A Study on the Activation of Carboxylic Acids by Means of 2-Chloro-4,6-dimethoxy-1,3,5-triazine and 2-Chloro-4,6-diphenoxy-1,3,5-triazine

Article abstract of DOI:10.1021/jo972020y

Activation of carboxylic function by means of 2-chloro-4,6-disubstituted-1,3,5-triazines 1 and 2 leading to triazine esters was found to be a multistep process with participation of quarternary triazinylammonium salts 3-6 as the intermediates, with the rate of reaction strongly dependent on the structure of the tertiary amine. The studies on alkylation of tertiary amines with CDMT revealed the two-step process A_N + D_N, and zwitterionic addition product 9 was identified by ¹H NMR spectroscopy. Semiempirical modeling of the reaction as well as measured nitrogen and chlorine isotope effects also support this mechanism.

Full text of DOI:10.1021/jo972020y

4248
J . Org. Chem. 1998, 63, 4248-4255
A Stu d y on th e Activa tion of Ca r boxylic Acid s by Mea n s of
2-Ch lor o-4,6-d im eth oxy-1,3,5-tr ia zin e a n d
2-Ch lor o-4,6-d ip h en oxy-1,3,5-tr ia zin e
Z. J . Kamin´ski,* P. Paneth, and J . Rudzin´ski
Faculty of Chemistry, Technical University of Lo´dz´, ul. Zwirki 36, 90-924 Lo´dz´ (Lodz), Poland
Received November 4, 1997
Activation of carboxylic function by means of 2-chloro-4,6-disubstituted-1,3,5-triazines 1 and 2
leading to triazine esters was found to be a multistep process with participation of quarternary
triazinylammonium salts 3-6 as the intermediates, with the rate of reaction strongly dependent
on the structure of the tertiary amine. The studies on alkylation of tertiary amines with CDMT
1
revealed the two-step process AN + DN, and zwitterionic addition product 9 was identified by H
NMR spectroscopy. Semiempirical modeling of the reaction as well as measured nitrogen and
chlorine isotope effects also support this mechanism.
In tr od u ction
mechanism in the process of activation of a carboxylic
group achieving highly enantioselective activation of
racemic carboxylic acids in the presence of chiral amines.5
These new, strongly stereodifferentiating effects of amines
acting as a chiral auxiliary prompted us to study all
stages of activation of the carboxylic function in the
process involving the substitution of chlorine in the
triazine ring with carboxylate anion.
The efficiency of 2-chloro-4,6-disubstituted-1,3,5-triaz-
ines in formation of the peptide bond,1 especially between
sterically hindered substrates,2 prompted us to undertake
a systematic study on the reaction of chloro-s-triazines
with carboxylic acids. The previous investigations docu-
mented the role of 2-acyloxy-4,6-disubstituted-1,3,5-tri-
azines3 as powerful acylating intermediates due to the
fact that they contain an excellent leaving group strongly
inclined for acyl-transfer reaction facilitated by intramo-
lecular catalysis involving nitrogen atoms of triazine ring
(triazine “superactive esters”).4 However, the details of
the activation step leading from the carboxylic acid to
2-(acyloxy)-4,6-disubstituted-1,3,5-triazines, which re-
sults in the substitution of the chlorine atom in the
triazine ring by a carboxylic group, remain unknown.
Our preliminary observations have shown that the
successful activation of carboxylic acids by means of
2-chloro-4,6-disubstituted-1,3,5-triazines requires the pres-
ence of a tertiary amine in the reacting medium. Thus,
the sodium, silver, or quaternary ammonium salts of
carboxylic acids, otherwise considered as good sources of
carboxylate anion, do not react with 2-chloro-4,6-disub-
stituted-1,3,5-triazines at all, or else react extremely
slowly.3 We concluded, therefore, that the reaction
proceeds with the participation of an intermediate,
formed in the first step, involving amine as the obligatory
component. J ust recently, we further confirmed this
Resu lts a n d Discu ssion
Generally, we found that the reaction of amine with
chloro-s-triazine could be considered as erratic because
only a few of all tertiary amines respond to treatment
with chloro-s-triazines. Moreover, the borderline be-
tween reactive amines (pKa in aqueous solutions at
25 °C are presented in brackets) N,N,N′,N′-tetrameth-
ylguanidine (TMG)6 [13.6], N-methylpiperidine7 [10.13],
trimethylamine8 [9.81], 4-(N,N-dimethylamino)pyridine
(DMAP)9 [9.61], N,N,N′,N′-tetramethylethylenediamine
(TMEDA)10 [8.1], N-methylmorpholine (NMM)11 [7,42],
4-picoline9 [6.25], pyridine [5.1912], and inactive amines:
triethylamine13 [10.87], tributylamine14 [10.63], N-eth-
ylmorpholine15 [7.8], N,N-diethylaniline14 [6.61.], and
N,N-dimethylaniline14 [5.06] does not correlate with the
basicity of amines in polar solvents.
It has been found that all amines involved in activation
of the carboxylic function are prone to the formation of
(5) Kaminski, Z. J .; Markowicz, W. S.; Kolesinska, B.; Matusiak,
W. XIV-th Polish Peptide Symposium, Polanica, 1997.
(6) Angyal, S. J .; Warburton, W. K. J . Chem. Soc. 1951, 2492-4.
