Novel Nano-Structured Silica-Based Electrolytes Containing Quaternary Ammonium Iodide Moieties

FULL PAPER

DOI: 10.1002/adfm.200700391

Novel Nano-Structured Silica-Based Electrolytes Containing Quaternary Ammonium Iodide Moieties** By Sophie Cerneaux, Shaik M. Zakeeruddin, Jennifer M. Pringle, Yi-Bing Cheng, Michael Grätzel,* and Leone Spiccia* A series of new hybrid organic-inorganic molecules were prepared either by grafting of aminopropyltriethoxysilane (APTS) on silica nanoparticles followed by quaternarization of the nitrogen with ethyl, heptyl and isopropyl iodides or by grafting of N,N,N-triethyl-3-(triethoxysilyl)propan-1-aminium iodide and N,N,N-tridodecyl-3-(triethoxysilyl)propan-1-aminium iodide onto the silica nanoparticles. These new materials were used as iodide sources in the preparation of electrolyte solutions for dye-sensitized solar cells (DSSCs). The performance of DSSCs was studied as a function of the nature of the solvent, the nature of the dye, the concentration of the modified silica in the electrolyte system and the silica content introduced during the hybrid synthesis. An efficiency of 8.5 % was obtained for solar cells containing the triethyl ammonium iodide salt at a concentration of 1 M in either acetonitrile (AN) or 3-methoxypropionitrile (MPN) under an illumination of 10 mW cm–2, the equivalent of 0.1 Sun at AM 1.5G. At 1 Sun (100 mW cm–2, efficiencies of 6.6 % and 5.1 % were recorded for the AN and MPN-based electrolytes, respectively.

1. Introduction Dye-sensitized solar cells (DSSCs) based on nanocrystalline films have received high recognition due to their good efficiency and low cost compared to inorganic or silicon-based other devices.[1] These photovoltaic cells usually consist of a TiO2 nanostructured semiconductor film on which is absorbed a monolayer of a RuII photoactive complex and finally, an electrolyte composed of an I–/I3– redox couple in organic solvents.[2] Both the high volatility and leakage of organic solvents used in liquid electrolytes remain major issues in the commercialization of DSSCs. As alternatives, quasi-solid state or solid state materials such as ionic liquids, polymeric electrolytes and plastic crystals have been developed and intensively studied.[3–9] An efficiency close to 7 % was achieved for modified quasi-solid electrolytes using poly (vinylidenefluoride-co-hexafluoropropylene) (PVDF-HFP) and silica nanoparticles to soli-

– [*] Prof. M. Grätzel, Dr. S. M. Zakeeruddin Laboratory for Photonics and Interfaces Ecole Polytechnique Fédérale de Lausanne 1015 Lausanne (Switzerland) E-mail: [email protected] Prof. L. Spiccia, Dr. S. Cerneaux, Dr. J. M. Pringle, Prof. Y-B. Cheng School of Chemistry and Department of Materials Engineering Monash University, Victoria, 3800 (Australia) E-mail: [email protected] [**] S.C., Y.B.C., and L.S. thank the Australian Research Council for financial support through the Discovery, Linkage International and Centre of Excellence (Australian Centre for Electromaterials Science) programs. S.M.Z. and M.G. thank The Swiss National Science Foundation for the financial support.

3200

dify solvent-based electrolytes, that contained an ionic liquid, 1-methyl-3-propylimidazolium iodide, as iodide source.[10] Xue et al. prepared a composite electrolyte consisting of LiI(CH3OH)4-I2 mixed with an ionic liquid and silica nanoparticles which gave an efficiency of 4.3 %.[11] These workers showed that the presence of silica nanoparticles prevented the crystal growth of LiI(CH3OH)4-I2. Lianos and co-workers have used a nanocomposite silica-based gel with an organic subphase containing propylene carbonate, 1-methyl-3-propylimidazolium iodide and Triton as surfactant which, when applied as electrolyte in DSSCs, gave a maximum efficiency of 5.4 %.[12] The same group achieved an efficiency of 5–6 % with a nanocomposite organic-inorganic sol-gel electrolyte which incorporated a hydrolysable alkoxysilane derivative.[13] We report an attractive strategy to enhance the stability of the solar cells that makes use of modified silica nanoparticles as iodide source in the electrolyte system. Indeed, the presence of nanosized particles, generally used as fillers in polymers, is expected to improve the photovoltaic properties of the cells by creating specific pathways for ion transportation and to prevent electron recombination at the working electrode. Nanostructured quaternary ammonium silica-based iodide salts were prepared following two different approaches. One involved the grafting of 3-aminopropyltriethoxysilane (APTS) onto the surface of activated silica nanoparticles, leading to aminopropylsilica particles that were further reacted with a series of alkyl iodides to prepare the new organic-inorganic materials described in this work. In a second approach, novel triethoxysilane precursors bearing quaternary ammonium iodide moieties[14] were grafted onto activated, nanosized silica particles. This second approach was expected of great interest in controlling the silica content of these hybrid materials. It also gives a wide range of possibili-

© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Funct. Mater. 2007, 17, 3200–3206

