17’5 XX.-Ort a Xisnpli’jicatiort of Begnnult’s Nethod for Determiizing Boiliq Points with small quantities of Substance. By H. CHAPMAN JONES, F.C.S. THE want of a method of determining boiling points by which a per- fectly definite result can be obtained by the expenditure of only a moderate amount of time and trouble, and the employment of only a small quantity of the substance under investigation, has probably been felt by many. The disadvantages of the ordinary process are several, the chief being the varying coiiditions under which different experi- menters work, and the necessity of vaporising a considerable quantity of the substance. The boiling point of a body may be defined as the temperature at which the tension of its saturated vapour is equal to the standard atmospheric pressure.The first method that suggests itself for deter- mining this temperature would be to introduce a small quantity of the substance into a barometer tube, filled with mercury, and inverted in a reservoir of the same metal, and then to heat the experimental tube, noting the temperature at which the mercury is at the same level outside and inside the tube. But such an experiment may be greatly simplified. If a syphon barometer tube is used, the separate cistern is dispensed with, and as it is only needed to observe the temperature at which the mercury stands at the same level in both limbs, the tube may be very greatly reduced in length. Practically, a piece of ordi- nary quill tubing, about 4 mm. internal diameter, and 200 mm. ( 8 inches) long, bent upon itself, so that the open end extends at least 1.5 mm.beyond the other, which is closed, is all that is necessary (vide Fig. 2). For the sake of convenience such a tube will in future be called a tension-tube. The filling of these tubes with mercury, and the introduction into the closed end of a drop of the liquid to be examined, was the first difficulty of importance. I found that it was possible, by the exercise of a considerable amount of patience and perseverance, to charge a tension-tube, and keep it pivetty free from air, by filling it with me]*- cury, except a small space at the open end, then adding the substance. closing the tube, and inverting, &c., exactly after the manner of work- ing with a bent eudiometer. Such an operation, however, is far too uncertain and tedious to be practically useful.I therefore tried to introduce the substance at the end of the tension-tube at which it was to remain, by means of a capillary opening. As this method has some advantages which will occasionally, perhaps, render it preferable to the one subsequently described, I will give the details of the mani- pulation.176 JONES ON A SlillPLIFICATION OF REGNAULT’S The tube is made as before directed, but instead of closing one end, it is drawn out to a capillary tube, about 60 mm. long, and bent as shown, Fig. 1. Mercury is then poured into the tube, until it is nearly full, when it is examined to see if any air bubbles remain in the FIG. 1. + size. limb that is t o be closed. Such bubbles can easily be removed by tapping or gently warming the tube, or, if they adhere obstinately, by emptying and refilling.When the tube is free from bubbles, mercury is added, till it begins t o drop from the capillary tube, and stands up above the walls of the tube at the other end. The damp thumb is now pressed upon the wide end, and should drive out more mercury from both ends of the tube. The apparatus is next inverted, and the‘ capillary opening introduced below the surface of the liquid for exami- nation ; then the thumb should be relaxed a little, and mercury allowed to flow out till the substance is drawn into the capillary tube, and occupies 30 or 40 mm. of it. A little air is then drawn in, and the capillary tube sealed in a small flame. The mercury is now removed from the open end down to the bend, by means of a pipette, and the bubble of air in the capillary tube reduced in size, by gently warming to expand it, and then re-sealing as close as possible to the liquid.This operation may be repeated, but by no means have I been able to get quite rid of the air. This bubble, which should not be larger than a pin’s head, entirely prevents any disposition of the substance t o become overheated in the bath employed, because of its adhesion to the glass, but at the same time it lowers the apparent boiling point. This method has one other great disadvantage, namely, that it cannot be used with safety when dealing with substances that are changed on ignition. The approximate amount of error to be expected from these causes will be seen from the three following cases, whereMETHOD FOR DETER3lINIKG BOILING FOISTS, ETC.17 7 the upper figures show the true boiling point (determined by the method subsequently described), and the middle row of figures the results obtained by the method just detailed. In the 1st. CS?, the bubble was not reduced at all ; in the 2nd. it was reduced as far as possible by the method given above. The greater difference in the case of ethylic iodide may be attributed to the change brought about in a small quantity of its vapour by the operation of sealing off the capillary tube. cs2. csp C2H.51. OH,. 