Optimization of a New ZnO Nanorods Hydrothermal Synthesis

Optimization of a New ZnO Nanorods Hydrothermal Synthesis Method for Solid State Dye Sensitized Solar Cells Applications Laurent Schlur,?Anne Carton,?Patrick Lévêque,?Daniel Guillon,?and Genevie v e Pourroy*,?

?Institut de Physique et Chimie des Mate r iaux de Strasbourg IPCMS,UMR7504,CNRS-Universite de Strasbourg-ECPM,23rue du Loess BP43,67034Strasbourg cedex2,France

?Institut d’E l ectronique du Solide et des Syste m es InESS,UMR7163,CNRS-Universite de Strasbourg,23rue du Loess BP20,67037 Strasbourg cedex2,France

ratio equal to1.74and a pH of8.2.The growth was performed at

surface,arising from zinc hydroxyacetate decomposition during the

(ranging from1to3μm)on solar cell e?ciency was tested.

length increases,the best photovoltaic parameters were measured

The use of quasi-in?nite solar energy through the development

of photovoltaic conversion is one of the most promising

solutions within the global context of increasing demand for stable,renewable,and sustainable energy sources.Besides the

conventional silicon-based solar cells and the promising organic

cells,dye-sensitized solar cells(DSSCs)incorporating both

inorganic and organic semiconductor materials give rise to an

increasing interest.1?4

The key components of a DSSC are a photoanode made of a ?lm of semiconductor nanoparticles(generally TiO2),coated onto a transparent conducting oxide substrate(usually?uorine-doped tin oxide,FTO),a dye,and a liquid electrolyte,which was initially based on the iodide/triiodide couple.5The nanoparticles?lm provides a large internal surface area for the chromophore anchoring to maximize light absorption.The photoelectrons generated by the dye excitation are injected into the semiconductor and undertake a random walk through the nanoparticles network.They may interact with a distribution of traps before reaching the collecting electrode.

A promising strategy for improving electron transport in

DSSCs is to replace the nanoparticles photoelectrode by a

single-crystalline nanorods photoelectrode.Electrons can then

be conducted within a nanorod instead of by multiple-

scattering transport within the nanoparticles network.6ZnO is considered as one of the most promising materials to grow up nanorods photoelectrode in DSSCs.Indeed,ZnO has a wide band gap(ca.3.37eV),its conduction band edge is found to be very close to that of TiO2(ca.?4.4eV),and ZnO presents a higher bulk electron mobility(200?300cm2V?1s?1)than

TiO2(0.1cm2V?1s?1).7The electron lifetime in ZnO is increased compared to TiO2with reduced recombination

losses.8Furthermore,numerous hierarchical ZnO structures have been fabricated,and large arrays of single-crystalline

nanorods were obtained at low temperatures(about90°C)on several substrates such as glass or silica.9?11They were

generally obtained by reaction of zinc nitrate hexahydrate (Zn(NO3)2·6H2O)in methenamine(hexamethylenetetramine, C6H12N4).9,12,13However,the geometries have to be improved and controlled to reveal the real potential applications of this material.Thus,this paper focuses on the control of ZnO seeds

and nanorods elaboration by a low-cost process using zinc acetate dihydrate and ethylenediamine as reactants in order to get the best photovoltaic properties.We have previously proved that the use of these reactants allows the growth of single-crystalline ZnO nanorods,having very few structural defects.14 The e?ects of ZnO morphology on the cell e?ciency are tested in TCO/ZnO/dye/spiro-OMeTAD/gold devices.Our goal is to estimate in what extent the precise control of ZnO elaboration parameters will allow the power conversion e?ciency to be improved.The dye is the organic D102dye often used in solid state hybrid solar cells15and the solid holes

Received:June13,2012

Revised:January11,2013

Published:January16,2013

conductor,the spiro-OMeTAD.16,17The traditional liquid electrolytes used in DSSC18,19are corrosive20and require careful packaging procedures which are not necessary with a solid hole transporter.

2.EXPERIMENTAL SECTION

Zinc acetate dihydrate(Zn(CH3COO)2·2H2O,98%)and tert-butyl alcohol were purchased from Alfa Aesar,ethylenediamine (C2H4(NH2)2,≥99%),tert-butylpyridine(tBP),and chlor-obenzene were from Sigma-Aldrich,anhydrous ethanol and acetonitrile were from Carlo Erba,spiro-OMeTAD(2,2′,7,7′-tetrakis(N,N-di-p-methoxyphenylamine)9,9′-spirobi?uorene) was from Merck KGaA,and lithium tri?uoromethyl sulfonyli-mide(Li-TFSI)was from Tokyo Chemical Industry(TCI). Indium tin oxide(ITO)(≤20Ω/square)and F-doped SnO2 (FTO)(7Ω/square)coated glass substrates were purchased from Precision Glass Optics and Solaronix,respectively.Their rms roughnesses are2.7and36.7nm,respectively.The D102 dye was purchased from Mitsubishi Paper Mills Limited Japan.

