480
THE J O U R N A L OF I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y .
The cost of production of alcohol by the sawdust process (Ruttan, J . SOG.Chew. Ind., 1909,p. 1290) is said to be about equal to t h a t of grain, although there is such an enormous difference between the cost of the raw materials. One company, however, claims t o be able t o manufacture alcohol by the sawdust process a t a cost of $0.07 per gallon. Without going into the details of the chemistry involved, it is striking that, although the patents claim a conversion of cellulose into sugar, the so-called true fibrous cellulose, absorbent cotton, does not yield t o this treatment. Also the material left after treatment contains cellulose which upon re-treatment yields but traces of fermentable sugar. Although authorities differ, it is claimed by some t h a t the acid acts as a catalyzer hydrolyzing the lignone complex and a proportion of the “easily attacked cellulose.” The oxycellulose and the “true cellulose” are said not to be seriously altered chemically. I t would seem t h a t there is a possibility t h a t a process might be evolved where the alcohol conversion of the wood pulp might be made first, and then the residue subsequently utilized for the production of paper pulp. ALCANHIRSCH. OCCUPATIONAL DISEASES.
This subject is now receiving the attention it deserves. Hitherto the matter of diseases peculiar t o industries has been in the hands of physicians. While medical men are competent to treat and cure such diseases i t can readily be seen t h a t they are unable b y nature of their training t o accomplish much in the elimination of the causes of occupational diseases. The physician is not the proper person to suggest t o a manufacturer such changes in his processes as will eliminate the causes of the peculiar occupational ills
I I
July, 1912
to which his employees are subject. This is necessarily the function of the chemical engineer and i t is to him t h a t the physician now turns for assistance. The June number of THISJ O U R N A L published an address delivered by Dr. W. Gilman Thompson, Professor of Medicine in the Cornel1 University Medical College, in New York City, before the New York Section of the American Chemical Society. The attention of our readers is called to the facts set forth and the points brought out in this address, and chemists will readily see the opportunity which presents itself to them, to be of assistance in this crusade of the Twentie t h Century. The New York Section of the American Chemical Society has appointed a Committee on Occupational Diseases, which is now cooperating with a similar Committee of the New York Association for Labor Legislation. The personnel of these committees is given in another column of this issue. A joint meeting of the two committees was held on May g ~ s t a, t the Chemists’ Club in New York City. This meeting was addressed by Dr. Alice Hamilton, whose work of investigating lead poisoning for the U. s. Department of Labor is well known. Dr. Hamilton spoke of the difficulties she had experienced in obtaining information, and laid particular stress on the importance of the cooperation of the chemist and physician. It will be readily seen t h a t this is a matter for organized and systematic work by our profession, and t h a t the campaign must be one of educating the public, the workingman, and the manufacturer. I t is evident t h a t statistics must be collected so proper lines of work can be designed. More will be accomplished by education than by inconsidered or hasty legislation, and b y this method the least possible hardship will be E. C. UHLIG. inflicted on established industries.
ORIGINAL PAPERS
I
The acid theory presumes t h a t corrosion is primarily caused by acid attack. It denies the solubility B y WILLIAM R. FLEMING. of iron in pure water and denies t h a t iron will rust I n a recent article in the Irovt and Steel Institute,a in acid-free air and water combined. Dr. J. Newton Friend again attacks the “Electrolytic” The electrolytic theory presumes that iron is soluble Theory of Corrosion. I n his Paper, a Carnegie Schol- in pure water, and t h a t rust is subsequently formed arship Memoir, he describes an experiment which through the influence of oxygen. The exposed surdemonstrates t h a t “Pure water and pure air combined faces of iron and steel contain segregated impurities are without visible action upon pure iron.” “This which are supposed to give rise to differences of poshows, Friend continues, t h a t the electrolytic theory tential, in the presence of an electrolyte, which is of corrosion is untenable, and definitely confirms the generally moisture. At those points where the metal older acid theory.” is anodic, iron passes into solution, assuming the I n this paper I shall attempt to show t h a t iron and ferrous ionic condition, while hydrogen is liberated steel will rust in pure water and air combined, if it a t the cathode. Oxygen plays its part by depolarizing is given a chance. Furthermore, I shall attempt to the hydrogen covered cathodes and by oxidizing the prove t h a t Friend’s experiment is a strong confirma- ferrous ions a t the anodes. In its purest sense the tlon of the electrolytic theory, and not destructive electrolytic theory presumes t h a t corrosion begins to it. in a pure electrolyte, e. g., water free from all traces We have to-day only two important theories of of impurities. corrosion, the “Acid” and the “Electrolytic” theories. The defendants of the acid theory deny the solu1 R e a d before the Cincinnati Section of t h e A C. S., February 14, bility of iron in pure water and are emphatic in their 1912. claim that corrosion is “primarily the result of acid 2 Iron am$ Steel Insfztufe, 3, 1 (1911). CORROSION OFIRON INPUREWATER AND AIR COMBINED; THE ELECTROLYTIC THEORY AGAIN CONFIRMED.’
