High yield of isosorbide production from sorbitol dehydration catalysed by Amberlyst 36 under mild condition

Journal Pre-proofs High Yield of Isosorbide Production from Sorbitol Dehydration Catalysed By Amberlyst 36 under Mild Condition Muhammad Ridzuan Kamaruzaman, Xiao Xia Jiang, Xiu De Hu, Sim Yee Chin PII: DOI: Reference:

S1385-8947(20)30177-7 https://doi.org/10.1016/j.cej.2020.124186 CEJ 124186

To appear in:

Chemical Engineering Journal

Received Date: Revised Date: Accepted Date:

20 September 2019 3 January 2020 21 January 2020

Please cite this article as: M. Ridzuan Kamaruzaman, X. Xia Jiang, X. De Hu, S. Yee Chin, High Yield of Isosorbide Production from Sorbitol Dehydration Catalysed By Amberlyst 36 under Mild Condition, Chemical Engineering Journal (2020), doi: https://doi.org/10.1016/j.cej.2020.124186

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2020 Elsevier B.V. All rights reserved.

High Yield of Isosorbide Production from Sorbitol Dehydration Catalysed By Amberlyst 36 under Mild Condition Muhammad Ridzuan Kamaruzamana Xiao Xia Jiangb,c Xiu De Huc Sim Yee China,d,*[email protected] aFaculty

of Chemical & Process Engineering Technology, Universiti

Malaysia Pahang, Lebuhraya Tun Razak, 26300 Gambang, Kuantan, Pahang, Malaysia bSchool

of Mechanical Engineering, Ningxia University, 750021, Ningxia, Yinchuan, China

cState

Key Laboratory of High-efficiency Utilization of Coal and Green

Chemical Engineering, Ningxia University, 750021, Ningxia, Yinchuan, China dCenter

of Excellence for Advanced Research in Fluid Flow, Universiti

Malaysia Pahang, Lebuhraya Tun Razak, 26300, Kuantan , Pahang, Malaysia *Corespondence Authors Highlights

Abstract Isosorbide (ISB), one of the important polyols, can be produced through the sequential intramolecular dehydration of sorbitol (SL) derived from an abundance renewable biomass resources. The advantages of its rigid structure have granted the ISB a wide application in the polymer industries. An acidic catalyst in the liquid phase was conventionally used in the dehydration process. This homogeneously catalysed reaction gave low ISB yield and required additional downstream processes to separate the catalyst. The present study employed solid acidic ion exchange resin, Amberlyst 36 in the SL dehydration at a mild condition. The effect of important operating parameters such as stirring speed, catalyst loading, temperature and reaction time was investigated. The increase of catalyst loading from 5 to 7 wt% did not significantly affect the ISB yield. A higher temperature increased the reaction rate whereas a prolonged reaction time increased the conversion of SL and yield of ISB to the

maximum. In terms of giving a higher ISB yield during SL dehydration, AM 36 was found to outperform the other resin catalysts reported in the literature. Both SL conversion and ISB yield of >99% were recorded after a 4 h reaction at 423 K with catalyst loading of 5 wt% and stirring speed of 300 RPM. The reaction kinetics was evaluated under a mass transfer resistances free condition at the reaction temperature ranged from 373 K to 423 K. The kinetic data well fitted to the Langmuir-Hinshelwood (LH2) model that took side reaction into account. The activation energy for dehydration SL to ST, dehydration ST to ISB and dehydration of SL to other side products such as humins were 109.22, 109.46 and 104.17 kJ/mol respectively. Keywords:

Sorbitol

Dehydration;

Kinetic

Modeling;

Sorbitan;

Isosorbide; 1. Introduction Sorbitol (SL), a polyol derived from carbohydrate is identified as one of the top renewable chemical building blocks with increasing importance in the manufacturing of bio-based material to substitute

petroleum-derived products [1]. Among the SL anhydrides, isosorbide (ISB) molecule has attracted great interest of the researchers and industries in the field of applied polymer chemistry. ISB with a V-shaped backbone composes of two fused furan units and two hydroxyl groups with unequal reactivity, exhibiting unique rigid scaffold associated with the bi-functionality that enables its potent applications in the polymer industries [2][3]. In particular, the use of ISB as a monomer in polyester production has significantly improved the glass transition temperature, transparency and mechanical property of the polymer [4]. ISB can be produced via sequential intra-molecular SL dehydration as depicted in Figure 1. Amidst the sorbitans (STs) produced from the first dehydration of SL, only 3,6-sorbitan (3,6-ST) and 1,4-sorbitan (ST) are the precursors of ISB through the second dehydration step. The other STs initiate the formation of coloured oligomeric side product species called “humins” after subsequent dehydration steps [5][6]. Acidic catalyst is commonly used to increase the reactivity of the SL molecule during the dehydration reaction. Commercially, ISB is

produced through homogeneously catalysed SL dehydration using strong Brönsted acid such as sulphuric acid. It was reported that the formation and yield of ISB highly reliant on the nature of acid sites and their acidic strength. In comparison the Lewis acid, a strong Brönsted acid catalyst exhibited higher activity and selectivity to ISB. The superior performance of strong Brönsted acid catalyst was ascribed to its higher affinity to protonate the hydroxyl group of SL, one of the crucial steps in the mechanism of ISB formation through SL dehydration. Moreover, the ISB synthesis via SL dehydration that employed the weak acidic catalyst like Lewis acid was also more susceptible to the formation of side products especially coke [7][8][9][10].

