Fischer-Tropsch Synthesis. Nitrides and Carbonitrides of Iron as

of e-iron nitride to e-iron carbonitride during normal synthesis. A study of activity and selectivity of partly nitrided catalysts as a function of in...
0 downloads 0 Views 697KB Size
I

J. F. SHULTZ, MANUEL ABELSON, LEONARD SHAW, and R. B. ANDERSON Central Experiment Station, Bureau of Mines, U. S. Department of the Interior, Pittsburgh, Pa.

Fischer-Tropsch Synthesis Nitrides and Carbonitrides of Iron as Catalysts The iron-carbon-nitrogen system offers a unique opportunity to prepare solid solutions of varying composition and to study their catalytic activity and selectivity as a function of composition

IN

THE Fischer-Tropsch synthesis nitrided iron catalysts (7, 5, 6, 74, 75) show improved stability against the deterioration caused by oxidation and free carbon deposition, and they preferentially catalyze the synthesis of alcohols, other oxygenated products, and lower boiling liquid hydrocarbons. After a short period of operation the nitride acquires appreciable amounts of carbon and becomes essentially carbonitride. Therefore, as the catalyst spends most of its lifetime in the carbonitride stage, this is the most important form to consider. The normal form of freshly nitrided iron is e-iron nitride. This phase is isomorphous with €-iron carbide, and a continuous series of solid solutions in the €-phase, ranging from pure carbide to pure nitride, is theoretically possible. But while phases corresponding to the two terminal compositions have been realized, compositions of carbonitride have been prepared only in the range between nitride and an atomic ratio of carbon to nitrogen of -2 to 1 (10, 77). Not only are carbonitrides with a wide range of carbon-nitrogen ratios experimentally realizable, but the ratio of iron to total interstitial elements (nitrogen plus carbon) can also vary between 2 to 1 and 3 to 1. There is apparently a critical carbonnitrogen ratio which, when exceeded, causes decomposition of e-carbonitride to X-iron carbide. I n this transformation the nitrogen from the €-iron carbonitride (1) may go into the X-phase, which then would really be X-iron car-

bonitride, (2) it may go into a phase of such small crystallite diameter that the diffraction pattern is too diffuse to be detected, or (3) it may be eliminated as gas. This critical ratio varies from about 2 to 1 a t 450' C. to 03 a t 180' C. This conclusion is based on the following experimental facts. Jack (77) and Hall, Dieter, Hofer, and Anderson (7) have shown that eiron nitride is converted to carbonitride by the substitution of carbon atoms for nitrogen atoms when treated with carbon monoxide, but the compositions accessible in this manner a t 400' C. range only from the nitride to a product with a carbon-nitrogen ratio of about 2 to 1. Further carburization leads to the formation of X-iron carbide. Direct carbiding of iron catalysts a t 190' C. with carbon monoxide leads to the formation of €-iron carbide, isomorphous with e-iron nitride (70, 77). Thermal conversion of pure e-iron carbide to X-carbide proceeds at a measurable rate a t 240' C., and probably a t lower temperatures. The critical carbon-iron ratio is not affected by gas phase composition, as the reactions involved can proceed entirely in the solid state. The iron-carbon-nitrogen system offers a unique opportunity to prepare by various methods a large number of solid solutions whose components can be varied over a wide range of composition, and to study their catalytic activity and selectivity as a function of composition. These same factors can be studied in terms of solid state reactions which occur in the composition range where

the e-carbonitrides are unstable. The present discussion records and discusses results of work bearing on both these objectives. Specifically, it presents: A detailed study of composition changes in fully nitrided catalysts during conversion of e-iron nitride to e-iron carbonitride during normal synthesis. A study of activity and selectivity of partly nitrided catalysts as a function of initial nitrogen content, correlated with a less detailed study of composition changes. A similar study of activity and selectivity of carbonitrides of various compositions directly synthesized by conducting, in proper sequences, carbiding with carbon monoxide, nitriding with ammonia, and simultaneous carbiding and nitriding by the use of methylamine.

