In both routes of fermentations and chemical methods described above, alkali sources are added to neutralize lactic acid, which results in the production of lactate salts. Lactate salts were transformed into lactic acid by using large amounts of mineral acids. In order to overcome this drawback, lactate esters were produced directly in sugar conversions by using Lewis acid catalysts.
4.5.2.1 Lactate Esters from Trioses FAU zeolite catalysts
Seles et al. reported the conversion of trioses into alkyl lactates (entry 3 in Table4.4) [50]. For various Y type zeolite catalysts with FAU type structure, the catalysts eventually were grouped based on their catalytic selectivity to form either dialkyl acetal or alkyl lactate (Fig.4.10). The former product would be formed by
Table4.4Thecatalyticconversionoftrioses:dihydroxyacetone(DHA)inalcohols EntryCatalystc Substrate conc.[M]bAlcoholCatalystconc. [mgmL-1 ]T[C]T[h]Yield[%] di-acetalYield[%]alkyl lactateRef. 1SnCl45H2Oa0.625MeOH10mol%a90298251 2SnCl45H2Oa 0.625EtOH10mol%a 901–8451 3YtypezeoliteZF2100.4EtOH40906286550 4H-USY-zeolite(6)0.25MeOH1611524–9652 5H-USY-zeolite(6)0.25MeOH16115157552 6Sn-bzeolite(125)0.25MeOH16802409954 7AmorphousSiAl(10)0.4EtOH4090612256 8Sn-MCM-41(49)0.4EtOH4090629857 9Sn-montmorillonite0.25EtOH101501509758 10Sn-montmorillonite0.25EtOH101201225358 11Sn-graft-MCM-41(460)0.2EtOH2090632359 12Sn-CSM-(15%C)d,e0.2EtOH20906010059 19Sn-CSM-(16%C)d 0.2C8OH20906–8359 20Sn-CSM-(16%C)d 0.2C14OH20906–5459 ReproducedfromRef.[55]bypermissionofTheRoyalSocietyofChemistry aHomogeneouscatalyst:inthatcasethecatalystconcentrationisinmol%tothesubstrate.bMonomericform.cValueinbrackets=bulkSi/Metalratio. d SitoSnratiooftheSn-MCMpart(=85wt%ofthecomposite)is460.e Post-oxidisedat300
strong Brứnsted acidity, while the latter would result from Lewis acidity. As consequently, ethyl lactate was thus produced with a yield of 65 % in ethanol with a Y type zeolite catalyst in DHA conversion [50], which was comparable with those over SnCl4homogeneous catalyst (entry 2 in Table4.4) [51].
Taarning et al. also performed a screening with several zeolite catalysts, such as USY zeolites, H-ZSM-5, H-BEA, H-MOR, a sulfated zirconia catalyst, and H-montmorillonite catalyst, both in water and methanol [52]. The H-USY with a bulk Si/Al ratio equal to 6 (H-USY6) outperformed all other catalysts and it showed markedly better catalytic performance than the other H-USY with a Si/Al ratio equal to 30 (H-USY30). A methyl lactate yield of 96 % at 99 % conversion was obtained at 115C after 24 h of reaction over H-USY6 catalyst (entry 4, Table4.4). The yield of 75 % was already obtained after 1 h (entry 5). The ratios of Brứnsted to Lewis acid sites were measured to be 1.8 for H-USY6 and 5.6 for H-USY30 catalysts, respectively, by an FTIR using pyridine and an NH3-TPD. On the other hand, addition of water generally lowered reaction rates, yield and lactate selectivity in the reaction media. In addition, the stability of the zeolite and its deactivation process were monitored in a promising continuous flow mode setup. It was found that lactic acid in water destroyed the catalyst structure even at a concentration as low as 0.3 M. In contrast, this was not observed for methyl lactate in methanol. Another cause for zeolite deactivation in water is catalyst coking, mainly attributed to an intermediate of pyruvic aldehyde. A kinetic study with H-USY6 catalysts in water showed an apparent energy barrier of 53 and Fig. 4.10 Mechanism for converting trioses into alkyl lactates in alcohols (R=alkyl) or into lactic acid in water (R=H). Reproduced from Ref. [55] by permission of The Royal Society of Chemistry
61 kJ mol-1 for dihydroxy acetone dehydration and pyruvic aldehyde to lactic acid reaction, respectively. These values are clearly lower than those obtained with soluble Al3+salts [53].
