Full Text
FROM CURRENT LITERATURE
NEW TYPES OF MASS SPECTROMETERS BASED ON TIME MEASUREMENT
Separation of isotopes by mass in mass spectrometers is usually based on differences in the deflection of ions in electric or magnetic fields.
The desire to simplify the design of the mass spectrometer in order to turn it into an accessible laboratory instrument, and also the effort to increase its resolving power (which is especially important for heavy isotopes), have prompted researchers to seek new principles for the design of these instruments.
Recently increasing attention has been drawn to the possibility of separating isotopes on the basis of differences in the times of flight of the corresponding ions in electric or magnetic fields.
Mass spectrometers constructed on this principle may be divided into three groups.
The simplest in concept and the crudest in accuracy are the so-called pulsed mass spectrometers1,2. In these instruments ions of different masses are accelerated in an electric field to a certain energy, then move freely in a vacuum chamber, reaching the detector after time intervals that depend on the ion masses. The time intervals are measured with the aid of a cathode-ray oscilloscope connected to the detector. Thus it is possible to record simultaneously ions belonging to an entire region of the mass spectrum, which is especially valuable when it is necessary to analyze gases with rapidly changing composition. Another merit of these instruments is their relative simplicity (the absence of a magnetic field). However, their resolving power is quite low, and the accuracy of measurements falls off in the region of large masses.
The second group includes the so-called radio-frequency mass spectrometers3,4. In these instruments a magnetic field is likewise not used. Ions, preaccelerated in an electric field to a certain energy, enter the region of an alternating radio-frequency electric field. The amount of additional kinetic energy acquired by the ions in this field depends on their mass and on the field frequency. At a definite relation between these two quantities the ions acquire the maximum amount of kinetic energy, enabling them to overcome the retarding electric field placed in front of the collector. There are also other variants of mass spectrometers that use the principle of ions acquiring additional kinetic energy in a radio-frequency electric field5.
The resolving power of these instruments is low in comparison with magnetic mass spectrometers; however, their simplicity and portability
allow us to think that in the near future such mass spectrometers will become widespread in chemical and factory laboratories.
The most promising from the standpoint of the possibility of increasing the accuracy of mass measurements appear to be mass spectrometers based on measuring the period of revolution of an ion in a homogeneous magnetic field. As is known, this period
\[ T=\frac{2\pi c}{e/m\,H}=655\,\frac{M}{H} \]
(\(M\)—in atomic mass units, \(H\) in gauss) depends neither on the velocity nor on the initial direction of motion of the ion, and in this sense we have a peculiar “double focusing.” This exceptionally valuable property has already been used in a number of new types of mass spectrometers, which we shall now consider in somewhat greater detail.
In the “omegatron”\({}^{6}\), which in many respects resembles a miniature cyclotron and is akin in its principle of operation to radio-frequency mass spectrometers, the frequency of revolution of an ion in a homogeneous magnetic field is determined directly (whence the name of the instrument). Ions of specific mass \(m/e\) are accelerated by a homogeneous radio-frequency electric field \(E_0\sin\omega t\), directed perpendicular to the magnetic field \(H\). At resonance between the frequency of the electric field \(\omega\) and the cyclotron frequency of the ions
\[ \omega_c=\frac{eH}{mc}, \]
the radius \(R\) of the trajectory of ions of the corresponding mass gradually increases until, at some value \(R_0\), the ions strike the detector.
In fact, ions whose cyclotron frequency differs from the radio frequency by some amount \(\varepsilon=|\omega-\omega_c|\) also reach the detector to some extent. The maximum possible value \(\varepsilon=\varepsilon'\) determines the resolving power of the instrument,
\[ \frac{M}{\Delta M}=\frac{\omega_c}{2\varepsilon'}. \]
This quantity is easily determined if one recalls that the ions move along a spiral of radius
\[ r=\frac{E_0}{H\varepsilon}\sin\frac{\varepsilon t}{2}. \]
If the detector is at a distance \(R_0\) from the starting point of the ion motion, then only those ions will reach it for which
\[ \varepsilon \leq \varepsilon'=\frac{E_0}{HR_0}. \]
Consequently, the resolving power is
\[ \frac{M}{\Delta M}=\frac{\omega_c}{2\varepsilon'}=\frac{R_0H^2e}{2E_0M}. \]
At a constant magnetic-field strength the resolving power decreases as the ion mass increases. If, however, the frequency of the electric field is kept constant and the magnetic-field strength is varied, then the absolute accuracy of the measurements \(\Delta M\) proves to be independent of mass. This latter method of measurement, however, is in practice less convenient.
