How To Make Starlite - The Miracle Insulating Material of Maurice Ward

Mick West

Staff member


Metabunk 2018-12-04 10-54-12.jpg

Starlite is one of those things that keeps coming up, and the story is repeated almost without question. An inventor, Maurice Ward, developed what he called a "plastic" that was flame resistant. He usually demonstrated this with a blowtorch, sometimes on an egg.

The material appears to be a white paste that sets. When a flame is applied it chars and expands a bit.
External Quote:
Under tests, Starlite was claimed to be able to withstand attack by a laser beam that could produce a temperature of 10,000 degrees Celsius. Live demonstrations on Tomorrow's World and BBC Radio 4 showed that an egg coated in Starlight could remain raw, and cold enough to be picked up with a bare hand, even after five minutes in the flame of a blowtorch. It would also prevent a blowtorch from damaging a human hand.[2] When heat is applied, the material chars, which creates an expanding low density foam of carbon which is very thermally resistant.[3]
Starlite's composition is a closely guarded secret, but it is said to contain a variety of organic polymers and co-polymers with both organic and inorganic additives, including borates and small quantities of ceramics and other special barrier ingredients — up to 21 in all. Perhaps uniquely for a material claimed to be thermal and blast-proof, it is claimed to be not wholly inorganic but up to 90 percent organic.[7]
Here's a video describing the story:


The results of the egg test are at 1:12

People have tried to replicate it, one Keith Lewis said he made some with "things you find in your own home"



He strongly suggests it involved PVA glue. Luckily I had a bottle full of PVA glue from an old book-binding experiment.

I was thinking about the simplest version you could make, and I decided that PVA glue plus baking soda would give a nice white paste. So I mixed up some of that. While doing that I found it's the ingredients of "slime", a DIY kids fun substance.

PVA Glue is a mixture of PVA and a solvent. When the glue dries it just leaves PVA.
External Quote:
Poly(vinyl acetate) (PVA, PVAc, poly(ethenyl ethanoate): best known as wood glue, white glue, carpenter's glue, school glue, Elmer's glue in the US, or PVA glue) is an aliphatic rubbery synthetic polymer with the formula (C4H6O2)n. It belongs to the polyvinyl esters family, with the general formula -[RCOOCHCH2]-. It is a type of thermoplastic.[1] There is considerable confusion between the glue as purchased, an aqueous emulsion of mostly vinyl acetate monomer, and the subsequent dried and polymerized PVAc that is the true thermoplastic polymer.
Mixing it with baking soda (sodium bicarbonate, NaHCO3 ) creates a spreadable paste. Both substances are white. Baking soda by itself is quite heat resistant.

I also tried it with aluminum oxide. This is grey, and more inert than baking soda.

Here's my current batches. Still wet from the glue, so I will wait until they dry. You can see I gave them a go, but it seems like the solvent in the glue is buring off in the flame, so I need to wait until that evaporates.
Metabunk 2018-12-04 11-25-17.jpg

Results to come....
Last edited:
Some other similar coatings:



So the question becomes what does Starlite claim to be beyond an egg fireproofer. Well, Ward made some pretty extreme claims in interviews later in life.
External Quote:
The defence establishment was watching. In July that year, Ward was invited to the British Atomic Weapons Establishment at Foulness, and the egg went nuclear. 'They'd been trying to get something to withstand a nuclear flash for 45 years, and we did it in five minutes.' Ward was reluctant to take part at first. 'I was happy with my egg. It was just a challenge and I didn't want to lose.' This was a different league. Starlite-coated eggs were subjected to light-energy sources that simulated a nuclear flash, equivalent to a temperature of 10,000 C. 'They did it twice and it was still there. Charred, but intact.' The Foulness equipment couldn't keep up. 'I said to one scientist, "Are we doing all right?", and he burst out laughing. He said, "Normally, we do a test every couple of hours because we have to wait for it to cool down. We're doing it every 10 minutes, and it's sat there laughing at us."' Most materials vaporise beyond 2,000 C. Pure carbon, which has the highest melting point of all elements, melts at 3,500 C. Starlite was withstanding temperatures and forces that physics and thermodynamics dictated it shouldn't. Even with tests from unquestionable authorities like AWE, people were sceptical. 'Some people called me a shyster. But they are blinkered. We've got video: We can show you.'
External Quote:
In fact, Ward let a sample out of his sight only once. In June 1991, a sample was sent to White Sands atomic weapons testing site in New Mexico, in the care of the SAS, and subjected to a simulated nuclear onslaught. 'It was classed as the biggest bang in town. I've seen a video [on which] it shredded forest to sawdust, rolled some tanks around, stripped an aircraft into pieces.' But Starlite survived. Further tests at Foulness had subjected it to the force of 75 Hiroshimas, and it survived that, too. NASA publicly raved about its potential, with spokesman Rudi Narangor revealing that 'We have done a lot of evaluation and … we know all the tremendous possibilities that this material has.'
External Quote:
Ward certainly believes in his product, claiming publicly that it could have prevented the space shuttle disasters. 'Starlite has a Q-value [an energy absorption rating] of 2,470. The space shuttle tiles have a Q-value of 1.' Not only that, but because Starlite is so lightweight – 1mm thick, compared to 75mm for the space tiles – it's actually '2,470 x 75 times better'.
Extraordinary claims presented without evidence can similarly be dismissed. Where is this public raving of NASA? Who is Rudi Narangor? If you search for NASA + Narangor all you get is Starlite stories. Narangor does not seem to exist.
Besides ridiculous things like "75 Hiroshimas," there's a rather more banal claim in there:

External Quote:
There were tests carried out at ICI by a contact in one of the labs, in which the still unnamed material passed the UL94 (VO) test – involving a calibrated Bunsen burner flame – with ease. Ward thought then that 'if it were in ICI labs right now it'd be worth 10 million quid.' But talks fell through. 'I know now it's because they were working on Victrex,' says Ward, inviting me to look it up. (It's a 'high-performance thermoplastic', but not revolutionary.)
UL94-V0 is a test for flammability of plastics, the VO requirement is laughably simple:
External Quote:

  • Length 125 mm (5 in) x Width 13 mm (0.5 in) x Thickness [typically 0.7 mm (1/32 in) or 1.5 mm (1/16 in) or 3.0 mm (1/8 in)].
  1. A total of 10 specimens (2 sets) are tested per thickness.
  2. Five specimens of each thickness are tested after conditioning for 48 hours at 23 degrees C and 50% RH.
  3. Five specimens of each thickness are tested after conditioning for 7 days at 70 degrees C.
  4. Each specimen is mounted with long axis vertical
  5. Each specimen is supported such that its lower end is 10 mm above Bunsen burner tube.
  6. A blue 20 mm high flame is applied to the center of the lower edge of the specimen for 10 seconds and removed. If burning ceases within 30 seconds, the flame is reapplied for an additional 10 seconds. If the specimen drips, particles are allowed to fall onto a layer of dry absorbent surgical cotton placed 300 mm below the specimen.

Requirements for V-0
  1. The specimens may not burn with flaming combustion for more than 10 seconds after either application of the test flame.
  2. The total flaming combustion time may not exceed 50 seconds for the 10 flame applications for each set of 5 specimens.
  3. The specimens may not burn with flaming or glowing combustion up to the holding clamp.
  4. The specimens may not drip flaming particles that ignite the dry absorbent surgical cotton located 300 mm below the test specimen.
  5. The specimens may not have glowing combustion that persists for more than 30 seconds after the second removal of the test flame.
Video of the Boeing test exists:

Last edited:
ho is Rudi Narangor? If you search for NASA + Narangor all you get is Starlite stories. Narangor does not seem to exist.

Seems like the Telegraph reporter was using a phonetic spelling, the NASA person was Rosendo "Rudy" Naranjo, who worked at NASA until 2001 and passed away in 2011.

A supposed NASA memo mentioned him, and details some of the tests:
Metabunk 2018-12-07 06-00-16.jpg

External Quote:

Laser testing was done by the Boeing High Energy Laser Laboratory in St. Louis, Missouri using an Nd:YAG laser and CO2 laser. duplicating earlier testing done in
the UK. but including tactical threads. Experts defined the parameters for the tests. The military laser test plans specified power densities ranging from under 10 W/cm2 for two seconds. These tests had absolulely no effect on Starlite. therefore. unplanned higher power densities were applied. but no specific power values that might have been applied were provided in the reports. From conversations with Ward we now know that the Nd:YAG laser was applied at 3.6 Giga-Watts and the CO2 laser was applied at 3750 W/cm2 for six minutes. but Starlite could not be destroyed.
Nd:YAG lasers operate in very short pulses, so it's not really clear how much energy that actually is. 3750 W/cm2 for six minutes is pretty significant. But since all we have is Ward's recollection (or rather his claim), then it's not of much use.

