The Effect of Pressure upon Arc Spectra.\#151 ; Titanium .
By B. Bossi , M.Sc .
( From the Physical Laboratories , the University of Manchester : communicated by Arthur Schuster , F.B.S. Beceived January 5 , \#151 ; Bead February 10 , 1910 .
) The Apparatus .
The apparatus used has already been fully described in a previous paper by Dr. W. G. Duffield.* It consists of a high-pressure cylinder in which the arc burns and the 21^-foot concave grating spectrograph of this laboratory .
The pressure in the cylinder was obtained by admitting air into it from a gas holder .
The second order spectrum was employed , giving a dispersion of 1*3 Angstrom units per millimetre on the photographic plate .
The arc was struck between two cylindrical poles : the top one , negative , being an ordinary carbon for arc lamps ; the bottom one , positive , being a graphite tube ( 1*5 cm .
outside diameter , 1 cm .
inside diameter ) filled with powdered titanium carbide .
The main impurity lines were due to calcium , iron , vanadium , and chromium . .
Behaviour of the Arc under High Pressures .
The arc under pressure between carbon poles containing the titanium carbide burns far better than an arc between metallic poles .
It is more steady and does not wander about the molten surfaces of the poles , consequently it is fairly easy to keep its image fixed on the slit of the spectrograph .
The titanium arc under pressure , however , is still quite intermittent , the exposure consisting of a series of flashes of short duration .
The brightness of the arc is greatly increased by pressure , especially at 50 and 100 atmospheres , with the result that , notwithstanding the broadening of the lines , the same exposure ( about 20 minutes ) was suitable for the different pressures used .
At atmospheric pressure the titanium arc is of a whitish colour , but it turns quite blue under high pressure ; a similar phenomenon is exhibited by the copper arc.f * W. G. Duffield , ' Phil. Trans. , ' A , vol. 208 , p. 111 .
t W. G. Duffield , ' Phil. Trans. , ' A , vol. 209 , p. 205 .
The Effect of Pressure upon Arc Spectra.\#151 ; Titanium .
415 The Photographs and the Measurements .
The comparison spectrum , at atmospheric pressure , was taken in the middle strip of the plate , with the one under pressure above and below it .
The comparison spectrum was photographed both before and after the one under pressure to ensure that no accidental displacement took place during the exposure .
Several photographs were taken at 15 , 30 , 50 , and 100 atmospheres , and the best two plates for each of the above pressures were chosen and measured .
Difficulties were encountered in getting good photographs at 100 atmospheres , as , owing to the shortness of the arc at that pressure , the images of the incandescent carbon poles could not be easily kept away from the slit , with the result that the continuous spectrum was very prominent on the background .
Imperial flashlight plates and Imperial pyro-metol standard developer were used .
The pressure displacements of the lines were measured with Kayser 's machine , the settings being made on the most intense portion of the lines .
Twelve settings were made upon each line on each plate , six with the plate in one position and six with the plate in the reversed position .
Since two plates were measured at each pressure , altogether 24 readings were made on each line at each pressure .
Broadening of the Lines .
All lines examined were found to be broadened by pressure , the broadening increasing with pressure .
The amount and type of broadening is different for different lines .
A few lines broaden out almost symmetrically ( as a rule those which broaden out least ) , while those which broaden out most give , as a rule , an asymmetrical broadening , greater towards the red end of the spectrum .
Reversal of the Lines .
Owing to the uncertain and varying amount of titanium present in the carbon pole , it was found impossible to note at what pressure the reversal of the lines took place .
On one plate , for instance , several lines were found to be reversed at 30 atmospheres , and the same lines were found to be still emission lines on other plates at higher pressures .
It could not be ascer- .
tained whether this was due to a variation in the amount of titanium present in the arc or to the fact found by Dr. W. G. Duffield , * in the case of iron , that above 25 atmospheres the reversals decrease in number and * W. G. Duffield , 4 Phil. Trans./ A , vol. 208 , p. 111 .
Mr. R. Rossi .
[ Jan. 5 , genei'ally in width .
A mere note was therefore made of the lines which have a tendency to reverse under pressure .
They are marked R , and r in .
Table I. Atmospheres .
15 .
30 .
50 .
100 .
Tendency to reverse .
Line .
Wave-length .
Displacements in thousandths of an A.TT .
1 4024 -71 42 86 148 281 2 55-18 35 98 \#151 ; 310 3 60 -42 37 79 ( 101 ) 207 4 78 -61 44 79 143 173 5 82-57 38 96 129 219 6 4112 -86 41 73 162 \#151 ; 7 23 -68 57 85 188 \#151 ; 8 37 -39 ( 31 ) 55 \#151 ; \#151 ; 9 51 11 43 \#151 ; \#151 ; \#151 ; 10 59 -79 78 ( 53 ) ( 86 ) \#151 ; 11 86 -27 57 ( 148 ) ( 213 ) ( 315 ) 12 4203 -58 42 64 ( 114 ) 13 38 -00 47 72 103 182 14 63 -28 52 88 159 291 15 74 -73 42 77 159 259 R , s 16 82 -85 45 93 159 303 r , s 17 86 -15 87 93 * 166 ( 299 ) 18 87 -55 42 81 158 ( 282 ) 19 4306 -07 47 91 160 ( 307 ) R , s 20 14 -95 49 86 185 287 r , s 21 21 -82 ( 116 ) ( 157 ) ( 264 ) \#151 ; 22 44 -47 37 85 129 \#151 ; 23 60 -60 56 ( 83 ) 170 ( 354 ) 24 4423 -00 45 85 141 324 25 27 -28 52 101 178 308 r , a 26 34 -15 64 108 174 317 27 40 -49 61 96 143 270 28 43 -97 59 75 147 \#151 ; 29 49 *32 67 107 166 297 r , s 30 51 -07 75 107 170 323 r , s 31 55 *48 79 160 234 438 r , a 32 57 *59 81 160 274 497 r , a 33 65 -96 68 105 166 305 34 68 -65 68 115 177 320 35 71 -40 62 101 147 333 36 75 00 96 ( 147 ) 207 \#151 ; 37 79 -86 68 99 143 \#151 ; 38 81 -41 68 103 189 322 r , s 39 82 -84 84 \#151 ; \#151 ; \#151 ; 40 89 -24 53 107 162 300 r , s 41 96 -33 57 88 162 275 42 4501 -43 68 96 197 470 r , a 43 12 -88 75 150 239 468 r , a 44 18 -18 65 137 27 2 464 r , a 45 22 -97 75 150 264 464 r , a 46 27 *48 76 157 262 457 r , a 47 4544 *83 76 153 275 493 r , a 48 48 -93 75 146 259 456 r , a 49 52 -62 79 165 252 468 r , a 50 55 -64 74 150 274 548 r , a 51 72 15 65 103 229 416 52 94 -28 39 78 137 1910 .
] Effect of Pressure upon Arc Spectra.\#151 ; Titanium .
417 the last column of Table I , according as they have a large or small tendency to reverse .
The letters s or a indicate a symmetrical or an asymmetrical reversal .
The Displacement of the Lines .
Table I gives the displacement at each pressure in thousandths of an Angstrom unit .
The first column contains a list of arbitrary numbers assigned to the different lines to facilitate reference to them .
The second 'Column gives the wave-lengths of the lines taken from Hasselberg 's tables .
\#187 ; The pressures are the excess above 1 atmosphere .
The measurements in brackets denote doubtful readings .
The displacements of the unreversed lines vary greatly from one line to another , and no grouping of the lines according to equal displacement seems possible .
In the case of the reversed lines , however , with the exception of line 42 , it seems possible to classify the lines in two groups : the mean displacements of the two groups being very nearly in the ratio 3 :5 .
The measurements of the displacements of line 42 were repeated , but were found to be practically the same as the former measurements .
