E I S CHARACTERIZATION

Transcription

1 147 Chapter V E I S CHARACTERIZATION Objective of this chapter To study the impedance characteristics of DSSCs 5.1. Introduction The steady-state J V characteristic of DSSCs provide the performance parameters of the cell, such as the short-circuit current (Jsc), open-circuit potential (Voc), fill factor (FF) and photocurrent efficiency (η). Nevertheless, more detailed properties cannot be extracted from steady-state measurements, and so dynamic techniques should be considered. These techniques allow the interpretation of the charge-transfer kinetics, mainly characterized by diffusion coefficients and lifetime of the different charge carriers. One of the most powerful characterization techniques of DSSCs involving transient probing is electrochemical impedance spectroscopy (EIS). EIS is a dynamic technique that has many advantages, not only because it is user-friendly but also for the reason that of its sensitivity and ability to separate different complex processes occurring in DSSCs (Wang, et al., 2005; Bisquert, 2002; Fabregat-Santiago, et al., 2002). Despite being a relatively easy method to apply to the study of electrochemical systems, the results may sometimes be difficult to interpret. The use of equivalent electrical analogs to fit the EIS experimental data is a useful tool in this respect, as it helps to identify and interpret the characteristic parameters of the system, such as, for instance, the internal resistances of DSSCs Electrochemical impedance spectroscopy EIS is a powerful technique for the characterization of electrochemical systems. It is one of the most useful experimental techniques as it permits a

2 148 simultaneous characterization of the different processes taking place in the cell (Bisquert and Mora-Ser, 2010; Walker, et al., 2006). In EIS measurements the potential applied to the solar cell is perturbed by a small amplitude sinusoidal modulation and the resulting sinusoidal current response is measured as a function of the modulation frequency (Oekermann, et al., 2005). The meaning of electrical impedance can be understood starting from the concept of resistance. The electrical resistance is the ability of a circuit element to resist the flow of electrical current. The well known Ohm's law defines resistance (R) in terms of the ratio between voltage (V) and current (I). This relationship is limited to only one circuit element, an ideal resistor. But usually the systems under study contain circuit elements that exhibit a much more complex behaviour. The simple concept of resistance needs to be replaced by a more general parameter: the impedance, which includes not only the relative amplitudes of the voltage and the current, but also the relative phases. Like resistance, impedance is a measure of the ability of a circuit to resist the flow of electrical current (Villanueva-Cab, et al., 2010). The excitation signal can be written as V(ω) = V 0 cos ωt where V(ω) is the ac potential applied to the system, V 0 is the amplitude signal and ω is the angular frequency (ω = 2πf rad s -1 ). The current response will be shifted with respect to the applied potential I(ω) = I 0 cos (ωt+ ) where I(ω) is the ac electrical current response signal, I 0 the amplitude and the phase shift. The phase factor contains the current lag with respect to the voltage. An expression analogous to Ohm s law allows for calculating the impedance of the system as

3 149 Z V I V cos t I cos t 0 0 Usually, it is convenient to use complex exponentials to express the impedance. Complex numbers allow for a simpler representation of the relative magnitude and phase of the input and output signal. Besides, it is a more powerful representation for circuit analysis purposes. Taking into account Euler s relationship, e jx cos x j sin x it is possible to express the potential as V V 0 e j t and the current response can be described as I I j t 0 e Since Z(ω)=V(ω)/I(ω) the exponential exp (jωt) terms cancel out, so that Z V I 0 0 e j Z 0 e j The impedance is, therefore, expressed in terms of a magnitude Z 0 (ratio of the voltage amplitude to the current amplitude) and a phase shift. Using Equation 5.4, it is possible to separate the real part and the imaginary part of the impedance Z 0 e -jφ = Z 0 cos (φ) - Z 0 j sin (φ) Real Imaginary By varying the frequency of the applied signal, one can get the impedance of the system as a function of frequency. The recorded data can either be represented as magnitude and phase vs. frequency (Bode plot) or on a complex plane (Nyquist plot).

4 150 In a Nyquist plot the real part of the impedance (Z') is plotted on the X axis and the imaginary part (Z ) is plotted on the Y-axis. EIS data are commonly analyzed in terms of an equivalent circuit model. Most of the circuit elements in the model are common electrical elements such as resistors, capacitors and inductors. These elements can be combined in series and in parallel to give complex equivalent circuits. The most common cell model for electrochemical systems is the Randles Circuit (Fig.5.1). This circuit models a cell where polarization is due to a combination of kinetic and diffusion processes. It includes a series resistance (R s ), a double layer capacitor (C rec ), a charge transfer resistance (R rec ), capacitance (C ct ) at electrolyte/ce interface, electron transport resistance (R ct ) at electrolyte/ce interface and Warburg diffusion impedance (Z d ). Fig.5.1: EIS equivalent circuit diagram The Nyquist plot for this circuit is shown in Fig.5.2. The double layer capacitance and the charge transfer resistance in parallel define the time constant (τ rec = RC) or electron life time of the system (O Regan and Durrant, 2009). The Nyquist plot for a RC combination is always a semicircle as the one indicated in Fig.5.2. The series resistance is expressed by the real intercept at high frequencies of this semicircle. Finally, the Warburg diffusion impedance appears as a straight line with a slope of 45 (Fig.5.2).

