Jour of Molecule ataat A Chemie 295 (2008) 52-55 Contents is ELSEVIER lable at ScienceDirect Journal of Molecular Catalysis A: Chemical journal homepage: www.elsevier.com/locate/molcata Efficient regioselective carboxylation of phenol to salicylic acid with supercritical CO) in the presence of aluminium bromide Takayuki lijima’, Tatsuaki Yamaguchi* Deporte of Ue and nvrnmenta Scenes, acy of rgnern. Chit Institue of Tea, 2-7-1 adam Narn, Chiba 275-20 jopon ARTICLE INFO ABSTRACT (atic story [ceived in vied for 20 July 2008, ‘Aecepend 2 ly 2008, ‘rae one 7 August 2008 Sale ac ie obtained in good to excellent yield by avbowylation af phenol with 3 carbon dioxide in the presence of Lewis acd catalysts at moderate temperatures (293-353K)- Among the catalyst invest- {ated aluminium bromice (Ar) was found tobe the most active. To optimize the reaction conditions {nd study the reaction mechanism, the experimental variables including the various Lewis acids, temper ature, carbon diosde pressure, amount of ABr,chiorotrimethyslane adaitive. etc were studied Under se {the optimized conditions, nearly complete carboxylation of phenol to salicylic acid was abserved with pt nd {Ormolaf yard oan aca On hts eeimetaldings teat ese acto eee mechanism proposed ins the famine an oanoauminune iz compound tock town Cope © 008 Published by Ese Al its eserve meer re 1. Introduction Aromatic hydroxyearbonfc acids have long been synthesized Carbon dioxide as 2 greenhouse gas contributes significantly to global warming [1]; the amount of greenhouse gases is con- tinuously increasing with the growth of industrial processes and energy-related activities. The Kyoto Protocol adopted in Decem- ber 1997 intends to restrict the emission of greenhouse gases e “The chemical fixation of carbon dioxide is a viable approach to mitigate the excessive carbon dioxide preferably by transformation tovaluable products[3-5]. Although several methodsare known for the chemical conversion of carbon dioxide, they are generally very energy demanding, limiting the practicality oftheir use. Therefore. the development of green processes based on the chemical fxation (of CO; has attracted considerable attention in the field of indus trial chemistry from the viewpoint of environmental protection (6 One of the most versatile processes for carbon-carbon bond for- ‘mation is the synthesis of aromatic hydroxycarboxylic acids such as salicylic acd, p-hydroxybenzoic acid, p-aminosalicylic acid, and 3 and 6-hydroxy-2-naphthoic acid via the coupling of COz with ‘phenols. These aromatic hydroxycarboxylic acids have been widely tused for various purposes, for example, as valuable raw materials and intermediates in the production of pharmaceutical and fine chemicals comespanding authors Tel: #81 47 4780420; fx: 18147 478 0120, Ema address abaya anmoenatmaliajp (Ra) via the Kolbe-Schmitt reaction (Scheme 1) |7-10}, an industrial synthetic process that utilizes CO; asa reactant. This reaction con- sists oftwo steps:(1 the production and purification of alkali metal Dhenoxides and (2) carboxylation, tn step 1, the alkali metal phe- oxides are generally prepared from alkali metal hydroxides and phenol, A mixture of alkali metal hydroxides, phenol, and deion- ized water is evaporated at 353K, and the precipitate is further dried in vacuo at 453K. However, the major disadvantages of this, process are its high-energy consumption and production costs that Stem from the requirement for the purification ofthe alkali metal phenoxide. In other words, this two-step manipulation isone ofthe ‘main drawbacks of this process. “The ather synthetic method (Scheme 2) [11.12] is via the attack ofatrichloromethyi cation (generated by acopper catalyst from car- bon tetrachloride) ofa phenoxide anion, followed by the hydrolyses of the C-Cl bonds in an intermediate ternary molecular complex with concentrated sodium hydroxide because itis fairly difficult to replace an aromatic hydrogen witha carboxyl function. "To eliminate the