]\gt ; sation of Extremely Deposits .
543 NOTES ON DIAGRAIsI .
Ordinates .
Left-hand numbers refer to upper curves , right-hand to lower .
Abscissoeratio alkali to acid .
Upper numbers to upper curves , lowel numbers to lower curves .
Upper plain line\mdash ; values calculated by formulae and at and dotted alues calculated by formula 1 Auerbach 's values at oken line\mdash ; values calculated by formulae owerO.upper .
Lower Curves\mdash ; Dilute carbonate 011 larger scale .
plain line\mdash ; values broken line\mdash ; values at experimental values as quoted .
N.B.\mdash ; The third points in Tables I and II have been erroneously plotted as and of Extremely Thin Radioactive Deposits .
By A. B. WOOD , M.Sc .
, Oliver Fellow and Assistant Lecturer in Physics , University of Liverpool .
( Communicated by Prof. Sir E. Rutherford , F.R.S. Received July 1 , 1915 .
) Introduction .
Within the last few years papers dealing with the volatilisation of radioactive substances have been published by various authors .
The principal aims of the experiments described in these papers may be classified under three heads : \mdash ; ( a ) To determine the temperatures of volatilisation of the various members of the active deposits of radium , thori , and actinium ; in some cases with a view to the ciassification of these substances in the periodic system .
( b ) To prove that these extremely small quantities of matter form definite chemical compounds in a manner similar to that of the commoner elements .
( c ) To separate the various members of the series from one another .
Other interesting results , throw light on the phenomena of volatilisation , have also been reported , but these have always formed a subsidiary part of the researches described .
Makower* determined the *Makower , Manch .
Lit. and Phil. Soc vol. 53 , Part II 1909 ) , and 'Le Badium , ' vol. 6 , p. .
Wood , more recently by the author .
S These authors have also shown the possibility of separatinCro radioactive substances which have not yet been separated by other methods .
possible sources of error in the determination of volatilisation temperatures .
facto.hich mnfluence tisation othese eyHence iddition tehaviour oubstance mubstance whose behaviour weated tarious hemperaturesbe there substances oharacter Barratt andOne fcommon txperiments mentioned , eference tecessary , nvestigation oature jentioned , hoose rhin films was desirable .
Wood , in studying the volatilisation of thorium and thorium , showed thatthecurveconnectingquitesimpleinthecaseofthoriumB , whereasthecurveforthoriumCwastheamountvolatilisedwithtemperaturewas more complex .
The author has also shown that the volatilisation curve for thorium is similar to that for thorium B. We may thus reasonably assume that the volatilisation curve of thorium is typical of that for any single constituent in the active deposits of radium , thorium , or actinium .
Consequently thorium is the substance employed in present experiments for an investigation of the various influences which aflect the volatilisation of the active deposits .
: : * Schrader , ' Phil. Mag vol. 24 , p. 125 ( 1912 ) .
Wood , Phil MDecember , Barratt aPhys .
Proc. JuneRussell , Phil. ' Extremely Thin Deposits .
Apparatus and Experimental Procedure .
The apparatus employed is shown diagrammatically in .
An electric furnace*whose temperature coul be measured accurately by means of a platinum thermometer and Callendar-Griffiths , was used to heat the active material .
The furnace and thermometer have been described in detail in a previous paper .
* aa .
Porcelain lining of furnace .
tinum thermometer .
bb .
Fused silica tube . .
Light nickel spoon . .
Tap connecting to pump and pressure gauge .
Temperatures above 1100o C. were estimated ) extrapolation of the curve connecting temperature and current through the furnace .
The active substance was placed inside a tube of fused silica , closed at one end , arranged centrally in the furnace and symmetrically with regard to the platinum thermometer see .
By means of a long nickel spoon , fixed to a ground glass joint which closed the cold end of the tube , the active material could be introduced or removed quickly from the silica tube .
The nickel spoon was of very small heat capacity , and attained the temperature of the furnace in a very short time .
This point is of importance , and will be referred to later ( see Section 3 ) .
The silica tube was connected by means of a side tube to a vacuum pump and pressure gauge , thus making it possible to heat the active substance at any desired pressure .
This arrangement also proved extremely useful in evacuating the tube between successive experiments , in order to remove any volatilised active material which would necessarily be present in the tube .
