UC-NRLF B 2 VS7 5fll OJ EXCHANGE =*= EXCHANGE IUL 6 ! '* 16 The Passivification of Iron by Nitric Acid A THL5I5 SUBMITTED TO THE FACULTY OF THE LELAXD STANFORD JR. UNIVERSITY IN PARTIAL FULFILLMENTTft 5 THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY BY ELTON MARION HOGG May, 1915 The Passivification of Iron by Nitric Acid A THL5I5 SUBMITTED TO THE FACULTY OF THE LELAND STANFORD JR. UNIVERSITY IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY BY ELTON MARION HOGG May, 1915 P. THE PASSIVIFICATION OF IRON BY NITRIC ACID BY EI/TON MARION HOGG Introduction The passive state -of certain metals may be induced by methods which resolve themselves into two general subdivisions : Passivity produced by strong nitric acid and other oxidizing agents, and that produced by anodic or cathodic polarization. The experimental work described in the present paper is con- fined to a study of the passivification of iron by nitric acid. This limitation was imposed upon the research in view of the fact that of all the former work done, that with nitric acid seemed to be less understood and there appeared to be some points of attack on this particular phase of the problem which might result in the clarification of our knowledge of the entire subject. Historical In the historical discussion in this paper both general methods of passivification will be briefly discussed and a general statement of the principal theories will be made. The study of the passivity of iron has engaged the atten- tion of scientific investigators for nearly one hundred years and yet our knowledge of the subject is still very limited. The recent activity in this field, recorded in the late symposiums before the British Faraday Society, l indicates that our informa- tion on the subject has, in the past, been greatly circum- scribed. Many new methods of attack are described which are distinct departures from those of the earlier students, and many new and important facts have been discovered. To explain these facts many theories are put forward which, in most cases, are merely modifications of older ones, so that Met. Chern. Eng., u, 12, 679 (1914). 335896 6i8 Elton Marion Hogg for present purposes an outline of the three most prominent theories will be sufficient. Theories of Passivity 1. The Oxide Film Theory of Faraday. 1 According to this theory the surface of the iron is either oxidized or the superficial particles of the metal are in such relation to the oxygen of the electrolyte as to be equivalent to oxidation. The formation of a layer of oxide is supposed to cause passivity by mechanically hindering the metallic ions from entering the solution. To this theory Hittorf raises the objection that no oxide could be found which would display the necessary prop- erties. The destruction of passivity by elevation of tempera- ture is also hard to explain by this means. Finkelstein states that if an oxide is formed it must conduct electricity like the metal itself. LeBlanc raised the objection that the reflecting power of the surface in the active and passive conditions is the same; hence, if a film is present, it must be of less than molecular thickness. 2. The Valency Theory of Kruger -Finkelstein 2 and Muller* According to this theory passivity is due to the change of the metal to a nobler modification. This change of state depends somewhat upon temperature. Otherwise the elec- trochemical behavior of the metal depends on the relative oxidation and reduction potentials of the electrolyte. 3. The Reaction Velocity Theory of LeBlanc. 4 In its general form this theory states that passivity is due to the slow rate of change at the anode. The passive metal sends out ions into the solution very lowly; that is, the reaction Fe + ' = Fe ' ' proceeds very slowly because the ionization of the metal is associated with chemical changes, and, when these changes are slow, passivity occurs. Several hypotheses have been 1 Phil. Mag., (3) 9, 57 (1836). 2 Zeit. phys. Chem., 39, 104 (1902). 3 Ibid., 48, 577 (1904). 4 Zeit. Elektrochem., 6, 472 (1900); n, 9 (1905). The Passivification of Iron by Nitric Acid 619 advanced regarding the mechanism of the reaction and the following are of importance : a. The Oxygen Charge Hypothesis of Fredenhagen 1 and Muthmann and Frauenberger. 2 The cause of passivity is sought in the slow rate of reaction between the anode and the oxygen liberated there, with the result that the anode becomes charged with the gas, or that a metal-oxygen alloy is formed. Grave objects to this hypothesis because it does not explain the fact that iron may be passive in alkalies and when heated in nitrogen. b. The Anion Discharge Hypothesis* By this theory the main change at the anode is not the formation of metallic ions, but the discharge of anions. The electrode, when active, is supposed to contain hydrogen and the discharged anion reacts with and removes the hydrogen. The slowness of this reaction allows oxygen to accumulate on the metal rendering it passive. c. The Hydrogen Activation Hypothesis of Foerster 4 and Schmidt. 5 According to this hypothesis the normal condition of iron is assumed to be passive and it becomes active under the influence of a catalyst. The reaction Fe + ' ' = Fe ' ' is reversible and is a retarded reaction in both directions. If this is true, the formation of a film may be considered the consequence, and not the cause, of passivity. The cause of passivity is taken as the absence or the removal of hydrogen and the deposition of oxide follows as a result of the inactivity of the metal. Foerster claims that hydrogen or an alloy of hydrogen and iron is the catalyst; but Grave and Schmidt favor the view that hydrogen ions are responsible for the change. d. The Retarded Hydration Hypothesis of LeBlanc. 6 According to this view the active iron sends out ions into the 1 Zeit. phys. Chem., 43, i (1903); 63, i (1908). 2 Sitzungber. bayr,, Akad., 34, 201 (1904). 3 Chem. News, 108, 249, 259, 271, 283 (1913). 4 Abhandlungen der Bunsen Gesellschaft, 2 (1909). 5 Zeit. phys. Chem., 77, 513 (1911). 6 Lehrbuch der Elektrochemie, 5th Ed., p. 285. 620 Elton Marion Hogg electrolyte and with metals tending to become passive, these ions combine slowly with the water ion + water = ion hydrate. The concentration of free ions at the electrode becomes great and finally the potential difference between the electrode and electrolyte becomes so great that the discharge of anions and development of oxygen begins. This view is based on the observation of Grave that when the ion concentration is sufficiently large, polarization begins at both anode and cathode. LeBlanc considers that the hydration and dehydra- tion of ions under certain circumstances may be a very slow process. To all of the above theories objections may be raised, many of them seemingly valid, and, on the other hand, the observed phenomena seem in many respects to fit one theory as well as another. Definition Up to the present time no single definition of passivity has been accepted by the many investigators who have busied themselves with this problem. In the present work the pass- ivity of iron is taken to mean that state of the metal in which it is not attacked immediately by dilute nitric acid. Normally nitric acid of specific gravity 1.250 will instantly and vigor- ously attack iron and finally dissolve it; but if the metal is first treated with acid of densities from 1.590 to 1.260 the attack is delayed for some time, frequently as long as seventy- two hours and even longer in some cases. During this period of inactivity the metal remains bright and no apparent action occurs. Preliminary Work A number of simple hand experiments were performed to determine, if possible, whether there was a measurable effect of the following factors on the passivity reaction : 1. Varying concentrations of nitric acid. 2. Different grades of iron. 3. Presence of iron salts. 