POLAROGRAPHIC CHARACTERISTICS OF a-FURILDIOXIME
2599
Polarographic Characteristics and Conti*olled-Potentia1 Electroreductiom of a-Furildioximela
by Robert I. GelbIband Louis MeiteslC Department of Chemistry, Polytechnic Institute of Brooklyn, Brooklyn, New York
(Received April 19, 1964)
A detailed description of the mechanism of the reduction of a-furildioxime at mercury electrodes at pH values ranging from very alkaline (no wave is obtained if the p H exceeds 11.5 in the presence of alkali metal ions) to very acidic (10 F perchloric acid) has been obtained with the aid of polarographic and chronocoulometric data.
In a recent study of the polarographic behavior of dimethylglyoxinie,2 only a single wave was obtained for the eight-electron reduction of this compound to 2,3-diarninobutane. 'By controlled-potential coulometry, however, the reduction was found to be accompanied by a large induced quantity of electricity. This was attributed to the reduction of some protonated species by either of two possible paths: one yielding the diamine, and the other yielding hydrogen gas. The present study of a-furildioxime was undertaken in the hope of obtaining further information about the intermediates formed in the reduction of vic-dioximes.
Experimental The polarograph and the apparatus for controlledpotential electrolysis and coulometry were the same as those used previously Supporting electrolytes used in coulometric experiments were always subjected to prior electrolysis at the potential of interest. This both removed reducible impurities and permitted accurate corrections for the continuous faradaic background current, Eastman a-furildioxinie was recrystallized by dissolving it in water, boiling with Norit A, filtering, and cooling. Four recrystallizations gave a product that, after drying at 80" zn uocuo, melted at 167-168" (uncor.; lit.4 166-168"). Solutions of the dioxime in water were prepared determinately, with gentle warming to ensure complete solution. Other chemicals were ordinary analytical reagents. All measurements were made a t 25.0". All potentials are referred to the saturated calomel electrode.
Results and Discussion Solubility and Dissociation Constants of a-Furildioxime. The solubility of a-furildioxime in water at 25" was found to be 0.010 F . Even 0,011 F solutions, prepared a t slightly higher temperatures, always deposited crystalline solute on standing for a few days at 25".
The formal acidic dissociation constants of a-furildioxime niay be defined by eq. 1, where F U - ~ denotes
the dianion. To evaluate pK1 and pKn, solutions of the compound containing 1.O F potassium chloride were titrated potentiometrically with standard sodium hydroxide, using glass and caloinel electrodes. The data used in evaluating pK1 were taken from the region of the titration curve in which the conceiit,ration of the dianion could safely be ignored. Those used in evaluating ~ Kwere z similarly taken from the region in which the concentration of the undissociated compound wa8 negligible. Appropriate stoichionietric equations were used to calculate the concentrations of HzFu, HFu-, (1) (a) This paper is based on a thesis submitted by Robert I. Gelb to the Faculty of the Polytechnic Institute of Brooklyn in partial fulfillment of the requirements for the degree of B.S. in Chemistry, June, 1963; (b) National Science Foundation Undergraduate Research Participant, summer, 1962; (c) to whom correspondence and requests for reprints should be addressed. (2) M. Spritzer and L. Meites, Anal. Chim. Acta, 2 6 , 58 (1962). (3) L. & k i t e s and S. A. Moros, Anal. Chem., 31, 23 (1959). (4) "Beilstein's Handbuch der Organischen Chemie," 1701. X I X , 4th Ed., Springer:Verlag, Berlin, 1934, p. 166.
