[CONTRIBUTION FROM THE COBBCHEMICAL LABORATORY OF THE UNIVERSITY O F VIRQINIA]
POTENTIALS OF A SERIES OF NITROSOBENZENEPHENYLHYDROXYLAMINE SYSTEMS
OXIDATION-REDUCTION
ROBERT E. LUTZ
AND
MARION R. LYTTON
Received December IS, 1036
Nitrosobenzene and phenylhydroxylamine under some conditions form a typical reversible oxidation-reduction system that is stable and gives rise to reproducible potentials which can easily be measured.'
We have determined the oxidation-reduction potentials of a series of derivatives and have obtained results which, in spite of the sensitivity of many of the compounds involved, are in the main sufficiently accurate and extensive to show the general reversibility of the various reductions, and to show the effect on the potentials of conditions and substitution in the benzene nucleus. These preliminary results are being reported a t this time because active work on the problem has been discontinued. I n Table I is listed the series of nitrosobenzene derivatives which have been prepared and used as oxidants, together with the oxidation-reduction potentials of the systems concerned, determined under standardized conditions, namely, 0.1 N hydrochloric acid in 50 per cent. acetone-water mixtures at 25'. This solvent was chosen rather than alcohol-water mixtures because the systems studied were relatively unstable in the latter. The measurements of the potentials were made by the method of titration of the oxidant with the reducing agent, titanous chloride, the end- and midpoints being determined graphically and checked by the amounts of standardized reagent consumed. I n one instance, namely, p-bromonitrosobenzene, the titration results were checked by determination of the potentials directly by the method of mixtures of oxidant and reductant. The curves representing change in potential plotted against increments of reducing agent followed with reasonable accuracy in most cases the theoretical curves calculated from the midpoint in the titration. Few of the systems measured were stable indefinitely and in many cases the titration had to 1
CONANT AND LUTZ,J . A m . Chem. SOC., 46,1059 (1923). 68
69
NITROSOBENZENE-PHENYLHYDROXYLAMINE SYSTEMS
TABLE I OXIDATION-REDUCTION POTENTIALS OF TEE VARIOUSNITROSOBENZENE-PHENYLEYDROXYLAMINE SYSTEMS IN 50 PERCENT. ACETONE-WATER-O.~ N HYDROCHLORIC ACIDAT 25' BUB8mUENT# OF THE NITROBENZENE DERIVATIVII USED A0 OXIDANT
0.5820 0.598' 0.579* 0.567' (0.595)' 0.587 0.579 0.589
3.methyl .......................................... 4.methyl .......................................... < 2,bdimethyl ...................................... 2.4.dimethyl ...................................... 3.5.dimethyl ...................................... 2-ethyl ............................................ .4.ethyl ............................................
Alkyl
0.560
'2-chloro , .......................................... 3-chloro . .......................................... 4.chloro ........................................... 2.bromo ........................................... 3.bromo .......................................... 4-bromo .......................................... 3-iOdO ............................................ 4-iOdO ............................................ .4.fluoro ...........................................
0.598 0.583 0.576 0.592 0.577 0.575 0.582 0.577 (0.574)
.
Halogens
(0.W) 0.573 0.554 (0.596) (0.575) 0 .550d
Alkoxy1
0.672 0.822' (0.593) 0.610 0.622' (0.586) 0.613 0.607
3-carbomethoxyl , ................................. Carboalkoxyl 4.carbomethoxyl .................................. and Nitro 2-carboethoxyl . ................................... 3-carboethoxyl . ................................... 4-carboethoxyl . ................................... 4-carboisopropoxyl . ...............................
....
I
. .
Values are expressed in terms of the hypothetical normal hydrogen electrode in the same solvent. assuming quinhydrone to be 0.699 v . d Deviation of the potentials from theoretical a t and f reduction is *2 mv unless otherwise indicated as follows: (b) deviation is -3 to -4 mv (e) -5 mv (d) +3 to +5 mv Values in brackets are of questionable accuracy and may be in error by f3-6 mv Other values unless indicated are probably accurate to f l - 2 mv . 0
+
.
.
.
. .
70
ROBERT E. LUTZ AND MARION R. LYTTON
be carried out rapidly. The determinations which are of questionable accuracy because of instability of the systems are indicated in brackets; the others are accurate to fl-2 mv. The values are expressed in terms of a hypothetical normal hydrogen electrode, the measurements actually being made against a calomel or quinhydrone electrode, the latter being the real standard of reference. The values listed are expressed as the difference between the potentials of the cell consisting of the oxidationreduction half-cell and the quinhydrone electrode in the same solvent, subtracted from the value 0.699 v. assumed for the quinhydrone electrode in this solvent (we have arbitrarily chosen to use the actual potential of TABLE I1 THEDIFFERENCES BETWEEN THE OXIDATION-REDUCTION POTENTIALS OF THE NITROSOBENZENE-PHENYLHYDROXYLAMINE SYSTEM AND ITS DERIVATIVES
1
SUBSTITUENT QROUP
1
POTENTIAL DIFFERENCE IN YV. REBULTINQ WHEN GROUP IS
ortho
meta
-3
(F..........................
C1. ......................... Br .......................
..........................
Alkoxy1
.I
...I
+16 +10
l -
-28 -32
-9
$14
-7
COOCHJ....................
+40
N O S . .......................
+40
--
-8 -6
-7
$18
.................. ................
-15 -22
+1 -5 0
....................... {OCHJ OClH ...................... I
para
-5
+28 +31 +25
+90
quinhydrone in water, rather than the value in alcohol [cf. ref. (l)]). The potentials are significant therefore only in relation to each other and to quinhydrone under the standard conditions we have chosen, and of course are only approximately comparable with potentials of these and other compounds determined in water or other solvents. The effect of substitution of groups on the potential of the arylnitrosoarylhydroxylamine system is outlined in Table 11. Alkyl groups when ortho to the nitroso group raise the potential, when para lower it, and when meta have little effect. The halogens, which are nearly equivalent to each other in effectiveness, in the ortho position also raise the potential,
NITROSOBENZENE-PHENYLHYDROXYLAMINE SYSTEMS
71
when para lower it, although only slightly, and in the meta positionare apparently without appreciable effect. The strongly ortho-para orienting groups, methoxyl and ethoxyl, show close agreement with each other, raising the potential when ortho, lowering it very slightly when meta, and lowering it very considerably when para. It is apparent from these facts that the ortho-para orienting groups when located meta to the nitroso group have very little or no effect on the potential, but when ortho consistently raise it, and when para consistently lower it. The substitution of the carboxylic ester group in either the ortho or the para position raises the potentials considerably, by 40 mv. when ortho, and 25-31 mv. when para. Ortho nitro-nitrosobenzene, the only nitro compound studied, gave the highest potential, 90 mv. above that of nitrosobenzene itself. The meta carboxylic esters gave somewhat uncertain values which were, however, qualitatively similar to the potential of nitrosobenzene itself, showing the meta substitution to be without significant effect. From the foregoing results it is evident that both ortho-para and meta orienting groups, located ortho to the nitroso group, raise the potential, the carboxylic ester and nitro groups having the greater effects. I n the para position, however, the effects of the two types of groups are opposite, the ortho-para orienting groups lowering and the meta orienting groups raising the potential. Meta substitution has consistently little or no effect on the potential regardless of the nature of the group. A comparison between nitroso compounds and quinones with respect to the effects of substitution on oxidation-reduction potential is of interest. However, meta and para substitution in nitroso compounds finds no counterpart in substitution in the para quinones. There is some structural analogy between the