1116
T H E J O U R N A L O F 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 Vol.
It will be noticed t h a t while the formation of dicyanodiamide is less with the smaller quantities of Cyanamid used i n t h e fertilizer, yet t h e change t o dicyanodiamide is still very considerable with 50 lbs. per ton. CONCLUSIONS
Further investigations are being conducted, b u t t h e results already secured seem t o warrant the following conclusions, a t least in so far as t h e conditions of concentration, temperature, and moisture here reported on are concerned: I-When Cyanamid is mixed with fertilizer materials containing acid phosphate and 5 t o I O per cent oE moisture, t h e cyanamide content decreases with great rapidity . 2-This change is represented partially by, and in t h e higher concentrations principally by, t h e formation of dicyanodiamide. 3-A given quantity of moist acid phosphate is able t o transform a limited amount of calcium cyanamide. 4-Cyanamid is not affected by dry acid phosphate. 5-Moisture alone is able t o cause the conversion of cyanamide t o dicyanodiamide, but the change is much slower than when acid phosphate is present. Since i t has been repeatedly shown t h a t dicyanodiamide is valueless as a fertilizer material, and, moreover, is toxic t o many plants,l t h e formation of this compound in fertilizer materials seems undesirable. On first thought, it would appear t h a t this conversion of cyanamide into dicyanodiamide could be avoided by employing dry fertilizer mixtures, but this overlooks the fact t h a t when such mixtures are added t o the soil, moisture conditions are a t once provided, and t h e transformation may possibly then take place. Preliminary experiments carried out in this laboratory indicate t h a t , under certain conditions at least, this is t h e case. I t should be noted t h a t these unfortunate reactions between acid phosphate and Cyanamid do not in any sense imply t h a t Cyanamid cannot be successfully used when mixed with other forms of phosphate. I n this connection i t should be noted t h a t the Fixed Nitrogen Research Laboratory of t h e Ordnance Department has called our attention t o the fact t h a t lime nitrogen (Cyanamid) can be mixed with calcined and basic phosphates without the excessive production of dicyanodiamide noted when moist acid phosphate is used. STUDIES ON THE CORROSIVE ACTION OF CHLORINETREATED WATER. I-THE EFFECTS OF STEEL ON THE EQUILIBRIUM: Clz H 2 0 HC1+ HC10, AND OF PRODUCTS OF THE EQUILIBRIUM ON~TEEL By George L. Clark and R. B. Iseley
+
VANDBRBILT UNIVFRSITY, NASHVILLE, TENNESSEE Received July 17, 1920
I n the entire range of the subject matter of chemistry there is scarcely a phase upon which more work has 1
Brioux, Ann. Sci. Agron., [ 3 ] 1 (1910), 241; Hovermann and Koch,
J. Landw., 64 (1916), 317; Pfeiffer and Simmermacher, Landw Vers.-Sta., 90 (1917), 415; Johnson, Tid. Kerns, Farm. Tempi, 16 (1918). 349; Cowie, J . Agr. Sci.. 9 (1919), 113.7
12,
No.
I
IL
been done, and concerning which there is greater deviation of opinion expressed in the literature, t h a n upon solutions of chlorine gas in water. The importance of a thorough knowledge of such solutions lies in t h e fact t h a t so many distinct fields are concerned. T h e physical chemist is interested in all phases of t h e equilibrium established between t h e products of t h e reaction Clz
+ HzO
-
HCl
+ HOC1
and how this may affect the properties of each substance. The bacteriologist and sanitary engineer are concerned with the amount of chlorine i t is necessary t o add t o natural waters in order t o render them potable. The physiologist is interested in discovering t h e effect upon bodily functions and metabolism of t h e introduction of chlorine-treated water into t h e alimentary canal. The fields of t h e engineer and metallurgist converge in considering t h e effect of chlorine water upon the corrosion of iron and steel pipes and upon steam boilers. Applications of this problem have been widely recognized, as is indicated by t h e numerous papers t o be found touching in varying fashion upon some phase of t h e subject. There are two outstanding facts t o be gained from a survey of t h e literature: first, t h e extremely qualitative nature of t h e great majority of these reported observations; and hence, second, t h e wide differences of opinion, even t o t h e point of diametrical opposition, on some of the most fundamental points of t h e chemistry, corrosive effects, and physiological action. It is not within the scope of this paper t o undertake any extensive critique of the work which has been done; the work which may be considered most authoritative will be referred t o later. With ‘these large discrepancies and rather serious omissions in t h e present knowledge of the solution of chlorine in water clearly in mind, this series of researches has been undertaken with t h e ostensible purpose of attempting t o bring order out of an indefinite and unsatisfactory state of a scientific subject, with especial reference t o t h e application t o t h e city water supply of Nashville, Tennessee. Of course, of t h e greatest importance from any stand-, point is t h e question whether or not city water which has been treated with chlorine is physiologically injurious when used for drinking purposes. There is not t o be found in the literature a record of any complete biological research of this kind, though work upon t h e effect of chlorine gas in respiratory processes is extensive and conclusive since t h e war. There is an almost even division of opinion among medical men as t o whether chlorine-treated water is, or is not, injurious in metabolism. One writer finds i t beneficial in t h e treatment of such diseases as infantile diarrhea, and another argues t h a t even very slight traces are injurious t o t h e organism. There must be established, therefore, by careful research upon living organisms just what t h e effect of chlorine-treated water is, and what t h e maximum allowable concentration of t h e gas may be without serious consequences. Before any adequate interpretation of such experiments can be made, however, i t is necessary t o know
