1840.] 



THE CIVIL ENGINEER AND ARCHITECT'S JOURNAL. 



425 



rate set of tables of results ; but these laborious investigations being yet in 

 progress, the author directs his special attention to so much only of the sub- 

 ject as may be necessary for their elucidation, divesting his remarks as much 

 as possible' of a purely chemical character, and contiuing them to those prac- 

 tical conclusions which are of immediate use and importance of the engineer. 



The tables of results are altogether twelve in number. The first five con- 

 tain the data and results of the chemical or corroding action of sea and fresh 

 water on cast and wrought iron under five several conditions, during a period 

 of a year and ten months ; and these five series of experiments are so co- 

 ordinate with each other as to form one connected and comparable whole, 

 whence the relative rates and absolute amounts of corrosion of cast and 

 wrought iron — by, 1. clear sea water, 2. foul sea water, 3. clear sea water at 

 temperature 115° F., 4. foul river water, and 5. clear river water — may be 

 ascertained. The corrosive action of water and au' combined produces on the 

 surface of csst or wrought iron a state of rust possessing one of the five fol- 

 lowing characteristics — 1. Uniform, 2. uniform with plumbago, 3. local pitted, 

 4. local pitted, 5. tubular — or of two or more of these characteristic condi- 

 tions in combination ; these facts for 82 different specimens of British and 

 Irish cast iron- -together with their original external characters, mode in 

 which they were cast, specific gi'avity, dimension and weight before and after 

 immersion, loss of weight per square inch of surface, this loss referred to a 

 standard bar, and the weight of water absorbed for clear sea water — compose 

 Table I. The four subsequent tables contain similar results for specimens of 

 iron immersed under the other four conditions mentioned above. These five 

 tables contain also the results of the corrosion of certain cast iron protected 

 by either of ten several paints or varnishes, the results of which are compara- 

 ble with those for the unprotected iron. Table VI. exhibits the general com- 

 parison of the results set forth in the preceding tables for specimens of iron 

 one inch thick, and reduced to one common or equal period of immersion. 

 Table VII. shows the average loss of all varieties of cast iron experimented 

 on per square inch of surface. Table VIII. the average calculated amount of 

 coiTosion (assumed uniform) of various specimens of east and wrought iron 

 per superficial foot of surface at the end of one century. From these tables 

 it appears, that the metallic destruction or con-osion of the iron is a maximum 

 in clear sea water of the temperature of 115' F. — that it is nearly as great in 

 foul sea water — and a minimum in clear fresh river water. 



Iron under certain circumstances is subject to a peculiar increase of corro- 

 sive action — as, for instance, cast iron piling at the mouth of tidal rivers— 

 from the following cause. The salt water being of greater density than the 

 fresh, forms at certain times of tide an under current, while the upper or 

 surface water is fresh ; these two strata of different constitutions coming in 

 contact with the metal, a voltaic pile of one solid and two fluid elements is 

 formed ; one portion of the metal will be in a positive state of electrical ac- 

 tion with respect to the other, and the corrosive action on the former portion 

 is augmented. The lower end of an iron pile, for instance, under the circum- 

 stances just mentioned, will be positive with respect to the other, and the 

 corrosion of the lower part will be augmented by the negative state of the 

 iipper portion, while the upper will be iVse^ preserved in the same proportion. 

 From this theoretical view may be deduced the important practical conclu- 

 sion, that the lower parts of all castings subject to this increased action should 

 have increased scantling. 



The increased corrosive action of Jbul sea water may be referred to the 

 quantity of hydrosulphuric acid disengaged from putrifying animal matter in 

 the mud, converting the hydrated oxides and carbonate of iron into various 

 sulphurets, which again are rapidly oxidized further under certain conditions, 

 and becoming sulphates are washed away. Hence the rai)id decay of iron in 

 the sewage of large cities, and of the bolts of marine engines exposed to the 

 bilge water. The corrosive action being least in fresh water may be partly 

 referred to this being a worse voltaic conducting fluid than salt water. 



It appears also that wrought iron sufl'ers the greatest loss by corrosion in 

 hot sea water ; which fact has led the author to inquiries, with reference to 

 marine boilers, at what point of concentration of the salt water, whether 

 when most dilute, after the common salt has begun to deposit, or at a farther 

 stage of concentration, the corrosive action on wrought iron is the greatest, 

 and he points out the important practical use which can be made of this in- 

 formation. It appears also, that the removal of the exterior skin of a casting 

 greatly increases the corrosive action of salt water and its combined air, so 

 that the index of corrosion under these circumstances is not much less than 

 that of wrought iron, and in clear river water is greater. 



It farther appears, that chilled cast iron corrodes faster than the same sort 

 of cast iron cast in green sand, and that the size, scantling, and perhaps form 

 of a casting, are elements in the rate of its corrosion in water. The explana- 

 tion of these facts is to be found in the want of homogenity of substance, 

 and the consequent formation of numerous voltaic couples, by whose action 

 the corrosion is promoted. It is also observable that the corroded sm-face of 

 all these chilled specimens is tubular. 