production and purification step of alkali metal pphenoxides in the two-step manufacturing process, the most dsir- able route for the synthesis of aromatic hydroxycarboxylic acids ‘would be the direct reaction of phenols and carbon dioxide: this could be regarded as the thied access to aromatichydroxycarbaxylic acid from CO. In our previous study, several types of basic metal ‘oxides, such as y-aluminas, zirconia, ceria, and alkali metal and alkaline earth metal salts, have been used in this process [13.4], ‘Among the base catalysts examined, potassium carbonate was the 1361-110 - see ont matte. Crown Copyright ©2008 Published by Ever Al sighs reserved 110%) most 2008 02017"ima Yamauchi oura ef Molecular Caras A: Chic 295 (2008) 52-86 2 1 oH oom scheme, ;0ONa ‘most effective, nd its carboxylation was 99% selective for salicylic acid; however, the reaction took ver 10h to reach completion, ‘Accordingly, the keys to performing a direct carboxylation reaction of phenol are the selection of active catalysts, and the tuning of, reaction parameters to proceed the reaction. Despite al attempts, no satisfactory conditions were found for the direct carboxylation ‘phenol with supercritical carbon dioxide (scCO, to hydroxyben- zoic acids, and this remained a long-standing challenge. ‘The purpose ofthis articles todemonstrate that the direct syn thesis of salieyic acid can be performed in scCO2 with the aid of 4 Lewis acid catalyst. The effects of catalyst usage and reaction, ‘conditions, such as reaction time, reaction temperature and COz pressure, on the reaction were then investigated in detail. Based on the experimental observations, a possible reaction mechanism is proposed. 2. Experimental 24. Materials Anhydrous AlBry, ACs, FeCs, SnCly, ZnCl, ZnBr2. CuCl, and HyS0, (Wako Pure Chemical, 99.5%) as well as phenol (Wako Pure Chemical, 99.5%) were used as supplied. CO, witha purity of 9.9% ‘was used without further purification, 22. General procedures and analyses CAUTION: Experiments using compressed gases such a5 scCOs are potentially hazardous and must only be conducted by using appropriate equipment and taking safety precautions. The direct catalyzed reaction was performed with a scCOs reac- tor system comprising an SCF-Get, SCF-Bps, and SCF-Sro JASCO). All equipments were assembled as outlined in Scheme 3. Phenol (10.0mmol) and AiBrs (2.0mmol) were placed in a 50-m stain less steel autoclave (SUS-316) with a stirring bar. Liquid CO, cooled to.263K was then introduced into the autoclave, which had been hheated to the requisite temperature and flushed with nitrogen, ‘using an HPLC pump. The resultant mixture was stired for 5h at the requisite temperature with the CO, pressure constantly main- {alned atthe requisite pressure. The autoclave was cooled in an ice bath, and the pressurized CO, was gradually released. The resulting, solid mixture was washed from the autoclave with deionized water, andan inner standard was added into the solution. The solution was + NaCl + 3140 3 oo} © Gases 4 ‘Scheme 3A schematic iustatin of experinetal apparatus othe sytess of Fyronybenzoic ac omer in spent Cy (1) Cs reser: (2) mp ‘anolled metering synge puns (3)ba- hair (4) themastat (5) rene ‘i (Oreator(7 stiles stot sringbar (8 er 9) back presse regu “thermocouple pressure wansducet analyzed by HPLC with column (ODS column 250 mm + 4.6mm0) ‘on a Shimadzu LC-10AD chromatograph. 23. Solubility measurement A saturation column was first packed with a solid compound of interest and placed in a constant-temperature bath. Liquid CO> at the ambient temperature was charged itoan HPLC pump and com- pressed tothe desired pressure. The pump was used tocontinuously deliver the solvent at flow rates ranging from 0.1 to 10 standard ‘em3/min and at a constant pressure of up to 30MDPa. After reach- ing thermal equilibrium with the constant-temperature bath, the solvent was fed into stainless steel high-pressure equilibrium cell ‘The entrainment of the solute was prevented by inserting