Skip to main content

Full text of "Deformed Wigner crystal in a one-dimensional quantum dot"

See other formats


Deformed Wigner crystal in a one-dimensional quantum dot 



Yasha Gindikin and V. A. Sablikov 
Kotel'nikov Institute of Radio Engineering and Electronics, 
Russian Academy of Sciences, Fryazino, Moscow District, 141190, Russia 

The spatial Fourier spectrum of the electron density distribution in a finite ID system and the 
distribution function of electrons over single-particle states are studied in detail to show that there 
are two universal features in their behavior, which characterize the electron ordering and the defor- 
mation of Wigner crystal by boundaries. The distribution function has a <5-like singularity at the 
Fermi momentum Uf- The Fourier spectrum of the density has a step- like form at the wavevector 
with the harmonics being absent or vanishing above this threshold. These features are found by 
calculations using exact diagonalization method. They are shown to be caused by Wigner ordering 
of electrons, affected by the boundaries. However the common Luttinger liquid model with open 
boundaries fails to capture these features, because it overestimates the deformation of the Wigner 
crystal. An improvement of the Luttinger liquid model is proposed which allows one to describe the 
above features correctly. It is based on the corrected form of the density operator conserving the 
particle number. 



I. INTRODUCTION 

One-dimensional (ID) quantum dots attract a great 
deal of attention as appealing model objects to study 
the effects of electron-electron (e-e) interaction, which 
is principally important for ID electron systems!^ The 
interest to the many-electron state in bounded ID sys- 
tems is presently increased due to the recent progress 
in m agnet otunnelling spectroscopy studies of such struc- 
tures 13I415J a s a conS equence of e-e interaction, electrons 
form the strongly correlated state, which is referred to 
as Luttinger liquid. Main distinctive features of a Lut- 
tinger liquid are the absence of fermionic quasiparticles, 
which manifests itself in the absence of the Fermi step in 
the momentum distribution function, the power-law be- 
havior of spectral functions near the Fermi energy with 
interaction-dependent exponents, and Wigner- like cor- 
relations of electronsP 

These properties were established for ideal, i.e. bound- 
less ID systems. However, whether they hold for realistic 
mesoscopic systems of finite length is not obvious. In- 
deed, the presence of boundaries may strongly affect the 
electron state and excitations since ID electron correla- 
tion functions are known to decay as a power of distance, 
hence there is no characteristic length. Many observa- 
tions performed on mesoscopic structures containing a 
finite ID system do not confirm the theoretical predic- 
tions made for ideal systems. Just recall the interaction- 
driven conductance renormalizatiorP in an infinite ID 
syste m that does not actually occ ur be cause of the con- 
tactsJ^^J^ the spin polarizatio n 1 13 1 14 1 that should not 
exist according to the Mattis-Lieb theorem, the '0.7' 
anomaly,^ many f acts of t he nonuniversality of conduc- 
tance quantizationjiSli^liSI which existing theories fail to 
explain, and so on. 

It is commonly believed that electrons in a bounded 
ID system form the Luttinger liquid. The boundary ef- 
fect consists in the change of electron correlations near 
the ends which are described by additional boundary ex- 
ponents. This conclusion was made in several theoretical 



works using the bosonization approach! 19 * 20 ! 21 ! Unfortu- 
nately, this approach is based on a number of model as- 
sumptions, which are not well justified. These are (i) 
the extension of the linear electron spectrum to infinite 
negative energ y, wh ich results in a violation of the con- 
servation lawsp^H (ii) the linearization of the electron 
spectrum, which can lead to a striking departure from 
the properties of a real system with quadratic disper- 
sion relationJ231 and (iii) neglecting some parts of the 2kp 
components of the electron densities in the e-e interac- 
tion Hamiltonian. 

The goal of our paper is to investigate the electron 
state in ID quantum dots beyond model assumptions. 
For this purpose the exact diagonalization method is em- 
ployed to calculate numerically the electron density dis- 
tribution in a finite ID wire and the distribution function 
of electrons over single-particle states. We have found 
that there are two unexpected features. First, the dis- 
tribution function has a <5-like singularity at the Fermi 
energy at the background of a smooth dependence on 
the energy. The second feature relates to the Fourier 
spectrum of the spatial distribution of the electron den- 
sity. The Fourier spectrum has a step-like form at the 
wavevector 2kp. Above this threshold, the harmonics 
are absent or vanishing. These properties are universal 
in the sense that they do not depend on the e-e interac- 
tion strength, interaction radius, wire length, mean elec- 
tron density. We argue that these features are caused 
by the Wigner ordering of electrons. We compare the 
obtained results with the calculations carried out within 
the frame of the common Luttinger liquid approach to 
find that it fails to capture the above results. The Lut- 
tinger liquid theory also gives a singularity of the dis- 
tribution function at the Fermi energy, but its form is 
incorrect. The Fourier spectrum of the electron density 
distribution turns out to have an incorrect singularity at 
2k f and not to vanish above this value. We have clarified 
that this discrepancy appears because the Luttinger liq- 
uid theory does not describe correctly the deformation of 
the Wigner crystal by the boundaries. The reason of this 



2 



shortcoming lies in the violation of the particle number 
conservation within the bosonization approach used. We 
find a way to remedy the Luttingcr liquid theory by in- 
troducing an improved expression for the electron density 
operator. The proposed approach allows one to describe 
correctly the above features of the distribution function 
and the Fourier harmonics of electron density. 



