woelen

I found a nice word of wisdom on Usenet:rec.pyrotechnics, and here follows my liberal phrasing of this: Playing with KClO3 and red phosporous is like playing with a deadly poisonous snake. Playing with KBrO3 and red phosporous is like playing with a deadly poisonous snake, which is slightly pissed-off.

Yes, you are right. Red P can be used with some oxidizers (KNO3, KClO4, but YT probably can give more authorative answers on that), but even then it is not the reductor of first choice. In the Netherlands, the use of red P in consumer fireworks is prohibited and I see no real reason why you want to use it for that. But I'm quite sure, that if you look at pyro-books, that there will be compositions with red P in them, but again, here I would like to refer to the real pyro-experts.

Jdurg, I did the test, I was wondering what happens: With KNO3 and C, finely powdered. The reaction is self-sustaining, but not as smooth as one would like. I took the melt and added some acid to it, and it bubbles a little and I saw small amounts of brown vapor, that must be due to NO2. So, I conclude that the mix contains at least some KNO2, also K2CO3. I only took a mix of let's say 200 mg, it might be that in larger quantities the reaction proceeds smoother, but this small quantity needed to be relighted another time. Next, I took KNO3, C and S, all finely powdered. When lit, this mix burns quickly and smoothly, however not violently (it's a huge difference, compared with the KBrO3 mix I also made, see other thread on KBrO3). After the reaction, I took the molten and again solidified mix and added some acid to it. Definitely, some H2S could be smelled, there was a smell of rotten eggs, but I hardly had any bubbling. This is not what I expected. If the molten solid was mainly K2S, then I would have had a lot of bubbling. I did a counter experiment with some of my commercially obtained Na2S. This indeed bubbles vigorously, when 10% HCl is added. So, I conclude that a small amount of K2S is formed, but the main products are different. I suspect most of the stuff formed is K2SO4, with only small amounts of K2S and K2CO3. Have a look at this: http://en.wikipedia.org/wiki/Black_powder Your reaction with K2S is mentioned, but another reaction with K2CO3 and K2SO4 is mentioned as well. The truth is even more complex, with a LOT of side reactions and one cannot give a good equation, representing all products of burning. This also is stated by the wikipedia article. I just did this little research with my own chems out of curiosity. This simple reaction still has a lot of interesting features to be investigated.

What language is this ??? Please keep the level of this forum at an acceptable level!

Lab application of what? I asked for the potassium bromate, not for the silver fulminates. I'm a little bit confused here. What could be the mechanism behind that? I doubt whether this can detonate without an oxidizer. I'll give it a try by making some and hitting it with a hammer. An another option is to heat it with a propane torch. Don't worry, I'll do this with only a few tens of mg. I asked these things just out of curiousity. I only wanted an opinion on the compound Ag(NH3)2BrO3. I do not agree with your comparison with ammonium chlorate. Here I'm talking about a complex of ammonia, not ammonium. I cannot imagine a compound, made of ammonium chlorate and then silver hinged in somewhere. The compound, I'm talking about contains ions [Ag(NH3)2](+) and BrO3(-). Of course, I agree with you that this might be a really dangerous compound. YT, if you think that this is going over the edge of SFN's rules, then I'll stop with this discussion and I'll respect the rules. In that case, feel free to remove this part of my response or edit my response. I'll wait and see what remains of this thread .

Another way to make really beautiful colloid gold (that is how it is called) is the following: Dissolve some HAuCl4 in water. Approximately 0.0001 gram per ml of water, so you need ridiculously small amounts. Prepare some sodium citrate solution, by dissolving citric acid in water, together with some sodium hydroxide. pH is not very critical, better a little on the acidic side. Add this to your dilute HAuCl4 solution. Your liquid slowly becomes deep blue/indigo. You also can get red, yellow, purple, almost any color you like. The color depends on the particle size. Typical particle size is in the order of nanometers, each particle consisting of a few hundreds to several 10000's of atoms. If you don't have HAuCl4, you can prepare some by dissolving gold in aqua regia. If you take 0.01 gram gold scrap, dissolve that in a few drops of aqua regia and dilute up to 100 ml with distilled water. With that 100 ml you can do many colloid experiments. As I mentioned in the pyro-thread already (very off-topic over there), I use gold in my photography dark room. What I do is preparing photographic pictures, in which the image is built up of microscopic gold particles. These pictures have a remarkable deep blue color, VERY nice. When a silver sulfide image is gold-toned, then you get a beautiful red/purple color. My toning solutions are extremely dilute. They contain up to a few % of thiocyanate, SCN(-), and somewhere between around 0.01% of gold. Even with 0.001% of gold you can do remarkable things. Now you can understand why 1 gram of HAuCl4, which seems very expensive, is not that expensive, given the use of it.

