Carbon fibre production during hydrogen plasma etching of diamond films

Abstract

Diamond thin films have showed outstanding performance when exposed to extreme conditions such as high power plasmas. However there are always concerns about the stability of the diamond structure in the presence of other materials deposited on the film surface by plasma diffusion. It is known that diamond films are etched by hydrogen plasma but in the presence of Si, carbon fibres are formed. In this report we show for the first time the effect of Si on the production of fibres under etching conditions and propose growth mechanisms based on the results of characterisation techniques. Carbon fibres have been synthesised on diamond films and characterised by scanning electron microscopy, X-ray photoelectron spectroscopy and Raman spectroscopy. In situ optical emission spectroscopy was performed during the experiments showing different concentration of growing species which may result in the observed variability of fibres growth rate and morphology. Furthermore, fibres varied in size and shape depending on the structure of the diamond films. The surfaces of the fibres contain silicon and are oxidised having COO and CO groups as seen by XPS analysis. Raman analyses revealed a spectrum typical for graphite combined with that from diamond that remains on the surface after hydrogen bombardment. The results of this study show the experimental conditions in which carbon fibres are produced under high hydrogen etching of diamond films and opens the possibility to other applications such as catalysis, sensors and the production of electrodes, since they combine the unmatchable properties of a diamond supporting substrate with the unique properties of carbon fibres.

---

1. Introduction

In applications such as plasma facing materials there are potential concerns about the stability of diamond films and questions are raised as to whether diamond films will be eroded, peel off or transformed to other structures.^1–3 It is known that hydrogen bombardment etches carbon films and the eroded material can contaminate the plasma and the chamber with hydrocarbons and dust. The production of other structures like fibres may worsen the problems of contamination with larger particles or bring other issues such as gases retention. Information is lacking as to whether the materials or metals arising from the chamber components could affect the structural integrity of diamond films.

We know that fibres as well as nanotubes and other nanostructures⁴ are synthesised in catalytic chemical vapour deposition (c-CVD) using catalyst particles such as Fe, Ni, Co, Cu or their alloys. Metal thin films have also been used as catalysts: the technique allows the formation of metal nanoparticles after thermal or plasma etching of the film.^5,6 In particular, the metal catalyst plays an important role during the growth of carbon fibres. It determines the fibre structure, since the growth is parallel to the crystalline planes of the catalyst. The metal particle can be solid or in a molten-state during the growth. The catalyst size affects the diameter of the fibre⁷ although Louis et al.⁸ reported the growing of carbon fibres with diameters of 30 nm using nickel particles of dimensions greater than 100 nm. The carbon source is commonly methane, benzene or acetylene and is mixed with hydrogen or carbon monoxide in a range of temperatures between 500 and 1200 °C (ref. 6 and 9–11). Hydrogen has an important role during the carbon fibres growth. It has been observed that it can accelerate the synthesis by decomposing the inactive metal carbides and graphitic deposits in the catalyst surface.¹² Furthermore, it can suppress the synthesis by competing with the carbon species during the adsorption on the metal surface and reacting with carbon to form methane. The temperature affects the growth rate and the morphology of the fibres. At low temperature both straight and tight helical fibres are evident. When temperature is raised the nanofibres are mostly twisted. Amorphous filaments would be obtained at low temperature, whereas isometric shells would be obtained at high temperature.¹² Carbon fibres have been produced for applications such as: energy storage,¹⁰ catalyst support materials,^13–15 composite materials,¹⁶ selective adsorption agents,^17,18 electrodes^19–21 and recently, the production of carbon nanofibre paper.²²

Although most of the reported experiments for the production of carbon fibres involved the use of catalyst, Kobashi et al.²³ reported the production of fibrous structures on diamond films under hydrogen etching conditions using only a DC bias.

The aim of this work is to study the formation of fibres on diamond film substrates during hydrogen etching conditions in the presence of silicon. Diamond films used as substrates have different characteristics such as morphology and diamond quality (sp³/sp² ratio) and the resulting fibres morphology are related to the substrate and the concentration of carbon species diffusing during growth.

2. Experimental

Carbon fibres were synthesised on nanocrystalline and polycrystalline diamond films. The materials exhibited different morphology and sp³/sp² ratio arising from adopting different growth conditions selected from previously reported experiments.²⁴ Diamond film deposition and fibre production were carried out in a 2.45 GHz microwave plasma enhanced chemical vapour deposition (MPECVD) reactor described elsewhere.²⁵ Graphite substrates for diamond deposition were 20 × 20 × 3 mm³ highly oriented pyrolytic graphite (HOPG) pieces (Ringsdorff Werke GmbH, Bonn, Germany).

