Skip to content

Conductivity ​

Methods ​

Thermal conductivity is defined as

GeoParams.MaterialParameters.Conductivity.ConstantConductivity Type
julia
ConstantConductivity(k=3.0W/m/K)

Set a constant conductivity

k=cst

where k is the thermal conductivity [W/m/K].

source
GeoParams.MaterialParameters.Conductivity.T_Conductivity_Whittington Type
julia
T_Conductivity_Whittington()

Sets a temperature-dependent conductivity following the parameterization of _Whittington, A.G., Hofmeister, A.M., Nabelek, P.I., 2009. Temperature-dependent thermal diffusivity of the Earth’s crust and implications for magmatism. Nature 458, 319–321. https://doi.org/10.1038/nature07818._ Their parameterization is originally given for the thermal diffusivity, together with a parameterization for thermal conductivity, which allows us to compute

Cp=a+bT−c/T2κ=d/T−e if T<=846Kκ=f−g∗T if T>846Kρ=2700kg/m3k=κρCp

where Cp is the heat capacity [J/mol/K], and a,b,c are parameters that dependent on the temperature T:

  • a = 199.50 J/mol/K if T<= 846 K

  • a = 199.50 J/mol/K if T> 846 K

  • b = 0.0857J/mol/K^2 if T<= 846 K

  • b = 0.0323J/mol/K^2 if T> 846 K

  • c = 5e6J/mol*K if T<= 846 K

  • c = 47.9e-6J/mol*K if T> 846 K

  • d = 567.3m^2/s*K

  • e = 0.062m^2/s

  • f = 0.732m^2/s

  • g = 0.000135m^2/s/K

This looks like:

Example

julia
julia> using GLMakie, GeoParams
julia> p=T_Conductivity_Whittington();
julia> T,k,plt = PlotConductivity(p)
source
GeoParams.MaterialParameters.Conductivity.T_Conductivity_Whittington_parameterised Type
julia
T_Conductivity_Whittington_parameterised()

Sets a temperature-dependent conductivity that is parameterization after Whittington, et al. 2009

The original parameterization involves quite a few parameters; this is a polynomial fit that is roughly valid from 0-1000 Celsius

k[W/m/K]=−210−9(T−Ts)3+610−6(T−Ts)2−0.0062(T−Ts)+4Ts=273.15K

where T[K] is the temperature in Kelvin (or the nondimensional equivalent of it).

The comparison of this parameterisation vs. the original one is:

source
GeoParams.MaterialParameters.Conductivity.TP_Conductivity Type
julia
TP_Conductivity()

Sets a temperature (and pressure)-dependent conductivity parameterization as described in Gerya, Numerical Geodynamics (2nd edition, Table 21.2). The general for

k=(ak+bkT+ck)(1+dkP)

where k is the conductivity [W/K/m], and ak,bk,ck,dk are parameters that dependent on the temperature T and pressure P:

  • ak = 1.18Watt/K/m

  • bk = 474Watt/m

  • ck = 77K

  • dk = 0/MPa

source
GeoParams.MaterialParameters.Conductivity.Set_TP_Conductivity Function
julia
Set_TP_Conductivity(name::String)

Returns the pre-defined temperature(-pressure) dependent conductivity law name.

Available laws:

  • "UpperCrust"

  • "LowerCrust"

  • "OceanicCrust"

  • "Mantle"

Example

julia
julia> k = Set_TP_Conductivity("Mantle")
T/P dependent conductivity: Name = Mantle, k = (0.73 + 1293.0/(T + 77.0))*(1 + 4.0e-5*P)
source
GeoParams.MaterialParameters.Conductivity.T_Conductivity_Moitra Type
julia
T_Conductivity_Moitra(; ρ=ConstantDensity(ρ=3000kg/m^3), Cp=T_HeatCapacity_Moitra())

Thermal conductivity k=ρcpα (Moitra et al. 2018, eq. 2) from the temperature-dependent thermal diffusivity of a basanite glass (T<Tg) and melt (T≥Tg), their eqs. 5 and 6:

