High-Efficiency Dye-Sensitized Solar Cells Based on the Composite Photoanodes of SnO2

High-Efficiency Dye-Sensitized Solar Cells Based on the Composite Photoanodes of...

(Parte 1 de 4)

High-Efficiency Dye-Sensitized Solar Cells Based on the Composite Photoanodes of SnO2 Nanoparticles/ZnO Nanotetrapods†

Wei Chen,‡ Yongcai Qiu,‡ Yongchun Zhong,§ Kam Sing Wong,§ and Shihe Yang*,‡

Departments of Chemistry and Physics, The Hong Kong UniVersity of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, China

ReceiVed: September 10, 2009; ReVised Manuscript ReceiVed: NoVember 9, 2009

We have deviseddye-sensitizedsolar cells (DSSCs)with >6% efficiencyby employingcompositephotoanodes of SnO2 nanoparticles/ZnO nanotetrapods. Benefiting from material advantages of both constituents, the composite photoanodes exhibit extremely large roughness factors, good charge collection, and tunable light scattering properties. Among the three composite photoanodes with widely differing compositions tested, the best performance (efficiency ) 6.31%) was obtained with a weight ratio of SnO2/ZnO 2:1 mainly due to the highest saturated Jsc achieved at a thinnest film thickness. Charge collection losses in composite films with more ZnO nanotetrapods content and thus necessarily larger film thicknesses appear to be a main limiting factor on IPCE and therefore Jsc, which undermines the gain from their favorable light scattering ability. An ultrathin layer of ZnO spontaneously shelled on SnO2 nanoparticles is found to enhance Voc primarily by lifting the band edges rather than by suppressing recombination. Finally, by intensity modulated photocurrent/ photovoltage spectroscopy (IMPS/IMVS), we have identified that recombination in SnO2/ZnO composite films is mainly determinedby the ZnO shell conditionon SnO2, whereas electron transportis greatly influenced by the morphologies and sizes of the ZnO crystalline additives. In particular, ZnO nanotetrapods have proved to be superior in electron transport and therefore charge collection over ZnO particles additives in the SnO2/ ZnO composite-based DSSCs.

1. Introduction

Dye-sensitized solar cell (DSSC), which accomplishes efficient solar-to-electric power conversion by virtue of nanostructures,1-3 is regarded to be a potentially low cost alternative to traditional silicon-based photovoltaics. A certified DSSC efficiency record has been pushed to >1% with a maximum IPCE up to ∼85%4,5 because of the introduction of a highly porous TiO2 nanoparticles photoanode by O’Regan and Gratzel in 1991.1 Many other semiconductors besides TiO2 have also been tested as photoanode materials, and some of them, for example, SnO2-ZnO composite (15 nm SnO2 nanoparticles/1 µm ZnO particles 1:1 by weight),6-8 SnO2-MgO core-shell nanoparticles,9 Zn2SnO4,10 and so on, have achieved promising performance.6-13 In parallel to the exploration of new photo- anode materials, novel 1D nanostructures,1,14-18 including ZnO nanowires19,20/nanorods21,2 and TiO2 nanorods23,24/nanotubes arrays,25,26 have also attracted serious attention. Although their overall performance is still far from satisfactory as photoanodes because of various practical problems, they are generally superior in certain aspects of efficient DSSCs. Examples include facilitated electron transport in 1D materials such as perpen- dicularly aligned ZnO nanorods arrays21,2 and TiO2 nanotubes/ nanowiresarrays25 as well as compositesof nanowires/nanotubes and nanoparticles.27,28

