Oxygenation of Ferrous Iron

亚铁的氧化

IIRON

WERNER STUMM and G. FRED LEE’ Department of Sanitary Engineering, Harvard University, Cambridge, Mass.

Oxygenation of Ferrous IronThis study of the reaction between ferrous iron and oxygen emphasizes the role oxygenation catalysts play in the deferrization process. Application of these catalysts may markedly increase the rate of over-all iron removal from natural watersBEARING WATERS are often deferrized by oxidizing the ferrous iron with dissolved oxygen and then removing the resulting ferric oxide floc by sedimentation or filtration. Ferrous iron in river water--e.g.. pollution by mine drainage or iron pickling wastes-not only disturbs the oxygen balance but reduces the p H and forms suspended matter as ferric oxide-Fe20, XH20, where X may vary. The capability of iron to undergo reversible oxidation or reduction reactions is rdated to the metabolic cycles of nearly all other elements of importance and especially to the distribution of oxygen in impounded surface waters. The corrosion of, and the subsequent occurrence of“red water” in a water distribution system with iron piping is influenced by the oxygenation rate of ferrous iron (22). A quantitative investigation of the reaction between ferrous iron and oxygen is needed in order to understand the many roles iron may play in natural waters. Special attention should be given to the kinetics of this reaction; the rate at which it proceeds depends on several parameters including pH, temperature, concentration of dissolved oxygen, and catalysts. The investigations reported here have shown that the rate of oxygenation of ferrous iron in bicarbonate solutions follows the law:

sions reached in these investigations have yielded valuable clues on the mechanism of the reaction between ferrous iron and oxygen, but they do not necessarily apply to bicarbonate-containing waters. Weiss (25) suggested that the oxygenation of ferrous iron proceeds in the following sequence:

+ 7 Fe(II1)+ HOi ( a j Fe(I1)+ HO.2+ Fe(II1)+Fe(I1)0 2

HzO? ( b )

+ HyOy+Fe(II1)+ H O’+ H2O (c) Fe(1I)+ H O .+Fe(II1)Fe(T1)

(2)

+ H20 ( d )

These reactions as written are not balanced with respect to hydrogen ions and are not intended to specify the dependence of the reaction rate on this variable. The rate determining step is believed to be Reaction 2.a thus indicating that the oxygenation rate is of the first order with respect to both the concentration of Fe(I1) and dissolved oxygen.-dFe(rl)= k[Fe(II)] (Os)

the hydrogen ion concentration is observed (70. 72), a very pronounced p H dependence is observed in less acidic solutions (1: 7 7, 13, 24). However, except for the very early study of Just (73), no quantitative investigations have been undertaken on the rate of ferrous iron oxygenation in the buffer system of natural waters. Solubility of Iron. Studies on iron oxidation in bicarbonate solutions are complicated by the slight solubility of ferrous and ferric iron in such solutions. Ferrous

constituents generally are more soluble than ferric constituents. Table I gives a survey of the equilibria that must be considered for the evaluation of their respective solubilities. In bicarbonate waters, the solubility of ferrous iron is controlled by the solubility of FeC03. and can be estimated by[Fe+*

+ Fe(OH)+]=

dt

(3)

--d[Fe(ll)l dt

= k[Fe(I1)]Po,(HO-)* (1)

Catalysts (especially Cu -2) in trace quantities, as well as anions which form complexes Lvith ferric iron (HzP(3-). increase the reaction rate significantly, while small amounts of Fe(III), C1-, S04-? have no effect on the reaction rate. Previous Studies Many reliable and extensive studies have been made in homogenous acid solutions (6, 72: 74, 20). The concluPresent address, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, Pa.

