This invention relates to the preparation of a cata¬ lyst suitable for the aldol-type condensation of a satu- rated aliphatic monocarboxylic acid compound and .formal¬ dehyde compound to yield an alpha, beta-ethylenically unsaturated aliphatic monocarboxylic acid compound of one more carbon atom than the starting saturated aliphatic monocarboxylic acid compound which comprises gelling an aqueous silica colloid and at least one water soluble alkali metal or alkaline earth metal compound.

The literature is replete with disclosures of the reaction of saturated aliphatic monocarboxylic acid com¬ pounds with formaldehyde to produce alpha, beta-ethyleni- cally unsaturated aliphatic monocarboxylic acid compounds of one more carbon atom than the saturated carboxylic acid compound. The catalysts disclosed are of two gen¬ eral types, those which are water tolerant and those that are relatively water intolerant. In those cases where water intolerant catalysts are employed, it is generally necessary or desirable to use anhydrous reactants. How¬ ever, for every molecule of alpha, beta-ethylenically unsaturated aliphatic monocarboxylic acid compound pro¬ duced there is one molecule of water by-product. Irrespective of whether or not the catalyst is water tolerant or not, Kirk-Othmer indicates in Volume 15, 3rd Edition (1981) at pages 364 and 374 that a catalyst for this reaction must provide high selectivity and high con¬ version and have at least 6 months life. Effective cata- lysts disclosed include alkali metal or alkaline earth metal aluminosilicates, potassium h-ydroxide or cesium hydroxide treated pyrogenic silica, alumina and lanthanum oxide. Kirk-Othmer indicates that the data obtained with these catalysts were in short runs and it appears that additional catalyst development is required for this method of producing alpha, beta-ethylenically unsaturated monocarboxylic acid (methacrylic acid) compounds. A

careful review of the prior art by us fails to disclose any examples of operations where the reaction has been on stream for more than a day or two. This is not sur¬ prising since our experiences have shown that catalyst life is generally low and there is a tendency for coke deposition on the catalyst with the result that the cata¬ lyst reactivity drops rapidly. It is not uncommon for coke deposition to reach unacceptable levels in 24 to 48 hours. Our studies have shown that when using approximately equal molar concentrations of propionic acid and formal¬ dehyde that silica catalysts provide a relatively high degree of conversion and selectivity based on propionic acid. However, our studies have also shown that silica supports tend to degrade over a period of time in the sense that surface area of the catalyst decrease while the average pore size increases, particularly in the presence of water. In other words, silica catalysts tend to be relatively water intolerant. Further, there is a tendency for the cations to be volatilized off over a period of time. As pointed out by Kirk-Othmer there is a need for a catalyst system which permits operation for longer periods of time. While Kirk-Othmer states that the catalyst should have at least 6 months life, we know of no prior art examples that have disclosed condensation reactions of more than 24 to 48 hours. Accordingly, there is a need for a suitable catalyst for the conver¬ sion of propionic acid compound to methacrylic acid com¬ pound which can be utilized for extended periods of time without substantial degradation during conversion and decoking operations.

Prior to this invention, silica catalysts were pre¬ pared in our laboratory by measuring the water absorption of the silica support having the desired porosity and surface area. A suitable alkali metal or alkaline earth metal compound was then dissolved in exactly the amount of water that would fill the pores of the silica support.

