Process optimization with acid functionalised activated carbon derived from corncob for production of 4-hydroxymethyl-2,2-dimethyl-1,3-dioxolane and 5-hydroxy-2,2-dimethyl-1,3-dioxane

In this article, a two-step activated carbon preparation technique from corncob has been elucidated. The derived catalysts AAC-CC has been characterized using various techniques for the determination of their structural properties and compared with AC-CC, already reported with another article. The conjugated boat structure of AAC-CC resulted in a very high surface area (779.8 m2/g) and high pore volume (0.428 cc/g). This unveils the suitability of AAC-CC as better among the two catalytic pathways for solketal production. The activated carbons so prepared have been used for the valorization of glycerol to produce 2,2-Dimethyl-1,3-dioxolane-4-methanol (solketal), oxygenated additives to fuel. The face-centered composite design (FCCD) of RSM was applied for the optimization of the reaction parameters for the ketalisation reaction using AAC-CC as a catalyst. From the optimized results, the acidic catalyst AAC-CC resulted in a glycerol conversion, i.e. 80.3% under the actual laboratory experiment. Moreover, the catalyst could be reused for three consecutive batch reactions without (< 5%) much reduction of activity and no distinctive structural deformity.

glycerol reforming using various catalysts derived from waste biomass materials 9,10 . One plausible pathway that the glycerol can be converted to five and six-membered oxygen-containing functionalized cyclic compounds reported by several authors is the production of 2,2-Dimethyl-1,3-dioxolane-4-methanol (solketal) via ketalisation. The chemical reaction for the production of two ketal species is shown in Fig. 1 11 . Solketal is a ring-shaped di-ether that has an additional hydroxyl group 12 . It can be used as an additive in gasoline, diesel, or biodiesel to increase ignitability and reduce particulate emission 12 .
There are various commercially developed catalysts reported in the literature having different physicochemical properties responsible for the conversion of glycerol to solketal. The nature of these catalysts may be either acidic or alkaline and the reactions may be either homogeneously or heterogeneously catalyzed 13 . Oprescu et al. reported the green solid superacid catalyst SO 4 2− /SnO 2 for the ketalisaton of glycerol to produce solketal with a yield of 97.5% 14 . Rodrigues et al. reported the activated carbons produced from olive stone functionalized with acid groups for solketal production with 97% conversion of glycerol 15 . The Nb-SBA-15 (Nb metal modified SBA-15 catalyst) gives 95% glycerol conversion as reported by Ammaji et al. 16 . Fe(NO 3 ) 3 ·9H 2 O exhibited the highest catalytic activity, virtually converting all the glycerol to solketal, with 95% selectivity as reported by Silva et al. 17 . The OTS-grafted HY (organosilane modified HY) catalyst reported having high catalytic activity (89% conversion) at low temperatures as stated by Rahaman et al. 18 . Shirani et al. reported a heterogeneous resin catalyst, i.e., Purolite PD206 for glycerol conversion to solketal with a 95% yield 19 . A mesoporous phenol sulfonic acid-formaldehyde polymeric acid catalyst was synthesized that provides glycerol conversion of 97% as reported by Laskar et al 20 . BEA zeolite and the hierarchical zeolite of MFI structure exhibited high catalytic activity in solketal production, i.e. almost 85% of glycerol conversion and 98% selectivity to solketal were achieved as reported by Kowalska-Kuś et al 21 .
Various acidic or basic homogeneously or heterogeneously catalyzed conversion processes amongst the reported literature unveil that basic catalysts are considered less active as they provide a lower conversion as compared to the acidic catalysts. As per the literature, the basic oxide catalysts partially leached during the reaction which hinders the separation of the catalyst from the reaction mixture 22 . Also, base catalysts can't give high conversion at lower temperature and for a shorter reaction time that led to the formation of salts 23 . Thus, the acidic catalysts show the highest catalytic activity as compared to the basic one and lead to better conversion due to the availability of the more active sites to the catalyst that will help in the reaction for higher conversion.
Physical activation, chemical activation, physical-chemical and microwave-assisted activation are the common activation methods for the preparation of catalysts 24 . Physical activation consists of heat and steam, chemical activation involves activation using chemical agents, physical-chemical activation involves both physical and chemical activation and microwave-assisted activation is done by microwave radiation 25 .
