The Oxidative Dehydrogenation of n- Butane in a Fixed-Bed Reactor and in an Inert Porous Membrane ReactorMaximizing the Production of Butenes and Butadiene

The Oxidative Dehydrogenation ofn-Butane in a Fixed-Bed Reactor and in anInert Porous Membrane ReactorsMaximizing the Production of Butenes andButadiene

David Milne, David Glasser, Diane Hildebrandt,* and Brendon Hausberger

Centre of Material and Process Synthesis, UniVersity of the Witwatersrand, Johannesburg, PriVate Bag 3,WITS 2050, South Africa

The oxidative dehydrogenation (ODH) ofn-butane (butane) produces three isomers of butene (1-butene,trans-2-butene, andcis-2-butene) which in turn are oxidized to form butadiene. Butane also is oxidized directly tobutadiene. In this simulation study, the authors have analyzed the operating conditions required to producethe maximum amount of butenes, i.e., all three isomers, and butadiene in a fixed-bed reactor (FBR) and inan inert porous membrane reactor (IMR). The theoretical maximum yields of butenes and butadiene werefound to be 0.119 and 0.800 carbon mass fractions, respectively. The reactor configuration in both instanceswas a large IMR operating at a low constant partial pressure of oxygen in the stream of reactants and products.It was found that 99.7% and 83% of the theoretical maximum yields of butenes and butadiene, respectively,can be achieved in an IMR with a constant oxygen partial pressure of 0.25 kPa. The corresponding residencetimes are 75 and 322 s. Candidate attainable regions have been identified for the system subspaces butane-butenes and butane-butadiene.

Introduction

Olefins and dienes are precursors for a wide range of usefulchemicals. A very attractive route to make them is via theoxidative dehydrogenation of hydrocarbons, as these are readilyavailable from crude oils and Fischer Tro¨psch synthesis. Theproblem with this route is to try to minimize the oxidation ofthese hydrocarbons to other products such as carbon monoxide,carbon dioxide, and water. However, such routes to olefins anddienes will only become practical when both the yield of productand the selectivity to the desired product are high.

1,3-butadiene is a high-volume and valuable intermediateorganic chemical used in many industrial processes to producerubber, resins, and plastics. It is involved in several differentreactions including addition, oxidation, and substitution reac-tions, but its main use is for polymerization. Most 1,3-butadieneis used in synthetic elastomer production and in adiponitrileproduction, the raw material for nylon-6,6 production. Theoverall demand for butadiene is expected to increase becauseof the growth of specialty uses for it.1 Butadiene is usuallyproduced by one of two processes (a) recovery from a mixedhydrocarbon stream and (b) by the oxidative dehydrogenation(ODH) of butenes.1

In this paper, we examine the ODH ofn-butane to butenesand butadiene. Butane is a readily available feedstock and isproduced from crude oils and Fischer Tro¨psch synthesis, andwe believe that its conversion to butadiene offers potentiallysignificant economic benefits. Another requirement is to achievea high selectivity of butane to butadiene allied to high yields ofbutadiene.

Once the kinetics of the reactions are known, it is importantto optimize the reaction system to ensure that the economics ofthe process make it an attractive industrial option. In this paper,we examine the possible maximum yields and selectivities andthen the ways of achieving them in practice.

In an earlier paper,2 the authors studied the ODH of 1-buteneto butadiene in a fixed-bed reactor (FBR) and in an inert porousmembrane reactor (IMR). It was found that, in an IMR wherethe inlet oxygen partial pressure was maintained at a constantlevel along the length of the reactor, the maximum yield ofbutadiene increased as the oxygen partial pressure was reduced.This earlier paper acknowledged the work done on the ODH ofbutane by Te´llez et al.3-5 and Assabumrungrat et al.6 Thecatalyst used in the FBR and IMR reactors was a V/MgOcatalyst containing 24% (by mass) of V2O5.

The reaction network for the ODH of butane was postulated3,4

as shown in Figure 1. The three isomers, 1-butene,trans-2-butene, andcis-2-butene, have been lumped together as C4H8

in reactions 7-9.The mathematical model created to describe and simulate the

ODH of butane assumed isothermal conditions and atmosphericpressure. Maintaining atmospheric pressure in the reactorimplied varying the size of the catalyst bed to attain the desiredyields of butenes and butadiene.

Matlab, Version 6, Release 13, was used for all the simula-tions. The kinetic rate expressions for the oxidation of butane,butenes, and butadiene were taken from Te´llez.3 These expres-sions have as variables the partial pressures of oxygen and thehydrocarbons, butane, butenes, and butadiene.

In principle, one would like to analyze the system using theattainable region (AR) method, because this would give resultsfor the optimum conditions and reactor structure to achieve adesired product. In this particular ODH study, the size of theproblem is too large to be currently analyzed using this approach.

* To whom correspondence should be addressed. E-mail:[email protected]. Tel.:+27 (11) 717 7557. Fax:+27 (11) 717 7557.

Figure 1. Reaction scheme for the oxidative dehydrogenation of butaneto butenes and butadiene.

2661Ind. Eng. Chem. Res.2006,45, 2661-2671

10.1021/ie050120l CCC: $33.50 © 2006 American Chemical SocietyPublished on Web 03/09/2006

However, when doing our analyses, some of the thinking behindthis method is employed.

