What Chemically makes Crackle/ Dragon Eggs Work?

[AustralianPyromaniac]

I was thinking today about the bismuth/ lead oxide crackle formulas. And what makes those metal oxides work and others not work. Copper Oxide + MgAl is known to crackle/ explode, if prepared correctly, it works quite well. It isn't quite as loud as the formula containing bismuth/ lead, and it is hard to light. I couldn't find any consensus online about what the actual reaction occurring in crackle is, clearly it is a thermite reaction between CuO and MgAl, but does the bismuth also undergo a thermite reaction as well? Some formulas contain no CuO and work, but these mixtures are expensive. I think there must be multiple reactions taking place. 

Lead and bismuth oxides and subcarbonates are used as "fluxes" in pottery and melt at low temps to create a small amount of liquid which can then go on to dissolve other higher melting point compounds, lowering the observed melting point of the mixture. Is it possible this liquifying behavior is also occurring in crackle? The bulk Bi/ Pb compounds melt yielding a slag, which the MgAl and CuO mix into. The melting of the oxides would keep the temperature low, even as the MgAl is "burning" and being oxidized. Once the melting is complete, the reaction runs away, the temp rapidly increases, and the mixture explodes. The intimate mixing of the compounds yields a more rapid reaction than with CuO and MgAl alone. 

This does not fully explain why the particle size of the MgAl affects the delay though. Assuming this is correct, I would theorize that the MgAl oxidation creates the heat that drives the reaction forward melting the Bi/ Pb, smaller particles burn faster and melt the mixture more rapidly. I am making the following predictions based on this theory which I'd like some input on. 

1) Magnesium metal alone should work for crackle (unsure about the validity of this one)

2) CuO + MgAl should work for crackle but should exhibit no delay between ignition and explosion 

3) Boric oxide (B2O3) and Zink Oxide (ZnO) should both work in place of Bi/Pb, but the zinc is strongly basic and hygroscopic so should not be used

4) In mixtures with CuO and oxide (bismuth/ lead), the more Bismuth/ Lead is added, the longer the delay is, as it all must melt. Mixtures with no CuO would exhibit the longest delay

Can anyone with experience with any of these speak to this? Or someone who knows how it does work, and has some better input on the true mechanism? 

[WSM]

My personal experience with crackling micro-stars (the dragon egg effect) go back over thirty years ago. I just reviewed Clive Jennings-White's article in Pyrotechnica XIV (August 1992) on "lead-free crackling microstars" and refreshed my thoughts of how this developed then.

One detail that came to light was that a larger proportion of aluminum in the magnalium (or simply added atomized aluminum to the magnalium) improved the effect. 

In observing the reaction during the burn-testing of individual microstars, they seemed to be similar to a strobe star in that they appeared to have a smolder phase followed by a flash phase (terms used by researchers explaining the strobe effect). In the best crackling microstars, this happened without repeating, i.e., one delay and then "Bang!".

Another observation was that the residue after the effect seemed to leave a metallic strike on the surface of the test plate. I believe the effect reduces the metal oxide (suddenly, loudly and exothermically) to the base metal. This observation lead me to believe the effect is a type of hybrid reaction between a strobe and thermite (thermitic or "Goldschmidt" reactions are sometimes used to "win" fairly pure metal samples from their respective oxides for laboratory samples).

I will leave to you to search more resources and consider for yourself if I'm right in my recollected observations or if I'm missing something. If the latter, please share with us here so we all can be edified.

WSM 😎

[AustralianPyromaniac]

I have looked into it a lot more, and I am not convinced at all about the mechanism given by Shimuzu. At the time, only lead oxide was known to work. I do agree the overall reaction is a thermite, with the copper/ lead/ bismuth oxide being reduced to the metallic form by the Mg/ Al. The reaction occurs rapidly, for whatever reason, causing the explosion. I've written below what I currently understand. 

Shimizu found aluminum alone is responsible for the detonation, magnesium powder alone does not work, but aluminum alone can work. The use of magnalium is important as the magnesium is more reactive, and helps keep the reaction going in the early stages, by reacting and liberating heat. 

