Selective Area Laser Deposition of FCC Beta Silicon Carbide

Title

Selective Area Laser Deposition of FCC beta Silicon Carbide

Authors

*James I. Paul, Marc J J. Schmidt, Timothy J. Abram

*Corresponding author

Affiliation

University of Manchester

School of Mechanical, Aerospace and Civil Engineering, The University of Manchester, Manchester, UK, M13 9PL

Abstract

Silicon Carbide (SiC) has been deposited onto an alumina substrate by the thermal decomposition of the gaseous precursor tetramethylsilane (TMS). A 500 W ytterbium fibre laser was used to heat the surface of an alumina substrate locally, resulting in deposition of SiC at the sample surface. The SiC deposit was analysed using energy dispersive X-ray spectroscopy and X-ray diffraction (XRD). The deposit was confirmed to be silicon carbide and found to be face centre cubic (FCC) crystal structure. Raman spectroscopy was used to measure the stoichiometry of the deposit which initially was found to be carbon rich. Further analysis by Raman spectroscopy suggests the deposit may be more stoichiometric following a two hour thermal treatment of the sample at 600 degrees celcius in an atmosphere of air.

Keywords

Nuclear Applications (E) SiC (D) Films (A) Joining

Introduction

It has been proposed that silicon carbide may be a future cladding material for pressurised water reactors (PWRs) however the joining of ceramic components is difficult in comparison to current PWR nuclear cladding materials such as metallic zirconium alloys. Conventional welding processes are not suitable due to the lack of a silicon carbide (SiC) liquid phase. Instead SiC sublimates between 2000ºC and 2700ºC, depending on the specific crystal structure [1]. Ceramics are difficult to join using mechanical fasteners such as bolts or rivets since the resulting stress concentrations at the join locations may lead to brittle fracture. Any joining techniques used for ceramics should avoid such stress concentrations in the material.

The chemical vapour deposition of SiC films from gaseous phase precursors has been carried out commercially for a number of years in the manufacture of TRISO coated particles for High Temperature Reactors (HTRs) and as a result it is an established technique for the manufacture of SiC and pyrolytic carbon layer. Localised deposition of SiC is a relatively immature technique having been demonstrated previously only on a laboratory scale [2,3] and has been applied to the joining of SiC with limited success as the join strength was found to be weak in comparison to the strength of the clay bonded SiC tube that was used for the joining experiments [4]. A new SALD method has been developed that uses tetramethylsilane as the precursor gas and nitrogen as the carrier gas. Deposition of SiC was observed and validated using EDX analysis. The crystal structure was determined to be face centre cubic (FCC) using X-Ray Diffraction (XRD) and stoichiometry was measured using Raman spectroscopy.

Material and Methods

A cylindrical chamber was constructed from mild steel to contain the sample and precursor gasses during the deposition process. The chamber included fittings to allow for the inlet and exhaust of process gasses. The top was sealed using a laser window which provided a gas tight seal to the chamber yet allowed the laser light to pass through and heat the surface of the substrate. A schematic of the SALD equipment set up is shown in Figure 1.

Samples of alumina were used as the deposition substrates. The samples were cleaned using ethanol before being coated with a layer of graphite powder to improve the absorption of the laser. Samples were placed in the base of the SALD chamber before it was sealed using a borosilicate laser window. A bead of silicon grease between the top of the chamber and the laser window ensured a gas tight seal.

The SALD chamber was purged with 150 cm3/s nitrogen gas for 60 minutes in order to ensure there was an inert atmosphere within the system. Nitrogen gas was heated to 30ºC using a heat exchanger and bubbled through TMS to pick it up in the nitrogen flow as vapour. A 500 W peak power ytterbium fibre laser with 1070 nm wavelength was used to heat the sample surface to a temperature sufficient to decompose TMS (1100 – 1250ºC [5]) into Methane and SiC according to Equation 1. Laser parameters that were found to reliably produce a deposit were a peak power of 200 W, a scan speed of 750 mm/s, a frequency of 10 kHz, a pulse length of 8 μs and a 150 μm laser spot size. The laser was scanned in a circular pattern, 6 mm in radius with a 0.5 mm hatch producing a 12 mm diameter SiC deposit on the surface of the sample.

