Dye-Sensitized Solar Cell

Terms Used on This Page

Short names are written out here once, with the source that defines them; where a source uses a name without defining it, the entry says so. Tags in brackets: (computed) calculated from the numbers given, (assumed) a value chosen by me because no source gives it, (my reading) my interpretation of what a source says, (background knowledge) standard material that none of the sources I read states.

Cell and measurement terms

• DSSC, DSC: dye-sensitized solar cell (Boschloo writes DSC).
• TiO${}_2$, anatase: titanium dioxide; anatase is the crystal form used in the film (Boschloo).
• FTO: fluorine-doped tin oxide, the transparent conducting coating on the glass (Boschloo; RSC Sustainability).
• NHE: normal hydrogen electrode, the reference against which potentials are quoted (background knowledge; Boschloo uses it without defining it).
• AM1.5G: air mass 1.5 global, the standard reference sunlight spectrum with an intensity of 1000 W/m${}^2$; “1 sun” means this light (background knowledge; the sources use it undefined).
• IPCE: incident photon to current conversion efficiency, the fraction of the photons of one wavelength that give an electron in the outer circuit (Boschloo).
• LHE: light-harvesting efficiency, the fraction of the photons that the dye absorbs (background knowledge).
• $V_{oc}$, $J_{sc}$, FF: open-circuit voltage, short-circuit current density and fill factor (Boschloo; Ma et al.; RSC Sustainability).
• PCE, $\eta$: power conversion efficiency (Ma et al.; RSC Sustainability).
• HOMO, LUMO: highest occupied and lowest unoccupied molecular orbital; LUMO+2 is the third-lowest unoccupied orbital (background knowledge; Ma et al. use the terms without defining them).
• TDDFT: real-time time-dependent density functional theory, the calculation Ma et al. use for the injection times.
• lux, CFL, LED, WW: lux is the unit of illuminance, and 1000 lux is a bright indoor level (background knowledge); compact fluorescent lamp; light-emitting diode; warm white (RSC Sustainability defines WW and uses CFL and LED without expanding them).
• Mirror case, both faces: my terms in the limit row, for a cell that emits light only from its front face (rear mirror) and a cell that emits from both faces.
• EPFL, AIST, Fraunhofer ISE: Ecole Polytechnique Federale de Lausanne; National Institute of Advanced Industrial Science and Technology (Japan); Fraunhofer Institute for Solar Energy Systems. The last two are the test centres named in the efficiency tables (background knowledge for the expansions).

Ligands (the molecules around the metal atom)

• bpy: 2,2’-bipyridine (Yella et al. call Co(bpy)${}_3$ a cobalt tris(bipyridyl) complex; the full ligand name is background knowledge).
• phen: 1,10-phenanthroline (background knowledge; no source defines it).
• bpy-pz: 6-(1H-pyrazol-1-yl)-2,2’-bipyridine (Yum et al.).
• tmby: 4,4’,6,6’-tetramethyl-2,2’-bipyridine (Cao et al. 2017; Saygili et al.). Boschloo writes Cu(tmp) in Table 1 and Cu(tmbpy)${}_2$ in the text and defines neither; I take both to mean tmby (my reading, see the last question).
• dcbpy: 4,4’-dicarboxy-2,2’-bipyridine, the anchoring ligand of N3 (Nazeeruddin et al. 2005 write dcbpyH${}_2$ for the acid).
• NCS: the isothiocyanate group, bonded to ruthenium through nitrogen (the sources write thiocyanato or SCN; the bonding is background knowledge).

Redox couples and hole conductor

• I${}^{-}$/I${}_3^{-}$: iodide and triiodide. Br${}_3^{-}$/Br${}^{-}$: tribromide and bromide.
• Co(bpy)${}_3$, Co(phen)${}_3$, Co(bpy-pz)${}_2$: the cobalt(III)/cobalt(II) complexes of three bipyridines, three phenanthrolines and two bpy-pz ligands, with six nitrogen atoms around the metal in each case (two per bpy or phen, background knowledge; three per bpy-pz, which Yum et al. call tridentate). The names are assembled from the ligand names, and Yum et al. write the last as [Co(bpy-pz)${}_2$]${}^{3+/2+}$.
• Cu(tmby)${}_2$: bis(tmby)copper(II)/copper(I), used as a mediator or as a solid hole conductor (Cao et al. 2017).
• Spiro-MeOTAD: 2,2’,7,7’-tetrakis(N,N-di-p-methoxyphenylamine)-9,9’-spirobifluorene, an organic solid hole conductor (background knowledge; Boschloo gives only the abbreviation).
• HTM, hole conductor: hole-transporting material, a solid that carries positive charge in place of the liquid electrolyte (Boschloo).

Dyes

• N3: cis-di(thiocyanato)bis(2,2’-bipyridyl-4,4’-dicarboxylate)ruthenium(II), Ru(dcbpy)${}_2$(NCS)${}_2$ (Nazeeruddin et al. 2001).
• N719: “the most commonly used ruthenium bipyridine dye” (Ma et al., no full name given); it is the tetrabutylammonium salt of N3 (background knowledge).
• Black dye: a ruthenium(II) sensitizer with an absorption threshold at 920 nm (Nazeeruddin et al. 2001); its full structural name is not in the text I read (background knowledge: a tri(thiocyanato)terpyridine-tricarboxylate ruthenium complex).
• Z-907: an amphiphilic ruthenium dye, cis-Ru(L)(L’)(SCN)${}_2$ with L = 4,4’-dicarboxylic acid-2,2’-bipyridine and L’ = 4,4’-dinonyl-2,2’-bipyridine (Wang et al.; my reading that the sensitizer described in the abstract is Z-907).
• YD2-o-C8: a zinc porphyrin dye of donor-$\pi$-bridge-acceptor type (Yella et al.).
• Y123: a metal-free organic dye with a cyclopentadithiophene bridge (Yum et al.).
• SM315: a porphyrin dye with a benzothiadiazole unit (Mathew et al.; Ma et al.).
• R6, WS72, S5: donor-$\pi$-acceptor dyes, the labels of Boschloo’s Table 1; the sources I read give no full names.
• Adeka-1, LEG4: an alkoxysilyl-anchor dye and a carboxy-anchor organic dye used together (Kakiage et al.).
• XY1b, D35: organic dyes used alone and together in the indoor paper; neither is defined there.
• SY404, D404, N1: SY404 and D404 have the same donor with an acyloin and a cyanoacrylic acid anchor; N1 is a simplified model dye used in the calculation (Ma et al.).
• Donor-$\pi$-acceptor (DpA): a dye built from an electron-donating group, a conjugated bridge and an acceptor group that also anchors it to the TiO${}_2$ (Boschloo; Ma et al.; the anchoring role is background knowledge).

1. Theory

• Definition: A dye-sensitized solar cell has a mesoporous film of TiO${}_2$ on a transparent conducting glass (coated with FTO), a monolayer of dye on the TiO${}_2$, a redox electrolyte and a counter electrode. Light is absorbed by the dye, not by the semiconductor, which has a gap of 3.2 eV and absorbs only below about 390 nm (Boschloo).
• Film: Boschloo gives an optimal thickness of about 10 $\mu$m for liquid cells and about 2 $\mu$m for solid-state cells. The black-dye cell of Nazeeruddin et al. used a film of 18 $\mu$m with a surface area of 75 m${}^2$/g and a porosity of 0.62, and Ito et al. passed 10 % by optimising the transparent layer and adding a light-scattering topcoat (abstract; no thicknesses given).
• Dyes: N3 is Ru(dcbpy)${}_2$(NCS)${}_2$; the black dye has an absorption threshold at 920 nm, an IPCE above 80 per cent and an efficiency of 10.4 % in the same paper (Nazeeruddin et al.); Z-907 is an amphiphilic ruthenium dye (Wang et al.). The porphyrin and organic dyes of the best cells are YD2-o-C8, Y123, SM315, R6, Adeka-1, LEG4 and WS72 (Boschloo; Yella et al.; Mathew et al.; Kakiage et al.). N719 starts to absorb at about 1.65 eV (750 nm) (Ma et al.).
• Mediators: Table 1 of Boschloo lists the redox potentials against NHE: +0.35 V (I${}^{-}$/I${}_3^{-}$), +0.56 V (Co(bpy)${}_3$), +0.62 V (Co(phen)${}_3$), +0.86 V (Co(bpy-pz)${}_2$), +0.91 V (Cu(tmby)${}_2$, written Cu(tmp) there) and +0.75 V (spiro-MeOTAD in solid-state cells); Yum et al. give 0.86 V against NHE for Co(bpy-pz)${}_2$. The iodide couple loses more than 0.5 V against a typical dye of about 1 V (Boschloo).
• Energetics: The conduction band of anatase is at about $-0.5$ V against NHE, or $-0.7$ V with Mg doping, and the highest voltage reported is 1.4 V (Mg-doped TiO${}_2$, coumarin dye, Br${}_3^{-}$/Br${}^{-}$ couple) (Boschloo).
• Limits: Boschloo gives 33.8 % at 1.3 eV, falling to 25 % for an onset of 650 nm (1.9 eV) with a current of about 17 $mA/c{\mathrm{m}}^2$; Ma et al. give 33.7 % at 1.37 eV and an ideal threshold of 1.3 to 1.4 eV.
• Kinetics: The injection takes 33 to 160 fs in the real-time calculations of Ma et al. on dyes and model dyes, the regeneration is on a sub-nanosecond time scale, the electron lifetime in liquid cells is 1 to 10 ms and the transport time about a factor of ten less; the IPCE of optimised cells is about 90 per cent (Boschloo; Ma et al.).
• Records (efficiency tables, version 64): Green et al. list a cell of 11.9 % (Sharp, 1.005 $\mathrm{c}{\mathrm{m}}^2$, 0.744 V, 22.47 $mA/c{\mathrm{m}}^2$, 71.2 %, AIST, September 2012), a cell of 13.0 $\pm$ 0.4 % (EPFL, 0.1155 $\mathrm{c}{\mathrm{m}}^2$, 1.0396 V, 15.55 $mA/c{\mathrm{m}}^2$, 80.4 %, Fraunhofer ISE, October 2020), a minimodule of 10.7 % and a submodule of 8.8 %.
• Other records: Boschloo lists 12.3 % (YD2-o-C8 and Y123, 2011), 13.0 % (SM315, 2014), 14.3 % (Adeka-1 and LEG4, 2015, open device) and 13.1 % (a copper mediator, 2018), and 11.7 % for solid-state cells (2018); Ren et al. report 15.2 % for cosensitized cells with hydroxamic acid adsorbed first on the TiO${}_2$, with 500 h of operational stability (Nature, published online 2022); Cao et al. (2017) measured 11.0 % for a solid-state cell with a copper(II/I) hole conductor.
• Indoor light: Boschloo gives 32 % at 1000 lux (citing Cao et al. 2018); RSC Sustainability reports 30.48 $\pm$ 0.23 % and 31.33 $\pm$ 0.62 % (co-sensitised) at 1000 lux, 29.35 $\pm$ 0.09 % for a cell of 1.92 $\mathrm{c}{\mathrm{m}}^2$ and about 95 per cent retention after 1000 h under LEDs; Ren et al. give 28.4 to 30.2 % for cells of 2.8 $\mathrm{c}{\mathrm{m}}^2$ over a range of ambient light intensities (press summary).

2. Working Principle

• Cycle: The dye absorbs a photon (S + h$\nu$ $\to$ S${}^{\ast }$), injects an electron into the TiO${}_2$ conduction band in femtoseconds, is regenerated by the iodide (S${}^{+}$ + e${}^{-}$ $\to$ S) and the triiodide is reduced at the counter electrode (I${}_3^{-}$ + 2e${}^{-}$ $\to$ 3I${}^{-}$) (background knowledge; Figure 1 of Boschloo shows the electron-transfer steps).
• Light harvesting: The porous film carries enough dye for an absorbance of 1 or more, which gives 90 % or more of the photons at the peak (computed, assumed absorbance); the scattering layer returns unabsorbed red light to the film (background knowledge; Ito et al. use such a topcoat).
• Injection and collection: With an injection time of 33 to 160 fs and an excited-state decay of 260 ps the injection yield is 0.9994 to 0.9999, and with a transport time one tenth of the lifetime the collection is 0.90 (computed). An IPCE of 90 % then needs each of four steps near 0.97 (computed).
• Voltage: The voltage is set by the Fermi level of the TiO${}_2$ and the redox potential of the mediator; the cells lose 0.11 to 0.21 V against the band edge for iodide and cobalt, and about 0.36 V for the copper and cobalt-pyrazole couples (computed, with the band edge assumed at $-0.5$ V).
• Fill factor and current: The fill factors of 71.2 and 80.4 % are 0.834 and 0.907 of the ideal values, and the current of 22.47 $mA/c{\mathrm{m}}^2$ is 0.661 of the radiative limit of a 1.37 eV absorber (computed).
• Limit: A dye with an onset of 1.65 eV is limited to 29.99 % and an onset of 1.9 eV to 25.05 % (computed, mirror case).

3. Applications

• Indoor and self-powered devices: The 1000 lux efficiencies of 30 to 32 % are measured under a lamp and cannot be compared with 1 sun values (Boschloo; RSC Sustainability; my reading). RSC Sustainability reports a sensor running on the cells with a power output of 73 $\mu$W at 120 lux (LED).
• Building-integrated modules: EPFL announced 1500 modules of 35 by 50 cm on a surface of 300 ${\mathrm{m}}^2$ for the SwissTech Convention Center, with an estimated 8000 kWh per year (EPFL news, 11 December 2012, written before the installation); the modules cover 262.5 ${\mathrm{m}}^2$ (computed).
• Low-cost and diffuse light: Dye cells use cheap oxides, and the cosensitized cells of Ren et al. keep 28.4 to 30.2 % over a wide range of ambient light intensities (press summary).
• Tandem and hybrid ideas: A perovskite sensitiser (CH${}_3$NH${}_3$PbX${}_3$) on mesoporous TiO${}_2$ reached about 15 % in 2013 (Burschka et al.); the Perovskite Solar Cell page treats such devices as perovskite cells.
• Next steps: The sources name higher redox potentials with low losses (Cu and Co complexes), solid-state hole conductors, cosensitization and indoor cells as the active directions (Boschloo; Cao et al. 2017; Ren et al.; RSC Sustainability).

Light Harvesting, Injection and Collection

Name	Formula	Related Formulae	Variables/Constants
Light-harvesting efficiency of the dyed film	$LHE=1-10^{-A}$	$A=\epsilon {\Gamma }_{dye}$	$\begin{array}{r}LHE:fractionofthephotonsabsorbedby\\ thedyeatonewavelength\end{array}$ $A:absorbanceofthedyedfilm$ $\epsilon :extinctioncoefficient$ $\begin{array}{r}{\Gamma }_{dye}:dyeamountperareaoffilm\\ (growswiththesurfaceareaoftheporous\\ Ti{O}_2),noneofthesourcesgivesAorthe\\ roughnessfactor,sothevaluesareassumed;\\ thefilmsareabout10\mu \mathrm{m}(liquid\\ cells)orabout2\mu \mathrm{m}(solid-state)\\ thick(Boschloo),and18\mu \mathrm{m}witha\\ surfaceareaof75m^2/gintheblack-dyecell\\ ofNazeeruddinetal.\end{array}$ $\begin{array}{r}(computed:A=0.5,1,2and3giveLHE=\\ 68.4,90.0,99.0and99.9percent)\end{array}$
Electron injection yield	${\eta }_{inj}=\frac{k_{inj}}{k_{inj}+k_d}=\frac{{\tau }_d}{{\tau }_d+{\tau }_{inj}}$	$k_{inj}=\frac{1}{{\tau }_{inj}}$	$\begin{array}{r}{k}_{inj}:injectionratefromtheexciteddye\\ intotheTi{O}_2\end{array}$ $k_d:decayrateoftheexciteddye$ $\begin{array}{r}{\tau }_{inj}:33fs(SY404),60fs(D404),87fs\\ (N1,LUMO+2)and160fs(N1,LUMO)inthe\\ real-timeTDDFTcalculationsreviewedbyMaet\\ al.,whichdefinethetimeasthemomentwhen\\ 63.2percentoftheelectronshave\\ transferred(N1isasimplifiedmodeldye)\end{array}$ $\begin{array}{r}{\tau }_d:theexcited-statefluorescence\\ lifetimeofR6onA{l}_2{O}_{3}of260ps\\ (Boschloo),injectioncanalsotakeupto\\ tensofpsorevenns(Maetal.)\end{array}$ $\begin{array}{r}(computed:{\tau }_{inj}=33,60,87and160fs\\ with{\tau }_d=260psgive{\eta }_{inj}=0.99987,\\ 0.99977,0.99967and0.99938;aninjection\\ timeof5pswouldgive0.981)\end{array}$
Charge collection efficiency	${\eta }_{cc}=1-\frac{t_{tr}}{{\tau }_n}$	$IPCE=LHE\cdot {\eta }_{inj}\cdot {\eta }_{reg}\cdot {\eta }_{cc}$	$t_{tr}:electrontransporttimeatshortcircuit$ $\begin{array}{r}{\tau }_n:electronlifetimeatopencircuit,\\ Boschloogives{\tau }_n=1to10ms(liquid\\ cells,1sun)andatransporttimeabout\\ oneorderofmagnitudeshorter\end{array}$ $\begin{array}{r}{\eta }_{reg}:regenerationyieldofthedye\\ (sub-nsregeneration,Boschloo),thefirst\\ expressionisbackgroundknowledge,IPCEis\\ about90percentforoptimisedcells\\ (Boschloo)\end{array}$ $\begin{array}{r}(computed:t_{tr}/{\tau }_n=0.1gives{\eta }_{cc}=0.90\\ and0.01gives0.99;anIPCEof0.90withfour\\ equalfactorsneeds0.90^{1/4}=0.974each;the\\ duplicated{\eta }_{reg}cellthatstoodherewas\\ replacedintheaudit)\end{array}$

