WO1997002223A2

WO1997002223A2 - Method for producing ethylene and other chemicals

Info

Publication number: WO1997002223A2
Application number: PCT/US1996/011217
Authority: WO
Inventors: Vitaly Lissianski; William Gardiner
Original assignee: Vitaly Lissianski; William Gardiner
Priority date: 1995-06-30
Filing date: 1996-06-29
Publication date: 1997-01-23
Also published as: WO1997002223A3

Abstract

A synergistic combination of two or more chain reaction promoters. Each promoter is chosen so as to modify the rate of a dynamically different component of the overall chain process on an appropriate time scale such that in combination the two or more promoters achieve improvement of the process with respect to features such as reaction yield, energy requirement, economics, safety, and environmental burden. An embodiment described in quantitative detail is enhancement of the process of pyrolysis of ethane to ethylene by the chain initiating promoter hydrogen peroxide in combination with the chain branching promoter oxygen.

Description

METHODFORPRODUCINGETHYLENEANDOTHERCHEMICALS

Background-Fieldof Invention

This invention relates to an improvedmethod for pyrolysis-based synthesis of valuable chemicals, including olefins, fromhydrocarbon or other feedstocks, specifically bymeans of controlled introduction of suitable homogeneous promoters in such a manner that chain reaction initiation, branching, andpropagation are regulated, togetherwith flowrates and temperature, so as to enhance the desiredconversion while suppressing side reactions that lead to undesiredbyproducts.

Background-Descriptionof PriorArt

Abroadvariety of process engineeringmethods is knownto theprior art of thepetrochemical industry. To take as a specific example themanufacture of ethyleneby heatingvarious feedstocks, also known as "cracking" or "pyrolysis" , theprior art recognizes many ways toprepare andheat the feedstock, togetherwithothergases or liquids that may be added, to a selectedreaction temperature, maintain that temperature for a suitable reactiontime, quenchthe hot gas to a temperature lowenough to halt thepyrolysis process, andthen separate the outflowinto its components. To design and operate an ethyleneplant one has manywell recognizedresources available to achieve optimal results in terms of plant cost and operating economics. Options for feedstockmixing, conservationof process heat with exchangers of various kinds, variable area ducts, recyclingpumps, measurement andcontrol of parameters, and awide range of separation technologies are known to engineeringpractice and are described in technical articles, monographs, and textbooks too numerous tomention.

The essence of the inventiondescribedherein lies in our discovery that there are fundamental chemical principles relatedto the dynamic behavior of chain reactions that canbe, but heretofore have not been, brought tobear on the design and operation of industrial pyrolytic processes formakingmorevaluable substances, such as olefins or acetylenic or aryl compounds, for example, from less valuable ones such as feedstocks derived fromnatural gas, petroleum, or industrial byproduct streams. Theseprinciples couldnot be applieduntil now because, as we have learned and are disclosing indetail herein, in order to discover howto use them in a beneficialmanner onemust take advantage of very recent gains in quantitative knowledge of the way that theprinciples of chain reactiondynamics are expressed in specificmolecular-level chemical processes. In this disclosure we explain these in terms of reactions that governpyrolysis-based petrochemistry.

Prior ar : Chain reaction theory.

The dynamic theory of chain reactions, developedstarting in the 1920s by Christiansen, Herzfeld, Semenov, andothers, takes as its startingpoint the notion of "chain center" , usually an atom or free radical R generated in somemanner, for example by thermal dissociation of a relatively weak chemical bondholding radicals and R' together

R-R' —-> R + R',

inwhat is called a "chain initiation reaction" . [Gardiner 1969; Kondratiev andNikitin 1974] (Hereinwill not need to distinguish carefully between reactions that are thought to be real molecular- level events, the so-called "elementary reactions" , and schematic ones that only illustrateprinciples or describe an overall effect of many molecular-level events . An example in the former categorywould be the decomposition of acetone upon heating into methyl and acetyl radicals according to

CH3COCH3 -→ CH3 + COCH3,

which can be effected eitherphotochemically or by heating. This distinction, well recognized in the science of chemical kinetics, may indeed become important when specific substances are considered for specific applications according to the discovery disclosed herein. ) A variety of fates may await the chain centers R andR' . One is disappearance fromthe chemical scene in "chain termination reactions" , whichmay result for example fromdiffusive loss to vessel walls ("heterogeneous loss") or from "recombination reactions" that remove two chain centers at once, such as recombination of methyl radicals to formethane

CH_{3 +}CH₃→C₂H₆.

Another fate is decomposition of one chain centermolecule to form another chain center molecule or atomand at the same time a stable molecule; the acetyl radical, for example, readily decomposes to the stable carbonmonoxide and the reactivemethyl radical

COCH3 — r CH₃ + CO

in an example of what is called a "chainpropagation reaction" , one which continues toprovide a chain center but of different chemical identity. Chainpropagation reactions can also involve a chain center anda stable species; methyl radical, for example, can react with ethane to transfer the chain center to an ethyl radical CH₃ +C₂H₆ -→CH₄ +C₂H₅.