Nov.,
1920
T H E J O U R N A L OF I N D U S T R I A L A N D ENGINEERING C H E M I S T R Y
exactly t h e chemistry involved-the compounds formed, and their concentration, activity, and relation t o each other, when chlorine is dissolved in water. This paper is, therefore, preliminary t o more practical work on t h e corrosive action upon living organisms and upon metal containers and pipes,l t h e cause and nature of which may then be adequately and scientifically explained. It involves essentially a study of t h e reaction between chlorine and water t o form hydrochloric and hypochlorous acids; t h e determination of equilibrium conditions and t h e disturbing of t h e m b y time, light, and t h e presence of steel, alkali, and excess products; methods of analysis of t h e products in t h e presence of each other; t h e effect upon t h e passive state arid t h e corrosion of steel; the settling of t h e controversy concerning the existence of “fixed chlorine;” and a study of some new oxidizing properties of t h e solution of chlorine in water. The work t o be described may be subdivided into t h e following heads: (u) The solution of chlorine in water: Methods of analysis Equilibrium in the system Effect of iron on equilibrium Interpretation of results ( b ) Corrosion of iron and steel: Nature, and extent Interpretation Secondary effects T H E SYSTEM
ClrH20-HOC1-HC1
-
It has been long known t h a t when t h e reaction L Clz H20 HCI HClO
+
+
takes place out of contact of light and in t h e cold, equilibrium is established. There is, however, very great disagreement as t o t h e exact conditions and concentrations which determine t h e equilibrium, and t h e constant has never been determined. Much better known is t h e action of sunlight upon chlorine water, i n which t h e following monomolecular reaction proceeds: J ZHOC1zHCl+ 0 2 Also well known is t h e extensive work on t h e reaction comprising t h e Deacon process ZHzO ZCIZ 0 2 4HC1 which is carried out a t high temperatures in t h e presence of bricks soaked with CuClz as catalyst. The equilibrium constant is about 2.4 a t 430’. A fourth equilibrium is t h e following: J HCIOs sHC1 3H2O 4-3C12 which has been fully investigated by Sands,2 Foerster,? Luther a n d McDougall,4 and Olson.5 The latter determined an equilibrium constant of 4.3 X IO+ at 3 6 4 O % . . , and also found t h a t t h e effect of hastening t o equilibrium with iron as catalyst was unsatisfactory .
+
-
+
+
1 Lillie [Science, 50 (1919), 2591 has shown that there are fortunately some close analogies between many of the properties of metal surfaces and living protoplasm. 2 2.p t y s i k . Chem., 50 (1904), 465. * J . Prakt. Chem., 63 (1901). 141. 4 2. Physik Chem., 62 (1908), 199. 6 J . Am. Chem. SOC., 42 (1920). 896.
1117
A N A L Y T I C A L METHODS
Perhaps t h e greatest difficulty in establishing definite relationships in t h e first of t h e four above equations, with which this entire series of studies is concerned, is t o be found in a method of analysis b y which i t would be possible t o determine chlorine, and both hypochlorous and hydrochloric acids, in t h e presence of each other. The usual method for thus determining chlorine and hypochlorous acid in t h e presence of each other takes hydrochloric acid into account not a t all, and thus introduces large errors. The determination is based upon t h e following reactions : L HOCl 2NaI NaCl NaOH IZ L Clz zNaI zNaC1 IZ
+ +
-
+ +
+
One mole of HOCl sets free one mole of 1 2 and produces a t t h e same time one mole of NaOH, while t h e chlorine simply sets free an equivalent amount of iodine. After neutralizing t h e alkali with an excess of hydrochloric acid and determining t h e iodine by titration with sodium thiosulfate, t h e excess of hydrochloric acid is titrated with standard alkali solution. The NaOH produced by t h e action of hypochlorous acid upon t h e iodide obviously requires half as much of, say, 0.01 N acid, for neutralization as is required of 0.01 N Na2S203 solution t o react with t h e iodine set free b y t h e action of t h e hypochlorous acid. From this is calculated how much of t h e NazS2Oa is required for t h e free chlorine. It is clear, therefore, t h a t this method depends entirely upon t h e fact t h a t t h e standard acid added is t h e only HCl present, where in reality there is in addition a n indefinite amount of HC1 formed in t h e reaction Clz H20 HOCl HC1, which would neutralize some of t h e NaOH formed when NaI and HOCl react. This fallacy has not been pointed out before. It is impossible, therefore, b y t h e use of one sample t o determine Clz, HOC1, and HC1 in t h e presence of each other. Most of t h e preliminary experimental work of this paper was concerned with various possible analytical devices, some of which may be briefly described as follows:
+
( I ) DETERMINATION
OF
TOTAL
+
ACIDITY
B Y USUAL
METHOD-The difficulty in this method lies in t h e use of indicators in t h e presence of chlorine and hypochlorous acid. Methyl orange, rosolic acid, and phenolphthalein were tried out in order t o determine which would be best in t h e titration for total acidity. It was found t h a t neither of t h e first two would produce any stable color in t h e presence of Clp and HOC1, no matter how much base or acid was added. It was found t h a t phenolphthalein would give a stable color and definite end-point, provided t h a t a t least I cc. was used. Furthermore, t h e end-point depended upon whether this relatively large amount was added drop by drop, or poured in as rapidly as possible, with best results seemingly from t h e latter method. As an illustration of t h e lack of exactness of this method, the following d a t a are quoted on t h e number of cc. of 0.01 N NaOH required t o neutralize successive z j cc. portions of the same sample: 17.5, 19, 16.2, 15.5, 1 6 . 7 . T h e
T H E J O U R N A L O F 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 Vol.
