Buschow K.H.J. (Ed.) Concise Encyclopedia of Magnetic and Superconducting Materials

Подождите немного. Документ загружается.

is Bi

2.15

1.92

0.75

8.1

with the ‘‘calcium site’’

occupied by 75% calcium, 19% strontium, and 6%

bismuth (Petricek et al. 1990, Coppens et al. 1991).

Other sites in the crystal structure also show cation

mixing, although to a lesser extent.

Nonstoichiometry is an important aspect of high

superconductors. For some phases, the adjustable

nonstoichiometry (solid solution) within a particular

building layer is explicitly stated, e.g., (La,Sr)

CuO

(Nd,Ce)

CuO

, YBa

7d

, and (Ba,K)BiO

.As

described below, nonstoichiometry in the building

layers provides charge carriers via chemical doping.

Pure Bi

CaCu

would not be superconducting,

since the average oxidation state of copper would be

exactly þ2 and there would be no charge carriers,

whereas Bi

2.15

1.92

0.75

8.1

is superconducting

and the average oxidation state of copper is þ2.21.

1.2 Chemical Doping

In addition to structural control, realizing the opti-

mal carrier concentration in the CuO

layers is a

crucial part of the synthesis of high T

materials. The

building layers surrounding the CuO

layers serve as

‘‘charge reservoirs’’ to dope the CuO

layers by

chemical means (Cava 1990). Doping the CuO

layers

with holes or electrons is a necessary, although not

sufﬁcient, condition for achieving superconductivity

in these materials. A convenient method for charge

counting to assess chemical doping is through the use

of the concept of oxidation states. Note that the ox-

idation state of an ion is not the same as the true

charge on the ion. For simplicity, the oxidation state

of an oxygen ion (O) is deﬁned to be –2; hence, in

order to quantify the doping of the CuO

layers, it is

sufﬁcient to know the oxidation state of copper. The

average oxidation state of copper is determined

from the phase composition, the standard oxidation

states of the remaining ions (e.g., Sr

2 þ

,La

3 þ

, etc.),

and the condition of charge neutrality. For example,

in La

2x

CuO

the average oxidation state of cop-

per is (2 þx), and in Nd

2x

CuO

(where Ce

4 þ

present) it is (2x).

In the parent compounds, La

CuO

and Nd

CuO

where no chemical doping is present (x ¼0), the ox-

idation state of copper is þ2 and the compounds are

not superconducting. As x increases and the parent

compounds are doped, charge carriers (holes) enter the

CuO

layers and the average oxidation state of copper

deviates from þ2. Some high T

compounds contain

elements that may assume a variety of oxidation states

(e.g., bismuth, lead, thallium, or mercury); these must

be known in order to determine the oxidation state of

copper in the CuO

layers. The same is true for com-

pounds with nonequivalent copper sites, e.g., those

that contain CuO

1–d

‘‘chains’’ (see Fig. 2(b)).

The oxidation state of copper is greater than 2 for

p-type superconductors and less than 2 for n-type

superconductors. The amount of charge transferred

from the surrounding (charge reservoir) building lay-

ers to the CuO

layers is given by the deviation of the

oxidation state of copper from þ2. One can also talk

of the deviation of the oxidation state of each CuO

unit from 2: [CuO

]

2 þd

, where d is positive for

hole-doped (p-type) and negative for electron-doped

(n-type) superconductors. (While the two jargons are

mathematically equivalent, the latter may be closer to

the physical reality: in high T

cuprates, the holes in

CuO

layers reside in states of predominantly O 2p

character, with relatively small admixture of Cu 3d

(Nu

cker et al. 1988).) Studies of the maximization of

with doping (Tokura et al. 1988, Torrance et al.

1988, Tokura et al . 1989, Shafer and Penney 1990,

Rao 1991) have revealed that a maximum T

for each

structure occurs at about 7d7E0.2, as shown in Fig. 3

for several p-type superconductors.

1.3 Minimum Oxygen Partial Pressure for

YBa

7d

Stability

While some high T

superconductors are thermody-

namically stable at high temperatures and may be

quenched to avoid decomposition, others (e.g.,

YBa

) are not stable at any temperature or

pressure (Sleight 1991). Their synthesis is commonly

achieved by ﬁrst forming a thermodynamically stable

parent structure (e.g., YBa

) and subsequently

Figure 3

as a function of hole concentration, n

, per CuO

unit

for various high T

superconductors as determined by

chemical titration (reproduced by permission of World

Scientific from Mod. Phys. Lett. B 1992, 6, 555–71).

1180

Superconducting Thin Films: Materials, Preparation, and Properties

altering its composition after the sample has been

cooled to a temperature where the parent structure is

kinetically prevented from decomposing (e.g., by

venting the deposition chamber with oxygen as the

ﬁlm cools from its growth temperature).

It is for this reason that the growth of YBa

7d

ﬁlms with dE0 involves a minimum of two

steps: (i) formation of the YBa

parent struc-

ture and (ii) oxidation of this parent structure to

YBa

. For YBa

7d

, temperatures in the

vicinity of 700 1C and 400 1C are used for these two

steps, respectively. The latter temperature is low

enough to kinetically prevent the decomposition of

YBa

7d

, but high enough that the kinetics of

oxygen uptake are rapid (Shi and Capone 1988).