46.2 46.2 72.2 100.0 45.9 46.0 70.7 99.3 Difference -3 -2 1.5 -7 The only way in which I have succeeded in charging tension-tubes without chance of failure, and with perfect elimination of air, is by driving out the air by the vapour of the substance itself.This is effected by introducing about two drops of the liquid to be examined into the tube, running it round to the closed end, then introducing the open end into some mercury contained in a small porcelain crucible. This arrangement, supported by a suitable sling, made of copper wire, is lowered into a water or paraffin bath, according t o circumstances (see Fig. 2), and the temperature gradually raised. The air of the tension-tube is thus driven out in bubbles, which rise through the mercui-y and the liquid in the bath ; the air is, of course, followed by ------- ‘L I FIG. 2. f size.178 JONES ON A SIMPLIFICATION OF REQNAULT'S the vapour of the substance. These bubbles come off with increased rapidity as the temperature of the bath nears the boiling point of the :)ody used, and at about from 4" to 6" above it, there is a very marked increase, the action becoming violent if pushed much further.By observing the temperatures at which these changes take place, the boiling point of the substance may be ascertained to within 2" or 3', a precaution which should never be omitted, as it saves subsequent trouble. When enough vapour has escaped to sweep out all the air, the bath is allowed to cool, or, if the temperature is not above 50" or 60°, the tube with the crucible containing mercury may be at once removed from the bath. In either case as the vnpour in the tube condenses, the mercury rises, and completely fills it, except the space occupied by the drop of liquid at the top of the bend.The tension- tube is now released from the sling, and then removedfrom the crucible of mercury, and turned over by one operation, taking care that the closed limb shall be uppermost, so that the liquid may rise into the closed, and not into the open end. If the water (or paraffin) and mer- cury are nowremoved from the open limb by means of filter paper (or cold wire) and a pipette, the tube is ready €or use. This method of filling a tension-tube, or rather of making it fill itself, is exceedingly simple in manipulation, and perfectly satisfactory. A tube so filled, after nine days had elapsed, during which it had been used two or three times, did not show the slightest evidence of ariy air having gained access to the closed limb, and could be heated a degree or two above the boiling point of the liquid without vaporisation taking place.We now pass to the method of using these tubes. The temperature at which the mercury is at the same level in both limbs, will be the boiling point of the substance. To ascertain that temperature, the tube must be introduced into a bath of transparent liquid contained in a glass vessel, so that the closed end is below the surface, and the open end is freely exposed to the air. The bulb of the chief thermometer should be brought as close as possible t o that portion of the closed limb which will be occupied by the vapour of the body, and a second thermometer arranged as described below. The temperature is then gradually raised, and the tube carefully watched as it approaches the boiling point of the substance. I f the level of the mercury does not sink as soon as you judge that it should, the open end of the tube, which projects from the bath, should be gently tapped vertically by a hard substance (glass or metal).This gentle jerk will entirely obviate the difficulty. But if by accident or neglect the temperature of the bath has risen 2" or 3" above the boiling point of the substance, and consequently the mercury sinks with increasing instead of with di- minishing rapidity, we can prevent, the contents of the tube from being driven out, and the experiment lost, by raising the closed end a littleMETHOD FOR DETERMINING BOILING POlSTS, ETC. 1'79 out of the bath, and lowering it into the bath again only as the tempera- ture falls.(It will be observed t'hat if the tube is filled by the second method and is free from air, the approximate boiling point is ascer- tained during the filling.) With common care, however, such an acci- dent will never occur, but the mercury in the closed limb will gradually sink, until it is lower than the mercury in the open limb ; on cooling, the mercury will rise again, and so by alternately heating the bathand allowing it to cool, the two mercury levels can be made to pass one another any number of times. Each time that the lercls correspond, the temperature is noted. Six is a very convenient number of such observations to make, three as the bath is rising in temperature, and three as it is cooling. It is perfectly easy to keep these oscillations within half a degree on each side of the boiling point, and If" on each side of it should be the maximum.The average of six such observations may be taken as the uncorrected boiling point, unless the first temperature evidently disagrees with the rest, when the first two observations are neglected, and two others substituted for them. This error will occur when the temperature of the bath has been raised too quickly, so that the tension-tube, with its contained mercury, has not been able to keep pace with it. The tube, when done with, may be labelled and preserved, if worth while, so that the observations can be confirmed a t any future time. The boiling point so determined can be regarded only as approxi- mate unti1 it is corrected, and as all the boiling points