Cost Projections for High Temperature Superconductors

Paul M. Grant and Thomas P. Sheahen

EPRI, Palo Alto, CA 94304 and SAIC, Gaithersburg, MD 20878

Abstract --- It is generally argued that for high-temperature

superconductors (HTS) to be cost-competitive in power applications, the wire will have to sell for about $10 per kiloamperemeter ($10/kAm) for operation at 77 K (e.g., NbTi costs
around $1/kAm and Nb3Sn around $8, each at 4.2 K). Given
what is already known about the critical current performance of
Pb-stabilized Bi-2223 (BSCCO), this cost target may be extremely difficult to realistically achieve for silver-sheathed
BSCCO produced by the oxide-powder-in-tube (OPIT) technique. In this paper, we examine the cost of component materials, add reasonable estimates for labor and related costs, and
arrive at a likely cost/performance (C/P) figure. We also estimate the capital cost of a factory to produce HTS conductor by
a particular coated conductor method, and calculate the necessary production-output and performance parameters necessary
to manufacture 10 km/yr of wire and its associated C/P. Our
results indicate that the real C/P seen by the customer will remain substantially above this $10/kAm target for some time to
come.

I. INTRODUCTION
For over a decade, the emphasis of applied research on
high-temperature superconductivity (HTS) has been placed
on improving the properties of the materials, especially the
critical current JC. Until recently [1], little attention has been
given to considerations of manufacturing cost, because other
concerns and obstacles have always been of much more immediate concern. There seems to be a basic presumption that
if researchers can find a way to do something at all, then others will find a way to drive down manufacturing costs and
make a profit from it.
For power applications [2] of HTS, where many amperes
are to be transported over sometimes meters, sometimes
miles, there exists competition with traditional lowtemperature superconductors (LTS) such as NbTi and Nb3Sn,
because the latter can perform the exact same functions
whenever the temperature can be reduced into the liquid helium range (4 K). The very high cost of such refrigeration
has kept many superconducting power applications off the
market for years, despite their demonstrated technical feasibility [3]. The attitude of utilities and power equipment
manufacturers has been one of reluctance to accept superconducting technology because of the prohibitive costs (from
their point of view) in light of performance gained.
For electrical wire, the figure-of-merit for comparing costs
of different materials at a particular operating point has been
dollars per kiloamperemeter ($/kAm). This reflects the
Paper presented at the Applied Superconductivity Conference, Palm Desert,
CA, September, 1998.

twin purposes of a wire, namely, to carry high currents over

great distances. Notice that it is current, not current density,
that matters here; from a practical point of view, any conductor carries an overhead burden of sheathing, insulation, or
(in the case of conductors coated onto a nonconducting substrate) the thickness of extra material. For aluminum and
copper wire, the distinction is a few percent; for superconductors, the distinction is at least a factor of two, and often
far greater, reaching up to 100 or more in adverse cases. The
usual critical current density, JC, must thus be set aside in
favor of the engineering critical current, JE, where all these
burdens are included. The $/kAm number is what the user
has to design to and pay for whatever kind of conductor that
takes electricity from one place to another.
In addition to the capital cost of buying wire, the user has
to spend money for installation, maintenance, repair, and the
continuing cost of cooling the wire to its operating temperature. Any reduction in the refrigeration expense due to running at a higher temperature is of value to the customer, and
this can result in an increase of the acceptable capital (manufacturing) cost of HTS wire.
Given what is known today about the comparative tradeoffs between refrigeration and wire manufacturing costs, and
recognizing that Nb3Sn costs $8/kAm, and NbTi $1/kAm,
it is generally thought throughout the HTS community [4]
that it will be necessary for HTS conductor to sell for about
$10/kAm in order to have substantial penetration of power
application markets. Of course, the exact price sensitivity
varies from one application to another, and there are wide
error brackets, but the general feeling is that anything greater
than $50/kAm will satisfy only niche requirements.
The remainder of this paper is devoted to presenting an
analysis of how well the cost/performance (C/P) target of
$10/kAm is likely to be met by BSCCO (Pb-stabilized
Bi2Sr2Ca2Cu3O10, or Bi-2223) wire tapes made by the OPIT
(oxide-powder-in-tube) method, and by YBCO (YBa2Cu3O7,
or Y-123) coated conductors. We anticipate many of our
conclusions will be controversial and perhaps some even in
error. This is to be expected given that much critical data is
presently being held close by the various companies and institutes involved for quite understandable propriety and competitive reasons. Nevertheless, we believe it is now time to
commence discussion of cost/performance challenges that
confront both process developers, manufacturers and end
users of high temperature superconducting wire.

