9 111.-Researches on the Vacuum. By HERMANN Ph.D. SPRENGEL 1. The Instruments. THE methods hitherto proposed for producing a vacuum may be divided into two classes the mechanical and the chemical. A gas wliich fills a space may either be removed mechanically or it may be converted by taking asvantage of its chemical or physical properties into a non-gaseous tensionless body. As however our atmosphere from its general presence is for the most part the only gas which has to be considered when a vacuum is to be formed and as the reduction of its consti-tuents to the non-gaseous form is attended with peculiar diffi- culties I am inclined to consider that all practical methods of producing a vacuum may be regarded as mechanical. Atmo-spheric air may in fact be expelled from a space firstly by a solid * Chem.SOC.J. vol. ix p. 1. SPRENGEL’S RESEARCHES ON THE VACUUM. body secondly by a liquid and thirdly by a gas which afterwards takes either a liquid or a solid form. Perfect vacua have hitherto been obtained only by this last method but with such attending difficulties that it appeared to me highly desirable to improve the other two which afford greater facilities of working. The dis- placement of a gas by means of a solid or a liquid body is effected by instruments commonly called air-pumps. The exhausting syringe of Otto von Guericke with its piston is the type of one class of these instruments while Toricelli’s baro- meter with its column of mercury is the type of the other class.The invention of the common air-pump was a consequence of Toricelli’s vacuum and it is partly due to this that the degree of exhaustion attained by it has been up to the present day com- pared with Toricelli’s vacuum as the standard of perfection. Hence it has followed that physicists have from time to time attempted to use Toricelli’s vacuum as a receiver or in other words to fill receivers with mercury and attach tubes to them as long as those of barometers through which the mercury was allowed to run off All so-called mercurial air-pumps are based upon this principle and differ very little from the famous experi- ment which Toricelli made 220 years ago. This however is not the only way in which the case may be viewed. If the top of a barometer were knocked off the air would enter and the mercury would sink or what is the same the mercury would sink and draw in the air.If however the experiment be so arranged as to allow air to enter together with the mercury and in such a manner that the supply of air shall be limited while that of mercury is unlimited the air will be carried away and a vacuum produced. It is upon this principle that I have constructed a pneumatic machine of which Fig. 1. represents the simplest form. cd is a glass tube longer than a barometer open at both ends and in which mercury is allowed to fall down supplied by the funnel A with which the tube is connected at c. The lower end d of this tube dips into a small glass bulb B into which it is fixed by means of a cork.This glass bulb has a spout at its side situated a few millimetres higher than the lower end of the tube c d. The first portions of mercury which run down will consequently close the tube and form a safeguard against the air which might enter from below if the equilibrium should be dis- turbed. The upper part of cd branches off at x into a lateral SPRENGEL’S RESEARCHES ON THE VACUUM. tube to which the receiver R is affixed. As soon as the FIG. 1 stop-cock at c is opened and A the mercury allowed to run down the exhaustion begins and the whole length of the tube from x to d is seen to be filled with cylinders of mer- C cury and air having a down-ward motion. Air and mer- I! cury escape through the spout of the bulb B which is above the basin H where the mer- cury is collected.This has to be poured back froin time to time into the funnel A to pass through the tube again and again until the exhaus- tion is completed. As the exhaustion is progressing it will be noticed that the en- closed air between the mer- d cury cylinders becomes less and less until the lower part of c d presents the aspect of a continuous column of mer- cury about 30 inches high. Towards this stage of the operation a considerable noise begins to be heard similar to that of a shaken water-hammer and common to all liquids shaken iii a vacuum. The operation may be considered completed wben the column of mercury does not enclose ariy air and when a drop of mercury falls upon the top of this column without enclosing the slightest air-bubble.The height of this column now corresponds exactly with the height of the column of mercury in the baro-meter; or what is the same it represents a barometer whose Toricellian vacuum is the receiver R. Fig. 2 represents the actual instrument with which I am in the habit of working. The funnel A supported by a wooden stand serves as the reservoir for the mercury. By means of the tube z y 3 it is connected with the tube c d which I call the SPRENBEL’S RESEARCHES ON THE VACUUM. “fall tube,” so that the mercury will run down as soon as the clamp z is opened which compresses a caoutchouc tube inserted FIG. 2. P there. The tube xp leads to the receiver which is to be exhaust ‘ed and is in connection with two tubes one of which is attached to SPRENGEL’S RESEARCHES ON THE VACUUM.the common exhausting syringe S while the other serving as a gunge dips into a glass of mercury containing a barometer. When the instrument is at work the rising of the mercury in this guage will consequently show the degree of exhaustion. The exhausting syringe is merely attached as an auxiliary to accelerate the operation because the fall-tube for a reason to be presently mentioned must he of a thin calibre. The greater portion of air is more quickly removed by the syringe and after this has been done as much as possible and the connection between the receiver and the syringe has been broken off by compressing an inserted caoutchouc tube with the clamp i the remainder of the air is carried off by the runniirg mercury.The bulb B and the basin H are exactly as in Fig. 1. The instrument is about 6 feet high so that the mercury collected iii the basin H cm easily be poured back into the funnel A. The use of a pump would facilitate the raising of the mercury and prevent the admixture of air. This latter inconvenience may however be pretty well overcome by gently pouring the mercury on a glass plate floating on the sur- face of that in the funnel. Should a few air-bubbles attach them- selves to the side of the funnel it may be best to remove them by means of a wire or a glass-rod though they are not perhaps of much consequence as they are not observed to pass along with the mercury.The connections between the glass tubes are made of well-fitting black vulcanised caoutchouc tubing sold under the name of “French tubing.” This is free from metallic oxides which mder the tubing porous. Besides this all these joints are bound with coils of copper wire which is easily accomplished with a pair of pliers. Moreaver the space between the inside of the caoutchouc tubing and the outside of the glass tubing is filled with a resinous cement made of fused caoutchouc To prevent this substance soiling the interior of the instrumeut I first after having put the instrument together tie the caoutchouc joint with the copper wire and then turn back the end of the caoutchouc tubing over the coil coat the inside of the end with the cement and turn it back again into its proper position.From this it is obvious that the fused caoutchouc can only penetrate as far as the copper wire coil. The conuection of the funnel with the tube x y o is made by means of a perforated caoutchouc cork. (These corks are easily made from a flat block of caoutchouc cut out with a sharp Mohr’s cork-borer well lubricated with oil). When the instrument has been put together in this manner it is ready for SPRENGEL'S RESEARCHES ON THE VACUUM. use. At first the mercury is allowed to enter the fall-tube in such quantities as to raise the mercury in the guage as quickly as possible. When however the operation approaches its comple- tion which is shown by the rattling noise it will be found useful to lessen the supply of mercury and to let it fall down drop by drop on the column of mercury in the lower part of the fall-tube and to proceed in this way till the exhaustion is completed.I am not able to give any definite statement as to the quantity of mer-cury to be employed as it is obvious that a small quantity say an ounce or two is c%pable of exhausting a receiver of an indefinite size if this mercury is only made to pass the fall-tube often enough; but I may remark that I have found 10 to 15 lbs. of mercury a convenient quantity to work with. In my endeavours to find out how to construct the instrument in order to exhaust a receiver with the greatest economy of time and mercury I have not met with satisfactory results. There of course exists a certain relation between the amount of air to be exhausted the quantity of mercury to be employed and the time of the operation.In order to make the instrument act at all the supply of mercurymust be at least so large that the fall-tube may become closed i. e. the running mercury may form drops of a cylindrical shape breaking off the communication bet.ween the receiver and the external air. As the supply of mercury is in- creased the rapidity with which the air is carried off also increases. But this soon reaches its limit as should the mercury he admitted too rapidly into the fall-tube it gains the preponde- rance and closes the aperture at x. The most favourable condi. tions under which the instrument might be used are those where the mercury is made to fall down drop by drop enclosing between every two drops as large a portion of' air as possible.Volume may be increased by extension either in height or in breadth. If the fall-tube be lengthened the bulk of the enclosed air will be increased and the time required to produce exhaustion will be shortened without increasing the weight of mercury employed. But as it is inconvenient to have the instrument higher than the height of a man I attempted to increase the second dimension Gz. the width of the tubes. I soon found however that it was difficult if not impossible to close tubes of more than R certain width by single drops. In order to clase a wide tube with a cylinder of mercury (or any other liquid) the mercury (or this other liquid) must run in freely and not in drops for the simple SPRENGEL’S RESEARCHES ON THE VACUUM.reason that drops cannot be formed of the diameter of the calibre of the tube. The size of the drops depends upon the specific attraction of the molecules of the liquid the form and surface of the vessel from which the liquid drops the attraction existing between the liquid and the vessel and the resistance offered by the greater or less density of the air through which the drop falls. I have not been able to form in a vacuum drops of mercury