Phase I Rinal Report: Ultra-Low Background Alpha Activity Counter

Abstract

In certain important physics experiments that search for rare-events, such as neutrino or double beta decay detections, it is critical to minimize the number of background events that arise from alpha particle emitted by the natural radioactivity in the materials used to construct the experiment. Similarly, the natural radioactivity in materials used to connect and package silicon microcircuits must also be minimized in order to eliminate ''soft errors'' caused by alpha particles depositing charges within the microcircuits and thereby changing their logic states. For these, and related reasons in the areas of environmental cleanup and nuclear materials tracking, there is a need that is important from commercial, scientific, and national security perspectives to develop an ultra-low background alpha counter that would be capable of measuring materials' alpha particle emissivity at rates well below 0.00001 alpha/cm{sup 2}/hour. This rate, which corresponds to 24 alpha particles per square meter per day, is essentially impossible to achieve with existing commercial instruments because the natural radioactivity of the materials used to construct even the best of these counters produces background rates at the 0.005 alpha/cm{sup 2}/hr level. Our company (XIA) had previously developed an instrument that uses electronic background suppression to operate at the 0.0005 0.005 alpha/cm{sup 2}/hr level. This patented technology sets up an electric field between a large planar sample and a large planar anode, and fills the gap with pure Nitrogen. An alpha particle entering the chamber ionizes the Nitrogen, producing a ''track'' of electrons, which drift to the anode in the electric field. Tracks close to the anode take less than 10 microseconds (us) to be collected, giving a preamplifier signal with a 10 us risetime. Tracks from the sample have to drift across the full anode-sample gap and produce a 35 us risetime signal. By analyzing the preamplifier signals with a digital signal processor we easily distinguish between these two risetimes and thereby count only alpha particles emitted by the sample. Alpha particles emitted from the sample tray are absorbed in the rear of the sample, so the tray's emissivity does not contribute to the background either. Extensions of the method to the counter's sidewalls similarly allow us to reject alpha particles emitted from the sidewalls. We can thus able obtain background rates over a factor of 1000 lower than in conventional instruments without active background rejection. Extending this principle to count at the 0.00001 alpha/cm{sup 2}/hour, level encounters difficulties because there will typically be only 2.4 alpha particles per square meter per day. Since about 6 counts are required to measure activity at the 95% confidence level, large sample areas are required to make measurements in reasonable times. Unfortunately, increasing the counter's anode area to a square meter raises its capacitance so much that the preamplifier noise levels swamp the alpha particle signals and make counting impossible. In this SBIR we worked to solve this dilemma by segmenting the single large area electrode into several smaller, lower capacitance electrodes that could still detect the alpha particles reliably. Each electrode would have its own electronic and we would capture signals from all of them in coincidence (since an alpha track might well deposit charge on more than one electrode), a technique in which XIA is experienced. Therefore, in Phase I we worked to show proof of principle by subdividing our original 1,800 cm{sup 2} electrode into 4 square segments, each 625 cm{sup 2} and demonstrating that signal noise on individual channels reduced as expected. Because the Phase II counter with a 1 m{sup 2} segmented anode would require 16 segments plus a segmented guard as well, we also designed low cost signal processing electronics to instrument it in Phase II. Our Phase I effort met our major proof of principle goals. In particular, reducing the anode size by a factor of 4 in area reduced electronic noise by 3. We also developed an analytical model of signal generation as the charges in the track drift across the counter and showed that the features observed in the real signals closely resembled those predicted by the model. When we captured events where the track charge was indeed collected by two electrodes, we showed that, by summing them, we could recover the preamplifier signal shape appropriate to collection by a single electrode. We thereby showed that we could achieve noise levels that would allow us to analyze our signals with even higher precision than in our current instrument while being able to increase the measured sample area to the values needed to attain the proposed 0.00001 alpha/cm{sup 2}/hour sensitivity.