![Thumbnail: [Volume 57, 1927 page 900]](/journals/rsnz/images/rsnz_57/thumbnails/rsnz_57_00_1008_0900_ac_02.gif)
A Problem in Boiler Corrosion.
[Read before the Philosophical Institute of Canterbury, 2nd December, 1925; received by Editor, 31st December, 1925; issued separately, 8th March, 1927].
A Large factory which derived its boiler water from a neighbouring creek flowing through granite country was considerably hampered by excessive corrosion of the boiler tubes. The water, although in itself apparently exceptionally good, was treated with a soda ash softener, yet the formation of scale persisted. An examination of a sample of scale revealed, even to the naked eye, bands of hard, dark material alternating with bands of almost pure iron oxide. It appeared that, whilst the salts were accumulating in the boiler prior to crystallization setting in, active corrosion had been going on—a not uncommon phenomenon—especially with a silicate scale.
| Silica | 11.3% |
|---|---|
| Iron oxides | 55.5 |
| Calcium oxide | 10.6 |
| Magnesium oxide | 4.7 |
| Sulphuric anhydride | 12.6 |
| Chlorine | .2 |
| Loss on ignition | 5.1 |
| 100.0 |
In order to obtain a truer conception of the nature of this scale, the table has been recalculated with the elimination of the iron, which, after all, has entered the scale from the boiler, and not from the water.
| Magnesium chloride | 0.7% |
|---|---|
| Magnesium silicate | 25.7 |
| Calcium silicate | 9.1 |
| Calcium sulphate | 47.9 |
| Free Silica | 5.2 |
| Loss on ignition | 11.4 |
| 100.0 |
The deposit is thus revealed to be essentially a silicate-gypsum scale. The presence of the latter calls for little comment, for its relatively low solubility will cause its precipitation during evaporation whenever a feedwater contains soluble sulphates and soluble calcium salts, but such a scale is in no way corrosive. On the other hand, silicate deposits are notoriously corrosive. To quote from a standard book on the subject (Paul, Boiler Chemistry, p. 86) “a silicate scale one-sixteenth of an inch thick will cause tubes to fail in a few days.” This arises, of course, from the action of the silicic acid in liberating its equivalent of mineral acids from the salts present. The formation
![Thumbnail: [Volume 57, 1927 page 901]](/journals/rsnz/images/rsnz_57/thumbnails/rsnz_57_00_1009_0901_ac_02.gif)
of this silicate scale postulates the presence in the boiler water of dissolved calcium and magnesium salts as well as of silicic acid.
The former of these may well have found its way into the boiler through leaky condenser tubes as the condenser was cooled with sea-water, but the comparative absence of silica in sea-water, as well as of sodium chloride in the scale, rules out such a possibility. The feed water itself becomes an object of suspicion.
[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]
| Permanent Hardness | 2 grains per Imp. Gallon |
| Temporary Hardness | less than 1 " " " " |
| Reaction to litmus | neutral |
On this analysis the water would pass as an eminently satisfactory boiler water, but it was deemed advisable to carry out a complete gravimitric analysis. The results were as follows:—
[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]
| Total Residue | 6.3 grains per Imp. Gallon. |
| Silica | 2.52 |
| Ferrous sulphate | 0.39 |
| Calcium sulphate | 0.03 |
| Calcium chloride | 0.19 |
| Magnesium sulphate | 0.71 |
| Organic and volatile matter | 1.47 |
| Nitrates | trace |
| Total | 6.31 grains per Imp. Gallon. |
The amount of silica present is roughly ten times that present in a normal feedwater, and this is in itself sufficient to account for the excessive corrosion experienced in the boiler. As evaporation has proceeded, the steady enrichment of silica and of the salts of calcium and magnesium in the boiler has led to the precipitation of the difficultly soluble and highly injurious silicates of these metals, accompanied by the relatively insoluble gypsum.
Treatment of the Feed-Water.—The unsatisfactory result given by the use of soda-ash may be attributed to the failure of this substance to remove from solution small quantities of magnesium salts. This can generally be more easily effected by means of an alkali, but the use of a lime-soda mixture would naturally be fraught with risk as the usual excess of the softener would enchance the possibility of precipitating calcium silicate. It was, therefore, decided to use a caustic soda-soda ash mixture. This would serve the double purpose of precipitating a good deal of the calcium and magnesium salts present and also impart sufficient alkalinity to the water to inhibit the corrosive action of the silicic acid. The formula of the softener calculated from the above feed water analysis is as follows:—
Soda ash (calculated as 100% pure) 2½ lbs. per 10,000 gallons.
Caustic soda (calculated as 100% pure and free of sulphides) 1½ lbs. per 10,000 gallons.
This formula has been in use in the factory for over a year and it is gratifying to record that the corrosion of the boiler has been entirely eliminated.