Results

An initial feed mixture of butane and oxygen was used, andthe partial pressure of oxygen was varied over the range 0.25to 85 kPa. The feed temperature and the reactor isothermaltemperature was 773 K. As in our earlier paper,2 all hydrocarbonconcentrations are expressed in terms of mass fractions ofcarbon.

Three scenarios were considered. The first was feeding butaneand oxygen, the latter at an initial specified partial pressure, toa stabilized (steady state) FBR and permitting the reaction tocontinue until either all the oxygen or all the butane wasdepleted. The effect of oxygen partial pressure in the feed streamupon the yields of butenes (Case 1) and butadiene (Case 2) wasstudied. In the second scenario, using a stabilized IMR, thepartial pressure of oxygen was maintained at a constant specifiedlevel by the addition of fresh oxygen along the length of theIMR. Again, the effect of oxygen partial pressure in the feedstream upon the yields of butenes (Case 3), butadiene (Case4), and butenes and butadiene combined (Case 5) was studied.In a third scenario, the authors have explored the effect uponthe candidate attainable region of deploying two very large IMRsin series and by incorporating a policy of bypass and mixing.

The effect of residence time upon yields of butenes andbutadiene was examined. In all instances, the reaction waspermitted to attain equilibrium, at which stage either the oxygenor the butane had been depleted. In effect, the stoichiometricratio of oxygen in the feed was varied to simulate differentreactant compositions.

Despite there being a spectrum of seven products other thanbutane and oxygen in the product stream, this study hasconcentrated only on butenes and butadiene. The yields ofcarbon monoxide, carbon dioxide, and water were not consid-ered.

Scenario 1, Case 1: Depletion of Oxygen in a FBRsProduction of Butenes. The reactor configuration for thisscenario is shown in Figure 2. Using the given rate equationsand the initial conditions, that is, of pure butane with thespecified oxygen concentration (i.e., partial pressure), a totaloperating pressure of 1 atm, and an isothermal temperature of773 K, one can integrate the differential equations to obtainthe results shown in Figure 3, where all butane and butenesconcentrations are expressed in mass fractions of carbon.

In Figure 3, and in subsequent figures of concentrationprofiles, the various points on the profiles represent theconcentrations of reactant and product were the reaction to bestopped at that point, i.e., after the concomitant residence time.

At initial oxygen partial pressures of 85 and 86 kPa, thereaction proceeds until, at equilibrium, all the oxygen has beendepleted. When this occurs, the residual butane and butenesconcentrations for an oxygen partial pressure of 85 kPa are 0.075and 0.017, respectively. The other components present oncompletion of the reaction, other than butane, butenes, andbutadiene, are carbon monoxide, carbon dioxide, and water. All

the oxygen has been utilized in the oxidation of butane, butenes,and butadiene.

If the initial partial pressure of oxygen is increased to 87kPa, at equilibrium, all the butane, butenes, and butadiene areoxidized and there is residual oxygen present on completion ofthe reaction. At this initial partial pressure of oxygen, the supplyof butane is the limiting factor.

At oxygen partial pressures<87 kPa, the reaction ceases withoxygen depletion. At an initial oxygen partial pressure of 65kPa, reaction cessation occurs after a residence time of 31 s (at45 kPa, cessation occurs after a residence time of 14 s). Oxygendepletion was defined as when its partial pressure had fallenbelow 0.001 kPa, and the commensurate reactor residence timeat this milestone was noted.

The selectivity (S) of butane to butenes was defined as

Usually, selectivity is calculated as the ratio of moles of productand moles of reactant consumed. In the case of the ODH ofbutane to butenes, carbon mass fractions can be used insteadof moles because of the presence of four carbon atoms in eachof the relevant hydrocarbon molecules, butane and butenes (andbutadiene). This implies that the difference in the molar massesof butane and butenes, which otherwise would render thisdefinition invalid, does not apply in this case.

A maximum yield of butenes, 0.109, occurs at an initialoxygen partial pressure of 49 kPa after a residence time of 16s. Residual butane has a concentration of 0.634. If we examineFigure 3 in more detail, we see that the selectivity of butane tobutenes (butenes formed divided by butane consumed) is givenby the slope of a straight line from the feed point. Thus, as theprofiles shown in Figure 3 are bounded by convex curves withthe greatest slope at the beginning (the feed point), the largestselectivity of butane to butenes occurs at small conversions.The partial pressure of oxygen present does not seem to affectthis value significantly. At 85 kPa, the initial slope is 0.65, andat 15 kPa, the initial slope is 0.60. Thus, to get high selectivitiescommensurate with reasonable conversions, one would need asystem with low conversions but embodying separation andrecycle.

It is of interest to examine the residence times necessary toobtain the results shown in Figure 3. Figure 4 shows that thereaction times to attain the maximum yield of butenes do notexceed 25 s for all oxygen partial pressures, implying that theODH reaction is a very fast one.

In Figure 4, the “kink” in the residence time profile for 65kPa (and for 85 kPa at a residence time of 160 s) is attributed

Figure 2. FBR configuration.

Figure 3. Profiles of butane and butenes at various oxygen partial pressuresin a FBR.

SButane)CButenes

(CButane0 - CButane)

2662 Ind. Eng. Chem. Res., Vol. 45, No. 8, 2006

to the low concentration of oxygen resulting in no further netdepletion of butenes. It was established that the reaction wasstill occurring and butadiene was still being formed. This meantthat the butenes were being oxidized to butadiene as fast asthey were formed and/or that the butane was being oxidized tobutadiene directly.

