HBT-based chemosensors for the detection of fluoride through deprotonation process: experimental and DFT studies

Abstract

When searching to develop fluoride chemosensors based on O–H⋯F, we discovered that an HBT-based fluorophore containing a hydroxyl group was easily synthesized and displayed excellent fluorescence properties. 4-(benzothiazol-2-yl)-phenol (L¹H) was found to facilitate the monitoring of fluoride and showed ratiometric fluorescence changes. It is worth noting that an aldehyde group in conjugation with the HBT-based fluorophore core at an adjacent position to the hydroxyl group (i.e. 5-(benzothiazol-2-yl)-2-hydroxybenzaldehyde, L²H) would elevate the sensitivity towards fluoride immensely. Spectroscopic studies indicated that L¹H and L²H interacted with a fluoride anion, which involved a two-step reaction: hydrogen bond formation and deprotonation. Deprotonation of the chemosensors by a fluoride anion enhanced the electron-donating ability of the phenolic O⁻ to the HBT core acceptor and facilitated an intramolecular charge transfer process, resulting in a red shift in both UV-vis absorption and fluorescence spectra. The mechanism of L²H binding with fluoride was confirmed by ¹H NMR titration experiments and DFT computational calculations.

---

1. Introduction

Anions are widely present in the natural environment and organisms, since they play crucial roles in chemical and biological processes. As the smallest anion, fluoride has special physical and chemical properties. It plays vital roles in biological processes as well as industry. In living organisms, fluoride can be used in preventing tooth diseases and osteoporosis.¹ In the nuclear industry, as a result of the principle that fluoride could react with uranium, fluoride was used to separate radioactive and non-radioactive substances safely and easily.² However, the excessive utilization of fluoride has inevitably led to severe environmental pollution. Subsequently, fluoride accumulated in people's bodies via the food chain and eventually resulted in severe disease or death. Therefore, the detection of fluoride has attracted great attention.³

In the past few years, chemosensors for fluoride have been successfully developed, due to their properties of high sensitivity, excellent selectivity and low limits of detection. Many fluoride chemosensors were based on the interactions between fluoride and Lewis acids. A strategy for fluoride recognition employed the formation of a hydrogen bond. Fluoride can interact with thiourea,⁴ urea,⁵ amide,⁶ sulfonamide,⁷ imidazole,⁸ indole,⁹ pyrrole,¹⁰ and Schiff bases.^11,12 However, to the best of our knowledge, most of the recognition mechanisms mentioned above were based on N–H⋯F, but not O–H⋯F. Recently, O–H⋯F has been considered of great concern.¹³ Cao et al. designed new fluoride sensors that were based on the reason that O–H has higher electronegativity and acidity than N–H.^13b However, both the receptors underwent an “OFF–ON–OFF” process, viz., the fluorescence intensity would be quenched when the amount of F⁻ reached more than 6 equiv., and this self-quenching, caused by F⁻ concentration, would affect the detection results and greatly limit its application. In our ongoing interest in developing fluorescent probes for fluoride based on O–H⋯F, we recently endeavored to explore the types of fluorophores, containing a hydroxyl group. We discovered that an HBT-based fluorophore was easily synthesized, cheaply produced and displayed excellent fluorescence properties. As expected, the compound, 4-(benzothiazol-2-yl)-phenol (L¹H), containing a hydroxyl group was found to facilitate the monitoring of fluoride and showed ratiometric fluorescence changes. Although L¹H displayed high sensitivity toward fluoride, it was necessary to add amounts of F⁻ (less than 10 equiv.) until the fluorescence reached a saturated state. It is worth noting that an aldehyde group in conjugation with the HBT-based fluorophore core at an adjacent position to the hydroxyl group (i.e. 5-(benzothiazol-2-yl)-2-hydroxybenzaldehyde, L²H) would elevate the sensitivity towards fluoride immensely. The interaction between L²H and fluoride may involve a two-step reaction: hydrogen bond formation and deprotonation. The structures and proposed binding mechanisms are shown in Scheme 1. Mechanistic studies were confirmed by spectroscopic studies, ¹H NMR titration experiments and DFT computational calculations.

Scheme 1 Structures and proposed binding mechanisms of L¹H and L²H with F⁻.

