 |
Here we look at
products containing
semiconductor technology, i.e. small but essential part of eletrical
engineering. Analyzing this products leaves us with
components and
finally with materials and processes for semiconductor technology
- we run into materials science at its finest. |
|
 |
Products we define simply as something you and I
do buy or at least could buy. We also include services in this
category. |
|
|
Components are whatever one finds
inside a product; e.g. "Chips",
light emitting diodes (LED's), or
liquid crystal displays (LCD's) |
|
|
At least some components of our products of interest are made from
semiconductors.
What we want to learn in this course then is simply
- Which Semiconductors do we use?
- How do we make the component we want?
|
|
These questions go deeper then it may appear on
first sight. Let's look at two examples: |
|
So we use Si for our component
"chip". But just saying Silicon is not good enough. |
|
|
Do we use single-crystalline Si, poly-crystalline Si or
amorphous Si? Or perhaps nano-crystalline Si with some amorphous
regions? |
|
|
If we use single crystalline Si, do we go for
Czochralski-grown (CZ) single-crystals or
for float zone (FZ) single-crystals, or
possibly just for an epitaxial layer? |
|
|
OK - we take the CZ wafer. What doping type would you like? p
- or n-type? All right, we take the n-type, well done - thank you
very much. |
|
|
Sorry, we're not done yet. Would you prefer P-doping or
As-doping? Or may we recommend today's special: Sb-doping? And
what kind of interstitial oxygen concentration may I offer to you? We have a
large selection for every taste. |
|
|
You get the drift. And as in any good restaurant, you will
"taste" the difference. What you get as a component depends on your
detailed specification. |
|
Now let's make a solar cell. From Si or
from something else? |
|
|
In fact, we make (and you and I can buy it) solar cells from all kinds of
Si mentioned above, but also from GaAs, from
CuInSe2, from CdTe, from TiO2 and
from a growing number of other semiconductors and combinations of different
semiconductors. |
|
|
|
|
Why, oh why are we doing this? It seems to make
life so complicated. Can't we decide on the best material and process for solar
cells and be done with it? |
|
|
Well, being the boss of a large solar cell company, you
actually must make this decision - you can not possible run a multitude of
factories, each with its own materials and processes; you must make a choice
for one, or maybe just two basic product lines. |
|
|
The same is true for your competitor. If his choice is
different from yours, the market will tel in good time, which one of you guys
made the better choice. |
|
In other words: if we look at products, we do not
just look at technical topics, we actually look at economical issues!
Money, not Nobel
prizes is the decisive factor in the end! |
|
|
|
As a human being, you encounter all
kinds of products and services all the time - and you rarely think about what
is hidden behind the obvious. You pick up your (cell)phone, dial a number or
press a button, and expect that within seconds you will be able to talk to the
person of your choice - whoever and wherever that person might be. |
|
|
If a regular human being gives the
"behind the obvious" any thoughts, he or she will probably conclude,
in the words of Dave
Barry, "that cell phones are operated by
magic". |
|
|
As a (budding) Materials Scientist and Engineer,
you know better. Behind the obvious is semiconductor technology. Not
exclusively, and not always, but "immer öfter" (ever more
often). |
|
Note that not all that long ago (for
elderly professors) - in the 1950 ties - the number of semiconductor
products was exactly zero. |
|
|
Now we have such a large diversity of products, components and
semiconductors all around us that we can hardly do more than scratch the
surface in this course. |
|
|
|
|
The very first semiconductor products
intentionally made (i.e. based on understanding what is going on) hit the
market in the late fifties / early sixties of the 20th century - in the
form of "transistors", a word not used for a transistor per se, but
for a transistor radio. |
|
|
This was an unbelievable big
product achievement, because, for the first
time in human history it enabled everybody to make a lot of noise in public -
without any skills and exertion. Of course, the transistor radio was an instant
success. |
|
|
A portable, battery-run "transistor"
contained about 20 (Ge) transistors, already some progress in
comparison to your big and heavy home radio, that may have contained about
10 vacuum tubes as active elements. |
|
Putting several transistors on one
piece of Si, i.e. making an integrated
circuit (IC), was the next big
(double) step in technological development; it consisted of switching to
Si as base material and in finding ways for integration. |
|
|
Since then semiconductor technology is an
unprecedented success story - it is now (2007) arguably the world-wide
biggest industry with respect to product penetration. |
|
The key word in this respect is
"Moore's law", simply stating that
any quantitative measure of progress in IC technology grows
exponentially "forever" with growth rates in the 30 % range.
|
|
|
Typical measures are, for example, the number of
transistors on one chip, almost the same as the number of bits one can store in
one memory chip. What that looks like is shown in the figure below - note the
logarithmic scale |
|
|
|
|
|
Memories and micro processors |
 |
Intel processors
Courtesy of Intel Corp. |
|
|
|
|
|
The implications of exponential
growth for by now more than 40 years are staggering - use the
link
if you want to have just a flavor of this. We will just look at two points in
this context: |
|
|
The world has changed in a major way in the last
20 years or so because of semiconductor technology. Think about this
yourself. Hint: Consider what hides behind
catch words like "Internet", "electronic warfare",
"Resonance tomography", "globalization", "energy
crisis", ...., in technical, social and political terms. |
|
|
Exponential growth never
continues forever.In fact, since about 1985 serious people
believed (or actually tried to prove) that it will be over soon. When it will
be over, meaning that growth rates come down to "normal" values of a
few perscent, all kinds of problems might occur - witness the bursting of the
(stock market) "Internet bubble" in 2000 or the bursting of
the USA real estate bubble right now (Aug. 2007) - all caused by the
sudden end of exponential growth. If we are lucky, we will have a "soft
landing" in the IC business; if we are even luckier, the
foreseeable slack in the IC business will be compensated for by growth
in other areas of semiconductor technology, e.g.
solar cells. |
|
Where does this leave you? If the 30% per year growth rate peters
off, will there be jobs? Is it sensible to learn about semiconductor technology
now when it soon will be "over"? |
|
|
For an answer, look at the German car industry.
Seen with semiconductor industry eyes, technical progress in making cars in the
last 40 years or so was close to zero - compared to semiconductor
products. A factor of two in total performance progress (top speed, gas
consumption, ..) in these 40 years is already seen as enormous progress.
Compare with memory chips: 1978: 16 kbit, 2007: 16 Gbit;
improvement factor 106. |
|
|
Yes - but: The car industry is still the largest
industrial branch in Germany with lots of jobs.... |
|
|
© H. Föll (MaWi für ET&T - Script)
