April, 19XS
THE REVERSBLESPLITTING [CONTRIBUTION FROM
THE
OF ORGANOMERCURIC
891
CYANIDES
DEPARTMENT OF CHEMISTRY, IOWA STATECOLLEGE ]
Electron-sharing Ability of Organic Radicals. E. The Reversible Splitting of Organomercuric Cyanides’ BY EDWARD C m , I. B. The present work was suggested by the work of Kharasch and co-workers2 on the irreversible splitting by hydrochloric acid of unsymmetrical organomercurials of the type RHgR’, where R and R’ are bound to mercury by carbon. The work of Kharasch led to a series of radicals arranged in the order of increasing ease of cleavage from mercury, the series ending in the most easily cleaved radical, namely, -CX. These authors found that RHgCN in no case was cleaved a t the R-Hg bond by HC1, but only at the Hg-CN bond. A series of RHgCN compounds treated with HC1 according to the method of Kharasch would give quantitative yields of RHgCl according to the equation RHgCN f HC1+ RHgCl HCN
5.
+
RHgCl (s)
The analytical method used required forcing the reaction to completion by removing RHgCl as an insoluble compound. Since hydragen cyanide is distinctly more reactive than HR, where R is a hydrocarbon radical, it is to be expected that the reaction rrpresented above actually is not irreversible but should be represented as RHgCN f HCl RHgCl + HCN
it
RHgCl(s)
If this reaction is reversible as represented and is carried out a t such concentrations that RHgCI cannot precipitate, the reaction mixture will represent a case of homogeneous equilibrium, and will be amenable to the methods available for studying such systems. In a series of compounds of the type RHgCN the Hg-CN bond should be affected differently by the electron “pull” of different radicals, R, exerted through the mercury atom. The influence of R on the Hg-CN bond should be reflected in the cleavage energy of this bond. A variation of the cleavage energy of the Hg-CN (1) This research was supported in part by the Industrial Sdencc Research Fund of Iowa State College. Original manuscript r+ ccivad September 30, 1935. (2) Kharasch, THISJOURNAL,41, 1948 (1925), Kharasch and Marker. ibid , 48, 3130 (1926); Kherasch and Flenner, ibtd , 54, 674 (1932)
JOHNS AND
R. M. HIXON
bond with a variation in R should serve as a basis for arranging various R’s in a series according to their relative electron “pu1Is.” The present paper demonstrates that the reaction between RHgCN and hydrochloric acid is reversible, that the equilibrium constant for the reaction varies with R, and that the order of the radicals is the same as previously established for other reversible proce~ses.~
Experimental Ethanol was purified by refluxing the 95% alcohol with calcium oxide for twenty-four hours and distilling. The middle one-third was retained. The conductivity of this ethanol was 2.8 X lo-’. Hydrogen chloride was prepared from c. P. sodium chloride and concentrated sulfuric acid and was passed through concentrated sulfuric acid and phosphorus pentoxide, thence into ethanol. Hydrogen cyanide was prepared by dropping 1:l sulfuric acid onto potassium cyanide and was passed through long tubes of caIcium chloride, thenee into ethanol. The hydrogen chloride solutions were analyzed for chlorine by the Volhard method. The hydrogen cyanide solutions were analyzed for cyanide by the method of Sharwood.* The organomercuric cyanides and chlorides were prepared by methods described in the literature6 and were recrystallized a t least twice. The cyanides were tested for halogen and were analyzed for nitrogen by the microDumas method. The specific conductivity of ethanol solutions of RHgCN, RHgCl and hydrogen cyanide are recorded in Tables I and 11, and of hydrochloric acid solutions in Table ILI. Conductivity measurements were made a t 25 0.1‘. using a Leeds and Northrup student potentiometer and resistance boxes with the usual type conductivity cell. The null point was determined by means of ear phones. f
The reaction between RHgCN and hydrochloric acid in ethanol was shown to be quantitatively reversible by preparing separate solutions of these reactants and of thecompounds RHgCl and hydrogen cyanide, at a given initial concentration and comparing the resultant specific conductivities. Identical conductivities were ob tained in both cases as shown for the benzyl and (3) Hixon and Johns, ib&i., 49, 1788 (19271, Johns and Bxon, J
P k W C h m . , $4, 2226 (1930); Johns, Peterws and Kixoa, ihid., &4, 2218 (1930): Johns and Hixon, THISJOURNAL, 66, 1333 (1934), Crak and Hiaon, tlnd , 68, 4387 (1931); Goedhue and Hixoa, i b d , ‘56, 1329 (1934). (4) Sharwood, tbtd , 19,400 (1897) (5) Whitmore, “OPganie Compound7 of htercury,” Cheolrnl C e d q Co N e w York, N Y , 1921
EDWARI) CAKR,I. B.