Diamond film deposition was achieved under the following conditions. For sample 1, the film was grown on an ultrasonically scratched substrate (Bandelin Sonorex Digital 10P ultrasonic bath, for 30 min) with a suspension prepared with 3% microdiamond in methanol [1 μm natural diamond, de Beers]. CVD deposition was carried out using 3 kW input microwave power for 2 hours. For sample 2, the film was grown on an ultrasonically scratched substrate with a suspension of 3% nanodiamond/methanol (detonation nanodiamond that was previously cleaned by the procedure described by Jiang and Xu²⁶) and deposited using 1.5 kW microwave power for a total of 6 hours. For sample 3, the film was grown on a substrate manually scratched with a slurry of diamond powder [1 μm natural diamond, de Beers] in methanol and deposited using 1.5 kW for a total of 6 hours. The gas mixture used in the experiments was 5% CH₄/15% H₂/80% Ar with a total gas pressure of 100 ± 2 Torr. For the growth of carbon fibres a crystalline silicon wafer piece of 5 × 5 mm² was placed on top of the films and a hydrogen plasma treatment was carried out for 15 min using 4.2 kW power at 50 ± 2 Torr. All substrates were heated by the plasma, without external heating, and the sample temperature was measured using a two-colour optical pyrometer.

Samples were characterised by scanning electron microscopy (SEM, Hitachi 2700, operated at 10 keV electron energy and with a secondary electron emission detector) and Raman spectroscopy (Renishaw Ltd, inVia Raman microscope) with an excitation wavelength of 514.5 nm. The Raman spectra were collected over the range from 1000 to 2000 cm⁻¹ and results were fitted using Gaussian profiles in OriginLab© Data Analysis and Graphing Software. Optical emission spectra of the plasma were measured using a Monolite 6800 spectrometer and corrected for detector and grating responses using a calibration factor obtained with a near-blackbody tungsten lamp. X-ray Photoelectron Spectroscopy (XPS) measurements, using Mg Kα X-radiation, were carried out in a Scienta ESCA300 spectrometer at the National Centre for Electron Spectroscopy and Surface Analysis (NCESS) UK. Deconvolution of XPS peaks was performed using XPSPeak41 software, using a linear background correction and fitted to Lorentzian–Gaussian function.

3. Results and discussion

Carbon fibres were only produced in our experiments when silicon was present, in contrast to experiments reported by other research groups which claim that the use of a catalyst is not necessary.²³ Hydrogen etching in diamond, under similar experimental conditions to the ones presented here, reveals steps and pits and removes primarily non-diamond material.²⁷ When a piece of silicon wafer is introduced on top of the samples, this silicon is etched (Fig. 1a) and the growth of fibres is seen in the vicinity in localised areas, especially on hills and not in valleys by a mechanism similar to that proposed by Castro et al.²⁸ This mechanism suggests that there is a non-locality caused by competition between diffusing particles, where the growing species have higher probability of attaching to the surface protrusions than to surface valleys. We believe that silicon is transported by the plasma to the film's surface inducing nucleation sites as observed in nanocrystalline diamond films (Fig. 1b). Volatile SiH_x (x ≤ 3) radicals and SiH₄ have been reported to be directly generated from the Si surface by atomic hydrogen etching^29,30 in addition to the decomposition of monosilane in the plasma. The OES evidence could not confirm the chemical transport of silicon since the SiH band (420 nm) is obscured by an adjacent CH band.

Fig. 1 SEM micrographs from (a) etched silicon piece on the top of diamond films during fibre growth. (b) Cumuli formation in nanocrystalline diamond films.

Fibres grown on different diamond film morphologies and qualities (sp³/sp² ratio obtained by deconvolution of Raman spectra, Table 1) have different size and thickness. Our experiments have shown²⁴ that CVD diamond growth on ultrasonic pretreated graphite substrates produce polycrystalline diamond film consisting of {111} and {110} oriented crystals in the size range of 1 to 3 μm. The Raman spectrum presents a sharp diamond signal and an sp³/sp² ratio of 0.06, much higher than samples 2 and 3. Fibres grown on this film (Fig. 2a) are straight and thin in the range of 100 to 200 nm in diameter and 5 to 10 μm of length (dimensions estimated by SEM micrographs) as shown in Fig. 2c. These fibres grow perpendicular to the film, possibly on the edges and corners of the {111} and {110} crystallite surfaces (Fig. 2b). The concentration of carbon species (CH) in the plasma during etching of this film is lower compared with samples 2 and 3, as evidenced by OES (Fig. 6).

