Vol. 85
OLAH,TOLGYESI, KUHN,MOFFATT,BASTIENAND BAKER
1328
temperatures (3z0.5’) indicated in Table I for 0.5-hour periods. After cooling, a known volume of chloroform was added and the intensity of the carboxyl carbonyl peak ( l i 3 6 ern.-' in V and 1750 cnl,-l in in the infrared was measured, The amount of unreacted starting material was determined using a calibration curve prepared from known concentrations of 1. and TI.
[CONTRIBUTIOS S O .
Acknowledgment.-We wish t o thank Professor H. and professorH, 0, H~~~~for helpful discussions and Professor House for informing us of his results Prior t o Publication. We also thank Air. Gerald Caple for determining the n.m.r. spectra.
w. Walborsky
87, EXPLORATORY RESEARCH LABORATORY, DOXY
CHEMICAL O F CANADA,
LTD.,SARNIA,
OTTARIO, C A N . ]
Stable Carbonium Ions. IV. la Secondary and Tertiary Alkyl and Aralkyl Oxocarbonium Hexafluoroantimonates. Formation and Identification of the Trimethylearbonium Ion by Decarbonylation of the tevt-Butyl Oxocarbonium Ion BI’GEORGEh.OLAH,MS~ILLIAMS.TOLGYESI, STEPHENJ. KUHN,MARYANNE E. MOFFATT,IVANJ. BASTIEN.4ND EDWARD B. BAKERib RECEIVED OCTOBER11, 1962 1 : 1-Addition compounds of di- and trialkyl( aryl-)-acetyl fluorides with antimony pentafluoride mere prepared and investigated. On the basis of infrared spectra i t was found t h a t in the crystalline state t h e ionic oxocarbonium structure (RCO+SbFc-) prevails, with minor amounts of donor : acceptor complexes also present. Diphenylacetyl fluoride : antimony pentafluoride is practically exclusively the oxygen coordinated coordination complex. Triphenylacetyl fluoride gives only a very unstable coordination complex decarbonylating t o the stable, ionic triphenylcarbonium salt. I n solution there is evidence for an increased amount of coordination complex being present in addition to t h e ionic oxocarbonium salts, according t o high resolution nuclear magnetic proton and fluorine resonance spectra. Decarbonylation of tert-butyl oxocarbonium hexafluoroantimonate could be followed spectroscopically and t h e trimethylcarbonium ion was identified. The isolated oxocarboniuni salts are reactive acylating agents. dlkylation takes place as well in the case of tertiary and t o some degree in secondary alkyl oxocarboniurn salts due t o prior decarbonylation.
Introduction As a continuation of previous workla,r on n-alkyl and aryl oxocarbonium salts it was considered that it would be of interest t o extend these investigations to secondary and tertiary alkyl (aralkyl) oxocarbonium systems. No previous data were available on any of these ions as stable entities. Results and Discussion Preparation of the new oxocarbonium complexes was achieved according to the previously reported “fluoride method.”’ RCOF
+ SbFj
KCO+SbFB- ( R C O F , S b F j )
Some of the required acyl fluorides were available from previous ~ o r kand ~ ,the ~ preparation of others will be reported in a future p ~ b l i c a t i o n . ~ The crystalline complexes were found by analytical data to be 1 : 1 addition compounds of high purity (see Experimental part). Their melting points (in sealed capillary tubes) are listed in Table I and are compared with those of the corresponding primary alkyl (aryl-)acetyl fluoride : antimony pentafluoride complexes. TABLE I MELTISCI’OISTS OF SECOSDARY AND TERTIARY ALKYL(;\RYL-).ICETYL FLUORIDE : ASTIMOST PESTAFLUORIDE COMPLEXES l1.p
CH,CH?COF,SbF; (CHs)?CHCOF.SbF:, (CH,hCCOF.SbF, CH,CHrCH,COF.SbF:, (CH,CH?)?CHCOF.SbFi (CHpCH?):ICCOF,SbF, CsHiCH?COF’SbF, (CsHs)aCHCOF.SbF, ( C,H;):jCCOF,SbF, ~
~
~~
oc.
110- 112 95 20 dec. 58 15 dec. - 5 t o 0 dec. 132-135 68- 69 Dec. to (CRH,)3C‘SbF6 , 211O
III.~.
~~
( 1 ) ( a ) Part 111, Revue d e ( ‘ k t m t e , C I ) . S e n i t z e s c u s (jOth B i r t h d a y Issue, IOii’L. ( b ) Physical Research Laboratory, T h e !.low Chemical Co., Midland,
llich. ( 2 ) G A . Olah, S . J . K u h n , U’.S. Tolgyesi and E. B. Baker, J . A m . Chem. .Sot., 84, 2733 (1962) f:i) G . Olah. S . K u h n and S. B e k e , Be?,.,89, 802 (19%). ( 1 ) (;. A Olah a n d S J , K u h n , J . Ovg. Chem , 26, 237 (1961). C.; A . O l a h , S. J K u h n and W. S. Tolgyesi, unpublished results.