Volume 68, Number 9 September, 196.4
2600
ROBERTI. GELBAND LOUISMEITES
and Fu+. The necessary corrections for the incompleteness of the reactions were made by assuming that the measured p H was equal to -log [H+]. The results were pK1 = 9-50 f 0.05 and pK2 = 10.6 f 0.1, To obtain an estimate of the lower limit of KO,the dissociation constant of the superacid H3Fu+, a 0.01 F solution of the dioxiine was titrated conductometrically with standard hydrochloric acid. On comparing the resulting titration curve with a similar one obtained by using water instead of the dioxinie solution, no difference of conductance could be detected. It may be concluded that K Ois a t least 0.1. The Course of Reduction of a-Furtldioxime. It will be shown below that a total of eight electrons can be consunied in the reduction of a molecule of a-furildioxinie under most conditions. In agreement with this value, the ultimate reduction product is found to be 1,2-bis(2-furyl)ethylenedianiine. This is analogous to the behavior of dimethylglyoxime. However, two well-defined steps can be discerned in the reduction of a-furildioxime. The first involves six electrons. Under some conditions it can be shown to be the sum of a four-electron process and a subsequent two-electron process. The six-electron step is responsible for a polarographic wave that is the most prominent feature of a polarogram a t any pH value. Its product undergoes a pseudo-first-order chemical transformation, whose rate is markedly affected by pH. The product of this transforination niay finally accept another two electrons, The complexity of this mechanisin has made it necessary to einploy chronocoulonietric theory6 to iiiterpret the results of controlled-potential electrolyses and reconcile them with polarographic observations. The Reduction of a-Furildzoxame in Alkaline Solutions, A single polarographic wave is obtained at pH values above about 7 in 0.10 F buffer solutions contain1.00 F potassium chloride. At pH values bctween 6.5 and 9.8 its half-wave potential is independent of pH. Since it is irreversible, however, its half-wave potential docs vary with drop time. In view of the equations given by Meites and Israel,B it appears most convenient to describe the behavior of the half-wave potential in tcrnis of the parameter Eol/B.This is the value of the half-wave potential normalized to a drop time of 1 sec., and is explicitly given by
Ed.e.
-
0.0542 ____ ana
[log i / ( i d
- i)
- 0.546 log t ] (2)
Under the above conditions the value of Eo,,, is - 1.43 The Journal of Physical Chemistry
f 0.01 v. That of an, may be deduced from the slope of a plot of E d . e . us. [log i j ( i d - i ) -0.546 log I ] and is approximately 0.3. Poorly formed waves and appreciably different values of an, were sometimes obtained, but could not be correlated with the composition of the supporting electrolytes. Nevertheless, it seems clear that the rate-determining step involves a single electron. Moreover, the species actually reduced must be the monoanion, HFu-, which predominates over the greater part of this range of pH values. The diffusion current constant in these media is 9.4 f 0.4. This is nearly the same as for the six-electron wave obtained in strongly acidic solutions. Hence the reduction of the monoaiiion in this pH range must also consunie six electrons a t the dropping electrode. At a large stirred mercury pool, however, exhaustive electrolysis yields a couloinetric n-value of 8.00 f 0.02. This is independent of the initial concentration of the dioxime, of the potential of the working electrode, and of the pH and composition of the buffer used. The coulometric result was confirmed by isolating the reduction product. Ether extraction followed by evaporation in uucuo yielded a clear colorless liquid whose refractive index and infrared spectrum coincided with those of authentic 1,2-bis(2-furyl)ethylenediamine. The latter was prepared by refluxing 5 g. of the dioxime for 3 hr. with 100 g. of amalgamated zinc in 200 inl. of 2.0 F sodium acetate-2.0 F acetic acid-0.1 F zinc chloride. After neutralizing to pH 12 with sodium hydroxide and filtering off the hydrous zinc oxide, the product was isolated as above. It had nZ51.550 and gave an infrared spectrum consistent with the expected structure. It appears not to have been described previously. A plot of log i us. t for this eight-electron reduction is shown in Fig. 1, which also shows'how this can be dissected into two linear portions. Extrapolating the more slowly decaying one to zero timr and completing the dissection in the customary graphical manner indicates that the slower process consumes 30 f 2% of the total quantity of electricity. It inay appear that this figure should be 25% of the total if the two steps represent the addition of six electrons followed by two more, but this is iiot so. The extrapolation actually yields a fictitious value for the current consunied by the slower process at the start of the electrolysis. There can be none of the product of the first step present then, and so the extrapolation grossly overestimates the current that is due to the second step a t the beginning and during the first portion of the electrolysis.