I'm going to let my coated egg cure overnight, and then try to duplicate the egg test. That's the demonstration that most people focus on, and I think it's really just a parlor trick.
Metabunk 2018-12-04 15-35-18.jpg


Last edited:
The "NASA memo" looks like a pretty poor fake to me: it looks as if someone has just used a basic MS Word template and pasted in a NASA logo. It's not even the correct logo: the font for the "NASA" letters is wrong, as is the orientation of the "orbit" in the logo, and the pattern of stars - they seem to have just used an amateur clipart version off the internet:


Last edited:
There are plenty of official NASA memos of various eras available online. It seems pretty clear the Starlite one is a fake - the layout and formatting are all wrong, even without the "smoking gun" incorrect logo.



even on that starlitetechnologies(.com) "History" page, the logo is different. (although the header is crooked and doesn't line up with left hand margin.. not sure how that would happen in a fax or printer)

that page does also links to the Dateline episode with "Rudy" Naranjo. timestamp 4:08. Although Naranjo doesn't say anything like
'We have done a lot of evaluation and … we know all the tremendous possibilities that this material has.'
, unless i missed it. but i skimmed through the piece 3 times.

(add: forgot the transcript sorry: it's very hard to hear, had to hook it up to my TV. sounds like he says "well the interest is because it does so many things and it is only one material" and a few seconds later "any place there's heat it will hav applications. any place where it's radiation [unintelligible] it has applications")


Chris Bennet in the about section of that video is claiming NASA stole Starlites formula.
External Quote:
Starlite is the most important invention of all time that you have never heard of. Stolen by the Boeing Company, NASA, and the FAA. Media organizations have refused to investigate and tell the story for over 10 years.
Last edited:
That was a mixture of PVA glue and Baking soda. It was also six minutes vs the BBC's three minutes, and I had the torch quite a bit closer.

Metabunk 2018-12-05 10-43-29.jpg

Of course, my coating also looks a bit thicker.

But I'd like to replicate it a bit better, so I'm going to try again, this time replicating the BBC distance (about 3 inches) and trying to replicate the intumescent properties.

I'm thinking what we need is more carbon, so I'm going to use some icing sugar (very finely powdered sugar). Sugar carbonizes and should also create the gas needed to puff up the material.
I'm thinking what we need is more carbon, so I'm going to use some icing sugar
in the Dateline video Wayne Maurice mentions flour, although im not sure if he was kidding.

also.. the torch doesnt heat up the surrounding air? i was surprised your egg didnt cook a little at least since you had no protection on the back of the egg.
Last edited:
I think this all popped up on my radar again because of the BBC Reel documentary, which was only just released. This segment gives a good explanation of how it works.


It also shows some tests on samples provided by Ward's daughter, which seem to show more intumesence than the egg test.
Metabunk 2018-12-05 10-59-32.jpg

But seems rather similar to the KingKote799 coating
Metabunk 2018-12-05 11-00-54.jpg
also.. the torch doesnt heat up the surrounding air? i was surprised your egg didnt cook a little at least since you had no protection on the back of the egg.

It did cook a little, but only on the side that was "protected."
Metabunk 2018-12-05 11-34-42.jpg

That's down to the things I mentioned above. The failure of the material to swell up (hence the decision fo more baking ingredients), the closeness of the torch, and the duration of the test (six minutes vs. three minutes).

Another variable might the TYPE of torch. I'm using a very standard propane torch head, Bernzomatic UL2317 running on propane. Described as a 1/2-inch wide pencil flame.
Metabunk 2018-12-05 11-52-15.jpg

The BBC torches seem to have larger nozzles.
Metabunk 2018-12-05 11-38-18.jpg

Metabunk 2018-12-05 11-40-27.jpg

I'm not sure that this means in terms of the amount of heat supplied. The BBC torch is like this, a Taymar LG 870 Butane Blowlamp.

Metabunk 2018-12-05 11-50-14.jpg

Which is butane, not propane. The BBC presenter says it's 1200°C.

Just doing a search for temperatures, the internet says:

Butane torch: 1,430°C
Propane torch: 1,995°C

Well, heck. No wonder my egg cooked a little. Hotter torch, much closer, twice as long.
Last edited:
This guy using his own composition "Aeromax Starlite" used a Butane/Propane torch
Metabunk 2018-12-05 12-25-25.jpg

He also has a supposed thermocouple set up, which registers an initial peak of 1997°C, then decreases to 1665°C as the thermocouple gets hotter
Metabunk 2018-12-05 12-30-25.jpg

Then goes back up to 1943°C after it has cooled down
Metabunk 2018-12-05 12-31-29.jpg

So obviously those numbers are not the actual temperature. The thermometer is a Vichy DM6802A+, set to T1-T2 mode. There's a thermocouple plugged into T2. So it kind of looks like he set T1 to some high number. Maybe he just did not understand how it worked.

