Green chemical incorporation of silicon into polyoxoanions of molybdenum: characterization, thermal kinetics study and their photocatalytic water splitting activity

Abstract

Cetylpyridinium silicomolybdate (CSM) nanorods were successfully synthesized by applying green chemistry principles using sodium molybdate and a structure directing cationic surfactant, cetyl pyridinium chloride (CPC) at room temperature. The composition and morphology of the nanorods were established by Fourier transform infrared spectroscopy (FT-IR), scanning electron microscopy (SEM), transmission electron microscopy (TEM), thermogravimetric analysis (TG) and inductively coupled plasma atomic emission spectroscopic (ICP-AES) techniques. The thermal decomposition kinetics of CSM nanorods were investigated by a non-isothermal thermogravimetric analyzer at various heating rates. The thermal decomposition of CSM occurred in two stages. The activation energies of the first and second stages of thermal decomposition for all heating rates have been estimated using the iso-conventional methods of Flynn–Wall–Ozawa (FWO) and Kissinger–Akahira–Sunose (KAS) and the results are found to be in good agreement with each other. The invariant kinetic parameter (IKP) method and master plot method were also used to evaluate the kinetic parameters and mechanism for the thermal decomposition of CSM. The photocatalytic water oxidation mechanism using the CSM catalyst in the presence of platinum (Pt) co-catalyst enhances the H₂ evolution and was found to be 1.946 mmol g⁻¹ h⁻¹.

---

Introduction

Interest in polyoxometalates (POMs) is increasing worldwide in contemporary solid-state chemistry due to their enormous variety of structures and electronic versatility.^1,2 The diversity of structures as well as the high number of elements that can make up the structure of a POM allows for a wide range of applications in fields such as catalysis, medicine, biology, material science and more recently in nanotechnology.^3,4 Polyoxomolybdate species (many of these are mixed-valence types) show an extreme variety of complicated structures and therefore great significance in structural details. In recent years, a variety of very large polyoxomolybdates have been synthesized and structurally characterized, for example the wheel-shaped anions of the type {Mo₁₅₄}, {Mo₁₇₆}^5,6 as well as the corresponding derivative {Mo₂₄₈},⁷ the spherical icosahedral capsule {Mo₁₃₂},⁸ and the hedgehog-shaped {Mo₃₆₈} type cluster species.^9,10 In terms of forming direct POM polymers linked by Mo–O–Mo units, it has been observed that in the synthesis of wheel and sphere-shaped nano sized molybdenum-oxide based clusters even under one-pot conditions, extensive linking via these types of units can occur. This means that the clusters primarily formed by self-assembly can become further linked in the same phase.¹¹ An example is that the paramagnetic Keplerate “necklaces” synthesized by a novel room temperature solid state reaction were characterized in the solid state as metal–oxide-based nano particles which is formed from controlled linking of {(Mo^VI)–Mo^VI}₁₂Fe₃₀ type Keplerate balls connected by inter-ball Mo–O–Mo bonds.

Due to their acidic and redox properties, POMs can act in synergy with various compounds, yielding new attractive catalytic systems.^2,3,12 However, their high solubility in water and polar solvents, low surface area and relatively low thermal stability precludes them as stable supports for heterogenization. Because of this, the preparation of POM-based materials is an active field of research and positively charged surfactants can replace the counter-ions of POMs, resulting in surfactant encapsulated clusters (SECs) with a core–shell structure and a well-defined composition.^13–16 This surface modification for POMs improves their stability in air and enhances their solubility in nonpolar solvents such as chloroform, toluene, and benzene. Moreover, synthetic parameters such as concentration of the reactants, pH of the reaction medium, organic template and temperature seem to influence the formation of a solid, its crystal structure as well as morphology including nanostructural features.^17,18 Recent research has shown that nanostructured vanadium and molybdenum oxides could be prepared by employing soft chemistry routes in the presence of long chain amines and surfactants.^19–21

However, a great deal of research has focused on the synthesis and characterization of POMs, while there have been few studies on the fundamental thermodynamic.^22,23 As far as we know, the thermodynamic properties of POMs, such as apparent activation energy (E), pre-exponential factor (A) and reaction mechanism functions of thermal decomposition process are important properties that reflect the structures and stabilities of compounds but were rarely reported. Currently non-isothermal decomposition data from TG is analyzed by two methods: model-fitting (Differential, Freeman–Carroll, Coats–Redfern) and model-free (Kissinger, Flynn–Wall–Ozawa (FWO), Kissinger–Akahira–Sunose (KAS)) methods.^24–29 Model-fitting methods consist in fitting different models to the data so that a model is chosen when it gives the best statistical fit as the model from which the kinetic parameters are calculated. Model-free (isoconversional) methods require several kinetic curves to perform the analysis and the calculations from several curves at different heating rates are performed on the same value of conversion, which allows to calculate the activation energy for each conversion point.

Historically, model-fitting methods were widely used for solid-state reaction because of their ability to directly determine the kinetic triplet (activation energy, frequency factor and reaction model) from a single TG measurement. It is well established that force-fitting non-isothermal data to different reaction models results in a widely varying Arrhenius parameters.³⁰ The kinetic analysis based on isoconversional method is frequently referred to as model-free because it is possible to obtain the apparent activation energy (E) as the function of conversion degree (α) and these methods are considered as the mostly reliable ones. According to the recommendations of International Confederation for Thermal Analysis and Calorimetry (ICTAC) Kinetics Committee,³¹ multi heating rate programs are recommended for computation of reliable kinetic parameters, while the methods that use a single heating program should be avoided.

