CYCLICSATURATED OXYPHOSPHORANES
Nov. 5 , 1963
bonds. The resulting Cz radicals proved to be effective in capturing the “hot” P32atoms for formation of chemically stable C2-P3’ derivatives. The possibility that the organic P3*compounds described in this paper are the chemical reaction products of P32-oxya~ids and glycerol was eliminated by the experimental observation that a mixture of glycerol and HaP320.1 a t room temperature yielded very little glycerophosphorous acid ; no other compounds were observed. ?-Irradiation products of a H3P3’04-glycerol mixture contained unchanged H3P3’04, some glycerophosphoric acid-P3’ (probably in an equilibrium concentration), and other unrecognizable substances which were not identical with the neutron activation products. None of the neutron activation products was detectable by paper chromatographic spray reagents, whereas the starting materials, H3P3104 and glycerol, were easily detected by these reagents. Thus the organic products are considered t o be products of “hot atom” reactions of P32with the medium. Nascent PSI atoms from related nuclear processes
[COXTRIBUTIOS FROM
THE
DEPARTMENT OF CHEMISTRY OF
THE
3465
have recoil energies comparable t o those of P32atoms. The products of their reactions would be organophosphorus compounds of P3I. One might consider the sources of the phosphate esters which could accumulate in extra-terrestrialZ0 environments during chemical evolutionz1 and prior to origin of life. Some of these might result from nuclear processes on silicates t o yield organophosphorus derivatives which could be oxidized by suitable minerals. The surprisingly rapid metabolism of phosphonic acids by primitive microorganismsz2 and the occurrence of a phosphonic acid (2-arninoethanephosphonic acid) 2 s in nature suggest synthetic applications and possible relationships of reactions revealed in this study. (20) I t is recognized t h a t terrestial conditions did not include an appreciable neutron flux during the era of chemical evolution. (21) (a) M . Calvin, American ScicntiJ1, 44, 248 (195G), (b) M. Calvin, Suturzoissenschajien, 17, 387 (1956). (22) 1,. D. Zeleznick, T. C. Myers, and E. B . Titchener, Federalion P r o c . , 80, 41 (1961). (23) ( a ) M. Horiguchi and M. Kandatsu, Nafure, 184, 901 (1959); ( b ) J. S. Kittredge, E. Roberts, and D. G . Simonsen, Biochem.,1 , 621 (1962).
STATEUNIVERSITY OF NEW YORKAT STONY BROOK,N. Y.]
Cyclic Saturated Oxyphosphoranes and their Hydrolysis to Cyclic Phosphate Esters. Diastereomeric 2 :1 Biacetyl-Trimethyl Phosphite Adducts1
The
BY FAUSTO R A M I R E Zh’. , ~RAMASATHAS, A N D N . B. DESAI RECEIVED J U N E 6, 1963 Trimethyl phosphite reacts rapidly with the a-diketone biacetyl t o form exclusively a 1 :1 adduct having a cyclic unsaturated oxyphosphorane structure with the new l&dioxaphospholene ring system. The 1 : 1 adduct reacts slowly with more biacetyl t o form t w o diastereotneric forms of a 2 : 1 adduct having a cyclic saturated oxyphosphorane structure with the new 1,3-dioxaphospholane ring system. The pentavalent-phosphorus structures are based o n : solubility in pentane, Pal n.m.r., H’ n.m.r., and infrared spectra, and on the results of hydrolysis in aprotic solvents. The diastereomeric 2 : 1 adducts undergo a very rapid and exothermic reaction with one mole eouivalent of water and vield diastereomeric cyclic phosphotriesters. A stereomutation at phosphorus was o b s e r k d in the meso-cyclic phosphotriester.
We have described new organic compounds of phosphorus in which the phosphorus atom appears t o be covalently bound t o five oxygen atoms I C Typical of these substances is the biacetyl- trimethyl phosphite 1:l adduct I This is formed rapidly and exothermally when the diketone biacetyl is added t o one mole equivalent of the phosphite ester a t IOo The properties of this pentane-soluble adduct are in better agreement with a cyclzc unsaturated oxyphosphorane structure, I , derived from the 1,3-dioxaphospholene ring system, than with ionic structures IC ai3 C H ~
r
CH,
CH~
Results The biacetyl-trimethyl phosphite 1 :1 adduct I reacted slowly with a second mole of biacetyl yielding a colorless, distillable oil. Analysis of this material agreed with the formula CI1H210;P, which corresponds t o a biacetyl-trimethyl phosphite 2:l adduct. The data discussed below show t h a t the oil is a mixture of the two possible diastereomeric forms of a cycltc saturated oxyphosphorane, I I a and IIb, a new type of organic compound with pentavalent phosphorus, derived from the 1,X-dioxaphospholane ring system. In one of the diastereomers, IIa, the two acetyl groups are cis t o each other; in the other, IIb, they are trans t o each other. The isomers are formed in the approximate proportion of 4 : 1 , respectively. CH3
1
c=c L
Experiments in which the biacetyl/phosphite mole ratio was greater than one, and in which the reaction was allowed t o proceed for several hours, yielded a new substance. This observation was investigated in greater detail and the results aredescribed in this paper.3 (1) ( a ) Organic Compounds with Pentavalent Phosphorus, Part X. ib) Part I X F. Ramirez, N . B . Desai, and N . Ramanathan, J . A m r h e i n . S n c . , 86, 1874 (1903); (c) P a r t V I I I : F Ramirez and N . B. Ilesai, ibid , 86, 3232 (19fi3). and relerences therein. ( 2 ) Alfred P. Sloan Fellow 1901-19B:i. Acknowledgment is made t o the Cancer Institute of the Kational Institute of Health (CY-1769), the National Science Foundation (G 19509): and the Petroleum Research Fund of the A.C.S. (28F-A) for generous support of this research. (3) A preliminary account of parts of this work appeared in ref. I b and in J . A m . Chem. S o c , 84, 1317 (19G2)
~
CHJ I
CH3 CH3
I + c---c
I
O\p/O CH3O’ ‘OCH? OCH3
IIa
I1
0
il
0
slow
---+ 200
I
IIb
Table I (expt. 3 and 1) shows t h a t an excess of biacetyl leads t o higher yields of 2 : 1 adducts, IIa t
FAUSTO RAMIREZ,N. RAMANATHAN, AND N. B. DESAI
3466
Vol. 85
P.