The rate of oxidation is dependent upon the nature of the anion present, and increases as the complexing affinity of the anion to ferric iron increases (12). Thus under acid conditions, the rate is found to decrease in the following series: hydroxide, pyrophosphate (72), phosphate (76), chloride (20),sulfate (12, 14), and perchlorate (70). The rate law is in accordance with Equation 3 for the first four media in this list, but a second order dependence of the rate on the ferrous iron concentration has been found in sufficiently acid solutions of H& 0 4 and HCIOI. In addition to complexing agents, many substances, especially those which hasten the decomposition of peroxides in the presence of ferrous ironplatinum, charcoal, and cupric saltsaccelerate the oxidation of ferrous salt by oxygen (73). While in acid media only a small dependence of the oxygenation rate upon

where K, is the ion product of water and alk is the alkalinity[(HCO,-)+ 2(C0,-z)+(OH-)] in equivalents per liter and the numbered K values refers to the respective equation in Table I . Equation 4 is valid up to a p H of about 9 (23). In order to estimate the solubility of ferric iron: equilibria 6 to 9 in Table I must be considered. Experimental investigations (7, 9) in basic and neutral solutions (pH values 6 to 11) have shown that solubility of ferric hydroxide is markedly independent of the hydroxyl ion concentration. This observation has led to the proposal that Fe(OH)3 (s) is in equilibrium with undissociated molecular F e ( 0 H ) j (equilibrium 9: Table I). Experimental Method It is not feasible to study, within the pH range of natural waters, the kinetics of oxygenation of ferrous iron in homogenous solutions (Table I). Kinetic investigations in heterogenous solution are less amenable to rigid interpretation owing to the super-imposition of interfacial factors which operate at the phase boundaries on to the homogenous reacVOL. 53, NO. 20

FEBRUARY 1961

143

亚铁的氧化

Table I.

Solubility Equilibria

No.1 2

ReactionFe(I1) Solubility Fe(OH)*(s)= Fe-2 20HFe(OH)?(s)= F e ( 0 H )+ OHFeC03(s)= Fe+* COa-2 FeCOds) OH-= Fe(0H) HC03-= H c03-'

Equilibrium Constan

t at250

c.

+

34

56

+

+

+

+

+

8 4 2.1 1 4.8

x xX

X

10-1610-1010-11

xx

10-3 10-11

789

10a

Fe(OHjs(s) Fe(OHh(s) Fe(OHh(s) Fe(OH)l(s) 2Fe(OH) tz

Fe+3 3 0 H= Fe(OH)? OH= F e ( 0 H )+2 20H= Fe(0H)r"= Fe2(0Hj2T4=+

+

Fe(II1) Solubility

++

1.76.82.9

-10-36 10-15

x

X10-7

30

Undissociated molecular ferric hydroxide.

tion. .4s the initial purpose of this investigation is to study the chemical reaction between dissolved ferrous iron and oxygen, the concentrations of reactants must be chosen in such a way as to minimize specific surface effects on this reaction. All experiments were conducted with dissolved ferrous iron concentrations of less than 5 X l O - 5 M, such that no precipitation of ferrous carbonate was encountered even though the conditions of the experimental runs that were made above pH 7 were such that the solubility product for F e C 0 3 was occasionally exceeded. T h e existence of these supersaturated solutions of ferrous carbonate can be explained by the tendency of carbonate to remain in solution in the absence of crystallization nuclei. During the course of reaction, no precipitation or floc was observed; colloidal ferric hydroxide was formed, but this appeared to have no effect on the kinetics. since incipient addition of ferric iron u p to lOP4M did not alter the reaction rate. T h e partial pressure of oxygen and the pH was maintained constant during

each run by bubbling a mixture of C O Z and air from a compressed gas cylinder through the test solution. T h e amount of ferrous iron remaining at a given time was determined photometrically in aliquots of the test solution, in Xvhich the reaction was quenched by the immediate addition of HYSO+ reducing the pH to about 1. Because of the low concentrations of ferrous iron used, a very sensitive analytical method was needed. Since the recommended methods (2, 3) lacked both sensitivity in the concentration range used and specificity for ferrous iron in the presence of ferric iron. a modified technique using 4,7-diphenyl-l,IO-phenanthroline (bathophenanthroline) was developed (76). IVith this reagent the colored Fe(I1)-bathophenanthroline complex can be extracted into n-hexyl alcohol, thereby enhancing sensitivity as well as specificity. Preparation of Reagents a n d Experimental Conditions. T h e water used for all the results reported here was purified by redistilling glass distilled water in a n all borosilicate glass still. Reagent grade chemicals were used in all test solutions.