given in this paper have been reduced to standard conditions, to make them com- parable, it may be well to state here the exact corrections that have been applied.It is necessary to observe the height of the barometer, and the temperature of the air, and thus to get the height of the barometer a t 0" C.; also to observe the mean temperature of that portion of the thread of mercury in the thermometer that is not im- mersed in the bath. To ascertain this temperature a second thermo- meter is used, the bulb of which is placed half way between the level of the liquid in the bath and the top of the mercury in the chief ther- mometer. Every time the temperature is to be taken, both thermo- meters must be read ; then the number of degrees of the mercury of the chief thermometer that are not in the bath is to be multiplied by the difference in temperature between the two instruments, and the product multiplied by *0001545 (the average coefficient of the appa- rent expansion of mercury in glass) ; the result is to be added to the temperature indicated by the chief thermometer, to obtain the real temperature.I f the work is carried out as above described, there will be six such actual temperatures obtained, the average of which gives the boiling point under the pressure observed. The correction for pressure is made by multiplying the number of180 JONES ON A SIMPLIFICATION OF REGNAULT'S mm. by which it differs from 760, by -037 (that is, the fraction of a degree by which the boiling point of water varies for each mm.of barometric pressure, starting at 760 mm.), and adding the result to or subtracting it from the temperature observed, according to whether the pressure is below or above the normal pressure. A temperature so obtained is designated simply the boiling point. I have made a good many experiments to ascert'ain the degree of accuracy that may be expected by the use of this method, and to show the effects of varying circumstances upon it. In these experiments I have carefully avoided any refinement of apparatus, or the spending of an undue amount of time over each operation, so that the results truly show the applicability of the process to the ordinary require- ments of a laboratory, where the observation d boiling points ia but of rare occurrence. The apparatus used was the following :- A thermometer by N e g r e t t i and Zambra, graduated from -5" to 105" C., each degree being divided into 5 parts.This was compared with a Kew standard at the ordinary temperature and found correct, and it registered 99*8-99%-99.9, instead of 100, for the boiling point of water. -2" is therefore added to its indications at about 100, and a corresponding correction is made for other temperatures. Its read- ings are given to *lo, but the last figure must not be taken as absolute ; the error may, however, be considered as less than -1". An ordinary thermometer divided into single degrees graduated up to 360°, which was found fairly correct. A beaker of a little more than half a litre capacity, as a water-bath, and one about half its capacity for the paraffin-bath. It would have been better if the size of the paraffin-bath had been increased. These were placed on a thick iron plate, or a piece of tinned iron (tinplate2 supported on a tripod, and warmed by means of a Bunsen burner.The two slings used for supporting the apparatus were made out of copper wire, the fingers being the only tools. Bisulphide of carbon was the first substance employed. It is stated to boil a t " 46.6" under ordinary pressure." A small quantiiy of it was digested with sodium and then distilled without visible ebul- lition, by placing the retort containing it in warm water. The great variations of the barometer in November last, enabled me to make observations a t very different pressures. The following are the tem- peratures observed for the difei-ent pressures, the lowest line being the boiling point deduced from each experiment :- Pressure in mm... . . . . 736.0 744.5 752.0 762.0 B.P. at above pressure. 45.1 45.5 46.0 46.3 B.P. a t i60 mm. . . . . 46.0 46.1 46.3 46.2METHOD FOR DETERMINIXG BOILIXO POIXTS, ETC. 181 Each of these temperatures is the average of from 5 to 10 observa- tions, made as before described, the greatest difference between any two observations in anyone series being -2". The observations in any one series are often all exactly alike, and could almost invariably be made so by working a little more slowly. The following is a speci- men of what may be expected, the second decimal place being caused by the correction for exposed t,hread of mercury in the thermometer.The lst, 3rd, 5th, &c., were taken as the temperature of the bath was rising ; the 2nd, 4th, 6th, &c., as it was falling :- 46*04A5*90A6*0446~0O-45~99-45~95-4.Eib99-45*99- 45-99-4599. A sample of e t h y l i c iodide, which I believe is pure, was tested by this method, and its boiling point found to be 72.2". A determination by Dr. F r a n k l a n d (when corrected for pressure) gave 72*1", when tlie thermometer bulb was in the vapour, and 72.7 when immersed in the liquid. The entire time occupied by the above determination (including the making of the tension-tube) was 19 hour, much of which time was occupied in heating the water-bath, no attention being then necessary. Two determi- nations of its boiling point made by fixing the thermometer with its bulb a little above the