be of different thicknesses at different points throughout the free opening, which would give a nonuniform field. At some points the plate, being thinner or thicker, would have less or more rotation than at other points. Since a standard should be at least as accurate, and preferably more so, than the instrument it is to control, or check, and since the best modern saccharimeters are capable of approaching an accuracy of 0.01° S, it is necessary that quartz control plates do not differ in thickness at any point within the free aperture by more than an amount corresponding to about 0.01° S. Since a 100° plate is 1.59 mm thick, 0.01° S corresponds to a thickness of 0.000159 mm, or between one-fourth and one-third wave length in air of the light usually used in testing them. The planeness of the faces can be tested most conveniently by observing the interference pattern formed, by reflections from the surface being tested, and from an optical flat. Either a mercury or a sodium light source is satisfactory for the purpose. The parallelism of the faces is most easily checked by observing the Haidinger rings formed by reflections from the two surfaces of the plate. Both of these methods as applied to quartz control plates are described in detail by Brodhun and Schönrock [38]. (3) AXIS ERROR.-It is important that the faces of the plate be accurately oriented at 90° to the crystallographic axis, i. e., when in use the light shall pass through the plate in a direction parallel to the principal or crystallographic axis of the quartz crystal. The accuracy of the construction of the plate in this respect is checked by observing the amount of displacement of the uniaxial interference figure when the plate is rotated in its own plane. This scheme was used by Gumlich [39] and later perfected by Schönrock [40], and by Brodhun and Schönrock [38]. The latter designed a special instrument for the purpose, which was built by Schmidt & Haensch and also by R. Fuess. The method is capable of a precision of a few seconds of arc, which is more than is needed, since the tolerance permitted between the crystallographic axis and the normal to the faces of the plate is 12 minutes of arc. (4) MOUNTING.--Another factor of great importance, and one that in the past has not been given the attention it deserves, is that of the mounting. The plate should be mounted loosely in a metal frame, the axis of which forms an angle of 90° with the plate. The amount of play between the faces of the plate and the frame should be as small as possible, but the metal should exert no pressure upon the plate under any conditions. A little play around the circumference of the plate does no harm and is desirable in checking whether the plate is free from pressure and yet has not too much play. One of the most sensitive and re FIGURE 16.-A, Axis-error apparatus for testing quartz control plates, and group of mounted quartz plates and the optical flat used for testing the faces for planeness; B, Bausch & Lomb type of mount; C, Hilger type of mount; (The steel ring, C, which carries the quartz plate, G, is ground parallel and lapped down until it is just 0.00015 inch thicker than the plate. It is held in place against the flat surfaces, A and D, by the screws, F. The shoulder, B, rests in the trough of the instrument in the usual manner.); D, Physikalische-Technische Reichsanstalt type of mount (dimensions in millimeters). liable tests is made by ear in a place free from disturbing noises. If the plate is placed close to the ear and shaken sharply in a direction parallel to the axis of the plate, no click should be heard, or at most only a very faint one. A shake parallel to the faces should produce a decided click, proving that the plate is not bound in position by pressure. (5) MEASUREMENT OF ROTATION.-Inasmuch as precision work of the highest type is frequently called for in meeting the demands for accuracy in standardizing quartz control plates, the mercury line, λ=5461 A, for the reasons previously outlined, is used exclusively. Temperature coefficient. In order to utilize this source to best advantage, the temperature coefficient, a, of quartz for this wave length (X=5461 A) and also the constant x-5892/PX-5461 Where is the rotation, were determined. It was thought advisable to measure a for this wave length, although Lang [41], Sohncke [42], and Le Chatelier [43] state that it has the same value for all wave lengths. We have $i=0 (1+at), where, is the rotation at temperature t and 40 at zero. For any other temperature, t1, we have Hence =40 (1+at1). α= The measurements [24] were made with the most improved types of apparatus. By computing a from the value of at temperatures between 4° and 50° C, the value obtained was a=0.000144. Then between 4° and 50° C ¢r=¢o (1+0.000144t). (13) Conversion factor.-In measuring x-5892-5/x-5461, a large number of determinations were made, practically all of which were concordant. However, in order to eliminate the personal equation and avoid, as far as possible, errors due to the character of the sodium source, the value of is computed from the measurements of five plates whose sodium values have been determined at the PhysikalischTechnische Reichsanstalt. The mean of these values and those of this Bureau was taken as the rotation for λ=5892.5. The rotations were measured in part with a sensitive-strip polarizing system. The greater number, however, were made with an exceptionally good Lippich system. The average value obtained was Thus any quartz rotation for the wave length 5892.5 may be obtained by measuring the rotation for the wave length 5461A and multiplying it by the constant 0.85085. By this method the errors due to the character of the sodium source of light are eliminated, and the measurements of one observer may be readily compared with those of another. National Bureau of Standards certificates show the rotation in circular degrees at 20° C for wave lengths λ=5461 and λ=5892.5 A. The latter is the so-called optical center of gravity of the two sodium D1 and D2. Measurement.-In carrying out the rotation measurement, the plate is placed in the air bath of the circular-scale polariscope described on page 46, and the optical rotation carefully measured at 20° C for spectrally purified mercury light of wave length 5461 A. In order to eliminate accidental errors as far as practible, at least three sets of readings are made on different days. Moreover, all readings are compared to, and corrected by readings on a set of standard plates which this Bureau maintains as its primary standards. Because of the refinements in the method of checking quartz control plates, it is believed that the sugar values certified are not in error by more than 0. 