New formylating agents - Preparative procedures and mechanistic investigations

Article abstract of DOI:10.1002/1099-0690(200108)2001:15<2947::AID-EJOC2947>3.0.CO;2-R

The reactivity of new formylating agents related to formamide has been investigated both experimentally and theoretically. The reaction in 1,2-dichloroethane between tris(diformylamino)methane (2) and several arenes, catalyzed by AlCl₃ or BCl₃, was shown to proceed in good yields to afford the corresponding para-substituted aldehydes. The nature of the active electrophilic species was also investigated theoretically. Thus, the relative stability of the O- and N-protonated forms, as well as those of AlCl₃ adducts, of several formylating agents - diformamide, triformamide, N,N,N′,N′-tetraformylhydrazine, and tris(diformylamino)methane - were determined in the gas phase and in water or DCE by means of DFT calculations at the B3LYP/6-311++G(d,p) level, the solvents being modeled with the IPCM method. The amide oxygen atom in all cases appeared to be the most basic site, both in the Bronsted and Lewis sense, constituting a first step towards the understanding of the mechanism of this reaction.

Full text of DOI:10.1002/1099-0690(200108)2001:15<2947::AID-EJOC2947>3.0.CO;2-R

FULL PAPER
New Formylating Agents ؊ Preparative Procedures and Mechanistic
Investigations[‡]
Alessandro Bagno,*[a] Willi Kantlehner,*[b] Oliver Scherr,[c] Jens Vetter,[c] and
Georg Ziegler[c]
Keywords: Basicity / Density functional calculations / Electrophilic aromatic substitution / Formylation / Lewis acids
The reactivity of new formylating agents related to formam- ing agents − diformamide, triformamide, N,N,NЈ,NЈ-tetrafor-
ide has been investigated both experimentally and theoretic- mylhydrazine, and tris(diformylamino)methane − were deter-
ally. The reaction in 1,2-dichloroethane between tris(difor-
mined in the gas phase and in water or DCE by means of
mylamino)methane (2) and several arenes, catalyzed by DFT calculations at the B3LYP/6-311++G(d,p) level, the solv-
AlCl3 or BCl3, was shown to proceed in good yields to afford
the corresponding para-substituted aldehydes. The nature of
ents being modeled with the IPCM method. The amide oxy-
gen atom in all cases appeared to be the most basic site, both
the active electrophilic species was also investigated theoret- in the Brønsted and Lewis sense, constituting a first step to-
ically. Thus, the relative stability of the O- and N-protonated
wards the understanding of the mechanism of this reaction.
forms, as well as those of AlCl3 adducts, of several formylat-
Introduction
mamides and phosphorus oxychloride or phosgene, and (b)
reactions yielding primary products which are immediately
oxidized to aldehydes, such as the synthesis of salicylal-
dehydes from phenols and paraformaldehyde in the pres-
ence of SnCl4/trioctylamine[1,4] or the Duff reaction,
yielding aromatic aldehydes from activated aromatic com-
pounds and urotropine in the presence of acids.[1,5]
Nowadays, the Vilsmeier (or VilsmeierϪHaack) reaction
is the most common method for formylation of aromatic
rings. Nevertheless, it, like the other methods, has some sig-
nificant disadvantages. The formylating agents in these re-
actions are often toxic (as in the cases of carbon monoxide,
hydrogen cyanide, formyl fluoride, and phosgene), and they
are also difficult to handle in large amounts. Furthermore,
the scope of these reactions is limited. The
VilsmeierϪHaack reaction is the method with the widest
scope, but it is incapable of formylating simple alkylarenes
(unless they are much more reactive than benzene). A fur-
ther difficulty is the occurrence of phosphorus compounds
in wastewater, which is a serious environmental problem,
Despite the great importance of (hetero)aromatic alde-
hydes as intermediates in the chemical and pharmaceutical
industries, and the continuing intense research in this field,
there is still a need for new synthetic methods for the intro-
duction of the aldehyde group into (hetero)aromatics. The
introduction of a formyl group by CϪC bond formation is
generally achieved by means of various kinds of elec-
trophilic aromatic substitutions, which can be subdivided
further into two types of reactions: (a) reactions involving
the acid-promoted generation of a formyl cation or a pre-
cursor thereof, such as formyl fluoride/boron trifluoride
(Olah formylation),[1] CO/HCl (GattermannϪKoch reac-
tion),[1,2] CO/HF,[1] HCN/HCl (Gattermann reaction),[3]
and the Vilsmeier reagent,[1] formed from disubstituted for-
[‡]
Presented in part at the Symposium of the Liebig Society for
Organic Chemistry and the subject section Ecotoxicology of the
GDCh, Tübingen, August 23Ϫ25, 1999, the 7th European
Symposium on Organic Reactivity (ESOR 7), Ulm, August and the formation of highly toxic, carcinogenic dialkylcar-
22Ϫ27, 1999, and the 4th Iminium Salt Conference,
bamoyl chloride in a side reaction. The dichloromethyl
methyl ether/Lewis acid formylating system has an even
Rechenberg/Stimpfach, September 14Ϫ16, 1999.
[a]
[b]
[c]
Centro CNR Meccanismi Reazioni Organiche, Dip. Chimica
wider scope than the VilsmeierϪHaack reagent, but it has
`
Organica, Universita di Padova
not yet found practical use.[1,3,6]
Via Marzolo 1, 35131 Padova, Italy
Fax: (internat.) ϩ 39-049/827-5239
E-mail: alex@chor.unipd.it
Formamide Derivatives as Formylating Agents
Universität Stuttgart, Institut für Organische Chemie und
Isotopenforschung
Some time ago, we reported on a new synthetic method
for the formylation of (hetero)aromatic compounds using
triformamide (1)/AlCl3 as the formylating system
(Scheme 1).[7,8] This reaction is useful for the formylation
of a wide range of substrates, including unsubstituted and
alkyl-substituted aromatics, aromatic ethers, tertiary aro-
matic amines, fused aromatic rings, and thiophenes.