A strongly emanating source of radio-thorium ( equivalent in activity to 1 mgrn ] .
of radium bromide ) was used as a supply of active deposit .
By a simple arrangement this could be obtained ou one side only of the plate exposed to the emanation , and by varying the conditions of exposure a considerable range of activity could be obtained .
In all experiments , except a few of those described in Section 2 , exposures were made without electric field , a thin layer of tissue paper separating the radio-thorium from the metal foil , so that the possibility of thorium X being deposited on the foil was reduced to a minimum .
The active deposit * A. B. Wood , the activity were taken and the mean time of observation noted .
The deposit was then placed in the nickel spoon introduced into the furnace for the desired interval .
When a period of not less than six hours had elapsed after the removal of the active deposit from the furnace , the -activity was measured and at the end of this time the atilised portion of the deposit had again attained .
equilibrium .
* From the observed and calculated values of the activity at the time of the second set of measurements the percentage volatilised is easily deduced .
At the end of every experiment similar to the one just outlined , the silica tube was evacuated to a pressure of about 2 cm .
of mercury in order to free it as far as possible from the volatilised portion of the deposit .
Factors Irlflu arlier experiments have shown that the qation of a thin of radioactive matter ries enormously with the conditions to which it is exposed .
For example , the kind of surface on which the film is deposited has a 1narked effect , whilst several experi nenters have recorded notable variations with different ases on the furnace tube .
It is unnecessary at this stage to ellter into a discussion of numerous other sources of influence on this phenomenon , but the following list will serve as a guide to some of the principal factors wlrich have been investigated in detail .
The effects of variation on the ving conditions have been studied:\mdash ; 1 .
Nature of the surface on which the active substance is deposited .
2 .
Thickness of the layer of active material .
3 .
Period of heating .
4 .
Temperature .
5 .
Pressure of the surrounding .
gas .
6 .
Nature of the surrounding gas .
The first five of these factors will be considered in the order given above .
The consideration of the sixth , i.e. the influence of the nature of the gas surrounding the active material , will be discussed in a later communication .
* See paper by Barratt and Wood , loc. Extremely Thin Deposits .
1 .
fluence of Surface.\mdash ; Before investigating the other factors which may the volatilisation of the thorium , it is necessary to choose a suitable surface as a " " support\ldquo ; or base\ldquo ; for the deposit .
Barratt and Wood showed the necessity for a perfectly clean surface , since a film of grease , or other substance hich easily volatilises , will carry away the active material with it .
It is essential then to use some substance as " " base\ldquo ; which can be cleaned easily and thoroughly , either by heating or by treatment with acids .
Platinum and quartz are both ideal in these respects .
It is interesting , however , to examine other substances as to their suitability .
Substances whose melting points are below 1000o C. can at once be missed ; , all substances the volatilisation of which is detectable below this tenlperature are unsuitable .
* in his work on volatilisation of radium from surfaces of nickel , platinum , and quartz has shown that volatilisaLoion of the active substance is appreciable in all cases between and C. All the radium was driven off nickel and platinum at 1200o C. bnt not from the quartz till a temperature of 1300o C. was reached .
Barratt and Wood , using thorium active deposit , confirmed this observation of Makower that the initial temperature of volatilisation was the same whether the active substance was deposited on quartz or platinum : but found it impossible to volatilise the whole of the deposit from either quartz or platinum even after heating at 130 C. These conclusions are supported by the results of the present experiments .
It was found that after heating the active deposit on platinum for 10 minutes at a white heat ( temperature estimated at C. in the blowpipe flame there still remains an appreciable amount of thorium unvolatilised .
A similar result was obtained wfien quartz was used as the base .
With to unvolatilised portion of thorium , the experiment is instructive .
A platinum foil coated with thorium active deposit was heated for about 60 minutes at 120 C. and approxinlately 99 per cent. of the thorium volatilised .
The portion remaining on the platinum foil , i.e. the 1 per cent. , was allowed to attain equilibrium , when it was heated for the same lime at 120 C. It was now that o1lly 18 per oent .
of this remaining portion of the original deposit was volatilised .
These observations seem to indicate that a certain fraction of the molecules of the active deposit are held to the surface of the platinum by forces of cohesion due to the molecules of platinum in their immediate neighbourhood .
Possibly this small percentage of the active deposit has penetrated below the * A list of elements with their melting points , boiling points , and temperatul.es of initial volatilisation is given in a table in Kaye 's ' ) Makower , ' Manch .
Lit. and Phil. Soc vol. 53 , Part II 1909 ) .
quite probable that a small portion of the active deposit is carried off by this means , in addition to the normal volatilisation .
When the active material was heated on a nickel surface , the effects observed were very conflicting .
The temperature at which volatilisation colnmenced reed fairly well with the temperatures noted when platinum and quartz surfaces were used .
At temperatures , however , divergences were sometimes observed\mdash ; the results seeming to depend on the condition of the nickel surface before exposure to .
the emanation .
Thus it was noticed that the results were more consistent when the nickel surface , , before exposure to the emanation , was oxidised by heating strongly in blowpipe flame .
The following shows the result of a series of experiments using nickel and platinum surfaces:\mdash ; ( 1 ) First Heating at 1200o C. quite that a small portion of the active deposit is carried off by this means , in addition to the normal volatilisation .
When the active material was heated on a nickel surface , the effects observed were very conflicting .
The temperature at which volatilisation colnmenced reed fairly well with the temperatures noted when platinum and quartz surfaces were used .
At temperatures , however , divergences were sometimes resulffi seeming to depend on the condition of the nickel surface before exposure to .
the emanation .
Thus it was noticed that the results were more consistent when the nickel surface , , before exposure to the emanation , was oxidised by heating strongly in blowpipe The following shows the result of a series of experiments using nickel and platinum surfaces:\mdash ; ( 1 ) First Heating at 1200o C. quite that a small portion of the active deposit is carried off by this means , in addition to the normal volatilisation .
When the active material was heated on a nickel surface , the effects observed were very conflicting .
The temperature at which volatilisation colnmenced reed fairly well with the temperatures noted when platinum and quartz surfaces were used .
At temperatures , however , divergences were sometimes resulffi seeming to depend on the condition of the nickel surface before exposure to .
the emanation .
Thus it was noticed that the results were more consistent when the nickel surface , , before exposure to the emanation , was oxidised by heating strongly in blowpipe The following shows the result of a series of experiments using nickel and platinum surfaces:\mdash ; ( 1 ) First Heating at 1200o C. quite that a small portion of the active deposit is carried off by this means , in addition to the normal volatilisation .
When the active material was heated on a nickel surface , the effects observed were very conflicting .
The temperature at which volatilisation colnmenced reed fairly well with the temperatures noted when platinum and quartz surfaces were used .
At temperatures , however , divergences were sometimes resulffi seeming to depend on the condition of the nickel surface before exposure to .
the emanation .
Thus it was noticed that the results were more consistent when the nickel surface , , before exposure to the emanation , was oxidised by heating strongly in blowpipe The following shows the result of a series of experiments using nickel and platinum surfaces:\mdash ; ( 1 ) First Heating at 1200o C. ; After equilibrium had again been attained , the foils were reheated .
Boberts , ' Phil. Mag. , ' 6 , vol. 25 , p. 270 ( February , 1913 ) .
Extremely Thin Radioactive Deposits .
( 2 ) Second Heating at 1200o C. A third heating in the blowpipe flame ( temperature between 1500o C. and 1600o C. ) showed that there was a of the active deposit left from ( 2 ) , still unvolatilised .
It will be clear from these results that it is with only very great difficulty that the last 2 or 3 per cent. of the active deposit can be volatilised .
Again , it is more difficult to remove the active deposit from nickel than from platinum .
As one would expect , the nickel , on removal from the furnace at these high temperatures , is always coated with a thin film of black oxide , which renders this substance unsuitable as a base for several reasons\mdash ; in particular , the complication of the -ray measurements due to the absorption in the oxide film .
Again , volatilisation of this film might introduce serious errors in the measurements .
A surface of iron or steel is objectionable for similar reasons .
Using copper and brass surfaces this trouble of oxidation was still further increased , the comparatively low point of copper ( 1084o being another serious objection to its use at temperatures .
The metals osmium , iridium , and tungsten would be excellent for the purpose , since volatilisation of these elements has not been detected below 1400o C. in account of .
expensive nature , however , they were not used in the investigation .
From the above considerations it is clear that platinum and quartz are the most suitable materials to employin a research of this kind .
On account of the ease with which it can be cleaned and manipulated platinum has been used as base for the active substance in all experiments subsequently described .
2 .
Density of the Deposit.\mdash ; When referring to such minute quantities of substance as the active deposits from the various radioactive emanations we cannot use strictly the term " " thickness Assuming a uniform distribution of particles , a simple calculation of the thickness of a very active layer of thorium active deposit , equivalent in -activity to mgrm .
of radium per square centimetre of platinum , shows that only one thousandth of the platinum surface is covered by a layer one molecule deep , i.e. for every now considering .
It ] be shown later that the amount volatilised is not a linear function of the time of There are twc methods open to us for varying the density of the deposit : ( a ) by the time of exposure to the etnanation , without employing an electric field ; ( b ) by exposing for different times with an electric field .
( a ) By comparing the -activity of the active deposit with that of a standard source of radium , a simple calculation gives us the approximate number of active deposit molecules per square centimetre .
Variations of exposure from 1 tc 40 hours have been made Yvithout electric field , representing a range of density from to active deposit molecules per square centimetre\mdash ; the number of platinum molecules in a surface layer being of the order No regular variations in the amount volatilised with change of density of the deposit were observed , all diHerences being well within the experimental error .
Hence , under these conditions , the volatilisation is independent of the time of exposure to the emanation .
( b ) If an electric field is used when exposing to the emanation , the density of the deposit can be increased to about one hundred times the maximum possible by method , i.e. to a density of about molecules per square centimetre .
, however , fairly large and almost unaccountable variations occur .
Several explanations of these variations can be suggested .
They may be due ( 1 ) to the collection of dust particles or other nuclei , coated with active deposit , ( 2 ) to the formation of charged aCes of active deposit , and to the non-uniformity of the layer of active deposit collected by this method , as compared with that obtained by the ordinary process of sion .
In all cases , however , the amount volatilised from a plate exposed , till equilibrium is reached , by method ( b ) is greater for short periods of than that volatilised from a plate exposed by method ( a ) .
Extremely Thin Radioactive Deposits .
This gives strong support to the view that dust particles or other nuclei coated with the active deposit have been collected on the plate by the electric field used in consequently , when such a plate is introduced into the furnace , the dust burns rapidly , and carries away the actiye deposit with it .
This question will be discussed ] ater ( see Section 3 ) .
The Table gives a comparison of a series of results obtained by the two methods just outlined:\mdash ; Percentage thorium voIntilised nt of heating .
electric field .
ctric { ield .
Time Platinum foil exposed to emnnation ithou to ation 2 hours ' exposure .
) ' exposure .
24 hours ' exposure .
The above remarks must be modified , of course , when we are considering the volatilisation of the portion which remains after the platinum and active deposit have been once heated to temperatures over 1000o C.\mdash ; this has been explained in Section 1 .
For experiments of this character 1he results are always the same , whatever the density of the original deposit .
It appears , then , that the most consistent results are obtained when the platinum foil is exposed to the emanation without electric field , the volatilisation in that case being practically independent of the time of exposure .
In all subsequent experiments the active deposit was collected by exposure to the emanation for six hours without electric field .
3 .
Variation of Pcriod of Heating.\mdash ; A point of great ) ortance in the study of the problem of volatilisation is to determine the rate at which the active material is removed from the surface .
The experimental procedure was as follows : After the furnace had attained a steady state ( usually after about 2 hours ' preliminary heating the active deposit , previously measured , was introduced into the furnace for a definite period , by means of the nickel spoon .
On removal its activity was measured as explained above .
The silica tube was then evacuated to remove volatilised active deposit , and a second source of active deposit introduced for another different period .
This process was repeated u1ltil a series of observations had been made the relation between the time of The result of such a series of observations is shown in fig. 2 .
Time of heating ( minutes ) .
prominent features of these curves will be noticed at once .
( 1 ) The a1nount volatilised increases at first rapidly , then more and more slowly as eating is continued , .
the rate of volatllisation decreases rapidly time of heating , particularly at high temperatures .
