Novel Reactor for Exothermic Heterogeneous Reaction Systems ...

0 downloads 0 Views 2MB Size Report
Apr 20, 2017 - with a static mixer, both mass and heat transfer in the reactor can be intensified .... Hydrodynamic cavitation in chemistry is an effective method.
Article pubs.acs.org/IECR

Novel Reactor for Exothermic Heterogeneous Reaction Systems: Intensification of Mass and Heat Transfer and Application to Vegetable Oil Epoxidation Zhenyu Wu, Ting Zheng, Lihang Wu, Hongfeng Lou, Qinglong Xie, Meizhen Lu, Lianzhong Zhang, Yong Nie,* and Jianbing Ji Biodiesel Engineering Lab of China Petroleum & Chemical Industry Federation, and Zhejiang Province Key Lab of Biofuel, Zhejiang University of Technology, Hangzhou, Zhejiang 310014, China ABSTRACT: A novel reactor for exothermic heterogeneous reaction systems was developed and used in a vegetable oil epoxidation process. When hydrodynamic cavitation is coupled with a static mixer, both mass and heat transfer in the reactor can be intensified effectively. The Sauter mean diameter of water/oil dispersion was 8.90 μm under hydrodynamic cavitation conditions with a relatively low inlet pressure (2.5 bar) as compared to 60.33 μm for conventional stirring approach at 500 rpm. The epoxidation temperature variation was controlled within ±1 °C even when the H2O2 was added at once, whereas the temperature rise was up to 19 °C in a stirred-tank reactor. The use of this reactor results in decreased reaction time and a simpler and safer process, with final products having quality similar to that obtained in a traditional process. pseudohomogeneous system9−11 or a heterogeneous system.7,12−14 Most studies ignored the complex mass- and heattransfer phenomena in the process or eliminated the mass- and heat-transfer effects. Heterogeneous catalysts, such as ion exchange resins15−17 or titanium-based catalysts,18,19 were used in the reaction to improve the conversion of double bonds and selectivity to oxirane. Process intensification focuses on improvement of product quality by improving mass and energy transfer. Chavan et al.20 studied the intensification of epoxidation of soybean oil using tetra-n-butyl ammonium bromide as a phase-transfer catalyst in the presence of a sonochemical device and found that the reaction rate can be shortened remarkably. Aguilera et al.21 find that the reaction rate of oleic acid epoxidation can be accelerated under microwave radiation conditions. Recently, the study of epoxidation of vegetable oil at high temperatures has attracted the interest of researchers because a high reaction yield for the formation of oxirane rings was obtained in a short time. Santacesaria et al.22 studied soybean oil epoxidation with performic acid in continuous reactors by experiments and mathematical methods and found that the best reaction yield would be obtained in less than 45 min with a temperature of 90 °C, instead of the 6−8 h required for traditional industrial methods with a temperature of about 60 °C. De Haro et al.23

1. INTRODUCTION In recent years, bioenergy resources and their comprehensive utilization have attracted much research interest due to the depletion of petroleum resources. The research and utilization of renewable biomass resources have a wide range of benefits, e.g., increasing energy supply and reducing environmental pollution. Vegetable oils and fats are promising biomass feedstock for the production of many competitive products. Among these, epoxidized vegetable oils and fats have been used for many commercial applications, including additives in lubricants,1,2 costabilizer and plasticizers in polymers,3 stabilizers in chlorine-containing resins,4 pharmaceuticals,5 and biofuel additives.6 In current practice, the industrial scale epoxidation of vegetable oils and fats, such as fatty acid methyl esters (FAMEs), is carried out by reacting the double bonds of oils with a percarboxylic acid (generally peracetic acid or performic acid) generated in situ in a stirred-tank reactor.7 This reaction involves two immiscible phase and is extremely exothermic with enthalpy up to −196 kJ/mol.8 Mass transfer and heat removal capacity are the two limiting factors for the improvement of the conventional stirred-tank reactor conversion rate. To prevent excessive temperature increase in conventional industrial reactors, H2O2 is added slowly (typically over 2 h) to the oil at a moderate temperature (normally 10 °C below the reaction temperature); hence, the reaction requires about 10−12 h to be completed. In the last decades, some papers covered the intrinsic kinetics of epoxidation reaction, considering the two-phase system as a © 2017 American Chemical Society