July,
1912
T H E J O U R N A L OF I N D U S T R I A L A N D EiZ‘GINEERING C H E M I S T R Y .
attack.” Since carbon dioxide is universally present in our atmosphere it is considered t h e chief cause of rusting. It attacks the iron forming ferrous carbonate and liberating hydrogen. Oxygen and moisture further react with the ferrous carbonate t o form rust, freeing the carbon dioxide which in turn reacts with fresh iron. But, according t o our latest notions of solutions, carbonic acid does not attack the iron directly, e. g., in its atomic condition. It must first pass into solution (become ionic) before the acid can attack it. Therefore, the acid theory is b u t a special case under t h e general electrolytic theory. This has been admitted. But, while admitting the electrolytic character of corrosion, when viewed from the general standpoint of ionization as announced by Arrhenius, the defendants of the acid theory insist t h a t iron will not assume the ionic state unless an acid is present in t h e electrolyte. Hence, they say, acid attack is the true beginning of corrosion. There can be no doubt t h a t the presence of acids (chiefly carbonic) has much t o do with the rapid corrosion of our iron and steel. Without acids in our air and water the corrosion problem would be less formidable. This, however, is only a quantitative view-point. I n attempting t o get at the facts concerning the true cause of corrosion we should not allow ourselves t o be blinded by mere quantitative ideas. Given pure iron in pure air and water, the factor of corrosion is very small. Likewise, with impure iron in impure air and water, the factor of corrosion is very great. But this does not interest us. We are concerned only with the true startiqzg point of corrosion. I t matters not how small the value of this “starting point,” quantitatively, if iron is soluble in pure water even to the slightest extent, or, if iron will rust in pure water and air combined be it ever so slight, then this weakness in the metal itself is the true cause or starting point of corrosion. All influences vhich enter later are mere retarders or accelerators. Below is given Friend’s experiment and conclusions in his own words: A-E is a hollow cylinder of iron or steel, closed at one end. The open end is plugged with a tightly fitting rubber stopper, bearing two glass tubes, arranged in such a manner t h a t cold water can circulate through them. The glass bottle. containing about I O O cc. of twice normal caustic potash solution, is placed in a deep cold water bath, and heated gradually t o r o o o C. I t s mouth is now closed with a large India-rubber hung, from which hangs the metal cylinder by the glass tubes-already alluded to-and the whole is thoroughly shaken t o remove every trace of carbon dioxide from the walls, air, and surface of the iron. A current of cold water is now passed through the tubes, and on immersing the bottle again in the hot water bath, pure water vapor condenses upon A-E and drips off,rapidly washing it free from alkali. That the cylinder is actually free from alkali may be c _
481
readily demonstrated by opening the apparatus, and testing with phenolphthalein, when no change in color occurs. Nevertheless, the iron remains perfectly bright for an indefinite number of days. An isolated spot of rust will often form here and there upon the surface of the metal, owing to the unavoidable traces of impurity always present in the purest metal obtainable, and if the apparatus is opened and the cylinder again polished with emery, the spots will appear again in the same places upon repetition of the experiment. Furthermore, where the iron is in contact with the India-rubber, a little corrosion takes place, owing to the presence of sulphur, etc. These points, however, do not affect the general result of the experiment. The conclusion is inevitable: “Pure water and pure air combined w e without r,i.rihla actiou upon pure iron.” This shows that the electrolytic theory of corrosion is untena-
ble, and definitely confirms the older acid theory. “The experiment was now repeated in every detail, save that the caustic potash was replaced by an equal amount of a saturated solution of barium hydroxide. Although in every other respect the experiments were preciseIy similar, the iron always rusted in the course of an hour or two. This shows that barium hydroxide is not sufficiently powerful t o remove every trace of carbon dioxide from the air in the flask. In other words, its aqueous solution is in equilibrium with a definite, although minute, partial pressure of carbon dioxide. Now Moody employed aqueous barium hydroxide solution as his source of water distilled on to the iron in his apparatus. Consequently, his metal must have rusted had it not been protected by alkali dissolving o u t of the glass tubing containing it. A clear