sonne mans 73 mechanism 2

7/30/2019 Sonne Mans 73 Mechanism 2

1/11

JOURNAL OF CATALYSIS 31, 220230 (1973)

The Mechanism of Pyridine Hydrogenolysison Molybdenum-Containing Catalysts

II. Hydrogenation of Pyridine to PiperidineJ. SONNEMANS, G. H. VAN DEN BERG, AND P. MARS

Twente University of Technology, Enschede, The Netherlands

Received March 12, 1973The kinetics of pyridine hydrogenation was studied at high hydrogen pressures

on a MO-Al oxide and a Co-MO-Al oxide catalyst. The rate equation was foundto be T = kP,,rP~2/Pp,,0, in which n is 1.5 at 300 and 375C and 1.0 at 250C. Thisrate equation can be derived assuming strong adsorption of pyridine and itsproducts with identical adsorption constants.

The (hydrojcracking of piperidine appears to have a low order in hydrogen,probably lower than 0.5.

The adsorption behavior of nitrogen bases and hydrogen on alumina and themolybdenum-containing catalysts was investigated by the gas chromatographicmethod. The adsorption of the nitrogen bases appeared to b.e very strong on bothcatalysts, and varied in the order piperidine > pyridine > ammonia.

Hydrogen also showed a strong adsorption. Hydrogen and nitrogen bases ap-peared to adsorb on different sites.

In Part I (1) the preparation of a mono-layer molybdena-alumina catalyst wasdiscussed. Its activity for cyclohexane de-hydrogenation and pyridine hydrogenolysiswas compared with a cobalt-promotedmolybdena-alumina catalyst. For the pyr-idine hydrogenolysis the product distribu-tions were investigated as a function ofthe temperature. The same activity andselectivity pattern was found for bothcatalysts. This paper deals with the firststep in the reaction scheme, namely thehydrogenation of pyridine. The reactionscheme can be simplified for this case into

kl kzpyridine + piperidine -+ consecut,ive products.Much work has been performed on thehydrogenolysis of thiophene in order tostudy the mechanism of the desulfuriza-tion process. However, only a little workhas been published about the hydrogenol-

ysis of pyridine and the mechanism of thehydrodenitrogenation process. hIost hydro-denitrogenation studies are carried out tostudy either the overall kinetics of theprocess or the difference in the rate ofnitrogen removal of several types of nitro-gen bases (2-5). Some authors investigatedthe hydrodenitrogenation process with theaid of model compounds like quinoline orpyridine (6-9). The catalysts used in thesestudies are mostly molybdenum-containingcatalysts, e.g., the COO-Mo03-ALOcatalyst.The overall order in nitrogen was foundto be unity by many authors. This wasthe case for oil fractions containing differ-ent nitrogen compounds and for modelcompounds, mostly diluted with an or-ganic solvent. The first order rate con-stant, however, appears to be dependenton the nitrogen content of the feed (3, 9).McIlvried (8) gives an explanation forthis phenomenon by taking into account220Copyright @ 1973 by Academic Press, Inc.

All rights of reproduction in any form reserved.


2/11

HYDROGENOLYSIS OF PYRIDIKE 221the strong adsorption of the reaction prod-ucts on the catalyst. He studied the con-version of pyridine and piperidine on asulfided catalyst containing nickel, cobaltand molybdenum on alumina. The nitro-gen compounds were diluted with xylene(nitrogen level: 100-4000 ppm, P,,) = lo-5 X lo4 N/m). McIlvried showed thatthe conversion of piperid,ine could be de-scribed by an equation similar toreaction rate

bgipfpip= corlstant 1 + bxylPxyl + bpipPpigoin which PDipa = piperidine partial pressureat the reactor inlet