S. Cerneaux et al./Novel Nano-Structured Silica-Based Electrolytes

2. Results and Discussion 2.1. Silica Nanoparticles Functionalization 2.1.1. Amination of SiO2 Nanoparticles Amino-modified silica nanoparticles A were synthesized as described in Scheme 1. The grafting of 3-aminopropyltriethoxysilane (APTS) onto the surface of activated silica was achieved by condensation of silanol groups (–OH) created at

activation of the silica, isolated silanols are generated at the surface of silica as depicted in Figure 1(1). In step 2 of Figure 1, APTS molecules can be linked to the silica surface via one (e), two (f), or three (g) Si–O–Si covalent bonds and unreacted silanols may remain (a). In addition to the APTS monomers reacting with the activated silica surface, self-condensation may also occur, which results in the formation of dimers, trimers and higher oligomers of APTS that may react or bind to the oxide surface. Consequently, it is possible that only a fraction of APTS is bound directly to the inorganic surface. Silylated materials are usually characterized by 29Si MAS NMR spectroscopy as the progress of hydrolysis and condensation reactions can easily be followed. The amino-silica nanoparticles A are characterized by a signal at –110 ppm (Fig. 2) corresponding to Si–O units linked to four silicon atoms, denoted as Q4 entities (Si(OSi)4), whereas the starting 3-amino-

FULL PAPER

ties in designing new molecules that can be used in many different areas, such as chromatography, specific recognition or biomaterials.

10 % (v/v) HCl solution

(1)

2h, RT

Silica nanoparticle

(2) +

Si

Ethanol

Si

48h, RT

A

APTS

Activated silica

Scheme 1. Synthetic route to A: 1) Activation of commercially available silica nanoparticles and 2) Grafting of 3-aminopropyltriethoxysilane (APTS) on activated silica nanoparticles.

the silica surface with the hydrolysable ethoxy groups (–OCH2CH3). SiO2 nanoparticles were first activated in an acidic media (10 % v/v HCl) for 1 h, following Step (1) in Scheme 1, prior to the dropwise addition of APTS in dry ethanol (40 mL) (Scheme 1, Step 2). A simplified overview of the different species that can be present during the condensation reactions between APTS and the activated silica particulates is presented in Figure 1. After

(2) (a)

(e)

(f)

SiO1.5

(g)

APTS addition

(1) (a)

(b)

(c)

SiO1.5

(d)

Figure 1. Stepwise grafting of APTS to silica nanoparticles: 1) Different types of surface hydroxyls species that may be present on the activated silica surface, a) and b) single isolated units, c) double units, and d) triple units; and 2) Different types of grafted species formed upon condensation with APTS, a) non reacted mono units or T0, e) single attachment or T1, f) doubly bridged units or T2, and g) triply bridged units or T3.

Adv. Funct. Mater. 2007, 17, 3200–3206

Figure 2. 29Si MAS NMR spectrum of amino-silica nanoparticles A.

propyltriethoxysilane is characterized by a signal at –46 ppm specific of T0 units or (Si(OEt)3R, R represents the aminopropyl pendant). The signals observed for nanoparticles A at –51, –60, and –68 ppm are representative of T1 [Si(OEt or OH)2(OSi)1R], T2 [Si(OEt or OH)1(OSi)2R], and T3 [Si(OEt or OH)0(OSi)3R] units, respectively, confirming that the APTS has been successfully grafted to the inorganic network (Q4 at –110 ppm). The disappearance of the vibration band at around 960 cm–1 characteristic of SiOEt groups of the starting chemical and the presence of a very broad vibration band from 1000 to 1200 cm–1, attributable to the silica network, in the FTIR spectrum confirmed the attachment of APTS to the silica surface. Moreover, the covalent bond between the oxide surface and the organic moiety is evidenced by the existence of a sharp vibration at 795 cm–1 and a broad band at around 1300 cm–1. 2.1.2. Quaternarization of Amino-Silica Nanoparticles A series of quaternary ammonium iodide compounds were synthesized by reaction of A with different alkyl iodides as described in Scheme 2. Nucleophilic attack of the amino group of

© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.afm-journal.de 3201

FULL PAPER

S. Cerneaux et al./Novel Nano-Structured Silica-Based Electrolytes

Scheme 2. Synthetic route to the quaternary ammonium silica-based iodide salts B, C and D. Reaction conditions: 48 h for B, 96 h for C and 72 h for D at reflux in acetonitrile in presence of excess of K2CO3.

the modified silica A onto the alkyl chain of ethyl, heptyl and 2-propyl iodide led to the desired quaternary ammonium salts N,N,N-triethyl-3-silica-propan-1-aminium iodide, N,N,N-triheptyl-3-silica-propan-1-aminium iodide and N,N,N-triisopropyl-3silica-propan-1-aminium iodide denoted B, C, and D, respectively. The reactions were performed in presence of K2CO3 in order to neutralize the H+ liberated during the coupling and the products were obtained following workup. The 13C CP/MAS NMR spectrum of the carbon side chain of A is characterized by three signals at 8.3, 22.9, and 41.2 ppm, whilst the spectrum of compound B shows two additional signals at 6.4 and 54.6 ppm, which are representative of ethyl group on the terminal nitrogen. This is further confirmed by both the disappearance of the weak vibrations at around 3380 and 3273 cm–1 in the FTIR spectrum and the appearance of new and intense vibrations at 2977, 2952, 2904, and 2810 cm–1, which are characteristic of CH3 and CH2 groups. Compound C was obtained as a viscous oil that did not permit characterization by solid-state NMR spectroscopy. However, the solution 1H and 13C NMR spectra provided evidence for the covalent chemical modification even though the signals were broadened because of the silica network presence. Signals at 0.9, 1.3, 1.4, and 2.4 ppm appear for compound C, which are characteristic of the heptyl chains attached to the quaternary nitrogen. The 13C solid-state NMR spectrum of D shows three signals at 13, 28, and 48 ppm, with different intensities when compared to A but it is not clear that the reaction went to completion. 2.1.3. Grafting of N,N,N-Triethyl-3-(triethoxysilyl)propan-1aminium Iodide E and N,N,N-Tridodecyl-3(triethoxysilyl)propan-1-aminium Iodide F on Activated Silica Nanoparticles N,N,N-triethyl-3-(triethoxysilyl)propan-1-aminium iodide and N,N,N-tridodecyl-3-(triethoxysilyl)propan-1-aminium iodide denoted E and F, respectively, were synthesized following a procedure described elsewhere.[14] Reaction of E with various amounts of activated silica in refluxing acetonitrile led to compounds G and H depicted in Scheme 3. The hybrid G was prepared using a 1:4 molar ratio and compound H in a 1:8 molar ratio of silane to silica nanoparticles (based on silica composition). Compound K was obtained from F in a 1:8 molar ratio following the same procedure. Characterization by 29Si solid-state NMR spectroscopy indicates that compound H contains more silica than G as a broad-

3202 www.afm-journal.de

Scheme 3. Hybrid quaternary ammonium iodide salts, G, H and K, prepared in a 1:4 molar ratio for G and in a 1:8 molar ratio for both H and K. Reaction conditions: 24 h at reflux in acetonitrile for G and H and in acetonitrile:chloroform (2:1 v/v) for K, under nitrogen atmosphere.

er and more intense signal at around –110 ppm and a weaker broad signal at about –70 ppm are observed. In addition, the organic composition of both compounds did not change as evidenced by the presence of similar signals in 13C CP/MAS NMR at 10, 16, 32, and 54 ppm shown in Figure 3.

Figure 3. 13C CP/MAS NMR spectrum of compound H.

2.2. Solar Cell Performance The silica based quaternary ammonium salts were used to prepare liquid electrolyte solutions, which were then used in solar cells. The influence of the solvent, the nature of the dye, the concentration of the hybrid in the electrolyte system and the content of silica in hybrid photovoltaic characteristics was studied. Testing was performed on cells with electrolytes based on compound B only as compounds C and D were not miscible with either acetonitrile or MPN. Compounds containing higher silica content (G, H, and K) could not be properly dissolved in any solvent at the concentrations required for our study.

© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Funct. Mater. 2007, 17, 3200–3206

S. Cerneaux et al./Novel Nano-Structured Silica-Based Electrolytes

Hybrid B, containing 33 % of silica, was mixed with 0.08 M I2, 0.1 M GuNCS and 0.5 M TBP, in acetonitrile solvent to prepare electrolyte 1. 31 wt % of B was present in the electrolyte system, corresponding to a concentration of hybrid of 1 M in acetonitrile. GuNCS was added as it has been shown to lead to improved VOC values.[15–17] An electrolyte based on MPN as solvent was also used as electrolytes prepared using this solvent can lead to cell with high fill factors and VOC values.[18] Electrolyte 2 was prepared by combining appropriate amounts of 1 M B (28 wt %), 0.08 M I2, and 0.5 M TBP. Figure 4 shows the photovoltaic performances of the solar cells prepared with Z907Na photosensitizer and electrolyte systems 1 and 2. At 1 Sun illumination, open-circuit voltages (VOC) of 768 and 830 mV were measured for cells prepared with 1 and 2, respectively, as reported in Table 1. A lower current (JSC) of 10.7 mA cm–2 is observed for 2 in MPN solvent when compared to 1, which gives a current of 15.2 mA cm–2. The two cells have similar fill factors, ff, but the overall efficiency is higher for the cells containing 1 (6.6 %) than for those containing electrolyte 2 (5.1 %). A lowering in the light intensity to 0.1 Sun led to much higher efficiency of 8.5–8.6 % for the two electrolytes tested, which also contained the same silica loading. This is mainly due to a higher fill factor value, approaching 0.8 for electrolyte 2, whilst maintaining a quite high potential value and slight relative improvements in photocurrent at this light intensity. The two sys-

tems have a similar behavior even though a different solvent is used. The I–V characteristics for cells constructed with the MPN based electrolyte, in particular the fill factor and photovoltage, are very impressive.

FULL PAPER

2.2.1. Influence of the Electrolyte Solvent

2.2.2. Influence of Nature of the Dye To study the interaction of the electrolyte system with the dye, solar cell testing was conducted using Z907Na and N719 dyes that are bearing different substituents on one bipyridyl ligand, viz. two carboxylate groups in N719 versus two nonyl substituents in Z907Na. Figure 5 shows that lower VOC and JSC are measured when N719 is used as photosensitizer instead of the amphiphilic Z907Na dye at both full light (Fig. 5b) and 10 % intensities (Fig. 5d), viz., 6.3 % (1 Sun) and 7.0 % (0.1 Sun) efficiencies have been measured for cells containing N719 in comparison to 6.6 % and 8.5 % at 1 Sun (Fig. 5a) and 0.1 Sun (Fig. 5c), respectively, for those containing Z907Na. The better performance of the cell constructed with Z907Na

Table 1. Characteristic parameters of solar cells containing electrolytes 1 and 2 measured under an illumination of 0.1 Sun and 1 Sun (Z907Na dye used to sensitize the working electrode; cell active area of 0.158 cm2). Electrolyte 1[a] 2[b]

JSC [mA.cm–2] VOC [mV] ff g [%] 0.1 Sun 1 Sun 0.1 Sun 1 Sun 0.1 Sun 1 Sun 0.1 Sun 1 Sun 1.6 1.3

15.2 10.7

701 773

768 830

0.72 0.79

0.56 0.58

8.5 8.6

6.6 5.1

[a] Electrolyte 1 consists of 1 M B (31 wt %), 0.08 M I2, 0.1 M GuNCS and 0.5 M TBP in acetonitrile. [b] Electrolyte 2 consists of 1 M B (28 wt %), 0.08 M I2 and 0.5 M TBP in MPN solvent.

