Return to the clickable list of items

After reading my item #148 Richard Eskimos sent me a set of comments. He wrote:

"I noticed questions about Storms' paper on your website. Since I've been looking at the paper in detail, I may have some of the answers for you. In any case, it will help with my explanation of the Storms/Shanahan question [see item #151]. I've attached the .html file with my answers."

WHAT FOLLOWS IT THE CONTENT OF RICHRD'S FILE. THE ONLY DIFFERENCE BETWEEN ITEM #148 AND THIS ITEM #150 ARE RICHARD'S COMMENTS AND LINKS. THEY FOLLOW MY QUESTIONS PRINTED IN RED. LET ME ADD THAT RICHARD ESKIMOS IS THE AUTHOR OF ITEM #151

150) ON DIFFICULTIES COMMUNICATING

Ludwik Kowalski (6/9/04)
Department of Mathematical Sciences
Montclair State University, Upper Montclair, NJ, 07043



Introduction
Chemists and material scientists are often more familiar with physics, than physicists are with chemistry or material science. That is why reports written by chemists or material scientists are likely to be frustrating reading to a physics teacher, such as myself. This is unfortunate because cold fusion is a field involving three disciplines. This essay illustrates the kind of frustration I encounter. It is not a criticism of the paper that I am using as an example; that paper was written by an expert--Edmund Storms--and it was presented at a scientific conference attended by other experts. I suppose that most of them could find the answers to my questions “between the lines” of the reports (they are interested in what is new and not in what has already been debated). Experts communicating across the borders of scientific disciplines should be aware that their messages are not always clear to others.

How It Started
In browsing the Internet I found the following e-mail message (January 28, 2000), authored by E. Storms.

“I would like to invite the resident skeptics and curious to view the results of a successful cold fusion experiment, in particular a study of the Pons-Fleischmann Effect. . . . To give you a little background, I have been studying the P-F effect for about 10 years starting at the Los Alamos National Laboratory. During that time I have seen anomalous energy production from Pd-D2O cells having values over a range up to 7.5 watts. Recently, I have constructed several flow-type calorimeters having much better sensitivity and stability than the isoperibolic type I was using. Consequently, I can detect and believe lower excess power levels, values I would have ignored before. Recently, several samples produced anomalous power and will be described in detail at ICCF-8. I hope this rare success will stir up some interest and suggestions. Please feel free to give it your best shot. “

Fortunately, Storms’ ICCF-8 paper can be downloaded from the Internet. Google pointed to that paper after the following search phrase was specified:

E. Storms, 2000. "Excess Power Production from Platinum Cathodes Using the Pons-Fleischmann Effect," in Eighth International Conference on Cold Fusion. 2000 Lerici (La Spezia), Italy: Italian Physical Society, Bologna,Italy.

What follows is the text of that important paper. I am showing it to identify places at which I faced conceptual difficulties. My numbered questions are inserted in red capital letters. The figures, and the list of references, are skipped to shorten the description. I notice the term “Pons-Fleischmann effect’ in the title of Storms’ paper. Presumably this refers to generation of unexplained exess heat. That is not the same thing as as the term “cold fusion,” invented long before 1989, and defined as a nuclear process taking place in condensed matter. The two are likely to be related but this should not prevent us from saying that the P-F effects belongs to chemistry while the CF effects belong to physics. Division of labor along traditional lines is useful.

= = = = = = = = = = = = = = = = = = = = = = = = = = =



EXCESS POWER PRODUCTION FROM PLATINUM CATHODES USING THE PONS-FLEISCHMANN EFFECT

Edmund Storms Energy K. Systems,
2140 Paseo Ponderosa, Santa Fe, NM 87501


ABSTRACT
Excess power was produced using a platinum cathode. Efforts to produce active cathodes by plating palladium onto various metals were largely unsucessful.

INTRODUCTION
Palladium has been the cathode of choice since Pons and Fleischmann made their original claims. Occasionally, anomalous energy has been claimed to result from other elements such as Pt[1; 2], Au[3], Ti[4]. Thin layers of palladium on various inert substrates have also been claimed to produce anomalous energy [5; 6; 7; 8]. From this collection of experience, one might conclude that any layer of Pd made to stick tightly to the surface of another material would produce energy with greater ease than the bulk metal. This assumption has been found to be false even though such material can achieve a D/Pd ratio greater than 1.5[9]. Layers of electroplated Pd can be just as difficult to reproduce as bulk material. Although several successful samples were made, this paper will describe only one example of uncoated platinum which produced excess energy after being electrolyzed for an extended time in LiOD+D2O.

