Minimal cold fusion experiment

Summary

Challenged by one of my physics friends to build a cold fusion demonstration for under £1000, I decided to attempt to replicate the work of Les Case (more here and SRI replication here and Storms talks about his attempts in this interview at 57 min in).

This is an informal log that I'm using to keep track of my progress and to share some of the details of my journey with others. More bite-sized updates can be found on:

Instagram - casualphysics

2018-01-14

• Got some inspiration for the fusion vessel from a cafetière that I saw whilst visiting a friend.

2018-01-26

• Took delivery of Parker stainless steel pipe fitting cross union. The idea is that we'll use one port for thermocouple, one for gas inlet and one for pressure sensor.

2018-02-08

• Took delivery of thermocouple equipment and copper pipe to connect the cross union to the vessel

2018-02-09

• Attempted to put the vessel together and attach the thermocouple. The high temperature sealant was however far too flexible - it seems unlikely to survive one we have 4 atm pressure inside

• So far I've spent £169 on
• Pressure transducer
• Cafetiere
• Solder
• Stop ends
• Brass Male/Female BSP/NPT Adaptor [1/8" BSPT Male x 1/8" NPT Female]
• Parker Stainless Steel Pipe Fitting Cross Union + Fitting Straight Coupler, 1/8 in BSPT
• Male x6mm OD
• copper pipe
• Thermocouple + circuit amplifier
• 6mm Compression x 1/4 BSP Male Adaptor
• Stainless Steel Ball Valve

One of my physics partners is becoming less available to help so perhaps need to rethink this approach of building everything from scratch on a shoestring budget.

2018-05-01

• Took delivery of this power source which can give max 30V 10A which cost £48. It will be used to power an electrical heater made of Kanthal wire which will wrap around the vessel and hopefully get it to 200C

2018-05-07

• In the work of Les Case he uses 55V, 1.67A. I don't have as much voltage but I can adjust the length of kanthal wire and voltage from the power supply to give the same output power (or at least close enough).
• I started with the entire kanthal wire and as you can see it's far too resistive to allow much current to flow - I only have about 0.75A at the max voltage of 30V.

• Using the voltage and current readings to get a sense of the resistance of the wire, I cut the wire so as to allow a more useful amount of current to flow. As you can see we have about 200W of power flowing - enough to cause the wire to glow red hot.

• I had a few issues to start with, because the cables connecting everything together were far too thin. Below is a link to a YouYube video showing my progress

2018-05-27

• Tried to wrap the kanthal wire around the vessel using some ceramic fibre blanket made of aluminium silicate as insulation to stop the current from going through the stainless steal. While the blanket was supposed to be able to withstand temperatures much higher than 200C it appears that it was not able to withstand the heat from the wires.

• In addition, after touching it with bear hands (big mistake) I felt like I had been cut. Also I began to cough a lot. Clearly not a safe material. Decided to not use this again. Need to find an alternative way to heat the reactor.

2018-06-26

• To connect the vacuum pump, I needed some supplier specific connectors and some tubing. It was cheaper to get these by buying a new vacuum vessel rather than go to supplier for individual parts!
• After thinking I would need to source some new connectors to create an air-tight seal between the vacuum vessel and the plastic tubing, I found that the size of the port and tubing allowed me to basically screw on the tubing to the vessel

• Although the autoclave reactor pressure sensor isn't sensitive enough to tell me how good the vacuum is, you can tell by using your ears that the vessel does hold vacuum. You can see and hear this in the video below

2018-07-01

• Vacuumed out autoclave at 13:00 - will test whether vacuum is still held in 24 hours

2018-07-02

• Opened the gas valve on the autoclave at 13:00 and heard a hissing sound. This suggests the vessel can hold some vacuum after 24 hours.

• The autoclave pressure gauge is in units of MPa whereas vacuum pump in units of mbar. This makes it impossible to see how good the vacuum is by looking at the autoclave gauge. To get a quantitative sense of how much pressure remains in the autoclave after some time I have tried the following hack

• close one gas valve on the autoclave while leaving the other open
• pump out the air
• turn off the pump
• for about 3-4 sec the pump holds the vacuum
• after this time the pump starts to let air in again, but it does this slowly
• when the vacuum pump gauge is not yet at 1 atm, close the second gas valve on the autoclave leaving a partial vacuum inside
• turn on the pump again, but this time because both valves on the autoclave are closed only the air in the connecting tube is removed.
• switch the pump off and quickly open the gas valve on the autoclave that is connected to the vacuum tubuing.
• for 3-4 sec the pressure gauge on the pump jumps up to partial vacuum and gives enough time to record the reading
• Using these ideas, it appears that we can get some quantitative sense of how good the vacuum in the autoclave is after some time.

• We must keep in mind the volume of the connecting tube will affect our readings - making the pressure seem lower than it actually is. We need to eventually take this into account.

• Vessel has been vacuumed down at 13:45. I'll return to this after work tomorrow, approx 19:00

2018-07-05

• Didn't get opportunity to measure the pressure the last couple of days. This morning at 09:00 I performed the pressure measuring procedure outlined in the last log entry (2018-07-02).
• The pressure read 660mbar after approximately 67 hours

2018-07-06

• 07:35 - vacuumed out the vessel again. Will measure pressure tomorrow morning

2018-07-07

• 08:45 - Measured pressure to be approximately 270mbar after about 25 hours.

• Unfortunately I was not able to get a great angle to take the picture quickly enough before so there is a bit more uncertainty on this measurement compared to the first one.
• I have vacuumed down the vessel again and will measure the pressure again tonight after about 12 hours.
• 20:45 - After 12 hours the pressure measures around 80mbar

2018-07-08

• 08:45 - I want to get an indication for how much of the pressure that I measure is coming from leaks vs out-gassing from the vessel walls. So, I have vacuumed down the vessel again and in 12 hours I will

• measure the pressure of the vessel as I have been doing
• but before the pump goes back to atmospheric pressure turn it on again to vacuum down the vessel once more
• then come back in a further 12 hours and measure the pressure to see if it is close to the 80mbar measurement. If so then we can attribute the loss of vacuum to leaks, if not then it suggests that there are gases coming out of the vessel walls.
• 21:15 - A little late, but the pressure measurement for 12.5 hours is close to 80mbar which is consistent with the previous 12 hour measurement done on 2018-07-07

• I will come back tomorrow at 09:15 and perform the pressure measurement again

2018-07-09

• 09:20 - Unfortunately I dropped my phone whilst trying to take the photo so got a blurry pic. However you can see that after a further 12 hours the pressure reads about 60mbar.

• We might therefore ascribe the 20mbar discrepancy to out-gassing coming out of the walls of the container? However it could be that I just tightened the gas valve more this time around? Difficult to tell at this point.
• I have vacuumed down the vessel again and will return in the morning to perform the same procedure, but rather than a 12 hour period it will be 24.

2018-07-10

• 09:15 - Measured pressure to be approximately 180mbar after about 24 hours. This is very different to the previous 25 hour measurement of 270mbar on 2018-07-07 - could this be because of outgassing? I have been continually vacuuming down the vessel for the last week so the vessel hasn't had time sitting open to the atmosphere for more than a few minutes.

2018-07-(16 -19)

• Similar measurements for roughly 24 hour periods yield 240mbar, 260mbar, 320mbar
• Without a way to be consistent with the way I tighten the valves I'm not sure what I can conclude from these numbers
• Below is the data gathered so far

2018-08-28

• Took two months, but the heating pad finally arrived from Omega!

2018-09-24

• The heating back runs on 120V AC which meant I needed to acquire a transformer to make it work without destroying it.
• The transformer I bought can handle up to 200W which is fine for this heating pad which uses 138W
• The heating pad got up to 135C

2018-09-27

• Luckily for me the heating pad bends just enough to make it wrap around the reactor.
• I presumed that because it is made of silicone that it would easily bend to any shape but turns out there is a maximum curvature that it will take.
• Although the pad does bend, it will spring back into flat shape unless held in place. I bought Jubilee/hose clips to accomplish that.

• After about 10 mins the temperature got up to 113C.

• I didn't leave it on for very long so next step is to leave it on for a few hours to let the whole thing reach equilibrium

2018-10-01

• The temperature outside got up to a max of 170C
• It was impossible to measure the inside of the chamber, i.e. the teflon inner cylinder, because the outside steel had expanded making it impossible to twist open.
• I had to wait while before I could open it.

• I initially thought that for this testing stage it would be enough just to use the infra-red gun for temperature measurements, but it appears that I'll need to get the thermocouple working for internal measurements.

2018-10-19

• After taking delivery of some small screw drivers, a breadboard and some Dupont/jumper cables I was able to connect the thermocouple to a Raspberry Pi that I bought for a non fusion purpose.

• Unfortunately, after installing the required software on the Pi and connecting the pins as per the AdaFruit instructions I was always getting a zero reading from the sensor using their example python code.
xxxxxxxxxximport timeimport boardimport busioimport digitalioimport adafruit_max31855​spi = busio.SPI(board.SCK, MOSI=board.MOSI, MISO=board.MISO)cs = digitalio.DigitalInOut(board.D5)​max31855 = adafruit_max31855.MAX31855(spi, cs)while True:    tempC = max31855.temperature    tempF = tempC * 9 / 5 + 32    print('Temperature: {} C {} F '.format(tempC, tempF))    time.sleep(2.0)
• I checked that power was indeed flowing to the thermocouple amplifier using my multimeter and all seemed fine.
• I decided to dig into the adafruit_max31855 package from Adafruit_CircuitPython_MAX31855. I found that on line 63 an all zero byte array is defined which is then supposed to be updated on line 67as the Pi reads in the data from the sensor. However my byte array is never updated. If I completely disconnect the Pi from the sensor (ie pull off all the cables from the Pi pins) I still get a zero reading.
• This makes me think that the Pi is not talking to the sensor chip at all. I need to learn more about how the Raspberry Pi talks to other devices before I can continue as I feel like I must be missing something quite fundamental.

2018-10-24

• For reasons that escape me today was the first day that I tried to test feeding the thermocouple into the reactor. It turns out that even when one of the valves is open it is not possible to feed anything through.

• This means that rather than simply attaching a compression gland to the valve I'll need to start modifying the reactor, i.e. remove one of the gas valves.

• To test the ability of the reactor to hold pressure I put some water inside and turned on the heating blanket.

• Previous heating experiments suggest that the inside of the reactor might reach 150C.
• At 150C the vapour pressure of water is 4.7 atm. The vessel can hold 60 atm at 250C so this seems safe.
• To figure out how much water to use I started by imagining 1 mole of water (i.e 18g) turning into to vapour and being held at 150C in the reactor vessel of volume 200ml. Without considering saturation, $PV = nRT$ gives a pressure of around 175 atm. This tells us that there will be more than enough particles in 18ml to generate the 4.7 atm vapour pressure when the vessel becomes saturated. So I'll use that.
• After sealing the vessel and heating the water for 1.5 hr there was no noticeable change to the pressure. I'll try again tomorrow when I have more time.

2018-10-25

• Starting at 07:30 with ambient temperature of 21C.
• It takes a couple of hours (09:45) for the pressure gauge to reach 3atm and the outside blanket temperature to reach 174C

• 11:15: Blanket T = 184C, P = 4.9atm (See below)

• 12:00: Blanket T = 182C, P = 5.2atm. At this point the morning sun was off the reactor and I got the impression that the reactor was reaching a plateau.
• At 12:15 I decided to insulate the reactor to see how this would change things

• 12:26 Blanket T = 213C, P = 5.5atm. The blanket temperature rapidly increased which worried me because the maximum temperature according to the manufacturer (Omega) (see also here) is 232C.
• 12:30 I removed the insulation.
• 13:45 I applied some insulation around the top of the reactor. At 14:13 the P = 6atm.

• 15:20 Blanket T = 188C, P = 6.9atm. Inferring the internal temperature from the saturated vapour pressure gives Internal T = 164C. I decided to insulate the bottom of the reactor at this point too.
• 15:49 Blanket T = 188C, P = 8.5atm, Internal T = 173C.

• 16:18 Blanket T = 194C, P = 10atm, Internal T = 180C.

