Supported catalytic packing prepared from ceramic packing for reactive distillation of ethyl acetate

Zhiwei Wang,Jiannan Shi,Xiaonan Liu,Zhikai Cao,Yong Sha*

Department of Chemical and Biochemical Engineering,College of Chemistry and Chemical Engineering,Xiamen University,Xiamen 361005,China

Keywords:Supported catalytic packing Ethyl acetate Reactive distillation Bonding strength

ABSTRACT In this work,a strategy of‘‘etching-modification filling-graft copolymerization”was proposed to load the acidic ionic polyionic liquid on the smooth ceramic surface.In this way,commercial ceramic Raschig rings were successfully transformed into the supported catalytic packing for the reactive distillation,and were further evaluated with esterification reaction of ethyl acetate by means of the fully mixed reactor,the ultrasonic destruction,the cyclic catalysis reaction and the lab-scale distillation column experiment.This catalyst coating has good adhesion with the substrate.It can withstand 24 h of ultrasound damage and shows good stability in three cycle catalytic experiments.This kind of coated catalyst has better catalytic activity than the commercial Amberlyst 15 dry.In the lab-scale reaction distillation,the supported catalyst Raschig ring can achieve a higher conversion in comparison with the tea bag catalytic packing of Amberlyst 15 dry under some conditions.

1.Introduction

An important research area of the reactive distillation (RD)technology is the improvement and optimization of the catalytic packing [1].In recent years,with the demand of industry,various forms of catalytic packings have been developed and investigated.Currently,the main approach of the catalytic packing is to combine catalysts and packings through stationary elements.There are two kinds of classical fixing methods of the catalyst in the RD column[2].One is in the wire mesh bag such as the tea bag[3],and another is in the inter-layer of the structured packing such as the sandwich packing [4].These methods can flexibly adjust the amount of the catalyst and the proportion of separation units,and are widely utilized in the industrial and theoretical RD research.

Different from the catalyst encapsulation described above,catalyst in the RD can be fixed or loaded directly on the surface of the mass transfer element through physical bonding or chemical bonding,or directly built to be a mass transfer element[2].This kind of packing is also called the supported catalytic packing (SCP).The characteristic of the SCP is that the catalyst is directly exposed to the bulk of the gas and liquid phase.In comparison with the encapsulated catalyst,there is no obstruction and redistribution of flow and diffusion by fixed elements such as the wire mesh and packing[5],but at present the SCP has no industrial application due to limitation of the catalyst activity and load amount,the bonding strength between the catalyst and the substrate as well as mechanical properties of the catalyst itself.According to results from literatures about it,it has a high efficiency factor and better mass transfer performance [6,7].It can achieve considerable conversion for some RD systems [8-10],and its special surface flow pattern also has high theoretical research significance [11,12].Therefore,it has a potential application value,and is necessary to carry out further research.

Nowadays,the RD catalyst mainly adopts strong acid ion exchange resin,so the early research focused on how to load it on the packing surface [12].The simplest idea is to build it into a mass transfer element with a specific shape [13].Due to the volume change caused by the ion exchange,it can cause a big internal expansion force,and the packing prepared in this way is very easy to deform or even crack [14,15].The internal expansion force is directly related to the material thickness,and reduction of the thickness can reduce the expansion force.Loading a thin layer of catalyst on the surface of the column packing can avoid this problem.However,a thin layer has a high requirement for catalytic activity due to the less catalyst mass.Meanwhile,for the acid ion exchange resin catalyst,it is difficult to avoid its falling off caused by swelling and meet the requirement of the large-scale production [15].For non-swelling acidic catalysts,such as molecular sieves [16-20] and metal oxides [21,22],they have been successfully loaded and applied in the lab scale RD experiments in recent years.However,generally they are suitable for medium and high temperature reaction systems due to the active temperature.For the RD system with the low and medium temperature,there is few of feasible loading method of ion exchange catalysts.

