May, 1913 T H E JOURNAL OF I.YDLTSTRIAL

ORIGINAL PAPERS

interesting point in this connection was the former universal condemnation of copper by steel-makers as a producer of red shortness but in very considera- ble amounts its effect is not now feared in this respect.

A development, or an extension of one originating in the industry itself, or with workers associated with iron and steel investigation, is tha t of heat treat- ment, which is more largely applied by fabricators of steel parts, especially of high-grade machine parts for automobiles, engines and many forms of apparatus \There high duty is requisite. The ubiquitous auto- mobile has contributed more to the extension and ap- plication of heat treatment than any other factor. Not tha t discoveries have come from this source, but in their aggressive and enterprising spirit the builders have availed themselves of every possible advance in applied science to perfect details of con- struction and quickly recognized the benefits of the proper heat treatment of steel. In this way, the auto- mobile industry has accelerated the application of heat treatment more than any other agency until now every progressive maker of parts for high duty

service in many lines of construction, gives great at- tention to this feature. The proper heat treatment of steel has added greatly to the safety and reliability of much of our modern equipment in various lines and it is difficult to imagine hon much of the exacting service could otherwise have been met.

No survey of progress, however brief, should neg- lect the human side of the industry-the efforts for betterment, and accomplishmeiits in this direction. While efforts of this kind are not distinctive of any in- dustry but represent a movement of the period in all lines, no branch of industry has more vigorously promoted or is more successfully advancing this work than the one under consideration. The work for sani- tation, safety and health is of amazing magnitude, practically every concern, not only generously spend- ing money, but what is more important, conscientiously working to conserve the health, safety and welfare of all classes of employees until no finer examples of re- sults exist in our industrial life than many that are to be found in the Iron and Steel Industry.

G. D. CHAMBERLAIN.

AN ELECTROLYTIC METHOD FOR THE PREVENTION OF THE CORROSION OF IRON AND STEEL’

By J. K. CLEMENT AND L. V. WALKER

INTRODUCTION

According to the electrolytic theory of corrosion, when iron is oxidized in the wet way it first goes into solution as ferrous ions. The ferrous ions are then oxidized by the oxygen present in the water t o ferric ions and precipitated as ferric hydroxide. Simul- taneously with the formation of ferrous ions, hydrogen is liberated on the surface of the iron. The passage of the iron from the atomic to the ionic state and the passage of the hydrogen fromthe ionic totheatomicstate are accompanied by a transfer of electricity. The fer- rous ions derive from the metal surface a charge of posi- tive electricity, the hydrogen ions give up a charge of positive electricity and a current flows through the metal from the point where hydrogen is liberated to the point where iron is dissolved.

If iron and a metal having a greater ‘ I solution ten- sion” than iron-for example, zinc-be immersed in a n acid solution and connected externally through a wire, zinc will go into solution, hydrogen ions will be liberated on the iron, and current will flow through the wire from iron to zinc. The potential difference be- tween zinc and iron being opposed to the potential difference between iron and electrolyte, neutralizes the force required to pull the positively charged iron ions away from the negatively charged surface of the iron electrode. Consequently iron does not go into solution.

In place of zinc, other conducting materials-for ex- ample, carbon-may be used as the anode, and the E. M. F. required may be furnished from an external source.

1 Paper presented a t the Eighth International Congress of Applied Chemistry, New York, September, 1912.

I n this way electric energy derived from either a bat- tery or a generator may be used to protect metals from corrosion under water.

A vast amount of work, both theoretical and experimental, has been directed to the study of the causes of corrosion and the development of methods for its prevention, and the literature on the subject is extensive.2 Very little attention, however, seems to have been given to the development of an electro- lytic method for the prevention of corrosion. An electrolytic method has been used to some ex- tent for the protection of boilers.3 This method con- sists in submerging a bar of zinc in the boiler water and connecting the zinc electrically with the boiler plates. After the experiments described in this paper were begun i t was found that an electrolytic process for the protection of metals had been patented by an Australian inventor.4 The writers have been unable to find any data published by the inventor.

