Preparation and Reactions of Carbonitrides on Iron

By W. Keith Hall, 1 W. E. Dieter, L. J. E. Hofer and Robert B. Anderson. Received April 14, 1952. The preparation and reactions of carbonitrides in an...
0 downloads 0 Views 727KB Size
1442

W. K. HALL,W. E. DIETER,L. J. E. HOFERAND R. B. ANDERSON

VOl. 75

[CONTRIBUTION FROM BUREAU OF MINES,SYNTHETIC FUELS RESEARCH BRANCH]

Preparation and Reactions of Carbonitrides on Iron BY W. KEITH HALL,^ W. E. DIETER,L. J. E. HOFERAND ROBERT B. ANDERSON RECEIVED APRIL14, 1952 The preparation and reactions of carbonitrides in an iron catalyst (synthetic ammonia-type) have been studied in the temperature range 250-450". Two reactions occur when nitrides are treated with carbon monoxide, or carbides with ammonia: A rapid completion reaction in which carbon or nitrogen enters the lattice until the ratio of carbon plus nitrogen atoms to iron atoms increases t o about 0.5, followed by a slower substitution reaction in which carbon replaces nitrogen or vice versa. When iron nitrides are treated with HZ CO mixtures, the rate of elimination of nitrogen from the catalyst increases with the hydrogen content of the gas, whereas the rate of incorporation of carbon exhibits a maximum a t about 1H2 1CO. Nitrides are rapidly reduced by pure hydrogen; carbonitrides are reduced more slowly, the rate varying inversely with the carbon content. Nitrogen is removed more rapidly than carbon, and the residual interstitial carbon appears as carbide.

+

The i r o n - n i t r ~ g e n , ~ , ~iron-~arbon,~.7~S~~ ,"~ and ir~n-carbon-nitrogen~~~~ lo systems are of considerable interest in metallurgy and in the catalytic syntheses of ammonia and hydrocarbons. The use of iron nitrides as catalysts and their transformation to carbonitrides during the Fischer-Tropsch synthesis have been studied in this Laboratory11-12p13

carbideJJ6)appeared a t 450' when fl had decreased below 0.18, that is, when about 64% of the nitrogen ?>tomsof FezN had been replaced by carbon. When metallic iron and iron nitrides are treated with carbon monoxide, the principal reaction is 2CO

--+

COZ

+

C(interstitia1 or elemental)

(1)

If the carbon monoxide is virtually completely conPHASES IDENTIFIED IN THE IRON-CARBON-SITROGEN sumed, for example, a t high temperatures and/or a t SYSTEM low flow rates, oxidation of the iron also occurs, reApproximate composition gardless of the phase in which it is present Symbol or Crystal Carboname Y' €

1

x

(Hagg or percarbide) Cementite a

structure

Kitrides

nitrides

Carbides

Face-centered cubic Close-packed hexagonal Orthorhombic

FetN FesN to FezN FezS

-FeLXa -FesX to FezX -FezX

.., .

..........

Orthorhombic

.....

... ..

-FezC

. ...

. . . .. . .

-FezC

.. .. . ..

Fed2

X = carbon plus nitrogen.

Two types of reaction may occur during the formation of iron carbonitrides : A completion reaction in which carbon or nitrogen enters the lattice until ?i has increased to about 0.5, the approximate upper limit of the e-phaseI4; any interstitial atoms already present are retained in the lattice. The second is a substitution reaction in which one interstitial atom replaces the other; this reaction is slower than the first and usually proceeds only after the completion reaction is finished. Phase changes may occur in both types of reaction. For example, Jack observed that Hagg iron carbide (or "iron per-

CO

+ 3/4Fe --+ '/aFesOl +

C(interstitial or elemental)

(2)

Jack6J,g studied reactions of the Fe-C-N system using an iron powder. The data presented in this paper were obtained on an iron synthetic ammonia catalyst, and include in addition to the reactions considered by Jack, the carburization of nitrides with synthesis gas and the reduction of carbonitrides with hydrogen. Experimental

