Thepoint-contact transistor was the first type oftransistor to be successfully demonstrated.[1] It was developed by research scientistsJohn Bardeen andWalter Brattain atBell Laboratories in December 1947.[2][3] They worked in a group led by physicistWilliam Shockley. The group had been working together on experiments and theories ofelectric field effects insolid state materials, with the aim of replacingvacuum tubes with a smaller device that consumed less power.
The critical experiment, carried out on December 16, 1947, consisted of a block ofgermanium, asemiconductor, with two very closely spaced gold contacts held against it by a spring. Brattain attached a small strip of gold foil over the point of a plastic triangle—a configuration which is essentially a point-contactdiode. He then carefully sliced through the gold at the tip of the triangle. This produced two electrically isolated gold contacts very close to each other.
The piece of germanium used a surface layer with an excess ofelectrons. When an electric signal traveled in through the gold foil, it injectedelectron holes (points which lack electrons). This created a thin layer which had a scarcity of electrons.[4]
A small positivecurrent applied to one of the two contacts had an influence on the current which flowed between the other contact and the base upon which the block of germanium was mounted. In fact, a small change in the first contact current caused a greater change in the second contact current; thus it was anamplifier. The low-current input terminal into the point-contact transistor is the emitter, while the output high-current terminals are the base and collector. This differs from the later type ofbipolar junction transistor invented in 1951 that operates as transistors still do, with the low-current input terminal as thebase and the two high-current output terminals as the emitter and collector.
The point-contact transistor was commercialized and sold byWestern Electric and others but was eventually superseded by the bipolar junction transistor, which was easier to manufacture and more rugged. The point-contact transistor did still remain in production until the 1960s, by which time the siliconplanar transistor was dominating the market.
While point-contact transistors usually worked fine when the metal contacts were simply placed close together on the germanium base crystal, it was desirable to obtain as high an α current gain as possible.
To obtain a higher α current gain in a point-contact transistor, a brief high-current pulse was used to modify the properties of the collector point of contact, a technique called 'electrical forming'. Usually this was done by charging acapacitor of a specified value to a specified voltage then discharging it between the collector and the base electrodes. Forming had a significant failure rate, so many commercial encapsulated transistors had to be discarded. While the effects of forming were understood empirically, the exact physics of the process could never be adequately studied and thus no clear theory was ever developed to explain it or provide guidance on improving it.
Unlike later semiconductor devices, it was possible for an amateur to make a point-contact transistor, starting with agermanium point-contact diode as a source of material (even a burnt-out diode could be used; and the transistor could be re-formed if damaged, several times if necessary).[5]
Some characteristics of point-contact transistors differ from the slightly later junction transistors: