4100
Journal of t h e Aniericun Chemical S o c i e t r
Chem. SOC.,97, 3888 (1975). (15)C. M. Criss and M. Salomon in “Physical Chemistry of Organic Solvent Systems”, A. K.Covington and T. Dickinson, Eds.. Plenum Press, New York,
1973.D 260. (16)R . L. BenoitandS. Y . Lam, J. Am. Chem. SOC.,96, 7385 (1974). (17)T. Nakamura, Bull. Chem. SOC.Jpn., 48, 1447 (1975). (18)M. K. Chantooni. Jr., and I. M. Kolthoff, J. Am. Chem. SOC., 89, 1582(1987). The value of K l + for Li+ is uncertain by fl unit. (19)H. B. Flora, 11, Ph.D. Thesis, University of South Carolina, 1971. (20)M. L. Junker, unpublished work, this laboratory. (21)Benoit has reported (ref 16)a value of K1’ = 0.2 for Li’ with H20in Me2S0
/
101:15
/ July 18, 1979
solvent. This value, when multiplied by the concentration of Me2S0 in MezSO solvent. [Me2SO] = 14 M, yields a value of K,, = 2.8for the displacement of Me2S0 by H20; the reciprocal of this yields the value listed in the text. This latter value is an order of magnitude less than those for the same reaction in PC and in AN. The vapor pressures of H20 above 1 M solutions of H20 in these three solvents (also reported in ref 16)indicate that H 2 0 interacts much more strongly with Me2S0 solvent than with either AN or PC. (22)R. H. Erlich and A. I. Popov, J. Am. Chem. SOC.,93, 5620 (1971). (23)V. Gutmann, Electrochim. Acta, 21, 661 (1976):Angew. Chem., Int. Ed. Eflgl., 9, 843 (1970).
(Cross-conjugated dieny1)tricarbonyliron Cations. 2. 4-Methyl Derivatives Benedict R. Bonazza, C. Peter Lillya,” Elaine S. Magyar, and Gary Scholes Contribution from the Department of Chemistry, University of Massachusetts, Amherst, Massachusetts 01003. Receiced August 3, 1978
Abstract: Tricarbonyl(cross-conjugated dienyl)iron cations have been generated from precursor alcohols in strong acid solutions and studied by NMR spectroscopy in the temperature range 0 to -50 “C. The 4-anti-methyl cation (11) coordinates reversibly, but slowly, with fluorosulfonate ion to give the adduct 12. Broadening of ‘ H and I3C N M R signals in the region -40 to -9 “C demonstrates rotation about the Cl-C3 bond of 11. The observations exclude 12 as an intermediate in the rotation process. Relief of steric strain associated with the 4-anti-methyl enhances the driving force for FSO3 coordination and lowers the barrier for C>-C, rotation. The 4-syn-methyl cation (15) does not give a fluorosulfonate adduct and fails to exhibit b M R signal broadening caused by C Z - C ~rotation at temperatures below 0 “C. Neither cation coordinates with carbon monoxide. Lack of evidence of coordinative unsaturation and the substantial barrier to C Z - C ~rotation rule out the $-allyl type structure predicted by the simplest (one interaction) frontier orbital model.
Introduction A simple frontier orbital model which treats bonding in terms of dominant frontier orbital interactions enjoys considerable success when used to predict structures of metal A complexes of ligands which offer several nonequivalent coordination sites.’ T h e title cations were conceived to make a severe test of a simple (one interaction) frontier orbital model. I f spatial overlap for donation of electrons from a n occupied frontier orbital of the Fe(CO)3 fragment to the LUMO of the dienyl cation is maximized ( l ) , the q 3 or “allyl” structure, 2,
n
+
t
I
+ Fe
iC0,
(COi,
1
2 “allyl”
structure
Fe (COi 3
of these orbitals (see 1) finds a suitable partner on the dienyl ligand in this structure. Alternatively, two coordinatively saturated structures, 3 and 4, can be written for which diene-^ and q4-trimethylenemethane-4$5Fe(CO)3 complexes offer structural precedent. T h e characteristics which should distinguish a cation of structure 2 from one of structures 3 or 4 a r e coordinative unsaturation and facile (almost free) ~ rotation about the C Z - C bond. Trimethyl-substituted (cross-conjugated dieny1)tricarbonyliron cations, 5, have been generated in strong acid media, and formation of isomerized quench products (e.g., 6 ) has demonstrated that c2-c3 rotation occurs (Scheme I).2a T h e possibility that rotation could occur during quenching owing to reversible formation of 8 led to a search for less ambiguous
Scheme I. Chemistry of the Tricarbonyl( 1,1,3-trimethyl-crossconjugated dieny1)iron Cation in Liquid Sulfur Dioxide
“diene” structure
Fe
(eo,,
(CO),
4 “
trimethylenemethane” structure
will result. However, this structure can be achieved only a t the expense of several other potentially important frontier orbital interactions and by sacrifice of a filled metal valence shell. The F e ( C 0 ) 3 fragment possesses a degenerate pair of low-lying frontier orbitals occupied by a single electron pair.2 Only one 0002-7863/79/1501-4100$01 .OO/O
7
4
Ho
Fe
(CO), 6
0 1979 American Chemical Society
Lillyn et ai. 1 (Cross-conjugateddieny1)tricarbonylironCations
4101
Table 1. ‘ H N M R Data for Species in Strong Acid/SOz(l) at -65 O C 5a
8
cvidence for structure. Evidence for coordinative unsaturation was mixed. When the temperature of the cation solution was raised above -50 OC in an attempt to study the barrier t o intcrconversion of the proposed species 5 , irreversible formation of the C - H insertion product,6 7, occurred. Ions which lacked methyl substituents a t carbons 1 and 5 , which would be incapable of C - H insertion. were essential, if thorough solution studies were to be carried out. W e report below the preparation and study of the 4-methyl cations. These studies clearly reveal that C,-C3 rotation does occur i n the cations but that it is opposed by a sizable barrier.’