explanation is thus offered for the failure of Dunstan and of Walker, each working independently of the other, to confirm Moody’s results, when a different kind of glass was employed in the construction of the apparatus-a glass which was slow t o yield up any soluble alkali t o liquid mater condensed upon its surface.” No mention is made of conditions which might influence rusting of the metal. The temperature of the metal and water on its surface, and the rapidity with which this water is changed, are conditions which influence the results. These, along with other importa n t conditions, will be taken up fully after describing my experiments. Suffice i t t o say here that my experiments prove t h a t Friend is wrong in his conclusions, and t h a t his experiment was concluded before he had proved anything. The apparatus used in all experiments is shown in Fig. I. I and S are hollow metal blocks, I X I x‘ with s/~’’hole I*/~’’deep. The samples in all cases received their final polishing on 0000 French emery paper. The metal samples are connected by tubes for the circulation of cold water within them, as in Friend’s experiment. The purifying train consists of: A , dilute sulphuric acid; K,, j o per cent. potassium hydrate; K,, 5 0 per cent. potassium hydrate; So, Soda lime; (the tube connecting bottle and train extends just to the bottom of the large rubber stopper); Sy, is a syphon; C1, clamp on purifying train; C,, clamp on syphon; C,, large iron clamp for bolting down the rubber stopper.
482
T H E JOURAiAL OF I i V D U S T R I A L A N D E N G I N E E R l N G C I I E M I S T K.'1
July,
1912
densed water vapor on their surfaces was probably but slightly higher. These conditions were maintained every minute during the thirty-five days.
0bscwations:-At
the end
of thirty-five days not the slightest spot of rust or dis-
coloration of any kind was visible on either the pure iron or the steel. Under the saniples a ring of rust was formed by contact with the rubber stopper. This, of course, must be disregarded. The surface area of the top and four sides of each sample was 7 sq. in. It is remarkable t o think that 14 sq. in. of iron and steel can be exposed for thirty-five days to pure air and water combined without developing the 2 slightest sign of a rust spot. Conclusion:-This proves that iron or steel will not rust in pure air and water combined, provided the temperature of the metal does not exceed 15' C., and provided the temperature of the pure water condensing on its surface does not much exceed 15'; provided, also, that the pure water which condenses on their surfaces is being constantly renewed. Beyond this, the experiment proves nothing. I t simply ilzdicates that pure iron or steel will not rust in pure air and water combined. EXP.No. I, PARTz .
TY
EXP.No. Metals Used-Genuine Open Hearth Steel.
I , PARTI . Open Hearth Iron and ordinary mild
-
Analyses.
C. Mn. S. Genuine Open Hearth Iron'. ............... 0 . 0 1 0.01 0 . 0 2 8 Open Hearth Steel. .. . .. 0.09 0.40 0.080
P.
Si.
0.
0,003 0,003 0 . 0 5 5 0.070 0.006 0.014
The bottle (2% liters) was completely filled yith saturated barium hydrate solution, recently boiled and filtered. The large rubber stopper carrying the four condensing tubes, (with samples attached), syphon, and purifying train, was fitted tightly and bolted down securely by means of clamp C3. No air bubbles were visible in the apparatus. The bottle was placed in a cold water bath and heat applied. C2 was closed, and C I opened. The expanding solution filled the tube connecting the bottle and purifying train. Thus all air was expelled, and the necessity of removing carbon dioxide from the apparatus by mere shaking, avoided. Cr was then closed. Cz was then opened to allow the expanding solution to fill the syphon. The solution was then syphoned off by opening C I just enough to allow the incoming air to bubble through the purifying train a t the rate of four or five bubbles per second. When syphoned to the level indicated in Fig. I, the clamp Cz, on syphon, was tightly closed. C I was left slightly open to avoid undue pressure or rarity of the air in the apparatus. The rubber stopper and all connections were well paraffined. The water in the bath began to boil and cold water was madc to circulate through the glass tubes and samples. Water immediately began dripping from the samples. The experiment was continued thirty-five days. The water in the bath was boiling ever?' minute during the whole period. Likewise, cold water was circulating rapidly through the condensing system. The temperature of the ingoing water was 4 to 5' C. The temperature of the water leaving the system varied from 10' to 15' C. At no time did it exceed 15'. From this the temperature of the metal samples was presumed to bc about 8' to 12' C. The temperature of the pure con-
* Made by Newport Rolling Mill Co.
-4tthe end of thirty-five days the source of heat was removed and the apparatus taken out of the water bath. The circulation of the cold condensing water was continued until the barium hydrate solution had reached room temperature. The condensing system was then shut off, In a few hours the entirc apparatus had reached room temperature, about z z o C. Thc metal samples were left covered with pure water, the flat tops completely, the sides with irregularly distributed patches or globules. This same water was destined to remain on the iron and steel until removed by natural evaporation a t Yoom temperature. What was the effect of allowing the same water (at 22') to remain on the metal (at 2 2 O ) for a protracted period? NC\V conditions had been created in the apparatus.
Observations:-In fifteen hours the steel had developed thirty-three distinct rust spots and all of thesc spots were on areas covered with the pure water. The pure open hearth iron was still spotless. After twenty-four hours contact with the same water a t 2 z 0 the steel contained about 50 spots; the iron was still rustless. At the end of seventy-two hours the water had entirely evaporated from the sides of both metals. The tops were still covered with a thin layer of water. A few new spots had developed on the steel, but no rust had yet appeared on the iron. At the end of eighty hours the water had evaporated
July. 1912
T H E J O L-R.V.4 L OF I-1-L)L-STRI.4 L ili1-D EAYGISEERI\'G
entirely from the tops. Ticso large glariizg spots of rust had d c w l o p c d 011 the f o p of the irott c'ithin tize last riglit itours. More than seventy-two hours contact n-ith the s c i i i i z ii8aicr a t 2 2 ' was necessary t o produce rust on the pure open hearth iron. Less than fifteen hours contact with the ~ n i i i ewatcr produced rust on the steel. i'iiiiclzisiolis:-Iron or steel xiZ1 rust in pure water and air combined if it is given a chance. The tTTo most important factors r h i c h influence the development of rust are: F-irsi, tlic tciiipL~rutiLve o j ilzc metal aizd oj tizc p u r e s;,ater oii i t s xiirjace; secoiid, the rate at whicit the pitre watil- i s r r i i e i i ~ c d ,cltaiiged, 012 thc sitrface of the metal. EXP.KO. 2 , P A R TI The same samples were repolished and Exp. KO.I repeated, except t h a t the apparatus was kept in boiling water for six da?s instead of thirty-five. -lgain, not the slightest sign of a rust spot developed on either sample. EXP.No. 2 , PART2 . .Ipparatus cooled to room temperature as before. Rust developed in almost the same time on both the iron and steel, the iron requiring ninety hours contact with the same water, the steel about sixteen hours. The location of t h e rust spots does not seem to be determined entirely by impurities in the metal, but also by the distribution of the water patches on the surface. This experiment confirmed Exp. No. I in every detail.
EXP. No.
3
Same samples used. 2-A\r potassium hydrate substituted for barium hydroxide. Condensation period, six days. This experiment confirmed Nos. I and 2 , and further showed t h a t either potassium hydrate or barium hydrate may be used irith equal effect. EXP.KO. 4, PARTI . Same samples used. Solution, saturated barium hydroxide. Condensation period, six days. Both metals, as usual, v-ere spotless a t the end of six days. EXP.KO.4, PART2 . The floir of cold \rater through the condensing tubes \vas almost checked. The apparatus was kept in the boiling water bath. The slo~vlydropping water leaving the system shoived a temperature of 5 j ' C. From this I assumed that the temperature of the metal samples and of the pure water on their surfaces was about j 5-60' C. L-nder these conditions the condensation of water on the samples \vas almost stopped. Thus new conditions were created in the experiment. The metal samples were suspended in a hot atrmsphere of air and water vapor, while the temperature of the metals themselves \!-as about j jo C. and the temperature of the almost stagnant pure water on their surfaces n-as about j j o
c.
C)~~rcri'ntioiis.---In one and one-quarter hours rust b e p n developing on the t o p and sides of the steel. I t developed so rapidij- t h a t i t could be easily observed. Small clouds of insoluble matter came up from the metal and could be seen floating through the water. I t had a faint greenish color, Presumably i t was ferrous hydrate. I n two hours the pure iron began t o develop rust. This. too, rusted sc) rapidly t h a t it could be observed. In a few hours the tops of the samples were covered n.ith rust, and the sides were badly spotted.
CHE21IISTRk'.
483
Co~zclztsio~?:-This experiment clearly demonstrates t h a t pure open hearth iron and steel will rust, and rust badly, in pure water and air combined, provided the temperature of the metal and water on its surface is sufficiently high. This', and all preceding experiments, further prove t h a t rust develops much faster on steel (impure iron) than on pure open hearth iron. This is in strict accordance with the electrolytic theory. EXP, KO. j , P.4RT I . Samples, same. Solution, barium hydrate: condcnsation period, ten days. S o sign of rust a t close of ten days. EXP.NO. j , PART2 . I n Part 2 , Exp. No. 4,the condensing water was only partially checked. In this experiment the condensing system was shut off entirely, the apparatus being kept in the boiling water. -4ceordingly, the temperature of the metals and pure water on their surfaces would be higher than in Exp. Eo. 4. From this increase in temperature should result an increased rate of corrosion. Such was the case. I n z j minutes the top of the steel was badly rusted and the sides considerably spotted. I n j o minutes the iron was badly rusted. This further demonstrated the influence of temperature and of stagnant water. EXP.Xo. 6 , PARTI . Different samples used. Analyses. C. hi. S. Genuine Open Hearth I r o n . . . . . . . . . . . . . . . . . 0 . 0 1 0.040 0 . 0 2 6 MildOpenHearth Steel.. 0 . 1 2 0.38 0.040
P. 0.004 0.076
Si.
0.
0.003
0.056 0.009
0.007
Solution, barium hydrate. Condensation period, twelve days. At the close of twelve days not the faintest suggestion of a rust spot was visible on either sample. EXP.No. 6, PART2 . After twelve days the source of heat was removed as in Exp. No. I , Part 2 . The steel developed twenty-four rust spots in twenty hours. I n ninety hours, over forty spots were visible. The iron developed only three spots on top after ninety hours. This rust development took place a t room temperature (about 2 2 ' ) while the same water remained on the metal until removed by natural evaporation. EXP.KO. 7 . Sample same as Exp. No. 6. Solution, potassium hydrate. Condensation period, six days. KO sign of rust, as usual. EXP.KO. j, PART2 . Exp. No. j , Part 2 , repeated to confirm results on different metals. The steel began t o develop rust in 3 j minutes; the iron in one hour. I n one and one-half hours both were badly rusted. EXP.No. 8. Other samples used. Analyses. C.
Genuine Open Hearth Iron . . . . _,.. . . . . . , . . , AcidOpenHearthSteel..
0 01 0.15
Af.
0.03 1 I5
S. 0 027 0.055
P. 0.003 0.105
Si. 0,003 0 009
0. 0.061 0.010
Solution, potassium hydratc. Condensation period, ten days. At the close of ten days not the slightest suggestion of rust
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 N G I N E E R I N G C H E M I S T R Y .
