Experiments with nickel and hydrogen

Curt Edström, Jan-Erik Nowacki

1. Summary ...... 3 2. Short theory concering nickel-hydrogen ...... 3 3. Our experiments ...... 5 3.1. General Method ...... 5 3.2. Used test equipment ...... 6 3.2.1. Reactor vessels used ...... 6 3.2.2. Heating and controlling power ...... 9 3.2.3. Evacuating and filling gases ...... 10 3.2.4. Measuring γ-radiation ...... 11 3.3. Preparates used ...... 12 3.3.1. Examples of different metal (nickel) powders tested ...... 12 3.4. Used pressures and temperatures ...... 13 3.4.1. 2011-04-21 Curt Edström Experiment 1 ...... 13 3.4.2. 2011-07-25 Curt Edström Experiment 25, reactor 4...... 14

1. Summary

The energy situation in the world demands that all possible new energy forms be investigated. The Swedish Defence Materiel Administration, FMV has therefore financed some very rudimentary experiments with nickel and hydrogen, trying to experimentally reproduce the excess heating power claimed by Andrea Rossi and Sergio Focardi and described e.g. in the technical newspaper Ny Teknik1.

Four different reactor chambers were built by us, in which different forms of nickel were tested in contact with hydrogen at different pressures and temperatures. Some of the nickel samples also contained other metal “catalysts” like lithium, potassium, or copper. In some if our samples the nickel was e.g. in micrometer large crystal grains, in other samples the nickel was in the form of nanometre grains embedded in zirconium oxide. Contacts wer taken with many active researchers in the field, including Andrea Rossi, asking for guidance to find a functioning solution. Andrea Rossi could not reveal his catalyst for us but thought that we would get a small indicative response using just nickel and hydrogen.

Neither excess heat, nor any radiation indicating nuclear reactions have however been observed in the experiments in such large amounts that we have been able to detect it. We cannot however say for sure that no reaction took place. If other verification experiments using nickel and hydrogen, going on around the world, also should fail, we suggest a focus on other more “conventional” nuclear reactions.

One such reaction is lithium liquid bombarded by hydrogen 7Li + 1p → 24He + 17.35 MeV. That process, using liquid lithium, has been claimed to produce a high reaction probability,2 (large enough for practical nuclear reactions?). Lord Rutherford’s students Cockcroft and Walton showed that this reaction works qualitatively already 1932 yielding 100 – 500 times more output energy than input3. They got the Nobel prize 1951. The problem using lithium in solid phase, is/was the low probability for the reaction to occur (low cross section).

Another such aneutronic reaction (without significant neutron emission) is 1p + 11B → 3 4He + 8.7 MeV – the excess energy is lower however there are other advantages. For a more common use of nuclear energy these reactions have the advantage of not creating large amounts of radioactive waste. Some more speculation about such experiments using proton accelerators instead of high temperature reactors are at the end of this report.

2. Short theory concering nickel-hydrogen

1 Mats Lewan, Cold fusion: 18 hours excludes burning (of hydrogen), Ny Teknik 2011-02-23 2 Hidetsugu Ikegami, Toru Watanabe, Roland Pettersson, Kjell Fransson, Ultradense Nuclear Fusion in Metallic Lithium Liquid, Energimyndigheten, ER 2006:42. 3 http://plato.stanford.edu/entries/equivME/ Anomalous heat production was shown practically by Focardi et al already in 19944. Focardi and Rossi made an attempt to a physical explanation of the phenomenon 20105. According to this there are many possible nuclear reactions between nickel and hydrogen leading to excess heat.

NiXX+→ p11 Cu +

Except for the stable isotopes of copper (63 and 65) the copper will then decay back into nickel again according to:

CuXX+11→ Ni ++ ++ e ν

Subsequently the e+ particle would be annihilated by:

ee+−+ →+γγ

This resulting gamma radiation should normally be easy to detect.

When going through the different isotopes of nickel it can be concluded that quite a lot of energy would be released per reaction [from Rossi Focardi6]. Of course the natural isotope concentration is important, as should also be the reaction rate of the isotope.

