312 LEWIS A NEW SECTOR SPECTROPHOTOMETER. XXXII. -A New Sector Spectrophotometer. By SAMUEL JUDD LEWIS. IN a paper by the author on “The ultra-violet absorption spectra of blood sera” (Proc. Roy. Soc. 1917 [B] 89 327) it was stated that the work described had been done with two sector spectro-photometers that these were not adequate to the exacting nature of the investigation although they were the best and most modern instruments available and t.hat a new photometer to the author’s design was under construction. That apparatus has now been in use for several months and the purpose of this paper is to describe it. A short account of the method of using such instruments is given in the reference mentioned above. As was 5he case with the work detailed in the paper cited the present development of the sector spectrophotometer has been generously supported by the Beit Research Fund Committee the trl-ustees of a fund which has been placed a t the dispwal of the British Homeopathic Association by Mr.Otto Beit for purposes of scientific research. The new instrument was designed by the author in 1915 with a view to meeting the requirementa of the work on blood serum. No instrument can be unnecessarily refined for this and should it be desired to employ the method of ultra-violet absorption spectre graphy for clinical purposes whether for blood serum or for any other subst.ances an instrument which is a t once trustworthy and easy to manipulate is essential. Incidentally an instrument which fulfils these conditions should satisfy most of the demands of scien-tific research and bring the practice of ultra-violet absorption spectroscopy within the range of applied chemistry.Unless or until the ideals set out are attained absorption spectroscopy can have little motre than academic significance ; but perfect easily adjustable spectrophotometers may be expected in course of time to occupy a place in the general laboratory not less important than that filled by the polarimeter or the refractmeter. Among the objects aimed a t in the new design were the following : ( :) The utmost accuracy and refinement in the resulting spectra, because for the serum work i t is required to discover with certainty very m a l l differences of detail in the absorption curve as explained in the paper cited.( b ) To work quickly as well as accurately since it is necessary to undertake the examination of a serum a t very short notice and there is no reasonable opportunity of revising the observation a D. ^ . a ; [ 7'0,jUce Imyc 313 LEWIS A NEW SECTOR SPECTROPHOTOMETER. 313 the serum will usually have changed in character by the time the absorption curve is drawn; this presumes quick and perfect adjust-ment of the instruments. ( c ) To maintain this excellent adjustment throughout an experi-ment or to restore it from time to time during the progress of an observation without the necessity of other disturbances. ( d ) Precise quantitative values for the extinction coefficients, since the amplitude of the curve has important significance. ( e ) To attain this high standard and yet to be able to use the one spectrograph alternately with the photometer and for other purposes as circumstances may require.The photograph exhibits the general appearance of the new photometer; it exposes the vanes in the upper sector the front of the sector box having been removed; aIso only one platform bear-ing a tube carrier is shown. The arrangement of parts is displayed in the diagram. Q represents a suitable lamp light from which falls on the two lenses L, L, which fender the light parallel. These pencils of light pass through the face aA,b in the reflecting prism P and through the face a& in the reflecting prism Pz in the manner shown and are reflected approximately a t right angles by the inclined faces alcl in p r i m P and a2c2 in the prism P2 so that in each case the light follows a course parallel to the main axis of the prism to a second inclined face b,d in prism P and b 2 4 in prism P, where i t is reflected again a t right angles.The parallel beam reflected from b,d passes through the sector D, the diaphragm 0, and the lens L to the inclined face elfl in reflecting prism Y, where it is reflected a t right angles along the main axis of the prism on to the inclined face g,h, where i t is again reflected a t right angles and passes out of the prism on to the face hi of the rhomb R placed in front of the slit S of the spectrograph or in the absence of the rhomb directly on to the slit. The pencil from b,d undergoes similar treatment and eventu-ally passes out of the photometer on to the face kj of the rhomb.The rhomb may be dispensed with if the pencils of light are directed so as to fill the angles of the prisms a t 9 and g., provided that one prism slightly overlaps the other as shown in the figure, so that' the prominent edge may define the line of juxtaposition. It will be seen that any &ray parts of the pencils of light would be refleched or escape out of the field. The essential part between the two prisms in either path is the sector B, D, capable of cutting off any desired portion of the light passing along that path 314 LEWIS A NEW SEUTOR SPEUTROPHOTOMI?ITER. It should