Song Hu, Jinlong Li, Qihua Wang, Weisheng Yang,*

1 State Key Laboratory of Green Chemical Engineering and Industrial Catalysis, SINOPEC Shanghai Research Institute of petrochemical Technology, Shanghai 201208, China

2 School of Petrochemical Engineering, Changzhou University, Changzhou 213164, China

Keywords:Propylene oxide purification Ionic liquids Separation Extractive distillation Hydrolysis 1,2-propylene glycol

ABSTRACT It is difficult to separate the methanol and hydrocarbons in the propylene oxide(PO)purification process due to their forming azeotrope.As for this, a novel PO separation process, in that the deionized water is employed as extractant and 1,2-propylene glycol(MPG)that is formed from the PO hydrolysis reaction is recovered, is presented in this work.The salient feature of this process is that both the non-catalyzed reactions of PO hydrolysis to form MPG and dipropylene glycol (DPG) are simultaneously considered and MPG by-product with high purity is obtained in virtue of the deionized water as reflux liquid and side take-off in MPG column.In addition, the ionic liquid (IL) extractant is screened through the conductorlike screening model for segment activity coefficient (COSMO-SAC) and the comparisons of separation efficiency between the IL and normal octane (nC8) extractant for the separation of PO and 2-methylpentane are made.With the non-random two-liquid (NRTL) thermodynamic model, the simulation and optimization design for the full flow sheet are performed and the effects of the key operation parameters such as solvent ratio,theoretical stages,feeding stage etc.on separation efficiency are detailedly discussed.The results show that the mass purity and the mass yield of PO can be up to 99.99% and 99.0%, and the condenser duty, reboiler duty and PO loss in the process with IL extractant can be decreased by 69.66%, 30.21% and 78.86% compared to ones with nC8.The total annual cost (TAC) calculation also suggests that the TAC would be significantly reduced if using IL in replace of nC8 for the investigated process.The presented results would provide a useful guide for improving the quality of PO product and the economic efficiency of industrial plant.

1.Introduction

Propylene oxide (PO), the second largest propylene derivative after polypropylene, is an important basic organic chemical raw material.The chemical property of PO is very active.It can easily open the ring and polymerize.It can react with water, ammonia,alcohol,carbon dioxide and other solvents to form the corresponding compounds or polymers [1].Nowadays, there are several production processes like chlorohydrination, propyleneoxide/styrene monomer (PO/SM), cumene hydroperoxide propylene oxide(CHPPO) and propylene oxide/methyl tert-butyl ether (PO/MTBE)for PO production [1-3].However, the method of chlorohydrination captures a characteristic of serious pollution.Thus, the other ones of PO/SM, CHPPO and PO/MTBE have gradually become the main technologies in producing PO.No matter which process for PO production is employed, the solvent like methanol is usually used to promote the reaction of propylene oxidation.However,methanol and PO can form azeotrope under certain conditions[4], and the PO product with high purity cannot be obtained by the conventional distillation method [5-8].Meanwhile, there are other impurities like aldehydes and hydrocarbon compounds(C5-C7)contained in the crude PO,and it is also difficult to separate them due to their boiling point be close to one of PO or forming azeotropic mixtures with PO.Therefore, the extractive distillation process with extractant to purify the crude PO is usually applied.In open literatures, the extractants ofn-octane (nC8) [9,10],1-methyl-2-ryrrolidone [11], triethylene glycol [12], 1-propanol[13], the mixture composed of triethylene glycol and a lower boiling cosolvent [14]and so on were employed to separate the impurities such as methanol, water, propionaldehyde (PRA),etc.from the crude PO successfully.Recently, a method to remove the impurity of acetaldehyde (AC) with adsorptive technology was also presented [15].For the removal of C5-C7hydrocarbon components, the open chain or cyclic paraffin with the carbons from 8 to 20 [16-18]or the mixtures of water and paraffinic hydrocarbons [19]were always used as extractant.Furthermore,the PO hydrolysis reaction [1,20]to produce 1,2-propanediol(MPG) and further form dipropylene glycol (DPG) by-products really occurring in the separation process was always neglected,and also no research about the separation process for the recovery of glycol produced from PO hydrolysis is reported so far.