α=a0T−0.41+b0T(T<Tg),α=a1+b1T(T≥Tg)

with T in K and

  • a0 = 5.98 mm²/s K^0.41, b0 = 9.23e-5 mm²/s/K

  • a1 = -0.06 mm²/s, b1 = 4.57e-4 mm²/s/K

  • Tg = 1000 K

The density ρ and heat capacity Cp are laws evaluated with the same arguments, so k follows the material of the model while keeping the measured diffusivity. The defaults are the sample density and heat capacity of Moitra et al. (2018), with which k reproduces their Table 3 (298-1273 K). k jumps at Tg; T_Conductivity_Moitra_smoothed replaces the jump by a smooth transition. Above 1273 K the melt branches are extrapolated. Moitra et al. (2018, p. 3627) attribute the rise of α above Tg to crystallization of the sample, so the melt branch is not that of a crystal-free melt: at 1473 K it gives k≈2.6 W/m/K, about twice the 1-1.5 W/m/K typical of basaltic melt.

References

  • Moitra, P., Sonder, I., Valentine, G. A. (2018), Effects of size and temperature-dependent thermal conductivity on the cooling of pyroclasts in air, G-cubed 19, 3623-3636, https://doi.org/10.1029/2018GC007510
source
GeoParams.MaterialParameters.Conductivity.T_Conductivity_Moitra_smoothed Type
julia
T_Conductivity_Moitra_smoothed(; ρ=ConstantDensity(ρ=3000kg/m^3), Cp=T_HeatCapacity_Moitra_smoothed())

T_Conductivity_Moitra with the jump of the diffusivity at Tg replaced by a smooth transition of width ΔTg:

k=ρcpα,α=(1−H)αglass+Hαmelt,H=11+exp⁡(−4(T−Tg)/ΔTg)

with the glass and melt diffusivities of T_Conductivity_Moitra and the same H as T_HeatCapacity_Moitra_smoothed, the default Cp. The default ΔTg is 20 K.

source
GeoParams.MaterialParameters.Conductivity.IAPWS_H2O_Conductivity Type
julia
IAPWS_H2O_Conductivity(; ρ=HalbachChatterjee_Density(), L0, L1)

Thermal conductivity of H2O from the IAPWS 2011 formulation (Huber et al. 2012; IAPWS R15-11, eqs. 15-17) without the critical-enhancement term λ¯2:

λ=λ∗λ¯0(T¯)λ¯1(T¯,ρ¯),λ¯0=T¯∑k=04Lk/T¯k,λ¯1=exp⁡[ρ¯∑i=04(1T¯−1)i∑j=05Lij(ρ¯−1)j]

with T¯=T/T∗, ρ¯=ρ/ρ∗, T∗=647.096 K, ρ∗=322 kg/m³ and λ∗=10−3 W/m/K (eqs. 1, 3, 4). The dimensionless coefficients L0 (Lk, Table 1) and L1 (Lij, Table 2, row i, column j) default to the values of IAPWS R15-11. The density is the law ρ evaluated with the same arguments (P, T).

IAPWS R15-11 validates the formulation (its eq. 14) up to 1173.15 K for P≤ 100 MPa, up to 874 K for 100-250 MPa, and up to 573 K above 250 MPa; most of the range 600-1200 °C, 0.1-400 MPa is therefore an extrapolation of the standard. Within that range, omitting λ¯2 changes λ by at most 1.99 % (600 °C, 80 MPa) and not at all above about 697 °C, and with the default density λ is within 1.5 % of the full (extrapolated) IAPWS formulation. The law is not valid near the critical point (374 °C, 22 MPa), where λ¯2 diverges. The reference quantities Tref, ρref and kref are T∗, ρ∗ and λ∗; they are not material parameters.

References

source
GeoParams.MaterialParameters.Conductivity.GasMixture_Conductivity Type
julia
GasMixture_Conductivity(; kh2o=IAPWS_H2O_Conductivity(), kco2=ConstantConductivity(k=0.071W/m/K))

Thermal conductivity of a binary H2O-CO2 gas mixture after Udoetok (2013), eq. 16, the mean of the parallel (eq. 14) and series (eq. 9) arrangements of the two gases:

k=12(XH2OkH2O+XCO2kCO2+kH2OkCO2XH2OkCO2+XCO2kH2O),XH2O=1−XCO2

The CO2 mole fraction of the gas X_co2 is passed at runtime, as for the H2O-CO2 solubility laws; kh2o and kco2 are conductivity laws evaluated with the remaining arguments (e.g. T). The result is NaN outside 0≤XCO2≤1. The default kh2o depends on P and T; the default kco2 = 0.071 W/m/K is the conductivity of CO2 at 1000 K and low pressure. Over 600-1200 °C and 0.1-400 MPa the conductivity of CO2 ranges from 0.06 to 0.15 W/m/K, so the default underestimates it by up to a factor of about 2 at high pressure.

References

source
GeoParams.MaterialParameters.Conductivity.Maxwell_Conductivity Type
julia
Maxwell_Conductivity(; kmatrix=T_Conductivity_Moitra(), kgas=GasMixture_Conductivity())

Effective thermal conductivity of a matrix containing non-interacting spherical gas bubbles, the Maxwell equation used by Heap et al. (2020), eq. 1, with β=3(1−m)/(2+m) for spherical pores:

kk0=(1−ϕgas)(1−m)+mβϕgas(1−ϕgas)(1−m)+βϕgas=(1−ϕgas)(2k0+kgas)+3ϕgaskgas(1−ϕgas)(2k0+kgas)+3ϕgask0,m=kgask0

where k0 is the conductivity of the bubble-free magma (kmatrix: melt, including any crystals) and kgas that of the gas (kgas). The second form, used here, stays finite at m=1. The gas volume fraction of the whole magma ϕ_gas is passed at runtime; both sub-laws receive the remaining arguments (e.g. P, T, X_co2; the default kgas needs P and T). kgas is only evaluated when ϕ_gas != 0. The result is NaN outside 0≤ϕgas≤1.

The Maxwell equation assumes dilute, non-interacting bubbles. For kgas<k0 it is the Hashin–Shtrikman upper bound, so it overestimates k of foams with connected or deformed bubbles; Heap et al. (2020) test it up to ϕgas≈0.6. The default kmatrix inherits the melt-branch limitation of T_Conductivity_Moitra above Tg.

References

source

Conductivity of magma ​

T_Conductivity_Moitra computes k=ρcpα from the measured diffusivity of glass and melt, so it jumps at Tg like the heat capacity; T_Conductivity_Moitra_smoothed uses the same transition.

julia
using GeoParams, CairoMakie
T = range(300, 1400, 1101)
fig = Figure(size = (800, 450))
ax = Axis(fig[1, 1]; xlabel = "temperature T (K)", ylabel = "conductivity k (W/m/K)")
vspan!(ax, 1273, 1400; color = (:gray, 0.2), label = "extrapolated (> 1273 K)")
vlines!(ax, 1000; color = :black, linestyle = :dot, label = "T_g = 1000 K")
for (name, law) in ("T_Conductivity_Moitra" => T_Conductivity_Moitra(), "T_Conductivity_Moitra_smoothed (ΔTg = 20 K)" => T_Conductivity_Moitra_smoothed())
    lines!(ax, T, [compute_conductivity(law, (; T = t)) for t in T]; label = name)
end
axislegend(ax; position = :lt)
fig

The conductivity of the gas phase: IAPWS_H2O_Conductivity is solid where IAPWS R15-11 validates the formulation and dashed where it is extrapolated. GasMixture_Conductivity mixes it with the default constant CO2 conductivity.

