Welcome to Madame Curious' Blog Page**

-----------------------------------------------------------------------------------------------------------------------------

I’m in Vermont now learning about how IBM makes the electronic chips that go into computers and other things like cell phone and video games. 

This is my host, Joe (I’ll explain about that thing in his hand later.)  Joe is an engineer in a group called enablement.  They study all of the individual pieces, like transistors, wires and capacitors that a computer chip is made of and describe it all with mathematics.  Then the people who design the chips can use a computer to simulate how the new chip will work before it is built.  Joe’s main job is to adjust the equations until they match data that comes from the lab. But today we’re going to visit the lab where the data comes from.

Joe’s from Wisconsin were he got a degree in nuclear engineering.  After time in the US Navy he came to IBM and got another degree in electrical engineering while working for IBM.

Everyone here is excited about a computer called Watson which will be on national TV in February when it competes on the game show “Jeopardy” against human champions.  Before Watson could be designed and built this lab measured the transistors and built the mathematical models the designers needed.

This is Phil, who grew up in Maine.  After high school he went to a technical college, then joined IBM.  Phil is the lead engineer for the lab, solving any problems with the measurements and creating software which he is explaining to me.  Phil is explaining to me how engineers can use a computer anywhere in the world to specify exactly what devices to measure and exactly what electrical measurements to make.  Their computer will send the instructions to a computer in the lab here.  After the measurements are done all the data is sent to the engineer’s computer so they can study it.  Many of the engineers who use the data from this don’t even work in Vermont. Some of them work for IBM in Fishkill, New York and some in Bangalore, India.

This is Moe, a Vermonter who went to Vermont Technical College. The two disks Moe is showing me are thin slices of silicon called wafers.  This is what computer chips look like while they are being made.  When they are finished, the wafers are cut into small pieces and each one is put in a plastic or ceramic package with wires to connect it to the other circuits in the computer.  The big wafer is 300 millimeters in diameter.  That’s about 1 foot, almost as big as I am tall.  It was made in IBM’s factory in New York which also made the chips for the Watson computer.  The smaller one is 200mm and was made here in Vermont.  Chips made in Vermont are used in cell phones, DVD players and GPS navigation systems. 

This is a good time to talk about sizes.  Scientists like to talk about sizes in “orders of magnitude”.  An order of magnitude is a factor of 10.  The big wafer is 300 mm (mm is the abbreviation for millimeters), that about the size I am. One lens of my glasses is about 30 mm in diameter, one order of magnitude smaller.  Joe and I found some pictures on the web getting smaller and smaller to show you how small the transistors they test are. 

This lady bug about 3 millimeters 2 orders of magnitude smaller than me.

Three human hairs side by side is about 300 micrometers (um) across.  That’s three orders of magnitude smaller than me.  1 micrometer is 1- 1 millionth of a meter.

When you see a sunbeam coming in a window, what you really are seeing is light reflecting of particles of dust floating in the air.  Those particles are just a little smaller than the human eye can see.  That’s about 30um.  That’s four orders of magnitude smaller than me.

The things that make your blood red are tiny red blood cells.  They are about 6um in diameter and 2 um thick.  There much too small to see without a microscope and five orders of magnitude smaller than me.

Even smaller than the cells in your body are bacteria.  These bacteria are about 300 nanometers (nm) in diameter.  That’s 6 orders of magnitude smaller than me.  I’m a million times bigger than that.

Viruses, like the virus that gives you a cold are even smaller than bacteria.  They are also too small to take a picture even with a microscope, so this is a drawing.  The cold virus is about 30 nm in diameter.  That’s 7 orders of magnitude smaller than me.  It’s also about the size of the smallest transistor IBM is making now on computer chips.

Now back to the lab where they measure very small transistors on large wafers.   Here I am helping Crystal and Radhika move a cart full of wafers across the lab.  The wafers are made in rooms where the air is filtered to remove any dust particles.  These boxes seal air tight so that no dust can get on the wafers when they are being moved.  That’s important because the dust particles in the air 3 orders of magnitude bigger than the smallest transistors.  Crystal is also a Vermonter with a degree from the University of Vermont.  Radhika grew up in Hyderabad, India and has degree from an India university and a masters degree from the University of Texas.

Here I’m working with Chris to find the right wafer to test.  Each wafer has a unique number written on the back of the wafer with a laser.  And each box has a lot number.  When an engineer needs a particular wafer measured, they use the lot and wafer numbers to tell the lab which wafer to measure.  A technician like Chris finds the exact wafer and puts it on the tester.  Chris is from Vermont with a degree from Vermont Technical College.

We found the wafer and Radhika and I are loading into a machine called a “wafer prober”.  The wafer goes on a flat plate called a vacuum chuck.  There are small holes in the chuck which are connected to a vacuum pump.   With a vacuum under the wafer the air pressure above the wafer pushes it down against the chuck so that it will not slide while being tested.  Once the wafer is on the chuck the chuck slides into the probe station under the microscope and the probes you can see in the next picture.

Here is Jay telling me how to connect the instruments to one particular transistor on the wafer.  We have already run electrical wires from the instruments to the “positioners”.  The wires are the grey and bright yellow lines in this picture.  Notice the microscope with eye pieces and a camera at the top.  We will use that line up the probes to the tiny structure we are going to measure.  Jay is from Vermont and has a degree form the University of Vermont.

Here is a close up view of the positioners.  Each one has three knobs for moving a tiny probe tip to line it up with the transistor on the wafer.  Turning the one on the “back” of the positioner (closest to the yellow triangle) moves the probe forward and backward.  Turning the one on the side moves the probe from side to side.  Turning the one on the top moves the probe up and down.  I will use that one last actually touch the wafer when I have the probe over exactly the right spot.

WOW!  I finally got all of the probes lined up correctly and gently touching the wafer. If the probes push too hard on the wafer the pads can be damaged.    The wafers are specially designed with transistors connected to special pads for this purpose.  In this picture you can see the pads on a television screen which shows what I could see through the microscope on the prober.  The large black shapes are very fine metal wires call “probes”.  The light colored rectangles are the aluminum pads which are built on the wafer and connect to the device being tested.  The pads are about the size of the end of a human hair.  With the positioners you can place the pads with a precision of a few micrometers (the diameter of a human blood cell).

That setup was for a probing a large device on the wafer called a transmission line. It is much bigger than most transistors and would not fit in a standard set of pads. That’s why I had line up each probe individually.  There’s an easier way to probe transistors in a standard pad set.  Here I am examining a probe ring for a standard pad set.  In a standard pad set the pads that the probes touch are all lined up in one or two rows.  Because the same arrangement of pads is used over and over the lab has these probe rings made.  The ring has 25, 32 or 50 pads all mounted so the tips are lined up and just the right distance apart.  Then you only have to line up the two probes on either end of the row and all the rest are automatically lined up.  That saves a lot of time.  Plus more than one transistor can be probed at the same time.

Here’s a close up of the ring.  The probes are at the very center but are too small to see without the microscope.

Crystal is explaining all of the test instruments to me.  The one at the top with the wires coming out is an “LCR” meter.  That machine uses AC electricity to measure how the test device responds to an AC signal.  The next one is called a semiconductor analyzer and it measures the device with DC electricity.  On the bottom is a switch matrix.  It has one wire coming in from every connection of the tools up above and one going out for every probe being used on the prober.  It can connect any wiring going in to any wire going out.  The computer uses it to automatically test run many tests on all of the transistors in one probe set without the technicians having to move any wires or probes.

After testing the data can be review on the computer in the lab.  Here I’m looking at a plot of data to be sure that nothing went wrong.  But most of the data analysis happens not in the lab but back in engineer’s offices.   So after we make sure the tests ran correctly we send the data to a computer where the engineers who ask for it can get it on their own computers and use it to make mathematical models which describe exactly how the device behave.  I’ll write another post later about what they do with the data.

I asked everyone in the lab what was the best part of their job.  The answers were all similar.  They like the people they work with, they like working on technology that’s the most advanced in the world and they like have new problems to solve every day.  If you like to figure out riddles and solve puzzles, you might like working in a place like this.