Therefore, a novel process of PO purification and glycol separation is presented in this work.The feeding of crude PO contains the impurities of methanol, light hydrocarbon with five or six carbon atoms and so on.The deionized water is used as extractant to purify PO and the solvent of normal octane(nC8)or ionic liquids(ILs)is selected as extractant to separate the impurities of light hydrocarbons.Meanwhile,an integrated process for MPG separation is proposed.The salient characteristic is that the non-catalytic hydrolysis of PO to MPG and DPG is considered.On the other hand,an energysaving process of separating light hydrocarbon impurities with ILs extractant is developed based on the conductor-like screening model for segment activity coefficient (COSMO-SAC) [21,22]screening method [23-25].

2.Development of Separation Process

The crude PO that is from the mixture being removed the light impurities like formaldehyde,acetaldehyde,etc.is employed as the feed for the development of separation process.In the feedstock,the impurities of water, diol, methanol, C5and C6hydrocarbons,isopropyl benzene and benzene are included.Among them, the components of methanol, water, 2-methylpentane (2-MET) can form azeotropic mixtures with PO [4], leading that the mixture is difficult to separate with ordinary rectification method.To separate PO and methanol mixtures, the deionized water is selected as extractant and the extractive distillation process is designed.In Fig.1, the effects of water on the relative volatility of PO to methanol are given.With the increase of water content, the relative volatility of PO to methanol is significantly increased, and the azeotrope of PO and methanol can be eliminated.Here it should be noted that thex1′ofx-axis represents the normalized liquid phase PO mole composition after the elimination of water.Actually, the interactions among PO, methanol and water have deeply been studied from the molecular level in our previous work[26].It has showed that the addition of water can change the interaction mode of PO and methanol and effectively improve the separation of them.

Fig.1. Effect of water on the relative volatility of PO to methanol.(x1′ is the PO composition without water; the values of 0.0, 0.1, 0.2 and 0.5 are the contents of water in mixtures (xwater)).

However,due to the addition of water and the active reactivity of PO,it could not be avoided that the hydrolysis reaction without catalyst between PO and water would occur and the product of MPG would be generated.The components of MPG and PO might continue to react with each other and generate DPG product.The more the water is added, the more the PO is lost due to these side-reactions.Therefore,if the MPG generated from PO hydrolysis could be effectively separated and recovered for sale, it can not only reduce the chemical oxide demand (COD) in wastewater,but also improve the economic benefits of the production unit.

Therefore, to obtain the PO product with high purity and MPG products, a new complex process that is composed of the extraction rectification , MPG recovery , MPG product , PO product and solvent recovery columns is presented.The process is illustrated in Fig.2.The crude PO containing methanol, alkane and other impurities is first introduced into the extractive distillation column.The extractant water is added from the upper section of the column to remove the impurities such as water, propionaldehyde, ethylbenzene and methanol.The crude PO containing the heavy impurities(C5-C6)is obtained from the top of the extraction and rectification coulumn.This mixture is then sent to the PO product column, where then-octane (nC8) is used as the extraction solvent to extract and remove the hydrocarbon impurities of C5and C6.The PO product with high-purity is obtained from the top of the column, and the bottom kettle liquid is sent to thenC8solvent recovery column.The extractant ofnC8is obtained from the bottom of solvent recovery column and returned to PO product column.The obtained hydrocarbon impurities at the top of the solvent recovery column are discharged.On the other hand, the mixtures containing MPG from the bottom of the extractive rectification column are pumped into the MPG recovery column.At the top of this column, the stream contains water and light organic impurities.A part of the mixture is used as reflux of the MPG recovery column and some are considered as wastewater for post-treatment, while the others that contain PO, propionaldehyde, benzene and water are recognized as fuel gas.The streams at the bottom of the MPG recovery column, that contains MPG, DPG and small amounts of heavy components and water, are introduced into the MPG product column.To obtain the MPG product, the deionized water is chosen as the backflow, instead of directly using the condensate at the top of this column.The effect of the trace metal ions or the polymers from PO or PO and diol polymerization on the chrominance of the MPG product can be avoided through this method.At last, the MPG product is obtained from the side line of the MPG product column.

Fig.2. Flowsheet of PO purification and recovery MPG process.

In addition, for the separation of PO and hydrocarbon impurities, the ionic liquids (ILs) are also employed as extractant in this work.Here,it should be noted that the process is slightly different from the one withnC8extractant when the ILs are applied.Because of the non-volatility of ILs, a flash tank is directly used for the ILs solvent recovery instead of column.On the other hand,it has been shown that ILs has a strong interaction with PO and a weak interaction with hydrocarbons[26]so that the addition of ILs results in an increase in the relative volatility of alkanes to PO.Therefore,the hydrocarbon impurities are obtained at the top of the PO product column while the PO product with high purity is obtained from the top of the flash tank.