(7) Hall, H. K. J . Am. Chem. Soc. 1957, 79, 5444-5445.
(8) Ibrahim, I. T.; Wiliams, A. J . Chem. Soc., Perkin Trans. 2 1982,
1459-1466.
(1) (a) Kaminski, Z. J . Synthesis 1987, 917-920. (b) Cook, G. K.;
Hornback, W. J .; J ordan, C. L.; McDonald, J . H., III; Munroe, J . E. J .
Org. Chem. 1989, 54, 5828-5830. (c) Taylor, E. C.; Gillespie, P. J . Org.
Chem. 1992, 57, 5757-5761. (d) Hipskind, P. A.; Howbert, J . J .; Cho,
S.; Cronin, J . S.; Fort, S. L.; Ginah, F. O.; Hansen, G. J .; Huff, B. E.;
Lobb, K. L.; Martinelli, M. J .; Murray, A. R.; Nixon, J . A.; Staszak, M.
A.; Copp, J . D. J . Org. Chem. 1995, 60, 7033-7036. (e) Taylor, E. C.;
Dowling, J . E. J . Org. Chem. 1997, 62, 1599-1603. (f) Kuciauskas,
D.; Liddell, P. A.; Hung, S. Ch.; Lin, S.; Stone, S.; Seely, G. R.; Moore,
A. L.; Moore, T. A.; Gust, D. J . Phys. Chem. 1997, 101, 429-440. (g)
Nayyar, N. K.; Hutchison, D. R.; Martinelli, M. J . J . Org. Chem. 1997,
62, 982-991. (h) Svete, J .; Aljaz-Rozic, M.; Stanovnik, B. J . Heterocycl.
Chem. 1997, 34, 177-193.
(9) (a) Alexander, A. J .; Kebarle, P. Anal. Chem. 1986, 58, 471-
478. (b) Castro, E. A.; Ureta, C. J . Chem. Soc., Perkin Trans. 2 1991,
63-68.
(10) Gustafson, R. L.; Martel, A. E. J . Am. Chem. Soc. 1959, 81,
525-527.
(11) Hall, H. K. J . Phys. Chem. 1956, 60, 63-66.
(12) Hall, N. F.; Sprinkle, M. R. J . Am. Chem. Soc. 1932, 54, 3469-
3478.
(2) Glo´wka, M. L.; Iwanicka, I.; Kaminski, Z. J . Acta Crystallogr.
1990, C46, 2211-2215.
(13) Kuna, S.; Pawlak, Z.; Tusk, M. J . Chem. Soc., Faraday Trans.
1 1982, 2685-2692.
(3) Kaminski, Z. J . J . Prakt. Chem. 1990, 332, 579-583.
(4) (a) Kaminski, Z. J . Int. J . Peptide Protein Res. 1994, 43, 312-
319. (b) Kaminski, Z. J .; Paneth, P.; O’Leary, M. J . Org. Chem. 1991,
56. 5716-5719.
(14) Almarzoqi, B.; George, a. V.; Isaacs, N. S. Tetrahedron 1986,
42, 601-608.
(15) Fife, T. H.; Natarajan, R. J . Am. Chem. Soc. 1986, 108, 8050-
8056.
S0022-3263(97)02020-3 CCC: $15.00 © 1998 American Chemical Society
Published on Web 06/05/1998

Activation of Carboxylic Acids
J . Org. Chem., Vol. 63, No. 13, 1998 4249
Ta ble 1. Syn th esis of Qu a ter n a r y Am m on iu m Sa lts of 2-Ch lor o-4,6-d isu bstitu ted -1,3,5-tr ia zin es 3-6
amine
triazine
formula
yield (%)
C, found (calcd)
H, found (calcd)
N, found (calcd)
2
3
4
5
NMM
NMM
TMG
TMG
CDMT
CDPT
CDMT
CDPT
C
C
C
C
10H17N4O3Cl (276.72)
20H21N4O3Cl (400.87)
10H19N6O2Cl (290.75)
20H23N6O2Cl (414.90)
79
62
83.5
35
43.18 (43.40)
59.85 (59.93)
41.08 (41.31)
57.63 (57.90)
6.20 (6.19)
5.30 (5.28)
6.87 (6.59)
5.80 (5.59)
20.46 (20.45)
13.98 (13.98)
28.12 (28.90)
18.62 (20.26)
Ta ble 2. Activa tion of Ca r boxylic Acid s by Mea n s of 3
triazine
product
yield (%)
mp (°C)
lit. mp (°C)
7a
7b
8a
8b
8c
8d
8e
8f
CDMT*NMM
CDMT*NMM
CDMT*NMM
CDMT*NMM
CDMT*NMM
CDMT*NMM
CDMT*NMM
CDMT*NMM
CDMT + TEA
(CH3)3CCOODMT
p-O2NC6H4COODMT
CH3CONHc-C6H11
C6H5CONHc-C6H11
m-ClC6H4CONHc-C6H11
o-O2NC6H4CONHc-C6H11
m-ClC6H4CONH-C6H5
o-O2NC6H4CONH-C6H5
C6H5CONHC6H5
44
92
17
47
58
73
63
67
0
47-49
50-523
122-243
102-10326
144-14627
12128
118-120
102-104
144-146
145-147
147-149
125-127
149-150
151-229
12128
151-152
16330
8g
Sch em e 1
F igu r e 1. Rate of activation of trimethylacetic acid in
acetonitrile-d3 solution by means of 1 relative to the structure
of the tertiary amine: for N-methylomorpholine (- + - + -);
N,N′-tetramethylethylenediamine (-∆-∆-); triethylamine
(-O-O-); N,N-dimethylaniline (-0-0-).
intermediate products 3-6 when treated with chlorotri-
azine. In most cases, isolated intermediates decompose
rapidly at room temperature, but derivatives of N-
methylmorpholine and tetramethylguanidine were found
to be sufficiently stable to be studied by spectroscopic
methods and elemental analysis (Table 1).
On the basis of spectroscopic data and elemental
analysis, structures of quaternary morpholinium salts 3
and 4 and appropriate guanidinium salts 5 and 6 have
been determined (Scheme 1).
The participation of morpholinium salt 3 in the activa-
tion step was confirmed by the synthesis of 2-(acyloxy)-
4,6-dimethoxy-1,3,5-triazines 7a ,b and amides 8a -f,
directly from 3, under conditions typical for the activation
of carboxylic acids by means of CDMT (see Table 2). In
all experiments, we found noticeably lower yields as
compared to the standard procedure involving CDMT,16
which is probably due to the prolonged activation proce-
dure caused by the poor solubility of 3 and the formation
of appropriate carboxylic acid anhydrides as byproducts
during the reaction course.
In our further experiments, we attempted to study the
ability of amines to activate carboxylic acids by means
of CDMT (see Figure 1). The rate measurements of
activation of trimethylacetic acid and benzoic acid by
means of 2-chloro-4,6-dimethoxy-1,3,5-triazine (1) re-
vealed that the ability of tertiary amines to promote the
activation of carboxylic dramatically decreased with the
increase of steric hindrance of amine substituents. Thus,
triethylamine, which did not form a quarternary am-
monium salt in the reaction with CDMT, was entirely
incapable of activating benzoic acid (compare Table 3, last
entry and Figure 1).
Moreover, the addition of a tertiary amine, which
otherwise is inactive toward CDMT, has a negligible
influence on the rate of activation of carboxylic acids.
Thus, the presence of N,N-dimethylaniline, added in
amounts equal or even exceeding the amounts of NMM,
does not influence the reaction rate (as evidenced by
entries 4-6 in Table 3).
Th e Str u ctu r e of Mor p h olin iu m Sa lt 3. Morpho-
linium salt 3 can exist as a mixture of conformers with
triazine ring in axial 3a or in an equatorial position 3e.
Attempts to establish the preferred conformation of
substituents in the morpholine ring were performed by
means of 1H and 13C NMR spectroscopy. For all solvents,
only a single set of signals of N-methyl, -CH2O, and
-CH2N groups, respectively, was observed (Table 4).
Also, the low-temperature experiments revealed no split-
ting of the above-mentioned signals upon lowering the
temperature from 25 to -55 °C. Thus, taking into
consideration the separation of the methylene hydrogen
signals in the less substituted, and therefore probably
more flexible, ring of NMM observed below -25 °C, we
postulate that 3 is characterized by the presence of one
dominant conformer rather than a fast equilibrating
mixture of conformers in the chloroform solution.
(16) Kaminski, Z. J . Tetrahedron Lett. 1985, 26, 2901-2904.

4250 J . Org. Chem., Vol. 63, No. 13, 1998
Kamin´ski et al.
Ta ble 3. Ra te Con sta n ts of Activa tion of 2,2-Dim eth ylp r op ion ic Acid by Mea n s of 1 in Aceton itr ile-d 3 in th e P r esen ce
of Ter tia r y Am in es
kIII ( ∆k
a
CPivA (M)
CCDMT (M)
CNMM (M)
Cadditive (M)
T (K)
(M-2 s-1
)
1
2
3
4
5
6
7
0.1
0.1
0.1
0.1
0.1
0.1
0.05
0.05
0.05
0.05
299.5
305
310
303.5
303.5
303.5
303.5
0.333 ( 0.038
0.663 ( 0.073
0.972 ( 0.13
0.723 ( 0.046
0.740 ( 0.043
0.621 ( 0.045
0.185 ( 0.013
0.05
0.05
0.05
0.05
0.1
0.05
0.05
0.05
0.05
0.1
DMA; 0.05
DMA; 0.10
TMEDA; 0.1
a
2,2-Dimethylpropionic acid.
Ta ble 4. 1H a n d 13C NMR (δ) Da ta of 3
solvent
NCH3
3.02
NCH2CH2O
OCH3
3.95
triazine
1H NMR, (CDCl3)
3.71 (d, 2H, J ) 4.4 Hz); 3.76(t, 2H, J ) 3.4 Hz), 3.85 (t, 2H, J ) 3.3 Hz),
3.87 (d, 2H, J ) 4.4 Hz)
3.60 (bt, J ) 4.5 Hz), 3.70 (bt, J ) 4.5 Hz)
44.43, 66.63
1H NMR (DMSO-d6)
13C NMR (DMSO-d6)
3.03
27.03
3.81
54.98
167.10, 172.77
Ch a r t 1
The rigid conformation, however, should yield a 0.5-
1.0 ppm downfield chemical shift of axial compared to
equatorial hydrogens. Such a differentiation of chemical
shift of axial and equatorial hydrogens was observed in
the case of NMM17,18 for both NCH2- and OCH2- at -45
°C and for NCH2-hydrogens in the more rigid N-methyl-
N-phenylpiperidinium iodide at room temperature.19 No
such differentiation was observed for the chemical shifts
of axial and equatorial signals of 3. We believe that this
can be explained by the deshielding effect of triazine in
the axial position 3a (see Chart 1), but further studies
on the conformation of morpholinium salts20 are required.