I use it for toning photographic images. Gold-toned pictures are among the most beautiful pictures. Besides that, I also use it for experiments . HAuCl4 is called tetrachloroauric acid. It is a fairly strong acid and there also are salts of this, the anion AuCl4(-) is called tetrachloroaurate (III). It is a deep yellow/orange anion and it is one of the most stable compounds of gold. Keep in mind, that flame tests do not necessarily yield the same color as when a metal is used in fireworks. The observed colors are affected by subtle interactions. E.g. copper salts frequently give green flame tests, but in pyrotechnic mixtures they may also be blue as I've understood. So, there is more to say about that.

I expect not very much, maybe you get some sparkles of molten gold drops, but these would not be spectacular. Iron, aluminium or magnesium sparkles are more spectacular, because they do not only glow, but they also burn, which gives much added light. If I ever win a kilo of gold powder or gold chloride on eBay for a few dollars, then I'll give it a try but chances are not good at all. I have some HAuCl4, but that stuff costs $30 per gram, so that's not something for a firework composition .

The argument of bromide toxicity seems fairly meaningless to me. Bromate is reduced to bromide and if some bromine were formed, then that is converted to bromide really fast under normal conditions. The average piece of fireworks contain barium salts for color, copper salts for color, chromates or dichromates for Mg-coating, and who knows what more. Then a little bromide does not seem a real concern to me. Bromide only is marginally toxic. The argument of NaCl being so easily accessible truly is a valid one. But, if one is looking a little bit further, then NaBr also is easy to find (swimming pool supplies, photography raw chemical shops and not to forget eBay at some times). The somewhat more serious pyrotechnician is not held back by this, such a person also has other harder to find chems, so why not get some NaBr? I have experienced that it is easier to make bromate than chlorate, due to much smoother reaction of hypobromite to bromate and due to the anode not being pulverized as quickly as with chloride electrolysis. The latter phenomenon is a real pain in chlorate production. The color part might indeed be the most important reason. But, does bromine not enhance color? Well pyrotechnicians, jump on this thread and show us what you know about this . Ah, while I was typing YT jumped in . Yes, I have seen the MSDS of KBrO3 also, but as I noted, toxicity of other pyro-chems sometimes is even worse. Heavy metals are not broken down, bromate and bromine are destroyed very quickly under outdoors-conditions and I expect their chemical halflife to be only in the order of hours, when coming in contact with soil, natural waters, etc. The reaction product will be bromide. With regards to stability, I have some of the stuff right now, and it is quite stable. KBrO3 on its own cannot be made to decompose violently. Even when it is hit with a hammer, nothing happens. On heating with a propane torch directly on the solid apparently nothing happens, but it looses its oxygen, making KBr. When treated with concentrated H2SO4, then O2 and Br2 are released in a vigorous bubbling reaction (somewhat like when H2SO4 is added to solid NaCl), no bromine oxides are formed. Mixes of KBrO3 and reductors indeed are quite sensitive. I made a few batches of 100 mg of KBrO3/S mix and wow, that stuff is energetic, it flashes with a bright greyish/purple light and then it is gone! It does not explode, however. I have the impression that KBrO3 mixes are more energetic than corresponding KClO3 mixes. Unfortunately I dare not light some larger batches, due to the very strict regulations on pyro in the Netherlands (but Dec 31 is coming, that gives a little more headroom for just one day). I also do not dare to store KBrO3/S mixtures, due to my lack of knowledge on stability of that kind of mixes. I mix the stuff in-place and burn it immediately.

I'm wondering why all those people want to make KClO3 as oxidizer for pyrotechnic compositions. NaClO3 can be made in electrolysis cells from NaCl and then be precipitated out as KClO3. I have done some experiments now with NaBr, instead of NaCl and I was really nicely surprised to see how easily one can make KBrO3 from this! Electrolysis of NaBr has quite some advantages over electrolysis of NaCl. 1) No trouble with gaseous chlorine gas, which corrodes all things like hell in the neighborhood of the cell and also inside the cell. Any bromine formed remains in solution, it sinks to the bottom and does not bubble. Sometimes the cell contents need to be mixed well in order to bring hydroxide from the cathode region to the bromine in the anode region. 2) The reaction, in which bromate is formed is almost instant. Chlorine gives hypochlorite, bromine gives hypobromite, but the latter VERY easily disproportionates to bromide and bromate. The cell does not need to be warm. 3) This is a very important one: My graphite anode hardly is corroded (pulverized) by the electrolysis process. Bromine comes off as liquid and no small graphite particles come in the liquid. This is something which I really like . When electrolysis is allowed for a long time, then at a certain point you'll see that quite some gas is produced at the anode as well. Then it is time to stop. This gas is oxygen and if that is produced in appreciable amount, then it means that most bromide is used up and bromate is formed. Potassium bromate can easily be precipitated from the liquid by adding potassium chloride solution. KBrO3 is only sparingly soluble, just like KClO3. One-time recrystallization gives purer KBrO3. The sodium otherwise may be a problem in compositions where color is required. Finally, NaBr is cheap. You can get it at a price tag of a few dollars per pound, when purchased in kilo quantities at technical grade quality. Why isn't this process used more in home-brew pyrotechnics? I personally did some experiments with KBrO3 as oxidizer and I must say, it is powerful, comparable to KClO3. I'm not sure about its long term sensitivity and stability though in compositions. KBrO3 itself, however, has excellent shelf life, it is not hygroscopic at all and crystallizes very well from aqueous solution. If someone can shed some light on this and on the usefulness of KBrO3, then I would be pleased. One remark: If you intend to repeat my experiment, please keep in mind that bromate might be carcinogenic. In the past it was used as a food additive in cakes, bread and so on, but that use is abandoned since the 80's.

Azides have very practical applications. Each modern car contains a canister of sodium azide in an airbag. An acceleration sensor triggers a heater, when the acceleration (better: deceleration) goes above a certain limit. The heat of the heater (usually a simple wire inside the pellets of sodium azide) causes this to soft-explode and within 50 ms or so, the entire airbag is filled with non-toxic nitrogen gas. The NaN3 decomposes to Na-metal and N2 gas. Usually some other chemical is molded around the canister to neutralize the sodium metal formed just in order to avoid fire. Most cars have a few ounces of NaN3. (Ah, jdurg was first while I was typing, well done jdurg !) Another application of azides is as primary detonator for situations, where the arms can become moist or wet. Lead azide remains explosive, also when it is wet. A small amount of lead azide is used as primary detonator to start the detonation of a larger less sensitive (but usually less toxic and more powerful) load. ------------------------------------------------------ The high temperatures in fireworks, explosives and so on, indeed make it difficult to describe the reactions with a nice stoichiometrically well-defined chemical equation. However, one usually can state that nitrogen in the compounds is released as N2, hydrogen almost invariably is released as H2O and carbon and sulphur mostly are released as CO2, SO2 or a carbonate salt and a sulfate salt. Of course, it also depends on whether sufficient oxidizer is present in a composition. E.g. if you have C, S and KNO3 as BP, when the mix is KNO3-deficient, then not all sulphur is oxidized and some sulfides are formed. If the KNO3 is present in excess amount, then part of it is not reduced further than KNO2 and in fact, this sometimes is used as a means to make nitrites from the esaier to obtain nitrates.

In fact, azides are made in multi-tonne quantities every year. Most people have a few ounces of azides around their home and some even almost a pound . Probably you (I think in your case your parents) also have. Think good and you'll find the answer!

