稀有金属(英文版) 2015,34(01),34-39

收稿日期：8 August 2014

基金：financially supported by the National Natural Science Foundation of China (No. 51202139);Shanghai Pujiang Program (No. 11PJ1403400);the Fundamental Research Funds for Central Universities (No. 2012LYB24);

Improved photovoltaic performance of dye-sensitized solar cells by carbon-ion implantation of tri-layer titania film electrodes

Jun Luo Wei-Guang Yang Bin Liao Hai-Bo Guo Wei-Min Shi Yi-Gang Chen

School of Materials Science and Engineering, Shanghai University

College of Nuclear Science and Technology, Beijing Normal University

Abstract：

TiO2tri-layer structure films were modified by C-ions implantation for improving the photovoltaic performance of dye-sensitized solar cells(DSSCs), in which the structure of TiO2 changes from rutile to anatase and the sizes of TiO2 particles increase. The optimal concentration of ions implantation for C-implanted cells is 19 1015 atom cm-2,and the maximum conversion efficiency of 5.32 % is achieved(luminous intensity of 1 sun, light irradiance of AM1.5G),which is 25.2 % higher than that of unimplanted cell. The significant improvement in conversion efficiency by carbonion implantation is contributed to reducing charge recombination and enhancing the light-harvesting ability, as indicated from incident photon-to-collected electron conversion efficiency(IPCE) and ultraviolet-visible(UV-Vis) measurements. Furthermore, the charge carrier's lifetime in the trilayer titania films is prolonged after carbon-ion implantations.

Keyword：

Dye-sensitized solar cell; Ion implantation; Carbon; Conversion efficiency; Charge transfer;

Author： Yi-Gang Chen,e-mail: yigangchen@shu.edu.cn;

Received： 8 August 2014

1 Introduction

Dye-sensitized solar cells (DSSCs) were extensively studied and technologically attractive because of their low-cost,simple-synthesis procedure and high-conversion efficiency[1–5]. Up to now, the highest conversion efficiency of15 % has been reported, which challenges conventionalsilicon-based solar cells [6]. DSSC devices have a sandwich structure, consisting of a charge-transfer dye absorbedon a mesoporous film (typically titania film) supported on atransparent conductive glass (FTO) as photoelectroanode, aredox system containing iodide and triiodide ions as electrolyte, and a Pt black coated on FTO as counter-electrode[7, 8]. The mesoporous titania films play a significant rolein DSSCs owing to their large dye-adsorption capacity andhigh electron transport efficiency. However, the filmsgenerally have various amounts of defects which serve asthe traps of charge and the places for the recombination ofelectrons and excited dye molecules, leading to low photovoltaic performance.

At present, one of the most effective ways of facilitatingcharge separations and electron transport is doping metal ornon-metal ions to the electrode materials. Liu et al. [9] dopedgroup-V (vanadium, niobium, and tantalum) ions in titaniafilms to positively shift the conduction band edge of titania,match the lowest unoccupied molecular orbital (LUMO)level of photosensitive dye, and improve photoelectricresponse of DSSCs. Hachiya et al. [10] reported that V ionsserved as recombination centers in titania films with lowdoping density, while they acted as mediator for electrontransport with high doping density. Pang et al. [11] found thatMg-doped Sn O₂with ultrathin Ti O₂coating layer enhancedcharge transfer and prolonged the lifetime of electrons. Ma’sgroup systematically investigated N-doped Ti O₂electrodesand found that nitrogen dopant decreased the charge-transferresistance and enhanced the incident photon-to-collectedelectron conversion efficiency (IPCE) [12].

However, there are very few papers on the photovoltaicproperties of DSSC electrodes modified by ion implantation. Techniques of ion implantation are widely applied insurface modification, as they have the advantages ofaccuracy and controllability in complex reactions andprocesses of doping [13–15]. In contrast, chemical dopingis generally confronted with difficulties of introducingunwanted impurities, controlling complex reactions, andhaving poor repeatability. In this work, ion implantationwas used to dope carbon ions with low energies into trilayer titania films. Variable densities of carbon ions wereimplanted in the films to study the effect of implantationdensity on photoelectric response (electron transport,electron’s lifetime, electron injection) of the devices.