II. EXACT DIAGONALIZATION STUDY OF ID 
CORRELATED STATE 

A. Statement of problem 

Consider N spinless electrons in a ID quantum box 
with zero boundary conditions for the many-electron 
wavef unction, 



*|x=0 

The Hamiltonian is 







(1) 



H 



2m 



XV(*i-*i) + I>te) s ( 2 ) 



l>3 



where V(x) is the e-e interaction potential, and U(x) is 
the potential of positively charged background, which is 
considered as jelly. 

Exact diagonalization method reduces to finding the 
Hamiltonian matrix in the appropriate basis <& p and solv- 
ing the eigenvalue problem by the standard methods of 
computational linear algebra.^ As a result we obtain the 
many-electron wave function ^(x\, .., Xn), expanded in 
this basis, 



..,x N ) = 2ja p $p(xi, ..,x N ) 



(3) 



and the spectrum. It is convenient to choose the basis 
functions in the form of the Slater determinants 



1 



V'Qi(^l) 



(4) 



built from the eigenfunctions of the non-interacting 
single-particle Hamiltonian, 



2 irqx 



(5) 



The quantum number q G N, labelling single-particle 
states, is analogous to momentum in translationally 
invariant systems, and will be called so for brevity. 
The many-particle state, labelled by the vector p = 
,(Xn), is obtained by occupying the single-particle 
states ip ai , i — 1 . . . N. We adhere to the ordering con- 
vention that 014 < Qffc for i < k. 



The details of the calculation of the Hamiltonian ma- 
trix elements are given in Appendix [A] The interac- 
tion potential is chosen in the form V{x\ — X2) — 
Ve 2 /(2ed) exp( — \x\ — X2\/d), with d being the interac- 
tion radius, e the permittivity of the medium. This al- 
lows us to find the matrix elements analytically to cut 
the calculation time. Using such form of the potential is 
not essentially a limitation, because in our calculations 
we are able to vary the parameters V and d in a broad 
range. The Coulomb interaction can be modelled by tak- 
ing d ~ L, and V ~ 2e 2 ln(i/r), where r stands for the 
quantum wire radius. 

Now let us express the observables we are going to find 
via the coefficients of expansion (J3|. The momentum 
distribution function is defined as 



n(q) 



(*l<c,|*) 



(G) 



where c q is electron destruction operator. Since |<& p ) is 
the eigenvector of c+c g , 



>(<z) = Ei a pi 20 



(7) 



where the function # p equals one if the many-particle 
state <i> p has the single-particle state tp q occupied, and 
zero otherwise. 

The average value of the particle density operator 



N 



/J(x - Xi) 



(8) 



(9) 



equals 

p(z) = (*|p|*> = 

TV J 1 i'*(x,X2 7 ■■Xn) 1 ^'(x 7 X2, --Xn) dX2---dXN . 

Using , one gets 

P{x) = a Pi a P27piP2( a; ) . (10) 



where for p! = p 2 = («i, ..,ajv) 



and 



N 



7p 1 p 2 (^) = (-l) fel+fe2 V'* (^a k Ax) 



(11) 



(12) 



for the case when pi and p 2 have only two different occu- 
pied states ip ai in positions h\ and respectively; 7 = 
otherwise. The cosine Fourier-transform of the density is 

I 7rqx 
p{q) — I p(x) cos ——dx, q <E N , (13) 
Jo L 

and sine Fourier-transform is zero, since according to 
Eqs. (Ill, (12 1, p{x) contains only cosine harmonics. 



3 



n{q) 



1 

0.8 
0.6 
0.4 
0.2 



A A 



12 16 20 24 



4 8 12 16 20 24 28 32 36 40 44 



▼ T 



-0.5- 



▼ T 



-1.5 



▼ Interaction on 
• Interaction off 



P(q) 



FIG. 1: Momentum distribution function, calculated for the 
system of N = 12 electrons. The value of q — 12 corresponds 
to kp. 



FIG. 2: The Fourier-transform of electron density. The value 
of q = 24 corresponds to 2/cf- 



B. Results 

Below we present the results obtained for N = 12 
electrons. The system length is L = 333 as, the in- 
teraction radius d — 33 as, with being effective 
Bohr's radius, V = 3.6. The corresponding value of the 
RPA parameter r s — (2nas)~ 1 is 13.9, the estimate of 
the Luttingcr liquid interaction parameter according to 
g = (1 + V q=0 /TThv F )-°- 5 gives 0.3. 

The distribution function n(q) over the single-particle 
states ijj q is shown in Fig. [T] Far from the Fermi surface, 
the form of the n(q) curve is smoothed by the interaction, 
which is familiar from the Luttinger model. However, 
right at the Fermi surface there appears an unexpected 
(5-type singularity, with the value of the distribution func- 
tion being close to 1 at this point. 