Received: Revised: Accepted: Published: 5231

March 21, 2017 April 11, 2017 April 20, 2017 April 20, 2017 DOI: 10.1021/acs.iecr.7b01186 Ind. Eng. Chem. Res. 2017, 56, 5231−5238

Article

Industrial & Engineering Chemistry Research studied the epoxidation of grape seed oil by peracetic acid and found the optimal reaction conditions of 90 °C and 60 min. In addition, De Quadros Jr. et al.8 found that the process of soybean oil epoxidation was simpler and faster but still safe when all reactants were added at once under the conditions of a highly effective heat removal system. Thus, it can be seen that higher requirements are put forward for the heat-transfer capacity of the reactor. Nevertheless, further investigation is needed to break through the bottlenecks (mass and heat transfer) of the conventional epoxidation process. According to the characteristics of the vegetable oil epoxidation reaction, a more efficient and safer process with new types of reactors should be developed. Two problems that need to be solved are how to improve the liquid−liquid rate of mass transfer and how to enhance reaction heat transfer. Hydrodynamic cavitation in chemistry is an effective method for process intensification. The cavitation is generated by the flow of a liquid through a simple geometry such as venturi tubes or orifice plates under controlled conditions. When the pressure at the throat falls below the vapor pressure of the liquid, the liquid flashes, generating a number of cavities, which subsequently collapse when the pressure recovers downstream of the mechanical constriction.24 Cavitational reactors have a widespread application for the chemical and physical effects of hydrodynamic cavitation.25−29 The major physical effects of cavitation include (1) generation of shock waves, (2) liquid microjet formation, and (3) interfacial turbulence,30 which are important to increase the mass- and heat-transfer rates, particularly in heterogeneous systems. Therefore, hydrodynamic cavitation is attractive for the design of an epoxidation reactor. The main contents of this research include (1) design of a novel reactor which is suitable for exothermic heterogeneous reaction system; (2) measurement of droplet size distribution (DSD) in the reactor using photomicrography techniques; (3) evaluation of the mass- and heat-transfer performance of the reactor; and (4) epoxidation performance of this novel reactor and comparison with conventional reactor.

Figure 1. Schematic diagram of the experimental apparatus.

produced while mixing the oil phase and aqueous phase. The temperature control of the reactor was realized by recirculating thermostatic water inside the reactor jacket, and the temperature was monitored using a thermocouple in the outlet of the reactor. The inlet pressure was measured by a pressure gauge in the outlet of the pump. 2.3. Experimental Methodology. The mass-transfer performance of this novel reactor can be examined by measuring the DSD in the system. There are many methods to measure the drop size in liquid−liquid dispersions such as laser granulometry, video technique, etc.31 In this research, a photographic technique based on a microscope (MX-50, Olympus) was used to measure the DSD. As a consequence of the high exothermic epoxidation reaction, the oil-phase composition changes with the reaction time; hence, it is difficult to measure the drop size in the real reaction system directly. Therefore, a similar system consisting of oil phase [FAMEs \epoxy fatty acid methyl esters (EFAMEs)\byproducts, separated from the reaction system at 4 h] and aqueous phase (water) was selected to replace the real reaction system for the measurement of DSD. To obtain a more stable emulsion, a moderate amount of nonionic surfactant (S-80) was added to the water/oil system, which will stabilized the extracted sample against further coalescence and breakup. To determine the DSD, sample taken from water/oil system was coated on a glass slide quickly and evenly and then shot through the eyepiece by camera. The DSD was determined from image processing as follows: A threshold was set on the gray value of the pixels to detect the droplet edges in the image. Subsequently, a count of 500 droplets was manually performed to obtain a statistically reliable DSD result. In this study, the DSD in the novel reactor was investigated at four different inlet pressures (0, 0.5, 1.5, and 2.5 bar) with circulation flux of 80 mL/min. Additionally, for comparison proposes, the DSD in a stirred-tank reactor (a round-bottom flask without baffle; the diameter of the two-impeller used in the work was 5 cm) was investigated at four different agitation speeds (200, 300, 400, and 500 rpm). The amount of oil used is 50 g. The hold up (volume fraction) of water is 0.3. The amount of S-80 used is 1.2 wt % of the oil weight. The laboratory temperature was maintained at 25 °C to measure the DSD. Figure 2 shows the typical drops taken by the MX-50 instrument. All experiments