You can't actually use a simple thermocouple to measure flame temperature like that. All you measure is how hot the flame makes the thermocouple, which is usually a lot less than the flame temperature.
i see he cracked it and the yolk fell out, but so did yours.

Yeah, that's part of the showmanship aspect. I could quite easily post a video and get people convinced I'd reinvented Starlite.

With the amount of moving around he did and the lower temperature, I think his egg probably did not cook much. I'm going to try again tomorrow with my 1/3 sugar composition and a torch setup closer to the BBC.
It also very closely resembles this:
Metabunk 2018-12-05 14-44-19.jpg

That's baking soda clay. 2 parts baking soda, 1 part cornstarch, 1 part water. Combine baking soda and cornstarch in a pan, add water then heat and stir until it looks like that. Then put it in a bowl, cover with a damp towel for 15 minutes, knead it a bit, and then put it in a sealed container.

I feel fairly confident that this will work just as well as Starlite at protecting an egg under BBC conditions.
I'm cooking up some miracles!
Metabunk 2018-12-05 17-35-41.jpg

My secret ingredient:
Metabunk 2018-12-05 17-36-11.jpg

Burn tests tomorrow. I'm letting some dry.
Just doing some tests on the effects of flame distance on a butane flame (micro-torch, pencil shaped). Remember the temperature of a butane torch flame is 1,430°C

At the base of the flame, 622°C
Metabunk 2018-12-06 09-29-24.jpg

0.25" 725°C
Metabunk 2018-12-06 09-30-50.jpg

0.5" 805°C
Metabunk 2018-12-06 09-32-02.jpg

1.0" 976°C
Metabunk 2018-12-06 09-32-48.jpg

1.5" 984°C
Metabunk 2018-12-06 09-33-59.jpg

2" 874°C
Metabunk 2018-12-06 09-35-01.jpg

At this distance the "hot spot" on the flame started to rise quite a bit, meaning some adjustments.

2.5" 717°C
Metabunk 2018-12-06 09-36-37.jpg

3" 691°C
Metabunk 2018-12-06 09-37-39.jpg

Between 3" and 4" there was a rapid drop-off, and the upwards curve of the flame made consistent measurements difficult. But at 4" it did get to around 300°C. I was also able to hold my hand at the 5" mark where it was a bit painful, but not injurious.
Metabunk 2018-12-06 09-38-40.jpg

Metabunk 2018-12-06 09-43-02.jpg
Metabunk 2018-12-06 09-50-44.jpg

Hard to say how that translates to the BBC torch. It has a more diffuse flame, and of course, I'm measuring how hot my torch gets a cheap thermocouple, not how hot it gets an egg. The heat delivery is totally different.
I wrapped the thermocouple in "Starlite" (Formulation: BSC+PJ.) After 1 minute it was at the boiling point of water.
Metabunk 2018-12-06 10-04-35.jpg

However, I'm not convinced three minutes of this would cook an egg more than in one small spot. Also, this is obviously a terrible formulation.
The BBC torches seem to have larger nozzles.


To get closer to the BBC experiment, I've bought a Blazer Big Shot butane torch, which has a flame about twice the size of the micro torch.
Metabunk 2018-12-08 11-21-54.jpg

This is probably reasonably close to the BBC's Taymar torch in terms of power output. Looks like it's about 2.5" away from the egg, and the longer more powerful flame makes aiming it easier.
Wondered if you'd seen this comment on The Guardian website?
External Quote:
Having read a recent article on Mr Ward I've decided to leave a brief & factual synopsis of why Mr Ward's product never came to fruition. Mr Ward came to my lab about a year before his death needing help to turn what was essentially a party trick into a useable & commercialy viable product. The problem he had was although the powder component did exactly as it said on the tin, he had found no way of applying a lasting coating. All he really has was some powder mixed with PVA glue, the problem being that although you could apply it to certain objects it's longevity was no more than 2 weeks. While testing we discovered that a sample he'd kept for almost 10 years could be destroyed in a matter of minutes under a methylacetylene-propadiene propane blowtorch. Unfortunately after many samples & tests we where unable to find a effective application method & we parted company on good terms. Sadly this is the true reason why Mr Ward was never able to sell or bring his incomplete product to market. But rest assured, as of this time I can say that there is at least 1 complete & superior product in testing, testing that so far is going remarkably well. So one day there will be a product on the market that will save life's while also having countless other uses. The inspiration behind this project.... Mr Maurice Ward,5753,-5575,00.html
One of those that's difficult to buy. But you never know...
Ha, a "party trick" with "some powder mixed with PVA glue"

The egg thing is absolutely a party trick. Pretty much ANY white paste you apply to an egg will protect it from cooking for three minutes.

