pyTEMlib 0.2025.4.2__py3-none-any.whl → 0.2025.9.1__py3-none-any.whl

pyTEMlib/kinematic_scattering.py

@@ -41,7 +41,7 @@ from pyTEMlib.diffraction_plot import *
 
 _version_ = "0.2022.1.0"
 
-print('Using kinematic_scattering library version {_version_ }  by G.Duscher')
+print(f'Using kinematic_scattering library version {_version_ }  by G.Duscher')
 
 inputKeys = ['acceleration_voltage_V', 'zone_hkl', 'Sg_max', 'hkl_max']
 optional_inputKeys = ['crystal', 'lattice_parameter_nm', 'convergence_angle_mrad', 'mistilt', 'thickness',

pyTEMlib/probe_tools.py

@@ -14,7 +14,7 @@ def make_gauss(size_x, size_y, width=1.0, x0=0.0, y0=0.0, intensity=1.0):
     size_y = size_y / 2
     x, y = np.mgrid[-size_x:size_x, -size_y:size_y]
     g = np.exp(-((x - x0) ** 2 + (y - y0) ** 2) / 2.0 / width ** 2)
-    probe = g / g.sum() * intensity
+    probe = g / (g.sum()+ 0.00001) * intensity
 
     return probe
 

pyTEMlib/test.py

@@ -0,0 +1,437 @@
+import sys
+
+from matplotlib.pyplot import plot
+sys.path.insert(0, './')
+import pyTEMlib
+print(pyTEMlib.__version__)
+
+filename = "C:\\Users\\gduscher\\Desktop\\ESL-506-15kV.esl"
+filename = "C:\\Users\\gduscher\\.pyTEMlib\\k-factors-Spectra300UTK200keV.csv"
+# pyTEMlib.eds_tools.read_esl_k_factors("C:\\Users\\gduscher\\Desktop\\ESL-506-15kV.esl")
+
+# print(pyTEMlib.eds_tools.read_csv_k_factors(filename))
+# pp =pyTEMlib.eds_tools.convert_k_factor_file(filename)
+# print('read bruker k-factors')
+#
+# print(pyTEMlib.eds_tools.read_k_factors('k_factors_Thermo_200keV.json')) 
+
+# print(pyTEMlib.eds_tools.get_k_factor_files())
+import os
+file = os.path.join(pyTEMlib.config_dir.config_path, 'Dirac_GOS.gosh')    
+import h5py
+
+"""
+element = 'Si'
+edge = 'L3'
+with h5py.File(file, 'r') as gos_file:
+    gos = gos_file[element][edge]['data'][:].squeeze().T
+    free_energies = gos_file[element][edge]['free_energies'][:][ :]  # two dimensional array
+    q_axis = gos_file[element][edge]['q'][:] # in  [1/m]
+    ionization_energy = gos_file[element][edge]['metadata'].attrs['ionization_energy']
+print (ionization_energy)
+
+import matplotlib.pylab as plt
+import scipy
+import numpy as np
+
+x = free_energies
+y = q_axis
+y = y/min(y)
+z = gos
+print(x.shape, y.shape, z.shape)
+X = np.linspace(min(x), max(x))
+Y = np.linspace(min(y), max(y))
+print(min(y), max(y))
+grid_x, grid_y = np.meshgrid(X, Y)
+
+interp = scipy.interpolate.griddata(zip(x, y), z, (grid_x, grid_y), method='linear', fill_value=0)
+print(interp.shape, print())
+
+plt.imshow(interp[:,:,0].T, extent=(min(x), max(x), min(y), max(y)), origin='lower')
+
+plt.colorbar()
+
+plt.show()
+
+
+plt.figure()
+plt.pcolormesh( q_axis/1e10,free_energies+ionization_energy, gos)
+plt.xlabel('q [1/A]')
+plt.ylabel('Energy above ionization [eV]')
+plt.colorbar()
+plt.title('GOS for ' + element + ' ' + edge)
+plt.show()
+"""
+import matplotlib.pylab as plt
+
+import scipy
+import numpy as np
+
+
+
+
+
+
+def getinterpolatedgos(E, q, E_axis, q_axis, GOSmatrix):
+    """
+    Gets the interpolated value of the GOS from the E and q value.
+    """
+    index_q = np.searchsorted(q_axis, q, side='left')
+    index_E = np.searchsorted(E_axis, E, side='left')
+
+    if index_E == 0:
+        return 0
+    if index_E == E_axis.size:
+        return 0.
+    if index_q == 0:
+        return GOSmatrix[index_E, 0]
+    if index_q == q_axis.size:
+        return 0.0
+
+    dE = E_axis[index_E] - E_axis[index_E - 1]
