In the jargon of semiconductor manufacturers, the back end is all the stages of fabrication that come after the etching of the transistors onto the wafer of silicon (the front end). A naked die serves no purpose. It has to be protected and connected to the outside world. To do this, the die is soldered into a case or package, and its microscopic interfaces are connected to pins of a size that can be more easily handled.
Rennes receives its wafers from other STMicroelectronics fabs around the world (mainly Singapore, Tours, and Catania). But they’re not ordinary wafers. They’re already suited to aerospace requirements. The Rennes site has no R&D center, but it does have a radiation expert who works with the laboratories at the front-end plants to design chips that are capable of withstanding the ionizing radiation present in space (the same radiation that was responsible for the failure of Curiosity’s memory chip). They’ve been through their own special development and testing track before leaving the front end.
Preliminary Wafer Selection Tests
Once in Rennes, the wafers are cut into individual dies. STMicroelectronics then takes a sampling of chips that will undergo all stages of manufacturing, and then be irradiated by cobalt-60 sources in special laboratories (not on-site, since the use of a radioactive source requires specific expertise). The goal is to verify that the wafer the chip was taken from meets nominal performance standards. Theoretically, there’s no problem, since it was designed and developed to meet those standards. But if the results are not up to standard, the entire wafer has to be rejected.
The chip maker makes a point of adopting highly demanding selection criteria. They can’t afford to run any risk with a component that must be functional for 15 to 20 years (the life expectancy of a satellite in space). So, it's better to junk an entire wafer than risk a malfunction in a satellite in orbit or a robot on Mars.
Once the wafer is accepted, it’s cut up (diced) into individual chips. After dicing, each chip is examined under a microscope. The operation requires many different handlings, which limits the production cadence. An operator can sort a maximum of 140 chips per hour, or approximately 1000 chips per day. The selection is rigorous. At this stage, the question is not just whether the chip is functional. A visual defect alone is enough to cause the die to be rejected, even if it’s operational. Here again, the approach is to take zero risk with the final component that will ship.
Rigorous Selection
If there are any rejects, they mostly occur before packaging of the die. The initial selection criteria are so demanding that the chips that enter the production chain rarely encounter problems. However, the various certifications require that items be taken out of the circuit for testing at each stage of production.
After dicing, the chip is soldered into a case or package. The quality of the assembly is tested—a percentage of the chips are separated out and subjected to a die pull-off test. The packaging used is different from that found in consumer components. Instead of plastic, ceramic is often used—generally alumina (Al2O3) or aluminum nitride (AlN)—or else metal (steel or kovar, which is an alloy of iron, nickel, and cobalt). The choice of the case depends mainly on the type of chip it will house. Chips that need no power will use a ceramic case. Signal processing components, for example, use ceramic. Dies with a Thermal Design Power (TDP) of a few watts, such as power components that require a current of 3 to 5 A and a tension of a few volts, use metal packaging.
Contrary to popular belief, the choice of packaging doesn’t depend on the environment in which the chip will operate. Whether it is used in space or in a drill head 20 km underground, the component will be identical. The high-reliability standards these chips must meet are intended to satisfy operation under any conditions. So the choice of the case depends only on the characteristics of the circuit it houses.



Maybe I'm missing something obvious, but isn't 'utopian' the exact opposite of the word you were intending to use there?
I'm thinking that "utopian" was used in a way to mean something highly unachievable, unlikely, or unfeasible. I'm not sure how that would apply though and I can't speak for the author. I'm not sure if this helps explain that...
More practically, you should be able to delete your own post by going to the forum version of the comments. Click on the blue icon with quotation marks, full-edit the duplicate post, and there should be a delete button.