Cu2(ZnxSn2-x)(SySe1-yh MONOGRAIN MATER lA LS FOR PHOTOYOL TAICS

E. Mcllikov1, M. Altosaar1, J. Raudoja1. K. Tinm1o1, 0 . Volobt~eva 1 , M. Kauk1, J. Krustok1, T. Varema1, M. Grossbcrg1, M. Danilson1, K. Muska 1

•2

, K. Emits2, F. Lehner 2, D. Meissneru 1 Dpt. Mat. Scicncc, Tallinn University ofTechnology. Ehitajatc tee 5 Tallinn 19086, Estonia, cnnm@staff.ttu.cc, ph: +372620-2798, fax. -3367 2crystalsol OÜ, Akadecmia tee 15a. 12618 T alli1m, Estonia

ABSTRACT ln monograin solar cells powders rcplace wafcrs or thin tilms. This allows for eheaper and much more efficient materials production and minimize materials loss. The separation of materials fonnation from the module tabrication - allowing for all temperatures and purity precautions- is th e very important advantagc of the monograin layer bascd tcchnology. Large area module fabrication proceeds without any high tcmperature steps in a continuous roll-to-roll proccss. No up-scaling problems arisc as is typical of thin-fi lm tcchnologies since a homogenous powder Ieads to homogenous modules. The influence of different tcchnological parameters to the materials properties was studied and from CZTS=Cu2(Zn,Sn2.,)(SySe1.y)4) monograin powders solar cclls havc produced with parameters Yoc up to 690 mV, fill factors up to 65 % and lsc morc than 20 mA/cn/. lt was foundcd that changing ofthe sulfur/sclcnium ratio allows a change of 1.04 and 1.72 eY the band gap ofCu2(Zn,Sn2.,)(SySe1-y)4.

lNTRODuCT!ON Climate changc and insecurity of power supplies arc among the greatest problems for

humankind at the moment and solutions are to be found within only a few decadcs. Large-scale introduction of rcncwable power systems in combination with a strong increase of power efficiency and saving will bc thc only way out. For example, in ordcr to stabilizc thc world's climate a eontinuous growth of PV by at least about 20% per year will be needed during thc coming decadcs that will Iead to PV installation o f at least ten TWp before 2050. To realize thesc plans new abundant matcrials and cheap technologies are necdcd [ l).

The technologies used today for manufacturing solar cclls are the wafer and the thin fi lm tcclmologies [2). The wafer technology is bascd on the growth and use of very I arge monoerystals or the casting of ultrapure silicon or A385 materials. This is a very expensive proeess not only in tenns of money but also in tcrms of energy input. Later these !arge singlc or polycrystals with a high degree of chemical purity and physical perfection have to be sawn into wafers which are subjected to sophisticatcd methods of Oxidation, diffusion and chemical treatment in order to form localized regions of different types of conductivity neecled for highly efticient solar cell stmctures. Thc idea of making crystall ine solar eells by growing very I arge crystals and thcn eutting thcm into thin wafers is in fact not the most intelligent way to obtain perfect materials and structures. T hin film technologies of solar cells production can be applied, but the technieal parametersofthin film solar cells are as a rule much worse than those of crystall ine solar cclls. On the other hand, thin film technologics are, in general, much eheaper than monocrystalline materials bascd technologies, both with respect to financia l expenses and energy costs.

Although silicon as semiconductor material is still dominaring even in the thin film solar cell produetion, photovoltaic cells based on more complex compound scmiconductor materials are bccoming increasingly important. Unconventional solar cell materials that are abundant but much eheaper to produce than sil icon could substantially reduce the overall cost of solar photovoltaics. Currently the most widely used compound solar cell materials, cadmium telluride (CdTe) and copper indium diselcnide, copper indium gallium diselenide or the respective sulfides (CIGS), contain resource-l imited elements (Tc and In) that are alreacly today about ten times moreexpensive than other metals and do not a llow for the provision photovoltaic energy in amounts ncedcd for the very !arge scale applications in future [3). Onc of thc eurrently available most promising new unconventional matcrials is CZTS, here meaning all coppcr zine tin sulfo-sclcnides including e.g.

