Anhydrous Diluents for the Isobutylene Oxidation Reaction to Methacrolein and Methacrolein Oxidation to Methacrylic Acid

1. Field of the Invention

This invention relates to a process for the manufacture of methacrolein or methacrylic acid from isobutylene. More specifically, it describes an improved process for producing methacrolein or producing methacrylic acid by the catalytic vapor phase oxidation of isobutylene in the presence of essentially inert essentially anhydrous diluents.

2. Summary of the Prior Art

Generally, isobutylene in its gaseous phase is oxidized to methacrolein in the presence of molecular oxygen-containing gases and steam, whose concentration is often as high as about 35 volume percent of the total feed stream, by contact at elevated temperatures with solid metal oxide catalysts. The methacrolein produced in this reaction stage can be recovered or can be directed without separation of the methacrolein to a second reactor operating in series with the first reactor to oxidize the methacrolein to methacrylic acid.

In the prior art, steam has been used in the starting reactant gas mixture in order to avoid flammable gas mixtures and because it was believed to be important to reaction selectivity, but there is no indication of composite heat capacity of the importance of the added diluent gas. For example, U.S. Pat. No. 4,147,885 states that in a related process for oxidizing propylene to acrolein, it is wide practice to incorporate steam to avoid burning the reactant gases and increase selectivity to acrylic acid. Similarly, U.S. Pat. No. 3,475,488 discloses that it is desirable to incorporate steam in the starting reactant gas since this increases conversion and selectivity when employed in the order of 1 to 60, and preferably 5 to 30, moles of steam per mole of propylene or propylene plus acrolein.

Other patents also describe steam as the preferred diluent. For example, U.S. Pat. No. 3,171,859 states that "addition of steam is obligatory . . . it acts not only as a diluent, but also favors the reaction in that combustion to carbon oxides is substantially reduced." Also, U.S. Pat. No. 4,267,386 reiterates the general understanding among those skilled in the art, that while inert diluents may be added to the reaction system, "water, in the form of steam is desirably present . . . in amounts of from 0.5 to 15, preferably 2 to 15, moles per mole of unsaturated hydrocarbon (i.e., propylene or acrolein)." Again, with no recognition of composite heat capacity value of the diluent.

Many oxidation catalysts have been disclosed for producing methacrolein in high yield by oxidizing isobutylene. Predominantly, these are catalysts containing mixed oxides of molybdenum, bismuth and iron with phosphorous or tungsten or antimony. Cobalt and/or nickel and alkali metals are common promoters.

Catalysts which have been found to be advantageous for use in oxidizing methacrolein to methacrylic acid generally contain mixed metal oxides. Such catalysts typically contain molybdenum, vanadium, tungsten, chromium, copper, niobium, tantalum and antimony.

It has heretofore been well recognized in the relevant catalyst art that water, in the form of steam, is an important component of the inert gas diluent (see, e.g., U.S. Pat. No. 4,267,385, which states that it is "preferable to carry out the oxidation reaction in the presence of steam.") The ultimate object of the teachings in the literature cited above is to obtain high performance catalysts which give high selectivities to methacrolein and methacrylic acid at high isobutylene conversions. Other factors which influence the economic viability or the improved performance of these processes are not considered in these prior art techniques. For example, they do not adequately address the impact on process variables of use of high isobutylene concentrations, how to avoid the danger of explosion, the impact of inert reaction process feeds on recovery and waste disposal, or maintaining high catalyst performance over an extended catalyst life. These are all extremely important for commercial operation.

In commercial operation, it is of economic and ecological importance to minimize the presence of the steam which is fed to the reactors, since it passes through the system and becomes a burdensome waste water load after product recovery steps; nevertheless, to the knowledge of the present inventors, no commercial process has been successfully operated below a steam: isobutylene mole ratio of about 1.5:1. Furthermore, it is extremely important to minimize by-products which are difficult to separate from useful product or which carry a high economic penalty for disposal. Process improvements which will provide high catalyst performance while simultaneously maximizing isobutylene feedstock usage, and improvements which may promote conditions for extended useful catalyst life are important for commercial operation. Equally important is the ecological problem encountered when consideration has to be given to the disposal of millions of pounds of waste water, which will vary from about 0.5 to about 1.5 pounds of water per pound of acrolein/acrylic acid produced, generated by a single multimillion pound commercial facility for the production of methacrolein and/or methacrylic acid. The typical commercial plant has an annual capacity of from about 100 million to over 750 million pounds, thus giving an indication of the waste water problem. The prior art does not adequately address these issues.