2. Experimental

2.1 Materials and instruments

All starting materials and reagents (analytical grade), including tetrabutylammonium salts of halides and sodium salts of other anions, were supplied by commercial suppliers and used without further purification. THF was used at HPLC grade. ¹H NMR titration experiments were recorded on a JNM-ECS 400 instrument. Fluorescence spectra were acquired on a Thermo Scientific Lumina fluorescence spectrometer. UV-vis absorption spectra were produced on an Evolution 220 UV-visible spectrometer. The fluorescence quantum yield was measured by the absolute value method on a FL sp920.

2.2 Synthesis and characterization of L¹H and L²H

The synthetic methods of L¹H and L²H were mentioned in our previous literature.¹⁴

2.3 Absorption and fluorescence measurements

Stock solutions of L¹H and L²H (1 mM) were prepared in THF while anion solutions (10 mM) were prepared in triple-deionized water. The solutions of anions (1 mM) were diluted by putting the concentrated 10 mM solutions into THF. All absorption and fluorescence measurements were carried out in a quartz cell (1 cm × 1 cm × 3.5 cm). UV-vis absorption titrations and fluorescence titrations were performed by adding F⁻ to 2 mL THF with L¹H or L²H diluted in it in advance. The fluorescence spectra were measured by excitation at 330 nm. The slit widths in both excitation and emission were 5 nm. The detection limit was calculated as the ratio 3σ/k,^14b,15 where σ stands for the standard deviation of 10 measurements of the emission spectrum intensity of L¹H or L²H and k is the slope obtained by plotting the fluorescence intensity of L¹H or L²H upon addition of various amounts of F⁻. The dissociation constant K_d was determined by the equation K_d = [X]ⁿ × (F_max − F)/(F − F_min), where F_min, F_max and F represent the emission intensities in the absence of fluoride, presence of fluoride in saturation and any varying concentration of fluoride, [X] is the concentration of fluoride, and n is the number of fluoride ions bound per L¹H or L²H molecule.

2.4 DFT computational studies

The optimized geometries for the structures in Scheme 1 were obtained by employing quantum mechanical calculations which were carried out at the B3LYP/6-311+G(d,p) level of theory using the Gaussian 09 suite of programs.¹⁶ The IEFPCM solvation model was used for the calculations and THF was used as the solvent, which was in agreement with the experimental conditions.

3. Results and discussion

3.1 UV-vis absorption and fluorescence spectrum studies of L¹H

In order to search for new fluoride chemosensors, we used L¹H (an HBT-based compound containing a hydroxyl group) as a novel prototype. It was found to facilitate monitoring of the fluoride anion by UV-vis absorption and fluorescence emission spectra. In Fig. 1a, L¹H has an absorption peak at 320 nm and a shoulder peak at 309 nm in THF. When fluoride was added from 0 to 95 μM, the peaks at 309 nm and 320 nm decreased gradually and a new peak at 376 nm appeared and increased simultaneously with an isosbestic point at 330 nm. The changes in absorbance at 320 nm and 376 nm are presented in Fig. 1a (inset), respectively. Moreover, the emission peak of L¹H at 368 nm decreased gradually and a new peak at 426 nm increased simultaneously in fluorescence spectra of L¹H (Fig. 1b). The detection limit of L¹H was calculated to be 6.70 × 10⁻⁸ M by fluorescence titration and the dissociation constant was found to be 8.28 × 10⁻¹⁰ M² (Fig. S1 and S2†). In terms of its sensitivity for fluoride, L¹H is superior to previous probes.¹³ However, L¹H needs more fluoride to achieve equilibration, which inevitably leads to low sensitivity. Hence, it is necessary to explore methods of modifying the structure of L¹H to improve its sensing ability.

Fig. 1 UV-vis absorption spectra (a) and fluorescence emission spectra (b) of L¹H (10 μM) upon addition of various amounts of F⁻ (0–95 μM) in THF. Insets: (a) changes in absorbance at 320 nm and 376 nm, respectively; (b) changes in fluorescence intensity at 368 nm and 426 nm, respectively.

