A similar change of structure can be effected in the diazocyanides by separating the "salt" from aqueous solutions and dissolving it in a non-ionizing solvent, when it behaves as a typical non-electrolyte, whereas in water it behaves just like the chloride

C2H ̧. ÑCN=CH ̧ . N:N.CN

Precisely similar considerations apply to the fixation of a hydroxyl-ion. Thus a solution of methylphenylacridinium chloride, when mixed with an equivalent quantity of sodium hydroxide, shows at first the normal conductivity of such a mixture; but this conductivity diminishes progressively to that of sodium chloride as the hydroxyl is "fixed" by the

The acridol, which is formed as a product, is a tertiary amine-alcohol, compare (CH3)2NMe and (CH),C.OH. It is therefore just as true a base as trimethylamine, but differs from other tertiary bases in that one of the alkyl radicles carries a hydroxyl-group, which is eliminated in the form of water as soon as the base has performed its characteristic function of accepting a proton from an acid. Although, therefore, it is commonly referred to as a pseudo-base,' it would perhaps be described more accurately as a "pseudoalkali," i. e., as a non-electrolytic isomer of a quaternary ammonium hydroxide.

Strong and Weak Acids.

66

Whilst the conception of a "strong electrolyte" is a very modern one, the existence of strong and weak acids has been familiar for 1000 years or more, and had already been interpreted in many other ways before the strength of an acid was finally correlated with its electrical conductivity in

dilute solutions. The theory of complete ionization therefore receives its most serious challenge from the existence of weak acids, which obey Ostwald's dilution law (a x c ̄1, where a is small), and cannot therefore also obey the dilution formula for strong electrolytes (1-a c, where a is large).

*

In order to account for the existence of these weak acids, Hantzsch has suggested that carboxylic acids can exist in two interconvertible isomeric forms, thus:

A dynamic isomerism of this type is actually observed in the nitro-paraffins,

or

CH,. CH,. NO, CH,. CH: N

CH3. CH: N H

OH

and is an obvious factor in reducing the acidity of the compound, which must be diminished in direct proportion to the extent to which the true acid is converted into the isomeric pseudo-acid. It is, however, by no means certain that the contrast between strong and weak acids depends in other cases on so definite a phenomenon. For instance, we need not suppose that phenol owes its weak acidity to conversion into an isomeric pseudo-phenol,

since it is sufficient to postulate that, whilst the -ONa radicle of sodium phenate is permanently ionized, the -OH radicle of phenol is held together by a real hond, just

*Ber. 1. p. 1438 (1917).

as in the case of the alcohols. Dissolution of the sodium compound in water would then give rise to a direct dissociation of ready-made ions as indicated by the equation,

whereas, in the case of phenol itself, ions could only be produced as the result of a destruction of neutral molecules, by the rupture of the hydrogen-oxygen bond, as in the equation :

Some care is needed in order to formulate these views in accordance with modern theories of atomic structure, since, as I pointed out in an article on "The Uniqueness of Hydrogen" exceptional difficulty is experienced in deciding whether an atom of hydrogen is linked to the rest of the molecule by a real bond or by a mere electrovalence. This difficulty arises from the fact that, since the hydrogen "shell" contains 0 or 2 instead of 8 electrons, the ordinary static symbolism makes no distinction between these two forms of union. Thus, in the case of H:C1:, the same electronic formula serves equally well to represent a pair of ions, HCl, with 0 and 2+8+8 electrons, or a covalent compound, H-Cl, in which 2 of the 18 electrons are shared by the two atoms. When, however, a dynamic model is used, and shared electrons are regarded as moving in binuclear orbits, a clear distinction between the two types of union again becomes, possible, according as the orbits of the electrons in question surround two nuclei or only one. If this distinction is admitted, it appears probable that hydrogen is covalent in all its compounds, and that the free hydrogen ion or naked proton represented by the symbol H does not exist except as a transient product in a vacuum-tube or the like. This conclusion follows, on the one hand, from the fact that there are no electrostatic forces to prevent such a nucleus from falling inside the electronic orbits of any other atom with which it may collide, and, on the other hand, from numerical data which show that the union of a proton with water to form a hydrated hydrogen ion would liberate not less than 260,000 calories †. The ionization of an acid

Lowry, 'Chemistry and Industry,' xlii. p. 43 (1923). ↑ Fajans, Ber. deut. physik. Ges. xxi. p. 709 (1919).

(which does not take place when the pure compound is merely melted) must then be formulated as depending on an interaction of the acidic hydride with water, or the like, giving rise to an oxonium ion, as iu the equation,

C ̧H ̧.0.H+OH⇒C ̧н ̧Ō+ÕH ̧.

It will be seen, however, that this scheme still possesses the characteristic which is now put forward as a typical property of" weak" electrolytes, that the opposite electric charges of the ions can be neutralized without any violation of the "octet" rule, giving rise to products in which the electrovalence of the ions is replaced by a real bond.

Strong and Weak Bases.

Bases resemble acids in exhibiting a very wide range of strengths. Thus the soluble metallic hydroxides (which include the alkalis) are all strong bases, and their strength is shared by the quaternary ammonium hydroxides, where the formation of a covalent bond between nitrogen and oxygen is prevented by the fact that the nitrogen is unable to attach itself by real bonds to five atoms simultaneously. Weak bases appear to be of two main types, since covalent compounds can be formed, either by fixing the hydroxylgroup by a real bond to an isomeric form of the kation, as in the acridinium compounds cited above, or by eliminating it in the form of a covalent molecule of water. The latter case is extremely common. Thus the amines generally behave as weak bases, since the concentration of hydroxylions in their solutions depends on an interaction with water:

e. g.,

compare

NH3+H,O[NH]+OH

2

CO2+H2O=[CO2] ̄ ̄H2++

the ready reversibility of which is shown by the complete removal of ammonia from its aqueous solutions by boiling, or of aniline by steam-distillation. Since, however, this dehydration has the effect of eliminating the unbalanced electric charges from the system, giving rise exclusively to covalent compounds, this explanation is in close agreement with the general proposition set out above as regards the essential characteristics of a weak electrolyte.

The problem of weak bases has, however, been carried one stage further by Moore and Winmill *, who have shown by

* Journ. Chem. Soc. ci. p. 1675 (1912).

distribution experiments that the weakness of the primary, secondary, and tertiary amines can only be partially attributed to dehydration of the hydroxide, and by Latimer and Rodebush*, who have suggested that the hydrate NH, H2O may be formulated as

Н

H:N:H:0 : H.

H

This formula represents the monohydrate as a covalent isomer of ammonium hydroxide,

in which the hydroxyl radicle is united to the ammonium radicle by a real bond, as required by our definition of a weak electrolyte. The fulfilment of this condition is not vitiated by the fact that, since quadrivalent nitrogen is positively-charged, and bivalent hydrogen is nega tively charged, the resulting molecule would exhibit the phenomenon which J. J. Thomson has described as "intramolecular ionization".

The bivalency of hydrogen is also of value in accounting for the fact that the polymerization of liquid water is not accompanied by ionization. Thus, if hydrogen were always univalent, the formation of a double-molecule would be represented by the scheme,

and the product would be an oxonium hydroxide, just like ammonium hydroxide; but, if the hydrogen is bivalent, the formation of double and triple molecules can be represented, without any formation of free ions, by the schemes,

H

H

H

2H :Ö: H➡:Ö: H :Ö: H and 3H :Ö: H≈:Ö: H :Ö: H :Ö:

Journ. Amer. Chem. Soc. xlii. p. 1431 (1920); cf. G. N. Lewis, 'Valence,' p. 110 (1923).

+ Phil. Mag. (6) xxvii. p. 757 (1914); cf. Lowry, Trans. Faraday Soc. xviii. p. 285 (1923), and Phil. Mag. (6) xlv. p. 1105 (1923).