Mech a n ism of Qu a ter n iza tion of N-Meth ylm or -
p h olin e. The established relation between the structure
of amine and its ability to react with 2-chloro-4,6-
disubstituted-1,3,5-triazines leads to discussion of the
mechanism of activation of a carboxylic group by means
of triazine condensing reagents. Without a doubt, the
crucial stage comprises the two subsequent substitution
reactions in the triazine ring. The first one, which
involves substitution of the chlorine atom by amine
leading to quarternal ammonium salt, has been found
to be extremely sensitive to steric hindrance of amine
substituents. The second reaction, which is exceptionally
tolerant to the steric hindrance of activated carboxylic
acid, involves substitution of the amine leaving group
F igu r e 2. Reaction of NMM with 1 in acetonitrile-d6 solution
at -5 °C (b-b-b, substrate 1; 9-9-9, final product 3;
b-b-b, intermediate 9).
by the carboxylate ion, affording triazine “superactive
esters”.
Even taking into account the increase of steric repul-
sion accompanying the quaternization of the nitrogen
atom and increased repulsion caused by the hybridization
change from sp2 to sp3 on the carbon atom of the triazine
ring and the decrease of hindrance in the reverse
hybridization change, the strong dependence during the
first stage on steric hindrance suggests a sterically
crowded transition state. Thus, we excluded the partici-
pation of the triazine carbonium ion as highly improb-
able. Instead we considered two possibilities; the syn-
chronous, SN2-like mechanism, postulated by Williams
for the triazine (a) and the classic stepwise SNAr mech-
anism involving two separate stages of addition-elimi-
nation (b).
Examination of the mixture of compounds yielded by
the reaction of CDMT with NMM in a chloroform solution
revealed the presence of the intermediate product exclud-
ing synchronous mechanism of substitution a and strongly
suggests the participation of classic addition-elimination
mechanism b. The characteristic pattern of the plot of
the intermediate concentration vs time (Figure 2) further
confirms the presence of an intermediate in the two
subsequent reaction steps.
(17) (a) Walker, D. G.; Brodfuehrer, P. R.; Brundidge, S. P.; Shih,
K. M.; Sapino, C. J . Org. Chem. 1988, 53, 983-991. (b) Mikhailovskii,
A. G.; Bakrin, N. I. Khim. Geterotsikl. Soedin. 1991, 27, 1361-1364.
(18) J ones, A. J .; Beeman, C. P.; Hasan, M. U.; Casy, A. F.; Hassan,
M. A. Can. J . Chem. 1976, 54, 126-135.
(19) Cerichelli, G.; Luchetti, L. Tetrahedron 1993, 49, 10733-10738.
(20) Kogai, B. E.; Gubernatorov, V. K.; Sokolenko, V. A. Zh. Org.
Khim. 1984, 20, 2554-2558.
At low temperature, it was possible to cumulate the
intermediate product 9 in the reacting mixture to reach
over 50% of the concentration of the starting reactant,
but all attempts to isolate it were unsuccessful because

Activation of Carboxylic Acids
J . Org. Chem., Vol. 63, No. 13, 1998 4251
F igu r e 3. COSY spectrum of the reaction mixture of 1 with NMM. The intermediate 9 cumulated to about 50%. Temperature
-45 °C.
of its rapid decomposition to 3 at temperatures exceeding
5 °C. Besides COSY experiments (Figure 3), unequivocal
interpretation of the spectrum has been confirmed by low-
temperature experiments (Figure 4). These show a
multiplet of methylene group at temperatures -30 to -50
°C that at room temperature is covered by signals of
substrates and 3. The resultant spectrum of intermedi-
ate 9 was characteristic for the symmetrically substituted
morpholine ring; two proton doublet of triplets at 4.90
ppm, J ) 12, 2 Hz; doublet at 4.54 ppm, J ) 12 Hz;
doublet of doublets at 4.20 ppm, J ) 12, 2 Hz, and broad
triplet at 3.72 ppm, J ) 12 Hz; singlet at 4.25 ppm
(OCH3); and singlet at 4.03 ppm NCH3 (see Table 5).
We were able to demonstrate in an experiment involv-
ing exchange of 4,6-dimethoxy-1,3,5-triazine ring in 9
with CDMT-d6 deuterated in both methyl groups that the
addition of N-methylmorpholine to CDMT is partially
reversible. When a mixture containing 21% of unreacted
CDMT and 51% of intermediate 9 was treated with a
5-fold excess of CDMT-d6 and allowed to completely
transform to 3, about 29% of CDMT remained unreacted
(if the addition step was irreversible 19% of unreacted
CDMT should be expected). Moreover, the signal ratio
of CH3O/CH3N protons in the final product changed from
2:1 observed for unlabeled compound to 1.7:1. Since the
difference in the ratios is beyond the error in the NMR
measurements, this proves the exchange of triazine ring
in 3 and the reversibility of the addition stage. The
extent of the exchange, however, was significantly lower
than that calculated for the fast exchanging system (k1
. k2).