4. Contact with platinum and zinc wire. The Passimfication of Iron by Nitric Acid 621 5. Electrolytes other than nitric acid. 6. Varying periods of time in passivifying acid. The work, while giving many important facts, yielded results so inconsistent that we decided to pass directly to a set of carefully controlled reaction velocity experiments. On the basis of information derived from the preliminary experiments, certain lines of investigation naturally suggested themselves as likely to give fruitful results. After careful consideration, it seemed important to first investigate the rate at which iron is dissolved by nitric acid in varying con- centrations and at various temperatures. It was hoped that the results of such investigations might throw some light on the character of the reaction in general, and show whether the development of passivity was gradual, or whether there was a definite state which might be looked upon as a passive state, and also whether this state was influenced by tempera- ture. Velocity of the Reaction between Iron and Nitric Acid The metal used was "Merck's Pure Iron Wire," in the form of pieces two inches long (5.08 cm) and bent horse-shoe in shape so as to be readily hung from a glass hook. The density of the wire was 7.850 and each sample weighed about 180 mg. The acids were made up from "Baker's C. P. Nitric Acid, sp. gr. 1.42" and triple distilled water, the last distillation being with potassium permanganate and potassium hydroxide. In all, sixteen strengths of acid were prepared with the follow- ing concentrations : TABLE i 1 Sp. gr. Grams HNO 3 in 100 cc Sp.gr. Grams HNO 3 in IOO CC I .025 4-7I50 .240 47.4796 1.050 9-4395 .250 49-7750 1-075 14.1362 .260 52.0884 I . IOO 18.8210 .270 54.4449 1 .200 38.8320 -275 55.6610 1 .210 40.9222 .285 58.0563 1 .220 43.0416 .300 61.7370 1.230 45.8210 .400 91.4200 Physico-Chemical Tables, John Castell-Evans, Vol. 2, 839-842 (1911). 622 Elton Marion Hogg The apparatus, shown in Fig. i, consisted of a constant temperature bath with its stirring apparatus, constant level attachment apparatus for stirring the acid in the test tubes, and thus preventing the accumulation of gas bubbles on the surface of the iron, and an attachment for holding four test tubes in the bath. In the figure, a indicates a covering of wool packing, b is a large battery jar, c is the stirring apparatus for the bath, d is the constant level attachment, e is the ther- mometer, / is a pulley which operates the stirring apparatus Fig. i for the acid in the test tubes, g is a glass rod which is attached to the pulley / and holding the iron sample in a hook at the lower end, h is a glass tube serving as a guide for g and entering the top of the test tube through a perforated cork stopper, i is a Jena glass test tube containing the acid and the iron sample, / is the motor which operates the stirring apparatus in the bath and raises and lowers the iron sample twenty times a minute, and k shows the shape and position of the iron. The method consisted in taking weighed samples of iron wire and allowing twenty-five cubic centimeters of nitric acid of The Passimfication of Iron by Nitric Acid 623 known concentration to act on them at a given temperature for periods of fifteen, thirty, forty-five, and sixty minutes, respectively. The samples were removed at the end of each period, washed, dried and reweighed. The loss of weight was taken as a measure of the amount of reaction. All samples were run in duplicate so that for each and every period of time, for every concentration of acid used, and for every given temperature, two observations were possible. Reaction Velocity Equations The reactions between iron and nitric acid are at present not well understood. Some investigators claim that with con- centrated acid there is no action on the iron; others state that the strong acid causes the evolution of nitrogen peroxide, while dilute acid gives nitric oxide. Mellor 1 is of the opinion that "with dilute nitric acid, hydrogen is not evolved; but the acid is reduced to ammonia With hot nitric acid ferrous nitrate and nitrogen oxides are formed." Gmeliri- Kraut 2 states that in strong nitric acid some peroxide of nitrogen is formed, while nitric oxide passes off from the action with dilute acids, and in intermediate concentrations mixtures of these oxides are evolved. It is, therefore, a difficult matter to choose any one equation to represent the reaction in ques- tion. Since, however, stronger acids act relatively slowly, the main reaction to be considered is that brought about by weaker acids, namely that which produces nitric oxide instead of the peroxide. The equation, assuming ferric nitrate to be the iron salt produced, is Fe + 4 HNO 3 = Fe(NO 3 ) 3 + 2 H 2 O + NO. If ferrous nitrate were formed instead of the ferric salt, the tendency would be to produce variations in the velocity con- stant in the opposite direction from those found, which only confirms the more strongly the conclusions to be drawn later. The formation of nitrogen peroxide along with nitric oxide would influence the variations in the same sense as those 1 Mellor: "Modern Inorganic Chemistry," 485 (1912). 2 Handbuch der anorganischen Chemie, Band i, Abteilung i. 624 Elton Marion Hogg found, thus tending to somewhat discount the conclusions. But evidence to be given will show that this fact may be ignored, because, as will be proven, the reaction stops going at any easily measurable rate when nitrogen peroxide is present. For these reasons the equation has been assumed to be sufficiently correct for its purpose. According to the equation, 56 grams of iron react with 252 grams of nitric acid, and i gram of iron is dissolved by 4 x /2 grams of the acid. The generally accepted Reaction Velocity Equation for a heterogeneous system, of which only one concentration is variable 1 (in this case the acid), is f=K.S.(A-*), where doc/dd is the rate of solution, K the reaction velocity constant, S the surface exposed, and (A -- x) the concentra- tion of the acid. For our purposes, doc measures the amount, in grams, of iron dissolved in the time do, and, since, each gram of iron dissolved uses up 4 x / 2 grams of nitric acid, the x in the right-hand member of the equation must be multiplied by 4*72. The equation then becomes On transformation and integration this equation becomes dx 7T- VT-T= K.S.<#, (A 4Vi *) and In A In (A 4 V 2 x) = K . S . 