Balandin1958_2.pdf

The Nature of Active Centers and the Kinetics of Catalytic Dehydrogenation A. A. BALANDIN N . D. Zelinskii Institute of Organic Chemistry, U.S.S.R. A c a d e m y of Sciences, and Moscow State University, Moscow, U.S.S.R.

I. QUASI-HOMOGENEOUS SURFACES In my report on the active centers of catalysts certain new conclusions will be considered, drawn from our kinetic studies on dehydrogenation. I n accordance with the multiplet theory of catalysis, proposed by the author of this report (1, 6,S), owing to the small effective radius of chemical forces, reacting atoms must be in contact with the catalyst. Of interest is the catalysis of complex molecules that are unable to superimpose on the surface. Klabunovskii and the author (4) have found that in spite of their complicated structure certain tripticene derivatives can be hydrogenated on Raney Ni already a t 45’ and 1 atmosphere. It follows from this that the active centers of the catalyst are elevations on its surface; it is on them that the reacting portion of the molecule locates itself, the rest of it falling into the depressions. (In Fig. 1 a and b represent the hydrogenation of the benzene ring and of the carbonyl bond in contrast to c, representing the hydrogenation of the double bond that can take place on a plane as well.) Even isopropanol and other secondary alcohols cannot superimpose on the surface during dehydrogenation (Fig. 2). Since the difference in the dehydrogenation of the secondary and primary alcohols is only a quantitative one, the latter should also react on the elevations in the form of Taylor’s peaks ( 5) or Volkenshtein’s “biographical” active centers (6). But the author (7) believes, moreover, these elevations to be carriers of crystal facets. Such islets are metastable and, depending 011 their differences in heights and surface areas and consequently in the deformations caused by the neighboring lattice atoms, are subject t o statistical treatment. If their distribution bears an exponential character, then this can explain the appearance of parameter h in Equation (1) (8-11). All data on the kinetics of the catalytic dehydrogenation of hydrocarbons, amincs, and alcohols obtained in our laborrttory are expressed by the equation ( I d ) : 96

ACTIVE CENTERS AND CATALYTIC DEHYDROGENATION

97

Here dx/dt is the rate of reaction, E the activation energy, T the temperature in OK, p the partial pressure, and r the substance number, being 1 for the reactant, 2 for the product, 3 for hydrogen, and 4 for a n extraneous substance; k d and zr are constants. The fractional factor with p of Equation (1) according to the accepted view corresponds to Langmuir’s adsorption isotherm for mixtures on an almost covered homogeneous surface or on like active centers and the z are the relative adsorption coefficients. This fraction was so interpreted for other cases by Hinshelwood ( I S ) , Schwab (9),the author ( I , 1 4 , Frost (15),Hougen ( I B ) , and others. Two questions now arise: 1. Is the factor in general an adsorption isotherm, i.e., do the z’s have the significance of relative adsorption coefficients? 2. How is one to reconcile the shape of the adsorption isotherm deduced for a uniform surface with the fact that, as we have seen here, the surface is heterogeneous?

FIG.1. Hydrogenation of tripticene derivatives (4)

98

A . A . UALANDIN

FIG.2. Dehydrogenation

: (a) isopropanol,

(b) cycloliexanol

Pshexhetskii et al. (17) have recently expressed the view that in the equation for the rate of dehydrogenation, containing the function p in the form of the Langmuir isotherm, z is the ratio of the rate constants of some partial reactions. Calculations by the author and Kiperman (18) have shown that both interpretations of x are limiting cases and for x to possess the usual meaning of the relative adsorption coefficient, it is necessary that the desorption rate constant k, of the unchanged reactant molecules exceed the rate constant k of the dehydrogenation. That z actually is an adsorption coefficient has been demonstrated by Balandin et al. (19).A mixture of butane and butylene was dehydrogenated, isotope. It one of the compounds being labeled with the radioactive mas found that in the consecutive reaction C,Hio

4

CaHs + C4Hs

(1)

the butylene formed from the butane is first practically completely desorbed so that only the butylene molecules adsorbed m e cv are dehydrogenated to hutadiene, This can be seen from the fact that on dehydrogenating a. I : 1 butane-active butylene mixture, the specific activities of the butylene and of the butadierie formed are found prttcticdly to coincide. In the case of a mixture of active hutane and inactive butylene, the specific activity of the butadiene was found to be somewhat lower than that of the butylene. Runs were made on a chromium catalyst a t 635" and times of contact from 0.77 to 3.22 sec.

ACTIVE CENTERS AND CATALYTIC DEHYDROGENATION

I

99

A I

I I

I I I I

I

I

I ------

B

FIG.3. Potential curve for the consecutive dehydrogenation of butane and butylene-l (19) : (1) CaHlogas, (2) C4Hlnads., (3) C4H,oactive compl., (4)C,Hs ads., ( 5 ) ClHs gas, (6) G H s ads., ( 7 ) C 4 H active ~ compl., (8) C4Hs acts., (9) C4H6 gas.

Figure 3 shows a profile of the potential surface of the consecutive stages of (I).Let k be the rate constant of the reaction. Then, in conformity with the results obtained, k45 >> ka7. It is clear, however, that the level 4 of the adsorbed butylene in the process of dehydrogenation of butane is the same as the level 6 of the butylene in the dehydrogenation of the latter. Hence, k~ = ~ C H , and consequently k65 >> k 6 7 . Thus, in dehydrogenation the desorption rate constant of the initial substance is considerably larger than that of its dehydrogenation. It is thus shown that x has the meaning of a relative adsorption coefficient. But

zr

=

a7/al

(2)

where a designates the absolute adsorption coefficients, these lat,ter being the equilibrium constants of adsorption (20):

A (gnfi) -I-

A!O(free surfnce)

Therefore, for exchange on the surface

*

AS(stcrface,

(1 1)

100

A. A. BALANDIN

Then according to thermodynamics one can obtain the change in the standard molar free energy (at constant pressure), in the enthalpy and in the entropy for adsorptional exchange (21): AF" = -4.57T log z AH" = 4.57T'd log z/dT

AS"

=

(AH" - A F o ) / T

(5) (6)

(7)

and by means of the same equations with a instead of z, for adsorption. It should be emphasized that the quantities a, z, AF", AH", and AS" determined from Equations ( 2 ) to (7) refer to the catalytically active centers and not to the surface as a whole, wherein lies the advantage of the method now being considered. Their determination is of importance in catalysis because all the molecules taking pnrt in a catalytic reaction must pass through the stage of adsorption. The contradiction in Equation (1) concerning the homogeneity and heterogeneity of the surface is eliminated by accepting the theory of yuasihomogeneous surfaces developed in some detail by the author (22). A quasi-homogeneous surface is one, the two different active centers of which are characterized by a constant ratio ar/a,o =

@

(8)

independent of the substance number r . If a,, is a , for the less active, predominant centers, then w is the degree of unsaturation. It follows from the theory that the AF" of adsorption is here the arithmetic sum of AF a t the basic level (this quantity depending upon the chemic:d nature of the catalyst) and of AF" ( w ) , depending only upon the genesis of the catalyst. Applying Langmuir's adsorption law to each kind of active centers, the theory leads to a hitherto unknown logari thmic adsorption isotherm for mixtures. It shows that the expression for adsorption a t saturation does not depend upon the shape of the adsorption isotherm, provided only that the exchange law remains valid. At the same time the x r are constant over the entire quasi-homogeneous .surface : z, = a,/al = a7w/alow = aro/alo= const.

(9)

ACTIVE CENTERS AND CATALYTIC DEHYDROGENATION

101

Therefore, in the general expression for the rate of reaction on heterogcnous surfaces dz

-=

dt

1

hTlds (10)

where

is the distributioll function, wc havc ds

=

@(a,,a2

, . . .,

UP)

daldaz

, . , dar = @(a,)dal * *

=

a(€)de ( 1 1 )

The last part of Equation (11) is true because by our method the adsorption measured is that taking place on the catalytically active centers. S e a r saturation and when @(t)

=

hr

cye

(121

Equation (10) become::

This integral is easily taken and we obtain

Thus, we have derived the fundamental equation (I), since the term in the first brackets of Equ:it,ion ( I 1) may tw considrred as the constant 1.id and the second exponential function, owing to its rapid decay, may be neglected if h > 1IRT, whereas the first one may be neglected if h < 1/RT. Equation (14) holds also for other cases of monomolecular catalytic reactions.

