Coexistence of Superconductivity and Antiferromagnetism in Multilayered High-$T_c$ Superconductor HgBa$_2$Ca$_4$Cu$_5$O$_y$: A Cu-NMR Study

0.5 MB·English

by  H. Kotegawa

Coexistence of Superconductivity and Antiferromagnetism in Multilayered High-T
c
Superconductor HgBa Ca Cu O : A Cu-NMR Study
2 4 5 y
H. Kotegawa1,4,∗,†, Y. Tokunaga1,‡, Y. Araki1, G. -q. Zheng1, Y. Kitaoka1,4
K. Tokiwa2,4, K. Ito2, T. Watanabe2,4, A. Iyo3,4, Y. Tanaka3,4, and H. Ihara2,3,4∗
4 1Department of Physical Science,
0 Graduate School of Engineering Science,
0 Osaka University, Toyonaka, Osaka 560-8531, Japan
2 2Department of Applied Electronics,
Science University of Tokyo, Yamazaki,
n
Noda, Chiba 278-8510, Japan
a 3National Institute of Advanced Industrial Science and Technology (AIST),
J
Umezono, Tsukuba 305-8568, Japan
2 4Core Research for Evolutional Science and Technology (CREST)
2
of the Japan Science and Technology Corporation (JST),

Kawaguchi, Saitama 332-0012, Japan
]
n
o (Dated: February 2, 2008)
c
- We report a coexistence of superconductivity and antiferromagnetism in five-layered compound
pr HgBa2Ca4Cu5Oy (Hg-1245) with Tc =108 K,which is composed of twotypesof CuO2 planes in a
unit cell; three inner planes (IP’s) and two outer planes (OP’s). The Cu-NMR study has revealed
u
s that the optimallydoped OP undergoes a superconducting (SC) transition at Tc =108 K, whereas
. the three underdoped IP’s do an antiferromagnetic (AF) transition below TN ∼ 60 K with the Cu
t
a moments of ∼(0.3−0.4)µB. Thus bulk superconductivity with a high value of Tc =108 K and a
m static AF ordering at TN = 60 K are realized in the alternating AF and SC layers. The AF-spin
- polarization at the IP is found to induce the Cu moments of ∼0.02µB at the SC OP, which is the
d AFproximity effect intothe SCOP.
n
o PACSnumbers: 76.60.Cq,71.27.+a,75.20.Hr,76.60.Es
c
[

I. INTRODUCTION La1.85Sr0.15CuO4 and AF La2CuO4. [5] Thus, these is-
1
sues have not been settled yet.
v
6 Multilayered high-Tc cuprates, which have more than
1 There remain a number of underlying issues to be threeCuO2planesinaunitcell,exhibitveryuniquemag-
4 resolved in high-T cuprates. One of underlying is- neticandSCpropertiesbecausetheyincludetwotypesof
c
1 sues is an interplay of antiferromagnetism and super- CuO2planes. AsindicatedinFig.1,anouterCuO2plane
0
conductivity in the antiferromagnetic (AF) - supercon- (OP) has a pyramidal five-oxygencoordination,whereas
4
ducting (SC) phase boundary, near vortex cores un- an inner plane (IP) has a square four-oxygen one. Note
0
/ der magnetic field, and AF-SC alternate layered struc- thattheIP∗isthemiddleplaneofthethreeIP’sasshown
at tures. Lake et al. reported that the AF correlations inthefigure. Nuclear-magnetic-resonance(NMR)exper-
m are induced in vortex cores and extend over the cores iments revealed that the OP and the IP differ in the
into the SC region in La Sr CuO under the mag- doping level.[7, 8, 9, 10] We reported unusual magnetic
- 2−x x 4
d netic field, that is, an AF proximity effect into SC and SC characteristics in multilayered CuO2 planes in
n state. [1] This result is supported by some theoretical Hg and Cu-based high-Tc cuprates through 63Cu-NMR
o approaches.[2,3]Inthesetheoreticalpredictionsbasedon measurements. [11] The Knight shift (63K) at the OP
c
the SO(5) symmetry model, [4] each AF and SC fluctu- and the IP exhibits different characteristic temperature
:
v ation canextend into other region,when both the states (T) dependence, consistent with its own doping level. It
Xi come across. Recently, however, Bozovic et al. showed was shown that the doping level Nh(OP) at the OP is
r that the superconductivity does not mix into the AF in- larger than Nh(IP) at the IP for all the systems and its
a sulator in the superconductor-insulator-superconductor difference ∆Nh =Nh(OP)−Nh(IP) increases as either a
heterostructures realized by stacking each layer of SC total carrier content or n increases. At ∆Nh’s exceed-
ing a critical value, the respective SC transitions do not
simultaneously set in at the IP and the OP. [12] Some
theoretical approaches predict the effect induced by the
∗† Present Address: Department of Physics, Okayama Uni- carrier inhomogeneity. [13, 14]
versity. Tsushima-naka 3-1-1, Okayama 700-8530 Japan; In this paper, we report 63,65Cu-NMR study on Hg-
Electronic address: [email protected]; ‡
1245 which evidences a coexistence of AF order at the
Present Address : Japan Atomic Energy Research Institute,
IP’sandIP∗ andbulksuperconductivityattheOP.Note
Tokai-mura, Ibaraki 319-1195, Japan; Electronic address:
[email protected] thattheIP∗isthemiddleplaneofthethreeIP’sasshown
2
Hg
(cid:7)
Ba (cid:20) (cid:29)(cid:26)(cid:31) # (cid:2)(cid:0)(cid:0)$(cid:22)
OP (SC) Ca (cid:19)(cid:14) !
Cu (cid:17)(cid:18) (cid:3)(cid:6)
IP (AF : 0.30µB) O (cid:15)(cid:16) %&
(cid:13)(cid:14) (cid:5)
(cid:14890)9.7(cid:904) IP* (AF : 0.37µ ) (cid:12)(cid:11) (cid:3) (cid:21) (cid:2)(cid:0)"(cid:30)
B (cid:10) ’%& !
(cid:9)
(cid:8)
IP (AF : 0.30µB) (cid:3)(cid:4)(cid:0) (cid:1)(cid:0) (cid:2)(cid:0)(cid:0) (cid:2)(cid:1)(cid:0)
(cid:21)(cid:22)(cid:23)(cid:24)(cid:22)(cid:25)(cid:26)(cid:27)(cid:28)(cid:25)(cid:22)(cid:29)(cid:30)(cid:31)
OP (SC)
(cid:20)7 (cid:29),(cid:31) $- .-/.-0
6
(cid:14890)12.5(cid:904) 85
(cid:16) )(cid:0)(cid:0)(cid:30)
<=
OP (SC) :;
(cid:9) (cid:2)+(cid:0)(cid:30)
9
IP (AF : 0.30µB) 867
5
(cid:14890)9.7(cid:904) IP* (AF : 0.37µ ) 65(cid:14) (cid:2)(cid:0)(cid:0)(cid:30)
B 4
3
IP (AF : 0.30µ ) 2 )(cid:0)(cid:30)
B 1
OP (SC) (cid:2)(cid:1)() (cid:2)(cid:1)(* (cid:2)(cid:1)(+ (cid:2)(cid:1)((cid:1)
#(cid:29)(cid:21)(cid:31)
FIG. 1: The crystal structure of Hg-1245 (a = 3.850 ˚A, c = FIG.2: (a) T dependenceof dcsusceptibility with fieldcool-
22.126 ˚A) (Ref.6). The OP undergoes the SC transition at ing (FC) and zero-field cooling (ZFC). A clear diamagnetic
Tc = 108 K, whereas the three underdoped IP’s do an AF signal can be observed below Tc = 108 K. (b) The NMR
transition belowTN ∼60Kwith therespectiveCumoments spectra for H ⊥ c at T = 200, 140, 100, and 20 K. The
of ∼0.30µB and 0.37µB at theIP and theIP∗. NMR signals at the IP and the IP∗ disappear below ∼ 150
K.The OP’s signal becomes significantly broader at temper-
atures lower than TN ∼60 K.
in Fig. 1. Measurements of the Knight shift 63K, the
nuclear-spin-lattice-relaxationrate (1/T ) and the inter-
1
nal field (H ) of 63,65Cu have revealedthat the OP un-
int was already reported in the previous literatures.[11, 16]
dergoes a bulk SC transition below T = 108 K and the
c TheNMRspectrumattheIPexhibitsasharperspectral
IP* and IP order antiferromagnetically below T ∼ 60
N width with a smaller Knight shift than those at the OP.
K with the Cu moments of 0.37µ and 0.30 µ , respec-
B B The spectral width at the IP is estimated to be ∼50 Oe
tively.
for H k c, comparable to the ∼ 60 Oe for YBa Cu O
2 3 7
under H ∼ 15 T, which is the narrowest among high-
T cuprates to date. This ensures that the IP is rather
c
II. EXPERIMENTAL DATA AND DISCUSSION homogeneously doped. The spectra at the IP and the
IP∗ overlapeachother,suggestingthattheirlocaldoping
Polycrystalline sample was prepared by the high- levelsarenotsomuchdifferent. TheNMRsignalsatthe
pressure synthesis technique as described elsewhere. [15] IP and the IP∗ disappear due to their short relaxation
Powder x-ray-diffraction experiment indicates that the time below T ∼150 K.
sample consists of almost a single phase, but includes a Figure 3 indicates the T dependence of 63K at the
ab
small fraction of Hg-1234.[15] A SC transition tempera- IP’s and the OP for H ⊥ c . In general, K(T) con-
tureofTc =108KwasdeterminedfromanonsetT below sists of the T-independent orbital part, Korb, and the
which diamagnetic signal appears in dc susceptibility as T-dependent spin part, K (T), that is proportional to
s
shown in Fig. 2(a). For NMR measurements, the pow- the uniform susceptibility χ . K (OP) decreases be-
s s
dered sample was aligned along the c axis at H = 16 T. low T∗ ∼ 160 K, followed by a rapid decrease around
TheNMRexperimentwasperformedbytheconventional T =108 K which is indicative of the bulk superconduc-
c
spin-echo method at 174.2 MHz (H ∼ 15.3 T). tivityattheOP,asalsoconfirmedinthemeasurementof
Figure 2(b) shows NMR spectra for H ⊥c. At 200 K, 1/T T. The behavior of K (T) suggests that the OP is
1 ab
two well-separated peaks arise from the OP and the IP. almostoptimallydopedfroma comparisonwiththe pre-
The assignmentofNMR spectrumto the OPandthe IP vious study. [11] In the inset of Fig. 3, K (T) vs K (T)
ab c
3
plots are presented at the OP and the IP’s. The spin MHz,respectively(notshown). Therefore,allthesespec-
part in the measuredshift K (T) at the CuO plane is tra are affected by the presence of internal field H as-
s,α 2 int
expressed as following, [17] sociated with the onset of AF order.
K =(A +4B)χ (α=a,b,and c axis),
s,α α s,α
whereAα andB aretheon-siteandthesupertransferred _Q‘aZb @A
horyipgeinrfiatnien-gcofurpomlingthceondsitpaonltes.anAdαtihseasnpisino-tororbpiitc,inmtaerinalcy- HPEDO cdef @AB
tions for Cu-3d orbitals, and B is isotropic, originat- MN
L
K
ing from the Cu(3dx2−y2)-O(2p)-Cu(4s) covalent bond- JI
ing. χ isthespinsusceptibility. Fromalinearrelation E
s,α HG
in the figure, (A +4B)/(A +4B) ∼ 0.267 and 0.379 D ^A
c ab F
are estimated at the IP’s and the OP, respectively. By ED
C
assuming A ∼ −170 kOe/µ and A ∼ 37 kOe/µ in
c B ab B
YBa Cu O ,[18,19,20]therespectivevaluesofB atthe
2 3 7
IP’sandtheOPinHg-1245areestimatedasB(IP)∼61 6 76 86 96 :6 ;6 <6 =6 >6 ?6 766776
kOe/µB and B(OP) ∼ 74 kOe/µB. These values of B QRSTUSVWX YZ[\]
are larger than the typical value of ∼ 40 kOe/µ ob-
B
tainedinLa Sr CuO ,YBa Cu O ,andYBa Cu O ,
[18, 19, 20, 22−1,x22x, 23] s4uggest2ing3tha7t the Cu(32dx2−4y28)- FIG. 4: A zero-field 63,65Cu-NMR spectra at 1.4 K. Four
O(2p)-Cu(4s) covalent bonding in Hg-1245 is stronger peaksinf =55−110MHz,whichconsistoftwosites(IPand
IP∗) and two isotopes [63Cu (solid arrow) and 65Cu (dashed
than in the La- or Y-based systems.
arrow)], are observed. The spectra in f =10−40 MHz cor-
respond to the OP. The solid lines are the simulation cal-
culated by using 63νQ(IP) = 8.37 MHz, 63νQ(OP) = 16.05
(cid:0)(cid:4)(cid:6)
MHz,andHint alongabplane. EachCumomentisestimated
(cid:16)(cid:23) (cid:13)(cid:14) as M(IP) ∼ 0.30µB and M(IP∗) ∼ 0.37µB (see text). The
internal field of Hint∼0.54 T exists even at the SCOP.
