Electric Field Control of Shallow Donor Impurities in Silicon
Extracted source text ยท arXiv cond-mat/0308575
\documentclass[preprint,aps,prb,showpacs,preprintnumbers,amsmath,amssymb,groupedaddress]{revtex4}
\usepackage{graphicx}
\begin{document}
\draft
\title {Electric Field Control of Shallow Donor Impurities in Silicon}
\author{A. S. Martins, R. B. Capaz, and Belita Koiller}
\affiliation{Instituto de F\'{\i}sica, Universidade Federal do Rio de
Janeiro,
Cx.P. 68.528, 21945-970, RJ, Brazil}
\date{\today}
\begin{abstract}
We present a tight-binding study of donor impurities in Si,
demonstrating the adequacy of this approach for this problem
by comparison with effective mass theory and experimental results.
We consider the response of the system to an applied electric field:
donors near a barrier material and in the presence of an uniform
electric field may undergo two different ionization regimes according
to the distance of the impurity to the Si/barrier interface. We show
that for impurities $\sim 5$ nm below the barrier, adiabatic ionization
is possible within switching times of the order of one picosecond,
while for impurities $\sim 10$ nm or more below the barrier, no
adiabatic ionization may be carried out by an external uniform electric
field. Our results are discussed in connection with proposed Si:P
quantum computer architectures.
\end{abstract}
\pacs{71.15.Ap, 71.55.-i, 71.55.Cn, 03.67.Lx}
\keywords{Impurities in Si,Tight-binding}
\maketitle
\section{Introduction}
Simple donors in Si have recently become the subject of renewed interest
due to proposals of quantum computer architectures in which P donors in Si play
the role of qubits.\cite{kane,vrijen,skinner} Logic operations in such
architectures involve the response of the bound electron wavefunctions to
voltages applied to a combination of metal gates separated by a barrier
material (e.g. SiO$_2$) from the Si host. The so-called A-gate, placed above
each donor site, pulls the electron wavefunction away from the donor, aiming at
partial reduction \cite{kane} or total cancelation \cite{skinner} of the
electron-nuclear contact coupling in architectures where the qubits are the
$^{31}$P nuclear spins. In a related proposal based on the donor electron
spins as qubits, \cite{vrijen} the gates drive the electron wavefunction into
regions of different $g$-factors, allowing the exchange coupling between
neighboring electrons to be tuned. Ideally, electric-field control over the
donor electron wavefunction requires all operations to be performed in the
adiabatic regime \cite{messiah}, which sets a lower bound for the time scales
involved in such processes.
Recent studies have demonstrated that the tight-binding (TB) approach,
traditionally adopted for deep levels,\cite{hjalmarson} provides a valid
description for intermediate \cite{form1,form2} and shallow levels
\cite{martins02} in semiconductors. Impurity states are calculated from a
sequence of supercell sizes and a finite-size analysis which provides
extrapolation to the bulk limit. Also, electric-field effects may be
easily incorporated within the TB scheme \cite{graf}, allowing estimates of
switching times in electric-field-tunable devices \cite{ribeiro}. In
this work we present a TB description for donors in Si, aiming at a
physical description of the relevant properties involved in the A-gate
operations mentioned above.
Donors in Si have been extensively and successfully investigated within
the effective mass theory (EMT), \cite{emt} thus providing a preliminary test
for the TB approach by comparison of wavefunctions predicted by
the two formalisms. This comparison is presented in the next section. In
Sec.~\ref{sec:field} we explore a simplified model of the A-gate operations
in the Kane quantum computer proposal \cite{kane} by considering the Si:P
system under an uniform electric field and near a barrier. In Sec.~\ref
{sec:times} we discuss operation times and restrictions imposed by the donor
positioning with respect to the Si/barrier interface in connection with the
adiabaticity of the A-gate operations. Our summary and conclusions are
presented in Sec.~\ref{sec:end}.