The mean displacements of all the titanium lines studied at each pressure are given in Table II , from which we get as mean displacement o per atmosphere of all lines examined the value 0*00365 Angstrom unit .
From Table II it is seen that , within the accuracy of experiment , the displacement varies directly with the pressure , as first found by Humphreys and confirmed later by Duffield .
Table II .
Pressure in atmospheres ( excess above 1 atmosphere ) 15 30 50 100 Mean displacement in A.u. 0*060 0*105 0 *182 0*344 Mean displacement per atmosphere in A.U. 0 *0040 0 *0035 0 *0036 0 *0034 Only five of the lines already examined under pressure by Humphreys* were studied in this work .
As in the case of copper , f their displacement is found to be larger than that found by Humphreys .
The results are compared in Table III .
* W. J. Humphreys , \#163 ; Astrophys .
Journ./ vol. 26 , p. 18 .
t W. G. Duffield , 4 Phil. Trans./ A , vol. 209 , p. 205 .
Mr. R. Rossi .
[ Jan. 5 , Table III .
Line .
\#165 ; Mean displacement per atmosphere in thousandths of an A.U. Previous measurements .
Present measurements .
4286 -15 2 45 3*13 4287 '55 2*07 2- 77 4306 -07 2 *32 3 -10 4427 '28 0*53 3-40 4544 -83 1 -90 5*20 Pressure Displacement and Zeeman Effect .
It has been suggested by Humphreys that lines which show a large Zeeman effect ought also to show a large pressure displacement .
Table IV .
* Mean pressure displacement per atmosphere in thousandths of an A.U. Zeeman effect .
Line .
Type of separation in magnetic field ( spark observed at right angles to lines of force ) .
dx/ A2 , spark observed at right angles to lines of force .
d\l\ % spark observed parallel to lines of force .
Observer .
4024 -71 2 *85 Triple ?
,0 , -2-01 Purvis 4078 -61 2-80 a + 2*19 , 0 , \#151 ; 2 *23 \#187 ; 4286 *15 3 13 Double + 1*32 , 0 , -1 *32 jj 4306 *07 3*10 Triple + 1 *64 , 0 , -1 *66 a 4314 *95 3-03 Quadruple \#177 ; 2*14 , \#177 ; 0*76 if 4321 *82 5 *25 ( ?
) Triple + 2*46 , 0 , -2*47 11 4427 *28 4440 *49 3*40 2*93 a Quadruple + 1-38 , 0 , -1 *36 \#177 ; 2*35 \#177 ; 2*95 King 4443 *97 2*70 Triple + 1-38 , 0 , -1 *37 Purvis 4449 *32 3 *30 ii + 1 *71 , 0 , -1 *73 a 4455 *48 4-92 if + 1-68 , 0 , -1 -66 )y 4457 -59 5 *30 ii + 181 , 0 , -1 83 a 4465 *96 3 27 ii + 2-14 , 0 , -2T5 it 4468 *65 4471 -40 3 *50 3*20 if Septuple + 1 *53 , 0 , -1 *53 Inner pair 1 *50 King 4481*41 3 *40 Triple + 2*24 , 0 , -2-25 Outer pair 4 T3 Purvis 4489 *24 3*32 Quadruple \#177 ; 2*13 \#177 ; 5 22 King 4496 *33 3*15 Triple + 2*37 , 0 , -2*37 Purvis 4501*43 4*07 }\gt ; + 1-21 , 0 , -1 T9 a 4512 *88 4-87 a + 2-01 , 0 , -1*99 a 4518 *18 4-72 + 2*20 , 0 , \#151 ; 2 18 a 4522 -97 4-97 j ?
+ 2T5 , 0 , \#151 ; 2 T6 if 4527 *48 5*10 Septuple + 1*93 , 0 , -1 *90 Inner pair 3 *62 King 4544 *83 5*20 p + 1-92 , 0 , -1 *96 Outer pair 7'96 7 '69 ( mean for if 4548 -93 4*92 Triple + ?
