Expand this Topic clickable element to expand a topic
Skip to content
Optica Publishing Group

Nonparabolicity of size-quantized subbands of bilayer semiconductor quantum wells with heterojunction

Open Access Open Access

Abstract

This paper presents a theory of size quantization and intersubband optical transitions in bilayer semiconductor quantum wells with asymmetric profile. We show that, in contrast to single-layer quantum wells, the size-quantized subbands of bilayer quantum wells are nonparabolic and characterized by effective masses that depend on the electron wave number and the subband number. It is found that the effective masses are related to the localization of the electron wave function in the layers of the quantum well and can be controlled by varying the chemical composition or geometric parameters of the structure. We also derive an analytical expression for the probability of optical transitions between the subbands of the bilayer quantum well. Our results are useful for the development of laser systems and photodetectors based on colloidal nanoplates and epitaxial layers of semiconductor materials with heterojunctions.

© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement

1. Introduction

Since the first demonstration of quantum-size effects in semiconductor quantum wells in 1974 [1] to the present day colloidal and epitaxial two-dimensional quantum systems have been attracting great interest of researchers, constantly opening new horizons for their use in devices of unique functionalities [27]. In recent years, special attention has been paid to intersubband transitions in semiconductor quantum wells, which underlie the operation of infrared detectors and quantum cascade lasers [710]. Quantum wells with asymmetric profiles, in particular step quantum wells, are especially appealing for such applications [1115]. The development of quantum-well devices with superior performance heavily relies on the accurate description of their electronic structure and the ability to control their optical properties.

In this paper, we analyse the size-quantized subbands of a bilayer step quantum well with infinitely high potential barriers and obtain an analytical expression for the probability of intersubband transitions in the conduction band of the quantum well. The analysis employs the standard formalism of the solid-state physics and can be readily extended to investigate the energy structure and hole dynamics in the valence bands of bilayer quantum wells. Our results may prove useful in the development of new devices based on epitaxial quantum wells and colloidal bilayer nanoplates.

2. Theoretical framework

Consider the conduction band $\Gamma _6$ of a bilayer step quantum well in the two-band approximation (taking into account two zones degenerated by spin) [16]. We shall assume that the electron affinity of the first-layer material (A) is greater than that of the second-layer material (B), and that the work functions of both materials are the same. In the approximation of infinitely high potential barriers, the motion of the conduction electrons inside the quantum well is determined by the potential shown in Fig. 1. The growth direction of the structure coincides with the $z$ axis while the origin coincides with the outer surface of layer A.

 figure: Fig. 1.

Fig. 1. Confining potential for conduction electrons inside a bilayer quantum well; $E_{\Gamma _6,\mathrm {A}}^0$, $E_{\Gamma _6,\mathrm {B}}^0$, $L_{\mathrm {A}}$, and $L_{\mathrm {B}}$ are the lowest energies in the conduction bands of bulk materials and the layer thicknesses of materials A and B; $L=L_{\mathrm {A}}+L_{\mathrm {B}}$.

Download Full Size | PDF

According to the Bloch theorem, the electron wave function $\Psi _{\Gamma _6,s_z}(\mathbf {r})$ in the conduction band $\Gamma _6$ of the quantum well is the product of the envelope function $f_{\Gamma _6}(\mathbf {r})$ and the Bloch factor $u_{\Gamma _6,s_z}(\mathbf {r})$, where $s_z=\pm 1/2$ is the projection of the electron spin. Within the framework of the $\mathbf {k}\cdot \mathbf {p}$-perturbation theory, we characterize the motion of an electron in each zone by an effective mass calculated using the Bloch amplitudes. Since the Bloch amplitudes and the effective masses of electrons in materials A and B are different, the eigenvalue problem for calculating the subband energies $E_{\Gamma _6}$ and the envelope function $f_{\Gamma _6}(\mathbf {r})$ inside the quantum well $0<z<L$ can be written in the form

$${\hat{H}}_{\mathrm{c}}f_{\Gamma_6}(\mathbf{r})=E_{\Gamma_6}f_{\Gamma_6}(\mathbf{r}),\quad {\hat{H}}_{\mathrm{c}}=\begin{cases}E_{\Gamma_6,\mathrm{A}}^0+{\hbar^2{\hat{k}}^2}/{(2m_{\mathrm{c,A}}^\ast)}, & 0\le z\le L_A,\\ E_{\Gamma_6,\mathrm{B}}^0+{\hbar^2{\hat{k}}^2}/{(2m_{\mathrm{c,B}}^\ast)}, & L_A\le z\le L, \end{cases}$$
where $\hat {\mathbf {k}}=-i\nabla$ is the wave vector operator, $m_{\mathrm {c,A}}^\ast$ and $m_{\mathrm {c,B}}^\ast$ are the effective masses of the conduction electrons in materials A and B, and the rest of parameters are defined in Fig. 1.