The minimum P

at which a desired high T

struc-

ture (or its parent structure) is thermodynamically

stable is a major constraint in thin-ﬁlm synthesis by

vacuum deposition methods. The lines d

, m

,andd

shown in Fig. 4 bound the region in which the oxygen-

deﬁcient parent structure of YBa

7d

is stable as

a function of substrate temperature and oxygen pres-

sure. The lines were established for bulk YBa

7d

(Lindemer et al. 1991); however, the region in which

YBa

7d

ﬁlms may be deposited is the same for

deposition techniques that use molecular oxygen

Figure 4

Temperature and oxidant pressure dependence of the region in which YBa

7d

is stable. In molecular oxygen,

YBa

7d

is stable in the region bounded by d

, m

, and d

. In atomic oxygen, it is stable in the region between the

lowest two lines. The horizontal line A shows where the incident oxidant ﬂux is just six oxygen atoms per unit cell for a

YBa

7d

growth rate of 1.3 nms

1

(reproduced by permission of Elsevier Science from J. Cryst. Growth 2001,

225, 183–9).

1181

Superconducting Thin Films: Materials, Preparation, and Properties

(Hammond and Bormann 1989). This constraint is

particularly restrictive in the case of molecular beam

epitaxy (MBE) growth, where one must keep the

background gas pressure low in order to maintain the

long mean free path necessary for MBE. This has ne-

cessitated the use of activated oxygen species such as

ozone, atomic oxygen, and NO

(Nonaka et al. 1995,

Schlom and Harris 1995).

The approximate lines that bound the region of

YBa

7d

stability when grown using atomic ox-

ygen are also shown in Fig. 4. In addition to provid-

ing a growth environment that lies within the region

in which YBa

7d

is stable, it is also necessary

for the oxidant ﬂux to be sufﬁcient to supply the

six oxygen atoms per unit cell needed to form

YBa

7d

. This minimum ﬂux depends on the

ﬁlm growth rate and is shown by line A in Fig. 4 for a

YBa

7d

growth rate of 1.3 nms

1

2. Deposition Methods

In what follows, the key techniques for deposition of

thin ﬁlms of high T

superconductors are described.

Note that the quoted deposition conditions are spe-

cific for these materials. The key limiting factor here

is the necessity to fully oxidize copper, which requires

ﬁlm growth either under a relatively high partial

pressure of oxygen, or using more reactive oxidants.

This requirement has prompted attempts at deposi-

tion under unusual thermodynamic conditions, as

well as the development of a range of new hardware

tools. The latter include special heaters that are ox-

idation resistant, mechanisms for shuttling the ﬁlm

during deposition back and forth between neighbor-

ing low-pressure (deposition) and high-pressure (ox-

idation) zones, ozone distillation systems and oxygen

plasma sources for delivery of activated oxygen to

high-vacuum chambers, differentially pumped ana-

lytical tools, etc.

2.1 Sputtering

For the thin-ﬁlm deposition of high T

superconduc-

tors by sputtering, one makes use of discharge in a

gas, typically argon, plus oxygen to oxidize the ﬁlm.

The gas pressure is most frequently kept in the range

50–700 mtorr. The plasma is generated by an electric

ﬁeld, either d.c. or a.c. The ions from the plasma hit

the surface of a target and eject from it mainly neutral

atoms. These are collected onto a substrate, which is

usually heated to promote crystallization. The com-

position of the ﬁlm is largely determined by the sto-

ichiometry of the target, which can involve a single

element or a mixture of elements.

Soon after the discovery of high T

superconduc-

tors, sputtering from a single compound target was

applied with success to grow high T

superconductor

thin ﬁlms, and this remains one of the simplest and

most popular methods to grow such ﬁlms. Good-

quality ﬁlms of La

1.85

0.15

CuO

, YBa

7d

(both c- and a-axis oriented), Bi

CaCu

, etc.,

have been deposited in this way (Suzuki and

Murakami 1987, Xi et al. 1989, Eom et al. 1989,

Wagner et al. 1993). A common problem is resput-

tering of the ﬁlm by negatively charged oxygen ions

and the resulting degradation of the ﬁlm, but this can

be overcome by using higher gas pressure (Poppe

et al. 1988) or off-axis sputtering (Eom et al . 1990) as

shown in Fig. 5.

More advanced sputtering systems contain several

targets. The substrate is moved from target to target

to make multilayers. Using multi-element targets

and switching from one to another, superlattices

have been made of (YBa

7d

)

(PrBa

7d

)

(Triscone et al. 1990, Jakob et al. 1994) and

(Bi

CaCu

)

(Bi

CuO

)

(Li et al. 1994) with

remarkable crystalline coherence.

2.2 Pulsed Laser Deposition

In pulsed laser deposition (PLD), a short-wavelength

laser—most frequently a KrF excimer laser

(l ¼248 nm)—is focused onto a target to ablate ma-

terial. Some of this ablated material reaches and de-

posits on the substrate (Dijkkamp et al. 1987). Like

sputtering, PLD is generally done under a substantial

pressure, typically 100–300 mtorr, during deposition.

Unlike sputtering, argon is not necessary for the

process so pure oxygen is usually used. Apart from

the stoichiometry of the target, the composition of

the resulting ﬁlm may depend on the target-to-subst-

rate distance, the pressure, the laser ﬂuence (energy

per illuminated area on the target per pulse), etc.

Nevertheless, PLD is also a very simple technique

that generally produces good results fast; in partic-

ular, it is well suited to switching from one material

to another. A frequent problem is the occurrence of

micrometer-sized particulates (‘‘boulders’’) that are

observed in the ﬁlms. Various solutions to this prob-

lem have been proposed, such as using off-axis PLD.