Sundmacher [12] proposed a method of loading polystyrene sulfonic acid on the macroporous glass Raschig ring.This kind of Raschig ring catalyst showed its advantages in its strength due to macroporous geometric limitation and cross-linking between polystyrene sulfonic acid and the surface.The efficiency factor of this SCP was close to 1 in the MTBE RD system.Luet al.[9] utilized cross-linking of Nafion and Tetraethyl orthosilicate to load Nafion-SiO2catalyst on the surface of the three-dimensional stainless steel felt.This supported catalyst was evaluated in the RD system of ethyl acetate,and its loading strength was very considerable due to the usage of the inorganic adhesive and the threedimensional carrier.Although above two packing materials were just studied in the lab,they showed the possibility and importance of the geometric structure of the support substrate for loading of the swell-able catalyst.It may be a feasible way to load the catalyst by means of the structure constraint in the substrate.

Here,by means of artificial creation of macroporous defects by etching and surface modification on the surface of commercial ceramic Raschig rings,a kind of SCP for the RD was prepared by graft copolymerization of acid polyionic liquid,and it was evaluated by the lab-scale RD experiment of ethyl acetate in comparison with the conventional tea bag catalytic packing.

2.Experimental

2.1.Materials

(3-Mercaptopropyl) trimethoxysilane,Dioctyl phthalate (DOP),sulphuric acid,sodium hydroxide,ethanol,azodiisobutyronitrile(AIBN),ethyl glycol dimethyl methacrylate (EGDMA) were supplied from Shanghai Aladdin Biochemical Technology Co.,Ltd.(China).Acetic acid,ethyl acetate,1,3-propanesultone and 1-ethenylimidazole were provided by Shanghai linen Technology Development Co.,Ltd.(China).All chemical reagents are analytically pure.The ion exchange resin catalyst Amberlyst 15 dry,Amberlyst 35 wet and Amberlyst-131 was bought from Rohm and Haas (USA),and packed and sealed by the wire mesh as the teabag catalytic packing (about 35 mm × 15 mm × 5 mm,weight of single package catalyst (0.72 ± 0.1) g,the mesh aperture is 150 μm).The stainless steel cannon packing (4 mm × 4 mm) and ceramic Raschig ring (3 mm × 3 mm) was bought from Jiangxi PX Rongjian Environmental Protection Chemical Co.,Ltd.(China).The composition of the ceramic Raschig ring provided by the supplier is shown in Table 1.

Table 1 Composition table of ceramic Raschig ring provided by the supplier

2.2.Preparation of SCP

The preparation process is shown schematically in Fig.1 and can be divided into the following steps.

Pretreatment of packing:After cleaning commercially purchased ceramic Raschig rings with ethanol,they were ultrasonicated at 80 °C for 6 h in 10% (mass) sodium hydroxide solution to etch the surface,then washed with water to neutral,and dried at 120 °C overnight.

1 g mercaptosilane was added to 100 ml of the ethanol aqueous solution with the volume ratio of 3:1 and was ultrasonicated for 30 min.Then,30 g of dry etched ceramic Raschig rings were mixed with the solution and ultrasonicated for 3 h at 70°C.After washing them with ethanol,modified Raschig rings were dried in a vacuum at 70 °C for 12 h.Raschig rings change color from light gray to white.This step mainly uses the hydrolysis of silane to fill gaps etched on the surface.

Preparation of catalytic packing: Acidic polyionic liquids were grafted and copolymerized on the ceramic surface by means of silicon balls produced on the surface in the previous step.20 g of modified Raschig rings,1.5 g of vinyl imidazole,1.5 g of ethylene glycol methacrylate and 0.1 g of AIBN were mixed under a N2protective atmosphere,and were ultrasonicated under a negative pressure for 1 h.Then,60 g acetonitrile and 6 g DOP were added and oscillated at 60°C for 8 h.The uncrosslinked powder and porogen DOP were cleaned with ethanol ultrasonic,and the canary yellow coated Raschig rings were dried in a vacuum at 70 °C overnight.

20 g dried Raschig ring,40 ml toluene and 1.8 g 1,3-propanesultone were added into the conical flask and oscillated at 70 °C for 24 h.Then 15 g sulfuric acid was added and oscillated at 40 °C for 12 h.It was washed in the ethanol aqueous solution until neutralization,and dried in a vacuum at 70 °C overnight.Finally,the light brown yellow catalytic packings were obtained.The active component of the coating is polyionic liquid,which is recorded as P[SO3H-(CH2)3-VIM][HSO4].