3.2 UV-vis absorption and fluorescence spectrum studies of L²H

To greatly improve the detection properties, L²H was obtained by conjugation of an aldehyde group to the HBT-based fluorophore core at an adjacent position to the hydroxyl group. Remarkably, the sensitivity and detection limit of L²H towards fluoride anion were elevated after the modification. As shown in Fig. 2, UV-vis absorption titrations were performed in THF at room temperature. There was only one absorption band, at 311 nm, in the situation where 2 mL THF only contained 10 μM L²H. Upon addition of fluoride from 0 to 100 μM, the absorption band at 311 nm gradually decreased, while two new bands at 364 nm and 409 nm continuously increased and were accompanied with an isosbestic point at 330 nm. The two new peaks were simultaneously formed, mainly caused by the reason that the deprotonated form of L²H existed as two tautomeric forms (enol and keto) after L²H interacted with fluoride anion. The absorbance changes at 311 nm, 364 nm and 409 nm, respectively, are shown in Fig. 2 (inset). The absorbance remained unchanged when the concentration of fluoride reached 20 μM, implying that L²H could react with 2 equiv. of fluoride.

Fig. 2 UV-vis absorption spectra of L²H (10 μM) upon addition of various amounts of F⁻ (0–100 μM) in THF. Inset: changes in absorbance at 311 nm, 364 nm and 409 nm.

Using the isosbestic point at 330 nm as the excitation wavelength, fluorescence titration spectra of L²H versus various amounts of fluoride are shown in Fig. 3. Along with the addition of fluoride, the fluorescence intensity at 480 nm increased remarkably and ultimately became saturated if fluoride was added up to 2 equiv. of L²H. The ratio of the fluorescence intensity at 480 nm could reach 102-fold in THF in the case where L²H/F⁻ was 1 : 2 (Fig. 3, inset). The fluorescence quantum yield (Φ) of L²H was 20.56%, while Φ of L²H with the addition of 2 equiv. of F⁻ was 52.23%. To further determine the ratio of L²H to fluoride as we predicted in Scheme 1, a Job's plot was obtained in conditions where different amounts of L²H were combined with relevant different levels of F⁻ in THF at 480 nm. The total concentration of L²H and F⁻ was 10 μM. As illustrated in Fig. 4, L²H could react with 2 equiv. of fluoride. The detection limit of L²H was calculated to be 4.23 × 10⁻⁸ M by fluorescence titration and the dissociation constant of L²H was found to be 7.07 × 10⁻¹¹ M² (Fig. S3 and S4†). The presence of the aldehyde group enhanced the properties of L²H towards fluoride anion, due to the aldehyde group forming a hydrogen bond with the hydroxyl group of the HBT core, which weakened the interaction between the oxygen and hydrogen of the hydroxyl group, so the hydrogen atom would be lost more easily. Focusing on the selectivity of L²H, many anions were tested for comparison. In Fig. 5, upon treatment of L²H with fluoride anion, a dramatic increase in fluorescence intensity was observed. In contrast, the addition of I⁻, Ac⁻ and PO₄³⁻ only led to slight fluorescence changes in L²H; other representative anions did not cause any significant fluorescence response. What is more, the fluorescence intensity upon subsequent addition of fluoride exhibited a disproportionate increase. The ratiometric changes and bathochromic shift of L¹H and L²H after binding with fluoride probably originated from the processes of hydrogen bond formation and deprotonation, which were attributed to the basicity of fluoride anion and its strong ability to form hydrogen bonds and stabilized [HF₂]⁻.^3d,10a The occurrence of this process further enhanced the electron-donating ability of the phenolic O⁻ to the HBT core acceptor, resulting in an intramolecular charge transfer process. To verify this process, L²H binding with fluoride was confirmed by ¹H NMR titration experiments and DFT computational calculations.

Fig. 3 Fluorescence emission spectra of L²H (10 μM) upon addition of various amounts of F⁻ (0–50 μM) in THF. Excitation was at 330 nm. Inset: changes in fluorescence intensity at 480 nm.

Fig. 4 Job's plot of different amounts of L²H combined with relevant different levels of F⁻ in THF at 480 nm. The total concentration of L²H and F⁻ was 10 μM.

