Keller_et_al-2016-Progress_in_Photovoltaics-_Research_and_Applications
SHORT COMMUNICATION
Potential gain in photocurrent generation for Cu(In,Ga)Se 2solar cells by using In 2O 3as a transparent conductive oxide layer
Jan Keller 1*,Johan Lindahl 1,Marika Edoff 1,Lars Stolt 1,2and Tobias T?rndahl 1
1Solid State Electronics,The ?ngstr?m Laboratory,Uppsala University,SE-75121Uppsala,Sweden 2
Solibro Research AB,Vallv?gen 5,SE-75151Uppsala,Sweden
ABSTRACT
This study highlights the potential of atomic layer deposited In 2O 3as a highly transparent and conductive oxide (TCO)layer in Cu(In,Ga)Se 2(CIGSe)solar cells.It is shown that the ef ?ciency of solar cells which use Zn-Sn-O (ZTO)as an al-ternative buffer layer can be increased by employing In 2O 3as a TCO because of a reduction of the parasitic absorption in the window layer structure,resulting in 1.7mA/cm 2gain in short circuit current density (J sc ).In contrast,a degradation of device properties is observed if the In 2O 3TCO is combined with the conventional CdS buffer layer.The estimated im-provement for large-scale modules is discussed.Copyright ?2015John Wiley &Sons,Ltd.
KEYWORDS
In 2O 3;ALD;Zn-Sn-O;alternative window layer;CIGSe;TCO
*Correspondence
Jan Keller,Solid State Electronics,The ?ngstr?m Laboratory,Uppsala University,SE-75121Uppsala,Sweden.E-mail:jan.keller@angstrom.uu.se
Received 13March 2015;Revised 15June 2015;Accepted 15June 2015
1.INTRODUCTION
The recently achieved advances in Cu(In,Ga)Se 2-based so-lar cell technology are mainly related to optimized window and buffer layer properties combined with interface-near modi ?cations at the heterojunction.A prominent example is the implementation of a post CIGSe-deposition potas-sium treatment (KF-PDT),which seems to promote Cu depletion at the absorber surface [1,2].Applying the KF-PDT,different research groups reported on an increase in open circuit voltage (V oc ),?ll factor (FF ),and/or J sc [2–7].Because the bene ?cial effect on the solar cell parameters is supposed to be caused by interdiffusion of Cd and Cu atoms at the interface [2],the bene ?t of a KF-PDT for solar cells using alternative buffer layers (e.g.,Zn-Sn-O,Zn-Mg-O,Zn-O-S or In-S compounds)is unlikely,or at least not proven yet.To sustain the competitiveness of non-toxic (Cadmium-free)buffer layers,different approaches might be preferred.
In state-of-the-art CIGSe solar cells using alternative buffer layers,the increased J sc ,induced by a reduced para-sitic light absorption,is usually offset by a drop in V oc and/or FF ,leading to,at best,only minor ef ?ciency im-provements [8,9].In order to fully utilize the potential of
high band gap buffer layers with E g ,buffer ≥3.3eV,there should be no absorption in the overlying transparent and conductive oxide (TCO)for E ≤E g ,buffer ,which is not en-tirely true for the commonly applied i-ZnO/ZnO:Al (AZO)bilayers [10].A suitable candidate ful ?lling this de-mand is the direct band gap semiconductor In 2O 3,which was suggested for utilization in photovoltaic devices al-ready in the early 1980s [11].
Despite the relatively small fundamental band gap en-ergy of E g ≈2.8eV,the main absorption onset of this ma-terial is increased to about 3.7eV because optical transitions from the energetically highest valence band states into the conduction band are forbidden [12–14].Fur-thermore,it was shown recently that highly conductive and transparent In 2O 3?lms can be deposited by atomic layer deposited (ALD)[15,16]even with relatively high growth rates (>1?/cycle)and in a temperature range that is appro-priate for deposition onto a CIGSe surface (i.e.,below 180°C)[17].
Because ALD often requires extended deposition times,this technique might not be applicable to industrial produc-tion,but because of its self-limiting character,the material properties are well controllable,which enables a funda-mental study of In 2O 3as a TCO.
PROGRESS IN PHOTOVOLTAICS:RESEARCH AND APPLICATIONS
Prog.Photovolt:Res.Appl.2016;24:102–107
Published online 17July 2015in Wiley Online Library (https://www.wendangku.net/doc/e813719029.html,).DOI:10.1002/pip.2655
Copyright ?2015John Wiley &Sons,Ltd.