Pfaffenwaldring 55, 70569 Stuttgart, Germany
Fax: (internat.) ϩ 49-711/685-4269
E-mail: willi.kantlehner@fh-aalen.de
Fachhochschule Aalen, Fachbereich Chemie/Organische
Chemie
Beethovenstraße 1, 73430 Aalen, Germany
Supporting information for this article is available on the
WWW under http://www.wiley-vch.de/home/eurjoc or from
the author.
Eur. J. Org. Chem. 2001, 2947Ϫ2954
WILEY-VCH Verlag GmbH, 69451 Weinheim, 2001
1434Ϫ193X/01/0815Ϫ2947 $ 17.50ϩ.50/0
2947

A. Bagno, W. Kantlehner, O. Scherr, J. Vetter, G. Ziegler
FULL PAPER
Results and Discussion
Formylations with Tris(diformylamino)methane
During the preparation of triformamide (1),[11,12] tris(di-
formylamino)methane[10] (2) is formed as a by-product.[8]
In an improved method, 2 can be prepared from tri-
formamide and sodium diformamide[13,14] in 80% yield.[8,15]
Tris(diformylamino)methane (2) is a stable, high-melting
(142 °C), nonhygroscopic, and nonetching compound. The
compound is nearly odorless and easy to handle. Although
toxicological data are not available, its manipulation
hazards may be estimated as similar to those of formamide.
Tris(diformylamino)methane (2) contains six formyl
groups, plus an additional one masked as an orthoamide
Scheme 1. New formylating agents based on formamide derivatives
In this paper we wish to introduce some other reagents function, which seems to set a record for this type of com-
based on formamide derivatives that can be used to formyl- pounds. In the presence of (strong) Lewis acids, it is capable
ate (hetero)aromatic compounds. We also present results of of formylating activated arenes.[8,15] The efficiency of 2 as a
computational investigations aimed at elucidation of the re- formylating agent is higher than that of triformamide (1),
action intermediates of these formylations.
since up to three out of the seven formyl groups in 2 can
The typical mechanism of electrophilic aromatic substitu- be used synthetically, whereas 1 can transfer only one of its
tion reactions involves, as a preliminary step, the formation three formyl groups.
of the actual electrophilic species. In the case of amides, it
In the presence of AlCl3 or BCl3, alkylarenes and aro-
is assumed that the Lewis basic site is the oxygen atom, by matic ethers are formylated by tris(diformylamino)methane
analogy with the behavior towards protonation.[9] However, (2) in moderate to good yields (see Table 1). With AlCl3 as
there is no guarantee that the behavior towards the proton the Lewis acid, yields are highest if a tris(diformylamino)-
exactly matches that towards other Lewis acids. Thus, al- methane/AlCl3 ratio of 1:6 to 1:8 is used (see Table 2). This
though one can easily formulate the generation of the elec- corresponds to a ratio of 1:2 with regard to the number of
trophilic species as a Lewis acid/base process involving the formyl groups transferred, a value coincident with the ratio
amide oxygen atom as the basic site, the actual nature of found in the FriedelϪCrafts acylation of esters and anhyd-
the active species must be elucidated independently. For this rides. 1,2-Dichloroethane proves to be the best solvent for
purpose, we have carried out quantum chemical calcula- the reaction, although other solvents, such as chloroben-
tions aimed at quantitative determination of the extent of zene and carbon disulfide, are also suitable. Solvents con-
preference of formylating agents for protonation or Lewis taining nitro groups, such as nitromethane and nitroben-
acid complexation at either the oxygen or nitrogen atom in zene, seem to suppress the formylating reaction, probably
the amide group.
because of complexation of the Lewis acid.
Table 1. Formylation of activated aromatic compounds with 2/Lewis acid in 1,2-dichloroethane
Substrate
Lewis acid[a]
Reaction conditions[b]
Procedure/workup[b]
Yield [%][c]
b.p. [°C/Torr]
Ref.
Toluene
Cumene
tert-Butylbenzene
Hexylbenzene
o-Xylene
p-Cymene
Biphenyl
Diphenyl ether
Resorcinol dimethyl ether
AlCl3
AlCl3
AlCl3
AlCl3
AlCl3
AlCl3
AlCl3
AlCl3
AlCl3
Ϫ15Ǟ0 (20)
I
55
38
33
55
42[e]
45[f]
54
20
45[g]
84/12
204Ϫ205 °C/760 Torr[26]
103Ϫ104 °C/10 Torr[26]
128Ϫ129 °C/760 Torr[27]
70 °C/2·10Ϫ5 Torr[28]
223Ϫ225 °C/760 Torr[26]
125 °C/20 Torr[29]
Ϫ13ǞϪ1 (14)
Ϫ15ǞϪ1 (16)
Ϫ13ǞϪ1Ǟ20 (15/2)
Ϫ19ǞϪ1 (15)
Ϫ15ǞϪ3 (15)
Ϫ13ǞϪ1 (16)
Ϫ13Ǟ1 (16)
II/A
II/B
II/B
I
II/B
II/A
II/B
II/B
41Ϫ42/0.2[d]
63/0.2
90Ϫ93/0.2
96/0.2
50Ϫ51/0.2
102/0.2
98Ϫ101/0.2
93/0.2 (m.p. 68 °C)
184 °C/11 Torr[30]
158Ϫ159 °C/4 Torr[31]
165 °C/10 Torr[26]
Ϫ13Ǟ2 (15)
(m.p. 70Ϫ71 °C[31]
125 °C/20 Torr[29]
)
p-Cymene
Anisole
BCl3
Ϫ13ǞϪ1 (15)
Ϫ10Ǟ1 (14)
I
18[f]
20
50Ϫ51/0.2
54/0.02
[h]
BCl3
II/B
118Ϫ120 °C/13 Torr[5]
[a]
[b]
Molar ratio substrate/2/Lewis acid (3:1:6). Ϫ
Initial and final reaction temperature [°C], and reaction time [h] in parentheses. See
Exp. Sect. for synthetic and workup procedures. Ϫ [c] Product is always the p-substituted benzaldehyde. Yields are based on tris(diformyla-
mino)methane under the assumption that 1 mol of the reagent supplies 3 mol of ‘‘formyl groups’’. Ϫ
Product: 3,4-dimethylbenzaldehyde. Ϫ
[d]
[e]
nD20ϭ1.5390 (1.5301[26]). Ϫ
[f]
[g]
[h]
Product: 5-isopropyl-2-methylbenzaldehyde. Ϫ
Product: 2,4-dimethoxybenzaldehyde. Ϫ
Molar ratio substrate/2/Lewis acid (3:1:4).