julia
using GeoParams, CairoMakie
# IAPWS R15-11 validates the formulation for T ≤ 1173.15 K at P ≤ 100 MPa, T ≤ 874 K at P ≤ 250 MPa, T ≤ 573 K above
validated(P, T) = (P ≤ 1.0e8 && T ≤ 1173.15) || (P ≤ 2.5e8 && T ≤ 874.0) || T ≤ 573.0
kh2o = IAPWS_H2O_Conductivity()
P = exp10.(range(5, log10(4.0e8), 200))
fig = Figure(size = (1000, 420))
ax1 = Axis(fig[1, 1]; xscale = log10, xlabel = "pressure P (MPa)", ylabel = "k (W/m/K)", title = "IAPWS_H2O_Conductivity (dashed: extrapolated)")
for (i, Tc) in enumerate((600, 850, 1100))
    T = Tc + 273.15
    k = [compute_conductivity(kh2o, (; P = p, T)) for p in P]
    lines!(ax1, P ./ 1.0e6, k; color = Cycled(i), linestyle = :dash)
    lines!(ax1, P ./ 1.0e6, [validated(p, T) ? ki : NaN for (p, ki) in zip(P, k)]; color = Cycled(i), label = "$Tc °C")
end
axislegend(ax1; position = :lt)
ax2 = Axis(fig[1, 2]; xlabel = "CO2 mole fraction of the gas X_co2 (–)", ylabel = "k (W/m/K)", title = "GasMixture_Conductivity, 850 °C")
X = range(0, 1, 101)
for p in (1.0e7, 1.0e8, 3.0e8)
    lines!(ax2, X, [compute_conductivity(GasMixture_Conductivity(), (; P = p, T = 1123.15, X_co2 = x)) for x in X]; label = "$(round(Int, p / 1.0e6)) MPa")
end
axislegend(ax2; position = :rt)
fig

Maxwell_Conductivity combines a matrix with gas bubbles. With its defaults, a Moitra matrix and an H2O gas, the bubbles lower the magma conductivity roughly in proportion to the gas fraction.

julia
using GeoParams, CairoMakie
k = Maxwell_Conductivity()
ϕ_gas = range(0, 0.8, 161)
fig = Figure(size = (800, 450))
ax = Axis(fig[1, 1]; xlabel = "gas fraction ϕ_gas (–)", ylabel = "conductivity k (W/m/K)", title = "Maxwell_Conductivity: Moitra matrix with H2O bubbles at 50 MPa")
for T in (900.0, 1200.0)
    lines!(ax, ϕ_gas, [compute_conductivity(k, (; P = 5.0e7, T, ϕ_gas = g)) for g in ϕ_gas]; label = "T = $(round(Int, T)) K")
end
axislegend(ax; position = :rt)
fig

Computational routines ​

To compute, use this:

GeoParams.MaterialParameters.Conductivity.compute_conductivity Function
julia
k = compute_conductivity(s::AbstractConductivity; P, T)
k = compute_conductivity(s::AbstractConductivity, args)

Returns the thermal conductivity k for the parameterization s, evaluated at the temperature T and pressure P passed either as keyword arguments or as a NamedTuple args.

Currently available:

  • ConstantConductivity

  • T_Conductivity_Whittington

  • TP_Conductivity

  • T_Conductivity_Moitra

  • T_Conductivity_Moitra_smoothed

  • IAPWS_H2O_Conductivity

  • GasMixture_Conductivity

  • Maxwell_Conductivity

Example

julia
julia> k = T_Conductivity_Whittington();

julia> compute_conductivity(k; T = 1000.0)
1.9014576517265882
source
GeoParams.MaterialParameters.Conductivity.compute_conductivity! Function
julia
compute_conductivity(k_array::AbstractArray{<:AbstractFloat,N},P::AbstractArray{<:AbstractFloat,N},T::AbstractArray{<:AbstractFloat,N}, s::ConstantConductivity) where N

In-place routine to compute constant conductivity

source
julia
compute_conductivity!(K::AbstractArray{<:AbstractFloat}, Phases::AbstractArray{<:Integer}, P::AbstractArray{<:AbstractFloat},Temp::AbstractArray{<:AbstractFloat}, MatParam::AbstractArray{<:AbstractMaterialParamsStruct})

In-place computation of conductivity K for the whole domain and all phases, in case a vector with phase properties MatParam is provided, along with P and Temp arrays. This assumes that the Phase of every point is specified as an Integer in the Phases array.


compute_conductivity!(k::AbstractArray{T,N}, PhaseRatios::AbstractArray{T, M}, P::AbstractArray{<:AbstractFloat,N},T::AbstractArray{<:AbstractFloat,N}, MatParam::AbstractArray{<:AbstractMaterialParamsStruct})

In-place computation of conductivity k for the whole domain and all phases, in case a vector with phase properties MatParam is provided, along with P and T arrays. This assumes that the PhaseRatio of every point is specified as an Integer in the PhaseRatios array, which has one dimension more than the data arrays (and has a phase fraction between 0-1)

source