(cid:15)(cid:14)
The nuclear Hamiltonian H=H +H at H =0 be-
(cid:0)(cid:4)(cid:5) (cid:16)(cid:17)(cid:18)(cid:19)(cid:20)(cid:21)(cid:22) Q Z
low T is described in terms of the Zeeman interaction
(cid:12) N
(cid:11) due to H and the nuclear electric quadrupole interac-
(cid:10) int
(cid:9) )%&’ tion as follows:
(cid:8)
(cid:7)
(cid:0)(cid:4)(cid:2) 45 )%& H =−γ ~ 1H (I +I )+H I ,
(cid:16)(cid:24) 132)%*’ Z n (cid:26)2 ⊥ + − k z(cid:27)
)%* whereH andH aretherespectivecomponentsperpen-
⊥ k
$%&’ $%( $%(’ $%’ dicularandparalleltothec-axisandγ istheCunuclear
n
+,-./0
gyromagnetic ratio, and
(cid:0)
(cid:0) (cid:1)(cid:0)(cid:0) (cid:2)(cid:0)(cid:0) (cid:3)(cid:0)(cid:0)
(cid:25)(cid:26)(cid:27)(cid:28)(cid:26)(cid:29)(cid:30)(cid:31) (cid:29)(cid:26)!"# H = e2qQ {[3I2−I(I +1)]+η(I2−I2)},
Q 4I(2I−1) z x y
whereη isthe asymmetryparameterofelectric-fieldgra-
FIG. 3: T dependence of Knight shift K for H ⊥ c. The
ab
dient. Here, note that the quadrupole frequency hν ≡
winistehtothfeFitge.m3(pbe)rasthuorwesaKsaabnvismKpclipciltotpsaartatmheetIePr.’sTanhdesteheplOotPs 3e2qQ/2I(2I−1) and η ∼0. Q
allow us to estimate the supertransferred hyperfine-coupling The spectra observed in f = 55 − 110 MHz corre-
constants B(IP) ∼ 61 kOe/µB and B(OP) ∼ 74 kOe/µB, spond to the case for νQ ≪ Hint due to the AF or-
respectively. der below T ∼ 60 K. Four peaks are understood as
N
the central peaks (1/2 ↔ −1/2 transition) of 63,65Cu at
Next, we present firm evidence for the occurrence of the IP and the IP∗. A ratio of 63,65Cu-NMR intensity
AF ordering at the IP and the IP∗. Figure 4 shows at low frequency to high frequency (I /I ∼ 2) sug-
L H
63,65Cu-NMR spectra at H = 0 and T = 1.4 K. Four gests that the two peaks at low (high) frequencies arise
and two peaks are observed in the frequency ranges of from the IP (IP∗). The satellite peaks (±1/2 ↔ ±3/2
f =55−110and10−40MHz, respectively. The nuclear transition)due to the electricquadrupoleinteractionare
quadrupole frequencies at the IP and the OP’s, ν (IP) not well resolved. By incorporating this intensity ratio
Q
and ν (OP), are estimated from the NMR experiments I /I ∼ 2, 63ν (IP) = 8.37 MHz, 63Q/65Q ∼ 1.08,
Q L H Q
athighT as63ν (IP)=8.37MHzand63ν (OP)=16.05 and 63γ /65γ ∼ 0.93, the NMR spectra at the IP and
Q Q n n
4
the IP∗are simulated as the solid line in the figure, giv- 160K.1/T (OP)isdistributedbelow∼90K.Inthenor-
1
ing rise to the respective values of H (IP) = 6.1 T mal state, a recovery curve of nuclear magnetization is
int
and H (IP∗) = 7.7 T. These values at the IP and consistentwithatheoreticalonefordeterminingasingle
int
the IP∗ allow us to estimate the Cu moments M(IP) ∼ value of T as seen in Fig. 6(a).[27] Below ∼90 K, how-
1
0.30µ and M(IP∗) ∼ 0.37µ by using a hyperfine- ever,ashortcomponentintherecoverycurveisobserved
B B
coupling constant (A − 4B) ∼ −207 kOe/µ where aspresentedinFig.6(b). Atentativefittingtothecurve,
ab B
B(IP) ∼ 61 kOe/µ . These Cu moments are one-half which is indicated by a solid line, allows us to estimate
B
smaller than 0.64µ estimated in La CuO .[24] We re- a short and a long component in T . Their T dependen-
B 2 4 1
mark that the N (IP) and N (IP∗) are tentatively esti- ciesareshowninFig.7,wherethe shortcomponentsare
h h
mated as [5δ −2N (OP)]/3 ∼ 0.057±0.02 by using presented by open square.
av h
anaverageholecontentδ =0.12evaluatedfromaHall
av
measurement.[25]NotethatN (OP)=0.212−0.217was
h +
estimated via the systematic experimental relation be-
tweentheN andtheK atroomtemperaturearguedin ,-. /012-345-56 +7(8
h s
the literature, [11] and hence it is shown that the OP is
optimally doped. Here we assumed K for Hg-1245 to
orb
()+
be0.19−0.20fromacomparisonwithothermultilayered
cuprates.[16, 26]
The spectra in f = 10−40 MHz suggest the case for @?
