July 20, 1957
hfECHANISTIC h P E C T S O F
REACTION OF
h 1 4 Y D R I D E S WITIl
NUCLEOPHILES 3701
ORGANIC AND BIOLOGICAL CHEMISTRY RUTGER?,T H E S T A T E U N I V E K S l TY LABORATORY, YALE UNIVERSITY]
[ C O N T R I B U T I O N F R O M T l i E SCHOOL O F C I I E M I S T R Y ,
A S D r l i E S r E K L I X G CIlEMIS TR
Y
Some Mechanistic Aspects of the Reaction of Anhydrides with Nucleophiles BY DONALD B. DGNNEY A N D MICHAEL A. G R E E N U A U M ~ RECEIVED J A N U A R Y 19, 19.57 Anhydrides, which were specifically labeled with oxygen-18, have been treated with various nucleophiles under several conditions. The oxygen-18 content of the products of these reactions coupled with data froin the literature can he utilized tv define a multi-step process for the course of these reactions. Several reactions have beeii devised to test this mechniiisiii and have been showii to be in agreement with the postulate.
Introduction Evidence from several sources can be used to construct a tentative mechanism for the reaction of anhydrides with nucleophiles. Berliner and Altschu13have made a study of the kinetics of hydrolysis of several aroyl anhydrides. Their data indicate that the rate-determining step in the hydrolysis involves attack by the nucleophile on the carbonyl carbon. The observed effect of ring substituents is in accord with this view. Bunton, Lewis and Llewellyn4 have shown by using an oxygen-18 tracer that 0
I1
an intermediate C ~ H ~ - C ( O H ) ~ - C - Cmust ~ H ~ be in equilibrium with the starting materials in the aqueous hydrolysis of benzoic anhydride. Gold and Jefferson5 have studied the tertiary amine catalyzed hydrolysis of acetic anhydride and have concluded 0
that
the
reactive
species
is
I1
i-
[CHI-C-N(R)3
0
II
CHS-C-0-1 although rigorous evidence for its formation was not obtained. Emery and Golds have studied the product ratios obtained from the reaction of anilines with mixed aliphatic anhydrides. They have found a marked effect of solvent on these ratios as well as an apparent steric factor which alters the ratio of the product^.^ These observations indicate strongly that the attack by the nucleophile, :NH, on an anhydride proceeds by a slow rate-determining step to give I which is in equilibrium with starting materials. Conversion of I to products will depend to some degree on the nature of the nucleophile. If, for example, it is an anion then proton transfer will not be required; on the other hand, amines, alcohols, etc., will require (1) A portion of this work has been reported as a Communication. D. B . Denney and M. A . Greenbaum, THIS J O U R N A L78, . 877
(19%).
( 2 ) Alfred P. Sloan Postdoctord~ Fellow at Rutgers, The State University. ( 3 ) E. Berliner and L. H. Altschul, THIS JOURNAL, 74, 4110 (1952). (4) C. Bunton, T. Lewis and D. Llewellyn, Chem. and Inn., 1154 (1984). 'i) V . Gold and E. Jefferson, J. Chem. Soc., 1409 (1953). /'/I 1. Emery and V. Gold. i b i d . , 1443 (1950). ( 7 ) A study in this Laboratory of the ratio of products obtained f r o i n mixed aroyl anhydrides as yet has not led t o the large solvent r f f r c t s found by Emery and Gold.
Further discussion of the points r.ii>cd by these experiments will be deferred until a later paper as thcy d o uut affect the arguments advanced in this paper.
