Cooper L.N., Feldman D. (Eds.) BCS: 50 Years

Подождите немного. Документ загружается.

September 14, 2010 9:46 World Scientiﬁc Review Volume - 9.75in x 6.5in ch16

The Evolution of HTS: T

-Experiment Perspectives 401

information in hand, we quickly found through partial magnetic substi-

tution of Y that Y in YBCO is electronically isolated from the super-

conducting component of the compound and serves only as a stabilizer

to hold the crystal structure together. The information led us to the

subsequent discovery of the whole series of RBa

(RBCO, R123 or

Cu1212) with a T

varying between 93 and 100 K where R = Y and

all rare-earth elements except Ce and Pr.

Two related structures were

later found by others, i.e. YBa

(Y124) and Y

(Y247),

with a T

around 80 K. The layer-stacking sequence for YBa

(BaO)(CuO)(CuO)(BaO)(CuO

)(Y)(CuO

). Y

can be consid-

ered as an intergrowth of Y123 and Y124.

The discovery broke the liquid nitrogen temperature barrier of 77 K for

superconductivity. It brings superconductivity technology a giant step closer

to applications by using liquid-nitrogen as the low-cost, plentiful, more eﬃ-

cient and easier-to-handle coolant and at the same time poses serious chal-

lenges to physicists concerning the origin of HTS. The excitement at the

time was amply demonstrated in the Special Panel Discussion on Novel Ma-

terials and High Temperature Superconductivity initiated by me and hastily

organized by the American Physical Society (APS) on March 18, 1987, in

the Sutton Complex at the New York Hilton and the July 1987 Federal Con-

ference at Washington, D.C., attended by President Reagan and his cabinet

members. The APS Panel Discussion started at 7:30 p.m. with short pre-

sentations by ﬁve panelists: Alex Mueller of IBM Zurich Lab; Shoji Tanaka

of the University of Tokyo; Paul Chu of the University of Houston; Bertrum

Batlogg of Bell Labs; and Zhongxian Zhao of the Physics Institute of the

Chinese Academy of Sciences, followed by short contributions and discus-

sions that lasted until the wee hours of the next morning. The 1,200-seat

meeting room was packed with more than 2,000 people. Many more who

could not get into the room watched on TV screens outside the room to

witness this exciting event, called the Woodstock of Physics by the late Bell

Labs physicist Mike Schluter, referring to the legendary 1969 Woodstock

Music and Art Festival in upstate New York. In spite of later discoveries

of many other cuprate HTS families, some of which have higher T

, YBCO

remains to be the most desirable material for HTS science and technology

due to its superior sample quality, current carrying capacity in the presence

of high magnetic ﬁelds and physical robustness in thin-ﬁlm form. A YBCO

puck together with a note was selected as an entry for the White House’s

National Millennium Time Capsule in 2000 (Fig. 6). The capsule was

created in the spirit of “honor the past – imagine the future” to contain

September 14, 2010 9:46 World Scientiﬁc Review Volume - 9.75in x 6.5in ch16

402 C. W. Chu

Fig. 6. Closing Ceremony for the White House Millennium Time Capsule “Honor

the Past — Imagine the Future” that included a YBCO puck, held at the National

Archives in Washington, D.C., December 6, 2000.

discoveries and achievements in all areas by Americans over the last 100 years

considered to be important to the present and/or the future. It will be

opened in 2100 to communicate to the future generations about the accom-

plishments and visions realized in the U.S. in the 20th Century.

3.1.3. The ﬁrst cuprate HTS family without rare-earth with a T

up to

110 K: Bi

n−1

2n+4

, where n = 1, 2, 3, . . . [BSCCO

or Bi22(n-1)n]

The discovery

of the Bi

n−1

2n+4

[BSCCO or Bi22(n-1)n] with

a T

up to 110 K in January 1988 by Hirosh Maeda et al. of the National

Research Institute for Metal suggested that HTS had a broader material-

base than originally thought and that more HTS materials with a higher T

might be discovered. The graphitic-like behavior of BSCCO due to its weak

inter-layer coupling makes the compound a good material for spectroscopy

studies and for the ﬁrst-generation HTS-wires, in spite of the weak ﬂux

pinning associated with the compound family.