It has been observed from the various reported literature that very few renewable precursors-based catalysts are developed 26 . Among all the catalysts investigated, acid and base activated carbons derived from agricultural waste, i.e. corncob has not been reported yet for the production of solketal. Keeping in view the introduction of cleaner production technology for solketal production using renewable precursor-based catalysts, a locally available, merely a waste lignocellulosic residue corncob was studied for the production of acid and base activated  Although it is challenging to use biomass in energy applications due to the NO x emissions during thermal processing of biomass via pyrolysis, there are several ways available for their migration. Some of the possible ways for their migration are chemical reduction of NO x , NO x oxidation, using a sorbent in the combustion chamber etc. 28 . Thus, using waste biomass as renewable precursors proves to be a cost-effective method.
Complete physicochemical characterization of the prepared acid-activated carbon (H 2 SO 4 ) (AAC-CC) has been accomplished, presented and discussed in this article. Process optimization under different sets of reaction conditions using response surface methodology (RSM) for ketalisation reaction using AAC-CC has also not been reported elsewhere except this article along with optimal production of 4-hydroxymethyl-2,2-dimethyl-1,3-dioxolane and 5-hydroxy-2,2-dimethyl-1,3-dioxane using virtual laboratory setup.

Results
Characterization of catalyst using various physical and adsorption techniques. The plausible structure of the derived catalyst is shown in supplementary section II (SS-II).
The surface groups present in the carbon material produced were characterized and identified using different techniques before their utilization for the ketalization reaction. The presence of different functional groups in the structure was identified from an extensive study of FTIR and XPS analysis. The XRD patterns of the catalyst showed the amorphous nature of the carbon material.
The microstructural evaluation was done using Scanning Electron Microscopy (SEM) by doing the sputter coating of the sample before the SEM analysis to make the sample conducting. This step involves the coating with a conductive material like gold for obtaining better quality images. SEM micrographs of the derived functionalised activated carbon have been shown in Fig. 2. It can be observed from the picture that AAC-CC has developed a conjugated boat-like structure having a rough or uneven surface with cavities and cracks in the external surface and random pore size distribution 29 . The elemental analysis was performed using energy-dispersive (SEM-EDX) and the elemental composition of the sample showed the presence of Carbon, Oxygen, Sulphur in AAC-CC. Carbon is present in the highest amount i.e. 90.99% in AAC-CC responsible for the higher glycerol conversion 24 . Thus, carbonization is confirmed in sample 30 (Table 2) with minimum carbon loss in AAC-CC. The sputter coating does not affect the elemental composition of the sample as the acid-activated carbon does not contain gold in it.
The TEM micrographs of the AAC-CC are shown in Fig. 3a. The TEM results are in good agreement with SEM results showing the porosity of the produced catalysts. The images revealed a complete porous carbon structure for both the activated carbons even at higher magnification. The existence of the pore structure of the catalyst is attributed to the NaOH activation adopted during the synthesis of the AC-CC catalyst 31,32 (Fig. 3b) and H 2 SO 4 activation AAC-CC as shown in the images 33 . The sizes and uniformity envisage that the glycerol molecule with a size of 0.5-0.6 nm can easily fit and escape from these pores where the active functional groups are present that facilitates the group transfer and cyclization reactions of glycerol.  www.nature.com/scientificreports/ FT-IR spectra show a broad peak at around 3426.50 cm −1 in AAC-CC ( Fig. 4) for stretching of -OH bond of alcohol. The 2924.72 cm −1 peak shows the aliphatic C-H stretching vibration in AAC-CC. The peak at 2856.82 cm −1 shows the vibration of the -OH group attached to ring carbon in the derived catalyst. There is a broad peak at 1694.28 cm −1 which shows the stretching of the C=C due to the influenced functionalities. The sharp band at 1595.93 cm −1 corresponds to a stretch probably from an amide in the acid-activated catalyst 35 . The peaks at 1457.57 cm −1 , 1381.38 cm −1 , and 1084.04 cm −1 show the C-C stretching vibration of the chain hydrocarbon parts due to different structural influences, C-H stretching vibration in alkanes or an alkyl group and C-S group, respectively in base-activated catalyst 34 . The peak at 779.62 cm −1 shows the absorption of the SiO 2 and the 548 cm −1 peak shows the C-H stretching of aromatic compounds 34,36 . The additional bands at 1171.17 cm −1 and 1123.8 cm −1 show the stretching of sulfonated groups in acid activated carbon 26 . The functional groups represent the chemically active components of the catalyst that accelerates the rate of reaction.