Figure 5. Comparison of I–V profiles for solar cells containing electrolyte 1 (AN based) at a) 1 Sun and c) 0.1 Sun light in presence of the Z907Na dye and b) 1 Sun and d) 0.1 Sun light, using the N719 dye.

Figure 4. Photocurrent density-voltage characteristics recorded under an illumination of a) 1 Sun at 1.5 G (100 mW cm–2) and b) 0.1 Sun for electrolyte 1: (31 wt %, 1 M), 0.08 M I2, 0.1 M GuNCS and 0.5 M TBP in AN (I) and electrolyte 2: B (28 wt %, 1 M), 0.08 M I2 and 0.5 M TBP in MPN (II).

Adv. Funct. Mater. 2007, 17, 3200–3206

© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.afm-journal.de 3203

FULL PAPER

S. Cerneaux et al./Novel Nano-Structured Silica-Based Electrolytes dye, which is most apparent at the low light intensity of 0.1 Sun, may be attributed to the presence of long alkyl chains on bpy ligands. These groups may reduce the probability of electron recapture (recombination) by the oxidized ruthenium complex due to a steric effects compared to the carboxylate entities attached on bpy ligands for N719. 2.2.3. Influence of the Concentration of Modified Silica B in AN In order to examine the effect of reducing the solvent content of the electrolyte system, the concentration of B in AN was increased from 1 M (31 wt %) to 1.5 M (41 wt %) in electrolyte 3 and 2 M (48 wt %) in electrolyte 4, containing still 0.08 M I2, 0.1 M GuNCS, and 0.5 M TBP. Changing the concentration of the modified silica in the electrolyte led to the variations in photovoltaic performance reported in Table 2. We report the results of testing at low light intensity of 0.1 Sun since this was the highest efficiency reached by cells containing the electrolyte prepared using hybrid B. The voltage was dramatically improved from 1 to 4, by 200 mV, but there

Table 2. Evolution of photovoltaic parameters of solar cells containing 1 M, 1.5 M, and 2 M B in AN, 0.08 M I2, 0.1 M GuNCS and 0.5 M TBP under an illumination of 0.1 Sun (Z907Na dye). Parameters JSC [mA.cm–2] VOC [mV] ff g [%]

1: 1 M[a]

3: 1.5 M[b]

4: 2 M[c]

1.6 701 0.72 8.5

0.9 829 0.76 6.2

0.8 909 0.44 3.6

[a] The modified-silica loading represents 31 wt % of the total electrolyte mass in electrolyte 1. [b] Electrolyte 3 contained 41 wt % of hybrid B to yield a concentration of 1.5 M in acetonitrile. [c] To prepare a 2 M solution of B in acetonitrile, addition of 48 wt % of the hybrid compound was necessary.

was counter-balancing decrease in photocurrent and hence a drop in efficiency of the corresponding cells. The lowest efficiency is observed for composition 4, which in addition to the low current density also resulted in cells with poor fill factors (0.44). This may be the result of a slower equilibration rate within the cell but there were also experimental difficulties encountered when trying to fill the cell with this highly viscous electrolyte. When electrolyte 3 is used to prepare the cell, a very good fill factor value of 76 % is attained as well as a high potential of 829 mV and a current of about 1 mA cm–2, giving an efficiency of 6.2 %. For this electrolyte, 3, the mass of the solid components (B plus other additives) is the same as the mass of solvent while for electrolyte 4, B is the dominant component (48 %, cf., 46 % acetonitrile and 6 % additives). As the modified-silica B becomes the major component of the whole electrolyte system, the high viscosity of the resultant electrolyte was found to cause problems with cell assembly. The lowest content of silica particles in the electrolyte system gives rise to the best solar cell efficiency, mainly due to a better short circuit current. Despite the increases in open circuit voltage in

3204 www.afm-journal.de

going from electrolyte 1 to 3 and 4, lower efficiencies are found for electrolytes 3 and 4 due their lower fluidity, which may reduce the electron transport but also slow down the I–/I3– exchange within the cell. 2.2.4. Influence of the Silica Content in the Organic-Inorganic Quaternary Ammonium Iodide Salts Increasing the hybrid concentration in the electrolyte solution should bring different results from varying the silica content of the compound itself. Indeed, when preparing electrolyte at the same concentration with B (1:1 molar ratio) and G (1:4 molar ratio) for example, the former should contain less silica than the latter and therefore modifications of the photovoltaic parameters of the corresponding cells should be encountered. Unfortunately, preparation of electrolyte solutions at a concentration of 1 M in acetonitrile failed, as too much silica was present in compounds G (54 %), H (67 %) and K (45 %). To prepare a 1 M solution of K in acetonitrile, 125 mg need to be dissolved in 78.6 mg of AN, which corresponds to a 58.5 wt % hybrid loading in AN. In this case, the solvent merely wets the hybrid giving a nonhomogeneous electrolyte system not exploitable for testing. It would have been judicious to lower the concentration of the hybrid ammonium iodide salts in the electrolyte system however no comparison would have been possible in the conditions studied here. Consequently, more work is required on compounds G, H, and K to be able to prepare totally solid electrolyte systems.

4. Conclusions A new type of nanosized quaternary ammonium iodide salts has been synthesized and characterized. They consisted of silica bearing propyl-triethyl, triheptyl and tridodecylammonium moieties that can be used as iodide sources in the preparation of electrolyte solutions for dye-sensitized solar cells. Two approaches were applied: (i) initial activation of silica was followed by the chemical grafting of aminopropyltriethoxysilane onto the surface, followed by the reaction of different alkyl iodides with the modified silica; (ii) new alkoxysilane precursors bearing ethyl, heptyl, and dodecyl side chains were synthesized first and then grafted on activated silica nanoparticles. The second pathway appears to give better control in preparing high silica content hybrids as it is easier to follow the reaction completion before grafting with the oxide network. This shows that a wide range of new materials can be prepared with different silica/organic moiety ratio, leading to a wide variety of new iodide sources that can be used as electrolytes in a variety of applications. Photovoltaic performances of DSSCs containing these quaternary ammonium iodide silica based compounds were studied as well as the influence of the solvent, the nature of the dye used, the concentration of hybrid and the silica content. The best efficiency of 8.6 % is obtained for solar cells containing the triethylammonium iodide salt at a concentration of 1 M either in AN or MPN solvent under an illumination of 0.1 Sun.

© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Funct. Mater. 2007, 17, 3200–3206

S. Cerneaux et al./Novel Nano-Structured Silica-Based Electrolytes

5. Experimental Materials: Nanosized silica particles (Aerosil 200, Degussa) with a mean particle size of 12 nm and specific surface area of 200 m2 g–1 were used in this work. 3-aminopropyltriethoxysilane (APTS) (> 97 %), ethyliodide (EI), 2-iodopropane (IP), heptyliodide (HI), dodecyliodide (DI), and guanidinium thiocyanate (GuNCS) were purchased from Aldrich and used without further purification under nitrogen atmosphere. Potassium carbonate was obtained from ABCR. 3-methoxyacetonitrile (MPN, Fluka), 4-tert-butylpyridine (TBP, Fluka) were distilled before use. Acetonitrile (AN), ethanol (99 %), chloroform and dichloromethane (DCM) were of analytical grade and used as received under a nitrogen atmosphere. The amphiphilic Z-907Na dye, [cis-Ru(HNadcbpy)(dnbpy)(NCS)2], where the ligands HNadcbpy and dnbpy correspond to the mono-sodium salt of 4,4′-dicarboxylic acid-2,2′-bipyridine and 4,4′-dinonyl-2,2′-bipyridine, respectively] was synthesized as described in the literature [19]. The homoleptic dye N719, (Bu4N)2[Ru(Hdcbpy)2(NCS)2] in which the second bipyridyl ligand also bears two carboxylate pendant arms instead of the long alkyl chains (dnpby), was used where specified. Characterization Techniques: Solid-state 13C and 29Si CP-MAS NMR spectra were recorded on a Brüker AM300 instrument equipped with a 4 mm solid-state probe operating at 75.5 MHz for 13C and 59.6 MHz for 29Si nuclei. Chemical shifts were referenced to external samples of glycine and tetramethylsilane. The 13C signal was enhanced using cross polarization techniques, where the 1H polarization is transferred to the 13 C nuclei. The spectra presented were collected using a magic angle spinning speed of 8 kHz, which was sufficient to eliminate spinning side bands. For the 13C NMR spectra, a contact time of 1000 ls was applied, with a 3.75 ls pulse, a recycle delay of 1 s and line broadening of 100 Hz. For the 29Si NMR spectra, a pulse length of 3.5 ls was used, with a delay time of 20 s and 100 Hz line broadening. Solution phase 1H and 13C NMR spectra were measured on an Advance DRX-400 MHz or a 300 MHz Brüker spectrometer at T = 298 K. The samples were analyzed in deuterated chloroform (CDCl3) under a nitrogen atmosphere and the chemical shifts d reported in ppm (parts per million) in reference to tetramethylsilane (TMS) used as an internal standard. The abbreviations for the peak multiplicities are as follows: s (singlet), d (doublet), dd (doublet of doublets), t (triplet), q (quartet), m (multiplet), and b (broad). Infrared spectra were recorded on a Perkin-Elmer 1600 series FTIR spectrophotometer as KBr pellets at a resolution of 4 cm–1. Thermogravimetric analyses (TGA) were carried out on a Perkin–Elmer apparatus under N2 flow at a scanning rate of 10 °C min–1 over a temperature range of 25 to 800 °C.