102
While tin-doped indium oxide has been frequently
applied in thin?lm solar cells[18,19],there are(to the
authors knowledge)no published studies that evaluate
non-extrinsically doped In2O3as a TCO in CIGSe solar
cells.In this work,we evaluate an all-ALD window layer
structure consisting of the wide band gap buffer layer
ZTO(E g≈3.4eV)[20–22]and a highly transparent In2O3 TCO layer and compare the photovoltaic performance with
solar cells in which conventional AZO is used as a TCO.
2.EXPERIMENTAL
The solar cells investigated in this study(total area
A=0.5cm2)were processed as a stack of SLG/Mo/CIGSe/
buffer/TCO/metal grid,in which the buffer layer was
interchanged between ALD ZTO and chemical bath depos-
ited CdS and the TCO between a sputtered i-ZnO/AZO
bilayer and ALD In2O3.The CIGSe?lms were deposited
by an inline co-evaporation process at Solibro Research
AB,exhibiting an overall composition of[Ga]/([Ga]+[In])
=0.43and[Cu]/([Ga]+[In])=0.83determined by X-ray ?uorescence spectrometry(Spectro X-Lab2000).The elec-trical and optical properties of the TCO layers were evaluated for?lms grown on soda-lime glass(SLG)substrates only.
The ZTO and In2O3layers were grown in a
Microchemistry F-120ALD reactor at the same deposition
temperature of135°C by using N2as a carrier gas.
ZTO buffer layers with a composition of Zn0.8Sn0.2O y
were deposited from diethyl zinc[Zn(C2H5)2],
tetrakisdimethylaminotin(IV)[Sn(N(CH3)2)4]and deion-
ized water by using a Sn/(Sn+Zn)pulse ratio of0.5,where
the length of the Sn/Zn precursor:N2:H2O:N2pulses were
400/400:800:400:800ms,respectively[22].Similarly,the
In2O3TCO was grown from cyclopentadienylindium(I)
[In(C5H5)]and a mixture of deionized water and oxygen
gas according to[17].In this study,the In(C5H5)precursor
was sublimated at50°C and transported into the reactor by
bubbling with N2.The water and oxygen gasses used sep-
arate gas lines,but were supplied in a combined H2O+O2
pulse that mixed the precursors close to the substrate inside
the reactor.The corresponding pulse lengths for the In2O3
layer were1200:800:400:1600ms for the In precursor:
N2:H2O+O2:N2pulses,respectively.Although not com-
mon practice for ALD,the H2O+O2precursor mixing
was carried out to decrease the length of the ALD cycle
by two pulses,and because improved electrical properties
have been obtained by using the gas mixing pulse
scheme[17].Still,self-limiting growth could be obtained
for the process,and the?lms were found to be homoge-
neous over the5×5cm2SLG substrate size.Microstruc-
tural investigations by X-ray diffraction and transmission
electron microscopy proved crystalline growth of In2O3.
A?lm density ofρ=6.7g/cm3was determined by X-
ray re?ectivity that is lower than the tabulated value for
bulk In2O3(7.1g/cm3).For low temperature water-based
ALD processes,slightly lower density values are not un-
common because of incorporation of hydrogen in the ?lms,which might lead to small deviations from2:3 In2O3stoichiometry.
The absorption data of the TCO?lms on glass were de-
rived by R-measurement and T-measurement using a
Perkin Elmer Lambda900spectrometer with an integrating
sphere.Film thickness was measured by a Dektak150Sty-
lus Pro?ler(Veeco,Plainview).The electrical properties of
the TCO layers on glass were measured in a Hall set-up,
applying a magnetic?eld of B=0.5T.
External quantum ef?ciency(EQE)measurements were
conducted for each sample variation in a home-built setup.
The J sc values,calculated from the EQE and AM1.5G stan-
dard spectrum,have been used to calibrate the light inten-
sity for measuring illuminated current–voltage(IV)
characteristics with a tungsten halogen lamp as a light
source at a cell temperature of T=25°C.To minimize the
in?uence of differences in re?ection,a MgF2anti-
re?ection coating(ARC)was evaporated on top of the de-
vices prior to the IV measurements.