+    dq = q_axis[index_q] - q_axis[index_q - 1]
+
+    distE = E - E_axis[index_E - 1]
+    distq = q - q_axis[index_q - 1]
+
+    r = GOSmatrix[index_E - 1, index_q - 1] * (1 / (dE * dq)) * (dE - distE) * (dq - distq)
+    r += GOSmatrix[index_E - 1, index_q] * (1 / (dE * dq)) * (dE - distE) * (distq)
+    r += GOSmatrix[index_E, index_q - 1] * (1 / (dE * dq)) * (distE) * (dq - distq)
+    r += GOSmatrix[index_E, index_q] * (1 / (dE * dq)) * (distE) * (distq)
+    return r
+
+
+def gaussian(E, integral, x0, sigma):
+    # A = integral / (np.sqrt(2 * np.pi) * sigma)
+    g = np.exp(-0.5 * (E - x0)**2 / sigma**2)
+    g = integral * g / g.sum()
+    return g
+
+def getinterpolatedq(q, GOSarray, q_axis):
+    """
+    Gets the interpolated value of the GOS array as a function of q.
+    Usefull for the bounded states
+
+    Parameters
+    ----------
+    q: float
+        The q from the GOS should be interpolated
+    GOSarray:
+        dddd
+    q_axis: numpy array
+        The q axis on which the GOS is calculated
+
+
+    Returns
+    -------
+    interpolated GOS matrix
+
+    """
+    index_q = np.searchsorted(q_axis, q, side='left')
+
+    if index_q == 0:
+        return GOSarray[0]
+    if index_q == q_axis.size:
+        return 0.0
+
+    dq = q_axis[index_q] - q_axis[index_q - 1]
+
+    distq = q - q_axis[index_q - 1]
+
+    r0 = GOSarray[index_q - 1] * (1/dq) * (dq-distq)
+    r1 = GOSarray[index_q] * (1/dq) * (distq)
+
+    return r0 + r1
+
+def correction_factor_kohl(alpha, beta, theta, min_alpha=1e-6):
+    """
+    STILL NEEDS TO BE VALIDATED
+    Calculates the correction factor when using a convergent
+    probe. For probes having is convergence angle smaller than
+    min_alpha no correction is applied.
+    Ultramicroscopy 16 (1985) 265-268:
+    https://doi.org/10.1016/0304-3991(85)90081-6
+     Parameters
+     ----------
+     alpha : float
+         Convergence angle in radians
+     beta : float
+         Collection angle in radians
+     theta : float
+         The angle for which the correction factor should be calculated
+     min_alpha : float
+        Minimum convergence angle for which the correction is applied
+
+     Returns
+     -------
+     corr_factor : float
+        correction factor used in the integration
+    """
+    if alpha < min_alpha:
+        corr_factor = 1.
+    elif theta <= np.abs(alpha - beta):
+        min_thetasq = min(alpha**2, beta**2)
+        corr_factor = min_thetasq / alpha**2
+    else:
+        x = (alpha**2 + theta**2 - beta**2) / (2. * alpha * theta)
+        y = (beta**2 + theta**2 - alpha**2) / (2. * beta * theta)
+        wortel = np.sqrt(4 * alpha**2 * beta**2 - (alpha**2 + beta**2 - theta**2)**2)
+        corr_factor = (1 / np.pi) * (np.arccos(x) + (beta**2 / alpha**2 * np.arccos(y)) - (
+                                      1 / (2 * alpha**2) * wortel))
+    return corr_factor
+
+def get_dirac_X_section(element = 'Si', edge = 'L3', file = file,
+                        q_steps=100, E0=200000, beta=0.100, alpha=0):
+    """ Calculates the cross section from the GOS matrix"""
+
+    
+    with h5py.File(file, 'r') as gos_file:
+        GOSmatrix = gos_file[element][edge]['data'][:].squeeze().T
+        free_energies = gos_file[element][edge]['free_energies'][:][ :]  # two dimensional array
+        q_axis = gos_file[element][edge]['q'][:] # in  [1/m]
+        ek = gos_file[element][edge]['metadata'].attrs['ionization_energy']
+
+    energy_axis = np.linspace(50, 850, int(800/5))
+    shell_occupancy = 1
+    pref = 1e28 * shell_occupancy
+    
+    e = scipy.constants.e
+    c = scipy.constants.c
+    m = scipy.constants.electron_mass
+    a_0 = scipy.constants.physical_constants['Bohr radius'][0]
+    gamma = 1 + e * E0 / (m * c ** 2)
+
+    effective_incident_energy = E0 * (1 + gamma) / (2 * gamma**2)
+    T = effective_incident_energy
+    R = scipy.constants.Rydberg
+
+    bool0 = free_energies < 0
+    Ebound = free_energies[bool0] + ek
+
+    dsigma_dE = np.zeros(energy_axis.shape)
+    dsigma_dE_bound = np.zeros(energy_axis.shape)
+    sigma = 2*(energy_axis[1] - energy_axis[0])
+    rel_energy_axis = free_energies + ek
+
+    for i in range(Ebound.size):
+        E = Ebound[i]
+        integral = 0
+        # the bounded states are differently interpolated
+        qa0sq_min = E ** 2 / (4 * R * T) + (E ** 3) / (8 * gamma ** 3 * R * T ** 2)
+        qa0sq_max = qa0sq_min + 4 * gamma ** 2 * (T / R) * (np.sin((beta + alpha) / 2)) ** 2
+        logqa0sq_axis = np.linspace(np.log(qa0sq_min), np.log(qa0sq_max),
+                                    q_steps)
+        lnqa0sqstep = (logqa0sq_axis[1] - logqa0sq_axis[0])
+        print(i, logqa0sq_axis, lnqa0sqstep)
+        for j in range(logqa0sq_axis.size):
+            q = np.sqrt(np.exp(logqa0sq_axis[j])) / scipy.constants.physical_constants['Bohr radius'][0]
+            theta = 2. * np.sqrt(np.abs( R * (np.exp(logqa0sq_axis[j]) - qa0sq_min) / 
+                                        (4. * gamma**2 * T)))
+            GOSarray = GOSmatrix[i, :]
+            df_dE = getinterpolatedq(q, GOSarray, q_axis)
+
+            # integral+= df_dE*lnqa0sqstep
+            integral += df_dE * lnqa0sqstep * correction_factor_kohl(alpha, beta, theta)
+
+        sig = 4 * np.pi * a_0 ** 2 * (R / E) * (R / T) * integral
+        dsigma_dE_bound += gaussian(energy_axis, sig, E, sigma)
+
+    
+    # the for loop over the bound states
+    for i in range(energy_axis.size):
+        E = energy_axis[i]
+        integral = 0
+        if (E > ek) & (E <= rel_energy_axis[-1]):
+            qa0sq_min = E**2 / (4 * R * T) + (E**3) / (8 * gamma**3 * R * T**2)
+            qa0sq_max = qa0sq_min + 4 * gamma**2 * (T / R) * (np.sin((beta + alpha) / 2))**2
+            logqa0sq_axis = np.linspace(np.log(qa0sq_min), np.log(qa0sq_max), q_steps)
+            lnqa0sqstep = (logqa0sq_axis[1] - logqa0sq_axis[0])
+            for j in range(logqa0sq_axis.size):
+                q = np.sqrt(np.exp(logqa0sq_axis[j])) / a_0
+                theta = 2. * np.sqrt(np.abs(
+                    R * (np.exp(logqa0sq_axis[j]) - qa0sq_min) / (
+                            4. * gamma ** 2 * T)))
+                df_dE = getinterpolatedgos(E, q, rel_energy_axis, q_axis, GOSmatrix)
+                # integral+= df_dE*lnqa0sqstep
+                integral += df_dE * lnqa0sqstep * correction_factor_kohl(alpha, beta, theta)
+            # dsigma_dE[i] = 4*np.pi*pc.a0()**2*(R/E)*(R/T)*integral*dispersion
+            dsigma_dE[i] = 4 * np.pi * a_0 ** 2 * (R / E) * (R / T) * integral
+        else:
+            dsigma_dE[i] = 0
+
+    cross_section =  dsigma_dE + dsigma_dE_bound * pref
+
+    return cross_section
+
+def energy2wavelength(e0: float) -> float:
+    """get deBroglie wavelength of electron accelerated by energy (in eV) e0"""
