Links between through-bond interactions and assembly structure in simple piperidones

Article abstract of DOI:10.1039/b808818g

3,5-Disubstituted piperidones provide an opportunity to explore donor-acceptor through-bond interactions in the context of molecular and supramolecular structure. The crystal structure of cis-3,5-dibenzyl-1- phenylpiperidin-4-one 3 is disordered and the lattice accommodates a ～3:1 ratio of the N-Ph equatorial (3-eq) and N-Ph axial (3-ax) epimers, based on refined values of occupancy factors. The fortuitous result allows a side-by-side comparison of the two configurations with respect to their donor-acceptor through-bond interactions. The energy difference between 3-ax and 3-eq (ΔE_ax-eq) has been evaluated in the gas phase using extensive first principles calculations, and for many levels of theory this difference parallels the experimentally-observed configurational ratio in the solid state (where the epimers share nearly identical packing environments). The calculations further show a difference in the through-bond stabilization for 3-ax and 3-eq, with larger contributions for 3-ax. Natural bond order (NBO) analysis quantifies the delocalization of the donor nitrogen lone pair into the adjacent carbon-carbon bonds and carbonyl acceptor for the 3-ax epimer. The work concludes that molecular-level structural perturbations that arise from or otherwise influence through-bond donor-acceptor interactions have consequences on solid-state and supramolecular assembly structure. The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008.

Full text of DOI:10.1039/b808818g

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/njc | New Journal of Chemistry
Links between through-bond interactions and assembly structure
in simple piperidonesw
Ling Yuan,a Bobby G. Sumpter,b Khalil A. Abbouda and Ronald K. Castellano*a
Received (in Gainesville, FL, USA) 28th May 2008, Accepted 12th June 2008
First published as an Advance Article on the web 16th July 2008
DOI: 10.1039/b808818g
3,5-Disubstituted piperidones provide an opportunity to explore donor–acceptor through-bond
interactions in the context of molecular and supramolecular structure. The crystal structure of
cis-3,5-dibenzyl-1-phenylpiperidin-4-one 3 is disordered and the lattice accommodates a B3 : 1
ratio of the N-Ph equatorial (3-eq) and N-Ph axial (3-ax) epimers, based on refined values of
occupancy factors. The fortuitous result allows a side-by-side comparison of the two
configurations with respect to their donor–acceptor through-bond interactions. The energy
difference between 3-ax and 3-eq (DEax–eq) has been evaluated in the gas phase using extensive
first principles calculations, and for many levels of theory this difference parallels the
experimentally-observed configurational ratio in the solid state (where the epimers share nearly
identical packing environments). The calculations further show a difference in the through-bond
stabilization for 3-ax and 3-eq, with larger contributions for 3-ax. Natural bond order (NBO)
analysis quantifies the delocalization of the donor nitrogen lone pair into the adjacent
carbon–carbon bonds and carbonyl acceptor for the 3-ax epimer. The work concludes that
molecular-level structural perturbations that arise from or otherwise influence through-bond
donor–acceptor interactions have consequences on solid-state and supramolecular assembly
structure.
The interplay between self-assembly and through-bond
Introduction
interactions emerging for the AATs encourages studies of
even simpler systems wherein s-coupled donor–acceptor inter-
actions might be linked to a supramolecular event or structure.
Considered here are piperidones 2 and 3 (Fig. 1), functional
fragments of 1a, for which the configuration at nitrogen that
optimizes donor–acceptor through-bond interactions is not
fixed. The equilibrium that describes interconversion between
the axial (ax) and equatorial (eq) forms of 2 is shown as an
The rigid covalent framework of b-aminoketones 1,
1-aza-adamantanetriones1 (AATs), fosters efficient through-
bond2 (hyperconjugative3) communication between the
bridgehead nitrogen donor and three carbonyl acceptors.
Consequences at the molecular level include changes in bond
lengths and angles within the donor–s-acceptor core,
emergence of a new absorption band in the UV/Vis spectrum
(the ‘‘s-coupled transition’’4), and dramatic changes in
chemical reactivity.1b Functionalized AATs have appeared as
vehicles to explore ‘‘strong’’ through-bond interactions
‘‘beyond-the-molecule’’5 for the first time in the context of
solution-phase self-assembly.6 Both 1a and 1b (Fig. 1), for
example, show extensive aggregation in organic solvents that
is manifested, at low concentrations (ca. 0.5% by weight for 1a
and 1b), in organogelation.6a,b The macromolecular properties
respond to the chemical substituents on the core in ways that
implicate its unique donor–acceptor configuration,6c and recent
theoretical studies find that an unprecedented electronic structure
emerges for 1-D arrangements of the molecules in the gas phase.7
a Department of Chemistry, University of Florida, P.O. Box 117200,
Gainesville, FL 32611-7200, USA. E-mail: castellano@chem.ufl.edu;
Fax: +01 352 846 0296; Tel: +01 352 392 2752
b Computer Science and Mathematics Division and Center for
Nanophase Materials Sciences, Oak Ridge National Laboratory,
Oak Ridge, TN 37831, USA
w Electronic supplementary information (ESI) available: Cartesian
coordinates and energies for all calculated structures. CCDC reference
numbers 686114 (2) and 686115 (3). For ESI and crystallographic data
in CIF or other electronic format see DOI: 10.1039/b808818g
Fig. 1 Substituted 1-aza-adamantanetriones 1 and their cis-3,5-di-
benzylpiperidin-4-one counterparts 2 and 3.
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008
1924 | New J. Chem., 2008, 32, 1924–1934

View Article Online
example. It is established from both experimental and
theoretical work that the axial configuration with respect to
the nitrogen substituent is required to optimize (‘‘turn on’’,
depicted graphically by a no-bond resonance structure)
donor–acceptor through-bond interactions in cyclic
b-aminoketones and related structures.4,8 Given that these
types of molecules (a) may hold potential for electronic
applications7,9 and (b) show relationships between their bulk
properties and unique electronic configurations, elucidating
the relationship between the ax : eq ratio and molecular/
supramolecular structure is important.
The cis-3,5-dibenzyl substituents of 2 and 3 (Fig. 1) allow
the molecules to be easily crystallized. The structure of 3 is
disordered, but can be suitably refined based on a two-site
occupancy that corresponds to the axial (3-ax) and equatorial
(3-eq) nitrogen configurations (B3 : 1 eq : ax). The result is
fortuitous in that it places the epimers in similar solid-state
environments and encourages a side-by-side analysis of their
energies and electronic structures; a line of investigation
typically precluded and with results that are obscured by
‘‘packing effects’’. Through high-level calculations that include
natural bond order analysis, to date sparsely performed on
similar donor–s-acceptor systems,10 a surprising relative sta-
bility for 3-ax is revealed, in part a consequence of hyper-
conjugative interactions. Several computational approaches
are explored, all that treat electron correlation well, and
several find gas-phase energy differences between 3-ax and
3-eq that on-average parallel the solid-state results. Further
revealed, by comparison of 2 to 3 and several other model
compounds, is that simple functionalization of piperidones
can substantially bias the ax : eq ratio. This combined experi-
mental and theoretical study shows that molecular-level struc-
tural changes linked to through-bond interactions can have
consequences on assembly and therefore, bulk properties.
Scheme 1 Synthesis of cis-3,5-dibenzylpiperidin-4-one derivatives.
used (i.e., 60 psi).14 Worth noting, no significant epimerization
at positions C(3) or C(5) has been detected in these molecules
after even prolonged storage at room temperature in the
laboratory.15
X-Ray crystal structure of 2: general features and packing
Fig. 2 shows the X-ray crystal structure of 2. Selected bond
lengths and angles are presented in Tables 1 and 2. The two
benzyl groups are cis diequatorial (Fig. 2(a)) and the piper-
idone ring adopts a standard chair conformation. The
Results and discussion
Synthesis of cis-3,5-dibenzylpiperidin-4-one derivatives
Synthesis of 2 and 3 follows classical condensation chemistry
(Scheme 1). Starting piperidones 4 were commercially avail-
able (4a) or readily obtained (4b11). Standard base-catalyzed
solution-phase conditions (NaOH, EtOH) were found to work
best to effect the aldol cross-condensation between N-methyl-
piperidone 4a and benzaldehyde to give dibenzylidene deriva-
tive 5a (in 54% isolated yield), echoing observations in the
literature.12 Other dibenzylidenes, including 5b, were only
prepared in comparable yield (B50%) using solvent-free
conditions and ytterbium triflate catalysis.13 Both 5a and 5b
were isolated in the expected E,E configuration.12b Hydroge-
nation of 5a and 5b in the presence of catalytic palladium on
carbon then provided dibenzylpiperidones 2 and 3, respec-
tively, with the cis relative stereochemistry of the benzyl side
chains (480% d.e.). Initial hydrogenation attempts using the
Adams platinum oxide catalyst as described for similar sys-
tems by McElvain12a unfortunately also resulted in facile
reduction of the carbonyl group to the corresponding second-
ary alcohol. Palladium on carbon worked quite efficiently as a
substitute catalyst provided that high hydrogen pressures were
Fig. 2 The X-ray crystal structure of 2: (a) An ORTEP diagram of 2.
