ed Single-Crystal Rutile TiO2 Nanowires
Solar Cells DOI:10.1002/anie.201108076 Rapid Charge Transport in Dye-Sensitized Solar Cells Made from Vertically Aligned Single-Crystal Rutile TiO2Nanowires**
Xinjian Feng,Kai Zhu,Arthur J.Frank,Craig A.Grimes,and Thomas E.Mallouk*
Over the past two decades,dye-and quantum dot-sensitized solar cells and polymer/inorganic hybrid solar cells have emerged as promising alternatives to solid-state p-n junction cells for addressing the urgent need for inexpensive,efficient solar power.[1]At the heart of these solar cells is a mesoporous TiO2nanoparticle(NP)electrode,which not only provides a high surface area for accommodating the light-absorbing sensitizer but also serves as the stable conductor for photo-generated electrons.Fast charge transport in the network is essential for effective charge collection,particularly in solid-state cells in which recombination is very fast.[2]To increase the electron mobility,TiO21D polycrystalline nanotube arrays[3]and single-crystalline nanowire(NW)arrays[4,5] have recently been proposed and studied as electrode materials for solar cells.However,there is little experimental evidence of substantially enhanced electron transport in these TiO21D nanostructure-based solar cells.For example,there is no substantial difference in electron transport between1D TiO2nanotube and NP-based dye-sensitized solar cells (DSSCs).[3c]
Herein,we report a ketone–HCl solvothermal process for the growth of single-crystal rutile TiO2NW arrays on F-doped tin oxide(FTO)substrates.The TiO2NWs begin to grow within15min and lengths up to10m m can be obtained within https://www.sodocs.net/doc/d4710402.html,ing these NW films,we show for the first time that the electron diffusion coefficient of single-crystal rutile TiO2NWs is more than two orders of magnitude higher than that of rutile NP films at the same photoelectron density.In light of the findings reported here,arrays of1D single-crystal rutile TiO2are attractive for solar cell and other optoelectronic applications.
Figure1a,b show,respectively,cross-sectional field-emis-sion scanning electron microscope(FE-SEM;JEOL6300) images taken at low and high magnification,which show arrays of uniform and well-separated NWs.NWs grow vertically from the substrate to an average length of about 1.6m m and a diameter of40nm during a reaction time of 30min.From the grazing angle X-ray diffraction(GAXRD) patterns(Figure S1in the Supporting Information),the crystal phase of the thin base layer and nanowire arrays were identified as tetragonal rutile(JCPDS file number21-1276).High-resolution transmission electron microscope (HR-TEM)images(Figure1c)confirm that the NWs are single crystalline and have a(110)interplanar distance of 0.327nm.The NWs grow with a preferred[001]orientation. The length of NWs increases almost linearly with the reaction time.For example,growth of15min results in around260nm long NWs and a two hour reaction gives around9.6m m long NWs(see Figure S2in the Supporting Information).The NW length can also be adjusted by using different amounts of precursor.For example,for a reaction temperature of2008C and a duration of45min NWs with lengths of around1.1and around4.4m m can be obtained by using0.4and0.8mL of tetrabutyl titanate,respectively(see Figure S3in the Support-ing
Information).
Figure1.a and b)FE-SEM cross-sectional images of NW arrays on FTO-coated glass substrate at low and high magnifications,respec-tively.c)HR-TEM image of the as-synthesized single NW.
[*]Dr.X.Feng,[+]Prof.C.A.Grimes,Prof.T.E.Mallouk
Department of Chemistry,Materials Research Institute
The Pennsylvania State University
University Park,PA16802(USA)
E-mail:tem5@https://www.sodocs.net/doc/d4710402.html,
Dr.K.Zhu,[+]Dr.A.J.Frank
National Renewable Energy Laboratory
1617Cole Boulevard,Golden,CO80401(USA)
[+]These authors contributed equally to this work.
[**]Work at Penn State was supported by the U.S.Department of Energy,Office of Science,Office of Basic Energy Sciences under
grant number DE-SC0001087.Work at NREL was supported by the U.S.Department of Energy,under grant number DEAC36-
08GO28308.The Penn State Nanofabrication facility is supported by the National Science Foundation under grant number ECS-
0335765.We would like to thank Dr.Bangzhi Liu at the Penn State Nanofabrication facility for his help with FE-SEM,TEM,and HR-
TEM
analyses.
Supporting information for this article is available on the WWW
under https://www.sodocs.net/doc/d4710402.html,/10.1002/anie.201108076.
2727 Angew.Chem.Int.Ed.2012,51,2727–2730 2012Wiley-VCH Verlag GmbH&Co.KGaA,Weinheim
For most transition-metal oxides,the challenge of con-trolling crystal orientation and texture arises from the fast hydrolysis and condensation of metal–organic precursors.To control the morphology,strongly acidic media are frequently required,which in turn lead to slow reaction rates.Con-sequently,reaction times longer than 20h are usually required.[4]We observe fast growth of NWs using n -butanone as solvent.Ketones have previously been used in the non-aqueous sol–gel synthesis of NPs,acting both as the solvent and as a source of oxygen atoms.[6]During the growth of TiO 2nanocrystals,Cl àions selectively adsorb onto the (110)crystal plane,suppressing further growth of that plane,and resulting in the oriented growth of NWs along the [001]direction.[7]
Figure 2a compares the photoelectron density (n )dependence of the electron diffusion coefficient (D )for DSSCs based on rutile NW and NP films.The values of D and
n were determined using procedures described in detail elsewhere.[8a,b]The nonlinear dependence of D on the photo-electron density is commonly observed and often ascribed to electrons undergoing multiple trapping–detrapping events within an exponential distribution of conduction band tail states.[8c]The diffusion coefficient in the single-crystal rutile NW-based DSSC is over 200times higher than that of the rutile NP-based cell [9]at the same photoelectron density (e.g.,at 1 1017cm à3).Over the past two decades,little attention has been paid to rutile TiO 2nanocrystalline films in DSSCs.The main reason that rutile TiO 2has been considered
uninteresting for DSSCs is its much lower electron diffusion coefficient than the anatase polymorph in both single-crystalline and polycrystalline forms.[10]For rutile nanoparti-cles in DSSCs,the transport rate is,in part,limited by the low number of interparticle connections.[10c]The dramatically increased electron transport rate for vertical rutile NWs relative to rutile NPs is presumably associated with the lack of grain boundaries along the oriented NWs.The lower density surface defects of the oriented NWs plays a role.
Figure 2b shows the dependence of the photoelectron density n on the voltage for rutile NW-and NP-based DSSCs.The voltage dependence of n (or chemical capacitance)is usually used as a measure of the distribution of sub-bandgap defect states of the electrode materials.[11]The same (parallel)dependence of n on the voltage for these two samples suggests that the shape of the distribution of sub-bandgap states does not depend on the electrode morphology.The photoelectron density in the single-crystal rutile NW-based DSSC is about four times smaller than that in the rutile NP-based cell at the same voltage,indicating that the total density of sub-bandgap defect states in the NW sample is also about four times less than that in the NP sample,which partially accounts for the observed transport properties discussed above.
Figure 3compares the recombination times (t r )for DSSCs based on rutile NWs and NPs as a function of the photoelectron density.At a given photoelectron density (e.g.,
1 1018cm à3)the lifetime in rutile NW-based cells is one order of magnitude shorter than that of rutile NP-based cells.The former can be understood in terms of the electron transport properties shown in Figure 2a.The higher the diffusion coefficient,the faster the photogenerated electrons can reach the anode contact,but faster diffusion also increases the recombination rate of electrons with species at the surface of the particles (e.g.,oxidized dye or iodine species),leading to a shorter electron lifetime.The charge collection property of DSSCs can be measured by the electron diffusion length (L n ),which is determined by L n =(D t r )1/2.A longer L n usually leads to a higher charge-collection
efficiency
Figure https://www.sodocs.net/doc/d4710402.html,parison of a)electron diffusion coefficients (D )as a func-tion of the photoelectron density (n )and b)the photoelectron density as a function of the voltage for rutile NW-and NP-based DSSCs with a laser illumination at 680
nm.
Figure https://www.sodocs.net/doc/d4710402.html,parison of recombination lifetimes for rutile NW-and NP-based DSSCs as a function of the photoelectron density with laser illumination at 680nm.
https://www.sodocs.net/doc/d4710402.html,
2012Wiley-VCH Verlag GmbH &Co.KGaA,Weinheim
Angew.Chem.Int.Ed.2012,51,2727–2730
(h cc).To ensure the optimum cell performance is obtained,L n is usually several times(e.g.,more than three times)greater than the film thickness.[11]Analyses of the data in Figures2a and3yield approximately L n=60m m for the NW-based DSSC and only13m m for the NP-based DSSC.Thus,the optimum film thickness would be about20and4m m for the NW-and NP-based cells,respectively.Thus,the device performance of NW-based cells can be increased by simply increasing the NW film thickness(up to20m m)without negatively affecting h cc.In contrast,extending the NP film thickness much beyond4m m would significantly reduce h cc and thus lower the device performance.
Figure4compares the J–V characteristics of NW-and NP-based DSSCs.The film thickness was4.5m m for both cells. The DSSC based on a single-crystal rutile NW array showed an open-circuit voltage(V oc)of0.78V,a fill factor(FF)of 0.68,and a short-circuit photocurrent density(J sc)of 6.95mA cmà2to give a solar conversion efficiency of3.68%. In contrast,the DSSC based on a rutile NP film showed a V oc of0.67V,a FF of0.64,and a J sc of8.70mAà2to yield an efficiency of3.74%.The20%less J sc value for the NW-based cell is primarily resulting from its much less surface area than the NP-based cell.However,as the electron diffusion length for cells using NW arrays(60m m)is more than a factor of four longer than that for cells using NP films(13m m)and is much longer than the film thickness(4.5m m)used in this study,we expect that J sc for NW-based cells can be substan-tially enhanced by increasing the length of NW arrays without lowering their charge-collection efficiencies.The larger FF value for the NW-based DSSC(relative to the NP-based DSSC)is presumably associated with the much increased(by more than two orders of magnitude)transport rate.The NW-based DSSC showed a110mV higher V oc than the NP-based DSSC,in agreement with the observed difference in the density of sub-bandgap states between NW-and NP-based cells(Figure2b).Thus,in comparison to polycrystalline rutile NPs,the single-crystal rutile NW architecture could give higher solar conversion efficiency and is attractive for DSSC applications.
In conclusion,we have developed a ketone-based solvo-thermal synthesis for single-crystal rutile NWs as vertically oriented arrays on FTO substrates,and for the first time we find that the electronic transport properties of the rutile phase are dramatically improved in DSSCs.In addition,we find that the density of sub-bandgap defect states in the NW-based DSSC is significantly lower than that in the NP-based DSSC. In light of these results along with a good stability and high refractive index of rutile TiO2,constructing aligned arrays of 1D single-crystal rutile TiO2is an attractive approach for solar cell and other optoelectronic applications.The data reported here also suggest that it may be possible,by using a proper texture and crystal orientation,to significantly improve the electrical transport properties of other oxide materials that currently have limited utility in solar cell and related applications because of their low carrier mobility. Experimental Section
A FTO-coated glass(TEC-8)was cleaned by sonication in acetone,2-propanol,and methanol.After cleaning,a20nm thick TiO2layer was deposited by dip-coating(0.3m tetrabutyltitanate in ethanol)fol-lowed by30min at5008C in air.The substrates were then loaded into a sealed Teflon-lined stainless steel reactor(23mL volume),contain-ing a mixture of n-butanone(6mL),37%hydrochloric acid in water (6mL),and different amounts of tetrabutyl titanate.The reactor was then heated at2008C for15min to2h.The NW samples were then washed with ethanol,dried in air,dipped in a H2O2(30wt%)/NH4OH (25wt%;v:v of10/1)solution for10min,rinsed with water and heated in air at4008C for30min.For use in DSSCs,NWs array samples were coated with dye by immersion overnight at ambient temperature in a0.5m m ethanolic solution of commercial N719dye. The electrolyte was composed of0.8m1-hexyl-2,3-dimethylimidazo-lium iodide and50m m iodine in methoxypropionitrile.A conductive glass slide sputter-coated with100nm of Pt was used as the counter-electrode.The thickness of the electrolyte layer between the NW array and counter-electrode was fixed by the use of a25m m thick SX-1170spacer(Solaronix).The photocurrent density and photovoltage of the DSSCs were measured with active sample areas of0.24–0.5cm2 using AM-1.5simulated sunlight produced by a500W Oriel Solar Simulator.Electron transport and recombination properties of DSSCs were measured by intensity-modulated photocurrent and photovoltage spectroscopies as described previously.[8]
Received:November16,2011
Published online:February2,2012
.Keywords:diffusion coefficients·metal oxides·nanoparticles·nanowires·solar cells
[1]a)B.O Regan,M.Gr?tzel,Nature1991,353,737;b)J.Boucl?,P.
Ravirajan,J.Nelson,J.Mater.Chem.2007,17,3141;c)I.Mora-Seró,J.Bisquert,J.Phys.Chem.Lett.2010,1,3046;d)D.A.R.
Barkhouse,R.Debnath,I.J.Kramer,D.Zhitomirsky,A.G.
Pattartyus-Abrahan,L.Levina,L.Etgar,M.Gr?tzel, E.H.
Sargent,Adv.Mater.2011,23,3134.
[2]a)N.Cai,S.J.Moon,L.Cevey-Ha,T.Moehl,R.Humphry-
Baker,P.Wang,S.M.Zakeeruddin,M.Gr?tzel,Nano Lett.2011, 11,1452;b)S.R.Jang,K.Zhu,M.J.Ko,K.Kim,K.C.Kim, N.G.Park,A.J.Frank,ACS Nano2011,5,8267.
[3]a)D.Gong,C.A.Grimes,O.K.Varghese,W.C.Hu,R.S.Singh,
Z.Chen,E.C.Dickey,J.Mater.Res.2001,16,3331;b)J.M.
Macμk,H.Tsuchiya,P.Schmuki,Angew.Chem.2005,117,2136;
Figure4.J–V characteristics of DSSCs based on a4.5m m long NW
array and a NP film under simulated AM1.5light(J=photocurrent
density).
2729 Angew.Chem.Int.Ed.2012,51,2727–2730 2012Wiley-VCH Verlag GmbH&Co.KGaA,Weinheim https://www.sodocs.net/doc/d4710402.html,
Angew.Chem.Int.Ed.2005,44,2100;c)K.Zhu,N.R.Neale,A.
Miedaner,A.J.Frank,Nano Lett.2007,7,69.
[4]a)X.J.Feng,K.Shankar,O.K.Varghese,M.Paulose,T.J.
Latempa,C.A.Grimes,Nano Lett.2008,8,3781;b)B.Liu,E.S.
Aydil,J.Am.Chem.Soc.2009,131,3985;c)Z.J.Zhou,J.Q.Fan, X.Wang,W.H.Zhou,Z.L.Du,S.X.Wu,ACS Appl.Mater.
Interfaces2011,3,4349.
[5]a)T.Kasuga,M.Hiramatsu,A.Hoson,T.Sekino,K.Niihara,
Adv.Mater.1999,11,1307;b)C.C.TsaI,H.Teng,Chem.Mater.
2006,18,367;c)J.E.Boercker,E.Enache-Pommer,E.S.Aydil, Nanotechnology2008,19,095604.
[6]M.Niederberger,G.Garnweitner,Chem.Eur.J.2006,12,7282.
[7]M.Adachi,Y.Murata,J.Takao,J.T.Jiu,M.Sakamoto,F.M.
Wang,J.Am.Chem.Soc.2004,126,14943.
[8]a)J.van de Lagemaat,N.G.Park,A.J.Frank,J.Phys.Chem.B
2000,104,2044;b)K.Zhu,N.Kopidaks,N.R.Neale,J.
van de Lagemaat, A.J.Frank,J.Phys.Chem.B2006,110, 25174;c)A.J.Frank,N.Kopidakis,J.van de Lagemaat,Coord.
Chem.Rev.2004,248,1165.
[9]N.G.Park,G.Schlichthorl,J.van de Lagemaat,H.M.Cheong,
A.Mascarenhas,A.J.Frank,J.Phys.Chem.B1999,103,3308.
[10]a)H.O.Finklea in Semiconductor Electrodes,Studies in Physical
and Theoretical Chemistry,Vol.2,Elsevier,Amsterdam,1988, p.3;b)L.Forro,O.Chauvet,D.Emin,L.Zuppiroli,H.Berger,
F.L?vy,J.Appl.Phys.1994,75,633;c)N.
G.Park,J.van de La-
gemaat,A.J.Frank,J.Phys.Chem.B2000,104,8989;d)S.
Kambe,S.Nakade,Y.Wada,T.Kitamura,S.Yanagida,J.Mater.
Chem.2002,12,723.
[11]L.M.Peter,J.Phys.Chem.C2007,111,6601.
https://www.sodocs.net/doc/d4710402.html, 2012Wiley-VCH Verlag GmbH&Co.KGaA,Weinheim Angew.Chem.Int.Ed.2012,51,2727–2730