ortho monosubstituted nitroso compounds and the quinones with substituents fl to one of the quinone carbonyl groups; but the adjacency of the substituent to the other carbonyl is a factor which renders comparison of uncertain significance. Chlorine and bromine raise the potential of the quinone-hydroquinone system about 13 mv. This is in degree and direction close to the effect of ortho halogen on the nitrosobenzene potential. Alkyl and alkoxy1 in the quinone series, however, lower the potential sharply in contrast to the opposite effect in the ortho nitrosobenzenes. On the other hand such groups as the carboxylic ester in quinone raise the potential in much the same way as in the ortho and paxa nitrosobenzenes. Of the many nitrosobenzene drivatives studied, only two, nitrosomesitylene, and o-iodonitrosobenzene, failed to give reproducible oxidationreduction potentials, and were not reduced by titanous chloride under the usual conditions. Common to both of these compounds is heavy
72
ROBERT E. LUTZ AND MARION R. LY’M’ON
substitution ortho to the nitroso group and steric influence is undoubtedly in some way responsible for the phenomenon. These results are consistent with the fact that 2,6dimethyl and 2,4,6-trimethylnitrosobenzenes do not undergo normal reduction easily as do the other homologs with one or no ortho substituent.2 A series of determinations of oxidation-reduction potentials of a number of typical nitroso compounds was made a t 35” and at 45” in order to obtain temperature coefficients, but in view of the magnitude of the experimental error, and the uncertainty of many of the measurements a t the higher temperatures only a qualitative statement of the results is warranted. It may be said that the temperature coefficients lie between -0.5 and -2.0 mv. per degree, nitrosobenzene and the methyl derivatives having high values (-1.1 to -2.0 mv. per degree) and the carboxylic ester, the o-nitro, halogens and the methoxy compounds consistently lower values (-0.8 mv. per degree or less). These values may be compared with the temperature coefficient of -1.1 mv. per degree for quinoneaa The effect of hydrogen-ion concentration on the potential of a typical nitroso compound, p-bromonitrosobenzene, was studied, but because of the many uncertainties the determinations are only qualitative. Over a range of pH up to approximately 5 the change in potential of the pbromonitrosobenzene-p-bromophenylhydroxylamine system very closely parallels the change in the quinhydrone electrode potential with approximately the anticipated “0.6” slope except in the extreme acid range where there appears to be some deviation corresponding perhaps more closely to the “0.9” slope characteristic of systems where the reductant may form salts. The oxidation-reduction potentials could not be followed beyond a pH of about 5 due to increasing instability of the systems under these conditions. EXPERIMENTAL
Of the nitroso compounds listed in Table I, all but eight are known. Many of them are unstable, particularly when impure. The usual method of preparation, namely, oxidation of the corresponding phenylhydroxylamine with chromic acid, with subsequent steam distillation, was found to be inferior in many cases t o the method utilizing ferric chloride where the product can usually be isolated directly in the crude condition and washed free from acids before steam distilling. The nitrosobenzene homologs were best purified by successive steam distillations, since crystallization from solvents gave relatively unstable products in many cases. Compounds in this series were the least stable, and were stored in an ice box and used as soon as possible after preparation. The carboalkoxy and halogeno derivatives were relatively stable and were handled adequately by ordinary crystallization methods. The alkoxy1 derivatives were the most difficult to prepare and were made
* BAMBERGER AND BRADY, Ber., 33,274 (1900). a
CONANT AND FIESER, J . Am. Chem. SOC.,44,2480 (1922).
