996
W. L. REYNOLDS AND I. M. KOLTHOFF
mental development is under consideration in our laboratory. Measurements might be made more rapidly in an aerosol stream which flowed through the cell without disturbing the sound field; e.g., through quarter-wave stubs terminating in low acoustic impedances in the inlet and outlet stream. By putting in the optical pathaa neutral light filter of linear gradient to vary the illumination along the direction of sonic motion, a phototube which viewed a particle moving rapidly across the cell would deliver an output signal of sonic frequency, with an amplitude proportional to that of the particle. A narrow band-pass amplifier to enhance the signal-to-noise ratio could be used in conjunction with a rapid electronic circuit to obtain com-
Vol. 60
bined particle count and size distribution, while the phase lag of the particle also could be studied with a rapid electronic phase meter, and correlated with the amplitude ratio as a further test of Sewell's theory and another method for determining particle size. Acknowledgments.-It is a pleasure to acknowledge our indebtedness to the Research Committee of the Indiana University Foundation for a grantin-aid of this work; to E. Eugenia Schmidt for making the numerical calculations of the theoretical amplitude ratios and phase angles given in Figs. 1 and 2; and to Jack Baird and Maurice Williams for the construction of the apparatus in the machine shop.
'
Y
THE REACTIONS OF FERROUS VERSENATE AND FERROUS PYROPHOSPHATE WITH CUMENE HYDROPEROXIDE1 BY W. L. REYNOLDS~ AND I. M. KOLTHOFF School of Chemistry, University of Minnesota, Minneapolis, Minnesota Received March 6, 1966
The rate constants of the reactions between cumene hydroperoxide and ferrous Versenate and ferrous pyrophosphate have been determined under various experimental conditions. The rate constant of the former reaction was 5.0 X 1O'O exp( --10,40O/RT) liter mole-' set.-' and was independent of pH over the range 3.7 to 10.3. The rate constant of the latter reaction was 2.0 X 108 exp( -8200/RT), 2.7 X 109 exp( -8900/RT), and 1.6 X 109 exp (-8400/RT) liter mole-' sec.-l at p H 4.8, 6.8 and 8.8, respectively. The stoichiometry observed in the presence and absence of acrylonitrile suggests that a termination reaction between polymer free radicals and the ferrous complex can occur a t low concentrations of reactants. In the presence of monomer the main products of reaction a t low reactant concentrations are the ferric complex, acetophenone and polymer.
Although the reactions between aquo ferrous iron and various organic hydroperoxides have received some attention very few investigations have been made of the reactions between complexed ferrous iron and organic hydroperoxides. Orr and Williamsa have studied the reactions between polyethylenepolyamine complexes of ferrous iron and cumene and p-butylcumene hydroperoxides. Boardman4 has reported on the reaction between potassium ferrocyanide and cumene hydroperoxide. From a polymerization viewpoint a knowledge of the kinetics and mechanisms of the reactions of ferrous Versenate and ferrous pyrophosphate with organic hydroperoxides is important since these reactions are frequently employed to supply free radicals for the initiation of polymerization in emulsion polymerization systems. Experimental Determination of Rate Constants.-The rate of disappearance of the ferrous complex was determined by measuring its diffusion current a t a rotated platinum wire electrode during the course of the reaction. This diffusion current was found proportional to the concentration of the ferrous complex. The diffusion current of the complex iron(I1) was measured at a potential where a film of oxide is formed on the (1) This work was carried out under the sponsorship of the Federal Facilitiea Corporation, Office of Synthetic Rubber, in connection with the Synthetia Rubber Program of the United States Government. (2) From a thesis submitted by W. Revnolda to the University of Minnesota in partial fuliillrnent of the requirements f o r t h e degree of Doctor of Pbilosopliy, June, 1955. (3) (a) R. J. Orr and H. L. Williarna, Disc. Faraday Soc., No. 14, 170 (1953): (b) J . Am. Chem. Soc., 76,3321 (1954). (4) H.Boardman, ibid., 71, 4268 (lD53).
