A Preparation of Tripotassium and Trisodium phosphate

These experiments were an attempt to make use of a description of an industrial process¹ to improve the quality of these two reagents when made in an amateur setting.

The aim was to produce 100g of K₃PO₄ and 100g of Na₃PO₄.

If each can be crystallised as the .12H₂O hydrate, that equates to 202g and 232g of crude product respectively.

As mentioned in the reference, a two-step approach is used to eliminate impurities in the phosphoric acid required for the reaction, which is good if you have low grade green acid, such as mine.



Tripotassium Phosphate

First the acid was partially neutralised using K₂CO₃ to obtain K₂HPO₄ and precipitate iron/other impurities present in the acid.

H₃PO₄ + K₂CO₃ => K₂HPO₄ + H₂O + CO₂

Stoichiometrically, to achieve a 100g final yield, 65g of K₂CO₃ is first required to react with 67g of 72w% H₃PO₄, representing a 5% acid excess (by weight).

First the K₂CO₃ was placed into a 250ml beaker, then 100ml distilled water is added along with a stir bar.

The mixture was stirred and the phosphoric acid added dropwise, ensuring that the fizzing did not go out of control.
Heat was generated, although not enough to reach boiling point.

Once all the acid had been added, heat was applied carefully to drive out all of the CO₂ produced.
Care must be taken to control the fizzing, or the liquid WILL escape from the beaker !



The resulting liquid was filtered directly into a KOH solution made by dissolving 28g of KOH in 100ml of water, which is a 5% KOH excess (by weight).



This second reaction substitutes the final hydrogen, creating the desired product:

K₂HPO₄ + KOH => K₃PO₄ + H₂O

No precipitate/solid contamination was seen, so a second filtering was not done.

The resulting liquid was heated in an evaporating dish to dryness.
This was a mistake, as the material bubbles up and out of the dish near the end, also it sticks to the dish too strongly to remove.
Better to evaporate it down then finish the drying in an oven or dessicator.

A small amount of water was added to enable the mass to be removed from the dish.



99.5g appears to be an almost perfect yield, however some contamination was introduced when the material bubbled out of the evaporating dish, and i am not totally convinced that soluble contaminants are not contributing to the weight.

The two-step process did eliminate all of the green colour in the product, which is an improvement over the green product made by simply neutralising K₃PO₄ with KOH.

Trisodium Phosphate

The same process was followed as above with the sodium reagents substituted for the potassium ones : 90g of Na₂CO₃, 121g 72% H₃PO₄, 26g NaOH.

The Na₂CO₃ was obtained by prolonged heating of 142g NaHCO₃. Directly using NaHCO₃ would have increased the fizzing a lot.

A major difference with Na₃PO₄ is that the precipitates from the first reaction are goopy solids.



Another difference is that it will readily crystallise, as was seen on filtering. This was fixed by gently heating the funnel with a blowtorch.



Once filtered, the liquid quickly crystallised, and was put out to dry on paper.



The final yield was 151.82g of probably dodecahydrate product, representing a yield of 65% based on expectation.

These crystals are predominantly white, qualitatively confirming that the two step process described in the paper has a useful advantage when using highly impure reagents to make either of the phosphates.

¹ Attachment: snell1931.pdf (755kB)
This file has been downloaded 209 times