11. FLOWMETHODKINETICS Our ospc>rirrieiit:il results on drhydrogc'ii:Ltiot~were ohtnincd by a flow method undcr statioiinry conditions. We shall briefly describe the method

44FIG.4. Cross section of catalytic tube

102

-4. A. BALANDIN

of calculation which we have had newly to derive (23, 24) The change in the number of moles of reactants when the gas has passed out of an elementary section dl of the tube with the catalyst (Fig. 4), owing to the reaction, is dm = krludl, where I’l is the adsorption of the reactant, k the rate constant of the reaction, both quantities referring to one ml of the catalyst filled tube, and u the cross section of the tube. At nearly complete surface coverage I’l = spl/ p , (where s stands for the surface) and, therefore,

c

But the ratio of the partial pressure p r of the rth substance in the given section to the total pressure P in the tube equals the ratio of the number of moles N r of the rth substance passing through this section per second Nr passing per second through the same to the total number of moles section :

c

pr/P

=

(16)

N r / C Nr

Let the substance T be passed into the tube a t the cons1,ant velocity A , nioles/sec. Then, owing to reaction, the given section will pass

Nr

=

A , - vrrn

(17)

moles of the substance T per second, where v, is the stoichiometric coefficient. lcor the reactant v1 = - 1; for the product of the dehydrogenation v2 = 1; for hydrogen v 3 varies for different cases and for the extraneous substance v4 = 0. From Equations (15) to (17), we obtain dm/dl = uks(Al - m ) / [ z z p ( A r

+ v,m)]

(18)

and integrating between the limits 0 and I, we obtain for the general case [Ai(zz V3Z.3) Azzz A323 A4241 111 IAi/(Ai - m)l

+

+

+

+

- (z2 + v3z3 - l ) m = KV

(19)

where thc constant K = ks arid the volume of the catalyst V = ~ l For . derivation of an exprcssion corresponding to Equation (19) when r is the Langmuir isotherm, see reference (24). If the reactant alone is passed through the tube, A P == A s = A 4 = 0 and Equation (19) assumes the form (22

+ v & 4 1In [Ar/(A1 - mo)] - (z2 + vjza - l)mo

=

Z(V

(20)

Introducing the designations mo/A1 = y and A1 = u0 , Frost applied Equation (20) to different flow rates, giving it the form uo In [1/(1 - 1/)1 = cx

+ PUOY

(21)

If uo In [l/(l - y)] is plotted against u0y, a straight line is produced, with

ACTIVE CENTERS AND CATALYTIC DEHYDROGENATION

103

+

+

the constants a = K V / ( z , Y&) arid /3 = (z2 v3zS - 1 ) . Such straight lines have actually been obtained mainly for the reactions of dehydration and cracking by Antipiria and Frost (25),Orochko (as),l’anchenkov (27) (who further developed the theory of kinetics undcr flow conditions) , Topchieva (28), Nagiev (29), and others, which also confirms Equation (20)* Since, when z 6 0.1, -In (1 -

(22)

2) M 2,

Equation (21) can not be applied for small yields, because a straight line will then always be formed with (Y = 0 and /3 = 1. On the other hand, working with small yields mo/A1 permits one, a t constant rates, to obtain the rate constant and the activation energy. Indeed, from Equations (20) and (22) it follows that in that case mo = KV

(23)

For yields > 10 % the results will be approximate. Small volumes of gas are not difficult to measure with sufficient precision. Special cases are met with when z2 = 1 and 2 3 = 0. Then Equation (20) transforms to A1 In [Al/(A1 - m,)] whence mo = AIKV/(A1

KV

(24)

+ KV/2)

(25)

=

since (up to 30% conversion) the relation In (a/b) M 2(a - b ) / ( a

+ b)

(26)

is applicable. One can pass binary mixtures containing p % reactant and measure m and m, a t constant flow rates. Then A1 A, = N , lOOAl/N = p , and from this and from Equations (20), ( 2 2 ) , and (23), there follows a formula for evaluating zr (30) :

+

(cf. Fig. 5). For large yields the next approximation or the complete Equation (20) should be used. In order to ascertain whether zr belongs to the substance r introduced to the mixture and not to a poison which may have been formed, constant zr values should be obtained on increasing the flow rates or on diminishing 1. Now we can introduce the values of z, obtained in the manner described for smalI values of m / A l , into Equation (19). Then already for any A1

104

A . A . BALANDIN

m0

1 m

I

Y

P

-

1

I

I 3

FIG.5. Schematic curve of ndsorlhve exchange at the active centers

and m we obtain K , which should remain constant (for t: given temperature). Thus, the use of the flow method allows one to obtain the usual kinetic expression for constant volume conditions :

dx -- K dt

Pl

pi

+ z2pz + zspa + 24p4

(28)

From the z and their temperature dependence, the AF", AH", and AS" of adsorption exchange can be obtained using Equations (5) to (17) and from the value of K for different temperatures, the quantities K O ,E , and h.

111.

EXPERIMENTAL lxEsULTf3 ON DEHYDROGENATION

The validity of the equations given and of the methods of computation is illustrated by the following examples. The independence upon the total pressure P within wide limits as is required by Equations (18) and (19) is evident from Fig. 6. The variations in m for binary mixtures of different composition can be seen on Fig. 7, where the experimental points fit the theoretical curves well. It can be observed in Fig. 8 that with decreasing 1, the quantity z gradusilly assumes a quite constant (true) value. Figure 9 shows the temperature dependence of z ; straight lines are obtained for log z as a function of l/T, demonstrating that AH" is practicalty independent of the temperature. Figure 10 gives the linear Arrhenius relationship for log K vs. l / T and Fig. 11 shows that this holds only for log K , since the

ACTIVE CENTERS AND CATALYTIC DEHYDROGENATION

2o

a)

t

200

0

200 0 0

mo m1/5'

600 Rmm

400

e

Y

"

200

0

b)

-

600

400 _c)

0

O

105

800

0

800

P,mm

FIG.6. The pressure P and rate of dehydrogenation mo : (a) methylcyclohexane on l't (31), (b) cyclohexanol on Cu ( 2 3 ) .

function log m vs. 1/T although linear and parallel to the first line in its lower portion is bent down somewhat in the upper, in conformity with theoretical expectations. Figure 12 shows the logarithmic relationship log K Ovs. E ; for groups of related reactions straight lines are obtained, the slopes of which equal h/2.3. Table 1 summarizes the experimental results obtained in our laboratory on the kinetics of the normal dehydrogenation of hydrocarbons (hexahydroaromatics to aromatics, the open chain compounds butylene to butadiene, and ethylbenzene to styrene), of amines to ketimines, and of alcohols to aldehydes or to ketones, respectively, in the presence of metallic or oxide catalysts. Equation (1) was found to apply in all cases. KOand h are given by log KO = log K f/4.57T (29

+

h

=

which follows from Eq. (1).

2.3 (log K O - B ) / E

(30)

I00

Aldohydr

50

100 Alcohd

m

20

1

10

0

25

50

15

100

Flu. 7 . 1~~xcti:iiige curv(:3: (a) met1iylc~clohes:~iie with t,olucnc! on Xi (52), (1)) et.liyI rilcohol with :icct,uldchyde o n (hi (SS), ( c ) ethyl ulcoliol with hydrogen 011 (!u ( S 4 ) , (11) I)utylt:iie with wat.er ( 9 6 ) . (e) isopropyl irlcohol with cicetonr! ant1 ( f ) with hydrogcr1 011 Z I I O ( 4 4 ) .

107

ACTIVE CENTERS AND CATALYTIC DEHYDROGENATION

Aldehyde

Alcohol

FIG. 8. Stabilization of the value z for ethyl alcohol and acetaldehyde on Cu, by shortening the length 1 of the catalyst, bed. Full curve, 0, low-space velocity; dashed curve, X, two high space velocities. 1.0

-

T

15

Ib

a)

16

17

I

--*

18

1

18

15

16

I

17

I

18

I/T. 10

-I b)

FIG.9. The independence of the heats of exchange on the temperature (a) alcohols on oxide catalyst (36).Reading from left to right: allyl, n-propyl, ethyl, i-propyl, i-amyl, n-butyl, phenyl ethyl alcohols. (b) Curve 1-isopropyl alcohol with acetone, curve 2-isopropyl

alcohol with hydrogen on MnO (66).

108

A. A. BALANDIN

n Pr

1.6 -

1.4 -

1.2 -

f

1.0 -

lg k o'8 -

0.6 -

0.4 -

0.2 01 13

'

'

I

I

I

I

( I

17 18 19 20 +I/T. lo4 FIG. 10. Arrheiiius lines for dehydrogcn:ttion of :ilcohols on an oxide catnlyst. (56) ; numerstiori corrcsponds t o Table 1.

0

16

15

I4

.

-

\ 21 2 1 22 23 I/T.104 FIG.11. <:omparison of the log k (curve 1) and log mu (curve 2) vs. l / T relationships for the dehydrogonation of drohol on ('u (33).

"17

l'8

I9

20

I09

ACTIVE CENTERS AND CATALYTIC DEHYDROGENATION

10

12 N&

Cal

14

16

I

I4

25

16

+E Cal

20

18

fb)

(a)

I

I

I

12

30

6E Cal

L

35

10

15

20

4 & Col

25

(c) (4 Vio. 12. Straight lincs confirming thc 0-rulc; numer:ition corresponds t o Table 1 : (:I) smines on Ni and Pd (41, is), (1)) alcohols on :in oxide catalyst ( S6 ) , (c) cyclic hydrocarbons on CrzOt (39, 4O), (d) iaopropyl alcohol on ZnO prepared by different methods (44).

I n view of the extensiveness of the material the values for m o, z, AFO, AS", and K are given for only one temperature, their temperature dependencies being apparent from the magnitudes of AH" and E. For comparison the values for E', i.e., E obtained using Equation (23) are given in italics. The data in Table I refer to the kinetic and not to the diffusion regions, as evidenced by random tests on catalysts of varying granule size.

