Ion Composition Experiment M. A. COPLAN, K. W. OGILVIE, P. A. BOCHSLER, AND J. GEISS Abstract-An investigation using a novel ion mass spectrometer for measuring the ionic composition of the solar wind from the ISEE-C spacecraft is described. The resolution and dynamic range of the instru- ment are sufficient to be able to resolve up to twelve individual ions or groups of ions. This will permit the solution of a number of funda- mental problems related to solar abundances and the formation of the solar wind. The spectrometer is composed of a stigmatic Wien filter and hemispherical electrostatic energy analyzer. The use of curved electric field plates in the filter results in a substantial saving of weight with respect to a conventional filter of the same resolution and angular acceptance. The spectrometer is controlled by a microprocessor based on a special purpose computer which has three modes of operations: full and partial survey modes and a search mode. In the search mode, the instrument locks on to the solar wind. This allows four times the time resolution of the full survey mode and yields a full mass spectrum every 12.6 min. 1. SCIENTIFIC AIMS ION COMPOSITION studies in the solar wind are essential for understanding the dynamics and energetics of the solar wind acceleration region, and are an important source of solar abundance information. The ion composition of the solar wind is, however, poorly known because of the technical difficulties associated with the separation of the different ions over the wide dynamic range involved. This is illustrated in Table 1. Manuscript received April 3, 1978. The design and calibration phases of this work were supported. in part by the Swiss National Science Foundation. M. A. Coplan is with the Institute for Physical Science and Tech- rlology, University of Maryland, College Park, MD 20742. K. W. Ogilvie is with the Laboratory for Extraterrestrial Physics, Goddard Space Flight Center, Greenbelt, MD 20771. P. A. Bochsler and J. Geiss are with the Physikalisches Institut, Universitat Bern, Bern, Switzerland. With electrostatic energy analyzers, the He+2/H+ abundance ratio has been studied extensively in the solar wind, and large variations found [9], [16], [25], [26]. During quiet low- temperature conditions, ions other than H+ and 4He+2 have been separated with this technique and the presence of heavy ions has been demonstrated. In addition, some information about abundances and charge state distributions has been obtained [3]-[5], [16]. The foil collection technique has been used to determine the abundances of the isotopes of He, Ne, and Ar in the solar wind [12]-[14]. Although this method has yielded precise abun- dance data, its time resolution is limited. Mass per unit charge (M/Z) spectrometers on Explorer 34 [24], GEOS 1 [2], and ISEE-A [27] have high sensitivities and large energy ranges but limited resolution ((M/Z)/ Delta(M/Z)=4). The latter instruments will nevertheless permit identification of several individual ions (H+, 4He+2 , 16O+6 , and perhaps 3He+2) and groups of ions such as the Fe group in the solar wind. The ISEE-C spacecraft will be continuously in the solar wind and thus affords an opportunity for a comprehensive uninterrupted composition study. Table I shows that a mass spectrometer with an M/Z resolution of ~30 and a dynamical range of ~10^4 can obtain the abundances of about one dozen different ions (marked by the asterisk in Table 1). With a con- tinuous record of such a large number of ions, fundamental problems concerning the origin of the solar wind and solar composition can be investigated. Some of these are discussed below. 1) A number of momentum transfer mechanisms for the acceleration of heavy ions into the solar wind have been proposed. Under the inferred conditions in the solar wind source region, coulomb collisions would be sufficient to TABLE I ION FLUXES IN THE SOLAR WIND AT I AU __________________________________________________________________________ Ion M/Z Flux (in cm^-2sec^-1) * 3He+2 1.50 4.3 x 10^3 * 4He+2 2.00 1.0 x 10^7 12C+6 7.5 x 10^4 14N+7 7.0 x 10^3 * 160+7 2.29 3.0 x 10^4 * 14N+6 2.33 1.6 x 10^4 * 28Si+12 9.5 x 10^2 * 12C+5 2.40 1.8 x 10^4 * 24Mg+10 7.2 x 10^3 2ONe+8 2.50 1.9 x 10^4 28Si+ll 2.54 1.3 x 10^3 24Mg+9 2.67 8.3 x 10^2 * 16O+6 9.9 x 10^4 14N+5 2.80 5.8 x 10^3 28Si+10 2.2 x 10^3 32S+ll 2.90 4.6 x 10^2 * 28Si+9 3.11 2.5 x 10^3 * 32S+10 3.20 1.1 x 10^3 * 28Si+8 3.50 7.9 x 10^2 * 32S+9 3.56 1.1 x 10^3 * 32S+8 4.00 4.6 x 10^2 * 36Ar+9 3.5 x 10^2 * 40Ca+10 5.4 x 10^2 * 56Fe+13 4.31 1.8 x 10^3 * 56Fe+12 4.67 3.8 x 10^3 * 56Fe+ll 5.09 3.1 x 10^3 * 56Fe+10 5.60 1.7 x 10^3 The elemental abundances are based on measured solar abundances with the exception of the noble gases where measured solar wind abundances were adopted. The charge states are calculated for a coronal equilibrium temperature of 1.6 x 10^6 K (Jordan, [21]). The following selection was made: only the major isotopes of an element (except for He) were included; because of low abundances relative to their neighbors, odd elememts other than nitrogen were omitted; charge states were included only if they contribute 10 percent of more of the element abundance; no element with abundance less than Ar and Ca were included. transfer momentum from protons to heavy ions with high Z^2/M ratios [13]. Momentum transfer by waves has also been proposed as an efficient mechanism [10], [11], [17]. To assess the relative importance of the various mechanisms, measurements of solar wind ions over a large range of M and Z under varying conditions in the corona and solar wind are necessary. 2) Bame et al. [6] have recently proposed that the He/H ratio of 0.04-0.05 typically found in fast solar wind streamers be equated with the He/H ratio in the outer convective zone of the sun. This is a factor of two lower than the generally ac- cepted value and, if correct, would reduce the calculated boron neutrino flux [7], [19]. A comprehensive investigation of the ion abundance variations should lead to a better under- standing of ion fractionation processes in the solar corona, and in turn should yield accurate estimates of the He/H ratio. 3) A correlation between the He/H ratio and solar wind speed has been observed [25], but the mechanism is unknown. By extending correlation studies to other ions it may be possible to uncover the mechanism. 4) Local temperatures and temperature gradients in the corona can be estimated from measurement of the charge distribution of Fe ions and the charge states of O and Si [4],[18]. To test the validity of the various solar wind expansion models, the experimentally derived temperatures and gradients can be compared to those derived from the models. 5) An average 4He/3He ratio of 2350 +/- 120 has been derived from the five Apollo solar wind foil experiments [14], however, the ratio appears to be highly variable [3]. In one instance, a ratio of -500 was observed over a two-day period [15]. With the mass spectrometer on ISEE-C, 3He will be continuously monitored, permitting an accurate determination of the solar surface abundance of this cosmologically important nucleus. The relation between anomalous 3He abundances in the solar wind and the small 4He/3He ratios in some solar flares can also be studied [1]. II. INSTRUMENTATION A. General Design The basic requirements of the instrument were that it be capable of measuring ionic abundances over the M/Z range of 1.5-5.6 (3He+2 - 56Fe+10) at velocities from 300 to 600 km/s. To