Voltage, Fill Factor and Efficiency Checks

Name	Formula	Related Formulae	Variables/Constants
Radiative limit at the absorption onset	${\eta }_{rad}=\frac{V_{oc,rad}{J}_{sc,rad}F{F}_0}{P_{in}}$	$J_{sc,rad}=q{\int }_{E_{onset}}^{\infty }\Phi dE$	$\begin{array}{r}{E}_{onset}:absorptiononsetofthedye(N719\\ about1.65eV,750nm;a650nmonsetis\\ 1.90eV)\end{array}$ $\begin{array}{r}\Phi :photonfluxoftheAM1.5Gspectrum,\\ P_{in}=1000.37W/m^2,k_{B}T/q=25.693mV,\\ emissioninto2\pi or,withamirror,\pi ,\\ stepabsorptanceasontheearlierpages;the\\ sourcesgive33.8percentat1.3eVand25\\ percentat1.9eVwith{J}_{sc}about17\\ mA/c{m}^2(Boschloo)and33.7percentat1.37\\ eV(Maetal.)\end{array}$ $k_B:Boltzmannconstant$ $\begin{array}{r}(computed,\pi (2\pi ):1.30eV33.19(32.57)\\ percent,1.34eV33.76(33.15)percent,1.37eV\\ 33.70(33.11)percent,1.40eV33.49(32.92)\\ percent,1.65eV29.99(29.57)percent,V_{oc,rad}1.3733\\ (1.3555)V,J_{sc,rad}24.05mA/c{m}^2,1.90eV\\ 25.05(24.76)percent,16.97mA/c{m}^2;the\\ maximumisat1.335eVinbothcases)\end{array}$
Maximum voltage from the redox level	$V_{max}=\frac{E_{CB}-E_{redox}}{q}$	$r_V=\frac{V_{oc}}{V_{max}}$	$\begin{array}{r}{E}_{CB}:conductionbandedgeofTi{O}_2,about\\ -0.5VagainstNHEforundopedanataseand\\ about-0.7VwithMgdoping(Boschloo)\end{array}$ $\begin{array}{r}{E}_{redox}:redoxpotentialofthemediator,\\ +0.35V(I^{-}/I_3^{-}),+0.56V(Co(bpy{)}_3),\\ +0.62V(Co(phen{)}_3),+0.86V(Co(bpy-pz{)}_2),\\ +0.91V(Cu(tmby{)}_2)and+0.75V\\ (spiro-MeOTAD,solidstate)fromTable1of\\ Boschloo,whowritesCu(tmp)forthecopper\\ couple,and0.86VforCo(bpy-pz{)}_{2}inYumet\\ al.;Boschlo{o}^{\prime }stextsaysthat{E}_{F}isabout\\ 0.1Vmorepositivethan{E}_{CB}at1sun,that\\ isabout0.1eVbelowthebandedge;theuse\\ of-0.5Vforeverycellismyassumption\end{array}$ $\begin{array}{r}(computedwith{E}_{CB}=-0.5V:V_{max}=\\ 0.85,1.06,1.12,1.36,1.41and1.25V,so\\ theloss{V}_{max}-V_{oc}is0.106V(blackdye,\\ 0.744V),0.125V(YD2-o-C8andY123,0.935\\ V),0.150V(SM315,0.91V),0.210V(R6,\\ 0.850V),0.110V(Co(phen{)}_3,1.01V),0.362\\ V(Co(bpy-pz{)}_2,0.998V),0.360V\\ (Cu(tmby{)}_2,1.05V)and0.420V\\ (spiro-MeOTAD,0.83V);therecordof1.4V\\ withB{r}_3^{-}/B{r}^{-}(+0.9V)onMg-dopedTi{O}_2\\ (-0.7V)is0.875of1.6V)\end{array}$
Ideal fill factor	$F{F}_0=\frac{v-\ln \left(v+0.72\right)}{v+1}$	$v=\frac{V_{oc}}{k_{B}T/q}$	$\begin{array}{r}F{F}_0:fillfactorwithoutresistivelosses\\ (backgroundknowledge,asontheearlier\\ pages),onejunction,thecellsarethe\\ entriesofGreenetal.(Sharp0.744V,\\ 71.2percent;EPFL1.0396V,80.4percent)andthe\\ cellsofTable1ofBoschloo\end{array}$ $k_B:Boltzmannconstant$ $\begin{array}{r}(computed:F{F}_0=0.8534for0.744Vand\\ 71.2percentis0.834ofit;0.8862for1.0396\\ Vand80.4percentis0.907;BoschlooTable1:\\ 0.844(0.935V),0.892(0.91V),0.869\\ (0.850V),0.873(1.01V),0.877(0.998\\ V),0.891(1.05V),0.844(0.83V,solid\\ state))\end{array}$
Efficiency from the reported parameters	$\eta =\frac{V_{oc}{J}_{sc}FF}{P_{in}},P_{in}=100mW/c{\mathrm{m}}^2$	$J_{sc}=\frac{\eta P_{in}}{V_{oc}FF}$	$\begin{array}{r}\eta :efficiency,V_{oc}inmV,J_{sc}inmA/c{m}^2,FF\\ asafraction(efficiencyinpercent=\\ V_{oc}[mV]\cdot J_{sc}[mA/c{m}^2]\cdot FF/1000),the\\ entriesarefromGreen\\ etal.(version64)andTable1of\\ Boschloo(1sun),theblack-dyerowof\\ Boschloo(0.744V,22.5,0.712)isthe\\ samecellastheSharpentryofGreenet\\ al.(0.744V,22.47,71.2percent),dated2011\\ byBoschlooandSeptember2012(AIST)by\\ Greenetal.\end{array}$ $\begin{array}{r}(computed:Sharp744,22.47,71.2percent\\ give11.90percent;EPFL1039.6,15.55,80.4\\ percentgive13.00percent;YD2-o-C8and\\ Y12312.25percent(12.3);SM31512.85per\\ cent(13.0);R612.63percent(12.6);Adeka-1\\ andLEG414.25percent(14.3);Co(bpy-pz{)}_2\\ 10.12percent(10.1,and10.08inYumet\\ al.);Cu(tmby{)}_{2}13.02percent(13.1);\\ spiro-MeOTADS57.82percent(7.81);theWS72\\ row(1.07V,11.7,0.79asprinted)gives9.89\\ percentagainst13.8percentprinted,but\\ withtheefficiencyandcurrentcolumns\\ interchanged(13.8mA/c{m}^2)itgives11.67per\\ cent,the11.7percentofthetextof\\ Boschloo;Caoetal.2017:1.08V,13.87\\ mA/c{m}^{2}and0.733give10.98percent(11.0);\\ theSharpcellhas0.661ofthelimitcurrent\\ ofa1.37eVcell)\end{array}$