Figure 5 shows the selectivity profile for butane relative tobutenes at a maximum yield of butenes as a function of theinitial oxygen partial pressure in a FBR.

The discontinuity in the selectivity at an oxygen partialpressure of 56 kPa is explained by reference to Figure 3. Atoxygen partial pressures from 85 to 57 kPa, the final butenesconcentration is less than the maximum butenes concentration.Below 57 kPa, the final and the maximum butenes concentra-tions are identical. As selectivity in Figure 5 is calculated forthe maximum yield of butenes, a shift occurs at an oxygenpartial pressure of 56 kPa. It is apparent from Figure 5 that, forinitial oxygen partial pressures in a FBR in excess of 56 kPa,the selectivity of butane to butenes is relatively unaffected bythe oxygen partial pressure.

Figure 5 indicates that butane selectivities for maximumbutenes vary widely over the range of partial pressures. At 85and 1 kPa, butane selectivities are 0.28 and 0.72, respectively.

Scenario 1, Case 2: Depletion of Oxygen in a FBRsProduction of Butadiene.At an initial oxygen partial pressureof 85 kPa, the reaction proceeds until all the oxygen has beendepleted. When this occurs, the residual butane and butadieneconcentrations are 0.075 and 0.059, respectively.

In Figure 6, the “kink” at the end of the concentration profilefor 85 kPa is attributed to the very low concentration of oxygenat that stage of the ODH process, resulting in the preferential

oxidation of butane to butadiene via reactionr4 rather than tobutene via reactionsr1, r2, andr3, as was shown diagrammati-cally in Figure 1.

If the initial partial pressure of oxygen is increased to 87kPa, all the butane, butene, and butadiene is oxidized and thereis residual oxygen present on completion of the reaction. Atthis initial partial pressure of oxygen, the supply of butane isthe limiting factor. At oxygen partial pressures of 85 kPa andless, the reaction ceases with oxygen depletion.

The maximum yield of butadiene from a FBR, 0.183, occursat an initial oxygen partial pressure of 70 kPa. The residualbutane has a concentration of 0.399. The residence time is 41s.

A characteristic of all the butadiene/butane profiles shownin Figure 6 is the presence of a concave region between thefeed point and the profile (at an oxygen partial pressure of 85kPa, the concave region extends from the feed point to thetangential point at a butadiene concentration of∼0.09).

Figure 7 shows that the reaction times to attain the maximumyields of butadiene do not exceed 49 s for all oxygen partialpressures up to 85 kPa, implying that the ODH reaction is afast one.

The selectivity (S) of butane to butadiene was defined in thesame manner as the selectivity of butane to butenes, namely,

We may use Figure 6 to examine the selectivity of the butane

Figure 4. Residence times for butenes at various oxygen partial pressuresin a FBR.

Figure 5. Selectivity of butane to butenes in a FBR as a function of initialoxygen partial pressure for conditions of maximum yield of butenes.

Figure 6. Profiles of butane and butadiene at various oxygen partialpressures in a FBR.

Figure 7. Residence times for butadiene at various oxygen partial pressuresin a FBR.

SButane)CButadiene

(CButane0 - CButane)

Ind. Eng. Chem. Res., Vol. 45, No. 8, 20062663

to butadiene. The maximum selectivity is given by the line ofmaximum slope from the feed point (pure butane). Because ofthe concavity of the profiles in Figure 6, this will occur whenthe line is tangential to the curve or, where no tangent pointexists, at the final point of the profile.

The discontinuity at an oxygen partial pressure of 80 kPa isexplained by reference to Figure 6. At oxygen partial pressuresfrom 85 to 81 kPa, the final butadiene concentration is less thanthe maximum butadiene concentration. Below 81 kPa, the finaland the maximum butadiene concentrations are identical. Asselectivity in Figure 8 is calculated for the maximum yield ofbutadiene, a shift occurs at an oxygen partial pressure of 80kPa.

Figure 8 indicates that butane selectivities for maximumbutadiene vary by 100% over the range of partial pressures. At85 and 1 kPa, butane selectivities are 0.24 and 0.18, respectively,with a maximum selectivity of 0.36 at an oxygen partial pressureof 50 kPa.

The consequence of this was that the supply of oxygen at anappropriate partial pressure was deemed to be an importantfactor for high yields of butenes and butadiene. To explore thishypothesis, the control of the oxygen supply to a different reactorconfiguration was examined. The reactor configuration was anIMR with oxygen injection along the length of the reactor tomaintain a constant oxygen partial pressure in the gas mixture.

Scenario 2, Case 3: Replenishment of Oxygen in an IMRsProduction of Butenes. The reactor configuration for thisscenario is shown in Figure 9. As before, we can integrate thesystem of differential equations describing this system. As wedid previously, we will limit the total reactor tube-side pressureto 1 atm and the isothermal temperature to 773 K. Furthermore,we will assume that we supply the oxygen in such a way as tomaintain its partial pressure in the reactor at a constant valueequal to that in the feed stream and to replenish that consumedin the ODH process. Because of the way we analyze our resultsin terms of carbon mass fraction, this addition does not affectour analysis unduly.

Figure 10 shows the effect of adding oxygen along the lengthof the IMR to maintain a constant oxygen partial pressure in

the stream of reactants and products. It is noticeable from Figure10 that the maximum yield of butenes increases but marginallydespite the significant reduction in oxygen partial pressure from85 to 0.25 kPa. At an oxygen partial pressure of 0.25 kPa, themaximum yield of butenes is slightly less than 0.119 with acommensurate residual butane value of 0.622. The associatedresidence time is 75 s (see Figure 11).

