Mechanism of the reaction of amines with 5-[(Aryl- or Alkylamino) hydroxymethylene]-2,2-dimethyl-1,3-dioxane-4,6-diones in the presence of chlorotrimethylsilane (Me3SiCl)

Article abstract of DOI:10.1002/hlca.201200326

Addition of chlorotrimethylsilane (Me₃SiCl) to the mixture of a carbamoyl-substituted Meldrum's acid, i.e., a 5-[(arylamino)hydroxymethylene]-2, 2-dimethyl-1,3-dioxane-4,6-dione of type 1 and a secondary amine as nucleophile strongly accelerated the rate of their reaction. The reason for this phenomenon observed, during our previous research, remained, however, unclear. To elucidate the mechanism of this reaction, we assumed and verified three possible pathways for the action of Me₃SiCl (cf. Scheme 2): The acceleration of the reaction is caused i) by formation of a O-trimethylsilylated Meldrum's acid of type 2, ii) by the silylated amine 3, or iii) by the presence of HCl liberated from Me₃SiCl. The performed experiments revealed that the faster course of reaction is caused by the formation of N-trimethylsilylated amines of type 3. Copyright

Full text of DOI:10.1002/hlca.201200326

978
Helvetica Chimica Acta – Vol. 96 (2013)
Mechanism of the Reaction of Amines with 5-[(Aryl- or
Alkylamino)hydroxymethylene]-2,2-dimethyl-1,3-dioxane-4,6-diones in the
Presence of Chlorotrimethylsilane (Me3SiCl)
by Karolina Janikowska, Sławomir Makowiec*, and Janusz Rachon´
Department of Organic Chemistry, Faculty of Chemistry, Gdansk University of Technology, Narutowicza
11/12, PL-80-952 Gdansk (phone: þ 48-58-3471724; fax: þ 48-58-3472694; e-mail: mak@pg.gda.pl)
Addition of chlorotrimethylsilane (Me3SiCl) to the mixture of a carbamoyl-substituted Meldrumꢀs
acid, i.e., a 5-[(arylamino)hydroxymethylene]-2,2-dimethyl-1,3-dioxane-4,6-dione of type 1 and a
secondary amine as nucleophile strongly accelerated the rate of their reaction. The reason for this
phenomenon observed, during our previous research, remained, however, unclear. To elucidate the
mechanism of this reaction, we assumed and verified three possible pathways for the action of Me3SiCl
(cf. Scheme 2): The acceleration of the reaction is caused i) by formation of a O-trimethylsilylated
Meldrumꢀs acid of type 2, ii) by the silylated amine 3, or iii) by the presence of HCl liberated from
Me3SiCl. The performed experiments revealed that the faster course of reaction is caused by the
formation of N-trimethylsilylated amines of type 3.
Introduction. – Meldrumꢀs acid derivatives have a broad scope of application in
organic synthesis [1]. The most explored feature of these compounds is the ability to
form ketenes in the course of thermal decomposition, which can be trapped with
various nucleophiles (Scheme 1). Depending on the type of Meldrumꢀs substrate,
pyrolysis can lead to formation of oxo ketenes [2], carbamoyl ketenes [3],
thiocarbamoyl ketenes [4], iminopropadienones [5], or even nitroso ketenes [6].
Practical application of the thus formed ketenes allow the preparation of various useful
compounds, as e.g., 3-substituted b-lactams [3][7], isoxazolols [8], pilicides [9], 1,3-
oxazinones [10], or derivatives of tetramic acid [11].