It appears also that, in castings of equal weight, those of massive scantling 

 have proportionately greater durability than those of attenuated ribs and 

 feathers. Hence appears also the great advantage of having all castings, par- 

 ticularly those intended to be submerged, cooled in the sand, so as to insure 

 the greatest possible uniformity of texture. The principles now stated afford 

 an explanation of the fact often observed, that the back ribs of cast iron sheet 

 piling decay much faster than the faces of the piles. It is also probable that 

 castings in dry sand and loam will, for these reasons, be more durable than 

 those cast in green sand. The general residt of all these experiments gives 



a preference to the Welsh cast iron for aquatic purposes, and to those which 

 possess closeness of grain. Generally, the more homogeneous, the denser 

 and closer grained, and the less graphytic, the smaller is the index of corro- 

 sion for any given specimen or make of cast iron. 



The author next proceeds to the important question of the protection 

 afforded by paints and varnishes. White lead perishes at once in foul water, 

 both fresh and salt ; and caoutchouc dissolved in petroleum appears the most 

 durable in hot water, and asphaltum varnish or boiled coal tar laid on while 

 the iron is hot under all circumstances. The zinc paint, which is now so 

 much noticed as an article of commerce, the author has analyzed, and states 

 its composition as — 



Sulphuret lead 



Oxide zinc 



Metalhc zinc 



Sesqui-oxide iron . 



Silica 



Carbon 



Loss 



905 

 415 

 81-71 

 0-14 

 1-81 

 1-20 

 1-94 



100- 



It may, a priori, be considered likely to produce a most excellent body for a 

 sound and durable paint under water. The black oxide of mangar\ese has no 

 advantages but that of being a powerful drier. The defects of all oil paints 

 arise from the instability of their bases ; the acids which enter into the con- 

 stitution of all fixed oils readily quit their weakly positive organic bases to 

 form salts with the oxides of the metal on which they may be laid. Hence 

 we must look for improvements in our paints to those substances among the 

 organic groups which have greater stability than the fat or fixed oils, and 

 which, in the place of being acid or Haloid, are basic or neutral. The heavy 

 oUy matter obtained from the distillation of resin, called " resenien." and 

 eupion, obtained from rapeseed oil, have valuable properties as the bases o£ 

 paints. 



Tables IX. and X. contain the results as to the corrosion of cast iron in 

 sea water when exposed in Voltaic contact with various alloys of copper and 

 zinc, copper and tin, or either of these metals separately, per square inch of 

 surface. It appears that neither brass nor gun metal has any electro-chemi- 

 cal protective power over iron in water, but on the contrary promotes its 

 coiTOsion. This question is only a particular case of the following general 

 question : viz. if there be three metals, A. B. C, whereof A. is electro-posi- 

 tive, and C. electro-negative, with respect to B., and capable of forming 

 various alloys, 2A + C....A-(-C....A + 2C; then if B. be immersed in a 

 solvent fluid in the presence of A., B. will be electro-cheraically preserved, 

 and A. corroded, and vice versa. If B. be so immersed in the presence of C, 

 B. will be dissolved or corroded, and C. electro-chemically preserved ; the 

 amount of loss sustained in either case being determined according to Fara- 

 day's " general law of Volta-equivalents." The tables show that the loss 

 sustained by cast iron in sea water, as compared to the loss sustained by aa 

 equal surface of the same cast iron in contact with copper, is 8'23: I1'37; 

 and when the cast iron was in contact with an alloy containing 7 atoms of 

 copper and 1 of zinc, the ratio was 8'23 : 13'21 ; so that the addition in this 

 proportion of an electro-/;osi7it'e metal to the copper produces an alloy (a 

 new metal, in fact) with higher electro-negative powers, in respect to cast 

 iron, than copper itself. The author discusses many results equally remark- 

 able, and is therefore enabled to suggest by its chemical notation the alloy of 

 " no action," or that which in the presence of iron and a solvent would 

 neither accelerate nor retard its solution, one of the components of this alloy 

 being slightly electro-negative, and the other slighly electro-positive, with 

 respect to cast iron. These results will also enable some advances to be 

 made towards the solution of the important problem proposed by the author 

 in his former report, viz. " the obtaining a mode of electro-chemical protec- 

 tion, such that while the metal (iron) shall be preserved, the protector shall 

 not be acted on, and the protection of which shall be invariable." 



Table X. exhibits especially the results of the action of sea water on cast 

 iron in the presence of copper and tin or their alloys. It appears that copper 

 and tin being doth electro-negative with respect to cast iron, all their alloys 

 increase or accelerate the rate of corrosion of cast iron in a solvent, though ia 

 very variable degrees ; the maximum increase is produced by tin alone, thus 

 indicating that this metal (contrary to what was previously believed) is more 

 electro-negative to cast iron than copper. Hence the important practical de- 

 duction, that, where submerged, works in iron must be in contact with either 

 alloy, viz. brass or gun metal ; common brass, or copper and zinc, is much to 

 be preferred. These experiments %viU also serve to demonstrate the fallacy 

 of many of the patented so-called preservatives from oxidation, which are 

 brought before the public with so much parade. 



The author lastly proceeds to the subject of the specific gravity of cast ii'on, 

 tables of which aie added to the preceding. The specific gravities here re- 

 corded were taken on equal sized cubes of the several cast irons cut by the 

 planing machine, from bars of equal size, cast at the same temperature, in 

 the same way. and cooled in equal times. Many of these results differ con- 

 siderably from those given by Dr. Thompson and Mr. Fairbairn ; which the 

 author refers to the probability that those of Dr. Thompson were taken from 

 pieces of the raw pig, and those of Mr. Fairbaurn by weighing in air equal 

 bulks cut from the mass by the chisel and file, by which latter process the 

 volume is liable to condensation. The experiments of Mr. Fairbairn and Mr. 



3 M 



