RESEARCH ARTICLE
Nanoimprinted Tio2sol–gel passivating diffraction gratings for solar cell applications
Jérémy Barbé*,Andrew Francis Thomson,Er‐Chien Wang,Keith McIntosh and Kylie Catchpole The Australian National University,Center for Sustainable Energy Systems,Canberra,Australian Capital Territory,Australia
ABSTRACT
We report the fabrication and characterization of TiO2sol–gel diffraction gratings on silicon substrates by using nanoimprint lithography.The gratings are homogeneous and free of defects and cover an area of25cm2.Minority carrier lifetimes of up to900µs for imprinted samples under illumination are reported,which Kelvin probe measurements indicate is due to light‐generated negative charge in thefilms.The structures reported here are very promising as light‐trapping,pas-sivating coatings for solar cells.Copyright©2011John Wiley&Sons,Ltd.
KEYWORDS
TiO2sol–gel;diffraction grating;solar cells;nanoimprint;passivation;transient photoconductance decay
*Correspondence
Jérémy Barbé,The Australian National University,Center for Sustainable Energy Systems Canberra,Australian Capital Territory, Australia.
E‐mail:jeremy.barbe@hotmail
Received5January2011;Revised4March2011
1.INTRODUCTION
Reduction of surface losses is a major issue for the devel-opment of cheaper and more efficient solar cells.In partic-ular,the reduction of optical and recombination losses at the surface plays an increasingly important role in the de-velopment of thinner and cheaper solar cells.
Conventional silicon texturing,that is,the use of pyra-midal structures of<111>crystal planes at the surface of solar cells,is less effective in providing anti‐reflection and light trapping for multicrystalline silicon than for sin-gle crystal silicon due to the range of crystal orientations. In addition,it is not applicable to thin‐film cells because of the reduced thickness of material.In response to this problem,it has been shown that diffraction gratings can form a very effective structure as an alternative technol-o
gy for light trapping[1–5].
Nanoimprint lithography is a suitable high resolu-tion,large area and cheap technique to realise these pat-terned structures in the nanometer range[6–8]. Whitesides and co‐workers have achieved structures with feature sizes ranging down to30nm[9].Hampton et al. have shown excellent patterning of sub‐500nm inorganic oxide structures and,in particular,TiO2sol–gel on glass or aflash layer[10].Verschuuren and van Sprang have
realised high‐quality three‐dimensional sub‐micron structures over square centimetres by using sol–gel imprint lithography[11].Moreover,since the early1980s,TiO2 has shown to be an efficient anti‐reflection coating for sil-icon solar cells[12–14].Its high refractive index and low absorbance at wavelengths relevant to solar cells make it
a well‐suited material for light trapping as well as for
anti‐reflection.
The other major surfConventional silicon texturing,that is,the use of pyraace loss mechanism is the recombination of photogenerated charge carriers.This is particularly impor-tant at the front surface whe
re the majority of incident light is absorbed.Passivation is currently achieved in most industrial solar cells by using SiN x thinfilms,which also act as a sin-gle‐layer anti‐reflection coating[15].SiN x is deposited by relatively expensive plasma‐enhanced chemical vapour de-position,and the use of TiO2deposited by cheaper techniques could reduce costs drastically.TiO2can be de-posited by a number of techniques including spin‐coating of sol–gelfilms and atmospheric pressure chemical vapour deposition(APCVD).Thompson et al.have recently shown that APCVD TiO2exhibits a photo‐induced effect, which creates negative charges in the TiO2thinfilm when illuminated and strongly enhances the passivation of crys-talline silicon[16–18].
PROGRESS IN PHOTOVOLTAICS:RESEARCH AND APPLICATIONS
Prog.Photovolt:Res.Appl.(2011)
Published online in Wiley Online Library(wileyonlinelibrary).DOI:10.1002/pip.1131 Copyright©2011John Wiley&Sons,Ltd.
