INDUCED OXIDATION OF ARSENIC(III)
Aug. 20, 1963 [CONTRIBUTION FROM
THE SCHOOL OF
CHEMISTRY, UNIVERSITY
OF
2385
MINNESOTA, MINXEAPOLIS 14, MINX]
Arsenic(1V) as an Intermediate in the Induced Oxidation of Arsenic(II1) by the Iron(I1)Persulfate Reaction and the Photoreduction of Iron(II1). I. Absence of Oxygen1 BY R . WOODS,I. M . KOLTHOFF AND E. J. MEEHAN RECEIVED J A N U A R Y 12, 1963 The persulfate-iron(I1) couple induces the oxidation of arsenic( 111) The induction factor in the absence of oxygen a t high dilutions of persulfate and iron(I1) can become greater than 1 Iron(II1) and copper(I1) increase the induction factor, which can approach infinity The mechanism is accounted for by the postulation of the intermediate formation of arsenic(1V) which can oxidize iron(I1) and reduce iron(III), FeOH2+ being 90 times as reactive as Fe3+ At 25' the rate constant of the reaction of the sulfate free radical with arsenic( 111) is 21 times as great as t h a t of its reaction with iron(I1) The kinetics of the induced reaction are in agreement with the proposed mechanism Additional evidence for the intermediate formation of arsenic( IV)was obtained in a study of the photosensitized oxidation of arsenic(II1) in solutions containing iron(II1)
The reaction of iron(T1) with persulfate is known to proceed through two steps, the first being the formation of a sulfate free radical.*-? Fez+
+ S Z O ~ ~ - - + Fe(II1) + SO4- + S012-
(1)
Fe(II1) denotes all forms of iron in the trivalent state. I n the absence of other oxidizable species, the sulfate free radical oxidizes another ferrous ion Fez+
+ SO4- -+
Fe(II1)
+ Sod2-
(2)
The sulfate free radical produced by reaction 1 is able to induce oxidations and polymerizations in organic substance^^-^ and its presence is confirmed by the incorporation of S3j, from S3j-labeled persulfate, into polymer chain^.^,^ N o studies are found in the literature on the induced oxidation of inorganic constituents by the couple persulfate (actor) and iron(I1) (inductor) under the conditions where the acceptor does not react or reacts only very slowly with persulfate. In the present paper, in which all the experiments have been carried out in the absence of oxygen, it is shown that the above couple induces the oxidation of arsenic(III), the oxidation being initiated by the reaction
so,- + A%s(III)----P As(1V) + SO42-
and/or by the hydroxyl free radical, formed in the reaction SO,-
+ H20
--f
HSO4-
+ OH'
Csanyi' postulated the reaction between SO4?-and As(II1) t o occur in the titration of a mixture of persulfate and hydrogen peroxide with one-electron transfer oxidants. He also proposed the formation of As(1V) in the induced reduction of chlorate by arsenic(III).8 Daniels and Weissgproposed the production of As(1V) by the reaction of a hydroxyl radical with arsenic(II1) t o occur during the oxidation of arsenite by the action of X-rays. Very recently Daniels'O also concluded that this reaction takes place in the photosensitized reaction between hydrogen peroxide and arsenite. In order to substantiate the interpretation of the induced oxidation of As(II1) by the persulfate-iron(I1) couple, it was thought desirable to investigate the induced oxidation of arsenic(IT1) upon photochemical reduction of iron(II1). I n the present paper, results are presented on the oxidation of arsenic(II1) upon ultra(1) This investigation was carried out under a grant from t h e National Science Foundation (2) J H. Mertz and W . A . Waters, Discussions Faraday Soc., 2 , 179 (1947). ( 3 ) J W . L Fordham and H L . Williams, J A m . Chem S o c . , 73, 1634, 4855 (1951) (4) I . M Kolthoff,A. I Medalia, and H P Raaen, ibid., 73,1733 (1951). ( 5 ) W. V . Smith, ibid., 71, 4077 (1949). ( 6 ) I M Kolthoff,P. R . O'Connor, and J L. Hanson, J Polymer Sci , 15, 459 (19.55) ( 7 ) L. J . Csanyi, Discussions Faraday Soc , 29, 146 (1960). (8) I> J Csanyi and ? V I . Szaho, T a i a n f a , 1, 359 (1958). (9) M . Daniels and J . Weiss, J . Chem. Soc., 2467 (1958) ( I O ) M Daniels, J . Phys Chem., 66, 1473, 1475 (1962).
