CHINESE JOURNAL OF PHYSICS VOL. 38, NO.

2-II APRIL 2000

Aspects of p-Wave Superconductivity

K. Maki1 , E. Puchkaryov1 , G.-F. Wang1 , and H. Won2

1
Department of Physics and Astronomy, University of Southern California,
Los Angeles, CA, 90089-0484, U.S.A.
2
Department of Physics, Hallym University, Chunchon 200-702, Republic of Korea
(Received August 17, 1999)

Growing evidence indicates the newly discovered superconductivity in Sr2 RuO4 is of p-

wave. Also, a few experimental data of the superconductivity in Bechgaard salts (TMTSF)2 X
with X = ClO4 and PF6 point toward p-wave superconductivity.
Here we shall review our recent works on p-wave superconductivity. First we describe
the effect of impurity scattering in p-wave superconductivity. Then the vortex state both in
a magnetic field parallel to the c-axis and within the a-b plane are considered. The latter
geometry will provide a new method to probe the p-wave symmetry of the superconductivity.

PACS. 74.20.-z Theories and models of superconducting state.

PACS. 74.25.-q General properties; correlations between physical properties in normal
and superconducting states.

I. Introduction

Perhaps the identification of d-wave superconductivity in the hole doped high-Tc cuprates
is one of the most important events in the recent history of superconductivity [1-3]. First of
all, this opens a vast vista that there are many unconventional superconductors [4]. Of course,
the unconventional superconductors have been already around us, since the discovery of heavy
fermion superconductors [5, 6] and the organic superconductors [7, 8]. However, at the present
moment, the systematic understanding of these new superconductors becomes more urgent. Also,
for detection of d-wave superconductivity a new method called phase-sensitive experiment has
been developed [1-3].
We believe p-wave superconductivity fits exactly in this new development. The supercon-
ductivity in Sr2 RuO4 was first reported in 1994 [9, 10]. In the last 2 years, we have gained
enormous understanding of this system [11]. First of all, from 17 O-Knight shift measurement we
know the superconductivity is of spin triplet paring with d^ parallel to the c-axis [12] very similar
to the one in superfluid 3 He-A [13].
Also the specific heat data suggest a superconductor with full energy gap [14; 15] but it
contains a persistent T -linear term. Also the superconductivity is very sensitive to impurities
indicating unconventional superconductivity [16]. Then all these fit very nicely the three orbital
band model [17], which assumes that the superconductivity resides mainly in the band, while
the and bands are in the normal state. Of course, at this moment, we cannot exclude a
possible superconductivity with much smaller order parameter in the and bands.

386 c 2000 THE PHYSICAL SOCIETY

OF THE REPUBLIC OF CHINA
VOL. 38 K. MAKI, E. PUCHKARYOV, G.-F. WANG, AND H. WON 387

All these suggest the most likely order parameter

^ k^1 + ik^2 )
(k) = d( (1)

^ and k is the quasi-particle wave vector within the a-b plane. We recall Sr2 RuO4 is the
where dkc
quasi-2D system with cylindrical Fermi surface and superconductivity between conducting layers
is coupled by Josephson tunneling.
The vortex lattice within Ginzburg-Landau theory based on Eq. (1) appears to predict a
square vortex lattice in the presence of a magnetic field parallel to the c-axis [18; 19]. Also the
linearized Gorkov equation is solved in the same geometry and the square vortex lattice is favored
[20; 21]. Indeed, the square vortex lattice is observed by small angle neutron scattering [22] in
the most part of the H-T diagram.
In a magnetic field within the a-b plane the p-wave order parameter is polarized parallel to
the magnetic field [18; 21]. This polarization is accessible to STM or other transport measurement.
Perhaps this will provide a unique mean to probe the p-wave superconductivity [23].
The effect of impurity scattering is studied in the work by Maki and Puchkaryov, where the
scattering is in the unitarity limit (i.e. the resonance at E = 0) [24]. Impurities first generate a
small island of density of states near E = 0, which may be detected by low temperature transport
measurement. More recently, the quasi-particle spectrum around a single vortex in a magnetic
field parallel to the c-axis is discussed [25].
Before entering the discussion of p-wave superconductors, we would like to point out that
the result described here should be applicable to the superconductivity in Bechgaard salts with
minor modification [21, 26].

II. Impurity scattering

Following the standard method, the effect of impurity is incorporated by replacing the
quasi-particle energy ! by !
~ , where
p
~2 + 2
!
!
~ = !+ (2)
!~
where ! is the Marsubara frequency and is the quasi-particle scattering rate. Here we as-
sumed that the scattering is in the unitarity limit [27-29]. Solving the gap equation within the
weak-coupling limit, we calculate Tc =Tc0 , ( ; 0)= 00 , and N (0)=N0 where Tc (Tc0 ) is the
superconducting transition temperature in the presence (absence) of impurities, ( ; 0) is the
superconducting order parameter at T = 0 K in the presence of impurities ( 00 = (0; 0)), and
N (0) is the quasi-particle density of states at E = 0. They are shown as functions of = c
in Fig. 1, where c is the critical scattering where Tc = 0 [24]. It is very remarkable that
Fig. 1 is very similar to the corresponding one in d-wave superconductors [30]. On the other
hand, the quasi-particle density of states shown in Fig. 2 for several = is quite different from
the one in d-wave superconductors [31]. A small impurity scattering produces an island of the
density of states near E = 0, which is separated from the main part. In particular, the effect
of this island is visible in the thermodynamical critical field shown in Fig. 3. Here we show
D(T =Tc ) = Hc ( ; T )=Hc ( ; 0) (1 (T =Tc )2 ) as a function of (T=Tc )2 . In the pure limit
D(T =Tc ) reduces to the one in s-wave superconductors as given in [32]. Similarly, the superfluid
388 ASPECTS OF p-WAVE SUPERCONDUCTIVITY VOL. 38

FIG. 1. ( ; 0)= 00 ( ), Tc =Tc0 ( ), and the FIG. 2. The quasi-particle density of states
residual density of states N (0)=N0 ( ) N (E)=N0 is shown as a function of
are shown as functions of = c where E= for several = .
c = 12 00 .

FIG. 3. D(T ) versus (T=Tc )2 . FIG. 4. Superfluid densities sk (solid line) and
s? (dotted line).

density within the a-b plane sk and parallel to the c-axis s? are shown as functions of T =Tc in
Fig. 4. Finally, the specific heat and the thermal conductivity for different impurity concentrations
are shown in Fig. 5 and Fig. 6, respectively. Again in the clean limit these results reduce to the
ones for s-wave superconductors [32-33].

III. Vortex state in H k "

We first determine the upper critical field by solving the linearized Gorkov equation [21-22]
following [34]. In contrast to s-wave superconductor, (k, r) is given by
VOL. 38 K. MAKI, E. PUCHKARYOV, G.-F. WANG, AND H. WON 389

FIG. 5. Specific heat versus T for several . FIG. 6. Thermal conductivity for several .

(k; r) = (ei + C(ay )2 e i )j0i (3)

where j0i is the reference Abrikosov state for an s-wave superconductor and ay is an analog of
the raising operator in a harmonic potential. The upper critical field for several is given in
Fig. 7. Perhaps one of the remarkable relations is [21]
Hc2 ( ; 0)=Hc2 (0; 0) = (Tc =Tc0 )2 (4)
which is unique to unconventional superconductors. From this we can deduce
v
( ; 0) = 0:1473 : (5)
Tc
This is very close to the one in the clean s-wave superconductor near T = Tc , (Tc ) =
0:16322 v=Tc Also, when Hc2 ( ; T )=Hc2 ( ; 0) is plotted for T=Tc the result for = c 0:3
converges to a single curve. Remarkably this prediction appeared to be obeyed except for the
purest Sr2 RuO4 crystals
p p with Tc 1:4 K [35]. Another remarkable feature of Eq. (3) is
3=2 1 C ( 3 1)=2. The Abrikosov parameter A is given by
r
hj (r)j4 i R 2 2 1
A = 2 2
= ( 1)mn e 2 R(n +m )
hj (r)j i 2 (1 + 2C 2 )2
Z 1
2
dxe x (1 + C 2 (f(x + x1 )
1 (6)
+f(x x1 ))(f(x + x2 ) + f (x x2 ))

+C 4 f (x + x1 )f(x x1 )f(x + x2 )f(x x2 ))

390 ASPECTS OF p-WAVE SUPERCONDUCTIVITY VOL. 38

FIG. 7. Upper critical field for H k c for several FIG. 8. j (k)j2 is schematically shown in a
. magnetic field parallel to the a-b plane.

where f (x) = x2 p1, x1 = 2R (n m), x2 = 2R (n + m), and R = 1 corresponds to the square

lattice while R = 3 to the triangular lattice. It is readily shown that the square vortex lattice is
favored when C 0:2615, which takes place for t < 0:95. But, perhaps, it is important to point
out there are no orientation energy, when the Fermi surface is a circle. Also, the quasi-particle
density of states and the transport properties should exhibit twofold symmetry [20-21]. However,
in order to make definite prediction it appears necessary to incorporate the tetragonal symmetry
breaking term in the Fermi surface [18, 21].

IV. Vortex state in H ka-b plane

In this particular geometry, the order parameter is given by

p
(k; r) = 2 cos( )(1 + C(ay )2 )j0i (7)

for a magnetiac field not far from the upper critical field Hc2 (T ) (in the GL region for Hc2
H H 0:45 Hc2 [18]). At H ' 0:45 Hc2 [18] the above order parameter transformed into a
1
more symmetric form like cos( ) ! p 1+ 2
(cos( ) + i sin( )) Here is the angle
H makes from the a-axis. So we can think of the order parameter polarized parallel to the field
(see Fig. 8).
Also, the normalized upper critical field h(t) = Hc2 =( @H @t jt=1 ) and C are shown as
c2

functions of t in Fig. 9. We note that h(t) is remarkably linear in t and we find h(0) = 0:806.
Unfortunately there is an error in Fig. 7 in Ref. [21] which is corrected in Ref. [23]. On the other
hand, we do not see any tendency for the reentrant behavior as discussed in Ref. [36]. Of course,
it is possible that the reentrant region may be separated from the vortex state we are considering
now.
VOL. 38 K. MAKI, E. PUCHKARYOV, G.-F. WANG, AND H. WON 391

FIG. 9. h(t) = Hc2 =( @H ~ versus x = 2E

@t jt= 1) and C versus t. FIG. 10. G(x; ) at T = 0 K for
c2
0
~
= 4 , 8 , and 16 .

The quasi-particle density of states in the vortex state is given by [23]

2
~ 2E
N(E; )=N0 = 1 + 2 G (8)
0 0

where
*Z 2
+
2 ~ 1 e u (1 + 2C(u2 21 ))2
~ = p cos () Re
G(x; ) du p (9)
(1 + 2C 2 ) 1 (x u s2 + s0 2 )2
p
where ~ = , 2 2 (T )(1 H=Hc2 ), 0 = 2vv 0 eB, s = sin , ~ s0 = sin , and h: : :i
~
means the average over The function G(x; ) is shown as a function of x with x = 2E 0 for
~ ~
= 4 , 8 , and 16 in Fig. 10. For = 2 , G(x; 2 ) = 0 i.e the effect on the superconductivity
~ reduces to
disappears completely. For E = 0, G(0; )

~ = 2((1 2C)2 2C 2 ) 1 S 2
G(0; ) p : (10)
(1 + 2C 2 ) jsj 1 + s2

We believe that this ~ dependence should be readily accessible by STM for example. Also, the
thermal conductivity is given by [23]
p
2 2
xx = n ' 1 (A0 + A1 cos(2)) (11)
0

p
2 2
xy = ' A1 sin(2) (12)
0
392 ASPECTS OF p-WAVE SUPERCONDUCTIVITY VOL. 38

where ( 0 ) is the quasi-particle scattering rate and A0 and A1 are given by

Z 1 1 + cos(2)
3 2 2 x 2Tx
A0 = 2 dxx sech p F p (13)
2 0 2 s2 + s0 2 0 s2 + s0 2

3
Z 1 x 1 + cos(2) 2Tx

2 2
A1 = 2 dxx sech p F p (14)
2 0 2 s2 + s0 2 0 s2 + s0 2
and

2
u2 1 2
F (u) = e 1 + 2C u =(1 + 2C 2 ) (15)
2

In these expressions s = sin , s0 = sin , and h: : :i means the average over both and . Here
we assumed that the heat current is flowing parallel to the a-axis. The function A0 and A1 are
evaluated numerically and shown in Fig. 11. We note A1 =A0 ' 0:6.
Similarly, the ultrasonic attenuation is also sensitive to this polarization of the order param-
eter [23]. Among many sound waves the transverse wave with (q k x and e k z) and with (q k z
and e k x) are of interest. Then the attenuation coefficient is given by
p
2 2
= n = 1 (B0 + B1 cos(2)) (16)
0
with

Z 1 x
dx 1 + cos(2 ) 2T
B0 = sech2 p F p x 0 (17)
0 2 2 s2 + s0 2 0 s2 + s 2

Z 1 x
dx cos(2)(1 + cos(2)) 2T
B1 = sech2 p F p x 0 : (18)
0 2 2 s2 + s0 2 0 s2 + s 2
The function B0 and B1 are also shown in Fig. 12.
When the magnetic field is rotated away from the direction of the heat current or the
propagation vector of the sound wave, the coefficient exhibits cos(2) or sin(2) dependence. We
believe this will provide a unique way to observe the underlying p-wave symmetry.

V. Concluding remarks

We reviewed recent theoretical results on p-wave superconductors. First of all, the impurity
scattering has a profound effect on the superconductivity. Impurities not only reduce the transition
temperature dramatically but also, generate the quasi-particles on the Fermi level, which should
be readily accessible by low temperature technique.
VOL. 38 K. MAKI, E. PUCHKARYOV, G.-F. WANG, AND H. WON 393

FIG. 11. A0 and A1 versus t. FIG. 12. B0 and B1 versus t.