A special fate of chain centers is the "chainbranchingreaction", bymeans of whichthenumber of chaincenters increases. One of the important reactions of methyl radicals in flames, for example, is with molecular oxygen to formoxygen atoms and themethoxy radical, bothof whichare then independent chain centers in the flame, i.e. , by

CH₂ +O2 →CH₃0 + 0.

This foregoing characterizationof chain reaction types gives the underlying concepts that compose the known art of analyzing complex chain reactions in terms of simple "steps" havingparticular types of chain character. What one does with these concepts to explain a natural chain reaction, such as the stratospheric chemistry of ozone or the formation of lignin inwood, or to control an industrial one, such as a free- radical typepolymerization, may vary greatly depending on one's purpose.

Promoters : The above considerations suggest that chain reactions maybe significantly affectedby changing the rates of chain center production anddestruction, for exampleby changing the temperature orby adding substances to the reacting system. The influence of an additive on aparticular chainreactionprocessmust dependof course on the specificnature of theprocess andthe additive. From the above discussion, however, it is clear that two components of a chain reactionmay beparticularly attractive targets for additives: chain initiation and chainbranching. Additives that enhance radical production in either of these steps may "promote" reaction, such that at a giventemperature the reaction runs faster.andproducts areproduced sooner. We also recognize that enhancement of chain branchingmaygive a different kindof advantagethanenhancingchain initiation: The latterproduces chain centers at a linear rate, while the formerdoes so exponentially. Sucheffects havebeennotedbefore in connectionwith sensitizationof fuel ignition, for exampleby Borisov et al. (1987) and Zamansky andBorisov (1992) .

Prior art : Computer modeling of complex chemical dynamics .

In the early history of chemical kinetics, insight into thenature of chain reactions was gainedby searching for the consequences of assuming that the concentrations of chain centers stabilize at "steady state" values. This prior art is still valuable today for analyzing all sorts of chemical processes where steady state assumptionspermit one tobetterunderstandandcontrol whatever underlying chemistry pertains. For thepurpose of describingour invention, however, there is amoremodernprior art that canbe exploitedtobetter advantage. This is the description of chain reactions in themathematical form known as an "initial valueproblem" .

What one does is transcribe the chemical descriptionof theprocess into a systemof ordinary differential equations, eachdescribing the dependence of the time rate of change of the concentrationof one of the chemical species participating in theprocess uponthe current concentrations of all of them. Tomake use of this transcriptionone adds an "initialization" statement of the startingvalues of all concentrations (in time, for chemical processes that proceed in a batchmode, or at a startingpoint in space for flowprocesses) and a "constraint" statement of thephysical conditions that pertain: Reaction at constant volume, constant pressure, or at a steadymass flowrate; at constant temperature orwithadiabaticwalls; andso on until thephysics of theproblem is defined. All of thismathematics together has come to be calleda "computermodel" of the chemical process.

In order tomakepractical use of computermodels, values of all of theparameters appearing in the equationsmust be given. These may include large numbers of thermochemical and transport property values as well as elementary or composite chemical reaction rate coefficients. Wewill show later that themodelingresourceswhich have become available in recent years can givevery valuable insight, andhave led to the discovery that is disclosedhere.

Methods of computermodeling for chemical purposes havebeen described extensively in the archival and report literature, for example by Lambert (1973), andRadhakrishnan andHindmarsh (1993). The discovery disclosedhereindoes not dependonmakinguse of any particular computermodelingmethodormethods; it wouldemerge inthe same formregardless of whichmodelingprocedures wouldbe applied. Similarly, the discovery disclosedhereindoes not depend onusingany particular source of chemical data for setting themodel parameters, as long as one adopts valid rate coefficient andthermochemical datavalues. The set used for the computations forwhichresults are presentedhere are those givenby Frenklachet al. (1994) .

Prior ar : Pyrolytic petrochemistry .

The earliest processingmethods forpetroleumandnatural gas were physical rather than chemical procedures, primarily distillations andother separations developed tomaximize themarketability of the available resources byproviding, for example, essentially odor-free andsmoke-free kerosene for illumination. In ongoing competitionwithother sources of rawmaterials, especially coal and wooddistillates, scientists and inventors steadily expandedthe variety of useful products and improved the technology formaking them. The more favorable economics of production andprocessing eventually led todominance of petroleumandnatural gas as industrial feedstocks for synthesis purposes; over the course of the 20th century further development of the technology of exploiting these rawmaterials continuedunabated. It resulted in the sophisticated chemical industrywe knowtoday, with its myriadproducts entering theplastics, textile, rubber, packaging andmany other downstream industries.

Along theway it became generally recognizedthat one of themost useful chemical reactions forprocessing the components foundin petroleum andnatural gas is simplyheatingthemto appropriate temperatures for appropriate times, a reaction known as pyrolysis. Pyrolysis reactions, as all other chemical changes, must be thermochemically favored in order to occur. The essential principles involvedcanbe seen in thepyrolytic decomposition of ethane, a relatively abundant component of natural gas, to ethylene and hydrogen, present in only small relative amounts in natural gas,

C₂H₆ →C₂H₄ + H₂.

This reactionhas a standardGibbs function change of +101 kJ/mol at 298 K, (77 °F) indicating that dehydrogenation to formethylene is disfavored at ambient temperature, i.e. , onlyminute traces of ethylene andhydrogenwouldbe formed fromdecompositionof ethane. [The fundamental physical- chemical principles that connect the signs andvalues of thermochemical quantitites to the feasibility of corresponding chemical reactions arewell recognized in theprior art; see for exampleAtkins 1990. ] Because of thepositive standard entropy change of reaction, however, (+121 J/mol K at 298 K) the reaction is more andmore favoredat higher temperatures. But if the ethanepyrolysis temperature is chosen too high, pyrolysismay continuewell beyond ethyleneproduction andbecome quiteunspecific, i.e., many chemical reactions besides the desiredone begin to be thermochemically significant. The thermochemical analysis also shows that ethanepyrolysis to ethylene andhydrogen is quite endothermic and is therefore self-limitingunless heat is suppliedto sustain the reaction: For the temperature range of interest for ethanepyrolysis to ethylene, 143 kJ of heat must be added to the system for eachmole of ethanepyrolyzedto ethylene.

Inparallel with these thermal considerations is one of process pressure. Because twomoles of products are formed for eachmole of ethanepyrolyzed, Lechatlier's principle implies that relatively moreproduct will bepresent at equilibriumwhen the total pressure is reduced. Again a compromisemust be found to achieve optimal results: If thepyrolysis reaction is run at too high apressure, themaximum attainable conversionpermole of ethane is too low, while if it is run at too lowapressure, the total rate of processing ethane is overly diminished.

Aside fromconsiderations of maximum theoretically attainable conversion one has to take account of the rates of pyrolysis reactions as well. Reaction rates increasewith reactant concentrations, with temperature, and in thepresence of catalysts. Compositions and temperature ranges that may appear to be optimal fromequilibrium calculations usually have tobemodified in order to adjust the rate of the conversionprocess. Furthermore, because of the endothermicity of the reaction, the rate at whichtheneeded heat canbe supplied can also be a limiting factor.

These considerations underlie a great deal of effort to improve the art of pyrolyticprocesses in the petrochemical industry: Onemust devisemeans to adjust the temperatureneeded to cause a desired reaction to occur to sufficient extent while suppressing the concurrent production of undesiredproductswasteful of the feedstock resource, whichmay be costly to separate fromthe desiredproduct and are of en environmentally hazardous as well. In the following paragraphs theprior art known to accomplishthis end is described. Primary focus is given to the just- citeddehydrogenationof ethane to ethylene. In theUnitedStates ethane is themost favoredfeedstock for ethyleneproduction, as its use lowers boththe cost ofproduction andcapital investment. Becausemethane ismore abundant thanethane innatural gas, methods forproducingethylene from it havebeen proposed, but none are in current usebecause of thehightemperatures and lowselectivity involved methane is more expensive to convert and producesmore undesiredbyproducts.

Prior art : Industrial practice.

The chain-reaction chemistry of pyrolyticpetrochemistry is well known andhas been described in detail, for example by Dente et al. [1983] . Throughyears of operatingexperience, thepractical effects of changing operating conditions onproduct distributions havebeen discovered. For ethyleneproduction, the temperatureof hydrocarbon feedstock entering the "cracking chamber" of an industrial furnace typically varies from750 to 1000 K, (890 to 1340 °F) dependingon the feedstockused, and the outlet temperature is maintainedbetween 1050 and 1300 K. (1430 and 1880 °F) Current industrial practice is to crack ethane at temperatureswhichconvert about 60% of the feedstock on a single-pass basis. This conversion level is consideredoptimal because it combines acceptableultimate ethyleneyields and furnace sizewithout excessive recycle flow. Theprocess is usually carried out inpresence of steam, withethane:steamratios of about 70:30 by weight. The steamsuppresses side reactions that formcoke, andby loweringthe hydrocarbon fraction it also improves the selectivity to ethylene formation. For heavier feedstocks suchas naptha or gas- oil thepyrolysis time ranges from 0.1 to 0.4 seconds, while for ethane feedstockheating times as long as 0.7 seconds may be used. Decreasing the heating time improves ethylene selectivity; for 60% reduction inpyrolysis time the net savings for butane, propane andethane feedstocks are 6, 4 and 3%, respectively. [Mol, 1982] As a result of such considerations, furnaces with small diameter heatingtubes have beenproposed to effect the crackingwithin a short pyrolysis time; because of the small diameters, however, enhancedcoking onthewalls offsets the advantage of the reducedheatingtime. Theundesired sootproducedas a byproduct of cracking is partially depositedon the furnacewalls as coke, which impedes heat transfer, decreases the reaction selectivity, andcauses ethyleneproductiontobe about 1.5%more expensive at the endof a run thanat the start. [Plehiers etal. 1990]. Coke formation is highest at thepoints of highest temperature andvery sensitive to that temperature increasing the outlet temperature from 1120 to 1140 K (1557 to 1593 °F) is reportedto shorten a furnace run from 65 to 35 days, while amaximumtemperature increase of 10 K (18°F) increases the coking rate by 18%. [Mol 1982; Goossens et al. 1980] Flowrates also affect coking decreasing the heating time suppresses the formation of aromatics andcoke. [Lichtenstein 1964] Decokingmust be done regularly, typically every 40 to 100 days, taking the furnace out of production for 12 to 48 hours. The carbonaceous solidwaste is sent to a sanitary landfill.