Displacement ellipsoids are drawn at the 50% probability level; (b)
close-packed dimers with near perfect edge-to-face interactions and
dipolar core alignment (d1 = 3.76 A; d2 = 3.95 A). Atom colors: blue
N, red O, white H.
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008 New J. Chem., 2008, 32, 1924–1934 | 1925

View Article Online
Table 1 Selected bond lengths (A) for 2, 3-eq, and 3-axa
searching, with cyclic b-aminoketones. It provides a rare
opportunity here to compare the two epimers in similar solid
state environments (vide infra).
Bond
2
3-eq
3-axb
N(1)–C(2)
N(1)–C(6)
N(1)–C(7)
C(2)–C(3)
C(5)–C(6)
C(3)–C(4)
C(4)–C(5)
C(4)–O(1)
a
1.4611(13)
1.4604(13)
1.4617(13)
1.5359(15)
1.5312(14)
1.5100(14)
1.5096(15)
1.2178(12)
1.4755(18)
1.4524(17)
1.4340(19)
1.519(2)
1.545(2)
1.521(3)
1.457(7)
1.407(6)
1.462(6)
1.487(9)
1.556(8)
1.517(11)
1.504(11)
1.191(14)
The AAT systems 1 (Fig. 1) are unusual in that they show
pronounced bond length changes consistent with hyperconju-
gative interactions; for example, a slight shortening of the
N(1)–C(2) and N(1)–C(6) bonds (to 1.45 A), and a lengthening
of the central C(2)–C(3) and C(5)–C(6) bonds (to 1.59 A).1b
Fig. 3(a) and (b) show the two nitrogen epimers of 3 as they
are found in the crystal. Tables 2 and 3 then summarize the
key bond angles and distances, respectively. The stereoelec-
tronic requirements for optimized through-bond interactions
(Fig. 1) predict that the piperidone ring bond lengths of 2 and
3-eq might be both similar and ‘‘normal’’ when compared to
3-ax. Such an analysis is unfortunately precluded here by the
disorder of 3, particularly with respect to 3-ax, the minor of the
two epimers. Evaluation of subtle bond angle changes with
respect to through-bond interactions could be equivalently risky.
High-level calculations using the crystal structures as a starting
point provide a better way to probe these stereoelectronic
effects (vide infra).
1.510(3)
1.209(3)
Standard deviations are shown in parentheses. Additional crystal-
lographic details are provided in the Experimental section.
b
Corresponding atom labels are primed (see Fig. 3(b)).
Table 2 Selected bond angles (1) for 2, 3-eq, and 3-axa
Angle
2
3-eq
3-axb
C(2)–N(1)–C(6)
C(2)–N(1)–C(7)
C(6)–N(1)–C(7)
110.09(9)
111.35(8)
110.65(9)
332.1
—
—
110.51(9)
111.36(8)
107.70(8)
108.47(9)
114.14(9)
113.36(9)
113.84(9)
123.48(10)
122.62(10)
110.33(11)
111.40(11)
116.19(12)
337.9
111.6(4)
119.5(4)
116.4(4)
347.5
P
CNC
N(1)–C(7)–C(22)
123.81(15)
117.55(18)
112.57(13)
109.89(11)
109.03(15)
108.37(15)
113.02(15)
110.90(14)
115.62(17)
121.3(2)
116.2(4)
123.6(4)
115.4(5)
113.6(4)
112.4(6)
109.3(5)
111.2(5)
108.6(6)
114.2(7)
121.1(10)
123.9(10)
The two epimers of 3 do share similar packing environments
(Fig. 3(c) and (d)); the aromatic rings in each experience a
number of readily characterizable contacts. The N-Ph groups
show close edge-to-face interactions with neighboring benzyl
substituents, defined here by the C(24)-centroid distances
(d1 = 3.70 A; d3 = 4.13 A). The angle between the aromatic
planes is also around the ideal 901 (3-eq = 85.61; 3-ax =
84.31). Slipped (offset) face-to-face interactions are found that
can be defined by the ring centroid to ring centroid distance
(d2 = 3.99 A; d4 = 4.15 A). The distance between the ring
planes is 3.60 A for 3-eq and a short 3.24 A for 3-ax in this
configuration. The packing forces certainly influence the pli-
able twist angle about the N(1)–C(7) bond (calculated from
the angle between the rms planes defined by atoms
C(4)–N(1)–C(7) and atoms C(7)–C(23)–C(25)) that is 35.21
in 3-eq and 33.71 in 3-ax.
N(1)–C(7)–C(26)
C(3)–C(2)–N(1)
C(5)–C(6)–N(1)
C(4)–C(3)–C(2)
C(4)–C(5)–C(6)
C(4)–C(3)–C(15)
C(4)–C(5)–C(8)
C(3)–C(4)–C(5)
C(3)–C(4)–O(1)
C(5)–C(4)–O(1)
a
123.0(2)
Standard deviations are shown in parentheses. Additional crystal-
lographic details are provided in the experimental section.
b
Corresponding atom labels are primed (see Fig. 3(b)).
equatorial N-CH3 epimer is observed (2-eq), a configuration at
nitrogen that allows individual molecules to form closely-
packed dimers (Fig. 2(b)). Several key interactions highlight
this assembly: near perfect edge-to-face aromatic interactions
are found with d1 (distance between the ring carbon C(17) and
neighboring ring centroid) = 3.76 A and an angle between the
aromatic planes of 89.71; this geometry optimizes the electro-
static interactions between the protons on the ring edge and
the p-electron density on the ring face.16 Stabilizing dipolar
(electrostatic) interactions come through an antiparallel align-
ment of the cores; more subtle electrostatic matching between
the piperidone nitrogens and carbonyl carbons is also evident
(Fig. 2(b); d2 = 3.95 A). The extended crystal packing for 2
(not shown) features additional edge-to-face aromatic interac-
tions between the rings of the dimeric units.
Not necessarily decoupled from the packing and important
for consideration of through-bond interactions is the pyrami-
dalization at N(1), analyzed thoughtfully in a variety of related
systems.8,10b As a baseline, the nitrogen of 2 is, as expected,
strongly pyramidalized based on the sum of the C–N–C bond
= 3321). The nitrogen is slightly flattened in
P
angles (
3-eq (
CNC
P
= 3381) and more significantly flattened in 3-ax
CNC
P
(
= 3481). The latter presumably reflects conjugation
CNC
between the nitrogen and phenyl ring in the axial epimer that
also minimizes steric repulsion; however, the N(1)–C(7) twist
angles are nearly the same in 3-ax and 3-eq (vide supra).
Weaker and longer aromatic contacts (e.g. edge-to-face
involving the benzyl groups) then define the packing structures
(Fig. 3(e) and (f)). Unlike the structure of 2 in which the
monomers are oriented into antiparallel dimers, 3 forms
abutting one-dimensional stacks that feature the carbonyl
groups all pointing in the same direction with respect to the
column axis. The macrodipole created within one pair of
stacks is cancelled by a neighboring pair. The intracolumnar
nitrogen-to-nitrogen distance is 5.5 A, a distance nearly iden-
tical to that postulated for 1-D stacks of 1.6c,7 Also observed
are C–Hꢂ ꢂ ꢂO interactions18 between the piperidone carbonyl
X-Ray crystal structure of 3: general features and packing
The X-ray structure of N-phenylpiperidone 3 at 173 K is
disordered; both the equatorial (3-eq) and axial (3-ax) epimers
of the compound are accommodated in the crystal lattice, in a
B3 : 1 ratio (76 ꢁ 1% 3-eq; 24 ꢁ 1% 3-ax) based on refined
values of occupancy factors. Configurational disorder at
nitrogen has been identified and studied before in other
N-heterocycles in the solid state,10b,17 but not, based on our
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008
1926 | New J. Chem., 2008, 32, 1924–1934

View Article Online
Fig. 3 The X-ray crystal structure of 3 is disordered and refined based on a 76% occupancy of 3-eq (a) and a 24% occupancy of 3-ax (b). The local
packing environments of 3-eq (c) and 3-ax (d) are similar with respect to their aromatic contacts (d1 = 3.70 A; d2 = 3.99 A; d3 = 4.13 A; d4 =
4.15 A). The molecules are further organized into 1-D columns with a nitrogen-to-nitrogen intracolumnar distance of 5.5 A (e and f). Atom colors:
blue N, red O, white H. The van der Waals surfaces shown were generated in Discovery Studio Visualizer 2.0 (Accelrys Software, Inc.).
oxygens and both the benzylic CH2 hydrogens and the benzyl
aromatic C–H hydrogens (dCꢂ ꢂ ꢂO = 3.5 A).
