Experimental Path after Hot-Cat Replication (Michael Lammert)

The following post was submitted by Michael Lammert (AKA Dr. Mike)


Experimental Path after Hot-Cat Replication

     Within the next 3-6 months many successful replications of a hot-cat type device are anticipated.  What experimental path should be taken after these replications?  The obvious answer to this question is to continue to run experiments both to improve the performance of the reactor and to gather data that may be useful in establishing a theory of LENR.  Improving the performance of the reactor will probably be a two-step process, first determining and optimizing key parameters of operation, then applying that knowledge to design a higher power reactor optimized for delivering heat to a particular system.  Hopefully, the experiments run for parameter optimization will lead to a universal theory of LENR, which will also aid in the development of many systems that can make use of LENR heat.

It should be noted that Andrea Rossi (and possibly a few others) have already run hundreds of experiments for optimizing LENR reactors.  However, there is little hope that any of his results or results of anyone else trying to develop a commercial product will be revealed.  Perhaps the replications will result in some public funding of LENR research.  Most publicly funded work requires a fairly extensive documentation of the work done and the results achieved.   Perhaps the replications will also lead to the acceptance of LENR patent applications (if they meet all patent disclosure rules) by the US Patent Office and to the acceptance of high quality LENR papers by major scientific journals.

Preparation for Future Experimental Work

The first step in preparing to run optimization experiments should be to review the results of all existing replication work.  The quality of this review will be totally dependent on how well each replicator has documented his/her work.  It certainly would be nice if most replicators posted their results on the e-catworld.com website for peer review.  (It would also be beneficial for each replicator to send copies of their reports to the editors of Science, Nature, Scientific American, and Science News and to the head of the US Patent Office.)  My assumption is that even if someone is trying to replicate either the Lugano or Parkhomov reactors, there will be differences in both the detailed design of their reactor and the exact fuel used in their reactor.  These details in differences among replicators might be very useful to someone just starting to run optimization experiments.  Also, improving the reactor design for experimental work should be a key factor in optimizing the reactor parameters.  Enhancements seem both in Parkhomov and MFMP reactors, such as adding pressure monitoring, moving the sealing points away from the reactor hot zone, and encapsulating the fuel are several things that should be included in an experiment friendly reactor.  Based on the results of replication efforts observed thus far, the design of a reactor to be used in optimization efforts should preferentially include the following features:

  1. A port for and means of measuring pressure

The internal reactor pressure may be an important parameter in the performance of a reactor.  Also, monitoring the pressure during a qualification test of a reactor or during an experimental run could be used to determine if the reactor is leak tight.  Any abnormality in the pressure verses time or temperature for a qualification or experiment would require the reactor to be rejected or the experiment rerun in a new reactor.

  1. A port to pressurize, pull vacuum, and analyze chamber atmosphere at the end of an experiment

As part of the qualification of every reactor and prior to each experimental run the reactor should be pressurized (probably with dry nitrogen) to the maximum expected operating pressure to check for leaks.  (The test criterion might be no measurable pressure drop in one hour after initial pressurization.)  Although this won’t guarantee that the reactor won’t leak at elevated temperatures, it will verify that no experiment starts with a leaky reactor.  Although previous reactor demonstrations have shown that it is not necessary to remove the air in the reactor prior to operation, optimization experiments would be more controlled if most of the air was initially removed from the reactor.  (The reactor atmosphere would be mostly nitrogen if the reactor was initially pressure tested to 150-250 psi.)   Therefore, it is preferable that optimization experiments start with using a vacuum pump to remove most of the nitrogen from within the reactor (perhaps to a base pressure of 1 torr or less).

Note that this same port can be used to measure the residual gas at the end of an experiment, perhaps checking for helium.  Also, this port can be used in a set of experiments that compare replacing LiAlH4 as the source of hydrogen with an initial pressurization of hydrogen gas.

