Sodium and potassium reacting with water, in the absence of air

As the piece of metal skitters across the surface of the water in a beaker and — particularly in the case of potassium — it appears to catch fire, it is not obvious that the explanation for both phenomena lies in the production of hydrogen gas. There is a simple and safe way to demonstrate the reaction. This can be accomplished by placing a layer of paraffin oil or turpentine above a layer of water in a test tube and then dropping a small piece of sodium or potassium into the upper layer. It sinks down, and on reacting with the water (Fig. 1 – 1) it releases bubbles which themselves carry the metal up to the surface (Fig. 1 – 2). Once they have been released, the metal sinks and the process is repeated. The test tube can be loosely closed with a stopper, and the gas ignited after a couple of minutes. Excess metal remaining at the end of the reaction can be destroyed by adding ethanol or propanol. Students themselves can carry out the reaction, provided that they are warned neither to shake the test tube nor stopper it tightly. The procedure cannot be carried out with lithium (why not?).

Diagram showing two Erlenmeyer flasks with test tubes and equilibrium arrows.

Fig. 1 A layer of paraffin oil above water in a test tube is shown with a small piece of sodium.

A graphic of a graduated cylinder filled with water and inverted into a basin of water.

Fig. 2 A graduated cylinder inverted in water.

A graphic of water-filled graduated cylinder inverted in a basin of water with a tweezers holding a small sphere in the water below the cylinder.

Fig. 3 A small piece of sodium is added to the water under the graduated cylinder.

However, there is another way with which to demonstrate the reaction.

A large, preferably glass, bowl or evaporating dish is three-quarters filled with water and a 100 mL glass measuring cylinder is filled with water— the mouth covered by a piece of a polythene bag* so that it can be inverted and placed in the bowl (Fig. 2). Note the sketch is not to scale. The polythene is removed, and the cylinder remains filled with water. The cylinder should be held by a clamp so that its mouth is just below the surface — for simplicity I have not drawn the clamp.

All safety precautions must be in place before starting. A small piece of sodium — a cube of about 4x4x4 mm — is cut on a Petri dish, and then re-immersed in paraffin oil. This is important, as the oil slows down the initial reaction with the water. Now comes the tricky bit. Using tweezers, the sodium has to be rapidly placed underneath the mouth of the cylinder as shown in Fig. 3. Teachers must practice this before demonstrating it! It is not a disaster if the sodium escapes and reacts on the surface of the water; another piece can be inserted, but first make sure that the tweezers are quite dry! Using tweezers or forceps that have an angled tip might help facilitate this maneuver.

The sodium reacts and floats on the surface; gas is released and forces the water level down. More sodium can be added. The sodium forms a ball, a sign that it has melted (exothermic reaction).

After all the sodium has reacted the gas can be tested by igniting it, using a long match or splint. It is best to hold the cylinder nearly horizontal, to enable the air to mix with the hydrogen. If the cylinder is almost full of hydrogen it burns quietly. If less sodium is used, then on lifting the cylinder out of the water, air goes in, and on lighting the gas one gets a very satisfying "bark" (not an explosion), along with a characteristically orange flame.

If potassium is used, the teacher has to be that much more skillful! On igniting the hydrogen, one gets a light purple flame.

I devised this demonstration some years ago but recently saw that a French chemistry teacher, M. Michel Bultingaire (previously of the Lycée Saint Exupéry, Fameck, France), also does it in much the same way.

A graphic of a graduated cylinder shown ¾ full of water inverted into a basic of water. On the top of the water in the cylinder is a small sphere, representing a piece of sodium metal.

Fig. 4 The collection of gas from a small piece of sodium reacts with water.

The first time I do the demonstration, I prefer not to add phenolphthalein, so that there will be more scope for inquiry. When I do add it, I add it to all the water before placing it in the cylinder, so that on reaction the colour change is associated with the reaction (in this video this is not the case).

The experiment can also be carried out quantitatively by removing the oil from the piece of sodium with petrol ether. The dry sodium can be rapidly weighed, then re-immersed in oil. The volume of hydrogen released is measured.

Here are my own results:

Three pieces of Na were used with a total mass 0.441 g — 0.0192 moles Na. Volume of hydrogen released — using a
250 mL cylinder as I wanted greater precision — was 234 mL.

To find the number of moles of hydrogen, I compared the mass of the 80%:20% butane:propane mixture — in our Campingaz C 206 GLS cartridges — which would have the same volume under the same conditions. This way I obviated the need to measure temperature and pressure, and to take into account the vapour pressure of water. I just had to unscrew the burner head above the tap and attach a flexible tube to the outlet and bubble gas into the same inverted measuring cylinder.

Mass of gas giving 234 mL was measured as 0.530 g — the mass difference of the butane:propane cartridge before and after.

Since the molecular weight of the 80%:20% butane:propane is calculated at 55.2 g mol-1, it was determined that there was 0.00960 moles of mixture in the 234 mL.

Hence 234 mL corresponds to 0.00960 moles of any gas (Avogadro's hypothesis).
So 0.0192 moles Na liberate 0.00960 moles H2

→ 2 moles Na liberate 1 mole H2


In this reaction, increase in temperature is predicted and therefore affects the volume of the gas collected. This temperature effect was considered negligible. This is an additional opportunity for inquiry. Have students predict the temperature increase and determine the effect on the volume of hydrogen.

From the website,,

2Na(s) + 2H2O(l) → 2NaOH(aq) + H2(g)   ΔH = -368.5 kJ

If 2 moles Na release 368.5 kJ, we can predict that 0.0192 moles release 3.538 kJ = 3538 J. We know that 4.18 J will heat 1 g of water by 1 °C. Therefore, there is enough heat to raise 100 g of water by about 8.5 °C. It can be assumed that most of the heat released in the reaction is absorbed by the water, which is a much better conductor than hydrogen gas — not to mention that the reaction takes place with the water. It should be noted that no noticeable temperature change was felt to the hand. Of course the initial temperature of the water near the reacting sodium is higher, but the heat will be rapidly dissipated to the rest of the water — given there is much more than 100 g. Heat would also be lost to the glass of the measuring cylinder. But let us assume that the temperature of the hydrogen goes up by about 8 °C — this would mean an error of about 306/298 = 3% compared with STP.

*Editor’s Note: A piece of filter paper (or any paper works even better. After the cylinder is inverted and placed below the water in the bowl, the filter paper will float off by itself. A proofreader gave this suggestion — it is one of his favorite tricks!