Dye sensitized solar cells

1. Physicalia (2008)3 pp.149-.j60
l,4as.30
CHARACTERIZATION TiO2TOWARDSHIGHER
OF
INCIDENTPHOTON-TO-CURRENT
EFFICIENCY DYE
IN
SENSITIZED
SOLARCELLS*
W. Moonsl,K. Vandewall,P. Robaeysl,G. Krishnal and J. Mancal'2
1lnstituutvoor
Materiaalonderzoek,
universiteitHasselt,
Wete nschapspark1, 3590Diepen eek.
b
z IMEC-IMOMEC
vzw, Wetenschapspark 3Sg0Diepenbeek
1,
Summaryof the work, awardedwith the prize ot the best poster presentation Wouter Moons at
by
the General
Scienfific
Meetingofthe BPS at the Universit6 Librede Bruxelles ly'av 200g
on 21,
Abstract
Worldenergyprovision probably
is one of the most challenging engagements the 2.1"t
of
century.
Amongihe novelgeneration solarcell concepts, so called"GrAEel ce " or
the
Dye Sensitized
SolarCell is likelyone of the most promising. is basedon the spectral
lt
sensitization a thin ceramic mesoporousmembraneby suitabletransitionmetal
of
complexes (=dye).
The film consists nanometer-sized
of colloidal
titaniumdioxideparticles
sinteredtogether allow for chargecarrierlransport. focusing the fundamenial,
to By on
thoughmostlyignored,
processes
like sintering
the anataseTiO, porouslayer and dye
soaking, incfeasein IPCE is observed.
an Duringthe brief heatingof the mdso-porous
film,calcination the TiO, particles
of takes place.This calcination monitored
is in-situby
electrical
measurements. tuningthe dye coverage porousTiO2,light absorption
By of is
enhanced,and suTfacetraps are decreased.In this work the adsorptionkjneticsof
ruthenium-based havebeenstudiedby measuring optical
dyes the absofbance spectraof
the sensitized
filmsusingUV-Vis
spectroscopy. resulting
The effective
aDsoroance
curyes
as functionof time were successfully
fltted as pseudosecond-order chemicalsorption
processes.Both presentedcharacterization techniquesresultedto be efficienttools
towards betterunderstanding further
a and improvements dye sensitized
of solarcells
149

2. lntroduction
effect by Becquerel[1]' researchers and
Ever since the discoveryof the photoelectric
sunlightcouldbe
engineers have been enthralled the idea that freelyavailable
with
captured converted electrical
and into power'
'thatgenerates photovoltage
a by
The mostcommonly known solarcells,i'€' a device
The semiconductor
shining lightuponit, incorporate pn-junction a semiconductor'
a in
solarspectrum' Dependent the
on
material to be ableto absorb largepartof the
has a
moreor lessclose
absorption properties the material lightis absorbed a region
ol the in
pairs are generated and if
to the surfiace. When photonsare absorbed'electronhole
their recombination prevented
is they can reachthe junctionwherethey are separated
absorbing semiconductors silicon
like mostcaniers
by an electric field.Evenfor weakly
are generated nearthe surface'The pn-junction whichseparates emitterand base
the
a high collection probability free
for
layer is very closeto the surfacein orderto have
carriers.. operating
The principtes beendescribed manypublications [3]'
have in [2]
Sofar,thescienceofso|arce||shasbeendominatedbydevicesinwhichthejuncti
betweeninorganicsolid-statemateria|s,usua|lydopedformsofcrySta||ineoram
silicon.lnrecentyears,novelclassesofphotovoltaicmaterialsystemshaveimer
solarcellscuFently become broader
of interest'
suchas, nano-crystalline organic
and
roughly molecular polymdr
into and organicsolarcellsor into
Thesecellscanbe divided
flallayer systems bulkheterojunctionsl4]'
and
cheapto fabricate(theexpensive
solarcells,whichare relatively
Thesethird generation
high-temperature high-vacuum
and processes neededfor the
and energy-intensive
on flexiblesubstrates'and can be
traditionaldevicescan be avoided),can be used
shapedortintedtosuitdomesticdevicesorarchitecturalordecorativeapp|icationsl5
developed molecular
a system solarlight
photovoltraic for
ln 1991GreEeland O'regan
of
basedon the speclralsensitization
harvesting Gonversion electricitylG] lt is
and to [4'
metalcomplexes' Jilmconsists
transition The of
a thin ceramicmembrane suitable
Dy
nanometer-sizedcol|oida|titaniumdioxideparticlessinteredtogethertoa||owforch
chromophore' thesemembranes give
Whenderivatized a suitable
transport'
carrier with
l
150

