close

Вход

Забыли?

вход по аккаунту

?

Carbonyl Oxides Zwitterions or Diradicals.

код для вставкиСкачать
Carbonyl Oxides : Zwitterions or Diradicals?
By Wolfram Sander *
Over the last few years new experimental and theoretical methods have made it possible to gain
a more detailed insight into the chemistry of short-lived reaction intermediates. In 1949 Criegee
postulated the intermediacy of carbonyl oxides in the mechanism of ozonolysis, and since then
these species have become the goal of much research effort. Even though the formation of
“Criegee zwitterions” during ozonolysis and carbene oxidations was proven by scavenger
experiments, the electronic structure -zwitterion o r diradical--of this short-lived species is still
a subject of debate. To date n o stable carbonyl oxide has been found to exist under “normal”
laboratory conditions, although by using matrix isolation and laser spectroscopy, it has been
possible to obtain highly resolved IR and UVjVIS spectra of carbonyl oxides as well as to
determine their dipole moments experimentally. The influence of substituents, exact kinetic
data on modes of formation, and the subsequent reactions of carbonyl oxides as well as their
photochemistry complete the picture. In accordance with a b initio calculations carbonyl oxides
are best viewed as polar diradicals. The zwitterionic state lies at higher energies and should be
stabilized by R donors.
1. Introduction
Carbonyl oxides 1 and their isomeric form, the dioxiranes
2, have become an intensively studied species over the last
few years owing to their role in oxygen-transfer reactions.
The occurrence of I and 2 as short-lived reaction intermediates has been discussed for combustion processes,[” many
“organic” oxidations (ozonolysis, Baeyer-Villiger reaction),” - 51 and enzymatic processes.[61Furthermore, the use
of 1 and 2 in selective oxidations is becoming increasingly
imp~rtant.[~
’1 ’ ~The
introduction of appropriate substituents causes these species to react with substrates by
transfer of 0 atoms, either in a nucleophilic or electrophilic
manner. In contrast to the extremely labile 1, which has so
far only been characterized indirectly or spectroscopically, a
number of derivatives of 2 have been isolated in solution or
in their pure form a t room t e r n p e r a t ~ r e . [ ’I~ , ~ .
Carbonyl oxides 1 were first postulated as intermediates in
the ozonolysis of olefins by Criegee in 1949[z.81and are also
referred to as Criegee zwitterions in the literature. Since then
the many experimental and theoretical studies in this area
have on the whole substantiated the postulated mechanism
in Section 2 [Eq. (2)J in its key points. Nonetheless, a direct
proof of the existence of a carbonyl oxide was not achieved
until 1983.[91To prove the existence of 1, the application of
relatively new experimental procedures such as matrix isolation and time-resolved laser spectroscopy was necessary.
Carbonyl oxides are also interesting species from a theoretical point of view. Their isomeric form, the dioxiranes, are
molecules with a complete valence shell and are thus easily
described in theoretical terms;” O , the electronic structure
of carbonyl oxides, on the other hand, is still a subject of
debate. In most of the experimental work the description as
an ylide, 1 A and 1 B [Eq. (I)], originally introduced by
Criegee, has been invoked. Quite a few important reactions
of carbonyl oxides such as 1,3-dipolar cycloaddition, dimerization, and addition to alcohols may be explained in terms
of this species. Some of the semiempirical calculations that
have been performed postulate a zwitterionic ground
state,[’21 while a b initio[13- 15] and other semiempirical calculations[I6*l’] predict a diradical ground state l C. Experi-
p=oepe
R\
R‘
-
R\e
c-0
R‘I
PeA
- R\.c-0 p’ ===R”
1B
1A
R\
R”
0
c:i
(1)
0
2
1c
mental evidence also favors a diradical ground state.“’] In a
review on ozonolysis in 1975, Criegee wrote:121
Can the view recently expressed by Goddard that the socalled carbonyl oxides are singlet diradicals also in solution be confirmed or disproved? Solution of this problem will be of crucial importance for our understanding
of the events occurring during the ozonolysis of olefins.
From 1980 onward the development of new experimental
and theoretical methods has led to an enormously increased
interest in carbene oxidations, carbonyl oxides, and dioxiranes. The use of these new methods in the field of carbonyl
oxides is the focus of discussion in this article.
2. The Criegee Zwitterion in Ozonolysis
In the three-step mechanism [Eq. (2)] describing ozonolysis,[’] the initial step involves the 1,3-dipolar cycloaddition of
ozone to an olefin, resulting in the formation of a 1,2,3-trioxolane (3; “primary” ozonide). In the second step, 3 decomposes to afford a carbonyl compound and 1. These components react in a further cycloaddition to yield a 1,2,4-trioxolane (4; “molozonide”), which may often be isolated from
the reaction mixture. The great importance of ozonolysis in
[*] Priv -Doz. Dr. W. Sander
Organisch-chemisches Institut der Universitit
Im Neuenheimer Feld 270, D-6900 Heidelberg (FRG)
344
t>
VCH V e r l u ~ s ~ e s e l l s r hmhH,
u ~ 0-6940 Weinheim,1990
3
0570-0833/90/0404-0344 $02.50/0
a
1
4
Angew. Chem. Inr. Ed. Engl. 29 (1990) 344-354
organic chemistry and recently also in the chemistry of the
upper atmosphere is reflected in the large number of publications and reviews covering this area.[18-201
It was possible to confirm the Criegee mechanism by detecting cross products, by carrying out trapping and isotopic
labeling experiments, and, finally, by direct detection of the
unstable primary ozonides 3.123 19] Recently, the parent
compound 3 was detected by microwave spectroscopy as one
of the products in the ozonolysis of ethylene.[”] A weak
point of the Criegee mechanism is the observed stereoselectivity of ozonolysis: If the proposed three-step mechanism is
to operate, cis and trans olefins should give the same ratio of
stereoisomeric molozonides 4 and thus lead to loss of stereochemical information. In fact, trans and cis olefins with
bulky substituents lead mainly to the trans and cis forms of
4,respectively. According to a proposal by Bailey, the observed stereoselectivity can be explained by a modified mechanism, which rests upon the following points: the cleavage of
3 is presumed to be a concerted process and 1 exists as syn or
anti isomer, the two of which d o not equilibrate at room
t e r n p e r a t ~ r e . [ ’ ~ ~With
~ ~ ~ .these
~ ~ . ~refinements,
~J
the stereochemistry of ozonolysis is determined by the conformation of the primary ozonides 3 and is dependent on the configurational stability of 1.
In some cases the products of ozonolysis are 1,2,4,5tetroxanes 5, which can be viewed as dimers of 1.[18] Performing the ozonolysis in the presence of alcohols affords
alkoxy hydroperoxides 6, which are formed by addition, no
definite spectroscopic evidence exists so far for the alcohol to
1 [Eq. (3)].[”l Apart from these indirect indications, no definite spectroscopic evidence exists so far for the existence of
1 as a reactive intermediate during ozonolysis.
3. Oxidation of Diazo Compounds
and Carbenes in Solution
Apart from ozonolysis, there are two alternative methods
for the preparation of 1 and 2: the direct reaction of free
“triplet path”) and the
carbenes 7 with triplet oxygen (30,,
oxidation of diazo compounds 9 with singlet oxygen (lo,,
“singlet path”[251).A formal “end-on” addition would result
in 1, a “side-on” addition in 2 [Eq. (4)].
R\.
0
R’
aids on
c-0
R./
2
R./
end on
7
P’
(4)
1
Kirmse et al. were the first to conduct a selective study in
this field.r261The photooxidation of diphenyldiazomethane
(9b) in a variety of solvents yielded up to 73 YOof benzophenone (8b); no other products containing oxygen could be
detected. It was postulated that the primary adduct arising
from photooxidation is benzophenone 0-oxide, which, in
the following step, transfers its terminal 0 atom to a further
molecule of carbene 7 b or diazomethane 9 b.
Bartlett and Traylor studied the photooxidation o f 9 b with
1 6 0 2 and 1 8 0 2 in solid chlorobenzene at - 78 0C.[271
Under
these reaction conditions it was possible to isolate 15 YOof
the tetroxane 5 b, a dimer of 1 b. From their experiments with
mixtures of 1602
and “ 0 2Bartlett
,
and Traylor were able to
conclude that the elimination of 0, from two molecules of
1 b does not proceed to a significant extent.[271More recent
results by Girard and Griller show that, if no trapping
reagents are present, the extrusion of 0, from two separate
molecules of 1 b can become the main reaction pathway for
1 b, and the formation of 5 L a t least in solution at room
temperature-does
not occur [Eq. (5)].1281The activation
barrier for this reaction amounts to only 1.8 kcal mol-
’.
