Clean 2008, 36 (8), 657 – 661 Research Article
Glycerin as a Renewable Feedstock for Epichlorohydrin Production.The GTE Process
BruceM. Bell1 ，John R. Briggs1，RobertM. Campbell1
SusanneM. Chambers1，Phil D. Gaarenstroom1
Jeffrey G. Hippler1，Bruce D. Hook1，Kenneth Kearns1 JohnM. Kenney1，WilliamJ. Kruper1
D. James Schreck1，Curt N. Theriault1，Charles P.Wolfe1
1
The Dow Chemical Company, Midland,MI, USA.
A significant improvement in a process to produce epichlorohydrin through the use of glycerin as renewable feedstock is presented. The glycerin to epichlorohydrin (GTE) process proceeds in two chemical steps. In the first step, glycerin is hydrochlorinated with hydrogen chloride gas at elevated temperature and pressure to a mixture of 1,3-DCH(1,3-dichlorohydrin,1,3-dichloropropan-2-ol) and 2,3-DCH (2,3-dichlorohy- drin,2,3-dichloropropan-1-ol), using a carboxylic acid catalyst. In the second step, the mixture of dichlorohydrins is converted to epichlorohydrin with a base. This solventless process represents an economically and environmentally advantageous, atom-efficient process to an existing commodity chemical that can employ a renewable resource for its primary feedstock.
Keywords: Glycerin;Renewable;Feedstock; Green Chmistry; Epichlorohydrin Production
Received:March 12, 2008; revised: April 8, 2008; accepted: April 9, 2008 DOI: 10.1002/clen.200800067
1 Introduction
Epichlorohydrin is a high volume commodity chemical used largely in epoxy resins, although smaller quantities have, until recently, been employed for the manufacture of synthetic glycerin [1]. Although several routes are known for epichlorohydrin manufacture, most is made from propylene and chlorine as primary raw materials in a multi-step process. Shown in Fig. 1, this requires the allylic chlorination of propylene to allyl chloride followed by hypochlorination to give a 3:1 mixture of 1,3-DCH and 2,3-DCH, which is then treated with base to yield epichlorohydrin [1]. Although practiced on a very large scale, this process suffers from some undesirable features, particularly the low chlorine atom efficiency. Only one of the four chlorine atoms employed in themanufacture of epichlorohydrin by this route is retained in the product molecule, the remainder emerging as by-product hydrogen chloride or waste chloride anion. Additionally, inefficiencies in the chlorination and hypochlorination steps lead to the formation of unwanted chlorinated organics that are expensive to dispose of. Such factors have prompted the search for alternative routes to epichlorohydrin that are more atom-efficient and environment-friendly. The escalating cost of petrochemical raw materials such as propylene has also contributed to the accelerated search for processes that employ less expensive rawmaterials.
One such route that has been recently examined by us [2] and others [3, 4] is based on the conversion of glycerin through dichlorohydrins to epichlorohydrin. This two-step process, shown in Fig. 2, appears significantly simpler than the incumbent process, but the historically high cost of glycerin has prevented its development as a commercial process. Recently, however, glycerin has become increasingly available as a by-product of the manufacture of biodiesel, particularly in Europe. As a result, the available volume of renewable glycerin has risen, and the price has declined to a point where its use in the manufacture of commodity chemicals, such as epichlorohydrin, has become feasible. Several companies have announced plans to commercialize technology to manufacture epichlorohydrin from glycerin (Dow Chemical, C&E News, August 14,2006, p. 3, Solvay www.solvaypress.com/pressrele-
ases/0,52477-2-0,00.htm,Spolchemiewww.spolchemie.cz/dwn/factsheet12.pdf).Epich-lorohydrin is but one of several new opportunities that have been recognized as a viable use of increasingly plentiful, low cost glycerin [5], and exemplifies a more general trend of an expanding use of natural polyols for themanufacture of commodity chemicals[6].
Figure 1. The dominant commercial route to epichlorohydrin is a multi-step process comprising the initial allylic chlorination of propylene to allyl chloride. This is then reacted with hypochlorous acid, made by dissolving chlorine in water, to yield a 3:1 mixture of 1,3-dichloropropan-2-ol and 2,3-dichloropropan-1-ol in dilute aqueous solution. This mixture is reacted with base to give epichlorohydrin. Of the four equivalents of chlorine atoms employed, only one is retained in the desired product, the remaining three equivalents appearing as by-product HCl or waste chloride ion.
Figure 2. A route to epichlorohydrin that employs renewable glycerin as feedstock is a two-step process comprising initial hydrochlorination of glycerin with hydrogen chloride to give a 30 – 50:1 mixture of 1,3-dichloropropan-2-ol and 2,3-dichloropropan-1-ol, followed by reaction with base.This process produces only one equivalent of waste chloride.