5 151 Fig.5.2: Transmission line model with Nyquist plot A dye-sensitized solar cell is a very complex system and its impedance response will be related to the response of the different components of the device. Generally, a transmission line model is used to describe the system, as shown in Fig.5.2. Different circuit elements are attributed to the different processes taking place in the cell. The processes occurring in the mesoporous oxide film are usually modeled by a diffusion-recombination transmission line. Its application to DSSCs was first proposed by Bisquert (Peter, 2009). This transmission line is composed of a network of resistive and capacitive elements, which describe the transport and interfacial transfer of electrons that take place in the oxide. Fig.5.3 shows the model Bode and frequency phase plots of DSSC. The mesoporous oxide film has two main contributions in the impedance spectrum. One feature is the intermediate frequency arc, which accounts for the parallel connection of the charge transfer resistance R rec and the capacitance of the

6 152 film C rec. The other feature is a Warburg-like diffusion element, a 45º phase shift at high frequencies related to the electron transport resistance R t in the mesoporous layer. The diffusion impedance of redox species in the electrolyte (Z d ) is usually modeled using a finite-length Warburg element of the type used for thin layer electrochemical cells. Generally, Z d in DSSC is small and it is difficult to identify it in the overall impedance solar cell response. However, it is clearly visible when electrolytes based on ionic liquids are used. In that case, a semicircle at low frequencies can be distinguished in the spectra. Fig.5.3: Impedance Vs frequency and phase Vs frequency plots A cathodic impedance due to the counter electrode is also present in the spectra and it is expressed as a parallel combination of the electron transfer resistance R ct (characterizing electron transfer to and from the redox system) and a capacitance, C ct. In the Nyquist spectra, this feature is represented by a semicircle at high frequencies. Finally, R s accounts for the series resistance of the conducting glass plus any other element that might be considered to be in series with the rest of the circuit.

7 153 The other useful electrochemical parameters of DSSCs are k eff (rate of recombination of the electrons in the photoanode film), τ rec (effective electron life time), L n (electron diffusion length), D eff (effective diffusion coefficient), µ (electron mobility), σ (electron conductivity), η cc (electron collection efficiency) and concentration of electrons in the TiO 2 /dye/electrolyte interface. These parameters are calculated from the following expressions (Adachi, et al., 2006). The lifetime of electrons in DSSC is a particularly important quantity. The measured lifetime is a strong function of the Fermi level or open-circuit photo voltage, Voc. The lifetime is a kinetic quantity that contains information not only on the rate constants of charge transfer but also on the distribution of electronic states and electronic transitions that intervene in the operation of the DSSC. rec s R rec C rec Eqn.(5.1) s Rct Cct Eqn.(5.2) d Where τ d is the electron transport time. k eff is the effective rate constant for recombination of charge carriers at photoanode and electrolyte interface. k eff is estimated from the peak frequency ω max of the central arc, which describes the rate of recombination of electron in the film. k eff (s -1 ) Eqn.(5.3) rec D eff is the effective diffusion coefficient of electrons is calculated as, eff 2 2 L cm s F rec D Eqn.(5.4)

8 154 The diffusion length L n is the distance that an electron travels through the film before it is recombined to the acceptor L n 1 m D Eqn.(5.5) rec eff Conductivity of the photoanode film D 2 eff cm V s D Eqn.(5.6) eff K T B e Electron mobility L F ms cm Eqn.(5.7) S 1 P R rec Electron collection efficiency d cc Eqn.(5.8) rec Concentration of charge carriers describes the charge transport resistance and its recombination rate along the entire thickness of working electrode and is calculated as Conc. cm s RrecLFkeff Eqn.(5.9) where, L F - the film thickness of the photoanode (15 µm); S-active area of the DSSCs; P-porosity of the photoanode film (0.6); (R ct x C rec ); e Charge of an electron; K B - Boltzmann constant; T Room temperature The electrochemical impedance spectra were fitted using the Zsimpwin software by means of equivalent circuit shown in Fig.5.1.

9 Experimental Electron impedance spectra of DSSCs were recorded with Gamry-300 impedance analyzer. The applied bias voltage and ac amplitude were set at opencircuit voltage of the DSSCs at 10 mv between the ITO/CE at the ITO/TiO 2 /dye working electrode, respectively, and the frequency range explored was 1 mhz to 10 5 Hz. The impedance spectra were analyzed by an equivalent circuit model interpreting the characteristics of the DSSCs through Zsimpwin software ( Fig.5.4 shows the Gamry impedance analyzer. Fig.5.4: Gamry impedance analyzer (Impedance analyzer (EIS) EIS 300 specifications) Overview The EIS 300 Electrochemical Impedance Spectroscopy allows us to measure impedances spanning 15 orders of magnitude over a frequency range of 10 µhz to 1 MHz. Data analysis in the Echem Analyst is easy. Spectra can be fit to standard cell models or custom models created with graphical Model Editor. Gamry EIS 300 data files ZSimpWin and Equivalent Circuit. Data Acquisition Features Continuous AC amplitude and DC offset correction for the most accurate potential and current measurement.