that a proton be transferred before the products are formed. I t is the latter situation which is of concern here. There are two attractive paths for the formation of products from I. Path A involves an ionization of I to give the fragments of 11. Reaction of some base, :B, i.e., solvent, nucleophile or R-COO- completes the process and one ohtains the observed products. If Path A is followed and one uses an anhydride which has been labeled as indicated, then the distribution of oxygen- 18 in the products will depend on the rate of return from I1 to starting materials as compared to the rate of collapse of I1 to products.* If return to starting materials is much faster than product formation, then the oxygen-I8 in the anhydride will become equilibrated, since in I1 R-COO- allows equilibration of the label. If kq is much larger than k-1 and k--2, then there will be no equilibration, as I1 will never return to starting materials. I t call be readily seen that product formation by Path R will not lead to equilibrated products. If one uses a labeled anhydride and if these two paths are the ones being followed, it is possible to get some measure of the relative importance of return from I1 as conipared to the non-equilibrating routes by measuring the oxygen-18 distribution in the products. Unfortunately] a further complication arises since benzoic acid, a product of the reaction, can react with the anhydride to give equilibrated material. I t has been possible t o show that this reaction contributes very little to the oxygen-18 equilibration. Results and Discussion Synthesis.-The preferred method of preparing the monocarbonyl labeled benzoic anhydride is to allow silver benzoate to react with labeled berizoyl chloride. This reaction affords good yields of specifically labeled anhydride. I t was by this procedure that IV was converted via the acid chloride to VI1 (see Table I). Another method used for preparing the monolabeled anhydride was to allow sodium benzoate in water to react with labeled benzoyl chloride (from 111) in chloroform. Using this procedure VI was prepared; however, because of the unpredictableness of the method, it cannot be recommended. When labeled benzoic acid, V, was allowed to react with benzoyl chloride and pyridine in ether, an excellent yield of VI11 was obtained. (8)The reasoning which is being applied bere depends upon the fact that the products are not in equilibrium with the starting materials; that this is the case is abundantly demonstrated by qualitative organic chemical data.
Path A
Path B
r
7
I
0
-0'8
k2
+ fk..
I
I
R-C-0-C-R
2
:B ka
--+t-
+ R-
k-r
0'8
- mr
1
I1 k 4 i
O'S/X
I1
R-C--N
R-C
I
:B x22
0'8
+ R-COO + 6 H
L
R-C
!I
k i i tk-1 0 11 C-R+
O-
1
--0 --C-R I
I
?;
I11 i k 5
0'8/2
I1
:NH
R-C-N
+ R-COO-
+
4-BH
'0'
It was shown t h a t VI11 was partially equilibrated with respect to its oxygen-18, and therefore this method is unreliable. TABLE I Compound
Benzoic anhydride (VI) Benzoic anhydride (VII) Benzoic anhydride (VIII) Benzoic acid (111) Benzoic acid (IV) Benzoic acid (V) P-Nitrobenzoic anhydride (IX)
Atom
Yo0 ' 8
dride which had been dissolved in ether gave the same oxygen-18 distribution as untreated material when they were allowed to react with aniline in water-acetone.
in labeled position
1.32;1.30 1.oo; 0 . 9 9 1.19;1.19 1.32; 1.31 1.02;1.01 1 . 1 9 ; 1.18 1.13;1.11
Position of Label.-The distribution of oxygen-18 in these anhydrides was elucidated by allowing them to react with ammonia a t -33". The oxygen-18 content of the benzamides and benzoic acids derived from VI and VI1 indicated that all of the excess oxygen-18 was incorporated in the carbonyl groups. The methods of synthesis limit this excess oxygen-18 to only one of the carbonyls. When VI11 was treated in this manner, the results showed that VI11 was about 357, equilibrated. This equilibration presumably occurred during the preparation of the anhydride. The reaction of these anhydrides with ammonia proceeds without any equilibration; therefore, Path B must be operating or no return from 11 is taking place. Such a finding is not unexpected since the excess ammonia will make the conversion of I1 to products and of I to I11 very rapid, thereby eliminating return. It is very interesting to note that a t - 78" the reaction of the anhydrides with ammonia leads to completely equilibrated products. At this temperature ammonia apparently exists as rather tightly hydrogen bonded oligmers. Diffusion of these large groups to the reaction site will be slow compared to the reaction a t -33", and therefore return from I1 can compete effectively with the non-equilibrating processes. Equilibration Paths.-When the labeled anhydrides were allowed to react with aniline in ether, the products showed complete equilibration of the oxygen-18. If the anhydride interacts very rapidly with the ether to give an ion pair, X, then collapse of X to the starting material would lead to equilibration. It has been shown that the equilibration does not take place through X since anhy-
I
CzHs
X
Another possible path for the equilibration involves a rapid attack by the benzoic acid being formed in the reaction on unreacted anhydride. T h a t this path contributes little to the equilibration has been shown by allowing molar ratios of benzoic anhydride, aniline, dimethylaniline and benzoic acid-C14to react in ether. The benzanilide contained 15y0of the possible carbon-14 calculated on the basis of complete exchange of all benzoyl species. This figure represents an unreal maximum since the concentration of benzoic acid under the normal reaction conditions is much lower than that used here. I n the absence of aniline the amount of exchange was considerably higher, 46%. The other path which can lead to equilibration involves return from 11. I n order to test this possibility, a molar ratio of anhydride to aniline of 2 : 1 was allowed to react in ether. The unreacted anhydride was recovered and treated with ammonia. The benzamide contained 387, of the excess oxygen-18 and the benzoic acid 587,. These figures correspond to about 69% equilibration and amply demonstrate that equilibration of the starting materials is taking place. Effect of Basic Solvent.-An inspection of the proposed reaction scheme shows that in all paths some base : B is required before the products can be formed. I n Path A benzoate ion can function as the base and, since it is intimately associated with the cation, it should have ample opportunity to remove the proton. It also has the opportunity to function as a nucleophile and cause return to I. If the benzoate ion acts primarily as a base, then one would expect t o find non-equilibrated products, which is not the case in most of the examples studied. The prime function of the benzoate ion therefore appears to be that of a nucleophile, and it nttacks on the cation to give return to I.
MECH.~NISTIC ASPECTSOF REACTIOK OF ANIIYDRIDESWITII NVCLEOPHILES
July 20, 10.77
3703
TABLE 11" Reactants and conditions
+ iYH3 a t -33' + NH3 a t -33" VI11 + NHs a t -33" V I 1 + NH3 a t -78" V I + CcH5-NH2 in ether VI1 (after C6HjNH2 in ether) + NHI VI + CeHj-?iHz acetone/H20(1:2) 1/11 + cyclo-C6HliNH2 in CeHj-N(CHa)* VI1 + cyclo-CeHIIND1in C&-N(CH3)2 1-11 + CHIOH IX + C6H6-NH2 in ether VI1 + c y ~ l o - C 6 H ~ ~ - N inHacetonitrile ~ VI1 + C H 3 0 H in acetonitrile VI1 + C6H5-XH2 in acetonitrile VI1 + C6H5-NH2 in D M F
VI VI1
Amide or ester-0'8
Renzoic acid-0'8
0.86,0.85 .57, .57 0.85 0.46,O. 46 .64, .64 0.50 n , 81, 0 . 8 0 0.51 I46 .4G
D.66,0.65 0.40 1.03 0.46 0.64,O. 64 0.43 n x , o . 56 0.44
.67 .47 .4G
.47 47
76
Equilibration
Dev. from total 0 ' 3
0
0 35 100 100 69 23 62 100 in0 74 100
0.00
-
I).os
-
0.03 0 00 0 [)(I
0.06 0.01
n .or)
+ o.ni + 0.01 + u.01 - 0.03 + 0.09
.46 .45 100 .47 .47 100 - 0.03 .46 .45 100 Analyzed by the method of \Ir. E. Doering and E. Dorfman, THISJOURSAL, 75, 5595 (1953), as slightly modified by D. B. Denney and M. A . Greenbaum, ibid., 78, 979 (1957). In the case of the anhydrides and products the observed oxygen-18 values appear t o be good to f0.01atom yo oxygen-18. I n general, in work of this sort the tendency is t o obtain low values for the oxygen-18 content since any dilution or exchange causes a decrease in oxygen-18 content. T h e deviation from the expected total oxygen-18 content of the products has been listed in the table, and in all cases but one it is equal t o or less than +0.03. These values represent the sum of the errors over three positions. The calculated per cent. equilibration is accurate t o about &8%, in the case of products from VII. In the case of products from VI i t is about =t5%. The reason for the difference between V I 1 and V I is because of t h e greater oxygen-18 content of V I ,
The presence of added base in the reaction mixture should lead to a net decrease in equilibration, all other things being equal. This will arise since the rate of conversion of I1 to products will be increased as will the rate of formation of 111. The reaction in ammonia demonstrates the importance of base concentration in controlling the amount of equilibration. I n order to find other examples of this behavior, the labeled benzoic anhydride was allowed to react with aniline in acetone-water (1:2). Analysis of the products showed that 23%.equilibration had occurred. Under these conditions the water is functioning as a base and therefore decreases the rate of return from 11. I1 is also probably much more stable in this solvent mixture than in solvents of lower dielectric constant, and the ions should therefore be much more loosely bound. Any diffusion away from the cation by benzoate ion will lead to a lowering in the rate of return from I1 to I. Prof. Philip A. Vaughan (Rutgers) has derived an equation which relates the ratio of the rate of return from I1 to the rate of those processes which lead to products, i.e., kd and Path B. This equation only applies to those situations in which an effective base is present in excess throughout the course of the reaction. The equation is li
= 0.333
0.166 + 0,75r
+
where V = the observed excess oxygen-18 content in the product benzoyl species normalized to a starting value of 1.00. r = the ratio of the rate of return from I1 to product formation by all paths, and r is also equal to the average number of returns from 11. Using this equation r can be calculated for the reaction in acetone/water and it is 0.22. This result is indicative of the delicate balance of rates between return and product formation. It was deemed desirable to have an example of the reaction in a basic organic solvent. The choice