larger than about 3 millimetres in diameter. Having failed in the use of a wide fall-tube I endeavoured to effect my object by the use of several small fall-tubes. Here how- ever another obstacle offered itself.It is exceedingly difficult to regulate the flow of mercury so evenly that exactly the same quantity shall run down in each separate fall-tube and I have found in practice that unless the flow of mercury can be so regu-lated simultaneous action cannot be ohtained in the fall-tubes. From these experiments I have found myself unable to produce a vacuum as quickly as with a common exhaustiirg syringe unless by the employment of inconveniently large quantities of mercury at a time. If speed is required I think the fall-tube should have the addition of the exhausting syringe which will take away more quickly the larger quantity of air and leave to the running mercury only the task of completing the exhaustion. By such a combination however the instrument loses much of its simplicity and offers by its numerous joints a far greater chance of leakage.For this reason where time is no object it will be preferable even to do without the gauge and to use the instrument in its simplest form as represented in Fig. 1. The operator will soon learn from observation of the way in which the drops fall down when the exhaustion is completed. Without the auxiliary air-pump the ex-haustion of a receiver of the capacity of about half a litre will take from 20 minutes to half an hour. Though this may appear to be a long time I have no doubt this method will be found after all the quickest and simplest for producing a vacuum as irearly perfect as I have been able to produce. The slowness of the action is obviously due to the smallness of the bore of the fall-tube.As soon as the calibre of this fmbe is increased the time of the operation rapidly decreases for the contents of two cylinders of the same height are to each other as the squares of their radii and the time of the operation ought to decrease in the inverse ratio. The proper size of the bore of the fall-tube is 2+ to 22 millimetres. As soon as I exceeded this SPRENGEL’S RESEARCHES ON THE VACUUM. limit I invariably found the vacua less perfect. It is not difficult with thesewider tubes to raise the mercury in the gauge to the height of the barometric pressure minus 1 or even 4 millimetre; but I have iiot been able to obtain with them vacua so near perfection as I have been enabled to obtain by the use of fall-tubes of 2$ millimetres calibre.The explanation of this fact must be sought for in the size of the drops which as it appears to me must in fall- ing down exercise a certain pressure against the side of the tube thereby preventing the denser air underneath the drop from finding its way again to the part of the tube above the drop where the air is more rarefied. I have not obtained better vacua by the use of fall-tubes of a calibre less than 29 millimetres. Before I proceeded to test the efficiency of the instrument I directed my whole attention to the construction of air-tight joints ; in this however I did not succeed. It is a well-known fact that barometers become inaccurate in time as air finds its way into the Toricellian vacuum between the glass and the mercury enclosed in it.The vacuum in my instrument is of course exposed to the same sources of imperfection. (To offer *a greater resistance to the air which might enter from the funnel I have given to the tube zyo the form of a U tube.) Leakage however happens in a far less degree from this cause than from the imperfection of the caoutchouc joints. Among the numerous modifications I have tried I con-sider the one which I have before described as the most practicable. With this joint the mercury in the gauge does not sink more than about + millirnetre in 24 hours. Williamson and Russell,* in constructing their admirable apparatus for the analysis of gases have met with the same difficulty and I am able to corroborate their statements.The porosity of solid bodies is astonishing and one is almost compelled to think that glass vessels are the only ones impenetrable to gases. The following degrees of exhaustion have been made with cold mercury which had merely been filtered through paper to free it from visible impurities but which had neither been dried nor freed from air by special means. I havk worked with heated mercury (100 to 15OoC.) but have not noticed much difference in the perfection of the vacua. Even if some slight advantage could be obtained by it its use would be objectionable prac- tically speaking from the risk of endangering the health of the * Chem. SOC.J. [Z] ii 238. SPRENGEL’S RESEARCHES ON THE VACUUM. operator.When the mercury is heated and allowed to run down quickly the instrument is at the same time converted into a sort of electric machine In the dark flashes of electric discharge are seeu to light up the exhausted tubes and sparks may be drawn at intervals from the basin in which the mercury collects as from an electrophorus. The fall-tube invariably becomes soiled after some time by some impurity in the mercury and particularly after the employment of heated mercury. I attribute this to the oxidation of the mercury arising from this electric action which must be favourable to the formation of ozone. I have to mention that to attaiil. high vacua the fall-tube must be clean as well as the mercury. The length of the fall-tube in the instrument before US is calculated for the use of a