(6) R. I. Gelb and L. Meites, J . Phys. Chem., 68, 630 (1964). (6) L. ,Meites and Y , Israel, J . Am. Chem. Soc., 83, 4903 (1961).
POLAROGRAPHIC CHARACTERISTICS OF CY-FURILDIOXIME
At the same time, of course, the gmphical procedure leads to a total quantity of electricity that exceeds the value obtained coulometrically. I n these experiments the excess was 6 f 2%,
I
I
I
2601
A
+ nle 3 c
B+%e-%D
(3a)
(3b)
Equation 50 of our earlier paper6 describes the value of QR in terms of the electrolytic rate constants P1 and pz QR =
+
nlVCAoe-B1t nzVC,oe-B"
(4)
If p1 > Pz, the first term on the right-hand side of this equation will vanish a t sufficiently long times, and one will obtain QR =
nzVCBOe-Bat
(5)
This describes the quantity of electricity that remains to be consumed in the reduction of B. At zero time it becomes QRO
=
~ZVCB'
(6)
which is of course an accurate description of the quantity of electricity consumed in reducing all the B present initially. The argument cannot be applied to stepwise processes. For example, if the mechanism is
+ nle -% B B + nze3 c
A
I
I
I
1500
I
1
I
3000
(7a) (7b)
eq. 12 and 13 of our earlier papef provide the following description of QR at long times
t, s e c . Figure 1. Variation of current during the controlled-potential reduction of 0.35 m F a-furildioxime a t - 1.60 v. in 0.05 F ammonia-0.05 F ammonium chloride-1.0 F potassium chloride. The inset shows the results of point-by-point subtraction of the extrapolated line (which corresponds to p = 9.2 X 10-6 sec.-l) from the experimental curve a t short times; it corresponds to ,9 = 1.44 X sec.-l.
I n general it is not possible to predict the magnitude of the error on a priori grounds, for it depends on the mechanism of the reduction and on the relative values of the rate constants involved. The extrapolation is based on establishing an equation for QR, the quantity of electricity that remains to be consumed, from date obtained a t times so long that all of the exponentiale except the most slowly decaying one have vanished. This equation is then used to evaluate the whole quantity of electricity coilsullied in the slowest process by simply solving it for zero time. Consider, for example, the reduction of a solution containing two unrelated reducible substances A and R. The half-reactions are assumed to be
At zero time this gives
whereas the quantity of electricity actually consumed in the reduction of B is of course simply n2VCAo. The error decreases as Pl/P2 becomes larger, but the traditional procedure always overestimates the quantity of electricity consumed in the second step. Similarly, if the mechanism is
A
+ nle A B
(104
B-%C
(lob)
c + n3e 3 D
(10c)
the extrapolation will yield a value of Q R O obeying the equation
Volume 68,Number 9 September, 1984
2602
ROBERT I. GELBAND LOUISMEITES
I””’
as the potential becomes more negative. No such variation was observed. Experiments similar to that of Fig. 1 were performed at varying stirring rates, and it was concluded that the data can be described by the mechanism of ey. 10 with PI = Pa. The analysis of typical data is illustrated by the discussion accompanying eq. 32-36 in our preceding paper.6 I n the case cited there we calculated p1 = P3 = 3.4 X lop3 see.-’ and kz = 3.4 X sec.-’. Assuming n3 = 2, the extrapolated value obtained from eq. 11 is then 2.47. This is in excellent agreement with the experimental value of 2.4. A final significant observation is that the extrapolated “n-value” is independent of pH. In the light of the above discussion this must mean that the pseudofirst-order rate constant kz does not vary with pH. Hence the rate-determining step in the chemical transformation intervening between the two electron-transfer steps does not consume hydrogen or hydroxyl ion. One mechanism that would account for all of the data obtained in the range of pH values from 6.5 to 9.8 is
t,
sec.