violet radiation of a mixture of iron(II1) perchlorate and arsenic trioxide. Evans, Santappi, and Uri" studied the photochemical reduction of iron(II1) by irradiation with the 3630 and 3130 A.lines of a mercury arc and its initiation of vinyl polymerizations. They concluded that the photochemically active species is an ion pair complex of the formula Fe3+X- (where X- = OH-, C1-, N3-, etc.), e . g . F e 3 + OH-
-+- F e z + + OH
This conclusion has been confirmed by other workers. l2-I4 The production of the hydroxyl radical was also postulated for the photoreduction of cerium( 1 ~ 15.16 ). Experimental Materials. Water.-Conductivity water was distilled from alkaline permanganate and redistilled under nitrogen. Arsenic Trioxide .-Mallinckrodt AS203 was recrystallized from dilute hydrochloric acid (1 : 2 ) ,washed, dried, and sublimed. The product was free of chloride. An approximately 0.1 M stock solution was prepared by boiling the oxide with water. Potassium Persulfate.-Merck K&08 was recrystallized twice. Perchloric acid was G . F . Smith reagent grade. Ferrous Perchlorate.--Xn excess of iron wire (99.9% F e ) was added t o standard perchloric acid in an atmosphere of nitrogen and, after the reaction was completed, the solution was filtered, acidified with standard perchloric acid, and diluted. The solution, which was 0.1 &' in perchloric acid and 5.21 X M in iron( I I ) , was stored under nitrogen. Ferric Perchlorate.-G. F. Smith reagent grade ferric perchlorate was recrystallized from perchloric acid solution. Cupric perchlorate was G. F. Smith reagent grade. Sodium perchlorate was G . F. Smith reagent grade recrystallized from water. Nitrogen.-Linde nitrogen (99.99% S 2 )was further purified from oxygen by passage over a column of finely divided copper deposited on infusorial earth and heated to 200"." The oxygen content was thereby reduced t o the order of 4 X Reaction Vessels .-Pyrex flasks were cleaned with concentrated nitric acid which was 5%; in hydrofluoric acid, rinsed thoroughly, and steamed out for 2 hr. Analytical Methods.-Iron( 11) was determined spectrophotometrically with o-phenanthroline,16the pH being adjusted where necessary to about 3.5 by addition of sodium citrate. I t may bz noted t h a t spectrophotometric titration with cerium( I V ) a t 0 using ferroin as indicator,lg which ordinarily yields good results in acid medium (pH < 2 ) , cannot he used in the presence of much perchlorate because of precipitation of ferrous phenanthroline perchlorate a t 0". Arsenic(V) was determined spectrophotometrically*" by the formation of 12-molybdoarsenic acid, extracting this species into 2-butano1, and subsequent reduction with stannous chloride t o molybdenum blue, the absorption being measured a t 740 mM. (11) M . G Evans, hl Santappi, and K. Uri. J . Polymer Sci., 7 , 243 (1951). (12) J . Saldick and 0 Allen, J A m Chem. Soc., 77, 1388 (1936). (13) J. H. Baxendale and J. hfagee, Traits. Faraday Soc., 51, 205 (1955)' (14) F. S. Dainton and M Tondoff, ibid , 53, 666 (1957) (15) J. Weiss and L). Porret, .\-aluve, 139, 1019 (1937) (16) T . J Sworski, J A m . Chem S o c , 77, 1074 11956). (17) F. R . Meyer and G . Ronge, 2: angeu'. Chem , 5 2 , 637 (1939)' (18) E . 13. Sandell, "Colorimetric Determination of Traces of Metals," 3rd E d . , Interscience Publishers, I n c , S e w Y o r k , X . Y . , 19.59, p ,537 (19) I hI Kolthoff and R Woods, Mzciociiem. J . , 5 , 669 (1961) (20) h l Daniels, A n n l y s l , 133, 82 (19.57)
2386
R. WOODS,I . M. KOLTHOFF AND E. J. MEEHAN
The extraction technique was necessary to avoid oxidation of arsenic( 111) by perchlorate which occurs when molybdoarsenic acid is reduced by hydrazine sulfate in the reaction mixture.21 In the absence of perchlorate such extraction is unnecessary. Procedure.-The ionic strength was kept constant a t 0.50 by addition of sodium perchlorate in all experiments with t h e exception of those of Table I1 (effect of ionic strength) and of Table I11 (in which a low ionic strength was desirable, oide infra). Induced Oxidation by the Persulfate-Iron(I1) Couple. Determination of Induction Factor .-Solutions of ferrous perchlorate, perchloric acid, arsenic trioxide, a n d , where relevant, ferric or cupric perchlorate, were pipetted into flasks covered with black adhesive tape to eliminate photoinduced reactions, deaerated with nitrogen, and placed in a thermostat bath at 25'. Potassium persulfate, also deaerated with nitrogen, was added from a syringe and t h e solution mixed. The reaction was allowed t o go t o completion and the solution analyzed for iron(I1) and As(\'). An excess of iron(I1) was always used. Rate of Reduction of Persulfate in the Presence of Iron(II1). -Solutions of ferrous and ferric perchlorate, perchloric acid, and arsenic trioxide were pipetted into tape-covered flasks, deaerated with nitrogen, and placed in a thermostat bath at 25'. Deaerated potassium persulfate was added as before. Aliquot samples were removed at desired times, entrance of oxygen being prevented by continuous bubbling of nitrogen, and analyzed polarographically for persulfate remaining. One drop of 0.5% o-phenanthroline was added t o the polarographic cell t o arrest the reaction. Sodium fluoride in a concentration of 0.1 M was used as supporting electrolyte t o eliminate interference by iron(II1) at the d.m.e. Polyacrylamide in a concentration of O . O l ~ owas used as a maximum suppressor. The current was measured at 0.15 volt 21s. s.c.e. and shown t o be proportional t o persulfate concentration. Oxygen did not interfere and was not removed from the polarographic cell. Photochemical Experiments .-The light source consisted of two 200-w. tungsten bulbs fitted with a reflector and a diffusing screen t o give a n evenly distributed illumination. The light passed through t h e glass wall of a thermostat bath before reaching the samples. Glass bottles were used as reaction vessels. The absorbance of glass approaches infinity for wave lengths less than 300 mp. Ferrous perchlorate does not absorb radiation of wave length greater than 300 mp and therefore the photooxidation of i r ~ n ( I I ) will ~ ~ -be ~ ~negligible during this experiment. Ferric perchlorate, however, is photoreduced by radiation of wave lengths greater t h a n 300 rnp.ll Solutions containing known concentrations of ferrous and ferric perchlorate, perchloric acid, and arsenic trioxide were placed in the water b a t h , deaerated with nitrogen, and the light source turned on. Aliquot samples were taken, and the iron(I1) and arsenic(V) formed were determined as described above.