In the instrument constructed\({}^{6}\), the electric field was applied to two flat plates \(3\times 5\) cm in size. The distance between the plates was 2 cm. The magnetic-field strength was \(H=4700\) gauss. The detector was placed on a circle of radius 0.9 cm.
Measurements were made only with light isotopes. The resolving power for the \(H^+\) ion proved equal to 10,000; for the heaviest of the measured ions, \(N_2^+\), the resolving power was 5000.
A direct measurement of the period of revolution of an ion in a homogeneous magnetic field was carried out in an ingenious and comparatively simple mass spectrometer, called by its author the “chronotron”\({}^{7,8}\).
In a vacuum chamber with a vertical magnetic field there is placed an ion source, and at a distance of 12 cm below it—a detector. Ions leaving the source move along a helical line and, after a certain number of turns \(N\), enter the detector. In this process there is “spatial focusing” of the ions, consisting in the fact that after each turn all the ions, irrespective of their velocity or direction of emission, intersect the line of force passing through the source. Consequently, it is sufficient to place the ion detector on one line of force with the source.
Ions emitted from the source with a suitable vertical component of velocity enter the detector. One turn before entering the detector is made by ions possessing the maximum admissible vertical velocity. Of practical importance is the circumstance,
Fig. 1.
that, in order to obtain the required vertical component of velocity, no special electric accelerating field is required; the thermal motion of the ions emitted by the source proves quite sufficient for this. As a result, pulses of different “orders” \(N\) differ in intensity in accordance with the distribution of the ions over thermal velocities. Pulses of the first order are of low intensity and are poorly visible on the oscillogram (see Fig. 1); the pulses of the third order proved to be the most intense.
To increase the accuracy of the measurements, it is advantageous to register ions of the highest possible orders. For a given distance between the source and the detector, the maximum number of observable turns is determined by the dimensions of the source. Ions moving along a helical line with a very small pitch, after the first turn, fall onto a given wall of the source and, consequently, are not registered. In the apparatus described, it was possible to observe ions up to the twelfth order.
The ion pulses were obtained from a source of the usual type. The duration of the pulses was of the order of 0.25–0.5 microseconds. The accelerating voltage reached 600 V. Owing to the comparatively short duration of the accelerating voltage, all ions lying within a definite mass interval passed through the exit slit of the source after the accelerating field had ceased to act. Consequently, in this interval all ions, irrespective of mass, possess the same horizontal component of momentum and, therefore, move along circles of the same radius
\[ R=\frac{mv_{\perp}c}{eH}. \]
This makes it possible simultaneously to measure the masses of ions lying within a rather broad interval
from the literature
masses, without the need to change the parameters of the source (the accelerating voltage or the pulse duration).
A number of instruments were tested as ion detectors. In addition to high sensitivity and low inertia, the detector is also required to operate satisfactorily in a magnetic field.
Initially a scintillation counter was used;[^9] the light from the phosphor, placed inside the vacuum system, was transmitted to the photomultiplier through a quartz “light guide.” The rather considerable duration of the phosphor glow, which distorted the shape of the ion pulse (see below), and also the need for a careful optical screening, forced the authors to abandon this type of detector and to use a magnetic electron multiplier specially developed for these purposes, with dynodes made of beryllium bronze.
A very important part of the apparatus is the equipment for measuring time. On the screen of an oscilloscope connected to the detector, a series of equally spaced pulses is obtained, corresponding to ions of different orders (Fig. 1). The distance between neighboring pulses determines the time of one revolution. Usually the time between the second and ninth pulses was measured.
In practice, the two pulses between which the time interval is to be determined were fed to two separate sweep traces of a second, fast oscilloscope. One of the pulses is then superposed on the other, and from the corresponding displacement of the beginning of the sweep one judges the initial distance between the pulses. For better superposition it is desirable that the pulse shapes be identical.
By this method it was possible to measure time intervals with an accuracy down to 0.01 microsecond. In the field used, \(H = 450\) gauss, this theoretically corresponds to 0.001 atomic mass unit. In experiment, the masses of isotopes were measured beginning with \(S^{32}\) and ending with \(Xe^{134}\). The average accuracy proved to be of the order of 0.002 mass unit. For heavy isotopes such accuracy is quite satisfactory.
To increase the accuracy of the measurements, the time of motion of the ions should be made as large as possible.
Reducing the magnetic-field strength for this purpose is undesirable, since, first, it entails the need to increase the radius of the vacuum chamber and, second, strongly increases the effect of ion scattering because the energy of the ions moving along a circumference of a given radius decreases:
\[ E=\frac{H^2R^2}{144^2M}. \]
An increase in the number of revolutions \(N\) can be achieved at the cost of lengthening the tube or reducing the size of the source; both of these possibilities are also unsatisfactory from the practical point of view.
An interesting solution of the problem of increasing the number of revolutions has been achieved in a mass spectrometer of a somewhat different type, called a “synchrometer.”[^10] In this instrument ions making \(N\) revolutions in a plane perpendicular to the magnetic field are recorded. Unlike the omegatron, no accelerating electric field is used here, and the ions move along one and the same circumference.
To prevent ions from striking the source after the first half-turn, the ions are slowed by an electric field. For this purpose, on the path of the ions leaving the source, a system of three slits is placed (Fig. 2). Slits \(S_2\) and \(S_4\) are grounded, while slit \(S_3\) is connected to a generator of rectangular pulses of short duration. The slowed ion pulses subsequently move along a circumference of smaller diameter, passing after each revolution through the slit system. After a certain known time interval \(t\), a second retarding pulse is applied to \(S_3\) …
slowing pulse, and those ions enter the detector for which \(t = NT\), where \(T\) is the period of one revolution.
In the model constructed, the ions moved in a field produced by a magnet with poles 38 cm in diameter and a gap of 5 cm. The diameter of the principal orbit was 25 cm. The field strength was \(H = 820\) gauss.
Ions of mass 18 were recorded after \(N = 70\) revolutions. The flight time \(NT \approx 1\) millisecond was measured with an accuracy of \(\Delta(NT) \approx 0.1\) microsecond (the width of the corresponding current pulse in the detector). The resolving power of the instrument
Fig. 2.
\[ \frac{M}{\Delta M} = \frac{NT}{\Delta(NT)} \]
was therefore equal to 10,000. It was also possible to measure the mass of the sulfur isotope with an accuracy of up to \(0.001\) atomic mass unit.\(^{11}\)
The author believes that in the large synchrometer now under construction it will be possible to increase the accuracy to \(10^6\).
The chief advantage of the synchrometer over the chronotron is the possibility of obtaining a large number of ion revolutions with a comparatively small vacuum-tube size; in addition, the overlap of ions of different orders also proves to be considerably smaller. Finally, the pulse intensity can be significantly increased by reducing scattering and also by using simple electrical focusing of the ions emerging from the source.
In comparison with the omegatron, a magnetic field of substantially greater extent must be used; however, this involves practically no spatial charge that would noticeably hinder the attainment of high accuracy in the omegatron.
L. B.
CITED LITERATURE
- R. Keller, Helv. Phys. Acta 22, 386 (1949); UFN XLIII, 299 (1951).
- A. Cameron, D. Eggers, Rev. Sci. Instr. 19, 605 (1948); UFN XL, 619 (1950).
- W. Bennett, J. Appl. Phys. 21, 143 (1950); UFN XLIII, 301 (1951).
- A. Henson, J. Appl. Phys. 21, 1063 (1950).
- P. Schissel, J. Appl. Phys. 22, 680 (1951).
- H. Sommer, H. Thomas, J. Hipple, Phys. Rev. 82, 697 (1951).
- S. Goudsmit, Phys. Rev. 74, 622 (1948); UFN XL, 619 (1950).
- E. Hays, P. Richards, S. Goudsmit, Phys. Rev. 84, 824 (1951).
- P. Richards, E. Hays, Rev. Sci. Instr. 21, 99 (1950).
- L. Smith, Rev. Sci. Instr. 22, 115 (1951).
- L. Smith, Phys. Rev. 81, 295 (1951).