surface of some water kept rapidly boiling in a flask, gave 100.1" and 100*0".These experiments were made quite independently at an interval of ten days. The boiIing point of water, taken by means of a tension-tube and a paraffin-bath, was found to be 100.0". Some phenol was prepared from solid carbolic acid by distilling and collecting when the distillate was homogeneous. The portion collected came over between 1 8 3 O and 185". The phenol was liquefied by immersing the bottle containing it in warm water; the tension- tube was slightly warmed, and a drop or two introduced by meam of a small pipette. The boiling point was found to be 183F, but a t this high temperature the tension of mercury-vapour becomes appreciable (= 11 mp.) ; so, deducting this from the barometric pressure, and recalculating, we get 184' as the boiling point.This agrees with a determination by Scrugham, who found the boiling point to be 184", whereas L a u r e n t made it 187" to 188". The next substance employed was benzoic acid. This body is slated to melt at 121.4" and boil at 250". The tension-tube was charged as usual, the acid vaporising abundantly a t 256". The snb- stance solidified on cooling a t the bend of the tube, but was driven to the top of the closed limb without the least difficulty by gently warm- ing it in a small flame. As the paraffin became practically opaque by W a t e r was the next substance experimented with. VOL. XYXIII, P182 JONES ON A SIXPLIFICATION O F REGXNAULT'S METHOD, ETC.heating it to this high temperature, the experiment could not be finished ; but if a suitable substance were found to replace it, the only difficulty of these experiments at such temperatures would be sur- mounted. Inasmuch as substances operated on are not alwajs pure, it is well to know the effect that solids in solution, liquid impurities, &c., have upon the boiling point of a body. The following experiments were undertaken with this elid in view. A strong but not saturated solution of calcic chloride gave a boiling point of 1 1 9 + O . A saturated solution of common s a l t was then experimented with as follows. An ounce or two was introduced into a capacious flask, some scraps of platinum dropped in, and the solution heated till it boiled briskly. A thermometer with its bulb immersed in the liquid was stationary at 108$", and when raised so that the bulb was about 1* inch from the surface of the liquid, it fell only or 6 a degree.The boiling point taken by means of a tension-tube was found to be 107r. A minute bubble of air was afterwards discovered in the tube, which accounts €or the discrepancy observed. The separation of solid salt tended to prevent a very satisfactory operation. These experiments show that solid substances in solution tend to raise the boiling point of the liquid, that is, the tension of the vapour of the solvent is lowered. A mixture of alcohol and water gave ft boiling point of 82.0°, showing that a mixture of liquids that mix with one another, or that are mutually soluble in each other, gives a boiling point intermediate between the boiling points of its constituents. Some carbonic disulphide and water mere introduced into a t,ension-tube, and the apparent boiling point determined.It was found to be 43.$", showing that a mixture of liquids which are not soluble in each other, gives an apparent boiling point lower than that of either separately, that is, the tensions of their vapours are added. By deducting from the pressure observed, the equivalent for the tension of water vapour at, the observed temperature, and recalculating for the reduced pressure, we obtain from this experiment 45.7" for the boiling point of CS2, its actual boiling point being 46.2". It, was not considered necessary to multiply these experiments, as they agree perfectly with well known laws.In conclusion, I think that the advantages of this method will be found to be:- 1st. The obtaining of definite, constant, and perfectly comparable results. 2nd. Diminution of error of observation, as several readings can be taken. Paraffin is suitable for use up to about 200".JOEINSON ON CERTAIN POLSIODIDES. 183 3rd. The fact that a singIe drop of any substance is all that is 4th. That the method can at any time be rendered more exact, by needed. litble refinements that will be obvious to any experimenter. Postscr@f.-On communicating the results of the work detailed above to Dr. Frankland, he was good enough to tell me that sper- maceti could be used at higher temperatures than paraflin. I find that spermaceti will bear heating up to nearly 300" C. without losing its transparency or evolving an inconvenient amoiint of vapour ; the only drawback to its use for all temperatures over 80' or 90" is that it would probably be more readily acted upon by the vapour of substances passing through it than paraffin. Dr. Frankland was also so kind as to give me small quantities of the following substances to examine by my method. No. 1 was a sample of methylic iodide, and No. 2 a sample of ethylic diethoxalat,e, ( {(cc2H5)2'0H)7 o. o(c2H,) B.P. stated as 175" C. (barometric pressure not known). 'Both these were the ordinary preparations of the'laboratory. No 3 was a, sample of commercial benzol, known to be impure. The boiling points I found were as follows :- 1. 2. 3. 42.8 1752 86.9 These results fully confirm the usefulness and trustworthiness of the method.