November 01, 2012
C 2012 American Chemical Society
Size-Controlled Anatase Titania Single
Crystals with Octahedron-like
Morphology for Dye-Sensitized
Solar Cells
Jia-Wei Shiu, Chi-Ming Lan, Yu-Cheng Chang, Hui-Ping Wu, Wei-Kai Huang, and Eric Wei-Guang Diau* Department of Applied Chemistry and Institute of Molecular Science, National Chiao Tung University, Hsinchu 30010, Taiwan
N
anostructured TiO2is of greatinter-est to researchers because the re-lated materials have been utilized in many applications such as photocatalysis, sensors, batteries, photovoltaics, water splitt-ing, and so forth.1,2In particular, TiO2plays a
key role as a mediator of electron transport in working electrodes for dye-sensitized solar cells (DSSCs);36in such cells, dye molecules are sensitized on the surface of TiO2films.
Photoexcitation of the dye produces an effective charge separation through rapid injection of electrons from the excited state of the dye into the conduction band of TiO2;
these injected electrons in TiO2
subsequ-ently migrate toward the electrode, with minimal loss through charge recombina-tion, to generate sufficient photocurrents for the device. An ideal TiO2film should thus
have adequate porosity and a large specific surface area to attain dye loading (DL) to a
sufficient extent and for feasible transport of the redox couple of the electrolyte across the network of mesoporous particles.4,5To
fulfill the functionality of such TiO2, the
effects of particle size,711particle shape,1218 film composition,1923morphology,2432
cry-stalline phase,3335and peptization condi-tion36on photovoltaic performance of the device have been investigated.
For typical highly efficient DSSCs, the TiO2
films feature a double-layer structure: a scat-tering layer (SL) with a thickness of 25 μm on top of a transparent TiO2 active layer
(AL) with a thickness of 1214 μm.59,3740
For an AL, anatase TiO2nanoparticles of size
∼20 nm exhibit optimal dye adsorption and light harvesting in the visible spectral region; for SL, the size of particles was increased to 200400 nm to enhance the light har-vesting in the wavelength region of 600 800 nm. Although a TiO2SL advantageously
* Address correspondence to [email protected].
Received for review September 14, 2012 and accepted November 1, 2012. Published online
10.1021/nn3042418
ABSTRACT A simple hydrothermal method with titanium tetra-isopropoxide (TTIP) as a precursor and triethanolamine (TEOA) as a chelating agent enabled growth in the presence of a base (diethylamine, DEA) of anatase titania nanocrystals (HD1HD5) of controlled size. DEA played a key role to expedite this growth, for which a biphasic crystal growth mechanism is proposed. The produced single crystals of titania show octahedron-like morphology with sizes in a broad range of 30400 nm; a typical, extra large, octahedral single crystal (HD5) of length 410 nm and width 260 nm
was obtained after repeating a sequential hydrothermal treatment using HD3 and then HD4 as a seed crystal. The nanocrystals of size∼30 nm (HD1) and ∼300 nm (HD5) served as active layer and scattering layer, respectively, to fabricate N719-sensitized solar cells. These HD devices showed greater VOCthan
devices of conventional nanoparticle (NP) type; the overall device performance of HD attained an efficiency of 10.2% power conversion at a total film thickness of 28μm, which is superior to that of a NP-based reference device (η = 9.6%) optimized at a total film thickness of 1820 μm. According to results obtained from transient photoelectric and charge extraction measurements, this superior performance of HD devices relative to their NP counterparts is due to the more rapid electron transport and greater TiO2potential.
KEYWORDS: anatase . crystal growth . dye-sensitized solar cells . hydrothermal . titania
improves the light-harvesting ability of the dye, parti-cles of large size and small surface area would decrease the DL on TiO2films. To combine both a dye-loading
ability and the light-scattering effect in one TiO2layer,
various hierarchical pore structures were developed. For instance, Caruso and Cheng2527reported meso-porous spherical TiO2beads of size 830 nm, Lee and
Park22reported mesoporous TiO2hollow spheres with
a spherical size of 13 μm and a wall thickness of ∼250 nm, and Hore et al.21
and Han et al.23reported interstitial voids in nanocrystalline films. These hier-archical TiO2 beads and spheres are composed of
nanoparticles in the anatase phase of size∼20 nm so that they exhibited an excellent bifunctional character (AL þ SL) for use as photoanodes to promote the overall performance of the DSSCs. Hierarchical anatase TiO2 mesospheres consisting of various
nanostruc-tures, such as nanorods,24 nanocubes,16 and nano-sheets,17were also reported for DSSC applications.
Various methods exist for synthesis of TiO2
nanos-tructures of controlled shape and tunable size.1,2As a SL for DSSCs, for instance, monodisperse mesoporous TiO2spheres were produced in a two-step synthesis
comprising a controlled solgel process and a hydro-thermal or solvohydro-thermal treatment with titania pre-cursor at a varied molar ratio in water under basic conditions.8,9In a typical hydrothermal approach, var-ious experimental conditions such as titanium precur-sor, concentration, peptizer, catalyst, surfactant or addi-tive, pH, autoclaving temperature, and duration must be considered to synthesize appropriate anatase TiO2
nanoparticles as AL for DSSCs. Grätzel and co-workers reported that the size and morphology of the TiO2
nanoparticles differ when an autoclave was used with varied pH;3,41 not only were the nanoparticles larger under basic condition than under acidic conditions, but also rod-like3,41and octahedral3,15nanostructures were observed in the former case. The formation of octahedral or truncated octahedral nanocrystals was observed also under acidic hydrothermal conditi-ons18,4244or non-aqueous solvothermal treatments in the presence of capping surfactants.4547The
mor-phology of a truncated octahedral single crystal is controllable on varying the concentration of the hy-drofluoric acid solution to adjust the fraction of reac-tive{001} facets with respect to the eight thermo-dynamically stable{101} facets;18,48,49the best perfor-mance of a device correlated directly with the greatest portion of exposed (001) facets.1618
Even though hydrofluoric acid played a key role as a surface-capping agent in the hydrothermal treatment to control the crystallography and morphology of TiO2,18,48,49using concentrated HF raises serious
con-cerns about hazard to the environment and health due to its toxic and corrosive nature. Hore and Durrant36 found that devices with TiO2 films prepared under
basic conditions exhibited smaller dark currents and
greater open-circuit voltage (VOC) than those prepared
under acidic conditions. Accordingly, we developed a simple two-step sol/hydrothermal approach using ti-tanium tetraisopropoxide (TTIP) as a precursor and triethanolamine (TEOA) as a chelating agent to retard the hydrolysis.50,51We thus grew anatase titania nano-crystals of controlled size in the presence of a base, diethylamine (DEA), that played a key role expediting the nanocrystal growth in the hydrothermal treatment; the produced titania single crystals show octahedron-like morphology and size of 30400 nm. The signifi-cance of the present work is emphasized on the shape control of the titania nanocrystals using the selective crystal face etching method under a basic condition. The∼30 nm nanocrystals (denoted HD1) and ∼300 nm nanocrystals (HD5) were fabricated into N719-based DSSCs as AL and SL, respectively. These HD devices showed significantly greater VOC than those of the
reference devices (denoted NP) for which the TiO2
films were fabricated according to a conventional acid-peptized hydrothermal procedure.40The perfor-mance of the HD devices improved on increasing the thickness of AL, attaining an efficiency η = 10.2% of power conversion optimized at totalfilm thickness 28 μm (SL ∼ 5 μm); this performance is superior to that of the NP device, for whichη = 9.6%, optimized at total film thickness of 1820 μm. To understand the elec-tron transport and charge recombination properties of the system, we undertook transient photoelectric mea-surements of three relevant devices.
RESULTS AND DISCUSSION
Morphology and Crystallinity. Figure 1 shows SEM images of the morphologies of titania nanocrystals generated from the two-step sol/hydrothermal procedures with varied autoclaving temperature and duration. When we examined the effect of autoclaving period at the fixed temperature of 230°C, we found that the average particle size increased from ∼30 nm at 2 h (HD1, Figure 1a) to 5070 nm at 8 h (Figure 1b). The average size did not change further at 16 h (Figure 1c), but some large crystals were observed at extended durations. Concerning the effect of autoclave temperature at fixed duration of 12 h, for temperature increase from 230°C (Figure 1d) to 250 °C (HD2, Figure 1e) to 270 °C (Figure 1f), the average particle size increased signifi-cantly from∼50 nm to over 100 nm. Moreover, the geometry of the titania nanocrystals varied from a cuboid-like shape at 230 °C to the octahedron-like shape at 270°C. As shown in Figure 1g for crystals grown at 270°C for 16 h, larger octahedral or truncated octahedral nanocrystals (HD3) were unambiguously observed. With HD3 as a seed crystal to repeat the same sol/hydrothermal experiment as for HD3, much larger titania single crystals (HD4, Figure 1h) were obtained. Repeating again the experiment with HD4 as a seed crystal, even larger truncated octahedral
crystals (HD5, Figure 1i) were formed; a typical HD5 single crystal highlighted in Figure 1i has diagonal size length of 410 nm and a width of 260 nm, which are appropriate for use as a SL for a DSSCs.
The formation of cuboidal TiO2nanocrystals was
observed also in conventional solgel experiments at intermediate pH,50,51but the shape became altered from cuboidal to ellipsoidal under basic conditions (pH ∼11.5),50
in the presence of ammonia52or a primary, secondary, and tertiary amine.51Although the genera-tion of octahedron-like TiO2nanocrystals is reported from
a surfactant-assisted hydrothermal or solvothermal meth-od under various experimental conditions,15,18,4247 the crystal sizes were small, <50 nm, even with auto-claving protracted up to 255 h.42From a solvothermal approach, for example, Wu et al.46reported the forma-tion of monodisperse rhombic-shaped anatase titania single crystals of length∼20 nm using a primary amine as the capping agent. Dinh et al.47reported the
shape-controlled synthesis of TiO2 nanocrystals using oleic
acid (OA) and oleylamine (OM) as distinct capping surfactants to generate rhombic and truncated rhom-bic anatase titania single crystals of length∼38 and ∼18 nm, respectively; to suppress the growth of (101) facets, Yan et al.15 synthesized octahedral anatase titania nanocrystals with ammonium bicarbonate as a
surface-capping reagent and obtained nearly uniform TiO2 nanooctahedra of size∼17 nm. A synthesis of
octahedron-like anatase titania single crystals with a length as great as 400 nm that we achieved in the present work has never been reported elsewhere.
The detailed morphologies of HD1 are visible in the TEM and HRTEM images in Figure 2a,b, respectively. Consistent with the SEM images shown in Figure 1a, the TEM images of the HD1 crystals in Figure 2a exhibit mainly a“cuboidal” shape, but several crystals showing the octahedron-like shape are highlighted with red circles. The HRTEM image shown in Figure 2b indicates the distance between adjacent lattice fringes to be 0.35 nm, which matches exactly the lattice spacing of the (101) planes of anatase TiO2.12,15,4447The
forma-tion of purely crystalline nanocrystals HD1HD5 in the anatase phase has been confirmed with the XRD patterns shown in Figure 3. The XRD signals increas-ingly sharpen from HD1 to HD5, indicating a greater crystallinity and a larger crystal. The three diffraction signals of the (103), (004), and (112) facets were well-resolved for the HD crystals, but those of the NP sample were congested and unresolved; this condition indi-cates that the HD crystals evolved along the [001] direction,47for which a mechanism of crystal growth is offered below. On the basis of this crystallographic
Figure 1. SEM images of TiO2nanocrystals in the HD series synthesized at varied hydrothermal conditions: (a) 230°C and 2 h
(HD1), (b) 230°C and 8 h, (c) 230 °C and 16 h, (d) 230 °C and 12 h, (e) 250 °C and 12 h (HD2), (f) 270 °C and 12 h, (g) 270 °C and 16 h (HD3), (h) repeated hydrothermal treatment using the same sol with HD3 as a seed at 270°C for 16 h (HD4); the red circles show the (001) facets. (i) Repeated hydrothermal treatment using the same sol with HD4 as a seed at 270°C for 16 h (HD5); the red circle shows a titania single crystal with a truncated octahedral morphology of a extraordinarily large size. The scale bars indicate 100 nm for each plot.
information, we constructed a speculative 3D octa-hedral geometry for a typical HD1 crystal shown in Figure 2b; the directions of both (101) and (001) facets are also indicated. According to the rhombic-shaped HRTEM images analyzed elsewhere,15,4447 we pro-pose that the quadrate TEM image of HD1 shown in Figure 2b is the result of a projection from an octahe-dral crystal with a short c-axis, consistent with the observation of other crystals showing an octahedral geometry with a longer c-axis highlighted in Figure 2a. Crystal Growth Mechanism. DEA served as a surface-capping agent in other solgel systems to promote the growth of ellipsoidal titania nanoparticles with small aspect ratios.51,53To examine the role of DEA played for the crystal growth in our system, we per-formed control experiments with the same synthetic procedure but without adding DEA in the third step of the sol preparation. We found that the crystal size, after hydrothermal treatment in the absence of DEA, is less
than that in the presence of DEA at 230 °C for 2 h (Figure S1a, Supporting Information). Figure S1b, Sup-porting Information, shows that the crystal grew slowly at 270°C for 16 h in the absence of DEA to attain a particle size of∼30 nm, which is much less than that of HD3 (Figure 1g) grown in the presence of DEA. DEA thus plays a key role not only as a catalyst to increase the rate of crystal growth but also as a shape controller to refine the nanocrystals with a beautiful truncated octahedral geometry. A mechanism for the crystal growth is presented in Scheme 1 to rationalize the formation of the octahedron-like anatase titania single crystals of extraordinary size.
In the first step, TEOA was mixed with TTIP at a molar ratio of 2:1; the solution turned from colorless to pale yellow, indicating the formation of a complex. We thus propose the formation of a TEOA-chelated tita-nium complex [Ti(OR2)n(OR1)42n], with n = 1 or 2, of
which the molecular structures of OR1 and OR2 are shown in Scheme 1. Sugimoto et al.5052indicated that TEOA plays a role as a chelating agent to suppress rapid hydrolysis of the titanium complex. We treated TEOA as a bidentate ligand (OR2), but it might also act as a tri-dentate ligand.54When water was added to the mix-ture of TTIP and TEOA, hydration and hydrolysis oc-curred to form an octahedral complex, [Ti(OH)m(OR2)n
-(H2O)6m2n]4m2n, with me 6, n e 3, and m þ 2n e
6. We expect that the bidentate ligand OR2 has a bridged chelating feature to prevent the approach of water molecules toward the central metal ion and thus may slow both the hydrolysis and the ensuing con-densation.55 For the solgel reaction performed at 100°C for 24 h, we expect that the required hydrolysis and condensation generated a sol suitable for the subsequent hydrothermal reaction to proceed. As the titania nanocrystals were produced in only the pure anatase phase, the formation of [Ti(OH)2(H2O)4]2þ
mono-mers is expected to generate nuclei of anatase type
Figure 2. (a) TEM images of HD1 showing the cuboidal morphology of the nanocrystals; the red circles highlight the images with rhombic shapes that represent the formation of octahedron-like nanocrystals. (b) HRTEM image of one specific HD1 single crystal that can be emulated with a 3D octahedral structure with the directions of (101) and (001) planes as indicated; the inset shows thefine structure of the lattice with a distance of 0.35 nm between the (101) facets in the anatase phase.