BecausethephotoanodesofDSSCsarenanostructuredbynature foreffectivelyferryingelectronsandholes,it is of greatimportance to explore new materials with different architectures on the nanoscale. This cannot only enhance our understanding of theinherentworkingprinciplesof DSSCsbutalsohelpto improve the cell performance,29,30 especiallyin emergingfrontierssuch as flexibleDSSCs31 andsolid-stateDSSCs.32 Mostrecently,ourgroup reporteda DSSCefficiencyof 3.2%basedon a networkstructured photoanode assembled from building blocks of ZnO nanotetrapods.3 Such kind of photoanode,albeitpreparedwithoutcalcination,has demonstratedincrediblylong effectiveelectrondiffusion lengths.Thissuccessencouragedus to blendappositenanoparticles intothenanotetrapodsto improvefurtherthecellperformanceand to elucidatetheworkingmechanismof thecompositephotoanodes in the hope to realize ultimately practicableand highly efficient flexible DSSCs. We had in mind that the ZnO nanotetrapods network acts as a global electron transport highway, while the nanoparticlesenclosed in and attached to the network serve to increase the roughness factor and take part in the local charge transport.27,28,30

Our surveyexperimentshave shown that blendingZnO, TiO2, or SnO2 nanoparticles with ZnO nanotetrapods generally leads to greatly different performance features of the corresponding

DSSCs. It is believedthat a proper combinationof the electronic structuresof the constituents,for example,conductionband edge positions, which for TiO2 is close to that of ZnO but ∼500 mV higher than that of SnO2,34 distribution of accepting states in the conduction bands, which for SnO2 and ZnO consist of s orbital contrary to d orbital for TiO2,30 is critical to the performance of composite photoanodes. In practice, different surface chemistry of the constituents has to be duly considered.

In this work, SnO2 nanoparticles were chosen on the basis of two considerations.First, good cell performancerecords have already been achieved for SnO2/ZnO composite photoanodes but with very different nanostructures.6-8 It is tempting to see whether a similar or even higher cell efficiency can be achieved by blending SnO2 nanoparticles with ZnO nanotetrapods.

† Part of the “Benoît Soep Festschrift”. * Corresponding author. E-mail: chsyang@ust.hk. ‡ Department of Chemistry. § Department of Physics.

J. Phys. Chem. A X, x, 0 A

10.1021/jp908747z X American Chemical Society

Second, the precise functions of each component in the composite film and their synergy are yet to be elucidated.7,35 This study aims to provide guiding principles for the search of efficient photoanodes based on binary or even ternary nanomaterials systems.

Indeed, our DSSCs based on the composite photoanodes of

SnO2 nanoparticles/ZnO nanotetrapods have reaped efficiencies comparable to or higher than the preceding record of SnO2/ ZnO composite photoanodes. The maximum photocurrent densities that can be extracted from the composite photoanodes vary only slightly when the ratio of SnO2 nanoparticles to ZnO nanotetrapods is changed in a wide range, manifesting the outstanding charge transport property of the composite photoanodes. The structuralforms and the functionalroles of the ZnO component in the composite films have been carefully studied by the IMVS and IMPS methods, lending direct evidence of the origins of the Voc enhancement and the Jsc variation of the DSSCs arising from the nanoscale blending. Moreover, ZnO nanotetrapods have proved to be excellent in electron transport (charge collection) in the composite photoanodes in comparison to particle forms of ZnO additives.

2. Experimental Section

2.1. Synthesis of SnO2 Nanoparticles and ZnO Nanotet- rapods. SnO2 ultrafine nanoparticles were synthesized via a modified hydrothermal method from literature.10 In a typical synthesis, 6 g SnCl4·5H2O was dissolvedin 80 mL of deionized water at 353 K. To this hot solution, diethylenetriamine was added dropwise under vigorous stirring until its pH reached 9.2. After the stirring was maintained at this temperature for 2 h, the initiallyformedwhiteprecipitateswere completelydissolved, and a nearly transparent, light-yellow sol was obtained. The sol was transferred to a 100 mL Teflon-lined stainless steel autoclave, which was then heated to 453 K and kept at this temperature for 48 h. A dark-yellow gel was formed with a thin liquid layer on the top. The SnO2 product was separated by repeated centrifugation and washing with excessive ethanol, stored in residual ethanol with the solid content of 40 wt % for later use. Details of the controlled synthesis of ZnO nanotetrapods by the metal vapor transport-oxidation method can be found elsewhere.3,36 2.2. Fabrication of Composite Films. For the preparation of pastes with three different SnO2/ZnO weight ratios of 2:1,

1:1, and 1:2, 1, 0.5, and 0.25 g SnO2 wet products (solid content ) 40 wt %) were, respectively, mixed with 0.2 g ZnO nanotetrapods powder. To each mixture was added 6 mL of an ethanol/H2O (5:1, volume ratio) solvent mixture to obtain a sol. After ultrasonic treatment of the sols for 10 min, 0.015 g PEO

2 0 0 and 0.1 g block copolymerPluronicP123 were added with thorough stirring, forming homogeneous viscous pastes.

Pure SnO2 paste was prepared in the same way. Doctor blade technique was employed to spread the as-prepared pastes onto conductive glass substrates (FTO-coated glass, 14 Ω/square, Nippon sheet glass, Japan). The films were calcined at 450 °C in air for 30 min and then treated in 0.03 M ammonia for 30 min. After being soaked in excessive deionized water and ethanol, they were finally calcined at 450 °C for 30 min for a second time. The ammonia treatment is critical to the formation of a homogeneous ultrathin ZnO shell on SnO2 nanoparticles (Supporting Information, Figure S1), which, as will be revealed below, plays an important role in determining the charge transport/recombinationkineticsin compositephotoanode-based DSSCs. For the sake of comparative studies, some films were treated with a solution of acetic acid/H2O/ethanol 1:3:6 by volume for 5 or 30 min, which are hereafter referred to as HAc5 and HAc30, respectively. 2.3. General Materials Characterization.Morphologies of the nanomaterialswere directlyexaminedby SEM usinga JEOL 6700F at an accelerating voltage of 5 kV. TEM observations were carried out on a JEOL 2010F microscope operating at 200 kV. BET surface areas and pore size distributions of the film samples were characterized using a Coulter SA 3100 surface area analyzer. The samples for BET characterization were obtained by scratching the as-prepared films off the conductive glass substrates.XPS measurementsof the film sampleson glass slides were performedwith PhysicalElectronicssurface analysis equipment(model PHI 5600). XRD of the film sampleson glass slides was characterized using a Philips high resolution X-ray diffractionsystem(modelPW1825).Diffusedreflectancespectra were carried out on the same film samples using a Perkin-Elmer UV/vis spectrophotometer (model Lambda 20). The film thickness was determined by a Tencor Alpha-Step 200 surface profiler system. 2.4. Solar Cell Assemblyand PhotoelectrochemicalCharacterization. The calcined films were immersed in a mixed acetonitrile/tert-butylalcoholsolution(1:1 in volume)containing 5 × 10-4 M N719 dye (cis-bis(isothiocyanato)bis(2,2′-bipyridyl- 4,4′-dicarboxylato)-ruthenium(I)bis-tetrabutylammonium,Suzhou Chemsolarism,China)for 4 h. These dye-coatedelectrodeswere assembled to make solar cells with Pt-sputtered FTO counter electrodes and acetonitrile/valeronitrile (85:15, volume ratio) electrolyte containing 0.6 M 1,2-dimethyl-3-propylimidazolium iodide,0.03 M I2, 0.1 M GuSCN,and 0.5 M 4-tert-butylpyridine. We determinedthe adsorbeddye amountby immersingthe films in 1 mM NaOH solution (water/ethanol, 1:1, volume ratio) and monitoring the desorbed dye concentration by UV-vis spectroscopy. J-V characteristic curves were recorded using a CHI 600A electrochemical analyzer. The light source (Oriel solar simulator, 450 W Xe lamp, AM 1.5 global filter) was calibrated to 1 sun (100mW cm-2) usingan opticalpowermeter(Newport, model 1916-C) equipped with a Newport 818P thermopile detector. The photoanode film area for the DSSC performance test was typically 0.25 cm2. IPCE was measured on the basis of a Jobin-YvonTriax 190 monochromator.Intensity-modulated photocurrent/photovoltagespectra (IMPS/IMVS) were taken by the Zahner Zennium C-IMPS system based on an IM6x electrochemical workstation, using green light-emitting diode (525 nm, LED) as the light source.