Scheme 1
In our laboratory, we have focused on the reactivity and synthetic application of
carbamoyl ketenes generated from 5-[hydroxy(aryl- or alkylamino)methylene]-2,2-
dimethyl-1,3-dioxane-4,6-diones. In contrast to acyl-substituted Meldrumꢀs acids, which
became the subject of many publications including detailed mechanistic studies [12],
ꢁ 2013 Verlag Helvetica Chimica Acta AG, Zꢂrich

Helvetica Chimica Acta – Vol. 96 (2013)
979
the reactivity of carbamoyl-substituted Meldrumꢀs acid is poorly described in the
literature. Excluding our own papers, there is only one example of a systematic study of
reactivity of carbamoyl-substituted Meldrumꢀs acids with nucleophiles presented by
Pak and co-workers [13]. However, this article dealt only with thermolytic reaction of
carbamoyl-substituted Meldrumꢀs acids with weakly basic N-nucleophiles such as
aromatic amines with electron-withdrawing groups or even amides. During our
experiments [14] with 5-[(aryl- or alkylamino)hydroxymethylene]-2,2-dimethyl-1,3-
dioxane-4,6-diones, we found that the use of stronger basic secondary amines caused
some problems such as low yield or a too long reaction time, difficulties that could be
bypassed in the presence of chlorotrimethylsilane (Me3SiCl). Thus, addition of
1.5 equiv. of Me3SiCl to the mixture of an amine and 5-[(aryl- or alkylamino)hydroxy-
methylene]-2,2-dimethyl-1,3-dioxane-4,6-dione led to the acylation product in nearly
quantative yield, concomitantly with a reduction of the reaction temperature by over
fifty degrees [14]. Initially, we used Me3SiCl intuitively as catalyst, treating this reagent
as an at least as good source of HCl as HCl itself. However, the actual effect of Me3SiCl
on the reaction remained unclear. We now wish to report on the elucidation of the
mechanism of the reaction of amines with carbamoyl-substituted Meldrumꢀs acids
promoted by Me3SiCl.
Results and Discussion. – At the beginning of this study, we assumed three
hypotheses that would explain the observed phenomenon (Scheme 2), i.e., the
acceleration of the reaction of 5-[hydroxy(phenylamino)methylene]-2,2-dimethyl-
1,3-dioxane-4,6-dione (1) with a secondary amine such as piperidine in the presence of
Me3SiCl to give the amide 4. Addition of Me3SiCl to 1 and the secondary amine may
cause the formation of two silylated species, namely 2,2-dimethyl-5-{(phenylamino)-
[(trimethylsilyl)oxy]methylene}-1,3-dioxane-4,6-dione (2) or (trimethylsilyl)amine 3.
Moreover, a stepwise addition of Me3SiCl to only one species, for example 1, does not
exclude the formation of the silylated amine 3 in a subsequent transsilylation process
after addition of the amine. Hence, we had to conceive a set of intelligent experiments
allowing to separate the reciprocal influence of all involved species.
In the first probing of our hypothesis, we assumed that initially formed 2,2-
dimethyl-5-{(phenylamino)[(trimethylsilyl)oxy]methylene}-1,3-dioxane-4,6-dione (2)
promotes the acceleration of the reaction (Scheme 2, Path A). This seems very likely
in view of the well known greater affinity of silylating agents towards an OH group
rather than to a N-nucleophile. However, such an O-silylated Meldrumꢀs acid should
rather slow down than speed up the reaction rate taking into account experiments,
where Meldrumꢀs acid derivatives which do not contain OH or SH groups were stable
up to 1308 for 4 – 5 h [15].
The second possibility for the acceleration of the reaction (Path B) might be the
presence of HCl liberated in the reaction of Me3SiCl with the residual water present in
the mixture. Grabowski and co-workers [12] have shown the influence of the acidity on
the decomposition rate of 1. Hence, we could not ad hoc exclude such a possibility for
the examined reaction mechanism.
The third assumption for the acceleration of the discussed reaction (Path C) implies
silylation of the secondary amine since silylated amines such as 3 are well known for
their strong nucleophilic properties which may speed up the reaction.

980
Helvetica Chimica Acta – Vol. 96 (2013)
Scheme 2. Hypothetical Paths for the Acceleration of the Reaction of 1 with Amines by Me3SiCl
The key issue to be elucidated was to determine which silylated species was formed
at the beginning of the reaction, O-silylated Meldrumꢀs acid, N-silylated amine, or
neither of them.
In a first experiment, we questioned Path A of the proposed mechanism, i.e., can 1
be O-silylated [16]. To answer this question, the mixture of 1, Et3N, and Me3SiCl in
1
(D6)benzene was analyzed by H-NMR spectroscopy after 17 h at room temperature.
Taking into account that a carbamoyl-substituted Meldrum acid has only three types of
H-atoms, we had to focus our observations on the Me groups of the acylal part. The
spectra showed only one Me signal at d(H) 1.24. Substitution at the O-atom, e.g., with
alkyl groups, has a weak, however, noticeable influence on the chemical shift of these
Me groups with a Dd of ca. 0.06 [17][18]. Hence, we recorded the 1H-NMR spectrum of
1 in (D6)benzene with the aim to reveal a difference between silylated carbamoyl-
substituted Meldrumꢀs acid 2 and 1. However, the 1H-NMR spectrum still showed only
one signal at d(H) 1.24. This fact strongly suggests that compound 2 was not formed.
The second diagnostic signal of Meldrumꢀs acid, which may exclude or confirm the
formation of 2, is the presence or disappearance of the signal of the acidic H-atom at
d(H) 16.4. However, in the presence of Et3N, this signal obviously disappears because

Helvetica Chimica Acta – Vol. 96 (2013)
981
of salt formation. Therefore, we performed another NMR experiment, in (D6)benzene
with a mixture of 1 and N,N’-bis(trimethylsilyl)urea, a silylating agent which allows for
silylation of carboxylic acids without the presence of a tertiary amine [19]. But again, in
1
the H-NMR spectrum only one Me signal was observed at d(H) 1.20, and even after
conducting this silylation reaction for 5 h, the unchanged signals of the OH and NH
groups were still present at d(H) 16.4 and 11.3, respectively. These results clearly
exclude Path A as a route for the acceleration of the reaction of carbamoyl-substituted
Meldrumꢀs acid with amines.
To obtain additional arguments, confirming or excluding the formation of 2 (Path
A), we performed a large-scale experiment with 0.1 mol of 1 in ClCH2CH2Cl, 1.1 equiv.
of Et3N, and 1.5 equiv. of Me3SiCl. After 24 h, ClCH2CH2Cl and all volatile products
were evaporated, and the residue was dissolved in dry Et2O and treated with an excess
of piperidine for additional 12 h. Then, the piperidinium salt of 1 was filtered, and Et2O
destilled off. The intention of this experiments is the following: if 2 was formed, it
should react with piperidine under formation of N-trimethylsilylated piperidine 3, or 2
would react with piperidine at elevated temperature by formation of malonamide 4.
However, distillation of the Et2O residue did not result even in traces of 3, and
malonamide 4 also was not formed. These results also exclude Path A.
The following two additional experiments also exclude Path A: When 1 was treated
in boiling ClCH2CH2Cl without and with addition of 1.5 equiv. of Me3SiCl, the time
required for decomposition of 1 was as long as 24 h, whereas, as previously reported
[14], in boilig ClCH2CH2Cl, 1 disappeared in the presence of Me3SiCl and amine within
2.5 h. This clearly indicates that Me3SiCl has no influence on the decomposition of 1 by
accelerating its decomposition to ketene or by any other way.