Figure 3. XRD patterns of base-catalyzed nanocrystals HD1 HD5 and acid-peptized nanoparticles (NP) exhibiting the well-defined signals assigned to the deflections of the lattice facets in the anatase phase (JCPDS no. 21-1272).
that serve as seeds in the solution for the succeeding condensation and nucleation.56Once nucleation gen-erated the seed nuclei of anatase type, as the two crystal growth units shown in Scheme 1, DEA would play a key role as a bifunctional reagent;catalyst and shape controller;during the hydrothermal process to produce initially small octahedral (cuboid-like) nano-crystals (HD1), from which the nano-crystals evolved to become large, truncated octahedral nanocrystals (HD3 HD5) as we have observed.
According to a mechanism involving a two-phase morphology evolution that Penn and Banfield pro-posed,42the growth of a crystal along the [001] facet is twice as rapid as that along the [101] facet in thefirst phase (hydrothermal duration 010 h), but it becomes 33.5 times as rapid along [001] than along [101] in the second phase (hydrothermal duration >10 h). The (001) faces hence expand during thefirst phase but shrink during the second phase. We observed that the crystal morphology evolved from an octahedral shape (HD1, Figure 2b) to a truncated octahedral shape with a larger exposed (001) facet (HD3, highlighted inside the red circles in Figure 1g), and then further evolved to a large truncated octahedron with a smaller exposed (001) facet (HD5, highlighted in Figure 1i). This evolu-tion from HD1 to HD5 in our case can also be inter-preted with a biphasic model, according to which the crystal growth along the [001] facet expanded from HD1 to HD3 and shrank from HD3 to HD5, as indicated in Scheme 1.
Why did such a (001) expanding and shrinking of the crystal evolution occur in our system? Dinh et al.47 found that the evolution of the crystal shape depends on the molar ratio OA/OM, for which the rhombic, truncated rhombic, and spherical morphologies were produced at OA/OM ratios of 4/6, 5/5, and 6/4,
respectively; only rhombic-type crystals were produced under more basic conditions (OA/OM < 4/6). It was pro-posed that the base OM tends to bind to the [101] facets, whereas the acid OA binds to the [001] facets to control the crystal growth in the corresponding direc-tions. In our case, we expect that the base DEA plays a role similar to that of OM at 230°C and hydrothermal period of 2 h to generate octahedral-type crystals of small size (HD1). At this stage, DEA functions as a surface-capping agent to inhibit the crystal growth along the [101] facets. When the temperature was increased to 270°C for 16 h, the surface-capping effect gradually diminished because of the greatly enhanced surface area when the crystal was growing. At this stage, the catalytic role of DEA is expected to accelerate the crystal growth along the [101] facets to form truncated octahedral crystals of moderate size (HD3). TEOA released from the solgel reaction50
might play a role similar to OA to inhibit the crystal growth along the (001) direction responsible for the formation of truncated octahedral crystals observed in this stage;a (001) expanding crystal evolution. When the crystals evolved further in the presence of a new sol solution, DEA in the fresh solution again played a role inhibiting the crystal growth along the (101) direction for the (001) shrinking to occur, producing the truncated octahedral crystals with a small exposed [001] area and an extraordinarily large size (HD5).
Photovoltaic Performance. Nanocrystals HD1HD5 served as AL for DSSC applications; the devices were fabricated with identical components;N719 dye, io-dide/triiodide electrolyte, and Pt counter electrode; except that TiO2AL was made of HD1HD5 with the
same film thickness (∼11 μm). To assess the effect of these nanocrystals as AL on photovoltaic performance, we added no SL at this stage. The currentvoltage
Scheme 1. Mechanism for the formation of anatase titania single crystals HD1HD5 with octahedron-like morphology.
(JV) characteristics and the corresponding IPCE ac-tion spectra of these devices are shown in Figure 4a,b, respectively; the photovoltaic parameters are summar-ized in Table 1. Integrating the IPCE over the AM-1.5G solar spectrum (Figure S2, Supporting Information) yielded the calculated JSC (shown in parentheses in
Table 1), similar to the collected values for all devices under investigation, thus validating the photovoltaic results listed in Table 1 obtained from the JV mea-surement.
The photovoltaic performances of these devices exhibited a decreasing order of JSCfrom HD1 to HD5,
but VOC showed an opposite trend, giving the best
device HD1 with an efficiency η = 8.0% power conver-sion under a thin-film condition without an added SL. The systematic variation of JSCis consistent with DL on
the TiO2films, for which DL = 73 nmol cm2onfilm
HD1, but it decreased to only 3 nmol cm2onfilm HD5. The specific surface areas (SBET/m2 g1) determined
with the BrunauerEmmettTeller (BET) method gave values 62.7, 29.2, 13.0, 7.4, and 5.1 for HD1HD5, respectively, which are consistent with the variation of DL showing the effect of particle size. The variation of VOCis remarkable in that VOCwas significantly
en-hanced from 829 mV (HD1) to 949 mV (HD5), indicating the superior performance of HD5 nanocrystals of large size. Such a large VOCin a DSSC based on an iodine
electrolyte and with a N719 dye is unprecedented. We expect that the injected electrons in such a large single
crystal would transport efficiently to the electrode, together with the small surface area to diminish the possibility of charge recombination for the observed appreciable VOCof the HD5 device, but the large crystal
size of HD5 leaves a small surface area available for dye adsorption, which leads to a small JSC and a poor
overall performance for the HD5 device.
According to the marked performance of HD1 at a thin-film condition, we further tested the effect of film thickness on device performance using HD1 as AL and HD5 as SL. As a reference for comparison, the experi-ments were also performed with NP as AL and CCIC as SL according to the same procedure as that of Ito et al.40For the HD1þ HD5 system, the total thickness (L) of the ALþ SL film was readily controllable across a broad range from 17 to 32μm; for the NP þ CCIC system, a thickfilm tended to crack, which limited the film thickness to a range of 1522 μm. In both cases, the thickness of the SL wasfixed at ∼5 μm. The JV curves obtained for the HD1 þ HD5 system (six devices) and the NPþ CCIC system (four devices) are shown in Figure S3a,b, respectively; the corresponding DL and photovoltaic parameters are summarized in Table S1, Supporting Information.
Figure 5ae shows parameters DL, JSC, VOC, FF, and
η as a function of L, respectively. The results indicate that DL increased as L increased for both systems, consistent with the variation of JSCfor the HD system
to attain the maximum JSCvalue, 16.27 mA cm2, at
L = 32μm. For the NP system, JSCattained its maximum
value, 16.18 mA cm2, at L = 20μm, in agreement with the results of Ito et al. that show the best performance of a device to occur at L = 1719 μm (1214 μm for AL).40The DL values are much greater for NP than for
HD at similar L because the particle size of NP was much smaller than that of HD, making the surface area of the former much greater than that of the latter. Larger amounts of loaded dye increased JSCfor the NP
devices to outperform the HD devices for L < 21μm, but the robust mechanical structure of the HDfilms enabled us to make thickerfilms to enhance DL and JSC
Figure 4. Optimized photovoltaic properties: (a) curren-tvoltage characteristics and (b) corresponding IPCE action spectra of N719-dye-sensitized devices sensitized with TiO2
active layers made of HD1HD5 nanocrystals at a constant film thickness of ∼11 μm without added scattering layer but with TiCl4post-treatment under irradiation (one-sun
AM-1.5G).
