YOUNG : THE SPECIFIC VOLUMES OF LIQUIDS, ETC. 37 IV.-A New MctJaod of Detemniwhzg the Spci$c ?Tolumes of L i q u i d s and of their X n t w a t e d Yapours. By SYDNEY YOUSG, D.Sc., Professor of Chemistry, University College, Bristol. IF a closed graduated tube, containing a known volume of liquid at a given low temperature to, be heated to some higher temperature To, the liquid will expand, but the apparent expansion will be smaller than the real, on account of the evaporation of a portion of the liquid. If the density of the saturated vapour, or the ratio of the specific volume of the vapour to that of the liquid at To is known, the necessary correction may be applied, but at temperatui?es above the ordinary boiling point of the liquid this is usually not tlie case. Suppose now that by some suitable arrangement one portion ca of the tube cd (p. 38) can be heated to To, the remainder of the tube, ad, being kept at a constant low temperature fo, and tlhat by moving the tube, or the heating apparatus, a greater length ca' of the tube can subsequently be heated t o To, then we shall have a second method of determining the apparent expansion of the liquid. For a known volume of t h e liquid aa' at to has been heated to TO, and has expanded from b to b'.But in this case the space above the liquid is always filled' with saturated vapour, and since the rolume cb' is less than cb, the apparent expansion is greater than the real, f o r the vapour originally occupying the volume bb' has riow condensed. Here again, the ratio of the specific volume of the saturated vapour to that of the liquid at To being usually unknown, i t is impossible t o calculate the true expansion from to to To.But since by each method we have the38 YOUNG: THE SPECIFIC VOLUMES OF LIQUIDS Rame two unknown values-the true volume of the liquid and the ratio of the specific volume of the vapour to that of the liquid-we can obtain two equations from the experimental data, from which both values may be calculated. Let V, be the true volume of the liquid at To, and r the ratio of the specific volume of saturated vapour t o that of liquid at To. In the first experiment, when the whole tube is heated to To, let V’, be the apparent volume of the liquid, and V, the volume of saturated vapour. Then the volume of liquid at To, formed by the condensation of the saturated vnpour V,, would be VJr, and the true volume of liquid at To = v c v, = V‘, + -.r I n the second experiment, let Vt be the total volume of liquid at f, Va the volume of liquid between a and a’, and VB the apparent expansion from b t o b’. Then, since V, volumes of liquid at to expand so as to occupy the volume V, f VB at To, the total volume of liquid Vt would, under the same conditions, occupy the volume But in conse- V T’, (V, -j VJ.AND OF THEIR SATURATED VAPOURS. 39 quence of the expansion a certain amount of the saturated vapour has condensed, and the error due to condensation-calculated for the total Vt VR volume of liquid-will be - . --. Therefore, the true volume of v, ?' liquid at To will be From these two equations we have and the ratio of the specific volume of saturated vapour to that of liquid at To will be IT, v, - VIT' Lastly, the specific volume (volume of 1 gram) of liquid at To r = will be VT x Sr s, = - Vt ' where St is the volume of 1 gram at to ; and the specific volume of the saturated vapour st To will be ST = r x ST. It is of course impossible to heat the whole of one portion of a, tube strongly, and to keep the whole of the remainder of the tube a t a low temperature; there must be an intermediate portion, one extremity of which is at the higher temperature, the other at the lower, the temperature falling gradually from one extremity to the other.If, however, in both stages of the second experiment there is a similar gradation of temperature, though of course in different parts of the tube, the results will not be affected provided that the tube is of even bore.I have succeeded in devising an apparatus by which the specific volumes of liquids and saturated vapours may be determined by the method described, and as the metlt,od i s applicable to liquids which ccttack mercury, and may be epnployed through a wide range of tempera- tures, it seems desirable to give a detaiied account of it. A piece of barometer tubing, of very even bore and of about 60 cm. length, is closed at the blowpipe at d (Fig. 2, p. 40), about 20 cm. from one end, the closed part being carefully rounded. The portion of the tube de serves as a handle, and the end e may conveniently be d o se d .40 YOUNG: THE SPECIFIC VOLUMES OF LIQUIDS The tube from d to c is then graduated in millimetres, and calibrated in the usual manner by weighing with mercuryr Fig.2. e d C Cc > The end c is then sealed to the apparatus shown in Fig. 3, and the whole apparatus is thoroughly dried by repeatedly exhausting and allowing dry air t o enter. FIG. 3. The freshly distilled liquid is then admitted iuto the apparatus at g, either by means of a fine funnel, or, if the liquid is hygroscopic, by means of a siphon arrangement such as that described by Thorpe (Trans., 1880, 37, ad), until the wide protuberance f is about one- third filled. A piece of thick-walled indiarubber tubing, provided with a screw clip, is then passed over g, and the apparatus is con- nected with a pump and exhausted. The clip is then closed, the apparatus removed from the pump, and the liquid in the tube cd made t o boil vigorously for some time, so as to completely remove air adhering to the walls of the tube.