electrode.6 When the current voltage curve is measured with a self registering apparatus the film formation is characterized by a wave with an apparent limiting current (see curve 1 of Fig. 1). Film formation is also responsible for the distorted shape of the current-voltage curve of the complexed ferrous iron (curve 2, Fig. 1) when measured with a polarograph. Upon subtraction of the residual current a diffusion current plateau was observed as is evident from curve 3 in Fig. 1 . Film formation did not interfere with the kinetic measurements when a potential was applied at whlch the film was formed and the limiting current of iron(I1) was measured. The current caused by film formation rapidly decays with time.6 Thus upon measuring a current-time curve the current is found to decrease rapidly to the value of the diffusion current of iron(I1). The apparatus employed has been described elsewhere .e Complexing agent was added to the buffer solution in the electrolysis cell, the solution deaerated, a known quantity of ferrous iron added and a potential of +0.7 v. us. the saturated calomel electrode (S.C .E.) applied to the electrode. After the limiting current had decreased to a constant value (the diffusion current of the iron(I1)) a known quantity of air-free standard cumene hydroperoxide( CHP) solution was injected into the electrolysis cell. A typical current-time curve obtained after addition of CHP is shown in Fig. 2 . From the value of the Fe(I1) diffusion current before C H P addition and the known concentration of the ferrous complex the value of i d / C for the electrode could be calculated. An excess of ferrous iron was always used and from the value of the Fe(I1) diffusion current a t the end of the reaction the final Fe(I1) was calculated. Hence the stoichiometry of the reaction was readily determined. The rate equation
(5) I. M. Kolthoff and N. Tanaka, Anal. Chem.. 26, 032 (1954). (6) I. M. Koltlioff and W. L. Reynolds, Disc.Faraday Soc., No. 17, 167 (1954).
..
REACTION OF FERROUS VERSENATE WITH CUMENE HYDROPEROX~DE
July, 1956
I
'
I
'
I
20-
997
1
-
-
4.8
4.0 -
15-
3.2-
52.4 -
f
-
(idla,
1.6-
-
08 -
Fig. 1.-1, residual current of reaction medium; p H 9.2; 2, limiting current of 1.00 x 10-4 M ferrous Versenate; 3, diffusion current of ferrous Versenate; 4, limiting current at constant potential. was tentatively adopted. In equation 1 (Fe)o is the initial total concentration of Fe(II), Ro is the initial concentration of CHP, x is the decrease in CHP concentration at time t , ko is the observed rate constant, ftnd 12 is the stoichiometric factor and was found equal to unity or two depending upon the experimental conditions. At the concentration levels employed for the measurement of ko,n was equal to two in the ferrous Versenate-CHP and ferrous pyrophosphateC H P reactions when acrylonitrile (AcN) was resent and in the ferrous pyrophosphate-CHP reaction wten monomer was absent. Rate measurements were not made for the ferrous Versenate-CHP reaction in the absence of monomer. I n the ferrous pyrophosphate-CHP reaction n was approximately unity when methyl methacrylate (MMAc) was present. Integration of equation 1 and substitution of ((Fe)o ns) = (Fe(I1)) = Kid where K (Fe)o/(id)o, 2 ((Fe)o'- Kid)/'?& and ((Felo nRo) = K ( i d ) m gives equation 2
-
-
designates a diffusion current. A plot of log us. t as in Fig. 3 gave a straight line in all The cases, the slope, s, of the line bekg k&(id),/2.30. observed intercept in all cases was in good agreement with the calculated value of log ((Fe)o/nRo). Therefore equation 2 was found to be valid and the rate constant, ko, was calculated from k~ = 2.30s/K(id),. Determination of Stoichiometry .-Since the final concentration of iron was determined as described and the initial CHP concentration was known the stoichiometry was readily determined. Determination of Acetophenone.-Acetophenone is one of the reaction products formed. When the concentrations of the reactants were s d c i e n t l y high (ca. 10-8 M ) the acetophenone formed from the C H P was determined polarographically with a dropping mercury electrode in alkaline medium. The total current measured at -1.7 v. us. S.C.E. was corrected for residual current and the diffusion current of the ferric complex to yield the diffusion current of acetophenone. By adding known quantities of acetophenone t o the electrolysis cell and measuring the increase of diffusion current the concentration of acetophenone in the reaction mixture could be calculated. When the initial concentration of CHP was too low (cu. 3 X 10-6M) to permit direct determination of the acetophenone concentration in the reaction mixture with a dropping mercury electrode two other methods of analysis were employed. One method, applicable when monomer was absent, consisted of the extraction of acetophenone with spectrophotometric grade isooctane followed by spectrophotometric determination of the extracted acetophenone. One hundred ml. of reaction mixture was taken; the initial Fe(I1) and CHP concentrations were 10 and 2.5 X 10-6 M , re[idlid
id
- (id)m]
0
Fig. 2.-(Fe)o
40
80
I20
I60
200
= 9.9 X M ; RO = 2.47 X AcN = 0.15 M ; pH 10.0, 0.0".