Stability and Indoor Operation

Name	Formula	Related Formulae	Variables/Constants
Retention and acceleration of stress tests	$R\left(t\right)=\frac{\eta \left(t\right)}{\eta \left(0\right)}$	$r=\frac{1-R}{t}$	$R:retainedfraction$ $\begin{array}{r}r:meanlossperhour(astraightlineis\\ assumed),Wangetal.report94percentof\\ theinitialperformanceafter1000hof\\ heatingat80{}^{\circ }CforaZ-907cellwith\\ apolymergelelectrolyte,and1000hoflight\\ soakingat55{}^{\circ }Cand100mW/c{m}^{2}witha\\ UVfilterwithoutapercentageinthe\\ abstract;Nazeeruddinetal.requirea\\ sensitizertosustainatleast10^{8}redox\\ turnovers,andGratzel(2006)quotes{}^{\prime }one\\ hundredmillionturnover{s}^{\prime }andaservicelife\\ ofatleast20years;RSCSustainability\\ reportsabout95percentafter1000hofLED\\ soakingindoors(lightlevelnotgiven)\end{array}$ $\begin{array}{r}(computed:6and5percentin1000hare\\ 0.006and0.005percentperhour;20years\\ are175200h,so1000his0.0057ofthe\\ requirement;10^{8}turnoversin20yearsof\\ continuouslightare13699perday,oneper\\ 6.3s)\end{array}$
Indoor efficiency and the implied voltage	${\eta }_{ind}=\frac{V_{oc}{J}_{sc}FF}{P_{ind}}$	$V_{oc,impl}=\frac{{\eta }_{ind}{P}_{ind}}{J_{sc}FF}$	$\begin{array}{r}{P}_{ind}:opticalpowerdensityoftheindoor\\ light,about0.3mW/c{m}^{2}at1000lux(my\\ assumption,foracompactfluorescentlamp\\ withaluminousefficacyofradiationofabout\\ 330lm/W,backgroundknowledge;thepaper\\ givesnovalue),RSCSustainabilitygives\\ 30.48\pm 0.23percentwith{J}_{sc}=137.8\\ \pm 4.99\mu \mathrm{A}/c{m}^{2}andFF=82.9\pm \\ 1.5fortheXY1bcellundera1000lux\\ warm-whitecompactfluorescentlampanddoes\\ notgive{V}_{oc};twolargecells(1.92c{m}^2\\ each)inseriesgive1.39V,546.02\\ \mu \mathrm{A},64percentand485.9\\ \mu \mathrm{W}\end{array}$ $\begin{array}{r}(computed:with0.3mW/c{m}^{2}and137.8\\ \mu \mathrm{A}/c{m}^{2}theimpliedvoltageis\\ 0.800V;1.39\times 546.02\times 0.64=\\ 485.7\mu \mathrm{W},printed485.9\\ \mu \mathrm{W};overtwocellsof1.92c{m}^2\\ thisis126.5\mu \mathrm{W}/c{m}^2,42percent\\ oftheassumedlight,andthecurrentof\\ 546.02\mu \mathrm{A}is284\mu \mathrm{A}/c{m}^2\\ percell,2.06timesthe137.8ofthesingle\\ cell)\end{array}$

Important Questions

Ques: What is the dye-sensitized record: 11.9, 13.0, 14.3 or 15.2 %?

Green et al. (version 64) list 11.9 % for a Sharp cell of 1.005 $\mathrm{c}{\mathrm{m}}^2$ (AIST, September 2012) and 13.0 $\pm$ 0.4 % for an EPFL cell of 0.1155 $\mathrm{c}{\mathrm{m}}^2$ (Fraunhofer ISE, October 2020). Boschloo lists 14.3 % for Adeka-1 and LEG4 (2015, an ‘open device’ that he does not explain), and the Kakiage abstract says only ‘over 14 %’. Ren et al. report 15.2 % under simulated sunlight with 500 h of operational stability, but the summary I read gives no area and no testing laboratory. The black-dye row of Boschloo (2011) has the Sharp cell’s parameters, so the dates differ. I regard 15.2 % as the highest reported value and 13.0 % as the highest in the independent tables (my reading). Reference: Green, Dunlop, Yoshita, Kopidakis, Bothe, Siefer, Hinken, Rauer, Hohl-Ebinger and Hao, ‘Solar cell efficiency tables (Version 64)’, Prog. Photovolt. 32(7), 425-441 (2024); G. Boschloo, ‘Improving the Performance of Dye-Sensitized Solar Cells’, Frontiers in Chemistry 7, 77 (2019); K. Kakiage, Y. Aoyama, T. Yano, K. Oya, J. Fujisawa and M. Hanaya, Chem. Commun. 51, 15894-15897 (2015) (abstract); Y. Ren, D. Zhang, J. Suo, Y. Cao, F. T. Eickemeyer, N. Vlachopoulos, S. M. Zakeeruddin, A. Hagfeldt and M. Gratzel, ‘Hydroxamic acid preadsorption raises efficiency of cosensitized solar cells’, Nature (published online 2022; read through the ScienceDaily summary of 27 October 2022 and the CHIMIA 2023, 77(4), 254 abstract).

Ques: Is the highest voltage 0.744, 0.91, 1.05, 1.08 or 1.4 V?

Boschloo lists 0.744 V for the iodide black-dye cell, 0.91 to 1.01 V for cobalt cells, 1.05 V for the copper mediator and 1.07 V for a copper hole conductor, and gives 1.4 V as the highest voltage (Mg-doped TiO${}_2$, coumarin dye, bromide couple). His text says that the best solid-state voltage is about 0.8 V, but Cao et al. measured 1.08 V in 2017 for a solid-state cell with a copper hole conductor, so I read the table as right and the text as not updated (my reading). Yum et al. report 998 mV at 10.08 % and 1031 mV at 7.52 % for a thinner film. With a band edge of $-0.5$ V (assumed) the loss from the redox level is 0.106 V for iodide and 0.360 V for the copper couple (computed). The efficiency of the 1.4 V cell is not given. Reference: G. Boschloo, ‘Improving the Performance of Dye-Sensitized Solar Cells’, Frontiers in Chemistry 7, 77 (2019); Y. Cao et al., ‘11% efficiency solid-state dye-sensitized solar cells with copper(II/I) hole transport materials’, Nature Communications 8, 15390 (2017); J.-H. Yum et al., ‘A cobalt complex redox shuttle for dye-sensitized solar cells with high open-circuit potentials’, Nature Communications 3 (2012), DOI 10.1038/ncomms1655.

Ques: Is the efficiency limit 33.15, 33.7 or 33.8 %?

Ma et al. give 33.7 % at a gap of 1.37 eV and Boschloo 33.8 % with an optimum of 1.3 eV. My detailed-balance model gives 33.76 % with a rear mirror and 33.15 % with emission from both faces, both at 1.335 eV (computed). The first agrees with 33.7 and 33.8 %, so the sources probably differ by emission model and spectrum (my reading); I did not obtain a primary table of the limit, so the 33.15 % rests on my calculation. The optimum is flat: 1.30 eV gives 33.19 % and 1.37 eV 33.70 % (computed, mirror). The dyes are far from it: an onset of 1.65 eV (N719, 750 nm) is limited to 29.99 % and an onset of 1.9 eV to 25.05 %, which agrees with Boschloo’s 25 % and the current of 16.97 $mA/c{\mathrm{m}}^2$ with his 17 $mA/c{\mathrm{m}}^2$ (computed). Reference: W. Ma, F. Zhang and S. Meng, ‘Atomic Scale Investigation of Dye Sensitized Solar Cells: Interface Structure and Dynamics’ (arXiv 1406.3111; the journal is not named in the text); G. Boschloo, ‘Improving the Performance of Dye-Sensitized Solar Cells’, Frontiers in Chemistry 7, 77 (2019); Solar Cell Fundamentals page of this vault.

Ques: How thick should the TiO${}_2$ film be: 2, 10 or 18 micrometres?

Boschloo gives about 10 $\mu \mathrm{m}$ for liquid cells and about 2 $\mu \mathrm{m}$ for solid-state cells. Nazeeruddin et al. used 18 $\mu \mathrm{m}$ for the black dye, with a surface area of 75 ${\mathrm{m}}^2/g$ and a porosity of 0.62; that dye has a threshold at 920 nm, so it absorbs the red part weakly (my reading). Ito et al. optimised a transparent layer under a light-scattering topcoat to pass 10 %, but the abstract gives no thickness. The sources give no roughness factor, dye loading or extinction coefficient. If the absorbance is 1, 2 or 3, the film absorbs 90, 99 or 99.9 % at the peak (computed, assumed). The solid-state film must be thinner because the hole conductor limits the transport (background knowledge). So 10 to 18 $\mu \mathrm{m}$ is the liquid-cell level and 2 $\mu \mathrm{m}$ the solid-state level (my reading). Reference: G. Boschloo, ‘Improving the Performance of Dye-Sensitized Solar Cells’, Frontiers in Chemistry 7, 77 (2019); M. K. Nazeeruddin et al., ‘Engineering of Efficient Panchromatic Sensitizers for Nanocrystalline TiO2-Based Solar Cells’, J. Am. Chem. Soc. 123(8), 1613-1624 (2001); S. Ito et al., ‘Fabrication of thin film dye sensitized solar cells with solar to electric power conversion efficiency over 10%’, Thin Solid Films 516, 4613-4619 (2008) (abstract).

Ques: How fast is injection and how complete is collection: 33 fs, 160 fs or 1 ns?

