Journal of Combinatorial Theory, Series A 95, 390-397 (2001)

doi: 10.1006/jcta2001.3182,available online at http://www.idealibrary.com on D E hl0

NOTE

Composite Fermions and Integer Partitions

Arthur T. Benjamin
Department of Mathematics, Harvey Mudd College, Claremont, Calijbrnia 91 711
E-mail: [email protected]

Jennifer J. Quinn
Department of Mathematics, Occidental College, 1600 Campus Drive,
Los Angeles, California 90041
E-mail: [email protected]

and

John J. Quinn and Arkadiusz W6js

Department of Physics, University of Tennessee, Knoxville, Tennessee 37996
E-mail: [email protected], [email protected]
Communicated by the Managing Editors
Received September 28, 2000; published online June 18, 2001

We utilize the KOH theorem to prove the unimodality of integer partitions with
at most a parts, all parts less than or equal to b, that are required to contain either
repeated or consecutive parts. We connect this result to an open question in quan-
tum physics relating the number of distinct total angular momentum multiplets of
a system of N fermions, each with angular momentum L, to those of a system in
which each Fermion has angular momentum d* = d - N + 1. 0 zoo1 ~cadnnic~ r r a
Key Words: unimodality; restricted integer partition; generating function.

1. INTRODUCTION

Mathematics and physics have a symbiotic relationship. Insights in one

field often lead to advances in the other. For example, the proof of a purely
mathematical construct, the alternating sign matrix theorem, was only
made possible with the tools developed by physicists studying the statistical
mechanics of square ice [2]. In this paper, tools from partition theory
(specifically, recent combinatorial proofs due to O'Hara and Zeilberger
390
0097-3165/01 $35.00
Copyright 2001 by Academic Fr'm
All rights of reproduction in any form reserved.
 NOTE 39 1

[3, 4, 12, 131) make it possible to answer an open question of quantum

physics involving the composite Fermion model of the fractional quantum
Hall effect.
The study of electronic properties of quasi-two-dimensional systems
under high magnetic fields and low temperatures has been a fruitful area of
investigation garnering the 1985 and 1998 Nobel Prizes in physics for the
discovery and explanation of the integral [lo] and fractional [6,9] quantum
Hall effects. For completeness, we provide a brief description of both of
these effects, though a true physical understanding is not required to
appreciate the partition results. Mathematicians may skip the physics
presented in the next two paragraphs.
The integral quantum Hall effect results from the highly degenerate
nature of the single particle energy levels E , = (n + $) Ao,, where n is a
non-negative integer and the cyclotron energy Ao, is proportional to the
applied magnetic field B,. Each level contains N4 = ABo/$, single particle
states, where A is the area of the sample and 4, = Ac/e is the quantum of
magnetic flux. Each particle state can be occupied by a single electron or
can be empty. When the ratio v of the number of electrons N to N , is an
integer, an energy gap h o , separates the ground state from the nearest
excited state. This gap causes the integral quantum Hall effect. When v
takes on a value corresponding to an odd denominator fraction (e.g.,
,=!3 , 1s, 23 , 2s, ...) the fractional quantum Hall (FQH) effect is found. The
energy gap causing the FQH effect results from the Coulomb interaction
between the electrons in the highly degenerate single particle states of the
partially filled energy level.
A simple intuitive understanding of the FQH effect has been obtained
from the composite Fermion (CF) model [5]. An electron constrained to
move on the surface of a sphere in the presence of a radial magnetic field
of constant magnitude can be described by an eigenvector It', m), where t'
is the angular momentum of an individual electron and m is its projection
onto a given direction. According to elementary quantum mechanics the
allowed values of m belong to the set A?[ = { - t', -t' + 1, ..., t' - 1, t'}
where for the present case t' is either an integer or half integer, i.e., 2t' E Z.
Many electron eigenfunctions can be constructed from antisymmetrized
products of single particle eigenfunctions fly-, lt',mi) in which every m, in
a given product must be a different member of the set A?/. We represent
such a state by m = (m,, m,, ...,m,) where t' 2 m, > m, > ... > m,> - l
and refer to it as an M-state if Cy-, m i = M. From the (U,+') distinct
products obtained by choosing N different values of mi from the 2t'+ 1
members of 4,we can form linear combinations ( L ,M , a ) which are
eigenfunctions of the total angular momentum operator i = Cy-, ti
(where tiis the angular momentum operator of the ith particle of angular
momentum t') and of M = Cr-, mi. The label a is used to distinguish
392 NOTE