The aqueous composition was then deposited carefully upon the silica support in a manner such that substantially all of the aqueous catalyst composition was taken up and the silica gel support was dried. For convenience, this technique is referred to as the "incipient wetness" method. As indicated above these catalysts are rela¬ tively water intolerant since the surface area of the catalyst decreases during the condensation of propionic acid and formaldehyde while the pore size increases. Further, the more water in the feed, the faster the sur¬ face area decreases and the faster the average pore size increases. Although the catalyst tends to be active as the surface area decreases and pore diameter increases, there is a drop off in the activity of the catalyst due to loss of pore structure which is not necessarily depen¬ dent upon the loss of the alkali metal or alkaline earth metal cation. As the surface area drops to about 10 to 20 /g and the number of pores having a diameter less than 1200A ⁰ decreases, catalyst activity falls off. Accordingly, there is a need for a new silica catalyst which is more water tolerant than those described above. Since it is relatively expensive to provide substantially anhydrous formaldehyde as a reactant and since water is produced in the condensation reaction, it is desirable to employ a water tolerant catalyst in these reactions. unless otherwise stated, pore volume, surface area and average pore diameter was determined by BET nitrogen adsorption (desorption test).

The general object of this invention is to provide a new method of producing a water tolerant catalyst sui¬ table for the aldol-type condensation of formaldehyde compounds with saturated aliphatic monocarboxylic acid compounds. Other objects appear hereinafter.

The objects of this invention can be attained by cofor ing a catalyst by gelling a silica colloid and a suitable alkali metal or alkaline earth metal compound. The silica supported catalysts of this invention are sub-

stantially more water tolerant than catalysts prepared by the incipient wetness technique described above. This technique has the additional advantage that it is easier to control decoking the catalysts of this invention. There is a substantially smaller exotherm during decoking and accordingly less chance of destroying the catalyst. It is believed that the reduced exotherm is due to the more even distribution of cations in the silica support. Briefly, the silica catalysts of this invention can be prepared by gelling an aqueous composition comprising a silica colloid and alkali metal and/or alkaline earth metal cation, drying the composition to remove substan¬ tially all of the moisture other than the water of hydra- tion, and calcining. The alkali metal and/or alkaline earth metal cations of the catalyst can be used in a concentration of .001 to .2 equivalents of cation per 100 grams of silica support on a dry solids basis. In general, it is preferred to have from about .004 to .1 equivalents (gram atoms) of cation per 100 grams of silica support on a dry solids basis since the higher the concentration of cation, the lower the temperature needed for condensation of the saturated aliphatic monocarboxylic acid compound and for¬ maldehyde compound and the greater the selectivity and life of the catalyst. The lower the concentration of the cation, the higher the condensation temperature necessary to obtain the desired degree of conversion to alpha, beta-ethylenically unsaturated monocarboxylic acid com¬ pound and the lower the selectivity of catalyst and life of the catalyst.

Suitable sources of Group I alkali metal and Group II alkaline earth metal cations include sodium hydroxide, potassium hydroxide, cesium hydroxide, lithium hydroxide, rubidium hydroxide, strontium hydroxide, magnesium hydroxide, lithium phosphate, trisodium phosphate, cesium phosphate, sodium borate, barium hydroxide, sodium carbo¬ nate, cesium fluoride, cesium nitrate, etc. Of these.

the alkali metal cations are preferred and particularly cesium.

While any commercially available colloidal silica can be used, it is preferred to use commercially avail- able colloidal silicas having an average particle diam¬ eter of 40 to 1000A ⁰ , particularly those having a par¬ ticle diameter of about 50 to 250A°. The preferred silica supported catalysts have a surface area of 20 to

275 i 2/g, a pore volume of .1 to .8 cc/g and an average pore diameter of about 75 to 200A°, which are the subject of co-pending S.N. 624,040 filed on even date in the name of Hagen et al, which is hereby incorporated by refer¬ ence. As pointed out in S.N. 624,040, silica catalysts comprising at least one cation of a Group I or Group II

2 metal having a surface area of 20 to 275 m /<_ , pore volume of .1 to .8 cc/g and an average pore diameter of