Ma et al. report real-time calculations with injection times of 33 fs (SY404), 60 fs (D404), 87 fs (N1, LUMO+2) and 160 fs (N1, LUMO), defined as the time for 63.2 % of the electrons to transfer, and say that the range extends from sub-picoseconds to tens of ps and even ns. N1 is a simplified model dye. Boschloo gives a sub-nanosecond regeneration, a fluorescence lifetime of 260 ps for R6 on alumina, an electron lifetime of 1 to 10 ms, a transport time about ten times lower and an IPCE of about 90 %. The injection yield is 0.9994 to 0.9999 for 33 to 160 fs, 0.981 for 5 ps, and the collection is 0.90 for a ratio of 0.1 (computed). An IPCE of 90 % then needs four factors of 0.974 (computed). Injection is probably not the limit in good cells (my reading). Reference: W. Ma, F. Zhang and S. Meng, ‘Atomic Scale Investigation of Dye Sensitized Solar Cells: Interface Structure and Dynamics’ (arXiv 1406.3111; the journal is not named in the text); G. Boschloo, ‘Improving the Performance of Dye-Sensitized Solar Cells’, Frontiers in Chemistry 7, 77 (2019).

Ques: Which mediator is best: iodide, cobalt or copper?

Boschloo’s Table 1 gives 11.9 % for iodide (0.744 V, 22.5 $mA/c{\mathrm{m}}^2$), 12.3 to 13.0 % for Co(bpy)${}_3$, 14.3 % for Co(phen)${}_3$ (Adeka-1 and LEG4, 1.01 V) and 13.1 % for the copper mediator (1.05 V, 15.7 $mA/c{\mathrm{m}}^2$). Gratzel quoted 11 % in full sunlight for the iodide cells in 2006, which agrees with the Sharp cell. Yum et al. reached 998 mV and 10.08 % with Co(bpy-pz)${}_2$ but only 13.1 $mA/c{\mathrm{m}}^2$. The voltage loss from the redox level is 0.106 V for iodide and 0.360 V for the copper couple, but the copper cell has a lower current (computed, band edge assumed at $-0.5$ V). Iodide loses more than 0.5 V against a typical dye (Boschloo), and cobalt and copper couples allow a smaller driving force (background knowledge). The highest values come from cobalt and copper (my reading). Reference: G. Boschloo, ‘Improving the Performance of Dye-Sensitized Solar Cells’, Frontiers in Chemistry 7, 77 (2019); M. Gratzel, ‘Photovoltaic performance and long-term stability of dye-sensitized mesoscopic solar cells’, Comptes Rendus Chimie 9(5-6), 578-583 (2006); J.-H. Yum et al., ‘A cobalt complex redox shuttle for dye-sensitized solar cells with high open-circuit potentials’, Nature Communications 3 (2012), DOI 10.1038/ncomms1655.

Ques: What is the solid-state record: 7.81, 11.0, 11.7 or 13.8 %?

Boschloo’s text gives 11.7 % (Zhang et al., 2018), but his Table 1 lists 13.8 % for the WS72 dye with a copper hole conductor, 1.07 V, 11.7 $mA/c{\mathrm{m}}^2$ and a fill factor of 0.79. That row gives 9.89 % as printed, and 11.67 % if the efficiency and current entries are interchanged, which is the 11.7 % of the text (computed). I regard the interchange as more likely, so the current is 13.8 $mA/c{\mathrm{m}}^2$ (my reading). Cao et al. (2017) report 11.0 % for 1.08 V, 13.87 $mA/c{\mathrm{m}}^2$ and 0.733, and the product is 10.98 % (computed). The 7.81 % row of spiro-MeOTAD with the S5 dye checks out at 7.82 %. A perovskite sensitiser on mesoporous TiO${}_2$ reached about 15 % in 2013 (Burschka et al.), but the Perovskite Solar Cell page treats such devices as perovskite cells. Reference: G. Boschloo, ‘Improving the Performance of Dye-Sensitized Solar Cells’, Frontiers in Chemistry 7, 77 (2019); Y. Cao et al., ‘11% efficiency solid-state dye-sensitized solar cells with copper(II/I) hole transport materials’, Nature Communications 8, 15390 (2017); J. Burschka, N. Pellet, S.-J. Moon, R. Humphry-Baker, P. Gao, M. K. Nazeeruddin and M. Gratzel, ‘Sequential deposition as a route to high-performance perovskite-sensitized solar cells’, Nature 499, 316-319 (2013) (abstract).

Ques: How stable is a dye-sensitized cell: 94 or 95 % after 1000 hours, and how many dye turnovers are needed?

Wang et al. report that a Z-907 cell with a polymer gel electrolyte kept 94 % of its performance after 1000 h of heating at 80 ${}^{\circ }\mathrm{C}$, and describe 1000 h of light soaking at 55 ${}^{\circ }\mathrm{C}$ and 100 $mW/c{\mathrm{m}}^2$ with a UV filter without giving a percentage. RSC Sustainability reports about 95 % after 1000 h of LED soaking and gives no light level. Nazeeruddin et al. require at least $10^8$ turnovers of the sensitizer, and Gratzel (2006) quotes ‘one hundred million’ and a service life of at least 20 years. The mean losses are 0.006 and 0.005 % per hour, 1000 h is 0.0057 of 20 years, and $10^8$ turnovers in 20 years of continuous light are one per 6.3 s (computed). The tests differ in temperature, light and electrolyte, so 94 to 95 % is the typical level, short against 20 years (my reading). Reference: P. Wang, S. M. Zakeeruddin, J. E. Moser, M. K. Nazeeruddin, T. Sekiguchi and M. Gratzel, ‘A stable quasi-solid-state dye-sensitized solar cell with an amphiphilic ruthenium sensitizer and polymer gel electrolyte’, Nature Materials 2, 402-407 (2003); George, Pradhan, Unni and Soman, ‘Engineered hole-free, spacer-free dye-sensitized light harvesters for indoor photovoltaic and self-powered applications’, RSC Sustainability 2, 2839-2843 (2024); M. K. Nazeeruddin et al., ‘Engineering of Efficient Panchromatic Sensitizers for Nanocrystalline TiO2-Based Solar Cells’, J. Am. Chem. Soc. 123(8), 1613-1624 (2001); M. Gratzel, ‘Photovoltaic performance and long-term stability of dye-sensitized mesoscopic solar cells’, Comptes Rendus Chimie 9(5-6), 578-583 (2006).

