to have a positive dn/dT and is considered to have large homopolar bonding, similar reasoning predicts that dn/dp is positive; i.e., Xo>l. In crystals containing radicals and in many glasses, positive dn/dT values are frequently obtained although their dn/dp values are negative. In these materials there are effects within the radical which contribute mainly to dn/dT and only slightly to dn/dp. A more complete treatment of these subjects will be presented in a forthcoming paper.i reported -h(Pu-Pn) for two mixed thallium halides in agreement with our results. Their data also gives a -\-{px\-pn) and -pa, for AgCl in agreement with Mueller's prediction for NaCl structures with small ratio of negative to positive ion polarizibilities. We do not agree with West and Makas concerning the sign of the diamond constants and believe them to be correct as given by Ramachandran. 3

The contact potential (c.p.) and the change of contact potential with illumination (Δc.p.)L, of several germanium surfaces have been measured. The reference electrode used was platinum. It was found that the c.p. could be cycled between two extremes about 0.5 volts apart by changing the gaseous ambient. Ozone or peroxide vapors gave the c.p. extreme corresponding to the largest dipole at the Ge surface. Vapors with OH radicals produced the other extreme. There is a one to one correlation between c.p. and (Δc.p.)L. For 12‐ohm cm n‐type Ge (Δc.p.)L was large and positive when the surface dipole was largest, decreased to zero and became slightly negative as the surface dipole decreased to its smallest value. The variation for 12‐ohm cm p‐type Ge was just opposite as regards both sign and dependence on surface dipole. The surface recombination velocity was found to be independent of c.p. For a chemically prepared surface it was 50–70 cm/sec and 180–200 cm/sec for n and p‐type surfaces respectively. A theory is given that explains the results in terms of surface traps, Na per cm2 donor‐type traps near the conduction band and Nb per cm2 acceptor‐type traps near the filled band. A quantitative fit with experiment is obtained with only one free parameter. The results are direct evidence for the existence of a space charge layer at the free surface of a semiconductor.

Measurements have been made of the electrode potential of p‐ and n‐type germanium in contact with aqueous solutions of KOH, KCl and HCl as a function of anodic and cathodic current and of incident light intensity. For anodic currents, the measured electrode potential can be separated into three parts: the reversible electrode potential corresponding to the anodic reaction, depending only on the solution; an overvoltage of the usual form; and a term (kT/e) In (p1/p), where p is the equilibrium hole concentration and p1 is the concentration just inside the space‐charge region of the germanium. The anodic current is determined by flow of holes to the surface, so that the current saturates for n‐type germanium but not for p‐type. The saturation current is determined by body and surface generation of holes and by creation of excess holes by light. There is a current gain of 1.4 to 1.8. In addition, there is a small “leakage” current not dependent on hole supply. Similar statements may be made for cathodic current, except that the electrode potential and current arc determined respectively by the concentration and supply of electrons instead of holes, the current gain is of the order of unity, and the leakage current is larger. Complicating time changes were observed for cathodic but not for anodic currents. The measurements may be understood in terms of simple thermodynamic considerations, based on the idea that the anodic reaction is with holes, the cathodic reaction with electrons, in the semiconductor; the behavior for very small currents depends on a competition between the anodic and cathodic reactions, which may be treated by simple rate process considerations. A comparison is made with experiments on the germanium‐gas interface by Brattain and Bardeen, to which similar considerations may apply.

The properties associated with the space-charge region and with surface states at a semiconductor surface are discussed. A theory of the space-charge region that takes into account charge-densities arising from immobile impurities and from both signs of mobile carrier is presented. The properties of the space-charge are discussed in terms of the surface potential and of the electrochemical potentials of holes and electrons, and related to the transport of added carriers in a homogeneous semiconductor. The change in surface conductivity arising from nonvanishing surface excesses of holes and electrons is treated. The space-charge systems at a free surface and at a p-n junction are compared, and the range of validity of the Mott-Schottky space-charge theory evaluated. The arrangement of surface states is discussed with reference to the Brattain-Bardeen model. Theories for the surface photoeRect and field-effect experiments are given, with and without surface states: it is concluded that the existence of surface states is without gross effect on the former, while relevant quantitative evidence from the latter is not yet available. The question of the relation between surface potential and contact potential is discussed. The properties of "channels" are discussed in terms of the theory. The paper concludes with a short section on long-time effects.

No abstract