0. Then, obtaining the value for K, we have TT - _1_ /*, A >.S' (A-4V2*)' To obtain an equation for the surface, it is only necessary to make use of the following equation: 1 On account of the great tendency to simplification in reaction rates, it is assumed that the reaction is proportional to the concentration of the acid, and not to its fourth or other power. The Passivification of Iron by Nitric Acid 625 where / is the length of the wire, m its mass, d its density, and TT has its usual value. Solving for S, we have _ 4 irlm ~d~' and S = Substituting this value in the Reaction Velocity Equation, we have i A . In /47r/w ' (A 4VW 6 . d in which A and x are measured in grams of nitric acid and not in gram equivalents. Thus K represents the number of grams of iron dissolved in one minute from one square centi- meter of surface under the given conditions and m is the average mass of the sample, i. e., where w is the original mass and mi the mass at time 6. Under these conditions V4 irlm ~~d~ represents the average surface. Experimental Results The first series of observations was made at o C, and the value of the reaction velocity constant calculated for each sample of iron taken. The results showed a distinct decrease of the value of the constant for increasing concentrations of acid, while with the stronger acids there was also a decrease with increasing time. Observations were also made at 10 C. and 20 C. Complete tables giving concentration, time, and temperature would be too large to present in full so we have chosen four representative ones, Table 2 gives a full summary of the reaction velocity constants obtained in the whole series of experiments. As it became evident, after a a time, that 626 Elton Marion Hogg data beyond certain limits of concentration would be super- fluous, we frequently omitted such observations. TABLE 2 Values of K T = oC = time in min. Sp. gr. of acids 15 30 45 60 1.025 0.00574 0.00661 0.00655 0.00680 0.00645 0.00685 0.00583 1.050 0.00431 0.00479 0.00442 0.00439 o . 00466 o . 00466 o . 00450 1-075 o . 00368 0.00373 o . 00366 0.00364 0.004II o . 00407 0.00424 0.00390 I. 100 0.00387 0.00305 0.00327 o . 0032 i 0.00389 0.00338 0.00322 0.00324 1.200 0.00157 0.00134 0.00151 O.OOI29 0.00152 0.00133 0.00149 O.OOI29 1.250 0.000825 o . 000736 0.000834 0.000814 0.000862 0.000797 0.000792 o . 000843 1.260 0.000254 0.000129 0.0000877 0.0000554 0.000229 0.000123 0.0000946 0.0000815 1.270 0.0001036 0.0000539 0.0000414 0.0000320 0.0000889 o . 0000584 0.0000389 0.0000253 1-275 0.0000561 0.0000372 0.0000238 0.0000160 0.0000994 0.0000280 0.0000256 0.0000228 1.300 o . 0000463 0.0000156 0.0000104 0.00000777 0.0000285 0.0000126 0.0000104 0.00000782 1.400 0.00000273 0.00000137 0.00000091 0.00000068 0.00000274 0.00000137 0.00000091 0.00000068 The Passivification of Iron by Nitric Acid 627 TABLE 2a l Values of K T = 10 C = time in min. Sp. gr. of acids 15 30 45 60 I. 100 0.00623 0.00569 0.00677 o. 0061 i 1.200 o . 00402 0.00332 o . 00366 0.00292 Entire sample of iron dis- I .210 0.00396 0.00314 solved in specified time o . 00402 I .220 0.00401 0.00300 0.00367 0.00301 1.230 0.002585 O.CKX)6l5 0.000331 o . 000484 0.000814 0.000751 0.000338 0.000277 I .240 0.000398 0.000193 0.000113 0.0001322 0.000384 0.000200 0.000124 0.0000785 1.250 0.000256 O.OOOI3O 0.0000887 o . 0000607 0.000266 0.000128 o . 0000867 o . 0000649 I .260 0.000176 0.0000885 o . 0000630 o . 0000468 0.000171 O.OOOIoSl 0.0000587 0.0000439 1.275 o . 0000880 o . 0000439 0.0000291 0.0000220 o . 0000880 0.0000436 0.0000250 0.0000219 1.300 0.0000437 0.0000218 0.0000146 0.00001097 0.0000438 O.OOOOI9I 0.0000146 0.00000958 I .400 0.00000273 0.00000136 0.00000089 0.00000068 0.00000272 0.00000136 0.00000091 0.00000068 1 With acids (sp. gr. 1.025-1.075) the entire sample of iron dissolved inside of 15 min. and with acids (sp. gr. 1.100-1.240) inside of 30 min. 628 Elton Marion Hogg TABUS 2& 1 Values of K T = 20 C 8 = time in min. Sp. gr. of acids 15 30 45 60 1.250 0.000251 O.OOOI3I 0.0000741 0.0000584 0.000278 0.000128 O.OOOIOOO o . 0000685 I .260 O.OOOI7I 0.0000888 o . 0000538 0.0000429 O.OOOI72 o . 0000885 0.0000537 0.0000401 1-275 0.0000936 o . 000044 T 00002 94 0.0000221 o . 0000883 o . 0000442 0.0000313 O.OOOO22O 1.285 O.OOOO828 o . 0000443 0.0000276 O.OOOO22O 0.0000828 0.0000495 0.0000240 o . 0000207 1.300 0.0000382 0.0000128 o . 00000964 0.0000192 0.0000164 0.00000273 1.400 0.00000545 0.00000137 0.00000181 . 00000068 0.00000272 0.00000136 0.00000091 . 00000068 In Tables 3, 4, 5 and 6 appear the complete data for con- centrations of acid corresponding to densities 1.050, 1.250, 1.260 and 1.400, respectively, the measurements being for o C. These tables are chosen because they present typical forms of conduct, and some considerable discussion of them will be given, from which, it is hoped, the significance of the results in Table 2 will become quite clear. TABLE 3 TABLE 4 Sp. Gr. = 1.050 Sp. Gr. = 1.250 T = oC T = oC A = 2.360 A = 12.444 e X m K e X m K 15 0.0376 0.1632 0.00431 15 0.0395 0.1643 0.000825 . 0403 o. 1606 o . 00466 o . 0408 o. 1610 0.000862 30 0.0758 0.1454 o . 00479 30 0.0668 o. 1489 o . 000736 0.0723 0.1488 0.000797 45 0.0973 0.1310 o . 00442 45 0.1052 0.1305 0.000835 o. 1024 0.1320 o . 00466 o. 1009 0.1325 o . 000792 6O I O . 1 2 1 2 0.1223 0.00439 60 o. 1294 0.1179 0.000814 0.1232 o. 1204 0.00450 0.1326 o. 1164 o . 000843 1 With acids (sp. gr. 1.025-1.240) the entire sample of iron dissolved inside of 15 min. The Passivification of Iron by Nitric Acid 629 TABLE 5 Sp. Gr. = 1.260 T = oC A = 13.022 TABLE 6 Sp. Gr. = 1.400 T = oC A = 22.855 e OC m K e x m ' K 15 0.0137 0.1767 0.000254 15 o.oooi o. 1827 0.00000273 0.0124 0.1767 0.000229 o.oooi o. 1822 0.00000274 30 0.0139 0.1776 0.000129 30 0.0003 0.1827 0.00000137 0.0134 o. 1760 0.000123 0.0003 0.1809 0.00000137 45 0.0140 . o. 1705 O.OOOO88 45 O.OOO2 O. 1834 0.00000091 0.0153 0.1756 o . 000095 1 0.0003 0.1842 0.00000091 60 0.0117 0.1733 0.000055 1 60 0.0003 0.1837 0.00000068 0.0176 0.1741 0.000081 ! o.oooi 0.1813 o . 00000068 From an inspection of Table 3, it is seen that since the value of K remains practically constant throughout, the loss of iron in the varying concentrations of acid is in exact agree- ment with the mass action theory of reaction velocity in hetero- geneous systems on the basis of the equation assumed, and from this fact we may conclude that there is no inhibition to re- action, and, hence, no tendency towards the development of passivity as time goes on. In Table 4 there is a similar agreement in the time-con- centration relation, but the value of the constant is about one-fifth that of the corresponding value in the preceding table. This means that in the stronger acid there is a marked de- crease in the rate of reaction. Notwithstanding this, how- ever, 1.250 acid does not develop a passivity which increases with the time of action. Table 5 gives data which show the first evidence of the gradual development of the passive state. The value of the constant for the fifteen-minute period is about one-third that of the corresponding value in Table 4; and for the thirty-, forty-five and sixty-minute periods in the same concentration of acid these values become still smaller. The attack on the iron is practically limited to the first fifteen minutes of re- action, as is shown by the values of oc, which, with slight varia- 1 Probable error. 