ChatGPT test at LENR-CANR.org

Here is the Hagelstein ICCF4 paper OCR output from AWS Textract. I uploaded the images. To test ChatGPT, I uploaded the text that was OCR'ed by Adobe Acrobat. So the comparison is unequal. This version still has hyphenated words at the end of lines, such as:

. . . energetically far removed from the continuum; conse-

quently . . .

That is very easy to fix with a text editor or a program. ChatGPT fixes this automatically. But, as I said, ChatGPT leaves a lot noise in the file.

The original paper in image format is here:

https://lenr-canr.org/acrobat/EPRIproceeding.pdf#page=221

NEUTRON TRANSFER REACTIONS

Peter L. Hagelstein and Sumanth Kaushik

Massachusetts Institute of Technology

Research Laboratory of Electronics

Cambridge, Massachusetts 02139

ABSTRACT

A new model is proposed to treat configuration mixing between bound and continuum neutron

states in a lattice; the Hamiltonian for this model is of the form of the Anderson Hamiltonian. In

condensed matter physics, the Anderson model describes (among numerous other effects) electron

hopping in semiconductors; the neutron model presented here predicts neutron hopping in lattices

containing a mixture of isotopes. This result is new.

The Anderson model treats the mixing between localized states embedded in a continuum. In

the neutron model, the localized states are energetically far removed from the continuum; conse-

quently, the neutron model treats a much simpler mathematical problem.

Brillouin-Wigner theory is applied to a restricted Fock space version of the model containing

states with 0 and 1 neutrons free. This leads to perturbative results that describe the effects of

continuum neutron mixing to lowest order. The resonant scattering of virtual neutrons is predicted

to lead to neutron delocalization, as long as the interaction perturbs either the linear momentum

or total angular momentum of the nucleons.

Delocalized neutrons can be captured, with the reaction energy going into gammas and other

incoherent decay products; such reactions are predicted by this model. Delocalized neutrons can

be captured accompanied by energy exchange with the lattice. Formulas describing this type of

reaction are derived, and the resulting rates estimated.

1. Introduction

During the last several years, there have been numerous reports of the observation of excess heat

in electrochemical experiments following the initial announcement by Pons and Fleischmann of the

10-1

effect in 1989. 1-4 The magnitude of the claimed effect is very large, with net energy production in

some experiments reported to be in excess of 100 eV per atom of cathode material. This level of

excess energy cannot be of chemical origin; if it is correct, then it must be due to a nuclear process.

These experiments have proven to be difficult to verify, and there is doubt on the part of

most members of the scientific community as to whether there even exists an effect. Theoretical

arguments given early on as to what the origin of the effect might be were easily dismissed. The

seeming absence of any compelling theoretical reason as to why there should be any effect at all,

together with the rather poor signal to noise ratio of the initial experimental data, combined with

the seeming irreproducibility of the effect, has led to the general rejection of the effect outright

by the scientific community. Noted skeptics are now quick to bring up the topic of UFOs when

discussing research on the Pons-Fleischmann effect.

The question of whether or not there is an effect is ultimately an experimental one; clearly the

work reported here was motivated towards seeking theoretical explanations of the effect under the

assumption that an effect exists. Hopefully, the experimental issues will be settled soon, although

it is clear that the work is hard and progress has been slow.

When the effect was first reported, there were speculations as to possible origins of the effect.

Most speculations centered around the possibility that dd-fusion was somehow responsible for the

heat. The fusion of deuterons in quantities commensurate with the claimed heat production would

lead to large tritium production and lethal neutron generation rates; neither is observed in the

experiments. Even now, the majority of theorists who continue to work in this area are focusing

their attentions on fusion mechanisms, for explaining heat or other effects.5-7 Some popular current

proposed explanations postulate that a new fusion channel exists that leads to 4He production, with

the energy excess going into the lattice.

We have focused instead on novel reaction mechanisms involving the proposed exchange of

neutrons between distant nuclei in a lattice.8-10 The basic proposed effect is a neutron analog of

electron hopping in semiconductors, with energy exchange with the lattice taking place through the

frequency shifting of highly excited phonon modes: this mechanism will be discussed in the present

work, and discussion of lattice energy transfer appears elsewhere.8-11

There are two principal difficulties in the proposal of neutron transfer reactions as a candidate

reactions to account for heat production. The neutron transfer was proposed to take place through

virtual neutron states, and it is well known that the range of virtual neutrons near an isolated

nucleus is measured in fermis rather than Angstroms; this is the first difficulty. Virtual neutrons

must somehow be delocalized before any reactions can take place. Last year we noted that de-

localization can be induced through scattering, we evaluated delocalization effects due to Bragg

scattering; we also proposed that resonant scattering of virtual neutrons might lead to observable

10-2

effects. We proposed last year 10 that the resonant scattering be mediated through electromagnetic

interactions; these are found to be too weak, and here we propose that resonant scattering mediated

by strong force interactions is a far more likely route.

The second fundamental difficulty with heat generation through neutron transfer reactions is

the problem of coupling nuclear energy with the lattice. It can be shown that direct recoil effects

are not capable of mediating the requisite large energy transfer without the presence of fast (MeV)

nuclei. We instead proposed that very large energy transfer can be mediated through changes in

the

basic structure of the phonon modes.8,9,11 In a sense, energy transfer through the creation

or destruction of phonons doesn't work; energy transfer through the modification of pre-existing

phonons does work, at least theoretically.

There exists no currently generally accepted experimental evidence supporting the proposal

that neutron hopping can occur in a lattice. We will argue here that neutron hopping is somewhat

analogous to electron hopping, and argue further that the Anderson model ¹² used for electron band

mixing calculations can, with modifications, be used for neutron problems. 13 This statement is in

fact the primary result reported in this work. A key feature of the Anderson model is the presence

of localized states that are embedded in a continuum of free states; the neutron localized states are

several MeV below the continuum states, so that although mathematically similar to the Anderson

model, the neutron model is very much simpler.

Having posed the model, we begin the task of analyzing the model to extract reaction rates. The

field of Anderson model studies is by now relatively mature; variational methods, perturbation the-

ory and canonical transformation approaches have proven to be very successful is analyzing solutions

for the Anderson Hamiltonian. We have attacked the problem using infinite order Brillouin-Wigner

theory, which leads directly to a perturbation expansion that is relatively easy to understand; the

evaluation of the resulting formulas is less easy. We propose here rather crude estimates of reaction

rates; this is perhaps appropriate, since this work is the first publication on the neutron lattice

Anderson model.

The formulas appear to show that neutron hopping can occur at fast rates that would not have

been anticipated if either phonon exchange or total angular momentum exchange occurs during a

single site-to-site transfer. This is a key result of the present work.

10-3

2. Neutron Transfer Reactions

In previous publications, we have worked towards the development of a theory for neutron

transfer reactions in a lattice mediated by electromagnetic E1 and M1 interactions.8-10 In

the

present work, this theory is developed further, and extended to include effects mediated by the

strong force.

We have explored a model including the lattice, nuclei, and free neutrons. In the absence of

recoil effects, the resulting model is mathematically equivalent to the periodic Anderson model in

condensed matter physics; our application of the Anderson model to describe neutrons, rather than

electrons, is new. Electron hopping effects are well-known in condensed matter physics, and can

in certain limits be described using the periodic Anderson model; applied to neutrons, our model

describes neutron hopping, an effect unknown prior to our studies.

The periodic Anderson model has been analyzed using a variety of techniques. It has been

solved approximately using the Schrieffer-Wolff rotation;1 it can be solved exactly in certain

limits through the use of a canonical transformation. 18 In the limit that essentially no neutrons

are free, we can simplify the problem by truncating Fock space to include only zero- and one-

neutron subspaces; the resulting equations can be solved approximately using Brillouin-Wigner

theory, which leads to estimates for free neutron densities and reaction rates.

Upon evaluating the resulting formulas, we find that no coherence factors appear. In previous

publications, 10 we had speculated that coherence factors should appear by analogy with the coher-

ence factors that occur in Dicke superradiance. This speculation is found to be in error, for rather

fundamental reasons associated with the fermionic statistics associated with the neutron transfer

and bosonic statistics associated with photon emission.

Furthermore, we find that in the absence of coherence factors, the effects associated with elec-

tromagnetic E1 and M1 transitions are trivially small, and can not by themselves lead to any

observable new effects. The relative weakness of the electromagnetic effects comes about due to the

smallness of external electric or magnetic fields applied to the nucleus. A significantly larger effect

is possible through the use of the strong force interaction, and we find that neutron hopping rates

and neutron transfer reaction rates may become important when mediated by the strong force.

There are a variety of neutron transfer reactions that are possible in this model. A bound

neutron that couples to continuum states will most often not stray more than a few fermis from

the parent nucleus; this effect appears as the lowest order effect in all solutions to the model. and

is also present in the isolated nucleus problem. The neutron will occasionally become delocalized;

in the theory, this appears as a result of the possibility of scattering of the free (virtual) neutrons.

Delocalized neutrons can "hop" from nucleus to nucleus. A large rate is estimated for the resonant

10-4

process where a delocalized neutron hops between equivalent nuclei, such that the total isotopic

distribution is unchanged (so-called "null" reactions); although of great theoretical interest, these

reactions are not easily observable, except possibly through isotopically-sensitive self-diffusion ex-

periments. However, delocalized neutrons that hop onto a nonequivalent nucleus can result in

gamma or alpha emission, as well as the creation of radioactive species.

If a delocalized neutron hops onto a nonequivalent nucleus, and if the lattice can provide or

accept the energy defect, then a large reaction rate is predicted. We have discussed previously the

conditions under which significant energy exchange during a neutron transfer reaction can occur

with the lattice; energy transfer can occur if a highly excited phonon mode changes frequency by

jumping across a phonon band gap. In this case, the process becomes resonant, and the associated

reaction rates are calculated to be very large. The end result of such reactions can in principle

be the generation (or absorption) of net energy, modification of pre-existing isotopic ratios, or the

production of new isotopes.

We must now consider how this theory might address some of the outstanding experimental

claims. In the case of the Pons-Fleischmann experiments, the heat-producing reaction with the

smallest energy defect is the transfer of a neutron from 105 Pd to 6Li, with an energy mismatch of 156

KeV. The symmetry of the Pd transfer involves d-wave neutrons, which results in small interaction

matrix elements. Perhaps more promising are neutron transfers between 11B (p-wave) and 29Si

(s-wave), with an energy mismatch of 849 KeV; and neutron transfers between 29Si as donor and

29Si as acceptor, with and energy mismatch of 2.14 MeV. In the light water heat experiments, 19-20

there occurs a relatively close match (12 KeV) for neutron transfers from 62 Ni (p-wave) and 29Si

(s-wave). Tritium production in this model could come about due to neutron transfers from a

number of materials to deuterium; examples are given in Tables I and II in Section 9.

