LIII.-On an Apparatus f o r Maintaining n Constant Pressure when BistiZling under Reduced Pressure. By W H. PERKIN Ph.D. F.R.S. THE great value of conducting distillations under reduced pressure in the separation of volatile products from non-volatile ones in fractioning substances decomposable at high temperatures &c. is now so well recognised that it is unnecessary to comment upon it, and as it is a method of working so much resorted to at the present time anything that can render it more simple and manageable is of service. Those who have conducted distillation under reduced pressure know that the constant watching which is usually required to keep the pressure uniform very much interrupts the attention which i t is necessary to give to the distillation itself and as the work I am engaged in frequently requires me t o resort to distillations of this kind a variety of experiments have been made t o obtain some simple and effective automatic arrangement for this purpose and, thinking the results might be useful to others I have ventured t o lay them before the Society.The first idea which naturally suggested itself was to regulate the pressure by means of a column of merciiry so arranged that air would be drawn through it as soon as the desired reduction of pressure was attained. Arrangements on this principle have been used by Lothar Meyer and others (AnnuZen 165 303 ; 198 218), but my experiments in this direction especially for large reductions of pressure were not satisfactory and some more mechanical method was sought for.The use of a valve opening inwards and regulated by a lever and weight was next tried. This however,a-as also found to work unsatisfactorily because when the regulating point was reached a state of equilibrium existed and directly the air commenced entering vibration set in the valve chattering and not coming to rest again anti1 a considerable change of pressure had taken place. This showed that what was wanted was a valve that would act not only with considerable force and decision both as to opening and closing when the desired reduction of pressure was obtained but also with very small variations of pressure. This was attained eventually by means of a valve weighted to make it close quickly and capable of being opened by an electr~ma~gnet the action of the latter being inflnenced by a column of mercury making electrical contact with a platinum mire as soon as it reached the desired height.An apparatus made on this principle has been in use in my laboratory for several VOL. LIII. 3 690 PERKIX APPARATUS FOR RTAINTAINING years and worked satisfactorily. The following is an account of it given in detail as there are several points which require attention to make it work well. The description of the apparatus will be under-stood more easily by refereiice t o the diagrams which are given. At first, valves of metal and also of caoutchouc were experimented with b u t as they did not work very well glass was eventually used. The shape of the valve is that of part of a sphere a (Fig. l) half an inch in diameter ; it fits into a glass seating b t o correspond both surfaces The first thing to be considered is the regulating valve.F I G . 1. being very carefully ground and polished together ; there is a hole passing through the seating for the inlet of air. The valve has a hole drilled through it into which a wire with hooks on both ends is cemented. One hook is to carry tt weight c the other to suspend the valve to an electromagnet. The valve and seating is fixed into the movable top of a brass cylinder d in which the weight attached t o the valve hangs. A tube e is inserted into the side of the cylinder to connect it with the part of the apparatus which is to be exhausted by the water-pump 01' other means. The upper hook on the valve is attached by means of a copper wire,f to a ball and socket joint provided with a long screw passing through the armature of th r-4 I d' Fe.3 A CONSTANT PRESSURE. 69 1 magnet a s seen in Fig. 2 ; this screw is provided with a milled head, g so that the valve may be adjusted in relation t o the armature. There is also a nut above this and working on this screw which when tightened up to the armature locks the screw so that the adjustment may not alter by vibration. The electromagnet and other parts of the apparatus are fixed on a rigid mahogany stand as seen in the diagram and need no explanation. The armature is kept in its place at one end by a steel spring la the other resting in a brass support provided with an adjusting screw i so that the distance of the armature from the electromagnet may be amanged when necessary.This apparatus is kept in a box so that it may not suffer from dust and laboratory fumes a hole beipg pierced in the side of the box t o receive the connecting tube whilst binding screws are fastened on its top to connect with the necessary wires. The arrangement by which the desired reduction of pressure can be regulated consists of a graduated glass tube about 7 or 8 mm. in diameter and about 800 nim. long. A piece of tube about 70 mm. long and 12 in diameter is fused on to the tup and from its side an ordinary piece of quill tubing passes being afterwards bent twice at right angles ( j j j Fig. 3). This is connected with a T-piece k one arm of which passes t o the receiver consisting of a strong bottle of about 4 or 5 litres capacity the other arm being connected with the regulating valve by the tube 1.A stout copper wire m about 600 mm. long passes down the graduated tube ; this is provided with a platinum point screwed into it at the lower end and a binding screw on the upper end. Near the lower end a piece of copper is screwed on with three projectirtg points so as t o keep the copper wire in a central position in the tube j . This wire fits air-tight into the enlarged upper part of the tube by means of a vulcanised cork. The lower end of the graduated tube is drawn off sideways as in Fig. 4 leaving an aperture of about 0.5 mm. diameter (the object of this will be described further on) and is placed in a vessel of mercury m"" Fig. 3. A thin copper wire m' passes from the mercury in this vessel t o one of the binding screws of the regulating valve.Another in" passes from the binding screw on the thick copper wire to the battery and a third m"' from the second binding screw of the battery to the second one on the regulating valve. The working of the apparatus will now be easily seen. The copper wire 132 is drawn up till its point corresponds pretty nearly to the height of the column of mercury necessary t o be used for the desired pres-sure. On setting the water-pump in action the air being gradually removed from the glass bottle the mercury will rise in the graduated tubej but the moment it touches the platinum point the circuit is completed and the electromagnet lifts the glass valve so that some 3 A 692 PERKIN APPARATUS FOR MAINTAINIKG air enters ; this causes the mercury column to fall but the moment it does so the circuit is broken and the valve closes.It then rises FIG. 4. again ; the circuit is again completed and the valve opens and so on. I n this way when all the adjustments referred to further on are attended to a nearly perfectly regular pressure may be obtained not varying more than a fraction of zi millimetre. When first trying this apparatus the graduated glass tube j j j, was not cont'racted at the bottom but it was found that owing t o the sparking which takes place when contact is made between the wire and the mercury thus heating the rarefied air in its vicinity and the action taking place a t nearly equal intervals oscillations were set up in the mercury column and the apparatus therefore worked un-steadily but by contracting the end of the tube sufficiently as in Fig.4 the oscillations are almost prevented and the fluctuations of the columns which take place are only sufficient to regulate the pressure that is provided the regulating valve be properly adjusted, because it is evident that if it opens too wide it will let in so niuch air that the mercury will fall some distance before it closes again and if it does not open suGciently there will not be enough air admitted to counteract the excess of pumping power of the water-pump ; but with a little attention this matter can be arranged by turning the milled head g Fig. 2 on the regulating valve so that the quantity of air admitted only slightly influences the column of mercury.When read-ing the height of the column of mercury in the graduated tube j j j , the depression of the mercury in the reservoir 172"" must not be forgotten to be taken into account. As the column of mercury in the graduated tube cannot obviously be kept quite quiet a second tube was originally placed a t its sid A CONSTANT PRESSURE. 693 standing with its open end in the mercury reservoir the other end being connected with the exhausted receiver and this mas used as a barometer to measure the pressure and when the apparatus was working well no appreciable fluctuations mere noticed in this tube ; when measuring the pressure in this way however it was always neces-sary to compare it with an ordinary barometer giving the atmospheric pressure and subtract the one from the other to get the pressure the apparatus was working at.It was therefore thought that it would be more simple to connect it with a barometer direct instead of using this second tube and thus get the correct pressure at once. The most convenient way of doing this is by means of a syphon barometer (n Fig. 3) with the return tube made more than long enough to pass the zero point of the graduation. To the side of this, and near its end a tube 0 is fused to connect it with the glass bottle used for the exhausted receiver the open end of the tube p being closed with a cork. To make the apparatus more easy to move about, the syphon barometer is provided with a glass stop-cock q in the lower part of the return tube so that by tilting i t fills with mercury ; the stop-cock being then closed it can be carried about without danger.* When first trying this barometer it was found to woi+ very unsteadily.Owing to the vacuum in it it was found to be excessively sensitive t o minute changes of pressure especially when occurring at regular intervals oscillations setting in but it was eventually found that by turning the stop-cock p partially off taking care not to prevent the action of the barometer these oscillations practically were stopped; or the same thing can be effected by drawing out the tube which connects the barometer with the exhausted receiver to a fine point; when well adjusted only slight fluctuations will be seen on the meniscus of the mercury. This apparatus works successfully at a pressure of 60 mm.this being about the limit of the water-pump with which it has been used, but no doubt it would do equally well for lower pressures. For most work a pressure of 200-210 mm. is most convenient as the boiling is then generally pretty steady and there is no need to make arrsnge-ments for the admission of minute quantities of air its when using very low pressures. This pressure gives a reduction of over 40" in the boiling point. The apparatus employed in conducting the distillations is practically the same as that described by my late assistant Dr. L. T. Thorne (Trans. 1883 301). It has been in use in my laboratory ever since it was first devised and is found to be extremely convenient especially for fractional distillations.It is attached to the tube s. Q This would probably be a convenient adjunct to ordinary syphon barometers 694 APPARATUS FOR MAIXTAINING A COSSTANT PRESSURE. The battery power used for the regulating valve apparatus should only be sufficient to cause the apparatus to work with decision other-wise if strong sparking takes place between the mercury and the platinum point the surface of the mercury becomes dirty and the apparatus does not work well ; but if the power be kept down this takes place but slowly if the tube is of the diameter given above. A bichromate battery as shown a t r is convenient the power being regulated by the distance the zinc is allowed i o go into the chromic mixture. The tube j j j in which this sparking occurs should be attached tc the stand in such a way that it can be easily removed for cleaning when necessary.Experiments were made with floats covered with platinum foil so that there should be no sparking on the mercury. It was found that the best floats to work in a tube were small flat discs of metal covered with platinum upright ones attach themselves t o the side of the tube. These work pretty well but not SO sharply as the mercury itself a certain amount of pressure being sometimes necessary to make good contact and this necessitates a greater rise in the mercury than is necessary with mercury only the fluctuation, however is but small ; nevertheless I prefer to use the mercury alone, and to clean the tube when necessary. The use of a good sized bottle for the exhausted receiver (about 4+ litres) is advantageous as it assists in keeping the pressure equal and prevents any considerable alteration in it when changing receivers as in fractional distillation the small amount of air which enters having but little influence.If using this apparatus in connection with any powerful arrange-ment for pumping out the air it would be necessary to restrict the connection with it by means of a stop-cock capable of pretty fine adjustment or by a tube of fine bore because it is evident that if the pumping power is in excess of the capability of the valve to let in air it can no longer act as a regulator. The use of the weight on the valve is not necessary when large reductions of pressure are being maintained as the atmospheric pressure is sufficient to close it but for small reductions it is believed to be of service. Since devising this apparatus my attention has been drawn to a very ingenious pressure regulator devised by F. D. Brown (Proc. Phys. SOC. Lond. 1879 3 68) which has some points in common with mine but it has the disadvantage of requiring mot'ive power to work it as well as a battery