Tab. 1 The energy per reaction from Rossi Focardi5 in MeV

To get a practical hint of how high the energy developed is, one can compare with the annual heating need of a typical Swedish house of 25 MWh/year. If we assume that each reacting nickel molecule gives 5 MeV totally, 25 MWh corresponds to about 55 g of Ni reacting annually with about 1 g of hydrogen.

There were experiments made already in 19967 that tried to explain the Forcardi measure- ments in another way.

4 S. Focardi, R Habel and F Piantelli, Anomalous Heat Production in Ni-H Systems, Il Nuovo Cimiento, Note Brevi, vol 107 A, N. 1, 1994 5 S. Focardi, A. Rossi, A new energy source from nuclear fusion, Physics Department, Bologna University and INFN Bologna Section and Leonardo Cort (USA), 2010. 6 S Focardi, A Rossi, A new energy source from nuclear fusion, Physics Department, Bologna University, 2010

The prerequisite for the reactions above is that the coulomb barrier can be overcome. The coulomb barrier would repel the positive nickel core from the positive proton in the hydrogen core and almost totally prevent them from reacting. Another way of saying this is that conventional nuclear physics would never allow a reaction between nickel and hydrogen at the temperature levels used. It has therefore been assumed that some unknown mechanism allows the reaction to take place.

As we cannot explain the nature of the reaction we chose to experimentally verify if we could achieve any reaction at all in the first place – and if such a reaction was achieved try to explain it theoretically later.

Rossis experiments as they are described by Ny Teknik1 seems to be have been performed according to figure 1 below. The amount of water fed into the reactor and then evaporated per second was measured and gave the resulting power knowing the heat of evaporation of water.

Fig. 1. Rossis measurements of power.

3. Our experiments 3.1. General Method

When heating a sample inside a test tube, in room temperature and using an internal electric heater, the tube temperature will follow a certain curve. If a higher power is added to the sample the curve will rise faster and the balance temperature difference to the room air will be higher.

7 E. Cerron-Zeballos, I. Crotty, D. Hatzifotiadou, J. Lamas Valverde, M. C. S. Williams and A. Zichichi, Investigation of anomalous heat production in Ni-H systems, LAA project, CERN - Geneva Switzerland, 1996

Fig. 2. The rise of temperature with a lower (continous) and higher (dotted) power used inside a test tube.

Another practical way that we mainly used was to maintain the temperature constant and then observe fluctuations in the power supplied to keep that temperature. An exothermal reaction would then lead to that less current had to be supplied to maintain temperature (and pressure) in the constant volume. One problem we had was also that hydrogen in chemically absorbed in nickel yielding a chemical heat up and also a pressure drop. The amount of hydrogen that the nickel can hold is affected by temperature. The higher the temperature the smaller amount of chemically dissolved hydrogen.

The calibration was done with a nickel power sample without any hydrogen (using inert argon-gas instead) and with a carefully measured electric power. When adding more electric power a higher curve results and from that the difference in temperature as a function of the power can be deduced. If the convection around the tube and radiation from the tube would be linear to temperature the curve would follow a very simple exponential function. That is however a too large simplification in this case. We of course also insulated the tubes exterior surfaces, which made the sensitivity much larger. In our later experiments an excess heat of 1 W would have been detected if the rise would be instantaneous. If the reaction would start smoothly we would require more power – say 5 W - to be able to say that we have been able to verify an extra power output.

In addition to the temperature measurements we used a gamma-meter to detect radiation. This gamma-meter was calibrated with a Ba 133 radioactive sample.

3.2. Used test equipment

3.2.1. Reactor vessels used We started with rather large reactors which then grew smaller and smaller as we succesively learnt from Rossi and others that higher temperatures and pressures should be used to achieve a reaction.

Fig. 3. Different reactor vessels that were used

The first vessel to the left was made for about 35 bar and the last one to the right for about 300 bar. The last reactor vessel “4” to the far right is shown in cross section below in figure XX. This was the last and best vessel we used.

Fig. 4. Reactor vessel 4 – layout principle

Fig. 5. Dimension of the outer tubing, OD 35 ID 30mm.

The heating wire was made of ceramic insulated Kanthal wire. The maximum power was 300 W, and it was wrapped around the bottom of the tube. The heating wire was insulated outside using glass wool. As can be seen from fig XX there was an extra container inside reactor 4 holding the nickel powder that was going to be tested.

Fig. 6. The container in the bottom of the tube.

The distance, 1mm around and 2.5 mm in the bottom, gave entry to the hydrogen gas. The Inside dimension of this inner container was 26 mm diameter. It was 40 mm high and the wall thickness was 1 mm. The chosen material was stainless steel. Two holes (8 mm) were drilled in the lid so that H2 gas could enter the container (not yet drilled in this lid). The thermocouple was held in position by the lid. Glass wool was placed also inside the container, mainly to keep the powder in place.

3.2.2. Heating and controlling power

The power was kept constant using a PID-regulator – it is shown in the figure below. For an experiment of this type it is of outmost importance that the electric power is known and can be controlled accurately.

Fig. 7. The PID-regulator for power – original and improved version

3.2.3. Evacuating and filling gases Before each experiment the reactor vessels were evacuated after which either argon (as dummy gas) or hydrogen was supplied. The system including valves looked like figure XX below.

Fig. 8. Supply of gas, pumping vacuum and safety valves

It was very difficult and took many hours to get the rig gas tight especially for hydrogen. The gas tightness could be measured in many ways – looking how fast the vacuum pressure increased or looking for how fast a high hydrogen pressure decreased while maintaining the same temperature.

3.2.4. Measuring γ-radiation We were warned for radiation dangers by e.g. Rossi. The radiation was said to be “soft” between 30 keV and 150 keV. A rather sensitive γ-meter was therefore bought and calibrated using a 1 microcurie Ba 133 source. The Ba 133 source gave a significant output. The photons emitted from Ba133 are of many various energies and from that point of view suitable for calibration. The energies emitted was between 3 and 383 keV8. We also wanted to bring the γ- meter as close as possible to the test tube. So we therefore shielded it with an aluminium foil to reflect thermal radiation and also cooled it with water.

Fig. 9. The γ-meter shielded from thermal radiation and water cooled

We however never got any γ-radiation from any of our tests above the background radiation.

8 http://ie.lbl.gov/toi/nuclide.asp?iZA=560133 25 Barium 133 source approaching from behind γ-detector 20 15

10

5 0 0 2000 4000 6000 8000 10000 12000 -5

Fig. 10. No γ radiation above background noise level was detected at any time in any experiment

3.3. Preparates used

3.3.1. Examples of different metal (nickel) powders tested To achieve a maximum contact surface between the metal and the hydrogen the samples used were in some form of powders. The grain size varied between a micron scale to a nano scale. Larger grains or surfaces such as pipe-walls were considered as less suitable for this type of reaction.

Tab. 2 Some examples of the powders used Type Main metals % by weight Grain size From Remark 1 Ni 100 nm Nano Dynamics Unknown quality 2 Ni Israel Carbonyl Nickel 3 Ni 90%, Fe 9%, Li<0,5% 100 nm Own make Nova1 4a C 4.18%, O 23%, Ni 5.4%, Si Ahern EDS- Undamaged particle 3 µm 0,21%, Zr 65%, Pd 0.86% analyze 1 surface 4b C 14%, O 17%, Ni 18,5%, Si Ahern EDS- Average value from 3 µm 1.5% Zr 47%, Pd 1,6% analyze 2 many grains 4c C 10% o 17,5 Ni 20%, Si Ahern EDS- 3 µm Very fractured piece 0.9%, Zr 46% Pd, 2,8% analyze 3

After our first attempts to find pure Ni powders reacting we resorted to ask people, who over the Internet, said that they had powders that had resulted in some form of reactions Brian Aherne9 was one such person (sample 4 a-c). The samples from Aherne were analyzed using EDS-analysis. On some specific spots the content of Pd vas very high. It is clear the Pd was very unevenly distributed.