TABLE 1. Amount of Dye Loading (DL), Specific Surface Area (SBET), and Optimized Photovoltaic Parameters of
Devices Made of N719 Dye and HD1HD5 TiO2Films with
Post-TiCl4Treatment under Simulated AM-1.5G
Illumina-tion (Power = 100 mW cm2) and Active Area of 0.16 cm2
with a Shadow Mask Area of 0.25 cm2a
TiO2film DL (nmol cm2) SBET(m2g1) JSCb(mA cm2) VOC(mV) FF η (%)
HD1 73 62.7 12.89 (12.78) 829 0.747 8.0 HD2 33 29.2 11.15 (11.13) 873 0.753 7.3 HD3 15 13.0 6.68 (6.32) 922 0.747 4.6 HD4 7.4 7.4 3.77 (3.87) 927 0.747 2.6 HD5 2.8 5.1 2.06 (2.22) 949 0.714 1.4 a
The devices contain only one active layer of the samefilm thickness (∼11 μm) without added scattering layer.bThe values shown in parentheses were obtained
from integration of the IPCE spectra with the solar irradianceflux spectra.
for the HD devices, to outperform the NP devices for L > 21μm.
The trend of VOCis opposite that of JSCbecause
charge recombination plays a key role infilm thickness (discussed in the next section). Upon increasing L for both systems, the VOCvalues decreased linearly and in
parallel with∼50 mV advance for the HD over the NP system. The FF values show no correlation with L, but the average value of HD is slightly greater than that of NP, which might indicate the advantage of HDfilms with conductivity greater than that of NPfilms. Because of the countertendency between VOCand JSC, upon
increasing thefilm thickness, the overall performance of the HD devices was dominated by JSC until a
maximum efficiency was attained at L = 28 μm, for which JSC= 15.87 mA cm2, VOC= 825 mV, FF = 0.776,
and η = 10.2%. For the NP devices, the best perfor-mance (η = 9.6%) occurred for L = 1820 μm.
Arakawa and co-workers19testedfilm composition with varied particle size on device performance as a function offilm thickness; when they varied the film thickness of a mixed particle system (particle sizes of 23 and 100 nm) in a broad range (1131 μm), they found an optimal thickness in the range of 1518 μm; both JSCand VOCdecreased for afilm thicker than 18 μm. To
test the effect of film thickness on device performance,
Adachi and co-workers13used TiO2films composed of
highly crystalline nanorods with a length of 100 300 nm and a diameter of 2030 nm; the optimal performance of the nanorod cell occurred at a film thickness greater than 16μm, whereas the P25 cell attained the best performance at a smaller thickness; that the FF values were greater for the nanorod devices than for the P25 devices was attributed to the high crystallinity of the nanorod films. In our system, the best performance of the HD devices occurred at an overallfilm thickness of 23 þ 5 μm, which is difficult to attain for a regular DSSC system but is feasible in our HD1 þ HD5 system because of the remarkable me-chanical strength of the films; in that manner, an enhanced DL can be achieved to overcome the short-age of HD1 single crystals with a small specific sur-face area.
Transient Photoelectric Measurements. To understand the discrepancy of the device performance between the two systems, we fabricated three devices, HD-18, HD-29, and NP-17, for transient photoelectric measure-ments. The HD-18 and HD-29 devices are based on the HD1þ HD5 system with the total film thicknesses of 18 and 29μm, respectively; the NP-17 device is based on the NPþ CCIC system with a total film thickness of 17 μm. All three devices involve a SL of the same film thickness,∼5 μm. The corresponding JV curves and IPCE action spectra of these three devices are shown in Figure S4a,b, respectively. The kinetics of electron transport and charge recombination of these devices were investigated via decays ofΔJSCandΔVOCversus
time, respectively, based on eight white-light (WL) intensities as bias irradiations (power densities P0in a
range of 4.298 mW cm2). The photoelectric transi-ents were obtained with a red pulsed LED (λ = 630 nm, duration∼50 ms) under these WL bias irradiations; the resulting photocurrent and photovoltage decays are shown in Figures S5S10, Supporting Information.
Decay curves ofΔJSCandΔVOCversus time of these
devices werefitted according to a single exponential function to determine time coefficients for electron collection (τC) and charge recombination (τR),
respec-tively. For the kinetics of electron transport, we showτC
and diffusion coefficients (D) as a function of JSCin
Figure 6a,b, respectively. Our results indicate that the electron collection periods of the NP-17 and HD-29 devices are similar but much greater than those of the HD-18 device when compared at the same JSClevel.
The largeτCvalues of the HD-29 device are due to its
thickerfilm, for which greater durations of transport are required to collect the electrons at the TCO elec-trode. When thefilm thickness is taken into account, the collection periods become convertible into di ffu-sion coefficients via D = L2/(2.77τC).
57
As shown in Figure 6b, the diffusion coefficients of both HD-18 and HD-29 devices are similar to each other but are 3 times those of the NP-17 device, indicating that the
Figure 5. (a) Amounts of dye loading (DL) and DSSC photo-voltaic parameters, (b) short-circuit current density (JSC), (c)
open-circuit voltage (VOC), (d)fill factor (FF), and (e)
effi-ciency (η) of power conversion, as a function of total TiO2
film thickness (L). The results were obtained from the HD1 þ HD5 system (filled squares) and the NP þ CCIC system (filled circles) for which HD1 or NP served as an active layer (AL) and HD5 or CCIC served as a scattering layer (SL). The thickness of the SL wasfixed at ∼5 μm for all devices.
electron transport of the HD nanocrystals is much superior to that of the NP nanoparticles as an AL for DSSCs.58
To make a fair comparison of charge recombination kinetics of the systems under investigation, we exam-ined the location of the edge of the conduction band of TiO2 with a charge-extraction method. For this
pur-pose, charge densities (Ne) at each VOCwere extracted
via rapid switching of the three devices to the short-circuit condition.57Figure 7a,b shows plots of VOCand
τR as a function of Ne, respectively. The results in
Figure 7a indicate that the TiO2potentials of the two
HD films are similar, with the edge of the potential band of thinnerfilm HD-18 being slightly greater than that of thickerfilm HD-29, but the potentials of both HD films are located significantly above the potential of the NPfilm. The potential shift by ∼50 mV between the HD-18 and the NP-17films is consistent with discre-pancy of VOCshowing a∼30 mV advance for the HD
devices over the NP devices (Figure S4a). Figure 7b shows that the electron lifetimes of the HD-29 device are much smaller than those of the HD-18 device compared at the same charge density, indicating that charge recombination was more significant in the thickerfilm than in the thinner film, so that the VOC
decreased upon increasing the film thickness, as shown in Figure 5c and Figures S3 and S4a. Both HD and NP systems were compared at a similar film thickness (HD-18 vs NP-17): the electron lifetimes of HD-18 were larger than those of NP-17 at smaller charge densities, but the trend reversed at greater
charge densities. We expect that the larger surface area of the NP film and greater DL retard charge recombination more efficiently than for the HD film at a larger Ne, but the electrons inside HD nanocrystals
transported more rapidly than inside the NP nanopar-ticles: the electrons trapped at the surface of the HD-18 device become fewer than for the NP-17 device, so that charge recombination is prevented more effectively for the former than the latter at a smaller Ne. As the
electron lifetimes of both HD-18 and NP-17 have the same order of magnitude, we believe that the discre-pancy of VOCis due mainly to the difference of the edge
of the potential band between the two systems as we observed according to the measurements of charge extraction.59
CONCLUSION
Octahedron-like anatase TiO2single crystals, HD1
HD5, with particle sizes in a range of 30400 nm were prepared according to a simple two-step sol/hydro-thermal approach: thefirst step prepared a suitable TTIP/TEOA/DEA sol at a basic condition (pH 9.6); the second step involved a hydrothermal treatment of varied temperature (230270 °C) and duration (2 16 h). At autoclave temperature of 230°C for a period of 2 h, cuboid-like and octahedron-like nanocrystals of size∼30 nm (HD1) were produced. At hydrothermal temperature of 270°C for 16 h, the crystals evolved to a
Figure 7. (a) Open-circuit voltage (VOC) and (b) electron
lifetime (τR) as a function of charge density (Ne) generated
under eight bias white-light irradiations for NP-17, HD-18, and HD-29 devices. HD and NP represent the devices fabricated based on HD1þ HD5 and NP þ CCIC films, respectively, with suffixed numeric labels representing total film thickness (L).