137

Cu2(ZnxSn2_,J(SySe1_y)4 Monograin Materials for Photovoltaics

Cu2ZnSnS4, Cu2ZnSnSc4 and Cu2ZnSn(S,Se t.x )4 with their abundant and nontox ic constituents. With a d irect band gap that depcnds on thc ratio of sulphur to selenium it can be tuned to the optimum for so lar energy conversion even for tanclem structurcs. These materials havc also a high optical absorption coefficient (> 104 cm-1

) [ 4] . This makes thcsc matcrials ideal as adsorbcr layers for low cost high efficiency terrestrial photovoltaic clevices [5].

A number of reports have been published on the prope11ies of CZTS materials prepared as thin fi lms by various vacuum as weil as low-cost chemical preparation techniques. The most important are spurtering and vacuum evaporation of metals or constituent meta I binaries fo llowed by chalcogcnization [6, 7, 8], electrodeposition [9, 10], chemical spray pyro1ysis [ 11], and soft chemical technologies [ 12, 13]. Thc cfficicncy of dcveloped polycrystalline films solar cells is lower than theoretically possible. However, in a rcccnt papcr Todorov. Reuterand Mitzi achieved a 9.6% solar efficiency confirmed by NREL [13]. As a rcsult thc optimizati.on of growth conditions are regularly discusscd by different authors [14, 15]. ln aclclition to the abovc, thc manufacturing of thin films w ith a largc surfacc area, suitablc stoichiometric composition and good reproducibility is a technically complicated task and is not up to end solved even on the Iabaratory Ievel.

ln addit ion to the monocrystalline and the thin film approach, thc third alternative to prcpare solar cell structures is the use of powder materials [16]. Powder technologies are the cheapcst teclmologics for materials production. At the same ti me, a lthough scvcral companies and rescarch institulians have made considerable efforts, powdcr mcthods for solar cell applications have not fo und widespread use yet . lt has been shown that isothermal rccrystallization of initial powdcrs in different malten fluxcs appears to bc a rclatively simple, inexpensive and a convenient method to produce CIS and CZTS powdcrs wirb an improved crystal structure, that are perquisite for solar cell use [16-19] . Powders consist of small single crystalline grains. ln monograin solar ce ll s powdcrs replace monocrystalline wafers or thin films. This allows for eheaper and much morc cfficient materials production that minimizcs matciials lass . Tbe Separation of materials formation from the modulc fabrication - allowing for all te.mperatures and purity prccautions- is an additional very important advantage. Largc arca modulc fabrication could be clone without any high temperature steps in a continuous roll-to-rol l proccss. Homogenaus powders leacl to homogenaus modules and do not result in up- scaling problems.

EXPERIMENT AL Monograin powdcr mateti als wcre syn thesized from meta! binaries (CuSe(S), ZnSe(S),

SnSe(S) and eiemental se1enium or sulphur, rcspcctively) in a malten flux using an isothermal rccrystallization process. The precursors wen: thermally anncalcd in evacuated quartz ampoulcs . The evolution of crystal shape and morphology of the monograin powdcrs was analyzcd by electron imaging using a high-rcsolution scanning electron microscope (SEM) Zeiss GL TRA55 with the compositional cantrast detcctor EbS. Thc chcmical composition a nd the distribution of components in powder crystals werc detcnnincd using an cncrgy dispersive X-ray analysis (EDX) system). XRD patterns were recordecl by using a Bruker AXS D5005 diffractometer using monochromatic Cu Ka­radiation.