νQ ∼ 16 MHz ≃ Hint. Actually, the calculated spectra =>
ttohebfiegcuorne,siastlleonwtinwgithustthoe eesxtpimeraitmeenνt a∼re1i6ndMicHatzedanind =>@CDB ()(++( ()* +
Q
Hint = H⊥ ∼ 0.54 T. These spectra are hence assigned @?A ,E.FG45-56 H(8
as arising from the OP. This Hint of 0.54 T is far larger =>
than the calculated dipole field in the OP of ∼ 70 Oe, <
whichisinducedbytheCumomentsofM(IP)∼0.30µ ()+
B
and M(IP∗) ∼ 0.37µ . H (OP) ∼ 0.54 T corresponds
B int
to the Cu moments of ∼0.02µ .
B
(cid:2)(cid:0) ()(+( + 9 : 7 *
5 ,246;.
(cid:26) (cid:1)(cid:4) (cid:29)(cid:28)
(cid:19)
(cid:24)(cid:25) (cid:30)%&!’"$ FIG. 6: The recovery curve of nuclear magnetization for de-
(cid:19)
(cid:24)(cid:23) termining T1 at the OP at (a) the normal and (b) the SC
(cid:21)(cid:22) state, respectively. A short component in the recovery curve
(cid:20) (cid:1)(cid:0) is observed as presented below ∼90 K in theSC state.
(cid:18)
(cid:18)(cid:19)
(cid:27)(cid:28)
(cid:17) (cid:30)(cid:31) !"#$
Generally in the SC mixed state under magnetic field,
(cid:16) (cid:4)
the shortcomponentin T is believed to arisedue to the
1
presence of the normal state in vortex cores. But the
large fraction of the short component reaching ∼80% is
(cid:0)
(cid:0) (cid:1)(cid:0)(cid:0) (cid:2)(cid:0)(cid:0) (cid:3)(cid:0)(cid:0) unusual,givingrisetoalmostasamefractionastheshort
componentatH =0indicatedbyopentriangleinthefig-
(cid:5)(cid:6)(cid:7)(cid:8)(cid:6)(cid:9)(cid:10)(cid:11)(cid:12)(cid:9)(cid:6) (cid:13)(cid:14)(cid:15)
ure. These results ensure that the shortcomponent dose
not arise due to the presence of vortex cores, but origi-
nates from the unexpected relaxation process at the SC
OP. Thus, some low-lying magnetic excitations survive
FIG.5: T dependenceof1/T1T forH kc. T1(OP)below∼90
at the SC OP even though the d-wave superconductiv-
K is the long component in the recovery curve. The pseudo-
gap behavior is observed at the optimallydoped OP below ity is formed well below Tc = 108 K. The short compo-
T∗ ∼160K.Ontheotherhand,theunderdopedIPdoesnot nent indicates two peaks at TN ∼ 60 K and T? ∼ 25 K,
showanypseudogapindication,revealingthatthelow-energy whereas the long component indicates a peak at T? ∼25
spectral weight in χ(q= Q,ω) is critically enhanced around K.T ∼60K is indicativeof anonsetofAF orderingat
N
ω∼0 toward AFordering at TN ∼60 K. theIP’s,corroboratedbytheincreaseofKs(OP). Recent
muonspin resonance measurementalso evidences an AF
Figure 5 indicates the T dependence of 1/T T of ordering below ∼ 60 K in this material. [28] T ∼ 25 K
1 ?
63Cu at the IP’s and the OP for H k c. Remarkably, might be related to the occurrence of H (OP) because
int
1/T T(OP) exhibits a pseudogap behavior below T∗ ∼ both the long and short components show peaks.
