Chemical and Technical Aspects of the Synthesis of Chlorohydrins

Review pubs.acs.org/IECR

Chemical and Technical Aspects of the Synthesis of Chlorohydrins from Glycerol E. Santacesaria,* R. Vitiello, R. Tesser, V. Russo, R. Turco, and M. Di Serio Dipartimento di Scienze ChimicheUniversity of Naples FEDERICO II − NICL − Naples Industrial Chemistry Laboratory, Complesso di M.te S. Angelo, Via Cintia, 80126 Napoli, Italy ABSTRACT: In the synthesis of biodiesel via the transesterification of vegetable oils, 10 wt % of glycerol is obtained as byproduct. This means that, by increasing the biodiesel production, the glycerol availability also increases and its cost goes down more and more. In order to consume the large amount of glycerol derived from biodiesel production in a profitable way, only two strategies are possible: (i) use glycerol as raw material to produce fuel additives and (ii) use glycerol as raw material to produce commodities. In the present work, we have briefly considered the first aspect while focusing, in particular, on the second opportunity by reviewing the production of chlorohydrins by glycerol hydrochlorination with HCl. Chlorohydrins are important intermediates in the production of epichlorohydrin used to produce epoxy−resins. The advantages of producing chlorohydrins by starting from glycerol instead of propenethat is, the classical routewill be discussed. The glycerol hydrochlorination reaction is catalyzed by carboxylic acids, and in this work, we describe (i) the reaction conditions normally adopted; (ii) the behavior of different catalysts proposed in the literature (concerning activity and selectivity); (iii) the reaction mechanism; (iv) the kinetic laws, reported by different authors, along with the related parameters; and (v) the role of mass transfer. A brief discussion on the best reactors for performing the reaction and some information about the different processes used to produce epichlorohydrin starting from glycerol will also be reported. Some catalysts, other than carboxylic acids, have also been briefly reviewed, although they have not been used in industrial plants until recently.

1. INTRODUCTION Biodiesel production has strongly increased recently as one of the possible substitutes of diesel from petroleum. In perspective, biodiesel could become convenient if the price of petroleum goes up, as a consequence of both an increase in the consumption and a decrease in the availability. This energy source, covering a relatively small segment of the human energetic needs, will be in concurrence also with diesel that can be obtained with more- or less-complicated processes also from coal, natural gas, or biomasses. As known, crude glycerol is an important byproduct of biodiesel synthesis, corresponding to ∼10 wt % of the produced biodiesel.1 Clearly, by increasing the biodiesel production, the availability of glycerol also increases and, consequently, its price goes down more and more. In the past, glycerol has found applications in several fields, such as foods, cosmetics and personal care, pharmaceuticals and drugs, polyethers/polyols, explosives, alkyd resins, triacetin, detergents, cellophane, and tobacco industries. However, the mentioned markets are almost saturated and cannot absorb the great amounts of glycerol coming from the increasing production of biodiesel. Glycerol often is burned by the biodiesel producers to obtain the energy to be used inside the biodiesel production process. Therefore, in the past decade, intense research activity has been developed worldwide to find new profitable uses for glycerol, and many different proposals can be found in some reviews on the topic that recently appeared in the literature.2,3 Clearly, finding new profitable uses for glycerol is also useful for decreasing the biodiesel production costs. Therefore, several possible new uses of glycerol as feedstock are briefly summarized in the next section. 1.1. New Potential Routes of Glycerol Chemical Transformation. Considering the large amounts of glycerol © 2013 American Chemical Society

that could potentially be obtained as byproduct from biodiesel plants, only two acceptable strategies can be followed:1 (i) the production in large amount of oxygenated additives for biofuels and (ii) the use of glycerol as raw material for obtaining commodities. However, any economical forecast about the convenience of using glycerol as raw material in both of the mentioned fields is a very complicated matter, because, while, biodiesel production is often sustained by government subsidies for strategic reasons, glycerol is rigorously subjected to the market laws. As it will be seen, glycerol can be used in some important processes as a raw material alternative to propene; therefore, the cost of propene can be considered as a reference point for establishing the economic convenience of using glycerol as raw material for producing fuel additives or chemicals. Ultimately, glycerol obtained as a byproduct of biodiesel by following the conventional route, based on the use of a homogeneous alkaline catalyst, is impure and requires an expensive purification procedure before any use as raw material. For this reason, many studies have been devoted worldwide to the development of new heterogeneous catalysts, promoting the transesterification reaction with the attainment of pure glycerol.1 Many different substances, synthesized from glycerol, can be used as blending components for fuels, such as ethers (glycerol isobutylethers),4−6 esters (triacetin),7 acetals,8,9 and ketals.9 Special Issue: Massimo Morbidelli Festschrift Received: Revised: Accepted: Published: 8939

October 2, 2013 December 2, 2013 December 4, 2013 December 4, 2013 dx.doi.org/10.1021/ie403268b | Ind. Eng. Chem. Res. 2014, 53, 8939−8962

Industrial & Engineering Chemistry Research

Review

an intermediate in epichlorohydrin synthesis. The production of epichlorohydrin occurs through the following reaction steps:

In particular, etherification to obtain a mixture of di- and triisobutyl ethers (GTBE) is the most promising reaction. GTBE is a good additive for diesel (both fossil and biodiesel) and also for gasoline as an octane booster. In diesel and biodiesel (7.5 wt %), it will lead to a reduction in the emissions of particulates, NOx, and unburned hydrocarbons. Moreover, blending diesel or biodiesel with GTBE also reduces the viscosity, cloud point, and pour point, but it reduces the calorific power somewhat. As previously mentioned, another possible use of the big amounts of glycerol potentially available in the world is the production of commodities. Examples of this type are (i) the hydrochlorination of glycerol to obtain chlorohydrins,10−12 which are useful intermediates for producing epychlorodrin to be used in the production of epoxy resins; and (ii) the dehydration to acrolein, followed by the oxidation to acrylic acid,13,14 via the following two-step reaction:

Instead of NaOH, Ca(OH)2 also can be used. Epichlorohydrin is an important raw material for the production of some polymers such as epoxide resins, synthetic elastomers, and sizing agents for the papermaking industry. Starting from glycerol, the reaction proceeds through a first hydrochlorination reaction, primarily forming 1-chloro-2,3-propanediol (1-MCH) and water, along with small amounts of the isomer 2-chloro1,3-propanediol (2-MCH) (monochlorohydrins). This is followed by a second hydrochlorination step from which 1,3DCH is obtained as the main product, together with modest amounts of 1,2-chloro-3-propanol (1,2-DCH). Another interesting aspect is that, in some cases, crude glycerol could probably be used in the reaction, with evident economic advantage, although this possibility has not been studied enough. Only one paper has been published on the subject by Kruper et al.39 However, it can be foreseen that some problems will arise in continuous plants for the necessity of a purge during the recycle of the catalyst for eliminating the impurities contained in crude glycerol coming from biodiesel production (normally potassium or sodium salts). Different carboxylic acids can be used as catalysts. The hydrochlorination of glycerol is very selective in giving 1,3DCH. In contrast, starting from propene and following the traditional technology, a mixture of 1,3-DCH and 1,2- DCH (30:70%) is obtained. This is an important advantage of the process via glycerol, because 1,3-DCH is much more reactive than 1,2-DCH and, consequently, the plants for obtaining epychlorohydrin, in this case, are much smaller in size and, hence, less expensive. In this review, the activity and selectivity shown by different carboxylic acids used as catalysts in promoting the reaction and the effect of the catalyst concentration have been examined and compared; then, the reaction mechanisms suggested in the literature by different authors have been compared and discussed; at last, the importance of using gaseous hydrochloric acid and the role of both the HCl pressure and the temperature on the reaction rate have been examined and discussed. In particular, all the kinetic aspects of the reaction, in relation with the suggested reaction mechanisms, will be discussed in detail. Finally, some suggestions will be given about the process scaleup and the possible structure of the hydrochlorination reactors.

The second reaction step of reaction scheme described by reaction 1 is a well-known technology. In fact, acrylic acid is commonly produced by propene in two oxidation steps: the first giving acrolein and the second one giving acrylic acid using two different catalysts. Therefore, starting from glycerol, the second step is exactly the same as that observed for propene. In both of the described examples, glycerol substitutes for propene as raw material and clearly the convenience of the new processes is related to the costs of, respectively, propene and glycerol. Other reactions are obviously possible and it has also been reported that glycerol can be thermochemically converted to propylene glycol,15−17 acetol, or a variety of other products. 18 An aqueous phase reforming process that transforms glycerol to hydrogen also has been developed.19 Another possibility is the biological conversions of crude glycerol, because glycerol represents a good feedstock in various fermentation processes. For example, glycerol has been used in the fermentation of Anaerobiospirillum succiniciproducens for the production of succinic acid,20 or it can be converted to citric acid by using the yeast Yarrowia lipolytica. In particular, it has been reported that this organism produces the same amount of citric acid when grown on glucose or on raw glycerol.21 The scheme described by Figure 1 shows the most relevant reactions in which glycerol can be involved as a new building block.1 1.2. Production of Chlorohydrins from Glycerol Hydrochlorination. In this work, we focused our attention to the glycerol hydrochlorination process for producing chlorohydrins with the aim to review the “state of the art” of this technology. The strong industrial interest for this production route is confirmed by the fact that recently some big companies have announced plans to commercialize their technology to manufacture epichlorohydrin starting from glycerol.22 Studies of glycerol hydrochlorination have focused mainly on the production of 1,3-dichloro-2-propanol (1,3-DCH), which is 8940

dx.doi.org/10.1021/ie403268b | Ind. Eng. Chem. Res. 2014, 53, 8939−8962


Review

Figure 1. Glycerol as a new building block; examples of the reactions involving glycerol as feedstock are shown.

2. COMPARISON OF THE NEW PROCESS VIA GLYCEROL FOR PRODUCING EPICHLOROHYDRIN WITH THE TRADITIONAL PROCESS VIA PROPENE The process for the synthesis of epichlorohydrin is a rather old process. The classical process starts from propene and, through a first step, consisting of a high-temperature chlorination, allyl chloride is obtained;23 subsequently, combining allyl chloride with hypochlorous acid gives glycerol dichlorohydrins isomers, and, finally, the reaction of the glycerol dichlorohydrins with sodium hydroxide or calcium hydroxide leads to epichlorohydrin. The main reactions involved in the epichlorohydrin synthesis23 are shown in Figure 2. As it can be seen, from the second step of this process, a mixture of dichlohydrins is obtainedmore precisely, the reaction product is a mixture composed of 1,2-DCH (70%) and 1,3-DCH (30%). As already mentioned, the reactivity of 1,2-DCH, which is the most abundant component, is much lower and the difference in the reactivity has already been explained, based on the reaction mechanism.24 This gives the following disadvantages: (i) an increase in reactor size (reactive distillation column) is necessary to obtain satisfactory conversion of the less-reactive reactant and (ii) the formation of undesired byproduct occurs, as a consequence of the long residence time.24,25 This process also has the disadvantage of using an abundant amount of chlorine. The availability of great amounts of glycerol, at low cost, stimulated the development of alternative processes, based on the use of this substance as raw material instead of propene. It must be recognized that the process based on glycerol hydrochlorination, to obtain chlorohydrins, has been known for a long time.26−33 The old process involved the reaction of glycerol with aqueous hydrochloric acid at high concentration, using acetic acid as a catalyst. The reaction was performed at 80−100 °C. The process, which is based on the use of aqueous HCl, is affected by many drawbacks, such as the following:

Figure 2. Reaction scheme for the traditional process via propene.

