-
Notifications
You must be signed in to change notification settings - Fork 2
/
Copy pathUnitCell.jl
252 lines (218 loc) · 12.5 KB
/
UnitCell.jl
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
using LinearAlgebra, Optim
struct UnitCell{D}
primitive::NTuple{D,NTuple{D,Float64}}
rl::Matrix{Float64}
basis::Vector{NTuple{D,Float64}}
interactions::Vector{Tuple{Int,Int,NTuple{D,Int},Matrix{Float64}}} #interactions specified as (basis1,basis2,offsetPrimitive,M)
interactionsOnsite::Vector{Matrix{Float64}}
interactionsField::Vector{Vector{Float64}}
UnitCell(a1::NTuple{1,Float64}) = new{1}((a1,), Vector{NTuple{1,Float64}}(undef,0), Vector{Tuple{Int,Int,NTuple{1,Int},Matrix{Float64}}}(undef,0), Vector{Matrix{Float64}}(undef,0), Vector{Vector{Float64}}(undef,0))
UnitCell(a1::NTuple{2,Float64}, a2::NTuple{2,Float64}) = new{2}((a1,a2), Vector{NTuple{2,Float64}}(undef,0), Vector{Tuple{Int,Int,NTuple{2,Int},Matrix{Float64}}}(undef,0), Vector{Matrix{Float64}}(undef,0), Vector{Vector{Float64}}(undef,0))
UnitCell(a1::NTuple{3,Float64}, a2::NTuple{3,Float64}, a3::NTuple{3,Float64}) = new{3}((a1,a2,a3),
2*pi*inv([[a1...] [a2...] [a3...]]),
Vector{NTuple{3,Float64}}(undef,0),
Vector{Tuple{Int,Int,NTuple{3,Int},Matrix{Float64}}}(undef,0),
Vector{Matrix{Float64}}(undef,0),
Vector{Vector{Float64}}(undef,0))
UnitCell(primitives...) = new{length(primitives)}(primitives, Vector{NTuple{length(primitives),Float64}}(undef,0), Vector{Tuple{Int,Int,NTuple{length(primitives),Int},Matrix{Float64}}}(undef,0), Vector{Matrix{Float64}}(undef,0), Vector{Vector{Float64}}(undef,0))
end
"""
Adds an interaction between spin1 located at basis site `b1` of the given `unitcell` and spin2 at basis site `b2` in a unit cell that is offset by `offset` lattice vectors.
The exchange energy is calculated as spin1'.M.spin2.
"""
function addInteraction!(unitcell::UnitCell{D}, b1::Int, b2::Int, M::Matrix{Float64}, offset::NTuple{D,Int}=Tuple(zeros(Int,D))) where D
size(M) == (3,3) || error("Interaction matrix must be of size 3x3.")
b1 == b2 && offset == Tuple(zeros(Int,D)) && error("Interaction cannot be local. Use setInteractionOnsite!() instead.")
push!(unitcell.interactions, (b1,b2,offset,M))
end
function setInteractionOnsite!(unitcell::UnitCell{D}, b::Int, M::Matrix{Float64}) where D
size(M) == (3,3) || error("Interaction matrix must be of size 3x3.")
unitcell.interactionsOnsite[b] = M
end
function setField!(unitcell::UnitCell{D}, b::Int, B::Vector{Float64}) where D
size(B) == (3,) || error("Field must be a vector of length 3.")