The presence of the singularity was checked by chang- 
ing the number N of electrons from 3 to 20, varying the 
length parameters and the interaction strength by the 
two orders of magnitude. The result proved to be per- 
fectly stable against the change in the system parameters 
(L,d,V, N). Hence, the (S-singularity in the momentum 
distribution n(q) at q = kp is a universal property of 
finite ID systems. Its origin is explained in section |TV) 

Electron density p(x) in the ground state is an oscil- 
latory function, the amplitude of which decays off the 
boundaries. To analyze the ordering in the electron state, 
consider the Fourier-transform of electron density p(q), 
which is shown in Fig. [2j For comparison, the results for 
non-interacting electrons are also provided. 

For free electrons, p(q) has a step-like structure, with 
p(q) = -0.5 for < q < 2k F , and p(q) = for q > 2k F . 
We emphasize that such threshold behavior holds even 
for a strongly interacting system, where p(q) remains very 
close to zero for q > 2k p. Electron-electron interaction 
modifies the values of density harmonics only for < q < 
2k F . 



The harmonics p(q) for < q < 2kp are suppressed by 
the interaction. The q = 2kp harmonic of the density is 
enhanced by interaction, reaching the values comparable 
to one half of the background density. This reflects the 
well-known fact that e-e interaction leads to the strong 
electron correlations on the scale of average inter-particle 
distance, or in other words, to the Wigner-type ordering 
in the system. 

The presented results are exact, since they are based 
on a precise many-particle wave function. In the next 
section we compare them to the results of the Luttingcr 
liquid theory. 



III. BOUNDED LUTTINGER LIQUID THEORY 

One of the most advanced analytical theories of a 
strongly correlated electron state in ID quantum dots 
is a Luttinger liquid theory, based o n the b osonization 
with zero (open) boundary conditions! 19 * 20 * 2 ^ In this the- 
ory, the spatial distribution of electron density and the 
distribution of electrons over single-particle states are ex- 
pressed through the Green function of chiral fcrmions 



G + (x,y) = (^+(y)^ + (x)) 



via 



n(q) = 



2L 



dxdyG + (x,y)e tq{y - 



(14) 



and p[x) = — e FX G+(— x, x)+c.c. The Green function 



G+(x,y) 



^{e 13 - 1)** 
f3L 



zM 9 ( 



2^r 



[(cosh 0- cos 2p) (cosh (3- cos ^VT ; 



cosh j3 



(a+i) 2 

4g 



cosh (3 



(15) 



4 




P(q) 



20 
10 


-10 
-20 



2k T 



FIG. 3: Momentum distribution in the Luttinger model with 
zero (filled circles) and periodic (empty circles) boundary con- 
ditions, the interaction parameter g = 0.3. 



FIG. 4: The Fourier-transform of electron density in the Lut- 
tinger model with zero boundary conditions, the interaction 
parameter g — 0.3. 



with the dimensionless cutoff parameter j3 ~ N 1 . 

The momentum distribution function is presented in 
Fig. [3j A comparison with exact diagonalization shows 
that this result is qualitatively incorrect. The momen- 
tum distribution function, calculated within bosoniza- 
tion, also has a singularity at q — kp, but instead of a 
single (5-peak at q = kp, n(q) deviates from a smoothed 
step in a finite band around kp, where the form of the 
curve is close to the derivative of the 5-peak!^ 



Electron density, calculated according to (151, equals 



p(x) 



N 

T 



l-25 S inh s (/3/2) 



with f(x) being 



cos(2kpx — 2f(x)) 
[cosh (3 - cos(27rx/L)]f/ 2 

(16) 



,, 1 sin(2-KX I L) 

fix) = - arctan — = ; — — 

JK ' 2 e0 - cos(27rar/L) 



(17) 



Its Fourier-transform is presented in Fig. Q. The qual- 
itative error of this result is that p(q) does not van- 
ish at q > 2k p, but, on the contrary, grows rapidly as 
q — » 2kp + 0. In other words, a whole branch appears in 
the region q > 2k p, which is absent in the exact solution. 

Thus bosonization breaks down for the momenta close 
to kp and multiples of it. This scale corresponds to 
the mean inter-particle distance. Hence, the bosoniza- 
tion with zero boundary conditions incorrectly treats the 
short-ranged electron correlations, responsible for the for- 
mation of the ordered, Wigner-like state in ID quantum 
dots. In fact, the results of this section describe the de- 
formed Wigner crystal. To demonstrate this, let us con- 
sider a simple model that takes into account the Wigner 
ordering from the very beginning. 



IV. INTERPRETATION OF THE RESULTS 

In this model, the many-electron wavefunction 
<J>(yi , ..,2/jv) is represented by the Slater determinant 



1 



'm 



<j>{Vi-x\) 



(j>{yi-x N ) 



(18) 



built on the single-particle wave functions 4>(x) = 
7r _1 / 4 a -1 / 2 exp(— x 2 /2a 2 ), localized at positions x^, k — 
1 . . . N. The wavefunction 'width' a is assumed to be 
smaller than the distance between the particles L/N, so 
the wavefunctions do not overlap and form an orthonor- 
mal set. 

The Fourier-transform of the density is 



P(?) 