CH~C~HI Ht

lcyclohexane lcyclohexane

imethylcyelone in

lcyclohexane

hexane

H%

aa Nos. 1-3

CH.CsHr

(CHddMh

I

I

1 CsHs

imethylcyclont.

lcyclohexane

hexane

245

245

236

1

I

1

0

I

1

0

1

I

0

1

I

I

1

0

-

1

0

0

1 -

1 "-> I

Same as Nos. 1-3

0

1

0

Ni-MrOa ; V = 60 ml (Ref. 46)

0

Ni-AlrOt ; V = 100 ml (Ref. 32)

1

I

Ni-AlrOa ; V = 72 ml (Ref. 45)

Pt-asbestas; V = 2 ml (Refs. 3:. 48)

1

1

1

25.6

31.2

34.9

28.8

1

1

1

I

0.84 0.85 0.96 0.92 0.35

415 472 446 442 425

-0.17 -2.08

0 0

115 1140

20.7

20.3

22.3 -0.08 0

55

21.2

0 240

-0.35

I

CrL)z-asbestos; V = 16 ml (Refs. 39, 4O).B = -2.68

I ii: 1

I

10.0

2.0

2.2 5.6 2.7

0.76

7.0

12.6

5.6

1

1

12.0

0.610

0.075

2.13

0.108

0.385

0.131

0.153

0.0852

0.081

I

I

2.49 6.83 2.96

0.77

7.87

15.3

6.08

22,9

5.92 10.5

25.8

69,7

32,3

22,2

2S, 7

7.08

25,9

18,2

-

15,6 14,7

15,6 12,3

28,4

1

8.03

7.44

l -

5.72

5.78

5.64

orption Coeficients ( z , ) , Free Energies ( A F " ) , Enthalpies (-AH"), and Entropies (AS") of Adsorptive Exchange Rate ation Energies (e), and the h Parameters i n Catalytic Dehydrogenation. AF", AH', and c cal./mole; A S " e.u.; A1 R pply, and mo Reaction Rate ml. Substance Vaporlmin., K-ml./(ml. m i n . ) .All Valuesf o r N . T . P . Original Data Are e Units.

TABLE I

I

0 0

302 298

[(CHa)zCHlzC=NH (CzHs)zN (CHt)r-C=NH

CHa

I

(CHa)rCHCHzC=NH

CHs

I

300

300 CHa(CHz)rC=NH inoheptane

CHI

I

CH:

20.7 20.7 -4.73 -4.57

I

=

1

3

7

0

0

0

0

O i

O

-

37.7

29.8

6.40

7.00

-

17.43

12.44 23.8

-

7.21 6.45

30.7

-30

9.20

8.20 37.7

1 9.64

8

11.1

8.30

1

9.4

29.8

1

22 ml (Ref. 49)

Pd-mbestw; V = 3 ml (Refs. 41-48)

0

300

I

CHI

21.2 22.3 20.3 -5.17 -5.37 -5.65

3 ml (Refs. 41-49) B = -0.96

(CHr)zCHCHC=NH

=

I

0

270-33

thyl-4-aminotane

0

Ni-Alas ; V

I

Copper-chromium catalyst; V 650

20.7

-2.20

2 ml (Refs. 80, 80)

0

0

0

300

I

I

=

3260 3160

0.092 0.100

Chromium catalyst; V

0

3580 3720 3900

0.073 0.067 0.058

0

0

1520

0.33

CHa(CHt)rC=NH

Cot

€la

425

HP

I

405

Hz Ht

CHFCH-CH%Hn Hz

466

Hz Hz HI 461 517

448

CHaCioHi

as Nw. 23-26

imethyl-3nopentane hylamino-4nopentane

hyl-4-aminotane

inoheptane

enzene

ne-1 ne-1 ne-1

hyl-5,6,7.8hydronaphe bexane leyclohexane methylcyclone in thyl-S, 6,7,8hydronaphe

I

I

10,7 4.87 5.81

2.33

2.06

4.16

4.03

9,9

9,7

8,6 3.74 2.74

-

9,5 4.14 3.07

-

9,3

I

-

33’w

90,40

4.09

I

-I

!

3.21

0.250

17,6

17,i

5.09

6.0.1

~

i

1

16.3

-14,00

12.8

5.59

16,3

ii.1n

5.06

-14,00

9,1 3.89

8.@ ; O . X

2.9

,

I 7.7

9,140

5.06

!

11,400

I

!

1

I

i

I

!

23501 -22.000~-31.9

i

1

I

12.2

I

I

5.09

I

12,200

~

,

1

I

!

!

I342

i

1 1

j 257

4.4@!0.63 4.20 10.60

ol-2 ol-2

' i

I

1 1

1

(R.ef, .% I3 i) 2, 37. -.

3.8 3.7

1 CH:CH=O

]

ol

-

11.7j 14.7

Hs

-.

= I 1 ml

I

CHaCOCH C Hs

-.

28,3301 44.2 -36.030 ,-58.5

CsHto=O

Oxide catalyst: T'

1

hexanol

360 2480

(CHr)eC=O

1'

Oxide catalyst; V = 11 ml (Ref. 35) B = 2, 37.

Ni-quartz (0 58% Ni); V = 7 ml (Ref. 3 6 )

ml (Refs. 53. 54)

Co; V

n /

I n

= 44

O

n n n 970

970 0.40

1

0.40

257

1250

n n

-1.83

IO.76 0.15

panol

385

! - I

3.89

O /

Ni-quartz (0.58% Nil: V = 7 ml (Rei. 36)

389

15.87

76.8:

~

~

12.3

4.9

84.7 -1.87

5.1

3.1

84.7

-1.83

6.0

2.8

88.0

0

-

84.3 85

73.4 83.7 0

n

n

n

-(o

=

0

(Refs. 53. 54)

H20

1

I

O /

panol (Ref.

Hs

= 44 ml

Hz0

CHaCOCHeCH3

1

panol (Ref.

1

5.1

(V = 5)

4.4

10.8

-

30

I

2 ml (Refs. 5 , S S 37,38,59)

i

i

250

Butanol-2 Butanol-2

0.97

2.2 2.5 4.4

11.5

-

5.29

15.87 11.63

23.8

n

[

I 1

HO CHaCHzCH=O CHaCHzCH=O

I

I I

12.50

ol (Ref. 50) ol (Ref. 61) ldehyde (Ref.

1

Co; V

I (CHr)eC=O

I

970 :

CHsCH=O H2

51

I

0.40

ol (Ref. 83) ol (Ref. S 4 )

50

I

CyclohPxanol

Ij

1;:

1

I 2.26 / j 5.29

6.18

30.7

I

I

ns Nos. 31-34

40

257

970

! 0.40

252

I

,

O 1

CHaCHzCK=O

HzO

1

panol (Ref.

51 ) Isopropanol ( F k f .

j

250

47

25i

!

1 I

1x20

Isopropnnol

252

01

-(o

970 , 0 01

V.4b

0.40

n-Propnnol (Ref.

i

1

0

i

I

0

! I

i

0'

257 257

46

48

30.7

CHsCHzCH-3

,

' 1

257

CIhC1IzCH=0 CII3CIIaCII'=

Cu; V

56)

261 257

ni

n-Propanol (Ref.

HAI

i

n

45

51)

CHsCH=O H z

1

I

1

Ethnnol (Kef, 50) Ethanol (Ref. 61) ! Acetaldehyde (Ref.

i

303

0

01

~

42 43 44

2.26 6.10

Ethanol (Ref. 33) Ethanol (Ref. 3 4 )

n

'

!

CH,

1

41

(CZH~)~N(CHI)IC=NH

Pd-asbestos; V = 3 ml (Refs. 41-43)--Continued

40

I(CHa)eCIIhC=N~

imethyl-3nopentane ethylnmino-4no-pentane

34

2,4-Dimethyl-3aminopentane 1 1-Dimethylamino-4i amino-pentane ,

TABLE I-Continued

38

6.78

I

i

TABLE I-Conlinuecl

3.07

I

9,9001

12.1

341

3.6

-3530

342

6.1

-4940

1.33 1.51 1.0

-16,400

-18.2

-11,300

-12.7

-

66.0

65.0 51.4 43.6

30.4 (30.4 130.4 30.4

i 0.60 1 624 1 -8,100 1-14.2 The values are close to No. 75.

-8,500 -13.4

2.6 28.8

44). B

I'

30.4

(30.1 t30.1

-

346 341 342

345

344.5 344.5

7 4 1-3, 5-7

36 4 5 2

I

1-3, 5-7

:" 1

I

1.19'

0.94 0.26

0

14.0 44.0

= -0, 59.

-8,500 -13.4

-215

=

ZnO: V

1650

191400 16:;

MnO: 1'

1 ml (Rei.

-215

1 I

0.60 1 624 1 -8.300 1-14.2 The values are close to No. 75.

319

342

341

4

=

1,W 19,400

The values are close t o No. 68

1.2

6.1

3.5

0.95

1 6

16i00

1 I

=

-27,500

-

l.lgf

-

-4940

5

-515

-5,100

1

1 I

ZnO; V

52. i

The values are close to No. 68

--

2.6 28.8

15 ml (Ref. 55)

-

-43.9

-20.3

-12.7 -11,300 -3530

-16,400 140

4

1 ml (Ref. LA). B = -0. 59.

-515

I

44.0 44.0

52.7

43.6

51.4

66.0

65.0 -18.2

-12.5

-19.3 -11,000 -4320

3e

7

336

7683

1.2

0.94 0.26

2 4.55

69 70 71 72 73 74 75

Impropmol lsopropanoi lsopropanol Isopropanol Impropanol Isopropanol Isopropanol Isuyrupanul laopropmol

352

68

Isopropanol Isopropanol

-11,200

67

349

0

-3570

66

1

Same us Nos. 52-58

3.6

59-65

8- Phenylet.hano1

342

58

60.3

1

10.2 8.3 7.9 9.5

140

60.3

-19.3

14.6 10.9 10.3 13.1 12.3 11.4 3.1

0.95

0.79

-

-11,000

-4320

336

-12.5

-

0.988

0.51

4.55

-11.200

-3570

8.1

9.4 9.1

14.6 10.8

3.6

6.0 6.0

0.85

1.33

4.36

1.22 13.3

3- Methyl-1-butanol

10.4

57

48.1

n-rlutmol

32.8

56

4.6

Iaupropnol

4.7

55

342 352

19.3

__

2-Propnol-1

19.2

54

0.41

1.76

n-Propanol ~

53

-18 m and K V are given for the temperatures of column 3, which are near the temperatures of determination of zr . K, AF", and AS" are given for one temperature (419". interpol). rogenation ( 4 7 ) . 3-39, z from the validity of Equation (23). 52-65. K = Kars' . gives the Nos. of the ZnO catalystes differing in the methods of preparation. stant8 have been found from the next approximation. 68-83 column 11 gives K / s , where s is the B.E.T. surface. varying from 0.3 to 15 m2. 68-83 column 12 contains log(Ko/s). The values of log& and h were obtained from and K according to Equations (29) and (30). Column 14 umber of points sufficient for the application of Equation (30).

-

-

-

6.35 7.03 10.20

4.34

11,4 12.28 12,5 15,0 17,6 19,4 26.80 -

4.06

8.13

6.19

5.61

16,3

23,6

19.10

5.75

12,8

5.09

JY.J

17,50

5.70

5.68

__

__

I _

19.2 4.7

19.3 4.6

__

~

1.76

15,600

5.68

I

4.87

13,700

17,500

5.70

0.41

(3.06)

17.100

32.8

48.1

4.36

5.00

12, 800

4.88

10.4

13.3

1.22

6.75

1.77

10.8

14.6

1.33

5.61

16,300 15,000 12,900

lb. 1w

9.1

9.4

0.85

6.19

18,000 17,600

6.0 6.0

8.1

0.54

8.13

23,650

0.981

~1.06

0.79

4.31

9.5 8.6 8.5 3.0

14.6 10.9 10.3 13.1 12.3 11.4 3.1

1.33 1.51 1.0

6.35 7.03 10.20

-

-

10.2 8.3

7.9

-

(5.77)

13,500

-

-

I

-I

*I'

11,400 i2.m 12.500 15,000 17,600 19,400 26.800 -

1.87

0.36

-

S.OB

-

I

9 .Oi 9.04 9.21

i

CHXH-CH=O

(CHa)zC=O

CHGHzCHtCH=O

(CHdzCHzCHzCH=O

CsHsCHtCH=O

HP

(CHa)zC=O Hz

(CHa)KO (CHa)tCO (CHa)zC0 (CHdzCO (CHa)zCO (C Ha)zC0 (CHa)zCO Hz Hz

enol-1

panol

anol

hyl-1-butanol

nylethanol

as Nos. 52-58

panol panol

panol panol panol panol panol panol panol panol panol

'

CHzCHzCH=O panol

1 6

In Nos. 9-18 m and K V are given for the temperatures of column 3, which are near the temperatures of determination of zr . R, d F ",and AS" are given for thesake of comparison far one temperature (319". ioterpol). From h3-drogenation (47). C I n Nos. 23-39, z from thc validity of Equation (23). For Nos. 8 - 6 5 . K 2 Karn' . Column 2 givaa t.hc Nus. of the Z n 0 cxtalystes differing iu thc mothoda of prepRru.tiun. 1 The z constants have been found from the next ayyroximatiun. 0 For Noa. 6&83 column 11 gives K i 8 , where 8 is the B.E.T. surface. varying from 0.3 to 15 m2. For 3 0 s . 68-83 column 12 contaias log(Ko/s). The values of lagKO and h were obtained from and K according to Equations ( B Y ) and (30). Column 14 gives tr for cases with a number of points sufficient fur the application of Equation (30).

114

A . A. BALANUIN

IV. DISCUSSION OF RESULTS AND

THE

MULTIPI,ETTHEORY

Table I shows the following: 1. The Nature of Catalysts. The kinetics of dehydrogenation depend upon the nature of the catalyst. The catalysts used divide sharply into two groups-metals and oxides. For the latter the reaction temperature and e are higher, whereas h is lower than for the former. This is nmtural, for catalysis proceeds under the influence of chemical forces, i.e., valence electrons. According to the model described further below (Fig. 18) metal atoms participate in the dehydrogenation process; in the oxides the electrons belonging to the metal are displaced toward the oxygen atoms located in the lower layer (or, less probably, laterally adjacent). It is for this reason that the method of preparation (Nos. 68 to 83) of the catalysts affects the kinetic characteristics; upon it depends the profile of the catalyst surface, i.e., the number and the location of the atoms adjacent to the active centers and influencing them. 2. The Nature of Reactants. No less effect on z and e is exerted by the nature of the reacting atoms of the molecule. The values of z and E are distinctly different for hydrocarbons, amines, and alcohols (on the same or similar catalysts). This complies with the multiplet theory of catalysis, according to which in the three reactions the following atomic groups, in which the vertical bonds pass over to the horizontal, are in contact with the catalyst: C-C

I 1

H

H

C-N

I 1

H

H

C-0

I 1

H

H

(IV)

I n harmony with t.he multiplet theory is the observed dependency on the nature of t$e catalyst, since the atoms of the latter are involved in the multiplet complex. 3. The Influence of Substituents. The substituents a t the groups indicated have a markedly smaller effect on z and 6 . This is evidence for the orientation of the molecules with their reacting atoms toward the catalyst (Fig. 13), as is required by the theory. The influence is particularly slight in the case of metals. For hydrocarbons (Nos. 1-3), amines (23-29), and alcohols z1 = 1 and 23 = 0. From a crisscross examination of the data of Nos. 40-47, it can be seen that in the case of copper not only the balues of z2 but also those of uz are practically the same for ethyl, n-propyl, isopropyl alcohols, acet- and propionic aldehydes, and also acetone. This was explained by the author (66) in a treatment of the above-mentioned model from the standpoint of statistical mechanics. It was found that

ACTIVE CENTERS AND CATALYTIC DEHYDROGENATION

115

H,\C CH, CH3-CH

CH

\/

CH-NH

w

/

Pd 9700

E = 9920

CH-NH

Pd

I I400

CH-CH,

C

-

crz 0,

28500

cr2 0, 26600

Cr, 03 24800

FIG.13. Diagram of molecular orientation in catalytic dehydrogenation

where the numerator contains the partition function of the adsorbed molecules and the denominator of the molecules in the gas. Owing t o the canceling of both the masses m and the moments of inertia I , to the close values of AH, and to the constant values of the other quantities, a: remains constant for all the above molecules. The activation energy E for the cyclic hydrocarbons on Ni-A1203 are practically independent of the number of methyl side groups (Nos. 1-3). For amines on Ni and Pd (Nos. 23-39) the E are also quite constant. Studying the dehydrogenation kinetics of ethyl, n-propyl, n-butyl, and n-amyl alcohols on Cu prepared by a different procedure, Palmer and Constable (8) did not measure z, presumably because they were working under conditions when Equation (23) was applicable. They found for these alcohols a constant e = 22,000 cal./mole, which they interpreted as the result of the same orientation of the OH group. In our experiments the values of E for C2H,OH and C3H70Hare also quite close, although there is a noticeable difference of 600 cal./mole, which is beyond the limits of experimental error. The close values between the reaction rates for the alcohols were explained by the author (56) from the standpoint of the theory of absolute reaction rates and the multiplet theory, which may serve as illustration of their combined application; from the former one can compute velocities using models of the latter. In the case of oxides 23 # 0. For MnO and isopropyl alcohol 2 3 increases from 0.1 a t 316" to 0.36 a t 366'. The substituents exert a stronger influence on zz. Thus, on CrZ03, zz has a tendency to increase with the methyl