630 Elton Marion Hogg tions, are practically constant after that time. Thus, at the end of the thirty-minute period, the inhibiting force reaches its maximum value, and the attack of the acid is practically stopped. At o C the passive point or what we will hereafter call the " passive break," occurs somewhere between acid concentrations of 1.250 and 1.260. The passive break is shown in the Curve ia, Plate I, where the acid concentrations, expressed in grams of nitric acid in twenty-five cubic centimeters of solution, are abscissae and the values of K X io 6 are ordinates. The curve is plotted from values of K obtained at o C, for sixty-minute periods, \ 9QOO rooo 5OOO 3000 /ooo O 2 <4 6 3 IQ 12 14 /G /8 2O 22 24 Plate I 1 using data for all concentrations of acid in Table i. The acid concentration, 12.444, corresponds to HNO 3 (1.250). The curve is made up of three parts, one (BC) showing the gradual decrease of the velocity constant with increasing con- centrations, the second (BA) showing the very rapid drop of the velocity constant to a very small value, and OA showing the very slow decrease of the constant in passivifying acids. ib shows the same curve except that the value of K is multi- plied by io 7 instead of io 6 . In these curves identical letters refer to identical points. The latter curve is introduced to show that the drop in the velocity constant is a gradual, although a very rapid one, in the portion BA. 1 Abscissae are grams HNOa per 25 cc. The Passivification of Iron by Nitric Acid 631 The portions of the curves represented by AB show the pas- sive break. This point has the characteristic that acids having greater concentrations will induce passivity, passivity always being considered as a somewhat variable quantity, whereas those acids having lower concentrations will activify. Even so, it is to be seen that neither passivity nor activity are perfectly definite states. All passive iron even in passivifying acids is still undergoing some solution. In acids more dilute than those corresponding to the passive break, iron is slowly and imperfectly activified, while in the still lower concentra- tions this is accomplished more rapidly and completely. Table 5 gives the results for acid of density 1.400, the highest concentration used in the experiments. The value of the constant has suffered a very marked decrease, and there is, nevertheless, as the table shows, a very slow rate of solu- tion. The 1.400 acid is, therefore, a strongly passivifying acid. The results indicated by the data in these tables is con- firmed and amplified by the results given in the complete sum- mary of all results, namely Table 2, and the complete data for the curves ia and ib in Plate I were taken from this table. It is to be noted that in some cases, namely in those ex- periments carried on with acids of high passivifying power, the whole measurable amount of reaction was over in the first fifteen minutes. In such cases, of course, the decrease of the values of K with increasing time is without particular signif- icance. From the results given above, we must conclude that an increase of concentration of nitric acid inhibits the rate of solution of iron very greatly even at concentrations which are not sufficient to produce visually complete passivity. In concentrations of acid just below the passive break, that is, from 1.200 to 1.25, it is also evident that the degree of inhibition increases very materially with the time of action. In no case was the development of the inhibition complete at the end of one hour, although in some cases the sample of iron had completely dissolved at that time. When the density Elton Marion Hogg of the acid used reaches 1.260, there occurs a break in the re- action velocity. The values of K fall off at a far more rapid rate than in more dilute acids, although even here the drop is not abrupt. Rather the values of K decrease at a measur- able pace to the very small values corresponding to visually complete passivity. Visually complete passivity is not per- fect passivity, but merely a very slow rate of reaction. Rate of Solution of Iron in Passivifying- Acids In order to be positively sure that there was solution of iron in passivifying acids, and to obtain some data on the solution rate, the following experiments were devised : Weighed samples of iron and ten cubic centimeters of nitric acid (1.300) were placed in Jena test tubes and the tops of the tubes drawn out to a capillary to prevent evaporation. Duplicate sets were prepared to allow of observations over a period of twelve weeks. At the end of the periods specified in the table the samples were removed from the acid, washed, dried, and re- weighed. The data appear in Table 7. The entire experi- ment was repeated with nitric acid (1.400) and Table 8 gives the results so obtained. TABLE 7 TABLE 8 Sp. Gr. = 1.300 Sp. Gr. = 1.400 A = 6.174 A = 9.142 e in weeks X m K .0 in wks. X m K ! 0.0043 0.1820 0.00279 I 0.0043 0.1793 0.00166 0.0043 o. 1813 0.00255 o . 0040 o. 1790 0.00157 2 0.0062 o. 1786 0.00187 2 0.0077 0.1778 0.00158 0.0060 o. 1808 0.00181 o . 0079 o. 1796 o. 00161 3 o . 0086 0.1792 0.00174 3 0.0123 0.1748 0.00168 o . 0088 o . i 804 0.00179 0.0115 o. 1748 0.00159 5 0.0150 0.1751 0.00183 5 0.0209 0.1719 0.00176 0.0143 0.1751 0.00174 0.0197 o. 1726 0.00165 6 i 0.0187 o. 1752 0.00191 6 0.0244 0.1678 0.00173 0.0177 o. 1729 0.00183 0.0243 o. 1712 0.00170 8 0.0258 o. 1694 O.OO2OI 8 0.0356 o. 1619 0.00196 0.0259 o. 1690 O.OO2O4 0.0367 0.1651 0.00196 10 0.0347 0.1668 0.00220 10 0.0452 0.1584 0.00197 o . 0340 0.1651 0.00218 0.0454 o. 1601 0.00197 12 0.0413 0.1618 0.00222 12 0.0544 0.1544 . 00202 0.0427 o. 1614 0.00230 0.0547 0.1540 o . 00203 The Passivification of Iron by Nitric Acid 633 From these tables it is evident that there is a continuous rate of solution of iron even in passivifying acids. The results further show that while the absolute amounts of solution (values of x) is noticeably greater in the stronger than in the less strong acid, nevertheless, this increase is not sufficiently great to make up for the theoretical effect (as demanded by the mass action law) of the increased concentration of the acid used. This accounts for the fact that while in the stronger acid the values of x are larger, the values of K are smaller. This points, of course, to an increased inhibiting effect with increasing concentrations, even with these extremely strong acids. The Time-Temperature-Concentration Function From the data given in Table 2, a number of cross com- pilations might be made. In Plate II are shown the values of K X io 6 plotted against the time for five concentrations 6O Plate II of acid at o, 10 and 20 C. Each individual curve shows the decrease in the velocity constant with increasing time. With decreasing concentrations of acid the curves assume a 634 Elton Marion Hogg steeper slope, showing a rapid falling off in the value of the constant in partially passivifying acids. It will be seen that, in the case of acids of higher concentrations, namely, 1.400, 1.300, and 