24.5% e$ciency PERT silicon solar cells on SEH-1

24.5%e$ciency PERT silicon solar cells on SEH MCZ substrates and cell performance on other SEH CZ and FZ substratesJianhua Zhao*,Aihua Wang,Martin A.Green Photo v oltaics Special Research Centre,Uni v ersity of New South Wales,Sydney NSW2052,Australia AbstractHigh-e$ciency passivated emitter,rear totally di!used(PERT)and passivated emitter,rear locally di!used(PERL)cells have been fabricated on magnetically con"ned Czochralski(MCZ) silicon wafers supplied by SEH,Japan.One PERT MCZ cell demonstrated24.5%energy conversion e$ciency con"rmed at Sandia National Laboratories under the standard 100mW/cm global AM1.5spectrum at253C.This is the highest ever reported e$ciency for a silicon cell made on a non-FZ silicon substrate.A light boron di!usion in the PERT cell structure not only reduced the cell series resistance but also improved the cell open-circuit voltage.Other experimental results are also discussed concerning cells fabricated on other SEH CZ(Ga),CZ(B)and FZ substrates. 2001Elsevier Science B.V.All rights reserved.Keywords:Silicon;Solar cells;FZ;MCZ;CZ;High e$ciency1.IntroductionThe e$ciency of single crystalline silicon solar cells has been rapidly improving over the last one and a half decades.To obtain the highest possible cell e$ciencies,the highest quality#oat zone(FZ)substrates have been commonly used for high-e$ciency silicon cell research[1}3].An e$ciency record of24.7%for a passivated emitter,rear locally di!used(PERL)silicon cell on an FZ substrate has been reported in a separate paper at the PVSEC-11conference[4].*Corresponding author.Tel.:#61-2-9385-5246;fax:#61-2-9662-4240.E-mail address:j.zhao@.au(J.Zhao).0927-0248/01/$-see front matter 2001Elsevier Science B.V.All rights reserved.PII:S0927-0248(00)00155-028J.Zhao et al./Solar Energy Materials&Solar Cells66(2001)27}36Also,as a single crystalline approach,the Czochralski-grown(CZ)material has produced considerably poorer cell performance.The CZ materials have markedly lower minority carrier lifetimes than FZ materials,attributed to the higher oxygen and carbon contents in the CZ materials.Hence,the highest reported CZ cell e$ciency is only22.0%[3],which is considerably lower than that from FZ substrates.The high oxygen content in the standard boron-doped CZ substrates has also resulted in cell performance degradation due to the reaction of boron with oxygen[5,6].Magnetically con"ned Czochralski-grown(MCZ)silicon material has been thought for a long time to be a good material candidate for high-e$ciency silicon cells.This material has high minority carrier lifetimes attributed to the low oxygen and carbon contents.Recent research interest at the University of New South Wales has been aimed at obtaining high cell e$ciencies from boron-doped MCZ silicon substrates, supplied by Shin-Etsu Handotai Co.(SEH),Japan.A research project has also been conducted to fabricate high-e$ciency cells on gallium-doped CZ(CZ(Ga)),boron-doped CZ(CZ(B)),and FZ silicon wafers with a variety of resistivities,also supplied by SEH.2.E4ective carrier lifetime testing on SEH MCZ,CZ and FZ silicon wafersThe measurement of e!ective carrier lifetime on the silicon wafers has been a useful tool to detect the material quality of silicon wafers.A Leo Giken photo-conductivity decay(PCD)e!ective carrier lifetime tester has been used for this experiment.Since the e!ective carrier lifetime,which includes the e!ect of surface recombination,is tested by such an equipment,the wafer surface had to be well passivated before lifetime testing.The surface passivation was done by thermally growing an SiO layer of2000A thickness in a trichloroethane(TCA)ambient.After the oxide growth,the wafers were sintered in forming gas,and then subjected to an alneal(aluminium anneal)treatment at about4003C.In the alneal process,a thin layer of aluminium was deposited onto the wafer surface,the wafer with the aluminium layer was sintered,and then the aluminium layer was removed by phosphoric acid[1].Ten di!erent SEH wafers of di!erent growth methods were selected for this experiment.The e!ective carrier lifetime of each wafer was tested at di!erent stages of the surface passivation process.The results of these tests are listed in