EMISSION CURRENT DEPENDENCE IN THE TIME AND ANGLE NOISE SPECTRA OF GALLIUM LIQUID METAL ION SOURCES

The frequency and i n v a angular power spectra of the emission of ions from a gallium liquid metal ion source has been studied as a function of emission current using an ion streak camera. The camera records the angle and time of arrival of individual ions with a chevron multichannel plate detector. Results indicate that the previously reported decrease in fluctuation power with increasing angular wave number is present at all emission currents between 5 and 75 PA and becomes only slightly more pronounced with increasing source current. There is also a distinct rolloff in fluctuation power with increasing frequency above 200 Mhz, the slope of this rolloff increasing slightly with increasing source current. The total power in the fluctuations is roughly proportional to the axial current density from the source. At emission cumnts greater that 100 PA the emission of individual droplets is evident as "shadows" in the angle-time images. The shapes of these shadows are influenced both by droplet size and charge. Introduction For several years it has been known that liquid metal ion sources (LMIS) produce a wide range of particle masses and charges. B varying source parameters, everythiig from atomic ions['] to ionized clusters[31 [41 1Y and large droplets[51 l6 have been observed. These phenomena cannot be completely described using a simple stationary Taylor cone modelr7]. The fact that a liquid is in motion is strongly suggestive that a more dynamic approach to modeling the behavior of these sources is needed, particularly at high emission currents. ~ a g u e r [ ~ ] models the emission by totally hydrodynamic consuaints, and explains droplet creation as a byproduct of stress-indud cavitation of the liquid metal at the emitting site. Swanson and ICinghamr91 have more recently developed a model for emission based on a fluid cusp which uses both the hydrodynamic 'prgperties of the LMIS and the space charge produced by the ions to limit the current produced by these sources. When an LMIS emits an ionized cluster of atoms and even more dramatically when it emits a liquid droplet, it is anticipated that the dominant emission of atomic ions will be momentarily disrupted by the geometrical perturbation of the liquid tip and of the electric field acting on it. If this picture is c m t , some non-random time varying behavior is anticipated in the emission of ions (or other species) from the source. This dynamic behavior may be characterized by the reformation time of the emitting region after such a perturbation. In order to clarify the importance of hydrodynamic and space charge effects in the emission process, we have examined the time varying emission properties of these sources to supplement what is already know about their time averaged behavior. This work extends our previously reported results['01 of frequency and inverse angle correlations in the LMIS both by including more detailed results and by including source emission currents from 5 to 125 PA. Apparatus The experimental apparatus is illustrated in figure 1. Atomic ion emission (and emission of clusters and droplets) is produced by applying a bias voltage of several kilovolts between the liquid metal ion source tip (la) and an extractor (lb). From the cylindrically symmetric pattern of total emission, a narrow slit (lc) selects a fan-shaped portion which is 1 mrad in the vertical direction by 140 mrad in the horizontal. Deflection plates (Id) then deflect the beam across a chevron microchannel plate (MCP) particle multiplier (If). The deflection time is adjustable and has been either 100 nsec or 4 vsec in the results reported here. The individual ion hits on the MCP are electron multiplied and appear as individual flashes on the output phosphor screen (lg). A CCD video camera (lh) records the image. The recorded image is then digitized Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1988630 C6-178 JOURNAL DE PHYSIQUE Figure 1. Mechanical layout of the streak camera showing the path of the ion beam through the instrument. Shown are the LMIS (a), extractor (b), difining slit (c), deflection plates (d), limiting baffle (e), channelplate (0, phosphor screen (g), and camera (h). and stored for later processing. The camera records all events that have occurred withiin the previous 1/60 sec and the fan of beam spends almost all of that time at one extreme or the other of its &flection. In order to decrease the background of ions striking the MCP during the long dwell time at the extremes, the gain of the MCP is gated to be on only slightly longer than the sweep time across the.plate. A baffle (le) also decreases significantly the number of spurious events that the MCP "sees" from the dwell time extremes. An image of a single sweep is shown in figure 2. The sweep time for. this image was 100 nsec. The digitized image is 128 columns (angle bins) by 64 rows (time bins). The actual resolution was 1 mrad in the angle axis (horizontal) by 1 nsec in time (vertical). The time and angle resolution are somewhat higher than 1/64 of 100 nsec and 11128 of 140 mrad because the beam traverses more than the vertical 0pening.h the baffle in 100 nsec and the total beam is slightly wider than the baffie opening's width. Then is no mass selection in the streak camera, but because the number of atomic ions exceed the total of all other charged particles by more than a factor of IOO[~], the image is formed by the atomic ions. Figure 2. An image produced by the streak camera at a sweep time of 100 nsec. The source current is 5 PA.