3.Thermodynamic Model and ILs Screening

Thermodynamic model is the basis of process simulation and solvent screening, and its accuracy directly determines the result accuracy of process simulation and solvent screening.Thus, it is very important to get the accurate thermodynamic data and select a reasonable thermodynamic model for the process development and simulation.All of the compounds like propylene oxide, water and methanol contained in the investigated mixture are strong polar substances.To obtain the reasonable interaction parameters, the NRTL model [27]is selected to describe the phase equilibrium data from experimental measurement.Meanwhile, some binary parameters for certain systems built in Aspen Plus software are modified according to experimental results[4,28-34].However, no related phase equilibrium data for the mixtures of PO + DPG and ethyl benzene (EB) + 2-MET are available in open literatures so far.Thus, for the two systems, the pseudo vapor-liquid equilibrium data are predicted with UNIFAC and COSMO-SAC [22]and then input into the Aspen Plus to obtain the NRTL parameters.To show the reliability and feasibility of the prediction methods, the prediction of vapor liquid equilibrium for PO + methanol is firstly carried out, and the ones for PO + DPG and EB + 2-MET are then performed, as shown in Fig.3.In Fig.3(a), the comparisons of isobaric vapor liquid equililbrium (T-x1(y1)) among experimental data, predictions from UNIF-DMD and COSMO-SAC for PO + methanol system are made.The corresponding experimental data from literature[34]and measurement by us are also given.Good consistency among them is observed.The VLE of EB + 2-MET is then predicted with UNIF-DMD and COSMO-SAC, as shown in Fig.3(b).For the binary mixture of PO + DPG, the prediction with UNIFAC model is made (Fig.3(c)) since the two components capture a large difference in volatility.One can find that both predictive methods give similar predictions for the investigated mixtures.Here, it should be noted that the prediction methods are just used for an initial attempt to develop a novel process.The experimental measurement for these missing data must be performed if the proposed process could be applied for industrial practice.

Fig.3. VLE for binary mixtures calculated from UNIFAC and COSMO-SAC at 101.3 kPa.

As we all know that the ionic liquids can be composed of different anions and cations leading that the number of ILs is numerous[35].It is impossible that the screening of suitable ILs extractant is carried out through the experimental method.It is also time- and labor-consuming.Thus, the priori predictive model of COSMOSAC is employed to sift the ILs extractant.This method has been verified to be effective in solvent screening,and the related details can refer to our previous work [24,25].With the same calculation method shown in the previous work [26], the solvation information of different anions and cations that form the ILs are firstly obtained.The influence of different anions and cations on the relative volatility of PO to hydrocarbon is then analyzed with the help of COSMO-SAC.In this work, 2-MET is chosen to represent the hydrocarbon impurities.The influence of various ILs on the relative volatility of PO to 2-MET is depicted in Fig.4.

Fig.4. Effects of ILs extractant on the relative volatility of PO to 2-MET at 101.3 kPa.(Abbreviation: [C2MIM]+, 1-ethyl-3-methylimidazolium; [C4MIM]+, 1-butyl-3-methylimidazolium; [C6MIM]+, 1-hexyl-3-methylimidazolium; [PYHM]+, 1-hexyl-3-methylpyridinium;[P4444]+,tetrabutylphosphonium;[MGuH]+,dimethyldihexylguanidine; [AMIM]+, 1-allyl-3-methylimidazolium; [BF4]-, tetrafluoroborate; [Cl]-,chloride; [DEP]-, diethylphosphate; [DBP]-, dibutylphosphate; [EtSO4]-, ethylsulfate; [NTf2]-, bis(trifluoromethylsulfonyl)imide; [OAc]-, acetate; [PF6]-,hexafluorophosphate).

Fig.4(a) shows the influence of cation (the anion is fixed as tetrafloroboric acid ([BF4]-)) on the relative volatility while Fig.4(b) the effect of anion (the cation is fixed as 1-ethyl-3-methylimidazole ([C2MIM]+)) on the relative volatility.As can be seen from Fig.4, both the cation of 1-allyl-3-methylimidazole([AMIM]+) and the anion of hexafluorophosphate ([PF6]-) have a greater influence on the relative volatility of PO to 2-MET.Therefore, the IL of 1-allyl-3-methylimidazoliumhexafluorophosphate([AMIM][PF6]) is considered as the extractant for PO purification.Here, it should be noted that the PO captures a greater volatility when ILs are free in the mixture(see the dash line in Fig.4),while the volatility of 2-MET is strongly improved leading to a decreasing relative volatility of PO to 2-MET(less than 1)over the whole range of PO concentration when introducing ILs into the mixture,so that making the separation of PO and 2-MET possible.Thus,the stream at the top of the PO product column is the mixtures of hydrocarbon impurities,while the PO product is obtained from solvent recovery column top (see Fig.2).In IL process, the solvent recovery column is replaced by a flash tank.