Here, it is a report on a simple, cost-effective and environmentally benign synthesis of the water insoluble polyoxomolybdate containing silicon hetero atom from aqueous solutions at room temperature by applying the principles of green chemistry resulted in nanorods and its characterization by FT-IR, SEM, TEM, EDAX and ICP-AES techniques. The present work also aims to understand the thermal decomposition kinetics of resulting material using non-isothermal multi-heating thermogravimetric (TG) data in terms of model-fitting and model-free isoconversional methods as recommended by ICTAC Kinetics Committee. In this article, we have also tried to explore the photocatalytic water splitting activity of CSM catalyst.

Materials and methods

Synthesis of CSM nanorods

All chemicals are of reagent grade and were used without any further purification. In the present work, sodium molybdate as molybdenum source, CPC as structure directing template and double-distilled water as the medium of reactions were used. The synthesis of CSM was carried out by preparing two different solutions. Sodium silicate (0.5 mL) was slowly added to the aqueous solution of Na₂MoO₄·2H₂O (0.02 mol) with constant stirring and heated to boiling (solution I). Required amount of the surfactant was dissolved in 1 M solution of HCl with vigorous stirring (solution II). Solution I was slowly added to solution II at room temperature with vigorous stirring by keeping the mass ratio as CPC/Na₂MoO₄·2H₂O = 0.33, a yellow green precipitate was formed at pH 5, which was adjusted to pH 4 by adding HCl. The above mixture was then stirred continuously for 24 h.^32,33 The precipitate is then washed continuously with double-distilled water and then with absolute alcohol thoroughly to remove the excess of surfactant. The formed precipitate was then dried below 50 °C.

Photocatalytic reaction

A 70 mL quartz cell was used for performing the water splitting activity. The CSM catalyst (0.025 g) was dispersed in 80 v/v% water and 20 v/v% methanol by means of a magnetic stirrer. To the reaction mixture about 2 μL of chloroplatinic acid solution was added as a co-catalyst and methanol as sacrificial reagent. The visible light source was 400 W high pressure mercury lamp and the reaction was carried out for 4 h. The amount of H₂ evolved was determined using Agilent Technologies 7890A gas chromatography (GC) system and the evolving gas mixture was taken in a gas tight syringe and injected into the GC.

Kinetic methods

Iso-conversional methods. It is well known that the iso-conversional method easily gives an estimate of activation energy regardless of the reaction mechanism. Two kinds of iso-conversional methods are applied in this article.

FWO method is an integral method which is based on the measurement of the adequate temperature to certain values of the conversion α, for experiments effectuated to different rates of heating.³⁴ The equation corresponding to this method is:

where g(α) is integral reaction model, α is extent of conversion, E is activation energy, A is pre-exponential factor, T is Kelvin temperature, β is heating rate and R is gas constant.

The KAS method³⁵ is another integral isoconversional method and is based on the equation:

The relations given by eqn (1) and (2) involve all three members of kinetic triplet. From these relations the apparent activation energy can be evaluated using the linear representation of ln β, ln(β/T²) vs. 1/T with a given value of the conversion.

Model fitting methods. These methods obtain the kinetic parameters with a single heating rate. In this work, the method that has been employed is the integral Coats–Redfern (C–R) method.³⁶

Algebraic expressions for g(α) for the most frequently used mechanisms (see Table S1 in ESI†). For each theoretical kinetic model, g(α), and at each heating rate, from the slope and the intercept of plots ln[g(α)/T²] versus 1/T, the parameters ln A and E can be evaluated. For a given model and heating rate, the linear plot of the left hand side versus 1/T permitted the determination of E and A from the slope and the intercept, respectively.

Invariant kinetic parameters (IKP) method. Furthermore, the ‘true’ kinetic model can be obtained using the Invariant kinetic parameters (IKP) method^37–39 and is based on the study of the compensation effect. If the compensation effect between ln A and E (obtained from CR method) exists, then by plotting ln A versus E straight lines should be obtained for each heating rate, according to

ln A = α* + β*E (4)

The lines of the plot should intercept in a point that corresponds to the ‘true’ values of E and ln A for the ‘true’ kinetic model, which were called by Lesnikovich and Levchik^37–39 the invariant kinetic parameters, E_inv and A_inv. Due to the fact that certain variations of the experimental conditions determine regions of intersections, the intersection is only approximate. Therefore, in order to eliminate the influence of experimental conditions on the determinations of E_inv and A_inv they were determined from the slope and intercept of the so called super correlation relation.

α_v* = ln A_inv − β_v*E_inv (5)

The slopes and intercepts of the plot give the compensation parameters. The straight line, α_v* and β_v*, allows us to determine the IKP (E_inv and A_inv) from the slope and intercept.

It has been reported that the values of the invariant conversion function are proportional to their true values.^40,41 Therefore, the IKP method aims to determine the invariant parameters independent of the kinetic model; comparing the invariant parameters to those obtained using other methods also allows us to determine which kinetic model is better for describing the process.