unitcell.interactionsField[b] = B
end
function addBasisSite!(unitcell::UnitCell{D}, position::NTuple{D,Float64}) where D
push!(unitcell.basis, position)
push!(unitcell.interactionsOnsite, zeros(3,3))
push!(unitcell.interactionsField, zeros(3))
return length(unitcell.basis)
end
function getDist(unitcell::UnitCell{D}) where D
nbond = length(unitcell.interactions)
dist = zeros(nbond,3)
for ind_bond = 1:nbond
bond = unitcell.interactions[ind_bond]
pos1 = unitcell.basis[bond[1]]
pos2 = unitcell.basis[bond[2]] .+ bond[3][1].*unitcell.primitive[1] .+ bond[3][2].*unitcell.primitive[2] .+ bond[3][3].*unitcell.primitive[3]
dist[ind_bond,1] = pos2[1] .- pos1[1]
dist[ind_bond,2] = pos2[2] .- pos1[2]
dist[ind_bond,3] = pos2[3] .- pos1[3]
end
return dist
end
function getFourier_iso(unitcell::UnitCell{D}, dist::Matrix{Float64}, q_crys::Matrix{Float64}) where D
natom = length(unitcell.basis)
Jq = zeros(natom, natom, axes(q_crys,1)).*im
Threads.@threads for ind_q in axes(q_crys,1)
q_lab = zeros(3)
q_lab[1] = q_crys[ind_q,1]*unitcell.rl[1,1] + q_crys[ind_q,2]*unitcell.rl[2,1] + q_crys[ind_q,3]*unitcell.rl[3,1]
q_lab[2] = q_crys[ind_q,1]*unitcell.rl[1,2] + q_crys[ind_q,2]*unitcell.rl[2,2] + q_crys[ind_q,3]*unitcell.rl[3,2]
q_lab[3] = q_crys[ind_q,1]*unitcell.rl[1,3] + q_crys[ind_q,2]*unitcell.rl[2,3] + q_crys[ind_q,3]*unitcell.rl[3,3]
nbond = length(unitcell.interactions)
for ind_bond = 1:nbond
bond = unitcell.interactions[ind_bond]
phase = q_lab[1]*dist[ind_bond, 1] + q_lab[2]*dist[ind_bond,2] + q_lab[3]*dist[ind_bond,3]
Jterm = bond[4][1]*exp(1im*phase) # no 1/2 factor, check logbook SCGA/A_further_check
Jq[bond[1], bond[2], ind_q] += Jterm
Jq[bond[2], bond[1], ind_q] += conj(Jterm)
end
end
return Jq
end
function solveLambda_iso(unitcell::UnitCell{D}, beta::Float64) where D
n_grid::Int64 = 8
k0::Vector{Float64} = [(1-n_grid)/(2*n_grid):1/n_grid:(n_grid-1)/(2*n_grid);]
qx::Array{Float64, 3} = [x_grid for x_grid in k0, y_grid in k0, z_grid in k0]
qy::Array{Float64, 3} = [y_grid for x_grid in k0, y_grid in k0, z_grid in k0]
qz::Array{Float64, 3} = [z_grid for x_grid in k0, y_grid in k0, z_grid in k0]
q_BZ = [vec(qx) vec(qy) vec(qz)]
dist = getDist(unitcell)
Jq_BZ = getFourier_iso(unitcell, dist, q_BZ)
# eigensolution
natom = length(unitcell.basis)
eig_val = zeros(natom, size(q_BZ,1))
for ind_q in axes(q_BZ,1)
eig_val[:,ind_q] = eigvals(Jq_BZ[:,:,ind_q])
end
lam_sum_chi2(lambda) = (sum(1 ./ (lambda[1] .+ beta.*eig_val))/size(eig_val,1)/size(q_BZ,1)-1/3)^2
lam_lb = maximum(-beta.*eig_val)
lower = lam_lb
upper = Inf
x0 = [lam_lb+10]
result = optimize(lam_sum_chi2, lower, upper, x0)
lambda = Optim.minimizer(result)[1]
return lambda
end
function getCorr_iso(unitcell::UnitCell{D}, Jq_calc::Array{ComplexF64,3}, beta::Float64, lambda::Float64) where D
natom = length(unitcell.basis)
correl = zeros(axes(Jq_calc,3))
Threads.@threads for ind_q in axes(Jq_calc,3)
# correl[ind_q] = sum(inv(lambda.*Diagonal(ones(natom)) .+ beta.*Jq_calc[:,:,ind_q]));
correl[ind_q] = real(sum(inv(lambda.*Matrix{Float64}(I,natom,natom) .+ beta.*Jq_calc[:,:,ind_q])));
end
return correl
end
function getChiT_iso(unitcell::UnitCell{D}, temperatures::Vector{Float64}) where D
natom = length(unitcell.basis)
beta_list = 1 ./ (temperatures ./ 11.6045)
chiT::Vector{Float64} = zeros(axes(beta_list))
for ind = 1:length(beta_list)
beta = beta_list[ind]
dist::Matrix{Float64} = getDist(unitcell)
Jq_calc::Array{ComplexF64,3} = getFourier_iso(unitcell, dist, [0.0 0.0 0.0])
lam_answer::Float64 = solveLambda_iso(unitcell, beta)
chiT[ind] = getCorr_iso(unitcell, Jq_calc, beta, lam_answer)[1]/natom/3
end
return chiT
end
function getFourier_aniso(unitcell::UnitCell{D}, dist::Matrix{Float64}, q_crys::Matrix{Float64}) where D
natom::Int64 = length(unitcell.basis)
Jq::Array{ComplexF64, 3} = zeros(natom*3, natom*3, axes(q_crys,1))
Threads.@threads for ind_q in axes(q_crys,1)
q_lab = zeros(3)
q_lab[1] = q_crys[ind_q,1]*unitcell.rl[1,1] + q_crys[ind_q,2]*unitcell.rl[2,1] + q_crys[ind_q,3]*unitcell.rl[3,1]
q_lab[2] = q_crys[ind_q,1]*unitcell.rl[1,2] + q_crys[ind_q,2]*unitcell.rl[2,2] + q_crys[ind_q,3]*unitcell.rl[3,2]
q_lab[3] = q_crys[ind_q,1]*unitcell.rl[1,3] + q_crys[ind_q,2]*unitcell.rl[2,3] + q_crys[ind_q,3]*unitcell.rl[3,3]
nbond = length(unitcell.interactions)
for ind_bond = 1:nbond
bond = unitcell.interactions[ind_bond]
phase::Float64 = q_lab[1]*dist[ind_bond, 1] + q_lab[2]*dist[ind_bond,2] + q_lab[3]*dist[ind_bond,3]
Jterm::Matrix{ComplexF64} = bond[4]*exp(-1im*phase) # no 1/2 factor, check logbook SCGA/A_further_check
# Jq[3*(bond[1]-1)+1:3*(bond[1]-1)+3, 3*(bond[2]-1)+1:3*(bond[2]-1)+3, ind_q] += Jterm
# Jq[3*(bond[2]-1)+1:3*(bond[2]-1)+3, 3*(bond[1]-1)+1:3*(bond[1]-1)+3, ind_q] += adjoint(Jterm)
Jq[3*bond[1]-2, 3*bond[2]-2, ind_q] += Jterm[1]
Jq[3*bond[1]-1, 3*bond[2]-2, ind_q] += Jterm[2]
Jq[3*bond[1], 3*bond[2]-2, ind_q] += Jterm[3]
Jq[3*bond[1]-2, 3*bond[2]-1, ind_q] += Jterm[4]
Jq[3*bond[1]-1, 3*bond[2]-1, ind_q] += Jterm[5]
Jq[3*bond[1], 3*bond[2]-1, ind_q] += Jterm[6]
Jq[3*bond[1]-2, 3*bond[2], ind_q] += Jterm[7]
Jq[3*bond[1]-1, 3*bond[2], ind_q] += Jterm[8]
Jq[3*bond[1], 3*bond[2], ind_q] += Jterm[9]
Jq[3*bond[2]-2, 3*bond[1]-2, ind_q] += conj(Jterm[1])
Jq[3*bond[2]-1, 3*bond[1]-2, ind_q] += conj(Jterm[4])
Jq[3*bond[2], 3*bond[1]-2, ind_q] += conj(Jterm[7])
Jq[3*bond[2]-2, 3*bond[1]-1, ind_q] += conj(Jterm[2])
Jq[3*bond[2]-1, 3*bond[1]-1, ind_q] += conj(Jterm[5])
Jq[3*bond[2], 3*bond[1]-1, ind_q] += conj(Jterm[8])
Jq[3*bond[2]-2, 3*bond[1], ind_q] += conj(Jterm[3])
Jq[3*bond[2]-1, 3*bond[1], ind_q] += conj(Jterm[6])
Jq[3*bond[2], 3*bond[1], ind_q] += conj(Jterm[9])
end
# interactionsOnsite
for ind_atom = 1:natom
Jq[3*ind_atom-2, 3*ind_atom-2, ind_q] += unitcell.interactionsOnsite[ind_atom][1,1]