1.275, the values of the constant for the sixty-minute periods are nearly the same at all temperatures, while for the shorter periods the value increases as we pass from o to 10 C, and then remains constant or decreases slightly from 10 to 20 C. With 1.250 acid at o C, the value of the constant is so large as to be entirely off the plate. Some of the points for the duplicates show considerable variation, but when we consider that the amounts of iron dissolved are extremely small, they are no greater than is to be expected. The facts, which are brought out above, seem to indicate a much slower development of passivity at o C than occurs at 10 and 20 C, and we have the curious phenomenon, between o and 10 C, of a reaction with a negative temperature coefficient. Thus, 1.250 acid shows no development of passsivity at o C. At 10 C the constants are smaller from the start, and rapidly decrease with time. (Compare values for o and 10 C in Plate II.) In the summary possible explanations of these phenomena will be given. (See Summary .B, 8.) The Temperature Function of the Passive Break Concentra- tion It has been previously pointed out that at a given tempera- ture there is a fairly definite concentration of acid, above which the reaction rate diminishes rapidly. This point in concen- tration we have called the "passive break." At different temperatures this passive break occurs at different concen- trations. The following data show this. The values for o, 10 and 20 C are taken from Table 2, while that for 100 C is taken from an experiment to be described later. Passive break density 1.260 1.230 1.250 1.300 Temperature o 10 20 100 Thus the passive break concentrations decrease from o to 10 C, and thereafter increase. The Passivification of Iron by Nitric Acid 635 Miscellaneous Experiments In view of the fact that a better understanding of some of the phenomena of passivity was obtained from the pre- ceding investigation, it was considered advisable to repeat certain of the preliminary experiments, not only as checks, but also with the hope that new phases of the problem would present themselves. Different Samples of Iron. Although many investigators claim that impurities do not influence the passivity reaction, it was found to be nevertheless true that different grades of metal gave different results. In the case of "Stubb's Drill Rod," nitric acid 1.300 failed to passivify. This was also true in a few cases with "Bessemer Steel Rod." ''Merck's Pure Iron Wire," however, gave better results, and acid of specific gravity 1.300 produced passivity in every case. In all of the following experiments Merck's wire was used exclu- sively. Reaction at Higher Temperatures. It was found that the amounts of iron dissolved were approximately the same in acid of density i .300 at all temperatures used in the reaction velocity experiments. With a view of obtaining data at a still higher temperature, the following experiment was made at the boiling point of the acid. Nitric acid (1.300) was heated to its boiling point (about 115.3 C) and the sample of iron dropped into the hot acid. Violent continuous action occurred for some time, after which the iron became partially passive, a slow evolution of nitrogen peroxide persisting. The acid was allowed to cool slowly and the tube was jarred at frequent intervals. TABLE 9 Temperature at which Boi,in g point of acid 115.0 C 115.5 115.3 101.0 C 99-0 100.0 8 4 .oC 80.0 83.0 636 Elton Marion Hogg This jarring always produced temporary action for a few seconds. As the tube cooled, a point was always reached at which jarring would not produce this temporary activity. Table 9 gives the results of the experiment. While there is a great probability that there is considerable solution of iron at temperatures below iooC, we may conclude that the passive state is fairly stable under these conditions, and that the passive break will occur somewhere about 100 C for acid of this strength. No explanation of the fact that jarring produces temporary activity can be offered at this time. Effect of Time of Immersion. A sample of iron was im- mersed in ten cubic centimeters of nitric acid (1.400) for five seconds and then quickly transferred to ten cubic centimeters of 1.050 acid until active, after which it was returned to the first acid for five seconds. This procedure was repeated until the iron remained passive in the dilute acid for a period of at least ten minutes, the same portions of the acids being used. Sets were also run for immersion periods of fifteen, thirty, and sixty seconds in the 1.400 acid and the results are re- corded in Table 10. TABLE 10 6 = 5 sec. 6 = 15 sec. 6' 0' N i 2 3 I 2 3 I 4 2 4 8 7 9 2 10 12 20 12 17 10 3 18 44 6 4 26 18 19 4 53 60 86 26 17 15 5 10 min. 37 21 30 6 93 42 29 7 6 3. 70 8 10 min. 9 10 ii The Passivification of Iron by Nitric Acid 637 6 = 30 sec. 6 = 60 sec. 0' 0' I 2 3 I 2 3 I 6 12 6 15 H 13 2 14 18 10 17 20 19 3 17 17 19 H 16 29 4 10 17 '5 15 18 35 5 18 14 18 18 17 46 6 19 26 27 23 21 50 7 24 35 54 21 22 45 8 44 37. 23 23 20 55 9 10 min. 30 30 10 1 53 33 ii 46 45 10 min. 6 is the time of immersion in 1.40x3 acid, 0' is the time in seconds required to destroy passivity in 1.050 acid, and N is the number of immersions. The results are taken from a complete series of observations on ten independent samples for each im- mersion period. Of these, the results for three samples only are reported here. Despite the slight disagreement in some instances, the experiment clearly shows that the time of immersion in passivi- fying acids has a marked influence on the degree of passivity. For shorter periods of immersion there is a decrease in the number of trials necessary to retain passivity for ten minutes, while the longer passivifying immersions give at the outset a longer persistence of passivity in the activifying acid. For these phenomena, we have no explanation at present. Experiments with the Same Portions and also with Fresh Portions of Acids. It was thought to be interesting to find if the time of activation after passivification was varied by the number of immersions in the same portions and also in fresh portions of both activifying and passivifying acids. The following experiments were carried out under laboratory con- ditions. The method of procedure was the same as that em- ployed in the preceding experiment, one minute passivifying 638 Elton Marion Hogg immersions being used. The work was divided into five parts and five samples of iron were used for each part. (a) Sample of iron passivified in same portion of 1.400 acid and activified in same portion of 1.050 acid. (b) Sample of iron passivified in same portion of 1.400 acid and activified in fresh portions of 1.050 acid. (c) Sample of iron passivified in fresh portions of 1.400 acid and activified in same portion of 1.050 acid. (d) Sample of iron passivified in fresh portions of 1.400 acid and activified in fresh portions of 1.050 acid. (e) A fresh sample of iron passivified and activified in the acids already used in part (a) . The results are given in Table 1 1 . TABLE n 9 = one minute 0' = time in seconds of persistence of passivity in 1.050 acid N = number of immersions N