3. extraordinaryefficienciesfor the conversionof incident photons into electric current
(IPCE).These cells are referred as "Gratzel
to cells"or .Dye Sensitized Solar Cells"
(DSSC).
operation the Dye-Sensitized
Figure 1: Schematic of SolarCell
presentation the operation
A schematic of principle givenin Figure The heartof the
is 1.
oxidefilm (see
deviceis a mesoporous Figure typically
2), 10pmthick,whichis made
of tiny nano-crystals,
interconnected allow electronic
to conduction take place.Oxides
to
suchas TiOzZnO,SnOz suchas CdSe,are the preferred
and Nb2O5or chalcogenides
,
compounds[8]. till now, titanium
Up dioxidehas been the material choicefor these
of
applications. becauseof its large band gap (3_3.2eV), TiO, intrinsicly
Unfortunately,
part of the solar emissionand so has low conversion
absorbsonly the ultraviolet
efficiencies. solutionlies in the separation the opticalabsorption
The of and charge-
151

4. generating using an electron
functions, absorbing the visible
transfersensitizer' in
gap' Photo-
region injectchargecaniersintoa substrate a wideband
spectral to with
excitalionofthesensitizer(S)resu|tsintheinjectionofane|ectroninthecond
Theoriginalstateofthe dyeis subbequently restored eleclron
by donation
bandof TiOz.
from the eleclrolyte, impregnated the porous
in '-TiO2'usuallyan organicsolvent
containing redoxsystem, coupleis used'Thusresults a
suchas the iodide/triiodide in
three dimensional networkwith an enormouscontactarea betweenthe two types of
materials, wherecharge tansportis bestdescribed a random
by walkmodel[9]' The
intercepts recapture the conduction
the of band
regeneration the sensitizer iodide
of by
electron the oxidized
by dye.A dense TiOzlayer(seeFigure is sputtered a hole-
3) as
in
blocking layerto reducerecombination the electrode' iodideis regenerated
at The
typically withPlatinum'
coated
tum by the reduction triiodide the counter-electrode'
of at
migration through external
the load' The
The circuitis beingcompleted electron
via
generated underillumination conesponds the difference
to between Fermi
the
voltage
overallthe
levelof the electron the solidand the redoxpotential the electrolyte'
in of
devicegenerates electricpowerfrom lightwithoutsuffering permanent
any chemical
transformation[10]'
efficiency this Dyesensitized
overall of AM1'5(Air
solarcell is 7% under
The reported
mass1.5)f6l [7t.
TheAMl.Sconversionpowerefficiency1lAMl.sofaphotovoltaicdeviceisgivenb
n
no*'"=P-ffm=FFv**
where
-- v.i,""
-
,-l'=-
v*J*
wherePo,uristheoutpute|ectricalpowerofthedeviceunderil|umination,Prnist
incident the deviceas measured a calibrated
on by reference cell' %" is the
intensity
cunent denslty;m is the spectral
ooen circuitvoftale, and Js" is the short-circuit
mismatch factorthat accounts deviations the spectraloutputof the solarsimulator
for in
152

5. with respectto the standardAM1.5 spectrumand deviationsin the spectralresponseof
the deviceunder measurewith respectto that of the referencecell;FF is the fill factor.
novel deviceconceptshave been
Great effort has been made to improvethe efficiency:
developed,new materialshas been tested, different layer techniqueswhere applied.
Theseeffortsresulted an efficiency approximately [8] [11][12] lt appears
in of 11% that
for one decade now the efficiency of nano-structuredcells has been essentially
slagnating researchprogress.['13]
despitesignificant
though
We investigatedwith an alternativeapproach,the effects of the fundamental,
mostlyignored,processeslike sinteringand soakingthe anataseTiO2porouslayer, on
the IPCE.
printed
Screen
Figure lresoporous
2: TiO, Figure3: Dense Ti02
sputtered
Experimental
For the productionof substrates,we distinguishtwo types: the first substrates are
regular microscopeglasses,the second type are patternedFluorinedoped Tin Oxide
ConductiveOxide (TCO),which will act as a
(FTO) glasses.The latter is a Transparent
153