Euphoria permeated the ﬁeld following the announcement of supercon-

ductivity above 90 K in YBCO and RBCO, and the sky seemed to be the

only limit to T

. As 1987 was drawing to an end, an impatient indus-

trial physicist told the Wall Street Journal that the accumulated man-hours

devoted to cuprate-HTSs in 1987 exceeded all previous eﬀort expended on

LTSs in the preceding 75 years since its discovery and that any cuprate

with a T

above 90s K should have been found. He went on to propose

September 14, 2010 9:46 World Scientiﬁc Review Volume - 9.75in x 6.5in ch16

The Evolution of HTS: T

-Experiment Perspectives 403

Fig. 7. ρ(T ) of Bi-Si-Ca-Cu-O (Bi1223) shows a T

up to 105 K in 1988 by Maeda

et al. [H. Maeda et al., Jpn. J. Appl. Phys. 27, L209 (1988)].

searching outside cuprates for superconductors with higher T

. The fal-

lacy of prophecy based on statistics immediately faced the truth in the ﬁrst

week of January 1988 when Maeda et al. of the National Research Institute

for Metal reported superconductivity above 100 K in Bi-Sr-Cu-O that later

appeared in the February 20 issue of Japanese Journal of Applied Physics in

an article entitled “A New High-T

Oxide Superconductor without a Rare

Earth Element.”

In their attempt to broaden the material base for HTSs, Maeda et

al. examined elements in the VB group of the periodic table, such as

Bi, that are trivalent and of similar ionic radii to those of the rare-

earth elements. They discovered superconductivity in samples of Bi-Sr-

Ca-Cu-O above 105 K as shown in Fig. 7. The resistivity data suggested

that the sample might consist of multiple phases. Indeed, shortly af-

terward, three members of the Bi22(n-1)n family with n = 1, 2 and 3

were isolated and the structures determined by several groups, including

the Carnegie Institute Geophysical Lab and our own. The layer stacking

sequence is, e.g. (SrO)(BiO)(BiO)(SrO)(CuO

)(Ca)(CuO

) for

with n = 3. Layer structural modulation appears in

the (BiO) double layers for all n. The maximum T

s for members of the

homologous series are ∼22, ∼80 and ∼110 K for increasing n = 1, 2 and 3,

respectively. However, for n > 3, T

starts to drop and this drop is attributed

September 14, 2010 9:46 World Scientiﬁc Review Volume - 9.75in x 6.5in ch16

404 C. W. Chu

to the combined inﬂuence of electrostatic shielding and proximity eﬀects of

the CuO

-layers. Aided by the experience on R123, it took less than two

days for Hazen et al. of Carnegie Geophysical Lab and Veblen et al. of Johns

Hopkins to crack the structure of BSCCO after receiving the samples from

us, in contrast to the more than 10 days required for Hazen et al. to do the

same for Y123 in 1987. From their structures, it was immediately evident

that the weak van der Waal force between the (BiO)-double-layers in BSCCO

is responsible for the graphitic-behavior of the compound, making cleaving

the sample in vacuum easy for spectroscopy studies and mechanical rolling

eﬀective for alignment of the (CuO

)-layers for the ﬁrst generation-HTS-wire

processing.

It is interesting to note that in the summer of 1987 Akimutsu et al. of

Aoyama-Gakuin University and Raveau et al. of University of Caen re-

ported an 8 K superconducting transition in the Bi-Sr-Cu-O, which was

later found to be associated with the n = 1 member of the BSCCO-family.

This important piece of information was lost at the time in the mad rush for

superconductors with a higher T

than YBCO. I even marked their preprints

with “exciting, more study needed!” in August 1987, but took no immediate

action. With greater attention to these early results, the discovery of the

110 K superconducting BSCCO-family without rare-earth could have been

advanced by half a year. A lesson from this episode is never to get trapped

in the turbulence of excitement presented by fashionable ideas.

3.1.4. The second cuprate HTS family without rare-earth with a T

up to 125 K: Tl

n−1

2n+4

, where n = 1, 2, 3, . . .

[TBCCO orTl22(n-1)n]

The discovery

of Tl

n−1

2n+4

[TBCCO or Tl22(n-1)n] with a T

up to 125 K in early February 1988 by Zhengzhi Sheng and Allen Hermann

of the University of Arkansas appeared to justify the optimism suggested

following the discovery of YBCO and BSCCO that more superconductors

with higher T

might be discovered soon. TBCCO was considered by some to

be a preferred material for HTS applications, especially for thin ﬁlm devices,

due to their higher T

, easier formation and higher stability when compared

with BSCCO. Unfortunately, the softness of TBCCO makes it less desirable

for thin ﬁlm device applications, although the experience gained from its

processing proved useful for the later processing of HTS with toxic elements.