Integral master plot method. Integral master-plot method^42–44 was used for the determination of the reaction model for the thermal decomposition of solids. Essentially the master plot method is based on the comparison of theoretical master plot, which are obtained for a wide range of ideal kinetic models, with the experimental master plot. The integral function of conversion in the solid state nonisothermal decomposition reactions is expressed aswhere u = E/RT. The temperature integral, has no analytical solution and can be expressed by an approximation. An approximate formula⁴⁵ of p(u) with high accuracy is used

p(u) = e^−u/[u(1.00198882u + 1.87391198)] (7)

From the integral kinetic equation the following equation can be obtained using a reference point at α = 0.5 and according to eqn (6), one gets

where u_0.5 = E/RT_0.5. The following equation is obtained by dividing eqn (6) by eqn (8)

Plotting g(α)/g(0.5) against α corresponds to the theoretical master plots of various kinetic functions (ESI, Table S1†). Using the predetermined value of E from isoconversional method, along with the temperature measured as a function of α, p(u) can be calculated according to the eqn (7). Then the experimental master plots of p(u)/p(u_0.5) against α can be obtained. Comparing the experimental master plots with theoretical ones, can conclude the kinetic model.

Results and discussion

Spectrochemical analysis

FT-IR spectrum technique provides information related to the maintenance of the Keggin structure after salt preparation. In the Keggin structure, the four types of oxygen provide four characteristic bands in the spectrum in the range 1100–550 cm⁻¹, called the fingerprint region.³ The exact position of these bands depends upon the degree of hydration and the type of counter ion present. In the present study Fig. 1 shows the FT-IR spectra of CSM which yielded four characteristic peaks related to a Keggin structure: ν_{(Mo–Oc–Mo)} = 789 cm⁻¹, related to asymmetric stretching of molybdenum with edge oxygens in Mo–O–Mo; ν_{(Mo–Ob–Mo)} = 864 cm⁻¹, attributed to the asymmetric stretching of corner oxygens in Mo–O–Mo; ν_(Mo=Od) = 949 cm⁻¹, indicative of the asymmetric stretching of the terminal oxygen; and ν_(Si–Oa) = 890 cm⁻¹, assigned to asymmetric stretching of oxygens with a central silicon atom.⁴⁶ In addition to these bands, there is also a band related to N–H extending vibrations of CPC moiety stretching observed at 1382, 1474, and 1484 cm⁻¹. Two absorption bands of 2926 and 2850 cm⁻¹ corresponded to aliphatic C–H vibrations of CPC. Thus the FT-IR analysis confirms the presence only of the specific bands of the Keggin unit in the synthesized CSM.

Fig. 1 FT-IR spectrum of CSM nanorods.

Fig. 2(a) and (b) shows the typical SEM and TEM images of the synthesized product. The particles were found to be nanorods of diameter ∼150–160 nm and length ∼0.5–1 μm.

Fig. 2 (a) SEM (b) TEM (c) EDAX images of CSM nanorods.

EDAX analysis (Fig. 2(c)) as well as ICP-AES analysis showed the presence of Mo, Si and O in the nanorods. ICP-AES analysis indicated the ratio of Si : Mo is 1 : 12. XRD pattern (Fig. 3) of the synthesized nanomaterial shows that the particles are below crystalline dimension and hence amorphous in nature.

Fig. 3 XRD pattern of CSM nanorods.

Thermogravimetric analysis

The TG-DTG curves of CSM nanorods at four heating rates 10, 15, 20 and 25 °C min⁻¹ in the temperature range of 35–700 °C are shown in Fig. 4(a) and (b) and it seems that TG curve exhibits mass losses in two stages. For 10 °C min⁻¹, the first stage of decomposition takes place in the temperature range of 42 °C to 334 °C with a mass loss of 13.6% and is due to the loss of 1.5 cetyl pyridinium group. The corresponding DTG peak is seen at 322.08 °C. The second stage of decomposition takes place in the temperature range of 334 °C to 698 °C with a mass loss of 31.8% and it is due to the loss of 3.5 cetyl pyridinium group, 1.5 water molecules and the collapse of Keggin-type structure. The final decomposition product is the mixture of MoO₃ and Si₂O₅. This is accompanied by DTG peak at 368.8 °C. There is a good agreement between the expected and observed mass losses for the loss of cetyl pyridinium cation and silicomolybdic anions.²³ As seen in Fig. 4(b), the DTG peaks were shifted to higher temperature with increase in heating rate. DTG peak of first stage observed at 322.08 °C for heating rate 10 °C min⁻¹ was shifted to 323.97 °C, 335.11 °C and 340.78 °C for 15, 20 and 25 °C min⁻¹, respectively. Similarly the DTG peak of second stage observed at 368.8 °C for heating rate 10 °C min⁻¹ was also displaced to 387.5 °C, 391.29 °C and 393.17 °C for 15, 20 and 25 °C min⁻¹, respectively. On the basis of TGA and EDAX, composition of the CSM nanorods was found to be around (C₂₁H₃₈N)₅H₃ [SiMo₁₂O₄₀]·xH₂0 (x < 1).

Fig. 4 (a) Thermal degradation (TG) and (b) derivative mass loss (DTG) curves at the different heating rates (β = 10, 15, 20 and 25 °C min⁻¹) for the thermal decomposition of CSM nanorods.

