DESCRIPTION CARBOXYLATION OF HYDROCARBONS TO TEREPHTHALIC ACID OR NAPHTHALENE DICARBOXYLIC ACID Technical Field This invention is related to the synthesis of commercially important aromatic acids and diacids, especially those which are comonomers for high performance polyester polymers. Further, this invention relates to a new method of synthesizing these raw materials at a reduced cost in significantly improved purity without the necessity of oxidation reactions or purification of alkyl hydrocarbon isomers, both of which are expensive and difficult. The present invention solves these problems by a new chemical reaction which, in effect, adds the elements of carbon dioxide or possibly carbon monoxide directly to an aromatic ring to form the desired product, for example, terephthalic acid (PTA) or naphthalene dicarboxylic acid (NDA), in high purity.

Background Art An extensive body of literature going back to the 1950s documents the so-called Henkel process for disproportionation/isomerization. In the 1960s and 1970s small scale terephthalic acid plants were operated commercially, finally being displaced by large (up to 1 billion lb/yr) plants based on variations of a process involving separation of p-xylene from mixed xylenes, followed by oxidation of the p-xylene in acidic media with manganese or cobalt catalysts in the liquid phase with oxygen or air, followed by reduction of any residual aldehyde with Pd/C and hydrogen, followed by

crystallization of the product terephthallic acid (PTA) away from the remaining species at high temperature in pressurized water (typically at 200-220°C). While complex, this type of process has benefits of scale. However, the global demand for para-xylene has raised its price far above that of benzene or xylene, and the complexity of the separation and oxidation processes further increase costs.

Similarly, the so-called Henkel process used only aromatic acids, viz. phthalic acid, phthalic anhydride, or benzoic acid as feeds. Since the cheapest of these acids is more than twice as expensive as benzene, the raw material costs of the Henkel process, in conjunction with the lack of scale advantages, contributed to the demise of the original Henkel plants.

A few of the Henkel patents for use of benzoic acid suggest that it might be possible to achieve slightly more than theoretical yield of terephthalic acid, and imply that some C02 is directly incorporated, however the evidence is spotty and the processes claimed to have that effect are the same as processes which are not claimed to have that effect, with minor differences in temperature, etc. which are within the normal range of Henkel processes. None of these patents identifies what we have identified as the essential factor comprising using basic carboxylating salts, and none of them starts from a hydrocarbon. Instead they claim a slightly higher conversion of benzoic acid to terephthalic acid.

It would constitute a substantial advance in the art if there were a method available for synthesis of commercially important aromatic acids and diacids, especially those that are comonomers for high performance polyester polymers.

It would be even more advantageous if the raw material cost was low and the purity was improved without the necessity of oxidation reactions or purification of alkyl hydrocarbon isomers, both of which are expensive and difficult.

The process of the present invention accomplishes these desirable objectives by a new chemical reaction which in effect adds C02 directly to an aromatic ring, to form the desired product, terephthalic acid or naphthalene dicarboxylic acid, in high purity directly.

Disclosure of the Invention In accordance with the foregoing the present invention is a process for direct carboxylation of aromatic hydrocarbons which comprises reacting aromatic hydrocarbons with a carboxyl source selected from basic carbonate containing salts and carbon dioxide in the presence of a metal catalyst at elevated temperatures under conditions providing said carboxyl source sufficient mobility to effect a net reaction to form the aromatic acids and diacid salts of the aromatic hydrocarbon. We have demonstrated that > 99% pure 2,6-naphthalene dicarboxylic acid can be formed by the reaction of potassium oxalate, potassium carbonate, and naphthalene in the presence of a metal oxide or metal salt catalyst at a temperature in the range of 400-500°C. In addition, terephthalic acid of very high purity (PTA) is formed by the similar reaction of phenol with the listed potassium compounds.

Detailed Description of the Invention In the present invention we have discovered methods by which non-aromatic carboxylic acid materials, under conditions of very high temperature and very strong basicity, may be used to add acid groups to a variety of aromatic hydrocarbon nuclei to make aromatic acids and

particularly diacids, such as terephthalic acid (PTA) or 2,6-naphthalene dicarboxylic acid (NDCA).

Although not necessary, an apparent mechanism has been worked out for the reaction, allowing a cyclic process which takes potassium oxalate, composed of CO, CO2, and KOH, plus hydrocarbon to PTA or NDCA, or potassium salicylate or phenol to PTA or NDCA in the presence of CO2, using CO2 as the source of carboxyl groups via intermediate reaction products of phenol or salicylate, giving a net reaction of hydrocarbon plus C02 to diacid.