More advanced PLD systems have also been built,

and many of these include multiple, interchangeable

targets. Efforts have been made to extend PLD to

large-area substrates, by rotation and translations of

both the target and substrate. Finally, by using ozone

or NO

, either of which is much more reactive than

, PLD has been performed under high-vacuum

conditions (Tabata et al. 1989, Koinuma and Yoshi-

moto 1994). Here, the key advantage is that one can

use in situ analytical tools such as reﬂection high-en-

ergy electron diffraction (RHEED) to monitor the

crystallinity and morphology of the surface of the

growing ﬁlm. This has been accomplished more sim-

ply at the pressures conventionally used for PLD by

differential pumping of the RHEED electron gun (see

Fig. 6) (Rijnders et al. 1997). The growth of individual

1182

Superconducting Thin Films: Materials, Preparation, and Properties

monolayers and the engineering of new structures at

the atomic-layer level have been achieved using PLD

in this way.

2.3 Reactive Evaporation

The simplest variant of reactive evaporation is the use

of resistively heated metallic ‘‘boats’’ as thermal

sources of metal atoms. The evaporation rates are

controlled by tuning the current, and monitored by

quartz crystal microbalances (QCMs). One effective

implementation of this technique, shown in Fig. 7,

uses a black-body-type rotating disk heater, which

has an ‘‘oxidation pocket’’ with a very narrow open-

ing (Kinder et al. 1996). This allows for a relatively

high pressure inside the oxidation pocket

(B10

2

torr), and a low pressure in the rest of the

chamber (B10

4

torr), which extends the lifetime and

stability of the thermal evaporation sources. This is

the preferred technique for the deposition of uniform

YBa

7d

ﬁlms over large-area wafers (up to 9 in

(23 cm) in diameter). The method also has a compar-

atively high throughput and is cost effective.

Figure 6

Schematic of a PLD system, equipped with an in situ

RHEED probe (reproduced by permission of the

American Institute of Physics from Appl. Phys. Lett.

1997, 70, 1888–90).

Figure 7

Schematic of a reactive thermal evaporation system,

with a high oxygen pressure pocket (reproduced by

permission of SPIE from ‘‘Oxide Superconductor

Physics and Nano-engineering II,’’ 1996, pp. 154–9).

Figure 5

Schematics of (a) 901 off-axis and (b) on-axis sputtering from a single target, onto an array of substrates, to study the

compositional spread (reproduced by permission of Elsevier Science from Physica C, 1990, 171, 354–82).

1183

Superconducting Thin Films: Materials, Preparation, and Properties

2.4 Molecular Beam Epitaxy

The coherence length in high T

superconductors is

very short, a couple of nanometers at most. To fab-

ricate good active devices such as Josephson junc-

tions, one thus needs a capability to deposit ultrathin

layers, with perfect interfaces. This becomes even

more critical when one is concerned with large-scale

integration where the devices have to get denser and

smaller, while the uniformity demands increase.

One of the most successful techniques that enables

one to control the ﬁlm deposition at an atomic-layer

level, for high T

superconductors as well as for other

complex oxides, is reactive MBE (Eckstein and

Bozovic 1995, Schlom and Harris 1995). It is simi-

lar to reactive evaporation, except that (i) the pres-

sure during deposition must be low enough such that

the mean free path of the depositing species (atoms or

molecules) is longer than the source-to-substrate dis-

tance and (ii) the chamber walls must be cold enough

such that the metal atoms or molecules that collide

with them are not re-emitted. Under such conditions

(typically 10

5

torr or less), the supply of the desired

species, in whatever overlapping or nonoverlapping

order desired, can be precisely controlled.

The molecular beams are chopped by mechanical

shutters, and the timing is computer controlled. Ra-

diatively heated effusion cells are typically used to

provide molecular beams of the elemental constituents

of the materials to be grown (see Fig. 8). For atomic-

layer engineering, it is not enough to monitor relative

ratios; one needs an absolute measurement of the

atomic ﬂuxes. For this purpose, spectroscopic tech-

niques such as measurements of absorption or emis-

sion of visible or UV characteristic radiation from the

atomic beams have been used with success, usually in

addition to the more conventional QCMs.

Using MBE, a variety of high T

superconductor

heterostructures have been fabricated, including tri-

layer Josephson junctions (known also as planar or

sandwich junctions). The superconductive electrodes

are made of Bi

CaCu

, YBa

7d

(both

c- and a-axis oriented), or La

1.85

0.15

CuO

. For the

barrier layers, various materials have been used in-

cluding titanates and manganites, with the layer

thickness down to one unit cell. High T

supercon-

ductor layers of one unit cell thickness have also been

fabricated with T

equal to bulk samples, as well as

various superlattices, and even some novel, artiﬁcial

high T

superconducting compounds (Bozovic et al.

1994, Schlom and Harris 1995).

2.5 Other Techniques

Some of the high T

superconductors with the highest

contain highly volatile (and toxic) elements, such as

mercury or thallium. The techniques described above

Figure 8

Schematic of an MBE system for atomic-layer synthesis of oxide superconductors (reproduced by permission of

Annual Reviews Inc. from Annu. Rev. Mater. Sci., 1995, 25, 679–709).

1184

Superconducting Thin Films: Materials, Preparation, and Properties

are not well suited for handling such elements. For this

reason, special approaches have been adopted: either a

two-step process, where the precursor ﬁlm is deposited

ﬁrst, and then postannealed (ex situ) under a high

vapor pressure of thallium or mercury, or a single-step

(in situ) process where sputtering of the precursor and

high-rate thermal evaporation of thallium are applied

simultaneously (Tsuei et al. 1994, Myers et al. 1994).

Several techniques have allowed ﬁlms with excel-

lent in-plane texture to be realized on polycrystalline

substrates of practical importance. One method

makes use of ion beam-assisted deposition (IBAD)

to create an in-plane oriented buffer layer of yttria-

stabilized cubic zirconia (Iijima et al. 1992) or MgO

(Wang et al. 1997) on the underlying substrate. Films

of high T

superconductors are then epitaxially

grown on top of this buffer layer. Another method,

which has achieved comparable in-plane texture to

IBAD, uses rolling and annealing to achieve texturing

in metal strips, e.g., nickel, that are then used as

substrates for ﬁlm deposition (Goyal et al. 1996,

Norton et al. 1996). The freedom from single-crystal

substrates that these techniques bring offers ﬂexibility

for the integration of high T

superconductor ﬁlms

with other materials and forms the basis of a growing

coated-conductor technology (see Superconducting

Wires and Cables: Low-ﬁeld Applications).

Other techniques, such as metal-organic chemical

vapor deposition (MOCVD), etc., have also been ap-

plied with success to grow high T

superconductor

thin ﬁlms, but they are less widespread.

3. Summary and Outlook

As discussed in this article, ﬁlms and multilayers

containing a great number of high T

compounds are

now grown routinely and have exceptional properties

(see High-temperature Superconductors: Thin Films

and Multilayers)—quite an accomplishment, given

that high T

superconductivity was discovered only in

1986. Multilayers have progressed from combining

known materials to literally engineering new materi-

als at the atomic-layer level.

The progress achieved and the knowledge acquired

will be applied in the future to the growth of ﬁlms and

heterostructures of other oxide or nitride compounds,

into which high T

superconducting ﬁlms will be em-

bedded. As in semiconductors, through doping and

material selection, the functionality of individual lay-

ers can be tailored in heterostructures. It is expected

that the challenges and the potential beneﬁts of such

heterostructures will stimulate further improvements

in the control of ﬁlm growth.

Acknowledgments

We gratefully acknowledge our many collaborators

with whom, beginning in early 1987, we have worked

on high T

superconducting thin ﬁlms and multilayers,

especially D. Anselmetti, M. R. Beasley, J. G. Bed-

norz, A. Catana, K. Char, Ch. Gerber, J. N. Eckstein,

C.-E. Eom, T. H. Geballe, R. H. Hammond, J. S.

Harris Jr., E. S. Hellman, A. Kapitulnik, M. E.

Klausmeier-Brown, J. Mannhart, and G. F. Virshup.

We also thank D. H. A. Blank, C. B. Eom, and H.

Kinder for providing us with original ﬁgures of their

work for use in this article. DGS acknowledges the

support of DOE through grant DE-FG02-97ER45638

and a research fellowship from the Alexander von

Humboldt Foundation during his sabbatical stay at

Augsburg University.

See also: Films and Multilayers: Conventional

Superconducting; Superconducting Materials: Types

of; Superconducting Thin Films: Multilayers

Bibliography

Bozovic I, Eckstein J N, Virshup G F, Chaiken A, Wall M,

Howell R, Fluss M 1994 Atomic-layer engineering of cuprate

superconductors. J. Supercond. 7, 187–95

Cava R J 1990 Structural chemistry and the local charge picture

of copper oxide superconductors. Science 247, 656–62

Cava R J, Batlogg B, Krajewski J J, Farrow R, Rupp L W Jr.,

White A E, Short K, Peck W F, Kometani T 1988 Super-

conductivity near 30 K without copper: the Ba

0.6

0.4

BiO

perovskite. Nature 332, 814–6

Coppens P, Lee P, Gao Y, Sheu H -S 1991 Application of the

selective atom diffraction method to the cation distribution in

high T

bismuth cuprates. J. Phys. Chem. Solids 52, 1267–72

Dijkkamp D, Venkatesan T, Wu X D, Shaheen S A, Jisrawi N,

Min-Lee Y H, McLean W L, Croft M 1987 Preparation of

Y–Ba–Cu oxide superconductor thin ﬁlms using pulsed laser

evaporation from high T

bulk material. Appl. Phys. Lett. 51,

619–21

Eckstein J, Bozovic I 1995 High-temperature superconducting

multilayers and heterostructures grown by atomic layer-by-

layer molecular beam epitaxy. Annu. Rev. Mater. Sci. 25,

679–709

Eom C B, Sun J Z, Lairson B M, Streiffer S K, Marshall A F,

Yamamoto K, Anlage S M, Bravman J C, Geballe T H,

Laderman S S, Taber R 1990 Synthesis and properties of

YBa

thin ﬁlms grown in situ by 901 off-axis single

magnetron sputtering. Physica C 171, 354–82

Eom C B, Sun J Z, Yamamoto K, Marshall A F, Luther K E,

Geballe T H, Laderman S S 1989 In situ grown

YBa

7d

thin ﬁlms from single-target magnetron sput-

tering. Appl. Phys. Lett. 55, 595–7

Fisk Z, Sarrao J L 1997 The new generation high-temperature

superconductors. Annu. Rev. Mater. Sci. 27, 35–67

Goyal A, Norton D P, Budai J D, Paranthaman M, Specht E

D, Kroeger D M, Christen D K, He Q, Safﬁan B, List F A,

Lee D F, Martin P M, Klabunde C E, Hartﬁeld E, Sikka V K

1996 High critical current density superconducting tapes by

epitaxial deposition of YBa

thick ﬁlms on biaxially

textured metals. Appl. Phys. Lett. 69, 1795–7

Hammond R H, Bormann R 1989 Correlation between the in

situ growth conditions of YBCO thin ﬁlms and the thermo-

dynamic stability criteria. Physica C 162–164, 703–4

Humphreys R G, Satchell J S, Chew N G, Edwards J A,

Goodyear S W, Blenkinsop S E, Dosser O D, Cullis A G

1185

Superconducting Thin Films: Materials, Preparation, and Properties

1990 Physical vapour deposition techniques for the growth of

YBa

7d

thin ﬁlms. Supercond. Sci. Technol. 3, 38–52

Iijima Y, Tanabe N, Kohno O, Ikeno Y 1992 In-plane aligned

YBa

7–x

thin ﬁlms deposited on polycrystalline metallic

substrates. Appl. Phys. Lett. 60, 769–71

Jakob G, Hahn T, Kluge T, Wagner P, Adrian H 1994 Char-

acterization of YBa

/PrBa

Superlattices. J.

Supercond. 7, 197–200

Jo W, Peng L S-J, Wang W, Ohnishi T, Marshall A F, Ham-

mond R H, Beasley M R, Peterson E J 2001 Thermodynamic

stability and kinetics of Y–Ba–Cu–O ﬁlm growth at high rates

in atomic and molecular oxygen. J. Cryst. Growth 225, 183–9

Kinder H, Berberich P, Prusseit W, Semerad R, Utz B 1996

Very large area YBa

ﬁlm deposition. In: Bozovic I,

Pavuna D (eds.) Oxide Superconductor Physics and Nano-

Engineering II, SPIE Proc. 2697. SPIE, Belingham, pp. 154–9

Koinuma H, Yoshimoto M 1994 Controlled formation of oxide

materials by laser molecular beam epitaxy. Appl. Surf. Sci.

75, 308–19

Li Z, Riﬁ H, Vaures A, Megtert S, Raffy H 1994 Synthesis of

CaCu

/Bi

CuO

superlattices with a T

enhance-

ment effect. Phys. Rev. Lett. 72, 4033–6

Lindemer T B, Washburn F A, MacDougall C S, Feenstra R,

Cavin O B 1991 Decomposition of YBa

7d

and

YBa

for p

p0.1 MPa. Physica C 178, 93–104

Mannhart J 1992 Changes in the superconducting properties of

high-T

cuprates produced by applied electric ﬁelds. Mod.

Phys. Lett. B 6, 555–71

Matijasevic V, Bozovic I 1996 Thin ﬁlm processes for high-

temperature superconductors. Current Opinion Solid State

Mater. Sci. 1, 47–53

Myers K E, Face D W, Kountz D J, Nestlerode J P 1994 Su-

perconducting and epitaxial (Tl,Pb)Sr

1x

thin

ﬁlms by in situ deposition. Appl. Phys. Lett. 65, 490–2

Nonaka H, Ichimura S, Shimizu T, Arai K 1995 Development

of oxidation sources in preparation of high-T

oxide super-

conductor thin ﬁlms using the molecular beam epitaxy meth-

od. Crit. Rev. Solid State Mater. Sci. 20, 285–338

Norton D P, Goyal A, Budai J D, Christen D K, Kroeger D M,

Specht E D, He Q, Safﬁan B, Paranthaman M, Klabunde C E,

Lee D F, Sales B C, List F A 1996 Epitaxial YBa

biaxially textured nickel (001): an approach to superconducting

tapes with high critical current density. Science 274, 755–7

cker N, Fink J, Fuggle J C, Durham P J, Temmerman W M

1988 Evidence for holes on oxygen sites in the high-T

su-

perconductors La

2x

CuO

and YBa

7y

. Phys. Rev.

B 37, 5158–63

Petricek V, Gao Y, Lee P, Coppens P 1990 X-ray analysis of the

incommensurate modulation in the 2:2:1:2 Bi–Sr–Ca–Cu–O

superconductor including the oxygen atoms. Phys. Rev. B 42,

387–92

Poppe U, Schubert J, Arons R R, Evers W, Freiburg C H,

Reichert W, Schmidt K, Sybertz W, Urban K 1988 Direct

production of crystalline superconducting thin ﬁlms of

YBa

by high-pressure oxygen sputtering. Solid State

Commun. 66, 661–5

Rao C N R 1991 The Blackett memorial lecture, 1991: chemical

insights into high-temperature superconductors. Phil. Trans.

R. Soc. London, Ser. A 336, 595–624

Rijnders G J H M, Koster G J, Blank D H A, Rogalla H 1997

In situ monitoring during pulsed laser deposition of complex

oxides using reﬂection high energy electron diffraction under

high oxygen pressure. Appl. Phys. Lett. 70, 1888–90

Schlom D G, Harris J S Jr 1995 MBE growth of high T

superconductors. In: Farrow R F C (ed.) Molecular Beam

Epitaxy: Applications to Key Materials. Noyes, Park Ridge,

NJ, pp. 505–622

Shafer M W, Penney T 1990 Hole, T

relationships in supercon-

ducting oxides. Eur. J. Solid State Inorg. Chem. 27,191–220

Shi D, Capone D W II 1988 60–90 K superconducting phase

transition in YBa

7d

. Appl. Phys. Lett. 53, 159–61

Sleight A W 1991 Synthesis of oxide superconductors. Physics

Today 44 (6), 24–30

Sleight A W, Gillson J L, Bierstedt P E 1975 High-temperature

superconductivity in the BaPb

1x

BixO

system. Solid State

Commun. 17, 27–8

Suzuki M, Murakami T 1987 Hall effect in superconducting

(La

1x

)

CuO

single crystal thin ﬁlms. Jpn. J. Appl. Phys.,

Part 2. 26, L524–5

Tabata H, Kawai T, Kanai M, Murata O, Kawai S 1989 Tai-

lored thin ﬁlms of superconducting Bi–Sr–Ca–Cu–O oxide

prepared by excimer laser ablation technique. Jpn. J. Appl.

Phys., Part 2. 28, L823–6

Tokura Y, Arima T 1990 New classiﬁcation method for layered

copper oxide compounds and its application to design of new

high T

superconductors. Jpn. J. Appl. Phys., Part 1. 29,

2388–402

Tokura Y, Takagi H, Uchida S 1989 A superconducting copper

oxide compound with electrons as the charge carriers. Nature

337, 345–7

Tokura Y, Torrance J B, Huang T C, Nazzal A I 1988 Broader

perspective on the high-temperature superconducting YBa

system: the real role of the oxygen content. Phys. Rev.

B 38, 7156–9

Torrance J B, Tokura Y, Nazzal A I, Bezinge A, Huang T C,

Parkin S S P 1988 Anomalous disappearance of high-T

su-

perconductivity at high hole concentration in metallic

2x

CuO

. Phys. Rev. Lett. 61, 1127–30

Triscone J-M, Fischer Ø, Brunner O, Antognazza L, Kent A D,

Karkut M G 1990 YBa

/PrBa

superlattices:

properties of ultrathin superconducting layers separated by

insulating layers. Phys. Rev. Lett. 64, 804–7

Tsuei C C, Gupta A, Trafas G, Mitzi D 1994 Superconducting

mercury-based cuprate ﬁlms with a zero-resistance transition

temperature of 124 kelvin. Science 263, 1259–61

Wagner P, Hilmer F, Frey U, Adrian H, Steinborn T, Ranno L,

Elschner A, Heyvaert I, Bruynseraede Y 1993 Prepara-

tion and structural characterization of thin epitaxial

CaCu

8 þd

ﬁlms with T

in the 90 K range. Physica

C 215, 123–31

Wang C P, Do K B, Beasley M R, Geballe T H, Hammond R H

1997 Deposition of in-plane textured MgO on amorphous

substrates by ion beam-assisted deposition and com-

parisons with ion beam-assisted deposited yttria-stabilized

zirconia. Appl. Phys. Lett. 71, 2955–7

rdenweber R 1999 Growth of high-T

thin ﬁlms. Supercond.

Sci. Technol. 12, R86–102

Xi X X, Linker G, Meyer O, Nold E, Obst B, Ratzel F, Smithey

R, Strehlau B, Weschenfelder F, Geerk J 1989 Supercon-

ducting and structural properties of YBaCuO thin ﬁlms de-

posited by inverted cylindrical magnetron sputtering. Z.

Phys. B 74, 13–9

I. Bozovic

OXXEL GmbH, Bremen, Germany

D. G. Schlom

Penn State University, University Park

Pennsylvania, USA

1186

Superconducting Thin Films: Materials, Preparation, and Properties

Superconducting Thin Films: Multilayers

Over the years superconducting multilayer technolo-

gy has been developed either for scientific curiosity or

for technological applications. One of the simplest

multilayer structures is the superlattice where super-

conducting layers and nonsuperconducting layers al-

ternate either in a periodic order or in a nonperiodic

special order. The nonsuperconducting layers are

usually normal materials that can allow some cou-

pling between the superconducting layers. Without

any further processing, these superlattices have been

used for investigation of their superconducting prop-

erties. The main scientific interest has been the inter-

action between the superconducting layers as a

function of their coupling strength. Most studies

were on artiﬁcially made two-dimensional supercon-

ducting structures before the discovery of high-tem-

perature superconductivity (HTS) in 1987.

Some of the investigations of the superlattices have

led to important technological discoveries, one of

which being Nb/Al:Al

/Nb tunnel junctions. While

the tunneling phenomena were being studied in Nb/Al

superlattices, the idea of using the proximity effect

between niobium and aluminum was considered and

theexcellentbarrierpropertiesofAl

were discov-

ered. This tunnel junction technology has been very

important in low T

superconducting (LTS) electronics

since the early 1980s.

While most low T

superconductors are based on

conventional metals, the discovery of HTS has brought

on a completely new class of materials, namely the

metal oxides in complex perovskite structures. These

new oxide materials require new deposition and

processing techniques.

This article describes the deposition and processing

technology of superconducting multilayers and their

applications for the case of high T

superconducting

materials. The LTS case will be briefly mentioned for

comparison.

1. Deposition Technology

1.1 LTS Multilayers

When lead was used before niobium was introduced

to LTS technology, the preferred thin-ﬁlm deposition

technique was evaporation, because the soft lead was

almost impossible to sputter. However, when the

more refractive niobium material is used, sputtering

techniques are possible, although the evaporation

technique continues to be used, especially for careful

epitaxial growth of niobium thin ﬁlms. The sputtering

technique for niobium is found to be very reliable

once ultrahigh vacuum and stress-free deposition

conditions are obtained.

The majority of niobium growth is for polycrystal-

line thin ﬁlms and the adjacent nonsuperconducting

layers are also polycrystalline. For many technolog-

ical applications SiO

is used as the insulating layer

between the niobium wiring layers, and chemical

vapor deposition techniques are often used to take

advantage of the developments made in the semicon-

ductor industry. LTS technology based on metals is

very similar to that of metallization in the semicon-

ductor industry.

The sputtering, evaporation, and chemical vapor

deposition techniques mentioned above are standard

methods used throughout the electronics industry

(see High-temperature Superconductors: Thin Films

and Multilayers).