+    ev = scipy.constants.e * e0
+    m_e = scipy.constants.m_e
+    c = scipy.constants.c
+    h = scipy.constants.h
+    return h / np.sqrt(2 * m_e * ev * (1 + ev / (2 * m_e * c**2)))*1e10
+
+
+def ddscs_dE_dOmega(free_energies, ek, E0, q_axis, GOSmatrix):
+    """scattering cross section as a function of energy loss and solid angle
+
+    Args:
+        free_energies: 1d numpy array
+            The energy axis on which the GOS table is calculated without the onset
+            energy [eV]
+        ek: float
+            The onset energy of the calculated edge [eV]
+        E0: float
+            The acceleration voltage of the incoming electrons [V]
+        q_axis: 1d numpy array
+            The momentum on which the GOS table are calculated. [kg m /s]?
+        GOSmatrix: 2d numpy array
+            The GOS
+
+    Returns:
+        np.array: scattering cross section as a function of energy loss and solid angle
+    """
+    R = scipy.constants.Rydberg
+    e = scipy.constants.e
+    e = scipy.constants.e
+    c = scipy.constants.c
+    m = scipy.constants.electron_mass
+    a_0 = scipy.constants.physical_constants['Bohr radius'][0]
+    gamma = 1 + e * E0 / (m * c ** 2)
+    energy_losses = free_energies + ek
+    
+    k0 = 2 * np.pi / energy2wavelength(E0)
+
+    scs_list = []
+    for idx, epsilon in enumerate(free_energies):
+        kn = 2 * np.pi / energy2wavelength(E0-energy_losses[idx])
+        scs = (
+            4
+            * gamma ** 2
+            / q_axis**2
+            * kn
+            / k0
+            * GOSmatrix[idx]
+            / energy_losses[idx]
+            * R
+        )
+        scs_list.append(scs)
+    scs_list = np.array(scs_list).squeeze()
+
+    return scs_list
+
+def plot_ddscs(element = 'Si', edge = 'L3', file = file,
+                        q_steps=100, E0=200000, beta=0.100, alpha=0):
+    with h5py.File(file, 'r') as gos_file:
+        GOSmatrix = gos_file[element][edge]['data'][:].squeeze().T
+        free_energies = gos_file[element][edge]['free_energies'][:][ :]  # two dimensional array
+        q_axis = gos_file[element][edge]['q'][:] # in  [1/m]
+        ek = gos_file[element][edge]['metadata'].attrs['ionization_energy']
+        for k in gos_file[element][edge]['metadata'].attrs.keys():
+            print(f"{k} => {gos_file[element][edge]['metadata'].attrs[k]}")
+        occupancy = gos_file[element][edge]['metadata'].attrs['occupancy_ratio']
+    k0 = 2 * np.pi / energy2wavelength(E0)
+    ddscs = ddscs_dE_dOmega(free_energies, ek, E0, q_axis/1e10, GOSmatrix)
+    plt.subplots(1, 2, figsize=(12, 5))
+    plt.subplot(121)
+    plt.pcolormesh(q_axis/1e10, free_energies, ddscs)
+    plt.ylabel('Energy loss [eV]')
+    plt.xlabel('Scattering vector [1/A]')
+    plt.colorbar()
+    plt.tight_layout()
+    plt.title('Double differential scattering cross section')
+    plt.subplot(122)
+    theta = np.arctan(q_axis/1e10 /k0)*1e3 # this is approximation
+    plt.pcolormesh(theta, free_energies, ddscs) 
+    plt.xlim([0, 50])
+    plt.xlabel('Scattering angle [mrad]')
+    plt.ylabel('Energy loss [eV]')
+    plt.title('Double differential scattering cross section')
+    plt.colorbar()  
+    plt.tight_layout()
+    max_q = np.searchsorted(theta, 30)
+    max_e = np.searchsorted(free_energies, 1000)
+    max_q_value = np.tan(30e-3)*k0*1e10
+    print(f"min_q: {q_axis[0]}, end_q: {q_axis[-1]}, max_q_value: {max_q_value}")