After examining the role of R in R123, Sheng and Hermann of the Univer-

sity of Arkansas started to replace the trivalent R in R123 with the trivalent

Tl with an ionic radius similar to R. They found superconductivity up to

September 14, 2010 9:46 World Scientiﬁc Review Volume - 9.75in x 6.5in ch16

The Evolution of HTS: T

-Experiment Perspectives 405

Fig. 8. ρ(T ) of Tl-Ba-Ca-Cu-O shows a T

up to 120 K (Tl2223) in 1988 by Sheng

and Hermann [Z. Z. Sheng and A. M. Hermann, Nature 332, 138 (1988)].

90 K in their mixed phase samples of Tl-Ba-Cu-O with a nominal compo-

sition of Tl

8+x

in early January 1988, about the same time as

Maeda et al. announced their ﬁnding of superconductivity up to 110 K in

Bi22(n-1)n. Sheng and Hermann soon partially replaced Ba with Ca and

observed superconductivity above 120 K in their multiphase samples during

the second week of February 1988 as shown in Fig. 8. The results appeared

in the March 10, 1988, issue of Nature in an article entitled “Bulk super-

conductivity at 120 K in the Tl-Ca/Ba-Cu-O system.”

The n = 2 and

3 members of the Tl22(n-1)n homologous series were quickly identiﬁed in

these samples by Hazen et al. of the Carnegie Geophysical Lab and Torardi

et al. of Du Pont Research Lab only two days after the announcement of the

120 K superconductivity. Parkin et al. of the IBM Almaden Lab obtained

pure-phase samples of Tl2223 and achieved a T

of 125 K, which was the

record-T

at ambient pressure until April 1993. The T

of Tl2223 was raised

to 131 K by Berkeley et al. at the Naval Research Lab by pressures up to

7 GPa in September 1992.

The homologous series Tl22(n-1)n display the layered stacking se-

quence, e.g. (BaO)(TlO)(TlO)(BaO)(CuO

)(Ca)(CuO

) for

September 14, 2010 9:46 World Scientiﬁc Review Volume - 9.75in x 6.5in ch16

406 C. W. Chu

with n = 3. Similar to Bi22(n-1)n, the T

of Tl22(n-

1)n increases with n up to 3 and decreases when n > 3. The maximum T

are 90, 110 and 125 K, respectively, for n = 1, 2 and 3. Depending on the

oxygen content, the T

can be varied by more than 10 K. The structure of

Tl22(n-1)n is similar to that of Bi22(n-1)n. In spite of the gross similarity

between Tl22(n-1)n and Bi22(n-1)n, there exists no structural modulation

along the double (TlO)-layers in contrast to that along the double (BiO)-

layers, suggesting that such a structural anomaly does not play a signiﬁcant

role in their high T

as was initially thought.

The single (TlO)-layer TBCCO homologous series Tl12(n-1)n was later

discovered. The layer stacking sequence is similar to that for Tl22(n-1)n

except that the double (TlO)-layers in Tl22(n-1)n are replaced by the single

(TlO)-layer. The T

of members of this series is, in general, lower than

that for the corresponding members of Tl22(n-1)n. It was found to increase

continuously with n up to n = 4 before it decreases, i.e. T

= 50, 82, 110

and 120 K for n = 1, 2, 3 and 4, respectively. The reasons for the lower

s and for the continuous increase of T

to n = 4 could be interesting for

understanding the inner atomic structure inﬂuence on T

, but unfortunately

remain unknown.

3.1.5. The cuprate HTS family with the highest T

of 134 K at

ambient and 164 K at 30 GPa: HgBa

n−1

2n+3−δ

where n = 1, 2, 3, . . . [HBCCO or Hg12(n-1)n]

The discovery of HgBa

n−1

2n+3−δ

[HBCCO or Hg12(n-1)n] with a

record-T

up to 134 K in mid April 1933 by A. Schilling et al. of ETH at

ambient

and to 164 K at 30 GPa in Fall 1993 by C. W. Chu et al. of

Houston and H. K. Mao et al. of Carnegie Geophysical Lab

provides new

challenges as well as opportunities for HTS science and technology. The

s of the optimally doped samples increase in parallel by about 30 K for

all members of Hg12(n-1)n up to ∼ 30 GPa in contrast to other HTSs.