Values of e' (a) (b) (c] (d} () I 15 H 8 9 16 8 15 8 ii ii 12 2 17 20 14 15 19 13 17 15 18 3 14 16 19 19 16 15 18 18 10 min. 4 15 18 21 21 18 17 17 19 5 18 17 24 22 18 20 15 18 6 23 21 25 26 15 20 19 21 7 21 22 28 30 16 27 18 20 8 23 20 32 36 23 30 20 24 9 30 30 33 37 27 39 25 26 10 53 33 45 39 29 42 23 29 ii 10 min. 35 52 4i 29 39 12 51 50 10 min. 31 40 13 67 57 40 39 H 78 64 38 48 15 79 78 48 55 16 88 93 61 73 17 102 96 70 68 18 1 06 1 02 69 84 Not Not passive passive The Passivification of Iron by Nitric Acid 639 These data show that the difference in effect produced by using the same portion or fresh portions of the passivifying acid is very slight, while in the case of the fresh portions of the activifying acid the iron was not made passive for a period of ten minutes even after eighteen immersions, while with used portions, a far less number of immersions accomplished this. The results shown in part (e) are most striking. Such results as these may be duplicated at will. In fact, where only two duplicate experiments are given in each part, five were actually performed. The interesting point is that activi- fying acids, even as dilute as 1.050, sooner or later lose their activifying power when repeatedly used for this purpose, while fresh acids will always activify. An explanation for this will be given later. (See Summary B, 3.) Influence of Different Strengths of Nitric Acid on Passivity. -The experiments were carried out under laboratory condi- tions. The samples of iron were subjected to the action of different concentrations of acids which were known to passivify and then transferred to acids which would activify. The time of immersion was sixty seconds in the case of the passivi- fying acids. Table 12 gives the data taken from five separate determinations in each case. From the table it appears that with decreasing concen- trations of passivifying acids the persistence of passivity in the same activifying acid always increases. In the case of passivifying acids of the densities 1.260 and 1.270, it would seem that, by use, these acids became diluted to a strength below that required to produce passivity, although up to this time the persistence of passivity was greater than with stronger acids. A possible explanation of the above phenomenon will be offered in the summary. (See Summary B, 9.) It will also be noticed that the more dilute the activifying acid, the less persistent the passivity. Effect of Contact with Platinum and Zinc. A piece of platinum wire was wound around the iron sample and the usual procedure followed. The results given in Table 13 are 640 Elton Marion Hogg TABLE 12 Activifying acids Passivifying acids 1.250 1.400 1.300 1-275 1.270 1.260 B' = time in seconds iomin. 10 min. 10 min. 10 min. 10 min. I .200 10 min. 10 min. 10 min. 10 min. 10 min. I . 100 8 9 18 18 15 15 19 23 36 39 17 18 43 44 10 min. 10 min. 10 min. 1.050 15 H 17 20 14 16 15 18 18 17 23 21 15 18 26 30 36 40 52 59 75 93 109 116 25 28 5i 35 100 97 135 H7 189 204 10 min. 39 36 69 66 160 138 196 231 Fails to make 56 75 153 H5 Fails to make passive 21 22 10 min. passive 23 20 30 30 53 33 46 45 10 min. for sixty-second immersion periods in 1.400 acid and 6' is the time in seconds in 1.050 acid. TABLE 13 0' = time in seconds ^64 157" 128 138 72 hours While no explanation of the effect of platinum on the passive state can be given at this time, the results show that the nobler metal exerts an inhibiting effect on the activifica- tion of iron in dilute nitric acid, a fact previously noted by both Faraday and Schoenbein. A peculiar phenomenon was noticed in connection with the effect of platinum. The The Passivification of Iron by Nitric Acid 641 nobler metal was wound around the middle of the iron sample and the couple placed in acid in which the iron alone would have remained active. It became passive. If one end of the iron wire was now touched with zinc, that end only became active and remained so as long as the zinc was in contact. In some instances after removing the zinc, the end formerly passive became active for a moment and as it became passive again the opposite end became active. The alternations often occurred four or five times before the entire wire be- came passive. No explanation of this phenomenon is offered at this time. Zinc always activified the iron immediately by contact in dilute acids in which it would have otherwise remained passive for some time. The Influence of Iron Salts. In order to ascertain the influence of iron salts on passivity, the following experiment was performed: Samples of iron, previously weighed, were placed in ten cubic centimeters of nitric acid (1.300), together with one cubic centimeter of ferric nitrate solution containing 0.0653 grams of iron and 0.5304 gram of nitric acid in each cubic centimeter. The density of the ferric nitrate solution used was 1.4459. The tops of the test-tubes containing the samples were drawn out to capillaries. After a specified time had elapsed the tubes were opened, the iron removed, washed, dried and reweighed. TABLE 14 A = 6.704 6 x m K I 0.0052 0.1813 0.00280 0.0062 o. 1804 0.00338 2 0.0079 0.1792 0.00220 0.0079 o. 1760 0.00222 3 o.ono 0.1759 o . 00209 0.0126 0.1776 O.OO236 5 0.0208 o. 1710 O.OO239 0.0199 0.1753 0.00226 6 0.0260 o. 1716 0.00247 0.0239 o. 1697 0.00231 642 Elton Marion Hogg The values of K were calculated and the data appear in Table 14, where 6 is in weeks, x is loss in grams, A is the acid concen- tration in grams of nitric acid in each ten cubic centimeters, and K is the reaction velocity constant, which measures the amount of iron dissolved from one square centimeter of sur- face in one week under laboratory conditions. At the time of starting the experiment it was noticed that there was a rapid initial action for a few seconds which accounts for the seemingly large amount of iron dissolved in the first week. The table shows a slight decrease in the amount of iron dissolved as com- pared with Table 7 where there was not an excess of iron. However, this is not so great as to be very significant and we may conclude that the presence of iron salts has little or no effect on the passivity reaction. The Influence of Nitrites. -To determine the influence of nitrites on the passivity reaction, a solution of sodium nitrite was prepared containing fifty grams of the salt in five hundred cubic centimeters of solution. Three-hundredths of a cubic centimeter of the solution was added to ten cubic centimeters of nitric acid (1.400) and this combination was used as the passivifying agent. The persistence of passivity was de- termined as usual in i .050 acid. The results appear in Table 1 5 . 