Kinetic analysis

Isoconversional methods. The results obtained from thermogravimetric analysis were elaborated according to model-free methods to calculate the kinetic parameters. In order to apply different kinetic methods on the thermal decomposition process of CSM, the dependence of fraction decomposed (α) on temperature (T) at different heating rates for stages I and II are plotted (Fig. 5). It is seen that all α–T curves have the same shapes. The sigmoid-shaped curves are shifted to higher temperatures with an increase of heating rates as reported earlier.⁴⁷

Fig. 5 Plot of fraction reacted (α), versus temperature (T) for stages I and II at four different heating rates using TG data.

The FWO plots of ln β versus 1/T for different values of conversion are shown in Fig. 6. KAS plots (ln(β/T²) versus 1/T) also give same pattern of curves and hence these are not included in the present paper. The apparent activation energies and frequency factors were obtained from the slope and intercept of regression lines and are given in Table 1. The calculated correlation coefficients, r, were higher for all cases. From Table 1, we can observe that values of E obtained by FWO method are in the ranges of 143.95–179.21 and 144.52–158.38 kJ mol⁻¹ for the first and second decomposition stages, respectively. And values of E obtained by KAS method are about 141.69–178.38 and 141.42–155.36 kJ mol⁻¹, respectively. The averages of E for two stages from FWO and KAS methods are 163.78 and 150.89, 162.33 and 147.80 kJ mol⁻¹, which are noted that the calculated values of the activation energy in the FWO method very well agree with KAS method.

Fig. 6 FWO plot for stages I and II in nitrogen atmosphere for the non-isothermal decomposition of CSM using TG data. obtained by FWO and KAS methods for stages I and II.

Table 1 Kinetic parameters for stages I and II for the non-isothermal decomposition of CSM by FWO and KAS methods

Degree of conversion (α)	FWO				KAS
	Stage I		Stage II		Stage I		Stage II
	E kJ^−1 mol^−1	ln A/S^−1	E kJ^−1 mol^−1	ln A/S^−1	E kJ^−1 mol^−1	ln A/S^−1	E kJ^−1 mol^−1	ln A/S^−1
0.1	143.95 ± 0.69	28.24 ± 1.17	144.52 ± 0.93	25.47 ± 1.45	141.69 ± 0.69	27.65 ± 1.18	141.42 ± 0.93	24.73 ± 1.45
0.2	151.82 ± 1.35	30.03 ± 2.28	145.78 ± 1.35	25.98 ± 2.10	149.89 ± 1.35	29.54 ± 2.29	142.66 ± 1.36	25.23 ± 2.11
0.3	155.15 ± 1.67	30.77 ± 2.80	146.43 ± 1.30	26.19 ± 2.01	153.34 ± 1.59	30.31 ± 2.67	143.26 ± 1.24	25.44 ± 1.92
0.4	162.23 ± 1.42	32.22 ± 2.38	149.03 ± 1.07	26.71 ± 1.64	160.74 ± 1.43	31.85 ± 2.39	145.93 ± 1.07	25.98 ± 1.64
0.5	166.14 ± 1.51	33.01 ± 2.51	150.07 ± 0.92	26.91 ± 1.41	164.81 ± 1.5	32.68 ± 2.52	146.95 ± 0.92	26.17 ± 1.40
0.6	167.45 ± 1.47	33.27 ± 2.44	152.26 ± 0.83	27.27 ± 1.25	166.15 ± 1.43	32.95 ± 2.36	149.16 ± 0.64	26.56 ± 1.32
0.7	173.25 ± 1.49	34.44 ± 2.45	154.94 ± 0.89	27.73 ± 1.33	172.19 ± 1.73	34.18 ± 2.84	151.91 ± 0.76	27.04 ± 1.13
0.8	174.80 ± 1.73	34.74 ± 2.83	156.58 ± 1.02	28.01 ± 1.51	173.79 ± 1.60	34.48 ± 2.62	153.55 ± 1.02	27.33 ± 1.51
0.9	179.21 ± 1.59	35.65 ± 2.61	158.38 ± 0.99	28.34 ± 1.47	178.38 ± 1.42	35.44 ± 2.35	155.36 ± 1.00	27.67 ± 1.47
Average	163.78 ± 1.44	32.49 ± 2.39	150.89 ± 1.03	26.96 ± 1.57	162.33 ± 1.42	32.12 ± 2.35	147.80 ± 0.99	26.24 ± 1.55

---

Fig. S1 in ESI† shows the dependence of the activation energy (E) on the degree of conversion (α). Because the values of activation energy worked out for both stages by the KAS and FWO methods were very close, we chose the average values of two as the value of activation energy used in the master-plot method.

Model fitting methods. The sets of ln A and E for stages I and II can be calculated at different heating rates for various kinetic functions (see ESI, Table S1†) by using eqn (3) and are given in Table 2. For a given model and heating rate, the linear plot of the left hand side of eqn (3) versus 1/T permitted the determination of E and A from the slope and the intercept, respectively. It can be observed that a wide variety of results depend on the applied mechanism. If the collected E values shown in Table 2 are compared with those determined previously using the isoconversional methods, the function that yields a better fit to the experimental results corresponds to Avrami mechanism, A₂ for stage I and A_3/2 for stage II.

