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

C. W. Chu

University of Houston,

Lawrence Berkeley National Laboratory

∗

Hong Kong University of Science and Technology

Dedicated to John Bardeen, Leon Cooper and Bob Schrieﬀer,

the architects of superconductivity.

The rise of the superconducting transition temperature T

has been

reviewed in three major superconducting systems: the cuprate, the

Fe-pnictide and the heavy fermion. While the ﬁrst two systems display

high T

s, heavy fermion superconductors show low T

but embody many

crucial features found in the others. The prospect of future superconduc-

tors with higher T

, preferably close to room temperature, is also discussed.

Those interested in the detailed physics of high temperature superconduc-

tivity are referred to the article by E. Abrahams in the next chapter of this

book and reviews published elsewhere.

1. Introduction

The search for superconductors with a higher transition temperature (T

) has

been one of the major driving forces in the long-sustained research eﬀort on

superconductivity ever since its discovery in 1911.

The eﬀort has led to the

rise of T

and to the discoveries of novel compounds, both superconducting

and non-superconducting; new physics, such as non-Fermi liquid behavior in

cuprates superconductors; and new phenomena, such as heavy electrons in

lanthanide compounds.

Traditionally, high temperature superconductor (HTS) is a relative term

referring to superconducting compounds with a T

comparable to the existing

-record, which rises with time. Before 1953, any superconductors with a

comparable to 9 K of Nb were considered to be HTSs. However, after

the discovery of the A15 inter-metallic superconductors, only those with a

∗

Until September 1, 2009.

391

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

392 C. W. Chu

Fig. 1. The evolution of T

with time and the three periods, i.e. low tempera-

ture superconductivity (LTS), high temperature superconductivity (HTS) and very

high temperature superconductivity (VHTS) or room temperature superconductiv-

ity (RTS).

close to 23 K of Nb

Ge were treated as HTSs. Following the discovery of

cuprate superconductors in 1986, HTSs have been more or less referred to

superconductors with a T

above 23 K and sometimes only to those with a

above 77 K, the liquid nitrogen boiling temperature.

The rise of T

with time is summarized in Fig. 1. In accordance with

the materials, the evolution of T

with time may be roughly divided into

three periods: low temperature superconductivity (LTS) before 1986 that

deals mainly with inter-metallic compounds or alloys with the highest T

23 K; high temperature superconductivity (HTS) between 1986 and 2009

(or later) that concerns mostly cuprates with the highest T

at the current

records of 134 K at ambient and 164 K under high pressure and the recently

discovered iron-pnictides with a highest T

at 57 K; and very high temper-

ature superconductivity (VHTS) after 2009 (or later when a new T

-record

or a new material system is found) in the hope of achieving a T

higher than

the current record and/or new insight to HTS, not to mention uncovering

possible new physics.

2. The LTS-Period: 1911

1986

Before 1986, extensive eﬀorts to search for higher T

had been carried out

mainly on elements, alloys and inter-metallic compounds. In 1955, Bernd

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

The Evolution of HTS: T

-Experiment Perspectives 393

Matthias

proposed the empirical rule according to which T

generally peaks

at the valence-electron-per-atom ratios (e/a) of ∼4.75 and 6.4. In the

ensuing 33 years, mainly due to Matthias’s eﬀort, a generation of super-

conducting inter-metallic alloys and compounds was born, giving rise to the

then-record T

of 23 K observed in Nb

Ge with an e/a = 4.75 in 1974,

attainable in a liquid hydrogen environment. The results had also demon-

strated the possible correlation of instabilities with superconductivity of high

, as evidenced by the simultaneous occurrence of lattice instabilities and

high temperature superconductivity in diﬀerent material families with a rel-

atively high T

within their respective families.

In 1957, John Bardeen, Leon Cooper and Bob Schrieﬀer developed the

comprehensive microscopic theory of superconductivity, the well-known BCS

theory.

According to the theory, the electrons, in the presence of a phonon-

mediated eﬀective attraction V between them, form pairs and undergo a

phase-coherent superconducting transition at a temperature

= 1.14Θ

exp[−1/N (E

)V ] , (1)

where Θ

is the Debye temperature and N(E

) the electron density of

states at the Fermi surface E

. According to this relationship, a higher Θ

N(E

) and/or V is expected to give rise to a higher T

. The descriptive

power of the theory has been amply demonstrated by the accounts of many

successful experiments in the ensuing decades. For instance, the Matthias

empirical rule can be understood in terms of the variation of N(E

) with

e/a and the simultaneous occurrence of lattice instabilities and high T

terms of the large N(E

)V . Unfortunately, BCS theory fails to provide a

deﬁnitive guide in the search for superconductors with high T

other than

the simplistic suggestion mentioned above. The failure can be attributed

to: (1) the small superconducting energy scale in comparison with the large

energy scales associated with the many other excitations in solids and (2)

the coupling between the three parameters, Θ

, N(E

) and V , in the BCS

-formula. The mismatch of energy scales makes an accurate T

-calculation

diﬃcult and the coupling between Θ

, N(E

) and V renders the simplistic

approach to raise T

mentioned above impractical.

The correlation between high temperature superconductivity and lattice

instabilities had therefore become one of the most investigated problems in

superconductivity in the 1970s and 1980s. The goal was to determine if

lattice instabilities would be detrimental to or helpful for superconductivity

at high temperature. This is because in almost all superconductors with

a relatively high T

, such as A15 superconductors, lattice instabilities were

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

394 C. W. Chu

detected. They reveal themselves as a martensitic transformation with a

phonon softening indicative of the large electron-phonon coupling favorable

to superconductivity. However, phonon stiﬀening returns upon the comple-

tion of the martensitic transformation, suggestive of a reduction in electron-

phonon interaction. The correlation had thus been studied extensively by

adjusting the lattice instabilities and T

via high pressure without introduc-

ing any chemical or physical complexity. It was found that lattice instabilities

do aﬀect the T

of the A15 superconductors, but only slightly by no more

than a few tenths of one degree.

The observation led us to conclude that

the incipient instabilities associated with large N and/or V do not present

an insurmountable obstacle to higher T

, provided a catastrophic instability

is avoided. This gave my group the conﬁdence that superconductivity at

higher T

might be achievable.

Unfortunately, T

stagnated at 23 K after 1973, which was reinforced

by the T

-limit of 30’s K predicted theoretically.

Indeed, one of the most

diﬃcult challenges superconductivity researchers faced then was the crisis

in conﬁdence. The ﬁeld was put in a very depressed state in 1986 as the

government and industry largely withdrew support for superconductivity

study, bringing the search for higher T

almost to a halt. However, the

discovery of cuprate HTSs in 1986 drastically changed the situation.

3. The HTS-Period: 1986 2009

In the past 23 years, many HTSs have been discovered in diﬀerent material

systems and many studies carried out. However, in this paper I shall re-

view only the T

-evolution of three systems, i.e. the cuprate system that is

the largest and possesses the highest T

, the newly discovered iron-pnictide

system that may provide important insight into the inner working of HTS

and the low T

-heavy fermion system that is rich in physics and bears great

similarities to the cuprate and iron-pnictide HTSs. Over the years, studies

have revealed exciting physics and challenges in all three compound systems

that have been covered by numerous review articles. Those whose interest

goes beyond T

may refer to the article by E. Abrahams in the next chapter

of this book and review articles cited later.

3.1. The cuprate system

Disappointed by the slow progress in raising T

in the inter-metallic com-

pounds over the years, Alex Mueller and Georg Bednorz of IBM Zurich Re-

search Lab decided to leave the beaten path of inter-metallics and dive into

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

The Evolution of HTS: T

-Experiment Perspectives 395

when the two (AO)-layers are absent.

Fig. 2. The layer structures of the three major cuprate HTS families,

n−1

2n+3

, A

n−1

2n+4

and E

n−1

2n+2

, as exempliﬁed

by members with n = 3.

the woods of oxides for higher T

. They were encouraged by the relatively

high T

of ∼ 14 K detected in the perovskite-like oxides of Li

1+x

2−x

and Ba(Pb

1−x

and wanted to raise the T

of oxide superconductors

further by increasing the electron-phonon interaction and the carrier con-

centration. Their seminal work on La-Ba-Cu-O in 1986 heralded the new

era of HTS of cuprates. Many cuprate HTSs have since been found. De-

spite the diﬀerences in details, all cuprate HTSs discovered exhibit a layer

structure as shown in Fig. 2 and can be represented by a generic formula

n−1

2n+m+2

, where A = Bi, Tl, Pb, Hg or Cu; E = Ca, Sr or

Ba; R = Ca, Y or rare-earth; m = 0, 1 or 2 and n = 1, 2, . . . . The generic

formula can be rewritten as A

n−1

2n+m+2

= [(EO)(AO)

(EO)]

+ [(CuO

n−1

(CuO

)

n−1

], which consists of two substructures: the ac-

tive block of [(CuO

n−1

(CuO

)

n−1

] and the charge reservoir block of

[(EO)(AO)

(EO)]. The space group for compounds with m = 2 or 0 is

I4/mmm but changes to P4/mmm when m = 1. The active block com-

prises n-square-planar-(CuO

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

charge reservoir block contains m-(AO)-layers bracketed by 2-(EO)-layers

which are relatively inert chemically. Superconducting current ﬂows mainly

in the (CuO

)-layers and doping takes place in the charge reservoir block,

which transfers charges without introducing defects into the (CuO

)-layers

(similar to modulation-doping in semiconducting superlattices). All cuprate

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

396 C. W. Chu

HTSs can simply be designated as Am2(n-1)n or just 02(n-1)n-E when the

(AO)-layers are absent.

We shall recall brieﬂy the discovery of the main cuprate HTS families

to which the following HTSs belong: (1) the ﬁrst 35 K HTS, the Ba-

doped La

CuO

[214 or La2001]; (2) the ﬁrst 93 K HTS, YBa

[123,

YBCO or Cu1212]; (3) the ﬁrst 110 K HTS, Bi

[BSCCO or

Bi2223]; (4) the ﬁrst 120 K HTS, Tl

[TBCCO or Tl2223];

(5) the current T

-record holder of 134 K at ambient and 164 K at 30 GPa,

HgBa

9−δ

[HBCCO or Hg1223]; (6) the base of HTS without a

toxic element, Ba

[BCCO or 02(n-1)n-Ba]; and (7) the simplest

inﬁnite layer HTS family, RCuO

[RCO or 0011-R]. Reviews on the experi-

mental results of these HTS families can be found in Ref. 7.

3.1.1. The ﬁrst cuprate HTS family with a T

up to ∼ 35 K: doped

CuO

, where R = La, rare-earth [R214 or R2001]

The seminal discovery

of the high temperature superconductors, Ba-doped

CuO

[La214 or La2001] with a T

of ∼ 35 K, in mid April 1986 by Alex

Mueller and Georg Bednorz of Zurich IBM Research Lab marked the start

of the new era of HTS of cuprates. It showed that HTS is possible in oxides.

A comprehensive account of the discovery of the Ba-doped 214,

(La

1−x

)

CuO

by Mueller and Bedorz has been published elsewhere.

In their search for superconductivity with a high T

in materials with a

strong electron-phonon coupling and higher carrier concentration, Mueller

and Bednorz started in 1983 by examining the perovskite oxides known to

display the Jahn–Teller eﬀect, which promotes electron-phonon interaction.

In January 1986, they detected resistively superconductivity up to 35 K in

their multiphase La-Ba-Cu-O samples, setting a new T

-record as shown in

Fig. 3. The resistive result was published in the September 1986 issue of

Zeitschrieft f¨ur Physik in an article entitled “Possible High T

Superconduc-

tivity in the Ba-La-Cu-O System.”