(i) The reaction is slow because of the abundant amount of water present in the system, coming from both the aqueous HCl and the water produced by the reaction; (ii) Separation of the dichlorohydrins from the reaction mixture is difficult; and (iii) Acetic acid, used as a catalyst, can be lost from the reaction mixture by evaporation, since its boiling point is relatively low (117 °C). The disadvantages listed above discouraged the possibility of further studies in this direction. In contrast, new studies have been directed toward processes using gaseous HCl instead of 8941



Review

Figure 3. Overall reaction scheme from glycerol to epichlorohydrin, proposed by Santacesaria et al.10−12,24,25

3. CATALYSTS PROPOSED FOR PROMOTING THE GLYCEROL HYDROCHLORINATION REACTION

aqueous solutions and new catalysts (less volatile than acetic acid, such as carboxylic acids of higher molecular weight). In this way, it is possible to keep the concentration of the catalyst inside the reactor constant during the reaction.10−12 The use of organic acids less volatile than acetic acid allows operation at higher temperatures, thus increasing the reaction rate significantly. The greater cost of the less-volatile organic acids, with respect to acetic acid, is probably compensated by the minor loss of catalyst for volatilization during the process. The use of gaseous HCl also allows easier recovery of the dichlorohydrins produced from the reaction mixture, because 1,3-DCH and 1,2-DCH are both more volatile than both the unreacted glycerol and the intermediate monochlorohydrins 1-MCH and 2-MCH. Gaseous HCl can be fed to the reaction system either pure or diluted with an inert gas.10−12 The process for producing dichlorohydrins, according to this method, is performed starting from pure or crude glycerol, as a byproduct of biodiesel production, treated with gaseous hydrochloric acid, in the presence of a carboxylic acid. The pressure of HCl has a positive effect on both the reaction rate and selectivity toward 1,3-DCH, and it is usually maintained in the range of 1−10 bar. The complete conversion of glycerol, however, involves a reaction time that is dependent on the pressure of HCl, the adopted temperature, and the catalyst concentration. This time varies from 2 h to 24 h, depending on the adopted operative conditions. However, different patents34−41 and publications10−12,42−49 have appeared in the literature on this subject in recent years, starting from 2005. In particular, Santacesaria et al.,10−12 based on experimental observations, suggested the reaction scheme reported in Figure 3, in which all of the occurring successive-parallel reactions are numbered. As it can be seen, glycerol reacts with HCl in two steps, giving monochlorohydrins in the first step and then dichlorohydrins in the second one. It is interesting to point out that very small amounts of 2-DCH and 1,2-DCH are obtained and, if the reaction time is long enough, 1,3-DCH is obtained with very high yields (85%−95%). Ultimately, both 1,2- and 1,3-dichlorohydrins gives epichlorohydrins. This last reaction is promoted by a basic environment, that is, by contacting dichlorohydrins with an aqueous solution of Ca(OH)2 or NaOH. Some patents have also been published in the literature in which an inert organic solvent is used; this solvent is nonmiscible with water and is souble in dichlorohydrins. In this type of process, the reaction is performed at temperatures below the boiling point of the mixture (usually less than 110 °C).28

As already seen in the scheme reported in the previous section (Figure 3), the reaction between glycerol and gaseous HCl gives monochlorohydrins (mainly 1-MCH and small amounts of 2-MCH) in a first step. Subsequently, hydrochlorination proceeds and the dichlorohydrins (mainly 1,3-DCH and small amounts of 1,2-DCH) are produced. In Table 1, some physicoTable 1. Properties of the Components Involved in the Glycerol Hydrochlorination Reaction glycerol derivate

molecular mass [g/mol]

density [g/cm3]

glycerol α-MCH β-MCH α,γ-DCH α,β-DCH acetic acid

92.09 110.54 110.54 128.99 128.99 60.05

1.261 1.322 1.303 1.364 1.360 1.050

refractive index

solubility in water at 25 °C

boiling point (°C)

1.475 1.480 1.473 1.483 1.4835−1.4855 1.3716

soluble soluble high 15.6 12.7 soluble

290 213 248.5 174 182 117

chemical properties of the chlorinated derivatives of glycerol and of glycerol itself are reported. Some properties of acetic acid, which is the most commonly employed catalyst, also are reported in the same table. As it can be seen, acetic acid as a catalyst has a relatively low boiling point (117 °C), which is near the temperature normally adopted for this reaction, and a loss of catalyst by evaporation can be predicted. The catalytic performances of acetic acid sometimes are also reported by different authors, in comparison with other carboxylic acids. Tesser et al.,12 for example, have used acetic acid as a catalyst with a concentration of 8 mol %, by operating in a hastelloy semibatch stirred reactor at 5.5 bar and 100 °C. Under these conditions, acetic acid gives place to a good performance, more precisely, after 4 h of reaction, glycerol was completely converted, mainly to 1,3-DCH (89.37 mol %). Bell et al.42 also used acetic acid as a catalyst. They have studied the reaction at a pressure of 7.6 bar and temperature of 110 °C with 5 mol % acetic acid, obtaining a total conversion of glycerol after less than 4 h of reaction. The mixture of the reaction products contained ∼93 mol % dichlorohydrins and 6 mol % monochlorohydrins. In this case, at the end of the reaction, the concentration of 2-MCH was higher than that of 1-MCH, because 2-MCH is much less reactive to further chlorination 8942



Review

Table 2. Summary Table Reporting the Performances of Some Catalysts Used in the Literaturea catalyst [organic acids]

ref

pKa

acetic acid malonic acid levulinic acid citric acid succinic acid propionic acid caprilic acid adipic acid pivalic acid benzoic acid oxalic acid fumaric acid tartaric acid maleic acid acetic acid monochloroacetic acid dichloroacetic acid trichloroacetic acid trichloroacetic acid

10 10 36 36 36 36 34 34 10 10 41 41 41 41 12 12 12 10 12

4.79 2.92 4.78 3.15, 4.77 4.24, 5.64 4.79 4.89 4.43, 5.41 4.94 4.2 1.38 3.03, 4.44 3.07 3.15, 5.13 4.79 2.85 1.48 0.7 0.7

reaction condition reflux reflux reflux reflux reflux reflux reflux reflux reflux reflux reflux reflux reflux reflux semibatch semibatch semibatch semibatch reflux

4.5 4.5 4.5 4.5

bar bar bar bar

reaction time [h]

1-MCH [mol %]

2-MCH [mol %]

1,3-DCH [mol %]

1,2-DCH [mol %]

glycerol conversion [%]

3 3 3 3 2.5 3 3 3 3 3 3 3 3 3 4 4 4 4 3

63.01 55.85 60.04 70.45 60.88 49.88

5.98 7.24 6.98 6.49 6.84 8.77

29.82 35.87 32.20 17.61 30.05 41.00

0.33 0.46 0.37 0.26 0.37 0.35

99.1 99.4 99.6 94.8 98.1 100 96.9 89.7 11.9 13.7 70.0 94.9 85.9 93.8 100 86.5 48.0 30.6 17.1

57.2 7.4 9.19 12.38

39.7 82.3 1.65 1.27

1.11 0.00

69.95 93.07 85.86 92.06 0.70 69.98 41.94 26.64 15.08

0.00 0.00 0.01 1.02 0.00 1.74

7.66 6.37 4.71 3.08 2.01

89.37 10.09 1.99 1.47 0.00

2.85 0.33 0.00 0.00 0.00

Some runs have been made under a flowing stream of HCl (24 g min−1) by refluxing the other components, some others in semibatch conditions feeding gaseous HCl at constant pressure. The reaction temperature was always 100 °C. a

A recent patent37 proposed hydrochlorination via the use of a reactor column consisting of a vertical cylinder with an external recirculation of the liquid reaction mixture. The column was filled with glycerol (97.5%), water (0.5%), and acetic acid as a catalyst (2%); such a mixture was recirculated with a flow rate of 5.0 kg/h. Gaseous HCl was fed from the bottom of the reactor column with a flow rate of 4.6 kg/h. In the external recirculation line, a vacuum rectification column had been inserted, downstream from the reactor, from which a stream composed of a mixture of dichlorohydrins, water formed during the reaction, and unreacted HCl were separated, using a flow rate of 9.3 kg/h. The distillation residue is then recycled to the reactor. Finally, a purge of the residue of the distillation products, containing unwanted byproducts, was collected in a tank for waste disposal. Krafft et al. reported, in another recent patent,34 the use of glacial acetic acid as a catalyst. The authors used glycerol and acetic acid (10 mol %) with gaseous hydrochloric acid at 110 °C. They used, for 2 h, a HCl flow rate of 5.2 mol/h, then a flow rate of 3.8 mol/h for an additional 100 min, and finally a flow rate of 1.3 mol/h for 317 min. They obtained, in this way, a glycerol conversion of 99.1% and 1,3-DCH resulted, also in this case, as the main product (>75 mol %). Also, Siano et al.36 reported the use of acetic acid as a catalyst in their patent. The test was carried out at 100 °C by flowing, at atmospheric pressure, gaseous HCl in the reaction environment. A second addition of catalyst during the reaction was made to compensate for the loss of catalyst caused by the stripping effect of HCl during the reaction. The duration of the reaction was 5 h, but already in the first hour, a high rate of 1,3DCH formation was observed. After 2 h of reaction, the conversion of glycerol was almost complete, but the amount of 1,3-DCH formed was 20 mol %; meanwhile, after 5 h of reaction, the amount of 1,3-DCH formed was ∼70 mol %. The advantage of operating in a flowing stream of gaseous HCl is related to the high volatility of 1,3-DCH under conditions that allow one to recover this product by condensation from the flowing stream. However, as mentioned previously, acetic acid

than 1-MCH; therefore, the amount of 2-MCH formed in the initial phase of the reaction (5%−6%) does not change any further with time. Kruper et al.39 also studied the hydrochlorination reaction of glycerol with gaseous hydrochloric acid in the presence of acetic acid as a catalyst. The test was conducted at a pressure of 5.6 bar and temperature of 93 °C for 90 min. After this time, the reaction products obtained were 1,3-DCH (92.6 mol %) and 1,2-DCH (1.7 mol %); both monochlorohydrins (4.4 mol %) and unreacted glycerol (1.0 mol %) also were observed. These results were obtained by using a purified commercial glycerol. Kruper et al.39 also studied the possibility to conduct the chlorination reaction with gaseous HCl, in the presence of acetic acid as a catalyst, but using crude glycerol coming directly from biodiesel plants. The test was conducted at a pressure of 8.3 bar and temperature of 120 °C for 90 min. The reaction products, discharged from the reactor, was a liquid with a suspension of a white solid. After filtration, the filtrate was analyzed by gas chromatography (GC), and the products were 1,3-DCH (95.3 wt %), 1,2-DCH (2.6 wt %), 2-acetoxy-1,3-dichloropropane (0.7 wt %), 1-acetoxy-2,3-dichloropropane (0.1 wt %), and acetoxychloropropanols (0.87 wt %). Luo et al.57 studied the formation of dichlorohydrins from glycerol and HCl, in the presence of acetic acid as a catalyst, but using an aqueous HCl solution. They performed the reaction in a well-stirred batch reactor with a capacity of 500 cm3. The authors evaluated the trend, relative to the time of reagent and product concentrations, and observed that the amount of glycerol decreased, reaching a minimum after ∼100 min, while 1-MCH increased, reaching a maximum at the same time and then was reacted by forming mainly 1,3-DCH and therefore decreased in concentration. After 100 min, the concentration of glycerol remained low but constant. In all the described cases, the observed behavior, characterized by a maximum monochlorohydrin concentration, clearly means that monochlorohydrins are intermediates of the formation of dichlorohydrins, confirming the reaction scheme presented in Figure 3. 8943



Review

Table 3. Results Reported by Schreck et al.,35 Using Different Catalystsa

8944



Review

Table 3. continued

8945



Review

Table 3. continued

8946



Review

Table 3. continued

The results are related to 4 h of reaction for runs performed at a temperature of 110 °C and HCl pressure of 7.48 atm, catalyst concentration = 3 mol %, with respect to loaded glycerol (330 mmol) containing 167 mmol water.

a

Schreck et al.,35 in their patent work, tested many catalysts. Among these, the most active, showing a glycerol conversion of >90 mol % in 4 h, were hexanoic acid, 4-trimethylammoniumbutyric acid, 4-dimethylbutyric acid, 4-aminobutyric acid, glycine, glycolic acid, lactic acid, 4-aminophenylacetic acid, 4-hydroxyphenylacetic acid, 4-methylvaleric acid, heptanoic acid, ε-caprolactone, and γ-butyrolactone. Therefore, all these catalysts have activities comparable to that of acetic acid, but not all are as selective to 1,3-DCH. All the results obtained by these authors,35 with the different tested catalysts, are reported in Table 3. It is interesting to observe from the data reported in Table 3 that some catalysts are not carboxylic acids and some others are carboxylic acids but also contain other more- or less-vicinal functional groups. Those functional groups sometime have a dramatic effect on the activity and selectivity. For example, benzoic acid (see Table 2) is not a good catalyst, despite a pKa value very near to that of acetic acid; a low activity is also shown by 2-aminobenzoic acid (see Table 3) and no activity is shown by 2-methylaminobenzoic acid. In contrast, phenylacetic acid has shown a high activity. Another interesting observation is that some catalysts that exhibit good activity could be derived from renewable sources, such as levulinic acid, for example. Bell et al.42 have also studied the performances of lesscommon carboxylic acids, such as 3-methylvaleric and 3,3dimethylbutanoic acid, giving a high HCl consumption rate, while 2-trimethylammoniumacetic acid chloride has showed a reaction rate slightly above that of blank run (without catalyst). Bell et al.42 compared these carboxylic acids with acetic acid, and a reaction rate higher than that of acetic acid was never observed. Some carboxylic acids that can be characterized by a high boiling point, such as adipic acid and caprylic acid, have been proposed as catalysts by Krafft et al.34 In this case, working at 120 °C, with HCl in solution, a conversion of glycerol above 95% was observed, using caprylic acid as a catalyst with a 1-MCH selectivity of 57.2% and a 1,3-DCH selectivity of 39.7%.

is subjected to some drawbacks, such as the catalyst loss by evaporation during the reaction, limiting the reaction temperature to less than 100−110 °C. In order to overcome these problems, many researchers have investigated the possibility of finding new catalysts to substitute in place of acetic acid. Siano et al.,36 for example, studied the use of malonic acid as a possible catalyst and compared its performance with that of acetic acid. This catalyst was studied under different operating conditions (flowing stream of HCl and total reflux) and at different temperatures. More precisely, the tests were conducted at temperatures of 80−110 °C, and from the results, it was observed that the yield in 1,3-DCH, after 3 h of reaction, increased by increasing the temperature, reaching a maximum of ∼57 wt % at 110 °C. The same authors36 also tested levulinic acid, operating at 100 °C under total reflux, with 8 mol % catalyst and a flowing stream of HCl (50 NL/h). Under these conditions, after 3 h of reaction, the conversion of glycerol was almost complete, but the yield in 1,3-DCH was slightly above 30 mol %, with 1-MCH being the main product (60 mol %). Citric acid was also found to be a good catalyst of glycerol chlorination.36 It was used again at a concentration of 8 mol %, temperature of 100 °C, and a flowing stream of HCl of 50 NL/h, for 3 h. After 3 h of reaction, the residual glycerol was slightly above 5 mol %, and the yield in 1-MCH was >70 mol %. This shows that this catalyst is more selective toward 1-MCH, with the second step of chlorination being very slow. The same authors also used succinic acid and propionic acid as possible catalysts. The operative conditions were the same as those of the previous tests. Using succinic acid, it has been observed almost complete conversion of glycerol in 150 min of reaction and the collected 1,3-DCH was more than 30 mol %. Propionic acid has shown an activity comparable to that of acetic acid and after 3 h of reaction, a complete conversion of glycerol and a 1,3-DCH yield of 41 mol % is observed. The same results have been reported in more detail by Tesser et al.10−12 All of the aforementioned results are summarized in Table 2, together with the performances reported in the literature, also using other catalysts for a useful comparison. 8947



Review

Working at 130 °C, with azeotropic HCl in solution, a glycerol conversion of >95% was obtained, using caprylic acid (96.7%) and adipic acid (99.4%), but a very high selectivity to 1,3-DCH was obtained in the presence of adipic acid (82.3%), greater than that in the presence of caprylic acid (60.3%). Tesser et al.12 also tested different carboxylic acids. All catalysts were tested at 100 °C, under conditions of total reflux except for HCl, with a HCl stream of 24 g/min and with a catalyst concentration of 8 mol %, as reported in Table 2. Among the catalysts tested, the most active were acetic acid, malonic acid, propionic acid, succinic acid, citric acid, and levulinic acid. In conclusion, it has been shown by several authors that there are many different catalysts that have high boiling points that can advantageously replace acetic acid. It is interesting to point out that some carboxylic acids are very selective toward monochlorohydrins, in particular, 1-MCH, as explained in a recent patent by Di Serio et al.41 These authors reported the use of dicarboxylic or hydroxycarboxylic acids as catalysts. All tests were carried out also in this case under reflux using a vertical condenser. The reaction was carried out for 3 h, at 100 °C, using a catalyst concentration of 8 mol %. From the results obtained, it can be seen that maleic and fumaric acids show similar catalytic behavior, with a conversion of glycerol slightly above 90 mol %, with an almost total selectivity to 1-MCH. Also tartaric and oxalic acids, after 3 h at 100 °C, using a catalyst concentration of 8 mol %, have shown a selectivity near to 100% toward 1-MCH, although the conversion of glycerol was not total (∼70 mol %). As previously mentioned, some catalysts selectively promote the first hydrochlorination step, mainly toward the formation of 1-MCH, while, dichlorohydrins start to form only after the complete conversion of glycerol. Clearly, the hydrochlorination mechanism for these type of catalysts would be different, as will be discussed later. In conclusion, an attempt could be made to classify the efficiency of the different catalysts tested by considering the influence of the following parameters: (i) the acidity strength, as in the case of the series, “acetic acid, mono-chloroacetic, dichloroacetic, and trichloro acetic acids”, tested by Tesser et al.12 by considering their pKa value; (ii) the length and branching of the alkyl chain of the monocarboxylic acids; (iii) the presence of different functional groups near the carboxylic group; and (iv) the presence in the molecules of more than one carboxylic group (dicarboxylic, tricarboxylic). These aspects will be considered and discussed in more detail in another section that will be devoted to the reaction mechanism. However, we can summarize here the most important experimental observations: (i) Some catalysts are active in glycerol hydrochlorination but promote only the first step to monochlorohydrins, while others promote both successive hydrochlorination steps to, respectively, monochlorohydrins and dichlorohydrins; (ii) Some dicarboxylic acids, such as adipic acid, are more active than monocarboxylic acids, provided that the alkyl chain between the two carboxylic acids is not too short to give a steric hindrance effect or too long to decrease its solubility in glycerol. Lastly, we have seen that, in some cases, catalysts different from carboxylic acids have been used with satisfactory results

(see Table 3). However, the compounds used are probably reactive substances that give the true catalyst in situ and normally are all organic substances. In contrast, Lee et al.44 investigated, for the same reaction, some inorganic compounds as catalysts. They studied, first of all, the behavior of a commercial polyoxometallate catalyst H3PW12O40. This catalyst was used after a thermal pretreatment at 300 °C for 2 h. The reaction between glycerol and HCl solution was carried out, at 100 °C, for times between 5 h and 30 h. H3PW12O40 showed a total conversion of glycerol in all the experimental runs, and the main products formed were dichlorohydrins. However, small amounts of acrolein, dichloropropane, propanediol, and dichloroethane were obtained as byproducts. The authors44 attributed the total conversion of glycerol and the high selectivity to DCH to the fact that the Brönsted acid sites of the catalyst are particularly active in promoting the glycerol chlorination reaction. The same authors45 have studied the reaction using many other commercial heteropolyacids catalysts, such as H3PMo12−XWXO40 (X = 0−12), H4SiMo12−XWXO40 (X = 0−12), H3+XPW12−XVXO40 (X = 0−3), and H3+XPMo12−XVXO40 (X = 0−3). All catalysts were tested after a thermal treatment at 300 °C for 2 h. H3PMo12−XWXO40 (X = 0−12) and H4SiMo12−XWXO40 (X = 0−12) showed 100% glycerol conversion. Even in these cases, monochlorohydrins are intermediate products formed in great amounts but, again, small amounts of acrolein, dichloropropane, propanediol, and dichloroethane were obtained as byproducts. Considering that the tungsten-containing HPA are more acidic than the molybdenum-containing HPA catalysts,46−50 it can be inferred that the acid property of HPA catalysts plays a very important role in determining the selectivity to DCH. For example, H3PMo12−XWXO40 (X = 0−12) showed a higher selectivity to DCH than H4SiMo12−XWXO40 (X = 0−12); this can be explained by the fact that the catalyst containing phosphorus as a heteroatom is even more acidic than the corresponding HPA with silicon. H3+XPW12−XVXO40 (X = 0−3), and H3+XPMo12−XVXO40 (X = 0−3) both showed 100% glycerol conversion in the reaction. The reactions was carried out in a liquid-phase batch reactor (200 mL); 12.6 g of glycerol (reactant), 78.9 g of aqueous HCl solution (37 wt %, chlorination agent), 20 g of H2O (reaction medium), and 15 g of HPA catalyst were charged into a batch reactor. After the homogeneous solution was heated to 110 °C with vigorous stirring (450 rpm), nitrogen was fed into the reactor to keep the reaction pressure at 10 bar. The catalytic reaction was carried out at 110 °C for 20 h with vigorous stirring.45 In both HPA catalysts containing vanadium, the selectivity to DCH decreases as the amount of vanadium is increased. However, H3+xPW12−xVxO40 (x = 0−3) catalysts showed a selectivity for DCH higher than H3+xPMo12−xVxO40 (x = 0−3) at the same level of vanadium substitution. The above results can be attributed to the acidic property of HPA-substituted catalysts decreasing in the following order: tungsten‐containing HPA > molybdenum‐containing HPA > vanadium‐containing HPA

The above results strongly support the observation that the acidic properties of HPA catalysts play a key role in determining the selectivity toward DCH. The more acidic the HPA catalyst, the more active the catalyst toward the formation of DCH from glycerol. Song et al.51 studied the effect of the operating conditions on the production of DCH. They studied, 8948



Review

Figure 4. Glycerol hydrochlorination reaction mechanism suggested by Tesser et al.10−12

that the selectivity to DCH was continuously increased by increasing the amount of catalyst. In conclusion, the strong Brönsted acid sites of H3PW12O40 catalyst favorably contribute to glycerol chlorination to obtain 1,3-DCH.

for example, the influence of the stirring rate on the formation of DCH in the presence of H3PW12O40 as a catalyst. The obtained results showed that the stirring rate is very important for an efficient formation of DCH in this solvent-free gas (hydrochloric acid)−liquid (glycerol/chlorohydrins) system. However, the selectivity to DCH was almost constant at the highest stirring rates (≥600 rpm), suggesting that mass transfer between gaseous HCl and glycerol could become a key factor in this reaction system. For this reason, the authors have stirred all of the reaction experiments at 900 rpm, in order to avoid any mass-transfer limitations. Another important aspect is the reaction temperature. The authors51 studied the effect of the reaction temperature on the formation of DCH from glycerol and gaseous HCl. In all of the experimental runs, which were performed at different temperatures, glycerol conversion was 100% but the selectivity to DCH gradually increased by increasing the reaction temperature until reaching a constant value of ∼98% at temperatures of >150 °C, while the selectivity to MCH gradually decreased by increasing the reaction temperature. It is interesting to note that trichlorohydrins (TCHs), which are a chlorination product of DCH, was formed in a very small amount only at very high reaction temperature. The dependence of product selectivity on the reaction pressure showed the same trend as those of the reaction temperature. In other words, selectivity to DCH increased by increasing both pressure and temperature. The same authors51 also reported that, through the addition of a substance that absorbs water, such as silica gel, the reaction favorably moves toward the formation of DCH.51,52 This means that water has a negative effect on glycerol chlorination. Finally, the same authors,51, using H3PW12O40 as a catalyst model, controlled the effect of the amount of catalyst on the formation of DCH and noted