Spectroscopic Investigations. Infrared Investigations.-A Perkin-Elmer model 421 grating spectrophotometer was used to record the spectra. Emulsions of the solids in mineral oil (Nujol) and a fluorinated hydrocarbon (Fluorolube, S30, Minnesota Mining and Manufacturing Co.) were pressed between silver chloride or Irtran plates, all operations being carried out in a drybox, as the compounds were extremely sensitive to moisture. No etching of these plates was observed, in marked contrast to the observation when sodium chloride plates were used. Even barium fluoride plates were slightly etched. The main characteristic dat.a obtained are summarized in Table 11. As illustrative examples, the spectra of (CH3)&HCO+SbF6- and (CH3)3CCO+SbF6- (as mixed mulls in Nujol-fluorolube) are shown in full in Fig. 1 and 2 . There has been general agreement6,’that the spectra of acetyl halide : Lewis acid metal halide complexes contain a strong band a t 2300 cm.-‘ due to the methyl oxocarbonium ion (CH3CO+)formed in the process. CHICOX
+ MX,,
CHaCO+MXn+1-
Similarly, aromatic acyl halide (benzoyl, etc.) complexes have given spectra containing a strong band at 2200 cm. - l , indicating the presence of phenyl oxocarbonium ion. The lower carbonyl frequency of the phenyl OXOcarbonium ion than that of the methyl oxocarbonium ion was attributed t o conjugation of the CO group with the ring (an effect often observed in infrared spectroscopy). Similarly, conjugation of the alkyl groups with the CO double bond caused a shift of the C=o stretching frequency to lower values in the secondary and tertiary alkyl substituted oxocarbonium ions. In addition t o the ionic oxocarbonium salts there are present also, but to a lesser extent, the polarized OXYgen donor coordination compounds, as shown by the presence of the shifted C=O stretching frequencies. Table I1 shows the shift of the carbonyl stretching frequency of the donor-acceptor complexes and, as may be seen, the shift increases with the branching of the (ti) B. P. Susz a n d J. J . Wuhrmann, Hela. C h i m . Acta, 40, 971 (1957). (7) I). Cook, C a n . J . C h e n i . . 37, 18 (1959).
May 3, 1963
1329
ALKYL OXOCARBONIUM HEXAFLUOROAIVTIRlONATES W A V E L E N G T H (MICRONS)
3
4
5
6
7
8
9
10
12
15
4
5
6
7
8
9
IO
12
IS
1
I
I
,
I
3
ICH&CHCOTsb F i
3
25 I00
( C H3,C
-
*
E,
I
CO'S b F i
1
80J
Li i
g
v
60-
aJ 0, A
.--
i Y e
40-
a
r
3
$
20-
0-
3500
4000
3000
2500
alkyl group in the acyl fluorides. The sizable carbonyl shifts also indicate the strong Lewis acid nature of antimonv pentafluoride. Of particular interest is the tert-butyl oxocarbonium hexafluoroantimonate complex, isolated from pivalyl
RCOF,SbFs
CHiCOF CH.jCO-ShF6CHiCHZCOF CH:jCHzCO-SbFs(CH~ZCHCOF' (
CH:)?CHCO'ShFB
(CHa)aCCOF (CHa)iCCO+ShFeCHaCH?CH?COF C H B C H ~ C H ~'SbFeCO I CH1CH?)>CHCOF
iCHICH,),CHCO 'ShFA f CH.ICHZ)~CCOI'
(CH::CH?),CCOF.SI,~'., C,H,CH,COI; CRH+CH?CO 'ShF,
"c= n
THE
CORRESPOSDISC ;\CYL "CEO
F L U O R I D E : SbF:, COMPLEXES "C-F
1848,s 826~809s 1621~ P294vs 1554~ 184517s 1068s 1610rn 229ovs 1840vs 1065, 1585111 2270vs 1823vs 1O6()v5 157ovs 2260vs 1824~s 1073s l6lO1n "283vs 1545, 1837~s 1070s 1 A78m 1832vs 1065S Dec. to (CH.;CH?hCFa i i d I CHXH?),C=CHCHs, 1813~s lO78vs 1699m "279VS
l516rn (CsH:)?CHCOF ( CsHi )yCHCOF,SbFi
f CsHa)aCCOI' ( Cti H >),tC CO F,Sh I'i
IO
1000
fluoride and antimony pentafluoride. The observation of the infrared spectrum a t room temperature as a Fluorolube-Nu jol mull (Fig. 2 ) is always accompanied by some decomposition. In preparative runs carbon monoxide was isolated from the decomposition product,
TABLE I1 ISFRARED FUNDAMESTAL S T R E T C H I N G FREQUENCIES O F ACYL FLUORIDES A6D RCOF
I500
2000
1847vs 15 7 8 ~ 5 1831vs Deconip. to (C6H6)aC.+SbF6-
1052s
"SbF8-
6655
6505 650s 660s
655s
661s resp.
650s
OLAH,TOLGYESI, KUHN,MOFFATT,BASTIENAND BAKER
1:MO
PROPIONYL FLUORIDE
J,.,.z.~~~, (-80.)
