The influence of pH on the stability of antazoline: kinetic analysis

Abstract

Degradation of the drug antazoline was studied in aqueous solutions by means of pH-rate profiling (pH 0–7.4). The novel approach of Runge–Kutta numerical integration in combination with multi-parameter optimisation was applied to UV-Vis spectral data to determine a valid kinetic model and kinetic parameters of the degradation process. The overall degradation mechanism was found to be dependent on the environmental pH. In the pH range of 3.0–7.4, the formation of the antazoline hydrolysis product (N-(2-aminoethyl)-2-(N-benzylanilino)acetamide) through three different pathways (acidic, non-catalysed, and semi-alkaline hydrolysis) was observed. In highly acidic media (pH 0–2), the degradation mechanism was found to be more complex. Although the same primary degradation product formed, a colourful (dark blue/violet) intermediate was also observed and further investigated by HPLC/TOF-MS.

---

Introduction

Degradation of pharmaceuticals represents a serious problem related to the loss of the pharmacological activity and the possibility of adverse effects induced by the resultant degradation impurities.¹

The International Conference on Harmonisation (ICH) guidelines² require stress tests for assessing the stability of drug substances and drug products, as well as to describe the nature of the degradation products and the operating degradation mechanisms.³ These tests also reveal useful information that is essential for improving manufacturing processes, validating expiration dates and selecting proper packagings.⁴

Degradation of a particular active pharmaceutical compound can be caused by various factors (temperature, humidity, pH, etc.). During API manufacturing, alternating pH conditions is a widely accepted strategy for improving drug solubility.⁵ However, this can also lead to API degradation, typically through hydrolysis,⁶ and depending on the environmental pH, different degradation products may form.⁷ Therefore, pH-rate profiling is commonly used for investigating API degradation rate and mechanisms.⁸

Degradation of pharmaceuticals under various pH conditions can be investigated by using a number of analytical methods. Spectrophotometric techniques are widely used due to the ease of spectral data acquisition, handling and interpretation. In addition, these techniques are highly sensitive and thus suitable for studying chemical reactions in solutions.⁹ The advantages of spectrophotometric methods also include improved selectivity due to the simultaneous measurement and evaluation of the absorbance with the reaction time, as well as possibility to avoid interference from coloured and opaque samples.^10,11

Antazoline, also known as imidamine or phenazoline (C₁₇H₁₉N₃, N-(4,5-dihydro-1H-imidazol-2-ylmethyl)-N-(phenylmethyl)-aniline) (ANT, Fig. 1) is a histamine H₁ receptor antagonist used mainly in the treatment of nasal congestion and allergic conjunctivitis.^12,13

Fig. 1 Chemical structures of antazoline (ANT) and antazoline hydrolysis product (ABA).

However, due to the poor solubility of its free base, only antazoline salts (hydrochloride and phosphate) are currently used in pharmaceutical practice.¹⁴ The drug also exhibits significant antiarrhythmic properties, therefore, can be applied in the treatment of ventricular and supraventricular tachycardia.¹⁵ Furthermore, several studies have confirmed the overall positive, but less common effect of intravenously administered antazoline in the treatment of paroxysmal atrial tachycardia, atrial flutter, and atrial fibrillation.¹⁶

Only one impurity of antazoline (its hydrolysis product) has been defined in European Pharmacopoeia – N-(2-aminoethyl)-2-(N-benzylanilino)acetamide (ABA, Fig. 1).¹⁷ This compound has been characterised using different analytical methods – UV-Vis spectroscopy, TLC-MS, FT-IR,¹⁸ as well as NMR and HPLC-MS.¹⁹ To the best of our knowledge, the degradation process of antazoline has been studied only at pH 7.4 (ref. 18) and no investigation of degradation process kinetics has been performed.

The present research expanded the degradation kinetics studies of antazoline to wider pH range (pH 0–7.4). It was found that the previously reported degradation process occurs only over a limited pH range (3.0–7.4). In strongly acidic media, a more complex degradation route with the formation of a colourful intermediate was identified. The degradation pathway was further investigated and characterised using UV-Vis spectrophotometry and HPLC/TOF-MS analysis.

The Runge–Kutta numerical integration method in combination with least squares multi-parameter optimisation was used to determine the kinetic model and parameters of antazoline degradation at various pH conditions.