The next possibilty for the acceleration of the reaction may be associated with the
acidity of the mixture (Path B). As aforementioned, Grabowski and co-workers [12]
demonstrated that the rate of formation of ketenes from acyl-substituted Meldrumꢀs
acids directly depends on the concentration of the free acid form of the Meldrumꢀs acid
derivative. Therefore, taking into account that the highest observed difference for the
reaction with strong basic secondary amines, i.e., a very slow reaction without Me3SiCl
and a very fast one with 1.5 equiv. of Me3SiCl [14], led us to suppose that addition of
controlled amounts of a strong acid to the mixture through hydrolysis of Me3SiCl with
residual water may have a decisive influence on the rate of the reaction, just by
changing the acidꢀbase equilibrium. However, as previously reported [14], the reaction
of 1 with piperidine in ethylbenzene saturated with HCl was slightly faster, but the yield
of malonamide 4 still remained unsatisfactory. Nevertheless, this result cannot exclude
Path B, because during saturation with excess of HCl, all amine is converted into its
hydrochloride so that the rate of decomposition of 1 to ketene is obviously higher, but
there is a too low concentration of free amine to efficiently react with the formed
ketene. The addition of 1.5 equiv. of Me3SiCl to the mixture containing 2 equiv. of
amine may produce only up to 1.5 equiv. of amine hydrochloride and thus leave still
free amine. Therefore, we performed an experiment with 1 equiv. of 1, 1.5 equiv. of
piperidine hydrochloride, and 0.5 equiv. of free piperidine in boiling ClCH2CH2Cl.
Under these conditions, we observed that total decomposition of 1 took as long as 24 h
instead of 2.5 h as in the presence of Me3SiCl, and the obtained yield of malonamide 4
was lower than 70% in comparision with 96% in the presence of Me3SiCl. A ten times

982
Helvetica Chimica Acta – Vol. 96 (2013)
longer reaction time clearly excludes Path B, i.e., a simple acid catalysis of the
decomposition of 1 to ketene.
The last possibility for the acceleration of the reaction is the intervention of N-
silylated amine 3 (Path C). To confirm this route, we performed the following
experiment: 1 equiv. of 1 and 1.5 equiv. of 1-(trimethylsilyl)piperidine (3) were heated
in boiling ClCH2CH2Cl. After 2.5 h, all starting 1 was consumed, and the yield of 4 was
as high as 96%. This experiment shows that the use of pre-silylated amine or the use of
Me3SiCl results in the same yield and time required for the decomposition of 1. This
finding finally confirms Path C.
Having established that the acceleration of the reaction between 1 and an amine is
caused by the formation of a silylated amine in the first step, we focused on the
interaction of silylated amines with 1. Some observed facts can be summarized as
follows: after addition of Me3SiCl, the acceleration of the reaction was enormous – in
some cases, the conversion was ten times faster and, occurred at temperatures more
than 508 lower than in the absence of Me3SiCl. The influence of Me3SiCl was more
pronounced in the cases of N-aryl derivatives such as 1 than in the cases of N-alkyl
derivatives of type 1 and, finally, reaction of 1 with a secondary amine in the absence of
Me3SiCl in low-boiling solvents was extremely slow, as established, e.g., by the
complete decomposition of 1 requiring ca. 100 h in the reaction with Et2NH in
ClCH2CH2Cl. These facts strongly suggest that in case of the use of Me3SiCl/amine or
just N-silylated amine, we indeed observe a change of the reaction mechanism, from
formation of ketene and addition of a nucleophile to the additionꢀelimination-type of
reaction. This idea is opposite to views presented by Grabowski and co-workers [12],
which showed no dependence of the reaction rate on nucleophilicity and pointed to
ketene as an undoubtedly existing intermediate. However, they examined acyl-
substituted Meldrum acids, which are more prone to thermal decomposition to ketenes,
whereas carbamoyl-substituted Meldrum acids are more resistant. Although Shtaiwi
and Wentrup [5] have observed the formation iminopropanediones, formed from 1, but
under harsh conditions of FVT (flash vacuum thermolysis) at 350 – 5508.
Therefore, we postulate that at low temperatures when the formation of ketenes
from 1 is virtually put off (100 h in boiling ClCH2CH2Cl), an N-silylamine acts as a
strong nucleophile and reacts directly with the 1,3-dioxane-4,6-dione system through an
additionꢀelimination mechanism (Scheme 3), whereas in high-boiling solvents such as
ethylbenzene in the absence of Me3SiCl, 1 may react through the established
mechanism of decomposition to ketene. The addition of N-silylamine to the 1,3-
dioxane-4,6-dione system is facilitated when an N-aryl substituent is present, because in
the case of N-alkyl derivatives of type 1, we did not observe such a significant
acceleration.