Table 2 Integral parameters determined by using CR Method for stages I and II for the thermal decomposition of CSM nanorods

Heating rate
Kinetic model	β = 10 °C min^−1			β = 15 °C min^−1			β = 20 °C min^−1			β = 25 °C min^−1
	E kJ^−1 mol^−1	ln A/s^−1	−r	E kJ^−1 mol^−1	ln A/s^−1	−r	E kJ^−1 mol^−1	ln A/s^−1	−r	E kJ^−1 mol^−1	ln A/s^−1	−r
Stage I
P2	112.4 ± 9.1	14.6 ± 4.3	0.9776	120.7 ± 11.5	16.4 ± 5.3	0.9697	127.5 ± 11.1	17.8 ± 5.5	0.9745	130.2 ± 11.8	18.5 ± 5.9	0.9723
P3	71.7 ± 6.1	5.9 ± 3.2	0.9756	77.2 ± 7.7	7.4 ± 4.2	0.9671	81.6 ± 7.4	8.4 ± 4.1	0.9724	83.5 ± 7.9	8.9 ± 4.7	0.9700
P4	51.3 ± 4.6	1.6 ± 2.6	0.9732	55.4 ± 5.8	2.7 ± 3.5	0.9642	58.7 ± 5.6	3.6 ± 3.7	0.9700	60.1 ± 5.9	4 ± 4.1	0.9675
A2	160.5 ± 4.4	24.8 ± 2.6	0.9974	172.8 ± 7.2	27.5 ± 4.0	0.9939	181.9 ± 5.9	29.2 ± 3.8	0.9963	186.0 ± 6.8	30.1 ± 4.4	0.9953
A3	103.7 ± 2.9	12.9 ± 1.9	0.9972	111.9 ± 4.8	14.9 ± 3.1	0.9935	117.9 ± 4.0	16.1 ± 3.0	0.9961	120.6 ± 4.6	16.8 ± 3.5	0.9950
A4	75.3 ± 2.2	6.9 ± 1.4	0.9970	81.5 ± 3.6	8.4 ± 2.6	0.9931	85.9 ± 3.0	9.5 ± 2.6	0.9958	87.9 ± 3.4	10 ± 3.0	0.9947
A3/2	217.2 ± 5.9	36.6 ± 3.2	0.9974	233.8 ± 9.6	39.9 ± 4.8	0.9941	245.9 ± 7.9	42.2 ± 4.5	0.9964	251.4 ± 9.1	43.2 ± 5.1	0.9954
R1	234.7 ± 18.3	39.7 ± 6.8	0.9795	251.4 ± 23.0	43.0 ± 8.3	0.972	264.9 ± 22.2	45.5 ± 8.4	0.9763	270.6 ± 23.7	46.5 ± 8.9	0.9743
R2	278.0 ± 13.8	48.1 ± 5.6	0.9914	298.4 ± 19.0	52.1 ± 7.4	0.9861	314 ± 17.5	54.9 ± 7.2	0.9893	320.9 ± 19.1	56.2 ± 7.8	0.9878
Stage II
A2	117.5 ± 0.0	13.6 ± 2.4	0.9876	117.2 ± 7.7	13.6 ± 4.0	0.9853	117.2 ± 7.3	13.6 ± 4.2	0.9868	125.2 ± 7.3	15.2 ± 4.4	0.9882
A3	74.8 ± 0.0	5.3 ± 3.3	0.9863	74.5 ± 5.1	5.4 ± 3.2	0.9838	74.5 ± 4.9	5.5 ± 3.1	0.9854	79.8 ± 4.9	6.7 ± 3.5	0.9871
A4	53.4 ± 0.0	1.0 ± 3.8	0.9849	53.1 ± 3.9	1.2 ± 2.6	0.9821	53.1 ± 3.6	1.4 ± 2.8	0.9838	57.1 ± 3.7	2.3 ± 3.1	0.9858
A3/2	160.3 ± 0.0	21.7 ± 1.6	0.9882	159.9 ± 10.2	21.7 ± 4.8	0.9859	159.9 ± 9.7	21.6 ± 4.9	0.9874	170.7 ± 9.8	23.6 ± 5.1	0.9887
R1	171.1 ± 19.8	23.3 ± 6.9	0.9561	170.5 ± 20.4	23.2 ± 7.4	0.9531	170.8 ± 20.0	23.1 ± 7.5	0.9553	182.3 ± 20.8	25.2 ± 7.8	0.9573
R2	204.7 ± 0.0	29.2 ± 0.6	0.9754	204.2 ± 18.3	28.9 ± 6.9	0.9729	204.3 ± 17.7	28.9 ± 7.0	0.9748	217.9 ± 18.3	31.3 ± 7.3	0.9763
R3	217.6 ± 0.0	31.3 ± 0.3	0.9806	216.9 ± 17.3	31.1 ± 6.6	0.9783	217.2 ± 16.6	30.9 ± 6.7	0.9799	231.6 ± 17.1	33.5 ± 7.0	0.9814
F3/2	224.3 ± 0.0	32.3 ± 1.1	0.9829	223.7 ± 12.9	32.1 ± 5.5	0.9806	223.9 ± 11.8	31.9 ± 5.5	0.9822	238.7 ± 11.4	34.5 ± 5.6	0.9836

---

IKP method. Other method using the calculation of kinetic parameters is IKP. The application of the IKP method is based on the study of the compensation effect. If a compensation effect is observed, a linear relation defined by eqn (4) for each heating rate, β is obtained. It is shown by plotting ln A versus E that a compensation effect is observed for each heating rate (Fig. 7) for the first and second stages of thermal decomposition. The slopes and intercepts of the plot give the compensation parameters α* and β* for both the stages and the values are presented in Table 3.