The homogeneous strong Brönsted acid catalysts such as HCl, H2SO4, H3PO4 and p-toluene sulfonic acid have shown exceptional performance in catalysing SL dehydration for the production of ISB under normal atmospheric pressure, elevated pressure and microwave-assisted condition [12][13][14][15][16]. Nevertheless, the pre-eminence of the

homogeneous Brönsted acid catalyst is offset by its severe drawbacks in resulting equipment corrosion and additional downstream separation processes that require extra capital and operating cost [13][17]. From the economic and environmental point of view, the catalysis of SL dehydration for ISB production has geared towards a green and sustainable heterogeneous catalytic system. In recent years, numerous supported strong Brönsted acids have been adopted as heterogeneous catalysts in the SL dehydration for ISB synthesis. Tungstophosphoric acids (PW) supported on various metal oxides were used as a catalyst to dehydrate aqueous SL into ISB. The best ISB yield attained at SL conversion of 95% was only 56% in a 30% PW/SiO2 catalysed reaction that operated at 523 K and atmospheric pressure. The formation of substantial side products such as humins under elevated reaction temperature has reduced the yield of ISB [7]. Considering the requirement of a milder operating condition, the supported sulfonic acid on a different type of supports has been researched extensively for the dehydration of SL to ISB. These catalysts encompass

sulfonic acid on silica such as hydrophilic sulfonic acid functionalized microsphere silica (SA-SiO2) and propyl sulfonic acid functionalized mesostructured SBA-15 silica [8][18], sulfonic acid on carbon-coated magnetic particle such as glucose derived magnetic sulfonic acid (GluFe3O4-SO3H) [19], mesoporous carbon-based sulfonic acid (MCPhSO3H) [6] as well as sulfonic acid resins include Purolite type resins, Amberlyst (AM) type resins and hydrophobic mesoporous polymer-based sulfonic acid catalyst (P-SO3H) [13][15][16][20]. Among these sulfonic acid solid catalysts, P-SO3H exhibited the highest activity and selectivity as indicated by the attainable SL conversion of 99% and ISB yield of 87.9% at 413 K for 10 h. The superiority of P-SO3H was attributed to its tuneable surface hydrophobicity, pore structure and a number of active sites. Nevertheless, it is noteworthy to state that the superiority of P-SO3H was also achieved at the expense of the use of vacuum conditions to continuously evacuate the water formed during the reaction [15]. Under the normal atmospheric pressure, commercial sulfonic resin, Purolite CT269 resulted in the

complete conversion of SL and reasonably high ISB yield (75%) when it was used as the solid acid catalyst for SL dehydration without simultaneous water removal at 413 K for 12 h [20]. Conversely to the researches on the solid acid catalyst development for the SL dehydration to ISB, the investigation on the product distribution and kinetics of this reaction is rather scarce despite its importance for process and equipment design. Yamaguchi et al. [21] carried out the kinetic analysis of the SL dehydration in high-temperature liquid water without adding any acid catalysts at 523-573 K and 100 bar. The water molecules acted as a proton donor to acid-catalyse the reaction at high temperature. Pseudo-homogeneous (PH) model was used to fit the kinetic data. It was predicted that the maximum ISB yield of 57% was achieved at 590 K after 1 h reaction. Meanwhile, Polaert et al. [16] evaluated the kinetics of microwave-assisted ISB synthesis through SL dehydration using a heterogeneous catalyst (AM 35) under vacuum condition. The obtained yield was up to 70% at 413 K for 5 h. The kinetic data were fitted with both PH and Langmuir Hinshelwood (LH) models.

LH model was found to best describe the reaction kinetics of SL dehydration to ISB. Differ from the reaction without using a catalyst, the energy barrier of the dehydration of ST to ISB was lower than the one for the dehydration of SL to ST with the introduction of AM 35 as a catalyst [16]. With the ultimate aim to reduce the cost of a typical vacuum or highpressure operation, the present study opted to use the commercial sulfonic acid resins Amberlyst 36 (AM36), a counterpart to Purolite CT269, as the catalyst for a batch-wise SL dehydration to ISB at atmospheric pressure with concurrent water removal. The parametric and kinetic studies of the SL dehydration were carried out to identify the best operating condition and evaluate the reaction mechanism. The kinetic model validated with the experimental data at different operating conditions can then be used for reaction optimisation and reactor scale-up purposes. It is also essential information used to determine the distribution of the SL anhydrides which defines the quality of the SL esters.

2. Materials and Methods 2.1 Materials The reactant for dehydration reaction, D-sorbitol (97%) was purchased from Sigma-Aldrich while the catalyst, Amberlyst 36 (AM36) was supplied by Acros. Isosorbide (dianhydro-D-glucitol, 98%, Santa Cruz) and 1,4 sorbitan (1,4-anhydro-D-sorbitol, 97%, TRC) were used as the standard for gas chromatography (GC) analysis. Prior to GC analysis, the samples underwent silylation by mixing it with reagents pyridine (99.5%), chlorotrimethylsilane (99%) and hexamethyldisilazane (98%) procured from Merck. All these chemicals were used as received without further purification. 2.2 Sorbitol Dehydration Reaction Study The SL dehydration was carried out in a 500 ml 3-necked round bottom flask placed in a rotamantle equipped with a temperature controller as shown in Figure 2. The flask was connected to a Liebig condenser circulated with hot oil (383 K) to condense SL. The Liebig condenser was followed by a Dean-Stark apparatus attached with Graham

condenser to condense the water vapour formed during the reaction. SL was first melted in the flask before it was further heated up to the desired reaction temperature. Then, the catalyst was introduced to initiate the reaction. Throughout the reaction duration, samples were withdrawn at the desired time interval. Prior to storage and GC analysis, the samples were cooled in an ice bath to cease the reaction. The important reaction parameters varied during the reaction study included the stirring speed (0 – 400 RPM), catalyst loading (1–7 wt%) and reaction temperature (373– 423 K). All experimental runs were repeated twice to ensure reproducibility.