75 to 200A° have relatively high activity (% conversion and selectivity) and relatively long life. Pore volume, surface area and average pore diameter are interdependent variables. Other things being equal, holding one vari¬ able constant, as the surface area increases, pore volume increases; as the surface area increases, average pore diameter decreases; and as the pore volume increases, average pore diameter increases. It is critical in S.N. 624,040 that the catalyst satisfy each of the pore volume, surface area and average pore diameter require¬ ments. For example, if the catalyst has a porosity greater than .8 cc/g, the catalyst lacks the strength to resist attrition necessary for use over extended periods of time. If the porosity is less than .1 cc/g, the sur¬ face area is too low and/or the average pore diameter is too high. However, the catalyst loses activity as it loses pores having a diameter under 1200A ⁰ . Accordingly, it is preferred to use a catalyst having a substantially smaller average pore diameter to insure that the catalyst has adequate life. If the average pore diameter of the starting catalyst is substantially higher than 200A°,

there is a substantial decrease in the life of the catalyst. Pore diameters of at least 75A° are necessary in order to permit gas diffusion of the reactants and reaction products. In somewhat greater detail, the silica supports of this invention can be prepared by forming an aqueous com¬ position comprising about 10 to 60% by weight colloidal silica on a dry basis and alkali metal and/or alkaline earth metal cation. The colloidal silica is gelled by adjusting the pH to a range of about 3 to 10, preferably about pH 6.0 to about 9.0, preferably with alkali metal or alkaline earth metal cations. Salts such as NH 4.NO3- can be used to accelerate gelation. While silica hydro- gels can be aged for two weeks or more, aging seems to have no effect on the properties of the catalyst and accordingly, aging is not necessary. The composition is then dried by any suitable means, such as in a microwave oven, to constant weight and apparent dryness, e.g., about 4 to 5% moisture on a dry solids basis. Apparently only the water of hydration is retained by the silica gel after drying to constant weight. The silica gel is then calcined at about 300 to 800°C, preferably about 300 to 600°C. As pointed out in S.N. 624,040 calcination tem¬ peratures above about 800°C, there is a tendency for the surface area to go down, the pore volume to go down and the pore diameter to go up with the result that the cata¬ lyst is outside the ranges required in S.N. 624,040. However, these catalysts are substantially more water tolerant than silica catalysts produced by the incipient wetness technique.

The catalysts of this invention can be used for the aldol-type condensation of saturated aliphatic monocar¬ boxylic acid compounds to alpha, beta-ethylenically unsa¬ turated monocarboxylic acid compounds of one more carbon atom than the starting saturated aliphatic carboxylic acid compound. Suitable aliphatic monocarboxylic acid compounds that can be converted in the aldol-type conden-

sation reaction include acetic acid, propionic acid, methyl acetate, methyl propionate, ethyl propionate, ace¬ tonitrile, propionitrile, etc. The preferred saturated monocarboxylic acid compounds are propionic acid com- pounds and particularly propionic acid since the catalyst of this invention has been designed primarily for large scale production of methacrylic acid.

While any suitable source of formaldehyde compound can be used, such as formalin, paraformaldehyde, metha- nolic formaldehyde, trioxane, etc., it is preferred to use substantially anhydrous formaldehyde, particularly cracked monomeric, gaseous, substantially anhydrous for¬ maldehyde.

Briefly, an alpha, beta-ethylenically unsaturated monocarboxylic acid compound can be prepared by con¬ densing under vapor phase conditions a saturated ali¬ phatic monocarboxylic acid compound and formaldehyde com¬ pound in the presence of a coformed silica gel catalyst preferably one comprising at least one cation of Group I or Group II metal and a silica gel support, said support

2 having a surface area of 20 to 275 m /g, a pore volume of

.1 to .8 cc/g and an average pore diameter of 75 to

200A°.

This reaction can be carried out advantageously in the presence of a water-immiscible hydrocarbon or halohy- drocarbon diluent of from about 6 to to 12 carbon atoms, which is the subject of co-pending S.N. 624,050, in the name of Smith filed on even date, which is hereby incor¬ porated by reference. As pointed out in S.N. 624,050, diluent is advantageous in increasing the percent yield by approximately 10% (e.g. from 30 to 33%). Further, as explained below, the diluent has additional functions in the overall unitary process for producing methacrylic acid from propionic acid. Suitable diluents include n-hexane, n-heptane, n-octane, 2-ethylhexane, n-decane, n-dodecane, o,p,m-xylene, benzene, toluene, etc. The concentration of diluent can range from about 10 to 50%

by weight of the reactants in the main reactor.

The molar ratio of monocarboxylic acid compound to formaldehyde can range from 25:1 to 1:25. However, best results with this catalyst in the production of metha- crylic acid can be obtained using a molar ratio of pro¬ pionic acid compound to formaldehyde of from about .5-2.0 to 1. In general, the lower the molar ratio, the higher the percent conversion based upon the amount of propionic acid converted. While the aldol-type condensation can be carried out at about 280 to 500°C, it is preferred to operate at about 280 to 350°C since selectivity goes up as the reac¬ tion temperature goes down. As pointed out in co-pending application S.N. 623,945 filed in the name of Smith et al on even date, which is hereby incorporated by reference, (1) the amount of undesirable unsaturated cyclic ketone by-product which is a catalyst for the polymerization of alpha, beta-ethylenically unsaturated monocarboxylic acids can be reduced from 4 mol % based on starting pro- pionic acid compound at 390°C to approximately 2.5 mol % at 350°C or less (about 1% at 325°C) or (2) over a 80-day period of alternate 24-hour periods of condensation fol¬ lowed by 24-hour periods of decoking that the loss of cation can be reduced from over 75% at 390°C with atten- dant loss of catalytic activity to about 10% at 350°C with constant activity.

In somewhat greater detail, the unitary process for the production of methacrylic acid comprises (1) feeding propionic acid and formaldehyde compound to a reactor containing the silica catalyst of this invention, (2) condensing under vapor phase conditions formaldehyde and propionic acid to produce a composition comprising water, formaldehyde, propionic acid and methacrylic acid, (3) distilling said reaction product to remove water, unreacted formaldehyde and at least some of the propionic acid from the reaction product, (4) passing an entraining agent comprising a water immiscible diluent of from 6 to

12 carbon atoms to the distillation column to remove water and at least some of the formaldehyde overhead.

In a still more preferred version of this process, a side draw is located at least part way up the distilla- tion column to remove a composition comprising substan¬ tially all of the propionic acid and at least some of the formaldehyde. The use of a side draw is disclosed and claimed in co-pending S.N. 624,049 filed in the name of Baleiko et al which is hereby incorporated by reference. As pointed out in S.N. 624,049, the side draw facilitates the removal of part of the formaldehyde from the aqueous mixture going overhead and thereby precludes polymeriza¬ tion of formaldehyde at the top of the distillation column thereby eliminating or reducing the possibility of plugging at the top of the distillation column. Irres¬ pective of whether a side draw is employed or not, it is contemplated that the formaldehyde is recovered from the aqueous formaldehyde taken overhead by reacting the aqueous formaldehyde with an alcohol of from about 6 to 12 carbons to form a hemiacetal, distilling water from the hemiacetal and then cracking the substantially anhyd¬ rous hemiacetal to recover the formaldehyde. The formal¬ dehyde is advantageously separated from the alcohol by adding a water immiscible diluent at the top of the column in order to facilitate the removal of the formal¬ dehyde from the alcohol used to form the hemiacetal. The formaldehyde and diluent are then recycled to the main reactor.

The process of this invention can be carried out at a weight hourly space velocity of about .1 to 10, prefer¬ ably .5 to 6.5. In general, the lower the weight hourly space velocity, the lower the reaction temperature neces¬ sary. The higher the weight hourly space velocity the higher the reaction temperature necessary. The catalysts of this invention are preferably decoked after about 12 to 72 hours on stream, preferably by the method disclosed and claimed in commonly assigned

application S.N. 624,048 in the name of Smith filed on even date, which is hereby incorporated by reference. In order to prevent sintering of the silica and/or loss of cation, dilute oxygen (1 to 5% by volume and 95 to 99% by volume inert gas) is contacted with the catalyst bed at about 450 to 650°F, preferably 450 to 550° F, while holding the exotherm to about 10 to 30°F, increasing the oxygen content incrementally while holding the exotherm to about 10 to 30°F until there is no exotherm with a mixture of 20% oxygen and 80% inert gas, e.g. air, fol¬ lowed by incrementally raising the temperatures by 25 to 75°F and controlling the exotherm to about 10 to 30°F until decoking is completed at about 650 to 800°F.

In the examples that follow, percent conversion, percent yield and percent selectivity are all based on propionic acid (PA) unless otherwise stated.

It should be noted that in Examples II to IX where mini-reactors were used, the reaction temperature had to be sufficiently high to crack trioxane to monomeric for- maldehyde (390°C). Accordingly, it is anticipated that yields and selectivities will be better when operating with monomeric formaldehyde at lower temperatures.

EXAMPLE I This Example illustrates the production of metha- crylic acid in a pilot plant reactor using a coformed cesium phosphate silica gel catalyst having a surface

2 area of 119 /g, porosity of .604 cc/g and an average pore diameter of 168A° containing 1.97 weight percent cesium based on the dry weight of the silica. A slurry of 29 parts by weight paraformaldehyde, 106 parts by weight propionic acid (PA:FA molar "ratio 3:2) and 47 parts by weight heptane was continuously vaporized to thermally decompose the paraformaldehyde to monomeric formaldehyde at 400°F. The composition was then conveyed to a reactor system comprising a 1" outside diameter by .834" inside diameter by 6 ¹ Inconel tube equipped with 0.25" outside diameter thermowell having a 4* long cata-

lyst zone containing 200 grams of catalyst and on each side of the catalyst a 1' zone of Denstone packing. The thermowell was equipped with thermocouples inserted at 6" intervals and electrical heating means were positioned along the reactor. Conversion of propionic acid and for¬ maldehyde to methacrylic acid was carried out for 24 hours while maintaining the pilot plant reactor at 660°F (350°C), 10 psig and a weight hourly space velocity of 1.55. The reaction product was collected in a heat exchanger and condensed. After 24 hours, feed to the reactor was turned off and the reactor temperature was reduced to 550°F. Two percent oxygen in nitrogen was added slowly to the reactor in order to limit the exot¬ herm during decoking to about 20°F. After that exotherm passed, the oxygen content was increased to 10% and after that exotherm was limited to 20°F, the nitrogen-oxygen mixture was replaced with air. After each exotherm passed, the temperature of the reactor was raised by 50°F increments by closely controlling the exotherm until the reactor was at 700°F. This decoking process typically takes 2 to 4 hours. Air was continuously flowed through the reactor at 700°F for a total of 24 hours. The air was turned off, the reactor temperature was reduced to 660°F and the condensation of propionic acid and formal- dehyde was begun. The sequential condensation and decoking operation was carried out for 80 days of which 40 days was methacrylic acid production and 40 days decoking. The molar ratio of PA:FA was varied from 1.5:1 to 1.34:1. The physical properties of the catalyst before and after the 80 days onstrea is set forth below in Table I. The yields after the first day and the average for the first 66 days also are set forth below in Table I.