U.S. Pat. No. 4,049,577 teaches an improved catalyst composition for making acrolein. The authors mention that recycle gas comprised of the noncondensable fraction of the product can be used in place of steam. They suggest that these recycled inerts are preferable to steam as diluent since they allow higher conversions of propylene and thus enable one to obtain higher yields, and also reduce the water load on the system; however, the use of recycled inerts is stated as being made possible by the characteristics of this particular catalyst composition. Nowhere does this patent suggest or teach that anhydrous diluents having hereinafter defined composite heat capacity have an improved effect on selectivity or product mix, or are useful with other catalysts. The relationships between various diluents and the heat capacity effects on selectivity are not suggested.

U.S. Pat. No. 3,801,634 teaches the use of inert solids mixed with active catalyst in the first and second stage reactors used to manufacture acrolein and acrylic acid. The authors indicate that noncondensable, second-stage effluent gases can be recycled to the first stage as inert diluting gas which can, at least partly, replace steam. The authors do not show any relationship between the inert anhydrous diluent gases and improvements in product selectivity; or the desirable effect of composite heat capacity value.

U.S. Pat. No. 4,031,135 presents a recycle process in which noncondensable gases, preferably and generally including steam, are recycled to the first-stage reactor and also to the interstage (second-stage) reactor feed. There is no recognition of the benefits in using anhydrous diluents having certain composite heat capacity values with respect to their effect on yield, conversion, by-product selectivity mix and waste water generation. The authors do, however, recognize an apparent improved acrylic acid efficiency, which they attribute partly to the use of recycled off gas employed as the inert diluent. In column 4, lines 13-15 the patent says the off-gas "has been substantially freed from condensable products, including water, and essentially consists of nitrogen and small amounts of" other named compounds. In column 6, lines 45 to 54 the general composition, in volume percent of the off-gas is stated in broad and "especially" terms as being:

Further, in the only two examples presented in support of their invention the patentees specifically disclose the presence of 2% by volume and 8.9% by volume of steam in the recycled off-gas clearly not an essentially anhydrous diluent off-gas stream. The patent clearly teaches the recycle of nitrogen, in impure form, and steam in the off-gas to the reactors. In the invention described in this specification the essentially inert diluent gas feed is an essentially anhydrous diluent gas composition, as hereinafter described. The gas feed differs significantly from the teachings of U.S. Pat. No. 4,031,135, it is essentially anhydrous. It is further to be noted that the patentees do not adequately establish improved efficiency to acrylic acid production, and that they do not recognize or disclose the effect of composite heat capacity of the diluent on product selectivities.

U.S. Pat. No. 4,365,087 refers to the recycling of dewatered residue gas, containing both inert and reactive gases, to increase the concentration of acrylic acid recovered. However, the authors not only consider this procedure unsatisfactory since the composition of the residue gas fluctuates but have no recognition of the concept of composite heat capacity of the diluent and its effect on the process.

U.S. Pat. No. 4,442,308 teaches the use of inert gases as diluent in the acrolein process; however, it specifies their use for a particular supported first-stage acrolein catalyst. Most common commercial catalysts for propylene oxidation to acrolein are neat (unsupported) and do not follow this patent's prescribed preparation. This patent also claims that 0.5 to 7 mole % steam is beneficial and its use is recommended. Nowhere in this patent do the authors teach the advantage of anhydrous diluents on product mix nor do they mention composite heat capacity or flowing heat capacity as major variables in controlling product selectivity to advantage and formation of undesired product streams.

U.S. Pat. No. 4,456,006 teaches a catalyst preparation for the propylene-to-acrolein reaction. It shows that nitrogen diluent presents an improvement over steam diluent when used with this catalyst. It does not recognize or disclose the composite heat capacity effect of diluent on product selectivity, nor does it show by-product and waste water reductions when using anhydrous diluents.