10 156 Lissajous (I vs. E) real-time display at each frequency. Bode or Nyquist real-time display of spectrum as it develops. Realtime monitoring of signal-to-noise ratio for deciding when a measurement is done for the best data in the shortest time. Data Analysis Features Impedance or Admittance plots. Graphic Model Editor for equivalent circuit models with the ability to define custom components. Choice of Levenberg-Marquardt or Simplex algorithms in fitting the data to equivalent circuits. Kramers-Kronig Transform to evaluate the quality of the EIS data. A unique Auto-Fit routine takes the guesswork out of estimating initial values for your model elements. System Requirements Gamry PCI4/Series G, Reference, or Interface Family Instrument Microsoft Windows XP SP3/Vista/7/8 (32-bit or 64-bit)

11 Results and discussion Electron transport properties of DSSCs were investigated using electrochemical impedance spectroscopy (EIS). In general, the Nyquist plot exhibited three semicircles, which are attributed to the redox reaction at the counter electrode in the high-frequency region, the electron transfer at the TiO 2 /dye/electrolyte interface in the middle-frequency region and carrier transport by ions within the electrolytes in the low-frequency region. The model equivalent circuit diagram (shown in Fig. 5.1) consisted of a series of resistance (R s, starting point of the first semicircle in Nyquist plot), electron transport resistance at the counter electrode/electrolyte (R ct, first semicircle in Nyquist plot), charge transfer resistance at the photoanode/electrolyte (R rec, second semicircle in Nyquist plot), the constant phase elements of capacitance C rec (R rec ), C ct (R ct ) and Warburg diffusion impedance Z d. The charge transfer kinetics has been explained on the basis of the diffusion recombination model proposed by Bisquert, 2002 and has been frequently employed to analyze similar structures (Bisquert, 2002; Tan and Wu, 2006 ; Wu et al., 2008). The internal impedances R s, R rec, R ct, Z d were determined from EIS analysis and listed in Tables 5.1, 5.3 and 5.5. The other useful electrochemical parameters of DSSCs, k eff (rate of recombination of the electrons in the film), τ eff (effective electron life time), L n (electron diffusion length), D eff (effective diffusion coefficient), µ (electron mobility), σ (electron conductivity), η cc (electron collection efficiency) and concentration of electrons in the TiO 2 /dye/electrolyte interface were calculated using Eqns. (5.1)-(5.9) and presented in Tables 5.1, 5.3 and EIS characterization of DSSCs made with core-shell photoanode Fig. 5.5 shows the Nyquist plots of DSSCs made with different types photoanodes and electrolytes. From the Nyquist plots it could be seen that first and third semicircles are weak when compared the second semicircle. From the Fig. 5.5,

12 158 comparison of the middle semicircles of Nyquist plot of the DSSCs made with different electrolytes indicated the decrease of the diameter in the order DSSC-1>DSSC-3>DSSC-2>DSSC-4, suggested that DSSC-4 contributed to the lowest charge transfer resistance at the TiO 2 /dye/electrolyte interface. When comparing the Nyquist plots of DSSC-4 and DSSC-5, the R rec value is increased in DSSC-5 due to uncoated ZrO 2 photoanode. Similarly when comparing the Nyquist plots of DSSC-4 and DSSC-6, the R rec value is increased in DSSC-6 due to uncoated blocking layer HfO 2. The DSSC-7 showed the highest value of R rec. In this study, higher charge collection efficiency is achieved in DSSC-4 based on TNPWs/ZrO 2 core-shell with HfO 2 blocking layer structure photoanode through quasi-solid state electrolyte as compared to the other cells. From the Table 5.1 it could be observed that the DSSC-4 made with novel photoanode strucuture illustrated higher values of C rec, τ rec, L n, D eff, µ, σ and η cc as comparing other DSSCs. In addition, k eff and concentration of electrons in the TiO 2 /dye/electrolyte interface are lower values as compared to the other DSSCs. Control of interfacial electron injection and recombination at the photoanode/dye/electrolyte interface is pivotal for the best performance of DSSCs. Modification at the metal oxide/dye interface has received increased attention among researchers to improve the injection efficiency and suppress the recombination as the interfacial energetic can be controlled by this strategy. Introducing shell layer on the surface of photoanode materials exerts larger dipole moment that changes the interface energetic in an energetically more favourable way for electron transfer (Watson and Meyer, 2004). Recombination rate was diminished because the insertion of ZrO 2 could shield electron back flow through TNPWs to the electrolyte. In the present investigation, the surface modification of photoanode with very thin layer as well as larger band gap shell material improved the photocurrent.

13 159 Hence, core-shell material provided faster electron injection, and suppressed recombination at photoanode/dye/electrolyte interface. In our examination on the performance of core shell structure-based DSSCs hints that the essential improvement in electron injection efficiency and thereby overall performance can be obtained by proper designing of the second metal oxide with suitable band gap, optimal thickness of shell layer and IEP. In addition, one-dimensional (1D) nanostructure, the grain boundaries effect could be restricted (Fujihara et al., 2007, Baxter and Aydil, 2006). Moreover, the loading of dye can be much higher in 1D nano structured material than the nanoparticles, for instance, the TNPWs has allowed for larger adsorption of dye. These results were indicated that the combining TNWs with TNPs having relatively high thermal stability and low aggregation property. The enhanced charge collection efficiency of the TNPWs/ZrO 2 core-shell DSSCs is achieved by optimizing device architectures, which enhanced light absorption and facilitated electron transport by determining and designing appropriate dimensions of TNPWs/ZrO 2, by optimizing the cation concentration in the electrolyte solution for promoting electron injection yield from sensitizing D149 dye molecules to core-shell TNPWs/ZrO 2 electrodes, or by synthesizing thermally stable TNPWs with a stable high surface area. In the coreshell TNPWs/ZrO 2 photoanode, the lower R rec suggested that each TNWs made more difficult for an electron to jump outside the nanostructure than to stay within the structure during diffusion; this explained the high charge collectin collection efficiency and thus high short-circuit photocurrent.