In a magnetic field, the vortex state exhibits the behavior characteristic of p-wave super-
condcutors. In particular, in a magnetic field parallel to the a-b plane, the polarization of the order
parameter should be accessible to the transport measurement. This will provide a unique way to
probe p-wave symmetry.

Acknowledgments

We would like to thank D. Agterberg, T. Ishiguro, Y. Maeno, Z. Q. Mao, A. C. Mota,

and M. Sigrist for their enlightening discussions on p-wave superconductors. K. Maki thanks P.
Fulde and Max-Planck Institut fur Physik Komplexer Systeme for their hospitality, where a part of
this work is done. H. Won acknowledges the support from Hallym Academy of Science, Hallym
University. This work is in part supported by the National Science Foundation under grant number
9531720.

References
[1] M. Sigrist and T. M. Rice, J. Phys. Soc. Jpn. 62, 4283 (1992).
[2] C. C. Tsuei and J. R. Kirtley, Physica C 282, 4 (1997).
[3] D. J. van Harlingen, Rev. Mod. Phys. 67, 515 (1995).
[4] K. Maki and H. Won, J. Phys. I (Paris) 6, 237 (1996).
[5] M. Sigrist and H. Ueda, Rev. Mod Phys. 63, 239 (1991).
[6] R. H. Heffner and M. R. Norman, Comments on Cond. Matt. Physics 17, 361 (1996).
[7] T. Ishiguro and K. Yamaji, Organic Superconductors (Springer-Verlag, Berlin 1990).
[8] M. Lang, Superconductivity Review 2, 1 (1996).
[9] Y. Maeno et al., Nature 372, 532 (1994).
[10] Y. Maeno, Physica C 282, 206 (1997).
[11] M. Sigrist et al., Physica C 317, 134 (1999).
[12] K. Ishida et al., Nature 396, 658 (1998).
[13] D. Vollhardt and P. Wolfle, Superfluid Phases of Helium Three (Taylor-Francis, London 1990).
394 ASPECTS OF p-WAVE SUPERCONDUCTIVITY VOL. 38

[14] S. Nishizaki et al., J. Phys. Soc. Jpn. 67, 560 (1998).

[15] Y. Maeno et al., J. Superconductivity 12, 535 (1999).
[16] A. P. Mackenzie et al., Phys. Rev. Lett. 80, 161 (1998).
[17] D. F. Agterberg, M. T. Rice and M. Sigrist, Phys. Rev. Lett. 76, 3374 (1997).
[18] D. F. Agterberg, Phys. Rev. Lett. 80, 5184 (1998).
[19] D. F. Agterberg and R. Heeb, Phys. Rev. B 59, 7076 (1998).
[20] G.-f. Wang and K. Maki, Europhys. Lett. 45, 71 (1999).
[21] K. Maki, G.-f. Wang and H. Won, J. Superconductivity 12, 551 (1999).
[22] T. M. Riseman et al., Nature 396, 242 (1998).
[23] K. Maki and H. Won Physica B (in press).
[24] K. Maki and E. Puchkaryov, Europhys. Lett. 45, 263 (1999); E. Puchkaryov and K. Maki, preprint.
[25] M. Matsumoto and M Sigrist, J. Phys. Soc. Jpn. 68, 724 (1999).
[26] K. Maki et al., Physica C 317, 353 (1999); K. Maki et al., Synth. Metals 103, 1933 (1999).
[27] L. J. Buckholtz and G. Zwicknagl, Phys. Rev. B 23, 5788 (1981).
[28] S. Schmitt-Rink, K. Miyake and C. Varma, Phys. Rev. Lett. 57, 2577 (1986).
[29] P. Hirschfeld, D. Vollhardt, and P. Wolfle, Solid State Comm. 59, 2575 (1986).
[30] Y. Sun and K. Maki, Phys. Rev. B 51, 6059 (1995).
[31] T. Hotta, J. Phys. Soc. Jpn. 62, 274 (1993).
[32] J. Bardeen, L. N. Cooper and J. R. Schrieffer, Phys. Rev. 108, 1175 (1957).
[33] J. Bardeen, G. Rickayzen and L. Tewordt, Phys. Rev. 113, 982 (1959).
[34] I. A. Lukyanchuk and V. P. Mineev, Soviet Phys., JETP 66, 1168 (1987).
[35] Z. Q. Mao, Y. Mori, and Y. Maeno, Phys. Rev. B 60, 610 (1999).
[36] A. G. Lebed and K. Yamaji, Phys. Rev. Lett. 80, 2697 (1998).