Passive Iron.

After dipping iron into strong nitric acid, weak acids do not attack it, and galvanic tests show that it has grown electronegative. It was believed that this change was due to a coating of oxide of iron or to occluded gases, or to an allotropic change. The various views and experiments are discussed by J. S. de Benneville, 'I. and S. I.,' 1897, ii. p. 40.

IV. Galvanic Action of Zinc.-If slabs of zinc are fitted in such a manner as to be in metallic contact with the iron (see fig. 100), a current passes from zinc to water, to iron, and back, zinc being dissolved and hydrogen evolved on the iron surface.

If the zinc is connected to the iron plates by means of a copper wire, there will be two circuits, viz. one from the zinc to water, to copper, and one from iron to water, to copper; the one current wastes the zinc, the other the iron. If, therefore, the zinc does protect the iron from corrosion, it must be due to some other cause than galvanic action. Copper pipes should produce a current from iron to water, to copper, and the iron should grow electronegative.

FIG. 100

The intensities of electric currents which might be expected in a boiler, if no polarisation took place, are very difficult to estimate, as the laws which govern the flow of electricity in conductors of large dimension have not yet been clearly formulated. The resistance of a conductor is proportional to its length, and inversely proportional to its sectional area. For a cubic centimeter and per cubic inch we have the following resistances measured in ohms :-

This table shows that the resistance in the metal may be neglected, but it is also evident that, except when the electrodes are almost in contact, the fluid resistance will reduce the intensity of the current excessively.

W

S D

Fe

FIG. 101

Take, for instance, a strip of iron 1 in. wide (fig. 101), from which part of the scale, S, has been removed, exposing the iron at Fe to the salt water, W; then a current will flow from Fe to W, to S, and back. Restricting the observation to the small zone of the diameter D, the available section of the conductor is-say, 3 in.-and its mean length from Fe to S is 32 in. Assuming that the water contains 5·5%

G

salt, the resistance of this small element would be (35.9 × 1·507=) 54.2 ohms + the resistance of the scale and iron, say 60 ohms.

Assuming also that the electromotive force of iron scale in salt water is the same as for copper, the electromotive force of such an element would be 276, which would give a current of 276=0046, which would be capable of corroding 0022 lb. per day, or at the rate of about in. in 60 days. At one inch distance from the edge of the scale (see fig. 101) the action would be about 3% of the above, showing how rapidly it diminishes with the distance.

Pitting has been explained in this way, and this view is to a certain extent corroborated by the experiments of Mr. Parker (I. and S. I.,' 1881, i. p. 39), who found that the average depth of corrosion in plates. from which the scale had been only partly removed amounted to 20% more than that of the bright discs. But then their average losses were not sufficiently uniform to permit of strong conclusions being based on them.

The above calculations will also make it clear that the necessity for isolating various samples while subjected to experimental corrosive influences does not exist, and therefore the very exhaustive experiments of the Admiralty Boiler Committee are worthy of more confidence than is generally placed in them.

The Galvanic Action of Zinc in Boilers.--If galvanic currents exist, the studs by which the slabs are secured ought to remain bright, at least within one-sixteenth of an inch of the metal; but it is well known to those who take much trouble in keeping up a good metallic connection that corrosion takes place even at the contact surfaces, which have to be scraped or filed after every voyage. The current produced by zinc, iron, and sea water is about twice as strong as that of scale and iron, but as it has no effect on its immediate surrounding, it is hardly probable that it will have much influence on the more distant parts, which are separated from it by several feet of salt

water.

In spite of these estimates, it can be shown experimentally that zinc and iron (wire) if placed together in aërated water, but without touching each other, will both corrode; if the iron wire be attached to zinc, only the zinc corrodes. This, however, is only true when certain salts are present in the water, and that may explain the varied experiences as regards zinc. Searching experiments on this subject are now being made for the Manchester Steam Users' Association.

Composition of Boiler Plates and Corrosion.-A subject to which sufficient attention has not yet been paid is the influence of the composition of iron and steel on their behaviour in the presence of corrosive fluids. It is believed that manganese increases corrosion, while there can be no question that nickel reduces it. Carbon behaves very strangely, and it would seem as if in one form-viz. as found in annealed steel-it increased corrosion, whereas in hardened steel it reduces it. However there is little scope for improvement, because its percentage is fixed by other considerations.

A few remarks on some of the results of the experiments carried out by the Admiralty Boiler Committee will not be out of place here. As in the case of Mr. Parker's experiments, small samples of steel and iron plates were exposed in various merchant and Government

steamers, and their losses ascertained. The unit of comparison is 1 grain per square foot per 10 days, equal to about in. per annum.

It was found that when using jet condensers the corrosion was very erratic, ranging from 35·5 to 281-3 grains; but there were only four of these experiments.