In this paper,we report the fabrication of TiO2diffrac-tion gratings on silicon substrates by using sol–gel imprint lithography and characterise their surface morphology and electrical properties.We show that defect‐free TiO2 sol–gel diffractions gratings can be achieved over areas of25cm2,which is the size of t
he nanoimprinting stamp used.These layers can provide excellent surface passiv-ation,as indicated by minority carrier lifetime measure-ments of up to900µs under illumination.The surface passivation is due to the presence of light‐generated nega-tive charge in the layers.
2.EXPERIMENTAL:TITANIA
SOL–GEL PREPARATION
Titania(TiO2)sol–gel was prepared following traditional sol–gel chemistry techniques[10].The precursor titanium (IV)butoxide,Ti(0CH2CH2CH2CH3)4,wasfirst mixed with acetylacetone in order to moderate its hydrolysis rate by chelation[19].After15min of stirring(120rpm),the solution was diluted with the solvent2‐propanol,and then glacial acetic acid was added dropwise to the stirring solu-tion as a catalyst.The solution was stirred for1h andfil-tered before use.The molar ratio of chemicals was optimised in order to obtain the desired viscosity and thick-ness.In particular,we noticed that a high amount of solvent was more appropriate for nanoimprints.The optimised mass ratio of titanium IV butoxide,Hacac,2‐propanol and glacial acetic acid that we used was10:6.8:30:1.
Thefiltered solution was then poured on a silicon sub-strate and spin‐coated following the recipe:resist dosing for10s at100rpm,resistflying for10s at1500rpm then 2s at1000rpm and pre‐drying for10s at300r
pm.To avoid a fast drying of the sol–gel while spinning because of the high amount of isopropanol,we used a cover adapted to the spin‐coater to prevent solvent removal before printing.
For the nanoimprinting,we used the soft stamp‘substrate conformal imprint lithography’method of Verschurren and van Sprang[11].The polydimethylsiloxan stamp was pat-terned with a square array of circular pillars with a diame-ter of343nm,a height of200nm and a pitch of513nm. The master stamp was provided by Philips and made by e‐beam lithography.After the spin‐coating,the cover was removed and the polydimethylsiloxan stamp quickly placed face‐down on the coated wafer.No high pressure was applied;stamp cavities were onlyfilled by capillarity‐fill processes,but it may be necessary to help this process by gently pressing withfingers.The TiO2sol–gel was printed and dried for2h by solvent removal through the stamp.The stamp can then be peeled off without damaging the nanoimprinted surface.
In this work,transient photoconductive decay was used to measure the effective minority lifetime at the surface of the semiconductor material.This technique described by Kane and Swanson[20]in1985is a contactless method,which involves the measurement of sheet photoconductivity over time[21,22].Kelvin Probe measurements were conducted to characterise both the magnitude and the polarity of the photoinduced charges at the surface of TiO2sol–gel on sil-icon.This technique used a vi
brating tip to measure the work function difference with the conducting sample[23].
3.RESULTS AND DISCUSSION
3.1.TiO2as a diffraction grating
Figure1shows four scanning electron microscope images of sol–gel TiO2features on silicon.The features are arrays of holes that are approximately50nm in depth and460nm in width.The thickness of the TiO2thin‐film is about 100nm.Samples are free of defects,the thickness is homo-geneous,and no cracks can be observed.The imprinted area is bigger than5×5cm square and is only limited by the size of the stamp that we used.The holes in titania are9.5%wider than the original stamp grating,which is due to shrinkage of the titania during the evaporation of the solvent.
To study the light‐trapping efficiency of such structures, we used the software G SOLVER4.20c(Grating Solver De-velopment Co.,Allen,TX,USA)to simulate the above grating.In parallel,the reflectance of this sample was mea-sured with a Cary5000(Varian Inc.,Palo Alto,California) spectrophotometer in the range400–1100nm.The simula-tion and measurement data are shown in Figure2.Between 500and1000nm,simulation and experimental resultsfit well.The shift in the infrared is due to the fact that we sim-ulate an infinite structure so that reflectance from the back surface is not taken into accoun
t.The weighted average reflectance calculated from the AM1.5G solar spectrum and spectrophotometric measurements is14.2%for the imprinted sample against36%for bare silicon.The grating is not optimised for anti‐reflection and light trapping,but is clearly promising in this regard.Future works will focus on the geometry of such structures to enhance light trapping.