1118
00
reactions between t h e indicator and chlorine and hypochlorous acid led t o results, therefore, which were a t once nonconcordant and doubtful, unless exactly t h e same method of titration was used throughout. Attention is directed t o t h e fact t h a t methyl orange cannot under any circumstance be used, though its use is advocated by such an authoritative text as Treadwell. ( 2 ) R E M O V A L O F H Y P O C H L O R O U S AcID-?vlethods dependent upon t h e removal of HOCl were not successful because of the incompleteness of the reactions. For example, a polution was shaken up with pure mercury in order t o test t h e reaction ZHg
+ 2HOCl
&
HgO.HgCl2
I
1
i
+ 2HOC1
2AgCI
Page 655.
2
Comfit. rend., 168 (1919), 1114.
C C 0 0 ~ N O N h 0 C W wm
-O *h 880CCOhO
88803c80c8~
oo~ooo~6ooo W
00
n a
16300
m
3 0
ooc 0 80 0 0 0 9? 0.90 0 0 0 0o000c~occc
00 0 0 0 0,0 0,0 0 0,N, 0, oCo00NONh0c W a m
m h 4
R%O008?qF NONOCOCNNCO
-
0
+ HzO.
+
-
W 3 N 00
iD
M
Urn Lo
g8000Cr\13
OO NW 0
3c
0,t"
0 C 0 0 0 0 C C ~ - 0
0000
2*
d o d c o "*-. ko&&do -?e W N C
0 2
and the decomposition in t h e presence of cobalt nitrate solution, with oxidation of t h e cobalt t o higher valence, were found entirely unsatisfactory as quantitative methods, since both chlorine and hydrochloric acid are also involved. (3) R E M O V A L O F cm,onINE-One of the most promising analytical methods seemed t o be found in t h e removal of chlorine by bubbling air through the solution and collecting t h e gas in another vessel containing sodium iodide solution, t h e HOCl remaining behind. de Mallman2 reports t h a t experiments upon t h e elimination of chlorine by a current of air showed t h a t free chlorine is totally removed in j min. from j o t o I O O cc. solution, and t h a t hypochlorous acid in t h e small quantities present is little, if a t all, affected by aeration. Unfortunately, no data were cited t o prove this. These experiments and others t o be found do not preclude the possibility, however, of large errors which might arise from difficulties in manipulation, t h e existence of "free" and "fixed" (possibly hydrated and inactive) chlorine, appreciable decomposition of hypochlorous acid with formation of chlorine, and a minimum concentration below which i t is impossible t o remove alzy chlorine from chlorine water even in I j min. As a matter of fact, as will be fully shown, the present researches have met with all of these contingencies and it is therefore impossible t o make use of this method alone in determining chlorine and hypochlorous and hydrochloric acids together. It was found possible by careful work t o remove t h e difficulties in manipulation so t h a t , as the following tables will show, i t was repeatedly found t h a t the sum of t h e two titrations, after air was bubbled through and t h e gas collected in a second solution of NaI, checked exactly t h e total amount of Na2S20arequired by t h e solution 1
0?000900910
e
T h e oxychloride HgO.HgC1, first formed is immediately decomposed by t h e HC1 present, and HgC12 passes into solution, leaving a precipitate of yellowbrown HgO. Subsequent treatment with X a I and titration with Na2S203for the determination of chlorine gives highly discordant results. Exposure t o strong sunlight in order t o decompose HOC1, of course, throws no light upon the relationship between C12 and HOCl since both disappear a t a monomolecular rate. The reaction AgZO
2 ,02 0 0 0
No. I I 0 C 0
-e
88
c 0 0 0 c 3 ~ &
00
12,
00
-6
0 88 OTF?
LO
a000
a+
v
n
0
32
LO
G * *c G d c Nn N6 &* -& o o W N C
8s
4
"E *
N-
urn g80000Nc
DO 0003000cc-c
Lo 2 2 WC? C00D00N^ ? 00?????973 ~ o w c ~ m ~
ON0
00
NC=C
**
069
t.*
"e-
m
m
~
c
bi
.. .. .. ..
.. ... ... .. .. .. .. .. .. .
.. .. .. ..
.. .. .. ..
.. .. .. ..
.. .. .. ... ... ... .. .. .. .. .. .. . . . ... ... ... . . . .. .. .. ... ... .. .. ..
.., ... ...
.,, .,
..
. ... ... .. ., ,.
... ... ...
... ... ... . . . . . .. .. ... ... ... . . . . . . . .. .. ... ... ... .. .. .. .. .. .. .. .. .. . ... ... . ) . u ... ... :, :.u .. .. . .. ^ , uii . .
