Monolayer molecular probes for detection of trace amounts of cyanide anions

Abstract

In this work, monolayer sensor molecules are firstly developed for the detection of trace amounts of cyanide ions in aqueous solution, which was inspired by the fact that in dye-sensitized solar cells, dye molecules adsorb onto the semiconductor surface by forming a monolayer. Compound P1 is a well-known p-type dye molecule in dye-sensitized solar cells, with a carboxyl group on the triphenylamine used as the anchoring point. Thus it could be adsorbed onto a metal oxide (such as TiO₂ and NiO) surface to form a monolayer M-P1. M-P1 detects cyanide through a nucleophilic addition reaction between the negatively charged CN⁻ and the dicyanovinyl group, resulting in a weaker intramolecular charge transfer from the donor to acceptor in P1. The detection behavior could be easily observed from a visual color change of the metal oxide films, which was also in agreement with their UV-vis absorption spectral changes. Since the adsorbed sensor molecules are in very small amounts (at the nM level), M-P1 could be applied to detect trace amounts of CN⁻ in aqueous solution (pH = 7.4) with high sensitivity, selectivity and anti-interference ability. The detection limit of the monolayer sensor is determined to be 2.99 nM, which is far below the WHO cyanide standard for drinking water (1.9 μM). Additionally, it is much lower than that obtained by dissolving P1 and the analytes in organic solution. The results demonstrate that the monolayer sensor has high sensitivity and selectivity, and is very efficient for detecting trace amounts of analytes in dilute solution, and the detection could be easily observed by the naked eye. Moreover, the idea of a monolayer sensor excludes the requirement of an organic environment for reaction-based sensors to function, and provides new possibilities for constructing probes in the future.

---

Introduction

The detection of cyanide anions has aroused wide interest in recent years, due to their fatal effects on living organisms and the environment.^1,2 Cyanide-containing substances are usually found in industrial processes, such as gold mining, metallurgy, electroplating, and the synthesis of fibers and resins.^3–6 Therefore, the development of convenient and efficient detection techniques for tracing small amounts of cyanide ions has gained considerable attention in recent years. Among the various techniques, molecular probes based on absorption or fluorescence changes are more attractive due to their high sensitivity, ease of operation, and rapid real-time monitoring properties.^7–11 Especially, the design of new cyanide-selective chemosensors with significant fluorescence and colorimetric changes is more feasible and applicable. In this regard, various mechanisms for the detection of cyanide have been used, for example, excited-state intramolecular proton transfer (ESIPT), hydrogen-bonding interactions, forming a complex with metal ions and boron derivatives, a metal-cyanide displacement approach and nucleophilic addition reactions.^12–17 Reaction-based chemosensors for CN⁻ have been paid much attention due to their high selectivity over other anions and the fact that they can effectively reduce interference from hydrogen bonding and the acidity of the media.^10,11,18,19 The detection limit of these sensors of cyanide is usually at the μM level, and very few reports have detection limits at the nM level.^20–22 Besides, organic solvents are often required for the reaction based sensor to work.^10,11 Thus, the development of a quick, facile, and chemodosimeter based sensor, which works in aqueous media with a low detection limit, is still a challenge and is in urgent need.

Probes with dicyanovinyl groups have been exploited for cyanide detection. The dicyanovinyl group works as a selective cyanide-reactive unit through a nucleophilic addition reaction, which could break the main conjugation and interrupt the electronic structure and optical properties of the sensor molecule, and realize specific selectivity and high sensitivity toward cyanide anions.^23,24 For example, a novel salicylideneaniline-based fluorescent sensor was developed, and was demonstrated to fluorescently sense CN⁻ with specific selectivity and high sensitivity. The sensor showed a large blue shift in the absorption spectra and fluorescence quenching in response to CN⁻.¹² A dicyanovinyl-functionalized phenothiazine was designed as a near-infrared “on–off” fluorescent chemodosimeter for CN⁻.⁹ It exhibited a rapid colorimetric and quenchable NIR fluorescence response toward CN⁻. Considering the fact that the dicyanovinyl group is usually incorporated as a typical acceptor unit for constructing D–π–A type dye molecules in DSSCs and OPVs, we tried to develop the typical p-type dye molecule 4-(bis-{4-[5-(2,2-dicyanovinyl)thiophene-2-yl]phenyl}amino)benzoic acid, known as P1, for the detection of CN⁻.^25–27 We expect that the reaction of CN⁻ with the dicyanovinyl group would break the conjugation structure of the acceptor, and decrease its electron withdrawing ability, thus the charge transfer process will be blocked, leading to prominent photophysical changes of P1.