The exhaust gasses were removed from the chamber using a PUREX fume extractor and the SiC was deposited at the focal point of the laser. Any unburned TMS in the exhaust was recaptured by precipitation as the exhaust gasses passed through a second heat exchanger cooled with salted ice to a temperature of -10ºC. The system was again purged and the extraction continued for thirty minutes before the sample was removed from the chamber. The surface of the sample was cleaned using a sharp blast of compressed air and the surface was wiped with an ethanol soaked cloth. This cleaning process removed any loose powder from the surface of the deposit.

Equation 1

Si(CH_3)_4 \rightarrow SiC+CH_4

Results

Characterisation of the deposit

The deposit was sputter coated with a gold-palladium alloy to improve the electrical conductivity and mounted in a scanning electron microscope. The deposit was observed and the interface between the deposited SiC and the Alumina was pronounced (Figure 2). Large surface cracks were observed on the SiC deposit.

The deposit was confirmed to be containing silicon and carbon using Energy Dispersive X-ray Spectroscopy (EDX). X-ray energy peaks corresponding to silicon, carbon, aluminium and oxygen were prominent corresponding to the alumina substrate (aluminium oxide) and the silicon carbide deposit. The elements were found to be in the abundance shown in Table 1 as a result of the EDX analysis of Figure 3.

The sample appears to contain FCC β-SiC, carbon and alumina from the use of X-Ray diffraction. The diffraction pattern of an as received sample is shown in Figure 4. Rietveld fitting was applied with a goodness of fit 1.611 and R (Weighted Profile) of 9.25. The XRD pattern of the same sample following thermal treatment is shown in Figure 5 with the peaks corresponding to FCC SiC annotated. Rietveld fitting was applied with a goodness of fit 2.10 and R (Weighted Profile) of 12.13.

\begin{figure}[htp]

\centering

\includegraphics[width=0.45\textwidth]{asrec.eps}

\caption{X-ray diffraction pattern of SiC deposited onto a graphite coated alumina substrate. Peaks corresponding to FCC SiC are annotated. Rietveld fitting was applied with a goodness of fit 1.611 and R (Weighted Profile) of 9.25.}

\label{sicxrd}

\end{figure}

\begin{figure}[htp]

\centering

\includegraphics[width=0.45\textwidth]{ann.eps}

\caption{XRD pattern of SiC deposited onto an alumina substrate following a thermal treatment process with peaks annotated corresponding to face centre cubic SiC. Rietveld fitting was applied with a goodness of fit 2.10 and R (Weighted Profile) of 12.13.}

\label{sicxrd2}

\end{figure}

.1 Soichiometry

The stoichiometry of the deposit was analysed using Raman spectroscopy. Figure [6] shows Raman traces of silicon rich, stoichiometric and carbon rich silicon carbide [6]. The 3C polytype of SiC is known to have two Raman peaks at 790 and 970 cm-1 corresponding to the TO and LO phonon respectively. Silicon rich silicon carbide will also have a sharp peak at 520 cm-1. Carbon rich silicon carbide has peaks approximately at 1350 and 1500 cm-1. If the Raman spectra does not contain these additional peaks it is said to be stoichiometric, possessing no free carbon or free silicon. The surface of the deposit was analysed at three random locations on the sample, the Raman traces area shown in Figure [7]. The presence of large carbon peaks at 1350 and 1500 cm-1 show that the SiC deposit is carbon rich. The sample was thermally treated at 600ºC to oxidise the free carbon. The Raman traces of the thermally treated sample at three random locations is shown in Figure [8]. The post thermal treatment Raman traces still possess the two carbon peaks, suggesting that there exists carbon contamination in the sample despite the attempt to remove it through thermal treatment.

The localised deposition of solid material has been successfully demonstrated onto an alumina substrate. The XRD and Raman spectroscopy studies support the hypothesis that this deposit may be 3C-SiC. The ability to deposit SiC locally has a number of potential uses such as a hard wearing coating for cutting tools, the surface coating of component surfaces to improve wear resistance, to increase hardness or increase friction, an additive manufacture process to manufacture three dimensional SiC components and, as previously mentioned a potential ceramic joining technique.

For nuclear applications, it is important to use β-SiC due to its stability when subjected to irradiation [7]. The SALD deposition of SiC was appears to be β phase through the use of X-ray diffraction. Selective area laser deposition may be a useful technique in the manufacture of components for use in nuclear environments where radiation resistance is an important consideration.

The surface of the SiC deposit was seen to contain large surface cracks when viewed using a scanning electron microscope. The presance of these surface cracks may be due to a difference in the thermal expansion of the deposit and the substrate. These cracks may have a detrimental effect on the mechanical strength of a material which may make it unsuitable for a nuclear joining technology where high reliability is required.