.., ... ... . . .
,
.
, .
... ... ...
.., ... ... . . .
.,, ... ... . . .
., . .., .. , ... ., I
. . ... .. . ... .. . .
.
. . ... .. . ... .. . .
.
u . :
J : .
.. ..
. ...
. ... .
. ... .
. ... .
. ... .
Lo
: : : : I %
.. .. .. .. .m. , .' vW, .. .. .. .. ... ... .. . . . . . . . . . . . . . . .
:. ..
v1
..: ..: ..: ..: ..: ..: ..: '': N3A . .. .. ... ... ... ... ... ... ... _ ... ... ... ... ... ... ... ... ... u .. .. .. .. .. .. .. .. . .. .. ,. .. . .. ., .. .: b i :,Li : : : : : : : % . a , ... . . .-, .. ^. .. . . ' .. .. .. .. ._ k
h
Nov.,
1920
T H E J O U R N A L OF I N D U S T R I A L A N D ENGINEERING C H E M I S T R Y
before air was passed through. These points will be considered in detail in t h e interpretation of t h e experimental results, where i t is shown t h a t i t is impossible t o establish any correct relationship between Clz and HOCl, and t h a t such work as de Mallman's is distinctly in error unless properly interpreted. A P P R O V E D METHOD-The method finally adopted as b y far t h e most accurate depends essentially upon the determination of HOCl by means of arsenious acid, and a combination of this result with Methods I a n d 3, together with t h e usual method for chlorine a n d hypochlorous acid in the presence of each other, from which t h e concentration of all three products may be calculated. The arsenious acid method, depending upon oxidation t o arsenic acid b y hypochlorous acid, is briefly described in such texts as Treadwell; but no work has been done on the effect upon its accuracy of the simultaneous presence of Clz and HOC1. The actual titration is remarkably simple, definite and rapid, and the results have been found t o be capable of repeated reproduction from the same solution, since there is no possibility of secondary reactions with indicators. I n order t o test out t h e accuracy of this method for determining HOCl alone and in the presence of the other equilibrium products in the solution, the following experiments were perf or med. A solution was carefully shaken for several minutes with freshly precipitated mercuric oxide. The oxide was dissolved b y t h e HC1, but t h e HOCl, being too weak t o react, was left free in t h e solution, which was then titrated with arsenious acid. I n a second test, some of the solution was shaken u p with pure calcium carbonate, which reacted only with t h e hydrochloric acid, This solution was then titrated. It is a singular fact t h a t in both of these cases t h e results checked exactly with those of a straight titration for HOCl in the presence of both Cla and HC1. The titration must, of course, be carried on very rapidly in order to obtain best results. following t h e changes in conE F F E C T O F TIME-In centration of Clz, HOCI, and HC1 with time, in the absence and presence of iron, the following process was adopted. T h e solution was very carefully made up in order to assure .uniform composition and kept in tightly stoppered bottles in the thermostat a t 2 j '. Three portions of 50 cc. each were removed a t the same time; one was titrated for HOCl by means of 0.01 ilr arsenious acid solution; the second was acidified with 2 5 cc. 0.01 N HC1, NaI added, and t h e iodine liberated b y both C12 and HOCl was titrated with Na2SzO3,after which the excess acid was titrated with' 0.01 N NaOH, using phenolphthalein as indicators; the third was placed in a Drechsel wash bottle, connected with another similar bottle containing S a 1 solution, and air bubbled through the original solution for 5 or I O min. in such a way t h a t the evolved gases would pass through the XaI solution. Both were then titrated with sodium thiosulfate. I n Table I are given the results of the best single series of determinations. Part A shows the results with the solution alone, and Part B in the presence of
1119
low-carbon steel. The solutions in both cases were exactly t h e same, and the three series of titrations on both solutions were made a t the same time, once every 24 hrs., using standard solutions which had been repeatedly checked in order t o assure accuracy. A great many relationships of rather remarkable nature are a t once apparent from an examination of this table, of which only the more important will be specifically pointed out. Part A will first be considered. As an indication of the accuracy and soundness of the method involved and a justification of the interpretation of results, i t is noted t h a t the results in several lines check perfectly; e. g., Line j giving t h e back titration with NaOH corresponds with Line v which is the same titration on a different solution after the air-bubbling process t o remove chlorine. Lines g and u for chlorine also agree. Another outstanding relationship is t h a t final equilibrium is reached after 48 hrs., and the amounts of chlorine, h y p o c h l o r o u s a c i d , and hydrochloric acid remai% constant for the duration of the experiment, 1 2 0 hrs. There is no reason t o doubt the fact t h a t this equilibrium would remain indefinitely in the dark bottle, if conditions were kept constant. This point of equilibrium represents a concentration of 0.008 mole HOCl per liter, 0.00012 equivalents or 0.00006 mole of chlorine per liter; and 0.003 j mole HC1 per liter, This gives an equilibrium constant of 0.467 a t 2 j 0 , considering the concentration of HzO as constant and t h e acids completely ionized, which is checked surprisingly well, starting from entirely different concentrations (see Table 11). I t is a matter of considerable surprise t h a t so relatively great a concentration of HOCl is found, but i t is verified by such experiments as titrating before and after exposure of the equilibrium solution t o strong sunlight for some time, during which the HOCl is converted t o HCI. TABLEI1
WITHOUT STEEL
A-SOLUTION
Clear Bottle -Aei
001 N NazSzOa
0.01 N
NazSzOs
0.01 N
cc.
cc.