2 Experimental

2.1 Preparation of carbon-ion implanted titaniatri-layer films

Three kinds of titania pastes made up of titania particles ofdifferent sizes were screen-printed on a F-doped Sn O₂conductive glass (FTO glass) substrate to make the titania trilayer photoanodes. The titania pastes were prepared following the procedures in Ref. [16]. A 6-lm-thick transparentlayer was printed by the paste of titania particles of 20 nm onthe FTO substrate, forming the bottom layer. This layer wascovered by a 6-lm-thick layer of the mixed paste (made up ofpastes of titania particles of 20 and 200 nm, in mass ratio of6:4). The top layer is a 2-lm-thick scattering layer using thepaste of titania particles of 200 nm. Before the screen printing, the FTO glass was treated by ultrasonic cleaning indeionized water and alcohol, and immersed in 40 mmol L^-1Ti Cl₄aqueous solution at 70 °C for 30 min. After sintered at500 °C for 30 min, the tri-layer films were placed in a metalvapor vacuum arc (MVVA) implanter (details of theimplanter being in Refs. [17, 18]). The accelerating voltagewas set to a low energy of 8 k V at high vacuum level(\1 9 10^-3Pa) at room temperature, to limit the damage tothe films induced by the ions. Variable contents of carbon ions1 9 10¹⁵, 2 9 10¹⁵, 4 9 10¹⁵, and 6 9 10¹⁵atom cm^-2were applied and compared with the unimplanted sample toinvestigate the influence of dose on the photoelectric properties of the devices. The samples were named C100-TiO 2,C200-TiO 2, C400-Ti O₂, and C600-TiO 2, respectively.

2.2 DSSC assembly

The carbon-implanted and un-treated samples were dippedinto 40 mmol L^-1Ti Cl₄solution at 70 °C for 30 min, andsintered at 500 °C for 30 min. Then the samples were cooledto 80–100 °C and immersed immediately in a 0.5 mmol L^-1dye solution of cis-bis (isothiocyanato) bis (2,2-bipryridyl4,4⁰-dicarboxylic acid) ruthenium (II) (N719) [19]. Forcompleting the dye adsorption, the procedure of sensitizationcontinued for 24 h at room temperature in the dark. Aplatinum paste was screen-printed on another FTO substrate,dried at 100 °C for 1 h, and annealed at 400 °C for 10 min,for the platinization of the counter-electrode. The dye-coatedworking-electrode and Pt counter-electrode were assembledinto a sandwich type cell with a sealing spacer (surlyn film,50 lm in thickness). A redox (I^-/I₃^-) electrolyte wasinjected between the two electrodes. The active area of theworking-electrode in the cells was 0.25 cm².

2.3 Characterization

X-ray diffraction (XRD) was performed using a D/Max2200 X-Ray diffractometer with Cu Ka radiation in 2hrang of 20°–40°. A JSM-6700F cold field electron-scanning electron microscopy (FE-SEM) equipped with anenergy-dispersive X-ray spectroscopy (EDX) detector wasemployed for measuring the morphology of the carbon-ionimplanted films and un-treated films. Photocurrent-voltagecharacteristics were measured with a solar light simulator(Oriel, 91160-1000) under intensity of 100 m W cm^-2, atlight irradiance of AM1.5. The monochromatic IPCEspectra were examined by a Qtest Station 2000AD in thewavelengths range of 200–800 nm. The ultraviolet–visible(UV–Vis) absorption spectra were performed with a Cary 5spectrophotometer. The electrochemical impedance spectra(EIS) were carried out by an electrochemical workstationPHI660E at a forward bias of 0.75 V in the frequencyrange of 1 Hz–100 MHz.

3 Results and discussion

3.1 Crystal structure and morphology

Figure 1 shows the XRD patterns of C100-Ti O₂and unimplanted-Ti O₂samples. The typical peaks for the anatasephase (101) and (004) are found in both samples, but thepeaks for the rutile phase at 2h = 29.2° are absent in C100Ti O₂. This illustrated that the carbon-ion implantationalters the crystalline structure of titania from rutile toanatase. This phase transfer could increase electron transport, because electron transport in anatase is faster than thatin rutile [20].