Figure 6. (a) Electron collection period (τC) and (b) electron
diffusion coefficient (D) as a function of short-circuit current density (JSC) generated under eight bias white-light
irradia-tions for NP-17, HD-18, and HD-29 devices. HD and NP represent devices fabricated based on HD1þ HD5 and NP þ CCIC films, respectively, with the suffixed numeric labels representing the totalfilm thickness (L).
truncated octahedral morphology with a length of ∼150 nm and a width of ∼100 nm (HD3). Performing the same hydrothermal treatment (270°C for 16 h) with HD3 as a seed with a fresh TTIP/TEOA/DEA sol, larger octahedral single crystals (HD4) evolved. Repeat-ing the same experiment with HD4 as a seed, extra-ordinarily large octahedral anatase titania single crystals (HD5) were generated, of a typical length of 410 nm and width of 260 nm. Base diethylamine (DEA) played a bifunctional role as both a catalyst and controller of shape for the crystal growth; on that basis, we proposed a biphasic mechanism of crystal evolu-tion. When HD1HD5 served as active layers for dye-sensitized solar cells with N719 dye as a photosensiti-zer, the photovoltaic performances exhibited a sys-tematic trend with JSCdecreasing from HD1 to HD5,
consistent with their dye-loading amounts reflecting their specific surface areas. In contrast, VOCshowed an
increasing order from HD1 to HD5; the best value
attained 0.949 V for HD5. Among these devices, HD1 was the best candidate as an active layer to attain η = 8.0% with the film thickness fixed at ∼11 μm. The device performance of HD1 was further optimized to attainη = 10.2% at a film thickness of ∼23 μm together with HD5 at ∼5 μm as a light-scattering layer; that result is superior to that,η = 9.6%, of a conventional nanoparticle-type DSSC (NPþ CCIC) optimized at a total film thickness of 1820 μm (CCIC ∼5 μm). To understand the electron transport and charge recom-bination kinetics of these devices, we performed tran-sient photoelectric and charge-extraction measure-ments. The outstanding performance of the HD de-vices relative to the NP dede-vices is attributed to the greater rate of electron transport and the more nega-tive potential for the HD than for the NP system, making the HD single crystals a promising photoanode material for further promotion of device performance for other sensitizers and co-sensitization systems.
EXPERIMENTAL DETAILS
Preparation of Base-Catalyzed TiO2Nanocrystals (HD1HD5). The
experimental procedures are summarized in Scheme S1, Sup-porting Information; the details are given below. Of three steps to make the required sol, in the first, TTIP (0.1 mol, 29.3 mL, 97%; Aldrich) was slowly added to a beaker containing TEOA (0.2 mol, 26.6 mL, 95þ%; TEDIA) and stirred for 5 min at 25 °C; the color of the TTIP/TEOA (1/2) mixture turned pale yellow, which indicated the formation of a TEOA-chelated titanium complex. In the second step, water (100 mL) was added to the beaker to form the TTIP/TEOA aqueous solution and stirred for 6 h at 25°C; the color of the solution became transparently light yellow. In the third step, DEA (0.02 mol, 2.1 mL, 99þ%; ACROS) and water (100 mL) were added to the TTIP/TEOA aqueous solution and then stirred 24 h at 100°C, resulting in the formation of a translucent solution; the prepared TTIP/TEOA/DEA sol was thus ready for the following hydrothermal reaction.
For the hydrothermal synthesis of HD1HD5, the TTIP/ TEOA/DEA sol was transferred into a titanium autoclave and kept at a constant temperature in a range of 230270 °C (autoclave pressure 4060 bar) for a duration of 216 h. For example, the HD1 nanocrystals were obtained at 230°C for 2 h, HD2 at 250°C for 12 h, and HD3 at 270 °C for 16 h. For HD4, we used HD3 as seed crystals and mixed them with the same sol already prepared for hydrothermal reaction at 270°C for 16 h; for HD5, with HD4 as seed crystals, the procedure was repeated as for HD4. These HD particles were then collected on centrifu-gation and washed several times with ethanol. To prepare a screen-printable paste, ethyl cellulose andR-terpineol were added to the ethanol solution of the TiO2nanocrystals HD1
HD5; ethanol was then removed from the solution with a rotary evaporator to obtain a viscous paste suitable for screen printing.60
Preparation of Acid-Peptized TiO2Nanoparticles (NP). The anatase
TiO2nanoparticles (NP) under acidic conditions were prepared
according to a conventional synthetic procedure reported by Grätzel and co-workers,3,40but with a minor modification; the
detailed procedure follows. First, TTIP (0.06 mol, 17.7 mL) was added to a flask containing ethanoic acid (0.06 mol, 3.4 mL, glacial; J.T. Baker) and stirred for 15 min at 25°C; a white precipitate was formed at this stage. Second, water (100 mL) was added slowly to the beaker and stirred for 1 h at 25°C; the produced white turbid solution indicates that the hydrolysis of the sol led to a solution of gel-type solution. Third, nitric acid (0.03 mol, 1.3 mL, 70%; J.T. Baker) was poured into the flask
containing the gel-type aqueous solution and heated to 80°C while the solution was continuously stirred for 1 h; nitric acid served as peptizer for the solgel reaction, and the solution appeared foggy in this stage. After peptization, the gel-type solution was transferred into a titanium autoclave and heated to 230°C for total autoclave period of 12 h to complete the hydrothermal process. The NP particles were eventually col-lected on centrifugation, washed several times with ethanol, and then prepared as the required paste for screen printing according to a procedure reported elsewhere.60
Characterization of Morphology. The morphology of the TiO2
samples was investigated with a field-emission scanning elec-tron microscope (FESEM, JSM-7401F, JEOL). The microstructures of the products were analyzed with a high-resolution transmis-sion electron microscope (FE-HRTEM, JEM-2100F, JEOL) with energy-dispersive X-ray (EDX) analysis of the composition. The crystal phases of the products were characterized with an X-ray diffractometer (XRD, X'Pert Pro, PANalytical, Cu KR radiation), in a 2θ range from 10 to 80°. The specific surface area and porosity of the TiO2 nanocrystals were determined with a nitrogen
adsorptiondesorption apparatus (model TriStar 3000 and VacPrep 061, Micromeretics).