Adv. Funct. Mater. 2007, 17, 3200–3206

Electrospray mass spectra of positive and negative ions were measured on samples dissolved in CHCl3 using a Micromass Platform mass spectrometer fitted with an electrospray source. Syntheses of Modified Siloxanes and Silica Materials: Amino-Modified Silica Nanoparticles A: 2.5 g (42.0 mmol) of silica nanoparticles were dispersed in 50 mL of HCl (10 % v/v) and left stirring for 1 h. The gel obtained following evaporation of the solvent under vacuum was washed several times with water until the pH of the filtrate was 4–5. The wet gel was suspended in 40 mL of dry ethanol and 9.8 mL (42.0 mmol) of 3-aminopropyltriethoxysilane (APTS) were added dropwise. The reaction was stirred 48 h at room temperature until evaporation of the solvent. The compound was then washed with 50 mL of chloroform to eliminate unreacted alkoxysilane molecules and dried under vacuum to yield a fine white powder after grinding. Characterization data: C3H8NSi2O5 Yield: 96 %, 7.8 g (40.0 mmol). FTIR m (cm–1): 3380 and 3273, w, NH2; 2974, 2952, 2904, 2810, m, CH2; 1591, s, CN; 1200–1000, vb, SiOSi and SiOC; 795, s, SiC. 13C CP/MAS NMR spectrum, d (ppm): 8.3 (Si–CH2), 22.9 (CH2–CH2–CH2), 41.2 (CH2–CH2–N). 29Si MAS NMR spectrum, d (ppm): –51 T1 [Si(OCH2CH3 or OH)2 (OSi)1], –60 T2 [Si(OCH2CH3 or OH)1(OSi)2], –68 T3 [Si(OEt or OH)0(OSi)3], –110 Q4 (SiOSi). N,N,N-Triethyl-3-silica-propan-1-aminium Iodide B: 6.7 g (34.5 mmol) of A were dispersed in 200 mL of acetonitrile (AN) in presence of K2CO3 (19.9 g, 144 mmol) and 5.4 mL of ethyliodide (67.5 mmol). The reaction vessel was covered with an aluminium foil to prevent photo-oxidation of iodide, and the mixture was refluxed 48 h at 70–75 °C to yield a highly hygroscopic yellow powder after successive extraction of the base with dichloromethane (DCM) and evaporation of the solvent at 60 °C. Characterization data: C9H21NISiO2 Yield: 53 %, 6.3 g (18.1 mmol). FTIR m (cm–1): 2977, 2952, 2904, 2810, m, CH2, CH3; 1200–1000, b, SiOSi and SiOC; 809, s, SiC. 13C CP/MAS NMR spectrum, d (ppm): 6.4 (N–CH2–CH3) 8.3 (Si–CH2), 22.9 (CH2–CH2–CH2), 41.2 (CH2–CH2–N), 54.6 (N–CH2–CH3). 29Si MAS NMR spectrum, d (ppm): –69 T3 [Si(OSi)3]. Elemental analysis: % C: 30.7; % H: 5.6; % N: 4.3; % I: 34.4. Calcd for C9H21NISiO2: % C: 31.2; % H: 6.1; % N: 4.0; % I: 36.6. TGA weight loss between 25 and 800 °C: 65.2 wt %. N,N,N-Triheptyl-3-silica-propan-1-aminium Iodide C: After suspension of 1.3 g of A (6.7 mmol) in 40 mL of AN with 4.0 g of K2CO3 (28.9 mmol) and addition of 2.3 mL (14.0 mmol) of C7H15I, the mixture was refluxed four days at 70 °C. Compound C was obtained after extraction of the base with 200 mL of DCM and evaporation of the solvent and drying under vacuum at 70 °C. Characterization data: C24H51NISiO2 Yield: 55 %, 2.0 g (3.7 mmol). 1H NMR spectrum (CDCl3), d (ppm): 0.5 (b, 2H, Si–CH2), 0.9 (t, 9H, N–(CH2)6–CH3), 1.3 (s, 24H, N–CH2–CH2–CH2–(CH2)3–CH3), 1.4 (s, 6H, N–(CH2)2–CH2– (CH2)3–CH3), 1.7 (b, 2H, CH2–CH2–CH2), 2.4 (b, 6H, N–CH2–(CH2)5– CH3), 3.4 (CH2–CH2–CH2–N). 13C NMR (CDCl3), d (ppm) spectrum: 14.2 (N–(CH2)6–CH3), 22.8 (N–(CH2)2–CH2–(CH2)2–CH2–CH3), 26.6 (N–CH2–CH2–(CH2)4–CH3), 29.0 (N–(CH2)3–CH2– (CH2)2–CH3), 31.8 (N–(CH2)4–CH2–CH2–CH3), 59.7 (N–CH2–(CH2)5–CH3). N,N,N-Triisopropyl-3-silica-propan-1-aminium Iodide D: 1 mL (10.0 mmol) of C3H7I was added dropwise to 0.5 g of A (2.6 mmol) dispersed in 15 mL of AN with 1.5 g of K2CO3 (108.5 mmol). The reaction was kept under reflux 72 h at 70–75 °C covered with an aluminium foil. Compound D was finally obtained following the same procedure as used for compound C. Characterization data: C12H27NISiO2 Yield: 21 %, 0.2 g (0.54 mmol). FTIR m (cm–1): 2969, 2913, 2848, 2783, m and w, CH2, CH3; 1200–1000, vb, SiOSi and SiOC; 792, s, SiC. 13C CP/MAS NMR spectrum, d (ppm): 13 (Si–CH2), 28 (CH2–CH2–CH2) and (N– CH– (CH3)2), 48 (CH2–CH2–N) and (N–CH– (CH3)2). 29Si MAS NMR spectrum, d (ppm): –69 T3 [Si(OEt or OH)0(OSi)3], –109 Q4 (SiOSi). TGA weight loss between 25 and 800 °C: 73.4 wt %. Characterization Data for N,N,N-Triethyl-3-(triethoxysilyl)propan-1aminium Iodide E: C15H36NISiO3 Yield: 91 %, 6.8 g (15.6 mmol). 1H NMR (CDCl3), D (ppm) spectrum: 0.6 (t, 2H, Si–CH2), 1.1 (t, 9H, O– CH2–CH3), 1.3 (t, 9H, N–CH2–CH3), 1.8 (m, 2H, CH2–CH2–CH2), 3.2 (t, 2H, CH2–CH2–N), 3.4 (q, 6H, N–CH2–CH3), 3.7 (q, 6H, O–CH2– CH3). 13C NMR (CDCl3), D (ppm) spectrum: 7.1 (Si–CH2), 8.2 (N– CH2–CH3), 16.3 (CH2–CH2–CH2), 18.4 (O–CH2–CH3), 47.1 (CH2– CH2–N), 54.1 (N–CH2–CH3), 58.8 (O–CH2–CH3). Electrospray mass

© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

FULL PAPER

At low irradiance, the functionalized nanoparticles facilitate ion transport through the cell leading to high currents and good fill factors. At higher light intensity mass transfer limitations are encountered for triiodide reduction at the counter electrode. The new quaternary ammonium silica-based nanoparticles are highly promising as iodide source in the solar cells as high voltages can be reached at low irradiance, suitable for indoor application in particular. The electrolytes prepared using these new iodide sources improve cell performance by reducing charge recombination (particularly evident for the Z907Na dye) and by increasing VOC presumably through tuning of either the I–/I3– potential or the semiconductor band edges. These aspects require further investigation but these preliminary results are very encouraging, since there is scope for optimization of electrolyte compositions. Indeed, the photocurrent can easily be improved by varying the iodide concentration and/or the I–/I3– ratio and other additives can also be introduced to subsequently ameliorate the solar cells efficiencies.

www.afm-journal.de 3205

FULL PAPER

S. Cerneaux et al./Novel Nano-Structured Silica-Based Electrolytes spectrum (m/z) (positive ion): 306.4/307.4/308.4 [M]+ (100 %). Electrospray mass spectrum (m/z) (negative ion): 126.8 [I]– (100 %) [12]. Characterization Data for N,N,N-Tridodecyl-3-(triethoxysilyl)propan-1-aminium Iodide F: C45H96NISiO3 Yield: 98 % (16.8 mmol, 14.4 g). 1H NMR spectrum (CDCl3), d (ppm): 0.7 (t, 2H, Si–CH2), 0.9 (t, 9H, N– (CH2)11–CH3), 1.2 (t, 9H, O–CH2-CH3), 1.3 (m, 12H, N– CH2–CH2–(CH2)3–(CH2)6–CH3), 1.4 (m, 6H, N–CH2–CH2–(CH2)3– (CH2)6–CH3), 1.7 (m, 2H, CH2–CH2–CH2), 1.8 (m, 2H, N–CH2–CH2– (CH2)3–(CH2)6–CH3), 3.3 (m, 4H, CH2–N–CH2–(CH2)10–CH3), 3.8 (q, 6H, O–CH2–CH3). 13C NMR spectrum (CDCl3), d (ppm): 7.2 (Si– CH2), 14.1 (N–(CH2)11–CH3), 18.3 (O–CH2–CH3), 22.2 (CH2–CH2– CH2), 22.6 (N–(CH2)10–CH2–CH3), 26.0–30 (N– (CH2)2–(CH2)7– (CH2)2–CH3), 31.9 (N–(CH2)9–CH2–CH2–CH3), 33.6 (N–CH2–CH2– (CH2)9–CH3), 58.7 (O–CH2–CH3), 59.8 (CH2–N–CH2–(CH2)10–CH3). Electrospray mass spectrum (m/z) (positive ion): 726.7/727.7/728.7 [M]+ (100 %). Electrospray mass spectrum (m/z) (negative ion): 126.7 [I]– (100 %). N,N,N-Triethyl-3-silica-propan-1-aminium Iodide G and H: 0.6 g of silica nanoparticles (10.0 mmol) were suspended in 10 mL of AN after the activation step in acidic media as described for the synthesis of compound A. 1.0 g (2.4 mmol) and 0.5 g (1.2 mmol) of functionalized alkoxysilane E were added under nitrogen atmosphere to the suspension of silica to synthesize materials G and H, respectively. The reactions were kept under reflux 24 h at 70–75 °C whilst covered with an aluminium foil and the resulting light yellow powders recovered by filtration after rinsing with acetone. Yield of G, corresponding to C9H21NISi5O11, 71 %, 1.0 g (1.7 mmol) and yield of H, corresponding to C9H21NISi9O19, 92 %, 0.9 g (1.1 mmol). 13C CP/MAS NMR spectrum, d (ppm): 10 (N–CH2–CH3) 16 (Si–CH2), 32 (CH2–CH2–CH2), 48 (CH2–CH2–N), 54 (N–CH2–CH3). 29Si MAS NMR spectrum, d (ppm): –58 T2 [Si(OCH2CH3 or OH)1(OSi)2], –67 T3 [Si(OEt or OH)0(OSi)3], –109 Q4 (SiOSi). TGA weight loss between 25 and 800 °C: 20–30 wt %. N,N,N-Tridodecyl-3-silica-propan-1-aminium Iodide K: To 1.0 g of silica (16.6 mmol), dispersed in 20 mL of acetonitrile, 1.8 g (2.2 mmol) of alkoxysilane F were added. Since F was not soluble in AN, 10 mL of chloroform were added to the reaction mixture that was refluxed at 70–75 °C for 24 h. 3.0 g of compound K, corresponding to C39H81NISi9O19, were collected after evaporation of the solvent under vacuum at 65 °C that were further cleaned by suspending it in acetone and filtering. Yield: 45 %, 1.2 g (1.0 mmol). Electrolyte Solution Preparation: Electrolyte 1 was prepared by mixing 40.6 mg of compound B (1 M), 2.0 mg of I2 (0.08 M), 1.2 mg of guanidinium thiocyanate (0.1 M) and 6.8 mg of 4-tert-butylpyridine (0.5 M) in 0.1 mL of acetonitrile (AN). Hybrid compound B, in which 33 weight percent (wt %) of inorganic loading was present, represented 31 wt % of the total electrolyte solution. When MPN solvent was used, electrolyte 2 was obtained, which was composed of 40.6 mg of B (1 M), 2.0 mg of I2 (0.08 M) and 6.8 mg of TBP (0.5 M) with a modified silica loading of about 28 wt % of the total electrolyte system. Two further electrolytes denoted 3 and 4 were prepared by mixing 2.0 mg of I2 (0.08 M), 1.2 mg of guanidinium thiocyanate (0.1 M) and 6.8 mg of 4-tert-butylpyridine (0.5 M) with 60.9 mg and 81.3 mg of compound B, respectively, in 0.1 mL of acetonitrile. The hybrid loading in electrolytes 3 and 4 was 41 wt % and 48 wt %, respectively, corresponding to increasing concentration of B of 1.5 M and 2.0 M, respectively. Solar Cell Fabrication: A nanoporous TiO2 layer was coated onto Fdoped SnO2 glass substrate as described in previous papers, and then sintered at 500 °C in air [20–22]. Dye absorption was carried out overnight at room temperature by dipping the working electrodes into 3 × 10–4 M solutions of either Z907Na (in acetonitrile:tert-butyl alcohol (1:1 v/v or N719 (in ethanol) where specified. An acidic solution of Pt (0.05 M H2PtCl6) was used to prepare the counter electrodes that were heated at 400 °C for 10 min before assembling. The devices were then filled with the electrolyte solution and sealed using a 35 lm-thick hotmelt ring Bynel (Dupont, USA).

