218) My dilemma -- to do or not to do it?

Ludwik Kowalski (5/6/05)
Department of Mathematical Sciences
Montclair State University, Upper Montclair, NJ, 07043

This morning I receive a message from a stranger, Walter. He wrote:

> "Oriani counted very dark tracks only" It is important to know
> whether "very dark" tracks were counted in a double-blind
> manner. Otherwise the experimenter can unconsciously --
> and innocently! -- bias the results towards what was expected.
> I.e., "very dark" might actually have to be darker for the control
> experiments. Best if the counting can be done automatically,
> e.g., by image-processing software.

I asked for the advice about the image-processing software but Walter said that it was only a general comment and that he never used such software. But his comment is highly appropriate for a situation in which the signal to noise ratio is as small as described in unit #210. The only way to decide if the signal is real (that is to decide if alpha-like particles are really emitted from the cathode) is to use photos, as described by Oriani at ICCF10. Walter’s message reminded me that I wanted to use this method; here is my reply to him:

> After the first etching I will take numerous photos of the entire
> surface. This will identify “old” tracks not by their sizes or darkness
> but by their positions (with respect to recognizable scratched lines).
>The same will be done after the second etching. “New” tracks will
> thus be identified unambiguously by comparing the photos “before”
> with the photos “after.”

Should I start this work right now or should I wait? I am hesitating because I know how much work is involved in photographing ten surfaces. At the minimum magnification of my microscope each photo covers the area of only 1.3 mm by 0.9 mm. Each of my CR-39 chips has the area of about 1 cm2. That implies about ~100 pictures per surface. With ten surfaces (5 for control chips and 5 for treatments) I must study about 1000 pictures. That is not trivial. Assuming I use the same exposure times (24 hours), and the same counting geometry (chips are ~5 mm away from the cathode) as Richard, I should expect, on the average, 33.6 alpha tracks per surface. This translates into about one track per 30 photos.

The purpose was to produce a simple method by which students and teachers can convince themselves about reality of nuclear particles due to a chemical process. To be useful the experiment should be feasible in three to four hours. Suppose that the rate of emission is ten times larger than what was reported by Oriani. In that case the expected number of tracks would 336 per surface. No photography, and no detailed statistical analysis, would be necessary to see that the effect is real. At the minimum magnification of my microscope the field of view has the diameter of about 5 mm. This would allow to count tracks at the rate of several minutes per surface. But now Oriani is in a situation in which photography is a must, as far as track detectors are concerned. This is not for students and teachers. In a week or two Oriani, and a nuclear physicist from the University of Minnesota, will start exploring the phenomenon with a silicon detector. I should wait for their preliminary results before deciding what to do next. The silicon detector, if the background can be lowered, as in Jones’ laboratory, seems to be (?) more appropriate than CR-39 for very low emission rates. But lowering the background might not be trivial; it calls for a set of good veto detectors.