M;
1.60 40
20
I ' Calculated Intercept =0.30 Observed Intercept = 0.3I
'
=i
E.
where
O
0
3
I
O0I 0.80
0.60
0.40 I ~
0.20
30
60
90
120
Time in Seconds. Fig. 3.-(Fe)o = 9.90 X 10-5 M ; Ro = 2.47 X ''01 AcN = 0.15 M ; pH 10.0; 0.0".
150
M;
spectively. This solution was divided into four equal parts and acetophenone was added to increase its concentration by 0, 2.5, 5.0 and 7.5 X 10-6 M , respectively. Each part was extracted thrice with 12.5-ml. portions of isooctane and the three portions combined. The absorbances of the resulting four isooctane solutions were measured with respect to the pure solvent at 238 mp in quartz cells with a Beckman D U spectrophotometer. These absorbances were corrected for the absorbance of a blank obtained by extracting a reaction mixture containing no acetophenone. The absorbance-acetophenone concentration plot gave a straight line. Hence the acetophenone concentration in the reaction mixture could be calculated. Appropriate blank experiments showed that acetophenone was completely extracted from simulated mixtures by the above procedure. The other method, applicable when acrylonitrile was present, consisted of the extraction of acetophenone with benzene so as to concentrate the acetophenone followed by the polarographic determination of the acetophenone in a non-aqueous solution. Three hundred ml. of reaction mixture was re ared. The initial concentrations of Fe(II), CHP and &$were 1.17 X 10-4, 4.90 X 10-6 and 0.01 M , respectively. This solution was extracted twice with 15-ml. portions of benzene
998
W. L. REYNOLDS AND I. M . KOLTHOFF
by vigorous shaking for 15 minutes, separation of phases with a centrifuge and collection of the organic layer in a small separatory funnel. The two benzene portions were combined, 15 ml. added to 15 ml. of absolute ethanol containing 0.2 M tetraethylammonium chloride and 0.005% gelatin in an electrolysis cell and the current-voltage curve determined. In this medium the diffusion current region of acetophenone began a t about -1.84 v. us. SCE. The diffusion current was determined a t -1.9 v. us. S.C.E. The error in the analysis was a t least & l o % since, in the presence of AcN, a large reduction current began at about -1.9 v. us. S.C.E. and interfered with the determination. The residual current a t -1.9 v. u s . S.C.E. was determined by extracting a simulated reaction mixture without acetophenone present and determining the current-voltage curve after addition of organic phase to ethanol as described above. Appropriate blank experiments in the absence of AcN showed that the acetophenone in simulated,reaction mixtures was about 100 10% recovered by this method. Retermination of Methanol.-One hundred ml. of reaction mixture containing 0.070 M Fe(I1) and 0.018 J4 CHP initially was extracted with three 25-ml. portions of benzene to eliminate acetophenone. Sufficient sodium hydroxide was added to give an alkali concentration of approximately 0.1 M. This solution was divided into four equal parts and methanol was added to give a concentration of 0, 0.5, 1.0, and 1.5 mM, respectively. Each part waa separately distilled to dryness under reduced pressure, the distillate diluted to volume in a 25-ml. volumetric flask, and tested for methanol by a procedure which already has been described.? The absorbance, measured a t 570 mN against a reagent blank, was plotted us. added methanol concentration. The result was a straight line which passed through the origin showing no measurable quantity of methanol present in the reaction mixture. Thus, methanol was not a reaction product in the above experiment. Appropriate blank experiments showed that methanol was not measurably extracted from simulated reaction mixtures by the benzene extraction. Determination of Gas Evolution.-High concentrations (ca. 0.05 M )of reactants were used to secure gas evolution. Helium was used for deaeration purposes since nitrogen interfered in the mass spectrometric analysis for ethane in the gas evolved. The gas was collected over the reaction mixture and its volume determined. Since the main gaseous component was ethane the small amount of water vapor present did not interfere with the analysis.