Sn-b zeolite catalysts
Taarning et al. prepared Al-free Sn-b zeolite catalysts according to a previous report. They proved that in particular Lewis acidicbzeolites were very active for the DHA conversion [54]. A comparison of a series of zeolites with Ti, Zr, and Sn incorporated in the framework revealed a correlation of the catalytic activity with the Lewis acid strength,bzeolites with tin showed the strongest in the series (entry 6, Table4.4). The Sn-b zeolites showed the catalytic selectivity to alkyl lactate, whereas, conventional Al-bzeolite and Brứnsted acidic ion-exchange resin cata- lysts yielded only dialkylacetals, which suggested that strong Brứnsted acidic zeolites are selective toward acetals and Lewis acidic zeolites are selective toward alkyl lactates. Interestingly, steaming a parent Al-b zeolite to produce extra- framework Al enhanced the lactate yield in regard to the pristine H–Al-b [54].
These results led to the same conclusion as with USY6 versus USY30 study [52].
In addition, the Sn-b zeolite catalyst performed more active per Sn site than the homogeneous halides: the initial turnover rate of methyl lactate formation was 45 mol mol-1Sn h-1, whereas a value of 4.2 mol mol-1Sn h-1was calculated for homogeneous SnIVCl45H2O catalyst.
Other heterogeneous catalysts
Several researchers investigated conversion of torioses to lactic acid under mild conditions with heterogeneous catalysts [55]. Pescarmona et al. screened widely heterogeneous catalysis with a high-throughput equipment [56]. They focused on different aluminosilicates with various types and strengths of both Brứnsted and Lewis acidity, since that appeared critical in the performance of H-USY zeolites.
An amorphous silica-alumina (entry 7, Table4.4) presented very selective results at low temperature, but insufficient activity.
Li et al. reported the triose reaction over substituted MCM-41 materials and Sn- MCM-41 appeared to be very selective to ethyl lactate and high catalytic activity (entries 8, Table4.4) [57]. For Sn-MCM-41 the initial turnover was around 8 mol mol-1Sn h-1. The catalytic advantage of Sn-MCM-41 was ascribed to a combination of Lewis acid and mild Brứnsted acid sites. The authors demonstrated the rate-accelerating effect of Brứnsted acidity on the formation of lactate esters, suggesting dehydration to be rate limiting with Sn-MCM-41, because the presence of Brứnsted acidity in Sn-MCM-41 is somehow correlated to the presence of Sn and Sn-free MCM-41 did not exhibit such high acidity. However, such bifunctional working hypothesis of Sn-MCM-41 was still not clear because of co-existence of both acid sites.
Onaka et al. reported that Sn-exchanged montmorillonite catalyst showed very high ethyl lactate yields for DHA conversion in ethanol at 150C, despite the strong Brứnsted acidity of the catalyst (entry 9, Table4.4). This example nicely illustrates the aforementioned advantageous effect of a high reaction temperature, since the yield is much lower at 120C (entry 10, Table4.4) [58].
The bifunctional catalytic mechanism for alkyl lactate synthesis from trioses has recently been revealed by de Clippel and Dusselier et al. [59]. They used Sn grafted carbon–silica composite catalysts to alter the number of Brứnsted acid sites inde- pendently from that of the Lewis acid sites. Lewis acidity was provided by grafting a mesoporous silica like MCM-41 with isolated SnIV. The oxygen containing functional groups like carboxylic acids and phenols functioned as weak Brứnsted acid sites. The density of the acid sites was controlled by the carbon loading, the degree of oxidation (depending on the pyrolysis temperature) and an optional post-synthesis oxidation.