bXsi, b,i, = adsorption constants forxylene and piperidine.(The term 1 + b,,,P,,, in the denominatorwas represented as a constant in the equa-tion given by McIlvried.)This equation can be derived by assum-ing equally strong adsorption of piperidineand its products on the catalyst. NIcIlvriedshowed that the first order rate constantsobtained by other authors were also in-versely proportional to the nitrogen con-tent of the feed (or parGal pressure inthe reactor).The rate equation given above couldnot describe the conversion of pyridine.McIlvried concluded that equally strongadsorption of t,he nitrogen bases may oc-cur on the cracking sites but not on thehydrogenation sites. Relatively very strongadsorption of ammonia on the hydrogena-tion sites could explain the results for thepyridinc conversion.The reaction order in hydrogen is notwell known. Regarding the order of thedenitrogenation reaction in respect to hy-drogen, orders of I and 2 were reported(4-5). It is not rlcar whether these reac-tion orders applv for the hvdrogenationreactions or for the (hydra) crackingreactions.WC investigated the hydrogenation ofpvridine, ring opening and dehydrogena-tion of piperidine and the cracking anddisproportionation of different amines, toobtain more information about the kinetics

of the different reaction steps. The resultsof the hydrogenation of pyridine are pre-sented in this paper together with the re-sults of the adsorption experiments.METHODS

Kinetic MeasurementsThe apparatus used for kinetic measure-ments has been described elsewhere (10).Besides the U-type reactor we used a stain-less steel reactor of a smaller size (I-type).The diameter was also 6 mm. A separateexperiment showed that the walls of thereactor did not show any important ac-

tivity for the pyridine hydrogenation. Thepyridine partial pressure at the reactorinlet and the composition of the productswere analyzed by means of gas chromatog-raphy.The reaction time t is defined as: t =?~LP/+~ in which m = mass of catalyst(kg) ; P = total pressure (N/m) ; +t =total moles fed to the reactor (moles/set).Adso .ption Measurements

Adsorption measurements were carriedout by the method of Roginskii (11). Boththe pulse method and the frontal methodwere applied. For the pulse experiments astainless steel reactor (di = 4 mm) wasused. A thermal conductivity detector andrecorder were used for a continuous regis-tration of the composition of the outputgas flow. For the frontal method a U-typeglass reactor (di = 9 mm) was used. Thebreakthrough curves were obtained bytaking samples at the reactor outlet andanalyzing them bv gas chromatography.In both methods the temperature of thecatalyst (l-3 g) was constant within lC,both as a function of the time and the bedlength. The carrier gas used in the experi-ments was either hydrogen or argon, de-oxidized by a copper catalyst and dried bymolecular sieves. The gas flow rate was1.44 or 3.6 ml/see. A constant partial pres-sure of pyridine in the carrier gas wasobtained with the aid of two saturators.The second one was thermostated. Am-monia was added to t,he carrier gas froma cylinder. A constant gas flow rate of


3/11

222 SONNEMANS, VAN DEN BERG, AND MARSammonia was obtained by maintaining aconstant pressure drop over a capillary.The partial pressures of the adsorbateswere determined by titration of the am-monia and by freezing out and weighingthe pyridine; the accuracy was 2-370.Materials

The catalysts used were: r-Al,O, (213m/g), 4% COO-12% MOO,-A1,Oj (Ketjen-fine, 235 m/g) and 22% Moo,--ALO, (195m/g). The y-Al,O, catalyst contained 1%SiO,, 0.8% SO,, 0.06% Na,O and 0.02%Fe. (The alumina (and Ketjenfine) werekindly supplied by Ketjen Amsterdam.Analysis data obtained from Ketjen.) The22% Moos-Al,O,catalyst was prepared byslow adsorption of molybdate on y-aluminaat pH 1. The MOO,-Al,O, catalyst is con-sidered to have a very extensive MoOasurface and a low ALO, surface (1). In allthe experiments the catalysts were pre-treated for more than 16 hr at 450C inhydrogen.