'
Vol. 60
the type and concentration of monomer and the effect of high concentrations of reactants is shown. TABLE I1 p H 5.4; p = 0.1; 25' Monomer
...
M
...
(CET;oMX
3060 40 ... ... 2.5 AcN" 1.0 2170 1.o 40 0.1 25 0.1 3.5 2.5 0.01 2.5 6X ACA~ 0.1 2.5 0.01 2.5 0.001 2.5 2.5 AA" 0.17 0.10 2.4 0.01 2.4 a Acrylonitrile. b Acrylic acid.
...
'
(Monomer).
...
(F;oY,-&X
5820 100 10 4800 100 100 10 10
R
1.2 1.0 ,0.84 1.2 1.7 1.5 1.9 1.6 1.4 1.9 1.8 1.2 1.6 1.1
10 10 10 10 10 10 10 0.8 c Allyl acetate.
In the absence of monomer the reaction ratio increased somewhat as the concentrations of reactants were increased. With a fixed concentration of a given monomer the reaction ratio increased as the concentrations of reactants were decreased. With fixed concentrations of reactants and with a given monomer the reaction ratio approached the value of two as the monomer concentration was increased. Allyl acetate was less efficient in increasing the reaction ratio than either acrylonitrile or acrylic acid; the latte'r two are quite similar in behavior in this respect. I n Table I11 are some results obtained by varying pH, temperature, the type of monomer, and the initial concentrations of reactants in the ferResults Stoichiometry.-The results of determinations rous pyrophosphate-CHP reaction. At pH 8.8 of stoichiometry of the ferrous Versenate-CHP in the absence of monomer and in the presence of reaction a t different p H and temperature values AcN the reaction ratio, R, was about 2 in the lowin the presence of AcN are given in Table I. It is est concentration range of reactants and decreased seen that the reaction ratio, R, defined as the to about 1.6 in the presence of AcN and to about unity in the absence of monomer as the concentrations of the reactants were increased. At pH TABLE I THEREACTION RATIOOF THE FERROUS VERSENATE-CHP 6.6 the reaction ratio was also equal to 2 in the absence of monomer and in the presence of AcN REACTION AS A FUNCTION OF pH AND TEMPERATURE a t the lowest Concentration level of reactants and 0.1 M AcN present in all cases; ,U = 0.1. decreased as the concentrations of reactants were (CHP)o X ( F e y % X R = A Fey-)/ increased above this level. At pH 4.8 the reaction pH P,C. 106,M 106, M *(&HP) n5 ratio was somewhat less in the absence of monomer 3.72 2.00i0.13 8 25.0 1.3-2.5 2.5-10 than in the presence of AcN but the difference is not 5.36 2.3-10 1 . 9 8 i 0.15 11 25.0 1.3-5.0 significant. At all three pH values methacrylate 5.36 2.4 4.5-10 1.92iO.12 6 0.0 (MMAc) as monomer gave a reaction ratio only 8.40 25.0 2.4 4.5-10 1 . 8 7 ~ O o . 0 9 10 slightly greater than unity. An important differ8.80 2.4 25.0 4.5-10 1.98 2 ence between the ferrous Versenate and ferrous 4.5-10 1.95i0.07 6 9.18 2.4 25.0 pyrophosphate reactions with CHP is that, a t the 4.5-10 1.86 4 10.0 2.4 25.0 low concentrations of reactants where the rate 10.0 4.5-10 l.9O=kO.O5 6 0.0 2.4 constants were determined, the reaction ratio of the 4.5-10 1.85 2 10.3 2.4 25.0 former reaction is less than unity whereas the ren is the number of measurements. action ratio of the latter reaction is equal to two in number of moles of ferrous Versenate oxidized per the absence of monomer. It is seen from Table I11 that the change of mole of CHP reduced, is approximately equal t o two in all cases. In Table I1 the effect of varying temperature does not affect the reaction ratio. results of Formation of Acetophenone.-The (7) W. L. Reynolds and I. M. Koltho5, THISJQURNAL, 60, 969 analyses for acetophenone are presented in Table (1956).