( 2 ) The initial rate of volatilisation increases very rapidly with small increases of temperature .
This point is shown more clearly in fig. 4 , which will be discussed in Section 4 .
Another , perhaps less tant , feature of these curves will be oticed from a consideration of the rate when a large percentage has bee volatilised .
A decided flattening of the curves becomes evident as the ioned iection 1amount viQthe turve ( hown iigure)Cruns pwith 9dinate.eriod oeating temov.traces ofThe first ation ourves wests isExtremdy Teposits .
that they are exponential .
It seems quite reasonable to suppose .the rate of volatilisation is proportional to the amount of active substance present on the platinum foil at any particular time , i.e. we are dealing , as in this case , with a comparatively small number of molecules .
The facts do not support this view , however , and it becomes necessary to look for some other explanation .
Of course , it is quite probable that several factors are nvolved in the complex behaviour of the active deposit ; hence a ) sideration of the following hypothetical case may throw some cvht on the problem .
Suppose we could obtain on a platinum surface a layer of active deposit of measurable thickness ( sa .
mm Let us consider what happens this layer is heated .
Up to a certain critical poit , in fig. 3 , Thickness of layer ( diminishing ) .
Maximum less at .
Infinitesimally small thickness at Zero thickness at would ) the rate of volatilisation to be constant , .
the amount volatilised is proportional to the time of heating , but when ) depth of the layer had decreased to one or two molecules the rate of volatilisation vould begin to decrease , at first slowly , then more rapidly , until only those molecules of the active deposit which are intermingled with molecules are left .
At this stage , in fig. :3 , the rate of volatilisation decreases more slowly , since the remaining nzolecules are subject to a different law of foroe\mdash ; the cohesion of the platinum molecules\mdash ; from that of the active deposit molecules volatilised in the earlier stages .
is a diagrammatic represeutatiom of the process just outlilled The latter portion Mr. A. B. Wood .
Volatilisation of of the curve , between the points and , represents the result of an actual experiment similar to the one described at the beginning of this seccion .
As was made clear in Section 2 , the thickness of the active deposits is correct wxpect , gating whick 1ayers octive depositto iortion ourve 1ying between tisation iever constant ibove suggestions a points and .
Unfortunately , we have no sources of thorium emanation at .
3 present available , from which we can collect such thick layers of active deposit by diffusion .
to the disturbing action of dust particles , when an electric field is used to increase the density of the deposit , the results obtained.in such cases cannot itimately be compared with those obtained by the diffusion method used in the above experiments .
We now come to the second and , perhaps , more important feature of the curves shown in .
the influence of temperature on the rate of volatilisation .
The problem bears a close similarity to that of the emission of ioDS from a hot metal .
Thus if we assume that the volatilised particles like a perfect gas , the calculation of the rate of volatilisation is analogous in practically all respects to that in the case just mentioned .
egarding the steady state as the result of a dynamical equilibrium between the molecules going from the layer of active deposit to the air and those going from the air to the deposit , we obtain the well-known relation connecting the rate with the absolute temperature Rate where A nd are constants which.can be determined experimentally .
The expression for the rate , just given , was first deduced by Richardson in his researches on emission of ions from hot bodies , and was subject to the condition that the number of free electrons per unit volume of the metal is independent of the temperature .
He has since modified his views* by assuming that the number of free electrons in the metal is proportional to , whence Rate The assumption just mentioned was made in order to explain the observed values of the specific heat of electricity .
More recently hasgiven another deduction of the second formula based on the quantum theory. .
* Richardson , 'Phil .
Mag vol. 23 , p. 604 Richardson , ' Phil. Mag vol. 28 , p. 633 ( 1914 ) .
Extremely Thin Radioactive Deposits .
Planck* also has obtained the relation by consideration of the vapour pressure of a solid at very low temperatures , using the assumption that the entropy of the solid at these temperature will be practically independent of the temperature , while that of the vapour can be considered like that of a perfect gas .
It seems very doubtful , however , whether the assumptions used by Planck will apply to the case under consideration , i.e. to the evaporation of a metal at high temperatures .
In this case , a lower power of the factor would be expected .
