Nanoscale device predictive simulations are undoubtedly vital for development of several branches of science and technology. Among approaches to the nanoscale device simulation, coherent quantum transport calculations can be pointed out. These calculations provide electric properties of given samples, namely resistance or conductance dependencies. Landauer-Büttiker formalism expresses system conductivity (specific conductance) through it is transmission coefficients [1]. Transmission coefficients are determined from solutions of Schrödinger equation, that describes system. Plane wave approximation [2] is usually used; it is based on time-independent Schrödinger equation:
$$ \hat{H}\psi_E(x) = -\frac{\hbar^2}{2m}\frac{\partial^2}{\partial x^2}\psi_E(x) + U(x)\psi_E(x,t) = E\psi_E(x). \tag{1} $$Potential U(x) is chosen constant at $x < x_{\sigma'}$ and $x > x_{\sigma}$ and has the following form
$$ U(x) = \theta(x_{\sigma'} - x)U_l + \theta(x - x_{\sigma'})\theta(x_{\sigma} - x)U_m(x) + \theta(x - x_{\sigma})U_r, \tag{2} $$where $\theta(x)$ stands for Heaviside theta function. As the potential can be chosen regardless to a constant term, so the case of $U_l = 0$ is chosen. Thus eigenfunctions $\psi_E(x)$ of equation (1) are as follows:
$$ \psi_E(x) = \theta(x_{\sigma'} - x)(1 + r(E)) \exp(i\frac{\sqrt{2mE}}{\hbar}x) + \theta(x - x_{\sigma'})\theta(x_{\sigma} - x)f(x) + \theta(x - x_{\sigma})t(E) \exp(i\frac{\sqrt{2m(E - U_r)}}{\hbar}x), \tag{3} $$where r(E) and t(E) stand for reflection and transmission amplitude for a particle with fixed energy E. Appropriate transmission coefficient is determined with respect to particle flux:
$$ T(E) = \sqrt{\frac{E}{E - U_r}} |t(E)|^2. \tag{4} $$However, this approach renders no possibility to represent particle states with wave packages. In order to do that, one has to solve time-dependent Schrödinger equation with the same potential as in (2) and initial conditions representing a wave package on the left to the system:
$$ i\hbar \frac{\partial}{\partial t} \psi(x,t) = \hat{H} \psi(x,t) = -\frac{\hbar^2}{2m} \frac{\partial^2}{\partial x^2} \psi(x,t) + U(x) \psi(x,t), \tag{5} $$$$ \psi(x,0) = \theta(x_{\sigma'} - x)\psi_0(x), \tag{6} $$and calculate the transmission coefficient as probability of a wave package to pass through the system
$$ T = \lim_{t \to \infty} \int_{x_{-}}^{\infty} |\psi(x,t)|^2 dx. \tag{7} $$However, solution of (5) is much more complicated than solution of (1). A novel approach is presented to link the non-stationary (7) and the stationary (4) transmission coefficients. Plane wave decomposition is done for initial wave package (6):
$$ \psi_0(p,0) = \frac{1}{\sqrt{2\pi}} \int_{-\infty}^{\infty} \psi(x,0) \exp(-\mathrm{i}px) dx. \tag{8} $$Final formula expresses non-stationary T with stationary T(E):
$$ T = \int_{0}^{\infty} T(E) |\psi_0(E,0)|^2 dE = \int_{-\infty}^{\infty} T(E(p)) |\psi_0(p,0)|^2 \frac{p}{m} dp, \tag{9} $$where momentum is chosen as the integration variable in the right part, replacing $E(p) = \frac{p^2}{2m}$, and, appropriately, $dE = \frac{pdp}{m}$. This formula is validated for cases of free motion and step potential and rectangular barrier potential, some other consistency checks are done.
Finally a formula was obtained that helps one to avoid solution of (5) when he only wants to determine transmission coefficients. However it still requires strict mathematical prove.