14 160 Fig.5.5: Nyquist plots of DSSCs 1-7 The core-shell TNPWs/ZrO 2 photoanode exhibited different electrochemical behaviours when they are employed in dye-sensitized solar cells. In this photoelectrochemical system, charge separation, transport and recombination strongly depend on the nanostructure and core-shell structure of the photoanode. From EIS data it could be concluded that for the cell containing core-shell TNPWs/ZrO 2, formation of a space charge layer at the surface of the electrode effectively promoted the separation of photo generated charge carriers and prevented recombination of the electron with the hole carriers. The uninterrupted long pathway with fewer boundaries is beneficial for transport of electron toward the conducting substrate. Simultaneously, the decrease of trap sites in the band gap and the surface state can effective enhance the charge-diffusion coefficient and suppression of the interfacial charge transfer (Qi et al., 2010). In addition, the holes transport to electrolyte via the different channel, which could prevented the charge recombination at the core-shell

15 161 TNPWs/ZrO 2 /redox electrolyte interfaces and led to the electrons diffusion and transport become easy. Therefore, the DSSCs based on the core-shell TNPWs/ZrO 2 film with QSSE has the lower values for R rec, R ct and the longer electron lifetime, consequently led to an enhanced charge collection efficiency of the cell. Electron transport in core-shell TNPWs/ZrO 2 photoanode is tortuous and often confront with dead end (Hagfeldt et al., 2010), which enormously increased the electron transport route. As revealed from Table 5.1, the presence of core-shell TNPWs/ZrO 2 layer increased the electron lifetime and made the electron transfer more easily. In other words, the charge transfer resistance in core-shell TNPWs/ZrO 2 interfaces and electron recombination are decreased, led to a positive influence on the improvement of solar cell performance. Hence with EIS investigation, it is observed that the enhanced electron transport properties, reduced charge recombination and increased electron life-time which are imperative for higher solar cell performance in DSSC-4. Therefore, DSSCs fabricated with core-shell TNPWs/ZrO 2, which provided direct and enhanced charge transport while minimizing charge recombination resulted in enhanced the charge collection efficiency. Use of core-shell nanostructure to lower the charge recombination is based on a hypothesis that a coating layer might build up an energy barrier at the semiconductor/electrolyte interface and, thus, retarded the reaction between the photogenerated electrons and the redox species in electrolyte. In core-shell nanostructure, an energy barrier is established at the semiconductor/electrolyte interface, it required that the conduction band potential of the shell material is more negative than that of the core material. As such, the back electron transport is retarded and the interfacial recombination could be reduced (Kanmani and Ramachadran, 2012). The D e and τ rec values of the DSSC-5 with Hafnium oxide layer are shown to be higher than those of DSSC-7 built with TNPWs

16 162 bare photoanode. The higher D e means that the TNPWs film becomes more conductive owing to the increased electron density in the conduction band by introducing Hafnium oxide barrier. Table 5.1: EIS parameters of DSSCs fabricated using core-shell photoanode structure DSSCs R rec Ω R ct Ω C rec µf C ct µf Z d Ω- s 0.5 k eff s -1 τ rec ms τ d ms L n µm D eff 10-6 cm 2 /s µ 10-4 cm 2 /V /s σ 10-6 S/cm η cc % Con c. Ωcm s -1 DSSC DSSC DSSC DSSC DSSC DSSC DSSC The increased τ rec value in DSSC-5 by introducing Hafnium oxide might be due to the Hafnium oxide layer formed on the TNPWs surface could effectively retarded the electron recombination process between photoanode and electrolyte. Hence, the application of HfO 2 as an efficient blocking layer material to modify the core-shell TNPWs/ZrO 2 electrodes in DSSCs. It is due to the more conductive coreshell TNPWs/ZrO 2 electrodes film (that is, the more accumulation of the electrons in the conduction band of the TiO 2 film) and the reduced recombination rate at the interface between the TNPWs/ZrO 2 and dye/electrolyte. In the present investigation, DSSC-4 with quasi-solid state electrolyte have relatively high ambient ionic conductivity, intimate interfacial contact with core-shell TNPWs/ZrO 2, and remarkable stability which led to higher efficiency as compared to the other electrolytes.

17 163 The cell performance confirms that suitable coating of ZrO 2 shell layer with TNPWs core film can suppress the charge recombination process effectively and improve the overall conversion efficiency. Fig. 5.5, which suggests that the improvement in η cc and τ rec result from a suppression of the charge recombination by ZrO 2 shell coating. The role and the mechanism of the wide-band semiconductor as shell coating on TiO 2 have been discussed in many researches (Chen et al., 2001 and Kruger et al., 2000). The present study shows that the thin ZrO 2 shell coating can suppress the charge recombination. Since, ZrO 2 has a conduction band more negative than that of TiO 2, ZrO 2 shell coating on TNPWs surface can form an energy barrier on the surface of photoanode, which can suppress the injected electrons from recombination with oxidized dye or I - 3, thus reduces the charge recombination process and increases the conversion efficiency (Chen et al., 2001; Kruger et al., 2000; Burnside et al., 1999; Lenzmann et al., 2001; Vilan et al., 2000). Generally, it can be found that reasonable thickness of the shell layer is beneficial to suppress the recombination process and achieve relative high conversion efficiency. Fig.5.5 shows EIS of the cells of DSSC-4 and 5 with and without ZrO2 shell coating. The result also confirms that DSSC with ZrO2 shell coating has less electron recombination at the photoanode/dye/electrolyte interface and longer electron lifetime compared with the uncoated DSSC-5. Thus the application of ZrO2 shell coating on TNPWs composite photoanode can promote the interfacial electron transfer and suppress the electron recombination effectively. From the frequency - phase plots in the lower frequency regime, it can be seen that the characteristic frequency peak of DSSC-4 shifted to lower frequency as compared to other DSSCs. The peak shifted to a lower frequency indicates an increased lifetime of the DSSC-4. The longer lifetime implied a lower recombination rate and enhanced charge collection efficiency. Therefore the interfacial charge recombination of TNPs between the photo-injected electron and electrolyte materials