2948
Eur. J. Org. Chem. 2001, 2947Ϫ2954

New Formylating Agents
FULL PAPER
another formamide derivative, N,N,NЈ,NЈ-tetraformyl-
hydrazine (4), are also encouraging.[19,20]
Table 2. Influence of molar ratio of reagents on yield of formylation
of o-xylene with tris(diformylamino)methane (2) in the presence of
AlCl3 (Procedure I)
Quantum Chemical Calculations
Molar ratio[a]
Yield%[b]
o-Xylene
AlCl3
The formamide derivatives discussed above (1Ϫ5) are ac-
tivated through complexation with a Lewis acid. All these
bases are structurally related to formamide, and contain at
least two basic sites: a nitrogen atom and an oxygen atom.
In order to provide information on the nature of the elec-
trophilic species thus formed, we determined the relative
stability of the species that can be formed by protonation
or adduct formation with a Lewis acid. Thus, the site and
energetics of protonation and AlCl3 adduct formation by
some formamide derivatives were investigated computa-
3
3
3
4
4
4
6
6
4[c]
5
42
46
42
47
59
65
44
63
72
75
64[d]
65
45
6[c]
5[c]
6
8
6[c]
6
6
6
6
8
10
12
12
4
14[e]
15.3[e]
[a]
[b]
With respect to 2. Ϫ
Product is 3,4-dimethylbenzaldehyde.
Yield calculated for transfer of 3 formyl groups per molecule 2. Ϫ
[c]
‘‘Aged’’ AlCl3 (partially hydrolyzed by repeated opening of the
[d]
container). Ϫ
tained as a by-product. Ϫ
1,2-Bis(3,4-dimethyphenyl)ethane (3 %) was ob-
[e]
Reaction performed without addi-
tional solvent.
The reaction features a high para selectivity in the sub-
strate, consistent with the high steric bulk of the attacking
species. As can be appreciated from Table 1 and Table 2, in
order to achieve satisfactory yields with formamide derivat-
ives bearing one or more diformylamino groups, it is neces-
sary to use at least 2 equiv. of Lewis acid. It is hence very
likely that the primary adduct reacts with further Lewis acid
to form the formylating species.
Lastly, we emphasize that the method we have presented
can be performed using only stoichiometric amounts of the
formamide derivative, and 2 equiv. of Lewis acid per formyl
group transferred. E-factors (kg of waste per kg of product)
of the new formylation procedures are in the range from 4
to 11; further investigations will show whether these values
can be reduced by use of other types of catalysts. The com-
parable method with dichloromethyl methyl ether often em-
ploys a large excess of the formylating mixture (up to 18
equiv.),[16,17] and the regioselectivity is rather low.[18]
Formylation with Other Formamide Derivatives
The perchlorinated derivative of triformamide, tris(di-
chloromethyl)amine (3),[12] can be prepared from N,N-di-
methylformamide and phosgene, followed by photo-
chlorination of the resulting N,N-dimethylformamide chlor-
ide. In the presence of various Lewis and Brønsted acids, 3
also proves to be a good formylating agent, and it is suitable
for the formylation of a wide range of aromatic com-
pounds.[8] Thus, toluene reacts with 3/AlCl3 to give tolual-
dehyde, in a yield of 70% and a p/o ratio of 30:1. Similarly,
anisaldehyde can be prepared from anisole and 3/ZnCl2 in
62% yield (p/o 24:1) (see Exp. Sect.). However, this complex
subject will be dealt with in more detail in a forthcoming
Scheme 2. Calculated structures of formamide derivatives 1؊5 and
paper. Preliminary studies on the formylating ability of their protonated forms
Eur. J. Org. Chem. 2001, 2947Ϫ2954
2949

A. Bagno, W. Kantlehner, O. Scherr, J. Vetter, G. Ziegler
Table 3. Calculated proton affinities and protonation sites
FULL PAPER
tionally by means of DFT calculations. All calculations
were performed with Gaussian 98;[21] the stabilities were de-
termined both for the isolated species and in solution, the
solvent effect on the equilibria being modeled by the IPCM
continuum method.[22] Atomic charges (see below) were cal-
culated with the NBO method.[23] The bases studied were:
diformamide [HN(CHO)2] (5), triformamide [N(CHO)3]
(1), N,N,NЈ,NЈ-tetraformylhydrazine [(CHO)2NϪN(CHO)2]
(4), and tris(diformylamino)methane {HC[N(CHO)2]3} (2).
The structures and conformations thus obtained are shown
in Scheme 2 and 3. Geometrical features are only given in
[b][c]
[b][c]
Species
PA[a]
∆E(g)
∆E(aq)
Formamide
HCONH3
HC(OH)NH2
196.0[d]
ϩ
ϩ17.4
(0.0)
ϩ11.3
(0.0)
ϩ
Diformamide (5)
187.8
185.4
ϩ
H2N(CHO)2
ϩ24.7
(0.0)
ϩ21.7
(0.0)
HN(CHO)(CHOH)ϩ
Triformamide (1)
ϩ
HN(CHO)3
ϩ28.5
(0.0)
ϩ28.4
(0.0)
N(CHO)2(CHOH)ϩ
the Supporting Information as PDB files, except where such N,N,NЈ,NЈ-Tetraformylhydrazine (4) 180.7
ϩ [e]
(CHO)2NHϪN(CHO)2
Ϫ
Ϫ
Ϫ
Ϫ
features are discussed in the text. Even though the analysis
(CHO)2NϪN(CHO)(CHOH)ϩ
Tris(diformylamino)methane (2)
HC[N(CHO)2]2[NH(CHO)2]ϩ
HC[N(CHO)2]2[N(CHO)(CHOH)]ϩ
for formamide had previously been carried out in our
192.5
laboratories,[9] we repeated the same calculations to ensure
maximum homogeneity in data comparison.
ϩ22.5
(0.0)
ϩ31.6
(0.0)
Absolute and relative proton affinities were calculated
(for the most stable conformer only) as follows. For
HCONH2, HN(CHO)2, N(CHO)3, (CHO)2NϪN(CHO)2,
and their protonated forms, gas-phase data were obtained
[a]
[b]
Proton affinity for O-protonation. Ϫ
Energies [kcal/mol] in
the gas phase calculated at the B3LYP/6-311ϩϩG(d,p)//B3LYP/6-
311ϩϩG(d,p) level, including zero-point vibrational energy correc-
tions (ZPE); energies in water at the IPCM-B3LYP/6-
at the B3LYP/6-311ϩϩG(d,p) level and geometry, includ- 311ϩϩG(d,p) level with the gas-phase geometry, assuming ε ϭ
78.5, except for the protonated forms of 2, calculated with the 6-
ing zero-point vibrational energy corrections (ZPE). All
species were thus characterized as minima on the respective
31ϩG(d,p) basis set. Ϫ [c] Energy difference between the two tauto-
meric protonated forms, referenced to the most stable O-pro-
potential energy surfaces, except where noted. Energies in
water were calculated at the IPCM-B3LYP/6-311ϩϩG(d,p)
level at the gas-phase geometry, assuming a dielectric per-
mittivity of ε ϭ 78.5. For the larger species HC[N(CHO)2]3
and associated protonated forms, the smaller 6-31ϩG(d,p)
basis set was used, other methods being the same. Absolute
and relative proton affinities (PA) (Table 3) were obtained
tonated form (c in the gas phase, a in water; see text). Ϫ [d] Experi-
[e]
mental PA: 196.5 kcal/mol (ref.[32]). Ϫ
Not a minimum on the
potential energy surface (expulsion of HCOϩ; see text).
Proton affinities (Table 3) were obtained with respect to
with reference to the most stable conformers of the neutral the most stable conformers of the neutral (5a) and the pro-
and the protonated forms, where applicable.
tonated forms (5Oc and 5N). The data indicate a stronger
gas-phase basicity of oxygen relative to nitrogen by 25 kcal/
mol, and thus by a larger margin than in formamide (17
Diformamide (5)
Conformational analysis of the simplest member revealed kcal/mol; cf. 15 kcal/mol[9]).
that the conformer with the carbonyl groups farthest apart
Calculations in water indicated a similar relative stability
(5a) is more stable (by 6 kcal/mol) than that with those of 5a and 5b, with 5a still being the most stable conformer
groups eclipsed (5b). In addition to these, the possibility of (with ∆E ϭ 4 kcal/mol). Again, the enol form 5c is much
the enol form (5c) being present should be considered. Even less stable. With regard to the conformers of 5O, we note
though amides are known to be enolized only to a very that the closed form 5Oc, somewhat more stable in the gas
small extent, species containing an N(CHO)2 moiety might phase (see above), is not particularly favored in water, being
be stabilized by intramolecular hydrogen bonding (see 1.7 kcal/mol less stable than the reference conformer 5OaЈ.
Scheme 2). However, 5c proved to be substantially less In any event, a marked preference for O-protonation is again
stable (17 kcal/mol) than 5a or 5b.
indicated, as judged from the relative PAs of 22 kcal/mol.
Various possible conformations of its O-protonated
At this point, we can provide an initial assessment of the
forms (5Oa؊c) were considered, including the closed spe- conformational and keto/enol effect on the proton transfer:
cies (5Oc), formally related to 5c. All lie within 6 kcal/mol (a) the enol form is much less stable than the keto form
of each other, the closed form 5Oc being the most stable both in the gas and aqueous phases, and (b) the various
thanks to its internal stabilization. However, 5OaЈ (related conformers of the O-protonated form differ by at most 6
to the most stable neutral conformer 5a, and presenting the kcal/mol in energy, the closed form being the most stable in
typical arrangement of protonated amido groups,[9,24]) is the gas phase but not in water. Hence, (a) the neutral enol
only slightly less stable (1.8 kcal/mol), whereas 5Ob (with form is disregarded in all further calculations, and (b) for
eclipsed carbonyl groups as in 5b) is the least stable. Geo- consistency, relative proton affinities in the gas or water
metry optimization of the other conceivable conformer de- phases are calculated as the energy difference between dif-
riving from 5b led to the 5OaЈ structure. AM1 semiempir- ferent conformers; given the small spread of their stabilities,
ical geometry optimization of all conceivable conformers of however, the final outcome in terms of O-protonation vs.
the N-protonated form (5N), on the other hand, resulted in N-protonation is not affected. Such considerations are also
the single structure depicted in Scheme 2.
applicable to the other bases.