For DHA conversion to ethyl lactate at 90C in ethanol, the carbon-free pure sili- ceous Sn-grafted MCM-41 catalyst (entry 11 in Table4.4) showed high ethyl lactate selectivity but low catalytic activity on Sn basis (TOF=41 mol mol-1Sn h-1). In contrast, introduction of the weak Brứnsted acidic carbon in the mesopores seriously boosted the catalyst activity without compromising the excellent ethyl lactate selec- tivity (TOF=289 mol mol-1Sn h-1). The best result was obtained with a composite containing 15 % of carbon, which was pyrolysed at 500C and post-treated at 300 C under oxygen flow (entry 12 in Table4.4). Table4.5 shows the initial DHA con- version rate against the total amount of COx released by heating the composite catalyst. The amount of released COx was estimated to be the total weak Brứnsted acidity. Both the amount of released COx and the initial conversion rate correlated linearly, pointing for the accelerating effect of the estimated weak Brứnsted acidity on the Lewis acid catalyzed conversion of DHA. From these results, the authors sug- gested that the bifunctional desige was focused as to maximize alkyl lactate and lactic acid formation rate by balancing the number of weak Brứnsted acid sites for a given content of Sn, whereas the strong Brứnsted acidity triggered unwanted competitive reactions.
Table 4.5 Overview of total released COx content and the initial dihydroxyacetone (DHA) conversion rate over catalysts
Catalyst mmol g-1
Total content COxa mmol g-1h-1
Initial DHA conversion rateb
1 Sn-Si-MCM-41 – 22
2 Sn-Si-CSM-773-11.9 0.25 36
3 Sn-Si-CSM-773-16.4 0.38 54
4 Sn-Si-CSM-773-24.3 0.55 65
5 Sn-Si-CSM-1073-15.7 0.24 39
6 Sn-Si-CSM-773-15.3/O2473 0.76 68
7 Sn-Si-CSM-773-14.9/O2573 1.02 108
Reprinted with the permission from Ref. [59]. Copyright 2012 American Chemical Society
aThe values for ‘‘Total content COx’’ were obtained from the TPD-MS experiment in heating from 303 to 1073 K at 10 K min-1,bthe initial conversion rates were typically calculated from the kinetic profiles, i.e. relationships between conversion and reaction time
4.5.2.2 Lactate Esters from Hexoses
The direct conversion of glucose, fructose, and sucrose to methyl lactate in methanol at 160C in an autoclave reactor was reported by Holm et al. [60].
Brứnsted acidic zeolite H-Al-bcatalyst gave almost no yield of methyl lactate and catalyzes the dehydration of the sugars, leading to HMF derivatives and methyl levulinate from fructose and predominantly methyl- d-pyranoside from glucose and sucrose, which were in agreement with early reports by Rivalier et al. [61]. On the other hand, the Lewis acidic Sn-bzeolite catalysts were found to induce high selectivity toward methyl lactate of 43, 44, and 64 % yields from glucose, fructose, and sucrose, respectively. Nanocrystalline SnO2is inactive, whereas SnCl4shows moderate selectivity toward methyl lactate of 31 % yields from sucrose. The nonacidic Si-bzeolite did not improve the yield of methyl lactate (5 and 6 % yield from glucose and sucrose, respectively), indicating that the catalytic ability is related to the Lewis acidity of Sn ions incorporated into the zeolite structure.
Roman-Leshkov and Davis reported that the Sn-b catalyst is unique among its fellow Zr, Ti, Si, and H–Al-b zeolites for this sugar conversion [62]. The capa- bility of Sn ions to activate carbonyls and to shift a hydride from one carbon to the adjacent one is not only useful in the internal Cannizzaro reaction of the hemi- acetal of pyruvic aldehyde to lactic acid, but also in the ketose-to-aldose isom- erisation of sugars in water.
de Clippel and Dusselier et al., for instance, also tested their bifunctional Sn-based carbon-silica composite materials for the conversion of hexoses [59]. Low fructose and glucose working concentrations are indeed important as to avoid competitive side reactions such as methylation to stable methyl sugars and dehy- dration ultimately to insoluble humins. An excess amount of Brứnsted acids would be not desired. With regard to the reaction mechanism, the one-pot conversion of hexoses into lactate esters may be regarded in part as a synthetic glycolysis including an isomerisation of glucose to fructose and the splitting into trioses, and in part as a synthetic glyoxalase system including formation of lactic acid from pyruvic aldehyde, followed by esterification of lactic acid with methanol or ethanol.