[Papyone]

Hello everyone,
First of all I want to inform you that I’ve not much knowledge with chemistry because my job was commercial diving. I’m now retired, but still very concerned about the security of my former colleagues and therefore I come to you with what some of you will maybe find a weird question, but which has some great and even vital importance to us.
One of the current things a commercial diver does is cutting steel underwater. Therefore we use a short thermal lance that works about the same way as a welding rod except that the lance is hollow to allow the passage of pure oxygen. Due to the DC current we use, as well as the heat of the flame (+/- 5500°c) we have electrolysis phenomena of the water that produce an explosive mixture merely composed of H2 and O2. For commercial divers, the formation of such gas mixture sometimes presents a lethal risk if it cannot escape freely to the surface and a lot of accidents did happen with deadly consequences.
Up to now, we were convinced that during the cutting process the percentage of hydrogen that was produced was situated above the H2 L.E.L which if I’m correct is situated around 3.9%
But, recently following another deadly accident it was decided to make a few cutting tests (6) to recover a few samples of residual cutting gasses for analysis and after seeing the results, I must say that I’m very astonished to see that the percentage of hydrogen present is much lower than this L.E.L of 3.9.
As for instance, the average values that the laboratory has measured are:
H2 0.0167% / O2 93.5% / N2 5.49% / CO 0.113% / CO2 0.855%. 
What is very confusing is that even with such a low hydrogen proportion we are confronted to explosions even if the trapped volume of gas is small (a few cubic centimetres).
So my question is: Why do explosions happen with such a low hydrogen %?
Does the combination of all these gases change the LEL of the mixture?
Sorry for that long description.
I also ad here a picture to help you to understand my question better.
Thanks in advance for your help.

[Papyone]

Hello everyone,
First of all I want to inform you that I’ve not much knowledge with chemistry because my job was commercial diving. I’m now retired, but still very concerned about the security of my former colleagues and therefore I come to you with what some of you will maybe find a weird question, but which has some great and even vital importance to us.
One of the current things a commercial diver does is cutting steel underwater. Therefore we use a short thermal lance that works about the same way as a welding rod except that the lance is hollow to allow the passage of pure oxygen. Due to the DC current we use, as well as the heat of the flame (+/- 5500°c) we have electrolysis phenomena of the water that produce an explosive mixture merely composed of H2 and O2. For commercial divers, the formation of such gas mixture sometimes presents a lethal risk if it cannot escape freely to the surface and a lot of accidents did happen with deadly consequences.
Up to now, we were convinced that during the cutting process the percentage of hydrogen that was produced was situated above the H2 L.E.L which if I’m correct is situated around 3.9%
But, recently following another deadly accident it was decided to make a few cutting tests (6) to recover a few samples of residual cutting gasses for analysis and after seeing the results, I must say that I’m very astonished to see that the percentage of hydrogen present is much lower than this L.E.L of 3.9.
As for instance, the average values that the laboratory has measured are:
H2 0.0167% / O2 93.5% / N2 5.49% / CO 0.113% / CO2 0.855%. 
What is very confusing is that even with such a low hydrogen proportion we are confronted to explosions even if the trapped volume of gas is small (a few cubic centimetres).
So my question is: Why do explosions happen with such a low hydrogen %?
Does the combination of all these gases change the LEL of the mixture?
Sorry for that long description.
I also ad here a picture to help you to understand my question better.
Thanks in advance for your help.

[Enthalpy]

Welcome, Papyone!

Of the measured gases, only H₂ and CO can burn, and both concentrations are very far below the explosion limit. No interaction known to me, nor do I expect any.

Explosions must result from other conditions than the measured composition. The dangerous conditions may not happen every time.

I don't see how an electric arc would produce significant hydrogen by electrolysis. The evolved hydrogen must burn immediately under these conditions, before any amount accumulates. That's even more the case if (?) you use AC current. But surprises happen.

A different scenario? From the item being cut or welded, or from the tools, metal still hot reacts with water when or where no oxygen is present, and produces metal oxide and hydrogen. If hydrogen accumulates in something like a sunken hull, it can detonate.

Still an other hypothesis: methane produced by bacterial decomposition of organic material already accumulated in a place where the diver adds oxygen. The mix can deflagrate.

In case (?) these explosions occur where hydrogen or oxygen accumulates, away from the tools, a solution would be to cap the operation site with a funnel and a pipe to evacuate the gas to the surface.

[Papyone]

Already thank you for your reply. I will comment more on it in two days.

[Papyone]

Hello enthalpy,
Here I’m again with a few more information concerning our cutting process.
To do what you see on the photo, we use a rod that needs about 150 Amps to create the electric arc which ignite the rod and once ignited such a rod burns during about 50 seconds and consumes about 250 litres (at atmospheric pressure) of pure oxygen per minute.
To limit the electrocution risks we always use DC current.
Concerning the formation of hydrogen by electrolysis we can see that it happens in salt water, because when we start a new job we have to make a polarity test to verify if the cutting torch is correctly connected to the negative pole of the welding group.
Therefore we need to use a bucked of salt water into which we approach the extremity of the electrode at close proximity of the earth clamp. Once the current is on, we can see that a lot of bubbles travel in the water from the clamp towards the electrode (test illustration on the photo).
In most commercial diving schools and manuals they tell the divers that this phenomenon (electrolysis) continuous during the cutting process, but apparently from what I have found during my research on the net this is not really the case.
If I’m correct, thanks to a Faraday law it is possible to calculate the theoretical value of the electrolysis flow generated during the consumption of the rod.
For instance by using one of his formulas I’ve learned that whatever the voltage, 1 ampere will produce 0.209 l of oxygen and 0.418 l of hydrogen, or a total of 0.627 l of gas per hour.
Knowing that when we are cutting, the thermal rod will burn for about 50 seconds with an intensity of 150 amperes, I’ve calculated that this amount of gas mixture would be produced.
(0.627 / 3600) x 150 x 50 = 1.3 l of O2 / H2 mixture
This amount of gas will actually be produced as long as the tip of the electrode does not touch the workpiece but and at contrary to what is often taught to the divers as soon as we start to cut, a large part of the current will pass directly through the electric arc and in this case only the leakage currents located at the periphery of the arc (and therefore still in contact with the water) will be able to create the electrolysis of the water.
I guess therefore that the production of our explosive mixture is not due electrolysis.
Going through the web, I’ve also learned that if water is brought into contact with a very high heat source greater than 2200 ° C, it will start to split into their atomic components hydrogen and oxygen.
At the same temperature of 2200 ° C about 3% of the water surrounding the heat source is decomposed into hydrogen and oxygen, and that this percentage of dissolution increases sharply when the temperature increase. As the combustion temperature at the extremity of our rod is close to 5534 ° C (10000° F) I ‘ve tried to find out  how much water is then decomposed, but unfortunately up to now nobody has given me an answer. 
Do you by chance know how to calculate this amount?
Anyway, I suppose that most of the hydrogen produced in our cutting comes from the heat and not from the electricity.
Concerning the production of methane by bacterial decomposition of organic material you are quite correct. In the latest accident that happened in Japan this was the cause (methane accumulation in a sealed sheet pile tube).
The use of your funnel would be difficult to install in most situation, but anyway thanks for the suggestion. Just to let you know that one of the first things a diver learns when he starts to cut is to make sure the residual gas can escape freely to the surface.
If not, they are taught to make vent holes above their cut or if not possible the must then vent the cut with air.
But unfortunately this is not always done correctly and accidents continuous to happen (35 in the latest forty years).

[Enthalpy]

Hi,

150A is the current when the user puts the rod in contact with the other electrode for ignition. It would be extremely difficult to pass this current through seawater, needing huge flat electrodes very near to an other and using deadly voltages. In industrial electrolysis, the current per cell is but bigger, with optimized shapes and solutions as concentrated as possible, not seawater. So as soon as the user separates the rod from the counter-electrode (does he?), the current must drop to a low value, maybe 0.1A or 0.01A.

In electrical engineering, I learned that electrocution is more probable with DC. I can't confirm it first-hand (so to say) as I never had an accident with 240V DC and have zero desire to try. Videos I saw of thermal lances on land used a 12V car battery to ignite the rod, and this may be a reason for DC. Spectacular.

I too would say that any current flowing in an underwater arc is inefficient at electrolysis. Most current flows through the vapour without producing hydrogen nor oxygen.

I didn't quite understand: I imagine that you need current to ignite the thermal lance, but thereafter none is needed. Or is it a different use, maybe when welding with current instead of cutting with oxygen and heat?

If the rod and the counter-electrode are separated by 0.1m or more after ignition, and the voltage is a few volts, any current flowing through the water would be small, like 10mA, and produce very little hydrogen.

Yes, heat decomposes water partly. This could be computed by hand or using software like Propep. At 5500°C it's efficient. However, hot oxygen and hydrogen recombine immediately to form water as soon as the temperature drops. As bubbles go away from the lance, they just make water back. I don't imagine a production of hydrogen by this process. It's the very problem that prevents production of hydrogen from concentrated sunlight.

To my eyes, the credible process is that the hot metal (Fe, Al...) reacts with water to produce hydrogen and metal oxides and hydroxides. This can happen at the outer surface of the lance, near the hot tip, where no oxygen can burn immediately the hydrogen. One potential parry would coat the lance's outer face with thick oxides and hydroxides stable to heat, so water can't reach quickly the hot bare metal.

Experiments aren't too difficult: use a lance horizontally in a tub, observe if gas appears at the outer face near the tip, then collect and analyse it.

The cut parts (hull...) too can produce hydrogen once or where the lance and its oxygen aren't any more but the metal is still hot. And possibly a better candidate: sparks of liquid metal ejected to the water where no oxygen is present. Or biological material attached to the hull: it gets hot at some distance from the oxygen, and its pyrolysis can produce a mix of flammable gases.

[wildfyr]

I too think its the very finely powdered (or even gaseous?) metal in the 0 oxidation state that is the most suspicious flammable thing around. Its a matter of surface area. A chunk of Fe and Zn 1 cm³ is not very flammable, but the same amount that has been reduced essentially to nanoparticles is an entirely different story.

[Papyone]

 Videos I saw of thermal lances on land used a 12V car battery to ignite the rod, and this may be a reason for DC. Spectacular.

Yes, heat decomposes water partly. This could be computed by hand or using software like Propep. At 5500°C it's efficient.
Experiments aren't too difficult: use a lance horizontally in a tub, observe if gas appears at the outer face near the tip, then collect and analyse it.

Hi again,
If it is true that a thermic rod can be lit with a 12 volt battery we seldom use that method under water because most of the time when we have a cutting job we cut during many hours a day and therefore we use a dc generator or an inverter. Also as you have probably seen on your video, above water it is possible to cut the current once the rod in burning. Under water this works also but most of the divers prefer to cut hot (current on) because the quality of those rods vary a lot and some of them extinguish easily once the current is of.
Thanks again for your explanations that confirm that the majority of the H2 is effectively produced by heat and not by the rod current. I will try to convince the few big commercial diving organization to make a few more gas analysis in order to find out what really kills so many divers.
Concerning the water decomposition by heat, could you give me an example on how to compute it by hand?