Preparation of Iron Carbonitrides All the catalysts were prepared from 6- and 8-mesh particles of a standard promoted fused magnetite: Fez04 93.4670, MgO 4.61%, Si02 0.71'%, Crz03 0.6570, and KaO 0.57%. The magnetite was reduced by treating with a stream of hydrogen (6, 8) at 450' to 550' C. for 20 to 40 hours, a t a space velocity of 2500 hr,-l The nitride was prepared by passing ammonia (6)over the reduced catalyst at 350' to 500' C. for 0.5 to 22 hours at a space velocity of 1000 hr.-l The variation of time and temperature was required to produce specific nitrogen contents in the catalyst. The carbonitride was prepared in four ways. A. The nitrided catalyst was placed in the testing apparatus (3) and operated in the synthesis with 1Hz f 1CO gas a t 100 or 300 pounds per square inch VOL. 49, NO. 12

DECEMBER 1957

2055

Table I.

Preparation and Properties of Carbonitrides Formed in Synthesis All synthesis tests at 21.4 atm. except where noted)

(All pretreatments a t atmospheric pressure.

Composition of Activity' Reduction in Hydrogen" Nitriding inAmmoniab Pretreated Catalysts Composition of Used Catalysts per Gram Temp., Time. ReducTemD., Time. Comaonent Comoonent Fe. Test ' C. hr. tion, % ' Cf6 hr, W/Fe phases" C/Fe N/Fe O/Fe phases A F* X517 550 24 96.5 None 0.0 M, cy 0.28 0.0 1.05 e', X, S, M 69 X274 550 20 99.0 500 1 0.07 a, y' 0.23 0.02 0.99 M, x 63 550 20 98.3 250 1 0.11 cy, y' 0.34 0.01 1.03 M, a, x 90 X447 X766 450 40 98.4 350 0.5 0.16 cy, y' 0.17 0.09 1.11 M, cy 75 x445 550 20 99.1 450 1.5 0.18 y', a 0.38 0.03 0.95 M, XI 99 550 20 97.4 385 4 0.26 y', a 0.35 0.1B 0.63 Y', e 109 x219 X617 550 20 100 250 10 0.35 y', e, a 0.51 0.08 0.80 e, M 88 X711 450 40 97.4 350 8 0.38 e, y' 0.32 0.06 0.94 M, 6 76 X225 550 20 96.4 350 6 0.41 I, E 0.40 0.14 0.56 e, M 115 X612 550 20 100 350 6 0.45 e 0.36 0.20 0.76 6, M, S 81 X620 550 20 100 350 12 0.46 e, M 0.41 0.23 0.62 e) M, S 126 X518 550 24 96.5 35 0 6 0.46 E 0.35 0.19 0.58 6, M, S 95 X670 550 20 99.8 350 22 0.47 E 0.36 0.26 0.36 € 9 M, s 111 X219 550 20 98.7 385 4 0.46 e 0.36 0.31 0.23 e 70 X218* 550 20 97.5 385 8 0.43 E 0.42 0.23 0.15 e, M 68 X361 500 24 97.2 None 0.0 01, M 0.18 0.0 1.14 M, a: 54 X194E 450 40 None 0.0 a 0.31' 0.0 1.09" M, XI 27 a Space velocity 2300 t o 2700 hr.-' in all tests. Space velocity 950 t o 1050 hr.-1 in all tests. For definition of activity see ( 4 , p . 392). a = a-iron, x = H&gg carbide, M = magnetite, S = MgC03 or FeC08, y' = ?'-nitride or carbonitride, E = cnitride or carbonitride, r-nitride, and e' = hexagonal close-packed carbide. Phases tabulated in order of decreasing intensity of x-ray diffraction pattern. Tested a t 7.8 atm. O/Fe and C/Fe after 95 days of synthesis. ff

...

gage and 100 or 300 hr.-I space velocity. The temperature a t any pressure and throughput was varied to maintain 65% contraction due to synthesis after absorption of carbon dioxide. B. The nitrided catalyst was treated with a stream of pure carbon monoxide at atmospheric pressure and at 350' to 450' C. for 6 to 10 hours, a t lOO-hr.-l space velocity. The higher temperature and longer time were required to produce a carbonitride rich in carbon and low in nitrogen. C. The reduced catalyst was carbided with a stream of pure carbon monoxide flowing a t a space velocity of 100 hr,-I, while the temperature of the catalvst was slowly raised from 150' to 250' C. The exit gas was continuouslv analyzed bv a thermal conductivity method, and the rate of temperature rise was manually controlled to keep the carbon dioxide produced by the carburization below 207& This precaution (9, 77, 72)

Table II.