A highly pure silicon carbide with no excess carbon or excess silicon is required for nuclear applications [8]. The stoichiometry of the deposit was measured using Raman spectroscopy. The deposit was found to contain excess carbon. For nuclear applications the most desirable is stoichiometric silicon carbide. Stoichiometric silicon carbide is known for its predictable irradiation swelling response. The intensity of Raman peaks reduced following a two hour thermal treatment at 600ºC suggesting that the carbon content of the sample had reduced.

Conclusions

The difficulty in joining of ceramics with high strength is one of the challenges with the use of SiC as a PWR cladding material. All ceramic joining methods to date are based on the addition of material to create a join. It has been proposed that selective area laser deposition could be used to deposit SiC at the join interface in order to create a SiC join. This study suggests that the deposited SiC from TMS using a Nitrogen atmosphere and carrier gas is of β phase which is important for nuclear applications due to its preferential irradiation performance. In addition, the thermal expansion and thermal conductivity of the join material would be similar to the SiC cladding tubes, so thermal stresses and operating temperatures of the join are likely to be similar to SiC-SiC cladding.

Although efforts have been made to remove the excess carbon through the thermal treatment of the deposit, excess carbon is still present in the sample. The carbon contamination has been significantly reduced by thermal treatment at 600ºC. Future work will focus on the manufacture of stoichiometric silicon carbide removing the necessity for a thermal treatment process.

Aknowledgements

The authors are indebted to the technical staff at The University of Manchester for their assistance, in particular Mr Daniel Wilson for his expertise in the manufacture and setup of equipment. This work was funded by the Engineering and Physical Sciences Research Council (EPSRC).

References

[1] M. Pons, E. Blanquet, J. M. Dedulle, I. Garcon, R. Madar, C. Bernard, Journal of The Electrochemical Society 143 (1996) 3727–3735. doi:10.1149/1.1837280.

[2] J. Mi, W. J. Lackey, Journal of Materials Processing Technology 209 (2009) 3818 – 3829.

doi:

[3] K. Jakubenas, J. Sanchez, H. Marcus, Materials & Design 19 (1998) 11 – 18. doi: 3069(98)00004-1.

[4] S. Harrison, H. Marcus, Materials & Design 20 (1999) 147 – 152. doi:

[5] K. C. Kim, K. S. Nahm, Y. B. Hahn, Y. S. Lee, H.-S. Byun, Journal of Vacuum Science & Technology A 18 (2000) 891–899.

[6] R. A. Shatwell, K. L. Dyos, C. Prentice, Y. Ward, R. J. Young, Journal of Microscopy 201 (2001) 179–188. URL: doi:10.1046/j.1365-2818.2001.00836.x.

[7] K. Sickafus, in: R. J. Konings (Ed.), Comprehensive Nuclear Materials, Elsevier, Oxford, 2012, pp. 123 – 139. doi:

[8] L. Snead, Y. Katoh, T. Nozawa, in: R. J. Konings (Ed.), Comprehensive Nuclear Materials, Elsevier, Oxford, 2012, pp. 215 – 240 doi:

Figure 1

A Schematic of the SALD equipment, showing the layout of the components in the system.

Figure 2

Interface between the deposited SiC and the Alumina substrate as viewed using scanning electron microscopy at 250x magnification

Figure 3

Energy Dispersive X-Ray Spectroscopy of the interface between the SiC deposit and the Alumina Substrate annotated showing the carbon, oxygen, aluminium and silicon peaks corresponding to the SiC deposit and the Al 2 O 3 substrate.

Figure 4

X-ray diffraction pattern of SiC deposited onto a graphite coated alumina substrate. Peaks corresponding to FCC SiC are annotated. Rietveld fitting was applied with a goodness of fit 1.611 and R (Weighted Profile) of 9.25.

Figure 5

XRD pattern of SiC deposited onto an alumina substrate following a thermal treatment process with peaks annotated corresponding to face centre cubic SiC. Rietveld fitting was applied with a goodness of fit 2.10 and R (Weighted Profile) of 12.13.

Figure 6

Raman spectroscopy of Carbon rich, stoichiometric and Silicon rich silicon carbide \cite{Shatwell2001}

Figure 7

Raman spectrum of the SiC deposit as deposited without thermal treatment at 3 locations on the sample surface.

Figure 8

Raman spectrum of the SiC deposit after a 2 hour 600ºC thermal treatment at 3 locations on the sample surface.

Table 1

1