Typeset with jpsj2.cls <ver.1.2.2b> Full Paper T Antiferromagnetic phase transition in four-layered high- superconductors c Ba Ca Cu O (F O ) with T = 55−102 K: 63Cu- and 19F-NMR studies 2 3 4 8 y 1−y 2 c Sunao Shimizu1∗, Hidekazu Mukuda1, Yoshio Kitaoka1, Hijiri Kito2, Yasuharu Kodama2, ParasharamM. Shirage2, and Akira Iyo2 9 0 0 1Graduate School of Engineering Science, Osaka University, Toyonaka, Osaka 560-8531, Japan 2 2National Institute of Advanced Industrial Science and Technology (AIST), Umezono, Tsukuba 305-8568, Japan n a We report on magnetic characteristics in four-layered high-T superconductors c J Ba2Ca3Cu4O8(FyO1−y)2 with apical fluorine through 63Cu- and 19F-NMR measurements. 4 Thesubstitutionofoxygenforfluorineattheapicalsiteincreasesthecarrierdensity(N )and h 2 T from55Kupto102K.TheNMRmeasurementsrevealthatantiferromagneticorder,which c can uniformly coexist with superconductivity, exists up to Nh ≃ 0.15, which is somewhat n] smaller than Nh ≃ 0.17 being the quantum critical point (QCP) for five-layered compounds. o The fact that the QCP for the four-layered compounds moves to a region of lower carrier c density than for five-layered ones ensures that the decrease in the number of CuO2 layers - makes an interlayer magnetic coupling weaker. r p u KEYWORDS: high-Tcsuperconductivity,copper-oxide,antiferromagnetism,NMR,phasediagram,apical- s fluorine . t a m 1. Introduction plingforanonsetofAFMorderisenhancedasanumber d- Remarkably high superconducting transition tem- of CuO2 layers increases. In order to establish how the n perature (T ) in the copper oxides hasbeen realizedina interlayermagneticcouplingaffectstheonsetofAFMor- c o multilayeredstructureofCuO2planes.Tcdependsonthe der, we deal with four-layered high-Tc superconductors [c number of CuO2 layers (n) in multilayered compounds Ba2Ca3Cu4O8(FyO1−y)2 with apical fluorine. We note withamaximumatn=3.1) Inparticular,thehighestT that all these compounds are in an underdoped state of c 1 was observed around 133 K in a Hg-based three-layered hole doping regime. 0v (n=3)copperoxideHgBa2Ca2Cu3Oy (Hg-1223).2)Cop- Ba2Ca3Cu4O8(FyO1−y)2 comprises a stack of four 1 peroxideswithmorethanthreelayerscompriseinequiva- CuO2 layers as illustrated in Fig.1(a). It is known as a 8 lenttypesofCuO2layers,anouterCuO2plane(OP)ina newfamilyofmultilayeredcopperoxideswithapicalflu- 3 five-foldpyramidalcoordination,andaninnerplane(IP) orine.10–12) Substitutionofoxygenforapicalfluorine,i.e. 1. inafour-foldsquarecoordination(seeFig.1asexample). adecreaseinnominalfluorinecontent(y)resultsindop- 0 Site-selective 63Cu-NMR studies have revealed that the ing holes into CuO2 layers, increasing Tc from 55 K at 9 local carrierdensity (N ) for the IP is smaller than that y = 1.0 to 102 K at y = 0.6 as indicated in Fig.1(b).11) h 0 for the OP. These results in turn revealed an intimate This system provides an opportunity to investigate the v: relationship between antiferromagntism (AFM) and su- characteristicsofCuO2layersoverawiderangeofcarrier i perconductivity(SC)inherenttoCuO layersintowhich density,especiallyenablingustofocusontheinterplayof X 2 mobile hole carriers were homogeneously doped.3–6) SCandAFMinanunderdopedregion.13–21)Wereported ar The recent systematic Cu-NMR studies on five- in the literature22) that self-doping