The fact that the electrons can move freely in the $xy$ plane makes possible the following factorization of the envelope function:

$$f_{\Gamma_6}(\mathbf{r})=\frac{1}{\sqrt S}e^{i\mathbf{q}\mathbf{r}_\parallel}f_{\Gamma_6}(z),$$
where $\mathbf {r}_\parallel =(x,y)$ is the radius vector of an electron in the plane of the quantum well, $\mathbf {q}=(q_x,q_y)$ is the two-dimensional wave vector of the electron, and $S$ is the normalization area. This factorization results in the ordinary differential equation for $f_{\Gamma _6}(z)$:
$$\frac{\mathrm{d}^2f_{\Gamma_6}(z)}{\mathrm{d}z^2}+q_z^2(z)f_{\Gamma_6}(z)=0,\quad q_z^2(z)=\begin{cases}q_{\mathrm{A}}^2, & 0\le z\le L_{\mathrm{A}},\\ q_{\mathrm{B}}^2, & L_{\mathrm{A}}\le z\le L, \end{cases}$$
where
$$q_\mathrm{A}^2=\frac{2m_\mathrm{c,A}^\ast}{\hbar^2}\left(E_{\Gamma_6}-E_{\Gamma_6,\mathrm{A}}^0-\frac{\hbar^2q^2}{2m_\mathrm{c,A}^\ast}\right),\quad q_\mathrm{B}^2=\frac{2m_\mathrm{c,B}^\ast}{\hbar^2}\left(E_{\Gamma_6}-E_{\Gamma_6,\mathrm{B}}^0-\frac{\hbar^2q^2}{2m_\mathrm{c,B}^\ast}\right).$$
It should be noted that $q_{\mathrm {A}}$ and $q_{\mathrm {B}}$ can be either real or purely imaginary.

The general solution of the above equation inside the quantum well $(0<z<L)$ is given by

$$f_{\Gamma_6}(z)=\begin{cases}A_\mathrm{A}e^{iq_\mathrm{A}z}+B_\mathrm{A}e^{{-}iq_\mathrm{A}z}\equiv f_{\Gamma_6}^{\mathrm{A}}(z), & 0\le z\le L_\mathrm{A},\\ A_\mathrm{B}e^{iq_\mathrm{B}z}+B_\mathrm{B}e^{{-}iq_\mathrm{B}z}\equiv f_{\Gamma_6}^{\mathrm{B}}(z), & L_\mathrm{A}\le z\le L. \end{cases}$$
Instead of the continuity of the full wave functions and their first derivatives at the interface of layers A and B, we use the Bastard boundary conditions. The first condition gives $f_{\Gamma _6}^{\mathrm {A}}(L_{\mathrm {A}})=f_{\Gamma _6}^{\mathrm {B}}(L_{\mathrm {A}})$ whereas the second can be represented by the equation [16]
$$\frac{1}{m_\mathrm{c,A}^\ast}\left.\frac{\mathrm{d}f_{\Gamma_6}^{\mathrm{A}}(z)}{\mathrm{d}z}\right|_{z=L_\mathrm{A}}=\frac{1}{m_\mathrm{c,B}^\ast}\left.\frac{\mathrm{d}f_{\Gamma_6}^{\mathrm{B}}(z)}{\mathrm{d}z}\right|_{z=L_\mathrm{A}}.$$
In addition, the wave functions must vanish at the left and right boundaries of the structure, $f_{\Gamma _6}^{\mathrm {A}}(0)=f_{\Gamma _6}^{\mathrm {B}}(L)=0$. These four boundary conditions lead to the following dispersion relation:
$$\frac{q_\mathrm{B}}{m_\mathrm{c,B}^\ast}\tan(q_\mathrm{A}L_\mathrm{A})={-}\frac{q_\mathrm{A}}{m_\mathrm{c,A}^\ast}\tan(q_\mathrm{B}L_\mathrm{B}).$$
The solution to this equation is a series of size-quantized subbands of energies $E_{\Gamma _6,n}(q)$, where $n=1,\ 2,\ 3,\ldots$ is the subband number. The electrons of subband $n$ are described by the envelope function
$$f_{\Gamma_6,n}(q;z)=2iA_{\mathrm{A},n}\begin{cases}\sin(q_{\mathrm{A},n}z), & 0\le z\le L_\mathrm{A},\\ b_n\sin[q_{\mathrm{B},n}(L-z)], & L_\mathrm{A}\le z\le L, \end{cases}$$
where the normalization constant is given by
$$A_{\mathrm{A},n}=\mathrm{sign}(q_{\mathrm{A},n})\left[2L_\mathrm{A}\left(1-\frac{\sin(2q_{\mathrm{A},n}L_\mathrm{A})}{2q_{\mathrm{A},n}L_\mathrm{A}}\right)+2L_\mathrm{B}b_n^2\left(1-\frac{\sin(2q_{\mathrm{B},n}L_\mathrm{B})}{2q_{\mathrm{B},n}L_\mathrm{B}}\right)\right]^{{-}1/2},$$
$\mathrm {sign}(x)=x/|x|$ is the complex sign function, and $b_n=\sin (q_{\mathrm {A},n}L_{\mathrm {A}})/\sin (q_{\mathrm {B},n}L_{\mathrm {B}})$. The full wave function of the electron in the $n$th subband of the conduction band is then given by
$$\Psi_{\Gamma_6,s_z;n,\mathbf{q}}(\mathbf{r})=\frac{1}{\sqrt S}e^{i\mathbf{q}\mathbf{r}_\parallel}f_{\Gamma_6,n}(q;z)u_{\Gamma_6,s_z}(\mathbf{r}).$$
We next consider intersubband optical transitions whose rates are determined by the matrix element
$$M_{f\!i}=\left\langle\Psi_{\Gamma_6,\alpha_f\!;n_f\!,\mathbf{q}_f\!}\middle|{\hat{H}}_\mathrm{eR}\middle|\Psi_{\Gamma_6,\alpha_i;n_i,\mathbf{q}_i}\right\rangle,\quad{\hat{H}}_\mathrm{eR}=\frac{e}{mc}\mathbf{A}\cdot\hat{\mathbf{p}},$$
where $\mathbf {A}$ is the vector potential of the electromagnetic field in the Coulomb gauge $(\nabla \cdot \mathbf {A}=0)$, $\hat {\mathbf {p}}=\hbar \hat {\mathbf {k}}$ is the momentum operator, and $e$ and $m$ are the charge and mass of a free electron. It should be noted that the vector potential is different in the two layers of the quantum well.