Both the relations just given require that the rate should be constant when the temperature is constant , but , as we have already seen , this is far from being the case , the rate varying considerably with the amount of material volatilised .
Consequently we must introduce a factor into the expression for rate , in order to make this quantity in agreement with observations\mdash ; the function expressing the rate as a function of the number of thorium molecules volatilised .
We now have Rate , ( 3 ) where or 2 .
The form of the function has already been discussed from the physical standpoint , hence we need only concern ourselves with a consideration of the remaining factor .
From observations of the initial rate of volatilisation at two different temperatures the values of A and in equation can be calculated , and the rates at any other temperatures deduced .
The following Table provides a comparison between the observed and calculated values of the initial rates at different temperatures .
*The observed rates at these temperatures were used in the calculation of the constants A and in the expressions for the rate .
PIanck , " " Die gegenwartige Bedeutung der Quantenhypothese fur die tische Gastheorie 'Vortrage uber die kinetische Theory der Iaterie der Electricitat \mdash ; Mathematische Vorlesungen an der Universitat Gottiugen , ' VI .
VOL. XCL\mdash ; A. 2 Mr. A. B. Wood .
sation of Considering the extreme sensibility of the expressions and to small changes of temperature , the reement between observed and calculated values is good .
It will be noticed , however , that the factor or has practically no effect on the value of the expression for the rate until the temperature reaches a high value\mdash ; thus both sions for the initial rate of yolatilisation agree equally well .
It is only possible , experimentally , to distinguish between the two forms when the temperature is in the neighbourhood of 1000o C. ; unfortunately at this temperature it is practically impossible to measure the initial rate .
Similar results are found if , instead of using the initial rates , we consider the rates of volatilisation at a later stage of the process .
Thus the rates of volatilisation at different temperatures whetJ 40 per cent. of the active material has been removed are in the same relative proportion as the initial rates at the same temperatures .
This only holds , however , up to a point where about 80 or 90 per cent. of the thorium is volatilised .
Beyond this point , the whole process is changed on account of the effects previously discussed .
The results of the experiments on the rate of volatilisation of the active deposit thus seem to support the view that the volatilisation proceeds on similar lines to the emission of ions ( positive or negative ) from a heated metal .
Other experimenters have shown*that there is a close si1mlarity between the laws of disintegration of a heated wire and the leak of positive electricity from it .
The present experiments support this view , although the only safe conclusion which can be given at present is that the formula for the volatilisation\mdash ; besides containing terms more slowly\mdash ; may be expected to contain the factor or , where is the gas constant and the latent heat of vaporisation : this result , of course , is well known from ordinary thermodynamical considerations .
4 .
Variation of the Temperature.\mdash ; From the poin of view of interest this factor should have been considered earlier , for several reasons it seemed advisable to defcr ib until the questions raised in Section 3 had been discussed .
Barratt and Wood ( loc. cit. ) the question by heating thorium active deposit for a constant time ( 15 minutes ) at various accurately measured temperatures .
Their work led to a result C. as the temperature of volatilisation of thorium B. In the experiments the procedure adopted was exactly the same\mdash ; the constant period of heating in this case again being 15 minutes .
Curve , fig. 4 , is a combination of the results obtained in the present work with those obtained in a previous research by Barratt and the author .
The observations made in the present research are indicated in the figure by * See J. J. TAomson 's ' Conduction of Electricity through Gases , ' p. 213 , Chap. VIII .
Extremely Deposits .
a circle\mdash ; those in previous work by Barratt and Wood by a combination of a circle and a cross .
It will be seen that the two sets of observations are in excellent agreement .
One important difference , however , should be noted .
Temperature ( degrees Curve A. At 760 mm. Curve B. At 3 mm. pressure .
Present paper .
Barratt and Woo Mr. A. B. Wood .
of The temperature of volatilisation , as given by Barratt and Wood , was slightly lower than 75 C.\ldquo ; \mdash ; it is evident from the figure that if the main portion of the curve is produced to cut the axis at , the temperature obtained , 74 C. , agrees very well with the value just given .
This temperature , C. , is that at which the thorium begin to volatilise very rapidly .
Below this temperature , however , a small amount of volatilisation is observed , but this quickly falls to a quantity too small for detection at about 68 C. Of course , the lowest temperature at which it is possible to detect yolatilisation is even lower than this , but in that case the time of heatir-g must be considerably increased .