• 17:51 Blanket T = 194C, P = 12.25atm, Internal T = 189C.

• After the reactor cooled down it loosened by itself. This has safety implications for the hydrogen that might leak out. The cause might be that the different parts of the reactor are cooling down at different rates and so the metal contracts at different rates.

2018-12-12

• I was advised back in November that the reason the thermocouple wasn't working was very likely due to a bad electrical connection.
• I wrongly assumed that since I could measure the voltage from the Pi that was powering the chip that everything was fine. However, it turns out that the voltages sent to the Pi to tell it to read the sensor are much smaller than the 5V that powers the chip. These smaller voltages are therefore more sensitive to "errors" coming from bad connections.
• So, I need to solder the connections like these below:

• My dad helped me do the soldering

• Below is the finished result.

• The thermocouple now produces sensible temperature readings as you can see below.

2018-12-13

• I also decided to do a double check of the thermocouple by comparing the readings to a traditional thermometer - they were in agreement (to the extent that I was able to read the thermometer by eye).

2019-02-09

• Today Dad and I aimed to insert the thermocouple inside the reactor. This required us to remove one of the gas valves.
• I needed some of dad's tools and equipment.

• Turns out a 11/16 inch spanner was good enough to remove the value without too much force once it was clamped.

• It was harder than expected to feed the theromcouple through the pipe. We had to feed a smaller wire through first, solder it to the thermocouple in order to drag it through.
• Upon dragging the thermocouple through, the outer protective coating appears to have become snagged on something inside the pipe, leaving the red and yellow wires exposed. We speculate that perhaps the pipe is rougher on the inside than outside.

• Despite the problems dragging the cable through, I had no problems connecting the thermocouple to the Raspberry Pi and start taking internal measurements of the air inside the reactor.

• Note, now that the gas value has been removed I can no longer vacuum out the reactor. The next step is either to fill the pipe with something like Torr Seal, or find a compression gland.

2019-02-20

• I found a compression gland from RS Componenets
• At the time, I could only find glands for small probe diameters with 1/8" BSP so I had to buy a 1/8" to 1/4" convertor (I determined the size I needed from this pipe guide)
• I also bought some PTFE tape to ensure I got an air tight seal

2019-03-10

• After making several mistakes along the way by getting confused between parallel vs tapered threads aka G vs R type (I found out that I need parallel G type), I got the right components and was ready to seal the vessel
• Since there are 2 thermocouple wires I decided to wrap them in PTFE tape in the hope it would create a more circular profile so the compression gland could create a good seal

• Unfortunately, the seal was terrible :-(
• Need to figure out where the leak is coming from next

2019-03-16

• I tried lots of different tests, e.g. putting PTFE tape around the end hole to check whether it was the thermocouple that was causing the issue…it was not the issue :-(

• I put PTFE tape around every possible gap/hole I could think of. Nothing improved the leak

2019-03-19

• I realised that the type of compression glad I had bought was supposed to be used with a solid metal thermocouple like this

• I bought a 1.5mm one from RS components along with a compression gland of the correct size with a 1/4" fitting
• When I tried to put everything together, I realised the problem was the fitting wasn't 1/4" BSP (G) but instead a metric M14. The mismatch was causing the fitting not to screw all the way to the plastic gasket i.e. this was the cause of the leak.
• I need to find a convertor to go from 1/4" BSP to M14.

2019-04-09

• Ive finally found some BSP to metric convertors from AliExpress. However they were R type, i.e. tapered so I had to get a convertor from RS components.
• Because I could only find the convertor from 1/2" R to 1/4" G I had to get the 1/2" from AliExpress.
• Now I wait

2019-05-10

• After waiting for the Chinese components to arrive and moving house I discovered that the 1/2" female to M14 male convertor that I thought was R type on the female side is actually G type.

• My RS components convertor was 1/2" R to 1/4" G so this won't work now.
• I have ordered another Chinese convertor from AliExpress - this time 1/4" female to M14 male convertor. This should be able to directly interface with the 1/4" compression gland that I bought.
• Now I wait... again

2019-05-16

• In preparation for putting the thermocouple in the vertical port of the reactor I've moved the pressure sensor to the side.
• It turns of that the gasket for the vertical and horizontal ports are not the same so I am currently testing whether vacuum will hold with the pressure sensor in the new location
• I am using a M14 bolt to seal the top port

• I vacuumed down the reactor at 13:30
• After 4.25 hrs (at 17:45) the pressure was 180 mbar
• I vacuumed down the reactor again at 17:50 (making more effort to tighten all fittings)
• After 5.5 hrs (at 23:20) the pressure was 130 mbar
• When compared to July 2018 these pressures today seem to be about 5 times higher if you extrapolate the historical data. I'll try doing a 10 hr test to better compare with July 2018 results.

2019-05-17

• I vacuumed down the reactor at 09:30
• After 13 hrs (at 22:30) the pressure was 260 mbar

2019-05-20

• I cleaned the vessel and thermocouple with isopropanol before putting the thermocouple in position. I did this because once I tighten the compression gland around the thermocouple I won't be able to move it easily again so I want to make sure it's free of grease that might mess up the experiments later.

• Then I sealed the vessel

• I then vacuumed it down at 09:20.

• 2.5hr later at 11:50 I measured the pressure to be 80 mbar

• I tightened up the fittings and vacuumed down again at 12:00

• 10.5 hr later at 10:30 the pressure read 200 mbar. this is about 2.5 times higher than the best vacuum experiments I did when I first got the vessel.

• The above pressure is similar to the 260 mbar I measured at 13 hrs on 2019-05-17. This is good news. It suggests that I'm not causing any leaks by using the compression gland.

• I could probably make the vacuum better, but in the end I will be working with high pressure where the vessel components will all tighten up and seal better. I therefore think that this level of vacuum is probably ok for the purpose of getting rid of the oxygen so that I can be sure I won't have any accidentally hydrogen+oxygen explosions

• Next I will repeat the water experiments from 2018-10-25

2019-05-21

• I put a shot glass of water inside the reactor and sealed it.