+    print(q_axis[max_q])
+    y_sparse = theta[:max_q]
+    x_sparse = free_energies[:max_e]
+    z_sparse = ddscs[:max_e, :max_q]
+    xnew = np.linspace(0, 1000, 1000)
+    ynew = np.linspace(0, 30, 100)
+
+    Xnew, Ynew = np.meshgrid(xnew, ynew)
+
+    # Flatten the input data for griddata
+    X_sparse, Y_sparse = np.meshgrid(x_sparse, y_sparse, indexing='ij')
+    points = np.column_stack([X_sparse.ravel(), Y_sparse.ravel()])
+    values = z_sparse.ravel()
+    znew_reggrid = scipy.interpolate.griddata(points, values, (Xnew, Ynew), method='linear', fill_value=0)
+    plt.figure()
+    plt.imshow(znew_reggrid, extent=(0, 400, 0, 30), origin='lower', aspect='auto')
+
+    plt.colorbar()
+
+    print(ddscs.shape, q_axis.shape, free_energies.shape, znew_reggrid.shape)
+
+    plt.figure()
+    #plt.plot(free_energies, ddscs[:, :max_q].sum(axis=1),)
+    plt.plot(xnew, znew_reggrid.sum(axis=0),)
+    plt.show()
+
+def plot_ddscs(element = 'Si', edge = 'L3', file = file, 
+               energy_scale = np.linspace(50, 850, int(800*3)),
+               q_steps=100, E0=200000, beta=0.100, alpha=0):
+    with h5py.File(file, 'r') as gos_file:
+        GOSmatrix = gos_file[element][edge]['data'][:].squeeze().T
+        free_energies = gos_file[element][edge]['free_energies'][:][ :]  # two dimensional array
+        q_axis = gos_file[element][edge]['q'][:] # in  [1/m]
+        ek = gos_file[element][edge]['metadata'].attrs['ionization_energy']
+        for k in gos_file[element][edge]['metadata'].attrs.keys():
+            print(f"{k} => {gos_file[element][edge]['metadata'].attrs[k]}")
+        occupancy = gos_file[element][edge]['metadata'].attrs['occupancy_ratio']
+    
+    
+    k0 = 2 * np.pi / energy2wavelength(E0)
+    ddscs = ddscs_dE_dOmega(free_energies, ek, E0, q_axis/1e10, GOSmatrix)
+    theta = np.arctan(q_axis/1e10 /k0)*1e3 # this is an approximation
+    
+    max_q = np.searchsorted(theta, beta*1e3)+1
+    max_e = np.searchsorted(free_energies+ek, energy_scale[-1])+1
+    max_q_value = np.tan(beta)*k0
+    print(f"min_q: {q_axis[0]/1e10}, end_q: {q_axis[-1]/1e10}, max_q_value: {max_q_value}")
+    print(q_axis[max_q])
+    y_sparse = q_axis[:max_q]/1e10
+    x_sparse = free_energies[:max_e]+ek
+    z_sparse = ddscs[:max_e, :max_q]
+    xnew = energy_scale
+    ynew = np.linspace(0, max_q_value, q_steps)
+
+    Xnew, Ynew = np.meshgrid(xnew, ynew)
+
+    # Flatten the input data for griddata
+    X_sparse, Y_sparse = np.meshgrid(x_sparse, y_sparse, indexing='ij')
+    points = np.column_stack([X_sparse.ravel(), Y_sparse.ravel()])
+    values = z_sparse.ravel()
+    znew_reggrid = scipy.interpolate.griddata(points, values, (Xnew, Ynew), method='linear', fill_value=0)
+    plt.figure()
+    plt.imshow(znew_reggrid, extent=(xnew[0], xnew[-1], 0, ynew[-1]), origin='lower', aspect='auto')
+
+    plt.colorbar()
+    delta_q = ynew[1]-ynew[0]
+
+    print(ddscs.shape, q_axis.shape, free_energies.shape, znew_reggrid.shape)
+
+    plt.figure()
+    #plt.plot(free_energies, ddscs[:, :max_q].sum(axis=1),)
+    plt.plot(xnew, znew_reggrid.sum(axis=0)*delta_q,)
+    plt.show()
+
+plot_ddscs(element='Si', edge='L3', beta=0.030, alpha=0.0)
+plt.show()
+#xsec = get_dirac_X_section(element='Si', edge='M2', beta=0.0001, alpha=0.0)
+#print('calculated cross section')
+#plt.figure()
+#plt.plot(np.linspace(50, 850, int(800/5)), xsec)
+#plt.xlabel('Energy loss [eV]')
+#plt.show()