U.S. Pat. No. 3,717,675 describes a process for recovery of acrylic acid where acrolein is expelled from the aqueous acid collected and returned to the reactors to increase subsequent yields of acrylic acid. This patent mentions the use of inert diluents such as carbon oxides and nitrogen, but does nothing to demonstrate their importance. In fact, it states that it is necessary to add steam to the reaction in order to increase selectivity. This addition of steam, however, only serves to aggravate the waste water disposal problem.

UK Patent 2,068,947 teaches a process for producing methacrolein and methacrylic acid whereby inert anhydrous diluent gases are used, also combined with water vapor, to produce a product with a reduced quantity of condensables compared to the typical steam diluent process. The authors fail to recognize the relationship between anhydrous diluents and acetic acid reduction, and they do not address composite heat capacity of the diluent or selectivity improvements resultant from use of various anhydrous diluents.

U.S. Pat. No. 4,147,885 describes a recycle process in which steam is an essential ingredient. The object of the patented invention is to recycle steam to the reactors. This is contrary to the techniques of the instant invention, since it has now been found that the reduction or absence of added steam to the reactors is beneficial.

U.S. Pat. No. 4,618,709 presents an attempt to remedy, or at least alleviate, the waste water problem common to the existing catalytic oxidation processes for producing methacrolein and methacrylic acid from, e.g., isobutylene. This is accomplished in this patent by evaporating the waste water solution and subjecting the waste water vapor to combustion with molecular oxygen, whereby the amount of liquid waste water discharged is reduced. As can be seen, this is an expensive procedure since it involves two additional costly steps. In discussing the isobutylene oxidation process the patent refers to the well known and common procedure of carrying out the oxidation in the coexistence of an inert gas for dilution (column 1, lines 22-35 and 49-53) to control temperature and prevent explosion. The patent then mentions "nitrogen, water vapor, exhaust gases, etc." as examples of inert gases added and further states water vapor as being the most frequently employed, in an amount as great as 10 to 50 moles of water, or water vapor, added per mole of methacrolein or its precursor (column 1, lines 54-58). These figures clearly evidence the intentional addition of significant quantities of water to the oxidation reaction, water that becomes contaminated with reactants and products of the reaction and must subsequently be disposed of in an ecologically accepted mode. The reference nowhere suggests or discloses the importance of using an inert gas having a composite heat capacity of certain value for dilution. Nor does the reference recognize or suggest the importance of avoiding the intentional addition of supplemental quantities of water to the reaction system.

UK Patent Specification 939,713 disclosed one of the earliest catalytic processes for preparing unsaturated monocarboxylic acids from olefins. In these early processes yields of acrolein and acrylic acid were low, as evidenced by the figures presented in the examples that show low conversion, and overall yields of less than about 75% from propylene charged. By comparison, today's processes operate at exceptionally high yields, and conversions that typically approach 95% to 98%. On page 2, lines 4 to 19, this UK Specification refers to the starting materials and indicates they need not be in a pure state and may contain quantities of paraffinic hydrocarbons, such as propane or butanes. It is to be noted it is present in the starting olefin reactant as an impurity and it is not intentionally additionally introduced into the reactant as a diluent. It is also noted that its quantity is nowhere clarified and that its function is described as an entraining agent; there is no recognition of its possible use as a medium for heat removal, nor does the reference state or suggest that the hydrocarbons have any effect on maintaining the desired temperature range. Though the reference states the process can be operated in the absence of water (page 2, lines 92-94), this statement is both immediately preceded and followed by the statement that water is preferably used in quantities of from 1 to 10 moles of water, preferably 3 to 7 moles for each mole of olefin initially introduced into the first reaction zone and all of the examples use significant amounts of water. The reference does state the reaction can be carried out in the absence of water but it nowhere indicates the water must be replaced. It merely states do not add the water. Any attempt to carry out this reaction without any added diluent would be catastrophic. In Examples I and II, 4.2 moles of water were intentionally added per mole of propylene, in Example III, 5.8 moles of water were intentionally added per mole of propylene and in Examples IV and V, 12 moles of water were added per mole of propylene. Thus, water comprised the majority of the bulk of the materials introduced in all examples. Further, nowhere in this UK Specification is there any mention of the use of any other coolant or temperature control medium. Nor is there any suggestion or recognition of the importance of the composite heat capacity of the gas diluent and its effect on conversion and yield.

None of the prior art suggests or recognizes the use of various inert anhydrous diluents in specific proportions so as to have the hereinafter defined composite heat capacity that will favorably affect the product mix obtained when using any of the commonly used catalysts.