Experimental

Materials

Antazoline phosphate (purity >99%) was supplied by JSC Grindeks (Riga, Latvia) and used without further purification. All the reactants were of analytical or HPLC grade and were procured from commercial sources. Water was deionised at the laboratory to <0.1 μS using Adrona Crystal 5 water purification unit (Adrona, Co. Ltd, Latvia).

Synthesis of antazoline base (ANT)

Antazoline phosphate (1.0 g) was added to 200 mL beaker and mixed with deionised water (50 mL) and stirred for 15 minutes until homogenous solution was obtained. 0.03 M sodium hydroxide solution (100 mL) was then added to the solution. A white crystalline precipitate was collected by filtration, rinsed with deionised water and dried for 24 hours in an airtight container at ambient temperature. PXRD method was used to identify and characterise the purity of the product.

Synthesis of antazoline hydrolysis product (ABA)

The hydrolysis of antazoline was performed in the same manner as described by Ruckmick et al.¹⁹ Antazoline base (0.3 g) was refluxed for 60 min in 50 mL of 0.5 M sodium hydroxide. The hydrolysis product was isolated by cooling the alkaline reaction solution, followed by acidification with concentrated hydrochloric acid and extraction into chloroform. The chloroform extract was concentrated on a rotary evaporator down to several milliliters and dispersed onto a Petri dish prior to drying and scraping off the residue.

The degradation of antazoline base at semi-alkaline, acidic, and highly acidic conditions

Eight samples of antazoline base (0.0100 g each) were dissolved in 30 mL aliquots of phosphate buffer solution (pH 7.4), phosphate–citrate buffer solutions (pH 6.0, 5.0, 4.0, 3.0), and 0.010–1.0 M hydrochloric acid solutions (pH 2, 1, 0). All samples were maintained at 50.0 °C temperature. UV-Vis absorption spectra were recorded on a PerkinElmer Lambda 25 spectrophotometer (PerkinElmer Inc., USA) in the wavelength range between 200 and 700 nm (10 mm quartz cuvette) at 24 h intervals for a total of 5000 hours. The most acidic sample (pH = 0) was analysed by HPLC/TOF-MS upon preparation, as well as after 500, 1000, 2500, and 5000 hours.

Powder X-ray diffraction (PXRD) analysis

PXRD data were obtained using a Bruker AXS D8 Advance powder diffractometer (Bruker AXS GmbH, Germany) with Cu Kα radiation (λ = 1.5418 Å), 40 kV, 40 mA. The diffraction patterns were recorded at room temperature with a 0.02° step and at a scan speed of 0.2 s per step. DIFFRAC^plus EVA (ver. 3.2) software (Bruker AXS GmbH, Germany) was used for the analysis of recorded diffraction patterns.

High performance liquid chromatography – time of flight mass spectrometry (HPLC/TOF-MS)

An Agilent 1290 infinity series instrument (Agilent Technologies, USA) equipped with an Agilent 1290 infinity DAD detector (Agilent Technologies, USA) and Agilent 6320 TOF-MS mass-spectrometer (Agilent Technologies, USA) was used for the HPLC/TOF-MS analysis.

Chromatographic separation was performed using a Waters XTerra MS C18 column (2.1 × 150 mm, 3 μm) (Waters Corp., USA) placed in a column oven at 30 °C. The UV-Vis spectra were recorded between 250 and 600 nm. The chromatographic profiles were registered at 293, 330, 425, and 590 nm.

The analysis was performed in gradient mode, with the mobile phase (90 : 10, v/v) composed of solvent A (HPLC grade water + 0.1% formic acid) and solvent B (HPLC grade acetonitrile + 0.1% formic acid), and gradually changed over 10 min to 5 : 95, and then maintained for an additional 5 min. The injection volume was 1 μL and the flow rate was 0.2 mL min⁻¹.

The mass spectrometry conditions were as follows: positive ionisation mode, scan range from m/z 50 to 1600 was used with the gas temperature of 325 °C, nitrogen flow rate was 10 L min⁻¹, nebulizer pressure 45 psi, capillary voltage 3000 V and the applied fragmentor voltage was 130 V.