  1. Ability to remove the ash at the end of an experiment and maintain spatial relevance

The fuel for the reactor needs to be held within a container that can be removed intact at the end of an experiment.  It would not be possible for a reactor to be reused for multiple experiments if any fuel/ash residue remained in the reactor from the previous experiment.  Also, for some experiments it may be desirable to analyze the ash as a function of position within the reactor (both along the length of the reactor and edge to center).  If the ash is not analyzed for some experiments, it would still be a good idea to keep the ash intact for possible future analysis.

  1. Reactor sealing points distant from high temperature zone

A leaky reactor will probably be one of biggest problems in causing inconsistent experimental results.  By keeping the reactor sealing points somewhat distant from the high temperature region of the reactor, the problem of leaks can be minimized by verifying room temperature sealing with a pressurized nitrogen test and by monitoring the pressure during all experiments.

  1. Ability to wind heating coils precisely

If experimental reactors are going to be fabricated to have nearly identical input power verses temperature profiles, the heater coils will need to be formed with a fairly high degree of precision.  This will require either an automated process or a manual process that uses some type of template to maintain precise spacing of the coil windings.

  1. Heater coils with the capacity to achieve no fuel temperatures of 1250-1300oC

Although there are only a few parameters that need to be examined at 1250-1300oC, every reactor should have the input power verses temperature characterized up to this temperature range without fuel as a pre-use qualification test.  The reactor would only need to be at this high current, high temperature condition for the time it takes for the reactor temperature to stabilize at a fixed power setting.


  1. A separate coil for supplying a high frequency electromagnetic pulse/signal to the reactor

For running the optimization experiments, particularly those investigating the power supply parameters, it is desirable to separate energy supplied to the reactor as just heat from that supplied in the form of electromagnetic energy.  My recommendation for designing a reactor for experimental tests is to have one coil for supplying heat using dc power (or 50-60 cycle sine wave ac) and a separate coil that supplies energy from a higher frequency power generator.   Separate winding for supplying heat and electromagnetic energy make it easier to measure the power supplied from each energy source and enable one power source to be turned off without affecting the other power source.  It should be noted that a second winding would not be needed on the experimental reactor if replication results show that only heat is needed to drive LENR at high temperature.  Also, a single winding could be used with an electronic superposition of a heating signal and a high frequency signal provided the power contribution from each source can be measured and each source can be turned on and off independently of the other source.

  1. Method for precise placement of one or more external thermocouples

It is anticipated that the only way to build reactors with nearly identical power verses temperature characteristics is to develop a precise procedure for attaching one or more thermocouples to the reactor.

  1. Internal thermocouple

It is desirable to be able to monitor the internal temperature of the reactor near the fuel, although some design effort may be required to achieve items #1, 2, 3 and 9 simultaneously in a single reactor.

  1. Alternative temperature measurement

Although thermocouples should be adequate for monitoring reactors, a secondary temperature measurement such as IR thermal imaging could be used as a secondary method of verifying temperature.  Some modification to the reactor design might be needed so that the IR thermal imager does not have to measure the temperature of the Al2O3 cement that covers heating coils.


  1. Ability to reuse the reactor for multiple experiments

The basic experimental reactor needs to be able to be reused for multiple experimental runs to reduce the cost of the experiments, to be able to verify the reproducibility of the reactor before it is used in an optimization experiment, and perhaps as a requirement for some sensitive experiments.

  1. Reactor fabrication procedure yields reproducible reactors

Just as it is desirable to be able to reuse a reactor many times, it is also necessary to be able to fabricate nearly identical reactors.  If an experimental matrix requires 12 separate runs, it would be good to have the ability to run 3 or 4 of the experiments simultaneously in separate, “identical” reactors.  Also, if a reactor fails in the middle of a series of experiments, it would be nice not to have to rerun all previously completed portions of that experiment.  Verification of reactor uniformity will permit the use of multiple reactors in a single experiment.