Fabrication of DSSCs.For the working electrode, a paste com-posed of TiO2HD1HD5 or NP as a transparent AL was coated
on a TiCl4-treated FTO glass substrate (TEC 7, Hartford, USA) to
obtain a film of the required thickness (active size 0.4 0.4 cm2)
with repetitive screen printing. Both our HD5 and the commer-cially available CCIC (PST-400C, JGC Catalysts and Chemicals Ltd., Japan) served as the SL to print on top of the transparent nanocrystalline TiO2AL with a film thickness of∼5 μm. The TiO2
films were annealed according to a programmed procedure: heating at 80°C for 15 min, at 135 °C for 10 min, at 325 °C for 30 min, at 375°C for 5 min, at 450 °C for 15 min, and at 500 °C for 15 min. The annealed films were then treated with TiCl4fresh
aqueous solution (40 mM) at 70°C for 30 min and sintered at 500°C for 30 min. The TiO2films as prepared were sensitized in a
solution (N719 dye, Solaronix, 3 104M) containing cheno-deoxycholic acid (CDCA, 3 104M) in acetonitrile/tert-butanol mixture (v/v = 1:1) for 18 h. The electrolyte solution contain-ing GuNCS (0.1 M), I2(0.03 M), PMII (0.6 M), 4-tert-butylpyridine
(0.5 M) in a mixture of acetonitrile and valeronitrile (volume ratio 85:15) was introduced into the space between the two electro-des, completing the fabrication of these DSSC devices.
Photovoltaic Characterization. The photovoltaic performance of a device was assessed through measurement of a JV curve with a solar simulator (AM-1.5G, XES-40S1, SAN-EI), calibrated
with a standard Si reference cell (Oriel PN 91150 V, VLSI stan-dards). The efficiency (η) of conversion of light to electricity was obtained viaη = JSCVOCFF/Pin, in which JSC/mA cm2is the
current density measured at short circuit and VOC/V is the
vol-tage measured at open circuit. Pinis the input radiation power
(for one-sun illumination Pin= 100 mW cm2), and FF is the fill
factor. For all measurements, the DSSC devices were covered with a black mask (aperture area 0.25 cm2) to ensure that the measured photocurrent was not exaggerated. The incident monochromatic efficiencies for conversion from photons to current (IPCE) spectra of the corresponding devices were measured with a system comprising a Xe lamp (A-1010, PTi, 150 W), monochromator (PTi, 1200 gr mm1blazed at 500 nm), and source meter (Keithley 2400, computer-controlled). A stan-dard Si photodiode (S1337-1012BQ, Hamamatsu) served as a reference for the calibration of the power density of the lamp at each wavelength.
Transient Photoelectric Characterization. The photocurrent and photovoltage decays were measured with eight steady-state light intensities as bias irradiations from a white LED.57A red
LED (λ = 630 nm) controlled with a pulse generator (DG535, SRS) generated a perturbation pulse of duration 50 ms. Both the pulsed red light and the steady-state white light irradiated the photoanode side of the cell. The pulsed-probe irradiation was controlled with a LED power supply to maintain the modulated photovoltage less than 5 mV in each measurement. The probe beams generated carriers causing a slightly increased photo-current (ΔJSC) near JSCof the cell at the short-circuit condition or
a slightly increased photovoltage (ΔVOC) near VOCof the cell at
the open-circuit condition, subjected to the white bias light; the current and voltage decays were thereby measured, respec-tively. The resulting photocurrent and photovoltage transients were recorded on a digital oscilloscope (MSO2014, Tektronix); the signals passed a current preamplifier (SR570, SRS) at a short-circuit condition.
Measurement of Charge Extraction. The extracted charges were measured with the same apparatus as that of the transient photoelectric measurements, but with only the white LED at eight intensities of bias light.57Under white LED irradiation, the system was initially set to an open-circuit condition for about 200300 ms for the photovoltage of the device to attain a steady state; the white LED was then terminated while the device was simulta-neously switched to a short-circuit condition with a rapid elec-tronic switch controlled with a pulse generator (DG535, SRS). The transient signals of voltage across 50Ω were concurrently re-corded with a digital oscilloscope (MSO2014, Tektronix) and converted into current transients with Ohm's law. The extracted charge (QCE) was obtained on integration of the current transient
versus time; the extracted charge density (Ne) is determined with
the equation, Ne= (QCE)/(eL(1p)), in which e is elementarycharge,
L is the thickness of the TiO2film, and p is the porosity of TiO2.
Conflict of Interest: The authors declare no competing financial interest.
Acknowledgment. National Science Council of Taiwan and Ministry of Education of Taiwan, under the ATU program, provided support for this project.
Supporting Information Available: Schematic illustration of the synthesis of HD1HD5 TiO2single crystals, additional SEM
images for the crystals synthesized in the absence of DEA, supplementary table andfigures for photovoltaic and transient photoelectric data. This material is available free of charge via the Internet at http://pubs.acs.org.
REFERENCES AND NOTES
1. Chen, X.; Mao, S. S. Titanium Dioxide Nanomaterials: Synthesis, Properties, Modifications, and Applications. Chem. Rev. 2007, 107, 2891–2959.
2. Khataee, A.; Mansoori, G. A. Nanostructured Titanium Dioxide Materials: Properties, Preparation and Applications; World Scientific: Singapore, 2012.
3. Barbé, C. J.; Arendse, F.; Comte, P.; Jirousek, M.; Lenzmann, F.; Shklover, V.; Grätzel, M. Nanocrystalline Titanium Oxide
Electrodes for Photovoltaic Applications. J. Am. Ceram. Soc. 1997, 80, 3157–3171.
4. Grätzel, M. Photoelectrochemical Cells. Nature 2001, 414, 338–344.
5. Grätzel, M. Recent Advances in Sensitized Mesoscopic Solar Cells. Acc. Chem. Res. 2009, 42, 1788–1798. 6. Hagfeldt, A.; Boschloo, G.; Sun, L.; Kloo, L.; Pettersson, H.
Dye-Sensitized Solar Cells. Chem. Rev. 2010, 110, 6595– 6663.
7. Koo, H.-J.; Park, J.; Yoo, B.; Yoo, K.; Kim, K.; Park, N.-G. Size-Dependent Scattering Efficiency in Dye-Sensitized Solar Cell. Inorg. Chem. Acta 2008, 361, 677–683.
8. Yu, I. G.; Kim, Y. J.; Kim, H. J.; Lee, C.; Lee, W. I. Size-Dependent Light-Scattering Effects of Nanoporous TiO2
Spheres in Dye-Sensitized Solar Cells. J. Mater. Chem. 2011, 21, 532–538.
9. Park, Y.-C.; Chang, Y.-J.; Kum, B.-G.; Kong, E.-H.; Son, J. Y.; Kwon, Y. S.; Park, T.; Jang, H. M. Size-Tunable Mesoporous Spherical TiO2as a Scattering Overlayer in
High-Perfor-mance Dye-Sensitized Solar Cells. J. Mater. Chem. 2011, 21, 9582–9586.
10. Kim, Y. J.; Lee, M. H.; Kim, H. J.; Lim, G.; Choi, Y. S.; Park, N.-G.; Kim, K.; Lee, W. I. Formation of Highly Efficient Dye-Sensitized Solar Cells by Hierarchical Pore Generation with Nanoporous TiO2Spheres. Adv. Mater. 2009, 21, 3668–
3673.
11. Chen, Y.; Huang, F.; Chen, D.; Cao, L.; Zhang, X. L.; Caruso, R. A.; Cheng, Y.-B. Effect of Mesoporous TiO2Bead
Dia-meter in Working Electrodes on the Efficiency of Dye-Sensitized Solar Cells. ChemSusChem 2011, 4, 1498–1503. 12. Adachi, M.; Murata, Y.; Takao, J.; Jiu, J.; Sakamoto, M.; Wang, F. Highly Efficient Dye-Sensitized Solar Cells with a Titania Thin-Film Electrode Composed of a Network Structure of Single-Crystal-Like TiO2Nanowires Made by the“Oriented
Attachment” Mechanism. J. Am. Chem. Soc. 2004, 126, 14943–14949.
13. Jiu, J.; Isoda, S.; Wang, F.; Adachi, M. Dye-Sensitized Solar Cells Based on a Single-Crystalline TiO2 Nanorod Film.
J. Phys. Chem. B 2006, 110, 2087–2092.
14. Fujihara, K.; Kumar, A.; Jose, R.; Ramakrishna, S.; Uchida, S. Spray Deposition of Electrospun TiO2Nanorods for
Dye-Sensitized Solar Cell. Nanotechnology 2007, 18, 365709. 15. Yan, K.; Qiu, Y.; Chen, W.; Zhang, M.; Yang, S. A Double
Layered Photoanode Made of Highly Crystalline TiO2
Nanooctahedra and Agglutinated Mesoporous TiO2
Mi-crospheres for High Efficiency Dye Sensitized Solar Cells. Energy Environ. Sci. 2011, 4, 2168–2176.
16. Zhang, H.; Han, Y.; Liu, X.; Liu, P.; Yu, H.; Zhang, S.; Yao, X.; Zhao, H. Anatase TiO2Microspheres with Exposed
Mirror-like Plane {001} Facets for High Performance Dye-Sensitized Solar Cells. Chem. Commun. 2010, 46, 8395–8397. 17. Yang, W.; Li, J.; Wang, Y.; Zhu, F.; Shi, W.; Wan, F.; Xu, D. A Facile Synthesis of Anatase TiO2Nanosheets-Based
Hier-archical Spheres with over 90% {001} Facets for Dye-Sensitized Solar Cells. Chem. Commum. 2011, 47, 1809– 1811.
18. Wu, X.; Chen, Z. G.; Lu, G. Q.; Wang, L. Z. Nanosized Anatase TiO2Single Crystals with Tunable Exposed (001) Facets
for Enhanced Energy Conversion Efficiency of Dye-Sen-sitized Solar Cells. Adv. Funct. Mater. 2011, 21, 4167– 4172.
19. Wang, Z.-S.; Kawauchi, H.; Kashima, T.; Arakawa, H. Sig-nificant Influence of TiO2Photoelectrode Morphology on
the Energy Conversion Efficiency of N719 Dye-Sensitized Solar Cell. Coord. Chem. Rev. 2004, 248, 1381–1389. 20. Dürr, M.; Schmid, A.; Obermaier, M.; Rosselli, S.; Yasuda, A.;
Nelles, G. Low-Temperature Fabrication of Dye-Sensitized Solar Cells by Transfer of Composite Porous Layers. Nat. Mater. 2005, 4, 607–611.
21. Hore, S.; Nitz, P.; Vetter, C.; Prahl, C.; Niggemann, M.; Kern, R. Scattering Spherical Voids in Nanocrystalline TiO2
-Enhancement of Efficiency in Dye-Sensitized Solar Cells. Chem. Commun. 2005, 2011–2013.
22. Koo, H.-J.; Kim, Y. J.; Lee, Y. H.; Lee, W. I.; Kim, K.; Park, N.-G. Nano-Embossed Hollow Spherical TiO2 as Bifunctional
Material for High-Efficiency Dye-Sensitized Solar Cells. Adv. Mater. 2008, 20, 195–199.
23. Han, S.-H.; Lee, S.; Shin, H.; Jung, H. S. A Quasi-Inverse Opal Layer Based on Highly Crystalline TiO2Nanoparticles: A
New Light-Scattering Layer in Dye-Sensitized Solar Cells. Adv. Energy Mater. 2011, 1, 546–550.
24. Liao, J.-Y.; Lei, B.-X.; Kuang, D.-B.; Su, C.-Y. Tri-functional Hierarchical TiO2Spheres Consisting of Anatase Nanorods
and Nanoparticles for High Efficiency Dye-Sensitized Solar Cells. Energy Environ. Sci. 2011, 4, 4079–4085.
25. Chen, D.; Huang, F.; Cheng, Y.-B.; Caruso, R. A. Mesoporous Anatase TiO2Beads with High Surface Areas and
Control-lable Pore Sizes: A Superior Candidate for High-Perfor-mance Dye-Sensitized Solar Cells. Adv. Mater. 2009, 21, 2206–2210.
26. Huang, F.; Chen, D.; Zhang, X. L.; Caruso, R. A.; Cheng, Y.-B. Dual-Function Scattering Layer of Submicrometer-Sized Mesoporous TiO2Beads for High-Efficiency Dye-Sensitized
Solar Cells. Adv. Funct. Mater. 2010, 20, 1301–1305. 27. Sauvage, F.; Chen, D.; Comte, P.; Huang, F.; Heiniger, L.-P.;
Cheng, Y.-B.; Caruso, R. A.; Gräetzel, M. Dye-Sensitized Solar Cells Employing a Single Film of Mesoporous TiO2
Beads Achieve Power Conversion Efficiencies over 10%. ACS Nano 2010, 4, 4420–4425.
28. Ghadiri, E.; Taghavinia, N.; Zakeeruddin, S. M.; Grätzel, M.; Moser, J.-E. Enhanced Electron Collection Efficiency in Dye-Sensitized Solar Cells Based on Nanostructured TiO2
Hollow Fibers. Nano Lett. 2010, 10, 1632–1638. 29. Hwang, K.-J.; Cho, D. W.; Lee, J.-W.; Im, C. Preparation of
Nanoporous TiO2Electrodes Using Different
Mesostruc-tured Silica Templates and Improvement of the Photo-voltaic Properties of DSSCs. New J. Chem. 2012, 36, 2096– 2100.
30. Chen, C.-C.; Chung, H.-W.; Chen, C.-H.; Lu, H.-P.; Lan, C.-M.; Chen, S.-F.; Luo, L.-Y.; Hung, C.-S.; Diau, E. W.-G. Fabrication and Characterization of Anodic Titanium Oxide Nanotube Arrays of Controlled Length for Highly Efficient Dye-Sensitized Solar Cells. J. Phys. Chem. C 2008, 112, 19151– 19157.
31. Li, L.-L.; Tsai, C.-Y.; Wu, H.-P.; Chen, C.-C.; Diau, E. W.-G. Fabrication of Long TiO2Nanotube Arrays in a Short Time
Using a Hybrid Anodic Method for Highly Efficient Dye-Sensitized Solar Cells. J. Mater. Chem. 2010, 20, 2753– 2758.
32. Li, L.-L.; Chen, Y.-J.; Wu, H.-P.; Wang, N. S.; Diau, E. W.-G. Detachment and Transfer of Ordered TiO2 Nanotube
Arrays for Front-Illuminated Dye-Sensitized Solar Cells. Energy Environ. Sci. 2011, 4, 3420–3425.
33. Enache-Pommer, E.; Liu, B.; Aydil, E. S. Electron Transport and Recombination in Dye-Sensitized Solar Cells Made from Single-Crystal Rutile TiO2Nanowires. Phys. Chem.
Chem. Phys. 2009, 11, 9648–9652.
34. Zhang, H.; Yu, H.; Han, Y.; Liu, P.; Zhang, S.; Wang, P.; Cheng, Y.; Zhao, H. Rutile TiO2Microspheres with Exposed
Nano-Acicular Single Crystals for Dye-Sensitized Solar Cells. Nano Res. 2011, 4, 938–947.
35. Sheng, J.; Hu, L.; Li, W.; Mo, L.; Tian, H.; Dai, S. Formation of Single-Crystalline Rutile TiO2Splitting Microspheres for
Dye-Sensitized Solar Cells. Sol. Energy 2011, 85, 2697– 2703.
36. Hore, S.; Palomares, E.; Smit, H.; Bakker, N. J.; Comte, P.; Liska, P.; Thampi, K. R.; Kroon, J. M.; Hinsch, A.; Durrant, J. R. Acid versus Base Peptization of Mesoporous Nanocrystal-line TiO2Films: Functional Studies in Dye Sensitized Solar
Cells. J. Mater. Chem. 2005, 15, 412–418.