Runge–Kutta method combined with the multi-parameter least squares optimisation method

In case of multiple step degradation with complicated pathways, the use of common analytical equations is not adequate for the determination of kinetic parameters. Therefore, simplified numerical integration methods such as Runge–Kutta method are attractive alternatives for solving the behavior of such systems.^20,21

In this study, the degradation process was mainly investigated using UV-Vis absorption spectroscopy. A series of absorption spectra was collected and processed as a function of time. The results were used to extract analytical information: the number of species present, the concentration profiles of specific components, the process kinetic model and kinetic parameters. First, the theoretical absorption values were modelled as a sum of the pure component contributions to the system. The spectral profiles of the individual components at various wavelengths were expressed using the general Beer–Lambert law (eqn (1)):

where l is the light path (mm), C_i is the molar concentration (mol L⁻¹), and ε_i is the molar absorption coefficient (mol⁻¹ L mm⁻¹).

The Runge–Kutta method was applied to different kinetic models that were potentially suitable for describing the degradation process. As a result, molar concentration plots for each component were constructed. Various pathways including different reaction orders and rates were evaluated. Kinetic parameters, reaction rate constants (k_i) were characterized for each degradation step.

The optimal ε_i and k_i values were calculated using multi-parameter least squares optimisation by finding the minimum value of the goal function with MS Excel Solver software (eqn (2)):

The calculations were repeated until a global minimum value was achieved. The optimisation process was validated by assigning different input values for the optimised parameters (ε_i and k_i).

The most appropriate kinetic model was then determined by using root-mean-square deviation (RMSD) criteria (eqn (3)):

where n is the number of experimental points and z is the number of optimised parameters.

The obtained RMSD values for each kinetic model were compared, and the models with the lowest RMSD values were considered as the most probable. In the cases with similar RMSD values (difference of less than two times), the simplest kinetic model was chosen for describing the degradation process.

Results and discussion

Synthesis of antazoline base

The comparison of synthesis product and antazoline base PXRD patterns (Fig. 2) showed mutual match and no residue of antazoline phosphate was detected.

Fig. 2 The PXRD patterns of antazoline phosphate, antazoline synthesis product, and antazoline base.²²

Degradation studies in semi-alkaline and acidic media

The degradation of antazoline base under semi-alkaline and acidic conditions resulted in the formation of the previously reported hydrolysis product (ABA). Due to the limited solubility of antazoline in alkaline media, pH = 7.4 was found to be the upper limit for pH-rate profile measurements.

Qualitative analysis of the system was mainly based on previously conducted research at pH = 7.4, where simple conversion of ANT to ABA was proven using chromatographic and spectroscopic techniques.¹⁸ Since our recorded UV-Vis profiles were practically identical in the pH range of 3.0–7.4, the same simple conversion of ANT to ABA was assumed under the applied conditions.

Although the degradation process of antazoline was monitored for a total of 5000 hours, the conversion of antazoline into ABA was complete only at pH = 7.4 within the experimental time period. Fig. 3 shows overall UV-Vis spectral changes of the sample over time (a) and the corresponding concentration profiles obtained from the most suitable kinetic model (b). In this case, only part of the obtained UV-Vis data is presented for clarity.

Fig. 3 (a) UV-Vis absorption spectra of antazoline solution at pH 7.4 over time, and (b) the corresponding concentration plots obtained using the kinetic model comprising two components – ANT and ABA.

The comparison of the experimental UV-Vis spectra with those of the pure antazoline and the main degradation product (ABA) suggested that only two components were present in the solution. An isosbestic point (at 310 nm) was also observed throughout the degradation process, indicating a conversion of one species into another unique species or the existence of an equilibrium between the two species.²³

The RMSD values (Table 1) obtained for various applied kinetic models clearly favored a simple first order reaction. The concentration profiles of ANT and ABA for this particular kinetic model were expressed with the following equations (eqn (4) and (5)) and solved according to the Runge–Kutta method using numerical integration:

where t_R is the reaction time (h); k₁ is the reaction rate constant (h⁻¹); C_ANT and C_ABA are the molar concentrations (mol L⁻¹) of ANT and ABA, respectively.

Table 1 The RMSD values of potential kinetic models for antazoline degradation at pH 7.4

Kinetic model	k1	k−1	RMSD
	Reaction order
	1		0.0012
	2		0.014
	3		0.048
	1	1	0.0012
	1	2	0.016
	1	3	0.065
	2	1	0.044
	2	2	0.038
	2	3	0.065
	3	1	0.059
	3	2	0.079
	3	3	0.090

---

Although the determined RMSD values for the reversible first order and simple first order kinetic models were the same, the obtained rate constants revealed that the reversible first order kinetic model was also describing a simple first order reaction. This was due to the reversible reaction rate constant (k₋₁) being close to zero after optimisation.

