Strong influence of intramolecular Si?O proximity on reactivity: Systematic molecular structure, solvolysis, and mechanistic study of cyclic N-trimethylsilyl carboxamide derivatives

Article abstract of DOI:10.1016/j.tetlet.2017.04.057

A comparative alcoholysis study of N-silylated derivatives of simple heterocyclic carboxamides (lactams, imides, ureas) is presented. The second-order rate constant values span a range as wide as three orders of magnitude. On the basis of DFT calculations, a good correlation between reactivity and the Si?O distance was found within each family of compounds. The viability of two different reaction pathways was evaluated using a detailed computational mechanistic study of the methanolysis of cyclic urea homologues. Peculiarities in the single-crystal X-ray diffraction structures of the trimethylsilyl and trimethylsiloxy phthalimides are also discussed.

Full text of DOI:10.1016/j.tetlet.2017.04.057

Accepted Manuscript
Strong influence of intramolecular Si...O proximity on reactivity: systematic
molecular structure, solvolysis, and mechanistic study of cyclic N-trimethylsilyl
carboxamide derivatives
Roland Szalay, Veronika Harmat, János Eőri, Gábor Pongor
PII:
S0040-4039(17)30499-9
DOI:
http://dx.doi.org/10.1016/j.tetlet.2017.04.057
TETL 48849
Reference:
To appear in:
Tetrahedron Letters
Received Date:
Revised Date:
Accepted Date:
30 January 2017
12 April 2017
18 April 2017
Please cite this article as: Szalay, R., Harmat, V., Eőri, J., Pongor, G., Strong influence of intramolecular Si...O
proximity on reactivity: systematic molecular structure, solvolysis, and mechanistic study of cyclic N-trimethylsilyl
carboxamide derivatives, Tetrahedron Letters (2017), doi: http://dx.doi.org/10.1016/j.tetlet.2017.04.057
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Strong influence of intramolecular Si...O proximity on reactivity:
systematic molecular structure, solvolysis, and mechanistic study of cyclic
N-trimethylsilyl carboxamide derivatives
a*
a,b
a
a
Roland Szalay , Veronika Harmat , János Eőri , Gábor Pongor
Institute of Chemistry, Eötvös Loránd University, Budapest
a
b
MTA-ELTE Protein Modeling Research Group, Budapest
th
Dedicated to Dezső Knausz for his 80 birthday
Abstract
A comparative alcoholysis study of N-silylated derivatives of simple heterocyclic
carboxamides (lactams, imides, ureas) is presented. The second-order rate constant values
span a range as wide as three orders of magnitude. On the basis of DFT calculations, a good
correlation between reactivity and the Si...O distance was found within each family of
compounds. The viability of two different reaction pathways was evaluated using a detailed
computational mechanistic study of the methanolysis of cyclic urea homologues. Peculiarities
in the single-crystal X-ray diffraction structures of the trimethylsilyl and trimethylsiloxy
phthalimides are also discussed.
Graphical abstract
Keywords
Silyl; Amide; Solvolysis; Reactivity; Intramolecular; DFT
Introduction
Silylated carboxamide derivatives represent a notable group within organosilicon
1
compounds since they are extensively used as silylating agents for hydrogen-bonded, non-
volatile analytes in gas-phase analytical techniques (e.g. gas chromatography and mass
2
3
spectrometry) and as protected/activated intermediates in multi-step organic syntheses.
Moreover, the unsilylated parent compounds (e.g. ureas) have gained increased interest in
4
applications such as organocatalysis and supramolecular chemistry, hence their controlled
release from Si-protected (and lipophilic) precursors by protodesilylation reactions (e.g.
1

