Carbon, 1976.Vol. 14, pp. 287-288. Par8amon Press. Printed in Great Britain zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFED LETTERS TO THE EDITOR Evidence that carbon formation from acetylene on nickel involves bulk diiusion zyxwvutsrqponmlkjihgfedcbaZ (Received 14 Ianuary 1976) The process of carbon formation from hydrocarbons on metals has been widely studied. The catalytic effect of nickel, cobalt and iron on the decomposition is well established (temperatures below 600°C). BEFORE REACTiON CSXMTH WRlNG REACTION Alternative mechanisms have been proposed to describe the process [ I, 21. Extensive kinetic observations using vacuum microbalance techniques have given support to the mecha~sm which involves bulk delusion of carbon in nickel[2-4]. Decom- position of the gas on the metal surface produces carbon atoms which diffuse through the metal to regions where growth takes place. At temperatures below 500-550°Cthe bulk diffusion step seems to be rate controlling. Studies using controlled atmosphere electron microscopy lead to the same conclusion [5,6]. RUN 1 Clean zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPON crystal fOll t al both sides n Both swfaces RUN 2 cwered ! No depose ticn byc=m 1 We performed a very simple sequence of experiments which prove that bulk diffusion is involved in the process. Single crystal foils (4 x 6 x 0. I mm) were used in the experi- ments. The crystals were preground with 400and 800grit paper in a METASERV rotary pregrinder, and eiectropoiished (20% perchioric acid, 80% ethanol, 100 mA, -40°C) and the faces were checked by X-ray diffraction for (I 11) orientation. II One surface RUN3 A B pre-coated by carbon I [I One slatace 1 II precosied RUN 4 A B + ameallng . On Seth sides for 5 ml”. at 900 “c Run 1. A single crystal nickel foil was hung in a vacuum microbalance at a tempera~re of 475°C. 100 ton (1 torr= 133.3 N me’) of acetylene were admitted and carbon was quickly formed on the two faces of the foil, as expected. Fig. 1. Catalytic growth of carbon from C,H, on nickel crystal foils under various pre-coating conditions (IO0 torr f&H,, 475°C). Run 2. A similar foil was covered on both sides by a 4OOA carbon layer before reaction. This pre-coating was performed in a Nanotech vacuum coating unit. No catalytic deposition could be subsequently observed under the same conditions as in Run 1.The foil was poisoned by the carbon pre-coating. nickel, presumably facilitating the nucleation process [8]. Appar- ently, pre-seeding is not critical here. Run 3. A foil was pre-coated as in Run 2, but on one face only. Under the same conditions as the previous runs, grow& ~oo~~iace on the coated side. The uncoated face remained ciean. Although more than 300 ia of carbon were deposited, equivalent to a layer of about 60,006 , no carbon is observed on face B, except near the edges (nucleation seems to be progressively spreading round the edges and over face B). In this experiment side A was poisoned, so the decomposition of C,H, must have occurred on side B. However, growth takes place on the opposite side A. It is obvious that carbon atoms migrate through the nickei crystal from face 3 to face A during the reaction. The energy of activation for the process of carbon formation from various hydrocarbons on nickel in the range of temperatures 40%5oo”C has been observed to lie in the range 29-34 kcai/mole (1 cai = 4.184J)[3]. In the case of acetylene the more reliable value was nrobabiy 31 kc~/moier41. The coincidence of these values _ _ with the activation energy observed for diffusion of carbon through nickel (E = 33-35k&/mole) was advanced as evidence that the process of bulk diffusion was rate controlling in that region of temperatures[2-51. The order zero observed for the reaction[4], and the independence of the kinetics observed with regard to the parent hydrocarbon used 131, support this view. Run 4. The foil was pre-coated on one side as in Run 3, followed by annealing of the foil at ca. 900°C for 5 min. Under the reaction conditions referred to above, growth of carbon was observed on both sides of the foil. However, the existence of a concentration gradient across the metal required the defi~tion of the limits of wncen~ation on the decomposition side and on the growth side. The change of solubility with temperature should therefore be taken into account (AH = 10 kcal/moie[9]). This would give an apparent activation energy of 33+ 10 = 43 kcal/mole [4]. The results of the above four runs are summa~s~ in Fig. 1. Run 5. A ~lyc~s~liine foil was polished and pre-coated on one face only as in Run 3. When reacting in the microbalance with 100torr of acetylene at 47s”C, carbon was again formed on both faces of the foil. It seems obvious that in Run 3 the pre-coating acts as seeding for carbon growth. The presence of carbon on face A prevents formation of nuclei on face B, probably keeping carbon supersat~ation in nickel sufficiently low from the be~nn~ of the reaction. In Run 4, dissolution and reprecipitation of carbon during annealing was enough to create nuclei on the opposite side. This process has been investigated[7]. Growth was observed on both faces in this case, although the decomposition must have occurred on the clean areas in face II, since face A was completely poisoned. In fact this contradiction seems to derive from incorrect data. The values of E = 33-35kcaltilmole mentioned by various authors[lO] are based on experiments performed at much higher ~mperatmes (8~12~C). A recent study in the tempera~re range 350-7OO’C aives a much lower value of E = 20 22 kcal/moie Ill]. The two values are probably related to two different modes of carbon diffusion: an interstitial mechanism at low temperatures (E = 20 kcal/mole) and a vacancy mechanism at higher temperatures (E = 33 kcal/mole)[ll]. The mechanism with the lower activation energy obviously predominates at low temperatures. A similar shift of mech~ism is known to occur in the C-aFe system[ll]. It is interesting to note here that surface diffusion was not observed in the C-N1 system. The calculation of the apparent activation energy for the process must be based on the appropriate value of 20t 2 kcal/mole and gives 30it 2 kcahmole, which is now in good agreement with the observed values. In Run 5, the presence of grain boundaries seems to make Ack~owfedgemen~s-The authors are indebted to Dr. D. L. nucleation much easier. In fact, grain boundaries were shown to Trimm, Imperial College, for providing the facilities for the have a marked influence on the activity for carbon formation on preparation of the samples, and for bringing Ref. [l l] to their 287