]\gt ; 515 On 's Rings formed by Metallic Reflection .
By RICHARD C. bIACLAIfRIN , , LL. D. , late Fellow of St. John 's College Cambridge , Professor of Mathematics , Wellington , New Zealand .
( Communicated by Professor J. Larmor , Sec. R.S. Received August 15 , \mdash ; Read November 16 , 1905 .
) In 1876 , Sir G. Stokes brought to ) notice of the British Association* phenomenon of metallic reflection which he regal'ded as very renlarkable .
" " When Newton 's rings he said , " " are formed between a lens and a plate of metal , and are viewed by light polarised perpendicularly to the plane of incidence , we know that , as the of incidence is increased , the rings which are at first dark-centred , disappear on passing the polarising of the glass , and then reappear white-centred , in which state they remain up a gauging incidence , when they can no longer be followed .
At a high incidence the first dark ring is much the most conspicuous of the " " To follow the beyond the limit of total internal reflection , we must employ a prism .
When the formed between oo.lass and glass are viewed in this way , we know that as the angle of incidence is increased the rings one by one open out , uniting with bands of the same respective orders which are seen beneath the limit of total internal reflection ; the limit or boundary between total and partial reflection passes down beneath the point of contact , and the central spot is left isolated in a field .
" " Now when the rings are formed between a prism with a slightly convex base and a plate of silver , and the angle of incidence is increased so as to pass the critical angle , if common light be used , in lieu of a simple spot have a ring , which becomes more conspicuous at a certain of incidence well beyond the critical , after which it rapidly contracts and passes into a spot .
" " As thus viewed the is , however , somewhat confused .
To study the phenomenon in its purity we must employ polarised light , or , what is more convenient , analyse the reflected by means of a Nicol 's prism .
" " When viewed by light polarised in the plane of incidence , the show nothing remarkable .
They are naturally weaker than with glass , as the 'Brit .
Assoc. Report , ' Glasgow , 1876 ; 'Math .
) .
Papers , ' vol. 5 , pp. 361\mdash ; 364 .
Professor Stokes was at that time inclined to see in the phenomena the influence of a layer of gradual transition or whatever cause produced similar effects with diamond .
The author learns from Prof. Larmor that Sir George Stokes had an intention for some time of mining theoretically the formation of Newton 's rings on a metallic surface .
[ The author is indebte Dr. .
Havelock for the correction of the Prof. R. C. Maclaurin .
[ Aug. 15 , interfering streams are so unequal In intensity .
They are black-centred throughout , and , as , they open out one after another on approaching the limit of total reflection and disappear , leaving the central spot isolated in the field beyond the limit .
The spot appears to be notably smaller than with glass under like conditions .
" " With light polarised perpendicularly to the plane of incidence , the rings pass from dark-centred to bright-centred on passing the polarising angle of the glass , and open out as they limit of total reflection .
The last dark ring to disappear is not , however , the first but the second .
The order to the first bright ring within the polarising angle of the glass , remains isolated in the bright field , enclosing a relatively , though not absolutely spot .
At the centre of the spot the glass and metal are in optical contact , and the reflection takes place accordingly , and is not total .
The dark ring , too , is not absolutely black .
As the angle of internal incidence increases by a few degrees , the dark ring umdergoes a rapid and remarkable change .
Its intensity increases till ( ill the case of silver ) the ring becomes sensibly black , then it rapidly contracts , squeezing out , as it were , the bright central spot and forming itself a dark , larger than with glass , isolated in the bright field .
When at its best it is distinctly seen to be fringed with colour , blue outside , red inside ( especially the former ) .
This rapid alteration taking place well beyond the critical angle is very remarkable .
Clearly there is a rapid change in the reflective properties of the metal , which takes place , so to speak , in passing a certain angle determined by a sine greater than unity .
" " I have described the phenomenon with silver , which shows it best ; but speculum-metal , gold , and copper show it very well , while with steel it is far less conspicuous It is with the object of accounting for these phenomena that the following investigation of Newton 's rings formed by metallic reflection has been undertaken .
Suppose that we have a piece of lass and metal separated by a thin layer 1905 .
] On Newton 's Rings formed by Reflection .
5 17 of air of thickness ( .
A wave of unit amplitude represented by ) will at the passage from glass to air , give rise to a reflected wave in the glass and a refracted wave in the air .
Let the quantities to and be and at the passage from air to metal , and and at the from air to glass .
Suppose , further , that a wave of unit amplitude , represented by , becomes after traversing a distance in air .
Waves will be reflected to and fro between the glass and metal , and the waves that return into the cvlass will constitute a reflected beam represented by ad imf .
, provided The values of the quantities , etc. , will be iven by the theory of reflection and refraction , and in this way we can easily prove that and , results that can also be obtained by means of Stokes ' principle of reversion or by noting that when the metal is replaced by glass ( so that and coincide with and ) and the thickness of the layel of air is indefinitely diminished and becomes unity , then all the is transmitted and none reflected .
Making use of these relations we obtain for the reflected beam In general , and will be complex quantities .
We shall We thus get The intensity of the reflected beam is , therefore , The numerator is the square of the resultant of two vectors , and inclined at an angle .
If , then , there is to be absolute blackness anywhere we must have , so that and where is an integer .
It will be necessary to distinguish two cases\mdash ; first , when the of incidence is less than the critical angle of the glass , and second when it is greater than that angle .
In each case vibrations parallel and perpendicular to the plane of incidence must be dealt with separately .
Prof R. C. Maclaurin .
[ Aug. 15 , ( A ) the Critical Angle .
If be the angle of incidence , that of retraction , and the efractive index of the glass , we have .
The surface of separation is , the plane of is that of incidence , so that everything is independent of .
The components of the light vector are proportional to where is the velocity of light in air .
Putting we see that where is the wave-length in air .
Hence we obtain and .
We thus have where For a given angle of incidence , and are constants and the only variable in the expression for I is , which depends on the thickness ) of the layer .
Differentiating I with respect to , we find that I is stationary when ; i.e. , Putting and ( it appear from the sequel that is always real ) , we get Thus where is any , so that being an integer .
By giving different values to we see that there be rings of maximum and minimum intensity .
polarised perpendicularty to the Plane of Incidence .
If we suppose the transition from glass to air to be abrupt , then as given by Fresnel 's formula ; viz. , When lies between and the polarising angle , this gives and 1905 .
] On 's Rings formed by Metallic Reflection .
while when is greater than the polarising angle it gives and In both cases I is stationary when being an integer .
The maximum value of I is and the minimum is When is less than the polarising angle I is a maximum when is odd , and a minimum when is even ; when is greater than that angle I is a minimum when is odd and a maximum when is even .
At the angle so that ( a constant ) , and there are no If we suppose the transition from glass to air to be and not abrupt , then we know that enters by degrees instead of suddenly becoming at the .
Except , however , within a few rees of the polarising angle , is sufficiently well represented by Fresnel 's formula above .
In the immediate neighbourhood of the polarising we have where , and , , being a small constant depending on the thickness of the layer of transition and the law of variation of the refractive index within that layer .
* It thus appears that will be a small quantity , and that will always be less than and that will be small .
, however , will not be small , but will vary from to , being when at the polarising angle , and when is .
I is a maximum when is even , in which case its value is ) and a minimum en is odd , in which case its value is .
Owing to the smallness of the variations of intensity will be very slight , so that the in the neighbourhood of the polarising will scarcely be visible .
For the determination of the radii of the rings and their intensity nothing remains but the calculation of the constants .
For this purpose we shall take , which ives a polarising angle of and a critical angle of .
We obtain from the formulae already given .
* See ' Roy .
Soc. Proc , vol. 76 , 1905 , p. 61 .
Prof R. C. Maclaurin .
38 15 38 30 38 46 [ Aug. 15 , 0.5257 R2 and are determined from the formula where is a complex quantity given by Here is the " " refractive index\ldquo ; of the metal , , the ratio of the velocity of light in air to that in the metal , and is the coefficient of absorption .
We shall put .
For all the metals is large ; according to Drude 's experiments it is least for copper , where its value is , and greatest for zinc , where its value is .
For the two metals\mdash ; silver and steel\mdash ; to which we shall apply the theory , is and respectively , so that the error in and higher powers of will be very small .
To this order of approximation we have Therefore Whence where ' and By the aid of these formulae the following values of and have been calculated for silver and steel .
