close

Вход

Забыли?

вход по аккаунту

?

Conductance distribution at criticality one-dimensional Anderson model with random long-range hopping.

код для вставкиСкачать
Ann. Phys. (Berlin) 18, No. 12, 891 – 895 (2009) / DOI 10.1002/andp.200910390
Conductance distribution at criticality:
one-dimensional Anderson model with random long-range
hopping
J. A. Méndez-Bermúdez1,∗ , Victor A. Gopar2 , and Imre Varga3,4
1
2
3
4
Instituto de Fı́sica, Universidad Autónoma de Puebla, Apartado Postal J-48, Puebla 72570, Mexico
Departamento de Fı́sica Teórica and Instituto de Biocomputación y Fı́sica de Sistemas Complejos (BIFI),
Universidad de Zaragoza, Pedro Cerbuna 12, 50009 Zaragoza, Spain
Elméleti Fizika Tanszék, Fizikai Intézet, Budapesti Műszaki és Gazdaságtudományi Egyetem,
1521 Budapest, Hungary
Fachbereich Physik und Wissenschaftliches Zentrum für Materialwissenschaften, Philipps Universität
Marburg, 35032 Marburg, Germany
Received 1 September 2009, accepted 17 September 2009
Published online 11 December 2009
Key words Metal-insulator transition, Anderson model, electronic transport, random matrix theory.
PACS 03.65.Nk, 71.30.+h, 73.23.-b
We study numerically the conductance distribution function w(T ) for the one-dimensional Anderson model
with random long-range hopping described by the Power-law Banded Random Matrix model at criticality. We concentrate on the case of two single-channel leads attached to the system. We observe a smooth
transition from localized to delocalized behavior in the conductance distribution by increasing b, the effective bandwidth of the model. Also, for b < 1 we show that w(ln T /Ttyp ) is scale invariant, where
Ttyp = expln T is the typical value of T . Moreover, we find that for T < Ttyp , w(ln T /Ttyp ) shows a
universal behavior proportional to (T /Ttyp )−1/2 .
c 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
The study of transport properties of systems at the Anderson metal-insulator transition (MIT) has been a
subject of intensive research activity for several decades. In particular, the interest has been focused on the
conductance T and its corresponding probability distribution w(T ) [1–8]. At the MIT, w(T ) has been found
to be universal, i. e. size independent, but dependent on the given model, dimensionality, symmetry, and
even boundary conditions of the system. w(T ) has been studied for systems in two and more dimensions
with a large number of attached single-channel leads. In fact, concerning T and w(T ), the regime of small
number of leads and the case of one-dimensional (1D) systems have been left almost unexplored [8–10].
In the present work we study numerically w(T ) for a 1D system at the MIT described by the Powerlaw Banded Random Matrix (PBRM) model at criticality with two single-channel leads attached to it (the
minimal configuration needed to measure T ).
The PBRM model [11, 12] at criticality describes 1D tight-binding samples with random long-range
hopping of length L represented by L × L real matrices whose elements are statistically independent
random variables drawn from a normal distribution with zero mean, Hij = 0, and a variance decaying as
a power law |Hij |2 ∼ (b/|i − j|)2 , where b is a parameter. There are two prescriptions for the variance
of the PBRM model: the so-called non-periodic,
|Hij |2 =
∗
Corresponding author
1 + δij
1
,
2 1 + (|i − j|/b)2
(1)
E-mail: jmendezb@venus.ifuap.buap.mx, Phone: +52 222 2295610, Fax: +52 222 2295611
c 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
892
A. Méndez-Bermúdez et al.: Scattering at the MIT
where the 1D sample is in a line geometry; and the periodic,
|Hij |2 =
1 + δij
1
,
2 1 + [sin (π|i − j|/L) /(πb/L)]2
(2)
where the sample is in a ring geometry. Field-theoretical considerations [11–13] and detailed numerical
investigations [12, 14, 15] verified that the PBRM model shows all the key features of the Anderson MIT
at the critical point. Thus the PBRM model possesses a line of critical points b ∈ (0, ∞).
Using standard methods [16, 17] we attach two semi-infinite single-channel leads to the 1D sample described by the non-periodic and the periodic versions of the PBRM model at criticality. Then, we calculate
the 2 × 2 scattering matrix in the standard form
r t
.
S(E) =
t r
The leads are attached to the first two sites of the sample. That is, in the non-periodic version of the PBRM
model, Eq. (2), we attach the leads at the boundary of the system. While in the periodic version, Eq. (1),
we attach them to the bulk. In the latter case, finite size effects are considerably reduced. However, the
quantities we analyze below are L-independent once L is much larger than the number of attached leads
for both versions of the PBRM model.
Once the scattering matrix is calculated we can compute the dimensionless conductance T = Tr(tt† )
and its probability distribution w(T ). Notice that for b → ∞, H reproduces the Gaussian Orthogonal
Ensemble. Then, in that limit we expect to recover the Random Matrix Theory (RMT) prediction for
w(T ) [18]:
1
w(T )RMT = √ ,
2 T
(3)
which is valid in the absence of direct processes.
In the following we focus on w(T ) for the PBRM model at criticality. We will compare RMT predictions
(at least for large enough b) with numerical simulations, where the statistics is collected by sampling over
different disorder realizations. All quantities reported here were obtained for L = 50 with 106 ensemble
realizations. Moreover, we have verified that our results are invariant when increasing L.
For b < 1, w(T ) is highly concentrated close to T = 0. So it is more convenient to analyze w(ln T )
instead. In Fig. 1(Left)(a) we show w(ln T ) for several values of b. Notice that the distribution functions w(ln T ) do not change their shape or width by varying b, thus being scale invariant. In fact, ln T clearly displays a linear behavior when plotted as a function of ln b, see Fig. 1(Left)(b). Then, all distributions w(ln T ) fall one on top of the other when shifting them along the x-axis by the typical value of
T , Ttyp = expln T ∝ b2 , as shown in Fig. 1(Left)(c). In Fig. 1(Right) we present the same quantities
as in Fig. 1(Left) but now for the non-periodic version of the PBRM model at criticality. We observe very
similar results in both cases.
Notice that for T < Ttyp , w(ln T ) is proportional to (T /Ttyp )−1/2 , see panels (c) in Fig. 1. Moreover,
we found that this behavior is universal
∞ for the PBRM model as can be seen in Fig. 2, where we plotted the
cumulative distribution wC (x) = x w(x )dx , x = ln T /Ttyp , for several small and large values of b.
In Fig. 3 we show w(T ) for large b (b ≥ 0.4) for both, the periodic and non-periodic versions of the
PBRM model at criticality. In the limit b → ∞, w(T ) is expected to approach the RMT prediction of
Eq. (3). However, once b ≥ 4, w(T ) is already well described by w(T )RMT . Moreover, we have found that
w(T ) ∝ T µ for T ∈ [0.5, 1], where, according to the RMT prediction, μ → −0.5 as b → ∞. In Fig. 4(a)
>
we plot μ as a function of b as measured from the histograms of Fig. 3. We noticed that for b ∼ 1, μ is
−2
proportional to b . This observation together with the analytical estimation for the correlation dimension
of the eigenfunctions [12] D2 (b) = 1 − (2πb)−1 , b 1, allowed us to conclude that μ(D2 ) ∝ (1 − D2 )2
c 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.ann-phys.org
-2
893
(a)
ln w(ln T)
ln w(ln T)
Ann. Phys. (Berlin) 18, No. 12 (2009)
-4
-6
-8
(a)
-4
-6
-8
-20
-25
-10
-15
0
-5
2
~ ln b
-10
0.01
(b)
1
0.1
<ln T>
-5
-20
-25
ln T
0
<ln T>
-2
0
-5
2
~ ln b
-10
10
0.01
-6
(b)
1
0.1
10
b
(c)
~ 1/2 ln(T/Ttyp)
ln w(ln T)
ln w(ln T)
-4
b = 0.01
b = 0.02
b = 0.04
b = 0.1
b = 0.2
b = 0.4
0
-5
ln T
b
-2
-10
-15
-8
-2
-4
-6
b = 0.01
b = 0.02
b = 0.04
b = 0.1
b = 0.2
b = 0.4
(c)
~ 1/2 ln(T/Ttyp)
-8
-10
-15
-5
0
5
-15
ln (T/Ttyp)
-10
-5
0
5
ln (T/Ttyp)
(Left)
(Right)
0
-3
~ 1/2 ln(T/Ttyp)
-6
-15
-10
-5
0
b = 0.01
b = 0.02
b = 0.04
b = 0.1
b = 0.2
b = 0.4
b=1
b=2
b=4
b = 10
ln wc(ln T/Ttyp)
ln wc(ln T/Ttyp)
Fig. 1 (online colour at: www.ann-phys.org) Left [Right]: (a) Conductance probability distribution ln w(ln T ) for the
periodic [non-periodic] PBRM model at criticality for several values of b < 1. (b) ln T as a function of b (symbols).
The black dashed line is the best fit of the data for b < 1 to the logarithmic function A + ln b2 , with A ≈ −1.35.
[A ≈ −1.44]. (c) ln w(ln T ) for b < 1 scaled to Ttyp = expln T ∼ b2 . The black dashed line has slope 1/2 and is
plotted to guide the eye.
0
-4
~ 1/2 ln(T/Ttyp)
-8
-18
5
-12
-6
0
6
ln (T/Ttyp)
ln (T/Ttyp)
(Left)
b = 0.01
b = 0.02
b = 0.04
b = 0.1
b = 0.2
b = 0.4
b=1
b=2
b=4
b = 10
(Right)
Fig. 2 (online colour at: www.ann-phys.org) Left [Right]: Cumulative conductance distribution ln wC (ln T /Ttyp )
for the periodic [non-periodic] PBRM model at criticality for several values of b. The black dashed line has slope 1/2
and is plotted to guide the eye.
www.ann-phys.org
c 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
894
A. Méndez-Bermúdez et al.: Scattering at the MIT
w(T)
4
4
b = 0.4
2
0
4
b=1
2
0
2
0
1 0
0.5
4
b=2
0.5
T
2
0
1 0
T
b=4
0
1 0
0.5
T
0.5
1
T
Fig. 3 (online colour at: www.ann-phys.org) Black [red] histograms: Conductance probability distribution
w(T ) for the periodic [non-periodic] PBRM model at criticality for several values of b. Blue dashed lines
are the RMT prediction for w(T ) given in Eq. (3).
-1
-1
-2
~ (1-D2)
μ
~b
-2
-2
(a)
-3
2
1
10
b
(b)
-3
0.5
0.75
D2
1
Fig. 4 (online colour at: www.ann-phys.org) Downwards (upwards) triangles: µ for the periodic (nonperiodic) PBRM model at criticality as a function of (a) b and (b) D2 . µ is extracted from the decay w(T ) ∝
T µ in the interval T = [0.5, 1]. Blue dashed lines indicate the RMT prediction µ = −0.5, expected for
b → ∞. Black dashed lines are best fits proportional to (a) b−2 and (b) (1 − D2 )2 . Error bars are not shown
since they are much smaller than symbol size.
for D2 approaching unity, as can be seen in Fig. 4(b). Thus, μ(D2 ) provides the possibility of evaluating
D2 by means of scattering experiments.
To conclude, we observed a transition from localized- to delocalized-like behavior in the conductance
distribution of the PBRM model by moving b from small (b 1) to large (b > 1) values. For small b,
we showed that w(ln T ) is scale invariant with the typical value of T , Ttyp , as scaling factor. Also, for
T < Ttyp and irrespective of b, we found the universal behavior w(ln T /Ttyp ) ∼ (T /Ttyp)−1/2 . Finally,
we showed that the RMT limit, expected for b → ∞, is already recovered for relatively small values of b:
b ≥ 4. Our conclusions are valid for leads attached to the bulk of the system as well as for leads attached
at the boundary.
Acknowledgements This work was supported by the Hungarian-Mexican Intergovernmental S & T Cooperation
Programme under grants MX-16/2007 (NKTH) and I0110/127/08 (CONACyT). Support of the Hungarian Research
Fund OTKA under contracts 73381 and 75529 is also greatfully acknowledged. V. A. G. acknowledges support from
the Ministerio de Educación y Ciencia through the Ramón y Cajal Program and European Social Fund.
References
[1] B. Shapiro, Phys. Rev. Lett. 65, 1510 (1990).
[2] P. Markoš, Europhys. Lett. 26, 431 (1994); Phys. Rev. Lett. 83, 588 (1999).
c 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.ann-phys.org
Ann. Phys. (Berlin) 18, No. 12 (2009)
895
[3] K. Slevin and T. Ohtsuki, Phys. Rev. Lett. 78, 4083 (1997); K. Slevin, T. Ohtsuki, and Kawarabayashi, ibid. 84,
3915 (2000); K. Slevin, P. Markoš, and T. Ohtsuki, ibid. 86, 3594 (2001).
[4] X. Wang, Q. Li, and C. M. K. Soukoulis, Phys. Rev. B 58, 3576 (1998).
[5] M. Rühländer and C. M. Soukoulis, Physica B 296, 32 (2001); M. Rühländer, P. Markoš, and C. M. Soukoulis,
Phys. Rev. B 64, 172202, 212202 (2001).
[6] I. Travěnec and P. Markoš, Phys. Rev. B 65, 113109 (2002).
[7] L. Schweitzer and P. Markoš, Phys. Rev. Lett. 95, 256805 (2005); J. Phys. A: Math. Gen. 39, 3221 (2006).
[8] M. Jansen, M. Metzler, and A. R. Zirnbauer, Phys. Rev. B 59, 15836 (1999).
[9] K. Senouci and N. Zekri, Phys. Rev. B 66, 212201 (2002).
[10] C. Monthus and T. Garel, Phys. Rev. B 79, 205120 (2009); J. Stat. Mech. P07033 (2009).
[11] A. D. Mirlin, Y. V. Fyodorov, F.-M. Dittes, J. Quezada, and T. H. Seligman, Phys. Rev. E 54, 3221 (1996).
[12] F. Evers and A. D. Mirlin, Rev. Mod. Phys. 80, 1355 (2008).
[13] V. E. Kravtsov and K. A. Muttalib, Phys. Rev. Lett. 79, 1913 (1997); V. E. Kravtsov and A. M. Tsvelik, Phys.
Rev. B 62, 9888 (2000).
[14] E. Cuevas, M. Ortuno, V. Gasparian, and A. Perez-Garrido, Phys. Rev. Lett. 88, 016401 (2002).
[15] I. Varga, Phys. Rev. B 66, 094201 (2002); I. Varga and D. Braun, ibid. 61, R11859 (2000).
[16] C. Mahaux and H. A. Weidenmüller, Shell Model Approach in Nuclear Reactions (North-Holland, Amsterdam,
1969); J. J. M. Verbaarschot, H. A. Weidenmüller, and M. R. Zirnbauer, Phys. Rep. 129, 367 (1985).
[17] J. A. Méndez-Bermúdez and I. Varga, Phys. Rev. B 74, 125114 (2006).
[18] P. A. Mello and N. Kumar, Quantum Transport in Mesoscopic Systems (Oxford University Press, Oxford,
2004).
www.ann-phys.org
c 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Документ
Категория
Без категории
Просмотров
1
Размер файла
200 Кб
Теги
random, anderson, dimensions, criticality, distributions, mode, one, long, conductance, ranger, hopping
1/--страниц
Пожаловаться на содержимое документа