Many inventors have reporteddiscoveries of means to increase yields and suppress coking by addition of catalysts andhomogeneous promoters. One suggestionwas to add large amounts, up to 50% by weight, of hydrogen to thepyrolysis mixture. [Okamoto andOhshima 1986; Austin et al. 1989] The effect of the hydrogenwas claimedtobe increasedheating rate, increase in concentrationof free radicals, and suppression of polycondensationreactions in the reactor. Another suggestion, whichhas attainedwidespreadcommercial use, was partial oxidation, whereby aportion of the hydrocarbon feedstock, or, alternatively, amore favorable according to chemical characteristics or on-site availability fuel than the one tobe pyrolyzed, is burned toprovide heat as well as additional reactant material. [Freundetal. 1970; KammandTanaami 1979; Font Freide et al. 1992] Various mixing arrangements may be usedto enhancemixing andheat transfer rates. Suchprocesses also generate significant amounts of carbonmonoxide andhydrogen, whichafter separation from the hydrocarbonproducts may be further utilized as synthesis gas forvarious purposes. Often theprocess heat is recycledtoprovide advantageous levels of preheat to the feedstockand the other gases. In addition to air and oxygen, other oxidizers suchas nitrous oxide, ozone andhydrogenperoxidehavebeenproposed. [Han etal. 1991] It has been discussedthat oxidizer-enhancedprocesses may optionallybe combinedwithheterogeneousmethods to enhanceburning orpyrolysis, althoughthis is not current operatingpractice.

It has also been recognizedthat addition of chlorine, as well as an oxidizer, to the feedstocktobepyrolyzed, orprior conversionof feedstock to chlorinatedhydrocarbon, opens chemical routes that may provide advantages. [Kurtz etal. 1977; Senkan 1987] These additives, however, have been foundnot to increasepyrolysis rates unless they are added in large amounts. For example, Okamoto andOhshima (1988) suggestedusingmethanol at 5% ormore of ethane feedstockto enhance ethyleneproduction. The reason for their loweffectiveness is that methanol only increases the rate of radical initiation somewhat, and thus does not enhance ethyleneproduction significantly. Oxygen increases the rate of chainbranching, but if it has tobepresent in a significant amount it convertspart of the feedstock to carbon monoxide and thereby reduces the effectiveness of theprocess, while in small amount it does not accelerate reaction significantly at the lowtemperatures employed in theprocess. We showbelowthat using combinedadditives which target twoprocesses chain initiationand chainbranching on an appropriate time scalepermits one to achieve significant enhancement with far smaller amounts of additive.

Other inventors haveproposed specialmeans to enhance thepyrolysis process by fluiddynamics. Kammet al. [1979] suggest to inject heavier feedstocks into a streamof combustionproducts, after whichthe flow is acceleratedto supersonic andthemixture further processed in a standing Shockwave. Hertzberg et al. [1994] suggest an apparatus to achieve rapidheatingby a standing shockwavewithout theprecombustion.

Objects andAdvantages of Our Invention

In accordancewith thepresent invention, enhancement of a chemical process to formvaluableproducts from feedstock is providedby arranging, with suitable equipment already knownto the art or tobe introduced into the art at some future time, to cause theprocess of chain initiation to proceedmore rapidly andundermore favorable conditions bymeans of introducing, at an appropriateplace orplaces in the flowstream, whichmay be arranged togetherwith the addition of other components to the flow stream, a supplementaryhomogeneous chain initiating substance, or a combinationof such substances, whose chemicalproperties leadto enhancedhomogeneous chain initiation, and furthermore to sustain theprocess of chain reaction at an enhanced level by introducingat an appropriateplace orplaces in the flow stream, which alsomay be in combinationwith the addition of other components to the flow stream, a supplementary substance, or substances, whose chemicalproperties sustainthehomogeneous chain reaction rate by enhancing the rate of chainbranching.

It is very important for the successful operation andoptimization of the combined "synergistic" effect that thepromoters andthe operating conditions be selectedso that the effective reaction rates of the two contributingpromoters bematchedto one another andto theprocess uponwhich theirpromotingeffect is exerted. Thus the chain initiatingpromoter shoulddecompose on a time scale similar to the reaction time employed in aprocess. Our invention is of a chemical nature, addressing the rates of components of chain reactions, anddoes not addresswhatever specificmeans maybe adopted to effect mixing, control reactor temperature orpressure, separate reactionproducts, and so on. Theprecedingparagraphs describe a method andprocess that can only be realized in industrial practice if real substances exist whichhave the requiredchemicalproperties. To demonstrate that such substances do exist, and to clarify the criteria that onepracticed in the art of chemical process engineering has to apply in order to select among alternatives formatching such substances to the relevant pyrolytic goals and conditions, we examine the case that we later describe as apreferredembodiment of this invention, namely, addition of hydrogenperoxide, (H 0₂) as a chain initiating substance, andaddition of oxygen, (0₂, which may be addedpure, present in air, or otherwise) as an enhancer of chainbranching, to aprocess streamcontaining ethane, (C₂Hg) and perhaps other substances as well, with the goal of converting the ethane into ethylene (C2H4) bypyrolysis. It is to be understood that the numerical values given for compositions, temperatures andpressures are for illustration only andthat optimal values for specific operating conditions have to be selectedaccording to prevailing feedstock and additive economics and the characteristics of theprocess equipment available or to be designedto implement the process. Description of Figures