He said the reaction proceeds by the mechanism which I have written below to inform future readers. 

1) The microstar is ignited by a flame

2) Mg and Al are oxidized to MgO and Al2O3 as the reaction begins to smolder and heat up. The oxygen for this reaction comes from some of the Pb3O4, which, acting as the oxidizer, is reduced to PbO. The oxygen liberated in this reaction then goes to join with the Mg and Al generating a large amount of heat. 

3) The magnesium has no further role than to generate heat. Some portion of magnesium remains and will continue to react keeping everything hot. Aluminum alone can be used but will be unreliable and require a hotter prime. 

4) The aluminum particles (ideally 60-100 mesh) now exist suspended in a liquid soup of lead oxide, which has melted from all the heat. Pb3O4 decomposes to PbO above 500c, so by this point decomposition has occurred. PbO then melts at 888c. 

5) Due to the reaction in step 2, the aluminum particles in the PbO soup are coated in a layer of solid Al2O3 (aluminum oxide) which doesn't melt till 2000c. This layer prevents further reaction with the liquid PbO. The reaction taking place at this point would be PbO -> Pb if it could occur.   

6) The temperature continues to rise as the remaining particles of Al and Mg continue to react in various ways. Many reactions are probably taking place at this point which continues to increase the temperature. Suddenly, once the temperature reaches 1500C, the PbO reaches its boiling point and vaporizes. The vapor, unlike the liquid, can diffuse quickly through the Al2O3 layer created in step 5. 

7) As the vaporized PbO hits the core of the aluminum/aluminum oxide particles, the PbO reacts with Al and is reduced to Pb extremely rapidly. The star explodes.  

---

To me, this explanation is weak. It was based on two observations. 

1) If the mixture is stopped in the smolder fase and analyzed by microscope, it is found that the Pb3O4 has decomposed to PbO (determined by color change). Indicating it has been reduced by reaction with Al. 

2) The particles of aluminum are found to be coated in a layer of oxide, supporting the hypothesis that this oxide layer is what prevents further reaction and creates the delay and rapid reaction 

Both of these observations could be explained by general burning reactions, and may not be related to the crackle phenomenon itself. 

---

We now know other metal oxides to work well, primarily bismuth, but also copper oxide. Clive Jennings-White established the function of bismuth crackle, which we now know to be excellent. Modern research has shown mixtures of CuO and Al to work well also. The boiling point of CuO is 2000C, which would not support the theory proposed by Shimuzu. We also know that other metal oxides work, Manganese and Iron, for example, exhibiting delay and then a flash, but no bang. The delay in these mixtures cannot be attributed to the time it takes to boil the metal oxides, as their boiling points exceed the flame temperature achieved even in the hottest pyrotechnic mixtures. 

I have spoken to Clive Jennings-White about this, he favors the self-confinement of the heavy atoms as being critical to creating the effect, which is why lighter metal oxides do not work as well. This would make sense.

I will continue to look into this and update. 

[rocket]

Will be interesting to see where you get with this as I’ve been wondering the same thing after playing around with dragon eggs over the last few weeks. Can't add much other than it looks to be some sort of thermite reaction to me.

Mesh size seemed to play a big roll weather they popped flashed or just smouldered.

Quality of chems helped as well a lot of my stuff is getting old and stale.

I wonder what the size limit would be with the right mesh size.

I made a small batch where I made 5mm eggs with a coarser MgAl and when it did work they went of like a fire cracker not the best to do in the shed😀

Edited January 3 by rocket

[AustralianPyromaniac]

Below is my current working theory. I have a few ideas of how to test it, and what I have written here is not complete, but it does fit the data well, better than the metal oxide boiling theory. It also offers some avenues for new formulas, which must be tested. 

