]\gt ; The given by of Sulphur By W. H. , F.R.S. , Cavendish Professor of Physics in the University of Leeds .
( Received December 1 , 1913 , \mdash ; Read January 29 , 1914 .
) A number of results of examination of crystals by means of the spectrometer , which have been recently obtained in this atory , are discussed in a paper by W. L. .
The cases which he has consi-lered are those in which the information has been sufficient for a solution of the crystal structure .
This note deals with two cases , , sulphur and quartz , which have not been completely solved , but which have nevertheless iven i results .
In one form sulphur is thorhombic .
According to raphic measurements of the usual kind it has three unequ axes at to each other , whose lengths are in the ratio ) ) : 1 : .
It be assumed at once therefore that the element of structure is based on a parallelepiped of the form given in the figure .
The examination of the spectra iven by ) the assumption justified .
It is convenient to say a word at this as to the method of crystal planes for examination .
Some , of course , occur as natural faces ; seyeral of the results given below were obtained from a magnificent sulphur crystal very kindly lent by Prof. Pope .
Others it is necessary to cut , but operation is a very simple one .
It is not in the least necessary that the surface should be true or smooth ; for the -rays are eflected by body of the crystal and not by the surface .
It is sufficient to rub down an FIG. 1 .
ordinary sulphur crystal on a piece of sandpaper , judging the result by eye .
There would be no need to prepare a surface at all were it not that it is then more convenient to place the crystal correctly in its holder .
Some crystals such as } ) inels , diamonds and rubies cannot of course be cut , and a little calculation is required 'Roy .
Soc. Proc p. 468 supra .
Prof. W. H. Brag .
The -ray if they are to be correctly mounted ; but if the mounting is difficult , and at first incorrect , the spectrometer readily gives indication as to how it is to be amended .
The -ray bulb which was used in the examination of the sulphul had a palladium anticathode .
The principal palladium line cn was reflected by rock salt ( 100 ) at the following angles : the references are to fig. 1 .
Plane 100 010 001 011 101 110 111 OCDB OAEC OAFB AODG BOEG OCGIl ABC Glancing ?
It is at once clear that there is a connection between these ures and the crystallographic ratios .
For example , the lancing a for the planes ( 100 ) and ( 010 ) are and .
These are in the ratio 1 : ; and the spacings of these planes , or rather sets of planes , are therefore in the inverse ratio , or .
This is very close to the ratio .
It is to be remembered that the connection between the glancing angle , the spacing and the wave length is given by ; when is small , the sine is of course nearly proportional to the itself .
Again the perpendicular from on the plane ABC is where OA , OB , OC The value of this expression is easily found to be .
Now , the glancing angles of the planes ( 010 ) and ( 111 ) are and , and these are in the inverse ratio of 1 : , which is very nearly twice the ratio just found .
In the same way the perpendiculars from the origin on the planes and BCEF are and ; or and .
These are in the ratio 1 : .
The of reflection by the corresponding planes are and .
These are in the inverse ratio of 1 : Let us proceed , therefore , to examine the results of assuming the various forms of space lattice which may be associated with this figure ( fig. 1 ) .
The choice is very limited , and the correct form is readily found to be that in which an atom*of sulphur is placed at each of the figure , and at the centres of the two faces perpendicular to the longer axis .
Everything ( as will presently appear ) then falls into place at once , with the exception of the It is here assumed that there is one diffracting centre to each atom .
This assumption seems to be very clearly in accord with present results .
given by of Sulphur Quartz .
angle of reflection by the ( 001 ) plane , an exception which is readily explained .
Let us assume this form of lattice and proceed to calculate its dimensions .
Two atoms are associated with each element of the lattice shown in fig. 1 .
Each atom at a corner is at the place of volume elements , and .
of it may be said to belong to each element .
Hence the whole of these count as one .
Each of the two atoms at the centres of faces belongs equally to two elements .
a face in common , and half its value is to be assigned to each nent .
The two therefore count as one .
up , there are two atoms to each unit of volume .
We have now to find how many .
lattices there are .
It may be convenient at to explain that the spectra iven by a set of lattices are the same as those given by a lattice if we consider only the positions of the first-order spectra in each , unless it happens.that the planes of one lattice exactly interleave those of the other .