The carboxylic acid catalyzed hydrochlorination of glycerin to dichlorohydrins has been known for over a century [7], and the mechanism of the reaction was delineated some 50 years ago [8]. An extensive summary of the early literature up to 1930s can be found in a series of papers by Gibson [9]. In this paper, we will describe elements of the glycerin hydrochlorination step of the glycerin to epichlorohydrin (GTE) process, a technology that represents an economically and environmentally advantageous route to epichlorohydrin froma renewable carbon resource.
2 Experimental
Experiments were performed in a 100 mL, Hastalloy C( Parr auto-clave equipped with a Magnedrive stirrer, a thermocouple, and internal cooling coils. Glycerol (Aldrich 99%, or Interwest Corporation) was added to the reactor, followed by carboxylic acid and water, and the reactor sealed. The mass of the reactor and contents were recorded. The reactor was stirred and water at ambient temperature cycled through the cooling coils. Hydrogen chloride gas (Airgas Corporation) at the desired pressure was admitted to the reactor, resulting in a 15–25OC exotherm. The reactor was heated to the desired temperature, and the reaction then allowed to proceed for the desired length of time, while hydrogen chloride gas was fed continuously at the set pressure as it was consumed by reaction. The mass of hydrogen chloride fed to the reactor was determined by the change in the mass of the cylinder throughout the reaction, or in some cases using a calibrated mass-flow controller. In some instances, samples were withdrawn from the reactor through a bottom valve for analysis. After the desired reaction time had elapsed, the hydrogen chloride feed was ceased, and the reactor and contents cooled to room temperature. The reactor was then vented and the mass of the reactor and contents were recorded.
All other chemicals used were of the highest purity commercially available and were used without purification.
Samples and standards were analyzed by gas chromatography (GC). Products for which no GC standard sample was available were identified by GC/mass spectrometry.
The sample and standards were analyzed using an Agilent 6890 GC system with the following conditions:
Column: DB-5.30m×0.25mm×1.0µL film, S/N:9118515 Injection: Split Injection Volume: 0.2µL Detection: FID Carrier Gas: He Carrier Pressure: 24 psi Split: 125 mL/min Hydrogen: 30 mL/min Air: 350 mL/min Makeup: 25 mL/min Injector Temperature: 200OC Detector Temperature: 300OC Temperature Program:
Initial Temperature: 40OC for 6 min Ramp Rate: 10OC/min Final Temperature: 180OC for 0 min Ramp Rate A: 30OC/min Final Temperature A: 300OC for 1 min
Response factors were calculated using the following equation:
RF(Analy WteeighAt)n/a(ly Atere)aPurity(Inter nSatlanda rWdeighItn)t/e(rn Satlanda Ardrea) （1）
ISTD weight1wt% concentration =RFArea100%ISTD AreaSample Weight （2）
The retention times of the major components (in minutes) are: 1,3-DCH (13.7); TCP (14.2); 2,3-DCH (14.6); 1-monochlorohydrin (1-MCH) (14.8); 2-monochlorohy- drin (2-MCH) (15.5); glycerin (15.6–16.2).
3 Results and Discussion
Treating glycerin containing 2 wt % of a carboxylic acid catalyst with hydrogen chloride at slightly above atmospheric pressure (20 psi) and 120OC in a sealed vessel results in its conversion initially to mainly 1-MCH (1-MCH, 1-chloropropane-2,3-d- iol).Much smaller amounts of 2-MCH (2-MCH, 2-chloropropane-1,3-diol) are formed. In a much slower, low conversion reaction, the 1-MCH is converted mainly to 1,3-DCH withmuch smaller amounts of 2,3-DCH. The evolution of the major products in such a reaction is shown in the first plot in Fig. 3. The low conversion to DCHs obtained in a glycerin hydrochlorination process operated at atmospheric pressure appears to have resulted in a number of approaches to improve this process. Efforts adopted in the literature have included sparging hydrogen chloride gas through the reaction solution where upon water is also removed from the solution [10], employing an azeotroping agent to facilitate water removal [11], and employingmultiple reaction stages with interstage water removal [4]. Sparging hydrogen chloride gas through the reaction solution, or the use of an azeotrope to remove water from the reaction medium is expensive and therefore less desirable on a commercial scale. In the first case, large excesses of hydrogen chloride gas are employed, and the hydrogen chloride that is recovered is contaminated with water and must be separated for reuse in the process, disposed of, or used for lower value applications. The use of an azeotroping agent to remove water results in amuchmore complex process in which the solvent must be separated from the water and hydrogen chloride and recovered for reuse. The use of multiple reaction stages results in increased equipment costs and process complexity.
Because the prior literature suggested that the glycerin hydrochlorination reaction exhibited an equilibriumlimitation, we were led to explore the effect of higher hydrogen chloride concentration on the reaction conversion, rate, and selectivity. This was achieved experimentally by employing higher applied pressures of hydrogenchloride gas. We expected that this would both speed up the hydrochlorination, and drive the reaction to higher conversion to the desired
dichlorohydrins while minimizing the stoichiometric excess of hydrogen chloride required to achieve the desired, high conversion. Figure 4 shows the effect of hydrogen chloride pressure on the rate of hydrogen chloride uptake by a glycerin solution containing 2 wt % of acetic acid as catalyst.