\section{TB description for donors in Silicon}
\label{sec:TB}
\subsection{Formalism}
The TB Hamiltonian for the impurity problem is written as \cite{form1}
\begin{equation}
H=\sum\limits_{ij}\sum_{\mu
\nu}h_{ij}^{\mu\nu}c_{i\mu}^{\dagger}c_{j\nu}+\sum_
{i,\nu} U(r_i)c_{i\nu}^{\dagger}c_{i\nu} \label{hamilton}
\end{equation}
where $i$ and $j$ label the atomic sites, $\mu$ and $\nu$ denote the
atomic orbitals and $r_{i}$ is the distance of the site $i$ to the impurity
site. The matrix elements $h_{ij}^{\mu\nu}$ define all the on-site energies
and first and second neighbors hoppings for the bulk material. The donor
impurity potential $U(r_i)$ is described by a screened Coulomb potential
($\epsilon=12.1$ for Si)
\begin{equation} U\left(r_{i}\right)=-\frac{e^{2}}{\epsilon r_{i}}.
\label{potential}
\end{equation}
At the impurity site $\left(r_{i}=0\right)$, the perturbation potential is
assigned the value $-U_{0}$, a parameter describing central cell effects
characteristic of the substitutional species. In the present calculations,
$U_{0}$ was kept as an adjustable parameter (previous estimates for this
parameter\cite{form1} are of the order of one to a few eV). We adopt here
the $sp^{3}s^{\ast}$ TB parametrization for Si proposed by Klimeck {\it et
al},\cite{boysi} which includes first and second neighbors interactions.
Inclusion of hopping matrix elements up to second neighbors provides a good
description of the effective masses at the conduction band minima. This
parametrization gives the $k$-space positions of the six band minima at the
six equivalent points along the $\Delta$ lines, at $\Delta_{min}=0.75
(2\pi/{\rm a})$, where $\rm{a} = 5.431$\AA~is the conventional cubic lattice
parameter for Si. We do not include spin-orbit corrections in our calculations.
The eigenstates of $H$ are determined for a system where a single impurity
is placed in a cubic supercell containing $N=8L^{3}$ atoms arranged in the
diamond structure, where $L$ is the length of the supercell edge in units of
${\rm a}$. The supercells are subject to periodic boundary conditions, and
full numerical diagonalization can be performed for $L\lesssim 6$. Much larger
supercells\cite{martins02} (up to $10^6$ atoms) may be treated within a variational
scheme\cite{araujo} where the ground state wavefunction and binding energy $E_{L}$
for a donor level is obtained by minimizing the expectation value of
$\left\langle\Psi\left|\left(H-\varepsilon_{ref}\right)^{2}\right|\Psi
\right\rangle$. For the donor ground state, $\varepsilon_{ref}$ is a reference
energy chosen well within the gap, but nearest to the conduction band minimum,
and excited states are obtained by tuning $\varepsilon_{ref}$ towards the conduction
band edge. Finite size scaling allows extrapolation to the bulk limit ($L\to\infty$)
according to the {\it ansatz}~\cite{form1,form2}
\begin{equation}
E_{L}=E_{b} + \widetilde{E}e^{-L/\lambda},
\label{ansatz}
\end{equation}
where $E_b$ is the binding energy for a single donor in the bulk.
The eigenfunctions of (\ref{hamilton}) written in the basis of atomic
orbitals $|\phi_{\nu}(\mathbf{r}-\mathbf{R}_i)\rangle$ are given by
$|\Psi_{TB} (\mathbf{r})\rangle = \sum_{i\nu} a_{i\nu}|\phi_{\nu}
(\mathbf{r}-\mathbf{R}_i)\rangle $ where the expansion
coefficients $a_{i\nu}$ give the probability amplitude of finding the
electron in the orbital $\nu$ localized at $\mathbf{R}_i$. We do not include
explicit expressions for the atomic orbitals; the overall charge distribution
is conveniently described through the TB envelope function squared,\cite{tania}
\begin{equation}
|\Psi_{EF}({\mathbf{R}_i})|^2 = \sum_{\nu} |a_{i\nu}|^2.
\label{envfunc}
\end{equation}
\subsection{Donor ground state}
In the proposed TB model, the only free parameter is related to the
on-site value for the impurity potential, $U_0$. In Fig.~\ref{fig1}(a) we
present the converged $(L\to \infty)$ binding energy of the lowest donor state
as a function of $U_0$. We also characterize the donor ground state by its
orbital averaged spectral weight\cite{tania} at $\Delta_{min}$
\begin{equation}
W(\Delta_{min})=\frac{2}{N}\sum_{\mu=1}^6\sum\limits_{ij\nu}e^{i{\bf{k}_\mu}
\cdot({\bf R}_{i}-{\bf R}_{j})} a_{i\nu}a_{j\nu},
\label{espectral}
\end{equation}
where $N$ is the number of atomic sites in the supercell, and the first summation
is over the six equivalent $\mathbf{k_\mu}$ at the conduction band minima. This
quantity is plotted in Fig.~\ref{fig1}(b) as a function of $U_0$.