Fig. 5 Fluorescence intensity changes of L²H (10 μM) upon addition of various anions. The black bar represents the fluorescence intensity of L²H; the blue bars represent the fluorescence intensities of L²H in response to different anions (50 μM): (1) F⁻, (2) Cl⁻, (3) Br⁻, (4) I⁻, (5) S²⁻, (6) HS⁻, (7) PO₄³⁻, (8) BrO₃⁻, (9) H₂PO₄⁻, (10) HCO₃⁻, (11) SO₃²⁻, (12) HSO₄⁻, (13) S₂O₃²⁻, (14) HPO₄²⁻, (15) CO₃²⁻, (16) Ac⁻ and (17) HSO₃⁻; the red bars represent the fluorescence intensity that occurs upon subsequent addition of 50 μM F⁻ to the above solutions.

3.3 Studies on the mechanism of L²H versus fluoride

In order to confirm the proposed process, ¹H NMR titrations were performed in CDCl₃–THF mixed solvent. As shown in Fig. 6, significant spectral changes were observed with the gradual addition of certain concentrations of F⁻. It was noticeable that the phenol OH proton signal (10.70 ppm) disappeared as soon as fluoride anion was added to the mixed solution of L²H. With increasing concentrations of fluoride, the signals of H (2, 3, 4, 5, 6, 7) protons were gradually shifted upfield while the other H (1, 8) protons were shifted downfield. Also a signal of [HF₂]⁻ was observed at low field and is shown in Fig. S5.† However, the signals of protons were very weak when the amount of fluoride reached 1.5 equiv., so the titration experiments were not completed as in the initial scheme, designed to reach 2 equiv. Even so, the results suggest that the interaction between L²H and fluoride anion underwent a two-step reaction: hydrogen bond formation and deprotonation. The final deprotonation of L²H by fluoride anion enhanced the electron-donating ability of the phenolic O⁻ to the HBT core acceptor and facilitated an intramolecular charge transfer process.

Fig. 6 Partial ¹H NMR spectra of probe L²H in the presence of different amounts of F⁻ in CDCl₃–THF (v/v, 2/3) mixed solvent.

3.4 DFT computational studies

To better understand the spectral behavior of L²H before and after binding with F⁻, theoretical calculations were performed. The results of these calculations are shown in Fig. 7 and 8. L²H intrinsically displayed two fluorescence emission peaks at 355 nm and 460 nm (Fig. S6†), which was due to L²H being present in tautomeric forms, enol and keto. After binding with F⁻, the O–H⋯F hydrogen bond was formed, which resulted from the weak interaction between hydrogen and fluoride. The optimized geometric structures of the four forms are shown in Fig. 7. In Fig. 8, in the process of L²H interacting with fluoride, the molecular orbital diagrams of both HOMOs and LUMOs are represented in four forms. The energy gap between the HOMOs and LUMOs in these four forms becomes smaller after interaction with fluoride. These changes are in good agreement with the red shift in the absorption observed upon treatment of L²H with fluoride.

Fig. 7 Optimized geometric structures of L²H (enol, a), L²H (keto, b), L²H–F⁻ (c) and (L²)⁻ (d).

Fig. 8 Frontier molecular orbitals (HOMO, LUMO) for (a) L²H (enol); (b) L²H (keto); (c) L²H–F⁻; and (d) (L²)⁻, and the HOMO–LUMO energy gaps obtained by DFT calculation.

4. Conclusions

In conclusion, we have developed two HBT-based chemosensors (L¹H, L²H), containing a hydroxyl group for the detection of fluoride anion. Spectroscopic studies indicated that an aldehyde group in conjugation with the L¹H core at an adjacent position to the hydroxyl group would elevate the sensitivity towards fluoride immensely. The studies suggested that the process of L¹H and L²H sensing fluoride anion involved a two-step reaction: hydrogen bond formation and deprotonation. The final deprotonation of the chemosensors by fluoride anion enhanced the electron-donating ability of the phenolic O⁻ to the HBT core acceptor and facilitated an intramolecular charge transfer process, resulting in a red shift in both UV-vis absorption and fluorescence spectra. The mechanism of L²H binding with fluoride was confirmed by ¹H NMR titration experiments and DFT computational calculations.