occurred at y = 1.0 layered (n = 5) compounds have unraveled that AFM to realize superconductivity. For a nominal content at order, which can uniformly coexist with SC, is robust y = 1.0, if the apical site of OP were fully occupied by up to N ≃ 0.17,a quantum critical point (QCP) where F−1,theformalCuvalencewouldbejust+2andhencea h the AFM order collapses.7) This result significantly dif- Mott insulator; however, it exhibits SC. The occurrence fers from the well-established results for mono-layered ofthe SCaty =1.0wasarguedintermsofaself-doping (n=1)La2−xSrxCuO4 (LSCO)8) andbi-layered(n=2) model wherein charge carriers were transferred between YBa Cu O (YBCO),9) in which the AFM order col- IPandOP.22–24) Thesediscussions,however,werebased 2 3 6+x lapses completely by doping extremely small amount of on a simple assumption that all the apical sites were holes of N ∼ 0.02 and 0.055, respectively. These re- occupied by F−1. Since then, a bi-layered apical-F com- h sults strongly suggest that the QCP of n ≤ 4 moves to pound Ba2CaCu2O4F2 has also been synthesized; it ex- a region of lower carrier density than that of n = 5. hibits SC withTc =73K.25) The self-doping mechanism Therefore, it is likely that the interlayer magnetic cou- cannot apply to this compound because it has no IPs. We, therefore, have conducted NMR studies on the bi- ∗E-mail:[email protected] 1 J. Phys.Soc. Jpn. Full Paper 2. Experimental Polycrystalline powder samples of all multilay- ered systems used in this study were prepared by a high-pressure synthesis technique, as described else- where.10–12) Powder X-ray diffraction measurements in- dicate that these compounds almost entirely comprise a single phase, and that the a-axis length continually changes with the nominal fluorine content y.11) The T c was uniquely determined by a sharp onset of diamag- netism using a dc SQUID magnetometer as summarized inTableI.ForNMRmeasurements,thepowdersamples were aligned along the c-axis at an external field (H) of ∼ 16 T, and fixed using stycast 1266 epoxy. The NMR experiments were performed by the conventional spin- echo method in a temperature (T) range of 1.5 − 300 K with H perpendicular or parallel to the c-axis. 3. Results and Discussions 3.1 Knight shift and local carrier density Figure 2 indicates typical 63Cu-NMR spectra of the central transition (1/2 ⇔ −1/2) for (a) y = 0.6, (b) y = 0.7, (c) y = 0.8, and (d) y = 1.0. The field-swept NMR spectra were measured at H perpendicular to the Fig. 1. (color online) (a) Crystal structure of c-axis.TheNMRspectralwidthsforthesamplesatroom Ba2Ca3Cu4O8(FyO1−y)2. This four-layered system in- cludes two crystallographically different CuO2 planes, namely, temperature areasnarrowasthat ofHg-1245,5,7) assur- IP and OP. (b) Tc vs. fluorine content y.11) It shows that the ingthehighqualityofthesamples.TheCu-NMRspectra substitutionofoxygenforfluorineresultsindopingholecarriers forOPandIPareseparatelydetectedaty =0.6,whereas intoCuO2 layers theIP’sspectrafory =0.7and0.8disappearatlowtem- peratures due to the development of AFM correlations upon cooling as well as the case for the five-layered Hg- layeredsystemtoconfirmthattheself-dopingmechanism orTl-basedcompounds.5,7) Surprisingly,theNMRspec- didnotoccur,andthatallmultilayeredcompoundswith trum of OP at y = 1.0 