Nested Acronyms, 1

Acronyms, love ‘em or hate ‘em. I say: both! Acronyms save you a lot of typing and create an inside language that unifies a community. Yet their mysterious capital-lettered wall reveals no clue to what’s behind it if you’re not in the know.

You have to be in the Princeton Nassoons for three years before you learn what SFC stands for. (I don’t know what it stands for. Never was in the Nassoons. But check them out on Spotify if you like men’s a capella!)

Nested acronyms are the Russian doll of acronyms. You open up the acronym only to find another acronym inside.

Two of these have come into my life: SNV and RBD. You’re going to hear about both of them if you keep reading my stories. SNV and RBD are dear to my heart, for very different reasons.

I’m sure there are more examples of nested acronyms. Maybe you know some; if so, add them in the Comments section below. I can also imagine, with some trepidation, double-nested acronyms. An acronym within an acronym within an acronym. Dastardly. Never encountered one in the wild, though.

Sidebar: Taken to an extreme, you can imagine an infinity mirror of acronyms. Every time you unfold one, there’s another one inside, ad infinitum. Or recursive acronyms! They’d circle back and join the beginning. What would a Möbius strip of acronyms look like? OK, I’ll stop. We’re just going in circles. End of sidebar.

Let’s look at SNV first. SNV = SEM Non-Visual. SEM = Scanning Electron Microscope. I see you’re intrigued. Here we go.

You know what an electron microscope (EM) is. They were invented in the 1930s by scientists at Siemens. It was the Great Depression, and there was not a lot to do, so a bunch of men went into the lab at Siemens and figured out how to make a completely novel microscope that had better resolution than any light microscope. They stood in the break room, smoking cigarettes and drinking sour coffee. They wrote on their blackboards with white chalk. Then the eyes of the shortest man lit up. He wiggled his bushy eyebrows up and down rapidly and said, Let’s try electrons!

Within a year they’d invented a lens to steer electrons. Pretty soon they had an electron microscope, which took tiny pictures. Woohoo! they said, and celebrated with more cigarettes.

I don’t know if it really happened this way, but I do believe it’s true, to quote my friend Rev Ben.

The Siemens men redoubled their efforts and soon produced a scanning electron microscope (SEM), where the electron beam was serpentined across the surface to build up a larger-area picture. Much more practical! A marketable product! They got sweet little plaques from Siemens for their lucrative patents. And to cap off a productive decade, they invented the transmission electron microscope (TEM), which shoots electrons straight through a tiny shaving of a sample and can image freakin’ atomic structure. Sheesh, a TEM can see how the atoms are arranged--that's how amazing the resolution is. Another patent plaque for each of their desks!

All this excitement happened at Siemens while their fellow countrymen were standing in bread lines waiting for a loaf of Dunkelbrot. In the US, our ancestors similarly waited for their daily allotment of Wonder Bread (WB). Our strong bodies were built twelve ways.

There must be something causal from hardship to creativity, because the 1930s also produced great works of literature. And the first Technicolor movie, the Oz part of The Wizard of Oz. My favorite scene: Dorothy and Scarecrow come upon the Tin Man, immobilized in the forest from rust. The Tin Man says, WAH-wah. Dorothy turns to Scarecrow and says, He said ‘oil can.’

My parents were little kids back in the 1930s. I can just picture them in black and white, walking to their tidy gray-brick elementary schools, Dad in Redbank and Mom in Nutley, New Jersey. So cute with their little overcoats and their book bags! A tiny necktie on my future dad, and a tiny nurse's cap pinned to the hair of my future mom.

Returning to the nest of SNV, we now know what a SEM is, but what is a SEM Non-Visual? Let’s set that up.

Beau and I built our fortunes, such as they are, by working at MNC for over twenty years. Close to thirty for him, because he started earlier and didn’t take a detour to CBF, as I did. (Callback to The Real Ruler of CBF.) My son, Max, exists because his father and I met at MNC.

MNC makes capital equipment that supports semiconductor manufacturing. Capital equipment means Costly, with a capital C. Copious Cash. I’m talking multiple millions of dollars for one machine.

MNC’s most successful product lines are optical defect inspectors. A chip manufacturer buys these things, and puts them in their manufacturing lines to check every layer of the chip—and there are hundreds of layers—to see if there are any defects. The fastest defect inspectors can find flaws as tiny as, oh, three nanometers in size. Nanometer? That’s zero point <insert eight zeroes> one meter. Yes, 0.000000001 m.

For some reason, everyone likes to use a human hair for reference. A human hair from a brunette is about zero point <insert four zeroes> one meter in diameter. Blond hairs are about a factor of three smaller, but blonds are vanishing from the face of the earth, so no one will care about the diameter of blond hairs in the long run.

I prefer to use a coronavirus for comparison, but my focus group says that’s in bad taste. Forging ahead anyway, our current pandemic is caused by spiky, spherical devils about 100 nanometers across. But if you’re trying to figure out if your N95 mask is going to trap them, it’s really the size of the gooey aerosol droplet that you care about, and that’s around the same diameter as a human hair, varying in size from approximately blond to approximately black. So the pores on your mask need to be smaller than that. Handy to know, right?

Anyway, the smallest worrisome defects that arise during semiconductor manufacturing might be ten thousand times smaller than a (brown) human hair. In order to find defects this tiny, at a speed fast enough to keep up with chip manufacturing, you pretty much have to use light for detection. Photons, the solar-powered Ferraris of physics.

Some of you geekier types are thinking, what the heck? How can light be used to detect something smaller than its wavelength? Doesn’t that violate inviolable physics principles? Well, good question, smarty. The answer is that lots of tricks are used, physics tricks and algorithm tricks. It also helps a great deal if you’re smart about where to really look. (Clue: you look in the inner city areas more than in the lush green suburbs.) I can’t tell you more because I signed a non-disclosure agreement (NDA). Also, I don’t understand more than that, the real reason that MNC’s trade secrets are safe with me.

One time when MNC hosted Bring Your Kids to Work Day (BYK2WD), I took a school group to our manufacturing floor. The kids got to dress in bunny suits, a cute name for horrendously uncomfortable full-body bags that keep humans from contaminating the chips. Think N95 masks from head to toe. I explained to the bunny-suited kids that these machines found dirt on the chips.

A very cute little girl asked a genius of a question. She said, Do they clean it off once they find it?

No adult had ever asked me that question before.

And the answer is, nope. Can you believe it? Chip manufacturers pay millions of dollars for one of these machines to tell you your chips are messed up, and they don’t even clean up the mess. Imagine how much the machines would cost if they did.

MNC is the perfect employer for people who like to point out errors and then walk away.

And BTW, BYK2WD is not as fun if you WFH.

Really what these machines do is tell you the precise location of the defect. What happens next is that the chip manufacturer has to find out what caused the defect, so they can fix their process. Then they’ll get more perfect chips with the next batch. Higher yields, better profits. Too bad about this batch, though.

To find the cause and fix it, you have to know what the defects look like. Light ain’t good at that when the defects are nanometer-scale. So you send the chips over to our friend the scanning electron microscope (SEM) to get more information. You need to leverage exactly what motivated the chain-smoking Siemens engineers back in 1930: the superior resolving power of electrons, woohoo! The coordinates of the defects found by the (light-based) inspection system are passed to the (electron-based) review SEM, which no one has acronymized to RSEM.

Also, no one has verbified acronym to acronymize before now, as far as I know. Let’s hope no one does ever again.

The first thing the SEM has to do is re-locate the defect, using the coordinates that the inspection system has supplied. Then the SEM can zoom in and take an image of the defect, after which the human operator or artificial intelligence (AI) algorithms can classify it. When you know what the thing is (missing material? extra material? foreign material? copper (gasp)?), a smart process engineer can probably figure out the cause and adjust the process equipment to fix it. Now AI is starting to take over that job, too.

I don’t mean to sound regretful about AI or other automation. We consumers want fast, small, inexpensive electronics, so chips need to be manufactured easily and cheaply. Are you with me?

The re-location process is where we encounter the nested acronym SNV. Yes, we are finally getting to that, thirty-five paragraphs in. Not to be defensive or anything, but some of these are very short paragraphs.

Sometimes the SEM drives over to the specified location and sees squat. Nothing. Actually, that happens disturbingly often, and not seeing something you want to classify wastes precious time during the highly cost-sensitive process of chip manufacturing. That category of undesirable event is called a SEM Non-Visual.

You may be thinking, why are SNVs so common? Two main reasons. The more obvious but less likely reason is, the defect inspector made a mistake. I don’t even work for MNC any more, but I can tell you that an inspection mistake is rarely the reason, especially with the later models. Perhaps the MNC Kool-Aid is still coursing through my veins.

Instead it comes down to the difference between a beam of light and a beam of electrons. Photons (light particles) and electrons are very different, more different than cats and dogs. Our cat would climb up on your desk and decide that the most comfortable place to fall asleep was halfway on top of a stack of papers and pens. The dog likes blankets and couches, the softer the better. The cat would see a grocery bag, run over to it and jump inside. If it was a plastic bag, he’d lick it. The dog runs away from grocery bags, because they rustle and move in unpredictable and frightening ways. Could be just our particular dog.

Electrons have charge and mass, and photons don’t. That’s one of the many places the cat-dog analogy breaks down. Cats and dogs both have mass. Cats are more charged than dogs, though.

I’ll give you a real example in which light detects a defect that electrons cannot. Imagine a defect that’s buried inside an oxide layer, a goodly twenty nanometers under the surface. Light is going to see that and report the heck out of it! Oxide is generally transparent to light (asterisk, asterisk, wavelength, optical properties, blah blah blah). Light can see inside an oxide layer, just like you can see inside jello and detect the maraschino cherries. And then not eat it. At least, I sure wouldn’t.

Electrons can’t see inside a surface (asterisk, asterisk, energy, electronic properties, blah blah blah). Those puppies just bounce right off and say, nope! Didn’t see a thing.

Just for completeness, let’s note that electrons are also not as good as photons at seeing a tiny bit of extra material on top of the same material, especially if the extra material doesn’t have sharp edges. In short, electron-based microscopes may have superior resolution to optical microscopes (thank you, Siemens), but they can’t see everything that light can see. And, to be fair to electrons, they can see things that light can’t, even when those things are big enough that you'd think light would see it. I think that’s enough said about that.

Cats are better at purring and dogs are better at fetching. Well, most dogs anyway.

This concludes our discussion of the nested acronym SNV. Pencils down; physics class is over for today. I’m going to go play with my dog now, because I feel like I’ve unfairly disparaged him.

The story of RBD shall be relegated to Nested Acronyms, 2. Coming one of these days to a website near you.

P. S. My friend Mork (of Friend For Life) used to be the keeper of acronyms for MNC. Some of the acronyms were industry-wide and some were unique to MNC. A shocking number were reused, i.e. had multiple meanings depending on context. The list was posted on the MNC intranet under the title ACK-ronyms.

[Found colorized SEM-like image of covid-19]