--- 

BURNING/ SMOLDER (BELOW AL MELTING POINT)

The reaction proceeds in phases, as has been established by Shimizu. The first stage is smoulder, where the metal oxide is reduced in the flame to the form ME₁O₁ ME₂O₁ and ME (ME meaning any metal), where the metal is reduced from higher oxidation states to the low +1 and +2, and in some cases even further to the metallic form. The oxygen removed in this stage forms particles of aluminium coated in aluminium oxide. Research has shown that this coating does develop and is it reasonable to infer this slows the reaction rate. The coating of oxide prevents further rapid reaction between the metal oxide and the aluminium. As the aluminium oxide is formed, it sinters together forming relatively “large” (50 um), embedded in which is the aluminium. During this stage, oxygen free radicals are also created. These radicals roam through the aluminium oxide matrix, diffusing into the alumina, and continuing to react with the aluminium raising the temperature of the star.

It is hard to explain why magnesium alone does not work for this reaction. It is likely that is too reactive. This is to say, even with a passivating coat of MgO, the reaction cannot be arrested, and continues to oxidiser very rapidly without a smoulder phase, i.e. flash powder. Most other properties of magnesium are the same as aluminium. The melting point of magnesium is very similar to that of aluminium. Its boiling point is lower than that of aluminium, approx. 1000c, but this temperature is not theorised to be reached in this stage of the reaction.   

 

EXPLOSIVE TRANSITION

Once the melting point of the aluminium is reached (660c), the star stops heating further and must pass the first exotherm (all the generated heat goes to melting the aluminium). As the aluminium core particles melt, they remain surrounded by a layer of solid aluminium oxide, which has an extremely high melting point. During this stage the start is extremely sensitive to shock, and this has been established by Shimizu’s research. All that is stopping a violent runaway reaction at this point is the thin shell of alumina. If it cracks, the liquid aluminium will react extremely quickly with the surrounding oxide an explode.

Once all the aluminium has melted, the star is again free to increase in temperature, and the temperature continues to rise as more oxygen diffuses into the alumina shell and reacts.

At some point, a localised breakdown of one or more alumina shells occurs. This is to say, aluminium seeps out into the bulk oxide and reacts near instantly to generate Cu metal and Al2O3. The liquid phase of the aluminium causes this reaction to occur extremely rapidly. A shock wave, and a large amount of heat is generated by this reaction. If the mixture is made well, the spread of this shock wave breaks the remaining alumina shells and the star detonates. This is a true detonation, above the speed of sound, and propagated by shock not heat diffusion.

A 2019 paper studying CuO-Al pyrotechnic crackle found an unreliable explosive transition temperature, ranging from around 700-1000 C. That is to say, the crack occurred when the smouldering phase had brough the star somewhere between 700 and 1000 C. This transition always occurred after the melting of aluminium, but before the boiling of the metal oxide. Shimizu, in his article in Pyrotechnica attempted to measure the transition temperature of the explosion of lead oxide crackle, and was only able to collect 1 data point, 850C, which is in line with this theory, and does not support the metal oxide boiling theory he ended up accepting.

EXPLOSION

Once the alumina shells break and liquid aluminium contacts the bulk oxide, the reduction reaction occurs near instantly and everything explodes. This liberates enormous amounts of heat. Research has shown CuO-Al thermite reactions to peak at over 3400 C. Thermite reactions are generally understood to create no gassous products. In the case of explosive thermites, we know this is not the case. CuO-Al thermite theatrically has no gaseous end products, but burning this mixture shows a definite “plume” of smoke characteristic of rapid gas generation. Confined CuO-Al is capable of creating a pressure wave which can burst a steel container. This supports the generation of products in vapor form. Aluminium oxide has a boiling point of 3000c and is unlikely to be vaporised. Bismuth and lead metal have very low boiling points and low heat of vaporisation, so are likely contribute to the explosive effect of these thermites by boiling. Copper has a high boiling point, and moderate heat of vaporisation, but is also an extremely energetic thermite. Gram for gram, copper thermite releases 3 times more energy than lead thermite, and 2 times more energy than bismuth thermite. This energy contributes to the boiling of the bismuth and lead metal, which is what causes the explosion. Copper thermite alone, while very energetic, does not have the energy to boil its metallic products to as great a degree.

[CHNO]

Crackers.pdf Crckers2.pdf

[AustralianPyromaniac]

50 minutes ago, CHNO said:

Crackers.pdf 499.38 kB · 3 downloads Crckers2.pdf 1.37 MB · 2 downloads

Thanks, I have read both of those papers. They are by the same guy and are both very good. They analyze data but do not provide a mechanism, and they investigate a somewhat more simple set of formulas. 