b
r
July, 1956
REACTION OF FERROUS VERSENATEWITH CUMENE HYDROPEROXIDE
999
TABLE I11 millimole of methyl free radical. The amount of C H P consumed was 3.15 millimbles and the amount THEREACTION RATIO,R, OF THE FERROUS PYROPHOSPHATEof acetophenone formed was 1.41 millimoles (45% CHP REACTION AS A FUNCTION OF pH, TEMPERATURE, conversion of CHP to acetophenone). Methanol TYPE OF MONOMERAND INITIAL CONCENTRATION OF was not found. Thus the amount of methyl free REACTANTS; p = 0.1 t, (Monomer), (CHP)o (Fe(I1))o radical produced, within the experimental error, OC. Mb X 106, M X 106, M R na PH is equal to the amount of acetophenone formed. 8.8 0 0.5AcN 1.3-2.6 4.4-20 2.08f0.15 9 In the ferrous pyrophosphate-CHP reaction 25 0 . 5 AcN 1.2-2.6 2.5-20 2.08 f 0:lB 13 when the initial concentrations of ferrous pyro1.OAcN 40 100 1.65 1 1750 7000 1.56 1 1.OAcN phosphate and CHP were 0.0700 and 0.0291 M , 1 . 0 MMAc 40 100 1.00 1 respectively, in the absence of monomer 1.23 2.5 10 2.2 1 millimoles of gas, consisting of 10 5% methane 40 100 1.2 1 and 90 f 5% ethane, was formed. Thus the gas 1 2910 7000 1.1 phase accounted for 2.34 millimoles of methyl free 6.8 0 2.5-5.0 7.0-10 1.97 4 25 0.6-4.0 2.2-20 1.98 zt 0.09 6 radical. The amount of CHP used was 3.31 milli10 6 20-40 1.76 i .09 moles. CHP was 96% converted to acetophenone, 0.2AcN 0.6-4.0 2.2-20 2.02 zt .14 10 so that 3.18 millimoles of acetophenone was formed. 0 . 2 AcN 10 20-40 1.79 2 Methanol was not found in the reaction mixture. 0 . 4 MMAc 1.2-3.4 4.0-15 1.22 3 0.2 MMBc 10 20-40 1.17 2 If the amount of CH3 radical again mere equal to 4.8 2 0 0.5AcN 24 50-75 1.93 the amount of acetophenone the difference between 24 50-75 1.82 2 the observed and calculated quantities of methyl 0.2 MRlAc 24 40 1.15 1 free radical was 26%. This difference may be due 4.8 25 0 . 5 AcN 4.0-7.0 13-18 1.97 3 to both experimental error and the occurrence of 2 6.0 14-17 1.75 side reactions. With the initial concentrations of 0 . 2 M M A c 4.0-11.0 9-16 1.18 3 a n is the number of measurements. AcN and MMAc ferrous pyrophosphate, C H P and AcN of 0.0700, designate acrylonitrile and methyl methacrylate, respec- 0.0175 and 1 M , respectively, a small amount of tively. gas (0.33 millimole) was formed. Thus the amount IV. I n the presence of 1 M AcN or MPr'lAc a large of gas evolved was considerably decreased by the reduction current began a t such potentials as to presence of AcN in both reactions. Rate Constants.-The results nf determinations seriously interfere with the acetophenone determination and hence the corresponding analyses for of the second-order rate constant oi the ferrous Veracetophenone were rather inatcurate. In the senate-CHP reaction are given in Table V for 0 ferrous pyrophosphate-CHP reaction, CHP was and 25" and for various pH values. The standard quantitatively converted to acetophenone within deviation of a single measurement was approxithe accuracy of the measurements a t all concentra- mately 13%. The observed second-order rate tion levels. In the ferrous Versenate-CHP reaction constant did not change significantly over the pH C H P was 80-100~0converted to acetophenone a t range investigated. Substitution of acrylic acid the lowest concentration level and less than 50% for acrylonitrile did not affect the observed rate constant. Changing the buffer constituents a t a t the high concentration levels. constant, pH and ionic strength did not affect the TABLE IV observed rate constant either. As a function of (AcN), (Fe(I1)o (CHPh Reaction M X 104, X 104, % temperature the rate constant of this reaction may M M APo R ka = 5.0 X 1Olo liter be written Ferrous 10 4.0 80 1.0 mole-' set.-'. Versenate-CHP 582 306 45 1.2