hydrolysis, alcoholysis, fluoride-induced cleavage) under less conventional (e.g. biphasic,
5
phase-transfer catalytic, micelle-forming) conditions may afford novel synthetic routes.
In nucleophilic substitution reactions the increased reactivity of penta- and hexacoordinate
6
silicon atoms over tetracoordinate ones has long been recognized. One of the first reports
concerning the unique reactivity of silylated amides, for example, over silyl amines, was
published by Klebe who presumed a switch from the pentacoordinated configuration of the
silicon center to the hexacoordinated one upon nucleophilic attack by a proton donor onto the
7
Si atom. In a later study, Lane and Frye determined the relative thermodynamic silylating
abilities of trimethylsilylated N-alkyl amides and observed significantly different stability for
the silylated pyrrolidone and ε-caprolactam, respectively; however this phenomenon was not
8
fully rationalized.
Figure 1. Compounds concerned in this work.
In our previous paper we reported the surprisingly different reactivity of N-silylated cyclic
ureas with 5- (TMS
2
Im), 6- (TMS
2
Pyr), and 7- (TMS Diaz) membered rings during
2
9
hydrolysis and alcoholysis reactions. Both experimental (single-crystal X-ray diffraction) and
quantum chemical studies revealed a strong relationship between the ground-state molecular
9,10
structure of the substrates and their silylating power.
Herein, we present an extension of
this study relating to silylated cyclic carboxamides, interpreting the new results in a wider
perspective (Fig. 1). The octanolysis rate of a series of silyl lactams (TMSBL, TMSVL,
TMSCL) and dicarboxylic acid imides (TMSSI, DMSBPI, TMSPI, TBDMSPI, TMSOPI,
9
TMSGI, TMSNI) was determined under similar conditions as previously described (Scheme
1
).
Scheme 1. Octanolysis reaction of silylated cyclic carboxamides.
In order to gain a deeper insight into the ring-size dependent solvolytic reactivity, a
detailed computational mechanistic study of the methanolysis of three N,N’-disilylated urea
homologues (TMS
2
Im, TMS
2
Pyr, TMS
2
Diaz) is also discussed.
Results and Discussion
Syntheses
2

The silylated lactams (TMSBL, TMSVL, TMSCL) and TMSPI were obtained from the
reaction between the corresponding N-H compound and trimethyl chlorosilane in the presence
of triethylamine as a proton scavenger. Unlike aromatic (benzene, toluene) or ether (Et
2
O,
1
1
dioxane) solvents commonly used for such reactions, CH
2
Cl
2
was selected as being one of
1
2
the best solvents for silylation. DMSPI and TBDMSPI were similarly prepared from
phthalimide and dimethyl or tert-butyldimethyl chlorosilane. In the case of TMSNI,
2
naphthalimide was first converted into its sodium salt using NaNH , then silylated using a
THF solution of Me SiCl in excess.
3
Both alicyclic silyl imides (TMSSI and TMSGI) were synthesized by silylation of the
corresponding imide with N,O-bis(trimethylsilyl)-trifluoroacetamide (BSTFA). Several
attempts with commonly used basic silylating agents (e.g. Me
3
SiCl/Et N, hexamethyl
3
disilazane, trimethylsilyl N,N-dimethyl carbamate) failed to give TMSOPI in satisfactory
yield since the reversibly formed red-coloured ammonium salt of the deprotonated N-hydroxy
phthalimide precipitated out of the mixture and therefore could not be efficiently silylated.
Using the more powerful reagent BSTFA led to the desired product. To the best of our
knowledge, TMSNI and TMSOPI are novel compounds.
Octanolysis study of silylated lactams and imides
Rate measurements were carried out at room temperature in 1-octanol/THF mixtures
containing the substrate in 100 times dilution compared to the alcohol. Solvolysis reactions
could be conveniently followed using gas chromatography by monitoring the relative peak
area of the trimethylsiloxy octane (OctOTMS) formed (Fig. 2).
0
0
0
.4
.3
.2
0
50
100
time/h
Figure 2. GC peak area of OctOTMS vs time for the octanolysis reaction of TMSPI.
The peak intensity vs time data were evaluated surmising a pseudo-first order kinetic
model:
−
k
t
exp
X = X (1− ae
),
∞
3