It is well-known that in DSSCs, the dye molecules are adsorbed onto a metal oxide (such as NiO, TiO₂, ZnO, etc.) surface through anchoring groups (usually a carboxylic acid, phosphoric acid, etc.). Thereby, the dye molecules are attached onto the semiconductor surface by chemical adsorption. As has been reported in many literature publications, the dye molecules adsorb onto the TiO₂ surface by covalent bonding to form a monolayer.^28–30 Inspired by the monolayer formation, we came up with the idea of exploiting a monolayer dye for detection. We expect that a single molecular layer of P1 could efficiently react with cyanide anions, meanwhile, the quite small amount of adsorbed P1 on the metal oxide films (usually 10⁻⁷ to 10⁻⁸ mol cm⁻² for DSSCs) may help to enhance the sensitivity.^31–33 To further investigate the sensing mechanism, the sensitivity of P1 for cyanide ions in solution phase is also investigated. It turned out that P1 shows a high selectivity toward CN⁻ in pure organic solvent such as THF, and the reaction mechanism between P1 and CN⁻ was also studied using NMR titration experiments. Herein, in this work, we report the preparation of monolayer sensor molecules on metal oxide films, and their application as efficient colorimetric sensors for cyanide anions in aqueous solution (Scheme 1). This method opens a new path for colorimetric sensing with high sensitivity and selectivity, which requires very few sensor molecules and is able to detect trace amounts of analytes in aqueous solution.

Scheme 1 The mechanism of the monolayer sensor M-P1 for detecting cyanide anions.

Experimental

Materials and instruments

All starting materials were purchased from commercial suppliers (Sigma-Aldrich and Energy Chemical) and used without further purification. Tetrahydrofuran (THF) for the synthesis was freshly distilled over a Na–K alloy under argon atmosphere prior to use. The ¹H NMR spectra were recorded using a BRUKER AVANCE III 600 MHz NMR Instrument in CDCl₃, using tetramethylsilane as an internal reference. UV-vis absorption spectra of the films with adsorbed P1 molecules were recorded using a spectrophotometer (UV-2450, Shimadzu).

Preparation of the metal oxide films and the P1 monolayer

P1 was synthesized according to literature procedures.²⁵ The metal oxide film was prepared by “doctor-blading” the precursor solution of NiO or the TiO₂ paste onto a glass substrate (the procedure is similar to that for DSSCs), and then sintering at 450 °C for 30 min to obtain the single-layer metal oxide films.³⁴ The NiO or TiO₂ film was immersed in a 0.1 mM P1 solution (EtOH) at room temperature for 1 min, followed by rinsing with CHCl₃ and drying under air. The monolayer of P1 adsorbed onto the NiO or TiO₂ film was then ready for use. It should be noted that the adsorption time would not affect the monolayer formation, but only affect the amount of loaded dye molecules.²⁸ Usually for DSSCs, the semiconductor film is soaked for 16–20 h in the dye solution (0.3 mM in general cases), to make sure that the maximum amount of dye is adsorbed. The dye molecules adsorb onto the NiO or TiO₂ nanoparticle surface, and the coverage is dependent on the dye amount adsorbed. As a result, the longer the soaking time, the deeper the color of the NiO or TiO₂ films, indicating more dye molecules have adsorbed (Fig. S4†). So we choose 1 min to balance the dye loading amount and the color of the films. With less than 1 min, the film color will be too pale, while more than 1 min will result in a larger dye loading amount and it could be envisaged that the reaction time needed for M-P1 with the analytes would be increased.