We excluded reversibility of the second step (Scheme
2b) because no exchange of the triazine ring between 3
and CDMT-d6 was observed (the ratio of CH3N/CH3O
remains unchanged).
The proposed two-stage, addition-elimination mech-
anism has been further confirmed by studies of the
kinetic isotope effects of chlorine and nitrogen.
Isotop e Effects. The chlorine kinetic isotope effect
was determined from the differences in isotopic composi-
tion of the product at low and full conversion levels using
dead-end kinetic conditions. An appropriate excess of
triazine was used so that after full consumption of
N-methylmorpholine the required fraction of triazine
conversion was achieved. Morpholinium salt 2 was
filtered and then dissolved in THF-water and treated
with an excess of silver nitrate to quantitatively precipi-
tate chloride ions as silver chloride, which was then
subjected to isotope analysis by a hybrid FAB-isotope
ratio mass spectrometer MI 1201E (PO Electron, Ukraine).
Samples of about 5 mg of silver chloride were deposited
on the silver tip of the direct insertion probe. Xenon
atoms of 6 keV hitting the surface of the probe at an
incidence angle of 45° were used for ionization. The
negative ions formed in this way were accelerated by the
potential of 5 kV and detected in a Faraday cup collector
system. The spectra contained almost exclusively chlo-
rine peaks at 35 and 37 m/z. The mean value of the (M

4252 J . Org. Chem., Vol. 63, No. 13, 1998
Kamin´ski et al.
F igu r e 4. 1H NMR spectrum of the reaction mixture of 1 with NMM. The intermediate 9 cumulated to reach 50% concentration.
Temperature range 5 to -50 °C.
Ta ble 5. 1H a n d 13C NMR (δ) Da ta of 9
solvent
NCH3
NCH2-
OCH2-
OCH3
4.25
triazine
1H NMR (CDCl3)
4.03 ppm
4.90 (double triplet, J ) 12, 2 Hz),
4.54 (doublet, J ) 12 Hz)
59.20
4.20 (double doublet, J ) 12, 2 Hz),
3.72 (triplet, J ) 12 Hz)
62.45
13C NMR (CDCl3)
57.43
56.99
170.74, 173.84
Sch em e 2
busted, and the isotopic composition of nitrogen was
measured. About 10 mg of sample was required per
measurement. Isotope ratio measurements were done
using a Finnigan Delta S isotope-ratio mass spectrometer
combined on-line with a Hereaus elemental analyzer. The
natural isotopic composition of nitrogen was obtained
from the original NMM sample.
The chlorine kinetic isotope effect was calculated from
eq 124
ln(1 - f)
ln(1 - fRPf/RP∞
k35/k37
)
(1)
)
where f is the fraction of reaction, and R is isotopic ratio
of the product after full conversion, R∞, and after small
conversion, RPf, respectively.
For the calculations of the nitrogen isotope effect, an
equation analogous to eq 1, but derived for the analysis
of the isotopic ratios of reactant, was used:
+ 2)/(M) isotopic ratio was obtained from up to 50
separate determinations, each being an average of 10
individual measurements. The total ion current under
the above-mentioned conditions is stable for 0.5-1 h,
yielding a precision of (0.03-0.07%.
In the case of the morpholine nitrogen isotope effect,
the low conversion samples were processed using an
excess of NMM (0.05 M) over triazine at 5 °C, so that
only a fraction of morpholine was reacted while the entire
amount of triazine was exhausted. The excess of NMM
was precipitated after treatment with dry HCl. The
precipitate was filtered, washed with dry THF, and dried.
The morpholinium salt was redissolved in chloroform and
separated by passing through a silica gel column. The
NMM samples obtained in the above way were com-
ln(1 - f)
ln[(1 - f)(RSf/RS0)]
k14/k15
)
(2)
It was further modified to include the δ values rather
than isotope ratios R according to
RSf/RS0 ) (1000 + δf)/(1000 + δ0)
(3)
(21) Paneth, P. J . Am. Chem. Soc. 1985, 107, 7070-7071.

Activation of Carboxylic Acids
J . Org. Chem., Vol. 63, No. 13, 1998 4253
where δ values are relative isotopic ratios, defined as δi
) (Ri/Rst - 1)1000 and δ0 and δf are relative isotopic
compositions of the remaining reactant at the beginning
of the reaction i ) 0) and at a fraction of reaction f (i )
f), respectively, compared to the isotopic ratio of a
standard Rst. The value depends on the isotopic composi-
tion of the standard used in the mass spectrometric
measurements, but this value drops out at the final
calculation of the isotope effect.
The equation can be rearranged to yield a linear
function of logarithm of isotopic composition on logarithm
of reaction progress:
1
ln(1000 + δf) )
ln(1 - f) + ln(1000 + δ0) (4)
k14/k15
The slope of this function gives the value of the isotope
effect. Both methods yield practically the same value of
the nitrogen kinetic isotope effect.
F igu r e 5. Experimental chlorine isotope effect (changes with
reaction progress).
is a trend in measured values of this isotope effect with
the reaction progress. This cannot be explained on the
basis of the synchronous SNAr mechanism.