Table 1 listed the lattice parameters (a, b, and c) andcrystallite sizes of the C100-Ti O₂and the unimplanted-Ti O₂samples. The lattice parameters of C100-Ti O₂(a = b =0.377 nm, c = 1.251 nm) are larger than those of the unimplanted-Ti O₂(a = b = 0.369 nm, c = 1.238 nm). Theexpanded lattice of C100-Ti O₂is evidence that carbon ionsenter the lattice of titania. The sizes of titania particlesincrease slightly after carbon-ion implantation (Table 1).This could enhance the scattering property of the scatterlayer and the mechanical strength of the tri-layer films.

Fig. 1 XRD patterns of unimplanted titania tri-layer films andcarbon-ion implanted films with 1 9 1015atom cm-2

Table 1 Lattice constants and average crystal sizes of C100-Ti O₂and unimplanted Ti O₂  下载原图

Table 1 Lattice constants and average crystal sizes of C100-Ti O₂and unimplanted Ti O₂

The mesoporous tri-layer structure of the titania films isshown in Fig. 2. This structure is designed to improve theadsorption of dye molecules [21]. There are two differentsizes of particles clearly observed in the films, with thelarger particles being approximately 200 nm in dimension,and the smaller ones being approximately 30 nm. Theparticles sizes have no significant change after carbon-ionimplantation. The images also exhibit similar morphologiesof the unimplanted films and C600-Ti O₂, indicating that nosignificant damages are induced by carbon implantationseven at the highest dose of 6 9 10¹⁵atom cm^-2in thisexperiment.

3.2 Solar energy conversion efficiency

The I–V curves of the DSSCs based on the carbon-ionimplanted and un-treated tri-layer titania films are presented in Fig. 3, and the detailed photoelectric parametersare listed in Table 2. The solar energy conversion efficiency (g) of the cells is related to photocurrent density(J_SC), open circuit photovoltage (V_OC), fill factor (FF), andpower of incident light (P, 100 m W cm^-2) as follows:

Fig. 2 SEM images of a unimplanted and b C600-Ti O2films

The efficiency of DSSC based on C100-Ti O₂electrodeis 5.32 %, which is significantly higher than that of theunimplanted device (4.25 %). However, with the doses ofcarbon-ion implantation increasing, the efficiencies of thecarbon implanted DSSCs reduce from 5.32 % (C100-Ti O₂)to 3.77 % (C600-Ti O₂). The same trend is found in J_SC(Fig. 3b). This indicates that carbon-ion implantation iseffective on electron transport and charge recombination.The low-content carbon implantation would improve theelectron transport, but the carbon ions also play a role ofrecombination centers, which deteriorate the photocurrentdensity when the dose of carbon-ion implantation exceeds1 9 10¹⁵atom cm^-2. Another reason for the enhancementof J_SCin C100-Ti O₂sample is the shift of conduction bandof titania by the carbon-ion implantation. It was reportedthat doped ions could positively shift conduction band oftitania to better match the LUMO of the dye, resulting inthe increase of electrons injected from the excited dye [9].

IPCE is defined as the incident monochromatic photonto-electron conversion efficiency. It is related to J_SCbyEq. (2)

where k is wavelength. The IPCE spectra for the dyesensitized (N719) unimplanted and C100-Ti O₂tri-layertitania films are shown in Fig. 4. According to the entirespectra region, C100-Ti O₂films exhibit enhanced IPCEvalues compared with the unimplanted sample. The resultindicates that the carbon-ion implanted tri-layer films havea large amount of dye adsorption, which improves the lightharvest. These factors probably lead to the high J_SCofDSSC based on C100-Ti O₂.

The ultraviolet–visible absorption spectra of the unimplanted and the C100-Ti O₂films are shown in Fig. 5. Theabsorption peaks correspond to electron-transition from thehighest occupied molecular orbital (HOMO) to the LUMOof the N719 dye [22]. The positions of the peaks have noobvious change between the unimplanted and the C100Ti O₂films, but the intensities apparently differ. Theintensities of C100-Ti O₂at peaks of 334, 402, and 550 nmare 54.1, 44.3, and 48.6, which are 8.6 %, 1.3 %, and7.1 %, higher than those of the unimplanted sample,respectively. The result also proves that the amount ofN719 dye adsorption is enhanced by carbon-ionimplantation.