Results and discussion

Response of the monolayer to CN⁻

As P1 is applied as a p-type dye in NiO-based photoelectrodes, NiO was chosen as the template for P1 to adsorb onto upon the first consideration. The NiO film on a piece of glass was prepared, and then immersed into a 0.1 M solution of P1. After quite a short time (about 1 min), the film was taken out from the P1 solution and rinsed with chloroform to wash off the physically adsorbed dye. Then the monolayer of P1 (M-P1) was ready for use. In the experiments, PBS solution (pH = 7.4) was used throughout the measurements. Under acidic conditions, the NiO layer will be destroyed, and the following detection would not be conducted. While for an alkaline environment, it is widely known that the dye molecules will be desorbed from the NiO surface.^27,32 Therefore, a neutral environment is required for the detection to proceed.

The UV-vis absorption spectra of the NiO film before and after adsorption of P1 were measured. As shown in Fig. 1, the NiO film shows a characteristic sharp absorption band at around 335 nm, and a very weak absorption at around 440 nm, making the NiO film colorless and semitransparent. After P1 is adsorbed onto the NiO film, a broad band from 400–600 nm in addition to the absorption band of NiO appears, attributed to an intramolecular charge transfer (ICT) transition from the triphenylamine donor to the dicyanovinyl acceptor in P1. The absorption band of P1 on the NiO film is broader and slightly red shifted compared to that measured in THF, since the dye molecules tend to aggregate on the semiconductor surface. As a result, after P1 adsorption, the colour of the NiO film turns to dark red, featuring the colour of compound P1 in the solid state. Unfortunately, we were unable to acquire emission spectra for M-P1, since in P1 the D–A interactions are very strong and thus its fluorescence emission is weak, and after being adsorbed onto the NiO film, the very few dye molecules (∼0.125 nmol cm⁻²) in the film state make the fluorescence emission unable to be observed.

Fig. 1 UV-vis absorption spectra of the blank NiO film and the P1 monolayer (M-P1) formed on NiO. The inset shows photographs of the NiO film colour before and after adsorption of P1.

According to the studies on DSSCs, P1 can form a monolayer on a NiO film, and the adsorbed dye amount is quite small (usually 10⁻⁸ mol cm⁻² for NiO-based DSSCs). Meanwhile, considering the specific reactivity of dicyanovinyl with CN⁻, we expect that this monolayer sensor molecule may bring about positive effects on the sensing behaviour. We next performed UV-vis measurements of M-P1 after treatment with various anions. The anions were dissolved in a PBS solution containing 40% EtOH with a concentration at the nM level. The M-P1 on the NiO film was immersed into the anion solution for a short while (∼15 min, the time was decided from the amount of dye molecules loaded on the NiO film as shown in Fig. S4†), and then taken out. After drying under air, the absorption spectra were recorded.

From Fig. 2, as the concentration of CN⁻ increases from 10 nM to 60 nM, the charge transfer band of M-P1 gradually decreases, and the red colour of the M-P1 on the NiO film fades to almost colourless. The phenomenon observed is consistent to that observed in THF solution. The absorbance of M-P1 at around 510 nm was plotted against the CN⁻ concentration, which showed that the absorption intensity decreases with increased CN⁻, and reaches a plateau even at a quite low concentration of CN⁻ (about 50 nM) (inset of Fig. 2). Obviously, the absorption intensity at 510 nm varied almost linearly with the concentration of CN⁻ in the range of 0–40 nM, exhibiting a coefficient of R² = 0.97437. The detection limit for CN⁻ was calculated using the formula 3σ/k, where σ is the standard deviation of the blank film, and k is the slope of the calibration curve obtained from the inset of Fig. 2. The detection limit was determined to be 2.99 nM, which is much lower than that obtained with P1 dissolved in THF. This result indicates that the P1 monolayer could operate well and the detection limit is far below the WHO cyanide standard for drinking water (1.9 μM).³⁶ Since 50 nM CN⁻ could completely react with the P1 on the NiO film, and quantitatively, one mole of P1 reacts with 2 equiv. of CN⁻, the dye loading amount of P1 could be derived from the concentration of CN⁻. As a result, the amount of P1 adsorbed onto the NiO film was calculated to be 0.125 nmol. Therefore, the quantity of the molecular sensors and analytes required in this system is a very small amount, and could be considered as a trace amount.

