cu(PY5)]TFSI1/2 as the redox shu(cid:425)le used, and for each 

redox  shu(cid:425)le  a  batch  was  fabricated  with  10  equivalents  of  TBP  rela(cid:415)ve  to  the  amount  of 

copper(II)  in  solu(cid:415)on,  and  another  batch  without  the  addi(cid:415)on  of  TBP.  The  devices  yielded 

sta(cid:415)s(cid:415)cally similar results for the JSC, around 6.71 mAcm-2, and eﬃciencies, around 1.26% for all 

the batches.  The  diﬀerence between the cells with  and  without TBP is no(cid:415)ceable in the open 

circuit voltage (VOC) where the devices with TBP have an increase of around 100 mV compared to 

their counterparts without TBP. While the devices with and without TBP show diﬀerences from 

one another the devices with diﬀering redox mediator showed sta(cid:415)s(cid:415)cally iden(cid:415)cal performance 

when compared, as shown in Table 4.5.  

75 

 
 
 
 
 
 
This similarity in performance across the diﬀerent redox shu(cid:425)les indicates that even though there 

was a diﬀerence in both the redox and kine(cid:415)c proper(cid:415)es of the two versions of the redox shu(cid:425)le, 

the choice of counter-ion did not signiﬁcantly aﬀect device characteris(cid:415)cs. When looking at the 

solu(cid:415)on poten(cid:415)al of the devices it can be seen that the solu(cid:415)on poten(cid:415)al is similar between the 

OTf and TFSI versions of the complex, -0.372 V and -0.398 V vs ferrocene respec(cid:415)vely, which for 

both complexes is slightly nega(cid:415)ve of the Nerns(cid:415)an poten(cid:415)al of -0.357 and -0.365 respec(cid:415)vely. 

The predominate form of the redox shu(cid:425)le that seems to be controlling the solu(cid:415)on poten(cid:415)al of 

the devices is the ﬁve-coordinate redox couple represented by the wave at ~-0.375 V vs ferrocene, 

with minimal contribu(cid:415)on from the acetonitrile bound wave at -0.662 V. When TBP is added to 

the devices the solu(cid:415)on poten(cid:415)al nega(cid:415)vely shi(cid:332)s to -0.496 V and -0.501 V vs ferrocene for the 

OTf and TFSI complexes respec(cid:415)vely, which matches well with the predicted -0.503 V vs ferrocene 

for the [Cu(PY5)(TBP)]OTf1/2 complex. This was expected due to the binding of TBP to the copper 

center,  so  both  versions of  the  complex would  form  the  [Cu(PY5)(TBP)]2+ complex.  Due  to  the 

similarity of the device results between the varia(cid:415)ons of the complex only [Cu(PY5)]OTf1/2 was 

used for subsequent experiments.  

76 

 
 
12

10

8

6

4

2

0

2
-

m
c
A
m
/
t
n
e
r
r
u
C

-2

-0.5

0 M TBP

0.055 M TBP

0.526 M TBP

-0.4

-0.3

-0.2

-0.1

0

Voltage / V

Figure 4.10: J-V curves of best performing devices containing [Cu(PY5)]OTf1/2 in dry acetonitrile 
and 0 M TBP (green), 0.055 M TBP (blue), 0.211 M TBP (gray), 0.526 M TBP (yellow). 

Table 4.6: Summary of device performance of devices containing [Cu(PY5)]OTf with various 
equivalents of TBP added, parameters represent the average of ﬁve devices. 

η / % 

Jsc / mA*cm-2 

Voc / V 

FF 

0 M TBP  2.00(±0.06)  8.40(±0.09) 

0.39(±0.00) 

0.61(±0.02) 

0.055 M 
TBP 

0.526 M 
TBP 

2.12(±0.10)  8.91(±0.39) 

0.41(±0.02) 

0.58(±0.04) 

2.01(±0.24)  7.81(±0.63) 

0.45(±0.01) 

0.57(±0.03) 

The eﬀect of the addi(cid:415)on of TBP in the DSSC devices was inves(cid:415)gated via fabrica(cid:415)ng devices with 

various  concentra(cid:415)ons  of  TBP  used.  All  of  the  devices  had  an  improved  power  conversion 

eﬃciency of around 2.04%,  which  can be  a(cid:425)ributed  to both  an increased short circuit current 

(JSC), around 8.38 mAcm-2, and a signiﬁcantly increased open circuit voltage (VOC), which is over 

100 mV improved over the previous a(cid:425)empt. However, it can be noted that the current of the 

devices  were  hampered  by  mass  transport  issues.  The  incident  photon-to-current  conversion 

eﬃciency (IPCE) spectra was measured, which measures the current at various wavelengths to 

77 

 
 
 
 
 
 
 
 
determine how much of the photons absorbed at each wavelength are converted to current. The 

IPCE  in  ﬁgure  4.12  shows  that  as  TBP  is  added  there  is  a  general  trend  towards  an  increased 

photon-to-current  conversion  between  400  and  650  nm.  This  increase  could  be  caused  by 

decreased recombina(cid:415)on to the redox shu(cid:425)le as TBP is being  added to the solu(cid:415)on. The IPCE 

spectra can also be integrated to determine what the theore(cid:415)cal current of the device is, with 

the  results  listed  in  table  4.7,  and  all  of  the  devices  show  higher  theore(cid:415)cal  current  than  the 

measure short circuit current, ge(cid:427)ng as high as 10 mAcm-2 when 0.526 M of TBP is added to the 

solu(cid:415)on. The high theore(cid:415)cal current and the nega(cid:415)ve redox poten(cid:415)al make this redox shu(cid:425)le a 

candidate for u(cid:415)liza(cid:415)on with narrow band gap dyes for high photocurrent DSSC devices.   

78 

 
 
%
/
E
C
P
I

90

80

70

60

50

40

30

20

10

0

0 M TBP

0 M TBP

0.055 M TBP

0.526 M TBP

0.055 M TBP

0.526 M TBP

12

10

2
-
m
c
A
m
/
t
n
e
r
r
u
C

8

6

4

2

0

400

450

500

550
Wavelength / nm

600

650

700

Figure 4.11: IPCE curves for the best performing [Cu(PY5)]OTf1/2 based DSSC devices with 
various amounts of TBP added. The devices represented by the green diamonds and curve 
contain 0 M TBP added, the blue square and curve represent 0.055 M TBP added, the gray 
triangles and curve represent 0.211 M TBP added, and the yellow circles and curve represent 
0.526 M TBP added. The points represent IPCE responses while the solid lines represent the 
integrated photocurrent. 

79 

 
 
 
 
 
 
 
 
 
 
 
Table 4.7: The measured current and the integrated current found via IPCE for devices 
containing [Cu(PY5)]OTf1/2 and various equivalents of TBP added. 

Jsc / mA*cm-2 

Integrated Current / 

0 M TBP 

8.40(±0.09) 

0.055 M TBP 

8.91(±0.39) 

0.526 M TBP 

7.81(±0.63) 

mA*cm-2 

8.92 

9.82 

10.64 

The  solu(cid:415)on  poten(cid:415)al  of  the  [Cu(PY5)]OTf1/2  solu(cid:415)on  without  TBP  was  -0.374  V  vs  ferrocene 

matching well with the calculated Nerns(cid:415)an poten(cid:415)al of -0.357 V vs ferrocene. Upon the addi(cid:415)on 

of 0.526 M TBP the solu(cid:415)on poten(cid:415)al shi(cid:332)ed to -0.490 V vs ferrocene as the TBP bound to the 

copper center, the calculated poten(cid:415)al was -0.503 V vs ferrocene. This shi(cid:332) in redox poten(cid:415)al and 

the change in the open circuit poten(cid:415)al from 0.39 V to 0.45 V as TBP is added to the solu(cid:415)on 

indicates that the fermi level of the devices increased from -0.764 V vs ferrocene to -0.940 V vs 

ferrocene. The addi(cid:415)on of TBP seems to have shi(cid:332)ed the fermi level ~200 mV which is a small 

shi(cid:332) 

in  the  fermi 

level  when  compared  to  the  [Cu(bpyPY4)]TFSI1/2  complex  and  the 

[Cu(dbmed)]OTf1/2 complex. When the electron life(cid:415)me is compared to the previous complexes 

it  can  be  seen  that  the  life(cid:415)me  of  the  [Cu(PY5)]OTf1/2  complex  is  similar  to  that  of  the 

[Cu(bpyPY4)]TFSI1/2 complex. The addi(cid:415)on of TBP shows an increase in the electron life(cid:415)me by a 

factor of 10. This is likely due to a combina(cid:415)on of low driving force for recombina(cid:415)on for both the 

complexes  due  to  their  nega(cid:415)ve  redox  poten(cid:415)al,  and  their  rela(cid:415)vely  slow  electron  transfer 

kine(cid:415)cs.   

80 

 
 
 
 
 
10

1

0.1

s
/
1
/
e
m

i
t
e
f
i
L

0.01

0

[Cu(dbmed)] without TBP

[Cu(bpyPY4)] without TBP

[Cu(PY5)] without TBP

[Cu(dbmed)] with TBP

[Cu(bpyPY4)] with TBP

[Cu(PY5)] with TBP

0.2

0.4

0.6

0.8

1

1.2

Voltage / V

Figure 4.12: A plot of the electron life(cid:415)me vs poten(cid:415)al plot for DSSC devices using 
[Cu(dbmed)]OTf1/2 (green), [Cu(bpyPY4)]TFSI1/2 (red), [Cu(PY5)]OTf1/2 (blue), with (pale color) 
and without (dark color) the addi(cid:415)on of TBP. 

4.4 Conclusion 

A  copper-based  redox  shu(cid:425)le  featuring  a  pentadentate  polypyridyl  ligand  with  a  labile  axial 

posi(cid:415)on  was developed for use  in DSSC devices. This complex was studied using two diﬀerent 

counter  ions  which  provide  diﬀerent  redox  and  kine(cid:415)c  proper(cid:415)es  for  the  copper  complex, 

including a 4-fold increase in self-exchange rate. Despite these diﬀerences in kine(cid:415)cs and redox 

behavior,  both  variants  of  the  copper  complex  showed  iden(cid:415)cal  performance  in  DSSC 

applica(cid:415)ons.  In  solu(cid:415)on,  the  copper(II)  complexes  predominantly  adopted  a  ﬁve-coordinate 

geometry  with  a  vacant  axial  site.  Further  analysis  revealed  that  while  there  is  an  interac(cid:415)on 

between the copper complex and TBP, this interac(cid:415)on is limited to the subs(cid:415)tu(cid:415)on at the labile 

axial posi(cid:415)on by the base, rather than involving the ligand itself. Notably, the addi(cid:415)on of TBP to 

81 

 
 
 
 
 
the DSSC’s electrolyte led to an increase in device performance. This suggests that the presence 

of  a  Lewis  base,  even  when  reac(cid:415)ng  with  the  copper  complexes,  is  beneﬁcial  for  the  overall 

eﬃciency of DSSCs. 

82 

 
 
 
 
Chapter 5 – Developing a model to predict ligand exchange in DSSCs 

5.1 Introduc(cid:415)on 

In the wake of Grätzel's pioneering 1993 report on achieving a 10% eﬃcient dye-sensi(cid:415)zed solar 

cell (DSSC) through the use of dye-sensi(cid:415)zed photoelectrodes, further advancements in eﬃciency 

were hampered by the reliance on the triiodide/iodide (I3-/I-) redox shu(cid:425)le.15 A primary constraint 

of  I3-/I-  u(cid:415)liza(cid:415)on  arises  from  the  considerable  overpoten(cid:415)al  required  for  eﬃcient  dye 

regenera(cid:415)on. 

Over  the  last  decade,  the  adop(cid:415)on  of  outer  sphere  redox  shu(cid:425)les  has  enabled  the  design  of 

systems that mi(cid:415)gate this overpoten(cid:415)al, resul(cid:415)ng in performance enhancement.23 However, the 

incorpora(cid:415)on  of  Co(III/II)  based  redox  shu(cid:425)les  has  encountered  challenges  due  to  substan(cid:415)al 

inner-sphere  reorganiza(cid:415)on  energy  accompanying  electron  transfer.  This  complica(cid:415)on,  arising 

from the transi(cid:415)on of a high-spin d7 Co(II) system to a low-spin d6 Co(III) system, directly aﬀects 

dye regenera(cid:415)on kine(cid:415)cs by limi(cid:415)ng the available electron transfer driving force.23 

In contrast, there has been a growing interest in copper-based redox shu(cid:425)les for DSSCs. These 

shu(cid:425)les have achieved record eﬃciencies of 15.2% under 1 sun illumina(cid:415)on and an impressive 

34.5%  under  1000  lux  intensity  ﬂuorescent  ligh(cid:415)ng.20,87  The  success  of  copper-based  redox 

shu(cid:425)les  can  be  a(cid:425)ributed  to  their  diminished  inner-sphere  electron  transfer  reorganiza(cid:415)on 

energies, placing them on par with conven(cid:415)onal I3−/I− and Co(III/II)-based counterparts. 

Copper complexes are versa(cid:415)le, adop(cid:415)ng diverse geometries inﬂuenced by oxida(cid:415)on state and 

ligand environment. Cu(II) d9 complexes typically exhibit six-coordinate, octahedral or tetragonal, 

ﬁve-coordinate,  square  pyramidal  or  trigonal  bipyramidal,  or  four-coordinate,  tetrahedral  or 

83 

 
 
square planar geometries. Similarly, Cu(I) d10 complexes predominantly assume four-coordinate, 

tetrahedral or square planar geometries.60–63 

Most copper redox shu(cid:425)les in DSSCs u(cid:415)lize bidentate bipyridine or phenanthroline ligands with 

bulky ﬂanking groups to minimize structural varia(cid:415)on between Cu(II) and Cu(I) states. Notably, 

recent experimental ﬁndings indicate that commonly employed Lewis base addi(cid:415)ves, like 4-tert-

butylpyridine,  can  coordinate to  Cu(II) species,  poten(cid:415)ally having  a detrimental  impact  on  the 

electrochemical  behavior  and  device  performance.38  This  coordina(cid:415)on-induced  complexity 

inﬂuences redox species and shi(cid:332)s formal redox poten(cid:415)als, limi(cid:415)ng photovoltage. However, not 

all copper redox couples undergo ligand subs(cid:415)tu(cid:415)on.39,41,94,95  

The focus of  this work  lies in harnessing a profound insight contributed by Dr. Rorabacher. Dr. 

Rorabacher demonstrated a strong correla(cid:415)on between the poten(cid:415)al of a copper redox couple 

and the stability constant of its Cu(II) form.96 It was found that the Cu(I) form of copper redox 

shu(cid:425)les tend to have a similar stability constant, where the stability constant is determined by 

the steric strain. Dr. Rorabacher was also able to show that, in aqueous solu(cid:415)ons, the stability 

constant of the Cu(II) form was found to be inversely propor(cid:415)onal  to the redox poten(cid:415)al of the 

Cu(II/I) complex. This rela(cid:415)onship was found to be linear, shown in Figure 5.1, with a slope of -

0.059, matching the electrochemical constant of the Nernst equa(cid:415)on, which indicates that this is 

a fundamental property. The trend was tested against a variety of ligand types from cyclic ligands 

to tripodal ligands, and the trend held true across all the diﬀerent types. This means that ligands 

that bind with copper to form complexes with more nega(cid:415)ve redox poten(cid:415)als should be able to 

subs(cid:415)tute  ligands  bound  to  copper  complexes  with  more  posi(cid:415)ve  redox  poten(cid:415)als.  This 

rela(cid:415)onship  between  redox  poten(cid:415)al  and  stability  constant  presents  a  novel  pathway  for 

84 

 
 
predic(cid:415)ng  ligand  subs(cid:415)tu(cid:415)on  in  copper  complexes,  speciﬁcally  in  the  presence  of  4-tert-

butylpyridine (TBP).  

Figure 5.1: A plot of copper redox poten(cid:415)als versus the log of the stability constant of the 
copper(II) form of the redox complex in aqueous solu(cid:415)ons. The ligands represent samples of 
various ligand types. The solid line has the Nerns(cid:415)an slope of -0.059, indica(cid:415)ng that this 
rela(cid:415)onship follows the fundamental trend.97 

Building upon Dr. Rorabacher's work, our study delves into the interplay between redox poten(cid:415)al 

and  stability  constant.  The  ability  to  predict  ligand  subs(cid:415)tu(cid:415)on  by  TBP,  a  crucial  factor  in  the 

85 

 
 
 
electrochemical behavior of copper complexes, opens the door to engineering redox mediators 

with enhanced performance in DSSCs. 

5.2 Experimental Details 

All  NMR  spectra  were  taken  on  an  Agilent  DirectDrive2  500  MHz  spectrometer  at  room 

temperature and referenced to residual solvent signals. All NMR spectra were evaluated using 

the  MestReNova  so(cid:332)ware  package  features.  Cyclic  voltammograms  were  obtained  using 

µAutolabIII poten(cid:415)ostat using BASi glassy carbon electrode, a pla(cid:415)num mesh counter electrode, 

and a fabricated 0.01 M AgNO3, 0.1 M TBAPF6 in acetonitrile Ag/AgNO3 reference electrode. All 

measurements  were  internally  referenced  to  ferrocene/ferrocenium  couple  via  addi(cid:415)on  of 

ferrocene  to  solu(cid:415)on  a(cid:332)er  measurements  or  run 

in  a  parallel  solu(cid:415)on  of  the  same 

solvent/electrolyte.  UV-Vis  spectra  were  taken  using  a  PerkinElmer  Lambda  35  UV-Vis 

spectrometer using a 1 cm path length quartz cuve(cid:425)e at 480 nm/min. Elemental Analysis data 

were obtained via Midwest Microlab. 

Synthesis of [Cu(dmbpy)2]OTf: 

Copper(I)  tetrakisacetonitrile  triﬂate  (1.0284  g,  2.73  mmol)  was  dissolved  in  minimal  dry 

acetonitrile.  Meanwhile  dmbpy  (1.0605  g,  5.76  mmol)  was  dissolved 

in  minimal  dry 

dichloromethane. The dmbpy solu(cid:415)on was added dropwise to the copper solu(cid:415)on, turning the 

solu(cid:415)on a red  color. Reac(cid:415)on  was s(cid:415)rred  for 30 minutes then precipitated  with dry ether and 

ﬁltered.  The  solvent  was  then  removed  via  vacuum  and  the  Cu(I)(dmbpy)2OTf  product  was 

collected (1.4716 g, 96.1% yield). 1H NMR (500 MHz, CDCl3) δ 8.26 (d, 2H), 8.02 (t, 2H), 7.44 (d, 

2H), 2.22 (s, 6H). 

Synthesis of [Cu(dmbpy)2]OTf2: 

86 

 
 
Copper(II) triﬂate (1.0325 g, 2.85 mmol) was dissolved  in minimal dry acetonitrile. Meanwhile 

dmbpy (1.1255 g, 6.11 mmol) was dissolved in minimal dry dichloromethane. The dmbpy solu(cid:415)on 

was added dropwise to the copper solu(cid:415)on, turning the solu(cid:415)on a dark green color. Reac(cid:415)on was 

s(cid:415)rred  for  30  minutes  then  precipitated  with  dry  ether  and  ﬁltered.  The  solvent  was  then 

removed via vacuum and the Cu(II)(dmbpy)2(OTf)2product was collected (1.7771 g, 85.4% yield).  

Synthesis of N,N(cid:3387)-Dibenzyl-N,N(cid:3387)-bis(6-methylpyridin-2-ylmethyl)ethylenediamine (dbmed): 

2-(Chloromethyl)-6-methylpyridine 

hydrochloride 

(0.5300g, 

3.0 

mmol) 

and 

benzyltriethylammonium chloride (0.0121 g, 0.05 mmol) were added to 20 mL dichloromethane 

and s(cid:415)rred un(cid:415)l fully dissolved.  N,N’-Dibenzylethylenediamine (0.35 mL, 1.48 mmol) was then 

added  to  the  solu(cid:415)on,  a(cid:332)er  which  a  cloudy  white  solu(cid:415)on  was  formed.  Meanwhile  sodium 

hydroxide  (20.1807  g,  500 mmol)  was  added  to  20  mL  of  water  which was  then  added to  the 

dichloromethane solu(cid:415)on and s(cid:415)rred vigorously. A(cid:332)er 2 hours the organic solu(cid:415)on became clear 

and the aqueous solu(cid:415)on turned white. The solu(cid:415)on was then allowed to reﬂux overnight. A(cid:332)er 

reﬂuxing addi(cid:415)onal water was added to the solu(cid:415)on to allow the layers to fully separate then the 

organic layer was then separated from the aqueous layer. The aqueous layer was then extracted 

with  dichloromethane  two  (cid:415)mes,  and  all  the  organic  layers  were  combined  and  dried  with 

magnesium sulfate. Then the solvent was removed under low vacuum. The crude solid was then 

recrystallized  in  acetonitrile  and  dried  to  obtain  the  product  (0.4270  g,  60.1%  yield).  1H-NMR 

(CDCl3 500 MHz) δ 2.50 (s, 6H), 2.67 (s, 4H), 3.56 (s, 4H), 3.67 (s, 4H), 6.96 (d, 2H), 7.25 (m, 12H), 

7.45 (t, 2H).  

Synthesis of [Cu(dbmed)]OTf: 

87 

 
 
Copper(I)  tetrakisacetonitrile  triﬂate  (0.1123  g,  0.30  mmol)  was  dissolved  in  minimal  dry 

acetonitrile.  Meanwhile  dbmed  (0.1487  g,  0.33  mmol)  was  dissolved 

in  minimal  dry 

dichloromethane. The dbmed solu(cid:415)on was added dropwise to the copper solu(cid:415)on, turning the 

solu(cid:415)on a bright yellow color. Reac(cid:415)on was s(cid:415)rred overnight then precipitated with dry ether and 

ﬁltered. The solvent was then removed via vacuum and the Cu(I)dbmed product was collected 

(0.1876 g, 94.9% yield). 1H-NMR (CDCl3 500 MHz) δ 2.67 (s, 2H), 2.72 (s, 6H), 2.82 (s, 2H), 3.57 (s, 

2H), 3.74 (s, 2H), 3.85 (s, 2H), 4.16 (s, 2H), 7.20 (m, 4H), 7.29 (m, 8H), 7.43 (d, 2H), 7.83 (t, 2H). 

Synthesis of [Cu(dbmed)]OTf2: 

Copper(II)  triﬂate  (0.808  g,  2.23  mmol)  was  dissolved  in  minimal  dry  acetonitrile.  Meanwhile 

dbmed (0.1030 g, 2.29 mmol) was dissolved in minimal dry dichloromethane. The dbmed solu(cid:415)on 

was added dropwise to the copper solu(cid:415)on, turning the solu(cid:415)on a deep blue color. Reac(cid:415)on was 

s(cid:415)rred overnight then precipitated with dry ether and ﬁltered. The solvent was then removed via 

vacuum, and the Cu(II)dbmed product was collected (0.1596 g, 87.9% yield).  

Synthesis of [Cu(TBP)4]OTf: 

Copper(I)  tetrakisacetonitrile  triﬂate  (0.2513  g,  0.67  mmol)  was  dissolved  in  minimal  dry 

dichloromethane. 4-tert-butylpyridine  (0.972  mL,  6.63 mmol)  was  then added  to  the  solu(cid:415)on, 

turning the solu(cid:415)on a bright yellow color. Reac(cid:415)on was s(cid:415)rred overnight then precipitated with 

dry ether, forming pale yellow crystals, and the solid was ﬁltered. The product was then dried via 

vacuum and was collected (0.4351 g, 86.% yield).  1H NMR (500 MHz, acetonitrile-d3) δ 8.47 (s, 

2H), 7.44 (s, 2H) 1.33 (s, 9H). 

Synthesis of [Cu(TBP)4]OTf2: 

88 

 
 
Copper(II) triﬂate (0.4135 g, 1.14 mmol) was dissolved in minimal dry acetonitrile. TBP (1.14 mL, 

8.00 mmol) was then added to the solu(cid:415)on, turning the solu(cid:415)on a deep purple color. Reac(cid:415)on 

was  s(cid:415)rred  for  2  hours  then  precipitated  with  dry  ether  and  ﬁltered.  The  reac(cid:415)on  was  then 

recrystallized  using  acetonitrile  and  ether.  The  product  was  then  dried  via  vacuum  and  was 

collected (0.9043 g, 87.5% yield).  

Synthesis of [Cu(bpyPY4)]OTf: 

Copper(I)  tetrakisacetonitrile  triﬂate  (0.1647  g,  0.43  mmol)  was  dissolved  in  minimal  dry 

acetonitrile. Meanwhile 6,6(cid:3387)-bis(1,1-di(pyridin-2-yl)ethyl)-2,2(cid:3387)-bipyridine (bpyPY4) (0.2545 g, 0.49 

mmol) was dissolved in minimal dry dichloromethane. The bpyPY4 solu(cid:415)on was added dropwise 

to the copper solu(cid:415)on, turning the solu(cid:415)on a dark red color. Reac(cid:415)on was s(cid:415)rred overnight then 

precipitated  with  dry  ether  and  ﬁltered.  The  solvent  was  then  removed  via  vacuum  and  the 

Cu(I)bpyPY4 product was collected (0.2440 g, 76.1% yield).  1H-NMR (CDCl3 500 MHz) δ 2.25 (s, 

6H), 7.20 (s, 4H), 7.76 (s, 4H), 8.10 (d, 2H), 8.18 (t, 2H), 8.37 (d, 2H) 

Synthesis of [Cu(bpyPY4)]OTf2: 

Copper(II) triﬂate (0.4215 g, 1.16 mmol) was dissolved  in minimal dry acetonitrile. Meanwhile 

BPYPY4  (0.6942  g,  1.33  mmol)  was  dissolved  in  minimal  dry  dichloromethane.  The  BPYPY4 

solu(cid:415)on  was  added  dropwise  to  the  copper  solu(cid:415)on,  turning  the  solu(cid:415)on  a  pale  blue  color. 

Reac(cid:415)on was s(cid:415)rred overnight then precipitated with dry ether and ﬁltered. The solvent was then 

removed via vacuum and the Cu(II)bpyPY4 product was collected (0.6571 g, 63.9% yield). 

5.3 Results and Discussion 

While Dr. Rorabacher already conﬁrmed the Nerns(cid:415)an rela(cid:415)onship between copper(II) stability 

constants  and  poten(cid:415)al,  the  rela(cid:415)onship  was  only  inves(cid:415)gated  in  aqueous  solu(cid:415)ons,  whereas 

89 

 
 
most DSSC electrolytes use acetonitrile as the solvent. To discern the rela(cid:415)onship between the 

redox poten(cid:415)al and stability constants of copper complexes in acetonitrile, data from previous 

studies was analyzed, speciﬁcally concentra(cid:415)ng on the stability constants of these complexes in 

acetonitrile.98,99 A(cid:332)er all the collated data was plo(cid:425)ed, Figure 5.2A, a linear regression, with a 

ﬁxed slope of -0.059, was applied, resul(cid:415)ng in the equa(cid:415)on y = -0.059x + 0.9002. This equa(cid:415)on 

was then used to simulate the stability constants of redox shu(cid:425)led u(cid:415)lized in DSSC by using their 

previous determined redox poten(cid:415)als, Figure 5.2B.38,39,41,94,95 

Figure 5.2: A) a plot of published copper redox poten(cid:415)als versus the log of the stability constant 
of the copper(II) form of the redox complex found in acetonitrile.98,99 The solid line was ﬁt while 
holding the slope at the Nerns(cid:415)an slope of -0.059, resul(cid:415)ng in an equa(cid:415)ons of y = -0.059x + 
0.9002. B) A simulated plot where the stability constant of the copper(II) form of various copper 
complexes used in DSSCs was simulated using the equa(cid:415)on from plot A.38,39,41,94,95 

The simulated plots reveal a pa(cid:425)ern: copper complexes that undergo ligand subs(cid:415)tu(cid:415)on display 

more  posi(cid:415)ve  poten(cid:415)als  and  subsequently  exhibit  lower  stability  constants  compared  to  the 

[Cu(TBP)4]2+/+ redox shu(cid:425)le, while those complexes that do not undergo ligand subs(cid:415)tu(cid:415)on have 

more nega(cid:415)ve poten(cid:415)als and thus higher stability constants. This correla(cid:415)on underscores the role 

of  redox  poten(cid:415)al  as  a  poten(cid:415)al  tool  for  predic(cid:415)ng  the  suscep(cid:415)bility  of  copper  complexes  to 

ligand subs(cid:415)tu(cid:415)on. 

90 

 
 
 
The valida(cid:415)on of the simulated data was pursued by examining four copper complexes u(cid:415)lized in 

DSSCs, based on their relevance to the study's goals and their dis(cid:415)nc(cid:415)ve redox poten(cid:415)als. The 

selected complexes were [Cu(TBP)4]2+/+, [Cu(dmbpy)2]2+/+, [Cu(dbmed)]2+/+, and [Cu(bpyPY4)]2+/+. 

The  [Cu(TBP)4]2+/+  was  chosen  as  a  reference  point  for  compara(cid:415)ve  evalua(cid:415)on,  forming  the 

baseline  against  which  other  complexes  were  assessed.  The  selec(cid:415)on  of  [Cu(dmbpy)2]2+/+  was 

mo(cid:415)vated by its posi(cid:415)ve redox poten(cid:415)al, placing it as one of the copper redox shu(cid:425)les with the 

highest  predicted  stability  constant  among  those  commonly  employed  in  DSSCs.  Conversely, 

[Cu(bpyPY4)]2+/+  was  singled  out  for  its  nega(cid:415)ve  redox  poten(cid:415)al,  reﬂec(cid:415)ng  its  status  among 

copper  redox  shu(cid:425)les  with  the  lowest  poten(cid:415)als  integrated  into  DSSCs.  The  selec(cid:415)on  of 

[Cu(dbmed)]2+/+  was due  to  its  redox poten(cid:415)al  closely  resembling  that  of  [Cu(TBP)4]2+/+,  which 

enabled  a  targeted  explora(cid:415)on  of  ligand  subs(cid:415)tu(cid:415)on  behavior  while  upholding  a  uniform 

poten(cid:415)al. 

For  [Cu(dmbpy)2]2+/+,  [Cu(dbmed)]2+/+,  and  [Cu(bpyPY4)]2+/+  the  stability  constants  were 

determined via the applica(cid:415)on of square wave voltammetry. Due to the high stability of copper(II) 

complexes  in  acetonitrile,  making  them  diﬃcult  to  measure,  the  stability  constants  of  the 

copper(I) form were experimentally determined. Copper(I) complexes have been shown to have 

a signiﬁcantly lower stability constant in acetonitrile than in water. This has been a(cid:425)ributed to the 

signiﬁcantly higher binding energy of copper(I) to acetonitrile than that of water, causing about a 

106  decrease  in  stability  constants  in  acetonitrile  rela(cid:415)ve  to  their  aqueous  values.99,100  Square 

wave voltammetry was u(cid:415)lized to determine the stability constant of the copper(I) form of the 

complexes. When using square wave voltammetry, assuming the sweep rate is held constant, the 

copper(I) oxida(cid:415)on peak can be directly related to the [CuIL] following the equa(cid:415)on 

91 

 
 
(cid:3117)
𝑖 = 𝑛𝐹𝐴𝐶(𝜋𝐷𝜎)
(cid:3118)𝛸(𝜎𝑡)  

(5.1) 

where i is the peak current, n is the number of electrons transferred, F is the Faraday constant, C 

is the bulk concentra(cid:415)on of CuIL, D is the diﬀusion coeﬃcient, σ  = (nFν/RT), ν is the sweep rate, 

R  is  the  gas  constant,  T  is  the  absolute  temperature,  and  Χ(σt)  is  the  normalized  current.  The 

concentra(cid:415)on  at  each  measurement  was  determined  using  the  propor(cid:415)onality  of  the  peak 

current, i, for the measurement compared to the peak current when all copper is bound to ligand, 

i∞, usually at 120% ligand added, following the equa(cid:415)on [CuL] = CCu(i/i∞). 

From the combina(cid:415)on of deﬁni(cid:415)on of equilibrium constant between copper and a ligand 

𝐾(cid:3004)(cid:3048)(cid:3013) =

[(cid:3004)(cid:3048)(cid:3013)]
[(cid:3004)(cid:3048)][(cid:3013)]

(5.2) 

and the known concentra(cid:415)ons of copper, CCu, and ligand, CL,  

𝐶(cid:3004)(cid:3048) = [𝐶𝑢] + [𝐶𝑢𝐿]   (5.3) 

𝐶(cid:3013) = [𝐿] + [𝐶𝑢𝐿]  

(5.4) 

a new equa(cid:415)on can be derived 

[𝐶𝑢𝐿] =  

(cid:3012)(cid:3252)(cid:3296)(cid:3261)(cid:3004)(cid:3261)((cid:3004)(cid:3252)(cid:3296)(cid:2879)[(cid:3004)(cid:3048)(cid:3013)])
(cid:2869)(cid:2878)(cid:3012)(cid:3252)(cid:3296)(cid:3261)((cid:3004)(cid:3252)(cid:3296)(cid:2879)[(cid:3004)(cid:3048)(cid:3013)])

(5.5) 

then subs(cid:415)tu(cid:415)ng equa(cid:415)on 5.5 into equa(cid:415)on 5.1 yields the equa(cid:415)on 

(cid:3004)(cid:3252)(cid:3296)
(cid:3036)

=

(cid:2869)

(cid:3026)

+

(cid:2869)
(cid:3026)(cid:3012)(cid:3252)(cid:3296)(cid:3261)((cid:3004)(cid:3261)(cid:2879)[(cid:3004)(cid:3048)(cid:3013)])

(5.6) 

where CCu is the total concentra(cid:415)on of copper in solu(cid:415)on, CL is the total concentra(cid:415)on of ligand 

in solu(cid:415)on, i is the peak current, KCuL is the stability constant of the copper(I) form of the complex, 

and Y is the product of several electrochemical constants. By plo(cid:427)ng CCu/i by 1/(CL – [CuL]) and 

taking a linear regression of the plot, you obtain a slope of 1/YKCuL and an intercept of 1/Y. By 

dividing the intercept by the slope, one is le(cid:332) with KCuL. 

92 

 
 
   
  
  
To get the stability constant of the copper(II) form, the Nerst equa(cid:415)on was used 

𝐸(cid:3004)(cid:3048)(cid:3258)(cid:3258)(cid:3013)

(cid:3033) = 𝐸(cid:3004)(cid:3048)((cid:3010)/(cid:3010)(cid:3010))(cid:3046)(cid:3042)(cid:3039)(cid:3049)

(cid:3033) −

(cid:2870).(cid:2871)(cid:2868)(cid:2871)(cid:3019)(cid:3021)

(cid:3007)

𝑙𝑜𝑔

(cid:3012)

(cid:3252)(cid:3296)(cid:3258)(cid:3258)(cid:3261)
(cid:3012)(cid:3252)(cid:3296)(cid:3258)(cid:3261)

  (5.7) 

where ECu(I/II)solv

f is the formal poten(cid:415)al of the solvated copper(I/II), [Cu(ACN)4]2+/+, redox couple, 

which was found to be 0.601 V vs ferrocene99.  

The log of stability constants of the copper(II) for of the complexes were found to be 9.05, 15.02, 

and 21.53 respec(cid:415)vely for the [Cu(dmbpy)2]2+/+, [Cu(dbmed)]2+/+, and [Cu(bpyPY4)]2+/+ complexes. 

These results match well with the log of the predicted stability constants found previously, 9.49, 

15.34, and 22.55 respec(cid:415)vely. However, when the [Cu(TBP)4]2+/+ complex was a(cid:425)empted to be 

measured via square wave voltammetry it was found that the electron kine(cid:415)cs of the complex 

were too slow at the sweep rate used for  a redox wave to show, so a diﬀerent technique was 

used to determine the copper(II) stability constant. 

93 

 
 
 
Figure 5.13: The square wave voltammetry plots for 2.0 mM of A) [Cu(dmbpy)2]OTf2, blue, B) 
[Cu(dbmed)]OTf2, orange, C) [Cu(bpyPY4)]OTf2, green, and [Cu(ACN)4]OTf2, in anhydrous 
acetonitrile containing 0.1 M LiOTf. D) The UV-Vis spectra of 2.0 mM [Cu(dbmed)]OTf2 (cid:415)trated 
increasing equivalents of TBP. For all spectra as the concentra(cid:415)on of ligand increases the darker 
the spectra. 

