Translator’s Note: The Aquidaban was a vessel of the Brazilian Navy somewhat similar to our old Texas, having two echelon 8-inch turrets, the forward one being on the port side and the after one on the starboard side. In addition she carried a battery of eight or ten 4.7 guns.
In 1906, carrying on board a commission of high officers, several admirals and captains, she steamed from Rio to Ilha Grande Bay, a distance of about seventy miles. It was the purpose of the commission to determine the feasibility of establishing a naval base in that bay. Opposed to this idea there was a strong political faction both within the navy and without.
The vessel anchored and at about eight o’clock at night she was destroyed by an internal explosion of the after eight-inch magazine. While we are not technically interested in the history of the affair, there is, however, a point brought out by the investigation which is of technical interest, and that is the remarkable stability and homogeneity of the cordite after being immersed in sea water for fourteen years. With a powder of similar insensibility to sea water it would be quite practicable to flood a magazine in time of danger and afterwards to pump out and go into action or renew the action again with the same powder. Can we do this with our smokeless powder? This article is translated from the Boletim do Club Naval of Brazil.
In connection with the previously made study of the cordite taken from the submerged hull of the Aquidaban, we return to the subject in order to mention new data, some of which it would not be proper to omit.
By the letter from Sir A. G. Hadcock, dated July 26, 1921, we verified that the cordite which was supplied to the Aquidaban had been originally manufactured for this battleship in 1898, and was from the factory of the Chilworth Smokeless Powder Company.
At the time the explosion occurred the powder had been kept in the magazine about eight years under conditions upon the extreme precariousness of which we have already commented.
The examination of the samples obtained showed that the chemical stability was not greatly altered as might have been expected of a powder twenty-two years old.
Admiral Severiano de Castilho, eminent naval engineer, sent to the firm of Armstrong of Newcastle a small portion of the cordite which had then been taken out by divers from the magazines of the Aquidaban. Sir A. G. Hadcock made an analysis of these samples, which were from the propellants designated respectively for the 8-inch and 4.7-inch guns.
The illustrious English technicist gives us the interesting results °f his inquiry, which we reproduce with much satisfaction as they are in complete accord with those which we had already arrived at. The confirmation follows:
Data given by the cordite at the time of its receipt in 1898.
Lot 396 for 8-inch gun.
Nitrocellulose, soluble 2.62%
Nitrocellulose, insoluble 36.10
Nitroglycerine 56.69
Vaseline 4.59
Total 100.00
Loss of weight by heating 1 hr. at 150° F. = 0.7% Density 1.565.
Lot 402 for 4.7-inch gun.
Nitrocellulose, soluble 3.47%
Nitrocellulose, insoluble 35.10
Nitroglycerine 56.86
Vaseline 4.57
Total 100.00
Loss of weight by heating 1 hr. at 150° F. = 1.0% Density 1.572. In 1921 the samples taken from the submerged hull gave, according to Sir Hadcock, the following results:
Cordite for 8-inch gun.
Nitrocellulose, soluble 5.27%
Nitrocellulose, insoluble 36.22
Nitroglycerine 53.40
Vaseline 5.11
Total 100.00
Humidity 0.22 per cent.
Test of Abel at 180° F. 9 minutes.
Density 1.567.
Point of ignition at 336° F.
Cordite for 4.7-inch gun.
Nitrocellulose, soluble 6.66%
Nitrocellulose, insoluble 40.21
Nitroglycerine 47.62
Vaseline 5.51
Total 100.00
Humidity 0.25 per cent.
Test of Abel at 180° F. 9 minutes.
Point of ignition at 334.4° F.
Noting these results, Sir A. G. Hadcock calls attention to the fact that both the samples examined stood the test of stability for nine minutes in spite of the advanced age of the powder.