4. REACTION SCHEME, REACTION MECHANISM, AND KINETICS 4.1. Hydrochlorination Reaction Mechanism. Glycerol hydrochlorination is a reaction that has been investigated many times in the past century, but only recently have some different reaction mechanisms and reaction conditions been proposed. Since the reaction takes place in the presence of a carboxylic acid as a catalyst, Santacesaria et al.10−12 hypothesized, in agreement with the previous literature,53,54 that, in the presence of a strong acid environment, because of HCl, the reaction occurs through an initial esterification. The reaction mechanism proposed by Tesser et al.10 is represented schematically in Figure 4. The first step of the scheme depicted in Figure 4 is a nucleophilic addition, in which one glycerol hydroxyl attacks the protonated carbonyl group. The first step is then followed by the formation of an oxonium group and, subsequently, by the addition of a chloride ion that leads to the formation of a monochlorohydrin. The last step is a nucleophilic substitution SN2 and occurs mainly in the α-position that is much more favored, with respect to the β-position. According to the proposed mechanism, the product obtained in a larger amount is always 1-MCH. In conclusion, according to the authors, the overall reaction, in practice, is a nucleophilic substitution of a carboxylic group with a chloride ion. It is well-known that carboxylic groups are better leaving groups than the hydroxyl groups. For this reason, carboxylic acids are good catalysts for this reaction, because, in an acidic environment, they easily 8949



Review

form esters and can then be substituted by chlorine more easily than the hydroxyl groups. But the acidity of the carboxylic acids cannot be too strong to avoid the fact that the corresponding esters are too stable to complete the reaction. As a matter of fact, by comparing the results reported in the last four examples reported at the bottom of Table 2, it is possible to observe that acetic acid is much more active and selective than monochloroacetic acid and the activities and selectivities decrease more and more for respectively dichloro and trichloro acetic acids. A useful parameter to be considered in the mentioned cases is the pKa. As it can be seen, pKa gradually decreases by introducing more chlorine in the acetic acid molecule, that is, monochloroacetic acid, dichloroacetic acid, and trichloroacetic acid have gradually increasing acidity, with respect to acetic acid, and this is detrimental for the hydrochlorination reaction. This is probably due to the increased stability of the corresponding ester. The reaction mechanism described above is in agreement with the experimental evidence that the amount of 1-MCH that is formed is always greater than the amount of 2-MCH. Moreover, 2-MCH does not react further to give 1,2-DCH. The absence of two vicinal OH groups, in the case of 2-MCH, probably prevents the formation of the intermediate oxonium, thus hindering the second chlorination reaction. In contrast, the 1-MCH can undergo a further chlorination with a mechanism similar to that shown above, forming mainly 1,3-DCH, accompanied by small quantities of 1,2-DCH. The reaction scheme suggested by Tesser et al.,10 based on the experimental observation is given by Figure 5.

Figure 6. Evolution with time of the experimental products distribution for the run with tartaric acid at T = 100 °C. Open box (□) corresponds to 1,3-DCH. Points are experimental data, while the solid lines are simulations except for 1,3-DCH. keq1

glycerol + catalyst HoooI ester1 + water keq2

1‐MCH + catalyst HoooI ester2 + water

(4) (5)

If Keq1 ≈ Keq2, the formation of monochlorohydrins and dichlorohydrins follows a reaction-in-series mechanism, according to that proposed. In contrast, if Keq1 ≫ Keq2, the catalyst gives mainly monochlorohydrin, because the catalyst gives mainly the ester1 as an intermediate. However, also a strong difference in the two direct kinetic constants or a significant influence of the HCl mass-transfer limitation, in the initial period of the reaction, can contribute to determining the selectivity to MCH shown by some catalysts. However, it is then interesting to observe that catalysts having pKa ≥4 are normally selective to dichlorohydrin, while the catalysts with pKa values in the range of 1.2−3 are more selective to monochlorohydrins. Finally, the selectivity to monochlorohydrins shown by different catalysts41 opens a perspective to the industrial production of 1-MCH and, consequently, also to the production of glycidol that both could become building blocks for other interesting syntheses. As seen previously, the more-acidic carboxylic acids, such as trichloroacetic acid, are not active in the reaction. Clearly, to promote the reaction, it is necessary to form esters that are not too stable to allow the subsequent reaction steps. However, as recently shown by Tesser et al.,12 the pKa value is not the only factor influencing the selectivity, because the presence of more than one carboxylic group in the same molecule and/or the presence of other functional groups could be important. More deepened mechanistic studies are clearly necessary for a finetuning of the interpretation of the selectivity−structure relationship. An improvement is represented by the application of the Taft equation.12 It has been shown, by applying the Taft equation to the homologous seriesacetic acid, monochloroacetic acid, dichloroacetic acid, and trichloroacetic acidthat both the polar and steric effect are important in determining the catalyst behavior, in terms of activity and selectivity. Other two different alternative mechanisms have been proposed in the literature.42,55,57−59 Bell et al.42 suggested a mechanism, derived from an old paper,55 in which esterification is again the first step but, after the ester formation, they suggest the formation of a tautomeric cyclic molecule that, in the acidic

Figure 5. Scheme of the hydrochlorination reactions (see Santacesaria et al.10−12).

From the data reported in the previous section, it is possible to observe that some catalysts are highly selective in promoting the formation of monochlorohydrins (in particular, 1-MCH). It seems that, in those cases, the second hydrochlorination step is prevented until almost all of the glycerol is converted to monochlorohydrins, as can be seen in Figure 6, relative to the behavior of tartaric acid used as a catalyst.11 This particular behavior requires an explanation based on the reaction mechanism. The first step of the reaction, that is, the formation of an ester between glycerol and the carboxylic acid used as catalyst is surely an equilibrium reaction more or less shifted to the right. The successive hydrochlorination again requires the formation of an ester between a monochlorohydrin and the carboxylic acid, and obviously this reaction also is an equilibrium reaction. Tesser et al.10−12 suggested that when the equilibrium constant of the first equilibrium step of esterification is much greater than the second one, only the first reaction can occur, while the second one is prevented, because all of the loaded catalyst is involved in the most favorable equilibrium until the concentration of the main reactant (glycerol) becomes very low. In conclusion, the reaction mechanism remains the same, as shown in Figure 4 or 5 but the following esterification reactions can have different equilibrium constants: 8950



Review

Figure 7. Mechanism of glycerol hydrochlorination catalyzed by carboxylic acids, according to Bell et al.42,55

Figure 8. A bimolecular SN2 mechanism.

Moreover, the chlorination reactions of Figure 5 (reactions 1 and 3) were considered irreversible. From their kinetic runs, it can be observed that the presence of water has a detrimental effect on the reaction rate as recently confirmed by Dmitriev et al.56 Ling et al.60 proposed a reaction scheme slightly different from that proposed by Tesser et al.,10 reported in Figure 5, assuming that the reaction is reversible only as reaction 3. They studied the kinetics of glycerol chlorination in the presence of different catalysts, such as acetic acid, propionic acid, malonic acid, succinic acid, and adipic acid, and they studied the behavior of adipic acid in particular. The authors collected the experimental data using the gas chromatography−mass spectroscopy (GC-MS) analysis method. Although these authors announced the proposal of a new mechanism in their work, they interpreted all their experimental data with pseudo-firstorder kinetics laws and the reaction scheme has been presented as a reaction mechanism. More recently, Salmi and co-workers61 reinterpreted the data from Tesser et al.10 and their own experimental results with a more general and detailed mechanism in which all the possible intermediate species are considered. Intermediates species considered are as follows: E1 (the first ester between glycerol and the catalyst), E2 (the second ester between MCH and the catalyst), I1+ (the first ionic intermediate (protonated E1)), and I2+

environment, is converted to an acetoxonium cationic ring, which is destabilized by a chloride ion and gives monochlorohydrins after the hydrolysis of the corresponding ester, as it can be seen in the scheme depicted in Figure 7. The further chlorination step would occur with the same mechanism and mainly gives 1,3-DCH. This mechanism is intriguing but, as it has been seen, requires a final hydrolysis to close the catalytic cycle; that is, according to this mechanism, the presence of water would be important for favoring the last indispensable reaction step. In contrast, it has been demonstrated that aqueous HCl is less active than gaseous anhydrous HCl and that the presence of water negatively affects both the reaction rate and the yields.56 Luo et al.57,58 and Lim et al.59 both studied the kinetics of glycerol hydrochlorination using aqueous HCl (37 wt %) and assumed a mechanism similar to that proposed by Tesser et al.10 but not considering any possible intermediate species, with the exclusion of the esters; that is, they considered only the double-sequence esterification−chlorination, applying a second-order kinetics law to all of the reactions that are occurring and a pseudo-steady-state condition to the ester compounds. This mechanism can be considered a direct bimolecular nucleophilic substitution SN2, in which Cl− anion substitute in one step RCOO− in the ester molecule (Figure 8): 8951



Review

where catalyst+ is the protonated catalyst, I1+ the first reaction intermediate, ester+ the ester in protonated form, and I2+ the second reaction intermediate in protonated form. At this point, it is easy to write the kinetic laws for each elementary step reaction as

(the second ionic intermediate (protonated E2)). The proposed mechanism is the same as that depicted in Figure 4, but the kinetic approach is based on the quasi-steady-bstate hypothesis applied to the four mentioned intermediate species. The authors applied their model to the runs performed by Tesser et al.10 in the presence of malonic acid and to their run made in the presence of acetic acid. 4.2. Reaction Kinetics. 4.2.1. Kinetics of Glycerol Hydrochlorination. As seen in the scheme depicted in Figure 3, the synthesis of epichlorohydrin from glycerol involves several reaction steps, and for a correct plant design, all of the kinetic laws and related parameters of the eight reactions appearing in Figure 3 must be known. For this purpose, a detailed kinetic investigation performed in the kinetic regime is required. Until now, different papers have been published regarding the kinetics of (i) the formation of chlorohydrins, starting from glycerol and gaseous HCl, in the presence of a carboxylic acid as catalysts10−12,60,61 or using a concentrated aqueous HCl solution;57−59 (ii) the formation of epichlorohydrin from dichlorohydrins and degradation of this product, as a consequence of the ring-opening reaction or other undesired side reactions.24,25 However, it is opportune to focus, first of all, only on the reactions involved in the hydrochlorination that, according to Santacesaria et al.,10−12 are those reported in the scheme described by Figure 5. The experimental data collected on these reactions by different authors have been interpreted with kinetic models that differ for the adopted reaction mechanism and/or for the number of physical phases considered. Tesser et al.,10 for example, have interpreted all their experimental data with a pseudohomogeneous kinetic model. The authors have studied in detail the kinetics of the hydrochlorination reaction in the presence of malonic acid as a catalyst, determining for this catalyst the best performing kinetic laws and related parameters. Then, they applied the same kinetic laws to many other catalytic systems and determining the related kinetic parameters for a fixed temperature of 100 °C. It is possible to recognize, in the scheme described by Figure 5, a sequence of four reactions:

r1* = k1[Cat][H+]

r2* = k 2[Gly][Cat+] r5* = k5[ester +] r3* = k 3[I1*]

k −1 k2

glycerol + HCl → 2‐MCH + H 2O k ‐3 k4

1‐MCH + HCl → 1,2‐DCH + H 2O

(6) (7)

(8) (9)

Each reaction reported here (reactions 6−9) corresponds to a sequence of elementary reaction steps and we can write, in a simplified way, for each hydrochlorination step, the following sequence of elementary reactions: step 1)

catalyst + H+ → catalyst+

(10)

step 2)

glycerol + catalyst+ → I1+

(11)

step 3)

I1+ → ester + H 2O + H+

(12)

step 4)

ester + H+ → ester +

(13)

step 5)

ester + → I 2+ + catalyst

(14)

step 6)

I 2+ + Cl− → 1‐MCH

(15)

⎡ ⎛ C HCGly ⎞ ⎤ K ⎟⎟ − 1 (C1‐MCH)⎥ r1 = CC⎢K1⎜⎜ K E1 ⎢⎣ ⎝ C W ⎠ ⎥⎦

(16)

⎛ C HCGly ⎞ r2 = K 2CC⎜ ⎟ ⎝ CW ⎠

(17)

⎡ ⎛C C ⎤ ⎞ K r3 = CC⎢K3⎜⎜ H 1‐MCH ⎟⎟ − 3 (C1,3‐DCH)⎥ K E3 ⎢⎣ ⎝ C W ⎥⎦ ⎠

(18)

⎛C C ⎞ r4 = K4CC⎜ H 1‐MCH ⎟ ⎝ CW ⎠

(19)

For Model B, assuming that the ester formation is the RDS, the following kinetic expressions resulted:

k3

1‐MCH + HCl ⇌ 1,3‐DCH + H 2O

r6* = k6[I+][Cl−]

Then, following the rate-determining step (RDS) approach, two different kinetic models were applied. In the first model (Model A), the oxonium ion formation (a three-membered ring, see the mechanism described in Figure 4) was considered as the RDS; in the second one (Model B), the ester formation was assumed to be the RDS. By mathematical manipulation, it is possible to define, for each hydrochlorination reaction, in reactions (6−9), a kinetic law expression for Model A was as follows:

k1

glycerol + HCl HooI 1‐MCH + H 2O

r4* = k4[ester][H+]

⎡ ⎤ K r1 = CC⎢K1C HCGly − 1 (C1‐MCHC W )⎥ ⎢⎣ ⎥⎦ K E1

(20)

r2 = K 2CCC HCGly

(21)

⎡ ⎤ K r3 = CC⎢K3C HC1‐MCH − 3 (C1,3‐DCHC W )⎥ K E3 ⎢⎣ ⎥⎦

(22)

r4 = K4CCC HC1‐MCH

(23)

The equations representing Model A (eqs 16−19) or Model B (eqs 20−23) have been used by the authors10 to solve the material balance equations that, for the liquid phase, in a batch reactor, by assuming a chemical kinetic regime, can be written as follows: dnG = (− r1 − r2)VR dt dn2‐MCH = (r2)VR dt dn W = (r1 + r2 + r3 + r4)VR dt

dn1‐MCH = (r1 − r3 − r4)VR dt dn1,3‐DCH = (r3)VR dt dn1,2‐DCH = (r4)VR dt (24)

8952



Review

Figure 9. Parity plot experimental versus calculated data for a kinetic run performed with malonic acid at T = 100 °C. Comparison between Model A and Model B (from Tesser et al.10).