A detailed analysis of Figure 11 shows that the residencetime for maximum yield of butenes decreases from a value of41 s to 9 s with reduced oxygen partial pressure over the range85-10 kPa. One can further see that, as the (constant) partialpressure of oxygen is reduced below 10 kPa, the residence timesfor the maximum yield of butenes gradually increase. For partialpressures<1 kPa, the residence time for the maximum yieldof butenes increases sharply.

Figure 12 shows this interesting result more clearly, i.e., theresidence times for the maximum yield of butenes at variousoxygen partial pressures. A possible explanation for the shape

Figure 8. Selectivity of butane to butadiene in a FBR as a function ofinitial oxygen partial pressure for conditions of maximum yield of butadiene.

Figure 9. IMR configuration.

Figure 10. Profiles of butane and butenes at constant oxygen partialpressures from 85 to 0.25 kPa in an IMR.

Figure 11. Residence times as a function of mass fraction of butenes atconstant oxygen partial pressures from 85 to 0.25 kPa in an IMR.

Figure 12. Residence times for maximum yield of butenes at constantoxygen partial pressures from 95 to 0.25 kPa in an IMR.


of this curve is that high oxygen partial pressures require longerresidence times because of the scarcity of other reactants. Oncethe oxygen partial pressure is reduced, so are the reaction rates.This implies a minimum in the curve, as was found to be thecase.

Maximum butenes yields, associated butane values, butenesselectivities, and residence times from an IMR operating at aconstant oxygen partial pressure are shown in Table 1. FromFigure 10, it is concluded that the maximum yield of butenesincreases with decreasing oxygen partial pressure. Figures 11and 12 show that the residence times associated with themaximum yield of butenes fall to a minimum and then increase.The maximum selectivity of butane to butenes is attained atlow oxygen partial pressures, but the profile of these selectivitiesis fairly flat, with the percentage difference between the observedminimum and maximum selectivities being but 10%.

We conclude from Table 1 that the selectivity of butane fora maximum yield of butenes in an IMR is but slightly influencedby the oxygen partial pressure. This observation that increasedbutenes yield is associated with low oxygen partial pressureraises the question as to what yield of butenes could be attainedat a very low oxygen partial pressure and in a very large reactor?

This question was answered by defining a very low oxygenpartial pressure as 0.000 001 kPa, and the results are shown inFigures 13 and 14. The maximum yield of butenes at a verylow oxygen partial pressure is 0.119 with a corresponding butaneconcentration of 0.623. The associated selectivity of butane tobutenes is 0.316. As before, the butane-butenes profile in Figure13 is convex over its entire length.

Figure 14 shows that the residence time at a very low oxygenpartial pressure for the total conversion of butane is 5.63× 107

s. The residence time for the maximum yield of butenes is 1.7× 107 s.

As has already been noted, for an IMR at a constant oxygenpartial pressure of 0.25 kPa, the maximum yield of butenes isslightly less than 0.119 with a residence time of 75 s (residualbutane 0.622). This, in a commensurately sized reactor, repre-sents an achievement of practically 100% relative to thetheoretical maximum butenes yield.

For a FBR with an initial oxygen partial pressure of 49 kPaand in which the oxygen is not replenished, the maximum yieldof butenes is 0.109 with a residual butane concentration of 0.634(see Figure 3). The residence time was 16 s. This represents anachievement of 92% relative to the theoretical maximum butenesyield of 0.119.

In Figure 15, we show the FBR profile for an oxygen partialpressure of 49 kPa. Also shown are the butane-butenes profilesfor an IMR in which the original oxygen partial pressures (0.25and 0.000 001 kPa) are maintained constant through the additionof fresh oxygen along the length of the reactor (with the twoprofiles being practically identical).

It is noteworthy that the butane-butenes profiles consideredin Figure 15 (depleted oxygen at 49 kPa and constant oxygenat 0.25 kPa) all lie below the profile for a very low oxygenpartial pressure. The profile for an oxygen partial pressure of0.25 kPa lies very close to, but nevertheless below, the profilefor an oxygen partial pressure of 1× 10-6 kPa.

From an analysis of Figure 15, we conclude that thetheoretical profile for maximum butenes yield at a very lowoxygen partial pressure represents the furthermost boundarywithin which all scenarios so far identified lie. Consequently,we believe that Figure 13 represents a candidate AR for thesystem subspace butane-butenes.

Table 1. Maximum Butenes Yields, Selectivities, and ResidenceTimes from an IMR at Various Constant Oxygen Inlet PartialPressures

oxygenpartial

pressure, kPa

maximumbutenesyield

associatedbutanevalue

butaneselectivity

residencetime, s

85 0.103 0.638 0.286 4165 0.104 0.634 0.283 1949 0.104 0.639 0.288 1345 0.104 0.641 0.290 1225 0.106 0.645 0.297 915 0.107 0.632 0.291 910 0.109 0.633 0.297 95 0.112 0.637 0.309 101 0.117 0.621 0.309 240.49 0.118 0.626 0.316 410.25 0.119 0.622 0.314 75

Figure 13. Profile of butenes and butane at a very low oxygen partialpressure and in a very large IMR.

Figure 14. Residence time as a function of butenes concentrations at avery low oxygen partial pressure and in a very large IMR.