When using high temperatures in hydrogen during the tests, powders sintered together, forming a still hydrogen-permeable lump. We always tried to prevent NiO forming on the grains. One way of preventing that was to heat the samples to high temperatures – over 360°C in a reducing hydrogen atmosphere.

9 Vibronic Energy Technologies Corp, http://www.scribd.com/mjstrand1/d/39076066-Vibronic-Energy- Technologies

Fig. 11. Powders where the grains have sintered together

3.4. Used pressures and temperatures

The pressures used ranged between a few bars to 200 bars. The temperatures varied between room temperature and about 600°C.

As no sure reactions were found during the experiments only a few tests, with comments of how the experiments were conducted, will be shown.

3.4.1. 2011-04-21 Curt Edström Experiment 1 Pressure 33-29 bar in the hydrogen. Ni 100 nm grains from Nanodynamics - 2.6 gram.

pwr.W temp AI0

160 510 158 508 156 506 154 504 152 502

W 150 500 148 498 146 496 144 494 142 140 492 0 2000 4000 6000 8000 10000 12000 14000 490 tid se c 0 2000 4000 6000 8000 10000 12000 14000

Fig. 12. Supplied power and resulting temperature

The room temperature was 22 +/- 1°C during the experiment. Maybe you could say that the temperature has risen has increased from 498 to 501 °C during the time interval 2 h and 20 min. That rise in this experiment is considered too small to be significant. 3.4.2. 2011-07-25 Curt Edström Experiment 25, reactor 4. . The pressure varied between vacuum and later 13.6 to 13.1 bar after hydrogen had been inserted.

Aherns ZrCuNi mixture (see table XXX). The sample has been baked at 500°C over night in vacuum at 62 W power. This power was kept very constant. The temperature in vacuum has been very stable at 479 °C the whole night (+/- 1deg.).

Hydrogen was inserted at to 13.6 bar, at 64 000 sec. The pressure lowered by 0.5 bar during less than 3 hours (rather small pressure drop). The temperature first increased to 505°C then slowly dropped to stable 498.8°C.

When vacuum was restored again the temperature returned and stabilized around 480°C.

510 14

13.8 505 13.6 500 13.4

495 13.2

490 13

12.8 485 12.6 480 12.4

475 12.2

470 12 60000 65000 70000 75000 80000 85000 60000 65000 70000 75000 80000 85000

Fig. 13. Temperature °C curve A0 (left) and pressure bar (right) after insertion of hydrogen (and then re-evacuated to vacuum).

4. Discussion of results 4.1. Detection limits of measurements

Let us for example take test number 23, which was first tested with a “dummy” gas i.e. argon that was assumed unable give rise to any nuclear reactions. First some data about the experiment:

Run 23, 2011-07-18, first test with reactor 4. 70 W continuous input power. Pressure 2.05 bar cold start. Mass of powder 8.3 g

Channels A0 internal probe, A1 air temperature, A2 probe outside at bottom, A3 outside at top of heater A4 GM tube, A5 Top of flange (near safety valve)

Observations:

It takes about 2 hours before it gets stable. The observed pressure decrease is from absorption only. There is also a pressure rise at 6000 s - this is due to a pressure check. The volume of the reactor was 137cm³ (195mm x 30mm diameter), The volume of the tubing is around 5 cm³ yielding a. total volume of 142cm³. (minus the volume of the test-powder).

Temperatures as function of time (dummy experiment)

600 Internal Probe 0 500

400 Probe outside at bottom 2 300 Probe outside at top 3 200

Temperatures Room temp- 100 erature 1

0 0 5000 1000 1500 0 0

Fig. 14. Test 23 with dummy gas (argon)

The question that arises is: If extra power was added due to a nuclear reaction would that be detected by us as a temperature rise? If we enlarge the steady state part of the figure above we can se that a temperature difference to the surrounding of 538°C to 22°C corresponds to a power of 70 W. That means that a step in power of 1 W would roughly give a step in tempera- ture of 7 K.