Well Ryan, here I am again . My remarks on the green color I did in the previous version of the thread already have been covered much better by YT, so I'll not elaborate on that again. I'm not a real fireworks expert, mainly due to the fact that pyrotechnics is forbidden in the Netherlands. Yes, forbidden, period. I did some small-scale experiments, however. With sulphur you can make beautiful blue blue. I once made some liquid sulphur, made it burn and then caused a small explosion in it. This gives a spray of beautiful ghostly blue flares of the burning molten sulphur. I'm not sure, whether this can be created in real fireworks, but the effect is neat. As far as I know the following colors can be obtained from the salts of these metals: barium - (greyish pale) green (YT knows how to make it brilliant ) strontium - beautiful bright red calcium - pale red lithium - red From metals themselves, brilliant flares can be produced: magnesium metal aluminium metal magnesium salts and aluminium salts do not add color to compositions and in fact are quite useless. I once did a small experiment with K2CO3 and KNO3. A really remarkable mix is K2CO3 / KNO3 and S. It can be made to explode, even in the open without confinement. You need a special preparation, but I'll not elaborate on that over here (k3wls also read this). This is a very remarkable reaction, because K2CO3 is not an oxidizer at all. I do not understand, why this mixture explodes and does not simply burn like BP. My last experience with energetic compounds was with azides. I was so lucky to obtain some sodium azide (NaN3), a very hard to obtain chem. Wow, that is nice stuff, not on its own, but it can be used to make compounds like Pb(N3)2 and Cu(N3)2. These are explosive compounds, which can be made to detonate, even when wet. These are a little scary though. Right now I'm doing a larger project on nitronic acid, nitrolic acid, nitronates, and the nitrolates. These are derived from the aci-form of nitromethane, a very common fuel enhancer. If I have interesting results, then I'll post them here. However, making these at home is not that easy, because they require a lewis acid-base reaction under totally anhydrous conditions.

If you look at the real numbers, yes, but when you extend to the complex plane plus {inf}, the Riemann-sphere, then the pole in 0 of 1/x has a meaning, while something like z/z* (z* means conjugate of z) is ill-defined in zero, one cannot speak of a zero or a pole. An entity like z/z (or sin(z)/z from my previous examples) can be defined in 0 without problem. What I want to point out is that the whole issue of division by 0 and the corresponding limits is a much more subtle matter than most people think of.

Yes, but that is a whole other story. Taking into account direction of approach opens up a whole bucket of new bizarre things. Related to this, there is a limit, which is finite and yet undefined: (A+B*i)/(A-B*i), with i*i = -1. Now choose (A, B), such that A+B*i equals eps*exp(i*phi) and let eps go to 0 from above. Now you see that the value of this limit depends on the direction from which 0 is approached. You can make it equal to 1, to -1, to i or even sqrt(2)+i*sqrt(2), just as you like. Choose a good direction and you get any value you like on the unit circle in the complex plane. So, the actual value of this limit is undefined and this even in a more bizarre way shows that 0/0 cannot be expressed correctly, not even in a sloppy way.

Correct. Correct These remarks are correct, up to some imprecise formulation. But I can accept them. A more formal saying is: Given a number N/Z, for non-zero finite N, and Z approaching zero, the quotient approaches infinity. Given a number N/I, for finite N, and I approaching infinity, this quotient approaches 0. This is not correct in the general case. You need more information about the nature of the limit. Suppose you have a function f(z)=sin(z) and g(z)=1/z. Now if you look at the product p(z)=f(z)g(z) and you let z approach the value 0, then you'll see that this product approaches a finite number, being 1. In fact, the funtion p(z) can be analytically extended to 0, by defining p(0) = 1. This function is infinitely differentiable at 0 without any anomality. Now suppose we have another function h(z)=1/(z*z). Now if we take the product q(z)=f(z)h(z), and you let z approach the value 0, then you'll see that this product approaches infinity. This function has a first order singularity in 0. Now we take the mad function m(z) = exp(1/z) (here exp() is the e-power of its argument). Now, if z is approaching 0 from above, then m(z) goes to infinitiy. If z is approaching 0 from below, then m(z) goes to 0. So, now you have a really weird situation. This function has a so called GESP (General Essential Singular Point) in 0, which can be regarded as an ordinary singularity of infinite degree. As you can see, this statement is too coarse. There are many subtleties in division by 0. E.g. sin(z)/z is perfectly valid in 0, cos(z)/z has a first order singularity in 0 and exp(1/z) has a GESP in 0. When the field of complex numbers is extended with one single point, called "infinity", then any arithmetic with singularities of final degree becomes perfectly legal. E.g. the function (1-z)/z has two zeros in that case, one zero being 1 and the other being infinite. The functoin (1-z)/(z*z) has three zeros, one zero being one and two coinciding zeros at infinity. The function sin(z)/z has zeros in pi, 2pi, 3pi, ... and -pi, -2pi, -3pi.... The function exp(z) has a GESP at infinity and the function exp(1/z) has a GESP in 0. Have a look at this: http://en.wikipedia.org/wiki/Riemann_sphere

