1598
VOl.
W. R. RICCLELLAN
s3
proximations can be very useful, even though the charged site inside a polyion had major contribumathematical approximation of localizing the tions from both the near neighbor and the rest of particles is not a t all in accord with the real, very the polyion. This suggests that we take the mobile molecular system. Of course, distinct multiply charged linear model and imbed it inside from the binding question, the use of Ising methods a polyion. Such a model would correspond to to compute the electrostatic energy and other choosing an electrostatic Kuhn element and surproperties remains a valuable method of accounting rounding this element with a more or less spherical for the discrete nature of the charges. The view charge distribution. Qualitatively, we expect the proposed satisfies all the available criteria by which potential due to the rest of the ion to be approxiion binding has been introduced: The high con- mately constant throughout the polymer domain.21 centration in the “bound monolayer” insures very Thus, we would anticipate that the activity coeffective charge-charge shielding; the division in efficient of the counterion would be further determs of the ionic distribution should, as in the pressed (over that calculated herein for the small case of simple electrolytes, be relatively field in- compounds) and that its value would be indedependent a t low fieldsls but yield a large Wien pendent of molecular weight. While these qualieffect (as has been observedlg); since we do not tative features are easily seen to follow from the appeal to discrete ion-pair formation, it is clear physical nature of the model, the mathematical that counterion binding and low counterion activity formulation of this problem is fraught with difcoefficients are but two descriptions of one phenom- ficulties, and inore details will be presented in a enon20; we suggest that lattice gas calculations, subsequent paper. 2 2 IX. Acknowledgments.--We wish to thank the which are in semiquantitative agreement with experiment, are useful mathematical approxima- United States Public Health Service, National Institions and that such a model should not be in- tutes of Health, the National Science Foundation and the Air Force Research and Development Comterpreted blindly in terms of pairs. We now turn briefly to the possible extension of mand for support of this work and Mr. George the models proposed herein to the case of coiled Thomson for assistance in the preparation of compolyions. The results of Nagasawa and Rice6 pounds. Our task would have been immeasurably indicated that the mean electrostatic field a t a more difficult without the advice of Dr. Alitsuru Nagasawa, whom we wish to thank for his aid. (18) F. T. Wall, H. Terayama and S. Techakumpuch, J . Polymer Sci., 20, 477 (1956). (19) F. E. Bailey, A. Patterson and R. &I. Fuoss, J. A m . Chem. Soc., 74, 1845 (1952). (20) hf. Nagasawa, ibid., 83, 1026 (1961).
[CONTRIBUTION No. 642 FROM
THE
(21) J. Herrnans and J. T. G. Overheek, Rec. 2rac. chi?%, 67, 761 (1949). (22) See, however, the related earlier calculation: F. E. Harris and S. A. Rice, J. Chenz. P k y s . , 2 6 , 955 (1956).
CENTRAL RESEARCH DEPARTMENT, EXPERIMENTAL STATION, E. I. DU PONT DE XEMOURS AND COMPANY, INC., ~ ~ ’ I L M I X G T O NDELAWARE] ,
Perfluoroalkyl and Perfluoroacyl Metal Carbonyls BY W. R . MCCLELLAN RECEIVED OCTOBER19, 1960 and Rr-Mn( CO)5, respecPerfluoroacyl and perfluoroalkyl derivatives of manganese pentacarbonyl, R-CO-Mn( tively, where Rr = CFI, n-C3F7 and iso-C3F7, and the perfluoromethyl, -ethyl and -72-propyl derivatives of cobalt tetracarbonyl have been prepared by reaction of the appropriate perfluoroacyl halide with the lithium derivative of manganese or cobalt carbonyl hydride. The perfluoroalkylcobalt compounds have outstanding stability compared t o their hydrocarbon analogs. The preparation of perfluoropropenylmanganese pentacarbonyl is also described.
Closson, Kozikowski and Coffieldl have reported the preparation of alkyl- and acylmanganese pentacarbonyls by reaction of alkyl and acyl halides with sodium manganese pentacarbonyl. Certain of the acyl compounds lose the keto carbonyl on heating to give the corresponding alkyl derivatives and, in most cases, this reaction is reversed by carbon monoxide gas under high pressure. Because of the positive character of the halogen atom, reaction of perfluoroalkyl halides with sodium manganese pentacarbonyl does not provide a route to pertluoroalkylmanganese pentacarbonyls Na+Mn(CO)s-
+ i s 0 - C ~ F 7 - 1-+ ~ IMn( CO), + CF~CF-CFZ + NaF
The preparation of trifluoromethylmanganese pentacarbonyl, however, by thermal decarbonyla(1) R D. Closson, J. Kozikowski and T H. Coffield,J O r g . Chem , 22, 598 (1957).
tion of trifluoroacetylmanganese pentacarbonyl has been reported.2
+ A411(cO)j- +CF~-CO-Lh(CO)s + CI90-1 10 CF3-COMn(CO): ----+ CF3-hln(CO)s + CO
CFa-CO-Cl
O
We have used this route to prepare several other perfluoroacyl and perfluoroalkylmanganese pentacarbonyl derivatives described in Table I. The only previously described cobalt tetracarbonyls have been the methyl,3 , 4 ethyl4 and benzyl derivatives which, because of their high degree of instability, have been prepared and isolated a t subzero temperatures. We have prepared several (2) T. H. Coffield, J. Kozikowski and R. D. Closson. International Conference on Coiirdination Chemistry (London, 1939), p. 126 (Abstract), Burlington House, London, 1959. (3) W. Hieber, 2. Nattrvjororsch., 13b, 192 (1958). ( 4 ) D. S. Breslow and R. F. Heck, Chent. upid I n d . ( L o f t d o n ) ,467 (1900).
April 5 , 1961
1599
PERFLUOROALKYL AND -ACYL METAL CARBONYLS 3
T 0
?.
F? F?
3
n
d
vi
-
v
3
1
h
01
-_
?
vi v m L m a
0"
I
I
I
h
h
vi
3
v
0
01 m
?
0
Y
3 .-
cl
2 Y ri
0
h
2
v
B
i
h
5
v
,r.-. v h
0
0 v
$
0
0
kV
m y. h
3 v
b
?? N.??
8
Y
$
0
0
L
h
k
Y 3
n
was stirred and held a t a temperature of 25". After standing overnight at O', the solvent was removed at room tcniperature under reduced pressure. By sublimation of the residue at 50-55' (1.5 mm.), 13.3 g. of I was obtained as pale yellow crystals, 1n.p. 55-56'. Bands found in the infrared for a T H F solution of I were a t 8.12, 8.57, 5.83, 11.69, 13.95 and 14.57 p . I n addition to these, assignments for earhoiiyl bands are listed in Table I . Trifluoromethylmanganese Pentacarbonyl (II).-On heating 10 g. (0.034 mole) of I a t 110' a t atmospheric pressure, 720 ml. (theory is 770 ml.) of carbon monoxide was eollected in 100 minutes. By sublimation of the residue a t 70" (20 mind), 5.3 g. of I1 was obtained as white crystals, n1.p. X2-83 . Bands found in the infrared (solid stat? spectrum-KBr pellet) wcre a t 9.47, 10.03 and 11.39 p ; the carbonyl bands are listed in Table I . Perfluoro-n-propylcobalt Tetracarbonyl (III).-Over a period of 15 minutes, 10.7 g. (0.046 mole) of perfluoro-nbutyryl chloride was adcird dropwise t o 80 ml. of a 0.2% i v solution of lithium cobalt tetracarbonyl in tetrahytirofuran under a n atiriosphere of nitrogen. Tlie reaetioii mixture CI~~=:C1~-CF2Cl hIn(C0);- --+ was stirred and lield a t a teinperature of -25 t o -20" [C1CF*--CI;-CF2--nIn( CO),] 4 during the dropwise addition of the acyl chloride. T h e .3 solvent ivns removed under reduced pressure at a teinpcr:tcr~2=cI;-cF,!hill~cO)j c i - --3 turc oF -15 to -20" with the use of a "Kiiico" cwnoratnr. cr~,--cI~---c~~--MIl( C O ) ~ In this temperature range, the evolution of gas froin tlie reaction mixture was slow enough that no trouble W:IS cnThe perfluoroallylriiarig~nese peiitacarbonyl re- countered with foaming during removal of solvent. The arranges under surprisingly mild conditions to the final residue was warmed to room temperature atid hehl under a blanket of nitrogen until no further carbon monoxide more stable isomer, perfluoropropenylmanganese was evolved. The residual product was extracted with penpentacarbonyl. tane. After removal of solvent from the combined pentane An improved method of preparing manganese extracts, 6.2 g. of I11 was distilled as a light amber liquid, pentacarbonyl and cobalt tetracarbonyl anions in b.p. 44' (16 mmj. Bands found in the infrared for the a t 7.58, 8.15, 8.35, 8.54, 9.13, 9.62, tetrahydrofuran solution in high yield by reaction pure liquid 111 were 12.34 and 13.80 p ; the carbonyl brrnds are listed in Table of lithium wire with the corresponding metal car- I. bonyls is described in the Experimental section Perfluoropropenylmanganese Pentacarbonyl (VIII ) .At -5O, 4.8 g. (0,029 mole) of perfluoroallyl chloride was 2Li h\In~(CO)l~ --+ ZLibln(CO)s condensed into a flask under nitrogen, and then 53 nil. of a 2Li 4- Co2(CO)8 -+ 2LiCo(CO)a 0.44 N solution of lithium manganese pentacarbonyl in tetrahydrofuran mas added with stirring. The reaction mixture was allowed to warm to room temperature. After Experimental Lithium Manganese Pentacarbony1.