Fluoride additive in epoxide-initiated sol–gel synthesis enables thin-film applications of SnO₂ aerogels

Abstract

Aerogels of SnO₂ were synthesized by an epoxide-initiated sol–gel method. Using ammonium fluoride in the precursor solution allowed for tunability of the aerogel morphology while no change in the conductivity was measured. In particular, aerogel shrinkage was decreased dramatically by the addition of the fluoride precursor. Unfluorinated aerogels showed severe shrinkage of 43% volume change upon supercritical drying compared to the original alcogel volume. Fluorinated samples exhibited a much less pronounced shrinkage at 7%. Multiple characterization methods converged to reveal the mechanism by which fluoride enables the morphological tunability. These findings enable the casting of SnO₂ aerogels as thin films (which in the absence of fluoride these crack and delaminate due to shrinkage), opening potential uses in many optoelectronic devices including solar cells.

---

Introduction

Sol–gel chemistry has become widely used for the synthesis of metal oxide nanoparticles that are cross-linked into a gel, and used ultimately either as a dry gel or a powder. Epoxides have been proven to be effective at initiating the formation of the gel by acting as mild proton scavengers, maintaining an elevated pH to promote hydrolysis of metal precursors. While the hydrated metal is deprotonated, it is linked with other hydrated metals via olation and oxolation to form metal oxide particles.^1,2 This facile sol–gel process has been widely used in metal oxide nanoparticle synthesis^2,3 due to its ease of preparation and relatively low cost of the metal salt precursors.

Tin oxide (SnO₂) is an n-doped, wide bandgap (3.6 eV at room temperature) semiconductor widely used in solar cells,^4–6 water splitting,⁷ optoelectronic devices,⁸ gas sensors,⁹ and transparent conducting oxides (TCOs).¹⁰ Heat treatment of mesoporous tin oxides is required for improving conductivity or crystallinity.^11–15 Heating SnO₂ materials made by the epoxide-initiated sol–gel method induces undesirable morphological features, in particular, shrinkage resulting in cracking and, in the case of films, delamination.^16,17 The morphology of nanostructures is known to affect critical material properties and has become a key component in the synthesis of nanoparticulate materials.^18,19 In order to take full advantage of such materials one must be able to manipulate the pore structure, surface area, particle size and crystallinity.

Here, we report on the use of fluorine to control SnO₂ aerogel morphology. We prepared SnO₂ alcogels, by the epoxide-assisted sol–gel process modified by the inclusion of ammonium fluoride (NH₄F), and dried them using supercritical CO₂ to form aerogels. The fluoride profoundly affected aerogel properties such as shrinkage, density, porosity and surface area. These changes were extensively characterized, and the mechanism that triggers this was unraveled. Critically, fluoride allows the aerogels to be cast as thin films on glass substrates by greatly reducing the shrinkage of the gel during supercritical drying, which in unflorinated samples results in cracking and delamination. Thin-film aerogels open a wide array of applications. For example, we have shown that thin-film SnO₂ aerogels, when also doped with Sb(V) for electrical conductivity, function as mesoporous TCOs, which we have used in dye-sensitized solar cells.¹⁶

Results and discussion

The effect of the precursor F : Sn ratio on gel time and gel volume change is shown in Table 1. The increase in F : Sn consistently decreased the gel time from 300 s for unfluorinated samples to 105 s for samples containing a 1 : 1 F : Sn ratio. Similar trends have been observed in silica sols, which gel faster with the addition of F precursors.^20–22 The effect of F in gelation times of silica has been studied and it is well known that F catalyzes the gelation process. However, the mechanisms that drive this are still not well understood. Based on the study by Gash et al.² it is possible that F⁻ ions are able to ring-open the epoxide readily. Vincent et al.,²³ have shown that the fluoride ion can be a good nucleophile when in the presence of water, which our precursor solutions contain. In addition, other groups^24,25 have shown that fluoride-containing compounds are excellent epoxide ring-openers. It is therefore possible that the use of fluoride ions can reduce the time required to induce olation and oxolation, which in turn yields the wet-gel.