Fig. 2 UV-vis absorption changes of M-P1 upon addition of CN⁻ with an increased concentration from 10 nM to 60 nM. The inset shows plots of the maximum absorption changes (510 nm) against the cyanide concentration.

Sensing mechanism

To confirm the detection mechanism, the response of P1 to CN⁻ in THF and THF–H₂O mixtures was also studied. The absorption spectrum of P1 was measured in THF at a concentration of 5 μM at room temperature. As shown in Fig. 3a, two absorption bands could be observed. The one at 350 nm was ascribed to the π–π* transition, while the other one at almost 500 nm corresponds to intramolecular charge transfer (ICT) from the triphenylamine donor to the dicyanovinyl acceptor. After the addition of CN⁻, the absorption band of P1 at the longer wavelength disappears and a new band at around 400 nm appears. The changes in the absorption spectra result in a significant colour change of the P1 solution. But when the solvent was changed to a THF–H₂O mixture (v/v/9 : 1), the addition of CN⁻ hardly caused any changes to the absorption of P1 (Fig. 3b), even with the addition of a large amount of CN⁻ (50 equiv.). It was after more than two hours that the colour of the P1 solution faded, indicating that in aqueous solution, it takes a longer time and a larger amount of the guest anions for the reaction to happen than in organic solvent. The sensing mechanism involves a nucleophilic addition reaction between the dicyanovinyl group and cyanide, which blocks the intermolecular charge transfer (ICT) process from the triphenylamine donor to the dicyanovinyl acceptor. As a result, the CT band of P1 gradually decreases with the increased CN⁻ concentration. We thought that the existence of water may lead to hydration of the CN⁻, and thus will reduce its reactivity. But when the sensors are adsorbed onto the metal oxide surface, the sensor molecules are at a nM level in this condition, and they are able to react with the cyanide anions in a short time to result in significant colour change, even with the existence of water.

Fig. 3 (a) UV-vis absorption of P1 (5 μM) before and after addition of 8.0 equiv. of CN⁻ measured in THF. (b) UV-vis absorption of P1 (5 μM) before and after addition of 8.0 equiv. of CN⁻ measured in THF–H₂O.

An absorption titration curve of P1 with CN⁻ in THF solution was also obtained. As shown in Fig. S1,† upon addition of CN⁻, the absorption band of P1 (5 μM) at the longer wavelength (483 nm) gradually decreased accompanied by increase of the absorption band at 346 nm during titration. When the amount of CN⁻ reached 8.0 equiv. (40 μM), the titration came to an end. The emission spectra of P1 indicate a quenching profile upon CN⁻ addition (Fig. S2†). The emission band at 590 nm gradually decreases with the increased CN⁻ concentration, and the near infrared (NIR) emission is quenched. We note that upon CN⁻ addition, both the maximum absorption band and the emission band of P1 decrease along with a gradual red shift. We used tetrabutylammonium cyanide as the cyanide anion source, which is weakly alkaline. The carboxyl group on P1 will be partly neutralized under the alkaline conditions, and its electron-accepting ability will thus be reduced. Actually in the original form of P1, the electron donating ability of the donor part (triphenylamine) has been decreased due to the existence of the carboxyl group, whereas, under the weak alkaline conditions, the donating ability of triphenylamine is increased to result in a bathochromic shift of the maximum absorption and emission bands of P1. To better visualize the sensing process, we plotted the fluorescence intensity at the emission maximum (590 nm for P1) with respect to the concentration of CN⁻ (Fig. S3†). In the range of 0–16 μM for the cyanide concentration, the fluorescence intensity at 590 nm decreased almost linearly with a coefficient of R² = 0.99329. The detection limit for CN⁻ was calculated using the formula 3σ/k,³⁵ where σ is the standard deviation of the blank solution, and k is the slope of the calibration curve obtained from the inset of Fig. S2.† The detection limit for CN⁻ was calculated to be 19 nM. We note that the absorption changes for P1 are irregular with increased CN⁻ concentration, which is probably due to the interference of the carboxyl group in P1. However, the colorimetric changes are more intuitive and could be observed conveniently by the naked eye.