Monograin layer (MGL) solar cells (graphite/CZTSSe/CdS/ZnO) wcrc madc from grains with diameters of 56-63 f.lll1 [ 19-2 1] . Powder crystals were covered with chcmically deposited CdS buffer layers. For the MGL formation a monolayer of CZTS powder crystals was glued tagether by a thin layer of epoxy. After polymerization of this epoxy, i-ZnO and ZnO:AI were deposited by RF­spuucring onto the open surface of the layer. Solar ccll structures were comp1cted by vacuum evaporation of 1-2 f.ll1l rhic k Tn grid-contacts onto the Znü window laycr. Subsequently, the layer was gluecl on to glass substratcs . The opening of the back contact areas of the crystals that were orig inall y insidc the epoxy was clone by etching the epoxy in H2S04 and by abrasive treatment. Graph ite paste was used for thc back contacts. Thc solar cell effi ciency was measured in an Oriel class A solar simulator. l- V curves were measured using a Kcithley 2400 sourcc mctcr.

138 Materials Challenges in Alternative and Renewable Energy

CuAZn.Sn2_,J(SySe,_y)4 Monograin Materials for Photovoltaics

R.ESUL TS AND DISCUSSION The structural, morphological, electrical and optical properties of Cu2ZnSn(S,Se)" materials

dcpcnd strongly on thc composition and on the additional thermal and chemica l trcatmcnts. Fig. l shows thc morphologies of CZTS powder crystals grown in the Kl flux. The crystals have tetragonal shape with rounded grain edges. A change in thc chemical nature of the flux Ieads to variations in crystals shape.

1.20

l,L5

1.\0

~ 1 .0~ g 1,00

" ~ 0,9$

~ 0.90

0.~

.• . ... _ ... ..-... { ' tl '·'

CLIIZn-Sn

o.so.l-...--~-~-~--~-...---1.00 1.04 1.08 1.12 1.16 1.20

Zn Sn m inilials

Fig I. SEM m1crographs of Fig. 2. Dcpcndcnce of Cu2ZnSn(S,Sc)4

Cu2ZnSn(S,Sc)4 monograin powdcr monograin powdcr compos ition on precursor crystals gr0\\11 in Kl flux. composition .

Fig. 2 shows the results of thc compositional analyses of as-grown Cu2ZnSn(S,Sc)• matcrials as a function of the initial precursor composition. With thc Zn/Sn compositions ratio increasing in the precursors thc obtained powders contain an increased ratio of Zn to Sn but a dccreased amount of Cu. At the same time thc content of chalcogcns (S+Sc) in the materials docs not depend on the Zn/Sn ratio. At stoichiomctric ratios of 25%, 12.5%, 12.5% and 50% fo r Cu. Zn, Sn. and S+Se, in Cu2ZnSn(S,Sc)4 respectively, all studicd materials were Zn-rich and Cu-poor and have an stoichiometri c content of (S+Se). An overstoichiometric chalcogen amount in the starting materials Ieads to an increase of thc relative contents of Zn and Se in the as-grown powdcrs (Fig. 3).

30 28

~ 26 Cu " 24 ö ~ 22

.g 20 18 e 16 c s 14

g 12 u 10 8

1.10

0

!.OS~ 7n.·sn

.2

0 5 Q,·ers10ichiometric Sc (molc%)

, 1.00 j§ c g

0.95 3

-" C3h

~~ 100 ...

Fig. 3. The influcnce of the chalcogcn (Se) Fig. 4. Surface composition of etchcd in content on thc composition of Cu2ZnSn(S,Se)• di fferent ctchants Cu1ZnSn(S,Se)4 powdcrs powders

By comparing as-grown powdcrs crystal surfacc compositions with thc bulk compositions it was found that thc surface of as-grown Cu2ZnSn(S,Sc)4 crystals was Sn-rich whilc thc bulk of crysta ls was Zn-rich. This was the result of thc contamination of thc powder crystals by