2.3 Gas Chromatography Analysis The compositions of SL, ST and ISB in the samples collected from dehydration reactions were analysed using gas chromatography equipped with a flame ionisation detector (GC-FID). The sample was pre-treated using a silylation method to increase its volatility before injecting into GC. The sample with a weight of 50 mg was first diluted in 3 ml of

pyridine. Under vigorous stirring, 0.4 ml of hexamethyldisilazane and 0.3 ml of chlorotrimethylsilane were then added consecutively. The sample was stirred for 30 sec before it was left for 15 minutes to complete the reaction. All samples were filtered and injected into GC-FID installed with CP-TAP column (25 m length x 0.25 mm i.d. x 0.1 μm film thickness). As adopted from the previous study [22], the oven initial temperature was held for 7 minutes before it started ramping up from 373 to 623 K at a rate of 8 K/min. The detector and injector temperatures were set at 523 K and 653 K respectively. Meanwhile, the carrier gas (helium) flow rate was adjusted to 0.1 ml/min with the split ratio of 10:1. The standard samples with different concentrations of SL, ST and ISB were prepared and analysed to generate the multiple point calibration curve for the respective compound. The concentration of SL, ST and ISB in the product sample was subsequently determined. The conversion of SL, the yield of ST, yield and selectivity of ISB were calculated using Eq. 1 to Eq. 4, respectively.

SL conversion = (NSLi - NSLt)/NSLi × 100 1 ST yield = NSTt/NSLi × 100

2 ISB yield = NISBt/NSLi

× 100

3

ISB Selectivity = ISB yield/ SL conversion

4

where, N is the number of moles, i denotes the initial state and t denotes certain time instant. 2.4 1H NMR (Nuclear Magnetic Resonance) Analysis The products withdrawn from the SL dehydration reaction study were subjected to 1H NMR analysis to further validate the presence of ISB. The analysis was performed using a Bruker Ultrashield 500 Plus spectrometer working at 500 MHz and ambient temperature. Deuterated

water was used as a solvent and while tetramethylsilane was used as the internal reference.

2.5 Kinetic Study The SL dehydration is represented by two reactions in series. SL is first dehydrated to ST through the chemical reaction as shown in Eq. 5 and subsequently ST is converted to ISB follows Eq. 6. The overall reaction by neglecting the by-product (water) is shown in Eq. 7.

k1

SL ST + H2O 5

ST O

k2

6

ISB + H2

k1

k2

SL ST I SB

7

The reaction mechanism can be elucidated using different types of kinetic models that covering the PH model, LH1 model and LH2 model as presented in Eq. 8, Eq. 9 and Eq. 10 respectively. LH1 model adopted the model proposed by Polaert et al. [16] in which the adsorption of ST was ignored during the dehydration of ST whereas the LH2 model considered both adsorptions of the SL and ST. The unknown rate and adsorption parameters (k1, k2, KSL and KST) of each model were determined using non-linear regression in Polymath 6.1.

rST = k1CSL - k2CST 8

rST = (k1CSL/(1 + KSLCSL)) ― k2CST 9

rST = (k1CSL/(1 + KSLCSL)) - (k2CST/1 + (KSTCST ))

10

where, k1 and k2 are the rate constants, Ci is the concentration for species i, Ki is the adsorption constant for species i, the subscript of SL, ST are sorbitol and 1,4 sorbitan, respectively, and rST is ST experimental reaction rate. The experimental reaction rate of each compound was determined by differentiating the respective concentration-time profile. The rate constant can be expressed as a function of temperature through Arrhenius equation in Eq. 11.

ln k = -Ea/RT + ln A

11

where, Ea is the activation energy while A is the pre-exponential factor. With the aim to generate the predicted concentration profile of each compound in the reaction, the kinetic model that well representing the SL dehydration was incorporated into the mole balance equation of ST species in a batch-wise system as shown in Eq. 12.

rSTV = dNST /dt

12

where, N is number of moles and V is reactant volume. Eq. 12 can be rewritten in terms of concentration, C as given in Eq. 13

rSTV = dCSTV/dt = VdCST/dt + CST dV/dt

13

The density of reaction mixture was assumed constant, allowing the volume changes of the system to be represented using Eq. 14.

V = V0 - v0 t

14

where, V0 is initial reactant volume in ml and v0 is the water outlet flow rate in ml/h.