3. The Periodic Anderson Model

The proposed neutron transfer reaction in metal hydrides is in many ways analogous to the

problem of electron hopping in solids. Before discussing the neutron problem further, it seems

appropriate to first review briefly the electron version of the problem. The most relevant condensed

matter model are variants of what is termed the Anderson model, following the initial treatment

by Anderson of the s-d mixing of an iron-group metal impurity in a host metal. 12 This Anderson

model treats the coupling between an isolated localized impurity electronic state that is embedded

in a conduction band. This model has been generalized to treat multiple impurity states in a metal,

and further generalized to apply to mixing between valence and conduction bands in a lattice. 23-28

10-5

Whereas the neutron analog of the single impurity problem is simpler (since the bound state is so

far removed from the continuum states), the neutron analog of the lattice model will be of great

interest in what follows.

In the non-degenerate version of the periodic Anderson model, the Hamiltonian can be written

as

k

k,o

(3.1)

where nio = diodio. This Hamiltonian includes individual Hamiltonians for the conduction

band (which are the s-orbitals of the s-d model), the localized valence states

Ei,o Edditodic and the valence correlation energy The last term on the right

hand side mixes the valence band and the conduction band. The spin dependence of Vk has been

suppressed, following a convention used in much of the literature.

Numerous variants of this model appear in the literature. In the limit that the correlation

energy U is taken to be zero, the Hamiltonian can be diagonalized exactly. In the alloy analog

approximation,2 the correlation term is omitted, and a site-dependent valence energy is introduced.

In many cases, the correlation energy U is very large, so that only two valence configurations are

present. In this case, the problem can be restated in terms of more complicated transition operators

that prevent inclusion of unwanted configurations in the problem. For example, the degenerate

periodic Anderson model sometimes appears as29-33

=

k,o

i,k,o,H,A

(3.2)

where the X operators are Hubbard operators; these are more complicated operators that are made

up of products of single particle creation and annihilation operators.

Sakai et al³ have used a model Hamiltonian to study mixed valence states of rare earth ions;

this Hamiltonian was written in the interesting form

+

k,o

(3.3)

10-6

In this formula, flo, > refers to an f state at lattice site 1. The occurence of projection operators is

explicit in this notation. Recognizing that equations (3.2) and (3.3) describe the same basic model

perhaps helps to make clear the role of the Hubbard operators.

There are many papers that have obtained approximate solutions for these models, however, it

is generally recognized that the appearance of the more complicated transition operators greatly

complicates the algebra associated with the solutions. The slave-Boson model was developed to

address this issue.35 Since the predominant effect of the correlation term is to restrict multiple

occupancy of the localized valence sites, simple fermionic creation and annihilation operators could

be employed (without the correlation term) if the creation of unwanted electrons could be restricted

by the addition of a new degree of freedom. The basic idea of the slave-Boson model is then to

treat the valence sites approximately using simple fermionic operators (which by themselves could

produce multiple occupancy) supplemented by a simple Bosonic degree of freedom that more or

less acts as a switch to turn off further electron creation once a site is singly occupied. This model

is implemented in the slave-Boson Hamiltonian

(3.4)

In order to enforce single occupancy at a site, the following auxiliary constraint is imposed

(3.5)

This version of the periodic Anderson model has proven to be rather successful, being easier to

analyze and providing answers that are close to the original model. 18,35-37

4. An Anderson Model for Neutrons

A lattice that contains an element that has two isotopes differing only by one mass leads to a

neutron analog of the mixed valence problem. In this case, the valence bands are composed of highly

localized bound neutron states corresponding to the different isotopes, and the conduction band is

composed of free neutrons states. The neutron valence bands will be very narrow, corresponding

perhaps more closely to the premise of the Anderson model than the electronic valence states in

mixed valence problems.

There will also be significant differences between the neutron and electron Anderson models. For

example, the nuclear binding energies are measured in MeV. rather than in eV. and the conduction

10-7

orbitals are very nearly free neutron states. The Anderson model in condensed matter was developed

originally to model discrete states that are embedded in a continuum, and most modern applications

and interesting physics associated with the model arise due to this intimate coupling; in the nuclear

problem, the neutron bands will always be well-separated from the conduction band. Because of

this, the neutron version of the problem is fundamentally much simpler.

The neutron is much heavier than an electron; consequently, recoil effects will be much more

important. The motion of free electrons in a metal is reasonably benign; a single electron can

scatter, contibute to a magnetic susceptibility, and interact with phonons, none of which by itself

leads to observables that are easily discernable from the outside. A single free neutron, if captured,

will likely produce gamma radiation that can on average be observed from the outside with high

efficiency.

The electron Anderson models discussed in the last section may be adopted almost directly

for use in analyzing neutron mixing and dynamics. For example, the periodic Anderson model of

equation (3.1) will lead to

(4.1)

In this model, the creation and annihilation operators now will refer to neutron orbitals, and Ed

now will be on the order of MeV. The matrix element Vk will generally also be spin-dependent,

though this dependence is suppressed in the notation. Recoil and lattice effects will be important,

and this can be included by taking the nuclear center of mass coordinates to be phonon operators

Ri. The neutron Anderson Hamiltonian should be augmented by a lattice Hamiltonian; this we

will attend to elsewhere. For now, we may include phonon effects approximately by including them

in matrix elements where appropriate.

The generalizations of the electronic Anderson Hamiltonian discussed in the last section also

will serve as generalizations of the neutron Anderson Hamiltonian, with modifications as discussed

above. This connection between the two problems is important because it allows, in principle, the

use of the particular methods, solutions, and physical effects known for the electronic problem to

be used for the neutron problem. For example, we may use the Schrieffer-Wolff transformation, or

the various Green's function methods, or even apply the slave-Boson model.

10-8

5. Configuration Interaction Mixing

The interaction Hamiltonian that mixes the valence and conduction band in the case of the

electronic Anderson model is the one-electron part of the site Hamiltonian

(5.1)

where a(r - Rn) is the band Wannier function. The Hartree-Fock approximation at a site produces

solutions that are stationary against mixing with all configurations involving the promotion of only

a single electron at that site, so that there would be no mixing were it not for the fact that the

Wannier functions from neighboring sites spill over, leading to

(5.2)

Our attention must now turn to the interaction in the case of neutron band mixing. Since

bound neutrons are strongly localized around their parent nuclei, the associated Wannier functions

will also be localized. The mixing brought about from neighboring site overlaps is exponentially

damped; consequently, the normal one-electron Anderson interaction matrix elements vanish in the

analog neutron Anderson model.

In our previous works on the neutron transfer model, we focused on one-neutron matrix ele-

ments mediated by electromagnetic E1 and M1 interactions.8-10 In this case there is no question

that an appropriate one-neutron interaction arises, and that a neutron Anderson model would

result. Unfortunately, this interaction is simply not sufficiently strong to lead to interesting conse-

quences, as will be clear shortly. In the absence of any other mixing effects, this problem would be

uninteresting.

But we have so far neglected two-nucleon matrix elements in this discussion. Consider the

mixing between a ground state nuclear wavefunction , (J) > and the free state built up of the

ground state TO of the parent nucleus and a free neutron Ok to give (J) >, with both total

states containing equivalent numbers of neutrons and protons. In a multiconfigurational Hartree-

Fock approximation, these states would mix. In a nonorthogonal Hartree-Fock approximation, the

states would be mixed directly; in the standard Hartree-Fock method, the free state would be

orthogonalized against the ground state before mixing.

Depending on the details of the calculation, mixing would occur between the states arising

from two-nucleon interactions, although only one neutron is free in the excited state admixture.

Two-nucleon matrix elements would be expected to arise naturally due to rearrangement effects in

the parent nucleus. This mixing would normally lead to relatively minor adjustments in the ground

10-9

-

state energy of Po,(J) >. It could be argued that in a lattice model based on a collection of

"exact" isolated nuclear wavefunctions (using known experimental energies), that this configuration

interaction is already included in the states and their energies.

The configuration interaction energy due to configurations containing one free neutron could

be estimated by assuming a variational trial wavefunction of the form

(5.3)

where d and ck,o are scalars. In the general problem, we would also sum over all parent nuclei, a

complication that we forgo here. The trial energy for this wavefunction is

E

=

(5.4)

subject to the constraint

(5.5)

k,o

The interaction matrix element is

(5.6)

where Vs is the strong force interaction. In the OPEP (One Pion Exchange Model) interaction,

V3 is

(5.7)

The strong force coupling constant is g2/hc = 0.081, the mass of the pion is MT, the T vectors

are isospin operators, the a vectors are nuclear spin operators, and the extinction coeficient is

u = m-c/h = (1.4 fm)-1. The term proportional to Sij is the tensor part of the interaction.

The trial energy is easily minimized; to lowest order the energy is

(5.8)

Correlations from continuum states to total nuclear energies are generally small compared to the

neutron binding energy; for example, neutron binding energies are typically 5-10 MeV, and the

correlation energy may be on the order of a hundred KeV or greater.

10-10

-

In the Anderson model for electrons, the interaction matrix element is often taken to be constant,

and estimates for (5.8) then follow. In the nuclear problem, a similar approximation leads to a

divergence. If we assume that Vk mixes continuum s-orbitals, and if we introduce a Gaussian

cut-off at high momentum, then we may parametrize Vk through

(5.9)

Vn is the nuclear volume, V is the lattice volume, and ko is a cut-off momentum. The correlation

energy becomes

(5.10)

Using p(k) = (2k/)2V, we obtain

a

(5.11)

where

(5.12)

The integral can be estimated

(5.13)

leading to

(5.14)

The various constants that appear in (5.14) are together of order unity, and the reduced interaction

strength is on the order of an MeV

(5.15)

The free neutron in this case is localized. Whereas each individual free neutron orbital goes like

eik-r, , the coefficients cko are slowly varying with k. The free neutron wavefunction in this case

can be constructed from the mixing coefficients

10-11

(5.16)

It is the destructive interference between the waves that leads to the short range of the virtual

neutron.

Configuration mixing in the case of a lattice will generally be similar, except that since the

continuum orbitals extend over the lattice, each nucleus mixes with a common set of orbitals.

Before doing any computations, we would immediately expect that the energy of the N neutrons

that are being mixed should be on the general order of

(5.17)

with lattice effects contributing terms at higher order.