Here, ni is the moles of component i in the liquid phase and VR is the volume of the reacting liquid mixture. The details for solving the ordinary differential equation (ODE) system described by eqs 24 are reported in ref 10. All the experimental data were submitted to mathematical regression analysis,62 and Model B exhibited the best results, with regard to fitting the available data, as shown in Figure 9. The two models have also been submitted to a different approach based on the “steady-state approximation”. At this purpose, the catalytic cycle for the chlorination of glycerol can be expressed according to the following simplified scheme (Figure 10):

Equation 27 has been applied to the runs performed with malonic acid as a catalyst, and a ratio of Kchlor/Kest ≫ 1 resulted. This suggests that the esterification step is much slower than the chlorination one, confirming the assumption that ester formation is the rate-determining step (RDS) and Model B would be the most reliable. In conclusion, the esterification of glycerol with a carboxylic acid used as catalyst seems to be the RDS of the overall hydrochlorination reaction. Tesser et al.12 very recently have implemented their Model B also considering the role of the gas−liquid mass transfer of HCl. At this purpose, they adopted the Whitman double-film theory assuming, as a first approximation, that no resistance to mass transfer is given by the gas-side film. In this way, it is possible to write a mass-transfer rate expression for HCl in the liquid film as follows: JHCl (mol/cm 3 min) = k1a([HCl]* − [HCl]) = β([HCl]* − [HCl])

(28)

where [HCl]* is the HCl equilibrium concentration at the gas/ liquid interphase, evaluated as Figure 10. Catalytic cycle according to Tesser et al.10

[HCl]* = HHClP

The two following reaction rate expressions can be written as overall esterification reaction: rest = KestCC*CGly

HHCl is the Henry constant of the HCl solubility, corresponding to HHCl =

(25)

overall chlorination reaction: rchlor = KchlorC E*C H

By applying the steady-state approximation to the ester concentration, and taking into account for the catalyst the material balance CC = CC* + CE* considering that the catalyst is partitioned between the original form of carboxylic acid and the ester. An overall reaction rate equation for the product formation can be derived: rchlor

( )

solubility P

(30)

with P being the pressure. The solubility in a multicomponent reaction mixture such as that considered here can be calculated as

(26)

⎤ ⎡ ⎥ ⎢ 1 = KchlorCCCGC H⎢ ⎥ ⎢ CG + Kchlor C H ⎥ K est ⎦ ⎣

(29)

solubility =

i ∑ xiKHCl

(31)

A complete mass balance can be written as follows. Liquid-Phase Mass Balance.

(27) 8953

dnHCl = −r1VR − FGly dt

(32)

dn1‐MCH = ( +r1 + r3 + r4)VR − F1‐MCH dt

(33)



Review

Table 4. Kinetic Constants for Reactions 6−9, and Corresponding Arrhenius Parametersa temp, T [°C] 80 90 100 110 120

k1 7667 11704 13274 19433 27411

± ± ± ± ±

Ea [kJ mol−1] ln A

k2 940 1272 1692 2216 2861 reaction 1

450 764 1089 1465 2215

± ± ± ± ±

k3 41 60 87 123 170

714 1109 1784 2383 2179 reaction 2

35.2 ± 0.3 20.9 ± 9

44.3 ± 0.2 21.3 ± 0.7

± ± ± ± ±

k4 227 307 407 532 685

8±3 13 ± 5 26 ± 7 32 ± 9 31 ± 13 reaction 3 34.9 ± 0.8 18.6 ± 2.2

KE1b

KE2c

3846 3064 2470 2015 1660

194 167 146 128 113 reaction 4 42.1 ± 1.0 16.5 ± 2.8

a

Kinetic constants are expressed in units of cm6/(mol2 min). Data have been collected by using malonic acid as catalyst. Data taken from Tesser et al.10 bEquilibrium constants for reaction 6. cEquilibrium constants for reaction 8.

Table 5. Kinetic Constants, at T = 100 °C, for Various Catalystsa Kinetic Constants [cm6/(mol2 min)] k1

catalyst acetic acid malonic acid citric acid levulinic acid succinic acid propionic acid tartaric acidb a

34619 13274 3307 16905 13549 25545 21712

± ± ± ± ± ± ±

k2 4012 1692 387 1801 1290 2871 2561

2342 1089 247 1315 1028 2265 2096

± ± ± ± ± ± ±

k3 198 87 21 119 95 201 201

1576 1784 269 1421 1354 1197

± ± ± ± ± ±

k4 302 407 41 85 78 91

17 26 0.5 17 14 12

± ± ± ± ± ±

4 7 0.2 6 5 4

Data taken from Tesser et al.10 bTartaric acid is selective to 1-MCH.

dn1,3‐DCH dt

= +r3VR − F1,3‐DCH

dn2‐MCH = +r2VR − F2‐MCH dt

dn1,2‐DCH dt dn H 2 O dt

= +r4VR − F1,2‐DCH

= ( +r1 + r2 + r3 + r4)VR − FH2O

dnHCl = ( −r1 − r2 − r3 − r4)VR + JHCl VR − FHCl dt

kinetic constants obtained for other catalysts are reported in Table 5. In the previous section, we have seen that some catalysts are selective in producing monochlorohydrins, because the second hydrochlorination step seems to be prevented. As mentioned previously, Santacesaria et al.11 have suggested that this is mainly due to a difference in the two equilibrium constants of esterification of glycerol and monochlorohydrins (reactions 4 and 5, respectively). By introducing this change, in the previously described model, it is possible on a theoretical basis to also correctly simulate the kinetic runs performed in the presence of catalysts that are selective in producing monochlorohydrins instead of dichlorohydrins. However, it is difficult to discriminate if this is the only reason for the observed selectivity or if other factors can contribute toward such as a strong difference in the kinetic constant between first and second hydrochlorination or a mass-transfer limitation for HCl migration from the gas phase to the liquid phase. This difficulty is related to the long time necessary, in those cases, to collect reliable kinetic data for the second hydrochlorination step for the low conversion level toward 1,3-DCH. Therefore, the kinetic analysis has been restricted only to the first hydrochlorination step. Figure 6 shows an example simulation of a kinetic run performed in the presence of tartaric acid, while the corresponding kinetic parameters are reported in Table 5. Another kinetic approach has recently been proposed by Ling et al.,60 based on a reaction scheme, such as that reported in Figure 5, but assuming only reaction 3 to be reversible. They particularly studied the kinetics of glycerol chlorination in the presence of adipic acid, proposing an unreliable model in which all the reactions are considered to be pseudo-first order. Consequently, the obtained fittings are not satisfactory, as can be seen in Figure 11. On the other hand, the work is interesting for the experimental results reported on both the kinetic behavior of adipic acid and the comparison of glycerol conversion in the

(34)

(35)

(36)

(37)

(38)

with r1 = k1[Cat][Gly][HCl] − k −1[Cat][H 2O][1‐MCH] r2 = k 2[Cat][Gly][HCl] r3 = k 3[Cat][1‐MCH][HCl] − k −3[Cat][H 2O][1,3‐DCH] r4 = k4[Cat][1‐MCH][HCl] (39)

Gas-Phase Mass Balance. dn HClgas = FHCl IN − JHCl VR dt

(40)

IN FHCl = KP(PSET − PTOT)

(41)

Clearly, the mass-transfer parameters kLaL depends on both the device used to perform the reaction and the adopted operative conditions (pressure, temperature, type of catalyst, and concentration). All the kinetic parameters, collected by Santacesaria et al.10−12 by applying the described Model B to the catalyst malonic acid, are reported in Table 4, while the 8954



Review

presence of different catalysts, such as acetic acid, propionic acid, malonic acid, succinic acid, and adipic acid. The kinetic parameters found by Ling et al.,60 along with their pseudo-first-order model, are collected in Table 6. We attempted to simulate a run reported in the work of Ling et al.60 Unfortunately, in that work, the initial amount of glycerol and the amount of catalyst used (adipic acid) were not reported. To make the simulation, we assumed 100 g of glycerol, 5 mol % catalyst, a temperature of 110 °C, and a pressure of 1 bar. The simulation has been performed with the most complete kinetic model proposed by Tesser et al.,12 and the results obtained can be appreciated in Figure 11, where the results are compared with the results obtained with the pseudofirst-order model suggested by Ling et al.60 The best-fitting parameters of the Tesser et al.12 model used for the simulation of the data reported in Figure 11 are as follows: k1 = 5.30 × 105, k2 = 2.62 × 104, k3 = 2.54 × 104, k4 = 32.1, k−1 = 3.03 × 10−11, k−3 is negligible, and β = KLa = 0.12 (see eqs 28−41). Dmitriev et al.63 studied the kinetics of both esterification and hydrochlorination of glycerol separately. Acetic acid has been used in the esterification of glycerol as a reagent, while, as always, acetic acid is used in hydrochlorination as a catalyst. The reactions considered for esterification are

Figure 11. Comparison of the performance of the kinetic models respectively proposed by Tesser et al.12 and Ling et al.60 The kinetic run has been performed by Ling et al.60

Table 6. Pseudo-First-Order Kinetic Constants for the Hydrochlorination of Glycerol in the Presence of Adipic Acida Pseudo-First-Order Kinetic Constants [min−1]

a

temp, T [°C]

k1 × 102

k2 × 104

k3 × 103

k4 × 105

90 100 110 120

1.23 1.35 2.01 2.56

3.16 4.59 6.59 9.07

2.42 4.18 5.39 5.03

3.12 8.70 11.10 11.37

Runs performed in flowing HCl at atmospheric pressure.60

Keq1 = 1.89

(47)

⎡ −(53000 ± 3000) ⎤ k 2 (L2 s2/mol2) = 1.60 × 104 exp⎢ ⎥ ⎣ ⎦ RT (48)

Keq2 = 1.00

From these parameters, it is possible to observe that reaction 42 is much faster than reaction 43. However, these data are not reliable, because the presence of HCl as a catalyst also induces the hydrochlorination reaction. Another observation is related to the negative role of water in both reactions. In the presence of water, the esterification reaction rate is lower and the concentration of esters at equilibrium also is lower. The same authors also studied the hydrochlorination reactions, assuming a bimolecular nucleophilic substitution mechanism SN2 to interpret the results. They considered all of the reactions of the scheme reported in Figure 5 irreversible and, therefore, used pseudo-second-order kinetic laws equations. The reactions were studied in the absence and the presence of acetic acid as a catalyst, observing that the reaction rate in the absence of acetic acid is only 3% of that in the presence of this catalyst. By applying their bimolecular model, they found that kinetic constants were strongly affected by the ratio CHCl/CH2O and introduced this term, in an empirical way, inside the kinetic law to interpret the kinetic data. For the formation of 1-MCH and 2-MCH, the following relations have been proposed:

In both cases, HCl has been introduced in a relatively small amount to act as a catalyst. Kinetic data of esterification have been interpreted with the kinetic laws: ⎛ CGly AcetateC H2O ⎞ ⎟⎟ r1 = k1C HCl ⎜⎜CGlyCAcetic Acid − Keq1 ⎝ ⎠

(44)

⎛ C1‐MCHC H2O ⎞ ⎟⎟ r2 = k 2C HCl ⎜⎜C1‐MCHCAcetic Acid − Keq2 ⎝ ⎠

(45)

(49)

The kinetic parameters determined by the authors are ⎡ −(58500 ± 3000) ⎤ k1 (L2 s2/mol2) = 6.29 × 105 exp⎢ ⎥ ⎣ ⎦ RT (46) 8955



Review

Table 7. Reinterpretationa of the Kinetic Data from Tesser et al.10 with the Kinetic Model Proposed by Salmi and Co-workers61 T = 80 °C k3′ [L mol−1min−1] k4′ [L mol−1min−1] k′7 [L mol−1min−1] k′8 [L mol−1min−1] β′ [mol L−1] δ′ [mol L−1] a

9.36 5.45 6.55 1.81 0.57 1.43

× × × ×

T = 90 °C

10−4 10−5 10−7 10−4

3.98 2.64 5.03 5,48 2.05 2.62

× × × ×

T = 100 °C

10−3 10−4 10−6 10−4

1.26 1.01 5.75 1.25 6.33 4.06

× × × ×

T = 110 °C

10−2 10−3 10−6 10−3

4.30 × 3.30 × 3.29 × 2.23 × 17.50 6.91

T = 120 °C

10−2 10−3 10−5 10−3

1.70 × 1.54 × 7.77 × 2.89 × 50.70 10.30

10−1 10−2 10−5 10−3

The best-fitting parameters are reported in the table. (10.5 − 0.023T ) ⎡ −(102000 ± 3000) ⎤ ⎛ C HCl ⎞ ⎟⎟ CGlyC HCl r3 = 2.1 × 109 exp⎢ ⎥ × ⎜⎜ ⎦ ⎝ CH O ⎠ ⎣ RT 2

r8 =

(50) ⎡ − (132000 ± 4000) ⎤ ⎛ C HCl ⎞ ⎟ r4 = 2.2 × 1012 exp⎢ ⎥ × ⎜⎜ ⎣ ⎦ ⎝ C H O ⎟⎠ RT 2

CGlyC HCl

(51)

while, for the formation of 1,3-DCH and 1,2-DCH,

r3 = k 3*CCatCGlyC HCl

⎡ −(86000 ± 1000) ⎤ r5 = 1.24 × 10 exp⎢ ⎥ ⎣ ⎦ RT 11

(52)

Table 8. Activation Energies and Pre-exponential Factors for the Hydrochlorination Reactions of Glycerol Catalyzed by Malonic Acida (53)

In contrast with Tesser et al.10 and others, these authors concluded that the esterification reaction is fast and must be always considered at equilibrium and hydrochlorination reactions are much slower. However, these results are not supported by experimental data, because the kinetic data of hydrochlorination are not reported. Moreover, we have seen that the esterification data are not reliable for the presence of HCl as a catalyst, which can become a reagent of the hydrochlorination in this reaction system. Salmi and co-workers61 improved the kinetic approach of Tesser et al.10 to describe the glycerol hydrochlorination reactions. They assumed a similar mechanism but applied the principle of quasi-steady state approximation to all of the intermediate species appearing in the mechanism, that is, the esters E1 and E2 and the charged species I1+ and I2+. After opportune mathematical manipulations, the authors obtained the following set of kinetic law expressions:

r4 =

Ea [kJ/mol] ln A

k ′c c c 2 r7 = 7 Cat 1‐MCH HCl D78

r4

r7

r8

159.4 33.5

132.1 20.0

80.7 8.1

A reinterpretation by Salmi and co-workers61 of kinetic runs performed by Tesser et al.10

shows the activation energies and pre-exponential factors determined from the parameters reported in Table 7. The same model has also been used by the authors to interpret their own data of a run performed in the presence of acetic acid at 105 °C, and the corresponding parameters are reported in Table 9. Table 9. Kinetic Parameters Obtained by Salmi and Co-workers61 for the Hydrochlorination Reactions of Glycerol, Catalyzed by Acetic Acid (11 mol %), for Runs Performed at 105 °C parameter k3′ k4′ k′7 k′8 β′ δ′

(54)

k4′cCatcGlyc HCl 2 D34

r3 147.5 32.3

a

k 3′cCatcGlyc HCl 2 D34

(58)

where k*i corresponds to k3′/β′, k′4/β′, k′7/δ′, and k′8/δ′. They successfully applied their model to the kinetic data collected by Tesser et al.10 for the runs performed at different temperatures in the presence of malonic acid, and the corresponding kinetic parameters are reported in Table 7. Table 8

CGlyC HClCAcetic Acid

⎡ −(91500 ± 1000) ⎤ r6 = 1.38 × 1010 exp⎢ ⎥ ⎣ ⎦ RT

r3 =

r4 = k4*CCatCGlyC HCl

r7 = k 7*CCatCMCHC HCl r8 = k 8*CCatCGlyC HCl

(25.63 − 0.056T )

⎛ C ⎞(13.36 − 0.026T ) CGlyC HClCAcetic Acid × ⎜⎜ HCl ⎟⎟ ⎝ C H 2O ⎠

(57)

where r3 represents the rate of 1-MCH formation, r4 the rate of 2-MCH formation, r7 the rate of 1,2-DCH formation, and r8 the rate of 1,3-DCH formation. D34 = CCatCW + (α′CW + β′) CHCl and D78 = CCatCW + (γ′CW + δ′)CHCl. By assuming CW ≈ 0, because water is continuously removed, the four expressions become

(14.7 − 0.034T )

⎛C ⎞ × ⎜⎜ HCl ⎟⎟ ⎝ C H 2O ⎠

k 8′cCatc1‐MCHc HCl 2 D78

(55)

value 2.16 1.39 1.29 1.01 17.4 1.91

× 10−2 [L × 10−3 [L × 10−9 [L × 10−3 [L mol L−1 mol L−1

mol−1min−1] mol−1min−1] mol−1min−1] mol−1min−1]

The mechanism assumed by Salmi and co-workers61 is quite similar to that proposed by Tesser et al.,10 being different in only the kinetic model approach. The kinetic model of Salmi

(56) 8956



Review

and co-workers61 is more complex and rigorous, but many simplifications have been introduced by the authors probably for limiting, at an acceptable level, the number of kinetic parameters to be determined from the experimental data. Therefore, to use the most rigorous model, a large number of kinetic runs is necessary. It is not possible to make a direct comparison between the kinetic constants reported in Table 7 and those reported in Table 4, because the units are different; however, it is possible to evaluate a ratio between the corresponding kinetic constants. It is interesting to note that these ratios are quite similar. The conclusion is that the two models are practically equivalent in simulating the experimental data collected at different temperatures, relative to malonic acid. The models proposed by other authors are less reliable. Lastly, it is worth remembering that Luo et al. and Lim et al.58−60 studied the kinetics of glycerol hydrochlorination in the presence of acetic acid but using concentrated aqueous HCl. They adopted the bimolecular substitution S N 2 mechanism of Figure 8 and considered only esterification as equilibrium reactions. On the basis of this mechanism, all the reactions are considered to be pseudo-second-order. The kinetic parameters determined by Luo et al.59 are summarized in Table 10. Clearly, the kinetic behavior in the presence of

Figure 12. Plots of HCl absorbed by a mixture of 85 wt % glycerin, 9 wt % water, and 7 wt % acetic acid at 90 °C at different pressures (from ref 42).

As previously mentioned, the pressure of HCl affects both the hydrochlorination reaction rate and the yield toward the formation of DCHs. This behavior can be simulated by using the more complete model suggested by Tesser et al.,12 also considering HCl solubility and mass-transfer limitations. For this purpose, Figure 13 shows a simulation of two runs performed by Bell et al., at 90 °C, in the presence of 7 wt % of

Table 10. Kinetic Constants Obtained by Luo et al.59 for the Reactions Occurring in Glycerol Hydrochlorination Performed with HCl in Aqueous Phase, Using Acetic Acid as a Catalyst kinetic constant [mol L−1 min−1] k1 k2 k3 k4 k5

= 1542.3 exp(−5702.7/T) = 175.6 exp(−4918.8/T) = 5.8 exp(−4381.2/T) = 40.34 exp(−7171.32/T) = 27.83 exp(−4950.5/T)

activation energy [kJ/mol]

reaction

47.4 40.9 36.4 59.6 41.2

glycerol consumption 1-MCH formation 1,3-DCH formation 2-MCH formation 1,2-DCH formation

aqueous HCl cannot be directly compared with that in the presence of pure gaseous HCl; however, it is possible to conclude qualitatively that the presence of water decreases the hydrochlorination reaction rate. As has been seen, in the kinetic model proposed by different authors, glycerol and monochlorohydrin hydrochlorination reactions are considered to be first order, with respect to HCl concentration, but using gaseous HCl; this concentration depends on the solubility of HCl in the reaction environment. This solubility obviously changes with pressure, temperature, and composition of the liquid mixture. Unfortunately, there is not a work in the literature in which this aspect has been examined deeply enough. The experimental observation is that pressure has a dramatic effect on both reaction rate and final yields of dichlorohydrins. Bell et al.42 reported the evolution with time of the number of moles of HCl adsorbed per mole of loaded glycerol for different pressures (from 1 bar to 5.5 bar). The obtained results are reported in Figure 12. As can be seen by increasing the pressure, the reaction rate increases very much but the HCl uptake also increases; this last observation means that hydrochlorination reactions are limited by the equilibrium. As a matter of fact, by operating at 120 °C with 2 wt % of acetic acid, at 1.36 atm of HCl pressure, a glycerol conversion of ∼50% is observed; at 2.04 atm, the glycerol conversion is >85%; and at 3.40 atm, the glycerol is completely converted.42

Figure 13. Simulation of the runs performed by Bell et al.42 at different pressures. The other experimental conditions are summarized in the legend of Figure 12.

acetic acid, at different pressures of 2.76 and 5.52 bar, respectively. The plot reports the evolution of the HCl consumption as a function of time in the used semibatch reactor. The kinetic parameters giving the best fit in these simulations are k1 = 31234, k2 = 3040, k3 = 2578, k4 = 9.9 × 10−6, k−1 = 1.18 × 10−12, k−3 is negligible, and β = KLa = 0.99 (see eqs 28−41). Tesser et al.12 also made and simulated some kinetic runs at different HCl pressures, but using monochloroacetic acid as a catalyst; the results are reported in Figure 14. The trends observed by Bell et al.42 for the effect of pressure on activity and yield are also confirmed in this case. 4.2.2. Kinetics of Epichlorohydrin Synthesis from Dichlorohydrins. The kinetics of epichlorohydrin synthesis with related side reactions (reactions 5−8 of Figure 3) has been studied in detail many years ago by Carrà et al.24 The reaction occurs by dehydrochlorination starting from both 8957



Review

the kinetic data related to the synthesis of epichlorohydrin with a pseudo-first-order kinetic law satisfactorily. Epoxides are reactive substances and, in a basic environment, epichlorohydrin is subjected to an undesired ringopening reaction, decreasing the yield, and occurring with the following mechanism, which has been suggested by Patai:64

Figure 14. Glycerol conversion as a function of time for runs 2, 5, and 6 of Table 2 for the monochloroacetic acid catalyst.

Therefore, the kinetic law for the ring opening (rro) will be

1,3- and 1,2-MCH in an aqueous basic environment created by dissolving in water Ca(OH)2 or NaOH. The main occurring reactions are: 1,2‐DCH → EPY + HCl

(59)

1,3‐DCH → EPY + HCl

(60)

rro = k roK ii[EPOX][OH−] ≈ keff [EPOX]

(65)

Again, it is possible to assume a pseudo-first-order kinetic law. On the basis of all the experimental data collected, Carrà et al.24 isolated the following reaction scheme:

Clearly, the developed HCl neutralize the basic catalyst used. Therefore, if NaOH is used, the pH changes rapidly and the reaction rate slows down. By using Ca(OH)2, which is poorly soluble in water, a buffer effect is operative, because of the presence of both Ca(OH)2 and CaCl2. As a consequence, the OH − concentration changes smoothly with the DCH conversion, according to the relationship [OH−]3 + XC0[OH−]2 − 2KS = 0

(61)

0

where X is the conversion, C the initial DCH concentration, and KS the solubility product of Ca(OH)2 in water. Reactions 59 and 60 can be classified as ring closure reactions, and both occur with the following mechanism (as reported in ref 24 and references therein): The kinetic parameters obtained for the pseudo-first-order reactions are reported in Table 11. The same table also gives Table 11. Kinetic Constants for Dehydrochlorination Reactions and Epichlorohydrin Degradation in an Alkaline Environment Created by Ca(OH)2a reagent

This mechanism can be regarded as an internal nucleophilic substitution (SN2), preceded by a base-catalyzed dissociation equilibrium. By assuming the ring closure (rrc) to be the RDS, it is possible to write the following rate law: −

1,3-DCH 1,2-DCH 1-MCH EPY a

−

rrc = k rcK i[MCH][OH ] = k′[MCH][OH ] ≈ kapp[MCH] (63)

pre-exponential factor [s−1]

activation energy [cal/mol]

× × × ×

11718 16984 13200 18852

1.00 6.40 3.74 4.92

107 108 107 108

Data taken from Carrà et al.24

the kinetic parameters for glycidol formation, by assuming a direct conversion from epichlorohydrin to glycidol, given that the dehydrochlorination of 1-MCH is very fast. It is very important to point out that 1,3-DCH is much more reactive than 1,2-DCH; therefore, the residence times for these two reactants are very different.

where [MCH] is the reagent concentration, Ki is the equilibrium constant for the formation of the ionic intermediate ion, k is the true reaction rate constant, while kapp is the apparent kinetic constant. By assuming a constant [OH−] value, corresponding to the Ca(OH)2 solubility, it is possible to interpret 8958



Review

Figure 15. A simplified scheme of glycerol hydrochlorination.

5. GLYCEROL HYDROCHLORINATION REACTORS AND PROCESSES Until now, three companies (Dow Chemicals, Solvay EPICEROL Technology, and CONSER SpA ECH-EF = Eco Friendly)) have developed their own process for producing dichlorohydrins from glycerol and HCl, using different catalysts (different carboxylic acids or mixture of them),22 and some plants have recently been constructed or are still under construction in China (Solvay, DOW Chemicals, CONSER SpA), Thailand (Solvay), and Korea (Dow Chemicals). All of these plants need the following, as unit operations: (i) a well-mixed gas−liquid reactor, in which glycerol, containing the catalyst in an opportune concentration, is contacted with gaseous HCl (under moderate pressure, 1−10 bar) at the reaction temperature (100−120 °C); (ii) a heat exchanger to remove the reaction heat, given that the reaction is moderately exothermic (the heat exchanger can also be located inside the reactor); (iii) a stripping unit, to separate the unreacted HCl and, eventually, also the water produced as a consequence of the occurred hydrochlorination reactions; (iv) the separation units having the scope of recovering pure dichlorohydrins as feedstock for the successive step of epichlorohydrin synthesis. Dichlorohydrins are much more volatile than monochlorohydrins and glycerol and a distillation column is probably sufficient for this separation. The residue of the column will contain monochlorohydrins, small amounts of glycerol, and the catalyst (a carboxylic acid) if its boiling point is high; otherwise, the catalyst must be separated beforehand (for example, in the stripping unit). The residue can be recycled in the gas−liquid reactor. Some undesired byproducts can be formed, and a purge is probably necessary. Based on the list of the previously described unit operations, a simplified scheme of general validity can be sketched (Figure 15). The operation units of this scheme are all conventional systems that do not require particular mention, with the exclusion of the reactor unit requiring (i) a system providing a large gas/ liquid interface area, to avoid HCl mass-transfer limitation; (ii) a good heat exchange, to eliminate the reaction heat maintaining the desired temperature; and (iii) a good mixing system, to warrant a good dispersion of the gas in the liquid phase and a uniform composition. There are some different solutions that can be adopted to satisfy the aforementioned requisites. A large gas/liquid interface area can be realized by using, for example, a falling film reactor with gaseous HCl flowing in countercurrent, a spray tower loop reactor, a Venturi loop reactor, a bubble tray column, or a perforated plate column.

The simplest and most economical solution would be a bubble column. Clearly, the optimal choice is related to the reaction conditions, that is, temperature, pressure, and catalyst concentration determine the reaction rate and, therefore, the HCl consumption. However, the choice is complicated by the corrosiveness of HCl in the presence of water that is formed as a byproduct of the reaction. It is well-known that few materials are compatible with HCl in the presence of moisture (for example, glass, Hastelloy, Teflon, enameled steel); therefore, some of the aforementioned possibilities could be realized with difficulty. The problem of the compatibility of the materials with moisturized HCl is important also for the stripping unit, and very probably, a neutralization is necessary after this unit, to eliminate any trace of this reagent before the separation by distillation of the dichlorohydrins. Finally, it is important to point out that the process via glycerol has an important advantage, with respect to the process via propene, because the hydrochlorination reaction is highly selective toward the formation of 1,3-DCH. To better emphasize the aforementioned advantage, a simulation has been made by Santacesaria et al.11 for comparing the yields of the two processes. A reactive distillation column has been simulated with a commercial process simulation package (Chemcad 5.2) for two different feed streams: the first stream was a mixture with the traditional composition (30% 1,3-DCH and 70% 1,2DCH), coming from a via propene process, while the second one was constituted by pure 1,3-DCH. The simulation of the epichlorohydrin reactive column has been performed by introducing the kinetic expression and parameters taken from Carrà et al.24 and summarized in Table 11, using the model published by Carrà et al.25 In both cases, the reactive column has been simulated by assuming the following: total pressure, 1 bar; reboiler heat duty, 17 000 kcal/h; 15 theoretical plates; liquid holdup, 0.02 L/stage. The comparison between the two different feeds to the column is reported in Figure 16, where epichlorohydrin yields are plotted as a function of the reflux ratio adopted in the reactive column. The yield is the ratio between the distilled epichlorohydrin and the dichlorohydrins used in the feed. As it can be seen, a higher epichlorohydrin yield is always obtained at low reflux ratios with a feed that is constituted by pure 1,3-DCH. Moreover, in order to obtain a satisfactory yield in epichlorohydrin by feeding a 30%/70% mixture, a very high reflux ratio must be used, with, obviously, a greater consumption of energy. 8959



■

REFERENCES

(1) Santacesaria, E.; Martinez Vicente, G.; Di Serio, M.; Tesser, R. Main technologies in biodiesel production: State of the art and future challenges. Catal. Today 2012, 195, 2−13. (2) Pagliaro, M.; Rossi, M. The Future of Glycerol: New Usages for a Versatile Raw Material; RSC Green Chemistry Book Series; Royal Society of Chemistry (RSC): Cambridge, U.K., 2008. (3) Zhou, C. H.; Beltramini, J. N.; Fan, Y. X.; Lu, G. Q. Chemoselective catalytic conversion of glycerol as a biorenewable source to valuable commodity chemicals. Chem. Soc. Rev. 2008, 35, 527−549. (4) Behr, A.; Obendorf, L. Development of a Process for the AcidCatalyzed Etherification of Glycerine and Isobutene Forming Glycerine Tertiary Butyl Ethers. Eng. Life Sci. 2003, 2, 185−189. (5) Jaecker-Voirol, A.; Durand, I.; Hillion, G.; Delfort, B.; Montagne, X. Glycerin for New Biodiesel Formulation. Oil Gas Sci. Technol. 2008, 63, 395−404. (6) Di Serio, M.; Casale, L.; Tesser, R.; Santacesaria, E. Development of a Process for the Acid-Catalyzed Etherification of Glycerine and Isobutene Forming Glycerine Tertiary Butyl Ethers. Energy Fuels 2010, 24, 4668−4672. (7) Melero, J. A.; Van Grieken, R.; Morales, G.; Paniagua, M. Acidic Mesoporous Silica for the Acetylation of Glycerol: Synthesis of Bioadditives to Petrol Fuel. Energy Fuels 2007, 21, 1782−1791. (8) Silva, P. H. R.; Gonçalves, V. L. C.; Mota, C. J. A. Glycerol acetals as anti-freezing additives for biodiesel. Bioresour. Technol. 2010, 101, 6225−6229. (9) Crotti, C.; Farnetti, E.; Guidolin, N. Alternative intermediates for glycerol valorization: iridium-catalyzed formation of acetals and ketals. Green Chem. 2010, 12, 2225−2231. (10) Tesser, R.; Santacesaria, E.; Di Serio, M.; Di Nuzzi, G.; Fiandra, V. Kinetics of Glycerol Chlorination with Hydrochloric Acid: A New Route to αγ-Dichlorohydrin. Ind. Eng. Chem. Res. 2007, 46, 6456− 6465. (11) Santacesaria, E.; Tesser, R.; Di Serio, M.; Casale, L.; Verde, D. New Process for Producing Epichlorohydrin via Glycerol Chlorination. Ind. Eng. Chem. Res. 2010, 49, 964−970. (12) Tesser, R.; Di Serio, M.; Vitiello, R.; Russo, V.; Ranieri, E.; Speranza, E.; Santacesaria, E. Glycerol Chlorination in Gas-Liquid Semibatch Reactor: An Alternative Route for Chlorohydrins Production. Ind. Eng. Chem. Res. 2012, 51, 8768−8776. (13) Atia, H.; Armbruster, U.; Martin, A. Dehydration of glycerol in gas phase using heteropolyacid catalysts as active compounds. J. Catal. 2008, 258, 71−82. (14) Katryniok, B.; Paul, S.; Bellière-Baca, V.; Rey, P.; Dumeignil, F. Glycerol dehydration to acrolein in the context of new uses of glycerol. Green Chem. 2010, 12, 2079−2098. (15) Dasari, M. A.; Kiatsimkul, P. P.; Sutterlin, W. R.; Suppes, G. J. Low-pressure hydrogenolysis of glycerol to propylene glycol. Appl. Catal., A 2005, 281, 225−231. (16) Alhanash, A.; Kozhevnikova, E. F.; Kozhevnikov, I. V. Hydrogenolysis of glycerol to propanediol over Ru: Polyoxometalate bifunctional catalyst. Catal. Lett. 2008, 12, 307−311. (17) Chiu, C. W.; Dasari, M. A.; Sutterlin, W. R.; Suppes, G. J. Removal of residual catalyst from simulated biodiesel’s crude glycerol for glycerol hydrogenolysis to propylene glycol. Ind. Eng. Chem. Res. 2006, 45, 791−795. (18) Johnson, D. T.; Taconi, K. A. The glycerin glut: Options for the value-added conversion of crude glycerol resulting from biodiesel production. Environ. Progress 2007, 26, 338−348. (19) Cortright, R. D.; Davda, R. R.; Dumesic, J. A. Hydrogen from catalytic reforming of biomass-derived hydrocarbons in liquid water. Nature 2002, 418, 964−967. (20) Lee, P. C.; Lee, W. G.; Lee, S. Y.; Chang, H. N. Succinic acid production with reduced by-product formation in the fermentation of anaerobiospirillum succiniciproducens using glycerol as a carbon source. Biotechnol. Bioeng. 2001, 72, 41−48.