We (at the time of our analysis) were only able to obtain sample of 4.7-inch ammunition. In connection with the powder of this ammunition however, the agreement of the greater part of the measurements which we found to those referred to us would extend also to the results of chemical analysis of different samples. Thus, for example, proportioning the ingredients by approved process; at the end of 1920 we got for one of the samples in question:
Nitrocellulose, soluble 6.8%
Nitrocellulose, insoluble 41.1
Nitroglycerine 46.8
Vaseline 5.3
Total 100.00
A result which practically equals that found by the director of the firm of Armstrong and which shows for the powder a loss of nitroglycerine of about 11 per cent in weight, a loss that that technical authority attributed to evaporation of the liquid and considered perfectly normal data for the existing conditions.
It is clear that we cannot make an exact analysis of samples whose state of damage was evident, as was proved by the tests for stability. We are not interested with the approximation of the second place decimal in the percentages of the constituents, in view of the irregularity of the measurements of the various samples.
It was proper that on three different trials of the best samples, we should investigate also for the possible presence of per- chloride of mercury which would, with the tiny dose of 1/1000 milligram, annul completely the significance of the Abel test.
Adopting the current technical German method of Kast (Post et Neuman, Analyse Chem. p 347), a search for mercury in .the nitro-cellulose extracted from the powder gave negative result in the three tests, as the pure gold leaf impregnated with the reagent presented no visible indication of amalgam, it not having been possible to form iodide of mercury.
Simultaneously we made a verification of this result by the employment of the spectroscope, as is preferred by English technicists.
Having communicated to Sir A. G. Hadcock this first evidence of the non-existence of mercury in the cordite, he replied by letter of 26 August, 1921, his entire confidence in the confirmation of such results.
We cannot then support here as conforming to our other study of the subject the hypothesis that there was an explosion of the magazine of the after turret of the Aquidaban provoked by an initial explosion of the boilers.
The chief of divers, Mr. Ogapito, who repeatedly penetrated the submerged hull, declared to us that the damages in the boiler compartment, owing to there having actually been an explosion of the magazine, were greater on the side of the magazine than on the other.
Lieutenant Ormando Roxo and First Lieutenant Manoel Barbosa de Santa Anna, witnesses present at the accident, did not see any escape of steam at the beginning of the explosion; ref erring to the latter, however, as a peculiar noise, like a rapidly rising shriek followed by strong explosions. This would explain itself easily by the burning of a charge of powder propagating itself to others, from which the liberation and escape of gases produced the noise mentioned. Then the great and rapid rising of temperature could have caused the successive explosions of the remaining charges.
From all of which it appears permissible for us to form the following conclusions:
I. The cordite first taken from the 4.7-inch ammunition from the submerged hull of the Aquidaban in October, 1920, showed some alteration as regards the percentage of its constituents, and as regards the homogeneity and consistency of the colloid; it was not, however, all in a bad condition of chemical stability, having strands which still presented results satisfactory in heat tests, in spite of the age of the powder, which according to Hadcock was more than twenty-two years; and, in spite of the high temperature to which it was frequently exposed before submersion.
II. It is very probable that the fact of having been submerged in the sea has contributed to moderate the progress of the damage of the powder, as during that time it has been subjected to temperatures very much lower and less variable than those in the magazines of the Aquidaban.
III. The fact of finding still more strands of cordite in a state of conservation above that which could be foreseen, does not authorize the dismissal of “spontaneous combustion,” as the probable cause of the disaster.
IV. The solution of the probable cause of the explosion is, we think, an indeterminable problem. It should not be left without impressing the fact, that the explosion was produced precisely where the conditions of temperature for it were exceptionally favorable, because the magazine of the after turret, which exploded, was interposed between the boilers and the engines. Furthermore, it was adjacent to a group of dynamos, in which compartment the thermometer registered at times more than 140° F.
V. The conditions in which the cordite was stored would have provoked a rapid decomposition in any smokeless powder, single or double base, and the risk of spontaneous combustion would have been practically the same for any.