As has been indicated above, the basic two-stage process for oxidizing isobutylene to methacrylic acid via methacrolein is well known and has been extensively described in the literature. In connection with the related acrylic acid process, it is also known that wet, overhead gases (noncondensables) from the acrylic acid scrubber can be recycled to the first reactor stage. By this recycling of unreacted propylene and acrolein, it is predictable that an improvement in overall yield is obtained in any chemical reaction. By use of such a recycle stream, it is also possible to provide a supplemental means of controlling the steam content to the first-stage reactor, as is taught in U.S. Pat. No. 4,147,885. In the process of that patent, the steam content of the first-stage feed is required to be 4 to 30% by volume, with all the steam, except that in the starting reactant gas mixture, being provided by the recycle stream. As discussed above, however, the presence of even as little as 4% steam is disadvantageous. This finding has not been addressed, nor even identified, by the prior art.

As has also been indicated above, nowhere in the prior art is there any disclosure or suggestion of the important role exerted on the process by the composite heat capacity of the diluent gas mixture and of the unexpected and unpredictable effect exerted by the composite heat capacity of said mixture on yield, conversion, by-product formation and waste water generation.

FIGS. 1 and 2 illustrate the application of recycle streams to methacrolein and methacrylic acid processes.

The present invention embraces two separate but related concepts, namely, the reduction or elimination of steam to reduce the water load on the process, and the improvement of selectivity to maximize the output of desired products. Both of these results are achieved by elimination of the steam diluent intentionally added in prior art processes and use of an essentially anhydrous gaseous diluent (optionally containing a minimal amount of steam which is generated from trace impurities of water that may be present in the reactants and diluents initially charged to the reactor) having a composite heat capacity in a selected range as hereinafter defined.

In a preferred embodiment, a portion of the non-condensable essentially anhydrous gases from the process, e.g., the overhead stream from the acid scrubber, is recycled back to the first-stage reactor feed stream. In addition, most preferred would be the use of enriched or pure oxygen in the two stages to replace the use of air.

It will be understood that the process of this invention can be applied not only to a combined isobutylene methacrolein-methacrylic acid process, but also to a separate methacrolein-methacrylic acid process, or to the methacrolein-methacrylic acid leg of an isobutylene-methacrylic acid process. Thus, a portion of the stream from the first-stage isobutylene-methacrolein reactor can be sent to a methacrolein recovery process, from which some or all of the non-condensable overhead gases from the methacrolein scrubber system can be recycled as diluent to the first and/or second-stage of the isobutylene methacrylic acid process.

For purposes of this invention, the term "isobutylene" is intended to include as well other compounds which can be converted in situ to isobutylene, e.g., t-butyl alcohol, t-butyl acetate, methyl t-butyl ether, and the like, and/or which respond to the same catalysts as does isobutylene and proceed through the same intermediates (i.e., isobutylene and methacrolein).

According to the invention, it has been discovered that essentially anhydrous diluent gases having high heat capacity can be used to reduce or essentially completely replace steam in the isobutylene oxidation reaction to efficiently produce methacrolein and methacrylic acid. (For purposes of this invention, a diluent is any gas which does not react in the reaction stage in which it exists.) Furthermore, when replacing the steam diluent with essentially anhydrous diluents, major by-products, e.g., acetaldehyde and acetic acid, are significantly reduced. This reduction in by-products is especially important since acetaldehyde forms acetic acid over the methacrolein-to-methacrylic acid reaction step, and acetic acid is a significant waste product. To make saleable quality methacrylic acid, considerable energy is required to more completely remove the acetic acid. Waste disposal costs for disposing of acetic acid are high. The present invention provides a means for reducing acetic acid disposal costs, decreasing separation costs for both methacrolein recovery and methacrylic acid recovery, and enables existing equipment to enact a better separation, thus saving methacrylic acid recovery losses and providing potential for higher quality refined product.