keep the dynamic range near 10^4, H+ is not measured. The geometric factor g of the instrument is related to the ion flux F and count rate N by g = N/F. (1) In order that N be at least one count per second for the least abundant ion species under consideration, g must be of order 10^-3 cm^2. The dimensions of the apertures and angular acceptances in the plane of the spacecraft rotation and perpendicular to it determine g. The combination of a Wien filter and an electrostatic energy analyzer is a well-known means for determining the velocity and the energy per charge and, hence, the mass per charge ofions. Low resolution mass spectrometers of this type have been flown in the solar wind [24] and in the inner magnetosphere [28]. Another type of low resolution mass spectrom- eter [2] is operating on GEOS I and on ISEE-A [27]. These instruments were primarily designed for the lower magnetosphere, where a low mass-resolution is adequate, but a high geometric factor and an energy range are required. B. Stigmatic Wien Filter It has been shown [20], [23] that the Wien filter can be made to focus along the direction of the magnetic field as well as perpendicular to it by using cylindrical electrical plates as shown in ifigure1. When this is done, the focusing length will be longer than for the parallel plate filter by a factor of 2^1/2; however, the height of the filter can be reduced by up to a factor of 2. The result is a filter with volume and approximate weight 2^-1/2 that of a conventional filter of comparable resolution and geometric factor. 1) Ion Optics: ifigure1 shows the coordinate system for the Wien filter. The equations of motions [23] for positive ions X'' = -omega_o*V_o*[1-(X_o/a)+c((X^2/a^2)-(Y^2/a^2))] (2) Y'' = -omega_o*V_o*[(Y/a)-2c(XY/a^2)] (3) Z'' = -omega_o*X' (4) where omega_o is the cyclotron frequency ZeB/m, V_o is the velocity to which the filter is tuned (equal to the ratio of electric field on the axis to magnetic field Eo/B), a is the radius of curvature of the zero equipotential, and c is related to the change of the radius of curvature of the equipotentials along the x axis. For concentric cylinders c=+1. These equations are solved subject to the following initial conditions: x(t=0) = X_o X'(0) = Vo[1+(delta(V)/V_o)]alpha*Cos(psi) (5) y(0) = Y_o Y'(0) = Vo[1+(delta(V)/V_o)]alpha*Sin(psi) (6) z(0) = 0 Z'(0) = Vo[1+(delta(V)/V_o)] (7) r and r' define the position and velocity of an ion at the entrance aperture, and delta(v) is the difference between the velocity of an ion and the velocity to which the Wien filter is tuned. Alpha is the polar angle and psi is the azimuthal angle with respect to the (x,z) plane as shown in ifigure1. To first order the solutions are X = X_o+[((X_o/a)+(delta(V)/V_o))/((omega_o/Vo)-(1/a))] *(1-Cos((omega_x*t))+(V_o/w)alpha*Cos(psi)*Sin(wt) (8) Y = Y_o*Cos((omega_y*t)+((V_o/(omega_y*alpha *Sin(psi)*Sin(omega_y*t)) (9) Z = V_o*t (10) Where omega_x = omega_o * [1-(V_o/(omega_o*a))]^(1/2) omega_y = omega_o * [(V_o)/(omega_o*a)]^(1/2) The above solution holds for the case (omega_x)^2 > 0. For the region (omega_x)^2 < 0, the sine and cosine terms in (8) are replaced by hyperbolic functions. It should be noted that c does not appear in the first-order solutions. 2) Transmission Properties: We define the transmission T as T = (particles passing exit aperture)/(particles passing entrance aperture) From (9), it can be seen that for alpha * sin(psi) = 0, an ion entering the filter at a distance yo above the axis will exit the filter at a distance y, where y < yo. Thus for alpha small, there is no loss of transmission in they direction. Applying (8), the overall transmission of the Wien filter for delta(v) = 0, alpha = 0, and equal entrance and exit aperture widths is T = [1-(V_o/(omega_o*a)] / [1 - (V_o/(omega_o*a) * Cos(omega_x)*L/Vo)] where L is the length of the Wien filter. This transmission is plotted in ifigure2 as a function of VM/Z for the Wien filter parameters listed in Table II, where V = ion velocity in km/s, M = mass in amu, and Z = charge in electron charges. 3) Velocity Resolution: Exposing the Wien filter to a broad, monoenergetic, and collimated beam, and changing the tuned velocity v_o will produce a trapezoidal response function for particles passing the exit slit. The width (FWHM) is given by the expression (delta(v)/V_o) = (s/a) * [[(a*omega_o)/Vo) - Cos(omega_x*L/Vo)] /(1-Cos(omega_x*L/Vo)] where s is the aperture width. TABLE 11 INSTRUMENT PARAMETERS ___________________________________________________________________________ Wien Filter (Stigmatic): Working Gap Length 180mm Working Gap Cross-Section 11mm x 33mm Average Magnetic Field 0.1454 W/m2 Zero equipotential Radius of Curvature in (x,y)-plane 80.8 mm Entrance and Exit Aperture Dimensions 2mm x 15mm Transmission 26-50% Resolution (Vo/delta(v)) 20-30 Angular acceptance in ecliptic plane 0.5 - 3.6 deg. Angular acceptance vertical to ecliptic plane 3 - 12 deg. Weight 2.7 kg Electrostatic Energy Analyzer (spherical) Radius of zero equipotential 50.0mm Radius of outer plate 54.7mm Radius of inner plate 46.1mm Sector Angle 167 deg. Entrance and exit aperture dimensions 2mm x 15mm Resolution (E/delta(E)) 50 Weight 0.178 kg Mass Spectrometer (Filter and Analyzer) Range 3He - Fe Dynamic range 10^4 - 10^5 Overall gemtric factor (eq. 1 and 15) (4 x 10^-4)-(1.7 x 10^-3)cm^2 Weight of sensor (without electronics) 4.28 kg Total weight of instrument 5.6 kg In ifigure2, (delta(V)/V_O)1/2 is plotted as a function of VM/Z. For cos (omega_x*L/v_o)=1 (cf. equation (12)), there is a pole at VM/Z=400 km/s. The parameters of the Wien filter were chosen in such a way that the pole does not appear in the VM/Z region of interest. 4) Angular Acceptance: From (9), it can readily be seen that the angular acceptance in the (y, z) plane is (delta(alpha)) = (h/2) / [ (V_o / omega_y *sin(omega_y*L/Vo)] (13) where h is the height of the apertures. A somewhat more com- plicated expression describes the angular acceptance in the x-z plane (ecliptic plane). Both are plotted in ifigure3 as functions of VM/Z for the parameters given in Table II. The radius of curvature of the zero equipotential a was chosen equal to (2^1/2)*L/pi so that angular focusing occurs for v_o/omega_o = a/2 in both planes. The parameter c appearing in (2) and (3) has been chosen to be -0.5, which produces a slightly elongated trace in the y direction [23]. C Electrostatic Energy Analyzer The spherical electrostatic energy analyzer is of conventional design [22]. Care was taken to insure that the height and spacing of the plates were sufficient to accommodate the maximum angular divergence of entering ions; the analyzer constant is 5.86 eV/V, and its energy resolution is 50. With this design, the energy analyzer will transmit virtually all ions at the tuned energy independent of the angle with which they leave the Wien filter and enter the analyzer. It was thus as- sumed that the transmission function for the tuned analyzer is unity. D. Instrument Description Table II gives a detailed description of the Wien filter and energy analyzer parameters. An approximation to the geo- metric factor is g = s * h[(delta(alpha)/180] * T (14) where s and h are the width and height of the Wien filter en- trance aperture, (delta(alpha)) is the acceptance angle in the (x-z) plane, and T is the transmission. Using the data from Table II and Figs. 2 and 3, at a speed of 450 km/s, g is 1.7 x 10^-3 CM^2 for 3He and 4 X 10^-4 CM^2 for the Fe ions, results which