The detection mechanism, as has been proposed in many other reports, involves a nucleophilic substitution reaction between the negatively charged CN⁻ and the dicyanovinyl unit.^10–12 The attack of CN⁻ interrupts the intramolecular charge transfer process from triphenylamine to dicyanovinyl, leading to the CT band decreasing and a significant color change of P1. This mechanism could be further supported by ¹H NMR titration experiments. As shown in Fig. S4,† the addition of 1.0 equiv. of cyanide results in a decreased signal for the vinylic protons (H_a) at 8.15 ppm, and a new signal at 4.63 ppm (H_a′) appears. A further addition of 2.0 equiv. of CN⁻ leads to the disappearance of the vinylic proton signal (H_a), and meanwhile an increased signal for H_a′ at 4.63 ppm, corresponding to the reaction product of CN⁻ with the dicyanovinyl group. Meanwhile, the aromatic protons on the thiophene rings (H_b and H_c) exhibit an upfield shift as a result of a sharp decrease in the electron-withdrawing effect of the dicyanovinyl groups.

Selectivity and competition experiments

To be qualified as an efficient sensor for CN⁻, a high selectivity and anti-interference ability toward the analyte over other competitive anions are essential. The absorption properties of M-P1 in response to different anions (including CN⁻, F⁻, Cl⁻, Br⁻, I⁻, S²⁻, HPO₄²⁻, SO₄²⁻, CO₃²⁻ and OAc⁻) were tested by immersing M-P1 into solutions of the different anions (10⁻⁴ M) in PBS (pH 7.4) containing 40% EtOH. As shown in Fig. 4b, only CN⁻ could cause decreased absorption bands for the M-P1 film, while all the other competitive anions, some of which often show strong interference in CN⁻ detection, did not bring about any distinct changes to the absorption spectra of M-P1. The competition experiments also indicated that the absorbance of M-P1 with CN⁻ is not disturbed by the addition of excess competitive anions (Fig. 5a). Meanwhile, no significant changes in the visual color were observed in parallel experiments with the other anions (Fig. 4a). These results further confirm that only treatment with CN⁻ could result in an apparent color change of M-P1. Overall, the monolayer P1 exhibits extraordinary selectivity and an anti-interference ability toward CN⁻ over high concentrations of the other investigated anions and can therefore be utilized as an excellent colorimetric probe for detection of trace amounts of cyanide ions.

Fig. 4 (a) M-P1 after immersion into various anion solutions. (b) UV-vis absorption changes of M-P1 in response to various anions.

Fig. 5 (a) Photographs of M-P1 and its color changes when soaked in a mixed anion solution in the presence and absence of CN⁻. (b) The effects of using TiO₂ films with different nanoparticle sizes on CN⁻ detection.

The effect of different metal oxide films

In order to verify the versatility of this method, single-layer TiO₂ films with 20 nm, 200 nm and 400 nm particle sizes were prepared. P1 was adsorbed onto the TiO₂ films with different sizes and then rinsed several times to obtain a monolayer. The TiO₂ based M-P1 was then immersed into a CN⁻ solution (PBS : EtOH v/v 3 : 2) for 1 min, and the colour changes are presented in Fig. 5b. Apparently, in all the situations, the colour of M-P1 was bleached quickly upon treatment with the CN⁻ solution. The colours of the different TiO₂ films are distinguished due to the different nanoparticle sizes, which also reflects the amount of dye molecules adsorbed. The results suggest that the sensitivity of M-P1 toward CN⁻ is irrespective of the metal oxide film as well as the size of the nanoparticles used. Therefore, it could be concluded that the formation of the monolayer sensor is a prerequisite for efficient detection of CN⁻, and the metal oxide film only works as a support for the sensor molecule to adsorb onto.