Once the design of the experimental reactor is complete and the process for building it is fixed, the reproducibility of the reactors must be established.  This can be accomplished by building 5 or so reactors to spec, verifying heater wires resistances are equal, leak checking the reactors with pressurization, verifying no-fuel control heating is nearly identical by running an input power versus temperature curve, and finally verifying that fuel-loaded reactors behave statistically the same.  Any reactor with measured data statistically “out of family” should be rejected from use and the cause of the out of family data investigated.  All future reactors can be evaluated in a similar manner with their measured data adding to the statistical data base.  Note that some reasonable specs should be set on what is the maximum allowable spread in measured data among “identical” reactors.  If the spread in the data on the initial reactors is larger than those specs, the manufacturing process of the reactors will need to be improved before starting any optimization experiments.

Another issue that should be decided before starting any experiments is what is going to be the measured parameter to determine if one operating condition is better than another.  Possible measured parameters include:

  • Excess heat as a function of temperature as measured by calorimetry
  • Delta temperature of the active run input power verses temperature curve as compared to the control (no fuel) input power vs. temperature curve
  • Time period of self-sustaining heat as a function of temperature
  • Temperature at which excess heat is first observed

Other parameters might become evident as optimization experiments are run.

Although good safety procedures should have been established in replication attempts, if they weren’t, they should be implemented before any optimization experiments.   Every experiment should be run with the assumption that the reactor can explode or melt down at any time, and some strange combination of operating conditions might produce health damaging high energy neutrons or other dangerous ionizing radiation.  Appendix 1 (pages 33-36) in the Lugano report seems to serve as a good reference for what should be done for monitoring neutrons and high energy ionizing radiation.

Reactor Performance Optimization

The optimization of the fuel components and the operating conditions for the LENR reactor should be done following standard methodologies of experimental design.  It might be beneficial to those involved in running LENR experiments to review some of the concepts of experimental design (such as taught in Six Sigma training) before beginning the experiments.  Even the most basic LENR reactor is a multi-factor, multi-level device from an experimental design viewpoint.  Some of the key factors and levels to be evaluated in the Lugano type reactor include the:

  1. Ni
  2. LiAlH4
  3. Ratio of Ni : LiAlH4 weight and ratio of total fuel weight to reactor volume
  4. Importance of Li and Al
  5. Effect of adding other elements to the fuel
  6. Effect of the heater power supply and other electromagnetic pulses
  7. Effect of time, temperature, and electromagnetic pulse variables on products found in the ash and residual gas and on the self-sustaining time period

Possible experimental work that should be done to optimize the above parameters is as follows:


Although the Ni powder is surely one of the most important parameters in the reactor, it won’t be surprising if most small particle nickel powders work well.  However, if one vendor’s Ni powder is definitely better than others, it will be necessary to determine the differences in the Ni powders.   The basic experiment for evaluating different Ni powders is just to run each powder in the reactor over a range of input powers, holding all other variables constant.  If differences are noted in the experimental runs, then the Ni powders in those experiments need to be compared in morphology, nominal particle size, nominal surface area, and the impurities present.  If one vendor’s Ni powder demonstrates better results than all other tested, it would probably be beneficial to test a second sample from that vendor that was manufactured in a separate batch to verify that the vendor can consistently fabricate a best performing Ni powder.  One other investigation that would be useful would be to evaluate the effect of the Ni powder surface area using Ni powder produced by the same vendor.  For example, Umicore’s “Ultrafine Powder” (1 micron) has a surface area that is about three times that of their “2M Powder” (2 micron).


There hasn’t been much information on the LiAlH4 powder in the replication reports other than the rather large mostly Al particle shown as “Particle #2” in the fuel analysis in the Lugano report on page 44.  The experiments needed to optimize the LiAlH4 would consist of comparing the LiAlH4 powder that works in the replications to other vendors LiAlH4 powder.  If differences are found in the reactor performance, the LiAlH4 powders should be evaluated for particle size and purity.