Furthermore, the ε_i values optimised using the first order kinetic model were in good agreement with those calculated from the experimental UV-Vis absorption spectra of the pure components (see ESI (Table S1†)).

At moderately low pH conditions (3.0–6.0), the first order reaction kinetic model was also favoured for describing antazoline degradation.

First of all, for antazoline sample at pH = 6.0 the overall degradation rate dropped significantly (approximately 10 times compared to degradation at pH = 7.4). At pH = 5.0 the degradation rate reached the minimum value and then started to increase at lower pH values (3.0–4.0) (Table 2).

Table 2 The kinetic rates of antazoline hydrolysis at the pH range of 7.4–3.0

pH	k1, h^−1
7.4	1.85 × 10^−3
6.0	3.28 × 10^−4
5.0	1.69 × 10^−4
4.0	1.80 × 10^−4
3.0	2.24 × 10^−4

---

Thus, antazoline degradation process at the pH range of 7.4–3.0 was proposed to consist of three different reactions: semi-alkaline (pH 7.4–6.0) (Fig. 4b), non-catalysed (pH 5.0) (Fig. 4c), and acidic hydrolysis (4.0–3.0) (Fig. 4d) of the imidazoline ring in antazoline molecule. Considering the distinct basic properties of antazoline (pK_a = 10.10 (ref. 24) at 25 °C), the previously reported alkaline hydrolysis reaction²⁵ (Fig. 4a) did not occur under applied pH conditions.

Fig. 4 The proposed mechanisms of imidazoline ring degradation through (a) alkaline,²⁵ (b) semi-alkaline,²⁶ (c) non-catalysed²⁶ and (d) acidic²⁶ hydrolysis.

Process kinetic parameters were in good agreement with the proposed mechanisms, emphasizing the differences between each rate limiting step of hydrolysis. In semi-alkaline and acidic media, the controlling step is the nucleophilic attack on the protonated form by a hydroxide ion or water respectively. However, as the nucleophilicity of hydroxide ion and water is different, the overall process kinetic rate varies. Furthermore, the effect of pH is more evident in semi-alkaline media, as the hydroxide ion concentration directly affects the formation rate of the unstable intermediate with imidol structure, whereas in acidic hydrolysis the decrease of pH would only shift the equilibrium towards the formation of the protonated imidazoline form.

The non-catalysed hydrolysis with the reported limiting stage of C–N bond cleavage is regarded as an intermediate process between semi-alkaline and acidic hydrolysis, and consequently exhibits the lowest overall reaction rate.²⁶

Degradation studies in highly acidic media

In highly acidic media, a colourful (dark blue/violet) intermediate X was observed during the conversion of antazoline to the main degradation product ABA. Fig. 5 illustrates the antazoline degradation process under the most acidic conditions (pH = 0). The UV-Vis spectra are divided into two parts (Fig. 5a and b) to emphasize the formation and depletion of the intermediate X and only part of the obtained UV-Vis data is presented for the clarity.

Fig. 5 (a and b) UV-Vis absorption spectra of antazoline degradation reaction mixture at pH 0, (c) the corresponding concentration plot obtained from the applied kinetic model with four components – ANT, X, Y, and ABA; (d) the HPLC/TOF-MS (TIC) data.

At the very beginning of the degradation process a rapid decrease of the antazoline absorbance peak (298 nm) was observed. It could not be rationalised by the simple formation of an intermediate X even if kinetic models of higher order reactions were applied. Consequently, an additional hypothetical intermediate Y was proposed in order to yield a valid kinetic model. The resulting kinetic model, as determined by the obtained RMSD values (Table 3), consisted of one reversible reaction of antazoline being in equilibrium with the intermediate Y, followed by two consecutive first order reactions. To facilitate the processing and interpretation of the obtained data, the intermediates X and Y were assumed to be single components with distinct molecular structures. However, due to the overall system intricacy, X and Y might also consist of multiple components.

