Curhon. 1972, Vol 13. pp. 172 Pergamon Presc.

Printed m Great Bntain

FORMATION OF CARBONACEOUS DEPOSITS

FROM THE REACTION OF METHANE OVER
NICKEL

R. T. K. BAKER, P. S. HARRIS, J. HENDERSONand R. B. THOMAS

Applied Chemistry Division, Atomic Energy Research Establishment, Harwell, Didcot, Oxfordshire, England

(Received 20 May 1974)

Abstract-Controlled atmosphere microscopy techniques have been used to study the formation of several forms of
carbon produced when methane was passed over heated nickel surfaces. The predominant forms of carbon were
found to be nodular clusters of polycrystalline material and graphite flakes. The latter type of deposit appeared to be
favoured under conditions of rapid cooling of specimens which had been held at 1000°Cfor a significant period.
Experiments were designed to demonstrate that a solution mechanism could account for a large fraction of the flake
deposit. The absence of filamentous carbon growth is accounted for in terms of a previously postulated mechanism.

1. INTRODUCTlON produced from commercial grade methane though not

Several forms of carbon deposits when
are produced with high purity methane at temperatures below 900°C.
carbon-containing gases are decomposed over heated All these studies were limited to post-reaction examina-
metal surfaces. Some of these deposits may be unique to a tion of the carbon products. Since the structures of these
particular gas-solid system whilst others consistently products are so complex the interpretation of such results
appear in widely differing systems. The role of the metal is difficult in the absence of a knowledge of the initial
surfaces on the decomposition reactions has, over the stages of deposition. In the present work we have used
past few years, attracted increasing interest and the controlled atmosphere microscopy techniques to follow
reaction of methane over nickel is among the most the reactions continuously. The use of post-reaction
intriguing[l-101 since this is one of the few endothermic scanning electron microscopy has enabled the observed
reactions where carbon is deposited. structures to be related to those reported by other
Two types of carbon produced from the pyrolysis of workers, and a definitive experiment has been designed to
600Torr methane over nickel and other transition metals test the role of solution processes in determining the
were identified by Robertson[l, 23, these he termed structure of the deposit.
“flake” and “polycrystalline”. Electron diffraction analy-
sis showed the first to be highly ordered graphite and the 2. EXPERIMENTAL
polycrystalline form was stated to have an essentially 2.1 Controlled atmosphere optical microscopy [CAOM]
non-graphitic structure, although it occasionally appeared The CAOM studies were carried out using a Leitz 1750
in a fibrous form which, from diffraction evidence, was heating stage fitted to a Vickers M41 Photoplan
shown to have the graphite basal planes parallel to the microscope[l2]. Strips of nickel foil, about 5 mm wide,
fibre axis. Fryer and co-workers[3,4] have examined were clamped between steel electrical conductors and the
similar fibres at high resolution and have resolved the foil heated by passing a current through it. The deposition
3.4 .& interlayer lattice period confirming the preferred temperature was measured with an optical pyrometer
orientation. They have shown that at the fibre tip the basal (Pyrowerke GmbH, optical micropyrometer) which had
plane layers are parallel to the surface of the metal been calibrated “in situ” against known melting points
particle which is usually found as an inclusion in the fibre. (Au, Ag, Cu and NaCI).
They also found that all forms of carbon produced from 1 The nickel foils were cleaned in distilled water and
to 600Torr methane over nickel at 7Oo”C,including the acetone by ultrasonic agitation. After the specimen had
flake variety, contained metal in a highly dispersed form. been clamped into position the cell was flushed with
Evans ef a[.[5], investigating the same system, reported hydrogen several times before the nickel was exposed to
that a further filamentous form, similar to that produced in 600 Torr hydrogen at 750°C for 2 hr. Finally the specimen
the decomposition of acetylene over nickel [ 111, could be was reacted with 600Torr methane at between 750 to

17
18 R. T. K. BAKER et a[.