RESULTSKinetics of Pyridine Hydrogenation

The reaction order in pyridine of thehydrogenation was determined by chang-ing the reaction time and the initial pyri-dine partial pressure. A first order in pyri-dine was found by changing the reactiontime. This was clearly demonstrated atthree temperatures, viz, 270, 310 and

350C (12). Experiments at constant reac-tion time and different initial partial pres-sures of pyridine showed a serious devia-tion of the first order behavior. Anincreasing conversion of pyridine wasfound at decreasing partial pressures ofpyridine (Table 1). The first order con-stant k, of the pyridine hydrogenation ap-pears to be inversely proportional to thepyridine pressure at the reactor inlet(Table 1).Next to pyridine and piperidine, am-monia and N-pentylpiperidine were ob-served in the products. The last two wereformed in more or less equimolar amounts.Pentylamine was not found. This indi-cates that the most important consecutivereaction is not the formation of pentyl-amine by the opening of the piperidinering, but the disproportionation of t,wopiperidine molecules into ammonia andIl-pentylpiperidine (10). Assuming an ap-parent first order reaction in piperidinethe reaction rate constant k, was calcu-lated (Table 1). This rate constant alsoincreases with decreasing initial partialpressure of pyridine.The order in hydrogen of the pyridinehydrogenation was determined at 250, 300and 375C. By multipIying the apparentfirst order (rate constants k, and Ic, bythe initial partial pressure of pyridine, thecorrected rate constants k, and Ic, wereobtained. The composition of the productsand these rate constants are presented in

TABLE 1CONVERSION OF PYRIDINF: ON THE COO-Mo03-Al& CATALYST IS .i FKNCTIONOF THE INITIAL PYRID~W; P.IRTIA~~ PRESSURK-

Seqllenre of experiments: 7 2 6 5 4 3 8P pylD 105N/m2)Pyridine (mole yc)Piperidine (mole %)Consecutive products (mole To)kIh(10-8 m* moles/kg N set)klP,,,,(10-3 moles/kq see)kJ(lO-8 m2 moles/kg N see)

1.28 0.87 0.81 O.i664 Fi3 49 4832 39 42 464 8 9 12

0.73 1.05 1.16 1.40.94 0.91 0.94 1 .050.35 0.61 0.56 0.69

0.72404713

1.51.060.72

0 57 0.4937 3350 5413 13

16 1.80 91 0.880.65 0.63

10.28

125929

3.50.971.04

a 7 = 300C: PH% = 60 X loj N/m2, t = 6.1 x 107 kg N sec/mP moles.kl k2b The reaction scheme used was: pyridine ti p iperidine -b conwcuiive products.


4/11

TE2

C

ROOPDNC

A23

A

3CAAF

O

OTHH

P

Po

cmon

Kecas

Jeoc

o

C

v

kP

=k

kPruo=k

Iu

PWT

Pdn

Ppdn

pos

(me

(me

(oNm

(ONW)

(me4C

(me7

(me5%)

ks

ks

!z

7=2C

1

07

9

7

13

58x1h

24Xo

3

75g2MOA~O

3

05

8

1

3

12x1d

19x1d

g

4

05

6

3

4

22x1d

11xlo

8

t=1OX1OkNsmme

6

04

5

4

4

23x1d

07x14

2

2=3C

1

10

8

1

24

35x1

09xIO3

35gCM

AO

3

10

6

3

6

2

11X1S

08x1S

r

4

10

4

5

7

18x1a

06xlo

$

t=45X1kNsmme

6

10

2

5

1

31X1S

12x1S

7=35%

2

15

03

9

23

27

19X13

35x1Z

?