, 0 , -2-26 two pairs ) Purvis 4552 *62 5 12 jj + 2T4 , 0 , \#151 ; 2 T5 a 4555 *64 4*97 + 2'12 , 0 , \#151 ; 2 *12 a 4572 15 4*40 a + 1-28 , 0 , -1 *29 a 1910 .
] Effect of Pressure upon Arc Spectra.\#151 ; Titanium .
419 Table IY gives the comparison between pressure displacement and Zeeman effect for some lines which were studied both in the magnetic field and under pressure ; but their limited number , and the limited accuracy of the pressure displacement readings do not enable one to draw any relation between the two effects .
The separations in the magnetic field were taken from the valuable papers of J. E. Purvis* and A. S. King !
on the subject .
Dr. King 's readings , which were made in a field of 12,500 units , were reduced to a field of 39,980 units in order to make them comparable with Mr. .
Purvis ' readings .
Changes in Relative Intensity under Pressure .
Some lines maintain at different pressures the same relative intensity that they possess at 1 atmosphere ; but most of them undergo a change , and are either enhanced or weakened by pressure .
Table Y gives those lines which are weakened or strengthened by pressure .
The last column of Table Y gives the relative intensities of the same lines in the arc and furnace spectrum as found by A. S. King .
J It may be noted that with one exception all lines which are weakened in the furnace spectrum are strengthened in the arc under pressure .
Table V. Line .
Behaviour under pressure .
Intensity in Line .
Behaviour under pressure .
Intensity in Arc. Furnace .
Arc. Furnace .
4024 -71 S 4468 *65 W _ 4186 -27 S \#151 ; \#151 ; 71 *40 s 22 15 4263 -28 s 17 9 75 -00 w 8 12 74 -73 s 34 18 81 *41 s 35 12 4306 -07 s 38 21 82 *84 w 8 9 14 -95 s 36 18 89 '24 s 19 9 21 -82 w 8 6 96 *33 s 22 18 44 -47 w \#151 ; \#151 ; 4501*43 w \#151 ; \#151 ; 60 -60 s \#151 ; \#151 ; 12 *88 s 36 15 4427 '28 s 38 18 18 T8 s 38 15 34 T5 ' w 14 15 22 -97 s 40 15 43 -97 s 38 12 27 -48 s 36 18 49-32 s 36 15 44 *83 s 38 15 51 -07 s 30 12 48 *93 s 38 15 55 -48 s 36 15 52 *62 s 42 15 57 -59 s 40 21 55 *64 s 38 15 65 -96 s 20 12 72 *15 s \#151 ; \#151 ; S = Strengthened .
W \#151 ; Weakened .
* J. E. Purvis , ' Proc. Camb .
Phil. Soc. , ' vol. 14 , p. 41 .
t A. S. King , 'Astrophys .
Journ. , ' vol. 29 , p. 76 .
J A. S. King , ' Astrophys .
Journ. , ' vol. 28 , p. 300 .
420 The Effect of Pressure upon Arc Spectra.\#151 ; Titanium .
Six of the lines given by Lockyer , Reese , and others as enhanced on passing from the arc to spark spectrum were studied in this work , and it was found that some of them are enhanced , while others are weakened under pressure { see Table VI ) .
Table VI .
Line .
Behaviour under pressure .
Intensity* in Spark .
Arc. 4055 -18 Strengthened ( at high pressures ) 2 1 4344 -47 Weakened 3 1 4443 -97 Strengthened 9 4 4468 '65 Weakened 9 4 4501 -43 a 8 4 4572 -15 Strengthened 9 4 * Tables of enhanced lines ( Publications of Solar Physics Committee , 1906 ) .
My best thanks are due to Prof. Eutherford for putting the necessary apparatus at my disposal and for the interest he has taken in this work ; also to Prof. Schuster for his valuable suggestions .
I also wish to express my indebtedness to Dr. W. G- .
Duffield for the experience I gained under him in high-pressure work , and for suggesting to me this research .

