Single-Isotope Light Sources in Metrology

N. R. Batarzhukova

In connection with the forthcoming transition to the light meter, it is timely to raise the question of the final choice of the spectral line whose wavelength is to define the unit of length. The idea of using a light wave as an invariant, reproducible standard of length was first expressed more than one hundred and twenty years ago (in 1827) by J. Babinet^1. However, the real possibility of such a use of wavelength appeared only in 1887, when Michelson and Morley^2 developed the first methods for applying the phenomenon of interference of light waves to length measurements. This work should be regarded as the beginning of serious investigations into the question of defining the unit of length by means of a light wave.

In 1892 Michelson^3, while establishing the limits of visibility of interference in various spectral lines, found that most of them have a complex structure, which he did not observe only in one red line of natural cadmium. This is why the latter was proposed by him as the primary one. Michelson’s successful choice provided spectroscopy and metrology with a standard of light wavelengths for more than fifty years^5. The red line of natural cadmium has up to the present served as a standard not only in spectroscopic measurements; it has also been adopted as the primary one in metrology for interference measurements of end measures of length. As subsequent investigations showed, this line, long regarded as simple, in fact also has a complex structure. The components of its hyperfine structure are so close to one another that the width of the line in the radiation of standard^6 light sources did not allow them to be resolved by an interference instrument. The hyperfine structure of the red line of natural cadmium was first observed by M. F. Romanova and A. A. Ferkhmin^4 in studying the spectrum of cadmium emitted by a hollow-cathode tube, which is a source of substantially narrowed spectral lines.

In 1928, during a detailed study of the spectral lines used for interferometric measurements of length,^5 it was established that the wavelength of a complex spectral line seems to change with the path difference in the interferometer. This apparent change is caused by a shift of the interference maximum of the complex line when the path difference changes, owing to a redistribution of intensities among the components of the hyperfine structure. The wavelength corresponding to the center of gravity of the complex line was given the name “effective wavelength of light.” The reproducibility of this value in measurements of measures of different length is a necessary condition for the use of spectral lines in interferometric measurements. The more complex the line and the more widely separated its structure, the worse the wavelength is reproduced as the path difference in the interferometer changes. Therefore the lowest reproducibility of wavelength is possessed by the green line of natural mercury, which also prevented it, despite the exceptional brightness of this line, from being chosen as the standard of wavelengths of light.

In recent years the reproducibility of the wavelength of the red line of natural cadmium has been studied in detail, and repeated comparisons of it with the length of the meter prototype have been carried out.^6 The effective wavelength of this line, emitted by a Michelson lamp under strictly standardized conditions, is reproduced with the highest accuracy ($\pm 0.0001 \div \pm 0.0002 \ \text{Å}$) in comparison with the wavelengths of any other complex spectral lines. This is explained by the fact that the red cadmium line has a very narrow structure, the resolution of which, as has already been mentioned, is possible only in light sources emitting narrowed spectral lines. The reproducibility of the red cadmium line fully satisfies modern high-precision measurements of length in light waves.*)

However, the desire for a further increase in the accuracy of reproducing the unit of length itself is natural. For this purpose it is necessary to extend the limit of absolute interferometric measurements of length, which can be achieved by using light sources of special design that emit narrowed spectral lines. Sources of this kind will not provide special advantages with respect to the reproducibility of the wavelength of the red line of natural cadmium, since the influence of the hyperfine structure, which previously could remain unnoticed, will inevitably begin to make itself felt.

The question of the reproducibility of the wavelength of simple lines, which have no hyperfine structure, is resolved quite differently. The very concept of the “effective wavelength of light” loses its meaning here.

*) See the article by M. F. Romanova, “The Present State of the Question of Defining the Meter by the Wavelength of Light,” printed above (p. 259).

With a symmetric distribution of intensity in a spectral line, the value of its wavelength, referred to the maximum of intensity, in principle cannot change with a change in the path difference in the interferometer. Asymmetry of the line or displacement of the maximum, leading to a decrease in the reproducibility of the wavelength, appears in a simple line only in light sources with a high vapor or gas pressure or in the presence of external electric or magnetic fields.

In the natural state no element emits simple lines. However, as a result of recent advances in nuclear physics it has become possible to obtain stable even isotopes of certain elements. The spectral lines emitted by even isotopes do not possess hyperfine structure and therefore will always have advantages for metrology over the corresponding lines of natural elements.