Ques: Is the best indoor efficiency 28, 30 or 32 %?

Boschloo gives 32 % at 1000 lux and cites Cao et al. (2018), which I did not read. RSC Sustainability reports 30.48 $\pm$ 0.23 % for XY1b and 31.33 $\pm$ 0.62 % for XY1b with D35 under a 1000 lux warm-white compact fluorescent lamp, and 29.35 $\pm$ 0.09 % for cells of 1.92 $\mathrm{c}{\mathrm{m}}^2$. Ren et al. report 28.4 to 30.2 % for cells of 2.8 $\mathrm{c}{\mathrm{m}}^2$ over a range of ambient light intensities, without values. No source gives the optical power density, so the lamps are not on one scale. With an assumed 0.3 $mW/c{\mathrm{m}}^2$ the cell data (137.8 $\mu A/c{\mathrm{m}}^2$, fill factor 82.9) imply 0.800 V, a plausible voltage (computed). I regard 30 to 32 % as the level of small cells and 28 to 29 % as the level of large cells (my reading), and the 32 % as the highest reported but unchecked here. Reference: G. Boschloo, ‘Improving the Performance of Dye-Sensitized Solar Cells’, Frontiers in Chemistry 7, 77 (2019); George, Pradhan, Unni and Soman, ‘Engineered hole-free, spacer-free dye-sensitized light harvesters for indoor photovoltaic and self-powered applications’, RSC Sustainability 2, 2839-2843 (2024); Y. Ren, D. Zhang, J. Suo, Y. Cao, F. T. Eickemeyer, N. Vlachopoulos, S. M. Zakeeruddin, A. Hagfeldt and M. Gratzel, ‘Hydroxamic acid preadsorption raises efficiency of cosensitized solar cells’, Nature (published online 2022; read through the ScienceDaily summary of 27 October 2022 and the CHIMIA 2023, 77(4), 254 abstract).

Ques: Do the module values of the indoor paper agree with its cell values?

The paper prints a module of two cells in series with 1.39 V, 546.02 $\mu \mathrm{A}$, a fill factor of 64 % and 485.9 $\mu \mathrm{W}$, and $1.39\times 546.02\times 0.64$ is 485.7 $\mu \mathrm{W}$ (computed). Each cell has 1.92 $\mathrm{c}{\mathrm{m}}^2$, so the current is 284 $\mu A/c{\mathrm{m}}^2$ per cell, 2.06 times the 137.8 of the small cell, and the power is 126.5 $\mu W/c{\mathrm{m}}^2$ over 3.84 $\mathrm{c}{\mathrm{m}}^2$, 42 % of the assumed light (computed). The module area is not stated. Either the module cells are larger than 1.92 $\mathrm{c}{\mathrm{m}}^2$ or the light at the module differed from the cell test; I could not resolve it, so the module numbers should not be used to rank cells (my reading). The units are microamps and microwatts as printed in the paper. Reference: George, Pradhan, Unni and Soman, ‘Engineered hole-free, spacer-free dye-sensitized light harvesters for indoor photovoltaic and self-powered applications’, RSC Sustainability 2, 2839-2843 (2024).

Ques: Is the copper ligand called tmp, tmbpy or tmby, and is the cobalt complex Co(bpy-pz)${}_3$ or Co(bpy-pz)${}_2$?

Boschloo writes Cu(tmp) in Table 1 and Cu(tmbpy)${}_2$ in the text and defines neither. Cao et al. (2017) define tmby as 4,4’,6,6’-tetramethyl-2,2’-bipyridine and use [Cu(tmby)${}_2$] as the hole conductor with Y123, and Saygili et al. define the same ligand and report a Cu(tmby) mediator with Y123 at 10.3 % and 1.04 V. I regard tmp and tmbpy as the same ligand as tmby (my reading), and this page uses tmby. Yum et al. write the cobalt complex as [Co(bpy-pz)${}_2$]${}^{3+/2+}$, with two tridentate ligands, 0.86 V against NHE and 10.08 % with 998 mV, while Boschloo’s table writes the couple without a subscript. This page uses Co(bpy-pz)${}_2$. Reference: G. Boschloo, ‘Improving the Performance of Dye-Sensitized Solar Cells’, Frontiers in Chemistry 7, 77 (2019); Y. Cao et al., ‘11% efficiency solid-state dye-sensitized solar cells with copper(II/I) hole transport materials’, Nature Communications 8, 15390 (2017); Y. Saygili et al., ‘Copper bipyridyl redox mediators for dye-sensitized solar cells with high photovoltage’, J. Am. Chem. Soc. (2016; volume and pages not read); J.-H. Yum et al., ‘A cobalt complex redox shuttle for dye-sensitized solar cells with high open-circuit potentials’, Nature Communications 3 (2012), DOI 10.1038/ncomms1655.

Related pages in this vault: Perovskite Solar Cell, Quantum-Dot Solar Cell, Perovskite-Silicon Tandem Solar Cell, Organic Solar Cell (Bulk Heterojunction), Solar Cell Fundamentals, Solar Cells and Modules (Crystalline Silicon and Thin Film).