*
0.1
Ferrous Pyrophosphate-CHP 0.01 1.0 1.0 1.0"
10 1.0 10 700 1.2 10 700 10
4.0 0.25 4.0 291 0 49 4.0 175 4 0
108 108 98 96 100 120b 10lb
1.1 2.2 1.2 1.1 2.0 1.7
High
1.0
TABLE V SUMMARY OF FERROUS VERSENATE-CHPRATEMEASUREMENTS
p =
0.1; A(Fe(II))/A(CHP) = 2; (AcN) = 0.1 M
1.6
A P .designates acetophenone. AcN interfered seriously in the acetophenone determination. MMAc. Presence of monomer interfered so that an approximate determination of acetophenone could not be made. However, the acetophenone wave was evident on the polarogram. a
Formation of Gas.-In the ferrous VersenateC H P reaction when the initial concentrations of ferrous Versenate, C H P and AcN were 0.0480, 0.0217 and 1 M , respectively, insufficient gas was evolved to saturate the solution and form a gas phase. With initial concentrations of ferrous Versenate and CHP of 0.0582 and 0.0306 M , respectively, in 103 ml. of solution in the absence of monomer 0.70 millimole of gas, consisting of 10 f 5% methane and 90 f 5Y0 ethane, was formed. Thus the gas phase accounted for 1.33 f. 0.04
t, o c .
25.0 f 0 . 1
0.0 f 0 . 2 a
PH
(Cl€I?hX 1.30-2.53 1.25-2.47 2.44 8.40 2.44 8.80 2.44 9.18 2.44 10.0 2.44 10.3 2.25 5.36 2.45 10.0 2.45
3.72 5.36
(Fe(1I))o X 105, M 2.30-10.1 2.30-10.0 4.98-9.95 4.95-9,90 4.86-9.90 4.95-9.90 4.95-9.90 4.95-9.90 4.95-9.90 4.95-9.90
ko, 1. mole-' sec. -1 ,1370 216 1200 f 157 1270 1140 f 157 1000 1050 146 1100+ 182 1170 240 =t 33 240 f 38
*
+
n
6 10 2Q 9 2 6
4 2 6 6
0.1 M acrylic acid was employed instead of acrylonitrile.
Values of the second-order rate constant of the ferrous pyrophosphate-CHP reaction are given in Table VI for three pH values a t 0 and 25". The rate constants were determined at the low concentration levels where the reaction ratio was equal to two in the absence of monomer and in the
1000
W. L. REYNOLDS AND I. M. KOLTHOFF
Vol. 60
TABLE VI SUMMARY OF FERROUS PYROPHOSPHATE-CHP RATEMEASUREMENTS ~
t. o c .
25 f 0 . 1
0.0f0.2
PH
(CHP)o X 108,M
(Fe(1I))o X 106,M
4.82
3.89-11.1
9.15-17.7 0.5AcN 0 . 2 MMAc
6.78
Av. la0 = 209 i 22 1. mole-' sec.-l 0.80-1.98 1.45-4.08 0.47-1.95 1.45-4.10 0 . 2 AcN 1.20 4.00 0 . 2 MMAc
837 825 780
6 11
8.80 4.82
Av. ko = 830 f 114 1. mole-' sec.-l 1.21-2.45 2.46-14.1 0 . 5 AcN 24.4 54.9-69.6 24.0 57.9 0 . 5 AcN 44.4 77.0 0 . 2 MMAc Av. ko = 59 f 7 1. mole-' set.-' 2.66-5 .16 7.35-10.9 0 . 2 AcN 2.66 7.35 Av. ko = 210 f 26 1. mole-' sec.-l 1.30-2.60 4.06-18.7 0 . 5 AcN
presence of AcN. I n the presence of MMac the reaction ratio had the values indicated in Table 111. To obtain good agreement between calculated and observed values of log ((Fe)o/nRo) (see Experimental section) the observed value (not the value of unity) of n had to be used in calculating the log term. However, the value of the rate constant does not depend upon the value of n employed so long as n is constant throughout the reaction. It is very important to note that, within the experimental error, the value of the rate constant is independent of the presence or absence of monomer and of whether the monomer yields a reaction ratio of two or unity. At pH 6.78 the rate constant mas independent of a variation of ionic strength over the range 0.091 t o 1.35 and of pyrophosphate concentration as long as the latter was in considerable excess over the iron. However, the experimental error was sufficiently large t o prevent the detection of a small effect. The observed rate constant was dependent on pH and increased with increasing pH. I n terms of temperature the rate constants a t p H 4.82, 6.78 and 8.80 were 2.0 X lo8 exp( -8200/RT), 2.7 X lo9 exp(-8900/RT), and 1.6 X log exp(-8400/RT) liter mole-'sec.-', respectively.
Discussion The complex ions F e y + and F e y - are stable in the pH range 3.5 t o 6.5.8 At pH values greater than 6.5 the equilibriag Fey-2 + OH- I _ Fe(0H)Y-8; K = 7.4 X lo4 F ~ ( O H ) S Y - ~K ; = 1.3 X lo'
have t o be considered for the ferrous species. Since the rate of reaction between ferrous Versenate and C H P was found t o be independent of pH the rate of reaction was not dependent upon the exact composition of the ferrous Versenate species. The formulas of the ferrous pyrophosphate species a t various p H values are not ( 8 ) I. (1952).
n
2 3 3
8.80
+ OH-
\
ko
227 21 1 197
6.78
Fe(0H)F-8
(Monomer), M
M. Kolthoff and C. Auarbaoh, J . A m . Chdm. SOC.,74, 1452
(9) G. Scliwarsenbaoh, Helu. Chim. Acta, 84, 676 (1951).
1
1200 i 200 55.2 66.3
8 2 2 1
201 240
3 1
60.0
3 0 0 2 ~ 38
10
known. The reaction of these species with C H P was found to be dependent on pH. All ferrous Versenate and ferrous pyrophosphate species were electroactive a t the electrode used. I n view of the facts that, in the ferrous VersenateCHP reaction a t low concentration levels, the reaction ratio was equal to two in the presence of AcN, unity (or less) in the absence of monomer, and approached the value of two as monomer concentration was increased, that C H P was converted to acetophenone in high yield in the presence and absence of monomer, that all three ferrous Versenate species reacted with C H P a t the same rate, and that the rate constant was not affected by a change of monomer the following mechanism is proposed. The reaction RO. Fe'lT-2 + RO-
+
+
+ ROOH -+kl Fe(OH),,Y-'-"I + RO. + OH-; m = 0, 1, 2 (1) kz RO. +CsHr,(CO)CH3 + CHa. (2) ki CHs. + M -+ Mi. (initiation) (3) k, Mi. + M -+ Mz. (propagation) (4) k, M.n-1 + M +Mn. kt.m Fe(OH),Y-Z-" + M,.--+ Fe(OH),Y-l-* + Mm-, m = Fe(OH),Y-Z-m
.
