What does an equipment engineer do in a chip fab?

An equipment engineer (often just "EE") in a semiconductor fab looks after the tools themselves: keeping them running, keeping them steady, and fixing them when they break. Whether the wafers coming out are right is mostly the process engineer's job; how the two split the work is in What's the difference between a process engineer and an equipment engineer?. This article is only about the equipment engineer's own day.

Think of a mechanic for a fleet of very expensive cars: regular servicing, rushing out when one breaks down on the road, and a test drive after every repair before the car goes back into service.

What does a fab tool look like?

Inside many tools, the place where wafers actually get processed is a sealed chamber, and a tool often has two or more. Around them are a robot arm that moves wafers, vacuum pumps, power supplies, gas lines and many sensors recording numbers all the time.

Equipment engineers usually look after one area, for example the etch tools in the etch area (the areas are in What do the seven process areas of a chip fab do?). After looking after the same kind of tool for a while, it gets easier to tell what's different about it today.

What does a normal day look like? Scheduled maintenance

With use, parts wear out and chambers get dirty. The most basic part of the job is scheduled PM (preventive maintenance): stopping the tool for servicing before it breaks. Some schedules follow the calendar, some follow usage, such as how many wafers have run; some are set when the sensor numbers start to change.

Typical maintenance work:

• Chamber open: open the chamber and clean off what has slowly built up on the inner walls. Before starting, the power and gas valves are locked, with a tag carrying your name. This is lockout/tagout: someone is working inside, don't touch.
• Replacing parts: an O-ring, for example, the rubber ring between two parts that keeps gas in. Sitting in plasma and heat for a long time, it slowly ages, seals less well, and can even shed bits.
• Writing it down: what was replaced and what was measured goes into the PM record, so whoever does the next PM can look it up.

While maintenance is going on, the tool can't run product. That counts as down time too, just the scheduled kind.

Why can't production start right after maintenance?

Because "put back together" isn't the same as "just like before". Once a chamber has been opened, the conditions inside have changed, so it goes through a series of checks:

1. Pump down: close the lid and pump the air out. While the chamber was open, water vapor from the air settled on the inner walls, and it only comes off a little at a time, so this often takes a long wait.
2. Leak check: one fitting that wasn't tightened properly will leak, so you go over them one by one to make sure there's no gap.
3. Season: freshly cleaned walls make the first few wafers come out differently from usual. So a few dummy wafers that aren't customer product run first, to let the walls settle.
4. Qual run: test wafers are run to measure the results, such as how fast the tool etches and how many particles were added. Only within spec does the tool get its tool release back to production.

Who does which step differs from fab to fab. Commonly, the equipment engineer finishes the hardware side and does the tool handback to the process engineer, who checks the results. Only when both pass is the repair done. Why the chamber walls make results slowly drift is in What is etching?.

When a tool goes down at night, who gets the call?

Chip fabs never stop (the reasons are in Why do chip fabs run 24 hours a day?). Engineers mostly work day shifts and take turns being on-call at night: when it's your week and a tool has a problem the night shift can't handle, the phone call comes to you.

A common sequence: the tool's light turns red and an alarm pops up (a more serious problem; the tool usually stops by itself). The technician on the floor first does what they can by following the OCAP (the out-of-control action plan), then calls the equipment on-call. How technicians and engineers work together is in What's the difference between a fab technician and an engineer?.

When you pick up, you have to decide: which part or which signal is the problem? Can it wait until morning, or do you go in now? Should you order a part right away?

One thing to resist: just pressing reset. Clearing the error makes the tool look ready to run again; but if the cause is still there, the same problem soon comes back.

Availability, down and up: which numbers does an equipment engineer watch?

A tool that can run product is up; one that can't is down. Availability is how much of a tool's time goes to production. Fabs calculate it a bit differently; there is a public international standard that sorts a tool's time into six states, such as standby, scheduled downtime (maintenance) and unscheduled downtime (failures).

Along with availability, two more numbers are commonly tracked: how long a tool runs between failures on average, and how long each repair takes on average. The first should get longer, the second shorter.

Why care so much? Tools are very expensive and lose value even when they're idle; when one stops, the lots behind it wait. But maintenance is downtime too: maintain too often and the tool is down more; too rarely and when it does break it's down longer, and may ruin a few wafers first. So researchers have also tried deciding when to do maintenance from a tool's monitoring data: fewer unnecessary PMs, and fewer surprise breakdowns.

That monitoring system is FDC: when a sensor reading crosses a threshold, it sends an alert to call someone. Set the threshold too tight and it keeps raising false alarms; too loose and it misses the kind of problem that gets worse slowly.

Who does an equipment engineer work with?

Who	What they do together
Technicians	Often the first on the floor to notice something is off; they follow the procedure first, then call and describe clearly what they see
Process engineers	Tool handback and qual run after a repair; they decide what happens to the lots inside and behind the tool
Manufacturing	Ask "When will it be up?" so they can plan the lots waiting behind
Facilities	Look after the power, water, gases and chemicals a tool uses; working out whose side a problem is on
Vendor engineers	Big problems and special parts, looked at together

The split with Facilities roughly follows the tool's connection points (see What does a fab facilities engineer do?). If a pump keeps getting hotter, for example, you first ask Facilities whether the main cooling water loop is fine; if the loop and the other tools are fine, the problem is in this tool's own part. How Manufacturing lines up the work is in What does the manufacturing department do in a chip fab?. To run tests on a tool that's in production, you also need a tool loan from Manufacturing: while it's lent out, it doesn't ship product.

In fab24hr

Tap "Clock in", choose "A day in the fab", then on the "Your role today" card tap "Production fab", swipe to the equipment engineer and tap "Etch" under the name: you're the etch equipment engineer. Which of the four shifts you get follows the time you open the game:

• Late-night shift, "Three Calls": you're the equipment on-call this week. You're at home, and the night shift's photos are your eyes: low liquid level on a wet clean tool, Chamber A on an etch tool that can't grip the wafer, a pump on a thin-film tool getting hotter. In each photo, you have to tell which thing is the cause and which is only a result.
• Morning shift, "Chamber open": Chamber A of ETCH-03 has been down since the small hours, and today you open it: lockout/tagout, vent and wait for it to cool, remove the old ESC (electrostatic chuck), tighten eight bolts in a star pattern, pump down, leak check. Every step has a shortcut, and whatever you save comes back for you later. When it's done, you hand the tool back, and Kai, your senior, says: "Fixed doesn't count. It counts when Process says it passed."
• Afternoon shift, "No Spare": the thin-film tool CVD-02 failed its particle test after a major PM, and the spare showerhead won't arrive until tomorrow. Borrow one from another section, let it run with a monitor wafer on every lot, or keep it down until tomorrow: your call. On some days, the afternoon shift is about a hot plate in lithography instead.
• Evening shift, "False Alarms All Night": you stay late to write the repair report for the ESC swap, and FDC keeps calling you, mostly false alarms. Whether to loosen the thresholds is your choice too.

The companies, people, tool numbers and figures in the game are fictional; what equipment engineers do and why comes from public, general sources, and the real split of work differs between companies. For other jargon, see Chip fab jargon in plain English.