Treatments to Reduce Aldol Condensation and Subsequent Polymerization in Diethanolamine Scrubbers

The present invention relates to the use of agents to reduce fouling caused by aldol condensation in diethanolamine (DEA) scrubbers. Suitable inhibiting agents include: hydrogen peroxide; hydroperoxides; nitroalkanes; salts of hypochlorous acid (commercial bleach, NaOCl and CaOCl); alkyl substituted aldehyde oximes; trialkyl phosphites; isothiuronium salts; 2-mercaptoethanol (C.sub.2 H.sub.6 OS); and, benzoic hydrazide. A preferred inhibiting agent for diethanolamine scrubbers is an excess of hydrogen peroxide.

Refineries employ atmospheric and vacuum distillation towers to separate crude oil into narrower boiling fractions. These fractions then are converted into fuel products, such as motor gasoline, distillate fuels (diesel and heating oils), and bunker (residual) fuel oils. Some of the low boiling fractions from various units of the refinery are directed to petrochemical plants, where they are further processed into highly refined chemical feedstocks to be used as raw materials in the manufacture of other types of products, such as plastics and basic chemicals.

During the refining of petroleum, hydrocarbon streams typically are treated in an amine scrubber, such as a DEA or MEA scrubber, to remove acid gases, such as hydrogen sulfide and carbon dioxide. In a petrochemical plant, hydrocarbon streams typically are treated for the same purpose in a caustic scrubber. All of these scrubbers herein are referred to as "acid gas" scrubbers. The hydrocarbon stream entering an acid gas scrubber may contain aldehydes and ketones, their precursors, such as vinyl acetate, or other impurities, that are hydrolyzed or otherwise converted to aldehydes and salts of organic acids in the highly alkaline environment of an acid gas scrubber. Such compounds will herein be referred to as "reactive compounds." These reactive compounds either (a) contain carbonyls, or (b) form carbonyls under highly alkaline conditions that are susceptible to classic aldol condensation reactions. Carbonyls that are susceptible to classic aldol condensation reactions hereinafter will be referred to as "reactive carbonyls."

Under highly alkaline conditions, lower molecular weight aldehydes, such as propionaldehyde (propanal) and especially acetaldehyde (ethanal), readily undergo base catalyzed aldol condensation at ambient temperatures. The result is the formation of oligomers and polymers which precipitate out of the scrubbing solution as viscous oils, polymeric gums, and solids. These precipitates can foul the processing equipment and result in the reduction of processing throughput and costly equipment maintenance or repair.

In the past, organic reducing agents or organic and inorganic oxidizing agents have been proposed to prevent such polymerization. These organic agents might successfully retard polymerization in acid gas scrubbers; however, the organic agents also tend to undergo other reactions which can reduce their effectiveness as aldol condensation inhibitors.

Effective and economical methods for retarding aldol condensation in diethanolamine scrubbers would be highly desirable.

The present invention provides a number of inhibiting agents that are useful to reduce aldol condensation in diethanolamine scrubbers. Suitable inhibiting agents include: hydrogen peroxide; hydroperoxides; nitroalkanes; salts of hypochlorous acid (commercial bleach, NaOCl and CaOCl); alkyl substituted aldehyde oximes; trialkyl phosphites; isothiuronium salts; 2-mercaptoethanol (C.sub.2 H.sub.6 OS); and, benzoic hydrazide. A preferred inhibiting agent for diethanolamine scrubbers is an excess of hydrogen peroxide.

The present invention is directed to reactions that cause fouling in diethanolamine scrubbers. Without limiting the present invention, it is believed that the red precipitate that forms in acid gas scrubbers is the result of several aldol condensation/dehydration steps. As used herein, the term "aldol condensation" is intended to refer to the reactions that ultimately result in the formation of a precipitate in acid gas scrubbers. The inhibiting agents of the present invention are believed to inhibit fouling by inhibiting such aldol condensation.

Hydrogen peroxide is a most preferred inhibiting agent for use in DEA scrubbers.

Other suitable agents for inhibiting aldol condensation in diethanolamine scrubbers include: hydroperoxides; nitroalkanes; salts of hypochlorous acid; alkyl substituted aldehyde oximes; and trialkyl phosphites.

Agents that are less preferred, because they are not as soluble in a diethanolamine scrubber solution, but that should inhibit aldol condensation in diethanolamine scrubbers are isothiuronium salts and 2-mercaptoethanol (C.sub.2 H.sub.6 OS).

Preferred hydroperoxides are tertiary-butyl hydroperoxide and cumene hydroperoxide. Tertiary-butyl hydroperoxide is commercially available from Arco Chemical Co., Chemicals Group, Newtown Square, Pa. Cumene hydroperoxide is commercially available from Elf Atochem North America, Inc., 2000 Market Street, Philadelphia, Pa.

Preferred nitroalkanes having the structure R--NO.sub.2, wherein R is selected from the group consisting of straight, branched, and cyclic alkyl groups having between about 1-8 carbon atoms. A preferred nitroalkane is nitromethane, which is commercially available from Angus Chemical Co., 1500 East Lake Cook Road, Buffalo Grove, Ill.

Salts of hypochlorous acid are commonly available commercially as a 3.62% solution of NaOCl (household bleach) and in solid form as Ca(OCl).sub.2. An example of a commercial source is Mayo Chemical Co., Smyrna, Ga.

Preferred alkyl substituted aldehyde oximes are those in which the alkyl groups are selected from the group of straight, branched, and cyclic alkyl groups having between about 1-8 carbon atoms. A most preferred alkyl substituted aldehyde oxime is acetaldehyde oxime, which is commercially available from AlliedSignal, Morristown, N.J.

Preferred trialkyl phosphites have the general structure (RO).sub.3 P, in which R is a straight, branched, or cyclic alkyl group having between about 1-8 carbon atoms. A most preferred trialkyl phosphite is trimethyl phosphite, which is commercially available from Akzo, Nobel Chemicals, Inc., 300 Riverside Plaza, Chicago, Ill. 60606.

A preferred isothiuronium salt is benzyl isothiuronium chloride, which may be manufactured as shown in Example 1. Persons of skill in the art can alter the starting materials in the procedures of Example 1 to manufacture other isothiuronium salts.

Preferably, the inhibiting agent should be injected into the scrubber slowly, on an "as-needed" basis. Due to substantially continuous mode of operation of most scrubbers, it is believed that the agent should reach a steady state during processing. An equimolar ratio of inhibiting agent to active carbonyl containing compound should be sufficient to inhibit aldol condensation. Even less than a 1:1 ratio may be sufficient. An excess of inhibiting agent may be added, if desirable.

The agents of the present invention will react with the reactive carbonyls, or a condensation product of two or more reactive carbonyls, in the acid gas scrubber at ambient temperatures. Therefore, the hydrocarbon stream need not be heated. In the following examples, the samples are heated to between about 50.degree.-55.degree. C. (122.degree.-131.degree. F.) to hasten the polymerization reaction for facile candidate evaluation. In actual use, heating is neither necessary nor advisable.

The invention will be more readily understood with reference to the following examples.

Thiourea (7.60 g; 100 mmoles) was added to 150 ml of ethanol in a 250 ml Erlenmeyer flask. The mixture was hearted to reflux to dissolve almost all of the thiourea. After cooling the contents to 40.degree. C., benzyl chloride (12.6 g; 100 mmoles) was added dropwise over three minutes. The slightly hazy solution was refluxed overnight, causing it to become clear. In the morning, the ethanol was stripped off on the rotary evaporator leaving a good yield of a white solid. NMR analysis of the white solid showed it to be benzyl-isothiuronium chloride.

A number of candidates were screened for use in DEA (diethanolamine) scrubbers. To perform the screening, 2 oz. bottles were dosed with 32,000 ppm of candidate. 25 ml of 20% DEA was added to each bottle, and the bottles were shaken. Subsequently, 300 .mu.l of vinyl acetate (4,000 ppm) was added to each bottle, and the bottles were shaken, placed in an oven at 55.degree. C. (131.degree. F.), and observed as a function of time for at least 24 hours for color and precipitate formation. Candidates that exhibited no color, or only slight color, and no precipitation were selected for further study. Approximately 24 candidates were selected for further testing.