also disappears at low tempera- the apical fluorine were doped by hole carriers irrespec- tures, as shownin Fig.2(d), suggestingan onset of AFM tive of y. For y =1.0, it was anticipated that a possible order at the OP. The systematic variation of the NMR replacementof O−2 for F−1 and/orexcessoxygenin the spectraindicatesthatBa2Ca3Cu4O8(FyO1−y)2 becomes BaF layers resulted in doping hole carriers into CuO 2 underdoped as the nominal fluorine content y increases, layers.26) which ensures that the substitution of oxygen for the In this paper, we report systematic 63Cu- and 19F- apical fluorine increases N . h NMR studies on Ba2Ca3Cu4O8(FyO1−y)2 with y = 0.6, 0.7, 0.8 and 1.0 as each nominal content. Measurements of 63Cu Knight shift (63K) have revealed that hole car- rier density (Nh) increases progressivelywith decreasing Table I. Lists of Tc and Nh at the OP and IP of y. The substitution of oxygen for fluorine at the apical Ba2Ca3Cu4O8(FyO1−y)2. Nh is estimated from the Ks(T) at site increases N and T from 55 K up to 102 K. The roomtemperature(seetext).Nhisanaverageofthecarrierden- h c sity, defined as (Nh(OP) + Nh(IP))/2. Note that y is nominal measurements of 63Cu-NMR spectra and nuclear-spin- fluorinecontents. lattice-relaxation rate of 19F-NMR (19(1/T1)) unravel y=0.6 y=0.7 y=0.8 y=1.0 that an AFM order, which can uniformly coexist with Tc 102K 91K 77K 55K SC, exists up to N ≃ 0.15 being a QCP for the four- Nh [OP] +0.207 +0.189 +0.167 +0.148 layered compounds.hFrom the fact that the QCP of the NhN[hIP] ++00..118665 ++00..117500 ++00..115464 ∗∗((++00..113420)) four-layered compounds moves to a region of lower car- rierdensitythanthatofthefive-layeredones,N ≃0.17, h it is ensured that the decrease in the number of CuO ∗)Nh=0.140aty=1.0isestimatedatNh(OP)=0.148onalinear 2 layers makes an interlayer magnetic coupling weaker. line in the plot of Nh(OP) vs Nh (see Fig.3). A linear extrap- olation in the plot in Nh(IP) vs. Nh gives a tentative value of Nh(IP)=0.132atNh=0.140,sinceitcannotbeestimateddirectly fromtheKs(T)atIPaty=1(seeanopencircleinFig.3andthe text). 2/?? J. Phys.Soc. Jpn. Full Paper intheplotofN (IP)versusN givesatentativevalueof h h N (IP)=0.132 at N = 0.140. As summarized in Table h h I, the increase of N (OP) and N (IP) with increasing a h h nominal oxygen content at the apical site increases T c from 55 K to 102 K. Fig. 3. (coloronline)PlotsofNh(OP)andNh(IP)againstaver- aged carrier density Nh, defined as (Nh(OP)+Nh(IP))/2. Here Nh(OP)andNh(IP)aredeterminedfromtheKnightshiftmea- surement (see text). Nh = 0.140 at y=1.0 is estimated from Fig. 2. (coloronline)63Cu-NMRspectraofthecentraltransition Nh(OP) = 0.148 on a linear line in the plot of Nh(OP) versus (1/2 ⇔−1/2) for (a) y=0.6, (b) y=0.7, (c) y=0.8, and (d) Nh.AlinearextrapolationintheplotsinNh(IP)versusNhgives y = 1.0. The temperature dependence of the Knight shift with atentative valueofNh(IP)=0.132atNh=0.140. H perpendicular to the c-axis for (e) y = 0.6, (f) y = 0.7, (g) y=0.8,and(h)y=1.0. 