24 Superfluid
3
He and the Cuprate Superconductors 1533
perimental quantities such as the specific heat which
are crudely proportional to the amount of normal
component will themselves have a power-law behav-
ior. Such power-law behavior is indeed observed in
the cuprates,in particular in theNMR properties and
the deviation of the penetration depth from its zero-
temperature limit. This is strong evidence for the ex-
istence of nodes of the gap which are also observed
more directly in ARPES experiments. To be sure, the
mere existence of nodes is not peculiar to exotic pair-
ing states, an s-wave (A
1g
) state, provided it is of the
so-called“extended”type,can also have them, but the
salient point is that a d
x
2
−y
2
state, because it must (by
definition!) change sign under reflection in any 45
◦
axis, necessarily has nodes at these points, which is
exactly where they are observed in ARPES.
In the above argument, it was taken for granted
that the energy gap (k) is simply proportional to
the orbital order parameter. This is indeed the case
for spin singlet pairing, where the order parameter
can be expressed as a simple product of a trivial spin
factor (the singlet state) and an orbital factor. For
spin triplet pairing the situation is in general a lit-
tle more complicated. In this case it turns out (at
least for the so-called “unitary” class of states which
includes the A and B states) that the gap (k) is pro-
portional to the quantity |d(k) |
1/2
. We see that the
B-phase gap is uniform over the Fermi surface and
thus has no nodes, while the A-phase gap is propor-
tional to (k · )2 and so has two nodes in the direc-
tions±
ˆ
.Consequently,we expect that the properties
of
3
He-B should behave as a function of temperature,
similarly to those of a classic (s-wave) superconduc-
tor, (e.g. the specific heat should decrease exponen-
tially in the limit T → 0), while for
3
He-A we should
get power-law behavior qualitatively similar to that
seen in the cuprates. But at the very lowest temper-
atures (lower than currently accessible) the thermo-
dynamics should be dominated by bosonic collective
excitations such as spin waves. This prediction ap-
pears consistent with existing experiments, although
lower-temperature measurements would be needed
to confirm it definitively [9].
A second consequence of exotic pairing in general
is that various experimentally measured quantities
may show an anisotropy over and above that pre-
scribed by the Hamiltonian. Examples of such be-
havior in superfluid
3
He include the A-phase ultra-
sonic attenuation,which depends on the angle of the
sound propagation vector relative to the characteris-
tic orbital vector
ˆ
, and the magnetic susceptibility,
which depends on the angle of the field relative to
the characteristic spin vector
ˆ
d.In
3
He-B the orbital
and spin properties are individually isotropic,but the
effect of the rather subtle anisotropy in the correla-
tion of the spin and orbital coordinates comes out
in the NMR behavior,which is sensitive to just such
correlations and indeed depends on the direction of
the static field relative to the“spin-orbit rotation vec-
tor” ˆ!. In the cuprates the most obvious signature of
anisotropy is in the ARPES data.While onewould ex-
pect (and finds) the spectra to be somewhat different
along the crystal axes and at 45
◦
even in the normal
phase, this difference is strongly accentuated in the
superconducting phase, since as already described a
gap develops which is finite along the axes but zero
in the 45
◦
directions.
A third class of phenomena related to exotic pair-
ing has to do with possible variations of the char-
acteristic “orientation”in space and/or in time. Here
there is a very substantial difference between
3
He and
the cuprates.In the former case,if for definiteness we
consider the A phase, the characteristic orbital vec-
tor
ˆ
and spin vector
ˆ
d are pinned only very weakly
in bulk (by effects such as the nuclear dipole–dipole
interaction,residual heat flows,etc., see [29], Sect.X),
and as a result such spatial and/or temporal variation
of the orientation is quite commonplace and of great
interest both experimentally and theoretically, see,
e.g. [9]. By contrast, in the cuprates there is no ques-
tion of spin “orientation” since the state is a singlet,
and the “orientation” of the orbital wave function
is firmly pinned by the crystal axes. The only re-
maining freedom is to exchange the + or – signs on
thefour“lobes”ofthepairwavefunction.Sincethis
merely multiplies the total pair wave function by a
– sign, one might think it would have no physical
significance.Indeed for a spatially uniform situation
this is true. However, it is precisely this subtle degree
of freedom which allows, in a spatially nonuniform
setup, the rather spectacular Josephson (quantum
phase interference) experiments which are the most