Best Paths

Demonstration 5

An MgB2 superconducting cable for gigawatt power transmission

Ten European partners built and tested a direct-current superconducting loop rated at 3.2 GW, using a wire that is cheap to make. The hard part turned out to be where the cable meets the grid.

Line drawing of a cable cutaway: concentric layers peeled back to a bundle of thin superconducting wires at the core
On this page
  1. Why a superconducting cable changes power transmission
  2. The MgB2 conductor and the partners behind it
  3. Building the MgB2 cable and its terminations
  4. Superconducting power transmission in the grid today
  5. What held the superconducting cable back

The wire at the heart of the fifth BEST PATHS demonstration, a superconducting power cable built for direct current, carried 700 amperes, roughly 500 times what a copper wire of the same size can take, and it did so without heating up at all. Scaled into a full cable system, the design was rated at 10,000 amperes and 320 kilovolts: 3.2 GW, about the output of three large power stations, down a superconducting conductor that an early description of the design put at 12.5 mm across. Demo 5, led by the cable maker Nexans, set out to show that a superconducting cable made from magnesium diboride (MgB2) could do this in direct current, at transmission voltage, as part of a real grid. According to Nexans, it was the first high-voltage system of its kind designed for DC; every earlier grid project had used alternating current.

The case for it was about bulk power transmission over distance. The project expected future grids to move 2 to 20 GW from remote wind and solar farms to cities several hundred kilometres away, and argued that long superconducting cables could carry 5 to 10 kA or more at up to 400 kV with no resistive loss.

The stated objectives were plain. Design, manufacture and test a superconducting cable system operating at 320 kV and 10 kA. Validate MgB2 as a conductor for high-power electricity transfer. Give transmission system operators and policymakers guidance on the technical side, the economics and the environmental impact. The demonstration ran from 2014 to 2018 and closed with a workshop in La Spezia on 5 July 2018.

Why a superconducting cable changes power transmission

Every conventional cable loses energy as heat because its copper or aluminium conductor has resistance. Christian-Eric Bruzek of Nexans, who coordinated the demonstration, put that loss at up to five per cent of the energy carried in conventional transmission; for long lines, an earlier article on the project gave a figure of up to ten per cent. A superconductor has no resistance once cooled below its critical temperature, so the conductor loss disappears. The cooling itself costs energy, but Bruzek’s point was that it costs less than the resistive loss it replaces.

The second advantage is size. The Demo 5 design needed an underground corridor about one metre wide. Carrying the same power with conventional technology, Bruzek said, would take a corridor up to ten metres wide with eight resistive cables laid side by side. A superconducting cable also does not warm the soil around it. For a grid operator trying to push several gigawatts through a city, under a river or along a crowded road, that difference in width is the whole case.

The idea is older than the grid it is meant to serve. Heike Kamerlingh Onnes identified superconductivity in 1911, when he found that some metals lose all resistance near absolute zero. In the 1960s the American physicist Richard Garwin proposed a single underground superconducting line, about 30 cm across including its cooling, to carry 100 GW over 1,000 km. Cooling a conductor to a few degrees above absolute zero made that too expensive. The ceramic high-temperature superconductors found in the 1980s work at up to about 70 kelvin above absolute zero, which makes cooling far cheaper, but they are hard to make and too expensive for long runs.

Direct current also suits a superconductor better than alternating current does. A superconductor carrying DC has no resistance at all, but an alternating current, which reverses fifty times a second, still causes small losses inside the superconducting material and in the metal around it, and that heat has to be removed by the cooling plant. A DC cable avoids those losses and also avoids the charging current that limits long AC cables. The combination is what made the project’s planners think about superconducting links hundreds of kilometres long, and not only short urban sections.

The MgB2 conductor and the partners behind it

In 2001 researchers in Japan found that magnesium diboride, a simple compound, becomes superconducting at 39 K. It only comes as a powder, but Columbus Superconductors in Genoa found that it could be made into long strands by filling copper or nickel tubes with the powder and sintering them. Giovanni Grasso, the company’s co-founder, summed up the effect on cost in one line: “MgB2 really changed the game.”

The route to Demo 5 ran through physics labs. In 2011 the Nobel laureate Carlo Rubbia started a research cluster at the Institute for Advanced Sustainability Studies (IASS) in Potsdam on superconducting lines for energy transport. In 2014 CERN tested MgB2 wire with a current of up to 20,000 amperes, a result the IASS physicist Adela Marian read as confirmation that the material suited long-distance transmission. CERN, IASS and Columbus all joined the demonstration. The full team of ten was Nexans as leader, CERN, Columbus Superconductors, ESPCI Paris, IASS Potsdam, the Karlsruhe Institute of Technology, the Italian research centre RSE, the French operator RTE, TU Dresden and the Universidad Politécnica de Madrid. The mix mattered: a wire maker, a cable maker, laboratories that could test at extreme currents and cold, a system operator that would one day have to connect the thing, and IASS, which took on the economic and environmental assessment and the public communication of the results.

By the end of the project Columbus had produced hundreds of kilometres of MgB2 wire, and the industry partners had made the cable conductor and its high-voltage insulation themselves. MgB2 has since found a use at CERN itself: the High-Luminosity LHC upgrade uses MgB2 superconducting links of 19 cables, each about 140 metres long, that together carry about 120 kA of direct current at 25 K. ASG Superconductors and CERN marked the end of production of 1,450 km of MgB2 wire for that project in October 2023.

Building the MgB2 cable and its terminations

The MgB2 cable demonstrator was a full-scale HVDC superconducting loop: a monopole cable with two terminations and the cooling systems that go with them, tested under high voltage. The team built a 30-metre prototype. The partners’ list of activities reads like a checklist for any new cable technology: make the MgB2 wires and the cable conductor; study how an MgB2 component behaves in the grid, both in normal operation and during faults; develop a safe and reliable HVDC insulation; design and build high-voltage terminations; model the grid with a superconducting cable in service; and explore links longer than 100 km.

The insulation is the clever part. It is lapped paper impregnated with liquid nitrogen, and if an electrical breakdown opens a gap, the nitrogen flows in and the insulation recovers. The team published the test results in a scientific journal.

Grasso, speaking at the close of the project, pointed out that MgB2 wire was not new to industry: it had already been proven in full-scale devices such as MRI systems, induction heaters and fault current limiters. What Demo 5 added was the step from those machines to power transmission, where the conductor has to be cheap enough to run for kilometres and reliable enough to be trusted with a grid.

Cooling was the least settled element. An early description of the design had an outer nitrogen circuit bringing the temperature down to about 70 K and an inner stage cooling the cable to about 20 K with either liquid hydrogen or helium gas. Bruzek said at the time that the cryogenic system was the main aspect still limiting the application.

What Bruzek called the hardest task was not in the cable at all. It was the termination, the joint where a conductor held far below minus 200 degrees Celsius meets a high-voltage grid at ambient temperature. He named it one of the most important insights of the demonstration.

The results were shown to about 60 participants at the La Spezia workshop, hosted by ASG Superconductors and co-organised by Nexans and IASS, with a visit to the Columbus wire plant. The partners reported that experts from ten European countries had designed the cable system, including its high-voltage insulation and terminations, and that the result set new records for superconducting power transmission.

Superconducting power transmission in the grid today

Superconducting cables have run in grids before, all of them AC and all of them short. The project pointed to a 0.6 GW AC cable operating on Long Island in the United States and a 1.4 GW AC prototype installed in a substation in Japan, both built from the costly ceramic superconductors. In Germany, the Munich utility and NKT inaugurated in October 2024 the test system for SuperLink, described as the world’s longest superconducting power cable system in development.

The economics depend on the place. The project’s assessment found that where rights of way are scarce or civil works are costly, a superconducting power cable is already cost-competitive with resistive cable, because the narrow trench and underground route cut the price of easements and construction. The earliest uses, Bruzek argued, would be short “bridges”: a superconducting section under a river or through a dense district, joined to ordinary cables or overhead lines at each end. Long links to remote wind farms, where several gigawatts have to travel hundreds of kilometres, would come later.

For comparison, the site’s HVDC explainer covers the resistive DC cables that do this job today, and the guide to grid congestion explains why corridor width matters so much in European cities.

What held the superconducting cable back

The limits were set out by the people who would have to buy it. At the La Spezia workshop, Jean-Yves Astic of the French operator RTE said that to invest in superconducting links, system operators would need customised use cases from 63 to 320 kV, with the exact costs, the practical operation and the long-term maintenance spelled out. Massimo Rebolini of Terna described the Italian operator’s criteria for any new cable as four: technology, environment, resilience and grid integration. Only part of that list, he said, had been addressed in the project.

The project’s own next step was to write testing guidelines for high-voltage DC superconducting cables, agreed by a consortium of manufacturers together with the operators, so that safety and quality could be checked the way they are for any other cable. Without such a standard no operator will sign for a first installation, and Bruzek was blunt that operators were still cautious. He asked policymakers for economic incentives to break the deadlock.

The honest verdict is that Demo 5 solved the materials question and left the market question open. The MgB2 cable works, the wire is cheap, and the system was built at the rating it promised. The superconducting power cable in a DC transmission grid is still waiting for its first owner.

The project’s white paper on superconducting links for very high power transmission, published on 15 October 2018, laid out a 3-gigawatt-class cable and a set of recommendations for industry and policymakers. In June 2018 Bruzek said he expected superconducting cables to enter the real grid within five years. Whether the first one at transmission voltage will run on direct current, as the La Spezia loop did, is still an open question.

From the project record

  1. Event ·

    Demo 5 Special Workshop: superconducting cables for very high power

    The Demo 5 special workshop at ASG Superconductors in La Spezia on 5 July 2018: results of a 3.2 GW, 320 kV direct current superconducting cable system.

  2. News ·

    New superconducting cable sets records for power transmission

    La Spezia, July 2018: Nexans, Columbus, Terna and RTE presented a 3.2 GW magnesium diboride superconducting DC cable system.

  3. News ·

    Superconducting cables: from the laboratory into the power grid

    June 2018: Nexans engineer Christian-Eric Bruzek on the 3.2 GW MgB2 DC cable, its 30-metre prototype, nitrogen-impregnated insulation and first applications.

  4. News

    New superconductive material for long-distance energy transmission

    How magnesium diboride, superconducting at 39 K, could make long underground power lines affordable, and the 3.2 GW cable test planned.

Questions

What is a superconducting cable?

It is a power cable whose conductor is made of a material that loses all electrical resistance when cooled below a critical temperature. With no resistance there is no heating loss in the conductor, so a thin superconducting cable can carry a current that would need several conventional cables. The cost moves to the cooling system that keeps it cold.

Why use MgB2 for a superconducting power cable?

Magnesium diboride becomes superconducting at 39 K and is a simple compound that can be drawn into long wires by filling metal tubes with powder. The ceramic high-temperature superconductors used in earlier AC cable trials work at warmer temperatures but are complex and costly to produce. MgB2 trades a colder operating point for a much cheaper conductor.

How much power could the BEST PATHS superconducting cable carry?

The demonstrator was a monopole system designed for 10 kA at 320 kV, which corresponds to 3.2 GW, roughly the output of three large power stations. The team reported that the MgB2 wires carried about 500 times the current of a copper wire of the same size, and that hundreds of kilometres of wire had been produced.

Are superconducting cables already used in power grids?

Only in short AC installations so far. The project cited a 0.6 GW AC cable in service on Long Island and a 1.4 GW AC prototype installed in a Japanese substation. In October 2024 NKT and the Munich utility inaugurated a test system for SuperLink, a planned superconducting cable for the city. The BEST PATHS loop was, by Nexans's account, the first such system designed for direct current.

What stops superconducting power transmission from spreading?

Three things, according to the people who built the demonstrator: there are no testing standards for high-voltage DC superconducting cables, the terminations that join the cold cable to the warm grid are hard to engineer, and system operators want use cases with exact costs and maintenance figures before they order one. Cooling is the other open engineering question.

Follow the line

  1. Direct current HVDC transmission explained: the cable, the converter station and why DC wins

    How HVDC transmission works: why DC wins on long and subsea routes, LCC vs VSC, what sits inside a converter station and which cable types run in Europe.

  2. Demonstration 4 Repowering transmission lines: reconductoring and dynamic line rating on Europe's AC grid

    Repowering transmission lines without new routes: reconductoring, dynamic line rating, insulated cross-arms and live-line work, as tested in Demo 4.

  3. Guide Grid congestion: why Europe pays billions to move power it already has

    What grid congestion is, what redispatch and curtailment cost in Germany and the EU, why Dutch firms wait for connections, and which fixes work.

  4. Glossary Power grid glossary: the terms behind HVDC and interconnectors

    Plain-English definitions of the terms used on Europe's power grid, from ampacity and bidding zone to VSC, XLPE and extra high voltage transmission.