(The narrow tube h, drawnAND OF THEIR SATURATED VAPOURS. 41 out and bent at the end, prevents the projection of the liquid against the indiarubber.) Lastly, the liquid is allowed t o fill the tube corn- pletely, and is then boiled from above downwards until only the iaequired amount remains in the tube, when the apparatus is tilted, s o that the greater part of the liquid above h flows intof. The tube is then at once sealed at c, the glass being allowed t o fall together, so as to withstand a high internal pressure. The volume of liquid is then nieasnred, either at 0" or with the tabe surrounded by running water of known temperature. The total volume of the tube, which is as yet unknown, is easily ascertained by inverting the tube, so that all the liquid flows to thc other end, and taking a second reading.The sum of the two readings gives the total volnme, but as the tube was calibrated with mercury, which has a convex meniscus, while that of the liquid is concave, the first reading must be corrected for this reversal of meniscus. When the tube is inverted, the concave meniscus of the liquid corresponds in position with the convex meniscus of mercury during calibration, and no correction is required. AB is a jacketing tube, containing a pure liquid which is t o be boiled under The arrangement for heating the tube is shown in Fig. 4.42 YOUNG: THE SPECIFIC VOLUNES OF LIQUIDS known pressures. The lower end of the jacketing tube is narrow, and is provided with a piece of indiarubber tubing, C, through which the volume tube cd passes.A current of cold water passes through the tube D, and keeps the loxer part of the volume tube at a constant temperature, which is measured by the thermometers GG’. If the indiarubber tube C has been kept under water for several hours before it is placed on the jacketing tube, the volume tube may easily be pushed up through it ; but, notwithstanding the reduced pressure in the jacketing tube, no water passes up between the indiarnhber and the volume tube. A ring of lead, L, is placed above the indiarubber cork I, so that the jacketing tnbe D may always be pushed up to exactly the same height.The position of the volume tube in the jacketing tube is shown by a horizontal line E, etched on the narrow part of the jacketing tube. The scale divisions on the volume tube corresponding to this etched line give the readings a and a’ (Fig. 1, p. 38). The jacketing tube is protected from drainghts by the outer tube HH, the space between A and H at the top being closed by a ring of asbestos cai=dboard. The experiment is begun with the volume tube in the position shown in the diagram. The liquid in B is made to boil, and when the vaponr has reached the top of the jacketing tube, and the pressure has been regulated to give the desired temperature, the volume tube is pushed up until the liquid in it is about 25 mm. above the level of the liquid F.The reading a may be taken at once, but tIhe height b of the liquid in the volume tube does not become constant until the liquid P (Fig. 4, p. 41), which has been cooled to some extent by pushing up the cold tube, has regained its original temperature. From 15 to 20 minutes are required before the reading of b becomes quit’e constant, but it is advisable to take readings every few minutes until constancy has been attained. The volume tube is now pushed LIP again until ihe bottom of the tube d is abowt 10 or 20 mm. below E, when the readings a and b’ are taken in the same manner as before. Lastly, the volunie tube is pushed up until d is about 25 mni. above F, when, of course, the whole of the liquid is heated, and the rending V’ is taken. In this case, the reading becomes constant after three or four minutes.The expansion of the liquid at low temperatures having been determined in the ordinary way, the volume Vt at the temperature t of the flowing water may be calculated or read from a curve, and we have now all the data required for calculating the specific volumes of the liquid and saturated vapoL1-r at To.AND O F THEIR SATURATED VAPOURS. 43 It is necessai-y, however, to apply the following correctioiis :- 1. The apparent volume of liquid at the ordinary temperature 01- 0" is slightly too small, as a certain quantity of the substance is present in the form of satumted vapour. The correction, which is very small, may be made on the assumption that the density of the satitrated vaponr is normal. 