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
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
|
# fmt: off
import numpy as np
from ase.io.jsonio import read_json, write_json
class STM:
def __init__(self, atoms, symmetries=None, use_density=False):
"""Scanning tunneling microscope.
atoms: Atoms object or filename
Atoms to scan or name of file to read LDOS from.
symmetries: list of int
List of integers 0, 1, and/or 2 indicating which surface
symmetries have been used to reduce the number of k-points
for the DFT calculation. The three integers correspond to
the following three symmetry operations::
[-1 0] [ 1 0] [ 0 1]
[ 0 1] [ 0 -1] [ 1 0]
use_density: bool
Use the electron density instead of the LDOS.
"""
self.use_density = use_density
if isinstance(atoms, str):
with open(atoms) as fd:
self.ldos, self.bias, self.cell = read_json(fd,
always_array=False)
self.atoms = None
else:
self.atoms = atoms
self.cell = atoms.cell
self.bias = None
self.ldos = None
assert not self.cell[2, :2].any() and not self.cell[:2, 2].any()
self.symmetries = symmetries or []
def calculate_ldos(self, bias):
"""Calculate local density of states for given bias."""
if self.ldos is not None and bias == self.bias:
return
self.bias = bias
calc = self.atoms.calc
if self.use_density:
self.ldos = calc.get_pseudo_density()
return
if bias < 0:
emin = bias
emax = 0.0
else:
emin = 0
emax = bias
nbands = calc.get_number_of_bands()
weights = calc.get_k_point_weights()
nkpts = len(weights)
nspins = calc.get_number_of_spins()
eigs = np.array([[calc.get_eigenvalues(k, s)
for k in range(nkpts)]
for s in range(nspins)])
eigs -= calc.get_fermi_level()
ldos = np.zeros(calc.get_pseudo_wave_function(0, 0, 0).shape)
for s in range(nspins):
for k in range(nkpts):
for n in range(nbands):
e = eigs[s, k, n]
if emin < e < emax:
psi = calc.get_pseudo_wave_function(n, k, s)
ldos += weights[k] * (psi * np.conj(psi)).real
if 0 in self.symmetries:
# (x,y) -> (-x,y)
ldos[1:] += ldos[:0:-1].copy()
ldos[1:] *= 0.5
if 1 in self.symmetries:
# (x,y) -> (x,-y)
ldos[:, 1:] += ldos[:, :0:-1].copy()
ldos[:, 1:] *= 0.5
if 2 in self.symmetries:
# (x,y) -> (y,x)
ldos += ldos.transpose((1, 0, 2)).copy()
ldos *= 0.5
self.ldos = ldos
def write(self, filename):
"""Write local density of states to JSON file."""
write_json(filename, (self.ldos, self.bias, self.cell))
def get_averaged_current(self, bias, z):
"""Calculate avarage current at height z (in Angstrom).
Use this to get an idea of what current to use when scanning."""
self.calculate_ldos(bias)
nz = self.ldos.shape[2]
# Find grid point:
n = z / self.cell[2, 2] * nz
dn = n - np.floor(n)
n = int(n) % nz
# Average and do linear interpolation:
return ((1 - dn) * self.ldos[:, :, n].mean() +
dn * self.ldos[:, :, (n + 1) % nz].mean())
def scan(self, bias, current, z0=None, repeat=(1, 1)):
"""Constant current 2-d scan.
Returns three 2-d arrays (x, y, z) containing x-coordinates,
y-coordinates and heights. These three arrays can be passed to
matplotlibs contourf() function like this:
>>> import matplotlib.pyplot as plt
>>> plt.contourf(x, y, z)
>>> plt.show()
"""
self.calculate_ldos(bias)
L = self.cell[2, 2]
nz = self.ldos.shape[2]
h = L / nz
ldos = self.ldos.reshape((-1, nz))
heights = np.empty(ldos.shape[0])
for i, a in enumerate(ldos):
heights[i] = find_height(a, current, h, z0)
s0 = heights.shape = self.ldos.shape[:2]
heights = np.tile(heights, repeat)
s = heights.shape
ij = np.indices(s, dtype=float).reshape((2, -1)).T
x, y = np.dot(ij / s0, self.cell[:2, :2]).T.reshape((2,) + s)
return x, y, heights
def scan2(self, bias, z, repeat=(1, 1)):
"""Constant height 2-d scan.
Returns three 2-d arrays (x, y, I) containing x-coordinates,
y-coordinates and currents. These three arrays can be passed to
matplotlibs contourf() function like this:
>>> import matplotlib.pyplot as plt
>>> plt.contourf(x, y, I)
>>> plt.show()
"""
self.calculate_ldos(bias)
nz = self.ldos.shape[2]
ldos = self.ldos.reshape((-1, nz))
current = np.empty(ldos.shape[0])
zp = z / self.cell[2, 2] * nz
zp = int(zp) % nz
for i, a in enumerate(ldos):
current[i] = self.find_current(a, zp)
s0 = current.shape = self.ldos.shape[:2]
current = np.tile(current, repeat)
s = current.shape
ij = np.indices(s, dtype=float).reshape((2, -1)).T
x, y = np.dot(ij / s0, self.cell[:2, :2]).T.reshape((2,) + s)
# Returing scan with axes in Angstrom.
return x, y, current
def linescan(self, bias, current, p1, p2, npoints=50, z0=None):
"""Constant current line scan.
Example::
stm = STM(...)
z = ... # tip position
c = stm.get_averaged_current(-1.0, z)
stm.linescan(-1.0, c, (1.2, 0.0), (1.2, 3.0))
"""
heights = self.scan(bias, current, z0)[2]
p1 = np.asarray(p1, float)
p2 = np.asarray(p2, float)
d = p2 - p1
s = np.dot(d, d)**0.5
cell = self.cell[:2, :2]
shape = np.array(heights.shape, float)
M = np.linalg.inv(cell)
line = np.empty(npoints)
for i in range(npoints):
p = p1 + i * d / (npoints - 1)
q = np.dot(p, M) * shape
line[i] = interpolate(q, heights)
return np.linspace(0, s, npoints), line
def pointcurrent(self, bias, x, y, z):
"""Current for a single x, y, z position for a given bias."""
self.calculate_ldos(bias)
nx = self.ldos.shape[0]
ny = self.ldos.shape[1]
nz = self.ldos.shape[2]
# Find grid point:
xp = x / np.linalg.norm(self.cell[0]) * nx
dx = xp - np.floor(xp)
xp = int(xp) % nx
yp = y / np.linalg.norm(self.cell[1]) * ny
dy = yp - np.floor(yp)
yp = int(yp) % ny
zp = z / np.linalg.norm(self.cell[2]) * nz
dz = zp - np.floor(zp)
zp = int(zp) % nz
# 3D interpolation of the LDOS at point (x,y,z) at given bias.
xyzldos = (((1 - dx) + (1 - dy) + (1 - dz)) * self.ldos[xp, yp, zp] +
dx * self.ldos[(xp + 1) % nx, yp, zp] +
dy * self.ldos[xp, (yp + 1) % ny, zp] +
dz * self.ldos[xp, yp, (zp + 1) % nz])
return dos2current(bias, xyzldos)
def sts(self, x, y, z, bias0, bias1, biasstep):
"""Returns the dI/dV curve for position x, y at height z (in Angstrom),
for bias from bias0 to bias1 with step biasstep."""
biases = np.arange(bias0, bias1 + biasstep, biasstep)
current = np.zeros(biases.shape)
for b in np.arange(len(biases)):
print(b, biases[b])
current[b] = self.pointcurrent(biases[b], x, y, z)
dIdV = np.gradient(current, biasstep)
return biases, current, dIdV
def find_current(self, ldos, z):
""" Finds current for given LDOS at height z."""
nz = self.ldos.shape[2]
zp = z / self.cell[2, 2] * nz
dz = zp - np.floor(zp)
zp = int(zp) % nz
ldosz = (1 - dz) * ldos[zp] + dz * ldos[(zp + 1) % nz]
return dos2current(self.bias, ldosz)
def dos2current(bias, dos):
# Borrowed from gpaw/analyse/simple_stm.py:
# The connection between density n and current I
# n [e/Angstrom^3] = 0.0002 sqrt(I [nA])
# as given in Hofer et al., RevModPhys 75 (2003) 1287
return 5000. * dos**2 * (1 if bias > 0 else -1)
def interpolate(q, heights):
qi = q.astype(int)
f = q - qi
g = 1 - f
qi %= heights.shape
n0, m0 = qi
n1, m1 = (qi + 1) % heights.shape
z = (g[0] * g[1] * heights[n0, m0] +
f[0] * g[1] * heights[n1, m0] +
g[0] * f[1] * heights[n0, m1] +
f[0] * f[1] * heights[n1, m1])
return z
def find_height(ldos, current, h, z0=None):
if z0 is None:
n = len(ldos) - 2
else:
n = int(z0 / h)
while n >= 0:
if ldos[n] > current:
break
n -= 1
else:
return 0.0
c2, c1 = ldos[n:n + 2]
return (n + 1 - (current - c1) / (c2 - c1)) * h
def delta(biases, bias, width):
"""Return a delta-function centered at 'bias'"""
x = -((biases - bias) / width)**2
return np.exp(x) / (np.sqrt(np.pi) * width)
|