was necessary to prevent oxidation. The carbided catalyst was then treated with ammonia a t 350' C. for 28 hours, at a space velocity of 1000 hr.+ D. The reduced catalyst was treated with monomethylamine at atmospheric pressure and 250' C. for 16 hours? at a space velocity of 1000 hr.-I At the end of procedure B, 6, or D, the catalyst was cooled to below 50' C. in the last gas used in the pretreatment (6). With the exception of ammonia, the gas was then swept out with a stream of carbon dioxide (6) and all subsequent transfers of the catalyst were made in a carbon dioxide atmosphere to prevent oxidation. Ammonia was swept out with prepurified nitrogen and the nitrogen replaced with carbon dioxide. The testing apparatus and method of conducting synthesis experiments, with methods for characterizing synthesis

products, have been described (3, 6). Pretreated and used catalysts were analyzed by chemical methods for iron, carbon, and nitrogen, and oxygen was determined by difference. Phases present were identified by x-ray diffraction. The nature and sequence of pretreatment steps, chemical composition, and phases identified in the pretreated catalyst are given in Tables I and II. Throughout this discussion C/Fe, N/Fe, and O/Fe represent the atomic ratio of carbon, nitrogen, and oxygen to iron.

Behavior of Nifride and Carbonitride Catalysts Composition Changes in Converting Nitrided Catalysts to &arbonitrides by Synthesis Gas. Completely Nitrided Catalysts. Figures 1 to 4 present the behavior of catalysts prenitrided but not

Preparation and Properties of Presynthesized Carbonitride Catalysts

(All pretreatments at atmospheric pressure.

All synthesis tests at 21.4 atm. except where noted) Reduction in Composition of Catalyst Composition ActivHydrogena Third Pretreated Catalysts after Synthesis ityC per Reduc- Second Pretreatment* Pretreatment5 CompoCompoGram Temp., Time, tion, Temp., Time, Re- Temp., Time, nent nent Fe, Test O C. hr. % Reagent C. hr. agent O C. hr. C/Fe K / F e O / F e phasesd C/Fe N / F e O/Fe phases Ape X342 500 24 97.2 CO 150-350 I8 0.57 0.0 0.07 x, a 0.49 0.0 1.48 M,a , S 95 X2948 500 24 96.6 CO 150-350 I8 ,I 0.58 0.0 0.03 X, 01 0.71 0.0 0.37 X, M 52 X287' 550 20 97.9 NH3 350 6 CO 450 10 0.62 0.02 0.16 x, M 0.63 0.02 0.32 x, M 65 X252' 550 20 96.4 NHB 400 1 CO 350 6 0.19 0.29 0.03 B e 69 X279 550 20 97.8 NHa 400 1 CO 350 6 0.30 0.23 0.08 e 0.42 0.15 0.60 e, M, x 108 X407 550 22 96.0 CO 150-250 12 NH1 350 28 0.21 0.29 0.13 E 0.44 0.08 0.81 e, M 93 X739 450 24 63.4 CHaNHz 250 16 0.15 0.05 0.85 M, e 0.24 0.02 1.09 M,e 114 Z , Space velocity of 2300 t o 2700 hr.-1 in all tests. Space velocity of NHs 950 t o 1050 hr.-L Space velocity of CO 95 to 105 hr.-1 Space velocity of methylamine 1000 hr.-I For definition of activity see (4, p. 392). Symbols for catalyst phases in Table I. * Tested at 7.8 atm.

... ... ... ..

... ... ...

2056

INDUSTRIAL AND ENGINEERING CHEMISTRY

FISCHER-TROPSCH S Y N T H E S I S COMPONENT PHASES,

Figure 2. Nitrogen content during synthesis as a function of time

USED CATALYST

05

-

.4

9

\

z 'P .m t .3 0

in.

P

u Q

Y IO

0

20

30

TIME, DAYS

I 40

Figure 1. Composition changes of a nitrided iron catalyst during synthesis lCOat21.4atm. with 1Hz

+

carburized prior to service in the synthesis. Similar curves for catalysts canditioned only by a standard reduction procedure are included in Figures 3 and 4. Pertinent details concerning the pretreatments, with descriptive analytical information obtained before and after service, are found in Table I. Figure 1 shows the composition changes that occurred in a typical completely nitrided &on nitride catalyst (test X-670, Table I) during a synthesis experiment at 21.4 atm. Carbonitride was formed during synthesis