1.2 High T

Superconducting Multilayers

High T

superconductors, discovered in 1987, turned

out to be very complex metal oxides that had to be

grown in high-oxygen environments (greater than

1 mtorr) and at high temperatures (greater than

700 1C). Laser ablation, or pulsed laser deposition

(PLD), is a newer technique that has gained much

popularity because it is ideally suited for deposition

at a high oxygen pressure. The relative ease of this

technique in depositing multicomponent oxides

makes it especially effective for exploring new mate-

rials for HTS electronics, such as epitaxial dielectrics

or barrier layers. A short-wavelength (170–260 nm)

excimer laser is focused onto a rotating target of the

material to be deposited. At the energy of the laser

beam (0.1–2 J per pulse) the matter emitted from the

target forms a plume that carries it to the substrate at

supersonic velocities. In general, a higher gas pressure

is required during laser ablation owing to the very

high energy of the vaporized material from the sur-

face of the target. The laser plume glows brightly

from the target and deposition is usually done near

the end of the plume, about 5 cm away. As in the case

of sputtering, the various species scatter differently

and therefore the ‘‘sweet spot’’ of the deposition

process is relatively small, usually 1–2 cm, depending

on the geometry. The deposition rate per laser pulse

ranges from a fraction of an angstrom to a few ang-

stroms. The technique is fairly reproducible and has

been used extensively for research and development

purposes. Although other techniques are potentially

more manufacture-friendly, laser ablation is very ac-

tively used for prototyping multilayer devices made

up of several complex materials, such as supercon-

ductors, dielectrics, ferroelectric oxides, magnetic ox-

ides, and other conducting oxides.

A particular problem associated with PLD is the

deposition of micrometer-sized droplets, so-called

‘‘boulders,’’ on the grown ﬁlm. These particles orig-

inate at the target and are emitted from the action of

the laser pulse. A variety of procedures have been

used to reduce this problem so that the boulder den-

sity can be very low in the best ﬁlms. Such procedures

1187

Superconducting Thin Films: Multilayers

include target preparation, such as frequent polish-

ing, defocusing the laser spot, mechanically chopping

the plume, or spatially ﬁltering the beam. The quality

of multilayers made by PLD depends on being able to

reduce these boulder problems.

Although PLD can produce relatively high depo-

sition rates (up to tens of nanometers per second), the

area on which one deposits is relatively small. A

straightforward way to increase the deposition area is

to scan wafers over the plume either by moving the

substrate vertically and horizontally or by rotating

the substrate. Another scheme uses a rotating cylin-

drical target with a linear laser proﬁle to obtain larger

deposition areas.

Another major difﬁculty of these techniques is the

uniform heating of large wafers. Most of the heating

methods for small-area laser ablation mount a subst-

rate on a heated metal surface with silver paste, which

is difﬁcult to extend to larger sizes. In order to over-

come this difﬁculty of heating a large wafer, an off-

axis laser ablation technique has been developed. A

large wafer (2–3 in (5–8 cm)) can be mounted parallel

to the direction of the plume inside a relatively simple

black-body-like heater. The deposition takes place

when the atoms collide with the background pressure

and are scattered to the surface of the wafer. By ro-

tating the wafer and selecting an appropriate pressure

for the geometry, a fairly uniform deposition can be

achieved over 2 in (5 cm) wafers. This technique al-

lows for simultaneous deposition on both surfaces of

the wafer, which is an important beneﬁt for micro-

wave applications requiring a ground plane. A draw-

back of this technique is its low deposition rate

because of the off-axis geometry. Typical conditions

for 2 in (5 cm) wafer deposition results in a deposition

rate lower by about a factor of 10 than that for on-

axis deposition. However, the added advantage of

off-axis laser ablation is the reduction of the boulder

density, especially important for multilayer work.

By using a laser ﬂuence just enough to evaporate a

few atomic layers of the target in a low-oxygen en-

vironment (less than 10

4

torr) and at the same time

by using a sequence of metal or metal oxide targets,

one can obtain a process similar to molecular beam

epitaxy (MBE) by evaporation, here called laser-

MBE. Some in situ diagnostic tools can then be used

to characterize the growth of the materials. This

technique has mainly been used to grow artiﬁcially

layered superconducting materials, such as the inﬁ-

nite-layered superconductor.

2. Processing Technology

2.1 LTS Multilayers

LTS multilayer technology has made significant

progress since the 1980s, mainly in Japan, based on

niobium thin ﬁlms primarily for SQUID and high-

speed digital circuit applications. The main challenge

has been to deﬁne the junction structure, the Nb/

Al:Al

/Nb trilayer, without degrading the uni-

formity of its properties. A new process of deﬁning

the junction area by anodization rather than etching

has been invented and is a very reliable process for

minimizing the damage along the edges of the trilayer

structure.

A reactive ion etching process of niobium uses a

ﬂuorine-based gas such as CF

. By using dielectrics

such as SiO

, multilayer structures including a ground

plane, a trilayer structure, and wiring layers have been

fabricated.

2.2 High T

Superconducting Multilayers

Unlike polycrystalline niobium, HTS multilayer struc-

tures need to be epitaxial mainly because of a large

anisotropy and the weak grain boundaries of HTS

materials. Epitaxial growth requires a high-tempera-

ture deposition of each material and this complicates

the heating methods. No large-area (41 in (2.5 cm))

HTS multilayer process is available, although fairly

complicated structures including a ground plane

and edge junctions have been successfully fabricated

(see High-temperature Superconductors: Thin Films

and Multilayers; Films and Multilayers: Conventional

Superconducting).

Of several HTS materials, YBa

7–x

(YBCO)

is the most frequently used material for multilayer

work. This is because the anisotropy of YBCO is the

smallest of all HTS materials and YBCO is least

susceptible to the damage caused by etching. It is

inevitable that a patterned multilayer structure will

require a current path perpendicular to the substrate.

