Decomposition of silicon carbide at high pressures and temperatures

Kierstin Daviau and Kanani K. M. Lee

Phys. Rev. B 96, 174102 – Published 3 November 2017

Abstract

We measure the onset of decomposition of silicon carbide, SiC, to silicon and carbon (e.g., diamond) at high pressures and high temperatures in a laser-heated diamond-anvil cell. We identify decomposition through x-ray diffraction and multiwavelength imaging radiometry coupled with electron microscopy analyses on quenched samples. We find that B3 SiC (also known as 3C or zinc blende SiC) decomposes at high pressures and high temperatures, following a phase boundary with a negative slope. The high-pressure decomposition temperatures measured are considerably lower than those at ambient, with our measurements indicating that SiC begins to decompose at $\sim 2000$ K at 60 GPa as compared to $\sim 2800$ K at ambient pressure. Once B3 SiC transitions to the high-pressure B1 (rocksalt) structure, we no longer observe decomposition, despite heating to temperatures in excess of $\sim 3200$ K. The temperature of decomposition and the nature of the decomposition phase boundary appear to be strongly influenced by the pressure-induced phase transitions to higher-density structures in SiC, silicon, and carbon. The decomposition of SiC at high pressure and temperature has implications for the stability of naturally forming moissanite on Earth and in carbon-rich exoplanets.

Received 13 June 2017
Revised 14 August 2017

DOI:https://doi.org/10.1103/PhysRevB.96.174102

Physics Subject Headings (PhySH)

Liquid-solid phase transition Phase transitions Pressure effects Solid-solid transformations Temperature

Raman spectroscopy X-ray diffraction

Condensed Matter, Materials & Applied Physics

Authors & Affiliations

Kierstin Daviau^* and Kanani K. M. Lee

Department of Geology & Geophysics, Yale University, New Haven, Connecticut 06511, USA

^*kierstin.daviau@yale.edu

Article Text (Subscription Required)

Click to Expand

Supplemental Material (Subscription Required)

Click to Expand

References (Subscription Required)

Click to Expand

Issue

Vol. 96, Iss. 17 — 1 November 2017

Reuse & Permissions

Access Options

Article Available via CHORUS

Download Accepted Manuscript

Author publication services for translation and copyediting assistance advertisement