2.3.Characterization

The reaction system contains four components.Water concentration was analyzed by the Karl Fischer titration method.Ethyl acetate,ethanol and acetic acid were analyzed by the gas chromatography (GC9160,Xiamen Jingkejie Intelligent Equipment Co.,Ltd.,China) with an FID detector and a capillary column (OV-17 with a length of 30 m).The injector and detector temperatures were set to 100 and 230°C,respectively.The GC column temperature was firstly set to 100 °C for 2 min,then increased at a fixed rate of 30 °C·min-1to 130 °C for 1 min.

Further,scanning electron microscopy (SEM)was performed to observe the morphology of the sample using a Hitachi SU4800(Tapan) scanning electron microscope.Attenuated total reflection infrared spectroscopy (ATR-IR) spectra was recorded by using a Thermo Scientific Nicolet(USA)iS50 FT-IR spectrophotometer with 450-4000 cm-1.Nitrogen adsorption-desorption isotherms were measured on a nitrogen adsorption apparatus(Micromeritics ASAP 2460,USA) at 77 K.The ion exchange capacity was characterized by acid-base titration.Before titration,the packing was exchanged by 1 mol·L-1sodium chloride solution for 24 h.

2.4.Catalyst performance evaluation

Catalyst activity test:The comparison of catalytic performance was carried out through a fully mixed reactor.A 250 ml threeneck glass flask fitted with a total reflux condenser was used as a reactor.In the experiment,1 mol acetic acid and a specific mass of catalyst were added to the flask and heated to 70°C.Then,ethanol at 70 °C was added quickly.The temperature error range was maintained within ± 0.5 °C.A sufficient stirring speed of 600 r·min-1was selected to eliminate the external diffusion resistance of heterogeneous catalysts.0.2 ml samples were withdrawn and analyzed by the gas chromatography and Karl Fischer titration method.The conversion is calculated according to analysis results to compare the catalytic performance.

Fig.1.Schematic diagram of design strategy.

Catalyst stability test:Add 2 g catalyst into 1 mol of acetic acid at 70 °C,then pour 1 mol of ethanol at 70 °C.Start timing and calculate the conversion after 12 h of reaction.Wash and dry the catalyst after reaction,and then follow the previous procedure to calculate the conversion for subsequent cycles.The catalyst stability can be illustrated by comparing the conversion retention after several cycles.

SCP stability test:The stability of the catalytic coating was evaluated by calculating the falling off rate of the coating in the mixed solution of ethanol and water (7:3).

2.5.Lab-scale RD experiment

The lab-scale distillation experiment(Fig.2)was carried out in a glass column with an inner 30 mm diameter at an atmospheric pressure.A 1000 ml four-neck flask with a stirring heating device was taken a reboiler,and a condenser fitted with a relay controller was taken for reflux.The height of rectifying and stripping section were 0.29 m,and the middle reactive section was 0.58 m.Two feed ports were located at the top and bottom of the reactive section.The height of each feed section is 100 mm,and there are no packings here.Acetic acid is fed into the upper feed port through the metering pump,and ethanol is fed into the lower feed port.The column wall is insulated through thermal insulation foam.The packings used in the experiment and the filling method of tea bag catalytic packing were shown in Fig.3.The rectifying and stripping sections were loaded with 4 mm × 4 mm cannon packings(Fig.3(c)).For the tea-bag catalytic packings,they were mixed with 4 mm×4 mm cannon packings in the reactive section as shown in Fig.3(e).For SCP,they are directly mixed with cannon packings and loaded into the reaction section.Since the column diameter,column height and packing type were determined,the appropriate total feed rate can be determined according to the F-factor method,which is 15-35 ml·min-1[23].

All experiments were carried out at an atmospheric pressure.The experimental steps are as follows: first add a certain amount of mixture of ethanol and acetic acid(V:V=1:1)to the column kettle,then start the heating jacket.When reflux liquid appears on the top of the column,start feeding according to the preset proportion.After full reflux for 20 min,adjust the reflux ratio and other parameters according to preset conditions.During the experiment,the liquid level of the column kettle is adjusted by the metering pump to keep constant.The pressure drop of the column is detected by an U-type press differences meter.Each group of experiments is sampled and analyzed by gas chromatography and moisture analyzer after reaching the steady-state conditions.At the steady-state,the inlet and outlet flow are stable,the tower pressure drop is stable,the column kettle liquid level and heating power are stable,and the temperature change of each section is stable within±1.0°C[24,25].Once the parameter setting is changed,the sampling analysis is carried out after steady-state conditions is established.The reaction conversion rate (Xi) of the RD process for the reactant is defined as Eq.(1).

Fig.2.Sketch of the RD column.1—reboiler;2—rotermeter;3—peristaltic pump;4—thermometer and sampling point;5—thermometer;6—condenser;7—magnetic valve;8—Metering tank.

Fig.3.Packings used in the experiment: (a) original ceramic Raschig ring,(b) SCP,(c) cannon packing,(d) tea bag packing,(e) filling method of tea bag packing.

wherexi,dis,xi,bot,xi,feedare the mass fraction of the componentiin the liquid phase of the column top distillate,the column bottom discharge and the feed respectively.

3.Results and Discussion

3.1.Characterization of catalytic packing

The original ceramic Raschig ring was prepared by sintering kaolinite at high temperature,and it has a high strength which is sufficient to resist the swelling force of the resin.Due to hightemperature sintering,the surface was relatively smooth (Fig.4(a)) and it was difficult to attach the catalyst.Therefore,an‘‘etching-modification filling-graft copolymerization” strategy was proposed and conducted as shown in Fig.1.Firstly,a rough outer surface is created to increase the contact surface with the catalyst.Ceramic Rasch rings are mainly composed of silicaaluminates,etc.According to the acid-base difference,alkali can preferentially etch SiO2.Part of SiO2was selectively etched through sodium hydroxide solution,and artificially manufactured porous structure can be produced for later catalyst loading.EDS data showed that the content of Si decreased from 26.83% to 13.79%(Table 2) after alkali washing,and the rough surface was also observed by SEM (Fig.4(b)).This treatment also contributes to the increase of a specific surface area (Fig.5).Further,through the hydrolysis of (3-mercaptopropyl) trimethoxysilane,silicon microspheres with sulfhydryl group were filled in the gap (Fig.4(c) and (d)),and microspheres hydrolyzed by silane were firmly bonded to the ceramic surface.This treatment provides a mortise-and-tenon like surface and increases the specific surface area (Fig.5).Meanwhile,it introduces reactive sulfhydryl groups which can be proved by ATR-IR in Fig.6 (2500-2600 cm-1) [26].

Table 2 EDS data (%)

Fig.4.SEM figures of:(a)original Raschig ring,(b)Raschig ring after alkali washing,(c) and (d) silane modified Raschig ring.

For the silane coupling agent combined with the small molecule catalyst,it is difficult to achieve a sufficient ion exchange capacity according to previous works [12],so the polymeric ion exchange catalyst was considered.The sulfhydryl group exists on the surface,so it can be used as the cross-linking functional group of polymerization to graft viscoelastic polymers.Nafion has been reported to have high esterification activity and can be applied to RD[9],but it is expensive.Acidic polyionic liquids have been widely utilized on the silica gel and other carriers for esterification catalysis,and have good activity [26],so the relatively cheap imidazole thiosulfate was selected as the active component.Through copolymerization of double bond and sulfhydryl group,P[SO3H-(CH2)3-VIM][HSO4]were successfully introduced on the ceramic Raschig ring as the active component.SEM (Fig.7) shows that the polymer fills the original gap,and the specific surface area test(Fig.5(d))shows that the specific surface is reduced due to polymer filling.At the same time,ATR-IR (Fig.6) also detected the peaks of imidazole and sul-fonic acid groups on the surface.These confirmed the strategy shown in Fig.1.

3.2.Performance of SCP

The purpose of the above preparation strategy is mainly to obtain catalytic packings with high binding strength and relatively high activity for subsequent reaction distillation.Since the catalyst of SCP is attached in a thin layer with a much lower loading amount than that in the tea-bag packing,it is necessary to determine whether the selected catalyst has a performance close to or beyond that of commercial catalysts or not.So,the activity of the catalytic coating was measured in a fully mixed reactor firstly.It was found that its activity at the low mass concentration was higher than that of commercial strong acid ion exchange resins such as Amberlyst 15 dry under the same test conditions (Fig.8(a)).The cross-section of the coating was measured by SEM.Its thickness was 30 μm (Fig.7(c)),and the catalyst loading amount was about 5.7% of the Raschig ring by weighing.Further,the performance of the supported Raschig ring packing was compared with Amberlyst 15 dry at the same mass,and it was found that its activity was slightly worse,but the final conversion was basically the same.Compared with autocatalysis (auto in the Fig.8(b)),the catalytic activity was greatly improved.Since P[SO3H-(CH2)3-VIM][HSO4] is lower than Amberlyst 15 dry in terms of ion exchange capacity(2.07 mmol·g-1vs4.7 mmol·g-1),the source of the SCP activity may be that P[SO3H-(CH2)3-VIM][HSO4] has a larger specific surface area than Amberlyst 15 dry (683 m2·g-1vs53 m2·g-1).In terms of catalyst stability,the conversion of P[SO3H-(CH2)3-VIM][HSO4]decreased after the first catalytic experiment,but it can remain relatively stable for the subsequent cycles(Fig.9).The activity of Amberlyst 15 dry remained relatively stable in 5 cycles.This may be due to the high ion exchange capacity of Amberlyst 15 dry.The catalyst performance can be restored after regeneration with 0.1 mol·L-1sulfuric acid,and it shows that the catalysts have certain regeneration.

Fig.5.Nitrogen adsorption-desorption isotherms of:(a)original Raschig ring,(b)Raschig ring after alkali washing,(c)silane modified Raschig ring,(d)grafted Raschig ring.

Fig.6.ATR-IR spectra.

Fig.7.SEM of: (a) and (b) grafted Raschig ring,(c) coating section of grafted Raschig ring,(d) Nitrogen adsorption-desorption isotherms of catalytic coating.

Fig.8.Comparisons of catalytic performance: (a) coated catalysts and commercial catalysts,(b) SCP and Amberlyst 15 dry.

In terms of stability,we evaluated it in two ways.Firstly,the bonding strength of the coating was evaluated by calculating the coating abscission rate through the long-term ultrasonic damage(70 °C,150 W,40 kHz).The results showed that the abscission was high in the early stage (8 h),about 20%,but there was almost no further abscission after 8 h (Fig.10(a)).Through SEM observation,it was found that the coating at the original smooth part basically fell off(Fig.10(c)and(d)).Due to limitation of surface defects and cross-linking combination,the coating mainly existed in the defects as small balls with close cross-linking bonding,and it was difficult to get out.The above preparation strategy was further proved,and the bonding strength can be attributed to limitation of the mortise-and-tenon like construction.Another way to evaluate stability of the coating was to test the stability of cyclic catalysis.Under the same condition as the catalyst activity test,the 24 h conversion of 2.5 g SCP was measured circularly,and the results are shown in the Fig.10(b).The conversion rate decreases slowly,and the conversion of the initial three cycles can maintain about 90% of the first cycle.The conversion decreases significantly after the third cycle,and this may be caused by the loss of ion exchange which decreased from (0.11 ± 0.01) mmol·g-1to (0.08 ± 0.01)mmol·g-1.

Fig.9.Comparisons of catalyst stability: (a) under cyclic conditions,(b) after regeneration.

Fig.10.Stability tests of catalyst: (a) ultrasonic damage,(b) cyclic catalysis,(c) and (d) SEM of packing surface after ultrasonic.

3.3.Results of RD experiment

Reaction distillation is a strongly coupled multi-variable nonlinear system.Since the main aim of this work is to examine the application possibility of SCP,the detailed influence of the distillation operation parameters is not conducted in experiments.Therefore,the length of each section of the column was remained constant to illustrate the SCP’s performance in comparison with the catalyst packages though it may not be an optimal configuration.In the experiment,the parameters which are adjusted are the catalyst mass,the reflux ratio,the reactant feed rate and the column kettle heating power which can be corresponding to thedistillate rate.Experiments with different parameter settings(Table 3),such as different reflux ratios (E1,E2 and E3),different distillate rates (E1,E4 and E5),and different feed rates (E1,E6 and E7),were conducted to show performance of SCP and tea bag catalyst packing (TBP).

Table 3 Experimental conditions of RD

Fig.11.Distillation experiment results of different content of SCP: (a) ethanol conversion,(b) acetic acid conversion.