In the dipole approximation, the matrix element has the form [17]

$$M_{f\!i}=\frac{e}{mc}\int_{0}^{L}{\mathrm{d}z\ \mathbf{A}(z)\cdot\int_{S}{\mathrm{d}\mathbf{r}_\parallel\ \Psi_{\Gamma_6,\alpha_f\!;n_f\!,\mathbf{q}_f\!}^\ast(\mathbf{r})\hat{\mathbf{p}}\Psi_{\Gamma_6,\alpha_i;n_i,\mathbf{q}_i}(\mathbf{r})}}.$$
This interaction cannot change the spin of the electron; therefore, only matrix elements with $\alpha _f\!=\alpha _i$ are nonzero. Using the symmetry considerations one can show that in the case of intersubband transitions the momentum operator $\hat {\mathbf {p}}$ acts only on the envelope wave functions. Consequently, we get
$$M_{f\!i}=\delta_{\alpha_f\!,\alpha_i}\frac{e}{mcS}\int_0^L\mathrm{d}z\ U_{f\!i}(z)f_{\Gamma_6,n_f\!}^\ast(q_f\!;z)[\mathbf{A}_\parallel(z)\cdot\hbar\mathbf{q}_i+A_z(z){\hat{p}}_z]f_{\Gamma_6,n_i}(q_i;z),$$
where $\delta _{ij}$ is the Kronecker delta, $\mathbf {A}_\parallel =(A_x,A_y)$, and
$$U_{f\!i}(z)=\int_S\mathrm{d}\mathbf{r}_\parallel\ u_{\Gamma_6,\alpha_f\!}^\ast(\mathbf{r})u_{\Gamma_6,\alpha_i}(\mathbf{r})e^{i(\mathbf{q}_i-\mathbf{q}_f\!)\mathbf{r}_\parallel}.$$
In order to calculate $M_{f\!i}$, we split the integral from 0 to $L$ into two parts: from 0 to $L_{\mathrm {A}}$ (over layer A) and from $L_{\mathrm {A}}$ to $L$ (over layer B), so that $M_{f\!i}=M_{f\!i}^{\mathrm {A}}+M_{f\!i}^{\mathrm {B}}$. We then apply the procedure described in Anselm’s book [18] to quantities $M_{f\!i}^{\mathrm {A}}$ and $M_{f\!i}^{\mathrm {B}}$. Consider first $M_{f\!i}^{\mathrm {A}}$ and replace the integral over the entire volume by the sum of integrals over the unit cells. Let the position of the electron be given by the radius vector $\mathbf {r}=\mathbf {b}_{\mathbf {m}}+\mathbf {r}^\prime$, where $\mathbf {b}_{\mathbf {m}}$ is the radius vector of unit cell $\mathbf {m}$ and $\mathbf {r}^\prime$ is the radius vector of the electron inside this cell. Vector $\mathbf {b}_{\mathbf {m}}$ can be decomposed into vector $\mathbf {b}_{\mathbf {m}}^\parallel$ in the $xy$ plane and vector $\mathbf {b}_{\mathbf {m}}^\bot$ perpendicular to this plane. Using the periodicity of the Bloch amplitudes and the fact that the envelope wave functions vary slowly over the unit cell, we obtain
$$M_{f\!i}^{\mathrm{A}}\approx\delta_{\alpha_f\!,\alpha_i}\frac{e\hbar}{mcS}\sum_{\mathbf{m}}{J_{f\!i}^{\mathrm{A}}(b_\mathbf{m}^\bot)e^{i(\mathbf{q}_i-\mathbf{q}_f\!)\mathbf{b}_\mathbf{m}^\parallel}}\int_{\Omega_{\mathrm{A}}}\mathrm{d}\mathbf{r}^\prime|u_{\Gamma_6,\alpha_i}^{\mathrm{A}}(\mathbf{r}^\prime)|^2e^{i(\mathbf{q}_i-\mathbf{q}_f\!)\mathbf{r}_\parallel^\prime},$$
where
$$J_{f\!i}^{\mathrm{A}}(\zeta)=f_{\Gamma_6,n_f\!}^\ast(q_f\!;\zeta)\left((\mathbf{A}_\parallel^{\mathrm{A}}\cdot\mathbf{q}_i)f_{\Gamma_6,n_i}(q_i;\zeta)-iA_z^{\mathrm{A}}\left.\frac{\mathrm{d}f_{\Gamma_6,n_i}(q_i;z)}{\mathrm{d}z}\right|_{z=\zeta}\right)$$
and $\mathbf {A}^{\mathrm {A}}=(\mathbf {A}_\parallel ^{\mathrm {A}},A_z^{\mathrm {A}})$ and $\Omega _{\mathrm {A}}$ are the vector potential and the volume of the unit cell in layer A. By going from summation over the lattice nodes to integration, we find
$$M_{f\!i}^{\mathrm{A}}\approx\delta_{\alpha_f\!,\alpha_i}\delta_{\mathbf{q}_f\!,\mathbf{q}_i}\frac{e\hbar}{mc}\int_{0}^{L_{\mathrm{A}}}J_{f\!i}^{\mathrm{A}}(z)\ \mathrm{d}z,$$
where we have considered that the Bloch amplitudes are normalized as $\int _{\Omega _{\mathrm {A}}}{\mathrm {d}\mathbf {r}^\prime |u_{\Gamma _6,\alpha _i}^{\mathrm {A}}(\mathbf {r}^\prime )|^2}=\Omega _{\mathrm {A}}$.

After a similar calculation of $M_{fi}^{\mathrm {B}}$ we find the total intersubband matrix element to be given by

$$M_{f\!i}\approx\delta_{\alpha_f\!,\alpha_i}\delta_{\mathbf{q}_f\!,\mathbf{q}_i}\frac{e\hbar}{mc}\left(\int_{0}^{L_\mathrm{A}}J_{f\!i}^{\mathrm{A}}(z)\ \mathrm{d}z+\int_{L_\mathrm{A}}^{L}J_{f\!i}^\mathrm{B}(z)\ \mathrm{d}z\right),$$
where $J_{f\!i}^{\mathrm {B}}(z)$ is obtained from Eq. (16) by replacing the vector potential in layer A ($\mathbf {A}^{\mathrm {A}}$) with the vector potential in layer B ($\mathbf {A}^{\mathrm {B}}$). This result shows that in contrast to bulk materials, where a ‘third body’ (e.g., phonons, free electrons or impurity centers) must be involved in the emission and absorption processes due to the conservation of the wave vector, direct intraband transitions between different size-quantized subbands of the conduction band are allowed in quantum wells.

When the boundary condition $\mathbf {A}_\parallel ^{\mathrm {A}}=\mathbf {A}_\parallel ^{\mathrm {B}}\equiv \mathbf {A}_\parallel$ is satisfied, integration over $z$ leads to the factor $\delta _{n_f\!,n_i}$ for the term that is proportional to $\mathbf {A}_\parallel$. This term vanishes for intersubband transitions with $n_f\!\neq n_i$. Therefore, $M_{f\!i}$ takes the form

$$M_{f\!i}=\delta_{\alpha_f\!,\alpha_i}\delta_{\mathbf{q}_f\!,\mathbf{q}_i}M_{n_f\!,n_i}^{(0)}(q_i),\quad M_{n_f\!,n_i}^{(0)}(q)={-}i\frac{e\hbar}{mc}\left(A_z^\mathrm{A}\Theta_{f\!i}^\mathrm{A}+A_z^\mathrm{B}\Theta_{f\!i}^\mathrm{B}\right),$$

where

$$\Theta_{f\!i}^\mathrm{A}=\int_{0}^{L_\mathrm{A}}{\mathrm{d}\zeta\ f_{\Gamma_6,n_f\!}^\ast(q;\zeta)\left.\frac{\mathrm{d}f_{\Gamma_6,n_i}(q;z)}{\mathrm{d}z}\right|_{z=\zeta}},$$
$$\Theta_{f\!i}^\mathrm{B}=\int_{L_\mathrm{A}}^{L}{\mathrm{d}\zeta\ f_{\Gamma_6,n_f\!}^\ast(q;\zeta)\left.\frac{\mathrm{d}f_{\Gamma_6,n_i}(q;z)}{\mathrm{d}z}\right|_{z=\zeta}}.$$
The evaluation of these integrals yields:
$$\Theta_{f\!i}^\mathrm{A}=A_{\mathrm{A},n_f}^\ast A_{\mathrm{A},n_i}\frac{4q_{\mathrm{A},n_i}}{q_{\mathrm{A},n_f\!}^2-q_{\mathrm{A},n_i}^2} [q_{\mathrm{A},n_f\!}^\ast(1-c_{\mathrm{A},f\!i})-q_{\mathrm{A},n_i}s_{\mathrm{A},f\!i}],$$
$$\Theta_{f\!i}^\mathrm{B}={-}A_{\mathrm{A},n_f}^\ast A_{\mathrm{A},n_i}\frac{4q_{\mathrm{B},n_i}}{q_{\mathrm{B},n_f\!}^2-q_{\mathrm{B},n_i}^2} [q_{\mathrm{B},n_f\!}^\ast(1-c_{\mathrm{B},f\!i})b_{n_f\!}^\ast b_{n_i}-q_{\mathrm{B},n_i}s_{\mathrm{A},f\!i}],$$
where $c_{\mathrm {A},f\!i}=\cos (q_{\mathrm {A},n_f\!}L_{\mathrm {A}})\cos (q_{\mathrm {A},n_i}L_{\mathrm {A}})$, $c_{\mathrm {B},f\!i}=\cos (q_{\mathrm {B},n_f\!}L_{\mathrm {B}})\cos (q_{\mathrm {B},n_i}L_{\mathrm {B}})$, and $s_{\mathrm {A},f\!i}=\sin (q_{\mathrm {A},n_f\!}^\ast L_{\mathrm {A}}) \times \sin (q_{\mathrm {A},n_i}L_{\mathrm {A}})$.

Finally, we calculate the rate of direct intersubband transitions per unit area of the quantum well. Using the golden rule of quantum mechanics, this rate can be represented in the form

$$W=\frac{2}{\hbar}\sum_{n_f\!,n_i}\int_{0}^{\infty}{\left|M_{n_f\!,n_i}^{(0)}(q)\right|^2\left[F_{n_i}(q)-F_{n_f\!}(q)\right]\delta\left[E_{\Gamma_6,n_f\!}(q)-E_{\Gamma_6,n_i}(q)-\hbar\omega\right]q\mathrm{d}q},$$
where $F_n(q)$ is the Fermi–Dirac distribution, $\delta (x)$ is the Dirac delta function, and we have replaced summation over $q$ by integration.

3. Results and discussion

It is easy to show that the above expression for $M_{f\!i}$ is reduced to the well-known result for a single-layer quantum well. This result is recovered by setting $L_{\mathrm {B}}=0$,

$$\left. M_{n_f\!,n_i}^{(0)}\right|_{L_{\mathrm{B}}=0}={-}i\frac{e\hbar}{mc}\frac{A_z}{L}\frac{{2n}_fn_i}{n_f^2\!-n_i^2}[1-({-}1)^{n_f\!+n_i}],$$
and coincides with the standard textbook expression [19]. It shows that only direct intersubband transitions between the states of different parity (i.e., with an odd sum $n_f\!+n_i$) are allowed in a single-layer quantum well. Note that Eq. (19) does not impose a similar restriction on the intersubband transitions inside a two-layer quantum well.

We illustrate our results using the Al$_{0.35}$Ga$_{0.65}$As/GaAs heterojunction as an example. Figure 2(a) shows the energies of the first four subbands of a bilayer Al$_{0.35}$Ga$_{0.65}$As/GaAs quantum well plotted as functions of $q^2$. The subbands are seen to be generally nonparabolic, yet with parabolicity in certain areas — where $E_{\Gamma _6,n}(q^2)$ is approximately linear. Two dashed lines $E_{\Gamma _6}(q^2)$ on the figure correspond to $q_{\mathrm {A}}=0$ and $q_{\mathrm {B}}=0$. The slopes of these lines are determined by the effective masses $m_{\mathrm {c,A}}^\ast$ and $m_{\mathrm {c,B}}^\ast$. It is easy to see that $q_{\mathrm {A}}$ is imaginary below line $q_{\mathrm {A}}=0$ and is real above it, and similar for $q_{\mathrm {B}}$. It should be also noted that for all the energies of the subbands and any $q$

$$E_{\Gamma_6}>\min{\left(E_{\Gamma_{6},\mathrm{A}}^0+\frac{\hbar^2q^2}{2m_\mathrm{c,A}^\ast},E_{\Gamma_{6},\mathrm{B}}^0+\frac{\hbar^2q^2}{2m_\mathrm{c,B}^\ast}\right)},$$
which implies that $q_A$ and $q_B$ cannot be imaginary simultaneously.

 figure: Fig. 2.

Fig. 2. (a) The first four subbands (colour curves labelled by $n=1,\ 2,\ 3,\ \textrm {and}\ 4$) of the conduction band of a bilayer quantum well made from GaAs (layer A) and Al$_{0.35}$Ga$_{0.65}$As (layer B); $L_{\mathrm {A}}=L_{\mathrm {B}}=10\ \textrm {nm}$; dashed are the lines where $q_{\mathrm {A}}$ and $q_{\mathrm {B}}$ vanish. (b) Effective masses of electrons in the four subbands $n=1,\ 2,\ 3,\ \textrm {and}\ 4$ as functions of wave number of electrons; dashed are the effective masses of bulk GaAs $(m_{\mathrm {c,A}}^\ast )$ and bulk Al$_{0.35}$Ga$_{0.65}$As $(m_{\mathrm {c,B}}^\ast )$. All the material parameters are taken from Ref. [20].

Download Full Size | PDF

We next analyse how the effective masses of electrons, defined as $m_{\mathrm {c,n}}^\ast (q)=\hbar ^2/E_{\Gamma _6}^{\prime \prime }(q)$ (here the double prime denotes the second derivative with respect to $q$), vary as functions of wave number. The effective masses for the first four subbands of the bilayer Al$_{0.35}$Ga$_{0.65}$As/GaAs quantum well are shown in Fig. 2(b). For small $q$ the following condition is satisfied for the first two subbands:

$$E_{\Gamma_6,\mathrm{A}}^0+\frac{\hbar^2q^2}{2m_{\mathrm{c,A}}^\ast}<E_{\Gamma_6}<E_{\Gamma_6,\mathrm{B}}^0+\frac{\hbar^2q^2}{2m_{\mathrm{c,B}}^\ast}.$$
Therefore, only $q_B$ is imaginary, the envelope function of the electron decays exponentially in layer B (see Fig. 3), and the electron is localized in layer A. As can be seen from Fig. 2(b), the effective masses in this case are close to $m_{\mathrm {c,A}}^\ast$.

 figure: Fig. 3.

Fig. 3. Probability $\left |f_{\Gamma _6,n}(0;z)\right |^2$ of finding an electron with $q=0$ inside a bilayer Al$_{0.35}$Ga$_{0.65}$As/GaAs quantum well for the first four subbands $(n=1,\ 2,\ 3,\ \textrm {and}\ 4)$. All the parameters are the same as in Fig. 2.

Download Full Size | PDF

The opposite situation occurs at sufficiently large $q$ for which

$$E_{\Gamma_6,\mathrm{B}}^0+\frac{\hbar^2q^2}{2m_{\mathrm{c,B}}^\ast}<E_{\Gamma_6}<E_{\Gamma_6,\mathrm{A}}^0+\frac{\hbar^2q^2}{2m_{\mathrm{c,A}}^\ast},$$
so that $q_{\mathrm {A}}$ turns imaginary and the electron is more and more localized in layer B, as evidenced by the effective mass of electrons approaching $m_{\mathrm {c,B}}^\ast$.

The inequality

$$E_{\Gamma_6}>\max{\left(E_{\Gamma_6,\mathrm{A}}^0+\frac{\hbar^2q^2}{2m_{\mathrm{c,A}}^\ast},E_{\Gamma_6,\mathrm{B}}^0+\frac{\hbar^2q^2}{2m_{\mathrm{c,B}}^\ast}\right)},$$
which signifies that $q_{\mathrm {A}}$ and $q_{\mathrm {B}}$ are real, holds even for small $q$ for the third, fourth and, obviously, higher subbands not shown in the figure. Accordingly, an electron can be localized in the entire structure (see Fig. 3). The effective masses in the respective domains oscillate and are not limited to the values between $m_{\mathrm {c,A}}^\ast$ and $m_{\mathrm {c,B}}^\ast$.

For all the subbands the function $m_{\mathrm {c,n}}^\ast (q^2)$ exhibits sharp peaks at points where the dispersion curves intersect with the straight lines $q_{\mathrm {A}}=0$ and $q_{\mathrm {B}}=0$. They correspond first to transitions to a state in which the electron is localized in the entire structure $(q_{\mathrm {B}}=0)$, and then to transitions to a state in which it is localized in layer B $(q_{\mathrm {A}}=0)$. For subbands whose energies at $q=0$ are greater than $\max (E_{\Gamma _6,\mathrm {A}}^0,E_{\Gamma _6,\mathrm {B}}^0)$, only the equation $q_{\mathrm {A}}=0$ can be satisfied and $q_{\mathrm {B}}$ is always real.

4. Conclusions

We have developed a theory of size quantization and intersubband optical transitions in a bilayer semiconductor quantum well with asymmetric profile. It was shown that the spatial confinement of electrons inside the quantum well leads to the appearance of size-quantized subbands, which are essentially nonparabolic due to the asymmetric profile of the well. The effective masses of electrons, defined through the second derivative of the energy with respect to the wave number, were found to depend not only on the wave number, but also on the number of the subband and be closely related to the conditions of the electron localization in the layers of the quantum well. We also derived analytical expressions for the matrix elements of optical transitions between the subbands and for the transition rates, which are needed for the design of new quantum-well-based optoelectronic devices. We believe that our results will benefit the development and implementation of such devices.

Funding

Russian Science Foundation (19-13-00332).

Disclosures

The authors declare no conflicts of interest.

References

1. R. Dingle, W. Wiegmann, and C. H. Henry, “Quantum states of confined carriers in very thin AlxGa1−xAs-GaAs-AlxGa1−xAs heterostructures,” Phys. Rev. Lett. 33(14), 827–830 (1974). [CrossRef]  

2. N. V. Tepliakov, I. O. Ponomareva, M. Y. Leonov, A. V. Baranov, A. V. Fedorov, and I. D. Rukhlenko, “Field-induced broadening of electroabsorption spectra of semiconductor nanorods and nanoplatelets,” J. Phys. Chem. C 120(4), 2379–2385 (2016). [CrossRef]  

3. N. V. Tepliakov, A. S. Baimuratov, I. A. Vovk, M. Y. Leonov, A. V. Baranov, A. V. Fedorov, and I. D. Rukhlenko, “Chiral optical properties of tapered semiconductor nanoscrolls,” ACS Nano 11(7), 7508–7515 (2017). [CrossRef]  

4. D. Liu, J. Jiang, Y. Cheng, and J. He, “Effect of delta doping on mid-infrared intersubband absorption in AlGaN/GaN step quantum well structures,” Phys. E (Amsterdam, Neth.) 54, 253–256 (2013). [CrossRef]  

5. N. V. Tepliakov, I. A. Vovk, A. S. Baimuratov, M. Y. Leonov, A. V. Baranov, A. V. Fedorov, and I. D. Rukhlenko, “Optical activity of semiconductor gammadions beyond planar chirality,” J. Phys. Chem. Lett. 9(11), 2941–2945 (2018). [CrossRef]  

6. I. A. Vovk, N. Tepliakov, A. S. Baimuratov, M. Y. Leonov, A. V. Baranov, A. V. Fedorov, and I. D. Rukhlenko, “Excitonic phenomena in perovskite quantum-dot supercrystals,” Phys. Chem. Chem. Phys. 20(38), 25023–25030 (2018). [CrossRef]  