In conclusion, the formation of N-trimethylsilylated amines is directly responsible
for the acceleration of the reaction of 5-[(aryl- or alkylamino)hydroxymethylene]-2,2-
dimethyl-1,3-dioxane-4,6-diones with amines. Me3SiCl has no influence on the rate of
the decompostion of 1. All facts strongly suggest that N-trimethylsilylated amines react
with 1 according to an additionꢀelimination mechanism rather that by formation of
free ketenes.

Helvetica Chimica Acta – Vol. 96 (2013)
983
Scheme 3. Reaction Mechanism of 1 with 1-(Trimethylsilyl)piperidine (3)
This work was financially supported by the Founds for Science in 2010 – 2011 (research project N
N204 088338).
Experimental Part
1. General. All solvents used in this study were dried over appropiate drying agents and distilled prior
to use. Commercially available reagents were purchased from SigmaꢀAldrich. Commercially unavailable
reagents were prepared by literature procedures: 5-[hydroxy(phenylamino)methylene]-2,2-dimethyl-
1,3-dioxane-4,6-dione (1) [17] and 1-(trimethylsilyl)piperidine [20]. TLC: Merck silica gel 60 F254. Flash
column chromatography (FC): Zeochem ZEOprep 60/40-63. M.p.: Warsztat Elektromechaniczny
(Warszawa); uncorrected. NMR Spectra (CDCl3): Varian Unity Plus 500; at 500 (1H): 500 and
125 MHz (13C); d in ppm rel. to Me4Si as internal standard, J in Hz. HR-ESI-MS: Micromass-Quattro-
LCT mass spectrometer; in m/z.
2. Attempts of Silylation of 5-[Hydroxy(phenylamino)methylene]-2,2-dimethyl-1,3-dioxane-4,6-
dione (1). To a stirred soln. of 1 (26.3 g, 0.1 mol) in CH2Cl2 under Ar and cooled to 08, Et3N
(15.29 ml, 0.11 mol) was added. After 15 min, Me3SiCl (16.27 g, 0.15 mol) was added dropwise. After
additional 15 min, the cooling bath was removed, and the mixture was stirred for 24 h. Then CH2Cl2 was
evaporated and the residue dissolved in dry Et2O. Piperidine (9.86 ml, 0.1 mol) was added to this soln.
and the mixture stirred at r.t. for 12 h. Thereafter, the crystalline precipitate of the carbamoyl-substituted
Meldrumꢀs acid salt with piperidine was filtered off. The Et2O soln. was concentrated without appearance
of any residue. The precipitate of Meldrumꢀs acid salt with piperidine was suspended in 2m HCl (200 ml),
and filtered, the solid washed with H2O and then dissolved in AcOEt, and the soln. dried (MgSO4) and
concentrated. Crystallization from AcOEt gave 10.73 g of recovered 1. M.p. 104 – 1068. Spectral data: in
agreement with [17].
3. 3-Oxo-N-phenyl-3-(piperidin-1-yl)propanamide (¼ b-Oxo-N-phenylpiperidine-1-propanamide; 4)
from Piperidine Hydrochloride. To a stirred soln. of 1 (263 mg, 1 mmol) in ClCH2CH2Cl, piperidine
(43 mg, 0.5 mmol) and piperidine hydrochloride (183 mg, 1.5 mmol) were added. The mixture was stirred

984
Helvetica Chimica Acta – Vol. 96 (2013)
and heated under reflux for 24 h. After decomposition of 1, the ClCH2CH2Cl was evaporated and the
1
residue purified by FC (AcOEt/hexane 5 :2): 172 mg (70%) of 4. M.p. 115 – 1178. H-NMR: 1.61 – 1.70
(m, 3 CH2); 3.49 (s, CH2); 3.52 – 3.65 (m, 2 CH2); 7.12 (t, J ¼ 7.3, 1 arom. H); 7.34 (t, J ¼ 7.8, 2 arom. H);
7.60 (d, J ¼ 7.8, 2 arom. H); 10.18 (s, NH). 13C-NMR: 24.5; 25.8; 26.7; 40.1; 43.6; 47.5; 120.3; 124.5; 129.2;
138.1; 164.6; 167.0. HR-ESI-MS: 246.1381 ([M þ Na]þ, C14H18N2NaO2þ ; calc. 246.1368).