Results The Induced Oxidation of Arsenic(II1) by the Persulfate-Iron(II1) Couple.-The persulfate oxidation of arsenic trioxide is very slow compared with its oxidation of iron(I1). However, the iron(I1)-persulfate reaction induces the oxidation of arsenic trioxide. The induction factor for the induced oxidation was found t o be dependent on perchloric acid concentration (Table I) but independent of arsenic trioxide concentration provided the latter was greater than twice the ferrous concentration. An increase in either [H+] or [Fez+]produced a decrease in the induction factor. I n the presence of initially added iron(II1) the induction factor is shown to increase to infinity. This "catalytic" effect of iron(Il1) is more pronounced with decreasing acidity. Reduction of ionic strength to 0.01 or 0.002 caused an increase in both I.F. and k4 !ka compared t o the values a t the same [H +I and p = 0.5 (Table 11). The rate of reduction of persulfate at an initial conJf in the presence of arsenic trioxide centration of and ferric perchlorate was measured polarographically after addition of small amounts of ferrous perchlorate, under conditions where the induction factor is infinity. (In this set of experiments the ionic strength was 0.07, rather than 0.5, in order to have the maximum fraction (21) E B Sandell, ref. 18, p . 282 ( 2 2 ) C . Heller. P h . D . Thesis, University Microfilms, Ann Arbor, Mich , Mic 80-490. 19,57 (23) E . Hayon a n d J . Weiss, J . Chem. S o c , 3866 (1960). Ilin, W. B. M a r t i n , J r , and A. M J . Beatty,
VOl. 85
TABLE I VARIATION OF INDUCTION FACTOR ( 1 . F . ) AT CONSTANT [S2082-], [As(III)], AND IONIC STRENGTH, p , WITH VARYIXCIROS(II) CONCENTRATION ASD ACIDITY equiv. As(II1) oxidized I . F . = equiv, Fe(II) oxidized [s2oS2-i = 2.5 x 10-5 M ; Ir = 0.5; T = 25" [Fez+], X 105
[Fe2+], X 105 [ F e ( I I I ) ] ,
[Fe(III)],
M
I.F.
'1.I
kr/kaa
M
0.50 Ak' HClO, 5.71 4 56 4 56 5 X 4 56
0.17 0 23 15 4 6
0.20 14 10 14
M
I .F .
ka/ks"
0.005 M HC10,
6 85
075 5 71 087 4 56 107 2 74 180 l4 2 74 5 X 5 75 0.1 MHClO4 2.74 10.4 5 71 0.20 0.26 Av. 4.56 0.32 27 4 . 5 6 5 x 10-4 2 . 7 .21 0.002 M HC104 4.56 10-3 5 . 2 .21 095 '4V 0 24 125 0 010 M HC10, 132 4 56 274 210 6 85 0 56 1 6 205 250 5 71 68 1 6 205 10-5 4 38 4 56 88 1 8 1 5 20525X10-567 2 74 1 4 16 2 05 5 X 10 1 2 74 6 8 2 05 10-5 2 74 m
2 6 2 5 2 6 2 9 2 6 2.6
~
-
2.6
;!
3 4 3 4 3 4 4 3
__ 4 0
-
AV
1 6
7 6 8 2 5 5 1 6
Av
See Discussion
[S2082-]
[HC104],
M
0.010
TABLE I1 EFFECT O F I O S I C STREXGTH = 2.5 x 10-5 M , [4s(III)] = M , T = 25" [Fez +], [Fe(III)1 , I.F. kdka x 105 M M 0.010 6.85 ... 0.83 3.1 5.71 ... 1.03 3.4 4.56 ... 1.23 3.3 2.74 ... 1.86 3.1 12.9 3.2 2.74 10-4 Av. 3 . 2
0.002
0.002
6.85 4.56 2.74 2.05 2.05
...
... ... ... 2 . 5 X 10-5
1.4 1.54 2.22 2.9 8.4
Av.
5.0 5.1 4.7 5.3 5.4
5.1
of iron(I1I) present as FeOH2+ rather than as Fe3+.) The reaction was found to be first order to persulfate with the rate constant proportional t o the iron(I1) concentration (Table 111). If the rate-determining step in this
lo-*
TABLE I11 RATEOF REDUCTION OF PERSULFATE Fe( II:), 0.01 A4 HC104, M S ~ O B ~0.01 - , M As( 111), 0.01 I.F. = m , p = 0.07, T = 25 [Fez+], 105 M
x
0 57 1 14 2 28 4 57
sec
-1
ki, X 10'
184 3 75 6 38 12 5
k , = k/[Fez'I, I mole-' sec - 1
32 33 28 27
system is the iron(I1)-persulfate reaction and the iron(11) is regenerated at the same rate a t which it is oxidized, the rate of reduction of persulfate, -d [S20e2-]/dt, will be
INDUCED OXIDATIONOF ARSENIC(III)
Aug. 20, 1963
pseudo first order with a first-order rate constant given by k = kl[Fe2+](where kl is the rate constant of the iron(I1)persulfate reaction). The rate constant kl can be calculated from the experimental data (Table 111) and gave the average value 30 1. mole-' set.-'. The rate of the iron(I1)-persulfate reaction in the absence of arsenic(II1) was measured in the presence of the same concentrations of ferric perchlorate and perchloric acid, the ionic strength of the solution therefore being the same as that in the experiment in the presence of arsenic(II1). The iron(I1) concentration was determined a t given periods of time. The rate constant was found to be 27 1. mole-' set.-', which agrees favorably with the rate constant obtained above. These results show that the iron(I1)-persulfate reaction is the rate-determining step in the induced oxidation of arsenic(III), that only iron(I1) reacts with persulfate, and that, in the presence of sufficient iron(III), iron(I1) is re-formed a t the same rate a t which it is oxidized. Copper(I1) behaves similarly t o iron(TII), producing an increase in the induction factor and giving a limiting value approaching infinity (Table IV). TJnlike iron(111), however, its effect is virtually independent of the hydrogen ion concentration. TABLE IV ON INDUCTION FACTOR EFFECTOF COPPER(II) M ,[S208z-]= 2.5 X 1,0-5M,[As(III)] = [ F e z + ]= 5.71 X M, 1.1 = 0.5, T = 25
Molar concn. [Cus+l0
HC104,M
10-6 M -Induction facto-
0.17 0 20 0.68 1.25
0.50 .10 ,010 ,002
10-4 M
1.6 1.7 1.9 1.9
12 13 10 9
2387
5.71 X M in iron(TI), and 0.01 M in arsenic(III), the induction factor decreased from 1.25 in the absence of fluoride to 0.82, 0.18, and 0.10, respectively, in the presence of 5 X loF5,5 X and 5 X Msodium fluoride. I t was found that fluoride did not affect the rate of the iron(1I)-persulfate reaction, and therefore this cannot be the cause of the reduction in the induction factor. The Photoinduced Reaction between Iron(II1) and Arsenic(II1) .-Radiation of wave lengths greater than 300 m p was found to induce a reaction between iron(II1) and arsenic(III), in the absence of oxygen, according to the stoichiometric relationship: 2Fe(III) f As(II1) + 2Fe(lI) As(V). The initial rate of formation of iron(I1) was measured in the initial presence of l o p 4 .If Fe(III), 5 X M Fe(II), 0.01 M HClOI, j~ = 0.5, and was found to be independent of the concentration of arsenic(II1) when this was varied from 0.025 to 10F3 M . When M, the initial rate [As(III)]was decreased to 5 X of formation was reduced to about half of the value found a t the larger concentrations of arsenic(II1). The initial rate of formation of iron(I1) also was measured in 0.01 M As(I1I) and the same [H+] and p as above, and varied [Fe(III)]. It was found to be proportional to [Fe(III)] up t o about M , and t o increase less than proportionally t o [Fe(III)] a t higher concentrations. The effect of initial [Fe(II)] upon d[Fe(II)]/dt was neasured in solutions M in Fe(III), 0.01 A4 in As(III), either 0.01 or 0.005 ,Win HC104, ( p = 0.5), and M initially in Fe(I1). Figure 1 from 0 to 1.25 X shows that the reciprocal of the initial rate increases linearly with initial [Fe(II)1.
+
The effect on the induced oxidation of reducing the ratio of concentrations of arsenic(III) to iron(I1) to values less than two was investigaled in the presence of iron(1II). The concentration of iron(II1) in each case was sufficient to yield an induction factor of infinity when theratio [As(III)]/ [Fe(II)]wasequal to or greater than two. TABLE V OF ARSENIC(III)IN PRESENCEOF EFFECTOF CONCENTRATION Fe( 111) ON I.F. M , [SzO,z-] o= 2.5 X M,1.1 = 0.5, [ F e z + ] = 5.0 X T = 25 [HClOrI,
[Fe(III)I,
M
0.10
M
2
x
10-2
0.010
2
x
10-3
0,002
5
x
10-4
[As(III)I, x 106 M
2.5 5.0 10.0 2.5 5.0 10.0 2.5 5.0 10.0
I.F.
kr/kn"
4.55 13.7
18
17
0
2 5
..
m
5.75 13.3
19 24
..
W
6.35 15.7
27 22
..
m
Av.
a
0
21
See Discussion.
Iron(II1) is produced during the reaction and is therefore present in all experiments in which the induction factor was determined (Table I). In order to deduce the course of the reaction in the absence of ferric ions, the effect of the addition of fluoride, which complexes the iron(III), was determined. Fluoride was found to reduce the induction factor to values approaching zero. For example, in an oxygen-free mixture 0.002 M in perchloric acid and originally 2.5 X M in persulfate,
5
75
10
12 5
[Fe?'J X 105 M
Fig 1.-The effect of initial [ F e z + ]on d [ F e 2 + ] / d t ; [Fe3+l = 10-4 M,As(II1) = 0.01 M ,[HClO,] = 0.005 M or 0.01 M . I , 0.01 M HC104; 11. 0.005 M HCIO,
Discussion The induced oxidation of arsenic trioxide by the persulfateeiron(l1) couple in the presence of an excess of iron(II1) can be explained by the intermediate formation of the 4+ oxidation state of arsenic according to the reaction mechanism Fez+
ki + &Oxz- --+ Fe(II1) 4-SOa2- + SO4? slow
As(II1)
ks + SO47 + .4s(IV) + Solzfast
(1) (3)
Vol. 85
R . WOODS.I. M. KOLTHOFF AND E. J. MEEHAN
2388
The stoichiometry of this mechanism is in accordance with the induction factor of infinity found when sufficient iron(II1) was added. ‘The mechanism is substantiated by the kinetics of the reduction of persulfate (Table 111) which show that reaction 1 is the rate-determining step, that only iron(I1) reacts with persulfate, and t h a t iron(I1) is re-formed at the same rate that it is oxidized. Iron(I1) and arsenic(II1) compete in reacting with the sulfate free radical Fez+
+ SOa-
kz
Fe(lI1)
+ SOa2-
(2)
Reaction 2 becomes insignificant compared with reaction 3 a t arsenic(II1) t o iron(I1) concentration ratios greater than two. Under these conditions the induced oxidation becomes independent of arsenic(II1) concentration. However, as this ratio is reduced to less than two, reaction 2 becomes significant and a reduction of the induction factor from its value of infinity occurs (Table V). Assuming steady state kinetics with respect to As(1V) and SO4 T, the rate of oxidation of iron(I1) will be given by
and the rate of oxidation of arsenic(II1) by
[ F e ( I I I ) ] = [Fe(III)IOf [ F e 2 + l o- [ F e z + ]
where [Fe(III)lo and [Fe2+loare the initial concentrations. Substituting this relationship in the differential equation and integrating gives the relationship
(1 - -?3([FeZ+]o- [ F e z + ] , ) + (where [Fez+] and [SZO8*--]oare the remaining iron(I1) concentrat ion and the initial persulf a t e concentrat ion, respectively), provided excess of iron(I1) is used and, therefore, all the persulfate is reduced. The only unknown in this equation is the ratio of the rate constants for the reactions of As(1V) with iron(I11) and iron(II), ka/kb, which can therefore be calculated. Table I shows that within the experimental error, the ratio k 4 / k 6 is a constant for varying [Fe2+]and [Fe( I I I ) ]a t a given pH, substantiating the proposed mechanism. The ratio increases with decreasing hydrogen ion concentration (vide znfra). The photosensitized Fe(II1)-As(II1) reaction is also interpreted in terms of the formation of As(IV) and its reaction with iron(I1) and -(HI). The photoactive species in the photoreduction of iron(II1) is the ion pair complex FeOHz+ l 1 and the reaction will proceed ks
Fe3+OH- +Fe2+OH.primary step
(6)
ki
Fe2+OH.--+- F e 3 + 0 H - back reaction
Therefore
ks
d [ A s ( I I I ) ] 1 ka [ A 4 ~ ( I I I ) ] d [ F e Z + ] = Zkz[Fe2+]
F e 2 + O H .+F e z +
kz
Fez+
In ( [ 4 s ( I I I ) ] o / [ . Z s ( I I I )m] ) In ( [ F e z + ] ~ / [ F e 2m+ 1
ks + F e z ++ .4slIII) + Fe(II1)
and the rate of reduction of persulfate by =
ki[Fe2’]
back reaction
+ OH. --+ As(IV) + OHAs(IV) + Fe(II1) +As(V) + Fez+ ka As(IV) + F e z + +As(II1) + Fe(II1)
(9)
klo
As(II1)
[s2oBZ-]
(10)
k4
(4) (5)
The over-all stoichiometry of this reaction is in accordance with the experimental result 2 F e ( I I I ) f As(II1) ---+ 2Fe2+ f As(\’)
The rate of formation of [Fez+]in the photochemical reaction 6 is given by d [ F e 2 + ] / d t = kl,t,,
(5)
Reaction 5 also explains the decrease in the induction factor brought about by an increase in iron(I1) concentration a t constant pH (Table I ) . In the induced oxidation of arsenic trioxide by the persulfate-iron(I1) couple, both ferrous and iron (111) are present and As(IV) reacts with both. Assuming steady state with respect t o As(1V) and SO4-, the rate of oxidation of iron will, therefore, be given by
-d[S2082-]/dt
+- OH. +Fe2’OH.
The hydroxyl radical produced in reaction 8 induces the oxidation of As(II1)
[As(III)]o, [Fe2+Ioand [As(II1)lm,[Fez+], being the initial and final concentrations, respectively. Table V shows t h a t kS/k2 = 21 and is independent of acidity. In all other experiments of this paper the arsenic(II1) concentration greatly exceeds the iron(I1) concentration and therefore reaction 2 need not be considered in the reaction mechanism. When iron(II1) is complexed with fluoride, the induction factor is reduced t o a limiting value approaching zero. In the absence of iron(III), therefore, the As(1V) produced is postulated t o react with iron(I1) to re-form As(II1). As(I1.)
(8)
kg
Integrating over the whole reaction gives k“,2
+ OH. secondary step
(7)
=
k l o ( l - exp(--ab[FeOHz+])j
where I o and I&s are incident and absorbed intensities, respectively, a is molar absorptivity of FeOH2+, and b, is the path length; k is a proportionality constant. At small [FeOH2+]this simplifies to d[Fez*l/dt
=
kloaO[FeOH2+]= k’[FeOHz+]
Hence a t small [FeOH2+]the net rate of formation of [Fez+]in the presence of As(II1) is given by d [ F e Z C ] / d t= k’lFeOH2+1
[FeOH2+]is related t o the total ferric iron present, [Fe(111)1, by the relationship
Combining these equations gives the relationship where The iron(II1) concentration is given by the initial concentration plus that formed by the oxidation of iron during the reaction
K,
=
[FeOH2+][H+1/[Fe3+]
The term containing [Fez+]/[As(III)] must be negligihle as no effect of varying As(II1) concentration a t con-
Aug. 20, 1963
2389
INDUCED OXIDATION OF ARSENIC(III)
stant Fez+ was found under conditions where the As(111) was in large excess. Theref ore
The initial rate of production of iron(I1) when [Fez+]-t 0 will be independent of [As(III)] and, a t small [Fe( I I I ) ] , proportional to [ F e ( I I I ) ] . I t was found experimentally that the former relationship is obeyed and that the production of ferrous iron increases with increase in [Fe(III)]; the deviation from exact proportionality is due to the difference between ( 1 - exp(ab. [FeOH2+]) } and ab [FeOHZC]. At constant [ F e ( I I I ) ]and [ H + ] ,the reciprocal of the rate of formation of iron(I1) is predicted from the mechanism to be linearly related to the iron(I1) present. Experimentally, this was found to be the case (Fig. 1). The ratio k 4 / k 6 can be obtained from the slopes and intercepts of the plots of Fig. 1, and has the values 1.4 and 2.3, respectively, in 0.01 M and 0.005 M perchloric acid ( p = 0.5). The values agree with experimental error with the entirely independent values (cf. Table I ) of 1.6 and 2.6 a t the same acidities obtained from the induced oxidation by the persulfate-iron(I1) couple, further substantiating the proposed mechanism. The ratio of the reaction rate constants k4/k5 is dependent on the hydrogen ion concentration. Table VI shows that the ratio increases as the hydrogen ion concentration is decreased. At hydrogen ion concentrations less than 0.01 the ratio is approximately proportional t o I / ( [H+I f K a ) ,taking K , t o be 1.9 X 10V3a t p = 0.5 in perchlorate medium,25i.e., ( [ H + ] 4-Ka). k4/k5 is constant a t [ H + ]less than 0.01. TABLE VI DEPENDENCE OF k4/kaON [H +] ; p HCIOI, M
k4/kS
0.50 .10 ,010 005 ,002
0.14 0.24 1.6 2.6 4.0
=
+
Fe3+
+ As(1V)
Fez+
--f
Fez+
+ As(\’)
+ As(V)
+
(4’) (4”)
This leads to the expression k4 = k4‘ _ _ k6
kb
k
K, ([H+l
[H+]
+ Ka) + L (W+I + K s ) 11
~~
k6
+
AS shown in Fig. 2, a plot of ( [ H + ] Ka)k4/k6 US. lHClO41 yields a straight line with a slope of k4jt/k5 = 0.1 and an intercept of K a k I f / k s = 0.017 corresponding = 9. t o a value of k4!/k5 = (1.7 X 10-2)/(1.9 X The fact that k4’ is 90 times as large than k 4 / /explains the constancy of ( [ H + ] K , ) k 4 / k s a t low acidity as the term in [ H + ]will then become negligible. The effect of ionic strength p on the ratio k4/k6 also substantiates the assumption that FeOH2+is the principal reacting species a t low [ H + ] . The dependence of the ratio [FeOH2+]/[Fe3+]upon p and [H+] can be
+
(25) R . hl. Milburn and W. C Vosburgh, J . A m . Chem. SOC.,7 7 , 1352 ( 1955)
0.5
+ K,)kr/kr against
[HC104].
calculated from the data of Milburn and V o ~ b u r g h . ~ ~ When [H+] = 0.01 M , this ratio of concentrations increases 1.9-fold when p is changed from 0.5 to 0.01 M ; and when [ H + ] = 0.002 ill, the ratio increases 1.5-fold when p is changed from 0.5 to 0.002. The observed values of k4/k5 (Tables I and 11) upon decrease of p increase by 2.0-fold and 1.3-fold a t [ H + ] = 0.01 M and 0.002 M , respectively, in excellent agreement with the predicted values. The effect of copper(I1) on the iron(I1)-persulfate induced oxidation of As(II1) is similar to that of iron(l11) and gives a limiting induction factor apparently approaching infinity which is explained by the mechanism As(IV)
kii + C u z ++ As(V) + Cu(1)
(11)
followed by either kin + Fe(II1) + Cuz+ + Fez+
kia
we must consider t h a t in the range of acidities used iron(II1) is present as FeOHz+ and Fe3+. I t is reasonable t o assume that both species react with arsenic(IV) but with different rates. kc”
of ( [ H + ]
Fig. 2.-Plot
CU(I)
0.070 024 ,019 ,018 ,016
kr‘
0.4
[ H C l o ~ lM , .
([Hi] K a ) k 4 / k a
--f
0.3
0.2
0.1
(12)
or
0.5
In order t o account for the constancy of ( [ H + ]
+ As(1V)
i .
0
Cu(1)
K a ) k r / k s a t low acidity and its increase a t higher acidity
FeOHz+
0’01 0
S Z O ~ ~ - + CUz+
+ SOaz- + SO47
(13)
The induction factor in the presence of copper(I1) was found to be practically independent of pH (Table IV). This is to be expected as copper(I1) does not form any species comparable to FeOH2+ at the acidities used. Also the independence of the induction factor of hydrogen ion concentration in the presence of copper(I1) excludes the possibility that the reaction in eq. 5 is [ H + ] dependent. I t is possible to estimate the value of kll/’k5; ix.,the ratio of the reaction rate constants for the reactions of As(IV) with Cu2+andFez+,from the effect of copper(I1) on the iron(I1) -persulfate induced oxidation of arsenic(111). From Tables I and IV it is seen that the effect of copper(I1) is approximately equivalent to that of the same concentration of FeOH2+and therefore kll/k5 must be approximately equal to k4)/k5. The over-all mechanism is therefore
+
ki
+ S042-+ SO4F e z + + SO4- +Fe(II1) + S O A ~ ks As(II1) + SO47 As(1V) + SO,’ko As(1V) + FeOHZ+-+ As(V) + Fez’ kdt? As(1V) + Fe3+ +As(V) + F e z + ks As(1V) + F e z + --+ As(II1) + Fe(II1) kii As(1V) -t C u z + --+ .4s(V) + Cu(1)
Fez+
&08z-
-+ Fe(II1)
(1)
kn
--f
(2) (3)
(4’) (4“)
(5) (11)
with the ratios k4’/k6 = 9, k 4 1 1 / k 6= 0.1, k l l / k s about 9, and kS/k2 = 21.
DONALD T. SAWYER AND
2390
Arsenic(II1) is not easily oxidized by one-electron oxidants without the addition of catalysts but is readily oxidized by two-electron transfer. Watersz6 suggests that a one-electron oxidation of arsenic(II1) can occur only when the oxidant is capable of withdrawing a hydrogen atom from the favored tautomer; e.g. OH
OH.
I + H--i\s=O
OH
-+- HzO +
JAMES
Vol. 85
E. TACKETT
the formulated Sod2-. This proposed structure of As(1V) could also explain why the ion pair complex FeOHz+ reacts faster with As(1V) than the Fe3+ does. The former reaction will be the direct transfer of a hydroxyl FeOHz+ f AsO(OH)Z+F e z +
+ AsO(OH)3
while the latter will involve electron transfer followed by the addition of hydroxyl ion
.?s=O
+ .AsO(OH)* --+ F e z + + +AsO(OH)2 + A S O ( O H ) ~ OH-
OH
OH
Feac
The sulfate free radical can also be considered as a hydrogen-extracting radical, producing HS04- instead of (26) W A Waters, D$scussions Faraday S O L ,29, 170 (1960)
[CONTRIBUTIoN FROM THE
DEPARTMEKT OF
I n conclusion, it is mentioned that oxygen has a great effect on the iron-arsenic-persulfate system, which is discussed in a subsequent paper.
CHEMISTRY, UNIVERSITY OF CALIFORKIA,
RIVERSIDE, CALIF.]