Initially, the resistive data did not attract too much attention except

by a few, since oxides are often insulators and not even metallic, let alone

superconducting at high temperature. Their magnetic data showing the

decisive Meissner eﬀect did not reach the U.S. until late January 1987. Little

action was taken after the news broke except by a lucky few. Among these

few, Shoji Tanaka and Koichi Kitazawa’s group at the University of Tokyo

and my group at the University of Houston quickly reproduced the Zurich

resistive results, and so reported on December 4, 1986, at the Materials

Research Society (MRS) Fall Meeting in Boston. The rapid reproduction

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

The Evolution of HTS: T

-Experiment Perspectives 397

Fig. 3. The ρ(T ) of La-Ba-Cu-O shows a T

of 35 K (La214) in 1986 by Bednorz

and M¨uller [J. G. Bednorz and K. A. M¨uller, Z. Phys. B 64, 189 (1986)].

by both groups was helped by their prior experience studying the oxide

superconductors Ba(Pb

1−x

and Li

1+x

2−x

. The Meissner eﬀect

was reproduced by the Tokyo group quickly. The same group also identiﬁed

the superconducting phase in the mixed-phase sample to be the Ba-doped

CuO

[La214 or La2001], as announced at the 1986 MRS Fall Meeting.

This compound (La

1−x

)

CuO

has the La

CuO

layer structure with a

layer stacking sequence La

CuO

= (LaO)(LaO)(CuO

After reproducing the results of Mueller and Bednorz, the Houston group

applied pressures to the La-Ba-Cu-O mixed-phase samples, focusing only on

the superconducting phase in an attempt to understand the cause for such an

unexpectedly high T

. The T

was quickly raised to 40 K and then to 52 K

at a rate more than 10 times that of the inter-metallic superconductors.

The observation of a T

higher than 40 K shattered the then-theoretically

predicted T

-limit of 30s K.

The unusually large positive pressure eﬀect

on T

observed also suggested that cuprate superconductors may form a

class of their own and warrant further study, although some still suggested

at the time that the 35 K cuprate superconductor was nothing more than a

conventional BCS-superconductor because its T

still fell within what theory

predicted.

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

398 C. W. Chu

It was later observed that superconductivity with a similar T

can also

be induced in La214 by Sr-doping or excess oxygen. The structure of these

La214 compounds displays the so-called 1T-phase and the carriers exhibit

hole-characteristics. A similar layer structure with a diﬀerent stacking known

as the 1T’-phase was stabilized in R124 with R = smaller rare-earth such as

Nd, Sm, Eu, Gd or Tb and can be electron-doped by partial Ce-replacement

for 1T’-Nd214 to induce superconductivity.

A 1T

∗

-phase, which is a mix-

ture of the T- and T’-phases, each ﬁlling half of the unit cell, was also

synthesized by properly adjusting the matching between the various atomic

radii and synthesis conditions. These new 214-phases display a lower T

but

oﬀer the opportunity to examine the symmetry between the roles of electron-

and hole-dopings in HTSs.

3.1.2. The ﬁrst liquid nitrogen cuprate HTS family with a T

∼ 93–100 K,

RBa

, where R = rare-earth [R123, RBCO or Cu1212]

The discovery

of YBa

[Y123, YBCO or Cu1212] with a T

∼ 93 K

in late January 1987 by Paul C. W. Chu, Maw-Kuen Wu and colleagues in

their respective groups at the University of Houston and the University of Al-

abama at Hunstville ushered in the modern era of HTS. It brought down the

liquid nitrogen temperature barrier of 77 K, representing a giant advance-

ment in modern science and technology and drastically changing the psyche

of superconductivity research. It is the most desirable HTS compound for

device applications to date.

As mentioned earlier, the Houston group was among the very few non-

skeptics who took seriously the initial report from the Zurich IBM Research

Lab.

This was because, since the mid-70s, they had actively investigated

the unstable superconductors, including the perovskite-type oxides, such as

BaPb

1−x

and Li

1+x

2−x

. Not only did they learn that supercon-

ductivity at higher temperature is possible, but they also mastered the oxide

synthesis skills crucial for later cuprate HTS studies. The observations of

an onset T

∼ 40–52 K in the mixed-phase La-Ba-Cu-O samples under pres-

sure,

above the then-theoretically predicted T

-limit of 30s K, enhanced

their conﬁdence in achieving higher T

and raised doubts about the old the-

oretical models. The unusually large positive pressure eﬀect on T

observed

further suggested to me that higher T

might be obtained through chemi-

cal pressures by replacing elements in the compound with those of smaller

ionic radii of the same valences, such as Ba by Sr or Ca and La by the

non-magnetic Y or Lu.