II. BSCCO WIRE

To put the HTS target number in perspective, we employ
the Sokolowski Plot shown in Fig. 1 [5]. This conveniently displays operating current vs. cost for various
combinations of material, temperature and magnetic field.
The plot is also divided into regions of constant C/P in
$/kAm as indicated by the various diagonal lines. Note
only Nb3Sn and NbTi lie to the right of the $10/kAm line,
and that several of the reported values are manufacturer's
targets, not achieved results. Presently, HTS tapes available
on the "open market" are quoted between $800-1200/kAm.
10000

$10 Bi-2223

1000

$1 /kAm

77K, 0T
Siemens
Target

$1,000

NbTi
4.2K, 2T

7.0

Nb3Sn
4.2K, 12-15 T

100
EURUS

100

Y-123 IBAD
Bi-2223
77K, 0T
NKT Target

Bi-2223
77K, 0T
IGC

10

1
1000

Bi-2212
4.2K, 0T
IGC

$/troy oz

Operating Current, Ic (A)

$100

For numerical simplicity, let's say the HTS material in a

given BSCCO tape has a critical current density JC = 2 104
A/cm2 [8]. Thus 200 amperes can be carried over one mm2.
The silver surrounding that much BSCCO is roughly 2 mm2,
and a segment one meter long therefore contains 2 cm3,
which weighs 21 grams. At an average of $5/troy-ounce, one
meter of silver sheath costs $3.38 yielding a C/P for 200 A of
$16.88/kAm. Depending on volume ordered or internally
produced, the cost of the superconductor material runs between $0.34-1.37/cm3 at stoichiometric density for Bi-2223
[9]. We then need to add to silver a C/P for the HTS material
on average of $4.28/kAm for a rounded-down total of
$21/kAm. This leaves the C/P over 100% above target for
materials alone, with no indirect manufacturing costs (labor,
capital, etc.) yet included.

6.5
6.0
5.5
5.0

10

0.1

Cost ($/m)
Fig. 1. The "Sokolowski Plot" of a number of HTS and LTS wire embodiments. Data were obtained courtesy of the manufacturers as labelled in the
plot. When comparing performance, note the different operating points, i.e.,
temperature and field. The diagonal lines represent demarcation of various
constant C/P values in $/kAm.

As mentioned at the Applied Superconductivity Conference, ASC-96, two years ago [6], the cost of silver sheathing
for BSCCO carries a very severe price penalty. The calculation is straightforward, and might readily be left as an exercise for the reader; but nonetheless we will present our version now.
Assume a typical powder-in-tube configuration has a silver
tube 2 mm in diameter, with an inside diameter of 1.15 mm
filled with BSCCO. The ratio of cross-sectional areas is such
that silver is 2/3 and BSCCO 1/3; so the volume fraction of
superconductor is = 1/3. As the tube is stretched and
thinned, that ratio doesnt change. When a bundle of tubes is
combined to form a multifilament wire and crushed into a
tape, still the ratio doesnt change.
The silver tubes could perhaps have a thinner wall, but the
need for thermal protection [6] against burnout during
quench argues for having more than half the wire made of
non-superconducting material [7]. It would be dangerous to
run above = 0.5. In most previous applications, such as
NbTi for accelerator magnets, = 0.4 is common. For any
wire with adequate thermal protection, the silver cost component will certainly be nontrivial. Figure 2 gives an idea of the
variation in silver futures for the early part of 1998.

4.5
8/97

9/97

10/97

11/97

12/97

1/98

2/98

3/98

4/98

Month
Fig. 2. Variation in May 1998 silver futures since August 1997. The peak in
February displays the "Warren Buffet Effect" (the large volume silver purchases that month by Berkshire Hathaway) the and is suggestive of the volatility of silver prices against moves by large investment groups.

What are the future prospects for silver supply and prices
[10]? In 1997, total worldwide demand for silver was 27
million kilograms, outstripping supply by 6 million kg. Demand for silver has exceeded supply since 1990, not exactly a
good sign for a new technology that will make even greater
demand on reserves, in the ground, recycled and stored as
bullion. One of the largest consumers of silver is the photography industry (7.2 million kg/yr of which 1.4 million kg/yr
is recycled), and it is sometimes remarked that the replacement of silver halide-based photographic technology by digital methods will lower silver prices to the benefit of other
uses, e.g., power applications of superconductivity. Yet in
1997, photographic uses of silver increased 3%, a trend
which is expected by industry analysts to continue for some
time for amateur and some professional applications.
The annual production of NbTi by IGC is about 20,000
km/yr [11]. Annual installation of underground transmission
cable (three phase circuits) is around 360 km/yr in the US.
Assuming a form factor of 20 to acount for layering and
helicity, this would require 3,600 km/yr of HTS tape. Let's
say the total potential annual market for practical HTS tape is
twice this total number (i.e., complete takeover of all commercial applications of LTS and new HTS opportunities), or
nearly 50,000 km/yr in the US. For BSCCO/OPIT tape of

the form factor we have been assuming (3mm2 and = 1/3),

this scenario would consume about all the silver currently
recycled by the photographic industry, a situation likely to
adversely affect silver prices, assuming other demand remains constant (perhaps the wisest investment strategy for a
utility is to insert Ag-based HTS wire technology as rapidly
as possible, subsequent to salvage at a huge profit 40 years
from now after the discovery of room temperature superconductors!).
Returning now to additional factors affecting BSCCO wire
C/P, estimates for labor and overhead (L&O) range from $15/m [9], which transform into $5-25/kAm in our JC =
20,000 A/cm2 example above. Given this very wide bracket
(right now, the production lines are manned mainly by
PhDs!), let's split the difference in L&O which now raises the
overall C/P for OPIT BSCCO to $37/kAm [12].
Next, some real bad news. There are further considerations which must be taken into account when arriving at a
practical number for an "end use" C/P, which we will call
"derating factors." These factors have to be inserted because
it is the operating current, not the laboratory IC, which, from
the viewpoint of the end user, must comprise the normalization unit in C/P. There are at least four elements which contribute to lowering the real current that the wire can carry,
and two more, one related to strength and ac loss and the
other to marketing, which can raise the cost. All are multiplicative. First of all, the voltage drop criterion by which
JC is defined is commonly taken to be 1 V/cm for HTS
wires, but in real applications, it will be necessary to stay
about 20% (depending on the particular value of n in E ~ Jn,
presently about 10-15) below this "conventional JC" in order
to lower the voltage drop below 0.1 V/cm, the criterion for
establishing IC in LTS wire, otherwise resistive losses become untenable. Second, It is a wise precaution to assume
that most applications will take place in a magnetic field of at
least 0.1 tesla, which reduces JC by 15% (80% if the field is
aligned perpendicular to the ab plane). Third, for manufactured lengths over 1 km, a factor of about 3 may be lost, at
least as indicated by publicly available data (American Superconductor [13] maintains they can achieve only a 5%
derating over long lengths, but to the knowledge of the authors, supporting data has not been published). Fourth, it is
generally assumed, especially in the transmission/distibution
cable community, operation under ac conditions will lower
the current capacity by a factor of as much as 2.5. Fifth, we
must remember that metallurgically pure silver (not even
sterling!) is not used in the actual BSCCO/OPIT manufacturing process for reasons of strength and reduction of ac loss.
This additional alloying and/or processing does not come for
free. We speculate such "finishing touches" will add perhaps
50% or more to the "commodity price" of silver. Finally, wire
manufacturers presumably want to make some profit -- 30%
over basic manufacturing cost seems fair to us at this time.
These various derating factors, and their products, are
summarized in Table I.