3.RESULTS AND DISCUSSION
The electronic properties of the TCO?lms on glass as they
are used in the solar cell devices are summarized in Table I.
The respective?lm thickness was tuned to an optimal sheet
resistance for the present cell geometry,of R sh=30Ω/sq.
While the mobility value for the In2O3?lm(thickness d=205nm)is about3.5times larger compared with the
AZO?lm(d=225nm),the free electron density(presum-
ably originating from oxygen vacancies)is heavily re-
duced,leading in total to a slightly lower resistivity for
In2O3.X-ray diffraction and transmission electron micros-
copy investigations(not shown here)revealed an almost
non-textured crystalline growth and a comparatively large
lateral grain size(~60nm)for the In2O3?lm,which most likely contributes to its high mobility.
Figure1shows the absorption spectra of the same?lms.
It is obvious that,despite its rather small fundamental band
gap energy,the In2O3?lm shows a blue-shifted absorption onset compared with the i-ZnO/AZO https://www.wendangku.net/doc/e813719029.html,paring the difference in absorption(gray dashed line)with the band gap energy of the buffer layer,it is evident that the photon ?ux for E ph≤E g,buffer reaching the buffer layer can be sig-ni?cantly enhanced by exchanging the i-ZnO/AZO struc-ture for the In2O3TCO.Furthermore,the reduced charge carrier density leads to an increasing transmission for λ>700nm,which should improve J sc independent on the choice of buffer layer.In the case of the CIGSe layers cho-sen in this study,the free charge carrier absorption(FCA) can be reduced by up to8%close to the band edge region of the absorber(λ≈1100nm).
In the next step,CIGSe solar cells were processed with
the mentioned variations in buffer/TCO structures,leading
to four different combinations.It turned out that the ALD
In2O3did not nucleate on CdS layers(as well as not on
pure CIGS)in this study,leaving three different window
103
Prog.Photovolt:Res.Appl.2016;24:102–107?2015John Wiley&Sons,Ltd. DOI:10.1002/pip
constellations (ZTO/In 2O 3;ZTO/i-ZnO/AZO;CdS/i-ZnO/AZO)all of which will be discussed in the following.The bene ?cial effect of the increased optical transmis-sion of the “new ”ZTO/In 2O 3-window structure is sup-posed to be restricted to J sc of the investigated cells.In order to spectrally resolve this gain independent on optical interference effects/losses and contact grid areas,the EQE was measured for one cell of each sample and corrected by the total re ?ectance,assigned here as the internal quan-tum ef ?ciency (IQE)(Figure 2(a)).As expected from the
optical results,the IQE for λ>700nm is signi ?cantly in-creased by applying the In 2O 3layer (less FCA).While the alternative buffer layer already leads to an improved IQE for λ<550nm compared with a cell with a CdS buffer,this value can be even more increased for λ<450nm if In 2O 3is employed.
Figure 2(b)illustrates the differences in IQE between the different con ?gurations and quanti ?es the potential gain in J sc by integrating ΔIQE multiplied by the AM1.5G spec-trum (dashed curves).If re ?ection losses and shading areas assumed to be zero (i.e.,perfect ARC and no grid),the gain in J sc due to less ultraviolet absorption (λ<550nm)sums up to (i)ΔJ sc ,IQE ≈+0.5mA/cm 2in the case of only a TCO exchange while keeping the buffer to ZTO and as high as (ii)ΔJ sc ,IQE ≈+2.6mA/cm 2compared with a CdS/i-ZnO/AZO window structure.The lower FCA in the In 2O 3would add another ΔJ sc ,IQE =1mA/cm 2,leading to (i)ΔJ sc ,IQE =+1.7mA/cm 2and (ii)ΔJ sc ,IQE =+3.6mA/cm 2,respectively,as a theoretical total improvement in photo current density.
A similar trend is seen for J sc ,EQE calculated from the measured EQE (corrected by shading grid area)of the cells where an ARC was applied to mitigate the effect of differ-ent re ?ection properties of the window con ?gurations.Here,the gain in short circuit current density by using the “new ”window layer structure compared with the ZTO/i-ZnO/AZO stack is slightly lower (ΔJ sc ,EQE =+1.0mA/cm 2),while the improvement related to conventional CdS/i-ZnO/AZO is similar (ΔJ sc ,EQE =+3.5mA/cm 2),com-pared with values calculated from IQE .The extracted values are summarized in Table I,together with the
Table I.Average IV parameters for devices with different window layer constellations and TCO thicknesses.The electrical properties of the different TCO layers are added,too.For J sc ,EQE and J sc ,IQE ,the absolute gains compared with the solar cells containing the CdS/i-ZnO/AZO top layers are given in brackets.
