Inside a Superconducting Power Cable: Cryostat, Vacuum and Cooling
Part 3 of our Superconductor 101 series. Background reading: Part 1 on what superconductivity is and Part 2 on the materials.
A superconducting power cable is really two machines in one: an electrical cable, and a cryogenic plant stretched out over its whole length. The conductor inside must stay around −200 °C for years, buried under a city street, while carrying thousands of amps at grid voltage a few centimetres away from soil at ordinary temperature. This article takes the cable apart layer by layer and explains how that is possible.
The problem the design must solve
Three requirements pull against each other:
- Electrical: thousands of amps must flow through superconducting tapes, insulated to withstand tens of kilovolts.
- Thermal: the conductor must be held at cryogenic temperature with minimal heat leaking in from the warm world. Every watt that sneaks in must be pumped back out by the refrigeration plant, at a cost of many watts of electricity.
- Mechanical: the whole assembly must survive being wound on a drum and pulled through ducts, and must cope with shrinking several millimetres per metre as it cools from room temperature to −200 °C.
Everything in the construction traces back to one of these three.
The anatomy, layer by layer
Working outward from the centre of a typical single-phase HTS cable:
The former. At the core is a flexible support, often a spiral of copper strands or a corrugated metal tube. It gives the cable its shape and bending behaviour, can serve as a channel for coolant, and provides an emergency path for current: if the superconductor ever quenches (see Part 1), the fault current momentarily flows in the copper instead of vaporizing the tapes.
The superconducting layers. HTS tapes, each a few millimetres wide (Part 2 covers REBCO and BSCCO), are wound in a spiral pattern around the former, usually in one to four layers. The spiral winding is what lets a cable containing brittle ceramic tapes bend around drums and corners: when the cable flexes, each tape shifts slightly along its helix instead of stretching. In multi-layer cables the winding angles are chosen carefully so the layers share current evenly.
The dielectric (high-voltage insulation). In most modern designs the electrical insulation also operates cold, soaked in liquid nitrogen. This is called a cold dielectric design. The standard material is PPLP (polypropylene laminated paper) or similar tape wound in many layers; impregnated with liquid nitrogen, it withstands grid voltages very well. The nitrogen is part of the insulation system as well as the coolant.
The superconducting screen. Outside the dielectric, cold-dielectric cables carry a second, thinner layer of HTS tape at earth potential. The screen carries an induced current equal and opposite to the core current, which cancels the magnetic field outside the cable almost perfectly. The consequences: neighbouring phases don’t heat each other, nothing magnetic escapes into the surroundings, and the cable’s inductance is unusually low. This near-zero external field is one of the quiet selling points of HTS cables in dense urban corridors.
Copper stabilization and binding layers complete the cable core before it enters its thermal armour.
The cryostat: a kilometre-long thermos flask
Everything above lives inside the cryostat, which works on the same principle as a vacuum flask, manufactured by the kilometre.
It consists of two concentric corrugated stainless-steel pipes with the space between them evacuated. The corrugation (like the bendy section of a drinking straw, running the whole length) is what makes a rigid steel pipe flexible enough to coil on a drum, and it absorbs the thermal contraction when the inner pipe cools by 200 degrees while the outer pipe stays at ground temperature.
Vacuum stops heat conduction and convection, but not radiation: a warm outer wall glows infrared at the cold inner wall. So the gap is also filled with multilayer insulation (MLI), meaning dozens of wraps of aluminized polyester film, each reflecting radiant heat back outward. This is the same shiny material used on spacecraft, doing the same job.
How good does the vacuum need to be? Roughly a millionth of atmospheric pressure or better. At that level, so few gas molecules remain between the walls that they carry negligible heat. Cryostat sections are evacuated at the factory and sealed. Maintaining that vacuum over decades is one of the technology’s real long-term engineering challenges, since even tiny leaks, or gases slowly released from the steel surfaces themselves, degrade the insulation. A well-made cable cryostat leaks heat at only about one watt per metre, which is remarkable for a pipe with a 200-degree temperature difference across a few centimetres.
The cooling system
The coolant in nearly all HTS grid cables is liquid nitrogen: cheap, abundant, non-flammable, and an electrical insulator to boot (Part 1 explains why 77 K is the magic threshold).
Cables don’t use nitrogen the way a laboratory flask does, quietly boiling at 77 K. Bubbles of gas are electrically weak spots inside a high-voltage dielectric, so the nitrogen is subcooled (chilled below its boiling point, typically to 65–72 K) and pressurized, guaranteeing it stays liquid everywhere in the circuit. Subcooling also buys margin: colder nitrogen means the tapes sit further below their critical temperature and can carry more current.
The circuit is a loop. A pump pushes subcooled nitrogen down the cable (through the former, through the annular space around the cable core, or through a separate return pipe, depending on the design), where it absorbs the heat leaking through the cryostat plus the small electrical losses. It then returns to a cooling station where refrigerators chill it back down. The refrigeration is the expensive part. A rule of thumb is that removing one watt of heat at 70 K costs on the order of twenty watts of electricity at the plug, which is why the cryostat’s watt-per-metre performance matters so much.
Cooling stations also set the practical length limit of a single cable section. Pressure drop and gradual temperature rise mean nitrogen can only be pushed so far before it must be re-chilled, typically a few kilometres with today’s designs. Munich’s planned 15 km SuperLink (see the tracker) is pushing exactly this boundary, with intermediate cooling along the route. Some designs take different approaches entirely: Russia’s St Petersburg DC line uses banks of cryocooler machines, and VEIR’s US pilot lines evaporate nitrogen deliberately for passive open-cycle cooling.
Terminations: the hardest metres of the cable
At each end, the cable must connect to an ordinary warm grid, and the termination is where all three design problems collide in one device. Within about a metre it must:
- bring thousands of amps from room-temperature busbars into the cryogenic conductor through current leads, which are an unavoidable heat highway (a metal that conducts electricity well conducts heat well too, and tens of watts flow in per lead);
- step the high voltage safely across the temperature gradient through a bushing;
- seal liquid nitrogen and vacuum against the outside world while parts of the assembly shrink and grow with temperature.
Terminations typically dominate the heat budget of a short cable and account for a significant slice of total system cost. This is one reason superconducting cables favour applications where a lot of power must travel between two points: the termination overhead is paid once, however long the cable.
One core or three?
AC grids run in three phases, and HTS cable makers arrange them in two main ways. Three separate single-phase cables (as at LIPA on Long Island) keep each phase in its own cryostat: simple and high-capacity, but three times the cryostat. Triaxial concentric designs wind all three phases around one former, separated by dielectric layers, inside a single cryostat: the most compact and nitrogen-efficient option, used at AmpaCity in Essen and in Shenzhen’s city-centre cable. DC cables (St Petersburg, and CERN’s MgB₂ links) sidestep phase geometry entirely and avoid AC losses altogether.
A note on those AC losses: superconductors are only perfectly lossless for steady (DC) current. Alternating current makes the magnetic vortices inside the tapes (see Part 1) shuffle back and forth fifty or sixty times a second, dissipating a small amount of heat. It amounts to fractions of a watt per metre, but at 70 K every fraction counts, and minimising AC loss drives much of the fine detail in tape layout and winding geometry.
What this looks like in the real world
None of this is hypothetical; the engineering described above has been in the ground for decades. From our tracker: Essen’s AmpaCity cable (1 km, three phases in one cryostat, 40 MVA at just 10 kV) ran seven years in a downtown grid and delivered around 20 GWh. Shanghai’s Xuhui cable has run a 1.2 km commercial link since 2021 with daily loads of 700–800 A. LIPA’s Holbrook cable put 574 MVA of transmission capacity in a 600 m trench back in 2008. And in Yongin, South Korea, KEPCO’s 1 km cable became the world’s first in commercial utility operation in 2019.
Next in the series: the tape inside those cables is its own manufacturing epic. Continue to Part 4: How REBCO superconductor tape is made (and why it’s expensive).