are close to the design goal. To achieve the theoretical resolution,the magnetic field in the working gap of the Wien filter was made homogeneous to better than +/- l percent by the use of shaped pole pieces. The magnets used are sintered samarium cobalt, and a contoured yoke of vanadium pennendur provides a flux path for the internal magnetic field and serves as the principal structural element of the filter. To reduce the joint area (which contributes to flux leakage), the yoke was fabricated with only one open end. A shield of high permeability material further attenuates the external magnetic field to acceptable levels. The electric field in the Wien filter is produced by plates of gold plated aluminum with shims on top and bottom of each to correct for fringing fields. The plates of the electrostatic analyzer were roughened and plated with copper sulfide black [2] in order to reduce ion scattering and UV reflectivity. A grounded high-transmission grid covers the exit aperture to reduce field distortion inside the analyzer. Ions leaving the exit aperture of the analyzer are detected by three identical channeltron electron multipliers (CEM's) mounted on ceramic circuit boards which also contain pre- amplifiers and discriminators. The multipliers are operated in the saturated pulse mode and the discriminators are set at a level corresponding to approximately 10^-13 C of charge per pulse. A grid, maintained at the potential of the CEM input, is placed directly in front of the detectors to increase counting efficiency by ensuring that all secondary electrons are collected. The relative count rates in each of the detectors depend on the angular distribution of the ions entering the filter and can be used to obtain the direction of the bulk motion of each species and some temperature information. Counts are only registered during 12 percent of the spin period of 3 s, giving a total background count rate of 0.02 per spin period. By comparison, the count rate for each of the Fe charge states (flux, 3 X 10^3 cm^-2 s^-1; geometric factor, 4 X 10^-4 cm^2) is estimated to be 4 counts per spin period. E Instrument Operation The input voltages to the power supplies which provide the potential to the electric field plates of the Wien filter and energy analyzer are obtained from 12-bit digital-to-analog converters. The digital inputs to the converters come from the microcomputer-based processor. ifigure4 shows the instrument and an electrical block diagram. The full parameter range covered by the experiment is 300-600 km/s in velocity, 840-11720 eV/Z in energy per charge, and 1.5-5.6 in mass per charge. The velocity range is divided into n steps and the energy per charge range into m steps, with both being logarithmically related. V_1s(n,m) = [R_1^(m-1)] * [R_2^2(n-1)] * V_1(1,1), energy analyzer (15) V_2s(n) = [R_2^(n-1)] * V_2(1), Wien Filter 1 <= n <= 24, 1 <= m <= 40, where V_1s and V_2s are the digital values of the voltages applied to the inputs of the power supplies, V_1(1,1) and V_2(1) are the voltages for the velocity step n=1 and the mass step m=1, and R_1, and R_2 are constants. There are three modes of operation which are selectable by ground command. In mode 1, m is varied from 1 to 40. For each value of m, n is varied over its complete range. This mode is useful for surveying the solar wind. In mode 2, minimum and maximum values of m and n are chosen, so that only a part of the full parameter range is scanned; this mode will be for particular M/Z ratio determinations. Mode 3 is a search mode. The value of n corresponding to the velocity of the solar wind is determined by varying n from 1 to 24 while holding m constant at the value corresponding to that for 4He++ until the value of n corresponding