Properties in solution
Conformation and relative stereochemistry. 1H NMR experi-
ments corroborate the solid-state conformations and relative
stereochemistry of 2 and 3 in solution. Provided in Table 3 are
The sensitivity of the 3-eq : 3-ax ratio to temperature and
crystal preparation has been investigated. A second X-ray
quality crystal was obtained from a different synthetic batch
of 3; the epimer ratio at 173 K (69 ꢁ 1% 3-eq; 31 ꢁ 1% 3-ax)
appeared similar to first sample where again the axial epimer
was the minor component, but well represented. The disorder
in this sample showed no dependence on temperature based on
data collection at 297 K (68 ꢁ 1% 3-eq; 32 ꢁ 1% 3-ax) and
123 K (69 ꢁ 1% 3-eq; 31 ꢁ 1% 3-ax). The conclusion is that
the epimers do not equilibrate in this temperature range in the
solid state—static disorder—although the energy barrier for
doing so in solution is small (vide infra).
the 1H NMR chemical shifts and JH,H (vicinal) coupling
3
3
3
constants for the compounds. From the Ja,c and Jb,c values
for 2 (in CDCl3) and 3 (in toluene-d8) it is clear that the benzyl
groups at positions C(3) and C(5) are cis diequatorial and that
the molecules adopt chair conformations. Taking 2 as an
3
3
example, the Jc,d (3.5 Hz) and Jb,c (8.8 Hz) coupling con-
stants support the solid-state benzyl substituent conformation
wherein the CH2Ph group is staggered with respect to the
piperidine ring and Hc. For 3 in CDCl3, Ha and Hb are
deshielded by the neighboring phenyl ring. The difference in
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008 New J. Chem., 2008, 32, 1924–1934 | 1927

View Article Online
Table 3 1H NMR data for compounds 2 and 3a
temperature to ꢃ55 1C; the nitrogen inversion barrier is simply
too low to be readily measured. The barriers are likely similar
to that of N-methylpiperidine (DGz B 6–9 kcal molꢃ1),21 even
lower for 3 given that the transition state for nitrogen inver-
sion is stabilized by conjugation with the phenyl ring. Addi-
tionally, the UV/Vis spectra of the molecules in solution
(CH3CN) show no evidence of the so-called ‘‘s-coupled
transition’’ (lmax 220–260 nm).4 For 2 this is presumably due
to the methyl group preferentially adopting the equatorial
position (vide infra);8a–d for 3, this relevant (and weak; e
B500–2500 Mꢃ1 cmꢃ1) absorption is otherwise masked by
the stronger absorption bands of the N-Ph chromophore.
2
3
3b
3.66
2.46
2.54
3.23
2.34
—
6.49
4.8
Ha (d, ppm)
Hb (d, ppm)
Hc (d, ppm)
Hd (d, ppm)
He (d, ppm)
N-CH3 (d, ppm)
ortho-CH (N-Ph) (d, ppm)
Ja,c/Hz
2.98
2.04
2.93
3.25
3.90
2.88
2.90
3.31
2.49
—
6.62
—
—
—
2.40
2.22c (1.74d, 2.49e)
—
5.1
11.1
3.5
8.8
Jb,c/Hz
Jc,d/Hz
Jc,e/Hz
12.1
4.5
8.4
—
a
All data recorded in CDCl3 at 21 1C unless otherwise specified and
chemical shifts are reported relative to TMS. In toluene-d8. The
b
c
N-CH3 protons of 1-methylpiperidine appear at d 2.24, 2.23, and
2.10 ppm in C6D6, CDCl3, and DMSO-d6, respectively. In C6D6.
Discussion and theoretical analysis
d
e
In DMSO-d6.
Configurational preferences at nitrogen in the solid state
Unlike the solid-state structures of the AATs (1), hallmark
(and more subtle) intramolecular structural changes as a
consequence of through-bond interactions are difficult to
assess in 3 due, in part, to crystallographic disorder. These
include perturbed bond lengths and angles, and pyramidaliza-
tion effects at nitrogen or the carbon of the carbonyl.1b,8f,h Still
intriguing is that 3-ax is highly represented in the solid state,
an appearance not obviously linked to a much differently
packed structure. Translating solid state conformer (or epi-
mer) ratios into meaningful energetic differences ranges from
risky to inappropriate; only in certain cases22 can these con-
nections be considered. In our case, the co-representation of
3-ax and 3-eq provides such an opportunity.
chemical shift between Ha and Hb speaks, in part, to the on
average positioning of the phenyl substituent with respect to
the symmetry plane of the piperidone ring; here the phenyl
group is likely twisted to some extent about the N-Ph bond
and Ha experiences a larger ring current effect.17e,19
The notion that through-bond donor–acceptor effects might
be significant enough to bias solution-phase and solid-state
conformations is credited to Verhoeven and co-workers who
have performed numerous seminal studies with 1,10-dicya-
noethylene-derivatized piperidones and tropinones.8f,i,j
Related findings have been reported by Jenneskens and
co-workers.8h On the other hand, Ogawa and co-workers have
shown that crystal packing effects can significantly influence
the nitrogen configuration of various N-phenylpiperidines;
there are certainly cases where the higher-energy epimers of
such compounds may selectively crystallize.10b The observa-
tion is a universal one.23 The disorder that defines 3-ax and
3-eq represents the first case for simple piperidones where the
packing argument is at least partially abated.
Solvent dependence of chemical shifts. Some donor–
s-acceptor molecules show substantial chemical shift
responses to solvent polarity, noted previously for 1a.6a The
N-CH3 group of compound 2 is a sensitive probe of this
behavior. Considerable downfield shifts (Dd = 0.75) are found
for 2 (Table 3) upon moving from C6D6 (e = 2.3) to DMSO-d6
(e = 47),20 consistent with stabilization of the core dipole
(polarization effects) and the nitrogen developing an increasing
partial positive charge, while the opposite trend is found for
simple 1-methylpiperidine (Dd = ꢃ0.14) that lacks an acceptor
group. Whether this behavior is coupled with variation of the
equilibrium 2-eq : 2-ax ratio is not known. For 3, the protons are
in general more deshielded in the more polar CDCl3 vs.
toluene-d8. Also interesting is that the chemical shift difference
between Ha and Hb increases slightly in 3 (vs. 2), possibly related
to a conformational change with respect to the N-Ph substituent.
To be complete, we have performed a Cambridge Structural
Database search of N-alkyl and N-arylpiperidines/-ones. The
query was restricted to include only organic structures with R
factors better than 0.05; also eliminated were piperidines/-ones
that (a) were not in chair conformations (e.g., those with fused
ring structures or vinylic C(3)/C(5) carbons) and (b) bore
Nitrogen configuration. Piperidones 2 and 3 show no sub-
stantial change in their 1H NMR spectra in CDCl3 from room
Table 4 Energy difference between the axial (ax) and equatorial (eq) epimers for molecules 2, 3, 6–11 at various levels of theorya
Level of theory
6
7
8
9
10
11
2
3
B3LYP/6-31G*
MP2/6-31G*
MP2/6-311++G**
a
3.4c
—
1.5c
—
–0.012
—
—
–0.997
—
—
–0.15
—
—
–1.99
—
—
–0.24
0.18
3.32b
3.37
2.36b
2.25
–2.16
–3.31
c
b
Reported as E(axial) ꢃ E(equatorial); a negative number means that the axial configuration is more stable in the gas phase. See ref. 10a. See
ref. 10b.
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008
1928 | New J. Chem., 2008, 32, 1924–1934

View Article Online
substituents at C(2) or C(6)). Of the 280 or so piperidine
candidates, there are only about a dozen N-C(sp2) derivatives
that are axially or pseudo-axially disposed (see, for example,
CSD codes CEQXUS, EZOPAK, HUFROP, LUPVUN and
VUCHIK). Most molecules of this type (e.g., N-acylpiperi-
dines) feature definitively sp2-hybridized N atoms, a conse-
quence of extensive p-conjugation between the nitrogen and its
substituent. No axially-configured N-alkylpiperidines were
identified. For the piperidones, only two N-alkyl (CSD codes:
DIKVUP and FEKNUE) and no N-aryl derivatives were
found; both of the former are in the equatorial configuration.
On the basis of CSD searching, the appearance of 3-ax is quite
uncommon.
Utilized here are levels of theory that treat electron correla-
tion well, a prerequisite for considering hyperconjugative
effects in such molecules (vide infra).10a These include MP2,
CCSD, CCSD(T), CASSCF and DFT. The results, in terms of
DEax–eq, report on electron correlation effects when compared
to those obtained from simple Hartree–Fock (HF) theory that
neglects electron–electron interactions. Even so, each non- or
post-HF method treats electron correlation (and for DFT,
electron exchange) differently; these subtleties are discussed in
the computational methods section (vide infra). Beginning with
the simplest piperidine case, 6, the MP2/6-311++G** level
predicts a 3.4 kcal molꢃ1 advantage for 6-eq vs. 6-ax, a value
similar to that reported by DFT (B3LYP/6-31G*)10b and
MP2/6-31G* calculations.10a Each of these values is close to
the experimentally reported one.21 This energy difference di-
minishes significantly, by 1.1 kcal molꢃ1, upon introduction of
the carbonyl acceptor (7) at the MP2/6-311G++** level.
Brouwer and co-workers report a similar 1.0 kcal molꢃ1 de-
crease at the MP2/6-31G* level.10a A comparable axial stabiliza-
tion is also observed for 8 vs. 9 (0.99 kcal molꢃ1) and 11 vs. 3
(1.3 kcal molꢃ1). The value is, however, reduced for 10 vs. 2
(o0.1 kcal molꢃ1), and this may speak to a difference in the
relative magnitude of stabilization between N-CH3 and
N-Ph in the 3,5-disubstituted derivatives. Finally, it is interesting
to note for 3 (and to a lesser extent 8) the quite different DEax–eq
values between the DFT and MP2 methods. Given the consis-
tency between the methods for 6, some of this disparity likely
arises from how they handle electron correlation (vide infra).
As first reported by Ogawa et al.,10b the phenyl substituent can,
despite its size, better approach an axial positioning than a methyl
group by maintaining some delocalization of the nitrogen lone
pair into the ring.23b At the DFT level, this amounts to a reduction
in the DEax–eq value by B1.9 kcal molꢃ1 for 8 vs. 6. Our MP2
calculations give this difference as a larger B3.4 kcal molꢃ1
such that Eax D Eeq at this level. Similar energetic differences
characterize 7 vs. 9 (3.2 kcal molꢃ1), 10 vs. 11 (1.8 kcal molꢃ1), and
2 vs. 3 (3.1 kcal molꢃ1). Therefore, conversion of the N-CH3
function to N-Ph results in a 1.8–3.4 kcal molꢃ1 swing in relative
stability toward ax.
Configurational preferences at nitrogen in solution
Solution-phase values for the energy difference between the
equatorial and axial configurations at nitrogen of N-methyl-
piperidine have been controversial, but most sources report a
DG1 B 3 kcal molꢃ1 in favor of the former.21 The equili-
brium constant is similar for N,cis-3,5-trimethylpiperidine
(B3 kcal molꢃ1), again in favor of the equatorial configura-
tion, that is substituted similarly to the molecules discussed in
this work.24 Experimental measurements with the corres-
ponding N-methylpiperidones, N-phenylpiperidines, and
N-phenylpiperidones are comparatively scarce, although some
aspects have been taken up computationally (vide infra). The
phenyl group of N-phenylpiperidine, for example, is known to
be preferentially equatorial in solution,15,25 although this has
not been rigorously quantified.
Theoretical studies
Little is known about the energetic differences between the ax
and eq configurations in simple piperidones, particularly for
N-Ph derivatives,10b,26 despite the fact that these molecules
lend themselves well to comprehensive high-level theory that
can evaluate energy differences between conformers and inter-
rogate their respective electronic structures.10a The molecules
considered, in addition to 3-ax and 3-eq, are shown in Fig. 4
(and the results are provided in Tables 4 and 5). The series will
first establish a baseline energy difference between eq and ax
epimers when N is substituted with methyl (6, 10) or phenyl
(8, 11), in the presence and absence of benzyl substituents at
positions 3 and 5 of the piperidine ring. The energy differences
(between ax and eq, DEax–eq) are then evaluated in the
presence and absence of the acceptor (CQO); that is, when
through-bond donor–acceptor interactions are ‘‘turned on’’ or
‘‘turned off.’’
A surprising finding emerges related to 3,5-dibenzyl substitu-
tion of the piperidine ring, where a further energetic tilt
toward ax is observed. The energy differences are: 6 vs. 10
(3.5 kcal molꢃ1),
8 7 vs. 2
vs. 11 (2.0 kcal molꢃ1),
(2.5 kcal molꢃ1), and 9 vs. 3 (2.3 kcal molꢃ1). Discussed earlier,
there is virtually no difference in the ax-eq ratio between
N-methylpiperidine and N,cis-3,5-trimethylpiperidine in solu-
tion. The result then implicates the p-systems of the phenyl
groups as potentially important stabilizing groups. The combi-
nation of all of the effects—the carbonyl acceptor, 3,5-dibenzyl
substitution, and N-Ph substitution—leads to a significant
predicted stability advantage for 3-ax at the MP2/
6-311++G** level (3.3 kcal molꢃ1). These structural modifica-
tions could be general ones to optimize through-bond interac-
tions in simple piperidones.
Various levels of theory treat the energy differences between
3-ax and 3-eq differently (Table 5), and a systematic study is
important for future computations that consider through-
bond interactions in b-aminoketones or related molecules.
The effects of electron correlation on DEax–eq is reinforced
Fig. 4 Structures of molecules considered for computational studies
(in addition to 3-ax and 3-eq).
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008 New J. Chem., 2008, 32, 1924–1934 | 1929

View Article Online
Table 5 Energy difference between the axial (ax) and equatorial (eq) epimers (with respect to nitrogen configuration) for 3-ax and 3-eq at various
levels of theorya
Basis
HF
MP2
CCSD
CCSD(T)
CASSCF (10,10)
LDA
PBE0
B3LYP
STO-3G
6-31G*
6-31G**
6-311G**
6-311++G**
cc-pvdz
cc-pvtz
—
—
1.1
—
—
—
—
–0.36
—
0.6
—
—
—
—
–1.3
—
—
3.0
—
—
—
2.34
2.26
1.91
2.42
2.31
2.27
–2.16
–2.52
–2.45
–3.31
–2.74
—
–3.3
–3.4
–3.2
—
–3.1
—
–0.57
–0.90
–0.08
—
–0.62
—
0.18
0.10
0.36
—
0.11
0.31
3.36, 2.3b
—
—
—
—
a
Reported as E(axial) ꢃ E(equatorial) in kcal molꢃ1; a negative number means that the axial configuration is more stable in the gas
b
phase. CASSCF using 12 orbitals (6 occupied and 6 unoccupied molecular orbitals) and 12 electrons.
by comparing results obtained from simple Hartree–Fock
(HF) theory to those obtained with methods that treat corre-
lation (such as MP2, CCSD, CI or DFT). HF predicts that
3-eq is significantly more stable (by 1.9–2.4 kcal molꢃ1) over a
range of basis sets. Adding in the effects of dynamical electron
correlation by implementing many body perturbation theory
(MBPT) significantly predicts 3-ax to be the most stable (e.g.,
DE(MP2/cc-pvdz) = ꢃ2.74 kcal molꢃ1). Similarly, DFT
methods (LDA and PBE0) predict 3-ax stability, although in
some cases with smaller energy differences (e.g., DE(DFT-
PBE0/cc-pvdz) = ꢃ0.62 kcal molꢃ1; DE(DFT-LDA/cc-pvdz) =
ꢃ3.1 kcal molꢃ1). There is a significant difference between the
PBE0 functional and simple LDA, implying that treatment of
exact electronic exchange is important and that there is likely a
large correction for self-interaction. Also, the B3LYP func-
tional, another hybrid functional that includes a small amount
of exact exchange, actually finds 3-eq to be more stable (e.g.,
DE(DFT-B3LYP/cc-pvdz) = 0.11 kcal molꢃ1), although mod-
estly so. Correction for the self-interaction error, implicit in
DFT calculations and particularly large for strongly corre-
lated systems, gives 3-eq as the more stable isomer (e.g.,
from coupled cluster, PBE0, and B3LYP calculations are, on
average, coincidentally consistent with the solid-state epimer
ratio of B2–3 : 1 eq : ax (B0.4–0.65 kcal molꢃ1 preference for
3-eq). In the absence of a solution-phase experimental value for
the energy difference between 3-ax and 3-eq, it is difficult to
determine which method performs best for these systems,
independent of basis set. Apparent is that the use of methods
that account for strong electron correlation is important,10a and
each, in general, predicts a significant stability for 3-ax.
The energy-minimized structures of 3-ax and 3-eq do bear
resemblance to those found in the solid state, although this
need not be the case. An overlay between the calculated (dark
gray), at the MP2/cc-pvdz level, and experimentally-
determined (light gray) 3-eq structures is shown in Fig. 5(a).
The twist angle about the N–C(phenyl) bond is slightly
diminished in the calculated structure, by B101 (to 25.81),
P
and the nitrogen atom is also slightly flattened by 21 (
=
CNC
3401). The excellent agreement nonetheless suggests that 3-eq
is able to nicely optimize intermolecular and intramolecular
interactions in the solid state. A similar analysis with the
calculated and experimental structures of 3-ax is inherently
DE(SICDFT-PBE0/6-311G**)
=
4.5
kcal
molꢃ1
;
less reliable (due to disorder). While the nitrogen is modestly
P
DE(SICDFT-B3LYP/6-311G**) = 6.1 kcal molꢃ1). Clearly
these systems require a very high level of theory to account for
the strong electron correlation effects that are fundamental to
the relative stabilities of the two structures.
less flattened in the calculated structure (by 81; CNC = 3401),
the N–C(phenyl) twist angle is now B01 (optimizing interac-
tion between the N lone pair and phenyl ring; this is also
shown by NBO calculations, vide infra). Along the same lines,
it is important to note that subtle differences in the N-Ph ax
We also explored full configuration interactions on a re-
stricted active space (CASSCF) and coupled cluster theory that
included up to perturbative triples excitations. For the CASSC-
F(orbitals, electrons) calculations we used an active space of 10
orbitals (5 OMOs and 5 UMOs) and 10 electrons. This
relatively small active space, 63504 determinants, is nonetheless
adequate to include both the donor (N lone pair) and acceptor
(carbonyl antibonding orbital) portions of the piperidones. The
CASSCF(10,10) results find 3-eq as the most stable epimer (e.g.,
DE(CASSCF(10,10)/6-31G*) = 3.0 kcal molꢃ1). A slightly
larger active space (12 orbitals and 12 electrons) lowers the
configuration do exist among the calculated structures re-
ported in Tables 4 and 5. Even so, the
P
values are
consistently 340–3431 and the N–C(phenyl) twist angles lie
between 0 and 261.
CNC
In order to rationalize, in part, the origin of the 3-ax
stability, we performed a natural bond order (NBO) analysis
based on the optimized wavefunctions (MP2/cc-pvdz) of the
energy difference to 2.3 kcal molꢃ1
.
Finally, there is a substantial difference in the energies when
coupled cluster calculations are used (e.g., DE(CCSD/STO-3G) =
1.1 kcal molꢃ1; DE(CCSD(T)/STO-3G) = 0.6 kcal molꢃ1).
Larger basis sets (such as the 3-21G or cc-pvdz) lead to a slight
preference for 3-ax. The largest basis set that yielded converged
results at the CCSD and CCSD(T) levels was cc-pvdz, which
gave DE(CCSD/cc-pvdz) = ꢃ0.36 kcal molꢃ1 and DE(CCSD/
cc-pvdz) = ꢃ1.3 kcal molꢃ1. The energy differences obtained
Fig. 5 Superimposed calculated (dark gray; at the MP2/cc-pvdz level)
and experimentally-observed (light gray; X-ray) structures for 3-eq (a)
and 3-ax (b).
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008
1930 | New J. Chem., 2008, 32, 1924–1934

View Article Online
Two important conclusions emerge from the studies. First,
the axial/equatorial ratio with respect to the nitrogen substi-
tuent in simple piperidones is easily perturbed by the nature of
the nitrogen (e.g. methyl vs. phenyl) and ring substituents. It
follows that it is particularly inappropriate to employ classic
cyclohexane ring substituent parameters to rationalize the
piperidine/-one results where ‘‘sterics’’ are outweighed by
electronic and stereoelectronic effects. Second, it is reasonable
to assume that molecular-level structural perturbations that
arise from or otherwise influence through-bond donor–
acceptor interactions should have a consequence on solid-state
assembly structure. Future studies will continue to explore
these interactions in the solution phase where they might
influence the reversible assembly of molecules.
Fig. 6 Plot of the HOMO for 3-ax (a) and 3-eq (b) (energy-minimized
structures at the MP2/cc-pvdz level). Delocalization of the nitrogen
lone pair into the phenyl ring, adjacent C–C bonds of the piperidone
ring, and the carbonyl is observed for 3-ax.
two epimers. This technique is among the best to probe
hyperconjugative interactions in molecular systems.8h,23b,27
The results show, for 3-ax, that the largest second-order
stabilization energy (a measure of the degree of hyperconjuga-
tion) is due to the delocalization of the N lone pair (donor)
orbital into one of the adjacent aromatic CQC p* orbitals
(33.52 kcal molꢃ1). There is also another relatively large
stabilization, 8.61 kcal molꢃ1, due to the N lone pair orbital
delocalized into the two s* C–C orbitals within the piperidone
ring. Additionally, and significantly, one of the principle
orbital delocalizations for the lone pair is into the p* orbital
of the carbonyl group. A plot of the HOMO of the axial epimer
shows these delocalizations (Fig. 6(a)). For 3-eq, one main
source of stabilization is also from donation of the N lone pair
into adjacent aromatic CQC p* orbitals (23.23 kcal molꢃ1);
stabilization from N donation into the two p* C–C orbitals is
a comparatively lower 2.34 kcal molꢃ1. Most importantly,
there is no notable delocalization of the N lone pair into the p*
orbital of the carbonyl group of 3-eq (Fig. 6(b)). The bond
lengths within the calculated structures parallel these findings,
where the C(2)–C(3) and C(5)–C(6) bonds in 3-ax (1.547 A)
are somewhat longer than those in 3-eq (1.536 A). The trend is
in fact detectable throughout the calculated structures when
the carbonyl acceptor group is present (i.e., 2, 7 and 9), but
falls off in its absence (i.e., 6, 8, 10 and 11).
Experimental
Materials and general methods
Reagents and solvents were purchased from commercial
sources and used without further purification unless otherwise
specified. THF, ether, CH2Cl2, and DMF were degassed in
20 L drums and passed through two sequential purification
columns (activated alumina; molecular sieves for DMF) under
a positive argon atmosphere. Thin layer chromatography
(TLC) was performed on SiO2-60 F254 aluminum plates with
visualization by UV light or staining. Flash column chroma-
tography was performed using Purasil SiO2-60, 230–400 mesh
from Whatman. Melting points (mp) were determined on a
Mel-temp electrothermal melting point apparatus and are
uncorrected. 1H (13C) NMR spectra were recorded at 300
(75) MHz on Varian Mercury 300, Gemini 300, or VXR 300
spectrometers. Chemical shifts (d) are given in parts per
million (ppm) relative to TMS and referenced to residual
protonated solvent (CDCl3: dH 7.27 ppm, dC 77.00 ppm;
toluene-d8: dH 2.09 ppm). Abbreviations used may include s
(singlet), d (doublet), t (triplet), q (quartet), quin (quintet), hp
(heptet), br (broad) and m (multiplet). UV/Vis absorption
spectra were obtained using a Cary 100 Bio spectrophotometer
and 1 cm quartz cells. ESI-MS spectra were recorded on a
Bruker APEX II FTICR spectrometer. EI-MS spectra were
recorded on a Thermo Trace GC DSQ (single quadrupole)
spectrometer.
Conclusions
Simple 3,5-disubstituted piperidones have been used to explore
donor–acceptor through-bond interactions in the context of
molecular and solid-state (supramolecular) assembly structure.
The crystal structure of cis-3,5-dibenzyl-1-phenylpiperidin-4-
one 3 is disordered and the lattice reflects a B 3 : 1 mixture
of the N-Ph equatorial (3-eq) and N-Ph axial (3-ax) epimers,
based on refined values of occupancy factors. The fortuitous
result has inspired a side-by-side comparison of the two con-
figurations with respect to their donor–acceptor through-bond
interactions. The energy difference between 3-ax and 3-eq has
been evaluated in the gas phase using extensive first principles
calculations, and for many levels of theory this difference
parallels the experimentally-observed configurational ratio in
the solid state (where the epimers share nearly identical packing
environments). The calculations further show a difference in the
relative through-bond stabilization for 3-ax and 3-eq, with
larger contributions for 3-ax.8h Natural bond order (NBO)
analysis quantifies the participation of the piperidone nitrogen
and carbonyl groups in stabilization of the 3-ax epimer.
Syntheses
(3E,5E)-3,5-Dibenzylidene-1-phenylpiperidin-4-one
(5b).
Benzaldehyde (3.6 mL, 36 mmol) was added to 1-phenyl-
piperidin-4-one 4b11 (3.09 g, 17.6 mmol) under argon protec-
tion. Ytterbium triflate (55 mg, 0.088 mmol) was added and
the mixture was then heated to 90 1C for 20 h. The mixture was
purified by silica gel column chromatography (hexane–EtOAc,
6 : 1) to yield 5b (2.7 g, 45%) as a yellow solid (Found: C,
85.09; H, 5.99; N, 3.92. C25H21NO requires C, 85.44; H, 6.02;
N, 3.99%); mp 160–162 1C; dH (300 MHz; CDCl3) 4.63
(s, 4H), 6.68 (d, J = 7.8 Hz, 2H), 6.79 (t, J = 7.2 Hz, 1H),
7.13 (t, J = 7.2 Hz, 2H), 7.45 (m, 10H), 7.90 (s, 2H); dC
(75 MHz, CDCl3) 51.4, 116.7, 120.2, 128.7, 128.9, 129.2, 130.3,
133.0, 135.0, 137.6, 148.8, 187.2; m/z (EI) 351.1610
(M+; C25H21NO requires 351.1623).
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008 New J. Chem., 2008, 32, 1924–1934 | 1931

View Article Online
cis-3,5-Dibenzyl-1-methylpiperidin-4-one (2). (3E,5E)-3,5-
Dibenzylidene-1-methylpiperidin-4-one 5a12a (3.0 g, 10 mmol)
in ethanol (100 mL) was hydrogenated (60 psi H2, 240 mg of
10% palladium on carbon). After being stirred for 36 h at
room temperature, the catalyst was filtered off, the solvent was
evaporated, and the mixture was purified by column chroma-
tography (hexane–EtOAc, 4 : 1). The desired cis diastereomer
eluted from the column first (1.7 g, 58%) followed by the trans
diastereomer (0.16 g, 5%); both were isolated as pale yellow
solids. The calculated d.e. of cis-2 is 83% based on isolated
mass. Single crystals of cis-2 suitable for X-ray analysis were
obtained by recrystallization from ethyl ether. For cis-2
(Found: C, 82.12; H, 4.81; N, 8.12. C20H23NO requires C,
81.87; H, 4.77; N, 7.90%); mp 68–69 1C; dH (300 MHz;
CDCl3) 2.05 (t, J = 8.7 Hz, 2H), 2.24 (s, 3H), 2.40 (dd, J =
11.1, 14.4 Hz, 2H), 2.97 (m, 4H), 3.25 (dd, J = 4.1, 14.1 Hz,
2H), 7.22 (m, 10H); dC (75 MHz, CDCl3) 33.2, 45.4, 51.3, 62.3,
126.5, 128.7, 128.8, 129.3, 140.0, 147.5, 210.4; m/z (ESI)
294.1858 ([M + H]+; C20H24NO requires 294.1852). For
trans-2; dH (300 MHz; CDCl3) 2.25 (s, 3H), 2.41 (dd, J =
6.3, 11.4 Hz, 2H), 2.62 (dd, J = 4.8, 11.4 Hz, 2H), 2.75 (t, J =
13.2, 2H), 2.86 (m, 2H), 3.15 (dd, J = 4.5, 13.2 Hz, 2H), 7.24
(m, 10H); dC (75 MHz, CDCl3) 22.3, 35.3, 45.9, 50.7, 60.3,
126.6, 128.7 129.4, 139.6, 212.3; m/z (ESI) 294.1829
([M + H]+; C20H24NO requires 294.1852).
atoms. Details are provided in Table 6. For 3, the central six-
membered ring and the phenyl ring bonded to N(1) are
disordered whereas the rest of the molecule falls in the
same positions. The disorder was refined in two parts
with the site occupancy factors dependently refined. The major
part [occupancy factor = 0.76(1)] corresponds to 3-eq;
the minor part [occupancy factor = 0.24(1)] corresponds
to 3-ax.
CCDC reference numbers 686114 (2) and 686115 (3).
For crystallographic data in CIF or other electronic format
see DOI: 10.1039/b808818g
Computational methods. The fundamental properties and
equilibrium structures of the axial and equatorial epimers of
the N-substituted piperidines/-ones were examined using ex-
tensive first principles calculations based on quantum many-
body theory (via many body perturbation theory (MBPT),
coupled clusters (CC), and configuration interactions (CI) on a
defined active space) and density functional theory (DFT).
All-electron calculations were performed at each level of
theory using the NWChem package.29 For the DFT calcula-
tions, both the local density (LDA) and generalized gradient
approximations (GGA) were employed. However, in general,
the available functionals that can be used in DFT calculations
of materials properties suffer from the self-interaction of the
electrons that stems from incomplete cancellation between
self-interaction in the Coulomb interaction and in the ex-
change interaction. A framework for the self-consistent calcu-
lation of self-interaction corrections (SIC) to the density
functional theory (DFT) based on the original method of
Perdew and Zunger30 can be implemented based on a number
of variants. One approach, a perturbation treatment, uses the
Kohn–Sham orbitals determined from the self-consistent cal-
culations and performs orbital localization with the Foster-
Boys algorithm.31 The self-interaction energy is added to the
total energy. The advantage of this approach is that all
exchange–correlation functionals implemented in NWChem
can be used with this option. We employ this approach in the
present study but have also considered the more accurate
treatment of SIC using the optimized effective potential
(OEP). For the many body quantum calculations, these in-
cluded at the simplest level Møller–Plesset second order
perturbation theory (MP2).32 MP2 can typically account
(depending on the basis set) for up to 90% of the electron
correlation energy and is among the most efficient methods for
including such effects. In order to investigate higher order and
non-dynamical correlation effects we used configuration inter-
action calculations on a restricted space (CASSCF).33 Addi-
tionally, higher level calculations based on coupled cluster
calculations that included singles, doubles and perturbative
triples, CCSD(T), were carried out.
cis-3,5-Dibenzyl-1-phenylpiperidin-4-one (3). (3E,5E)-3,5-Di-
benzylidene-1-phenylpiperidin-4-one 5b (1.8 g, 5.2 mmol) was
hydrogenated as described for 2 (60 psi H2, 150 mg 10%
palladium on carbon) to afford 3 (0.75 g, 41%) as a pale yellow
solid and a mixture of cis and trans diastereomers (80% d.e. by
1H NMR integration) that could not be separated by column
chromatography. Single crystals of cis-3 suitable for X-ray
analysis were obtained by recrystallization from diethyl ether
(Found: C, 84.19; H, 7.36; N, 3.88. C25H25NO requires C,
84.47; H, 7.09; N, 3.94%); mp 80–81 1C; dH (300 MHz;
toluene-d8) 2.34 (dd, J = 8.4, 14.1 Hz, 2H), 2.46 (dd, J =
11.4, 12.1 Hz, 2H), 2.54 (dddd, J = 4.5, 4.8, 8.4, 12.1 Hz, 2H),
3.23 (dd, J = 4.5, 14.1 Hz, 2H), 3.66 (dd, J = 4.8, 11.4 Hz,
2H), 6.49 (m, 2H), 6.67 (m, 1H), 6.98 (m, 2H); dC (75 MHz,
CDCl3) 32.8, 50.8, 55.3, 115.5, 119.9, 126.6, 128.8, 129.1,
129.2, 129.6, 139.7, 148.6, 209.9; m/z (EI) 355.1938 (M+;
C25H25NO requires 355.1936).
X-Ray crystal structure determination and refinement. Data
were collected at 173 K on a Siemens SMART PLATFORM
equipped with a CCD area detector and a graphite mono-
chromator utilizing Mo-Ka radiation (l = 0.71073 A). Cell
parameters were refined using up to 8192 reflections. A full
sphere of data (1850 frames) was collected using the o-scan
method (0.31 frame width). The first 50 frames were re-
measured at the end of data collection to monitor instrument
and crystal stability (maximum correction on I was o1%).
Absorption corrections by integration were applied based on
measured indexed crystal faces. The structure was solved by
the Direct Methods in SHELXTL6,28 and refined using full-
matrix least squares on F2. The non-H atoms were treated
anisotropically, whereas the hydrogen atoms were calculated
in ideal positions and were riding on their respective carbon
For all calculations, we have used atom centered, contracted
Gaussian basis sets, ranging from the Pople split-valence basis
sets (i.e., 6-31G*, 6-311G**, etc.)34 to the Dunning correlation
consistent basis sets (cc-pVXZ and aug-cc-pVXZ)35 during the
calculation of the self-consistent solution. In all cases, full
geometry optimization was performed for each epimer. Tight
convergence criteria were used for the MP2 calculations:
Convergence for the SCF was 10ꢃ8, the AO and MO integrals
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008
1932 | New J. Chem., 2008, 32, 1924–1934

View Article Online
Table 6 Crystallographic data and processing parameters for cis-3,5-dibenzylpiperidin-4-one derivatives 2 and 3. For the methods used in the
structure determination, see the experimental sectiona
2
3
Formula
Mr
T/K
l/A
Crystal system
Space group
a/A
b/A
c/A
C
293.39
173(2)
0.71073
Monoclinic
P21/n
9.5059(7)
12.3970(9)
20H23NO
C25H25NO
355.46
173(2)
0.71073
Monoclinic
P21/n
15.1125(12)
5.5285(4)
14.7096(11)
102.928(2)
1689.5(2)
23.3378(18)
91.899(2)
1948.8(3)
b/1
V/A3
Z
4
1.153
0.070
632
4
1.212
0.073
760
Dc/g cmꢃ3
m/mmꢃ1
F(000)
Crystal size/mm
0.25 ꢄ 0.23 ꢄ 0.11
0.30 ꢄ 0.21 ꢄ 0.18
y range/1
2.17–27.50
10 768
1.58–27.50
12 085
Reflections collected
Independent reflections
Observed reflections (I 4 2s(I))
Parameters
3786 (Rint = 0.0427)
2040
4385 (Rint = 0.0315)
2662
199
0.786
285
0.919
Goodness-of-fit on F2
Final R indices
Dr (max., min.)/e Aꢃ3
R1 = 0.0331,a wR2 = 0.0661b
0.115, –0.143
R1 = 0.0379,a wR2 = 0.0890b
0.115, –0.157
P
P
P
P
a
b
2
R1 = (||Fo| ꢃ |Fc||)/ |Fo|. wR2 = [ [w(Fo ꢃ Fc2)2]/ [w(Fo2)2]]1/2; w = 1/[s2(Fo2) + (mp)2 + np]; p = [max(Fo2,0) + 2Fc2]/3, m and n are
constants.
10ꢃ11, and the CPHF 10ꢃ5; energy convergence for the
geometry optimization was 10ꢃ7 with a maximum gradient
of 10ꢃ5. For compounds 2 and 10 of Table 4, this level of
convergence was not reached on the geometry optimization.
For 2-ax, the energy convergence was to 6.8 ꢄ 10ꢃ6 (max grad =
2.7 ꢄ 10ꢃ4); for 2-eq, the energy convergence was to
1.8 ꢄ 10ꢃ5 (max grad. = 1.7 ꢄ 10ꢃ4). Likewise, for compound
10: 10-ax energy convergence to 3.8 ꢄ 10ꢃ5 (max grad. =
1.1 ꢄ 10ꢃ3); 10-eq energy convergence to 1.7 ꢄ 10ꢃ6 (max grad =
7.1 ꢄ 10ꢃ4). For the DFT, CASSCF, and coupled cluster
calculations, the NWChem default tolerances were used.
The electronic structure of 3 was probed by the natural
bond orbital (NBO) method of Weinhold and co-workers.36
This method uses the one electron density matrix to define the
shape of the atomic orbitals in the molecular environment, and
to derive molecular bonds from the density between atoms. In
short, the optimized wavefunction at a given level of theory is
localized to obtain canonical molecular orbitals that are
transformed into an orthonormal set of localized NBOs. In
this representation, formal core orbitals, s and p orbitals
(both bonding and antibonding), lone pairs and Rydberg
orbitals are explicitly given and it also permits the assignment
of hybridization both to the atomic lone pairs and to each
atom’s contributions to its bond orbitals. Since the NBOs do
not diagonalize the Fock or Kohn–Sham operator, off-diag-
onal elements will be non-zero and second-order perturbation
theory has shown that these off-diagonal elements give an
energy that is directly related to the stabilization energy
due to that interaction. In the present study, the stabilization
energy resulting from the interactions of the electron lone pair
on the nitrogen with other orbitals on the molecule was
evaluated and the influence of through-bond interactions
quantified.
Acknowledgements
This work was financially supported by the National Science
Foundation CAREER program (CHE-0548003) and the
University of Florida. B. G. S. thanks the Center for Nano-
phase Materials Sciences (CNMS), sponsored by the Division
of Scientific User Facilities, US Department of Energy. The
extensive computations were performed using the resources of
the National Center for Computational Sciences at Oak Ridge
National Laboratories (ORNL). R. K. C. is grateful to the
CNMS User Program (CNMS2004-016 and CNMS2007-029)
for resources. K. A. A. thanks the National Science Founda-
tion and University of Florida for funding the X-ray crystallo-
graphy equipment. We are additionally grateful to Juergen
Koeller for his assistance with processing the X-ray data, Dr
Ion Ghiviriga for help obtaining the 1H NMR spectra of 2 and
3, and Dr Edo Apra for useful discussions and help with the
CCSD(T) calculations.
References
1 (a) N. Risch, J. Chem. Soc., Chem. Commun., 1983, 532–533; (b) A.
J. Lampkins, Y. Li, A. Al Abbas, K. A. Abboud, I. Ghiviriga and
R. K. Castellano, Chem. Eur. J., 2008, 14, 1452–1463.
2 (a) R. Hoffmann, A. Imamura and W. J. Hehre, J. Am. Chem. Soc.,
1968, 90, 1499–1509; (b) R. Hoffmann, Acc. Chem. Res., 1971, 4,
1–9; (c) R. Gleiter, W. D. Stohrer and R. Hoffmann, Helv. Chim.
Acta, 1972, 55, 893–906; (d) R. Gleiter, Angew. Chem., Int. Ed.
Engl., 1974, 13, 696–701; (e) M. N. Paddon-Row, Acc. Chem. Res.,
1982, 15, 245–251; (f) R. Gleiter and W. Schafer, Acc. Chem. Res.,
1990, 23, 369–375.
3 (a) J. B. Lambert, Y. Zhao, R. W. Emblidge, L. A. Salvador, X.
Liu, J.-H. So and E. C. Chelius, Acc. Chem. Res., 1999, 32,
183–190; (b) I. V. Alabugin and T. A. Zeidan, J. Am. Chem.
Soc., 2002, 124, 3175–3185; (c) I. V. Alabugin and M. Manoharan,
J. Org. Chem., 2004, 69, 9011–9024.
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008 New J. Chem., 2008, 32, 1924–1934 | 1933

View Article Online
4 (a) R. C. Cookson, J. Henstock and J. Hudec, J. Am. Chem. Soc.,
1966, 88, 1060–1062; (b) R. C. Cookson, Proc. R. Soc. London Ser.
A, 1967, 297, 27–39.
5 J.-M. Lehn, Supramolecular Chemistry, VCH, Weinheim, 1995.
6 (a) H. Li, E. A. Homan, A. J. Lampkins, I. Ghiviriga and R. K.
Castellano, Org. Lett., 2005, 7, 443–446; (b) A. J. Lampkins, O.
Abdul-Rahim, H. Li and R. K. Castellano, Org. Lett., 2005, 7,
4471–4474; (c) B. G. Sumpter, V. Meunier, A. Vazquez-Mayagoitia
and R. K. Castellano, Int. J. Quantum Chem., 2007, 107,
2233–2242.
Sect. C, 1990, 46, 410–414; (d) A. Bouchemma, P. H. McCabe and
G. A. Sim, Acta Crystallogr., Sect. C, 1990, 46, 671–674; (e) A.
Bouchemma, P. H. McCabe and G. A. Sim, Acta Crystallogr.,
Sect. C, 1991, 47, 1219–1222; (f) D. Adam, P. H. McCabe, G. A.
Sim and A. Bouchemma, Acta Crystallogr., Sect. C, 1993, 49,
837–841; (g) D. Adam, P. H. McCabe, G. A. Sim and A.
Bouchemma, Acta Crystallogr., Sect. C, 1995, 51, 246–249; (h)
R. Glaser, I. Adin, D. Shiftan, Q. Shi, H. M. Deutsch, C. George,
K.-M. Wu and M. Froimowitz, J. Org. Chem., 1998, 63,
1785–1794; (i) R. Glaser, A. Novoselsky, D. Shiftan and M.
Drouin, J. Org. Chem., 2000, 65, 6345–6353.
7 B. G. Sumpter, V. Meunier, E. F. Valeev, A. J. Lampkins, H. Li
and R. K. Castellano, J. Phys. Chem. C, 2007, 111, 18912–18916.
8 (a) A. W. J. Dekkers, J. W. Verhoeven and W. N. Speckamp,
Tetrahedron, 1973, 29, 1691–1696; (b) P. Pasman, J. W. Verhoeven
and T. J. de Boer, Tetrahedron, 1976, 32, 2827–2830; (c) A. M.
Halpern and A. L. Lyons, J. Am. Chem. Soc., 1976, 98, 3242–3247;
(d) G. Spanka and P. Rademacher, J. Org. Chem., 1986, 51,
592–596; (e) B. Krijnen, H. B. Beverloo and J. W. Verhoeven,
Recl. Trav. Chim. Pays-Bas., 1987, 106, 135–136; (f) B. Krijnen, H.
B. Beverloo, J. W. Verhoeven, C. A. Reiss, K. Goubitz and D.
Heijdenrijk, J. Am. Chem. Soc., 1989, 111, 4433–4440; (g) T.
Scherer, W. Hielkema, B. Krijnen, R. M. Hermant, C. Eijckelhoff,
F. Kerkhof, A. K. F. Ng, R. Verleg, E. B. van der Tol, A. M.
Brouwer and J. W. Verhoeven, Recl. Trav. Chim. Pays-Bas, 1993,
112, 535–548; (h) W. D. Oosterbaan, R. W. A. Havenith, C. A. van
Walree, L. W. Jenneskens, R. Gleiter, H. Kooijman and A. L.
Spek, J. Chem. Soc., Perkin Trans. 2, 2001, 1066–1074; (i) D. J. A.
De Ridder, K. Goubitz, H. Schenk, B. Krijnen and J. W.
Verhoeven, Helv. Chim. Acta, 2003, 86, 799–811; (j) D. J. A. De
Ridder, K. Goubitz, H. Schenk, B. Krijnen and J. W. Verhoeven,
Helv. Chim. Acta, 2003, 86, 812–826.
9 (a) M. Rubiralta, E. Giralt and A. Dıez, Piperidine. Structure,
Preparation, Reactivity and Synthetic Applications of Piperidine and
its Derivatives, Elsevier, Amsterdam, 1991; (b) W. Schuddeboom,
B. Krijnen, J. W. Verhoeven, E. G. J. Staring, G. Rikken and H.
Oevering, Chem. Phys. Lett., 1991, 179, 73–78; (c) K. Srinivas, S.
Sitha, V. J. Rao, K. Bhanuprakash, K. Ravikumar, S. P. Anthony
and T. P. Radhakrishnan, J. Mater. Chem., 2005, 15, 965–973.
10 (a) A. M. Brouwer and B. Krijnen, J. Org. Chem., 1995, 60, 32–40;
(b) K. Ogawa, J. Harada, M. Endo, Y. Takeuchi and H. Kagawa,
Tetrahedron Lett., 1997, 38, 5663–5666.
18 R. K. Castellano, Curr. Org. Chem., 2004, 8, 845–865.
19 M. D. Rozeboom, K. N. Houk, S. Searles and S. E. Seyedrezai, J.
Am. Chem. Soc., 1982, 104, 3448–3453.
20 All of the proton chemical shifts of the piperidone core respond to
solvent polarity; only Hd and He remain essentially constant.
21 (a) P. J. Crowley, M. J. T. Robinson and M. G. Ward,
Tetrahedron, 1977, 33, 915–925; (b) A. R. Katritzky, R. C. Patel
and F. G. Riddell, Angew. Chem., Int. Ed. Engl., 1981, 20, 521–529;
(c) T. A. Crabb and A. R. Katritzky, Adv. Heterocycl. Chem., 1984,
36, 1–173.
22 (a) M. Bixon, H. Dekker, J. D. Dunitz, H. Eser, S. Lifson, C.
Mosselma, J. Sicher and M. Svoboda, Chem. Commun., 1967,
360–362; (b) G. I. Birnbaum, G. W. Buchanan and F. G. Morin,
J. Am. Chem. Soc., 1977, 99, 6652–6656; (c) G. A. Sim, J. Chem.
Soc., Chem. Commun., 1987, 1118–1120; (d) E. Alessio, E.
Zangrando, R. Roppa and L. G. Marzilli, Inorg. Chem., 1998,
37, 2458–2463; (e) K. Ananda, S. Aravinda, P. G. Vasudev, K.
Muruga Poopathi Raja, H. Sivaramakrishnan, K. Nagarajan, N.
Shamala and P. Balaram, Curr. Sci., 2003, 85, 1002–1011.
23 (a) G. R. Desiraju, Angew. Chem., Int. Ed., 2007, 46, 2–17; (b) I. V.
Alabugin, M. Manoharan, M. Buck and R. J. Clark,
THEOCHEM, 2007, 813, 21–27.
24 M. J. T. Robinson, J. Chem. Soc., Chem. Commun., 1975, 844–845.
25 (a) M. Aroney and R. J. W. Le Fevre, J. Chem. Soc., 1960,
2161–2168; (b) S. F. Beach, J. D. Hepworth, J. Sawyer, G. Hallas,
R. Marsden, M. M. Mitchell, D. A. Ibbitson, A. M. Jones and G.
T. Neal, J. Chem. Soc., Perkin Trans. 2, 1984, 217–221.
26 E. Kleinpeter, Adv. Heterocycl. Chem., 2004, 86, 41–127.
27 I. V. Alabugin, J. Org. Chem., 2000, 65, 3910–3919.
28 SHELXTL6, Bruker-AXS, Madison, WI, USA, 2000.
29 R. A. Kendall, E. Apra, D. E. Bernholdt, E. J. Bylaska, M.
Dupuis, G. I. Fann, R. J. Harrison, J. Ju, J. A. Nichols, J.
Nieplocha, T. P. Straatsma, T. L. Windus and A. T. Wong,
Comput. Phys. Commun., 2000, 128, 260–283.
30 J. P. Perdew and A. Zunger, Phys. Rev. B, 1981, 23, 5048–5079.
31 J. M. Foster and S. F. Boys, Rev. Mod. Phys., 1960, 32, 300–302.
32 Chr. Møller and M. S. Plesset, Phys. Rev., 1934, 46, 618–622.
33 B. O. Roos and P. R. Taylor, Chem. Phys., 1980, 48, 157–173.
34 (a) W. J. Hehre, R. Ditchfield and J. A. Pople, J. Chem. Phys.,
1972, 56, 2257–2261; (b) J. S. Binkley, J. A. Pople and W. J. Hehre,
J. Am. Chem. Soc., 1980, 102, 939–947.
11 D. R. Tortolani and M. A. Poss, Org. Lett., 1999, 1, 1261–1262.
12 (a) S. M. McElvain and K. Rorig, J. