High-Temperature Superconductivity:

History and Outlook

Shoji TANAKA
Director General
Superconductivity Research Laboratory
1-10-13 Shinonome Koto-ku, Tokyo 135-0062, JAPAN

HISTORY OF HIGH- conductivity in charge density wave systems,

0.010
TEMPERATURE SUPER- and in 1964 Little 2) proposed a so-called exci-
CONDUCTIVITY tonic superconductivity model. No such super-
After superconductivity was conductors, however, have actually been found.
discovered by Onnes in 1911, many Sleight 3) reported in 1975 that supercon-
0.008
superconductors were discovered ductivity was seen in BaPb(1-x)BixO3 and that the
and the critical temperature rose critical temperature of this material, which can
year by year. More than 2000 su- be as high as 13 K, changes with the Bi/Pb ra-
0.006 perconducting materials had been tio. This was just the beginning of research on
discovered by 1975, and the criti- the oxide superconductors. Tanaka and others
ρ(Ωcm)

4)
cal temperature had reached 22.3 immediately studied this material and found
K with the discovery of Nb3Ge in that the carrier concentration in it is more than
0.004
1973. After that, however, no an order of magnitude smaller than that in or-
7.5 A/cm2
2.5 A/cm2 higher critical temperature was dinary metals. At that time there was a little
0.5 A/cm2
obtained for more than 10 years. bit of hope that the superconductivity seen in
0.002 When the BCS theory in 1957 this material is non-BCS superconductivity, but
provided an elegant explanation for today it is still thought to be BCS superconduc-
superconducting phenomena, tivity.
0
many scientists believed that higher In the early 1980s investigators all over
0 10 20 30 40 50 60
T(K) critical temperatures could not be the world began looking for new types of su-
reached without finding new super- perconductivity. In Switzerland a new super-
Fig. 1 Low-temperature Resistivity of a Ba-doped LaCuO3 Sample conducting phenomena. Already conductor, PbMo6S8 was found by Chevrell. It
with x (Ba)=0.75, Recorded for Different Current Densities
in 1954 Fröhlich 1) had proposed a had high upper magnetic critical field, but its
model of high-temperature super- critical temperature was still only about 16 K.
In Japan a government project called “New Su-
perconducting Material” started in 1984, and
in the United States a new conference on the
“Material and Mechanism of Superconductiv-
ity” was held in 1985.
2
.5
Bednorz and Müller5) found early in 1986
0
that in Ba-doped LaCuO3 the temperature de-
0

-1 pendence of the conductivity in the transition

-2
region from the normal state to the supercon-
-2
Magnetic Susceptibility ( 10- 4 emu/g )

ducting state changes with the current density

-4
-3 (Fig. 1) and they pointed out this might indi-
19 21 23 25 27 cate the possibility of high-temperature super-
T( K)
-6 conductivity. Tanaka, Kitazawa, Uchida, and
Takagi 6) at the University of Tokyo immediately
-8 started to explore this possibility and at the end
of 1986 reported that measurements of resis-
H=1.0 0e Fig. 2 Temperature
- 10
Dependence of the tivity and diamagnetism confirmed that the
Magnetic Susceptibility critical temperature for superconductivity in Ba-
- 12
in Ba-doped La2CuO4
0 5 10 15 20 25 30 doped La2CuO4 was as high as 30 K (Fig. 2).
Temperature( K)
Furthermore, in early 1987 Tanaka 7) pointed