NaOH cc.
cc.
0.0 24.0 48.0 72.0 96.0 120.0
42.70 42.60 38.40 36.50 34.50 31.00
50.00 46.00 41.00 39.00 35.50 30.50
23.50 17.00 18.00 19.50 20.00 20.00
0.0 24.0 48.0 72.0 96.0 120.0
46.60 49.00 46.00 43.00 42.00 34.00
Time Elapsed Hrs.
0 01 N As203
-ation---
0.01 N NazSzOa
001
N
NaOH
41.00 39.20 33.50 34.00 31.00 26.50
(2) cc. 6.40 7.80 7.50 4.00 4.50 3.50
40.50 41.50 37.50 38.00 37.00 37.00
14.00 8.50 10.50 6.40 6.50 4.00
17.80 17.70 17.70 19.00 17.50 17.50
35.00 21.00 10.00 2.50 0.00 0.00
4.00 2.00 1.50 0.00 0.00 0.00
10.00 18.50 26.50 35.50 35.50 35, SO
31.00 20.50 10.00 2.50 0.00 0.00
2.50 3.50 2.00 0.00 0.00 0.00
9.00 15.50 23.50 31.80 32.00 32.00
(1)
cc. 18.00 18.00 19.00 19.50 21.00 20.50
Dark Bottle
B-IN
PRESENCE O F STEEL
Clear Bottle 0.0 24.0 48.0 72.0 96.0 120.0
30.50 18.00 5.00 0.00 0.00 0.00
34.50 26.00 14.00 2.50 0.00 0.00
0.0 24.0 48.0 72.0 96.0 120.0
30.50 20.00 7.50 0.00 0.00 0.00
33.50 24.00 12.50 2.50
10.00 18.00 25.50 35.00 35.20 35.20
Davk Bottle
0.00 0.00
10.00 '
15.50 24.00 35.00 35.00 35.00
Another observation which' may be made on Part A of Table I is the formation of chlorine by decomposition of HOCl with air. After 48 hrs. this total
I120
--
T H E J O U R N A L O F 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 Vol.
amount of HC1 remains constant (Lines j and v ) a n d also the concentrations of HOCl and Clz (Lines d and i, and e and Y ) , but the amount of chlorine which i t is possible t o remove by bubbling air through the equilibrium solution becomes gradually less. By comparing Lines b, p , q , s , and t , it is a t once apparent t h a t HOC1, a t first decomposing t o form chlorine t o the extent of 16.5 per cent, gradually becomes more stable towards the action of air as time passes, until after 1 2 0 hrs. only 3.75 per cent is decomposed t o form chlorine. Just what is the explanation of this anomaly is difficult t o say. It may be caused by hydration or polymerization, but most likely is connected intimately with the formation and existence of “fixed” chlorine, which will be presently discussed. This same result has been observed with original solutions of all concentrations, but t h e more dilute t h e solution, the more stable i t is, It is thus possible t o prepare a solution of such concentration t h a t it will be impossible t o remove any chlorine by the bubbling of air, either as the free gas or by decomposition of HOC1. Solutions which require a maximum of 2 2 cc. 0.01N NatSzOa per 50 cc. in determining both hypochlorous acid a n d chlorine will not give even from the beginning the slightest trace of chlorine in the second bottle. If the solution is strong enough t o require 2 4 cc., i t is possible t o remove enough chlorine a t once t o require a fraction of a cc. of NazSz03for titration, b u t in 2 4 hrs., though the titer of the original solution is still t h e same, no chlorine is driven over the series of experiments shown in t h e table the concentration is over twice the minimum for appearance of chlorine, and i t is easy t o see why there should be some decomposition of hypochlorous acid. As chlorine is removed, of course t h e equilibrium is displaced in such a way t h a t HOCl and HC1 react to form more chlorine. But HCZ remailzs constalzt, which would, therefore, point t o t h e fact t h a t t h e greatest factor seems to be the increasing concentration of “fixed” chlorine, which cannot be re; moved from solution b y aeration. I n reality, therefore, the decomposition of HOCl is apparent only and indeterminate. This is even more clearly shown when iron is present and all the HOCl disappears. Part B of Table I shows the extremely interesting effect of iron upon the equilibrium just discussed. Some time after 2 4 hrs. all t h e hypochlorous acid disappears, and after 7 2 hrs. the chlorine disappears, forming hydrochloric acid, which thereafter remains constant in concentration, showing t h a t finally only one equilibrium is involved :
+
-
1
+
FeCla 3H20 Fe(OH)3 3HCI Lines j and v again show remarkable agreement.
It is t o be observed t h a t the ratio of concentrations of HOCl t o Clz is a t such a value after 2 4 hrs. t h a t aeration apparently causes . complete separation; for Lines b and 0, and g and Y show this. At the next reading, however, HOCl has disappeared, giving n o apparent trace with arsenious acid, but 8.g cc. 0.01 N NazSz03 are still required for the original solution, which can be only chlorine. However, no length of aeration will serve t o remove any of this chlorine from solution. This entirely disappears in another 48 hrs.,
12.