01°S, or at most 0. 02°S. Repeated determinations made several years apart seldom differ by more than this amount. (b) NATIONAL BUREAU OF STANDARDS SPECIFICATIONS In line with the requirements cited above, this Bureau has drawn up the following specifications for quartz control plates: (1) PURITY. The plate shall be made of quartz which is optically homogeneous, i.e., it shall be free from twinning, striae, strain, and other optical defects. (2) PLANENESS, PARALLELISM. The faces of the plate shall be both plane and parallel within the following limits: There shall be no departure from true flatness by more than that corresponding to one-half wave length in air of sodium light anywhere within the free aperture of the plate, nor shall there be a radius of curvature of less than 100 m. The plate shall not differ in thickness between any points comprised within the free aperture by more than 0.00015 mm. This thickness corresponds to 0.01°S. (3) AXIS ERROR. The faces of the plate shall be accurately orientated at right angles with respect to the crystallographic axis. The axis error, i.e., the angle between the normal to the faces and the crystallographic axis, should be as small as possible and shall not exceed 10 or 12 minutes of arc. (4) MOUNTING.-The plate shall be mounted loosely in a metal frame, the axis of which forms an angle of 90° with the faces of the plate. The amount of play between the plate and its surrounding frame shall be as small as possible in the direction parallel to the axis of the plate, but the metal shall exert no pressure upon the plate. (5) DIMENSIONS.-The plates should be 15 to 17 mm in diameter and, after mounting, should have a free aperture not less than 10 mm in diameter. Plates having low sugar values may be, and it is recommended that they be, composed of two thicker plates, one rightrotating and one left-rotating, mounted in separate mounts of the same type, preferably one on each end of the tube. (c) SPECIFICATIONS OF THE INTERNATIONAL COMMISSION FOR UNIFORM METHODS OF SUGAR ANALYSIS FOR QUARTZ CONTROL PLATES The following resolutions were adopted by the International Commission for Uniform Methods of Sugar Analysis at its Eighth Session at Amsterdam in 1932: 1. Quality tests of saccharimeter quartz control plates. (a) Optical purity.-At least the central 9 mm of the plate must be sufficiently optically homogeneous. Especially plates from 90° to 102°S must not show any striae. During the purity test, the plate should be rotated in its own plane at least 360° (about its own axis). (b) Plane parallelism. In general the wedge angle between the faces may at most amount to not more than 20 seconds. For plates between 90° and 102°S this limit must be 10 seconds. The radius of curvature of each of the faces shall be not less than 50 meters. (c) The axis error, that is, the angle between the optical axis and the normal to the plate, may at the highest amount to 10 minutes of arc. 2. Dimensions of the plates.-The plates must be between 15.0 and 17.0 mm in diameter. After mounting they must have a free aperture at least 10 mm in diameter. Plates having values below +24°S should be composed of two thicker plates, one plus and one minus, the combined thickness of which, however, must be less than 1.6 mm. The edges of the plates shall be slightly beveled. 3. Identification marks. Near the edge of the plate shall be engraved "IP" (International Plate), the number of the plate, and the year. 4. Plate mountings. It is proposed to accept the form of mounting prescribed by the Physikalisch-Technische-Reichsanstalt wherein the plate is free from pressure, but the clearance is a minimum. For plates below 24°S, the two plates are to be mounted in separate mounts of the same type, one on each end of the holder tube. 5. Rotation measurement. The rotation in circular degrees shall be made at 20°C, using spectrally purified light either of wave length 5461 or 5892.5 A, obtained, respectively, from a mercury vapor arc of suitable design, or from a sodium light source, such as the Pirani or Osram sodium arc lamp. 6. The plates shall be tested in one or more of the four national physical laboratories, viz., National Bureau of Standards, Washington, D. C.; National Physical Laboratory, Teddington, London; the Physikalisch-Technische Reichsanstalt, Charlottenburg, Berlin; and the Laboratoire National, Paris. All details of the tests and measurements shall be left to these four institutions. At the Ninth Session at London in 1936, the Referee on Quartz Control Plates made the following recommendations which were unanimously adopted by the Commission: 1. The resolution of the Eighth Session shall be amended to read as follows: (a) Optical purity.-The central 9 mm at least of the plate must be sufficiently optically homogeneous. It is essential that plates from 90° to 102° S should not show any striae. During the purity test, the plate should be rotated in its own plane at least 360° (about its own axis). It is recommended that the four National Physical Laboratories should investigate tests for fixing a quantitative limit to permissible defects in homogeneity. (b) Identification marks.-Near the edge of the plate shall be engraved "I. P." (International Plate), the number of the plate, the year, and the sign of the testing National Physical Laboratory. (c) Plate mountings.-The form of mounting to be acceptable for testing shall be such that the plate shall be free from compression and the clearance a minimum. For plates below 24° S, the two plates are to be mounted in separate mounts of the same type, one on each end of the holder tube. 2. The four National Physical Laboratories, viz., National Bureau of Standards, Washington, D. C.; National Physical Laboratory, Teddington, London; the Physikalisch-Technische Reichsanstalt, Charlottenburg, Berlin; and the Laboratoire National, Paris, be requested to collaborate and determine: (a) The rotation of the 100° S plate in circular degrees for the wave length (optical center of gravity) produced by the Osram sodium vapor arc lamp. (b) The rotation in circular degrees for the wave length (optical center of gravity) produced by the Osram sodium vapor arc lamp for plates of the following approximate values, 25, 50, 75, and 100° S. It is suggested that sets of such plates be interchanged between the above laboratories for comparative tests to be made. (d) CERTIFICATION OF QUARTZ CONTROL PLATES The usual procedure at this Bureau in testing quartz control plates is: 1. Examination of the mounting to see if the plate is satisfactorily mounted. 2. Examination between crossed nicols as a test of purity. |