Right now I have an egg coated in baking soda clay mixed with PVA glue. This will probably be the last test I do.
I think it would be interresting to also try it with an unprotected egg with the same parameters to see how much it cook.
I think it would be interresting to also try it with an unprotected egg with the same parameters to see how much it cook.
An unprotected egg will crack open in a few seconds.

I wrapped an egg in three layers of aluminum foil to give a minimal amount of protection. It still cooked to 100°C in three minutes.
Metabunk 2018-12-09 07-36-31.jpg
This is what happens if you heat it directly (which I did just 20 minutes ago). About 5-10 seconds.
Metabunk 2018-12-09 10-07-30.jpg
All [Ward] really had was some powder mixed with PVA glue, the problem being that although you could apply it to certain objects it's longevity was no more than 2 weeks.

Looking forward to seeing if your two-week-old samples fare better than Ward's. :)
I came across this video today:

The author, Nighthawkinlight, made his (if I recall correctly) with 10 parts corn starch, 1 part baking soda, and then enough PVA to make something that looks like Play-Doh or plasticine. He has some interesting comments about dry vs. wet, the insulating properties of carbon, and the use of sugar. Not a lot of details, but well done nonetheless.

Most interesting is the showmanship, however. He literally formed it into a pancake, held it in his hand, and lit up the propane torch. He has another video where he melts 3 pennies on the pancake while holding it in his hand.
The author, Nighthawkinlight, made his (if I recall correctly) with 10 parts corn starch, 1 part baking soda, and then enough PVA to make something that looks like Play-Doh or plasticine.
Yeah, I needed more corn starch to make more carbon.
Baking soda and baking powder is used in cooking to "leaven" the product before or during cooking. Meaning, it will add air bubbles to a raw cookie or cake mix, and "set", usually from the cooking (hardening) of the included beaten eggs.

Baking soda is sodium bicarbonate, and will foam with a somewhat acidic catalyst (think of those school science-fair volcanoes).

Baking powder is partly... baking soda, monocalcium phosphate, and sodium aluminum sulphate (plus a little corn starch). When mixed with water, it will become fizzy (activated).

Cream of tartar is sodium bitartrate, and is used in stabalizing egg whites for merangue (yummy)....and a version is included in baking powder as the catalyst.

The point is, all of these can become an airpocket-filled barrier, aka - used as an insulator if stabalized in their foam state..

Aerogel is a type of industrial "foamed" open celled silica-type insulator, with remarkable properties...

In the Hollywood special effects biz, we would add baking soda to Plaster of Paris, to make fragile "break-a-way" bricks or tiles, for stunt work or exploding miniatures .... they would break easily because the plaster would set, but was full of air bubble.
Last edited:
Interesting notes....
PVA glue is supposedly a neutral PH adhesive (non acidic), so how would it create a bubble foam insulator if mixed with baking soda ? (Heating ? )

Fire proof (intumescent) paint is mostly misunderstood. Such paint is not "fire proof", but "fire retardant", meaning it WILL char, but will self-extinguish itself within a few seconds or less, and not continue burning or smoldering.

Sodium Silicate (or other salts) is added to the paint, or applied to or over flammable substrates, to retard flame.

Often, we had to adhere to fire regulations, and our painted sets or Hollywood backdrops had to be fire retardant, applied by a certified applier. Fire Marshals would sometimes check by using a flame test (in a safe controlled manner)

- were the Starlite tests done on corregated paper cardboard, or aerospace corregated epoxy panels ?
(I have some epoxy samples, not easy to find...)
Last edited:
Fumed silica (aerosil, cab-o-sil), is usually a powder, and will not burn at reasonable temps.
If I were to try to make a thin barrier and heat dispersive material, I'd use this to thicken it....its microns in particle size (5-50nm) and thermally dissapative.
Last edited: