L'ONDE ÉLECTRIQUE

THE TRANSISTRON TRIODE TYPE P.T.T. 601 BY R. SUEUR Chief Engineer P.T.T. Head of the Department of Service des Recherches et du Contrôle Techniques P.T.T. Translation into English Copyright Mark P D Burgess March 2011

On Wednesday, May 18, 1949, the Minister of P.T.T. presided over the presentation of the Transistron triode P.T.T. 601 and some instruments equipped with this device at the laboratories of Service des Recherches et du Contrôle Techniques (S.R.C.T.) of P.T.T. It was similar to a presentation held in America at Bell Telephone Laboratories in 1948. Work on semiconductors conducted in France in recent years in collaboration between the Administration des PTT and the Société des Freins et Signaux Westinghouse has produced similar results to those of the Americans. Building on previous work, Doctors Welker and Matare and a team of researchers prepared germanium of high resistivity and started manufacturing high back voltage detectors, a prelude to the development of the of germanium triode or Transistron triode. During the same year the first germanium Transistrons manufactured in France left the Laboratories. In French we could call this device “transistance” which is the literal translation of the American term “transistor.” However transistance in French would be like resistance, an electrical quantity. Thus we have the name “Transistron”, or resistance of transfer, the suffix “tron” indicating active elements involving electrons or ions.

1o Semiconductors At room temperature, solids can be divided into three classes according to their electric conductivity. - Conductors - Insulators - Semiconductors The phenomenon of conductivity is related to the electronic organization of atoms of the material being considered, the organization of its crystal lattice and its physical crystalline imperfections and chemical properties. We know that the latest theories on the constitution of matter shows the atom consists of a central nucleus and electrons with fixed energy levels. Of these electrons, we distinguish two kinds: electrons bound to atoms and free electrons. The energy of a bound electron is insufficient for it to separate from its atom and its energy is quantized. The energy of a free electron is sufficient for it to be separated from its atom and its energy is not quantized.

According to the Pauli exclusion principle which states that “In an atom there cannot be two electrons defined by the same quantum coordinates” (1) it is not possible to have several electrons with the same energy in an atom. We say that each electron has a defined “energy level” and that energy level can vary by a quantum jump under the influence of X-rays, for example. We know that this quantum energy is equal to dw dw = h v where h is Planck's universal constant equal

Fig. 1 Unit cell of Ge. All other Projection on a horizontal atoms in the network are plane, the atoms numinferred from these by bered in the unit cell. translations ha1 + ka2 + la3

to 6.55 10-27 erg sec. and v is the frequency of the electromagnetic wave radiated in the quantum jump of the electron. Germanium is the 32nd element in the Mendeleev table normally with 32 electrons around its nucleus. The atoms form a solid body in a simple arrangement known as a three dimensional crystal lattice. Fig. 1 gives the example of the theoretical crystal lattice for germanium. In the crystal lattice of a particular solid there may be free lattice electrons liberated from atoms that are chemical impurities or from atoms of the material itself occupying energy levels. In a structure without impurities or physical defects energy levels are grouped into bands and each band

(1) There are four quantum coordinates: n, l, m, s n: characterizes the position of the electron, between 1 and 7 l: is related to the momentum m: is the magnetic quantum number s: is the quantum number that characterizes the spin

L'ONDE ÉLECTRIQUE

can generally be understood to have a maximum of twice as many levels as there are atoms in the crystal lattice according to the Pauli exclusion principle. Several bands may exist and they are separated by regions called "forbidden bands" where there cannot be any electrons. A full band has all levels occupied by electrons and an empty band has no electrons. The higher energy bands are occupied by electrons of high kinetic energy (Fig. 2). At room temperature a solid conductor has its higher energy bands partially filled with electrons, thus an electric field applied to the conductor easily causes a change of electron energy levels in this band and this explains the high conductivity of con-

ductors. At absolute zero they still have electrons in the upper band. On the other hand insulators have no electrons in their upper bands, their lower bands are filled and the forbidden bands may be several electron volts. A very intense electric field can only move the low-energy electrons from the lower bands with great difficulty. The electrons would have to move from one energy band to another and this is very unlikely. Thus the conductivity of insulators is very low. At absolute zero, there are no electrons in the upper band. Semiconductors are intermediary materials, their higher energy band is empty and lower band is filled. Between these the bands are partially filled, and unlike insulators the energy difference between the empty band above and the full band below is quite low. At room temperature, they have very low conductivity. At absolute zero they are insulators and their temperature coefficient is negative. There are two types of semiconductors. The first is known as "intrinsic" such as pure germanium and the second is known as "extrinsic" and is the result of physical defects or chemical impurities in the crystal structure of intrinsic semiconductors. According to their type these impurities may receive or donate electrons in the semiconductor and they create additional energy levels that increase the conductivity.

The very low conductivity of intrinsic semiconductors at room temperature (10 ~ 2 mho / cm for germanium) make them unsuitable for practical applications. We say an extrinsic semiconductor is N or P type depending on whether the impurities add or donate electrons in the crystal lattice and it appears to be a function of the chemical valency of the intrinsic semiconductor relative to the impurities. Thus phosphorus and antimony produce an Ntype germanium extrinsic semiconductor and boron and aluminum make a P-type silicon extrinsic semiconductor. Copper oxide Cu2O is a P-type extrinsic semiconductor. When impurity atoms are inserted in an intrinsic semiconductor two cases may occur: -The valency of the impurities is less than that of the intrinsic semiconductor. In this case, if an impurity atom takes the place of an atom in the crystal lattice, one or more electrons in the semiconductor can simply fill the vacant bond leaving a hole in the band they leave. This hole can be treated as a charge of equal and opposite sign to the electron. It causes the appearance of an energy level related to the corresponding band. -The valency of the impurities is greater than that of the intrinsic semiconductor. In this case the impurity atoms promote electrons in the permitted bands and they are found to have energy levels in the band gap. These respective energy levels are located close to the boundary (0.1 eV.) of the normal bands and increase the conductivity of the body. In the first case conduction is caused by pseudo electrons, as though positive electrons were involved and in the second case the conduction is electronic.

The type and quality of a semiconductor is conveniently determined by the Hall effect. A strip of material M simultaneously subjected to the influence of a magnetic field H normal to its thickness e and a longitudinal electrical current I causes a transverse emf E (Fig. 3). For a body at temperature θ the quantities H, e, I, E are connected by the expression:

L'ONDE ÉLECTRIQUE where R is called the “Hall coefficient” named after the physicist Hall who demonstrated this effect in 1879. With the direction given to the current I and the field H, the emf E can appear positively or negatively oriented. This sign has a direct impact on the coefficient of R and depends on the material subjected to the experiment. Electronic conductive bodies have a negative Hall coefficient, (N type); those conductive by holes have a positive coefficient (Ptype). In general, germanium at room temperature has electronic conductivity. The value of R can be between 10-7 and about 10-4 for germanium. The origin of the Hall Effect can be explained by the deviation of the trajectories of free electrons liberated by the electric current by the magnetic field. The sign of the Hall coefficient enables the determination of the type of conductivity of the material and its magnitude: - The number of conducting electrons per unit volume. - The mean free path of electrons. - Their mobility. The conductivity, σ, of a semiconductor with conductivity due to electrons or holes is given by the expression:

In addition the measurement of σ and R allows the calculation of b:

and consequently that of l. Knowledge of these different quantities is essential to define the best properties of semiconductors, the existing methods of chemical and spectrographic analysis being insufficient to make an adequate determination of impurities. In addition, defects in the crystal structure have a huge influence on the behavior of semiconductors and these defects have an impact on their physical properties, particularly on mobility. For different kinds of French N type germanium

at room temperature, typical data is given in the following table:

where: e = charge of an electron l = mean free path of electrons or holes N = number of liberated electrons or holes per unit volume m = electron mass k = Boltzmann constant T = absolute temperature. The mobility, b, of a stream of electrons or holes is defined as the speed of the electron or hole in a unit electric field.

where:

The height of the barrier layers of a rectifier is more pronounced when the mobility is lower. The theory of solid bodies permits us to write:

Figure 4 shows equipment for measuring the Hall coefficient. Germanium was discovered in 1886 by the German chemist Winkler.