3.2.TiO2as a passivation layer
In the second part of this work,we have studied the effec-tiveness of TiO2sol–gel as a passivation layer for crystal-line silicon.In the following experiments,we used5Ωcm, 700‐µm‐thick,<100>,one‐side‐polished,n‐type,flat zone silicon wafers.These wafers werefirst etched and RCA cleaned.One side was oxidised at1050°C for30min, whereas the other side was coated with non‐imprinted TiO2sol–gel as described earlier.It is necessary to passiv-ate one side with SiO2in order to reduce the effect of the surface on the minority carrier lifetime.The high quality of the wafers means that bulk recombination will have a negligible effect.At the end of the process,samples were baked for30min at275°C.
The effective minority lifetime of the wafers’surface was measured using transient photoconductive decay. Samples were measured both with and without a light bias with an intensity of0.5suns to observe the effect of Prog.Photovolt:Res.Appl.(2011)©2011John Wiley&Sons,Ltd.
DOI:10.1002/pip
illumination on the minority carrier lifetime.It was found that TiO 2layer passivates silicon much better under illumi-nation than in the dark.In Figure 3,the effective carrier lifetime at the surface coated with non ‐imprinted TiO 2sol –gel is plotted over time as the bias light is switched on and off.The lifetime is plotted for an excess carrier den-sity of 1015cm −3.
From this graph,we see that relatively high lifetimes can be reached in crystalline silicon coated with TiO 2sol –gel.Moreover,we observe a reversible light ‐induced effect which signi ficantly increases the carrier lifetime:
The sample has an initial lifetime of 140µs in the dark.Af-ter the light bias is switched on,the lifetime increases rap-idly in the first few seconds and then increases exponentially with a time constant of 70s,saturating at 700µs after 5min.Then,after the bias light is switched off,the lifetime decays to its initial value,also with a time constant of 70s.These results for sol –gel TiO 2are consistent with those for APCVD TiO 2[16–18].
Annealing is necessary to enhance the TiO 2passivation effect.Figure 4shows the effect of different annealing temperatures between 250°C and 400°C on the carrier lifetime for an excess carrier density
of 1015cm −3.In all cases,the samples were baked for 30min under nitrogen flow and the lifetime was measured soon after.We observe that the highest lifetime up to 900µs under illumination,is reached for a 250°C annealing.We also notice that the light ‐induced effect is reduced with increasing temper-atures and almost disappears after 300°C.If we measure the sample after 3days and more,the lifetime is somewhat deteriorated.However,after 6weeks,the lifetime under illumination is still high,with a decrease from 900to 400µs for imprinted and annealed samples.
As a first step towards our goal of producing light ‐trapping,passivating coatings,we have studied the passiv-ation ef ficiency of imprinted samples.The same grating as shown in Figure 1was used.Table I shows lifetime mea-surements for imprinted TiO 2before and after annealing.First,we observe that when imprinted samples are not annealed,light no longer induces any effect.However,af-ter baking at 200°C for 30min,light ‐induced surface passivation returns.These observations are not true
for
Figure 2.Simulation and re flectance measurements for TiO2
sol –gel on silicon patterned with a hole
grating.
Figure 1.Scanning electron microscope image of TiO2sol –gel on silicon patterned with a hole grating.
Prog.Photovolt:Res.Appl.(2011)©2011John Wiley &Sons,Ltd.DOI:10.1002/pip
non ‐imprinted samples for which light induces passiv-ation both before and after annealing.
Kelvin probe measurements were then performed to as-sess the effect of illumination on the charge associated with the TiO 2.Figure 5plots the measured voltage V KP as a function of time when the illumination is switched on and off for a sample annealed at 200°C.Because of dif fi-culties in the calibration,the absolute V KP is unknown,and we have offset V KP such that V KP ~0mV when the sample is in steady state under illumination.Although this offset does not affect the change in V KP due to the chang-ing light conditions,the relationship between V KP and illu-mination is not trivial because it depends on the location of the charge in the TiO 2,the interface ‐trapped charge at the TiO 2–Si interface,the degree of band ‐bending in the sili-con and the location of the electron and hole Fermi levels
characterisein the Si;furthermore,all of these parameters interrelate and have a complicated dependence on illumin
ation [24].Nevertheless,it is possible to construe some information about charge in the TiO 2from Figure 5.
Firstly,after the light is switched off in region 1of the graph,there is an immediate decrease in V KP of about 400mV.It arises because of (i)the change in the quasi ‐Fermi levels that occurs under illumination and,possibly,(ii)an increase in the band ‐bending caused by a negative charge in the film (as band ‐bending is signi ficantly greater at equilibrium than under illumination).Thus,from this ef-fect alone,one cannot determine whether there is a charge in the TiO 2(or at the TiO 2–Si interface),but one can con-clude that if there is a charge,it is not large and positive,which would have caused an increase in V KP when the light is switched off.
A second salient feature of Figure 5is the slow increase in V KP of ~300mV after the light is switched off in region 2.The increase in V KP is approximately exponential with a time constant of a similar magnitude (30s)to that of the decay in surface recombination (70s).Because the light remains off,there can be no change in the Fermi level;in-stead,the increase in V KP must come from a ‘downward ’movement of the energy bands relative to the Fermi level,as occurs if the charge in the TiO 2becomes less negative (more positive).We can conclude,therefore,that the TiO 2charge must be negative and decreasing with time,or else there would not be a corresponding reduction in lifetime (increase in surfa
ce recombination)in
the
Figure 3.Effect of bias illumination on the effective minority carrier lifetime for 5Ωcm,n ‐type crystalline silicon wafers coated with non ‐imprinted TiO 2sol –gel.The effective lifetime is determined for an excess carrier density of 1015cm −3
.
Figure 4.Effective minority carrier lifetime of TiO 2‐passivated silicon wafers as a function of annealing temperature with and without light bias for an excess carrier density of 1015cm −3.The samples were measured soon after the annealing process.
Table I.Effect of annealing (200°C for 30min)on the effective
minority carrier lifetime for nanoimprinted samples.
Dark
Light τeff before annealing (µs)122129τeff after annealing (µs)
276
902
Figure 5.Kelvin probe voltage versus time measured with a Kelvin probe for patterned TiO 2on silicon after an annealing at 200°C.
Prog.Photovolt:Res.Appl.(2011)©2011John Wiley &Sons,Ltd.
DOI:10.1002/pip
non‐diffused samples of Section3.2and,more particu-larly,in the boron‐diffused samples of[16–18].
Finally,it is evident in Figure5that when the light is switched back on,V KP increases by just150mV in region 3rather than the decrease of400mV in region1;this is consistent with the increase in V KP due to the change in Fermi levels being countered by an immediate decrease in V KP due to the introduction of negative charge.An im-mediate increase in charge passivation is apparent from the photoconductance data.Unlike the photoconductance signal,however,only a barely discernible reduction in V KP occurs over time when illuminated.
In summary,it can be concluded from the Kelvin probe measurements that the illumination introduces a negative charge into the TiO2(or at the TiO2interface) that dissipates over minutes when the illumination is extinguished.
4.CONCLUSION
In conclusion,we have produced titania sol–gel diffraction gratings on silicon substrates by using nanoimprint lithogra-phy.The patterned layers show promise as an anti‐reflection and light trapping stru
cture for solar cells and can provide very good passivation of crystalline silicon.A light‐induced effect creates negative charges in the TiO2layer,which strongly enhances the carrier lifetime at the surface of sili-con,with the sample lifetime being high as900µs for both non‐imprinted and imprinted samples. REFERENCES
1.Eisele C,Nebel CE,Stutzmann M.Periodic light cou-
pler gratings in amorphous thinfilm solar cells.Jour-nal of Applied Physics2001;89(12):7722–7726.
DOI:10.1063/1.1370996
2.Zeng L,Yi Y,Hong C,Liu J,Feng N,Duan X,
Kimerling LC,Alamariu BA.Efficiency enhancement in Si solar cells by textured photonic crystal back re-flector.Applied Physics Letters2006;89(11): 111111–111111‐3
3.Catchpole KR,Green MA.A conceptual model for
light‐coupling by pillar diffraction gratings.Journal of Applied Physics2007;101:063105–063105‐8.
DOI:10.1063/1.2710765.
4.Zhao L,Zuo YH,Zhou CL,Li HL,Diao HW,Wang
WJ.A highly efficient light‐trapping structure for thin‐film silicon solar cells.Solar Energy2010;84: 110–115.DOI:10.1016/j.solener.2009.10.014
5.Heine C,Morf RH.Submicrometer gratings for solar
energy applications.Applied Optics1995;34: 2476–2482.DOI:10.1364/AO.34.002476
6.Guo LJ.Nanoimprint lithography:methods and
material requirement.Advanced Materials2007;19: 495–513.DOI:10.1002/adma.200600882
7.Chou SY,Krauss PR,Zhang W,Guo L,Zhuang L.
Sub‐10nm imprint lithography and applications.Jour-nal of Vacuum.Science&Technology1997;15(6): 2897–2904.
8.Gates BD,Xu Q,Stewart M,Ryan D,Grant Wilson C,
Whitesides GM.New approaches to nanofabrication: molding,printing,and other techniques.Chemical
Reviews2005;105(4):1171–1196.DOI:10.1021/ cr030076o
9.Xia Y,Whitesides GM.Soft lithography.Angewandte
Chemie International Edition1998;37:550–575. 10.Hampton MJ,Williams SS,Zhou Z,Nunes J,Ko D,
Templeton JL et al.The patterning of sub‐500nm inorganic oxide structures.Advanced Materials2008;
20:2667–2673.DOI:10.1002/adma.200702495 11.Verschuuren M,van Sprang H.3D photonic structures
by sol–gel imprint lithography.Materials Research Society2007;1002:N03–N05.
12.Richards BS.Comparison of TiO2and other dielec-
tric coatings for buried‐contact solar cells:a review.
Progress in Photovoltaics:Research and Applications 2004;12:253–281.DOI:10.1002/pip.529
13.Hovel HJ.TiO2antireflection coatings by a low tem-
perature spray process.Journal Electrochemical Soci-ety:Solid‐State Science and Technology1978;125: 983–985.
14.Kern K,Tracy E.Titanium dioxide antireflection coat-
ing for silicon solar cells by spray deposition.RCA Re-view1980;4:133–180.DOI:10.1016/0038‐092X(83) 90133‐0
15.Dekkers HFW,Beaucarne G,Hiller M,CharifiH,
Slaoui A.Molecular hydrogen formation in hydroge-nated silicon nitride.Applied Physics Letters2006;
89:211914–211914‐3.DOI:10.1063/1.2396900 16.Thomson AF,Lynn SZ,McIntosh KR.Passivation of
silicon by negatively charged TiO2.35th European Photovoltaic Solar Energy Conference,Valencia, 6–10of September2010.
17.Thomson AF.Studies of silicon oxide ageing and tita-
nium oxide passivation of silicon for use in solar cells, PhD thesis2010,The Australian National University, Canberra.
18.Thomson AF,McIntosh KR.Light‐enhanced surface
passivation of TiO2‐coated silicon.Submitted to Prog-ress in Photovoltaics2010.
19.Schubert U,Husing N.Synthesis of Inorganic
Materials.Wiley‐VCH:Weinheim,2000.
20.Kane DE,Swanson RM.Measurement of the emitter
saturation current by a contactless photoconductivity decay method.Proceedings of the18th IEEE
Prog.Photovolt:Res.Appl.(2011)©2011John Wiley&Sons,Ltd. DOI:10.1002/pip
版权声明：本站内容均来自互联网，仅供演示用，请勿用于商业和其他非法用途。如果侵犯了您的权益请与我们联系QQ:729038198，我们将在24小时内删除。
发表评论