Also, what is the total length of the interconnect (including poly-silicon, vias, and metal tracks) on a high-end silicon chip? Hmmm, interesting...
I received a thought-provoking email from EDA expert, analyst, and "stylist to the stars" Gary Smith the other day posing a variety of questions to ponder, including the following:
- How big is a bacterium compared to a transistor on a modern silicon chip?
- How many die do you get on a 300 mm diameter wafer?
- What is the total length of the interconnect (poly-silicon and metal tracks) on a high-end silicon chip?
Why did Gary want to know these facts? Who knows? Over the years I've learned that it's usually better not to ask, otherwise there's a danger of my being sucked into incredibly convoluted situations that "grow in the telling".
The answers I provided to Gary are as follows, but I'd be very interested in any additional information you may have to offer (just email me at the address provided at the end of this blog):
#1 How big is a bacterium compared to a transistor on a modern silicon chip?
When I started designing ASICs in 1980, we were working somewhere around the 5 µm technology node as I recall, where "µm" is the abbreviation for microns or micrometers, meaning "a millionth of a meter". (When we use the terms "technology node" or "process" or "geometry" in this context, we're alluding to the size of the smallest structures on the silicon chip, such as the pitch of the tracks or the length of the channel forming an individual transistor).
As an aside, the µ symbol stands for "micro" from the Greek micros, meaning "small" (hence the use of µP as an abbreviation for "microprocessor"). In the metric system, µ stands for "one millionth part of", which is why 1 µm means "one millionth part of a meter".
While I think about it, one thing that "niggles" me is when I see this sort of thing written as "5µm". The rule is that if you have a single-letter unit qualifier (like "5V", meaning "five volts"), then there is no space between the number and the qualifying letter. However, if the unit qualifier has two or more letters, then there should be a space (like "5 µm", meaning "five microns"). But we digress. . .
Ah, how I remember when the 1 µm technology node arrived circa 1990. This was pretty exciting at the time. Later we moved to 0.8 µm, then 0.6 µm, then 0.5 µm. Many folks at that time believed that this was as small as we could go, but we proceeded on to 0.35 µm followed by 0.25 µm. By the year 2000 we'd reached 0.18 µm, and by 2001 this had shrunk to 0.13 µm.
Actually, things started to get a bit awkward once we dropped below 1 µm, because it's a bit of a pain to keep on having to say things like "zero point one three microns." For this reason, in conversation it became common to talk in terms of "nano", where one nano (short for nanometer, which is abbreviated to "nm") equates to a thousandth of a micron – that is, one thousandth of a millionth of a meter. Thus, instead of mumbling "zero point one three microns" (for 0.13 µm) under our breaths, we could now stridently proclaim "one hundred and thirty nano" (for 130 nm). In reality, of course, these both end up meaning exactly the same thing, but if you insist on talking about these things in mixed company, it's best to use the vernacular of the day and present yourself as hip and trendy as opposed to an old fuddy-duddy from the last millennium.
And, of course, things moved on apace. The 130 nm node was followed by the 90 nm node. Now we're at the 65 nm node, with increasing numbers of devices at the 45 nm node expected to start appearing in 2008. Good Grief Charlie Brown, when will this all end?
Sorry, what was the original question? Oh yes, the size of a bacterium compared to a transistor. Well, that sort of depends on which bacterium we are talking about, because these little rapscallions can range from being as small as the largest viruses to large enough to be visible by the naked eye! Some members of the genus Mycoplasma, for example, are around 100 to 200 nm in diameter, which means they are approximately two to four times the size of the channel of a current state-of-the-art transistor at the 45 nm technology node.
By comparison, a typical bacteria is about 2 µm to 6 µm long (an E.coli bacterium, for example is about 2 µm long). Let's average this out at 4 µm for a typical bacterium. Thus, if we say that 45 nm is about 1/20 of 1 µm (give or take), then a transistor at 45 um is about 1/80 the size of a typical bacterium.
As an aside, the largest bacterium known is Epulopiscium fisthelsoni, which can grow to 600 µm, which is about the size of the period at the end of a printed sentence. These little scamps, known as Epulos for short, live in the guts of surgeonfish, which live near tropical coral reefs.
As another aside, it's common to see articles in non-technical publications making meaningless comparisons, such as saying: "the transistors on modern silicon chips are smaller than the diameter of a human hair". Although true, this is sort of a "Duh!" statement, because transistors are so much smaller.
How much smaller? Well, human hairs range in thickness from around 0.07 mm to 0.1 mm. I am informed that the hair from a typical blond lady's head is approximately 0.075 mm (three-quarters of one-tenth of a millimeter) in diameter. Of course 0.075 mm = 75 µm = 75,000 nm, which is approximately 1,670 times bigger than the channel of a transistor at the 45 nm technology node.
Eeeeek, I was going to post my answers to all of Gary's questions today, but I seem to have gotten a little carried away (just wind me up and watch me go), so I'll answer question #2 tomorrow and #3 the day after that... this is of course my way of making you return for more (I'm a sneaky little rascal that way).
Questions? Comments? Feel free to email me – Clive "Max" Maxfield – at email@example.com). And, of course, if you haven't already done so, don't forget to Sign Up for our weekly Programmable Logic DesignLine Newsletter.