Hydrogen Bonding to the Nitrobenzene Anion Radical
2649
Equilibrium Studies by Electron Spin Resonance. V. The Role of the Cation in Hydrogen Bonding to the Nitrobenzene Anion Rad ea Luis Echegoyen, Hector Hidalgo, and Gerald R . Stevenson* Department of Chemistry, University of Puerfo Rico, Rio Piedras, Puerfo Rico 00937
(ReceivedJune 73. 7973)
Publication costs assisted by the University of Puerto Rico
The alkali metal cation was found to have a pronounced effect upon the thermodynamics of the hydrogen bond exchange reaction between hydrogen bonded hexamethylphosphoramide (HMPA) and the nitrobenzene anion radical. This effect was found to be due to the participation of the hydrogen-bonded HMPA in the solvation sheath of the cation. When ammonia is used as a proton donor, a similar interaction was found to exist between the cation involved in ion pairing with the nitrobenzene anion radical and the hydrogen-bonded HMPA. This interaction leads to the formation of a hydrogen-bonded ion pair.
The nitrobenzene anion radical when formed by alkali metal reduction in hexamethylphosphoramide (HMPA) exists in two forms: the “free” ion ( a ) and the ion pair ( p ) . Both of these species can be observed simultaneously by esr, and the enthalpies controlling the ion pair dissociation reaction are known (eq l].1,2 p.T-)a+M+
( 1)
It has been previously reported that the addition of small amounts of water or alcohol to the system PhN02HMPA-Li affords increases in the nitrogen coupling constant for the formally “free” ion due to hydrogen bonding of the acidic proton of the alcohol with the nitro group of the PhNOz anion r a d i ~ a l However, .~ no change in the nitrogen splitting for the ion pair ( p ) was observed. These results were explained by eq 2, in which a hydrogen bond is formed with a. This hydrogen bonding increases the charge and spin density on the NO2 group, thus increasing the nitrogen isotropic coupling constant. The ion pair, however, is too tightly complexed with the cation to participate in hydrogen bonding. H-OR
Due to the fact that the nitrogen splitting of the hydrogen-bonded ion (a’) increases smoothly with the addition of hydrogen-bonding donor, it was concluded that equilibrium 2 was fast on the esr time scale, and that the observed nitrogen coupling constant was the averaged coupling constant (AN) for the hydrogen bonded ion (a’) and the “free” ion ( a ) . This conclusion was supported by the fact that dramatic line width alternation was observed a t lower temperatures (-loD), and by the fact that the concentration of the time-averaged species increased relative to that of 3’/ as the proton donor concentration was increased.3 Utilizing time averaging equations, expression 3 was developed for the thermodynamic equilibrium constant for reaction 3.3
K,,
=
(Zh
- 8.48)(HMPA)/(Ah’ - x,)(HMPA’)
The Journal of Physfcal Chemistry. Voi 77. No. 22, 1973
(3)
is the coupling constant for the completely hydrogenbonded ion, and 8.48 is the coupling constant for a. To better understand the role that the counterion plays in hydrogen bonding to anions and to obtain valuable information as to the structure of hydrogen bonded anions, we wish to report the effect of the cation upon the thermodynamic parameters controlling equilibrium 2 and the formation of the first hydrogen-bonded ion pair.