JSAP International No.4 (July 2001) 17

JSAPI_Vol4.p65 Page 17 01.7.17, 10:11 AM Adobe PageMaker 6.5J/PPC

 out the possibility of two-dimensional super-
conductivity due to the layer structure of this
material. At almost the same time, Anderson
8)
proposed the mechanism of the high-tem-
perature superconductivity based on the two-
dimensional resonating valence band model.
At that time, many scientists began to think
that the superconductivity of Ba-doped La2CuO4
cannot be explained by the BCS theory.
In February of 1987 Chu and others 9)
found a new superconducting material,
YBa2Cu3O7, with a critical temperature above
90 K. Many scientists all over the world then
Fig. 3 OCMG-processed Gd-Ba-Cu-O Bulk (Single-domain 48 mm ø sample)
Mixing ratio Gd123:Gd211=10:5 began to search for new superconducting ma-
Trapped magnetic field 2.2T at 77K
terials with higher critical temperatures. On
March 15 of 1987 the American Physical Soci-
ety held a special symposium on high-tempera-
ture superconductivity. Several thousand sci-
entists gathered in the main hall of the Hilton
Hotel in New York City, where the enthusiastic
atmosphere of the symposium continued until
the dawn of the next day. One physicist called
Fig. 4 Characteristics of High-Jc Bi2223 Conductor
this meeting a Woodstock in physics.
Tremendous advances were made in theo-
Space insertion for the retical and experimental research, and new
absorption of heat contraction Filled-core fine-strand former
force for eddy current reduction materials were found in quick succession. As a
XLPE cable Cooling system
result the critical temperature reached 112 K
for terminations
in a Bi-compound, 126 K in a Tl-compound and
Spiral pitch adjustment 135 K in a Hg-compound. The physical prop-
for AC loss reduction
erties of these compounds were also investi-
150mm duct gated very intensively, and it was confirmed that
3 splitting structure
Short connection of 3 shielding layers in all cuprate superconductors the supercon-
Current transformer
100m 3-phase HTS cable ductivity occurred in very thin layers including
Thermal insulation
property at corner CuO2 planes. Many kinds of peculiar proper-
Splitter box Specification
Voltage 66kV
ties were observed, but except the d-wave sym-
Cooling system Termination
for cable Current 1000Arms metry of the superconductivity, most of them
Transformer for Capacity 114MVA
voltage apply Length 100m are not explained with consistency. The most
Cable type 3-phase, cold dielectric type
Cooling Closed circulation of
peculiar property may be the appearance of the
sub-cooled LN2
stripe structure observed recently in neutron dif-
Fig. 5 Superconducting Transmission Cable System & Design Concept fraction experiments. In this structure, the elec-
tric charges exist in stripes that are separated
by the array of magnetically ordered stripes.
The physical properties in the normal state
of the high temperature superconductors are
so complicated that any theory can not yet ex-
plain them in a consistent way. Deeper under-
standing on “the strongly correlated system”
may be necessary in order to solve the origin
of the high temperature superconductivity.

OUTLOOK FOR THE FUTURE

OF HIGH-TEMPERATURE
SUPERCONDUCTIVITY

[1] Material developments

Fig. 6 Cryogen-Free R&D Magnet There have in the past ten years been

18 JSAP International No.4 (July 2001)

JSAPI_Vol4.p65 Page 18 01.7.17, 10:11 AM Adobe PageMaker 6.5J/PPC

 Refrigerator
posed, but even today there is no consensus
Power lead
among theoretical physicists. There are so many
various kinds of interactions in such compli-
LHe tank
cated systems-electron-phonon interactions,
LN2 tank
spin-spin interactions, charge density waves,
Service port spin density waves, and so on-that it may be
considered we are just beginning to understand
Outer vessel
the physics of these complex systems. Explain-
Supporter
ing this phenomenon clearly will take a long
SC coil Radiation shield time.
Super insulator

Persistent [3] Applications

current switch
Since room temperature superconductiv-
Vacuume
Railway Technical Research Institute
ity has not been found yet, the applications of
high-temperature superconductivity 12) are re-
Fig. 7 Superconducting Magnet for MAGLEV Train stricted to temperatures around that of liquid

many attempts to obtain high-temperature

superconductivity in materials other than (1) Silver-sheathed Bi-Compound Tape
cuprates. Superconductivity was observed in Silver
alkali-ion doped C60 at 33 K 10), and Akimitsu
very recently found superconductivity in MgB2
at of 39 K 11). But the superconductivity in both
Bi-compound
layers
these materials is explained by the BCS theory,
so it can be said at present that all the non-
(2) Tape of Next Generation (1-2-3 compounds)
BCS superconductors belong to cuprate fam-
ily. We therefore need to find new non-BCS Superconductor (axis aligned)
Pinning Centers
superconductors outside the cuprate family if
we raise the critical temperature beyond room Buffer layer
temperature. Room temperature superconduc- (axis aligned) Substrate
tivity is still a dream of many scientists, but there [Multi-layer Structure]
is no guideline to reach it at present. A little
Fig. 8 Structure of Superconducting Tape
bit of hope, however, may be found when we
consider organic compounds. The methods for
analyzing very complicated biological materi-
als, like the human genome, are developing
very fast and it may soon be possible to ana-
lyze the interactions between complicated ge-
107
nomes and proteins. This would enable us to
simulate new organic materials on high-speed
106
supercomputers and evaluate the interaction NbTi(4.2K)
between electrons and molecules in new ma-
terials. This is a way to realize excitonic super- 105
Nd123(Expected, 77K, B//c)
Jc(A/cm 2 )

conductivity suggested by Little in 1964. This

Y123877K, B//c)
means the progress of future computing sci- 104
ence will open a new way for science and tech- Bi-2223(77K, B//c)

nology of new superconducting materials.

103

Tl-1223(77K, B//c)
[2] Theoretical Developments
After the discovery of the high-tempera- 102
0 2 4 6 8 10 12 14
ture superconductivity in cuprates, many kinds
Magnetic Field(T)
of theoretical models of the mechanism of the
high-temperature superconductivity were pro- Fig. 9 Magnetic Field Dependence of Critical Current (prospect)