Samples of 6- to 8-mesh, fused iron oxide, syntheticammonia type catalyst (Bureau of Mines number D-3001), were used for all experiments. The unreduced catalyst contained 67.4% total iron, 21.7% ierrous iron, 4.61% MgO, 0.57% KzO, 0.71% Si02 and 0.65% Cr203. The gases employed were: Electrolytic hydrogen which was further purified by passage over hot copper or through a Deoxo purifier and over Anhydrone; anhydrous ammonia, which was passed over freshly fused potassium hydroxide; cylinder carbon dioxide; and carbon monoxide prepared by the dehydration of formic acid in the presence of phosphoric acid and dried over Anhydrone. Hydrogen-carbon monoxide mixtures (Fischer-Tropsch synthesis gas) were (1) Now with hlellon Institute of Industrial Research, Pittsburgh, prepared by reforming natural gas, compressed in steel Pa. cylinders at 130 atmospheres, and freed of small quantities ( 2 ) E. Lehrer, 2. Eleklrochem., 36, 383, 460 (1930). of iron carbonyl by passage through glass beads a t about (3) 0. Eisenhut and E . Kaupp, $bid., 86, 892 (1930). 200". As used, one-component gases were more than 99.8% (4) S. Brunauer, M. E. Jefferson, P. H. Emmett and S. Hendricks, pure (mass spectroscopic analyses); the synthesis gas conTHIS JOURNAL, 68, 1778 (1931). tained less than a total of 0.5% nitrogen, methane and car( 5 ) K. H. Jack, Proc. Roy. Soc. (London), A196, 34 (1948). bon dioxide. Gas flows, controlled by micrometric needle (6) L. J. E. Hofer, E. hI. Cohn and W. C. Peebles, THISJOURNAL, valves, were measured by small capillary flowmeters. 71, 189 (1949). The experiments were made in glass systems. Two types (7) The early studies of higher iron carbides by Hlgg, Halle and of reaction tubes, both equipped with special four-way Herbst, and Pichler and Merkel have been reviewed in referenre 6. stopcocks,'6 were used. During reactions, their tempera(8) K. H. Jack, Proc. R o y . Sac. (London), A198, 56 (1948). tures were controlled to h 3 " . After treatment in the first (9) K. H. Jack, ibid., A196, 41 (1948). type tube16 and subsequent cooling to room temperature, (10) The reader is referred to an informative review of nitrides, the entire charge was transferred in a stream of carbon dicarbonitrides, and carbides by H. L. Riley, Quart. Rev. (London), S, oxide to a glass tube attached to the filling arm by a ground 160 (1949). joint, and the portion of the tube containing the sample was (11) R . B. Anderson, J. F.Shultz, B. Seligman, W. K. Hall and H. H. sealed a t both ends. I n this and other transfers, due preStorch, THISJOURNAL, 73, 3502 (1950). cautions were taken to preclude atmospheric oxidation as (12) J. F. Shultz, B. Seligman, J. Lecky and R. B. Anderson, ibid., well as heating of the sample during sealing off. The 74, 037 (1952). second type reaction tube was similar to the tirst, but it con(13) J. F. Shultz, B. Seligman, L. Shaw and R. B. Anderson, Znd. tained a larger charge (about 50 cc. as compared with 10 cc.

c+

Eng. Chenz., 44, 397 (1982). (14) Throughout this paper, the carbon and nitrogen - contents are expressed as the ratios of these atoms to iron atoms: C = atom ratio of carbon to iron, 5 = atom ratio of nitrogen to iron.

(15) R. B. Anderson, I n d . Eng. Chem., Anal. E d . , 18, 156 (1946). (16) R. B. Anderson, W. IC. Hall and L. J. E. Hofer, THISJOURNAL, 70, 2162 (1918).

PREPARATION AND REAC TIONS

March 20, 1953

OF

CARBONITRIDES ON IRON

1443

in the other reaction vessel) and was provided with a helix of glass tubing which extended out of the furnace zone. A representative portion of the catalyst was transferred into the helix by proper manipulation, and the section of the helix containing this sample was sealed off. From the known weight of catalyst plus glass removed, weight of glass, and weight changes during reaction, the weight change per gram of charge was calculated. Both types of reaction tube were weighed on an analytical balance. Composition data for reduction of raw catalysts or nitrides and for nitriding of reduced catalysts were obtained from weight changes only. For carburization of reduced and nitrided samples, ammonia treatment of carbides and hydrogenation of carbonitrides, chemical analyses for iron, carbon and nitrogen were obtained in addition to weight changes. When the sample had not been treated with synthesis gas, the glass ampoules containing the catalyst were scratched with a file and introduced through an "air lock" into a drybox filled with carbon dioxide a t a slight positive pressure. The ampoules were broken, and the samples were ground to finer than 80 mesh, quartered, and placed in small tared vials with screw caps equipped with neoprene gcjskets for weighing and transferring to apparatus for analysis. Before this dry-box procedure, samples treated with synthesis gas were freed of wax by continuous extraction with boiling toluene in an A.S.T.M. extraction apparatus for rubber products. The toluene was removed by evacuation a t 100". The method of transferring and extracting the samples was usually satisfactory as indicated by low oxygen values calculated by difference. Nitrogen was determined by a modification of the standard Kjeldahl method, and the carbon by a method similar t o the gravimetric determination of carbon in steel. Iron was determined by a standard dichromate titration method. Raw catalyst samples were essentially completely reduced with hydrogen a t a space velocity'? of 1,000 in 20 hours a t 550'. For use in the reaction tube in which only one sample was obtained per experiment, a relatively large quantity of catalyst (150 cc.) was converted t o the desired initial phase-nitride, carbonitride or carbide-in an aluminum block reactor and transferred in carbon dioxide to a special glass storage bulb"; portions of this material were transferred in carbon dioxide to the reaction tubes. With the other type of tube, the initial phases were prepared in situ. At the end of a period of treatment, the gas flow was stopped by closing the four-way stopcock on the reaction tube, the furnace was removed, and the sample was cooled rapidly in the presence of the residual gas. Evacuation at reaction temperatures was avoided to prevent the possible loss of nitrogen. The methods of preparing iron nitrides," Hagg carbide,'* and cementitel9 have been described previously.

space velocity of carbon monoxide of 2500, carbon was deposited a t a somewhat greater rate than a t the lower flow; the rate was not sufficient, however, to lead to oxid_ation_according to equation 2. In the first hour C N increased to 0.59 and then remained about constant. Although carbon deposition was more rapid, the nitrogen content decreased a t a slower rate. The principal phase was c-carbonitride, with Hagg carbide appearing Cnly after the nitrogen content had decreased below N = 0.16.

Results Carburization of €-Iron Nitride with Carbon Monoxide.-Table I presents the composition and phase changes during treatment of an €-nitride ( N = 0.46) with carbon monoxide a t 250 to 450'. At a space velocity of 80 and a t temperatures below 400', carbon eniered t h e catalyst without replacing N had increased to about 0.5. nitrogen until C At 350' and 400°, following the completion reacN remaintion, carbon replaced nitrogen, with ing about 0.5. At 450' the carbon content increased more rapidly in the first hour than possible by reaction 1,based on the flow of carbon monoxide, and the presence of magnetite confirmed that reaction 2 had occurred. Under these conditions the catalyst lost nitrogen more rapidly than it gained carbon, and Hagg carbide was the principal phase after 5 hours. At the same temperature but a t a

Carburization of €-Iron Nitride with Synthesis Gas.-With low space velocities of 1Hz +_ 4CQ gas a t 300°, carbon replaced nitrogen with C N remaining a t 0.45 (Table 11). At 350', substitution and completion reactions occurred simultaneously. At 400 and 450', carbon was deposited a t a greater rate than possible for reactions 1and 3

+

c+

(17) Space velocity is defined as volumes of gas (S P.T.) per volume of catalyst space per hour. (18) W. K Hall, W. E. Tarn and R. B. Anderson, J. Phys. Chem., 6 6 ,

688 (1952). (19) E.M. Cohn and L. J. E. Hofer, THISJOURNAL, 72,4662 (1950); J. Chrm. P h y s . , 18, 766 (1950).