NITROSOBENZENE-PHENYLHYDROXYLAMINE SYSTEMS
73
by oxidation of the corresponding alkoxyaminobenzenes by means of Caro's acid. The ortho compounds were reasonably stable, but the para less so, while the meta compounds decomposed on drying and had to be used immediately after preparation. o-NitrosotoZuene4was prepared as follows: 15.6g. of o-nitrotoluene, 24 cc. of water, 42 cc. of ethanol, and 1.5 g. of calcium chloride, heated to boiling, with mechanical stirring, were treated with 20 g. of zinc dust (added slowly). After 10 min. the residue was filtered off and the solution poured into 600 cc. of ice water containing 37 g. of ferric chloride. The o-nitrosotoluene separated as a solid and was filtered off, washed on the filter with water, and then steam-distilled twice (yield 20%, m.p. 72.5'). m- and p-Nitrosotoluenes4 and $,a- and IJ-dimethylnitrosobenzenes were prepared by the above procedure. Three new homologs, namely, 3,b-dimethylnitrosobenzene (m.p. 59'), o-ethylnitrosobenzene (m.p. 61°), and m-ethylnitrosobenzene (m.p. 22O), were also obtained by the above procedure from the corresponding nitro compounds in yields of 13-16% of purified material. They were difficult to get into a condition of analytical purity and were not analyzed. They each reacted nearly quantitatively with two equivalents of titanous chloride. In view of the mode of preparation from the corresponding known nitro compounds, their reduction by two hydrogen equivalents of titanous chloride, and their sharp oxidation-reduction potentials which were of the correct order, their structures are certain. Nitrosomesitylene was prepared by the above method and also by the oxidation of mesidine with Caro's acid.', 4b The 0-,m-, and p-nitrosomethylbenzoates and 0-,m-, and p-nitrosoethylbenzoates were prepared according to the directions of Alway and Walker.6 p-Nitrosoisopropylbenzoate(new) was obtained also by this method in a yield of 40%. It melted a t 61-62" and reacted practically quantitatively with two equivalents of titanous chloride. p-Fluoronitrosobenzene' was prepared by the Alway and Walkereo. method using however ferric chloride as the oxidizing agent. o-Chloronitrosobenzene."ine grams of o-nitrochlorobenzene was reduced by adding 10 g. of zinc dust to a boiling solution in 50 cc. of ethanol, 12 cc. of water, and 0.5 g. of calcium chloride, heating being continued for 10 minutes with stirring. The filtered solution was added to a solution of 18.5 g. of ferric chloride in 600 cc. of water and ice, the crude nitrosochlorobenzene being filtered off, washed, and steamdistilled (yield 40%, m.p. 56'). The in- and p-chZoronitrosobenzenesh were prepared similarly in yields of 29 and 41%, respectively, with melting points of 72"and 89.5', respectively. o-Bromonitrosobenzene~was prepared by reducing 11.5 g. of o-nitrobromobenzene in 100 cc. of alcohol, 15 cc. of water, and 0.5 g. of calcium chloride with 10 g. of zinc dust a t boiling temperature, and adding the filtered solution to a solution of 18.5 g. 4 (a) BAMBERGER, Ber., 28, 245 (1895); ( b ) VON PECHMANN AND NOLD,ibid., 31, 561 (1898); (c) BAMBERGER AND RISING,Ann., 316,277 (1901). 6 BAMBERGER AND RISING,Ber., 33,3632 (1900). * (a) ALWAY AND WALKER,Ber., 36, 2312 (1903); (b) ALWAY AND PINCKNEY, Am. Che,m. J . , 32,399 (1904). 7 RINKES, Chem. Zentr., 1914, I, 2036. 8 cf. EAWORTH AND LAPWORTH, J . Chem. SOC.,119,768 (1921). BAMBERGER, Ber., 28,1222 (1895).
74
ROBERT E. LUTZ A N D MARION R . LYTTON
of ferric chloride in 600 cc. of water and ice. The crude nitrosobenzene was filtered off, washed, and steam distilled (yield 35%, m.p. 97'). m- and p-Bromonitrosobenzenea, were prepared similarly. p-Hydroxylaminobromobenzene was prepared in the usual way in a yield of 47% (m.p. 92°).0 o-lodonitrosobenzene (new) was prepared by reducing 14.5 g. of iodonitrobenzene in 200 cc. of alcohol, 15 cc. of water and 0.5 g. of calcium chloride a t boiling temperature with 10 g. of zinc dust, the filtered mixture being diluted with 100 cc. of alcohol, boiled for a half hour and filtered. The solution was added slowly t o a solution of 18.5 g. of ferric chloride in 600 cc. of water and ice, the crude product being filtered, washed, and steam distilled (yield 30%; m.p. 117"). Anal. Calc'dfor CaHJNO: C, 30.9; H, 1.7; I, 54.4. Found: C, 31.4; H, 1.8; I, 54.6. m-lodonitrosobenzene (new) was prepared as above (yield 19%; m.p. 77"). Anal. Calc'd for CaHJNO: I, 54.4. Found: I, 54.4. p-lodonitro~obenzene~~ was also prepared as above. o-Methoxynitrosobenzene (o-nitrosoanisole)lo was prepared in the following manner: 13.5 g. of potassium persulfate was added to 15 g. of concentrated sulfuric acid in a small mortar and the mixture was ground to homogeneity and allowed to stand for one hour. The solidified product was dissolved in 300 cc. of water and 400 cc. of ice, made alkaline with sodium carbonate and acidified with acetic acid; 6.15 g. of o-anisidine was added and the mixture was stirred for 25 min. and filtered, the solid then being washed and steam-distilled (yield 7%; m.p. 101.5"). m-11 and p-'OMethoxynitrosobenzenes and 0-,m- and p-ethoxynitrosobenzenes (nitrosophenetoles)l* were prepared in exactly the same way in yields of 13-23010. In the preparation of p-ethoxynitrosobenzene the mixture after oxidation was directly steam distilled since the product was low melting (30"). The m- and p-ethoxynitrosobenzenes are new but were not obtained sufficiently pure for analysis. They reacted with two hydrogen equivalents of titanous chloride and gave oxidationreduction potentials of the expected order of magnitude. The apparatus and measurement of potentials.-The usual procedures' were followed and need be only briefly summarized here. For the titration half-cell a 400cc. wide-mouthed bottle was used, fitted with a rubber stopper carrying a mercurysealed stirrer, an inlet tube for introducing titrating agent and pure nitrogen for sweeping, and an agar-agar-potassium-chloride salt bridge. A thermostat controlled the temperature a t 25 & 0.1". The electrode was of the platinized platinum type. The standard reference half-cell was either quinhydrone in the standard solvent or calomel which had been checked against quinhydrone during and after use. In many instances both quinhydrone and calomel were used in different runs on the same compound with identical results. The actual runs were made as follows: 0.0005 moles of the nitroso compound was dissolved in 100 cc. of acetone, and 100 cc. of standard hydrochloric acid was added. The apparatus was swept with nitrogen and titration was carried out with standardized titanous chloride, the equilibrium potentials being noted. The variation in concentration of acetohe on addition of the aqueous titanous chloride was shown experimentally to have a negligible effect. The necessary small corrections for the change in acid concentration during titration due t o the excess of acid in the standardized titanous chloride solution were made, and
BAEYERAND KNORR, ibid., 36,3034 (1902). BAUDISCH AND FURST, ibid., 48,1665 (1915). 11 RISING,ibid., 37, 43 (1904). 10
11
NITROSOBENZENE-PHENYLHYDROXYLAMINE SYSTEMS
75
the results all referred to 0.1 N hydrochloric acid solution, the actual average concentration involved in these measurements. The validity of these corrections was checked by controlled experiments on p-bromonitrosobenzene by the method of mixtures and titrations under the various conditions actually used in the experiments. The observed potentials were plotted against increments of standardized reducing agent and the midpoints were determined by the graphical method with the amount of reducing agent consumed serving as a check. SUMMARY
The oxidation-reduction potentials of a series of substituted nitrosobenzene-phenylhydroxylamine systems have been determined, and the effects of substitution of different types of groups in the various positions are discussed and compared with the effects of similar groups on the quinone-hydroquinone potential. Eight new nitrosobenzene derivatives are described.