1
5
of the presence of AF state. Hg-1245 is a good candi-
(cid:0)(cid:1)(cid:7)
(cid:18)(cid:20) (cid:21)(cid:22) date to address this issue. Both the measurements of
(cid:18)(cid:19)
T and Knight shift evidence that the OP is in the SC
1
(cid:0)(cid:1)(cid:6) (cid:27)(cid:22) (cid:23)(cid:24)(cid:25)(cid:1)(cid:26) (cid:27)(cid:22) state. If the AF-spin polarization at the IP induced the
H (OP)=0.54 T via the hybridizationbetween 4s(IP)
int
(cid:18)(cid:28) and 4s(OP) and/or 3d3z2−r2(OP) [not 3dx2−y2(OP)],
(cid:0)(cid:1)(cid:3) a ratio of [T (OP)/T (IP)] ∼ [H (IP)/H (OP)]2 ∼
1 1 int int
(cid:17) [6.1T/0.54T]2 ∼ 102 would be expected. It is, however,
(cid:11)(cid:16)(cid:15) surprising that the T at the SC OP is 103 times shorter
(cid:14) 1
(cid:13) (cid:0)(cid:1)(cid:4) thanattheantiferromagneticallyorderedIPatlowT far
(cid:12)
(cid:10)(cid:11) below TN, being [T1(OP)/(T1(IP)] ∼ 10−3. This result
(cid:9) demonstrates the existence of low-lyingmagnetic excita-
(cid:0)(cid:1)(cid:5)
(cid:8) tions inherent to the OP associated with a possible in-
(cid:21)(cid:22) (cid:23)(cid:24)(cid:25)(cid:1)(cid:26) terplay with the superconductivity. It suggests that the
(cid:0)(cid:1)(cid:2)(cid:4) weak AF order with small moment 0.02 µB is respon-
(cid:27)(cid:22) (cid:29)(cid:30)(cid:31) ! "(cid:31)#$(cid:31)%&%! sible for the low-lying magnetic fluctuations at the OP
(cid:27)(cid:22) ’(cid:31)%( "(cid:31)#$(cid:31)%&%! and coexists with the SC state at the OP. We note that
(cid:0)(cid:1)(cid:2)(cid:3) this coexistence seems to be similar to the phenomenon
(cid:0) (cid:0)(cid:1) (cid:0)(cid:1)(cid:1)
near vortex cores where the AF correlations originating
(cid:18)&#$& )!* & (cid:23)+(cid:26) from the vortex cores extend over the cores into the SC
region.[1] On the other hand, it is quite interesting issue
whether the SC order parameter exists at the metallic
FIG. 7: T dependence of 1/T1 for H k c. 1/T1 below TN
AF IP in Hg-1245,but this is a future issue,because the
can be measured at zero field. 1/T1(OP) shows the peak at
TN ∼ 60 K associated with AF ordering at the IP. 1/T1(IP) NMR signal from the AF IP disappears in T =40−150
shows a T1T ∼ const relation far below TN, indicating that K.
the IP is metallic. [T1(OP)/(T1(IP)]∼ 10−3 at low T shows FinallywementionedsomeresultsofTlBa2Ca4Cu5Oy
the existence of low-lying magnetic excitations inherent to (Tl-1245) with T = 100 K, which is slightly much
c
theOP,whichisassociatedwithapossibleinterplaywiththe overdoped than Hg-1245. As shown in Fig. 8(a), the
superconductivity.
1/T T(OP) of Tl-1245 does not show the pseudogap be-
1
havior, indicating that the OP is in overdoped regime.
1/T T(IP) increases monotonically with decreasing T
1
Ontheotherhand,1/T T(IP)increasesmonotonically, and the signals of the IP and the IP* disappear below
1
whereasK (IP)decreaseswithdecreasingT downto150 ∼140 K as well as Hg-1245. The low-lying magnetic ex-
s
K. Instead of the pseudogap, unexpectedly, the NMR citations are induced at the SC OP also in Tl-1245, and
signals at the IP and the IP∗ disappear below ∼ 150 K. thus T1(OP) distributes below ∼60 K. Its short compo-
Thisisbecausethelow-energyspectralweightindynam- nents show a peak at ∼45 K, correspondingto the peak
ical response function χ(q=Q,ω) is critically enhanced of TN ∼ 60 K of Hg-1245, as shown in Fig. 8(b). This
aroundω ∼0. Here Q is the AF wavevector(π/a,π/a). TN ∼ 45 K at the IP of Tl-1245 suggests that the IP
Eventually, the IP’s order antiferromagnetically below ofTl-1245has somewhatmuchcarriercontentthanthat
T as evidenced from the zero-field AF NMR experi- of Hg-1245. Interestingly, however, the second anomaly
N
ment that probes the Cu moments of M(IP) ∼ 0.30µB corresponding to T? disappears in Tl-1245, and the line
and M(IP∗) ∼ 0.37µ . The measurement of T at the width of the OP does not change between 150 K and
B 1
IP reveals a behavior of T T ∼ const below ∼ 20 K in 20 K, which is quite contrast to Hg-1245 as shown in
1
the SC state at H = 0 as seen in Fig. 7. In magneti- Figs. 8(c) and 2(b). This ensures that the internal field
cally orderedmetals, a nuclear-spin-relaxationprocess is at the OP of Tl-1245 is quite tiny even below TN ∼ 45
mediated by the interaction between nuclear spins and K at the IP, which is different from Hint(OP) ∼ 0.54 T
conduction electrons via spin-wave excitations, which is of Hg-1245. These things imply that Hint(OP)∼0.54 T
called as the Weger mechanism leading to a behavior of is induced below T? ∼25 K in Hg-1245.