Figure 2. Variation of current during the controlled-potential reduction of 0.70 m F a-furildioxime at -0.59 v. in 0.020 F hydrochloric acid-1.0 F potassium chloride ( p H 1.73). Circles show experimental points; the solid line is the calculated curve for the mechanism described by eq. 10 with p1 = 1.99 X sec. -l.
if PS is the smallest of the three rate constants. This is again larger than the quantity of electricity actually consumed in the last step, which is of course n3VC,i0. The error decreases as the smallest of the rate constants approaches zero. Qualitatively, therefore, it may be concluded that the “n,-value” of 2.4 obtained for the final step by extrapolation is not inconsistent with the integral value of 2. It may further be concluded that the extrapolated “n-value” cannot be reproduced by eq. 9. I n this mechanism one would have to assume P1/P2 = 15 to account for the ratio of slopes in Fig. 1. Together with nz = 2, this would require that the extrapolated n-value be only 2.14. I n addition, it is improbable that a-furildioxime and its six-electron reduction product should have values of the mass-transfer constant P differing by a factor as large as 15. To be sure, this would be possible if the electrolysis were performed at a potential near the foot of the wave for the two-electron reduction of the intermediate. Then, however, the ratio PI/,& should be potential-dependent, decreasing The Journal of Physical Chemistry
R-C-C-R II II HON NO-
+ e +.R-C-C-R I II
e
~
R-y-f-R+5e+ HON NO-
7H+=
slow, (124
Y H YH-R qh ,NH N H ~ (fast)
H’ R-CH-CH-R’+
I
I
HONH NH2
(12b)
‘0
‘OH
2e +2 H+ = R- CH- CH-R I I NH2 NHz
(fast) (124
in which R represents the 2-fury1 group. The pHindependent transformation of the six-electron reduction product into the species that undergoes further reduction is represented by eq. 12c. From the chronocoulometric data, its rate constant can be assigned a value of the order of loe4 set.-'. The transformation must arise from an interaction involving one of the 2fury1 groups, for nothing analogous to this behavior is observed with dimethylglyoxinie. There is no plausible interaction except hydrogen bonding, and this must involve the hydroxylaiiiino group in order to hinder further reduction. Simple scission of the hydrogen bond would surely be much more rapid than the process in question here. Two other possibilities may be discerned. One is the loss of water from the hydroxyl-
POLAROGRAPHIC CHARACTERISTICS OF WFURILDIOXIME
amino group followed by its readdition in a different configuration, and is shown in eq. 12c. The other, which seems less probable, is the loss of water to form an imine. Although it is independent of pH up to 9.8, the value of EO1,% becomes more negative on increasing the pH still further. At pH values between 10.2 and 11.5, AEol,,/ApH = -230 mv., while an, = 0.25 f 0.01. If the value of kol,h in any one medium is represented by Icof,h
=
k[H+]”
(13)
where IC is the true higher-order heterogeneous rate constant a t zero potential while p is the number of hydro-, gen ions consumed in the rate-determining step, eq. 2 becomes AEol/, -
--
ApR
0.05915 ana
P
(14)
Substituting the values given above into this equation, one obtains p = 0.97. Consequently, in this region where the dianion predominates, the first step must be its rapid protonation to the monoanion, and then the latter must be reduced in accordance with eq. 12. At any pH value above about 11.5, the value of Eo1,* predicted by extrapolating the data obtained in less alkaline solutions is very negative. For example, it is - 1.89 v. a t pH 12. Consequently, no wave is obtained for the reduction of the dioxime in such alkaline solutions in the presence of sodium or potassium ions. The Reduction of a-Furildioxame in Acidic Solutions. The dioxime yields two polarographic waves in support ing electrolytes having pH values between 0 and 4 and ionic strengths between 1.0 and 1.2. The diffusion current constant of the first wave is essentially independent of pH. It i3 9.3 f 0.3 in 1.0 F hydrochloric acid, and is 8.8 f 0.3 in media having pH values between l and 4. These values are nearly identical with those obtained in alkaline media for the six-electron process that occurs there. The second wave is much smaller. Its diff usilon current constant decreases from 1.6 in 1.0 F hydrochloric acid to 1.0 a t pH 1.7, and then remains constant on increasing the pH further. The half-wave potentials of the two waves are, for example, -0.4’7 and -0.74 v. a t pH 1.7. The values of ana, obtained in the same way as in alkaline media, are 0.70 for the first wave and 1.3 for the second. These imply that the rate-determining steps involve one electron for the first wave and two for the second. Unfortunately, the waves occur in the region of potentiala where specific adsorption of the dioxiine is most to be
