Yale fusion seminar (Moshe Gai)

[Keith Calvert Ivey]

have something to contribute.
Today (28 April 1989) at 2 pm the physics department here at Yale
had a special seminar by Moshe Gai (a nuclear physics professor here)
with the title "Does Cold Fusion Exist?". My notes on the seminar fol-
low.

Prof. Gai gave a seminar yesterday with an almost identical title as
part of an undergraduate research colloquium. Greg Howard posted notes
on this yesterday. Today's seminar was, I believe, more technical. I
guess yesterday's seminar drew off some of the crowds of fusion enthusi-
asts. I arrived at the room only twenty minutes before the talk, and
was just the third person there. By shortly after 2 there were about
200 people, only a couple of whom seemed to be with the media.
Prof. Gai began by saying that his results would be written up in
about two weeks. He is speaking about them at the American Physical
Society conference on Monday, and this seminar was a trial run for that.
He said that his results are negative and that Princeton and MIT have
come out with similar results within the last week or so, and that there
are rumors that the Cal Tech experiments agree with his too. He admit-
ted that he might have been "out of line" in saying some of the things
he did to undergrads yesterday. The work involves "seven people and at
least ten opinions". The workers are nuclear physicists from Yale and
chemists from Brookhaven National Laboratory (in Upton, New York).
He gave the chemical equations for the process of absorption of
deuterium into a metal lattice during electrolysis:

- -
D O + e --> D + OD
2 ads

-
D + D O --> D + OD
ads 2 2


D --> D
ads lattice

where D(ads) is a deuterium atom adsorbed on the electrode surface and
D(lattice) is one absorbed into the lattice. When fully saturated, the
lattice contains one hydrogen molecule per titanium atom or one hydrogen
atom per palladium atom.
Possible nuclear reactions between two deuterons are:

{ 3
d + d --> { He (0.82 MeV) + n (2.45 MeV) ~50%
{
{ t (3.03 MeV) + p (1.01 MeV) ~50%
{
{ 4
{ He (0.08 MeV) + gamma (23.77 MeV) ~10^-4

The probabilities given are for "normal" conditions. There has been
much theorizing about the possibility that they change when the deuter-
ons are confined in a lattice.
Since heavy water will always have a considerable amount of normal
H2O contamination and since protium diffuses into palladium much faster
than deuterium does, fusion of a deuteron and a proton must also be
considered:

3
p + d --> He (0.005 MeV) + gamma (5.49 MeV)

A paper by Koonin and Nauenberg to be published in _Nature_ suggests
that this p+d reaction occurs about 10^8 times faster than the d+d reac-
tions given above. This combined with the isotope effect on diffusion
mentioned above means that the proportion of p+d to d+d reactions will
be much greater than one would expect based simply on the ratio of pro-
tons to deuterons in the electrolyte.
Gai is "very bothered" by the neutron measurements given in the
Fleischmann and Pons paper, especially the 4x10^4 neutrons/sec given as
the flux, which he thinks is remarkably high. He would like to see the
raw data so he can figure out how they treated the background. He also
complained about the strange nonlinear scale given on their spectrum of
beta emissions, which seems to indicate negative energies on the left
side of the graph. The F&P paper was overall very difficult for him to
figure out. He likes the Jones paper much better because they give
"real numbers".
The Yale-Brookhaven setup consists of four(?) electrolytic cells
partially surrounded by six neutron detectors and two sodium iodide
crystal detectors for gamma rays. This is enclosed in ~15 cm of borated
concrete and ~15 cm of borated paraffin, and topped by two cosmic ray
detectors so that possible muon-catalyzed fusion resulting from cosmic
rays can be "vetoed". They had several tons of lead to use as shielding
for a while, but got rid of it because the lead itself was causing a
high background. A neutron coming from the experiment interacts with
the first neutron detector (#0), which sits directly below the cells,
and then scatters to one of the other five which are arranged in a ring.
They require coincident signals from two detectors (#0 and one other) to
give a neutron count. They can get some energy information about the
neutrons with this setup, but the placement of the detectors requires a
compromise between efficiency of detection and precision of energy in-
formation. Signals from gamma rays and neutrons can be distinguished
easily by the shapes of the pulses.
Nitrogen gas is cycled through the cells to remove hydrogen gas,
keeping it below the 4.8% required for an explosive mixture with air.
The nitrogen is wetted with D2O to replace that lost by electrolysis.
They used nine electrodes and nine electrolyte solutions in various
combinations. They wanted to check every suggestion they had heard for
getting things to work. I didn't get all the information about the
electrodes and electrolytes, but here's what I have:

Electrodes:

1) Pd plate - cold-worked (pounded with a sledge hammer to create dislo-
cations in the lattice structure), then heated in D2 (300 degrees C, 120
psi) and anodized.