Table 1 Effect of NH₄F in sol–gel processing and shrinkage (ΔV/V₀) of aerogels after supercritical drying (SCD) and calcination at 450 °C

F : Sn	Gel time (sec)	ΔV/V0 (%) after SCD	ΔV/V0 (%) 450 °C
0 : 1	300	43%	90%
0.25 : 1	240	39%	88%
0.50 : 1	180	32%	85%
0.75 : 1	150	17%	78%
1 : 1	105	7%	73%

---

Unflorinated SnO₂ gels made in this study suffered from visible shrinkage during the ethanol and acetone exchange step even before supercritical drying (SCD). This effect is due to the reorganization of clustered nanoparticles as solvents diffuse through the pores,²⁶ and was mitigated by the addition of F in solution. The measured shrinkage for samples with varying concentration of F after the SCD and calcination steps is presented in Table 1. Shrinkage following SCD decreased markedly with increasing F content from 43% in fluorine-free samples to only 7% shrinkage in samples made with a 1 : 1 F : Sn ratio. Calcination of aerogel materials at 450 °C induced severe shrinkage for all samples, although it too was diminished by the presence of fluorine (Table 1). Even though samples made with 1 : 1 F : Sn exhibited a 73% reduction from their initial volume when calcined at 450 °C, that level of F addition enabled the aerogels to be cast as thin films without delamination, unlike unflorinated gels.¹⁶

Fig. 1 shows TEM (a–c and e–g) and SEM (d and h) images, performed to examine the porous structure of unfluorinated (Fig. 1a–d) and fluorinated (at a F : Sn precursor ratio of 1 : 1, Fig. 1e–h) aerogels. A clear effect of fluorine on morphology is observed. TEM images of uncalcined unfluorinated materials show large agglomerations with no clear mesoporosity (Fig. 1a). In contrast, the uncalcined fluorinated samples (Fig. 1e) exhibited less agglomeration and more defined mesoporosity. Both materials, after calcination at 450 °C for 30 min (Fig. 1b, c, f and g), show well-defined particles and porosity, and a branched structure composed of spherical interconnected particles, as seen by others.³ The aerogels exhibit mesopores in the range of 10–70 nm, similar to an earlier report.² A clear decrease in particle size is observed with the addition of F in the materials. The crystallites exhibit a highly crystalline nature with well-defined lattice fringes (Fig. 1c and g).

Fig. 1 Morphological study by electron microscopy for unfluorinated (a–d) and fluorinated at 1 : 1 F : Sn (e–h) aerogels. TEM images of aerogels uncalcined (a and e) and calcined at 450 °C for 30 min (b, c and f, g). SEM images of aerogels calcined at 450 °C for 30 min (d and h).

SEM images of the calcined unfluorinated and fluorinated SnO₂ aerogels (Fig. 1d and h, respectively) show an open pore structure typical of aerogels. Qualitative changes in morphology from fluorine addition are apparent, which are consistent with results shown in Table 1. Fluorinated samples have larger pore openings, whereas the unfluorinated analogues show more compact structures. These images suggest that the shrinkage exhibited by the unfluorinated samples is caused by the contraction of the larger pores, which are preserved in the fluorinated samples. Macroscopically, the aerogels also exhibit large differences with respect to F addition. As seen in Fig. 2, unfluorinated samples suffered from large shrinkage in the bulk and cracking and delamination in thin films deposited on FTO glass. Addition of F at a F : Sn of 0.5 : 1 yielded less pronounced shrinkage in bulk samples but cracking and delamination were still prevalent. Samples with a F content of 1 : 1 of F : Sn yielded little shrinkage and no cracking or delamination of the SnO₂ films deposited on FTO glass. The films were able to withstand SCD, sintering at 450 °C and dye sensitization to yield monolithic aerogel films, which were semi-transparent in nature.

Fig. 2 Photographs of bulk aerogels and deposited as thin films on FTO glass. The figure shows the optimum 1 : 1 F : Sn to avoid structural shrinkage which in turn allows film formation, heat treatment and dye-sensitization.

The isotherms for both uncalcined and calcined samples were type IV with a H1 hysteresis typical of mesoporous materials with spherical particles, as seen in Fig. 3a and b. The pore size distribution confirmed the materials are in the mesoporous range with high surface area for uncalcined aerogels, the addition of F consistently decreased the surface area and pore volume. The addition of F had a pronounced effect in eliminating pores (Fig. 3c), especially in the smaller pore size range (<8 nm). This suggests that fluorinated particles are more interconnected than the unfluorinated counterparts. In contrast, the presence or absence of F yielded no significant change in BET surface area for aerogel samples calcined at 450 °C (Fig. 3d). Calcination at 450 °C changes the pore size distribution dramatically. The smaller pores seen in uncalcined samples almost completely disappear while pores in the mesoporous region become dominant for calcined samples.