No.
11
with attendant increase in hydrochloric acid. The Arrhenius theory of active and passive molecules in equilibrium with each other might be very well applied here, b u t t h e uncertain catalytic effect of H+ ions, the formation of hydrogen during reaction with iron, and various possible effects of increasing concentrations of ferrous a n d ferric salts introduce too many complications. Another plausible explanation of fixed chlorine would be hydrated molecules. T h e pressure of chlorine gas above a solution is proportional t o both the concentration of unhydrated and hydrated molecules, b u t t h e proportionality factor is quite different.‘ There is undeniable evidence of t h e existence of hydrated chlorine in solution. Now, a t equilibrium, simply expressed: C12
+ HzO x
-
C12.H20
CHiO = CC1z.H.J0
cClz
But C H ~ O may be considered constant, and hence: CClZ
cClz.Hz0 = KIor cc1, = KICCI~.HZO
The distribution ratio of unhydrated molecules gives by Henry’s law P = K’Cci2
(where P is pressure of gaseous chlorine). Combining with the above: P = K’KICC~~.H =~KO& I ~ . H ~ o Hence K’ and Kz may be greatly different. Furthermore, t h e equilibrium is constantly being disturbed because of the existence of another: 1 Clz HzO HCl HOCl Under certain conditions of concentration, therefore, a final total effect may be represented by:
+
Clz.HZ0
-
-
1
HC1
+
+ HOCl
The gaseous pressure would therefore be regulated t o the hydrated molecules, and the fugacity of unhydrated molecules would be negligible. This is merely another way of saying, as de Mallman‘ does, t h a t chlorine is fixed in a quantity proportional t o t h e ratio C12 : HzO, and decreasing with increased ratio and also with increasing acid concentration. If we therefore consider as “free” chlorine only t h a t which can be removed by aeration-a sort of excess above t h e proportionality t o the ratio Clz : HzO-de Mallman’s generalities agree with these experimental results. The impossibility of removing chlorine gas completely from solution in a short time is verified by the work of Olson,’ just published, on the equilibrium in chlorinetreated water a t 91’. This author says t h a t it may be removed iiin a few hours,” as contrasted with t h e 5min. periods of de Mallman. Olson, however, does not take into consideration any possible influences which might cause the chlorine t o disappear in a few hours, such as formation of the acids, without being removed by aeration. THE C O R R O S I O N O F L O W - C A R B O N STEEL B Y C H L O R I N E TREATED WATER
Table I11 discloses several interesting relationships. Low-carbon steel bars of similar composition were
’
L O C . Cit.
NoY., 1920
THE J O U R N A L O F 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 TABLE I11
Kind of Water and Amount of Chlorine River 0 Clz.. River 2 p. p. m. Clz..
+ ............................ + ..................... River + 5 p . p. m. Clz.. ..................... River + 10 p. p. m. C h . . .................... ..........................
Reservoir f 0 Clz. Reservoir 2 p. p. m. C h . . . . . . . . . . . . . . . . . . .
+
Re'servoir f 5 p. p. m. Clz.. Reservoir
.................
+ 10 p. p. m. C h . . ................
Distilled -1- 0 Clz.. .......................... Distilled 1- 2 p. p. m. Clr..
..................
Distilled
- 5 p. p. m. Clz.. ..................
Distilled
T
10 p. p. m. Clz.. .................
Diameter Length No. In. In. 1 0.501 3.031 2 0.501 3.14 3 0,501 3.14 4 0.501 3.14 5 0.501 3.14 0.501 3.125 6 7 0.501 3.14 8 0.501 3.14 9 0.339 3.094 10 0,339 3.156 11 0.339 3.156 12 0.339 3.125 0.339 3.047 13 14 0.339 3.156 15 0,339 3.125 0.339 3.125 16 17 0.339 3.125 0.339 3.156 18 19 0.339 3.125 0,339 3.125 20 21 0.339 3.156
placed in solutions of different kinds, and kept in sealed bottles in t h e light or dark for I O O days. The adhering rust was then carefully removed without abrasion of the unaffected metal, dried a t IOZ', and reweighed. The loss in weight could then be determined; though more pertinent for purposes of comparison are t h e results in the weight lost per square inch of original surface. I t is a t once apparent t h a t greater loss has been sustained in the light t h a n in t h e dark, due of course t o t h e greater completeness of reactions involving t h e decomposition of HOC1 in light, t o form HC1 and oxygen. I t is also evident t h a t , considering each type of solution separately, rusting or oxidation usually increased with increase in chlorine content, and t h a t corrosion increased in t h e order: river, reservoir,l distilled water. The least rusting effect is shown by reservoir water containing no chlorine, in t h e following ratio: reservoir 0.0666, river 0.0771, distilled. 0.0994. For t h e solutions containing chlorine, the least active is river water with I O p. p. m. chlorine in the dark, a n d t h e most active is distilled water with the same concentration of chlorine in light. This wide difference is a clear indication of t h e profound effect of dissolved electrolytes upon t h e chlorine equilibrium a n d upon t h e iron oxidations. Solutions having apparently t h e same effects are: 2 parts in 0 part in 10 parts in 2 parts in 2 parts in light
river water, light, and 10 parts in distilled, dark river water, light, and 10 parts in river, light river water, light, and 2 parts in reservoir, dark reservoir water, light, and 10 parts in reservoir, light distilled water, both light and dark, and 5 parts in distilled,
Because of t h e extremely great influence of variations of t h e structure of t h e steel, rigorous conclusions are not, of course, justified, aside from t h e general points already considered. After t h e rust coating was removed a t t h e end of t h e experiment, t h e steel bars presented a peculiar spotted appearance, varying from one specimen t o another in t h e ratio of dark area t o bright. T h e corrosion clearly took place on t h e bright areas, while the dark areas presented very largely t h e appearance of t h e original bar with a very tenacious layer of oxide. This differential corrosion in a solu1
tanks.
