Far infrared radiated energy-proficient rapid one-pot green hydrolysis of waste watermelon peel: optimization and heterogeneous kinetics of glucose synthesis

Abstract

For the first time, a one-pot green hydrolysis of waste watermelon (Citrullus lanatus) peel (WWP) was optimized for a maximum glucose yield employing a heterogeneous Amberlyst-15 catalyst. The effects of energy-proficient far infrared radiation (FIRR) on intensification of pretreatment and subsequent solvent-free hydrolysis reactions in the one-pot system have been maximized. The optimal process variables for pretreatment and consequent hydrolysis were 20 min and 10 min batch times, 70 °C and 60 °C reactor temperatures, and 5 and 10 (w/w) water to WWP ratios, respectively. Optimal 2.5 (w/w) NH₄OH loading and 2.5 wt% catalyst concentration for pretreatment and hydrolysis under FIRR resulted in a maximum glucose yield (89.87 mol%), which was superior to that obtained (59.86 mol%) using a conventional thermal source. In comparison with pseudo-homogeneous and Langmuir–Hinshelwood models, the Eley–Rideal model described the hydrolysis kinetics more accurately. Significantly, a higher hydrolysis activation energy (92.02 kJ mol⁻¹) in the conventional system compared to the FIRR mode (activation energy, 59.69 kJ mol⁻¹) clearly demonstrated the superior energy-efficiency of the FIRR system. The energy-proficient fast hydrolysis process is expected to be sustainable and applicable to similar lignocellulosic biomasses.

---

1. Introduction

Lignocellulosic biomass (LB) is comprises lignin, hemicellulose and cellulose.¹ The relative concentrations of these components depend on the nature of the LB resource. Cellulose is one of the key components from which many important platform chemicals, e.g. glucose, 5-HMF, levulinic acid and bioethanol, can be derived.^2,3 Globally, almost 93 600 million tons of waste watermelon peel (WWP) are available.^4,5 WWP contains a significant amount of cellulose,⁶ indicating its enormous potential as a feedstock for glucose synthesis.

Memon et al. and Huang et al. reported the use of WWP as an adsorbent for the removal of methyl parathion pesticide and Pb(II) from water, respectively.^7,8 The use of watermelon rinds for production of polygalacturonase and xylanase by Trichoderma species was also reported.⁹ Recently, Sandeep et al. synthesized cutin from WWP.^5,10 Notably, no scientific literature has been published on WWP hydrolysis for glucose production, in spite of the presence of a large amount of cellulose in WWP.¹¹

Attempts have been made to conduct homogeneous, acid catalyzed (HCl, H₂SO₄, CH₃COOH) hydrolysis of LB; nevertheless, the attempts have suffered from several difficulties including product purification, apparatus corrosion and waste stream management.^12–14 On the other hand, recent studies have revealed the advantages of heterogeneous catalysts (Nafion-SAC-13, Zr/P/O, PrSO₃H–SiO₂) over homogeneous acid catalysts for cellulose hydrolysis. Nonetheless, heterogeneous hydrolysis requires elevated temperature (150–190 °C),¹⁵ long reaction time (10–24 h)¹⁶ and poor glucose yield (β_G) (5.8–11.3 wt%).¹⁷

In recent years, ionic liquids have been applied as co-solvents for LB hydrolysis to augment the glucose yield.^18–20 However, owing to the additional costs, the overall process is economically unattractive. On the other hand, the hydrolysis of LB for glucose production could be accelerated by microwave radiation (MWR) at the expense of a high power input (1 kW).^21,22 Recently, time-saving and energy-efficient far infrared (FIR) radiation was employed by our research group to accelerate production of biodiesel and glyceryl laurate.^23,24 Notably, the application of FIR radiation (FIRR) to the hydrolysis of LB for glucose production is not available in any scientific literature.

The Taguchi orthogonal design array (TODA) has been increasingly used for process optimization and for the evaluation of parametric interactions governing the response or process output.²⁵ The effects of individual process variables on the response variable were also estimated by TODA.²⁶

Aguilar et al. and other researchers reported on the conventional pseudo-homogeneous (PH) hydrolysis kinetics of sugarcane bagasse.^27–29 Notably, the initial glucose content in the fruit waste was not considered in the formulation of the hydrolysis kinetics. Sanz et al. reported a comparative study among several heterogeneous reaction mechanisms viz., PH, Langmuir–Hinshelwood (LH) as well as Eley–Rideal (ER) models, for the hydrolysis reaction of methyl lactate using the acidic cation-exchange resin, Amberlyst-15 catalyst.³⁰ Bozek-Winkler et al. verified both PH and LH models for trans-esterification of methyl acetate with n-butanol using Amberlyst-15.³¹ Moreover, PH, ER and LH kinetic models for esterification of propionic acid with various alcohols (methanol, ethanol and 1-butanol) over Amberlyst-15 have been reported.³² Notably, no scientific report on heterogeneous hydrolysis kinetics of LB for glucose synthesis are available to the best of our knowledge.

To date, no research work has been reported on WWP hydrolysis for glucose production. In this report, the optimization of pretreatment and consequent hydrolysis of pretreated-WWP (PWWP) to maximize glucose yield has been explored through a one-pot^33,34 conversion employing a novel, energy-saving, far infrared radiation assisted reactor (FIRRAR). The optimal process variables for pretreatment and subsequent hydrolysis of PWWP corresponding to the maximum glucose yield were evaluated by TODA. The performance of the FIRRAR in a one-pot batch glucose production was assessed and compared with the conventional thermal source assisted reactor (CTSAR) in terms of glucose yield and kinetic parameters. The conventional pseudo-homogeneous, pseudo-first order (PHF) hydrolysis kinetic model²⁹ was modified by considering the initial glucose concentration in WWP. Furthermore, under the evaluated optimal conditions, the heterogeneous ER and LH kinetic models were formulated and validated to assess the hydrolysis kinetic parameters.