JSAP International No.4 (July 2001) 19

JSAPI_Vol4.p65 Page 19 01.7.17, 10:11 AM Adobe PageMaker 6.5J/PPC

 (b) Superconducting Tapes
The first generation of superconducting
Biaxial In-plane Alignment
Homogeneity
tape using high-temperature superconductors
High Production Rate is the so-called Silver-Sheathed Bi-Compound
Thickening
Tape. The Bi-compound usually used is
Alignment Control for SC Layer Superconducting Layer Bi2Sr2Ca2Co3O10, and the cross section of the
Prevention of Reaction Y-123, Nd123, Sm123,
Suppression of Crack Formation RE(mix)-123
tape is shown in Fig. 4. At 77 K the critical
Homogeneity, Flatness (Thickness; 1-10µm) current of this tape is more than 110 A, and
High Production Rate
the length of commercial tape is more than
Buffer Layer
High Strength YSZ, MgO, CeO2, NiO 1000 m. Trials of the actual applications of
Non-magnetic Materials ZrO2, BaZrO2, etc.
Thinning (Thickness; <3µm) this tape in various fields have already begun.
Alignment Control
for Buffer Layer Substrate A high-current transmission cable is very im-
Ni-Alloy(Hastelloy), Ni,
Ag, Ni-Cr Alloy, Zr, etc. portant when large amounts of electric power
(Thickness; 25-125µm) are sent to the central part of a big city. Be-
cause digging new underground tunnels is very
Fig. 10 Structure of Coated Conductor expensive, it is necessary to use the old tunnel
and new superconducting cables that can carry
large amounts of power even though they have
a small cross section. These trials are being
made in Detroit in the United States and in To-
nitrogen. As the refrigeration systems are well formly distributed throughout the bulk. But kyo, Japan. In Fig. 5 is shown the scheme of
developed recently, those applications at liquid after the bulk is cooled to below the critical the superconducting transmission cable which
nitrogen temperature will greatly contribute to temperature, the magnetic field is quantized is under construction in Tokyo Electric Power
the society. and quantized flux is pinned by strong pinning Company.
The fundamental technologies for appli- centers. Then when the external field is re- Another application is in the supercon-
cations are classified into those of (a) super- moved, the quantized flux is left inside and ducting magnet for the pulling system used to
conducting bulk, (b) superconducting tapes and behaves like a permanent magnet. The make large-diameter crystals of silicon. To pre-
(c) superconducting devices. strength of the trapped magnetic field in vent oxygen from the crucible of fused quartz
GdBa2Cu3O7 bulk reaches more than 2 T at 77 from entering the molten silicon, a strong mag-
(a) Superconducting Bulk K and more than 3.5 T at 30 K (Fig. 3). This is netic field must be applied. The photograph
Materials for superconducting bulks are several times the strength of the magnetic field in Fig. 6 shows a magnet developed for this
REBa2Cu3O7, where RE is Nd, Sm, Gd, or Y. The of an ordinary permanent magnet. system by Toshiba and Sumitomo Electric Co.
bulk is made by using the quenched melt One very fruitful application of this bulk The newest application trial is the development
growth (QMG) method in the case of would be in a water-cleaning system using the of a race-track-shaped magnet for the mag-
YBa2Cu3O7 and in other cases is made by using magnetic separation effect. This system would netically levitated transportation system
the oxygen controlled melt growth (OCMG) be more than 100 times as efficient as the mag- (MAGLEV) train in Japan. As shown in Fig. 7,
method under a low partial pressure of oxy- netic-separation cleaning systems available to- this magnet is about 1 m long and 50 cm in
gen. The diameter of the circular plate of bulk day. wide, and at 20 K the strength of the mag-
reaches more than 10 cm and crystal c-axis is
perpendicular to the plate. In the bulk, the
pinning force of magnetic flux in the supercon-
106 105
ducting state is very strong and at 77 K the
critical current density is more than 104 A/cm2. Generator Generator

105 104
Magnetic field X Inner diamiter of Coil (T-cm)

The applications of the bulk are of two kinds. Bi-based (20K)

SMES SMES
(1) As the pinning force of the magnetic flux is
Current X Wire length (A•km)

104 103
so strong, the outside magnetic field cannot Laboratory
MRI
Laboratory
MRI
Levitation
Levitation
Magnet Magnet
penetrate the bulk in the superconducting
Steel Current Steel Current
103 102
state. This results in a strong levitation force Plant Limter Plant Limter
Chemical
Si single crystal Chemical Si single crystal
when the bulk is close to an ordinary perma- Reaction
(Laboratory)
Reaction
(Laboratory)
102 Magnetic Separator Magnetic Separator 101
nent magnet, and at 77 K this force usually
Electric Power Cable
reaches 15 kg/cm2. This phenomenon can be Electric Power Cable

101 100
exploited to make friction-free flywheel-type Y-based (77K)
Current Lead
electricity storage systems that can store, say, Current Lead

10 kWh and be used as emergency power sup- 100 10-1

1990 1995 2000 2005 2010
plies. (2) When the bulk is in the normal state, AGE

a magnetic field applied from outside is uni- Fig. 11 Expectations of the Development of Superconducting Tape

20 JSAP International No.4 (July 2001)

JSAPI_Vol4.p65 Page 20 01.7.17, 10:11 AM Adobe PageMaker 6.5J/PPC

 netic field at the center of the magnet is 5 T. It one and zero, and logical circuits can be made method is not well established yet.
is hoped that this magnet can replace the mag- by combining SFQ devices. The most impor- With regard to the production process, on
net made of Nb-Ti wire and operated at 4.2 K, tant features of this device are that its opera- the other hand, the SFQ circuits have some
since no quenching phenomena in it is expected tion time is only a few picoseconds and its advantages over semiconductor circuits (Table
and this is very robust against electromagnetic power dissipation is only of the order of 1 1). The most important thing is that SFQ cir-
disturbance from outside. nanowatt. As a result, the circuits comprising cuits with a 1-micrometer line width are good
Possible applications of superconducting these devices have very high operating frequen- enough that operating frequencies greater than
magnets of this type are appearing all the time: cies, of the order of 100 GHz, and consume 100 GHz can be obtained. This means that we
in the continuous casting systems in steel mills, very little power, of the order of 1 mW. This can use 10-year-old lithography equipment in
in high-power motors for ship propulsion sys- operation speed is almost 100 times faster than the semiconductor industry and that we will
tems, and in the superconducting magnetic that of ordinary semiconductor circuits and the be able to produce better characteristic circuits
energy storage (SMES) system. power consumption is about 1/100 that of the in the future.
The development of the next generation semiconductor circuits. Of course there are A SFQ circuit consisting of several thou-
of superconducting tape is underway in Japan, many kinds of difficulties in the development sand Nb junctions has already been tried and
the U.S., and Europe. This tape, shown in cross of the circuits, as the operating principle of this proven to work at very high frequencies at 4.2
section in Fig. 8, has three layers (metal sub- type of circuit is completely different from that K. High-Tc junctions have been made using
strate, oxide buffer layer, and superconducting of semiconductor circuits and the design YBCO, and very recently high-quality Joseph-
layer) and is called a coated conductor. The
superconducting layer is made of REBa2Cu3O7,
and the critical current density (Jc) and its mag-
netic field dependence expected from the re-
sults of testing the materials are shown in Fig. SET

9. The question of what kind of combination

ø0
of three kinds of layers is the most preferable
is quite serious at present as is shown in Fig.
10. If the critical current is to be large, the c- Absence of Presence of
RESET SFQ SFQ
axis of superconducting layer must be perpen- logic value="0" logic value="1"

dicular to the surface of substrate. Therefore Control

the alignment of crystal grains in the a-b plane
Magnetic Field or Current
becomes important, as the critical current de- SQUID
dc-SQUID
creases rapidly when the grain boundary angles SFQ: (Single Flux Quanta)

are more than 10 degrees.

Superconducting
As mentioned above, in the developments Weak Link
rf-SQUID

of the coated conductors there are still difficult

problems left unsolved, but this superconduct- Fig. 12 Controlling SFQ with Use of SQUID
ing tape is expected to reach the market at
around the year of 2005 (Fig. 11).

(c) Superconducting Electronic

Devices
The most prominent phenomenon in su-
perconductivity is Josephson tunneling, and
Table 1 Comparison between Semiconductor and Superconductor circuits
most applications of the superconductivity in
Semiconductor Superconductor
electronics are based on this phenomenon. The
1. Line Width < 0.1µm 0.8µm
most well known application is the supercon-
2. Lithography after 2005 completed
ducting quantum interference device (SQUID).
3. Structure Three Dimensional Two Dimensional
Its extremely high sensitivity to magnetic fields
4. Number of Marks > 20 ≈ 10
is exploited in many fields: medical electronics,
Multi-layered ; high resistance Multi-layered ; no resistance
5. Wiring
mass production (e.g., as non-contact defect heating no heating

detectors), and so on. 6. Wafer

8 inchs ; surface roughness 3 inchs ; surface roughness
several nm's several nm's
Another important application is the single
7. Design Efficient Tool ; Margin. large Small Tool ; Margin. small
quantum flux (SFQ) device. As shown in Fig.
8. Frequency 5 GHz (2005) 50~100 GHz
12, the principle operation of the SFQ device is
9. Output ≈ 1V ≈ 0.1mV
rather simple. The presence and absence of a
Which is easier to make, semiconductor devices or superconductor devices?
single magnetic flux quantum in the SQUID ring
respectively correspond to information signals

JSAP International No.4 (July 2001) 21

JSAPI_Vol4.p65 Page 21 01.7.17, 10:11 AM Adobe PageMaker 6.5J/PPC

 tronics will greatly be changed.
High-quality superconducting microwave
filters operating in the gigahertz region are used
X : 2.0 mV/div. X : 1.0 mV/div. in the base stations of wireless communication
Y : 1.0 mA/div. Y : 0.5 mA/div.
systems, and some of them have already been
commercialized. Wireless communication is be-
coming increasingly important in the informa-
tion society, and therefore it must play an im-
portant role in connection with the develop-
ment of future software communication sys-
at 4.2K at 40K tems.

