Replication failure is not replication

A reader recently mentioned Coolessence. As the linked web site shows

Coolescence LLC was a privately funded research company located in Boulder, Colorado. The company was originally formed to rigorously examine repeated experimental reports of so-called ‘cold fusion’ (low energy nuclear reaction – LENR), generally manifesting themselves in the form of unexpected or ‘excess’ heat, from a number of scientists around the world.  Over the past 12 years the Coolescence team has replicated the most celebrated of these experiments, with no positive results that have not been attributable to measurement artifacts or chemical effects.

This page will introduce the study of the work done by Coolessence. The relevant papers are linked below. If readers are interested, reading those papers before I study them will increase experience and comprehension. If I know anything about LENR, it is because I have studied materials in the field over and over. It’s not magic.

I was quite impressed by Coolessence, in many ways. When I was planning a tour of the U.S. and Canada in 2015, I hoped to visit them … but that trip was cancelled when my Subaru, in the first fifty miles of the trip, broke a timing belt and the engine was destroyed. So, as far as I know, I have never met the principals. Late in 2016, there were private discussions with them on the CMNS list.

From my point of view, Coolessence demonstrates how to take high risk of wasting time and money . I intend, here, to review the projects they undertook. Mostly, these would not be projects I would have chosen for first work. To be sure, this is hindsight, and it took me a few years in the field to develop perspective.

Writing several years ago, I laid out Plans A and B for LENR breakthrough. Plan A was to have Rossi (or someone like him) save us by making products appear in Home Depot, or the like.

As I pointed out, Plan A was risky, but had the benefit of not requiring Any Actual Work by anyone (other than the inventor, of course).

Given the possible importance of LENR, we needed a Plan B, and Plan B was to undertake what had been recommended by both United States Department of Energy reviews (1989 and 2004). Basic science, to nail down and confirm or clearly disconfirm earlier findings.

Plan B began with Phase I. Phase I was to confirm, ideally with increased precision, what had already been confirmed. The point was not to reinvent the wheel, but rather to start with what is far more likely to succeed. Failure is a damned nuisance, unless it leads to learning. I had identified the work showing a correlation between anomalous heat and helium production as not only rather widely confirmed, but as much more strongly probative than simple isolated findings of anomalous heat, or tritium, etc., without correlations.

Looking ahead from there, Phase II would be work to create one or more “lab rats.” I.e., protocols with adequate reliability to be readily reproducible most of the time. This could actually create a “product,” such as standardized, prepared, and pre-conditioned cathodes for FP class experiments or the like. There are indications that these cathodes could be pretested and would later work as tested.

Phase II would study already-reported and, where possible, already-confirmed protocols, not new ones. The reason is, again, to make success more likely.

Phase III would be a wide variety of investigations, using the lab rats where possible, or creating new ones. Phase III would include attempts to replicate isolated reports of interest.

Phase IV would be blue sky. By this time, if the first phases are handled (with Phase I and II being completed), there will be plenty of money for wider explorations and playing hunches, etc.

This proposal did encounter some opposition in the field, because Phase I was considered to be a waste, since “we already knew that helium was being produced.” Tonto: “What you mean, “we”?

Many of us may know this (the evidence is actually strong, though there is much room for improvement), but “we,” i.e., the human and the scientific communities, don’t have this as collective knowledge. Yet. What will it take?

The DoE reviews laid it out: replications with improved methods, published in the “journal system.” The LENR community doesn’t trust the journal system, so there you go. That’s a self-maintained trap.

In any case, Coolessence describes five projects. I will study each in dedicated pages. Here is the list:

Studies
POSSIBLE NUCLEAR REACTIONS MECHANISMS
AT GLOW DISCHARGE IN DEUTERIUM (1992)
Intensification Of Low Energy Nuclear Reactions Using Superwave Excitation (2003)
Results
Glow Discharge Loading of Pd (2007)
Update on results at Coolescence, LLC (2008)

Studies
RADIATION PRODUCED BY GLOW DISCHARGE IN DEUTERIUM (2007)
Results
Partial Replication of Storms/Scanlan Glow Discharge Radiation (2008)

Studies
Use of CR-39 in Pd/D co-deposition experiments (2007)
Characterization of tracks in CR-39 detectors obtained as a result of Pd/D Co-deposition (2009)
Results
Search for charged particle emissions resulting from Pd-D Co-Deposition (2011)

Studies
Establishment of the “Solid Fusion” reactor. (2008)
Hydrogen/deuterium adsorption property of Pd fine particle systems and heat evolution associated with Hydrogen/deuterium loading (2009)
Results
MECHANISM OF HEAT GENERATION FROM LOADING GASEOUS  HYDROGEN ISOTOPES INTO PALLADIUM NANOPARTICLES (2012)
Origin of excess heat generated during loading Pd-impregnated alumina powder with deuterium and hydrogen (2012)
Mechanisms for Heat Generation during Deuterium and Hydrogen Loading of Palladium Nanostructures (2012)
Using Bakeout to Eliminate Heat from H/D Exchange During Hydrogen Isotope Loading of Pd-impregnated Alumina Powder (2012)
Effect of temperature gradient on calorimetric measurements during gas-loading experiments (2012)
Measurement Artifacts in Gas-loading Experiments (2012)

Studies
Data from Melvin Miles’ July 2016 experiment and My Recent Kitchen Experiment (2016)
Results
Miles Summer 2016 Ridgecrest Experiment – Coolescence Analysis (2016)

Impressions

According to the classification in my Introduction, Coolessence chose what would be Phase 3 or Phase 4 projects. Given the difficulties in the field, the probability of failure was high. In reviewing this, I will be looking for behaviors and approaches that may have fostered failure. Notice: “failure” means not finding a definitive conclusion. At first glance, the Storms/Scanlan study may have been successful. The others, as far as I have seen so far, did not find the same results as the original reports, so these would be “replication failures.”

“Failure” is not defined as not confirming LENR. If an experiment confirms earlier findings, it is a successful replication, but “findings” does not mean “conclusions.” If the work is left there, fine. It’s a successful confirmation of earlier results.

If it goes on, after that, and demonstrates with controlled experiment that the original results were misleading, i.e., artifact, that is a success (and to be careful, it should, itself, be confirmed. Sometimes, historically, that step has been skipped and premature conclusions drawn). And, of course, if it nails results, eliminating possible artifacts, or increasing precision, it is also successful.

Looking for what is wrong with an experiment or analysis is not the first step. Not ever, except in one way. If one can look at anomalous results and see an obvious artifact, one may not want to put in the effort to actually confirm, and that is a reasonable personal (or organizational) choice. Ordinary skepticism is there to keep us from wasting time. Taken too far, though, it can blind us.

The most recent “Replication” is misnamed. They did not attempt to replicate Miles’ results. Rather, they analyzed his data and came to different conclusions.

Importance

Again, from my point of view, most of this work was of low value, compared to other possibilities. Gas-loading is nowhere near the center of what has been well-confirmed. Glow discharge has always been iffy (and is quite dissimilar to the original findings). There is a general fuzziness that lumps together anything that might be nuclear.

I was originally quite excited over the SPAWAR work, but the neutron results, not the charged-particle results that Coolessence studied, which have always been shaky in some ways (with replicators showing a lack of precision in defining what have been called “SPAWAR tracks). I didn’t like CR-39, it is messy and difficult to interpret, LR-115 might be much easier (but Pam Boss told me that the absorption curve for LR-115 would not be as sensitive as CR-39. Maybe.)

Remarkably, When I opened a box Coolessence sent me (they donated a large cold fusion library to Infusion Institute), stuck in the box was a plastic ziplock bag with what looks like a sheet of LR-115.

No, the neutron findings are more interesting! But still there is a huge problem. The protocols SPAWAR used do not look for excess heat. And you run this experiment for five weeks or more and then pull and develop the detectors. There is no other indication of whether or not the original effect (heat) was present. It’s a small experiment and would not be expected to produce much heat, but this makes a SPAWAR study, even if it shows a radiation effect, close to anecdotal. And from other evidence, if radiation is being produced, it’s at very low levels and has little or nothing to do with the main reaction. All it does is increase the mystery and confusion.

In my 2015 paper, I suggested further study of one protocol other than measuring the heat/helium ratio, and that was the dual-laser stimulation approach of Dennis Letts. It appears that others agreed with the importance of this work. It is known that Industrial Heat worked with this, but that work was discontinued when they closed their lab and released the staff. There was an attempted replication by ReResearch, also published in JCMNS, vol. 20, 2016.

It failed. I notice an acknowledgement from the authors:

we would like to thank Coolescence LLC for the contribution of Pd
material to test in this experimental campaign.

Eek! It is known that the source of palladium can be crucial. There can be unknown impurities or structure present from manufacturing. “Perfect palladium” apparently does not work.

ReResearch showed that replicating Letts wasn’t easy. Letts has claimed high reliability, but that was in his own practice. He might be carrying just the right mojo hand.

McKubre laid out how to run replication; it starts with seeing the effect, where possible, in the original lab and then, step-by-step, this is moved to the replicating lab. As necessary, the original reporter participates in the new lab, the replicators want to see the effect in their own lab. Eventually, the work becomes completely independent and eventually, controls are added. This is painstaking work, done properly.

For future work, my hope is that helium measurement be added. This is difficult and expensive, but … consider the ReResearch work. They clearly did not obtain the FP Heat Effect (or it was not at adequate levels). With helium measurement, this could be confirmed. In the Letts work, the primary study is of the effect of laser stimulation and laser frequency. (This is dual-laser and the effective frequency is thought to be the beat-frequency of the two lasers, in the THz region, as predicted by Hagelstein.)

There are many details where failure is possible.

(One of the supporting activities in the field is and will be the development of more precise helium measurement methods and sampling protocols. My sense is that what already exists is adequate for work where there is significant heat, but if sensitivity and precision can be increased, this will allow the extension of reaction confirmation into lower heat levels.

Other work that might be classified in Phase II would be the identification of additional signals of the reaction. These do not need to be “nuclear” if they are shown to be associated with the nuclear effect. An example: suppose it turns out that the acoustic signals reported by SPAWAR are distinct and associated with reaction “success.” It could then become easier and faster to identify the reaction.

There are fire alarms that depend on heat. (Sprinker systems activate when a plug melts in the sprinkler, and then the movement of water in the piping triggers an alarm. But fire alarms can also detect smoke!)

With a strong signal like helium, if there appears to be heat, and there is no helium, this would then be additional grounds to suspect calorimetry error. Ultimate assessment should be based on extensive experimental series, and hopefully many measures, not just anecdotes and single measures, as has happened too often.