4.Simulation and Optimization

According to the industrial scale for PO production[1],the scale of the PO unit designed is set to about 400 kt·a-1PO product based on the annual running time 8000 h, with PO mass purity ≥99.99%and water mass content ≤0.0030%[1,36].The mass purity of MPG product is ≥99.50%.The feed temperature and pressure of crude PO are 40 °C and 300.0 kPa, respectively.The mass contents of PO,water, propylaldehyde, benzene, ethylbenzene, methanol and 2-MET in the feed are supposed to be 98.6%, 1.0%, 0.03%, 0.12%,0.01%,0.10%and 0.06%respectively according to the characteristics of the process for the epoxidation of propylene.The simulation and optimization are carried out by Aspen Plus software (Version 11).All columns are modeled with the RadFrac module.All of the conventional compounds contained in the investigated process are selected from the databank of the Aspen Plus package, while the ILs are defined as new conventional ones.The required thermodynamic properties like critical properties,boiling temperature,acentric factor and heat capacity of ILs in simulation are determined according to the literatures’ method [37-41].The optimization order is determined according to the process flow and the importance of operation parameters.All other parameters,those are considered as initial values (see Table 1) determined from the converged calculation for the initial designed process, would be fixed when one parameter is optimized.

Table 1Optimized parameters for the investigated flowsheet

4.1.Extractive distillation column

Fig.5. Optimization of theoretical stages (a), extractant flowrate (b) and feed position (c) of crude PO for extractive distillation column.

Fig.6. Relationships between MPG flowrate and variables of extractive distillation column.

The extraction distillation column is mainly used to remove the impurities of water,propylaldehyde,methanol,etc.under the help of water entrainer.In the separation process,it must try to improve the efficiency of the impurity removal and simultaneously minimize the PO losses.For the column, the variables of the number of theoretical stages, the flow rate of extractant, the feeding position and so on capture important effects on the separation efficiency.In optimization, the reflux ratio is considered as a control variable to match with the design specification of the PO recovery yield.Fig.5(a)-(c) shows the relationship of the PO loss and the heat duty with each variable.The number of the theoretical stage is firstly optimized with all other fixed parameters, as shown in Fig.5(a).With the increase of the number of theoretical stages,the PO loss generally increased linearly, while the heat duty of reboiler rapidly decreased firstly and then become unchangeable.The reason would be contributed to the enhancement of separation efficiency with the increase of the number of theoretical stages.It makes the reflux ratio decrease when the same recovery yield of PO is kept, resulting in that the heat duty of the reboiler would reduce.However, when the number of the theoretical stage reaches a certain number, the separation capacity of the column would approach a limit.Therefore, the reflux ratio and heat duty would be unchangeable.According to the above analysis,the optimal number of the theoretical stage for the extractive distillation column is determined as 74.

The flow rate of water extractant is then optimized at the fixed number of theoretical stage and other variables and the result is illustrated in Fig.5(b).One can see that the PO loss decreases linearly with the increase of water flow rate.It also shows that the heat duty slightly decreases firstly and then increases.However,the overall change of the heat duty is not large and there is a minimize point at 2630 kg·h-1.Fig.5(c) shows the effect of feed position on the PO loss and reboiler duty.It can be seen that the PO loss linearly decreases with the feed position moving down from up,while the curve for the heat duty has a lowest point at the 53th theoretical stage.Therefore, the number of theoretical stages for the optimal feed position of crude PO is 53.

It has been explained above that the PO recovery yield is defined as the design specification in the optimization.For the conventional distillation column,the PO loss yield should be kept constant.However, the reactions of PO hydrolysis to generate MPG and DPG are considered in the simulation so that the PO loss is closely related to the reaction of PO + water →MPG occurring under different conditions.In Fig.6, the curves to show the relationship between the number of the theoretical stage, the flow rate of extractant and the feed position and the flow rate of MPG are provided.One can see that the opposite changes for the flow rate of MPG and the PO loss are observed along with the variation of the above three parameters.This is because the concentrations of PO and water and the temperature distributions in the column changes with the changing of the parameters of the extractive distillation column.

