Han Dongmin; Chen Yanhong

(Department of Chemical Engineering, Shengli College, China University of Petroleum, Dongying, 257061)

Abstract: In the present work, a comparative study of the extractive distillation and pressure swing distillation for methanol-acetonitrile separation is performed for the first time. Different separation alternatives, including the conventional extractive distillation, the extractive distillation with vapor or liquid side-stream, the pressure-swing distillation with or without full heat integration, and the heat-pump assisted pressure-swing distillation are rigorously simulated and optimized based on the minimum total annual cost (TAC) via the sequential iterative strategy. The results show that TAC and CO2 emission of the new extractive distillation with vapor side-stream (Vapor-SED) are similar to those of the extractive distillation with liquid side-stream (Liquid-SED). Furthermore, the Vapor-SED and Liquid-SED can achieve 30.01% and 30.56% reduction in TAC and 23.32% and 23.49% reduction in CO2 emission, respectively, over the most competitive fully heat-integrated PSD con figuration. Hence, the extractive distillation with vapor or liquid side-stream appears to be a better option economically and environmentally for the separation of methanol and acetonitrile.

Key words: azeotrope; extractive distillation; pressure swing distillation; TAC; methanol/acetonitrile

1 Introduction

Methanol and acetonitrile, which are important raw materials in chemical industry, have been widely used as extraction solvents, synthetic organic materials, etc.Since methanol and acetonitrile form a minimum boiling homogeneous azeotrope at atmospheric pressure and 63.5 °C with the mixture containing 81% of methanol,it is impossible to separate methanol and acetonitrile mixture by conventional distillation method. Hence,some special distillation technologies such as pressureswing distillation (PSD), or extractive distillation (ED)are needed for this separation. Perhaps there are several methods capable of separating an azeotropic mixture.However, for different azeotropic system, the most appropriate separation method must achieve a best economics and environmental effect of the system.

Many researchers have made comparisons between extractive distillation process and pressure-swing distillation process for various azeotropic systems.For examples, Ghuge, et al.[1]studied the separation of THF-water system using extractive and pressure-swing distillation methods. He found that extractive distillation with DMSO as the entrainer appeared to be a better option for this system. Luo, et al.[2]studied the separation of isopropyl alcohol and diisopropyl ether mixture using these two methods. It was revealed that the fully heatintegrated pressure-swing distillation process was more attractive in terms of steady-state economics. Similar studies were carried out for separation of acetone and chloroform[3], methanol and chloroform[4], di-n-propyl ether andn-propyl alcohol[5], acetonitrile andn-propanol[6]by extractive distillation and pressure-swing distillation. It can be concluded that for different azeotropic system, the performance of pressure-swing distillation and extractive distillation varies from system to system.

There is a problem of high energy consumption in the distillation-based processes. Consequently, it is important to reduce the energy consumption and improve the economic performance[7]. Lots of energysaving technologies have been proposed and applied to the distillation processes, such as the heat-integrated distillation[8-10], the heat-pump assisted distillation[11-12],the dividing-wall column[13-14], the reduced-pressure distillation[15-19], etc. As for extractive distillation, the extractive dividing wall column (EDWC)[14]has drawn much attention in recent years. However, the design and control of the EDWC are complicated. Tututi-Avila[20]proposed a novel liquid side-stream extractive distillation system based on the thermally coupled distillation system.And this configuration is more energy-efficient than the EDWC system. Wang[21-22]and Ma[23]further proved the efficiency and controllability of this configuration. But there is no report on the simulation of the vapor sidestream extractive distillation system. As regards pressureswing distillation (PSD), the partial or full heat integration technology[24]and the heat-pump technology[25]can be applied to reduce the energy cost. But these integration technologies may also increase the capital cost. Thus, it is necessary to investigate the PSD processes modi fied by different heat integration technologies in more detail for different azeotropic mixtures.

To the best of our knowledge, the direct comparison between extractive distillation and pressure-swing distillation for methanol and acetonitrile separation is not reported in the open literature. The purpose of this work is to compare these two methods for the separation of methanol and acetonitrile mixture. As for the extractive distillation system, the optimum entrainer was firstly chosen based on the VLE curves and the residue curve.Two energy-efficient extractive distillation processes,including the liquid side-stream extractive distillation system (Liquid-SED) and the vapor side-stream extractive distillation system (Vapor-SED), are developed based on the conventional process. In regard to the pressure-swing distillation, con figurations with full heat integration (HIPSD) and heat-pump assisted pressureswing distillation (HPAPSD) processes are analyzed. The parameters of all the processes are optimized based on the minimum TAC. Furthermore, all the processes have been compared based on the environmental and economic performance. Finally, the most economic and energy efficient process is identified among various process schemes.