The integration of Eq. 14 was substituted in Eq. 13, generating Eq. 15. Eq. 15 is also applicable to other compounds by replacing the CST with CSL and CISB, respectively.

dCST/dt = rST + (v0CST )/V

15

The kinetic model expressed in Eq. 8, Eq. 9 or Eq. 10 was incorporated into Eq. 15 before it was solved using the Runge-Kutta method embedded in Ordinary Differential Equation (ODE) solver, Polymath 6.1. The predicted concentration of each compound at a certain time instant was generated and compared with the experimental concentration data. 3. Results and Discussion 3.1 Effect of Important Operating Parameters 3.1.1 Effect of Mass Transfer Resistance

The SL dehydration reactions were carried out at different stirring speeds (0 -400 RPM) for 4 h to identify the best stirring speed that could minimise the external mass transfer limitation. Figure 3 shows that the conversion and product yield did not change significantly at 4 h when stirring was introduced at 200 RPM and above during the reaction. The average repeatability standard deviation for different stirring speed for SL conversion, ST yield and ISB yield were 0.6465, 0.5980 and 0.3907, respectively. The catalyst was dispersed in the liquid phase even at low stirring speed and hence reducing the thickness of the boundary layer on the catalyst surface that resulted in minimum external diffusion resistances of the reactant and products at the surface of the catalyst [16]. In addition, the negligible external mass diffusion resistance was also affirmed by the Mears Parameters, CM in Table 1 that fulfilled the criteria as stated in Eq. 16.

𝐶𝑀 = ―𝑟𝑆𝐿,𝑜𝑏𝑠𝜌𝑏𝑅𝑐𝑛/(𝑘𝐶𝐶𝑆𝐿,𝑏 ) < 0.15

16

where, rSL,obs is the reaction rate of SL, n is the reaction order, Rc is the catalyst particle radius, b is the bulk density of the catalyst, CSL,b is the bulk concentration of SL and kC is the mass transfer coefficient. In addition, Weisz Prater Parameter (CWP) calculated using Eq. 17 was used to validate the significance of internal diffusion resistances during the SL dehydration. Table 1 shows that CWP at different stirring speeds satisfied the criterion given in Eq. 17. Therefore, the internal diffusion resistances were trivial.

𝐶𝑊𝑃 = ―𝑟𝑆𝐿,𝑜𝑏𝑠𝜌𝑏𝑅2𝑐 /𝐷𝑒𝑓𝑓𝐶𝑆𝐿 < 1

17

where, Deff and CSL represent the effective diffusivity and the SL concentration in the mixture, respectively. The initial reaction rate at different stirring speeds as displayed in Figure 4 assured that the further increment of speed to above 200 RPM was not necessary for the SL dehydration. Nevertheless, 300 RPM was

chosen for subsequent studies to ensure a definite mass transfer free condition.

Table 1: CM and CWP for sorbitol dehydration for different stirring speed Stirring CM (x

CWP (x

10-4)

10-3)

0

2.73

4.92

100

2.65

4.76

200

2.67

4.81

300

2.63

4.73

speed (RPM)

400

2.64

4.75

3.1.2 Effect of Catalyst Loading The catalyst dosage is an important parameter that affects the rate of reaction and needs to be optimised to achieve a desired catalytic behaviour. The increasing catalyst loading can increase the accessible active sites for the reactant which leading to satisfactory catalytic reaction [23]. The SL dehydration was carried out at various catalyst loadings (19 wt%) by retaining the other operating conditions. The blank experiment without using catalyst did not produce any sorbitol anhydrides (ST and/or ISB) yield. Figure 5 reveals a progressive increase of SL conversion (nearly 100%), ST yield and ISB yield. The average standard deviation for SL conversion, ST yield and ISB yield were 0.9201, 0.9089 and 0.9699, respectively. Further addition of catalyst loading from 5 wt% to 7 wt% did not significantly impact the SL conversion but reducing the ST and ISB selectivity of circa 10%. The excessive acid sites have caused the

formation of a polymer by self-polymerisation of SL anhydrides of crosspolymerisation SL anhydrides with other products [21][23]. The catalyst loading of 5 wt% was chosen for the subsequent studies. Regardless of the dissimilarity in the acidity of various types of solid catalyst and operating conditions of the SL dehydration adopted in other research studies, the best catalyst loading identified was within the range of the catalyst

loadings

(4-20

wt%)

reported

in

the

literature

[15][16][17][20][24][25].

3.1.3 Effect of Reaction Temperature Figure 6 demonstrates the influence of reaction temperature to the reaction performance of the SL dehydration when it was varied from 373 to 423 K. The maximum temperature used in the study was 423 K due to the thermal stability of AM 36. The reaction was prolonged until the dehydration reaction reached the maximum and constant conversion. All reaction temperatures led to a final SL conversion of nearly 100% but at

a different time frame. At a reaction temperature of 423 K, the SL conversion reached 100% within 4 h, giving the corresponding ST yield and ISB yield of 50% and 100%. ST required about 12 h to be completely converted to ISB. On the other hand, the reaction at a temperature of 373 K could only reach 100% SL conversion at 120 h, resulting in the ST yield and ISB yield of 38% and 54% respectively. The reproducibility of the experiments was ascertained by the average standard deviation for SL conversion, ST yield and ISB yield of 0.9665, 0.8081 and 0.8852, respectively. Albeit the status of the highest temperature in the present study, a reaction temperature of 423 K is a relatively mild condition comparing to the typical SL dehydration process [7][17].

No ISB

degradation was evidenced as the ISB yield remained consistent even though the reaction time was extended until 276 h after the yield reached peak value. Therefore, a reaction temperature of 423 K was chosen as the best temperature.

Table 2 compares the reaction performance of the SL dehydration catalysed by different types of sulfonic acid resin catalysts. All these sulfonic acid resin catalysts resulted in the SL conversion of 100%. Nevertheless, the catalyst employed in the present study, AM 36 outperformed the other resin catalysts by giving a significantly higher ISB yield. The outstanding performance of AM 36 could be ascribed to its higher ion exchange capacity and activity. The activity retained during the course of the reaction because the water was removed continuously from the reaction mixture using a Dean-Stark apparatus. The resin catalyst deactivated when the water molecule produced a hydrating shell that could prevent reactant molecules from reaching the active sites [6]. Under a continuous water removal condition, the reversible reaction of ISB dehydration could also be suppressed although its likelihood was low but plausible as confirmed by Cubo et al. [18]. Moreover, the smaller pore size of AM 36 also served as a conducive confined space for the formation of 1,4 sorbitan, a precursor for the formation of ISB [24].

Table 2: Comparison of ion exchange resin performance for SL dehydration Characterist Reaction conditions

Performance ic Acid

Catalyst

Catalys

Con

Yield/

Ti tLoadin Temperat

versi selecti me

g

ure (K)

on

vity

(%)

(%)

(h) (wt%)

capa

Pore

city

size

(me

(Å)

Reference

q/g) Purolite 5

413

12

100

75

4.1

-

[20]

5

413

10

100

87.9

0.52

-

[15]

8.6

413

5

100

69

5.0

300

[16]

CT269 P-SO3H

AM 35

Present AM 36

5

423

4

100

>99

5.4

240 study

3.2 Validation of the Presence of Isosorbide in the Product The final product from the SL dehydration carried out at the best condition was in dark brown due to the presence traceable amount of black-colour impurities like humins [26]. Following the GC results as shown in the supplementary document (S1), the corresponding purity of ISB in the reaction product was 96 wt%. The presence of high purity of ISB in the reaction product was further validated using 1H NMR spectroscopy. Figure 7 demonstrates that the 1H NMR spectra of both standard ISB and reaction product were identical, indicating that the reaction product was dominated by ISB. The peaks occurring between 3.50 and 4.50 ppm were assigned to the bicyclic methylene protons of ISB [27], whereas the peak at 4.71 ppm was ascribed to the deuterated water solvent [28]. It was observed that the spectrum of the reaction product was

negatively shifted due to the presence of impurities as detected through the peaks appeared at 3.6-3.8 ppm and 4-4.3 ppm.

3.3 Kinetic Modeling The SL conversion, ST and ISB yield data obtained from the reaction studies at various temperatures was used to identify the best fitted kinetic model of the sorbitol dehydration using the model discrimination approach. PH, LH1 and LH2 models were compared in the present study. Table 3 shows the kinetic parameters and their 95% confidence interval (CI). The kinetic data fit to the LH2 model resulted in a coefficient of determination R2 that closest to unity, implying that the adsorption of both SL and ST remarkably takes place during the dehydration of SL and ISB. The SL molecule is first adsorbed and protonation on the active sites via its hydroxyl groups at C1 and C6 positions. Then, the SL is dehydrated to form ST and water. The hydroxyl group at the C6 position of the

adsorbed ST molecule is further protonated and dehydrated to form ISB and water [6][21]. As commonly found, the k1 was higher than k2, justifying the presence of by-product of ST as the first dehydration step (SL to ST) proceeds faster in comparison to the second dehydration step (ST to ISB) [21]. Adsorption affinity is generally related to polarity and polar surface area of the compound in which adsorption increases with polarity but decreasing with the polar surface area. ST possesses a lower polar surface area than SL [29], causing a higher adsorption affinity of ST as indicated by the higher adsorption constant of ST as given in Table 3 [29]. Table 3: Kinetic parameters of the models used to fit experimental data for SL dehydration using AM 36 as catalyst Temperature

k1±CI

k2±CI

(K)

(L/mol.h)

(L/mol.h)

373

0.0469±0.0015 0.0077±0.0017

393

0.1230±0.0088 0.0107±0.0090

k3±CI (L/mol.h)

KSL±CI

KST±CI

(L/mol)

(L/mol)

413

0.4720±0.0487 0.1499±0.0955

423

0.8921±0.0635 0.2928±0.1643

373

0.0639±0.0099 0.0096±0.0020

0.0501±0.0294

393

0.2953±0.0661 0.0396±0.0085

0.1825±0.0733

413

1.1408±0.0730 0.2558±0.0667

0.1896±0.0235

423

2.2682±0.7749 0.658±0.1657

0.2041±0.1145

373

0.0680±0.0097 0.0373±0.0066 0.0045±0.0014 0.0591±0.0281 0.7053±0.122

393

0.2454±0.0311 0.1600±0.0210 0.0105±0.0037 0.1212±0.0100 0.6951±0.145

413

1.8141±0.0912 0.8909±0.1012 0.1048±0.0354 0.2712±0.0001 0.6760±0.131

423

4.0375±0.2389 2.4235±0.3518 0.1975±0.0536 0.4178±0.0001 0.6374±0.162

The parity plots compare the experimental and predicted concentration of the respective compounds at various reaction temperatures using the LH2 model are shown in Figure 8. The significant deviations of the predicted CSL and CST from the experimental data

initiated the simulation considering the side reaction involving dehydration of SL to humins. The equations for chemical reaction and the corresponding reaction rate are shown in Eq. 18 and 19 respectively. 𝑘3

𝑆𝐿 ℎ𝑢𝑚𝑖𝑛𝑠

18

𝑟𝑆𝐿 = ―𝑘1𝐶𝑆𝐿/(1 + 𝐾𝑆𝐿𝐶𝑆𝐿) ― 𝑘3𝐶𝑆𝐿 19

The rate constant, k3 as shown in Table 3 was determined using nonlinear regression function in Polymath 6.10. Figure 8 reveals that the deviation between the predicted and experimental concentration of especially the SL and ST compounds was significantly reduced when k3 was taken into account in the LH2 model. This implied the occurrence of side reaction that producing humins due to the oligomerisation and/or self-polymerisation [6][30]. Figure 9 demonstrates that the predicted concentration is in good agreement with the experimental data for a reaction at an arbitrary operating condition. The concentration of ISB after 3 h reaction was not accurately predicted because of its very low absolute value compared to

the other compounds. It was more susceptible to errors during experimental studies.