01g2MOAJ&

3

03

8

1

8

56Xo

28x1Z

t

4

04

7

1

8

92x1a

20x1Z

t=14X1VkNsmme

6

03

6

3

1

15X1Z

14X1Z

6

05

6

2

0

19X1Z

21X1Z

75

03

3

4

2

24x1Z

14x1Z

uTemsa3a3Cswh#

raecasthnmyoan

ah

emaue.Ts

ohemswa

35C631545761Nma3OC46316oNmaa2C31461Nm

Inhsea2Canem

daereuo

ohcaya4Cwn

yboethE

a4X1Nmtoadthne

oavydnothdemno

oheood


5/11

224 SONNEMANS, VAN DEN BERG, AND MARSTable 2. The order in hydrogen of the ringhydrogenation was calculated by means ofthe least squares method. With a 90%confidence level the following values wereobtained: 1.06 + 0.40 at 25OC, 1.54 +0.21 at 300C and 1.57 -C 0.15 at 375C.There may be a possibility of 5% that thereaction order in hydrogen is the same atthe three temperatures. However, takinginto account the sequence of the experi-ments at 250C a lower order in hydrogenfor the pyridine hydrogenation at 250Cmay be concluded. Other experiments sup-port this conclusion.The hydrogen pressure appears to havelittle effect on the rate of the piperidinehydrogenolysis. At 300 and 375C a moreor less conslant value for k, was obtained.At 259C the lc, even seems to decreasewith increasing hydrogen partial pressure.So it may be concluded that the order inhydrogen of the opening of the piperidinering is about zero. However, only a roughestimate of the reaction order was possibledue not only to the inaccuracy of the de-termination of the partial pressures of con-secutive products but also because theconversion of piperidine turns out to berather complex. For instance, piperidine isformed again from one of its products,namely N-pentylpiperidine (101. More ex-periments have to be carried out in orderto formulate in detail the kinetics for thepiperidine conversion.

Adsorption MeasurementsNitrogen Bases

We were not able to determine the iso-therms for the adsorption of the nitrogenbases integrally due to several reasons.Gravimetric measurements showed alreadya maximum coverage of pyridine, on forinstance the MOO,-ALO, catalyst, at par-tial pressures of less than lo3 N/m2 (7 mmHg). The pulse technique described byRoginskii (11) also failed due to the strongadsorption of the nitrogen bases, even attemperatures of 450C. Moreover, severalnitrogen compounds like piperidine orpentylamine reacted above 250C.

With the pulse technique some qualita-tive results were obtained. An estimate ofthe heat of adsorption on the molybdena-alumina catalyst was possible by compar-ing the retention time for pyridine at450C and benzene at about 200C. Theratio AHPYr.ads/AHbenzene.ads as found tobe 1.5. From adsorption isotherms a valueof 14 kcal/mole was calculated for theAH benzene.ads.ence, the heat of adsorptionfor pyridine will exceed 20 kcal/mole.Hydrogen appeared to have little effect, ifany, on the adsorption of pyridine and am-monia on the MOO,-Al,O, catalyst. Iden-tical gas chromatographic curves were ob-tained using either hydrogen or heliumas carrier gas.With the aid of the frontal technique(11) the number of moles of nitrogenbases adsorbed per square meter (u) wasdetermined for the different catalysts. Theresults are given in Table 3. The valuesare characteristic for the maximumamounts chemisorbed (except for the valueu = 1.63 at PNH = 2.1 mm Hg). However,only a part of the surface is covered bythe nitrogen compounds (lO-j moles/m* =160 AZ/molecule). The total number ofmoles adsorbed per square meter appearsto be the same for ammonia and pyridine.(Preliminary experiments carried out withpiperidine showed that its adsorption is ofthe same order, viz, 2.4 X 10eGmole/m2 at225C on the 22% MOO,-Al,O, catalyst.)The amount adsorbed on the molybdena-alumina catalyst is about twice as high asfor the alumina. The temperature did notappear to be an important variable forthe amounts adsorbed. At a pyridine par-tial pressIre of 3.75 mm Hg the valuesfound at 350C were only 10% less thanat 300C for all catalysts. At 400C we ob-tained for the pyridine adsorption on theKetienfine catalyst the values 0.93 x 1O-6moles/m* at 3.75 mm Hg and 1.15 x 1O-6moles/m* at 8.7 mm Hg. Hence, even athigh temperatures maximum adsorption ofthe nitrogen bases takes place a t low par-tial pressures of the bases.

At the end of the adsorption experi-ments some information about the desorp-tion was obtained by flushing the catalyst


6/11

HYDROGENOLYSIS OF PYRIDIXE

TABLE 3ADSORPTIONFPmm: AND hn1hf0~1.1ONDIFFERENT.ITM,TSTSu 3OOV

225

Pyridine r\TH3- -

(10-6 rIGles,m)P (1OP mZles/mz) P(mm Hg) (mm %I

hi203 1 .02 3.75 0.x.5 4.00.95 x.7 0. k;: 6.25coO-?\loOs-.41&j 1.21 3.75 -22;; hlOO$ A1rOa 1.78 3. 7.3 1 6:: 2.11.81 8.7 1.90 4.9

u The reproduc ibilit,?: is ;;,I. 111 comparing different catalysts the accrrracy of the swface area determina-tion has to be taken into awcrunt (37;).with pure hydrogen. The amounts of pyri-dine or ammonia still adsorbed after dif-ferent times of desorption are given inTable 4. The table shows that a part ofthe nitrogen compounds desorbed ratherfast; a part did not desorb, even afterflushing the catalyst for 18 hr. The valuesfor the Ketjenfine catalyst are in betweenthose for alumina and the monolayerMOO,-ALO., catalyst. Similar experiments,carried out at higher temperatures, showeda decrease of the amounts adsorbed ; forthe COO-MOO,-Al,O, catalyst the values0.65 and 0.3 were obtained for desorptionof pyridine at 400C during 30 min and18 hr, respectively. It may be stated thata part of the pyridine is irreversibly ad-sorbed. However, discussion about irre-versible adsorpt,ion appears to be some-what arbitrary because Fransen (13) inour laboratory has observed by ir spec-troscopy that. already at low temperaturesdesorption of the strongly chemisorbed

molecules took place when other nitrogencompounds were present in the gasphase (Isj .With the aid of the frontal method wealso investigated the competitive adsorp-tion of the nitrogen bases by supplying amixture to the catalyst and analyzing thebreakthrough curves. A typical curve isgiven in Fig. 1, from which the ratio of themean b values of ammonia and pyridinewas calculated assuming a Langmuir typeof adsorption. The ratio of t,he amounts ofpyridine and piperidine adsorbed equalsbpy,P,y,./bpipPp,p. The results are given inTable 5. Pyridine showed a stronger ad-sorption than ammonia on both aluminaand the molybdena-alumina catalyst Thisstronger adsorption was found at 300 and400C indicating that the mean heats ofadsorption of both compounds only differslightly. Piperidine, in its turn, showed astronger adsorption than pyridine. The ad-sorption was measured at lower tempera-

TABLE 4IVIOLICS OF PYRIDINK AND bk~~~~~I~ STILL ADSORHED AFTER I)IFFEKBNT Tr~~ss OF I~ESOKPTJOW

gpYridine(lOWmoles/m2) u~~~(10-~ moles/m2)Time of desorption : 0 min 20 min* 18 hrc 0 min 20 min 18 hrc

Al& 1.0 0.6 0.3, KS:, 0.5 0. li,Coo-Moo,-Al,03 1.2 1 .(I 0.522yc MoOa-AlrOB 1.x 1.5 1.0 1.9 1.45 0.9*

6,Hs, 3.6 liters!hr, rl = SOOC.*After a desorption time o f 20 min the rate of desorption was below the detection level (0.04 mm Hg).c Analyzed by the Tkrmas-method.


