File: RoughMultiLayerContribution.cpp

package info (click to toggle)
bornagain 23.0-4
  • links: PTS, VCS
  • area: main
  • in suites: forky, sid
  • size: 103,936 kB
  • sloc: cpp: 423,131; python: 40,997; javascript: 11,167; awk: 630; sh: 318; ruby: 173; xml: 130; makefile: 51; ansic: 24
file content (178 lines) | stat: -rw-r--r-- 8,130 bytes parent folder | download
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
//  ************************************************************************************************
//
//  BornAgain: simulate and fit reflection and scattering
//
//! @file      Sim/Computation/RoughMultiLayerContribution.cpp
//! @brief     Implements class RoughMultiLayerContribution.
//!
//! @homepage  http://www.bornagainproject.org
//! @license   GNU General Public License v3 or higher (see COPYING)
//! @copyright Forschungszentrum Jülich GmbH 2018
//! @authors   Scientific Computing Group at MLZ (see CITATION, AUTHORS)
//
//  ************************************************************************************************

#include "Sim/Computation/RoughMultiLayerContribution.h"
#include "Base/Util/Assert.h"
#include "Resample/Element/DiffuseElement.h"
#include "Resample/Flux/ScalarFlux.h"
#include "Resample/Processed/ReSample.h"
#include "Sample/Interface/Roughness.h"
#include "Sample/Multilayer/Layer.h"
#include "Sample/Multilayer/Sample.h"
#include <cerf.h>
#include <numbers>

using std::numbers::pi;

// As we say in our 2020 paper (Sect 5.6), diffuse scattering from rough interfaces
// is modelled after Schlomka et al, Phys Rev B, 51, 2311 (1995). They give credit
// for the basic modelling idea to Sinha et al (1988), and for the matrix elements
// to Holy et al (1993), Holy and Baumbach (1994), Sinha (1994) and de Boer (unpublished).
//
// Specifically, we implemented the differential cross section according to lines 2-3
// in Eq 3 of Schlomka et al. Unfortunately, this equation has an incorrect prefactor k^2,
// and so had our implementation up to release 20.0. The correct prefactor is k^4, as in
// Eq 17 of Holy et al (1993). This was corrected in release 20.1 (issue #553).

namespace {

complex_t h_above(complex_t z)
{
    return 0.5 * cerfcx(-mul_I(z) / std::sqrt(2.0));
}
complex_t h_below(complex_t z)
{
    return 0.5 * cerfcx(mul_I(z) / std::sqrt(2.0));
}

complex_t get_refractive_term(const ReSample& re_sample, size_t i_layer, double wavelength)
{
    return re_sample.avgeSlice(i_layer).material().refractiveIndex2(wavelength)
           - re_sample.avgeSlice(i_layer + 1).material().refractiveIndex2(wavelength);
}

complex_t get_sum8terms(const ReSample& re_sample, size_t i_layer, const DiffuseElement& ele)
{
    // Abbreviations:
    //   i/f : initial/final beam
    //   A/B : above/below the interface
    const auto* i_A = dynamic_cast<const ScalarFlux*>(ele.fluxIn(i_layer));
    const auto* f_A = dynamic_cast<const ScalarFlux*>(ele.fluxOut(i_layer));
    const auto* i_B = dynamic_cast<const ScalarFlux*>(ele.fluxIn(i_layer + 1));
    const auto* f_B = dynamic_cast<const ScalarFlux*>(ele.fluxOut(i_layer + 1));
    if (!(i_A && f_A && i_B && f_B))
        throw std::runtime_error(
            "Rough interfaces not yet supported for polarized simulation (issue #564)");

    const complex_t kiz_A = i_A->getScalarKz();
    const complex_t kfz_A = f_A->getScalarKz();
    const complex_t qz1_A = -kiz_A - kfz_A;
    const complex_t qz2_A = -kiz_A + kfz_A;
    const complex_t qz3_A = -qz2_A;
    const complex_t qz4_A = -qz1_A;

    const double thickness = re_sample.avgeSlice(i_layer).thicknessOr0();
    const complex_t T_i_A = i_A->getScalarT() * exp_I(kiz_A * thickness);
    const complex_t R_i_A = i_A->getScalarR() * exp_I(-kiz_A * thickness);
    const complex_t T_f_A = f_A->getScalarT() * exp_I(kfz_A * thickness);
    const complex_t R_f_A = f_A->getScalarR() * exp_I(-kfz_A * thickness);

    const complex_t kiz_B = i_B->getScalarKz();
    const complex_t kfz_B = f_B->getScalarKz();
    const complex_t qz1_B = -kiz_B - kfz_B;
    const complex_t qz2_B = -kiz_B + kfz_B;
    const complex_t qz3_B = -qz2_B;
    const complex_t qz4_B = -qz1_B;