PIVALYL F L U O R I D E
L
NEAT
so.
L
NEAT
i" I
J H . , . O l cpr lCH,),CCO+ S b F'. 6 (PPYIFROM T M S
sq(-eo*c)
1
i
(I
I
I '
I A,
-5
-4
-3
-2
-I
b(pRc)FROMTMS
Fig. 3.-H' resonance spcctra of niethylacetyl fluorides and their antimony pentafluoride complexes at 60 Mc.
E1,CCOF
NEAT
Et,CCOF.SbFa
25. SO, -40.
SO,
S I P P Y I FROM T M S
Fig. I.-"
$
-9
-3
-2
I'&
[LI'
r
*I'
-!
-0
I
resonance spectra of ethylacetyl fluorides and their antirnoriy pentafluoride coniplexes at 60 Mc.
together with polyisobutylene and HF. Thus it is believed that partial decarbonylation took place also during the infrared investigations. The strong, broad peak observed in the infrared spectra a t 2835 cm.-' is unique in that it is not present in the spectra of other alkyl oxocarbonium complexes, pivalyl derivatives, isobutylene or polyisobutylenes. The CH3 stretching frequency in isopropyl oxocarbonium hexafluoroantimonate (Fig. 1) is a t 2910 ern.-', -($1cm.-' higher than t h a t observed for the same vibration in the case of the trrt-butyl derivative. I t therefore is suggested that the peak a t 28.73 cm.-I in the spectrum of the latter represents the methyl stretching frequency in the formed and relatively stable trimethylcarbonium ion. (CHa)aCCO+SbFs-
VOl. 83
Jr C O + (CH:I).jC+SbFs-
The corresponding asymmetrical CH3-C deformation vibration is a t 1333 cm.-', the symmetrical deformation a t 12%)cm.-'. The tert-butyl oxocarboniuni ion consequently can be considered as a precursor to the trimethylcarbonium ioii. Decarbonylation of the former represents a feasible way t o generate the latter under moderate, acid-free conditions. The decarbonylation of pivalyl chloride under Friedel-Crafts conditions has been observed by Rothsteiqn and by Balaban and N e n i t z e s ~ u . ~ (8) E. Kothstein and R. W. Saville, J . Cheiir. SOL.,I Y X j , IYJO, 19.54, l Y X , I9Iil (lY4Y); E. Rothstein, ibid., I!X>Y (19.51): E. Rothstein, el n i . , i b i d , , 1,558, 4 5 l i l ( l < I X j ) ; 381 (1956). ( 9 ) A 1'.R a l a h a n and C . I ) . Senitzescu, A u u . , 615, (jci ( l U > < l ) ,
-8
-7
8
-5
-9
-3
-2
I
Fig. 5.-Resonance spectra of phenylacetyl fluorides and their antimony pentafluoride complexes at 60 Mc.
High resolution nuclear magnetic proton resonance investigations in sulfur dioxide solutions of the tert-butyl oxocarbonium ion, to be discussed later, also show the formation of a second, highly ionic and electron-deficient species, which further substantiates the present observation. Even more predominant is the decarbonylation of the triethylacetyl fluoride : antimcny pentafluoride and triphenylacetyl fluoride : antimony pentafluoride complexes. Although a t low temperature 1: 1 addition compounds are formed, they lose CO (and H F ) so easily t h a t the spectra of the complexes obtained show no carbonyl absorption and were identical with those of 3-ethyl-2-pentene ((C2H5)*C=CHCH3) and triphenylcarbonium hexafluoroantimonate ((CeH5)3C+SbFe-), respectively. Spectroscopic data obtained on crystalline addition compounds, however, cannot be extrapolated to give information on the structure of these compounds in solutions. Ultraviolet Investigations.-No systematic investigation of the ultraviolet absorption spectra of oxocarbonium complexes has yet been undertaken. HOWever, in view of the particular interest in the possibility of generating trimethylcarbonium ion from the tertbutyl oxocarbonium complex, i t was felt useful to investigate the ultraviolet absorption of pivalyl halides under conditions allowing ionization and thus subsequent decarbonylation. Rosenbaum and SymonsIohave reported the surprising stability of the trimethylcarbonium ion in concentrated sulfuric acid solutions. Solutions of both tert-butyl alcohol and isobutene gave in sulfuric acid a single measurable ultraviolet band having Xmax 293 i 2 m p ( E 6.4 X lo3) with a half-height width of 490 em.-'. As an alternate Dossibilitv for the absorbing ion, Syrnons considered ihe oxidized buteny! cation CHz= C (CHa)-C Hs +. It has been found now that both pivalyl chloride and fluoride give in 10070 sulfuric acid solutions a single ultraviolet band a t 2$2 =t 1 mp, The same maximum was observed in antimony pentafluoride Folution of pivalyl fluoride or in the same solution of tert-butyl fluoride. The observed extinctiou coefficierlts were, however, small (-30(1), Deiiol' recently questioiled the assignment of the 293 mp maximum in the sulfuric acid ( I O ) J . Rosenbaurn a n d 51. C . K . S y r n o n s , P r f . r . ('hurri .Sot., 9 2 (19.59); . I l o l P f i y s . , 3 , 20; ( l ! l I ~ U ) , ( 1 1 ) X , C. Ueno, Abstr. Paixrh. 7 7 Q , I 4 2 n d A C.S. Nati. hleetin6, Atlantic City, S . J . , September, I Y ( i 2 ; N . C. Ileno, H. G. Richey, J r . , J. S. Liu, J. U . Hodge, J . J . Hauqer and 11. J . Wisotsky, J . A m Chrtn. SOC., 84, P O l t i (l!lti2),
1Iay 3, 19.33
133I
ALKYLOXOCARBONIUM HEXAFLUOROANTIMONATES
Fig. 6,-( C H D ) ~ CinF SbFs at 60 Mc.
solution of butyl alcohol to be due to the trimethylcarbonium ion or the isobutenyl cation and suggested, based on n.m.r. investigations, that it is entirely that of the heptamethylcyclopentenyl cation. It is difficult from our observation of the ultraviolet spectrum to come a t this time to final conclusions. It is possible but not probable to have the oxidized allylic ion in small concentration in our SbF5 system, not detectable by n.m.r. investigations, but sufficient care was exercised to exclude oxygen from the system. Although our observed extinction coefficient is only 1/10 of that reported by Symons'? we suggest that the 293 mF absorption with the observed small extinction coefficient could be assigned to the electron transfer of the trimethylcarbonium ion involving quasi n-electrons and the vacant p-orbital of the positive carbon. Recovery experiments and other investigations of the tert-butyl fluoride : SbF5 and pivalyl fluoride : SbFj systems (to be reported later in detail) gave no evidence of Deno's heptamethylcyclopentenyl ion. Nuclear Magnetic Resonance Investigations.-To obtain further information on the structure of the acyl fluoride: antimony pentafluoride 1: 1 addition complexes in solution, high resolution nuclear magnetic proton and fluorine resonance spectroscopic investigations were carried out. X modification of the high resolution n.m.r. spectrograph of Baker and BurdI3 was used. Part of the proton resonance spectra (when working a t slightly elevated temperature, -40') were obtained on a Varian A G O spectrograph. Liquid sulfur dioxide (also anhydrous hydrogen fluoride) was used as solvent, the samples being sealed in quartz probe tubes. The H1 and F19 spectra obtained are shown in partially schematic form in Fig. 3, 4, 5 , 6 and 7. The corresponding acyl fluorides are shown for reference in the same spectrum, and the FI9-H1 splitting (to be discussed later) is removed, for compactness. Proton Spectra.-The proton spectra of the complexes obtained a t GO AIc. are displayed in Fig. 3 , 4, 5 , and 6. The spectra are given in p.p.m. relative to (CH3)dSi (TLIS) as standard. The acyl fluorides exhibit H1-FI9 coupling of varying magnitude. None of the complexes exhibits HI-FI9 coupling, which would of course not be expected in the ionic forms, but this can hardly be used as conclusive ( 1 2 ) Prof. 11.R. C. Symons in a personal communication kindly informed t h a t his reported extinction coefficient value should be modified, being not more t h a n 2000. (13) E. B. Baker a n d L. W.Burd, Rep. Sci. I w k , 28, 313 (1957). Us
H S b h YO (SOL".) SbF8 (GUTOWSKY.H O P W A N ) CH,CH,COF I CH,CH,CO'Sb< (CH,), CHCOF 1 (CH,), CHCO'SbF; (CH,),CCOF 1 (CH,),CCO'SbF; -80. CH,CH, CH,COF C H, C He CH,CO'S b F c (CH,CH,), CHCOF (CH,CH,), CHCOF,SbFs (CH,CH& CCOF I (CH,C H , ) , C C O F ~ S b F ~ ~ ~ o ~ L Y P CeH8 CH, COF C,H,CH,COF SbFs (CeHa)zCHCOF I (CeH8),CHCOF SbEs (CeHo), CCOF I ( Ce HE), C+ S b F i S ( P P U ) A T ~ ~ ~ M - I -100 ~O
A I
I A 1
1
1 1
1 I 1
I -50
CF, COOH
t50
*loo
F i g . 7.-F19 resonance spectra of acyl fluoride: SbFj complexes in SO? solution.