3.2 Zero-field NMR evidence of AFM order In general, the Knight shift (K) comprises the T- WedealwiththeAFM ordertakingplaceinthe ud- dependent spin part (Ks(T)) and the T-independent or- erdopedCuO2 layers.Theobservationofzero-fieldNMR bitalpart(K );K =K (T)+K .Here,K wasde- (ZFNMR) spectra enables us to assure an onset of an orb s orb orb terminedas0.23(±0.02)%,assumingK ≈0 inthe T = AFM order,since magnetically ordered moments induce s 0 limit. The T dependences ofKs(T) with H perpendic- internal magnetic field Hint at nuclear sites. Generally, ular to the c-axis are displayed in Figs.2(e),(f),(g), and the Hamiltonian for Cu nuclear spin with I =3/2 is de- (h)fory =0.6,0.7,0.8,and1.0,respectively.TheK (T) scribed by the Zeeman interaction due to magnetic field s decreases upon cooling down to Tc for all samples in as- H (HZ) and the nuclear-quadrupoleinteraction(HQ) as sociationwithanopeningofpseudogap,27,28)whereasits follows: steep decrease below T evidences the reduction of spin c susceptibility due to the formation of spin-singlet pair- e2qQ ing. We note that the empirical relation between Ks(T) H=HZ+HQ =−γN~I·H+4I(2I−1)(3Iz2′−I(I+1)), at room temperature and the Nh in a CuO2 plane29,30) (1) allowsustoevaluateNhsatOPandIPforthefoursam- whereγN istheCunucleargyromagneticratio,eQisthe ples, which are summarized in Table I along with the Tc nuclear quadrupole moment, and eq is the electric field andtheaveragedcarrierdensityNh,definedas(Nh(OP) gradient (EFG) at the Cu nuclear site. Here, in the HQ, + Nh(IP))/2. Figure 3 indicates the plot of Nhs at IP an asymmetric parameter (η) is zero in the tetragonal and OP against Nh. The Nh(IP) at y = 1.0, however, symmetry. Note that the nuclear quadrupole resonance was not directly estimated from the Ks(T), because the (NQR) frequency νQ =3e2qQ/2hI(2I−1). The nuclear Cu-NMR spectrum was not detected at room tempera- Hamiltonian given by eq.(1) is described with H in- int ture. Instead, Nh = 0.140 at y = 1.0 is estimated from stead of H for zero-field experiments. Nh(OP)=0.148 on a linear line in the plot of Nh(OP) Figure 4(a) indicates the Cu-NQR spectrum at y = versus Nh in Fig.3. Furthermore, a linear extrapolation 0.6.Respective63νQsareevaluatedas9.7and15MHzat 3/?? J. Phys.Soc. Jpn. Full Paper smallerthanthosefortheintrinsicphase,whichsuggests thattheseNQRspectraarisefromsomeimpurityphases containingcoppersuchasstartingmaterialspreparedfor the sample synthesis, and intermediate products in the high-pressure synthesis.10,11,35) Figure 4(b) indicates the Cu-NQR/ZFNMR spectra at y =0.7. The NQR spectrum for OP is observed with almostthe same63ν asthatofthe OPaty =0.6.Note Q that the respective NQR and ZFNMR spectra at IP(i) and IP(ii) arise from IP. The NQR spectrum at IP(i) is observed at 9.1 MHz that is close to 63ν = 9.7 MHz Q of the IP at y = 0.6, whereas the ZFNMR spectrum at IP(ii) is observed at ∼ 18 MHz. Assuming 63ν = 9.1 Q MHz, H ∼1.5 T is estimated for the IP(ii). The H int int atCuO planeisgenerallygivenbyH =|A |M = 2 int hf AFM |A−4B|M , where A and B are the on-site hyper- AFM fine field and the supertransferred hyperfine field from the four nearest neighboring Cu-AFM moments, respec- tively, and M is the AFM moment.36) Here A∼ 3.7 AFM T/µ , B(OP) ∼ 7.4 T/µ , and B(IP) ∼ 6.1 T/µ are B B B assumed to be the same as those for Hg-1245.5) Using these values,a uniform AFM momentat the IP(ii)is es- timated at M (IP) ∼ 0.08 µ for an AFM phase at AFM B y = 0.7. The fact that IP(i) and IP(ii) originate from a paramagnetic phase and an AFM phase, respectively, suggests that the phase separation takes place