Reservoir water is simply the river water treated with alum in settling
Surface Ss. In. 4.77 4.94 4.94 4.94 4.94 4.91 4.94 4.94 3.29 3.36 3.36 3.33 3.25 3.36 3.33 3.33 3.33 3.36 3.33 3.33 3.36
Weight Weigh! before Being after Being Loss in Acted on Acted on Weight Grams Grams Grams 76.3380 75.9700 0.3680 76.6648 76.2920 0.3728 78.1100 0.4150 78.5250 78.2050 0.3876 78.5926 78.4000 0.3923 78.7923 77.9700 0.3565 78.3265 77.6500 0.3765 78.0265 78.5790 78.2500 0.3290 36.5230 0.2505 36.7735 37.5197 37.2300 0.2897 37.1500 37.4256 0.2756 36.6960 0.3015 36.9975 36.4015 36.1260 0.2755 37.0060 0.2908 37.2968 37.0810 36.7500 0.3310 37.1517 36.8100 0.3417 36.9910 0.3400 37.3310 38.0679 37.7430 0.3249 36.8830 37.2250 0.3420 36.930d 37.2095 0.2795 37.1720 0.3945 37.5665
Loss in Weight Bottle Per cent Keot in Light 0.48 0.47 Dark Light 0.53 0.49 Dark Light 0.50 0.45 Dark Light 0.48 Light 0.42 0.68 Dark 0.77 Light 0.74 Dark 0.82 Light 0.76 Dark Light 0.78 0.89 Liaht 0.92 Dirk Light 0.91 0.86 Dark 0.92 Light 0.75 Dark 1.06 Light
II2f
Weight Lost per
S a . In. 0.0771 0.0754 0.0840 0.0784 0.0794 0.0726 0.0762 0.0666 0.0764 0.0862 0.0820 0.0905 0.0878 0.0865 0.0994 0.1026 0.1021 0.0966
0.1026 0.0839 0.1174
tion of uniform constitution in which there was uniform contact can be accounted for only by t h e heat treatment of t h e original steel, which was probably cold-rolled, so t h a t electrolysis resulted. There was no evidence of any marked pitting due t o segregation in t h e steel structure; hence the effect must be due t o heat treatment, This very delicate test of t h e uniformity of steel was even more markedly shown by t h e following experiment. Two highly polished steel bars from the same piece and of t h e same size were suspended by means of platinum wire in a solution of chlorine in water. One of t h e bars had been previously rendered passive by immersing in fuming nitric acid, while the other was left active. After 5 min. small bubbles of gas were apparent on the surface of t h e active bar, and in 1 5 min. there appeared quite sharply the peculiar mottled condition just described, in which the bright and dark areas were extremely well defined, though the differentiation gradually faded somewhat. The passive bar presented no such phenomenon, though corrosion ultimately set in t o much t h e same extent as with the active bar. The general average effect of t h e city water of Nashville upon low-carbon steel containers, such as pipe, is t o be found between Nos. g and I O in Table I11 (reservoir water containing on the average z p. p. m. of chlorine), or a corrosive effect of perhaps something over 0 . 7 per cent loss in I O O days. This supposes t h a t there is t o be found no great segregation, which would result in much greater local corrosions, and t h a t outside electrolytes in ground water are not in t h e action. As a matter of fact, however, the presence of strong local circuits about underground pipes will result in a much greater and much less uniform corrosion t h a n is t o be observed in this study made under t h e mosb ideal conditions. Flow of water tending t o remove adhering rust and presenting new surfaces will also have a large effect. However, in this case of water flowing through a pipe continuously, the electrical conditions are very much more uniform t h a n outside t h e pipe, and rusting from t h e inside is always negligible compared with t h a t on t h e outside, where there may be intermittent exposure t o water. It has been further confirmed t h a t iron subjected t o
I122
T H E J O U R N A L O F 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 Vol.
strain or uneven treatment corrodes much more readily t h a n t h a t which has been treated uniformly, t h e strained p a r t being electropositive t o t h e unstrained. In addition t o t h e differential corrosion due t o cold rolling of t h e bars, there was indication of small grooves parallel t o the long axis, due t o drawing and resultant minute differences in mechanical strain. NOW,in order t h a t rust may be formed, t h e metal must first go into solution and hydrogen be given off in t h e presence of oxygen or certain oxidizing agents, t h u s presuming electrolytic action, as every metallic ion t h a t appears a t a certain spot demands t h e disappearance of a hydrogen ion a t another. Iron possesses a specific electrolytic solution tension with regard t o a solution of chlorine in water, so t h a t solution results until the osmotic pressure coufiterbalances t h e solution pressure. This latter is influenced by segrdgation, crystal form as related t o heat treatment, impurities, temperature, a n d , of course, t h e strength of t h e chlorine solution, and t h e relative quantities of Clz, HOC1, HC1, and, under some conditions as previously shown, hydrated chlorine. I t scarcely requires mention t h a t in steel different portions of t h e same bar, even a very few centimeters apart, possess different solution tensions and behave like different metals if immersed in an ionized solution. I n the consequent electrolysis, the ferrous ions liberated a t the anodes undergo secondary chemical changes in t h e presence of such activating agents as chlorides or dissolved oxygen or chlorine itself-the latter two of which act in t h e very important role of depolarizers-resulting in hydrated ferric oxides, which are formed with a speed of reaction infinitely great in comparison with t h e other processes. SUMMARY
Rusting occurs in distilled water without chlorine because of t h e presence of dissolved oxygen, which is of course present in greater quantity t h a n in water containing dissolved electrolytes. I n t h e presence of
I-
PROFESSOR O F PHARMACOLOGY, UNIVERSITY OF WISCONSIN, MADISON, WISCONSIN THE SERVICE TO BIOLOGY AND MEDICINE O F STUDIES ON OXIDATION
All biological science rests on a tripod of physics, chemistry, and a residual leg of biology which includes all the factors which cannot a t present be reduced to the more exact sciences. So intimate is the relation of chemistry to biology and medicine that it is inconceivable that any great advance could be made in chemistry without being reflected in the superimposed structures. I should like to point out the effect on these subjects of two great advances in chemistry. Modern chemistry may be said to have begun with the overthrow of the phlogiston theory of combustion by Lavoisier and the establishment of the mechanism of combustion and oxidation. With this, chemical physiology became a possibifity. Oxidation is the only energyyielding process known to occur to any extent in the body. 1 Presented a t the General Session of the 60th Meeting of the American ChemicalSociety, Chicago, Ill., September 6 t o 10, 1920.
No.
11
chlorine, the most active agent is of course HOC1, which, it has been shown, disappears first from t h e solution. There is no question but t h a t t h e rust is formed rapidly during t h e disappearance of HOC1, and then practically stops when t h e concentration of H C l becomes constant, as shown in Table I . This indicates a n equilibrium: FeC13
+ 3H20
-
Fe(0H)g
+ 3HCl
T h e presence of FeCL in solution is of course easily verified. It has been shown t h a t t h e reaction HzO
+ Clz
+ HC1
I
HOCl
proceeds only very slowly and comes t o equilibrium, t h e constant of which has been roughly determined. It has further been shown t h a t iron has a n apparent catalytic action on t h e reaction, resulting in t h e disappearance of both chlorine a n d hypochlorous acid. This has been verified by t h e very recent work of Olson1 on t h e same reaction, carried out a t 91' so t h a t HC103 would be formed. I t is pointed out in t h a t work t h a t in order t o act as a catalyst t h e iron must of course exist in two stages of oxidation, a n d Fe++ must reduce ClOT ion faster t h a n C1= does, together with t h e necessity for a fast enough reduction of Fe+++ by C1- ion in order t o maintain an effective concentration of Fe++. I n deciding whether or n o t t h e iron has actually acted as a catalyst in materially hastening t h e final attainment of the equilibrium
+
+
HOCl HC1, Clz Hg0 i t is a t once apparent t h a t , in the presence of chloride, t h e iron is almost completely in the form of t h e ferric ion, and hence the effective concentration of ferrous ion which can react with C10- is negligibly small. The effect is, therefore, practically entirely chemical rather t h a n catalytic in nature, since t h e equilibrium is entirely destroyed a n d two of t h e members disappear. 1
J . A m Chem. SOC.,4 1 (19201, 896.
ADDRFSSES AND CONTRIBUTED ARTICLE3
CHEMISTRY'S CONTRIBUTION TO THE LIFE SCIENCES' By A. S.Loevenhart
12,
I
The hydrolyses and hydrosyntheses which occur in the body are practically isotherrnic reactions. Lavois'er was not a physiologist, but the relation between oxidation and respiration is so very close that Lavoisier associated with himself the physiologist Sequin, and in their joint article they stated that the bronchi exude a hydrocarbonous fluid and that this is burned in the lungs. They believed that vital oxidation is exactly similar to combustion, t h a t oxidizable material in the body is brought to the lungs and there burned in contact with the oxygen inhaled, the lungs acting as a sort of furnace. Spallanzani first showed that all tissues absorb oxygen and produce carbon dioxide, and through the work of many men it was established that oxygen is absorbed by the blood in passing through the lungs and carried to the tissues where oxidation really occurs. In the effort to determine why such stable substances as proteins, carbohydrates, and fats are burned in the body a t a temperature of 37' the group of catalytic agents known as the oxidizing enzymes or oxidases was discovered. The work on the digestive enzymes and the oxidases proved to be the greatest stimulus to the study of the whole subject of catalysis, which has become so immensely important in every phase of chemistry. It is obvious that it was necessary to understand the nature of