evidence of ionic dissociation, but only as supporting evidence. Fast exchange in a highly polarized donor : acceptor complex could equally well result in the absence of observable H1-F19 coupling. The proton spectra of solutions of the acyl fluoride: SbF6 complexes generally show two species, labeled I and 1', with varying chemical shifts to less shielding. Species I, which is the least shielded, is thought to be the ion RCOf present, however, as a solvated ion pair (cryoscopic measurements show very little, if any, ion separation). Species I' is suggested to be in all probability the polarized coordination complex 6+
RCO * I
6-
. . SbFj
Due to the possible fast fluorine exchange, the H1-F19 coupling is not observed in the complex and substantial polarization can account for the considerable shift, which is nonetheless smaller than is observed in the case of the truly ionic species I. The proton resonance spectra of propionyl fluoride and its complex with antimony pentafluoride were discussed previously.2 The proton spectrum of isobutyryl fluoride shows the characteristic separate multiplets for the CH3 and C H groups. I n the spectrum of the SbF5 complex the
OLAH, TOLGYESI, KUHN,hIOFFATT, BASTIENA N D
1382
methyl doublet is a t -1.71 p.p.m. and the C H septuplet a t -4.33 p.p.m., indicating considerably less shielding of the protons in the complex than in isobutyryl fluoride, in accordance with the ionic isopropyl oxocarbonium structure of the complex. The proton spectrum of pivalyl fluoride shows a methyl line a t - 1.27 p.p.m., with a small ( J H F= 0.79 c.P.s.) coupling by the fluorine. The complex of pivalyl fluoride with SbF5 shows three methyl lines a t --SOo, all a t lower shielding than with pivalyl fluoride alone. The least shielded is a t -3..95 p.p.m., the second a t -2.02 p.p.m. and the third a t -1.43 p.p.m. Xs the positively charged carbon is twice removed from the methyl protons in the tert-butyl oxocarbonium ion it cannot be expected that the shift of -3.93 p.p.m.
[ -.I;
for the least shielded line corresponds to this ion. In other cases in the present work chemical shifts of protons in methyl groups of secondary and tertiary oxocarbonium ions, in which the electron-deficient carbon is two carbons away, are found to be much smaller-of the order of - 0 . 3 p.p.m. from the parent compound. Even the primary methyl oxocarbonium ion [CH3-C=O]+, where the positive charge is on the neighboring carbon (see subsequent discussion of the C'3-spectrum of this ion) shows a shift of only -1.90 p.p.m. I t therefore is suggested that the chemical shift a t - 3.93 p.p.m. corresponds to a considerably less shielded electron deficient species, namely the trimethylcarbonium ion CH.! CHa-CS
I
CH.,
which is formed from the oxocarbonium ion by the loss
co.
TABLE 111 PROTOX
LfAGNETIC KESOSANCE I N V E S T I G A T I O S O F THE METHYLCARBOXIU3I Solvent
(CHa)?CCOF (CH:OaCCOF SbF; ICHa)aCF (CH3)aCF 4- SbF,
+
TRI-
10s
5 ( p p , m . TSIS) J " ~ ( c , p5 . )
-1.27 - 3 95 - 1 .30 -4.35
Seat
SO? Seat
SbFj
0,7 20 ..
F19 Spectra.-The FI9 spectra were obtained a t 56.5 Mc. and are shown in p.p.m. (parts per million of applied frequency or field) relative to external CF3COOH (Fig. 7). The antimony pentafluoride complexes were reasonably soluble in both solvents ( H F and SO*)and the spectra indicated only the ion SbF6- in all cases. In the H F solutions, exchange of fluorine from SbF6- to H F is slow enough to give separate lines. However, there is no evidence to exclude the possibility of an exchanging, highly polarized -F SbF5 system. The H F lines are quite broad in some cases. There is a large solvent effect for the SbF6- ion in SOz of IT.6 p.p.m. from that in H 2 0 , and -3.8 p,p.m. in H F , or a difference of 13.8 p.p.m. of H F solutions relative to SO2 solutions. The possibility of fluorine exchange in a highly polar-
-
CHa-C-C-
of
Vol. 85
BAKER
-
The proton line a t -2.02 p.p.m. is assigned to the tert-butyl oxocarbonium ian, and the line at - 1.45 p.p.m. to the donor: acceptor complex (CHJ3CCOF SbFS. To substantiate that the line at -3.93 p.p.m. is indeed assigned correctly to the trimethylcarbonium ion, we investigated solutions of tert-butyl fluoride in excess SbFj as solvent.'l The resulting proton spectrum showed a single line a t -4.33 p.p.m. and it thus appears that the trimethylcarbonium ion is formed directly from tert-butyl fluoride in excess of the Lewis acid SbF5, and is the same species as that formed from pivalyl fluoride by CO elimination (Fig. 6). Table 111 summarizes observations of proton resonances relating