Make chlorine disinfectant

Three hazards, each with its rule:

• Chlorine gas. Hypochlorite mixed with an acid — vinegar, lemon juice, toilet or descaling cleaners — releases chlorine; mixed with ammonia it releases chloramines (ATSDR). Below pH 4 the dissolved chlorine is mostly chlorine gas (PLOS 2024). Chlorine irritates the nose at 1–3 ppm, causes chest pain and vomiting at 30 ppm and fluid in the lungs at 40–60 ppm. The only acid in this step goes into a 10 mL titration sample, after the iodide.
• Hydrogen. The cell gives it off for the whole run, and it is explosive in air from 4.0 % to 74.2 % (Casson and Bess, WEFTEC 2006). Run the cell open, in moving air, with no flame, lamp or spark near it.
• The solution itself is corrosive and irritates skin, eyes and airways; rinse any contact with clean water (WATA-Standard manual).

Pass direct current through salt water in a single vessel with no membrane. Chlorine forms at the anode and is at once hydrolysed to hypochlorous acid and hypochlorite in the brine; hydrogen leaves at the cathode (PLOS 2024). Nothing is collected or compressed.

• Graphite or carbon rods are the anode this tier can get. They wear away, and only about 17 % of the current makes chlorine (PLOS 2024).
• Titanium coated with mixed metal oxides (the dimensionally stable anode) is what modern on-site generators use (Casson and Bess); the WATA-Standard's titanium plates make 6 g/L. Change to it as soon as you can obtain it.
• Not chlorine gas. Dosing gas from cylinders needs compressed-gas handling and trained staff (PLOS 2024). Industrial cells that separate the gas belong to Electrochemistry.

The tested batch cell (Naranjo-Soledad and others, 2024): a 1.5 L plastic bottle holding 1.35 L of brine, and two bare carbon rods 9.5 mm across and 305 mm long, one the anode and one the cathode. Keep clips and wires above the liquid. Never use a metal container (WATA).

Brine: 30 g of salt per litre of water, the strength both WATA and the Berkeley cell use. Make it with:

• The cleanest, whitest salt you have from Salt. All salt carries some bromide, which the cell oxidises to bromate; WHO limits bromate in drinking water to 0.01 mg/L. WATA asks for refined salt.
• Clear water with no organic matter, which uses up the chlorine as it forms (PLOS 2024), and as soft as you can find: hardness scales the cathode, and on-site generators keep their feed below 50 mg/L of hardness as calcium carbonate (Casson and Bess).
• Water at 20–27 °C at the start (WATA).

About 5 V DC for 90 minutes gave 962 mg/L of free chlorine in 1.35 L at an average 3.8 A (PLOS 2024). Lead-acid cells rest at about 2.1 V charged, so that is two or three in series; each batch draws about 5.7 Ah.

• Watch for steady bubbles, the sign the cell is working, and check at least every hour. Cover the cell with mosquito net, not a lid, so the gas escapes (WATA).
• Keep the brine below 35 °C. On-site generators run at 15–35 °C to limit chlorate, and cells are damaged above 60 °C (Casson and Bess). Running longer or adding salt heats the bath without raising the strength (WATA).
• Wear: bare carbon anodes lasted about 45 hours, and reversing the polarity, by swapping the two wires, roughly doubles the pair's life. Falling current and a visibly thinning anode mean replace it (PLOS 2024).
• Filter out the black carbon dust through two paper filters: in the Berkeley tests the chlorine fell only from 451 to 439 mg/L.

Iodometric titration (IGNOU laboratory method) gives grams of available chlorine per litre:

1. Into a glass flask put 10 mL of the product, then 10 mL of 10 % potassium iodide solution, then 10 mL of acetic acid.
2. Stand it in the dark for 5 minutes.
3. From a burette, run in 0.05 M sodium thiosulphate (12.4 g of sodium thiosulphate pentahydrate crystals, Na₂S₂O₃·5H₂O, or 7.9 g of the anhydrous salt, and 0.1 g of sodium carbonate per litre) until the colour is pale yellow.
4. Add 2 mL of 0.5 % starch solution (starch heated in water at about 80 °C for 5 minutes, then cooled) and titrate drop by drop until the violet colour just disappears.
5. Strength in g/L = 0.177 × mL of thiosulphate. A graphite batch near 1 g/L takes about 5.6 mL.

Thiosulphate made by weight is only approximate; standardise it against a primary standard wherever a laboratory can.

Chlorine works only on clear water. Settle or filter first, as in Make water safe. WHO wants turbidity below 1 NTU, or below 5 NTU with higher doses or longer contact where 1 NTU cannot be met, and pH below 8. Chlorinate after a slow sand filter, never before it: chlorine kills the bacteria that do the filtering (WHO Technical Note 11).

Jar test: fill a row of buckets with the same water, dose each a little more than the last, wait the contact time, and choose the dose closest to the target residual (Emergency WASH Compendium). Without a residual test, use WHO's household doses: 2 mg/L for clear water (below 10 NTU) and 4 mg/L for turbid.

Millilitres of product per 20 L bucket = 40 ÷ strength in g/L for 2 mg/L: about 42 mL of a 0.96 g/L batch, twice that for 4 mg/L.

Contact: at least 30 minutes at 18 °C or warmer, longer in colder water (WHO TN 11).

Targets (WHO Guidelines for drinking-water quality, 4th edition with the third addendum, 2026):

• At least 0.5 mg/L free chlorine after 30 minutes at pH below 8 in a piped supply, and at least 0.2 mg/L at every point of delivery.
• 0.2–0.5 mg/L where people collect water (WHO TN 11), and 0.2 mg/L still present in stored household water.
• No more than 5 mg/L, WHO's health-based value. Some people taste chlorine from 0.3 mg/L, and strong-tasting water discourages people from drinking it (WHO TN 11).

Test with DPD: a tablet colours the sample red, read against the comparator's colour scale. Test daily while a system is new, then at least weekly, at the nearest and the farthest taps (WHO TN 11). Without tablets, a faint chlorine smell after 30 minutes shows that some free chlorine remains, but not how much.

Opaque, clean, sealed and labelled, in a cool shaded place out of children's reach, never in metal (WATA). The label gives the contents, the word poison, the strength in g/L and the date made.

• Decay: hypochlorite breaks down faster when warm, into chlorate and chlorite; WHO limits chlorate in drinking water to 0.7 mg/L. Kept dark at room temperature, the Berkeley solution lost under 5 % in 34 days; WATA re-tests any batch older than 24 hours. Titrate an old batch before dosing from it, and use each within a month, as WHO advises for bought hypochlorite.
• Never top up old solution with new: it speeds chlorate formation (WHO).
• Never acidify a drinking-water stock: at pH 6 the Berkeley solution lost 16–20 % in 34 days, and lower still it releases chlorine gas.
• Clean the cell after each run by rinsing it with water; only after that rinse, soak off white scale overnight in half vinegar, half water (WATA).

• It does not kill Cryptosporidium oocysts (WHO); where they are the risk, filter or boil as well.
• It does not remove chemical contaminants (WHO TN 11); the water guide covers them.
• It cannot protect an intermittent pipe supply: when the pressure drops, dirty water enters at every leak, and no acceptable residual can deal with it. Treat such water at the point of use (WHO TN 11).