Neutron Detection Problems "Bubble Fusion"

[DaveC]

Just read a blurb in the Whatsnew (American Physical Society) from Robert L. Park, on a report coming out in Science on Bubble Fusion. The work is being done at Oak Ridge NL. I extracted the relevant part here - the full piece can be read on Whatsnew@aps.org (What's New).

Robert Park writes ..." a report out of Oak Ridge of d-d fusion events in collapsing bubbles formed by
cavitation in deuterated acetone, is scheduled for publication in the March 8 issue of Science magazine. Taleyarkan et al. observe 2.5 MeV neutron peaks, evidence of d-d fusion, correlated with sonoluminescence from collapsing bubbles. Pretty exciting stuff huh? It might be, if the experiment had not been repeated by two experienced nuclear physicists, D. Shapira and M.J. Saltmarsh,
using the same apparatus, except for superior neutron detection equipment. They found no evidence for 2.5 MeV neutron emission correlated with sonoluminescence. Any neutron emission was many
orders of magnitude too small to account for the tritium
production reported by the first group. ... "

There's no information in this piece about the detection equipment being used. Would anyone (Richard Hull??) have more infomation? Especially, what the first group was doing wrong, if anything?

Dave Cooper

[guest]

The neutron deficit might be a problem with how the cavitation pulses are generated. Neutron detection is extremely fussy, any high impulse pulse can fool the detector (BF3 is what most people use). Most sonoluminesence at OleMiss died out because the methods of detection were in error. Estatic promises rendered by early experiment with simple gear pumped up hope to the sky. Dreams of Nobel prizes dancing in their heads. When the experiments were rerun with a shielded setup there were no neutrons! Plenty of uv but the power levels didn't get even get close enough to the fusion temp... missed by 2 orders of magnitude.
A Tig welder puts out plenty of uv but it is about 3 magnitudes of energy below fusion energies.

Larry Leins
Physics Teacher

[ijv]

Heres a link to the story on "bubble fusion"

http://www.eurekalert.org/pub_releases/ ... 030102.php

One thing I noticed in the story was that they were using a neutron source to seed the buuble, could this be leading to some false readings?

IJV

[Richard Hull]

Larry is 100% right on! Most counts received in neutron detectors ain't neuts. The art is in knowing the counter and your experiment. BF3 counters are particularly suseptable to noise, sparks, arcs, pops, booms, temperature gradients and things that go bump in the night!

The DC fusor is a particularly nice device for use with the BF3 as nothing changes. There are no discharges or burst energies associated with a DC operated gas device and besides, you can run the system with other gases and see no neutron production in similar operation as a control test.

Most experiments either can't have such a control or the in-experienced experimenters consider the neutron counter as the Lord God almighty and if it clicks, you just gotta' have neuts! WRONGO!

Some very few of you will remember on the old songs forum how paranoid I was in double and triple blind verifying neutron emissions from my early fusor experiments. It was just the greatest thing to see Joe Zambelli's fusor mime mine with my counter, in his presence, and reaffirm his counter's calibration as well. This, was good experiment.

There are many things that make a neutron counter count.... only one of which is neutrons. Countrawise, when a neutron counter is clickin' away, it may have little or nothing to do with the actual presence of neutrons. Hold these admonitions close to your hearts, me buckos.

Richard Hull

[Richard Hull]

I sort of republished this announcement in our fusion news forum today. I felt it should also be there as well.

Ivan asked about the seed neutrons. I would imagine the scientists would turn there counters off and most of their instrumentation is surely computer controlled.

The neutron seed source is probably shut off with a thick window mechanically or, if generated by an accelerator, the power removed. As the bubble collapses, the computer would only then start the counter. This could all reasonably be accomplished in a few microseconds.

One thing that they must surely consider is the natural tendency for an ionized counter with high voltage applied to have a few false counts the minute a hot neutron source is removed. (excited gas)

If this work was done by responsible people, let us hope they employed a good neutron metrologist specialist to setup and take their measurements. Remember the problems with Pon and Fleishman's neutron metrology in 89? Not even physicists have a complete knowledge of such issues. This requires a specialized guy where this is all he does. Especially, when stunningly amazing results are found in an environment that is noisey and the counts are low.