to the trimethylcarbonium ion. Small differences in chemical shifts are due to the different solvents used (SO2and SbF,). This suggestion was substantiated by observations of the proton magnetic resonance spectrum of (CH3)4r\;-SbF6- in the same solvents, where similar solvent effects were found. If spectra of the pivalyl fluoride :antimony pentafluoride complex (in SO?)were obtained a t --4O0,the amount of the trimethylcarbonium ion present was decreased (as shown by the small -3.93 p.p.m. peak). At roam temperature the ion disappears after standing for some hours and a substantially broadened peak indicates polymer formation. ( 1 I) l ) e l a i l \ t o h e ouhliihed in a suhsequent paper of this series. Partly Ixeiented in talk5 given h y G. A. Olah a t t h e 4th Brookhaven Reaction hlechanivn Conference. U p t o n , N . Y , August. I W t 2 , a n d a t t h e Stahle Carhrmium Ion S y m l i o \ i u m a t the I 1 2 n e S a t i o n a l Sleeting of t h e American C1irrnic:il Society A t l m t i c C i t y , S .[ , Seplernl,er. l ! # l i Z ,
6-
6+
ized complex of the type RCO
+
I
SbFS, where the C-F
F band must be considerably weakened (and in the limiting case ionized) must be considered. There is also a possibility of exchange involving solvent H F and SbF6ions. Fluorine resonance probably cannot differentiate between the SbF6- line and the one corresponding to a n exchanging F-SbFj system. Attempts were made to see if a t lower temperature, due to the decreased exchange, differences are observable, but this was not the case. C13 Resonance Investigation of CH3C 130+sbF6-.-To further substantiate t h a t in the ionic oxocarbonium complexes the positive charge is a t least partially located on the carbon atom (and not-as alternatively possible-entirely on the oxygen), the complex CH3C 13O+SbF6- was prepared and investigated. Due to the fact that the solubility of the complex in the used solvent employed (anhydrous H F ) is less then lo%, it was necessary to prepare the complex with the highest CI3 isotopic concentration (53YG) available to use, as the sample is tenfold diluted in solution. Owing to the low concentration of the complex in the solution it was not possible to observe the C13resonance directly on a 4-mm. sample, even though enriched. However, this resonance was observable indirectly by the INDOR (internuclear double resonance) method'j and showed a 1 : 3 : 3 : 1 quartet a t 15.00231 Mc. with the proton methyl resonance centered a t ti0.009769 Mc. The coupling of the enriched carbonyl carbon to the methyl protons was JCCH = 6.0 C.P.S. The C l 3 0is shifted -30.9 p.p.m. relative t o C'3 in CF3C1300H,or -43.4 p.p.m, relative to C13 in acetyl fluoride, CH3C'30F. A direct C'" spectrum of an enriched sample of CH3Cl30F consisted ofa pair of quartets with J C C H= 7.6 C.P.S. and J C F= 335.2 C.P.S. Thus the C1zOin the complex is considerably less shielded than in acetyl fluoride, which is consistent with a t least a partial positive charge on the carbonyl carbon atom in the ion [ CHI-CII&O]+
Chemical Reactivity.-The isolated oxocarbonium complexes are highly effective acylating agents in c-. 0-, N- and S-acylations of a variety of organic c o n pounds. [I.?)I,:
E!
Baher, ./.
('/ic,iii
P h j ? , 3 7 , YI I l l ! # l , 2 )
May 5$ 1963
A L K Y L OXOCARBONIUM
Aromatic hydrocarbon when allowed to react with oxocarbonium hexafluoroantimonates, preferably in nitromethane solution, gave the corresponding ketones ArH
+ RCO+SbFB- +ArCOR + H F + SbFs
Yields of ketones obtained are listed in Table IV. TABLE I\' C--%CYLATIOSO F .%ROhfATICS W I T H O X O C A R B O K I U M HEXAFLUOROASTIMOXATES
___-
I
Aromatic compound
CH3CH2COtSbF6-
Benzene Ethyltevt-ButylFluoroChloroToluene
93 96
Yield ketone, %----(CHa)zCHCO-SbFs-
(CH3)3COiSbFa-
87
93 89 92
3 3 27
90 68 89
2
All yields reported were obtained from preparative runs. Although acylations seem to give high yields, data do not necessarily represent optimum conditions, owing t o losses during preparative operations. Reaction of tert-butyl oxocarbonium hexafluoroantimonate with beniene yielded predominantly tertbutylbenzene. The reactions with toluene, ethylbenzene and tert-butylbenzene gave varying amounts of pivalophenones and tert-butylated products. tert-Butylations are believed to take place through primar y decarbonylation of the tert-butyl oxocarbonium ion (CH3)3CCO+ S ( CH3)3CC
+ CO
as no decarbonylation of pivalophenones was observed under the reaction conditions used. These data are in accordance with previous osbervations on the FriedelCrafts reaction of pivalyl halides with aromatics, investigated by Rothstein* and Balaban and N e n i t z e s ~ u . ~ It was interesting t o observe that in the reaction of isopropyl oxocarbonium hexafluoroantimonate with toluene and benzene, although the corresponding isopropyl ketones were predominantly formed, there was a smaller amount (3.5% in the case of the reaction of benzene, 1.5% in the case of toluene) of isopropylated hydrocarbons formed (as analyzed by gas-liquid chromatography, using a Golay-type capillary column and hydrogen flame ionization detector). Thus a limited decarbonylation of the isopropyl oxocarbonium ion according to ( C H 3 ) z C H C O ' Z (CHa)zCH+
+ CO
was taking place. Table V summarizes the ratios of acylations compared with alkylations in the reactions of isobutyryl and pivalyl fluoride complexes with aromatics. TABLE 1-
REACTIOXSO F METHYL-A N D ETHYLACETYL FLUORIDE COMPLEXES WITH AROMATICS
K A T I O O F -%CYLATION
Aromatic
Benzene Toluene Benzene Toluene Ethylbenzene Icrt-Butylbenzeiie Benzene Toluene Benzene Toluene
TO -ALKYLATION I N THE
ArCOR: ArR
29 8 82 2 0.1 03 .09 .67 2 3 11.2 0 01 0 01
Oxocarbonium salts are also highly effective acylating agents for 0-acylation of alcohols giving esters
HEXAFLUOROANTIMONATES ROH
4- R'CO'SbFs-
ROOCR'
1333
+ H F + SbFj
for acylation of mercaptans giving thiolesters RSH
+ R'CO-SbFa-
+KSOCR'
+ H F + SbFj
and for N-acylation of primary and secondary amines yielding amides 2RNH2
+
R'CO'SbF6-
+ RKHOCR'
+
RSH2.HSbFs
Results of acylations are shown in Tables VI, VI1 and VIII, respectively. TABLE 1.1 0-.%CYLATION OF XI COHOLS WITH OXOCARBOSIUM SALTS _ _ ~Yield- alkyl ester, 7'-Alcohol
Methanol Ethanol 1-Propanol 1-Butanol teit-Butyl alcohol 1-Octanol
CHaCHKO SbFa-
A-
I C H d K H C O -- (CHdaCCO-SbF6SbF6-
70 78
65 74 72 80 81 82
53 58 63
79
TABLE 1-11 S-.%CYLATIOS O F
MERCAPTANS 11 ITH
OXOCARBOVIUM S A L T S
Yield thiolester, GG ----CH3CHzCOt- ( C H d z C H C O - ( C H d C C O SbFsSbFsSbFb-
I - -
Mercaptan
Methanethiol Ethanethiol n-Octanethiol
52 50 70
57 59 62
47 51 48
TABLE TI11 WITH OXOCARBOSKM SALTS
S-XCYLATIOX OF MIXES
Amine
A%mmonia Ethylamine Diethylamine dniline
---CHCHaCO SbFa-
+-
Yield amide, (&-----(CH3)gCHCO (CHd3COO'SbFsSbF6-
90 89 84 90
81 87 90 82
63 70
Experimental Acyl fluorides were prepared by the method earlier described. Antimony pentafluoride was obtained from the Harshaw Chemical Co., Cleveland, O., and was freshly distilled before use. BaC130s was obtained from Merck, Sharp a n d Dohme L t . , Montreal, Que. Preparation of Oxocarbonium Hexafluoroantimonates.Acyl fluoride (0.2 mole) was dissolved in 70 ml. of 1,1,2-trifluorotrichloroethane (Freon 113) and t h e solution was allowed t o react at -5 t o 0" with an equimolar solution of freshly distilled antimony pentafluoride, adding this solution t o the stirred cold acyl fluoride solution. .%fter half a n hour of continued stirring, the white, crystalline precipitate t h a t formed was filtered, washed with cold Freon 113 and dried under vacuum. All operations were carried out under the usual conditions for excluding moisture, preferably in vacuum systems. Tields are generally close t o quantitative. The prepared oxocarbonium salts could be recrystallized from liquid SOP solutions and were obtained as colorless crystalline compounds with high purity. Melting points (in sealed tubes) are given in Table I . Elementary analysis of the complexes are summarized in Table IX. The fluoride analyses were carried out by neutron activation, in sealed polyethylene sample tubes. This method is particularly recommended for the analysis of hydrolytically sensitive or otherwise not too stable complexes. CH3C130F.-CH3C1300Sa (1.66 g., 0.02 mole) with 53% CI3 content (obtained from Merck, Sharp and Dohme of Canada Ltd., Montreal, Que.) was placed in a semimicro distillation apparatus, provided with a 10 ft. Vigreux column and protected in the usual way from atmospheric moisture. Benzoyl fluoride (25 g., 0.2 mole) was added through a dropping funnel t o the reaction flask, which was gently heated t o regulate a smooth evolution of the CH3C130F. When the distillation stopped (b.p. 20-21') t h e collected acetyl fluoride (cooled receiver) was found (infrared and n . m . r . spectra, gas-liquid chromatography) to be of sufficient purity t o be used nithout further purification in the preparation of the antimony pentafluoride complex. CH3C'30LSbFs-.-The preparation of the labeled methyl oxocarbonium salt was carried out under identical conditions as
DIETRICHHEINERTAND ARTHURE. MARTELL
1334
Vol. 83
TABLE IX
---Fluorine, 7*-Mol. wt.
Calcd.
