Generation of Intense Negative Ion Beams

This invention relates to a method and apparatus for the generation of atomic and molecular negative ion beams, either pulsed or continuous beams.

The production of beams of atomic and molecular negative ions is of considerable interest in diverse areas of atomic, molecular, and plasma physics. Such beams are required for fusion plasma heating, heavy ion inertial-confinement fusion, and in basic atomic and molecular scattering studies. The types of sources for ion production are many and varied, involving both plasma and surface-plasma interactions.

In accordance with the present invention, an ion source utilizes a beam of electrons and target molecules. The source includes an electrode which reverses the electron beam, producing electrons at their turning point having a distribution of velocities centered at zero velocity. A gas which attaches zero velocity electrons or some near-zero velocity electrons is introduced at or near this turning point. Negative ions of the gas are produced by an attachment, or a dissociative attachment, process. For many of the thermal electron-attaching molecules, the attachment cross section at zero electron energy can be quite large, varying as (electron energy).sup.-1/2, or just the s-wave threshold law.

The novel features of the invention are set forth with particularity in the appended claims. The invention will best be understood from the following description when read in conjunction with the drawings.

FIG. 1 is a schematic diagram of the apparatus and method for generation of intense beams of negative ions by zero-energy electron attachment, or dissociative attachment, in accordance with the present invention.

FIG. 2 illustrates the grid pulsing sequence for pulsed operation of the source shown in FIG. 1.

FIG. 3 illustrates the relative Cl.sup.- yields for continuous operation of the source shown in FIG. 1 as a function of the ring potential V.sub.R, and at electron energies E.sub.o of 20 eV and 30 eV.

FIG. 4 illustrates the relative Cl.sup.- yields for continuous operation at 40, 50, 60, and 80 eV electron energies.

FIG. 5 illustrates the distribution of Cl.sup.- pulses during several pulsing cycles of the ring electrode voltage V.sub.R.sbsb.max.

In the following description, the cross sections referred to are "attachment cross sections" defined in the usual sense of the probability per molecule that an electron traversing a gas will undergo an attaching collision with the gas molecules.

Many molecules, such as SF.sub.6, CFCl.sub.3, perfluorinated carbon compounds, and chlorocarbon compounds, have extremely large cross sections for attachment of zero-energy electrons to form negative ions such as F.sup.-, Cl.sup.-, Br.sup.-, and I.sup.-. In conventional ion sources, electron energies are too high (of the order 2-3 eV or greater) to attach to these molecules. In the apparatus of the present invention, "zero energy" electrons are generated momentarily by pulsing a reversing electrode R negatively with respect to an electron-gun cathode F,P. In this preferred embodiment, the reversing electrode is comprised of a conductive ring R and a conductive mirror plate M. These electrons can then attach to molecules, such as CFCl.sub.3, which effuse from a 1 mm diameter stainless-steel tube B placed at the center of the electrostatic reversal field. This field is produced by applying appropriate negative potential on the ring R and mirror M. The location of the electron turning point is governed by the magnitude of potentials on the electrodes R and M. Applying larger negative potentials E.sub.R and E.sub.M has the effect of moving the turning point to the left in FIG. 1, and vice versa. The "mirror" effect produced by the ring R and mirror M can be either concave, convex or planar, depending upon the potential ratio. However, the invention is not limited to such an arrangement of these electrodes R and M as the "reversing electrode" for production of zero energy electrons.

The present invention takes advantage of the extremely large, threshold (zero-energy) electron attachment cross sections in several molecules to generate, for example, intense beams of F.sup.-, Cl.sup.-, Br.sup.- and I.sup.- ions. Thus, an ion source is provided in which an electrode or grid, which reverses electron trajectories, is placed just beyond an ion extraction aperture comprised of ion lenses L.sub.1 and L.sub.2. When the reversing electrodes are pulsed negatively, this pulse voltage reverses the direction of electrons momentarily, giving the electrons zero kinetic energy. At this zero-energy turning point the electrons attach to the ambient gas, and generate negative ions as the product of the attachment process. The method described herein as an example is applied to the generation of Cl.sup.- ions from CFCl.sub.3. Other ions from effusing molecules may be formed and extracted through the aperture formed by the ion lenses L.sub.1 and L.sub.2 in the same manner.

Recent experimental evidence has shown that the zero-energy cross section to CFCl.sub.3, SF.sub.6, CCl.sub.4, and other molecules is much larger than previously expected. Thus, the probability of electron attachment during the field-reversal pulse will be much larger, making this apparatus an even more efficient ion source. Generally, the production yield of negative ions is the product of the electron energy distribution function and the dissociative attachment cross section, integrated over all electron energies. The reversing electrode momentarily creates a maximum in the distribution function at the maximum of cross section (zero eV), thus maximizing the integral and therefore the yield of negative ions.