1905 .
] On Newton 's Rings formed by Reflection .
Silver .
Steel .
; ; We are now in a position to calculate the intensity I at any point .
For our purposes it will be sufficient to determine the maximum and minimum values of I and its values at and very near the centle .
We have already seen ( p. 519 ) that the maximum value of I is , and the llinimum Very near the centre , and we have is less than the polarising angle , and ' where is than that angle .
From these formulae and the numerical results on pp. and 521 the following table has been made Silver .
Steel .
522 Prof. R. C. Maclaurin .
The value of the intensity the centre cannot be obtained from these , for at that point there is no air space between the glass and metal , so that we have direct reflection from the metal .
The intensity in this case can be obtained from the mula for on p. 520 by replacing and by and respectively .
In this way we obtain the following values of the intensity for different values of ranging from to The radii of the rings can be readily obtained .
If be the radius of a ring , the curvature of the lens , we have very approximately .
We have seen that at any point on a ring of maximum or minimum intensity ; Hence we have The smallest ring corresponds to , the next to , the next to and so on .
From this formula we shall calculate the values of , and ( corresponding to the first four ) , for diffelent values of ranging from zero to the critical angle .
Silver .
Steel .
From these results it will appear that the rings open out as the of incidence increases .
The expansion of all the rings except the first is ver .
rapid when hearing the critical angle , and the radius of each increases to 1905 .
] On Newton 's Rings formed by Metallic Reflection .
infinity when the critical angle is reached .
With the first ring , however , there is very little change as the critical angle is approached , and the limiting value of is for silver and for steel .
Our formulae cannot , in strictness , be applied when is exactly equal to the critical angle , for then , and the fundamental series on p. 517loses its convergence .
We may , however , approach the critical angle as near as we wish .
We have seen that When is nearly is very small , and we have approx. Therefore the limit of which is .
Whence ' which gives the values for silver and steel recorded above .
From the formula ' we see that is a function of , so that , unless homogeneous be employed , the rings will be coloured .
and consequently are functions of , but the change in due to this fact will be inappreciable , and very approximately we shall have .
Thus , if , and be radii corresponding to Fraunhofer 's lines , and in the spectrum , A being in the extreme red and in the violet , we have , approximately , and In what way this will aflect the appearance of the rings will be seen most clearly by drawing curves to represent the intensity for various angles of incidence .
The following tables give the value of corresponding to the lines VOL. LXXVI.\mdash ; A. 2 Prof R. C. Maclaurin .
[ Aug. 15 , , and , the values of in the tables on p. being taken to correspond to the line Silver .
Steel .
The general effect of his dependence of the radius on the wave-length will be that each bright ring will change in colour from violet on the inside to red on the outside , whereas in the dark rings the order of colours will be reversed .
As , however , in the case of silver , the difference between the maximum and minimum intensities is nowhere great , the rings will not be conspicuous enough to show the colours well .
Another effect , well illustrated by the changing wonsidered allustrative fetals oThese rglect thange iptical constants oetals w types .
The ring system has in fact been employed by P. Glan to determine the change on reflection for different colours .
1905 .
] On 's Rings formed by Metallic Reflection .
tables above , is that\mdash ; for all the rings except the first\mdash ; the red of one will overlap the violet of the next .
This will diminish the sharpness of the rings , and make the first appear more conspicuous than the others .
Following Stokes , we shall compare these results with the case when the metal is replaced by glass of the same refractive index as the first piece .
The phenomena in this case are very well known and have been frequently discussed mathematically .
We have only to put and in the formulae of p. 517 , and we for the reflected beam Except near the polarising angle we may put or , so that and we get This has maxima and minima when , where is an .
If is even , , and if is odd , The radii are given by Where , we and , so that there is a black spot at the centre .
The following table gives the value of for the first four rings corresponding to , 2 , 3 , 4 respectively , the first one being bright .
Glass . .
, The maximum value of I , , its value at the first , third , etc. , is as follows:\mdash ; Prof. R. C. Maclaurin .
[ Aug. 15 , It will be well to illustrate most of these matters graphically .
Figs. 1 , 2 , 3 , and 4 represent the first four rings for values of FIG. 3 .
FIG. 2 .
FIG. 4 .
1905 .
] 's Rings formed by ffection .
below the polarising .
A glance at these figures will show that the rings are all dark-centred , that they gradually expand as increases , and that the rings are much less distinct with silver than with steel or glass .
The rings are most distinct with glass , but the difference between that and steel is not very marked .
Fig. 4 shows that as the angle is approached the rings rapidly disappear .
After passing the polarising angle the rings pass from darkcentred to bright-centred in the case of the metals ; but with glass they remain bright-centred throughout .
This is clearly brought out in ligs .
5 , 6 , 7 , and 8 which follow .
The rings are still very much more distinct with glass and steel than with silver .
There a close resemblance between the rings with glass and with steel\mdash ; the main difference that the rings with steel are bright-centred , but the brightness at the centre is not very appreciable compared with that at the rings .
The rings with steel are slightly more distinct than glass slightly larger .
After the polarising angle the have ail expanded so much that it is convenient to use a smaller scale to represent the radius , and in the fJures that follow ( 5 to 8 ) the scale of is half that in the earlier figures .
FIG. .
FIG. 7 .
FIG. 6 .
Prof. R. C. Maclaurin .
[ Aug. 15 , The colour effects are brought out by figs. 9 and 10 , the first of which deals with the case where and the second with , i.e. , beyond the polarising angle .
The dotted curve represents the intensity for red , and the continuous curve that for violet .
FIe .
9 .
FIG. 10 .
Light to the Plane of Incidence .
In this case the influence of the transition layer on the phase is always very small , so that we may employ Fresnel 's formula throughout and put and We then ( see p. 518 ) have so that maxima and minima occur when being any integer .
The maximum intensity is and the minimum is The rings will be dark-ceutred throughout , instead of passing from darkcentred to bright-centred as with polarised perpendicularly to the plane of incidence .
We have the following values of obtained from the formula above : 1905 .
] On Newton 's Rings formed by Metallic Reflection .
0.530 and are determined from the formula ( cf. p. 520 ) .
This makes whence which gives where and By the aid of these formulae the following values of and have been obtained:\mdash ; Silver .
Steel .
530 Prof. R. C. Maclaurin .
[ Aug. 15 , From these results we derive the following values for the intensity at the bright and dark rings , and near the centre ( cf. p. 521):\mdash ; Silver .
Steel .
The value of the intensity the centre ( p. 521 ) is obtained from the formula for on p. 529 by replacing and by and respectively .
In this way we obtain the following values of the intensity as ranges from to For the radii of the rings we have the following values cf. p. Silver .
SteeL The influence of colour on the size of the is exhibited in the following tables:\mdash ; 1905 .
] On Newton 's Rings rmed by Metallic Reflection .
Silver .
Steel .
If we wish to compare these results with those for glass we have merely to substitute the values of in the formula of p. ) .
The radii of the rings are the same as for the case of perpendicularly-polarised .
The only difference is that the maximum value of I is now given by the following table :\mdash ; The graphical representation of these results is undertaken in the next three figures .
It will be seen that the rings with silver are very indistinct .
Prof. R. C. Maclaurin .
[ Aug. 15 , With steel , however , the rings begin , when is small , by being almost as marked as with glass .
They remain fairly distinct as increases , but the with glass gain in intensity much more markedly than with steeL In all cases the rings are dark-centred ; but , as will be seen from fig. 13 , we approach the critical angle the difference between the intensity at the centre and at the first ring is\mdash ; with the metals\mdash ; very slight , so that there will appear to be a bright spot at the centre .
FIG. 11 .
FIG. 12 .
FIG. 13 .
( B ) Beyond the Critical Angle .
Before we can proceed beyond the critical angle it will be necessary to discuss briefly the reflection and retraction of the " " surface waves\ldquo ; due to total reflection . .
1905 .
] On Newton 's Rings formed Reflection .
Supposing light to be due to displacements in a rotational ether , we have the following expressions for the kinetic energy and work function in an isotropic medium:\mdash ; ; Here is an element of volume , the density , the displacement ( fgh ) .
curl , the velocity of light in free ether , a constant which represents the refractive index in the case of a transparent medium .
The dynamical equations and boundary conditions are most simply obtained from the Principle of Action , which makes .