Thus measurement the Hall coefficient, R, allows the calculation of N

It is available from several sources: a) Naturally, from germanite, pyrite

L'ONDE ÉLECTRIQUE comprising 30 to 40% copper and 1 to 4% germanium, the richest deposits are located in South-West Africa near Tsumeb. They are in the form of pockets that seem to accompany deposits of zircon. It is also found in some argyrodites. Pyrite containing up to 1% in deposits have been found near Freiberg (Saxony). It is also contained in some sulfur coals, such as those of Durham. b) As a metallurgical by product from the processing of zinc and cadmium. The crude germanium does not generally have a suitable chemical composition nor a homogeneous structure.

Strong ionisation (n2 ≈ N1)

TYPE P Weak ionisation (n2 ≤ N2)

Strong ionisation (n2 ≈ N2)

where: v = is a function of (h, k, T, n); e = Electron charge; ʋ = Contact potential difference; xs =Work function of the semiconductor; xm = Work function of metal; k = Boltzmann constant; T = Absolute temperature; It still must undergo physical and chemical treatment to give suitable P or N type characteristics and resistivity appropriate to the required end use. Figure 5 shows a germanium processing facility.

2° Contact between semiconductor.

a

conductor

and

a

It is known that two different materials in contact with one another create a potential difference in the vicinity of the point of contact called the "contact potential difference" related to the difference in the work function of electrons in each material. The work function is the work done to remove an electron from the body and make it free. The work function xn in a extrinsic semiconductor depends on the level of impurities found there. The difference in work function between a metal and a semiconductor is about 0.2 to 0.5 eV. Formulas for the difference in work function were given by Fowler for different cases. For an extrinsic semiconductor that is:

TYPE N Weak ionisation (n1 ≤ N1)

N1 = Number of donator energy levels per unit volume, located a level below the band gap. n1= number of electrons excited to the empty band per unit volume. N2 = number of acceptor energy levels per unit volume and located at a certain energy level above the full band. n2 = number of free holes [in the full band] per unit volume. The difference in contact potential causes the emergence of barrier potentials located in the vicinity of the surfaces in contact. Between these barriers there is an area known as insulating barrier layer with high dielectric constant. Explanations on the formation and existence of barrier potentials have been given by many authors (Schottky, Mott etc). They refer to the difference of work function that is to say the difference in energy levels that may exist between the two bodies involved. Following the conventional representation, Figure 6 indicates the position of the bands and energy levels that exist in the vicinity of the surface of a metal and an N-type semiconductor before and after contact. In the semiconductor energy levels of electrons created by the impurities are located around the normally empty upper band and they are higher than the upper band, of the conductor. When contact is made it is found that the electrical current flows more easily in the direction

L'ONDE ÉLECTRIQUE of "conductor to semiconductor” than in the opposite direction. At the moment of contact we can say that because of the difference in the energy of the electrons from each material involved, the electrons of the semiconductor migrate to the conductor

In a P-type extrinsic semiconductor equilibrium is established differently. (Fig. 7). Upon contact, the holes arising from the partially filled band flow from the semiconductor to the conductor, creating a positive barrier and allowing a negative barrier to form in the semiconductor. The overall conductivity is not electronic and it can be seen that the easy flow of electric current is from semiconductor to conductor. [See translator’s notes] 3° Operation of the Transistron (Fig. 8) N-type high resistivity germanium was polished and then etched with acid to expose a suitable crystalline surface structure which was then chemically treated to make (or enhance) a P type semiconductor layer. Wire point contacts of bronze, tungsten or molybdenum were then positioned on this surface.

allowing a positive surface layer to form on the surface of the semiconductor while a negative layer forms in the conductor; a double potential barrier is thus formed, and between them appears the barrier layer. An electric potential difference moves the highenergy electrons more easily from the semiconductor to the conductor and direct current flow in the easy direction is observed. This overview of the probable mechanism of formation of potential barriers and barrier layers does not give an accurate view and the theory is incomplete. But note that the difference in conductivity caused by the direction of flow of an electric current in a contact between a conductor and a semiconductor is exploited in detectors and is one of physical phenomena seen in the Transistron.

The most recent solid state theories and experimental results indicate that the barrier layer has a thickness of about 10-4 mm.

It is found that the forward direction of current flow is observed for N-type semiconductors. Then apply an electric current (I) of 1-2 milliamps in the forward direction that is in the direction "from the contacts towards the semiconductor." It appears that the electrical conductivity around the point contact depends on the value of I (current). This is the second phenomenon called transistance used in the Transistron.

If we then place a second contact on this surface at a distance d from 20 to 50 microns from the first, we see that the resistance R measured between the two points depends on I according to a law given qualitatively in Figure 9. We also know that an amplifier is essentially an energy valve where the energy ratio Ws/Wc between energy output Ws released by a valve and the input and control energy, Wc, is larger than unity. Now arrange the assembly of Figure 10 where

L'ONDE ÉLECTRIQUE the direct current IE is of the order of a milliampere, the variable resistor r a few hundred ohms, a battery p of voltage e of 1-2 volts. Point “a” is the control point and the resistance of the layer is around a few tens of thousands of ohms.

A very high voltage E (50 to 100 volts) is applied through a resistance R (20,000 to 30,000 ohms) to the point “c” in the reverse direction so the resistance of the semiconductor is very large. We note that the current Is depends on IE and the current Is must certainly flow in the layer between the points "a" and "c". The valve is then represented by the resistance of the layer, it is controlled by the current IE and it is found that the control power required is much less than that which appears in the output resistance, Rs. The magnitude of the input to output power ratio given by

can now reach 100 to 200 and output power Ws is of the order of several tens of milliwatts. Explanations or physical phenomenon of variation of surface conductivity attributed to the P layer and the bulk N semiconductor as a function of

the polarization of the control electrode are provided by Bardeen and Brattain (see bibliography) and we refer the reader to articles of these authors. The calculations and tests conducted in France so far seem to confirm the role of a layer where the conductivity occurs by pseudo electrons [holes]. Their verification require the implementation of very difficult procedures, particularly the precise measurements of interelectrode capacity. An electrical equivalent circuit most convenient in our opinion for the Transistron follows (Fig. 11). It results from the mathematical analysis of the operation of the Transistron. One can propose the following equations where the parameters correspond to the four elements of the equivalent circuit.

Where we define: RE = input resistance, Rr = feedback resistance between the output and input Rc= coupling resistance which loads the voltage gain of the circuit Rs = output resistance. These two equations allow us to establish the following equivalent circuit (Fig. 12). Where:

where by applying the Thevenin Theorem we readily obtain from the schematic in Figure 11: R1 = R E - Rr R2 = R r R3= Rs - Rr Es = (Rc- Rr) dIE The family of curves in Figure 13 found for a type 601 Transistron shows, as might be expected, (see fig. 9) that Is increases with IE.

Furthermore the voltage Er = Rr dIs is the load voltage.

L'ONDE ÉLECTRIQUE We can now establish a convenient method for calculating the various characteristics of an amplifier in terms of the previously defined variables. Place the Transistron between an EMF generator where e = E(sinωt) for example and input impedance RG and output impedance equal to RR. Figure14 shows the equivalent circuit in this case excluding the bias circuits.

For convenience of calculation we put:

For RG ≠ ZE the first term is negative and the second for RE ≠ Rs, they are equal to zero for: R G = ZE RE = R s For good power gain the source and output impedance should be matched to ZE and Rs. respectively and the gain will be:

Around an operating point identified in Figure 13 and defined by UE ≈ + 0.35 volt Us ≈ -45 volts We find for the Transistron 601 the following key characteristics: RE ≈ 170 ohms Rr ≈ 70 ohms Rc ≈ 30,000 ohms Rs ≈ 20,000 ohms

And then we find:

Which gives the equivalent circuit in figure 15.

a) Input impedance ZE ZE = RE (1- μβ) b) Ouput impedance Zs Zs = Rs + Rr (1- μoβ) ≈ Rs c) The voltage gain in Nepers

If RR is very large

d) The overall gain in Nepers

We find for RR = Rs and RG = ZE μo = 30,000/170 = 176 μ = 176/2 = 88 β = 3.5 10-3 μβ = 0.308 ZE = 170 (1 – 0.308) ≈ 188 ohms Zs = 20,000 ohms Gt ≈ 5.5 Nepers GoM ≈ 5.15 – 2.38 – 0.1 ≈ 2.7 Nepers ≈ 24.5 dB We note in passing that Transistron is a very good voltage amplifier. Circuits for measuring Transistron gain are easily deduced from the definitions. The overall stage gain of the Transistron is related to the input and output transformers.

L'ONDE ÉLECTRIQUE 4° Making Transistrons. The key to the production Transistrons lies in the preparation of germanium, in the selection of bars where pellets should be cut, in the search for points of contact and optimal adjustment of the spacing of the point contacts. These last two operations are carried out moreover under the microscope and are made easier in the type 601 by the mechanical arrangements that are used (Fig. 16).

Legend Fig 16 the ceramic

-A 4 Transistron television video amplifier bandwidth. 40-10000 p/s, gain of 5.2 Neper and 20 milliwatts of output power.

Ceramic body ... ... ..Soldered bronze caps with

The two bronze rods a and c are connected to a point contact P of tungsten wire. The rod b supports the Ge pellet. Each rod slides in a bronze cap and its position can be fixed by a screw. The set of three rods can be adjusted while checking the electrical characteristics on the surface of the pellet: a and c can slide laterally and b can be adjusted by translation and rotation. The translation of b allows such precise control of the spacing of the point contacts.

-A medium range telephone line repeater (Fig. 19); -A long distance telephone repeater for a 4 wire loaded circuit.

This latter unit will be inaugurated in Paris on a Paris-Nancy circuit and its schematic is given in figure 20.

A pilot production run has already been completed enabling the study of methods of manufacturing and control. From November the Westinghouse Company will manufacture sufficient quantities for the state agencies that sponsored the research at the company. Figure 17 shows a photograph of Transistron type 601.

Its maximum gain is equal to 3.5 Neper and available output power is 15 milliwatts. Total energy consumption is about 0.9 watt per direction of amplification. A pentode repeater of guaranteed 10,000 hours life consumes 4 watts per direction.

5° Transistron Applications. Of test equipment currently operating in the laboratories of S.R.C.T. there is: -A broadcast receiver (Fig. 18); -A transmitter for 300 metres longwave.

Due to the simplicity of the possible source of polarization of its electrodes and low consumption, Transistrons are conveniently powered remotely via the telephone line they are installed on. To the extent that the life-time projections of this equipment are borne out they will reduce the cost

L'ONDE ÉLECTRIQUE of telephone lines by using more amplifiers on thinner lines reducing raw materials usage. Figure 21 shows the diagram of the amplifier remotely powered with two push-pull Transistrons shown in Figure 19.

The S.R.C.T. Laboratories current applications research is particularly directed to applying the Transistron for telephone circuit electronics targeting the simplification of equipment, reducing the hardware footprint, increasing the security of service and reducing the annual costs of the circuits. The team conducting applications research on semiconductors include: For the Administration of P.T.T.: MM. JOB, Ingénieur des P. T. T. MOLL, Ingénieur Contractuel CHALHOUB, Ingénieur Contractuel PERINET, Inspecteur des I. E. M. GANET, Inspecteur des I. E. M. LE FLOCH, Contrôleur des I E. M. VALIÉNET, Contrôleur des I. E. M. COULON, Contrôleur des I. E. M. POTET, Contrôleur des I. E. M. For the Westinghouse Company:

MM. WELKER, Docteur Physicien MATARE, Docteur Physicien. PETIT-LEDU, Physicien BETHGE, Ingénieur. POILLEAUX, Technologiste. CALON, Technologiste. PHILIPPOTEAUX, Technologiste.

The authors particularly thank M ENGEL, Technical Director of the Westinghouse Company, for the valuable assistance he gave in all circumstances during the work on the Transistron. Refer Translator’s Notes overleaf

REFERENCES Atomistique et Chimie Générale, by R RENAULT. Dunod, Paris. Modern Theory of Solids by SEITZ - McGraw Hill, N. Y. and London. Electronic Processes in Ionic Crystals by N. F. MOTT and RW. GURNEY - Oxford University, London. Crystal Rectifiers, by TORREY and WHITMER - M.I.T., McGraw Hill N.Y. and London. Microwave Mixers by POUND - M.I.T. Mc. Graw Hill N.Y. and London. Die Elektronenleitung des Kupferoxyduls by W. SCHOTTKY and F. WAIBEL, Physikalische Zeitschrift N ° 23, 1933, (Translation CNET N° 38). Uber die Elektriche leit fahrigkeit des kupfer oxyduls im gleichgewicht mit weinen nachbarphalen by F. WAIBEL. Zeitchrift für technische Physik N° 11, 1935 (Translation CNET No 428). Détecteurs à Pyrite pour ondes décimétriques by H. WELKER - (Translation N° 5441 du Ministère de l'armement S.E.F.T.). Schottky's Theories of dry solid Rectifiers by JOFFE Electrical Communications Vol. 22, N° 3, 1945. Electrical resistance of the contact between a semiconductor and a metal, by JOFFE,- J. Phys. U. R. S. S., Vol. 10, N° 1, 1946 (Translation CNET, N° 622). Sur l’intérêt et les possibilités d'application des semiconducteurs électroniques dans la technique des hautes fréquences by M. TEZNER - Note Technique CNET, N° 1047. Note relative aux redresseurs à contact ponctuel sur semiconducteur - Note S.R.C.T. - Département Transmission 1010-48. Le courant électrique, le photon et l'électron par M. G. POCHOLLE, Ingénieur en Chef des P,T.T. - Bulletin de Documentation du Secrétariat aux Forces armées «Guerre ». The Transistor a crystal triode, by D. G. F. and F. H. R., Electronics, September 1948. The Transistor A semi-Conductor Triode, by J BARDEEN and W. H. BRATTAIN, the Physical Review, July 15, 1948. Nature of the Forward Current in Germanium point Contacts, by W. H. BRATTAIN and J. BARDEEN, The Physical Review, July 15, 1948. Modulation of Conductance of Thin Films of SemiConductors by Surface Charges, by W. SHOCKLEY and G.L. PEARSON, The Physical Review, July 15, 1948. Les détecteurs à Germanium by R. SUEUR, Information Technique - janvier-février 1949. Germanium, important new Semiconductor, by Dr. W. Crawford DUNLAP VR, General Electric Review, February 1949. Temperature Dependence of the Work Fuction. of Semiconductors by A. H. SMITH, Physical Review, 15 March 1949. The Effect of Surface States on the Temperature Variation, of the Work Function of Semiconductors by Jordan, J. MARKHAM and PH MILLER Jr. The Type-A Transistor by RM. RYDER, Bell Laboratories Record. March 1949. Some Novel Circuits for the Three 'Terminals Semiconductor amplifier by W. M. WEBSTER, E. EBERHARD and L.E. BARTON, R.C.A. Review, March 1949. Physical Principles Involved in Transistor Action by J. BARDEEN and W. H. BRATTAIN, Physical Review 15 April 1949.

L'ONDE ÉLECTRIQUE Translators’ Notes

(3) But these contemporary sources show Transistrons without a cap:

The available copies of the original publication are only available in relatively low resolution and in particular the originals of Figures 18 and 19 are poorly reproduced. In this facsimile Figure 18 has been copied from an identical picture in Aberdam 1949 (courtesy Christian Adam) and Figure 19 has been copied from Aisberg 1949.

Early Versions of the Transistron 601 There are three versions of the Transistron 601: 1. 2. 3.

Three adjustable stems One adjustable stem and a window cap One adjustable stem and no window cap

Pictures above from Aberdam 1949 courtesy Christian Adam.

(1) In his text Sueur describes what we presume to be an early version in which the stems that hold the emitter and collector point-contacts can slide and rotate in their end-caps for the purposes of adjustment and then fixed in place with a grub screw. The crystal holder is equipped with the same facility. The pictures of the Transistron clearly show this arrangement for the crystal holder but not for the emitter and collector indicating that early in 1949 the method of production was simplified. Contemporary pictures also show two versions of this Transistron.

(2) In figure 17 a Transistron is shown with a cap placed over the window used to adjust the point contacts on the crystal. This version of the Transistron is used in the broadcast receiver and the telephone circuit repeater (figs 18 and 19) and other equipment such as the long wave transmitter shown in Aberdam 1949:

Above: Publicity picture of the Transistron on its official release. [Aisberg 1949] Forward and Reverse Bias The treatment by Sueur is somewhat confusing in relation to point contacts on N or P germanium. The following provides a consistent explanation from Pfann 1950. Conductivity Type of the Semiconductor N-Type P-Type

Polarity of Point Forward Reverse Direction Direction + +

Thus the bias arrangements for each case is given as follows where the emitter is forward biased and the collector is reverse biased:

L'ONDE ÉLECTRIQUE References Aberdam H 1949 Transistor et Transistron Ingénieurs et Techniciens 12 213-18 Aisberg E 1949 Transistron = Transistor+ ? Toute la Radio 137 218-20 Pfann W Scaff J 1950 The P-Germanium Transistor Proc IRE 38 1151-54