7 8 otm. X 218

z .3

-

A

A 1

A

8

X 670 .2

slope of the O/Fe curve and the appearance of magnetite (M) and possibly siderite or magnesite (S) as major phases after 20 to 30 days of operation show that oxidation of iron is progressive in nitrided catalysts, as in the reduced ones. Both nitrided (tests X-218, and X-670, Table I ) and reduced catalysts (tests X-194 and X-361) showed increased rates of oxygen absorption with increasing pressure (Figure 3). Comparison of the performance of a nitrided with a reduced catalyst tested a t the same operating pressure shows that nitriding substantially retarded oxidation (2, 72, 73). As shown in Figure 4, the nitrided catalysts (tests X-218 and X-670, Table I) absorbed carbon at least as rapidly as reduced catalysts (tests X-194 and X-361). I n fact, a t the higher operating pressure the nitrided form showed both the highest initial absorption rate and the greatest ultimate carbon content observed in these four experiments. The apparent decrease in carbon uptake of the reduced catalyst a t the higher pressure is too small to be considered signifi-

by reaction with carbon monoxide in the synthesis gas. The comparatively rapid rate of nitrogen loss during the first few days of the experiment, and the rapid carbon uptake and persistence of the e-phase as determined by x-ray diffraction, are evidence of warbonitride formation. Carbon absorption, however, was considerably greater than nitrogen depletion, which is interpreted to mean deposition of free carbon during this period. With the passage of time, the rate of loss of nitrogen decreased substantially, so that a t 35 days of operation N/Fe was still more than half the initial value. As shown by Figure 2, the retention of nitrogen was not significantly affected by operating pressure during synthesis at 7.8 to 21.4 atm. The data a t 21.4 atm. are reproduced from Figure 1;those at 7.8 atm. refer to test X-218 (74), Table I. Iron catalysts are gradually oxidized during Fischer-Tropsch synthesis. I n Figure 1, the rate of this reaction may be compared with rates of absorption of carbon and loss of nitrogen. The

7 8 otm. x 218

Nitrided

I

I

0

Figure 3. Oxygen content during synthesis as a function of time Figure 4. Carbon content during synthesis as a function of time

j

-

LL

0-

IO

=. 20

I

I

1

I

H 21.4atm

X361

30 TIME, DAYS

VOL. 49, NO. 12

7.8atm. X 194

40

1

50

60

DECEMBER 1957

2057

$

/-I

120

100

t 0

a an

I/.

*I

A

c/

4

~~~~~

40

0

02

01

03

04

05

INITIAL N/Fe

Figure 5. Variation of activity with initial nitrogen content

0

A

Unrieved catalyst Sieved catalyst

cant without additional experimental verification. PARTLYKITRIDEDCATALYSTS. The influence of initial nitrogen content of catalysts on their performance in the synthesis was studied at 21.4-atm. operating pressure, with 1Hz 1CO synthesis gas. Time and temperature of nitriding were varied over wide limits to attain the desired range of nitrogen content. I t is doubtful if this temperature variation in itself influenced the behavior of the catalysts, as reduction in any given instance was effected at a higher temperature than nitriding. At the lower nitrogen-iron ratiosup to 0.26-nitriding produced y '-iron nitride and a-iron as the major components [Table I, tests X-517 (control

+

40t-

test), X-274, X-447, X-766, X-445, and X-2191. In the synthesis, the y' phase proved unstable; in every experiment except X-219, it disappeared completely and is assumed to have been transformed to magnetite, X-iron carbide, and a-iron, In the two tests a t intermediate K/Fe ratios (X-617 and X-711) where the y' phase was initially present, it was also destroyed during the synthesis. EIron carbide, also present a t the beginning, was detected as a major phase after the tests were ended. Upon nitriding to N/Fe values above 0.40, €-nitride was the principal phase formed (Table I, tests X-225, X-612, X-620, X-518, and X-670). During synthesis, this was converted to €-iron carbonitride, which persisted as a major component throughout the experiments. Despite the fact that the nitrogen content of the catalysts decreased with the passage of time, the &on carbonitride structure was maintained and the products of synthesis retained the characteristics associated with nitrided catalysts even at N/Fe values as low as 0.14. Activity and Selectivity of Partly Nitrided Catalysts. In Figure 5, activity, defined as milliliters of synthesis gas (HZ CO) converted per gram of iron per hour a t 240' C., is plotted as a function of the initial nitrogen-iron ratio, The activity was computed by an empirical rate equation (4). In a number of tests with nitrided iron catalysts, more variability in activity was observed than in other synthesis tests. Examination of the catalyst after reduction and nitriding indicated that some particles had disintegrated to a fine powder, and because the activity increases with decreasing particle size, marked deviations of this function from normal values were observed. Fines are also formed in the reduction of catalysts, but to a smaller

+

extent, and reproducibility of activity is not seriously affected. Particle disintegration during pretreatment was fully appreciated only after the first part of the data for Figure 5 had been obtained. Subsequently, several tests were made with catalysts that had been sieved under heptane to remove fines produced in the pretreatment. In both sets of data, activity increased with nitrogen content u p to N/Fe values of 0.20 to 0.25 and then remained substantially constant. X-ray diffraction examination of catalysts after synthesis showed that those in which the initial N/Fe was 0.25 or greater generally contained the €-phase as the major component, whereas in those containing smaller amounts of nitrogen, magnetite predominated; Product distribution or selectivity as a function of initial N/Fe is shown in Figures 6 and 7. In these curves, the term "products" includes oxygenated organic compounds dissolved in the oil phase. Figure 6 is a plot of synthesis products exclusive of carbon dioxide, water, and water-soluble chemicals. The curves are based on mass spectrometric analysis of product gases and a simple distillation of liquid fractions. Figure 7 shows the weight per cent of C=C in the a and p positions and of total alcoholic hydroxyl in the fraction boiling from room temperature to 185' C. These determinations were made by the infrared spectrometer. Figure 8 shows the oxygen content of the catalyst after about 6 weeks of synthesis as a function of initial nitrogen content. Composition Changes in Presynthesized Carbonitrides during Operation. T o study further the effect of preparational variables (Table 11), carbonitrides were prepared by nitriding reduced catalysts in ammonia and then carburizing with carbon monoxide (X287, X-252, X-279): by carburizing with carbon monoxide and then nitriding with

zoby Froctim

10

Froctim

0 0

01

02

03

04

0 5

N/ Fa

Figure 6. Effect of initial nitrogen content of catalysis on distribution of products

2058

Figure 7. Effect of initial nitrogen content of catalysts on alcohol and olefin production

INDUSTRIAL AND ENGINEERING CHEMISTRY

Figure 8. Effect of initia! nitrogen content of catalysts on oxidation of catalyst during synthesis