Temperatures as function of time (dummy experiment)

540 Internal Probe 0 539 538 Probe outside at bottom 537 2 536 Probe outside at top 3 535 534 533 Room temp-erature 1 Temperatures 532 531 530 10000 10500 11000 11500 12000

Fig. 15. Test 23 above magnified

It is possible to see a small ripple in the temperature reading of the innermost internal probe. The amplitude peak to bottom of that ripple is in the order of 0.2°C. As “inert” gas was used in this dummy test it is highly improbable that the ripple stems from any nuclear reaction. From an inertial point of view only a steps in power that is long enough, meaning that contains enough energy, would be detected. An upper boundary for that sensitivity is when you take the mass of the whole probe times its specific thermal capacity, which is about 300 J/K. A lower boundary would be just looking at the powder itself – 8.3 g · say 500 J/kg K (the powder is very inhomogeneous) means 4 J/K. If we assume that 1 K would be a detectable temperature rise that would correspond to 0.14 W. For such a step to be detected it must have prevailed for a minimum time of say 30 s and if it is not a instantaneous step but rather a slow ramp – say 2000 s. The fusion heat is said to come in “bursts” but the length of the bursts are largely unknown to us.

The same sample was then tested first in vacuum and then with hydrogen gas. Heating started now using 60 W (instead of 70 W). When the temperature had stabilized in vacuum at 477 °C, 10 bar of hydrogen vas loaded. Thereafter the temperature was stabilized for a couple of hours. Due to large amount of data the curves have been divided into three sets - set 1, 2 and 3.

A0

600

500

400

300

200

100

0 0 2000 4000 6000 8000 10000 12000 14000

Fig. 16. Test 23 temperature in powder sample continued in vacuum

At 2600 sec in set 2, H2 was injected with a pressure of 10 bar. The temperature in the sample then rapidly rose to 497 °C and thereafter slowly lowered again to 490 where it stabilized. From 7000-10000 s in set 2, the sample temperature is constant. This indicates no more H2-absorption in the sample explaining the stability. When the gas pressure after that is lowered to vacuum, the temperature immediately drops back to the original temperature in vacuum of 477 °C.

A0

500

495

490

485

480

475

470 0 2000 4000 6000 8000 10000 12000 14000

Fig. 17. Test 23 continued (set 2) introduction of H2

The vacuum curve in figure XX can be explained by three coefficients: - a m·Cp term symbolizing the thermal mass - an U·Cp term for the convective heat transfer from the test tube 4 4 - an A·f·ε·σ term which should be multiplied by (TA0 – Troom ) symbolizing radiation A curve fit gives the following coefficients: mCp= 183 J/K UA= 0.129646 W/K A·f·ε·σ= 7.50E-12 W/K4 The expected temperature in vacuum is then according to the figure below.

Fig. 18. The expected temperature in vacuum When hydrogen is chemically absorbed in a metal 14.6 MJ of heat /kg hydrogen is released10. The pressure in the 142 cm³ hydrogen compartment fell from 10.7 bar to 9.5 bar during the test see fig XXX. According to the normal gas laws:

10 L. Kit Heung, Using Metal Hydride to Store Hydrogen, Savannah River Technology Center, WSRC-MS- 2003-00172, http://sti.srs.gov/fulltext/ms2003172/ms2003172.pdf m pV⋅= ⋅⋅ RT (1) M ∆m ∆⋅pV = ⋅ RT ⋅ M ∆=p (10,7 − 9.5) ⋅ 1055 = 1.2 ⋅ 10 Pa −63 Vm=142 ⋅ 10 M= 2/ kg kmol R= 8314 J / kmol⋅ K TK≈+273 500 1.2⋅⋅⋅ 1056 142 10− ⋅ 2 kg m3 kg s 3⋅ kmolK ⋅ 1 ∆= = ⋅−6 ⋅ ⋅ ⋅ ⋅ = m 5.3 10 22[kg] 8314⋅ 773 m ⋅⋅ s kmol kg m K

The heat developed from that is then 14.6·106 · 5.3·10-6 = 77 J which is far to small a power to have any influence. During the long time of the experiment the absorption power is in the magnitude 10 mW.

Pressure

11.5

11

10.5

10

9.5

9

8.5

8 0 2000 4000 6000 8000 10000 12000 14000

Fig. 19. The pressure during the hydrogen filled period