How REBCO Superconductor Tape Is Made (and Why It's Expensive)
Part 4 of our Superconductor 101 series. Background reading: Part 2 explains what REBCO is, and Part 3 shows where the tape ends up.
Nearly every recent project in our tracker, from Shanghai’s grid cable to SPARC’s fusion magnets, is built from the same feedstock: REBCO coated-conductor tape, a steel-backed ribbon a few millimetres wide and about a tenth of a millimetre thick. A single 4 mm-wide tape, thinner than a credit card, can carry over a hundred amps at liquid-nitrogen temperature. A household copper cable needs a hundred times the metal to do the same.
That performance comes from a manufacturing process closer to semiconductor fabrication than to wire-drawing. Understanding it explains the single biggest question hanging over the HTS industry: why does the wire cost so much, and when will it get cheaper?
The problem: a ceramic that demands perfection
REBCO is a ceramic. You cannot melt it and draw it into wire like copper, and even if you could, the result wouldn’t work. The material’s superconducting current flows well only along particular directions of its crystal lattice, and current passing from one crystal grain to a misaligned neighbour is throttled dramatically. Tilt two adjacent grains by even a few degrees and the current they can exchange collapses.
A useful conductor therefore needs REBCO in which billions of crystal grains, over kilometres of length, are aligned to within a couple of degrees of each other. In effect, a single crystal in ribbon form. Nobody can grow a kilometre-long single crystal, so the industry found a workaround: grow the REBCO as a thin film on top of a flexible template that imposes the alignment.
The result is called coated conductor, or second-generation (2G) HTS tape.
The sandwich, layer by layer
A finished tape is a stack of layers, each with one job.
1. The substrate. The foundation is a ribbon of high-strength nickel alloy (commonly Hastelloy), typically 30–100 micrometres thick, polished to near-mirror smoothness. It contributes nothing electrically. Its job is mechanical strength and flexibility, and it is why a tape full of brittle ceramic can be wound around a spool.
2. The buffer stack. A series of oxide layers, together well under a micrometre thick, is deposited on the substrate. This is the clever part: the buffers create crystallographic texture (a common crystal orientation) on top of a metal ribbon whose own grains point every which way. The dominant technique is IBAD (ion-beam-assisted deposition). While an oxide layer, often magnesium oxide, is deposited, an ion beam bombards it at a precise angle, preferentially destroying crystals in unwanted orientations so that only correctly aligned ones survive and grow. An alternative route, RABiTS, textures the metal substrate itself by heavy rolling and annealing, then buffers over it. Either way, the top of the buffer stack presents an atomically ordered template.
3. The REBCO layer. On that template the superconductor itself is grown, as a film just 1–3 micrometres thick, by one of several deposition methods: MOCVD (metal-organic chemical vapour deposition, in which precursor gases decompose onto the heated tape), PLD (pulsed laser deposition, in which a laser blasts material off a ceramic target onto the tape), or RCE (reactive co-evaporation). Each manufacturer champions its own route, and all face the same challenge of holding temperature, chemistry and growth rate steady while kilometres of tape stream through the reactor. During growth, manufacturers also engineer in nanoscale defects: deliberate imperfections that pin magnetic vortices and boost current capacity in magnetic fields (the flux pinning described in Part 1).
4. Silver and copper. The ceramic is capped with a thin silver layer (a couple of micrometres, for electrical contact and oxygen processing), then usually surrounded with 20–50 micrometres of copper. The copper is the safety net. If the superconductor quenches, current shunts into the copper for the moments it takes protection systems to react.
The finished tape is slit to width (commonly 4–12 mm), tested, and shipped on spools, in single pieces that today reach from hundreds of metres to kilometre class.
Why it costs what it costs
The industry prices superconducting wire in dollars per kiloamp-metre ($/kA·m): the cost of enough conductor to carry 1,000 amps over one metre. This is the honest metric, because tapes differ in how much current they carry, and a dearer tape carrying twice the current can be the cheaper conductor.
REBCO’s cost drivers are structural:
- It is thin-film processing by the kilometre. The value sits in a ceramic layer a few micrometres thick, deposited in vacuum reactors at semiconductor-grade precision. Unlike a chip fab, which patterns billions of devices on one small wafer, a tape plant must sustain that precision continuously over kilometres.
- Yield is unforgiving. One bad metre can downgrade a whole piece, and the long single-piece lengths that cable and magnet makers want are precisely the hardest thing to make.
- Throughput is slow. Film growth proceeds at nature’s pace. Reactors process tape at metres per hour, not metres per second.
- Testing is total. Every metre is characterized, because a magnet builder cannot tolerate surprises at tape position 847 of 900.
The encouraging part is that none of these is a raw-materials problem (contrast BSCCO below), so costs fall as volumes rise and plants scale up. They have indeed been falling as demand from fusion, led by SPARC’s magnet programme (the largest REBCO consumer to date; see the tracker), pulls manufacturers into genuine mass production for the first time. Watching $/kA·m decline is watching the industry’s future arrive.
How the other wires are made
For contrast, the two other production superconductors from Part 2 take a completely different and more traditional route called powder-in-tube (PIT): pack superconductor powder (or its ingredients) into metal tubes, bundle them, then draw and roll into wire or tape, with heat treatments to form the final compound.
- BSCCO (1G tape) requires silver tubes, silver being the only practical metal that tolerates the oxygen heat treatment, and the finished tape ends up roughly two-thirds silver by volume. That puts a permanent precious-metal floor under its price, which is the core reason the industry moved to REBCO.
- MgB₂ uses cheap iron or nickel sheaths around two abundant elements, which is why it wins on price wherever its lower operating temperature is acceptable, such as CERN’s 120 kA links.
There is some irony in the history here: the first-generation HTS wire used the conventional wire-making process, while the second generation had to invent something without precedent.
Who makes it
Coated conductor is produced by a small global club of specialist manufacturers. Among them: THEVA (Germany), whose tape flew in the EcoSwing wind turbine and feeds Munich’s SuperLink; AMSC (US), supplier to Chicago’s grid-resilience cable and the LIPA cable; SuperPower (US), whose wire made history in the Albany cable as the first 2G wire ever installed in a live grid; Shanghai Superconductor, supplier to Shanghai’s commercial cable; and Sumitomo Electric (Japan), the standard-bearer of 1G BSCCO. You’ll find these names throughout the Key Suppliers column of the tracker, and fusion demand is drawing new capacity into the market worldwide.
From spool to street
A spool of tape is only the beginning. It gets spiral-wound into cables and buried under cities, or wound into pancake coils and stacked into 20-tesla fusion magnets. For the cable half of that story, see Part 3: Inside a superconducting power cable. For where it is all being deployed, continue to Part 5: Where superconductors are actually used.