Fig. 7 IKP plots obtained from the CR method at different heating rates (β = 10, 15, 20 and 25 °C min⁻¹) for stages I and II for the thermal decomposition of CSM nanorods.

Table 3 Values of integral compensation parameters calculated from data in Table 2 for stages I and II for the thermal decomposition of CSM nanorods

Heating rate/°C min^−1	Stage I			Stage II
	−α*/min^−1	β*/mol kJ^−1	r	−α*/min^−1	β*/mol kJ^−1	r
10	0.2060 ± 0.0016	8.642 ± 0.256	0.9998	0.1831 ± 0.0023	8.322 ± 0.425	0.9994
15	0.2039 ± 0.0015	8.181 ± 0.256	0.9998	0.1809 ± 0.0026	7.945 ± 0.425	0.9994
20	0.2016 ± 0.0014	7.858 ± 0.256	0.9998	0.1788 ± 0.0026	7.682 ± 0.425	0.9994
25	0.2004 ± 0.0014	7.624 ± 0.256	0.9998	0.1775 ± 0.0024	7.400 ± 0.425	0.9994

---

A plot α* versus β* is actually a straight line proving the existence of a super correlation relation, whose parameters allow evaluation of the invariant activation parameters (Fig. 8) for stages I and II. The obtained values of E_inv and ln A_inv are 176.60 kJ mol⁻¹ and 27.77 s⁻¹ for the stage I and 160.2 kJ mol⁻¹ and 21.01 s⁻¹ for the stage II are in agreement with the values of E and ln A estimated for models A₂ and A_3/2 for stages I and II respectively, presented in Table 2.

Fig. 8 Supercorrelation relationship for stages I and II for the thermal decomposition of CSM nanorods.

Integral master plot method. The reaction model for the thermal decomposition of CSM was confirmed by integral master-plot method. Using the predetermined value of E from isoconversional method, along with the temperature measured as a function of α, p(u) can be calculated according to the eqn (7). Then the experimental master plots of p(u)/p(u_0.5) against α under various heating rates for the first and second stages of thermal decomposition of CSM can be drawn. Both of the theoretical master plots of various kinetic functions and experimental master plots are shown in Fig. 9. The comparisons of the experimental and theoretical master curves show that the kinetic process of the thermal decomposition of CSM agree with the A₂ master curve for the first stage of decomposition and A_3/2 master curve for the second stage of decomposition. In the literature A₂ and A_3/2 models are known as Avrami–Erofeev mechanism which describes the decomposition through nucleation and nuclei growth. These observations are in agreement with the mechanism obtained by CR method in the previous section.

Fig. 9 Theoretical (lines) integral master plots of g(α)/g(0.5) versus α and the experimental master curves (symbols) for stages I and II for the thermal decomposition of CSM nanorods.

Photocatalytic activity of CSM

There are various challenging methods employed for solar energy conversion. Among them, the most attractive one was the photocatalytic decomposition of water into hydrogen and oxygen which is known to be a clean renewable source. Many of the photocatalysts were UV active, thus wasting the abundance of visible light irradiance in sunlight falling on earth's surface. Photocatalytic reactions on semiconductors are initiated by the absorption of photons with energy hν equal to, or greater than, the semiconductor band gap.

The presence of surface active sites on the photocatalyst is also an important parameter for the production of hydrogen. In the absence of surface active sites the photogenerated conduction band electrons can recombine with valence bond holes very quickly and release energy in the form of unproductive heat or photons. This is because the photo-excited electron and hole possess an electrostatic attraction between each other. Availability of surface active sites for the redox reactions by the photogenerated electrons and holes minimize the chances of recombination processes. The photocatalytic decomposition of water before recombination process can be achieved by introducing photocatalytic active sites in the nanostructured particles. This can be achieved by loading them with noble metals as co-catalysts like Pt, Au, Pd, Rh, Ni, Cu and Ag which are very effective for the enhancement of photocatalytic decomposition of water.^48–53 These activities greatly reduce the possibility of electron–hole recombination, resulting in efficient separation and stronger photocatalytic reactions in the CSM photocatalyst. The rate of photocatalytic hydrogen production can be further increased by sacrificial reagents like methanol, ethanol and also organic compounds, organics derived from renewable sources like biomass which may be profitably employed as hole scavengers. But, the mechanisms of the role of cocatalysts, sacrificial agents and photocatalytic water splitting have been under debate with many researchers. Kudo et al.⁵⁴ first studied the photocatalytic properties of many Ta-based catalysts, among them La-doped NiO/NaTaO₃ showed high efficiency. Zou et al. carried out direct water splitting on In–Ni–Ta oxides under visible light irradiation. The photocatalytic water splitting activity was found to be 3.3 times higher for doped TiO₂ than the bare TiO₂.

In present work, the CSM acts as catalyst in the presence of platinum co-catalyst. In the absence of platinum cocatalyst, CSM does not display any photocatalytic water splitting activity due to its large band gap in the UV region. There are reports suggesting that the band gap issue can be minimized with proper cocatalysts which serve as the reaction sites and catalyze the reactions. The role of Pt as co-catalyst is to reduce the band gap of CSM catalyst so that visible light can be harvested by the catalyst to act as an electron donor for H₂ evolution through the splitting of water.