Similar to the investigation of the NiO based M-P1, the selectivity and anti-interference ability of the TiO₂ based monolayer probe were also studied (Fig. S6 and S7†). It turned out that in this case the P1 monolayer could also exhibit extraordinary selectivity and anti-interference abilities for CN⁻ over high concentrations of the other competitive anions. Unfortunately, the UV-vis spectra of M-P1 on the TiO₂ films were unable to be recorded due to interference from the absorption of TiO₂.

To confirm the significance of the monolayer, a contrast experiment was conducted. A TiO₂ film was prepared on a glass substrate and then was immersed into a P1 solution to form a monolayer of P1 on the TiO₂. Onto the same substrate, a solution of P1 in EtOH (0.3 mM) was dropped and dried under air to form a multilayer-film with considerable size compared to the TiO₂ film. As shown in Fig. S8,† the multilayer P1 film exhibited a much deeper colour than the monolayer film, suggesting that the amount of P1 molecules in the multilayer film was much more than that on the TiO₂. After the glass substrates were soaked in a CN⁻ solution, the colour of the monolayer P1 disappeared quickly (in 10 min) to obtain a white TiO₂ film; while the colour of the multilayer P1 hardly changed, even after deposition for two hours. This comparative test could well explain the advantage and necessity of the monolayer sensor.

Conclusions

In conclusion, a new method for preparing a monolayer sensor molecule M-P1 has been developed and applied for the detection of trace amounts of cyanide anions in aqueous solution. A metal oxide film (NiO or TiO₂) was prepared on a glass substrate, and immersed into the sensor solution to generate the monolayer M-P1. It was then dipped into various anion solutions. The detection behavior could be easily observed from a visual color change of the metal oxide films, as well as the UV-vis absorption spectra. Since the adsorbed sensor molecules are quite small in amount (nM levels) and could form a monolayer on the metal oxide surface, the probe M-P1 is able to detect trace amounts of CN⁻ in aqueous solution with high sensitivity, selectivity and an anti-interference ability. However, there are still some shortcomings of this new detection system. For example, a neutral pH environment is required for the monolayer sensor to work, and a small portion of ethanol is also needed during the detection. But this method still provides a new way to sense analytes in very dilute solutions. The TiO₂ film has potential to be prepared on a flexible substrate using a screen-printing method and thus could be manufactured on a large scale. The features of a low cost and facile detection make the monolayer sensor more applicable and practical.

Acknowledgements

This work is financially supported by the National Natural Science Foundation of China (No. 51203046) and the “Fundamental Research Funds for the Central Universities” (XDJK2014C145 and XDJK2014C052).