I've once written such a program too. It required one point to be entered on the curve by hand and subsequent points are computed numerically. A curve is described by an implicit relation f(x,y) = 0 So, you have to compute one pair (x0, y0), which satisfies this equation. That is your starting point. Next, taking differentials of this equation results in the following: ∂f/∂x * dx + ∂f/∂y * dy = 0. Computation of ∂f/∂x and ∂y/∂x in the point (x0, y0) can be done numerically for a generic function f, but if f has a special known form, it is best to compute the partial derivatives by hand and use that in your program. Now, given your point (x0, y0) and the value of the partial derivatives, you can find a vector (dx dy), which is along the curve: any vector a*(-∂f/∂y ∂f/∂x) satisfies the differential equation, given above. Now you compute x1 = x0 + dx and y1 = y0 + dy and you plot a line segment from (x0 y0) to (x1 y1). Suppose, you want to step along the curve with steps of size h, then you can compute your factor a accordingly. A faster algorithm is to use a 1-norm instead of a 2-norm for determining a. It is not important whether you follow the curve with exactly h or with steps which at most deviate from h by a factor of sqrt(2). So, use the 1-norm |∂f/∂y| + |∂f/∂x| as the length and compute h from that. An even better approach is to make the size of step h dependent on the radius of curvature. A strongly bent curve in point (x0, y0) requires h to be small, a curve, which almost is like a straight line in (x0, y0) allows a large h to be chosen. A formula for the radius of curvature can be found here for a curve, described as x=x(t), y=y(t). This is easy to transform to an implicit form in terms of f(x,y), its partial derivatives ∂f/∂x and ∂f/∂y and its second partial derivatives ∂2f/∂x2, ∂2f/∂y2 and ∂2f/∂x∂y. That is left as an exercise for you. http://mathworld.wolfram.com/RadiusofCurvature.html You see, it is not that difficult. Most of the effort will be to transform the math to a good program, especially if you program in a language like C or C++. Computing first and second order derivatives is cumbersome numerically, due to round-off noise. Another problem may be if you want to be able to define your functions in a flexible way, without the need to recompile your program all the time. For the latter thing I've made a solution. I've developed an easy to use string parsing routine with general variables and all kinds of mathematical functions, which allows you to type expressions like "sin(x)+sin(y)+x*y" on stdin and evaluate this as a function with variables x and y, which you set. Have a look on my web site: http://woelen.scheikunde.net/science/software/misc.html You need to take the module parser.tgz. It is written for UNIX/Linux, but I expect it to compile on Windows without any problems with Visual Studio, Pelles C and DevC++. A way of drifting control is to add -k*h*grad(f*f) to the vector (dx dy). This vector steers your next point (x1, y1) towards the curve if it drifts away. You have this vector almost for nothing, you need to compute f(x,y) for each step and you already have the two partial derivatives ∂f/∂x and ∂f/∂y. This drifting control really is necessary, otherwise your drawn curve slowly moves away from the real curve. This correctionvector -k*h*grad(f*f) is zero, when your point (x,y) is on the curve, but it steers towards the curve if the point is somewhat besides the curve. If you assure that your step h is never larger than a pixel size, then this drift control assures that you will remain very close to the curve and the drift is limited such that you will not see it in your graphs. I implemented this algorithm and it is VERY robust, especially when combined with step-size control. If your curves are reasonably smooth without kinks and sharp bends then it also is very robust even without step-size control. This method is so robust, that even when your initial point is not on the real curve, but a little besides it, then you see that the drawn curve nicely starts in your point (x0 y0) and moves towards the real curve while moving forwards with small steps h.