In a paper presented before the Sydney, Australia, Section of the Society of Chemical Industry, G. Harkers describes experiments on the prevention of the cor- rosion of iron by electrolysis. He determined the cur- rent density necessary to prevent the corrosion of steel plates submerged in N / 2 j H,SO,, t ap water, sea water and “dilute acid.” The results of Harker’s experiments, which furnish the only data on the sub-

1 Author’s abstract of an address delivered as chairman of the Pitts- burgh Section of the American Chemical Society, January, 1913.

2 Cushman, A. S., and Gardner, H. 4 . . “Corrosion and Presenation of Iron and Steel,” N e w York. McGraw-Hill, 1910. “List of References to Books and Magazine Articles on Metal Corrosion and Protection,” Bulletin, Carnegie Library of Pittsburgh, July, 1909.

J. H . Paul: “Corrosion in Steam Boilers,” Trans. SOC. Enoimeers, 31, 147.

4 See Elec. Review and W e s t . Electn., 58, 326 (191 1 ) . 5 G. Harker and J. McPiamara, “Electrolysis as a Means of Preventing

the Corrosion of Iron and Steel,” J O U I . SOC. Chem. Id., 29, 1286 (1911).

1-

1

362 T H E J O U R N A L OF I N D U S T R I A L A N D E-VGINEERISG C H E J f I S T R Y Vol. 5 , No. 5

ject which the writers have been able to find, will be discussed in connection with the results of their own experiments.

E X P E R I M E N T S BY F. M. STANTON

Experiments to devise an electrical method for the prevention of the corrosion of steel immersed in acid water were undertaken a t the Pittsburgh Experi- ment Station of the Bureau of Mines, in May, 1911, by Mr. F. M. Stanton.

Stanton's work included laboratory experiments and tests on plates immersed in the Monongahela River. In the laboratory experiments, steel plates 6 mm. X 19 mm. X 5 1 mm. were suspended in sepa- rate vessels in N / 5 H,SO,. In one vessel a car- bon rod was suspended near the steel plate and con- nected to the positive pole of a storage battery, the negative pole of which was connected with the steel plate. In the second vessel the steel plate was not protected. After 24 hours exposure it was found that the unprotected plate had lost in weight I . 8100 grams and the protected plate only 0.0012 gram.

I n the tests with plates submerged in the Monon- gahela river a much smaller current density was used than in the laboratory experiment, and the protec- tion against corrosion was much less complete. The plates used in these tests had the following dimensions: 6" X 8" X I//. The protecting current was fur- nished by a 2-volt accumulator. The loss in weight of the two plates a t the end of fifteen days immersion was for the protected plate 0 . 7 gram and for the un- protected plate I O . 9 grams.

E X P E R I M E N T S TO D E T E R M I N E CURRENT DENSITY

REQUIRED TO P R E V E N T CORROSION

Mr. Stanton left the service of the Bureau in July, 1911, and the work was interrupted until December,

E XP ERIM E N TS w IT H A PP A R A T u s I .-The first method of determining the protective effect of small currents was to suspend an iron plate, of the dimen- sions shown in Fig. I, in an 800 C.C. flask parallel to a I/, inch carbon rod and I / ~ inch from it. Using storage bat- teries as a source of E. M. F., a current was passed from the carbon rod through dilute sulfuric acid to the iron plate. The current regulation was accomplished by the use of variable resistances and the current measured by the drop in potential over a known re- sistance, by means of a Siemens and Halske milli- volt meter; electrical connections were always com- pleted before any acid was introduced into the flask and the current was regulated immediately thereafter. I n order to save time six such flasks were used in each experiment.

The results of the first series of experiments are shown graphically in Figs. 2 and 3. The experi- ments represented by the curves in Fig. 3 were made

7 m u

3 b ; ea0

1

2 im

0 I 4 3 e 7 0 m C N I E N S T I - M L L W C F f S X R Io N

FIG. 2-RELATION BETWEEN LOSS O F TVEIGHT AND CURRENT DENSITY Apparatus I. Duration of Test, 45 hrs. N/5 HzS04, Electro1,te

Unstirred

under uniform conditions, except that the duration of the test was varied. These experiments indicate that a certain definite current density must be main-

FIG. APPARATUS I

1911, when experiments were resumed by the authors FIG. 3-RKLATlOiX BKTWEEN LOSS O F TITEIGHT AND CURREXT DENSITY

Apparatus I. X, 10 H2S04, Electrolyte Unstirred

and Mr. A. E.-Hall. tained in order to reduce the corrosion loss to a mini- The object of the experiments to be described was to mum. On account of the poor agreement between

determine the current density required to prevent, the results of individual experiments, a second type under various conditions, the corrosion of steel plates of apparatus was designed. in acid water, EXPERIMESTS WITH A P P A R A T U S 11.-In order to

May, 1913 T H E JOC7R.V.1L OF I-VVDiYSTRIAL

eliminate as fast as possible any error due to differ- ences in acid concentrations about the iron plates, all the test plates were im-xersed in a common ves- sel. Another improvement was the employment of an acid reservoir, and a constant level device by means of which fresh acid could be fed into the con- taining vessel a t any desired speed, and the excess electrolyte could be siphoned off through the level bottle into the overflow. The vessel was a porcelain-lined cylindrical dish of about 2 2 jo c c. capacity and the plates were supported by a thin wooden cover a t points on a circle 4" in diameter, and a t whose center the carbon rod anode was suspended. The plates were set equi-distant from each other and, of course, from the anode.

Six experiments were conducted according to this scheme and the acidity of the electrolyte was deter- mined a t the end of each run. There was always a

9 15 CLRRENT [XNSIPI-Mf&F€RES F€W:WN

FIG 4-RELATION BETWEEV LOSS OF WEIGHT AND CURRENT DESSITY Apparatus I1 N/10 H2SOI, Electrolyte Vnstirred

considerable decrease in acid strength which could be eliminated to a large degree by increasing the rate at which the acid was fed. For example, a t the end of a 27-hour run, started with *V/IO H,SO,, the acidity was S ~ / I O O O normal. As all the plates, however, were subjected to the same variation in the strength of the electrolyte, the loss in weight of the plates should be comparable.

The curves in Fig. 4 represent the plate losses for experiments 11 and 1 2 , in which apparatus I1 was used. Though the results do not check exactly, they are more consistent than those of the previous experiments. The indications from the experiments ofsSeries I and I1 are that a current of 0 . 4 milliampere per square inch (figuring I / 3 of the back of the plate as is customary) will prevent 93 to 94 per cent. of the corrosive action of N/IO H,SO, when the solution is not stirred.

E X P E R I M E N T S W I T H APPARATUS 111. - Apparatus 111, the one finally adopted, is shown in Fig. 5 . It differs from apparatus I1 chiefly in having a greater capacity-4. j liters--and in being provided with an attachment for stirring the electrolyte. The plates were arranged as before, except that they were set inches from the carbon rod, instead of 2 inches, in order to keep them under water when the electrolyte was rota- ting rapidly. For stirring, a low power, hot air engine gave fairly constant speed and could be run over night. I n order to be able to calculate the current densities with greater accuracy, all surfaces of the

' *I 1.

AND ENGINEERING CHESIISTR I7 363 1'

plates except the one facing the anode were protected by being painted with Bakelite Lacquer. An applica- tion of four or five coats gave a good. smooth surface that was impenetrable by acids.

The first few experiments using the final apparatus

FIG. 5-APPARATUS I11 A . acid reservoir; B , electrolytic tank; C. level bottle: D , known

resistance; a, belt t o motor: b , stirrer; c , to overflow

were run with the stirrer rotating a t 660 R. P. M. and under these conditions, CUI-rents that had pro- tected the iron plates almost completely in still water, were absolutely useless. On reducing the stirring rate to 3 j R. P. M. in N/ IOO acid, small currents were again effective, and runs could be checked with a fair degree of accuracy.