Table 3 The RMSD values of various first order kinetic models for antazoline degradation in highly acidic media

Kinetic model	RMSD	Kinetic model	RMSD
	0.18		0.0073
	0.0082		0.0077
	0.055		0.0080
	0.068		0.0072

---

Similar minimum RMSD values were obtained for first order kinetic models with various levels of complexity. The improvement of RMSD values achieved by elaborating the applicable kinetic model was insignificant. Additionally, when higher order kinetics were used, the overall RMSD values for all kinetic models increased by 2–10 times. Therefore, the simplest model with the lowest RMSD value was determined to be most suitable. Consequently, the degradation process was described with a kinetic model consists of one reversible first order reaction followed by two consecutive first order reactions. Concentration plots for the selected kinetic model were expressed with the following equations (eqn (6)–(9)) using a similar approach as previously described:

where t_R is the reaction time (h); k₁, k₋₁, k₂, k₃ are the reaction rate constants (h⁻¹); C_ANT, C_ABA, C_X and C_Y are the molar concentrations (mol L⁻¹) of the ANT, ABA and intermediates X and Y, respectively.

Comparable RMSD values were also obtained for the less acidic samples of antazoline solutions (pH 1 and 2); the same kinetic model was favoured for describing the degradation process. A comparison of the obtained kinetic parameters for antazoline degradation in highly acidic media (Table 4) revealed the substantial effect of the applied pH conditions. The conversion rates for both of the consecutive reactions (ANT → X → ABA) exhibited exponential dependence on pH. The kinetic parameters also showed that a decrease of pH in the medium promoted faster formation of intermediate X and main degradation product ABA.

Table 4 Kinetic rates of antazoline hydrolysis under acidic pH conditions

pH	k1, h^−1	k−1, h^−1	k2, h^−1	k3, h^−1
0	5.51 × 10^−3	9.15 × 10^−3	9.05 × 10^−4	6.19 × 10^−4
1	3.21 × 10^−3	1.54 × 10^−2	8.13 × 10^−5	1.23 × 10^−4
2	7.85 × 10^−4	8.79 × 10^−1	3.11 × 10^−5	2.41 × 10^−5

---

HPLC/TOF-MS investigation of antazoline degradation process in highly acidic media

In order to investigate antazoline degradation mechanism at highly acidic conditions, the most acidic antazoline sample (pH 0) was also analysed using HPLC/TOF-MS at different stages of degradation.

Both antazoline and the main hydrolysis product (ABA) were identified from chromatographic data by their mass spectral peaks (Fig. 6). The respective chromatographic peak intensity changes were in a good agreement with the selected kinetic model (Fig. 6, TIC). A slight peak shift was observed during the antazoline degradation process and was investigated further with EIC (extracted ion chromatography). Two different mass spectra profiles were separated by their most characteristic mass to charge ratio (m/z) peaks and were identified as antazoline and intermediate X (Fig. 6, EIC).

Fig. 6 HPLC/TOF-MS analysis of the most acidic antazoline sample (pH 0) (upper left side, TIC), EIC analysis of antazoline and intermediate X (upper right side, EIC) and the respective mass spectra for individual components (ANT, X, and ABA).

Chromatographic identification of the intermediate X was continued by analysing data obtained with DAD (diode array detector). In both EIC and DAD, the chromatographic peaks of antazoline and intermediate X had similar retention time differences between the components. Additionally, their respective chromatographic UV-Vis spectral profiles generated from DAD chromatographic data (Fig. 7) matched those observed in degradation studies using UV-Vis spectrophotometry.

Fig. 7 The extracted UV-Vis spectra profiles of the intermediate X (left side) and antazoline (right side) from the chromatographic data obtained with DAD.

Based on the experimental data obtained from chromatographic analysis, intermediate X and antazoline were proposed to be structurally similar. The characteristic deep blue/violet colour of intermediate X may be attributed to the formation of charge transfer complex, strong chromophore or conjugation in the molecule. However, complete structure identification could not be achieved within this study.

Conclusions

This work has presented the suitability of Runge–Kutta method in combination with multi-parameter optimisation for processing kinetic data obtained from experimental UV-Vis absorption spectra profiles. Using this approach, kinetic parameters describing the antazoline degradation process under various pH conditions were obtained.

It was found that within the pH range of 0–7.4, degradation of antazoline resulted in the formation of previously reported hydrolysis product, however, the reaction rate strictly depended on the environmental pH. Within the pH range of 3.0–7.4, three distinct pH intervals were distinguished and exhibiting different imidazoline ring hydrolysis mechanisms (acidic, non-catalyzed, and semi-alkaline) with distinct rate limiting steps for these degradation processes.