Another key discovery of this invention is that by increasing the flowing heat capacity of the reactant gas mixture, the yield of useful products can be increased significantly. The flowing heat capacity is increased by the introduction of an essentially anhydrous diluent having a relatively high composite heat capacity (as defined herein), comprising one or more essentially inert essentially anhydrous gases with relatively high molar heat capacities. Flowing heat capacity is the composite heat capacity of the essentially anhydrous diluent, plus the heat capacity of the reactants, i.e., the flowing heat capacity is the composite heat capacity of the total gas stream. However, flowing heat capacity does not change appreciably as a result of reaction, since various reaction products have a higher heat capacity than that of the reactants, and some have a lower heat capacity. In general, the flowing heat capacity will not be expected to typically change by more than about one heat capacity unit as a result of reactions. Thus, the composite heat capacity of the diluent is a dominant variable for process control purposes.

As the flowing heat capacity of the reaction feed gas mixture is increased, yield to methacrolein, and methacrolein plus methacrylic acid increases, and the flammable gas range is reduced, enabling higher productivity operations. Simultaneously, the peak temperature in the catalyst bed due to the exothermic heat of reaction is lessened and the heat of reaction that is released is absorbed more efficiently in the bulk gas stream. This, in turn, should increase catalyst life by decreasing thermal stresses within the catalyst pellets' structure, reducing potential carbon build-up within catalyst pores, and by reducing pressure drop, since there will be lower volumetric flow of reactant gas feeds necessary to meet a given production level.

The invention is advantageous for recycling diluent gases and unreacted isobutylene back to the reactors. The resulting low-steam-containing product streams provide an ample source of noncondensable diluent, so that separation of useful product is simplified. This is particularly advantageous for methacrolein recovery, since methacrolein, which is more volatile than water, can be effectively separated from the reaction-produced water without loss of diluent. By using essentially anhydrous diluents with higher volatilities compared to methacrolein, the present invention permits operation on a methacrolein recovery system with recycle of diluent, unreacted isobutylene, and unrecovered methacrolein back to the reactor for further efficiency gains and cost reductions. Such a system using steam diluent is not possible when adapting prior art methacrolein recovery equipment and techniques. It also enables implementation of recycle processes in which components such as acetic acid and acrylic acid and other minor, heavy by-products are excluded from the recycle stream. This is significant since the acids and heavy by-products are suspected of adversely affecting catalyst life, and furthermore, it aids in minimizing recycle handling problems, such as compressor corrosion.

The composition of the process feeds must be comprised so that flammable gas mixtures are not formed. According to this invention, the starting reactant gas mixture to the first-stage reactor typically contains up to about 16 g-moles per hr. of isobutylene, preferably up to about 8 g-moles per hr. of isobutylene, per liter of first-stage catalyst; about 1.1 to about 2.2 moles of molecular oxygen per mole of isobutylene, and an essentially inert diluent gas having a composite heat capacity of at least about 6.5 calories/gram-mole (.degree.C.) which comprises about 40 to about 94% by volume of the feed stream. The oxygen source can be air, oxygen-enriched air, essentially pure oxygen or a mixture of oxygen and essentially inert anhydrous gases. As used in this specification by the term essentially inert anhydrous diluent gas (or variant thereof) is meant the inert gas stream of one or more gases introduced into a reactor to which additional water in any form has not intentionally been added before the inert gas is introduced into the reactor but which inert gas stream may contain trace impurities of water, or which water may have been introduced into the reactor as a trace impurity present in the oxygen feed, or formed during the reaction. It is desirable that the mole ratio of composite diluent to isobutylene be in the range of about 2 to about 32. The essentially anhydrous diluent gas typically comprises a mixture of nitrogen, carbon dioxide, methane, ethane, propane and butanes; however, any other essentially anhydrous inert gas can be included. Some other useful inert gases include helium, argon, saturated hydrocarbon gases, N.sub.2 O, and carbon monoxide. When water is present as a trace impurity in any of the materials introduced into the reactors, at the elevated temperature required for these reactions the water is immediately converted to steam. The materials used should preferably be free of water, but in those instances in which water may be present as an impurity, the total amount thereof in all materials added should be no more than about 0.4 mole per mole of isobutylene, preferably 0 to about 0.3 mole per mole of isobutylene, and most preferably zero. The inert diluent should be of sufficient quantity to avoid flammable mixtures when combined with the isobutylene and molecular oxygen. Air or oxygen stream or pure oxygen can be used as the molecular oxygen source. Of course, if air is used, the contained nitrogen acts as a supplemental diluent. In the process of this invention the intentional introduction of extraneous steam to a reactor is not contemplated.

For each inert essentially anhydrous diluent gas mixture there is a relationship which can be determined by experiment and which describes the limiting compositions of oxygen, isobutylene, and diluent for which flammable mixtures exist. Most commercial applications are most desirably operated in a "fuel-rich" mode, whereby the oxygen content is the limiting factor from a flammability standpoint. The isobutylene concentrations will be determined by catalyst performance and by commercial cost effectiveness factors.

It is a distinct advantage of this invention that, since diluent gas mixtures with high composite heat capacities have a tendency to broaden the operable range due to shrinkage of the flammable gas envelope, high isobutylene concentrations are possible. It is theorized that first stage isobutylene feed concentrations as high as about 30 mole % will be achievable using the method of this invention.

Typically approximate ranges for feed compositions are defined based on the generalized operating constraints discussed above. First-stage feeds in the following ranges are typically particularly useful:

Isobutylene: up to about 16 g-mole per hr. per liter of first-stage catalyst, preferably up to about 8 g-mole per hr. per liter of first-stage catalyst;

Oxygen: 1.1 to 2.2:1 O.sub.2 /C.sub.4 H.sub.8 ratio, such that there is up to about 33.6 g-mole O.sub.2 per hr. per liter of first-stage catalyst, preferably up to about 21 g-mole O.sub.2 per hr. liter of first-stage catalyst;

Diluent: About 2.0 to 32:1 diluent/C.sub.4 H.sub.8 ratio, preferably 3.5 to 12 diluent/C.sub.4 H.sub.8 ratio. Nevertheless one can use amounts slightly below or above the above ratios, e.g., one can go as low as about 0.5:1 and as high as about 33:1.

The process of the invention is particularly advantageous in that it is not dependent upon any particular catalyst, as is much of the prior art, and will provide its benefits for any catalyst of choice. Any molybdenum, iron-based mixed metal oxide oxidation catalyst, such as those disclosed in U.S. Pat. Nos. 3,825,600; 3,649,930; 4,339,355; 4,267,385; and 4,306,090; (catalysts relevant to the first stage are also shown in U.S. Pat. Nos. 4,012,449 and 4,354,044) can be used in the isobutylene-to-methacrolein oxidation reactor. Any Mo, P-based mixed metal oxide oxidation catalyst, such as described in U.S. Pat. Nos. 4,419,270; 4,444,907; and 4,051,179 can be used effectively in the second stage (i.e., the methacrolein oxidation to methacrylic acid reaction).

The general reaction conditions are not narrowly critical, and are those known to the art. The first-stage reaction operates at temperatures of 250.degree. C. to about 450.degree. C., although temperatures of about 300.degree. C. to about 400.degree. C. are preferred. The second-stage reaction requires temperatures of about 200.degree. C. to about 450.degree. C., with a preferred range of about 250.degree. C. to about 375.degree. C.

Operating pressures of about 1 to about 4 atmospheres are typical, although this process improvement will apply for all operating pressures, whether subatmospheric, atmospheric, or superatmospheric. Preferred commercial modes of operation will minimize pressures, but pressures are typically held in the 2 to 3 atmosphere range due to system pressure-drop constraints.

Flow rates can be varied from about 0.5 to about 15 seconds contact time; however, typical commercial flow provides about 1.5 to about 4 seconds contact time. Contact times of about 1.7 to about 3 seconds are preferred.

As indicated above, selection of proper heat capacity of the essentially inert anhydrous diluent gas or gases is critical to the proper performance of the invention. Since the essentially inert essentially anhydrous diluent gas stream may comprise a mixture of several individual gases, it is convenient to refer to a composite heat capacity for the total stream. The term "composite heat capacity," as used herein, means the sum of the products of the volumetric fraction of each gas in the diluent gas mixture and its heat capacity. (Heat capacity, as referred to herein, is the ideal gas heat capacity determined at 330.degree. C. for purposes of the composite heat capacity definition.) The composite heat capacity for the essentially inert anhydrous diluent gas going to the first stage reactor should be at least about 6.5 calories/gram-mole (.degree.C.). Below this value, the product selectivity benefits of this invention are minimal. There is no known upper limit on composite heat capacity; however, it is theorized that above a value of about 40 there may be an unrecoverable heat loss through absorption of reaction heat into the process stream, which would result in an economic penalty. In addition, there could be a problem with increased after-burning at the exit of the first-stage reactor. It is preferred that the composite heat capacity be maintained from about 8 to about 30, and most preferably about 10 to about 20. Assume the presence of four gases in the inert diluent gas stream, A, B, C, D, in volumetric presence of 20% A, 40% B, 30% C, 10% D. Assuming heat capacity in degrees centigrade of w cal/ gram-mole for gas A, x cal/gram-mole for gas B, y cal/gram-mole for gas C and z cal/gram-mole for gas D. Then the "composite heat capacity" (CHC) of the inert anhydrous diluent gas stream is expressed by the equation:

The sum of these should be at least about 6.5 calories/gram-mole (.degree.C.) as indicated above.

The flowing heat capacity of the second-stage reactor feed gases is determined predominantly by the choice of essentially inert anhydrous diluent gas fed to the first-stage reactor. The first-stage product mix has only a minor influence on the second-stage feed flowing heat capacity, since the products typically account for only about ten to twenty percent of the total stream volume. For example, a typical operation with 6% isobutylene and 13% oxygen produces water plus methacrolein, methacrylic acid, acrolein, acrylic acid, acetaldehyde, acetic acid, and carbon oxides. The average heat capacity of the feed isobutylene and oxygen is very nearly the same as the average heat capacity of the resulting products (approximately 0.65 cal/g-mole (0.degree. C.) more for products versus reactants).

The essentially anhydrous diluent gas of this invention introduced into the reactor can be a single gas or a multi-component mixture of gases, provided that certain criteria are observed. Each gas must be essentially inert to the oxidation reactions of the process, and each gas must be non-condensable under typical operating conditions and readily separable from the reaction products.

Since each plant installation will have specific constraints that affect the energy usage for the entire plant, attention must be paid to the impact on the plant energy balance for use of particular diluents which alter the current heat recovery schemes. For example, a high heat capacity diluent will retain more of the heat evolved through reaction, whereas now the bath temperature is relied on more to remove and recover the heat of reaction. High heat capacity diluents will require more attention to recovering heat after the reaction. Furthermore, if process off-gas is disposed of via combustion, the recovery of heat will be affected by a major change in diluents.

In addition, one should avoid catalyst poisons, e.g., sulfur dioxide, and gases that react to form unwanted by-products, as would unsaturated hydrocarbon compounds (e.g., propylene); or NH.sub.3, which produces methacrylonitrile.

It is another advantage of this invention that the steam component typically contained in the feed to the first stage can be minimized, even eliminated. While there is controversy among those skilled in the art as to the precise function of steam, e.g., whether it is truly an inert diluent or whether it somehow participates in the oxidation of isobutylene and methacrolein, it is accepted practice in the art as is currently performed that a significant concentration of steam is required in order to successfully operate the first-stage and second-stage reactions. Contrary to this holding of the art, it was a surprising, unexpected and unpredictable discovery of the present invention that intentional addition of steam is desirably eliminated entirely. This is accomplished by substituting for the steam the essentially inert essentially anhydrous gas diluent of selected composite heat capacity as described in this invention. Accordingly, it has been found that the steam content of the feed gas can be essentially zero. While not preferred, the steam content of the feed gas can range as high as about 3% by volume of the feed gas when the materials comprising the feed gas have not previously been treated to remove water. However, it is preferred that the steam content resulting from water as an impurity in the feed stream be kept below about 2% more preferably below about 1% by volume and most preferably zero. Since steam is not intentionally added to the reactors this waste water disposal problem is significantly reduced.

While not an absolute requirement of this invention, it is highly preferred that the essentially inert anhydrous diluent gas used be, at least in part, an essentially anhydrous recycle stream from within the process. Preferably, this will comprise a portion of the non-condensable, overhead gas mixture from the methacrolein or methacrylic acid recovery scrubber train which removes water and methacrylic acid from the product mixture. In particular, the use of low-boiling, anhydrous diluents makes recycle from a methacrolein recovery process possible. Besides providing a beneficial system which allows recovery of current methacrolein separation efficiency losses and reuse of unconverted isobutylene, recycling process gases from a methacrolein recovery process is desirable and advantageous to recycling process gases from a methacrylic acid recovery process as described in the prior art. Generally, in the process of this invention additional high heat capacity inert diluent is added to the recycle stream.

In order to minimize the water vapor carried overhead from the scrubbers, they should be operated within the ranges of conditions shown in Table A.

Under most operating conditions, it will be necessary to take off a purge stream, the size and location of which will be determined by the specific process being used. If pure oxygen is used as the source of oxygen, the purge can be relatively small. If air is used as the oxygen source, there will be a build-up of inerts, e.g., nitrogen, so that a substantial purge will be required, and will be controlled to maintain the desired composite heat capacity. The use of pure oxygen (i.e., oxygen not burdened with substantial concentrations of inert gases) permits maximization of diluents with a heat capacity higher than that of nitrogen. This permits minimization of the purge, which in turn, makes feasible the use of high heat capacity gases, such as propane or butane, which might be too expensive for use as inert diluents if they were to be substantially lost through purging.

The examples which follow illustrate the invention, but are not intended to limit it in any way. In these examples, all concentrations are in mole percent.

A pilot plant experimental set-up consists of two single tubular reaction vessels of typical commercial reactor tube dimensions. The first reactor tube contains a commercial catalyst comprising molybdenum, bismuth, iron and several promoter metals typical of first-stage catalyst, as described above. The second reactor tube is filled with a commercial second-stage catalyst, similar to those described previously, and is connected in series with the first. The gaseous reaction products are sampled and separated into condensable and noncondensable portions. Each phase sample is measured and analyzed by gas chromatograph. The resultant measurements are used to calculate reaction yields and isobutylene conversions. These sampling procedures are accomplished for both first-stage product and second-stage product, so that process performance for methacrolein production and process performance for methacrylic acid production are both determined. A jacket surrounding each tube is filled with a heat transfer fluid which circulates to remove heat of reaction. Thermocouple and sample taps are provided along the length of each reactor and at the bottom of each reactor. Gas feeds are metered into the first reactor using mass flowmeters. The first-stage effluent is then conducted directly into the second-stage reactor. The condensable portion of the second-stage effluent is recovered as a liquid tails streams from a water-based scrubber. Non-condensable gases are conducted out the top of the scrubber and could be returned, if desired, to the reactors to supply additional diluent gas. The system outlet pressure is controlled at 7 psig in order to control reactor feed pressures on the system. Isobutylene feed concentration is set at 6.0% and air feed concentration is set at 60.2%. The additional (to nitrogen in the air feed) feed gas diluent contains 2.6% nitrogen and 30% steam. (Also present is about 0.2% inert impurities in the isobutylene.) The system outlet pressure is set at 7 psig, and reactor temperature is set to 350.degree. C.

A typical isobutylene conversion of 98% with 85% selectivity to methacrolein plus methacrylic acid is achieved.

Experiment 1 is repeated, but with 32.6% nitrogen and 0% steam as the diluent gas. Temperature is adjusted to give a first-stage isobutylene conversion of 98%.

A reduction of about 50% in acetaldehyde and acetic acid byproducts is observable. Methacrolein plus methacrylic acid selectivity is increased by about 1%.

The conditions of Example 1 are repeated with methane instead of nitrogen in the diluent gas, the remainder being steam.

Acetaldehyde and acetic acid byproduct selectivities are reduced by about 50%. Methacrolein plus methacrylic acid selectivity increases by about 3%.

The terms "conversion," "yield," "selectivity," "space velocity," and "contract time" are defined as follows: ##EQU1##

Recycling of process streams is well known in the chemical process arts, and is usually implemented to improve reaction efficiencies and process economics. More specifically, recycling of product or a portion of a product stream enables efficient use of feed material not reacted in a single pass or reuse of feed material which is costly to make up in the reactor feed stream. Use of essentially anhydrous diluents has a particularly advantageous effect on the operability of recycle. It enables using a recycle stream which has less acid, thus increasing compressor operability. Furthermore, prior art recycle processes require more elaborate sampling mechanisms in order to reliably measure recycled oxygen concentrations. The control of oxygen is essential to safe operation of these recycle processes due to concerns over flammable gas mixtures. The essentially anhydrous streams of this invention, however provide for reliable and accurate monitoring of oxygen, thereby increasing the recycle process reliability and operability as well as safety.

Furthermore, an essentially anhydrous diluent process allows simple, efficient recovery and recycle of methacrolein in a methacrolein production unit.