1060°Cfor periods of up to 12hr. Micrographs were taken possible to discern the grain boundaries. When the
of the metal surface at regular intervals throughout the specimen was exposed to 600Torr methane and the
heating and cooling periods. temperature raised to 850°C deposition of carbonaceous
After the deposition reaction has been terminated some material was apparent after about 5 min and collected at
specimens were examined in a Cambridge Stereoscan grain boundaries in the form of clusters. Over a period of
scanning microscope and others by transmission electron 5 hr the deposit coverage and size of clusters increased. In
microscopy. The transmission specimens of the carbona- Fig. 2 is shown a scanning electron micrograph of the
ceous deposit were obtained by allowing the specimen to deposit.
float in approximately 5M HCI for a period of 15hr, when
the carbonaceous film was found to separate quite readily
from the nickel surface.

2.2 Nickel tube experiments

This was a definitive series of experiments designed to
allow the examination of graphite produced from solution
in the nickel on a surface to which the hydrocarbon did
not have access. The experimental arrangement is shown
in Fig. 1. The tube was made from spectrographically pure
nickel sheet, 0.25 mm thick, bent and welded to shape.

-
5,um

Fig. I. Experimental arrangement for nickeltube experiments.

Fig. 2. Scanning micrograph of a specimen showing the detailed
Stainless steel end plates were welded to the tube so that appearance of the deposit.
gas could be led to the tube through stainless steel pipes.
This whole assembly was placed in a silica outer jacket. In experiments where the upper temperature was raised
The tube was cleaned with water and acetone before use to 900°C the deposit initially took on the same form as at
and then heated in flowing hydrogen at 750°C at lower temperatures but after 2 hr or so long fibrous
atmospheric pressure for 2 hr. The apparatus was cooled growths (2pm width and up to 50pm length) began to
and then methane allowed to flow through the centre tube form at grain boundaries. On continued reaction these
whilst argon was passed through the jacket and the growths expanded in width and increased in length.
temperature raised to 1000°C and maintained there for More dramatic effects were produced by temperature
periods of up to 8 hr. When cool the tube was split open cycling. The temperature was maintained alternately at
and the deposits on both the inner and outer surfaces were 1000°C and 500°C to 200°C for periods of 5 min. No
examined. deposition was usually observed after the first cycle.
Methane (Matheson Research Grade, 9999% pure) was After subsequent cycles small nuclei of deposit appeared
used in all these experiments. The argon and hydrogen on cooling, although at the temperatures used these did
were supplied by British Oxygen Ltd. with a stated purity not appear to increase appreciably in size with increasing
of better than 99.9% and were used without further time. On reheating, the nuclei grew very rapidly and the
purification. The nickel, 0.25 mm thick sheet or wire, was whole surface of the metal was covered, though as the
Johnson-Matthey spectrographically pure grade (analy- temperature increased to 1000°Cthis deposit disappeared
sis: Fe, 10ppm; Si, 3 ppm; Ag, 2 ppm; Al, Ca, Cu, Mg, to leave a bright surface again. This sequence of events
< 1ppm, others not stated). was too fast to record continuously with the facilities
available. The nuclei of deposit which appeared at the
3. RESULTS lower temperature are shown in Fig. 3. Figure 4 is a
3.1 CAOM Experiments stereoscan micrograph of a specimen after an extended
After treatment with hydrogen at 750°C the surface of run in methane at 1000°C and cooled before the
the nickel appeared to be relatively smooth and it was appearance of clusters. Depressions in the crystalline
 Formation of carbonaceous deposits from the reaction of methane over nickel 19

surface layer of the type shown were quite common.

Figures 5A and 5B are transmissiori micrographs of the
stripped deposit from a similar specimen together with its
electron diffraction pattern. The deposit is evidently

Fig. 3. Optical micrograph of a specimen after temperature cycling Fig. 4. Scanning micrograph showing depressions in the laminar
from 1000°C to 200°C in 600 Torr methane, showing the formation deposit produced from cooling a specimen which had been
of deposit nuclei at the lower temperature. exposed to 600 Torr methane at 1000°C for 35 min.

Fig. 5. (a). Transmission electron micrograph of the stripped laminar deposit. (b). Selected area diffraction pattern
indicating the graphitic nature of this material. (c). Diffraction pattern with 12 spots in the innermost ring.
20 R. T. K. BAKER
et al.

graphite of high crystallographic perfection. There is

evidence of some rotational mismatch as shown by Moire
fringes in the micrograph and satellite spots in the
diffraction pattern. Diffraction patterns where there were
12 spots in the innermost ring in a symmetrical
arrangement, Fig. X, were frequently encountered.