Table 1 sp³/sp² ratio determined by Raman and XPS analyses of the sample surfaces from the areas of the core line spectra (% atomic)

Sample	sp^3/sp^2 ratio determined by Raman spectroscopy	Carbon (sp^2 C1s) determined by XPS (% atomic)	C–O bonds (O1s) determined by XPS (% atomic)	Silicon (2p) determined by XPS (% atomic)
1	0.060 ± 0.006	4.3 ± 0.5	82.6 ± 8.6	13.0 ± 1.4
2	0.006 ± 0.0002	12.4 ± 1.3	31.8 ± 3.3	55.8 ± 5.8
3	0.004 ± 0.0003	11.5 ± 1.2	75.7 ± 7.9	12.8 ± 1.3

---

Fig. 2 SEM micrographs from sample 1 showing: (a) polycrystalline diamond film before carbon fibre growth; (b) fibre growth in local areas, (c) straight and thin carbon fibres.

Nanocrystalline diamond films are produced on graphite ultrasonically scratched with nanodiamond or manually scratched with microdiamond.²⁴ Both films present cauliflower-like structures (Fig. 3a and b), however deconvolution of Raman spectra showed they have different qualities which has an effect on the later fibre growth. Fibres in sample 2 (Fig. 3a) have different diameters which are in the range of 0.2 to 0.4 μm. As opposed to the fibres in sample 1 (Fig. 2b and c) these do not have a preferential growth direction, they are bent and present complex shapes. It has been reported³¹ that this type of structural development and ‘octopus-like’ growth is due to a large catalyst particle size, where several fibres grow simultaneously on the different catalyst faces. However, in our samples we have seen (Fig. 1b) that during the catalyst nucleation phase, cumuli are formed in different sizes and orientations. We believe that re-nucleation and the fibre collapsing during growth may induce the production of ‘octopus-like’ structures and not large catalyst particle size as it has been proposed.

Fig. 3 SEM micrographs from sample 2 showing: (a) nanocrystalline film, (b) after hydrogen etching showing fibre growth. Micrographs from sample 3 showing: (c) nanocrystalline film before etching and (d) short and thick fibres.

Fig. 3c and d show SEM results from sample 3. The fibres were grown on a nanocrystalline film pre-treated by manually scratching with diamond powder. These fibres differ from those produced in sample 1 and 2, being thick and short in the range of 1 to 2 μm diameter and 3 to 7 μm length. The production of thicker fibres may be related to high concentration of CH species in the plasma, as a consequence of high etching rate of non-diamond material. A high concentration of growing species may lead to re-nucleation on the growing fibres which in turn, limits or prevents homoepitaxial growth. The high growth rate may influence the thickness of the fibres as proposed by Bower et al.³² for nanotubes.

The growth mechanism of the fibres presented here may be similar to that for nanotubes.³² The growth of the fibres is from the bottom and stops when the metal particle (catalyst) is saturated by carbon species; from that point the fibres begin to increase in diameter.

Wunderlich³³ proposed that low concentrations of carbon growing species produce thin nanotubes (e.g. single wall carbon nanotubes) that have fast reaction velocity. Multi-wall carbon nanotubes and graphite fibres have lower reaction velocities since they have to fill the three-dimensional volume. These features of the mechanism explain the variation of the thickness of fibres seen in samples 1, 2 and 3. Thick and short fibres in sample 3 are the result of low reaction velocities, compared to the thin and long fibres in sample 1.

XPS analysis (Table 1) shows the presence of Si in the surface of each of the etched samples, supporting the idea of silicon species being transported to the fibre tips during growth. XPS also shows high content of oxygen bonded to carbon as C–O–C, C–OH, COOH or COOR. Spectra show contributions from carbon sp² and carbon bonded with oxygen. The presence of peaks from ether and especially carboxylic acid (or ester) functional groups at ∼288 eV, can be compared with carbon fibres that have been oxidised in strong acid mixtures.³⁴ Since oxygen is not deliberately added to the gas mixture it is conceivable that either the samples were oxidised on exposure to air following removal from the vacuum chamber, or during plasma exposure to oxygen arising from the native oxide or the bulk oxygen in the wafer. A typical deconvoluted C1s peak is shown in Fig. 4.

Fig. 4 Typical deconvolution of C1s peak of the prepared carbon fibres, this spectrum corresponds to sample 2.

Fig. 5 shows the results from the Raman spectroscopy analysis for all the samples. The spectra present contributions from the broad D-band at ∼1350 cm⁻¹, and the G-band at ∼1550–1590 cm⁻¹ due to graphite. Samples 1, 2 and 3 also present the diamond peak at around 1332 cm⁻¹ which arises from the substrate. The Raman analyses of the samples have contributions that correspond to the reported Raman spectra for carbon fibres which exhibit two Raman-active modes. The E_2g mode at around 1350 cm⁻¹ has been related to polycrystalline graphite and its intensity depends on the particle size.³⁵ The second mode is the A_1g at around 1580 cm⁻¹ which has been identified with the doubly degenerate deformation mode of the hexagonal ring structure, observed in graphite single crystals. The 1332 cm⁻¹ signal from the diamond phase suggests that the spectra are a convolution of contributions from the carbon fibres and the diamond film.

Fig. 5 Raman spectra of the prepared carbon fibres.

OES spectra obtained during the hydrogen plasma treatment for samples 1, 2 and 3 after 15 min of treatment are shown in Fig. 6. The spectra show emission bands from the Balmer series at 656.5, 485 and 587 nm corresponding to Hα, Hβ and Hγ. The strong lines at 471 and 516 nm from C2 were not seen in any of these etching experiments although C2 radical has been reported by Mori et al.³⁶ to be responsible for the carbon fibre growth. Contributions from CH bands at 388, 418, 430 and 776 nm were seen in the three samples. These bands are weak in sample 1 and 2, but very intense in sample 3 due to greater surface erosion.

Fig. 6 Optical emission spectra acquired during fibre growth under hydrogen plasma conditions.

4. Conclusions

The production of carbon fibres, reported in this study, occurs when diamond films are exposed to hydrogen plasmas in the presence of silicon. The fibre morphology depends on the structure of the diamond film used as substrate, and the concentration of growth species during CVD conditions. Silicon plays an important catalytic role in the formation of fibres and is transported to the fibre during growth by a mechanism similar to that proposed for nanotubes. Fibre surfaces are oxygen terminated evidenced by the presence of COO and CO functional groups. Raman analysis confirmed the presence of graphitic structures but peaks from polycrystalline diamond from the diamond films also appeared.

The results of this work open the possibility to other applications such as catalysis, sensors and the production of electrodes; since they combine the unmatchable properties of a diamond supporting substrate with the unique properties of carbon fibres.

Acknowledgements

I. V. was supported by CONACYT and by the Programme Alβan, the European Union Programme of High Level Scholarships for Latin America, scholarship no. E05D056416MX. Funding from EPSRC (E/035868/1) is also gratefully acknowledged. M. C. Victor Valles-Gomez for proof reading the article.

References

1. S. Porro, G. De Temmerman, P. John, S. Lisgo, I. Villalpando and J. I. B. Wilson, Phys. Status Solidi A1, 2009, 206, 2028.
2. G. De Temmerman, R. P. Doerner, P. John, S. Lisgo, A. Litnovsky, L. Marot, S. Porro, P. Petersson, M. Rubel, D. L. Rudakov, G. Van Rooij, J. Westerhout and J. I. B. Wilson, Phys. Scr., 2009, T138, 014013.
3. S. Porro, G. De Temmerman, D. A. MacLaren, S. Lisgo, D. L. Rudakov, J. Westerhout, M. Wiora, P. John, I. Villalpando and J. I. B. Wilson, Diamond Relat. Mater., 2010, 19, 818.
4. S. Porro, S. Musso, M. Giorcelli, A. Tagliaferro, S. H. Dalal, K. B. K. Teo, D. A. Jefferson and W. I. Milne, J. Non-Cryst. Solids, 2006, 352, 1310.
5. N. M. Rodriguez, A. Chambers, R. Terry and R. T. K. Baker, Langmuir, 1995, 11, 3862.
6. R. T. K. Baker, Carbon, 1989, 27, 315.
7. S. S. Tzeng, K. H. Hung and T. H. Ko, Carbon, 2006, 44, 859.
8. B. Louis, R. Vieira, A. Carvalho, J. Amadou, M. J. Ledoux and C. Pham-Huu, Top. Catal., 2007, 45, 75.
9. M. Endo, K. Takeuchi, S. Igarashi, K. Kobori, M. Shiraishi and H. W. Kroto, J. Phys. Chem. Solids, 1993, 54, 1841.
10. A. Chambers, C. Park, R. Terry, R. T. K. Baker and N. M. Rodriguez, J. Phys. Chem. B, 1998, 102, 4253.
11. A. Oberlin, M. Endo and T. Koyama, J. Cryst. Growth, 1976, 32, 335.
12. W. X. Lu, Z. J. Sui, J. H. Zhou, P. Li, D. Chen and X. G. Zhou, Chem. Eng. Sci., 2010, 65, 193.
13. J. H. Bitter, M. K. Van der Lee, A. G. T. Slotboom, A. J. Van Dillen and K. P. de Jong, Catal. Lett., 2003, 1–2, 139.
14. F. Rodriguez-Reinoso, Carbon, 1998, 36, 159.
15. E. Boellaard, P. K. de Bokx, A. J. H. M. Kock and J. W. Geus, J. Catal., 1985, 96, 481.
16. G. G. Tibbetts, M. L. Lake, K. L. Strong and B. P. Rice, Compos. Sci. Technol., 2007, 67, 1709.
17. G. A. Kovalenko, E. V. Kuznetsova, Y. I. Mogilnykh, I. S. Andreeva, D. G. Kuvshinov and N. A. Rudina, Carbon, 2001, 39, 1033.
18. H. Suda and K. Haraya, J. Phys. Chem. B, 1997, 101, 3988.
19. B. T. Hang, M. Eashira, I. Watanabe, S. Okada, J. I. Yamaki, S. H. Yoon and I. Mochida, J. Power Sources, 2005, 143, 256.
20. C. A. Bessel, K. Laubernds, N. M. Rodriguez and R. K. T. Baker, J. Phys. Chem. B, 2001, 105, 1115.
21. S. Flandrois and B. Simon, Carbon, 1999, 37, 165.
22. Z. Zhao, J. Gou and A. Khan, J. Nanomater., 2009, 325769.
23. K. Kobashi, T. Tachibana, Y. Yokota, N. Kawakami, K. Hayashi and K. Inoue, Diamond Relat. Mater., 2001, 10, 2039.
24. I. Villalpando, PhD thesis, Heriot-Watt University, Edinburgh UK, 2010.
25. J. R. Rabeau, P. John, J. I. B. Wilson and Y. Fan, J. Appl. Phys., 2004, 96, 6724.
26. T. Jiang and K. Xu, Carbon, 1995, 33, 1663.
27. I. Villalpando, P. John, S. Porro and J. I. B. Wilson, Diamond Relat. Mater., 2011, 20, 711.
28. M. Castro, R. Cuerno, M. Nicoli, L. Vazquez and J. G. Buijnsters, New J. Phys., 2012, 14, 103039.
29. T. Yamada, H. Ohmi, K. Okamoto, H. Kakiuchi and K. Yasutake, Jpn. J. Appl. Phys., 2012, 51, 10NA09.
30. H. Ohmi, T. Funaki, H. Kakiuchi and K. Yasutake, J. Phys. D: Appl. Phys., 2016, 49, 035202.
31. W. Xia, J. H. Bitter, D. Su, J. Qian and M. Muhler, Diamond Relat. Mater., 2009, 18, 13.
32. C. Bower, O. Zhou, W. Zhu, D. J. Werder and S. Jin, Appl. Phys. Lett., 2000, 77, 2767.
33. W. Wunderlich, Diamond Relat. Mater., 2007, 16, 369.
34. S. D. Gardner, C. S. K. Singamsetty, G. L. Booth, G. R. He and C. U. Pittman, Carbon, 1995, 33, 587.
35. S. N. Chaudhuri, R. A. Chaudhuri, R. E. Benner and M. S. Penugonda, Comput. Struct., 2006, 76, 375.
36. S. Mori and M. Suzuki, Diamond Relat. Mater., 2008, 17, 999.

---

Footnotes

† Present address: Antigua Normal Rural de Salaices, López, Chihuahua, México, C. P. 33941, Fax +52 629 534 6023, Tel. +52 629 534 6048.

‡ Present address: Department of Applied Science and Technology, Politecnico di Torino, Corso Duca degli Abruzzi 24, 10129 Torino, Italy.

---