4.2.PO product and recovery columns

The mixture stream derived from the top of the extractive distillation column mainly includes PO,2-MET and other trace impurities.The compounds of PO and 2-MET can form azeotrope under certain condition so that they must be separated through special extractive distillation.For the separation of PO and hydrocarbons, the isooctane ornC8was often employed as extractant.HerenC8is used.On the other hand, based on the special characteristics of ILs such as non-volatile and non-combustible compared with ones of organic solvents, the ILs is also used as an extractant to separate PO and 2-MET for the first time.The basic physical properties(boiling point, critical properties, acentric factor,etc.) of ILs are calculated according to the method reported in open literatures[37-41].

4.2.1.Process with nC8 extractant

The process ofn-octane extractant is mainly composed of the PO product column and the solvent recovery column.The PO product with high purity is obtained at the top of the PO product column and the 2-MET impurity is extracted at the top of the solvent recovery column.ThenC8extractant at the bottom of the column is recovered and pumped back to the PO product column as recycling solvent.

For the PO product recovery column,the solvent ratio,the number of theoretical stage and the feed position ofnC8are optimized in order, and the corresponding optimal parameters of them are 2.165, 50 and 13, respectively.The effects of the solvent ratio and thenC8feed position on the purity of PO and the reboiler heat duty are depicted in Fig.7.As can be seen from Fig.7(a), with the increase of solvent ratio, the heat duty rapidly increases linearly,while the purity of PO product first increases rapidly and then slowly rises.According to the design specification for the purity of PO product 0.9999, the optimum solvent ratio in mass is 2.165.In Fig.7(b), one can see that the purity of PO product increases rapidly and then basically remains unchanged as thenC8feed position moved down along with the column.The heat duty of the reboiler captures a similar trend,but with a small overall change.Therefore, the 13th theoretical stage is determined as the optimalnC8feed position.

Fig.7. Effects of extractant ratio (a) and feed position (b) on PO product column.

The same method is applied to optimize the parameters of the solvent recovery column.The results are shown in Fig.8.Fig.8(a) shows the change ofnC8solvent loss and heat duty with the number of theoretical stages.With the increase of the number of theoretical stages, thenC8loss rapidly decreases within a certain range.ThenC8loss is basically kept constant when the number of theoretical stages is greater than 22.In Fig.8(b), the effect of the reflux ratio on thenC8purity and heat duty is given.Both thenC8purity and the heat duty increases with the increase of reflux ratio.In order to minimize the energy consumption and at the same time the influence of the purity ofnC8on the operation of the PO product column,the best reflux ratio is optimized as 603.07 in mass.The value of the reflux ratio could ensure thenC8extractant with high purity 0.99999.

Fig.8. Effects of theoretical stage (a) and reflux ratio (b) on solvent recovery column.

4.2.2.Process with IL extractant

The process of IL extractant is different from the one withnC8extractant.Due to the characteristic of IL’s non-volatility, the solvent recovery column can be replaced by a flash tank for the IL recovery.So,the process of IL extractant is composed of an extraction column(for removing 2-MET impurities) and a flash tank (for recovering ILs).The operation conditions for the flash tank can be determined according to the design specification of the PO product purity and the nature of the vapor liquid equilibrium.For the operation conditions of the extractive distillation column,the identical pressure with one ofnC8process is employed.

According to the operation characteristics of the extractive distillation column,the solvent ratio, the number of theoretical stage and the feed position are analyzed in detail.The results are illustrated in Fig.9.The effects of the solvent ratio on the removal of 2-MET and the heat duty are given in Fig.9(a).With the increase of solvent ratio, the heat duty of reboiler increases linearly, and the flow rate of 2-MET approximately rises linearly too.To make the purity of PO product meet the design specification,the removal yield of 2-MET must satisfy the specified requirement.Thus, the optimal solvent ratio is set to 9.16 in mass based on the balance between the 2-MET removal yield and the energy consumption.Fig.9(b) shows the effect of the number of theoretical stages on the separation efficiency.The change trends of heat duty and 2-MET removal yield with the number of the theoretical stage are consistent.They rapidly increase first and then basically keep constant when the number of the theoretical stage is greater than 29,which is considered as the optimum value.The effect of feed position on the separation is provided in Fig.9(c).With the feed position moving from top to bottom, the number of the theoretical stage in rectifying section will increase leading the improvement of separation between 2-MET and PO.However,the separation efficiency cannot continuously be enhanced when the separation efficiency of the rectifying section reaches to one limitation.Thus,based on the curves shown in Fig.9(c), the best feed position is the 19th theoretical stage.