116

A. A. BALANDIN

groups introduced into cyclohexane (Nos. 9-11). The value of x2 falls to % of its initial value when the cyclohexane ring is condensed with benzene to tetralin. Formation of the conjugate systems of but,adiene (No. 19) raises the value of zz . An extensive study has recently been made with respect to the effect of structure on the dehydrogenation of alcohols in the presence of an oxide catalyst. It is seen from Table I that the structure influences greatly the 22 value. The formation of conjugate bonds in acrolein again raises 22 (No. 54). With increase in open chain length z2 rises ( C z H 6 0 H , z2 = 2.3; C 3 H 7 0 H , z2 = 3.4; C4H90H,x2 = 3.5; n-C6Hl10H, zz = 5.8 (Nos. 52, 53, 56, 57); for the secondary, isopropyl alcohol z2 is lower (22 = 0.2) (No. 55). The activation energy E also changes here regularly. E falls by 4900 cal. on substituting the a-hydrogen in ethyl alcohol by a CH3 group and by 2100 cal. for substitution in the P-position. On the contrary, substitution by phenyl and methylene radicals somewhat raises the E value. Such regularities in dehydrogenation remind one of the rules of Dohse (57) and Bork and Tolstopyatova (58, 59) in the dehydration of alcohols (on bauxite and on A1203). The singular position occupied by the dehydrogenation of alcohols on oxides can perhaps be explained by intermediate formation of hydrogen bonds. 4. Enthalpy and Entropy of the Adsorption Exchange. Experimental evidence shows that, in dehydrogenation the calculated A H o is practically independent of the temperature over the range of 50 or 100" (see, e.g., Fig. 9). The entropy change ASo does not vary markedly with the temperature in the cases which were investigated. The AHn and ASo of adsorption exchange approximate those for chemical reactions. The large values for A H o speak of the possibility that the adsorption required for catalysis originates from the formation of a one-electron bond. With accumulation of experimental material it would be of interest to treat the entropy of adsorptional displacement from the standpoint of statislkal mechanics as an approach t o the construction of adsorption models in catalysis. As in the case of some chemical reactions (GO), a parallelism is observed between the AH0 and ASn of adsorptional displacement (Fig. 14). 5. Activation Energy. If one works with small yields, there should be no difference between E and E' [Equation (23)]. In practice cine has sometimes to deal with larger yields. As is shown by Table I the clifference between E and E' will then not exceed several kilogram calories and, if this is taken into account, E' may also be used in suitable cases. The values of h remain sufficiently constant for groups of related reactions, which confirms the 8-rule. That the 8-rule holds true is especially clear from Fig. 12b, where all experiments were carried out on a single sample of catalyst of constant activity. The total surface in this case was,

ACTIVE C E N T E R S AND CATALYTIC DEHYDROGENATION

0-

I

I

I

I

-

t

-AHIDk

I

I

I

-

117

0

- -10

/

f

--m-AS

- -30 -40

-

-50

therefore, strictly constant. The points fit the straight line well, except for ally1 alcohol, which is not surprising if one considers the specific structure of the acrolein formed, and for isoamyl alcohol. Figure 12c shows that a straight line is obtained if log (Ko/s)be taken, where s is the specific surface area of Z n O determined by the B.E.T. adsorption method. It is characteristic that for dehydrogenation often h > 1/RT. According to Equation (14), this means that dehydrogenation is carried out predominantly not by a. few highly active centers, but, on the contrary, by centers of lowest activity, which, however, are very large in number. 6. Application of the Structure Correspondence Principle. The kinetic parameters, z, AFO, AHo, K , and E confirm the difference in the dehydrogenation mechanism of six-membered rings over Ni on A1203 and over Cr2O3 (Table I, Nos. 1-6 and 9-18). We shall enumerate now the experimental evidence in support of the concept that the elevated active centers end in crystalline sites. This includes the sextet mechanism for dehydrogenation of the cyclohexane ring on the (111) faces of Ni, Pt, etc. (Fig. 15) ( 1 ) . The sextet mechanism has been confirmed by Long et al. (61);Emmett and Skau ( 6 d ) , Beeck (63)) Trapnell (64), Rienacker and Unger (65), and others. In contrast to the sextet mechanism is the duplet niechanism for the dehydrogenation of cyclohexane on Cr2O3and other oxides, in which the ring is turned with its edge t o the surface (66) (Fig. 16). I n respect of this reference is made t o

118

.4. A. RALANDIN

FIG.15 FIG.16 FIG.15. The sextet model for the dehydrogenation of cyclohexune ( 1 ) FIG.16. The duplet mechanism for t,hc dehydrogenation of cyclohexane (66)

the recent review by Trapnell (64) concerning the structural aspect of the multiplet theory. However, certain reservations should he made. Firstly, it is preferable not to modify the sextet model according to Trapnell. In Trapnell’s variant the hydrogen atoms do not permit the cyclohexane ring to superimpose on the lattice. Secondly, according to the inultiplet theory catalysis is caused not by van der Waals’ but by chemical valence forces, which is the basis of the energetics aspect of the multiplet theory, not included in Trupnell’s review. The sextet mechanism is also confirmed by the following new facts. Rubinshtein and associates (67) found that as Y t is dispersed more and more on carbon, the rate of dehydrogenation of cyclohexane parallels the intensity of X-ray reflections from the (111) faces and not from the others. Sachtler et al. (68) established by careful electron difrraction and electron microscopic studies that in hydrogenation (and consequently in dehydrogenation) on Ni the (111) [and possibly the (lO0)l faces are active, but not the (110) ones as was found previously by Beeclr (69). The author and Isagulyants (70) investigated the dehydrogenation of cyclohexane (I) and decalin (11) A

A

over Ni on A1203 and over Crz03. In the presence of 1 5 about twice as many molecules of I dehydrogenate as 11, because if the molecules lie flat on the surface, then more molecules of I can occupy the same surface area than of I1 (Fig. 17). On the contrary, in the presence of the Cr203both I and I1 are dehydrogenated a t about the same rate, because the molecules are located on the catalyst edgewise with their sides A B , the hydrogenation of which is the slowest stage of the reaction, determining the over-all rate. The dehydrogenation activation energies of I and I1 h.sve been found to

ACTIVE CENTERS -4ND CATALYTIC DEHYDROGENATION

119

FIG. 17. Diagram of the planar orientation of cyclohexane (I) and decalin (11) on the (111) faccs of P t (70).

FIG.18. Model for the dehydrogenation of n-butyl alcohol. The model is of general significance for the hydrogenation of double bonds and the dehydrogenation of hydrocarbons, amines, and alcohols.

be the same for the same catalyst, equaling 12,500 cal./mole on Ni arid 26,000 cal./mole on Cr203. This is evidence for the same orientation of the molecules of I and I1 on each of the catalysts. The well-known studies of Rideal ( 7 l ) ,Twigg (721, and Herington (73), in which the rate of hydrogenation on Ni is treated as a function of the cross section of the molecules being hydrogenated, a t complete surface coverage by the latter, also speak in favor of catalysis proceeding on planes. Recently the author ( 3 , 7 4 )has proposed a duplet model for the catalytic dehydrogenation of open chains (Fig. 18). Here four atoms of the catalyst are active, in the valleys between which are situated four reacting atoms. The intermediate complex is similar to a surface alloy. The reacting atoms do not get into the valleys simultaneously but in stages, for example, for the dehydrogenation of alcohol : (1)

(2) (3)

+ VzWp = RHCHOV + HW + VW RHCHOV + VW = RRCVOV + HW RIICVOV + ZHW = RHCO + H P + VzWo.

RNCflOII

120

A . A. BALANDIN

The slowest stage is the disruption of the C-H bond in the alcohol. This stage possesses the highest potential barrier. Stronger adsorption in the deep pits M (Fig. 18) leads to poisoning. One of the atoms of the active center (the shaded portiori on Vig. 18) differs in its nature or in the surroundings. Therefore, distinction should be made between two types of valleys V and W . The V sites adsorb less strongly than the W and hence in dehydrogenation, when the temperature is high, they are practically free. The model on Fig. 18 covers a very extensive experimenical material considered elsewhere (75). It is mentioned here because it d s o requires the presence of a planar group of four atoms for the active center. This model forms the basis for the construction of Fig. l b . Thus, it appears that the catalytically active centers of dehydrogenation are peaks on the one hand and crystal faces on the other. This can be explained only by assuming them to be flat islets with steep banks. Such is the result obtained hy the method considered here, which may be called the method of kinetic molecular probing. With the aid of this method the author and Rubinshtein (76) have found that the activity of a mixed Ni-AlzOS catalyst is located a t its interfaces. Indeed, on specimens of this catalyst prepared by different methods, the activation energies of the dehydrogenation and dehydration of isoamyl alcohol were falund to be related by a direct proportionality. But dehydrogenation takes place on Ni, whereas dehydration occurs on A1203. Consequently, the catalysic; must proceed a t places where the Ni and AlzOa are geometrically connected, i.e., a t the boundaries between the phases. Therefore, one may believe that the active center islets on mixed catalysts are adjacent to the boundaries of the other solid phase. 7. Application of the Enerqy Correspondence Principle. The kinetics of dehydrogenation and dehydration may also be applied to obtain bond energy values between the reactants and the active centers and to determine the effect upon them of neighboring atoms. I n the catalytic active complex (Fig. 19) six layers (3, 77) are to be distinguished: (I) the bulk catalyst atoms adjacent to the active centers and differing in nature, number, and position; (11) the active center; (111) the reacting group of atoms in the molecule; (IV) the substituents at thefre atoms; (V) the adsorption layer, considered in the present report; (VI) the layer of molecular adsorption and lateral diffusion. Consider a duplet reaction of the type: A