2 Design Basis

In this work, a feed flow rate of 3 000 kg/h with the feedstock containing 50% of methanol and 50% of acetonitrile was taken as the basis for the simulation.The product purity specification of acetonitrile and methanol was specified at 99.5%. Aspen Plus 7.2 was used to simulate all the processes. The Wilson model for thermodynamic properties study was used to describe the non-ideality of liquid and ideal vapor phase behavior[21-22].

2.1 Basis of economic analysis

The economic analyses are evaluated in terms of total annual cost (TAC), which is the sum of the operating cost and total capital cost divided by 3 years (payback period).The more calculation details can be found in references[14] and [26].

2.2 CO2 emissions

CO2emissions can reflect the environmental impact of different processes. It can be calculated for a given amount of fuel burnt. CO2emissions (kg/h) are related as[27],

where the fuel net heating value (NHV) is 39771 kJ/kg and the mass percentage carbon in fuel (C) is 86.5% when heavy oil is used as the fuel; α is the ratio of CO2and carbon molar masses (3.67). QFuelrepresents the heat duty from fuel burnt (kJ/h), which is calculated through the following expression.

where QProcrepresents the process heat duty (kW), λProc(kJ/kg) is the latent heat of utilized steam; hProc(kJ/kg)is the mass enthalpy of utilized steam; TFTB(°C) is the flame temperature of the boiler flue gas; Tstack(°C) is the stack temperature; and T0(°C) is the ambient temperature(25 °C). As for the steam boiler, the flame temperature(TFTB) and stack temperature (Tstack) are adopted as 1 800 °C and 160 °C, respectively.

3 Methanol-Acetonitrile Separation Using Extractive Distillation

3.1 Selection of entrainer

Figure 1 Effect of different entrainers for the separation:(a) effect of differernt entrainers on VLE of methanol/acetonitrile, (b) residue curve maps for methanol/acetonitrile/chlorobenzene system and methanol/acetonitrile/aniline system at 1 atm

It is important to select an appropriate entrainer for the extractive distillation process. Organic solvents such as aniline, ethylene glycol, dimethyl formamide (DMF)and chlorobenzene are generally used as entrainers in the extractive distillation process. Figure 1(a) shows the effect of different entrainers on VLE of methanol/acetonitrile with an entrainer/feed molar ratio of 1. It can be seen that aniline and chlorobenzene can both greatly enhance the relative volatility between methanol and acetonitrile. In order to compare these two entrainers in more detail, the residue curve maps for both methanol-acetonitrile-aniline and methanol-acetonitrile-chlorobenzene systems at 1 atm are analyzed (Figure 1(b)). It can be seen from Figure 1(b) that pure methanol and acetonitrile are saddle points while aniline and chlorobenzene are stable nodes. There is no distillation region in the residual curves for both the two systems. The curves illustrate the feasibility of aniline and chlorobenzene as entrainers for methanol and acetonitrile separation. In addition, when using aniline as the entrainer，the intersection point of the isovolatility curve and the methanol-entrainer edge of the triangle is closer to the methanol corner, which means that aniline is a more effective entrainer than chlorobenzene for this system to some extent[26,28]. Therefore, aniline is chosen as the entrainer in the simulation.

3.2 Flowsheet of the conventional extractive distillation system (CED)

The conventional extractive distillation process consists of an extractive distillation column (EDC) and an entrainer recovery column (ERC) (Figure 2). High purity methanol and acetonitrile products are obtained as the distillates of the EDC and ERC, respectively.For CED system, the various design variables including the operating pressure of EDC (P1), the entrainer flowrate (S), the total number of stages of EDC (NT1)and ERC (NT2), the feed stage locations (NF1, NF2and NFS), and the molar re flux ratio of EDC (RR1) and ERC(RR2) need to be determined for achieving an optimum performance. The optimization work was done through a sequential iterative strategy reported in our previous work[28]. Figure 2 presents the flowsheet of the optimal system.

3.3 Flowsheet of the vapor side-stream extractive distillation system (Vapor-SED)

The vapor side-stream extractive distillation system(Vapor-SED) for methanol and acetonitrile separation was simulated and optimized in this section. Figure 3 shows the flowsheet of the Vapor-SED process. Methanol is withdrawn at the top of the first column (T1). A vapor side stream is withdrawn near the bottom of T1and is fed to the second column (T2). Acetonitrile is removed from the top of T2. The entrainer obtained from the bottom of T1and T2is recycled to T1.In the Vapor-SED process, the optimization variables include the entrainer flowrate (S), the total number of stages of T1(NT1) and T2(NT2), the feed stage locations(NF1, NF2and NFS), the molar re flux ratio of T1(RR1) and T2(RR2), the side vapor location (NVS), and the flowrate of the side vapor stream (VS). The optimization work was done through a sequential iterative strategy (Figure 4) to find the optimal designs.