7/11

226 SONNEB'IANS,AN DEN BERG,AND MARS

5P(mm Hg)

I3

1 (mV)

FIG. 1. Breakthrough curves obtained by supply-ing continuously both ammonia and pyridine to1 g of the 229& Mo03-A1203 catalyst (T = 3OOC,+Ar = 0.45 ml/set); (O)pyridine, (0) ammonia.tures to avoid the decomposition of thepiperidine.Hydrogen Adsorption

Hydrogen chemisorption was also investi-gated with the pulse technique. Mostly thistechnique is used to determine adsorptionisotherms, e.g., by Roginskii (11) ; theestablishment of the adsorption equilib-rium is one of the most essential points inthis method. We demonstrated that thispulse method also can be used to determinethe rate of adsorption, if the adsorption isfar from equilibrium. The method is semi-quantitative and has the advantage, incomparison with the usual volumetrictechnique, that the effect of other com-pounds, e.g., nitrogen bases, on the rate ofhydrogen chemisorption can easily be in-vestigated. In a future paper we intend topublish details of this method.

FIG. 2. Breakthrough curves obtained after in-jection of pulses of 0.76 ml hydrogen on 2.73 gcatalyst iT = 350C; +.kr = 1.0 ml/set; p.4, =lo5 N/mZ); (-) 4y0 COO-AlTOa; (---) 4% CoO-12% Mo0,-.41?03.

Typical curves obtained after injectionof hydrogen are given in Fig. 2. The CoO-Al,O, catalyst did not show adsorption ofthe hydrogen: injection of helium gave thesame type of curve. The curve for theCOO-Moos-Al,O, catalyst showed a lowerpeak height indicating that adsorption ofhydrogen takes place. The decrease inpeak height is related to the rate ofchemisorption.By varying the temperature, the volumeof the hydrogen pulse, and the composi-tion of the carrier gas, the following re-sult,s were obtained. Above 250C the

TABLE 5RELATIVE ADSOWTION CONSTANTS OF THE NITROGEN BZSE:S

Temp (C)Catalyst 245 300 400

bpyridinelbNHi AM% 4.7bpyritcxJbNH~ 223, Mo03-A120$ 4 3 4.2bviperidine/bpyridine 22y0 Mo03-Al203 6.1bpiperidinelbpyridine 229& MOOS-A1203 6.P

a In this case the method was slightly different. A mixt,ure of pyridine and piperidine was added to thecatalyst which already cont,ained the maxim um amount of piperidine adsorbed. Desorpt,ion of a part ofthis piperidine took place. The t.emperature was 225C.


8/11

HYDROGENOLYSIS OF PYRIDINE 227chemisorption of hydrogen was observed ;the rate of chemisorption increases withincreasing temperature. The hydrogenchemisorption was found to have a loworder in hydrogen (> 1 the equation may be simpli-fied to:(4)

In Fig. 3 the data from Table 1 havebeen plotted as t/in (P.,,JP,k) as a functionof PA, in order to test whether Eq. (3) or(4) applies for the pyridine hydrogenation.It appears that the intercept practicallyequals zero. Hence, the ring hydrogenationof pyridine can be described by Eq. (4).The kinetics support the supposition of theequally strong adsorption of pyridine and

FIN. S. Plot to test F,q. (a) for the pyridine hydrogenation (see also Table I); (-) line calculated bymeans of the least, squares method; (p. m) line calculated by means of Eq. (2) assuming bB/bA = 2; (- _ -)line calculated by means of E:q. (2) asstrming ba/b.k = 5; (-4 = pyridine, B = piperidine; bAPA >> 1: thereaction rate constant was calculated from the point at P,,,, = 1.28 X 10 N/n?).


9/11

228 SONNEIIANS, VAN DER BERG, AND MARSpiperidine. ,Figure 3 shows that curvedlines have to be expected when the adsorp-tion constants differ. The figure suggeststhe ratio bpi,/b,,, to be smaller than two.