2603
feared. However, the difference between their halfwave potentials is not large, and it is a little difficult to believe that a would vary so widely over so narrow‘ a range. In addition, the value n, = 1 for the. first wave is attractive because it is the same as in alkaline media. Employing eq. 14 gives p = 0.95 for the first wave and p = 2.4 for the second. If these waves corresponded to a six-electron reduction followed by a two-electron reduction, the diffusion current constant of the second wave should be onethird as large as that of the first, or about 3.1. It actually never exceeds half of this. The discrepanc-y might possibly be due to adsorption of the eight;electron reduction product onto the drop surface, but this possibility is elimirated by the fact that even 10 m F afurildioxinie had no detectable effect on the electrocapillary curve at pH 2 . The observed behavior can be accounted for by supposing that the product of the six-electron reduction niust undergo a chemical transforination before the subsequent two-electron step becomes possible. Bccause some of the six-electron reduction product diffuses away from the drop surfacc before undergoing the transformation, the height of the second wave is smaller than it would be if the transforination were very fast. The theory of kinetic currents further predicts that the height, of the second wave should vary as some power of h, the corrected height of the column of niercury above the capillary, that is smaller than 1/2. Experimentally, a plot of log i us. log h for the second wave was found to be linear and to have a slope of 0.39 a t pH 1.7 or 2.4. The manncr in which the current varies with time during a controlled-potential electroreduction of CYfurildioxinie in this range of acidities depends on the pH and on the potential a t which the electrolysis is performed. At pH values between 1 and 2, and at potentials very near the beginning of the plateau of the first wave, plots of log z us. t have the form shown! in Fig. 2. The coulometric n-value under these conditions is 6.010 f 0,009. As far as is now known, the shape of this curve is uniquely characteristic of the mechanism described by eq. 10 in the special case where the value of is intermediate between the values of kz and p3. The detailed analysis of the data obtaired in a typical experinient was described in the discussion accompanying eq. 30 and 31 in our earlier paper.5 Here it is necessary only to summarize the results obtained. In 1.0 F potassium chloride containing 0.02 F hydrochloric acid and having a pH value of 1.73, we obtained 11.1 = 4, n2 = 2, p,/p, = 1.04 f 0.02, and k2 = (6.0 k 0.5) X lop3 sec.-l. These values were used to draw the solid curve in Fig. 2 , which agrees very well with V o l u m e 68, N u m b e r 9
Septemher, 1964
2604
ROBERT I. GELBAND LOUISMEITES
the data. The value of kz increases with increasing acidity. Insufficient time was available for a detailed examination of this dependence, but it appeared to be consistent with the consumption of a single hydrogen ion in the rate-determining step of the process described by eq. lob. Several possible products of a four-electron reduction of the furildioxime can be formulated, but the above facts require, for example, that the product undergo a reduction characterized by a value of kof,h comparable to that for the reduction of the original dioxime. Hence the only reasonable intermediate is the amine-oxime resulting from a four-electron attack on a single oxime function. On this basis the mechanism responsible for the first wave may be written
+ e + H+,+
R-C-C-R
/I /I
HON NOH R-6-C-R
II
I
I1
I
HZN
ll
I
+ 3e + 4H+ = R-CH-C-R
I
HZN ----t
ll
NOH (fast) (15b)
R-CH-C-R
I I1 \
R-CH-CH-R