Surnram and Lovas were able to detect the unstable dioxirane 2 a‘*’ during the gas-phase reaction between ethylene
and ozone. Its structure was determined by microwave spect r o ~ c o p y . [ ~Presumably,
~]
the thermal relaxation of the
“hot” formaldehyde oxide 1 a, formed in the gas phase, is too
slow to inhibit the rearrangement to 2a. Ozonolysis in solution does not produce dioxiranes 2.
[*] The substituents are given in Scheme 1 (Section 4.2).
2
9.
,c-0
R’
1
P’ -
2
R\
c=o
+02
(5)
R’/
8
Apart from 8 and 5, other oxidation products that arise
from the reaction of 9 are esters (or lactones) 10. The mechanism of ester formation was elucidated in a number of pub-
Wolf..m Sander was born in 1954 in Heidelberg. From 1973 to 1979 he was a student at the
Universitat Heidelberg, where he subsequently received his Ph.D. in 1982 under R. Gleiter and
completed his Habilitation in 1989. He was introduced to the matrix isolation technique as a
postdoc with 0 . L . Chapman in Los Angeles (1982 t o 1984). His interests include the characterization of reactive intermediates, oxidation reactions with molecular oxygen, and eflect of’electron spin on chemical reactivity. In 1988 he was awarded the Karl-Freundenberg-Preis of the
Heidelberger A kademie der Wissenschaften, and in 1990 he received a Heisenberg-Stipendium
and a Winnacker-Stipendium.
A n g m (‘hem fnt Ed En81 29 (1990) 344-354
345
lications by Sawakiet al.[291At first, the formation of 10 was
thought to proceed through an intramolecular rearrangement of 2.
11
1
2
10
Since mixed isotopomers of 10 with one atom of l60and
each are formed when mixtures of l60,and I8O2 are
used, two molecules of 0, must participate to form 10. This
Sawnki et al.
excludes 2 as the direct precursor of
proposed an alternative explanation in which they presume
the presence of 1 as an intermediate. Species 1 reacts with a
further molecule of 0, and then decomposes via different
peroxidic species to yield 10 [Eq. (7)].[29a1
l80
R,.
c-0
,o-
4 2
0-0-
R!
R'
0
"
1
(7)
15
However, this mechanism is also incompatible with the
experimental findings, since the yield of 10 is nearly independent of 0, pressure. In recent work a radical chain mechanism seems to give a plausible explanation.[29b1According to
these results, ester 10 is formed by radical oxidation of a
diazo compound (9) and not of a free carbene (7). The l8O
labeling experiments also showed that the rearrangement
1 + 2 is not a thermal process [Eq. ( 6 ) ] .
In the reaction of difluoromethylene (7s) with O,, McKee
et al. found a species that transfers 0 atoms stereospecifically
onto a l k e n e ~ . [ This
~ ~ 1 species was not characterized in detail,
yet a b initio calculations make formation of 2 s from the
primarily formed compound 1s plausible.
Besides mechanistic studies, much research has been directed toward the use of the intermediates resulting from the
oxidation of 9 as reagents for oxidation. Predominantly the
formation of 1 is presumed in these complex reaction mixtures. The most important reactions are the transfer of oxygen onto alkanes,"21 alkenes,[". 3 1 . 321 sulfides and sulfox341 and aromatic compound^.^^^^ 361 In the
presence of H donors, 1 can react to form autoxidation products.["] Furthermore, cycloadditions with carbonyl compounds[37.381 and ole fin^[^^-^^] ha ve been reported. The
inherent instability of 1 makes it difficult to exclude the
possibility that the oxidation proceeds through other peroxidic species, expecially 2. Adam et al. developed an elegant
method to distinguish between oxygen-transfer reagents by
examining their different n u c l e ~ p h i l i c i t i e s . [Thianthrene
~~~
5-oxide (1 la) is prone to nucleophilic oxidation at the sulfoxide S atom, but is oxidized in an eiectrophilic manner at the
sulfide S atom [Eq. (S)].
1 la
12
From the relative amounts of thianthrene 5,Sdioxide (11)
and thianthrene 5,lO-dioxide (12), an X,, scale (X," = 0 for
electrophiles, X,, = 1 for solely nucleophilic reagents) was
developed.[3s331 Generally, dioxiranes 2 are less nucleophilic
(X," z 0.7) than carbonyl oxides 1 (X," > 0.8). Typical electrophilic oxygenation reagents have even smaller values on
the X,, scale (m-chloroperbenzoic acid in CH,CI,, for instance, has X,, = 0.36). According to Murray,[41the nucleophilic properties of dioxiranes are definitely smaller than
indicated by an X,, value of 0.7 and, therefore, the difference
to the carbonyl oxides is larger.
From sulfide oxidations[341and oxidations of naphthawith a number of p-substituted arylcarbonyl O-oxides, Agarwal and Murray obtained positive Hammett constants and came to the conclusion that these oxidations are
electrophilic in character. Ogata et al. also obtained positive
Hammett constants for the oxidation of diphenyl sulfoxide,
which indicates the presence of an electrophilic 0-transfer
process.[I2' Obviously, the properties of carbonyl oxides depend strongly on the experimental conditions. Therefore, it
became necessary to employ direct spectroscopic methods to
come closer to an understanding of the electronic properties
and bonding of carbonyl oxides.
4. Spectroscopic Identification of Carbonyl Oxides
To date, all spectroscopically identified carbonyl oxides
have been generated by oxidation of free carbenes or diazo
compounds. Two techniques-matrix
isolation and timeresolved laser spectroscopy-have played an important part
in the detection of short-lived species. In both methods the
free carbenes are generated by photolysis of either a diazo
compound (9) or a diazirine (14). In contrast to the matrix
technique, where the carbene is produced under conditions
at which it is stable (low temperature, inert media), the method used by time-resolved spectroscopy is fast detection (ps to
ps) under normal laboratory conditions (room temperature,
organic solvents). Thus, the two methods go hand in hand.
4.1. Matrix Isolation Technique
33b7
346
Pimentel et al.[431and Norman and Porter[441were the first
to use inert-gas matrices to study reactive intermediates. In
contrast to organic glasses, which react more o r less quickly
with free carbenes, the noble gases used in this technique,
such as Ne, Ar, Kr, and Xe o r nitrogen, are all inert.
Through matrix isolation in inert gases it is possible to
record highly resolved IR, UVjVIS, luminescence, and ESR
spectra and thus to carry out an extensive characterization of
reactive intermediates.
A prerequisite for the practicability of the matrix technique as a method to be used extensively in organic chemisAngew. Chem. Int. Ed. Engl. 29 (1990) 344-354
try was the development of closed-cycle helium cryostats.
They allow temperatures as low as 10 K to be attained on a
routine basis.r45.461
Photolysis of 9 and 14 in a matrix (usually Ar or N, at
10 K) generates carbenes 7, which are stable if they satisfy
the following requirements: (1) They do not rearrange intramolecularly. This is why simple alkyl-substituted carbenes that rearrange by a 1,Zshift to give olefins cannot be
isolated in a matrix. (2) They are stable to the irradiation that
is necessary to photolyze the precursor. Many carbenes are
photolabile (phenylmethylene, for instance, easily rearranges photochemically to ~ycloheptatetraene[~'~); therefore, the photolysis should be carried out using the longest
wavelength possible and irradiating monochromatically in
order to inhibit these side reactions.
The temperature of the matrix influences diffusion processes in a complicated mannerf4*].To induce the reaction of
7 with O,, the matrix is first doped with 0, (0.2-5%) and
then annealed (30-45 K). As long as the matrix is kept at
10 K. even a large excess of 0, will not lead to a reaction.
Once 35 K is reached, however, the oxidation is complete
within a few minute^.[^^-^^] An argon matrix containing
small molecules is shown schematically in Figure 1.
light and a photomultiplier (for single-channel detection) or
an array of diodes (for multichannel detection).
Scheme 1 provides a summary of the carbonyl oxides for
which spectroscopic or theoretical data are available.
H\
c=oz
H'
la
cl$Loz
CI
C' I f
A
Ph\
c=oz
F, C'
lk
0%
1P
lr
18
(52
c=02
'
H
lu
I
Scheme 1.
4.3. Cyclopentadienone 0-Oxide
Fig. 1. Schematic representation of an Ar matrix. Open circles, argon atoms;
dark circles, small molecules such as N, or 0,. At 10 K, diffusion of Oz is so
slow that, even after hours, no oxidation products of carbenes are formed. At
35 K . on the other hand, oxidation IS complete within a few minutes.
4.2. Laser Spectroscopy
The great advantage of time-resolved spectroscopy is that
it allows the measurement of absolute kinetics; its disadvantage is that the spectroscopic possibiiities for structural determination are restricted to the near-UV and visible range
with relatively low resolution.r531Recently, the range has
been extended to cover the infrared as we11.r54-571
In timeresolved laser spectroscopy, carbonyl oxides 1 are easily detected by observation of their strong band in the visible part
of the spectrum, whereas their photoproducts with absorption maxima below 300 nm cannot be monitored.
In order to study carbene oxidations in solution, time resolution in the ns range will suffice. The sample (generally a
solution at room temperature) is photolyzed by a short laser
pulse. the time resolution being defined by the length of the
pulse. The nitrogen (337 nm) and excimer (249 or 308 nm)
lasers usually employed have pulse lengths of 5-10 ns.[531
The detection system consists of a source of intense white
Angeu. C'hem. In!. Ed. Engl. 29 (1990) 344-354
Independently of one another, Chapman et al.[58-601and
Dunkin et al.[49352a1 reported the formation of cyclopentadienone 0-oxide (1 c) from the matrix oxidation of cyclopentadienylidene (7c). Chapman's group studied the photooxidation (irradiation with light at 1 > 418 nm) of diazocyclopentadiene (9c) in matrices with high 0, content (20%).
They discovered a product which was assigned structure 1 c
owing to its intense IR bands at 1395, 1385, 895, and
741 cm-'.[601
On the other hand, Dunkin et al. described the formation
of a different product with IR bands at 1398,1321,888, and
758 cm- This species, generated by monochromatic irradiation (A. = 296 nm) of 9c isolated in a matrix in the presence
of 10% O,, was also assigned structure 1 c.[~']The same IR
spectrum was observed when a matrix doped with 1 % 0,
and containing carbene 7c was warmed to 35 K[491[Eq. (9)].
'.
7c
lc
2c
In a later publication Dunkin et al. showed that l c is extremely photolabile and that the molecule described by
Chapman et al. must be an isomer of 1 c, the dioxirane ~ c . [ ~ ' " J
To prove the differing symmetry of l c (C,) and 2c (C2J,
341
isotopic labeling with 8O was used.[52a1These experiments
indicate clearly that 1 c contains equivalent and 2 c nonequivalent 0 atoms. A further proof for the structure of 2 c is that,
on UV irradiation ( A > 220 nm), it rearranges to a-pyrone.[46. 52al
The largest I8O isotopic shifts in l c are found for the
bands at 1014 (-28) and 947(-22) cm-' (Table
These bands lie in a region typical for stretching modes of
single bonds and can approximately be assiged to 0-0 or
C-0 stretching vibrations. In the region of visible light the
spectrum shows an intense absorption with a maximum at
420 nm (Table
Table 1. Characteristic IR and UV spectroscopic data for some carbonyl oxides 1.
~(0-0
[cm-'1
)
Id]
Matrix
Ib
IC
Id
le
896 s
1014 m
947 m
912 m
898 m
A [b]
Matrix
- 35 (3.9)
-28 (2.8)
-22 (2.3)
422 [d]
420 [gl
-
If
1g
Ih
li
11
lk
I1
Im
In
lo
890 s
91Sm
890 s
901 s
931 m
931 s
1009 s
943 s
1049 m
997 s
1045 m
1034 vs
-50
-30
-21
-30
-41
-42
-49
-20
-35
-18
-35
-60
-57
(5.6)
(3.3)
(2.4)
(3.3)
(4.4)
(4.5)
(5.3)
(2.0)
(3.7)
(1.7)
(3.5)
(5.7)
(5.5)
445 [h, i]
460 [h, i]
395 [i]
400 [k]
387 [I]
Laser
4.4. Benzophenone 0-Oxide
and Other Arylcarbonyl Oxides
450 ti]
1W
lx
1Y
[fl
427
M
468
[fl
582 [n]
378 [o]
-
[Ol
462 [PI
410 [ql
425 [r]
546 [fl
493 M
504 [fl
IP
lu
lv
489
406 [m]
Lq
Lr
it
[nm] [c]
Calculation
i
i+ ii*
[cm- '1
401 [el
435 [s]
405 [el
390 [el
428 [el
418 [el
[a] 0-0 stretching vibration. [b] 160-180
isotopic shift; in brackets, relative
shift in %. [c] Absorption maxima. [d] Ref. 1511. [el Ref. (681. [fl Ref. 1171. [g]
Ref. [52a]. [h] Ref. [52b]. [i] Ref. [61 a]. b] Ref. [62]. [k] Ref. [65]. [I] Ref. [51].
[m] Ref. [64]. [n] Ref. [73]. [o] Ref. [SO]. [p] Ref. [72]. [q] Ref. [71]. [r] Ref. [9].
[s] Ref. [70]
Dunkin et al. also studied indenone 0-oxide (1 d),[52b361a1
fluorenone 0-oxide (1 e),[52b*
61a1 and tetrachlorocyclopentadienone 0-oxide (1 f).[611
The strongest IR bands for 1d are
observed at 912cm-' and for l e at 898cm-1.[s2b1Even
though n o I8O labeling experiments were performed with 1 d
and I e, comparison with other carbonyl oxides 1 (see below)
convincingly shows that these bands can best be interpreted
as 0-0 stretching modes (Table 1).
Carbonyl oxides 1 c - 1 f show strong absorptions in the
visible part of the spectrum between 395 and 460nm
Absorption in this region is characteristic for
(Table
1 and can be used to detect such species in laser photolysis
experiments. Of the variety of substituted cyclopentadienone
0-oxides, only 1 e has been characterized in laser photolysis
348
experiments by Scaiano et al.[621To generate the carbonyl
oxide l e , diazofluorene (9e) was either irradiated in solution
a t room temperature in the presence of 30,(triplet path) or
allowed to react with '0, (singlet path). Independent of the
method of generation, both paths led to a transient species
with an absorption maximum at 450 nm.r621The half-life is
strongly dependent on the solvent used and ranges from
> 800 ps in benzene to only 9 ps in CH,Cl,. On the basis of
the kinetic data and the observed absorption maximum,
structure l e was proposed for the transient species. The
above results are in accordance with the data obtained in a
matrix and show that the same reactive species is formed
both in solution at room temperature and in an inert gas at
10 K. Just as Criegee had postulated, 1 e is quickly trapped
by reaction with aldehydes.['] At large enough concentrations of aldehyde, pseudo-first-order kinetics with rate constants between 107 and 109 mol-'sC1 (in Freon-I 13) are
observed.[621
The chemistry of benzophenone 0-oxide (1 b) is well established by trapping experiments in solution, but has also
been studied by IR and UV spectroscopy in matrices and
with the aid of laser photolysis. In the thermal reaction of
diphenylmethylene (7 b), isolated in a matrix containing ' 0 ,
(Ar, 30-40 K), 1 b is produced in high yield. The product
was characterized by UVjVIS and I R spectroscopy as well as
by its subsequent photochemistry (Scheme 2).[63.5 1 1 In
agreement with the findings from laser photolysis experiments, an intense broad band with I?,, = 422 nm was observed in the visible region of the spectrum (Table l).[631
This
can be taken as proof for the existence of the same products
in solution at room temperature and in a matrix at cryogenic
temperatures.
The structure of 1 b has been further elucidated by carrying out IR experiments in connection with isotopic labeling
analysis. The reaction of [1-12C]-7b and [1-13C]-7b with
l60, and "0, leads to four isotopomers of l b , thereby
h,
CsN,
P i
9b
h\
c-0
P h'
+
hv
543nm
h\
C:
ph' 7b
I3
Ph\
+to1
Ph/
c-0
*
8b
LO]
hv
Ph\.
c-0
630nm ph'
8b
-
p
hv
630nm
Ph\ 0
C c l 2b
ph'
0
lb
1 Ob
15b
Scheme 2.
Angew. Chem. In!. Ed. Engl. 29 (1990) 344-354
allowing a nearly complete assignment of the IR
The most intense band in the spectrum is observed at
896 cm- and exhibits an " 0 isotopic shift of 35 cm(Table I). Because of the small shift on labeling of the carbony1 C atom with 13C, one can conclude that this band
arises from the 0-0 stretching vibration. The 0-0 deformation mode is a weak band at 550 cm-'; in the region typical
for the carbonyl vibration, no band is observed.
Like all other carbonyl oxides that have been matrix isolated to date. 1 b is extremely photolabile. Irradiation with
light far to the long-wavelength side of the absorption band
(1> 630 nm) leads to rearrangement (main reaction) to afford diphenyldioxirane (2 b) or results in cleavage of 0 atoms
(side reaction) to form 8 b (Scheme 2). Presumably, there
exists a long-wavelength band of weak intensity that has not
been detected so far.1171
Under these conditions, dioxirane 2b is stable. On shortwavelength irradiation (1 > 400 nm), however, it rearranges
to the phenyl benzoate (lob). On oxidation of 7 b in the
matrix. one can aIso detect an intense chemiluminescence
that has been identified as the phosphorescence of 8b.["]
Two further carbonyl oxides with absorption maxima at
about 400 nm are benzaldehyde 0-oxide (1 h)'"' and orthochlorobenzaldehyde 0-oxide (1 i).[641The absorption spectra
of both compounds exhibit distinct vibrational fine structure; for I i, l8O isotopic labeling experiments indicate that
the vibrational progression corresponds to the 0-0 stretching mode of its electronically excited state.[641The vibrational frequency of 901 cm- in the ground state (Table 1) is
reduced to 790 cm- in the excited state.
Other arylcarbonyl oxides that have been characterized in
the matrix are benzoyl chloride @oxide (1g)f65Jand the
CF,-substituted carbonyl oxides 1k and 1 l.f501Whereas the
chloro substituent in 1 g hardly has any effect on the characteristic spectroscopic data, the strongly electron-withdrawing CF, groups in 1k and 11 bring about an increase in the
0-0 stretching mode-an effect attributed to a strengthening of the 0-0 bond. On top of that, the UVjVIS absorption
maximum is shifted to shorter wavelength (Table l)[sO1in the
latter two compounds. The reaction of 7 g with 30,is a
formally "spin-forbidden'' process, because the carbene 7 g
has a singlet ground state. As expected, the oxidation proceeds considerably slower than that of triplet carbenes.t6'. 641
Scaiano et al. were able to generate 1 b by laser photolysis
in solution at ambient temperatures, via both the tripletpath[661and the ~inglet-path.'~~]
Its existence as a short-lived
intermediate was proven by the observation of an intense
absorption at 410 nm.166'671
If no special trapping reagent is
present, I b will react according to second-order
A similar result was obtained by Girurd and Griller with the
aid of optical modulation spectroscopy (this method can
achieve time resolution into the ps range).1281Just like Scaiuno, they found that 1 b disappears in a reaction with bimolecular kinetics, yet they were not able to isolate the dimer
5b. This observation was made plausible by an alternative
reaction path of two molecules of 1 b to yield 8 b and 0,
I E ~(511.
.
Like 1 e, species 1 b is rapidly trapped by aldehydes. The
rate constants of the pseudo-first-order reaction in acetonitrile amount to about lo6 mol-' S - I f681 and are thus considerably smaller than the corresponding values for 1 e.16'1
'
'
'
Angeu, Chrm. I n t . Ed. Engl. 29 (1990) 344-354
Using time-resolved dielectric-loss spectroscopy, Fessenden and Scaiano estimated the dipole moments of 1 b and 1 t
to be 4.0 D and 3.8 D, respectively.[691These dipole moments are larger than expected for a diradical but considerably smaller than anticipated for a zwitterion, and they give
an indication of the complex electronic structure of carbonyl
oxides.
Scaiuno et al. later studied the absorption spectra and
kinetic data of a series of substituted arylcarbonyl 0-oxides
(Table 1).162,68,
701 With the exception of le, all absorption
maxima fall into the range of 410 _+ 20 nm and are hardly
dependent on the solvent used. Because of steric interaction,
the formation of 1 w from 7 w and 30,proceeds more slowly,
and the lifetime of 1 w is longer than that of the other carbony1 oxides.f681
Ztoh et al. studied the photooxidation of methyl a-diazophenylacetate (9n) and were able to generate carbonyl
oxide 1 n.1711The reaction of 1 n with an excess of methanol
was pseudo first order; without methanol, bimolecular kinetics were observed.
4.5. Other Carbonyl Oxides and Analogous Compounds
By laser photolysis of the diazo compound 90 at room
temperature in cyclohexane, Zwamuru et al. obtained 10,lOdimethyl-I O-silaanthracene-9(10m-ylidene (To), which has
a lifetime of 870 r1s.1~~
If the photolysis was carried out in the
presence of 0,, carbonyl oxide 1 o was produced, which has
a lifetime of 45 ps and exhibits an absorption maximum at
425 nm [Eq.
Another cyclic carbonyl oxide that has been monitored in
a matrix is para-benzoquinone 0-oxide (I m).[721Since 1 m
not only contains a carbonyl group but also a carbonyl oxide
functionality, a direct comparison of these two functional
groups becomes possible. The 0-0 stretching mode at
1034cm-' shows an intensity similar to that of the C-0
stretching vibration of the carbonyi group and exhibits an
" 0 isotopic shift of 57 cm-'. This large isotopic shift indicates that little coupling with other vibrations occurs, so that
a force constant corresponding to an 0-0 single bond may
be taken as a good estimate.[721In contrast to 1 m with its
A-acceptor group, bicyclo[6.3.0]undeca-2,4,6,8,1l-pentaen10-one 0-oxide (I j) contains a A-donor group.[731This effects a strong redshift of the 7~ + A* band (see Section 5.3.2).
The photochemistry of 1 m and 1j is analogous to that of
other carbonyl oxides (Scheme 4): long-wavelength irradiation (630nm for 1 m or 480 nm for 1 j) yields the spirodioxiranes 2m and 2j, respectively. These compounds rearrange
to the lactones 10m and 1Oj if irradiation is continued at
shorter wavelength (436nm for 2m or 350 nm for 2j).172,731
349
The reaction of methylene (7a) with 0, in an Ar matrix
was studied by Lee and Pimentel.[741They were not able to
detect 1a or 2a, but found formic acid as the main product,
which in part is produced in an electronically excited state.
In contrast to the reaction of carbenes 7, little is known
about the reaction of silanediyls ("silylenes") 16 with 0,.[751
The formation of an adduct of dimesitylsilanediyl (16a) and
0, was described by Ando et al. (Scheme 3).[761On the basis
of isotopic labeling experiments and a b initio calculations,
the adduct was assigned the structure of dimesitylsilanone
0-oxide (17a). Compound 17a is photolabile, but the photoproducts were not characterized. In contrast to 16a, the thermal reaction of dimethylsilanediyl(16b) with 0, leads to the
formation of dimethyldioxasilirane (18 b) as the sole product, which on irradiation rearranges to methoxymethylsilanone (19) [Eq. (12)].[77a1
9
R'C=N*
11
ISC
.c: e
3rR,
R'
>.I
R
14
' rR,
R'
s-7
T-7
R'
faat
+02!
R,.
R'
.c-0
R
hu
,o
-c
R'
1
R,
A/
R,
Si:
+
R'
.o
I
O2
?\
-
R
R,@
'si:
R'
18
R'
16
hv
Si. I
0.
R
0-
1
0"
RO,
/
Si-0
R'
a: R=2,4.6-(CH3)3C6H2
S i -0
19
b: R=CH3
c:
R=F
R
d: R=CI
Scheme 3.
The halogenated silanediyls 16c and 16d d o not react
thermally with 0, in the gas phase or in the matrix to an
appreciable extent, yet on irradiation they are photooxidized
to the dioxasiliranes 18c and 18d, which are stable towards
UV irradiation.[77b1
OR'
4'
10
Scheme 4.
R'
17
..c.
R'
15
conversion, IC) or by extrusion of an 0 atom and subsequent
formation of a carbonyl compound (8, Scheme 4). The ratio
of 1 to 8 is strongly dependent on the substituents at the
carbene center. Especially in the case of H atoms as substituents (1 h, 1 i), the extrusion of 0 atoms predominates and
1 is formed only in low yield^.[^'.^^]
According to a b initio calculations (MP2/6-31C*) performed by Cremer et al., the rearrangement 1a + 2 a shows
an activation barrier of 22.8 kcal mol- and is exothermic by
34.5 kcalmol-' (according to Kajafi et al., 44.1 kcal
mol- [ I 5 ] ) (Fig. 2).114d1
The activation barrier depends on
22.8
5. Formation and Properties of Carbonyl Oxides
5.1. Mechanism of Carbene Oxidation
Carbenes 7 formed by photolysis of 9 or 14 are primarily
generated in the singlet state; in the case of triplet carbenes
the triplet ground state is reached through fast intersystem
crossing (ISC). Thermal reactions in the matrix generally
occur from the ground state, whereas, in reactions in solution at room temperature, low-lying excited states may play
an important part (Scheme 4).[7x]
The reaction of triplet carbenes (T-7) with 30,is an extremely fast process (controlled by diffusion); the formally
"spin-forbidden'' reaction of singlet carbenes, on the other
hand, is very
6 5 1 and cannot be detected by timeresolved spectroscopy in solution.[791
Since the reaction of 7 with 0, is very exothermic (for the
formation of I a, ca. 47 kcalmol-'),1'51 the primary products of carbene oxidations, carbonyl oxides 1, are generated
in vibrationally excited states. Their stabilization occurs either through relaxation to the thermal ground state (internal
350
63.8
-
-
Fig. 2. MP2/6-31G*-calculated barrier of activation and reaction energies
[kcal rnol-'1 for the rearrangements I a 2a (Ref. [14d]) and 17 18. R = H
(Ref. [77a]).
the substituents and is diminished by n-donor ligands. In
agreement with these findings, the thermal rearrangement
does not occur in solution or in the matrix. In contrast, the
rearrangement of silanone 0-oxides 17 to dioxasiliranes 18
exhibits an activation barrier of only 6.5 kcalmol-' and,
consequently, the formation of the three-membered ring 18
is observed.[77d1
A n g m Chem Int Ed. Eng1 29 11990) 344-354
Since 1 reacts quickly with alcohols[7'I and aldehydes,[681
an important step in the Criegee mechanism of ozonolysis
could be verified by direct observation. A reaction competing with the formation of trapped species is that of 1 with
itself to yield 5 or 8 and 0,.127.281
Irradiation of 1 with red light (presumably into an n + n*
transition) will either lead to the rearranged species 2 or to
the formation of 8 with concomitant extrusion of 0 atoms.
The ratio of 2 to 8 is again influenced by the substituent on
1 . For example, 1 m rearranges completely to 2m, whereas
the aldehyde 0-oxides (1 h and 1 i) afford mainly 8. Therefore, one can selectively generate either 1 , 2 , or free 0 atoms
just by choosing the appropriate substituents and reaction
conditions. 0 atoms in their O(3P) state (ground state) are
only detected indirectly in the matrix. A reaction easily monitored by IR spectroscopy is the oxidation of arylaldehydes
to carboxylic acids. This insertion presumably proceeds via
intermediate formation of benzoyl and hydroxyl radicals,
which recombine instantly to yield carboxylic acids. The activation barrier for the reaction of O(3P) with acetaldehyde
was determined in molecular beam experiments (gas phase)
to be 2.4 kcal mol- by Kleinermanns and Luntz.["] In analogy to the H-atom abstraction reactions of free carbenes at
cryogenic temperatures[821described by Platz, these H-atom
abstractions may also be explained by invoking a quantummechanical tunneling effect.
Dioxiranes 2 are much more photostable than carbonyl
oxides I , so that irradiation with blue or UV light is necessary to form 10. The rearrangement proceeds by primary
cleavage of the weak 0-0 bond (activation barrier 1215 kcal mol- ')[' 'I and subsequent migration of a substituent
R (Scheme 4). An extrusion of 0 atoms from 2 and formation of 8 was not observed.
5.2. Chemiluminescence
An interesting phenomenon in carbene oxidations, discovered by Trozzolo et al., is chemiluminescen~e.[~~I
On warming a solution of carbene 7 b (generated by photolysis of 9b)
in an organic glass in the presence of O,, an intense blue light
is emitted. A similar phenomenon was observed by Wasserman et al. when bis(trifluoromethy1)diazomethane (91) was
p h o t o ~ x i d i z e d . [ ~In~ ]both cases the chemiluminescence
spectra match the phosphorescence spectra of the corresponding carbonyl compounds 8 b and 81. Turro et al. were
able to show that chemiluminescence is a general occurrence
in the reaction of arykarbenes with organic glasses as long as
the corresponding carbonyl compound 8 also shows phosphore~cence.[*'~
To explain these observations, a mechanism
involving the transfer of a terminal 0 atom of 1 onto a
molecule of 0, was proposed [Eq. (14)], although it was not
possible to detect ozone and 1 in these experiments.t851
An alternative explanation for this chemiluminescence
was provided by Sawaki and I w a m ~ r a . [ ~They
~ " ' proposed
that the decomposition of cyclic peroxides [Eq. (7)] may be
the cause of chemiluminescence. However, their hypothesis
was not substantiated by any experiments.
In all cases the chemiluminescence spectrum is equivalent
to the phosphorescence spectrum of the carbonyl compound
generated by carbene oxidation. It should be noted that for
A n g w (%cm In[. Ed. En,qI. 29 (1990) 344-354
aryl ketones spin inversion (ISC) of the electronically excited
singlet state to the triplet state is extremely effective, so that,
on excitation with light, only phosphorescence and no fluorescence is emitted.[*"' Therefore, the observation of phosphorescence does not prove that the carbonyl compound is
formed directly in the triplet state. Considering the case of
bis(trifluoromethy1)methylene (7 I), where, on oxidation, the
chemiluminescence corresponds solely to the phosphorescence of 81 (on excitation with light, both fluorescence and
phosphorescence are observed),[871one can conclude that
hexafluoroacetone (8 1)-and most probably also the other
ketones-is formed directly in its triplet
Possible reaction sequences that can lead to triplet carbony1 compounds-and
thus to chemiluminescence-are
listedin Equations (1 1)-(14). Reactions (1 1)-(13) areexamples of the transfer of 0 atoms from a variety of sources onto
a free carbene. Reaction (14) corresponds to the mechanism
proposed by Turro et al.[851
-
RR'C:
+
RR'C:
+ O('P)
RR'C:
+ 02=CRR'
RR'C=O,
'02
+ '02
+
[O]
'RR'C=O
+
'RR'C=O
-
'RR'C=O
-
'RR'C=O
O=CRR'
+ 03
In the matrix, chemiluminescence occurs immediately after warming to temperatures above 10 K. At temperatures
lower than 25 K, even the diffusion of small molecules such
as 0, is
The observation of chemiluminescence at
temperatures below 20 K cannot be explained by reactions
(II), (13), and (14). Reaction (14) can also be excluded on
other grounds. First, ozone is formed only as a byproduct at
high 0, concentrations, whereas reaction (14) requires the
formation of equimolar amounts. Second, chemiluminescence already commences at temperatures below 20 K, when
no carbonyl oxide has been formed in the matrix yet, and
subsides after only a few minutes at 30 K, even though plenty
of carbonyl oxide is now present in the matrix and the mobility of 0, is very Iarge.[50*51*641
To exclude 0, as a reaction component, the rate of formation of carbonyl oxides (a reaction in which 0, participates)
was compared to the rate at which the intensity of chemiluminescence decreases a t constant temperature. It could be
shown that chemiluminescence decreases four times as fast
as carbonyl oxides are formed. It seems reasonable to conclude that the occurrence of chemiluminescence arises from
diffusion of a species smaller than 0, .I' 'I Finally, reaction
(11) is endothermic according to thermochernical estimates-at least for 7h-and may also be
The best explanation for all of the observations made is
the presumption that the reaction of an 0 atom with 7 is the
cause of chemilurninescence [Eq. (12)]. The multiplicity of
the carbene (triplet or singlet) does not have much influence
on this reaction and the carbonyl compounds are formed
directly in their triplet states.
351
5.3. Influence of Substituents
Table 2. Comparison of MP2/6-31G* vibrational frequencies of 1a with experimental frequencies of 1 m.
5.3.1. I R Spectra
A characteristic feature of the IR spectrum of 1 is its
intense bands between 890 and 1050cm-’, which show a
large 160-‘80
isotopic shift (30-60 cm-I) and are assigned
to 0-0 stretching modes (Table 1). In some cases, a splitting
of the bands into two components is observed. This can be
attributed either to Fermi resonance or to matrix cage effects. When 1 is unsymmetrically substituted, the occurrence
of two bands could also be caused by the existence of syn and
anti isomers. However, there is no experimental evidence for
this suggestion.
According to the simple model for a diatomic harmonic
oscillator, the highest possible isotopic shift for the 0-0
vibration amounts to 5.7 YOif both l60atoms are replaced
by I8O atoms. This maximum value is almost reached by the
values monitored for l g (5.6%) and l m (5.5% and 5.7%;
Table 1). These carbonyl oxides exhibit nearly pure 0-0
vibrations without coupling with other vibrations. For 1 b
with its isotopic shift of 3.9% for the band at 896 cm- the
calculated contribution of the 0-0 vibration is 68%. The
estimated force constant is typical for 0-0 single bonds.[721
Therefore, 0-0 bonds in 1 should best be described as single
bonds.
The relationship between the position of the 0-0 stretching mode and the substituents of 1 becomes evident from
Table 1. Electron-withdrawing groups shift the bands to
higher frequencies. Compared to a phenyl group, a CF,
group induces a shift of 50 cm-’. Even more effective is the
dienone moiety as x acceptor in the quinone 1 m: here, the
band is shifted by 138 cm- to higher frequencies as compared to 1 b. On the other hand, replacing a phenyl group by
an H or C1 atom hardly influences the 0-0 vibration at all.
The 0-0 bond is stabilized by electron acceptors, x acceptors giving rise to larger effects than o-acceptors.
A weak IR band in the region between 550 and 740 cm-’
also shows a marked 60-‘8Oisotopic shift (0.7- 1.9 %) and
is assigned to an 0-0 deformation ~ i b r a t i o n . ~721~
Whereas the 0-0 stretching vibration of 1 is a characteristic group frequency, it is difficult to unequivocally assign
the C-0 stretching vibration to any of the observed bands.
In the “carbonyl region” no absorptions are registered. The
C-0 stretching vibration is strongly coupled with other vibrations, is of low intensity, and-in comparison to C = O
vibrations of carbonyl compounds-is shifted to lower wave
numbers. Thus, the electronic structure of a carbonyl oxide
group is characterized by a C-0 bond of low polarity and a
C-0 bond order of less than 1.5 and is fundamentally different from that of a normal carbonyl group.
To verify the assignments mentioned above, a comparison
of the experimental data with quantum-mechanical calculations seems in order. Unfortunately, the calculation of IR
spectra by exact a b initio methods can only be achieved
through great computational effort, so that, to date, only
calculations on 2a have been performed on this scale
(Table 2).[14c1The comparison with the experimental data
for l m indicates that not only the position but also the
amount of the isotopic shift of the 0-0 stretching mode and
the 0-0 deformation vibration is well reproduced by the
calculations.
’,
’
‘ 7
352
Assignment
v,/G(COO)
rs/v(C0)
v,/v(OO)
r,/&(COO)
la
v ( 0 - 0 ) [a]
[cm-’1
A [b]
[cm-’1
lm
v ( 0 - q ) [cl
[cm- ]
1222 (7)
1116 (52)
1028 (100)
522 (1.2)
-32
-22
-49
-18
1241 (m)
1327 (w)
1034 (vs)
570 (w)
(2.6)
(2.0)
(4.8)
(3.4)
A [bl
[cm-’1
-6
(0.5)
-16 (1.2)
-57 (5.5)
-11 (1.9)
[a] Ab initio frequencies of 1a scaled by 0.93; in brackets, relative intensities. [b]
160-’80
isotopic shift; in brackets, relative shift in %. [c] Experimental data for
1 m; in brackets, relative intensities.
5.3.2. UV Spectra
In the near-UV and visible region of the spectrum of 1
(2 = 380-460 nm) one observes intense and broad absorptions that, in somecases, also exhibit fine structure (Table 1).
For l i the fine structure is caused by the 0-0 stretching
vibration of the COO group.1641Owing to this absorption in
the visible region of the spectrum, carbonyl oxides isolated in
a matrix are yellow to orange, except for 1 j, which is blue.
In order to calculate the UVjVIS transitions, the semiempirical CNDOjS method[881is appropriate, since it is ideally
suited to compute x x* and n + n* transitions (Table 1).
According to the CNDOjS calculations, one expects two
characteristic absorptions for carbonyl oxides in the visible
region of the spectrum: a very strong x + x* transition and
a weak n + n* transition.[’71
The x x* transition with IgE zz 4 is the strongest transition in the spectrum and is situated between 378 and 582 nm
depending on the substituents (Table 1). Because of the
strong intensity and the favorable position of this band in the
spectrum, it is especially important for the detection of carbony1 oxides. The comparison of the calculated data with
experimental results derived from matrix and laser photolysis studies (Table 1) reveals an excellent agreement between
theory and experiment. The largest discrepancy in the calculated and experimental maxima is observed for 1b (33 nm).
The position of the x + n* transition depends on the size
of the x system and the electronic properties of the substituents (n donor or n acceptor). Large conjugated x systems such as those in l e , l c , l m , o r I p induce a redshift
when compared to 1 a. However, a large redshift is found for
1 r, which features only a small x system (Table 1). Consequently, x donors (like the 0 atoms in l r ) also lead to a
redshift of the x + x* transition. The opposite is true for
electron acceptors: the x + x* transition of shortest wavelength is observed at 378 nm for 1 k.
The influence of the n system becomes evident from the
comparison of 1 c and 1 p with their benzoannelated derivative 1 e and 1 q (Table I).[’ 71 Compared to 1 a, the tropylidene
group (a x donor) in 1 p induces a redshift of 162 nm. In
contrast, the cyclopentadienyl group in l c shifts the band
into the red by only 40 nm. Benzoannelation leads to reverse
effects in both compounds: in 1 p to a blueshift (from 546 nm
in 1 p to 493 nm in 1 q) and in 1 c to a redshift (from 424 nm
in 1 c to 489 nm in 1 e). This can be explained by the larger
x systems in the benzoannelated derivatives, in that they can
delocalize the charge more effectively and consequently reduce the influence of the tropylidene o r cyclopentadienyl
group on the COO moiety. The calculated redshift of the
-+
-+
Angew. Chem. Inl. Ed. Engl. 29 (1990) 344-3S4
n + n* transition induced by n
was impressively
verified by experiment: the carbonyl oxide 1j, which contains the bicyclo[6.3.0]undecapentaenyl framework as a %
donor, is blue and exhibits the highest long-wavelength absorption, with a maximum at 582 nm.[73J
The calculated, extremely weak n + n* transitions are situated in the red part of the spectrum (598 -895 nm) and have
not actually been measured u p to now. An indirect indication of their existence is the photochemistry carried out a t
long wavelengths. The rearrangement of 1 to 2 on irradiation
at wavelengths lying far from the n + n* transition can best
be explained by a long-wavelength transition of low intensity.
5.3.3. Caicuiations
Whereas only a few simply substituted carbonyl oxides
have been subject to calculation by a b initio methods
(la[13. 14a.14d3 15.89.901 and 1 ~ [ 1 4 d . 3 0 1as well as some alkyl51),
the semiempirical
substituted derivatives['4a.
M I N D 0 / 3 method[9'*921 allows the inclusion of larger substituents." 71 Nonetheless, just as in the a b initio methods, the
semiempirical calculations require that one exceed the R H F
level and perform 2 x 2 CI['61 or U H F calculations.['7, 931
The calculated spin densities indicate that in n-substituted
carbonyl oxides-like 1 b and 1h or the cyclic carbonyl oxides 1 m, 1c, and1 p-the wave functions have diradical character.['71 The spin polarization of the n system and the 0-0
bond length (Table 3) are influenced by the substituent only
to a minor extent. The influence of the phenyl group largely
manifests itself in the extension of the n system and transfer
of some n-electron density onto the COO group. The result
is an enhanced p,-p, repulsion in the 3-center 4n-electron
system and thus a lengthening of the C1-01 and 01-02
bond.
14d3
calculated at 1.28 A (Table 3) but is in agreement with the
experimental results. The 0-0 stretching vibration is shifted
to higher wavelengths by electron acceptors, which indicates
a stabilization of the 0 1 - 0 2 bond.
A more dramatic influence on the wave function is calculated for electronegative elements such as 0 in 1 r and F in
1s. In these cases only a small amount of spin polarization of
the n system remains and the U H F result is identical to the
R H F result to a high degree; the wave function is that of a
zwitterion. Carbon atom C1 bears a large positive charge in
1r and 1s (Table 3).
From the M I N D 0 / 3 calculations some general conclusions can be drawn:
n acceptors lead to a transfer of n-electron density from
the COO group onto the substituent. In consequence, the
COO group is stabilized.
n donors lead to a transfer of n-electron density of the
substituent onto the COO group. Thus, the COO group is
destabilized.
The C-0 bond length is more strongly influenced by substituents than the 0-0 bond length.