The figures describedbelowshowcomputedresults basedupon a specific set of values for thermochemical and reaction rate parameters, namely, those of the "GRI-Mech 1.1" model, a combination of 177 elementary reactions of 31 chemical species, whichwas optimizedandvalidatedtodescribe the oxidationof natural gas. (Frenklachet al., 1994; thismodel is in thepublic domain andcanbe obtainedon the Internet by anonymous ftp to crvax.sri.comor onthe WorldwideWeb at theUniversal Resource Locatorhttp: //www.gri.orgby navigating to TechnicalAreas and then Basic Research. ) Conventional methods of numerical integrationwere used. (Lambert 1971; Radhakrishnan andHindmarsh 1993) The inferences drawn fromthese calculations, however, would followwith only inconsequential numerical changes fromany similar calculations carried out with any set of valid thermochemical andrateparameters and anyvalidmethod of numerical integration.

Figure 1 shows the computedpercent conversion of ethane to ethylene for a reaction time of one second at a temperature of 1000 K (1340 °F) and atmospheric pressure for various small additions of hydrogen peroxide. Hydrogenperoxidewas chosen as additive because the reaction times for H 0₂ thermal decompositionat 800-1000 K (980- 1350 °F) are in the range 0.02-0.2 s, (Frenklachet al. , 1994) comparable to residence times in industrial furnaces for pyrolyzing ethane to ethylene. One sees that very small amounts of hydrogen peroxide additivehave large effects upon the conversion to ethylene. One also sees that the onset of the enhancement occurs at very short heating times, less than 0.02 seconds. Inspectionof the computed species concentration and reactionrateprofiles gives insight into why the enhancement of ethyleneproduction occurs in this manner. The hydrogenperoxide additive decomposes into hydroxyl radicals

^H2°2 —^ OH + OH

that can attack ethane toproduce ethyl radicals andwater

OH + C₂H₆→C₂H₅ + H₂0,

theunstable ethyl radicals thenmostly decomposing to formethylene andhydrogen atoms

^C2H₅ → C₂H₄ + H that can generate ethyl radicals againby attackupon ethane

H + C₂H₆-→C₂H₅+H₂.

The last two elementary steps comprise an endless chain that would convert unlimited amounts of ethane to ethylenewere it not for numerous inevitable loss processes that remove hydrogen atoms and other reactive intermediate species fromparticipation in this and other chain reactions.

As noted earlier, it is of advantage to keep the pyrolysis temperature low in order to suppress undesirable side reactions that produce unwantedbyproducts including eventually coke. The temperature dependence of the enhancement by hydrogenperoxide additive is illustrated for a one second heating time in Figure 2. It can be seen that to attain a given amount of conversion the effect of just 0.4% hydrogenperoxide additive is to lower the neededpyrolysis temperature by many tens of degrees. While such aprocess temperature decrease implies only a modest direct saving in energy cost, it also implies very important indirect cost reductions by allowing dramatic suppression of the very temperature-sensitive coking rate with consequent prolongation of production runs and reduction of environmental burden.

It is clearly seen in Figure 2 that as temperature increases the effectiveness of hydrogenperoxide as additive decreases, i.e., curves 1 and 2 approach one another at higher temperatures. This happens because its decomposition time decreases to values somuch less than the pyrolysis time that the promotion effect is lost.

The enhancement of ethylene production described in Figures 1 and 2 arises from enhanced chain initiationby the hydrogenperoxide additive. We nowconsider enhancement by increasing the concentration of chain centers by means of an additive that can contribute chain branching. Such an additive is oxygen, which can react withhydrogen atoms to increase the number of chain centers in the elementary reactions

H + 0₂ —> OH + O O + C Hg — OH + C₂H₅.

In Figure 3 the effect of 1% oxygen addition on the ethylene yield fromethane after a one-secondpyrolysis time is shown for temperatures from 1100 to 1400 °F. (866 to 1033 K) It canbe seen that the increase is significant at theupper end of the temperature range. In Figure 4 the combined effects of hydrogenperoxide and oxygen adition topure ethane are illustrated for the relatively low pyrolysis temperature of 1340 °F. (1000 K) The synergistic effect of combining the chain initiatingadditive hydrogenperoxidewiththe chainbranching additive oxygen is evident. It is also evident from Figure 4 that when oxygen is present to serve as a chainbranching agent the amount of hydrogenperoxide actingas chain initiatingagent may be chosen to beverymuch less thanwhen oxygen is absent. This illustrates the advantage of oxygen addition to the economics of the process, for hydrogenperoxide and other chain initiating substances are likely to be the costliest of the components of thepyrolysis mixture. Figure 5 shows the dependence of the conversionenhancement upon the amount of hydrogenperoxide for two different residence times. It canbe seen that very small amounts of hydrogenperoxide are sufficient to generate large enhancements of the conversionyield.

The computed results shown in the figures describe idealized pyrolysis situations that presume instant heating, completemixing, isothermal reaction andno interferenceby reactions at the reactor walls. None of these is realizable in industrial practice. Thus theenhancements shown are expectedto differ fromresults obtained using the same nominal conditions in a real reactor. Whether actual enhancements are greater than computedor less than computedwill depend ondetails of reactor designand operation as these affect the mixing, heating, and subsequent coolingprocesses.

The computed results shown in the figures describe, in idealized fashion, a single preferredembodiment of our invention: Only one choice of chain initiatingadditive (pure hydrogenperoxide) , only one choice of feedstock (pureethane) , only one choice of chain propagation enhancing additive (oxygen) , andonly one choice of combinedadditive (hydrogenperoxide andoxygen) . To those skilled in the art of chemical process design it will be obvious that many alternative choices couldbemade to serve eachof these essential functions. Such choices coulddepend onwhat feedstocksmaybe locally available, what levels of per-pass conversion and recyclingmay be thought advantageous for aparticularplant, what equipment maybe available or what costs maybe incurredinpurchasing and installing newequipment, andnumerous other factorspertinent to a specific application.

Crucial to the invention disclosedherein is recognitionand exploitation of the special, in concept separate but in ultimate effect synergistic dynamic roles of modified chain initiation rates andmodifiedchain center concentrations so as to enhance the conversion of a feedstockto desiredproducts and, simultaneously or as an alternative, to effect economically competitive conversions underprocess conditions that suppress undesiredside reactions, whichmay include suppression of theproductionof environmentally undesirable byproducts. It is particularly crucial to the invention disclosedherein that the reactionrates resulting fromthe combined introduction of twopromoter substances bewellmatched to one another for thepyrolysis conditions employed. If the time scales of the twopromoting effects are far different fromone another, thenno synergismensues. This is illustrated in the example by hydrogen peroxidedecomposing on the same time scale as ethanepyrolyses in the temperature range of interest. For chain initiationat other temperatures or on other time scales, chain initiatingpromoterswith higher or lower decomposition rates will bemore suitable.

References Cited

U.S. PATENT DOCUMENTS

Austin, F.D., Baclay, J.L., and Hunter, L.O. "Hydrocarbons pyrolysis" , U.S. Patent 4,840,723, 6/1989.

Font Freide, J.J.H.M., Howard, M.J., andLomas, T.A. "Process for the production of mono-olefins" , U.S. Patent 5, 105,052 , 4/1992.

Freund, M. , Nagy, Z. , andSzepesy, L. "Method for thermally decomposing saturatedhydrocarbons to produceunsaturated hydrocarbons employing oxygen alongwith a fuel gas" , U.S. Patent 3,499,055, 3/1970.

Han, S., Martenak, D.J. , Palermo, R.E., andWalsh, D.E. "Process for direct oxidative conversion of methane to higher hydrocarbons at high ressure andmoderate temperature" , U.S. Patent 5,068,486, 11/1991.

Hertzberg, A., Mattick, A.T. and Russell, D.A. "Method for initiating pyrolysis using a shockwave" , U.S. Patent 5,300,216, 4/1994.

Kamm, G.R. , and Tanaami, K. "Integratedprocess for thepartial oxidation-thermal cracking of crude oil feedstocks", U.S. Patent 4,134,824, 1/1979.

Kamm, G.R. , Milks, D. , Kearns, J.D. , Britt, H. I. , andKhavarian, CR. , "Process for the thermal cracking of hydrocarbons" , U.S. Patent 4,136,015, 1/1979.

Kurtz, B.E., andSmalley, E.W. "Process for producing ethylene from ethane", U.S. Patent 4,051,193, 9/1977.

Okamoto, T. , andOhshima, M. "Thermal crackingprocess for producing olefins fromhydrocarbons", U.S. Patent 4, 613,426, 7/1986.

Senkan, S. "Production of higher molecular weight hydrocarbons from methane", U.S. Patent 4, 714,796, 12/1987. OTHERDOCUMENTS

Atkins, P.W. Physical Chemistry, FourthEdition, Freeman, 1990.

BorisovA.A., Zamansky, V.M., Lissianski, V.V., Skachkov, G.I., andTroshin, K.Ya. "PromotedHigh—TemperatureReactions", 22nd Symposium (International) onCombustion, 1987, p. 903.

Dente, M. , andRanzi, E. "Mathematical modelingof hydrocarbon pyrolysis reactions", in Pyrolysis theory and industrial practice, Aibright, L.F., Crynes, B.L., andCorcoran, W.H. , Eds, Academic Press, Inc., NewYork, 1983.

Frenklach, M. , Wang, H. , Gardiner, W.C. , Lissianski, V.V. , Smith, G.P., and Bowman, C.T., "GRI-Mech 1.1 Natural Gas Combustion Mechanism" , publicly available on the Internet by anonymous ftp to crvax.sri.com, or on theWorldWideWeb at http://www.gri.org/tech/res, March 1994.