We determine the value of $U_0$ so that the binding energy of the donor
results to be in good agreement with the experimental value which,
for P in Si, is $E_b=45.6$~meV. As indicated in Fig.~\ref{fig1}, $U_{0}=U_{\rm P}
\cong 1.48$ eV gives the correct binding energy for the P donors in Si.
This value for $U_0$ is used in the calculations below.
\subsection{Comparison with EMT}
EMT exploits the duality between real and reciprocal space, where
delocalization in real space leads to localization in $k$-space, e.g. for
shallow donors around the $k$-vector at the minimum of the conduction
band. Within EMT in its simplest formulation,\cite{emt} the ground state
for donors in Si is six-fold degenerate, due to the six-fold
degeneracy of the Si conduction band. Valley-orbit interactions \cite
{baldereschi} lead to a non-degenerate ground state wavefunction of $A_1$
symmetry,\cite{foot-symmetry}
%
\begin{equation}
\psi ({\bf r}) = \frac{1}{\sqrt{6}}\sum_{\mu = 1}^6 F_{\mu} ({\bf r})
\phi_\mu(\bf r)\,,
\label{eq:sim}
\end{equation}
%
where $\phi_\mu({\bf r}) = u_\mu({\bf r})e^{i {\bf k}_{\mu}\cdot{\bf r}}$
are the pertinent Bloch wave functions, and the envelope functions given by
(e.g. for $\mu = z$)\cite{emtsi}
%
\begin{equation}
F_{z} ({\bf r}) = \frac{1}{\sqrt{\pi a^2 b}}\ e^{-[(x^2+y^2)/a^2 +
z^2/b^2]^
{1/2}} \,.
\label{eq:envelope}
\end{equation}
%
The effective Bohr radii for Si from a variational calculation are $a=2.51$~nm
and $b=1.44 $~nm.\cite{koiller01} In Fig.~\ref{fig2} we present the TB envelope
function squared calculated from (\ref{envfunc}) along three symmetry
directions with the corresponding EMT results obtained from (\ref{eq:sim}),
where the periodic part of the Bloch functions have not been explicitly
included, consistent with not explicitly including the atomic orbitals in the
TB description. Note that the oscillatory behavior coming from the interference
among the plane-wave part of the six $\phi_\mu$ is well captured by the TB
envelope function.
The good agreement between TB and EMT is limited to distances from the
impurity site larger than a few lattice parameters ($\sim$~1~nm). Closer
to the impurity, particularly at the impurity site, the TB results become
much larger than the EMT prediction, in qualitative agreement with experiment.
\cite{emtsi} This reflects central cell effects, not included in the EMT
expressions (\ref{eq:sim}) and (\ref{eq:envelope}). In the central cell region,
the discrepancy between TB and EMT wavefunctions is significantly larger than
those reported for donors in GaAs, \cite{martins02} a result that could have
been anticipated from the spectral weight given in Fig.~\ref{fig1}(b).
EMT rests on the assumption that the impurity eigenstate is highly
localized in $k$-space, so that only Bloch states near the conduction
band minima enter in the expansion, as implied in Eq.~(\ref{eq:sim}).
This is the case for GaAs,\cite{martins02} where for a range of values of
$U_0$ ($U_0<1.8$ eV) we find $W(\Gamma)$ essentially equal to one, in agreement
with the EMT assumption. In Si, even small values of $U_0$ yield spectral
weights at $\Delta_{min}$ well below one. For $U_0=U_{\rm P}$ in particular,
$W(\Delta_{min})\cong 0.3$.