[Boophoenix]

 Richard has done a lot of work trying to satisfy his curiosities. The reason for the simpler formulas I would assume to make it easier to find the mechanism involved. More chemicals more complex reactions. Reach out to him if he has time I bet he’d enjoy a good conversation on the subject. Some of his testing is public I believe in the UK forums at http://pyro-gear.co.uk/forum/index.php

[CHNO]

4 Bi2O3 + 3 CuO + 6 Al + 6 Mg

77.33% / 9.90% / 6.72% / 6.05%

 

My Theory:

 

CuO + Mg = Cu + MgO

2 Bi2O3 + 2 CuO = 2 Bi2CuO4

2 Bi2O3 + 2 Al + 3 Mg = 4 Bi + Al2O3 + 3 MgO

2 Bi2CuO4 + 4 Al + 2 Mg = 4 Bi + 2 Cu + 2 Al2O3 + 2 MgO

[AustralianPyromaniac]

20 hours ago, Boophoenix said:

 Richard has done a lot of work trying to satisfy his curiosities. The reason for the simpler formulas I would assume to make it easier to find the mechanism involved. More chemicals more complex reactions. Reach out to him if he has time I bet he’d enjoy a good conversation on the subject. Some of his testing is public I believe in the UK forums at http://pyro-gear.co.uk/forum/index.php

Do you have his email? I'd rather not make an account. Awesome to know he's around in amateur pyro and contactable. 

2 hours ago, CHNO said:

4 Bi2O3 + 3 CuO + 6 Al + 6 Mg

77.33% / 9.90% / 6.72% / 6.05%

 

My Theory:

 

CuO + Mg = Cu + MgO

2 Bi2O3 + 2 CuO = 2 Bi2CuO4

2 Bi2O3 + 2 Al + 3 Mg = 4 Bi + Al2O3 + 3 MgO

2 Bi2CuO4 + 4 Al + 2 Mg = 4 Bi + 2 Cu + 2 Al2O3 + 2 MgO

Not sure I am following . What makes you think of this " 2 Bi2O3 + 2 CuO = 2 Bi2CuO4 " ? I'd love some more thoughts on if you're able to explain your thinking! Very interesting pathway. 

The reaction can occur without Mg, so the first step is known not to be critical. Aluminum could work in its place.  Also, these mixtures work without copper, if the compound Bi2CuO4 is proposed as critical to the mechanism, there must be some other step taking place when copper isn't used. 

[Mumbles]

I've had the best luck trying to wrap my head around this concept as thinking about the dragon egg reaction as sort of a hybrid between a strobe effect and a glitter spritzle.  I think the strobe effect is fairly well understood conceptually, or at least ammonium perchlorate strobes.  Nitrate based strobes are probably function a lot more similarly to a glitter.  AP and metal smolder until a sufficient heat can be generated to initiate the sulfate and metal flash phase, which blows off the smolder shell.  Repeat until the star is consumed.  Glitter shares some characteristics with this in that there is often a sulfide melt phase on the surface of the particle which is believed to oxidize as it's falling through the air.  When a sufficient amount of heat or sulfate, or something builds up, the particle initiates the flash phase.  I referenced "Glitter: Chemistry and Techniques" by Lloyd Scott Oglesby in another thread.  It may be an interesting read to get a different perspective to view this reaction from.

I feel like a few things are understood about the properly functioning mixtures.  For one, it seems that metal oxides with a stable or meta-stable partially reduced oxidation state is favorable.  This tracks with lead, bismuth (meta stable), and copper most commonly seen.  There are of course a wide variety of materials that fit this bill, particularly within the transition metals and heavier metaloids.  It's also known that MgAl tends to work better than either individual metal, but a higher aluminum than magnesium content seems to be beneficial.  The mesh size of the metal used can influence the delay.  Larger metal particles, generally longer delay.  Larger particles react slower and are less intimately mixed, which may be a simple explanation.  There is also some knowledge about certain binders or materials leaching into the DE granule that tends to impede this effect.  I've often heard this basically being attributed to providing a sustained reaction instead of allowing for the two discrete smolder/flash phases to operate in concert.

I've preferred to think of this as a mix of the two between glitter and strobe.  The MgAl or one of the component metals smolders and reacts with the fully oxidized oxide reducing it to something comprising the reduced metal oxide.  Whether it is a shell, or a slag, or a melt, or some sort of complex spinel or molecular composition is somewhat unclear.  At a certain point the remaining MgAl, or the other component metal, reacts with this partially reduced smolder phase to initiate the flash/crack phase.  

[WayneJ]

On 1/9/2024 at 6:18 PM, Mumbles said:

I've had the best luck trying to wrap my head around this concept as thinking about the dragon egg reaction as sort of a hybrid between a strobe effect and a glitter spritzle.  I think the strobe effect is fairly well understood conceptually, or at least ammonium perchlorate strobes.  Nitrate based strobes are probably function a lot more similarly to a glitter.  AP and metal smolder until a sufficient heat can be generated to initiate the sulfate and metal flash phase, which blows off the smolder shell.  Repeat until the star is consumed.  Glitter shares some characteristics with this in that there is often a sulfide melt phase on the surface of the particle which is believed to oxidize as it's falling through the air.  When a sufficient amount of heat or sulfate, or something builds up, the particle initiates the flash phase.  I referenced "Glitter: Chemistry and Techniques" by Lloyd Scott Oglesby in another thread.  It may be an interesting read to get a different perspective to view this reaction from.

I feel like a few things are understood about the properly functioning mixtures.  For one, it seems that metal oxides with a stable or meta-stable partially reduced oxidation state is favorable.  This tracks with lead, bismuth (meta stable), and copper most commonly seen.  There are of course a wide variety of materials that fit this bill, particularly within the transition metals and heavier metaloids.  It's also known that MgAl tends to work better than either individual metal, but a higher aluminum than magnesium content seems to be beneficial.  The mesh size of the metal used can influence the delay.  Larger metal particles, generally longer delay.  Larger particles react slower and are less intimately mixed, which may be a simple explanation.  There is also some knowledge about certain binders or materials leaching into the DE granule that tends to impede this effect.  I've often heard this basically being attributed to providing a sustained reaction instead of allowing for the two discrete smolder/flash phases to operate in concert.

I've preferred to think of this as a mix of the two between glitter and strobe.  The MgAl or one of the component metals smolders and reacts with the fully oxidized oxide reducing it to something comprising the reduced metal oxide.  Whether it is a shell, or a slag, or a melt, or some sort of complex spinel or molecular composition is somewhat unclear.  At a certain point the remaining MgAl, or the other component metal, reacts with this partially reduced smolder phase to initiate the flash/crack phase.  

This is an interesting thread! I think the effect is more like thermite/strobe, especially if there's more than one crack. Currently, I keep pieces small (around 12 mesh), trying to avoid scattering unreacted product.  Years ago I made crackle with variable success, but lately when I needed quite a lot for comets, and didn't want to use lead or a lot of Bi trioxide, I found and modified a very useable mix of CuO, MgAl and only 5-6% Bi2O3 (which I used for two consecutive PGI firsts in competition}. So much of what is being discussed is theory, which is fine, and everyone can argue the "how", but when we're making pyro, sometimes you just have to experiment, get it in the air and let the audience decide. Not to say I don't appreciate or understand the chemistry, I certainly do, but it's artistry in the end.  Richard Harrison's articles referenced in above posts are just great! You mention nitrate strobes being maybe like glitter.  Having played with these things to varying success, I think it's slow vibrational burning.  My 2 cents.

[CHNO]

1. Smoulder Reaction: Bi2O3 + 3 Mg = 2 Bi + 3 MgO

2. Flash Reaction: Bi2O3 + 2 Al = 2 Bi + Al2O3

 

Strobe Chemistry.pdf

[CHNO]

0.5 Pb3O4 + 2 Mg = 2 MgO + 1.5 Pb

3 CuO + 2 Al = Al2O3 + 3 Cu

Crackers3.pdf