Fig. 6 shows the results of experiments using -\-/Ioo H,SO, as electrolyte and with the stirrer running a t 35 to 40 revolutions per minute. I n this series of ex- periments the electrodes were removed from the elec- trolyte at frequent intervals and their loss in weight determined. The curves in Fig. 6 show that the plates having a current density of 0 . 4 milliampere

F I Q . &EFFECT OF CURRENT DENSITY ON RELATION BETWEEN Loss OF WEIGHT AND DURATION OF TEST

Apparatus 111. N/100 H2SOl, Electrolyte Stirred, 35 R. P. M.

per square inch or less, continued to lose weight dur- ing the entire period of the experiments. Plates pro- tected by current densities greater than 0.4 milli- ampere per square inch exhibited a marked initial corrosion and thereafter no further loss in weight.

364 THE JOURNAL OF I N D U S T R I A L

The initial loss in weight of the plates protected by current densities greater than 0 . 4 milli-ampere is probably due t o the fact that a t the start of the ex- periment the plates were covered by a film of air and that the oxygen of the air film acted as a depolar- izer. Some time elapsed, therefore, before the plates became completely polarized, and during this period they were incompletely protected by the current. The difference between the rate of loss in weight of the unprotected plate during the initial period and the rate of loss during the remainder of the experiment is probably due to a difference between the structure of the original surface of the metal and the interior, as well as to the action of the film of air surrounding it a t the beginning of the experiment.

LYTE.-on account of the depolarizing effect of oxygen it was to be expected that the amount of oxygen present in the electrolyte would be an important factor in determining the speed of cor- rosion. Fig. 7 shows the results of a series of four experiments made to ascertain the effect of increas- ing the amount of oxygen in the electrolyte. The composition of the electrolyte, N/IOO H,SO,, and-the

EFFECT O F PASSING OXYGEN THROUGH ELECTRO-

"

FIG. CU EFFECT OF OXYGEN ON RELATION BETWEEN Loss OF WEIGHT AND CURRENT DENSITY'

Apparatus 111. Duration of Test, 24 Hrs. N/100 HzSOa, Electrolyte Stirred, 35 R. P. M.

speed of the stirrer, 35 R. P. M., were the same in all four experiments. The lower curves repre- sent two experiments with the usual amount of oxygen present. The electrolyte was prepared by adding sulphuric acid to distilled water without any special precautions to exclude the air or to aerate the solution. In experiments 38 and 39, represented by the two upper curves of ,Fig. 7 , the electrolyte was saturated with oxygen by rapidly bubbling oxygen through it during the experiment. The curves show that increasing the amount of oxygen in the solution accelerates the speed of corrosion very materially, the loss of weight in 24 hours being about twice as great when oxygen was passed through the electrolyte as when oxygen is supplied by diffusion only. It follows from this that the rate of corrosion in any acid solution depends on the degree of aeration of the solution.

EFFECT OF STIRRING.-The enormous acceleration of corrosion produced by extremely rapid stirring, 660 R. P. M., has been noted on page 363 . On ac- count of the magnitude of the effect, a series of ex-

A N D ENGINEERING CHEililISTRY Vol. 5 , No. 5

periments was made in which the speed of the stirrer was the only variable. The results are shown in Fig. 8. An examination of the curves shows that, the other conditions being constant, the loss in weight by corrosion varies directly as the rate of stirring. The increase in the speed of corrosion with the increase in the rate of stirring is further illustrated by Fig. 9, in which the loss in weight of the plates for several different current densities is plotted against the num- ber of revolutions of the stirrer per minute.

These results demonstrate that the rate of flow of acid solution over the metal surface is an important factor in determining both the amount of corrosion and . the current density necessary to protect the metal. The decided increase in the speed of corrosion

FIG. &-EFFECT OF SPEED O F STIRRER ON RELATION BETWEEN LOSS O F WEIGHT AND CURRENT DENSITY

Apparatus 111. Duration of Test, 24 Hrs. N/100 H2SOI

produced by circulation of the electrolyte is not sur- prising, since the flow of solution over the metal not only hinders the exhaustion of hydrogen ions in the vicinity of the anode, but provides~ a continual supply of oxygen and tends t o destroy any film of hydrogen which may form on the surface of the cathode.