1

B

D I-+IiI-+ C

A--I>

B-C

(V)

for example, one of those schematically represented in (IV) or on Fig. 18. According to the energy part of the multiplet theory the interaction he-

ACTIVE C E N T E R S AND CATALYTIC DEHYDROGENATION

121

V

IV

m I1 I FIG.19. Stratification of the active complex in catalysis (77)

tween the atoms of layers I1 and I11 (that essentially determines the chemical reaction) in the limit is given by the equations for the heats of formation E‘ and of disruption E” of the multiplet complex M:

E’

=

E”

+

(-QAR

=

QAK

-

(&AD

QAK

+

-

QBK) QDK)

+ +

+

(-Qco (QAC

-

QCK

QBK

+

-

QDK)

QCK)

(32) (33)

where Q is the bond energy and K the catalyst. The more easily the reaction proceeds, that is, the higher the specific rate, and the lower the temperature a t which i t takes place, the smaller the potential barrier ( - E ) . E is the smaller of the two quantities E! and E”. Thus, the reaction should proceed more easily, the larger the value of E. The specific activity and selectivity of catalysts depends upon the nature of the atoms A, B, C, D, and K, since on them depends the value of Q. But the Q are variable quantities; QdB , Q c o , Q A D , and Q B C (layer IJI), having their tabular values as basic, vary with the substituents a t A, B, C . and D (layer IV) ; Q A x , Q B K , Q C K , and Q D K , analogously vary with the neighboring catalyst atoms (layer I), i.e., with the method of preparation, the promotors etc. (cf. Figs. 12c and 12d). Introducing the designations, u

=

-&A

s =

QAB

q =

QAK

+ + + + + + +

B

- QCD

+

QAD

Qnc

(34)

Q ~ D

&AD

Qnc

(35)

QBK

QCR

QDK

(36)

where u is the heat of reaction ( V ) ,s the sum of the energies of the dis-

122

A . A. BALANDIN

0

- 10 - 20 - 40 - 50 - 60

- 70 1 FIG.20. Volcano-shaped curves for the dehydrogenation of hydrocarbons (I j and alcohols (11) and dehydration of alcohols (111) on C r r 0 3 (85). For E arid q, the units are kg. cdories.

rupting bonds and of the bonds being formed in the molecule, and q the absorption potential. Equations (32) and (33) then give

B’

=

(42)

E” = ( 4 2 )

- ($2)

+q

+ (42) - q

(37) (38)

This is graphically represented by a broken volcano-shaped line (Fig. 2 0 ) . ?he cuordinates of the peaks of these curves are Eo = u/2, qo = s / 2 . From the peaks straight lines extend a t an angle of f 4 5 ” to the axis of abscissae. Each reaction has its particular volcano-like curve, arid each catalyst (as well as the group of atoms A , B , C , and D)its secant E, for example, FG, Fig. 20. It is found that the activation energy EZ

-XE

(39)

Since the optimal ordinate E, equals u/2, this permits one from the value of E to judge how far the given catalyst stands from its maximum specific activity. Since the optimal abscissa qo is s/2, then, by selecting the proper sum of s in Equation (37) and b y varying the bond energy values Q A K, QBK, Q C K ,and Q D K , we can expect to obtain the optimum catalyst for a given reaction. The mean bond energies with the catalyst may be obtained from thermochemical measurements. By means of these and Equa,tion (30) the author (2, 3) had already some time previously calculated the sequence in the ease of rupture of bonds on Ni, namely, N-0

>

C-Cl

>

C-0

> C-N > C--C

where > stands for “easier.” This permitted calculation of the sequence of hydrogenolysis for hundreds of complex conipounds, almost without excep-

ACTIVE CENTEES AND CATALYTIC DEHYDROGENATION

123

tions (2, 3 ) . Recently, the author and Ponomarev (79) by means of Equations (37) and (38) have calculated the analogous sequence for approximately a hundred conversions of furane compounds on Ni. Thus, the theory in agreement with experimental facts easily shows that on hydrogenating over Ni, 1-(alpha-furyl) pentene-1-one-3

there are formed consecutively: (1) 1-(alpha-furyl) pentanone-3; (2) 1(alpha-furyl) pentanol-3; (3) l(a1pha-tetrahydrofuryl) pentanol-3; (4) 2-ethyl-l , 6-dioxa(4,4) spirononane; (5) n-nonane and (6) methane. We have also been able to predict and carry out new reactions (1.4, 21, 41, 80). Quite recently in the investigation by Balandin et al. (81) it was shown that hydrogenation of the peroxide bond on Ni precedes that of the olefin and the latter that of the C-0 bond, in accord with the theory. For example, cyclohexene hydroperoxide a t 30" and 1 atm. first yields cyclohexenol and then only cyclohexanol. Besides the thermochemical method the author (78) has proposed a kinetic method for measuring the energies of binding with the catalyst, possessing the advantage that the results refer to the atoms of the active centers. The essence of the method is the following: Consider three types of duplet reactions, for example, dehydrogenation of a hydrocarbon (I), dehydrogenation of an alcohol (11),and, dehydration of an alcohol (111). From data on runs with tracer atoms, it was shown (82) that on certain oxides these reactions possess the same mechanism which is atomic but riot ionic. The basic process underlying them is concentrated in the groups: C-C 1 1

H

I

H

c-0 11

I

H

I

H

C-C

111

I

H

I

O

(VII)

Here there are three atomic species and in accordance with the number of reactions three equations may be written of the type of Equation (32). l+om kinetic measurements, determining cI , eII , and eIII and making use of the expression (39) we can solve these equations with respect t o QaK , QcK , and Q o ~ since , the energies of the bonds between the atoms H, C, and 0 are known. The method is treated more fully elsewhere (82). This method was applied in our laboratory to a number of oxides (82), in particular to ZrO:, in the work of Balandin et al. (83).I n the investigations of the author and Tolstopyatovtl (84,85,86)the effect of the method of preparation of Crz03on the bond energy with the catalyst was deter-

35

29

29 29

129.6

129

155

178

33.2

23.8

-

145

44.9

11.9

145

152

39.0

152 152

I

158

184 183

149 147

147

I

130.2 186.9 147.6 50

135.0 187.1 143.8 46

14.1 66.2

50 50 51 51 192.6 192.0 192.6 192.0

150.0 147.3 150.0 147.3 130.4 130.2 129.6 129.4

134.8 192.4 149.8 46 135.0 192.2 147.5 46

12.4 70.0 11.4 68.6

11.3 73.5 10.2 72.0 11.1 74.1 10.0 72.6

_______ __

TABLE 11 rimental Activation Energiesa t; Bond Energies of Atoms i n the Reacting Molecules with the Catalyst Q a ;~Adsor d the Heights of the Potential Barriers E k g . cal./mole on Chromias of Diflerent Methods of Preparation; the Subsc I-Dehydrogenation of Hydrocarbons, 11-Dehydrogenation of Alcohols and Acids, 111-Dehydration of Al

TABLE I1

52.9 51.5

130.1 130.: 129.t 129.'

50.6 50.6 51.5 51.5

8.5 8.5 8.9

17.5 20.0 17.5 20.0

53.4

14.1

66.2

135.( 187.1 143.8 46.0

14.0

19.1

52.2

12.9

69.6

130.:

186.C 147.6 50.6

14 .O

19.7

55.8 56.1

20.0 19.8

52.9

152 152

184 183

16.5 18.6: 17.6 2 0 . 3

43.6

32.2

39.0

152

158

147

29.3

60.7

11.9

44.9

145

178

129

35.8 't2.8 38.1

48.8

23.8

33.2

145

155

129.6 35.9

13.1 15.0 13.1 15.0

150.0 147.3 150.0 147.3

43.6

on asbestos, from chromium nitra te precipitated with sodium carbonate (Ref. 88)

88

__

r,O3

51.5

192.f 192.( 192.t 192.1

__

73.5 72.0 74.1 72.6

48.8

11.3 10.2 11.1 10.0

60.7

53.9 54.9 53.7 54.7

32.2

20.0 19.8 55.8 56.1

__

69.6

12.9

134.1 192.4 135.( 192.2 147.5 46.0

53.4

70.0 68.6

53.9 54.9 53.7 54.7

52.2

on asbestos, from chromium nitrate preeipitated with sodiumcarbonate (Ref. 88)

4. (:rpOa

12.4 11.4

r,03 from chromium nitrate precipitated with ammonia (Ref.

88 1

QO K

- -.

atalyst No. 7

3. Cr,Ol from chromium nitrate precipitated with ammonia (Ref.

QCK

Fi5.0 56.1

6.4 6.7 6.4 6.7

2. Catalyst X o . 7

55.0 56.1

nitrate preripitat,ed with sodium carbonate

r,03 from chromium nitrate precipkated with sodium carbonate

I . CrzC)3 from chromium

QCK

- - - -. -

Qaic __

-

Catalysts

Catalysts

- -

h'zperiinenlal Bctivation Energies" e; Bond Energies of Aton~si n the Reacting Molecules with the Catalyst Q.aK ; Adsorption Pvlentials q and !he Heights of the Polenlid Barriers E k g . cal./mole on Chromim of Diflerent Methods of Preparation; the Subscripts Designate: I-Dehydrogenation of Hydrocarbons, II-Dehydrogenation o Alcohols and A c i d s , IIZ-Dehydration of Alcohols:

8.5

8.9

8.9

42.6

46 . O

17.5 20.0

20.4

38.1

-

35.9 36.1 122

45

13.0 20 40.6 147

165

140.4 188

i-CSHrOH 9.8 15.3

I

\

I

145

14.8 35.1

62.1

9.8 15.3 i-CsH70H 9.S 15.3 9.8i-C3yH15.3

57.7

1 1

-

14.8

i-CsH7OH 9.8 15.3 i-CJ170H 26.9

1

; 56.9

33.9

16.3

57 . 9

146.5 188.01 147.4) 34.5

13.0

20 .o

i 57.8

17.0

55.7

149.6 188.3) 147.5 31.6

13.0

20.0

I 52.2

17.6

54.0

151.4 188.01 147

29.6

13.0

20.0

141.6 188.11 147.3 39.6

13.0

20 .O

40.6

13.0

20.0

36.1

35.9

45.2

I

1; 55.8

14.8

157.7

14.8

I

I

The are obtained by applying Eqiiation (23). * Tetralin.

ained by applying Equation (23)

l

33.9

os, oblcining bi-

1

9.gi-C3f;10Hi5.3 i-CaH70H 269.8 , g i - c a ~ ~ 15.3 oH

1

I

i-C,HTOH 9.8 1 15.3 i-CaH7OH 9.8 15.3 i-CaH7OH 9.8 15.3

bestos, ium bieduced in ipiammo-

15.0

1 55.4 i

'

I

I

61.5 62.1

140.4 188

/ 147 1 122 I I

i

a

__

I

i

I

1

i-CaH,OH

55.4

20 13.0 141.6 188.1 147.3 39.6 61.5 15.0 55.8

13.0 20 29.6 151.4 188.1 147 54.0

17.6

17.0 57.8

56.9

6. Cr& on asbestns, obtained by calcining ammonium bichromate

16.3

55.7

nia

52.2

with CH,OH in H&h , precipitated with ammo-

20 13.0 149.6 188.: 147.5 31.6

20 13.0

I

147.4 34.5

5 . Cr,Oa on asbestos, from ammonium bichromate reduced

-

35.1

145

165

-

126

A . A . BALAKDIN

mined. I n Table I1 are reproduced the results of the last study (86). On Fig. 20 the experinierital volcano-like curves and a number of secants for this case are presented, drawn to scale. Table I1 shows the following: 1. The method does enable determination of the bond energies Q H C r , QCCr, QOCr, and they are of reasonable value, since, for example, therniochemical methods give for nickel QHNi = 55, Q C N i = 19, QONi = 59 kcal. The reactions (VII) are actually retarded owing to the formation of M, for if Equation (33) instead of Equation (32) is taken for the disruption of M, then improbable values would be obtained. 2. Saturated hydrocarbon suhstituents in the molecule have little effect on the bond energies with the catalyst. On the contrary, the appearance of stabilization energy, for example, of the carboxyl groups, has a profound influence on Q A K . 3. The method of preparation of chromia greatly affects the bond energy with the catalyst. This can not be explained otherwise than by the influence on the active center of adjacent catalyst atoms. Especially sensitive is QOCr, perhaps because of the influence of the oxygen a t o m of the chromia. Thus, the active centers are to be conceived of as small, separated parts of the surface and not as the smooth continuous surface as a whole, for in case of the latter, such factors as change in the microrelief of the surface or possible unnoted microimpnrities could not exert such strong action on the bond energy. Hence, the kinetic measurements on deliydrogenatiori and dehydration also confirm the concept of the active centers as developed in the present report.

V. SUMMARY The geometry and energy aspects of the multiplet theory are useful in interpreting the kinetics of catalytic dehydrogenation and in throwing light on the nature of the active centers. The possibility was shown of the experimental determination of change in free energy, enthalpy, and entropy of adsorption on the active centers and of the bond energies between the reacting atoms of the molecule and the atoms of the active centers of the catalyst.

REFERENCES 1. Balandin, A. A . , Zhur. Russ. F i z . Khim. Obshchcstva, Chast’ Khini. 61, 909 (1929); 2. physik. Chem. (Leipzig) B2, 289 (1929). 2. Balandin, A. A., Zhur. Russ. Piz. Khim. Obshchestva, Ch,ast’ Khim. 62, 703 (1030); 2. physik. Chem. (Leipzig) B3, 167 (1929). 3. Balandin, A. A . , Voprosy Khim. Kinetiki, Kataliza, i Reaktsionnoi Sposobnosti, Akad. Nauk S . S. S . R., Otde2. Khim. Nauk (l’roblems in Chemical Kinetics, Catalysis and Reactivity), p. 461 (1955). 4. Balmdin, A. A . and Klabunovskii, E. I . , Doklady Akad. Nauk, S . S. S . R . 110 571 (1956); 113, 585 (1957).

ACTIVE CENTERS AND CATALYTIC DEHYDROGENATION

127

6. Taylor, H. S., J . Phys. Chem. 30, 145 (1926). 6. Volkenshtein, F. F., Zhur. Fiz. K h i m . 23, 917 (1949). 7. Balandin, A. A. and Marukyan, G. M . , Doklady. A k a d . Nauk S. S . S . 12. 66, 219 (1947). 8. Constable, F. H., Proc. Roy. Soc. A108, 355 (1925). 9. Schwab, G. M., “Katalyse vom Standpunkt der chemischen Kinetik,” Springer, Berlin, 1931. fa. Balandin, A. A . , Zhur. Fiz. K h i m . 4, 257 (1933); 2. Physik. Chem. (Leipzig) B19, 451 (1932). 11. Roginskii, S. Z., “Adsorptsia i kataliz na neodnorod. poverknostyakh,” (Adsorp-

tion and Catalysis on Heterogeneous Surfaces) Izdatelstvo Akad. Nauk S.S. S. R., Institut Fix. Khim., Moscow and Leningrad, 1948. 12. Balandin, A . A., Doklady Akad. N a u k S . S . S. R. 93, 55 (1953). f3. Hinshelwood, C. N., “Kinetics of Chemical Change in Gaseous Systems.” Oxford Univ. Press, London and New York, 1929. 14. Balandin, A. A., Z h u r . Fiz. K h i m . 14, 1160 (1940); Acta physicochim. U . R . S . S . 14, 223 (1941). 26. Frost, A . V., Vestnik Moseov. Univ. No. 3 4 , 111 (1946). 16. Hougen, 0. A . , I n d . E n g . Chem. 36, 529 (1943). 17. Pshezhetskii, S. Ya., Lubarskii, G. D., Shcheglova, N. A , , and Merilyainen, S. K., Z h u r . F i z . K h i m . 28, 1459 (1954). 18. Balandin, A . A . and Kiperman, S. L., Zhur. F i z . Khim., 31, 139 (1957); J. chim. phys., in press. 19. Balandin, A. A., Neiman, M. B., Bogdanova, 0. K., Isagulyants, G. V., Shcheglova, A. P., and Popov, E. I., Izvest. A k a d . Nauk S.S.S.R., Otdel. K h i m . N a u k 2, 157 (1957). 20. Balandin, A. A., Zzvest. Akad. Nauk S.S.S.R. Otdel. K h i m . N a u k 4, 723 (1955). 21. Balandin, A. A . , Doklady A k a d . Nauk S.S.S.R. 63, 33 (1948). 22. Balandin, A. A . , Doklady Akad. N a u k S.S.S.R. 93, 55, 272, 475 (1953). 23. Balandin, A. A., Sbornik Kataliz, T r u d y 111 Fiziko-Khimicheshoi Konferentzu. Catalysis, Proceedings of the Third Physico-Chemical Conference. Nauchnoe Khimico-Technicheskoe Izdatelstvo. Leningrad, 1930 p. 193. 24. Balandin, A. A., Zhur. Obshchei K h i m . 12, 135, 160, (1942). 26. Antipina, T. V., and Frost, A. V., Uspekhi K h i m . 19, 342 (1950). 26. Orochko, D. I., “Teoreticheskie Osnovy Vedenia Sinteza Zhidkikh Topliv”, (Theoretical Principles of the Syntheses of Liquid Fuels), Gostoptekhixdat, Moscow-Leningrad, 1951. 27. Panchenkov, G. M., Zhur. Fiz. K h i m . 22, 209 (1948). 28. Topchieva, K . V . , and Panchenkov, G. M., Dokludy A k a d . Nauk S . S . S . R. 74, 1109 (1950). 89. Nagiev, M. F., “Osnovy Khimicheskol Kinetiki Promyshlonnykh Sistem” (Fundamentals of Chemical Process Kinetics) Izdatelstvo. A kademii Nauk Azerbatdzhansko‘l S.S.R., Baku, 1950. 90. Balandin, A . A . , Bogdanova, 0. K., and Shcheglova, A . P., Izvesl. A k a d . Naiik S.S.S.R. Otdel. Khinz. N a u k 6,497 (1946). 31. Tolstopyatova, A . A,, Kheifets, 1,. A . , and Patrikeev, V. V., Vestnik Moskov. Univ. 2, 47 (1951). 32. Balandin, A. A,, and Shuikin, N. I., Z h u r . Piz. K h i m . 6, 707, 1934; Aeta physicoehim. U . R . S . S . 1, 200 (1934); Uehenye Zapiski Moskov. Gosudarst. Univ. im. M . V . Lomonosova 6, 281 (1936). 33. Bork, A. Kh., and Balandin, A. A., Z. physik. Chenz. (Leipzig) B33, 54 (1936).

128

rl. A . BALrlNDIN

34. Bork, A. Kh., Acta physicochim. U . R . S . S . 9, 409 (1939). 55. Bogdanova, 0. K. , Balandin, A . A . , and Shcheglova, A , P., lzvest. A k a d . Nuick

S . S . S . R., Otdel. K h i m . N a u k , 7: 787, 795; 8, 909 (1957). 96. Balandin, A. A . , and Klabunovskii, E. I., Doklady A k a d . N a u k S . S . S . R . 98, 783 (1954). 97. Balandin, A. A., and Bork, A . Kh., Uchenye Zapiski Moskov. Gesudarst. Univ. im. M . V . Lomonosova 2, 217 (1934). 38. Balandin, A. A . , and Bork, A . Kh., 2. phys. Chem. 33, 73 (1936). 59. Balandin, A. A . , and Isagulyants, G. V., Doklady A k a d . N a u k S . S . S . R. 63, 139 (1948). 40. Balandin, A. A , , and Isagulyants, G. V., Doklady A k a d . N a u k S. S . S . R . 63, 261 (1948). 41. Balandin, A. A . , and Vasyunina, H. A., Zhu,r. Obshche?:K h i m . 18, 398 (1948).

4%. Balandin, A. A . , and Vasyunina, N . A., Doklady A k a d . N a u k S . S . S . R . 103, 831 (1955).

43. Balandin, A. A . , and Vasyunina, N . A., Doklady A k a d . N a u k S. S . S .

IZ. 106,

981 (1955). 44. Agronomov, A. E., Vestnik Moskov. Univ., Se,. K h i m . 2 , 109; 11, 41 (1951). 46. Balandin, A. A., arid Itubinshtein, A. M., Z h u r . Fiz. K h i m . Ei, 12 (1934); 2 . phys. Chem. (Leipzig) A167, 431 (1933). 46. Balandin, A . A,, and Yur’ev, Yu. K., Zhur. P i z . K h i m . 6, 393 (1934); Acta phisicochim. 1, 183 (1934). 47. Balandin, A . A , , Zhur. Obshchel; K h i m . 16, 608 (1945). 48. Zelinskii, N . D., and Balandin, A. A,, Izvest. A k a d . N a u k S . S . S . R . p. 20 (1929) ; 2.pkys. Chem. (Leipzig) 126, 267 (1927). 49. Balandin, A . A , , and Tolstopyatova, A. A,, Zhur. ObshcheZ K h i m . 17, 2182 (1947). 50. Bork, A. Kh., and Darykina, M. I., Zhur. F i z . K h i m . 9, 330 (1937); Acta physicochim. U . R . S . S . 6, 443 (1936). 61. Bork, A. Kh., Zhur. Fiz. K h i m . 13, 421 (1939). 52. Bork, A . Kh., and Balandin, A . A., 2. phys. Chem. (Leipzig7) 33, 435 (1936). 55. Balandin, A . A., Marushkin, M. N., and Ikonnikov, B. ,4.,Uchenye Zapiski Moskov. Gosudarst. Univ. im. M . V . Lomonosova 2, 221 (1934). 64. Bork, A . Kh., Zhur. F i z . Khim.8, 709 (1936); 2. phys. Chern. (Leipzig) B33, 443 (1936). 55. Tolsto:iyatov:i, A. A., Balandin, A. A., and Dmitriev, E. A . , lzvest. A k a d . N a u k S . S . S . R . , Otdel K h i m . N a u k , 11, 1404 (1956). 56. Balandin, A . A., Doklady A k a d . Nauk S . S . S . R . 63, 535 (1!)48). 67. Dohse, H., Z . phyy. Chem. B6, 351 (1930). 68. Bork, A. Kh., and Tolstopyatova, A . A., Zhur. F i z . Khim. 12, 227,236, 245 (1038) ; Acta phgsicochim. U . R . S . S.6, 577 (1938). 59. Bork, A . Kh., Zhur. F i z . K h i m . 14, 1288 (1940). 60. Parks, G . S., and Huffman, H . M., “Free Energies of Some Organic Compounds,” (Am. Chem. Soc. Monograph ser. no. 60), Reinhold, 1932. 61. Long, J. A., Fraser, J . C . W . , and Ott, F., J. Am. Ch.em. floc. 66, 1101 (1934). 62. Emmett, 1’. N., and Skau, N . , J . A m . Chem. Soc. 66, 1059 (1943). 65. Beeck, O., Revs. Modern P h v s . 17, 79 (1945); Beeck, O., and Richie, A . W., Discussiom Paraday SOC.8, 159 (1950). 64. Trapnell, B. M. W., Advances in Catalysis 3, 1 (1951). 66. Rienacker, G., and Unger. S., 2 . anorg. Chem. 274, 47 (19%). 66. Balandin, A . A., and Brusov, I. I., Zhur. ObshcheZ K h i m . 7, 18 (1937); Z . phys. Chevi. (Leipzigl B34, 96 (1936).

ACTIVE CENTERS AND CATALYTIC DEHYDROGENATION

129

67. Rubinuhtein, A . M., Shuikin, N . I., and Minachev, K h . M., Doklady Akad. Nauk S . S . S . R. 67, 287 (1948). 68. Sachtler, W. M. H., Dargelo, I,., and van der Knaap, W., J. chim. phys. 61, 491 (1954). 69. Beeck, O., Smith, A. E., and Wheeler, A., Proc. Roy. SOC.A177, 62 (1940); Beeck, O., Discussions Faraday SOC.8, 118 (1950). 70. Balandin, A. A . , and Isagulyants, G. V., Doklady Akad. Nauk S. S . S . R. 64, 207 (1949). 71. Rideal, E. K., Chem. & Ind. (London) p. 335 (1935). 72. Twigg, G. H. T., and Rideal, E. K., Trans. Faraday SOC.36, 533 (1940). 73. Herington, E. F. G., Trans. Faraday SOC.37, 361 (1941). 7.4. Balandin, A . A . , Doklady Akad. Nauk S . S . S . R. 97, 449 (1954). 7 5 . Balandin, A. A , , Zhur. Obshcher Khim. 16, 608, 619, 770, 781 (1945); 16, 557 (1946). 76. Balandin, A. A . , and Rubinshtein, A . M., Zhur. Fiz. Khim. 6, 576 (1935); J . A m . Chem. SOC.67, 1143 (1935). 77. Balandin, A . A . , Doklady Akad. Nauk S . S . S . R. 97, 667 (1954). 78. Balandin, A. A . , Zhur. Obshche'l Khim. 16, 793 (1946). 79. Balandin, A . A , , and Ponomarev, A . A . , Doklady Akad. Nauk S. S . S . R. 100, 917 (1955); Zhur. Obshch. Khim. 26, 1146 (1956). 80. Balandin, A . A . , and Vaskevich, D. N., Zhur. Obshchei Khim. 6, 1878 (1936). 81. Balandin, A . A . , Freidlin, L. Kh., and Nikiforova, N. V., Izuest. Akad. Nauk S . S. S. R., Otdel. Khim. Nauk 4, 443 (1957); Doklady Akad. Nauk S . S . S . R. 112, 649 (1957). 82. Balandin, A . A,, Izuest. Akad. Nauk S . S . S. R., Otdel. Khim. Nauk 4, 624 (1955). 83. Balandin, A . A . , Tolstopyatova, A . A . , and Ferapontov, V. A., Doklady Akad. Nauk S . S . S. R. 103, 611 (1955). 84. Balandin, A . A . , and Tolstopyatova, A . A . , Doklady Akad. Nauk S . S . S . R . 94, 49 (1954). 86. Balandin, A . A., and Tolstopyatova, A. A . , Zhur. Fiz. Khim. 30, 1367 (1956). 86. Balandin, A. A . , and Tolstopyatova, A . A . , Zhur. F i z . Khim., 30, 1636 (1956). 87. Balandin, A. A., and Kukina, A . K., Doklady Akad. Nauk S. S. S . R. 64, 65 (1948). 88. Tolstopyatova, A. A . , Balandin, A. A., and Dulitskaya, K. A . , Zzvest. Akad. Nauk, Otdel. Khim. Nauk, 10, 1256 (1956).