distinct angular momentum multiplets (the 2L + 1 states IL, M ) with M

belonging to JaC, are referred to as a multiplet of angular momentum L ) with
the same value of L. Let f,(N, M ) count the number of M-states using N
Fermions of angular momentum t. The number of distinct multiplets of
angular momentum L is defined as

f,(N,L)-f,(N,L+I) for LBO

(1)
f,(N,L+l)-f,(N,L) for L<O.

For the non-physicist reader, we now summarize the essential ideas.

A fermion of angular momentum t is a particle that takes on a value from
+
the set A, = { - t , - t 1, ..., t ) , where the number 2t' is an integer. In an
N-fermion system of angular momentum t, all ferrnions will have different
values from A,. We represent such a system by m = (m,, m,, ...,m,) where
t B m l > m , > ... >m,B -{andcall itan M-stateifzr-,mi=M. Welet
f,(N, M ) count the number of M-states using N fermions of angular
momentum t. Notice f,(N, M ) = f,(N, -M). In an N-fermion system
where each particle has angular momentum 8, the number of distinct multiplets
of (total) angular momentum L, denoted gAN, L), is the difference between the
number of M states when M equals L and L + 1, as defined in Eq. (1).
In this paper we show g,(N, L) 2 0 and establish the conjecture of Quinn
and Wojs [8] that the number of distinct multiplets for N Ferrnions of
angular momentum t is greater than or equal to the number of distinct
multiplets for N Fermions of angular momentum 4 - N + 1. In other
words, for all L,

This conjecture was useful in understanding why the mean field composite
Fermion picture correctly predicted the lowest energy band of multiplets
for any value of the applied magnetic field, because the transformation
from N electrons to mean field composite Fermions [7, 81 involves chang-
ing the angular momentum from t to t * = t - N + 1.
We illustrate these ideas through a small example. For a given t and N,
define the function G,(N) = C g,(N, M ) qM. Table I displays G,(3) for dif-
ferent angular momenta t in 3 particle systems. The term 2q3 in G4(3)
means that there are 2 more ways to express 3 as a sum of three distinct
integers from -4 to 4 than there are to express 4. Notice that simply
decreasing the angular momentum t does not guarantee containment of
distinct multiplets. Comparing G5(3) to G4(3), containment violations
occur for 3-states and l-states. While 3-states occur at both angular
momenta, the number of distinct multiplets when t = 4 exceeds the number
of distinct multiplets when t = 5. The 1-state is allowed only when t = 4 and
 TABLE I
Calculations of the Number of Distinct Multiplets for a 3 Ferrnion System
Using Angular Momenta of C = 5,4, 3,2, 1

not when r! = 5. However, by reducing the angular momentum by N - 1 = 2

we see that term-by-term, G,(3) 2 G,(3) > Gl(3) and G,(3) > G,(3).
Studying the relationships between distinct total angular multiplets effec-
tively reduces to a problem of restricted integer partitions. Let Pa(b, c)
denote the set of partitions of c into at most b positive parts, with all parts
less than or equal to a, and let pa(b, c) = IBa(b, c)l. For a, b > 0, the
generating function Cpa(b, c) qc is precisely the Gaussian polynomial [ I ]

The coefficients of this polynomial are unimodal and symmetric about ab/2.
Many elegant proofs of this fact have been discovered (see [12]), but it
was not until 1990 that a direct combinatorial proof was established by
Kathy O'Hara [4]. Subsequent work by Zeilberger [ 12, 131 distilled
O'Hara's combinatorics into the powerful KOH theorem. We utilize the
KOH theorem to prove the unimodality of integer partitions with at most
a parts, all parts less than or equal to b, that are required to contain either
repeated or consecutive parts. This further restriction will be the key to
unlocking the Fermion conjecture.

