arXiv:cond-mat/9907153v1 [cond-mat.

mes-hall] 9 Jul 1999

Quantum Hall effect today

V. J. Goldman 1
Department of Physics, SUNY, Stony Brook, New York 11794-3800, U.S.A.

Abstract
I present a brief survey of important recent developments in the quantum Hall effect. The review covers both
fractional and integer regimes, from an experimentalists perspective. The topics include direct measurements of
fractional charge, composite fermion Fermi surface, spin textures, and edge state (chiral Luttinger liquid) dynamics.
Key words: quantum Hall effect; integer; fractional

1. Introduction. The last two decades since the

discovery of the integer [1] and fractional [2] quantum Hall effects (QHE) have witnessed remarkable
increase in our understanding of low-dimensional
strongly correlated systems. Many surprising and
many beautiful experiments were performed (some
of which are reviewed here), new theoretical understanding was gained, new elegant and sophisticated techniques were developed, both experimental and theoretical.[3] 1985 and 1998 Nobel prizes
in physics [4] were awarded to von Klitzing, and
Tsui, Stormer and Laughlin for the discoveries.
The standard of resistance is now based on the
integer QHE as h/e2 = 25, 812.807 exactly. The
exactness of quantization of the Hall conductance
is understood as a consequence of the gauge invariance of electromagnetic field and the quantization of the charge of current carriers.[5,6] Consider
a gedanken experiment on a Corbino disc of twodimensional (2D) electrons of areal density n with
quantizing magnetic field B applied normal to the
1

E-mail: [email protected]

Preprint submitted to Physica B

disc, so that Landau level filling is = hn/eB. At

a low temperature T the system of electrons condenses into a QH state, integer or fractional, depending on and on how much disorder is present.
First consider the case of the integer QHE, where
the integer i = 1, 2, 3, ... is the number of occupied Landau levels; exact filling occurs at = i.
Now add adiabatically a flux quantum 0 = h/e
in the inner hole of the disc; while 0 is added one
electron per occupied Landau level is transferred
between the inner and the outer edges of the disc
(provided they are connected by a wire). Since 0
is added in the hole, the state of the electron system must be exactly the same as that before flux
was added (gauge invariance). Thus, adding 0 every t, the Hall current is ie/t, the voltage 0 /t,
and the Hall resistance Rxy = 0 /ie (in 2D xy =
Rxy ). What is important here is that Rxy remains
quantized exactly even as is varied from the exact
filling, because disorder localizes extra electrons or
holes and the diagonal conductivity xx = 0.
i
can be
The fractional QHE at = f = 2pi+1
24 July 2011

mapped onto the integer case using composite

fermion (CF) theory.[79] A composite fermion
is an electron bound to an even number 2p (p =
1, 2, 3, ...) of vortices of the many-particle wave
function. The binding results from Coulomb interaction between the electrons, and it has been
shown that the exact FQH ground states are very
close to those in the CF theory, also the CF theory
predicts the hierarchy of the FQH states observed
in nature.[10] Since on the average, in an area,
the number of vortices of the many-particle wave
function is equal to the number of the flux quanta,
in mean field theory CFs can often be thought of
as electrons each binding 2p0 of applied B. Thus
CFs experience effective magnetic field Bcf =
B 2pn0 and the filling of the CF pseudo Landau
levels cf = n0 /Bcf gives = cf /(2pcf + 1).
For p = 1, for example, Bcf = 0 occurs at = 21 ,
thus the system of interacting electrons looks like
zero-field metal of CFs. Also, the FQHE of interi
acting electrons at = 2i+1
maps onto the IQHE
of non-interacting CFs at cf = i, with i pseudo
Landau levels occupied by CFs.