M,-+H++M,H
0, 1, 2 ( 5 ) (6)
FeY - was omitted since in the absence of monomer the reaction ratio was less than unity a t the low concentrations of interest in measuring the rate constant. This reaction may be important in the absence of monomer a t the higher concentration levels of reactants employed since the reaction ratio increased somewhat and the acetophenone concentration decreased with increasing reactant concentrations. Termination reaction (5) has been written as reduction of the polymer free radical by a ferrous
I
1
1001
ADSORPTION AND HEATOF IMMERSION OF IRON OXIDE
July, 1956
Versenate species because of the observations made concerning the reaction ratio in presence and absence of monomer and because the reaction ratio decreased when monomer concentration decreased. Although ethane formation occurred a t high reactant concentrations it cannot contribute significantly at low reactant concentrations and high monomer concentration because the termination reaction 2CH3. + CzHawould lead to a reaction ratio less than two. As monomer concentration is decreased this termination reaction may become important and so result in a reaction ratio of less than two. Termination by combination of two polymer free radicals must also be excluded under the conditions where the reaction ratio was equal to two. Termination by disproportionation results in a double bond in one of the two polymer molecules formed. It is hardly possible that this double bond is reduced by iron(I1). Termination by chain transfer can occur since one polymer free radical is replaced by another until one eventually reacts with iron(I1). According to the proposed mechanism the methyl free radicals must be 100% efficient in initiating polymerization when the reaction ratio is two. It must be emphasized that the analyses for acetophenone were not sufficiently accurate to determine whether a few per cent. of RO. initiated polymerization also. The reaction ratio of less than unity in the absence of monomer may be the result of some decomposition of CHP by a free radical formed by H-atom extraction from the Versene molecule by the methyl free radical. The reaction CH3. ROOH + CH30H RO. cannot be proposed to account for this since methanol was not found and since this reaction was not found to occur in the reaction of aquo ferrous iron with CHP.? Considering that ferrous Versenate is such a strong reducing agent it does not seem likely that reduction of ferric Versenate by some intermediate is responsible for this low reaction ratio. Making the usual assumption about the steady states of free radicals and the assumption that the propagation rate constants are all equal, the rate of decrease of total ferrous iron is found to be
+
+
--d(Fe(ll)) dt
- 2kl(ROOH)(Fe(II))
or ds = lcl(ROOH)(Fe(II)) ai
since (Fe(II)), the total iron(I1) concentration a t time t, was equal to (Fe)o - 2x (see Experimental). The observed rate constant, ka, given in Table V, is equal to kl, the rate constant of the reaction between C H P and ferrous Versenate. I n the absence of monomer the CH3 free radicals formed in reaction ( 2 ) combine to form ethane. For the reaction between ferrous pyrophosphate and CHP in the presence of AcN a mechanism similar to that for the ferrous Versenate-CHP reaction may be postulated. Since the formulas of the ferrous pyrophosphate species present a t various pH values are not known a rate equation cannot be derived to give the observed rate constant as a function of hydrogen ion concentration. Although polymer formation was not observed in the ferrous Versenate-CHP reaction considerable polymer formation was observed in the ferrous pyrophosphate-CHP reaction in the presence of both AcN and MMAc. Since, in the latter reaction, the reaction ratio was equal t o two in the absence of monomer a t the low concentrations where the rate constants were determined the methyl free radical, or a free radical derived from it, must oxidize a second mole of ferrous pyrophosphate by a reaction such as Fe(I1)
+ CHs. + H +
--3
Fe(II1)
+ CHU
At very high concentrations of reactants ethane is formed but the reaction ratio under these conditions is 1.1 (see Table IV). When the ferrous pyrophosphate-CHP reaction occurred in the presence of MMAc the reactions ratio was about 1.15. I n this system polymer formation was observed. Termination must occur mainly by reactions not involving iron(I1). Since the observed rate constant was the same when monomer was absent as when either AcN or MMAc was present the rate determining steps must be the reactions of the ferraus pyrophosphate species with CHP. In conclusion it is interesting to note that the two ferrous species under our experimental conditions can act as polymerization termination agents as well as part of the initiation system.
THE ADSORPTION AND HEAT O F IMMERSION STUDIES OF IRON OXIDE BY F. H. HEALEY,J. J. CHESSICKAND A. V. FRAIOLI Contribution from the Surface Chemistry Department, Lehigh University, Bethlehem, Penna. Received March 6. f 656
The surface of an iron oxide powder was investigated by water adsorption and calorimetric measurements. Two thirds of the surface was found t o physically adsorb water in the region of relative pressure where monolayer coverage is normally encountered. The remainder of the surface chemisorbed water whioh could be released only by activation a t temperatures up to 450'. The average heat of hydration of the chemisorbed water was calculated to be about -24 kcal./mole HzO. Evidence for the release of internal water and other gases a t elevated temperatures was found. In addition, thermodynamic calculations reveaIed that the water physically adsorbed at 25' on iron oxide had the properties of ice in the first layer.
Introduction The adsorption of water vapor on porous iron oxide samples was measured previously by Foster' (1) A. G.Foster,
J. Chsrn. Soc., 360 (1945).
and Rao.2 Here, a systematic study of the surface characteristics of a non-porous iron oxide has been made using adsorption and heat of immer(2) K. 8. Rao, THIB JOURNAL, 46, 600 (1941).