The values of specific surface area calculated using the BET equation are presented in Table 3 together with the values of the total pore volume, Vp, taken at P/Po = 0.99, and average pore radius pore size distribution curves for activated carbon samples have been obtained from BJH calculation method 37 . The significance of the bold in Table 3 is that it refers the results of the present study while all other results are reported from literature.
The adsorption-desorption isotherm of AAC-CC is classified as type IV and reveals a hysteresis loop of H4 type implies to narrow slit pores including the pores in the micropore region as shown in Fig. 5 38 . H4 loops are often found with micro-mesoporous carbons.
A very high surface area and higher pore volume of AAC-CC unveil its suitability as the better of the two catalytic pathways during the solketal production as shown in Table 3.    Figure 5. For acid activated carbon (a) N 2 adsorption-desorption isotherm (b) Pore size distribution.  34 . In AAC-CC, the weight loss from 200-300 °C corresponds to the decomposition of the sulfonated groups. The weight loss from 300-400 °C is due to the loss of cellulosic materials. The prominent desorption occurs between 400-481 °C 44 . The weight loss at 600-700 °C shows the decomposition of the produced carbon and the weight loss at 800-900 °C shows the decomposition of the organic components. After 600 °C, the fusion of the carbon occurs till 800 °C and 800-900 °C is the ash fusion temperature.
Although Boehm titration (BT) is also a method to determine quantitatively the oxygen-containing groups in activated carbons, simple and cost-effective, it is time consuming and variations while opting measuring protocol that influences the results of analysis 45 . Though TPD and BT provide similar trends in results, a number of groups determined by TPD is higher than those obtained by BT. BT cannot determine the groups that can form on the surface such as ketones, ethers, aldehydes, and pyrones, groups that contain nitrogen, phosphorus, or sulfur 46 . Thus, the acidic sites of the catalysts are determined by NH 3 -TPD as these are the active sites responsible for the conversion of glycerol to solketal at a modest temperature 47 . The acidic strength of solid acid catalysts in the NH 3 -TPD profiles can be classified into three regions depending on their strength in the temperature range. These acidic sites are denoted as weak (150-300 °C), moderate (300-450 °C), and strong (450-650 °C) 16 . The desorption peak of NH 3 -TPD for AAC-CC shows a weak acidic site at 147.1 °C, while showing a sharp peak at 605 °C with higher intensity having strong acidic strength (Fig. 7). The total acidic amount of 0.187 mmol/g was obtained for the acidic site.
The characteristic peaks and their confirmation showing the presence of carbon, oxygen and sulphur groups attached with the AAC-CC and the XPS spectra showing the surface chemistry are shown in Fig. 8a,b,c,d. XPS detects the composition of the species on the surface and their binding energies, along with the oxidation states 39 . The main C1s peak of AAC-CC was observed at the binding energy of 284.5 eV belongs to C-O and one at 288.4 eV belongs to C-C bonding. The signals at 531.95 eV of O1s are associated with oxygen-carbon bonds (C=O). The XPS analysis confirms the presence of -SO 3 H groups, based on a Sulphur peak at ~168.11 eV. The results of characterisation and the elemental analysis obtained by XPS of the groups present in AAC-CC are listed in Table 4.
Finally, a comparison of the physicochemical characteristics of both the catalysts viz. AC-CC and AAC-CC has been put together in Table 5. It is clearly established from the table that AAC-CC has better catalytic properties than AC-CC.

Discussions
Catalyst Activation and functionalization. The carbon percentage in AC-CC was 78%, whereas in AAC-CC it is 91%. Moreover, the sodium, oxygen and silica content are drastically reduced in the AAC-CC. This is confirmed from the EDX and XPS data that the percentage of C increases by 13% after acid functionalization of the NaOH activated carbon. Moreover, the Rib shaped AC-CC has been converted to a conjugate boat structure in AAC-CC, resulting in high porosity and 50-60-fold increased surface area. This may be attributed to the basic reactions for the removal of silica and Na bonded with AC-CC in step -II production of activated carbon as shown in Fig. 9.