484
k, appeared on either sample. I n this connection it is interesting t o note the high percentage of impurities (especially the manganese) in the steel. EXP.No. 8, PART2. Cooled to room temperature as in Exp. No. I , Part 2 . I n eight hours the steel was covered with rust spots. The iron, as usual, required about ninety hours to develop a few rust spots. Aside from confirming the results of preceding experiments, this experiment further indicates the influence of impurities on the rate of rust development. As we should expect, from the electrolytic theory, increased impurities greatly increase the rate of corrosion. EXP.No. 9, PARTI . Sample, same as Exp. No. 8. Solution, barium hydrate. Condensation period, ten days. No rust developed, as usual. EXP.No. 9, PART2 . Condensing system entirely stopped, as in Exp. No. 5, Part 2 . In fifteen minutes the top of the steel developed rust. I n 2 5 minutes it was covered with rust. The sides were also badly spotted. I n 45 minutes the iron began to develop rust. I n one and one-half hours the top and sides were badly rusted. EXP.No. IO, PARTI . Sample, same as Exp. No. I . Solution, 2-N potassium hydrate. Condensation period, six days. No sign of rust, as usual. EXP.No. IO, PART2. Exp. No. I , Part I , demonstrated t h a t iron or steel will not rust, while the temperature of the metal and water on its surface does not exceed 15' C., and while the surfare mater is constantly changing. In this experiment the object is to determine whether iron or steel will rust while the metal and surface water remain below IO', with the same uater lodging on the metals. This condition was created in the bottle by removing the apparatus from the boiling water, and continuing the flow of cold condensing water through the system. I n a few hours the potassium hydrate solution had reached room temperature. The temperature of the incoming condensing water was 5', the outgoing water 7 Therefore the temperature of the metals and of the water on their surfaces was about 6'. This condition was maintained for thirty days. At the end of this period not the slightest indication of rust was visible. This experiment further shows the great influence of temperature on the development of rust in pure water and pure air combined. It demonstrates t h a t pure open hearth iron or ordinary mild steel will not rust in pure water and air combined, provided the temperature of the metal and water is constantly below 7' C. This has been demonstrated in P a i t I of each experiment, but this experiment further shows that the same water remaining on the metal has little effect if the temperature remains sufficiently low. EXP. No. IO, PART3. After thirty days the cold samples were still covered with water. The condensing system was stopped and heat applied t o the mater bath until the water had reached 60' C. I n two hours after the water had reached 60' the steel was spotted with rust. In four hours the steel was badly rusted. The pure open hearth iron was still spotless. The temperature was increased to 80'. I n thirty minutes the iron hegan to rust. One hour later the top was covered with rust and the sides badly spotted.
-
'.
kP
-Tub, -
-
IQIZ
b - LE:
i
Starting from a low temperature (6O), as in this experiment, and gradually raising the temperature is a n excellent means of determining the temperature necessary to produce rust. It also indirectly shows the solution pressure of the metals a t different temperatures. And in this way I believe we can test in a most reliable way the ability of a metal t o resist corrosion in service. There can be no doubt that metals of high solution pressure will rust much faster than metals of low pressure. By this experiment we can compare these pressures, indirectly. The carrying out of the experiment, however, is not extremely simple. It requires great care and in the hands of a careless operator it might become misleading. I m u s t call a t t e n t i o n t o the w a y in which r u s t dev e l o p e d i n these e x p e r i m e n t s at h i g h e r t e m p e r a t u r e s . At t h e s e t e m p e r a t u r e s r u s t d e v e l o p e d so r a p i d l y that o b s e r v a t i o n s c o u l d be made at different stages. When rust b e g a n to develop, the metal, o v e r a small a r e a , assumed a light grayish color. This gradually gave way t o green, which increased in i n t e n s i t y f o r a f e w m i n u t e s , t h e n gradually changed t o yellow o r
rust color. These o b s e r v a t i o n s s t r o n g l y i n d i c a t e that r a t i n g , i n a pure e l e c t r o l y t e ( w a t e r free f r o m acids, bases, e t c . ) is d u e t o c a u s e s w h i c h have b e e n p r o p e r l y exp l a i n e d as follows: Pure water is dissociated t o a c e r t a i n extent. I t contains dissociated hydrogen and h y d r o x y l ions.
-
For t h e p r e s e n t p u r p o s e let H r e p r e s e n t the h y d r o g e n
-
ion, OH the h y d r o x y l i o n , Fe the i r o n i o n , e t c . , and H, OH and Fe r e p r e s e n t the a t o m i c state of the e l e m e n t . A t the a n o d i c points i r o n p a s s e s i n t o solution,
-
Fe, and at c a t h o d i c p o i n t h y d r o g e n is l i b e r a t e d ; Fe
=
-
H = H I r o n and hydroxyl ions then u n i t e chemically, t o f o r m ferrous h y d r o x i d e ;
-
-
+
zOH = F e ( O H ) , F e r r o u s h y d r o x i d e then r e a c t s with w a t e r and dissolved Fe
oxygen,
+
+
zFe(OH), 0 H,O = 2 F e ( O H ) , , w h i c h is i m m e d i a t e l y c o n v e r t e d t o the h y d r a t e d oxide, 2Fe(OH), I r o n Dissolves
=
Fe,0,(3H20).
