Notes: • Next exam scheduled for April 13; if no conflicts can move to Friday (15th) same as before. • I have a meeting trip May 2&3 (last 2 class days). As noted before, the plan will be to start 25 minutes early on Fridays (starting April 1). Recall, lattice modes: LO ω Optical: 3m-3 branches; ω ≠ 0 at k = 0. TO LA Acoustic – always 3 (in 3D); ω → 0 as k → 0. TA m = number in basis k Phonon modes H = ∑k ,ν !ωkν ( 1 2 ) ( + akt ν akν ⇒ E = ∑ ω k,ν nk,ν + 1 2 k,ν § Each classical oscillator mode amplitude nkv = 0 to ∞ , “number of phonons present” § Raising and lowering operators add/remove a phonon. ) Equilibrium phonon occupation ( E = ∑ ω k,ν nk,ν + 1 2 k,ν ) nkν = 1,2,3 … = quantum number for each of 3N modes. Sum over all modes for energy average: − β ∑ (nk,ν +1/2)ω k,ν ∂ with Z = ∑ e k,ν E =− ln Z ∂β Same as, {nkν } ⎛ ∞ − β ( n +1/ 2 ) !ωk ,ν Z = ∏⎜∑e k ,ν ⎝ n = 0 − β!ω ⎡1 e k ,ν E = ∑ !ωk ,ν ⎢ + − β!ω k ,ν 2 1 − e k ,ν ⎣ β = 1 kT All sets of 3N quantum numbers 1 − β!ω k ,ν 2 e ⎞ = ⎟ ∏ − β!ω k ,ν 1 − e ⎠ k ,ν ⎤ ⎡1 ⎤ = ! ω + n ⎥ ∑ k ,ν ⎢ kν ⎥ 2 ⎣ ⎦ ⎦ k ,ν Mean occupation number at equilibrium nk ,ν = 1 e β!ω k ,ν −1 Equilibrium phonon occupation 1 ⎡1 ⎤ ⎡1 ⎤ E = ∑ !ωk ,ν ⎢ + β!ωk ,ν ⎥ = ∑ !ωk ,ν ⎢ + nkν ⎥ − 1⎦ k ,ν ⎣2 e ⎣2 ⎦ k ,ν nk ,ν = 1 e β!ω k ,ν −1 • Planck formula, sum over all normal modes • Phonons act as Bose particles; photon statistics • Note, photons maximum frequency unlimited, unlike phonons kT << !ω : E ≈ !ωk ,ν e kT >> !ω : E ≈ kT − β!ω k ,ν Low-T occupation vanishingly small, except for lowest acoustic modes. (Note omitted zero-point term here.) for each mode: Classical result. Equilibrium phonon occupation ⎡1 ⎤ E = ∑ !ωk ,ν ⎢ + nkν ⎥ ⎣2 ⎦ k ,ν nk ,ν = 1 e β!ω k ,ν −1 kT << !ω : E ≈ !ωk ,ν e kT >> !ω : E ≈ kT − β!ω k ,ν Low-T occupation vanishingly small, except for lowest acoustic modes. (Note omitted zero-point term here.) for each mode: Classical result. Dulong-Petit law (required by correspondence principle). 3N total modes: CV = 3Nk B / V Specific heat & phonon density of modes: ⎡1 ⎤ E = ∑ !ωk ,ν ⎢ + nkν ⎥ ⎣2 ⎦ k ,ν nk ,ν = 1 e β!ω k ,ν −1 General relation: g (ω ) = 1 (2π ) 3 ∫ω shell @ d 2k ∇ kω E = V ∫ g (ω )!ωk ,ν (12 + nkν )dω g (ω ) ∝ ω 2 Low T, Debye method works E = 3Nk BT High T Phonon Density of Modes: • density of k-points = V/(2π)3 • per point, 3Nm branches • For quantities depending only on energy use density of modes = g(ω) Saw this before g (ω ) = 1 (2π ) 3 ∫ω shell @ d 2k ∇ kω Same argument as for electrons Or: ⎡ 1 ⎤ × ( k − space volume d ⎢ ⎥ (2π ) 3 g (ω ) = ⎥ dω ⎢ converted to ω units ) ⎥⎦ ⎢⎣ Note, g(ω) is same as density of quantized modes (phonons) Debye approximation: • Assume ω = kc “average sound velocity” represents all modes. • Assume 3 branches, with cut off sphere containing # k-points = 3N. Result, Debye wavevector kD = 3 6π 2 n ω D = c3 6π 2 n Θ D = !ω D / k B = (!c / k B )3 6π 2 n 3 ω2 g (ω ) = 2 3 c 2π Debye temperature Einstein approximation: • Assumes that ω = ωo fixed value for all modes. Specific Heat (Debye model): E = V ∫ g (ω )!ωk ,ν (12 + nkν )dω 3 ω2 g (ω ) = 2 c3 2π Debye model ∂ C= ∂T ωD ∫ 0 3ω ω dω 3k T = 2 3 β ω 2π c e −1 2π c kT >> !ω : kT << !ω : 2 4 3 B 2 3 3 CV = 3Nk B / V ΘD ∫ 0 T 4 3 x dx e T = 9kB n 3 x 2 (e −1) ΘD ΘD ∫ T 0 x 4 dx e x (e x −1)2 classical limit, easy to show. ∞ x 4 dxe x 4π 4 1 / T → ∞; ∫ x = 2 (e − 1) 15 0 Standard notation x Debye function Can look up, solve numerically, etc. C = (12 / 5)π 4 nk BT 3 / Θ3D ≡ βT 3 phonons C ≡ βT + γT 3 (Exact low-temperature limit for both contributions) electrons Debye approximation: ▪ Common measure of phonon behavior ▪ But note g(ω) can be calculated even more precisely (e.g. using accurate electron codes) when needed. from “The Specific Heat of Matter at Low Temperatures” [Tari, 2003]. X Zheng et al. Phys. Rev. B 85, 214304 (2012) [my lab]: Einstein Anomalous Specific heat of thermoelectric crystal. Anharmonic lattice effects: ► Real potential not precisely quadratic; effects can normally be treated as perturbation. Results: • Thermal expansion occurs as higher vibrational states/numbers become populated. • Vibrational states not stationary states; finite lifetime. • Leads to phonon-phonon scattering: energy and crystal momentum conserved.