Cells with “cell-thick ”TCO
Cells with “module-thick ”TCO
ZTO/In 2O 3
ZTO/i-ZnO/AZO CdS/i-ZnO/AZO ZTO/In 2O 3ZTO/i-ZnO/AZO CdS/i-ZnO/AZO V oc (mV)
631±2638±2672±3624±3637±3673±2J sc ,IV (mA/cm 2
)36.3±0.234.6±0.232.1±0.235.7±0.432.6±0.131.8±0.1FF (%)72.7±0.472.8±0.974.0±0.970.9±0.771.4±0.775.0±0.4η(%)
16.6±0.216.1±0.3
16.0±0.3
15.8±0.2
14.8±0.3
16.0±0.1J sc,EQE (mA/cm 2
)35.834.832.3
35.732.731.8(+3.5)(+2.5)(+3.9)(+0.9)J sc,IQE (mA/cm 2
)38.636.93538.935.233.4(+3.6)(+1.9)(+5.5)(+1.8)R sh,TCO (Ω/sq)29282810
1111Number of cells 32323227(5excluded)
1232ALD Cycles
2000——6000——n (1/cm 3
)
2.3e207.3e207.3e20———μ(cm 2
/Vs)461313———ρ(Ωcm)
5.9e-4
6.4e-4 6.4e-4———
R s (Ωcm 2
)
0.34±0.060.69±0.180.45±0.130.31±0.050.37±0.070.32±0.07R p (Ωcm 2
)1148±1441224±871136±86805±1231548±587978±69n
1.73±0.03 1.61±0.09 1.73±0.08 1.75±0.07 1.79±0.03 1.61±0.03J 0(nA/cm 2
)
29±8
11±7
14±8
44±21
64±13
3±1
TCO,transparent and conductive oxide;FF,?ll factor;ALD,atomic layer
deposited.
Figure 1.Optical absorption spectra of the atomic layer depos-ited In 2O 3and sputtered i-ZnO/AZO layers on glass.The gray
curve illustrates the difference in absorption.
104
Prog.Photovolt:Res.Appl.2016;24:102–107?2015John Wiley &Sons,Ltd.
DOI:10.1002/pip
averaged results of the IV measurements.To illustrate the potential gain,the theoretical J sc without re ?ection and shading losses (J sc ,IQE )is added as well.
As reported previously,the transition from the CdS to the alternative buffer layer (same TCO)causes a drop in V oc and a slightly reduced FF [21],here roughly 35mV and 1%,respectively.Because of the gains in J sc ,the ef ?-ciency in the case of a buffer exchange remains more or less the same,considering the standard deviations (given as error estimation in Table I).The solar cells with a ZTO/In 2O 3window structure show a minor drop in V oc of 7mV and a similar FF compared with those with ZTO plus the common ZnO-based TCO.But considering the in-creased J sc ,this leads in total to the highest ef ?ciency among the investigated samples (Δη=+0.6%points com-pared with the common CdS/i-ZnO/AZO window structure).
For the sake of completeness,it should be mentioned here that we also deposited In 2O 3on CIGSe/CdS/i-ZnO cells.In contrast to the bare CdS ?lms,the additional i-ZnO layer (~60nm)allowed for a nucleation and a homog-enous and conformal growth of In 2O 3.But it turned out that the V oc of the corresponding solar cells was heavily re-duced by about 100mV (V oc ≈565mV).At this stage,we can only speculate that this deterioration might be related to an unfavorable band alignment or an incoherent lattice match at the i-ZnO/In 2O 3interface.