Materials Challenges in Alternative and Renewable Energy 139

Cu2(Zn.Sn2_,J(SySe1.y)4 Monograin Materials for Photovoltaics

components of Cu2ZnSn(S,Se)4 dissolved in the fiux. In order to remove thesc contaminations (other phases) from thc surfaces and to improve the developed Cu2ZnSn(S,Se)4 solar cells performance, the chemical trcatments in different etchanrs (HCI, KCN, Br-MeOH and NH40H) wcre performed. 1t was found that eiehing with KCN or HCI increases the ratio of Zn to Sn on the surface. At the same time HCI etching Ieads to a slightly chalcogen-poor surfacc. Treatment with KCN dissolves mainly Cu, Sn and chalcogen. and ammonia solution remove selectively Cu and chalcogen in an approximate ratio of I :2. Fig. 4 shows the results of this influencc of different ctchants on the surface composition of powder crystals.

Fig.5 . EBIC picture of MGL solar cell from material composed of Cul.ssZn~,oSno.9s

(So.sSeo.l)3.9;

~ 0,2

~ 0.1

c

.(),l 0,0 o.s 1.0

Voltagc (V)

Fig.6. I-V dependence of Lhe Cu2ZnSn(S,Se)4 monograin layer cell (certified in the Calibration Lab of Fraunhofer ISE, Freiburg, 2009)

Electron beam indueed current (EB!C) investigations of MGL solar cclls indieate that all crystals in :v!GL operate with ncarly the same efficiency. Fig. 5 represents EB IC pieture for the MGL solar eell from the material composed of Cul.ssZn~,oSno.9s (So.sSco.2h9s- Current-voltage dependenees ofCu2ZnSn(S,Se)4 MGL solar cells show V oc values of over 690 mV, fill factors ofup to 65 %, and short circuit currents of up to 20 mA/cm2. Efficiencics of the active area of these solar cclls of up to 6.4 % wcre detcnnincd, total area efficiencies of 5.9 % were certified by calibration Iab measurements (fig. 6). Varying the S/Se ratio in thc material a llows one to control the band gap of Cu2ZnSn(SSc)4 materials and with this the quantum efficiency of the solar ccll made on their base.

CONCLUSIONS 1t is shown that the Cu2ZnSn(S,Se)4 monograin technology allows solar ccll production with

conversion efficiencies of currently up to 6 %. The solar cells developed hcre have Y oc up to 690 mV, fill factors of up to 65 % and I,c of about 20 mNcm2. Changing the sulfur/selcnium ratio in materials allows a changc thc change of the band gap ofCu2ZnSn(S,Se)4 between 1.04 and 1.72 eV.

ACKNOWLEDGEMENTS This work was supported by the target fioancing by HTM (Estonia) project SFOI40099s08

and by Lhe Estonian Science Foundation grants G-8147, G-8282 and G-7678.

REFERENCES I. Avision for photovoltaic technology for 2030 and beyond, European Commission, 2004 2. A. Luque, S. Hegedus, Handbook of Photovoltaic Science and Engineering (Wiley), 2006 3. A. Feltr in, A. Freindlich, Materia l considerations for terawatt Ievel deployment ofphotovoltaics,

Renewable Energy, 33, 180 - 185 (2008). 4. K. Ito, J. Nakazawa, Elcctrieal and optical properlies of stannite-type quaternary semiconductor

thin ftlms, Jpn. Appl. Phys, 27, 2094 - 2097 (1988).

140 · Materials Challenges in Alternative and Renewable Energy

Cu2(ZnxSn2_x)(SySe1_y) 4 Monograin Materials for Photovoltaics

5. K. Jimbo, R. Kimura, T. Kamimura, S. Yamada, W. S. Maw, H. Araki, K. Oishi, H. Katagiri, CuzZnSnS.-typc thin film solar cclls using abundant matcrials. TSF, 515, 5997 - 5999 (2007).

6. 0 . Volobujeva, J. Raudoja, E. Mellikov, M. Grossberg, S. Bereznev, R. Traksmaa, Cu2ZnSnSe4 films by selenization of Sn- Zn- Cu sequcntial films, J. Phys. Chem . Solids 70, 567 - 570 (2009).