McIlvried showed that an equationsimilar to Eq. (3) very well describes theconversion of piperidine. A less strong ad-sorption of the nitrogen bases on his cat-alyst compared with our catalyst mightbe concluded.The adsorption experiments can giveevidence on the application of Eq. (3) or(4). Table 3 shows that at 300C and apartial pressure of pyridine or ammoniaof 24 mm Hg (500 N/m2) at least 90%of the maximum adsorption was obtained ;this means that bNebases 10m2m2/N, if aLangmuir type adsorption is assumed. Thekinetic experiments were carried out at apyridine partial pressure of >104 m2/N.bPYI.PBY,.mounts therefore to >lO?, so theadsorption data show that in our case Eq.(4) is principally correct. The reason thatEq. (4) failed to describe the piperidineconversion of McIlvried is not due to aless strong adsorption of piperidine on hiscatalyst but to the presence of xylene inthe feed. The intercept for this case is(1 + b,,,P,,-,)/Icbpip. From his results wecalculated a value of 3 X 10e3 for the ratiob,,,/bpi,; this corresponds to a differencein the heat of adsorption of about 7 kcal/mole. We found about the same value forthe difference in heat of adsorption ofpyridine and benzene in our experiments.If the relation between the kinetics andthe adsorption data is studied in greaterdetail a problem arises. The assumptionbetween Eq. (2) and (3), namely, equalvalues for the adsorption constants of thenitrogen bases, is supported by the kinetics(Fig. 3)) but is not in accordance with theresults of the adsorption measurements(Table 5). The b value for piperidine ap-peared to be about 6 times as high as thevalue for pyridine which in its turn was4 times as high as the b value for am-monia. The ratio bpyr/bNH2 was found tobe the same at 300 and 400C ; hence thequotients of the b values may be onlyslightly dependent on the temperature.A possible explanation for the discrep-

ancy may be the following. Desorptionexperiments reported above as well as in-frared spectroscopic measurements of thepyridine adsorption (13) point to a strongheterogeneity of the catalyst surface. TheLangmuir type adsorption will thereforecertainly only be a rough approximationof the adsorption behavior of the nitrogenbases on the MOO,-Al,O, catalyst. Themolybdena surface may only be partlyactive in the pyridine hydrogenation.Till now we did not discuss the resultsMcIlvried obtained for the pyridine hy-drogenation. He observed a deviation ofthe first order kinetics for this reaction(Fig. 3 in Ref. (9) ; P,,,.O was kept con-stant). He concluded that there was avery strong adsorption of ammonia on thehydrogenation sites of his sulfided cata-lyst, which will have separate hydrogena-tion and cracking sites. However, we didnot find evidence for a stronger adsorp-tion of ammonia than of pyridine. Thedeviation of the first order in pyridine forthe pyridine hydrogenation can be ex-plained by the fact that some of theirmeasurements were too close t,o the equi-librium of pyridine, piperidine and hydro-gen. He neglected to take into accountthe reverse reaction.The Order of the Reaction in Hydrogen

The order in hydrogen of the pyridinehydrogenation was 1.0 at 250C and 1.5at 300 and 375C. Lipsch and Schuit (14)also observed at about the same tempera-ture a change in the hydrogen order forbutene hydrogenation on COO-MoOj-Al,O,. An explanation cannot yet be given.It was shown that the observed activitydecline and the inaccuracy in the deter-mination of the equilibrium constant ofthe pyridine-piperidine equilibrium (12)did not have an important effect on theobserved order in hydrogen. The effect ofdiffusion limitation was also examined.Calculations showed that the influence ofinternal diffusion on the rate of hydro-genation is very small. Moreover, thiseffect will decrease the reaction order.For the calculation of the order in hy-