I
+ 2e + H +
I
NH,
I
+ H + -+-
HRN+ HNOH (slow)
(15c),
--f
(15e)
More accurately, since the half-wave potential data appear to show that the rate of this process varies as the 2.4th power of the hydrogen ion concentration, there must be some protonation of the hydroxylamine group prior to this reduction. But in addition to this reaction there must be another, whose rate increases with increasing acidity, by which the amine-hydroxylamine is transformed into a product that cannot be reduced to the diamine within the attainable range of potentials, and which is probably
I
,//
H
I I/ HSN+ NOH \
I
H3N+ HXOH
R-CH-CH-R
HzN+ NOH
NOH
R-CH-C-R
I
+ 2e + 2H+ +
H3Nf
HONH YOH (slow, rate-determining) (15a)
+ H+
R-CH-CH-R
R-&-C-R
HONH NOH R-CH-C-R
may be obtained by stepwise reduction. In 1 F potassium chloride-0.02 F hydrochloric acid, reduction a t a potential near the beginning of the plateau of the first wave gives n= 6.01 f 0.01, as stated above. If the potential is then changed to a value on the plateau of the second wave, an additional 1.78 f. 0.01 f . is consumed for each mole of dioxime present originally. The principal reaction under these conditions can be written
R-CH-CH-R
I
+ H&
1
H3N+ +NH
.1
R-CH-CH-R
///
H
2H+
R-CH-C-R
I
(fast) (15d)
1
H3N+ HNOH where the slowness of the reaction dcscribed by eq. 15 can be considered to reflect the extremely weakly basic character of the oxime group as well as the limited acidity of the protonated amine group. The effect of the reaction is presumably to alter the electron density in the C=N bond, thereby facilitating the addition of two electrons across it. Controlled-potential electrolysis a t potentials on the plateau of the second wave gave n = 8.00 f 0.02 a t pH 3.7. This represents reduction to the diamine. But the n-value decreases with increasing acidity, and is 7.80 f 0.02 a t pH values near 1.5. The same result The Journal of Physical ChemidTy
+
1
HW-NH
1
(15f)
Polarograms of a-furildioxime in concentrated perchloric acid solutions consist of three waves. Their half-wave potentials vary irregularly with acid concentration because of the large liquid-junction potentials involved. In 6.0 F acid they are approximately -0.1, -0.3, and -0.7 v., and their heights are very nearly in the ratio 2 : l : l . In 8.0 F acid the half-wave potentials are about -0.2, -0.5, and -0.8 v., and the wave-height ratio is approximately 4 : l : l . In 10.0 F acid the half-wave potentials of the first two waves are approximately -0.05 and -0.4 v. and the ratio of their heights is roughly 8: 1. In the last of these niedia the third wave is obscured by the onset of hydrogen evolution.
POLAROGRAPHIC CHARACTERISTICS OF a-FURILDIOXIME
2605
On the plateau of the first wave in 6 F perchloric acid, the variatjon of current with time during a controlledpotential electrolysis conforms to the theoretical predictions for the mechanism of eq. 7. A detailed analysis of typical data is given in the discussion following eq. 13 of our earlier papera5 At lower acidities a chemical step can be discerned between the two electron-transfer steps. Its rate increases with increasing acidity, and a t these high acidities it is so fast that it is chronocoulometrically invisible. Thus eq. 15a-d serve to represent the course of the reaction in these strongly acidic media as well as in less acidic ones. The plot of log z vs. t is concave downward when the electrolysis is performed at a potential such as -0.19 v. in 6 F perchloric acid, which is on the plateau of the first wave. It is perfectly linear at -0.40 v. in the same medium, on the plateau of the second wave. Nevertheless, there is no difference whatever between the coulometric'n-values, both of which are equal to 6, This must mean that the second wave corresponds to a direct six-electron reduction of the dioxime. Neglecting protonation of the product, this may be represented by
R-C-C--R
/I /I
+ 6e + 6Hf +
t , sec.
HON NOH
R-C-C-R
I I
HzN HNOH
(16)
Figure 3. Variations of current during the controlled-potential reduction of (a) 2.45 m F a-furildioxime at -0.85 v. in 1.0 F hydrochloric acid, and ( b ) 1.23m F 1,2-bis(2-furyI)-2-hydroxylaminoethylamine, formed by the six-electron reduction of a-furildioxime a t -0.40 v. in the game medium, a t -0.80 v. in 6.0 F perchloric acid.
That the slow step described by eq. 15c escapes chronocoulometric detection a t this acidity means t like those shown in Fig. 3. Curve a was obtained in merely that its pseudo-first-order rate constant is at the direct reduction of the furildioxime and correleast 0.05 sec.-I, which is considerably larger than the mass-transfer constants used. Means are a ~ a i l a b l e , ~ . ~sponds to an n-value of 9.0. Curve b was obtained in reducing the amine-hydroxylamine formed in a prior however, by which mass-transfer constants of the order six-electron reduction of the dioxime, and corresponds of 0.05 sec.-l can be obtained. These should make it to an n-value of 3.9, or a total of 9.9 for the reduction of possible to detect intervening chemical steps and the dioxime. Reduction at these potentials in 1 P evaluate their rate constants even if these are as large as hydrochloric acid gave a brownish black precipitate, about 1 sec.-l. A rate constant of the order of 0.05 freely soluble in concentrated hydrochloric or sulfuric set.-' is far from instantaneous on the dropping-elecacid, but only very slightly soluble in ethanol, diethyl trode time scale, and therefore an appreciable fraction ether, or acetone. Its physical properties and infrared of the dioxime undergoes only a four-electron reduction spectrum were identical with those of the product oba t potentials on the first wave. This reduction is detained by allowing a small amount of 1,2-bis(2-furyl)scribed by eq. 15a-c. The second wave reflects the ethylenediamine to stand in excess concentrated hydroaddition of six rather than four electrons to this fracchloric acid in the absence of air. I n each case the tion. As would be predicted from this description of peak a t 790-820 cm.-' was absent. This peak is comthe mechanism, the sum of the heights of the first and mon to the spectra of furil, furfuryl alcohol, furoin, second waves is independent of perchloric acid concenand 2-furaldoxiine, and may be ascribed to C-0 stretch tration after correction is applied for the effect of viscosity. Electrolyses at potentials on the plateau of the third (7) A. J. Bard, Anal. Ckem., 35, 1125 (1963). wave a t pH values below about 1 give plots of log i vs. (8) S. Karp and L. Meites, unpublished results, volume 68, Number 0 September, 1964
2606
ROBERTI. GELBAND LOUISMEITES
in the furan ring. There mas, however, a shoulder a t 1700 cm-I that suggested the formation of a carbonyl group by ring cleavage. As the electrolysis is prolonged, the n-value increases and the precipitate becomes less soluble in concentrated hydrochloric acid. A reduction was performed in 6 F perchloric acid and allowed to proceed until 9.9 f . had been consumed per mole of the dioxime. Neutralization then gave a precipitate that appeared to be completely insoluble in ethanol, diethyl ether, acetone, benzene, or water, although it was freely soluble in concentrated sulfuric acid or 70% perchloric acid. It is presumably a polyamine resulting from acid cleavage of the furan ring, reduction of the resulting a,P-unsaturated ketone, and polymerization of the free radicals. The Reduction of a-Furildioxime in Neutral Solutions. At pH values above 5 , the heights of the waves observed in more acidic solutions decrease, and the wave, for which Eoi,, = - 1.4 v., that is characteristic of more alkaline solutions appears. The total wave height is essentially constant and independent of pH although the wave-height ratios are pH-dependent. The total diffusion current constant is approximately 10. This indicates that a total of six electrons is consumed a t the dropping electrode, as is true everywhere else in the attainable range of pH values. In the absence of a maximum suppressor, the plateau of the first wave is marred by a minimum, and 0.0015% Triton X-100 was added to eliminate this and permit the precise measurement of the height of this wave. We may imagine that the height of the first wave reflects the extent to which a reaction of the form A+qH+-+B
(17)
occurs during the drop life. Here A denotes the dioxime, which is neither appreciably protonated nor appreciably dissociated in this range of pH values, while B denotes the reducible species. One then expects the wave height to be diffusion-controlled a t low pH values but to become increasingly kinetic in nature as the pH increases. In the limit thc wave height will be proportional to the concentration of A in the bulk of the solution, to m'/ata/8,and to the quantity I c f / k b ' / ' . The pseudo-first-order rate constant ICf for the transformation of A into B is proportional to the qth power of the hydrogen ion concentration, whereas the pseudofirst-order rate constant for the reverse transformation of B into A is independent of the hydrogen ion concentration. Hence the height of the pure kinetic wave ultimately obtained should be proportional to the qth power of the hydrogen ion concentration. Similarly, the slope of a plot of log i us. pH should apThe Journal of Phyeical Chemistry
PH. Figure 4. Effect of pH on the logarithm of the wave height of a-furildioxime; see text for experimental details. The dashed line has a slope of - 1.