Fig.12.Distillation experiment results of different catalyst packing types: (a) at different reflux ratios,(b) at different distillate rates,(c) at different feed rates.

Fig.13.Long term stability tests of distillation experiment: (a) conversion,(b) and (c) SEM of packing surface.

The 580 mm height of the reaction section used in the distillation experiment can be filled with about 290 g SCP at most.To facilitate comparison with TBP,other spare volume of the reaction section were filled with Canon packing when the amount of SCP in the reaction section was below 290 g SCP.The SCP was evenly mixed with Canon packing in the reaction section.Several parameters of the experiment are shown in Table 3.Results (Fig.11) showed that the conversion above 60 g SCP (3.42 g coated catalyst) did not change significantly.This shows that under these configurations,60 g SCP can basically satisfy the reaction requirement.

Further,the performance differences between the SCP and the TBP was compared.For TBP,13 tea bags were filled in the reaction section with a total mass of 9.421 g Amberlyst 15 dry while the left spare volume in the reaction section were also filled with Canon packing.For SCP,60 g SCP has the same volume as TBP,and it means that the volume of Cannon packing filled in the remaining space is the same,so the conversion rate of 60 g SCP with 3.42 g coated catalyst can be used for comparison (Fig.12).In general,SCP shows the same trend as TBP.For example,the conversion increases with the increase of reflux ratio and distillate rate.This is also in line with the general situation of reactive distillation.However,when the feed rate of ethanol or acetic acid is increased to E6 or E7 condition,the SCP conversion rate decreases significantly and it is even lower than TBP.Considering this feed rate is close to the maximum flow rate that the experimental device can handle(35 ml·min-1),the reason may be that the large liquid rate in the column could cause the significant change of the flow state in the TBP and SCP,and further affect the performance of RD[5,8].

It is found that the performance of SCP is a little better than that of TBP under a lot of conditions (Fig.12) for esterification distillation of ethyl acetate.It could be attributed to better combination between mass transfer performance and high activity surface of SCP[27,28].As mentioned before,the catalyst activity of SCP is better than Amberlyst 15.At the same time,SCP has no fixed components such as wire mesh,and its surface renewal of liquid under the same flow rate is faster than that of TBP.Combined with the faster reaction rate,SCP can achieve high conversion under a certain configuration with a less amount of catalyst than TBP[9,10,27].It should be noted that TBP is better than SCP in some configurations(E6 and E7),and this show the complexity of the RD.

It is worthy of noting that the catalytic packing prepared in this work showed a relatively stable conversion rate in the distillation experiment of 72 h(Fig.13(a)).From the SEM of the catalyst packing surface after 72 h,the surface coating was only slightly swollen and broken without large-area falling off,and the failure position was mainly due to self-bursting rather than interface falling off(Fig.13(b)and(c)).This shows that the structure similar to mortise and tenon can greatly improve the bonding strength of the catalyst through the geometric constraint.The SCP preparation strategy of this work can be utilized for other kinds of ceramic packings.

4.Conclusions

The strategy of ‘‘etching-modification filling-graft copolymerization”is effective to prepare the ceramic packing into the catalytic packing with considerable bonding strength due to the structure constraint in the substrate.For the esterification RD of ethyl acetate,this kind of the catalytic packing can achieve a better conversion under a lower catalyst loading in comparison with the tea bag packing with Amberlyst 15 dry in the lab-scale distillation column,and it had a stable performance in the long term test.This method can prepare the ion exchange catalytic coating with high stability on the smooth ceramic surface,so it can be utilized to other ceramic packings with different structures.Results from this work could be helpful for development and research of catalytic packings.

CRediT Authorship Contribution Statement

Zhiwei Wang:Writing-original draft,Data curation,Investigation,Conceptualization.Jiannan Shi:Visualization.Xiaonan Liu:Visualization.Zhikai Cao:Writing -review &editing.Yong Sha:Methodology,Conceptualization,Supervision,Writing -review &editing.

Declaration of Competing Interest

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.

Acknowledgements

The authors acknowledge financial support provided by the National Natural Science Foundation of China (No.21978243).

Nomenclature

Bbottom of column liquid flow rate,ml·min-1

Ddistillation flow rate,ml·min-1

Ffeed flow rate,ml·min-1

Rrmolar reflux ration,mol·mol-1