In some cases the production of simple lines can be accomplished by using interference monochromatization[^7]. This method makes it possible to use, for interference measurements, bright complex lines of natural elements[^8] along with the simple lines of even isotopes.

Elimination of hyperfine structure does not completely solve the question of choosing some spectral line as a primary one, since when a line is chosen requirements are also imposed on its width. All spectral lines emitted by modern sources of monochromatic light are considerably broadened. One of the main causes of the broadening of spectral lines is, as is known, the thermal motion of the emitting atoms. The magnitude of this broadening is proportional to \(\sqrt{\frac{T}{M}}\), where \(T\) is the absolute temperature and \(M\) is the molecular weight of the substance. Proceeding from this law of thermal broadening, when choosing a primary line it is considered necessary to investigate only the radiations of heavy elements (for example, mercury) or of such light elements whose vapor pressure proves sufficient for excitation of the spectrum at very low temperatures (\(-196^\circ\text{C}\) and below)[^9],[^10]. In doing so it is completely forgotten that the Doppler broadening of a line can be appreciably reduced by another method as well, for example by using a light source of special construction that permits observation of the radiation in a direction perpendicular to the motion of the emitting atoms (directed atomic beams)[^11].

Undoubtedly, when choosing a primary spectral line, definite requirements should be imposed not only on the spectral line itself but also on the conditions of its excitation. In most works[^9],[^12],[^13] devoted to this question, the requirement of minimum Doppler broadening is placed first, while the reproducibility of the wavelength when the path difference in the in-

...interferometer remains in shadow. Doppler broadening, being symmetrical, has no effect on the reproducibility of the wavelength. Therefore, from our point of view, it is not the main obstacle to the choice of a line. Its presence must be taken into account when establishing the excitation conditions.

Of considerably greater importance is the displacement of the maximum of the spectral line and its asymmetric broadening under the influence of the pressure of vapor or gas in the light source. Pressure broadening depends on the temperature, density, and kind of gas surrounding the radiating atom[^14]. It is caused by collisions of the radiating atom with atoms of the same kind (resonance broadening), and also as a result of collisions with atoms of another gas, which is necessarily introduced into the light source together with a small amount of the substance under investigation in order to facilitate excitation and maintain the discharge.

The interaction of the radiating atom with its neighbors occurs differently in different substances. Moreover, even for a given substance, not all spectral terms of the radiating atoms are equally subject to the influence of pressure. In the spectrum of any substance there are lines for which, under certain conditions, self-reversal readily appears, and there are lines for which it has never been observed. Self-reversal in lines must undoubtedly be taken into account when choosing a new standard of wavelengths, which must represent the wavelength of a spectral line corresponding to terms least subject to the influence of pressure.

For convenience in measurement, the line must be sufficiently bright, be easily isolated from the spectrum, and, in addition, ensure the best continuity in the determination of the unit of length. The latter means that the accuracy of comparison of the wavelength of the new line with the wavelength of the red cadmium line must be the highest.

In interference measurements of lengths, a great advantage is the use of such a source of monochromatic light which, in addition to the standard line, emits a number of auxiliary simple lines located in the visible region of the spectrum and convenient for finding the order of interference by the method of coincidences of fractional parts. The wavelengths of the auxiliary lines may play the role of secondary standards.

During the last ten years, work has been carried out in the metrological laboratories of a number of countries to create single-isotope light sources with the aim of selecting in their emission a simple spectral line for determining the unit of length. Such sources are light sources filled with even isotopes of mercury (Hg¹⁹⁸, Hg²⁰²), isotopes of krypton (Kr⁸⁴, Kr⁸⁶), and isotopes of cadmium (Cd¹¹², Cd¹¹⁴, Cd¹¹⁶).

LIGHT SOURCES WITH A MERCURY ISOTOPE

The production of the mercury isotope Hg$^{198}$ had already been developed in 1940$^{15,16}$. The pure stable mercury isotope 198, in quantities sufficient for spectroscopic investigations, can be obtained by bombarding chemically pure gold (99.99%) with a powerful beam of neutrons. In this process the following nuclear reaction is used:

\[ {}^{197}_{79}\mathrm{Au}+n\rightarrow{}^{198}_{79}\mathrm{Au} \qquad \text{(with a half-life of 2.7 days),} \]

\[ {}^{198}_{79}\mathrm{Au}\rightarrow{}^{198}_{80}\mathrm{Hg}+\beta^-. \]

Mercury isotopes are also separated in a mass spectrograph, but the degree of purity of isotope 198 proves to be lower.

The first lamp with a mercury isotope, made at the National Bureau of Standards (NBS), was an electrodeless tube (10 cm long and 5 cm in internal diameter) of Pyrex glass. At one end of such a tube there was a bulb for observing the radiation along the discharge. The tube contained the mercury isotope Hg$^{198}$ in a negligible amount and pure argon. The bright spectrum of mercury was excited in it by means of a high-frequency generator (110 MHz). Because of the rapid absorption of the molecules of mercury vapor by the Pyrex, the glow of the tube ceased. However, photographs were obtained of interference rings of equal inclination in all lines of the visible region of the spectrum, which made it possible to conclude that there was no hyperfine structure in these lines. The green line of natural mercury is the most complex of all spectral lines. Therefore its simplicity in the radiation of the isotope Hg$^{198}$ is especially clearly visible in the comparative photograph shown as an example (Fig. 1) of interference rings of equal inclination with a split field. On the basis of these preliminary data alone, the green line $\lambda = 5461$ Å, as the brightest and simplest, was proposed by Meggers for replacing the red line of natural cadmium$^9$.

Fig. 1.

Further improvement of the technique for obtaining stable isotopes in sufficiently large quantities made it possible to create lamps with the mercury isotope for carrying out detailed studies of the radiation of Hg$^{198}$. Such studies were carried out in various laboratories in America and England$^{19,20}$, and also at the International Bureau of Weights and Measures (IBWM) in France$^{21,22}$.

Fig. 2.

The HBS lamps are ordinary gas-discharge tubes (of the Geissler-tube type), with cold cathodes (Fig. 2, a) or electrodeless tubes, like Meggers’ first lamp, but not of Pyrex, rather of special glass (96% silica). As a result of lengthy experiments it was found that 1–3 mg of the isotope at a pressure of 5 mm Hg of pure argon is quite sufficient for filling electrodeless tubes when their glow is excited by a high frequency of 30–100 MHz. It was also established that the higher the generator frequency and the greater the argon pressure, the brighter the mercury spectrum and the longer the service life of the tube. Therefore Meggers used a frequency of 100 MHz to excite the glow of his tubes. The literature contains indications of the possibility of further increasing the frequency for exciting an electrodeless discharge up to 2700–3000 MHz$^{23,24}$ by using magnetron generators. Such a frequency greatly increases the brightness of the lines with very small amounts of luminous substance, but data on how it affects the width of the spectral line are lacking.

By comparison with the wavelength of the red line of natural cadmium, the wavelengths of six lines of the mercury isotope Hg$^{198}$ were determined in the interval from $\lambda = 5790$ Å to $\lambda = 4046$ Å. It was established that the accuracy of their reproduction in repeated measurements is of the order of $\pm 0.0000_9 \div \pm 0.0001$ Å. In these studies special attention was paid to the green line $\lambda = 5461$ Å. The maximum path difference for this line, at which interference is still visible, is 500 mm.

When determining the limit of visibility of interference in the lamp with Hg$^{198}$, it was found that it is greatly reduced if the lamp is not cooled with running water during operation. As soon as the temperature of the luminous source exceeds 30°C, self-reversal appears in all lines except $\lambda = 4358$.$^{21}$ The phenomenon of self-reversal is especially sharply expressed for the green line 5461 Å. The interference rings of equal inclination (Fig. 3) were photographed in the light of the green line emitted by an electrodeless tube without cooling it with running water, at a path difference of 42.2 mm.$^{25}$ Each ring is split into two as a result of pronounced self-reversal of the line. The interferogram shows that, in order to see interference with this line at the greatest possible distance for it (500 mm), it is necessary to maintain the light source at a temperature not higher than 20°C.

Fig. 3.