AN’
Experimental Section
The nitrobenzene (Eastman Organic Chemicals) was vacuum distilled before use. The HMPA was distilled under reduced pressure from calcium hydride and stored over molecular sieves 4A. The HMPA was distilled from the solvated electron under high vacuum directly into the reaction vessel. Quantitative additions of methanol were added to the HMPA solutions containing the PhNOz anion radical in the same manner as previously de~ c r i b e d .The ~ ammonia was carefully degassed by the freeze-thaw method and distilled from the solvated electron into an evacuated bulb for storage. Quantitative additions of KH3 to the PhNO2 anion radical solutions in HMPA were carried out with the use of a toepler pump connected with a gas buret. Ammonia was allowed to enter the calibrated gas buret under high vacuum Torr). The number of moles of ammonia was then calculated using the ideal gas equation. Initial pressures of NHs in the gas buret were between 10 and 400 Torr. This ammonia was then allowed to come in contact with the stirred HMPA solution of the PhN02 anion radical for a period of about 2 hr. The stopcock to the apparatus containing the HMPA solution was then closed and the remaining ammonia was toepled back into the gas buret. The difference between the initial and final number of moles of ammonia in the gas buret was taken as the number that had dissolved in the HMPA solution. The correction for the gaseous ammonia in the solution containing apparatus was found to be negligible. The esr system and the method used to form the anion radicals were identical with those previously described.1 Results and Discussion
Addition of small amounts of methanol or ammonia to the system PhNOz-HMPA-Na or PhN02-HMPA-Li affords an increase in the nitrogen coupling constant for the
2650
L. Echegoyen, H . Hidalgo, and G . R. Stevenson
I
I '
Figure 2. Esr spectra for the system PhN02-HMPA-Na. The lower spectrum is for the system containing 0.01 M NaC103. 3
I
1
j ,
18
1
II
Figure 1. Modified van't Hoff plot for the systems P h N 0 2 H M P A - M + For the LI system the numbers are given in parentheses and t h e plot is represented by The line represented by 0 is for the Na system TABLE I: Thermodynamic Parameters Controlling Equilibrium 2 for the Systems
-
Systems
PhNOZ-HMPA-Na-CH30H PhN02-HMPA-Li-CHSOH PhN02-HMPA-Na-N H3 Ph NOn- H M PA-Li-N H 3
Keg
AH",
AS",
at 25"
kcal 'mol
eu
2.0 rt 0.5 3.5 zk 0.2
-1.9 f 0 . 4 b 1.2 rt O . Z b
-5.1 1.6
a
OC
9,0 f 2 C
methanol and those containing ammonia. However, for the systems containing ammonia, the thermodynamics are complicated by other factors, which will be discussed later. There are two possible simple explanations for this metal effect. Either a and/or a' are not completely free of interaction with the counterion, or the counterion is complexing with the HMPA and HMPA' that are involved in equilibrium 2. To check for the former possibility, sodium chlorate was added to the system PhNOz-HMPA-Na. The addition of even 1 mg of NaClO3 resulted in a dramatic increase in the ion pair to "free" ion ratio (Figure a), but even relatively large concentrations of added iXa+ did not alter the nitrogen coupling constant for LY (8.48 G). Further, the addition of dicyclohexyl-18-crown-6-ether (DCC) had no effect uDon the coupling constants of either a , 13, or the time-averaged species. Both of these experiments supported the original report that n is essentially free of interaction with the counterion.1 constants tor @,either A N Or The fact that the A N a , did not vary with the addition of DCC means that either the DCC cannot compete with the strongly cation solvating HMPA or that the ion pair of PhN02 remains intact even though the cation is complexed with the DCC. This latter explanation has been found to be the case for the tetracyanoethylene anion radical in dimethoxyethane,* where the crown ether complexed the potassium cation in the ion pair, but the ion pair remained. For the case of Phn'Oz equilibrium 4 describes the process. -
a K,, for the systems containing ammonia cannot be determined since it is unknown how many of the three possible protons are interacting with the PhN02 anion radical or with the HMPA The modified Van t Hoff plots for these systems are shown in Figure 1 AHo f g the systems with added ammonia are taken from a plot of In / ( A N - 8 48) 1 ( A N ' - A N ) ! vs I I R T
formally free ion due to hydrogen bonding of the acidic protons with the nitro group of the anion radical. However, for the methanol addition no change in the nitrogen splitting for the ion pair ( p ) is observed. These results are essentially the same as previously r e p ~ r t e d ,however, ~ upon quantitative analysis there remains a striking differenre.
Utilizing- the previously described technique and eq 3,3 the thermodynamic parameters controlling the hydrogenbond exchange reaction (eq 2) were determined for the systems PhNOz-HMPA-M+ with added methanol and ammonia as proton donors. The results are shown in Table I. From Figure 1 and Table I it is obvious that equilibrium 2 is very dependent upon the counterion. The same striking effect is observed for the systems containing The Journal of Physical Chemistry, Vol. 77, No. 22, 1973
I
DCC f i\'a+,PhKO,.- === DCC,Sa+,PhN02.-
(4)
All of this indicates that the effect of the counterion upon the hydrogen bonding exchange equilibrium (eq 3) is due to the complexing of the counterion with the HMPA, and the HMPA that are involved in equilibrium 3. To account for this effect, eq 3 should be written to include this metal perturbation as shown in eq 5 . a
+ €JMPA'(Mf)
F=+=
a'
+
HIVIPA(M+)
(5)
2651
Hydrogen Bonding to the Nitrobenzene Anion Radical
0
Figure 3. First two lines of the esr spectra due to the ion pair in the system PhN02-HMPA-Li. Top spectrum shows only t h e first two lines of the ion pair with Li. Middle shows spectrum after NaC103 has been added, and an extra line due to the ion pair with Na is observed. Bottom shows spectrum after methanol has been added (0.3 M ) to the same solution, and the line due to the ion pair with Na has practicaly disappeared. The line marked 0 is the first line of the hydrogen bonded ion, which is shifted over d u e to the added methanol. The arrows indicate the lines d u e to the sodium ion pair.