In theory,the gain in J sc by exchanging AZO for In 2O 3as a TCO should be even more increased for thin ?lm photovoltaic modules where usually a lower sheet re-sistance is needed to compensate for the longer lateral path length of electrons.To evaluate the potential,we processed solar cells with the same con ?guration as be-fore,but with thicker TCO layers,aiming for a typical TCO sheet resistance applied in modules (R sh ≈10Ω/sq).The averaged IV parameters are added in Table I.Considering the differences in J sc ,EQE (as well as in J sc ,IQE ),it is con ?rmed that there is an even larger potential for short circuit current density gains in CIGSe modules,compared with laboratory cells.Actually,only a very minor loss in J sc ,IV was measured when going from “cell-thick ”to “module-thick ”In 2O 3layers (ΔJ sc ,
IV =à0.6mA/cm 2
).The cells with the thick In 2O 3show a slightly reduced V oc (and FF )compared with the ones with thinner TCOs,while there is no such detrimental thickness effect in the case of cells with a ZnO-based TCO.However,the distinct bene ?t to J sc ,IV (ΔJ sc ,IV =+3.1mA/cm 2vs ZTO/i-ZnO/AZO and ΔJ sc ,IV =+3.9mA/cm 2vs CdS/i-ZnO/AZO)leads at least to a similar con-version ef ?ciency compared with cells with the common CdS/i-ZnO/AZO structure.
As a further particularity,we found several shunted cells with thick In 2O 3(~15%,excluded from averaging),which re ?ects the trends we saw at preliminary experi-ments (not included here).We believe that this is owing to the conformal growth and high coverage of conductive In 2O 3by the ALD process.If defect regions,which might induce shunt paths (e.g.,exposure of back contact),are present prior to TCO deposition,the probability of shunting in this area is increased if ALD is used instead of sputter deposition.
The origin of the deterioration in V oc for the cells with In 2O 3might be related to the slightly larger work function,W ,compared with ZnO [12,23];this affects the band align-ment at the TCO/buffer interface and might deteriorate the junction quality.We extracted the diode parameters (as-suming a 1diode model [24])from the IV characteristics under illumination (Table I).Considering the deviations in between measurements,no distinct changes in ideality factor n and dark saturation current density J 0could be de-tected.This indicates a similar main recombination path for both TCOs and cannot explain the decrease in V oc .
4.CONCLUSIONS
In summary,the potential of In 2O 3as a TCO for CIGSe so-lar cells using Zn-Sn-O as an alternative buffer layer
was
Figure 2.(a)Internal quantum ef ?ciency (IQE)spectra of the so-lar cell samples for each of the three different window layer con-?gurations.(b)Spectrally resolved difference in IQE related to the cell employing the ZTO/In 2O 3window structure and corre-sponding gain in J sc ,IQE .The light gray graph in the background
shows the AM 1.5photon ?ux (in arbitrary units).
105
Prog.Photovolt:Res.Appl.2016;24:102–107?2015John Wiley &Sons,Ltd.DOI:10.1002/pip
demonstrated.Owing to its higher mobility,suf?cient charge carrier density,and reduced optical absorption in the ultraviolet region,the layer properties outperform the common ZnO-based TCO.It could be shown that the de-creased optical absorption can lead to the desired increase in short-circuit current density and thereby in ef?ciency of processed CIGSe solar cells.However,we observed a minor loss in V oc when introducing ALD In2O3into the cell structure,which seems to become more pronounced for thicker layers and has to be investigated more in detail. Furthermore,the long-term stability of such devices is not proven yet and needs to be evaluated.It is notable that the derived electrical properties of the In2O3layer are very similar to those published earlier for sputtered boron-doped ZnO[25],which showed an improved mobility compared with AZO?lms.However,only considering the bare layer properties,the In2O3layer should still be the preferred choice due to its higher total transmission.
While the performance of CIGSe solar cells with a ZTO/In2O3window structure is promising,the introduc-tion of an In2O3TCO,made by a chemical deposition method,for cells using a CdS buffer seems to be challeng-ing and should be further investigated.
Because ALD is a relatively time-consuming process,it might not be suitable for industrial mass production. Therefore,we suggest studies on alternative deposition methods like chemical vapor deposition or sputtering to utilize In2O3as a TCO in CIGSe solar cells.Furthermore, the scarcity of indium and the role of photovoltaic modules in this regard are still under debate,and a further introduc-tion of indium into the device structure has to be evaluated with respect to ef?ciency,cost,and capacity. ACKNOWLEDGEMENTS
Financial support from the Swedish Energy Agency and Vinnova are gratefully acknowledged.Furthermore,we thank Tomas Nyberg for supporting the Hall measure-ments and Timo W?tjen for conducting TEM investigations.