JV 

£< 

k=l 



nqxk 
L 



(19) 



The momentum distribution function n(g) can be ex- 
pressed through the electron Green function, similarly to 



Eq. (14) 



n(q) 



dxdyG(x,y)ip*(x)ip q (y) 



(20) 



The Green function G(x,y) = (tl> + (y)'4'(x)} is related to 
the one-particle density matrix 



p(x,y) = / z 2 ,..,z N )$(x, z 2 , .., z N )dz 2 ..dz 



N ■ 



(21) 



via G(x, y) = Np(x, y) Substitute this into Eq. Q to 
get 



n(q) = Njdz J dy<$>*(y,z)*(; q (y) 



(22) 



5 



0.8 
0.6 
0.4 ■ 
0.2 ■ 




n{q) 



n(q) 



1 

0.8 
0.6 
0.4 
0.2 




2k r 



10 



15 



FIG. 5: Momentum distribution function in the model of lo- 
calized electrons. 



where z = (z 2l --,z N ). Using Eq. (18 1, we finally obtain 



(?) = E / dy ^ ( y ~ x k)^g(y) 
k 

N 

9 2 2 



(23) 



= A^Jn — e 
1^ 



fe=i 



2 irqxk 
L 



First consider the case of a Wigner crystal, i.e. when 
Xk = (k — h)jj- The Fourier-transform of the density 
p(q) is non-zero only for multiples of 2fef , 

p(n • 2fc F ) = (-l)"Ar e -« 2 4™ 2 . ( 2 4) 

The momentum distribution function n(q) equals 



n(q) 



q,mN 



exp 



L 2 



(25) 

Fig. [5] shows that n(q) is fully consistent with the re- 
sults of exact diagonalization, including the ^-singularity 
at the Fermi surface. Moreover, the additional 6- 
singularities at multiples of hp, predicted by the model, 
do exist in the exact solution for stronger interaction 
(flwO.l), see Fig. [6] 

This model proves that the origin of the singularities of 
n(q) is the ordering of electrons in bounded ID systems. 
In infinite ID systems, the ordering is destroyed by fluc- 
tuations, and short-range electron correlations manifest 
themselves only in dynamic response to external pertur- 
bation.^ If the system is finite, the boundaries pin the 
charge density waves, giving rise to Friedel oscillations, 
the amplitude of which is enhanced by e-e interaction. 
This results in the increase of the weight of the state with 
q = hp, which is reflected in the momentum distribution 
function, as well as in the high value of 2fc F -harmonic 
of density. The conclusion about the important role of 
short-range electron correlations in bounded ID systems 



FIG. 6: Momentum distribution function, calculated by exact 
diagonalization for the system of N = 5 electrons. The value 
of q — 5 corresponds to ftp- 




FIG. 7: Momentum distribution of the deformed Wigner crys- 
tal. 



P(q) 



20 
10 


-10 
-20 



FIG. 8: The Fourier-transform of electron density for the de- 
formed Wigner crystal. 



6 



is confirmed by the calculation of momentum distribution 
of electrons in a ring, i. e. in a system without boundaries. 
In this case the singularity of the momentum distribution 
disappears, as can be seen from Fig. Q3J) . 

Now let us squeeze electrons down to the center of the 
system by introducing a displacement Sx k oc — (k — 
The momentum distribution function and the density 



Fourier-spectrum, calculated according to ( 19]), (|23j) for 
this case, are presented in Figs. [7J [8j The curves are 
clearly identical to those from the previous section. This 
confirms our suggestion that bosonization describes the 
electron state that is deformed in comparison with the 
exact solution. 



V. BOSONIZATION WITH PARTICLE 
NUMBER CONSERVATION 

The origin of such deformation is quite simple. 
Bosonization introduces an uncompensated positive 
charge that attracts electrons to the center of the sys- 
tem. In order to demonstrate it, let us trace the steps 
that led to Eq. (pi 



Formalism 



In the theory of the bounded Luttinger liquids EL 9 I 2 ° I 21 I 
the electron field operator 



(26) 



is decomposed into the fields tp r (x) of the so-called r- 
fermions (r = ±), 



ip(x) 



Akpx 



where 



'2L 



E 



c k e 



The fields are not independent, since 
^+0) = -tp-(-x) , 



(27) 



(28) 



(29) 



so we will deal with ip+(x) alone. The latter has the 
property that 



1>+(L) = M-L) , 



(30) 



i.e. it satisfies periodic boundary conditions on the inter- 
val [— L, L]. Hence, it can be bosonized in a conventional 
way. 

The main assumptions of bosonization are the lin- 
earization of r-fcrmion spectrum near the Fermi surface 
and its extrapolation to infinity. They allow one to solve 
the model exactly. 



Introduce the r-fermion density operator, 
P(Q) = E : c i+g Ck := E { c t+g c k ~ 8q,o (4 c fe>) ( 31 ) 

k k 

for q 0. Zero harmonic p(q = 0) is the number of 
particles operator AN. The density operator obeys the 
commutation relation 



[p(q),p(-<i')] 



qL 



which allows one to introduce bosons 
bq =iJ—p q , q>0 



(32) 



(33) 



The r-fermion field operator has the following boson rep- 
resentation, 



where the bosonic phase cf> equals 



-iqx—a\q\ 



-Pi 



(34) 



(35) 



9#0 



U stands for the ladder operator, a denotes the ultravi- 
olet cutoff, which by the order of magnitude equals kp . 

The r-fermion spatial density is related to bosonic 
phase via 

p+(x) =: xl>X{x)M*) ^ + ^ . (36) 

and p~{x) = p+(—x). 