The remaining alternative is a stepwise mechanism
with an adduct intermediate. If the reaction of this type
can be described by the scheme
Kin etic Isotop e Effects. Kinetic isotope effects are
used as a tool for distinguishing alternative mechanisms
of organic reactions. In a typical approach, the experi-
mental values are compared with the theoretical predic-
tions for alternative pathways. A more sophisticated
method has been postulated but never tested experimen-
tally.21,22 It was shown theoretically that the reaction
complexity can be traced by the dependence of the
observed isotope effect on the reaction progress even
when chemical detection of an intermediate fails due to
its low concentration. The same is true for those reac-
tions that are so fast that the determination of the
intermediates is not possible. This is the case of the
reaction under investigation here. We observed a sig-
nificant chlorine kinetic isotope effect, k35/k37 ) 1.0058
( 0.0005, and no morpholine nitrogen kinetic isotope
effect, k14/k15 ) 1.0001 ( 0.0006. These results could be
easily explained on the basis of the SN1-type mechanism,
in which the C-Cl bond breakage occurs in the transition
from the reactant to the transition state, and there is no
involvement of morpholine in the rate-determining pro-
cess. This mechanistic possibility has been, however,
excluded on the basis of kinetic observations of selectivity
toward different amines and first order in N-methylmor-
pholine.
Intuitively, if the alternative synchronous SN2 like
mechanism was correct then a smaller chlorine isotope
effect should be expected. In accordance with this
expectation, model calculations at the semiempirical
level, using AM1 Hamiltonian, yielded nitrogen and
chlorine isotope effects of 0.9939 and 1.0023, respectively.
The nitrogen kinetic isotope effect is an incoming group
isotope effect, in which case the two main factors deter-
mining the magnitude of an isotope effect partially cancel
each other. We have shown for similar quarternization
of para-substituted N,N-dimethylaniline that in fact this
isotope effect is small, being in the range 0.9985-1.0036,
depending on the substituent.23 Thus, the nitrogen
isotope effect can be explained within this mechanism.
The chlorine kinetic isotope effect, on the other hand, is
substantially larger than expected. Furthermore, a
closer inspection of individual results reveals that there
k3
9
S {kk
\
1} I
8 P
(5)
2
and the steady-state approximation is valid, then the
observed isotope effect obs(kL/kH) is related to the isotope
effects on the individual rate constants (kL/kH)i (i ) 1-3)
by the equation
(kL/kH)3/(kL/kH)2 + k3/k2
obs(kL/kH) ) (kL/kH)1
(6)
[
]
1 + k3/k2
The observed isotope effect lies between the values for
the two extreme cases of k3/k2 being equal to zero (k2
much smaller than k3) or infinity (opposite relation of rate
constants). These limiting values of the isotope effect are
given by
obs(kL/kH) ) (kL/kH)1(kL/kH)3/(kL/kH)2
for k3/k2 ) 0,
(7)
obs(kL/kH) ) (kL/kH)1
and for k3/k2 f ∞.
We estimated the range for the kinetic isotope effects
to be 0.9780-1.0012 for nitrogen and 1.0082-0.9996 for
chlorine, respectively. As can be seen, both experimental
values are within the predicted ranges. However, while
the nitrogen isotope effect is close to the upper limit,
suggesting that k3 , k2, the chlorine isotope effect is
much closer to the limit, indicating the oposite relation
of the rate constants. Thus, these two values cannot be
simultaneously explained on the basis of the model.
Reconciliation of both results is only possible when it
is recognized that the experimental chlorine isotope effect
changes with the reaction progress as shown in Figure
5. Such a dependence means that the reaction is complex
but the steady-state approximation is not valid.21 The
(22) Gawlita-Kamzurska, E.; Paneth, P. Pol. J . Chem. 1992, 66,
209-213.
(23) (a) Szylhabel-Godala, A.; Madhavan, S.; Rudzinski, J .; O′Leary,
M. H.; Paneth, P. J . Phys. Org. Chem. 1996, 9, 35-40. (b) Gawlita, E.;
Szylhabel-Godala, A.; Paneth, P. J . Phys. Org. Chem. 1996, 9, 41-49.
(24) Melander, L. Isotope Effects on Reaction Rates; Ronald Press
Co.: New York, 1960.

4254 J . Org. Chem., Vol. 63, No. 13, 1998
Kamin´ski et al.
calculations, although the stability is reversed; i.e.,
reactants are more stable than products. Analysis of this
PES indicates an intermediate formation with apparent
activation energy, followed by decomposition to the
product proceeding over a very shallow barrier.
All stationary points (reactants, products, intermedi-
ate, and transition states) were optimized, and force field
calculations confirmed them to be either stable molecules
(no imaginary frequencies) or transition states (exactly
one imaginary frequency). We were also successful in
optimizing the structure of the intermediate using gas-
phase calculations and PM3 Hamiltonian. Analogous
calculations with the AM1 Hamiltonian proved unsuc-
cessful. We found three conformations, 3a , 3e (Scheme
2), and a twisted one, to be stable with heats of formation
being closely similar for all three.
Con clu sion s
The activation of carboxylic acids by means of CDMT
is a multistep process proceeding via triazinylammonium
salts 3-6 formed in situ in the presence of the appropri-
ate amine. The rate of formation of 3-6 strongly depends
on the steric hindrance of N-substitutents. For more
hindered amines, the rapid loss of reactivity is observed.
The efficiency of 3 in the condensation reaction was
confirmed by its successful use in the synthesis of triazine
esters 7a ,b and amides 8a -f.