Properties and Infrared Spectra of Ethylenediaminetetraacetic Acid Complexes. V. and Structure of Several Metal Chelates in Solution BY DONALD T. SAWYER ASD
JAMES
Bonding
E . TACKETT
RECEIVED JANUARY 14, 1963 The ethylenediaminetetraacetic acid complexes of the alkaline earth ions, cobalt( 11), nickel( 11), copper(I I ) , zinciII), cadmium(II), manganese(11), lead(II), bismuth(III), cerium(III), aluminum(III), iron(III), chromium( I I I ) , thorium( I V ) , and vanadium( I V ) in aqueous solution have been studied by infrared spectroscopy using deuterium oxide as the solvent. The effect of solution acidity upon the spectra has been determined and from this stability constants have been estimated for several of the chelate systems The shift in peak position and area for the carboxylate-carboxylic acid absorption band has been used t o propose structures for the complexes in solution. The data indicate t h a t there is some tendency for metal-nitrogen coordination in the alkaline earth-EDTA chelates, but t h a t such bonding is present t o a much greater extent with the complexes of the divalent transition metal ions The structures proposed for the iron(III), aluminum(III), and chromium(II1) complexes are more complex and involve hydroxide groups as well as the E D T A ligand.
Although there have been a number of previous infrared studies of the metal-ethylenediaminetetraacetic acid (EDTA) complexes, these have been restricted to the solid phase; a previous paper in this series summarizes the literature. The major importance of metal chelates is their behavior in solution, and in the case of EDTA, in aqueous solutions. Absorption of infrared radiation by water and the limitations imposed by cell materials have restricted studies of aqueous solutions However, by using deuterium oxide and barium fluoride cells, certain regions of the infrared spectra for metal-EDTA chelates may be investigated, particularly the carbonyl region. The present study is concerned with the effect of solution pH upon the peak position and area for the carboxylate-carboxylic acid absorption band for a number of metal-EDTA complexes in aqueous solution. The metal ions considered include the alkaline earths, the divalent transition metals, lead(II), iron(111), aluminum (I I I ) , chromium (I1I ) , bismuth(II1), cerium(III), thorium(IV), and vanadium(1V). On the basis of the spectral changes with pH and the conclusions made earlier with respect to the structure of the free ligand,2a.bstructures are proposed for the various metal-EDTA complexes in solution. Experimental Equipment.-The infrared spectra were recorded with a Perkin-Elmer model 421 spectrophotometer equipped with a high resolution grating. Some preliminary spectra were recorded with a Perkin-Elmer model 221G. The aqueous solutions of the metal chelates were contained in barium fluoride cells with 0.025m m . spacers. The low solubility of this material in water prevented significant attack by the solutions; barium fluoride trans(1) D T . Sawyer and J M McKinnie, J . A m . Chem S O L . ,82, 4191 (1960) (2) (a) D . T . Sawyer and J . E. Tackett, ibid , 81, 314 (1963); (b) K. Nakamoto, Y. Morimoto, and A. E . Martell, ibid , 81, 309 (1963).
mits radiation down to 800 cm.-'. T h e p H of the solutions was measured with a line-operated Leeds and Northrup p H meter equipped with microelectrodes. With such electrodes the total solution volume could be as small as 0 . 5 ml. T h e meter was standardized with N.B.S. buffers. Reagents.-Deuterium oxide ( DzO) was used for the preparation of all solutions t o avoid the interfering absorption bands of the H20 molecule in the carbonyl region. The material was obtained from the Bio-Rad Laboratories, Richmond, Calif ., and had a n assay of 99.9c,b DzO. The disodium salt of 'EDTA ( a s the dihydrate from the J. T . Baker Co.) was used without further purification and had a n assay of 99.9yo. All other materials were reagent grade. The metal ions were introduced as the metal chloride or nitrate salts, with the exception of thorium( I V ) perchlorate. The metal-EDTA4 solutions were prepared by combining weiahed oortions of the disodium salt of E D T A ISazH2Y) and t h e m e t a i salt with DzO to give a n equimolar stock solution of ligand and metal ion. Concentrated S a O H or HCI was added t o aliquot portions of the prepared metal-EDTA solution t o give the desired p H . Because the major acidic species present in the solutions was the solvated deuterium ion, the indicated p H values on the p H meter had t o be corrected t o give true acidity by using the equation of Mikkelson and Nielson3 p D = "meter reading" 0.40 By measuring the peak areas for differing carboxylate groups, the relative number of such differing groups was determined. The peak areas were measured using the base-line technique and ascertaining the absorbance a t the base and at the peak. By taking the peak width a t the half-height (in terms of absorbance) and multiplying by the peak height the approximate concentration of a given absorbing species was determined.
+
Results The spectral shifts as a function of solution acidity (pD) for the alkaline earth-EDTA chelates in solution are quite similar, especially for the carbonyl region. A representative example of the effect of solution p D upon the carboxylate absorption band is given in Fig. 1 for the barium-EDTA complex. At low values of p D two peaks are observed, one a t 1720 cm.-' ( 3 ) K . Mikkelson and S. 0. Nielson, J . Phys. Chem., 64, 6 3 2 (1960).
