Roger Baker:
"Recently," Baker writes, "I learned of a governing principle known as Murphy's first law of biology. It states: 'Under any given set of environmental conditions an experimental animal behaves as it damn well pleases.' The same law appears to govern the behavior of thin metallic films, at least those I make. Some of my 'transistors' make dandy thermistors, and an occasional photocell works better as a fluorescent screen. There might be fewer surprises if I used better tools and had more experience, but some of the fun might also be lost. The techniques used by industry for making thin films are not beyond the reach of amateurs, but they require vacuum pumps, electronic heating device and controlled sources of high voltage that are costly and inconvenient to use. Thin films can also be deposited chemically. I use this method.
Roger Baker's apparatus for depositing metallic films on glass
"Most of my films are deposited on substrates of glass. Usually I heat the glass and spray the surface with a solution of selected chemicals. The sprays react immediately to form the film.
"Films can be annealed in various atmospheres and at various temperatures that alter their composition, structure and properties. The properties of a deposited film can also be modified by recrystallization, by solid-state diffusion or by a vapor-phase displacement reaction. These procedures are much simpler to perform than their imposing names suggest. The properties of films can be radically altered by the addition of minute amounts of impurities, either when they are formed or by subsequent diffusion.
The microstructure of the substrate can also influence the properties of a film. For example, calcium sulfide forms an amorphous film when sprayed on a metal surface, but on glass it becomes a crystalline film.
"The required tools include an electric hot plate, a diamond point for cutting thin glass, an atomizer and a microammeter. Desirable accessories are a fume hood, which can be improvised if you have an exhaust fan, an oven thermometer for measuring the temperature of the hot plate, a triple-beam chemical balance, a pair of tweezers and chemical glassware for preparing solutions. For substrates I use mostly cover glasses of 35-millimeter Kodak slides, which I cut into small rectangles with the diamond point. These glasses can be heated (up to 600 degrees Celsius) and sprayed without breaking. Thin disks of alumina can be used at higher temperatures. I salvage them from discarded vacuum tubes. I immerse the glass slides for three days in a solution of one part by volume of nitric acid to 12 parts of distilled water. The acid leaches sodium and calcium ions from the glass, exposing a surface layer of relatively pure silica.
"A great variety of semiconducting oxide films can be made by thermally decomposing the resinate salts of metals. Resinate salts are prepared by stirring an excess of pure granulated resin into a one-normal (1 N) solution of sodium hydroxide. The solution turns milky as it cools. Pour off and retain the milky solution. To make a metal resinate, reheat the milky sodium-resinate solution and combine it with a weak solution of the metal salt, stirring the mixture vigorously.
"A relatively large volume of sodium resinate reacts with a small volume of metal salt. An excess of sodium is indicated by a pH of 8 or more. Add metal salt to lower the pH. The desired metal resinate appears as a thick precipitate.
"Filter the solution to recover the precipitate and wash it thoroughly with hot distilled water. Spread the moist filter cake and dry it at a temperature of about 50 degrees C. Dissolve the dried material in an organic solvent such as turpentine. Allow the sediment to settle. Use the clear upper layer for experiments.
"With a disposable capillary tube apply a few drops of the clear fluid to the center of a prepared cover glass and rock the glass to spread the fluid into a uniform film that extends to the edges. Heat the coated glass on the hot plate. The film will smoke and turn dark. In time, at a temperature that depends on the nature of the resinate, the dark color will clear, leaving a thin film of metallic oxide. The cover glass can then be scribed with the diamond point and broken into rectangles of convenient size for further processing and experimentation.
"Sulfide films can be formed directly from a number of oxide films. Sprinkle a few milligrams of sulfur on the back side of the coated substrate, wrap it with several layers of aluminum foil, fold the ends of the foil over the package and heat the package. The hot vapor of the sulfur will react with many oxides to form adherent films with interesting electrical properties. Two drops of different resinates can be allowed to diffuse partially together so that the properties of various ratios of the two can be explored.
"So far I have experimented with gold, nickel, cobalt, copper, iron, manganese, silver, indium, chromium, zinc and cadmium resinates. The salts of noble metals decompose into metallic films instead of oxides. They can be used for making electrical connections between various films of oxide previously applied to a substrate.
"The preparation of a field-effect transistor illustrates a typical experimental procedure. The substrate, which has been treated with nitric acid solution, is first coated with a film of cadmium sulfide. With distilled water prepare 500 milliliters of a stock solution containing .01 molar (.01 M) thiourea and .01 M cadmium chloride.
"Place the substrate inside a 250-milliliter beaker so that it rests diagonally against the side of the beaker. Cover the substrate with stock solution and slowly add concentrated ammonium hydroxide until the mixture turns faintly cloudy and then clears. Cover the beaker and put it in a double boiler. Heat the vessel slowly and boil for about 15 minutes. The contents of the beaker will turn yellow-orange, indicating the precipitation of cadmium sulfide.
'Pour off the contents and replace them with distilled water. Swab the substrate lightly with a tuft of absorbent cotton to remove adhering particles of cadmium sulfide. Rinse the substrate with distilled water. Repeat the entire procedure to double the thickness of the film, after which you can clean the beaker with hydrochloric acid.
"Bake the substrate in air at 500 degrees C. for 30 minutes. The color of the hot substrate will gradually change from yellow to red and, as it cools, to a deeper shade of orange. With the diamond point scribe and break the cooled glass into rectangular chips 1/4 inch wide and 1/2 inch long.
"The transistor requires two contacts that function as electrodes, one a source and the other a drain. The electrodes are conveniently made of indium, a soft metal that can be pressed into firm contact with the film. Indium is available from dealers in chemicals. Place a small pellet of indium on clean plate glass and roll it into thin foil with a short length of clean glass tubing. Transfer the foil to a yielding surface such as glossy white cardboard and, by pressing straight down with a sharp razor blade, cut the metal into strips about 1/32 inch wide and 1/4 inch long.
Sequential steps in making a thin-film transistor
"With a sewing needle maneuver two of the strips to a clear portion of the paper so that they are parallel and spaced about 1/16 inch apart. Lay one of the coated chips over the strips so that the ends of the strips are even with one end of the chip. Press the chip firmly and evenly against the metal. The strips will adhere lightly to the film. Burnish them firmly into place by turning the chip strip side up on the plate glass, covering it with a glossy magazine cover and rubbing it with a fingernail. Place a small dab of conductive silver paste on the outer end of each indium strip [see illustration at right]. The dabs serve as terminals for connecting the electrode device to a power source.
"A layer of insulation is applied to the film and the indium strips in preparation for adding the third electrode, which is known as the gate. With a sewing needle apply a thin, uniform layer of vinyl cement by stroking the cement across the upper surface of the device. Do not coat the silver terminals. When the cement dries, apply a coat of silver paste over the insulation. Do not let the silver make contact with the cadmium sulfide film, the indium foils or the source and drain terminals. This completes the gate electrode.
"Finally, to protect the active region of the device coat the upper surface with a layer of silicone rubber that cures in air. This material is available from dealers in hardware. Leave one small region of the gate electrode exposed. This small area will be used for making electrical contact with the gate. Do not coat the source and drain terminals with rubber.
"To operate the device improvise a test fixture such as the one shown in the accompanying illustration [left] for holding the transistor and connecting it to a battery. Power is applied to the source and drain electrodes by a ninevolt transistor battery that is connected in series with a 10,000-ohm resistor and a 0-50 microammeter. If the transistor is reasonably good, the meter will indicate a current of about 10 microamperes. This is called the leakage current.
"Connect a one-megohm resistor between the gate electrode and the positive terminal of the battery. The positively charged gate will attract free carrier electrons into the cadmium sulfide film. Current through the film should rise to about 50 microamperes, indicating that the transistor is a so-called N-channel device and that it is operating in the enhancement mode. The gate electrode draws little current.
"If negative charge is now applied to the gate by transferring the one-megohm resistor to the negative terminal of the battery, current in the source-drain circuit should fall below 10 microamperes. The transistor is now operating in the depletion mode. I do not know why some homemade transistors work better than others. I suspect that their performance may be related to the crystalline structure of the films.
"Capacitors can be made by sandwiching insulation between films, resistors by etching away portions of film to form narrow conducting paths, photocells by doping cadmium sulfide with trace amounts of silver, copper or manganese. Films of zinc sulfide fluoresce strongly. Of course, devices are available on the market that work better than those one can make at home, but mine are better playthings.
"Certain hazards must be mentioned. Metallic salts and acids are toxic. Work either in a fume hood or outdoors when you spray chemicals onto a hot substrate. Wear gloves and an apron of neoprene when you handle acids. Remember that chemicals are hazardous and handle them accordingly."
Bibliography
THIN FILM MICROELECTRONICS: THE PREPARATION AND PROPERTIES OF COMPONENTS AND CIRCUIT ARRAYS. Edited by L. Holland. John Wiley & Sons Inc., 1965.
M-x grep-find in dired. Advance through matches with C-x `
In GNU Emacs:
1] M-x octave mode
2] Use C-c TAB C-l to evaluate the last Octave statement (eg. help)
[also used to start the inferior Octave process which we can interact with in the other window/frame.]
3] Ctrl-up-arrow/down-arrow to cycle through history at Octave prompt.
4] Opening a file as raw 8 bit character (unsigned) and plotting a histogram of values:
myfile = fopen("testrnd_delay_none", "r+")
x =fread(myfile, "uchar");
hist(x)
hist(x(1:50))
plot(x)
fclose(myfile)
3] Help/info issues:
At first there were some problems with help/GNU info in Emacs with Octave. These were solved with:
(require 'octave-hlp) (in Emacs) and apt-get install octave3.0-info