liquid having the specific gravity of mercury.Of course as the specific gravity of the liquid employed becomes less the fall-tube must be longer. Practically however water is the only liquid that need be considered in place of mercury and I have no doubt that an instrument adapted for water would fiirnish a simple and most efficient exhausting machine. It is cot unlikely that snch an instrument might possess advantages wltich air-pumps of other constructions have not particularly in hilly countries where the large volurrie of a natural waterfall might be rendered available. i.now come to consider the way in which this instrument acts. It is obvious that it stands in a near relation to the Trompe or Catalonian bellows the old arid well-known contrivance for pro-ducing a blast.My iristrunierit is merely the reverse of the trompe with this addition that the supply of air is limited while that in the trornpe is unlimited. The theory which explains the action of the trornpe will at the same time explain the action of my instru- ment. The theory of the trotnpe has repeatedly been treated by distinguished philosophers as Venturi Magnns Buff and others. It would lead me too far to enter upon a criticism of their opinions which appear to me partly erroneous partly not to the point. In my opinion the action of both instruments may in all cases be satisfactorily deduced from Kepler’ s law of the iiniformly accele- rated motion of bodies. When the clamp z is opened ouly n certain and almost uniform quantity of mercury (or any other liquid) can pass at a given time.As soon as a particle of mercury has arrived at x it is under the influence of the general law of gravitation. It must sink and move with uniformly accelerated VOL. xvirr. C SPRENQEL’S RESEARCHES ON THE VACUUM. velocity. The same may be said of the second or third particle of mercury; but while the second one is starting the first one has accomplished a portion of its way and when the third is starting the distances between one and two and two and three are not equal but unequal. A vacuum must therefore have been formed between them and hence the tendency of the air to restore the disturbed equilibrium ie. by rushing in if the instrument is open or by expanding if a receiver is attached.If the tube into which the liquid runs is larger than the column of liquid which the atmospheric pressure can support the air in the receiver will of course expand to its last degree. If the mercury is allowed to fall down in c d in drops it will act in exactly the same manner as the piston in a common air-pump. These drops are so to speak liquid pisto 11s. The chief excellence of this instrument appears to me to be its simplicity and the great perfection with which it performs its work. To ascertain the degree of exhaustion I had at first to resort to a comparison with the barometer but 1 have not been able to make a barometer in which the merciirial column stood higher than in the gauge of my instrument though the barometers were constructed with care and the readings made by means of a cathetometer.Though my instrument was not air- tight and consequently riot perfect this apparent equality of the levels in the gauge and iii the barometer is easily accounted for upon reflection by the fact that the liuman eye is not able to dis- tinguish between TB$Tc or even i.&T part of a millimetre. But being curious to see how ninch the instrument even in its present imperfect state can do I have taken particular pains to ascertain it; I have tried diffcrent WitYS but T will describe only that which appears to be the hest and most efficient. It is simply the application of D urna3’ method for the determination of vapour- densities.I took a receiver of the form R (Fig. Z),a bulb extended on both sides into a capillary tube one of which was open and attached to the instrument while a portion of the other was broken off and the aperture sealed. The part taken of and having consequei:tly the same calibre as the portion left attached to R was preserved. The receiver was now exhausted and taken off by sealing it at A. This point was broker1 under the mercury which had just run through the instrnunent. I did this to meet the objection that boiled mercury might absorb the remainder of the air in R while on the other hand mercury containing more air SPRENGEL'S RESEARCHES ON THE VACUUM. 19 might give off some of it and allow it to enter the vacuum. If now the receiver had been perfectly exhausted the mercury would have filled it completely.This however was not the case a very small air-bubble always remaining at the end of the sealed capillary tube. The capillary tube was broken cjff at f and the mercury contained in R was collected and weighed. Inh the capillary tube first broken off from tile receiver I now introduced by suction a small particle of meicury of exactly the same length as the particle of air in f. This mercury was then placed in a delicate balance and weighed. The weight of this particle of mercury bears the same proportion to the weight of mercirry in R as the weight and volume of air remaining in R after ex- haustion bears to the weight and volume of air in R before exhaustion. The highest proportion I have attained in this way is +g5- and upon the average I consider it not a difficult thing with the present means to exhaust a receiver to Ts&aco From this it is obvious that a barometer may be made by simply exhausting and sealing a tube one end of which is then broken under mercury.I have actually made some in this way. I have applied the finest reaction for the presence of gases viz. the absence of any electric discharge in a perfect vacuum.