37. Nazeeruddin, M. K.; Kay, A.; Rodicio, I.; Humphry-Baker, R.; Müller, E.; Liska, P.; Vlachopoulos, N.; Grätzel, M. Conver-sion of Light to Electricity by cis-X2bis(2,20-bipyridyl-4,40 -dicarboxylate)ruthenium(II) Charge-Transfer Sensitizers (X = Cl, Br, I, CN, and SCN) on Nanocrystalline Titanium Dioxide Electrodes. J. Am. Chem. Soc. 1993, 115, 6382– 6390.
38. Nazeeruddin, M. K.; De Angelis, F.; Fantacci, S.; Selloni, A.; Viscardi, G.; Liska, P.; Ito, S.; Takeru, B.; Grätzel, M. Com-bined Experimental and DFT-TDDFT Computational Study
of Photoelectrochemical Cell Ruthenium Sensitizers. J. Am. Chem. Soc. 2005, 127, 16835–16847.
39. Wang, Q.; Ito, S.; Grätzel, M.; Fabregat-Santiago, F.; Mora-Seró, I.; Bisquert, J.; Bessho, T.; Imai, H. Characteristics of High Efficiency Dye-Sensitized Solar Cells. J. Phys. Chem. B 2006, 110, 25210–25221.
40. Ito, S.; Murakami, T. N.; Comte, P.; Liska, P.; Grätzel, C.; Nazeeruddin, M. K.; Grätzel, M. Fabrication of Thin Film Dye Sensitized Solar Cells with Solar to Electric Power Conver-sion Efficiency over 10%. Thin Solid Films 2008, 516, 4613– 4619.
41. Burnside, S. D.; Shklover, V.; Barbé, C.; Comte, P.; Arendse, F.; Brooks, K.; Grätzel, M. Self-Organization of TiO2
Nano-particles in Thin Films. Chem. Mater. 1998, 10, 2419–2425. 42. Penn, R. L.; Banfield, J. F. Morphology Development and Crystal Growth in Nanocrystalline Aggregates under Hy-drothermal Conditions: Insights from Titania. Cosmochim. Acta 1999, 63, 1549–1557.
43. Neale, N. R.; Frank, A. J. Size and Shape Control of Nanocrystallites in Mesoporous TiO2Films. J. Mater. Chem.
2007, 17, 3216–3221.
44. Wen, P.; Itoh, H.; Tang, W.; Feng, Q. Single Nanocrystals of Anatase-Type TiO2Prepared from Layered Titanate
Na-nosheets: Formation Mechanism and Characterization of Surface Properties. Langmuir 2007, 23, 11782–11790. 45. Jun, Y.-W.; Casula, M. F.; Sim, J.-H.; Kim, S. Y.; Cheon, J.;
Alivisatos, A. P. Surfactant-Assisted Elimination of a High Energy Facet as a Means of Controlling the Shapes of TiO2
Nanocrystals. J. Am. Chem. Soc. 2003, 125, 15981–15985. 46. Wu, B.; Guo, C.; Zheng, N.; Xie, Z.; Stucky, G. D. Nonaqueous
Production of Nanostructured Anatase with High-Energy Facets. J. Am. Chem. Soc. 2008, 130, 17563–17567. 47. Dinh, C.-T.; Nguyen, T.-D.; Kleitz, F.; Do, T.-O.
Shape-Controlled Synthesis of Highly Crystalline Titania Nano-crystals. ACS Nano 2009, 3, 3737–3743.
48. Yang, H. G.; Sun, C. H.; Qiao, S. Z.; Zou, J.; Liu, G.; Smith, S. C.; Cheng, H. M.; Lu, G. Q. Anatase TiO2Single Crystals with a
Large Percentage of Reactive Facets. Nature 2008, 453, 638–641.
49. Fang, W. Q.; Gong, X.-Q.; Yang, H. G. On the Unusual Properties of Anatase TiO2Exposed by Highly Reactive
Facets. J. Phys. Chem. Lett. 2011, 2, 725–734.
50. Sugimoto, T.; Zhou, X.; Muramatsu, A. Synthesis of Uniform Anatase TiO2Nanoparticles by GelSol Method: 3.
For-mation Process and Size Control. J. Colloid Interface Sci. 2003, 259, 43–52.
51. Sugimoto, T.; Zhou, X.; Muramatsu, A. Synthesis of Uniform Anatase TiO2Nanoparticles by GelSol Method: 4. Shape
Control. J. Colloid Interface Sci. 2003, 259, 53–61. 52. Sugimoto, T.; Okada, K.; Itoh, H. Synthesis of Uniform
Spindle-Type Titania Particles by the GelSol Method. J. Colloid Interface Sci. 1997, 193, 140–143.
53. Yun, H. J.; Lee, H.; Joo, J. B.; Kim, W.; Yi, J. Influence of Aspect Ratio of TiO2Nanorods on the Photocatalytic
Decomposi-tion of Formic Acid. J. Phys. Chem. C 2009, 113, 3050–3055. 54. Bockmeyer, M.; Löbmann, P. Formation, Densification and Properties of SolGel TiO2Films Prepared from
Trietha-nolamine-Chelated Soluble Precursor Powders. J. SolGel Sci. Technol. 2008, 45, 251–259.
55. Lan, C.-M.; Liu, S.-E.; Shiu, J.-W.; Hu, J.-Y.; Lin, M.-H.; Diau, E. W.-G. Formation of Size-Tunable Dandelion-like Hier-archical Rutile Titania Nanospheres for Dye-Sensitized Solar Cells, unpublished results.
56. Zheng, Y.; Shi, E.; Chen, Z.; Li, W.; Hu, X. Influence of Solution Concentration on the Hydrothermal Preparation of Titania Crystallites. J. Mater. Chem. 2001, 11, 1547–1551. 57. Li, L.-L.; Chang, Y.-C.; Wu, H.-P.; Diau E. W.-G. Characteriza-tion of Electron Transport and Charge RecombinaCharacteriza-tion Using Temporally Resolved and Frequency-Domain Tech-niques for Dye-Sensitized Solar Cells. Int. Rev. Phys. Chem. 2012, DOI: 10.1080/0144235X.2012.733539.
58. The electron transport and charge recombination proper-ties of the devices are related to the structure of the TiO2
films. To investigate the intrinsic kinetic effect of TiO2on
photovoltaic performance, the transient photoelectric
and charge-extraction measurements were also per-formed for the NP and HD1 devices fabricated with similar film thickness (1213 μm) but without adding a scattering layer and TiCl4post-treatment. The transient photocurrent
decay data (Figure S11) show similar kinetic behavior as shown in Figure 6, confirming the superior electron transport ability of the HD1film over that of the NP film as an active layer for DSSCs.
59. The effects of potential shift of TiO2conduction band edge
and charge recombination in the dye/TiO2/electrolyte
interface for the same NP and HD devices as mensioned in ref 58 were also investigated. The charge-extraction data (Figure S12a) exhibit a potential shift of∼ 30 mV between the HD1 and the NPfilms, which is consistent with the data shown in Figure 7a with added SL and TiCl4
post-treatment. The transient photovoltage decay data (Figure S12b) show a similar behavior of charge recombi-nation between the NP and HD1 devices, confirming that the discrepancy of VOCbetween the two devices is due to
the position of the TiO2potential band edge discussed in
the text.
60. Ito, S.; Chen, P.; Comte, P.; Nazeeruddin, M. K.; Liska, P.; Pechy, P.; Grätzel, M. Fabrication of Screen-Printing Pastes from TiO2Powders for Dye-Sensitised Solar Cells. Prog.
Photovoltaics 2007, 15, 603–612.