glass ‘wool plugs at the top of the packed column, Once the saturated solution exited the equilibrium cel, it was flashed to atmospheric pressure across a heated metering valve, andthe heavy component ‘was collected in a cold trap held at ice temperature. The amount collected was determined by weighing and by HPLC with an ODS column, and the corresponding volume of CO; was measured with ‘8 wet-test meter. The equilibrium pressure was measured at the exit of the equilibrium cell Fluctuations in the pressure due to ‘the HPLC pump were less than £25 si over the entire pressure range. The temperature was measured within 1 K witha calibrated copper-constantan thermocouple referenced to an ice bath, 3, Results and discussion BU, Selection of catalysts First, the direct synthesis of salicylic acd from phenol and scCOz ‘was performed at 353 K and 8MPa inthe presence of several types of Lewis acids such as AlBrs, AICI, FeCl, SnClz, ZnCl, ZnBry, and CuCl, as catalysts. Some typical experimental results from catalyst screening are tabulated in Table 1. Under the reaction conditions used here, no p-hydroxybenzoic acid could be detected, and sal- loylic acid was the only Islable product in all cases. Without a catalyst, the coupling of scCO, with phenol afforded no product: while in the presence of catalytic amounts of Lewis acid halides, suchas entries 4—10inTable ,acarboxylated product was obtained in good yield, and their carboxylation was 100% selective for sal cylic acid Although not shown in Table 1, the reaction was also performed with protic acids such as H,S0 and CH)COOH. How- ever, these compounds showed no activity forthe reaction. The acidic metal oxides Al,O2, SiO, and Zr0-SO,?- (Table 1, entries,Pa “ina. Yameguh Jura of Molecular Cota A Chemis 295 (2008) 5256 (Gata actives of various Leis and pete ai catalyst for he det carbon Jaton fphesel sai ac supereical carbon doe my caine omvenion ——_altAyield pA ios (va) tot trot a a ° ° er) ° 5 Guth 3 ° 5 Sach 1s ° 2 Za 9 29 ° 5 AS 27 27 ° oA so 389 ® Reaction conditions are as follows: phenol 1 mmol catalyst usage 2.0 mano: vac 1-3), which are weak inorganic acids, were not effective catalysts for the direct synthesis of salicylicacid. As forthe Lewis acd, AIBr, which combines the highest Lewis acidity and good solubility in onpolar solvents [14}, showed the best performance among the Lewis acids examined (Table 1, entry 10). With AIBry as the cat- lyst, salicylic acid yield reached 55.9%. On the other hand, other Lewis acid halides, which were less strong in Lewis acidity than. Alli, demonstrated relatively less activity toward the carboxyla- tion reaction. Tis is similar to the pioneering work of Olah and. ‘coworkers for the carboxylation of benzene with an activated car- bon dioxide-Al,Cls/A system [15], where aluminium halides were able to promote the effective carboxylation reaction; however. the ‘other Lewis acids were inactive, Because AIBI itself also inactive in its hydrated forms (AlBrs H0 to AlBry 6H,0), it became clear that the optimum range of acid catalysts for the reaction is very narrow. After screening many Lewis acid halides, we pursued the carboxylation of phenol by AlBrs in detail 32, Effects of reaction rime on the reaction behavior ‘Te learn more about the carboxylation reaction and to find the ‘optimized conditions, we extended the study to experimental vari- ables such as reaction time, reaction temperature, CO, pressure, and catalyst usage. First, the effect of reaction time was studied Fig. 1 shows the changes in the yield of salicylic acid with reaction, Semen --m-- YYla (mot) re ee Reaction tie (min) Fig 1. Time dependence of product yi of Ary ctlyed carbonation of pe nolst 35 ane MPs Reaction andi are as follows: Phenol. 1Omanol: AB {uaniy. 2mm reaction tempeatre, 353K; COy presse, SNPs (phen (Shsshtieada. 