Another 50th Anniversary
Recently, I’ve written about a number of anniversaries, notably
the 50th anniversary of the launching of Sputnik, which began
the space age and has gotten a lot of media attention in the past
few weeks. I was reminded of another significant anniversary
when I received my copy of the Fall issue of Interface, a
quarterly publication of The Electrochemical Society (ECS).
The focus of the issue, as headlined on the cover, is “Frosch and
Derick: Fifty Years Later”. Carl Frosch and Lincoln Derick were
two of my favorite Bell Labs colleagues and friends. Eight years
ago (7/13/1999), I wrote about their achievement that ECS
deemed important enough to warrant a special issue devoted to
them and the developments following their invention.
Here, I’ll repeat some of that earlier column but with more
details on the events surrounding the invention of oxide-masking
and more about Frosch and Derick themselves. Some of the
material is taken from an adaptation of an article by Nick
Holonyak, Jr. in this issue of Interface. (I’ve written before of
Holonyak’s achievements in the field of light-emitting diodes.)
In September 1957, Frosch and Derick published a paper in the
Journal of The Electrochemical Society titled “Surface Protection
and Selective Masking during Diffusion in Silicon”. Derick and
Frosch’s patent, “Oxidation of Semiconductive Surfaces for
Controlled Diffusion” issued in August that year. Let’s go back
to 1954. Following the invention of the transistor in 1947,
various methods were used to make the so-called p-n junctions in
germanium and silicon transistors or other devices. Making
these p-n junctions required control of the incorporation of
impurities such as aluminum (for p-type material) or phosphorus
(for n-type material). By 1954, the preferred way to control
impurities was to diffuse them into germanium or silicon from a
vapor containing the particular impurity.
It was also clear that silicon had properties that made it
preferable to germanium for transistors – if the diffusion of the
impurities could be controlled. Frosch and Derick, his very
capable technical assistant, had been working on diffusion in
silicon for some years and were called upon by Holonyak and
others to attempt the fabrication of various device structures.
The temperatures required for impurities to diffuse far enough
into silicon were quite high, sometimes 1100 degrees Centigrade
or higher. Taking phosphorus as an example, Frosch and Derick
would try diffusing by placing silicon wafers in a furnace with a
stream of a dry gas such as hydrogen, nitrogen or argon
containing phosphorus vapor flowing over the silicon wafer.
However, with the dry gas the silicon wafers became pitted, n-
layers were lost or, as Holonyak puts it, “ Frosch regularly
reduced many of our silicon wafers to cinders ”! However, in
the spring of 1955, Holonyak walked into Frosch’s lab and found
Frosch with the usual long face saying, “Well, we did it again.”
Then Frosch broke into a smile and showed Holonyak the silicon
wafers, smooth and green in color. The green color was an
interference color due to a thin film covering the silicon, the
color arising from the same interference effect as the colors of a
thin film of oil on water.
What had happened? As Linc Derick told me, he had left the lab
while passing hydrogen over some silicon wafers. On his return,
he found that the regulator valve was faulty and the hydrogen,
which was burned as it exited the furnace, had stopped flowing.
The burning hydrogen had flashed back over the silicon wafers.
Expecting the worst, when Derick pulled the wafers out of the
furnace he found they looked like “jewels” with their bright
interference. When hydrogen burns it forms water or, in this
case, steam. When the steam passed over the hot silicon, it
formed a silicon dioxide film.
Frosch immediately recognized the significance of what had
happened. They deliberately added water vapor in following
experiments and got the same thin oxide films. They then
showed that the oxide films blocked some impurities, while other
impurities could pass through the film and into the silicon. They
also found that patterns could be etched in the oxide films, thus
allowing selective diffusion of impurities directly into well-
defined regions of a silicon wafer. California didn’t know it but
the stage was set for the birth of Silicon Valley.
Patents resulted and 160 companies obtained licenses.
Commenting on the oxide-masking invention Holonyak states,
“He (Frosch) appreciated almost immediately the oxide’s layer’s
importance. It is debatable whether any other contribution has
had as much to do with the microchip’s existence and today’s
microelectronics industry as the silicon-dioxide layer.” “Silicon
itself is, of course, the critical ingredient, followed by its unique
natural oxide, without which little of today’s thriving
semiconductor industry would ever have begun to exist.”
What kind of men were Frosch and Derick? Frosch was a
delightful human being, modest and unassuming. I will always
remember his generosity in allowing me to go over to his house
and collect black walnuts, which I loved, from under tree in his
yard. For years I had lunch with a group that included Carl, a
chain smoker. The Bell Labs cafeteria served quite tasty pies,
my favorite being the cherry, and dessert was customary. Carl
invariably lit up and smoked a cigarette before dessert. Late in
his career, he had some sort of surgery and was told he had to
quit smoking. We were all shocked when he gave up the habit
cold turkey and said he didn’t miss it at all! Sadly, in his last
years he suffered from severe emphysema, barely able to walk
fifty feet before having to sit down and rest.
Linc Derick was a fighter pilot in World War II and was hard as
nails. His father, if I recall correctly, was a chemist and Linc
must have picked up a lot from his dad. He was a meticulous
experimentalist. When I became a supervisor in the development
area at Bell Labs I was told I could recruit individuals to join me
in moving from the research area. The first one I chose was Linc
Derick. He and Frosch had become experts in the growth of
gallium phosphide and I put Linc in charge of the effort to supply
gallium phosphide for our device area to use in their LED work.
He did a great job.
Linc also was a golfing buddy and we played in many Bell Labs
tournaments together. In addition, he had a boat and loved
diving and exploring underwater wrecks. He was also an avid
skier and somehow talked me into buying a pair of ski boots. I
only used them once, giving up skiing after finding myself
sailing downhill completely out of control at a New Jersey
venue. Our StocksandNews editor, Brian Trumbore, also will
not forget Linc, who offered to take Brian skiing when Brian was
still in high school. That was the last skiing trip for Brian. Linc
took him up to the top of the expert trail at Great Gorge here in
Jersey. Brian came home completely soaked and permanently
disinterested in ever going up another lift, having fallen his way
down the long trail.
Linc was a man of many interests. As a hobby, he reclaimed
gold from various sources, making exotic gold rings, some
copies of Egyptian jewelry he had seen. Just within the past
month or so, I saw in the Star-Ledger that Linc’s widow had
died. Time goes on. If I had the capability of adding pictures, I
would finish this column with a photo of a transparency Linc
gave me. Carl Frosch used this photo in his first talk describing
oxide-masking. The photo is of an oxide-coated silicon wafer
with the oxide layer etched in a pattern spelling “THE END”.
Far from being the end, oxide-masking was the beginning of an
era of electronic marvels that continues today.
Allen F. Bortrum