548 CHEMISTRY [ORGANIC. The globe, after being cleaned and allowed to cool, is again weighed, the tempera ture and height of the barometer being at the same time observed. The capacity of the globe is measured by breaking the point of the neck under mercury, when the metal rushes in to supply the place of the condensed vapour. As the expulsion of air by the escaping vapour is seldom complete, there usually remains a bubble of residual air, which must be allowed for by running in a known quantity of mer cury from a burette. The total quantity of ^, mercury is then poured FIQ. 8.-Dumas s Vapour-density Apparatus. out and its volume mea- A, glass globe, supported by wire claw; 6, project- sured. The calculation ing port ion of neck; BB, bath; C, thermometer; , c ,-, j , D, gas-burner for heating bath. is made from these data by the following approximative method : Let m = weight of globe + air at the temperature t of weighing and height of barometer b ; m = weight of globe + vapour at the temperature t of seal ing and height of barometer b ; V = capacity of globe in cubic centimetres ; A = weight of V cubic centimetres of air at t and b. Then m A = weight of vacuous globe ; m - (m - A) = weight of substance. Let H = wt. of V c.c. of hydrogen at t and b , then the vapour- density (d) referred to hydrogen is A~ - h When residual air (r) is found in the globe, V r = capacity of globe. In very exact determinations cor rections must be made for (1) the expansion of glass, (2) the difference of temperature and pressure between the first and second weighings of the globe, and (3) the difference in den sity between the drop of fluid remain ing in the globe and the density of mercury. For most chemical pur poses, however, the above-given ap proximation formula is sufficiently accurate. For high temperatures the globe is immersed in the vapours of boiling mercury, cadmium, or zinc, and the apparatus is modified accordingly. Gay-Lussac and Hofmqnris Methods. These methods have for their object the measurement of the volume of a known weight of vapour. Gay- Lussac s method, being available only for substances boiling below 100 C., has been gradually replaced by Hof- mann s modification (fig. 9). _ A glass tube about 1 metre in length and 20 mm. diameter, closed at one end, is graduated and calibrated. The tube being filled with mercury, and inverted in a vessel of the same liquid, is practically a barometer with an ex- Fio. 9. Hofmann s Vapour- aggerated Torricellian vacuum. Sur- density Apparatus, rounding this tube is a wider tube, AA, graduated barometer tube through which the vapour of any standing in funnel ; h, height of liquid boiling at a constant tempera ture can be passed, and thus the baro meter tube and its contents kept at that temperature. The substance of which the vapour- density is to be determined is weighed (about -fa gram.) in a minute stoppered bottle, and passed up into the Torricellian vacuum. According to the boiling-point of the substance (which is, of course, much lowered by the reduced pressure), the vapour of alcohol, water, aniline, or amylic alcohol is passed through the space between the mercury column; BB, outer glass cylinder enclosing baro meter tube; t, tube by which hot vapour is introduced; t , tube by which hot vapour and overflow of mercury escape; t is connected with the flask of boiling liquid, and t with a condenser. two tubes till the temperature and volume of vapour remain con stant. The height of the mercury column, the temperature to which the vapour is heated, and the height of the barometer in the room being observed, all the necessary data are obtained. Let m weight of substance in grammes ; V volume (in c. c. ) occupied by vapour at temperature t ; h = height of mercury in tube above mercury in reservoir ; b = height of barometer in room. Then b - h = pressure upon vapour. Let H = weight of Vc.c. of hydrogen at a pressure b-h and temperature t. Then the vapour-density (d) referred to hydrogen is ,.5. For exact determinations at high temperatures the tension of mer cury vapour (e) at the temperature (t) must be allowed for, and tha pressure upon the vapour then becomes b-h-c. In some cases the substance of which the molecular formula is to be determined does not admit of vaporization, being decomposed by heat. With such substances, some method other than the determination of the vapour-density must consequently be resorted to. In the case of acid or basic compounds, the problem admits of easy solution. Thus, supposing we desired to determine the molecular weight of acetic acid without having recourse to a vapour- density determination. Having ascertained that the acid contains one atom of hydrogen replaceable by metals, or, in other words, that it is monobasic, the silver salt is pre pared, and the amount of silver determined. All the necessary data are then obtained. Thus, supposing the analysis to give 64 67 per cent, of metal, the molecular weight of the salt, i.e., the weight containing one atom of silver, will be given by the proportion Whence 64-67 : 100 :: 108 : x
- -167.
The weight of the " acid-radicle " is therefore 167-108 = 59. And as one atom of H is replaced by the Ag, the molecular weight of the acid is 60. The empirical formula deduced from the ultimate analysis would be CH 2 O = 30, so that the molecular formula is 2(CH 2 O) = C 2 H 4 O 2 . With polybasic acids the problem is somewhat more complex, but the solution is effected in a similar manner, i.e., by estimating the metal in a normal salt. Silver salts are employed when obtainable, as they are generally anhydrous and easily purified by crystallization. As a further illustration we now give an example of the determin ation of the molecular weight of a basic substance. Supposing an analysis of the base triethylamine to have given the following results : Carbon, 71 "29 Hydrogen, U f Nitrogen * 100-00 The base is monacid, forming a hydrochloride containing one mole cule of HC1, and this hydrochloride forms a double platinum salt containing two molecules of the hydrochloride to one molecule of platinum ; 100 parts of the platinum salt left, on ignition, 32 U parts of platinum, so that, to find out the amount of salt containing one atom of platinum, we have 32-14 : 197-5 : : 100 : x .: x-6U-5. Putting x for the unknown molecular weight of the base, the mole cular weight of the salt is 2HC1= 73-0 Pt=197 5 C1 4 = 142-0
2x + 412 S