Table 1

Initial Catalyst and Final First Day Analysis of Catalyst Products and Avg. Analysis of

Products

Cesium Content 1.97 wt. % 1.76 wt. % Surface Area Pore Volume .604 cc/g .570 cc/g

Average Pore Diameter 168A ⁰ 302A° Molar Ratio of PA:FA 1.5 1.34 Conversion Based on PA 32.1% 32.1% Selectivity Based on PA 91.3% 85.9% MA/PA Yield 29.3% 27.6%

Conversion Based on FA 51.9% 50.0% Selectivity Based on FA 84.9% 74.2% MA/FA Yield 44.1% 37.1%

The catalyst employed in this Example was prepared by intensely stirring a solution of 10302.9 grams Nal- coag, 1034-A colloidal silica (34% solids, 200A° particle diameter) and a solution of 111.66 grams cesium phosphate (having an average of 5 molecules of water per mol) in. 500 cc deionized water. After intense stirring for 10 minutes, a solution of 100 grams ammonium nitrate in 150 grams deionized water was added to the sol and the mix¬ ture was stirred for 2 minutes at which time it began to thicken. The silica gel was permitted to harden after standing at room temperature overnight. The gel was dried in a microwave oven to constant weight, sized to 20 to 40 mesh and calcined according to the following incre¬ ments: 2 hours at 165°C, followed by increasing the tem¬ perature gradually to 540°C over 4 hours and then main¬ taining at 540°C for an additional 8 hours. All of the steps were carried out in flowing air.

Substantially the same results can be obtained by replacing the cesium phosphate with 2% by weight cesium

as the hydroxide or carbonate.

EXAMPLE II This Example illustrates that coformed catalysts are more water tolerant than catalysts prepared by the inci- pient wetness technique. Each of these catalyst runs was carried out in a vertical laboratory 14 to 18" long quartz mini-reactor having an inside diameter of about 1/2" equipped with a thermowell having a diameter of about 3/16 to 1/4". Each of the reactors contained a quartz spun plug to support approximately a 2 1/2 to 3" bed of catalyst weighing approximately 2 to 3 grams. The mini-reactor was enclosed in an electrical furnace main¬ tained at about 390°C. A liquid feed comprising either anhydrous 3:2 molar ratio of propionic acid to trioxane or aqueous mixture comprising 3:2 molar ratio of pro¬ pionic acid and formalin (23% by weight water in the feed) was dripped into the vertical reactor and vaporized therein. In order to achieve comparable initial conver¬ sion, the weight hourly space velocities were adjusted as needed prior to the start of the run. Sparge samples were taken after every 24 hours on feed and the catalyst was decoked for 23 hours at 390°C after every 48 hours on feed. The coformed catalyst used with the anhydrous feed was the catalyst prepared in Example I of this applica- tion and the weight hourly space velocity was 2.3. The catalyst for the aqueous coformed catalyst feed contained 1.6% by weight cesium as cesium phosphate was prepared in the same manner as the catalyst in Example I. The weight hourly space velocity of this catalyst was 1.6. The catalyst prepared by the incipient wetness technique con¬ tained 2.1% by weight cesium as cesium phosphate on silica gel support prepared as follows. A sample of Ludox AS-40 Brand colloidal silica, 40 wt. % silica, was treated by the dropwise addition of concentrated nitric acid until the pH changed from about 10.5 to 3.0. The pH was then raised to about 6.0 by the dropwise addition of concentrated ammonium hydroxide. The mixture was stirred

for 8 hours, at which time a thick gel had formed. The gel was dried overnight at 120°C, then crushed and sized to 18-40 mesh. The material was washed three times to remove sodium by being submerged in 0.10 N nitric acid for fifteen minutes at 75°C. It was then washed five times with deionized water at 50°C and dried overnight in an oven at 120°C. The incipient wetness technique was used to impregnate the dried silica support with an aqueous solution of cesium phosphate. The catalyst was then dried overnight at 120°C and calcined as described in Example I. The resulting catalyst had a BET surface

2 area of 96 m /g, pore volume of 0.4971 cc/g, and average pore diameter of 180 A°. The weight hourly space velocity was 1.6 for the anhydrous feed and .99 for the aqueous feed. The results are set forth below in terms of grams of propionic acid contacted per gram of catalyst on a dry solids basis.

Table 2 A Coformed Catalyst Incipient Wetness Catalyst

Feed Anhydrous Aqueous Anhydrous Aqueous

g PA/g Con- Yield Con- Yield Con- Yield Con- Yield Catalyst ver- ver- ver- ver¬ sion sion sion sion

10-20 37 22 35 16

40-60 37 22 37 21

105 35 23

140-160 35 21 40 22 37 22 22| 11

185 37| 22 18 10

195 39 21

210 34 20

225 11 10

240-250 37 19 37 21

290 28 17

310-330 25 18 28 14 34 22

The change in the BET pore volume distributions for the used coformed and incipient wetness catalysts are summarized in Table 2 B. This example illustrates that

coformed catalysts are substantially more water tolerant than catalysts prepared by the incipient wetness tech¬ nique.

Table 2 B

Coformed Incipient Wetness Catalyst

Anhydrous Aqueous Anhydrous Aqueous

Surface Area (m ² /g) 67 57 66 16

Pore

Vol. (cc/g) 0.6173 0.5501 0.5048 0.0729

Avg. Pore Diameter (A) 262 310 228 304 EXAMPLE III

This Example illustrates the preparation and use of other coformed silica gel catalysts. Unless indicated otherwise, the catalysts were prepared in the manner described in Example I and tested in a mini-reactor in the manner described in Example II for the anhydrous feed. The pore size, pore volume and surface area were similar to those of the catalysts of Example I when all of the cation material set forth in Table 3 was water soluble.

Table 3

Description %C %Y %S WHSV

2.1% Cs as Cs ₃ P0 ₄ 41 29 70 2.5

1.7% Cs as Cs-ZrO- 43 27 62 2.5

1.7% Cs as Cs-TiO., 48 27 55 2.5

2. 2% La as La (NO ₃ ) ₃ 34 15 43 2.6

0.73% La as La(NO ₃ )- + 2.1% Cs as CsF 42 24 56 2.6

0.73% La as La(NO ₃ ) + 2.1% Cs as CsOH 42 24 58 2.6

1.0% Zn as Zn(NO ₃ ) ₂ 28 18 65 2.5

0.51% K as K4.P2.O-7 31 24 77 2.0

1.7% Cs as Cs,CO ₃ 42 28 66 2.5

C = Conversion Y = Yield S = Selectivity

WHSV = Weight Hourly Space Velocity. _.

EXAMPLE IV Example III was repeated except that samples were taken for analysis 24 hours after completion of the first decoking step.

Table 4

Description %C %Y %S WHSV

1.7% Cs as CsF 36 23 63 2.76

1.7% Cs as CsNO. 40 25 62 £, O18

1»7% Cs as Cs ₂ SO ₄ 36 21 58 2.29

1.7% Cs as CsOH 38 24 64 2.12

1.7% Cs as CsOH 40 26 64 2.54

1.6% Cs as Cs.CO ₃ 36 24 67 2.14

2.0% Cs as Cs3.PO4. 35 22 61 2.26

2.0% Cs as Cs ₃ PO, 37 23 62 2.13

2.0% Cs as Cs 3-PO4. 36 22 59 2.02

1,6% Cs as Cs ₃ PO ₄ 40 26 65 2.57

10% Ba as Ba ₃ (PO ₄ ) ₂ 39 24 62 1.50

10% Li as Li ₃ PO ₄ 36 22 61 1.17

1.7% Cs as CsOAC 40 25 64 2.54

C = Conversion Y = Yield S = Selectivity

WHSV = Weight Hourly Space Velocity

EXAMPLE V This Example sets forth the physical properties of coformed cesium catalysts and coformed silica gel cata¬ lysts having cation material of the preceding two exam¬ ples.

