Comment on "Anomalous proximity effect in underdoped YBa_2Cu_3O_{6+x} Josephson junctions"

1 Pages·0.08 MB·English

by  J. Quintanilla

Comment on “Anomalous proximity effect in underdoped YBa Cu O Josephson
2 3 6+x
junctions”
1 2 1
J. Quintanilla, K. Capelle, and L.N. Oliveira
1 Departamento de F´ısica e Inform´atica, Instituto de F´ısica de S˜ao Carlos,
3 Universidade de S˜ao Paulo, Caixa Postal 369, 13560-970 S˜ao Carlos, S˜ao Paulo, Brazil
0 2Departamento de Qu´ımica e F´ısica Molecular, Instituto de Qu´ımica de S˜ao Carlos,
0 Universidade de S˜ao Paulo, Caixa Postal 780, 13560-970 S˜ao Carlos, S˜ao Paulo, Brazil
2 (Dated: February 2, 2008)

n
a
J RecentlyDeccaetal. havereportedanunusuallylong- 8

8
rangedproximity effect (PE) between the superconduct-
2
ing and insulating phases of YBCO [1]. These beautiful 4

] experiments have receivedlots of attention because they
n may help to discriminate between alternative scenarios 0
ζ 2
o /
forcupratesuperconductivity[1,2]. However,laterwork ζ
c
- [2] took for granted the claim [1] that such long range is
r anomalous. Herewepointoutthatthisisnotwarranted. 1
p
u At the heart of our argument is the formula used in
s [1] to estimate the expected range of the PE, ζ ≈ ζ0 ≡ 0.5
. 0 0.2 0.4 0.6 0.8 1
mat tphi¯hsDe/x2pπreksBsiTo,nw[1h]eroebDtainisstζhe=di1ff5usnimon, cwohniscthanist.inUdseinedg NNVN/NSVS
much smaller than the experimentalvalue ζexp =90nm. FIG. 1: Dependence of the range of the proximity effect
d- However, the same formula agrees with earlier experi- on the parameters of the interface. Solid lines: results ob-
n ments on YBCO [3], which poses the question why it tained from Eq. (1), valid for NS = NN ≡ N. Dashed
lines: results obtained from Eq. (4.15b) of [5], valid also
o fails to account for the data of [1]. To answer we first
[c note that ζ ≈ ζ0 is only the VN → 0 limit of the more foofrNNSSVS6==N0.N1,.0.2E,a0c.h3,0li.n4e(fcroormresbpootntodms ttooatopd)iff.erent value
general expression [4]

1
∞ ′ ′
v ζ dη Σ NV Σ 2
= − + N + (1) In summary, although there are unconventional fea-
45 ζ0 Z0 πη (cid:20) Σ 1−VN/VS +NVNΣ η(cid:21) tures in the experiments, the claim that the long range
5 for an interface of a superconductor and a normal con- of the PE is one of them [1] seems presently unfounded.
1 ductor with coupling constants V and V , respectively. Conventional theory [4, 5] predicts comparably long
0 S N
Σisafunctiondefinedin[4]andN isthedensityofstates ranges, once the novel nature of the experiments is ac-
3
0 at the Fermi level. A numerical evaluation of Eq. (1) is counted for. We thus recommend performing similar ex-
/ presented in Fig. 1. The key point is that in the exper- periments on conventional superconductors. The lines
t
a iments of [1] superconductivity is induced in part of an of thought on cuprate superconductivity stemming from
m underdoped sample by photodoping. The PE thus takes theassumptionthatζexp isanomalouslylarge[1,2]must
- placenotbetweentwodifferentmaterials(asinstandard also be critically reexamined.
d
experiments), or in an inhomogeneous material a region This work was funded by FAPESP and CNPq.
n
o of which has been severely damaged in the creation of
c the normalregion(asin[3]),butbetweentwodifferently
v: doped parts of one and the same material. The limit
i VN →VS is thenmoreappropriatethanVN →0. Inthis
X [1] R.Decca,H.Drew,E.Osquiguil,B.Mairov,andJ.Guim-
limit Eq. (1) does predict ζ ≫ ζ0, suggesting that the pel, Phys. Rev.Lett. 85, 3708 (2000).
r
a long range is not an unconventionalfeature of YBCO. [2] Q.-H.WandandA.V.Balatsky,Phys.Rev.B64,092504
At VN = VS Eq. (1) predicts ζ → ∞. This infinity (2001). H.-J. Kwon, Phys. Rev. B 63, 134511 (2001).
reflects the fact that in Eq. (1) the density of states is J. Burgy et al., Phys.Rev.Lett. 87, 277202 (2001).
assumed to be the same on both sides of the interface [3] W.Booijetal.,Phys.Rev.B55,14600(1997).A.J.Pauza
(N = N = N), so that for V = V both materials et al., J. Appl. Phys. 82, 5612 (1997). A. Katz et al., J.
N S N S Appl.Phys. 87, 2978 (2000).
are identical. The restriction to N = N can be over-
N S
[4] G. Deutscher and P. de Gennes, in Superconductivity, ed.
come in the one-frequencyapproximation[5]. Numerical
R. Parks(Dekker,New York,1969).
evaluation of Eq. (4.15b) of [5] yields the dashed curves [5] P. deGennes, Rev. Mod. Phys. 36, 225 (1964).
in Fig. 1. Clearly, a large, but finite value of ζ/ζ0 is [6] NS >∼ NN is suggested by the low Tc and small relative
obtained, e.g., if VN ≈VS and NS >∼NN [6]. change in reflectancebetween theN and S regions.