Figure 15. Profiles of butane and butenes at different oxygen partialpressures for an IMR and a FBR.


Butenes Yields

The best yields of butenes from the reactor configurationsstudied were compared with the theoretical best yield of butenesof 0.119 from an IMR of very large size. Ranked in order oftheir closeness to the theoretical best yield, the results from thereactor configurations are shown in Table 2.

From Table 2 it is concluded that an IMR with a residencetime of 75 s operating under a constant oxygen partial pressureof 0.25 kPa gives a maximum butenes yield of 0.1188 carbonmass fraction, which is 99.7% of the theoretical maximum yieldof 0.1191.

The second highest yield also is from an IMR. The butenesyield of 0.1182 (99.2% of the theoretical maximum yield) wasachieved at a residence time of 41 s and at an oxygen partialpressure of 0.49 kPa.

In practical terms, all the reactor configurations shown inTable 2 produced maximum yields of butenesg90% of thetheoretical maximum. If 90% is accepted as the minimumcriterion, the preferred reactor configuration is an IMR with aconstant oxygen partial pressure of 5 kPa and a residence timeof 10 s. The resulting maximum yield of butenes, 0.112, is94.1% of the theoretical maximum.

No concave sections were observed in any of the butane-butenes profiles investigated, and consequently, no mixingstrategies were applied.

Effect of the Temperature upon the Yield of Butenes.Allthe analyses conducted have been at the isothermal temperatureof 773 K,5,6 and consequently, our candidate AR shown inFigure 13 is applicable only at that temperature. Figure 16 showsthe effect of temperature upon the butane-butenes profile in avery large IMR when the oxygen partial pressure is very low.

Examination of Figure 16 shows that increasing the reactortemperature from 773 to 823 K reduces the maximum theoreticalyield of butenes from 0.119 to 0.105 with an associated butaneconcentration of 0.665. The associated residence time was 5.39× 106 s, and the associated selectivity of butane at this

temperature is 0.313. Decreasing the operating temperature from773 to 748 K marginally increases the maximum theoreticalyield of butenes (from 0.119 to 0.124) with an associated butaneconcentration 0.596. The associated residence time was 3.15×107 s with a selectivity of butane at 748 K of 0.307.

In the butane concentration range of 0.76 to 0.90, both anincrease and a decrease in temperature result in slightly loweryields of butenes, as the two profiles for 748 and 823 K lieunder the profile for 773 K. Refer to Figure 17 for a magnifiedview of this. Consequently, we maintain that each of the threeprofiles shown in Figure 16 represents a candidate AR for thesystem subspace butane-butenes at the respective temperature.

To conclude our analysis, we investigated the circumstancesat which the maximum yields of butenes from a FBR and anIMR are equivalent. A detailed analysis of Figures 3 and 10shows that, at high oxygen partial pressures, a greater yield ofbutenes is obtained from a FBR than from an IMR and that, atlow oxygen partial pressures, the converse is applicable. Thecritical value of oxygen partial pressure was found to be 39kPa. At this pressure and greater, the maximum yield of butenesis greater from a FBR than from an IMR. Below 39 kPa, themaximum yields of butenes are greater from an IMR. Thegreatest percentage difference between the maximum yields ofbutenes, 5%, is at an oxygen partial pressure of 49 kPa.

Table 3 shows the respective values at oxygen partialpressures close to 39 kPa.

Scenario 2, Case 4: Replenishment of Oxygen in an IMRsProduction of Butadiene.Figure 18 shows the effect of addingoxygen along the length of the reactor to maintain a constantoxygen partial pressure in the stream of reactants and products.It is noticeable from Figure 18 that the maximum yield ofbutadiene increases with the reduction in oxygen partial pressurefrom 85 to 0.25 kPa. At an oxygen partial pressure of 0.25 kPa,the maximum yield of butadiene is 0.665 with a commensuratebutane concentration of 0.042. The associated residence timeis 322 s (see Figure 19).

Table 2. Best Butenes Yields from the Various Reactor Configurations Ranked According to Their Closeness to the Theoretical MaximumYield of Butenes

sourcemaximum

butenes yieldassociated

butane yieldresidencetime, s

selectivity:butane to butenes

% of theoreticalC4H8 yield

O2 partialpressure, kPa

reactorconfiguration

Table 1 0.119 0.622 75 0.314 99.7% 0.25 IMRTable 1 0.118 0.626 41 0.316 99.2% 0.49 IMRTable 1 0.117 0.621 24 0.309 98.3% 1 IMRTable 1 0.112 0.637 10 0.309 94.1% 5 IMRTable 1 0.109 0.633 9 0.297 91.6% 10 IMRFigure 3 0.109 0.634 16 0.298 91.6% 49 FBRTable 1 0.107 0.632 9 0.291 90.0% 15 IMR

Figure 16. Effect of temperature upon the theoretical maximum yield ofbutenes. Figure 17. Magnified section of Figure 16.


Again, it should be noted that each of the profiles shown inFigure 18 exhibits a concave section. These concave regionscan be removed through an appropriate mixing scenario (alonga straight line from the feed point that is tangential to the profile)involving fresh reactant (butane) and reaction products. Themaximum butane selectivity is found at the tangential point ofthe relevant profile. A detailed analysis of Figure 19 shows thatthe residence time for the maximum yield of butadiene decreasesfrom a value of 108 s to a minimum of 23 s over the oxygenpartial pressure range of 85-15 kPa.

As the partial pressure of oxygen is reduced below 15 kPa,the residence times for the maximum yield of butadienegradually increase. For partial pressures<1 kPa, the residencetime for the maximum yield of butadiene increases sharply.These results are illustrated in Figure 20.

Figure 20 is a synthesis of Figure 19 and shows that theresidence time associated with the maximum yield of butadienefalls to a minimum and then increases. As the oxygen partialpressure is decreased further below 0.25 kPa, the maximum yieldof butadiene obtainable from an IMR tends asymptotically to a

value of 0.8. However, to attain this value, residence times haveto be increased dramatically.

Figure 21 shows the effect of oxygen partial pressure onbutane selectivity for the maximum yield of butadiene. Thereis a wide variation of selectivities over the range of partialpressures from 0.24 at 85 kPa to 0.70 at 0.25 kPa.

To conclude our analysis of the production of butadiene, weinvestigated the circumstances at which the maximum yieldsfrom a FBR and an IMR are equivalent. An examination ofFigures 6 and 18 shows that, at high oxygen partial pressures,a greater yield of butadiene is obtained from a FBR than froman IMR and that, at low oxygen partial pressures, the converseis applicable. The critical value of oxygen partial pressure wasfound to be 50 kPa. At this pressure and greater, the maximumyield of butadiene is greater from a FBR than from an IMR.Below 50 kPa, the maximum yields of butadiene are greaterfrom an IMR. The greatest percentage difference between themaximum yields of butenes, 36%, is at an oxygen partialpressure of 70 kPa.

Table 4 shows the respective values at oxygen partialpressures close to 50 kPa.

The question as to what yield of butadiene could be attainedat a very low oxygen partial pressure and a reactor of very largesize was answered by defining a very low oxygen partialpressure as 0.000 001 kPa, and the results are shown in Figures22 and 23.

The maximum yield of butadiene at a very low oxygen partialpressure is 0.800. At this point, the initial butane feed has beendepleted. The butane-butadiene profile in Figure 22 is concaveover its entire length, and the maximum selectivity is given bythe slope of the line from the feed point (1, 0) to its point oftangential contact with the profile (0.8, 0).

Figure 23 shows that the residence time at this very lowoxygen partial pressure for the total oxidation of butane is 5.6

Figure 18. Profiles of butane and butadiene at constant oxygen partialpressures from 85 to 0.25 kPa in an IMR.

Figure 19. Residence times for butadiene at constant oxygen partialpressures from 85 to 0.25 kPa in an IMR.

Table 3. Comparison of Maximum Yields of Butenes from an IMRand a FBR at Different Oxygen Partial Pressures

IMR FBR

oxygenpartial

pressure,kPa

maximumbutenes

associatedbutane

residencetime, s

maximumbutenes

associatedbutane

residencetime, s

36 0.105 0.631 11 0.102 0.740 1037 0.105 0.634 11 0.103 0.732 1038 0.105 0.637 11 0.104 0.724 1039 0.104 0.640 11 0.105 0.717 1140 0.104 0.642 11 0.106 0.709 1141 0.104 0.628 12 0.107 0.701 11

Figure 20. Residence times for the maximum yield of butadiene at constantoxygen partial pressures from 85 to 0.25 kPa in an IMR.

Figure 21. Selectivity of butane to butadiene in an IMR as a function ofoxygen partial pressure for conditions of maximum yield of butadiene.


× 107 s. As has already been noted, for an IMR at a constantoxygen partial pressure of 0.25 kPa, the maximum yield ofbutadiene is 0.665 with a residence time of 322 s (residualbutane at this maximum yield of butadiene was 0.042). Thisrepresents an achievement of 83% relative to the theoreticalmaximum butadiene yield of 0.800.

For a FBR with an initial oxygen partial pressure of 70 kPaand in which the oxygen is not replenished, the maximum yieldof butadiene is 0.183 (see Figure 6). This represents anachievement of only 23% relative to the theoretical maximumbutadiene yield of 0.800. Residual butane concentration was0.399.

As well as the FBR profile for 70 kPa, Figure 24 also showsthe butane-butadiene profiles for an IMR in which the originaloxygen partial pressures (0.25 and 0.000 001 kPa) are main-tained constant along the length of the reactor. It is significantthat the butane-butadiene profiles considered in Figure 24(depleted oxygen at 70 kPa and constant oxygen at 0.25 kPa)all lie below the profile for a very low oxygen partial pressure.

We have commented upon the concave shape of all thebutane-butadiene profiles so far identified. The significance

of a concavity is that, in these instances, it can be removedgeometrically by a straight line from the feed point that istangential to the profile. This is akin to taking fresh feed andmixing it with reactor products at the tangent point. The tangentline, therefore, represents the locus of all possible mixingconfigurations. Consequently, we can extend the area beneaththe theoretical butane-butadiene profile by drawing the tangentfrom the feed point (point A) to the curve (point B).

We believe that the resulting expanded area represents acandidate AR for the system butane-butadiene in the subspaceshown. In terms of normal AR theory, it might be thought thatthe reactor configuration necessary to attain this candidate ARis a continuous stirred-tank reactor (CSTR) from point A topoint B followed by an IMR from point B. This is not correct,because Figure 25 is but a projection from the full space andonly those reaction vectors in the subspace are collinear withthe mixing vectors in the subspace.