7. 0. Volobujcva. E. Mcllikov, J. Raudoja, M. Grossberg, S. Bereznev, M . Altosaar, R. Traksmaa, SEM analysis and selenization of Cu- Zn-Sn sequential films produced by evaporation of metals, in: Proceedings: Conference on Optoelecrronic and Microelecrronic Marerials and Devices, JEEE Pubfishing, 257-260 (2009) .

8. P. M. P. Salome, F. A. Fernandes, A. F. de Cunha, Morphological and Strucrural characterization of Cu2ZnSnSc4 thin films grown by thc sclcnization of eiemental precursor layers, TSF, 517, 2531 - 2534 (2008).

9 . J. J. Scragg, P. J . Oale, L. M. Peter, G. Zoppi, I. Forbes, New routes to sustainable photovolta ics: evaluation of Cu2ZnSnS4 as alternative absorber material, Phys . Srm. Sol . (b ) 245, 1772 - 1778 (2008).

10. A. Ennaoui, M. Lux-Steincr, A. Weber. 0. Abou-Ras, I. Kötschau, H. -W. Schock, R. Schurr, A. Hölzing, S . Jost, R. Hock, T. Voß, J. Schulze, A. Kirbs, Cu2ZnSnS. thin film solar cclls from electroplated precursors : novel low-cost perspective, TSF, 517, 2511 - 2514 (2009).

J 1. N . Kamoun, H. Bouzouita, B. Rezig, Fabrication and characterization ofCu2ZnSnS. thin films deposited by spray pyrolysis technique, TSF, 515, 5949 - 5952 (2007).

12. T. Todorov, M. Kita, J. Carda, P. Escribano, Cu2ZnSnS4 films depositcd by soft-chemistry method, TS F, 517, 2541-2544 (2009).

13 . T . K. Todorov, K. B. Rcutcr, D . B. Mitzi, High-Efficicncy Solar Ccll with Earth-Abundant Liquid-Processed Absorber, Adv. Marer. 22, 1 -4 (201 0).

14. H. Katagiri, K. Jimbo, W. S. Maw, K. Oish i, M. Yamazaki, H. Araki , A . Takeuchi, Dcvclopmcnt of CZTS-bascd thin film solar cclls, TSF, 517, 24 55 - 2460 (2009).

15. A. Goetzberger, C. Hebling, H. \V. Schock, Photovoltaic materia1s, history, status and outlook, Mar . Sei . Engin. Rev. 40 (2003), I - 46.

16. M. Altosaar, A. Jagomägi , M . Kauk, M. Kmnks, J. Krustok, E. Mellikov, J. Raudoja, T. Varema, Monograin layer solar cells, TSF, 431-432, 466 - 469 (2003).

17. M. Altosaar, E. Mcllikov, CulnSe2, Monograin Growth in CuSc-Sc liquid phase, Jpn. J. Appl. Phys., 39, 65 - 66 (2000).

18. Mellikov, E., Altosaar, M ., Krunks, M., Krustok, J ., Varema, T., Volobujeva, 0., a .o., Research in solar cell techno1ogies at Tall inn University ofTechnology. TSF 516,7125- 713 (2008)

19. D. Mcissncr, E. Mcllikov, M. Altosaar, Monograin Powdcr and Monograin Membrane Production, Europeon Patent J 097262, Aug. 21 , 2002.

20. M . Altosaar, J. Raudoja, K. Tinuno, M. Danilson, M. Grossberg, J. Krustok, E. Mellikov, Cu2Zn1_xCd, Sn(Scl -ySy)4 solid solutions as absorbcr matcrials for solar cclls, Plzys . Starus Solidi A 205, 167- 170 (2008) .

21. E. Mellikov, 0 . Meissner, T. Varema, M. Altosaar, M. Kauk, 0 . Volubujeva, J. Raudoja, K . Timmo, M. Danilson, Monograin matcrials for solar cclls, Solar Energy Mar . Solar Cells 93, 65 - 68 (2008).

Materials Challenges in Alternative and Renewable Energy 141