where X, X , kexp, t and a denote the actual and final peak area of the OctOTMS (normalised
∞
to the internal standard), the pseudo-first order rate constant, the reaction time and a constant
parameter, respectively (see the literature for a physical interpretation of the latter
9
parameter ). The bimolecular rate constant (k
2
) independent of the alcohol concentration was
= kexp/[OctOH].
Considering the wide range of several orders in magnitude spanned by k s, the pk ≡ -
/h mol dm ) values are instead listed, together with the Gibbs free energies of
determined in order to characterise the reactivity of substrates where k
2
2
2
-1
-1
3
log10(k
2
activation calculated from the Eyring–Polányi equation wherein the transmission coefficient
is taken as the unity (Table 1).
-1
Firstly, it should be mentioned, for example, that the difference of 1.7 kcal mol between
‡
the ∆ G values for the five-membered (TMSBL) and the six-membered (TMSVL) silyl
lactams is reflected in the 16 times faster alcoholysis of the latter compound, whereas the
-1
difference of at least 3.4 kcal mol between the corresponding ring-membered silyl ureas
indicates a 300 times difference in reaction rates. Upon considering the pk values themselves,
2
it should also be emphasized that even the most reactive compounds examined in this work
are not necessarily anticipated to be as efficient as commercially available silylating reagents
(
e.g. BSA, BSTFA). The original aim of this study was only to show how a slight
modification in molecular structure, such as ring expansion remote from the reaction center
Si atom), can influence the silyl transfer capability. On the basis of our results, more efficient
(
and tailor-made silylating agents can be systematically designed, potentially outperforming
commonly used reagents.
-
1
-1
3
Table 1. pk
2
≡ -log10(k /h mol dm ) values for the octanolysis reaction and the Gibbs free
2
‡
-1
energies of activation (∆ G/kcal mol ) calculated from the Eyring–Polányi equation (with the
transmission coefficient of the unity).
‡
‡
Lactams
TMSBL
TMSVL
TMSCL
Ureas
pk
2
∆ G
Imides
pk
3.0
2.7
1.9
0.8
0.6
> 4.0
2.9
2
∆ G
4.0
2.8
2.5
27.8
26.1
25.7
TMSPI
TMSSI
TMSGI
TMSNI
DMSPI
TBDMSPI
TMSOPI
26.4
26.0
24.9
23.4
23.1
‡
pk
2
∆ G
TMS
TMS
TMS
2
Im9
Pyr9
Diaz9
> 3.5
1.0
1.3
> 27.1
23.7
2
> 27.8
26.3
2
24.1
Secondly, it is apparent from Table 1 (comprising of relevant data for the urea rings for
comparison) that there is also a significant but somewhat smaller difference in reactivity
between the five- and six-/seven-membered cyclic silylated carboxamides, as was previously
reported for silyl ureas. The five-membered rings are less sensitive to nucleophilic attack by
the alcohol than their six- or seven-membered congeners; however, in contrast to lactams, the
reactivity of the two alicyclic imides (TMSSI and TMSGI) show a diminished dependency
upon ring-size. For the six-membered imide ring, a switch from the pure alicyclic skeleton
(
TMSGI) having only one out-of-plane carbon at the β-position to the all-planar naphthalene-
condensed one (TMSNI) significantly enhances the rate of protodesilylation (>10 times).
Furthermore, within the series of N-silylated phthalimides, the reactivity decreases in the
order DMSPI >> TMSPI >> TBDMSPI, corresponding with the increasing bulkiness of the
silyl group as expected. The only compound bearing an O-Si bond, TMSOPI, shows a
reactivity pattern typical to aliphatic O-silylating agents.
Molecular structure studies
4