2. EXPERIMENTAL SECTION 2.1. Materials. Refined FAMEs (iodine value, 125 g I2/100 g oil) was produced from soybean oil which was locally purchased. H2O2 (50 wt %) was provided by Jiangyin Xiangyang Technology Co., Ltd. Sorbitan fatty acid ester (S80, nonionic surfactant) was purchased from Sinopharm chemical reagent Co., Ltd., and all reagents used in the experiments, including formic acid (88 wt %), were analysisgrade. Unless otherwise stated, all chemicals were directly used without further treatment. 2.2. Experimental Setup. The schematic diagram of the experimental setup is shown in Figure 1. The epoxidation reactions were carried out in a circulatory system. This system includes a jacketed glass tube (11 mm diameter and 270 mm length); hence, the heat-transfer area is equal to 0.0093 m2, and the pall ring (PTFE, ⌀ = 3 mm, void fraction is 0.69) was used in the tube as the fillers. Therefore, this structure can be considered to be a static mixer. A tundish (40 mm diameter and 80 mm length) is necessary to the feeding and operation, and a peristaltic pump and a ball valve (⌀ = 6 mm, equal to the external diameter of the pipeline, the inner diameter of the pipeline is 4 mm) are also used. During experimentation, the open angle of the ball valve was adjusted to keep the inlet pressure to a fixed value, and then cavitation would be 5232

DOI: 10.1021/acs.iecr.7b01186 Ind. Eng. Chem. Res. 2017, 56, 5231−5238

Article

Industrial & Engineering Chemistry Research

certain amount of FAMEs was added into the reactor through the tundish, followed by the addition of formic acid. The mixture of FAMEs and formic acid was circulated in the pipeline driven by the pump. The inlet pressure was adjusted by controlling the opening of the ball valve. Aqueous H2O2 was added into the system within less than 1 min when the temperature reached the reaction temperature. The temperature was maintained within ±1 °C by recycling water of a thermostatic bath. Samples were periodically taken from the reactor, with the aqueous phase separated with the oil phase by centrifuge (10 000 rpm). The oil phase was then exhaustively washed with deionized water and dried using a rotary evaporator, followed by measurement of iodine and epoxy values. 2.4. Analytical Techniques. 2.4.1. Iodine Value. The iodine value describes the amount of double bonds in fatty acids and is expressed as grams of iodine consumed by 100 g of oil. It was determined according to the GB/T 1676-2008 standard method. Briefly, the oil first reacted with excessive iodine in the form of iodine monobromide. The potassium iodine solution was then added and reacted with remaining iodine monobromide to form simple iodine, which was titrated with sodium thiosulfate solution. The iodine value was calculated by the subtraction of the total iodine added and the iodine titrated by the sodium thiosulfate solution.32 2.4.2. Epoxy Value. The epoxy value describes the amount of epoxy groups in the oil and is expressed as grams of oxygen in epoxy ethane groups consumed by 100 g of oil. It was determined according to GB/T 1677-2008 standard method. Briefly, a direct titration of the epoxy groups was conducted with hydrochloric acid−acetone solution.33

were performed three times to test reproducibility and repeatability of the experiments.

Figure 2. Drops in the water/oil system.