+

a

TABLE I CARBURIZATION OF E-IRON NITRIDEWITH CARBONMONOXIDE

Temperature, OC.

Time, hours

'Td

f2 7

Phases0

A. Space velocity = 80 hr.-' Original nitride, A 0 0.00 0.46 0.46 e 1 .02 .47 .49 E 5 .02 .46 .48 E 250 10 .03 .45 .48 E 1 .02 .46 .48 E 300 5 .04 .45 .49 E E 10 .04 .47 .51 1 .04 .45 .49 E 350 5 .06 .44 .50 E 10 .08 .43 .51 E 1 .03 .46 .49 E .50 € 400 5 .14 .36 10 .30 .18 .48 1 .18 .20 .38 E, M 5 .41 .02 450 .43 x, M, a(?) 10 .46 .01 .47 x, a(?),M(?)

1 1' 1 t

r

r

B. Space velocity 2500 hr.-I 1 0 . 3 2 0.27 0.59

i

E

5 .42 .16 .58 E 10 .53 .08 . 6 1 x 5 E = e-nitride or carbonitride, p = r-carbonitride, a = a-iron, M = magnetite and x = Hagg carbide. Phases in order of decreasing intensities of X-ray diffraction patterns. 450

€9

+

HI

+ CO +HzO +

C(interatitialorelementa1)

(3)

and the diffraction pattern of magnetite was observed in addition to Hagg carbide and a-iron. At 350" and a space velocity of 1000, both carbon deposition and nitrogen elimination were more rapid than a t the lower-flow, and magnetite was not found. Although C increased to ve-g high values, Hagg carbide did not appear until N was less than 0.18. Similar results were obtained with 0.7H2 1CO gas (Table 111), except that some magnetite was formed upon carburization a t 350' and low space velocity. At a space velocity of 1000, no oxidation occurred a t 350', but the carbon content increased greatly = 1.52 in 11 kours). Again, Hagg carbide appeared only when N was less than 0.18. With 2Hz 1CO gas a t 350°, carbon was

+

(e +

W. K. HALL,W. E. DIETER,L. J. E. HOFERAND R. B. ANDERSON

14.4‘1

deposited a t a greater initial rate than possible for a combination of equations 1 and 3; nitrogen was

removzd faster than i t was replaced by carbon, so that C N decreased from 0.45 to 0.37. At a given temperature the rate of removal of nitrogen increased with the hydrogen content of the synthesis gas; the rate of deposition of carbon showed a iliaximum a t about lHz 1CO.

+

+

TABLE I1 CARBCRIZATIOX OF E-IROSKITRIDEWITH lHr GAS Time, .-

Temperature, OC.

C

hours

+ 4C0

+ AT

.V

E

1i

E

e e

I\

400

450

250

E E

300

E

,52 E .13 e , h1, x .53 e , hl .70 E , x

350

.38 e , M, .45 x, bl, . 5 i x, bl,

x

400

i1

.45 .23 .50

.004(?) .14 ,[I7

E

B. Space velocity = 1000 hr.-l 0.46 ( 1 0.18 0 . 2 8 1.0i 350 3 0.80 .18 10 1.47 .14 1.61

1 I

1.60

.1:I

1.79

a

e,

x

Sec footnote of Table I .

TABLE 111 CARBURIZATION OF €-IRONSITRIDE WITH 0.7Hz 2H? I C 0 GASES

+

Timc, ~iours

‘Temperature, OC.

+

-

c

i’

C

+ IC0

+ fi

AND

~tiasesa

-

0.7H2

+ ZCO gas a t space velocity 1

350

(11 C.