Gardiner, W.C. , Rates andMechanisms of Chemical Reactions, Benjamin, NewYork, 1969.

Goossens, A.G., Dente, M. , andRanzi, E. "Improve steamcracker operation" , in Process control and optimization handbook for the hydrocarbonprocessing industries, L.A. Kane, Ed., Gulf Publishing Company, Houston, 1980, pg. 30.

Kondratiev, V.N. andNikitin, E.E., Kinetika i mechanizmgazofaznyx reaktsii, Izdatel'svonauka, Moscow, 1974, Chapter 11.

Lambert, J.D. , Computational methods in ordinary differential equations, JohnWiley & Sons, NewYork, 1973.

Lichtenstein, I. "Design cracking furnaces", Chemical Engineering Progress, 60, 64 (1964) .

Minkoff, G.J., andTipper, C.F.H. , "Pyrolysis of hydrocarbons and their derivatives" , Chapter 9 of Chemistry of combustion reactions, Butterworths, London, 1962. Mol, A. "Lowresidence time enhances NGLcracking" , Oil andgas journal, March29, 1982, pg. 87.

Plehiers, P.M., Reyniers, G.C., an Froment, G.F. "Simulationof the run length of an ethane cracking furnace", Ind. Eng. Chem. Res., 29, 636 (1990) .

Radhakrisnnan, K. , andHindmarsh, A.C. Descriptionanduse of LSODE, the Livermore solver for ordinary differential equations, NASA Reference Publication 1327, December 1993.

Stratton, A., Hemming, D.F., andTeper, M. Ethyleneproductionfrom oil, gas and coal-derived feedstock, EconomicAssesment Service, London, 1983.

Zamansky, V.M., andBorisov, A.A. "Promotion of High-Temperature Self- Ignition", Progress in energy and combustion science, 18, 297 (1992) .

Summaryof the Invention

The inventiondisclosedhereinconsists of ourdiscoverythat there exist conditions of practical interest where a synergistic effect of adding two ormore homogeneous promoters, eachcontributing in a different dynamic role to an overall chemical process of interest, can lead tomore favorable reactionoutcomes thanprevail in absence of this synergistic effect.

Description of the Invention

The inventiondisclosedherein, derived from the theory of chemical reaction rates, moreparticularly the rates of chain reactions, comprises theuse of two ormore homogeneouspromoters under conditions appropriate to achieve abeneficial synergistic effect. A complex chemical process involvingchain reactions canbedecomposed into subprocesses that have distinctivedynamic behavior, including chain initiation, chainpropagation, chainbranching, andchain termination. For given reaction conditions, each subprocessproceeds at a characteristic reaction rate. Additionof different homogeneous promoters may influence these subprocess rates in different ways. The discovery disclosedherein is that a remarkably synergistic effect canbe achievedby combiningtwo ormorepromoters, each influencingdifferent subprocesses, under conditions of temperature and composition such that the reactionrates of the homogeneous promoters are appropriate to the time scale of theprocess that one seeks to enhance. As we discoveredthroughuse of detailedcomputer simulations basedupon awell validated chemical model for small hydrocarbon oxidation, promotionof theyieldofpyrolytic conversion of hydrocarbon feedstock, exemplifiedby thepyrolytic conversion of ethane to ethylene, may be so enhanced in a synergisticmanner if promoters influencingdifferent subprocesses are introduced. Practical application of this discoverymaybedirectedtoward different goals. For example, onemayutilize the synergistic effect to increase theyield of product, shorten the reaction time, reduce the temperature or pressure of theprocess, or some combination of all of these. Implicit in these applications will be the opportunity to suppress the formation of undesiredbyproducts of reaction, some of whichmaybeunwantedenvironmentalburdens of theprocessespresently inuse. Process design exploitingour inventionmaybeguidedby computer simulations, as we have illustrated for the case of ethane pyrolysis to ethylene. On the other hand, it is alsopossible tomake use of our invention on an empirical basis. This is possiblebecause many other substances acting as chain initiating, chainbranching and chainpropagation agents arewell recognized in the science of chemistry. For example, dimethyl ether and a variety of peroxides are known as chain initiating agents withpractical applications. It is be recognizedthat our inventiondoes not consist of simply adding more than one homogeneous promoter to a chemical process. Indeed, many industrial as well as other chemical processes couldbe classified as involving combinations of substances that enhance someprocess of interes . Our invention specifically targets the synergistic enhancement that can be achievedby appropriatelypromoting, by whatever additives may be chemically and economically suitable, the rates of at least two dynamically different parts of a chainreaction process. The synergistic effect permits the amounts of promoter requiredtobe dramatically reduced, suchthat one ormore of the promoters canbe relatively expensive compared to the feedstockor the desiredproduct. Operation of the Invention

The inventiondisclosedherein operates within the ordinary context of chemical engineeringpractice. It is not restrictedby anyparticularphysicalmeans thatmaybeusedto carry out mixing, heatingor any other aspect of theprocess tobe improvedby adoption of our invention; its nature is chemical only, inparticular, chemical-dynamic. It canbe appliedto reactions startingwithmany different feedstocks or combinations of feedstocks. It canbeused withchain initiatingsubstances or chain center concentration sustainingsubstances other than thosenamedas specific examples for purpose of illustrationof theprinciples involved. It canbecombined withanymethods nowknown to the art, or later tobe introducedinto the art, for achieving flowstreamcontrol, reactant mixing, heat supply, product separationandpurification, andall of the attendant process operations as maybe thought advantageous for aparticular installation.