A T

of 134 K is above the temperature in the cargo bay of the Space

Shuttle when orbiting in Earth’s shadow and above the boiling point of

liquid natural gas (LNG) at ∼ 110 K. The former makes HBCCO a possible

material for HTS devices operable on the Space Shuttle without cryogen and

the latter may enable the development of the combined HTS/LNG system

for the eﬃcient delivery of electrical and chemical energies simultaneously.

In addition, superconductivity at 164 K in HBCCO under pressure can be

achieved by the use of the common air-conditioning technology.

In late 1989, the T

of cuprates appeared to have stagnated at 125 K

since the Spring 1988. A prominent chemist speculated that the T

September 14, 2010 9:46 World Scientiﬁc Review Volume - 9.75in x 6.5in ch16

The Evolution of HTS: T

-Experiment Perspectives 407

Fig. 9. ρ(T ) of Hg-Ba-Ca-Cu-O shows a T

up to 134 K (Hg1223) by Schilling et

al. in 1993 [A. Schilling et al., Nature 363, 56 (1993)].

cuprates could not exceed 160 K based on some physical chemistry ar-

guments. However, HgBa

7−δ

was discovered in 1993 to dis-

play a superconducting transition at ∼133 K by Schilling et al. of

ETH at Zurich as shown in Fig. 9 and the results appeared in the

May 6, 1993, issue of Nature in an article entitled “Superconductivity

above 130 K in the Hg-Ba-Ca-Cu-O system.”

The homologous series

of HBCCO, Hg12(n-1)n, was quickly identiﬁed to have the layer stack-

ing sequence, e.g. (BaO)(HgO

1−δ

)(BaO)(CuO

)(Ca)(CuO

) for

HgBa

9−δ

with n = 3. When Hg1223 is subjected to high pres-

sures, its T

rises to 164 K at ∼30 GPa before exhibiting signs of saturation

as shown in Fig. 10, clearly contradicting the predicted 160 K-limit made in

1989 for cuprate HTSs. A T

= 134 K and 164 K remain the records to date

at ambient and high pressure, respectively.

It is interesting to note that attempts were made as early as in 1991 to

substitute the linearly coordinated Hg

for the similarly coordinated Cu

in the (CuO)-chain-layer of R123. Compounds of HgBa

EuCu

with a

structure similar to Tl1212 were made but found not superconducting. A

small superconducting signal due to an impurity phase at 94 K was detected

in 1991 in one of our HgBa

EuCu

samples without recognizing that the

impurity phase was Hg1201. The whole story of HBCCO did not start to

unfold until September 1992 when Antipov of Moscow State University and

Marezio of CNRS at Grenoble succeeded in synthesizing Hg1201 with a T

94 K. Knowing that increasing the number of (CuO

)-layers per cell will lead

September 14, 2010 9:46 World Scientiﬁc Review Volume - 9.75in x 6.5in ch16

408 C. W. Chu

Fig. 10. ρ(T ) of Hg12(n-1)n for n = 1, 2, 3 under pressures up to 45 GPa shows

a T

up to 164 K in 1993 by Gao et al. [L. Gao et al., Phys. Rev. B 50, 4260(R)

(1994)].

to an increase of T

, A. Schilling et al. of ETH added Ca to the compound

to achieve n = 2 and 3 for Hg12(n-1)n and reported an enhanced T

up to

∼133 K. After overcoming issues with the complex chemistry of HBCCO,

we and several other groups isolated the diﬀerent phases and showed the

maximum T

’s to be 97, 128 and 134 K, for n = 1, 2 and 3, respectively.

While Hg12(n-1)n has a layer structure similar to Tl12(n-1)n, there ex-

ists a subtle diﬀerence, presumably arising from the linear oxygen coordina-

tion of Hg

-ions in HBCCO as reﬂected in the relatively short Hg-O bond

length along the c-axis and the large number of voids in the HgO

1−δ

-layer.

We therefore proposed that HBCCO could be diﬀerent from other cuprate

HTSs in that Hg might contribute to the density of states near the Fermi

surface and that the O-doping range might be large. Band calculations also

showed that large density of states associated with a van Hove singularity

existed near the Fermi surface. Higher T

is therefore expected in HBCCO

under pressures. Experiments on optimally doped pure Hg1201, 1212 and

1223 were carried out under pressures. We quickly found that the T

grows

with pressure without saturation up to 18 GPa. Later, in October 1993, we

extended the pressure to 45 GPA in collaboration with Mao et al. of the

Carnegie Geophysical Lab. The T

was found to peak at ∼118 K in Hg1201

at ∼24 GPa, at ∼154 K in Hg1212 at ∼29 GPa and at 164 K in Hg1223

at ∼30 GPa.

The present day record T

of 164 K was thus obtained. The

September 14, 2010 9:46 World Scientiﬁc Review Volume - 9.75in x 6.5in ch16

The Evolution of HTS: T

-Experiment Perspectives 409

pressure-dependent T

is shown in Fig. 10. The T

s of the three members

of the Hg12(n-1)n series appear to vary with pressure in a parallel fashion

independent of n and to show no sign of pressure-induced structure change

up to 45 GPa. The observation strongly suggests that superconductivity of

members of the Hg12(n-1)n homologous series arises from a common ori-

gin. However, the unusually large pressure-induced T

-enhancement, even in

their optimally doped states, cannot be accounted for by models commonly

used to explain the pressure eﬀect on T

for other HTSs. A modiﬁed rigid

band model has been proposed for the observations.

3.1.6. The interstitially doped HTS family that forms the base of

TBCCO and HBCCO without a volatile toxic element with a T

of 126 K at ambient: Ba

n−1

[BCCO or 02(n-1)n]

The discovery

of Ba

n−1

[BCCO or 02(n-1)n] with a T

up to

126 K at ambient by C. W. Chu et al. of the University of Houston in 1997

demonstrates that a formal multi-substructure in a HTS serving as a charge

reservoir may not be essential, provided that alternative means, such as

interstitial doping outside the active block, can be found.

It has been demonstrated that the layer cuprate HTSs consist of two

substructures, namely, the active block of [(CuO

n−1

(CuO

)

n−1

] and the

charge reservoir block of [(EO)(AO)

(EO)]. The active block comprises

n-square-planar-(CuO

)-layers interleaved by (n−1)-R-layers and the charge

reservoir block contains m-(AO)-layers bracketed by two chemically inert

(EO)-layers. It is known that processing of cuprate HTS materials into

practical forms is a challenge due to the chemical and physical complexity of

the materials. We therefore tried to synthesize layer cuprate HTSs of simpler

structure, e.g. without the toxic elements such as Tl and Hg in the charge

reservoir block. We attempted to deactivate the charge reservoir block by

removing the m-(AO)-layers and allowing interstitial doping to take place

in the two (EO)-layers that usually bracket the m-(AO)-layers. The com-

pound does not form at ambient. However, through high pressure synthesis

processing in a C- and O-free environment, we succeeded in stabilizing the

n−1

[BCCO or 02(n-1)n-Ba]. The XRD shows a layer lattice

following the I4/mmm space group symmetry. It superconducts with a T

126 K for the n = 4 member and the T

increases to ∼150 K under pressure.

The results (Fig. 11) were published in the August 22, 1997, issue of Science

in an article entitled “Superconductivity up to 126 Kelvin in Interstitially

Doped Ba

n−1

[02(n-1)n-Ba].”

September 14, 2010 9:46 World Scientiﬁc Review Volume - 9.75in x 6.5in ch16

410 C. W. Chu

Fig. 11. χ(T ) of Ba

n−1

shows a T

up to 126 K in 1997 by Chu

et al. [C. W. Chu et al., Science 277, 1081 (1997)].

Although Ba

n−1

[BCCO or 02(n-1)n] meet our original goal

to possess a simpler structure without toxic elements and are interesting

from chemistry and physics points of view, the compounds are not very

stable and degrade in the presence of humid air, and thus of no application

value. Its homologues Sr

n−1

(SCCO) were found also to exist with

high T

3.1.7. The simplest inﬁnite layer HTS family defected RCuO

where

R = Ca with a T

∼ 40–110 K [RCO or 0011-R]

The discovery

21,22

of superconductivity in defected CaCuO

, the building

block of cuprate HTSs, with a T

∼ 40–110 K by John Goodenough et al. of

the University of Texas and Takano et al. of Kyoto University in 1991–1992

could provide the basis to unravel the mysteries of HTS in cuprates if the

microstructure of the defected CaCuO

is determined to ascertain that the

superconductivity observed is caused by the inﬁnite-layer phase.

After the discovery of R123, it was suggested that a higher T

might

be achieved by increasing the number of (CuO

)-layers per unit cell (n) in

cuprates. Indeed, T

was later found to increase with n, although only up

to 3 or 4. It seemed to be interesting to synthesize a cuprate with large n,

preferably for n = ∞, the building block of cuprate HTSs, i.e. RCuO

[RCO