2) Pd cylinder - annealed in flowing argon at 1000 degrees C.

3) Pd cylinder - annealed in flowing argon at 1000 degrees C.
(There was a difference between this and #2 that I didn't get.)

4) Pd cylinder - annealed in vacuum ("super electrode").

5-8) Ti parallelepipeds, cold-worked.

9) TiFe Mn powder, "hydrided" at 120 psi D at 900 degrees C,
0.7 0.2 2
charged on 19 Dec 87 and recharged on 04 Apr 89. Contained in a 2x20 cm
cylinder pressurized to 120 psi. (I don't understand how exactly this
was used as an electrode, if it was.)


Electrolytes:

1) 0.1 M LiOD, 97.5% D2O

2) 0.1 M LiOD, 99.8% D2O

3) 1 M LiOD, 97.5% D2O

6
4) 1 M LiOD, 97.5% D2O

5-8) the solution of Jones et al. (100 g D2O plus 0.125 g each of var-
ious salts. Someone asked if they included "a small amount of AuCN".
He said no.)

9) 0.1 M LiOD, 99.3% D2O

In order to test the hypothesis that "ignition" by energetic parti-
cles was necessary to start the fusion, Gai disassembled the smoke alarm
from his home and spot-welded its americium source to electrode #1 for
some of the experiments, thus providing 5 MeV alpha particles.
The neutron detection employed "state-of-the-art" pulse-shape detec-
tors not yet commercially available. The threshold for neutron detec-
tion was ~0.5 MeV. Efficiency of detection, taking into account coinci-
dence was ~1%. The signal was filtered by software to remove gamma ray
signals in counting neutrons and to exclude neutron counts with energies
greater than 3 MeV.
The most thoroughly analyzed data comes from the last seven hours of
the experiment. Gai displayed data showing that during this time detec-
tor #1 counted a grand total of 2 neutrons, which the group named
"Fleischmann" and "Pons". (I think that was a bit of a cheap shot.)
The total number of neutrons counted in the 7 hours was 17 with the
cells off and 13 with all cells on (some had been running for two weeks
at this point). There was thus no statistically significant difference.
In 31 hours on 18 April the sodium iodide crystals detected no p+d
or d+d gamma rays. Background gamma rays from decay of radioisotopes in
the concrete were detected at 2.1 and 2.6 MeV. Gai wonders if this may
explain the 2.5 MeV signal seen by Fleischmann and Pons with their NaI
crystals.
Gai gave the three-standard-deviation upper limits on fusion yields
as < 2x10^-25 fusions/deuteron pair/sec for d+d (based on neutron
counts) and < 2x10^-22 fusions/pair/sec for p+d (based on gamma ray
counts). He says the first compares favorably with the number given by
Jones et al., 10^-23 (what's a factor of 50 between friends?).
At one point they saw a shoulder in the gamma ray spectrum which
ended at about 5.5 MeV. This shoulder "weighed on their shoulders" for
quite a while, but the eventually determined that it was residual radio-
activity from their calibration of the NaI crystals with a Pu-Be source.
Na-23 in the detector captures a neutron to give Na-24 which decays with
a half-life of ~15 hours. The shoulder remained even when shielding was
interposed between the detector and the cells, so it could not have been
coming from the experiment.
Gai stressed several times that he was "not making any statement
whatsoever about nonreproducibility of the result of Pons and Fleisch-
mann". He said that they had done their best to repeat what F&P had
done -- though it was somewhat difficult to figure out some of the de-
tails -- and had seen no remarkable results.
After this there was a question-and-answer session. The first ques-
tion was from someone who claimed that palladium actually absorbs deute-
rium faster than protium. Gai said that papers by Pons had shown the
opposite.
Then there was a question (from an inorganic chemistry professor)
about the heat F&P had gotten and how their electrode had melted. Gai
said his measurements were only relevant to the emission of neutrons and
gamma rays, and that he was making no statement concerning heat. Howev-
er, he claimed that putting lithium in palladium decreases the melting
point by "about an order of magnitude". (I find this difficult to be-
lieve unless he's talking about a significant percentage of lithium,
essentially an alloy.) He also said that on of the Brookhaven chemists
had a completely chemical explanation for the melting of the electrode,
but he didn't reveal what this involved.
To a question about the possibility that He-4 could lose energy
gradually by a cascade of low-energy gamma rays, Gai responded that this
was impossible since the first excited state of He-4 has an energy of
~20 MeV. There is no way for it to lose energy in small steps.
The next question came from a physical chemistry professor who was
at the University of Utah before he came to Yale, and who said he had
visited there yesterday. He said that the metallurgy department at U of
U has reproduced the heat production of F&P's experiment. According to
what they told him, cold-working the electrodes was about the worst
thing one could do. He said they prepared their electrodes in a special
way. Gai asked him, "What's the secret?", but he wouldn't talk about it
with the press in the room. He talked about it after the seminar broke
up. It turned out the secret was using cast electrodes; this has been
on the net for several days at least. The other things was that F&P ran
their electrolysis with the electrodes completely submerged. Also,
supposedly there is only one guy who can make electrodes that work. One
wonders if they've taken out an insurance policy on him with Lloyd's of
London.
Gai refused to comment on Hagelstein's theory about dissipation of
He-4 energy by loss to the lattice, but several in the audience said
they couldn't imagine how the lattice could absorb 20 MeV without melt-
ing. Gai did point out that Stanford has found a lot of He-4 in their
electrode.
To a question about the Mossbauer effect, Gai replied that it would
not apply in the case of He in a metal lattice, since the Mossbauer
effect requires that the energetic nucleus be part of the lattice and He
would be in the open space.
After the seminar a nucleus of chemistry and physics faculty formed
about Gai and the chemistry professor recently returned from Utah. This
nucleus was surrounded by a small cloud of graduate students, including
me. There was much trading of rumors, most of which I had already heard
on the net. The chemistry professor said that one of the main reasons
that things were such a mess was the behavior of the U of U administra-
tion (one of the reasons he left there). He said that U of U was trying
to keep as many details secret as possible because of patents. He also
said that the administration had kept F&P from putting the name of their
student (Hawkins(?)) on the paper as an author, so as to decrease the
number of people involved in possible patent claims, but that in the
errata to the article this had been corrected, along with the strange
scales in the figures and such. "We are sorry that the name of one of
the main contributors to this work was inadvertently left off the list
of authors"? He also said F&P believe that the Italians have the best
information about the F&P setup because they sent people over very early
before the administration cracked down on communication. Also, he said
F&P were going to Los Alamos to help with replication there.

I don't think Gai's title for the seminar was a good one, since he
didn't even pretend to give an answer to this question of whether cold
fusion exists. It seems that his negative results are irrelevant since
he did not measure heat and did not use cast electrodes. Why can't
anyone measure everything on the same experiment: neutrons, gamma rays,
and heat? What were the chemists for if not to help with calorimetry.
I guess to help with electrochemistry. I hope Gai brushes up on his
chemistry before going to the APS meeting. He thought LiOH was lithium
_hydride_ and he repeatedly used "hydriding" or, worse, "hydration"
instead of "hydrogenation". (Maybe it should be "hydridation" in analo-
gy with "oxidation".) I guess he doesn't need to know this stuff to do
the experiments, but it makes a bad impression.
Well, that's about it. I hope I didn't make too many errors.
========================================================================
Keith Calvert Ivey <ive...@yalevm.BITNET>
Yale University Department of Chemistry
Box 6666, Room 1 SCL, New Haven, CT 06511

[Jorge Stolfi]

Keith Calvert Ivey quotes from M. Gai's talk:
>
> Since protium diffuses into palladium much faster than
> deuterium does...

I hope that cold fusioners who are trying to confirm the F&P claims
keep this fact in mind when designing their light-water control
experiments. Looking at the non-nuclear explanations for F&P's extra
heat that have seen so far, it seems that they all depend somewhat on
the diffusion rate of deuterium in palladium.

For one thing, if electrical current in the hydrogen-saturated
palladium electrodes is carried primarily by H+/D+ ions, then I suppose
that the electrode's resistance will be a greater with D+ than with H+.
This difference could affect not only the Ohmic heating of the
electrode, but also the voltage gradient at its surface and in its
interior, and the ease with which the hydrogen enters and leaves the
lattice (hence its concentration in the lattice). It should also
affect the size and distribution of the hypothetical H+/H- domains, the
resistivity of the solution, the magnitude of any "heat pump"
effect, and so on.

Thus, merely observing that F&P-style cells with heavy water get hotter
than identical cells with light water is not enough. For the
experiment to be conclusive, one must carefully measure the electrical
input and the heat output. I wonder whether the Stanford team did this...

Jorge Stolfi (sto...@src.dec.com, ...!decwrl!stolfi)

=========================================================================
DISCLAIMER: Don't bother reading this posting; the author doesn't know
what he is talking about.