Images

Figure 1
(a) Diffraction patterns of temperature-quenched samples from before, during and after double-sided laser heating B3 SiC at ∼20 GPa (XRD_002). The hkl reflections of B3 SiC and Ne are labeled, as well as those from C diamond that appear after heating to high temperatures. In this case diamond appeared at upstream (downstream) temperatures of 2160 K (2333 K), but the temperature was increased to over 2400 K before taking a temperature-quenched pattern. Reflections from the Re gasket are labeled with an asterisk, reflections from 6H-SiC are labeled with a star, and an unidentified peak appearing at high temperature is labeled with an empty circle. (b) Optical images of the same sample taken after heating. The left image is of the sample under pressures at ∼22 GPa while the right image is after decompression. There are three different regions of the sample: (1) gray, unheated material on the outer edge of the sample chamber, (2) translucent material surrounding the hottest region, and (3) a black, opaque center where the laser was focused and temperatures, presumably, the hottest. We do not observe evidence for diamond-anvil damage upon unloading the sample.
Reuse & Permissions
Figure 2
(a) Photomicrograph of 4CLR_003, heated at 39 GPa, displaying characteristic optical changes across the heated location using both transmitted and reflected light. (b) BSE image of the corresponding FIB’ed cross section across the hot spot. Compositional map of sample cross sections measured by EPMA EDS of (c) C and (d) Si. Brighter areas denote higher relative abundance, while darker areas denote lower relative abundance. We observe concentrated C-rich and Si-poor grains as well as a larger central region with slight Si enrichments.
Reuse & Permissions
Figure 3
Progression of the absorption changes observed in a 42-GPa sample laser heated with temperatures measured by multiwavelength imaging radiometry [24]. The top row of images is of the sample in the 640-nm wavelength captured by the CCD camera during the heating experiment. The bottom row contains the corresponding emissivity and temperature maps found by fitting three additional wavelength images (580, 766, 905 nm). Top row from left to right: (a) transmitted white light through the sample after initial annealing at low temperatures; (b) in situ thermal emission from midway through heating (52 W of laser power); (c) white light image after 52 W heating (note enhanced transparency as compared to surrounding areas); (d) in situ thermal emission of final heating at 68 W; (e) white light after final 68 W heating. Bottom row from left to right: (f) in situ emissivity and temperature map from the 52 W heating; (g) in situ emissivity and temperature map from the 68 W heating. Notice that the sample becomes significantly more transparent over the heated region (green outline) and that a darker region becomes apparent in the white light image after the final heating. The emissivity and temperature maps match up well during the early heating but become uncorrelated later in the heating run. Green outlines correspond to the region of the image where temperature and emissivity are mapped.
Reuse & Permissions
Figure 4
Raman spectra from 4CLR_006 after heating at 81 GPa using the multiwavelength imaging radiometric temperature mapping system. The top two patterns are of the compressed sample while the bottom two patterns are of the same sample after unloading from the diamond cell. We see a small additional peak at $\sim 1470 c m^{- 1}$ in the center of the heated area while at pressure that is absent from the surrounding region as well as from the unloaded pattern. After unloading we see no evidence of decomposition throughout the hottest region but do see the characteristic $D$ and $G$ band signals of CDC in the surrounding, annealed area. The lack of decomposition in the hot spot is supported by in situ measurements, as throughout the entire heating and up to the highest temperatures of ∼3200 K, the temperature and emissivity maps continue to display symmetric behavior across the hot spot.
Reuse & Permissions
Figure 5
The room-temperature equations of state for SiC [20, 63], Si [44], and C [43, 64] including high-pressure phase transitions (solid curves) as well as the curve for across the SiC to Si+C decomposition reaction (dashed curve). The $Δ V$ for decomposition is large and positive at low pressures ( $< \sim 12$ GPa) but decreases to negative values upon the transition of Si I to Si II. The $Δ V$ remains negative until B3 SiC transitions to the higher-density B1 structure. At equilibrium, this reaction occurs at ∼58 GPa (i.e., [63]) but is kinetically hindered experimentally and is not seen before 100 GPa at room temperature [19]. The shaded region of the SiC and $Δ V$ curve represents the region across which the SiC volume, and therefore the $Δ V$ at decomposition, is dependent on the experimental conditions.
Reuse & Permissions
Figure 6
Phase diagram for the high-pressure decomposition of SiC. Conditions at which the first sign of diamond diffraction appears are represented by solid circles, conditions at which absorption changes are observed by multiwavelength imaging radiometry are represented by solid triangles, and the conditions at which B1 SiC shows no evidence of decomposition is represented by an open square. Temperature error bars are as described in Tables 1 and 2. Pressure error bars for the diffraction data are from errors in the Ne volume and are smaller than the symbols [20, 28]. Pressure error bars for the multiwavelength imaging radiometric measurements are from the spread in pressure across the sample chamber before and after heating as determined by the Raman edge of diamond [27]. Previous data observing decomposition are represented by solid diamonds [11] and by asterisks for the previously reported extrapolated phase boundary [17]. Previous data that did not observe decomposition are represented by plus signs [16]. We show the regions of phase stability through colored shading with blue corresponding to B3 SiC, green to Si+C, pink to B1 SiC, and the mixed purple region representing the area where the B3 to B1 transition is kinetically hindered. Temperature profiles for Earth's mantle (dash-dot [54], dotted [55]) as well as for a subducting slab (dashed [56]) are plotted. Both mantle geotherms cross the decomposition boundary for SiC, indicating that moissanite has a region of instability within the mantle.
Reuse & Permissions

Physical Review B

covering condensed matter and materials physics