The detrimental effect of Pt co-catalyst on the CSM catalyst is an interesting observation. The presence of Pt has definitely enhanced the H₂ evolution and was found to be 1.946 mmol g⁻¹ h⁻¹.

Conclusions

Synthesis under suitable conditions based on the requirements of green chemistry using cetylpyridinium chloride as template led to the formation of cetylpyridinium silicomolybdate nanorods. Non-isothermal thermogravimetric measurements at various heating rates enable calculation of kinetic parameters characterizing the decomposition process. The average values of activation energy of the thermal decomposition of cetylpyridinium silicomolybdate nanorods obtained by Flynn–Wall–Ozawa and Kissinger–Akahira–Sunose methods were found to be 163.1 and 140.4 kJ mol⁻¹ for the first and second stages, respectively, over the range of α = 0.1–0.9. For the first and second stages of decomposition, the invariant activation energy obtained by the invariant kinetic parameter method was in good agreement with the average value obtained by integral isoconversional methods in the conversion range of 0.10 to 0.90. The master plot method revealed that the first stage of thermal decomposition takes place through Avrami–Erofeev A₂ mechanism, g(α) = [−ln(1 − α)]^1/2 while the second stage of decomposition occurs through Avrami–Erofeev (A_3/2) mechanism, g(α) = [−ln(1 − α)]^2/3. In the present work, we have shown that the photocatalytic water splitting activity of cetylpyridinium silicomolybdate in the presence of platinum co-catalyst enhances the H₂ evolution.

Conflicts of interest

The authors declare no competing financial interest.

Acknowledgements

Bessy D'Cruz thanks the University of Kerala for the award of research fellowship and financial support. The authors are grateful to NIIST at Thiruvananthapuram, Government Arts College at, University College at Thiruvananthapuram and SAIF at Cochin for lending the analytical facilities.