Fig. 3 Nitrogen physisorption isotherms and pore distribution analysis for SnO₂ aerogels as a function of F : Sn precursor ratio, uncalcined (a and c) and calcined at 450 °C (b and d).

To understand the effect of F addition to the electrical properties of the SnO₂ aerogels we carried out 4-point probe measurements. The resistivities measured for uncalcined materials were very high, approaching 10¹⁰ Ω cm with high variability, regardless of fluorine content. This is most likely due to the amorphous nature of the nanoparticle structure and the organic by-products present. A 3 order of magnitude decrease in resistivity was seen for materials calcined at 450 °C for 30 min. No change in resistivity was observed with the increase of fluorine content in the aerogels. All calcined samples exhibited resistivities in the range of 10⁶ Ω cm. This is not uncommon for SnO₂ nanoparticle arrays which typically exhibit resistivities in the 10⁵ to 10⁷ range.^12,13 Avadhut et al.²⁷ also found that the addition of fluorine did not change the electronic properties of the SnO₂ nanoparticles unless they were processed in a reducing environment in order to create oxygen vacancies and thus reduce the resistivity.

Solid state MAS NMR spectroscopy is shown in Fig. 4 and was used to identify the unfluorinated and fluorinated SnO₂ species before and after calcination. The ¹¹⁷Sn NMR spectra (Fig. 4) of calcined materials show sharp peaks at −605.4 (Fig. 4a) and at −605.0 ppm (Fig. 4b) for unfluorinated and fluorinated materials, respectively. This slight shift in the Sn signal is attributed to the interaction of Sn with F in fluorinated samples, as shown by Avadhut et al.²⁷ The position of the peaks at −605 ppm corresponds to the cassiterite SnO₂, in agreement with the XPS spectra (Fig. S1†) showing Sn(V) and the XRD patterns (Fig. S2†) showing the cassiterite structure.^27,28 The intensity of the peak for fluorinated SnO₂ samples, both uncalcined (Fig. 3b and d) and calcined, is consistently lower than that for unfluorinated samples, and this can be attributed to the doping-induced disorder in the nanoparticle crystal structures.²⁷ Uncalcined materials show broad peaks at −609.1 (Fig. 4c) and −623.8 ppm (Fig. 4d) for unfluorinated and fluorinated samples, respectively. This is also attributed to the amorphous nature of the uncalcined samples, as was indicated by X-ray diffraction.

Fig. 4 ¹¹⁷Sn MAS NMR (left panel) of unfluorinated (a and c) and fluorinated samples at a precursor ratio of F : Sn of 1 : 1 (b and d) for calcined (a and b) and uncalcined (c and d) materials. ¹³C MAS NMR (right panel) of uncalcined unfluorinated (bottom) and fluorinated (F : Sn precursor ratio of 1 : 1, top) samples, with a detailed view of the spectrum for fluorinated samples.

The ¹³C NMR in Fig. 4, showed clear patterns for uncalcined samples of fluorinated and unfluorinated SnO₂ aerogels. However, no signal was detected for calcined samples suggesting that most carbon had volatilized during the calcination process. Only two sharp peaks were detected at 67.6 and 18.2 ppm for unfluorinated samples, which match closely the peaks for propylene oxide (green structure in Fig. 4). No other significant C peaks were detected for these materials. The results are very different for fluorinated materials, where several additional peaks were detected. Using MNova NMRPredict software, two additional structures were predicted for the fluorinated samples. First, the peaks at 207.7, 52.6 and 25.1 ppm yielded a structure containing a three-carbon chain with fluorine on one end and a carbonyl (blue structure). The additional peaks at 64.1, 51.0 and 30.1 ppm correspond to a similar carbon structure containing a hydroxyl group (red structure). Both structures are expected from the ring-opening process during the nucleophilic attack from the fluoride ion in solution.²

To identify the species evolved during heating, temperature-programmed desorption-mass spectrometry (TPD-MS) was performed (Fig. 5). A significant portion of the mass lost from the fluorinated sample (Fig. 5a) is due to water evolution (m/z = 18). The broad signal between room temperature and 250 °C is due to surface adsorbed water. This is followed by a second less intense peak that is due to water being removed from the pores of the material as well as surface hydroxyl groups condensing to form Sn–O–Sn linkages. A similar desorption profile is observed for m/z = 17 due to the fragmentation of water to hydroxyl groups when exposed to the electron impact ion source, with an additional peak at 400 °C attributed to the desorption of ammonia. Sharp signals are also observed centered at 170 °C that indicate the release of a carbonaceous species consistent with a ring-opened PPO derivative. The parent ion (m/z = 58) is observed concurrently to two major signals at m/z = 43 and m/z = 15 attributed to methyl group fragmentation. Additional signals were seen at 190 °C (m/z = 38) and 550 °C (m/z = 85) and are attributed to the desorption of a small amount of diatomic fluorine and residual carbonaceous species, respectively. The former is supported by the XPS data showing a large decrease in F from uncalcined to calcined at 450 °C (Fig. S1†).