Acknowledgements

We are grateful for the project (51474118, 21402077) supported by the National Natural Science Foundation of China, the Fundamental Research Funds for the Central Universities (lzujbky-2014-178), and Gansu Provincial Science and Technology Support Program (1104FKCP120).

References

1. M. Kleerekoper, Endocrinol. Metab. Clin. North Am., 1998, 27, 441–452.
2. (a) M. Gavrilescu, L. V. Pavel and I. Cretescu, J. Hazard. Mater., 2009, 163, 475–510; (b) G. H. John, I. May, D. Collison and M. Helliwell, Polyhedron, 2004, 23, 3097–3103; (c) J. Seneda, F. Figueiredo, A. Abrao, F. Carvalho and E. Frajndlich, J. Alloys Compd., 2001, 323, 838–841.
3. (a) Y. Zhou, J. F. Zhang and J. Yoon, Chem. Rev., 2014, 114, 5511–5571; (b) L. S. Natrajan, Coord. Chem. Rev., 2012, 256, 1583–1603; (c) K. Kikuchi, Chem. Soc. Rev., 2010, 39, 2048–2053; (d) T. Gunnlaugsson, M. Glynn, G. M. Tocci, P. E. Kruger and F. M. Pfeffer, Coord. Chem. Rev., 2006, 250, 3094–3117.
4. (a) E. J. Jun, K. M. K. Swamy, H. Bang, S. J. Kim and J. Yoon, Tetrahedron Lett., 2006, 47, 3103–3106; (b) S. Devaraj, D. Saravanakumar and M. Kandaswamy, Sens. Actuators, B, 2009, 136, 13–19; (c) M. Vazquez, L. Fabbrizzi, A. Taglietti, R. M. Pedrido, A. M. Gonzalez-Noya and M. R. Bermejo, Angew. Chem., Int. Ed., 2004, 43, 1962–1965; (d) W. Lu, M. Zhang, K. Liu, B. Fan, Z. Xia and L. Jiang, Sens. Actuators, B, 2011, 160, 1005–1010; (e) V. Thiagarajan, P. Ramamurthy, D. Thirumalai and V. T. Ramakrishnan, Org. Lett., 2005, 7, 657–660; (f) L. E. Santos-Figueroa, M. E. Moragues, M. M. Raposo, R. M. Batista, S. P. Costa, R. C. Ferreira, F. Sancenon, R. Martinez-Manez, J. V. Ros-Lis and J. Soto, Org. Biomol. Chem., 2012, 10, 7418–7428.
5. (a) M. Boiocchi, L. Del Boca, D. E. Gomez, L. Fabbrizzi, M. Licchelli and E. Monzani, J. Am. Chem. Soc., 2004, 126, 16507–16514; (b) S. K. Kim and J. Yoon, Chem. Commun., 2002, 770–771; (c) E. J. Cho, J. W. Moon, S. W. Ko, J. Y. Lee, S. K. Kim, J. Yoon and K. C. Nam, J. Am. Chem. Soc., 2003, 125, 12376–12377; (d) Y. Wu, X. Peng, J. Fan, S. Gao, M. Tian, J. Zhao and S. Sun, J. Org. Chem., 2007, 72, 62–70.
6. (a) S. Wang, W. Shen, Y. Feng and H. Tian, Chem. Commun., 2006, 1497–1499; (b) Y. Li, L. Cao and H. Tian, J. Org. Chem., 2006, 71, 8279–8282; (c) B. Ke, W. Chen, N. Ni, Y. Cheng, C. Dai, H. Dinh and B. Wang, Chem. Commun., 2013, 49, 2494–2496; (d) B. Liu and H. Tian, J. Mater. Chem., 2005, 15, 2681–2686; (e) H. C. Hung, Y. Y. Chang, L. Luo, C. H. Hung, E. W. Diau and W. S. Chung, Photochem. Photobiol. Sci., 2014, 13, 370–379.