Notes and references

1. M. T. Pope and A. Müller, Polyoxometalate Chemistry: From topology via Self-Assembly to Application, Kluwer, Dordrecht, The Netherlands, 2001.
2. M. T. Pope and A. Müller, Polyoxometalates: From Platonic Solids to Anti-Retroviral Activity, Kluwer Academic Publishers, Dordrecht, 1994.
3. (a) M. T. Pope, Heteropoly and Isopoly Oxometalates, Springer-Verlag, New York, 1983; (b) M. T. Pope and A. Müller, Angew. Chem., Int. Ed., 1991, 30, 34–48.
4. (a) X. Luo and C. Yang, Phys. Chem. Chem. Phys., 2011, 13, 7892–7902; (b) T. Hsu-Yao, K. P. Browne, N. Honesty and Y. J. Tong, Phys. Chem. Chem. Phys., 2011, 13, 7433–7438.
5. A. Müller and C. Serain, Acc. Chem. Res., 2000, 33, 2–10.
6. A. Müller and S. Roy, Coord. Chem. Rev., 2003, 245, 153–166.
7. A. Müller, S. Q. N. Shah, H. Bögge and M. Schmidtmann, Nature, 1999, 397, 48–50.
8. A. Müller, E. Krickemeyer, H. Bögge, M. Schmidtmann and F. Peters, Angew. Chem., 1998, 37, 3360–3363.
9. A. Müller, E. Beckmann, H. Bögge, M. Schmidtmann and A. Dress, Angew. Chem., 2002, 41, 1162–1167.
10. A. Müller, B. Botar, S. K. Das, H. Bögge, M. Schmidtmann and A. Merca, Polyhedron, 2004, 23, 2381–2385.
11. A. Muller, S. K. Das, M. O. Talismanova, H. Bogge, P. Kogerler, M. Schmidtmann, S. S. Talismanov, M. Luban and E. Krickeymeyer, Angew. Chem., Int. Ed., 2002, 41, 579–582.
12. N. Mizuno, K. Yamaguchi and K. Kamata, Coord. Chem. Rev., 2005, 249, 1944–1956.
13. D. G. Kurth, P. Lehmann, D. Volkmer, H. Cölfen, M. J. Koop, A. Müller and A. Du Chesne, Chem.–Eur. J., 2000, 6, 385–393.
14. D. Volkmer, A. Du Chesne, D. G. Kurth, H. Schnablegger, P. Lehmann, M. J. Koopanda and A. Müller, J. Am. Chem. Soc., 2000, 122, 1995–1998.
15. D. G. Kurth, P. Lehmann, D. Volkmer, A. Müller and D. Schwahn, J. Chem. Soc., Dalton Trans., 2000, 21, 3989–3998.
16. (a) T. Lunkenbein, M. Kamperman, Z. H. Li, C. Bojer, M. Drechsler, S. Forster, U. Wiesner, A. H. E. Muller and J. Breu, J. Am. Chem. Soc., 2012, 134, 12685–12692; (b) H. B. Wang, Y. Yan, B. Li, L. H. Bi and L. X. Wu, Chem.–Eur. J., 2011, 17, 4273–4282.
17. S. Mann, Angew. Chem., Int. Ed., 2000, 39, 3392–3406.
18. D. L. Long, E. Burkholder and L. Cronin, Chem. Soc. Rev., 2007, 36, 105–121.
19. M. E. Spahr, P. Bitterli, R. Nesper, M. Müller, F. Krumeich and H. U. Nissen, Angew. Chem., Int. Ed., 1998, 37, 1263–1265.
20. M. Niederberger, F. Krumeich, H. J. Muhr, M. Müller and R. Nesper, J. Mater. Chem., 2001, 11, 1941–1945.
21. G. R. Patzke, F. Krumeich and R. Nesper, Angew. Chem., Int. Ed., 2002, 41, 2446–2461.
22. V. Sasca, M. Stefanescu and A. Popa, J. Therm. Anal. Calorim., 2003, 72, 311–322.
23. (a) V. Sasca, O. Verdes, L. Avram, A. Popa, P. Barvinschi and M. Mracec, Rev. Roum. Chim., 2011, 56, 501–516; (b) B. D'cruz, J. Samuel, M. K. Sreedhar and L. George, New J. Chem., 2014, 38, 5436–5444.
24. (a) S. J. Pas, M. S. Dargusch and D. R. MacFarlane, Phys. Chem. Chem. Phys., 2011, 13, 12033–12040; (b) P. Simon, J. Therm. Anal. Calorim., 2004, 76, 123–132.
25. K. Chandran, M. Kamruddin, P. Muralidaran and V. Ganesan, J. Therm. Anal. Calorim., 2013, 112, 63–71.
26. C. G. Mothe and I. C. de Miranda, J. Therm. Anal. Calorim., 2013, 112, 497–507.
27. N. N. Nassar, A. Hassan, G. Luna and P. Pereira-Almao, J. Therm. Anal. Calorim., 2013, 114, 465–472.
28. A. K. Galwey, J. Therm. Anal. Calorim., 2012, 109, 1625–1635.
29. N. Sbirrazzuoli, L. Vincent, A. Mija and N. Guigo, Chemom. Intell. Lab. Syst., 2009, 96, 219–226.
30. S. Vyazovkin, J. Therm. Anal. Calorim., 1997, 49, 1493–1499.
31. S. Vyazovkin, A. K. Burnham, J. M. Criado, L. A. Perez-Maqueda, C. Popescu and N. Sbirrazzuoli, Thermochim. Acta, 2011, 520, 1–19.
32. B. D’cruz, J. Samuel and L. George, Thermochim. Acta, 2014, 596, 29–36.
33. J. Thomas and A. J. Ramanan, Inorg. Chim. Acta, 2011, 372, 243–247.
34. T. Ozawa, Bull. Chem. Soc. Jpn., 1965, 38, 1881–1882.
35. H. E. Kissinger, Anal. Chem., 1957, 29, 1702–1706.
36. A. W. Coats and J. P. Redfern, Nature, 1964, 201, 68–69.
37. A. I. Lesnikovich and S. V. Levchik, J. Therm. Anal. Calorim., 1983, 27, 89–94.
38. A. I. Lesnikovich and S. V. Levchik, J. Therm. Anal. Calorim., 1985, 30, 237–262.
39. S. V. Levchik, G. F. Levchik and A. L. Lesnikovich, Thermochim. Acta, 1985, 92, 157–169.
40. P. Budrugeac, Polym. Degrad. Stab., 2005, 89, 265–273.
41. P. Budrugeac, E. Segal, L. A. Pérez-Maqueda and J. M. Criado, Polym. Degrad. Stab., 2004, 84, 309–315.
42. C. D. Doyle, J. Appl. Polym. Sci., 1961, 5, 285–292.
43. S. Vyazovkin, Thermochim. Acta, 2000, 355, 155–163.
44. M. E. Brown, M. Macijewski and S. Vyazovkin, Thermochim. Acta, 2000, 355, 125–143.
45. T. Wanjun, L. Yuwen, Z. Hen, W. Zhiyong and W. Cunxin, J. Therm. Anal. Calorim., 2003, 74, 309–315.
46. C. Rocchiccioli-Deltcheff, M. Fournier, R. Franck and R. Thouvenot, Inorg. Chem., 1983, 22, 207–216.
47. J. A. Conesa, A. Marcilla, R. Font and J. A. Caballero, J. Anal. Appl. Pyrolysis, 1996, 36, 1–15.
48. K. Rajeshwar, R. McConnell and S. Licht, Solar Hydrogen Generation: Toward a Renewable Energy Future, Springer Science, New York, 2008.
49. A. Fujishima, X. Zhang and D. A. Tryk, Int. J. Hydrogen Energy, 2007, 32, 2664–2672.
50. A. Fujishima and K. Honda, Nature, 1972, 238, 37–38.
51. M. Gratzel, Energy Resources through Photochemistry and Catalysis, Academic Press, New York, 1983.
52. (a) A. Kudo, Int. J. Hydrogen Energy, 2006, 31, 197–202; (b) Y. Horiuchi, T. Toyao, M. Takeuchi, M. Matsuoka and M. Anpo, Phys. Chem. Chem. Phys., 2013, 15, 13243–13253.
53. K.-i. Shimizu, Y. Tsuji, T. Hatamachi, K. Toda, T. Kodama, M. Sato and Y. Kitayama, Phys. Chem. Chem. Phys., 2004, 6, 1064–1069.
54. A. Kudo and Y. Miseki Hyomen, Chem. Soc. Rev., 2009, 38, 253–278.

---

Footnote

† Electronic supplementary information (ESI) available: One figure showing dependence of activation energy E with conversion α obtained by FWO and KAS methods for stages I and II and one table giving the expressions for reaction models have been made available as ESI available. See DOI: 10.1039/c4ra12331j

---