We remark that the sharp shallow-to-deep transition obtained for GaAs in
Ref.~\onlinecite{martins02}, with kinks in the curves of $E_b$ and $W$
versus $U_0$, is not reproduced here (see Fig.~\ref{fig1}). We attribute this
to the lack of a strictly shallow region, with the spectral weight of the donor
state concentrated in one or a few $k$-points. Therefore, while the binding
energy of shallow donors in GaAs is essentially constant, independent of the
species ($\sim 6$~meV for C, Si and Ge, in excellent agreement with the EMT
estimate), in Si it varies according to the donor species (45 meV for P, 53 meV
for As and 42 meV for Sb, to be compared with the EMT estimate of 30 meV). It is
interesting to note in Fig.~\ref{fig1}(a) that, as the impurity level
becomes shallower by decreasing $U_0$, $E_b$ approaches the EMT
estimate for the binding energy.\cite{emt}
\section{Donors under an uniform Electric Field}
\label{sec:field}
The formalism presented in Sec.\ref{sec:TB} is easily extended to include
an uniform electric field in the system. Assuming a constant field $\mathbf{E}$
applied along the $[00\bar 1]$ direction, it is incorporated in the TB
formalism by modifying the on-site energies in Eq.~(\ref{hamilton}) as
follows:\cite{graf,ribeiro}
\begin{equation}
h_{ii}^{\nu\nu}(E) = h_{ii}^{\nu\nu}(0) - |e|Ez_{i}.
\label{campo}
\end{equation}
Periodic boundary conditions lead to a discontinuity in the
potential at the supercell boundary $z_i=Z_B$, where $Z_B$ is half of the
supercell length along [001] or, equivalently, the distance from the
impurity to the Si/barrier interface. The potential discontinuity, $V_B=2|e|EZ_B$,
actually has a physical meaning in the present study: It models the potential
due to the barrier material layer above the Si host\cite{kane} (see inset in Fig.~\ref{fig3}).
A description for the A-gate operations may be inferred
from the behavior of the TB envelope function squared at the impurity site
under applied field $E$, normalized to the zero-field value:
%
\begin{equation}
A/A_0 = |\Psi_{EF}^E(0)|^2/|\Psi_{EF}^0(0)|^2
\label{ratio}
\end{equation}
%
The notation here indicates that this ratio should follow a behavior similar
to that for the hyperfine coupling constants between the donor nucleus and
electron with $(A)$ and without $(A_0)$ external field. Since the
hyperfine interaction $A$ is proportional to $|\Psi (0)|^2$, and
we are using here the envelope rather than the full TB eigenfunctions, this
equivalence is not rigorous. The ratio in (\ref{ratio}) is plotted in
Fig.~\ref{fig3}(a) for three values of the impurity depth with respect to the
Si/barrier interface. Calculations for $Z_B$=10.86~nm were performed with
cubic supercells $(L=40)$, while for $Z_B$= 5.43 and 21.72~nm tetragonal
supercells with $L_x=L_y=40$ and $L_z=20$ and 80 respectively were used.
At small field values we obtain a quadratic decay of $A/A_0$ with $E$,
in agreement with the perturbation theory results for the hydrogen atom.
\cite{schiff} At large enough fields, $|\Psi_{EF}^E(0)|^2$ becomes vanishingly
small, and the transition between the two regimes is qualitatively different
according to $Z_B$: For the largest values of $Z_B$ we get an abrupt transition
at a critical field $E_c$, while smaller $Z_B$ (e.g. $Z_B=5.43$~nm) lead to a
smooth decay, similar to the one depicted in Ref.~\onlinecite{kane}. In this latter
case, we define $E_c$ as the field for which the curve $A/A_0$ vs $E$ has an
inflection point, where $A/A_0\sim 0.5$, thus $E_c(5.43 \rm{nm}) = 130$ kV/cm.
We find that the decrease of $E_c$ with $Z_B$ follows a simple rule
$E_c \propto 1/Z_B$, as given by the solid line in Fig.~\ref{fig3}(b).