Figure 3. Evolution of major reaction products in a solution of 2 wt% acetic acid in glycerin (
1,3-DCH) with time and as a function of applied
hydrogen chloride pressure at 120OC, showing that at the lower pressures, the reaction has essentially achieved equilibrium by the end of the 4 hreaction time, leaving substantial unconverted glycerin and 1-MCH. Only as the hydrogen chloride pressure is increased to 50 psi is substantial 1,3-DCH produced.
Two features of the plots in Fig. 4 are striking. After a very rapid, initial consumption of hydrogen chloride by each reactionmixture,which we attribute to dissolution of hydrogen chloride in the glycerin, the rate of consumption of hydrogen chloride by the reaction increases with the applied pressure of hydrogen chloride. At each pressure, the rate of consumption of hydrogen chloride is initially fast but slows
and largely ceases after a certain time. Thereafter, little or no additional hydrogen chloride is consumed by the reaction solutions even at extended reaction times. The total amount of hydrogen chloride taken up by the solutions also increases as the applied pressure increases. This behavior appears consistent with an equilibrium limited reaction that is driven to higher conversion at higher pressure.
Figure 4. Plots of hydrogen chloride gas absorbed by a mixture of 85 wt% glycerin, 9 wt % water, and 7 wt % acetic acid at 908C at different applied constant hydrogen chloride pressures show faster and greater uptake at higher pressure, suggesting an equilibrium limitation on the hydrochlorination reaction.
Figure 3 shows the evolution of the major components of three reactions run at different applied hydrogen chloride pressures determined by gas chromatographic analysis of samples taken during the reactions. At P(HCl)= 20 psi, the initial consumption of glycerin and the formation of 1-MCH is rapid, but the reaction slows dramatically, and the formation of 1,3-DCH is limited. As the pressure is increased to 30 and 50 psi, it can be seen that the conversion of 1-MCH to 1,3-DCH becomesmore
significant. These plots confirmthat the reaction is equilibrium limited at low pressure, and that this limitation can be largely overcome by increasing the applied hydrogen chloride pressure.
At even higher pressures, the reaction is fast and the formation of dichlorohydrins becomes very efficient. At 110 psi and 110OC, the hydrochlorination of wet glycerin (9 wt % water) in the presence of 5 mol% of acetic acid as catalyst is both rapid and efficient, so that after 4h in a sealed vessel to which hydrogen chloride gas is fed on demand, the reaction product mixture comprises 93 mol% of dichlorohydrins and their acetate esters, in a ratio of 46:1 of 1,3-DCH to 2,3-DCH, and 6 mol% monochlorohydrins and their esters in a ratio of 1:2 of 1-MCH to 2-MCH. 2-MCH attains a higher concentration than 1-MCH at high conversions because, although it is formed more slowly than 1-MCH, it is also hydrochlorinated much more slowly than 1-MCH.
Figure 5. The mechanism of the carboxylic acid catalyzed hydrochlorination of glycerin to mono and dichlorohydrins, as presented in ref. [7]. Initialesterification preceeds ring-closure and loss of water to form an acetoxonium cation. Ring-opening of the acetoxonium cation by a chloride
generates 1-MCH. Repetition of the process at the remaining two hydroxyl groups gives the desired 1,3-DCH product. This mechanism precludes the formation of 1,2,3-trichloropropane since two hydroxyls are required to form the acetoxonium cation.
A significant advantage of the GTE process compared with the incumbent propylene process of Fig. 1 is the improved regiochemistry observed in the dichlorohydrin products of the glycerin hydrochlorination reaction over that achieved during the hypochlorination of allyl chloride. This is manifested in the 1,3-DCH/2,3-DCH ratio, which is 30 –50 in the GTE process, but only about 3:1 in the incumbent process. The value of improved 1,3-regioselectivity is that 1,3-DCH undergoes cyclization with base to form epichlorohydrin about 300 times faster than does 2,3-DCH [12]. This translates into smaller process equipment and/or shorter residence times. Shorter residence times can result in fewer reaction by-products, higher process efficiency, and a purer final product.
A second advantage of the high pressure GTE process is the intentional accumulation of water mitigates the formation of some by-products, including chlorinated ethers and chloroacetone [13]. This reduces the processing and waste disposal costs for these materials. Retaining water in the GTE process, while still achieving high levels of conversion of MCH to DCH is possible because the increased hydrogen chloride concentration significantly counteracts the equi-librium limitations seen at atmospheric pressure.
An important feature of the carboxylic acid catalyzed hydrochlorination of glycerin that makes the development of a commercial process viable is that the reaction effectively stops at the dichlorohydrin stage – there is no exhaustive hydrochlorination to TCP (1,2,3-trichloropropane). Should a significant rate to TCP occur in this chemistry, an efficient process to DCH would be very difficult to operate commercially because it would require the recovery of the desired DCH at relatively low concentration from the product mixture. Additionally, the formation of TCP would represent a loss of glycerin efficiency, and a significant economic loss because of the need for its disposal.