Jq[3*ind_atom-2, 3*ind_atom-1, ind_q] += unitcell.interactionsOnsite[ind_atom][1,2]
Jq[3*ind_atom-2, 3*ind_atom, ind_q] += unitcell.interactionsOnsite[ind_atom][1,3]
Jq[3*ind_atom-1, 3*ind_atom-2, ind_q] += unitcell.interactionsOnsite[ind_atom][2,1]
Jq[3*ind_atom-1, 3*ind_atom-1, ind_q] += unitcell.interactionsOnsite[ind_atom][2,2]
Jq[3*ind_atom-1, 3*ind_atom, ind_q] += unitcell.interactionsOnsite[ind_atom][2,3]
Jq[3*ind_atom, 3*ind_atom-2, ind_q] += unitcell.interactionsOnsite[ind_atom][3,1]
Jq[3*ind_atom, 3*ind_atom-1, ind_q] += unitcell.interactionsOnsite[ind_atom][3,2]
Jq[3*ind_atom, 3*ind_atom, ind_q] += unitcell.interactionsOnsite[ind_atom][3,3]
end
end
return Jq
end
function solveLambda_aniso(unitcell::UnitCell{D}, beta::Float64) where D
n_grid::Int64 = 8
k0::Vector{Float64} = [(1-n_grid)/(2*n_grid):1/n_grid:(n_grid-1)/(2*n_grid);]
qx::Array{Float64, 3} = [x_grid for x_grid in k0, y_grid in k0, z_grid in k0]
qy::Array{Float64, 3} = [y_grid for x_grid in k0, y_grid in k0, z_grid in k0]
qz::Array{Float64, 3} = [z_grid for x_grid in k0, y_grid in k0, z_grid in k0]
q_BZ = [vec(qx) vec(qy) vec(qz)]
dist = getDist(unitcell)
Jq_BZ = getFourier_aniso(unitcell, dist, q_BZ)
# eigensolution
natom = length(unitcell.basis)
eig_val = zeros(natom*3, size(q_BZ,1))
for ind_q in axes(q_BZ,1)
eig_val[:,ind_q] = eigvals(Jq_BZ[:,:,ind_q])
end
lam_sum_chi2(lambda) = (sum(1 ./ (lambda[1] .+ beta.*eig_val))/size(eig_val,1)/size(q_BZ,1)-1/3)^2
lam_lb = maximum(-beta.*eig_val)
lower = lam_lb
upper = Inf
x0 = [lam_lb+10]
result = optimize(lam_sum_chi2, lower, upper, x0)
lambda = Optim.minimizer(result)[1]
return lambda
end
function getCorr_aniso(unitcell::UnitCell{D}, q_crys::Matrix{Float64}, Jq_calc::Array{ComplexF64,3}, beta::Float64, lambda::Float64) where D
correl::Vector{Float64} = zeros(axes(Jq_calc,3))
Threads.@threads for ind_q in axes(q_crys,1)
q_lab = zeros(3)
q_lab[1] = q_crys[ind_q,1]*unitcell.rl[1,1] + q_crys[ind_q,2]*unitcell.rl[2,1] + q_crys[ind_q,3]*unitcell.rl[3,1]
q_lab[2] = q_crys[ind_q,1]*unitcell.rl[1,2] + q_crys[ind_q,2]*unitcell.rl[2,2] + q_crys[ind_q,3]*unitcell.rl[3,2]
q_lab[3] = q_crys[ind_q,1]*unitcell.rl[1,3] + q_crys[ind_q,2]*unitcell.rl[2,3] + q_crys[ind_q,3]*unitcell.rl[3,3]
Q_norm::Vector{Float64} = q_lab/norm(q_lab)
F = eigen(Jq_calc[:,:,ind_q])
eig_val::Vector{Float64} = F.values
U::Matrix{ComplexF64} = F.vectors
for ind_eig in axes(U,2)
Fsum = zeros(3)*im
for ind_atom in axes(unitcell.basis,1)
UQ::ComplexF64 = U[ind_atom*3-2, ind_eig]*Q_norm[1] + U[ind_atom*3-1, ind_eig]*Q_norm[2] + U[ind_atom*3, ind_eig]*Q_norm[3]
Fsum[1] += U[ind_atom*3-2, ind_eig] - UQ*Q_norm[1]
Fsum[2] += U[ind_atom*3-1, ind_eig] - UQ*Q_norm[2]
Fsum[3] += U[ind_atom*3, ind_eig] - UQ*Q_norm[3]
end
F2::Float64 = real(Fsum[1].*conj(Fsum[1]) + Fsum[2].*conj(Fsum[2]) + Fsum[3].*conj(Fsum[3]))
denom::Float64 = lambda + beta*eig_val[ind_eig]
correl[ind_q] += F2/denom
end
end
return correl
end