Derating Factor

TABLE I
VARIOUS C/P DERATING FACTORS
This Paper
ASC [13]

1.0 0.1 V/cm

0.1 T Magnetic Field
Length 1 km
ac Operation
Ag Treatment
Profit

1.2
1.8
3.0
2.5
1.5
1.3

Cumulative Derating

31.6

1.15 (B ab plane)
1.05

7.1

The right hand column in Table I represents informal

communication from American Superconductor on what are
felt by them to be more appropriate numbers for field and
length derating. The rows left empty do not necessarily represent their agreement or disagreement with our estimates in
the middle column.
Thus the cumulative derating applied to our previous figure of $36/kAm now raises the true C/P to somewhere between approximately $230-1040/kAm. Admittedly, there is
room for manuever here. For example, not all applications
involve ac, and a considerable reduction in C/P may be in
reserve, enough perhaps to offset the cost of ancilliary
equipment for dc/ac conversion for some utility transmission
/distribution system applications. In addition, a factor of
about two improvement can be accomplished simply by reducing the operating temperature to around 66 K (77 K was
assumed the cryogenic operating point throughout our discussion so far). Nonetheless, the bottom line is that the
greatest opportunity for reducing C/P lies in significantly
increasing the fundamental JC. What are the prospects this
can be done?
Some idea can be obtained through a study of Fig. 3 which
shows the results of magnetoptically imaging the flux penetration of an externally applied magnetic to a BSCCO/OPIT
filament followed by an inversion of the Biot-Savart integral
equation to obtain the actual critical current paths and distributions therein.
Polarized Micrograph

Magneto-Optic Image

J C Paths @ 12 K

> 3.6 10 5 A/cm

> 2.6 10 5 A/cm
> 1.6 10 5 A/cm
=
> 0 A/cm 2

2
2
2

Fig. 3. Cross-sectional microscopic and magnetoptical (polar Kerr effect)

images (MOI) of BSCCO filaments in OPIT tape. The lowest picture shows

Slope = 9211 A/cm^2/yr

400000

350000

Jc (A/cm^2, 77 K, 0 T)

The truly startling consequence of Fig. 3 is that the overwhelming bulk of the filament (shown in blue) carries zero
critical current! Most of the critical current is preferentially
transported near the Ag/BSSCO interface, a result wellknown in the field [2, 15]. In this particular sample, the contiuous current path exceeded 160,000 A/cm2 (green and yellow), with a few small regions (red) reaching 360,000 A/cm2.
Can we use the example of Fig. 3 to estimate an "as good
as it gets" outcome for BSCCO/OPIT wire? The data were
taken at 12 K, but a reasonable way to extrapolate to 77 K
might be as follows: 1) the relative JC at all temperatures and
fields is distributed more or less as shown in Fig. 3, i.e., with
the maximum at the Ag/HTS interface; 2) the maximum
JC(77 K, 0 T) for epitaxial BSCCO films is roughly one million A/cm2; 3) one might assume it is possible to achieve
perhaps 1/3 this magnitude near the Ag surface in BSCCO
tape; and 4) let's just assume that would be 360,000 A/cm2,
the same value seen in Fig. 3 for 12 K (after all, the data of
Fig. 3 is now three years old, and recently JC values near this
figure have been observed by at least one institution [16]).
The "as good as it gets" scenario for BSCCO/OPIT would
then be played out if this 3.6 105 A/cm2 could be realized
throughout the entire cross-section of the filament.
How long might it take before such could be accomplished? To date, JC in meter-scale lengths has increased
linearly in time since the birth of BSCCO wire technology
around 1991a kind of linear "Moore's Law" which in fact
has been christened "Malozemoff's Law" after the ASC scientist who first observed this trend. The slope of JC vs. t is
about 9200 A/cm2/yr, and this allows, under an assumption
Malozemoff's Law will continue to hold, an estimation of
when "as good as it gets" for BSCCO is realized. This speculative scenario is exhibited in Fig. 4.

ASC Short Rolled Multifilament

(Bi,Pb)-2223/Ag

300000

250000

200000

150000

100000

50000

0
1990

1995

2000

2005

2010

2015

2020

2025

2030

2035

Year

Fig. 4. "Malozemoff's Law" showing the linear increase in JC between 1991

and the present. Assuming its continuing trend, maximum JC of 360,000
A/cm2 for BSCCO/OPIT will arrive near 2030.