The heat-transfer performance of this novel reactor also needs to be examined. The heat-transfer coefficient is a direct embodiment of the heat-transfer capacity of a device; therefore, it is an intuitive method for measuring the heat-transfer coefficient of the apparatus and for comparison with the traditional reactore. As mentioned in section 2.2, the heattransfer area of the reactor is 0.0093 m2, and a simple method was used to determine the heat-transfer coefficient of the reactor. The reaction medium and cooling water flowed past the tube under conditions similar to those used in the reaction. Then the heat-transfer coefficient could be determined by measuring the temperatures of the reactor medium and cooling water at the import and export of the reactor. Temperature is an important factor that influences the rate of epoxidation of vegetable oil. Increasing temperature increases reaction rate of epoxidation and also the reaction rates of side product formation. Therefore, there must be a balance between maximizing oxirane formation and minimizing side reactions when selecting optimal temperature. Selecting and maintaining a constant temperature is important, but it is not easy because epoxidation is exothermic. Generally, the H2O2 is added slowly to the reaction system to avoid out-of-control temperatures. In the new apparatus, a calculated amount of H2O2 will be added in less than 1 min to the system with three different reaction temperatures (50, 60, or 70 °C); the temperature variation is then recorded using paperless recorder. The calculated amount of H2O2 is added to the stirred-tank reactor in three different times (30, 60, or 90 min), then the temperature variation is recorded. When the mass- and heat-transfer performance of this reactor is confirmed, the epoxidation of FAMEs will be conducted in this novel reactor. The epoxidation of vegetable oil have been widely investigated,10,15,16 including the effect of temperature, the carboxylic acid to ethylenic unsaturation molar ratio, and hydrogen peroxide to ethylenic unsaturation molar ratio. On the basis of the information from literature and industry parameters,8 all reactions were performed at the temperature of 60 °C, formic acid to ethylenic unsaturation molar ratio of 0.23, and H2O2 to ethylenic unsaturation molar ratio of 1.5. The epoxidation reaction was conducted in the hydrodynamic cavitation reactor, and the effect of inlet pressure (0, 0.5, 1.5, and 2.5 bar) on the reaction was examined at circulation flux with 80 mL/min. The experimental procedure was as follows: A

3. RESULTS AND DISCUSSION 3.1. Droplet Size Distribution in This Novel Reactor. Several probability density functions, such as linear normal distribution,34 logarithmic normal distribution,35 and Frechet distribution,36 can be used to describe the DSD. The DSD dispersion with dominant droplet breakage is described by a logarithmic normal distribution assuming binary breakage, which agreed well with a large amount of experimental data.37,38 The water/oil system is a viscous dispersion with low interfacial tension, which most matched the characteristics of breakage-dominated system. Thus, the logarithmic normal distribution (eq 1) was applied: ⎛ (In(X /X̅ ))2 ⎞ 1 exp⎜ − ⎟ 2π σX 2σ 2 ⎠ ⎝

fd (X ) =

(1)

where X is the droplet size, σ the standard deviation, and X̅ the median of distribution. The coefficients of each distribution were obtained by nonlinear least-squares data fitting by the Gauss−Newton method. Then the Sauter mean diameter (d32) is determined as defined in eq 2:37

d32 =

∫d ∫d

dmax 3

d p(d)dd

min

dmax 2

d p(d)dd

min

(2)

where dmax and dmin denote the maximum and the minimum droplet size, respectively, and p(d) is the probability density function of the droplet size. 5233

DOI: 10.1021/acs.iecr.7b01186 Ind. Eng. Chem. Res. 2017, 56, 5231−5238

Article

Industrial & Engineering Chemistry Research

Figure 3. Measured DSD and fitted DSD of water dispersions with different inlet pressures: (a) 0 bar, (b) 0.5 bar, (c) 1.5 bar, and (d) 2.5 bar.