To obtain the density operator of real electrons, one 
has to express tj){x) in terms of ip r { x )- Using Eq. (27 1, 
one gets 



p{x) = pi w (x) + pcbw(x) , 



(37) 



where the first term is the long-wave component of the 
density, 



Piw(x) = p+(x) + p-(x) , 



(38) 



and the second term is the charge density wave (CDW) 
component 



Pcdw(x) = e- 2ik » x xl}%{x)i>-{x) + h.c. 



(39) 



which describes short-range electron correlations. In 
terms of bosonic phase, the electron density operator 
equals 

p(x) = —--^- — cos(2k F x-2cp(x)-2f(x)), (40) 

7T TT OX 7T 

where <p(x) — \ (<j>(— x) — <fi(x)), and function f(x) is an 
additional phase due to zero boundary conditions, 

1 



f(x) = - n [^(x),4>(-x)] 



4/ 



1 4 sin(27ra/i) 

2 e$ -cos(27ra;/L) 



(41) 



7 



with dimensionless cutoff f3 — wa/L « JV 1 . 
The kinetic energy 

Ho = v F ^2 k '■ c t c k '■ 



is bosonized using Kronig's identity to give 



H Q = vp^q : b+b q 

q>0 



TTVp 

^2L 



(AN)' 



(42) 



(43) 



The interaction part of the Hamiltonian equals 



V = - J dxdyp(x)V(x - y)p(y) , 



where 
ere 

taining only the direct 



e re p(x) is given by Eq. ( 40 1 . In the model consid- 
j| i9 | 20 | 2i ] ^ ne interaction operator is simpli 



(44) 
lsid- 

implified by re- 



f 



V d = ^ J dxd v (p+( x )p+(y) + p-{x)p-{v)) v{x - y) 

(45) 

and cross terms 

V c = ^ J dxdy (p+(x)p-(y) + p-(x)p+(y)) V(x - y) . 

The cross term contains the non-local contribution 
p+(x)p-(y) = pj r (x)pj r (—y), which makes the diagonal- 
ization of the Hamiltonian impossible, unless the inter- 
action potential is assumed short-ranged, that is of the 
form V8(x — y). Under this assumption, the direct and 
cross terms transform into 



V f 

Vd = y J ^dxp + (x) , 

V f L 

V c =— J ^dx p + (x)p+(-x) 



(47) 
(48) 



Combining the terms, and using bosonic representa- 
tion p6) of the density operator, we arrive at the follow- 
ing model Hamiltonian: 

H=H + V d + V c = {v F + y-)Y,1 (K b i + \ 



V 
4tt 



£ q (b+b+ + b q b q ) + ^ L (vf + ")(AA0 2 

9>0 



q>0 
TV 



(49) 



The Hamiltonian is diagonalized by Bogoliubov transfor- 
mation 



b q = e^bge 



(50) 



where 



q>0 



The Hamiltonian is diagonal if e 2 ^ = (1 + V/ttvf)^ 1 ^ 2 ■ 
The interaction parameter g = e 2 ^ belongs to (0,1) for 
repulsive interaction, and equals unity for free electrons. 
The diagonal form of the Hamiltonian is 



ff = X>(«) b 

q>0 



Pq 



2L 



v N (AN)< 



with 



v(q) 



v F + V/2tt V 

, v N = v F + — 

cosh 2£ 7r 



(52) 



(53) 



The bosonic representation (34 1 of the field operator, the 



density operator ( 40 1 and quadratic Hamiltonian ( 52 ) are 
sufficient to obtain the Green function (151 with all the 



ensuing problems for the momentum distribution (14 1 
and density (16 1. The problem lies in the violation of 



the particle number conservation within the bosonization 
approach. 



B. Density operator 

To see this, just notice that the integral of the density 
fluctuation (the second term in Eq. (16l) over the sys- 
tem length is not zero. This is a well-known problem, 
which exists not only in the case of zero boundary condi- 
tions, but also in a standard bosonization on the ring.ES 
The problem arises at the level of the electron density 
operator (40), the CDW component of which does not 



conserve the number of particles in an isolated system. 
The physical reason of the violation of the particle num- 
ber conservation is that the CDW component of (40 1 in- 
cludes the response of the infinite positron sea, which is 
not completely eliminated when using the approximate 
relation (27 1. 



In the case of zero boundary conditions, the situation 
is reacher because now there appear problems even with 



the long-wave component of the density. Indeed, Eq. ( 16 1 
of the common Luttinger liquid theory does not give a 
correct transition to the case of non-interacting electrons 
in the box. The density of free electrons 



Ptree(%) 



N 



1 

2L 



sin(2k F + j)x 
2sinf^ 



(54) 



contains an additional term 1/2L, missing in (16 1. This 



term, being integrated over the length of the system, 
gives an extra charge of e/2. 

Thus the microscopic theory leads to the density oper- 
ator that violates the particle number conservation. We 
will obtain the correct operator, following the harmonic- 
fluid approach by HaldanePSl 

Let us introduce the phase 9{x) that increases by 7r 
each time x passes the location Xk of a A:-th particle. 
Then the particle density operator pi becomes 



JY 



(51) p(x) = d x 6j2 S{9(x)-kw) = d x 9 ^ e 



2im8(x) 



(55) 



k=l 



8 



n{q) 



2kv 



0.5 
0.4 
0.3 
0.2 
0.1 





-2 
-4 
-6 



P(q) 



FIG. 9: Momentum distribution function, the violation of the 
particle number conservation is fixed. 