Influence of Condenser Tubes.-In four cases copper-tubed surface condensers were used, and the corrosion was slight. There were 25 cases of brass condenser tubes. The corrosion in their boilers was slight, except in two cases, with which respectively colza and Rangoon oil were used as lubricants in the cylinders. With these two exceptions the maximum and minimum losses were 128.8 and 16.8 grains In only two of these cases was zinc used, and its influence was not marked. The cylinder lubricants were nearly always mineral oils. There were seventeen cases of tinned condenser tubes. Their influence was very marked, the corrosion being far severer than under the other conditions. Mineral oil was used in four of these ships, and the corrosion varied from 2046 to 362-2 grains. In these cases the use of zinc alone does not seem to do much good, the losses in three cases being 129-6, 215-2, and 503-2 grains. On the other hand, when chalk, Portland cement, and, strange to say, tallow were placed in the boilers whose condenser tubes were tinned, the losses were much reduced, two cases being as low as 327 and 397 grains. In two boilers soda was used. Its action is not a decided one, for the losses were 306 and 376-3 grains.

Consumption of Zinc. In the reports of the Boiler Committee information will also be found as to the loss of zinc when fitted in boilers. In the case of H.M.S. Crocodile' 425 to 630 lbs. of rolled zinc were consumed per boiler per annum, while in H.M.S.Serapis the quantity was 236 to 420 lbs. This is only the actual loss as found by weighing the zinc before and after each voyage. J. Morris, N. A.,' 1882, vol. xxiii. p. 151, states that one slab 12 x 12 x 13 inch was completely consumed in 6 weeks. The use of zinc as an anti-incrustant is first mentioned in 1875.

·

In addition to the general remarks about preventing corrosion (p. 68) the following have suggested themselves. Strong chemicals, either alkalies or acids, should not be used. Certain salts increase or reduce the speed with which iron is attacked. Zinc salts appear to reduce the action, tin salts to increase it. Air, being the probable cause of pitting, should be carefully excluded. The existence of galvanic currents in boilers has not been proved.

84

CHAPTER IV

FUELS AND COMBUSTION

Combustion. When two substances, having a chemical affinity for each other, are brought together under favourable conditions, they combine, forming a new compound. During this combination heat is evolved, which, if the chemical affinity is sufficiently strong, will raise the temperature of the substances to such a height that they grow luminous. This is called combustion. The term is not restricted to the burning of coal, wood, oil, or gas in air, but is applicable to the burning of these or other substances in any other gas, such as chlorine or hydrogen, and is sometimes used to denote a similar process in which only solids or only fluids combine. It may be illustrated by igniting gunpowder, by dropping potassium metal into mercury, or by heating a mixture of iron filings and sulphur. In any of these cases luminous heat is generated.

Slow Combustion is a term which is commonly restricted to express decay of organic matter, but the process is met with on a large scale in every coal mine, where it is the cause of the increased temperature of the ventilating air as it leaves the shaft. In our lungs a process of slow combustion is continually proceeding, and in the tarnishing of metals the same action is manifested. Thus magnesium wire if lighted burns, if exposed to atmospheric influences it tarnishes. In both cases oxide of magnesium has been formed.

When burning coal in a boiler furnace, the main object is to obtain heat; slow and imperfect combustion have, therefore, to be prevented. The one takes place generally, but not always, if the air is in excess and too cold, while the other is caused by an insufficient supply.

It

Heat, as is well known, is not the same thing as temperature. is a quantity and not a condition. It is measured in units of heat (called calories), of which each one will raise the temperature of one pound of water one degree Fahrenheit. It is often more convenient to use another measure, viz. the evaporative unit. This is equal to the amount of heat which will evaporate one pound of boiling water from and at 212° F. One evaporative unit equals 966 calories, or heat units, or thermal units (J. K. Cotterill, London, 1878, p. 314); it is also equal to 22.63 horse-power during one minute, while one calorie per second equals 1·4054 horse-power.

Heats of Combustion.-Exhaustive experiments have been made on numerous elements and chemical compounds to determine the

amount of heat generated during the processes of combustion, chemical combination, absorption, and solution. Most of these results are contained in M. M. P. Muir's 'The Elements of Thermal Chemistry,' 1885. In this book the kilogram and the degree centigrade are employed, and in order to reduce the values to English measure they have to be multiplied by, or divided by 5363, to convert them respectively into thermal or into evaporative units.

For convenience in calculating the heats of complicated chemical processes the values in that and similar books are not stated per unit of weight of each element or substance, but per atomic weight. Thus the value for pure carbon is given for 12 kils. and not for 1 kil. of carbon, because the atomic weight of this element is 12.

The following table contains a few of the most important determinations. See also W. Ostwald, 1887.

=

Partial Combustion.-The heat generated by the combustion of 28 lbs. of carbonic oxide (CO 12 C + 16 O) amounts to 126.9 English evaporative units. If this heat be added to the 539, which are generated by imperfectly burning 12 lbs. of carbon, 180-8 units are obtained, which are exactly equal to the heat evolved by burning these 12 lbs. of carbon to carbonic acid in one operation. This shows that, as regards the final result, it does not matter whether carbon is burnt in one or in two stages, or, in other words, whether it is completely burnt at once, or is first converted into carbonic oxide gas and then burnt. Another feature of interest is that the partial burning of 12 carbon, to form 28 carbonic oxide, only produces 42.5% of the heat which will be evolved when 28 carbonic oxide are burnt to carbonic acid. The remaining 57.5% might be looked upon as the latent heat of evaporation of 12 lbs. of carbon. This would imply that it requires four and a half times as much heat to evaporate