Attempts to cool the source to a lower temperature led only to a sharp decrease in the brightness of the spectral lines, since already at 15°C strong condensation of mercury vapor begins and the vapor pressure falls so much that the source ceases to emit the mercury spectrum. Thus, eliminating Doppler broadening in mercury lines by cooling the source below 15°C proves unacceptable. It should be noted that, when observing the glow in the tube not across but along the discharge, the green line always appears self-reversed.$^{21}$

Barrell at the National Physical Laboratory (NPL)$^{25}$ investigated the radiation of mercury 198 in tubes filled with pure isotope without an admixture of argon, and in lamps with cold cathodes containing only 0.25 mg of Hg$^{198}$ at 10 mm Hg pressure of pure argon. This investigation enabled him to determine the effect of the argon pressure on the width of the spectral lines of Hg$^{198}$.

On the basis of Barrell’s data, a graph was constructed (Fig. 4) of the dependence of the mean displacement of the maximum of six mercury lines (4046–5790 Å) on the argon pressure in the lamps. In the graph, the argon pressure in mm Hg is plotted along the abscissa axis, and along the ordi-

are the differences between the frequencies of the lines emitted by various HBS tubes (with argon pressure of 3–6 mm) and the frequencies of the lines emitted by an HFL tube containing no argon. The graph also includes a point corresponding to the radiation of a lamp with cold cathodes at 10 mm Hg argon pressure. The values of the wavelengths emitted by the HBS tubes were taken as averages from the works of various authors. The points (for all lamps) for the green

Fig. 4.

mercury line are plotted separately. The average magnitude of the displacement of the maxima of all lines (toward the red side of the spectrum) is proportional to the pressure and amounts to 0.0001 Å (0.0004 cm\(^{-1}\)) per 1 mm Hg. It follows from this that, in order to attain the reproducibility accuracy of the wavelengths of Hg\(^{198}\) indicated in Barrel’s work (\(\pm 0.00005\) Å), it is necessary to control the pressure in manufacturing the lamp with an accuracy of \(\pm 0.5\) mm Hg, or to switch to lamps containing no argon at all. The latter have a very short service life, require large quantities of the isotope mercury—expensive at present—to fill them, and, in order to excite luminescence in them, require a higher frequency from magnetron generators.

Already beginning in 1950, the wavelength of the green Hg\(^{198}\) line

\[ \lambda = 5460.7532\ \text{Å} \]

has been used as the principal one for measurements of other wavelengths. In the work of Meggers and Kessler\(^{26}\), comparison with the wavelength of this line yielded wavelength values for the remaining 27 lines of the Hg\(^{198}\) spectrum with an accuracy of \(\pm 0.0001\) Å. Accuracy of this order is usual for comparison of the wavelengths of simple lines by spectrometric methods with the wavelength

red line of natural cadmium. The measurements of Meggers and Kessler are of great interest, since it was precisely they that led to a revision of the formulas for determining the dispersion of air and to the derivation of new averaged expressions ^27, ^28, one of which was adopted at the International Conference on Spectroscopy in 1952. Also based on the wavelength of the green line of Hg^198, measurements have recently been made in America of sixty wavelengths of the spectra of Hg^198 and Hg^202, as well as twenty lines of Hg I in an electrodeless discharge ^29, ^30. A light source filled with mercury 202 was proposed in the form of an electrodeless tube in 1949 ^31. The mercury isotope 202 for this purpose was obtained by the mass-spectrograph method with a purity of 98.06%.

Light sources with simple lines of the even isotope Hg^198 are already being used also for interferometric measurements of plane-parallel end standards of length up to 100 mm ^32, ^33, ^34.

LIGHT SOURCES WITH THE ISOTOPES KRYPTON 84 AND 86

Krypton lines were proposed for replacing the red line of natural cadmium even before it became possible to separate even stable isotopes. Thus, the yellow-green line \(\lambda = 5650\) Å was proposed by Kester ^35, the line \(\lambda = 5562\) Å by Perard ^25 and, finally, the infrared line \(\lambda = 9751\) Å by Ponizovskii ^36. All these three lines of natural krypton had no substantial advantages in comparison with the red line of cadmium. But when in 1942 ^42 it became possible to obtain pure even stable isotopes of krypton Kr^84 and Kr^86, its lines were again proposed for the choice of a new standard of wavelengths of light. Krypton belongs to the group of medium elements. The visibility limit of interference with its lines, emitted by ordinary light sources, is small, but excitation of krypton lines is possible at very low temperatures. Therefore krypton lines can be strongly narrowed by cooling the light source to the temperature of solidification of oxygen.

Fig. 5.

A special light source, developed by Kester and Engelhardt ^37, enabled them to observe interference in the krypton lines at an optical path difference of 700–800 mm. The external appearance of this source is shown in Fig. 5. The lamp has heated electrodes ^38.