Ratio of Ni : LiAlH4 Weight and the Ratio of Total Fuel Weight to Reactor Volume

These two factors are grouped together since a single set of experiments can probably optimize these parameters.  LiAlH4 weights at 5%, 10%, and 15% of the Ni weight could be used as one set of parameters in an experimental matrix.  The other set of factors would be 2 or 3 total weights of fuel, the total weight depending on the exact volume of the reactor (which needs to be determined to an accuracy of at least 0.1cm3).

Importance of Li and Al

Obviously, one important purpose of the LiAlH4 is to supply hydrogen, but what other roles do Li and Al play in the LENR reactor?  Possible experiments that might be run to evaluate the roles of Li and Al include:

  1. Replace the LiAlH4 with a different hydride that contains no Li or Al.
  2. Replace the LiAlH4 with hydrogen gas at a pressure that has the same hydrogen yield as the LiAlH4 (after removing residual nitrogen with a vacuum pump).
  3. Add Li and Al individually and in combination to (1) and (2) above.
  4. Repeat (1), (2), and (3) but pressure test the reactor with dry air and do not remove the air with a vacuum pump prior to beginning the active run (start runs at one atmosphere pressure of dry air).


Effect of Other Elements in the Fuel

It is quite possible that the addition of other elements to the baseline fuel mixture could significantly enhance the reactor performance.  Three elements that might be considered for investigation are Fe, C and Ti.  However, other elements a may be better initial choice for initial investigation, especially as more data is acquired.  Since it is unknown what quantity of these extra elements might be beneficial, the experimental design should encompass a wide range of quantities, perhaps 1%, 5%, and 25% of the total fuel weight.  A full matrix experiment would require 9 test runs, a baseline control run, and perhaps 2 or 3 replications.  If one wanted to also investigate combinations of any two of the elements with all combinations of weight percentages, an additional 27 test runs would be required without any replications.  An experiment such as this shows the need for reusable and reproducible reactors.  Also, it shows why one needs to use techniques from experimental design to run a partial matrix of an experiment rather than the full matrix, but still be able to acquire most of the information that would come from running the full matrix.

Effect of the Heater Power Supply and Other Electromagnetic Pulses

As mentioned earlier, a reactor with one coil winding for supplying heat and one winding for electromagnetic pulses would be a very desirable feature of an experimental reactor.  Assuming that the reactor is designed with two windings, the optimization of the signal to the second winding would have to be completed prior to running any fuel optimization experiments. Every new reactor would need to be validated using both windings.  The first experiment would need to demonstrate that just supplying power to the heater winding with a triac chopped ac signal is sufficient to demonstrate LENR replication.  The next experiment would be to run the same reactor first with dc power to the heater winding, then to run variac controlled ac power to just the heater winding.  These experiments probably will show no LENR excess heat effects.  The same reactor would then have dc or variac controlled ac supplied to the heater winding to achieve a reactor temperature of about 900oC.  A signal generator/pulse generator would be hooked up to the second winding of the reactor.  The parameters that might be examined for the signal generator are:

  1. Square wave
  • 100Hz, 1KHz, 10KHz, 100KHz, and 1MHz
  • 3-5 Different amplitudes
  1. Sawtooth wave
  • 100Hz, 1KHz, 10KHz, 100KHz, and 1MHz
  • 3-5 Different amplitudes
  1. Sine wave
  • 100Hz, 1KHz, 10KHz, 100KHz, and 1MHz
  • 3-5 Different amplitudes
  1. Triangular wave
  • 100Hz, 1KHz, 10KHz, 100KHz, and 1MHz
  • 3-5 Different amplitudes

Experimental parameters that might be evaluated for a pulse generator include:

  1. Pulse rates of 100Hz, 1KHz, 10KHz, 100KHz, and 1MHz
  2. Pulse durations of 3% and 10% of the pulse period
  3. 3-5 different amplitudes

For at least one or two of the above high frequency inputs, the frequency should be swept slowly over the entire frequency range to determine an exact optimum frequency.   All signal waveforms should be closely examined in a frequency range near this optimum frequency.  It should be noted that the power going to the second winding from the signal generator / pulse generator will have to be monitored and added to the power supplied to the heater wire to give the total input power going to the reactor.  If the reactor temperature is greater than the temperature on the control (no fuel) run for that reactor at that total input power, excess heat is being generated in the reactor.  This is one experiment where it might be possible to create a thermal runaway in the reactor.  The signal generator/ pulse generator should be turned off immediately if a rapid temperature rise is noted in any TC reading.  Lower the signal/ pulse amplitude and recheck frequency effects if rapid temperature rises are noted.  For all previously mentioned fuel optimization experiments the reactor should be run at the frequency producing the largest excess heat effect and at signal amplitudes well under any that might produce a runaway condition.

It should be noted that measuring the optimum frequency of the signal going to the second winding should be one measurement taken during the initial reactor qualification and should be a criterion for accepting the reactor for experimental use.  Another interesting experiment would be to determine if fuels with different Ni particle sizes (or distinctively different morphology) have dramatically different optimum second winding signal frequencies.  In fact, the frequency of the signal supplied to the second winding in every fuel optimization experiment should be scanned over a narrow range to determine if the optimum frequency shifts any in each experiment.  Another test that should be run with at least some of the fuel experiments that showed optimum results is to turn off the power to the heating coil and determine if the reactor temperature can be controlled by only the amplitude (or pulse rate) of the signal going to the high frequency winding.  If the reactor’s temperature can be controlled with just the high frequency signal and its power level is very small relative to the output power of the reactor, it should be possible to run the reactor at a very high COP continuously from start up until fuel exhaustion.

Effect of Time, Temperature, and Electromagnetic Pulse Variables on Products Found in the Ash and Residual Gas and on the Self-Sustaining Time Period

This last proposed experiment probably won’t benefit the optimization of the performance of the reactor but may greatly help in establishing a theory for LENR.  This experiment would use the optimized fuel mixture from previous experiments in all test runs.  Reactors would be run under the following conditions:

  • Two amplitudes of optimum high frequency pulses going to winding #2
  • Three temperatures, perhaps 850oC, 1050oC, and 1250oC
  • Four time periods, perhaps 12 hours, 2 days, 8 days and 32 days

The dc power (or variac controlled ac) going to the heater winding would be adjusted to achieve the desired set point temperature for each of the two high frequency amplitudes.  This experiment would require 24 runs plus a few extra runs for replication verification.  At the end of the time period for each run the power would be turned off and the time that the reactor maintained temperature should be measured (self-sustaining time).  The reactor residual gas would be analyzed at the end of each run and a complete analysis would be made of the “ash” at the end of each run.  For a few of the runs it may be desirable to determine the uniformity of the “ash” along the length of the reactor and from edge to center.  The results of this experiment should show isotropic changes in the fuel as a function of time for 6 operating conditions (3 temperatures and 2 high frequency amplitudes).   The self-sustaining times can also be correlated to the isotropic and composition changes in the fuel.

Scaling to a Larger Size Reactor and Operating under Load

      The information gained in optimizing the fuel and operation conditions in the experimental reactor should be helpful in building a higher output power reactor.  It is anticipated the optimum fuel components and their weight ratios would be about the same in a small reactor as a larger reactor, especially if the higher power reactor operates by primarily radiating heat.  However, building a higher output reactor will not be as simple as building a bigger reactor and adding more fuel.

One feature from the experimental reactor that needs incorporated into a higher power reactor is the ability to refuel (reuse) the reactor.  Perhaps, all that is needed to accomplish this is a port to remove the atmosphere (a large percentage He?) of the used reactor, then refill the reactor to the proper pressure of hydrogen.  The reactor design would have to be considerably more complex to enable complete refueling.