T h e ferrous iron solutions were freshly prepared by dissolving Fe(C104)2 (G. F. Smith Chemical Co.) in 0.1M HC104. Bicarbonate-COz buffers\vere prepared from NayCOzsolution by bubbling a n airCOZ mixture through the solution. Various pH values\vere obtained by using different air-CO? mixtures. These air-COZ compressed gas mixtures\\'ere prepared and analyzed by General Dynamics Corp., Eo;ron. hIass. The test solutions were\vel1 thermostated (*0.2' C.) and we

re vigorously stirred during the run. All pH measurements were made with a Beckman Model G pH meter. During any one run the pH was kept constant within 0.04 pH units. To a 1.5-liter borosilicate glass beaker was added a N a 2 C 0 3 solution. after temperature equilibrium\vas attained, a n air-CO1 mixture was bubbled through this stirred solution. I\'hen the pH of the solution had reached the equilibrium value for the particular partial pressure of COz in the gas mixture being used, the ferrous iron solution was added and a stopwatch was started. T h e volume of test solution a t this point was 1.O liter. At appropriate time intervals, aliquots of the test solution were withdrawn and placed in sulfuric acid solution to quench the reaction. Colloidal ferric hydroxide was dissolved by boiling this acid solution. The ferrous iron content of each aliquot was determined colorimetrically after neutralization with sodium acetate, using a Beckmann DC' spectrophotometer.Results

Stoichiometry. T h e reaction of Fe(I1) with oxygen generally leads directly IO ferric oxides or hydroxides. T h e stoichiometric relationship beingFe(l1)

+ 1/401 f 20H-+1/2Hz0-+

Fe(0H)a (s)

(5)

Effect of O x y g e n a t i o n o f Ferrous I r o n o n Surface Waters Prime examples are the rivers draining the coal mining areas of Pennsylvania, Ohio, and West Virginia. As a result of acid mine drainage, these waters contain large quantities of iron and have low pH values. Fifty per cent of the acidity in acid mine drainage arises from the oxygenation of ferrous iron; the remainder arises from the oxygenation of sulfide or polysulfide. In many cases, the drainage from these mines contains appreciable quantities of ferrous iron which, upon entering a stream, reacts with the dissolved oxygen present to form unsightly ferric oxide deposits. The accompanying reduction in pH may inhibit the selfpurification of the river b y providing an unfavorable environment for microorganisms. Water treatment operators who derive their raw water from reservoirs are also well aware of the common occvrrence of high iron concentration in the hypolimnetic zones during periods of stagnation and stratification of the reservoir. Water supplies derived from ground waters that have passed through iron bearing strata frequently contain appreciable quantities of ferrous iron which, upon exposure to the atmosphere, are oxidized to insoluble ferric oxide.

The oxidation may occur under certain conditions-especiallk in buffer solutions with p H values greater than 6in a stepwise fashion over the mixed iron( 11)-(111) hydroxides and magnetite (7, 8 ) . T h e ferrous iron in these intermediates is only very slowly oxidized, if at all (8). I t was experimentally ascertained that under the condition of the kinetic investigation reported here, the oxygenation of ferrous iron was in agreement with Equation 5. These measurements were made by quantitatively determining the oxygen uptake with a Warburg manometer appa

ratus. Oxygenation Rates. Several representative reaction runs at a constant partial pressure of oxygen and at various pH values are plotted in Figure 1. T h e rate of oxidation\vas first order with respect to Fe(I1) and independent of the Fe(II1) concentration. This is evident from the time dependence of any one reaction (linear relationship in

144

INDUSTRIAL AND ENGINEERING CHEMISTRY

亚铁的氧化

FERROUS I R O N O X Y G E N A T I O Nsemilogarithmic plot of Figure 1). This first order dependence with respect to Fe(I1) of the reaction rate was also confirmed by experimental runs in which the initial Fe(I1) concentration was ividely varied. Incipient addition of ferric iron up to concentrations of 10-4AW did not alter the reaction rate. The oxygen concentration in the reaction solution was kept in excess and constant during each run. Since differe n t investigators agree with respect to the dependence of the rate of reaction on the oxygen concentration, this dependence\vas not investigated in a rigorous manner. The partial pressure of oxygen\vas varied only between 76 and 155 mm. Hoii-ever. the results support the finding of others, that the rate of oxygenation of ferrous iron is dependent on the partial pressure of oxygen to the first power. Thus the rate law accounts for theTable 0. Dependence of Rate on Partial Pressure of Oxygen and (OH-)

Alkalinity 2.9 to 3.9 X 10-2 equiv. per liter. Temp. 20.5' C .kd

3 1 h - l A t r n . -1

Po2 dtm. 0.107 0.107 0.153 0.153 0.153 0.174 0.184 0.184 0.195 0.1951

(OH -1"

x

k,b

105

x

.\Iin. -110'

x

k'C 102

Liter2 3Iole-2

x

10-13

1.86 2.51 3.16 4.56 5.02 7.25 9.12 12.0 18.2 22.4

0.063 0.090 0.233 0.478 0.492 1.71 2.44 4.65 9.13 17.8

0.589 0.841 1.53 3.13 3.21 9.82 13.2 25.3 46.8 91.4

1.70 1.33 1.53 1.50 1.28 1.87 1.59 1.76 1.41 1.83

Computed froni measured pH ( K u= (1.7 X 10 1 4 1 . 2.3 k,= .lope of firbt order plots -din[Fe(II)I k,= k,'Pn,. k= k'/(OH-)*. (Figure 1): k,=di'

change in Fe(I1) concentration during the course of the reaction a t a constant pH value as a function of Fe(I1) concentration and partial pressure of oxygen, PO2. The change in rhe rate of reaction with p H as shoivn in Figure 1 can be explained by the influence of hydroxyl ions on the reaction rate. The logarithm of the .: rate constants X, (k,= - d l n[Fe(II)] J'dt) pided by the partial pressure of oxygen, is plotted as a function of p H of the test solution in Figure 12. The line through the points as plotted has a slope of 2. thus indicating a 100-fold increase in the rate of reaction for an increase of one pH unit; there is a second order relationship betiveen the rate of reaction and the hydroxyl ion concentration:

'Table I1 lists k, and the over-all rate constant k for a typical set of experimental runs. The value for kat20.5' C. has been found to be 1.5 (*0.5) X l@ 1 3 liter2 mole-2 atm.-' min.-' (or k,,= 1,1.*0,5 X 10'" literY if oxygen is expressed in mole per liter rather than in a m . ) The data p

resented in Figure 2 were obtained tvith solutions 2(C03-*) with a n alkalinity[(HCOs) (OH-)] between 9 X lo-" and 3.9 X equivalent per liter. Figure 2 shows that the temperature dependence of this reaction is large. For a given pH, the rate increases about IO-fold for a 15' C . temperature increase. Since the p H has been measured at solution temperature, this apparent temperature dependence can! however. be accounted for by the variation in OHion activity with temperature for a given pH, owing to temperature variation in K,. Essentially the same rate constants k,,, have been obtained for the three temperatures investigated, thus indicating an

+

+

\0

, o

\

o

\I

I0

20

30

40

50

OH

TIME, min

Figure 1. Oxygenation rate of ferrous iron i s propottional to Fe(ll) and i s strongly influewed b y pH20.5'

Figure 2. An increase b y one pH must increase oxygenation rate 100fold

C. P o=

const.