This leads at once to three dynamical equations of the form , and shows that the boundary conditions are satisfied if , , are continuous , where ( lmn ) are the direction cosines of the outward normal to the bounding surface .
The displacement must also be continuous to avoid rupture of the material .
We take to represent the light vector .
The surface of separation is , the plane of is that of incidence , so that everything is independent of Light Polarised Perpendicvlarly to the Plane of lnci In this case is a vector parallel to the plane of incidence , so that is at right angles to that plane .
Thus In the glass ( incident ) ( reflected ) .
In the air ( refracted ) , where is the velocity of light in the glass .
The amplitudes of the three waves are in the ratio 1 : : The dynamical equations require .
Therefore the sign of the radical being chosen so as to make diminish as increases .
The boundary conditions give , and 534 Prof. R. C. Maclaurin .
[ Aug. 15 , where From these we get so that ] ) .
This esult could also be obtained by transforming Fresnel 's formula where is complex and given by the equation .
Moreover , from the above we derive In exactly the same way , by considering a wave going from air to glass , we could prove that and .
Hence , in accordance with the statement on p. 517 .
The following table gives the values of obtained from the formula as increases from the critical angle:\mdash ; To determine and we have to consider a wave going from air to metal .
In the air we have ( incident ) ( reflected ) .
In the metal , ( refracted ) , where ; has the same value as before , and to satisfy the dynamical equations we must have .
the velocity of propagation in , metal .
Thus 1905 .
] On Newton 's Rings formed by Metallic Reflection .
There will be two distinct cases to discuss .
For metals of one class ( such as silver ) , so that is imaginary .
For metals of another class ( such as steel ) , and is real .
We shall first take the class to which silver belongs .
In this case and , say .
The boundary conditions give , and ' whence This leads at once to where and is greatest when , in which case , and is infinite .
These formulce lead to the following values of and In the case of a metal , like steel , for which is real , and we get where is a real determined by the relation ( so that is less than ) .
Putting the boundary conditions give and ' whence This gives and Prof. R. C. Maclaurin .
, where [ Aug. 15 , is greatest when , in which case and is infinite .
From these results we derive the following table:\mdash ; Steel .
In the formula on p. 517 we found In the present case we have where , as before , is the distance from the centre .
We thus have The numerator is the square of the resultant of two vectors 1 and QB2 inclined at an angle .
This resultant always be least when , and its least value will be zero if .
Hence , if we are seeking for places of absolute blackness we must take and .
The former condition , if it can be satisfied at all , will confine us to a particular angle of incidence ; the latter will the us to a ring of 1905 .
] On Newton 's Rings formed by Reflection .
definite radius .
It may be , however , that these conditions cannot be satisfied , but there may still be\mdash ; at different angles of incidence\mdash ; a dark ring where the intensity I is a minimum .
For a given of incidence and are fixed , so that to find the dark we have to differentiate I with respect to Q. In this way we learn that I is stationary when .
This makes ( see p. 517 ) , so that the fundamental series cease to converge ; but although we cannot thus go to the limit when , we can make as near this as we please .
The corresponding value of I is a if is negative , which is always the case since I cannot be greater than unity .
The minimum value of I is The accompanying figure 14 , drawn from the tables on pp. 535\mdash ; 536 , will show how varies in the neighbourhood of the polarising angle .
FIG. 14 .
It will be at once obvious , on looking at the figure , that the valiation of I is very much more rapid in the case of silver than of steel .
With silver the rise of from zero to 18 takes place almost entirely between the critical angle and , i.e. , within a space of only 13 ' .
As we have seen , there is absolute blackness when .
This occurs with silver when , and with steel when The radius of the dark ring is obtained from the formula which gives Prof. R. C. Maclaurin .
[ Aug. 15 , From this we get the following values for the radii of the rings:\mdash ; Silver .
Steel .
Fig. 15 embodies these results , and shows in a graphic way how the ring expands and contracts as the angle of incidence increases .
FIG. 15 .
1905 .
] On Newton 's Rings formed by Metallic Reflecb In order to show the variation clearly , it is necessary to represent on a different scale between the critical angle and from what is employed beyond that .
The crosses in the figure indicate the positions of the absolutely black ring .
It will be observed that with silver the contracts very rapidly after the black has been reached , and that the contraction is not nearly so marked in the case of stee ] .