Table1.It is seen that the e!ective carrier lifetime was increased considerably by the forming gas anneal and by the alneal.The cell open-circuit voltages were also increased signi"cantly after these two processing steps under similar conditions.It was a surprise that the high-resistivity gallium-doped CZ(Ga)substrate gave the highest carrier lifetime of the SEH wafers,although the cells on such wafers did not give the best performance. Other higher resistivity MCZ and FZ wafers also showed high carrier lifetimes longer than1ms.Only the boron-doped CZ(B)and the lowest resistivity CZ(Ga)wafers had e!ective carrier lifetimes of less than100 s.Accordingly,the cells made on these wafers gave the poorest cell performance.J.Zhao et al./Solar Energy Materials&Solar Cells66(2001)27}3629 Two high-resistivity n-type FZ wafers from another supplier were also passivated asa monitor during the passivation process.Much higher carrier lifetimes were tested onthese monitor wafers due to their lower doping levels.Due to the low overall energy level of the laser beam in the Leo Giken system,thelifetime signals were not strong enough for lifetime tests of some low-resistivity wafersbefore the forming gas sintering,as shown in Table1.However,after the forming gassintering,the e!ective carrier lifetime was signi"cantly increased.This increasedcarrier lifetime also increases the concentration of the light-generated carriers in thesubstrates.This in turn results in stronger photo-conductivity signals for correctcarrier lifetime measurements.After storage in a low-level illumination area for8days,the e!ective carrier lifetimesof these wafers were re-tested,as shown in Table1.All the wafers showed stable carrierlifetimes except for Box3MCZ wafers with high oxygen content.It must also be notedthat the apparently stable carrier lifetime for the CZ(B)wafers may not be reliable dueto the low signal and high noise levels during the lifetime measurement.3.PERL,PERT(s)and PERT cell structures used in this researchThree di!erent cell structures have been used in the present research.Fig.1showsthe structure of a standard high-e$ciency PERL cell,which has been used mostly onFZ substrates.The features of PERL cells have been discussed elsewhere[1].Themajor advantage of the PERL cells is passivation of most of the cell surface areas withhigh-quality TCA-grown SiO .Both front and rear metal contact areas are designed to cover very little cell areas.These contact areas are further passivated by heavilyphosphorus-and boron-di!used regions.The recombination at the heavily boron-doped regions was further reduced to a minimum by designing these regions intosmall localised areas.The inverted-pyramid front surface and the rear surface mirrorform an excellent light-trapping scheme.Hence,the PERL cells give very highopen-circuit voltage and short-circuit current density on FZ substrates with highpost-processing minority carrier lifetimes[1].Standard PERL cells were initiallyfabricated on the SEH MCZ substrates.This resulted in a high cell e$ciency of22.5%and a high cell open-circuit voltage of691mV.The SEH CZ(Ga)wafers were single-side mechanically polished,which was notalignable for the rear boron and contact patterns.Hence,a passivated emitter,reartotally(singly)di!used(PERT(s))cell structure,as shown in Fig.2,was used tofabricate cells on such wafers,since these PERT(s)cells do not require any alignmentto process the cell's rear side.The total rear boron di!usion in the PERT(s)cell canalso reduce the current crowding e!ect around the rear contact[7].Due to the highrear surface recombination expected from a heavy boron di!usion,a lighter totalboron di!usion was used to provide the rear contact passivation than in the standardPERL cell.However,this lighter boron di!usion would result in poorer passivation ofthe rear metal contact area.Hence,it has to compromise the loss due to the increasedrecombination at the rear metal contact and other rear surface areas,and the gain inthe cell"ll factor.T a b l e 1T h e e !e c t i v e c a r r i e r l i f e t i m e t e s t e d o n t h e S i Os u r f a c e p a s s i v a t e d S E H w a f e r s o f d i !e r e n t g r o w t h m e t h o d s B o x S E H G r o w t hR e s i s t i v i t y ( c m )O x y g e n c o n t e n t (p p m a )a f t e r T C A S i O g r o w t h ( s ) a f t e r F G a n n e a l i n g ( s ) a f t e r a l n e a l i n g ( s ) a f t e r 8d s t o r a g e ( s )1C Z (B )0.813N o s i g n a l 2116182C Z (B )0.819N o s i g n a l 1913113M C Z 5.211}15341024002354M C Z 1.21.5N o s i g n a l 274905135M C Z 4.81.2}2.33395164016416F Z 0.6(0.1N o s i g n a l 15880N o s i g n a l 7F Z 4(0.121602001958C Z (G a )1.61951159589C Z (G a )5.31917671070102510C Z (G a )9.1203215018701913T e s t 1n -F Z 100*3600330050003376T e s t 2n -F Z 100*160018002520198830J.Zhao et al./Solar Energy Materials &Solar Cells 66(2001)27}36Fig.1.Passivated emitter,rear locally di !used (PERL)cellstructure.Fig.2.Passivated emitter,rear totally (singly)di !used (PERT(s))cell structure.This structure was used for processing of the single-side polished wafers since it does not require rear side alignment.This structure also reduced the cell series resistance.A full passivated emitter,rear totally (doubly)di !used (PERT)cell structure,as shown in Fig.3,was also employed to further improve the cell performance for the rear alignable wafers.Since di !erent boron di !usions were separately processed for the total rear surface and for the rear metal contact area,this allows fully optimised doping pro "les for the two regions.Hence,the cell can be optimised for maximum open-circuit voltage as well as the best "ll factor.This PERT structure was "rst reported to improve the "ll factors for very high-resistivity substrates of 100 cm in the early stages of the PERL cell development in early 1991[8].A high cell "ll factor of 0.786was demonstrated for a PERT cell on a high-resistivity Wacker p-type FZ substrate of 100 cm at the time [8].In the present experiments,it was found that the PERT cells on SEH MCZ substrates have given open-circuit voltages as high as standard PERL cells on Wacker FZ wafers.They also gave very high e $ciencies.It was clear that the rear total light boron di !usion of the PERT(s)cell was not enough to passivate the rear metal contact areas.J.Zhao et al./Solar Energy Materials &Solar Cells 66(2001)27}363132J.Zhao et al./Solar Energy Materials&Solar Cells66(2001)27}36Fig.3.Passivated emitter,rear totally(doubly)di!used(PERT)cell structure,an improved cell structure from the PERT(s)structure.4.Cell performanceTable2lists the performance of di!erent cell structures on SEH MCZ and CZ(Ga) substrates,as independently con"rmed at Sandia National Laboratories.Table3lists the performance of cells on SEH FZ and CZ(B)substrates tested at UNSW,where no anti-re#ection coating nor periphery shading was applied to these cells.The anti-re#ection coating is expected to increase the cell current density by about15%.The peripheral collection for low-resistivity CZ and FZ cells listed in Table3was very small.This peripheral collection is very high for the high-resistivity FZ cells,and hence the cell current density is not directly comparable to the peripheral shaded measurement for this case.4.1.PERT and PERL cells on MCZ substratesThe total rear surface di!usion in the PERT cells not only reduced the cell series resistance,but also increased the cell open-circuit voltages.As listed in Table2,PERT cells on SEH MCZ substrates have demonstrated very high performance as con"rmed by Sandia.Cell Ws9-4b gave24.5%cell e$ciency.This is very close to the24.7% e$ciency which is the best achieved by standard PERL cells on Wacker FZ wafers [4].Some of these MCZ cells also showed very high open-circuit voltages of up to 707mV,which is as high as those for standard PERL cells on Wacker FZ wafers. Standard PERL cells were made on MCZ wafers of5.2 cm resistivity.These PERL cells had marginally lower open-circuit voltages and lower"ll factors than the PERT cells on similar4.8 cm resistivity substrates.It might be the PERT rear surface passivation or the higher minority carrier lifetime which has improved the open-circuit voltage of the cells on the4.8 cm resistivity substrates.The substrates of cells Ws8-3b,Ws8-2a and Ws8-4b were chemically thinned down by100 m.This wafer thinning process may not only reduce the wafer thickness,but also change the wafer surface properties.Previously,thinning Wacker FZ wafers usedT a b l e 2T h e p e r f o r m a n c e o f 4c m c e l l s o f d i !e r e n t c e l l s t r u c t u r e s t e s t e d a t S a n d i a N a t i o n a l L a b o r a t o r i e s u n d e r t h e s t a n d a r d g l o b a l A M 1.5s p e c t r u m (100m W /c m )a t 253CC e l l n a m e E !(%)< (m V )J(m A /c m )F F B o x S E H G r o w t h O x y g e n c o n t e n t (p p m a )R e s i s t i v i t y ( c m )W a f e r t h i c k n e s s ( m )C e l l t y p eW s 8-3b 23.570241.20.8113M C Z 11}155.2300P E R L W s 8-2a 23.470740.90.8074M C Z 1.51.2300P E R T W s 8-4b 23.670641.20.8144M C Z 1.51.2300P E R T W s 9-4b 24.570441.60.8355M C Z 1.2}2.34.8400P E R T W s 6-5b 14.356332.40.7838C Z (G a )191.6200P E R L W s 4-2b 20.064340.60.7679C Z (G a )195.3250P E R T (s )W s 5-3a 20.963541.00.80210C Z (G a )209.1250P E R T (s )J.Zhao et al./Solar Energy Materials &Solar Cells 66(2001)27}3633Table3The UNSW-tested open-circuit voltage performance of4cm cells without anti-re#ection coating and without peripheral shadingCell name< (mV)Box SEH Growth Oxygencontent(ppma)Resistivity( cm)Waferthickness( m)Cell typeWs10-2f6471CZ(B)130.8400PERL Ws10-4a6362CZ(B)190.8400PERL Ws10-6f6996FZ(0.10.6400PERL Wh25-3c7056FZ(0.10.6400PERT Ws10-7f7027FZ(0.14400PERL Wh25-5c7047FZ(0.14400PERT Ws11-3b6769CZ(Ga)19 5.3200PERTto reduce the cell open-circuit voltages[9].However,these thin SEH MCZ cells have demonstrated the highest open-circuit voltages.The surface conditions seem to be slightly di!erent for SEH MCZ wafers compared to those for Wacker FZ wafers.Also, it is clear that the thinner substrates have resulted in lower current densities than the thicker ones.4.2.PERT(s)cells on CZ(Ga)substratesCZ(Ga)material has a particular interest for potentially higher radiation resistance as a candidate for space silicon cells.Hence,no alneal process was used for these CZ(Ga)cells due to the consideration of the future space radiation damage. As shown in Table2,the cell performances on the CZ(Ga)substrates were rather poor.Considerably lower open-circuit voltages and e$ciencies were measured for these cells.Ws6-5b was mechanically polished on both sides at UNSW to allow a full PERL process.However,it might be this polishing process or the originally lower carrier lifetime,as seen in Table1,that has resulted in very low cell performance.Also,all the cells on CZ(Ga)substrate gave considerably lower< performance for Sandia tests than for UNSW tests.However,no performance degradation was found in cells stored at UNSW for over2months under low illumination.Hence,we believe that something happened during the transportation from UNSW to Sandia which caused these degradations.Later,some PERT cells were also fabricated on the SEH CZ(Ga)substrates with the alneal treatment.These cells have demonstrated open-circuit voltages of over 670mV.Unfortunately,the rear metallisation of these cells was damaged during the processing of these particular cells.Hence,these cells gave very low"ll factors and low e$ciencies.However,the potential of these CZ(Ga)substrates is obvious,not only from these high open-circuit voltages but also from the highest tested minority carrier lifetime(Box10)among all the SEH wafers.34J.Zhao et al./Solar Energy Materials&Solar Cells66(2001)27}36J.Zhao et al./Solar Energy Materials&Solar Cells66(2001)27}3635 4.3.PERT and PERL cells on SEH FZ and CZ(B)substratesAs shown in Table3,the PERT and PERL cells on SEH FZ wafers have demonstrated very high open-circuit voltages of up to705mV before anti-re#ection coating.It is expected that these cells should give a performance similar to that of Wacker FZ wafers after the application of the anti-re#ection coating.Also,it has been very clear that the PERT structure has considerably improved the cell open-circuit voltages,"ll factors and e$ciencies over the standard PERL structure on SEH FZ substrates.However,a similar PERT structure when applied to Wacker FZ wafers only resulted in lower open-circuit voltages and cell e$ciencies.There is a clear di!erence between Wacker FZ wafers and SEH FZ and MCZ wafers in this regard. The PERL cells on SEH CZ(B)substrates only gave a modest cell performance.It is possible that the PERT approach may also improve the performances of these cells.5.Stability of cell performanceStability problems