Fig.9. Effects of ILs extractant ratio(a),theoretical stage(b),feed position(c)of PO feeding on separation.

Fig.10. Materials balance diagram for flowsheet.(The unit of the flow and compositions are kg·h-1 and percent in mass respectively; 1 ppm = 10-6 (mass)).

4.3.MPG recovery and product column

In virtue of the same method above,the number of the theoretical stage,the feed position and the reflux ratio in mass of MPG recovery column obtained by optimization are 12, 6, and 0.08 respectively.For the MPG product column, the number of the theoretical stage is 12,the feed position for MPG feedstock and deionized water are 8 and 2 respectively,and the side line position for extracting MPG product is 4.The finial mass purity of MPG product is ≥99.50%.

5.Operation Parameters and Steady-state Simulation

The optimum operation parameters obtained through the optimal method above are listed in Table 1,in which the parameters of the number of the theoretical stage,feed position,mass reflux ratio and operation pressure are included.Based on these optimum parameters, the steady-state simulation for the whole process is performed.The mass purities of the obtained PO and MPG products are ≥99.99% and ≥99.50%, respectively.The obtained cooling and heat duties are also given in Table 1.In Fig.10,the detailed balance of materials is given, whereFis the mass flow rate (kg·h-1), ppm represents the unit for one in a million in mass and the composition is the mass fraction.Fig.10(a)is the results of the process withnC8extractant and Fig.10(b) the ones of the process with ILs extractant.From Table 1, one can see that the total heat and cooling duties for the PO product column and solvent recovery column are 10.336 and 3.969 MW respectively when employing ILs as extractant.The corresponding values are 30.21%and 69.66%lower than ones fornC8extractant, in which they are 14.811 and 13.081 MW,respectively.Meanwhile,the application of ILs extractant can make the number of the theoretical stage of PO product column reduce from 50 to 29, and of the solvent recovery column from 22 to 1 (flash tank).On the other hand, it can be seen from Fig.10 that the PO loss reduces from 20.48 to 4.33 kg·h-1and the solvent loss from 5 kg·h-1ofnC8to 0.78 kg·h-1of ILs when using ILs replacenC8extractant.In general, the separation efficiency could be effectively improved when the ILs extractant were employed.

To compare the difference in cost betweennC8and ILs used as extractants for the PO purification and recovery process, the total annual cost (TAC) calculation is performed according to the literature’s method [42,43].Here, the payback period is set to 3 years and the prices ofnC8and [AMIM][PF6]are supposed to be 1895 USD·t-1and 104USD·t-1, respectively.The detailed results for TAC calculation are given in Table 2.One can see that both the total capital cost and the total operating cost would be significantly reduced, leading to much decreasing in TAC, if the selected ILs is used as extractant although the supposed price of ILs is much greater than one ofnC8.This is because the applied ILs made the separation of PO and 2-MET easier and fewer theoretical stages and energy for separation process are needed.The result showsthat there is a good potential application of ILs in the investigated system.

Table 2TAC calculations for PO purification system

6.Conclusions

A novel process for the PO purification and the production of by-product MPG is proposed.The innovation point of this process is that the reaction of PO hydrolysis to MPG and DPG without catalyst is considered and the ILs extractant is employed to separate the mixture of PO and hydrocarbons.The binary interaction parameters in NRTL model are fitted from the experimental results or the pseudo phase equilibrium results predicted from UNIFAC and COSMO-SAC.The parameter optimization and process simulation of the whole process are successfully carried out, and the PO product with the mass purity of 99.99%and the MPG product with the mass purity of 99.50% are obtained.One extractant of ILs is screened with COSMO-SAC model for the separation of PO and hydrocarbon mixture.The total cooling and heat duties are 69.66%and 30.21%lower than ones innC8process and the PO loss reduces from 20.48 to 4.33 kg·h-1when using ILs extractant.Generally, the application of ILs extractant makes the investment and operation costs (TAC) reduce and simplify the whole process of PO purification.

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

Financial support for this work was provided by the National Key Research and Devolopment Program of China(2017YFB0702800) and the National Natural Science Foundation of China (21878025, 22078026).