In both cases the contraction after is very slow .
The colour effects will be similar to those discussed earlier ( cf. p. 524 ) .
For the black ring the value of obtained above is in the case of silver and with steel .
If , as before , we take these to correspond to the line in the spe , ctrum , then for the lines ( red ) and ( violet ) we shall have for silver , and for steel .
We have merely to plot the intensity for different colours , as is done in fig. 16 above , to see that the black ring is coloured red on the inside and violet on the outside .
The intensity of the violet on the outside is greater than that of the red on the inside , so that the outside colour is more marked , as Stokes pointed out .
The minimum value of I is .
Its value very near the centre is found by putting , and so getting The value of I at ) centre is given on p. 522 .
From the above formula for I we obtain the following results:\mdash ; VOL. LXXVI.\mdash ; A. Prof. R. C. Maclaurin .
[ Aug. 15 , Silver .
SteeL If we wish to compare the phenomena in the case of reflection from metals with that from glass , we have ( see p. 525 ) merely to put which gives and .
We thus get In this case there are no rings , but there is a dark spot at the centre .
The spot is absolutely black at the point of contact where .
The intensity of its blackness diminishes rapidly as we proceed outwards , as is illustrated in ligs .
17 , 18 , and 19 below .
In these figures the ordinates represent the blackness for different angles of incidence .
FIG. 17 .
Fig. 17 represents the state of affairs wheu , within 13 ' of the critical angle .
Here there is a distinct dark ring .
The ring is much more marked with silver than with steel , partly because the intensity of its 1905 .
] On Newton 's Rings formed by JIetallic Reflection .
blackness is twice that with steel , and partly because the darkness near the centre is much smaller with than with steel , so that the contrast between the ring and its centre is increased .
FIG. 18 .
FIG. 19 .
Fig. 18 mustrates the appearance when has increased to .
The dark with steel has now become very much more marked than formerly .
Its intensity has increased , while the blackness at its centre has diminished The ring with silver has contracted so much as to squeeze out its bright centre almost completely , so that it now presents the appearance of a dark spot at the centre , larger than with glass , but not so intense .
When has increased to ( fig. 19 ) the dark ring with steel has squeezed out the centre , and produced a dark central spot larger than with glass .
The central spot with silver has become so taint as to be scarcely appreciable .
A glance at tables on p. 540 will show that when is or greater the maximum intensity of is practically the same as that at the centre .
There will thus be no ring , but a central dark spot , so faint , however , as to be barely visible .
ht polarised paralld to the Plane of Incidjnce .
In this case it is convenient to introduce a new vector of which is the curl .
Then we have ; To determine and we must consider the passage of vibrations from glass to air .
542 Prof R. C. Maclaurin .
In the glass incident ) ( reflected ) .
In the air ( refracted ) , where has the same value as on p. 533 .
The boundary conditions give and whence where [ Aug. 15 , Similarly , and will be determined by considering the passage of vibrations from air to metal .
In the air ( reflected ) .
In the metal ( refracted ) , where has the same value as on p. 635 .
The boundary conditions give and whence In the case of a metal like silver is real and equal to , so that we get so that and Since , it follows that so thai the numerator and denominator in the expression for I are the same .
Thus everywhere , and there are no rings .
The only point where there is any darkness is where the glass and metal are in contact .
The intensity in this case is given on p. 530 .
There will thus be a small da\amp ; spot at the centre , much smaller than in the case of glass .
For metals , like steel , for which is not real , we have and ' whence and 1905 .
] On Newton 's Rings formed by Metallic Reflection .
Putting in the formula for I we get This is a minimum , i.e. , at the centre , and its value there is There will be absolute blackness at the centre when The following table gives the values of and the minimum intensity for different values of Steel .
With glass we have so that there is absolute blackness at the centre .
The following table enables us to compare the intensity with glass and steel at different distances from the centre:\mdash ; These results are represented graphically in , from which it appears that there is little difference between the phenomena with steel and glass .
On Newton 's Rings formed by Reflection .
FIG. 20 .
It would seem , then , that all the remarkable phenomena of Newton 's rings formed by metallic reflection , to which Stokes drew attention , are in accordance with the known principles of metallic reflection , without the introduction of effects of gradual transition between the media .