2950
Eur. J. Org. Chem. 2001, 2947Ϫ2954

New Formylating Agents
FULL PAPER
Triformamide (1)
The two possible conformers for neutral 1 (1a, 1b) differ
in the orientation of a CO group, and are remarkably sim-
ilar in energy both in the gas phase and in water (∆E ϭ
1Ϫ3 kcal/mol). The protonation site in 1 is, again, oxygen
rather than nitrogen, with a slightly larger energy gap (28
kcal/mol) in both phases.
N,N,NЈ,NЈ-Tetraformylhydrazine (4)
Because of the size of this system, conformational ana-
lysis on the neutral species was carried out at the AM1
level, in order to pinpoint the most stable conformer for
further calculations (Scheme 2). Only the O-protonated spe-
cies could be characterized as an energy minimum (two
conformations, 4Oa and 4Oc). In contrast, geometry op-
˚
timization of 4N resulted in a major lengthening (Ͼ 2.8 A)
of an H(O)CϪN bond involving the protonated nitrogen
atom, essentially corresponding to the expulsion of a formyl
cation. Thus, the partially optimized structure of 4N may
be viewed as a loose complex between neutral N,N,NЈ-tris-
formylhydrazine and the HCOϩ ion. The lengthening of the
bond between N and the acid residue is a general phenom-
enon in protonated amides; expulsion of the ionic acyl
group has been predicted for N-protonated nitrosamine
Scheme 3. Calculated structures of complexes between formamide,
(H3N···NOϩ)
and
phosphorous
acid
triamide
diformamide (5), and AlCl3
[H3N···P(NH2)2ϩ], as well as for O-alkyl-protonated formic
acid and methyl formate (HOR···HCOϩ).[9] While this oc-
currence prevents us from quantitatively estimating the ni-
trogen basicity of 4, it nevertheless further points out the
higher stability of the O-protonated form. By comparison
with 5 and 1, we can infer that the same conclusions hold
in water as well.
Scheme 3). To test the viability of this, we also ran calcula-
tions on such a structure.
On the basis of the results reported in the preceding sec-
tion, which showed a remarkable similarity of behavior be-
tween the various derivatives investigated, we restricted the
calculations to formamide itself as the Lewis base model,
and to diformamide (5). The calculations were run at the
same level as seen before, except that the estimation of the
absolute AlCl3 affinity was performed correcting for the ba-
sis set superposition error (BSSE) by the counterpoise
method.[25] We chose DCE (ε ϭ 10.4) for IPCM calcula-
tions, since this was the solvent in which the reactions of
interest had been carried out. These results are presented in
Table 4, and the pertinent structures in Scheme 3. Com-
plexation occurs with a small (20Ϫ24°) but appreciable de-
viation from planarity of the AlCl3 moiety; most other
structural features resemble the corresponding protonated
form.
Tris(diformylamino)methane (2)
This base can be envisaged as three closely spaced but
unrelated diformamide (5) systems. For this large and flex-
ible species, no conformational analysis was carried out, but
each individual formamido group was assumed to lie in the
conformation previously found to be more stable for 5
(Scheme 2). Not surprisingly, the picture is similar to that
of 5, the oxygen again being the preferred protonation site
(by 22 and 32 kcal/mol for the gas and aqueous phases, re-
spectively).
Aluminum Chloride Affinities
Protonation Sites, Proton and AlCl3 Affinities
Complexation with a Lewis acid such as AlCl3 is a
method widely adopted for carbonyl group activation, in-
The absolute calculated PAs for O-protonation of 1, 2, 4,
cluding that of amides, towards nucleophilic addition. The and 5 (180Ϫ190 kcal/mol) are all lower than that for forma-
complexation is normally thought of as taking place at the mide (ca. 196 kcal/mol), despite their larger molecular size.
oxygen atom; nevertheless, this needs to be confirmed inde- This indicates a strongly reduced basicity for these com-
pendently. To this end, we ran the same series of calcula- pounds relative to a simple amide group, although the gen-
tions as seen for proton transfer in order to determine the eral features are preserved; thus, all compounds considered
relative stability of N- and O-adducts with AlCl3. Further- are oxygen bases rather than nitrogen bases. The least basic
more, the ability of 2 to transfer several formyl groups (see is 4, presumably thanks to the known base-weakening effect
above) suggests the intermediate formation of a complex in of the hydrazo group,[9] whereas the most basic one is 2,
which AlCl3 loses a chloride ion (as AlCl4Ϫ) and acts as a which agrees with its larger molecular size (and polariz-
bidentate acceptor towards
a
diformamide unit (see ability). The strong base-weakening effect of the added for-
Eur. J. Org. Chem. 2001, 2947Ϫ2954
2951

A. Bagno, W. Kantlehner, O. Scherr, J. Vetter, G. Ziegler
FULL PAPER
Table 4. Calculated affinities for AlCl3
Species
Gas phase[a]
DCE[a]
AlCl3 affinity[b]
∆E(g)
AlCl3 affinity[b]
∆E(DCE)
[c]
[c]
Formamide
40.6
34.3
42.1
HCONH2···AlCl3
H2NϪCHO···AlCl3
Diformamide (5)
(OHC)2NH···AlCl3
HN(CHO)CHO···AlCl3
ϩ14.8
(0.0)
ϩ19.6
(0.0)
28.5
ϩ20.6
(0.0)
ϩ18.8
(0.0)
[a]
Energies [kcal/mol] in the gas phase calculated at the B3LYP/6-311ϩϩG(d,p)//B3LYP/6-311ϩϩG(d,p) level; energies in 1,2-dichloro-
ethane (DCE) at the IPCM-B3LYP/6-311ϩϩG(d,p) level with the gas-phase geometry, with ε ϭ 10.4. All results are corrected for basis
[b]
[c]
set superposition error. Ϫ
referenced to the O-adduct.