18 164 reduced after the TNWs decoration/shell layer on the core TNPWs. This consequently leads to a significantly enhanced energy conversion efficiency of the cell. The electron life time was calculated to be 32 msfor the DSSC-4, which is highest among other DSSCs. The characteristic frequency peak of DSSC-4 in higher frequency regime is shifted to lower frequency region when comparing to other DSSCs. This result could be attributed to the higher electron transport rate, faster electron diffusion coefficient and large surface area of core-shell structure photoanode in the DSSC-4. These results indicate that the DSSC-4 built with coreshell photoanode through QSSE has fast electrons transport and lower recombination rate at TiO 2 /dye/electrolyte interface and the same property leading to higher charge collection efficiency. Fig. 5.7 shows the Bode plots of DSSCs 1-7. Fig. 5.7 indicates that the DSSC-4 has lower impedance at electrolyte/counter electrode interface and photoanode/dye/electrolyte interface as compared to other DSSC. This might be to higher electrons transfer & lower recombination of charge carriers at photoanode/dye/electrolyte interface and faster electron transport at redox couple electrolyte/counter electrolyte interface. Hence in the present investigation, it could be concluded that, in the core-shell DSSCs, the electrons need to diffuse several micrometers into the TNPWs layer (core) surrounded by electron acceptor (shell) at a distance of only several nanometers. The mesoporous structure of the TNPWs layer provided a large surface area, allowing absorption of enough D149 dye molecules to achieve significant charge collection efficiency. Core-shell photoanode consisting of a mesoporous TNPWs covered with a shell of ZrO 2 slows the recombination processes by the formation of an energy barrier at the TiO 2 surface.

19 165 Fig.5.6: Frequency Vs phase plots of DSSCs 1-7 Fig.5.7: Bode plots of DSSCs 1-7

20 166 Table 5.2: Frequency and phase of DSSCs fabricated using core-shell photoanode structure DSSCs f 1 (KHz) f 2 (Hz) υ1 (Hz) PEAK 1 PEAK 2 θ1 (dec) υ2 (KHz) θ2 (dec) DSSC DSSC DSSC DSSC DSSC DSSC DSSC The conduction band potential of the shell should be more negative than that of TiO 2 in order to generate an energy barrier for the reaction of the electrons present in TiO 2 with the oxidized dye or the redox mediator in solution. An energy barrier forms not only at the electrode/electrolyte interface but also between the individual TiO 2 nanoparticles. The TiO 2 nanoparticles are connected directly to each other allowing electron transport through TiO 2. The shell having more negative conduction band potential acted as an energy barrier that slowed recombination reactions. Photo excitation of dye molecules anchored to ultra thin outer shell of insulators or semiconductors on n-type semiconductor crystallites resulted in electron transfer to the inner core material.

21 EIS characterization of DSSCs made with doping photoanode Electron transport properties of DSSCs were investigated by EIS. Fig. 5.8 shows the Nyquist plots of DSSCs fabricated with different types photoanodes and electrolytes. The internal impedances R s, R rec, R ct, Z d were determined from the Nyquist plot and listed in Table 5.3. The other useful electrochemical parameters of DSSCs, k eff (rate of recombination of the electrons in the film), τ rec (effective electron life time), τ d (electron transport time), L n (electron diffusion length), D eff (effective diffusion coefficient), µ (electron mobility), σ (electron conductivity), η cc (electron collection efficiency) and concentration of electrons in the TiO 2 /dye/electrolyte interface was calculated and presented in Table 5.3. From the Nyquist plots of DSSCs, it could be seen that first and third semicircles are weak as compared the second semicircle. From the Fig. 5.8, comparison of the middle semicircles of Nyquist plots of the DSSCs made with different electrolytes indicates the decreased of the diameter in the order DSSC-8>DSSC-10>DSSC-9>DSSC-11. This observation suggested that DSSC-11 contributed to the lowest charge transfer resistance at the photoanode/dye/electrolyte interface as a result of the quasi-solid state electrolyte. When comparing the Nyquist plots of DSSC-11 and DSSC-12, the R rec value is increased in DSSC-12 because of undoped Al photoanode. Likewise, when comparing the Nyquist plots of DSSC-11 and DSSC-13, the R rec value is increased in DSSC-13 due to the absence of HfO 2 blocking layer in the photoanode. The DSSC-7 showed the highest value of R rec among the other DSSCs as a consequence of sluggish electron transport. In the EIS study, higher η cc was achieved in DSSC-11 fabricated based on Al-doped TNPWs with HfO 2 blocking layer structure photoanode through quasi solid state electrolyte as compared to the other cells. The DSSC-11 illustrated higher values of C rec, τ rec, L n, D eff, µ, σ and η cc as compared to other DSSCs. In addition, k eff and concentration of electrons in the photoanode/dye/electrolyte interface were lower values as compared