Normalized experimental DSD with fitted distributions at various inlet pressures are shown in Figure 3. The regressed parameters of the size distribution function are listed in Table 1. The DSD plots showed that increased inlet pressure led to

confirmed by the model of bubble motion proposed by Gogate and Pandit24 and Moholkar and Pandit.40 Clearly, the cavities were not generated at inlet pressure 0.5 bar, and this is also confirmed by the large droplet size and wide size distribution. The droplet size at 0.5 bar is smaller as compared to that at 0 bar, which could be due to the presence of dissolved gases, which act as weak spots in the liquid providing nuclei for cavitation.24 Kumar et al.41 reported the flow regime maps on the basis of radial motion of the bubbles in the flow; for all values of orifice to pipe diameter ratios considered in simulations, cavitation bubbles show only moderately transient behavior (transient inducing sonophysical effect) for Cv = 1 and p2 = 101.3 kPa. This result indicated that the small droplet size and narrow size distribution at inlet pressure 1.5 and 2.5 bar were caused by the sonophysical effect of collapse of cavities. Amin et al.26 optimized a hydrodynamic cavitation reactor using salicylic acid dosimetry and found that absolutely no free radical formation can be seen at pressures below 6.8 bar. This result indicated the pressure pulses generated at 2.5 bar in the present work are too small to release free radicals. From the aspect of the DSD at various inlet pressures, the physical effects contributed by this relatively low cavitation intensity is sufficient for the mixing requirement of the oil and aqueous phases. Additionally, large droplet size was observed for aqueous/oil dispersion without hydrodynamic cavitation, i.e., when the inlet pressure was 0 bar (Figure 3a). In this case, the reactor was equivalent to a packed-bed reactor. This result indicated that it was difficult to achieve thorough mixing for this liquid−liquid system relying solely on disturbance of packing in the reactor. The normalized experimental DSDs with fitted distributions at various agitation speeds are shown in Figure 4. The regressed parameters of the size distribution function are listed in Table 2. It is noteworthy that a significant decrease in droplet size can be found when the agitation speed increased from 200 to 300 rpm, yet further increase in agitation speed led to small change

Table 1. Fitted Parameters of Lognormal Distribution Function inlet pressure (bar)

0

0.5

1.5

2.5

median droplet size X̅ (um) standard deviation σ (um) R2

105.79 0.3726 0.9750

76.02 0.3065 0.9897

9.82 0.4131 0.9993

7.71 0.1566 0.9982

smaller droplet size and narrower size distribution. The median of the DSD at relatively high inlet pressure (1.5 bar, 2.5 bar) (Figure 3c,d) was much smaller than that generated at the pressure of 0.5 bar (Figure 3b). This phenomenon can be explained based on the cavitation number, which is expressed as27 C V = (p2 − pv )/(0.5ρu 2)

(3)

where p2 is the fully recovered pressure downstream of the orifice; pv is the saturation vapor pressure of water, which is equal to 3.17 kPa at 25 °C; u is the average velocity of liquid at the orifice; and ρ is the density of water. In the present work, p2 is always equal to ambient pressure (101.3 kPa), and u can be determined by applying the Bernoulli equation between upstream and downstream, resulting in pinlet + 1.5 × 0.5ρu 2 = p2

(4)

The value of 1.5 in eq 4 is local resistance coefficient of the channel of valve. Then the value of u was determined, which was 8.16, 14.14, and 18.26 m/s at 0.5, 1.5, and 2.5 bar, respectively. Then the cavitation number was determined as 2.94, 0.98, and 0.59 at 0.5, 1.5, and 2.5 bar, respectively. Under ideal conditions, cavitation occurs for CV < 1.39 This is also 5234

DOI: 10.1021/acs.iecr.7b01186 Ind. Eng. Chem. Res. 2017, 56, 5231−5238

Article

Industrial & Engineering Chemistry Research

Figure 4. Measured DSD and fitted DSD of water dispersions with different agitation speeds: (a) 200 rpm, (b) 300 rpm, (c) 400 rpm, and (d) 500 rpm.

import and export of the reactor. The heat-transfer rate of the device was then expressed as

Table 2. Fitted Parameters of Lognormal Distribution Function agitation speed (rpm)