15 N 6' in seconds N 6' in seconds I 13 10 6 28 44 2 18 16 7 37 47 3 20 25 8 37 57 4 21 29 9 10 mm. 5 23 35 These results are but little different from those obtained without the addition of nitrites. The experiment was re- peated, adding the nitrite solution to the activifying acid instead of the 1.400 acid. The influence of nitrites on the reaction is very evident when added to the activifying acid, in which case the iron remained passive for more than seventy- two hours in every case. The Passivification of Iron by Nitric Acid 643 Experiments with Nitrogen Peroxide and Iron. Our at- tention was called to the fact that those strengths of nitric acid, which were capable of passivifying iron, gave off, on standing, reddish brown fumes of the higher nitrogen oxides while activifying acids are water-white, and it seemed possible that this gas might be a factor in the passivity reaction. To obtain some data on this point and also to find whether a measurable amount of this gas was occluded, the following experiment was carried out : Nitrogen peroxide was generated by the action of concentrated nitric acid on copper, the gas dried over phosphoric anhydride, and liquefied at atmospheric pressure in a bulb cooled with a freezing mixture. The liquid so obtained furnished the gas used in the experiment. It was soon noticed that when much of the gas escaped into the room, files and other iron instruments became passive. Iron samples passivified in the dry gas remained for many hours in dilute nitric acid without any inclination toward activification. Since the metallic surface remained bright and also since exposure to the air had little effect, we were led to suspect that there was a possibility that the metal occluded much of the gas and formed something of the nature of a ni- trogen peroxide-iron alloy, or in some other way entered into reaction with the iron. Following this idea an apparatus of the design shown in Fig. 2 was prepared. The bulb b was Fig. 2 filled with "Merck's Pure Iron Wire for Standardization" and the entire apparatus thoroughly dried by desiccation over phosphoric anhydride and aeration with pure, dry air. The apparatus was then connected with a supply of nitrogen per- oxide and a stream of the gas allowed to flow through the tube for one hour. The gas was displaced by dry air for the same length of time and the following data gathered: 644 Elton Marion Hogg Weight of tube, iron and nitrogen peroxide 71 . 2900 grams Weight of tube and iron 71 . 2674 grams Weight of nitrogen peroxide occluded 0.0226 grams Weight of tube and iron 71 . 2674 grams Weight of tube 66 . 6001 grams ^ Weight of iron 4-6673 grams Weight of nitrogen peroxide occluded by one gram of iron = 0.0226/4.6673 = 0.00484 grams Immediately after the dry iron was removed from the tube, and exposed to air (the atmosphere was moist) small liquid globules appeared on its surface, which proved to be strongly acid. It was undoubtedly either nitric acid or a mixture of this with nitrous acid. This point will be further investigated. These results lead us to suspect that passivity is due, not to nitric acid itself, but to nitrogen ' peroxide, or at least to oxides of nitrogen higher than nitric oxide. Whether the active agent is nitrogen peroxide or possibly nitrous acid is virtually impossible to determine, because nitrous acid always breaks down readily yielding the oxide and peroxide of nitro- gen, and the peroxide on dissolving in water always yields nitrous acid. Thus, in all cases, in aqueous solution, nitrous acid, nitrogen peroxide and litric oxide presumably exist together. It is perfectly possible that both nitrous acid and nitrogen peroxide may be passivifying agents. That nitric oxide is not will be shown in the following experiment. Experiment with Nitric Oxide and Iron. Nitric oxide was prepared by passing the gases from the reaction of dilute nitric acid on copper through a freezing mixture and then through water, thus removing the peroxide. The gas so obtained had no effect on iron as far as passivification was concerned. History of the Effect of Nitrogen Peroxide on Iron It is of interest in this connection to note that there are at least four instances, in former work on the subject of passivity, where nitrogen peroxide was noted by investigators. The Passivification of Iron by Nitric Acid 645 Schoenbein 1 found that when the temperature of nitric acid (1.36) was raised to 70 C, a gas was given off which he called the "deutroxide of nitrogen," and that up to this point the iron remained inactive. Herschel 2 noticed that after nitric acid (1.399) was re- peatedly used for passivification, it became unfit for use. He ascribes the reason, "because it was impregnated with nitrous gas." Varenne 3 found that when passive iron was made active in vacuo, there was an evolution of an orange-colored gas which he believed to be the "peroxide of nitrogen." Grave 4 found -that iron heated in nitrogen up to white heat was passive and also that ionized nitrogen was a good passivifying agent. These phenomena are quite referable to traces of oxygen which under these conditions might yield some nitrogen peroxide. From a consideration of any of these four statements, it seems strange that the possibility of passivification in nitrogen peroxide directly has never occurred to investigators long be- fore this time, since it appears to be the next logical step in each of the above instances. Summary A. The results of an extensive set of velocity determina- tions of the reaction between iron and nitric acid are given. These were carried out with acids of densities ranging from 1.025 to 1.400, by small intervals, and the measurements were duplicated for three different temperatures, namely, o, 10 and 20 C. Isolated observations at higher tempera- tures are also given. From these results are drawn the follow- ing conclusions, the reference data being taken from results at o C. As these results do not differ very materially from those at 10 or 20 C, the conclusions are generally applicable to all results. 1 Phil. Mag., (3)9, 259 (1836). 2 Ibid., (3) ii, 329 (1836) 3 Ibid., (5) 9, 76 (1836). 4 Loc. cit. 646 Elton Marion Hogg 1. The value of the velocity constant for the solution of iron by nitric acid decreases with increasing concentration of acid throughout the whole range investigated. 2. This decrease in rate is not uniform, but shows a rapid, though not vertical drop in the velocity constant at the con- centration corresponding to a density of 1.260. 3. In all concentrations of acid of 1.260 or greater, the value of the velocity constant shows a decrease with the time in a given experiment. That is, the rate of solution of iron is most rapid at the start and falls off as the reaction continues. This is to be interpreted as a gradual development of passivity, increasing as the reaction proceeds. 4. The concentration 1.260 is thus characterized by two things: first, it is the point, in concentration of acid, at which progressive passivification begins; second, it is the point at which the rapid drop in the reaction rate takes place. For want of a better term, we have called this point the "passive break." 5. The passive break is not independent of the tempera- ture, but seems to fall somewhat from o to 10 C, thereafter to rise to 20 C and even to 100 C, as shown by a few ex- periments. 6. Passivifying acids, no matter how concentrated, bring about a slow but steady solution of iron. B. Passivification by Nitrogen Peroxide. It is shown that a degree of passivity, far greater than any produced by the strongest nitric acid, is brought about by exposing iron to dry nitrogen peroxide gas. Nitric oxide is without effect. This fact explains many of the phenomena of passivification which we, as well as others, have observed. i. Passivifying acids are more or less yellow or red, i. e., contain some of the higher oxides of nitrogen, 1 while those acids which activify, in general, remain water- white. 