had been previously reported for cells made on boron-doped CZ substrates[5,6].It has been suggested that boron reacted with oxygen to reduce the carrier lifetime in the substrate[6].Hence,reducing the substrate oxygen content or reducing the substrate boron doping level is expected to reduce this stability problem. All the CZ(B)cells show considerable degradation after4d storage in a lowly illuminated area.The illumination level was far less than0.1sun.The degradation resulted in10mV lower open-circuit voltage,1mA/cm lower short-circuit current density,and0.8%lower absolute e$ciency,which is similar to"gures reported elsewhere[5,6].However,all the cells fabricated on other SEH materials,except these CZ(B) substrates,have shown a virtually stable performance under storage and testing.Due to the lack of a suitable testing station for the cell stability testing,only a few cells were illuminated for a few hours to check their stability.Many more cells were tested only after a few days of storage under the low illumination.No clear performance degrada-tion was observed for these cells.Although the Box3MCZ substrate with a high oxygen content showed a small carrier lifetime reduction after storage,as apparent in Table1,cell Ws8-3b on such a material showed a very stable performance compared to that of CZ(B)substrates. An early PERL cell Ws3-1e fabricated on the same high-oxygen MCZ substrates showed merely a1.2mV reduction in open circuit after5.3sun illumination for5h. The higher substrate resistivity of this cell is also expected to improve the cell stability. Hence,we cannot conclude whether it was the MCZ growth method or the higher resistivity which improved the stability of this particular cell.6.Conclusions24.5%e$ciency has been demonstrated by a PERT cell fabricated on an SEH MCZ silicon substrate.This is the highest e$ciency ever reported for a non-FZ silicon36J.Zhao et al./Solar Energy Materials&Solar Cells66(2001)27}36solar cell.The PERT cell structure improved not only the"ll factor but also the open-circuit voltage of these cells.This PERT cell structure also improved the performance of cells on SEH FZ substrates.The cells fabricated on CZ(Ga)substrates have also demonstrated reasonably high open-circuit voltages of676mV.Future experiments are expected to result in high cell performance on these CZ(Ga)substra-tes.All the cells fabricated on SEH substrates,except SEH CZ(B)substrates,have demonstrated stable performance.AcknowledgementsThe authors would like to acknowledge the contributions from other members of the Photovoltaics Special Research Centre,University of New South Wales.Many thanks to Dr.Takao Abe of SEH Co.for supplying a variety of high-quality MCZ wafers and other materials.The contributions of members of Sandia National Labor-atories in the area of cell characterisation are gratefully acknowledged,particularly those of Barry Hansen and James Gee.The Photovoltaics Special Research Centre was established and supported under the Australian Research Council's Special Research Centres Program.Dr.J.Zhao is also supported by an Australian Research Council's Australian Research Fellowship.References[1]J.Zhao,A.Wang,P.Altermatt,M.A.Green,Appl.Phys.Lett.66(1995)3636.[2]P.J.Verlinden,R.A.Sinton,K.Wickham,R.A.Crane,R.M.Swanson,Proceedings of the14thEuropean Photovoltaic Solar Energy Conference,Barcelona,1997,p.96.[3]S.W.Glunz,J.Knobloch,D.Biro,W.Wettling,Optimized high-e$ciency silicon solar cells withJsc"42mA/cm and "23.3%,Proceedings of the14th European Photovoltaic Solar Energy Conference,Barcelona,1997,p.392.[4]J.Zhao,A.Wang,M.A.Green,High e$ciency PERL and PERT silicon solar cells on FZ and MCZsubstrates,Technical Digest of the11th International Photovoltaics Science and Engineering Confer-ence,Sapporo,Japan,1999,p.557.[5]J.Knobloch,S.W.Glunz,D.Biro,W.Warta,E.Scha!er,W.Wettling,Proceedings of the25th IEEEPhotovoltaic Specialist Conference,Washington,DC,1996,p.405.[6]J.Schmidt,A.G.Aberle,R.Hezel,Proceedings of the26th IEEE Photovoltaic Specialist Conference,Anaheim,CA,1997,p.13.[7]J.Zhao,A.Wang,M.A.Green,Sol.Energy Mater.Sol.Cells32(1994)89.[8]A.Wang,Ph.D.Thesis,University of New South Wales,1992,p.111.[9]A.Wang,J.Zhao,S.R.Wenham,M.A.Green,Prog.Photovoltaics4(1996)55.。