    const double rms = re_sample.averageSlices().bottomRMS(i_layer);
    const complex_t term1 = T_i_A * T_f_A * ::h_above(qz1_A * rms);
    const complex_t term2 = T_i_A * R_f_A * ::h_above(qz2_A * rms);
    const complex_t term3 = R_i_A * T_f_A * ::h_above(qz3_A * rms);
    const complex_t term4 = R_i_A * R_f_A * ::h_above(qz4_A * rms);
    const complex_t term5 = i_B->getScalarT() * f_B->getScalarT() * ::h_below(qz1_B * rms);
    const complex_t term6 = i_B->getScalarT() * f_B->getScalarR() * ::h_below(qz2_B * rms);
    const complex_t term7 = i_B->getScalarR() * f_B->getScalarT() * ::h_below(qz3_B * rms);
    const complex_t term8 = i_B->getScalarR() * f_B->getScalarR() * ::h_below(qz4_B * rms);

    return term1 + term2 + term3 + term4 + term5 + term6 + term7 + term8;
}

} // namespace

double Compute::roughMultiLayerContribution(const ReSample& re_sample, const DiffuseElement& ele)
{
    if (ele.alphaMean() < 0.0)
        return 0;

    const size_t n_slices = re_sample.numberOfSlices();
    const double spatial_f = ele.meanQ().magxy() / (2 * pi);
    const double wavelength = ele.wavelength();
    double autocorr_sum = 0;
    double crosscorr_sum = 0;

    std::vector<Slice> roughStack; // only slices with rough top interfaces
    std::vector<complex_t> rterm;
    std::vector<complex_t> sterm;

    roughStack.reserve(n_slices - 1);
    rterm.reserve(n_slices - 1);
    sterm.reserve(n_slices - 1);

    for (size_t i = 0; i < n_slices - 1; i++)
        if (re_sample.avgeSlice(i + 1).topRoughness()) {
            roughStack.push_back(re_sample.avgeSlice(i + 1));
            rterm.push_back(::get_refractive_term(re_sample, i, wavelength));
            sterm.push_back(::get_sum8terms(re_sample, i, ele));
        }

    const size_t n_interfaces = roughStack.size();
    std::vector<double> spectrum(n_interfaces, 0);
    std::vector<double> crosscorr_with_interface_below(n_interfaces, 0); // the last is zero

    // precompute autocorrelation terms
    for (int i = n_interfaces - 1; i >= 0; i--) {
        auto* autocorr_model = roughStack[i].topRoughness()->autocorrelationModel();
        if (auto* k_corr = dynamic_cast<const SelfAffineFractalModel*>(autocorr_model)) {
            spectrum[i] = k_corr->spectralFunction(spatial_f);
        } else if (auto* lin_growth = dynamic_cast<const LinearGrowthModel*>(autocorr_model)) {
            ASSERT(i < int(n_interfaces - 1));
            const double thickness = roughStack[i].hig() - roughStack[i + 1].hig();
            spectrum[i] = lin_growth->spectralFunction(spectrum[i + 1], thickness, spatial_f);
        }
    }

    // auto correlation in each layer (first term in final expression in Eq (3) of Schlomka et al)
    for (size_t i = 0; i < n_interfaces; i++)
        autocorr_sum += std::norm(rterm[i] * sterm[i]) * spectrum[i];

    // precompute crosscorrelation terms
    for (int j = n_interfaces - 2; j >= 0; j--) {
        const double thickness = roughStack[j].hig() - roughStack[j + 1].hig();
        if (auto* spat_freq_cc = dynamic_cast<const SpatialFrequencyCrosscorrelation*>(
                roughStack[j].topRoughness()->crosscorrelationModel())) {
            crosscorr_with_interface_below[j] =
                spat_freq_cc->crosscorrSpectrum(spectrum[j], spectrum[j + 1], thickness, spatial_f);
        } else if (auto* lin_growth_autocorr = dynamic_cast<const LinearGrowthModel*>(
                       roughStack[j].topRoughness()->autocorrelationModel())) {
            crosscorr_with_interface_below[j] =
                lin_growth_autocorr->crosscorrSpectrum(spectrum[j + 1], thickness, spatial_f);
        }
    }

    // cross correlation between layers (second term in loc. cit.)
    for (size_t j = 0; j < n_interfaces - 1; j++) {
        if (crosscorr_with_interface_below[j] == 0)
            continue;
        double crosscorr_spectrum = crosscorr_with_interface_below[j];
        for (size_t k = j + 1; k < n_interfaces; k++) {
            crosscorr_sum +=
                (rterm[j] * sterm[j] * std::conj(rterm[k] * sterm[k])).real() * crosscorr_spectrum;
            if (crosscorr_with_interface_below[k] == 0 || spectrum[k] == 0)
                break;
            crosscorr_spectrum *= crosscorr_with_interface_below[k] / spectrum[k];
        }
    }
    const double k0 = (2 * pi) / wavelength;
    return pow(k0, 4) / 16 / pi / pi * (autocorr_sum + 2 * crosscorr_sum);
}