Photocurrent–Voltage Measurements: The current–voltage (I–V) curves were obtained by measuring the photocurrent of the cells using a Keithley model 2400 digital source meter (Keithley, USA) under an applied external potential scan. The irradiation source for the photocurrent–voltage measurement is a 450 W xenon light source (Osram XBO 450, USA), which simulates the solar light. The incident light from a 300 W xenon lamp (ILC Technology, USA) was focused through a Gemini-180 double monochromator (Jobin Yvon Ltd., UK) onto the cell under test (illuminated area of 0.158 cm2). Light intensity was varied from 100 mW cm–2, the equivalent of 1 Sun at AM 1.5G, to 0.1 Sun with neutral wire mesh attenuators in front of the light source. The overall conversion efficiency g of the photovoltaic cell is calculated from the integral photocurrent JSC, the open-circuit photovoltage VOC, the fill factor of the cell ff and the intensity of the incident light. Received: April 5, 2007 Revised: May 24, 2007 Published online: September 5, 2007

– [1] B. O’Regan, M. Grätzel, Nature 1991, 353, 737. [2] M. Grätzel, J. Photochem. Photobiol. C 2003, 4, 145. [3] S. Tan, J. Zhai, M. Wan, Q. Meng, Y. Li, L. Jiang, D. Zhu, J. Phys. Chem. B 2004, 108, 18 693. [4] R. Kawano, H. Matsui, C. Matsuyama, A. Sato, Md. A. B. H. Susan, N. Tanabe, M. Watanabe, J. Photochem. Photobiolol. A 2004, 164, 87. [5] N. Yamanaka, R. Kawano, W. Kubo, T. Kitamura, Y. Wada, M. Watanabe, S. Yanagida, Chem. Commun. 2005, 740. [6] M.-S. Kang, J. H. Kim, J. Won, N.-G. Park, Y. S. Kang, Chem. Commun. 2005, 889. [7] K. Hayamizu, Y. Aihara, H. Nakagawa, T. Nukuda, W. S. Price, J. Phys. Chem. B 2004, 108, 19 527. [8] Y. J. Kim, J. H. Kim, M.-S. Kang, M. J. Lee, J. Won, J. C. Lee, Y. S. Kang, Adv. Mater. 2004, 16, 1753. [9] J. Xia, F. Li, C. Huang, J. Zhai, L. Jiang, Sol. Energy Mater. Sol. Cells 2006, 90, 944. [10] P. Wang, S. M. Zakeeruddin, M. Grätzel, J. Fluorine Chem. 2004, 125, 1241. [11] B. Xue, H. Wang, Y. Hu, H. Li, Z. Wang, Q. Meng, X. Huang, L. Chen, O. Sato, A. Fujishima, C. R. Chim. 2006, 9, 627. [12] E. Stathatos, P. Lianos, S. M. Zakeeruddin, P. Liska, M. Grätzel, Chem. Mater. 2003, 15, 1825. [13] E. Stathanos, P. Lianos, V. Jovanovski, B. Orel, J. Photochem. Photobiol. A 2005, 169, 57. [14] S. Cerneaux, S. M. Zakeeruddin, M. Grätzel, Y.-B. Cheng, L. Spiccia, unpublished. [15] Z.-S. Wang, H. Sugihara, Langmuir 2006, 22, 9718. [16] N. Kopidakis, N. R. Neale, A. J. Frank, J. Phys. Chem. B 2006, 110, 12 485. [17] D. Kuang, P. Wang, S. Ito, S. M. Zakeeruddin, M. Grätzel, J. Am. Chem. Soc. 2006, 128, 7732. [18] M. Jijitori, M. Ikegami, T. Miyasaka, Chem. Lett. 2007, 190. [19] P. Wang, B. Wenger, R. Humphry-Baker, J. E. Moser, J. Teuscher, W. Kantlehner, J. Mezger, E. V. Stoyanov, S. M. Zakeeruddin, M. Grätzel, J. Am. Chem. Soc. 2005, 127, 6850. [20] C. J. Barbé, F. Arendse, P. Comte, M. Jirousek, F. Lenzmann, V. Shklover, M. Grätzel, J. Am. Chem. Soc. 1997, 119, 3157. [21] P. Wang, S. M. Zakeeruddin, P. Comte, R. Charvet, R. Humphry-Baker, M. Grätzel, J. Phys. Chem. B 2003, 107, 14 336. [22] D. Kuang, S. Ito, B. Wenger, C. Klein, J.-E. Moser, R. Humphry-Baker, S. M. Zakeeruddin, M. Grätzel, J. Am. Chem. Soc. 2006, 128, 4146.

______________________

3206 www.afm-journal.de

© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Funct. Mater. 2007, 17, 3200–3206