Fig. 5 TPD-MS curves for fluorinated (a, F : Sn 1 : 1) and unfluorinated (b) SnO₂.

The unfluorinated sample (Fig. 5b) exhibits a similar water desorption profile, though the ratio of surface-adsorbed water to pore-bound water and surface hydroxyl groups is reversed. In stark contrast to the fluorinated sample, however, no signals indicating desorption of PPO or a PPO-derivative were observed. The only carbon-containing species observed were CO and CO₂. The observation of oxidized carbon in an inert atmosphere suggests that the PPO decomposes on the surface of the unfluorinated material by removing oxygen from the structure, inducing cracking and shrinkage.

We propose a model, depicted in Fig. 6, to explain the results we have obtained in this study, as follows. Residual PPO at the surface of unfluorinated SnO₂ nanoparticles (Fig. 6, left) reacts with surface oxygen as the temperature is raised, removing oxygen from the structure and damaging it in the process. In fluorine-containing SnO₂ particles (Fig. 6, right), much of the fluorine is associated with the SnO₂ surface²⁷ and reacts with the PPO, leading to fluorinated PPO derivatives adsorbed to the SnO₂ via the F atom. During heat treatment, the C–F bond is severed, and the dehalogenated PPO derivatives escape, as does much of the fluorine. Some fluorine atoms deeper inside the nanoparticles remain in the structure. The result of these different mechanisms is that fluorine inhibits the removal of oxygen from the SnO₂ structure during calcination, reducing the degree of structural shrinkage.

Fig. 6 Schematic of unfluorinated (left) and fluorinated (right) samples depicting their respective interactions with carbonate species before (black) and during (red) calcination.

Conclusions

In summary, this study shows how to control morphological features of gels and aerogels by the addition of fluoride in solution to an epoxide-assisted sol–gel process. The mechanism that triggers this was unveiled by different characterization techniques. Incorporation of fluorine inhibited shrinkage after SCD and reduced the shrinkage during calcination. NMR results, taken together with the TPD data, suggest that a ring-opened and F-bonded PPO derivative is adsorbed via the F atom to the SnO₂ surface before calcination in the fluorinated materials. Upon heating, these molecules detach from the structure without taking oxygen atoms in the process. This mechanism prevents additional shrinkage that is induced by the removal of oxygen atoms from SnO₂, as seen in the unfluorinated materials. Finally, this facile method is promising for use in other sol–gel processes to control the morphological features of metal oxide structures or to control shrinkage in aerogels.

Experimental

Synthesis of aerogel materials

All chemicals were purchased from Sigma-Aldrich and were ACS grade or better. In a typical procedure, 0.6 M SnCl₄·5H₂O is dissolved in absolute ethanol and 7 M NH₄F in water.²⁹ The tin solution is placed in a water bath at 60 °C for 5 hours and the fluoride solution is added to it dropwise (varied over a precursor F : Sn ratio from 0 : 1 to 1 : 1). A white precipitate forms but redissolves after 2 hours. The sol–gel and supercritical drying procedures are explained in detail in our previous publication.³⁰ In a common procedure propylene oxide (PPO) is added (in 2 aliquots to avoid boiling over the synthesis solution due to rapid heat generation in the reaction) to the solution in an ice bath at a PPO : Sn ratio of 11 : 1.² The gelation occurs in 150 seconds after PPO addition, and the solution changes from a clear sol to a semi-transparent white gel. The reaction is allowed to proceed 20 min in a capped vial, which is subsequently filled with ethanol to age for 24 hours. The solvent and by-products are then washed with acetone 3 times over 3 days. The gels are transferred to a custom-built pressure vessel which is filled with liquid CO₂. The acetone is washed out by the liquid CO₂ over the course of 3 days and the vessel is heated above the critical point of the CO₂ (32 °C, 7.38 MPa). The temperature and pressure are held constant at about 45 °C and 9 MPa, respectively, for 60 minutes and then depressurized over 180 minutes to room temperature and pressure. The samples are then calcined at 450 °C in air for 30 minutes after heating at a rate of 3 °C min⁻¹.