Effect of reaction parameters on solketal production. In the present work, Response Surface Methodology (RSM) was employed to optimize the values of process parameters for the ketalization reaction to get the improved glycerol conversion with derived catalysts by correlating the experimental and predicted values; numerical optimization method and analysis of standard error in the obtained values; quadratic regression in the model through desirability function as 0.999 for the precision of the results. Also, possible interactions between process factors were determined for their overall influence on the progress of the process. After completion of optimization of the process via RSM, the obtained 30 sets of experiments were conducted in actual laboratory   Table 2 in the results and discussions in the supplementary information (S-II) in the form of observed values of RSY against the coded values. The effect of reaction parameters (molar ratio, time, temperature and catalyst amount) on RSY were also studied by plotting three-dimensional (3-D) surface curves shown in S-II, Fig. 6. The conversion of glycerol via ketalisation reaction was observed to be influenced by parameters within the range: molar ratio (1:2:2 to 1:8:8), time (1 to 3 h), temperature (50 to 150 °C) and catalyst amount (1 to 5 wt.%). The highest value of RSY was found to be 80.3% against the predicted value of 73.807% for solketal production corresponding to the parameters with molar ratio-1:8:8, for 2 h at 100 °C and 3wt.% catalyst loading. It was observed that a further increase in the reaction temperature associated with lower glycerol conversion because high temperature is associated with a higher rate of dissociation and lower stability of the product and was confirmed in Table 2 in S-II. Also, solketal production is an exothermic reaction and an increase in temperature reduces the product formation. From the shape of the 3-D surface plots, it was concluded that the effect of molar ratio on glycerol conversion dominates when compared with the effect of other variables like conversion increases with an increase in molar ratio but time has not shown a significant change in conversion as shown in Fig. 6a, S-II. These results have been found in concordance with the optimized results obtained by RSM employing Analysis of Variance (ANOVA).
The catalytic activity was also studied for the ketalization reaction i.e. conversion of glycerol to solketal using the derived functionalised catalyst obtained from corn cob. To check the stability of the catalyst and effectiveness of the active sites of the catalyst, a reusability study was carried out 34 . The catalyst was separated from the reaction mixture after the completion of the reaction by filtering the reaction mixture using filter paper. It was then washed with ethanol 3-4 times and then dried in an oven at a temperature of 90 °C. The study showed that the catalyst can be reused up to three consecutive batch reactions without much loss of the activity (< 5%) and no distinctive structural deformity. This shows the utility of the derived catalyst.

Conclusions
Conversion of glycerol to solketal using acid-functionalized activated corncob-based carbon as a catalyst is a greener pathway and possible. The acid-activated carbon has a higher surface area, pore-volume, and strong acidic sites that catalyze the reaction with a higher conversion rate as compared to alkali-activated carbon. The molar ratio of the reactants (glycerol: acetone: methanol) has a significant effect on the ketalization reaction. The derived activated carbon can be used for hydrogen storage and electrode fabrication material for the fuel cells. It can be reused to produce various valuable chemicals from glycerol and a prototype can be developed in near future for cleaner production of solketal from crude glycerol.

Experimental section
The chemicals used were of analytical grade. Glycerol (Sigma Aldrich-97% purity), Methanol (Sigma Aldrich-99.5% purity), Acetone (Sigma Aldrich-99.5% purity) were procured from the Merck Specialists Limited, Mumbai in the chemical conversion division of Sardar Swaran Singh National Institute of Bio-energy, Kapurthala, India. The corn cob used for the preparation of the catalyst was collected from the local mandi of the Kapurthala, Punjab, India.
Catalyst preparation. The activated carbon was prepared using two-step activation methods by following the protocol as described below: Step I: The received corncob was dried in the open air for 2-3 days to remove the excess moisture. The dried corncob was milled into small pieces of 75 mesh sizes. The ground corncob was calcined in the muffle furnace using the prepared crucible to provide the inert atmosphere during calcination. The calcination temperature was kept at 500 °C for 1-2 h. The calcined biomass was activated using NaOH as an activating agent with a molar ratio of 1:3 (Biomass: NaOH). After preliminary investigation and experimentation with the different impregnation ratio, 1:3 molar ratio (Biomass:NaOH) was adapted in this process as a higher ratio of activating agent develops AC with higher pore volume until optimal range 48,49 . The activated carbon produced was again heated in a muffle furnace at a temperature of around 450-500 °C. After heating the activated carbon, a suspension was obtained which was then filtered using Whatman filter paper while washing it continuously with distilled water to remove the excess alkali present in it.
Step-II: The filtered activated carbon was then dispersed in a beaker of 1000 ml containing concentrated sulphuric acid (2 M) solution. The activated carbon remains dipped in the acidic solution for the proper functionalisation of the catalyst for 8-10 h with proper stirring of the solution at an RPM 200. It was then filtered using Whatman filter paper. The collected functionalised carbon was then heated in an oven at 80-100 °C for 8-10 h for removal of the moisture. It was then stored in an airtight bottle which was used as an acid-activated carbon catalyst (labelled as AAC-CC) for further characterization and ketalization reactions. The flask was then kept in a beaker of 1000 ml containing water which acted as a water bath. This beaker was then placed on a magnetic plate with a stirrer and temperature controller to control the temperature and uniform mixing of the reaction mixture during the reaction. Two necks of the flask were blocked by a flask glass stopper and in one reflux and condenser were also attached to the flask with the help of the burette stand, to maintain the uniformity of the reaction mixture as methanol and acetone were volatile and may evaporate. The reaction was preceded at 600 revolutions per minute.
The reactions set were performed as per the process optimization described below.
Process optimization. The ketalization reaction of the glycerol using AAC-CC was optimized using face centred composite design (FCCD) of RSM, Design-Expert software version 8.0 (STAT-EASE Inc., Minneapolis, USA) in terms of process parameters, the molar ratio (X1); time (X2); temperature (X3); catalyst amount (X4) to provide us with the maximum glycerol conversion (RSY) as a response (Y) 50 . Software provided 30 no. of reactions consisting of 24 non-central and 6 central axial points (α = 1). Central (axial) points were defined for the authenticity of the model via pure error variance. After completion of each reaction, the samples were collected in microcentrifuge tubes and then centrifuged using a 5430 R centrifuge to remove the traces of catalyst present in the sample. The centrifuged samples were then diluted 100 times and were analyzed using high-performance liquid chromatography (HPLC-Agilent Technologies model 1260 Infinity and glycerol conversion was calculated using equation (Eq. 1), from the graph obtained from HPLC. FTIR experiments were conducted to identify the functional groups. These were interpreted using a PE IR SUBTECH SPECTRUM ASCII PEDS 4.00 infrared spectrometer with a resolution of 1 cm −1 . The materials were mixed with KBr powder pelletized and the pellets were scanned in the IR range from 4000 to 400 cm −1 .
The carbonization behaviour of the AC-CC was determined by a simultaneous TG/DTA i.e. DTA with the proven capabilities of the TG measurement capabilities, providing thermal property information for a variety of samples. This was done using an SII 6300 EXSTAR thermal analyzer. Samples were heated in the temperature range from 35 to 900 °C at a constant heating rate of 5 °C/min in nitrogen with a 200 ml/min flow rate.
Nitrogen adsorption-desorption measurement was carried out using a BET surface area analyzer by St 2 on the NOVA touch 2LX instrument. The samples were degassed for 3 h at 180 °C on the degassed port. The linear part of the BET equation was used to determine the specific surface area.
The wide-angle X-ray diffraction pattern of the AC-CC was observed on Bruker XRD diffractometer using Cu-Kα radiation with a wavelength of 1.54 and F as a filter. The scanning angle (2θ) range was kept between 10° to 80°.
Scanning electron micrographs (SEM) and elemental analysis of the activated carbon were performed using Nova Nano FE-SEM 450 (FEI) coupled with an EDX analyzer of ultra-high-resolution characterization in an accelerating voltage of 15.0 kV.
Transmission Electron Microscopy (HRTEM) studies were done with a TEM TECNAI G2 20 S-TWIN (FEI) electron microscope. A drop of the powdered sample was dispersed into ethanol and then dropped onto a carboncoated copper grid. The high resolution of the analyzer was provided with a voltage of 200 kV.
The presence of surface groups with Carbon, Sulphur, Oxygen, Silicon, and Sodium was identified and confirmed by X-ray photoelectron spectroscopy (XPS) using PHI 5000 Versa Probe III model.
The acidity of the catalyst surfaces was studied using Temperature program desorption (TPD) and performed using BEL's new fully-automated catalyst analyzer (BELCAT II) with NH 3 as investigating molecule and He as a carrier gas. The desorption of NH 3 was performed after flushing using carrier gas up to temperature 600 °C at a heating rate of 10 °C/min.