7. (a) S. V. Bhosale, S. V. Bhosale, M. B. Kalyankar and S. J. Langford, Org. Lett., 2009, 11, 5418–5421; (b) H. Lu, W. Xu, D. Zhang, C. Chen and D. Zhu, Org. Lett., 2005, 7, 4629–4632; (c) H. G. Im, H. Y. Kim, M. G. Choi and S. K. Chang, Org. Biomol. Chem., 2013, 11, 2966–2971; (d) L. Fu, F. F. Tian, L. Lai, Y. Liu, P. D. Harvey and F. L. Jiang, Sens. Actuators, B, 2014, 193, 701–707.
8. (a) H. N. Lee, N. J. Singh, S. K. Kim, J. Y. Kwon, Y. Y. Kim, K. S. Kim and J. Yoon, Tetrahedron Lett., 2007, 48, 169–172; (b) J. Yoon, S. K. Kim, N. J. Singh and K. S. Kim, Chem. Soc. Rev., 2006, 35, 355–360.
9. (a) X. He, S. Hu, K. Liu, Y. Guo, J. Xu and S. Shao, Org. Lett., 2006, 8, 333–336; (b) G. Pereira, E. M. S. Castanheira, P. M. T. Ferreira and M. J. R. P. Queiroz, Eur. J. Org. Chem., 2010, 2010, 464–475.
10. (a) A. K. Mahapatra, R. Maji, K. Maiti, S. S. Adhikari, C. Das Mukhopadhyay and D. Mandal, Analyst, 2014, 139, 309–317; (b) S. Rivadehi, E. F. Reid, C. F. Hogan, S. V. Bhosale and S. J. Langford, Org. Biomol. Chem., 2012, 10, 705–709.
11. (a) K. Liu, X. Zhao, Q. Liu, J. Huo, H. Fu and Y. Wang, J. Photochem. Photobiol., B, 2014, 138, 75–79; (b) K. K. Upadhyay, R. K. Mishra, V. Kumar and P. K. R. Chowdhury, Talanta, 2010, 82, 312–318; (c) X. Bao, J. Yu and Y. Zhou, Sens. Actuators, B, 2009, 140, 467–472.
12. (a) A. B. Sariguney, A. O. Saf and A. Coskun, Spectrochim. Acta, Part A, 2014, 128, 575–582; (b) B. Sui, B. Kim, Y. Zhang, A. Frazer and K. D. Belfield, ACS Appl. Mater. Interfaces, 2013, 5, 2920–2923; (c) P. Zhao, J. Jiang, B. Leng and H. Tian, Macromol. Rapid Commun., 2009, 30, 1715–1718; (d) J. S. Chen, P. W. Zhou, G. Y. Li, T. S. Chu and G. Z. He, J. Phys. Chem. B, 2013, 117, 5212–5221; (e) J. S. Chen, P. W. Zhou, S. Q. Yang, A. P. Fu and T. S. Chu, Phys. Chem. Chem. Phys., 2013, 15, 16183–16189; (f) J. S. Chen, R. Z. Liu, Y. Yang and T. S. Chu, Theor. Chem. Acc., 2014, 133, 1411; (g) J. S. Chen, P. W. Zhou, L. Zhao and T. S. Chu, RSC Adv., 2014, 4, 254.
13. (a) J. T. Ma, Z. Li, Y. Q. Zong, Y. Men and G. W. Xing, Tetrahedron Lett., 2013, 54, 1348–1351; (b) Q. Y. Cao, M. Li, L. Zhou and Z. W. Wang, RSC Adv., 2014, 4, 4041–4046; (c) A. K. Mahapatra, K. Maiti, P. Sahoo and P. K. Nandi, J. Lumin., 2013, 143, 349–354.
14. (a) S. D. Liu, L. W. Zhang and X. Liu, New J. Chem., 2013, 37, 821–826; (b) S. Liu, L. Zhang, W. Zan, X. Yao, Y. Yang and X. Liu, Sens. Actuators, B, 2014, 192, 386–392.
15. (a) L. Zhang, X. Cui, J. Sun, Y. Wang, W. Li and J. Fang, Bioorg. Med. Chem. Lett., 2013, 23, 3511–3514; (b) L. Zhang, D. Duan, X. Cui, J. Sun and J. Fang, Tetrahedron, 2013, 69, 15–21.
16. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery Jr, J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, N. J. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox in Gaussian 09, Revision, C. 01, Vol., Gaussian, Inc., Wallingford CT, 2009.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra13532f

---