duction is possible at T = 0, xx = 0. At a finite

T quasiparticles are thermally excited in electronhole pairs across the gap while maintaining overall
charge neutrality. The thermally excited quasiparticles can carry dissipative current and, because
xx is approximately proportional to their concentration, the activated transport experiments are
used to determine the energy gap. The integer gaps
were thus measured in a variety of samples in different materials, and a number of fractional gaps, in
5 4 3 2
, 9 , 7 , 5 and 31 .[3,11]
particular at f = 23 , 35 , 47 , 95 , 11
1
The 3 FQH gap was recently measured in inelastic light scattering experiments, a technique which
also allowed to trace magnetoroton minimum in
the excitation spectrum at finite wave vector.[12]
3. Direct observation of fractional charge.
On a QH plateau quasiparticles are localized and
their charge is well defined. In the case of the integer QH plateau with xy = h/ie2 the quasielectrons are simply electrons in the (i + 1)st Landau
level, and the quasiholes are the holes (unoccupied
states) in the ith level. In the case of the FQH
i
plateau at f = 2pi+1
quasielectrons are CFs (an
electron binding 2p vortices) in the (i + 1)st Landau level of composite fermions, and the quasiholes are the holes in the ith CF level. It has been
predicted theoretically that the electric charge of
e
these quasiparticles is q = 2pi+1
, positive for the
quasiholes and negative for quasielectrons.[7,13
15] This charge fragmentation is a fundamental
property of the FQH quantum fluid.
The fractional charge of the quasiparticles has
been measured directly in recent experiments.[16]
In the experiment, a 300 nm quantum antidot
(potential hill) was etched into the layer of 2D electrons. In strong B electron states circling the antidot are quantized by the Aharonov-Bohm condition that a state must contain an integer number
of 0 . The occupation of the anitdot was detected
via resonant tunneling: there is a peak in tunneling
conductance each time a state becomes occupied,
shown in Fig. 1. The surface area of the antidot
S was measured independently from the B-sweep
data: the total flux through the antidot BS changes

2. Energy gaps. The QHE itself, that is, the

existence of plateaus in Rxy = h/f e2 as a function
of , centered at exact filling = f , depends critically on existence of a gap in the excitation spectrum. The physical nature of the gap can be distinguished between the single-particle gaps characteristic of the IQHE and the many-body interaction gap necessary for the FQHE. The examples
of integer gaps are the kinetic energy gap between
Landau levels, the spin flip gap, and the valley
gap in Si. Experimentally, the fractional gap is always due to Coulomb interaction. It is quite common for the magnitude of the gap to be affected
by several physical mechanisms, for example, spin
flip gaps are often enhanced by the spin-dependent
Coulomb exchange, and the fractional Coulomb
gaps are reduced by mixing of several Landau levels.[3] When is varied from the exact filling quasiparticles are created (electron-like for > f and
hole-like for < f ). The quasiparticles are localized by disorder and therefore no dissipative con2

Tunneling Conductance (a.u.)

0
-4.0

transformation somehow makes the many-body

Coulomb interaction energy of 2D electrons look
like kinetic energy of weakly interacting CFs, in a
way not completely understood at present. However, various experiments give strong evidence for
a reasonably well defined Fermi surface of CFs of
Fermi wave vector kFcf near = 12 . If Bcf = 0,
CFs experience no external magnetic field and
move in straight lines. At small Bcf , CFs execute cyclotron motion on the Fermi surface. The
experiments detect, by different techniques, the
resonance occuring when the CF cyclotron radius
Rcf = ~kFcf /eBcf is commensurate with an external length L.Note that for 2D spin-polarized
particles kFcf = 4n depends only on density.
In surface acoustic wave (SAW) experiments
the velocity of propagation and the attenuation
of SAW is affected by their interaction with 2D
electrons. Anomalies in propagation of high frequency SAW, in disagreement with what was
expected from dc conductivity, were observed
near = 12 .[19] These anomalies were successfully
explained in terms of interaction of small wave
vector SAW with a gapless Fermi sea of CFs, at
a mean field level.[9] In subsequent experiments
a geometric resonance has been observed, Fig. 2,
when the CF cyclotron diameter is equal to the
SAW wavelength.[20] In effect, the SAW sets up
a lateral superlattice potential, and similar commensurability geometric resonances were observed

=1
= 1/3

T = 14 mK
electrons

0
quasiparticles

-2.0
0.0
Global Gate Voltage (V)

2.0

Fig. 1. The quantum antidot electrometer. Tunneling conductance peaks occur each time the occupation of the antidot changes by one particle: an electron for the IQH and a
quasiparticle for the FQH regime. The charge of the particle is proportional to the global gate voltage; it takes same
voltage to attract three quasiparticles as one electron.[16]

