COMMUNICATIONS TO THE EDITOR
Dec. 5, 1963
to cis-2-phenyl-2-butene in deuterated hydroxylic solvents also supports the concept that hydrogen-bonded carbanion collapse to hydrocarbon occurs a t rates comparable to the rates of solvent molecule interchange a t the carbanionic site.5 Exchange occurred with high retention with either ammonia or propylamine in tetrahydrofuran (runs 1 and 2), or ammonia in tert-butyl alcohol (run 3), but gave racemization with ammonia in either methanol or dimethyl sulfoxide (runs 9 and 10). In these latter solvents, ion-pair dissociation is apparently faster than other processes (kz > k3 and possibly k - ] ) . Thus in nondissociating solvents, asymmetric solvation of a planar or near-planar carbanion accounts for the results, whereas in dissociating solvents, a symmetrically solvated carbanion is formed. Exchange went with high retention in benzene-pheno1 with either potassium phenoxide or tripropylamine as base (runs 4 and 6), but gave racemization in tertbutyl alcohol (run 11) with potassium phenoxide. Stereospecificity is undoubtedly dependent on very rapid proton transfers, which are in turn dependent on large pKa differences between the solvent (or proton donor) and substrate. With phenol (or ammonium ions), the difference is about 13 pKa units, whereas with tert-butyl alcohol, it is only about 4 pKa units. Retention in runs 4 and 6 is probably dependent on the fact that greater charge separation would result from capture of the carbanion from the rear than from the side of the leaving group.6 The low dielectric constant of the medium would suppress such a process. The potassium ion also probably plays a solvent-orienting role at the front side.
Asymmetrically solvated ion pair
Ion pair
Ion separated by inverted product
With tripropylamine as base in methanol, high inversion was observed (run 12). In this dissociating solvent, charge separation by back-side proton capture is not serious. The tripropylammonium ion shields the front side, thus promoting back-side proton capture.
+
\i
CH~OH.~*C-.D--N(Pr)3
I
-
CH3T) H-C
,+
+
DN(Pr)S
\ Run 13 was conducted in tetrahydrofuran with triethylamine as base, and lithium bromide was added so a reasonable temperature could be employed. Although attempts were made to exclude moisture, some undoubtedly was present, since k , / k , = 0.23. In the complete absence of moisture, k , / k , should become exceedingly low valued, since tetrahydrofuran or tripropylamine should be very poor proton source^.^ (5) D. J. Cram and R . T. Uyeda, J . A m . Chem. Soc., 84, 4358 (1962). (6) The author is indebted t o Dr. A. Streitwieser f o r first pointing out this relationship (private communication). ( 7 ) This research was sponsored by the U. S. Army Research Office, Durham, N. C .
DEPARTMENT OF CHEMISTRY THE UNIVERSITY OF CALIFORNIA AT Los ANGELES
DONALD J . CRAM LAWRENCE GOSSER
Los ANGELES,CALIFORNIA RECEIVED SEPTEMBER 5, 1963
389 I
Reactivities of the Primary Reducing Species Formed in the Radiolysis of Aqueous Solutions
Sir: It is currently recognized that two primary reducing species are produced by the action of ionizing radiations on aqueous systems, viz. the electron (or negative polaron]), represented here by (HzO)-, and a dehydrogenating specie^,^,^ presumed to be a hydrogen atom, which is designated in the following by H a . Considerable interest has recently centered around the use of pulse radiolysis techniques to determine the absolute rates of reaction of (H20)- with different s o l ~ i t e s . ~ In this communication we give some relative rates of reaction of the two primary species toward different solutes. All of these experiments have been carried out with Co60 7-rays in the absence of oxygen. 1. Reactivity of (H20)--.-It has been shown5 that NzO has a relatively high reactivity toward (H20)-, leading to the formation of nitrogen according to X2O
+ (H20)- --+
Nz
+ OH + OH-
(1)
Substances which compete efficiently with NzO for (HzO)- lower the nitrogen yield, and the relative rates can thus be obtained. Solutions of NzO (1.6 x lop2 M ) a t neutral pH have been used for these experiments, the yields of nitrogen being determined mass spectrometrically. Table I, column a, shows some of the results obtained, taking k [ ( H 2 0 ) NzO] = 1.00. Comparison with relative rate constants calculated from the data of Gordon, et ~ 1 (Table . ~ I , column c), shows good agreement for several of the solutes in-vestigated. Table I shows that nitrate ions also have a high reactivity toward the radiation-produced electrons. The reduction of nitrate to nitrite has been studied in sodium nitrate solutions containing 2-propanol (lo--] M ) , nitrite being determined by the method of Endres and Kaufman.G The observed yields of nitrite were G(NOZ-) = 2.52 and G(NOz-) = 2.86 a t sodium nitrate concentrations of and M , respectively. Addition of other electron acceptors to the nitrate2-propanol solutions led to a reduction in the radiolytic yield of nitrite. Relative rates compared to nitrate were then calculated assuming simple competition, in which nitrite is formed only as a consequence of reaction of (H20)- with nitrate. Some results normalized to k [ ( H z O ) NzO] = 1.00 are given in Table I, column b. It can be seen that there is a good measure of agreement with the corresponding relative rates obtained by the NzO system. 2. Reactivity of Hydrogen Atoms (Ha).-In the radiolysis of deaerated aqueous solutions of sodium deuterioformate containing another organic solute (RH) which can be dehydrogenated, the ratio G(HD)/ G(H2) will be governed by the competing reactions H + DCOO-+ H D -t COO(2)
+
+
H
+ R H +H2 + R.