W = 4µm W = 4µm SUMMARY

Fifteen years have already passed since
high-temperature superconductivity was dis-
Fig. 13 Typical Characteristics in Ramp-edge YBCO Josephson Junction covered in 1986. Many new materials have
been found and the critical temperature of su-
perconducting materials has increased to 135
son junctions have been made by using an ion- of distribution in 100 junctions and 10% in K. To obtain superconductivity at room tem-
bombarded barrier. These types of junctions 1000 junctions have already been obtained. perature, however, we must find materials other
are called ramp-edged junctions, and the typi- The relation between possible circuits and the than those in the cuprate family. It is very dif-
cal characteristics of this type of junction are distribution 1σ of junctions is shown in Fig. 14. ficult to say when that can be done and who
shown in Fig. 13. This figure indicates that cir- This figure shows that a high-speed (40 GHz) can do it. The theoretical basis of high-tem-
cuits consisting of this type of junction can work sampler has already been made and that some perature superconductivity is still uncertain and
at 40 K. circuits of small-scale integration can be made it must remain in the forefront of solid state
In order to integrate the junctions, the dis- in the very near future. If very inexpensive su- physics.
tribution 1σ of the critical current through junc- perconducting chips with excellent character- On the other hand, we can see the future
tions is critically important and until now 8% istics appear on the market, the features of elec- of the superconducting technologies more
clearly than ever and can expect the supercon-
ductivity industry to take off around the year
2005. It must play an important role in the
20
new industrial revolution now underway.
Circuit margin : 20%
: 50% circuit yield
: 99% circuit yield
Digital demos.
15
Required σIc (%)

Digital sampler

10 Counting A/D
High dynamic range A/D
References
4X4 network switch
1) H. Fröhlich: Proc. Roy. Soc. A223, 296 (1954).
32bit DSP
5
2) W.A. Little: Phys. Rev. A134, 1416 (1964).
3) A.W. Slight, J.L. Gillson and F.E. Bierstedt:
MUX/DMUX Solid State Commun. 17, 27 (1975).
4bit DSP
4) T.D. Thanh, A. Kuma and S. Tanaka: Appl.
Spread spectrum MODEM 128X128 switch Phys. 22, 205 (1980).
0 5) J.G. Bednorz and K.A. Müller: Z. Phys. B64,
100 101 102 103 104 105 106 189 (1986).
6) H. Takagi, S. Uchida, K. Kitazawa and S.
Number of junctions in circuit
Tanaka: Jpn. J. Appl. Phys. 26, L 123 (1987).
7) S. Tanaka: Jpn. J. Appl. Phys. 26, 218 (1987).
Fig. 14 Relation between Possible Circuits and the Distribution of Junctions 8) P.W. Anderson: Science 235, 1196 (1987).
9) C.W. Chu et al.: Phys. Rev. Lett, 58, 908
(1987).
10) R.C. Haddon, A.S. Hebard, et al. : Nature 350,
320 (1991).
11) J. Nagamatsu et al.: Nature 410, 63 (2001).
12) As to Applications, see S. Tanaka; Proc. Int.
Conf. On M2SVI / Physica C 341-348, 31
(2000).

22 JSAP International No.4 (July 2001)

JSAPI_Vol4.p65 Page 22 01.7.17, 10:11 AM Adobe PageMaker 6.5J/PPC