Characterization

SEM (FEI Nova NanoSEM 450) was used to study the morphology of the aerogels and TEM (JEOL 2010 FasTEM operating at 200 kV accelerating voltage) to study particle shape and size. XRD was used to study the crystallinity of the samples using a Rigaku Ultima IV X-ray diffractometer using Cu Kα radiation (λ = 0.154178 nm) at a scanning rate of 0.02° s⁻¹ in the 2θ range from 20° to 60°. The crystallite size was determined by the Scherrer equation,³¹ , where D is the crystallite size, λ is the X-ray wavelength, β is the full width at half maximum (FWHM) and θ is half of the angle of diffraction. A Micromeritics ASAP 2020 accelerated surface area and porosimetry analyzer was used to determine the aerogels' Brunauer–Emmett–Teller (BET) specific surface area and Barrett–Joyner–Halenda (BJH) pore volume and size.

The aerogel monolith resistivity (ρ) was measured by the common 4-point probe technique.³² Tungsten needles were placed atop the monolithic aerogel disks (1.2 cm diameter by 1 cm in height) and room temperature resistance was obtained from low voltage IV curves (−0.5 to 0.5 V) through the aerogel structure using the formula , where s is the spacing of the probes (0.6 mm), F is the correction factor dependent on sample geometry (taken as unity due to the large size of samples compared to the probe spacing), and V/I were extracted from the measured data. Diffuse reflectance measurements of aerogel powders diluted 10 times in barium sulfate were made using a Shimadzu UV-2450 UV-vis spectrometer with an IRS-2200 integrating sphere. Thermogravimetric analysis (TGA) was completed using a TA Instruments TGA Q-500. Temperature programmed desorption-mass spectrometry (TPD-MS) was completed using an MKS PPT quadrupole Residual Gas Analyzer equipped with a vacuum sampling manifold that allows for the analysis of gas streams at or near atmospheric pressure. Powderized samples were loaded into a quartz tube and placed inside a programmable tube furnace. A 50 sscm flow of argon was maintained while the sample temperature was ramped at a rate of 10 °C min⁻¹ from room temperature to 800 °C.

Elemental composition was determined by X-ray photoelectron spectroscopy (XPS) with a PHI 595 Multiprobe system using an aluminum anode with K_α = 1486.6 ev. Ground aerogel samples were placed and analyzed on a double-sided carbon tape. The spectra were corrected for sample charging effects using the C 1s peak at 284.6 eV.

Solid state nuclear magnetic resonance

All solid-state NMR experiments were performed on a Bruker Avance III 400 MHz wide bore spectrometer operating at the frequencies of 100.6 MHz for ¹³C and 142.5 MHz for ¹¹⁷Sn (magnetic field strength B₀ = 9.4 T) using a 4 mm H/X–Y triple resonance DVT MAS probe at a spinning frequency of 15 kHz at room temperature. ¹¹⁷Sn MAS NMR spectra were acquired with a 90° pulse length of 4.0 μs, spectral width of 200 kHz, and a recycle delay of 15 s. Chemical shifts in ¹¹⁷Sn NMR spectra were referenced to neat tetramethyltin with respect to the signal of SnO₂ sample at −604.7 ppm. ¹H–¹³C CPMAS spectra were acquired with a ¹H 90° pulse length of 4.0 μs, contact time of 2 ms, spectral width of 30 kHz, recycle delay of 5 s and high power two-pulse phase modulation (TPPM) proton decoupling. Chemical shifts in ¹H–¹³C CPMAS spectra were referenced to tetramethylsilane (TMS) using the methyl carbons of 3-methylglutaric acid (MGA) at 18.84 ppm.

Acknowledgements

This material is based upon work supported by the National Science Foundation under Grant No. CBET-1332022. J. P. C. B was also supported by NSF Grant No. DGE-0947869. SLS and DAK acknowledge the support of the US Department of Energy, Office of Basic Energy Sciences, Division of Chemical, Geochemical, and Biological Sciences under grant DE-FGO2-86ER13622.A000. The authors thank Drs Ali Gokirmak and Helena Silva, and Lhacene Adnane for use of and help with their four-probe measurement apparatus; and Dr Heng Zhang for XPS assistance. XPS and TEM were measured at the Institute of Material Science and SEM, XRD and N₂ physisorption were performed at the Center for Clean Energy Engineering, at the University of Connecticut.

References

1. T. F. Baumann, S. O. Kucheyev, A. E. Gash and J. H. Satcher, Adv. Mater., 2005, 17, 1546–1548.
2. A. E. Gash, T. M. Tillotson, J. H. Satcher, J. F. Poco, L. W. Hrubesh and R. L. Simpson, Chem. Mater., 2001, 13, 999–1007.
3. M. Davis, K. Zhang, S. Wang and L. J. Hope-Weeks, J. Mater. Chem., 2012, 22, 20163–20165.
4. S. Ferrere, A. Zaban and B. A. Gregg, J. Phys. Chem. B, 1997, 101, 4490–4493.
5. A. Kay and M. Grätzel, Chem. Mater., 2002, 14, 2930–2935.
6. H. J. Snaith and C. Ducati, Nano Lett., 2010, 10, 1259–1265.
7. R. Saito, Y. Miseki and K. Sayama, Chem. Commun., 2012, 48, 3833–3835.
8. C. Tatsuyama and S. Ichimura, Jpn. J. Appl. Phys., 1976, 15, 843–847.
9. Z. Miao, Y. Wu, X. Zhang, Z. Liu, B. Han, K. Ding and G. An, J. Mater. Chem., 2007, 17, 1791–1796.
10. K. Ellmer, Nat. Photonics, 2012, 6, 809–817.
11. S. Wu, S. Yuan, L. Shi, Y. Zhao and J. Fang, J. Colloid Interface Sci., 2010, 346, 12–16.
12. V. Müller, M. Rasp, G. Štefanić, J. Ba, S. Günther, J. Rathousky, M. Niederberger and D. Fattakhova-Rohlfing, Chem. Mater., 2009, 21, 5229–5236.
13. T. Nütz, U. Zum Felde and M. Haase, J. Chem. Phys., 1999, 110, 12142–12150.
14. C. Goebbert, M. A. Aegerter, D. Burgard, R. Nass and H. Schmidt, J. Mater. Chem., 1999, 9, 253–258.
15. V. Müller, M. Rasp, J. Rathouský, B. Schütz, M. Niederberger and D. Fattakhova-Rohlfing, Small, 2010, 6, 633–637.
16. J. P. Correa Baena and A. G. Agrios, J. Phys. Chem. C, 2014, 118, 17028–17035.
17. J. P. Correa Baena and A. G. Agrios, ACS Appl. Mater. Interfaces, 2014, 6, 19127–19134.
18. A. S. Aricò, P. Bruce, B. Scrosati, J.-M. Tarascon and W. van Schalkwijk, Nat. Mater., 2005, 4, 366–377.
19. Y.-G. Guo, J.-S. Hu and L.-J. Wan, Adv. Mater., 2008, 20, 2878–2887.
20. D. Y. Nadargi, R. R. Kalesh and A. V. Rao, J. Alloys Compd., 2009, 480, 689–695.
21. J. Mrowiec-Białoń, L. Paja̧k, A. B. Jarzȩbski, A. I. Lachowski and J. J. Malinowski, Langmuir, 1997, 13, 6310–6314.
22. E. J. A. Pope and J. D. Mackenzie, J. Non-Cryst. Solids, 1986, 87, 185–198.
23. M. A. Vincent and I. H. Hillier, Chem. Commun., 2005, 5902–5903.
24. J. A. Kalow and A. G. Doyle, J. Am. Chem. Soc., 2010, 132, 3268–3269.
25. D. Albanese, D. Landini and M. Penso, Synthesis, 1994, 1994, 34–36.
26. F. Ehrburger-Dolle, J. Dallamano, E. Elaloui and G. M. Pajonk, J. Non-Cryst. Solids, 1995, 186, 9–17.
27. Y. S. Avadhut, J. Weber, E. Hammarberg, C. Feldmann, I. Schellenberg, R. Pöttgen and J. S. A. der Günne, Chem. Mater., 2011, 23, 1526–1538.
28. C. Cossement, J. Darville, J.-M. Gilles, J. B. Nagy, C. Fernandez and J.-P. Amoureux, Magn. Reson. Chem., 1992, 30, 263–270.
29. T. Kawashima, H. Matsui and N. Tanabe, Thin Solid Films, 2003, 445, 241–244.
30. J. P. Correa Baena and A. G. Agrios, ACS Appl. Mater. Interfaces, 2014, 6, 19127–19134.
31. A. L. Patterson, Phys. Rev., 1939, 56, 978–982.
32. D. K. Schroder, Semiconductor Material and Device Characterization, Wiley-Interscience, New York, 2006.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra01015f

‡ JPCB and DAK contributed equally to this work.

---