Id0 v) g I20

: L 100 I

(0 5 0 0

k

d b 6 0

9 40 z 1 20 0

0 200 300 400

SPEED OF STIRRER-RPM 100

FIG. EFFECT OF CURRENT DENSITY ON RELATION BETWEEN Loss OF

WEIGHT AND SPEED OF STIRRER Apparatus 111. Duration of Test, 24 Hrs. N/1000 HzSOa

ACID CONCENTRATION.-That the rate Of COrrOSiOn of iron or steel in sulfuric acid solutions increases with the acid concentration is well known. Though the acid concentration in the waters of nearly all streams

May, I 9 1 3 T H E JOL‘R.V-4L OF I .VDVSTRIAL

in which the corrosion of iron ordinarily occurs is much less than N/Ioo-the concentration used in most of the experiments described-the application of the method to more concentrated solutions seemed of sufficient interest to justify experiments with stronger solutions. Accordingly a series of experiments was made in which the concentration of sulfuric acid was varied from N / ~ o o o to K / I o . The results are represented by the curves in Fig. I O . They indicate that, although a considerably greater current density would be required, the method would be effectual even in N/IO H,SO,. The curves for experiments with N / I O O

FIG. 10-EFFECT O F ACID COXCEXTRATION UPON RELATION BETWEEN

LOSS O F WEIGHT AND CURRENT DENSITY Apparatus 111. Duration of Test, 24 Hrs. Electrolyte Stirred,

35 R. P. M .

and N / ~ o o o acid are practically identical, showing that between these limits of acid concentration there is no change in the rate or corrosion greater than the limit of error of observation. This result, though unexpected, periments.

Current density Current milliamperes

milliamperes per sq. in 0 . 5 0 0 . 2 1 .oo 0 . 4 2 .oo 0 . 8 4.00 1.6

10.0 4 . 0 Rate of stirring = 35 R Acid - N/100 HSOl

was confirmed by repeated ex-

A!VD EXGINEERING CHEA14I5TR Y 365

especially in view of the variations found between the results of different experiments made under like con- ditions bf acid concentration and rate of stirring, to be of value as a measure of the degree of protection against corrosion afforded by the current.

CURRENT DExsITY.-In the experiments which have been described, the following factors were found to influence the current density required to protect the corrosion of steel plates submerged in sulfuric acid solutions :

( a ) ( b ) Amount of oxygen present. (c) Acid concentration. Of these the last named is probably least important,

especially in dilute solutions. The rate of stirring of the electrolyte is by far the

most important factor. Under normal conditions the supply of oxygen in the immediate vicinity of sub- merged metal surfaces depends largely on the rate of flow of solution over the metal surface. With the lowest rate of stirring used in the experiments, 35 R. P. M., and in acid concentrations not greater than N / I O O a current density of from 0 . 5 to 0.8 milliampere reduced the corrosion loss to a negligible quantity, whereas with the stirrer rotating a t a speed of 450 R. P. M. and a current density of 2. o milliamperes the corrosion loss amounted to z j per cent. of the loss on the unprotected plate.

Rate of stirring of electrolyte.

HA RR E R ’ s EXPERIMENT s From his experiments referred to above, Harker con-

cluded that in employing the electrolytic process the cur- T.4BLE I.-ELECTRODE POTENTIALS

POTENTIAL DIFFERENCE BETWEEN PROTECTED AND UNPROTECTED PLATE Exp. 30 Exp. 3 1 Exp. 32 Exp. 35 Exp. 36 Exp. 37 Exp. 38 Exp. 39 Exp. 40

0 . 0 2 1 0,004 -0.021 -0.024 0.004 0.012 0.027 -0.005 -0.004 . . . -0.003 -0.012 0.003 0.022 0 ,008 -0.018 -0.051 -0.024

0 ,198 0.025 0.102 0.078 0.073 0.044 0.012 0.000 0.015 0 .433 0 . 1 1 4 0.237 0.186 0.187 0.10s 0.102 0.081 0 .090 0,723 0 289 0 429 0.372 0.392 0 .304 0 .315 0 .295 0.303

P. 11,

POTENTIAL O F I R O I i ELECTRODES.-In the majority of the experiments, in addition to measuring the cur- rent flowing into the different electrodes, the poten- tial of the individual electrodes was measured against a zinc electrode. Potential readings were made with a Siemens & Halske millivolt meter of 4jo ohms resist- ance in series with a resistance of 175jo ohms. The results are given in Table I. The agreement between the potentials of electrodes having the same current density is only fair. In N/IOO H,SO,, and with a rate of stirring of 35 R. P. M., a current density of 0.8 milliamperes per square inch generally produced an increase in the potential of the iron electrode of about 0.04 volt. (In the table, potential differences are positive when the potential of the protected plate is positive to the unprotected plate, i. e . , when the potential difference between the zinc electrode and the protected plate is decreased by the protecting cur- rent.) As small an increase as 0.04 volt in the po- tential of the test plates sufficed, therefore, to protect them against corrosion. This effect is too small,

rent required to prevent corrosion might be calculated from the rate of corrosion of iron or steel under the given conditions and the electrochemical equivalent of iron. This conclusion he confirmed by one set of experiments in which, however, iron anodes were used. I n experiments with a platinum anode, Harker found that the protection afforded by the current to the iron cathode was somewhat less than when an iron anode was used. He concluded that “ the electrolytic process does not actually prevent metal from entering the solu- tion from the cathode, for where the rate remains con- stant, deposition is taking place from the anode, yet the tendency of the metal to pass into solution is much diminished.”

APPLICATION OF HARKER’S RmE.-In all of the ex- periments made by the present authors, a carbon anode was used. Harker’s method of calculating the cur- rent density required to prevent corrosion was based on experiments with an iron anode. If, then, as Harker suggests, the protective action of the current is largely due to the transfer of iron from the anode

366 T H E JOURNAL OF I N D U S T R I A L A N D ENGINEERING C H E M I S T R Y Vol. 5 , No. 5

to the cathode it could not be expected that his method would hold for experiments with a carbon anode.

The method offers, however, such a simple pr6cedure for arriving a t the proper current density required to prevent corrosion that it seemed worth while to ap- ply it to the results of the authors’ experiments. I n Table I1 are given the results of these experi- ments, to which have been added the values for the current densities calculated from the loss in weight of the unprotected plates and the electrochemical equivalent of iron. The “observed” values for the current densities in column 6 were obtained graphically from the curves in Figs. 2 , 3, 4, 7 , 8 and IO.

The agreement between the observed and calculated values is remarkably good. The differences are well within the experimental error.

I t may, therefore, be concluded that Harker’s method of calculating the current density required to prevent corrosion, from the loss in weight of the metal under the given conditions, is not limited to cases in which an anode of the same metal is used.

TABLE I1

d

z . . . . 4 3 . . . . 6 3. ... 7 3 . . . . 9 3 . . . . 8

4 . . . . 1 1 4 . . . . 12 7 . .... 39

7.. . . 38 7 . . . . 40 7 . . . . 30 S . . . . 48 8 . . . . 45 S . . . . 46 8. ... 42 8.... 41

10 .... 51 lo.... 49 10.. .. 50

d 8

s a

N/5 N/10 N/10 N/10 N/lO

N/lO N/10 N/100

N/lOO(a) N/100 N / l O O N / l O O O N/1000 N/1000 N / l O O O N / l O O O N/10 N/50

Apparatus I: .. . 45 .0 . . . 4 7 . 0 . . . 2 0 . 0 ... 23.5 .. . 2 2 . 0

Apparatus 11: . . . 22.5 . I . 2 2 . 0 35 2 4 . 0

Apparatus 111: 35 2 4 . 0 35 24 .0 35 24 .0

450 2 4 . 0 110 24 .0 110 24 .0 35 2 4 . 0 35 2 4 . 0 35 2 4 . 0 35 2 4 . 0

Current density milliamp.

per sq. inch -. Obs. Calc.