7. M. P. Touse, G. Karunasiri, K. R. Lantz, H. Li, and T. Mei, “Near- and mid-infrared detection using GaAs/InxGa1−xAs/InyGa1−yAs multiple step quantum wells,” Appl. Phys. Lett. 86(9), 093501 (2005). [CrossRef]  

8. J. Faist, F. Capasso, D. L. Sivco, C. Sirtori, A. L. Hutchinson, and A. Y. Cho, “Quantum cascade laser,” Science 264(5158), 553–556 (1994). [CrossRef]  

9. B. T. Diroll, M. Chen, I. Coropceanu, K. R. Williams, D. V. Talapin, P. Guyot-Sionnest, and R. D. Schaller, “Polarized near-infrared intersubband absorptions in CdSe colloidal quantum wells,” Nat. Commun. 10(1), 4511 (2019). [CrossRef]  

10. D. Liu and C. He, “Theoretical study of optical absorption in nonpolar AlGaN/GaN step quantum well structures,” J. Comput. Electron. 18(1), 251–259 (2019). [CrossRef]  

11. D. Liu, Y. Cheng, and J. He, “Nanogroove control of intersubband absorption in AlGaAs/GaAs step quantum wells,” Semicond. Sci. Technol. 31(3), 035006 (2016). [CrossRef]  

12. H. X. Wang, Z. L. Fu, D. X. Shao, Z. Z. Zhang, C. Wang, Z. Y. Tan, X. G. Guo, and J. C. Cao, “Broadband bias-tunable terahertz photodetector using asymmetric GaAs/AlGaAs step multi-quantum well,” Appl. Phys. Lett. 113(17), 171107 (2018). [CrossRef]  

13. H. Dakhlaoui, “The effects of doping layer location on the electronic and optical properties of GaN step quantum well,” Superlattices Microstruct. 97, 439–447 (2016). [CrossRef]  

14. E. Ozturk, “Linear and total intersubband transitions in the step-like GaAs/GaAlAs asymmetric quantum well as dependent on intense laser field,” Eur. Phys. J. Plus 130(11), 237 (2015). [CrossRef]  

15. E. A. de Andrada e Silva, G. C. La Rocca, and F. Bassani, “Spin-orbit splitting of electronic states in semiconductor asymmetric quantum wells,” Phys. Rev. B 55(24), 16293–16299 (1997). [CrossRef]  

16. E. L. Ivchenko, Optical Spectroscopy of Semiconductor Nanostructures (Alpha Science Int’l Ltd., 2005).

17. P. Y. Yu and M. Cardona, Fundamentals of Semiconductors: Physics and Materials Properties (Springer, 1996).

18. A. I. Anselm, Introduction to the Theory of Semiconductors (Nauka, 1978).

19. E. Vorobiev, L. E. Golub, L. E. Danilov, and S. N. Ivchenko, Optical Phenomena in Semiconductor Quantum-dimensional Structures (SPbGTU, 2000).

20. The Ioffe Institute, “NSM Archive - Aluminium Gallium Arsenide (AlGaAs),” http://www.ioffe.ru/SVA/NSM/Semicond/AlGaAs.

Cited By

Optica participates in Crossref's Cited-By Linking service. Citing articles from Optica Publishing Group journals and other participating publishers are listed here.

Alert me when this article is cited.


Figures (3)

Fig. 1.
Fig. 1. Confining potential for conduction electrons inside a bilayer quantum well; $E_{\Gamma _6,\mathrm {A}}^0$ , $E_{\Gamma _6,\mathrm {B}}^0$ , $L_{\mathrm {A}}$ , and $L_{\mathrm {B}}$ are the lowest energies in the conduction bands of bulk materials and the layer thicknesses of materials A and B; $L=L_{\mathrm {A}}+L_{\mathrm {B}}$ .
Fig. 2.
Fig. 2. (a) The first four subbands (colour curves labelled by $n=1,\ 2,\ 3,\ \textrm {and}\ 4$ ) of the conduction band of a bilayer quantum well made from GaAs (layer A) and Al $_{0.35}$ Ga $_{0.65}$ As (layer B); $L_{\mathrm {A}}=L_{\mathrm {B}}=10\ \textrm {nm}$ ; dashed are the lines where $q_{\mathrm {A}}$ and $q_{\mathrm {B}}$ vanish. (b) Effective masses of electrons in the four subbands $n=1,\ 2,\ 3,\ \textrm {and}\ 4$ as functions of wave number of electrons; dashed are the effective masses of bulk GaAs $(m_{\mathrm {c,A}}^\ast )$ and bulk Al $_{0.35}$ Ga $_{0.65}$ As $(m_{\mathrm {c,B}}^\ast )$ . All the material parameters are taken from Ref. [20].
Fig. 3.
Fig. 3. Probability $\left |f_{\Gamma _6,n}(0;z)\right |^2$ of finding an electron with $q=0$ inside a bilayer Al $_{0.35}$ Ga $_{0.65}$ As/GaAs quantum well for the first four subbands $(n=1,\ 2,\ 3,\ \textrm {and}\ 4)$ . All the parameters are the same as in Fig. 2.

Equations (29)

Equations on this page are rendered with MathJax. Learn more.