892
JOHNS AND
R. M. HIXON
Vol. 60
TABLE I MELTINGPOINTS,ANALYSES, AND CONDUCTIVITIES OF RHgCN AND RHgCl
of the mixture other than hydrogel1 chloride were actually exerted to an extent within the precision of these measurements in the presence of highly N Content. yo Compound hLp., ' C . Calcd. Obsd. Conductivitya ionized hydrochloric acid it would be necessary CsHiiHgCN 146 4.52 4.66 0.0 to correct for these in order to obtain the conc1 163 .3 x 10-7 ductivity of hydrochloric acid alone. But a CzHsHgCN 56.5 5.48 5.47 .O highly ionized acid tends to suppress the ionizac1 192 .6 X lo+ tion of a weaker acid, the more so the greater CsHsCHzHgCN 127 4.41 4.36 . 2 X lo-' c1 103 2 . 5 x 10-7 the difference he!ween the strengths of the two o-ClCgH4CH2HgCN 138 3.98 4.11 0 . 2 X 10' acids. Assuming a dissociation constant for c1 111 2 . 3 x 10-7 hydrogen cyanide in alcohol of the order of magp-CHaC&dHgCN 221 4.41 4.40 0 . 2 X lo-' nitudes of 10-lo, it can easily be calculated that c1 233 2.3 x 10-7 hydrogen cyanide a t a concentration of lo-' N CsHsHgCN 209 4.61 4.56 0.0 c1 251 1.0 x 10-7 furnishes a hydrogen-ion concentration of only a-CloHrHgCN 236 3.96 4.03 0 . 3 X lo-' N in the presence of hydrochloric acid a t a Specific conductivity of 0.005 N solution minus the concentration as low as N . This contribuspecific conductivity of themlvent (2.8 X tion of hydrogen cyanide to the hydrogen-ion concentration is below the precision of the presTABLE I1 ent measurements, which precision corresponds CONDUCTIVITY OF ETHANOL SOLUTIONS OF HCN AND oC1CeH4CH2HgC1. SPECIFIC CONDUCTIVITY OF ETHANOL: to a difference in hydrochloric acid concentra2.8 x 10-7 tion of 1 X N. It would require a conHCX, Specific conductivity normality of soln. centration of hydrogen cyanide of 1N in the pres0.112 7 . 5 x 10-7 ence of hydrochloric acid a t a concentration of .0657 6.0 x 10-7 N to contribute a detectable concentration .0131 4.8 x 10-7 of hydrogen ion. The same situation exits in .005 4 . 3 x 10-7 the case of RHgC1, the ionization of which would o-ClCsHLHzHgCI be suppressed by the common ion action of the 6.0 x 10-7 .Ol chlorine ion. 5 . 1 X lo-' .GO6 The independence of the conductivity of 4.0 x 10-7 .0036 hydrochloric acid was verified for a mixture conphenyl compounds in Table IV. The conduc- taining hydrogen cyanide a t 0.1 N and hydrotivity reported for the forward and reverse re- chloric acid a t 0.005 N. The constancy of K actions was established immediately upon mix- in Table V indicates that the contribution of ing the reactants and remained constant for RHgCN, RHgCl and hydrogen cyanide to the twenty-four hours, which is as long as the con- conductivity of the reaction mixture is not ductivity of alcoholic hydrochloric acid remains greater than the precision of determining the constant. The constancy of the conductivity hydrochloric acid concentration. proves that no irreversible side reactions occur The procedure for carrying out a measuresuch as the cleavage of the R-Hg bond. Further ment consisted of preparing an ethanol solution evidence that the reaction comes to equilibrium of either RHgCX and HC1 or of RHgCl and is the constancy of the expression for K in Table hydrogen cyanide, and measuring the specific V. TABLE I11 I n the reaction between RHgCN and HC1 as CONDUCTIVITY OF HYDROGEN CHLORIDEIN ETHANOL formulated by the equation RHgCN
+H
C 1 s RHgCl
+ HCN
only one of the components of the reaction mixture, namely, hydrogen chloride, is highly ionized and therefore the conductivity should be preponderantly due to hydrogen chloride. The magnitude of the conductivities of RHgCN, RHgCl and hydrogen cyanide is shown in Tables I and 11. If the conductivities of the components
25
* 0.1'.
Pl-ormality of HCI
7.07 4.24 2.54 2.12 1.50
x
10-4
SP. COND.OF ETHANOL 2.8 X lo-' Specific
Conductivity Equivalent
4.66 X 2.83 1.73 1.45 1.03
(6) The value in water is of the order of somewhat less in ethanol.
65.5 66.0 66.9 66.9 66.6 10-10
and would be
THEREVERSIBLE SPLITTIC
April, 1938
conductivity of the solution. The concentration of hydrochloric acid in the reaction mixture was then taken as that corresponding to the measured conductivity by referring to the concentration~onductivitydata of Table 111, which represent a straight line. Most of the reactions were carried out with RHgCN and hydrochloric acid as the reactants due to the greater facility and safety of working with hydrochloric acid than with hydrogen cyanide. An initial concentration of 0.005 N was chosen for the reactions reported in Table I V since this concentration is very nearly the limit of solubility of the aromatic mercuric chlorides. Reactions of the benzyl and phenyl compounds were carried out starting with RHgCl and hydrogen cyanide a t 0.005 N as recorded in Table IV.
ORCANOMERCURIC CYANIDES
OF
893
TABLE V CONSTANCY OF EXPRESSION FOR EQUILIBRIUMCONSTANT o-CICsH&HzHgCl+ HCN Initial concn. RHgCl HCN
0.00500 .0657 ,0131 ,112
0.00500 .OlOO .0131 ,00500
o-CIGHrCHzHgCN +HCI Sp. cond. reaction mixt.
1.52 X 4.31 X 2.80 X 4.05 X
2.15 2.17 2.17 2.08
K X I@
X lo6 X 10' X 10'
The equilibrium constants recorded in Table I V vary in a regular manner from radical to radical and are in the order previously established in this Laboratory. I11 Fig. 1 are plotted curves for (1) the reaction PRHgI R2Hg HgI2, (2) the ionization of primary amines, (3) the reaction RHgCN H C l z RHgCl HCN. The ordinate measures log K while the abscissa is the axis of radicals established by Hixon and J o h ~ i s . ~
+ +
+
TABLE IV EQUILIBRIUM DATAAT 25' FOR THE REACTION RHgCl f HCN KHgCN f H + f C1REACTANTS INITIALLY AT 0.005 N [RHgCll [HCNI Radical
,
Sp. conductivity of soln.
[HCII [RHgCNI [ H 1' [Cl-] [RHgCNl
CbH11-
0.86 x 10-5 .85 X 10-5
1.18 x 10-4 1.18 X 10-4
1.45 X 10' 1.45 X IO'
CiHr
1.12 x 10-5 1.13 X
1.69 x io-' 1.60 X 10-4
5.71 X 10'
1.33 X 1.33 X 1.32 X 1.33 X
1.91 X 1.81 X 1.90 X 1.91 X
3.32 X 3.32 X 3.37 X 3.32 X
C~HCHF
10-6 10-8 10-6' 10-b'
10-4 10-4
lo-' lo-'
5.83
x
10s
10s 10'
IO'
1.52 X 1 O - I 2.02 X 10-6 2.01 X 10-6
2.20 X 10-4 2.98 X 10-4 2.97 X 10-4
2.15 X 106 8.35 X 106 8.44 X 105
Cdh-
2.02 X 10-5 2.03 X 10-6 2.03 X 10-6"
2.98 x IO-' 3.00 X IO-' 3.00 X 10-4
8.35 X 106 8.17 X IO3 8.17 X 105
a-CwH-
2.23 x
3.34 x 10-4
5.84
x
-4
108
c-ClCdhCHaf4X3Xd34-
10-6
-2
-6
-8
106
Initial reactants RHgCl and HCN. Percentage conversion t o RHgCl ranges from 97% for cyclohexyl to 93% for a-naphthyl.
Table V records the reaction between 0-CICBHdCH2HgCI and hydrogen cyanide a t various initial concentrations. The equilibrium constant has been formulated as follows
- 10 0
-2
-
5
8 3
4
2
ti
4
x"x"
20 q
Opq
Q%
x
OY
7
Fig. 1.-Radicals plotted against log K for the reactions (I) 2RHgI 4 RzHg HgI2; (11) ionizaand against -log K for the reaction tion of R"z; (111) RHgCN 4- H + f C1- S RHgCl HCN.