EXPERIMENTAL

Calorimeter Design:
The calorimeter, shown in Fig. 1, consists of a Pyrex glass cell surrounded by a watercooled jacket. This assembly is contained in a vacuum dewar, thereby allowing most of the energy lost through the lid to be picked up by the cooling water. A magnetic stirrer is used to stir the electrolyte, thereby reducing temperature gradients.

(1*) AT WHAT RATE WAS THE LIQUID HEATED BY THE STEERER?
(2*) WHY SHOULD I ASSUME THAT IT WAS A NEGLIGIBLE RATE IN COMPARISON WITH THE EXCESS POWER (EP~ 0.1 W)?




Since the stirrer was used at the same rate for active and calibration runs, it shouldn't matter. I imaging the heating effect of the stirrer is difficult to measure, since most of the power generated by the driving electromagnet is wasted elsewhere.

The entire assembly along with all reference resistors is contained in a constant temperature environment. Table I lists values and uncertainties for the various quantities. The electrolytic cell contains three linear thermistors within the electrolyte, one near the top of the solution, one near the bottom, and the third just above the cathode. The anode is equidistant (0.5 cm) from the flat plate cathode (1 cm x 2 cm). Temperature of distilled water flowing through the jacket is measured just as it enters the jacket and just as it leaves. Data are recorded every 15 min. using a National Instruments data acquisition system after averaging 15000 values. The flow rate is measured by allowing the water after it leaves the calorimeter to fill a container on a balance while the weight and time are recorded every 120 sec. In addition, the cell contains a Pt-coated-carbon recombiner catalyst

(3*) TO “RECOMBINE” MEANS TO MAKE H2O FROM H2 AND O2 THAT ARE PRODUCED IN THE CELL. (AN EXOTERMIC PROCESS).




Absolutely. Since an electrolytic cell generates hydrogen and oxygen gas while operating, there can be alot of argument about how much heat was lost ( the chemical energy, possibly the latent heat of vaporization, etc). That makes a Pin vs Pout comparison very suspect. By recombining any gases produced, the energy stays in the calorimeter. It makes the equipment more complex, but the understanding simpler.

and an exposed Pt wire heater for calibration. Luggin capillaries

(4*) WHAT ARE LUGGIN CAPILARIES?


allow the voltage between a platinum reference electrode and the cathode to be measured.

(5*) WHAT IS THE THIRD (REFERENCE) ELECTRODE FOR?
(6*) WHERE IS THE ELECTRIC DIAGRAM?
(7*) WHICH VOLTAGE WAS USED TO CALCULATE THE INPUT POWER?



I'm confused by the reference electrode too. I assume the input power calculation was done correctly, but an electric diagram would be useful.

Because the cell is connected to an oil reservoir, any gas generated within the cell can be detected by weighing the oil displaced onto a balance.

(8*) WHICH GASSES WERE GATHERED IN THE CELL AND AT WHAT RATES?
(9*) WAS ANY CHEMICAL ENERGY GENERATED IN PRODUCING THESE GASSES?




The purpose of the oil is to check for failure of the recombiner. My understanding is that recombiners are very sensitive to surface contamination - any oil displaced onto the balance makes that run invalid because of recombiner malfunction.

Samples can be quickly changed or replaced by an inert cathode for calibration.

(10*) SAMPLES OF WHAT?
(11*) WHY ARE SUCH CHANGES OR REPLACEMENTS NECESSARY?
(12*)WHAT ELSE CHANGES WHEN A SAMPLE IS REPLACED?



"samples" are platinum cathodes. Because this experiment is the comparison of an active run against a control run (rather than relying on absolute accuracies), it is critical to change the equipment as little as possible between runs. Storms has specifically designed it for changing cathodes with minimum disturbance of the setup.

Lengthy studies of inert platinum show a stability of ±75 mW.

(13*) WHEN IS PLATINUM NOT INERT?



The $24000 question in cold fusion. I would have loved to see Storms use this equipment as a tool to study just that question. Probably, he has, but hasn't come up with a definitive answer. My biggest complaint with this experiment is the circular argument: "this sample is inert" because it had a low Pout reading, and "this sample is active" because it had a high Pout reading. The excess heating claim is made by selecting runs based on the results. Ideally, you should specify your run as calibration or active beforehand, then compare the results to prove excess heat.

Calibration and Error
A typical calibration for the flow-mode is shown in Fig. 2, using a clean piece of platinum for the cathode.

(14*) WHAT IS FLOW MODE? WHAT ARE OTHER MODES?
(15*) TO CALIBRATE USUALLY MEANS TO ESTABLISH A RELATION BETWEE TWO OR MORE VARIABLES. WHAT VALUES ARE INVOLVED IN THIS CALIBRATION?




There are two possible measurement modes that can be made with Storms' calorimeter: flow mode and isoperibolic. Storms compares them in this review.

Isoperibolic mode means this: the calorimeter cell is surrounded by a constant-temperature jacket. The thermal conductivity "U" from the heat-generating device to the jacket should not change. You graph the temperature of the inside of the cell vs Pin. Pout is U(Tinside - Tjacket). (See eqution 5.4 in this paper.) The benefit of isoperibolic mode is that it is simple (just a temperature probe) and the cell is almost at a constant temperature. The drawback is in electrolytic cells: as you increase the power, more hydrogen and oxygen bubbles are produced. This drastically reduces the thermal conductivity.

The flow mode is: measure the flow of water, and the temperature of in-flowing water and out-flowing water. Any dependence on constant thermal conductivity is removed. However, you are now very dependent on the heat capacity of the water not changing.

"Calibration" here really means control run. You plot Pin (electrical measurement) vs Pout (heat flow out of the water) on a graph for your inert platinum cathode. You haven't really "calibrated" anything in the normal sense of the word - as in making sure a measured value is traceable to some NIST standard. You've just characterized the behaviour of the calorimeter with one cathode ("inert") for later comparison with another cathode.

Values are taken both going up and going down in applied power in the same manner as the sweeps described later. The standard deviation of the electrolytic values from the least-squares line is ±30 mW

(16*) WHAT IS THE “ELECTROLYTIC VALUE” (EXPRESSED IN WATTS)?
(17*) “THE LEAST SQUARES LINE” OF WHAT VERSUS WHAT?



The graph of Pin (electrical) vs Pout(heat flow in water) should be a straight line. Also, for a perfect calorimeter, (100% heat collection efficiency) the slope should be 1W/W.

Storms generated heat inside the calorimeter in two ways for his "calibration" runs: with a simple resistor, and with the inert cathode. The simple resistor obviously doesn't generate hydrogen and oxygen gas (I think he calls these the joule-heating values). The graph for the inert cathode does generate gases: this data is the electrolytic values and is the data he uses for comparison.

If the graph of Pin vs Pout was not a straight line, Storms would have to explain why is calorimeter changed in behaviour from one power level to another. His statement of ±30 mW least squares fit simply says that the calorimeter did not change performance from an ideal one (except for the fact that the slope was 0.98W/W rather than 1).

Note thate the Pin vs Pout graph was done on values going up in power, and going down in power. This is because one could say that chemical energy is being stored in the electrolytic cell at some point, and what looks like excess heat is actually just the release of the stored energy. The Pin vs Pout graph lies on a nice straight line going up in power and down: energy storage would cause hysteresis (the up and down lines would not overlap).

which is the same as the standard deviation from a constant value when stable excess energy is being observed at low applied power.

(18*) HOW WAS “STABLE EXCESS HEAT” DEFINED OPERATIONALLY?



Good question. There is really nothing "stable" or "repeatable" about cold fusion cathodes in general. Specifically, here he probably means that at constant voltage and current settings, the Pout stays constant.

Figure 1. Drawing of the calorimeter. The electrolyte is 65 ml of 0.3 N LiOD and the anode is Pt mesh. The cell lid is Lucite and the Dewar lid is expanded foam insulation. All thermistors are glass covered. Time to reach a steady temperature is 50 min.

TABLE I
Summary of uncertainties in measured quantities
Water temperature entering the jacket = 20±0.02°
Environment temperature = 20±0.03°
Flow rate = 31.00±0.05 g/min (long term variation)
Precision of current measurement = ±<0.001 A
Precision of voltage measurement = ±<0.001 V
Precision of temperature measurement = ±<0.005°
Absolute accuracy of temperature measurement = 0.1°
Stirring rate = 300 rpm ± 1 rpm
Average heat capture efficiency = 98±0.5%

(19*) WHICH VOLTAGES WERE MEASURED AND HOW WERE THEY USED?