in P u r e Water and Oxygen.-If,
ac-
c o r d i n g t o t h e electrolytic t h e o r y , i r o n p a s s e s i n t o s o l u t i o n i n p u r e w a t e r , a h i g h l y polished surface should b e "etched" i n s p o t s d u r i n g corrosion and i t s h o u l d reveal t h e s t r u c t u r e of the metal s o m e w h a t as b y e t c h i n g w i t h d i l u t e a c i d s or o t h e r r e a g e n t s . We k n o w that i r o n passes i n t o s o l u t i o n w h e n i n c o n t a c t w i t h acids. We know that structure is d e v e l o p e d o n h i g h l y polished surfaces b e c a u s e the metal dissolves more r a p i d l y along the b o u n d a r y lines of the grains. Therefore, if i r o n p a s s e s i n t o s o l u t i o n i n pure w a t e r the s u r f a c e will be e t c h e d and s t r u c t u r e developed. P u r e w a t e r will thus a c t as an exceedingly d i l u t e acid. Microscopic e x a m i n a t i o n of both the pure i r o n and steel after e a c h of my e x p e r i m e n t s i n v a r i a b l y showed the true s t r u c t u r e of the m e t a l in many places. Most of t h e surface, of course, was c o v e r e d w i t h rust and
July, 1912
T H E J O U R N A L O F I!VDUSTRIAL
the structure obscured, hut large spots of rust invariably showed the structure of the metal in many places. Fig. 2 shows the surface of the iron before using in Exp. No. IO. Fig. 3 shows the ferrite structure which was developed in several places in the large spot of rust on top of the sample. The grains are large. A portion of the metal which was free from rust was later etched with dilute nitric acid and the same structure appeared. Pig. 4 shows the stecl beiore Exp. No. I O ; Fig. 5 , the structure in places after corrosion. The ferrite grains are small and mingled with pearlite, which accounts for the structure not being as well defined as i t is in the case of pure iron. The small dark areas are pearlite; some of them show lamellar structures under higher powers. It will be observed t h a t the 0 0 0 0 emery scratches have been almost entirely eroded. This proves conclusively that iron is soluble in pure water, free from all traces of acids and containing only dissoIved air. I t further confirms the electrolytic theory since the structure of the metal is developed only in spots-anodic points where iron passes into sollition
nificance. Tie every o w o) niy ~ z ~ c ~ i ~ + t mPart ~ . i . rI ,, six diflerent kinds of iron aiid steel were cxposed f r o m six to thirty-
EZG. 3
x
120
FiO. 4
x 120.
and ?lot the slightest spot of rust developed, not one minute spot on 42 sy. i n . of metal. Compare this with Friend's "herc and there" spots. Why did not a single spot develop in all my experiments? Simply because, in Part I of each, the temperature of the metals and pure water on thcir surfaces was kept tielow a certain fixccl point ( 1 s " C ) . As the
fi7ie days,
a t these points? Yes, partially. But i t was probably due mostly to the increase in solution pressure of the iron at these points, resulting from increased temperature. The surface of each metal contained points of lowest and points ai highest solution pressure, and other points between these extremes, Now, with increasing temperature, the points of highest solution pressure dcveloped rust first while the points of lowest pressure were slowest to rust. Along this line of reasoning, increasing temperature should result in a n increased number of rust spots on the metal. My experiments bore this out completely. Friend gives us no information about conditions in his experiment, but, irom my experiments I am convinced that his "here and there" spots were formed because the temperature was higher than in Part I oi my experiments, or because his samples were so impure t h a t the solution pressure at places caused rust to develop even at low temperatures.
T H E J O U R N A L OF IiL’DUSTRIAL ALYD E A ‘ G I S E E R I S G C H E M I S T R Y .
486
While temperature has great influence on the development of rust in these experiments there are other influences which affect the results. The purity of the metal determines t o a great extent the rate a n d amount of corrosion. This is definitely indicated in m y experiments. Of t h e impurities in steel, manganese seems t o have the greatest influence on the rate of corrosion. My experiments are not exhaustive enough t o warrant any sweeping conclusions regarding the influence of manganese, b u t as far as they go they strongly indicate t h a t this element is a n accelerator of corrosion. This is by no means new. For years, metallurgists and other scientists have held the opinion t h a t manganese is the chief instigator of corrosion. Of this, however, there can be no doubt: The purer the the w o r e it will resist corrosion. I believe I a m justified in drawing this conclusion from m y experiments. Indeed i t follows directly from the electrolytic theory of corrosion, and t h e great majority of scientists hold this opinion. Friend’s apparatus is a n ingenious device for investigating the corrosion of iron and steel in a pure electrolyte, e. g., water containing only dissolved air, free from all traces of acids or bases. Since corrosion in a pure electrolyte is very probably the true starting point, much can be learned b y using this apparatus. The conditions of the experiment are ideal since only the three substances concerned-iron, air and wa,terare brought in contact; no soluble matter coming from glass vitiates the experiment. S o theory ever advanced t o explain the corrosion o f iron has been so satisfactory as the “electrolytic” theory, announced b y Dr. Whitney, and so ably championed b y Drs. Walker and Cushman. Practically all known facts connected with corrosion are explained in the light of this theory. The “acid” theory explains many phenomena of corrosion, b u t it is narrow in its application and, after all, is only a special case under the general electrolytic theory. When carbonic acid enters the pure water electrolyte, i t greatly increases the number of hydrogen ions, and, in addition, pollutes the electrolyte with CO, ions. The solution pressure of t h e metal is thus greatly increased and corrosion is accelerated. Acids are only accelerators of corrosion. They are not the cause. The true ‘starting point of corrosion is the solubility of iron in pure water, its electrolytic solution pressure. This property was given iron b y nature, and with all our controversy we cannot take away t h a t which nature gave.
.
G E S E R 4 L C 0N CL U SI O ? S i
of each experiment confirms Friend’s statement t h a t iron or steel will 7zot rust in pure water and air combined. The failure 01 the wzetals to rust, hoz,evcr, uras sittirely d u e to temperature cofiditiotzs awd to yapidly clha?zgiutg pure water. 2 . Expts. No. I , 2 , 3, 6 and 8, Part 2 , prove conclusively t h a t pure iron or steel w i l l rust in pure water and air combined, provided t h e temperature of the metal and pure mater is not below 2 2 ’ C . , and proI.
Part
I
July, 1912
vided the same water remains for a sufficient time on the metals. 3. Exp. No. 4 , Part 2 , proves t h a t rust is developed rapidly if the temperature of the metals and pure water is about j 5 O C. 4. Expts. Nos. 5 , 7 and 9 , P a r t 2 , prove t h a t still further increase in temperature results in a decided increase in the rate of corrosion. j . Exp. No. I O , Part 2 , demonstrates t h a t the same pure water may remain on the metal for an indefinite period, and no rusting takes place, provided the temperature of water and metal is sufficiently lo^. 6. With m y modification of Friend’s apparatus, equally good results are obtained b y using either barium hydroxide, or potassium hydroxide. Barium hydroxide is to be preferred since i t is not as liable t o render the metal passive. 7. I n general, pure iron or steel will rust in pure water and air combined, free from all traces of acids. The amount of rust produced is a function of the temperature and of the purity of the iron. 8. The “acid” theory of corrosion is untenable. 9. All phenomena observed in these experiments are in perfect harmony with the electrolytic theory. I O . The electrolytic theory of corrosion is further confirmed b y these experiments. THE RELATIVE CORROSION OF IRON AND STEEL PIPE AS FOUND I N SERVICE.’ B y WILLIAMH. WALKER. Received February 5 , 1912.
There are few subjects relating t o the corrosion of metals which have received so much attention, or around which there has centered so spirited a discussion, as the relative merits of iron (meaning thereby wrought iron) and steel. The fact t h a t this matter is one still receiving attention, notmithstanding the great volume of accumulated and available literature, is due t o a number of causes, among which may be mentioned: First, t h a t although the words “iron” and “steel” carry with them a definite idea as t o general methods of manufacture and some of the more easily discernible properties, they convey no idea as t o standards of value. I t is possible t o make very poor iron and very good steel, and i t is just as possible t o make the reverse. Hence when an investigator compared the corrosion of a poor iron with a good steel, he obtained results which favored steel ; when the material under study was the reverse, iron w a s shown t o be the more resistant metal. Second. there is a woful lack of uniformity of conditions obtaining in many, if not most of the experiments which have been carried on for the purpose of comparing resistance t o corrosion. Some specimens were large, some small; some cleaned of scale, others n o t ; some immersed in deep water, others in shallow water; etc., etc. The corrosion of iron is so sensitive t o changing conditions of surface, oxygen concentration, salts in solution, and the like, t h a t only when the most careful preparation is made t o maintain all conditions 1 h b s t r a c t of a paper read before t h e New England \ V a t e r TYorks Association, Dec. 13, 1911.