by 0 between two conductance peaks. There is a

large global gate on the other side of the GaAs sample of thickness d; the gate forms a parallel plate
capacitor with the 2D electrons. A gate voltage VG
thus produces uniform electric field E = VG /d,
and induces a change of 0 E in the surface charge
density. The charge of one particle q is thus directly given by the electric field needed to attract
one more particle in the area S: q = (0 S/d)VG ,
where VG is the change of the global gate voltage
between the successive conductance peaks. The results of the quantum antidot electrometer experiment were as follows: in the integer QH at i = 1
and 2 the charge of the electron was obtained as
1.57 1019 Coulomb, the charge of the f = 13
quasiparticles as 5.20 1020 C, and of the f = 52
quasiparticles as 3.1 1020 C.[17] Subsequently,
two groups reported determining the charge of e/3
quasiparticles using quite different technique: measuring shot noise as a function of current in constrictions defined in 2D electron layer.[18]
4. Fermi surface of composite fermions.
The physics is different at even denominator fillings. As mentioned in Sect. 1 the effective field experienced by composite fermions Bcf vanishes at
= 21 and thus the system of interacting electrons
looks like zero-field metal of CFs.[9] The analogy is not exact, but still, the vortex attachment

Fig. 2. Relative change of propagation velocity vs. CF effective magnetic field for 8.5 GHz SAW. The dashed line
is a model fit including inhomogeneous broadening.[20]

Focusing Resistance (kOhm)

25 mK

+V

3.0

=1/2
2.0

-V

Vbias

9.2

9.0
1.0

Current
Amp

35 mK

n+GaAs
2DEG AlGaAs

0.0

-0.1

0.0

0.1

0.2

Fig. 4. The tunneling I V power law exponent vs. the

inverse filling 1/ for five samples. The dashed line
gives the LL-theoretical prediction. A schematic of the
cleaved-edge overgrown sample and the I V measuring
circuit are shown above.[28]

Magnetic Field (Tesla)

Fig. 3. Magnetic focusing of composite fermions near = 21

compared with focusing of electrons near B = 0. Composite
fermions experience Bcf = B B( = 12 ). The two B

axes differ by 2 to account for spin polarization at high

B. Inset shows the focusing sample geometry, where two
possible focusing paths are shown by arrows.[22]

cause edge confining potential prevents localization of electron states by disorder, and applied
current is carried in a sample by the delocalized
1D edge states.[6,3] Dissipationless conduction by
edge states has been established in experimental
observation of dramatic nonlocal resistance in
macroscopic QH samples.[24,25] The direction of
circulation of edge excitations is determined by
the applied magnetic field, and the theoretical picture of QH edge is based on chiral Luttinger liquid (LL) models.[26,3] LL theories make several dramatic predictions for edges of the FQHE
at = f . For example, for f = 13 the width of
2
a resonant tunneling peak should scale as T 3
in the low voltage Ohmic regime eV < 2kB T , in
contrast to the familiar Fermi liquid T 1 scaling.
The range of applicability of the LL behavior is
not well known however; a T 1 dependence was
reported in experiments on e/3 quasiparticle resonant tunneling.[27]
Another dramatic prediction of LL theories is
that for electron tunneling into a FQHE at = f ,
at high bias eV > 2kB T the tunneling current
is non-Ohmic, with a power law dependence I
i
the exponent is predicted to
V . For f = 2pi1
be = 2p + 1 for plus, and = 2p + 1 2i for
minus in the denominator, shown in Fig. 4. The
experiment was performed on a cleaved edge 2D
heterostructure overgrown by a metallic n+ GaAs

in experiments with lateral superlattices defined

by lithography, as well as in samples with lithographic 2D arrays.[21]
Another type of experiment where CFs were detected is magnetic focusing.[22] Magnetic focusing
of 2D electrons has been studied extensively.[23]
As shown in Fig. 3, a current is passed through the
left (emitter) constriction and the voltage across
the right (collector) constriction is measured. In
the linear regime, the voltage is proportional to the
applied current, and the ratio is defined as nonlocal focusing resistance. Classically, current carriers emitted out of the left constriction execute cyclotron motion on the Fermi surface, and for most
angles of injection are focused onto the collector
when B is such that the cyclotron diameter 2R is
equal to the constriction separation L. Allowing
for j 1 specular reflections, a series of focusing
peaks occurs when 2jR = 2j~kF /eB = L. As a
function of B the peaks are nearly periodic, Bj =
jB, spaced by B = 2~kF /eL, Fig. 3, both for
electrons near B = 0 and for CFs near = 21 . In
the opposite direction of B the current carriers are
deflected to the left, and no focusing is observed.
5. Edge state transport. The edge of a QH
system plays a central role in charge transport be4