Results and Discussion The 4-anti-Methyl Cation (11) and Its Fluorosulfonate Adduct (12). Sodium borohydride reduction of ketone 98 produced a single epimer isolated after purification in ca. 50% yield. T L C of the crude product produced no evidence for a second epimer. The *-endog relative configuration, 10, is assigned on the expectation that the borohydride reagent will approach the ketone (most stable conformation 9) from the side opposite the large F e ( C 0 ) 3 group, a phenomenon familiar in F e ( C 0 ) 3 complexes of conjugated die none^.^.'^ Treatment of an SO1 solution of 10 with 2 equiv of F S 0 3 H a t -78 “C gave a single species 11 ( N M R a t -65 “ C ) . Warming of this solution to -40 “ C caused formation of a second species. 12 (observed by N M R ) , which continued to form a t the expense of 11 until they were in a ca. 1 : l ratio. Further temperature changes in the range -30 to -65 O C had little effect on this ratio. Addition of a large excess of FSO3H caused complete conversion of 12 to 11. Use of 2 equiv of 1 :1 S b F S / F S 0 3 H mixture in place of FSO3H gave only 11. These data strongly suggest initial production of a cation (11) which slowly comes into equilibrium with its fluorosulfonate adduct. Both in excess F S 0 3 H and in 1:l S b F j / F S 0 3 H the activity of FS03- ion should be much lower than when a stoichiometric amount of FSO3H is used.I1.l2This change will decrease the equilibrium concentration of a fluorosulfonate adduct.I3 Low-temperature hydrolysis or methanolysis of SbFS/ F S 0 3 H / S O z solutions which contained only 11 produced moderate yields of 10 and 13, respectively. Examination of the crude products using T L C and N M R produced no evidence for formation of epimers. T h e relative configuration of 13 is assigned by analogy to that of 10. Stereospecific exo departure of water from protonated 10 would give the 4-anti-methyl cation 11. T h e vicinal coupling J 3 . 4 = 9.5 H z (see below) of 11 is consistent with a 3.4-cis c o n f i g ~ r a t i o n T . ~h~e r e is ample precedent for stereospecific exo departure of the leaving group in generation of organometallic Stereospecific exo attack by the nucleophile during quenching is almost a corollary of exo departure and also has considerable precedent.I6 W e sought to confirm stereospecific exo ionization by preparation of the *-ex0 alcohol 17, which is epimeric to 10.
(CO,
Fe (CO),
17 18 Ionization of 17 with exo departure of the leaving group would produce the 4-syn-methyl ion 15. Several methods of reduction,
IH
shifts, 6 ppmu 1s
12
2.40, t 2.50, d, bd 4.93, s. bd -4.93, ind ( 1 3 0 . d) 3.72, s 4.68, d
5.20, s. bd 2.94, d 5.83. d of d (2.28, d of d ) ? -6.1, md 2.94, d 5.20, s, bd
11
IH
3.44, d of d h 2.79, d of d 4.87, d. bd (2.25. d of d ) C 6.28, d of q d 3.90, s 4.82, d
la Is
3 ?a 4s 5s
5a
‘H-IH coupling constants, H z
JHH
JI~I,
3.0
J1~3
2.2
JlaSa
4.2 9.5 0 7.5
J34s JS~S,
JKH,
-
NaOHIEtOH
Fe,(CO), Et,O
ol -+
CH
Ac,O/pyridine
Fe I (CO),OAC 20
eneacarbonyl produces only the q - e n d o acetate 20 in 57% yield. N o trace of a second epimer could be detected in the crude reaction mixture. The structure of 20 was confirmed by comparison with an authentic sample produced by acetylation of 10 and by saponification to give 10. This route, which starts with mesityl oxide (see Experimental Section), is now the best method for preparation of 10. The strong asymmetric induction apparent in this reaction suggests a direct interaction between the polar acetoxy group and iron a t some stage prior to 7r bonding of iron to the ligand. An attempt to equilibrate 10 and 17 in the presence of aluminum isopropoxide, 2-propano1, and a trace of acetone a t 100-1 I O “ C produced no change. ‘ H and I3C N M R data for 11 and 12arepresented inTables I and 11. Generation of 11 and 12 deuterated a t position 4a (see T a b l e I ) enabled us to assign the H3 and Hj, signals unambiguously. Long-range j J “ W ” coupling,” so characteristic of trimethylenemcthane F e ( C 0 ) 3 complexes, provides sufficient basis for the remaining assignments in 11. The H I , signal was assigned on the basis of its 2.2-Hz “ W ” coupling to H3. T h e H l a doublet of doublets a t 6 3.44 exhibits the same 3.0 2 J l c , , ~splitting \ as H I , . T h e doublet a t 6 4.82 is Hsa, “W” coupled to H while H5,, which cannot exhibit “W” coupling, appears as a singlet a t 6 3.90. The small 2J values (0-3 Hz) are characteristic of 1 ,3-diene-,I8 trimethylenemethane-,5 and conjugated dienyl cation-Fe(CO)3I9 complexes. T h e value of J3.4, = 9.5 H z is in accord with the 3,4-cis c o n f i g ~ r a t i o n l ~ proposed above. Further, the strong deshielding ( 1 .O-1.4 ppm)
Journal of the American Chemical Society
4102
Table 11. I3C NMR Data for Species in Strong Acid/SOz(l) at -70 "C I3C shift, 6 ppma I 'C
11
15h
12
I 2 3 4 5 CH3 CO-
58.5 t (165)" 123.4 s 104.2d (165) 110.8 '?
55.4 t (165) 108.0 s 106.9 '? 121.0d(160)"
77.41(174) 16.8 q
77.3 t (170) 18.2 q ( 1 30)
68.3 t ( 1 70) 126.9 s 136.8d(157) 134.4d(153) 68.3 t ( 1 70) 14.2q(124) 201.1' 20 I .6/
197.7 200.6 205.8
196.2 199.0 203.0
'\
0
I'
"*(' Fe (CO),
S02,-65"C
*
21
0 0 22b
Hoffmann's exposition of substituent effects for pentacoordinate transition metals2* predicts just these structures and provides the basis for predicting a geometry for 12. W e may view 12 as a d* complex with allyl cation (a strong x acceptor through its nonbonding carbonyl (Tacceptor) and fluorosulfonate (0donor) ligands.30 T h e site preference for these ligands will clearly be allyl equatorial, carbonyl equatorial, and fluorosulfonate axial. Steric interaction between X ( O S 0 2 F ) and R (cis-1 -propenyl) should destabilize 22b. Thus, 12 should possess the structure shown in Scheme 11. This structure requires two I3CO NMR signals a t the slow exchange limit in a ca. 2:1 ratio as observed.3' Evidence reported here (Table I l l ) and in the l i t e r a t ~ r e I * ~ *suggests ~ - ~ ~ that two conformers will be seen only in the absence of 2 substitution (R = H ) and multiatom X groups which destabilize conformer 22b. Slow interconversion of 22a and 22b is consistent with the Scheme 11. Chemistry of (Cross-conjugated 4-methyldieny1)tricarbonyliron Cations
Fe (CO),
Fe
(eo),
9
10 2-3 equiv F S 0 , H SO,. -65 'C
If
-9
:$temp
;'-.
I
F0,SO-Fe-CO
(CO),
(CO)
13
11
-4 0r'\%
r 0
12
a
+ $e
Fe
(CO),
Fe--.OSOLF (CO),
pound) a r e virtually identical with those of corresponding carbons of 12. T h e correspondence between the analogous
18, 1979
'j,
0 22a
2-3 equiv F S 0 , H
equiv F S 0 , H
/ July
4 c c
d c i .
[
) ! , 1-1.5
101:15
chloride adduct (Table 111) and 12 is almost as close. Additional spectra reported in T a b l e I l l confirm that these I3C shifts are typical for g3-allyl-Fe(CO)3X compounds. T h e ' H and I3C spectra of 12 provide no evidence for existence of two slowly interconverted conformational isomers such as those evident in the IH spectra of g3-allyl-Fe(C0)3X (X = Cl,'5 Br,2s 125,26)and the I3C spectrum of the iodide (Table I l l ) . A crystal structure of the iodide2' has revealed what is essentially a trigonal bipyramid with the allyl and two C O ligands occupying equatorial positions (22a, R = H; X = I). The conformational isomer has been proposed as 22b.25Rossi and
a Relative to M q S i as calibrated by internal CDC13 at 76.9 ppm. A positive shift is a downfield (from Me4Si) shift. Coupling constants &8 H z owing to a poor signal/noise ratio in the coupled spectrum. 'Jl3C. H (Hz)in parentheses, &4 Hz. Assignment confirmed by deuterium labeling.