activation energy of approximately zero.[For constant (H') rather than (OH-) an activation energy of 23 kcal. can be calculated.] The Weiss radical mechanism (25) can be invoked to describe the oxidation kinetics\vith respect to the rate dependence on (FeL2) and Poi. As Abel ( 7 ) has shown, the 1,Yeiss mechanism cannot be enlarged, to account for the second order dependence on (OH-). The second order dependence on (OH-) is in accord\vith the earlier publication of Just (73) and the theoretical discussion of hbel ( I ) . I n systems of phosphoric acid, Cher and Davidson ( 6 ) obtained a second order dependence on (H?P04-). Hydroxyl ions can be considered as complex formers having a stronger affinity for Fe(II1) than (H?POd-). Close examination of the data presented in Figure 2 sho\vs that reaction rates obtained in solution of lower alkalinity tend to be of smaller magnitude and more scattered than those obtained in solutions of higher alkalinity. This effect is especially pronounced a t the higher p H values where the reactions are rapid. This discrepancy could be caused by the slow response of the HCOa--COZ buffer system to localized acidity changes induced by the oxygenation reaction. However, the possibility of a slight base catalysis by C O - * cannot be excluded. Catalytic Effects. Cupric ion is an efficient catalyst in many oxidation reactions and its catalytic effect has been studied in the oxygenation of ferrous iron in acid solutions (6. 72. 7 4 ) . Figure 3 shows that even trace quantities of C u+ * have a pronounced effect on the reaction rate. The catalytic effect is considerable= for C U -~ 3 X 10-izVf, but increases only sloivly with increasing C U P . These results are qualitatively in agreementVOL. 53, NO. 2 FEBRUARY 1961

145

亚铁的氧化

401

1

\

&

5-

I3b 1’0

\io3Io

40

50 1

610

70 1

*b

TlME.min

Figwe 3. C U+~catalyzes the oxygenation reaction20.5’

C. Po,= const.

pH= 6.0

with results obtained in acid systems by Cher and Davidson (6). The results of these authors have been accounted for by the following m

echanism: Fe(I1) CU(I)

+ Cu(I1) 7+ Fe(II1)CuV) ( a )

tion rate. The addition of various amounts of complex formers such as glycine, tetramine, and ethylene diamine tetraacetate did not affect the reaction rates. The latter two complexing agents markedly retard the oxygenation reaction in the absence of CuT2. Copper complexes appear to be as effective catalysts as the aqueous copper ions. Some results of a preliminary survey on the effects of other catalysts are represented in Figure 4 illustrating that other metal ions (MnT2, Co-2) as well as anions (besides OH-) which form complexes with F e ( I I I ) ( H? P 0 4 - . metaphosphate) are capable of hastening the oxygenation reaction. This figure shows that the reaction in Cambridge tap water which contains no analytically detectable C U+ is considerably faster than~ the reaction in synthetic water of similar mineral composition. Addition of chloride and sulfate (10 mg. per liter) had no effect on the oxidation rate. Practical Aspects Oxidation of ferrous iron is only one step in the deferrization process. Flocculation and sedimentation (or filtration) certainly may codetermine the over-all iron removal rate in surface waters and in iron removal plants. Many constituents present in natural waters may accelerate or decelerate the oxidation as well as the flocculation reaction. It can be inferred from our experiments, however. that the oxidation reaction may be the controlling factor up to pH values around 7. In the more alkaline p H range, flocculation may be the slowest step. At pH values above 8, diffusion or the oxygen rather than the chemical reaction determines the oxidation rate. In many natural waters organic matter may stabilize ferric oxide colloids or may increase the Fe(II1) solubility by complex formation. Generally, iron removal is very slow under such conditions (5, 79). Most organic impurities are known to hasten oxygenation reactions; under these conditions, difficulties in iron removal are apparently due to slow flocculation and cannot

necessarily be overcome by an increase in pH. Coagulation of iron(II1) is fast in many natural waters within the pH range 5 to 7 (4>5, 27) and ferric oxide removal by sand filtration is best within the same p H range (27). Under such conditions, however, oxidation rates are slow. Here the application of oxidation catalysts-e.g. Cu+2 may prove very efficient. For a given p H and oxygen concentration, the addition of as little as 0.02 mg. per liter reduces the oxygenation time by a factor of 5. Acknowledgment The authors wish to thank Carol Sinclair and N. S. Graham for excellent technical assistance. Literature Cited(1) Abel, E.. Z . Elektrochem. 59, 903 (1955). (2) American Public Health Association,