1 On account of the reversibility of the reaction, N^Os = NC>2 + NO, in aqueous solution, it is impossible to state with certainty whether the nitrogen peroxide or the nitrous acid is the passivifying agent, or whether both perform this function. All statements that are made here concerning the effects of nitrogen peroxide must, thus, be considered subject to this reservation. The Passivifiation of Iron by Nitric Acid 647 2. The great preponderance of opinion in the literature is to the effect that acids of low concentrations yield nitric oxide alone, while those of higher concentration yield some nitrogen peroxide also. Our results would make it seem prob- able that those acids which passivity are such as yield nitrogen peroxide in sufficient amounts, while activifying acids are those of such dilution as do not yield sufficient amounts of this substance. We have not, as yet, been able to investi- gate this point any further. 3. Activifying acids (even quite dilute ones such as 1.050) lose their activifying power when repeatedly used. This is presumably due to the accumulation of dissolved nitrogen peroxide (or nitrous acid) as a result of such use. 4. Experiments, which we have performed and which are too simple to need description, showed that acids in which iron is ordinarily active, rapidly produce passivity when small quantities of nitrogen peroxide are bubbled through near to the iron which is being attacked. 5. A similar effect is produced by the addition of nitrites. 6. When iron is passivified in nitrogen peroxide, a con- siderable, easily weighable, amount of the peroxide is absorbed by the iron. On exposure to moist air, the absorbed peroxide combines with the moisture and is converted into drops of a strongly acid liquid, presumably nitric acid, probably mixed with nitrous. 7. While we have, as yet, been unable to determine the minimum concentration of nitrogen peroxide capable of pro- ducing noticeable passivity, it seems likely that but little is necessary, since all iron and steel articles in a room in which a little of the gas is allowed to escape, assume a very con- siderable degree of passivity. The observation of Grave (loc. cit.) that iron which is exposed to the silent discharge in tubes evacuated from nitrogen, is probably to be explained as due to the formation of some nitrogen peroxide in the tube. 8. Schoenbein 1 records that when a bar of iron is only partially immersed in a passivifying acid, the unimmersed 1 Phil. Mag., (3) 9, 53 (1836). 648 Elton Marion Hogg portion also becomes passive. We have found that this is wholly due to exposure to the higher oxides of nitrogen (nitric anhydride, N2O.5 passivifies as well as nitrogen peroxide 1 ). If the unimmersed portions are protected from contact with the vapors and gases evolved, they do not become passive, as we have determined experimentally. 9. It has been pointed out that while nitric acid of density 1.250 shows markedly the phenomenon of the development of passivity with time at 10 and 20 C, it does not show this at o C. In the light of the foregoing discussion, the probable explanation is that such acid develops nitrogen peroxide at 10 C, but does not do so at o C. 10. Another curious fact brought out in the body of this paper is that the more dilute the passivifying acid the greater is the persistence of the passivity produced when it is brought into the activifying acid. No very positive explanation of this fact seems possible without further investigation. It may be, however, that, owing to the more gradual develop- ment of the passive state in the weaker acids, time is offered for a deeper penetration of the iron by the gas, or by whatever condition constitutes the passive state. C. The best conception of the passive state, as induced by nitric acid, which we are able to formulate, as a result of our investigations, is somewhat as follows: 1. The passive state is not a definite one. There maybe an indefinitely great number of degrees of passivity, taking the rate of solution of iron as a measure of the degree of passivity. 2. The passive state seems to be the result of an equi- librium between iron and nitrogen peroxide. Iron is capable of absorbing nitrogen peroxide from any solution in which it is being produced, and the rate of reaction is thereby in- hibited. The degree of inhibition will, therefore, be de- termined by the concentration of the nitrogen peroxide which the reaction itself is capable of maintaining. In strong acid this concentration is great, and in weak acids it is small, or perhaps zero. Thus, when a piece of iron, which has ad- justed its rate of solution to a strong passivifying acid, is 1 R. Weber: Jour, prakt. Chem., (2) 6, 342 (1873). The Passivification of Iron by Nitric Acid 649 brought into a more dilute one, it is charged with a higher concentration of nitrogen peroxide than can be maintained in the more dilute acid, and a new adjustment takes place, whereby the iron slowly gives off some of its store of inhibiting peroxide. This results in an increased rate of solution, which is what is usually called "activification." The slow develop- ment of activity in such cases is thus explained, as is also the fact that the "lag" in activification is less the more dilute the activifying acid. 3. Since iron passivified by dry nitrogen peroxide gas is much more persistently passive in activifying acids than that which is passivified in nitric acid, it would seem probable that the amount of nitrogen peroxide adsorbed by iron, from even very concentrated nitric acid, is relatively quite small. 4. What has been called the "passive break" in the concentration of nitric acid is, from our point of view, to be looked upon as that concentration at which the reaction be- gins to develop relatively large amounts of nitrogen peroxide. At concentrations below the passive break, it would seem probable that nitrogen peroxide in decreasing amounts is still being developed, since the values for the velocity constants continue to increase with dilution, although, as has been pointed out, it is difficult to say how much of this effect is due to increased electrolytic dissociation of the acid. We have not, as yet, been able to investigate this point any fur- ther, but hope to do so at some future time. 5. As to why nitrogen peroxide should produce passivity, or as to what the ultimate mechanism of the process is, we do not feel that we can say much. Whether passive iron is a solid solution of nitrogen peroxide in iron, or whether the result of the action of the peroxide on the ions is the reversible production of some highly oxidized condition of iron, are questions which we do not feel competent to discuss. We may state, however, that we have observed nothing which seems to indicate the existence of anything like a true gas film in any case. Laboratory of Physical Chemistry Stanford University May, U. C. BERKELEY LIBRARII UNIVERSITY OF CALIFORNIA LIBRARY