AlCl3 affinity for adduct formation through oxygen. Ϫ
Energy difference between the two adducts,
myl group is made evident by the smaller PA of 1 as com-
pared to 5. This decrease in basic strength is readily inter-
preted if the canonical resonance form of an O-protonated
amide is considered, with the positive charge delocalized
onto to the nitrogen atom bearing a partial positive charge.
Such a resonance form is diminished in importance for the
bases under consideration, since the nitrogen atom bears
one or two strongly electron-withdrawing formyl groups,
which obviously destabilize such resonance forms and, as a
consequence, the overall stability of the ion (however, see
below).
In the case of simple amides, the stability of the N-pro-
tonated form is enhanced in water. Thus, for HCONH2, the NBO atomic charges (boldface) for formamide, diformamide (5),
Scheme 4. Selected calculated bond lengths (normal typeface) and
and their protonated forms
energy gap between O- and N-protonated forms decreases
from 17 kcal/mol in the gas phase (cf. 15 kcal/mol[9]) to 11
ϩ
kcal/mol in water (Table 3). It is similar for MeCONHMe that for HC(OH)NH2 [∆q(O) ϭ ϩ0.031], or in other
(in which the substitution pattern at nitrogen is the same as words the charge was developed at the protonated oxygen
for 5), with the energy gap decreasing from 14 kcal/mol in atom to a relatively smaller extent. The CϪN bond (rCN
)
the gas phase to 7 kcal/mol in water.[24] Although the level was, again, shortened by a similar amount (∆rCN ϭ Ϫ0.090
˚
of stabilization was found to be overestimated because of and Ϫ0.067 A, respectively). There is, however, one struc-
the limitations of the IPCM method,[24] it was consistently tural feature that sets 5OaЈ apart from HC(OH)NH2ϩ: The
encountered, and ascribed to stronger hydration of the N- bond between N and the unprotonated CO group (rCЈN) is
˚
protonated form. Thus, for example, the data in Table 3 al- markedly lengthened (rCЈN ϭ 1.389 to 1.479 A, with
ϩ
˚
low the hydration energy of HC(OH)NH2 (Ϫ74 kcal/mol) ∆rCЈN ϭ ϩ0.090 A), with the carbonyl oxygen atom acquir-
ϩ
to be estimated as less exothermic than that of HCONH3
ing a large positive charge [∆q(OЈ) ϭ ϩ0.123]. Thus, 5OaЈ
(Ϫ81 kcal/mol). A similar relationship holds for the hydra- can be viewed as an incipient complex between the neutral
tion energies of MeC(OH)NHMeϩ (Ϫ63) and imido tautomer of formamide and a formyl cation,
MeCONH2Meϩ (Ϫ70).[24]
HOϪCHϭNH···[HCO]ϩ (although the CϪN bond length
For compounds 5, 1, and 2, in contrast, the relative is much too short to indicate a real bond break). It is, how-
stabilities of the two tautomeric ions either are unaffected ever, not obvious whether the charge distribution in this
by water or the N-protonated forms become even less stable. structure would in fact result in enhanced solvation.
For example, the hydration energy of 5OaЈ (Ϫ75 kcal/mol)
On the other hand, the structure and charge distribution
is only slightly less exothermic than that of 5N (Ϫ78 kcal/ of 5N are also peculiar. In addition to the expected
mol). In principle, this might be due to enhanced solvation lengthening of both CϪN bonds, a substantial positive
of 5OaЈ, or to diminished solvation of 5N (or both). An charge is formed on the carbonyl oxygen atoms (from
analysis of the geometry and atomic charges of 5OaЈ and Ϫ0.550 and Ϫ0.558 to Ϫ0.363 and Ϫ0.368), indicative of
O-protonated formamide (Scheme 4) showed no great dif- substantial participation of both carbonyl groups in charge
ferences.
delocalization in 5N, which probably accounts for its dimin-
In both cases, the charges residing on the acidic hydrogen ished solvation.
atom [q(H) ϭ ϩ0.537 and ϩ0.535, respectively] were very
With regard to AlCl3 complexation, for both formamide
similar, as were the charges on the acyl oxygen atom and 5, O-complexation is favored by 15Ϫ21 kcal/mol in
[q(O) ϭ Ϫ0.553 and Ϫ0.570, respectively]; in fact, the vari- both phases. That is, the general behavior closely resembles
ation in q(O) for 5OaЈ [∆q(O) ϭ ϩ0.003] was smaller than that observed towards protonation. The solvent effect is
2952
Eur. J. Org. Chem. 2001, 2947Ϫ2954

New Formylating Agents
FULL PAPER
relatively small, consistent with the low dielectric permittiv- Experimental Section
ity of DCE. We would also like to point out that the bident-
ate complex in Scheme 3 is also a stable species on the po-
tential energy surface, as confirmed by vibrational analysis.