2. Experimental

2.1. Materials

The waste watermelon peels (WWP) were collected from a local fruit market in Kolkata, India. Analytical reagent grade chemicals viz., acetone, DNS (di-nitro salicylic acid), aqueous NH₄OH (25%), etc. were procured from Merck (India), and Amberlyst-15 was purchased from Sigma Aldrich.

2.2. Effects of process variables on glucose yield

The effects of process independent variables on the response variable (glucose yield, β_G) as well as optimal process variables corresponding to a maximum β_G were estimated by TODA through evaluation of the signal to noise (S/N) ratios and analysis of variance (ANOVA).²⁶ MINITAB-17 (Minitab Inc. USA for Windows 7) provided a set of nine experimental runs (L₉) with different combinations of variables. Each of the nine runs was triplicated to obtain the mean (signal, S) of the response variables (β_G) and the corresponding standard deviation (noise, N). Eqn (1) represents the S/N ratio in terms of the glucose yield (y_i) of each experimental run.Where, n is the total number of runs executed in any particular process variable combinations as per Table 2, and the experimental trial is denoted by i.

For pretreatment, time (μ_tP), reactor temperature (μ_TP), NH₄OH loading (μ_N) and water to WWP weight ratio (μ_WP) were the four self-governing variables affecting the response variable (β_G) (Table 1). Similarly, in the subsequent hydrolysis step, by keeping β_G as the response variable, the hydrolysis time (μ_tH), reactor temperature (μ_TH), catalyst concentration (μ_C) and water to WWP weight ratio (μ_WH) were selected as the self-governing variables (Table 1). The response variable (β_G) was optimized employing the ‘larger is better’ algorithm for the S/N ratio (eqn (1)) for both pretreatment and hydrolysis.

Table 1 Self-governing process variables for one-pot pretreatment-hydrolysis of WWP^a

Process variable	Pretreatment				Hydrolysis
	μTP (°C)	μtP (min)	μN (w/w)	μWP (w/w)	μTH (°C)	μtH (min)	μC (wt%)	μWH (w/w)
a L1 = lower level value, L2 = middle level value, L3 = higher level value.
L1	50	10	0.5	5	60	10	2.5	10
L2	70	20	1.5	10	70	15	5	20
L3	90	30	2.5	15	80	20	7.5	30

---

2.3. Reactor configurations and reaction procedure

The one pot pretreatment-hydrolysis of WWP was performed in a uniquely designed batch reactor provided with an FIRR (150 W; wavelength: 15 μm to 1 mm) system. A centrally fitted mechanical stirrer (400 rpm) was attached to a three-neck 500 mL flask along with a PID temperature controller to keep a constant temperature throughout the reactor mix. A similar design was maintained in the CTSAR except that a 500 W electrical heating mantle was provided as the thermal source instead of the FIRR. The FIRR was applied to modify the conventional AFEX pretreatment³⁵ in the present one-pot conversion. In a representative run, a measured amount of WWP was placed in the flask, followed by addition of a measured amount of aqueous NH₄OH and deionized water at the pre-set, controlled temperature. Over a specified time span, the mix was stirred at 400 rpm (as per Table 2). After pretreatment, the purging of N₂ through one neck of the FIRRAR was carried out to remove any remaining ammonia. Afterwards, a measured amount of Amberlyst-15 catalyst and water were added to the pretreated WWP (PWWP) present in the reactor for the subsequent hydrolysis over a specified time (Table 2). Subsequently, vacuum filtration was conducted to separate the WWP residue and catalyst. Afterward, the filtrate was processed to measure the glucose concentration by a standard DNS method.³⁶ The FIR radiated pretreatment product was also subjected to hydrolysis under conventional heating to compare the performance between the FIRRAR and CTSAR in terms of glucose yield. Moreover, in order to identify the effects of the catalyst on hydrolysis, the optimal PWWP was hydrolyzed in the absence of Amberlyst-15 catalyst using the FIRRAR at TODA predicted optimal conditions. The hydrolysate was analyzed for the glucose yield to compare with the yield from catalytic hydrolysis.

Table 2 TODA for one-pot pretreatment-hydrolysis of WWP

Trial no					Pretreatment			Hydrolysis
					βG	Std	S/N ratio	βG	Std	S/N ratio
1	L1	L1	L1	L1	43.85	±0.17	32.84	89.87	±0.13	39.07
2	L1	L2	L2	L2	45.66	±0.11	33.19	69.27	±0.22	36.81
3	L1	L3	L3	L3	50.82	±0.03	34.12	52.89	±0.17	34.47
4	L2	L1	L2	L3	51.20	±0.22	34.19	74.69	±0.21	37.47
5	L2	L2	L3	L1	61.50	±0.02	35.78	68.55	±0.02	36.72
6	L2	L3	L1	L2	49.27	±0.15	33.85	63.97	±0.14	36.12
7	L3	L1	L3	L2	57.89	±0.21	35.25	61.44	±0.21	35.77
8	L3	L2	L1	L3	52.15	±0.01	34.35	50.84	±0.02	34.12
9	L3	L3	L2	L1	54.09	±0.26	34.66	61.19	±0.13	35.73

---

2.4. X-ray diffraction of WWP

An X-ray diffractometer (INTEL CPS 120 hemispherical detector, Rigaku miniflex Co. Japan) at a scanning speed of 1 min⁻¹ at 40 kV and 30 mA was used to detect changes in the crystalline structure of the cellulose from raw WWP to optimally pretreated and hydrolyzed WWP residue samples at 1.5418 Å wavelengths in a 2θ range between 0° and 50° using a Cu Kα detector.