4. Propanamide 4 from 1-(Trimethylsilyl)piperidine (3). As described in Exper. 3, with 1 (263 mg,
1 mmol), ClCH2CH2Cl, piperidine (43 mg, 0.5 mmol), and 1-(trimethylsilyl)piperidine (3; 235 mg,
1.5 mmol). The mixture was stirred and heated under reflux for 2.5 h. Workup as described in Exper. 3:
240 mg (96%) of 4. Data: identical to those of 4 from Exper. 3.
REFERENCES
[1] A. S. Ivanov, Chem. Soc. Rev. 2008, 37, 789.
[2] Y. Yamamoto, Y. Watanabe, S. Ohnishi, Chem. Pharm. Bull. 1987, 35, 1860.
[3] K. Janikowska, N. Pawelska, S. Makowiec, Synthesis 2011, 69.
´
[4] K. Janikowska, S. Makowiec, J. Rachon, Helv. Chim. Acta 2012, 95, 461.
[5] M. Shtaiwi, C. Wentrup, J. Org. Chem. 2002, 67, 8558.
[6] H. Matsui, E. J. Zuckerman, N. Katagiri, C. Kaneko, S. Ham, D. M. Birney, J. Phys. Chem. A. 1997,
101, 3936.
[7] Y. Yamamoto, Y. Watanabe, Chem. Pharm. Bull. 1987, 35, 1871.
[8] U. S. Sorensen, E. Falch, P. Krogsgaard-Larsen, J. Org. Chem. 2000, 65, 1003.
[9] H. Emtenas, L. Alderin, F. Almqvist, J. Org. Chem. 2001, 66, 6756; M. Sellstedt, F. Almqvist, Org.
Lett. 2008, 10, 4005.
[10] N. Pemberton, H. Emtenꢃs, D. Bostrçm, P. J. Domaille, W. A. Greenberg, M. D. Levin, Z. Zhu, F.
Almqvist, Org. Lett. 2005, 7, 1019.
ˇ
ˇ
ˇ
ˇ
[11] S. Pirc, D. Bevk, R. Jakse, S. Recnik, L. Golic, A. Golobic, A. Meden, B. Stanovnik, J. Svete,
Synthesis 2005, 2969.
[12] F. Xu, J. D. Armstrong, J. D. Zhou, B. Simmons, D. Hughes, Z. Ge, E. J. Grabowski, J. Am. Chem.
Soc. 2004, 126, 13002.
[13] H. L. Lee, J. P. Lee, G. H. Lee, C. S. Pak, Synlett 1996, 12, 1209.
[14] K. Janikowska, S. Makowiec, Synth. Commun. 2012, 42, 975.
[15] F. Ye, B. Chen, X. Huang, Synthesis 1989, 5, 317.
´
[16] K. Janikowska, S. Makowiec, P. Punda, N. Pawelska, J. Rachon, ꢄ15th International Electronic
Conference on Synthetic Organic Chemistryꢀ, http://www.sciforum.net/conf/ecsoc-15.
´
[17] J. K. Mukhopadhyaya, S. Sklenak, Z. Rappoport, J. Am. Chem. Soc. 2000, 122, 1325.
[18] B. E. Fulloon, C. Wentrup, J. Org. Chem. 1996, 61, 1363.
[19] W. Verboom, G. W. Visser, D. N. Reinhoudt, Synthesis 1981, 10, 807.
[20] H. Yoshida, T. Morishita, H. Fukushima, J. Ohshita, A. Kunai, Org. Lett. 2007, 17, 3367.
Received June 26, 2012