REFERENCES
1.Pistor P,Greiner D,Kaufmann CA,Brunken S,Gorgoi
M,Steigert A,Calvet W,Lauermann I,Klenk R, Unold T,Lux-Steiner MC.Experimental indication for band gap widening of chalcopyrite solar cell ab-sorbers after potassium?uoride treatment.Applied Physics Letters2014;105:063901.DOI:10.1063/
1.489288
2.
2.Chiril?A,Reinhard P,Pianezzi F,Bloesch P,Uhl AR,
Fella C,Kranz L,Keller D,Gretener C,Hagendorfer H,Jaeger D,Erni R,Nishiwaki S,Buecheler S,Tiwari AN.Potassium-induced surface modi?cation of Cu(In, Ga)Se2thin?lms for high-ef?ciency solar cells.
Nature Materials2013;12:1107–1111.
DOI:10.1038/nmat3789.
3.Pianezzi F,Reinhard P,Chiril?A,Bissig B,Nishiwaki S,
Buecheler S,Tiwari AN.Unveiling the effects of post-deposition treatment with different alkaline elements on the electronic properties of CIGS thin?lm solar cells.Physical chemistry chemical physics:PCCP 2014;16:8843–8851.DOI:10.1039/c4cp00614c.
4.Jackson P,Hariskos D,Wuerz R,Kiowski O,Bauer A,
Friedlmeier TM,Powalla M.Properties of Cu(In,Ga) Se2solar cells with new record ef?ciencies up to
21.7%.Physica status solidi(RRL)-Rapid Research
Letters2014;1–4.DOI:10.1002/pssr.201409520.
5.Reinhard P,Pianezzi F,Bissig B,Chiril?A,Bl?sch P,
Nishiwaki S,Buecheler S,Tiwari AN.Cu(in,Ga)Se2 thin-?lm solar cells and modules—a boost in ef?ciency due to potassium.IEEE Journal of Photovoltaics2015;
5:656–663.DOI:10.1109/JPHOTOV.2014.2377516.
https://www.wendangku.net/doc/e813719029.html,emmle A,Wuerz R,Powalla M.Investigation of the
effect of potassium on Cu(In,Ga)Se2layers and solar cells.Thin Solid Films2015;582:27–30.
DOI:10.1016/j.tsf.2014.10.088.
https://www.wendangku.net/doc/e813719029.html,emmle A,Wuerz R,Powalla M.Ef?ciency en-
hancement of Cu(In,Ga)Se2thin-?lm solar cells by a post-deposition treatment with potassium?uoride.
Physica Status Solidi-Rapid Research Letters2013;
7:631–634.DOI:10.1002/pssr.201307238
8.Naghavi N,Abou-Ras D,Allsop N,Barreau N,
Bücheler S,Ennaoui A,Fischer CH,Guillen C, Hariskos D,Herrero J,Klenk R,Kushiya K,Lincot D,Menner R,Nakada T,Platzer-Bj?rkman C,Spiering S,Tiwari AN,T?rndahl T.Buffer layers and transpar-ent conducting oxides for chalcopyrite Cu(In,Ga)(S, Se)2based thin?lm photovoltaics:present status and current developments.Progress in Photovoltaics: Research and Applications2010;18:411–433.
DOI:10.1002/pip.955.
9.Witte W,Spiering S,Hariskos D.Substitution of the
CdS buffer layer in CIGS thin-?lm solar cells.Vakuum in Forschung und Praxis2014;26:23–27.
DOI:10.1002/vipr.201400546.
10.Richter M,Schubbert C,Eraerds P,Riedel I,Keller J,
Parisi J,Dalibor T.Optical characterization and modelling of Cu(In,Ga)(Se,S)2solar cells with spectro-scopic ellipsometry and TCAD.Thin Solid Films 2013;535:331–335.DOI:10.1016/j.tsf.2012.11.078.
11.Pan CA,Ma TP.High-quality transparent conductive
indium oxide?lms prepared by thermal evaporation.
Applied Physics Letters1980;37:163–165.
DOI:10.1063/1.91809.
12.Klein A,K?rber C,Wachau A,S?uberlich F,
Gassenbauer Y,Harvey SP,Prof?t DE,Mason TO.
Transparent conducting oxides for photovoltaics:
106Prog.Photovolt:Res.Appl.2016;24:102–107?2015John Wiley&Sons,Ltd.