+
+
This formulation is in accord with the high degree of ionization of hydrochloric acid indicated by the constancy of the equivalent conductance in Table 111, and is verified by the constancy of K in Table V.
The range of pK variation in the reaction involving RlIgCN is 1.3 units which is a much smaller range of variation than in the case of the ionization of amines. The narrow range of variation in the case of RHgCN is entirely in
R. L. SHRINERAND K. E. DAMSCIIILODER
894
accord with the conception that the mercury atom is 80 large that it absorbs much of the electron “pull” of the radical which otherwise would be more pronouncedly imparted to the -CN bond. The wide range of $K variation in the case of the amines is in agreement with the fact that the nitrogen atom, through which the effect of the radical is transmitted, is small.
Summary 1.
It is shown that the cleavage of -CN from
[CONTRIBUTION FROM
THE
Vol. 60
RHgCN by hydrochloric acid in ethanol is reversible. 2. Equilibrium constants a t 25’ have been determined for this reaction in the case of the radicals cyclohexyl, ethyl, benzyl, 2-chlorobenzyl, 9-tolyl, phenyl and alpha-naphthyl. 3. The equilibrium constants vary from radical to radical. 4. The order of the radicals is that previously established for other reversible processes. AMES,IOWA
RECEIVED JANUARY 22, 1938
CHEMICAL LABORATORY OF THE UNIVERSITY OF ILLINOIS]
Derivatives of Coumaran. I. 2-Benzyl-5(and 6)-methoxyco~amaran-3-one BY R. L. SHRINER AND R. E. DAMSCHRODER The synthesis of coumaran derivatives is of interest as bearing on the problem of the structure of certain naturally occurring substances, such as tectorigenin, which may be a substituted coumaran-3-one’ or an isoflavone,2 quebracho tannin, which may be a substituted coumaran3-o13 or a f l a ~ p i n a c o l . ~Rotenone5 contains a coumaran nucleus with an unsaturated side chain in one portion of its structure.
aldehydes.6 If these unsaturated ketones could be reduced readily a t the double bond, a useful method for the synthesis of coumaran-3-ones substituted in the 2-position would be available. The reduction of benzalcoumaranones has been reported in which platinum was used as the catalyst’ but Freudenbere found that side reactions occurred with platinum as the catalyst and hence used nickel on kieselguhr. The present paper is concerned with 4 the catalytic reduction of 2-benzal-5CH~O-~)--COCHX~ + methoxycoumaran-3-one IV, which was prepared by the method of von Auwers \/-OH and PohLg Hydroquinone dimethyl I1 111 ether (I), chloroacetyl chloride and ClCHZCOC1 aluminum chloride yielded w-chloro-2AlCh hydroxy-5-methoxyacetophenone (11). CHaO-/ The latter underwent ring closure upon with alcoholic sodium acetate \)-OCH, c H 3 0 f ~ ~ ~ C H C e , H treatment ~ and the resulting hnethoxycoumaran3-one (111) was then condensed with benzaldehyde to yield IV. It was found that, in glacial acetic acid solution, the latter was reduced readily by platinum CHsO--//\-CO and hydrogen to the saturated ketone + o/LHCHzceH6 (V). An independent synthesis of V was carried out not only to establish the VI v fact that reduction occurred as indicated Since coumaran-3-ones contain active methyl(13) Kesselkaul and von Kostanecki. B ~ Y . , 19, 1890 (1896); Mameli, chim. ital., 61, I, 322 (1922); Feist and Siebenlist, Arch. ene goups, they readily condense with aromatic Gam. Pharm., 166, 196 (1927).
T
Y
,)(
(1) Shibata. J. Phavm. Soc. J a p a n , 47, 380 (1927). (2) Asahins, Shibata m d Ogawa, ibid., 48,1087 (1928). (3) Nufiez, Axales asoc. qlaim. avgexlinca, 24, 159 (1936). (4) Russell, Ckem. Rev., 17, 155 (1935). ( 5 ) 1,s Parge, Haller and Smith, ibid.. 11, 181 (1QW
(7) Drumm, MacMahon and Ryan, Proc. Roy. Irish Acad., S6B, 149 (1924); Bargellini and Monti, Cues. chim. ital., 44,11, 33 (lQl4). (8) Preudenberg, Fikentscher and Harder, A n n . , 441, 157 (1925). (9) Von AiiwPrC and Pohl. ibid., 405, 24.7 (1914): R P Y . ,48, 8: (1915).
