Blog thread, for discussion visit APC forum.

---

I received my chems! Finally i have KCl and bunch of other interesting chemicals.

Copper salts. Purpose: blue flame. Aim: explore ways to bind coper to either fuel or oxidiser to avoid dead mass in the composition.
Organic cathions (acids) can serve as a fuel. Such cathions include: citrate, benzoate, sorbate, ascorbate, stearate. It is desired that the final copper salt is insoluble and non-hygroscopic.
My source of soluble copper is copper oxide dissolved in acetic acid (vinegar). In first batch, I tried benzoate and stearate. Benzoate is obtained by precipitatuon from solution of sodium benzoate and copper acetate. Precipitate forms somewhat voluminous gel, but it is not that bad, filtering and washing is easy. The product as well as solution smells funny. I am not sure where this smell is comming from. Copper benzoate is not new in pyro. Even videos arw available detailing it's synthesis.
Flame test showed very positive results. Test composition with chlorine based oxygen burned very very fast, with a bright flash.

Inspired by success of copper benzoate, i went ahead and repeated a similar procedure for stearate. Common soap is sodium stearate. I just dissolved it in warm water and added copper acetate. But due to acidity of my solution, the solubility of soap decreased. Stirring helped. A very volumious light blue slimy precipitate formed. Despite the goey form, filtration proceeded and rest of the solution was squeezed out. Washing the precipitate was not very effective and flame test showed lots of yellow with hint of blue. Copper stearate is like soap, sticky and waxy, but unlike soap it does not wash away. This makes it quite unpleasant to work with. Test composition with chlorine based oxygen resulted in a unstable burn wirh a hint of blue.
I will not persue copper stearate further.

Copper bound to oxidisers.
Copper chlorate: water soluble, hygroscopic, unstable. This salt is close to being a explosive, and therefore I want to stay avay from that.
Copper nitrate: water soluble and binds up to 6 molecules of water.
Copper perchlorate: is hygroscopic and I do not have perchlorates.
Stop.

For turning PC on/off/reset using gpio. The switches are biased to 3.3V. This voltage is comptatible with raspberry. Hopefully the grounds can be tied together. The LEDs are at 5V so they may not be connected directly. 3v3 rail from psu can be used instead.
This is to remotely reset the PC if it freezes and I am not around.
Today i tested such connection. The ground lanes are compatible. Both reset and power operates around 3.3V and are active-low. I included a safety resistor on the high side between pi and the computer. The gpio control works and I can issue both reset and power signals from the Pi. Just there is a bug in the gpiochip chardev driver, the io-line is not released to default (input, tristate) as it should, when the userspace program exits.

Meanwhile:
Dissolving KCl and removing the red impurity.
I precipitated and filtered all KClO3 from the liquid electrolyte. It remains to redissolve the solid from the cell and precipitate that too.

Pressure filtration. Filtration works much faster over pressure gradient, either via vacuum on the output or elevated pressure on the input. Easy source of pressure is height and gravity.
Experimental apparatus has been assembled: input contained about 1 m above the filter, rubber hose (d=6mm) leading from the tank down to filter, collection tank and a filter assembly. The filter assembly consists of pet bottle cut like a funnel and a filter membrane attached to the wide end with rubber band, which consists of: cloth, 3 layers of paper tissue, and cloth again. The hose is air-tight attached to a hole in the bottle cap.
This works: both fast and clean filtration. But the plastic bottle is quite flimsy and can deform under the rubber band pressure, which rips the tissue.

Cell design considerations:

How to protect electrical connections to the electrodes from corrosion? Either the electrodes must have titanium strips welded to it that lead outside of corrosive environment. Or the electrodes themselves must lead thru the lid to make connection on the outside. Alternatively the connection can be made inside the cell and then insulated from the corrosive fumes, but this has proven very difficult and failed multiple times.

Considering electrode leads: It seems titanium pipes are more available than titanium sheet. Plus I do not have equipment to easily cut such sheets, but tubes can be gotten in fitting size. One end of the tube can be clamped to provide flat surface for the weld and the rest can be (optionally) filled with Zinc to increase it's conductivity.

Sizing the Anode: First of all, it must fit into the reactor chamber. Second, it must support the current I want to run. I am not what current should I aim for. Some say the highest possible. But with high power the heat becomes a issue and it may produce more chlorine based oxygen that I am ready to deal with. I want between 5 to 20 amperes.

The bulk of electrolyte in the reactor acts as a buffer. The ratio between reactor volume and current determines how quickly the buffer depletes. WSM suggests to not go below 10% of chloride saturation, I do not want to let the chloride level to drop below 50% saturation. This can be calculated:

100% saturation of KCl at 60°C is 458 g/l, that is 6.143 mol/l.
It takes 9 moles of electrons (=faradays) to convert one mole of KCl (without pH control). Thus it gives the formula (comptible with Qualculate):
t= ( (6.143 mol/l) * (9 faraday) * V ) / ( 2 * I )

For example 1 Liter reactor running at ridiculous 20 Amperes would consume half of the buffered KCl in 3 days and 2 hours.

I started recrystallization (temperature swing) of the crude chlorine based oxygen obtained previously. It is not very nice, because the process involves hot solutions that do cool down and also evaporate which forms unwanted crust.

After 3 weeks, I was told that the titanium items I ordered were unavailable. Not great! Yes, I will get my money back, but I needed those items yesterday already. Now I have to look for another source.

Potassium Persulfate is a safe cell additive. It is an alternative to the carcinogenic potassium dichromate, which is carcinogenic. Sodium persulfate is available in electronics supply stores, because it is used as PCB etching agent. It should be easy to convert it to K-version, which has low solubility in cold water.
For the electrodes, I am in touch with a supplier.
Potassium Persulfate is also available.

I received colophony resin powder, along some chemicals. This powder will be used as a reducing agent instead of flour for experimental compositions.

Flour based stars.
Organic pyrotechnic stars obviously use some kind of organic fuel. Common organic fuels are shellac, colophony resin, red gum, dextrin, parlon and pvc. The later two serving dual role as a chlorine donor. Now these chemicals, while unrestricted, are not particularly common. Because of that, I decided to look at commonly available organic materials. First, let's look at he unsuitable ones.
Sugar: it is a potent organic reducer, but it is hygroscopic, even more so when mixed with sodium chlorate.
Citric acid: low temperature reducing agent, but also very hygroscopic.
Candle wax: this mixture cannot be ignited with just a flame, and thankfully, because it is a secondary high explosive of FO type.
Wheat Flour appears to be good candidate: it is not as strong in reducing potential, but it is not hygroscopic. It comes conveniently powdered in multiple degrees of fineness and is cheap. When mixed with sodium chlorate, it burns tamely with well-formed flame envelope. The flour, when activated with water, acts as a binder too. The chemical composition of flour can be approximated to starch with the chemical formula of C6H10O5. When calculating amount of oxidiser, combustion to CO and H2O should be considered, not CO2. When mixed with just chlorate, it does not burn very fast and is hard to ignite. Charcoal powder was found to be good additive to improve burn speed and flame sensitivity. Added charcoal should be balanced with extra oxidiser.
Some years ago I developed a formula of 4:1:8 (flour:carbon:chlorate), but recently I changed the ratio because it was not sensitive enough.

Wheat Flour ... 12
Charcoal ... 4
Sodium Chlorate ... 28

Wet with small amount of water. This composition can be pressed into commets or moistened more and cut. Obviously, it burns yellow.

Reposted: https://www.amateurpyro.com/forums/topic/13343-flour-based-stars/

Personal Note: https://boinc.progger.info/odlk/forum_thread.php?id=49
Also communication.

After 3 weeks, I was told that the titanium items I ordered were unavailable. Not great! Yes, I will get my money back, but I needed those items yesterday already. Now I have to look for another source.

No reply from store owner. Boo. Do not do business with Šíma.

Todo: detail about custom made Chinese electrodes.

More chemicals:
https://www.poolman.sk
https://www.pgchem.sk/

Container.
Glass is said to be the idea material. While there are glass jars of various sized available for free, finding a corrosion resistant airtight lid is very hard. Lids for them are exclusively made of coated mild steel. Plastic one would have to me manufactured.
There are 5l paint/food buckets available with lid that snaps on. At first look such lid does not appear to be airtight, but it prevents the paint from drying and food from spilling, so it must be airtight. The material is usually polypropylene, which is not indestructible, but fairly resistant.
Cabbage/sauerkraut containers: 5l they are quite expensive, but come with a wide screw-on lid. I cant find details about the material or whether the lid comes with seal.
Bottles from protein supplement: of course, buying that supplement just for the bottle is unwise, but some fitness clubs may have one for free.
Building a container from glued pvc sewage pipes: again, a airtight lid is a problem. If I glue a plug on shut, I would not be able to inspect the electrodes or remove precipitate.

Chlorinated hydrocarbons.
The colour emitting species in the flame are usually chlorides. The metal ions (Cu,Sr,Ba) are less colorful in flame. For this reason chlorine donor is often added to the composition. Pyrodata lists handful of chlorine donors, ranging from common to exotic.
What about chloinating copper benzoate?
It turns out, common bleach can attach a chlorine atom to the ring of sodium benzoate! No special conditions are required for the reaction, just mix sodium benzoate and sodium hypochlorite in room temperature solution, and after a while, the product can be removed.
Copper Acetate (or sulphate) would be then added to the solution to precipitate copper chloro-benzoate. If the leftover bleach is found to interfere with the copper donor, acid can added to precipitate the chlorobenzoic acid in more traditional manner. Literature mentions attachment of the chlorine at different positions on the ring, or even chlorination to higher degree, but that's of little importance to my application.

Hopkins; Chisholm: CHLORINATION BY AQUEOUS SODIUM HYPOCHLORITE