In order to analyze the different regimes illustrated in
Fig.~\ref{fig3}(a), we study the overall behavior of the envelope squared
profile along the $z$-axis,
%
\begin{equation}
\rho (z) =\sum_{s=1}^2\sum_{x_i^s y_i^s}|\Psi_{EF}^E(\mathbf{R}^{s}_i)|^2,
\label{profile}
\end{equation}
%
where the first summation is over the two fcc sublattices, with
$\mathbf{R}^{s}_i$ corresponding to the atomic sites in sublattice $s$, thus
$z$ labels each monolayer in the diamond structure, and $\rho(z)$
quantifies the $z$-projected charge distribution for the electron states under
applied field $E$. Fig.~\ref{fig4} gives $\rho(z)$ for the electron ground
state and also for the first excited state with $Z_B=10.86$~nm as the applied
field increases. Up to fields very close to $E_c~(\sim 53$~kV/cm), the ground
state distribution retains essentially the bound donor character, with the
electronic charge accumulating predominantly around the impurity $(z=0)$. For
$E>E_c$ we observe an abrupt charge transfer towards the barrier, with some residual
charge remaining at the impurity site. The first excited state displays a
complementary behavior, with charge transfer from the barrier into the impurity
region as $E$ increases. The binding energies (energy eigenvalues relative to the
bottom of the conduction band) are calculated here taking into account
the dependence of the conduction band edge under applied field. The binding energies
of the two lowest electron states are given in Fig.~\ref{fig5}(a). Note that they
cross at $E_c$.
The binding energies of the two lowest eigenstates for $Z_B=5.43$~nm are presented in
Fig.~\ref{fig5}(b). They do not cross, but rather display an anticrossing
behavior, confirmed by the corresponding doubled-peaked charge distributions in
Fig.~\ref{fig6}, with wavefunctions extending over the attractive wells
of the impurity and of the electric field potential. This is consistent of eigenstates
which are superpositions of bound states in each potential well. Note that for $E=E_c$ in
Fig.~\ref{fig6}(c), the two states have essentially the same charge distribution,
as expected at the anticrossing point. The anticrossing in Fig.~\ref{fig5}(b) is such that
for $E<E_c$ the lines giving the two states are essentially parallel, converging
asymptotically at zero field to the binding energies 45.6 meV, for the $A_1$ ground
state, and 32.4 meV for the first excited state. This is very close to the experimental
binding energy of the excited $E$ (32.6 meV) and $T_2$ (33.9 meV) states, which can not
be individually resolved within our variational scheme.\cite{foot-symmetry} Note
that this was independently obtained with the same value of the parameter $U_0$,
chosen to fit the $A_1$ state binding energy alone. Near and above $E_c$ a typical
2-level anticrossing behavior is obtained, with the excited state eventually merging
into the conduction band at $E=150$ kV/cm.
The above results may be understood within a simple picture of the electron
in a double well potential, the first well being most attractive at the
impurity site, $V(\mathbf{R}=0)=-U_0$, and the second well at the barrier
interface, $V(z=Z_B)= -V_B/2=-|e|EZ_B$ neglecting the Coulomb potential
contribution (\ref{potential}) at the interface. An internal barrier separates
the two wells and, for a fixed $E$, this internal barrier height and width
increase with $Z_B$. Deep donor positioning leads to a weaker coupling between
the states localized at each well, even close to level degeneracy, resulting the
level crossing behavior illustrated in Fig.\ref{fig5}(a). For donor positioning
closer to the interface the internal barrier gets weaker, enhancing the coupling
between levels localized in each well and leading to wavefunction superposition
and to the anticrossing behavior illustrated in Fig.\ref{fig5}(b). The scaling of
$E_c$ with $1/Z_B$ may also be understood assuming that the critical field
corresponds to the crossing of the ground state energies of two wells: The
Coulomb well and an approximately triangular well at the barrier. Since the
relative depths of the wells increases with $EZ_B$, and assuming that the ground
states energies are fixed with respect to each well's depth, leads to the
$E_c\propto 1/Z_B$ behavior.