In the sense of our simple static (structure-based) model previously developed for
1
0
predicting the silylating power of O- and N-silylated carbamic acid derivatives, we explored
the relationship between the molecular structure of the title-compounds and their silyl transfer
capability, thus providing a starting point to elucidating the detailed mechanism using
computational methods (see next section). Out of the five solid compounds, single-crystal X-
ray diffraction structures could only be determined for the three phthalimide derivatives
(
TMSPI, TBDMSPI, TMSOPI; see Fig. 3-5, ESI, and Table 2), mainly due to the extreme
moisture-sensitivity of the other two compounds (TMSNI, DMSPI). For all N-
trimethysilylated lactams and imides, the ground state equilibrium structures were calculated
by quantum chemical (DFT) methods.
In the crystal phase all determined molecular structures are dominated by the largely planar
phthalimido moiety. In TMSPI (Fig. 3) the aromatic and five-membered rings, together with
their directly attached atoms and the C2 methyl carbon, lie within the crystallographic mirror
plane bisecting the C-Si-C angle (C1 Si1 C1) formed by the other two (C1) out-of-plane
1
methyl carbons, thus rendering half of the molecule as the asymmetric unit (P2 /m space
group). The considerably smaller (by 9°) exocyclic C-N-Si angle (C3 N1 Si1) may lead to a
more pronounced Si...O interaction between the silicon center and the corresponding carbonyl
oxygen (O1). Accordingly, the aforementioned oxygen, together with the in-plane methyl
carbon (C2) located on the opposite side represent the quasi apical ligand atoms around a very
slightly distorted tetra-to-penta-coordinate silicon center. It is worth noting that based upon
1
3
examining the Cambridge Structural Database (CSD), such a significant difference between
pairs of the aforementioned angle is unique to N-substituted phthalimide derivatives wherein
the “isolated” substituent has C (n > 1) local symmetry (only 4 hits out of 2506 structures). In
n
two of these cases (CSD codes: ICIMAK and IMIGUH) the phthalimido moiety substituent
was Sn and one of the three substituents of the tin atom lies within the phthalimido plane,
quite similar to the TMSPI structure, suggesting the steric effect discussed below may be
effective in Sn-containing analogues. In one of the other two exceptions (CUBJUG),
geometric restraints applied for resolving structural disorder might cause the large difference
in angles (23.0º), while the other (IMIHIW), is a phthalimido Au complex with the N-Au-X
angle being close to 180°. In the crystal all the phthalimido moieties are parallel with the a-c
plane of the unit cell and intermolecular interactions being dominated by aromatic stacking
interactions, as well as C-H…O weak, bifurcated hydrogen bonds established by the
phthalimido C-H moieties and oxygen atoms not involved in Si…O interactions along the unit
cell axis a.
Figure 3. ORTEP drawing of TMSPI at the 30% probability level. Selected interatomic
distances [Å] and angles [°]: Si1 N1 1.779(2), Si1 C1 1.801(3), Si1 C2 1.808(4), O1 C3
1
1
.205(4), O2 C4 1.196(4), N1 C3 1.371(4), N1 C4 1.376(4); N1 Si1 C1 107.9(1), C1 Si1 C1
11.2(2), N1 Si1 C2 108.9(2), C1 Si1 C2 110.4(1), C3 N1 C4 108.8(2), C3 N1 Si1 121.0(2),
C4 N1 Si1 130.2(2), O1 C3 N1 124.4(3), O2 C4 N1 125.4(3).
5