--A solution of 25 g. holding at room temperature for 2 hr., it was stored overnight a t 0'. The reaction mixture was filtered t o remove (0.OG4 mole) of dimanganese decacarbonyl in 250 ml. of the LiCl precipitate, and the solvent was then removed by anhydrous tetrahydrofuran containing 1.4 g. (0.2 mole) of distillation under reduced pressure at 25". The dark short lengths of lithium wire was stirred with a high speed was heated with an oil bath at 55' (1 rnin.). b'hen stirrer for 3 hr. During this time, the reaction mixture residue 0.35 g. of liquid had been collected, a white solid suddenly was blanketed with argon and a cooling bath was used t o formed the walls of the condenser and eolurnn of the disA tillation on prevent the temperature from rising above 45-50'. equipment. The 0.35 g. of liquid in the receiver stirrer with loose wire ends was used to provide cutting changed t o a mushy, low-melting solid. The infrared specaction on the lithium. The deep green solution of lithium indicated that this mushy product was roughly OO(i; manganese pentacarbonyl was siphoned from the excess trum of VI11 and 40Yc of a terminal CFx olefin product. I t w:ts lithium. This solution was assumed t o have a lithium assumed that this latter w a ~ perfluoroallylnianganese pentamanganese pentacarbonyl normality of 0.51, carbonyl. Ether was added to dissolve the residual product Lithium Cobalt Tetracarbony1.-A solution of 34.8 g. in the distillation flask. Charcoal was added t o thc ether (0.10 mole) of dicobalt octacarbonyl in 350 inl. of anhy- solution and it was then filtercd. On cooling the filtrate to drous tetrahydrofuran containing 2.65 g. (0.38 mole) of -15", 2.15 g. of crystals with a light brown tinge, m.p. 69short lengths of lithium wire was held in the temperature 73", were obtained. By sublimation a t 60" ( 3 rnm).;2.0 g. range of -10 to -20' while stirring with a high speed VI11 wits obtained as white crystals, m.p. 76-76.5 . stirrer as described above. The initial vigorous reaction of Bands founcl in the infrared were at 6.09 p (for internal subsided in 20 t o 30 minutes, and the solution then was C=C), 7.56, 8.42, 8.90, 9.67 and 12.35 p ; carbonyl bands warmed t o room temperature while continuing the stirring. are listed in Table I. There was no 5.64 p band for the The excess lithium pieces were removed and the solution CF2=CF function. The I719 magnetic resonance spectrum of was considered t o have a lithium cobalt tetracarbonyl 50% solution of VIII in tetrahydrofuran showed three difnormality of 0.58. ferent kinds of fluorine in an intensity ratio of 3: 1: 1. The Typical conditions for preparing the perfluoro derivatives CF3was a doublet split into doublets. The individual pxdis of manganese and cobalt carbonyls are describecl below. located a t -11.7, -11.4, -11.1 alid -1O.S p.p.m. Analyses, yields and other details are given in Table I . were (trifluoroacctie acid refereiice). Each of the olefinic CF's Trifluoroacetylmanganese Pentacarbonyl (I).-Over a was a doublet split into quadruplets. The chemical shifts pwiod of 1 hr., 7.9 g. (0.068 mole) of trifluoroaeetyl fluoride gas was bubbled into 120 ml. of a 0.51 N solution of lithium were +16.8 p.p.m. for one C F and 4-89.4 p.p.m. for the manganese pentacarbonyl in tetrahydrofuran under a other. The Jvalue for the doublct separation was 135C.P.S. nitrogen blanket. During this time the reaction mixture Acknowledgments.-I wish to acknowledge help-
perfluoroalkylcobalt tetracarbonyls and, in contrast to the alkyl derivatives, the perfluoro compounds listed in Table I have excellent thermal and oxidative stability; the trifluoroinethyl and pentafluoroethyl derivatives can be distilled (b.p. 91 and 1l o o , respectively) a t atmospheric pressure with no decomposition. The intermediate acyl compounds, which decarbonylate a t temperatures below Oo, were not isolated. The behavior of the allylic halide, perfluoroallyl chloride, is different from that of the saturated perfluoro-isopropyl iodide in its reaction with lithium manganese pcntacarbonyl. I t appears that the Mn(C0)a anion adds to the >CF2 of l,l-difluoro-3halopropenes in a nianner analogous to that observed for fluoride anion5
+
I ,
+
+
( 5 ) J , H. Fried and (19SY).
W. T. Miller, Jr., J . A m . C h e m . Soc., 81, 2078
ful suggestions made in connection with this work by Dr. E. L. Muetterties and Dr. B. W. Howk.