Formylation of Toluene with 2 and Various Amounts of AlCl3. ؊
(Typical) Procedure I: Dry AlCl3 (90 mmol) was added with stirring
to a mixture of toluene (45 mmol) and dry 1,2-dichloroethane (25
mL), cooled in an ice/salt bath. After a few minutes, tris(diformyla-
mino)methane (2) (15 mmol) was added. The reaction mixture was
stirred in the cooling bath with exclusion of moisture for 20 h, dur-
ing which the temperature rose to 0 °C. The viscous, reddish brown
mixture was carefully hydrolyzed by addition of 100 mL of ice-cold
water, and steam-distilled. The organic layer of the distillate was
separated, and the aqueous phase extracted three times with 10 mL
of 1,2-dichloroethane. The combined organic layers were dried with
sodium sulfate and filtered, and the solvent was evaporated at or-
dinary pressure. p-Tolualdehyde (b.p. 84 °C/12 Torr) (204Ϫ205 °C/
760 Torr[26]) was isolated by fractional vacuum distillation. Yield
25 mmol (55%).
Electrophilic Species and Formylation Mechanism
All the calculated data discussed above point to greater
basicity of the oxygen, rather than the nitrogen, in the for-
mamido group, both in the Brønsted and in the Lewis
senses. It is therefore possible to point to an O-complexed
structure such as those depicted in Scheme 3 for the reactive
electrophilic species in the formylation reaction. We can
thus propose a general mechanism such as that shown in
Scheme 5, which also highlights the coordination of a fur-
ther carbonyl group to AlCl3, thus accounting for the abil-
ity of formamide derivatives to transfer more than one for-
myl group, as found experimentally (and predicted theoret-
ically).
Formylation of Aromatic Compounds with 2/Lewis Acid. ؊ (Gen-
eral) Procedure II: A stirred mixture of the aromatic compound
(40Ϫ60 mmol) and 1,2-dichloroethane (25Ϫ40 mL) was cooled in
an ice/salt bath. The Lewis acid was added first, over 5 min, fol-
lowed by tris(diformylamino)methane (2). During the reaction time
stated, the temperature of the reaction mixture rose from ca. Ϫ15
°C to 0 °C (see Table 1). After hydrolysis by careful addition of 100
mL of water, the organic layer was separated and the aqueous
phase was extracted three times with 10 mL of 1,2-dichloroethane.
Ϫ Workup A: The organic layers were combined, and the solvent
was evaporated at reduced pressure. The residue was treated with
100 mL of a saturated sodium hydrogen sulfite solution, 3 mL of
methanol, and ca. 250 mg of tetrabutylammonium hydrogen sulf-
ate. The bisulfite adduct separated upon stirring and was isolated
by filtration. The adduct was cleaved by addition of either 100 mL
of 8% NaHCO3 solution or 20 mL of 10% HCl. The aldehyde was
extracted from the mixture three times with 10 mL of ether. The
combined organic layers were dried with sodium sulfate and fil-
tered, and the solvent was evaporated. The aldehyde was obtained
by fractional distillation of the residue through a 15-cm Vigreux
column. Ϫ Workup B: The combined organic layers were dried with
sodium sulfate and filtered, and the solvent was evaporated at re-
duced pressure. The aldehyde was obtained by fractional distilla-
tion of the residue through a 10-cm Vigreux column.
Scheme 5. Proposed mechanism for the electrophilic formylation
of arenes with formamide derivatives and Lewis acids
Formylation of Aromatic Compounds with 3/Lewis Acid. ؊ (a) Tolu-
aldehyde from 3 and AlCl3: Compound 3 (7.3 g, 30 mmol) was ad-
ded over 10 min, at 5 °C and with stirring, to a cooled mixture of
dry toluene (25 g, 0.27 mmol) and AlCl3 (11 g, 80 mmol). The cool-
ing bath was removed and the mixture was stirred for 15 h at 20
°C. The reaction mixture was hydrolyzed with 50 mL of water and
steam-distilled. The organic phase of the distillate was separated
and the aqueous phase was extracted three times with CH2Cl2. The
organic phases were combined and dried with sodium sulfate. The
Conclusion
Arene formylation can conveniently be carried out in sat-
isfactory yields with formamide derivatives in the presence
of Lewis acids. Quantum chemical calculations provide co-
gent indications of the nature of the electrophilic species
involved in this reaction. Thus, in the attack by a proton drying agent was removed by filtration. The solvents were removed
from the filtrate by distillation at ordinary pressure. Tolualdehyde
(2.31 g, 70%; 97% p-isomer, 3% o-isomer) was isolated by distilla-
tion (b.p. 80 °C/12 Torr). Ϫ (b) Anisaldehyde from 3 and ZnCl2:
ZnCl2 (2.18 g, 31.6 mmol) was added to a solution of 3 (4.2 g,
15.8 mmol) in anisole (20 mL, 187 mmol), whereupon the mixture
turned red. The mixture was heated with stirring for 3 h at 60Ϫ65
°C, hydrolyzed with 100 mL of water, and steam-distilled. Dichloro-
methane (30 mL) was added to the distillate in order to produce
better phase separation. The organic phase was separated and the
aqueous phase was extracted three times with 50 mL of dichloro-
methane. The combined organic phases were dried with sodium
or by AlCl3 (protonation or Lewis acid adduct formation,
respectively), the amide oxygen atom is the most basic site,
both in the gas phase and in water or DCE as solvents. The
formation of a chelate complex between the two adjacent
carbonyl oxygen atoms in a diformamide group and AlCl3
is also supported. All these results reinforce the proposition
that the attacking electrophilic species is formed by attach-
ment of the Lewis acid to the oxygen atom of the formam-
ide derivatives, and endorse the proposed reaction mechan-
ism (Scheme 5).