2N? 33J

t2

1

{

5 10

See footnote of Table 1.

LV

-

C

+

Phases”

4

.19

10 1 5 10 1

.25 .11 .27 .29 .21 .33 .36 .lo .32 .53

i

(Y

x,a

0.08 .. 01, 7‘ .06 0.13 a , -I’ .06 .25 a , *I‘ .04 .29 a, y’ .07 .18 a , y‘ .05 .32 a , y’ .08 .37 y’, a .07 .31 a , y’, h l .07 .40 e , x, h l , a .08 .44 x, a .05 .15 a , y’, h i .04 .3B x, 11 .02 .55 X

0 0 0.21 0..21 1 0.17 .21 .38 ,5 .26 .21 .47

y’, a, e ( ? ? ) y‘,

E,

h1

x

E,

a 01 = a-iron, M = magnetite, y’ = 7’-nitride or carbonitride, e = €-carbide or carbonitride, and x = Hagg carbide. Phases in order of decreasing intensities of X-ray diffraction patterns.

CARBURIZATION OF y’

+ E-SITRIDES iv1.m CARBONMor0.7& + 1CO GAS

OXIDE ASD

‘Time, hours

Temperature, ‘c.

C

N

C

+ ,V

P1iawi;u

A. Carbon ~nonosidca t space velocity = 100 hr.-’ 0 . 0 0.29 0.29 y’,E Original nitride, G 0 .ll .29 .40 E, y ’ 3.50

400

1i’i

!

1;) 16 24

1;

48

21.3 48

B.

E

E

x, E

+ 1CO gas at space velocity = 100 hr.-’ (

..

1 0.07

5 10

Original nitride, F 350

= 1000 hr.-I

0 . 2 2 0 . 2 6 0.48 ,123 1.05 1.3’7 .08 1.60

4 5 . 7 0.8;

C

WITH C A R B O N

TABLE V

A . 0.7H2 1CO gas a t space velocity = 100 hr.-l Original nitride, C n 0.02 0.43 0 . 4 5 t 1 .0:3 .40 .43 E i .04 .38 .42 E, x ( ? ? J 235 1 2 .37 .42 E , x(??) ;, 10 .05 .39 ’ 1 .04 . 4 3 E , x!??) .43 E .12 .3I I O 300 E ’ 7 . 8 , 1 4 .30 .41 .44 e .IF .28 \in 1 .06 .39 .44 E , M 350 i 5 .2i .23 .50 e , a , hl l~ ID .15 .19 .6-1 E li,

0

1

01

e

x

I 1‘ I (

e E,

hours

O C .

Original nitride, E

.26 .31 .56

MONOXIDE Time, -

(Space velocity = 80 hr.-l) Reduced catalyst, D 0 0 0 0 2 0.11 0 0.11 6 .24 0 .24 250 12 .28 0 .28

.50

[ 10

(24.6

TABLE IV CARBURIZATION O F CY-IRON AND ? ‘ - N I T R I D E S

.3i

.08

IIi

+

.13

.26 .35 .22 .11

1;

‘ I i 3

+ €-Nitrides + lC0.-

Tables IV and V present data for the carburization of a-iron and y’- and 7’- €-nitrides. At all temperatures studied, only the completion reaction

e

\

:!so

Carburization of a-Iron,y’-, and y’with Carbon Monoxide and 0.7Hz

Temperature,

Phasesa

A . Space velocity = 80 1 x - I Original nitride, €3 0 0.02 0.43 0.45 ’ 1 .03 .4’ .44 2950 5 .04 .40 ,44 10 .04 .42 .46 .Os . 4 1 .46 i ! 3011 2 a , 1CJ .35 .45 12.2 .15 .31 :IG i 1 .05 .41 .46

Vol. 75

1;

a

.25 .25 .26

,109 .26 .2i .27

.29

.2A

.14

.27 .26 .24 .21 .20 .I8

.27 .30 ..‘I5 .%3 .39

.50 .51

E E

.52 .53 .55 .41

t E E

E,

y’,nl

.5a .51 . -56 .55

E E E E

.57

0.7H2 f I C 0 gas at space velocity = 100 hr.-I .05 .27 .32 .17 .?? .39 E, y’ 3110

0.37 0 . 4 3 E, X l .I8 .IO .37 t, XI ,19 ,IS . 3 i e , Ii

0.06

.21

I

24

::i

.28 .?!I .43

18 .74 See footnote of Table IV.