The electrons in the turning point region B will have a peak at about zero eV in their energy distribution. These electrons can then attach to an admixed gas having a peak in its dissociative attachment (DA) cross section at zero eV. Cl.sup.- was generated as an example from CFCl.sub.3 for several reasons: the behavior of the cross section for Cl.sup.- production at zero eV (greater than 10.sup.-13 cm.sup.2) is now well known, and is many orders of magnitude greater than for H.sup.- production from H.sub.2 at higher energies, an effect due to the divergent nature of the s-wave attachment cross section; Cl.sup.- is by far the major ion produced in CFCl.sub.3 at any electron energy; and CFCl.sub.3 is inexpensive and inert.

Other examples of the effect of reversal of electron trajectories is seen in a high-pressure electron swarm and the diffuse discharge plasmas. Electrons in a swarm, for example are backscattered by elastic and inelastic gas collisions. These electrons move against the drift field until their trajectories are reversed, and are accelerated along the field lines again. At their turning point, they have essentially zero kinetic energy, and can be removed from the swarm by attachment to SF.sub.6, I.sub.2, or an admixed perfluorocarbon compound.

The overall process can be described, for molecules AB and electron energy E, as ##STR1## From the Wigner threshold law, the cross section to form the atomic ions Cl.sup.- or F.sup.-, or the corresponding parent negative ions, behaves as E.sup.-1/2 in the limit E.fwdarw.0, thus providing an extremely efficient path for negative ion formation. Use of CFCl.sub.3 had the additional virtue that the cross section for Cl.sup.- production greatly exceeded that for production of other possible ions (F.sup.-, Cl.sub.2.sup.-, and CCl.sub.3.sup.-) at energies below 3 eV. This situation arises in other molecular targets as well, and makes for an efficient source with minimal interference from other species during extraction and acceleration in an ion beam transport system.

The theoretical calculation of Henkelman and Ottensmeyer, J. Phys. E7, 176 (1974), is used to fix the placements of the C.sub.3, M, and R electrodes, and to determine the aperture diameters in C.sub.3 and R. The distance between C.sub.3 and R was 9.8 mm in our experiments.

Assuming the origin as the center of the aperture in C.sub.3 and neglecting aberrations in the reversal field, the longitudinal energy E.sub.y and transverse energy E.sub.x are

where .epsilon. is the electric field intensity, and E.sub.0x and E.sub.0y (E.sub.0 =E.sub.0x +E.sub.0y) refer to launched values of electron energy at the origin. The total energy at any location is E=E.sub.x +E.sub.y and diminishes with increasing electron penetration into the field. The mean transverse energy E.sub.x is E.sub.0 sin.sup.2 .theta., where the .theta. is the mean divergence angle in the electron gun. The particular case chosen in E.sub.y =0 located at y=.vertline.E.sub.0y /e.epsilon..vertline., although one may choose other energies as well.

The grids G.sub.1 and G.sub.2 serve to extract out of the reversal field the negative ions (here Cl.sup.-) generated. The bias on G.sub.1 is positive, and that on G.sub.2 negative relative to ground. The extracted ions are accelerated by several lenses formed by electrodes L.sub.1 and L.sub.2, and focused at the entrance aperture of a utilization device (here a quadrupole mass filter 1 tuned to the mass of the particular negative-ion fragment in the experiments). After mass filtering the ions are accelerated to 1 keV into a channel-type electron multiplier 2. The output of the multiplier is amplified, and stored by multichannel analyzer 3 as a function of either sweep voltage on the ring element R, or as a function of incident electron energy E.sub.0. The base pressure of the vacuum system (about 2.times.10.sup.-5 Pa) rose to a background pressure of 2.times.10.sup.-3 Pa during operation.

The extraction of Cl.sup.- ions may be carried out in either a continuous or pulsed mode. In the case of continuous operation, steady-state potentials are applied to the electron gun element C.sub.1 and extracting electrodes G.sub.1,G.sub.2. In practice there is negligible penetration of the G.sub.1, G.sub.2 fields into the region B so that ions extracted in this way are very likely only a small part of the total ion production in the turning point region.

In pulsed operation ions are formed by applying the appropriate potentials to elements C.sub.1, R and M to turn the ion beam on, and reverse it in the region B, for periodic intervals, and then applying the appropriate potential at a grid G.sub.1 for extraction of the negative ions while the electron beam and reversal field are pulsed off. The sequence of applied pulses is shown in FIG. 2. Elements G.sub.1 and G.sub.2 are held at ground potential during the ion generation interval. The Cl.sup.- ions are created via a dissociative attachment process e(E.about.0 eV)+AB.fwdarw.A+B.sup.- in this step. The electron gun is then pulsed off by applying a -100 V potential to electrode C.sub.1 and grounding the mirror M and ring R. A positive pulse is applied to grid G.sub.1 to extract the Cl.sup.- ions (while a negative pulse is applied to grid G.sub.2). These ions are extracted through an aperture formed by ion lenses L.sub.1 and L.sub.2, and mass analyzed. For these experiments, the ion pulses from the channel-type detector were routed to the multi-channel analyzer where the ion signal is stored as a function of E.sub.0, or the ring and mirror voltages, as part of an experimental test of the invention.

We have measured the negative ion yield as a function of ring potential at incident electron energies E.sub.0 of 20, 30, 40, 50, 60 and 80 eV. In the case of continuous operation the mirror potential V.sub.m at each energy was just -E.sub.0 /e (where e is the magnitude of electron charge) and the potential V.sub.R on R was obtained from the calculations of Henkelman and Ottensmeyer, supra for the ratio V.sub.R /V.sub.M.

Results of Cl.sup.- signal vs V.sub.R for continuous-mode operation are shown in FIGS. 3 and 4 at the indicated E.sub.0. It is clear from these spectra that, as expected, increasingly greater ring voltages V.sub.R are required to reverse increasingly energetic electron (higher E.sub.0). Also, the highest signal count rates and narrowest widths were encountered as E.sub.0 =30 eV. While this effect of width is not clearly understood, several effects which could give rise to this behavior with E.sub.0 are: optimum electron-molecule spatial overlap (e.g., focusing at region B the smaller "disk of least confusion" rather than the Gaussian image), a minimal transverse energy E.sub.x, thus bringing more electrons into the peak of the attachment cross section at zero electron energy (see Eq. 2), minimal aberrations in the reversal field, or optimal extraction and focusing efficiency by G.sub.1, L.sub.1, and L.sub.2.

All count rates in FIGS. 3 and 4 are given relative to the 30 eV count rate, so that relative efficiencies at the different energies can be compared. In the following table we list at each energy the peak counting rate, integral counting rate, and the mirror ratio V.sub.R.sbsb.max /V.sub.M corresponding to the peak signal.

For pulsed operation of the source described hereinbefore with reference to FIG. 2 there is shown in FIG. 5 an oscilloscope image of the distribution of Cl.sup.- pulses during several cycles of the ring voltage V.sub.R.sbsb.max. The electron energy is 40 eV and the value V.sub.R.sbsb.max /V.sub.M is 0.75. The maximum yield and peak-integrated yield for this pulsed operation is found to be about ten times greater than for continuous operation, even after a 30% duty cycle is taken into account. This increase is almost certainly due to the fact that Cl.sup.- ions are now being extracted from a field-free region, and not from the field of the ring and mirror which tends to drive the ions from the turning point region back toward C.sub.3.

The production rate I of Cl.sup.- ions can be written as

where N(cm.sup.-3) is the CFCl.sub.3 target density, V(cm.sup.3) the overlap volume between the electron and target beams, .phi.(s.sup.-1 cm.sup.-2) the incident electron flux, and .sigma..sub.A (E)(cm.sup.2) the attachment cross section at E.about.0 energy. Estimates of the quantities in Eq. (3) are as follows: A pressure of CFCl.sub.3 in the beam was taken as 0.133 Pa, or N.about.3.3.times.10.sup.-- cm.sup.-3. The interaction volume is taken as that of a cylindrical electron beam of 0.2 cm diameter intersecting a target beam of 0.2 cm diameter, or V.about.6.3.times.10.sup.-3 cm.sup.3. The electron flux is taken as a 1.5 .mu.A electron beam in a cylinder of 0.2 cm diameter or 3.0.times.10.sup.14 s.sup.-1 cm.sup.-2, and an average attachment cross section is taken as 2.times.10.sup.-14 cm.sup.2 for E less than approximately 0.1 eV. Assuming that the entire 1.5 .mu.A incident current is reversed to give a sum of longitudinal and transverse energies E of less than about 0.1 eV, then the production rate I(Cl.sup.-) is 1.2.times.10.sup.12 s.sup.-1, or 0.2 .mu.A. Thus, it appears that 13% of the incident electron beam can be converted to Cl.sup.-. This corresponds to a current density .rho. of .rho..about.0.2 .mu.A/3.times.10.sup.-2 =6.7 .mu.A/cm.sup.2 for emission from a 0.2 cm diam region of the target. This density can be increased by several orders of magnitude through use of (a) more intense electron beams (milliamperes or amperes as opposed to microamperes currently used) with a correspondingly larger beam size, and (b) denser targets.

It should also be possible to extend the use of the source to the generation of negative ions through attachment resonances located at nonzero electron energies, since a continuous distribution of electron energies exists between C.sub.3 and R. While for the present demonstration of the source we have chosen E.sub.y =0, one can choose (see Eq. 2) any other energy E'.sub.y located at the same y, but at a value of

This energy could correspond, for example, to the 3.7 eV.sup.2 .SIGMA..sub.u resonance in H.sub.2 to produce H.sup.-. Moreover, any given energy E'.sub.y is attained twice: once when .epsilon.' and y are parallel (electrons decelerating towards the turning point), and once when antiparallel (electrons accelerating away from the turning point). Thus, the electron beam is "used" twice.