because oftheclosenesstotheQCPatwhichtheAFMcollapses. Thepresenceofthephaseseparationprobablyimplythat the AFM critical point could be close to N ∼ 0.15. Fig. 4. (color online) Cu-NQR/ZFNMR spectra at H = 0 and h Figure 4(c) indicates the Cu-NQR/ZFNMR spectra T =1.5 K. (a) The respective Cu-NQR spectra at IP and OP for y =0.6with 63νQs = 9.7 and 15 MHz.Two sharppeaks at aty =0.8.Aspectrumobservedaround14.4MHzarises 20−25 MHz arise from unknown impurity phases. (b) The Cu- fromOPsince its peakfrequencyis almostthe samefre- NQR/ZFNMR spectra at y = 0.7. The NQR spectrum for the quencyas63ν =15MHzfortheOPaty =0.6.Accord- OP is observed with almost the same 63νQ as that of the OP ingly, anotherQspectrum around 28 MHz is assigned to aty=0.6.TheNQRspectrum atIP(i)isobservedat 9.1MHz that is close to 63νQ = 9.7 MHz of the IP at y = 0.6, whereas arisefromIP.Usingabove-mentionedparameters,Hint ∼ theZFNMRspectrumatIP(ii)probestheinternalfield Hint = 2.4TandMAFM(IP)∼0.12µB areestimatedforthe IP 2.4Tdue totheAFMorder.Thefact thatthe IP(i)andIP(ii) at y =0.8. originatefromaparamagneticphaseandanAFMphase,respec- Figure 4(d) indicates the Cu-ZFNMR spectra ob- tivelysuggests that thephaseseparation takes placebecause of served around 30 and 45 MHz at H = 0 for y = 1.0. the closeness to the QCP at which the AFM collapses. (c) The When noting that ZFNMR spectra are absent around Cu-NQR/ZFNMR spectra at y = 0.8. The NQR spectrum for the OPisobserved atalmostthesamefrequencyasthat ofthe 63νQ(IP) = 8 ∼ 10 MHz and 63νQ(OP) = 14 ∼ 16 OP at y = 0.6, revealing that no spontaneous moment is in- MHz, the observation of the NMR spectra around 30 duced at low temperatures. The spectrum observed around 28 MHz and 45 MHz demonstrates that H s are present MHzarisesfromthe IP withHint ∼ 2.4T andhence theAFM at the respective IP and OP with H ∼int3.8 T and 2.7 momentMAFM∼0.12µB.(d)TheCu-NMRspectraaty=1.0 int with Hint ∼ 2.7 T and 3.8 T for the OP and IP, respectively. T. Since Nh(OP) > Nh(IP) due to the chargeimbalance Here, MAFM = 0.11 µB and 0.18 µB are estimated for the OP betweenOPandIPandhenceMAFM(OP)<MAFM(IP), andIP,respectively. M (OP) ∼ 0.11 and M (IP)∼0.18 µ are evalu- AFM AFM B atedattheOPandIP,respectively,usingtherelationof H = |A |M = |A−4B|M . Notably, the OP, int hf AFM AFM the IP and OP, which are comparable to ∼ 8−10 MHz which is mainly responsible for the SC with T = 55 K, c and ∼ 16 MHz for five-layered systems.3–5) Here, the manifeststheAFMorder,leadingustoaconclusionthat sharp 63Cu- and 65Cu-NQR spectral widths at 63ν = the uniform mixing of AFM with M = 0.11 µ and Q AFM B 22.5 and 65ν = 20.8 MHz are as narrow as about 400 SC atT =55Koccursinthe OPaswellasinthethree Q c kHz. These NQR spectra qualitatively differ from those IPs of the five-layeredsystems.7) reported for various copper oxides.31–34) Integrated in- tensitiesoftheseNQRspectraareanorderofmagnitude 4/?? J. Phys.Soc. Jpn. Full Paper in Fig.5. In the present case, the relaxation processes in 1/T compose of quasiparticle contributions probing the 1 onsetofSCandthemagneticoneprobingmagneticfluc- tuations. However, we consider that the former is negli- gible in the case of 1/T at the apical site; in fact, it has 1 been reported 1/T at the apical oxygen by 17O-NMR 1 did not change drastically at T 37) because of the very c smallhyperfine-couplingconstantwiththequasiparticles in CuO layers. In this context, 19(1/T ) is expected to 2 1 be dominated by magnetic fluctuations. Respective figures 5(b) and (c) show the T depen- dences of 19(1/T )s at y = 0.7 and 0.8, exhibiting the 1 peaks at∼ 30K and50K.This resultensuresthe AFM order at T = 30 and 50 K for the IP(ii) at y =0.7 and N the IP at y = 0.8 with the spontaneous AFM moment of M (IP(ii)) ∼ 0.08 µ and M (IP) ∼ 0.12 µ , AFM B AFM B respectively. Note that the absence of a peak in 1/T at 1 y = 0.6 evidences that this compound is in a param- agnetic state down to 4.2 K. As shown in Fig.5(d) for y = 1.0, there are a distinct peak in 1/T at T ∼ 80 1 N ∗ K and a significant one at T ∼ 30 K. The AFM order N inherent to the OP responsible for SC is presumably de- ∗ veloped below T ∼ 30 K, exhibiting the spontaneous N AFM momentofM (OP)∼0.11µ atlowtempera- AFM B tures.This suggeststhatthe SC uniformly coexists with the AFM orderinasingle CuO planewith N ∼ 0.148. 2 h Since N (IP) < N (OP) and M (IP) > M (OP), h h AFM AFM Fig. 5. (coloronline)Thetemperaturedependencesof19(1/T1)s the T at the IP becomes larger than at the OP. It is N for (a) y = 0.6, (b) y = 0.7, (c) y = 0.8, and (d) y = 1.0 ∗ noteworthythat T ∼ 30K for the OPat y=1.0is com- at an NMR frequency 174.2 MHz and H parallel to the c-axis. N parabletotheT ∼30KfortheIP(ii)aty =0.7because For (a) y = 0.6, a reason that 1/T1 increases upon cooling is N associatedwiththedevelopmentofmagneticcorrelationsatlow both layers possess almost the same Nh. temperatures. For (b) y =0.7 and (c) y =0.8, 1/T1 exhibits a peakatTN∼30KandTN∼50K,respectively.(d)Therearea 3.4 Phase diagram of AFM and SC distinctpeakin1/T1atTN∼80KandasignificantoneatTN∗ ∼ Figure6revealsa phasediagramofAFMandSC as 30 K. The AFM order inherent to the OP responsible for SC ∗ is presumably developed below TN∗, exhibiting the spontaneous afunctionofNhwhereTcandTN(TN)areplottedagainst AFMmomentofMAFM(OP)∼0.11µB atlowtemperatures. Nh fortheOPsandIPsofthefour-layeredsuperconduc- tors Ba2Ca3Cu4O8(FyO1−y) at y=0.6, 0.7, 0.8, and 1.0. We remark that the uniform mixing of AFM (T = 30 N 3.3 19F-NMR probe of N´eel temperature K) and SC (Tc = 55 K) was observed for the OP at y N´eel temperatures (T s) at y = 0.7, 0.8 and 1.0 = 1.0, which strongly suggests that it is a general prop- N are determined by the 19F-T measurement, which sen- erty inherent to a single CuO2 plane in the underdoped 1 regime for hole-doping.It has been reportedin the liter- sitively probes critical magnetic fluctuations developing atures4,7) thatabulk T inmultilayeredcompoundswas at OP and IP as the system approaches an AFM order. c determined by the T of OP and that the T of IP was Generally, 1/T is described as follow; c c 1 significantlylowerthanabulkT due to the lowerN at c h 1 2γ2k T Im[χ(q,ω )] T1 = (γNe~B)2 Xq |Aq|2 ω0 0 , (2) IlaPy.eIrnedthseysptheamseisdioabgtraaimneidnaFtigN.6h, t≃he0Q.1C5Psminaltlheer ftohuarn- N ≃ 0.17 for the five-layered system,7) suggesting that h where A is the wave-vector (q)-dependent hyperfine- q theinterlayermagneticcouplingofthefour-layeredcom- coupling constant, χ(q,ω) is the dynamical spin suscep- poundissmallerthanthatofthefive-layeredcompound. tibility,andω istheNMRfrequency.TheT dependence 0 ThephasediagramsofAFMandSCinmultilayeredsys- of1/T showsapeakatT becausethelow-energyspec- 1 N tems are remarkably different from the well-established tral weight in χ(q = Q,ω) is strongly enhanced around ones for LSCO (n = 1)8) and YBCO (n = 2),9) where ω ∼0 in associationwith a divergence of magnetic cor- 0 the AFM order totally collapses by doping very small relation length at T ∼ T . Here Q is the AFM wave N amountofholeswithN ∼0.02andN ∼0.055,respec- vector(π/a,π/a).19(1/T )for allsamplesare presented h h 1 tively. The reason that the AFM phase exists up to N h 5/?? J. Phys.Soc. Jpn. Full Paper ≃0.15and0.17inthefour-andfive-layeredcompounds, ies have highlighted the intimate evolution of AFM and respectively,isbecausetheinterlayermagneticcouplings SC in the phase diagraminherent to the homogeneously arestrongerthaninLSCOorYBCOduetotheexistence dopedCuO plane,whichdependsontheinterlayermag- 2 of the homogeneously underdoped IPs. netic coupling significantly. Acknowledgments TheauthorsaregratefultoM.Mori,T.Tohyama,H. Eisaki and M. Ogata for their helpful discussions. This workwassupportedbyaGrant-in-AidforSpeciallyPro- moted Research (20001004)and in part by Global COE Program (Core Research and Engineering of Advanced MaterialsScience),fromtheMinistryofEducation,Cul- ture, Sports, Science and Technology (MEXT), Japan. One of the authors (S.S.) is financially supported as a JSPS (the Japan Society for the Promotion of Science) Research Fellow. 1) B.A.Scott,E.Y.Suard,C.C.Tsuei,D.B.Mitzi,T.R.McGuire, B.-H.Chen,andD.Walker:PhysicaC230(1994) 239. 2) A.Schilling,M.Cantoni,J.D.Guo,andH.R.Ott:Nature(Lon- don)363(1993)56. 3) Y.Tokunaga,K.Ishida,Y.Kitaoka,K.Asayama,K.Tokiwa, A.Iyo, andH.Ihara:Phys.Rev.B61(2000) 9707. 4) H.Kotegawa,Y.Tokunaga,K.Ishida,G.-q.Zheng,Y.Kitaoka, H.Kito,A.Iyo,K.Tokiwa,T.Watanabe,andH.Ihara:Phys. Rev.B64(2001) 064515. Fig. 6. (color online) A phase diagram of AFM and SC as a 5) H.Kotegawa,Y.Tokunaga,Y.Araki,G.-q.Zheng,Y.Kitaoka, functionofholecarrierdensityNh.Tc(indicatedbyopencircle) K.Tokiwa,K.Ito,T.Watanabe,andA.Iyo:Phys.Rev.B69 andTN(closedcircle)areplottedagainstNhfortheOPsandIPs (2004) 014501. ofBa2Ca3Cu4O8(FyO1−y)aty =0.6,0.7,0.8and1.0.Nh was 6) H. Mukuda, M. Abe, Y. Araki, Y. Kitaoka, K. Tokiwa, T. determined from the Knight shift measurement (see the text). The Tc ∼ 70 K for the IP at y = 0.6, that is shown by black Watanabe, A. Iyo, H. Kito, and Y. Tanaka: Phys. Rev. Lett cross (+) was determined from a small peak at T ∼ 70 K in 96(2006) 087001. 7) H.Mukuda,Y.Yamaguchi,S.Shimizu,Y.Kitaoka,P.Shirage, the T derivative of the Knight shift (dK/dT) as well as in the literatures.3,7) NotethattheuniformmixingofAFM(TN =30 andA.Iyo:J.Phys.Soc.Jpn.77(2008) 124706. 8) B.Keimer,N.Belk,R.J.Birgeneau,A.Cassanho,C.Y.Chen, K)and SC(Tc =55 K)takes place atthe OPat y = 1.0.This M.Greven,andM.A.Kastner:Phys.Rev.B46(1992)14034. resultstronglysuggeststhatitisageneralpropertyinherenttoa singleCuO2planeintheunderdopedregimeforhole-doping.6,7) 9) SP.hSyasn.nRae,vG..LAetltlo.d9i3,G(2.0C0o4n)c2a0s,70A0.1D..Hillier,andR.DeRenzi: 10) A.Iyo,Y.Tanaka,M.Tokumoto,andH.Ihara:PhysicaC366 (2001) 50. 11) A.Iyo,M.Hirai,K.Tokiwa,T.Watanabe,Y.Tanaka:Physica 4. Summary C392-396(2003)140. The extensive Cu-NMR/NQR and F-NMR mea- 12) A. Iyo, Y.Tanaka, H.Kito, Y.Kodama, P.M.Shirage, D.D. Shivagen,H.Matsuhata,K.Tokiwa,andT.Watanabe:J.Phys. surements on the four-layered high-T superconductors c Soc.Jpn.76(2007) 094711. Ba2Ca3Cu4O8(FyO1−y) have unraveled the systematic 13) M.Inaba,H.Matsukawa,M.Saitoh,andH.Fukuyama:Phys- evolution of AFM and SC as the function of hole carrier icaC257(1996) 299. densityN ;T andT arecontrolledby thesubstitution 14) S.C.Zhang: Science275(1997) 1089. h c N of oxygen for fluorine at the apical site. It is demon- 15) A.HimedaandM.Ogata:Phys.Rev.B60(1999)R9935. 16) Y.S.Lee,R.J.Birgeneau, M.A.Kastner,Y.Endoh,S.Waki- strated that the AFM order, which can uniformly coex- moto, K. Yamada, R.W. Erwin, S.-H. Lee, and G. Shirane: ist with SC, exists up to N ≃ 0.15,reinforcing that the h Phys.Rev.B60(1999)3643. uniformmixingofAFMandSCisageneralpropertyin- 17) Y.Sidis,C.Ulrich,P.Bourges,C.Bernhard,C.Niedermayer, herentto a single CuO plane in the underdoped regime L.P.Regnault,N.H.Andersen,andB.Keimer:Phys.Rev.Lett. 2 for hole-doping. N ≃ 0.15 at QCP for the four-layered 86(2001) 4100. h 18) R.I.Miller,R.F.Kiefl,J.H.Brewer,J.E.Sonier,J.Chakhalian, compounds is somewhat smaller than N ≃ 0.17 for the h S.Dunsiger,G.D.Morris,A.N.Price,D.A.Bonn,W.H.Hardy, five-layered compounds. The fact that the QCP for the andR.Liang:Phys.RevLett.88(2002) 137002. four-layered compounds moves to a region of lower car- 19) B. Lake, H.M. Ronnow, N.B. Christensen, G. Aeppli, K. rier density than for the five-layeredcompounds ensures Lefmann, D.F. McMorrow, P.Vorderwisch, P. Smeibidl, N. thatthedecreaseinthenumberofCuO layersmakesan Mangkorntong,T.Sasagawa,M.Nohara,H.Takagi,andT.E. 2 Mason:Nature(London) 415(2002) 299. interlayer magnetic coupling weaker. The present stud- 6/?? J. Phys.Soc. Jpn. Full Paper 20) T.K. Lee, C.-M. Ho, and N. Nagaosa: Phys. Rev. Lett. 90 29) G.-q.Zheng,Y.Kitaoka,K.Ishida,andK.Asayama:J.Phys. (2003) 067001. Soc.Jpn.64(1995) 2524. 21) E. Demler, W. Hanke, and S.C. Zhang: Rev. Mod. Phys. 76 30) Y.Tokunaga,H.Kotegawa,K.Ishida.G.-q.Zheng,Y.Kitaoka, (2004) 909. K.Tokiwa,A.Iyo,andH.Ihara:J.LowTemp.Phys.117(1999) 22) S. Shimizu, H. Mukuda, Y. Kitaoka, A. Iyo, Y. Tanaka, Y. 473. Kodama, K.Tokiwa, and T.Watanabe: Phys. Rev. Lett. 98 31) K.Ishida,Y.Kitaoka,N.Ogata,T.Kamino,K.Asayama,and (2007) 257002. J.R.Cooper:J.Phys.Soc.Jpn.62(1993) 2803. 23) Y. Chen, A. Iyo, W. Yang, X. Zhou, D. Lu, H. Eisaki, T.P. 32) S.Ohsugi,Y.Kitaoka,K.Ishida,G.-q.Zheng,andK.Asayama: Devereaux, Z. Hussain, and Z.-X.Shen: Phys. Rev. Lett. 97 J.Phys.Soc.Jpn.63(1994)700. (2006) 236401. 33) S. Ohsugi, T. Tsuchiya, T. Koyama, and K. Fueki: J. Low 24) W.Xie,O.Jepsen,O.K.Andersen,Y.Chen,andZ.-X.Shen: Temp.Phys.105(1996) 314. Phys.Rev.Lett. 98(2007)047001. 34) G.-q. Zheng, Y. Kitaoka, K. Asayama, K. Hamada, H. Ya- 25) A.Iyoet al.:unpublished. mauchi,S.Tanaka: PhysicaC260(1996) 197. 26) S.Shimizuet al.:unpublished. 35) N. Kobayashi, Z. Hiroi, and M. Takano: J. Sol. State Chem. 27) H. Yasuoka, T. Imai, and T. Shimizu: in Strong Correlation 132(1997) 274. and Superconductivity, ed. H. Fukuyama, S. Maekawa, and 36) F.MilaandT.M.Rice,Phys.Rev.B.40(1989) 11382. A.P.Malozemoff(Springer,Berin,1989)Vol.89,p.254. 37) P.C.Hammel,M.Takigawa,R.H.Heffner,Z.Fisk,andC.Ott, 28) R.E. Walstedt: The NMR Probe of High-Tc Materials Phys.Rev.Lett.63(1989)1992. (Springer,Berin,2008). 7/??