10/11

HYDROGENOLYSIS OF PYRIDINE 229drogen we assumed that the hydrogenpressure was an independent variable inthe rate equation. Investigations into thechemisorption of hydrogen and of theN-bases support this view. The ammoniaand pyridine adsorption was not affectedby hydrogen (between 0 and 1 atml . Thenitrogen bases were shown to have someeffect on the hydrogen chemisorption. How-ever, it has not to be expected that thisinterferes with the determination of theorder in hydrogen because this effect couldonly be found at very low partial pressuresof the nitrogen compounds.Preliminary results indicate that, thehydrogen chemisorption is not rate de-termining for the pyridine hydrogenationat temperatures above 250C. The relationbetween the hydrogen chemisorption andthe change in the kinetics of the pyridinehydrogenation will be subject of furtherresearch.Several problems arise when comparingour results with Merature data. First, thecatalysts are different. However, in ourlaboratory we found that the kinetics ofpyridinc hydrogenation were similar onoxidic and sulfidic catalysts (12). Secondly,the interference of other products may beimportant ; Roscnheimcr and Kiovsky (5)reported even the presence of liquid intheir reactor.The order in hydrogen reported by

Wilson, Voreck, and Malo (4) and byRosenheimer and Kiovsky (5) was ob-tained by following the decrease of thenitrogen content as a function of the hy-drogen pressure. It is not clear whether theorder applies to the ring hydrogenation ort,he ring opening reaction. Assuming areaction scheme k-1 krsH,+A=B+C,

k1n, as the order in hydrogen of the firstreaction and n2 as the order in hydrogenfor the second reaction, the following pos-sibilities may be present.

The first reaction is rate determining:order in hydrogen of the denitrogenationreaction is n,.The second reaction is rate determining:t,he dependence of the denitrogenation rateon the hydrogen pressure mill be

(k = equil. const. of the first step).Therefore experimentally an order in hy-drogen in between nz and 3 + n, may befound depending on the value of the equi-librium constant.Calculation on thc pipcridine-pyridineequilibrium showed that some of the ex-periments of the investigators of t.he de-nitrogenation process were performed at

Order in hydrogen of

Allthors nenitrogenation of PyridinePiperidine

hydrogenat,ion (hydro&acking I>enitrogenat,ionThis workMcIlvried (8)McIlvried (8)Wilson, Voreck, and

Malo (4 )Rosenheimer and

Kiovsky (5)Flinn, Larson, and

Beuther (2)

PyridinePyridineIiperidineMe-isoqllinoline in naphthaCracked diesel oilCracked light furnace oil

a Calculated by us from two experiments far from the pyridine-piperidine equilibrimn.h Calculated by us from the last, three lines of the upper section of Table 1 of McIlvrieds paper.e Calclllated by us from Fig. 3 of t.he paper of Flinn, Larson, and Beut.her (2).


11/11

230 SONNEMANS, VAN DEN BERG, AND MARSreaction conditions in which the hydro-genation equilibrium might be on the sideof the aromatics. Depending on the ratedetermining step an order in hydrogen ofn, or in between n, and 3 + n, has to beexpected.In Table 6 the orders in hydrogen aresummarized. The table shows that ourconclusions are in agreement with the re-sults of McIlvried: the hydrogenation hasa high order in hydrogen and the ringopening reaction has a low order in hydro-gen. Different orders in hydrogen for the,denitrogenation process were found. Thehydrogenation of the aromatic ring as wellas the ring opening may be rate determin-ing; in the last case the position of thehydrogenation equilibrium may influencethe rate of denitrogenation.

CONCLUSIONS1. The kinetics for the pyridine hydro-genation aredP,,r _ k fp&~sn---

clt r DYR(n = 1.0 at 250C and 1.5 at 300-375C).2. The dependence of the hydrogenationrate on the pyridine partial pressure canbe derived assuming strong adsorption ofpyridine and its products on the catalyst;the adsorption constants (b values) haveto be equal.3. Adsorption experiments support theconclusion concerning the strong adsorp-tion o f the nitrogen bases, but failed to

demonstrate the equally strong adsorptionof the bases. The strength of adsorptionwas found to be piperidine > pyridine >ammonia.

4. The order in hydrogen of the ringopening was found to be low (

sonne mans 73 mechanism 2

Documents