The Ba–Sr replacement to raise T

was quickly

conﬁrmed, but the Ba-Ca substitution was unfortunately found to lower

the T

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

The Evolution of HTS: T

-Experiment Perspectives 399

In many scientiﬁc discoveries, “a kick of luck,” as Mueller put it, often

plays a role. At the 1986 Fall MRS Meeting in Boston on December 4,

Kitazawa from Tokyo announced the identiﬁcation of the superconducting

La214-phase in the La-Ba-Cu-O mixed-phase samples, after my presentation

on BaPb

1−x

and disclosure of duplicating the observation by Mueller

and Bednorz. It was natural for people, including ourselves, to make pure

214-phase samples, preferably single crystals, and to examine the origin of

the 35 K-T

before contemplating how to raise T

. We tried but failed to

grow La214 single crystals, following the destruction of two of our three ex-

pensive crystal-growing Pt-crucibles. I made a crucial decision to turn our

attention to stabilizing the high temperature resistivity anomalous drops,

indicative but not yet a proof of superconductivity, that we detected spo-

radically in the multiphase La-Ba-Cu-O samples but not in the pure 214

ones, by changing the stoichiometry of La:Ba:Cu and/or replacing La with

Y and Lu.

It should be noted that the ﬁrst sign of superconductivity

evidenced by a resistivity drop at a temperature of ∼ 75 K was detected

on November 25, 1986, in one of our La-Ba-Cu-O mixed phase samples,

although it was too ﬂeeting to make a deﬁnitive characterization due to the

unstable nature of the samples. However, I showed the preliminary data to

M. K. Wu at the Fall MRS meeting in Boston and successfully convinced him

to join the search. On January 12, 1987, we observed a large diamagnetic

shift or Meissner signal up to ∼ 96 K in one of our mixed-phase La-Ba-

Cu-O samples, representing the ﬁrst deﬁnitive superconducting signal de-

tected above the liquid nitrogen temperature of 77 K, as displayed in Fig. 4.

Unfortunately, the sample degraded and the diamagnetic signal disappeared

the following day. No eﬀort of ours in the ensuing two weeks succeeded to

reproduce and stabilize this high temperature superconducting signal. I de-

cided to report details of the observation and let other groups stabilize and

identify the high temperature superconducting phase. No sooner than half of

the paper was drafted, M. K. Wu called with great excitement from Alabama

in the afternoon of January 29, 1987, and said that a resistive drop indicative

of a stable superconducting transition above 77 K was detected in the mixed-

phase Y-Ba-Cu-O samples. On January 30, we observed the Meissner eﬀect

in their sample. Stable superconductivity at ∼ 93 K was ﬁnally achieved

(Fig. 5), nearly tripling the T

of La214. The 93 K superconductivity was

attributed to a non-214 phase, based on the high pressure studies carried out

immediately afterward. The results appeared in the March 2, 1987, issue of

Physical Review Letters in an article entitled “Superconductivity at 93 K in

a New Mixed-Phase Y-Ba-Cu-O Compound System at Ambient Pressure.”

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

400 C. W. Chu

Fig. 4. χ(T ) of La-Ba-Cu-O shows the ﬁrst Meissner signal above 77 K (La123)

(but unstable) by Chu et al. in 1987 [C. W. Chu, AIP Conf. Proc. 169, 220 (1988)].

in 1987. [C. W. Chu, AIP Conf. Proc. 169, 220 (1988).]

Fig. 5. ρ(T ) and χ(T ) of Y-Ba-Cu-O shows the T

above 93 K (Y123) in 1987 by

Wu/Chu et al. [M. K. Wu et al., Phys. Rev. Lett. 58, 908 (1987)].

Contrary to the earlier R214-case, labs all over the world reproduced

our results soon after the news broke out. In about a month, the su-

perconducting phase YBa

(known as Y123, YBCO or Cu1212)

was identiﬁed and its structure resolved in collaboration with Bob Hazen

et al. from the Carnegie Geophysical Laboratory in Washington, D.C. The

Y123-phase has the layer structure with layers stacking in the sequence:

YBa

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

)(Y)(CuO

). It was interesting

to note that the La123-phase with a high T

had already been made in

the La-Ba-Cu-O sample on January 12, 1987, when the X-ray diﬀraction

taken then was compared to the Y123 pattern later. With the structure