Fig. 3 Photocurrent-voltage curves for unimplanted DSSC and DSSCs based on carbon-implanted tri-layer titania films with variableimplantation doses and incident light intensity of 100 m W cm^-2(luminous intensity of 1 sun) a; photovoltaic parameters of DSSCs as a functionof implanted carbon-ions dose b

Table 2 Photovoltaic parameters of DSSC based on carbon-ionimplantations with different contents  下载原图

Table 2 Photovoltaic parameters of DSSC based on carbon-ionimplantations with different contents

Fig. 4 IPCE spectra of DSSC with unimplanted tri-layer titania filmsand C100-Ti O2as photoelectrodes

Gao et al. [23] calculated the band structure of C-dopedTi O₂using the first-principles density functional theory(DFT). The results show that the Ti O₂energy structure isinfluenced by C-doping. Three isolated C2p impurity statesare located in band gap of Ti O₂and would act as the chargerecombination centers, which would be unsuitable foroptoelectronic materials [24]. However, the presence ofimpurity states narrows the band gap and obviouslyexpands the light absorption range from UV to visible light.This electronic structure model explains the increasing J_SCof C-ion implanted cells with the concentrations of C ions,and agrees well with the results of UV–Vis absorptionmeasurements.

Fig. 5 UV–Vis absorption spectra of unimplanted tri-layer Ti O2films and C100-Ti O2films

3.3 Charge transfer resistance and lifetime of electrons

To further investigate the enhancements of the photoelectric properties of the DSSCs by the carbon-ion implantationwith 1 9 10¹⁵atom cm^-2, charge-transfer resistance wasmeasured for a series contents of carbon-ion implantedcells and un-treated cells by EIS. The Nyquist plots arepresented in Fig. 6. Generally, there are three semicirclesin Nyquist plots, where the first semicircle (high-frequencyregion) is related to charge transfer at Pt/electrolyte interface, the second semicircle (in mid-frequency region) tocharge transfer at titania electrode/electrolyte interface, andthird semicircle (in low-frequency region) to a Warburgimpendence of electrolyte [25]. In this study, our mainfocus is on the titania electrodes; therefore, the measurements were in the frequency region of 1–100 k Hz. The singlesemicircle of the Nyquist plots is attributed to the electrontransfer at the titania/electrolyte interface. The diameter ofthe semicircle corresponds to the magnitude of chargetransfer resistance [26]. The Nyquist plots show that the cellbased on C100-Ti O₂has much lower resistance comparedwith the un-treated cell. This indicates that the electron hashigher charge-transfer rate in the C100-Ti O₂films. However, the resistance of the cells increases with the amount ofcarbon ions increasing. One possible reason is that carbonions act as recombination centers in the titania films at highcontents of carbon-ion implantation.

Fig. 6 Nyquist plots of electrochemical impendence spectra ofDSSCs with and without carbon-ion implantations. Inset being aBode plot of DSSCs with C100-TiO₂ and unimplanted-TiO₂ film aselectrodes

Bode plots (inset of Fig. 6.) of the cells based on C100Ti O₂and unimplanted-Ti O₂are presented to extract theelectrons lifetime by the following equation [27]:

where f is frequency. Larger values of s_effmean that moreelectrons can transport into the outside circuit and lesscharges recombination [28]. The peak frequency of thecells based on the C100-Ti O₂(4,570 Hz) is lower than thatof the unimplanted electrode (5,620 Hz). This indicatesthat the short-circuit current density of the device isenhanced by carbon-ion implantation with the dose of1 9 10¹⁵atom cm^-2.

4 Conclusion

In conclusion, the ion-implantation technique wasemployed to modify the tri-layer titania films for DSSCdevices. With the carbon-ion implantation at the dose of1 9 10¹⁵atom cm^-2, the conversion efficiency of the untreated titania electrodes increases to 5.32 % from 4.25 %.This substantial improvement is contributed to enhancingdye adsorption, lowering charge-transfer resistance, andprolonging electron’s lifetime, as indicated from the measurements of IPCE and UV–Vis. However, carbon-ionimplantations at doses of higher than 1 9 10¹⁵atom cm^-2lead to the decrease of conversion efficiencies. This couldbe attributed to the increased density of carbon ions thatserve as charge-recombination centers. These positiveresults indicate that carbon-ion implantation is a newapproach and has the potential to improve the performanceof DSSCs. It requires further studies to exploit newapplications of this method for DSSC devices in the future.