The glow in it is observed, just as in an ordinary Geissler tube, along the capillary. The voltage between the electrodes during the discharge reaches 160 V at a current of 20 mA.

For observing interference at large path differences, the lamp is cooled to the temperature of melting nitrogen. The krypton pressure in it must then be \(h = 0.03\) mm Hg. The service life of lamps containing such small quantities of gas is very short—only a few minutes. To extend the service life, Kesters and Engelbart\(^{37}\) subjected them to prolonged conditioning, which consisted in passing through a lamp filled with krypton to a pressure of 1–2 mm Hg an alternating current of several milliamperes at a voltage of 2000 V for many hours without heating the electrodes. During such conditioning the electrodes are cleaned owing to sputtering. A large part of the krypton atoms, as a result of ionic bombardment, is absorbed by the electrodes. Some of them, because of sputtering of the electrode surfaces, return back. After some time a dynamic equilibrium is established: the number of atoms absorbed by the electrodes under ionic bombardment becomes equal to the number of atoms returning back during sputtering of the electrode surfaces. At the end of the conditioning, the pressure in the lamp becomes equal to several hundredths of a millimeter of mercury. The service life of such lamps reaches 100 hours. A major drawback of such conditioning is the noticeable destruction of the electrodes. They become so brittle that the slightest shaking can destroy the lamp. This drawback forced recourse to another method of increasing the service life. The lamps are filled at normal temperature to a comparatively high krypton pressure (3–5 mm Hg), and then the pressure is regulated by changing the temperature of the light source. For this purpose the latter is immersed in a cylindrical Dewar vessel (Fig. 5). At the boiling temperature of nitrogen, the krypton pressure is reduced to 1.8 mm Hg. In order to reduce the krypton pressure in the light source further, the liquid air in the Dewar is cooled to the solidification point of nitrogen (63° K) by pumping. The Dewar is placed in a metal hermetically closing chamber. In the upper cover of the latter there are feedthroughs for the lamps and two connecting tubes: one for filling the Dewar with liquid air, and the other for connection to a pump. Opposite two silver-free apertures on the lateral surface of the Dewar, two windows are arranged in the chamber. Adjusting devices make it possible to align the center of the lamp capillary with the axis of the condenser and the collimator of the setup.

Having filled the Dewar with liquid air, one of the tubes is hermetically closed and, through the second tube, by means of a pump with a pumping speed of \(5\ \text{m}^3/\text{hour}\), the pressure inside the vessel is rapidly lowered. As a result of the reduction in pressure, the liquid air is cooled to the solidification point of nitrogen (63° K) and even lower—to the point-

freezing of oxygen (55° K). When the liquid air in the Dewar has all cooled to the freezing temperature of nitrogen, the pump may be replaced by another, less powerful one. The temperature inside the Dewar is measured by means of a thermometer inserted through the cover of the chamber. From the extrapolation equation relating the pressure of krypton to the temperature, one finds the value of the pressure in the light source, 0.03 mm Hg, corresponding to 63° K. This pressure is the most favorable for working with the radiation of krypton. According to the investigations of Kesters and Engelgart[^37], the asymmetric broadening and displacement of the maxima of the spectral lines of krypton under these conditions are very small.

Figure 6 gives the curves of the dependence of the displacement of the maxima of the spectral lines of krypton on the pressure in the light source[^37].

Fig. 6.

The displacements \(\Delta n\) are expressed in fractions of the order of interference per meter, and the pressure \(p\) in mm Hg. From these data it can be calculated that the mean value of the displacement for the krypton lines toward the red at a pressure of 3 mm Hg is \(\sim 0.001\)—\(0.004\ \mathrm{cm}^{-1}\), i.e., somewhat greater than for the mercury lines. These experimental curves correspond to the following analytical expression: \(\Delta n = C \cdot p^{2/3}\), i.e., the displacement of the maximum of the krypton lines is caused by the elementary Stark phenomenon, since it is known[^39] that the mean intensity of the electric field in an ionized gas is proportional to the number of ions per unit volume to the power \(2/3\).

The krypton lines under ordinary radiation conditions possess considerable brightness, especially the yellow-green line \(\lambda = 5649\ \text{Å}\). When

at negligible pressures (0.03 mm Hg) in the lamp, their brightness should be still less. For ordinary interference measurements such a light source is hardly simple and convenient in handling.

In none of the published works are there exhaustive data on studies of the reproducibility of the wavelengths of krypton isotope lines. In the papers by Kösters and Engelhardt \(^{37,38}\) there is only an indication that the wavelengths of the lines of \(\mathrm{Kr}^{84}\) coincide completely with the wavelengths of natural krypton. This made it possible for them, despite the lower percentage abundance of \(\mathrm{Kr}^{84}\) as compared with \(\mathrm{Kr}^{86}\), to propose, as a new standard of light wavelengths, the wavelength of the yellow-green line of krypton \(\mathrm{Kr}^{84}\).

Quite recently there appeared a communication \(^{43}\) from the International Bureau of Weights and Measures stating that, in the krypton 86 line \(\lambda = 9856\) Å, emitted by a lamp of the above-described construction, it had been possible to obtain a photograph of interference fringes of equal inclination at a path difference of 1000 mm. However, without a detailed study of the reproducibility of the wavelengths emitted by this light source, and especially of the infrared lines, the visibility of the interference pattern in them at large path differences does not yet indicate its definitive superiority.

LIGHT SOURCES FILLED WITH CADMIUM ISOTOPES

As already mentioned, the reproducibility of the wavelength of the red line of natural cadmium is sufficiently high. It should be assumed that it will be still higher if the radiation of one of its even stable isotopes is used. In the Optical Laboratory of VNIIM, studies were carried out of the radiation of three even stable cadmium isotopes, \(\mathrm{Cd}^{112}\), \(\mathrm{Cd}^{114}\), and \(\mathrm{Cd}^{116}\), in an electrodeless discharge. These studies made it possible to propose \(^{40}\), in place of the red line of natural cadmium, the red line of \(\mathrm{Cd}^{114}\). Such a replacement will ensure the best continuity in the definition of the unit of length, with all the other advantages of monoisotopic radiation.

In the first investigations, the light source with cadmium isotopes was an electrodeless tube similar to the tube with the mercury isotope (Fig. 7, a). This source was filled with cadmium isotope reduced from the oxide, in an amount of 1–3 mg, at 0.2–2 mm Hg pressure of pure argon or hydrogen.

To excite the luminescence of cadmium, the tube was heated in a special furnace (Fig. 8) to \(300^\circ\)C. The temperature inside the furnace was monitored by means of a thermocouple (chromel—copel) with an accuracy of \(\pm 5^\circ\). Such monitoring of the constancy of the temperature made it possible completely to avoid self-reversal in the cadmium lines.

Fig. 7.

Fig. 8.

The excitation of the glow in the tubes was carried out by means of a high-frequency generator of 60–75 MHz.

The presence of additional heating of the lamp is a certain complication of an exceptionally simple light source. Therefore, in the latest models of electrodeless tubes cadmium is excited without additional heating. These tubes (Fig. 7, b) are small quartz ampoules 5 mm in diameter and 30–40 mm long. The ampoule is filled with 1–2 mg of a cadmium isotope at 0.2 mm Hg pressure of pure argon and is placed in a glass bulb with silvered walls. The air is pumped out of the bulb. Excitation of the glow in the lamp is produced by the same high-frequency generator. The cadmium is heated in the argon discharge, and gradually, in place of the argon spectrum, a bright cadmium spectrum appears. Such lamps without an additional furnace are especially convenient for use in interference measurements of end standards of length on modern interference comparators. A study of these lamps showed that their service life is sufficiently long and that the cadmium lines they emit do not have self-reversal. Lamps of the Michelson type (Fig. 7, c) and lamps with heated electrodes (Fig. 7, d), filled with the isotope Cd¹¹⁴, were also made in the laboratory. In the latter, a certain improvement was introduced in the form of additional heating of the inner bulb of the lamp. The heating made it possible to reduce the current passing through the discharge to several tenths of an ampere. This provides a certain narrowing of the cadmium lines emitted by a lamp of this design.

In the photograph, given as an example in Fig. 9, of interference fringes of equal inclination for the green cadmium line

Fig. 9.

(on an enlarged scale), spectrum a corresponds to the radiation of natural cadmium, and spectrum b to the radiation of Cd¹¹⁴. The absence in spectrum b of subsidiary interference maxima corresponding to the hyperfine structure of the line clearly demonstrates the simplicity of the lines of isotopic cadmium.

In measurements of the wavelengths of the spectral lines Cd¹¹², Cd¹¹⁴, and Cd¹¹⁶ by the interference method of comparison with the wavelength ...

red line of natural cadmium, a high precision of their reproduction was obtained (of the order of \(\pm 0.9 \cdot 10^{-4}\ \text{Å}\) for the red line and \(1.7 \cdot 10^{-4}\ \text{Å}\) for the others). The value obtained for the wavelength of the red line of \(\mathrm{Cd}^{114}\) differs from the wavelength of natural cadmium by only \(0.0018\ \text{Å}\). Comparison of two such close lines can be carried out with very high accuracy, since the possible systematic errors connected with the introduction of corrections for the dispersion of air and for the dispersion of the phase jump upon reflection from the light-dividing layers of the interferometer mirrors are reduced to a minimum.

In the photograph of the interference rings (positive) with a divided field (Fig. 10), the right-hand side corresponds to the interference rings in the light of the red line of the isotope \(\mathrm{Cd}^{114}\),

Fig. 10.

and the left-hand side to the red line of natural cadmium. The weakening of the background at the minima and the increase in the contrast of the interference rings (right-hand side) once again confirms the presence of some narrowing of the red line of the isotope \(\mathrm{Cd}^{114}\) in comparison with natural cadmium. The photograph was obtained at an optical path difference of \(200\ \text{mm}^{41}\).

Preliminary investigations of the influence of the pressure of the foreign gas in the light source on the width of the red line of \(\mathrm{Cd}^{114}\) showed that the displacement of the maximum occurs toward the red, and the magnitude of this displacement in any case does not exceed \(1 \cdot 10^{-4}\ \text{Å}\) \((0.0004\ \text{cm}^{-1})\) per \(1\ \text{mm Hg}\).

If up to the present time cadmium sources have been used only for measuring small lengths, not exceeding \(100\ \text{mm}\), then cadmium mono-isotopic light sources make it possible to carry out meas—

...measurements of end standards up to 200 mm long. In the light of monoisotopic cadmium lines, measurements of end standards up to 175 mm long have already been made on an interference comparator.

CONCLUSION

The survey of monoisotopic light sources that has been carried out shows that any line emitted by these sources, given the present state of interferometric length measurements, is suitable, on an equal footing with the red line of natural cadmium, for being chosen as the fundamental one, and the choice of a line should not delay the transition to defining the unit of length through a light wave.

A higher reproducibility of the standard of length in light waves is ensured by choosing the best of the radiations of even isotopes of even elements. For such a choice, the results of the experimental investigations cited are still insufficient. Undoubtedly, investigations of all the proposed light sources should be carried out in national laboratories and at the International Bureau of Weights and Measures. Although all three monoisotopic light sources are already being used in interferometric measurements of length, up to the present time no comparison has yet been made of the radiations of these sources with one another in order to determine the metrological advantages of one of them.

Certainly, the production of monoisotopic light sources need not be limited only to the three elements mentioned above (mercury, krypton, and cadmium). It is possible to create monoisotopic light sources filled with even isotopes of other elements, for example lead 206 and 208, xenon 132, and even isotopes of some rare earths.

In order to further increase the accuracy of reproducing the unit of length in light waves in comparison with the existing one, it is desirable to extend the limit of visibility of interference in the selected line. In this respect, in choosing a line, a transition to spectral lines with large wavelengths, located in the far red and near infrared regions of the spectrum, is of some interest.

Line narrowing can be obtained by using absorption lines. Absorption lines have not yet been used for interferometric length measurements.

Extension of the limits of visibility of interference, as already mentioned, is accomplished by the use of narrow lines emitted by sources of special design.

In creating such sources, it seems to us that it would be correct to proceed from the following requirements:

1. Sufficient brightness of the glow at a comparatively low pressure of the luminous gas.

1. The influence of thermal expansion reduced as far as possible.

2. Low voltage for exciting the glow.

3. Minimum current density.

4. Long service life of the source.

5. Simplicity of the light source and absence of complex auxiliary equipment for exciting the glow.

6. Small dimensions, permitting its use in work on modern interferometers.

7. Safety in use.

The development of these light sources is a task for the nearest investigations.