200

300

400

500

median droplet size X̅ (um) standard deviation σ (um) R2

137.35 0.59 0.9523

83.91 0.29 0.9671

62.39 0.28 0.9594

56.75 0.20 0.9710

out in Q htr = m ·c P·(Trec − Trec ) = K ·Δtm·A

(5)

where Δtm is the logarithmic mean temperature difference, calculated as Δtm = (Δt 2 − Δt1)/In(Δt 2/Δt1)

(6)

where Δt2 = − Δt1 = − and Tinwat, out and Twat are 62, 39, 28, and 28 °C, respectively (the water flow rate in the jacket ≫ the flow rate of reactant in the reactor, so the temperature of water was approximately constant). Based on the values of mass flow rate of content in the reactor (m = 1.2 g/s), specific heat capacity of reaction mixture [cP = 1.98 kJ/ (kg·K)],8 and heat-transfer area (A = 0.0093 m2), the heattransfer coefficient was estimated as K = 288 W/(m2·K). This value is extremely large for a liquid−liquid heat-transfer process without phase change. The large heat-transfer coefficient in this hydrodynamic cavitation reactor can be attributed to two Tout rec

in droplet size. However, the drop size produced by cavitation was much smaller than the one produced by agitation. In the present study, the practical power of the agitator (500 rpm) and pump (1.5 bar, 80 mL/min) approximately calculated based on operation conditions were 7.20 and 3.92 W per kilogram of oil, respectively. Therefore, it can be said that the efficiency of introducing energy into the system of hydrodynamic cavitation is higher than that of mechanical agitation. 3.2. Determination of the Heat-Transfer Coefficient. The heat-transfer coefficient was determined by measuring the temperatures of the reactor medium and circulating water at the

Tout wat;

Tinrec

Tinwat;

Tinrec,

Tout rec ,

Figure 5. Temperature changes in a stirred-tank reactor (a) and hydrodynamic cavitation reactor (b). 5235

DOI: 10.1021/acs.iecr.7b01186 Ind. Eng. Chem. Res. 2017, 56, 5231−5238

Article

Industrial & Engineering Chemistry Research aspects. On the one hand, the cold and hot medium in the static mixer heat exchanger device were in a state of turbulence, which improve the heat-transfer coefficient; on the other hand, the presence of hydrodynamic cavitation and filler in the tube reduced the particle size of the dispersion phase, which increased the turbulence intensity and enhanced the heattransfer coefficient. In a traditional stirred-tank reactor, the reaction heat is removed by cooling water coil which is submerged in a reaction medium. The heat-transfer coefficient, in a traditional industry reactor, is normally equal to about 100 W/(m2·K); this value is estimated based on the data provide by Anhui Tianyi Environmental Protection Technology Co., Ltd. In such an industry-scale stirred-tank reactor, reaction medium flows outside the cooling water coil in a low velocity (the agitation speed is controlled between 60 and 120 rpm) with big and thick volume. Hence, the heat-transfer coefficient is very low. Meanwhile, such heat-transfer mode is not sensitive to operating conditions. Even in a laboratory scale (a 500 mL glass reactor), with an intensified heat-transfer condition of heat transfer, the heat-transfer coefficient of a stirred-tank reactor is about 186 W/(m2·K).8 3.3. Temperature−Time Profile of This Novel Epoxidation Reactor. The temperature behavior for reactions in a stirred-tank and this novel reactor is shown in Figure 5. Figure 5a shows the temperature behavior in a stirred-tank reactor with three different H2O2 addition rates at reaction temperature 60 °C. The temperature−time profiles are similar irrespective of the addition rate; however, the temperature profiles are higher and steeper at the fastest addition rate (H2O2 was added in 30 min) compared to the slower addition rates (H2O2 was added in 60 or 90 min). Similar results were reported in the literature.8,42 The maximum temperature represents the point at which the heat release rate equals the heat release rate. The steepest slope before maximum height indicated that the highest reaction rate throughout the reaction. It should be noted that the temperature profiles were obtained under labscale conditions; however, under industry-scale conditions, the temperature variation would be even larger because of the amplification effect. Figure 5b shows the temperature behavior in the novel reactor at three different reaction temperatures with H2O2 addition in less than 1 min. The variation of reaction temperature can be controlled within 1 °C even with H2O2 added to the system at once. This means the heat release rate equals the heat release rate throughout the reaction with reaction conditions of this work. The determination of heat-transfer area is crucial to the operation of a reactor. In the present work, heat release rate can be calculated based on the declining rate of the iodine value. Figure 6 shows the iodine value variation at reaction temperature of 60 °C with H2O2 added within less than 1 min. The slope of the red line in Figure 6 corresponds to the maximum iodine decline rate and also maximum heat release rate. The corresponding heat release rate can be calculated as Q hrr = 10·moil ·ΔH ·dIV /(M I2 ·d t)