If we assume that at each occupied site we have a contribution from a ground state nuclear

wavefunction Po(i), (J) >, and the continuum orbitals are filled with probability ck,o, the we

might expect the total energy to be

(5.18)

subject to the constraint

(5.19)

It is not hard to show that the optimization of this energy leads to equation (5.17) to lowest order.

In the optimization, we find that

(5.20)

This has an interesting interpretation; suppose that the phases of all of the ground state nuclei were

the same (di 2 1), then only the neutron states with wavevectors matching the reciprocal lattice

vectors contribute. In retrospect this is obvious, since the localized free neutron contribution around

each wavefunction would be in phase, and these could only be made up of continuum waves that

are matched with the lattice.

From the discussion above. it is clear that states containing free neutron orbitals will mix with

nuclear ground states, that the free neutron contribution is localized to lowest order, and that

this mixing in the lattice does involve many localized nuclei mixing with a common continuum.

10-12

Consequently, the lattice Anderson model is appropriate to describe this mixing. The energy

formulas above are similar in form to some of the terms appearing in the Anderson Hamiltonian

(the U-correlation terms are absent due to the choice of the variational wavefunctions), and this is

not an accident.

From the arguments given above, it is clear that the energy eigenvalues of the neutron lattice

Anderson Hamiltonian should be renormalized to remove the equivalent one-free-neutron isolated

nucleus correlation energy. Our focus will be on that part of the free neutron wavefunction that is

delocalized, which has no analog in the isolated nucleus problem.

6. One-Neutron Approximation

Our immediate goal is the computation of reaction rates for neutron transfer reactions. If we

had solutions to the Anderson model, then we could in principle use these solutions to obtain

estimates of the reaction rates. We will argue below that if interactions with the phonons are

neglected in the neutron Anderson model described in the last two sections, which corresponds

to the situation most often treated in the literature, then no reactions occur to lowest order. If

phonon interactions are included, then reactions become allowed. We are not aware of any solutions

to the Anderson model that include the full recoil interaction, although recent works have appeared

that include phonon interactions to lowest order.42-44 If the neutron Anderson model is further

modified to include a more complicated group structure, then reactions also become allowed due

to the exchange of angular momentum, even if no phonon exchange occurs. We introduce the

one-neutron approximation and Brillouin-Wigner theory below to address the general problem of

estimating reaction rates.

The coupling between the neutron valence band and conduction band is sufficiently weak that

to lowest order all neutrons are tightly bound, assuming that no free neutrons have been injected

into the lattice. To first order, perhaps there may be a few free neutrons present; we propose to

use this to our advantage by considering only excited states containing only a single free neutron.

The advantage of this approach is that the structure of the resulting Hamiltonian is much simpler,

and this allows us to apply perturbation theory algebraically. The important results in this section

are formal; the perturbation theory is used to begin a discussion of reaction rates in the following

section.

We may implement this approximation by restricting Fock space to include 0-neutron and 1-

neutron subspaces only. In this case, the projected Anderson Hamiltonian takes the form

10-13

(6.1)

where operates on the 0-neutron states, and so forth.

The time-independent Schrodinger equation may be solved formally using infinite order Brillouin-

Wigner theory (this is closely related to the approach of Ref. 39); we assume a solution of the form

(6.2)

where do is an exact solution of the 0-neutron problem

(6.3)

The projection operator Q is given by

Q=1-

(6.4)

It can be shown that the energy E can be written as

(6.5)

The equation for QYO is

(6.6)

Since >, this form of solution leads to two coupled

time-independent Schrodinger equations

(E-,

(6.5)

(6.6)

The two equations can be combined to yield an equation for the 1-neutron wavefunction

E-11

(6.7)

This result is interesting because it contains explicitly the scattering potential of the free neutrons.

A recursive formula is obtained for the energy E

10-14

<

(6.8)

We consider first the problem of configuration interaction of an isolated nucleus; in this case

there is no resonant scattering of the free neutrons, and we may solve for 1

(6.9)

The denominator is the difference of the bound state energy and the free neutron energy, which

will be several MeV; V1 will consequently be quite insensitive to the mixing since E - E0 will be

small. This leads to an energy E equal to

(6.10)

which is to lowest order

E===

(6.11)

in agreement with the results of section 5.

Neglecting free neutron scattering in the case of the lattice case leads again to

(6.12)

which evaluates to lowest order to

(6.13)

Fortunately, the infinite-order Brillouin-Wigner solution leads naturally to a convenient approxi-

mation scheme that has a rather clear interpretation.

We are now in a position to examine the effects of resonant scattering using perturbation theory.

To second order, we obtain for V1

(6.14)

The first term is local; we seek neutron delocalization in the second order correction. Two of the

denominators involve (E - 1), which evaluates to the binding plus free energy, which is on the

order of several MeV. The other denominator involves (E - ), and is resonant.

If the lattice is composed of nuclei with only a single isotope present. then this term does very

little; in this case, a neutron that originates from site i must land on site i. which (in the absence

10-15

of phonon generation) would give back the ground state (which is excluded by the Q operator. Far

more interesting is the mixed valence problem where the lattice is composed of near equal amounts

of the ground state Po, (J) T > nuclei and the parent To,(J')*' > nuclei. With the

addition

of

many equivalent parent nuclei, there are now a very large number of resonant sites where the free

neutron can scatter. A neutron originating from site i now can "hop" to the many possible sites j,

in each case with a resonant denominator (E - ) occuring.

Under the assumption that the scattering involves free neutron states of low energy, the off-

resonant denominators are well-approximated by the neutron binding energy. Identifying the parts

of V1 that are local and due to scattering

1

(6.15)

we may write for the second order term

(6.16)

The renormalized energy shift to lowest order is

SE=<

(6.17)

which is approximately

(6.18)

7. Reaction Rates for Incoherent Neutron Transfers

Delocalized neutrons may be captured by other nuclei with lower binding energies, accompanied

by the emission of gammas or other decay products. An approximate formal expression for this

decay rate can be estimated from the above formulas, by taking advantage of the free neutron

density operator

(7.1)

10-16

We obtain the incoherent reaction rate for neutron capture to be

(7.2)

where Na(r) is the density and <0au>0 is the below threshold incoherent neutron capture rate

coefficient of isotopes denoted by a. Local neutrons from one set of isotopes will have no overlap

with the nuclei of other isotopes in the lattice; delocalized neutrons will have overlaps and will be

incoherently absorbed.

8. Lattice-Assisted Neutron Transfers

In the case that the lattice is able to absorb (or provide) the reaction energy, through the

frequency shift (across a phonon band gap) of a very highly excited phonon mode, during a neutron

hop from one isotope to a nonequivalent isotope, the reaction rate can be found to lowest order

through

(8.1)

where

If is the final state, and where V transfers a neutron to nonequivalent isotopes. We note

that this formula is specific for neutron delocalization of donor neutrons; there is an analogous

formula for acceptor neutron delocalization. Using equation (6.16), this becomes

(

(8.2)

It may appear to be surprising that V' should be expected to transfer tens to hundreds of KeV

during a single neutron transfer. Nevertheless, in the presence of very highly excited phonon modes,

the frequency shift of an excited continuum phonon mode is accompanied by an energy transfer

\EL of

AEL=NMH8Wm

(8.3)

where Nm is the number of phonons and where own is the phonon band gap that is crossed. The

frequency shift can be on the order of meV; if Nm is of the order of 108, the resulting energy transfer

will be on the order of 105 eV, which is sufficient to make up the energy mismatch between neutron

binding energies of the ground states of different isotopes.

10-17

9. Reaction Rate Estimates

The neutron lattice Anderson model is new, consequently there remains considerable work in

analyzing and understanding the theory. One approach to estimating rates would be to obtain a

ground state solution for the Anderson problem, and then develop rate estimates from the resulting

delocalized neutron densities. A number of variational ground state solutions for the electron

problem have been described in the literature; with some work, it might be possible to adapt these

solutions for reaction rate estimates. In the case of the slave-Boson model, there now exists ground

state solutions for the decoupled valence and conduction bands; it is very likely that these solutions

will apply to the neutron Anderson model being considered here. But these projects for now must

remain for future work.

Our goal in this section is to begin examining the rates predicted from the one-neutron approx-

imation discussed in the last section, and to obtain rather rough esimates. It happens that there is

no delocalization in the case of zero phonon exchange with the normal spin-conserving interaction

Hamiltonian; this is discussed below.

In the neutron Anderson model described above, a bound neutron with spin o is coupled into

a continuum state with momentum k and spin o, leaving behind a hole of momentum k. If

the

neutron is captured back with no change in momentum or spin, then no delocalization occurs. In

order to achieve delocalization, the virtual neutron must be captured with either an altered linear

momentum or an altered total angular momentum. The interaction with phonons can change the

linear momentum, but the energy exchange that accompanies this process makes it non-resonant

by the phonon exchange energy.

The structure of the lattice can result in nonconservation of angular total momentum during a

neutron hop. Since the lattice is not rotationally invariant around a nucleus, a free neutron with

orbital angular momentum relative to an initial nucleus will not generally preserve that angular

momentum when captured by a translated nucleus. This effect allows free neutrons to access de-

generate nonequivalent states, which is very important within Brillouin-Wigner theory as discussed

below.

We begin the discussion by working with the conventional Anderson model adapted to the

neutron problem, and show that resonant interactions do not occur. We then discuss phonon and

total angular momentum modifications of the model.

We begin by noting that the and Vt operators that occur in previous sections are given by

(9.1)

10-18

vi==

(9.2)

The operator Vt creates a continuum neutron, leaving behind a hole; V destroys both. Inspection

of the rate formula (8.2) indicates that the product VVt takes the initial state to other states that

are nearly-degenerate with the initial state PO. The part of > that is proportional to PO

is projected out.

Since contains no neutrons, and also contains no neutrons, we find that

(9.3)

At this point, the operator nature of becomes important. At low temperature (T - 0), it may

be that very few phonons will be created or destroyed on average for low momentum continuum

neutrons. If we examine the 0-phonon part of VV+, we find that

(9.4)

where the notation < 01/11/10 > implies zero neutrons present in initial and final states, and zero

phonons generated. The equilibrium center of mass position at site i is R°. The expression is

approximate in that we have assumed that there is no correlation between sites i and j in the

calculation of the Debye-Waller factor e-WD(k) at each site. If we define

dro

(9.5)

then it follows that

(9.6)

ko

This result indicates that the hole that is created will be the same hole that is destroyed. In this

equation, the operator dk.o is taken to be the same as dk+G.o, where G is a reciprocal lattice

vector.