The activation energy (Ea) for the individual reactions was then determined using the Arrhenius plot as shown in Figure 10. The activation energies for SL dehydration to ST, ST dehydration to ISB and side reaction producing humins were 109.22, 109.46 and 104.17 kJ/mol, respectively. Typically, the Ea of uncatalysed ST dehydration is higher than the Ea of uncatalysed SL dehydration because of its more complicated reaction [18][31][32]. Relatively, AM 36 significantly brought down the Ea of ST dehydration for approximately 44%, benchmarking to the Ea of the uncatalysed reaction in Table 4. AM 36 catalysed ST dehydration showed higher Ea than the reaction catalysed by other catalysts reported in the literature due to the more superior properties of these catalysts, for instance, higher acidity of the silicotungstic acid catalyst and higher surface area of AM 35 catalyst [16][33][34]. Although ST dehydration

catalysed by AM 36 required higher Ea, the reaction performance of the present study was the best in terms of ISB yield in comparison to the reaction catalysed by other catalysts, ascribing to the effect of continuous water removal during the process.

Table 4: Kinetic studies for different catalyst of sorbitol dehydration Conditions

Ea (kJ/mol)

Stirri Catalys t

Performan Mod

ng Temperat

Pressur speed

ure (K)

Refere ce (ISB

el

e (bar)

k1

k2

k3

nce yield)

(RP M) 55% at 523No

na

100

PH

127

195

-

573 K

573 after 3 h

[21]

40% at Silico433tungsti

140. 70.6 1000

50

PH

473

473 K -

3

[34] after 6.7

c acid h 72. AM 35

393433

300-

PH

70% at

73.8 1

0.7-1

-

413 K for

[16]

600 LH

6.7 h

69.4 65.9

99% at AM

373-

109.2 109.4 104. 300

36

423

1

LH

Prese 423 K for

2

6

17

nt 4h

4. Conclusion The product distribution and kinetic behaviour of the sorbitol dehydration catalysed by AM 36 were investigated. AM 36 was found to surpass the other resin catalysts by producing ISB with a significantly higher yield. Under a continuous water removal condition, the highest SL

conversion, 100% and ISB yield, 99.8% were attained after 4 h in the reaction catalysed by 5 wt% of AM 36 at 423 K with the stirring speed of 300 RPM.

The kinetic data were best correlated with LH2 that

considering the side reaction. Disparate from the reported literature, the present study also found that ST adsorption played an important role in the reaction mechanism of ST dehydration. The activation energy was 109.22 kJ/mol, 109.46 and 104.17 kJ/mol for the respective SL dehydration to ST, ST dehydration to ISB and SL dehydration to other side products like humins. These reaction kinetics are crucial to providing an insight to control the distribution of SL, ST and ISB in a scaled-up process for the production of sorbitol anhydrides or other related products like sorbitol esters. Acknowledgement This research was supported by an internal grant RDU1803108, offered by Universiti Malaysia Pahang.

References [1] H. Kobayashi, H. Yokoyama, B. Feng, A. Fukuoka, Dehydration of sorbitol to isosorbide over H-beta zeolites with high Si/Al ratios, J. Green Chem. 17 (2015) 2732-2735. https://doi.org/ 10.1039/c5gc00319a. [2] X. Zhang, R. Abdallah, I. Mark, L. Adam Fraser, W. Karen, Sulfated zirconia catalysts for D-sorbitol cascade cyclodehydration to isosorbide: impact of zirconia phase, ACS sustain. Chem. Eng. 6 (2018) 1470414712. https://doi.org/10.1021/acssuschemeng.8b03268. [3] A.A. Dabbawala, S.M. Alhassan, D.K Mishra, J. Jegal, Solvent free cyclodehydration of sorbitol to isosorbide over mesoporous sulphated titania with enhanced catalytic performance, Mol. Catal. 454 (2018) 7786. https://doi.org/10.1016/j.mcat.2018.05.009 [4] M. Gu, D. Yu, H. Zhang, P. Sun, H. Huang, Metal (IV) phosphates as solid catalysts for selective dehydration of sorbitol to isosorbide, Catal. Lett. 133 (2009) 214-220. https://doi.org/10.1007/s10562-009-0142-5

[5] J. Li, A. Spina, J.A. Moulijn, M. Makkee, Sorbitol dehydration into isosorbide in a molten salt hydrate medium, Catal. Sci. Technol. 3 (2013) 1540-1546. https:// doi.org/10.1039/c3cy20809e [6] Y. Zhang, T. Chen, G. Zhang, G. Wang, H. Zhang, Efficient production of isosorbide from sorbitol dehydration over mesoporous carbon-based acid catalyst, Appl. Catal. A. Gen. 575 (2019) 38-47. https://doi.org/10.1016/j.apcata.2019.01.014 [7] P. Sun, D.H. Yu, Y. Hu, Z.C. Tang, J.J. Xia, H. Li, H. Huang, H3PW12O40/SiO2 for sorbitol dehydration to isosorbide: high efficient and reusable solid acid catalyst, Korean J. Chem. Eng. 28 (2011) 99-105. https://doi.org/10.1007/s11814-010-0324-2 [8] J. Shi, Y. Shan, Y. Tian, Y. Wan, Y. Zheng, Y. Feng, Hydrophilic sulfonic acid-funtionalized micro-bead silica for dehydration of sorbitol to

isosorbide,

RSC

Advan.