0900 00 = « as mee, Yee (mo!) 0 oa 8 CO, pressure (MPs) Fig 2 stating heeft of CD, pressure onthe carbonation of penal over ‘Ny 3¢ 353K by varying he COs pressure fom 7 to 17MPa Reaction conaitians {re afl: phenol, Oma Ale, quae, 20 mal eatin temperate, 555K; reacton meh () pheno: (©) sale. time over AIBr; catalyst. The reactions were performed in the pres- enceofAlBry (2.0 mmol) at353 KunderaCO pressure of8 MPa. and ‘were monitored in the range of 20-240 minat regular intervals The results indicated thatthe carboxylation reaction proceeded rapidly within the fist 20 min. It is amazing to note that within 60 min, the salicylicacid yield reached a maximum of 55.9 mob, which is com parable to the result for the reaction time of 240 min, suggesting a deactivation of the carboxylation activity. Notably, the selectivity, remained 100% forthe entire course of the reaction. Tis suggests that a reaction time of his appropriate for the direct synthesis of salicylic acid from phenol and scCO2. 433, Beets of (Oz pressure on the reaction behavior Generally, the solubilizing power ofa substance increases with an inerease in its density 16), Therefore, ifthe poor solubility of ‘phenol in scCO2 retards the reaction, an increase in the CO> density should enhance the reaction rate. ‘We systematically investigated a number of reaction pressures ranging from 7 to 17MPa, while the reaction temperature was ‘maintained constant 353K in order to increase the COz density “The effect of the CO» pressure on the yields of salicylic aid for the ‘Airy catalystisillusrated in Fig, 2. thas been reported that, when (Cop is used as a solvent or a reactant, the reaction rates are maxi- ‘mal near the critical pressure of CO [8.17.8]. Inthe other words, ‘excessively high COp pressures may retard the interaction between 8 substrate and a catalyst [19]. As is obvious from Fig. 2, however, increasing the CO, density through an increase in the CO, pressure sradually increases the activities ofthe AlBr, catalyst, demonstrat- {ng the preferential effect of supercritical conditions for promoting, the reactivity of CO». The increase in the CO» pressure and density, could inerease the solubility of phenol, and may cause a high con- centration of phenol in the vicinity ofthe catalyst thus resulting in high yield ‘The solubilities of phenol and salicylic acid in scCO, have been ‘measured over a pressure range of 8-30MPa at 323 K to assist in the study ofthe carboxylation reaction. The experimental solubility, data for phenol and salicylic acid in seCO as a function of pres- sure at 323K are ploted in Fig 3. The solubility isotherms in Fig. 5 for phenol and salicylic aid continue to increase at elevated pres- sures, However, the solubility of phenol in COs is two orders of magnitude higher than the binary system solubility of salicylic acid1 ima 7 Yagi Jura of leur Ces A: Chemicl 295 (208) 52-585 8 10 0 goo 08 0? of a pot oo 4 r an oe = a wo 4 08 a 4 108 ee rr CO, pressure (MPa) i. 5. Solubility others of phenol (2) and site ald (2) a a unetion of, pres. ‘over the same range of pressures. This is consistent with the lites ‘atures reports of these compounds [20,21]. The key advantage of this method is that the high solubility of phenol in seCO> enhances the concentration of phenol inthe vicinity ofthe catalyst. 34, Effects of reaction temperature onthe reaction behavior We performed similar experiments in the range of 273-373 Kin ‘order to investigate the catalytic activity at different temperatures. Fig. 4 shows the effect of raising the reaction apparatus tempera~ ture whilst the CO» pressure was maintained constant at 8MPa. A relatively high temperature was found to be favorable for an effi- ‘dent reaction, Unlike the KCO}-catalyzed carboxylation of phenol, to hydroxybenzoic acids, where the reaction occurs at tempera~ tures approaching 473K. the salicylic acid yield increased steadily from 0 to approximately 70% with a reaction temperature up to 373K, However, a further increase inthe reaction temperature to 393K led to a considerably slower reaction, possibly due to the ‘decomposition of salicylic acid at the higher temperature. Olah et performed