FIG. 10: The Fourier-transform of electron density, the vio- 
lation of the particle number conservation is fixed. 



According to Eq. ( 40 ) , 



0(x) = 2k F x - 2ip(x) - 2/0) 
Thus the density operator takes the form 
k F d x ip d x f 



(56) 



p(x) 



IT IT IT 

-M.) C0S (2k F x - 2<p(x) - 2f(x)), 

7T 7T 7T 



(57) 



where we retain only m = 0, ±1 harmonics, and halve the 
amplitude of the CDW component to obtain the correct 
transition to the non-interacting case^l 



The density operator of Eq. (57 1 has the form of a 



full differential, which guarantees that the integral of the 
density fluctuation over the length of the system is zero, 
i.e. the number of particles is conserved. The long- wave 
part of the operator contains an additional term — d x f /tt 
that gives the 1/2L component, missing in (16 1, since in 
the bulk of the system f(x) s» 



77.; 

2L- 



Hamiltonian and observables 



The interaction part (44) of the Hamiltonian, having 



been calculated with the density operator of Eq. (57 1, 
gets additional terms of the forrrp3 



V 



2tt 2 

By shifting bosons 



[ dxd x ipd x f = iV^J^=-(b% q -b 2q ). 



q>0 



d 2q = b 2q + i 



9° 

4u F V 2irqL 



(58) 



(59) 



we obtain a diagonalizcd full Hamiltonian H + Hi , 



H + H 1 = J2'lv(q)d+d q . 

q>0 



(60) 



The bosonic phase ( |35[ ) is transformed into 4>(x) = 
4>o(x) + (f>i(x), where 4>o{x) is linear in new bosons, 

c -se- tqx )d q + {ce- iqx -se lqx )d+] , 

(61) 

with c and s being, respectively, cosh£ and sinh£. The 
function 4>\{x) is the new phase, specific to the case 
of zero boundary conditions, 4>\{x) = Af(x), A — 
Vg 2 /2irv F . 




As a result, the field operator (34 1 acquires the factor 



of exp(iAf(x)), and the Green function (151 transforms 
as 

G ncw (x, y) = G+(x, y)e iA «M-Hv» . (62) 
The average value of the density equals 

(p(x)) = ^-(l-A) d *l 

7T 7T 

sinh 9 (/3/2) d sin{2k F x - 2(1 - A)f(x)) 



2 1 -9/ 2 tt dx [cosh /3 - cos ^] a/ 2 



(63) 



Note that the additional phase (1 — A)f(x), which ap- 
pears in Eq. ( 63 1 , changes the period of density oscil- 
lations. Since d x f < everywhere in the system ex- 
cept narrow regions near the ends, the Wigner crystal is 
squeezed by the boundaries. The deformation is deter- 
mined by the coefficient (1 — A). In the common Lut- 
tinger liquid theory A = 0. Our approach with corrected 
expression (57) for the density operator gives A > 0. 



Hence, restoring total neutrality results in the reduction 
of the Wigner crystal compression. 



9 




XIL 



FIG. 11: The spatial distribution of the density, calculated 



according to Eqs. (16 1 (dashed line) , Eq. (541 (thin solid line), 
Eq. |63f (solid lineK 



The distribution function n(q) and the Fourier spec- 
trum of the density calculated with the use of the above 
expression for A coincide with the numerical calculations 
quite well. However this formula is justified for weak 
interaction (1 — g <C 1). We propose a generalized ex- 
pression for any g taking into account that A is known 
for three limiting cases: 

i) For g = 1, A = to provide the correct transition 
to the case of non-interacting electrons. 

ii) For weak interaction, A should be proportional to 
A = V/2t:vf, in agreement with our model. 

iii) For strong interaction g — > 0, A — > 1 to provide the 
transition to the limiting case of the Wigner crystal with 
strictly periodic density. 

The simplest choice of A = l—g satisfies these require- 
ments and proves to be highly successful, as is demon- 
strated below. Figs. [9j [10] show the momentum distri- 
bution function and the density Fourier-transform, cal- 
culated according to Eqs. (62 1, (63 1. They are seen to 



agree nicely with the exact results of section |XTJ 

Fig. [TT] shows the spatial distribution of the density, 
calculated according to Eqs. (16 1, (54), (63). It is seen 



that according to the standard bosonization the electron 
density maxima almost coincide with those of free elec- 
trons, even for strong interaction. In contrast, accord- 
ing to Eq. ( 63 1 the electron locations are shifted towards 



the periodic positions as Wigner ordering prescribes at 
strong interaction. Thus the bosonization approach that 
does not respect the particle number conservation leads 
to the picture of the state with electrons squeezed to the 
center of the system. 