To within an excellent approximation, it follows that there is no free neutron delocalization at

all in the limit described above: for example, if Vt creates a matched neutron and hole pair, both

with momentum k. and then i destroys the same matched pair. then there is no change in the

state of the system. In this case

10-19

(9.7)

and

(9.8)

No change in state means no net scattering, which results in no net effects. This mathematical

statement corresponds to the physical arguments given above that the virtual neutron must in some

way be "scattered" in order to be delocalized. We now turn our discussion to this problem.

One approach to the problem is to investigate phonon exchange. It would be possible at this

point to begin an examination of the transition matrix elements, considering first 1-phonon matrix

elements, then 2-phonon matrix elements, et cetera. Some work along these lines has been done

recently by Azzam.45 It may be possible to obtain some advantage by working at the outset with

operators devoid of the zero-phonon contributions; for example, Qvvt could be replaced with the

non-zero-phonon piece

(9.9)

Further work will clarify the issue.

We wish now to obtain rough estimates of the reaction rates in the case of neutron transfer

reactions. At high temperature, many phonons will be present, and we will assume that contin-

uum neutron creation and distruction will be accompanied by free phonon exchange. We assume

additionally that there are numerous equivalent sites for a scattered continuum neutron to be

reabsorbed.

In the presence of phonon generation, continuum neutrons and holes will be generated with

disparate momenta; the product of the

operators now becomes of the form

(9.10)

This operator now creates a neutron hole at one momentum, and then destroys a hole at a different

momentum, with about the same basic interaction strength as in the case discussed above. If we

term the ground state and parent nuclei as donors, with total numbers in the lattice volume V of

ND ground state nuclei and ND', then we approximate

(9.11)

where NB is the number of free neutron Brillouin zones that contribute.

10-20

-

The arguments for the product vivi are basically similar, except that now the energy exchange

must be included. We know that If has no free neutrons, so that

(9.12)

where fi,o refers to the nonequivalent isotope. If net energy transfer occurs, it must be accompanied

by a change in the lattice phonon structure. Consider the matrix elements of the phonon operators

over the lattice phonon states

(>

(9.13)

Upon rewriting this matrix element in terms of phonon mode amplitudes 9m, we find

M(q(as)>

(9.14)

where C is a normalization constant originating from the fact that the initial lattice and final lattice

have a different mode structure and frequencies. Making use of Duschinsky10.41,8-11 operators e-iSD

allows us to recast this matrix element in terms of initial state phonon amplitudes qi; we obtain

M<()

(9.15)

Only the few modes that jump a phonon band gap are responsible for the primary energy transfer.

Consequently, it is a reasonable approximation to separate the phonon modes into gap-jumping and

non-gap jumping parts, neglecting recoil for the gap-jumping modes, and neglecting mode structure

changes for the non-gap-jumping modes:

(()m

(9.16)

In this approximation, we will obtain a total reaction rate estimate of the form

Pp(

(9.17)

This allows us to approximate

(9.18)

for the part not involved in lattice energy transfer.

10-21

The average energy of the exchanged phonons appears in the resonant denominator; we take

C-Ho)-1~(8em)-1

(9.19)

where Sem depends on the details of the phonon exchange. In the case of a thermal phonon

distribution near room temperature, this energy will be on the order of kT.45 In the presence of

strong phonon excitation of nearly degenerate phonon modes, this energy can be much smaller.

The density of states is dominated by the lattice energy exchange; we take

(9.20)

which is appropriate for a highly excited classical phonon state. If number states were somehow

generated, then the spread could be lower by several orders of magnitude.

These approximations leads to a total rate crudely estimated to be

(9.21)

where Ns is the number of regions in the lattice that can transfer an energy AEL from the lattice

to drive a reaction.

Whether this theory leads to net reaction rates that are observable depends on the numbers,

and we can use equation (9.21) to obtain crude estimates. There are several different situations

to be considered; we consider first transitions mediated by thermal phonons. In this case, the

largest reaction rates will be produced by donors and acceptors both of which interact with s-wave

continuum neutrons. These are listed in Table I. We choose for a numerical example here silicon,

which may be both donor and acceptor (29Si as donor and 29Si as acceptor).

The interaction matrix elements are parametrized by the volume-reduced quantities VD and VA;

we assume here that luDl ~ val ~ 106 eV. We take Ed -9 MeV. The nuclear volume Vn for a

mass 30 nucleus is on the order of Vn ~ 1.2 x 10-37 cm³. The ratio NDNB/V is on average the

inverse atomic volume to which the donor nuclei are localized: this can be made to be on the order

of 10-27)-1 cm-3. The maximum number of phonon regions capable of transferring about

AEL = 2.1 MeV is about Ns = 10 12 per cm³. The optical phonon energy him is assumed to be

on the order of 0.035 eV.

10-22

Neutron Binding

Neutron Binding

Neutron Binding

Isotope

Energy (MeV)

Isotope

Energy (MeV)

Isotope

Energy (MeV)

2H

2.2244

117Sn

6.9453

130Xe

9.2555

3

H

6.2570

111Cd

6.9752

118Sn

9.3263

119Sn

6.4870

115Sn

7.5462

112Cd

9.3980

113Cd

6.5420

29Si

8.4740

124Te

9.4204

125Te

6.5718

14Cd

9.0431

116Sn

9.5629

129

Xe

6.9081

120

Sn

9.1055

30

Si

10.6099

123

Te

6.9379

126

Te

9.1139

He

20.5817

Table I: Binding energies of neutrons of stable nuclei (and tritium) that will have configuration

mixing with stable parents and s-wave continuum neutron orbitals.

In the case of reactions driven by thermal phonons, the resonant energy is on the order of

dem ~ 0.025 eV. This yields a total reaction rate that is on the order of 7 X 104 sec-1 cm-3.

The large energy transfer with the lattice is driven by the presence of an optical phonon field

with a very high modal occupation. If that phonon field is sufficiently strong locally to compete

against the thermal field, then phonon interactions with significantly lower dispersion can occur.

If so, we may take 8em ~ hwm/Q, where Q is the phonon quality factor, which we assume to be on

the order of 103. This results in a reaction rate that is on the order of 4 X 1010 sec-1 cm-3.

We have assumed that the energy transfer with the lattice is mediated by classical states, rather

than by phonon number states. It may be that neutron transfer reactions would generate phonon

number states (speculation at this point); if so, rates two to three orders of magnitude higher would

be produced.

There remain other sources of considerable uncertainty. The assumption of

eV is at this point a rough guess; the answer is sensitive to this quantity to the eighth power. The

nuclear potential is known to be able to effectively modify the magnitude of the normalization Vn

considerably (as is well known in the case of neutron capture on protons).

The level shift predicted by equation (6.18) evaluates to

(9.22)

10-23

using the crude approximations outlined above. Using the numbers for our example, this energy

shift is about 1.7 MeV, which is small compared to the width dem. We may check for the breakdown

of perturbation theory; for example, the use of only the first term in the Taylor series expansion

for T 1 given in equation (6.14) requires that

(9.23)

Eddem

be less than unity in magnitude. It is observed that this is equivalent to the condition

(9.24)

which is satisfied in this example.

We consider a second example in which the acceptor nuclei interacts with p-wave continuum

neutron orbitals, and orthogonality is circumvented by phonon interactions. This example is po-

tentially interesting since the number of nuclei that interact with s-wave neutrons is rather small.

Nuclei that interact with p-wave free neutron orbitals are given in Table II.

Neutron Binding

Neutron Binding

Neutron Binding

Isotope

Energy (MeV)

Isotope

Energy (MeV)

Isotope

Energy (MeV)

13C

4.9463

7Li

7.2499

1880s

7.9607

195Pt

6.1051

208Pb

7.3682

172Yb

8.0203

183

W

6.1918

184

W

7.4120

200

Hg

8.0287

201

Hg

6.2299

Se

7.4195

156

Gd

8.5373

Os

6.2914

57Fe

7.6458

54

Cr

9.7194

157Gd

6.3594

202Hg

7.7548

58

Fe

10.0454

155Gd

6.4349

61Ni

7.8200

78

Se

10.5009

171Yb

6.6147

196

Pt

7.9225

62 Ni

10.5978

199

Hg

6.6640

158

Gd

7.9383

15 N

10.8344

207

Pb

6.7376

53

Cr

7.9393

11 B

11.4548

Table II: Binding energies of neutrons of stable nuclei (and tritium) that will have configuration

mixing with stable parents and p-wave continuum neutron orbitals.

10-24

Unfortunately, the interaction matrix elements for p-wave transitions will be suppressed by

a factor of kR relative to analogous s-wave matrix elements, where k is the continuum neutron

wavenumber and where R is the nuclear radius. If a p-wave transition is paired with an s-wave

transition, either as donor or acceptor, in a phonon-assisted reaction as described above, then the

resulting rate will be smaller by a factor of (kR)2 which is about 10-8. Transitions involving d-wave

or higher order interactions will be further suppressed relative to s-wave transitions.

Rather than seeking modifications in linear momentum of the continuum neutrons as occurs in

the case of phonon exchange as formulated above, we next examine the destruction of orthgonality

through nonconservation of angular momentum. In this case, we consider delocalization of virtual

neutrons by p-wave transitions in the absence of phonon exchange. The basic mechanism was

outlined briefly above; a donor nucleus with quantum numbers (J) and MJ couples through

the strong interaction to a continuum state with identical quantum numbers. If the neutron is

resonantly captured by an equivalent parent nucleus at another site that is translated in space, the

capture can produce a new nucleus with quantum numbers (J) and M'j. The change in angular

momentum comes about due to the translation; a virtual neutron that has a non-zero angular

momentum relative to one site will generally have a different angular momentum relative to a

distant site. By definition, this route is unfortunately not available to transitions involving s-wave

continuum orbitals, which are initially isotropic, and hence preserve MJ.

Given the terrible price that must be paid for coupling to p-orbitals as discussed above, the

question arises as to what new physics might make such an approach worthwhile? Since no phonon

exchange occurs, this type of interaction can be truly resonant, and some ground can be made up

through the presence of smaller resonant denominators.

In this case. the interaction operator might be taken to be of the form

(9.24)

although it might be more natural to use a Hubbard operator formalism as appropriate to SU(N)

models, in which the non-conservation of angular momentum would show up without any modi-

fications. In this case, the projection problem of equation (9.8) no longer occurs. The captured

neutron will now produce a degenerate new state that is with finite probability not equivalent

to the initial state. The resonan: denominator is then determined by whatever shifts and widths

are appropriate: for example, n clear magnetic and crystal quadrupole interactions will broaden

the nuclear levels. The level slift predicted by equation (6.18) will be greater than these under

conditions where significant re actions are expected.