of such a solvent is rather difficult since they can in general react with the anhydride to cause equilibration. It was found that the reaction of cyclohexylamine with labeled anhydride in dimethylaniline afforded products which were G2% equilibrated. I n this example a t least some of the equilibration may have been caused by the dimethylaniline. A calculation of r for the above reaction shows it t o be 1.4. In this case it is interesting to note that although dimethylaniline is a stronger base than water, more equilibration takes place than occurred in the aqueous solution. At least a part of this effect is due to the lower dielectric constant of dimethylaniline which will cause I1 to be "tighter" and will facilitate return. When methanol was used as the solvent as well as the nucleophile, the products contained completely equilibrated oxygen-18. Since methanol is a weaker base than either water or dimethylaniline and since its dielectric constant is rather low as compared to water, it is not surprising that one finds complete equilibration in the reaction. Effect of Deuterium Substitution.-Inspection of the proposed reaction scheme indicates that any diminution in the rate of conversion of I1 to products or in the rate of Path B should lead to increased oxygen-18 equilibration. Deuterium substitution for hydrogen to give : N-D will lead to intermediates containing deuterium and presumably to a diminution in rate of hydrogen transfer since the energy required to transfer the deuteron will be larger than that required for the proton. When cyclohexylamine-ND2 was allowed to react with labeled anhydride in dimethylaniline, the products showed complete equilibration of the label. These results are to be compared with the undeuterated material in which the amount of equilibration was 62%. Clearly there is an isotope effect which leads to the predicted behavior. Effect of Substituent.-When p-nitrobenzoic benzoic anhydride-p-ni trobenzoyl-carbonyl-O18(IX)
370-1
DONALD B.
D E N N E Y AND AIICHAEL
was allowed to react with aniline in ether, tlie 9nitrobenzanilide contained 5 2 7 , of the excess oxygen-18. Several factors can contribute to the difference found between the amount of ouygen-lS equilibration in the substituted and unsuhstituted cases. One difference is that in the substituted case one can isolate product containing only the original labeled position, i.e., p-nitrobenzanilide. I n this manner the dilution factor which was present in the symmetrical case is removed The presence of the +-nitro group will enhance the rate of attack a t the carbonyl adjacent t o the ring bearing the p nitro group. Increased attack a t this position causes a slower equilibration than an equal number of statistical attacks. There are considerable complications attendant with any attempt to calculate the number of returns from I1 in this case, since I1 now consists of two different ion pairs wh1ch are formed at different rates and which have different k - I ' S , etc. It is possible to make a very naive calculation if several assumptions are made. I t has been assumed that the ratio of the rate of attacks a t the substituted benzoyl moiety as compared to the benzoyl is 3: 1. This assumption is based on the product ratio, there being 7570 9nitrobenzanilide to 25y0 benzanilide. I t is also necessary to assume that ka and ka are zero until the proper degree of equilibration is reached and then Lccome infinite. Such a n assumption leads to a value for the number of returns (the ratio of the ratis is not being calculated here) that is smaller t h ~ i the i actual value. The equ,ition usecl in this calculation is one that arose during the derivatiljn of the equation reported abow. 11 is
where P x = observed excess oxygen-18 in the benzoyl product normalized to a starting value of 1.0.
,..-
pr