to the maximum counting rate is found. A scan is then made from n = n (solar wind) - 3 to n = n (solar wind) + 3 varying m between specified minimum and maximum values. This mode is intended for normal operations, as it yields the most data per unit time. In general, after each spin of the spacecraft a step is taken, but provisions have been made for holding a step for s = 2, 4, or 8 rotations of the spacecraft. The telemetry contains details of the mode, specified by the values of V_1(l, 1), V_2(l), m_min, m_max, n_min, n_max, s, R_1, R_2, and the counts from each of the detectors (separately logarithmically compressed above 127) for the corresponding values of m and n. At the minimum time resolution, a spectrum is obtained every 24 X 40 X 3 s (spin period = 3 s), or 48 min. Use of the search mode reduces this to approximately 12 X 3 + 6 X 40 X 3 s = 12 min, 36 s. III. CALIBRATION AND PERFORMANCE The sensor was calibrated at the test facility of the Physikalisches Institut of the University of Bern. This facility consists of a vacuum chamber in which a homogeneous monoenergetic mass separated beam of ions can be produced and directed at the sensor mounted on a test platform which can be rotated about two axes perpendicular to the direction of the ion beam [2]. Sweep circuits were used to synchronously change the voltages on the plates of the filter and analyzer while the outputs of each of the three CEM's strobed a multi-channel analyzer operating in the pulse-height-analysis mode. Calibrations were performed with H2+ and He+ at nominal velocities of 300, 400, 500, and 600 km/s. The orientation of the sensor with respect to the incident ion beam was changed in small increments over a range of +/- 8 degrees in the horizontal plane and +/- 14 degrees in the vertical plane. Angular acceptances derived from these functions are in good agreement with theory, whereas the experimental resolution always exceeds that predicted by (12).Integration of the contour plots over the angle in the x-z plane yielded graphs like the one shown in ifigure5. The absolute values of the transmission determined at normal incidence for H2+ and 4He+ at the four different values of ion velocity are in excellent agreement with theory. The ability of the instrument to reject stray ions which create ghost peaks by scattering from the outside plates of the analyzer and contribute to the background was measured. In all cases, the ratio of ghost peak counts to counts in the main peak was less than 10^-4 with the ghost peaks always appearing at energy values 7 percent less than those of the corresponding main peaks. Because of the large difference in M/Z between 4He and 3 He, interference of 4He+2 at the position of the 3He+2 peak is less than 10^-5 of the 4He signal. A typical mass spectrum obtained during calibration is shown in ifigure6. ACKNOWLEDGMENT The authors wish to acknowledge valuable assistance in the design and construction of the sensor and processor by F. Hunsaker, C. Moyer, J. Sites, T. Skillman, S. Weinberger, and G. Thompson. During the calibration at the Physikalisches Institut, Dr. H. P. Walker and R. Banninger contributed important technical assistance. REFERENCES [1] J. P. Anglin, W. F. Dietrich, and J. A. Simpson, "Super enrich- ments of Fe-group nuclei in solar flares and their association with large 3He enrichments," in Proc. 15th Int. Cosmic Ray Conf., vol. 5, Plovdiv, Bulgaria, p. 43, 1977. [2] H. Balsiger, P. Eberhardt, J. Geiss, A. Ghielmetti, H. P. Walker, and D. T. Young, "A satellite-borne ion mass spectrometer for the energy range 0-16 keV," Space Sci. Instrum., vol. 2, p. 499, 1976. [3] S. J. Bame, A. J. Hundhausen, J. R. Asbridge, and I. B. Strong, "Solar wind ion composition," Phys. Rev. Lett., vol. 