This Message will Self Destruct… as You Read It?

A group of Harvard chemists have come up with a novel use for fire. Through experimentation, they have been able to build what they call an InfoFuse. As the name implies, it’s essentially a burning fuse that can “transmit” information.

The fuse is made from flash paper, which is paper made from nitrocellulose. Flash paper burns at a relatively constant speed and leaves no smoke or ash, making it ideal for this type of project. The chemists developed a method of conveying information by changing the color of the flame on the paper. You might remember from high school chemistry class that you can change the color of fire by burning different metal salts. For example, burning copper can result in a blue flame. This is the key to the system.

The researchers dotted the flash paper with small bits of metal salts. As the flame reaches these spots, it briefly changes colors. They had to invent an algorithm to convert different color patterns to letters and numbers. It’s sort of like an ASCII table for fire. Their system uses only three colors. The three colors represent eight possible combinations of color at any given time. Just two quick pulses allow the researchers to convey all 26 letters of the English alphabet as well as ten digits and four symbols. In one test, the researchers were able to transmit a 20 character message in less than 4 seconds.

[Ben Krasnow] found the Harvard project and just had to try it out for himself. Rather than use colors to convey information, he took a more simple approach. He started with a basic strip of flash paper, but left large tabs periodically along its length. As the paper burns from end to end, it periodically hits one of these tabs and the flame gets bigger momentarily.

[Ben] uses an optical sensor and an oscilloscope to detect the quantity of light. The scope clearly shows the timing of each pulse of light, making it possible to very slowly convey information via fire. Ben goes further to speculate that it might be possible to build a “fire computer” using a similar method. Perhaps using multiple strips of paper, one can do some basic computational functions and represent the result in fire pulses. He’s looking for ideas, so if you have any be sure to send them his way! Also, be sure to check out Ben’s demonstration video below.

27 thoughts on “This Message will Self Destruct… as You Read It?

    1. I think it might be possible, even though it would be rather impractical. Burning two strips at the same time would basically provide an OR or AND function, depending on the threshold of the sensor used. If the sensor can only distinguish between “no tabs” and “one or more tabs” burning, this would be an OR function. If the threshold is set higher, so the sensor will only respond to two tabs burning at the same time, it would be an AND function.

      One practical problem would of course be the synchronization; the strips will not burn at the exact same rate, even if they seem identical, and the tabs will probably influence the burn rate as well, so I think it would be hard to use long strips. Maybe this could be solved by introducing bridges between the two strips at regular intervals, but these bridges would also take some time to burn. Another possible solution would be to make the two strips converge and diverge again; the flame to each of the sections of strips would more or less come from a single point.

      Another issue is the fact that the output of the process (the signal from a light sensor) cannot directly be used as the input for another operation.

      If you would use a full-width strip (base plus width of the tabs), and clamp it between two comb-like metal plates, such that the base strip would be covered, and small teeth would separate the possible locations for the tabs, you could invert the “signal” from one tab. You burn the strip you want to invert, and only the tabs on this strip would ignite and burn away the same tabs on the blank strip (the metal plates would prevent the rest of the blank strip from burning). This would leave the new strip with the inverse pattern of tabs. A slight detail is that the small sections separating the possible tab positions would have to be removed.

      This way, a NOR function would be possible; if one of the two input strips has a tab at a specific location, the corresponding tab would be removed from the output strip. Note that the two input strips could be burned simultaneously, or separately. Also, more than 2 input strips could be used. To make a NAND function, it would of course be possible to invert both input strips, and then use the inverted strips as the inputs to the NOR function.

      Another way could be to use a set of metal plates with rectangular holes at each tab position, where the tabs on the first input strip would ignite the tabs of the second strip, which would, in turn, transmit the flame further sideways to ignite the tab of the output strip.To clarify, the first strip would burn completely, and only the sections of the second strip where both input strips have a tab would burn. If the first input strip doesn’t have a tab at a certain position, that section of the second strip would never be ignited. On the other hand, if the first strip does have a tab, but the second strip doesn’t have a corresponding tab, there would be no material to ignite and carry the flame to the output strip.

        1. Indeed; the idea of the metal masks wasn’t to exclude oxygen, but to conduct heat away from the strip, to prevent it from reaching the ignition temperature in the parts you want to keep.

          As for “why”; just to see if you can, and just for the fun of it a perfectly valid reasons to do anything in your spare time.

          Obviously, this method of encoding doesn’t really have much in common with the original method, since the positions of the tabs can be read visually without burning. The original method also relies on the assumptions that your adversary can’t simply make a replica after reading the strip, and that it is impossible to determine the doping without actually burning it. Also, it would be possible to “duplicate’ the message by simply cutting the strip lengthwise, therefor making it possible to read it twice, but it would probably be obvious to the received that the strip has been tampered with.

  1. Is it just me or is this somewhat useless? I can encode infomation by throwing small and big rocks at you, I can produce AM and FM modulated farts, I can burn differently colored fireworks…….. I mean, it’s possible to encode infomation in so many ways and this one is just one of them and not even a particularly useful one, since it can only be read once and it’s difficult to write.

    1. Somewhat, yes, though perhaps not entirely. Printing or writing stuff down on flash paper is sometimes used (even thought kind of infrequently) to make it convenient to destroy after use. Mostly works because of novelty too – people wonder what will happen so they actually go through the step of burning it (as they’re supposed to). Here, it’d be another step less convenient to avoid destroying it as you have to either burn it to extract the information or resort to analyzing the paper. It’d also be a step less convenient for Eve, there’s no trivial way to take a snapshot of the information but leave the original appearing untampered-with.

      But yes, if you’re planning on keeping this level of security going without having the end-users be reliable enough to guard their data and destroy it when told to, this isn’t particularly feasible as is. It’s not a field overflowing with options though, most read-once things like reactive to oxygen things (those self-destructing cds/dvds, disappearing ink, etc) can be recovered after destruction and are easier to copy than this, even though taking a video of it burning isn’t a particularly high bar.

    1. Not really, a spectacular demonstration would be a ICP-OES testing for elemental composition by ionizing the constituents in a 1500 watt RF field generated argon plasma ignited by tesla coil and using a series of refractory lenses with CCD’s to analyse light output then ultimately determining concentration based upon calibration curves determined from known concentrates.

  2. It’s not really a one-time system though, is it? The message is encoded on the strip for as long as you want, provided you don’t burn it. Burning it is just one way of reading it.

    Could the message be read with a raman spectrometer or something similar? Just don’t set it on fire!

  3. Morse code anyone?

    Dots and dashes on a plain strip of paper, for a read once, somewhat encrypted message (well not encrypted, but very few know morse code and would understand what was going on before it was all over)

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.