+

0 2

f

Cu(l1) HO,

+( b ) (8)

The HO’? is assumed to react further with Fe(I1) as in the Weiss mechanism (Equation 2). The data obtained do not permit the formulation of a meaningful rate expression which would account satisfactorily for

the catalytic effect of copper. However, severalqualitative statements can be made. The same proportionate acceleration of the oxygenation rate was observed at any pH value within the pH range investigated. The runs made in the presence of Cu-2 show the same temperature dependence as those made in the absence of C u f 2 . The catalytic effect of C U+~ dependent upon both is the initial ferrous and ferric iron concentrations. Higher initial ferrous iron concentrations and/or an incipient addition of ferric iron causes a n increased rate of reaction. There is a saturation concen~ tration of C U+ for any given initial ferrous iron above which further addition of Cu+2 has little or no effect on the reac-

New York.“Standard Methods for the Examination of Water Sewage and Industrial Wastes,“ 9th and 10th Cdition, 1946. 1955. (3) American Society of Testing Materials, Philadelphia, Pa..“Manual on Industrial Water.’‘ D. 264. 1954. (4) Bartow, E.,’Black, A. P., Sandsbury, W. E., Trans. Am. Soc. Czd Engrs.

100, 263 (1935). (5) Camp, T. R., Root, D. A, Bhoota, B. V., J . Am. Water W o r k s Assoc. 32,1913 (1940). (6) Cher, M., Davidson, N.. J . Am. Chem. SOC. 793 (1955). 77, (7) Feitknecht, W., Z. Elektrochem. 63, 34 (1959). (8) Feitknecht, W., Keller, G.; Z. anorg. u. allgem. Chem. 262, 61 (1950). (9) Gayer, K. H.? Woontner, L., J . Phys. Chem. 6C, 1569 (1956). (IO) George, P., J . Chem. SOC. 1954,4349. (11) Holluta, J., Eberhardt, M., Jb. Vom Wasser 24, 79 (1957). (12) Huffman, R. E., Davidson, N.. J . Am. Chem. Soc. 78, 4836 (1956). (13) Just: G., Z., Physik. Chemie 63, 385 (1908). (14) Lamb, A . P., Elder, L. W., . Am. J Chem. SOG. 137 (1931). 53, (15) Latimer, W. M.,”Oxidation Potentials,” p. 222, Prentice Hall, New York, 1953. (16) Lee, G. F., Stumm. W.,. A m . W a t e r J M’orks Assoc. (in press). (17) Leussing, 0. L.. Kolthoff, J. M., J . Am. Chem. SOC. 2476 (1953). 75, (18) Milburn. R. M.. Ibid.. 79, 537 (1957). (19) Moore, E. W., Snow. E.. A, j .:Yew Engl. Water Works Assoc. 56, 320 (1942). (20) Posner, A. M., Trans. Faraday SOC. 49, 382 (1953). (21) Stanley, D. R.: P h . D . thesis. Harvard University. Cambridge. Mass., 1952. (22) Stumm, W.: Proc. Am. Soc. C i d Eners. San. Engr. D i t . 86, 21 (November1 o .. . ., m

Figure 4. Insidious trace quantities of impurities may strongly accelerate the oxidation reaction

(23) S;Umm,\V.. Lee. G. F., Rerm. suisse d’Hydrologie 22, 295 (1960). (24) Todt, F., Stoklassa, K.. Z . Elektrochem. 5 8, 354 (1954). (25) Weiss, J., .~atlirreis~enschaften 23, 64 (1935).

j0.3 0

,5

,+g~

,20

,40

160

RECEIVED review April 12, 1960 for ACCEPTED October 31, 1960 Division of Water and Waste Chemistry, 137th Meeting, ACS. Cleveland. Ohio, April 1960. Work supported by the Public Health Service, Division of Research grants (Research grant 5831).

IO

T I M E, min.

146

INDUSTRIAL AND ENGINEERING CHEMISTRY

本文来源：https://www.bwwdw.com/article/c2xq.html