\section{Adiabatic processes driven by an uniform Electric Field}
\label{sec:times}
Coherent manipulation of electrons by the A-gates requires
that the switching time between different electron states be slow enough to
guarantee adiabaticity of the process. Instantaneous eigenstates of $H(t)$ may
thus be defined at any time $t$. In the present case, we assume a linear
increase of the external field from 0 to a maximum value $E_{\rm max}$ so that
$H(t)= H(0)-|e|E_{\rm max} z t$, with $0<t<T$, where $T$ is the total switching
time. A lower bound for $T$ is obtained from the adiabatic theorem,
\cite{messiah,farhi} following Ref.\onlinecite{ribeiro}:
\begin{equation}
T_a=\frac{\hbar|e|E_{\rm max} Z_B}{g_{\rm min}^2},
\label{adiabatic}
\end{equation}
where $g_{\rm min}$ is the minimum gap between the two lowest electron
states. In the anticrossing case illustrated in Fig.\ref{fig5}(b), we
get $g_{\rm min}\cong 9.8$ meV. Assuming that a totally ionized state is
required as the final state, we take $E_{\rm max} = 180$ kV/cm, leading to
$T_a\sim0.5$~ps. This is a perfectly acceptable time for the operation of
A-gates in spin-based Si QC, given the relatively long electron spin coherence
times (of the order of a few ms) in Si.\cite{sousa}
As the impurity distance from the barrier increases, one eventually reaches
the crossing regime, when $g_{\rm min}\to 0$, meaning that $T_a\to\infty$
and no adiabatic ionization is possible. Ionization would still occur for
$E>E_c$, but as a stochastic decay process from the first excited
state. From Fig.~\ref{fig3}(a) we see that the A-gate might be used to
partially reduce the contact interaction, in the case of $Z_B=10.86$~nm to
about 20\% of its value at zero field. For larger $Z_B$ the range for
adiabatic variation in $A/A_0$ is even smaller. Therefore $Z_B\sim 5$~nm seems
to be a favorable positioning for the donors, since it allows adiabatic
reduction of $A/A_0$ to any desired final value, with this ratio varying
smoothly from one (at $E=0$) to zero (for $E=E_{\rm max}\sim 2E_c$).
\section{Summary and Conclusions}
\label{sec:end}
We have presented a TB study of donor levels in Si. The reliability of
the TB approach for the present study was verified by comparison of the
TB and EMT envelope functions as well as by the value predicted for the
$A_1 - \{E, T_2\}$ energy splitting in agreement with experiment within
our numerical accuracy. Previous TB studies of intermediate and shallow
impurity levels in semiconductors\cite{form1,form2,martins02} dealt with
materials with band extrema at ${\bf k}= 0$, and the present results show
that the oscillatory behavior of the wave function due to interference
effects in the plane-wave part of the Bloch wavefunctions, typical of
degenerate band extrema at $\bf k \ne 0$, is well is well captured
by the TB approach.
In the presence of an increasing uniform electric field, the donor states
respond in different ways according to the donor depth $Z_B$ below the
Si/barrier interface. For deeply positioned donors, i.e, for $Z_B>>a,b$,
where $a$ and $b$ are the Bohr radii for P in Si, abrupt ionization occurs
at a critical field $E_c$, while for $Z_B$ greater but of magnitude comparable
to the Bohr radii, a smooth electronic charge transfer from the donor site
towards the barrier interface is obtained, eventually leading to complete
ionization. The different regimes were identified in three ways: (i) From the
decrease in electronic charge at the donor nucleus [Fig.~\ref{fig3}(a)]. This
behavior implies an analogous dependence of the electron-nucleus hyperfine
coupling constant $A$ as a function of the increasing external field. (ii) From
charge distributions (Figs.~\ref{fig4} and \ref{fig6}), where the superposition of
donor-like and barrier-like bound states is inhibited for deeply positioned donors.
(iii) From the behavior of the binding energies of the two lowest electron states
as the applied field increases (Fig.~\ref{fig5}), changing from a level-crossing into
an anticrossing regime as $Z_B$ decreases. The donor excited states in the S-like
manifold also play a role in the anticrossing regime, as illustrated for
$E\lesssim E_c$ in Fig.~\ref{fig5}(b).
The minimum gap $g_{\rm min}$ in the anticrossing regime is a key ingredient
determining the possibility of an adiabatic evolution of the electron state
under the action of the A-gates. Given that the product $E_c Z_B$ is approximately
constant [see fit Fig.~\ref{fig3}(b)], the {\it adiabatic} time $T_a$ in
(\ref{adiabatic}) is expected to depend very weakly in the product $E_{\rm max} Z_B$,
assuming one aims at complete ionization.\cite{skinner} Therefore $T_a$ should
not depend explicitly on $Z_B$, but only implicitly through $1/g_{\rm min}^2$.
We have shown that for $Z_B\sim 5$~nm, i.e., about twice the largest Bohr radius $a$
in Eq. (\ref{eq:envelope}), electric field switching times smaller than 1 ps may be reached,
which is a favorable operation time given the long electronic spin coherence times in Si.