Three more candidates were screened in this experiment. Two oz. bottles were dosed with 1.5 g of candidate (except for candidate 2) followed by 25 ml of 20% DEA and then 300 .mu.l of vinyl acetate. The samples were shaken and placed in an oven at 55.degree. C. (131.degree. F.) and observed at one minute and 18 hours. The source and amount of the candidate are shown in Table I:

The results are shown in Table II:

Based on the foregoing, candidates 1 (hydrogen peroxide) and 2 (NaOCl) are preferred agents for inhibiting aldol condensation in DEA scrubbers because the solution has only minimal color and no haze or precipitate is present. Candidate 3 (iodine) appears to be a suitable agent for inhibiting aldol condensation in DEA scrubbers; however, a large amount of presumably iodoform was formed in this sample. Potential users should be informed of the potential for formation of this precipitate.

Ten of the candidates from Example 2 were tested under stoichiometric conditions in this experiment. Two oz. bottles were dosed with 3.56 mmoles of candidate (except for candidates 3 and 4, which inadvertently were misdosed), 25 ml of 20% DEA, and 300 .mu.l of vinyl acetate (280 mg; 3.24 mmoles). The samples were shaken and placed in an oven at 50.degree. C. (122.degree. F.) and observed after 24 hours. The amount of candidate added is shown in Table III:

The results are shown in Table IV:

Based on the foregoing, candidates 3 (t-butyl hydroperoxide) and 8 (acetaldehyde oxime) appear to be suitable agents for inhibiting aldol condensation in DEA scrubbers, and candidates 1, 2, 4, and 7--benzyl isothiuronium chloride, 2-mercaptoethanol, t-butyl perbenzoate, and benzoic hydrazide, respectively--may be suitable agents for DEA scrubbers provided that the presence of some precipitate in the scrubber is acceptable, or the candidate is dosed at a lower soluble level.

Ten of the candidates from Example 2 were tested under stoichiometric conditions in this experiment. Two oz. bottles were dosed with 3.56 mmoles of candidate, 25 ml of 20% DEA, and 300 .mu.l of vinyl acetate (280 mg; 3.24 mmoles). The samples were shaken and placed in an oven at 50.degree. C. (122.degree. F.) and observed after 24 hours. The amount of candidate added is shown in Table V:

The results are shown in Table VI:

Based on the foregoing, candidates 5 (nitromethane), 6 (trimethyl phosphite), and 10 (NaOCl) appear to be a suitable agents for inhibiting aldol condensation in DEA scrubbers.

The successful candidates from the DEA testing were evaluated for their ability to perform in a mixture of "lean" and "fat" DEA from a potential customer's DEA scrubber. 600 ml of "lean" DEA and 200 ml of "fat" DEA from a customer were vacuum filtered, mixed, and then filtered again through #1 paper. Two oz. bottles were dosed with 3.56 mmoles of candidate, to result in 1.1 moles of candidate per 1.0 mole of vinyl acetate, except for candidate 1 which was misdosed. 25 ml of haze-free "fat/lean" DEA was added to each bottle, and the bottles were shaken 100 times. Each bottle then was dosed with 300 .mu.l of vinyl acetate (280 mg; 3.24 mmoles), and the bottles were shaken 50 times. The bottles were then placed in a 50.degree. C. (122.degree. F.) oven and observed after 24 hours. The amount of candidate added is shown in Table VII:

At time zero, all samples were hazy because it was impossible to remove all of the haze by filtration. The results are shown in Table VIII:

Based on the foregoing, candidates 1 (t-butyl hydroperoxide), 2 (acetaldehyde oxime), and 3 (nitromethane) appear to be a suitable agents for inhibiting aldol condensation in DEA scrubbers. Candidate 4 (trimethyl phosphite) produced some red precipitate (but less than that produced in the blank), produced a lighter color solution, and may be a suitable agent for some DEA scrubbers.

Persons of skill in the art will appreciate that many modifications may be made to the embodiments described herein without departing from the spirit of the present invention. Accordingly, the embodiments described herein are illustrative only and are not intended to limit the scope of the present invention.