To confirm the fact that there is a strong interaction between the cation and the hydrogen-bonded HMPA, sodium chlorate was added t o the system PhN02-HMPA-Li containing both a and 0.The esr spectrum of this resulting solution indicated the presence of the “free” ion and two ion pairs. One of the ion pairs is due to the interaction with the sodium cation and one is due to the interaction with the lithium cation, Figure 3. Addition of methanol (0.3 M ) to this solution containing both ion pairs resulted in the complete disappearance of the sodium ion paired species, Figure 3. All of this leads us to conclude that a is free of interaction with the counterion and that the HMPA’ and HMPA t h a t are involved in equilibrium 5 are complexed with the alkali metal cation. Similar dependence upon the cation is observed for the systems PhNOz-HMPA-M- with added ammonia. For these systems, however, the nitrogen coupling constant for the ion pair does not remain constant with addition of the ammonia. This indicates that the ammonia not only interacts with the “free” ion, but also forms a hydrogen bond with the ion pair resulting in the formation of a hydrogen bonded ion pair ($’). In fact the interaction of the ammonia with 3 is stronger than t h a t with 01 as evidenced by the fact that the plot of the
hydrogen-bonded “free” ion over the hydrogen-bonded ion pair concentration us. the NH3 concentration has a negative slope for the systems PhN02-HMPA-Li and - S a . This is in direct contrast with these same systems when methanol is acting as the hydrogen-bonding donor, where the slope is positive, and A N for the ion pair is independent of the proton donor c ~ n c e n t r a t i o nThe . ~ only way a negative slope for this plot and the dependence of A s for the ion pair upon the NH3 concentration can be interpreted is to assume an exothermic interaction between the NH3 and the ion pair to form the hydrogen-bonded ion pair ( p ’ ) . Small additions of NH3 to the PhK02-WMPA-M+ systems results first in small increases in AI, due to the ion pair. At about 0.3 M NH3, A N decreases and reaches a minimum a t about 0.6 M NH3. It is interesting to note that Ah a t this minimum is less than that for the ion pair in pure HMPA. As the concentration of the ammonia further increases A h increases to approach its value in pure ammonia (10.9 G at room temperature). These results are nicely explained by the formation of the hydrogen-bonded ion pair and gradual increasing of the hydrogen bonding and simultaneous receding of the cation. At very low concentrations of NH3 hydrogen bonding of the ion pair compliments the cation in pulling charge and spin to the NO2 group. With further increases in the NH3 concentration, the cation becomes displaced by the ammonia and a decrease in A N is noted. As the cation migrates further from the electronegative center, it is replaced by stronger interactions with the ammonia, and A N again increases. Equation 6 accounts for the formation of the hydrogen bonded ion pair. /7
+ HMPA’
?=+ B’
+
HMPA
I t has been previously reported that a t lower temperatures and/or a t higher concentrations of proton donor, the time-averaged spectra of e and a’ exhibit line width alternation.5 If eq 7 is valid, line width alternation of the ion pair spectra is expected a t higher concentrations of ammonia. This is exactly what is observed experimentally. At higher concentrations of ammonia and/or lower temperatures, pronounced line width alternation of the esr spectra due to the ion pair is observed. This is due to the rapid formation and breaking of the hydrogen bond to the ion pair. Further additions of ammonia lead to further increases in the concentration of p’ and the loss of a detectible concentration of e’.At very high concentrations of XHs (ca. 0.9 M ) , the line width alternation becomes less evident due to the displacement of equilibrium 6 far to the right. At the same time, the metal splitting for the sodium cation falls to zero. At 100% ammonia a single anion remains due t o a strongly hydrogen-bonded PhNOz anion radical, which exhibits weak line width alternation as previously r e p ~ r t e d . It ~ , is ~ interesting to note that this radical in 100% NH3 results from the ion paired species in HMPA as ammonia is added. The lack of formation of the hydrogen-bonded ion pairs from alcohol or water proton donors, despite the fact that 0 has a more highly negative nitro group than does a due to the presence of the cation, is due to the fact that the ion pair (a) is too highly complexed with the cation to allow the interference of the proton donor. The ammonia molecule hydrogen bonds to this NO2 group and a t the same time complexes with the metal cation to further stabilize the hydrogen-bonded ion pair as shown in I. This The Journal of Physical Chem/sfry, Voi. 77, No. 22, 1973
2652
Elizabeth V. Patton and Robert West H
conclusion is further supported by the earlier observation t h a t there is close association between the hydrogen-bonded HMPA and the cation. A similar association is observed here with the ammonia HMPA complex and the cation involved in ion pairing. Acknowledgment. The authors are grateful t o the Research Corporation for the support of this work. References and Notes (1) G . R . Stevenson, L. Echegoyen, and L. R . Lizardi, J , Phys. Chem., 76, 1439 ( 1972). ( 2 ) (a) G. R. Stevenson, L. Echegoyen, and L. R. Lizardi, J, Phys. Chem.. 76, 2058; ( b ) G. R. Stevenson, and L. Echegoyen, ibid.. in press. (3) G. R . Stevenson and H.Hidalgo, J . Phys. Chem., 77, 1027 (1973). (4) M. T. Watts, M . L. Lu, and M . P. Eastman, J. Phys. Chem., 77, 625 ( 1973). (5) F . J. Smentowski and G. R. Stevenson, J. Amer. Chem. Soc.. 90, 4661 (1968)
New Aromatic Anions. IX. Anion Radicals of the Monocyclic Oxocarbons' Elizabeth V. Patton and Robert West* DePdrtment of Chemistry University ot Wisconsin, Madison, Wisconsin 53706
(Received June 27, 7973)
Publication costs assisted by the National Science Foundation
Electrolytic oxidation of bis(tripheny1phosphine)iminium salts of squarate, croconate. and rhodizonate anions in dichloromethane gives electron spin resonance spectra attributed to C 4 0 4 .-, C 5 0 5 .-, and c606' -. respectively. Electrolytic reduction of the same species gave a radical only for rhodizonate, believed to be & & - . A 13c hyperfine splitting constant of 4.10 G was found for C 5 0 5 - - . The implications of this value, and of the g values for all the radicals, for electron distribution in these species is discussed.
The properties of the monocyclic aromatic anions, C,O,-", have been studied in some detail over the past 10 years.2 Four members of the series are known. The dianions with n, = 4 (squarate), n = 5 (croconate), and n = 6 (rhodizonate) are all extremely weak bases which form very stable, high-melting salts. The tetraanion C60s4has also been isolated as the tetrapotassium salt,3 but it is much more reactive, undergoing rapid oxidation in air. The existence of intermediate one-electron oxidation and reduction products of the oxocarbon dianions has been predicted.2 These radicals could be stabilized by electron delocalization in the same way as the parent dianions. Two groups have reported tentative evidence for the existence of oxocarbon anion radicals in the solid state. Air oxidation of the tetrapotassium salt of C6Oa4--, which leads ultimately to C6O&, produces a n intermediate paragmagnetic solid surmised to be CSO6.3-.3 Also, Buchner and Lucken treated potassium metal with carbon monoxide a t 250" and obtained a mixture of products containing a radical giving a single-line esr ~ p e c t r u m This .~ was attributed to the species c606.5 - . The Journal of Physical Chemistry, Vol 77, NO. 22, 1973
We wish to report the generation of oxocarbon anion radicals in solution by electrolytic oxidation or reduction of oxocarbon salts in dichloromethane a t -85". Repeated attempts to obtain esr signals of oxocarbon anion radicals upon electrolytic oxidation or reduction in aqueous solution were uniformly unsuccessful. We then wished to study oxocarbon anions in nonaqueous solvents, in which temperatures low enough to stabilize the radicals might be possible. Most oxocarbon salts are highly insoluble in solvents other than water, or water-alcohol mixtures, and it proved quite difficult to find oxocarbon salts with sufficient solubility in nonhydroxylic solvents. Ultimately, however, salts of the oxocarbon dianions with bis(tripheny1phosphine)iminium cation5 were prepared and found to be soluble in organic solvents. Upon electrolytic oxidation, the squarate, croconate, and rhodizonate salts all formed radicals giving single-line electron spin resonance spectra which we attribute to c404* -, ' 2 5 0 5 - -, and C6O6. -, respectively. Confirmation of this assignment for C505.- is given by the observation of 13C sidebands with a hyperfine splitting of 4.10