.20 ,I7 . 16 .I3

.IX ,40

c , y’

e,

y’

.59 E , y’ .89 e , y’, x

PREPARATION AND REACTIONS OF CARBONITRIDES ON IRON

March 20, 1953

1445

c+

was observed-until fl equaled or exceeded 0.5. TABLE VI In series E ( N = O.OS), a-iron was the most promi- AMMONIATREATMENT OF CEMENTITEAND HBGG CARBIDE nent phase indicated by X-ray diffraction until Time, -C - -C + % Phasesn Temperature, OC. hours N exceeded 0.27, and the $-phase was the only inter(Space velocity = 1000 hr,-1) stitial phase observed a t € less than 0.32. Hagg carbide was the major phase produced upon ex- Cementite, H 0.35 0 0.35 C tesded carburization a t 350 and 400'. In series F .30 0.07 .37 C, E ( N = 0.21), the y'-phase was the principal cornpo300 .09 6.25 .31 .40 C, E nent after-one hour of carburization although C = .30 .10 .40 C, e 0.17 and C N = 0.38. After 5 hours, the X-ray 350 15 .27 .20 .47 E patterns of e-carbonitride and Hagg carbide were Mixture of a-iron and Hagg found. The carburization with carbon monoxide carbide, I 0.29 0 0.29 a, x of series G (Table V) containing the y'- and E345 4 .29 0.18 .47 e, x phases ( N = 0.29) produced e-carbonitrides. 350 12 .26 .24 .50 E, x 1CO gas a t Treatment of this sample with 0.7Hz 28 .19 .33 .52 e, x(??) 345 300' yielded a mixture of E- and y'-carbonitrides 350 52 .14 .38 .52 E with the former predominant, and nitrogen was 345 117 .07 .43 .50 e removed more rapidly than with carbon monoxide a t 400'. Mixture of a-iron and Hagg In Fig. 1 the rate of carburization with carbon carbide, J 0.48 0 0.48 x, a monoxide a t 350' is related to the initial nitrogen 8 .46 .06 .52 x, E content of the catalyst. The initial rate of carburi24.2 .44 .12 .56 x, E zation varied inversely with the initial nitrogen 48.2 .37 .21 .58 x, E 350 content. This result was also observed in carburi72.2 .33 .25 .58 x, c zations with carbon monoxide a t 250' and with 120.2 .26 .34 .60 x, E . 1CO gas a t 300'. Carburization with 0.7Hz 1158.2 .18 .42 .60 e carbon monoxide a t 350' of samples containing the a = a-iron, E = e-nitride or carbonitride, x = Hagg cara- and 7'-phases usually yielded the X- and e- bide, and C = cementite. Phases in order of decreasing intensities of X-ray diffraction patterns. phases together with magnetite and iron. Ammonia Treatment of Cementite and Hagg I I Carbide.-When Hagg carbide and cementite were

c

1:

+

+

I

+

(I

.5

-. d

!A

z

+ .4

0 0 l-

a a .3 I

2

4 - 0 - .21

.5 r

0

I

5 T I M E , HOURS.

I

I

IO

Fig. 1.-Variation of the rate of carburization of iron nitrides with carbon monoxide at 350' and space velocities of 80 and 100 hr.-1 as a function of initial nitrogen content.

Fig. 2.-Variation of the rate of nitriding of a-iron and Hagg carbide a t 350' and a space velocity of ammonia of 1000 hr.-l as a function of initial carbon content.

ifr. K. HALL,W. E. DIETER,L. J. E. HOFERAND I