是什么使动力半导体准备进入太空

全文内容 Full Text

Infineon has announced that its radiation-hardened HiRel power devices are integrated on NASA’s Nancy Grace Roman Space Telescope. The engineering question we try to understand in this article is what makes a power semiconductor suitable for a particular space mission. The answer depends on the circuit it serves, the conditions it will encounter, and what happens if it fails. Power MOSFETs switch current in converters that turn solar array and battery power into usable supply rails. They also connect and isolate loads and support redundant power paths. For instance, a switch that fails to turn off leaves a faulted circuit connected. One that fails to turn on deprives a payload of power. Either failure has consequences that extend beyond the efficiency of a single converter. Choosing the right power device requires considering the voltage rating or radiation figure. The devices must also meet the electrical specification on day one and survive the radiation environment expected over the mission. In an interview with Power Electronics News, Eric Faraci, principal product definition engineer at Infineon IR HiRel, said it is important to start by understanding the mission profile, such as the destination or orbit, operating life, shielding, redundancy, and the consequences of losing a spacecraft. In the case of long-duration space hardware, it requires more component-level evidence than a constellation whose satellites can be changed. Neither choice is always appropriate for every mission. Radiation exposure depends on how long the device operates and where it is integrated inside the spacecraft’s shielding. Engineers then translate these mission details into requirements for the switch’s voltage and current ratings, electrical characteristics, package, and radiation tolerance. Faraci says most mission requirements can be met with existing catalog MOSFETs rather than a custom die (Figure 1). “Generally, what we’re able to do is satisfy the need with many of our catalog products,” he says. A source control drawing narrows an accepted gate-threshold range or requires additional screening for an established device. Those changes select and test parts differently, but they do not change the underlying die. JANS screening under MIL-PRF-19500 addresses another requirement of reducing the risk of early-life failure in discrete semiconductors. Faraci describes more than 200 hours of production burn-in for the MOSFETs in the flow he discusses. The testing takes time and adds cost. It uncovers manufacturing defects, but does not provide radiation hardness. The mission and circuit define what a part needs to survive, and understanding radiation effects shows how it might fail. Radiation can gradually degrade a power MOSFET or damage it in a single strike. Total ionizing dose (TID) builds up over time. In a silicon MOSFET, ionizing exposure generates charge in the gate oxide and surrounding insulating regions. Some charge becomes trapped, and defects can form where silicon meets the oxide. These changes can change the gate-threshold voltage. If the threshold falls far enough, a switching in a converter may leak current or fail to turn off. A dose rating in a power device does not mean the transistor remains electrically unchanged. But Faraci explains that a device rated to 100 or 300 krad(Si) can experience parameter shifts while remaining within its specified limits through that dose. The circuit may work across those limits, including at the end of the expected exposure (Figure 2). Single-event effects (SEE) are a different problem. In this case, an energetic particle deposits charge along its path in one strike. Linear energy transfer (LET), expressed in MeV·cm2/mg, describes how much energy the particle loses per path length in a material. Drain voltage and gate bias establish electric fields inside an off-state device. Charge from the strike intensifies a local field to cause single-event burnout or rupture the gate oxide.