Unlike TMSPI, neither TBDMSPI (Fig. 4) nor TMSOPI (Fig. 5) possess a crystallographic
mirror plane, however, a quasi-mirror plane perpendicular to the phthalimido skeleton can be
fairly defined which is indicated by the close-to-the-right angles (87.4°/85.8°) between the C-
N-C and the N-Si-Ctert/N-O-Si planes, respectively. Furthermore, the latter planes virtually
also comprise one more carbon atom since the N-Si-Ctert-C/N-O-Si-C torsion angles are anti-
periplanar to within -178.3°/178.0° in TBDMSPI/TMSOPI. The crystal packing shows C-
H…O weak hydrogen bonds between the phthalimido C-H moieties and oxygen atoms, and
C-H…π type interactions between the methyl groups and aromatic rings in both TBDMSPI
and TMSOPI. The phthalimido planes of TMSOPI are quasi parallel to the b-c plane of the
unit cell.
Figure 4. ORTEP drawing of TBDMSPI at the 30% probability level. Selected interatomic
distances [Å] and angles [°]: Si1 N1 1.783(7), Si1 C3 1.816(11), Si1 C1 1.838(12), Si1 C2
1
.825(11), O1 C7 1.179(9), O2 C8 1.186(8), N1 C7 1.403(9); N1 C8 1.407(9); N1 Si1 C3
1
06.7(3), N1 Si1 C1 107.8(3), C3 Si1 C1 112.2(5), N1 Si1 C2 107.8(4), C3 Si1 C2 112.9(5),
C1 Si1 C2 109.2(4), C7 N1 C8 108.0(6), C7 N1 Si1 124.9(4), C8 N1 Si1 126.9(4), O1 C7 N1
24.2(6), O2 C8 N1 123.8(6).
1
Figure 5. ORTEP drawing of TMSOPI at the 30% probability level. Selected interatomic
distances [Å] and angles [°]: Si1 O3 1.703(2), Si1 C3 1.843(3), Si1 C1 1.847(3), Si1 C2
1
.844(3), O1 C4 1.192(3), O2 C5 1.199(3), O3 N1 1.387(3), N1 C4 1.400(4), N1 C5 1.403(3);
O3 Si1 C3 100.6(1), O3 Si1 C1 110.6(1), C3 Si1 C1 113.2(2), O3 Si1 C2 109.9(1), C3 Si1 C2
12.8(2), C3 Si1 C1 113.3(2), N1 O3 Si1 115.8(1), O3 N1 C4 121.1(2), O3 N1 C5 120.8(2),
1
C4 N1 C5 113.8(2), O1 C4 N1 125.6(2), O2 C5 N1 125.6(2).
Upon comparing the sums of angles around nitrogen (TMSPI: 360.0°, TBDMSPI: 359.8°,
TMSOPI: 355.7°), a significant deviance from planarity is observed in TMSOPI, similar to
other N-O bonded compounds bearing an amide-type nitrogen. This phenomenon is due to the
1
4
widely accepted lone pair repulsion between the N and O atoms. Based upon examining the
Cambridge Structural Database (CSD), structures with a nitrogen attached to two carbonyl
6

groups, i.e. dicarboxylic acid imide derivatives, showed the sum-of-angles value of TMSOPI
to be the 16th smallest out of the 240 hits. This extent of non-planarity is especially
noteworthy in light of the partial delocalization of the O electron density to Si (N-O-Si angle
is 115.8°) which might be expected to reduce the stereochemical activity of the oxygen lone
pair. Moreover, when taking all N-substituted derivatives of phthalimides and their congeners
into consideration, the sum-of-angles value for TMSOPI proved to be the 24th smallest out of
4
076 hits.
The DFT calculation based (i.e. gas phase) molecular structures exhibit some common
features (see ESI). First it should be noted that, due to the extensively conjugated
carboxamide moiety and the cyclic connectivity, the molecules are fully, or at least closely
flat. Furthermore, in all N-TMS derivatives, except for TMSGI, one of the three methyl
carbons of the TMS group being farthest away from the oxygen atom virtually lies in the
carboxamide plane, similar to the N-silylated ureas previously reported. In a steric and
stereoelectronic sense, this arrangement of the (C)-Si-N-C=O atoms makes a Si...O
1
0
interaction (see the literature for a detailed theoretical discussion ) feasible with some Si-N
bond lengthening which is crucial to the silyl transfer ability of substrates. Within the same
group of compounds (N-silylated lactams, imides, and ureas), a positive correlation is clearly
manifested between the higher reactivity (i.e. lower pk
longer Si-N bond length (Fig. 6), while a similar and even more correlated relationship with
opposite sign is shown by the pk vs Si...O distance data points (Fig. 7). These trends provide
further evidence to the applicability of our structure-based model previously reported for
2
) in octanolysis and the calculated
2
1
0
predicting the silylating power of silylated carboxamide derivatives.
lactams
ureas
imides
4
3
3
2
2
1
1
0
.0
.5
.0
.5
.0
.5
.0
.5
1
.79
1.80
1.81
1.82
1.83
1.84
1.85
1.86
1.87
d(Si-N)/Å
Figure 6. Correlation between pk
2
values for the octanolysis reaction and the Si-N bond
lengths in substrates.
7

lactams
ureas
imides
4
3
3
2
2
1
1
0
.0
.5
.0
.5
.0
.5
.0
.5
2
.8
2.9
3.0
3.1
d(Si...O)/Å
2
Figure 7. Correlation between pk values for the octanolysis reaction and the Si...O distances
in substrates.
Computational mechanistic study of the methanolysis of N,N’-disilylated ureas
In nucleophilic reactions of organosilicon substrates, enhanced reactivity of the
1
5
pentacoordinated silicon over the tetracoordinated one is widely accepted. Previously we
reported large differences in the rate of alcoholysis between silylated five- and six-/seven-
9
membered cyclic ureas. On the basis of our results related to the experimental and theoretical
molecular structures, we concluded that the Si atom adopts a more pronounced (pseudo)
pentacoordinate geometry in the more reactive six-/seven-membered rings (e.g. TMS
2
Pyr,
TMS Diaz) than in the less reactive five-membered ones (e.g. TMS Im). In order to refine
2
2
and extend our mechanistic view on the alcohol-induced protodesilylation, a detailed picture
of the reaction between methanol, for the sake of simplicity, and disilylated cyclic ureas is
presented using quantum chemical (DFT) calculations.
Considering the aforementioned close vicinity of the silicon and oxygen atoms, two
fundamentally different pathways for the methanolysis reaction can be surmised: (1) the
“
direct” route via a single transition state connecting the MeOH and N,N’-bis(trimethysilyl)
urea molecules directly to methoxytrimethylsilane (TMSOMe) and the enol form of N-
trimethylsilyl urea; and (2) an “indirect” two-step route, where an N-to-O silyl migration step
occurs in the N,N’-disilylated urea reactant prior to methanol-induced O-Si bond cleavage of
the N,O-bis(trimethylsilyl) isourea intermediate which results in the keto form of N-silyl urea
beside TMSOMe (Fig. 8-10; for TMS
2
Diaz see ESI).
8

2
Figure 8. B3LYP/6-31+G* energy profile of the reaction of TMS Im with MeOH
-1
(
Route 1 shown in blue, Route 2 in red). Values are displayed in kcal mol .
2
Figure 9. B3LYP/6-31+G* energy profile of the reaction of TMS Pyr with MeOH
-1
(
Route 1 shown in blue, Route 2 in red). Values are displayed in kcal mol .
9

Figure 10. B3LYP/6-31+G* energy profile of the keto-enol tautomerisation reaction of
TMS Im and TMS Pyr (unimolecular pathways shown in red, methanol-assisted pathways in
2
2
-
1
blue). Values are displayed in kcal mol .
First, the relative energies of the N,N′- and N,O-disilylated urea isomers having 5-, 6-, and
7
-membered ring were determined. The obtained values show that the latter isomers are less
stable in all cases, however, the very close values for 6- and 7-membered rings (6.0 and 5.9
-1
-1
kcal mol ) are significantly lower than that of the 5-membered ring (9.3 kcal mol ). During
separate calculations concerning the relative stability of monosilylated non-cyclic urea
3
rotamers bearing the Me Si-N-C=O moiety, “cis” arrangement of the Si and O atoms was
found to be preferable to the “trans” one, and furthermore, the energy gap increased along
with a decreasing N-C=O angle. By widening the N-C-N angle due to ring expansion, the
prominent strength of the Si-O bond becomes dominant in the 6- and 7-membered rings.
Interestingly, in contrast to the discussions above, bond critical point analysis was not able to
detect any bond between Si and O even in these higher homologues. For the N-to-O silyl
migration, much higher differences in energy of the relevant transition states were observed
-1
between the 5- and 6-/7-membered rings: 28.5 kcal mol for TMS
mol for TMS Pyr/TMS
2 2
2
Im, and 17.8/15.8 kcal
Diaz. These results are in accordance with our structure-based
reactivity model which predicts higher silyl group mobility for substrates bearing higher
degree of (pseudo) pentacoordination around the Si center (cf. TBP% values for TMS Im,
-1
2
9
2 2
TMS Pyr, and TMS Diaz are 5, 20, and 17, respectively).
Complexation of the disilyl urea/isourea substrate with one molecule of MeOH is
-1
exothermic (6-8 kcal mol ) in all cases. The transition states of the Me
reaction built-up from the silyl urea–MeOH adduct were determined for both the direct (1)
and indirect (2) route. In route 1 the rather high activation barriers (TMS Im: 30.4, TMS Pyr:
Diaz: 34.2 kcal mol ) give little possibility that the reaction proceeds in this
3
Si–H exchange
2
2
-1
3
1.0, and TMS
2
fashion, and moreover, the trend in values is opposite to what is expected from the rate
measurements. On the contrary, route 2 provides transition states with significantly lower
-
1
2 2 2
activation energy (TMS Im: 20.5, TMS Pyr: 18.6, and TMS Diaz: 18.4 kcal mol ), making
this pathway much more favourable for the methanol-induced desilylation. These findings are
supported by the enhanced silyl transfer ability of O-silyl isoureas with the =N...Si-O moiety
1
0
over O-silyl carbamates with the =O...Si-O moiety as previously reported. In sharp contrast
to the 6- and 7-membered rings (TMS Pyr and TMS Diaz), the high barrier of the preceding
N-to-O silyl isomerisation step dramatically reduces the overall reaction rate of the 5-
membered cyclic compound (TMS Im). In summary, the enhanced reactivity of higher
2
2
2
1
0