From what I can figure out in the raw data, the voltage used for Pin is the voltage from the cathode to the anode. I'm not sure yet how the "reference voltage" fits in here. The measurement of Pin should be so straightforward that I'm not inclined to think Strorms' screwed up here. It is a bit of faith on my part, because I don't want to delve deeper.

This scatter increases to ±0.1 W at the upper limit of applied power (27 W). A zero drift as much as -0.05 W has been observed over an extended time. Consequently, changes in excess energy production are more accurate than absolute values. Good agreement between the electrolytic- and Joule-based calibrations shows that the location of heat production does not affect the accuracy of the device. Doubling the fluid flow from 22.3 g/min to 45.3 g/min caused a change in the calibration constant from 0.0732 W/degree-g/min to 0.0738 W/degree-g/min, indicating that good thermal mixing is achieved in the exiting cooling water.

(20*) WHAT WATTAGE IS USED IN THE UNIT OF CALIBRATION CONSTANT?


Wattage: the Pin (electric power) and the Pout (water flow rate and temperature in and out)

(21*) HOW WAS THE CALIBRATION CONSTANT DETERMINED?


OK. Here is the crux of the whole Storms/Shanahan discussion: unfortunately, it can't be explained without going into some detail.

The water flowed at 31 grams/minute. The heat capacity of water is 4.18188 J/g/degC at 20C. 1 Watt is 1 Joule/second. So the theoretical "calibration" constant for a flow-mode calorimeter with 100% heat collection efficiency would be 4.18188 J/gC * 31g/min * 1min/60sec = 0.069698 W/degree-g/min * 31 g/min. This is the same value Shanahan calculates for m*Cp in Table 1. It is the value for a calorimeter with 100% heat collection efficiency.

Storms has is an abstruseobtuse way of presenting the calibration data, but I guess he does it because he multiplies the flow rate x (Tout - Tin) later and he wants to get Watts out.

In any case, Storms never calculates a theoretical 100% efficiency value. He gets his value of m*Cp experimentally from his graph of Pout vs Pin. Storms' value of 0.0732 W/degree-g/min to 0.0738 W/degree-g/min above corresponds to 95% to 94% heat collection efficiency.

From his two "calibration" runs with the inert cathode, Storms gets values of 0.071221 and .070892 (See Table 1 of Shanahan or, worse, go to Storms' raw data [html][txt][zip] courtesy of that 8th wonder of the world, the Wayback Machine!) which correspond to 97.8 and 98.3% efficiency. Storms concludes he has excess heat because some active cathode runs produce values like .0680 (run 4), .685 (run1), etc, (Table 1) which clearly shows that Pout has increased by a few percent, as compared to the inert cathode. Shanahan argues that .0685 is not that different than his calculated value of .069698, and concludes no excess heat. Shanahan has made the comparison against the wrong "calibration".


Because samples can be easily changed, the cathode is frequently replaced by clean platinum when the need arises to recalibrate. Good stability is shown by a scatter of only ±1.6% in the calibration constant when measured many times over three months.

Figure 2. Comparison between electrolytic and heater calibrations before and after the study using the flow method. The heater and electrolysis agree within 1.2%.

(22*) I SUPPOSE THAT DJ (SEE FIGURE 1) IS THE TEMPERATURE DIFFERENCE BETWEEN WATER-OUT AND WATER-IN. BUT WHY IS IT DJ AND NOT DT?


Confused me too, for a while. I guess delta T is for the temperature difference between the 3 thermal probes inside the electrolytic cell, and that left "delta Jacket". Confusing choice!.

(23*) IN FIGURE 2 T FOR THE TEMPERATURE “ACROSS THE JACKET.”
IS IT THE SAME AS DJ?


The cell can also be used as a rough isoperibolic calorimeter by measuring the average temperature between the electrolyte and the cooling jacket.

(24*) OK, THIS BE THE SECOND MODE OF OPERATION. (FIRST WAS THE FLOW MODE).
(25*) FORTUNATELY, I KNOW WHAT “ISOPERIBOLIC” IS. BUT MANY PHYSICISTS
ARE LIKELY TO BE CONFUSED BY THIS ADJECTIVE.