CHsCHiCO +SbFs293 38.93 (CHa)?CHCO+SbF307 37.15 (CH3)aCCO-SbFs32 1 35.52 C H B C H ~ C H ? C'SbFsO 307 37.15 (CHaCH2)2CHCO'SbFs335 34,03 CsHsCH2CO+SbFs355 32.12 (CsHj)aCHCO+SbFs43 1 26.45 S o t determined due t o relative instability of complex a t room described previously for CH3CO+SbFs- and CD,CO+SbF,-,* by treating Freon 113 solutions of CH3C130F and SbFe a t - 5 t o - 10" and subsequently isolating the stable, crystalline oxocarbonium salt. Reaction of Oxocarbonium Salts with Aromatic Compounds. ( a ) Without Solvent.-The appropriate oxocarbonium salt (0.2 mole) was added into 0.5 mole of well-stirred aromatic. T h e complex salts are generally not soluble in the aromatics. In most cases gentle heating was necessaqy t o start the reaction. The formed ketones give complexes with t h e by-product Lewis acids and separate from the excess aromatic as a lower layer. After washing t h e reaction mixtures, they were dried over S a ? S 0 4and the products isolated (fractionation or crystallization). ( b ) I n Solution.-In these experiments the reaction was carried out in nitromethane solutions in which the aromatics and the oxocarbonium salts are both soluble. The reactions are much slower in solvent, and owing t o the partial decomposition of the oxocarbonium salts in nitromethane the yields are lower. Reaction of Oxocarbonium Salts with Alcohols.-Oxocarboniutn salt (0.3 mole) was added as nitromethane solution or in
[COSTRIBUTIOS FROM
THE
Found
38.59 37.07 35.44 37.03 33.87 31.98 26.13 temperature.
--Carbon, Calcd.
12.30 15.65 18.71 15,65 21. El 27,07 39.01
R--Found
12.18 15,78 l5,39 21.18 26.80 38.28
-Hydrogen, Calcd.
%--
1.72 2.30 2.82 2.30 3.31 1.98 2.57
1.79 2.27
Found
2.21 3.17 1.78 2.39
small portions as a solid into 0.6 mole of the appropriate stirred and cooled alcohol. X fast reaction takes place. The resulting mixture was washed with water, dried over S a 8 0 4 and fractionated. Reaction of Oxocarbonium Salts with Mercaptans.-Oxocarbonium salt (0.3 mole) was added in nitromethane solution or in small fractions as a solid to 0.6 mole of well stirred and cooled mercaptan. The reaction is very fast. After completion of the reaction, the mixture was washed with water, dried over SaZSO1 and fractionated. Reaction of Oxocarbonium Salts with Amines.-The solution of 0.3 mole of oxocarbonium salt in nitromethane or SO2 solution was added t o 0.6 mole of t h e stirred and cooled primary or secondary amine. The products, after water washings, were isolated either b y distillation or crystallization.
Acknowledgments.-Dr. 0. U. Anders, Radiochemistry Laboratory, The Dow Chemical Company, Midland, Mich., is thanked for neutron activation fluorine analyses.
DEPARTMENTS OF CHEMISTRY OF CLARKUNIVERSITY, WORCESTER, MASS.,A N D ILLINOIS INSTITUTE OF TECHSOLOGY, CHICAGO, ILL.]
Pyridoxine and Pyridoxal Analogs. VIII. Synthesis and Infrared Spectra of Metal Chelates' BY DIETRICHHE INERT^
AND
ARTHURE. MART ELL^
RECEIVEDSOVEMBER 21, 1962 The preparation and purification of a number of metal chelates derived from the amino acid Schiff bases of pyridoxal analogs 3-hydroxypyridine-4-aldehyde and 3-hydroxypyridine-2-aldehyde are described. Copper( 11) chelates containing a 1 : 1 : 1 molar ratio of Cu(I1):hydroxyppridinea1dehyde:aminoacid and iron(II1) and nickel(I1) chelates having a 1 : 2 : 2 molar ratio of constituents were synthesized. Glycine, glutamic acid and valine were employed as the amino acids. Probable structures of these compounds were deduced from their stoichiometry, properties and infrared spectra
This paper describes the synthesis and properties of Fe(III), Cu(I1) and Ni(I1) chelates of the 3-hydroxy-4pyridinealdimines (XVI) and 3-hydroxy-2-pyridinealdimines (XVII) containing residues of the natural amino acids glycine, valine, phenylalanine and glutamic acid. Interest in such metal chelate compounds arises from their analogy to pyridoxylideneimine chelates, whose role in vitamin Recatalyzed reactions has been the subject of considerable investigation. The role of metal chelates of pyridoxal derivatives of the natural amino acids has been reported by Metzler and Snel13 and a general mechanism for a wide variety of reactions catalyzed by these compounds was described by Metzler, et u Z . ~ The first crystalline metal chelate of pyridoxal Schiff bases reported was the 1: 1 Cu (11) chelate of pyridoxylidenetyrosine described by Baddi1ey.j Christensen isolated and measured spectrophotometrically the formation in solution of the Cu(II), AZn(II), Ni(II), Zn(II), and Mg(I1) chelates of pyridoxylideneglycine,6 containing a 1: 1 molar ratio of (1) This investigation was supported by research grants A-1307 and A,5217 from t h e National I n s t i t u t e of Arthritis and Metabolic Iliseases, U . S. Public Health Service. (2) Department of Chemistry, Illinois Institute of Technology, Chicago 10, Ill. (3) D. E. Metzler and E. E . Snell, J . A m C h e m . Soc., 74, 979 (1952). (4) D. E. Metzler, .\I Ikawa . a n d E. E. Snell, ibid., 76, 048 (1954). (.5) J. Baddiley, A-uli