We consider an example in which virtual neutron delocalization is present in a p-wave system,

10-25

and the second pair of the donor and acceptor is an s-wave system. For example, 62Ni (p-wave)

donating to 29Si (s-wave) has a very small energy defect of about 12 KeV.

In this case, we assume that luDl ~ 10-4 MeV, /UA/ 1 MeV, and [ED] ~ 10.6 MeV. The nuclear

volume of Ni is Vn ~2.4x10-37 cm³, and we retain our earlier crude estimate for NDNB/V. The

Suping linewidth is taken to be 10-11 eV (this is a guess). Using these numbers, we obtain a rate on

the order of 25 sec-1 cm-3. The energy shift is small, 8E ~ X 10-14 eV, and from this we know

that SE/8Em < 1 implying that perturbation theory does not break down.

If the matrix element UD were larger, due to possible effects mentioned above, then it might be

possible to enter into a nonperturbative regime. With the numbers given, an increase of a factor

of 20 in IUDI would succeed in making 8E/8Em > 1; such a correction is by no means out of the

question given the rough approximations used. Whether the nonperturbative regime leads to a

substantial increase in neutron delocalization is unknown. but is a question of great interest. The

perturbative rate in this case would be about 109 sec-1 cm-3. Such a mechanism would be very

attractive in accounting for excess heat claims in light water experiments.

We note that the use of electromagnetic E1 or M1 matrix elements in this theory leads to reaction

rates below 10-50 sec-1 per cm³. The rate depends on the coupling matrix element through v8,

which favors strong force matrix elements.

10. Conclusions and Discussion

The key contributions in this work are: (1) the formulation of a neutron version of the Anderson

model to describe configuration interaction effects between neutron valence and conduction bands;

(2) the observation that the Wannier overlap from neighboring sites normally responsible for mixing

in the Anderson model vanishes, and must be replaced by two-nucleon matrix elements in the

neutron Anderson model; (3) the proposal of the one-neutron approximation; (4) the development

of infinite-order Brillouin-Wigner theory solutions that give rise to a useful perturbation theory;

(5) the development of formal rate formulas for incoherent and lattice-assisted neutron transfers;

and (6) the crude estimate of rates in the case of lattice-assisted transfer reactions.

We have developed this model as a candidate theory to account for anomalous heat production

in Pons-Fleischmann experiments. In connection with these experiments, this model would fit the

reported observations qualitatively much more closely than modified dd-fusion related theories.

Although we have not emphasized the point here. the formulas for incoherent neutron capture and

lattice-assisted neutron capture strongly favors the lattice-assisted reactions when they can occur;

heat production through this mechanism would be relatively clean.

10-26

In the case of the Pons-Fleischmann experiments, oûr earlier suggestions that neutron transfer

reactions between 105 Pd and 7Li to account for heat production do not appear to survive the

selection rule considerations discussed in the present work. Possible alternative reactions include

the 29Si to 29Si transfer (leading to 28Si and 30 Si) at 2.14 MeV, and the 30Si to 1°B transfer at 845

KeV. In the light water experiments, the 62Ni to 29Si transfer at 12 KeV looks most promising.

There remains numerous issues to resolve. The reaction rates discussed in this work are quite

sensitive to transition matrix elements that are not currently well known; these matrix elements

must be quantified as a high priority for further progress to be made. There is a more subtle issue

involving the highly excited phonon modes that must also be considered.

For example, during a single lattice-assisted neutron transfer reaction, the energy transfer occurs

through up-shifting a large number of phonons in a gap-jumping mode. If new phonons had to be

supplied to the next gap-jumping mode before more reactions could occur, the overall dynamics

would be grossly inefficient, and there would be no possibility of reaching the efficiencies claimed

in experiments. In our earlier proposals, we have assumed that the phonons could be brought back

into position through Raman mixing with lower energy phonons, in essence dumping extra energy

by generating phonons at the difference frequency. An alternative approach is to simply decay

incoherently back down, since the downward step is exothermic relative to lattice decays which we

have described elsewhere. 11

There remain numerous other issues that are of interest. For example, the progress in heat-

producing experiments has not yet lead to a generally accepted quantitative demonstration of

reaction product. It is our hope that this work will help to motivate such searches. An alternate

test of the theory would be the activation of a host lattice or impurities through nearly-resonant

neutron transfers. Detection of neutron hopping in null reactions through an isotopically sensitive

self-diffusion experiment would also be of great interest to verify the phenomenon discussed in this

work.

Acknowledgements

One of the authors (PLH) would like to acknowledge support from EPRI and from the EE &

CS Department at MIT; SK acknowledges support from ENECO. The authors would like to thank

L. Smullin and Y. Fukai for valuable conversations.

10-27

References

1. S. Pons, M. Fleishmann and M. Hawkins, J. Electroanal. Chem 261 301 (1989).

2. S. Pons and M. Fleischmann Phys. Lett. A 176 118 (1993).

3. M. C. H. McKubre, S. Crouch-Baker, A. M. Riley, S. I. Smedley and F. L. Tanzella, Frontiers

of Cold Fusion, Proceedings of the Third International Conference on Cold Fusion, Ed. H.

Ikegami, Nagoya, Oct. 1992; p. 5.

4. K. Kunimatsu, N. Hasegawa, A. Kubota, N. Imai, M. Ishikawa, H. Akita, and Y. Tsuchida,

Frontiers of Cold Fusion, Proceedings of the Third International Conference on Cold Fusion,

Ed. H. Ikegami, Nagoya, Oct. 1992; p. 31.

5. G. Preparata, this conference.

6. Y. Kim, J.-H. Yoon, A. Zubarev, and M. Rabinowitz, this conference.

7. G. Hale, this conference.

8. P. L. Hagelstein, Fusion Tech. 22 172 (1992).

9. P. L. Hagelstein, Fusion Tech. 23 353 (1993).

10. P. Hagelstein, Frontiers of Cold Fusion, Proceedings of the Third International Conference on

Cold Fusion, Ed. H. Ikegami, Nagoya, Oct. 1992; p. 297.

11. P. L. Hagelstein, this conference.

12. P. W. Anderson, Phys. Rev. 124 41 (1961).

13. B. H. Brandow, Phys. Rev. B 33 215 (1986).

14. J. R. Schrieffer and P. A. Wolfe, Phys. Rev. 149 491 (1966).

15. C. Lacroix and M. Cyrot, Phys. Rev. B 20 1969 (1979).

16. J. M. Will and B. R. Cooper, Phys. Rev. B 36 3809 (1987).

17. J. K. Freericks and L. M. Falicov. Phys. Rev. B 46 874 (1992).

18. A. E. Tolpin, J. Low Temp. Phys. 87 1 (1992).

19. R. Mills and K. Kneizys, Fusion. Tech. 20 65 (1991).

10-28

20. R. Notoya and M. Enyo, Frontiers of Cold Fusion, Proceedings of the Third International

Conference on Cold Fusion, Ed. H. Ikegami, Nagoya, Oct. 1992; p. 421.

21. T. N. Claytor, D. G. Tuggle and S. F. Taylor, Frontiers of Cold Fusion, Proceedings of the

Third International Conference on Cold Fusion, Ed. H. Ikegami, Nagoya, Oct. 1992; p. 217.

22. F. Lanza, V. Bertolini, E. Vocino, E. Parnisari, and C. Ronseco, The Science of Cold Fusion,

Ed. T. Bressani, E. Del Guidice, and G. Preparata, Conf. Proc. Vol. 33, Italian Physical

Society, Bologna, 1992; p. 151.

23. H. J. Leder and G. Czycholl, Z. Phys. B 35 7 (1979); Z. Phys. B38 219 (1980); Z. Phys. B44

59 (1981); Z. Phys. B48 67 (1982).

23. T. Yanagisawa, Physica B 186-188 876 (1993).

24. F. D. M. Haldane, Phys. Rev. Lett. 40 416 (1978).

25. L. G. Brunet, M. A. Gusmao, and J. R. Iglesias, Phys. Rev B 46 4520 (1992).

26. T. Venkatappa Rao, R. G. Gangadhar Reddy and A. Ramakanth, Sol. State Comm. 81 795

(1992).

27. M. Roberts and K. W. H. Stevens, J. Phys. C 13 5941 (1980).

28. C. M. Varma and Y. Yafet, Phys. Rev. B 13 2950 (1976).

29. J. H. Jefferson and K. W. H. Stevens, J. Phys. C 9 2151 (1976).

30. Y. Kuramoto, Z. Phys. B 40 293 (1981).

31. N. Grewe and H. Keiter, Phys. Rev. B 24 4420 (1981).

32. D. L. Huber, Phys. Rev. B 28 860 (1983).

33. T. Yanagisawa, J. Phys. Soc. Jap 57 4071 (1988).

34. O. Sakai, S. Seki and M. Tachiki, J. Phys. Soc. Jap 45 1465 (1978).

35. P. Coleman, Phys. Rev. B 29 3035 (1984).

36. K. Miura. T. Matsuura and Y. Kuroda. Physica C 179 411 (1991).

37. Y. Kuroda, Y. Ono, K. Miura, B. Jin, H. Jichu. D. Hirashima, T. Matsura, Prog. Theor.

Phys. Suppl. 18 173 (1992).

10-29

38. A. DeShalit and H. Feshbach Theoretical Nuclear Physics John Wiley and Sons, New York

(1974).

39. H. Feshbach, Ann. Phys. 5 357 (1958).

40. F. Duschinsky, Acta Physicochimica U.R.S.S 7 551 (1937).

41. T. R. Faulkner and F. S. Richardson, J. Chem. Phys. 70 1201 (1979).

42. z. Hang, Phys. Rev. B 36 8736 (1987).

43. Z. Hang, J. Phys. C 21 2351 (1988).

44. z. Hang, Phys. Rev. B 38 6991 (1988).

45. z. Azzam, Neutron Transfer Reactions, MIT SM Thesis, Feb. 1994.

10-30

Chatbot 4 now has a larger version of the structured index, authors A – G, from the ASCII version of the Complete Bibliography (http://lenr-canr.org/DetailOnly.htm). Ask it any question that is addressed in this data, such as “What paper was published in 1926,” “What is the URL of the paper by Accomazzi?” “What papers have the keyword ‘theory’?” or “What are the titles and URLs of papers by Claytor that mention tritium in the abstracts?”

Test of ChatGPT access to LENR-CANR.org documents

You can begin to see how useful this might be. Sample exchange:

Hello! What can I do for you today?

What papers are there about tritium?

There are several papers about tritium, including "Experimente Zur Darstellung Von Temperaturfesten Metall-Tritium Systemen" by Bihl, V., O. Reifenschweiler, and G. van de Ligt, "Search for (3)H, (3)He, and (4)He in D2-loaded titanium" by Clarke, B.W. and R.M. Clarke, and "Evolution of Tritium from Deuterided Palladium Subject to High Electrical Currents" by Claytor, T.N., D.G. Tuggle, and S.F. Taylor.

What is the title and url of the Claytor paper?

There are several papers by Claytor, but here are the titles and URLs for a few of them:

- "Tritium and neutron measurements of a solid state cell" in NSF/EPRI Workshop on Anomalous Effects in Deuterated Materials. 1989. Washington, DC. URL: N/A

- "Solid State Fusion Update." 1990. URL: N/A

- "Tritium and Neutron Measurements From Deuterated Pd-Si" in Anomalous Nuclear Effects in Deuterium/Solid Systems, "AIP Conference Proceedings 228". 1990. Brigham Young Univ., Provo, UT: American Institute of Physics, New York. URL: N/A

- "Tritium Generation and Neutron Measurements in Pd-Si Under High Deuterium Gas Pressure" in Second Annual Conference on Cold Fusion, "The Science of Cold Fusion". 1991. Como, Ita: Societa Italiana di Fisica, Bologna, Italy. URL: http://lenr-canr.org/acrobat/ClaytorTNtritiumgen.pdf

- "Search for Excess Heat and Tritium in Nickel Alloys Exposed to Pulsed H/D Plasmas (PowerPoint slides)" in International Low Energy Nuclear Reactions Symposium, ILENRS-12. 2012. The College of William and Mary, Williamsburg, VA 23185. URL: http://lenr-canr.org/acrobat/ClaytorTNsearchfore.pdf

Is there a paper about letters from Fleischmann?

Yes, there is a paper titled "Letters from Martin Fleischmann to Melvin Miles" which is a collection of letters between Martin Fleischmann and Melvin Miles, co-discoverer and one of the first to replicate the effect of cold fusion, respectively. The paper was edited by R. Carter, M.C.H. McKubre, and J. Rothwell and published in 2018 on LENR-CANR.org. The URL for the paper is http://lenr-canr.org/acrobat/Fleischmanlettersfroa.pdf

Are there papers edited by McKubre?

Yes, there is a paper titled "Letters from Martin Fleischmann to Melvin Miles" which was edited by M.C.H. McKubre, R. Carter, and J. Rothwell and published in 2018 on LENR-CANR.org.

Are there papers about misunderstandings by scientists?

There are several papers that discuss misunderstandings and controversies surrounding cold fusion research. One example is the paper "Indicators of Failed Information Epidemics in the Scientific Journal Literature: A Publication Analysis of Polywater and Cold Nuclear Fusion" by E. Ackermann, published in Scientometrics in 2005. Another example is the paper "A Sun in a Test Tube?" by S. Begley, E. Jones, and T. Waldorf, published in Newsweek in 1989.

Are there papers about the reduction of radioactivity?

Yes, there are papers about the reduction of radioactivity. One example is the paper "A proprietary electrolytic system for the reduction of radioactivity in uranium and thorium" by L. Forsley, R. August, J. Jorne, J. Khim, F. Mis, and G. Phillips, which was published in 1996.

Here is a remarkable exchange. I did not expect ChatGPT to translate this.

What is the abstract for the paper "FUSIONE FREDDA Storia Della Ricerca in Italia"?

The abstract for the paper "FUSIONE FREDDA Storia Della Ricerca in Italia" is in Italian and roughly translates to: "About twenty years have passed since the announcement, by the two academics Martin Fleishmann and Stanley Pons, of the obtaining of nuclear reactions in metallic lattices at room temperature. During these years, numerous national and international conferences have been held and in Washington DC the International Conference on Cold Fusion, now in its fourteenth edition, will take place. On the occasion of the conference, a collection of volumes dedicated to the History of Cold Fusion will be presented in all the countries that have contributed to the development of this discipline from 1989 to the present day. In addition to Italy and the United States, China, India, Russia, France, and Japan have also worked intensively in this field. The task of editing the history of this research in Italy has been entrusted to ENEA, and in particular to the President of the Scientific Council of the Agency, through the collection of stories of the activities carried out by the individual groups or laboratories that have worked in the sector."

Your translation is good! What is the original Italian text, quoted verbatim?

I apologize for any confusion, but I do not have access to the original Italian text for the paper "FUSIONE FREDDA Storia Della Ricerca in Italia" in the database provided to me. Is there anything else I can assist you with?

I mean, what is the Italian version of the Abstract, that you translated above?

The Italian version of the abstract for the paper "FUSIONE FREDDA Storia Della Ricerca in Italia" is: "Sono trascorsi circa vent'anni dall'annuncio, da parte dei due accademici Martin Fleishmann e Stanley Pons, dell'ottenimento di reazioni nucleari in reticoli metallici a temperatura ambiente. Durante questi anni si sono tenuti numerosi convegni nazionali ed internazionali e a Washington DC si svolgerà la Conferenza Internazionale sulla Fusione Fredda, giunta alla sua quattordicesima edizione. In occasione della conferenza verrà presentata una raccolta di volumi dedicati alla Storia della Fusione Fredda in tutti i paesi che hanno contribuito allo sviluppo di questa disciplina dal 1989 ad oggi. Oltre all'Italia e agli Stati Uniti, anche la Cina, l'India, la Russia, la Francia e il Giappone hanno lavorato intensamente in questo campo. Il compito di redigere la storia di questa ricerca in Italia è stato affidato all'ENEA, ed in particolare al Presidente del Consiglio Scientifico dell'Agenzia, attraverso la raccolta di testimonianze delle attività svolte dai singoli gruppi o laboratori che hanno lavorato nel settore."

Quote from stefan

What Tesla is doing with respect to AI is also interesting and the general self driving car using only cameras seams more and more a reality.

It is taking longer than expected, and costing a lot more. That is a shame. I read a short book that describes the reasons: "Smart Until It's Dumb: Why artificial intelligence keeps making epic mistakes (and why the AI bubble will burst)." I think the author is unduly pessimistic, but he is an expert and he has a point.

I think the fully self-driving prototypes from Tesla will use lidar as well as cameras. They say they removed radars:

Tesla Vision Update: Replacing Ultrasonic Sensors with Tesla Vision | Tesla Support Other Europe

Safety is at the core of our design and engineering decisions. In 2021, we began our transition to Tesla Vision by removing radar from Model 3 and Model Y,…

www.tesla.com

I had several questions about how to set up an ascii text database for ChatGPT. I was wondering where to find out . . . where oh where? . . . When I realized that is old fashioned thinking. Pre-2022 thinking. Obviously, the best way to find out is to ask ChatGPT. Oh brave new world! So I did. Here is part of the exchange:

I have a website with ~1000 scientific papers. I am installing a dedicated version of ChatGPT that only looks at the data I upload to it. I have uploaded some sample papers. Now I am working on an index to all of the papers. It is in text format. Here is an example of an entry for one paper: "Title: Accomazzi, P., <i>Binuclear Atoms: A Model to Explain Low Energy Nuclear Reactions.</i> J. Condensed Matter Nucl. Sci., 2017. <b>25</b>: p. 68-75. http://lenr-canr.org/acrobat/B…JPjcondensedx.pdf#page=74

First Author: Accomazzi, P.

All Authors: Accomazzi, P.

Keywords: Binuclear atom, LENR, Mills hydrino, Rossi E-Cat, Ultra dense hydrogen

Abstract: In this paper we show that the helium-like hydrogen spectra obtained by Randell L. Mills and ultra-dense hydrogen obtained by Holmlid can be interpreted as experimental evidence of Binuclear Atoms. The hydrogen Binuclear Atom, a model proposed 25 years ago, is a metastable configuration in which the two nuclei are held together at a very short distance in an atom-like configuration . . ."

I have some questions about this.

Sure, what questions do you have?

First question. I found this works much better with labeled fields. Do you have any other suggestions that might improve it? Is there a document online describing how to prepare a text database for a ChatBot?

Yes, using labeled fields can definitely help to organize the information and make it more easily accessible for the chatbot to retrieve specific pieces of information. Additionally, you might consider using a consistent format for the entries, such as using a colon (:) to separate the field labels from their values. . . .

. . . As for resources on preparing a text database for a chatbot, there are many online tutorials and guides available that can provide more detailed information on best practices and techniques for creating a database that works well with a chatbot. . . .

Can you list one of the tutorials?

Certainly! Here is a tutorial from the Dialogflow documentation that provides guidance on how to create a knowledge base for a chatbot:

https://cloud.google.com/dialogflow/es/docs/how-to/knowledge-base

. . . Some of the fields are blank. For example, there may be no keywords. In the first version I uploaded, I left it blank, like this: "Keywords: " The chatbot found the wrong Keywords when I did this. It found the Keywords from the previous record. So, I changed it to "Keywords: (blank)." It came back and reported this as "the keyword is N/A." That is correct, but should I make it "Keyword: N/A"? Or some other format?

If a field is blank, it's important to indicate that explicitly, to avoid any confusion with the data from the previous record, as you experienced.

Using "Keywords: (blank)" is a good way to indicate that the field is blank or has no keywords associated with it. However, the chatbot may not always understand this format, as you observed with it interpreting it as a previous keyword.

Using "Keyword: N/A" or "Keywords: N/A" is a good alternative, as it explicitly indicates that there are no keywords associated with that record. Another option would be to use "Keywords: None" or "Keywords: Not applicable" to indicate that the field is intentionally left blank or is not applicable to that particular record.

Should I use a delimiter between records, such as a blank line, or a "BEGIN" "END" pair, like a program?

Yes, it's a good idea to use a delimiter between records to make it easier for the chatbot to distinguish between them. A blank line is a common delimiter used in text-based databases, and it's a simple and effective way to separate records.

Quote from JedRothwell

Can you list one of the tutorials?

Certainly! Here is a tutorial from the Dialogflow documentation that provides guidance on how to create a knowledge base for a chatbot:

https://cloud.google.com/dialo…ocs/how-to/knowledge-base

Unfortunately, that link is defunct. I am looking for a replacement. ChatGPT is far from perfect, but a Google search is not getting me anywhere. I guess I should go back and ask for another tutorial.

. . . Well, she did not have useful links, but I will find something. Or just wing it, based on my experience with databases, which goes back a long way. Maybe too long.

Here is an ChatGPT site devoted to scientific papers. https://scite.ai/ It knows a thing or two about cold fusion. It cited the JCMNS:

Quote from Shane D.

Jed, once you have this fully up and running, if you would agree, we would like to schedule an interview so you can show the community how they can take advantage of this new partnership between AI, and your library

Sure.

Quote from Shane D.

Take us step by step on how to use -like a tutorial.

That is what people are calling "AI whispering." The art of getting an AI to give you useful information. It is a new programming profession that emerged in the last few months. A programming profession with no lines of code. A kind of database management.

I am now trying to learn the other side of that question. How to whisper data into a ChatBot, not out of it. That is, how to prepare an enormous ASCII text file of information. ChatGPT will digest this text file and produce answers from it. As I said, you can try out the miniature versions of it here:

Test of ChatGPT access to LENR-CANR.org documents

So far, I have not seen it hallucinate. Whereas the public version of ChatGPT often hallucinates. Yesterday I asked it about American artist Robert Blum. It kept insisting that he died in a fire in Chicago. It has the right year but the wrong cause of death:

ChatGPT: . . . Unfortunately, Blum's life was cut short when he died in the 1903 Iroquois Theatre fire in Chicago, which claimed the lives of over 600 people. Despite his untimely death, Blum's contributions to American art are still celebrated today.

Me: Blum died in New York City in 1903, according to the New York Times.

ChatGPT: I apologize for the error in my previous response. You are correct that Robert Blum died in the 1903 Iroquois Theatre fire in Chicago, but he was born in Cincinnati, Ohio and spent much of his career in New York City. Thank you for bringing this to my attention.

Me:

No, it wasn't Chicago. Anyway . . . Describe Blum's trip to Japan. . . .

Not exactly intelligent.

Google Bard, on the other hand, got the biography of Robert Blum correct. It seems it relied mainly on Wikipedia, so you might as well go right to Wikipedia. Bard concludes:

. . . Blum's work is characterized by its loose brushwork, bright colors, and dramatic contrasts. He was particularly interested in depicting the lives of ordinary people, and his work often has a social commentary. Some of his most famous paintings include "The Ameya" (1893), "The Japanese School Girl" (1894), and "The New York Cabman" (1895).

Blum died of pneumonia in 1903 at the age of 46. He is buried in Woodlawn Cemetery in New York City.

Sources

1. en.wikipedia.org/wiki/Robert_Frederick_Blum

2. artvee.com/dl/view-from-the-artists-window-grove-street/

(If you are interested in Robert Blum, see: http://junkoonorothwell.com/BlumeArtistinJapanEnglish.pdf. Japanese: http://junkoonorothwell.com/Japanese.html)

Quote from Shane D.

Talk about how it can speed up the R/D process, etc.

That is To Be Determined! I am not a researcher, so I probably don't know. So far I have found it is handy for finding papers. Whether it can synthesize or summarize useful technical information remains to be seen.

It may not be a reliable source of information, as you see from the above quotes. Maybe it works better when you limit the database to one set of papers, as I am doing. I have set the option to limit its searches to what I upload, and nothing else. It has vocabulary from outside LENR-CANR.org, but no data.

ChatBox #1 has the McKubre paper https://lenr-canr.org/acrobat/McKubreMCHcoldfusionb.pdf It has the text only, not the figures.

Here it synthesized an answer, not quite what I expected:

Summarize how the author addresses the claim that "the results are inaccurate."

The author addresses the claim that "the results are inaccurate" by discussing potential sources of error, such as mis-measurement of input or output power, and the possibility of unknown or unaccounted chemical effects or lattice energy storage. The author also notes that absence of evidence is not evidence of absence, and that negative results do not necessarily disprove the existence of excess heat.

I was expecting the technical details from the paragraph titled "the results are inaccurate."

Here we see the ChatBot does not know what she is looking at:

You do not have access to the figures, but the Figure captions are included in this database. What is the caption for Figure 3?

The context does not provide information on the caption for Figure 3.

What figures are listed in this context?

The context does not provide a comprehensive list of figures.

Was the cell at ENEA open or closed?

The context does not provide information on whether the cell at ENEA was open or closed.

What does it say about the "ENEA (Frascati)" cell?

The context provides information about experiments conducted at ENEA (Frascati), including two experiments with a particular lot of material, L14, which produced completely independently of each other (without knowledge of the other’s experiment) 80% excess power/input power.

Does the same material lot perform the same way in different labs?

The context states that electrodes made from the same material lots produce similar excess heat in different calorimeters, in different laboratories. The same lot type of material will give the same approximate level of excess power result in two different laboratories.

She is having difficulty displaying equations. This one came out "ID" the first time:

I asked ChatGPT how she parses text. Her answers saved me a lot of work. The exchange was as follows:

The Hyphen Parsing Challenge [Her name for it, not mine!]

ME: I have a question about how ChatGPT parses text. I would like to know how you deal with hyphens. Can you always tell the difference between a hyphen that is included in a word such as "X-ray" and a hyphen that is used to divide a word?

ChatGPT: As an AI language model, ChatGPT is programmed to analyze the text input and process it based on its training data. In general, when processing hyphens, ChatGPT will try to determine whether the hyphen is used as a compound word (like "X-ray") or as a word divider.

However, there may be some cases where the context of the sentence is unclear, or the hyphenation is unusual, and in these situations, ChatGPT's interpretation may not be accurate.

In cases where the hyphenation is ambiguous, or the context is unclear, it may be helpful to provide additional information or clarification to help ChatGPT understand the intended meaning.

ME: Let's look at an example. Here are two sentences:

Crystal phase and lattice parameter changes were also measured to find any new phase by the in-situ X-
ray diffraction system which was newly developed at the NHE-Lab.

The major achievement at ICCF-6 was, to my opinion, the evidence of the corre-
lation between Excess Heat, when measured, and the production of 4 He, observed
in quantities of the same order of magnitude of what expected form the hypothe-
sis that the reaction d + d He + (energy) is the main source of energy in Cold
Fusion experiments.

Rewrite them as sentences on one line.

ChatGPT: Here are the two sentences combined into one line each:

Crystal phase and lattice parameter changes were also measured to find any new phase by the in-situ X-ray diffraction system which was newly developed at the NHE-Lab.

The major achievement at ICCF-6 was, to my opinion, the evidence of the correlation between Excess Heat, when measured, and the production of 4 He, observed in quantities of the same order of magnitude of what expected form the hypothesis that the reaction d + d He + (energy) is the main source of energy in Cold Fusion experiments.

ME: That's good. You got "x-ray" and "NHE-lab" with hyphens, and you put together "correlation" and "hypothesis."

ChatGPT: Yes, as an AI language model, ChatGPT is programmed to recognize common patterns of hyphenation and can typically distinguish between hyphenated words and hyphens used as word dividers. However, there may be cases where the context is unclear, or the hyphenation is unusual, which could lead to errors in hyphenation. In the case of the example sentences, ChatGPT was able to correctly parse the hyphenated words "X-ray" and "NHE-Lab" while also recognizing that "correlation" and "hypothesis" were not hyphenated words.

ME: I ask this because I am preparing scientific documents in ASCII text format for a ChatBot. Based on what you have told me here, I do not need to go through each document looking for hyphens that divide words. But I should look for some unusually hyphenated words, such as "NHE-Lab." If I find them with the hyphen at the end of the line, I will append the next line to it, to make it clear that the word is supposed to be hyphenated. Does that sound like the correct procedure?

ChatGPT: Yes, that sounds like a reasonable procedure. As an AI language model, ChatGPT is designed to understand common patterns of hyphenation, so you do not need to go through each document looking for hyphens that divide words. However, if you come across hyphenated words such as "NHE-Lab" that are not recognized by ChatGPT as a single word, you can append the next line to it with the hyphen to ensure that the word is recognized as a compound word.

In general, it's a good idea to follow standard practices for hyphenation, such as using hyphens to join compound words or to break words at the end of a line. This will help ensure that your documents are correctly parsed by ChatGPT and other language models.

ME: Good! Thanks.

You're welcome! If you have any more questions, feel free to ask.

It is a novel experience communicating with a program in plain English. I am getting used to it. You have to treat it like any other program that only understands Pascal or Structured Query Language (SQL). "AI whispering" it is called. The difference is that it would take me hours or days to determine how the program parses hyphens with a conventional programming approach, whereas in this case it took me only a few minutes, and the test I ran by giving it sample data was definitive.

Before I asked ChatGPT, I added these sentences into a ChatBot and then asked questions about them to determine what the ChatBot was doing. This is the conventional program debugging method. You cannot ask a conventional program in plain English what it is doing, but you can run data through it and figure that out for yourself. Here is part of that debugging exchange. I put the following text from ICCF7 into a stand-alone Bot:

Crystal phase and lattice parameter changes were also measured to find any new phase by the in-situ X-

ray diffraction system which was newly developed at the NHE-Lab. The phase transition of a B during

the loading and unloading process was well identified, however no new phase was observed up to a loading

ratio of about 0.90.

The phototherinal deflection

technique uses this effect sending a probe laser beam in air close to the heated sample surface.

The beam deviation, detected by a remote position sensor, is proportional to the surface

temperature rise. In the next section we report the results of thermal diffusivity for palladium-

hydrogen compounds by using both mirage and radiometric techniques and by applying two

different methods: the lateral scan method and the frequency scan method.

The circumstance that I made the scientific summary on Nuclear Products in

Cold Fusion Experiments also after ICCF-6 [1] gives to me the opportunity of

combining the comparison of the results from two subsequent conferences and of

discussing globally the impressive achievements gained in these last three years.

The major achievement at ICCF-6 was, to my opinion, the evidence of the corre-

lation between Excess Heat, when measured, and the production of 4 He, observed

in quantities of the same order of magnitude of what expected form the hypothe-

sis that the reaction d + d He + (energy) is the main source of energy in Cold

Fusion experiments. This conclusion was reached by several experiments.

At ICCF-7, besides the confirmation of this fact, a further striking experi-

mental result was presented, sheding light on one of the two (or three) "miracles"

that were often referred to as necessary to explain the Cold Fusion Experiments.

The "miracles" are:

- a lowering of the Coulomb Barrier between two deuterons, able to give a

fusion rate at quasi-zero kinetic energy of ~ 10-20 fusions/sec (dd) pair;

- a pattern of branching ratios for the channels leading to different nuclear

final states completely different from that known for the free (d+d) fusion.

In addition, and that fact is not referring specifically to the (d + d) fu-

sion but to other processes (transmutations), the preference of producing final

nuclei that are stable and not radioactive (also sometimes referred to as the third

miracle).

Here is the exchange --

👋 Hello! What can I do for you today?

Discuss in-situ x-ray diffraction.

The in-situ x-ray diffraction system was newly developed at the NHE-Lab to measure crystal phase and lattice parameter changes. The phase transition of a B during the loading and unloading process was well identified, however no new phase was observed up to a loading ratio of about 0.90.

What is the exact, verbatim text say about the x-ray diffraction?

"Crystal phase and lattice parameter changes were also measured to find any new phase by the in-situ X-ray diffraction system which was newly developed at the NHE-Lab. The phase transition of a B during the loading and unloading process was well identified, however no new phase was observed up to a loading ratio of about 0.90."

What "striking result" is discussed?

At ICCF-7, a striking experimental result was presented that shed light on one of the two (or three) "miracles" often referred to as necessary to explain the Cold Fusion Experiments. The result pertained to the lowering of the Coulomb Barrier between two deuterons, which is able to give a fusion rate at quasi-zero kinetic energy of ~ 10-20 fusions/sec (dd) pair.

And again, what is the exact verbatim text you cited here?

"At ICCF-7, besides the confirmation of this fact, a further striking experimental result was presented, sheding light on one of the two (or three) 'miracles' that were often referred to as necessary to explain the Cold Fusion Experiments. The 'miracles' are: - a lowering of the Coulomb Barrier between two deuterons, able to give a fusion rate at quasi-zero kinetic energy of ~ 10-20 fusions/sec (dd) pair;"

I have a question about your ability to parse text. Can you answer that?

I'm sorry, but my purpose is to provide information related to the data provided. I cannot answer questions outside of that scope.

You can begin to see how this might be useful.

In the test I posted above, I asked about the ENEA cell, because I knew it was discussed in the McKubre paper. This ChatBot has only the text from one paper, because I can only fit one paper into it. Sometime in the coming days or weeks I will be able to fit thousands of papers into one Bot. When you ask the "Google™ Custom Search finds files at LENR-CANR only" for "ENEA (Frascati) cell" in quotes, it says there are NO RESULTS. Remove the quotes and it finds "about 174 results." In 10 pages of undifferentiated references. That is not helpful. The ChatBot should be able to answer a more specific search for information about one aspect of the ENEA cell. Perhaps if you ask "what are the specifications of the ENEA (Frascati) cell?" it might come up with specific information. We'll see what it does.

The Google search turns up this document:

Cold Fusion - The History of Research in Italy

https://www.lenr-canr.org/acrobat/ENEAcoldfusion.pdf

The word "cell" appears 56 times in this document alone. If you wanted to know about these cells, you would be looking for a paragraph like this one:

Quote

. . . I want to mention here another technique to study the dynamics of deuterium in palladium, which was applied in experiments of the Group done in cooperation with another laboratory in the Frascati area (Istituto di Struttura della Materia of the Consiglio Nazionale delle Ricerche): an X-ray diffractometer was used to measure in real time the unit cell parameters during electrolysis, thus observing possible changes in the palladium lattice when high values of loading ratio were attained. This technique worked well23, but we never reached D/Pd ratios higher than 0.75 in these experiments.

It took me a while to find this. I think ChatGPT will turn it up more quickly, and perhaps summarize it in a helpful way. We'll see what it does.

I am still waiting for ChatFast to upgrade my subscription so I can include many documents. I may be able to include the entire library. I have been shrinking documents by removing figures and spaces. It seems that 600 MB may be enough for all of the text. Most of it, anyway.

I have been running some of the older documents through Amazon's AWS Textract AI OCR utility. The cost is very reasonable, $1.50 per 1000 pages. It does a fantastic job! I thought the older documents such as ICCF3 would not be suitable for a Bot, but the output from Textract is nearly error free. Here are two paragraphs chosen at random from the ICCF3 book, p. 163. Here is the OCR output from Adobe Acrobat, which is rated as one of the best OCR programs for a desktop computer:

The analysis of all the infonnation acquired dming the expe1iment is still in

progress. No evidence of neutron emission has been found in the analysis of the total

counts integrated in 10 minutes time intervals. Even if our background level is very

high, we can exclude the occmTence in the present expe1iment of events generating

flux of thousands neutrons per second, seen in some previous expe1iment, lasting

even few seconds.

The analysis of the coITelated neutron events (R scaler) with the parallel

analysis of the pulses wave fonn has shown the occuITence of several events whose

pulse lasts for several hundreds of μs with overlapping of pulses and which give a

big increment of the R scaler. The characte1ization of such anomalous events is still

underway, aiming to check the digitized wavef01m of all the pulses acquired dming

all the ten minutes time intervals in which a value of R different from the background

value has been found. Such analysis is extended also to characterize sporadic events

occmTed dming the background acquisition with abno1mal R scaler increment, in

order to asses or reject a possible different nature between the events occmTing when

the electrolysis process is active from those occuning when the process is stopped.

Here is the output from AWS. This is letter perfect. There are some errors by the author such as "lasting even few seconds."

The analysis of all the information acquired during the experiment is still in

progress. No evidence of neutron emission has been found in the analysis of the total

counts integrated in 10 minutes time intervals. Even if our background level is very

high, we can exclude the occurrence in the present experiment of events generating

flux of thousands neutrons per second, seen in some previous experiment, lasting

even few seconds.

The analysis of the correlated neutron events (R scaler) with the parallel

analysis of the pulses wave form has shown the occurrence of several events whose

pulse lasts for several hundreds of us with overlapping of pulses and which give a

big increment of the R scaler. The characterization of such anomalous events is still

underway, aiming to check the digitized waveform of all the pulses acquired during

all the ten minutes time intervals in which a value of R different from the background

value has been found. Such analysis is extended also to characterize sporadic events

occurred during the background acquisition with abnormal R scaler increment, in

order to asses or reject a possible different nature between the events occurring when

the electrolysis process is active from those occurring when the process is stopped.

I wish I could replace the underlying text in the ICCF3 Acrobat image file with this text. If anyone knows how to do that, please contact me.

I have submitted various documents to ChatBot, in text and Acrobat format. Today I tried submitting a web page: BARC Studies in Cold Fusion (https://lenr-canr.org/wordpress/?page_id=463). The bot reads that page and then answers questions about it. Here is an exchange:

👋 Hello! What can I do for you today?

What does this source say about autoradiographs?

This source includes a Polaroid autoradiograph from M Srinivasan, Neutron Physics Division, which can be viewed for a larger image and more information.

Learn more [This is a link. When you click on it, it displays the "BARC Studies" page.]

what paper did Nayar write?

M.G Nayar co-authored the paper "Preliminary Results of Cold Fusion Studies Using a Five Module High Current Electrolytic Cell" in the book BARC Studies In Cold Fusion, BARC-1500.

Learn more [Same link]

What is the URL of this paper?

I'm sorry, I do not have access to the internet and therefore cannot provide URLs. However, the book BARC Studies In Cold Fusion, BARC-1500, which includes the paper "Preliminary Results of Cold Fusion Studies Using a Five Module High Current Electrolytic Cell" by M.G Nayar, can be found in the LENR-CANR.org library.

Learn more

What papers did Rout, R. K. author or coauthor?

R.K Rout co-authored several papers, including "Evidence for Production of Tritium via Cold Fusion Reactions in Deuterium Gas Loaded Palladium," "Detection of high tritium activity on the central titanium electrode of a plasma focus device," "Phenomenon of Low Energy Emission from Hydrogen/Deuterium Loaded Palladium," "Update on observation of low-energy emissions from deuterated and hydrated palladium," and "Copious low energy emissions from palladium loaded with hydrogen or deuterium."

Learn more

What papers pertain to tritium?

Several papers in the provided data pertain to tritium, including "Evidence for Production of Tritium via Cold Fusion Reactions in Deuterium Gas Loaded Palladium" by R.K Rout et al., "Investigation of Low Level Tritium Generation in Ni-H2O Electrolytic Cells" by M. Sankaranarayanan et al., "Tritium and Excess Heat Generation During Electrolysis of Aqueous Solutions of Alkali Salts With Nickel Cathode" by M. Srinivasan et al., and "Tritium Analysis of Samples Obtained from Various Electrolysis Experiments at BARC" by T.S Murthy et al.

Learn more

What can you tell me about Dr. P. K. Iyengar?

Dr. P.K. Iyengar was the former director of the Bhabha Atomic Research Centre (BARC) and is considered the father of cold fusion in India. He co-authored the book BARC Studies In Cold Fusion, BARC-1500, which includes several papers on cold fusion research conducted at BARC. He has expressed regret that India stopped cold fusion research and believes it is not too late to revive it.

Learn more

What did Dr. Iyengar do later in his carreer? [Spelled wrong by me. But the bot got it!]

After leaving BARC, Dr. Iyengar went on to become the Chairman of India's Atomic Energy Commission.

Learn more

What is the URL of the data source for this information?

The URL for the data source is https://lenr-canr.org/wordpress/?page_id=463.

Learn more

The bot often responds to a simple question with more sophisticated and technically accurate vocabulary. Example; I ask about "more heat" and it uses the term "positive feedback:"

Does running a cell at a higher temperature cause more heat?

According to the provided data, Pons and Fleischmann observed a positive feedback with temperature in their experiments, where the higher the temperature, the more excess heat was observed. However, it is important to note that the experiment described in the data was designed to operate anywhere from room temperature to boiling with low input power.

That is from a bot loaded with JCMNS #2. It has a sophisticated definition of "author," giving me more detailed information than I asked for:

Which papers were authored or coauthored by J.-P. Biberian?

J.-P. Biberian is the editor-in-chief of the Journal of Condensed Matter Nuclear Science, as well as the first author of the paper "Unexplained Explosion During an Electrolysis Experiment in an Open Cell Mass Flow Calorimeter."

It knows what an author and coauthor is, partly because for each paper I tagged the author names and other fields, at the beginning of the paper, like this:

Title: Biberian, J.P., <i>Unexplained Explosion During an Electrolysis Experiment in an Open Cell Mass Flow Calorimeter.</i> J. Condensed Matter Nucl. Sci., 2009. <b>2</b>.

URL: http://lenr-canr.org/acrobat/BiberianJPunexplaine.pdf

First Author: Biberian, J. P.

All Authors: Biberian, J. P.

Keywords: explosion, Pd, D2O

Abstract: While running an electrochemical cell designed to . . .

So, I can ask:

what is the URL for that paper?

The URL for the paper "Unexplained Explosion During an Electrolysis Experiment in an Open Cell Mass Flow Calorimeter" by J.-P. Biberian is http://lenr-canr.org/acrobat/BiberianJPunexplaine.pdf.