The greening of Infineon
Peter Wawer talks to Nick Flaherty about the transition to Green Industrial Power (GIP), one of the four divisions at Infineon Technologies.
“When I took over in 2016 the business was dominated by the industrial space, and that means automation and drives and we were very much focussing on the power, the switch and the drive. We then realised that the big growth opportunities cam from renewables, energy generation, solar and wind but especially photovoltaics (PV). 350GW is being installed this year, it’s an incredible number but the infrastructure needs to grow. Then there is e-mobility starting to grow and the need for the charging infrastructure and that’s a very nice opportunity for GIP.”
“With these driving the growth and becoming a relevant portion of the business the renaming sends a message to stakeholders and reflects where the growth is coming from. That is generating more positive momentum than I expected, it is seen as doing cool things.”
The majority of the business, renamed from Industrial Power Control (IPC) in April 2023, is based around existing silicon IGBT power technology, with some parts running up to 3300V, there is a significant shift happening to silicon carbide (SiC).
“We realised very early on that SiC would be a relevant topic,” said Wawer. We introduced the first SiC diode two months before Cree in 2001, 22 years ago and from then onwards it stayed in a high performance niche. Our early developments were around the JFET as the issues of gate oxide reliability made us think that was the way to go. Then the SiC MOSFET arrived and Infineon was a late starter there. In hindsight we should heave intensified the activities earlier. We knew that the defect density would be very tricky and we would only release products into the market with the high levels of reliability that our customers require.
“As we were a late starter we went for trench structure so after a couple of intense years of development we brought this to the market five years ago. We developed a testing scheme where we are able to remove defective devices before they leave for customers. These kind of defects can grow in the field and this effects the reliability in the field. The compound growth of 50 to 60% a year is proof that we have very competitive technology and the associated portfolio which is important, it enables a huge variety of products
Fast switching IGBT with SiC diode
IGBT sits very well alongside. The majority of our revenue comes from IGBT and it is not declining. For fast switching applications with SIC the switching losses are much lower, say 80% than for IGBT. But SiC comes at a higher price.
“We are on 6in wafer today and 8in will come in this decade. We have the first mechanical samples in the fab and we are turning them into electrical samples soon.”
However the generic drives market does not need high switching speed, so its about the overall cost reduction he says/
The majority of development resource is being shifted to wide bandgap technology, both SiC and gallium nitride (GaN).
“We expect that the annual investment in grid in this decade will double €300bn to €600bn globally and in this area we see increasing demand based on IGBT but now customers are also looking at SiC. We are already having discussions on high power GaN but today there is no obvious way forward.”
The volume market for SiC is currently up to 1200V, but Infineon has also developed 3300V SiC parts.
“Traction market for trains, we see requests for SiC for the opex costs for energy efficiency and recuperation and lower noise from shifting the frequency and saving space and weight, that’s worth a lot. We are sampling the first 3.3kV module to customers.”
“SiC penetrates each and every application and its developing in a way where we only find out in close collaboration with customers. We now have a shrink path and cost reduction opportunities for SIC but it’s not an apples to apples comparison with silicon IGBTs.”
“On the system side SiC provides a massive cost advantage in certain applications. If you look at the pure unpackaged die then for the foreseeable future it will remain more costly. 12in wafers for IGBTs have huge economies of scale plus the raw material of SiC is more costly. The epitaxial layer is also a big part of the process cost so a SiC wafer will always be more costly than a silicon wafer.
“We keep a very close eye on GaN which is coming from the 650V space, and the first devices at 1200V are on sapphire as far as I know.”
The GaN activity is in the Power, Sensor Systems division that will absorb the recent GaN Systems acquisition.
“We have close collaboration with PSS and offer samples to customers and we see interest for modules and we have prototypes. Will there be an interest, for sure The big thing for us at GIP is definitely SIC.
Its up to us to do proper forecasting and corporate makes the judgement on the investment and then we have to execute to our forecasts. If we request more we have a fair share process looking into the history and profitability of the divisions. That means I have the volume I can promise to customers and that is reliable.”
“We are really stretching the company to the limit and we invest everywhere, the €5bn in module 4 in Dresden, we continue to invest in the 12in facility in Villach, and converting the legacy silicon to SiC.
“From a strategic perspective Kulim is perfect as it is 8in and SiC will transition from 6in to 8in and that’s why Kulim is a great fit. We have tangible volumes in Villach and with Kulim we have second sourcing,” he said. “Overall the EU CHIPS Act is very positive,” he said. “There is the IRA and CHIPS Act in the US and in Asia a lot of money is flowing from governments into these industries and it is appreciated that the EU sees this and supports it from a political perspective. We are not naïve, and we know things move slowly.”
There is also a more urgent need for the renewal of the power grid infrastructure.
“We are moving too slowly, but the will and the good intention is there. Building infrastructure takes decades and we have to move faster, we need a longer term vision and build on it, for example in Germany there are plans to bring the wind energy from the north to the south but we are very much behind schedule.”
“The next thing we are looking at is electrolysers for hydrogen and fuel cells. The industry is trying it out, and if the business case is not there yet, then let’s think about tomorrow and tap into the gas grid infrastructure. These things will come and customer interest is really there and as it starts to develop, we want to be early with a foot in door,” he said.