* The few tubes I have made hitherto always showed a slight discharge which however I have chiefly att.ributed to the presence of mercurial vapour though I could have ascertained it by means of spectrum analysis. I have tried to reiiiove this vapour by introducing between the instrument and the elect rical tuhe a tube exposed to a cold of -10' C.or filled with fiiiely divided gold or freshly ignited charcoal but the intensity of the whitish- green electric light appeared to be not much diminished by these means. From this I infer that the mercurial vapour has either riot been condensed or the supporter of light is due to another body. Being aware of the subtlety these experiments require and of my shortcomings in their perforoxmce I should not like to say more about them now but I hope to repeat and study them with care hereafter. At any rate I have learned this much that duriiig the exhaustion with my instrument the colours of the electric discharge change from intense red and blue to faint white and green; that I have passed the limit at which the density of the air is most favourable to the electric discharge ; that the stratifications are exhibited in an admirable * Gassiot Phil.Trans. 1858-9. c2 SPRENGEL’S RESEARCHES OF THE VACUUM. manner; and that the instrument will prove useful for the per- formance and study of these experiments which on account of their beaiity have excited so large an interest. Before concluding this paper T sl?orild like to say one word more about the theory of the action of the instrument. Is the action entirely due to the accelerated velocity of the mercury and the elasticity of the air? I answer yes. Struck hy the extraordiriary attenuation of the air and misled by Ventnri’s theory I was at first much inclined to attribute the action partly to two other agencies the attraction of gases to liquids and their abwrptioii by liquids.But the follming experiments showed me that I was mistaken. I forced water through short T-pieces at different velocities using from a slight pressure up to that of several atmospheres and I have not been able to raise water in a tulle connected with the other branch of the T-piece to a greater height than that corresponding to the length of the tube from which the water was expelled. But this does not show so much as the fact that when the calibre of the fall-tube is larger than 2+ milliinetres it is impossible to raise the mercury in the gauge so high as it starids in the barometer however long the mercury ma>* be allowed to run and whatever quantity may be used.This shows that the action is entirely mechanical that tli= air expands and is cut off portion by portion by thy falling drops of mercury and that when the air is highly attenuated these drops must entirely fill the tube and even exercise a slight pressure against the sides of the fall-tube otherwise the action will cease as in a common air-pump through the rron-action of the valves. I am under the impression that the use of better materials and the application of greater skill than I have hitherto employed will be followed by still bettcr results and will not improbably furnish instruments capable of producing vacua perfect to our senses ; and even if we should not succeed in riiaking a perfectly air-tight joint the end will still be attained if we can only succeed in carrying off the air more quickly than it can enter.At any rate the immense elasticity of air is here displayed in a stinilting manner and there is a very wide interval between the attenua- tion of 1,300,000 to the density of gases which must exist in the powder-chambers of carmons or mines at the nioment of explosion. Another striking fact is that Bxhaustioa of mB$mThas been made with cold common mercury doubtless containing a considerable quantity of air and moisture which one would expect to be set free BROUGHTON ON ANHYDRIDES AND ETHERS. and enter the vaciium as soon as the mercury so violently agitated passes along the fall-tube. But the particles of these absorbed gases which are set free on boiling must at common temperatures be so intimately connected with the mercury that their expanding or gaseous properties are lust as in the oxygen of oxide of mercury.The main fact which I have established in this paper may be shortly stated to be that ij'a liquid be allowed to run down a tube to the upper part of which a receiver is attached by means of a lateral tube and if the height at which the receiver is attached be not less than that of the column of the liquid which can be sup-ported by the atmospheric pressure a vmuum will be formed in the receiver minus the tension of the tiquid etnployed. The properties of highly rarefied gases and the ccrnditions of that remarkable space it1 which there is nothing have hitherto been scantily examined though this subject is suggestive of in-teresting questions the solution of which I hope to treat of in my further investigations that border on the so to speak negative side of naturaI philosophy.The above experiments * were performed in the laboratory of St. Bartholomew's Hospital and I am glad to have the oppor- tunity of acknowledging most gratefully the facilities which have been offered to me by Dr. Odling in the prosecution of them. The instrument may be seen at Elliott Brothers Strand London from whom it may be obtained.