Diacid synthesis is observed when relatively larger amounts of carboxylating basic salts are used. A dominance of mixed isomers is observed if the reaction is run at relatively lower temperatures with relatively less catalytic material. For purposes of this invention,"lower temperatures"will generally be from about 350°C to about 420°C. Lower amounts of catalytic materials will generally be from zero to 10 percent weight based on the hydrocarbon feed. A single isomer is generally observed if the reaction is run at higher temperatures and with higher loadings of catalytic materials. For the purposes of this invention, "higher"temperatures will be from about 420°C to about 500°C. The higher levels of catalytic material will be those above 10% weight (basis the hydrocarbon feed). The terms for higher and lower temperature and catalytic material are, however, relative and may overlap according to the exact carboxylic basic carboxylating salts and aromatic hydrocarbons used.

The catalysts employed are isomerization catalysts, and as will be shown in the Examples, the carboxylation of hydrocarbons with CO2 will proceed even without them.

The aromatic hydrocarbons which are suitable as starting materials in the present invention include aromatic hydrocarbons containing one or more benzene rings, including, but not limited to benzene, toluene, xylene, and tetralin, and condensed aromatic hydrocarbons such as naphthalene, anthracene, phehanthrene, etc., or a mixture thereof or a fraction containing one or both of them.

Benzene and naphthalene were preferred due to the commercial value of the corresponding products, terephthalic acid and naphthalene dicarboxylic acid. The synthesis of PTA and NDCA is demonstrated in Examples 5-16.

The basic carboxylating salts consist of materials such as potassium oxalate, potassium acetate, potassium formate, potassium malonate, potassium sorbate, potassium citrate, potassium salicylate, potassium phenolate, potassium resorcinolate, potassium naphtholate, potassium cresolate, dipotassium carbonylate, potassium hydride, and the like. In the above list,"potassium"may be taken as <BR> <BR> <BR> <BR> either"monopotassium"or"dipotassium","tripotassium", etc. up to the limit of available hydroxy (or hydroxy and carboxy groups) according to the particular material.

Sodium, cesium, or rubidium can also be used as the counter ion, but potassium is generally preferred due to the best balance of cost and efficacy. The basic carboxylating salts may also be mixed with potassium carbonate, and may be formed by the mixture of the original carboxylating anion feed (e. g., oxalic acid, acetic acid, etc.) with potassium carbonate. If the carboxylating anion feed is a solid, the reaction of the two solids (anion feed and K2CO3) will normally proceed well under the reaction conditions.

Where potassium carbonate is used, we have discovered that the presence of oxalate, acetate, etc., salts, for example the"basic carboxylation salts"listed above, are

necessary in combination with potassium carbonate.

Potassium carbonate alone, as is typically used in the Henkel process for converting aromatic acids to PTA, is not effective, even in the presence of the catalysts.

Especially preferred in one mode of the invention are carboxylating anion sources such as oxalate, formate, and the like salts which may be formed in a"cyclic"process from C02 or CO. For example, oxalate can (at least in principle) lose C02 to make a strong anion (dipotassium formate) which can deprotonate the hydrocarbon feed in the presence of the catalyst and initiate the carboxylation of the feed; the protonated formate material can in principle regenerate oxalate. To the extent that such cycles do not operate in a self renewing condition under the actual reaction conditions, the basic potassium salts remaining can be converted back to oxalate or formate by well known reactions. ( e. g., KOH + CO at 120°C-> potassium formate; potassium formate at > 200°C K2 Oxalate; etc.) In another mode of the invention, it is especially preferred to use a eutectic mixture of alkali carbonate salts which are readily regenerated from the base left over by recovery of the product acids or diacids by reaction with carbon dioxide. In this mode of operation, it is preferred to use carbon dioxide as the acid to recover the product aromatic acids or salts.

The present invention is operable, in fact it may be preferred for some feeds, when all the materials including the anion feed, hydrocarbon feed, catalyst, and K2CO3 are infusible solids, although it will be appreciated that a certain amount of solid-solid molecular change must occur for the reaction to proceed, and that some intermediates may in effect become molten during the reaction process,

even if the starting materials and final products are not molten.

In addition to the use of such salts as carboxlating anion sources, some salts may be thermally stable under the particular conditions used, and may, therefore, if in the molten state, be used as solvents for the reaction, facilitating the interaction of the other components.

Potassium salts of alkyl carboxylic monoacids would be expected to be useful for this purpose, while acids such as oxalate, malonate, formate, etc., which might more easily decompose would be expected to be better carboxylating anion sources. Salicylic acid would be expected to have a labile C02, and also to be readily carboxylated by C02 when in the anionic decarboxylated form, and hence may be an especially useful carboxylating anion source cyclically regenerated from C02 provided it is dissolved or otherwise intimately in contact with the hydrocarbon feed. Otherwise it apparently prefers self carboxylation and loss of water to give TPA itself.