---
3.0 .
_-
4.0
5.0
6.0
7.0
8.0
9.0 10.0 l l . @
VL.
0
- \36 L I Q U I D MIXTUQE 3500 3000 2000
1800 1600 Cm.-'.
1400
1200
t
IIb. The reaction time is important, as seen in expt. 3 vs. 5, and 1 vs. 2. The 1: 1 adduct I need not be isolated in the preparation of the 2 : 1 adducts IIa and IIb. This is shown in expt. 6-9 of Table I. For example, three moles of biacetyl were added, dropwise, with stirring, to one mole of trimethyl phosphite, the mixture being kept a t 10' during the exothermic phase of the reaction. The second phase of the reaction was then allowed to proceed a t 20' for 5 days (cf. expt. 7 ) . TABLE I REACTIONO F BIACETYLWITH THE BIACETYL-TRIMETHYL PHOSPHITE 1 : 1 Adduct I, AND WITH TRIMETHYL PHOSPHITE,AT 20°, IN A NITROGEN ATMOSPHERE Mole ratioa
Expt.
Reacn. time, hr.
2:l adduct,b
1 120 67 1 96 52 3 120 91 4 3 48 87 5 3 24 40 6 2.5 72 83 7 3 120 87 8 3 170 87 9 4 96 92 a Biacetyl/l:l adduct ( I ) in expt. 1-5; biacetyl/trimethyl phosphite in expt. 6-9. Distilled mixture of mem and racemic isomers (IIa I I b ) of the biacetyl-trimethyl phosphite 2 : l adduct, ~ Z S D 1.4471, collected a t 65-75" (ca. 0.1 mm.); moisture must be excluded in handling the 1: 1 and the 2: 1 adducts. 1 2 3
u
-
60.0M C .
L, 4FIELD
1000
Figure 1.
_I-
-a
- ...('
I...
Fig. 2.-Proton n.m.r. spectrum of the original mixture of diastereomeric biacetyl-trimethyl phosphite adducts in CClr solution at 60 Mc. The line of the cyclohexane used as internal reference has been omitted. The shifts relative t o tetramethyl~ c.P.s.), 7.86, and 8.69 T ; silane are: major isomer, 6.51 ( J H 13 minor isomer, 6.40 ( JHP 13 C.P.S.), 7.75, and 8.83 T .
~ . p . m . ~to" high field of the external reference, 85% H3P04. The liquids were examined neat, and the solid in a concentrated benzene solution. The P31chemical shift of the 1: 1 biacetyl-adduct I is +53 f 2 ~ . p . m . ~ ~ as. 85% H3P04.1C These relatively large positive cheniical shifts in the P31n.m.r. spectra suggest pentacovalent phosphorus in the adducts, ;.e., oxyphosphorane structures. IC (b) The infrared spectra of the three materials, (l), (a),and ( 3 ) , in cc14 solution were very similar. The spectrum of the crystalline major isomer IIa is reproduced in Fig. 1. The split band a t 5.81 and 5.84 p is due to C=O stretching vibrations. The very strong bands a t 9.12 and 9.26 p are due to POCH3 stretching vibrations. The corresponding bands in the spectrum of the 1: 1 biacetyl adduct I are found a t 9.18 and 9.30 p . l C (c) The H1 n.m.r. spectra of the three materials, (l), (2), and (3), in cc14 solution showed signiTcant differences. These spectra are reproduced in Fig. 2 , 3, and 4, respectively. Figure 2 shows that the original
+
The adducts I I a and I I b are remarkably soluble in pentane; for instance, 40 g. of the original liquid mixture dissolved in 10 ml. of pentane. Crystals of the major isomer I I a separated from this solution in ca. 60y0 yield (or 53y0 yield based on trimethyl phosphite, in the case of expt. 7, Table I ) . The minor isomer I I b has not been isolated in a stereochemically pure form. The viscous oil which remained after removal of the crystalline isomer I I a consisted of approximately equal parts of both isomers IIb. The yield of this enriched isomer mixture IIa was 35% after distillation (or 30YG yield based on trimethyl phosphite in expt. 7 ) . For reasons discussed below, the meso-configuration IIa with cis-acetyl groups is assigned to the major crystalline diastereomer. The racemic or DL configuration IIb with trans-acetyl groups is assigned to the minor diastereomer. Structure of the Biacetyl-Trimethyl Phosphite 2 :1 Adducts Based on Physical Data.-Three types of materials must be discussed: (1) the original liquid mixture of isomers; (2) the crystalline major isomer which was obtained from this mixture by crystallization from pentane; (3) the liquid mixture of isomers which remained after the crystallization process. (a) The P31n.m.r. spectra of these three materials, ( l ) , (?), and ( 3 ) , gave one strong peak a t +51 f 2
16.0
INTERNAL
..
-131
q!s
10% IN
-44.8
2: I
cc \4
BIACETYL-
vl
@
- I19
--
INTERNAL