Figure 2 Flowsheet of the optimal conventional extractive distillation process (CED)

Figure 3 Flowsheet of the optimal Vapor-SED process

Figure 5 shows the optimization procedure of the Vapor-SED system. It is observed that the optimal flowrate of the entrainer is 4 800 kg/h, the optimal total number of stages is 48 for T1and 21 for T2,respectively. As for T1, the best feed position of entrainer is at the 5th stage, the best feed position of azeotrope is at the 31th stage and the best side-draw vapor location is at the 42th stage. The optimal feed location of T2is at the 14th stage. Figure 3 presents the flowsheet of the optimal system.

Figure 4 Optimization procedures for Vapor-SED process

Figure 5 Optimization data of Vapor-SED

3.4 Flowsheet of the liquid side-stream extractive distillation system (Liquid-SED)

The flowsheet of the liquid side-stream extractive distillation system (Liquid-SED) is similar to that of the Vapor-SED system, except that the vapor side stream is replaced by a liquid side stream. Figure 6 presents the flowsheet of the optimal system.

4 Methanol-Acetonitrile Separation Using Pressure-Swing Distillation

The PSD process includes a low pressure column(LPC) and a high pressure column (HPC) (Figure 7). The raw material and the recycled distillate stream from the LPC are fed to the HPC. High purity acetonitrile is obtained from the bottom of HPC and the distillate stream is fed to the LPC. Then, the high purity methanol is drawn from the bottom of the LPC and the distillate stream, the composition of which is close to the azeotrope, is recycled to the high pressure column.

Figure 6 Flowsheet of the optimal Liquid-SED process

Figure 7 The optimal flow sheet of the conventional PSD process

4.1 Selection of pressure

Figure 8 exhibits the influence of pressure on the azeotropic composition and azeotropic temperature for methanol and acetonitrile binary system. It can be seen that the mole fraction of methanol in the azeotropes significantly increases from 0.62 to 0.95 when the pressure changes from 0.1 atm to 5 atm. It means that the pressure-swing distillation (PSD) is feasible for the separation of methanol and acetonitrile.

When the higher pressure is performed in HPC, the less reboiler duties of the columns are needed. But higher pressure would lead to high temperature requirement for the reboiler. In order to use the low pressure stream(433 K) and ensure the temperature difference between the reboiler and the stream should be greater than 20 K,while the pressure of HPC is set at 5 atm. The pressure of the LPC is optimized by minimizing TAC. Figure 9 displays the effect of pressure of the LPC on TAC. As shown in Figure 9, the TAC first decreases and then increases with a decreasing pressure of the LPC. This can occur when the pressure of the LPC is below 0.6 atm, the expensive chilled utility for the condenser operation is needed. Thus, the pressure of the LPC is set at 0.6 atm.

Figure 8 Effect of pressure on azeotropic composition and temperature

Figure 9 Effect of pressure of the LPC on TAC

4.2 Process optimization

4.2.1 PSD without heat integration

The PSD process is optimized by sequential iterative strategy reported by Wang[6], and the flowrate of bottom products in the two columns is adjusted to maintain the purity speci fication for methanol and acetonitrile. Figure 7 exhibits the detailed information of the optimized PSD process.

4.2.1 PSD with full heat integration

It can be seen from Figure 7 that the condenser duty of HPC is 2.146 MW and the reboiler duty of LPC is 1.777 MW. Meanwhile, the temperature difference between the condenser of HPC (385.1 K) and the reboiler of LPC (325.2 K) is large. It indicates that the full heat integration can be used to reduce the TAC. Figure 10 shows the detailed information of the optimized PSD process with full heat integration (HIPSD). The TAC of the HIPSD process is 0.913×106$/a, which is by 45.88%lower than that of the PSD process.

Figure 10 The optimal flowsheet of the PSD process with full heat integration

4.2.2 PSD with heat pump technology

Given that the temperature difference between the condenser of LPC (313.9 K) and the reboiler of LPC(325.2 K) is small, the PSD process coupled with the heat pump technology (HPAPSD) is investigated in this work. In the HPAPSD process, the vapor stream from the LPC is compressed to heat the liquid stream of the reboiler. Figure 11 shows the flowsheet and the optimized parameters in the HPAPSD process. The operating cost of the HPAPSD process is 0.676×106$/a which can save 0.658×106$/a as compared with the conventional PSD process. The TAC of the HPAPSD process is1.067×106$/a, which is by 36.75% lower than that of the PSD process. The results demonstrate that the use of heat pump technology shows a better economic performance.

