Supporting Information
Coupling the High-Throughput Property Map to Machine Learning for Predicting
Lattice Thermal Conductivity
Rinkle Juneja
†,‡
, George Yumnam
†,‡
, Swanti Satsangi
†
, and Abhishek K. Singh
†
‡
Equal contribution
†
Materials Research Centre, Indian Institute of Science, Bangalore 560012, India
Lattice thermal conductivity, phonon dispersion, and Gr¨ uneisen parameter of 116 compounds
For all the screened compounds, lattice thermal conductivity, phonon dispersion, and Gr¨ uneisen parameter are calculated. The 2 ultralow and 2
ultrahigh κ
l
compounds are discussed in the main manuscript. The remaining 116 are presented here. Figures S1 - S5 show the lattice thermal
conductivity, Figures S6 - S10 show the phonon dispersion, and Figures S11 - S15 show the Gr¨ uneisen parameter, respectively.
GaP GeC InP MgS MgS
CaTe CdS CdSe CdTe GaAs
BeTe BiB CaS CaSe CaSe
BaS BaSe BaTe BeS BeSe
AlAs AlP AlSb BP BSb (F-43m) (F-43m) (F-43m) (F-43m) (F-43m)
(Fm-3m) (Fm-3m) (Fm-3m) (F-43m) (F-43m)
(F-43m) (F-43m) (Fm-3m) (F-43m) (Fm-3m)
(Fm-3m) (F-43m) (F-43m) (F-43m) (F-43m)
(F-43m) (F-43m) (F-43m) (F-43m) (Fm-3m)
500 1000 100 500 1000 100 500 1000 100 500 1000 100 500 1000 100
500 1000 100 500 1000 100 500 1000 100 500 1000 100 500 1000 100
500 1000 100 500 1000 100 500 1000 100 500 1000 100 500 1000 100
500 1000 100 500 1000 100 500 1000 100 500 1000 100 500 1000 100
500 1000 100 500 1000 100 500 1000 100 500 1000 100 500 1000 100
100
200
300
10
20
250
500
750
10
20
30
250
500
10
20
50
100
250
500
1000
2000
250
500
0
10
15
5
25
50
75
25
50
75
100
200
100
200
300
10
20
10
20
30
50
100
0
2500
1000
0
25
50
50
100
150
25
50
500
1000
1500
500
1000
Figure S1: Calculated DFT lattice thermal conductivity as a function of temperature for the nonmetallic and dynamically stable compounds.
S-1