(3)
If the experimental conditions are such that H" is the only dehydrogenating species entering the above competition, measurement of the yields of H D and of H? should lead directly to the rate ratio k ( ~ iu I)CO~~-):I k ( ~+ R H ) . These conditions are attained if NzO is (1) J. Weiss, Nolure, 186, 75 (1960). (21 J. T. Allan and G. Scholes. i b i d . , 197, 218 (19RO) , 868 (19621, J . Rabani and G. Stein, (3) J. Rabani, J . A m . Chem. S O L .84, J . Chem P h y s . , 37,1867 (1962). (4) S . Gordon, E . J. Hart, M . S. Matheson, J Rabani, and J . R Thomas. J. A m . Chem. Soc., 86, 1375 (1963); Discussions Faraday S O L ,Noire Dame, Paper No. 14, (1963). (5) F. S. Dainton and D . B Peterson, S a l u r e , 186, 878 (1960); P r o c . R o y . SOC. (London), A867, 443 (1962) ( 6 ) G. Endres and L. Kaufman, A n n . , 172, 530 (1940)
3 892
COMMUNICATIONS TO THE EDITOR
Vol. 85
TABLE I REACT~VITIES OF DIFFERENT SOLUTES TOWARD ( H s ~ ) -NORMALIZED , RELATIVETO
-
~
_
_
~
a
I n the
cu2
+
N 2 0
[ s ~ o=]
Solute, S
El, m M 0.28-4.75 1-63 23 10
( CuSOa)
H+ Thymine NO3 N*O Fe( CN)& Acetone NOS -
...
10 10-100 100 2Od 5200
cos
systems
IO m,M
R
[SI,m M
3 . 8 5 f 0 30 1 7OfO.15 1 . 6 7 5 0 15 1.17f 0.15 1.00 0.89 f 0.07 0 . 6 7 i 0 05 0 . 4 9 f 0.05 0.46 f0.04 0 19 f 0 . 0 1
...
(CH3)COH
+ Hz
(4)
(5)
TABLE I1 REACTIVITY OF DIFFERENT SOLUTES TOWARD HYDROGEN ATOMS( H a ) NORMALIZED RELATIVE TO THE REACTION WITH DCOO(Solute], mM
Fe(CS)63Ally1 alcohol Benzyl alcohol Cu*-(CuSO,) HCOO-
1 3 4 0 28-4 7.5 10
NOS -
100
2-Propanol GI yoxalate NO8 -
10 10 100
ncoo -
[DCOO-1, m A4
mM 100 100 100 100
10
100 10 5 100
10
3
Ethylene glycol Chloroacetate Methanol Acetone Acetate
10 10 100 100 100
20 2 10 2 2
I-Butyl alcohol
100
1
Ethanol
(Isopropyl],
Relative reactivity
+
180 40 104 rt 23 29 i 6 28 i 7 66zk07 6.1
A l l solutions contained N 2 0 (16 r n l f ) .
*10
225*02 1 6 rt 0 . 1 1 1 i0.4 1.0
0.7
. .
+ 0.1 +
8.4 7,4 2 8 1.2
+
X lo-' 1 . 2 X IO-*
+
i o +0
*
3 1
x x
... . .
1\;20
-
C
From absolute rate measurementse
z?& 3 81 f 0 . 6 2 72 f 0 . 5 ...
1.27 f 0 . 2 1.oo 0.92 i0 . l j c 0.68 f0.07 . .
0.88 f0 . 2
5 0 14 f 0 . 0 2 ... 300 2.22 + 0 . 3 x 10-3 ... Solution contained 16 rnM S Z O . .It the corresponding ionic strength.
From the measured hydrogen yields, the rates Of reaction of the solutes relative to Z-propanol (k4"kd have been obtained and these then normalized to k(HOL + DCOO-1 = l.0 '1). It is not necessary to have NzO present in these experiments, since the solute itself may scavenge the (HzO) -. This latter method is, in fact, somewhat similar to t h a t used by Rabani and Stein3using either ferricyanide or nitrite and Organic which can be dehydrogenated.
Solute'
...
2 08 f 0 . 3 0 1.19 i0 . 1 5 1.00 0.88 f0.08 1 . 0 1 f O 10
...
which is in competition with hydrogen atoms reacting with a solute ( X ) . H+X+XH
. .
. . .
+
--f
R
...
1 2
also present, since possible complications arising from reactions of (HzO)- with the organic solutes, and with water to give hydrogen atoms, will be suppressed owing to the occurrence of reaction 1. Table 11 includes some relative rates obtained using this technique, k ( ~ +a IICOO-) being taken as 1.0. I n those instances where the solutes do not yield hydrogen on reaction with H" (ex., allyl alcohol, 2Fe(CN)e3-) the system (NzO (1.6 X lo-' propanol (10-1 M) } has been used. Here, hydrogen arises from the molecular process and from the reaction
+H
REACTIOSWITH
0.44 lo* ..,
Chloroacetate H2P04. . ... a Maximum solute concentrations in the range 0.1-0.3 m X . Solution contained 8 mill N20;pH 3.5-4.0. e See ref. 4.
(CH3)ZCHOH
THE
Relative reactivity, b I n the NOa- systems [SOa-] = 1 m M
10-2 10-2 lo-'
These authors have reported that the relative reactivities of ferricyanide, nitrite, formate(HCOO -), and 2-propanol toward €3" are in the ratio 1 : 0 . 2 2 :0.055: 0.h13: The present work gives the corresponding ratio
as 1:0.033: 0.037: 0.012 which, apart from the case of nitrite, is in quite good agreement. Acknowledgment.-We thank Professor J . J . Weiss for his encouragement of this work and hfr. p. Kelly for mass spectrometry. mre are indebted to the u. s. Army for financial support, through Contract No. DA91-591-EUC-27j(). OF SCHOOL OF CHEMISTRY
THEUNIVERSITY XEWCASTLE LIPON T Y N E , 1 ENGLAND RECEIVEDACGUST19, 1963
A . APPLEBY
G. SCHOLES M . SIMIC
A Neutron Diffraction Study of a-Lead Azide Sir: Earlier studies of lead azide were made by hfiles,l Hughes,2 Pfefferkorn, Azaroff,4 and Saha,s The orthorhombic unit cell has dimensions a = 6,B3, = 11.31, and = 16.25 A, m:ith 12 molecules per unit cell, the density is 4,71 g,/ml, On the basis of the diffraction effects, the space group can be either Pnam or Pna21, The latter was chosen because a threedimensional of the lead parameters+showed t h a t the &factor comes down considerably if the noncentrosymmetric space group is assumed. Azaroff4 determined the lead positions from a two dimensional X-ray analysis, but failed to locate the nitrogen atoms because the lead atoms dominate the intensities of the X-ray data. Neutron data for the Okl reflections were collected by Danner and Kay a t the Brookhaven National Laboratory. Although the neutron scattering lengths for lead and nitrogen are approximately equal, an attempt to obtain a trial structure from a Patterson synthesis of this data failed, due to a considerable overlap of vector peaks. However, a three-dimensional X-ray analysis a t Pennsylvania State University by Saha5 resulted in a trial structure despite the fact that many difficulties were encountered both in collecting and analyzing the data. Using the neutron data, a two-dimensional refinement of these parameters was begun. Initially, a Fourier difference synthesis was used. However, after a few cycles i t became apparent that further refinement using Fourier projections would bebdifficult because of the considerable overlap of lead and nitrogen peaks. Least-squares refinement was not satisfactory (1) F. D. Miles, J . ciaem 506 , 2532 (1931) (2) E W . Hughes. P h I ) Thesis, Cornell University, 19% (3) G Pfefferkorn, Z S a l u v f o v s c h . , 3, 364 (1948). 14) L ,v, Azarofi, z , Krisl,, 362 (,9561, ( 5 ) P Saha, private communication