the carboxylation of benzene with an AIC catalyst [1 ‘They showed that no benzoic acid could be observed above 383K. ‘ . « . oa . os Zz 0 i ° 7 . ie ° a . . = ° ° oa Tonpeatre(t Fe 4. plot of yt fsa ai rm the canton of phenol t MPa Teens he temperature fom 299 t 39K over Al, eta Renton cond tons aes fe phepa, Ome: Ar quan. 20 mmo reaction time, 1: Or presi, MPa (a) penal (0) sage ae © Yield (mot) ‘ABA quantity (mmo!) Fie. Reaction ate—Albry quay profes forthe dec carboylton of pheno InscCo, Reaction condtensareas follows: phenol 0 mmol reac temperature, 539K: reseton ime, CO, presse, BMPa (wena (3) sabeye 2 Benzoic acid yield became significant at 363K and was formed exclusively at 363 K: Dunn etal. reported that the decarboxylation of, salicylic acid to phenol was proceeded in the temperature range of '363-503 [22]. Therefore, 373K could be the optimal temperature for tis reaction 35, Effects of AlBr; quantity on the reaction behavior Suzuki etal. found that a certain amount of carboxylated prod ‘ucts obtained could be dependent on the catalyst usage |. Thus, the influence of the amount of AlBrs catalyst on their carboxyla~ tion reaction was studied over the previously optimized reaction temperature and CO2 pressure. The effects of the AlBrs quantity added to the autoclave for the formation of salicylic acid are pre~ sentedin Fig. 5.Nearly complete carboxylation of phenol toslicylic acid was observed with 1 mmol of AlBrs added to the autoclave. ‘Although the salicylic acid formation significantly increased with an increase inthe AlBrs quantity, the amount of salicylic acid pro- {duced did not linearly increase with the amount of AIBrs. Another ‘observation was that the yields calculated on the basis of the reactantAlBr~ 1:1 ratio exceed the theoretical 100 mol%. There fore, these observations strongly suggest that AlBr; is considered tobe a catalyst inthis system rather than a reagent inthis system. 436. Possible reaction mechanism forthe salicylic acid synthesis Jrom phenol and CO, with AlBr3 [Nemoto et al. reported thatthe Lewis acid-mediated carboxy! tion of aromatic compounds with CO is promoted by the addition of large excess ofthe chlorotrimethylsilane (TMSCI) to yield aryl- carboxylic acids [23]. Under the optimized reaction conditions, we conducted the carboxylation of phenol over the 2mmol of AiBr3 with TMSCI additives, in order to probe the ability of TMSC! addi tive for this reaction and to obtain further insight into the nature of the reaction, The obtained results are collectively presented in Table 2. The efforts to increase the yields toward salicylic acid for- ‘mation using a large excess of TMSCI additives, however, failed, These results may be explained by a steric hindrance-the interac- tion of phenol and AlBr, catalyst sinhibited by a trimethyisilylated hydroxy! group. Trimethyislylation was a frequently used pro- tection method in several chemical conversions and synthesis of, natural produets due to the enhanced chemical and thermal stabil-6 "ima 7. Yamaguchi Jura of Maer oat A Chemin 295 (2008) 52-55 "ect of TMSCLadecves onthe ye of carbzayltion of phen with COs in the presence of AB tas ‘ey TASCimmey—Gamvenion—o¥Ayleld—_—prBAYioa (ts) re (rote) 3 2 1 16 ° 4 > 18 18 6 & © a3 35 6 ity. Because phenols are reaily transformed into the corresponding silyl ethers, which are prepared by the dehydrogenative silylation, of hydroxy! groups, in excellent yields|24], these observations were expected. However, the yield could be improved by increasing the amount of TMSCI Table 2, entries 2-6). Thisis due togenerate ary- silanes, which is in-situ silylation of the resultant phenol at the ortho position, under the Lewis acidic conditions, followed by the carboxylation at the ipso carbon of the silyl moiety. The addition, ofa large excess of the silylating reagent shifts the equilibrium to, the right to promote the second reaction, giving the arylsilanes in, higher viel, ‘Although the complicated characterizations are necessary