Eur. J. Org. Chem. 2001, 2947Ϫ2954
2953

A. Bagno, W. Kantlehner, O. Scherr, J. Vetter, G. Ziegler
FULL PAPER
[19]
[20]
W. Eitel, Dissertation Universität Stuttgart 1971, p. 42.
Proceedings of the 4th Iminium Salt Conference, Rechenberg/
Stimpfach, September 14Ϫ16, 1999.
sulfate. The solvents were evaporated, and the residue was distilled
in vacuo through a 20-cm Vigreux column. Yield: 1.33 g (62%) anis-
aldehyde (96% p-isomer, 4% o-isomer), b.p. 80 °C/0.01 Torr.
[21]
M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M.
A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgom-
ery, Jr., R. E. Stratmann, J. C. Burant, S. Dapprich, J. M. Mil-
lam, A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J.
Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Po-
melli, C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P.
Y. Ayala, Q. Cui, K. Morokuma, D. K. Malick, A. D. Rabuck,
K. Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz, B.
B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R.
Gomperts, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham,
C. Y. Peng, A. Nanayakkara, C. Gonzalez, M. Challacombe,
P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, J. L. Andres,
C. Gonzalez, M. Head-Gordon, E. S. Replogle and J. A. Pople,
Gaussian 98, Revision A.7, Gaussian, Inc., Pittsburgh PA, 1998.
J. B. Foresman, T. A. Keith, K. B. Wiberg, J. Snoonian, M. J.
Frisch, J. Phys. Chem. 1996, 100, 16098.
[1]
G. Simchen, Methoden Org. Chem. (Houben-Weyl), 1983,
vol. E3.
[2]
E. Winterfeldt, Methoden Org. Chem. (Houben-Weyl), 1983,
vol. E3.
G. A. Olah, S. J. Kuhn, FriedelϪCrafts and Related Reactions
(Ed.: G. A. Olah), Interscience Publ., New York, 1964, vol.
III/2.
G. Casiraghi, G. Casnati, G. Puglica, G. Sartori, G. Terenghi,
J. Chem. Soc., Perkin Trans. 1 1980, 1862.
J. C. Duff, J. Chem. Soc. 1941, 574.
A. Rieche, H. Groß, E. Höft, Chem. Ber. 1960, 93, 88.
W. Kantlehner, M. Vettel, A. Gissel, E. Haug, G. Ziegler, M.
Ciesielski, O. Scherr, R. Haas, J. Prakt. Chem. 2000, 342, 297.
W. Kantlehner, E. Haug, G. Ziegler, O. Scherr, M. Ciesielski:
Neue, umweltfreundliche, gewerbetoxikologisch unbedenkliche
Aldehydsynthesen (Varianten der VilsmeierϪHaack-Reaktion),
Final Report to BMBF Project (Förderkennzeichen
01 Z 9502), 1998.
[3]
[4]
[5]
[6]
[7]
[22]
[23]
[24]
[25]
[26]
[27]
[8]
A. E. Reed, L. A. Curtiss, F. Weinhold, Chem. Rev. 1988, 88,
899.
A. Bagno, B. Bujnicki, S. Bertrand, C. Comuzzi, F. Dorigo, P.
Janvier, G. Scorrano, Chem. Eur. J. 1999, 5, 523.
F. B. van Duijneveldt, J. G. C. M. van Duijneveldt-van de
Rijdt, J. H. van Lenthe, Chem. Rev. 1994, 94, 1873.
R. C. Weast (Ed.), Handbook of Chemistry and Physics, 60th
ed., CRC Press, Boca Raton (Florida), 1979؊1980.
G. Y. Han, P. F. Han, J. Perkins, H. C. McBay, J. Org. Chem.
1981, 46, 4695.
J. Frahn, A. D. Schlueter, Synthesis 1997, 1301.
C. T. Lester, R. E. Donaldson, J. C. Oswald, J. Am. Chem. Soc.
1949, 71, 1502.
L. Rousset, Bull. Soc. Chim. Fr. 1897, 17, 810.
A. Kreutzberger, Arch. Pharm. Ber. Dtsch. Pharm. Ges. 1969,
302, 828.
[9]
A. Bagno, G. Scorrano, J. Phys. Chem. 1996, 100, 1536.
Crystal structure: W. Frey, W. Kantlehner, G. Ziegler, O. Scherr,
[10]
Z. Kristallogr. 2001, 216, 97.
[11]
E. Allenstein, V. Beyl, W. Eitel, Chem. Ber. 1969, 102, 4089.
[12]
K. Grohe, E. Klauke, H. Holtschmidt, H. Heitzer, Justus Lie-
bigs Ann. Chem. 1969, 730, 140.
[28]
[29]
[13]
E. Allenstein, V. Beyl, Chem. Ber. 1967, 100, 3551.
[14]
H. Yinglin, H. Hongwen, Synthesis 1990, 122.
[15]
W. Kantlehner, G. Ziegler, M. Ciesielski, O. Scherr, M. Vettel,
[30]
[31]
Z. Naturforsch., Teil B 2001, 56, 105.
A. Arduini, G. Manfredi, A. Pochini, A. R. Sicuri, R. Ungaro,
J. Chem. Soc., Chem. Commun. 1991, 936.
T. Shimizu, S. Hiranuma, T. Watanabe, M. Kirihara, Hetero-
[16]
[32]
[17]
E. P. Hunter, S. G. Lias, J. Phys. Chem. Ref. Data 1998, 27,
413. Available on the Internet as http://webbook.nist.gov
Received January 22, 2001
cycles 1994, 38, 243.
[18]
G. A. Olah, L. Ohannesian, M. Arvanaghi, Chem. Rev. 1987,
87, 671.
[O01008]
2954
Eur. J. Org. Chem. 2001, 2947Ϫ2954