A weak current-handling capability in this direction,

namely a large anisotropy, will cause a problem in

such a multilayer structure. In order to minimize the

sudden current ﬂow in the perpendicular direction,

edges have to be patterned in such a way that the

edge angle is very small (o201), although this will

make the edge area wider.

Unfortunately, no good reactive etching method

for these complex metal oxides has been found, and a

dry ion beam milling (usually by an argon ion beam)

technique has been the primary method for etching.

Several wet etching methods have been found but

they have not been very popular because of severe

undercut below the photoresist during the etching

process. During dry ion beam etching, the YBCO

material is cooled by a thermal contact with a water-

cooled plate in order not to lose oxygen when heated

in vacuum. A line width as small as 1 mm has been

found to carry high critical current density by this

method.

YBCO is a semistable material at room tempera-

ture, and when it is bombarded with high-energy

particles, such as argon ions, it starts to lose oxygen

and then the materials becomes amorphous. This has

1188

Superconducting Thin Films: Multilayers

been observed during transmission electron micro-

scope sample preparation. However, the lost oxygen

can be restored by annealing the material in an ox-

ygen ambient at about 400–500 1C where the oxygen

intake is maximum. One can also restore the ion

beam-damaged crystallinity by annealing the material

around 800 1C in the appropriate oxygen environ-

ment (usually around 100 mtorr). Especially for mul-

tilayer work in which one relies on the epitaxial

growth of the subsequent layer on top of the edges,

the restoration of the crystalline quality after etching

is a very crucial step.

YBCO loses oxygen when heated but regains it

during cooling in an oxygen environment at 400–

500 1C. This means that if one grows a second epi-

taxial layer, through which the oxygen diffusion is

very slow at 400–500 1C, on top of the ﬁrst YBCO

layer, the resulting structure will have an oxygen-

deﬁcient YBCO layer. This is because the oxygen

cannot diffuse through the dielectric layer to get to the

bottom YBCO layer. This problem gets worse as one

develops a more perfect dielectric layer because de-

fects such as grain boundaries or oxygen vacancies in

the dielectric layer are often the channels for oxygen

diffusion.

The most frequently used dielectric layers are

SrTiO

, LaAlO

,Sr

AlTaO

,orSr

AlNbO

. These

all have a perovskite structure, similar to YBCO, al-

lowing epitaxial growth adjacent to YBCO. Of these

dielectric materials, SrTiO

is the easiest to deposit

and the ﬁrst that was developed. However, the high

dielectric loss of SrTiO

does not allow its use for

high-speed circuits. Other similar dielectric materials

with low dielectric loss, such as LaAlO

,Sr

AlTaO

and Sr

AlNbO

, were later developed. However, all

these dielectric materials have low oxygen diffusion

rate problems especially when a better material with

fewer defects is grown.

Materials that have exactly the same structure as

YBCO but are not superconducting have been used

in multilayer structures. These include PrBa

7–x

and doped YBCO. They have the advantage of hav-

ing an identically layered structure, which allows for

perfectly matched growth from the edges. In addi-

tion, they have an almost identical oxygen diffusion

rate, resulting in a fully oxygenated YBCO layer un-

derneath. However, their insulating properties are far

from ideal, making them impossible for use as di-

electric layers. For this reason, these materials have

been used mainly for barriers for Josephson junctions

(see Josephson Junctions: High-T

Because of its anisotropy, a Josephson junction

process for YBCO has been developed in an edge

junction geometry. The process requires forming a

low-angle slope for the ﬁrst YBCO layer and then

depositing the barrier layer and the second YBCO

layer on top of it. As mentioned above, restoration of

the crystallinity after etching is critical in allowing for

good epitaxial growth of the subsequent barrier and

YBCO layers. A technique to modify the edge and

convert the interface region into an insulating phase

by annealing them in a different environment has

been developed.

It remains to be seen how much further HTS mul-

tilayer technology will be developed. The answer to

this may lie in the prospects for its application in the

future.

3. Applications

3.1 LTS Multilayers

LTS multilayer technology has been used for devel-

oping SQUID sensors. Multilayer ﬂux transformers

or gradiometers coupled with trilayer tunnel junc-

tions have been successfully developed and are used

in practical SQUID instruments. In the case of high-

speed electronic circuits, both the voltage state logic

and the single-ﬂux quantum logic require multilayer

technology and niobium mutilayer technology is be-

ing used to demonstrate many high-speed circuits (see

SQUIDs: The Instrument).

3.2 High T

Superconducting Multilayers

Unfortunately, the use of HTS technology for

SQUID applications is by its single-layer process.

Multilayered ﬂux transformers for SQUID applica-

tions have been demonstrated, but the extra low-fre-

quency noise does not improve its performance over

that of single-layered ﬂux transformers.

For microwave applications, double layers of HTS

materials on both sides of a wafer are used. One of

these layers is the ground plane in a microstrip geo-

metry. It is very likely that many microwave

applications of this multilayer technology will be

developed with the proliferation of wireless commu-

nication. It remains to be seen how this will affect

HTS multilayer technology.

A few very simple multilayer HTS digital circuits of

the single-ﬂux quantum logic kind have been demon-

strated in spite of the many problems described above

(see Superconductors: Rapid Single Flux Quantum

(RSFQ) Logic).

See also: Superconducting Microwave Applications:

Filters; Superconducting Thin Films: Materials,

Preparation and Properties

Bibliography

Shinjo T, Takada T (eds.) 1987 Metallic Superlattices: Artiﬁ-

cially Structured Materials. Elsevier, Amsterdam

K. Char

Seoul National University, South Korea

1189

Superconducting Thin Films: Multilayers