FISCHER-TROPSCH SYNTHESIS ammonia (X-407), or by simultaneous carbiding and nitriding with methylamine (X-739). For comparison, carbide catalysts containing no nitrogen are included (X-342, X-294). In the ensuing synthesis tests, the e-iron carbonitride once formed persisted throughout the experiments. The catalyst of test X-739 was not completely reduced before treatment with methylamine; however, based on the portion of the catalyst that was reduced, N/Fe and C/Fe were, respectively, 0.08 and 0.24. As only the outer layer of the catalyst is effective in the synthesis, this test is probably comparable despite the lower extent of reduction. Activity and Selectivity of Directly In Presynthesized Carbonitrides. Figures 9 and 10, the activities and selectivities of the carbonitride catalysts of Table I1 are shown for experiments conducted at 7.8- and 21.4-atm. operating pressure. The performance of carbide (tests X-294 and X-342, Table 11) and nitride catalysts (tests X-215 and X-225, Table I) is included for comparison. Increased production of alcohols and decreased production of high-boiling products were characteristic of the catalysts containing nitrogen, with the exception of test X-287, where nitriding was insufficient to produce the €-phase when subsequently carburized with carbon monoxide. As a result, only X-iron carbide was formed, and therefore this catalyst behaved more like the carbide form in test X-294. Activity values for enitrides and ecarbonitrides prepared by carburizing nitrides were usually greater than for reduced or carbide catalysts, or for e-carbonitrides prepared from carbide.

I

TEST NUMBER

X 287

Fe

0.02

I

0.29

I

G / Fe

062

I

0.19

I

INITIAL N / INITIAL

TEMPERATURE, *C

227

X252

I

222

X 215 0.46

0

I

I

222

SPACE VELOCITY, hr - I CONTRACTION, %

INITIAL

63 3

I COMPONENT PHASES 1

ACTIVITY.

AS.

53 v.a I

I

CI

80

F c3+

I0 W z -

a W a

70

c I

CI

Cl

-

c4

= 79

1

I

c 3 + c4

COOHf

I

27'- 185'C Er IO OH 135

i

185"-352OC Er 45 OH 0.4

Br

26

%OHf

I L 2464°C.

6 *

0

352O-464OC'

Figure 9. Activity and product distribution from synthesis at 7.8 atm., carbonitride catalysts, 1 Hz 1 CO gas

+

Values for unsaturated hydrocarbons ( = 1, alcoholic hydroxyl (OH), carboxyl group (COOH), and carbonyl group (CO) are in.weight per cent of fraction where found. Br denotes bromine number

Discussion

"I

Both nitrided ( 7 3 ) and reduced catalysts were oxidized more rapidly at 21.4 than a t 7.8 atm., but the nitrides absorbed oxygen at lower rates a t both pressures. Carburized catalysts were previously oxidized more rapidly a t the higher pressures (17, 72). The depletion in nitrogen was not noticeably affected by pressure. Specifically, this means that the percentage of iron present as ecarbonitride decreases more rapidly at the higher pressure and the carbon-nitrogen ratio in the €-iron carbonitride increases more slowly a t the higher pressure. Therefore, nitrogen is replaced by carbon more slowly relative to the oxidation reaction at the higher pressure, or carbonitrides rich in carbon are more rapidly oxidized than those that are carbon-poor. Although all catalysts studied were oxidized to an appreciable extent during synthesis, their activity and selectivity

jl

[with the exception of reduced (a-iron) catalysts a t 7.8 atm.] remained essentially constant for long periods. Apparently only a thin layer of catalyst below the geometric surface of the particle is effective in the synthesis. The pores of the catalyst are probably filled with liquid hydrocarbons at synthesis temperatures. Mass transfer probably involves diffusion of reactants and products through liquid hydrocarbons. In the liquid-filled pores, the concentration of products probably increases rapidly with distance inward from the geometric surface to large constant values at the depth where either hydrogen or carbon monoxide is completely consumed. In this region, the H20/H2 and COn/CO ratios are sufficiently large to oxidize either metallic iron or interstitial phases.

Thus the catalyst is probably oxidized from the inside, involving first the deeper layers, which would not be effective in the synthesis in any event. The lower rate of oxidation of nitrided catalysts may result from an intrinsic resistance of the €-phase to oxidation; however, two other factors should be considered : Water production from nitrided catalysts is lower, and the presence of molecules of lower molecular weight in the pores of the catalyst may lead to a greater rate of diffusion and more effective ventilation of the pore system. If this model of the synthesis process is correct, sharp differences between behavior of the carbonitrides of different composition can hardly be expected. Various portions of the catalyst-e.g. VOL. 49, NO. 12

DECEMBER 1957

2059

X 342

TEST N U M B E R INITIAL N / F e

I

INITIAL C / Fe TEMPERATURE. "C S P A C E VELOCITY,

hr.-I

I

CONTRACTION, % ACTIVITY,

I I

A F ~

X 279

0

317

I1

632

1

057 243

0 23

z;

",;

~

1 313 646 643 1 632 1 294

IO4

I N I T I A L COMPONENT PHASES1