2. The expansion of the glass must be allowed for in the calculn- tion of VB, VrT, and V,. It is generally too small t o affect the value of V,.I n order to test the accuracy of the method, as regards the deter- mination of the specific volumes of liquids, experiments were made with benzene and carbon tetrachloride, for which substances accurate d a h have been obtained by the method employed in the investiga- tions of the benzene derivatives (Trans., 1889, 55, 486). In the experiments with carbon tetrachloride, however, aird in the earlier ones with benzene, the water jacket was not employed. A com- parison of the two series of determinations with benzene shows that more accurate results are obtained by the addition of the water jacket :- The results by the old and new methods are given in the following tables :- CCL; Temperature.100 200 210 220 230 250 260 Tolume of 1 gram liquid. Old method from curye. 0 '6972 0 *8408 0 '8637 0 -31 99 0 *999s 1 -0390 (J *8898 0 .GO6 1 0 -8418 0-8664 0 *sag9 0 -9219 1 -0029 1 -0CGO Difference per cent. -0 -16 + 0 '18 + 0 -30 + 0 *01 +0*24! + 0 -48 + 0 *6644 YOUNQ : THE SPECIFIC VOLUMES OF LIQUIDS Old method from curve. Temperature. Nem method. 100 130 160 190 220 24G 260 160 170 240 260 Eenzent: (witJLout TVafey Jackst). Volume of 1 gram liquid. I 1 -2616 1,3214 1 -3918 1 -4797 1 -5973 1 '7092 1 -8'7'70 L -2607 1 -3191 1.3877 1 *4815 1 -6006 1 9'174 1 W345 Beizaene (with TVater Jacket). 1 *3918 1 *4186 1 TO92 1-87'70 1 -3931 1 -4202 1 *ti123 1 -8809 Difference per cent. - 0 -07 -0 -17 -0 -29 +0-12 f 0 '21 -+ 0 -48 i- 0.40 + 0 '09 +0*11 +Om18 -t 0 -21 It will be seen that in the four determinations with the water jacket the greatest error is only 0.21 per cent., and that even without the water jacket the errors do not exceed 0.3 per cent., except at the highest temperatures.The chief value of the method depends on the fact that it can bc employed for the determination of the expansion of liquids such as nitrogen peroxide, bromine, 01' chlorine compounds which attack mercui-y. Xpec$c Volumes of Saturated Vapozw. The ratio of the specific volume of saturated vapour to that of liquid at To is given by the equation where V, is the volume of saturated vapour in the tube, V, the true volume of liquid, supposing all the rapour to be condensed, and V', the observed volume of liquid.It is clear that accurate results can only be obtained when the value V, - V', is fairly large, otherwise a, small error in either V, or V', would mean a relatively very much larger error iii their difference, and therefore in the value of r.AKD O F THEIR SATURATED VAPOURS. 45 Tempera- ture. ---- 160 170 180 190 200 2 10 220 230 240 250 260 270 280 -45 With a tube of the form employed, the results are sufficiently accurate only at high temperatures, at which the density of the saturated vapour is considerable. At low teniperatures the volume of the saturated vapour relatively to that of the liquid should be greatly increased. Of the three values V,, V'T, and VT, the two first are obtained by direct reading, and may therefore be considered as very accurately determined.On the other hand, the calculation of V, depends on a nniiiber of readings, and this value is therefore subject to greateia chalices of error. I f , therefore, better values of V, than those afforded by the new method are available, it is better to make use of them. Now the specific volumes of liquids, as determined by the old method (Phil. Trans., 178, 573, are calculated from a single reading in each case, and the values of V, can be calculated directly from them. I n the case, therefore, of substances which do not attack mercury, the best results mill be obtained by determining the specific volumes of the liquid by the old method, and the values V, a n d VIT by the new. There is also the great advantage that the time required for each determination of V', and V, is 1-ery small, for as the whole of the liquid is heated, the position of the voli-r.me tube in the jacketing tube does not require t o be altered during the experiments.I have determined the values of VfT and V, for carbon tetra- chloride in this manner at other temperatures in addition t o those given in the previous tables, and though I have as yet no means of testing the accuracy of the results, i t may be worth while to give d l those obtained at high temperatures. QC. . 0.8592 0 *8534 0 * 8479 0 *8425 0 -8369 0 -8313 0 '8259 0 %206 0 -815'7 0.8116 0 *8094 0 -8120 0 %400 Car4 ow Tetrcrch 1 oyide. QT. 1 -- 0 *5633 0 -5743 0.5864 0 -5998 0 -6149 0 -6316 0 -6507 0 -6'727 0 *698'7 0 5'311 0 '7743 0 -8395 0.9675 V'T. 0 -5402 0-5464 0 *5523 0 -5581 0 -5640 0.5699 0 '5'758 0.5814 0 '5866 0 '5911 0.5936 0 -5914 0 *5638 YT. 37 *00 30 -50 24 *90 20.20 16 '40 13 -50 11 .OO 8.99 7 -28 5 -80 4 '48 3 '27 2 *08 Voluine of 1 gram P a t urated vapour. 28 *50 24-00 19 * 90 16 *60 13 -80 11 -60 9 -81 8 '27 6 -95 5 *80 4 -74 3 -76 2 -7546 O’SULLIVAN AND TOMPSON : THE ESTIMATION Thz last two values of VT are somewhat uncertain, as carbon tetra chloride attacks mercury at high temperatures, and the values of VT were calculst,ed from the results by the old method. I hope before long to be able t o obtain some deterniinations of the volumes of a gram of the saturated vapours of the haloid derivatives of benzene, in order t o find whether the generalisation of Van der Wads holds good for these snbstances in the state of saturated vapour as well as in the liquid state.