VI. CONCLUSION 

To conclude, we investigated the ground state of in- 
teracting electrons in a ID quantum dot using the ex- 



act diagonalization. An unexpected <5-like singularity 
was discovered in the momentum distribution function 
at the Fermi energy. A threshold behavior was found for 
the spatial Fourier-spectrum of electron density, with the 
step at 2kp . These effects are stable against the change 
in the system length, interaction radius, the number of 
electrons, and the interaction strength. Thus we suggest 
that they are inherent in finite ID electronic systems. We 
proposed a simple model which proved that these effects 
originate from the formation of the Wigner molecule in a 
ID quantum dot. Comparison of exact results with the 
Luttingcr model with zero boundary conditions shows 
that the latter does not correctly describe both the mo- 
mentum distribution near the Fermi energy and the den- 
sity Fourier-spectrum. The problem is that bosoniza- 
tion overestimates the deformation of the Wigner crystal 
caused by the boundaries by introducing the fake posi- 
tive charge into the ID system, which squeezes electrons 
to the center of the system by attractive forces. This 
is a result of using the density operator that violates the 
number of particles conservation. We derived the density 
operator devoid of the mentioned shortcoming, corrected 
the Hamiltonian, and calculated the observables to find 
a nice agreement with the exact results. 



Acknowledgments 

This work was supported by Russian Foundation for 
Basic Research (project No. 05-02-16854), Russian 
Academy of Sciences (programs Quantum Nanostruc- 
tures and Strongly Correlated Electrons in Semiconduc- 
tors, Metals, Superconductors, and Magnetic Materials), 
RF Ministry of Education and Science, and Russian Sci- 
ence Support Foundation. 



APPENDIX A: EXACT DIAGONALIZATION 
METHODOLOGY 

The solution of the non-interacting problem is the 
Slater determinant Q built from N lowest energy single- 
particle states. For interacting electrons, the expan- 
sion (3j) of the many-particle wave function contains in- 
finitely many terms, because the interaction scatters elec- 
trons among the excited states. The coefficients in ^ 
tend to zero for high-energy states. Therefore it is suf- 
ficient to retain only the terms with energies E p < A, 
where A is the energy cut-off, by the order of magnitude 
defined by e-e interaction energy. The specific value of 
A should be determined from the requirement that the 
observables reach the saturation with the increase of A. 

Once A is fixed, the number of functions in expan- 
sion (|3| and the size of the Hamiltonian matrix are de- 
termined. The task then reduces to the calculation of the 
Hamiltonian matrix elements. 



#piP2 = / $p 1 H$ P2 dxi..dx N 



(Al) 



10 



The matrix elements are calculated by substituting 
Eqs. @, Q into Eq. (KJ\ , expanding the Slater deter- 
minants, and using the orthogonality property of single- 
particle functions ^ during integration. 

The matrix clement of the kinetic energy equals 



_ h 2 TT 2 

FpiP2 ~ 2^L2 



P2 • 



(A2) 



For the matrix elements of the pair e-e interaction, 
four situations are possible, depending on pi and P2. 

1. If Pi = P2 = {ot\, --iOln), i-e. diagonal matrix ele- 
ments arc calculated, then 



N 



^PlPl — ^QiQj.QiQj 7 

1>J = 1 



(A3) 



where 

( f>*a ia] (xi,X 2 )V(x 1 ~ X 2 )(f>a k a l (x 1 ,X 2 )dx 1 dx 2 , 



and 

<t>ot i a j (Xl,X2) 



1pa z (x 2 ) lp aj (x 2 ) 



(A4) 



(A5) 



2. If pi and P2 are identical except two states, 
numbered k\ and fc 2 , respectively, i.e. if pi = 
{ax, -.a*!, ..ajv), P2 = («i, --Pk 2 , -ajv), then 



JV 



(_!)*+<* £ y QfciQiA2Qi , (A6) 



i=l 



3. If pi and p 2 are identical except four states, 
numbered ki, i — 1 . . . 4, i.e. if pi = 

(ax,..ajfe 1 ,..ajfe a ,..ajv), P2 = (ai, -A 3 , • A 4 ) -a/v), 
then 



^PiP2 — ( I) 1 1 V akl a k2 ,0 k3 l3, 



(A7) 



4. If pi and p 2 contain more than four not coincident 
states aj, then 



Vp lP2 = 



(A8) 



Matrix elements of the interaction between electrons 
and positive background are as follows. 



1. If pi = p 2 = (ai, ..,a N ), then 

N 

fpipi = y ] /a,o, j 



where 



fa iak = / 1pai( x W(x)ljj ak (x)dx . 



(A9) 



(A10) 



2. If p! and P2 are identical except two states, 
numbered k\ and k 2 , respectively, i.e. if pi = 
(ai, ..a kl , ..ajv), P2 = ("l, ../3k a) -ajy), then 



U PlP2 - (-l) fel+fe2 /a fcl/3 . 