The following observations at temperatures below 68 will illustrate this point:\mdash ; The form of the curve , is interesting .
It will be that the amount volatilised increases at first slowly , then much more rapidly as the temperature increases .
When a fairly large percentage has been removed , however , the rate of volatilisation rapidly falls off with further heating at higher temperatures .
Of course , a series of curves similar to the one just described could be constructed from fig. 2 directly .
If the form of the function in equation ( 3 ) , Section 3 , were accurately known it would be possible to calculate the total amount volatilised in a yiven time at desired temperature .
5 .
Vctriation of Pressure of the \mdash ; All the experiments described up to this stage have been carried out with the air at atmospheric pressure in the silica tube .
We shall now consider the effects observed when the active deposit was heated at pressures considerably lower than this .
After its -activity had been measured the active deposit on the platinum foil was placed on the nickel spoon and introduced into the silica tube .
The air in the tube was immediately expanded into a bottle of large volume exhausted to the required pressure .
Hence it was possible to obtain almost instantaneously any desired pressure in the tube by that the volume of the receiver was very large compared with the volume of the silica tube .
A series of observations of this kind , keeping Extremely Thin Radioactive Deposits .
the time of heating constant ( 15 minutes ) , was made at different temperatures and pressures .
The result of such a series of observations at a pressure of 3 mm. is shown in curve , fig. 4 .
Comparing the two curves it is evident that the process of volatilisation takes place on very much } same lines in each case .
Thus the slow initial rise , followed by the much steeper pol.tion and ended by another slow increase , is characteristic of both curves .
In the " " low pressure\ldquo ; curve , however , the central portion is considerably steeper than the portion of the " " atmospheric pressure\ldquo ; curve , indicating that volatilisation is much more rapid in the former case .
Another ence b the curves is the lowering ( about C. ) of the initial temperature of volatilisation at low pressures .
It should be noticed , too , that at low pressures it is still found practically impossible to volatilise the last tlaces of the active deposit .
The following Table gives a few of the results of a series of experiments carried out on the variation of the amount volatilised at different temperatures , the pressures from 7 mm. to 3 mm. of mercury:\mdash ; The more important factors which influence the volatilisation of extremely thin films of the radioactive deposits have been ated experimentally .
Each of these sources of influence has been considered separately by eliminating , as far as possible , the disturbing effects of the others .
for reasons given in the earlier part of the paper , has been chosen as typical of all similar constituents of the active deposits of radium , thorium , and actillium .
560 tilisation of Extremely Thin Radioactive Deposits .
been examined:\mdash ; ( 1 ) The surface on which the active substance is deposited .
collected on the surface .
ised \ldquo ; withtimeofheating\ldquo ; series oifferent temperatures aVariation oeriod oeating .
Curves connecting percentageThe ience oowing factors oroce , olatilisation hVariationoftheamountpersquarecentimetre\ldquo ; oftheactivedeposit ( 4 ) Variation of temperature\mdash ; the time of heating remaining constant .
( 5 ) Variation of the pressure of the surrounding gas .
It has been shown to be extremely difficult to remove , by heatin , alone , the last traces of the active deposit from surfaces of quartz , nickel , or platinum .
A suggested explanation of this is given and verified experimentally .
The most consistent results are obtained when the surface is exposed to the emanation without electric field , in which case the volatilisation is independent of the time of exposure to the emanation .
The volatilisation is shown to obey practically ) same law as the rate of emission of ions from a hot body , this expression\mdash ; Rate \mdash ; being modified by the introduction of a factor which depends on the amount of active deposit which has been volatilised .
The function is considered from a physical standpoint , the mathematical relation being extremely complex .
A comparison of the volatilisation curves at pressures of 760 mm. and 3 mm. is given , a lowering of the initial temperature of volatilisation at low pressure being observed .
The volatilisation curve at atmospheric pressure rrees extremely well with that given in a previous paper by Barratt and Wood ( loc. cit In conclusion , I should like to express my warmest thanks to Prof. L. B. Wilberforce for his interest and encouragement , and to Prof. Sir E. Rutherford and Dr. N. Bohr for kind help and criticism .