3.2 Tube experiments

In the first experiment methane was passed through the
inside of the tube and argon around the outside. After one
hour at 1WC the tube was cooled to room temperature
and split open. The inner surfaces were black and had a
sooty accumulation of carbon. The outer surfaces were
matt grey and covered with a thin adherent film. Fig. 7. Transmission electron micrograph showing the overgrowth
This film was stripped from the outside of the tube and of well aligned crystallites on the laminar graphite deposit (Nickel
examined by transmission electron microscopy. Electron Tube Experiments).
diffraction analysis showed it to be graphite. The
morphology of the deposit reflected the nickel grain
structure; deposit was heaviest on grain boundaries and,
although the average thickness was the same over an
individual grain, it did vary from grain to grain. On some
grains the deposit was thick and much overgrowth had
occurred, Fig. 6, on other grains rows of well aligned
crystallites overlaid a thin substrate of good quality
graphite, Fig. 7. These two structures and various forms
distinguishable as intermediates between the two accoun-
ted for the whole deposit structure. The small crystallites
also appeared to possess borders of different contrast to
the rest of the crystallite, Fig. 7. The appearance of these
crystallites is shown in the scanning electron micrograph,
Fig. 8.
In a second experiment with a new tube, argon was first
passed both through and around the tube to verify that

Fig. 8. Scanning micrographof the crystallites shown in Fig. 7.

graphite was not produced in the absence of methane. In

the third experiment methane was re-introduced into this
tube and graphite was again obtained on the outer
surfaces after the tube had been cooled to room
temperature.

4.DISCUSSION
The formation of crystalline or “flake” graphite on iron
Fig. 6. Transmission electron micrograph of laminar deposit which
and nickel surfaces from the decomposition of hydrocar-
had been stripped from the outer wall of a nickel tube after bons has been widely reported[l-lo]. Suggested growth
passing 600Torrmethanethroughthe tube at lOUU”C. mechanisms have involved
 Formation of carbonaceous deposits fromthe reactionof methane over nickel 21

