Radiometric temperature measurements in non-gray ferropericlase
Abstract
Accurate temperature (T) measurements in laser-heated (LH) diamond anvil cells (DACs) are crucial to determine physico-chemical properties of minerals in deep planetary interiors. Fitting Planck's curve, scaled by material emissivity, to sample's thermal emission spectrum yields T and emissivity (ɛ). Virtually all LH DAC studies assume wavelength-independent ɛ (graybody approximation). However, iron-bearing mantle minerals are semi-transparent and may show distinct absorption bands at 600-900 nm, the spectral range of radiometric T measurements in LH DACs. We explore the potential of the graybody approximation to introduce systematic errors into the T determination of ferropericlase, the second most abundant lower mantle mineral, in a LH DAC. We model thermal emission/absorption in a sample at a peak T (Tmax), using realistic axial T profiles and variable sample/thermal insulation filling fraction. Our models are based on the experimentally-determined pressure- (P) and T-dependent absorption coefficient α(l) of ferropericlase across the 600-900 nm spectral range.
In the first scenario, assuming T-independent α(l), we model ferropericlase laser-heated at 40, 75 and 135 GPa to 900 K ≤ Tmax ≤ 2300 K. Without thermal insulation, the graybody approximation causes a systematic error of ΔT = T - Tmax ~±200 K. The sign and magnitude of ΔT is sensitive to both the wavelength-dependence of α(l) and filling fraction. In the second scenario (P = 75 GPa, 900 K ≤ Tmax ≤ 2300 K), α(l) is spin-dependent and is changing continuously with T due to a temperature-induced low-to-high spin transition. Again, we find that the variable filling fraction may result in an error of -50 K < ΔT < 200 K. At T > 2300 K, however, the sample's T may be underestimated by more than 200 K. Finally, we model the full P- and T-dependence of α(l) of ferropericlase at 135 GPa and Tmax = 4000 K, that is at core-mantle boundary conditions. In this scenario the effect of sample geometry is extreme, causing -300 < ΔT < 100 K with thermal insulation and -300 < ΔT < 600 K without. Our model results are applicable to the experimental solidus of ferropericlase at lower mantle pressures, and show that experimentally-constrained optical α(l) of lower mantle minerals can correct for systematic errors in temperature determination in LH DAC experiments.- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2019
- Bibcode:
- 2019AGUFMMR23D0143L
- Keywords:
-
- 3919 Equations of state;
- MINERAL PHYSICS;
- 3924 High-pressure behavior;
- MINERAL PHYSICS;
- 3939 Physical thermodynamics;
- MINERAL PHYSICS;
- 3994 Instruments and techniques;
- MINERAL PHYSICS