However, this method is not stable, in spite of active stirring, because of changes in the convection currents in the electrolyte and in the jacket. Figure 3 shows how the average temperature across the jacket changed between the first and final calibration using the electrolytic method and clean Pt.

Figure 3. Comparison between applied electrolytic watts and the average temperature across the jacket for two calibration runs.

Excess Energy Measurement
Figures 4, 5, and 6 show the time history for excess power (EP) production using a special platinum cathode.

(26*) WHAT MAKES THIS ELECTRODE “SPECIAL”?

Current sweeps consist of stepping the current up in value, waiting for the calorimeter to achieve steady-state (50 min), taking five values, and repeating the process.

(27*) HOW WERE THE VALUES OF EXCESS POWER (EP) CALCULATED?
(28*) THE ERROR BARS ARE GIVEN IN TABLE 1. HOW SHOULD I USE THEM TO ESTIMATE ERROR BARS FOR THE VALUES OF EP?


After reaching 3 A, the current is reduced in steps. Notice in Fig. 4 that the expected EP was achieved at 0.5 A and 1.0 A, but decayed away when 1.5 A was applied.

(29*) WHAT IS THE “EXPECTED” EXCESS POWER?
(30*) IF IT MEANS “EXPECTED ON THE BASIS OF ANOTHER REPORT,” THEN WHAT REPORT? IF IT MEANS “THEORETICALLY EXPECTED,” THEN BY WHAT THEORY?


I think "expected" is a poor choice of words and the cause of some concern that Storms is not unbiased. It sounds like he "expects" to see excess power, and then he sees it. I think the point of the statement is just that higher current (1.5A) somehow killed the excess power reaction.


A sweep taken after the decay showed very little EP. After the current was turned off for a brief time, the study was resumed in Fig. 5. Notice that the EP again gradually increased after 0.5A was applied. Sweep #3 again showed EP and this continued while 0.75A and 1.0A were applied. However, application of 1.5A again caused the EP to decay away.

Figure 4. Time history of excess power production from a Pt sample.

Once again the current was turned off. Application of 0.5 A, shown in Fig. 6, again showed EP. The calorimeter was calibrated at 430 h using the internal heater and later using an inert Pt cathode. This experience shows a consistent pattern of behavior which was repeated once again, but is not shown here. The first sweep is shown as applied current vs EP in Fig. 7. Notice that excess power is indicated by the isoperibolic method during this initial sweep, but later sweeps do not show an effect because of a change in calibration, as indicated in Fig. 3. Also notice that the excess power falls on a higher line upon reduction in applied current. Subsequent cycles, as shown in Fig. 8, produce excess power that fall on this higher line. The final relationship is linear and extrapolates to zero EP at zero applied current, in contrast to the behavior of palladium, which requires a critical applied current before EP is produced. Figure 9 shows the sweeps taken later using different numbers of values to produce the plotted average and these are compared to the heater calibration.

Figure 5. Time history of excess power production from a Pt sample.

DISCUSSION
A large average composition within the cathode is thought required to produce excess energy. Yet many researchers have failed to produce excess power after achieving large compositions, for example Nakata et al.[10]. Now, metals that do not even dissolve hydrogen are found to make excess energy. Clearly, additional variables are operating.

Figure 6.
Time history of excess power production from a Pt sample.

Figure 7. Comparison between excess energy measured using flow method and isoperipolic method during first current sweep.

In the case of platinum, this study suggests that an energy-active layer of unknown composition can deposit on a Pt surface. This observation might also be related to the frequent detection of Pt on the Pd cathode after excess energy is observed. Such a layer is slow to form, which is consistent with the observed long delay in producing excess energy, and, for this sample, it is unstable. Is it possible that such a layer might be the active material in all studies, even when palladium is used?. McKubre et al. [11] also suggest that a critical layer is required to maintain the required high composition in Pd. Perhaps this is the actual active material. If this is the case, the bulk properties of palladium are only important in that they must support a critical concentration within this layer. Since palladium can easily permit loss into the metal, such material might be the worst substrate to use. Platinum, gold, and other inert materials would appear to be better choices. Use of such materials would only require forming the active layer without a need to use special batches of the substrate. If this suggestion is true, significant changes in various theories will be required.

Figure 8.
Applied current vs excess power for the various sweeps, compared to Pt calibrations taken before and after the study.

Figure 9. Applied current vs excess power for sweeps using different number of values to average are compared to the heater calibration.



Return to the clickable list of items