DOI:10.1002/pip
manipulation of fermi level,work function and energy band alignment.Materials2010;3:4892–4914.
DOI:10.3390/ma3114892.
13.Walsh A,Da Silva JLF,Wei SH,K?rber C,Klein A,
Piper LFJ,Demasi A,Smith KE,Panaccione G, Torelli P,Payne DJ,Bourlange A,Egdell RG.Nature of the band gap of In2O3revealed by?rst-principles calculations and X-ray spectroscopy.Physical Review Letters2008;100:2–5.DOI:10.1103/PhysRevLett.
100.167402.
14.Piper LFJ,Demasi A,Cho SW,Smith KE,Fuchs F,
Bechstedt F,K?rber C,Klein A,Payne DJ,Egdell RG.Electronic structure of In2O3from resonant X-ray emission spectroscopy.Applied Physics Letters 2009;94:2–4.DOI:10.1063/1.3070524.
15.Nilsen O,Balasundaraprabhu R,Monakhov EV,
Muthukumarasamy N,Fjellv?g H,Svensson BG.Thin ?lms of In2O3by atomic layer deposition using In (acac)3.Thin Solid Films2009;517:6320–6322.
DOI:10.1016/j.tsf.2009.02.059.
16.Maeng WJ,Choi D,Chung K,Koh W,Kim G,Choi S,
Park J.Highly conducting,transparent,and?exible in-dium oxide thin?lm prepared by atomic layer deposition using a new liquid presursor Et2InN(SiMe3)2.ACS Ap-plied Materials and Interfaces2014;6:17481–17488.
DOI:10.1021/am502085c.
17.Libera JA,Hryn JN,Elam JW.Indium oxide atomic
layer deposition facilitated by the synergy between oxygen and water.Chemistry of Materials2011;23: 2150–2158.DOI:10.1021/cm103637t.
18.Mereu RA,Marchionna S,Le Donne A,Ciontea L,
Binetti S,Acciarri M.Optical and electrical studies of transparent conductive AZO and ITO sputtered thin?lms for CIGS photovoltaics.Physica Status
Solidi C2014;11:1464–1467.DOI:10.1002/ pssc.201300631.
19.Fortunato E,Ginley D,Hosono H,Paine DC.Trans-
parent conducting oxides for photovoltaics.MRS Bul-letin2007;32:242–247.
20.Hultqvist A,Edoff M,T?rndahl T.Evaluation of Zn–
Sn–O buffer layers for CuIn0.5Ga0.5Se2solar cells.
Progress in Photovoltaics:Research and Applications 2011;19:478–481.DOI:10.1002/pip.1039.
21.Lindahl J,W?tjen JT,Hultqvist A,Ericson T,Edoff M,
T?rndahl T.The effect of Zn1-x Sn x Oy buffer layer thick-ness in18.0%ef?cient Cd-free Cu(In,Ga)Se2solar cells.
Progress in Photovoltaics:Research and Applications 2013;21:1588–1597.DOI:10.1002/pip.2239.
22.Hultqvist A,Platzer-Bj?rkman C,Zimmermann U,
Edoff M,T?rndahl T.Growth kinetics,properties, performance,and stability of atomic layer deposition Zn–Sn–O buffer layers for Cu(In,Ga)Se2solar cells.
Progress in Photovoltaics:Research and Applications 2012;20:883–891.DOI:10.1002/pip1153.
23.Yeh TC,Zhu Q,Buchholz DB,Martinson AB,Chang
RPH,Mason TO.Amorphous transparent conducting oxides in context:work function survey,trends,and fac-ile modi?cation.Applied Surface Science2015;330: 405–410.DOI:10.1016/j.apsusc.2015.01.026.
24.Hegedus S,Shafarman WN.Thin-?lm solar cells:
device measurements and analysis.Progress in Photovoltaics:Research and Applications2004;12: 155–176.DOI:10.1002/pip.518.
25.Hagiwara Y,Nakada T,Kunioka A.Improved J sc in
CIGS thin?lm solar cells using a transparent conducting ZnO:B window layer.Solar Energy Mate-rials and Solar Cells2001;67:267–271.DOI:
10.1016/S0927-0248(00)00291-9.
107
Prog.Photovolt:Res.Appl.2016;24:102–107?2015John Wiley&Sons,Ltd. DOI:10.1002/pip