Pentaphosphonitrilic chloride, P5N5C110.-This body, carefully purified by fractional distillation, as described above, gave on analysis:

Pentaphosphonitrilic chloride fuses at 40.50-410, and boils at 223°224.3° (corrected) at 13 millimeters. Its vapor is without the pronounced and characteristic aromatic odor possessed by that of triphosphonitrilic chloride. At its melting point it is miscible in all proportions with benzene, gasoline, ether, and carbon disulphide, and can not be recrystallized from any of these solvents; in fact, small fragments liquefy instantly in their concentrated vapors. Glacial acetic acid also dissolves it quite readily, and from this solution water throws it out as an oil, solidifying at once on touching. It shows a decided tendency to superfusion, especially when not absolutely pure; when left by evaporating its ether or benzene solution, it may remain liquid for days, but solidifies at once on touching with a glass rod, usually to a decidedly crystalline mass, at other times to a transparent glass. The pure substance, when fused, slowly solidifies, long, flat crystals shooting out through the liquid, which are limited only by the size of the vessel, crystals of 10 centimeters in length being readily obtained. It contracts greatly on solidifying. When pure the solidified mass is naturally dry, but the least contamination with other members of the series causes a portion to remain liquid, which is easily detected by crushing on a piece of filter paper; this is a very good test of its purity. This tendency to superfusion must be borne in mind in separating it by fractional distillation; a nearly pure sample will remain liquid much longer than the higher or lower fractions. In ether solution it is perceptibly more easily attacked by water than the preceding chloronitride, but hot water alone is almost without action.

Hexaphosphonitrilic chloride fuses at 90° (corrected), and boils at 2610-263° (corrected) at 13 millimeters, and at 2810-282° (corrected) at 26 millimeters. It may be recrystallized from benzene, in which, however, it is more soluble than triphosphonitrilic chloride; ether, gasoline, and carbon disulphide also dissolve it readily; in alcohol it dissolves somewhat slowly, with decomposition. It shows no tendency to superfusion. It crystallizes well, in rather large crystals, which were examined by Mr. Wirt Tassin, to whom I am indebted for the following statement:

mm

.b

Fig. 7.

แ

"P6N6C112 crystallizes in the orthorhombic system in long, prismatic crystals, showing the following forms: c (001), b (010), m (110), o (111), n (011). Of these c is the dominant form; b large and well developed; m fair, though usually narrow, and o and n small and usually in similar development. c, b, m, o is the combination occurring most frequently; less often c, b, m, n, o; and rarely c, b, m. Angles m: m 57° 28′, b: m 61° 16′, b: n 40° 23', n: c 49° 37'. Axial ratio a : b : c = 0.54824 1: 1.17568. The crystals are optically positive. Plane of the optic axes (100);

1A contamination with PN,Cl, may be detected by treatment with gasoline, when the much smaller crystals of the latter are seen to dissolve much more slowly.

colorless to white; transparent, and have a perfect basal cleavage." (See fig. 7.)

It is scarcely attacked by boiling water, but if kept in moist air it very slowly evolves hydrochloric acid. Its ether solution, shaken with water, slowly gives a metaphosphimic acid; sirupy chlorhydrines are formed as intermediate products.

Heptaphosphonitrilic chloride, PN-C114.—After many distillations, this gave:

Heptaphosphonitrilic chloride is a nearly colorless, rather viscous liquid, which does not solidify at -18°, and which boils at 2890–294° (corrected) at 13 millimeters, undergoing some polymerization. It is readily miscible with benzene, gasoline, and ether, and toward water shows the stability manifested by the preceding chloronitrides.

Residual oily phosphonitrilic chlorides.-The residue, which did not distill at 13 millimeters when the temperature of the bath was 370°, was boiled with water to remove the solid polyphosphonitrilic chloride, filtered, and carefully dried in vacuo over sulphuric acid. It is a thick liquid which gave on analysis:

1 It is unlikely that these consist entirely of the original depolymerization products; a portion is doubtless formed by polymerization during distillation.

Although the oil is doubtless a mixture, the above figures show that the constituents are members of the series (PNC). A determination of the mean molecular weight was made, with the following results:

11 11

Mean 1,308. This does not lie far from that required by the formula P1NiCl2 (calculated, 1,276). This result is interesting in as far as it shows that a phosphonitrilic chloride of this molecular weight may exist; that it is stable and miscible with benzene, gasoline, and ether, and that the molecular weight of the solid polymer described below, which is insoluble, is probably very much higher. The oil has a reddish-brown color, due to dissolved impurities, which are destroyed by heating with strong nitric acid. It can not be distilled even at 13 millimeters, as it polymerizes almost instantly. In its behavior toward water it resembles the preceding members of the series.

Polyphosphonitrilic chloride (PNC12).-This remarkable body, frequently alluded to above, is formed when any of the lower members are heated, slowly at 250°, and very rapidly at 350°. As the change is reversible, complete transformation can not be effected, but reaches perhaps 90 per cent, the remainder consisting not only of the original phosphonitrilic chloride, but of others. These can be extracted by anhydrous benzene. The sample, the analysis of which is given, was prepared by heating pure triphosphonitrilic chloride in a sealed tube at 3500-460°. The transparent elastic product was repeatedly extracted with benzene dried over sodium and the absorbed benzene removed in a vacuum with constant exhaustion over paraffin, and finally by heating in vacuo at 110°. Analysis gave:

Polyphosphonitrilic chloride, when perfectly pure, is colorless and transparent, but is generally somewhat discolored by traces of organic matter. Its most striking property is its elasticity. It may be drawn out like rubber, and shows even a greater tendency to rebound from hard surfaces. It is readily cut with the shears. It is insoluble in all

neutral solvents, but absorbs benzene, swelling to many times its original volume and forming a jelly of but little coherence; on evaporating the benzene it returns to its original condition. Ether is absorbed, but less readily, and other chloronitrides are taken up in a similar manner. Hot water slowly dissolves it with decomposition; in warm dilute ammonia it swells, gelatinizes, and finally dissolves; hot caustic soda does not dissolve it readily, apparently insoluble sodium salts being formed. It begins to depolymerize toward 350°, and this change is rapid just below red heat, and is accompanied by partial fusion, the products being, as described above, a mixture of lower phosphonitrilic chlorides. These transformations are not modified by heating in an atmosphere of hydrochloric acid. When perfectly pure, it leaves no residue whatever on distilling. No difference could be detected in the product formed from different phosphonitrilic chlorides. Nitrilo-hexaphosphonitrilic chloride, P&N,Cl9.-The separation of this chloronitride, which does not belong to the phosphonitrilic chloride series, is described above. After four crystallizations from benzene, it gave:

A molecular weight determination, kindly made for me by Dr. H. C. Jones, of the Johns Hopkins University, with a limited quantity of material, in benzene solution, gave 667; calculated for PN-C19, 603.5. This, with the analysis, suffices to establish the above molecular formula.

Nitrilo-hexaphosphonitrilic chloride strikingly resembles the phosphonitrilic chlorides. It fuses at 237.5° (corrected) and boils at 251°-261° (corrected) at 13 millimeters without change. The boiling point coincides closely with that of hexaphosphonitrilic chloride (261°–263°, corrected at 13 millimeters), hence it is found associated with the latter. Heated in small quantities on foil, it volatilizes without residue, but at a higher temperature in a sealed tube it undergoes a change the exact nature of which has not been determined, but which involves the form ation of a substance resembling polyphosphonitrilic chloride, which yields lower phosphonitrilic chlorides on distilling. It crystallizes in transparent prisms of not more than 1 millimeter in length, apparently rhombic, which are often united to acicular forms. When pulverized it becomes electrified. It dissolves in about 20 parts cold and 5 parts boiling benzene,' is more soluble in carbon disulphide, but less soluble

Approximate only.