2.1.ZnO Seeds Layer on Substrates.A solution of zinc acetate dihydrate(Zn(CH3COO)2·2H2O)5×10?3M in anhydrous ethanol was prepared and stirred in a glovebag?lled with N2gas.The solution was deposited on the transparent conducting oxide(TCO)coated glass substrates according to the method developed by Greene et al.21

The substrates(2×1cm)were?rst cleaned under sonication successively in acetone and ethanol,then washed in water,and?nally dried in an oven at50°C for15min.The substrates were oxygen plasma treated for5min in order to improve their wettability and to eliminate last organic traces before the zinc acetate solution deposition.22Then,the substrates were placed in the glovebag under N2gas.40μL of the prepared solution was deposited on the substrate. Thanks to the surface treatment,the drop covers all the substrate.After the solvent evaporation,the substrate was annealed at400°C for20min in air in order to transform zinc acetate into zinc oxide.The deposition of the solution and annealing steps were repeated four times again to get a su?cient ZnO layer thickness.

2.2.ZnO Hydrothermal Growth.ZnO nanorods were obtained after submitting the previous substrate to a hydro-thermal treatment.In a typical procedure,an aqueous solution of10mL of ethylenediamine(C2H4(NH2)2,20vol%)and24 mL of Zn(CH3COO)2·2H2O(0.72M)was prepared and then poured into a homemade Te?on-lined stainless autoclave(V= 110mL);i.e.,zinc acetate dihydrate and ethylenediamine concentrations in the autoclave are equal to C ref Zn=5.05×10?1 mol/L and C ref En=8.8×10?1mol/L,respectively.The seed-coated substrate was suspended horizontally upside-down in the autoclave.The later is sealed and placed in an oven at110°C for2h.Then,it was immersed for10min in a water bath in order to cool it.The substrate was?nally rinsed with distilled water and dried in an oven at50°C for30min. Temperature,reaction time,ethylenediamine,and zinc acetate dihydrate concentrations were varied in order to describe the synthesis mechanism and to determine the most adapted parameters to the realization of the solar cell.The temperature was varied from65to130°C and the reaction time from30min up to18h,and the concentrations of reactants were reduced and increased.

2.3.Preparation of the Cells.Solid-state DSSCs were built using the following procedure.Substrates covered with nanorods were immersed for1h at50°C in D102dye(0.5mM)dissolved in1:1acetonitrile/tert-butyl alcohol and rinsed in anhydrous acetonitrile.The molecular glass2,2′,7,7′-tetrakis(N,N-di-p-methoxyphenylamine)-9,9′-spirobi?uorene (spiro-OMeTAD)was dissolved in chlorobenzene at di?erent concentrations depending on the length of ZnO nanorods.For nanorods length of1,2,and3μm,concentrations were210, 300,and370mg/mL,respectively.tert-Butylpyridine(tbp)and lithium tri?uoromethyl sulfonylimide(Li-TFSI)were used as spiro-OMeTAD dopants:tbp was added to the spiro-OMeTAD solution(3.2:40w/w tbp/spiro-OMeTAD).Li-TFSI,previ-ously dissolved in acetonitrile at170mg/mL,was added to the hole transporting material(1.3:40w/w Li-TFSI/spiro-OMe-TAD).Then70μL of spiro-OMeTAD was deposited on the top of the substrate before spin-coating at a speed of1500rpm for a nanorods length of1μm and2000rpm for2or3μm nanorods length.Finally,a120nm thick gold electrode was evaporated on top of the cell under low vacuum pressure,in order to complete the device with two0.16cm2diodes.

2.4.Characterization.The seeds layer and nanorods morphologies were observed by scanning electron microscopy (SEM)using a JEOL electron microscope6700.XRD measurements were performed on a Siemens D5000X-ray di?raction system in re?ection mode with a quartz mono-chromator(Cu Kα1=1.54056?).The roughness of the seeds layer was determined by atomic force microscopy(AFM)using a Digital Instrument3100microscope coupled to a Nanoscope IIIa recorder.Measurements were done in the tapping mode. TG and DT analysis were performed using a TA Instruments apparatus SDT Q600with a heating rate of2°C/min in air. The absorbance of the solar cells was measured with a PerkinElmer Lambda950spectrophotometer.

The?lling fraction of spiro-OMeTAD between the nanorods was determined by using the method developed by I-Kang Ding et al.23The?lling fraction is calculated by dividing the volume of spiro-OMeTAD by the free volume between the rods.The volume of spiro-OMeTAD and ZnO nanorods has been determined by absorbance measurements and by ICP-AES,respectively.The presence of the dye has been neglected. The total volume(ZnO nanorods and vacuum between the rods)has been calculated by knowing the surface of the substrate covered by the rods and their length,which has been determined by SEM images.

The current?voltage(I?V)characteristics of the solar cells were measured using a Keithley2400source measurement unit. The Oriel150W AM1.5G solar simulator using a?ltered Xe lamp was calibrated before each measurement with a THORLABS optical power meter to obtain standard100 mW/cm2illumination conditions.

3.RESULTS AND DISCUSSION

The morphology of ZnO nanorods,the distribution of the rods on the whole substrate surface,and the?atness of the upper surface,i.e.,the homogeneity of the rods length and the lack of sediments,are very signi?cant for the quality and the e?ciencies of solar cells.Indeed,the hole conductor has to be homogeneously spread out inside and over the ZnO layer. The thickness of spiro-OMeTAD at the top of the rods must be homogeneous and as thin as possible.ZnO nanorods elaboration occurs in two steps:the seeds deposition and annealing on the substrate and the nanorods growth from the seeds in hydrothermal conditions.Both steps must be accurately controlled to get homogeneous distribution of nanorods on all the substrate and homogeneous nanorods

lengths.Reactants concentrations,growth time,and temper-ature were varied in order to have several rod lengths and consequently several solar cells thicknesses.3.1.Step 1:The ZnO Seeds Layer.The quality of the

seeds layer is critical to obtain homogeneous and oriented nanorods but also to prevent any contact between the transparent conducting glass and the dye/spiro-OMeTAD.Seeds have to be spread out on the whole surface,and the layer must be homogeneous after the solvent evaporation and the annealing at 400°C.The solution is prepared by mixing zinc acetate dihydrate with ethanol and may involve adsorbed water.24As ethanol evaporates faster than water,24,25droplets of water/zinc acetate dihydrate may be formed on the substrate,leading to an inhomogeneous seeds repartition after annealing.This hypothesis has been checked by comparing seeds layers obtained after addition of water to the solution (1vol %)and obtained in a dried N 2atmosphere (Figure 1).In the ?rst case,the seeds are not regularly distributed on the surface.Zinc acetate dihydrate is concentrated in the droplets,leading to the formation of uneven aggregates after the annealing.Similar behaviors are observed when a solution of ethanol/zinc acetate dihydrate without water was deposited on the substrate in ambient air due to the adsorption of water from the external environment.24Under N 2gas,the thickness of zinc oxide seeds layer is homogeneous and was estimated to be 55±5nm (Figure 1b,c).A rms roughness of 2.45±0.2nm has also been determined thanks to AFM measurements on seeds formed on ITO substrate (?gure not shown).These two values indicate that the ZnO seeds layer will prevent the contact between the transparent conducting oxide on the substrate and the dye/spiro-OMeTAD.The di ?raction pattern recorded on the seeds layer exhibits the (0002)di ?raction peak of the hexagonal wurtzite (JCPDS Card No.[36-1451]),showing that the (000l)planes are parallel to the substrate surface (Figure 2).This orientation will allow ZnO nanorods to grow perpendicularly to the substrate:the bottom plane of ZnO nanorods is (000?1)considering the fact that the (000?1)plane of the rods will grow on the (0001)plane of the seeds.Although the (000l)planes have a high surface energy,they are largely developed.According to Greene et al.,21it can be due to acetates which can adsorb onto nascent (0001)planes,to a decrease of their surface energy for very thin seeds,or to a di ?erent con ?guration of the ?rst few atomic layers that can convert into (000l)orientation thanks to a minor structural trans-formation. 3.2.Step 2:ZnO Nanorods Morphology According to

the Hydrothermal Reaction Parameters.Let us consider

the evolution of ZnO nanorods when the zinc acetate dihydrate volume V Zn (0.72M)is varied while the ethylenediamine volume (20vol %)is kept equal to 10mL.For these samples,the hydrothermal growth was performed at 110°C for 2h.The morphologies of ZnO either on the substrate or in the solution are presented in Figure 3.Below 18mL,no ZnO is recovered from the solution by centrifugation,and no nanorods growth

is

Figure 1.SEM images of ZnO seeds layer on an ITO substrate,dried under N 2gas (glovebag)and

calcination at 400°C:(a)top view when 1vol %of water is added in the seed solution,(b)top view,and (c)

cross section without additional

water.Figure 2.X-ray di ?raction pattern of ZnO seeds layer on an ITO substrate,dried under N 2gas (glovebag),and air calcined at 400°

C.Figure

3.SEM images of ZnO nanorods grown in the solution and on the substrate using 10mL of ethylenediamine (20vol %)solution with added

zinc acetate dihydrate (0.72M)volume of (a)18,(b)19,

(c)

20,(d)21,(e)22,(f)23,and (g)24mL corresponding to initial pH

of 9.95,9.80,9.55,9.25,8.80,8.55,and 8.20,respectively (each scale bar corresponds to 1μm).

observed on the substrate.The ZnO seeds layer deposited on the TCO surface was totally dissolved.When V Zn ≥18mL,i.e.,when ethylenediamine concentration in the autoclave [En]divided by zinc acetate dihydrate concentration in the autoclave [Zn]is less than or equal to 2.32,nanorods are observed on the substrate,but their length decreases as V Zn increases from 19to 24mL.Except for 18mL,ZnO precipitate is present in the solution and drop down on the nanorod layer.As V Zn increases,the ZnO quantity recovered in the solution increases according to the variation presented in Figure 4.Zinc oxide rods are observed in the solution for 19mL ≤V Zn ≤21mL and rods

and brushes for 21mL ≤V Zn ≤24mL,showing that two di ?erent phenomena are involved.[En]/[Zn]=2.32(V Zn =18mL)is a limit for nanorods growth as no zinc oxide is visible in solution,and nanorods length and density are lower than for V Zn =19mL.It corresponds certainly to the limit between formation and dissolution of ZnO,in other words,to a pH at which zincate species are soluble at 110°C.As explained in previous studies,several chemical equili-briums are involved in the formation of ZnO nanorods.When mixing the reactants at room temperature,the majority of Zn 2+cations are chelated by ethylenediamine.26,27Shapnik et al.28,29proved that ethylenediamine and Zn 2+cations do not form only bis-and tris-ethylenediamine complexes (g =0and r =2or 3in eq 1)but also protonated ethylenediamine zinc complexes (g ≠0).The ethylenediamine hydrolysis provides OH ?anions in the solution according to eqs 2and 3.During the hydrothermal treatment,zinc ethylenediamine complex formed in eq 1partly decomposes,resulting in the increase of ethylenediamine,protonated ethlyenediamine,and Zn 2+in solution (eq 4).The presence of OH ?anions (eqs 1,2,and 3)and the condensation reaction (eq 5)result in the formation of ZnO either as nanorods on the substrate or in solution.+++?++++?g r g g Zn ()NH (CH )NH H O [Zn(NH (CH )NH )(NH (CH )NH )]OH g r g 22222222232222(2)(1)+?++?NH (CH )NH 2H O

[NH (CH )NH ]2OH 2222232232(2)+?+++?NH

(CH )NH H O

[NH (CH )NH ]OH 2223232232(3)

?++++++

r g [Zn(NH (CH )NH )(NH (CH )NH )]Zn NH (CH )NH NH (CH )NH g r g 22232222(2)222222223(4)

+?++?Zn 2OH ZnO H O 22(5)

For V Zn >21mL,a precipitate is formed immediately when ethylenediamine and zinc acetate dihydrate are mixed together.The precipitate was recovered by centrifugation and washed three times in water.It was shown to be isomorphous to zinc hydroxyacetate (Zn 5(OH)8(CH 3COO)2·n H 2O)(Figure 5,JCPDS Card No.[56-0569]),and the pattern was indexed in the hexagonal system.30

Thanks to the elementary analysis and TG analysis,the formula of the precipitate was determined to be Zn 5(OH)7.6(CH 3COO)2.4·3.7H 2O.It forms according to eq 6and decomposes into zinc oxide (eq 7)when heated.Similarly to previous studies showing that forced hydrolysis on cloudlike phase Zn 3(OH)8(NO

3)2·2H 2O allowed ZnO nanowhisker ?lms or twinlike ZnO nanoarrays assembly to crystallize,the decomposition of zinc hydroxyacetate provides nucleation sites which lead to ZnO brushes according to eqs 1?5.31,32However,part of hydroxyacetate can be dissolved during heating participating to the nanorods growth.

+++?·+??5Zn 7.6OH 2.4CH COO 3.7H O Zn (OH)(CH COO) 3.7H O 23257.63 2.42(6)·→++Zn (OH)(CH COO) 3.7H O 5ZnO

2.4CH COOH 6.3H O 57.63 2.4232(7)The decrease of nanorods length on the substrate and the increasing mass of ZnO in suspension observed when V

Zn

Figure 4.

Evolution

of

zinc

oxide mass present

in the hydrothermal

solution after reaction with the volume of zinc acetate dihydrate (0.72

M)added to 10mL of ethylenediamine (20vol

%).Figure

5.XRD pattern of the Zn 5(OH)7.6(CH 3COO)2.4·3.7H 2O precipitate formed immediately after mixing ethylenediamine and zinc acetate dihydrate.

increases appear to be inconsistent.That means that a competition takes place between ZnO nucleation and growth in the solution on one side and growth on the substrate on the other side.The zinc concentration increase seems to favor the growth of ZnO in solution rather than the growth of nanorods on the substrate.For 19mL ≤V Zn ≤21mL,zinc cations were released in solution according to eq 4.According to La Mer ’s theory,33when the supersaturation reaches the critical super-saturation,the formation of stable nuclei in solution begins,and when the supersaturation becomes lower than the critical supersaturation,the nucleation stops and the growth starts.Increasing Zn 2+amounts are available for the nucleation and the growth of ZnO in solution.In addition,for V Zn >21mL,ZnO crystallizes by decomposition of zinc hydroxyacetate.Therefore,growth of rods on the substrate and growth of ZnO brushes from nuclei in solution occur simultaneously leading to smaller nanorods.The most regular ZnO nanorods layer and ?attest upper surface was observed when V Zn =24mL (Figure 3),i.e.,[En]/[Zn]=1.74and pH =8.2.In that case,[En]=C ref En =8.8×10?1mol/L and [Zn]=C ref Zn =5.05×10?1mol/L.For V Zn >24mL,the length decrease is too high.As explained previously,the ?atness of the surface is signi ?cant for the photovoltaic devices e ?ciency,so the e ?ect of matter concentration,reaction temperature,and time on nanorods length,nanorods diameter,and ZnO in suspension (rods or brushes)were studied starting from these conditions.The concentrations of zinc acetate dihydrate and etylenedi-amine were varied in the same range;i.e.,[En]/[Zn]was kept constant and equal to 1.74,and the total volume was equal to 34mL.[Zn]to C ref

Zn ratios varied between 0.25and 2.0.The pH of the starting mixture varies between 8.1and 8.4.Figure 6a represents the evolution of ZnO nanorods length and diameter after the hydrothermal reaction at 110°C for 2h for several [Zn]/C ref Zn ratios ([En]/[Zn]=1.74).The length of the rods increases when the reactants concentration increases,the maximal length 8.08±0.27μm is obtained for a ratio of 2.0,and the minimal length is 3.09±0.25μm for a ratio of 0.25.Diameters around 110±40nm are generally encountered except for the 0.25ratio,for which the diameter is equal to 77±16nm.The nanorods length and diameter variations versus the reaction time are illustrated in Figure 6b for a reaction at 110°C.In the ?rst 90min of hydrothermal growth,the nanorods length increases linearly versus the time and reaches up to 4.77±0.20μm at 90min.For times longer than 90min,the length increase is slowed down.This behavior has been already observed by Wang et al.34At the beginning of the reaction,the growth was rapid as a lot of matter (Zn 2+,OH ?)was available.When the equilibrium is reached,the growth stops.The nanorods diameter increases during the ?rst 60min and stabilizes around 102nm for a longer reaction time.For a very long growth time,the diameter seems to increase slightly up to 131±26nm.For a reaction time of 30min,the rods length was 0.13±0.03μm.They are very short,so it can be supposed that for a growth time of 30min (approximately the time the autoclave needs to reach the reaction temperature)the rods have only begun to grow.This can also explain the fact that the diameter is smaller than for other reaction times.The reaction time was so short that the rods have not the time to reach a diameter around 102nm,which seems to be

the

Figure 6.Evolution of ZnO nanorods length and diameter as a function of (a)[Zn]/C ref

Zn ratio with [En]to [Zn]ratio ?xed at 1.74,(b)the growth time,(c)the synthesis temperature for a growth time of 2h,and (d)the synthesis temperature for a growth time of 18h.

equilibrium diameter.The e ?ect of temperature synthesis on the rods length and diameter is illustrated in Figures 6c and 6d for two reaction times of 2and 18h,respectively.A quasi-linear relationship between the length of ZnO nanorods and the synthesis temperature for a reaction time of 2h is observed.The rods length varies from 1.09±0.10μm at 65°C to 7.33±0.49μm at 130°C.For a longer reaction time (18h),the rods length is around 7μm for every synthesis temperature.Therefore,the growth rate depends on the temperature.For a lower synthesis temperature,the growth rate is smaller.In both cases,the rods diameter does not change signi ?cantly with an average diameter of 106and 125nm for 2and 18h,respectively.However,ZnO brushes are observed on the surface for a temperature of 110°C while they are absent when the reaction is performed at 65

°C (Figure 7).After 2h at 65°C,zinc hydroxyacetate is not decomposed yet into ZnO,and it

does not ?x on the surface of ZnO nanorods (Figure 7b).3.3.Solar Cells Characteristics Using Optimized ZnO Nanorods Growth.The e ?ect of ZnO morphology on solar cell e ?ciency was tested for various nanorods length.According to the previous observations,[En]/[Zn]=1.74was chosen since the upper surface of nanorods layer is ?at for every matter concentration reaction temperature and reaction time.The reaction was performed at 65°C for 2h in order to avoid the zinc hydroxyacetate decomposition into ZnO and the deposition of rods and brushes on top of the nanorods.[Zn]/C ref Zn ratio =0.5was chosen in order to get a nanorods length of exactly 1.0μm.The reaction was repeated on the same substrate with a new hydrothermal solution 2or 3times in order to get 2.0or 3.0μm rods,respectively (Figure 8).The photovoltaic characteristics of these solar cells were measured two times,namely immediately after their fabrication and one month later.Between the two measurements,the devices were stored without encapsulation at room temperature in the dark.Nanorods of 3μm length and covered with brushes were also obtained

with a growth temperature of 80°C with [Zn]to C

ref Zn =1.0(called S3b ).Figure 9exhibits the current ?voltage curve for devices which were not aged,built with nanorods of 1μm (S1),2μm (S2),3μm (S3),and 3μm with brushes (S3b ).The photovoltaic parameters are tabulated in Table 1.An important decrease of the photovoltaic parameters is

observed when impurities cover the surface (S3b ).Figure 10shows that ZnO brushes pass over the blocking spiro-OMeTAD layer.The presence of brushes is responsible of the formation of cracks in the spiro-OMeTAD when chlorobenzene evaporates.Nano short circuits occur between ZnO brushes or rods (below the cracks)and the gold electrode.These short circuits induce a strong increase of the leakage current (100-fold higher in the case of S3b compared to the S3counterpart),which can be at the origin of the slight decrease of

the shunt resistance observed when brushes are present.

Furthermore,a dramatic reduction of the open circuit voltage (V OC )is also observed for the S3b device.This strong V OC reduction could be due to enhance charge carrier recombina-tion at the surface in S3b devices.

The S1sample exhibits the best photovoltaic characteristics with an open circuit voltage (V OC )of 557mV,a short circuit current density (J SC )of 2.0mA/cm 2,and a ?ll factor (FF)of 0.38leading to an overall power conversion e ?ciency (PCE)of 0.42%.For the S2and S3samples,the V OC value is similar to S1.On the other hand,J SC ,FF,and consequently the PCE decrease when the nanorods length increases.Looking at the UV ?vis absorbance spectra,we can see that the absorbance increases when the rods length varies from 1to 3μm as shown in Figure 11.An absorbance increase is expected to lead to a current increase contrary to what is observed.Several hypotheses can be put forward to explain this behavior.First,it has already been observed that for solid state dye sensitized

solar cells made of nanoporous electrodes the pore ?lling fraction is getting smaller when the device thickness increases.35,36Consequently,the power conversion e ?

ciency

Figure 7.SEM tilted view (85°)of the

top of the substrate for a reaction time of 2h and a synthesis temperature of (a)110°C and (b)65°

C.Figure 8.SEM images of ZnO nanorods grown by an hydrothermal treatment of 2h at 65°C with a ratio of 1.0.These growth conditions have been repeated:(a)once,(b)twice,and (c)three

times.Figure 9.J ?V curve of solid state solar cells having nanorods length of 1μm (black),2μm (red),3μm (blue),and 3μ

m with brushes

(green)(light intensity 100mW/cm 2

,AM 1.5G illumination conditions).These devices were not aged.

decrease observed when the device thickness increases can be at least partially attributed to the pore ?lling fraction evolution.In our case,for samples S1,S2,and S3,the ?lling fraction between the rods are almost equal and close to 100%(Table 1).Therefore,the decrease of the power conversion e ?ciency is not linked to the pore ?lling fraction.Interestingly,the modi ?cation of the semiconductor oxide morphology from a nanoporous structure to a nanorod structure can increase the value of the ?lling fraction which was estimated to be 60?65%for a TiO 2nanoporous structure of 2.8μm thick by Ding et al.23Second,the increase of the spiro-OMeTAD overlayer thickness (Table 1)with the increasing ZnO nanorods length can also be partially responsible of the decrease of the J SC ,FF,and PCE.37The problem is that the spiro-OMeTAD overlayer thickness could not be decreased because of the increasing length inhomogeneity (Table 1)due to the repetition of hydrothermal syntheses at 65°C.Third,increasing the nanorods length (S1,S2,and S3)leads to a slight increase of the series resistance (calculated from the J ?V curve in the dark)that is not su ?cient to explain the measured decrease of both J SC and FF when the nanorods length increases.On the other hand,the hole mobility in doped spiro-OMeTAD is around 10?3cm 2V ?1s ?138while the electron mobility in ZnO is more than 3-fold higher,so that the charge carrier mobility is highly unbalanced.This may be the

main reason for the observed ine ?cient charge extraction (the highest ?ll factor measured for S1diodes is as low as 0.38).10,39

As the nanorods length is increased,the charge extraction limitations become more and more critical,and consequently,both J SC and FF decrease when the nanorods length increases.We strongly believe that the third hypothesis is the main explanation of the photovoltaic parameters evolution as a function of the nanorods length even though the second one cannot be completely ruled out.All these results show that despite a more direct pathway for the charges and a better ?lling fraction for a nanorods structure compared to a nanoporous one,the thickness of the solar cells cannot be increased more than for the nanoporous structure for which the best thickness is around 2μm (for TiO 2particles).The aging e ?ect on the J ?V response of the solar cells has also been tested.Figure 12shows the evolution of the

Table 1.Photovoltaic Parameters of Solid State Solar Cells Sensitized with D102as a Function of the Nanorods Length,Growth Conditions,Annealing Temperature,and Solar Cell Ageing a

sample ZnO nanorods length (μm)spiro-OMeTAD overlayer thickness (μm)J SC (mA/cm 2)V OC (mV)FF PCE (%)?lling fraction (%)1month aging presence of brushes S1 1.04±0.080.22±0.04 2.05570.380.4297no no S2 2.01±0.130.35±0.09 1.85460.330.3396no no S3 3.22±0.200.64±0.11 1.45460.320.2595no no S3b 3.04±0.110.72±0.27b 1.22230.370.10b no yes S1old 1.04±0.080.22±0.04 2.46040.430.6197yes no S2old 2.01±0.130.35±0.09 2.06150.320.4096yes no S3old 3.22±0.200.64±0.11 1.96260.300.3595yes no a

The relative PCE uncertainty estimated from diode to diode variations is below 15%.b Spiro-OMeTAD thickness determined far away from the

brushes.The ?lling fraction was therefore not

calculated.Figure 10.SEM top view of the S3b solar

cell.Figure 11.UV ?vis absorption spectra of ZnO nanorods sensitized with D102and in ?ltrated by spiro-OMeTAD.The rods have a length

of 1μm (black),2μm (red),and 3μm

(blue).Figure 12.J ?V curve of solid state solar cells stored in the dark without encapsulation for 1month and having nanorods length of 1

μm (black),2μm (red),and 3μm (blue)(light intensity 100mW/cm 2,AM 1.5G illumination conditions).

photovoltaic characteristics of the solar cells after an aging time of1month for nanorods having a length of1μm(S1old),2μm(S2old),and3μm(S3old).After aging,a V OC and a J SC increase are visible,compared to the samples without aging and having the same nanorods length.The S1old sample exhibit the best photovoltaic performances with an open circuit potential (V OC)of604mV,a short circuit current density(J SC)of2.4 mA/cm2,a?ll factor(FF)of0.43,and a power conversion e?ciency(PCE)of0.61%.To our knowledge,a PCE of0.61% is the highest reported to date for(TCO/ZnO nanorods/dye/ spiro-OMeTAD/metal electrode)devices.For the same device architecture,Plank et al.reported an e?ciency of0.25%.The PCE increase after aging has already been observed for hybrid solar cells including a ZnO or TiO2layers.13,40,16This increase has not been explained yet.Two di?erent hypotheses are often suggested.First,the aging can improve the contact between spiro-MeOTAD and gold.Second,ZnO can spontaneously lose oxygen at the surface when it is in contact with some organic material or when the cell is placed under vacuum(as for the top electrode evaporation).The ZnO surface is then electron-rich, converting the oxide into an electron donor,and disrupting the device.Therefore,an air exposure(like during the aging)allows the ZnO surface to be regenerated.

4.CONCLUSION

The elaboration of the oxide is one of the key parameters for increasing the hybrid solar cells e?ciency.We have pointed out the conditions for having a homogeneous ZnO seeds layer regularly spread out on the whole substrate surface.Nanorods perpendicular to the substrate surface have to be grown at65°C for2h with an ethylenediamine to zinc acetate dihydrate molar ratio equal to1.74and a pH of8.2to avoid impurities deposition and to obtain a regular upper surface.As a consequence of elaboration control,the power conversion e?ciency has been increased in TCO/ZnO nanorods/dye/ spiro-OMeTAD/metal electrode devices to reach an overall power conversion e?ciency of0.61%.The e?orts have now to be carried on the hole conductor to further improve the

photovoltaic parameters.

■AUTHOR INFORMATION

Corresponding Author

*E-mail pourroy@ipcms.unistra.fr.

Author Contributions

The manuscript was written through contributions of all authors.All authors have given their approval to the?nal version of the manuscript.

Notes

The authors declare no competing?nancial interest.■ACKNOWLEDGMENTS

We acknowledge the?nancial support of CNRS and Region

Alsace.

■REFERENCES

(1)Weickert,J.;Dunbar,R.B.;Hesse,H.C.;Wiedemann,W.; Schmidt-Mende,L.Adv.Mater.2011,23,1810?1828.

(2)Boucle,J.;Ackermann,J.Polym.Int.2012,61,355?373.

(3)Hagfeldt,A.;Boschloo,G.;Sun,L.;Kloo,L.;Pettersson,H. Chem.Rev.2010,110,6595?6663.

(4)Hardin,B.E.;Snaith,H.J.;McGehee,M.D.Nat.Photonics2012, 6,162?169.

(5)O’Regan,B.;Gra t zel,M.Nature1991,353,737?740.

(6)Wang,M.;Bai,J.;Le Formal,F.;Moon,S.-J.;Cevey-Ha,L.; Humphry-Baker,R.;Gra t zel,C.;Zakeeruddin,S.M.;Gra t zel,M.J. Phys.Chem.C2012,116,3266?3273.

(7)Zhang,Q.;Dandeneau,C.S.;Zhou,X.;Cao,C.Adv.Mater. 2009,21,4087?4108.

(8)Quintana,M.;Edvinsson,T.;Hagfeldt,A.;Boschloo,G.J.Phys. Chem.C2007,111,1035?1041.

(9)Vayssieres,L.;Keis,K.;Lindquist,S.-E.;Hagfeldt,A.J.Phys. Chem.B2001,105,3350?3352.

(10)Law,M.;Greene,L.E.;Johnson,J.C.;Saykally,R.;Yang,P. Nat.Mater.2005,4,455?459.

(11)Baruah,S.;Dutta,J.Sci.Technol.Adv.Mater.2009,10,0130001.

(12)Baruah,S.;Dutta,J.J.Cryst.Growth2009,311,2549?2554.

(13)Plank,N.O.V.;Howard,I.;Rao,A.;Wilson,M.W.B.;Ducati,

C.;Mane,R.S.;Bendall,J.S.;Louca,R.R.M.;Greenham,N.C.; Miura,H.;et al.J.Phys.Chem.C2009,113,18515?18552.

(14)Kovalenko,A.;Pourroy,G.;Cre g ut,O.;Gallart,M.;Honerlage,

B.;Gilliot,P.J.Phys.Chem.C2010,114,9498?9502.

(15)Schmidt-Mende,L.;Bach,U.;Humphry-Baker,R.;Horiuchi,T.; Miura,H.;Ito,S.;Uchida,S.;Gra t zel,M.Adv.Mater.2005,17,813?815.

(16)Burschka,J.;Dualeh,A.;Kessler,F.;Baranoff,E.;Cervey-Ha, N.-L.;Yi,C.;Nazeeuddin,M.K.;Gra t zel,M.J.Am.Chem.Soc.2011, 133,18042?18045.

(17)Boucharef,M.;Di Bin,C.;Boumaza,M.S.;Colas,M.;Snaith,H. J.;Ratier,B.;Boucle,J.Nanotechnology2010,21,205203.

(18)Nazeeruddin,M.K.;De Angelis,F.;Fantacci,S.;Selloni,A.; Viscardi,G.;Liska,P.;Ito,S.;Takeru,B.;Gra t zel,M.J.Am.Chem.Soc. 2005,127,16835?168347.

(19)Yella,A.;Lee,H.-W.;Tsao,H.N.;Yi,C.;Chandiran,A.K.; Nazeeruddin,M.K.;Diau,E.W.-G.;Yeh,C.-Y.;Zakeeruddin,S.M.; Gra t zel,M.Science2011,334,629?634.

(20)Snaith,H.J.;Schmidt-Mende,L.Adv.Mater.2007,19,3187?3200.

(21)Greene,L.E.;Law,M.;Tan,D.H.;Montano,M.;Goldberger, J.;Somorjai,G.;Yang,P.Nano Lett.2005,5,1231?1236.

(22)Lu,D.;Yu,Y.;Guo,J.;Lu,G.;Wang,Y.;Shen,J.Mater.Sci.Eng., B2003,97,141?144.

(23)Ding,I.-K.;Te t reault,N.;Brillet,J.;Hardin,B.E.;Smith,E.H.; Rosenthal,S.J.;Sauvage,F.;Gra t zel,M.;McGehee,M.D.Adv.Funct. Mater.2009,19,2431?2436.

(24)Innocenzi,P.;Malfatti,L.;Costacurta,S.;Kidchob,T.;Piccinini, M.;Marcelli,A.J.Phys.Chem.A2008,112,6512?6516.

(25)Shi,L.;Shen,P.;Zhang,D.;Lin,Q.;Jiang,Q.Surf.Interface Anal.2009,41,951?955.

(26)Gao,X.;Li,X.;Yu,W.J.Phys.Chem.B2005,109,1155?1161.

(27)Carlson,G.A.;McReynolds,J.P.;Verhoek,F.H.J.Am.Chem. Soc.1945,67,1334?1339.

(28)Shapnik,M.S.;Safina,L.R.;Bagautdinov,R.U.;Petrova,T.P. Koord.Khim.1982,8,1076?1079.

(29)Starodubets,E.E.;Borisevich,S.V.;Shapnik,M.S.Russ.J.Inorg. Chem.2009,54,453?458.

(30)Hosono,E.;Fujihara,S.;Kimura,T.;Imai,H.J.Colloid Interface Sci.2004,272,391?398.

(31)Hu,X.;Masuda,Y.;Ohji,T.;Kato,K.Thin Solid Films2009, 518,621?624.

(32)Hu,X.;Masuda,Y.;Ohji,T.;Kato,K.Cryst.Growth Des.2009, 5,3598?3602.

(33)LaMer,V.K.;Dinegar,R.H.J.Am.Chem.Soc.1950,72,4847?4854.

(34)Wang,S.F.;Tseng,T.-Y.;Wang,Y.-R.;Wang,C.-Y.;Lu,H.-C.; Shih,W.-L.Int.J.Appl.Ceram.Technol.2008,5,419?429.

(35)Snaith,H.J.;Humphry-Baker,R.;Chen,P.;Cesar,I.; Zakeeruddin,S.M.;Gra t zel,M.Nanotechnology2008,19,424003.

(36)Schmidt-Mende,L.;Gra t zel,M.Thin Solid Films2006,500, 296?301.

(37)Kim,H.-S.;Lee,C.-R;Jang,I.-H.;Kang,W.;Park,N.-G.Bull. Korean Chem.Soc.2012,33,670?674.

(38)Snaith,H.J.;Gra t zel,M.Appl.Phys.Lett.2006,89,262114/1?262114/3.

(39)Park,W.I.;Kim,J.S.;Yi,G.-C.;Bae,M.H.;Lee,H.-J.Appl. Phys.Lett.2004,85,5052?5054.

(40)Greene,L.E.;Law,M.;Yuhas,B.D.;Yang,P.J.Phys.Chem.C 2007,111,18451?18456.