6. cathode,respectively anode in the DSSC.'TWopattemswere etched uslng
an
Iithography, with inter{igital structures otherwlth planesufaces' All substrate-s
one the
-de-ionizeA
were lhoroughly cleaned in soap water, rinsed Mth water, ultrasonically
rinsedIn Acetoneand boiledin lsopropanol, this to ensurea fat anddustfree substrate
for good adhesion.The FTO substtateswere sputtered in a home made sputter
installation establish denseTio2-antase
to a layer(seeFigure of approximately00nm
3) 1
thick and will act as a holeblocklnglayer'All subsfates werethen screenprinted a
with
@mmerdalTiO2paste,purchased from Solaronix. The pasteis dried and sinteredin a
fumaceto ensurepercolation the TiO2nanoparticles'
of
0.2nE Fusg+lbl€A
a
a
B
tr|drdrtrd
Ru-dyes Rr535 (C)Ru-s3s.bis
Figure 4: (A)Absorbance (B) TBA
The obtainedsampleswere then dividedin seven groupssach group containing ten
samoles. w€re immersed an
The substrates in gthanolsoluiioncontalning commercial
a
Ruthenium bas€ddye (seeFigure4). Foralt groups,concentration temperature the
or of
solution changed.
was
pure
Afrer time periods,variatingfrom th to 196h, all substrateswere rinsed wllh
ethanolto withdrawthe excEssof dye.The absorbance was measured usinga W-Vis
meter.
The cathodelayers(FTO)were sputtered with a thin (3onm)Platinum layer,whichwill
layerfor elecirondonation
act as a catalyzing to the elecfolyte' The anodeandcathode

7. are assembled with a hot meltspacer.The complete
together cellswere vacuumfilled
with an electrolyte,
consistedof 50m[.4tri-iodidein acetonitrile,
sealed and ready to be
tested.
Results
and Discussion
Annealing
Fluorinedoped Tin Oxide (FTO) substrates
were patterned that 25pm interdigital
so
structureswere obtained.These substraleswere screen-printedwith a TiO2paste and
annealed
whilein-situresistance wereconducted.
measurements
profiles,see Figure 5 and can be
This resultedin time vs. temperature/current
interpretedin such way that the ideal temperatureprofile can be produced to get
sufficient percolationwithout destroying the meso-porousnetwork. The ideal
profile incorporate different
temperature will the usedin the Ti02 pasteand the
materials
thesechemical
timeto burn-out binders
without
creating
cracksin the layer.Cracksdue
to a rapidheatingwill clearlybe shownin the measurement a suddendecrease
as in
conduction.
when an increase conduction followed
Percolation the nanocrystals observed
of is in is
by a slow decreasewherethe crystalsgrow.
best results are obtainedwhen the surface area availablefor dye
Theoretically,
chemisorption maximum
is while still havinga continues
network allowfor electron
to
conduction.
While plotting currentin function temperature,
the of plols can be made-By
Arrhenius
fittingthe activation modelcan be derived. this caseconduction
energya conduction ln
is governed thermal
by hopping
activated [14][15][16].
155

8. * T€mperd$.re pllltb
o Offial'r€flt
g
, g
E
I
6
g
for annealing TiO2
Figure5: Timevs.Temperature/Current of
?
l {r {av}
10Vand 100V)
Figura6: Arrhenius for annealing TiO2at several.voltage.s-(1v' line ths
plot of
and the thin solid
where the thick solid r,n" t"pt"""ni" the measurement
fit.
Arrhenius
Dye soaking
pholons
should be
Lightabsorption a key factorof solarcellssinceabsorbed
is
converted electrons. Sensitized
into Dye SolarCells(DSSC) soaked a dye
are in
Ru-based andsoabsorption betuned
dyes, can bythechosen dye'
typically
solution,
156

9. Adsorption
kinetics TiOzwere undertaken, determination dye concentration
for the of
was done with a UV-Visspectrometer measuring
by absorbance 1,a"(535nm).
at The
time profileof dye adsorption a singleand continues
is curve leadingto saturation,
suggesting the possiblemonolayercoverageof dye on the surfaceof TiOz. The
adsorptionexperiments
were conductedat varioustemperatures
C18oC,20oC and
37.50C)and at various concentrations
(0.8m9/ml,
0.2mg/mtand 0.02mg/ml).
The
amount of adsorptionvariationis shown. lt is seen that temperature significanfly
influencesthe adsorptioncapacity.High temperature results in high adsorption,
suggesting endothermic
an property.
The adsorption function timecan be fittedwitha pseudo
in of second-order
rate:
1 't 1.
i=ne q.'
qr (andqe)is the amount dye adsorbed TiO2at timet (andat equilibrium).is the
of on k
pseudosecond-order constant dye adsorption, is expressed a function
rate of it as of
temperature the Arrhenius-
by typerelationship
[17].
lnK=lnA-3
RT
IPCE
Incident
Photon current
to (IPCE)is a technique measure conversion
Efficiency to the of
incidentphotonsand the percentage those photonsthat effectively
of generated
an
electron. halogenlamp in combination
A with a monochromator used to produce
is
monochromatic
light.This light is passedthrougha beam splitter that the incident
so
lighton the samplesolarcellis known.
IPCEis determined [18]:
by
4pce = 1Lue Dcr ' lcce
qrne:Lightharvesting (amount absorbed
efficiency of photons)
qcrE:
Chargeinjection
efficiency
rtccE:
Chargecollection
efficiency
t57