be observed that the prisms and sectors are enclosed in dustrproof metal boxes provided with quartz windows. The sector systelm is placed in a part of the path where the light is parallel.It consists of four vanes Vl V2 V3 V4 as shown in the side elevations in the figure. Each vane has t3wo edges a t right angles and when the four vanes are disposed to one another in one plane so that the four angles meet a t a point the system is closed and no light can pass. This arrangement is repre-sented in the upper path in the figure. The common point of the four angles is on the optic axis to which the plane is a t right angles. Each vane can be turned by means of suitable mechanism about its bisectors m, m2 m3 m4. When all the vanes are turned simultaneously through an angle of 90° about their respective bisect'ors lightl can pass in the direction of the optic axis without any interruption except -that caused by the slight obst.ruction due 60 the thickness of the material of the vanes as shown in the figure for the lower path.By turning t h e vanes through any other given angle about their bisectors a known proportion of the light may be allowed to pass. Each of the four vanes is carried on a wheel by means of a spindle which coincides with the bisector of the vane and also forms the axle of the wheel. These wheels are mounted on t(he outside of the walls of the box enclosing the sector a t right angles t o one another and fit into one another by bevelled cogs. They move simultaneously and the fitting is so close that back-lash is reduced to an insignificant minimum. The amount of rotation of the vanes is measured by a pointer I , mounted on the front wheel and moved against an arc graduated in half degrees from 0 to 90.A diaphragm or stop O, 0, is placed in front of each sector to reduce the section of the beam of light to suitable dimensions, say 9 mm. in diameter when the observation tubes or cells have a lumen of 12 mm. I n order to correct the error causeld by the obstruction due t o the thickness of the material of the vanes their edges are reduced to knife edges and the thickiiess along the bisector is graduated from the minimum a t its extremity at the optic axis to what is necessary say 1 mm. a t a distance of 10 nun. The two sides of each vane are equally made and each of the four surfaces is plane or concave. Hence a sectdon through a vane a t right angles to its bisector has the form of a rhombus having two very obtuse and two very acute angles; also a section which includes the bisector has the form of an isosceles triangle.It follows that when the sector system is open as shown in the lower path in the figure the obstruction or shadow produced b LEWIS A NEW SECTOR SPECTROPHOTOMETER. 316 each vane has the form of a geometrical sector of known dirnensions, and therefore the sectional area of the pencil of light passing through the sector system is reduced by four times the area of one small sector shadow ; also each of the four apertures has the form of a geometrical sector. Compensation for this reduction in area is made by employing for the path in which thel light is to be of whole or 100 per cent. intensity khat is for the path in which the substance under observa-tion is placed a diaphragm the aperture of which is appropriately larger than that of the diaphragm in the other path.In the author’s instrument the diameter of the larger aperture is 9.3 mm and that v€ the smaller 9.0 mm. These diaphragms are loose and may be used with either sector according to whether it is desired to use the upper or lower sector. The only limitation to this arrangement is that the sectors do not operate for those angles which are included by the small angle covered by the thickness of the vanes. In the author’s instrument this is 2*6O so that the range of operation is from 2 ’ 6 O to 90’. The lenses L, L, placed between the sectors and the second set of prisms focus the light on the slit the focus being that for light of a wavelength of about 3000.The edge of the rhomb is placed close to the slit so that the two pencils of light when they emerge from either side of it into the slit may be a t the focus of the collimating lens; the line of juxtaposition between the two spectra is thus v0ry clearly defined without the least overlapping or intervening space with the result that the lines in the two spectra whether visual or photographic, can be compared for their relative intensities a t the best advantage. The cell containing the substance under observation is placed in the parallel beam of light between the sector and the central lens; roam is provided for tubes up to 100 mm. in length and tubes of this length were used successfully in an investigation relating to Lambert’s law.As the pencil of light is 9.3 mm. in diameter it is desirable for long tubes that the lumen of the tube should be a t least 12 mm. in diameter so as to avoid serious reflec-tion from ths inner walls of the tube. When the lumen is small and the layer of substance thin it may be placed in the convergent beam immediately in front of the prism P3 as shown a t w and a still smaller cell might be interposed between th4 prism P and the face ki of the rhomb but only with due regard to existing con-ditions. Inwmucb as the system in the lower path of light is identical with that in the upper the cell may be placed.in the lower with equal advantage or as explained later cells may be placed in both beams simultaneously 316 LEWIS A NEW SECTOR SPECTROPHOTOMETER.The new instrument is characterised chiefly by its sectors; these are distinguished both in their construction and in the principles upon which they operate. I n comparison and contrast with &her sectors they accommodate the whole beam of light and not merely half as is the case with some others; this enables exposures to be reduced to one-half in order to produce a photograph of the same intensity ; the exposures being continuous calibration of the photo-graphic plate is avoided; the direct function of the new sectors is to cut down the intensity of the light and not to do so indirectly by altering the exposure as is the case with the Hilger instrument; the sector is sbill which is a great convenience. The whole aper-ture of the sector system is made up of four sector-shaped apertures arranged diagonally about the optic axis that is they are dis-posed a t an angle of 4 5 O to the vertical.This has the advantage of utilising all parts of the circular beam of light in proper pro-portion ‘whatever the magnitude of the effective aperture a t the moment. I n this it differs from the iris and other forms of stop or sector which reduce the int’ensity of the light by cutting down the light from the periphery of the beam inwards or by cutting it down in some other way which assumes that every part of the field is equally illuminated. Under the oonditions of ordinary practice this assumption lacks sufficient justification where precise photometry is required. The size of the sector aperture is measured in terms of the angle which it forms a t the optic axis.The angle 4 5 O represents an aperture of 100 per cent. and the angle formed at the optic axis by the shadow cast by one half of a vane in any given position is 4. Hence the sector aperture is proportional to 45”-4. Thjs method of measuring the size of t.he aperture is exceedingly con-venient both for simplicity of calculation from the angle 0 which is read on the graduated arc and for the fact that apertures of any odd value may be produced a t will. This will become evident from the following explanation of the manner of calculation. When the vane acb in the figure ( a ) (front elevation) is turned through an angle 8 as measured by the graduated arc so as to take up the position ecf each half of the vane ec or cf creates a sector-shaped shadow in a beam of light) in the direction of the arrow ; e ’ d is a projection of ed.It is shown again in the figure ( B ) in side elevation where e’d’ is the projection of ed. If the beam of light were rectangular in section the shadow would be triangular in shape (e’d’o) and proportional in area to sin@ since ed which subtends the angle 8 is equal to e,dl which subtends the angle 4 a t the optic axis 0 for edl and d’o are equal since they form the right angle in the isosceles triangle ed’o; therefore ed/edl L’IPWIS A NEW SECTOR SPECTROPHOTOMETER. 317 e’d’/d’o that is sin 8= tan 4 whence 8 being k n m 9 may be found directly from the tables. The beam of light however is circular in section; hence the area of any sector in the circle is proportional to the angle which it cont,ains that is to + for the shadow and 4 5 O - 4 for the sector aperture.Only oneeighth of the whole circular aperture has been con-sidered but itl will be seen on cancelling common factors that the whole aperture a t any moment is measured by 4 5 O - + where 4 5 O is taken to represent the fully open sector. Itl is convenient to express the apertures as percentages of the whole aperture that is, (45 - 4)100. The percentage apertures and values for logI,/Z’ as 45 corresponding with each half-degree on the graduated arc have been calculated by Miss Gartha Thompson t’o whom’the author is indebted for valuable assistance throughout the inquiry. Examples are given in the following table: 8. 10 25 40 40-5 41 55 70 80 80.5 81 85 85.6 88 Sin 8 or tan 9.0.1 736482 0.4226183 0.6427876 0.6494480 0.6560590 0-8191520 0.9396926 0.9848078 0-98@2856 0.9876883 0.9961947 0.9969173 0.9993908 9. 9.85 108 22.90981 32.73241 33.00163 33.27239 39.32269 43.2191 8 44.56143 44,60441 44.64512 44.89078 44.9 1 155 44.9 82 55 45 - @. 35.14092 22.0901 9 12.26759 11-99837 11.72760 5-67731 1.78082 0.43857 0.39559 0.35488 0.10922 0.08846 0.01745 Sector aperture, Der cent. (46-$4 100 45 78-11 49.08 27.27 26.67 26.07 12.62 3-958 0-975 0.870 0.789 0.243 0.197 0.0888 Log I / I ’ , 45-t$ log L* 0.1073 0.3091 0.5644 0.5740 0.5839 0.8991 1.4026 2.01 11 3.0559 2.1031 2.6150 2.7065 3.41 14 It is evident that the values in the table can be applied to graduating the arc on the instrument so that it may read directly in terms of logZ/Z/ as has been done with other instruments.This is convenient where it is intended to use the instrument for routine work only but for versatile research and especially where it is desired occasionally to elaborate a particular part of an absorption curve the freedom conferred by the ordinary scale and tables will be appreciated. There is however no difficulty in providing both scales on the same arc. The effect produced by the size of the shadow cast by the vane being a function of the sine of the angle through which the vane has been turned is significant. From the above figures it will b 3 18 LEWS A NEW SECTOR SPECTROPHOTOMETEB.seen that the difference in the sector aperture produced by a constant difference of 1 5 O in 8 decreases continuously thus : Per cent. Between PO" and 25' the aperture decreases by (78.11-49.08) 29-93 Y Y 25" Y 40" ? Y 9 ) (49.08-27-27) 21.81 Y 55" Y 70" ? 