[Enthalpy]

Computing an equilibrium by hand takes some training and time. I've run Propep on water at 1 atm (but 1+1 atm at 10m depth would change little) and varied temperatures. Here are molar fractions:

  H    H2    H2O    O    OH    O2
0.14  0.18  0.42  0.06  0.14  0.59 at 3000°C
0.58  0.06  0.01  0.29  0.05  0.14 at 4000°C
0.66  0.01   ppm  0.33  0.01  0.1% at 5000°C
0.66  0.1%   ppm  0.33  0.1%   ppm at 6000°C

- It's complicated, and here I didn't even include Al nor Fe. Hand computation is half-way reasonable with 2 or 3 possible compounds at most.
- The amount of H₂O that really reaches 5000°C is essentially decomposed, more to atomic H and O than to H₂ and O₂.
- But when the bubble cools down, everything recomposes snappily and fully to H₂O. I expect no H₂ production by this process.

But where should water be heated to that temperature? If it's by contact with hot metal, then the reaction is to oxidise the metal, not to decompose water by heat.

----------

Where does the permanent current pass? Is it within the lance, through oxygen, between the outer and some inner electrode? I still can't imagine 150A through seawater.

[Papyone]

Enthalpy,
I‘ve another question.
Making a gas analysis is quite a difficult and probably expensive operation and therefore I suppose some diving organisation may be reluctant to do it.
But I was thinking to ask some colleagues to make a flame test above the residual gasses that are reaching the surface (for instance by holding a gas burning torch above the bubbles that are coming up).
What would then happen if the residual gas contains less than the 3.9% H2 L.E.L?
What would happen if the residual gas contains more than the 3.9% H2 L.E.L?

[Enthalpy]

I would not make the test over a diver. Instead, collect the gas, for instance by filling with water a container, the letting the gas bubble into its wide opening. Move the closed container to an other location, and try to ignite. Prefer a plastic container, not glass nor metal - something like a Tupperware. Prefer a petrol engine spark plug and coil to ignite the gas from distance, or use a trail of flammable liquid.

The boom test tells you more directly whether the gas can make boom, rather than sub-details about a composition from which you may attempt deductions. And because it's faster and cheaper, you can test more varied situations, which is probably the key to finding the cause. My bet is that explosive mixtures form only under rare conditions, that's why the cause is still unknown.

Maybe methane is already present in the seabed from rotting vegetables, and the diver releases it by walking?

[Papyone]

Ok I will ask some colleagues to do something like that. Thanks for the tip and explanations.

[Enthalpy]

I've run propep on hot Fe₂O₃ (and also on hot Al₂O₃), and

• 3000K or 3500K look more plausible than 5500°C
• At such temperatures, the combustion produces much Fe (g), FeO (g), O (g), O₂ and other half-burnt species, but nearly no Fe₂O₃.
• So depending on how quickly iron evaporates or erodes from the rods, the flue gas may be fuel-rich sometimes. When little oxygen flows? Or can a permanent arc ignite the iron at the lance's beginning?
• This holds also for iron in the part being cut! If it melts and half-burns but has too little oxygen to achieve Fe₂O₃. I already suggested droplets of liquid Fe, but products of Fe and FeO in the cooled gas are unavoidable if the oxygen doesn't suffice.
• Flue gas rich in Fe or FeO, with too little oxygen, would inevitably react with water to make Fe₂O₃ and its hydrated variants, releasing much hydrogen.
• Such a hydrogen proportion would make boom if air is available somewhere or if the thermal lance releases oxygen-rich flue gas earlier or later, for instance when not cutting a steel plate.

I suggest to experiment that on the ground. No depth needed, no suit - but a tub full of water, steel plates, thermal lances and some sort of explosion-proof wall. Operate behind the wall with a long lance, try some combinations of plate thickness and oxygen throughput, observe in a mirror, collect the gas. If some flue gas can detonate with air or is fuel-rich, it's the explanation.

[Papyone]

Hello Enthalpy
Thanks to all your replies I’ve finally made a series of tests to calculate the production of hydrogen by electrolysis and thermolysis and have published the result in a document.
If you are interested by it send me a message with your email or another address where I can send you the link to download it (for free).
Regards
Francis

[billnotgatez]

@Papyone

Is there a reason you do not post the link here?