3 . 0 2 . 2 3 . 0 3 . 4 3 . 0 2 . 0 1 .5 2 . 7 1 . 5 2 . 0

0.45 0 . 5 0 .45 1 .2 1 . 6 1 . 3

1 . 6 1 . 0 0 . 8 0 . 7 0 . 8 0 . 7 2 . 0 2 . 7 1 . 6 1 . 4 1 .6 1 . 2 0 . 8 0 . 7 0 . 8 0 . 6 1 . 6 2 . 2 1 . 2 1 . 1

iu 1: n a I $ 3 33s .3 gg B

R -

+ v

344.6 551 .5 140.5 221 .o 151.5

43 .0 93 .5 7 5 . 9

6 4 . 3 44 .6 4 1 . 4

170.0 8 8 . 3 72 .8 44 .5 41 .O

133.8 7 1 . 3

N/50 35 24 .0 1 . 2 1 .3 8 0 . 9 (a) Oxygen saturated.

THE XATURE O F THE A C T I O N O F THE ELECTRIC CURRENT

Harker’s view that the current does not actually prevent metal from entering the solution from the cathode, and that its protective action is due to the deposition of metal from the anode, is contradicted by the authors’ experiments in which carbon anodes only were used. The action of the current, then, is t o pre- vent the formation of ferrous ions at the cathode by increasing its negative charge. In the absence of any depolarizer, especially of dissolved oxygen, an in- finitesimal current should produce an electromotive force large enough to neutralize the solution tension of the metal. When oxygen is present, it combines with the hydrogen ions liberated a t the cathode and hinders polarization. The supply of hydrogen ions furnished by the current must be equivalent to the

supply of oxygen in the vicinity of the cathode before the solution tension of the cathode can be completely neutralized.

C O N C L U S I O N

I t has been shown that the corrosion of iron sub- merged in sulfuric acid solutions may be prevented by imposing a counter E. M. F. The density of the cur- rent required to prevent corrosion depends on various factors, the more important being acid concentration, amount of dissolved oxygen and degree of circulation of electrolyte. The influence of these factors has been studied and curves have been given showing the cur- rent density necessary under various conditions. I t has been found that the current density required can be calculated, within the limit of experimental error, from the loss in weight of the unprotected metal under the given conditions.

u. s. BUREAU OF MINES PITTSBURGH, PA.

PAINT AS AN ENGINEERING MATERIAL’ By MAXIMILIAN TOCH

The progress that paint chemistry has made since 1905 is by far greater than the progress that was made from its earliest invention up to that date. It is very difficult for me to imagine that my first book on “The Chemistry of Paints” stimulated others to continue the work which I had started, and if the little that I have done to enlighten the manufacturers and consum- ers has brought about the progressive results, I cer- tainly have been rewarded for all the work I have ever done on the subject.

The first skyscraper ever built was the Gillender Building, corner of Wall and Nassau Streets, which was razed two years ago. Chemists knew before this building was demolished that linseed oil paint was not the best material for the protection of steel of large buildings. The question as to whether our monu- mental buildings are permanent has been a source of great worry to many chemists and engineers. For- tunately, if any of the steel contained in buildings like the Woolworth Building, Metropolitan Tower, the Singer Tower and dozens of others should show signs of corrosion and disintegration, the process is so slow that preventive methods could be applied, for a beam could not corrode in a masonry wall without cracking or bulging the wall. I have in mind one building in Maiden Lane where this actually occurred; the wall of the fifteenth floor was cut away, the cor- roded beam exposed, thoroughly scraped, painted and reinforced, surrounded by concrete, and the brick wall replaced.

From the street level up, every skyscraper in the world is safe, but from the street level to the grillage beams is the dangerous point. Of course, a small building could be “jacked up” and a grillage beam replaced. In large buildings-two of which I have in mind-where the grillages were affected by leaky electrical currents, the foundation beams were un- covered, scraped clean and painted, and then a grout

1 Abstracted by the author from an address presented before the N. Y. Section of the American Chemical Society, Chemists’ Club, Feb. 7. 1913.