• At 14:30 I turned on the electric heater. The ambient temperature was 22.8C

• After 1 hr (at 15:30) the temperature is 125C but the pressure gauge still remains at atmospheric. In the previous experiments of this kind it took about 2 hrs for the pressure to start to change.
• At 15:45 the ambient temperature had risen to 25.7C
• At 16:50 the ambient temperature had risen to 26.5C
• At 19:30 the ambient temperature had risen to 27.4C
• I'm almost ready to start putting cold fusion materials inside - I just need to buy some insulation in order to reach 200C

2019-05-27

• I took delivery of some essentials

• disposable gloves
• In addition I also took delivery of a digital scale capable of measuring milligrams

2019-06-01

• Just like the ceramic fibre I received on 2018-05-27, this one was very glassy to touch and made me uncomfortable about the prospect of handling it. I've seen people use ceramic blankets before so I must not be buying the right one - back to the drawing board.

2019-06-11

• I visited Alan and Russ at the fusion farm. They sold me a hydrogen generator for £400. Alan is "shipping" the generator to me as it was too heavy to carry home.
• In addition he sold me a Geiger counter for £40

• I can use the counter to detect radiation coming from any tritium produced in the reactions. While the radiation coming from cold fusion is always small, I'm hoping I'll at least me able to detect something above background - we'll see.

2019-06-13

• After being advised by Alan and Russ, I bought some ceramic fibre blanket from a pottery supplier
• This blanet says it is bio-soluble. After doing some digging I finally appreciated that this type of ceramic blanket is best for humans

• It is indeed much better - not glassy and not scratchy
• I also took delivery of some fire bricks from the same pottery supplier

• Now I'm ready to try for 200C in my reactor

2019-06-18

• I took delivery of the hydrogen generator I got from Alan

2019-06-20

• Today I will try and reach 200C in the reactor using the ceramic insulation and firebricks I recently bought

2019-06-27

• Because The coconut shall based palladium catalyst is no longer available to buy I'm going to try using similar materials to Mizuno's most recent work
• I've ordered some nickel palladium a week or so ago.
• Mizuno uses nickel mesh but I am curious to try both mesh and foil
• Today I delivery of 1g of palladium which cost £50

• I also took delivery of some deionised water for the hydrogen generator.

2019-07-03

• Took delivery of some Nickel 200 mesh with 180 wires per inch at 0.055mm diameter. This was provided by Alan Smith from the LENR forum. He has contacts in China who he believes are the people supplying Mizuno with his mesh.

2019-07-05

• In order to get the hydrogen from the generator into the reactor so that no air accompanies it I need to have the vacuum pump and generator connected to the reactor gas port simultaneously.
• I have decided to take the hit and buy Swagelok components, including port connectors, t-junctions, needle valves etc. The total cost is £226.

2019-07-30

• I decided to bite the bullet and buy some connectors and valves from Swagelok. The idea is to connect my reactor to a T like junction so that I can vacuum out the air from one side and pump the hydrogen in from the other

• I spent £226 in total on the following components (ordered on 2019-07-05)

• The final results of putting this all together are below

• I was also advised that degassing the water before putting it into the hydrogen generator was a good idea - apparently this speeds up the electrolysis (see comments on this Twitter post). The easiest way to do this is to boil it
• Notice in the above animation that I'm now using distilled water not deionised water. The hydrogen generator manual actually calls for this. The difference appears to be that deionised can still contain things like organic molecules which doesn't sound like a good idea to have inside the generator. That being said, the generator cap has a small hole in the top so who knows what kind of dust is inside it!
• The hydrogen generator manual called for 180g of KOH mixed with 400ml distilled water. Mixing the two generates some heat so you have to wait until it cools down. I aded a further 1200ml of distilled water, making a total of 1600ml. This filled the tank 2.5 levels above the minimum (denotes by L on the generator). Unfortunately it was hard to see what the minimum amount of water was. I suspect it is around 1L.
• To check the hydrogen generator is working as expected and has no leaks, the output is closed and the pressure is allowed to build until it reaches a maximum. At this point, the digital flow meter should read zero. The generator passes the test with a maximum pressure or about 4.5atm as you can see in the YouTube video below.

2019-08-01

• Attaching the flexible pipe to the barbed Swagelok fittings was somewhat successful

• The silicone pipe intended for connection to the hydrogen generator provided sufficient resistance but was able to be pushed on all the way

• The pipe that connects to the vacuum pump was extremely tight and its ability to flex was much less than the silicone. I had to put significant force on the Swagelok fittings to get it over a few of the barbs. As you can see, it is on on all the way. At present I'm not sure what to do about it. I did conduct some simple vacuum tests which suggests the seal was high. More tests need to be done.

2019-08-03

• Took delivery of fire protection equipment.

• I do not expect to have any problems with the hydrogen since I evacuate the air from the reactor before any hydrogen goes in. It is nevertheless a good precaution.

2019-08-04

• I recalled recently that I bought a cheap pressure transducer back at the start of 2018 (a bit like this) that can measure up to 1.2MPa. Turns out it had a 1/4 fitting that I could use with my 1/4 -> M14 adapters! So, inspired by this YouTube video I bought an Arduino Uno and connected the pressure sensor to it

• To turn the voltage into a pressure reading I am assuming a linear response between voltage and pressure. Unfortunately I do not have access to the official documentation to confirm whether this is true or not.
• I also needed to appreciate how the arduino deals with analogue voltage signals. It takes a reading from 0 to 5V and turns it into a signal value between 0 and 1023, i.e. it can measure voltage in 4.8mV incremenets.
• My task was to find the coefficients for the following equation $P = mS+c$ , where $P$ is pressure, $S$ is sensor value.
• I vacuumed out the reactor and measured $S = 33$
• At normal atmospherical pressure, $S = 102$
• Therefore, $m=10^5/69$ and $10^5 = \frac{10^5}{69}\times 102 +c$ implies that $c = 10^5\left( 1-\frac{102}{69} \right)$
• Therefore $P = \frac{10^5}{69}S+10^5\left( 1-\frac{102}{69} \right) = 10^5\left[ \frac{1}{69}(S-102) +1 \right]$
• From this you can see that the increment in sensor value corresponds to an increment of pressure (i.e. an uncertainty) of $10^5/69 \approx 1500Pa$
• I have recently purchased a high temperature pressure transducer from China that can measure up to 6MPa. The reason I went so high is that this is the highest pressure that my analogue sensor can measure and is the limit of the reactor. However, this comes at a cost. The larger pressure range of the new sensor, i.e 6/1.2 = 5 times larger, will be mapped to the same voltage range and so the same sensor values. This means my pressure increments will be 5 times larger, i.e. 7500Pa (0.075 atm)
• Because this Ardunio doesn't have in built wifi, I am connecting it to the USB port of my Raspberry Pi, transferring the data via the serial port and then sending it to firebase.

• At 22:10 I vacuumed down the reactor and then left it for 35 hours to see how leaky it was

• From the gradient of the line of best fit I can say that I gain about 5.4mbar of pressure per hour.
• This should be compared to the manual version of the same graph that was made on 2018-07-(16 -19). Superficially it appears that the vacuum we are now obtaining is about twice as good.

2019-08-06

• Today was the first hydrogen test.

• Over the course of the day I established and tested the following protocols I made to evacuate, fill and vent the reactor.

Key

H - Hydrogen generator valve

R - Reactor valve

V - Vacuum pump valve

Evacuate reactor

1. Close H
2. Open R
3. Open V
4. Switch on vacuum pump
5. Wait for 2 minutes (note, I should create an automatic notification when pressure sensor get's low enough )
6. Close R
7. Close V
8. Switch off vacuum pump

Fill reactor with hydrogen

1. Evacuate reactor (follow the protocol - H, V and R should be closed at the end of this)
2. Open H
3. Switch on the hydrogen generator
4. Wait until the generator pressure gauge reads 2 atm
5. Switch on the vacuum pump
6. Slowly open V a little bit and let the generator pressure gauge drop to somewhere between 0.5atm and 1atm. Hold for a 1 min. Don't open it full otherwise the generator will be exposed to full vacuum and it might break. This process allows the air present in the hydrogen tube to get flushed out.
7. Close V
8. Switch off vacuum pump
9. Open R
10. Wait until pressure gauge reaches the desired value
11. Close R
12. Switch off the hydrogen generator
13. Disconnect vacuum pump tube from the pump
14. Place pump connecting tube out of the window
15. Open V (You will hear a hissing sound as the high pressure hydrogen in the tubes vents to atmosphere)
16. Re-connect pump tube
17. Switch on the vacuum pump very briefly (until pressure goes down to 0.25 atm). This is to suck out most of the hydrogen remaining in the tubes. Don't let the hydrogen generator be exposed to full vacuum as it might break it.
18. Switch off the pump
19. Close V
20. Close H

Vent reactor

1. Wait until it reaches low temperature (NOT SURE HOW LOW YET)
2. Close H
3. Open V
4. Disconnect vacuum pump tube from the pump
5. Place pump connecting tube out of the window
6. Open R - this will vent the hot high pressure hydrogen gas to the outside
7. Re-connect pump tube
8. Switch on vacuum pump
9. Wait for 1 min (this will remove remaining hydrogen that's inside the reactor and connecting fittings)
10. Switch off pump
11. Allow reactor to reach atmospheric pressure
12. Close V
• Before I had arrived at the above protocols I discovered that I need a clamp around the silicone tube connected to the swage log fitting. At 4 atm pressure reading on the hydrogen generator, the tube popped off the reactor. I need to order some larger clamps (10mm or 11mm) - the ones I bought for the hydrogen generator end of the tube are too small (9mm)
• I also testing the vacuum pump for "continuous" use. I left it on for 15 minutes and it seemed perfectly happy. It did however rattle quite a lot (not that it was rattling even in the early days when I bought it)
• Some kind of smoke occasionally appears when using the pump. I read online and this is usually when evacuating lot of air. It mixes with the oil to form a vapour that it observed.
• Vacuum pump on sample chamber seems to only get to -29.5 "Hg, i.e. 0.5"Hg above vacuum ie 1700Pa. Considering the pump should get to 0.01Pa this is 100,000 poorer than it should be.

• I also tested this on the small vacuum chamber I bought from BacoEng some time ago (using the original plastic tubing that it came with) and found similar readings.

• I then took off the pressure sensor from the BacoEng chamber and attached it to my reactor. When I vacuumed down again I got better readings in several orientations (apparently the pressure reads differently if it's horizontal or vertical)

• This strange behaviour did not persist. Later the reading was again consistent with the -29.5"Hg saw at the start of the day. This suggests that these pressure sensors with oil inside need time to settle down before giving an accurate reading.

• Could it be dirty oil? No, the oil appears to be as clear now as when I bought it. This make sense as I've not used it that much.

• What I can say about my vacuum pump is that it was never the same after it was shipped to Canada and then back again. It is therefore very possible it's just a bit broken inside.
• The pressure transducer behaved in a strange way (at least it seems strange to me right now). When the hydrogen was filling the reactor, the sensor would read 1 atm pressure for a long time (while the pressure sensor on the hydrogen generator was reading beyond 2 atm). Then suddenly the pressure reading would jump suddenly by 1 to 2atm.
• There was an obvious hydrogen leak that can be seen in the data

• In addition, when I first filled the reactor it appeared not to be possible to go above about 3 atm. Later on I refilled it and it went over 4atm with no problems. This inconsistent behaviour is encouraging me to take things slowly with the hydrogen tests.
• On the leakiness of the reactor - It appears that the reactor is reasonably happy at 2atm at room temperature. This reactor is designed to work at high temperature. I therefore expect that it'll be less leaky when we heat it up a bit.

2019-08-07

• Today I wanted to see if I could create an insulated "container" to allow the reactor to go above 200C (this is the target from the original Les Case work). My previous best was 194C
• I cut out sections of the bio-soluble ceramic fibre I bought recently, making sure to wear a mask and disposable gloves for extra protection. This fibre is safe for humans but I wanted to avoid irritation as much as possible

• Despite being careful, the fibre does inevitably go everywhere. Because the main cutting was done in the lounge I was able to vacuum up the little bits and also keep the room well ventilated.
• When I transferred the fibre to the physics bench in my room, I was careful to avoid fibre getting on my bed. However, in constructing the "container" bits continue to fall off. I was wearing shorts and T-shirt at the time and I did feel irritated by the fibre. Clearly this material shouldn't be handled regularly. I will consider buying a metal bin sufficiently large to house the fibre and the reactor. This means I never need to handle the fibre, I can just lift the reactor in and out of the bin when I want to conduct experiments.
• The final result looks like this:

• I started at around 10:45.
• The maximum temperature reached was 222C, beating my last record by 28C!