We see that BSCCO/OPIT wire, subject to our assumptions and the continuance of Malozemoff's Law, will top out
around the year 2030quite a long time to wait. We suspect
JC vs t will deviate from linearity long before then. Which
way, of course, will prove crucial to its future application.
Some estimation of expected trends for the "real" C/P can
be made by combining our derating exercise above with the
extrapolation of Malozemoff's Law just discussed and displayed in Fig. 4. Figure 5 shows the result.
Derated C/P for BSCCO/OPIT
120.00

100.00

C/P ($/kA xm)

the lateral cross-section of a typical multifilamentary (~80 here)

BSCCO/OPIT tape such as manufactured by IGC and ASC. The upper three
stacked images are of a longitudinal "take out" of one such filament, the
topmost being a polarized micrograph (note 50 m scale bar) clearly showing
the granular nature of the filament. Next is the MOI of the flux penetration
of a small external magnetic field applied perpendicular to this same area
cooled to 12 K, followed by a map of the numerical inversion of the BiotSavart equation yielding critical current paths. Multifilamentary tape photo
courtesy of ASC and the rest of the figure is taken from the U. Wisconsin
Applied Superconductivity Center home page [14].

80.00

60.00

40.00

1998 ASC JC
$10.30/kAm

20.00

$10/kAm

0.00
1990

1995

2000

2005

2010

2015

2020

2025

2030

2035

Year

Fig. 5. Prediction of the decrease in BSCCO C/P per year employing the data
of Fig. 4 since about mid-1995, the geometrical, Ag-filling, materials, and
L&O numbers of our earlier example, all derated by a factor of 7.1. The
"call-out" displays the "unburdened" C/P associated with the record value of
JC = 70,500 A/cm2 announced by ASC in 1998 [17].

If we view $10/kAm as merely a "laboratory benchmark,"

one could plausibly claim it has already been reached (see the
balloon call-out in Fig. 5 pointing to the 1998 70,500 A/cm2
record announced by ASC earlier this year [17]). We see
that achievement of the generally accepted target price of
$10/kAm as a practical market objective will remain out of
reach for a very long time, perhaps many decades into the
future. Even if the ac loss burden (2.5) is removed by some
future large demand for dc cable, the $10 figure is unlikely

be reached before 2020and rememberup to now, our

analyses have not even included capital-cost-of-plant amortization. Therefore, it appears that the real cost of BSCCO
wire will remain far above the target price of $10/kAm
for a very long time, certainly many decades into the future.
III. YBCO COATED CONDUCTORS
The hope for this second generation of high temperature
superconductor wire relies on thin films of high JC performance YBCO deposited on inexpensive substrates. A very thin
veneer (a few microns) of silver may be needed on top of the
YBCO, but the thickness (and therefore cost) would be much
less than for BSCCO. The main advantage to YBCO coated
conductors is that the JC values in the YBCO film are hoped
to be ultimately around 106 A/cm2, a full order of magnitude
above the current BSCCO range.
As there is not presently any production of YBCO coated
conductors ongoing, not even in pilot line mode, it is thus
necessary to imagine a future manufacturing plant, using a
particular process, and then estimate the costs associated with
materials, labor, capital equipment, etc., in order to arrive at
the total expense of buiding, running and maintaining the
plant. For any given total output, it is then an easy calculation to arrive at the $/kAm cost of this conductor. Alternately, since this is an imaginary plant to begin with, it is
possible to hold the cost fixed at $10/kAm, and then ask
how much the total annual output must be -- in other words,
how fast must the production lines run to achieve break-even.
The latter approach has been taken in a report by Chapman
[18, 19], who looked at two potential YBCO manufacturing
process in considerable detail. Envisioning a hypothetical
factory where YBCO is vapor-deposited onto a nickel substrate following the ORNL RABiTS process, the various
components of the production cost were carefully estimated.
In that report, both pulsed laser deposition (PLD) and electron-beam (e-beam) methods were evaluated. Chapman
found that PLD turned out to be far too expensive, but the ebeam process might be viable.
To achieve the $10/kAm target cost, it was necessary to
produce 18,000,000 meters per year of coated conductor.
This is comparable the typical production, mentioned earlier,
of Intermagnetics General Corporation of 20,000 km/yr of
NbTi superconducting wire in their large scale production
facility in Waterbury, CT [11]. It is worth pointing out that
this fabrication rate for NbTi by IGC is the result of nearly 30
years of diligent engineering and manufacturing R&D.
It will be instructive to examine the Chapman Report scenarios in more detail, because they provide valuable insight
into the enormity of the scale-up problem that must be solved
if YBCO coated conductors are to become commercially viable. The 18,000,000 meter plant envisions 30 parallel 1 cm
tapes of nickel continuously moving across a 1/2 meter distance inside a chamber, all the while in a vapor of atomic Y,
Ba, and Cu. In addition, the chamber must operate in a par-

tial pressure of a few Torr of O2, unless oxygenation is done

in a subsequent process step.
The e-beam approach evaporates elemental atoms from
targets within in the chamber. To gain necessary production
capacity per year, Chapman assumes the rate of barium deposition in this hypothetical factory will be about 1 gram/sec,
with copper and yttrium rates yielding an additional net mass
transfer of 2 grams/sec. This is comparable to painting a
wall, and, incidentally, does not include material deposited
on the sides of the vacuum chamber and elsewhere. A typical
e-beam target would be exhausted within a minute, and even
very massive targets weighing several kilograms would need
to be renewed roughly every hour, thus a continuous feed of
stock material through vacuum-tight thrust bushings to the
hearths would be required (this may not be as implausible as
it sounds, as some methods of actinide isotope separation and
enrichment for nuclear weapons and power reactor fuel are
actually carried out this way). Both sides of the nickel tape
are to receive a net deposition of 2 microns thickness during
the 1/2 meter traverse of the chamber. That corresponds to a
deposition rate of 1168 Angstroms/second. Low grain-angle
boundary growth upon the substrate must be maintained for
that full thickness.
As mentioned previously, Chapman's approach has been to
fix the target price and scale the factory throughput to meet it.
If instead one reverses this thinking and imagines an ebeam based factory capitalized at perhaps $33 million, producing a more modest throughput of 10 km/year, it is possible to then derive a $/kAm figure as follows: Using an opportunity cost of capital [20] of 18% (typical today), the
plant costs $6,000,000 annually even if nothing goes out the
door. Adding reasonable estimates for labor and other operating costs (on a per-year basis, not on a per-meter basis as in
section 2 above), as well as materials, brings the annual expense up to $7,500,000. With an output of 10 km/yr, we then
have an average cost of $750/meter. If the conductor carries
perhaps 400 Amps, again a typical number for coated conductors of the geometry commonly discussed, we arrive at
$1,875 /kAm.
Each of these numbers can be massaged somewhat: lower
capital cost, higher labor cost, more precise materials estimates, etc. No combination of changes can erase the twoorder-of magnitude discrepancy between this and a factory
producing competitive conductor. The key to improving
price performance has to be to increase throughput of the
factory. This implies speeding up the deposition process
from what is feasible today at least tenfold, and throughput
speed tenfold as well. That means developing a continuous
process that gives uniformity and consistency of product to a
degree as yet unapproached for high temperature superconductors.
In our opinion, it is difficult to conceive how either PLD or
e-beam manufacturing methods will be able to realize a customer C/P in the range of $10/kAm. On the other hand,
several groups have proprietary programs underway that use
"wet and dry" chemical coating techniques, which, if success-