22 168 to the other DSSCs. Introducing cations as dopant on the surface of photoanode materials exerts larger dipole moment that changes the interface energetic in an actively more favourable way for electron transfer (Warson and Meyer, 2004). Consequently, Al-doped TNPWs photoanode provided faster electron injection at the photoanode/dye/electrolyte interface. Recombination rate was diminished because of HfO 2 layer could shield electron back flow through TNPWs to the electrolyte. In our examination on the performance of Al-doped TNPWs photoanode structure-based DSSCs hints that the essential improvement in electron injection efficiency and thereby overall performance could be obtained by proper designing of dopant with suitable band gap,optimal thickness of blacking layer. The improved photoelectrochemical performance of DSSCs is attributed to two main factors:(1) increased light harvesting efficiency caused by the large amount of sensitizing dye D149 adsorbed on the large surface area of Al-doped TNPWs composite and (2) increased electrical conductivity due to Al ions doped into the TiO 2 lattice at the divalent Ti 2+ site, allowing electrons to move easily into the Aldoped TiO 2 conduction band. The optical response of material is largely determined by its underlying electronic structure. The electronic properties of a material are closely related to its chemical composition, its atomic arrangement, and its physical dimension. The chemical composition of TiO 2 can be altered by doping. It is desirable to support the integrity of the crystal structure of the photocatalytic host material and to produce favourable changes in electronic structure. As the Fermi levels of Al are lower than that of TiO 2, photo-excited electrons can be transferred from conduction band (CB) of TiO 2 to Al band, while photo-generated valence band holes stay on the TiO 2. These activities greatly reduce the possibility of electron-hole recombination, resulting in proficient η cc (Patel et al., 2008; Wang and Caruso, 2011).

23 169 The presence of dopant metal can act as an electron sink and significantly decreased the electron transport time and improved mobility of photogenerated electrons (Carneiro et al., 2009). In DSSC-11, the lower R rec suggested that each TNW made more difficult for an electron to jump outside the nanostructure than to stay within the structure during diffusion; this explained the higher values of η cc and D eff. The D e and τ rec values of DSSC-13 with Hafnium oxide layer are shown to be higher than those of DSSC-7 built with bare electrode. The higher D e illustrated that the TNPWs film becomes more conductive owing to the increased electron density in the conduction band by introducing Hafnium oxide barrier. In addition, the elevated τ rec value of the by introducing Hafnium oxide also proved that the Hafnium oxide layer formed on the TNPWs surface could effectively retarded the electron recombination process between photoanode and electrolyte. Hence, the Al-doped TNPWs with blocking layer photoanode with QSSE exhibited enhanced photoelectrochemical behaviours in DSSC-11. In this photoelectrochemical system, charge separation, transport and recombination strongly depend on the nanostructure of the photoanode and nature of electrolytes. Therefore, the DSSCs based on the Al-doped TNPWs film has the lower values for R rec, R ct and the longer electron lifetime, led to enhanced η cc of the cell. In the present investigation, D149 dye is used as the efficient sensitizer for DSSCs. The Al-doped TNPWs photoanode demonstrated a successful photoelectrode within DSSCs which keep the desired specific surface area for dye-molecule adsorption and enough light-harvesting from prolonged light travelling. In addition, QSSE had relatively high ambient ionic conductivity, intimate interfacial contact with Al doped TNPWs-HfO 2, and remarkable stability which led to higher efficiency as compared to the other electrolytes. The η cc improvement can result from retardation of interfacial charge recombination and/or from possible upward shift of the CB edge position of TiO 2 due to development of a dipole at the interface of TiO 2 /metal oxide

24 170 (Law et al., 2006; Butler and Ginley, 1978; Diamant et al., 2004). The CB edge for all composite photoanodes, by extension is the same leaving retardation of charge recombination as the origin of the observed η cc increase. According to Fig.5.8, in DSSC-11, Al 2 O 3 presented the highest blocking ability for interfacial charge recombination because of its lowest R rec, which also resulted in the highest electron lifetime as determined from the EIS measurements. The blocking ability of Al is most likely due to its higher CB edge position. It is notable that the mechanism of charge transfer determining R rec is via direct transfer from the CB of TiO 2 (Fabregat-Santiago et al., 2005). However, electrons may recombine not only from the CB but also from localized intra band states as well as surface states in the band gap (Fabregat-Santiago, 2005; Bisquert et al., 2004). Therefore, it is also interesting to probe the region of lower Fermi levels (Zaban et al., 2003). Thus, the EIS technique is employed as it allows us to monitor the kinetics of recombination in the domain of low photo-voltages (Bisquert, et al., 2004; Zaban et al., 2003]. Fig.5.9 shows the frequency phase plots of DSSCs The characteristic frequency peak of DSSC-11 in higher frequency regime shifted to lower frequency region as compared to DSSCs. This result could be attributed to the higher electron transport rate, electron diffusion coefficient and large surface area of doping structure photoanode in the DSSC-11. These results indicated that the doping photoanode DSSC-11 had fast electrons transport and lower recombination rate at photoanode/dye/electrolyte interface and the same property leading to higher photocurrent efficiency.

25 171 Fig.5.8: Nyquist plots of DSSCs 7-13 Table 5.3: EIS parameters of DSSCs fabricated using doping photoanode structure DSSCs R rec Ω R ct Ω C rec µf C ct µf Z d Ω- s 0.5 k eff s -1 τ rec ms τ d ms L n µm D eff 10-6 cm 2 /s µ 10-4 cm 2 /V /s σ 10-6 S/cm η cc % Conc. Ωcm s -1 DSSC DSSC DSSC DSSC DSSC DSSC DSSC

26 172 Fig.5.9: Frequency Vs phase plots of DSSCs 7-13 Fig.5.10: Bode plots of DSSCs 7-13

27 173 Table 5.4: Frequency and phase of DSSCs fabricated using doping photoanode structure DSSCs f 1 (KHz) f 2 (Hz) υ1 (Hz) PEAK 1 PEAK 2 θ1 (dec) υ2 (KHz) θ2 (dec) DSSC DSSC DSSC DSSC DSSC DSSC DSSC Fig.5.10 shows the Bode plots of DSSCs The Bode plot of DSSC-11 explicated that in the lower frequency region the impedance of the cell is lower than the other DSSCs. These results indicated that the DSSC-11 had lower impedance at electrolyte/counter electrode interface and photoanode/dye/electrolyte interface. This is might be to higher electrons transfer & lower recombination of charge carriers at photoanode/dye/electrolyte interface and faster electron transport at redox couple electrolyte/counter electrolyte interface. Hence in the present investigation, it could be concluded that the Al doped TNPWs had better η cc observed in DSSC-11 built with QSSE. The electronic properties of a material are closely related to its chemical composition (chemical nature of the bonds between the atoms or ions), its atomic arrangement, and its physical dimension (confinement of carriers) for nanometer-sized materials. The chemical composition of TiO 2 can be altered by doping. It is desirable to maintain the integrity of the crystal structure of the photocatalytic host material and to produce favourable changes in electronic structure. It appears easier to substitute the Al cation

28 174 in TiO 2 strucutre and it is more difficult to replace the O 2- anion with other anions due to differences in charge states and ionic radii. The small size of the nanoparticle is beneficial for the modification of the chemical composition of TiO 2 due to the higher tolerance of the structural distortion than that of bulk materials induced by the inherent lattice strain in nanomaterials (Burda, et al., 2003; Chen, et al., 2003). In the present investigation, Al doped TiO 2 showed enhanced photocurrent due to the compatible energy level between Al and TiO 2. The presence of Al metal ion dopants in the TiO 2 matrix significantly influenced the charge carrier recombination rates and interfacial electron-transfer rates. In the present investigation, QSSE had relatively high ambient ionic conductivity, intimate interfacial contact with Al doped TNPWs-HfO 2, and remarkable stability which led to higher efficiency as compared to the other electrolytes. This may be attributed to the high dielectric constant and low viscosity of the organic solvent. The stronger polarity of solvent with high dielectric constant is beneficial to the enhancement of ionic dissociation in the QSSE, leading to the higher conductivity (Wang Miao et al., 2006). Furthermore, the corresponding Jsc value listed in Table 1 also have the same trend as the change of ionic conductivity data. This indicated that the organic solvent have great effects on the performance of QSSE solar cells. Fig shows time-dependent change in the PCE of DSSCs made with different electrolytes through aging test. It was observed that the PCE of the cell based on liquid electrolyte decreased immediately with increasing time and after 30 days was close to zero. On the other hand, the PCE of the cell based on quasi-solid state electrolyte containing PEG polymer decreased much slowly and after 100 days still remained 5.4%, which kept 66% of its initial conversion efficiency, because the addition of the polymer into the electrolyte changed the state

29 175 of electrolyte from liquid state to quasi-solid state and made the organic solvent sealed in the polymer network, which suppressed the evaporation of solvent and hence improved the stability of the cell. The potential problems caused by LE and ILE are leakage and volatilization of organic solvent, considered as some of the critical factors limiting the long-term performance and practical use of DSSC-8 and DSSC-9. The lesser stability of DSSC-10 might be due to the high temperature deposition process of SSE would certainly degrade the sensitized dye on the surface of nanocrystalline TiO 2 (Li et al., 2006). Thus quasi-solid state electrolyte with higher ionic conductivity of S/cm was obtained by solidifying the organic solvent-based liquid electrolyte with PEG polymer. It was applied in the dye sensitized solar cells and acquired PCE of 7.26%. Furthermore, the stability of the cell based QSSE is much better than that of the cell based on the other types of electrolytes, and even after 100 days the PCE still remains 66%. Fig Stability of DSSCs with different electrolytes

30 EIS characterization of DSSCs made with bi-layer photoanode Nyquist plots of DSSCs made using bi-layer photoanode is shown in Fig From the Fig.5.12, comparing the middle semicircles of Nyquist plot of the DSSCs made with bi-layer photoanode structure indicates the decrease of diameter in the order DSSC-14>DSSC-16>DSSC-15>DSSC-17, suggesting that DSSC-17 contributed to the lowest charge transfer resistance at the photoanode/dye/electrolyte interface. When comparing the Nyquist plots of DSSC-17 and DSSC-7, the R rec value increased in DSSC-7 due to uncoated Nb 2 O 5 photoanode. When comparing the Nyquist plots of DSSC-17 and 18, the R rec value increased for DSSC-18 due to bare Nb 2 O 5 photoanode layer. The DSSC-7 shows the highest value of R rec. In this study, higher η cc is achieved in DSSC-17 based on D149 sensitized TNPWs-Nb 2 O 5 bi-layer structure photoanode through quasi-solid state electrolyte as comparing to other cells. The DSSC-17 made with novel photoanode illustrated higher values of C rec, τ rec, L n, D eff, µ, σ and η cc when comparing other DSSCs. In addition, k eff and concentration of electrons in the TiO 2 /dye/electrolyte interface were lower values compared to other DSSCs. The bi-layer TNPWs-Nb 2 O 5 photoanode exhibited better electrochemical behaviours when they are employed in dye-sensitized solar cells. In this photo electrochemical system, charge separation, transport and recombination strongly depend on the bi-layer structure of the photoanode. From EIS data it could be concluded that for the cell containing TNPWs-Nb 2 O 5 bi-layer, formation of a space charge layer at the surface of the electrode effectively promoted the separation of photo generated charge carriers and prevented recombination of the electron with the hole carriers. In addition, the holes transported to electrolyte via the different channel, which could prevented the charge recombination at the bi-layer TNPWs- Nb 2 O 5 /redox electrolyte interfaces and led to the electrons diffusion and transport

31 177 become easy. Therefore, the DSSCs based on the composite core-shell TNPWs- Nb 2 O 5 film had the lower values for R rec, R ct and the longer electron lifetime, consequently leading to an enhanced conversion efficiency of the cell. As revealed from Table 5.5, the presence of bi-layer TNPWs-Nb 2 O 5 layer increases the electron lifetime and makes the electron transfer more easily. In other words, the charge transfer resistance in bi-layer TNPWs-Nb 2 O 5 interfaces and electron recombination are decreased, leading to a positive influence on the improvement of solar cell performance. Hence with EIS investigation, it was showed the enhanced electron transport properties, reduced charge recombination and increased electron life-time which are imperative for higher solar cell performance could be achieved in DSSC-17. Therefore, DSSCs fabricated with bi-layer TNPWs-Nb 2 O 5, which provided direct and enhanced charge transport while minimizing charge recombination resulted in enhanced efficiency in bi-layer TNPWs-Nb 2 O 5 based DSSCs compared to other DSSCs. Use of bi-layer nanostructure the Nb 2 O 5 coating layer may build up an energy barrier at the semiconductor/electrolyte interface and, thus, retard the reaction between the photogenerated electrons and the redox species in electrolyte. Instead of using only nanocrystalline TiO 2 particle films, it has been proposed that a bi-layer structure composed of light scattering layer and nanocrystalline semi-transparent layer can improve photocurrent density substantially due to the fact that the confinement of incident light by light scattering particles can gain more photons (Ito, et al., 2006; Hore, et al., 2006). When light collides with the large TiO 2 particle having submicrometer size, the light scattered strongly, this increased the path length of the incident light in the nanocrystalline TiO 2 films. Eventually, the scattering effect by introduction of the large TiO 2 particles is expected to enhance the photocurrent density and thereby overall

32 178 conversion efficiency. The scattering effect is known to be dependent on size (Vargas, 2000), refractive index (Hore, et al., 2006) and position (Wang, et al., 2004), of the scattering particles. The D eff and τ rec values of the DSSC-17 shown to be higher than DSSC-7. The higher D eff revealed that the TNPWs film becomes more conductive owing to the increased electron density in the conduction band by introducing Tantalum oxide barrier. The increased τ rec value of the photoelectrode by introducing Hafnium oxide also proved that the blocking layer formed on the TNPWs surface can effectively retard the electron recombination process between nanocrystalline TiO 2 and electrolyte. The enhancement of the electron injection might be attributed to the shift of conduction band edge of TiO 2. Previously, it has been reported that the acidic property of Nb 2 O 5 shifted the TiO 2 flat band energy potential (E fb ) toward positive values and increased the driving force for electron injection (determined as the difference between E fb and the LUMO state of the dye), thereby enhancing the electron injection efficiency. Here, in EIS spectra, the Nyquist plot revealed a large semicircle in the frequency range between 1 mhz to 10 5 Hz. The size of the arc depended on the concentration of I 3 and on the rate of back-electron transfer at the TiO 2 /electrolyte interface. The impedance spectrum was analyzed using an equivalent circuit for the TiO 2 /electrolyte interface in the conductive state. The recombination resistance (R rec ) is obtained by the curve of the large semicircle. The EIS is sensitive to the interfacial electron flow in the electrode/electrolyte interface. The EIS curves show the reduction peaks attributed to the charging/discharging at the interface and the peak was shifted positively upon the Nb 2 O 5 coating. This result confirms the positive shift of the conduction band edge in Nb 2 O 5 -coated TiO 2 electrodes. The much larger τ rec in the bi-layer DSSC-17 as compared to DSSC-18 may be explained by an internal radial electric field that is

33 179 developed within the walls of the bi-layer. This electric field drives the electrons away from the interface, preventing recombination with the electrolyte (Xu et al., 2010). The primary reason for longer lifetimes in DSSC-17 is due to the intrinsic material properties. It has been reported that the conduction band of TiO 2 is composed of d-orbital electrons that are less susceptible to recombination than the electrons in Nb 2 O 5 (Bandaranyake et al., 2004). Among the different electrodes, the bi-layer TNPWs Nb 2 O 5 electrode exhibited the slowest recombination rate, lower than the bare Nb 2 O 5 (DSSC-18). This could be due to the energy barrier formed by the Nb 2 O 5 coating, forcing the electrons to flow toward the current collecting surface and preventing the electrons from flowing in the opposite direction. Introduction of Nb 2 O 5 may have also passivated the reactive low-energy surface states. These factors lead to further lowering of the recombination rates, resulting in enhanced η cc for DSSC-17 based on the bi-layer TNPWs Nb 2 O 5 electrodes. The appliance of Nb 2 O 5 is an efficient layer material on the bi-layer TNPWs-Nb 2 O 5 electrodes in DSSCs. The diffusion coefficient (D eff ) and lifetime (τ rec ) of the DSSC-17 were increased by the Nb 2 O 5 energy barrier. It is due to the more conductive bi-layer TNPWs-Nb 2 O 5 electrodes film (that is, the more accumulation of the electrons in the conduction band of the TiO 2 film) and the reduced recombination rate at the interface between the TNPWs-Nb 2 O 5 and dye/electrolyte. In the present investigation, DSSC-17 with quasi-solid state electrolyte had relatively high ambient ionic conductivity, intimate interfacial contact with core-shell TNPWs-Nb 2 O 5, and remarkable stability which lead to higher efficiency.