Figure 16. Comparison between the epichlorohydrin yields of two different feeds plotted as a function of the reflux ratio adopted in the reactive column.25

6. CONCLUSIONS The synthesis of epichlorohydrin starting from glycerol instead of propene occurs through two steps: the preparation of dichlorohydrins by glycerol hydrochlorination and the successive step of dichlorohydrins dehydroclorination to epichlorohydrin. The second step is well-known, because it is the same for both of the aforementioned processes. The only difference is that, starting from propene, a mixture of the two dichlorohydrins (1,2-DCH (70%) and 1,3-DCH (30%)) is obtained and used as feedstock for the successive dehydrochlorination step, while, starting from glycerol, 1,3-DCH is the main component obtained (more than 95%). This is an advantage of the glycerol route: the residence time can be reduced and, hence, the reactor volume for the dehydrochlorination step also is reduced. In this review, we have collected all data available from the literature, for what concerns the glycerol hydrochlorination reaction, intensively studied in the last years. These data are related to the catalysts that can be used to promote the reaction, their performances, the kinetics of the reactions involved in the hydrochlorination with the related mechanisms, and some aspects that are particular to this gas−liquid system. All this information is fundamental for the choice of the most suitable reactor. The catalysts used until now are carboxylic acids with the preference for compounds of high boiling point, which allows one to recycle the catalyst without problems. Many kinetic data are available but are rather scattered, because they have been collected on different catalysts and interpreted with different kinetic models, based on different reaction mechanisms. Further work is necessary to identify the most-reliable catalytic mechanism and the best catalyst. For example, the use of polyoxometallates as a catalyst is a new promising route that is worth of deeper examination. Lastly, more investigation about the role of mass transfer of HCl and of HCl solubility in the reaction environment would be necessary for correct reactor modeling.

■

Review

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS Thanks are due to CONSER SpA and Eurochem Engineering srl for funding the research. 8960



Review

(21) Papanikolaou, S.; Aggelis, G. Lipid production by Yarrowia lipolytica growing on industrial glycerol in a single-stage continuous culture. Bioresour. Technol. 2002, 82, 43−49. (22) (a) Ullmann’s Encylopedia of Industrial Chemistry, 7th ed.; Wiley−VCH: Weinheim, Germany, 2006. (b) Dow Chemical Company. C&E News 2006 (Aug. 14), 3. (c) Solvay. EPICEROL Process; http://www.solvaychemicals.com/EN/Sustainability/Issues_ Challenges/EPICEROL.aspx. (d) CONSER SpA. ECH-EF Process (Eco Friendly); http://www.conserspa.com/frame_pr-de.htm. (e) de Guzman, D. Growing glycerine-to-ECH plants. ICIS Green Chemistry Jan. 20, 2011; Available via the Internet at: http://www.icis.com/ blogs/green-chemicals/2011/01/growing-glycerine-to-ech-plant/. (23) Sienel, G.; Rieth, R.; Rowbottom, K. T. Epoxides. In Ullmann’s Encyclopedia of Industrial Chemistry; Wiley−VCH: Weinheim, Germany, 2005 (DOI: 10.1002/14356007.a09_531). (24) Carrà, S.; Santacesaria, E.; Morbidelli, M.; Schwarz, P.; Divo, C. Synthesis of Epichlorohydrin by Elimination of Hydrogen Chloride from Chlorohydrins. 1. Kinetic Aspects of the Process. Ind. Eng. Chem. Process. Des. Dev. 1979, 18 (3), 424−427. (25) Carrà, S.; Santacesaria, E.; Morbidelli, M.; Schwarz, P.; Divo, C. Synthesis of Epichlorohydrin by Elimination of Hydrogen Chloride from Chlorohydrins. 2. Simulation of the Reaction Unit. Ind. Eng. Chem. Process. Des. Dev. 1979, 18 (3), 428−433. (26) (a) Ger. Patent 197308, 1906. (b) Ger. Patent 197309, 1906. (c) Ger. Patent 238341, 1908. (27) Hill, A. J.; Fischer, E. J. A; Synthesis of beta-chloro-allyl chloride. J. Am. Chem. Soc. 1922, 44, 2582−2595. (28) Britton, E. G.; Heindel, R. X. Preparation of glycerol dichlorohydrin. U.S. Patent 2,144,612, 1939. (29) Britton, E. C.; Slagh, H. R. Glycerol dichlorohydrin. U.S. Patent 2,198,600, 1940. (30) Clarke, H. T.; Hartman, W. W. Epichlorohydrin. Org. Synth., Coll. 1923, 3, 47. (31) Conant, J. B.; Quayle, O. R. Glycerol α-monochlorohydrin. Org. Synth., Coll. 1922, 2, 33. (32) Conant, J. B.; Quayle, O. R. Glycerol α,γ-dichlorohydrin. Org. Synth., Coll. 1922, 2, 29. (33) Thompson, W. P. Recovery of aliphatic chlorohydrins having from 3 to 5 carbon atoms from aqueous solution. Br. Patent GB 931,211, 1963. (34) Krafft, P.; Gilbeau, P.; Gosselin, B.; Claessens, S. (Solvay). Process for producing dichloropropanol from glycerol, the glycerol coming eventually from the conversion of animal fats in the manufacture of biodiesel. PCT Patent WO 054167 A1, 2005. (35) Schreck, D.; Kruper, W.; Varjian, R. D.; Jones, M. E.; Campbell, R. M.; Kearns, K.; Hook, B. D.; Briggs, J. R.; Hippler, J. G. (Dow Global Tech., Inc.). Conversion of multihydroxylated-aliphatic hydrocarbon or ester thereof to a chlorohydrin. PCT Patent WO 2006/020234 and Eur. Patent EP 1 771 403 B1, 2007. (36) Siano, D.; Santacesaria, E.; Fiandra, V.; Tesser, R.; Di Nuzzi, G.; Di Serio, M.; Nastasi, M. (Eurochem Engineering srl). Process for the production of α,γ-dichlorohydrin from glycerin and hydrochloric acid. WO 111810 A2, 2006; U.S. Patent US 2009/0062574, 2009; Eur. Patent EP 1879842 B1, 2012; Chin. Patent ZL 2006, 80012864.9, 2006. (37) Kubicek, P.; Sladek, P.; Buricova, I. (Spolek Pro Chemickou). Method of preparing dichloropropanols from glycerine. World Patent WO2005/021476 and U.S. Patent US 2007/0167659, 2007. (38) Santacesaria, E.; Di Serio, M.; Tesser, R. (Eurochem Engineering srl). Process for monochlorohydrins production from glycerol and hydrochloric acid. PCT Patent WO 132770 A1, 2008. (39) Kruper, W. J., Jr.; Arrowood, T.; Bell, B. M.; Briggs, J. ; Campbell, R. M.; Hook, B.; Nguyen, A.; Theriault, C.; Fitschen, R. Batch, semi-continuous or continuous hydrochlorination of glycerol with reduced volatile chlorinated hydrocarbon by-products and chloroacetone level. U.S. Patent US 7,906,690 B2, 2011. (40) Tirtowidjojo, D.; Merenov, A. S.; Kneupper, C. D.; Hook, B. D.; Metha, A. L. (Dow Chemical Co.). Multi-stage process and apparatus

for recovering dichlorohydrins. U.S. Patent US 2010/0137652A1, 2010. (41) Di Serio, M.; Santacesaria, E.; Tesser, R. (Eurochem Engineering srl). Process for monochlorohydrins production from glycerol and hydrochloric acid. World Patent WO 2008/132770 A1. (42) Bell, B. M.; Briggs, J. R.; Campbell, R. M.; Chambers, S. M.; Gaarenstroom, P. D.; Hippler, J. G.; Hook, B. D.; Kearns, K.; Kenney, J. M.; Kruper, W. J.; Schreck, D. J.; Theriault, C. N.; Wolfe, C. P. Glycerin as a Renewable Feedstock for Epichlorohydrin Production. The GTE Process. Clean 2008, 36, 657−661. (43) Lee, S. H.; Park, D. R.; Kim, H.; Lee, J.; Jung, J. C.; Woo, S. Y.; Song, W. S.; Kwon, M. S.; Song, I. K. Korean J. Chem. Eng. 2008, 25 (5), 1018−1021. (44) Lee, S. H.; Park, D. R.; Kim, H.; Lee, J.; Jung, J. C.; Woo, S. Y.; Song, W. S.; Kwon, M. S.; Song, I. Catal. Commun. 2008, 9, 1920− 1923. (45) Lee, S. H.; Park, D. R.; Kim, H.; Lee, J.; Jung, J. C.; Park, K. M.; Cho; Song, I. K. React. Kinet. Catal. Lett. 2008, 94, 71. (46) Misono, M. Heterogeneous Catalysis by Heteropoly Compounds of Molybdenum and Tungsten. Catal. Rev. Sci. Eng. 1987, 29, 269. (47) Hill, C. L.; Prosser-McCartha, C. M. Homogeneous catalysis by transition metal oxygen anion clusters. Coord. Chem. Rev. 1995, 143, 407−455. (48) Mizuno, N.; Misono, M. Heterogeneous Catalysis. Chem. Rev. 1998, 98 (1), 199−218. (49) Okuhara, T.; Mizuno, N.; Misono, M. Catalytic Chemistry of Heteropoly Compounds. Adv. Catal. 1996, 41, 113−252. (50) Kozhevnikov, I. V. Catalysis by Heteropoly Acids and Multicomponent Polyoxometalates in Liquid-Phase Reactions. Chem. Rev. 1998, 98 (1), 171−198. (51) Song, S. H; Lee, S. H.; Park, D. R.; Kim, H.; Woo, S. Y; Song, W. S.; Kwon, M. S.; Song, I. K. Direct preparation of dichloropropanol from glycerol and hydrochloric acid gas in a solvent-free batch reactor: Effect of experimental conditions. Korean J. Chem. Eng. 2009, 26 (2), 382−386. (52) Lee, S. H.; Song, S. H.; Park, D. R.; Jung, J. C.; Song, J. H.; Woo, S. Y.; Song, W. S.; Kwon, M. S.; Song, I. K. Solvent-free direct preparation of dichloropropanol from glycerol and hydrochloric acid gas in the presence of H3PMo12−XWXO40 catalyst and/or water absorbent. Catal. Commum. 2008, 10, 160−164. (53) Graham Solomons, T. W. Organic Chemistry, 7th ed.; John Wiley & Sons: New York, 1999; pp 741−744. (54) March, J. Advanced Organic Chemistry, 5th ed.; Wiley−Interscience: New York, 2001; pp 308−312. (55) Boschan, R.; Winstein, S. The Role of Neighboring Groups in Replacement Reactions. XXI. Front-Side Participation of the Acetoxy Group. Catalytic Effect of Acetic Acid on the Reaction of Glycols with Hydrogen Chloride. J. Am. Chem. Soc. 1956, 78, 4921−4925. (56) Dmitriev, S. G.; Zanaveskin, N. L. Hydrochlorination of GlycerolThe Role of the Water on the Process. J. Chem. Chem. Eng. 2011, 5, 1179−1182. (57) Luo, Z. H.; You, X. Z.; Li, H. R. Direct Preparation Kinetics of 1,3-Dichloro-2-propanol from Glycerol Using Acetic Acid Catalyst. Ind. Eng. Chem. Res. 2009, 48, 446−452. (58) Luo, Z. H.; You, X. Z.; Li, H. R. A kinetic model for glycerol chlorination in the presence of acetic acid catalyst. Korean J. Chem. Eng. 2010, 27 (1), 66−72. (59) Lim, J. H.; Song, W. S.; Woo, S. Y.; Lee, D. H. Kinetic model of glycerol chlorination with hydrochloric acid. Korean J. Chem. Eng. 2010, 27 (3), 785−790. (60) Ling, D. Q.; Lu, J.; Wang, M. X.; Liang, S.; Zhang, W.; Ren, J.; Ouyang, C. P. Investigation of the kinetics and mechanism of the glycerol chlorination reaction using gas chromatography−mass spectrometry. J. Serb. Chem. Soc. 2010, 75 (1), 101−112. (61) De Araujo Filho, C. A.; Salmi, T.; Bernas, A.; Mikkola, J. P. Kinetic Model for Homogeneously Catalyzed Halogenation of Glycerol. Ind. Eng. Chem. Res. 2013, 52, 1523−1530. 8961



Review

(62) Macey, R.; Oster, G. Berkeley Madonna’s Modeling and Analysis of Dynamic Systems, release 8.3; University of CaliforniaBerkeley: Berkeley, CA, 2001. (63) Dmitriev, S. G.; Zanaveskin, N. L. Synthesis of Epichlorohydrin from Glycerol. Hydrochlorination of Glycerol. Chem. Eng. Trans. 2011, 24, 43−48. (64) Patai, S. The Chemistry of Functional Groups; Interscience: New York, 1967.

8962


Chemical and Technical Aspects of the Synthesis of Chlorohydrins

Recommend Documents