Conclusion, Ramifications, andScope of Invention

We disclose herein aprocess forenhancedconversionof feedstockto valuableproducts. It maybe accompaniedby significant suppression of the formationof environmentallyburdensome or otherwise undesirablebyproducts of reaction. It is basedupon recognizingthe potential of a synergistic combinationof additiveswhose significant effects upon the reactionaddress different aspects of the chain reactions involved. Thepreferredembodiment of our invention, based upon economics of the hydrocarbonprocessing industryprevailing today, is for improving theproductionof ethylene fromethane feedstockby additionof hydrogenperoxide andoxygenor air. Our discovery, however, has muchwider scope, as the synergistic effect that we discloseherein canbe realizedbymany other combinations of substances, some of whichmaybe favoredover hydrogenperoxide andoxygen for useunderdifferent conditions of temperature or composition, or for economic reasons, or for reasons relatedto process safety, materials transport or handling

Claims

Claims :

We claim as our invention (1) Amethod andprocess for improving ayield of one ormore desired products, or suppressing formation of one or more undesiredproducts, or achieving both goals together, in a chemical process that involves one or more chain reactions, in whichmeans are used to introducemore than one homogeneous promoter in a controlledmanner so as to compose, on an appropriate time scale, a synergistic combination of altered chain initiation rates and one or more other altered chain reaction rates, whereby the economics of the overall chemical process may be modified, the range of suitable operating conditions for the overall chemical process may be modified, the scale of the overall chemical process may be modified, the energy requirement for the overall chemical process may be modified, the safety of the overall chemical process may be modified, or one or more environmental burdens may be eliminated or lessened.

(2) The method and process of claim (1) wherein the overall chemical process uses a predetermined feedstock substantially comprising one or a mixture of hydrocarbons.

(3 ) The method and process of claim (1) wherein the overall chemical process is substantiallypyrolytic in character.

(4) The method and process of claim (1) wherein the overall chemical process effects the reduction of the degree of saturation of carbon- carbon bonds in the products of the overall chemical process, on average, compared to the degree of saturation of carbon-carbon bonds in the feedstock.

(5) The method and process of claim (1) wherein a major product of the overall chemical process is substantially ethylene.

(6) The method and process of claim (1) wherein a major component of the feedstock is substantially ethane.

(7) The method andprocess of claim (1) wherein the homogeneous promoters used are substantially hydrogenperoxide, whichmay be substantially pure or added in the form of a liquid solution, and oxygen, substantiallypure, dilutedwith other gases, or substantially air.

(8) The method andprocess of claim (1) wherein one of the homogeneous promoters consists entirely or inpart of apredetermined mixture of products from a homogeneous combustionprocess introduced prior to or in concert with theprocess modifiedby thehomogeneous promoters .

(9) The method andprocess of claim (1) wherein one of the homogeneous promoters consists entirely or inpart of a predetermined mixture of products from an electrical discharge introducedprior to or in concert with the process modifiedby the homogeneous promoters.

(10) Themethodandprocess of claim (1) wherein one of the homogeneous promoters consists entirely or inpart of a predetermined product stream resulting froma photolyticprocess introducedprior to or in concert with theprocess modifiedby the homogeneous promoters.

(11) Themethod andprocess of claim (1) wherein one of the homogeneous promoters consists entirely or inpart of apredetermined product streamresulting fromapyrolytic process introducedprior to or in concert with the process modifiedby the homogeneous promoters.

(12) The method andprocess of claim (1) wherein one of the homogeneous promoters consists entirely or inpart of a predetermined product stream resulting from a heterogeneous combustionprocess introducedprior to or in concert with theprocess modifiedby the homogeneous promoters.

(13) The method andprocess of claim (1) wherein other substances besides the two or more promoters are addedto the feedstock or during the course of reaction.

(14) Themethod andprocess of claim (1) whereina majorproduct of the overall chemical process is acetylene.

(15) The method andprocess of claim (1) wherein a majorproduct of the overall chemical process is benzene.

(16) The method andprocess of claim (1) wherein the feedstock is composedsubstantially of methane.

(17) The method andprocess of claim (1) wherein one ormore predeterminedphysical means are used for thepurpose of altering the reaction conditions, includingbut not limited to thepurpose of altering themixingprocess for one or more of the homogeneous promoters, supplyingprocess heat, modifying the turbulence level, or changing flow characteristics by variable area ducts.

(18) The method andprocess of claim (1) wherein recycling of product stream is employed.

(19) Themethod andprocess of claim (1) whereinmethods of active control are used to adjust the operating conditions.