Physica 108B (1981) 1069-1070 RA 3

North-Holland Publishing Company

EXACT GROUNDSTATE OF A QUANTUM MECHANICAL ANTIFERROMAGNET

B. Sriram Shastry and B i l l Sutherland

Department of Physics, University of Utah, Salt Lake City, UT 84112

We present some exact results f o r the ground state of a quantum mechanical a n t i f e r r o -

magnetic model in the two dimensions with next-nearest neighbor i n t e r a c t i o n s .

We have found some exact results f o r the
ground state(s) of an anisotropic quantum spin
Heisenberg-Hamiltonian with next neighbor i n -
teractions in two and three dimensions. For
c l a r i t y we present the two dimensional case in
detail and postpone the three dimensional
results to a Forthcoming paper. We consider a
two dimensional l a t t i c e (Fig. 1) where each
dashed l i n e represents a bond with strength 2a"
and each nearest neighbor bond has strength
unity. The Hamiltonian is
z Hence a use of the v a r i a t i o n a l p r i n c i p l e implies
H= ~ h + 2a ~ h - h ~ S (1)
<ij> i j <¢m> ~m i i
where <ij> represents a sum over a l l nearest
neighbors and <~m> represents a sum over the
diagonal bonds. Also
z z x x YS y
h =J S S + J S S + J S (2)
ab za b xa b ya b The nature of ground state c o r r e l a t i o n s in
We r e s t r i c t our a t t e n t i o n to the n o n - t r i v l a l
case Jx,Jv,Jz ) O. This model resembles the
triangula~ l a t t i c e Heisenberg antiferromagnet,
which has been studied extensively in recent
l i t e r a t u r e , [1,2,3J in t h a t the Ising l i m i t
e x h i b i t s f r u s t r a t i o n [ 4 ] . I t is our hope t h a t
the exact solution sheds l i g h t on the i n t e r p l a y
between quantum effects and spinglass order-
ing [ 5 ] .
A. QUANTUMCASE Our solution is inspired by
the surprising solution of Majumdar's, [6] f o r
a l i n e a r chain Heisenberg antiferromagnet with
a nearest and next nearest neighbor coupling
which is h a l f as strong as the f i r s t . We f i r s t
observe t h a t the state
I~> = . [~,m] (3)
<zm>
is an eigenstate of the Hamiltonian Eq. (1) with
eigenvalue
E /N = W z - a ~ S ( S + 1)[J + J + J ] (4)
o 3 x y z allowed than f o r ~ > 1. We have established
(We use the notation [~,m] f o r a s i n g l e t combin-
ation of the spins ~ and m.) This is seen most
r e a d i l y by using the i d e n t i t y s i a ( s l ~ + Sk~)
x [ j , k ] = O and hence the f i r s t term in Eq. (3)
annihilates I~>. We f u r t h e r show t h a t Eo
saturates a lower bound to the ground state
energy f o r suitable values of the parameters
and hence I~> is the rigorous ground state.
We w r i t e H = ~t ht, where the sum is over
the N t r i a g l e s (each diagonal bond supports
two t r i a n g l e s ) . Each t r i a n g u l a r Hamiltonian
has the form
h = h + h + a h - _h (S z + S z ) (5)
t i¢ i ~ ~m 2 ~ m
Eo ) N et where et is the ground state energy of
Eq. (5). We have shown that et equals W and
hence I~> is the ground state of H, provided
) ab, where ~b is: (a) l + h/Jz f o r s = I / 2
and f o r a r b i t r a r y Jx, Jv and Jz" TFor the
i s o t r o p i c case we have ~ b e t t e r bound
~b = l + h/2J.) (b) 1 + s + h/2J f o r s ) I
in the i s o t r o p i c l i m i t (Jx = Jy = Jz = J)-
I~> is seen to be l i q u i d l i k e , with only short
ranged order and hence we have termed i t the
"quantum spin l i q u i d " phase (Q.S.L.). This wave
function is of the type suggested by Anderson
[1] and is f a m i l i a r from valence bond theory in
chemistry. For a smaller than the bounding
value ab, we have not succeeded in
solving the model exactly. However, we can
show that the character of the ground state
must change at some value of ~ between ~b and
2s/(2s + 1) since at the l a t t e r value, the N ~ l
(Ising) state provides a b e t t e r upper bound to
the ground state energy.
B. ISING LIMIT We have studied the ground
state degeneracy of the Hamiltonian in the
Ising l i m i t Jx = Jy = 0 and h = 0 f o r s = 1/2.
I t is readily established that f o r ~ < 1, the
ground state entropy is 0(1) whereas f o r a ) 1,
i t is of O(N). The case ~ > 1 is quite simple
since in t h i s case the only constraint is that
spins at the ends of a diagonal bond must be
a n t i p a r a l l e l . The entropy then is S = kB
N(0.3466). The case a = 1 is much more complex,
since in t h i s case, more configurations are
that the case ~ = 1 can be mapped on to a 10
vertex model containing N/2 sites with two
s u b l a t t i c e s , which is a generalization of the
ice problem solved by Lieb [ 7 ] . We have
obtained a rigorous lower bound to the entropy
s ~ kB N(.4812) by using a braiding technique.