To make this clear , let ns consider the corresponding case of the diffraction grating .
A set of rulings , , diffract light of a length a given .
If an interpenetrating similar set of lines , , is ruled on the same ) parallel to the first , the first-order occurs in the place as before , though with altered intensity , unless the second set of lines happens to fall exactly half way between the first , in which case it disal ) and the second spectrum becomes the first .
This pl.inciple is easily extended to the spac , lattice .
Generally , intersimilarly placed lattices a first-order spectrum in the same place as any one of the lattices would do alone , unless they are so that they can bc derived from one anothel ) movemenls .
equal successive translations perpendicular to any set of planes , in which case all the spectra given by that set of ) lanes will disappear except the nth , 2 th , and so on .
If the translations referred to are not equal , the intensities of the spectra of the various orders given by that set of planes will not be the same , eithcr absolutely or relatively to each other , as those given by one lattice ; but the position of the lirst-order spectrum will be the sa1ue as if there were only one lattice .
Let us assume that there are lattices .
every unit of volume ab there are , therefore , atoms ; for each lattice contributes two .
Taking the weight of the atom as , the volume contains a mass The density of sulphur is 2.07 , and Hence .
( i ) Prof. W. H. Brag .
The -ray Spectra The glancing angle of the first-order spectrum of the ( 010 ) planes is and the spacing of these planes is for the form of lattice we have assumed .
Thus .
( ii ) Comparing ( i ) with ( ii ) it is clear that must be equal to 8 .
this value in ( i ) we cm .
This gives the actual dimensions of the lattice of fig. 1 , and it is now easy to find the relative spacings of the various sets of reflecting planes , and therefore the glancing .
at which the first-order spectra should appear .
The next table compares the values so calculated with those actually observed .
Plane 100 010 001 011 101 110 111 Except in the case of the ( 001 ) plane , the agreement is quite satisfactory .
The exception is very interesting ; the angle of reflection of the first spectrum is eight times what it should be were there only one lattice .
The meaning of this is clearly that the eight lattices are to be derived from each other by movements involving equal translations parallel to the axis ; this causes all the spectra to disappear up to the eighth , in the manner explained above .
The spectra iven by sulphur are all weak , particularly in the case of the ( 101 ) plane , and in no case have any been found so far of higher order than the first .
Consequently there are no available data at present for proceeding further with the calculation of the relative positions of the eight lattices .
The analysis of the quartz spectra may be carried out on much the same lines as those of the preceding ument , and a briefer discussion will be sufficient .
The results show , as would be anticipated , that it is necessary to use a hexagonal lattice , as in the figure , where Ii , IF , and according to crystallographic data .
given by of The palladium is reflected by a prism at a lancing angle Hence We may take , as the arallcl i ) an { llonl at each corner .
Its volume and since the density of qnart is , the mass associated with it If is the number of molecules associated with the } } ) , the mass must also be given by or the atomic weight of Si and are respectively , th of the atom is .
The value is therefol etllat is to there are three space lattices of the assumed form ] , of which ) one molecule is associated .
The of the various sets of plaIlesn ) now , and 580 -ray given by Crystals of Sulphur and Quartz .
thence the angles at which the palladium line should be reflected .
These are compared with the observed values in the following table:\mdash ; Plane EDde Id FDdf i 'd BGed AHed BDF Observed Calculated 495 It is curious that so far no reflection has been found from the plane BDF , the plane which is perpendicular to the principal axis .
, 1914.\mdash ; Since this was written a weak reflection from this face has been found .
The value of the glancing angle shows that the spacing of the planes parallel to BDF is one-third of Ii .
This means that the three lattices can be derived from each other successively by move- ments involving equal translations parallel to the axis .
The result was to be expected .
] Many of the planes give very good reflection , and spectra of more than one order , so that it will no doubt be possible to complete the analysis of the quartz structure eventually .
So far , it is clear that there are three interlattices of the iven form for every atom in the molecule .
The three must no doubt be arranged as ested by Groth , so that they can be derived from each other successively by a translation equal to along the axis accompanied by a rotation of about a line parallel to that axis .
This constitutes a screw motion , which may be either right-or lefthanded .
It remains to complete the determination of the positions of the silicon and the oxygen lattices with reference to the line of rotation .