Butadiene Yields

The best yields of butadiene from the reactor configurationsstudied were compared with the theoretical best yield ofbutadiene of 0.800 from an IMR of very large size. Ranked inorder of their closeness to the theoretical best yield, the resultsfrom the reactor configurations are shown in Table 5. FromTable 5 it is concluded that an IMR with a residence time of322 s operating under a constant oxygen partial pressure of 0.25kPa gives a maximum butadiene yield of 0.665 carbon massfraction, which is 83% of the theoretical maximum yield of0.800.

Figure 22. Profile of butane and butadiene at a very low oxygen partialpressure and in a very large IMR.

Figure 23. Residence times for butadiene production at a very low oxygenpartial pressure and in a very large IMR.

Table 4. Comparison of Maximum Yields of Butadiene from anIMR and a FBR at Different Oxygen Partial Pressures

IMR FBR

oxygenpartial

pressure,kPa

maximumbutadiene

value


residencetime, s

maximumbutadiene

value


residencetime, s

48 0.138 0.440 33 0.130 0.643 1449 0.137 0.438 34 0.133 0.634 1550 0.137 0.436 35 0.136 0.625 1651 0.137 0.440 35 0.140 0.616 1652 0.137 0.438 36 0.143 0.606 1853 0.137 0.436 37 0.146 0.597 18

Figure 24. Profiles of butane and butadiene at different oxygen partialpressures for an IMR and a FBR.

Figure 25. Profile of candidate AR for the system subspace butane-butadiene.


Effect of the Temperature upon the Yields of Butadiene.All the analyses conducted have been at the isothermaltemperature of 773 K,5,6 and consequently, our candidate ARshown in Figure 25 is applicable only at that temperature. Figure26 shows the effect of temperature upon the butane-butadieneprofile in a very large IMR when the oxygen partial pressure isvery low.

Examination of Figure 26 shows that increasing the reactortemperature from 773 to 823 K increases the maximumtheoretical yield of butadiene from 0.800 to 0.925. Themaximum selectivity of butane at 823 K is 0.925. Decreasingthe operating temperature from 773 to 748 K decreases themaximum theoretical yield of butadiene from 0.800 to 0.661.The maximum selectivity at 748 K is 0.661

From Figure 26, we conclude that the theoretical maximumyield of butadiene and the selectivity of butane increase withtemperature over the range 773 to 823 K. The maximum yieldand selectivity decrease as the temperature is reduced from 773to 748 K.

Scenario 2, Case 5: Replenishment of Oxygen in an IMRsProduction of Butenes and Butadiene.Finally, we answeredthe question as to what was the maximum combined yield ofbutenes and butadiene from an IMR operating at a constantoxygen partial pressure. Figure 27 shows the profiles for butenes,butadiene, and butenes plus butadiene as a function of butaneconcentration. The constant oxygen partial pressure was 85 kPa.Whereas the profile for butadiene shows a concave section andthe profile for butenes does not, the profile for butenes andbutadiene is convex over its entire length.

Figure 28 shows the IMR residence time profiles for butenes,butadiene, and butenes plus butadiene at an oxygen partialpressure of 85 kPa. The residence time for the maximum yieldof butenes plus butadiene, 77 s, is greater than that for butenes(41 s, Table 1) and less than that for butadiene (108 s, Figure19).

Figure 29 shows the IMR concentration profiles for butenesplus butadiene as a function of butane concentration at constant

oxygen partial pressures from 0.25 to 85 kPa. At an oxygenpartial pressure of 0.25 kPa, the maximum yield of butenes plusbutadiene is 0.677 with a butane selectivity of 0.716 and aresidence time of 307 s. The corresponding residence times atthe same oxygen partial pressure are 75 s (butenes, Table 1)and 322 s (butadiene, Figure 19).

At oxygen partial pressures of 15 kPa and less, a concaveregion exists in the profiles at low values of butane concentra-tion. These regions could be extended by using a CSTR in serieswith the IMR. Figure 30 exhibits the same pattern noticed inFigure 19, namely a drop in residence time for the maximumyield of butenes plus butadiene from 77 s at 85 kPa to aminimum of 17 s at 15 kPa. At oxygen partial pressure<15kPa, the residence times for the maximum yield of butenes plusbutadiene increases to 307 s at 0.25 kPa. For the reason ofclarity, the 15 kPa profile has been omitted from Figure 30.Over the range of oxygen partial pressures studied, the greatestselectivity of butane to butenes and butadiene combined was

Figure 26. Effect of temperature upon the theoretical maximum yield ofbutadiene.

Table 5. Best Butadiene Yields from an IMR and a FBR RankedAccording to Their Closeness to the Theoretical Maximum Yield ofButadiene

source

maximumbutadiene

yield

associatedbutaneyield

residencetime, s

% of max.theoreticabutadiene

yield

oxygenpartial

pressure,kPa

reactorconfiguration

Figure 18 0.665 0.042 322 83% 0.25 IMRFigure 18 0.534 0.112 138 67% 0.70 IMRFigure 6 0.183 0.399 41 23% 70.0 FBR

Figure 27. Profiles of butenes, butadiene, and butenes plus butadieneagainst butane at a constant oxygen partial pressure of 85 kPa in an IMR.

Figure 28. IMR residence times for butenes, butadiene, and butenes plusbutadiene at a constant oxygen partial pressure of 85 kPa.

Figure 29. IMR profiles for butenes plus butadiene against butane atconstant oxygen partial pressures.


0.72 at an oxygen partial pressure of 0.25 kPa and the leastwas 0.46 at an oxygen partial pressure of 85 kPa.