a. direct formation of epitaxial layers from decomposi- appe~ing at the surface is fairly uniform and the sites at
tion of an adsorbed precursor, which they emerge are determined by the crystal structure
b. formation and decomposition of metastable carbide of the metal.
intermediates, and, At temperatures below 4O@‘Cthe thermal expansion
c. the mobility of carbon atoms generated by the coefficients for nickel and graphite have opposite
decomposition of the hydrocarbon on the metal surface. signsflfi]; graphite expands in the basal plane directions
These explanations can only be tentative because they with decreasing temperature. The resultant stresses are
lack firm evidence for their validity. This work shows that a mostly removed by formation of twin ridges as observed
further mechanism must also be considered: in Figs. 4 and SA. The formation of depressions in the
d. the precipitation of carbon from solution in the metal graphite film, Fig. 4, might also arise for the same reasons.
during the cooling stage of the experiment 1131. Although this work and that of Derbyshire et al.[14]
For most deposition studies the cooling down of the establishes the general characteristics of this underlying
specimen from the reaction temperature to room tempera- layer it certainly cannot be stated that all crystalline
ture prior to examination is an essential part of the graphite is formed in the same manner and it is not yet
experimental technique. Derbyshire et at. [14] showed that possible to distinguish graphite formed by a precipitation
such precipitated films could consist of extremely good mode from any formed by other high temperature
quality graphite and suggested that these processes could processes.
be operative in pyrolytic systems. The failure to find carbon ~laments of the type
We have found that graphite flakes grew from the produced from nickel catalysed decomposition of
metal”deposit interface. and that crystaliites of graphite acetylene[ll] in the present series of experiments is not
were produced on nickel surfaces which had not been in unexpected. In the suggested mechanism for carbon
contact with methane. Both these observations are filament production[l 11 one of the postulated require-
consistent with a mechanism involving the formation of ments is that the decomposition of the hydrocarbon
graphite by precipitation of dissolved carbon in the nickel. imparts heat to the exposed face of the metal particle.
The possibility of carbide intermediates being involved is Some of the carbon produced from this reaction is taken
discounted on the grounds that the temperature ranges in into solution by the metal and filament growth proceeds
which carbides are known to be persistent, < 250°C and by a diffusion process of carbon from the hotter leading
> 12OO”C, were not entered for a su~~iently long period face of the particle to be precipitated at the cooler regions.
to enable a carbide phase to establish itself [lS]. In the case of methane, the decomposition reaction is
It is now necessary to discuss what part of the overall endothermic (A~,~ = +21~4kcal.mole~‘)~l7] and to a
deposit structure we can expect to be due to reorganisat~- greater extent than the solution of carbon in nickel is
ons following the cooling down of the specimen. The exothermic, and so the postulated conditions for filament
fraction of the carbon deposited on the nickel surface formation are never attained. This seems to suggest that
which enters into solution is determined by both the the filaments in the deposits from commercial grade
solubility and the rate of diffusion of carbon in the metal. methane[5] are indeed due to the presence of an impurity.
On cooling the amount of carbon which is precipitated It should also be borne in mind, however, that oxygen
will depend on the degree of supersaturation achieved and could in principIe achieve the same effect as a depositing
the rate of diffusion back to the surface. These factors impurity through the release of heat from the exothermic
cannot be quantified on the basis of present knowledge oxidation of methane at the metal surface.
but the optical microscopy results indicate that precipita-
tion is a rapid process even compared to the rate of
cooling of small laboratory specimens and the structure of
the carbon deposit could be very much dependent on the 1. Robertson S. D., Carbon 8, 365 (1970).
detail of the way an experiment is performed. In general 2. Robertson S. D., Carbon 10, 221 (1972).
the amount of graphite formed will increase with (a) the 3. Baird T., Fryer J. R. and Grant B., Nature 233, 329 (1971).
duration of the ex~riment, probably to some limit, (b) 4. Baird T., Fryer J. R. and Grant B.. Int. Carbon Conf.,
Baden-Baden, June 1972,p. 266.
increasing deposition temperature, fc) increased cooling 5. Evans E. L., ThrowerP. A., Thomas.I. M. and Walker P. t.
time, and (d) may be influenced by the thickness of the Jr., Int. Carbon Conf., Baden-Badeo. June 1972,D. 284.
foil-either directly. or indirectly through the increased 6. Presland A. E. B. and Walker P. L. Jr., Carbon 7, I’( 1959).
heat capacity. This crystalline deposit will underlie 7. Presland A. E. B., Roscoe C. and Walker P. L. Jr., 3rd Conf.
fnd. Carbon and Graphite, London (1970).
whatever carbon remains on the surface. it is not 8. Presland A. E. B. and Biau 6.. 3rd Conf. Ind. Carbon and
unreasonable to suppose that the graphite layer will be Graphite, London (1970).
formed epitaxially particularly if the flux of carbon atoms 9. Karu A. E. and Beer M. J., J: Apgl. Phys. 37, 2179 (19%).
22 R. T. K. BAKERet al.

10. Rostrup-Nielsen J. R., J. Cnfal. 27, 343 (1972). 15. Hansen M., Constitution ofBinary Alloys. McGraw Hill 1958)
ii. Baker R. T. K., Barber M. A., Feates F. S., Harris P. S. and p. 374; R. P. Elliot, Constitution of Binary Alloys, McGraw
Waite R. J., f. Catal. 26, 51 (1972). Hill I%5 p. 223.
12. Baker R. T. K. and Thomas R. B., J. Cry~fu~G~o~f~ 12, 185 16. Reynolds W. N., physical Properties of graphite. Elsevier
(19772). Press 1968p. 80.
13. Austerman S. B., Chemistry and Physics of Corbon Vol. 4, 17. Stuff D. R., Westrum E. F. Jr. and Sinke G. C., The Chemicaf
Marcel Dekker, New York (1968). Thermodynamics of Organic Compounds. J. Wiley and Sons
14. Derbyshire F. J., Presland A. E. B. and Trimm D. L., Carbon 1969p. 243.
10, 114(1972).