Path: santra!tut!draken!kth!mcvax!uunet!cs.utexas.edu!tut.cis.ohio-state.edu!cwjcc!gatech!prism!loligo!mccalpin From: mccalpin@loligo (John McCalpin) Newsgroups: sci.physics Subject: Impressions from APS Meeting Summary: Shaken, not Stirred Message-ID: <710@loligo.cc.fsu.edu> Date: 12 May 89 20:07:15 GMT Sender: mccalpin@loligo.cc.fsu.edu Reply-To: jac@scri1.scri.fsu.edu (Jim Carr) Organization: Supercomputer Computations Research Institute Lines: 172 >From: jac@scri1.scri.fsu.edu Shaken, not stirred: The juxtaposition of the Tuesday night lecture by James "The Amazing" Randi on defects in the Benveniste experiments regarding the observation of a homeopathic effect in highly diluted solutions produced by a mystical shaking process and the Monday night presentation by Nathan Lewis of CalTech that emphsized the importance of F&P(&H) failing to stir their "cocktail" inspired the title of this little note. I was at both the Monday evening session AND the session Thursday morning when Jones (BYU) gave the invited talk that now seems to have been the catalyst for this media event. (My hypothesis is that the invitation to talk at Baltimore forced the hands of the Utah researchers. I also think if no big announcement with claims of lots of energy had been made by Utah, Jones' talk would have attracted little more than the few hundred at his session on Thursday.) I will give only a short reaction to the facts heavily reported in the media and emphasize those new facts brought out in the scientific sessions. In order of importance: (1) Lewis presented a clear, coherent description of how calorimetry is done and what the numbers and incantations mean. Since F&P did not do this in anything we have seen, Lewis answered the main question in my mind about the arithmetic magic that turns an 11% excess into 189% of breakeven. He extracted "raw data" from the F&P table which clearly show, if they are correct, that the cell was actually absorbing energy. It just was not as cool as one would expect IF the cell is always in thermal equilibrium AND all the evolved gases escaped. This is the meaning of excess heat, according to Lewis. We can await the meeting on 5/8 to see if F&P will show their data book entries on this point. (They did not.) It is a "fusion refrigerator". Further, the voltage used to get the final column in their tables is an assumed value. Lewis measured 3.5 to 4.0 Volts, not 0.5 V, to put D2 back into the cell. He remarked that since the reference electrode is at 0.8 V, the 0.5 V is impossible. In addition, the IR losses during electrolysis and recombination were not accounted for at all. After being told of this, Fleischmann told Congress these columns are hypothetical. The "four times the energy put in" words still rattle around in the media without comment from Utah, however. (They now claim 50 times, but this is presumably the same cooked numbers.) Finally, the failure to stir is crucial, as emphasized by the Stanford team (talk # 9). There are large thermal gradients in the cell which will bias the results, depending on where the thermometer is located, IF the cell is not well stirred. Any remnant of this gradient means that the effect of power into cell will not be the same as equivalent power put into the resistive heater located elsewhere in the cell, or by dumping in hot D2O as F&P describe. Stanford showed that D2O gives more apparent heat than H2O due to different gradients produced by differences in the electrochemical potentials of D2O and H2O with the cell operating conditions used. (2) Besides choosing not to attend this meeting, F&P have also consistently refused to share with their colleagues the data relevant to the points raised above. Their public assertions that they would be glad to give advice on the experiments have been refuted by the experience of Lewis and others. There are other defects worth noting, given the discussions in various places. The failure of F&P to show the full gamma spectrum and give calibration information is crucial. I have never seen a single peak presented in this fashion in an experimental nuclear paper. The use of a big NaI detector instead of a GeLi detector with 0.002 MeV resolution is also inexplicable until one knows the measurements were done by a radiation safety officer. They have the full spectrum, because they showed it at the news conference, but have not shared it so we can see what background peaks they might have had. They also do not show the mass spectrum used to deduce the presence of fusion products. This would show us if there was contamination from room air. Lewis warned that Tritium measurements can be thrown off completely if one does not properly neutralize the LiOD before mixing in the material used as a scintilator. Enough said --- None of this is addressed in the published paper, and lack of such information apparently led to the rejection by Nature. (The comments on page 701 of volume 338 (27 April), although misstating that BYU had not done a control experiment, were clear on the point that F&P(&H) have not produced "records of cell voltage and current for the full duration of each run". I assume this must be one of the onerous demands placed on them which led to the withdrawl of the paper from Nature.) Koonin is quite correct when he says the F&P work does not meet the usual standards of scientific discourse. I also think he is correct to be concerned that continued failure to address them indicates the possibility of misconduct on the part of the researchers. There are examples in the past, of which N-rays and homeopathy are only a small subset, where scientific error turned into alleged misconduct as researchers fought the criticisms of their colleagues. It is interesting that statements analogous to "you have an incorrect cell" were quite common in the N-ray business, where only Frenchmen could see the effect. (Perhaps the secret is that one has to anaesthetize the palladium rod so it does not give off N-rays!!) (3) Jones was very defensive, and rightly so. One man from William and Mary called him a scoundrel (but not in public). His opinion seemed based on hearing stories about Jones' role as a referee, but obviously without reading the article in Science about same. Jones showed his lab books at every opportunity. (When did F&P get the idea to look for neutrons and gammas, I wonder to no one in particular?) He made no great claims for a breakthrough, even telling an amusing story about how the Idaho Lab wanted to make a big thing of scientific breakeven with muon-catalyzed fusion a few years back until Jones and colleagues talked the administration out of it. I repeat that he used a 10% detector with 10% solid angle for total 1% efficiency. The few people who tried the BYU cocktail were not so clear on their efficiency or energy resolution. (Essentially all negative results were with the Utah cocktail.) Since he reports a typical fusion rate 2 to 4 times lower than the best rate, and the runs must be 3 to 8 hours long at most, it remains unclear if his experiment has been fully tested. It will be when the results from Gran Sasso (Italy), done under a mountain with low background rates, come in. A few questions were addressed to the background and the significance of the peak, but most of this came up Thursday. Jones repeatedly emphasized that everyone should read the published version of the paper (Nature 338,737 in the 27 April issue) because it contains more information than the preprint. I also suggest that you read the comments of one of the referees in a letter on page 711. It is a valid criticism, but would be addressed by the experimental protocol I outline in item 5 below. (4) Jones was strongly attacked at the session Thursday morning. Eric Adelberger (U of Washington, and an expert at measuring small signals such as the "fifth force" experiments he has been conducting) was adamant that using ratios to background is wrong. BYU followed this procedure in constructing their Figure 5 (but not Figure 2) because they had different background levels after they replaced burnt up electronics. This did not sit well with the audience. I asked him to show the individual spectra for a background and regular run. (Figure 2 is integrated over all of the runs and all of the background runs, with the latter scaled to equal time. In Figure 3 an additional scale factor is used as described in the text.) As I expected, they are very weedy BUT you CAN see evidence of a peak so it less likely that the signal comes from data massaging. You would not get a 3-sigma effect from them individually. The crowd appears to have convinced Jones that we really do want to see the raw spectra for his background runs (room, D2O without current, and H2O with current) compared to the real runs so we can make up our minds. Seeing the real data lower my confidence level to about 50% but also assuage some larger residual fears. (5) The experiment that needs to be done is for Jones to take data for 3 hours with the cell running followed by 3 hours of background, and keep this up for as long as needed. Then subtract the average of the background on either side of a give run and accumulate the result. It would be easy to include power off "runs" at random in this series so that a control spectrum could be accumulated at the same time. The same thing could be done with H2O and a 50-50 mix. Even better if this were done double-blind. A comparison of the resulting 6 spectra would settle the issue, one way or the other. (6) Commentary: I was not alone in feeling insulted that neither P nor F would show up at the meeting. I know the session chairman well, and first learned of his troubles over wonderful mexican food at El Azteco in East Lansing (MI) while attending a meeting the week before. Suffice it to say they were willing to go to extraordinary lengths to get F&P to attend. I was even more upset when I read that a U of Utah admin- istrator compared an open meeting of the APS to a press conference. Koonin can make all the disparaging remarks he wishes (such comments are more common at theoretical discussions than most would be led to believe) about someone's science, but all scientists should take offense when science itself is attacked in this fashion. (7) Finally, it is VERY IMPORTANT to note that the F&P(&H) paper in the Journal of Electroanalytic Chemistry (261,301 in issue #2A, 10 April) is a "preliminary note" and that one of the editors is Fleischmann's colleague at Southhampton. You should read the statement inside the cover that defines this form of communication, and observe that it does not have to meet the standards of a full paper. It is not surprising that a similar paper was not accepted by Nature's referees. -- Jim Carr "A nuclear theorist, who proudly notes that JAC@FSU.BITNET the APS videotaped its scientific session while the Electrochemical Society banned all recording devices of any kind. Hmmmm..."