20, p. 393, 1968. [4] S. J. Bame, J. R. Asbridge, W. C. Feldman, and P. D. Kearny, "The quiet corona: temperature and temperature gradient," Sol. Phys., vol. 35, p. 137, 1974. [5] S. J. Bame, J. R. Asbridge, W. C. Feldman, M. D. Montgomery, and P. E. Kearny, "Solar wind heavy ion abundances," Sol. Phys., vol. 43, p. 463, 1975. [6] S. J. Bame, J. R. Asbridge, W. C. Feldman, and J. T. Gosling, "Evidence for a structure-free state at high solar wind speeds," J. Geophys. Res., vol. 82, p. 1487, 1977. [7] S. P. Bhavsar and R. Harm, "The neutrino flux of inhomogeneous solar models," Astrophys. J., vol. 216, p. 138, 1977. [8] D. J. Faulkner, G. S. Da Costa, and A. J. R. Prentice, "Hydrogen- helium inhomogeneities and the solar neutrino problem," Mon. Notic. Roy. Astron. Soc., vol. 170, p. 589, 1975. [9] W. C. Feldman, J. R. Asbridge, S. J. Bame, and J. T. Gosling, "Solar wind He abundance variations during solar cycle 20," (A) EOS, vol. 58, p. 1222, 1977. [10] U. FeUer, "Stochastische beschieunigung schwerer ionen im sonnenwind," Ph.D. dissertation, Univ. of Bern, Bern, Switzer- land, 1978. [11] J. Geiss, "Solar wind composition and implications about the his- tory of the solar system," in Proc. 13th Int. Cosmic Ray Conf, vol. 5, p. 3375, 1973. [12] J. Geiss, P. Eberhardt, F. Buhler, J. Meister, and P. Signer, "Apollo 11 and 12 solar wind composition experiments: Fluxes of He and Ne isotopes," J Geophys. Res., vol. 75, p. S972, 1970. [13] J. Geiss, P. Hirt, and H. Leutwyler, "On acceleration and motion of ions in corona and solar wind," Sol. Phys., vol. 12, p. 458, 1970. [14] J. Geiss, F. Buhler, H. Cerutti, P. Eberhardt, and C. Filleux, "Solar wind composition experiment sec," 14, Apollo 16 Prel. Sci. Rep. NASA SP-315, p. 14, 1972. [15] H. Grunwaldt, "Solar wind composition from the HEOS-2 plasma experiment," Space Res., vol. 16, p. 681, 1976. [16] J. Hirshberg, "Composition of the solar wind: Present and past," Rev. Geophys. Space Phys., vol. 13, no. 3, p. 1059, 1975. [17] J. V. Hollweg, "Alfvenic accelerating of solar wind helium and related phenomena," J. Geophys. Res., vol. 79, p. 1357, 1974. [18] A. J. Hundhausen, H. E. Gilbert, and S. J. Bame, "Ionization state of the interplanetary plasma," J. Geophys. Res., vol. 73, p.5485,1968. [19] I. lben, "The 37 Cl solar neutrino experiment and the solar helium abundance," Am. Phys., vol. 54, p. 164, 1969. [20] D. Ioanoviciu, "Ion optics of a Wien filter with inhomogeneous fields," Int. J. Mass Spectrom. Ion Phys., vol. 11, p. 169, 1973. [21] C. Jordan, "The ionization equilibrium of elements between car- bon and nickel," Mon. Notic. Roy. Astron. Soc., voL 142, p. 501, 1969. [22] C. E. Kuyatt and J. A. Simpson, "Electron monochromator de- sign," Rev. Sci. Instr., vol. 38, p. 103, 1967. [23] W. Legler, "Ein modifiziertes Wiensches Filter also Elektronen- monochromator," Z Phys., vol. 171, p. 424, 1963. [24] K. W. Ogilvie, T. McIlrath, and T. D. Wilkerson, "A mass energy spectrometer for space plasmas," Rev. Sci. Instr., vol. 39, p. 441, 1968. [25] K. W. Ogilvie and J. Hirshberg, "The solar cycle variation of the solar wind helium abundance," J. Geophys. Res., vol. 79, p. 4959, 1974. [26] D. E. Robbins, A. J. Hundhausen, and S. J. Bame, "Helium in the solar wind," J Geophys. Res., vol. 75, p. 1178, 1970. [27] E. G. Shelley, R. D. Sharp. R. G. Johnson, J. Geiss, P. Eberhardt, H. Balsiger, G. Haerendel, and H. Rosendauer, "Plasma Composition Experiment on ISEE-A," this issue, pp. 266-270. [28] E. G. Shelley, R. G. Johnson, and R. D. Sharp, "Satellite observa- tions of energetic heavy ions during a geomagnetic storm," J. Geophys. Res., voL 77, p. 6104,1972. [29] R. V. Wagoner, "Big-Bang nucleosynthesis revisited,"Astrophys. J., vol. 179, p. 343, 1973.