CITED LITERATURE

1. Babinet, Ann. chim. et phys. 40, 177 (1829).
2. A. A. Michelson and E. W. Morley, Am. J. Sci. 34, 427 (1887).
3. A. A. Michelson, Phil. Mag. 34, 280 (1892).
4. M. F. Romanova and A. A. Ferkhmin, DAN 2 (1933).
5. A. Pérard, Rev. d’Opt. 7, 1 (1928).
6. M. F. Romanova, UFN 37, No. 2, 163 (1952).
7. N. R. Batarchukova, DAN 63, No. 6, 1013 (1947).
8. N. R. Batarchukova and A. I. Kartashev, Izv. AN SSSR, ser. phys., 14, No. 6, 753 (1950).
9. V. F. Meggers, UFN 34, No. 1, 105 (1948).
10. J. Cabanes, Tirage à part des Proc. Ver. C. I. Pds et Mes. 2nd series, XXIV, M—55 (1954).
11. Minkowsky und H. Bruck, Zeits. f. Phys. 95, No. 5, 284 (1935).
12. A. Pérard, Comptes rendus 237, 364 (1953).
13. R. Lennuier, Astronomie 67, 337 (1953).
14. H. Margenau and W. Watson, Rev. Mod. Phys. 8, 22 (1936).
15. J. H. Wiens and L. W. Alavarez, Phys. Rev. 58, 1005 (1940).
16. J. H. Wens, Phys. Rev. 70, 910 (1946).
17. Keim, Phys. Rev. 76, 12, 70 (1949).
18. W. F. Meggers and F. O. Westfall, J. Res. NBS 44, 5, 447 (1950).
19. H. Barrell et M. Puttock, C. R. IX Conf. Gen. Pds. Mes. 77 (1949).
20. J. M. Blank, J. Opt. Soc. Am. 40, 6, 345 (1950).
21. A. Pérard et J. Terrien, J. phys. et rad. 11, No. 6, 249 (1950).
22. A. Pérard et J. Terrien, Comptes rendus 228, 964 (1949).
23. E. Jacobsen and Harrison, J. Opt. Soc. Am. 39, 12 (1949).
24. M. Zelikoff, P. Wickoff, Z. Aschenbrand, R. Loomis, J. Opt. Soc. Am. 42, No. 11, 818 (1952).
25. H. Barrell, Proc. Roy. Soc. 209A, 132 (1951).

1. W. Meggers and K. Kessler, J. Opt. Soc. Am. 40, No. 11, 737 (1950).

2. H. Barrell, J. Opt. Soc. Am. 41, No. 5, 295 (1951).

3. B. Edlen, J. Opt. Soc. Am. 43, No. 5, 339 (1953).

4. K. Burns and K. Adams, J. Opt. Soc. Am. 42, No. 1, 56 (1952).

5. G. Fowles, J. Opt. Soc. Am. 44, No. 10, 760 (1954).

6. J. Mc Nally, J. Griffin and Z. Burkhart, J. Opt. Soc. Am. 39, No. 12, 1036 (1949).

7. H. Barrell, Proc. Verb. C. I. Pds et Mes., XXIII—A, 149 (1952).

8. J. Terrien, Tirage à part Proc. Ver. C. I. Pds et Mes., XXIV, M—95 (1953).

9. J. Claesen, Bul. de Métrologie, 168, 227 (1954).

10. W. Kösters, Mitt. Phys. Techn. Reichsanstalt 11, No. 4, 51 (1927).

11. L. B. Ponizovsky, Priroda, No. 2, 27 (1945).

12. W. Kösters et M. Engelhard, Proc. Verb. C. I. Pds et Mes., XXII, 137 (1950).

13. M. E. Engelhard, Proc. Verb. C. I. Pds et Mes., XXIII—A, 665 (1952).

14. P. Debye, Phys. Zeits. 20, 160 (1919).

15. N. R. Batarchukova, A. I. Kartashev and M. F. Romanova, DAN 90, No. 2, 153 (1953).

16. N. R. Batarchukova, A. I. Kartashev and M. F. Romanova, “Measurement of the wavelengths of the radiation of C$^{114}$ in the visible region of the spectrum.” Proceedings of VNIIM, special issue (1954).

17. K. Clusius und G. Dickel, Zeits. f. phys. Chem. 52, 348 (1942).

18. J. Terrien et J. Hamon, Comptes rendus 239, 586 (1954).