2.5. FTIR analyses of WWP

The chemical characteristics of raw WWP, optimally pretreated WWP and optimally hydrolyzed WWP residue and the hydrolysate were analyzed by FTIR (FTIR-Shimadzu Alpha, from 400 cm⁻¹ to 4000 cm⁻¹).

2.6. HPLC of hydrolysate

The compositions of the hydrolysates were analyzed by a 300 × 7.8 mm Bio-Rad HPXP, 9 μm column (Agilent 1200 series HPLC with refractive index detector) at a flow rate of 0.6 mL min⁻¹ of deionized water (mobile phase). Using calibration plots of the respective standard constituents, the concentration of each constituent in the hydrolysate was determined.

2.7. Formulation of hydrolysis kinetic models

The heterogeneous ER and LH mechanisms, along with a modified PHF hydrolysis kinetic model of pretreated-WWP, were developed in the present study to evaluate the kinetic parameters at TODA derived optimal process conditions.

2.7.1. Pseudo-homogeneous hydrolysis kinetics. The development of the (PHF) hydrolysis kinetic model of PWWP under optimal process conditions has been formulated using conventional assumptions, viz. the hydrolysis rate is independent of excess water concentration (pseudo-first order approximation);³⁷ and the PWWP hydrolysis is an irreversible first order consecutive series reaction.^28,38

In the present study, the usual PHF kinetics have been reformulated by accounting for the initial glucose concentration (C_M0 mol L⁻¹): i.e.

C_M0 ≠ 0 (2)

where, r_M is the rate of glucose formation (mol L⁻¹ min⁻¹); C_M, C_H and C_WW are concentrations of glucose (mol L⁻¹), water (mol L⁻¹) and PWWP (mol L⁻¹) at time t (min), respectively; k_X and k_M are the reaction rate constants for decomposition of glucose (min⁻¹) and formation of glucose (L mol⁻¹ min⁻¹), respectively.

Considering the initial glucose concentration: t = 0, C_M = C_M0 to t = t, C_M = C_M (5)

Integrating eqn (4) and imposing eqn (5) gives:

C_M = [(k_MC_WW0C_H)/(k_X − k_M)][e^−k_Mt − e^−k_Xt ] + C_M0e^−k_Xt (6)

where, C_WW0 is the initial concentration of cellulose (mol L⁻¹).

2.7.2. Formulation of an Eley–Rideal (ER) type hydrolysis mechanism. In developing the ER model, it was considered that on the catalyst surface (Amberlyst-15), the reactant ‘WW’ (PWWP) does not get adsorbed, and the surface reaction was the rate controlling step.

The suggested mechanism for the ER irreversible kinetics is described below:

H + S → HS (7)

WW + HS → MS (8)

MS → M + S (9)

The adsorption rate of water (H) on the catalyst surface:

−r_H = k_H(C_HC_V − C_HS/K_H) (10)

The surface reaction rate between adsorbed H and unadsorbed WW:

−r_S = k_S(C_WWC_HS − C_MS/K_S) (11)

The desorption rate of F:

−r_M = k_−M(C_MS − C_MC_VK_M) (12)

where k_S, k_H and k_−M are the surface reaction constant, adsorption rate constants of ‘H’ and desorption rate constant of product ‘M’, respectively. K_S, K_H, and K_M are the surface reaction equilibrium constant (L mol⁻¹), adsorption equilibrium constant of reactant ‘H’ (L mol⁻¹), and desorption equilibrium constant of product ‘M’ (L mol⁻¹), respectively. C_MS and C_HS are the occupied surface concentrations of glucose and water, respectively.

Based on the consideration that both the adsorption and desorption rate constants (k_H, k_−M) are quite large, meaning that r_H/k_H and r_M/k_−M tend to zero (as the surface kinetics (eqn (11)) were recognized as the rate controlling step), from eqn (10) and (12) we get,

C_HS = K_HC_HC_V (13)

C_MS = K_MC_MC_V (14)

The total active site concentration can be written as:

C_T = C_V + C_HS + C_MS (15)

Hence, C_V = C_T/(1 + K_HC_H + K_MC_M) (16)

Thus, eqn (13) can be rewritten using eqn (16) as:

C_HS = K_HC_HC_T/(1 + K_HC_H + K_MC_M) (17)

Assuming irreversible surface kinetics,³⁷ eqn (11) can be rewritten as:

−r_S = k_SC_WWC_HS (18)

So, using eqn (17) and (18) can be written as:

−r_S = k_SK_HC_WWC_HC_T/(1 + K_HC_H + K_MC_M) (19)

Rearranging eqn (19) in a suitable form:

r ≈ −r_S = k_EMC_WWC_H/(1 + K_HC_H + K_MC_M) (20)

where, k_EM = k_SK_HC_T = observed reaction rate constant (L mol⁻¹ min⁻¹). (21)

Eqn (20) presents the ER kinetic model for hydrolysis.

2.7.3. Formulation of the Langmuir–Hinshelwood (LH) type hydrolysis mechanism. In developing the dual site LH kinetic model, it is considered that both the reactants, ‘WW’ (WWP) and ‘H’ (water), are adsorbed on the Amberlyst-15 surface and the surface reaction rate is the rate controlling step.

The hypothetical mechanism for LH irreversible kinetics can be presented through the following equations:

WW + S → WWS (22)

H + S → HS (23)

WWS + HS → MS (24)

MS → M + S (25)

−r_WW = k_WW(C_WWC_V − C_WWS/K_WW) (26)

−r_H = k_H(C_HC_V − C_HS/K_H) (27)

−r_S = k_S(C_WWSC_HS − C_MS/K_S) (28)

−r_M = k_−M(C_MS − C_MC_VK_M) (29)

where, k_S, k_H, k_WW and k_−M are the surface reaction constant, adsorption rate constant of ‘H’, adsorption rate constant of ‘WW’ and desorption rate constant of product ‘M’, respectively. K_S, K_H, K_WW and K_M are the surface reaction equilibrium constant, adsorption equilibrium constant of ‘H’, adsorption equilibrium constant of ‘WW’ (L mol⁻¹) and desorption equilibrium constant of product ‘M’, respectively, and C_WWS is the occupied surface concentration of PWWP.

Assuming that both the adsorption and desorption rate constants (k_WW, k_H, k_−M) are considerably large, making r_WW/k_WW, r_H/k_H and r_M/k_−M negligible (as the surface kinetics, eqn (24)) has been considered as the rate controlling step) gives the following equations:

r_WW/k_WW ≈ r_H/k_H ≈ r_M/k_−M ≈ 0 (30)

Thus, eqn (26), (27) and (29) give:

C_WWS = K_WWC_WWC_V (31)

C_HS = K_HC_HC_V (32)

C_MS = K_MC_MC_V (33)

The total active site concentration can be written as:

C_T = C_V + C_WWS + C_HS + C_MS (34)

Hence, C_V = C_T/(1 + K_WWC_WW + K_HC_H + K_MC_M) (35)

Considering irreversible surface kinetics,²⁸ eqn (28) can be rewritten as:

−r_S = k_SC_WWSC_HS (36)

Thus, eqn (36) can be rearranged using eqn (31) and (32) as:

−r_S = k_S(K_WWC_WW)(K_HC_H)C_V² (37)

Thus, −r_S = k_SK_WWK_HC_WWC_HC_T²/(1 + K_WWC_WW + K_HC_H + K_MC_M)² (38)

Rearranging eqn (38) in a suitable form yields:

r ≈ −r_S = k_LMC_WWC_H/(1 + K_WWC_WW + K_HC_H + K_MC_M)² (39)

where, k_LM = k_SK_WWK_HC_T² = the observed reaction rate constant (L mol⁻¹ min⁻¹). (40)

Eqn (39) presents the LH kinetic model.

The Arrhenius equation represents the temperature dependence of k_M, k_X, k_EM and k_LM as expressed in eqn (41):

where, T₂ and T₁ represent two hydrolysis temperatures at optimal conditions. R = 8.314 × 10⁻³ (kJ mol⁻¹ K⁻¹) is the gas constant. The hydrolysis activation energy (kJ mol⁻¹) is denoted by E_WW, and E_WW0 is the pre-exponential factor.

The observed kinetic rate constants, as well as the adsorption–desorption equilibrium constants, were determined by the Levenberg–Marquardt algorithm^39,40 applying MATLAB R2014a.

3. Results and discussion

3.1. Effects of process variables on WWP pretreatment and hydrolysis

The major effects of the governing process variables on the WWP pretreatment obtained in the FIRRAR were evaluated by Analysis of variance (ANOVA) provided in the ESI (Table S1†).

For the pretreatment process, the reaction temperature (μ_TP) and water to WWP ratio (μ_WP) were statistically significant process variables at the 95% confidence level (p-value < 0.05). Notably, the process variable corresponding to a larger Δ-value exhibited (Table 3) a superior effect on the response variable (β_G) compared to other variables. Moreover, the relative significance of the variables on the response (β_G) in pretreatment were μ_N > μ_TP > μ_tP > μ_WP (Table 3). Furthermore, from Table 3 it may also be concluded that, for the subsequent hydrolysis process, at a 95% confidence level (p-value < 0.05), the reaction temperature and catalyst to PWWP ratio were statistically significant. Besides, the relative significance of the process variables on the response (β_G) in hydrolysis were μ_tH > μ_WH > μ_TH > μ_C (Table 3).

Table 3 S/N ratios and Δ of process variables in pretreatment and consequent hydrolysis of WWP

Level	WWP pretreatment				PWWP hydrolysis
	μTP	μtP	μN	μWP	μTH	μtH	μC	μWH
1	33.38	34.09	33.68	34.43*	36.78*	37.44*	36.67*	37.18*
2	34.75*	34.44*	34.01	34.10	36.75	35.88	36.61	36.23
3	34.60	34.21	35.05*	34.22	35.21	35.44	35.65	35.35
Delta (Δ)	1.37	0.35	1.37	0.33	1.57	2.00	1.02	1.83
Rank	2	3	1	4	3	1	4	2

---

The asterisk marks represent the maximum S/N ratio values corresponding to a particular level of the process variables (Table 3), implying that maximum the β_G (61.5 mol%) could be attained in pretreatment variable values, i.e. 20 min (μ_tP), 70 °C (μ_TP), 5.0 (μ_WP) and 2.5 (μ_N). Similarly, the indicated maximum β_G (89.87 mol%) could be achieved in the subsequent hydrolysis process at 10 min (μ_tH), 60 °C (μ_TH), 10.0 (μ_WH) and 2.5 (μ_C). Notably, the FIR radiation has superior effects in terms of the glucose yield (89.87 mol%) over the conventional heating method (59.86 mol%) for the hydrolysis of PWWP at the TODA derived optimal conditions (Table 3).

3.2. Interactive effects of process variables on the pretreatment-hydrolysis process

Retaining other variables at optimum conditions, an increase in the pretreatment temperature for a certain time resulted in an augmented β_G and vice versa (Fig. 1a), which indicated that the medium was thermally sensitive and the pretreatment was endothermic. Similarly, a gradual increase in temperature resulted in augmented β_G and vice versa over the entire range of the NH₄OH loading (Fig. 1b). Fig. 1c depicted that the water loading had negative effect on β_G over the entire range of pretreatment temperature. From Fig. 1d, it could be inferred that, at higher NH₄OH loading ≥1.5, β_G increased monotonically with time because of the enhanced accessibility of cellulose to NH₄OH.³⁵ On the other hand, no appreciable change in β_G was observed over a gradual increase in water loading with time and vice versa (Fig. 1e). From Fig. 1f, it was observed that the augmented yield of glucose could be attained at a higher NH₄OH loading over the entire range of water loading.

Fig. 1 Interaction plots for (a)–(f) pretreatment of WWP; (g)–(l) hydrolysis of PWWP for glucose yield at optimal conditions.

In the one-pot process after pretreatment and in the subsequent hydrolysis step, keeping other variables at an optimal level, a gradual increase in temperature resulted in lower β_G and vice versa owing to the formation of glucose decomposed products (Fig. 1g). On the other hand, an increase in the hydrolysis temperature could result in a gradual decrease in glucose yield with an increase in the catalyst concentration (Fig. 1h). Whereas, Fig. 1i indicated that the water loading had a negative influence on the glucose yield with an increase in temperature because of a possible enhancement of the side reaction. A decrease in the glucose yield was observed with a gradual increase in the catalyst concentration at a given hydrolysis time (Fig. 1j). Similarly, an increase in hydrolysis time resulted in poor β_G at a certain water loading and vice versa (Fig. 1k). Moreover, from Fig. 1l it could be observed that the yield of glucose decreased with water loading at a higher catalyst concentration (>6.0). The individual effect of each process variable has been provided in the ESI (Fig. S1†).

3.3. Validation of predicted results

Pretreatment of WWP and hydrolysis of PWWP were carried out at the TODA predicted optimal variable values (Table 3) in order to validate the TODA derived optimal process variables for both pretreatment and hydrolysis processes. In pretreatment and the subsequent hydrolysis process, acceptable standard deviations of ±0.158 and ±0.187 of β_G (mol%), respectively, were evaluated in the case of FIRRAR.

3.4. Comparison between FIRRAR and CTSAR in pretreatment and hydrolysis

Under the same optimal process conditions, a one pot pretreatment-hydrolysis process conducted in CTSAR resulted in 35.74 mol% and 59.86 mol% glucose yields for pretreatment and subsequent hydrolysis, respectively, indicating the superiority of FIRRAR over CTSAR. A lower glucose concentration of 0.023 mol L⁻¹ was achieved in the pretreatment of WWP using conventional heating in comparison with FIR radiation (0.041 mol L⁻¹) at TODA derived optimal conditions.

A comparative study of the variation in glucose concentration with time in both FIRRAR and CTSAR at TODA predicted optimal hydrolysis conditions is depicted in Fig. 2. It can be observed that a maximum concentration of 0.092 mol L⁻¹, corresponding to 89.87 mol% β_G, was achieved within 10 min of hydrolysis time. On the other hand, a slow increase in the glucose concentration with time was observed (Fig. 2) in the case of CTSAR due to a slower hydrolysis rate, attaining a maximum glucose concentration of 0.043 mol L⁻¹ corresponding to a β_G of 59.86 mol% within an equal period of hydrolysis time (10 min). Notably, keeping all other factors at optimal levels, the hydrolysis of PWWP in FIRRAR without Amberlyst-15 catalyst rendered significantly lower glucose yield (79.87 mol%) compared to that obtained (89.87 mol%) through catalyzed hydrolysis.

Fig. 2 Evolution of glucose concentration during hydrolysis (a) FIRRAR, (b) CTSAR.

It was evident (Fig. 2) that the conversion of PWWP to glucose was accelerated remarkably due to the effect of FIR radiation (150 W); thus, within a short pretreatment-hydrolysis time, maximum β_G could be achieved in comparison with the conventional heating system (500 W). The conversion of PWWP to glucose could be immensely amplified and intensified remarkably as FIR radiation²³ can deeply disrupt the lignin structures and intensely penetrate the pretreatment-hydrolysis reaction mix. Besides, FIR radiation enhanced the stretching vibration of the reactant molecules, rendering severe molecular collisions to promote the reactions. Hence, the application of FIR radiation can be considered as an energy efficient, time-saving and economically attractive process for glucose synthesis from WWP.

3.5. Hydrolysis kinetics of PWWP

The hydrolysis kinetic data were fitted in the modified PHF model (eqn (4)), ER model (eqn (20)) and LH model (eqn (39)) employing MATLAB R2014a at the TODA predicted optimal conditions. For both FIRRAR (Table 4) and CTSAR (Table 5), RMSE, the goodness of fit along with the model p-value were calculated. From Tables 4 and 5, it can be observed that the highest value of R_adj² (0.98 for FIRRAR and 0.974 for CTSAR) and the lowest value of RMSE (3.81 × 10⁻⁴ for FIRRAR and 4.16 × 10⁻⁴ for CTSAR) assured that the ER model could best represent the hydrolysis kinetics in FIRRAR as well as in CTSAR. The observed rate constants for various mechanisms for FIRRAR (Table 4) and CTSAR (Table 5) have been tabulated with their respective kinetic parameters (activation energy and pre-exponential factor).

Table 4 Kinetic parameters of different hydrolysis mechanisms pertinent to FIRRAR

		Temperatures			KH	KM	KWW	Kinetic parameters		Statistical parameters
		333 K	343 K	353 K				EWW	EWW0	Radj^2	RMSE	Model p-value
ER eqn (20)	kEM	2.40 × 10^−3	4.40 × 10^−3	8.15 × 10^−3	7.27 × 10^−2	44.62		59.69	5.49 × 10^6	0.980	3.81 × 10^−4	1.76 × 10^−7
LH eqn (39)	kLM	9.20 × 10^−3	2.44 × 10^−2	6.10 × 10^−2	3.62 × 10^−2	19.14	3.14	92.61	3.09 × 10^12	0.964	5.85 × 10^−4	4.68 × 10^−7
PHF eqn (4)	kM	4.30 × 10^−4	4.82 × 10^−4	5.64 × 10^−4				13.18	4.99 × 10^−2	0.961	5.02 × 10^−4	2.61 × 10^−8

---

Table 5 Kinetic parameters of different hydrolysis mechanisms pertinent to CTSAR

		Temperatures			KH	KM	KWW	Kinetic parameters		Statistical parameters
		333 K	343 K	353 K				EWW	EWW0	Radj^2	RMSE	Model p-value
ER eqn (20)	kEM	2.9 × 10^−4	7.45 × 10^−4	1.90 × 10^−3	2.73 × 10^−3	61.52		92.02	7.81 × 10^10	0.974	4.16 × 10^−4	2.45 × 10^−7
LH eqn (39)	kLM	7.69 × 10^−4	2.57 × 10^−4	7.36 × 10^−3	1.44 × 10^−2	20.49	0.87	110.38	1.61 × 10^14	0.953	6.87 × 10^−4	5.31 × 10^−7
PHF eqn (4)	kM	1.57 × 10^−4	1.97 × 10^−4	2.28 × 10^−4				18.31	1.19 × 10^−1	0.951	5.69 × 10^−4	2.84 × 10^−8

---

From Tables 4 and 5, it is evident that under FIRRAR the observed reaction rate constants are much larger in comparison with CTSAR at all temperatures. Moreover, the activation energies obtained in FIRRAR for various kinetic models are much lower compared to CTSAR under optimal conditions, indicating a lower energy requirement in FIRRAR while attaining significantly superior and rapid glucose yields.

3.6. XRD analyses of WWP

The crystalline phase of cellulose at 2θ = 21.82° is depicted by the X-ray diffractogram of raw WWP (Fig. 3a), optimally pretreated WWP (Fig. 3b) and optimal WWP residue after hydrolysis (Fig. 3c) in FIRRAR.⁴¹ It may be concluded that because of the effect of pretreatment and subsequent hydrolysis the intensity of the cellulose peak became progressively sharper.⁴²

Fig. 3 XRD analyses of (a) WWP residue after optimal hydrolysis in FIRRAR; (b) WWP residue after optimal pretreatment in FIRRAR and (c) raw WWP.

3.7. Analyses of WWP through FTIR

FTIR spectroscopy was used to analyze the characteristic functional groups present in raw WWP (Fig. 4a), optimally pretreated WWP (Fig. 4b) and optimal WWP residue after hydrolysis (using the FIRRAR) (Fig. 4c). The peak around 1408.03 cm⁻¹ represents H–C–H stretching vibration bands while the existence of C=O vibrations was detected at 1055.01 cm⁻¹.⁴³ The peak at 1321.22 cm⁻¹ confirmed the CH deformation stretching. Besides, the presence of the O–H vibration of adsorbed water molecules was recorded at 1604.77 cm⁻¹. The existence of CH₂ asymmetric stretching was detected by the peak located at 2916.37 cm⁻¹.⁴⁴ The stretching vibration of the O–H functional group was found at 3302.13 cm⁻¹. Careful observations of Fig. 4 revealed that the conversion of raw WWP through pretreatment and hydrolysis resulted in a gradual increase in the peak areas and intensities.⁴⁵

Fig. 4 FTIR analyses of (a) WWP residue after optimal hydrolysis in FIRRAR; (b) WWP residue after optimal pretreatment in FIRRAR; (c) raw WWP.

3.8. Analyses of hydrolysate through FTIR

Fig. 5a–c represented FTIR spectra of standard glucose and hydrolysate obtained through FIRRAR and CTSAR, respectively. The prominent peaks at 1619.22 cm⁻¹ and 1162.85 cm⁻¹ indicated the presence of C=C and C–O–C bonds, respectively.⁴⁶ The vibration bands at 1460.52, 1426.13, 1381.47 and 1207.39 cm⁻¹ confirmed the presence of the H–C–H group.⁴⁷ Peaks at 2965.31 and 2937.11 cm⁻¹ indicated the CH stretching vibrations.⁴⁸ On the other hand, peaks at 987.15 and 889.22 cm⁻¹ are characteristic of CH₂ vibrations.⁴⁸ The bands observed at 743.69 and 610.37 cm⁻¹ represent vibrations of some functional groups in anomeric form.⁴⁹ Furthermore, the stretching vibration at 3443.78 cm⁻¹ can be attributed to the existence of the O–H group. Thus, from Fig. 5, it is evident that the presence of the characteristic functional groups corresponding to glucose (Fig. 5a) was more prominent in the optimal hydrolysate produced in FIRRAR (Fig. 5b) compared to the hydrolysate obtained from CTSAR (Fig. 5c).