5 Comparison of PSD and Extractive Distillation for Methanol-Acetonitrile Separation

Figure 11 The optimal flowsheet of the PSD process with heat pump technology

The key economic performance and CO2emission of all of the processes are summarized in Table 1. The results show that the total annual cost of the extractive process (0.68×106$/a) is substantially smaller than that of the pressure-swing distillation process (1.687×106$/a). Heat integration and heat pump technology can be applied to the pressure-swing process to reduce the operating cost. As shown in Table 1, in comparison with the conventional PSD process, the HIPSD process and the HPAPSD process can save TOC by 55.62% and 49.33%, can reduce TAC by 45.88% and 36.75%, and can decrease CO2emission by 45.06% and 42.23%, respectively. The PSD configuration with full heat integration is more economical as compared to the heat pump technology, because the capital cost increases significantly in case of HPAPSD system. In comparison with the conventional extractive distillation, the Vapor-SED system and the Liquid-SED system can save TAC by 6.03%and 6.76%, and can reduce CO2emission by 13.95% and 14.13%, respectively. This can occur because the use of the side stream column can reduce the remixing degree of the components in the column so that the operating cost is saved.The Liquid-SED system requires 30.56% less TAC and 23.49% less CO2emission than those of the HIPSD system.In addition, in order to evaluate the economics of relevant processes more comprehensively, we further investigate the effect of payback period on TAC (Figure 12). It can be seen that when the payback period increases from 1 year to 15 years, the TAC drops significantly in the first 4 years and later flattens. The TAC of the conventional PSD process is the highest at each point on the payback period, while the TAC of the Liquid-SED and Vapor-SED processes is the least. These results further indicate that the Vapor-SED system and the Liquid-SED system are much more attractive for methanol/acetonitrile mixture separation in comparison with conventional extractive distillation and PSD processes.

Figure 12 The effect of payback period on the TAC■—CED; ●—Vapor-SED; ▲—Liquid-SED; ▼—PSD; ◆—HIPSD;◄—HPAPSD

Table 1 Economic analysis results of different processes

6 Conclusions

In this paper, different processes for separating methanol and acetonitrile mixture have been developed and optimized.Also, the comparison has been carried out between the extractive distillation with vapor or liquid side-stream and the pressure swing distillation with heat integration or heat pump based on the environmental and economic performance.The results show that the TAC and CO2emission of the PSD process without integration are the largest among all the processes. Even though the heat integration technology and heat pump technology require by a 45.88% and a 36.75% less TAC, respectively, as compared to those of the conventional PSD process, the energy cost and TAC are still much larger than those of the extractive distillation processes.Furthermore, the Vapor-SED and Liquid-SED give a 30.01%and a 30.56% reduction in TAC, a 23.32% and a 23.49%reduction in CO2emission, respectively, as compared with the most competitive fully heat-integrated PSD process.Therefore, upon considering the apparent economic and environmental benefits, the proposed extractive distillation with liquid or vapor side-stream is an attractive choice for the separation of methanol and acetonitrile.

Novel Cyclohexanol Dehydrogenation Catalyst for Manufacture of Cyclohexanone Passed Appraisal

On July 29, 2020 the project “Commercial application of novel cyclohexanol dehydrogenation catalyst for manufacture of cyclohexanone” undertaken by the SINOPEC Nanjing Chemical Research Institute (NCRI) has passed the appraisal of research achievements organized by the SINOPEC Science and Technology Division. The experts attending the appraisal meeting have admitted that the performance indicators of the dehydrogenation catalyst NDH6 have reached the advanced level of similar international catalysts.

Currently the SINOPEC Nanjing Chemical Industry Company has applied this catalyst to manufacture cyclohexanone from cyclohexanol with the capacity of process unit reaching 160 kt/a. The heat-conducting oil system is adopted in both benzene hydrogenation and cyclohexanol dehydrogenation processes, in which the outlet temperature of heat-conducting oil is about 240 °C for preheating benzene, while the temperature of heat-conducting oil for preheating the cyclohexanol dehydrogenation system is 230 ℃ at the initial stage and then increases to 260 ℃ at the final stage. NCRI has independently developed the NDH6 type cyclohexanol dehydrogenation catalyst, which with the addition of cocatalyst can further improve the low-temperature catalytic activity and extend the cycle length of catalyst during the lower preheating temperature stage (＜210 ℃).The surplus heat from benzene hydrogenation can be used to maintain the heat needed by the dehydrogenation system, resulting in reduced operating cost and energy consumption for dehydrogenation of the recycle stream.The result of commercial application of the NDH6 type catalyst in the 60 kt/a cyclohexanol dehydrogenation unit has revealed that under standard conditions the catalyst operated smoothly, with the cyclohexanol conversion reaching more than 55.0% and the cyclohexanone selectivity exceeding 99.0%. Compared with the traditional Cu-Zn dehydrogenation catalyst, the NDH6 type catalyst can increase the cyclohexanol conversion by 4%—5% along with a reduction of steam consumption equating to 1.34 t/h.