By the way, I consider fusion in a liquid or water environment, effectively, a solid state system.

Richard Hull

[r_c_edgar]

Looks like Science published it on-line because of all the news.

http://www.sciencemag.org/feature/data/ ... ndex.shtml

This has the article itself, along with a couple responses.

--Ryan Edgar

[Richard Hull]

Thanks ever so much to Ryan for posting the detailed accounts and first cannon rounds fired over the bow of this ship.

I am always struck by Robert Park's truly knee jerk reaction to anything new which seems too simple to be real science. He is not alone either. Many take the tack that if it ain't big science, it ain't science. Park, has become the self-annointed naysayer-watchdog for the modern day, well ensconced scientific preisthood. Barking loudly at the mere approach of any new idea brought forth on a budget of under 100 million dollars.

Too bad his bark is so loud, for it scares off consideration by the weaker elements of science whose own work might be turned on should they not all join in lockstep behind the preisthood.

Thank God Science magazine in the little blurb responding to criticisms had the guts to recognise Park's efforts as somewhat blind and irresponsible, saying as much in the piece. Also, they noted how they had issues prior to publishing, but felt the paper deserved airing to at least distribute the idea.

I feel this was wise on their behalf for if this thing flys, they might be THE first public notice of the work. Which turns into prestige. If it doesn't fly, it will be forgotten. Empirical experiment will be the ultimate arbiter.........not Parks. "Let the experiment be done!"

Parks usually, if he can't intimidate verbally, will try another underhanded methodology to prevent anything he finds offensive from moving forward. I wonder if a campaign directed at his cronies could have them cancelling their subscriptions to Science in protest?

More and more, bizarre goings on in bulk matter are coming to the fore and with apparent processes that the hot fusioneers would not like to hear about.

All this might not be the new energy we all seek (low Carnot efficiency), but it could blow the bottom out of 20th century nuclear physics and force an alternative view. Regardless, if it does, all the naysayers will say they sort of knew it or suspected it all along.

While not wanting to go out and do it myself (too many irons in the fire already), I will keep close tabs on this just as with the CF work.

Thanks again to all you web crawlin' folks who take the time to research the news and post it here. You perform as valuable a function as any one of us cookin' up neutrons in a fusor. Everyone contributes in their own way.

Richard Hull

[DaveC]

With no particular love lost on large intstitutional science programs, I am however, inclined to be cautious in siding with either the underdog or the mainstream wisdom. Especially so in a case like this one. The cold fusion work is another good example...Let us all remember the hall mark of sound science is the validation of theory with repeatable experimentation. The chief problem Pons and Fleischman had was in non-repeatablilty of their initially reported work. Regardless of the predictions of existing "theory", a set of well documented and repeatable experiments can force a rewrite and correction to flawed theory, however plausible it once appeared to be. But.... a set of non -repeatable experiments unless wisely set aside until such time as they can be performed repeatedly, takes matters from the realm of fact to the realms of whimsy and fancy, even belief.

The historical record establishes beyond any reasonable doubt, that we are all quite poor in a priori anticipations of what should or should not be.

If the Sono-Luminescence -fusion experiment IS producing netrons, then a set of runs specifically to address the issue of false counts... should eliminate all controversy. This has yet to be done.

Richard's observations and cautions about proper self checking of the detectors should be heeded by all operating fusors and counting neutrons.


Dave Cooper

[ijv]

Just came accross this short paper on insturmentation problems in the bubble fusion experiments. Could be of interest to people working with neutron detection gear.

http://www.arxiv.org/ftp/cond-mat/paper ... 204065.pdf

[Richard Hull]

Thanks for posting the latest paper on this topic Ivan. I have printed it, read it and included it in my portfolio on this interesting announcement.

While not totally damning, this well presented paper does point out some inconsistencies privy to folks who reviewed the original submissions.

Having been through peer review myself, I know that authors of papers with problems will often resubmit after removing or revising parts of there paper to which a reviewer has commented negatively. Often, this is months or even years after the work is done. Even with good notes, it is often not possible to "align the data" if a major rewrite is in order to get approval from a reviewer.

The major problem that the authors of this paper had with the original paper revolved around two points.

The time differential between the light pulse and the neutron detection pulse. They are correct if the neutrons are indeed coincident with the light pulse. But are they?

The other point hinges around the channel numbers of the energy spectrograms provided by the original authors. This is more serious. Assuming these rebuttal authors were given a single spectrogram with all the pulses present, they are right to question the validity of the data.

It is important to note here for those not familiar with nuclear energy spectrograms that the channel number is not necessarily the energy of the particle! Yes, most folks doing gamma spectroscopy try and make the channel number equal the kev of the gamma particle, but no such rigid operating conditions are demanded. In fact, in some situations, it is just not possible. An example would be in Alpha spectrometry where there are no 1mev alphas and the range for normal alphas is between about 2.5 and 7 mev. on a normal 1024 or 2048 channel device (relates to the memory points avaialble), such an alpha span would occupy too small a slice of the useable area (i.e. 250 - 700 channel numbers).

One has many adjustments on a multichannel analyzer where the span, resolution and upper and lower setpoints are adjustable. Thus, one can get a tremendous increase in resolution about a point of interest, by chasing the upper and lower setpoints over the entire span of channels closer together. Again, an example... Suppose I were interested in only two alpha signals at 4.85mev and 5.16 mev I could set the SCA and upper and lower setpoint on the span so that channel 1 could be 4.5mev and channel 1000 could be 5.5 mev.
To do this, I would have to have a "cal source" of radioactive that had a well known alpha in this span. This would mean that each channel would now represent 1kev of energy but my spectrograph would never even show a 5.8mev particle. (off screen)

It is often common for some papers to NOT include details like this!!! The old boys club is often expected to do the math for themselves. However to be meaningful, at a minimum, a reference pulse is supplied, with an indication of kev/channel or two known reference signals are given with no channel calibration stated. It sounds complicated, but to guys in th' biz, it is old hat.

The authors of this rebutal paper are having problems with the references not lying where they are supposed to based on the stated info. for the actual data pulses! This is a problem, and these guys point it out. It is possible all is well and that some charts got mixed up when a different calibration was used and were published in error, but the authors of the origianl paper should respond if this is the case or offer up some form of clarification.

Here is where the rub comes. They have to sift through their old notes and pray they have good cal info about the data in question here.

Again, thanks for the paper Ivan. More of what science does well within the peer review system.

Richard Hull

[DaveC]

I have read Putterman et al's response to the Taleyarken rebuttal paper. I think the points they make are valid. In the pico to nanosecond time regime of sono-luminescence, a microsecond delay requires explanation.

Further, wtih regard to the bin numbers of a MultiChannel Scaler, once the limits are set, the bin numbers correspond to definite voltages. As long as the calibration remains in place for the experiment, output from an energy proportional detector thus corresponds to actual particle energies, and the bin counters will contain the number of particles detected at each energy. The MCA I've used ( a Norland ) is really a quite stable and predictable instrument. I used it for partial discharge detection at the femto coulomb level, where signal detection can be tricky, and when properly setup, calibration pulse were so predictable, they rarely differed more than a couple bin numbers in all runs. And why shouldn't they? A 4096 or 8192 count scaler, scales the incoming data in increments of voltage that are exact and constant. For a 10 volt maximum signal, the 12 bit digitizer gives 2.5 mV resolution. What IS an important and subtle issue is the digitizer speed. Some scalers have a slow clock. The clock speed sets the lowest time increment limit for digitiizing process. A 200 MHz clock speed (5 nS per clock tick) will take about 65 nS (13 ticks = 1 tick for each bit plus one more) to complete a full range 12 bit digitizing of a signal sample. This is necessary BEFORE any bin value can be assigned. Thus data actually only gets logged at a rate equal to 1/65nS or about 16MHz, a far cry from the basic clock speed. Smaller signals ovbiously take less time..

One thing that should be obvious from this is that sequential samples of various sizes will be taken a various times, not at the same intervals, unless specifically gated. If gated, the gating rate cannot exceed the maximum digitizing time without the risk of missing large signals.

I believe that instead of rushing to be first to publish questionable data, researchers should exercise the wisdom to attempt to explain their results in terms of every imaginable mistake and vagueness of their experimental process. Science Mag did not come off in this case looking particularly astute.

Dave Cooper

.

[ijv]

Glad to see that you could get so much out of 3 pages!