proach - q as the wave height i decreases and as the current becomes more and more nearly equal to the kinetic current alone. Such a plot is shown in Fig. 4. The data were obtained by varying the pH from 5.2 to 6.0 while keeping all other experimental variables constant. Its slope clearly approaches -1 as a limit. This simple technique seems not to have been described in the earlier literature. It thus appears that a single proton is involved in the rate-determining step in neutral as well as in acidic media. One might account for this by assuming that the reaction mechanism is HzFu H3Fu+
+ H+
HaFu+ (fast)
(Ha)
(rate-determining)
(18b)
=
+ e +products
It would then be necessary to assume that the rate of the protonation becomes comparable to the rate of electron addition a t a pH value near 5 . This is not inconsistent with a reasonable estimate of the rate constant of the protonation. A nearly but not quite equivalent supposition is that the first electron is added through a hydronium ion which serves as a bridge between the electrode and the nitrogen atom of one of the oxime groups. The bridging ion would doubtless be polarized to such an extent that the reduction would
POLAROGRAPHIC CHARACTERISTICS OF a-FURILDIOXIME
addition of electrons across the C=N bond. The electronic structures of the oxime groups in these compounds are probably not very different. It is attractive to imagine that hydronium ion bridging to the dioxime facilitates the attack by eliminating the need €or the molecule to assume a particular orientation with respect to the electrode surface. Scheme I summarizes the information obtained on the inechanisiii of the reduction of a-furildioxime under all of the conditions investigated.
proceed by a hydrogen atom transfer mechanism rather than by electron transfer. I n this connection it is interesting to note that the reductions of a-furildioxime and of 2-furaldoxime occur in markedly different ways. In 1.0 F potassium chloride containing 0.04 F hydrochloric acid, the half-wave potential of the dioxime is -0.440 v. and its an, is 0.70. For 2-furaldoxime under the same conditions these parameters are --0.892 v. and 1.5, respectively. The latter compound is apparently reduced by 1,2Scheme I
R-7-t-R 4 HS -ON NO-
PKs 10.6
+
R-C-C-R ' H HOA 40+e
3.
0CH-CH--R
c--fapt
0) AH
2607
pg,
- v,s
O 'N
' A
I
n
v
OCH-CH-R "OH
I
I
+
2e+ 2H"
R-CH-C-R
I
II
+ 2ei-t
H2N$ 'POH H
NH*
1""
R- CH- CH-E I 1 H2N NH:,
R-CH-CH-R I t HsN +"OH
+ H+
-C=CH-C-CHpolymer
,+e
d
AH*+
21
+7Ht
HS
d0W
PH4
v r m i n i n g
"[I
-
+ 2HS
+ ringcleavalle 2.9
-
(polarizedat electrodesurface)
R-~-c-R +3e+ 4HS HONH OH
it/ 0 1
I1 II
slow at p H >S
+ 5e+ 7HS
R-5-C-R I II HON. NO-
+ e+ H ~ O +
R-C-C-R
c . +
R-CH-CH-R
I
t
HsN3 N&
R-CH-CH-R I I H~N ++NH
+ H20
I
R-CH-OH-R + 2H" I I HN-NH (electrolvtidv inert)
Volzime 68,Sztmber 9 September, 1964