Table 5

Description Surface Pore Pore

Area In Volume Radius

1.3% Cs as Cs2-SO4. 164 0.5195 56

1.9% Cs as CsNO- 161 0.5041 56

0.78% Cs as CsNO. 140 0.5183 70

1.6% Cs as CsOAC 161 0.6323 70

1.6% Cs as Cs ₂ C0 ₃ 137 0.7338 93

1.7% Cs as CsOH b 135 0.6695 91

1.6% Cs as Cs ₃ P0 ₄ 136 0.584 77

2.0% Cs as Cs ₃ P0 ₄ 119 0.604 84

1.0% Cs as CsF 117 0.6893 102

2 a - After 38 days on stream surface area was 119 m /g,

0.755 cc/g and 97A° average pore radius

2 b - After 13 days on stream surface area was 117 m /g,

.683 cc/g and 97A° average pore radius. EXAMPLE VI

This Example illustrates the preparation of a silica gel catalyst from a colloidal sol having 50A° diameter particles. The coformed silica gel catalyst was prepared in the same manner as in Example I except that Nalco 2326 containing 50A° diameter particles and 14.5% by weight solids was used in place of the Nalco 1034A. Gelation was carried out by adjusting the pH to 9 with nitric acid. After calcination the catalyst had a BET surface

2 area of 275 m /g, a pore volume of .753 cc/g and an average pore diameter of 78A°. The catalyst was run in a mini-reactor under the conditions set forth in Example II for the anhydrous feed. After .9 days on stream at 2.06 weight hourly space velocity, the yield was 21%, percent propionic acid conversion 44% and percent propionic acid selectivity 48%. After 3.1 days on stream and one regen¬ eration of the catalyst after 48 hours on stream, the percent yield was 19% , percent propionic acid conversion 40% and percent propionic acid selectivity 48%.

EXAMPLE VII The catalyst preparation of Example I was repeated using (1) PQ 2034DI and (2) PQ 2040NH ₄ , in place of the

Nalco 1034A. The silica gel catalysts had, respectively, (1) a BET surface area of 137 m 2/g, a pore volume of .602 cc/g and an average pore diameter of 148A° and (2) 146 2 m /g BET surface area, a pore volume of .621 cc/g and 142

A° diameter. Each of these catalysts were used in the mini-reactor under the conditions set forth in Example II for anhydrous feed. Percent yield was approximately 23%, percent conversion about 41% and percent selectivity about 56% after one and three days on stream.

EXAMPLE VIII This Example illustrates the preparation of coformed catalyst from a silica colloid having a diameter of 1000A°. The preparative process of Example I was repeated using PQ sol 9950 which yielded after calcina-

2 tion a catalyst having a surface area of 51 m /g, pore volume .417 cc/g, and an average pore diameter of 340A°. This catalyst was then run in a mini-reactor in the manner described in Example II. After one day on stream the propionic acid base yield was 22%, propionic acid base conversion 36%, and propionic acid base selectivity was 62%. After three days on stream the yield dropped to 16%, conversion to 30% and selectivity to 55%.

EXAMPLE IX This Example illustrates variations in physical pro¬ perties of silica gel catalysts due to variations in cal¬ cination temperature. The catalyst preparation method of Example I was repeated except that the calcination tem¬ perature was incrementally raised to the temperatures set forth below in Table 6 and held at said temperature for eight hours. The catalysts were then tested in a mini- reactor in the manner described in Example II.

Table 6

Calcination

Temp. 540 ^α C 700°C 800°C 850°C 900°C

Surface Area 115 57 20 10 1.4 in m2/g m ² /g m ² /g m ² /g m ² /g ² /g

APD in A° 174 A ⁰ 276 A° 890A° >900A° >900°A

PV in cc/g .58 .55 .32 .05 .02 cc/g cc/g - cc/g cc/g cc/g

% PA Selectivity 62% 56% 65% 61% 36%

% PA Conversion 33% 32% 31% 27% 25%

% Yield on PA 22% 18% 21% 16% <10%

The table clearly shows that as calcination tempera¬ ture increases, the surface area of the catalyst goes down, average pore diameter goes up and pore volume goes

2 down. Further, as the surface area goes below 20 m /g and the average pore diameter increases, the percent yield based on propionic acid decreases.