0378-4363/81/0000-0000/$02.50 © North-Holland Publishing Company 1069

1070
C. CLASSICALLIMIT As s + ®, the singlet
state is the ground state only as a ÷ ®.
However, for arbitrary a, we have succeeded in
determining the ground state exactly in the
isotropic l i m i t . We observe that the classical
planar ( i . e . , x-y) and Heisenberg model share a
ground state for the triangular Hamiltonian (5).
Each triangle has a two-fold discrete "chiral"
degeneracy, over and above the continuous
degeneracy, of the sort discovered by V i l l a i n
in similar systems [8]. The optimum twist
angle between a spin at the base and apex of
a triangle is (~ ± cos-I I/2~) and the two
c h i r a l i t i e s correspond to the two choices of the
sign. One may then assign arrows to the bonds
indicating the direction along which rotations
are anti-clockwise. The constraint on a square
containing a diagonal bond is that the arrows
on parallel sides be in the same direction.
The constraints on empty squares is that the
line integral of the arrows vanish ( i r r o t a t i o n -
al flow). There are precisely four degenerate
ground states which satisfy these constraints.
Each ground state contains one preferred direc-
tion, say one of the four points of the compass,
such that there is an average flow of arrows in
this direction. Thus, the spins exhibit helical
ordering as we proceed in this preferred direc-
tion. for a ~ 1/2, the Neel state is the ground
state.
D. SUMMARY Our understanding of the model in
the isotropic l i m i t is summarized in the "phase
diagram" (Fig. 2). The l i n e bounding the Q.S.L.
phase is the point of highest frustration in
the Ising l i m i t . In the Q.S.L. phase the Ising
l i m i t has macroscopic degeneracy but the quantum
effects freeze this degeneracy out for arbi-
t r a r i l y small Jx and Jy. I t is interesting that
similar behavior is found in approximate treat-
ments of similar systems [2,3,5].
Finally we would l i k e to point out the
remarkable property of the ground state I~> for
s = I/2 in the case a ~ 1 and Jx ~ Jv ~ Jz. In
this case the ground state possesses-rotation
invBriance (being a singlet) even though the
Hamiltonian does not. This is the only example
of symmetry br--r-ea-kTh-g in reverse that we are
aware of. This provides an exmple in non-
r e l a t i v i s t i c physics where the invariance of
of the vacuum exceeds the invariance of the
world [9].
References
[1] P.W. Anderson, Mat. Res. Bull. 8, 153
(1973), P. Fazekas and P.W. Anderson,
Phil. Mag. 30, 432 (1974).
[2J D.C. Mattis, Phys. Rev. Letts. 42, 1503
(1979); L.G. Marland and D.D. B~ts, Phys.
Rev. Letts. 43, 1618 (1979).
[3] R. J u l l i e n , R-Z.A. Penson, P. Pfeuty and K.
Uzelac, Phys. Rev. Letts 44, 1551 (1980).
t4] G. Toulouse, Comm. Phys. E 115 (1977).
[SJ R. K1emm, J. Phys. C. 12,--L735 (1979).
L6J C.K. Majumdar, J. Phys~--C. 3, 911 (1970).
[7] E. Lieb, Phys. Rev. Letts. I-8, 692 (1967).
[8] J. V i l l a i n , J. de. Physique-~IB, 385 (1977).
[9] S. Coleman, J. Math. Phys. 7_,---787 (1966).
\
\ \
\ \
\ \
\ \
\ \
\
/ /
/ /
/ /
/ /
/
/ /
/
\
\
\ \
\
\ \
\ \
N \
/
/ /
/
/ /
/ /
/ /
J
Fig. I. The Lattice
2
C{
.5
NEEL L,R,O,
!
l
I° 1/s
Fig. 2. The Phase Diagram