H ^ c f Γ 6 ( r ) = E Γ 6 f Γ 6 ( r ) , H ^ c = { E Γ 6 , A 0 + 2 k ^ 2 / ( 2 m c , A ) , 0 z L A , E Γ 6 , B 0 + 2 k ^ 2 / ( 2 m c , B ) , L A z L ,
f Γ 6 ( r ) = 1 S e i q r f Γ 6 ( z ) ,
d 2 f Γ 6 ( z ) d z 2 + q z 2 ( z ) f Γ 6 ( z ) = 0 , q z 2 ( z ) = { q A 2 , 0 z L A , q B 2 , L A z L ,
q A 2 = 2 m c , A 2 ( E Γ 6 E Γ 6 , A 0 2 q 2 2 m c , A ) , q B 2 = 2 m c , B 2 ( E Γ 6 E Γ 6 , B 0 2 q 2 2 m c , B ) .
f Γ 6 ( z ) = { A A e i q A z + B A e i q A z f Γ 6 A ( z ) , 0 z L A , A B e i q B z + B B e i q B z f Γ 6 B ( z ) , L A z L .
1 m c , A d f Γ 6 A ( z ) d z | z = L A = 1 m c , B d f Γ 6 B ( z ) d z | z = L A .
q B m c , B tan ( q A L A ) = q A m c , A tan ( q B L B ) .
f Γ 6 , n ( q ; z ) = 2 i A A , n { sin ( q A , n z ) , 0 z L A , b n sin [ q B , n ( L z ) ] , L A z L ,
A A , n = s i g n ( q A , n ) [ 2 L A ( 1 sin ( 2 q A , n L A ) 2 q A , n L A ) + 2 L B b n 2 ( 1 sin ( 2 q B , n L B ) 2 q B , n L B ) ] 1 / 2 ,
Ψ Γ 6 , s z ; n , q ( r ) = 1 S e i q r f Γ 6 , n ( q ; z ) u Γ 6 , s z ( r ) .
M f i = Ψ Γ 6 , α f ; n f , q f | H ^ e R | Ψ Γ 6 , α i ; n i , q i , H ^ e R = e m c A p ^ ,
M f i = e m c 0 L d z A ( z ) S d r Ψ Γ 6 , α f ; n f , q f ( r ) p ^ Ψ Γ 6 , α i ; n i , q i ( r ) .
M f i = δ α f , α i e m c S 0 L d z U f i ( z ) f Γ 6 , n f ( q f ; z ) [ A ( z ) q i + A z ( z ) p ^ z ] f Γ 6 , n i ( q i ; z ) ,
U f i ( z ) = S d r u Γ 6 , α f ( r ) u Γ 6 , α i ( r ) e i ( q i q f ) r .
M f i A δ α f , α i e m c S m J f i A ( b m ) e i ( q i q f ) b m Ω A d r | u Γ 6 , α i A ( r ) | 2 e i ( q i q f ) r ,
J f i A ( ζ ) = f Γ 6 , n f ( q f ; ζ ) ( ( A A q i ) f Γ 6 , n i ( q i ; ζ ) i A z A d f Γ 6 , n i ( q i ; z ) d z | z = ζ )
M f i A δ α f , α i δ q f , q i e m c 0 L A J f i A ( z ) d z ,
M f i δ α f , α i δ q f , q i e m c ( 0 L A J f i A ( z ) d z + L A L J f i B ( z ) d z ) ,
M f i = δ α f , α i δ q f , q i M n f , n i ( 0 ) ( q i ) , M n f , n i ( 0 ) ( q ) = i e m c ( A z A Θ f i A + A z B Θ f i B ) ,
Θ f i A = 0 L A d ζ f Γ 6 , n f ( q ; ζ ) d f Γ 6 , n i ( q ; z ) d z | z = ζ ,
Θ f i B = L A L d ζ f Γ 6 , n f ( q ; ζ ) d f Γ 6 , n i ( q ; z ) d z | z = ζ .
Θ f i A = A A , n f A A , n i 4 q A , n i q A , n f 2 q A , n i 2 [ q A , n f ( 1 c A , f i ) q A , n i s A , f i ] ,
Θ f i B = A A , n f A A , n i 4 q B , n i q B , n f 2 q B , n i 2 [ q B , n f ( 1 c B , f i ) b n f b n i q B , n i s A , f i ] ,
W = 2 n f , n i 0 | M n f , n i ( 0 ) ( q ) | 2 [ F n i ( q ) F n f ( q ) ] δ [ E Γ 6 , n f ( q ) E Γ 6 , n i ( q ) ω ] q d q ,
M n f , n i ( 0 ) | L B = 0 = i e m c A z L 2 n f n i n f 2 n i 2 [ 1 ( 1 ) n f + n i ] ,
E Γ 6 > min ( E Γ 6 , A 0 + 2 q 2 2 m c , A , E Γ 6 , B 0 + 2 q 2 2 m c , B ) ,
E Γ 6 , A 0 + 2 q 2 2 m c , A < E Γ 6 < E Γ 6 , B 0 + 2 q 2 2 m c , B .
E Γ 6 , B 0 + 2 q 2 2 m c , B < E Γ 6 < E Γ 6 , A 0 + 2 q 2 2 m c , A ,
E Γ 6 > max ( E Γ 6 , A 0 + 2 q 2 2 m c , A , E Γ 6 , B 0 + 2 q 2 2 m c , B ) ,
Select as filters


Select Topics Cancel
© Copyright 2024 | Optica Publishing Group. All rights reserved, including rights for text and data mining and training of artificial technologies or similar technologies.