In highly acidic media (pH 0–2), however, the overall degradation process was found to be more complex due to the formation of highly colourful (dark blue/violet) intermediate which was later characterised by its individual mass spectra using HPLC/TOF-MS analysis. Nevertheless, complete explanation of the reaction mechanism requires more specific research of the degradation process in highly acidic media.

Acknowledgements

This work was supported by the European Regional Development Fund (Grant No. 2011/0014/2DP/2.1.1.1.0/10/APIA/VIAA/092).

References

1. S. Görög, Identification and Determination of Impurities in Drugs, Elsevier Inc., 2000.
2. ICH, Q1A(R2) – Stability Testing of New Drug Substances and Products, Geneva, 2003.
3. J. Swarbrick, N. Carolina, L. L. Augsburger, H. G. Brittain, A. J. Hickey and C. Hill, Pharmaceutical Stress Testing, Taylor & Francis Group, LLC, 2005.
4. K. M. Alsante, A. Ando, R. Brown, J. Ensing, T. D. Hatajik, W. Kong and Y. Tsuda, Adv. Drug Delivery Rev., 2007, 59, 29–37.
5. A. T. M. Serajuddin, Adv. Drug Delivery Rev., 2007, 59, 603–616.
6. K. C. Waterman, R. C. Adami, K. M. Alsante, A. S. Antipas, D. R. Arenson, R. Carrier, J. Hong, M. S. Landis, F. Lombardo, J. C. Shah, E. Shalaev, S. W. Smith and H. Wang, Pharm. Dev. Technol., 2002, 7, 113–146.
7. A. S. Kearney, L. F. Crawford, S. C. Mehta and G. W. Radebaugh, Pharm. Res., 1993, 10, 1461–1465.
8. G. Alibrandi, S. Coppolino, N. Micali and A. Villari, J. Pharm. Sci., 2001, 90, 270–274.
9. J. M. Calatayud, Encyclopedia of Analytical Science, 2005, pp. 373–383.
10. I. A. Darwish, M. A. Sultan and H. A. Al-Arfaj, Spectrochim. Acta, Part A, 2010, 75, 334–339.
11. I. A. Darwish, M. A. Sultan and H. A. Al-Arfaj, Int. J. Res. Pharm. Sci., 2010, 1, 43–50.
12. M. Parvez, Acta Crystallogr., Sect. C: Cryst. Struct. Commun., 1997, 53, 506–508.
13. M. Figus, P. Fogagnolo, S. Lazzeri, F. Capizzi, M. Romagnoli, A. Canovetti, M. Iester, A. Ferreras, L. Rossetti and M. Nardi, Eur. J. Ophthalmol., 2010, 20, 811–818.
14. A. Dravniece, A. Actiņš, K. Krūkle-Bērziņa and I. Sarceviča, Mol. Cryst. Liq. Cryst., 2015, 606, 154–164.
15. M. Andersen, P. F. Hansen and E. Sandøe, Acta Med. Scand., 1965, 177, 761–763.
16. E. W. Reynolds, W. M. Baird and M. E. Clifford, Am. J. Cardiol., 1964, 14, 513–521.
17. European Pharmacopoeia, Council of Europe, Strasbourg, 7th edn, 2010.
18. N. Alfredo, A. Spadaro and K. Annunziaia, Int. J. Pharm., 1990, 65, 137–140.
19. S. C. Ruckmick, D. F. Marsh and S. T. Duong, J. Pharm. Sci., 1995, 84, 502–507.
20. L. G. Arnaut, S. J. Formosinho and H. D. Burrows, Chemical Kinetics: From Molecular Structure to Chemical Reactivity, Elsevier Inc., 1st edn, 2007.
21. J. C. Butcher, Numerical methods for ordinary differential Equations, John Wiley & Sons, Inc., 2nd edn, 2008.
22. M. Pavez, CCDC 128589: Experimental Crystal Structure Determination, Cambridge Crystallographic Data Centre, 1997.
23. D. N. Sathyanarayana, Electronic Absorption Spectroscopy and Related Techniques, Universities Press Ltd, 2001.
24. A. C. Andrews and J. K. Romary, J. Chem. Soc., 1964, 405–408.
25. B. G. Harnsberger and J. L. Riebsomer, J. Heterocycl. Chem., 1964, 1, 188–192.
26. M. M. Watts, J. Am. Oil Chem. Soc., 1990, 67, 993–995.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra09043a

---