If one aims at a final state where only partial reduction of the electronic
charge at the nucleus occurs\cite{kane,vrijen}, values of $Z_B$ of this order
of magnitude are still the most convenient, since any final value of the
nuclear charge may be attained.
The Bloch phases interference behavior in the donor wavefunctions has been
previously shown to lead to oscillatory behavior of the exchange coupling between
two donors,\cite{koiller01} affecting the two-qubit operations in exchange-based
architectures in Si. We remark that such oscillations are well captured in the
TB wavefunctions, and that the present study demonstrates that electric field
control over single donor wavefunctions, such as proposed in A-gate operations,
\cite{kane,vrijen,skinner} do not present additional complications due
to the Si band structure. The only critical parameter is the donor positioning
below the Si/barrier interface, which should be chosen and controlled according
to the physical criteria presented here.
\acknowledgments
We thank Daniel Loss and Bruce Kane for interesting discussions.
We also thank Fabio Ribeiro for useful suggestions. This work is partially
supported by Brazilian agencies Conselho Nacional de Desenvolvimento Cient\'\i fico
e Tecnol\'ogico (CNPq), CAPES, Funda\c c\~ao de Amparo \`a Pesquisa do Rio de
Janeiro (FAPERJ), Funda\c c\~ao Universit\'aria Jos\'e Bonif\'acio (FUJB),
PRONEX-MCT, and Instituto do Mil\^enio de Nanoci\^encias-CNPq.
\begin{thebibliography}{100}
\bibitem{kane} B. E. Kane, Nature \textbf{393}, 133 (1998)
\bibitem{vrijen} R. Vrijen, E. Yablonovitch, K. Wang, W. H. Jiang, A.
Balandin, V. Roychowdhury, T. Mor, and D. DiVicenzo, Phys. Rev. A \textbf{62}, 012306 (2000)
\bibitem{skinner} A. J. Skinner, M. E. Davenport, and B. E. Kane, Phys. Rev. Lett. \textbf{90}, 087901 (2003)
\bibitem{messiah} A. Messiah, \textit{Quantum Mechanics} (Dover-New York, 1999)
\bibitem{hjalmarson} H. P. Hjalmarson, P. Vogl, D. J. Wolford, and J. D. Dow, Phys. Rev. Lett.
\textbf{44}, 810 (1980)
\bibitem{form1} J. G. Menchero, R. B. Capaz, B. Koiller, and H. Chacham,
Phys. Rev. B \textbf{59}, 2722 (1999)
\bibitem{form2} J. G. Menchero and T. B. Boykin,
Phys. Rev. B \textbf{59}, 8137 (1999)
\bibitem{martins02} A. S. Martins, J. G. Menchero, R. B. Capaz, and B. Koiller,
Phys. Rev. B \textbf{65}, 245205 (2002)
\bibitem{graf} M. Graf and P. Vogl, Phys. Rev. B \textbf{51}, 4940 (1995)
\bibitem{ribeiro} F. J. Ribeiro, R. B. Capaz, and Belita Koiller, Appl. Phys. Lett.
\textbf{81}, 2247 (2002)
\bibitem{emt} J. M. Luttinger and W. Kohn, Phys. Rev \textbf{97}, 869 (1955)
\bibitem{boysi} G. Klimeck, R. C. Bowen, T. B. Boykin, C. S.-Lazaro, T. A. Cwik, and A. Stoica, Superlattices and Microstructures \textbf{27}, 77 (2000)
\bibitem{araujo} R. B. Capaz, G. C. de Araujo, B. Koiller, and J. P. von der Weid, J,. Appl. Phys. \textbf{74}, 5531 (1993)
\bibitem{tania} T. G. Dargam, R. B. Capaz, and B. Koiller, Phys. Rev. B \textbf{56}, 9625 (1997)
\bibitem{baldereschi} A. Baldereschi, Phys. Rev. B \textbf{1}, 4673 (1970)
\bibitem{foot-symmetry} The six-fold degenerate S-like donor states in Si split into
three components due to the tetrahedral symmetry of the host: A non-degenerate
ground state of $A_1$ symmetry given in Eq.~(\ref{eq:sim}), an excited triplet
of symmetry $T_2$ followed by an excited doublet of symmetry $E$. The TB treatment
adopted here correctly describes the symmetry and ordering of the states within
this manifold when full diagonalization of the TB Hamiltonian is performed.
Within the variational scheme we have not resolved the nearly-degenerate
$T_2$ and $E$ levels, which are only 1.3 meV apart.
\bibitem{emtsi} W. Kohn, Solid State Physics Series \textbf{5}, Academic Press (1957)
\bibitem{schiff} L. I. Schiff, Quantum Mechanics, Third Edition - McGraw-Hill (chapter 8), (1968)
\bibitem{koiller01} B. Koiller, X. Hu, and S. Das Sarma, Phys. Rev. Lett \textbf{88}, 27903 (2002)
\bibitem{farhi} E. Farhi, J. Goldstone, S. Gutmann, J. Lapan, A. Lundgren, and D. Preda, Science
\textbf{292}, 472 (2001)
\bibitem{sousa} R. de Sousa and S. Das Sarma, cond-mat/0211467 (2003), to appear in PRB
\end{thebibliography}
\begin{figure}
\includegraphics[height=5.5in]{figure1.eps}
\caption{\textbf{(a)} Binding energy of the ground impurity state as a function
of the on-site perturbation strength $U_0$, obtained from the $L\to\infty$
extrapolation ansatz. The dotted line indicates the value $U_0 = U_P$ that
reproduces the experimental Si:P $A_1$ state binding energy. \textbf{(b)}
Calculated spectral weight at the conduction band edge for the ground state.
Notice that as the perturbation $U_0$ becomes weaker, $E_b$ approaches the
EMT binding energy, while $W$ does not approach 1.
\label{fig1}}
\end{figure}
\begin{figure}
\includegraphics[height=6in]{figure2.eps}
\caption{The dots give the TB envelope function squared for the lowest impurity
state along three high-symmetry directions. The lines are the corresponding
effective mass $|\psi|^2$ results. Note that the TB approach captures the
oscillations of the EMT wave function in the asymptotic region.
\label{fig2}}
\end{figure}
\begin{figure}
\includegraphics[height=6in]{figure3.eps}
\caption{\textbf{(a)} TB envelope function squared at the impurity site
under applied field $E$, normalized to the zero-field value, for the indicated
values of the impurity-Si/barrier interface distance $Z_B$. \textbf{(b)} Dependence
of the critical field $E_c$ on $Z_B$. The solid line is a best fit of the form
$E_c \propto 1/Z_B$. The inset gives a schematic representation of the perturbation
potential added to the bulk Si hamiltonian due to the impurity at ${\bf R}=0$ and
to a uniform electric-field in the negative $z$ direction.
\label{fig3}}
\end{figure}
\begin{figure}
\includegraphics[height=6in]{figure4.eps}
\caption{Tight binding envelope function squared projected along $z$ for
$Z_{B}=10.86$~nm and the indicated values of the field $E$ applied in the
negative $z$ direction. The solid (dashed) line gives the donor ground
(1st excited) state. Note in (b) and (c) the exchange among the $\rho(z)$
for the lowest energy states [{\it ground}] $\leftrightarrow$ [{\it excited}]
which occurs over a narrow rage of electric field increase, a signature of the
crossing behavior in Fig.~\ref{fig5}(a).
\label{fig4}}
\end{figure}
\begin{figure}
\includegraphics[height=6in]{figure5.eps}
\caption{Calculated binding energies versus electric field intensity
of the two lowest donor electron states. \textbf{(a)} For $Z_{B}=10.86$~nm
the energies reveal a crossing regime. \textbf{(b)} Anticrossing of the
two lowest electron states for $Z_{B}=5.43$~nm. The open symbols correspond
the zero field calculated values: 45.6 meV and 32.4 meV.
\label{fig5}}
\end{figure}
\begin{figure}
\includegraphics[height=6in]{figure6.eps}
\caption{TB envelope function squared projected along $z$ for
$Z_{B}=5.43$~nm and the indicated values of the applied field $E$.
The solid (dashed) line gives the ground (1st excited) state.
At the critical field in {\bf (c)} the two states have similar
charge distributions, typical of a superposition of states localized in
each well and a signature of the anticrossing behavior
in Fig.~\ref{fig5}(b).
\label{fig6}}
\end{figure}
\end{document}