It is obvious that a reactor with ten times the output of the experimental reactor can be build by placing 10 experimental size reactors in a single box. However, each small reactor would need its own control system and further scaling up in size requires another controller for each added reactor.  Another fairly straight forward method of scaling the reactor is to form a large reactor from cascaded reactors.  (Andrea Rossi has hinted that he is using some form of cascading to build larger “hot-cat” type reactors.)  In this configuration a small reactor provides the heat for a larger reactor, which perhaps provides heat for a third larger reactor.  If each reactor has a gain of 5, the 3-stage reactor would have a total gain of 125. Although the control system would be more complicated for a 3-stage reactor than for a single stage reactor, it would be much less complicated than a control system for 125 individual reactors. If future high temperature LENR reactors are primarily driven by heat, it is quite likely that some form of cascading will be used in most larger reactors.

Thermal loading is one issue that will have an impact on the scaling of LENR reactors.  Most replication reactors and experimental reactors used in optimization experiments will be run under a “no load” condition, that is, all heat released goes into the “infinite” sink of the atmosphere.  The reactor will reach an equilibrium temperature such that that heat generated by the input power plus the heat generated by the LENR effect is equal to the heat being lost through radiation (and smaller amount through convection and conduction) to the surroundings.   In a useful device the output heat from the reactor will actually be heating something, that something being a thermal load on the reactor.  If most of the heat is transferred to the load by means of radiation, it is quite likely that the optimum operating conditions for the experimental reactor can be directly applied to a high power reactor.   If a higher power output is desired from a reactor than can be achieved through radiative energy, heat transfer through conduction and convection must be added to the system design (using the talents of thermo engineers).  It is quite likely that some of the reactor parameters that were optimized for reactors operating in a radiative mode will not be valid for a scaled reactor in which conduction and convection are important components of the heat transfer to the thermal load.  For example, more fuel would probably be needed in a reactor with a high thermal load as compared to the same reactor operating with a small thermal load.

There are several other engineering problems that will probably show up in scaled higher power reactors.  First, heat transfer from the fuel to the reactor chamber wall seems like it will be a critical factor in designing higher power reactors, particularly for reactors having higher thermal loads.  The fuel most distant from the reactor chamber wall must be kept well below its melting point.  Another possible problem in scaled up reactors is localized hot spots in the fuel mixture.  If more heat increases the LENR reaction rate, then a hot spot will continue to heat up until the fuel melts.  Localized hot spots in a scaled up reactor seems to be very similar to the problem of scaling up the size of power transistors using parallel devices (or multi-emitter devices), in which when one emitter begins to get hotter than the other emitters, the current in that emitter increases causing it to get even hotter.  This positive feedback eventually results in all of the current wanting to flow through only one of the emitters, often destroying the transistor.  The “current hogging” problem can be fixed in power transistors by adding a series resistor to each emitter.  Any rise in one of the emitter currents results in a larger voltage drop across the emitter resistor, lowering the voltage to that emitter, thereby reducing the emitter current.  A higher power reactor may need a similar negative feedback mechanism added to its design to prevent hot spot formation and propagation analogous to adding emitter resistors to paralleled power transistors.

One other problem that would probably show up in the design of a high power reactor is that if more of the heat is being carried away from the reactor by conduction (or perhaps convection) the reactor could melt down in seconds if the heat transfer from the reactor to the load stopped or decreased rapidly.   The rapid feedback needed to control the reactor temperature might be the most difficult problem to overcome in developing a high power reactor, especially if the thermal load on the reactor is not constant.   This problem could limit the maximum operating temperature of the reactor and limit the useful maximum output power of a large reactor.  Also, if the ability to shut down the reactor quickly is a necessity, the operating conditions of the reactor may need to be adjusted so that some input electromagnetic signal is always required to keep the reactor running.  Turning off this signal would have to immediately stop the LENR produced heat.

     Although it appears that considerable information and data will be available to the public on efforts for replicating LENR heat generation in high temperature reactors, it is not clear that much information and data will be available to the public from efforts to optimize reactor performance and develop higher power reactors.  Perhaps some of the ideas in this post (and ideas from those commenting on the post) will be useful to those who will be running post-replication experiments and who can share their results with the public.