For [Cu(TBP)4]2+/+ the redox couple was electrochemically irreversible at the sweep rates used for 

the square wave  measurement, so instead UV-Vis was used  to determine the rela(cid:415)ve stability 

constant  between  [Cu(TBP)4]2+/+  and  [Cu(dbmed)]2+/+.  Since  the  diﬀerence  of  the  predicted 

stability constants for the  [Cu(dbmed)]OTf2 and the [Cu(TBP)4]OTf2 complexes is small, only 10.6, 

UV-Vis spectroscopy can be used to determine the equillibrium constant for the reac(cid:415)on 

[Cu(dbmed)](OTf)2 + 4 TBP

[Cu(TBP)4](OTf)2 + dbmed

(Scheme 5.1) 

The  wavelength  of  1100  nm  was  used  due  to  the  minimal  contribu(cid:415)on  from  the  [Cu(TBP)4] 

complex, free TBP, and free dbmed to the absorbance at this wavelength. Due to this we can use 

the Beer-Lambert law 

𝐴 = 𝜀(cid:3004)(cid:3048)(cid:3013)𝑏[𝐶𝑢𝐿] 

(5.8) 

94 

 
 
 
  
where εCuL is the ex(cid:415)nc(cid:415)on coeﬃcient of the [Cu(dbmed)]OTf2 complex, b is the path length in 

cm,  and  [CuL]  is  the  concentra(cid:415)on  of  the  [Cu(dbmed)]OTf2  complex  in  solu(cid:415)on.  By  using  the 

deﬁni(cid:415)on of an equilibrium constant  

𝐾(cid:3004)(cid:3048)(cid:3013) =

[(cid:3004)(cid:3048)((cid:3021)(cid:3003)(cid:3017))(cid:3120)][(cid:3031)(cid:3029)(cid:3040)(cid:3032)(cid:3031)]
[(cid:3004)(cid:3048)((cid:3031)(cid:3029)(cid:3040)(cid:3032)(cid:3031))][(cid:3021)(cid:3003)(cid:3017)](cid:3120) 

(5.9) 

and the deﬁni(cid:415)ons of the analy(cid:415)cal concentra(cid:415)ons of copper(II), CCu, and dbmed, Cdbmed, and TBP, 

CTBP, in solu(cid:415)on  

𝐶(cid:3004)(cid:3048) = [𝐶𝑢(𝑇𝐵𝑃)(cid:2872)] + [𝐶𝑢(𝑑𝑏𝑚𝑒𝑑)] (5.10) 

𝐶(cid:3031)(cid:3029)(cid:3040)(cid:3032)(cid:3031) = [𝑑𝑏𝑚𝑒𝑑] + [𝐶𝑢(𝑑𝑏𝑚𝑒𝑑)] 

(5.11) 

𝐶(cid:3021)(cid:3003)(cid:3017) = [𝑇𝐵𝑃](cid:2872) + [𝐶𝑢(𝑇𝐵𝑃)(cid:2872)] 

(5.12) 

under  the  assump(cid:415)on  that  all  the  copper(II)  is  complexed  with  a  ligand,  the  equilibrium 

coeﬃcient calcula(cid:415)on can be rearranged to be 

𝐾(cid:3032)(cid:3044) =

((cid:3004)(cid:3252)(cid:3296)(cid:2879)[(cid:3004)(cid:3048)((cid:3031)(cid:3029)(cid:3040)(cid:3032)(cid:3031))])((cid:3004)(cid:3279)(cid:3277)(cid:3288)(cid:3280)(cid:3279)(cid:2879)[(cid:3004)(cid:3048)((cid:3031)(cid:3029)(cid:3040)(cid:3032)(cid:3031))])
[(cid:3004)(cid:3048)((cid:3031)(cid:3029)(cid:3040)(cid:3032)(cid:3031))]((cid:3004)(cid:3269)(cid:3251)(cid:3265)(cid:2879)(cid:3004)(cid:3252)(cid:3296)(cid:2879)[(cid:3004)(cid:3048)((cid:3031)(cid:3029)(cid:3040)(cid:3032)(cid:3031))])

  (5.13) 

Using this rela(cid:415)onship the equilibrium constant for the rela(cid:415)onship between [Cu(dbmed)]OTf2 

and  [Cu(TBP)4]OTf2  was  determined  to  be  1.02  (±  0.27)  which,  when  referenced  against  the 

stability constant for [Cu(dbmed)]OTf2, gave a stability constant of 16.74 for [Cu(TBP)4]OTf2, which 

is close to the predicted stability constant of 16.21. 

95 

 
 
Table 5.8: Simulated and experimental copper(II) stability constants and the percent error in 
between the measurements. 

Experimental 
Cu(II) stability 
constant 

Simulated Cu(II) 
stability constant 

Percent Error 
(%) 

Cu(dmpby)2 

9.05 

9.49 

Cu(dbmed) 

15.02 

Cu(bpyPY4) 

21.53 

Cu(TBP)4 

16.68 

15.34 

22.55 

16.21 

4.92 

2.01 

4.73 

2.80 

The stability constants determined for all the copper complexes are listed in Table 5.1 along with 

the predicted stability constant and the percent error for each copper complex. The comparison 

between  these  experimental  results  and  the  simulated  data  revealed  that  all  the  complexes 

displaying  devia(cid:415)ons  of  less  than  5%  from  the  simulated  outcomes.  This  agreement  to  the 

predic(cid:415)ve model denotes the poten(cid:415)al of the redox poten(cid:415)al-stability constant rela(cid:415)onship to 

be used as a predic(cid:415)ve tool for an(cid:415)cipa(cid:415)ng ligand subs(cid:415)tu(cid:415)on behavior. 

96 

 
 
 
 
 
 
 
V
/
e
n
e
c
o
r
r
e
F
s
v
l

a
i
t
n
e
t
o
P

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

Cu Complexes potential vs Stability constants in ACN

y = -0.059x + 0.900
y = -0.060x + 0.900
R² = 0.987

Cu(bpyPY4)

Cu(bpyPY4)

Cu(PY5)

Cu(TBP)4

Cu(dbmed)

Cu(TBP)4

Cu(dbmed)

Cu(dmbppy)2

Cu(tmbpy)2

Cu(dmbpy)2

Cu(ACN)4

5

10

15
log KCu(II)L

20

25

Figure 5.4: A plot of copper redox poten(cid:415)als versus the log of the simulated, orange, and 
experimental, blue, stability constant of the copper(II) form of the redox complex found in 
acetonitrile. The linear ﬁt equa(cid:415)on for the experimental results was found to be y = -0.060x + 
0.900. 

Addi(cid:415)onally,  to  deepen  insights,  the  poten(cid:415)al  of  each  redox  shu(cid:425)le  was  plo(cid:425)ed  against  the 

measured  stability  constant,  enabling  a  subsequent  linear  regression  analysis.  This  analysis 

yielded a slope of -0.060, closely mirroring the an(cid:415)cipated fundamental value of -0.059. Although 

it is important to note that this rela(cid:415)onship should only hold true as long as the stability constants 

of  the  copper(I)  form  of  the  redox  couple  have  similar  values.  This  adherence  to  theore(cid:415)cal 

expecta(cid:415)ons further underscores the legi(cid:415)macy of the correla(cid:415)on, oﬀering a compelling avenue 

for leveraging redox poten(cid:415)al to predict ligand subs(cid:415)tu(cid:415)on behavior in copper complexes. 

97 

 
 
 
 
 
 
 
5.4 Conclusions 

This study capitalizes on Dr. Rorabacher's work, revealing a compelling connec(cid:415)on between redox 

poten(cid:415)al and stability constants in copper complexes. This correla(cid:415)on enables a predic(cid:415)ve model 

of poten(cid:415)al-stability constant rela(cid:415)onships. Assuming the stability constants of the copper(I) form 

are consistent, this model facilitates the accurate predic(cid:415)on of copper complexes. Speciﬁcally, it 

helps  dis(cid:415)nguish  between  those  that  undergo  ligand  subs(cid:415)tu(cid:415)on  and  those  that  do  not. 

Valida(cid:415)on with four relevant copper complexes shows the predic(cid:415)ve model's accuracy. Given the 

knowledge that ligand subs(cid:415)tu(cid:415)on can be beneﬁcial for DSSCs, enhancing eﬃciencies could be 

achieved by enabling ligand subs(cid:415)tu(cid:415)on with a base that induces a smaller nega(cid:415)ve shi(cid:332) in the 

solu(cid:415)on  poten(cid:415)al.38  This  predic(cid:415)ve  method  could  allow  for  the  selec(cid:415)on  of  targeted  redox 

shu(cid:425)le/base combina(cid:415)ons, aiming to maximize the eﬃciency of DSSC devices. This work paves 

the way for informed design and op(cid:415)miza(cid:415)on of copper-based redox shu(cid:425)les in dye-sensi(cid:415)zed 

solar cells, oﬀering new dimensions for enhancing their performance. 

98 

 
 
WORKS CITED 

(1) 

Friedlingstein, P.; O’Sullivan, M.; Jones, M. W.; Andrew, R. M.; Gregor, L.; Hauck, J.; Le 
Quéré, C.; Luijkx, I. T.; Olsen, A.; Peters, G. P.; Peters, W.; Pongratz, J.; Schwingshackl, C.; 
Sitch, S.; Canadell, J. G.; Ciais, P.; Jackson, R. B.; Alin, S. R.; Alkama, R.; Arneth, A.; Arora, V. 
K.; Bates, N. R.; Becker, M.; Bellouin, N.; Bi(cid:427)g, H. C.; Bopp, L.; Chevallier, F.; Chini, L. P.; 
Cronin, M.; Evans, W.; Falk, S.; Feely, R. A.; Gasser, T.; Gehlen, M.; Gkritzalis, T.; Gloege, L.; 
Grassi, G.; Gruber, N.; Gürses, Ö.; Harris, I.; Hefner, M.; Houghton, R. A.; Hur(cid:425), G. C.; Iida, 
Y.; Ilyina, T.; Jain, A. K.; Jersild, A.; Kadono, K.; Kato, E.; Kennedy, D.; Klein Goldewijk, K.; 
Knauer, J.; Korsbakken, J. I.; Landschützer, P.; Lefèvre, N.; Lindsay, K.; Liu, J.; Liu, Z.; 
Marland, G.; Mayot, N.; McGrath, M. J.; Metzl, N.; Monacci, N. M.; Munro, D. R.; 
Nakaoka, S.-I.; Niwa, Y.; O’Brien, K.; Ono, T.; Palmer, P. I.; Pan, N.; Pierrot, D.; Pocock, K.; 
Poulter, B.; Resplandy, L.; Robertson, E.; Rödenbeck, C.; Rodriguez, C.; Rosan, T. M.; 
Schwinger, J.; Séférian, R.; Shutler, J. D.; Skjelvan, I.; Steinhoﬀ, T.; Sun, Q.; Su(cid:425)on, A. J.; 
Sweeney, C.; Takao, S.; Tanhua, T.; Tans, P. P.; Tian, X.; Tian, H.; Tilbrook, B.; Tsujino, H.; 
Tubiello, F.; van der Werf, G. R.; Walker, A. P.; Wanninkhof, R.; Whitehead, C.; Willstrand 
Wranne, A.; Wright, R.; Yuan, W.; Yue, C.; Yue, X.; Zaehle, S.; Zeng, J.; Zheng, B. Global 
Carbon Budget 2022. Earth Syst Sci Data 2022, 14 (11), 4811–4900. 
h(cid:425)ps://doi.org/10.5194/essd-14-4811-2022. 

(2) 

Interna(cid:415)onal Energy Agency. CO2 Emissions in 2022; 2023. www.iea.org. 

(3) 

Hausfather, Z.; Peters, G. P. Emissions – the ‘Business as Usual’ Story Is Misleading. Nature 
2020, 577 (7792), 618–620. h(cid:425)ps://doi.org/10.1038/d41586-020-00177-3. 

(4) 

Center for Sustainable Systems, U. of M. Photovoltaic Energy Factsheet; 2023. 

(5) 

Interna(cid:415)onal Energy Agency. Solar PV; 2023. 

(6) 

Gueymard, C. A.; Myers, D.; Emery, K. Proposed Reference Irradiance Spectra for Solar 
Energy Systems Tes(cid:415)ng. Solar Energy 2002, 73 (6), 443–467. 
h(cid:425)ps://doi.org/10.1016/S0038-092X(03)00005-7. 

(7) 

U.S. DOE. The SunShot Ini(cid:415)a(cid:415)ve. 

(8) 

(9) 

Jung, H. S.; Lee, J.-K. Dye Sensi(cid:415)zed Solar Cells for Economically Viable Photovoltaic 
Systems. J Phys Chem Le(cid:425) 2013, 4 (10), 1682–1693. h(cid:425)ps://doi.org/10.1021/jz400112n. 

Becquerel, E. Mémoire Sur Les Eﬀets Électriques Produits Sous l’inﬂuence Des Rayons 
Solaires. Comptes Rendus 1839, 9, 561–567. 

(10)  Moser, J. No(cid:415)z �ber Verst�rkung Photoelektrischer Str�me Durch Op(cid:415)sche 

Sensibilisirung. Monatshe(cid:332)e f�r Chemie 1887, 8 (1), 373–373. 
h(cid:425)ps://doi.org/10.1007/BF01510059. 

99 

 
 
(11)  Gerischer, H.; Tributsch, H. Elektrochemische Untersuchungen Zur Spektralen 

Sensibilisierung von ZnO-Einkristallen. Berichte der Bunsengesellscha(cid:332) für physikalische 
Chemie 1968, 72 (3), 437–445. h(cid:425)ps://doi.org/10.1002/bbpc.196800013. 

(12)  ANDERSON, S.; CONSTABLE, E. C.; DARE-EDWARDS, M. P.; GOODENOUGH, J. B.; 

HAMNETT, A.; SEDDON, K. R.; WRIGHT, R. D. Chemical Modiﬁca(cid:415)on of a Titanium (IV) 
Oxide Electrode to Give Stable Dye Sensi(cid:415)sa(cid:415)on without a Supersensi(cid:415)ser. Nature 1979, 
280 (5723), 571–573. h(cid:425)ps://doi.org/10.1038/280571a0. 

(13)  Matsumura, M.; Matsudaira, S.; Tsubomura, H.; Takata, M.; Yanagida, H. Dye Sensi(cid:415)za(cid:415)on 

and Surface Structures of Semiconductor Electrodes. Industrial & Engineering Chemistry 
Product Research and Development 1980, 19 (3), 415–421. 
h(cid:425)ps://doi.org/10.1021/i360075a025. 

(14)  Vlachopoulos, Nick.; Liska, Paul.; Augustynski, Jan.; Graetzel, Michael. Very Eﬃcient 

Visible Light Energy Harves(cid:415)ng and Conversion by Spectral Sensi(cid:415)za(cid:415)on of High Surface 
Area Polycrystalline Titanium Dioxide Films. J Am Chem Soc 1988, 110 (4), 1216–1220. 
h(cid:425)ps://doi.org/10.1021/ja00212a033. 

(15)  Nazeeruddin, M. K.; Kay, A.; Rodicio, I.; Humphry-Baker, R.; Mueller, E.; Liska, P.; 
Vlachopoulos, N.; Graetzel, M. Conversion of Light to Electricity by Cis-X2bis(2,2’-
Bipyridyl-4,4’-Dicarboxylate)Ruthenium(II) Charge-Transfer Sensi(cid:415)zers (X = Cl-, Br-, I-, CN-, 
and SCN-) on Nanocrystalline Titanium Dioxide Electrodes. J Am Chem Soc 1993, 115 (14), 
6382–6390. h(cid:425)ps://doi.org/10.1021/ja00067a063. 

(16)  Feldt, S. M.; Gibson, E. A.; Gabrielsson, E.; Sun, L.; Boschloo, G.; Hagfeldt, A. Design of 

Organic Dyes and Cobalt Polypyridine Redox Mediators for High-Eﬃciency Dye-Sensi(cid:415)zed 
Solar Cells. J Am Chem Soc 2010, 132 (46), 16714–16724. 
h(cid:425)ps://doi.org/10.1021/ja1088869. 

(17)  Daeneke, T.; Kwon, T.-H.; Holmes, A. B.; Duﬀy, N. W.; Bach, U.; Spiccia, L. High-Eﬃciency 

Dye-Sensi(cid:415)zed Solar Cells with Ferrocene-Based Electrolytes. Nat Chem 2011, 3 (3), 211–
215. h(cid:425)ps://doi.org/10.1038/nchem.966. 

(18)  Bai, Y.; Yu, Q.; Cai, N.; Wang, Y.; Zhang, M.; Wang, P. High-Eﬃciency Organic Dye-
Sensi(cid:415)zed Mesoscopic Solar Cells with a Copper Redox Shu(cid:425)le. Chemical 
Communica(cid:415)ons 2011, 47 (15), 4376. h(cid:425)ps://doi.org/10.1039/c1cc10454c. 

(19)  Cao, Y.; Liu, Y.; Zakeeruddin, S. M.; Hagfeldt, A.; Grätzel, M. Direct Contact of Selec(cid:415)ve 

Charge Extrac(cid:415)on Layers Enables High-Eﬃciency Molecular Photovoltaics. Joule 2018, 2 
(6), 1108–1117. h(cid:425)ps://doi.org/10.1016/j.joule.2018.03.017. 

(20)  Ren, Y.; Zhang, D.; Suo, J.; Cao, Y.; Eickemeyer, F. T.; Vlachopoulos, N.; Zakeeruddin, S. M.; 
Hagfeldt, A.; Grätzel, M. Hydroxamic Acid Pre-Adsorp(cid:415)on Raises the Eﬃciency of 

100 

 
 
Cosensi(cid:415)zed Solar Cells. Nature 2023, 613 (7942), 60–65. 
h(cid:425)ps://doi.org/10.1038/s41586-022-05460-z. 

(21)  Ondersma, J. W.; Hamann, T. W. Conduc(cid:415)on Band Energy Determina(cid:415)on by Variable 
Temperature Spectroelectrochemistry. Energy Environ Sci 2012, 5 (11), 9476. 
h(cid:425)ps://doi.org/10.1039/c2ee22926a. 

(22)  Boschloo, G.; Hagfeldt, A. Characteris(cid:415)cs of the Iodide/Triiodide Redox Mediator in Dye-

Sensi(cid:415)zed Solar Cells. Acc Chem Res 2009, 42 (11), 1819–1826. 
h(cid:425)ps://doi.org/10.1021/ar900138m. 

(23)  Mathew, S.; Yella, A.; Gao, P.; Humphry-Baker, R.; Curchod, B. F. E.; Ashari-Astani, N.; 

Tavernelli, I.; Rothlisberger, U.; Nazeeruddin, Md. K.; Grätzel, M. Dye-Sensi(cid:415)zed Solar Cells 
with 13% Eﬃciency Achieved through the Molecular Engineering of Porphyrin Sensi(cid:415)zers. 
Nat Chem 2014, 6 (3), 242–247. h(cid:425)ps://doi.org/10.1038/nchem.1861. 

(24)  Shi, J.; Zhang, Y.; Luo, B.; Chen, L. Synthesis and Structural Characteriza(cid:415)on of the 

Copper(II) Complex with N, N(cid:3387)-Bis-(1-Benzimidazolylethyl)Ethylenediamine. Inorganica 
Chim Acta 1989, 162 (1), 29–31. h(cid:425)ps://doi.org/10.1016/S0020-1693(00)83115-1. 

(25)  Randall, D. W.; Gamelin, D. R.; LaCroix, L. B.; Solomon, E. I. Electronic Structure 

Contribu(cid:415)ons to Electron Transfer in Blue Cu and CuA. JBIC Journal of Biological Inorganic 
Chemistry 2000, 5 (1), 16–29. h(cid:425)ps://doi.org/10.1007/s007750050003. 

(26)  Ha(cid:425)ori, S.; Wada, Y.; Yanagida, S.; Fukuzumi, S. Blue Copper Model Complexes with 

Distorted Tetragonal Geometry Ac(cid:415)ng as Eﬀec(cid:415)ve Electron-Transfer Mediators in Dye-
Sensi(cid:415)zed Solar Cells. J Am Chem Soc 2005, 127 (26), 9648–9654. 
h(cid:425)ps://doi.org/10.1021/ja0506814. 

(27)  Bai, Y.; Yu, Q.; Cai, N.; Wang, Y.; Zhang, M.; Wang, P. High-Eﬃciency Organic Dye-
Sensi(cid:415)zed Mesoscopic Solar Cells with a Copper Redox Shu(cid:425)le. Chemical 
Communica(cid:415)ons 2011, 47 (15), 4376. h(cid:425)ps://doi.org/10.1039/c1cc10454c. 

(28)  Colombo, A.; Dragone(cid:427), C.; Magni, M.; Roberto, D.; Demar(cid:415)n, F.; Caramori, S.; Bignozzi, 
C. A. Eﬃcient Copper Mediators Based on Bulky Asymmetric Phenanthrolines for DSSCs. 
ACS Appl Mater Interfaces 2014, 6 (16), 13945–13955. 
h(cid:425)ps://doi.org/10.1021/am503306f. 

(29)  Cong, J.; Kinschel, D.; Daniel, Q.; Safdari, M.; Gabrielsson, E.; Chen, H.; Svensson, P. H.; 

Sun, L.; Kloo, L. Bis(1,1-Bis(2-Pyridyl)Ethane)Copper( <scp>i</Scp> / <scp>ii</Scp> ) as an 
Eﬃcient Redox Couple for Liquid Dye-Sensi(cid:415)zed Solar Cells. J Mater Chem A Mater 2016, 
4 (38), 14550–14554. h(cid:425)ps://doi.org/10.1039/C6TA06782D. 

101 

 
 
(30)  Freitag, M.; Giordano, F.; Yang, W.; Pazoki, M.; Hao, Y.; Zietz, B.; Grätzel, M.; Hagfeldt, A.; 
Boschloo, G. Copper Phenanthroline as a Fast and High-Performance Redox Mediator for 
Dye-Sensi(cid:415)zed Solar Cells. The Journal of Physical Chemistry C 2016, 120 (18), 9595–9603. 
h(cid:425)ps://doi.org/10.1021/acs.jpcc.6b01658. 

(31)  Saygili, Y.; Söderberg, M.; Pellet, N.; Giordano, F.; Cao, Y.; Muñoz-García, A. B.; 

Zakeeruddin, S. M.; Vlachopoulos, N.; Pavone, M.; Boschloo, G.; Kavan, L.; Moser, J.-E.; 
Grätzel, M.; Hagfeldt, A.; Freitag, M. Copper Bipyridyl Redox Mediators for Dye-Sensi(cid:415)zed 
Solar Cells with High Photovoltage. J Am Chem Soc 2016, 138 (45), 15087–15096. 
h(cid:425)ps://doi.org/10.1021/jacs.6b10721. 

(32)  Hu, M.; Shen, J.; Yu, Z.; Liao, R.-Z.; Gurzadyan, G. G.; Yang, X.; Hagfeldt, A.; Wang, M.; Sun, 
L. Eﬃcient and Stable Dye-Sensi(cid:415)zed Solar Cells Based on a Tetradentate Copper(II/I) 
Redox Mediator. ACS Appl Mater Interfaces 2018, 10 (36), 30409–30416. 
h(cid:425)ps://doi.org/10.1021/acsami.8b10182. 

(33)  Long, H.; Zhou, D.; Zhang, M.; Peng, C.; Uchida, S.; Wang, P. Probing Dye-Correlated 

Interplay of Energe(cid:415)cs and Kine(cid:415)cs in Mesoscopic Titania Solar Cells with 4- Tert -
Butylpyridine. The Journal of Physical Chemistry C 2011, 115 (29), 14408–14414. 
h(cid:425)ps://doi.org/10.1021/jp202826m. 

(34)  XIONG, B.; ZHOU, B.; ZHU, Z.; GAO, T.; LI, L.; CAI, J.; CAI, W. Adsorp(cid:415)on of 4 -tert- 

Butylpyridine on TiO 2 Surface in Dye-Sensi(cid:415)zed Solar Cells. Chin J Chem 2008, 26 (1), 70–
76. h(cid:425)ps://doi.org/10.1002/cjoc.200890040. 

(35)  Kashif, M. K.; Axelson, J. C.; Duﬀy, N. W.; Forsyth, C. M.; Chang, C. J.; Long, J. R.; Spiccia, L.; 

Bach, U. A New Direc(cid:415)on in Dye-Sensi(cid:415)zed Solar Cells Redox Mediator Development: In 
Situ Fine-Tuning of the Cobalt(II)/(III) Redox Poten(cid:415)al through Lewis Base Interac(cid:415)ons. J 
Am Chem Soc 2012, 134 (40), 16646–16653. h(cid:425)ps://doi.org/10.1021/ja305897k. 

(36)  Nakade, S.; Kanzaki, T.; Kubo, W.; Kitamura, T.; Wada, Y.; Yanagida, S. Role of Electrolytes 

on Charge Recombina(cid:415)on in Dye-Sensi(cid:415)zed TiO 2 Solar Cell (1):  The Case of Solar Cells 
Using the I - /I 3 - Redox Couple. J Phys Chem B 2005, 109 (8), 3480–3487. 
h(cid:425)ps://doi.org/10.1021/jp0460036. 

(37)  Kusama, H.; Konishi, Y.; Sugihara, H.; Arakawa, H. Inﬂuence of Alkylpyridine Addi(cid:415)ves in 
Electrolyte Solu(cid:415)on on the Performance of Dye-Sensi(cid:415)zed Solar Cell. Solar Energy 
Materials and Solar Cells 2003, 80 (2), 167–179. 
h(cid:425)ps://doi.org/10.1016/j.solmat.2003.08.001. 

(38)  Wang, Y.; Hamann, T. W. Improved Performance Induced by in Situ Ligand Exchange 
Reac(cid:415)ons of Copper Bipyridyl Redox Couples in Dye-Sensi(cid:415)zed Solar Cells. Chemical 
Communica(cid:415)ons 2018, 54 (87), 12361–12364. h(cid:425)ps://doi.org/10.1039/C8CC07191H. 

102 

 
 
(39)  Fürer, S. O.; Milhuisen, R. A.; Kashif, M. K.; Raga, S. R.; Acharya, S. S.; Forsyth, C.; Liu, M.; 

Frazer, L.; Duﬀy, N. W.; Ohlin, C. A.; Funston, A. M.; Tachibana, Y.; Bach, U. The 
Performance-Determining Role of Lewis Bases in Dye-Sensi(cid:415)zed Solar Cells Employing 
Copper-Bisphenanthroline Redox Mediators. Adv Energy Mater 2020, 10 (37). 
h(cid:425)ps://doi.org/10.1002/aenm.202002067. 

(40)  Hu, M.; Shen, J.; Yu, Z.; Liao, R.-Z.; Gurzadyan, G. G.; Yang, X.; Hagfeldt, A.; Wang, M.; Sun, 
L. Eﬃcient and Stable Dye-Sensi(cid:415)zed Solar Cells Based on a Tetradentate Copper(II/I) 
Redox Mediator. ACS Appl Mater Interfaces 2018, 10 (36), 30409–30416. 
h(cid:425)ps://doi.org/10.1021/acsami.8b10182. 

(41)  Rui, H.; Shen, J.; Yu, Z.; Li, L.; Han, H.; Sun, L. Stable Dye-Sensi(cid:415)zed Solar Cells Based on 
Copper(II/I) Redox Mediators Bearing a Pentadentate Ligand. Angewandte Chemie 
Interna(cid:415)onal Edi(cid:415)on 2021, 60 (29), 16156–16163. 
h(cid:425)ps://doi.org/10.1002/anie.202104563. 

(42)  Michaels, H.; Benesperi, I.; Edvinsson, T.; Muñoz-Garcia, A.; Pavone, M.; Boschloo, G.; 

Freitag, M. Copper Complexes with Tetradentate Ligands for Enhanced Charge Transport 
in Dye-Sensi(cid:415)zed Solar Cells. Inorganics (Basel) 2018, 6 (2), 53. 
h(cid:425)ps://doi.org/10.3390/inorganics6020053. 

(43) 

Jeﬀrey D. Sachs. The Road to Clean Energy Starts Here. Sci Am 2007. 

(44)  Bierbaum, R. M.; Matson, P. A. Energy in the Context of Sustainability. Daedalus 2013, 

142 (1), 146–161. h(cid:425)ps://doi.org/10.1162/DAED_a_00191. 

(45)  El Chaar, L.; lamont, L. A.; El Zein, N. Review of Photovoltaic Technologies. Renewable and 

Sustainable Energy Reviews 2011, 15 (5), 2165–2175. 
h(cid:425)ps://doi.org/10.1016/j.rser.2011.01.004. 

(46)  Parida, B.; Iniyan, S.; Goic, R. A Review of Solar Photovoltaic Technologies. Renewable and 

Sustainable Energy Reviews 2011, 15 (3), 1625–1636. 
h(cid:425)ps://doi.org/10.1016/j.rser.2010.11.032. 

(47)  Hagfeldt, A.; Boschloo, G.; Sun, L.; Kloo, L.; Pe(cid:425)ersson, H. Dye-Sensi(cid:415)zed Solar Cells. Chem 

Rev 2010, 110 (11), 6595–6663. h(cid:425)ps://doi.org/10.1021/cr900356p. 

(48)  Liu, Q.; Jiang, Y.; Jin, K.; Qin, J.; Xu, J.; Li, W.; Xiong, J.; Liu, J.; Xiao, Z.; Sun, K.; Yang, S.; 

Zhang, X.; Ding, L. 18% Eﬃciency Organic Solar Cells. Sci Bull (Beijing) 2020, 65 (4), 272–
275. h(cid:425)ps://doi.org/10.1016/j.scib.2020.01.001. 

(49)  Lu, M. N.; Su, T.; Pylnev, M.; Long, Y.; Wu, T.; Tsai, M.; Wei, T. Stepwise Op(cid:415)mizing 

Photovoltaic Performance of Dye-sensi(cid:415)zed Cells Made under 50-lux Dim Light. Progress 

103 

 
 
in Photovoltaics: Research and Applica(cid:415)ons 2021, 29 (5), 533–545. 
h(cid:425)ps://doi.org/10.1002/pip.3406. 

(50)  Cao, Y.; Liu, Y.; Zakeeruddin, S. M.; Hagfeldt, A.; Grätzel, M. Direct Contact of Selec(cid:415)ve 

Charge Extrac(cid:415)on Layers Enables High-Eﬃciency Molecular Photovoltaics. Joule 2018, 2 
(6), 1108–1117. h(cid:425)ps://doi.org/10.1016/j.joule.2018.03.017. 

(51)  Freitag, M.; Teuscher, J.; Saygili, Y.; Zhang, X.; Giordano, F.; Liska, P.; Hua, J.; Zakeeruddin, 
S. M.; Moser, J.-E.; Grätzel, M.; Hagfeldt, A. Dye-Sensi(cid:415)zed Solar Cells for Eﬃcient Power 
Genera(cid:415)on under Ambient Ligh(cid:415)ng. Nat Photonics 2017, 11 (6), 372–378. 
h(cid:425)ps://doi.org/10.1038/nphoton.2017.60. 

(52)  Kakiage, K.; Aoyama, Y.; Yano, T.; Oya, K.; Fujisawa, J.; Hanaya, M. Highly-Eﬃcient Dye-

Sensi(cid:415)zed Solar Cells with Collabora(cid:415)ve Sensi(cid:415)za(cid:415)on by Silyl-Anchor and Carboxy-Anchor 
Dyes. Chemical Communica(cid:415)ons 2015, 51 (88), 15894–15897. 
h(cid:425)ps://doi.org/10.1039/C5CC06759F. 

(53)  Grätzel, M. Dye-Sensi(cid:415)zed Solar Cells. Journal of Photochemistry and Photobiology C: 
Photochemistry Reviews 2003, 4 (2), 145–153. h(cid:425)ps://doi.org/10.1016/S1389-
5567(03)00026-1. 

(54)  Sapp, S. A.; Ellio(cid:425), C. M.; Contado, C.; Caramori, S.; Bignozzi, C. A. Subs(cid:415)tuted 

Polypyridine Complexes of Cobalt(II/III) as Eﬃcient Electron-Transfer Mediators in Dye-
Sensi(cid:415)zed Solar Cells. J Am Chem Soc 2002, 124 (37), 11215–11222. 
h(cid:425)ps://doi.org/10.1021/ja027355y. 

(55)  Nusbaumer, H.; Zakeeruddin, S. M.; Moser, J.; Grätzel, M. An Alterna(cid:415)ve Eﬃcient Redox 
Couple for the Dye-Sensi(cid:415)zed Solar Cell System. Chemistry – A European Journal 2003, 9 
(16), 3756–3763. h(cid:425)ps://doi.org/10.1002/chem.200204577. 

(56)  Yum, J.-H.; Baranoﬀ, E.; Kessler, F.; Moehl, T.; Ahmad, S.; Bessho, T.; Marchioro, A.; 

Ghadiri, E.; Moser, J.-E.; Yi, C.; Nazeeruddin, Md. K.; Grätzel, M. A Cobalt Complex Redox 
Shu(cid:425)le for Dye-Sensi(cid:415)zed Solar Cells with High Open-Circuit Poten(cid:415)als. Nat Commun 
2012, 3 (1), 631. h(cid:425)ps://doi.org/10.1038/ncomms1655. 

(57)  Boschloo, G.; Hagfeldt, A. Characteris(cid:415)cs of the Iodide/Triiodide Redox Mediator in Dye-

Sensi(cid:415)zed Solar Cells. Acc Chem Res 2009, 42 (11), 1819–1826. 
h(cid:425)ps://doi.org/10.1021/ar900138m. 

(58)  Saygili, Y.; Söderberg, M.; Pellet, N.; Giordano, F.; Cao, Y.; Muñoz-García, A. B.; 

Zakeeruddin, S. M.; Vlachopoulos, N.; Pavone, M.; Boschloo, G.; Kavan, L.; Moser, J.-E.; 
Grätzel, M.; Hagfeldt, A.; Freitag, M. Copper Bipyridyl Redox Mediators for Dye-Sensi(cid:415)zed 
Solar Cells with High Photovoltage. J Am Chem Soc 2016, 138 (45), 15087–15096. 
h(cid:425)ps://doi.org/10.1021/jacs.6b10721. 

104 

 
 
(59)  Feldt, S. M.; Wang, G.; Boschloo, G.; Hagfeldt, A. Eﬀects of Driving Forces for 

Recombina(cid:415)on and Regenera(cid:415)on on the Photovoltaic Performance of Dye-Sensi(cid:415)zed 
Solar Cells Using Cobalt Polypyridine Redox Couples. The Journal of Physical Chemistry C 
2011, 115 (43), 21500–21507. h(cid:425)ps://doi.org/10.1021/jp2061392. 

(60)  Mosconi, E.; Yum, J.-H.; Kessler, F.; Gómez García, C. J.; Zuccaccia, C.; Cin(cid:415), A.; 

Nazeeruddin, M. K.; Grätzel, M.; De Angelis, F. Cobalt Electrolyte/Dye Interac(cid:415)ons in Dye-
Sensi(cid:415)zed Solar Cells: A Combined Computa(cid:415)onal and Experimental Study. J Am Chem 
Soc 2012, 134 (47), 19438–19453. h(cid:425)ps://doi.org/10.1021/ja3079016. 

(61)  Rorabacher, D. B. Electron Transfer by Copper Centers. Chem Rev 2004, 104 (2), 651–698. 

h(cid:425)ps://doi.org/10.1021/cr020630e. 

(62)  Kaim, W.; Rall, J. Copper—A“Modern” Bioelement. Angewandte Chemie Interna(cid:415)onal 
Edi(cid:415)on in English 1996, 35 (1), 43–60. h(cid:425)ps://doi.org/10.1002/anie.199600431. 

(63)  Birker, P. J. M. W. L.; Helder, J.; Henkel, G.; Krebs, B.; Reedijk, J. Synthesis and 

Spectroscopic Characteriza(cid:415)on of Copper(I) and Copper(II) Complexes with 1,6-Bis(2-
Benzimidazolyl)-2,5-Dithiahexane (BBDH). X-Ray Structure of Trigonal-Bipyramidal 
Coordinated [Cu(BBDH)Cl]Cl.2C2H5OH. Inorg Chem 1982, 21 (1), 357–363. 
h(cid:425)ps://doi.org/10.1021/ic00131a064. 

(64)  Freitag, M.; Daniel, Q.; Pazoki, M.; Sveinbjörnsson, K.; Zhang, J.; Sun, L.; Hagfeldt, A.; 
Boschloo, G. High-Eﬃciency Dye-Sensi(cid:415)zed Solar Cells with Molecular Copper 
Phenanthroline as Solid Hole Conductor. Energy Environ Sci 2015, 8 (9), 2634–2637. 
h(cid:425)ps://doi.org/10.1039/C5EE01204J. 

(65)  Rodrigues, R. R.; Lee, J. M.; Taylor, N. S.; Cheema, H.; Chen, L.; Fortenberry, R. C.; 

Delcamp, J. H.; Jurss, J. W. Copper-Based Redox Shu(cid:425)les Supported by Preorganized 
Tetradentate Ligands for Dye-Sensi(cid:415)zed Solar Cells. Dalton Transac(cid:415)ons 2020, 49 (2), 
343–355. h(cid:425)ps://doi.org/10.1039/C9DT04030G. 

(66)  Boschloo, G.; Häggman, L.; Hagfeldt, A. Quan(cid:415)ﬁca(cid:415)on of the Eﬀect of 4- Tert -

Butylpyridine Addi(cid:415)on to I - /I 3 - Redox Electrolytes in Dye-Sensi(cid:415)zed Nanostructured TiO 
2 Solar Cells. J Phys Chem B 2006, 110 (26), 13144–13150. 
h(cid:425)ps://doi.org/10.1021/jp0619641. 

(67)  Koops, S. E.; O’Regan, B. C.; Barnes, P. R. F.; Durrant, J. R. Parameters Inﬂuencing the 

Eﬃciency of Electron Injec(cid:415)on in Dye-Sensi(cid:415)zed Solar Cells. J Am Chem Soc 2009, 131 
(13), 4808–4818. h(cid:425)ps://doi.org/10.1021/ja8091278. 

(68)  Long, H.; Zhou, D.; Zhang, M.; Peng, C.; Uchida, S.; Wang, P. Probing Dye-Correlated 

Interplay of Energe(cid:415)cs and Kine(cid:415)cs in Mesoscopic Titania Solar Cells with 4- Tert -

105 

 
 
Butylpyridine. The Journal of Physical Chemistry C 2011, 115 (29), 14408–14414. 
h(cid:425)ps://doi.org/10.1021/jp202826m. 

(69)  Giordano, M.; Volpi, G.; Bonomo, M.; Mariani, P.; Garino, C.; Viscardi, G. Methoxy-

Subs(cid:415)tuted Copper Complexes as Possible Redox Mediators in Dye-Sensi(cid:415)zed Solar Cells. 
New Journal of Chemistry 2021, 45 (34), 15303–15311. 
h(cid:425)ps://doi.org/10.1039/D1NJ02577E. 

(70)  Deng, Z.; Yang, X.; Yang, K.; Zhang, L.; Wang, H.; Wang, X.; Sun, L. Helical Copper Redox 
Mediator with Low Electron Recombina(cid:415)on for Dye-Sensi(cid:415)zed Solar Cells. ACS Sustain 
Chem Eng 2021, 9 (15), 5252–5259. h(cid:425)ps://doi.org/10.1021/acssuschemeng.0c08195. 

(71)  Kavan, L.; Saygili, Y.; Freitag, M.; Zakeeruddin, S. M.; Hagfeldt, A.; Grätzel, M. 

Electrochemical Proper(cid:415)es of Cu(II/I)-Based Redox Mediators for Dye-Sensi(cid:415)zed Solar 
Cells. Electrochim Acta 2017, 227, 194–202. 
h(cid:425)ps://doi.org/10.1016/j.electacta.2016.12.185. 

(72)  Kannanku(cid:425)y, K.; Chen, C.-C.; Nguyen, V. S.; Lin, Y.-C.; Chou, H.-H.; Yeh, C.-Y.; Wei, T.-C. Tert -
Butylpyridine Coordina(cid:415)on with [Cu(Dmp) 2 ] 2+/+ Redox Couple and Its Connec(cid:415)on to the 
Stability of the Dye-Sensi(cid:415)zed Solar Cell. ACS Appl Mater Interfaces 2020, 12 (5), 5812–
5819. h(cid:425)ps://doi.org/10.1021/acsami.9b19119. 

(73)  Addison, A. W.; Burke, P. J.; Henrick, K.; Rao, T. N.; Sinn, E. Pentacoordinate Copper 

Complexes of Nitrogen-Sulfur Donors: Structural Chemistry of Two Complexes of Bis(2-(2-
Benzimidazolyl)Ethyl) Sulﬁde with the Sulfur Alterna(cid:415)vely in Equatorial and Axial 
Coordina(cid:415)on Modes. Inorg Chem 1983, 22 (24), 3645–3653. 
h(cid:425)ps://doi.org/10.1021/ic00166a030. 

(74)  Yang, L.; Powell, D. R.; Houser, R. P. Structural Varia(cid:415)on in Copper( <scp>i</Scp> ) 
Complexes with Pyridylmethylamide Ligands: Structural Analysis with a New Four-
Coordinate Geometry Index, τ 4. Dalton Trans. 2007, No. 9, 955–964. 
h(cid:425)ps://doi.org/10.1039/B617136B. 

(75)  Niu, J.; Zhou, H.; Li, Z.; Xu, J.; Hu, S. An Eﬃcient Ullmann-Type C−O Bond Forma(cid:415)on 

Catalyzed by an Air-Stable Copper(I)−Bipyridyl Complex. J Org Chem 2008, 73 (19), 7814–
7817. h(cid:425)ps://doi.org/10.1021/jo801002c. 

(76)  Linfoot, C. L.; Richardson, P.; Hewat, T. E.; Moudam, O.; Forde, M. M.; Collins, A.; White, 
F.; Robertson, N. Subs(cid:415)tuted [Cu(i)(POP)(Bipyridyl)] and Related Complexes: Synthesis, 
Structure, Proper(cid:415)es and Applica(cid:415)ons to Dye-Sensi(cid:415)sed Solar Cells. Dalton Transac(cid:415)ons 
2010, 39 (38), 8945. h(cid:425)ps://doi.org/10.1039/c0dt00190b. 

(77)  Zimmer, K. D.; Shoemaker, R.; Ruminski, R. R. Synthesis and Characteriza(cid:415)on of a 

Fluxional Re(I) Carbonyl Complex Fac-[Re(CO)3(Dpop(cid:3387))Cl] with the Nominally Tri-Dentate 

106 

 
 
Ligand Dipyrido(2,3-a:3(cid:3387),2(cid:3387)-j)Phenazine (Dpop(cid:3387)). Inorganica Chim Acta 2006, 359 (5), 
1478–1484. h(cid:425)ps://doi.org/10.1016/j.ica.2005.11.042. 

(78)  Zhang, J.; Siu, K.; Lin, C. H.; Canary, J. W. Conforma(cid:415)onal Dynamics of Cu(i) Complexes of 
Tripodal Ligands: Steric Control of Molecular Mo(cid:415)on. New Journal of Chemistry 2005, 29 
(9), 1147. h(cid:425)ps://doi.org/10.1039/b509050d. 

(79)  Colombo, A.; Dragone(cid:427), C.; Magni, M.; Roberto, D.; Demar(cid:415)n, F.; Caramori, S.; Bignozzi, 
C. A. Eﬃcient Copper Mediators Based on Bulky Asymmetric Phenanthrolines for DSSCs. 
ACS Appl Mater Interfaces 2014, 6 (16), 13945–13955. 
h(cid:425)ps://doi.org/10.1021/am503306f. 

(80)  Ha(cid:425)ori, S.; Wada, Y.; Yanagida, S.; Fukuzumi, S. Blue Copper Model Complexes with 

Distorted Tetragonal Geometry Ac(cid:415)ng as Eﬀec(cid:415)ve Electron-Transfer Mediators in Dye-
Sensi(cid:415)zed Solar Cells. J Am Chem Soc 2005, 127 (26), 9648–9654. 
h(cid:425)ps://doi.org/10.1021/ja0506814. 

(81)  Saygili, Y.; Söderberg, M.; Pellet, N.; Giordano, F.; Cao, Y.; Muñoz-García, A. B.; 

Zakeeruddin, S. M.; Vlachopoulos, N.; Pavone, M.; Boschloo, G.; Kavan, L.; Moser, J.-E.; 
Grätzel, M.; Hagfeldt, A.; Freitag, M. Copper Bipyridyl Redox Mediators for Dye-Sensi(cid:415)zed 
Solar Cells with High Photovoltage. J Am Chem Soc 2016, 138 (45), 15087–15096. 
h(cid:425)ps://doi.org/10.1021/jacs.6b10721. 

(82)  Le Fur, M.; Beyler, M.; Le Poul, N.; Lima, L. M. P.; Le Mest, Y.; Delgado, R.; Platas-Iglesias, 
C.; Pa(cid:415)nec, V.; Tripier, R. Improving the Stability and Inertness of Cu( <scp>ii</Scp> ) and 
Cu( <scp>i</Scp> ) Complexes with Methylthiazolyl Ligands by Tuning the Macrocyclic 
Structure. Dalton Transac(cid:415)ons 2016, 45 (17), 7406–7420. 
h(cid:425)ps://doi.org/10.1039/C6DT00385K. 

(83)  Cong, J.; Kinschel, D.; Daniel, Q.; Safdari, M.; Gabrielsson, E.; Chen, H.; Svensson, P. H.; 

Sun, L.; Kloo, L. Bis(1,1-Bis(2-Pyridyl)Ethane)Copper( <scp>i</Scp> / <scp>ii</Scp> ) as an 
Eﬃcient Redox Couple for Liquid Dye-Sensi(cid:415)zed Solar Cells. J Mater Chem A Mater 2016, 
4 (38), 14550–14554. h(cid:425)ps://doi.org/10.1039/C6TA06782D. 

(84)  Yang, K.; Yang, X.; Deng, Z.; Jiang, M. High Stability Tetradentate Ligand Copper 

Complexes and Organic Small Molecule Hybrid Electrolyte for Dye-Sensi(cid:415)zed Solar Cells. 
Electrochim Acta 2022, 432, 141108. h(cid:425)ps://doi.org/10.1016/j.electacta.2022.141108. 

(85)  Noviandri, I.; Brown, K. N.; Fleming, D. S.; Gulyas, P. T.; Lay, P. A.; Masters, A. F.; Phillips, L. 
The Decamethylferrocenium/Decamethylferrocene Redox Couple: A Superior Redox 
Standard to the Ferrocenium/Ferrocene Redox Couple for Studying Solvent Eﬀects on the 
Thermodynamics of Electron Transfer. Journal of Physical Chemistry B 1999, 103 (32), 
6713–6722. h(cid:425)ps://doi.org/10.1021/jp991381+. 

107 

 
 
(86)  Yang, E. S.; Chan, M.-S.; Wahl, A. C. Electron Exchange between Ferrocene and 

Ferrocenium Ion. Eﬀects of Solvent and of Ring Subs(cid:415)tu(cid:415)on on the Rate. J Phys Chem 
1980, 84 (23), 3094–3099. h(cid:425)ps://doi.org/10.1021/j100460a025. 

(87)  Michaels, H.; Rinderle, M.; Freitag, R.; Benesperi, I.; Edvinsson, T.; Socher, R.; Gagliardi, 
A.; Freitag, M. Dye-Sensi(cid:415)zed Solar Cells under Ambient Light Powering Machine 
Learning: Towards Autonomous Smart Sensors for the Internet of Things. Chem Sci 2020, 
11 (11), 2895–2906. h(cid:425)ps://doi.org/10.1039/C9SC06145B. 

(88)  Bechlars, B.; D’Alessandro, D. M.; Jenkins, D. M.; Iavarone, A. T.; Glover, S. D.; Kubiak, C. P.; 

Long, J. R. High-Spin Ground States via Electron Delocaliza(cid:415)on in Mixed-Valence 
Imidazolate-Bridged Divanadium Complexes. Nat Chem 2010, 2 (5), 362–368. 
h(cid:425)ps://doi.org/10.1038/nchem.585. 

(89)  Bechlars, B.; D’Alessandro, D. M.; Jenkins, D. M.; Iavarone, A. T.; Glover, S. D.; Kubiak, C. P.; 

Long, J. R. High-Spin Ground States via Electron Delocaliza(cid:415)on in Mixed-Valence 
Imidazolate-Bridged Divanadium Complexes. Nat Chem 2010, 2 (5), 362–368. 
h(cid:425)ps://doi.org/10.1038/nchem.585. 

(90)  Raithel, A. L.; Kim, T.-Y.; Nielsen, K. C.; Staples, R. J.; Hamann, T. W. Low-Spin Cobalt( 

<scp>ii</Scp> ) Redox Shu(cid:425)le by Isocyanide Coordina(cid:415)on. Sustain Energy Fuels 2020, 4 
(5), 2497–2507. h(cid:425)ps://doi.org/10.1039/D0SE00314J. 

(91)  Swarts, P. J.; Conradie, J. Redox Data of Ferrocenylcarboxylic Acids in Dichloromethane 

and Acetonitrile. Data Brief 2020, 30, 105650. 
h(cid:425)ps://doi.org/10.1016/j.dib.2020.105650. 

(92)  Baillargeon, J.; Xie, Y.; Hamann, T. W. Bifurca(cid:415)on of Regenera(cid:415)on and Recombina(cid:415)on in 
Dye-Sensi(cid:415)zed Solar Cells via Electronic Manipula(cid:415)on of Tandem Cobalt Redox Shu(cid:425)les. 
ACS Appl Mater Interfaces 2017, 9 (39), 33544–33548. 
h(cid:425)ps://doi.org/10.1021/acsami.7b01626. 

(93)  Kannanku(cid:425)y, K.; Chen, C.-C.; Nguyen, V. S.; Lin, Y.-C.; Chou, H.-H.; Yeh, C.-Y.; Wei, T.-C. Tert 
-Butylpyridine Coordina(cid:415)on with [Cu(Dmp) 2 ] 2+/+ Redox Couple and Its Connec(cid:415)on to 
the Stability of the Dye-Sensi(cid:415)zed Solar Cell. ACS Appl Mater Interfaces 2020, 12 (5), 
5812–5819. h(cid:425)ps://doi.org/10.1021/acsami.9b19119. 

(94)  Yang, K.; Yang, X.; Deng, Z.; Jiang, M. High Stability Tetradentate Ligand Copper 

Complexes and Organic Small Molecule Hybrid Electrolyte for Dye-Sensi(cid:415)zed Solar Cells. 
Electrochim Acta 2022, 432, 141108. h(cid:425)ps://doi.org/10.1016/j.electacta.2022.141108. 

(95)  Devdass, A.; Watson, J.; Firestone, E.; Hamann, T. W.; Delcamp, J. H.; Jurss, J. W. An 

Eﬃcient Copper-Based Redox Shu(cid:425)le Bearing a Hexadentate Polypyridyl Ligand for DSCs 

108 

 
 
under Low-Light Condi(cid:415)ons. ACS Appl Energy Mater 2022, 5 (5), 5964–5973. 
h(cid:425)ps://doi.org/10.1021/acsaem.2c00344. 

(96)  Bernardo, M. M.; Heeg, M. J.; Schroeder, R. R.; Ochrymowycz, L. A.; Rorabacher, D. B. 

Comparison of the Inﬂuence of Saturated Nitrogen and Sulfur Donor Atoms on the 
Proper(cid:415)es of Copper(II/I)-Macrocyclic Polyamino Polythiaether Ligand Complexes: Redox 
Poten(cid:415)als and Protona(cid:415)on and Stability Constants of CuIL Species and New Structural 
Data. Inorg Chem 1992, 31 (2), 191–198. h(cid:425)ps://doi.org/10.1021/ic00028a013. 

(97)  Rorabacher, D. B. Electron Transfer by Copper Centers. Chem Rev 2004, 104 (2), 651–698. 

h(cid:425)ps://doi.org/10.1021/cr020630e. 

(98)  Chaka, G.; Kandegedara, A.; Heeg, M. J.; Rorabacher, D. B. Compara(cid:415)ve Study of Donor 

Atom Eﬀects on the Thermodynamic and Electron-Transfer Kine(cid:415)c Proper(cid:415)es of Copper( 
<scp>ii</Scp> / <scp>i</Scp> ) Complexes with Sexadentate Macrocyclic Ligands. [Cu II/I 
([18]AneS 4 N 2 )] and [Cu II/I ([18]AneS 4 O 2 )]. Dalton Trans. 2007, No. 4, 449–458. 
h(cid:425)ps://doi.org/10.1039/B612252C. 

(99)  Aronne, L.; Yu, Q.; Ochrymowycz, L. A.; Rorabacher, D. B. Eﬀect of Macrocyclic Ligand 
Constraints upon the Kine(cid:415)cs of Complex Forma(cid:415)on and Dissocia(cid:415)on and Metal Ion 
Exchange. Copper(II) Complexes with Cyclohexanediyl Deriva(cid:415)ves of the Cyclic 
Tetrathiaether [14]AneS4 in 80% Methanol. Inorg Chem 1995, 34 (7), 1844–1851. 
h(cid:425)ps://doi.org/10.1021/ic00111a036. 

(100)  Deng, H.; Kebarle, P. Bond Energies of Copper Ion−Ligand L Complexes CuL 2 + Determined 
in the Gas Phase by Ion−Ligand Exchange Equilibria Measurements. J Am Chem Soc 1998, 
120 (12), 2925–2931. h(cid:425)ps://doi.org/10.1021/ja973814x. 

109 

 
 
  
 
APPENDIX A: CHAPTER 2 SUPPLEMENTARY DATA 

Figure 2.1A: 1H-NMR spectrum (500 MHz, CDCl3) of dbmed. 

110 

 
 
Figure 2.2A: 1H-NMR spectrum (500 MHz, CDCl3) of [Cu(dbmed)](OTf). 

111 

 
 
 
 
3

l

c
d
c

6
2
.
7

M
C
D

0
3
.
5

r
e
h
t
E

9
4
.
3

r
e
h
t
E

7
4
.
3

r
e
h
t
E

1
2
.
1

19

18

17

16

15

14

13

12

11

10
f1 (ppm)

9

8

7

6

5

4

3

2

1

Figure 2.3A: 1H-NMR spectrum (500 MHz, CDCl3) of [Cu(dbmed)](OTf)2. 

250

240

230

220

210

200

190

180

170

160

150

140

130

120

110

100

90

80

70

60

50

40

30

20

10

0

-10

-20

112 

 
 
 
 
 
 
 
 
Figure 2.4A: Crystal structure of the [Cu(dbmed)]OTf species. 

Table 2.1A: Selected bond lengths and angles for the [Cu(dbmed)]OTf complex. The bond 
lengths are reported in angstroms (Å) and bond angles are in degrees (°). The standard 
devia(cid:415)ons of each value are shown in parenthesis. 

Atom 

Atom 

Length/
Å 

Atom 

Atom 

Atom 

Angle/° 

Cu1 

Cu1 

Cu1 

Cu1 

N1 

N2 

N3 

N4 

2.268(5)  N1 

1.950(5)  N2 

2.302(5)  N2 

1.953(5)  N2 

N4 

N4 

Cu1 

Cu1 

Cu1 

Cu1 

Cu1 

Cu1 

N3 

N1 

N3 

N4 

N1 

N3 

83.46(18
) 
120.1(2) 

82.8(2) 

151.0(2) 

83.8(2) 

118.5(2) 

113 

 
 
 
 
 
 
 
 
 
 
 
APPENDIX B: CHAPTER 3 SUPPLEMENTARY DATA 

Table 3.1A: Selected bond lengths and angles for the [Cu(bpyPY4)]OTf complex. The bond 
lengths are reported in angstroms (Å) and bond angles are in degrees (°). The standard 
devia(cid:415)ons of each value are shown in parenthesis. 

Atom  Atom 

Length/Å  Atom  Atom  Atom  Angle/° 

Cu1 

Cu1 

Cu1 

Cu1 

N1 

N2 

N3 

N4 

2.019(2)  N1 

Cu1  N2 

87.61(9) 

2.086(2)  N1 

Cu1  N3 

135.49(10) 

2.078(2)  N3 

Cu1  N2 

78.64(10) 

2.011(2)  N4 

Cu1  N1 

126.30(9) 

N4 

Cu1  N2 

137.50(9) 

N4 

Cu1  N3 

89.34(9) 

Table 3.2A: Selected bond lengths and angles for the [Cu(bpyPY4)]OTf2 complex. The bond 
lengths are reported in angstroms (Å) and bond angles are in degrees (°). The standard 
devia(cid:415)ons of each value are shown in parenthesis. 

Atom  Atom 

Length/Å  Atom  Atom  Atom  Angle/° 

Cu1 

Cu1 

Cu1 

Cu1 

Cu1 

N1 

N2 

N3 

N4 

N5 

1.969(3)  N1 

Cu1  N2 

81.88(11) 

2.208(3)  N1 

Cu1  N3 

88.16(12) 

2.053(3)  N1 

Cu1  N4 

81.95(12) 

1.993(3)  N1 

Cu1  N5 

172.85(13) 

1.980(3)  N3 

Cu1  N2 

86.34(11) 

N4 

Cu1  N2 

133.89(12) 

N4 

Cu1  N3 

135.76(12) 

N5 

Cu1  N2 

96.68(11) 

N5 

Cu1  N3 

98.76(12) 

N5 

Cu1  N4 

94.20(12) 

114 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
r
e
h
t
E

6
4
.
3

r
e
h
t
E

5
4
.
3

r
e
h
t
E

3
4
.
3

r
e
h
t
E

2
4
.
3

e
n
o
t
e
c
A

4
2
.
2

n
c
3
d
c

7
9
.
1

n
c
3
d
c

6
9
.
1

N
C
A

9
9
.
1

r
e
h
t
E

5
1
.
1

r
e
h
t
E

4
1
.
1

r
e
h
t
E

3
1
.
1

19

18

17

16

15

14

13

12

11

10
f1 (ppm )

9

8

7

6

5

4

3

2

1

Figure 3.1A: : 1H-NMR spectrum (500 MHz, CD3CN) of [Cu(bpyPY4)](TFSI)2. 

400

350

300

250

200

150

100

50

0

115 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
APPENDIX C: CHAPTER 4 SUPPLEMENTARY DATA 

Figure 4.1A: 1H-NMR of [Cu(PY5)]OTf in deuterated acetonitrile. 

116 

 
 
 
Figure 4.2A: 1H-NMR of synthesized [Cu(PY5)]OTf2 in deuterated acetonitrile. 

117 

 
 
 
M
C
D

7
4
.
5

r
e
h
t
e

6
4
.
3

r
e
h
t
e

5
4
.
3

r
e
h
t
e

3
4
.
3

r
e
h
t
e

2
4
.
3

n
c
3
d
c

9
9
.
1

n
c
3
d
c

8
9
.
1

n
c
3
d
c

7
9
.
1

n
c
3
d
c

7
9
.
1

n
c
3
d
c

6
9
.
1

n
c
3
d
c

6
9
.
1

n
c
3
d
c

5
9
.
1

r
e
h
t
e

6
1
.
1

r
e
h
t
e

4
1
.
1

r
e
h
t
e

3
1
.
1

19

18

17

16

15

14

13

12

11

10
f1 (ppm)

9

8

7

6

5

4

3

2

1

Figure 4.3A: 1H-NMR of dispropor(cid:415)onated [Cu(PY5)]OTf2 in deuterated acetonitrile. 

2400

2300

2200

2100

2000

1900

1800

1700

1600

1500

1400

1300

1200

1100

1000

900

800

700

600

500

400

300

200

100

0

-100

-200

118 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 4.4A: 1H-NMR of [Cu(PY5)]TFSI in deuterated acetonitrile. 

119 

 
 
 
 
Figure 4.5A: 1H-NMR of [Cu(PY5)]TFSI2 in deuterated acetonitrile. 

Table 4.1A: Selected bond lengths and angles for the [Cu(PY5)]OTf complex. The bond lengths 
are reported in angstroms (Å) and bond angles are in degrees (°). The standard devia(cid:415)ons of 
each value are shown in parenthesis. 

Atom  Atom 
Cu1 
Cu1 
Cu1 
Cu1 

N15 
N31 
N37 
N44 

Length/Å 
2.057(3) 
2.100(4) 
1.980(4) 
1.938(4) 

Atom  Atom  Atom  Angle/° 
N15 
N37 
N37 
N44 
N44 
N44 

90.20(14) 
88.69(15) 
90.51(15) 
97.32(14) 
118.19(15) 
150.50(16) 

Cu1 
Cu1 
Cu1 
Cu1 
Cu1 
Cu1 

N31 
N15 
N31 
N15 
N31 
N37 

120 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Table 4.2A: Selected bond lengths and angles for the [Cu(PY5)]OTf2 complex. The bond lengths 
are reported in angstroms (Å) and bond angles are in degrees (°). The standard devia(cid:415)ons of 
each value are shown in parenthesis. 

N1 
N2 
N3 
N4 
N5 
N6 

2.082(3) 
2.035(3) 
2.160(3) 
2.038(3) 
2.090(3) 
2.369(3) 

Atom  Atom  Length/Å  Atom  Atom  Atom  Angle/° 
87.45(10) 
Cu1 
175.41(11) 
Cu1 
90.63(11) 
Cu1 
82.47(11) 
Cu1 
87.50(11) 
Cu1 
176.29(11) 
Cu1 
96.93(11) 
91.25(12) 
177.83(11) 
99.16(11) 
89.23(11) 
81.17(11) 
92.07(12) 
87.98(11) 
93.94(12) 

Cu1 
Cu1 
Cu1 
Cu1 
Cu1 
Cu1 
Cu1 
Cu1 
Cu1 
Cu1 
Cu1 
Cu1 
Cu1 
Cu1 
Cu1 

N1 
N1 
N1 
N2 
N2 
N2 
N2 
N2 
N3 
N4 
N4 
N4 
N4 
N5 
N5 

N3 
N5 
N6 
N1 
N3 
N4 
N5 
N6 
N6 
N1 
N3 
N5 
N6 
N3 
N6 

Table 4.3A: Selected bond lengths and angles for the [Cu(PY5)]TFSI2 complex. The bond lengths 
are reported in angstroms (Å) and bond angles are in degrees (°). The standard devia(cid:415)ons of 
each value are shown in parenthesis. 

Atom  Atom 
N(2) 
Cu(1) 
N(3) 
Cu(1) 
N(4) 
Cu(1) 
N(5) 
Cu(1) 
N(6) 
Cu(1) 

Length/Å 

2.0224(14) 
2.0675(14) 
2.1130(14) 
2.0491(14) 
2.0262(14) 

Atom  Atom  Atom 
N(3) 
Cu(1) 
N(2) 
N(4) 
Cu(1) 
N(2) 
N(5) 
Cu(1) 
N(2) 
N(6) 
Cu(1) 
N(2) 
N(4) 
Cu(1) 
N(3) 
N(3) 
Cu(1) 
N(5) 
N(4) 
Cu(1) 
N(5) 
N(3) 
Cu(1) 
N(6) 
N(4) 
Cu(1) 
N(6) 
N(5) 
Cu(1) 
N(6) 

Angle/° 

82.00(6) 
92.78(6) 
98.97(6) 
175.34(6) 
87.35(6) 
175.30(6) 
88.00(6) 
97.08(6) 
91.74(6) 
82.32(6) 

121