The absence of TCP in these hydrochlorination reactions is an inevitable consequence of the mechanism that is believed to be operative in this chemistry. A proposed mechanism of the carboxylic acid catalyzed hydrochlorination of polyols has been available for 50 years, and is shown in Fig. 5. In thismechanism, one hydroxyl group is esterified by the carboxylic acid catalyst. The cyclization of this ester with an adjacent hydroxyl group is followed by proton-assisted loss of water, and results in the formation of an acetoxonium cation. Ring opening of this acetoxonium cation with chloride ion, regioselectively at the primary position gives the observed major products. The intermediacy of the acetoxonium cation requires two available adjacent hydroxyls for this mechanism to be operative, and thus provides an explanation for why the glycerin hydrochlorination stops at the dichlorohydrin stage. Further confirmation of thismechanismis that under the hydrochlorination conditions either ethylene glycol or propylene glycol is only monohydrochlorinated to the chlorohydrin. This is particularly significant since these chlorohydrins can be employed for the synthesis of the corresponding epoxides by reaction with base. This makes the hydrochlorination reaction of a number of polyols, such as those derived from sugar hydrogenolysis, for example, particularly valuable for the conversion of the polyol into higher valued products which are currentlymanufactured using the epoxides.
It is known fromthe prior art that a variety of carboxylic acids are effective catalysts for the hydrochlorination reaction, although the vast majority of studies have employed acetic acid [6–8]. While acetic acid performs well in laboratory-scale batch reactions, and can be used in appropriately configured commercial processes, its volatility is too high for some recycle process configurations. Among the major products from the hydrochlorination of glycerin, the desired dichlorohydrins, 1,3-DCH and 2,3-DCH are themost volatile. This makes a recycle process, in which the DCHs are stripped from the product streamafter reaction, and recycle of any unreacted glycerin or MCHs and catalyst back to reaction, particularly attractive. For this option to be viable, it is preferred that the catalyst, and its esters with the products or intermediates, should be less volatile than the DCHs, so that they remain in the
stripper bottoms for easy recycle. Carboxylic acid catalysts containing six or more carbonatomsmeet this requirement.We have examined the structural features of a number of such carboxylic acids to delineate those structural features that lead to satisfactory catalytic performance in the hydrochlorination reaction.
Figure 6 shows the rate of hydrogen chloride uptake versus time in glycerin hydrochlorination reactions employing four different carboxylic acids, along with the corresponding plots for acetic acid and a reaction with no added catalyst. Of these five reactions, the acetic acid catalyzed reaction is fastest and total consumption of hydrogen chloride is greatest after 4 h. n-Hexanoic acid is initially slower, but after 4 h has consumed about the same amount of hydrogen chloride. Gas chromatographic analysis of the reaction products after these reactions confirms that the conversion to dichlorohydrins is essentially the same in these two cases. The plot of HCl uptake for the n-hexanoic catalyzed reaction appears to show an induction period which we attribute to mass-transfer limited esterification of the glycerin by the n-hexanoic acid catalyst, which is immiscible with the glycerin initially. The rate of hydrogen chloride uptake is slower for 3-methylvaleric acid and almost at the uncatalyzed rate in the case of 3,3-dimethylbutanoic acid. These rates appear to correlate with the steric bulk [14] of the carboxylic acids, with the less sterically hindered acids giving faster reactions. However, the pKas for these acids also correlate with this steric parameter. To break this correlation, we also examined the performance of trimethylammonium acetic acid chloride (pKa = 1.76) as a catalyst.The rate of HCl uptake in a reaction using this material as a catalyst is very slow, and quite similar to that observedwith 3,3-trimethylbutanoic acid, even though its pKa indicates it is by far the most acidic carboxylic acid catalyst of those acids shown in Fig. 6. This leads us to propose that it is the steric bulk of the catalyst that largely determines the catalyzed rate of hydrochlorination of glycerin.
Figure 6. Plots of hydrogen chloride gas absorbed by 3 mol% carboxylic acid in glycerin at 1008C and 110 psi applied HCl pressure. As the carboxylic acid becomes more sterically hindered, the rate of HCl uptake declines.
4 Conclusions
The interest in the commercialization of a GTE process demonstrates the ascendancy of renewable feedstocks in the manufacture of commodity chemicals. The increasing availability of cheaper glycerin, as a result of the rapidly growing biodiesel industry,makes it aviable feedstock in this application. As recently as 2006, synthetic glycerin wasmanufactured fromepichlorohydrin, but this has now completely ceased. Further development of technology for utiliza-tion of alternative, low-cost, renewable feedstocks, for example, sugar or cellulose, will likely result in the emergence of additional feedstock switches for existing commodity petrochemicals.
The GTE process is an economically advantaged route to epichlorohydrin that additionally features a number of environmentally desirable attributes compared to the incumbent process. These include a switch of chloride feedstock from elemental chlorine to hydrogen chloride, which can be obtained as the by-product from a number of commercial processes, including the incumbent epichlorohydrin process.
Additionally, the GTE process exhibits improved atom efficiency, less waste water, and lower levels of chlorinated organic by-products, and, of course, the use of competitively priced, renewable glycerin feedstock.
Acknowledgements
We thank Erin O'Driscoll and Ernesto Occhiello for supporting this Work.
References
[1] K.Weissermel, H.-J. Arpe, Industrial Organic Chemistry, 3rd Edn.,Wiley-VCH,Weinheim 1997, p. 294, 299.
[2] PCT application WO 2006020234, published February 23, and assigned to 2006 to Dow Global Technologies, Inc.
[3] PCT application WO 2005054167, published June 16, 2005 and assigned to Solvay. [4] PCT application WO 200521476, published March 10, 2005 and assigned to to Spolek. [5] A. Behr et al., Green Chem. 2008, 10 (1), 13 – 30. [6] M. Schlaf, Dalton Trans. 2006, 39, 4645 – 4653. [7] J. Carius, Justus Leibigs Ann. Chem. 1862, 122, 73. [8] R. Boschan, S.Winstein, J. Am. Chem. Soc. 1956, 78, 4921. [9] G. P. Gibson, Chem. Ind. 1931, 20, 949 – 975.
[10] J. B. Conant, O. R. Quayle, Org. Synth., 1941, CV1, 292 – 297. [11] Dow Chemical Company, US Patent 2,144,512, January 24, 1939.
[12] S. Carra, E. Santacesaria, M. Morbidelli, P. Schwarz, C. Divo, Ind.Eng.Chem. Proc. Des. Dev. 1979, 18 (3), 424 – 427.
[13] Dow Global Technologies, Inc., US Patent Appl. Publ. 20080015369, January 17, 2008. [14] T. Komatsuzaki, I. Akai, K. Sakakibara, M. Hirota, Tetrahedron 1992 48(9), 1539.
附录（二） 环氧氯丙烷相关英文文献资料翻译
研究性论文
可再生原料甘油合成环氧氯丙烷的工艺
BruceM. Bell1 ，John R. Briggs1，RobertM. Campbell1 SusanneM. Chambers1，Phil D. Gaarenstroom1
Jeffrey G. Hippler1，Bruce D. Hook1，Kenneth Kearns1 JohnM. Kenney1，WilliamJ. Kruper1
D. James Schreck1，Curt N. Theriault1，Charles P.Wolfe1
通过利用可再生原料甘油来合成环氧氯丙烷能够显著改善其生产工艺这个方法已经被人们所提及。甘油合成环氧氯丙烷的过程需经过两个化学步骤。第一步，甘油在较高的温度和压力下和氯化氢气体发生反应生成1，3-二氯2-羟基丙醇和2，3-二氯1-羟基丙醇的混合物，用羧基酸来作为催化剂。第二步，这些氯氢化混合物在加上一个羟基以后转变成环氧氯丙烷。这种无溶剂的反应过程可以利用可再生资源作为它的原材料来合成所需的化学品，而且它能够有效利用参与反应的原子，更为经济与环保。
关键词：甘油；可再生；原料，绿色化学；环氧氯丙烷生产
收录时间:2008.3.12 修订时间:2008.4.8 通过时间:2008.4.9 编号:10.1002/clen.200800067
1.介绍
环氧氯丙烷是一种大体积的商品化合物，被广泛应用于合成环氧树脂，不过量比较少，而且直到最近还多被用于生产甘油。虽然合成环氧氯丙烷的方法有很多种，但是它主要还是以丙烯和氯气为主要原材料通过一个多步骤的反应合成。如图1所示，丙烯高温氯化生成氯丙烯，氯丙烯次氯酸化生成1，3-二氯丙醇和2，3-二氯丙醇，体积比为3：1，最后二氯丙醇皂化生成环氧氯丙烷。但是在很大程度上，这个过程会产生一定的副产品，而且对于氯原子的利用率也很低。只有四分之一的氯原子参与到了环氧氯丙烷的合成并被保留在分子当中，其余的却都形成了副产品氢氯化物和没用的氯阴离子。另外，由于在氯化和次氯酸化过程中氯原子会被吸收形成含氯的有机化合物，这些需要花费昂贵的代价去进行处
理。这些因素促使人们去研究其它可以更为有效的利用氯原子和更加环保的工艺来合成环氧氯丙烷。而且由于石油化合物类似丙烯消耗的急剧增加，也迫使我们尽快研究出能够不使用价格昂贵的原料也能合成环氧氯丙烷的方法。
图1.合成环氧氯丙烷的主要路线还是通过丙烯氯化生成氯丙烯，然后与次氯酸反应，生成体积比为3：1的1，3-二氯丙醇和2，3-二氯丙醇的稀溶液，参与反应的四个氯原子只有一个被保留在合成产物中，其余的氯原子生成副产品盐酸或成为没用的氯阴离子。
最近我们已经研究出一种合成路线，就是基于利用甘油的氯化来转换生成环氧氯丙烷。这个只需两步骤的反应如图二所显示，它相比现有的其他合成方法过程更为简单，但是由于甘油的价格过于昂贵，阻碍了它变成一种商品化生产过程的发展。然而近来，尤其是在欧洲，甘油作为生产生物柴油过程中的副产品，产量已经越来越多。所以，能利用的可再生的甘油原料量在不断增加，因而导致它的价格不断下降，利用它来合成制造大宗化学品比如环氧氯丙烷的想法已经可行。一些公司已经宣布了利用相关的工业技术由甘油来生产环氧氯丙烷的计划 (陶氏化工，2006.8.14。相关网站www.solvaypress.com/pressreleases/0，52477-2-0，00.htm，www.spolchemie.cz/dwn/factsheet12.pdf)。环氧氯丙烷就是几种中的一个已经被认知的新机会，能够利用日益增多的低价的甘油，而且这也代表了一种趋势，那就是我们能够利用天然的多元醇来生产大宗化学品。
图2.另一种合成环氧氯丙烷的路线是利用可再生资源甘油做原料，通过一个两步骤的反应合成，这个反应包括甘油与盐酸反应生成体积比为30-59：1的1，3-二氯丙醇和2，3-二氯丙醇，在与氢氧化钠反应之后环氧化。这个过程只会产生一个废弃的氯离子。
羧酸能够作为甘油氯化反应的催化剂早在一个世纪以前就被知晓了，而这个反应机理在50年以前已经被研究出来。对此较早的文献记录可以追溯到1930年吉布森发表的一系列文献。本文中，我们将会探讨如何利用甘油来合成环氧氯丙烷的工艺，这项技术描述了如何利用那些可再生的含碳原料去更为经济和环保地合成环氧氯丙烷的工艺路线。
2.实验
实验在一个容积为100毫升的装置中进行，该装置配备一个搅拌器，一个热电偶,和内部冷却线圈。依次加入甘油，羧酸催化剂和水，开始反应。记录实验过程反应物的加入量以及相关现象。反应过程中不断进行搅拌，通过内部冷却装置将水温控制在室温左右。在一定压力下通入氯化氢气体，反应会在15-25C时放热。当反应器加热到所需温度，反应一段时间，在反应过程中不断通入氯化氢气体。控制反应过程中氯化氢的加入量，在一些情况下需要用到标准气量控制器。有时候，需要从反应装置的底部取出样品进行分析。达到一定反应时间使反应完全后，停止氯化氢气体的通入，让反应装置冷却至室温，排出气体，记录实验内容与反应过程。
所有其他的化学品都是纯度很高的，所以不需要再净化。
样品以及其标准试样用气相色谱来进行分析。对于没有提供标准试样的样品，需用气相色谱质谱仪鉴定。分析样品及其标准试样所要用到的6890型号气相色谱系统条件如下：
色谱柱： DB-5. 30 米60.25 毫米61.0 lm film, S/N:9118515
o
注入状态： 分离状态 注入体积： 0.2微升 检测器： FID 载 气： He
载气压力： 24 psi (磅每平方英寸) 分离速度： 125ml/min(毫升每分) 氢气流速： 30ml/min(毫升每分) 空气流速： 350ml/min(毫升每分) 组分速率： 25ml/min(毫升每分) 喷嘴温度： 200oC 检测器温度： 300oC
温度控制系统： 初始温度： 6min 40oC 渐变率： 10oC/min 最终温度： 0min 180oC 匝道利率: 30oC/min 最终温度： 1min 300oC 响应因子用下面的方程进行计算：
响应因子样品质量/样品峰面积浓度
内标物质量/内标物峰面积fiAAsWi样品质量/样品峰面积响应因子纯度kA
内标物质量/内标物峰面积fsAiWs质量浓度响应因子峰面积内标物质量1100%
内标物峰面积试样质量主要成分的保留时间（每分钟）为：1,3-DCH（13.7）；TCP（14.2）；2,3-DCH（14.6）；1-MCH（一氯丙醇）（14.8）；2-MCH（15.5）；甘油（15.6-16.2）
3．结果与讨论
在略高于常压（20psi）和120摄氏度的条件下，将含有2%羧酸催化剂的甘油与氯化氢在一个密闭容器中进行转化反应，首先将会生成大量1-MCH(1-MCH, 1-chloropropane-2,3-diol)，以及少量的2-MCH(2-MCH, 2-chloropropane-1,3-diol)。在一个缓慢低效的反应中，大部分的1-MCH转换成1,3-DCH，少量的转换成2,3-DCH。该反应生成的产物如图3所示，由于在常压下，甘油氯化工艺中DCH的转化率较低，这就促使人们去寻求其它方法来改善该工艺。相关文献中提到可以将氯化氢气体喷射到反应溶液中，并将生成的水从溶液中移除，可以利用该过程中目的产物和水形成共沸物促使水的移除，然后经过一个多元反应的步骤移除产生的水。将氯化氢气体通入溶液中反应，或者从形成的共沸物中除去水所需的费用相当高，因此不是商品化生产所期望的。在第一种情况下，需要用到大量的氯化氢，而且从生成的废水中还得回收氯化氢分离出来后再次用于合成，或处理之后用于其他低价值的应用。通过利用形成共沸物除去水分将会导致过程更为复杂，这样需要将溶剂与水分离，还得回收氯化氢用于再利用，这个多反应过程将会增加设备的成本费用以及过程的复杂性。
图3.在120摄氏度的条件下，一定的反应时间内，含2wt%的醋酸的甘油溶液吸收氯化氢气体的主反应的产物表明，在较低压力下,反应会在4小时之后基本上达到平衡，留下大量未转化的甘油和1-MCH。只有当氯化氢气体的压力增加到50psi时，才会产生大量的1，3-DCH。
图4.氯化氢气体被85wt%的甘油，9wt%的水，以及7wt%的醋酸的混合物在90摄氏度的条件下吸收，在不同的压力下的吸收情况表明，压力越高，吸收速率越快，吸收量也越多，由此可以证明甘油氢氯化反应受到平衡的限制。
因为之前的文献资料里提到甘油氯化法的反应具有平衡上的局限性，因此我们致力于研究高浓度氯化氢在反应中的的转换率，反应速率以及反应选择性。要在实验中做到这一点就得利用压力更高的氯化氢气体。我们期望这将会加快氢氯化反应速率，以及提高生成所需氯化产物的反应转化率，同时希望减少氯化氢的用量。图4显示了氯化氢的压力对含有2%醋酸催化剂的甘油吸收氯化氢速率的影响。
图4有两个明显的特点，在快速反应的情况下，最初的反应混合物中氯化氢的消耗，也就是我们加入到甘油中被吸收的氯化氢，它的消耗速率随着氯化氢气
体的压力的增加而增加。不同的压力下，氯化氢的消耗一开始会很快，但在一定的时间之后会慢下来，消耗大量减少，此后，即使延长反应时间氯化氢也极少量或几乎没有消耗。随着压力的提高总的氯化氢的吸收量也会提高，这种情况符合平衡反应中压力越高转化率越高的原则。
图3显示了在不同的氯化氢压力下这三个反应中的主要物质是如何转化的，这是通过对反应中提取的样品用气象色谱进行分析得到的。在压力P（HCl）=20psi时，最初消耗的甘油形成1-MCH的反应是很迅速的，然后会渐渐慢下来，然而1,3-DCH的生成会受到限制。当压力提高到30和50psi时，可以看到1-MCH明显的转化成1,3-DCH。这种情况可以证实反应达到平衡受到低浓度的限制，我们可以通过提高氯化氢的压力来克服这种限制。
甚至在更高的压力下，反应还能更快，而且形成氯化物的效率也会更高。在110psi和11摄氏度的条件下，含有9%水的湿甘油在5mol%的醋酸的催化下氢氯化反应快速高效，经过在密闭容器中4个小时的反应，通入所需的氯化氢气体，该反应将会生成含93mol%的氯化物以及相应的醋酸酯，其中1,3-DCH 和 2,3-DCH的含量比为46:1，另外生成的一氯化物包含含量比为1:2的1-MCH 和 2-MCH。2-MCH比1-MCH的浓度更高的原因在于，它相对于1-MCH的形成速率更为缓慢，而且的氢氯化速度也更慢。
图5.羧酸催化甘油氢氯化形成一氯或多氯化物的机理在文献（7）有所提到。最初的酯化过
程会形成一个闭环化合物同时脱去一分子水，最终形成一种阳离子。该阳离子遇到并结合一个氯离子会开环形成1-MCH。保留两个羟基，重复该过程会得到目的产物1，3-DCH。这种机制下也会形成1,2,3多氯产物，因为两个羟基团也会形成相应的阳离子。
相比于现有的如图1所示的丙烯工艺，甘油氯化法制环氧氯丙烷的工艺有一个显著的优势，也就是甘油氢氯化反应得到的氯化物比丙烯基氯化的选择性更高。这种优势体现在，甘油氯化法中1,3-DCH/2,3-DCH的比达到了30-50，而现有的工艺中只有3:1。1，3-DCH生成的选择性更高的价值在于它环化形成环氧氯丙烷的速率是2，3-DCH的300倍。这样一来就可以使用更小型的工艺设备，或者反应的保留时间也越短。反应保留时间越短会产生更少的副产品，过程越高效，得到的产品也越纯。
高压下的甘油氯化法还有第二个明显的优点。在反应中产生并积累下来的水分会减少副产品的形成，这些副产品中包括一些氯醚或其他氯化物。这样一来就会减少工艺消耗以及废液的处理费用。甘油氯化法工艺中生成的水同时也会提高MCH到DCH的转化，因为在大气压下，氯化氢的浓度越高，会抵消部分反应平衡的限制，使反应向着有利的方向进行。
羧酸催化甘油氯化向商品化生产工艺发展的可行性有一个重要的特点，也就是当反应到达生成氯化物时会有效地停止，所以就不会氢氯化产生很多TCP。如果在化工生产中产生较多的TCP，那么要想有效的合成DCH就会变得难以操作，因为这就需要从产物中回收浓度相对较低的DCH，另外，生成的TCP还会减少甘油的转化效率，更重要的损失是还得去额外处理这些TCP。
之所以在氢氯化反应过程中不会产生TCP是一个必然的结果，该机理可以通过已有的化学实验证实。一种关于羧酸催化多元醇氢氯化的有效的机理早在20年前就被提出，如图5所示。在这个反应机制中，一个羟基团结合羧酸催化剂形成相应的酯，相邻的羟基之间脱去一分子水，最后形成一个阳离子。遇到一个氯离子后该阳离子会开环，选择性的生成我们需要的主要产物，这个阳离子形成过程中的中间产物需要用到两个羟基，这就解释了为什么甘油氢氯化过程会在形成二氯丙醇后停止，进一步的确认该反应机理，在氯化氢存在的情况下，不论是乙二醇或丙二醇只会单氯化形成氯醇。这是一个很重要的机理，因为它可以用来合
成相应的环氧化合物。这就可以用来氢氯化那些多元醇，例如，尤其可以应用于将一些多羟基化合物转化成那些高价值的环氧化合物产品。
图6.在100摄氏度和110psi的条件下，部分氯化氢气体会被甘油中3mol%羧酸吸收，由于羧酸的阻碍，HCl的吸收会减少。
从已有的工艺可以知道有很多种羧酸可以有效地作为甘油氯化的催化剂，但是绝大多数研究中都用到醋酸。因为在实验室规模的合成中醋酸的催化效果很好，而且也能用于合适的工业生产工艺中，不过它不是很稳定，难以用设备进行回收。在甘油氯化生成的主要产物中目的产物1,3-DCH和2,3-DCH也最不稳定。这就需要一个回收设备，从产物中蒸馏分离出DCH以及一些没有参与反应的甘油和MCH以及催化剂，然后重新用于反应中，这就需要我们再去研究相关回收工艺。基于这个选择的可行性，醋酸是首选的的催化剂，而且它的酯化物或中间产物，相对DCH来说比较稳定，所以蒸馏时它们会停留在蒸馏塔底部，回收起来很方便。羧酸催化剂包含六个或以上的碳原子会满足需求。我们已经从一些羧酸中通过实验了解到能让它在氯化反应中起到催化作用的结构特点。
图6显示了在四种不同的羧酸催化下甘油氯化反应中氯化氢的吸收速率，同时与加入醋酸作催化剂以及不加催化剂做比照。在这五个反应中，四个小时过后，
醋酸催化的反应速度是最快的，而且反应消耗的氯化氢的总量也是最多的。正己酸最初反应很慢，但是在四个小时过后消耗的氯化氢与醋酸的同样多。用气象色谱对反应产物进行分析，我们能够知道这两种催化剂催化得到二氯丙醇的效率是一样的。只是用正己酸催化吸收HCl需要一定的反应时间，这就限制了它不能一下子处理大量的甘油去完成转化。3-甲基戊酸催化吸收HCl的速度相对较慢，而3,3-二甲基丁酸的催化效果与不加入催化剂的效果一样。这些反应速率的不同与羧酸的不同相关，而且少量的酯还会阻碍反应的进行。另外，酸的不同的分子结构也与此相关。我们用醋酸氯聚合物作催化剂（pka=1.78）来验证其中的相关性。当使用这种原料作催化剂时HCl的吸收非常慢，差不多与3,3甲基丁酸相同，然而它的酸性是目前为此酸性最强的羧酸催化剂，如图6所示。这样我们可以提出决定这些催化剂的催化甘油氯化速率的是它们的分子大小的观点。
4.总结
对甘油氯化法制环氧氯丙烷的研究证实了，我们可以使用大量的可再生原料来生产工业化学品。由于生物柴油产业的快速增长，多余可用的价格便宜的甘油的日益增多，使得把它当做原材料来应用成为可能。直到最近的2006年，甘油还是通过环氧氯丙烷来制得，但是现在这种方法已经停止了。我们需要进一步发展利用可替代，低成本，可再生的原料的技术，例如，糖和纤维素，也许可以作为原材料来取代现有的工业中的石油化合物。
相对于现有的工艺，甘油法制环氧氯丙烷具有更为经济与环保的特点。这里需要用到氯化氢中的氯来进行转化，而它可以从其它工业生产的副产品中得到，这其中包括现有的环氧氯丙烷合成工艺。另外，甘油法制环氧氯丙烷能够有效利用氯原子，产生更少的废水以及更少含氯有机物的副产物，当然，可再生原料甘油作为原材料在价格上更具有竞争性。
致谢
感谢Erin O'Driscoll 和 Ernesto Occhiello对我们工作的帮助。
因篇幅问题不能全部显示，请点此查看更多更全内容