to raw a detailed reaction mechanism for the salicylic acid syn- thesis under the current conditions, a rough reaction pathway is, suggested in Scheme 4, The reaction as exemplified with phe- ‘ol is initlated by the acid system providing the salt 1, w! is depicted as the phenoxyaluminium dibromide. Generally the presence of a phenolic group kinetically favors ortho-alkylation, And ortho-acylation. For example, Naeimi etal. performed several ‘exploratory reactions that only produced ortho-acylated products, [25]. The intermediate 1 pictured in Scheme 4 is similar to that proposed by Casiragiet al. [26] and accommodates the regiose- lective ortho-carboxylation of phenol. in their original paper [25), Casiragi and coworkers already proposed the possibilty of the for- ‘mation of phenoxymagnesium chloride. The aluminium salt is a requirement for further reaction; no reaction occurred when asolu~ nw Co on oH 9 oo evra Caan es Chrous Schemed, Pssineeaion mectansm orsalclc acd synthesis fom phenol ant CO, with Lewis ac eta tion of phenol was heated under a CO pressure of MPa. The intermediate 1 reacted with CO2 through the cyclohexadienone structure 2 furnishing the alminium salt of salicylate 3. Conse~ ‘quently, on the basis ofthe presently available data, the formation ‘of an organoaluminium-like compound appearsto be the preferred pathway. 4. Conclusions Inthe present study, the direct carboxylation of phenol with car- bon dioxide proceeded smoothly in the presence of various Lewis acids, especially ABrs, at moderate temperatures under supercrti- ‘al conditions. This method isa new, mild efficient. easy, and clean ‘method for the preparation of salicylic acid in excellent yields with high regioselectivity in the substitution of a carboxyl group in an ortho situation. Yo evaluate the mechanistic aspects ofthe reaction and to find the optimized conditions, the effect of several experi- ‘mental variables was also studied, including.a reaction using large ‘excess of TMSCI additives. These reactions have some advantages in comparison with other methods, such as: short reaction time, high yield and regioselectvity ofthe product, and mild reaction condi- tuons. In accordance with the experimental results, the formation of the phenoxyaluminium dibromide and its reaction with carbon dioxide is proposed as the most viable pathway for the studied carboxylations. References (3). kes, A. istzon, Nature 402001) 755. [2] Kom esol tote Une Nation Ramework Convention on Climate ‘Change. 1957 [3] Bj core). eget Owens 10. Glennon. AE Graham, 10. Holmes, Green hem (2005 15, (4) tye al Sng, M. Polo Green Chem. 7(2005)357 {5] jap ar, Neyor ene 289 1095) 1s. [6] Hr Aseawa, Me Avesta JN Armor MUA Bare, Eckman, AT. Bel JE Bescam, ©, Creu EDinus,Dat bao K.Demen De bubs. eke E Fit DF ctson WA crdaré DW. Goodman). Kalle Kab ne JE hanger Marks X Morakama, KM Nichols, etaa, Qbe ona Nielson WAH Sache LD. ehmdt A Sen,GA Samora PC Sti Se stu, W. Tuma, Chem He 101 (200) 953, (7) tims BTstagh 1 Yarsagaeh Jpn feta ns 51 (200865. {a] isima TamapurhTecahedsoa et 48 (2007) 530. | na Laase Yamaguchi Jpn. Pere nt 9 (2006) 206, (00) As tindsey H pesky Chem, R57 (1957) 383, [it] aston Rasnsy,|chem Soc chem commen. (198) 1124 {12 Nexomiyama.t Halt Am Chem Se 06 (1988) 74 [03] ijt tamagoen, Appl cata Ac cen 2452008) 12. thal sue Haters 1 Otten, © Mano, Chem Lee 3 (2002)102, 105] GA Gah 8 Tero ji osehek. cs ML Eteves asl, GS Paka ‘in Cem. soe 124200 11579. 96] 01 Kaimoes them ev 95 (1585) 355 {ir Sra Bo, Rbaage shims M Aral Creen Chem, 3(2001) 87 {is stahura Sato, Okano, hot Sak, Org Cher 63 (1888) 7055, [Ha] ¥. bus ca Da hong LN. He Gree hem 7 (2005) 18. [Jone ki dru Nad Chem Eng ste (200) (GS cua vs oer Ew he 201990175. [22] GE Dunn EC Jansen WRodenal, ca) Chem 6988) 2005 [23] Semoun Sur Morohat Hatori hers Let 352006), [24] D zareen Karim Tetrhedio et 48 (2007) 127. [25] Nas rad} MoL Catt A 262006) 282. [a5] casirgcCaaatMCormia A Pot. Pai, Serato, Ungaro Goom Soe erin rans 1 (1978) 318