10. + 37.5oC0.2m9/mlRu535
+ o.2mglml20qc Ftltes + Zo"Co.2mg/nrlRu535
+ 0.02mg/m120rc Ru535 ---- laoo 0.zmg/rnlnu535
+ 20'C R!535
o.8mgy'ml
-18'C0 2mg/mlRu535
2o6Co 2s€/ml Fu535
37.5"CO-2mSln Ru535
0.8 mg/ml20"C Ru535
c2 mg/ml2CfC8u535
o.o2 mg,'ml2o"c Ru535
3ff
S?lil?;^
*;ff ;:?:*"#:nU'""tJlHH*1.l"':':i's"JT.':i3
;,,ff
Belerence
Gall
Test Cdl
Monoafi€mator
(lPcE)testsetup
Photon currentEfficiency
Figure8r Incident
to
be
absorbtion'there should
on the rLHEby saturating dye
the
Since we focusedmainly "'"'nu showsthat
other parameter Frg::..n
without"n"ng;ng ;nv
an increasein lPcE will increase'
dye the efficiency DSSC's
of
by changing the amountof adsorDed
158

11. |lJ
Wav€lengdx{r:m}
Figure g: lpCE for DSSCas functionof soakinqtime
Conclusions
An effort has been made to increasethe Incidentphoton to
Current Efficiencyof Dye
Sensitized Solar Cells by focusing on the sinter process and
on the soaking time
process Positiveresurtswere obtainedby tuningthe sinteringprofirein such way that
a
conductive network of TiO2 was obtained without destroying
the pore size of the
mesoporousnetwork. These pores promote the active surface where
dye molecules
couldanchorand providea directchargeinjectionin the TiO2layer_
substrates were syslematicallyimmersed in solutionswith
differentconcentrationat
variabre temperaturesresurting a moderto describethe dye uptake.This resurted
in in a
positivechange of lpCE by a factor of 3. The proposed
approachthereforeresultedto
be efficienttowards a better understanding and improvementof dye sensitizedsolar
ce s-
159

12. Acknowledgements
Theauthor
likesto acknowledge universiteit
the Hasselt xios Hogeschool financial
/ for
support'Furthermore poryspec to the FWo
thanksis due to the rwr-sBo project an
(Phd.Koen Vandewal)
Refurences
['l] E.Becquerel, Acad.
C.R. i4S -
Sci.9, (1839). .
[2] A. Goetzberger, Knobtoch, B.Voss,Ctystaltine
J. and Siticon
So/arCels (Wtey,199g).
[3] M.A.Green,SiticonSolarCe s: Advancedprinciples practice
& (Bridgeprintery,
199S).
[4] H. Hoppe, N.S.Sariciftci,
and Joumat materials
of research1S,1S24 (2004).
[5] A. coetzberger, C. Hebting,
and SotarEnergyMaterials Sotar
and Celts I (2000).
62,
[6] B. O'Regan, M. Gr.itzet,
and Nature 353,737(1991).
f/l M. Gratzel,CoordinationChemistry
Reviews1tl, j67 (j991).
[8] J-M.Kroon, Bakker, H.J.p.Smit,progress photovoltaics (2OO7l.
N.J. and in 15,.1
[9] J. Nelson,
Physical Review phys.Rev.B pRB59, 15374
B (j999).
[10] M. cratzel,Journal photochemistry photobiotogyA:
of and Chemistry 164,3(2004).
[1ll K.D. Benkstein, Kopidakis, van de Lagemaat, A.J. Frank,Joumatof phvsical
N. J. and
Chemistry 107,7759(2003).
B
[12J B. C.J,andA. Journat
F, oftheAmerican
Ceramjc Society80,3157 (1997).
[13] H. Tributsch,
Coordination
Chemistry
Reviews
248, 1511(2004.
[14] H. Tang,K. Prasad, R. Sanjines,
and Joumal Applied
of physics 2042
75, (1gg4r.
[15] R. O'Hayre, Nanu, Schoonman, A. Goossens,
M. J. and Journal physical
of Chemistry
C
111,4809 (20071.
[16] J. vande Lagemaat, park,andA.J.Frank,
N.G. Journat physical
of Chemistry 1O4,2044
B
(2000).
[17] S.Wang, H. Li,Journat Hazardous
and of Materiats.t26,71
(2OOS.
[18] J. Halme,G. Boschtoo, Hagfetdt, p. Lund,Journat physical
A_ and of Chemistry 1.12,
C
5623(2008).
160