on top of an AlGaAs barrier.[28] I V measurements found power law tunneling, however with
a continuously increasing function of 1/. This behavior is not understood at present, in particular,
why varies on a QH plateau and what plays role
of edge states in gapless regions, as near = 21 .
6. Spin textures. In the picture of independent
spins, each Landau level is split into spin up and
spin down levels, so that electrons are fully spin polarized for < 1, completely unpolarized at even
filling and partially polarized otherwise. Electronelectron interaction, however, is spin-dependent,
which leads to spin-dependent correlations of the
many electron states. This effect is particularly
strong in GaAs, where the band g-factor is small,
0.44. The resulting spin depolarization depends self
consistently on , which affects the Coulomb correlations, as was studied in recent photoluminescence experiments.[29] Since spin Zeeman energy
responds to total magnetic field, while the filling
of 2D electrons is determined only by the normal
component of B, certain QH states undergo phase
transitions from un- or partially-spin polarized to
fully polarized in tilted B.[30]
Evidence for even more exotic topological spin
textures [3] has been reported in recent experiments. In double electron layer samples interlayer spin-dependent correlations can be nearly
as strong as intra layer correlations, even when
tunneling between layers is small. A remarkable
softening of long wavelength intersubband spin
excitations, Fig. 5, was observed at = 2.[31]
This occurs when each layer has only one spinsplit Landau level occupied. Probed by resonant
inelastic light scattering, these spin density wave
excitations soften to as much as 0.1 of the B = 0
values. Even though further studies are required,
these observations indicate magnetoroton instability precursor to a phase transition to an ordered
spin phase, such as quantum antiferromagnets,
long predicted for coupled QHE systems.
In single electron layers at 1 electrons
are fully spin polarized for large g-factors or in
the limit B . For small g-factors topolog-

Fig. 5. The width and the energy of the spin-density excitation (SDE) in a double layer sample. SDE is a linear
combination of Sz = 0 transitions, dashed lines; the spin
flip excitations are |Sz | = 1. Ez is the Zeeman splitting,
and SAS is the symmetric-antisymmetric gap.[31]

ical charged defects in spin orientation called

skyrmions are possible.[32,3] To visualize a
skyrmion one can think of all far spins pointing
up, then rotate smoothly downward at the position of the skyrmion. The size of a skyrmion is
controlled by the competition between the Zeeman energy, which tends to reduce the number of
flipped spins, and the Coulomb exchange energy,
which tends to spread the extra charge over a
large area. In = f QHE, skyrmion excitations
20

10

K (kHz)

15

0
0.6

0.8

1.0

1.2

1.4

1.6

1.8

Fig. 6. Electron spin polarization measured by NMR

Knight shift near = 1 at B = 7 T, T = 1.5 K. The solid
line assumes non-interacting electrons, the dashed line is a
fit for S = 3.6 finite size skyrmions.[33]

are predicted to have net charge of f e, in general

different from charge of a Laughlin quasiparticle.
Evidence for skyrmions was observed in NMR experiments measuring Knight shift of 71 Ga nuclei,
Fig. 6, which is sensitive to the spin polarization
of 2D electrons. The spin polarization exhibits a
maximum at = 1, and falls off steeply on either
side. This behavior is understood as due to a finite
size (S 4) skyrmions excited as is varied from
the exact filling. More recently, evidence for formation of collective spin textures, though even of
smaller size, was obtained for FQHE, = 13 .[33]

[12] A. Pinczuk, B. S. Dennis, L. N. Pfeiffer, and K. W.

West, Phys. Rev. Lett. 70, 3983 (1993).
[13] F. D. M. Haldane, Phys. Rev. Lett. 51, 605 (1983).
[14] B. I. Halperin, Phys. Rev. Lett. 52, 1583 (1984).
[15] A. S. Goldhaber and J. K. Jain, Phys. Lett. A 199,
267 (1995).
[16] V. J. Goldman and B. Su, Science 267, 1010 (1995).
[17] V. J. Goldman, Surf. Sci. 362, 1 (1996); Physica E 1,
15 (1997).
[18] L. Saminadayar, D. C. Glattli, Y. Gin, and B. Etienne,
Phys. Rev. Lett. 79, 2526 (1997); R. de Picciottoet al.,
Nature 398, 162 (1997).
[19] R. L. Willett, M. A. Paalanen, K. W. West, and L. N.
Pfeiffer, Phys. Rev. Lett. 65, 112 (1990).

Acknowledgments: The author benefited

greatly from many discussions with B. I. Halperin,
S. M. Girvin, N. Read, J. K. Jain, D. C. Tsui, A.
H. MacDonald, S. A. Kivelson, H. L. Stormer, I.
L. Aleiner, X. G. Wen and many other colleagues.
This work was supported in part by the NSF
under Grant No. DMR-9629851.