Figure 6. Profile of iodine value with reaction time.

as 30 °C). This method to determine the heat-transfer area is useful to the scale-up of the novel reactor. 3.4. Comparison of Reaction Performance between This Novel Reactor and Stirred-Tank Reactor. It is worthwhile to compare the results of conventional and the hydrodynamic cavitation approach in order to find the efficacy of one over the other. Vegetable oil epoxidation was conducted in hydrodynamic cavitation and stirred-tank reactors, and the results are shown in Table 3. As stated earlier, all these runs were carried out at 60 °C, H2O2 to unsaturated bonds ratio of 1.5:1, and formic acid to unsaturated bonds ratio of 0.23:1. The large temperature rise for epoxidation reaction in a stirred-tank reactor could be an obstacle to industrial application of the system, because larger reactor size and quantities of reactants would result in even higher temperature. Different from the stirred-tank reactor, the heat removal of this novel reactor was efficient even at the beginning of the reaction with the H2O2 added to the system at once, indicating an effective temperature control. The superior heat-transfer performance of the hydrodynamic cavitation reactor leads to a simpler and safer system, making it more suitable for largescale application. The Sauter mean diameter of droplet size was much smaller with the effect of hydrodynamic cavitation as compared to mechanical agitation. The conversion values, in terms of epoxy value and iodine value, were experimentally measured to assess the quality of the final product. Normally, a product with high epoxy value and low iodine value is considered to be better. Table 3 presents these values at different reaction conditions. The final iodine value of the product is 12.34 with inlet pressure 0 bar, i.e., without the presence of cavitation in the system. This result indicated that the presence of the cavitation element is required for the generation of cavitation and the useful physical effect of the same. Although the total reaction time in the novel reactor is lower, the quality expressed by the epoxy and iodine indexes are similar to those of a typical traditional process. These results further illustrate that the collapse of cavities will have scarcely any chemical effect, while the physical effect caused by it can satisfy the mixing requirements of oil and aqueous phases. More importantly, the hydrodynamic cavitation reactor has immediate and realistic potential for industrial scale applications, and the scale up of these reactors is comparatively easy because vast amounts of information about the fluid dynamics downstream of the constriction are readily available; in addition, the operating efficiency of the circulating pumps, which is the only energy dissipating device in the system, is always higher in large-scale operation.

(7)

where moil is the quantity of the oil phase in the reaction system (kg), ΔH the reaction enthalpy (196 kJ/mol), and dIV/dt the maximum iodine value decline rate (0.8 min−1). The value of the maximum heat release rate was calculated to be Qhrr = 6.2 kJ·kg−1·min−1. Then the heat-transfer area needed is 0.3588 m2/kg (the heat-transfer temperature difference was calculated 5236

DOI: 10.1021/acs.iecr.7b01186 Ind. Eng. Chem. Res. 2017, 56, 5231−5238

Article

Industrial & Engineering Chemistry Research Table 3. Vegetable Oil Epoxidation in Stirred-Tank and Hydrodynamic Cavitation Reactors inlet pressure and agitation speed

total reaction time (min)

peroxide addition time (min)

Sauter mean diameter (μm)

temperature rise (°C)

epoxy value (%)

iodine value (%)

0 bar HC-0.5 bar HC-1.5 bar HC-2.5 bar MS-400 rpm MS-400 rpm MS-400 rpm

480 480 480 480 600 600 600