Suitable catalyst materials for the present invention include the salts and oxides of Group IIB, Group IB, Group IIIB and perhaps Group VIIIA metals, including, but not limited to zinc, cadmium, copper, silver, lanthanum, scandium, manganese, and cobalt. Cadmium salts may be used to increase the rate of reaction over the uncatalyzed rate.

Zinc was preferred in the present invention on the basis of innocuousness and effectiveness.

The hydrocarbon material may be added all at once, or in stages. Likewise, the basic carboxylating salts may be added all at once or in stages. The materials are often solids, in which case it is convenient to add mixed powders to the reaction zone. Liquids, and even gases, may be used for the reaction. It will be obvious to those skilled in

the art that in the event of gaseous reagents being used, higher pressures will generally lead to higher rates and higher activities. Normal practice will consist of performing the reaction under a moderate pressure of C02 to increase the rate of carboxylation. A pressure in the range of 100 psig to 1500 psig is suitable. An especially useful range is from about 200 to 1000 psig.

Suitable temperatures are in the range of 300-550°C. A preferred temperature range for producing mixed isomers of diacids is from about 350°C to about 420°C. A preferred range for producing predominantly single isomers of diacids will be from about 420°C to about 500°C.

In the present invention suitable metallurgy in reactors will be required to avoid excessive corrosion under the severe operating conditions. Nickel alloys may be preferred over steel, although in many instances steel will be acceptable. Where liquid solvent or diluent media are employed, those such as KOH which react with C02 to form infusible solids are less desired than those such as carboxylic acid salts, which remain liquid over the range of interest. The presence of large amounts of free liquid KOH also leads to stress corrosion cracking in stainless steel, whereas the carboxylic materials such as potassium carbonate, bicarbonate, acetate, oxalate, etc. are comparatively benign and allow simpler reactor design.

The use of a molten solvent or diluent medium may facilitate transfer of the reacting mass from vessel to vessel and improve mixing, however it is not necessary to the practice of the invention. If a diluent medium is used, it should be liquid, stable at the temperatures employed, and, if inert, a material which does not undesirably affect the reaction.

In some preferred examples a eutectic mixture was employed. A eutectic mixture provides the lowest melting point of a mixture of two or more alkali metals that is obtainable by varying the percentage of the components.

Eutectic mixtures have a definite minimum melting point compared with other combinations of the same metals. For example, though the melting point of Li2CO3 is 622°C, in a eutectic mixture of alkali carbonates the melting point can be 400°C. What is required in the present examples, where a eutectic mixture is employed, is the right mixture of alkali metal carbonates where the melting point is less than about 400°C. Generally the ratio of alkali metal carbonates in the eutectic mixture is about 1: 1: 1, but it can vary. One eutectic mixture used as a solvent was K2CO3, Rb2CO3, Cs2CO3, and optionally Na2CO3.

In a preferred embodiment, a eutectic mixture of K, Rb, and Cs carbonates is used as the sole basic carboxylating salt as well as the medium for the reaction.

It is unexpected that it should be possible to add C02 directly to a hydrocarbon to produce an aromatic acid or diacid. Thermodynamic calculations indicate that for the free acids, the hydrocarbon + C02 is strongly favored (by about 13 kcal/mole) at Henkel reaction temperatures (400- 500°C). Thus under those conditions, the aromatic acids would be expected to decompose by decarboxylation. This is readily seen by pyrolysis of the free acid at 450°C, even under CO2 atmosphere, which in the case of naphthoic acid results in >90% conversion to char, with no 2,6-NDA detectable (0.2% detection limit). Thermal decomposition of the aromatic acids in the absence of C02 results in even faster destruction, with the generation of CO2. Condensed aromatic hydrocarbons, aldehydes, and ketones are formed as

well as the parent hydrocarbon. Crosslinked species represent a large portion of the total products. Further, if napthalene or benzene is heated with ZnO and K2CO3 in the presence of CO2, no detectable naphthalene carboxylic or dicarboxylic acids, or benzoic acid or benzene dicarboxylic acids are formed. There is no precedent in the literature for direct addition of C02 to a hydrocarbon, even in the presence of a potassium salt with a Henkel catalyst.