[20] R. L. Willett, K. W. West, and L. N. Pfeiffer, Phys.

Rev. Lett. 75, 2989 (1995).
[21] W. Kang, H. L. Stormer, L. N. Pfeiffer, K. W. Baldwin,
and K. W. West, Phys. Rev. Lett. 71, 3850 (1993); J.
H. Smet et al., Phys. Rev. Lett. 77, 2722 (1996).
[22] V. J. Goldman, B. Su, and J. K. Jain, Phys. Rev. Lett.
72, 2065 (1994).
[23] H. van Houten et al., Europhys. Lett. 5, 721 (1988);
Phys. Rev. B 39, 8556 (1989).

References

[24] P. L. McEuen et al., Phys. Rev. Lett. 64, 2062 (1990).

[1] K. von Klitzing, G. Dorda, and M. Pepper, Phys. Rev.

Lett. 45, 494 (1980).

[25] J. K. Wang and V. J. Goldman, Phys. Rev. Lett. 67,

749 (1991); Phys. Rev. B 45, 13 4795 (1992).

[2] D. C. Tsui, H. L. Stormer, and A. C. Gossard, Phys.

Rev. Lett. 48, 1559 (1982).

[26] X. G. Wen, Intl. J. Mod. Phys. B6, 1711 (1992).

[27] I. J. Maasilta and V. J. Goldman, Phys. Rev. B 55,
4081 (1997).

[3] For reviews see: The Quantum Hall Effect, 2nd Ed.
(Edited by R. E. Prange and S. M. Girvin) Springer,
NY (1990); Perspectives in Quantum Hall Effects,
(Edited by S. Das Sarma and A. Pinczuk) Wiley, NY
(1997); The Quantum Hall Effects, T. Chakraborty
and P. Pietilainen, Springer series in Solid State
Sciences 85 (1995); S. M. Girvin, The Quantum Hall
Effect, Les Houches Lecture Notes (1998).

[28] A. M. Chang, L. N. Pfeiffer, and K. W. West, Phys.

Rev. Lett. 77, 2538 (1996); M. Grayson, D. C. Tsui,
L. N. Pfeiffer, K. W. West, and A. M. Chang, Phys.
Rev. Lett. 80, 1062 (1998).
[29] I. V. Kukushkin, K. von Klitzing, and K. Eberl, Phys.
Rev. Lett. 82, 3665 (1999).

[4] http://www.nobel.se/prize/index.html

[30] R. G. Clark et al., Phys. Rev. Lett. 62,1536 (1989);

J. P. Eisenstein, H. L. Stormer, L. N. Pfeiffer, and K.
W. West, Phys. Rev. Lett. ibid 1540; R. R. Du et al.,
Phys. Rev. Lett. 75, 3926 (1995).

[5] R. B. Laughlin, Phys. Rev. B 23, 5632 (1981).

[6] B. I. Halperin, Phys. Rev. B 25, 2185 (1982).
[7] R. B. Laughlin, Phys. Rev. Lett. 50, 1395 (1983).
[8] J. K. Jain, Phys. Rev. Lett. 63, 199 (1989); Science
266, 1199 1994.

[31] V. Pellegrini, A. Pinczuk, B. S. Dennis, A. S. Plaut,

L. N. Pfeiffer, and K. W. West, Phys. Rev. Lett. 78,
310 (1997); Science 281, 799 (1998).

[9] B. I. Halperin, P. A. Lee, and N. Read, Phys. Rev. B

47, 7312 (1993).

[32] S. L. Sondhi, A. Karlhede, S. Kivelson, and E. H.

Rezayi, Phys. Rev. B 47, 16 419 (1993).

[10] J. K. Jain and V. J. Goldman, Phys. Rev. B 45, 1255

(1992).

[33] S. E. Barrett, G. Dabbagh, L. N. Pfeiffer, K. W. West,

and R. Tycko, Phys. Rev. Lett. 74, 5112 (1995); P.
Khandewal, N. N. Kuzma, S. E. Barrett, L. N. Pfeiffer,
and K. W. West, Phys. Rev. Lett. 81, 673 (1998).

[11] R. R. Du, H. L. Stormer, D. C. Tsui, L. N. Pfeiffer,

and K. W. West, Phys. Rev. Lett. 70, 2944 (1993);
D. R. Leadley, R. J. Nicholas, C. T. Foxon, and J. J.
Harris, Phys. Rev. Lett. 72, 1906 (1994).