• The earliest rapid temperature rise was due to an error on my part - I accidentally connected the heating pad to 240V rather than 120V.
• Let's see how closely the corresponding pressure curve (below - ignore wrong labels if you follow the link to the original data) matches up the the ideal gas law PV = nKT. Specifically, at around 10:58 the temperature was $160C$ and the pressure was $1.5\times 10^5 Pa$. From ideal gas law, $\frac{P_1}{T_1} = \frac{P_2}{T_2}$, we have $\frac{10^5}{300} = \frac{1.5\times 10^5}{T_2}$ giving $T_2 = 450K = 177C$ . This corresponds to an error of 10%. I need to look into this to see whether this is reasonable considering the cheapness of the sensor.

• When the reactor reached its hottest, the inside of the insulation reached 105.5C, the outside of the insulation reached 50.6C. The top of the insulating block that the reactor was sitting on was at 100.9C, the bottom of the block was at 41C and the black plastic of the pressure sensor was at 110C. This was all taking place under ambient temperature of 24.2C
• At 14:40 I decided to remove the lid of the insulation to see how hot the reactor would stay. It dropped below the previous record of 194. This shows that the top insulation is critical. This is to be expected - heat rises after all.
• At 15:30 I turned off the heating blanket and let it cool.
• Somewhere below 60 but above 36 the vessel loosened to the point where I could easily twist it with my hands. I need to do some more tests to find out more precisely when this loosening happens so that I don't have hydrogen leaking unexpectedly.
• In other news, I also took delivery of the Japanese Washington liquid that Mizuno uses to clean his samples

2019-08-13

• The first test with hydrogen suggested that the reactor is leaky. While hydrogen is a particularly leaky gas, my previous water vapour tests suggested the reactor was leak tight. So, what's going on?

• There are some key differences between the water tests and the hydrogen ones:

• Hydrogen was pumped into the reactor cold - my reactor is designed to seal best when hot
• It would have been hard to detect a water vapour leak by just looking at the analogue pressure sensor because there was ample liquid water in the reactor that could have replenished the vapour if some was escaping. No such replenishment could occur with hydrogen.
• I'd like to limit the amount of water that goes into the reactor as I've been told it takes a while to bake out. Therefore, the only choice is, in principle, to try hydrogen in the reactor when hot and see if it continues to leak....but....hot leaking hydrogen doesn't sound like a good idea.

• Instead of testing with hydrogen in the hot reactor I will instead test with helium.

• Asda sell Helium for £20 and once you take off the balloon adapter it has a 1/4" fitting that matches my vacuum pump cable perfectly!

• In general, using a gas tank without a regulator is a bad idea. However, I was told that these party tanks don't hold so much pressure. On the box it says total volume $0.2m^3$. The tank dimensions are approximately radius = 0.1m, height = 0.28m, which gives a compressed volume of approximately $0.0088m^3$. This gives a compression factor of about 23 and so we can expect the gas inside to be at a around 23 atm.
• My reactor could comfortably withstand 22atm of pressure, but I will obviously not expose it to the fully opened cylinder. Instead I will carefully open the valve a little bit at a time to attain some fraction of this.
• I will obviously have very little control over the final pressure, but this is not important for this leak test.

• I started out with the reactor cold, just to compare to the hydrogen tests the other day. As you can see, the pressure rapidly drops within minutes.
• When I opened the helium tank more, I noticed a hissing sound coming form the top of the reactor, near where the metal pipes go into the lid. When I placed my hand near the sound, I could even feel the motion of air around it. This is a very good candidate for the leaks. Interestingly, no such leaks were coming from the swagelok/vessel interface - I have been concerned that leaks might be due to the ability of the "Christmas tree" to swivel at this interface...but seems like it might be fine.

• I decided that while I had the 3atm of helium inside I would start heating the reactor and see how the pressure responded. The pressure should climb as the gas heats and indeed it does.

• Unfortunately my data acquisition from the Arduino is still not quite right.

• The gap in the data came from the python code crashing when it failed to turn a string to integer. The Arduino was sending numbers like "12345.67.89".
• The stepping up of the pressure appears to be due to some kind of mismatch between the sending of data by the Arduino and the reading by the Raspberry pi. I believe this because later in the day when I knew the pressure had risen and the pressure sensor was not responding, I stopped the Raspberry pi python code and restarted it. This immediately solved the problem. This suggests some kind of data backlog is occurring. I need to investigate further.
• At around 15:30 I let helium out in preparation for vacuuming and refilling with helium

• This is the first time I have vacuumed down the vessel when hot. There were no issues. I did notice the vacuum I could pull was not as good as when it was cold. This is to be expected due to outgassing. I will take a proper look at vacuum conditions when the reactor is hot another day.

• I took the reactor to around 731kPa at 16:05 and left it alone for a few hours.

• After 1 hour, the reactor had lost 2% of its pressure, i.e. around 15kPa / hour. This contrasts with the loss of 200kPa (from 500kPa to 300kPa) in 1min earlier in the day when the reactor was cold. Clearly the hot reactor is many orders of magnitude better in terms of leakiness.

• After 2 hours, pressure was at 707kPa - a loss of 3.3%, i.e. 24 kPa, (12kPa/hour)

• After 3 hours, pressure was at 697kPa - a loss of 4.7%, i.e. 34 kPa, (11kPa/hour)

• At 19:36 (just before I turned the heater off), pressure was 692kPa - a loss of 5.3%, i.e. 39 (11kPa/hour)

• You can see something interesting happened at around 21:30 - the reactor suddenly loses 200kPa of pressure in the space of about a minute as you can see below

• There is also a slight temperature dip a few minutes before. Remembering that, at the moment, there appears to be a delay in the pressure data due to some kind of mismatch between Arduino and Raspberry Pi, the little dip around 60C could be the moment that the reactor loses its containment due to different cooling rates of the lid and body.
• On 2019-08-07 we found that the loss of containment was below 60 but above 36 - it now appears that it might be just below 60. Note that when I went to check the reactor before bed the lid was not super loose as it has been in the past. This suggests that reactor containment fails before this point. I should definitely vent the reactor before we get too close to 60C
• Note that there is also a jump in pressure data around 20:00. I suspect this is again due to data lag, i.e. the reactor starts cooling down but the data is lagging behind and then all of a sudden it catches up and misses some data in between. This is speculation at this point - I will try and find out more tomorrow. The reason I am more convinced about the second jump is that I can see many data points as the pressure reduces, i.e. this is not a discontinuity.

2019-08-14

• Upon reading some articles online about Arduino talking to Raspberry Pi, I realised that error in my python code. I was using something like this:
xxxxxxxxxximport timeimport serialwhile True:pressure = serial.readline() # this brings data from the Arduinotemperature = max31855.temperature # this brings data from the thermocoupletime.sleep(0.5)
• The time.sleep came originally from the thermocouple code that reads the temperature. When I only measure the temperature, this is needed to regulate the data capture. However, if I'm measuring the pressure from the Arduino as well as the temperature, then the serial.readline() should regulate everything, i.e. no need for time.sleep, the data capture rate is then determined by the code put onto the Arduino.
• After removing the sleep, I set the Arduino data capture time to be every 500ms. I then vacuumed down the reactor a couple of times and also filled the reactor with helium (when the reactor was cold). The data come through in real time with no delays like I've seen before. Hopefully this has fixed the problem.

2019-08-15

• I received larger clamps capable of fitting over the swagelok barbed fitting - the tube is now secure
• I let the reactor heat up with air inside, then vacuumed out and then put around 355 kPa of hydrogen inside at around 11:30

• Initially the reactor was losing around 5kPa / hour. This is half the loss rate that was seen with helium. It is hard to tell whether this is because helium is a more leaky gas than hydrogen (well known) or whether this is due to the pressure being half of what was used in the helium tests. Unfortunately I have very little control over the helium so cannot do a more detailed comparison
• From 4-6pm, the loss rate was around 4kPa per hour
• At 22:00 the pressure was 314kPa. Over the 10.5hr the reactor lost 41kPa, i.e. 3.9kPa/hour on average (1% lost per hour)
• Overall, I'm happy that the losses are half what they were using helium. I would like to improve the containment further so I will attempt lower pressures and also clean the reactor to see if this improves things.
• After I vacuumed out the hydrogen, I left the reactor in its vacuumed state for 3 days. I expected to see the pressure increasing towards atmospheric in a similar way to what was seen on 2019-08-04. However, the vacuum level remained constant over the 3 days - this was very surprising.

• I also noticed on the morning after the hydrogen was evacuated that the screw mechanism on the lid was loose. When I inspected the reactor after 3 days it was event looser and yet the vacuum was held. It is quite evident that the steel screw mechanism isn't as important for vacuum as the interface between the teflon inner chamber and the steel lid. This makes sense - the outside air pressure will be pushing on the interface making it higher.

2019-08-19

• It is very common to use the above chemicals for stainless steel, but not necessarily for Telfon (inner chamber of my reactor). From this table we can see that teflon will not be damaged by either substance even when hot.
• I did the following:
1. Clean with deionised water using a toothbrush to get in between small spaces - this is to remove any big bits of dirt
2. Clean with isopropanol - this is the first attempt to remove any grease
3. Soak in 10% citric acid solution hot for 10 min (Alan soaks overnight, but the advice on "passivation" of stainless steel suggests a lot less time - I will be conservative in this stage)
4. Rinse with water
5. Soak in acetone for 10 min - removes any remaining grease and also kind of "mops up" any remaining water.
6. Heat the components to make sure all water and other chemicals have been taken out
• After heating the reactor for several hours I vacuumed it down and left it overnight. I achieved a better vacuum that previously - even getting negative pressures from my transducer (this means I need to recalibrate the sensor)

2019-08-20

• Today I will try a hot hydrogen test at a reduced pressure of 1.5 atm.
• At around 6:30 I turned the heater on while the reactor heated up I set up the MQ-8 hydrogen sensor that I bought a few days ago

• Inside the stainless steel mesh there is a tin dioxide layer on an aluminium oxide tube. The electrical resistance of this layer is sensitive to the hydrogen concentration in the air around it. For more details on the inner workings including calibration tables etc, go here
• Using several sources of information, including Alan Smith, a guy on youtube (who shows the connections and shared his ardunio code) and Sandbox Electronics (This is referred to by Alan) I was able to figure out the connections and get the Arduino code working.
• I also figured out how to use the breadboard to allow me to connect both the pressure sensor and hydrogen sensor to the 5V supply.
• The wiring is a bit of a mess

• But the pins on the bottom allow me to hook it onto the thermocouple wire which is in a good position to test for leaks from the reactor

• At 10:50, the hydrogen sensor was ready and the reactor temperature was more or less at its peak. I decided to test the vacuum pump for 30 min (twice as long as previous longest). The vacuum was even better than just after I'd cleaned the reactor. The sensor on the vacuum pump was below 0mbar!

• It was initially tricky to get 1.5 atm of hydrogen in the reactor - I overshot. I decided to alter the hydrogen filling protocol at point 6, i.e.

1. Slowly open V a little bit and let the generator pressure gauge drop to somewhere between 0.5atm and 1atm the desired pressure. Hold for a 1 min. Don't open it full otherwise the generator will be exposed to full vacuum and it might break. This process allows the air present in the hydrogen tube to get flushed out.
2. Close V Open R
3. Switch off vacuum pump Wait until reactor reaches the right pressure
4. Open R Close R
5. Wait until pressure gauge reaches the desired value Close V
6. Close R Switch off pump
7. ...
• The above made it easier to control the reactor temperature

• The leak rate was much better than for the first hot hydrogen test. 169kPa to 163kPa over 5 hours, i.e. a loss of 1.2kPa per hour (0.7% per hour).
• I covered the top of the reactor with a tea cosy just before 18:00 in order to see whether it would increase the temperature significantly without me having to use the ceramic wool. It only gave 2 C increase. It was only loosely covered with plenty of space for heat to escape so perhaps that's why. More tests required.

• During this hydrogen fill stage, I noticed that the hydrogen sensor was detecting much more hydrogen that I considered reasonable. Initially I noticed that flowing air (e.g. from the fan) had the effect of raising the level of hydrogen, but even in still air there was a problem when I was filling the reactor using the above protocol.
• The normal level was 20ppm
• When I moved the sensor away from the reactor I saw an increase. I thought that maybe the hydrogen was blowing back though the window after initially being blowed out by the pump.
• When I detected 100ppm not too close to the window I realised the blowback hypothesis wasn't right.

• By carefully moving the sensor around I detected 800pm around the vacuum pump tube. Upon close inspection I found tiny holes.

• It is unclear at this point if the holes where there all along or whether the hydrogen has eaten its way through the tube. Either way I need to get some more tube.

• It is worth noting that although the sudden increase in hydrogen concentration in the air might be cause for alarm, 10% of the lower explosive limit of hydrogen in air is 4100 ppm - i.e. we are now where close to the danger zone.

• Considering that the hose I bought was for refrigeration it is probably one of these industrial hoses. Upon inspection of the different types I see that the outer layer is desired as pin-pricked which sounds exactly like what I am observing on my tube. Although these pin-pricks are designed to vent any build up of gases inside, the inside is not supposed to leak. It should be noted that

• The working temperature is lower than what I have sometimes put through the tubing
• When I installed the tubing I had to use significant force to get the tube over the swagelok barb.
• Both of the above could have caused damage which allowed the inner tube to connect to the outer one - hence allowing me to detect the hydrogen through the pin-pricks.

2019-08-21

• Today I prepared the nickel mesh as closely as I could to Mizuno's procedure. For a play-by-play account of the day, watch my YouTube video
• I bought some sandpaper from a local shop. It's silicon carbide just like Mizuno, but not quite the right assortment of grades
• I could't find any ethanol so I used isopropanol instead.

• I took the nickel mesh that Alan Smith sold me and cut 1/3 off to work with

• After the initial wash in soapy water, the mesh weighed 3.612g

• I decided to do the sanding out in the park to more effectively avoid any harmful nickel dust.

• After sanding the mesh weighed 3.5g
• After the second wash the weight was 3.49g
• After a soak in hot water the weight was 3.495g
• After a soak in alcohol the weight was 3.497g

• The 1g of palladium actually weighed 1.018g

• I then proceeded to try and burnish the mesh with palladium. I found this to be very difficult. At first I burnished lightly and the palladium weight didn't change but the mesh did. This was presumably because the mesh is fragile and bits were breaking off.
• I heard that annealing the metal (ie heating it) could make it easier to work with. I put the Pd into the flame on the hob until it glowed red. An oxide layer was formed which I sanded off. Burnishing after this was more successful, but only until the palladium was mad shiny by the nickel - further burnishing didn't do much.
• I decided that I would anneal again (but not to red hot) and then sand. This was not as effective as the first time. I have since learnt from Ed Storms that because I was burnishing on top of blue lab paper the procedure would have been difficult because you need a hard surface which to burnish on (he has suggested a steel sheet)
• In the end I had to burnish using significant force so that I could really feel the resistance to my burnishing motion. I then managed to loose a few mg of palladium.
• Overall I'm unsure of how much Pd I really put on the mesh - my scale varied from place to place. I should have done some calibration before I started using it.
• Below is the mass in g after each round of burnishing (front and back twice)
• So, in total the mesh lost 29mg and the palladium lost 12mg (through burnishing)
• The final mesh does have a shiner appearance, as noted by Mizuno, so there is hope that I have actually done something right.

2019-08-22

• Yesterday evening I placed the mesh inside the reactor and vacuumed it down to start the degassing process
• Today, I heated the mesh up in vacuum to continue the degassing.

• The reactor top was wrapped in a tea towel and reached a maximum temperature of 185.5 max at 12:15 at which point I put the tea cosy on
• The ambient temperature in the room was 22.1C

2019-08-23

• Today is my first cold fusion experiment - I'm finally here! Watch the start of the day on my YouTube video.
• I set up the Geiger counter next to the reactor to keep track of any unexpected gamma rays - there was nothing out of the ordinary. The counts per minute (cpm) hovered around 20 throughout the course of the experiment.

• I also set up a pan full of water in case I needed a quick way to cool the reactor down.
• The hydrogen went into the reactor at around 12:45. At that time, the temperature was peaking at 185C. The temperature rose once the hydrogen was put in and continues to very gradually rise. At 01:53 (time of writing) it was 192.25 C

• The ambient temperature was warmer today than the last few days, starting at around 24C when the hydrogen went into the reactor at around 12:45 and reaching a peak of 26C at 18:30 where it hovered around there for hours. At 23:00 it was 25.4 and at 02:00 it was down to 24.7C
• Although you might think that higher ambient temperature could be the cause of the elevated temperatures inside the reactor, the temperature continues to increase while the ambient air is decreasing. There is also quite a strange undulating pattern in the temperature that doesn't seem consistent with the normal temperature rise o equilibrium that I've seen before. I need to do some investigations into the accuracy of my thermocouple before reading too much more into this.

• On the advice of Ed Storms, I took some of the pressure out of the reactor to encourage hydrogen to come out of the palladium and encourage the palladium to go from beta phase to alpha phase. In Ed's view this will create cracks that will enable the lenr reaction.
• Ed sent me equations from an old article about palladium and hydrogen that helps you figure out what phase you are likely to be in given particular temperatures and pressures.
• I create a Jupyter notebook and found that my pressure was too low to form sufficient amount of beta phase. Turns out that 3atm at 190C is that is required to make all the Pd go into beta phase. So I put the reactor at 3atm for 30 min and then took the pressure down to 0.5 atm at which point everything should transition to alpha phase and create the most stress in the material. We'll see if this has an effect.