(All) 



3. Otherwise, U PlP2 = 0. 



The matrix element of the pair e-e interaction (A4| 
equals 



Ve 2 d 



[f(l,j,k,m) - f(l,j,k,-m)- 



lj,km ~ 2 eL 2 

f(l,j, ~k,m) + f(l,j, -k, -m) - f(l, -j, k,m)+ 
f(l, -j, k , -m) + f(l, -j, -k, m) - f(l, -j, -k, -m)- 
f(j, I, k, m) + f(j, I, k, -m) + f(j, I, -k, m)- 
f(j, I, -k, -m) + f(j, -I, k, m) - f(j, -I, k, -m)- 
f(j,-l,-k,m) + f(j,-l,-k,-m)], 

(A12) 

where f(l, n, k, m) = g(l + k,n + m), and g(p, q) = for 
odd (p + q) , while for even (p + q) 



g{ Pl q) = {e- a {-lY -1) 



1 



rPQ 



(l + ^p 2 )(l + ^q 2 ) 
a , 

°p, — q 1 



(A13) 



with a = L/d. Matrix element (AlOl equals 



Ve 2 d 

fin = N—{g{l + n, 0) - g(l - n, 0)) . (AH) 
eL z 



1 L.P. Kouwenhoven, D.G. Austing, and S. Tarucha, Rep. 
Prog. Phys. 64, 701 (2001). 

2 S.M. Reimann and M. Manninen, Rev. Mod. Phys. 74, 
1283 (2002). 



3 G.A. Fiete, J. Qian, Ya. Tserkovnyak, and B.I. Halperin, 
Phys. Rev. B 72, 045315 (2005). 

4 H. Steinberg, O.M. Auslaender, A. Yacoby, J. Qian, 
G.A. Fiete, Y. Tserkovnyak, B.I. Halperin, K.W. Baldwin, 



11 



L.N. Pfeiffer, and K.W. West, Phys. Rev. B 73, 113307 
(2006). 

5 O.M. Auslaender, H. Steinberg, A. Yacoby, 
Y. Tserkovnyak, B.I. Halperin, K.W. Baldwin, L.N. Pfeif- 
fer, and K.W. West, Science 308, 88 (2005). 

6 F.D.M. Haldane, J. Phys. C 14, 2585 (1981). 

7 J. Voit, Rep. Prog. Phys. 58, 977 (1995). 

8 H.J. Schulz, Phys. Rev. Lett. 71, 1864 (1993). 

9 C.L. Kane and M.P.A. Fisher, Phys. Rev. B 46, 15233 
(1992). 

10 S. Tarucha, T. Honda, and T. Saku, Solid State Commun. 
94, 413 (1995). 

11 D.L. Maslov and M. Stone, Phys. Rev. B 52, R5539 (1995). 

12 V.A. Sablikov and B.S. Shchamkhalova, Phys. Rev. B 58, 
13847 (1998). 

13 C.-K. Wang and K.-F. Berggren, Phys. Rev. B 57, 4552 
(1998). 

14 L.P. Rokhinson, L.N. Pfeiffer, and K.W. West, Phys. Rev. 
Lett. 96, 156602 (2006). 

15 K.J. Thomas, J.T. Nicholls, M.Y. Simmons, M. Pepper, 
D.R. Mace, and D.A. Ritchie, Phys. Rev. Lett. 77, 135 
(1996). 

16 R. Crook, J. Prance, K.J. Thomas, S.J. Chorley, I. Farrer, 
D.A. Ritchie, M. Pepper, and C.G. Smith, Science 312, 
1359 (2006). 

17 A. Shailos, A. Ashok, J.P. Bird, R. Akis, D.K. Ferry, 
S.M. Goodnick, M.P. Lilly, J.L. Reno, and J. A. Simmons, 
J. Phys.: Condens. Matter 18, 1715 (2006). 

18 R. de Picciotto, L. N. Pfeiffer, K.W. Baldwin, and 
K.W. West, Phys. Rev. Lett. 92, 036805 (2004). 

19 M. Fabrizio and A.O. Gogolin, Phys. Rev. B 51, 17827 
(1995). 

20 A.E. Mattsson, S. Eggert, and H. Johannesson, Phys. Rev. 
B 56, 15615 (1997). 

21 J. Voit, Y. Wang, and M. Grioni, Phys. Rev. B 61, 7930 
(2000). 

22 V.A. Sablikov and Y. Gindikin, Phys. Rev. B 61, 12766 
(2000). 

23 O.A. Starykh, D.L. Maslov, W. Hausler, and L.I. Glazman, 
in Interaction and Quantum Transport Properties of Lower 
Dimensional Systems, Lect. Notes Phys., Vol. 544, p. 37, 
edited by T. Brandeis, Springer, New York, (2000). 

24 M. Khodas, M. Pustilnik, A. Kamenev, and L.I. Glazman, 
|cond-mat/0702505| 

26 G.H. Golub and CF. van Loan, Matrix computations, 
Johns Hopkins University Press, Baltimore, (1996). 

26 Since bosonization is applicable only for low-energy ex- 
citations, the results for n(q) can only be considered in 
the vicinity of the Fermi surface, at \kF — q\ <C fcf. The 
suppression of n(q) at &f — q ~ &f is the consequence 
of ultraviolet cutoffs, performed in the theory to exclude 
the high-energy contribution of the positron sea. Cutoffs 
should not affect the low-energy region near kp. 

27 G.D. Mahan, Many- Particle Physics, 2nd ed., Plenum 
Press, New York, (1990). 

28 F.D.M. Haldane, Phys. Rev. Lett. 47, 1840 (1981). 

29 The terms linear in the bosons for the Hamiltonian of elec- 
trons in a ID box were also obtained in by different argu- 
ments in V. Meden, W. Metzner, U. Schollwock, O. Schnei- 
der, T. Stauber, and K. Schonhammer, Eur. Phys. J., 16, 
631 (2000).