Scenario 3: Extension of the Attainable RegionsTwoIMRs in Series.Our previous studies of AR systems7 have ledus to expect that filling in a concave region through a processof bypass and mixing sometimes can result in a further extensionof the AR by feeding this mixture to another IMR. Referringto Figure 25, were an IMR to be added to the process flowdiagram with a feed taken from any point on the line AB, itmight be possible to extend the AR beyond the line AB.However, in a two-dimensional subspace, it is not alwaysapparent that the region can be extended. To establish whethera region can be extended, it would be necessary to considerhigher-dimensional profiles.

The reactor configuration for such an arrangement in the two-dimensional subspace is shown in Figure 31. The reactorconfiguration studied in Figure 31 was a very large IMRfollowed by a second equally large IMR. In this configuration,the output from IMR1 (i.e., point B in Figure 25) is mixed withbutane (point A in Figure 25) in the volumetric ratio ofq/(1 -q). The constant oxygen partial pressure in each IMR was0.000 001 kPa.

In Figure 32, we show that such an arrangement has notresulted in any further extension of the AR beyond the tangentline AB. By taking a range of mixtures from the first IMR withvarying mixing ratios,q, of final products and fresh reactantsand supplying each feed to a second IMR, we have shown thateach of the resulting butane-butadiene profiles lies whollybeneath the tangent line AB.

In Figure 25, the area enclosed by the straight line AB, thex-axis, and they-axis represents the boundaries of the regionwithin which all scenarios so far identified lie. Consequently,we believe that Figure 25 represents a candidate AR for thesystem subspace butane-butadiene.

Conclusions

The best yield of butenes identified in this study from areactor of finite size is slightly less than 0.119 with acorresponding residence time of 75 s. This yield of butenes

represents almost 100% of the theoretical maximum quantityfrom an IMR of very large size with a very low oxygen partialpressure. The reactor configuration for this example was an IMRwith a constant oxygen partial pressure of 0.25 kPa.

A candidate AR has been identified for the system subspacebutane-butenes at an operating temperature of 773 K. Thiscandidate AR is shown in Figure 13.

In a realistically sized reactor, the best yield of butadieneidentified in this study is 0.665 with a corresponding residencetime of 322 s (Figure 18). This yield of butadiene represents83% of the theoretical quantity from an IMR of very large sizewith a very low oxygen partial pressure. The reactor configu-ration for this example was an IMR with a constant oxygenpartial pressure of 0.25 kPa.

A candidate AR has been identified for the system subspacebutane-butadiene at a temperature of 773 K. This candidateAR is shown in Figure 25.

In the ODH ofn-butane, an increase in temperature reducesthe maximum yield of butanes. A reduction in temperatureincreases the maximum yield of butanes. In the ODH ofn-butane, an increase in temperature increases the maximumyield of butadiene. A reduction in temperature reduces themaximum yield of butadiene.

The maximum yield of butenes plus butadiene found was0.677 with a butane selectivity of 0.716. The reactor used wasan IMR with a constant oxygen partial pressure of 0.25 kPa.The residence time was 307 s.

Nomenclature

C ) carbon mass fraction of speciesi

Ci0 ) initial carbon mass fraction of speciesi

ri ) rate of reaction of reactioni, mol/kg sSi ) conversion selectivity of speciesi

Literature Cited

(1) International Network for Environmental Compliance and Enforce-ment, Washington, DC, USAAnon, Industrial Processes. Web sitewww.inece.org/mmcourse/chapt1.pdf.

(2) Milne, D.; Glasser, D.; Hildebrandt, D.; Hausberger, B. Applicationof the Attainable Region Concept to the Oxidative Dehydrogenation of1-Butene in Inert Porous Membrane Reactors.Ind. Eng. Chem. Res. 2004,43, 1827-1831.

(3) Tellez, C.; Menendez, M.; Santamarıá, J. Kinetic Study of theOxidative Dehydrogenation of Butane on V/MgO Catalysts.J. Catal. 1999,183, 210-221.

Figure 30. IMR residence times for butenes plus butadiene at constantoxygen partial pressures.

Figure 31. IMR series configuration.

Figure 32. Butane-butadiene profiles from two IMRs in series.


(4) Tellez, C.; Menendez, M.; Santamarıá, J. Simulation of an IMR forthe Oxidative Dehydrogenation of Butane.Chem. Eng. Sci. 1999, 54, 2917-2925.

(5) Tellez, C.; Menendez, M.; Santamarıá, J. Oxidative Dehydrogenationof Butane using Membrane Reactors.AIChE J.1997, 43 (3), 777-784.

(6) Assabumrungrat, S.; Rienchalanusarn, T.; Praserthdam, P.; Goto, S.Theoretical Study of the Application of Porous Membrane Reactor toOxidative Dehydrogenation ofn-Butane.Chem. Eng. J. 2002, 85, 69-79.

(7) Glasser, D.; Hildebrandt, D.; Crowe, C. A Geometric Approach toSteady Flow Reactors: The Attainable Region and Optimisation inConcentration Space.Am. Chem. Soc.1987, 1803-1810.

ReceiVed for reView February 1, 2005ReVised manuscript receiVed January 18, 2006

AcceptedJanuary 31, 2006

IE050120L


Documents

The Oxidative Dehydrogenation of n- Butane in a Fixed-Bed Reactor and in an Inert Porous Membrane ReactorMaximizing the Production of Butenes and Butadiene