I’ve described SkyWater Technologies as a specialist in custom chips and advanced semiconductor chip packaging. But after recently receiving state and Federal funding, CEO Thomas Sonderman has ambitions to grow. He wants the company to become a much larger foundry for radiation hardened (rad-hard) chips, and at the same time he wants to be a trusted foundry, part of a U.S. Department of Defense program that provides secure manufacturing capabilities for military and intelligence operations. With the growth in applications like unmanned aerial vehicles and commercial space launches that will need rad-hard electronics, he hopes to do what foundries like TSMC did for commercial chips: drive a separation of design from manufacturing and aggregate demand at his fab.

Anybody who has watched recent rocket launches or marveled at the pictures that come from satellites and deep space probes quickly realizes the central role played by electronic systems. These systems not only control things like rocket engines and thrusters, but they acquire data from sensors, and provide communications links back to the ground. The challenge for electronics in space is that it is a harsh environment filled with ionizing radiation, energy that is emitted in the form of x-rays or gamma rays, and electrons and charged particles. The earth’s atmosphere and its magnetic field protect us on the ground from most of this radiation, but the amount an electronic system flying on an airplane or above the atmosphere gets exposed to will vary depending on the altitude and conditions on the sun and in space. For example when the sun emits a solar flare, it sends particles shooting into space that generate lots of radiation. Electronic devices in aircraft, missiles, and spacecraft, or for that matter near nuclear reactors or nuclear power stations have to be able to withstand these and still produce the desired outputs.

Microelectronic chips are sensitive to radiation

Microelectronic chips are quite susceptible to damage from ionizing radiation. This is understandable when you consider that these chips rely on moving small amounts of electrical charges around in carefully structured lattices of silicon. Charged particles that come zooming in from space can cause permanent damage, or they can cause so-called single event effects such as when a high energy particle goes through a device and leaves an ionized track in its wake. Such events can lead to electronic noise, signal spikes, and malfunctions resulting from inaccurate or unintelligible functioning of a circuit.

The problem is different at different altitudes. There is an inner radiation belt that ranges from 6,000 – 12,000 kilometers above the earth’s surface, and an outer belt ranging from 25,000 – 45,000 km. The outer belt is where most geostationary satellites fly, and there are a lot of high energy electrons and radiation to deal with. The inner belt is where a lot of heavier charged particles can be found in high concentrations, and it is where many lower orbiting satellites and missiles fly.

Radiation hardening is the process for making chips resistant to damage and able to function in harsh environments. Shielding is an obvious approach, though one has to be careful about adding weight, especially for space applications. An alternative is to use different materials and processes to make the chips, for example by using a silicon on insulator (SOI) substrate. In this approach, a silicon thin-film layer sits on top of an electrical insulator, which might be a buried layer of silicon dioxide (called a buried oxide layer), or sapphire – in which case it is called silicon-on-sapphire (SOS). The interconnected CMOS circuits are fabricated in the silicon thin-film layer, and the insulator layer reduces what is known as parasitic capacitance, an unintentional buildup of electrical charge that happens when conductive materials are close to each other. It also means charges induced by radiation don’t migrate into the active regions of the transistors, which makes them less vulnerable. An advanced version of SOI called fully depleted silicon in insulator (FDSOI) is now one of the preferred approaches for making rad hard chips.

Many Defense Contractors Have Their Own Fabs

A number of defense contractors manufacture rad-hard microelectronics. BAE Systems has a large defense electronics business spanning space systems, electronic warfare, navigation, surveillance systems, sensors, radar systems, and display and targeting systems. It has its own rad-hard fab for making microprocessors, memories, application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs) and other standard components. Last year it received $35.5 million in funding from the CHIPS Act to modernize its trusted foundry in Nashua, New Hampshire. Honeywell offers a number of chips manufactured on 0.8 to 0.15 micron geometries (800 – 150 nm). It sells catalog offerings of their own design for processors, memories, and communications applications, as well as offering foundry services to customers who want their own designs manufactured. The Department of Defense awarded Honeywell $25.8 million last September to support their development of a 90 nm process at their fab in Plymouth, Minnesota. Other defense contractors like Sandia National Laboratories’ Microsystems Center and Northrup Grumman also have their own dedicated in-house rad-hard manufacturing.

On the other side of Minneapolis from Honeywell is Bloomington-based SkyWater, who has also been racking up government support. In December 2024, it announced proposed CHIPS funding of up to $16 million. Coupled with a pledge of $19 million from the State of Minnesota, the company is using the incentives to modernize the facility and increase capacity. The CHIPS funding is in addition to a 2019 DOD $170 million investment for a multi-phase project to enhance microelectronics capabilities for the strategic rad-hard market. SkyWater’s 90 nm FDSOI technology platform can support different levels of hardening – strategic rad-hard for the most extreme conditions, rad-hard for resistance to radiation but not necessarily the extremes of temperature or other conditions found in deep space, and plain vanilla radiation tolerant for less demanding applications.

SkyWater CEO Thomas Sonderman sees the field as ripe for conversion to a foundry model. His argument is that rather than maintaining their own smaller fabs and take the execution risk on getting processes working while bearing potentially low utilization rates for a lot of expensive equipment, SkyWater can pool demand at its Bloomington foundry. “When you look at design efficiency, you look at competitive efficiency, having a foundry that’s serving the market means that defense contractors can focus on design,” he explained. “We can centralize the manufacturing, they centralize the development in their product space.”

Sonderman wants SkyWater to be the demand aggregator with its trusted domestic foundry. He’s using the same playbook that TSMC pioneered in the commercial chip space with some twists. For years, companies like Qualcomm and Nvidia chose to focus on design, and contracted out the capital-intensive, risk-prone business of manufacturing their designs to foundries like TSMC. And for years many people didn’t believe TSMC could aggregate demand and reap scale economies across so many different customer needs in such a highly capital-intensive business and make good money in the process. Until they did. SkyWater will face a similar challenge. While there is a boom in low-earth orbit satellites, military hardware, and other products that need rad-hard chips, the volumes are still miniscule compared to what TSMC handles. SkyWater will have to figure out how to reap scale economies by aggregating across a much smaller and highly specialized market. Sonderman is convinced he can do it.

Read the full article here

Share.
Leave A Reply

Exit mobile version