T T = const at low T.[29, 30] Thus the IP is suggested
1
to be not in an insulating regime but a metallic one,
which is consistent with the estimated hole content of III. SUMMARY
∼ 0.057±0.02. The small value of T T = const indi-
1
cates that a large gap opens in the magnetic excitation, Insummary,the63,65Cu-NMRmeasurementshaveun-
indicating that the AF ordering in the IP is in a static
raveled that the disparate electron phases emerge at
regime.
the outer two CuO planes and the inner three ones in
2
When the SC state is closely faced to the AF state HgBa Ca Cu O . The T dependencies of Knight shift
2 4 5 y
realized in the doped CuO plane, it is not obvious to and 1/T have revealedthat the optimallydoped OP un-
2 1
what extent the superconductivity is affected because dergoes the bulk SC transition at T = 108 K and the
c
6
underdoped IP’s do the AF transition below T ∼ 60
(cid:10)(cid:29)(cid:15)(cid:16) (cid:27)(cid:1) (cid:23)"(cid:25) (cid:20)(cid:21) (cid:20)%&(cid:0)(cid:27)’( K63,6w5Cithuo-NutMaRnyexinpdericimateinontsoaftplsoewudToghaapv.eTprhoevizdeNerdo-fifiermld
(cid:10)(cid:15)(cid:13)(cid:14) (cid:19)(cid:18) evidence that the respective AF moments at the IP and
(cid:11)(cid:12) (cid:0)(cid:1) theIP∗ areM(IP)∼0.30µB andM(IP∗)∼0.37µB. The
(cid:9)
(cid:9)(cid:10) (cid:17)(cid:18) bulk superconductivity with the high value of Tc = 108
(cid:8) K and the static AF ordering at TN = 60 K take place
(cid:7) even though the AF and SC layers are alternatively
(cid:1)
(cid:1) (cid:0)(cid:1)(cid:1) (cid:27)(cid:1)(cid:1) (cid:28)(cid:1)(cid:1) stacked with the respective thickness being comparable
(cid:20)(cid:30)(cid:31) (cid:30)!"#$!(cid:30)(cid:23)(cid:26)(cid:25) with∼9.7˚Aand∼12.5˚A. TheAF-spinpolarizationat
the IP is found to induce the Cu moments of ∼ 0.02µ
(cid:0)(cid:1)(cid:6) (cid:23)(cid:24)(cid:25) (cid:20)(cid:21) (cid:19)(cid:18) (cid:23)-(cid:25) at the OP that fluctuates faster than the AF moment aBt
the IP does, evidencing the AF proximity effect into the
(cid:20)(cid:22) (cid:0)((cid:1)(cid:26)
SC OP.
(cid:16) (cid:0)(cid:1)(cid:5) (cid:17)(cid:18) (cid:17)(cid:18) (cid:19)(cid:18)*(cid:19)(cid:18)+
(cid:10)(cid:15)(cid:14)
(cid:13)
(cid:12)
(cid:11)
(cid:0)(cid:1)(cid:4)
(cid:9)(cid:10)
(cid:8)
(cid:7)
(cid:0)(cid:1)(cid:3) (cid:17)(cid:18) (cid:27)(cid:1)(cid:26)
(cid:0)(cid:1)(cid:2)
(cid:0)(cid:1) (cid:0)(cid:1)(cid:1) (cid:0)()(cid:27) (cid:0)()’
(cid:20) (cid:23)(cid:26)(cid:25) ,(cid:23)(cid:20)(cid:25) Acknowledgments
FIG. 8: T dependence of (a) 1/T1T and (b) 1/T1 for H k c
in Tl-1245, which is slightly overdoped compared with Hg- TheauthorsaregratefultoKenjiIshidaforhishelpful
1245. The short component of T1(OP) shows the peak at discussions andShotaro Morimotofor experimentalsup-
∼ 45 K, indicative of AF ordering at the IP. T? is absent in ports in dc susceptibility measurement. This work was
Tl-1245. Figure (c) presents NMR spectra at 150 K and 20 supported by the COE Research (Grant No.10CE2004)
K. No change is observed in the linewidth of the OP signal, in Japan NEXT Grant. One of authors (H.K.) has been
indicating that the internal field at the OP is quite tiny, in supported by JSPS Research Fellowships for Young Sci-
contrast with Hg-1245. entists.
[1] B.Lake,H.M.Rønnow,N.B.Christensen,G.Aeppli,K. [10] Y. Tokunaga, H. Kotegawa, K. Ishida, G, -q, Zheng, Y.