3. Results and Discussion

3.1. Nanostructural Characterizations. Shown in Figure

1A,B are TEM images of the as-synthesizedSnO2 nanoparticles. The nanoparticlesare well dispersedwithoutseriousaggregation

(Figure 1A), a feature that is important for film fabrication. High-resolution image and the corresponding selective area diffraction pattern (Figure 1B) show that the nanoparticles are well-crystallized with sizes in the range of 6-10 nm, much smaller than the commercial SnO2 nanoparticles (15 nm in size) normally used in the literatures.6,8 A typical SEM image in

Figure1C revealsa good uniformityand the uniquesymmetrical branching morphology (40 nm in arm diameter and 500 nm in length) of the ZnO nanotetrapods. Importantly, the branching trait of the ZnO nanotetrapots make them apt to forming 3D interpenetrating networks imbedded with small SnO2 nanoparticles. The ZnO nanotetrapods feature four arms growing symmetricallyfrom a common center along the 〈0001〉 direction (Figure 1D,E); such 1D arms not only offer a mechanically robust scaffold but also provide rapid electron transport pathways due to the minimized grain boundaries.

B J. Phys. Chem. A, Vol. x, No. x, X Chen et al.

SEM images (Figure 2) reveal that pure SnO2 film and SnO2 nanoparticles/ZnO nanotetrapods composite films are all highly porous. In composite films, there are additional larger pores caused by the integration of ZnO nanotetrapods into the SnO2 nanoparticlesmatrix with incompatiblemorphologiesand length scales: some interspaces between nanotetrapods are not fully filled by nanoparticles. The more ZnO nanotetrapods are blended, the more hollow the films become, and the more optically opaque the films are because of the light scattering effect, as will be discussed below. In addition, the branched nanotetrapods are beneficial to stabilizing the composite film by releasing any locally accumulated, capillary-force-induced stress during the film drying process when freshly deposited on a FTO glass substrate. From our experience, with the higher content of ZnO nanotetrapods, thick (>10 µm), well-adhered, and crack-free composite films could be more easily fabricated

(Figure 2D,F,H). For pure SnO2 nanoparticles, in contrast, a crack-free film (Figure 2B) could be made to adhere to a conductive glass substrate only when it was relatively thin (<5 µm). From XRD data (Supporting Information, Figure S2), the

SnO2 nanoparticles are in the tetragonal rutile phase, and the average crystal size is calculated to be 8.0 ( 0.5 nm by the

Scherrer’s equation, which is in accord with the TEM observations. The composite films exhibit only XRD peaks of the rutile

SnO2 and the wurtzite ZnO without sign of any new crystal phase such as ZnSn(OH)6, ZnSnO3,Z n2SnO4, and so on, which is suggestive of physical blending of the two for the most part. However, when the composite film samples were interrogated by TEM-EDX (Supporting Information, Figure S3), the Zn elementcouldalwaysbe foundin the SnO2 nanoparticlesregions with a Zn/Sn atomic ratio of 2-5%, which indicates a situation that is more complicated than the plain blending of the SnO2 nanoparticles and the ZnO nanotetrapods. No Sn was detected in the ZnO nanotetrapods regions.