Further, it is found that salts such as K2-2,6-NDA slowly decompose under Henkel reaction conditions (250psig of CO2, 450°C), typically at the rate of a 2-6% per hour. Therefore it is seen that even the Henkel product salts are not truly stable under the Henkel reaction conditions. The Henkel process, such as the disproportionation of potassium naphthoate to form K2-2,6-NDA, is thus a balancing act between rates of formation and loss of the desired product.

A net production of naphthalene always occurs in this reaction, even in the presence of excess base which results in some direct carboxylation of the naphthoic acid to K2- 2,3 NDA followed by isomerization of that salt to K2-2,6- NDA. Therefore, it is unexpected and novel in the literature and practice of Henkel chemistry that it should be possible to add CO2 directly to a hydrocarbon to produce an aromatic acid or diacid. However, we have determined that this novel and useful reaction will indeed occur under conditions of contact of molten strongly basic carbonate salts with aromatic hydrocarbons, with or without Henkel catalysts such as ZnO. The rate of formation is influenced by temperature, C02 pressure, hydrocarbon pressure, and surface area of the salt/hydrocarbon interface.

The following examples will serve to illustrate specific embodiments of the invention disclosed herein.

These examples are intended only as a means of illustration

and should not be construed as limiting the scope of the invention in any way. Those skilled in the art will recognize many variations that may be made without departing from the spirit of the disclosed invention.

Example 1-Comparative 10g of naphthalene, 10g of K2CO3, and 2g of ZnO is charged into a 100cc Hoke vessel, which is pressured to 250 psig with C02 and heated to 450°C for 3 hours. On cooling, no potassium salts of naphthoic acids or naphthalene dicarboxylic acids are detectable in the product mixture.

Example 2-Comparative A mixture of 10% benzene in C02 is flowed through a flow reactor packed with 5g of ZnO at 450°C for 3 hours. At the end of this time, no benzoic acid or TPA acid or their salts are detectable in the product bed.

Example 3-Comparative The experiment of Example 2 is repeated with A1203 as the catalyst bed. No benzene acids or diacids are generated.

Example 4-Comparative The experiment of Example 3 is repeated with K2CO3 as the catalyst bed. No detectable amount of potassium benzoate or TPA salt or acid is formed.

Example 5-Inventive In Example 5 a eutectic mixture of 100g of a 1: 1: 1 molar mixture of K2CO3, Rb2CO3, and Cs2C03 is charged to a 500 cc autoclave with 200g of naphthalene and 10g of ZnO.

The autoclave is fitted with a heater and high efficiency stirrer capable of thoroughly agitating a dense liquid. The autoclave is heated to 175°C, and pressured and depressured with C02 to dry the charge. It is then pressured to 500 psig with CO2. It is then observed that when it is heated above

ca. 400°C with rapid stirring, the carbonate salt charge has the appearance of small frozen"shot"on cooling, indicating melting or at least softening of the eutectic salt mixture. When the autoclave is then charged as described, and heated to about 450°C for 3 hours with vigorous (2000 rpm) stirring, the product frozen carbonate salt"shot"is observed on analysis by quantitative NMR or LC to contain significant quantities of naphthalene carboxylic acid salts, of from 0.5 to 3% by weight of the carbonate salt, the exact amount depending on the mixing and temperature. It is found that the initial product observed is the alkali metal salt of 2-naphthoic acid.

Quenching at later times results in the observation of alkali metal salts of 2,3 NDA and 2,6 NDA as increasing products. By short contact time or rapid quenching, a material containing largely alkali salt of 2-naphthoic acid can be produced. The alkali napthalene acid salts so produced, if concentrated by reducing the large excess of carbonate salt, are then excellent feeds for conventional Henkel I and II processes, resulting in primarily alkali salts of 2,6 NDA. If the hydrocarbon is removed, the Henkel reaction can be performed (albeit inefficiently) on the dilute napthalene acid in carbonate salt mixture as formed directly from the hydrocarbon, CO2, and carbonate salts.

Example 6-Inventive A flow reactor is set up as in Comparative Example 2.

A mixture of about 10% benzene in C02 is flowed over a bed of alumina containing about 30% by weight of the K, Rb, Cs eutectic carbonate salt mixture of Example 5. The flow reactor is heated to 450°C and the flow continued for 6 hours at a GHSV of about 1000 with the result that about 5% by weight of the resulting supported eutectic has become

alkali benzene acid and diacid salts at the end of the reaction period.

Example 7-Inventive The reactor of Example 6 is heated to 500°C and the experiment repeated with essentially the same result except that some dark decomposition products are apparent in trace quantities as well as the desired products. The concentration of TPA salt is seen to be higher than in Example 6. In the case of examples 5 and 6, calculations indicate that a thin surface layer of organic acid salts is formed at the interface of the supported eutectic and the hydrocarbon/C02 vapor. The reaction appears to slow when this"crust"contains enough diacid salt to reduce fluidity below a desirable level. It is seen that the amount of organic acids formed from the hydrocarbons, carbonates, and C02 increases with the surface area of the support/eutectic system.

Example 8-Inventive The supported eutectic of example 6 is added to an autoclave, which is charged with an equal amount of naphthalene and pressured to 300 psig with CO2. On heating to 475°C for 4 hours, it is observed that about 2-5% of the eutectic material is now naphthalene carboxylic acids, in similar ratio to that obtained in Example 5.

Example 9-Inventive A eutectic of Na, K, Rb, Cs carbonates is used instead of the ternary blend used in example 5. Similar results are obtained.

Examples 10-16 Examples 10-16 demonstrate the carboxylation of benzene. The experiments were run at 430-460°C, 250 psi C02 pressure prior to heating, 3 hours at temperature, no

mixing, 150 cc Hoke vessel as reactor, and a band heater. A Hastelloy C vessel, barricaded for potentially corrosive mixtures was used. The final pressure was 400-1100psig, preferably 600-800 psig. The following abbreviations were used: BA = Benzoic Acid; BTA = Benzene Tricarboxylic Acid (usually as 1,3,5, BTA or Trimesic Acid). Results for the carboxylation of benzene of examples 10--16 are shown in Table 1: Table 1 D2O Sol. Feed Bulk Rate Final/ Products Charged Density Production Initial solid Components & LR No. Mg/g Mg/g G/cc g/L/hr Wts. Ex. 10 BTA 1.02 37 0.9 0.3 0.951 BA 0.35 powder 0.11 Li2CO3 1.65 Na2CO3 3.36 K2CO3 3.11 K2Oxalt 2.13 Al2O3 10.19 ZnCO3 2.07 PhH 7.87 Ex. 11 BTA 0.70 30 1.1 0.3 1.004 Li2CO3 1.70 Na2CO3 3.26 K2CO3 3.10 K2Oxalt 2.03 Al2O3 10.10 ZnO 2.0 PhH 7.87 Ex. 12 TPA 0.29 8 1.83 0.2 1.33 Li2CO3 1.69 BA 0.30 Na2CO3 3.75 K2CO3 3.04 K2Oxalt 2.02 Ag2O 2.01 PhH 7.87 D2O Sol. Feed Bulk Rate Final/ Products Charged Density Production Initial solid Components & LR No. Mg/g Mg/g G/cc g/L/hr Wts. Ex. 13 TPA 3.34 16 1.63 1.81 1.19 on K2CO3 3.37 IPA 0.21 0.11 solid charge Rb2CO3 3.99 BA 1.09 0.59 Cs2CO3 4.98 (1.4 in Al2O3 15.23 meas. PhH 7.87 TPV) Ex. 14 BTA 0.39 4 1.6 0.21 0.917 K2CO3 3.13 TPA 0.72 0.4 Rb2CO3 4.2 BA 0.18 0.1 Cs2CO3 5.37 ZnCO3 3.14 Al2O3 15.18 Ex. 15 BTA 4.3 4 1.7 2.4 1.15 on solid K2CO3 3.13 TPA 0.39 (1 mg in 0.2 Rb2CO3 4.2 BA 0.47 TPV) 0.3 0.932 Cs2CO3 5.37 on total ZnCO3 3.14 Al2O3 15.18 K2 Oxalate 2.19 Ex. 16 BTA 0.45 23 1.1 0.2 1.3 on solid K, Rb, Cs, Zn BA 12.1 4.4 CO3 eutectic 3.11 (8.5 in 0.932 on Al2O3 15.03 TPV) total

If the feed is a vapor (such as benzene, critical point 290°C) in the reactor, the vapor is assumed uniformly distributed throughout the reactor, and therefore the mg/g value will be the number of mg in the reactor divided by the cc/volume of the reactor, and multiplied times the open volume per g of solid phase. In such a case the estimate is obviously very crude in a chemical sense, but relates well to the engineering production capacity of a reactor full of solid phase.

Since benzene is a supercritical gas (Tc = 289°C) under the reaction conditions. Density of Cs carbonate is ca. 3.5 as a pure phase; assume it is 1.1 as a powder; that means the interstices between the powder are 2/3 of the volume; so the amount of benzene feed present interspersed with lg of solid is 2/3 of 0.909cc of supercritical gas.

The total volume of gas is ca. 160cc; 7.87 g of benzene charged, gives 30 mg of benzene. This would be the amount of benzene around the solid in a reactor full of solid.