The Fredholm integral equation of the laser intensity modulation method is solved with the application of the Monte Carlo technique and a least-squares solver. The numerical procedure is tested on simulated data. In the LIMM experiment, a thin sample with electrodes on both surfaces is irradiated with a modulated laser. The light is absorbed at the irradiated electrode and the heat diffuses into the sample. Space-charge and polarization in the material respond to a non-uniform temperature distribution, and produce a pyroelectric current density j. The real j r and imaginary j i parts of the current are recorded for each modulation frequency ν. The typical frequency range is between 10 Hz and 100 kHz. Lang [2] and Mellinger [3] have recently illustrated how to reconstruct the space-charge and polarization profiles with polynomial and Tikhonov[4] regularizations, respectively. In this letter, an alternative numerical method is presented and tested.The measured current density in LIMM is expressed as a Fredholm integral equation [2],where g(z) is the unknown distribution of space charge or polarization, T(ν, z) is the solution of one-dimensional heat-conduction equation, z is the coordinate in the thickness direction, and ı = √ −1. The integral in Eq. (1) is evaluated over the samples thickness s. For a free standing film,Here, A is a constant, κ is the thermal conductivity of the sample. β is the complex thermal wave [β = (1 + ı) √ πνD −1 and D is the thermal diffusivity of the material].A numerical method based on the Monte Carlo method have been proposed by Tuncer and Gubański [6] to solve integral equations in the form of Eq. (1). The method * Electronic address: enis.tuncer@physics.org has previously been applied to extract the distribution of relaxation times from dielectric spectroscopy data [6,7] and the spectral density function (distribution) of dielectric mixtures [8]. In those problems, the distributions were probability densities of relaxation times and spectral parameters. Therefore, a constrained least-squares algorithm was implemented which yielded only positive valued outputs. In the present problem, the space-charge and the polarization distributions g can have both positive and negative magnitudes. As a result, we employ a similar numerical procedure as before [6]. However, here a linear least-squares solver is adopted without any a-priori assumptions instead of the constrained leastsquares method.The numerical procedure proposed is as follows, 1. Eq. (1) is written as a summation over some number n of z values for each experimental frequency point ν m (m = 1, . . . , M , M is the number of experimental data points),2. The z n values are preselected randomly from a linear distribution, 0 < z n < s.
Eq. (3) is written as a matrix equation,and solved for g. In Eq. (4), T is the m × n 'kernel' matrix, where T = T(ν m , z n ), g is a column vector with length n (g = g n ), and j is a column vector with length m.4. The minimization then yields the desired distribution gSince the problem is ill-conditi...