6

(2016)

13514-13521.

https://doi.org/10.1039/C5RA27510E [9] M. Yabushita, Acid-catalyzed dehydration of sorbitol to 1,4-sorbitan, in: A study on catalytic conversion of non-food biomass into chemicals,

Springer Theses (Recognizing Outstanding Ph.D. Research), Springer, Singapore, 2016, pp. 127-140 [10] A.A. Dabbawala, D.K. Mishra, G.W. Huber, J-S. Hwang, Role of acid sites and selectivity correlation in solvent free liquid phase dehydration of sorbitol to isosorbide, Appl. Catal. A. Gen. 492 (2015) 252-261. https://doi.org/10.1016/j.apcata.2014.12.014 [11] C. Dussenne, T. Delaunay, V. Wiatz, H. Wyart, I. Suisse, M. Sauthier, Synthesis of isosorbide: an overview of challenging reactions, Green Chem. 19 (2017) 5332-5344. https://doi.org/10.1039/c7gc01912b [12] A.A Dabbawala, J.J. Park, A.H. Valekar, D.K. Mishra, J-S. Hwang, Arenesulfonic acid functionalized ordered mesoporous silica as solid acid catalyst for solvent free dehydration of sorbitol to isosorbide, Catal. Commun.

69

(2015)

207-211.

https://doi.org/10.106/j.catcom.2015.06.017 [13] D.Yuan, N. Zhao, Y. Wang, K. Xuan, F. Li, Y. Pu, F. Wang, L. li, F, Xiao, Dehydration of sorbitol to isosorbide over hydrophobic polymer-

based solid acid, Appl. Catal. B Environ. 240 (2019) 182-192. https://doi.org/10.1016/j.apcatb.2018.08.036 [14] I.G. Fleche, I.M. Huchette, Isosorbide, preparation, properties and chemistry,

Starch.

38

(1986)

26-30.

https://doi.org/10.1002/star.19860380107 [15] J. Zhang, L. Wang, F. Liu, X. Meng, J. Mao, F-S. Xiao, Enhanced catalytic performance in dehydration of sorbitol to isosorbide over superhydrophobic mesoporous acid catalyst, Catal. Today 242 (2015) 249-254. https://doi.org/10.1016/j.cattod.2014.04.017 [16] I. Polaert, M.C. Felix, M. Fornasero, S. Marcotte, J-C, Buvat, L. Estel, A greener process for isosorbide production: kinetic study of the catalytic dehydration of pure sorbitol under microwave, Chem. Eng. J. 222 (2013) 228-239. https://doi.org/10.1016/j.cej.2013.02.043 [17] I. Ahmed, N.A. Khan, D.K. Mishra, J.S. Lee, J-S. Hwang, S.H. Jhung, Liquid-phase dehydration of sorbitol to isosorbide using sulphated titania as a solid acid catalyst, Chem. Eng. Sci. 93 (2013) 91-95. https:// doi.org/10.1016/j.ces.2013.01.068

[18] A. Cubo, J. Iglesias, G. Morales, J.A. Melero, J. Moreno, R. SanchezVazquez, Dehydration of sorbitol to isosorbide in melted phase with propyl-sulfonic

functionalized

SBA-15:

Influence

of

catalyst

hydrophobization, Appl. Catal. A Gen. 531 (2017) 151-160. https://doi.org/10.1016/j.apcata.2016.10.029 [19] R.S. Thombal, V.H. Jadhav, Facile O-glycosylation of glycals using Glu-Fe3O4-SO3H, a magnetic solid acid catalyst, RSC Advan. 6 (2016) 30846-30851. http://doi.org/10.1039/C6RA03305A [20] M.J. Gines-Molina, R. Moreno-Tost, J. Santamaria-Gonzalez, P. Maireles-Torres, Dehydration of sorbitol to isosorbide over sulfonic acid resins under solvent-free condiions, Appl. Catal. A. Gen. 537 (2017) 6673. https://doi.org/10.1016/j.apcata.2017.03.006 [21] A. Yamaguchi, N. Hiyoshi, O. Sato, M. Shirai, Sorbitol dehydration in high temperature liquid water, Green Chem. 13 (2011) 873-881. https://doi.org/10.1039/c0gc00426j [22] S.Y. Chin, C.S. Hong, M.R. Kamaruzaman, G.K. Chua, Chromatography strategies for the composition determination of bio-

based sorbitol azelaic acid ester, Polym. Test. 65 (2018) 270-280. https://doi.org/10.1016/j.polymertesting.2017.12.004 [23] D. Cao, B. Yu, S. Zhang, L. Cui, J. Zhang, W. Cai, Isosorbide production from sorbitol over porous zirconium phosphate catalyst, Appl. Catal.

A.

Gen.

528

(2016)

59-66.

https://

doi.org/10.1016/j.apcata.2016.09.017 [24] O.A. Rusu, W.F. Hoelderich, H. Wyart, M. Ibert, Metal phosphate catalysed dehydration of sorbitol under hydrothermal conditions, Appl. Catal.

B.

Environ.

176-177

(2015)

139-149.

https://doi.org/10.1016/j.apcatb.2015.03.033 [25] J. Zou, D. Cao, W. Tao, S. Zhang, L. Cui, F. Zeng, W. Cai, Sorbitol dehydration into isosorbide over a cellulose-derived solid acid catalyst, RSC

Advan.

6

(2016)

49528-49536.

https://doi.org/10.1039/C6RA05214B [26] F.J. Stevenson, Humus Chemistry: Genesis, Composition, Reactions, second ed., Wiley, New York, 1994

[27] H-J. Kim, M-S. Kang, J.C. Knowles, M-S. Gong, Synthesis of highly elastic biocompatible polyurethane based on bio-based isosorbide and poly(tetramethylene glycol) and their properties, J. Biomater. Appl. 9 (2014) 454-464. https://doi.org/10.1177/0885328214533737 [28] N.R. Babij, E.O. McCusker, G.T. Whiteker, B. Canturk, N. Choy, L.C. Creemer, C.V. De Amicis, N.M. Hewlet, P.L. Johnson, J.A. Knolbelsdorf, F. Li, B.A. Lorsbach, B.M. Nugent, S.J. Ryan, M.R. Smith, Q. Yang, NMR chemical shifts of trace impurities: Industrially preferred solvents used in process and green chemistry, Org. Process Res. Dev. 20 (2016) 661-667. https://doi.org/10.1021/acs.orprd.5b00417 [29] S. Lux, T. Winkler, G. Berger, M. Siebenhofer, Kinetic study of the heterogeneous catalytic esterification of acetic acid with methanol using Amberlyst 15, Chem. Biochem. Eng. Q. 29 (2015) 549-557. https://doi.org/10.15255/CABEQ.2014.2083 [30] W. Ni, D. Li, X.Zhao, W. Ma, K. Kong, Q. Gu, M. Chen, Z. Hou, Catalytic dehydration of sorbitol and fructose by acid-modified zirconium

phosphate,

Catal.Today

319

(2019)

66-75.

https://doi.org/10.1016/j.cattod.2018.03.034 [31] M. Yabushita, H. Kobayashi, A. Shrotri, K. Hara, S. Ito, A. Fukuoka, Sulfuric acid-catalyzed dehydration of sorbitol: mechanistic study on preferential formation of 1,4-sorbitan, Bull. Chem. Soc. Jpn. 88 (2015) 996-1002. https://doi.org/10.1246/bcsj.20150080. [32] J. li, W. Buijs, R.J. Berger, J.A. Moulijn, M. Makkee, Sorbitol dehydration in a ZnCl2 molten salt hydrate medium: molecular modeling, Catal.

Sci.

Technol,

4

(2014)

152-163.

https://doi.org/10.1039/c3cy00686g [33] N.A.Khan, D.K. Mishra, J-S. Hwang, Y-W. Kwak, S.H. Jhung, Liquid-phase dehydration of sorbitol under microwave irradiation in the presence of acidic resin catalysts. Res. Chem. Intermed. 37 (2011) 12311238. https://doi.org/10.1007/s11164-011-0389-5 [34] J.U. Oltmanns, S. Palkovits, R. Palkovits, Kinetic investigation of sorbitol and xylitol dehydration catalysed by silicotungstic acid in water,

Appl.

Catal.

A.

Gen.

456

(2013)

168-173.

https://doi.org/10.1016/j.apcata.2013.02.023

Highlights:  Sorbitol dehydration catalysed by Amberlyst 36 was carried out at mild condition.  The continuous water removal system accelerated the attainment of 100% isosorbide yield within 4 h.  The sorbitol dehydration was well represented by LangmuirHinshelwood Kinetic Model.  Both the adsorption of sorbitol and sorbitan played important role in the reaction mechanism.

Declaration of interests ☒ The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. ☐The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

Figure 1: The possible sorbitol dehydration pathway [11].

Figure 2: Experimental setup for SL dehydration.

Figure 3: (a) SL conversion (b) ST Yield and (c) ISB Yield profiles for SL dehydration at different stirring speeds. The reactions were carried out at 413 K, for 4 h with 1 wt% of AM 36.

Figure 4: Initial reaction rate for different stirring speed for SL dehydration using 1 wt% of AM 36 at 413 K for 4 h.

Figure 5: (a) SL conversion (b) ST yield and (c) ISB yield profiles for SL dehydration at different catalyst loading. The reactions were carried out at 413 K for 4 h with stirring speed of 300 RPM.

Figure 6: (a) SL conversion (b) ST yield and (c) ISB yield profiles for SL dehydration at different temperatures. The reactions were carried out at stirring speed of 300 RPM for 4 h with 5 wt% of AM 36.

Figure 7: The 1H NMR spectrum of (a) Standard ISB and (b) Product withdrawn from the SL dehydration reaction catalysed by 5 wt% of AM 36 for 5 h at 423 K with stirring speed of 300 RPM.

Figure 8: Parity plot of experimental and predicted concentration of SL, ST and ISB using LH2 model (Figure (a), (c) and(e)) and LH2 model with side reaction (Figure (b), (d) and (f)). (○) 373 K, (□) 393 K, (◊) 413 K, (Δ) 423 K and (----) ±10%.

Figure 9: Comparison between experimental and predicted concentration profile of SL, ST and ISB using LH2 model with side reaction. Reaction conditions of 300 RPM, AM 36 of 5 wt% and temperature of 423 K.

Figure 10: Arrhenius plot for sorbitol dehydration using AM 36 as catalyst at different reaction temperature.

s