ful, could dramatically lower production costs and quite possibly approach this number. As always, time will tell.
IV. SUMMARY
We believe we have made a strong case in our paper for
thoughtful reconsideration of $10/kAm as a target marketentry cost/performance criterion for high temperature superconductivity wires and tapes. Indeed, if the only result is for
us to have provoked controversy and discussion of this vital
issue in HTS power applications, then our purpose has been
served. It just doesn't look possible to achieve this number
as a practical C/P result for a very long time to come, if
everpractical meaning manufacture and sale at a profit for
a wide variety of power uses. We suspect, in fact we are
convinced, there is no single C/P market-entry value whose
realization would constitute a declaration of victory. Competition with LTS wires and devices will remain for a substantial period, especially, in our view, in very large applications
such as high energy physics, and large generators and motors
where helium cryogenics is both mature and improving.
Nevertheless, certain applications definitely benefit from a
higher temperature refrigeration system, cables being the
most dramatic example. The authors are aware of several
instances where installation of low voltage, equivalent power
distribution cables could enable a given utility to release urban real estate occupied by intermediate voltage step-down
substations. The enormous savings and cash return therefrom
could justify a C/P of perhaps as high as $1000/kAm (although this would surely be a niche business!).
We urge our colleagues in manufacturing companies to
seriously consider issuing wire "specification sheets" so those
of us in the end-user community can begin to intelligently
engineer and financially plan our respective potential applications. This is most difficult if all we are given are artificial
targets which have meaning only under laboratory conditions. Let's get going. There's a lot of work and, as always
the case with applied superconductivity, a long road ahead of
us.
ACKNOWLEDGEMENT
The authors have drawn on a large reserve of colleagues
and network associations for the material used in this paper,
and it would be impossible to cite them all. It would be also
inappropriate to expose them to perceived support of our
conclusions. Nonetheless, we are grateful to all, especially
those willing to let us have a "peek under the covers" at otherwise proprietary issues.
REFERENCES
[1]
[2]
[3]

P.M. Grant, IEEE Trans. on Appl. Sup. 7, 112 (1997).

T.P. Sheahen, Introduction to High Temperature Superconductivity
(Plenum Press: 1994).
See, for example, E.B. Forsyth, Science 242, 391 (1988).

[4]
[5]

[6]
[7]
[8]

[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]

[19]

[20]

J.W. Muehlhauser et al, R&D Roadmap to Achieve Electrical Wire

Advancements from Superconducting Coatings, University of Tennessee Space Institute Report to DOE, July 1997.
To the best of our knowledge, this method of plotting cost/performance
relative to HTS wires was first used by Robert Sokolowski of IGC in
his plenary talk at ISS-96 held in Sapporo, Japan, October 1996. We
have updated it accordingly using such information various manufacturers are willing to release.
See Chapter 17 of Ref. 2.
L. Dresner, Stability of Superconductors, (Plenum Press: 1995).
We realize this figure is about a factor of 3 below the present maximum
announced by American Superconductor Corporation. Nonetheless,
we deem it a more realistic number for the industry as a whole, given
that ASC's data is for "meter scale lengths."
NKT Corporation, private communication.
A "wealth" of data on silver uses and prices can be "mined" from the
home page of the Silver Institute at www.silverinstitute.org.
C. Rossner, private communication.
We have not even estimated the capital cost of the manufacturing facilities; it certainly isnt zero, so the $31 figure is low. This matter will be
revisited in Section III.
G.B. Riley, American Superconductor Corp., private communication.
University of Wisconsin/Applied Superconductivity Center web site at
www.engr.wisc.edu/centers/asc/moi.html
X.Y. Cai, et al., Nature 392, 906 (1998); A. Pashitski, et al., Physica
C246, 133 (1995).
T.B. Peterson, et al., Appl. Phys. Letters 71, (1) July, 1997, and citations therein.
G.B. Riley, presentation at the 1998 DOE Wire Development Workshop, St. Petersburg.
J.N. Chapman, Life Cycle Cost Study for Coated Conductor Manufacture by Electron Beam and Pulsed Laser Deposition Systems, draft
DOE Topical Report # DOE/PC195231-16, University of Tennessee
Space Institute, Tullahoma TN, June 1998. Hereafter referred to as the
"Chapman Report."
R.H. Hammond, 'Thick Film YBCO for Wires and Tapes: Scale-up
Issues and Cost Estimates,"Proceedings of the 8th International Symposium on Superconductivity (ISS-95), published in Advances in Superconductivity VIII, edited by M. Hayakawa and Y. Enomoto (SpringerVerlag, Tokyo, 1996).
R. Brealey and S. Myers, Principles of Corporate Finance, (McGrawHill: 1984).