Strong o-acceptors (e.g., 0 or F) can compensate for the
n-electron effects and, in particular, may lead to a more
stable zwitterionic state.
On comparing theoretical and experimental results it is
important to consider the influence of solvent effects. Polar
solvents can stabilize zwitterionic states more strongly than
diradical states and can therefore counteract the substituent
effects discussed above.['31
6. Conclusion and Outlook
[a] Number o f n electrons in the COO group. [b] Positive values correspond to an excess
of a-spin density. negative values to an excess of b-spin density.
The experimental and theoretical research discussed in this
review gives a relatively consistent picture of the electronic
structure of carbonyl oxides. Thus, carbonyl oxides in the
ground state in the matrix are polar diradicals. This form
should not be mistaken with a zwitterionic state, which lies
a t higher energy. the 0-0 bond is best described in terms of
a polar single bond, while the C-0 bond is less polar and
exhibits a bond order of ca. 1.5. Considering that the absorption spectra of carbonyl oxides are very similar in the matrix
and in solution at room temperature (Table 1 ) and that the
absorption maxima are only slightly influenced by the polarity of the solvent, it can be concluded that, also in solution,
the ground state is best described as a diradical species. Once
n-electron donors are introduced (e.g., as in l r ) carbonyl
oxides with a zwitterionic ground state should arise. The
experimental verification of a carbonyl oxide with such a
zwitterionic ground state has yet to be achieved.
In the cyclic carbonyl oxides the n-electron density in the
COO group can either be enhanced (ring as n donor as in 1 p
or 1 g) or diminished (ring as n acceptor as in 1e, 1c, or 1m);
n donors effect a lengthening of the C1-02 bond, n acceptors
a reduction. The stabilization of the COO group by n acceptors and its destabilization by 7[: donors is not reflected by the
calculated 0 1 - 0 2 bond lengths, which are almost constantly
I thank m y co-workers G. Bucher, A . Patyk, and A . Runge,
whose particular dedication-often at unearthly times of' the
day-made a completion of our work in the field of matrix
spectroscopy possible. M y thanks are also due to Prof: Dr. R.
Gleiterfor generous support over the last years, Dr. P. Bischof
for assistance in theoretical matters, Prof. Dr. D. Cremer,fi?r
excellent cooperation on the calculation of the properties of'
carbonyloxides, and Prof: Dr. B. Giese,for his help concerning
instrumental aspects.-This
work was supported by the
Table 3. Ground-state properties of some cdrbonyl oxides. calculated with the
MINDOI3-UHF procedure (bond lengths R, bond angles, atomic charges, p,-spin
densities).
R(CO) R ( 0 O ) <COO
[A1
IAI
["I
1269
1.280
1.280
1.280
1.279
1.278
1.279
I277
1.293
1270
124.0
124.9
122.9
124.0
125.2
123.0
123.9
130.0
115.5
132.6
Atomic charge
O(1)
O(2)
N , fa1
~~~~~
1.281
1.313
1.305
le
1.305
s v n - l h 1.298
lm
1.325
lp
1.329
lq
1.321
Ir
1.348
Is
1.294
la
Ib
Ic
0.074 0.183
0.124 0.087
0.203 0.084
0.150 0.090
0.113 0.118
0.167 0.069
0.127 0.071
0.126 0.065
0.769 -0.056
1.015 -0.123
-0.331
-0.298
-0.256
-0.281
-0.296
-0.232
-0.273
-0.285
-0.282
-0.111
A n g m Clirm. In!. Ed. EngL 29 (1990) 344-354
4.000
4.004
3.928
3.969
4.006
3.928
4.038
4.010
4.276
4.202
0.653
0.640
0.528
0.576
0.619
0.528
0.549
0.585
0.026
0.044
-0.253
-0.289
-0.278
-0.286
-0.285
-0.278
-0.312
-0.302
-0.015
-0.034
-0.400
-0.497
-0.582
-0.521
-0.491
-0.582
-0.536
-0 521
-0.006
-0.008
353
Deuische Forschungsgemeinschaft ( “Noncovalent Interactions” and SFB 247), the Stiftung Volkswagenwerk, and the
Fonds der Chemischen Industrie.
Received: October 11, 1989 [A757 IE]
German version: Angew Chem. 102 (1990) 362
[l] J. A. Miller, G. A. Fisk, Chem. Eng. News 65 (1987), No. 35 from August
31, 1987, p. 22.
[2] R Criegee, Angew. Chem. 87 (1975) 765; Angew. Chem. Inr. Ed. Engl. 14
(1975) 745
W. Adam, R. Curci, J. 0. Edwards, Acc. Chem. Res. 22 (1989) 205.
R. W. Murray, Chem. Rev. 89 (1989) 1187.
R. W. Murray, V. Ramachandrdn, Photochem. Photobiol. 30 (1979) 187.
T. A. Dix, S. Benkovic, Ace. Chem. Rex 21 (1988) 101.
R. W. Murray, Mol. Struct. Energ. 6 (1988) 311.
R. Criegee, G. Wenner, Justus Liebigs Ann. Chem. 564 (1949) 9.
T. Sugawara, H. Iwamura, H. Hayashi, A. Sekiguchi, W Ando. M. T. H .
Liu, Chem. Lett. 1983, 1261.
D. Cremer, M. Schindler, Chem. Phys. Lett. l33 (1987) 293.
W. Adam, Y Y Chan, D. Cremer, J. Gauss, D. Scheutzow, M. Schindler,
J. Org. Chem. 52 (1987) 2800.
Y. Sawaki. H . Kato, Y. Ogata, J Am. Chem. Soc. 103 (1981) 3832.
L. B. Harding, W. A. I. Goddard, J. Am. Chem. Soc. 100 (1978) 71 80.
a) D. Cremer, J Am. Chern. Soc. 101 (1979) 7199; b) ibid. 103 (1981) 3627;
c) J. Gauss, D. Cremer, Chem. Phys Lert. 133 (1987) 420; d) D. Cremer,
T. Schmidt, J. Gauss, T. P. Radhakrishnan, Angew. Chem. 100 (1988) 431;
Angew Chem. Int. Ed. EngI. 27 (1988) 427
S. A. Kafafi, R. 1. Martinez. J. T. Herron, Mol. Struct. Energ. 6 (1988) 283.
L. A. Hull, J Org. Chem. 43 (1978) 2780.
D . Cremer, T. Schmidt, W. Sander, P. Bischof, J Org. Chem. 54 (1989)
251 5.
R. L. Kuczkowski (Ozone and Carbonyl Oxides) in A. Padwa (Ed.): 1.3Dipolar Cycloaddition Chemistry, Wiley, London 1984, Chapter 11.
P. S . Nangia, S. W. Benson, J Am. Chem. Soc. 102 (1980) 3105.
P. S. Bailey: Ozonution in Organic Chemistry. Academic, New York 1982.
J. Z. Gillies, C. W Gillies, R. D. Suenram, F. J. Lovas. J Am. Chem. Soc.
110 (1988) 7991.
P. S. Bailey, A. Rustaiyan, T. M. Ferrell, J Am. Chem. Soc. 98 (1976) 638.
D . Cremer, Angew. Chem. 93 (1981) 934; Angew Chem. Int. Ed. Engl. 20
(1981) 888.
R. D. Suenram, F. J. Lovas, J. Am. Chem. Soc. 100 (1978) 51 17.
D. P. Higley, R. W. Murray, J Am. Chem. Soc. 96 (1974) 3330.
W. Kirmse, L. Horner, H. Hoffmann, Chem. Ber. 614 (1958) 19.
P. D. Bartlett, T. G. Traylor, J Am. Chem. Soc. 84 (1962) 3408.
M. Girard, D. Griller, J Phys. Chem. 90 (1986) 6801.
a) K . Ishiguro, K. Tomizawa, Y. Sawaki, H . Iwamura, TerrahedronLett. 26
(1985) 3723; b) K . Ishiguro, Y. Hirdno, Y Sawaki, J Org. Chem. 53 (1988)
5397.
M. Rahman, M. L. McKee, P. B. Shevlin, R. Sztyribicka, J Am. Chem.
Soc. 110 (1988) 4002.
W. A. Pryor, C. K. Govindan. J Am. Chern. Soc. /U3 (1981) 7681.
T.A. Hinrichs, V. Rarnachandran, R . W Murray, J Am. Chem. Soc. 101
(1979) 1282.
a) W. Adam, W. Haas, G. Sieker, J Am. Chem. Soc. 106 (1984) 5020; b) W.