1

I19

1

i

6 4 1

~

,y,a I

I

I

I

ci

70 C3f c 4

60 27'- 185°C 14 OH 124

3

7.9 Br

OH

50

co

COOH

f

I

=67

I

i I

ggOHj

40

30

Br OH

77

Br

(OH

27 1.6

OH

The authors acknowledge the assistance of R. A. Friedel, A. G. Sharkey, Jr., M. E. Kundick, W. E. Dieter, Walter Oppenheimer, and W. C. Peebles in performing analyses of catalysts and synthesis products. Literature Cited

co

I

7 "

57

Acknowledgment

185O-352'C

3 52%464°C

20

67

the same activity as carbonitrides produced during synthesis, as shown in all the experiments here reported. €-Carbonitride prepared by nitriding 2-iron carbide with ammonia showed lower activity and formed products with a slightly higher molecular weight than carbonitrides produced in the reverse order. More extensive carburization of a nitride to remove the nitrogen fairly completely produced X-iron carbide that had the same catalytic properties as the X-iron carbide produced by direct carburization of a reduced catalyst. The e-carbonitride from methylamine produced the low molecular weight product characteristic of e-carbonitrides; however, a smaller amount of oxygenated compounds and more olefins were formed.

1

IO

i

9

(1) Anderson, R. B., Advances in Catalysis' Academic Press, 355 (1953). ( 2 ) Anderson, R. B., Hofer, L. J. E., Cohn, E. M., Seligman, B., J . Am. Chem. SOC. 73, 944 (1951). ( 3 ) Anderson, R. B., Krieg, A., Seligman,

185"-352"C b Br 8 OH 25

-

3520-464OC

0

>464°C

> 464'C'

Figure 10. Activity and product distribution from synthesis at 2 1.4 atm., carbonitride catalysts, 1Hz 1 CO gas

+

Values for unsaturated hydrocarbons ( =), alcoholic hydroxyl (OH), carboxyl group (COOH), and carbonyl group (CO)are in weight per cent of fraction where found. Br denotes bromine number

(4) (5) (6)

(7)

the bottom compared with the top of the bed and the pore mouths of the particles as compared with pore bottoms-will be exposed to different gas compositions, and these areas may therefore promote different reactions. As the initial nitrogen content of the catalyst was increased, the activity increased up to about N/Fe = 0.25 and then remained relatively constant. Also, at about this composition, a major change in selectivity occurred. Specifically, increased production of alcohols and products of lower molecular weight associated with nitrides and carbonitrides were observed. At N/Fe less than 0.25, the products were characteristic of reduced carburized catalysts. Phase changes during synthesis (Tables I and 11) indicate that the principal phases present in used catalysts with initial N/Fe less than 0.25 were magnetite and Hagg carbide, whereas for initial values greater than 0.25 e-carbonitride

2060

was generally the principal phase. I n previous carburization experiments (7) a t atmospheric pressure with synthesis gas or carbon monoxide, a sample with N/Fe = 0.08 was converted to Hagg carbide, another with N/Fe = 0.21 to c-carbonitride, and a third with N/Fe = 0.29 to e-carbonitride. During synthesis the carbon and oxygen contents of the partly nitrided samples increased more rapidly than for fully nitrided €-iron nitrides under the same conditions. Thus the experiments on partly nitrided catalysts show that the e-iron carbonitride does not form in synthesis unless nitriding has already established the €-iron nitride phase. Other forms of nitrides revert to a-iron, iron carbides, magnetite, and perhaps other phases which are more readily oxidized and are characteristic of the operation of reduced or carbided iron catalyst. €-Carbonitrides prepared by treating nitrides with carbon monoxide had about

INDUSTRIAL A N D ENGINEERING CHEMISTRY

(8) (9) (10) . .

B., O'Neill, W. E., IND. ENG. CHEY.39, 1548 (1947). Anderson, R. B., Seligman, B., Shultz, J. F., Kelly, R.E., Elliott, M. A., Ibid., 44, 391 (1952). Anderson, R. B., Shultz, J. F., U. S. Patent 2,629,729 (Feb. 24, 1953). Anderson, R. B., Shultz, J. F., Seligman, B., Hall, W.K., Storch, H. H., J . Am. Chem. Sac. 73, 944 (1951). Hall, W. K., Dieter, W. E., Hofer, L. J. E., Anderson, R. B., Zbid., 75. 1442 11953). Hal1,'W. K:, Tain, W.H., Anderson, R. B., Zbid., 72, 5426 (1950). Hall, W. K., Tarn, W. H., Anderson, R. B., J . Phys. Chem. 56, 688 (1952). Hofer, L. J. E., Cohn. E. M., Peebles. W.'C., J . Am. Chem. Sac: 71, 189 (1949'1. Jack, K: H., Proc. Roy. Soc. (London) A195, 34, 41, 56 (1948). Shultz, J. F., Hall, W. K., Dubs, T. A , , Anderson, R. B., J . Am. Chern. Sac. 7 8 , 282 (1956). Shultz, J. F., Hall, W. K., Seligman, B., Anderson, R. B., Zbid., 77, 213 (1955). Shultz, J. F., Seligman, B., Lecky, J., Anderson, R. B., Zbid., 74, 637 (1952). Shultz, J. F., Seligman, B., Shaw, L., Anderson, R. B., IND.EKG.CHEM. 44, 397 (1952).

RECEIVED for review May 11, 1956 ACCEFTED January 14, 1957 Division of Industrial and Engineering Chemistry, Symposium on Techniques of Preparing Catalysts in the Laboratory, 129th Meeting, ACS, Dallas, Tex., April 1956.