2. RESTRICTED INTEGER PARTITIONS AND UNIMODALITY

We represent a partition in P,(b, c) by a b-tuple y = (y,, y,, ..., yb)

where a 2 y, 2 y, 2 -.2 yb 2 0 and C;-, y, = c. We say y has repeated or
394 NOTE

consecutive parts if yi+ = yi or yi - 1 for some 1 < i < b. Notice that a par-
tition y with fewer than b - 1 positive parts falls in this category since
y b - l = yb=O. Let ra(b,c ) count the number of partitions in Ba(b,c )
that have repeated or consecutive parts. We wish to show that for a, b > 0,
the coefficients of the generating function R(a, b; q ) = C ra(b,c ) qc are
unimodal and symmetric about ab/2.
Observe that the subset of partitions in Ba(b, c ) without repeated or
consecutive parts can be put into one-to-one correspondence with
9 , - , + , ( b , c - b(b - 1 ) ) via the bijection

Thus

The fact that [";fb ] q and q b 2 - b [ a - ~ + 2 ] q are both unimodal and symmetric
about ab/2 does not guarantee that their difference has the same feature.
Fortunately, we can appeal to the KOH Theorem, which we state here
using the formulation from Bressoud [ 3 ] .

THEOREM
2.1 [ K O H ] .

wherey,+,=O, Y O = 0 ,and Y i = y , + - . -+ y i f o r i = l , ..., b + 1 .

As shown in [ 131, all summands of the KOH identity are unimodal and
symmetric about ab/2 and consequently so is their sum. The partition
(b, 0, ..., 0) E Bb(b,b ) contributes the term qd-b[(a+z)-b]q
to the sum.
Consequently R(a, b; q ) is the sum of the remaining terms and hence is also
unimodal and symmetric about ab/2, as desired.
Interestingly, the generating function whose coefficients enumerate
restricted partitions with repeated parts need not be unimodal, nor will the
generating function that enumerates restricted partitions with consecutive
parts. For example when a = b = 2, the generating functions are 1 + q2 + q4
and q + q3, respectively.
 NOTE 395

3. CONNECTIONS TO QUANTUM PHYSICS

Returning to our original problem, let B,(N, M ) denote the set of

M-states using N Fermions of angular momentum t. For r n B,(N,
~ M)
the bijection

establishes a one-to-one correspondence between F,(N, M) and 8,-,+, (N,

M + Nf - N(N - 1)/2). This correspondence is valid even when / is a half-
integer. Hence we have the generating function

Since [ " i b ]is, unimodal and symmetric about ab/2, it follows that
C fC(N,M ) qM is unimodal and symmetric about 0. Thus for M < 0,
f,(N, M ) <f,(N, M + 1) and for M 2 0 , f,(N, M ) 2 f,(N, M + 1); so
g,(N, M ) 2 0.
We are now ready to establish the Quinn and W6js conjecture.

THEOREM 3.1. The number of distinct multiplets for N Fermions of total

angular momentum / is greater than or equal to the number of distinct multiplets
for N Fermions of angular momentum / - N + 1. In other words, for all M,

Proof: This is equivalent to proving for N, / > 0, and for M 2 0, that

or equivalently,

Hence it suffices to show that for fixed N, t', S(f, N, q) = C (f,(N, M ) -

fC-N+l(N, M ) ) qMis unimodal and symmetric about M = 0. From Eq. (2),
we have
396 NOTE

+
Substituting a = 2 t - N 1, b = N reduces the braced factor to R(a, b; q),
which from Section 2 is unimodal and symmetric about ab/2 = N t -
N(N- 1)/2. Thus, S ( t , N; q) is unimodal and syrnrnetric about 0, as
desired. I

Repeatedly reducing the angular momentum by N- 1 gives the

immediate corollary:

COROLLARY 3.1. For a positive integer k, the number of distinct multi-

plets for N Fermions of total angular momentum t is greater than or equal
to the number of distinct multiplets for N Fermions of total angular momen-
tum t - k(N - 1 ). In other worcls, for all M,

The preceding analysis can also be applied to quasi-particles called

bosons. A boson of angular momentum t takes on a value from the set Jlt,
where / is again an integer or half-integer. An N-boson system of angular
momentum t differs from a N-Ferrnion system of angular momentum t in
that bosons may take on repeated values from .All. We represent such a
system by z = (z, ,z,, ..., z,) where t 2 z, 2 z, 2 . -. 2 Z N > - t. Here
M-states and the number of distinct multiplets of a given total angular
momentum are defined as before. There is a one-to-one correspondence
between N-Fermion systems with angular momentum t and N-boson
systems with angular momentum t - ;(N - 1), via the bijection

( m , - ; ( N - I ) , m, - $(N- 1) + 1, ..., m ~ 1 -

So the number of distinct multiplets for N Fermions each with angular

momentum t exactly equals the number of distinct multiplets for N bosons
each with angular momentum t - $ ( N - l).' Therefore, analogous results
for bosons hold.
'The same result was arrived at independently by B. Wybourne [ 1 1 ] using group
decompositions.
 NOTE 397

ACKNOWLEDGMENTS

Special thanks to George Andrews for giving us encouragement and Dennis Stanton for
showing us the power of KOH. The support of Grant DE-FG02-97ER 45657 from the
Material Science Program-Basic Energy Sciences of the U.S. Department of Energy is
gratefully acknowledged.

REFERENCES
1. G. E. Andrews, 'The Theory of Partitions," Addison-Wesley, Reading, MA, 1976.
2. D. M. Bressoud, "Proofs and Confirmations: The Story of the Alternating Sign Matrix
Conjecture," Math. Assoc. of America, Washington, DC, 1999.
3. D. M. Bressoud, In the land of OZ, in "gSeries and Partitions" (D. Stanton, Ed.), IMA
Volumes in Mathematics and Its Applications, Springer-Verlag, New York, 1989.
4. K. O'Hara, Unimodality of the Gaussian coefficient: A constructive proof, J. Combin.
Theory Ser. A 53 (1990), 29-52.
5. J. K. Jain, Composite Fermion approach for the fractional quantum Hall effect, Phys. Rev.
Lett. 63 (1989). 199-202.
6 . R. B. Laughlin, Anomalous quantum Hall effect: An incompressible quantum fluid with
fractionally charged excitations, Phys. Rev. Lett. 50 (1983), 1395-1398.
7. J. J. Quinn and A. Wojs, Composite fennions in fractional quantum Hall effect: Essential
role of the pseudopotential, Phys. E 6 (2000), 1 4 .
8. J. J. Quinn and A. Wojs, Composite fermions in fractional quantum Hall systems, J. Phys.
12 (2000), R265-R298.
9. D. C. Tsui, H. L. Stormer, and A. C. Gossard, Two dimensional magnetotransport in the
extreme quantum limit, Phys. Rev. Lett. 48 (1982), 1559-1562.
10. K. von Klitzing, G. Dorda, and M. Pepper, New method for high-accuracy determination
of the line-structure constant based on quantized Hall resistance, Phys. Rev. Lett. 45
(1980), 494-497.
11. B. Wybourne, private communication, June 5, 2000.
12. D. Zeilberger, A one-line high school proof of the unimodality of the Gaussian polyno-
mials [;I for k<20, in "q-Series and Partitions" (D. Stanton, Ed.), IMA Volumes in
Mathematics and Its Applications, Springer-Verlag, New York, 1989.
13. D. Zeilberger, Kathy O'Hara's constructive proof of the unimodality of the Gaussian
polynomials, Amer. Math. Monthly 96 (1989), 590-602.

Printed by Tiger Publishing Systems, Tempelhof 41, B-8000 Brugge, Belgium