F igu r e 6. Potential energy surface for reaction of 1 with
NMM calculated for polar solvent using COSMO model.
value of the isotope effect extrapolated to the Y axis
equals
obs(kL/kH)0 ) (kL/kH)1(kL/kH)3
(8)
On the basis of NMR studies, we postulate for com-
pound 3 the presence of one dominant conformer with a
triazine located as an axial substituent of morpholine
ring.
Identification of the intermediate 9 excludes synchro-
nous mechanism and strongly suggests the classic SNAr
addition-elimination mechanism.
Semiempirical modeling of the reaction, as well as
measured nitrogen and chlorine kinetic isotope effects,
also support this mechanism.
From the curve, it can be estimated that this value
should be close to 1.008-1.009. This is in excellent
agreement with the value 1.0082 predicted theoretically.
Further support of the proposed mechanism comes
from the differences in the behavior of nitrogen and the
chlorine isotope effects. The nitrogen kinetic isotope
effect does not exhibit the dependence on the reaction
progress that is seen for the chlorine isotope effect. This
is in accordance with the proposed explanation. Mea-
surements of the nitrogen isotope effect were performed
using the isotopic composition of the remaining reactant,
while the chlorine isotope effect was calculated from the
isotopic ratios of the product. If the sequential reaction
described in the Scheme 3 is irreversible (k2 ) 0) or the
reversibility is negligible (k2 , k3) then one should expect
such a different behavior of isotope effects. The changes
in the chlorine kinetic isotope effect reflect the complexity
of the mechanism and changes of the concentrations of
all reagents throughout the reaction progress. At the
same time, if the formation of the reactant from the
intermediate is negligible, the changes in the isotopic
composition of the reactant reflect only the first reaction
(with the rate constant k1) and thus behave normally.
Ca lcu la tion s. Theoretical calculations of the reaction
(Scheme 3) were performed at the semiempirical level.
The simple gas-phase calculations are not suitable for
this reaction since the product is ionic and presumably
highly stabilized in a solvent of considerable polarity. The
grid calculations revealed that the two Hamiltonians
used, AM1 and PM3, predict different mechanisms for
the reaction in question. Using the continuous solvent
model COSMO and the AM1 Hamiltonian, we were able
to optimize a structure of the transition state correspond-
ing to the SN2-like mechanisms. This was possible if the
dielectric constant was larger than 6.
Exp er im en ta l Section
Gen er a l Meth od s. Melting points were determined in
open tubes and were uncorrected. Proton magnetic resonance
spectra (1H NMR) were measured with a Bruker DPX 250
MHz spectrometer, using TMS as internal reference.
4-(4,6-Dim et h oxy-1,3,5-t r ia zin -2-yl)-4-m et h ylm or p h o-
lin iu m ch lor id e (3). The solution of 2-chloro-4,6-dimethoxy-
1,3,5-triazine (1) (8.80 g, 50 mM) in benzene (30 mL) was
cooled to 0-5 °C, and N-methylmorpholine (5.5 mL, 50 mM)
was added dropwise. After 1 h, the white precipitate was
diluted with petroleum ether (70 mL), filtered, washed with
ether (3 × 30 mL), and dried in a vacuum desiccator under
KOH, P2O5, and paraffin turnings, affording 3 as a white
amorphous powder: yield 10.9 g, 79%; mp 120-122 °C dec
(lit.25 mp 90-98 °C); 1H NMR (DMSO-d6) δ 3.04 (s, 3H, NCH3),
3.60 (t, 4H, J ) 4.2 Hz, CCH2O), 3.71 (t, 4H, J ) 4.2 Hz,
CCH2N), 3.82 (s, 6H, OCH3); 13C NMR (DMSO-d6) δ 27.028
(NCH3), 44.424 (NCH2C), 54.975 (OCH3), 66.632 (OCH2C),
167.095 (NCN), 172.950 (NCO).
(25) Patent Sandoz Ltd.; BE 636 144; 1962; Chem. Abstr. 1965, 62,
1675f.
(26) Trost, B. M.; Vaultier, M.; Santiago, M. L. J . Am. Chem. Soc.
1980, 102, 7929-7932.
(27) Hamer, N. K. J . Chem. Soc. 1965, 46-51.
(28) Cooley, J . H.; Stone, D. M.; Oguri, H. J . Org. Chem. 1977, 42,
3096-3097.
(29) Partridge, M. W.; Stevens; M. F. G. J . Chem. Soc. 1964, 3663-
The PM3 Hamiltonian potential energy surface (PES)
for the dielectric constant equal to 20 is illustrated in
Figure 6. A similar PES was obtained in the gas-phase
3669.
(30) Derick, C. C.; Bornmann, H. J . Am. Chem. Soc. 1913, 35, 1269-
1289.

Activation of Carboxylic Acids
J . Org. Chem., Vol. 63, No. 13, 1998 4255
4-(4,6-Dip h en oxy-1,3,5-t r ia zin -2-yl)-4-m et h ylm or p h o-
lin iu m Ch lor id e (4). The solution of 2-chloro-4,6-diphenoxy-
1,3,5-triazine (2) (14.96 g, 50 mM) in benzene (40 mL) was
cooled to 0-5 °C, and N-methylmorpholine (5.5 mL, 50 mM)
was added dropwise. After 1 h, the white precipitate was
diluted with petroleum ether (70 mL), filtered, washed with
ether (3 × 30 mL), and dried in a vacuum desiccator under
KOH, P2O5, and paraffin turnings, affording 4 as a white
amorphous powder: yield 12.43 g, 62%; mp 140-153 °C dec;
1H NMR (CDCl3) δ 3.00 (s, 3H), 3.64 (m, 8H), 7.18 (m, 6H),
7.31 (m, 4H).