Fig. 5 FTIR analyses of (a) standard glucose; (b) hydrolysate of FIRRAR; (c) hydrolysate of CTSAR.

3.9. HPLC analyses of hydrolysate

The quantitative analyses of hydrolysates containing glucose and other products obtained via hydrolysis of PWWP through both FIRRAR and CTSAR at the optimal conditions were assessed with HPLC analyses (ESI, Fig. S2†). The retention times corresponding to 5-HMF, fructose and glucose were 23.97 min, 13.37 min and 8.52 min, respectively.⁵⁰ The concentrations of 5-HMF (12 mol%) and fructose (27 mol%) were significantly lower in comparison to the glucose concentration (61 mol%).

3.10. Mechanism of Amberlyst-15 catalyzed hydrolysis

In the present study, the ER model was found to be the best (according to the statistical criteria of Tables 4 and 5) to represent the hydrolysis kinetics. According to the ER kinetic model, one of the reactants, i.e. water, gets adsorbed onto the surface of the Amberlyst-15 catalyst⁵¹ while the other reactant (PWWP) remains unadsorbed. Moreover, water molecules (0.96 Å) can easily diffuse into the pores (300 Å) of the Amberlyst-15 catalyst, facilitating further chemisorption. The mechanism of the hydrolysis reaction (Fig. 6) involves the molecular interactions between the water and catalyst, facilitating proton (H⁺) release from the catalyst, resulting in the generation of hydronium ions.⁵² The H₃O⁺ ion gives rise to the formation of H⁺ and OH⁻ ions. Subsequently, the H⁺ ion attacks the 1,4′-β-glycosidic linkage of cellulose leading to the formation of a cyclic carbonium ion with a half-chair configuration. Finally, a glucose molecule is formed due to OH⁻ ion transfer to the carbonium ion.⁵³

Fig. 6 Proposed mechanism of the Amberlyst-15 catalyzed hydrolysis of pretreated waste watermelon peel.

4. Conclusions

The article presents a significantly improved, energy-efficient, one-pot hydrolysis of waste watermelon peel under far infrared radiation to render an augmented and accelerated glucose yield in comparison with the conventional thermal source. The optimal conditions in pretreatment and consequent hydrolysis of waste watermelon peel were efficiently estimated by the Taguchi orthogonal design along with evaluation of the interactions among the process variables. The heterogeneous Amberlyst-15 catalyzed one-pot hydrolysis was found to be best validated by the Eley–Rideal mechanism in comparison with the Langmuir–Hinshelwood mechanism. The estimated kinetic parameters of far infrared radiated hydrolysis protocol can be utilized to scale-up such efficient reactors. Thus, the developed method provides a promising, economically attractive, green technology for the synthesis of glucose from waste watermelon peel and is expected to be applicable to similar lignocellulosic biomasses.

Acknowledgements

The corresponding Author is grateful to University Grants Commission, New Delhi, India for financial support through Major Research Project [(F. No. 43-161/2014 (SR)].

References

1. L. Yan, A. A. Greenwood, A. Hossain and B. Yang, RSC Adv., 2014, 4, 23492–23504.
2. X. Qian, J. Lei and S. R. Wickramasinghe, RSC Adv., 2013, 3, 24280–24287.
3. E. R. E. Hassan, F. Mutelet, J. Moise and N. Brosse, RSC Adv., 2015, 5, 61455–61464.
4. M. P. Tarazona-Díaz, J. Viegas, M. Moldao-Martins and E. Aguayoa, J. Sci. Food Agric., 2011, 91, 805–812.
5. J. P. Maran, V. Sivakumar, K. Thirugnanasambandham and R. Sridhar, Carbohydr. Polym., 2014, 101, 786–791.
6. P. Kumar, Indian Food Packer, 1985, 39(4), 49–52.
7. G. Z. Memon, M. I. Bhanger, M. Akhtar, F. N. Talpur and J. R. Memon, Chem. Eng. J., 2008, 138, 616–621.
8. K. Huang, L. Liu, B. Jiang and H. Zhu, Electrometallurgy and Materials Characterization, John Wiley & Sons, 2012, pp. 671–680.
9. S. A. Mohamed, A. L. Al-Malki, J. A. Khan, S. A. Kabli and S. M. Al-Garni, J. Microbiol., 2013, 51, 605–611.
10. S. A. Chaudhari and R. S. Singhal, Int. J. Biol. Macromol., 2015, 79, 398–404.
11. H. M. A. Al-Sayed and A. R. Ahmed, Ann. Agric. Sci., 2013, 58(1), 83–95.
12. B. P. Lavarack, G. J. Griffin and D. Rodman, Biomass Bioenerg., 2002, 23, 367–380.
13. H. M. Baudel, C. Zaror and C. A. M. de Abreu, Ind. Crops Prod., 2005, 21, 309–315.
14. L. Wang, Z. Luo, S. Xiu and A. Shahbazi, Energy Sources, Part A, 2011, 33(13), 1230–1238.
15. J. Hegner, K. C. Pereira, B. DeBoef and B. L. Lucht, Tetrahedron Lett., 2010, 51, 2356–2358.
16. G. Gliozzi, A. Innorta, A. Mancini, R. Bortolo, C. Perego, M. Ricci and F. Cavani, Appl. Catal., B, 2014, 145, 24–33.
17. B. Wiredu and A. S. Amarasekara, Catal. Commun., 2014, 48, 41–44.
18. H. L. Cai, C. Z. Li, A. Q. Wang, G. L. Xu and T. Zhang, Appl. Catal., B, 2012, 123–124, 333–338.
19. H. X. Guo, X. H. Qi, L. Y. Li and R. L. Smith, Bioresour. Technol., 2012, 116, 355–359.
20. M. Liu, S. Y. Jia, Y. Y. Gong, C. S. Song and X. W. Guo, Ind. Eng. Chem. Res., 2013, 52, 8167–8173.
21. Z. Zhang and Z. K. Zhao, Carbohydr. Res., 2009, 344, 2069–2072.
22. J. Biana, P. Pengb, F. Penga, X. Xiaoa, F. Xua and R. Sun, Food Chem., 2014, 156, 7–13.
23. P. Pradhan, S. Chakraborty and R. Chakraborty, Fuel, 2016, 173, 60–68.
24. R. Chakraborty and E. Mandal, J. Taiwan Inst. Chem. Eng., 2015, 50, 93–99.
25. A. B. Engin, O. Ozdemir, M. Turanb and A. Z. Turanc, J. Hazard. Mater., 2008, 159, 348–353.
26. Y. Lin, A. Wang, D. Wang and C. Chen, Mater. Manuf. Processes, 2009, 24, 667–674.
27. R. Aguilar, J. A. Ramirez, G. Garrote and M. Vazquez, J. Food Eng., 2002, 55, 309–318.
28. A. R. Chong, J. A. Ramirez, G. Garrote and M. Vazquez, J. Food Eng., 2004, 61, 143–152.
29. S. Kumar, P. Dheeran, S. P. Singh, I. M. Mishra and D. K. Adhikari, Renewable Energy, 2015, 83, 850–858.
30. M. T. Sanz, R. Murga, S. Beltran, J. L. Cabezas and J. Coca, Ind. Eng. Chem. Res., 2004, 43, 2049–2053.
31. E. Bozek-Winkler and J. Gmehling, Ind. Eng. Chem. Res., 2006, 45, 6648–6654.
32. S. H. Ali, Int. J. Chem. Kinet., 2009, 41, 432–448.
33. E. Nikolla, Y. Roman-Leshkov, M. Moliner and M. E. Davis, ACS Catal., 2011, 1, 408–410.
34. S. De, S. Dutta and B. Saha, ChemSusChem., 2012, 5, 1826–1833.
35. F. Teymouri, L. Laureano-Perez, H. Alizadeh and B. E. Dale, Bioresour. Technol., 2005, 96, 2014–2018.
36. G. L. Miller, Anal. Chem., 1959, 31, 426–428.
37. I. S. M. Rafiqula and A. M. M. Sakinah, Chem. Eng. Sci., 2012, 71, 431–437.
38. A. I. Malester, M. Green, S. Kimchie and G. Shelef, Biol. Wastes, 1988, 26, 115–124.
39. D. Marquardt, SIAM J. Appl. Math., 1963, 11, 431–441.
40. K. Levenberg, Q. Appl. Math., 1994, 2, 64–168.
41. S. Xia, G. A. Baker, H. Li, S. Ravulab and H. Zhao, RSC Adv., 2014, 4, 10586–10596.
42. H. Zhao, J. H. Kwak, Z. C. Zhang, H. M. Brown and B. W. Arey, Carbohydr. Polym., 2007, 68, 235–241.
43. E. J. Johnson, O. Dorot, J. Liu, B. Chefetz and B. Xing, Commun. Soil Sci. Plant Anal., 2007, 38, 2461–2478.
44. L. Espana, J. A. Heredia-Guerrero, P. Segado, J. J. Benitez, A. Heredia and E. Dominguez, New Phytol., 2014, 202, 790–802.
45. T. F. Meyabadi, F. Dadashian, G. M. M. Sadeghi and H. E. Z. Asl, Powder Technol., 2014, 261, 232–240.
46. G. Socrates, Infrared and Raman characteristic group frequencies: tables and charts, John Wiley & Sons Inc., New York, 2004.
47. S. Pavlovic and P. R. G. Brandao, Miner. Eng., 2003, 16, 1117–1122.
48. H. A. Wells Jr and R. H. Atalla, J. Mol. Struct., 1990, 224, 385–424.
49. M. Dauchez, P. Derreumax and G. Vergoten, J. Comput. Chem., 1992, 14, 263–277.
50. W. Namchot, N. Panyacharay, W. Jonglertjunya and C. Sakdaronnarong, Fuel, 2014, 116, 608–616.
51. D. Lorenzo, E. Simon, A. Santos and A. Romero, Ind. Eng. Chem. Res., 2014, 53, 19117–19127.
52. P. Bhaumik and P. L. Dhepe, Biomass Sugars for Non-Fuel Applications, 2015, ch. 1, pp. 1–53.
53. C. H. Zhou, X. Xia, C. X. Lin, D. S. Tong and J. Beltramini, Chem. Soc. Rev., 2011, 40, 5588–5656.

---

Footnote

† Electronic supplementary information (ESI) available: Table S1 presents the effects of process variables on one-pot WWP pretreatment-hydrolysis through analysis of variance (ANOVA). Fig. S1 depicts the individual effects of each process variable on one-pot (a) pretreatment, (b) hydrolysis of WWP in FIRRAR at optimal conditions. Fig. S2 exhibits the HPLC analyses of hydrolysates obtained by employing (a) FIRRAR, (b) CTSAR. See DOI: 10.1039/c6ra13391f

---