Adam, H. Diirr, W. Haas, B. Lohray, Angew. Ckem. 98 (1986) 8 5 ; Angew.
Chem. Inr. Ed. Engl 25 (1986) 101
S. K. Agarwal, R. W. Murray, Isr. J Chem. 23 (1983) 405.
S. Kumar. R. W. Murray, J Am. Chem. Soc. 106 (1984) 1040.
R. W. Murray, S. Kumar (Oxidarion of Phenanrhrene with a Carhonyl
O-xide) Polynucl. Aromat. Hvdrocarbons: Int. Symp. 7th. 1982, Butterworth, Boston, USA 1983. pp. 575-581.
W. H. Bunnelle, E. 0. Schlemper, J Am. Chem. Soc. 109 (1987) 612.
K. Griesbaum, W. Volpp, R. Greinert, J Am. Chem. Soc. 107(1985) 5309.
H. Keul, R. L. Kuczkowski, H.-S. Choi, J Urg Chem. 50 (1985) 3365.
M. Mori, M. Nojimd, S. Kusabdyashi, J Am. Chem. Soc. 109 (1987) 4407.
P. R Story, J. R. Burgess, J. Am. Chem. Soc. 89 (1967) 5726.
S. K. Agarwal, R. W. Murray, Photochem. Photobiol. 35 (1982) 31.
E. Whittle, D. A. Dows, G. C. Pimentel, J Chem. Phys. 22(1954) 1943.
1. Norman, G Porter, Nature 174 (1954) 508.
A. J. Barnes, W. J. Orville-Thomas, A. Miiller, R. Gaufres: Matrix Isolation Spectroscopy: Nato AdvancedStudy Insritutes Series, Series C , Vo1. 76,
Reidel, Dordrecht 1981.
3 54
1461 B. Meyer: Lon Temperature Spectroscopy. American Elsevier, New York
1971.
[47] R. J. McMahon, C. J. Abelt. 0 . L. Chapman, J. W. Johnson, C. L Kreif.
J. P. LeRoux, A. M Mooring, P. R. West, J Am. Chem. Soc. 109 (1987)
2456.
1481 G. C . Pimentel (Radicul Formation and Trapping in the Solid Phase) in
A. M. Bass, H. P. Broida (Eds.): Formation and Trupping ofFree Radicals,
Academic, New York 1960. pp. 69- 1 15.
[49] G. A. Bell, 1. R. Dunkin, J. Chem. Soc. Chein. Commun. 1983, 1213.
[SO] W Sander, J. Org Chem. 53 (1988) 121.
[51] W. Sander. J Org. Chem. 54 (1989) 333.
[52] a) I. R. Dunkin. C. J. Shields, J Chem. Sol. Chem. Commun. 1986, 154; b)
I. R. Dunkin. G. A. Bell, Tetrahedron 41 (1985) 339.
[53] D. Griller. A. S. Nazran, J. C. Scaiano, Ace. Chem. Res. 17 (1984) 283.
1541 E. P. Wasserman, R G. Bergman. C. B. Moore. J Am. Chem. Soc. 110
(1988) 6076.
[55] M. A Young, G. C. Pimentel, Appl. Opt. 28 (1989) 4270.
1561 E. Weitz, J Pkvs. Chem. 91 (1987) 3945.
[57] a) H. Hermann. F.-W. Grevels, A. Henne. K . Scbaffner, J Phys. Chem. 86
(1982)5151; b) K.Schaffner, F-W. Grevels,J Mol.Struer. 173(1988)51.
1581 0 . L.Chdpmdn, Pure Appl. Chem. 51 (1979) 331.
[59] T. C. Hess: Low-Temperature Phorochemistry: Photooxidation of Kerenes
and Diazo Compounds Isolated in Ox-vgen Doped Argon Matrices at 10 K ,
Ph.D. Thesis, University of California 1978.
1601 0. L. Chapman, T. C. Hess, J Am. Chem. Soc. 106 (1984) 1842.
1611 a) G. A. Bell, I. R. Dunkin, C. J. Shields, Specrrochim. Acta. Part A 4fA
(1985) 1221; b) I. R. Dunkin, G . A. Bell, F. G. McCleod, A. McCluskey,
ibid. 42A (1 986) 567.
1621 H . L. Casal, M. Tanner, N. H. Werstiuk, J. C . Scaiano, J Am. Chem. Sot.
107 (1985) 4616.
[63] W. Sander, Angew Chem. 98 (1986) 2 5 5 ; Angew. Chem. Int. Ed. Engl. 25
(1986) 255
1641 W. W. Sander, Snecrrochim. Acru. Part A 43A (1987)
.
, 637.
G. A. Ganzer, R. S. Sheridan, M. T. H. Liu. J Am. Chem. Soc. 108 (1986)
1517.
N. H. Werstiuk, H. L. Casal, J. C. Scaiano. Can J Chem. 62 (1984) 2391.
H. L. Cdsal, S. E. Sugamori, J. C. Scaiano, J Am. Chem Soc. 106 (1984)
7623.
J. C. Scaiano, W. G. McGimpsey, H. L. Casal, J. Org. Chem. 54 (1989)
1612.
R. W. Fessenden, J. C. Scaiano, Chem. Phys Lett. 117 (1985) 103.
R. L. Barcus. L. M. Hadel, L. J. Johnston, M. S. Platz. T. G. Savino. J. C.
Scaiano. J Am. Chem. Soc. 108 (1986) 3928.
Y Fujiwara, Y Tanimoto, M. Itoh, K . Hirai, H . Tomioka, J Am. Chem.
Soc. 109 (1987) 1942.
W. Sander, J. Org. Chem. 53 (1988) 2091
S . Murata, H. Tomioka, T. Kawase, M. Oda, unpublished.
Y-P. Lee, G. C. Pimentel. J Chem. Phw. 74 (1981) 4851.
P. P. Gaspar, D. Holten, S. Konieczny, Arc. Chem. Res. 20 (1987) 329.
T. Akasaka, S. Nagase, A. Yabe. W. Ando, J. Am. Chem Soc. 110 (1988)
6270.
a) A. Patyk, W. Sander, J. Gauss. D. Cremer, Angew. Chem. 101(1 989) 920;
Angew Chem. Int. Ed. EngI. 28 (1989) 898; b) A. Patyk, W. Sander,
J. Gauss, D. Cremer, Chem. Ber. 123 (1990) 89.
D. Griller, A. S. Nazran, J. C. Scaiano, Tetrahedron 41 (1985) 1525
I. R. Could, N. J. Turro, J J. Butcher, C. J. Doubleday, N. P. Hacker, G. F.
Lehr, R. A. Moss, D. P. Cox, W. Guo, R. C. Munjal, L. A. Perez, M.
Fedorynski, Tetrahedron 41 (1985) 1587.
W Sander, Angew. Chem. 97 (1985) 964; Angew. Chem. Int. Ed. Engl. 24
(1985) 988.
K. Kleinermanns, A. C. Luntz, J Chem Phvs. 77 (1982) 3774.
M. S. Platz, Acc. Chem. Res. 21 (1988) 236.
A . M . Trozzolo. R. W. Murray, E. Wdsserman, J. Am. Chem. Sor. 84
( I 962) 4990.
E. Wasserman, L. Barash, W. A. Yager, J Am. Chem. SOC.8 7 (1965) 4974.
N J. Turro, 1. A. J. Butcher, G. J. Hefferon, Photochem. Pholobiol 34
(1981) 517
R. S. Becker: Theory und Interpretation of Fluorescence and Phosphorescence, Wiley. New York 1969.
A Gandini, K. 0. Kutschke. Proc. R. Soc. A. 306 (1968) 511.
J. Del Bene, H. H. Jaffi, J Chem. Phys. 48 (1968) 1807.
G. Karlstrom, S. Engstrom, B. Jonsson. Chem. Phys. Lett. 67 (1979) 343.
K. Yamaguchi, S. Yabushita, T. Fueno, S . Kato, K . Morokuma, S . Iwata.
Chem. Phw. Letr. 71 (1980) 563.
R. C . Bingham. M. J S. Dewar, J. Am. Chem. Soc 97 (1975) 1285.
D. F. V. Lewis, Chem. Rev. 86 (1986) 1111.
P. Bischof, J Am. Chem. Soc. 98 (1976) 6844.
Angew. Chem. Int. Ed. Engl. 29 (1990) 344-354
Документ
Категория
Без категории
Просмотров
1
Размер файла
1 253 Кб
Теги
oxide, carbonyl, diradicals, zwitterion
1/--страниц
Пожаловаться на содержимое документа