To confirm the inference above, we turned to XPS. The presence of Sn-O-Zn bonds between the SnO2 core and the ZnO shell is evidencedby the comparisonof the Sn 3d5/2 spectra of the composite films and the pure SnO2 film. As the XPS spectra in Figure 3 show, the Sn 3d5/2 peak is shifted to a lower energy for the composite films (486.20 eV) from that of the pure SnO2 film (486.65eV). A similarshift has been consistently observed in SnO2-ZnO core-shell nanostructures compared with that of pure SnO2.35 This led us to believe that an ultrathin ZnO shell must have formed spontaneously on the SnO2 nanoparticles in our SnO2 nanoparticles/ZnO nanotetrapods composite film. The lower electrophilicity of Zn2+ in the shell than that of Sn4+ in the core should be responsible for the down shift of the Sn 3d5/2 peak upon the formation of the ultrathin ZnO shell. The peak position at 486.20 eV is independent of the composition for the three composite films, indicating a similar ZnO shell coverage and thickness on SnO2 under the same treatment conditions.

We suspect that the formation of the shell layer of ZnO on the SnO2 nanoparticles stems from the chemical instability of ZnO in a polar solvent; the alkaline ZnO tends to be deposited on the relativelyacidicsurfaceof SnO2.8 The ammoniatreatment is considered to promote this process for the full coverage of the SnO2 surface with ZnO by the facilitated mass transport through the Zn(NH3)4 2+ complex in solution. Using an atomic ratio of 2-5% for Zn/Sn in the SnO2 nanoparticles region (estimated from the TEM-EDX measurement) and assuming ideal spheres of 8 nm diameter for the SnO2 nanoparticles, we estimatethe ZnO shell to be only one to two atomic layers thick, which explains why the shells could not be imaged clearly, even with high-resolution TEM.7

Pore size distributions of film samples peeled off from conductiveglass substratesare shown in Figure 4. In going from pure SnO2 film to composite films with increasing ZnO content, the pore size distribution gradually becomes wider, which is consistent with their SEM images in Figure 2. For pure SnO2 film, the pores are narrowly distributed at ∼13.9 nm, associated with long and narrow pore channels arising from nanoparticle hierarchical aggregation. This is inferred from its nitrogen adsorption-desorption isotherm (the inset in Figure 4), which is of the typical type IV with H1 hysteresis loop, very similar to that of mesoporousSnO2 obtainedby P123 surfactantdirected self-assembly.37 The long and narrow cylindrical channels are constructed by SnO2 nanoparticles and templated by P123 micelles. The fitted BET surface area is 98.7 m2 g-1. For composite films, the pore size distribution becomes broader, but still ∼14 nm, with the increase in the ZnO nanotetrapod content because more irregular, polydisperseinterspacesare formed and incompletely filled by the SnO2 nanoparticles, for which the pure nanotetrapods film presents an extreme case. The broader pore size distributions are embodied by the shape of their nitrogen adsorption-desorption isotherms: (1) the inflection points of isotherms shift toward high relative pressure (P/P0), corresponding to the capillary condensation of larger pores; (2) the adsorption-desorption hysteresis loops gradually transform from H1 type to H3 type associatedwith irregularpores because

Figure 1. (A) Low-resolution and (B) high-resolution TEM images of SnO2 nanoparticles.The inset in part B is the correspondingselective area electrondiffractionpattern.(C) SEM imageand (D,E)TEM images of ZnO nanotetrapods.

High-Efficiency Dye-Sensitized Solar Cells J. Phys. Chem. A, Vol. x, No. x, X C

of the loss of long and narrow cylindrical channels when more and more ZnO nanotetrapods are integrated into the SnO2 nanoparticles matrix.

By blending SnO2 nanoparticles with more and more ZnO nanotetrapods, BET surface areas of the composite films are gradually increased from 85.1 (SnO2/ZnO 2:1), to 109.1 (SnO2/ ZnO 1:1), to 128.5 m2 g-1 (SnO2/ZnO 1:2). The latter two BET

surface areas are even higher than that of the pure SnO2 nanoparticles film (98.7 m2 g-1). This is unexpected because the BET surface area of the pure ZnO nanotetrapods film is relatively low (18.9 m2 g-1),3 and blending more ZnO nanotetrapodsinto a compositefilm shouldhave only trimmeddown the BET surfacearea.Such abnormalphenomenonagainimplies that the composite films are not simply a physical blending of the ZnO nanotetrapods and the SnO2 nanoparticles. In fact, as can be seen from Figure 4, more and more pores are distributed

Figure 3. Sn 3d5/2 XPS spectra of pure SnO2 nanoparticles film and the three composite films.