Lefmann,D.F.McMorrow, P.Vorderwisch,P.Smeibidl, Kitaoka, K. Tokiwa, A. Iyo, and H. Ihara, J. of Low
N. Mangkorntong, T. Sasagawa, M. Nohara, H. Takagi, Temp. Phys. 117 473 (1999).
and T. E. Mason, Nature415, 299 (2002). [11] H. Kotegawa, Y. Tokunaga, K. Ishida, G. -q. Zheng,
[2] J. P. Hu,and S. -C. Zhang, cond-mat/0108273. Y. Kitaoka, K. Asayama, H. Kito, A. Iyo, H. Ihara, K.
[3] E.Demler,A.J.Berlinsky,C.Kallin, G.B.Arnold,and Tanaka,K.Tokiwa, andT.Watanabe,Phys.Rev.B64,
M. R.Beasley, Phys. Rev.Lett. 80, 2917 (1998). 064515 (2001).
[4] S.-C. Zhang, Science 275, 1089 (1997). [12] Y. Tokunaga, K. Ishida, Y. Kitaoka, K. Asayama, K.
[5] I.Bozovic,G.Logvenov,M.A.J.Verhoeven,P.Caputo, Tokiwa, A. Iyo, and H. Ihara, Phys. Rev. B61 (2000)
E.Goldobin,andT.H.Geballe,Nature422,873(2003). 9707.
[6] J.Akimoto, K.Tokiwa, A.Iyo,H.Ihara,K.Kawaguchi, [13] M.Mori,T.Tohyama,andS.Maekawa,Phys.Rev.B66
M. Sohma, H. Hayakawa, Y. Gotoh, and Y. Oosawa, (2002) 064502.
Physica C 281, 237 (1997). [14] S. Chakravarty, H. -Y. Kee, and K. V¨olker,
[7] A. Trokiner, L. Le Noc, J. Schneck, A. M. Pougnet, R. cond-mat/0309209.
Mellet, J.Primot, H.Savary,Y.M. Gao, andS.Aubry, [15] K. Tokiwa, A. Iyo, T. Tsukamoto, and H. Ihara, Czech.
Phys.Rev.B 44, 2426 (1991). Jour. of Phys. 46, 1491 (1996).
[8] B. W. Statt, and L. M. Song, Phys. Rev. B 48, 3536 [16] G. -q. Zheng, Y. Kitaoka, K. Asayama, K. Hamada, H.
(1993). Yamauchi, and S. Tanaka, Physica C 260, 197 (1996).
[9] Y. Piskunov, K. N. Mikhalev, Yu. I. Zhdanov, A. P. [17] F. Mila, and T. M. Rice, Physica C 157, 561 (1989).
Geraschenko, S. V. Verkhovskii, K. A. Okulova, E. Yu. [18] A. J. Millis, H. Monien, and D. Pines, Phys. Rev.B 42,
Medvedev,A.Yu.Yakubovskii,L.D.Shustov,P.V.Bel- 167 (1990).
lot, and A.Trokiner, Physica C 300, 225 (1998). [19] H.Monien,D.Pines,andM.TakigawaPhys.Rev.B43,
7
258 (1991). Vol.131(2003)637-641 J. of Low Temp. Phy.
[20] T. Imai, J. Phys.Soc. Jpn. 59, 2508 (1990). [26] M. -H. Julien, P. Carreta, M. Horvati´c, C. Berthier, Y.
[21] G.-q.Zheng,Y.Kitaoka,K.Asayama,Y.Kodama, and Berthier, P. S gransan, A. Carrington, and D. Colson,
Y.Yamada, Physica C 193, 154 (1992). Phys. Rev.Lett 76, 4238 (1996).
[22] H. Monien, P. Monthoux, and D. Pines, Phys. Rev. B [27] A. Narath, Phys.Rev. 167, 162 (1967).
43, 275 (1991). [28] K. Tokiwa, S.Itoh, H. Okumoto, S. Mikusu, T. Watan-
[23] K. Ishida, Y. Kitaoka, K. Asayama, K. Kadowaki, and abe,W.Higimoto,K.Nishiyama,A.Iyo,Y.Tanaka,and
T. Mochiku, J. Phys. Soc. Jpn. 63, 1104 (1994). H. Ihara, to be published.
[24] T. Tsuda, T. Shimizu, H. Yasuoka, K. Kishio, and K. [29] M. Wegar, Phys. Rev. 128, 1505 (1962).
Kitazawa, J. Phys. Soc. Jpn. 57, 2908 (1988). [30] M. Matsumura, andK.Asayama,J. Phys.Soc. Jpn.43,
[25] K. Tokiwa, S. Ito, H. Okumoto, S. Mikusu, A. Iyo, 1207 (1977).
Y. Tanaka, and T. Watanabe, J. of Low Temp. Phys.