membered (>5) cyclic N-silylated ureas in the desilylation process can be interpreted in terms
of the “springboard” role of the carbonyl oxygen acting as an intramolecular nucleophilic
catalytic site towards the silicon center.
Conclusion
The structure-based reactivity model previously developed by our group to explain the
high silylating power of silyl carbamates, and refined here using a detailed computational
mechanistic study of silylated cyclic ureas, proved to be a good starting point to predict the
relative rates of silyl transfer reactions for a wider range of substrates. The structural
constraints imposed on the accessibility of the carbonyl oxygen for the silicon center play a
decisive role in the ease of desilylation. On the basis of the calculated molecular structures,
the reactivity of silyl protected/activated intermediates becomes finely tuneable in chemo-
and/or regioselective organic transformations which is especially important in the design of
highly valuable synthetic targets containing several functional groups.
Acknowledgements
R. Sz. is very grateful to Oldamur Hollóczki for his generous disclosure of computational
results relating to the methanolysis reaction. R. Sz. wishes to thank Veronika Kudar and
Mária Darvas for their laboratory efforts as undergraduate students, and Zoltán Nyiri for GC-
MS measurements. R. Sz. and G. P. acknowledge the OTKA Research Foundation of
Hungary F030821 & T037658, and Chemical Works of Gedeon Richter Plc. for financial
support.
References
1
1

Strong influence of intramolecular Si...O proximity on reactivity:
systematic molecular structure, solvolysis, and mechanistic study of cyclic
N-trimethylsilyl carboxamide derivatives
a
a,b
a
a
Roland Szalay , Veronika Harmat , János Eőri , Gábor Pongor
Institute of Chemistry, Eötvös Loránd University, Budapest
a
b
MTA-ELTE Protein Modeling Research Group, Budapest
th
Dedicated to Dezső Knausz for his 80 birthday
Highlights
•
•
•
•
Wide range of reactivity in alcohol-induced protodesilylation
Good correlation between reactivity and the Si...O proximity
Two plausible concurrent reaction pathways
Noteworthy XRD structural features of silylated phthalimides
1
Reviews: (a) El Gihani MT, Heaney H. Synthesis. 1998; 357-375. (b) van Look G, Simchen G, Heberle J.
Silylating Agents, Fluka Chemie AG: Buchs, Switzerland, 1995. (c) Wuts, P. G. M. Greene’s protective groups
th
in organic synthesis, John Wiley & Sons, Inc., Hoboken: New Jersey, 5 ed., 2014.
2
Some representative examples: (a) Schettgen T, Dewes P, Kraus T. Anal Bioanal Chem. 2016; 408: 5467-5478.
(
b) Molnar B, Fodor B, Boldizsar I, Molnar-Perl I. J Chromatogr, A. 2016; 1440: 172-178. (c) Manzo V, Ulisse
K, Rodriguez I, Pereira E, Richter P. Anal Chim Acta. 2015; 889: 130-137. (d) Raeppel C, Fabritius M, Nief M,
Appenzeller BMR, Millet M. Talanta. 2014; 121: 24-29. (e) Cacho JI, Campillo N, Vinas P, Hernandez-Cordoba
M. J Chromatogr, A. 2012; 1247: 146-153. (f) Kovacs A, Mortl M, Kende A. Microchem J. 2011; 99: 125-131.
3
Some representative examples: (a) Homerin G, Baudelet D, Dufrenoy P, Rigo B, Lipka E, Dezitter X, Furman
C, Millet R, Ghinet A. Tetrahedron Lett. 2016; 57: 1165-1170. (b) Liu GJ, Zhang XT, Xing GW. Chem
Commun. 2015; 51: 12803-12806. (c) Wahler K, Ludewig A, Szabo P, Harms K, Meggers E. Eur J Inorg Chem.
2
014; 807-811. (d) Blanck S, Geisselbrecht Y, Kraling K, Middel S, Mietke T, Harms K, Essen LO, Meggers E.
Dalton Trans. 2012; 41: 9337-9348. (e) Kardon F, Moertl M, Csampai A, Ujszaszy K, Knausz D. Synth
Commun. 2011; 41: 914-924. (f) Trost BM, Zhang T. Angew Chem, Int Ed. 2008; 47: 3759-3761.
4
Volz N, Clayden J. Angew Chem, Int Ed. 2011; 50: 12148-12155.
Szalay R. personal note.
Reviews: (a) Chuit C, Corriu RJP, Reye C, Young JC. Chem Rev. 1993; 93: 1371-1448. (b) Wagler, J.; Bohme,
5
6
U.; Kroke, E. In Functional Molecular Silicon Compounds I: Regular Oxidation States; Scheschkewitz, D., Ed.;
Springer: New York, 2014, pp 29-105. Some representative examples: (c) Nikolin AA, Kramarova EP, Shipov
AG, Baukov YJ,Negrebetsky VV, Arkhipov DE, Korlyukov AA, Lagunin AA, Bylikin SY, Bassindale AR,
Taylor PG. Rsc Adv. 2016; 6: 75315-75327. (d) Steinhauer S, Stammler H-G, Neumann B, Ignat'ev N, Hoge B.
Angew Chem, Int Ed. 2014; 53: 562-564. (e) Azhakar R, Ghadwal RS, Roesky HW, Granitzka M, Stalke D.
Organometallics. 2012; 31: 5506-5510. (f) Metz S, Theis B, Burschka C, Tacke R. Chem—Eur J. 2010; 16:
6
844-6856. (g) Holloczki O, Nyulaszi L. Organometallics. 2009; 28: 4159-4164.
Klebe JF. Acc Chem Res. 1970; 3, 9: 299-305.
Lane TH, Frye CL. J Org Chem 1978; 43: 4890-4891.
Szalay R, Pongor G, Harmat V, Bocskei Z, Knausz D. J Organomet Chem. 2005; 690: 1498-1506.
Pongor G, Kolos Z, Szalay R, Knausz D. J Mol Struct-THEOCHEM. 2005; 714: 87-97.
Hua DH, Miao SW, Bharathi SN, Katsuhira T, Bravo AA. J Org Chem. 1990; 55: 3682-3684.
7
8
9
1
1
0
1
1
2

1
1
2
3
Knausz D, Meszticzky A, Szakacs L, Csakvari B, Ujszaszy K. J Organomet Chem. 1983; 256: 11-21.
a) The Cambridge Structural Database (CSD), ver. 5.37. (2016 Feb.). (b) Allen FH, Acta Crystallogr, Sect B:
(
Struct Sci. 2002; 58: 380-388.
1
4
(a) Mitzel NW, Hofmann M, Waterstradt E, Schleyer PV, Schmidbaur H. Dalton Trans. 1994; 2503-2508. (b)
Mitzel NW, Angermaier K, Schmidbaur H. Organometallics. 1994; 13: 1762-1766. (c) Mitzel NW, Schmidbaur
H. Z Anorg Allg Chem. 1994; 620: 1087-1092.
1
5
Bassindale, A. R.; Glynn, S. J.; Taylor, P. G. In The Chemistry of Organic Silicon Compounds; Rappoport, Z.;
Apeloig, Y., Eds.; Wiley: New York, 1998, Vol. 2, pp 495–511.
1
3