7 Y Y ? 9 , (12.62- 3.96) 8.66 , 700 ? 850 y 7 , ( 3.96- 0-24) 3.72 Y Y 9 , Again for half -degree int'ervals in 8 : Between 20" and 20.5O the aperture decreases by (58.03-67.09) 0-94 ,¶ 40" , 40.5" ? , ,) y y (27.27-26.67) 0.60 80" ? 80*5" y y ,? , , (0.975-0.879) 0.096 , 85" ) 85.5" ) , ) , (0.243-0.197) 0.047 For any of the apertm-es of small size i t is evident therefore, that a nioveinent of the pointer on the arc over one or more divisions alters the size of the aperture by only a small fraction so that any probable error in construction or manipulation can have no detectable effect on the measurement of the size of the aperture.The possibility of regulating the size of the m a l l apertures with such precision is a valuable advantage since it is with these that much of the more important and delicate work is done. I n an investigation of uric acid undertaken to see i f it obeyed Lambert's law it. was proved thatl the results for all the small apertures down to 0.29 per cent. (reading 84-5O on the graduated arc and giving the value 2.53 for logI/Z') harmonised perfectly with those found for the larger apertures. The still smaller ones were not" quite so true owing t o a slight imperfection in the setting of the vanes but this can be avoided in reproducing the instru-ment.That it is not only so but that i t is also precise follows from the examination of a standard piece of glass which was supplied by Messrs. Adam Illilger with their photometer for calibration purposes. The figures obtained are tabulated below. The two sets of values for logI/I' are as follows ( n ) those obtained by the new photometer; ( t ) ) the figures given by Hilger : The instrument is absolute in all its measurements. Wave-leng th. (a). ( b ) . 2751 0.281 0.278 2636 0.665 0.810 2564 0.919 0.888 2478 1.318 1.330 2435 1.605 1-608 2389 1.970 1-940 Of the two methods the one with the new photometer has the advantage of being direct and of not depending on the assumptio LEWTS A NEW SECTOR SPECTROPHOTOMETEk 319 that photometry in the ultra-violet is uniform with that in the visible region and on the use of accessory apparatus as i n the m&hod published by Hilger.I n any case the two series of results are sufficiently similar to one another to call for a discussion as to which is the more accurate expression of the phenomena. Considerable advant'age is gained by having two sectors which are equal in all respects as in the new photometer. As already stated the substance can be placed in tjhe first and second paths alternately so thatl any slight imperfections in the sectors or in any part of the optical train will express themselves in opposite senses in the tlwo series of spectra; also the work will be confirmed and experimental error corrected a t the same time,.The mean of the two closely concordant results must be a very near approxim-ation to the truth. Further opportunity is provided for eliminating the effect of the solvent directly. Most of the solvents alcohol for example, give feeble absorption spectra which spoil the accuracy of the extinction coefficients of the dissolved substance under investiiga-tion. It is not altogether satisfactory to correct the absorption constants of the solution by subtracting those due to the solvent, which have been ascertained separately and in any case it is laborious to do so; for example it is not safe to assume that a standard curve for absolutely pure alcohol applies to commercially pure spirit. Indeed some of the impurities commonly occurring in rectified spirit are strongly absorbent of ultra-violet light. It is better t o place i n the one path a tube of the solution and in front of the constant sector a similar tube filled with the same solvent as that used in making the solution. It is a good plan to have a tube of the given solvent in each of t.he two paths when adjusting the instrument and then to replace the solvent in one of them by the solution in question. It is perhaps unnecessary tlo do so but it lends a sense of satisfaction while it adds little or nothing to the experimental work. It is certain that the differ-ences then observed in the two spectra are due entirely t o the substance in the dissolved state and hence its absorption curve can be derived directly. There is however the possibility of t.he absorption spectrum being modified by the association of the solute with the solvent but that is a matter for other inquiry in each particular case and does not affect either the general truth of the proposition or the operation of the instrument. STAPLE INN BUILDINGS, HIGE HOLBORN W.C. 1. [Received March 39-4 1919.