Notes and references

1. B. Vennesland, E. E. Comm, C. J. Knownles, J. Westly and F. Wissing, Cyanide in Biology, Academic Press, London, 1981.
2. D. S. Bendall and W. D. Bonner, Plant Physiol., 1971, 47, 236–245.
3. S. Vallejos, P. Estévez, F. C. García, F. Serna, J. L. Peña and J. M. García, Chem. Commun., 2010, 46, 7951–7953.
4. K. W. Kulig, Cyanide Toxicity, U.S. Department of Health and Human Services, Atlanta, GA, 1991.
5. L. Peng, M. Wang, G. Zhang, D. Zhang and D. Zhu, Org. Lett., 2009, 11, 1943–1946.
6. G. Qian, X. Li and Z. Y. Wang, J. Mater. Chem., 2009, 19, 522–530.
7. J. Isaad and F. Salaün, Sens. Actuators, B, 2011, 157, 26–33.
8. J. F. Xu, H. H. Chen, Y. Z. Chen, Z. J. Li, L. Z. Wu, C. H. Tung and Q. Z. Yang, Sens. Actuators, B, 2012, 168, 14–19.
9. L. Yang, X. Li, J. Yang, Y. Qu and J. Hua, ACS Appl. Mater. Interfaces, 2013, 5, 1317–1326.
10. Z. Xu, X. Chen, H. N. Kim and J. Yoon, Chem. Soc. Rev., 2010, 39, 127–137.
11. X. Cheng, Y. Zhou, J. Qin and Z. Li, ACS Appl. Mater. Interfaces, 2012, 4, 2133–2138.
12. W.-C. Lin, S.-K. Fang, J.-W. Hu, H.-Y. Tsai and K.-Y. Chen, Anal. Chem., 2014, 86, 4648–4652.
13. V. Bhalla, S. Pramanik and M. Kumar, Chem. Commun., 2013, 49, 895–897.
14. S. Saha, A. Ghosh, P. Mahato, S. Mishra, S. K. Mishra, E. Suresh, S. Das and A. Das, Org. Lett., 2010, 12, 3406–3409.
15. S. C. Wei, P. H. Hsu, Y. F. Lee, Y. W. Lin and C. C. Huang, ACS Appl. Mater. Interfaces, 2012, 4, 2652–2658.
16. S. Y. Chung, S. W. Nam, J. Lim, S. Park and J. Yoon, Chem. Commun., 2009, 2866–2868.
17. V. Bhalla, H. Singh and M. Kumar, Dalton Trans., 2012, 41, 11413–11418.
18. Q. Zou, X. Li, Q. Xu, H. Ågren, W. Zhao and Y. Qu, RSC Adv., 2014, 4, 59809–59816.
19. L. Yang, X. Li, Y. Qu, W. Qu, X. Zhang, Y. Hang, H. Ågren and J. Hua, Sens. Actuators, B, 2014, 203, 833–847.
20. R. Manivannan, A. Satheshkumar and K. P. Elango, Tetrahedron Lett., 2014, 55, 6281–6285.
21. S. Pramanik, V. Bhalla and M. Kumar, ACS Appl. Mater. Interfaces, 2014, 6, 5930–5939.
22. S. Goswami, A. Manna, S. Paul, A. K. Das, K. Aich and P. K. Nandi, Chem. Commun., 2013, 49, 2912–2914.
23. W. Zheng, X. Z. He, H. B. Chen, Y. Gao and H. M. Li, Spectrochim. Acta, Part A, 2014, 124, 97–101.
24. X. F. Wu, B. W. Xu, H. Tong and L. X. Wang, Macromolecules, 2011, 44, 4241–4248.
25. P. Qin, H. Zhu, T. Edvinsson, G. Boschloo, A. Hagfeldt and L. Sun, J. Am. Chem. Soc., 2008, 130, 8570–8571.
26. L.-Y. Lin, Y.-H. Chen, Z.-Y. Huang, H.-W. Lin, S.-H. Chou, F. Lin, C.-W. Chen, Y.-H. Liu and K.-T. Wong, J. Am. Chem. Soc., 2011, 133, 15822–15825.
27. L. Zhu, H. Yang, C. Zhong and C. Li, Chem.–Asian J., 2012, 7, 2791–2795.
28. L.-M. Liu, S.-C. Li, H. Cheng, U. Diebold and A. Selloni, J. Am. Chem. Soc., 2011, 133, 7816–7823.
29. L. E. Lee, M. A. Gomez, S. Elouatik and G. P. Demopoulos, Langmuir, 2010, 26, 9575–9583.
30. F. Angelis, S. Fantacci, A. Selloni, M. Grätzel and M. Nazeeruddin, Nano Lett., 2007, 7, 3189–3195.
31. Y. Hua, S. Chang, D. Huang, X. Zhou, X. Zhu, J. Zhao, T. Chen, W.-Y. Wong and W.-K. Wong, Chem. Mater., 2013, 25, 2146–2153.
32. L. Cai, H. N. Tsao, W. Zjang, L. Wang, Z. Xue, M. Grätzel and B. Liu, Adv. Energy Mater., 2012, 3, 200–205.
33. L. Zhu, H. Yang, C. Zhong and C. Li, Dyes and Pigments, 2014, 105, 97–104.
34. S. Sumikura, S. Mori, S. Schimizu, H. Usami and E. Suzuki, J. Photochem. Photobiol., A, 2008, 199, 1–7.
35. Q. Zou, X. Li, J. J. Zhang, J. Zhou, B. B. Sun and H. Tian, Chem. Commun., 2012, 48, 2095–2097.
36. Guidelines for Drinking-Water Quality, World Health Organization, Geneva, 1996.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra21989b

---