Figure 2. High-resolution and low-resolution SEM images. (A,B) Pure SnO2 nanoparticles film. (C-H) SnO2 nanoparticles/ZnO nanotetrapods composite films with different weight ratios: (C,D) 2:1, (E,F) 1:1, (G,H) 1:2. Scale bars: (A) 100 nm, (B) 1 µm, (C,E,G) 100 nm, (D,F,H) 10 µm.

D J. Phys. Chem. A, Vol. x, No. x, X Chen et al.

in the range of 3-6 nm as the ZnO content is increased in the composite films, which may be partially responsible for the abnormal increase in BET surface area. One of the reasons for the formation of the small pores may be hindered hierarchical aggregation of the ∼8n mS nO2 nanoparticles as a result of the increasing incorporation of the ZnO nanotetrapods such that the pore size is mainly determined by the interspaces of the primary nanoparticle assembly. Moreover, the roughening of the ZnO and SnO2 surfaces together with the formation of the ZnO ultrafine nanoparticles in composite films with an increasing

ZnO content should also increase the BET surface area.

The difference in pore structure has a direct influence on the light scatteringpropertyof the compositefilms, as can be clearly perceived from their diffuse reflectance spectra in Figure 5. It should be pointed out that the dramatic reduction of reflection below 380 nm (except the pure SnO2 film) and below 340 nm (except the pure ZnO film) is caused by the band gap transitions of ZnO (Eg ) 3.2 eV) and SnO2 (Eg ) 3.6 eV), respectively. Overall, the reflectance spectra appear to abide by the Mie scattering theory, which states that pores with a comparable size to the light wavelength can act as effective light scattering centers.38,39 Understandably, light scattering in the pure SnO2 nanoparticles film is nearly negligible because the pores inside the film are too small (<20 nm) to act as effective scattering centers. However, the diffuse reflectance of the composite films increases dramatically and monotonously with the increase in the ZnO content,with the utmostfor the pure ZnO nanotetrapods films. Notably, a substantial fraction of visible light is scattered back. This is not surprising because more ZnO nanotetrapods blended with SnO2 nanoparticles will generate more pores with sizes closer to ideal for light scattering, namely, 50-100 nm, as suggested by Hore et al.38 Similar diffuse reflectance spectra arising from the scattering of voids (pores) were previously reported.16,40 3.2. Performance Appraisal of the Dye-Sensitized Solar

Cells. Presented in Figure 6 are the performance test results of the DSSCs based on the three series of composite films under 1 Sun AM 1.5G simulated solar light. The corresponding data for the pure SnO2 nanoparticles film and for the pure ZnO nanotetrapods film33 are also included for comparison.

The dependence of roughness factor (RF) on film thickness for the different series of films is plotted in Figure 6A. RF was calculated by multiplying a BET surface area by the weight of a film peeled off from the glass substrate, which is then divided by the area of the film. Significantly, the RF values for the three series of composite films can exceed 1600, much higher than that of the pure ZnO nanotetrapods film and also comparable to those of the most efficient TiO2 nanoparticles photoanodes.41

The dependence of dye-adsorbing amount on film thickness shows roughly similar increasing trends as those of RF (Figure 6A). It is noticed that the dye-adsorbing amount per RF for the pure SnO2 film is obviously lower than those of composite films probably because of the more basic surfaces of the latter, that is, the ZnO ultrathin layer modified SnO2 surfaces.8 More to the point, because the dye-adsorbing time of4hi n this work was optimizedwith respect to the compositefilms, it is probably too short for saturated dye adsorption on the pure SnO2 nanoparticles film.

(Parte 1 de 4)