4-(4,6-Dim eth oxy-1,3,5-tr ia zin -2-yl)tetr a m eth ylgu a n i-
d in iu m Ch lor id e (5). The solution of 2-chloro-4,6-dimethoxy-
1,3,5-triazine (1) (8.80 g, 50 mM) in benzene (30 mL) was
cooled to 0-5 °C, and TMG (6.25 mL, 50 mM) was added
dropwise. After 1 h, the white precipitate was diluted with
petroleum ether (70 mL), filtered, washed with ether (3 × 30
mL), and dried in a vacuum desiccator under KOH, P2O5, and
paraffin turnings, affording 5 as a white amorphous powder:
yield 7.26 g, 35%; mp 171-179 °C; 1H NMR (CDCl3) δ 2.91 (s,
3H), 3.06 (s, 9H), 3.96 (s, 6H), 7.36 (s, 1H).
4-(4,6-Dip h en oxy-1,3,5-tr ia zin -2-yl)tetr a m eth ylgu a n i-
d in iu m Ch lor id e (6). The solution of 2-chloro-4,6-diphenoxy-
1,3,5-triazine (2) (14.96 g, 50 mM) in benzene (40 mL) was
cooled to 0-5 °C, and TMG (6.25 mL, 50 mM) was added
dropwise. After 1 h, the white precipitate was diluted with
petroleum ether (70 mL), filtered, washed with ether (3 × 30
mL), and dried in a vacuum desiccator under KOH, P2O5, and
paraffin turnings, affording 6 as a white amorphous powder:
yield 7.26 g, 35%; mp 199-212 °C; 1H NMR (CDCl3) δ 2.83 (s,
3H), 3.04 (s, 9H), 7.18 (m, 6H), 7.31 (m, 4H, o-C6H5), 8.82 (s,
1H).
Acyla tion of 4-(4,6-Dim eth oxy-1,3,5-tr ia zin -2-yl)-4-m e-
th ylm or p h olin iu m Ch lor id e. Gen er a l P r oced u r e. The
suspension of 4-(4,6-dimethoxy-1,3,5-triazin-2-yl)-4-methyl-
morpholinium chloride (3) (1.38 g, 5 mM) in dichloromethane
(10 mL) was vigorously stirred, cooled to 0 °C, and treated with
carboxylic acid (5 mM). The stirring was continued at 0 °C
overnight and then at room temperature for an additional 10
h. The resulting mixture was diluted with dichloromethane
(10 mL) and washed successively with water, 10% citric acid
solution, water, 0.5 M sodium bicarbonate solution, and then
water again, dried with anhydrous magnesium sulfate, filtered,
and evaporated to dryness. The solid residue was recrystal-
lized from cyclohexane, affording 7a ,b, which were found
chromatographically and spectroscopically to be identical to
authentic samples prepared by the standard procedure.
Syn th esis of Am id es 8a -f In volvin g th e Use of 4-(4,6-
Dim e t h oxy-1,3,5-t r ia zin -2-yl)-4-m e t h ylm or p h olin iu m
Ch lor id e (2) a s Cou p lin g Rea gen t. Gen er a l P r oced u r e.
The suspension of 4-(4,6-dimethoxy-1,3,5-triazin-2-yl)-4-me-
thylmorpholinium chloride (3) (1.38 g, 5 mM) in dichlo-
romethane (10 mL) was vigorously stirred, cooled to 0 °C, and
treated with carboxylic acid (5 mM). The stirring was con-
tinued at 0 °C overnight, an appropriate amine (5 mM) was
added dropwise, and stirring was continued for 3 h and then
at room temperature for an additional 10 h. The resulting
mixture was diluted with dichloromethane (10 mL) and
washed successively with water, 10% citric acid solution,
water, 0.5 M sodium bicarbonate solution, and then water
again, dried with anhydrous magnesium sulfate, filtered, and
evaporated to dryness. The solid residue was recrystallized
from ethyl acetate/petroleum, yielding 8a -f, which were found
chromatographically and spectroscopically to be identical to
authentic samples prepared by the typical procedure.16
Kin etic Meth od s. The rate of reaction of trimethylacetic
acid with CDMT was measured in CD3CN solution. Stock
solutions containing (a) a mixture of 1 M CDMT and 1 M
trimethylacetic acid, (b) 1 M N,N-dimethylaniline, (c) 1 M
triethylamine, (d) 1 M NMM, and (e) 1 M TMEDA in aceto-
nitrile-d3 were prepared. The reagents were conditioned at
the reaction temperature for 30 min before use, and then a-c
were injected into an NMR tube by microsyringe and diluted
with acetonitrile-d3 to obtain the required concentrations after
the addition of amine (solutions d and e). The samples were
conditioned for an additional 5 min, and then the reactions
were initiated by adding the appropriate solution of NMM or
TMEDA.
Ack n ow led gm en t. Financial support from the State
Committee for Sciencific Research (KBN) through Grant
No. 3.T09A.067.08 and Technical University of Lodz
Funds is gratefully acknowledged.
J O972020Y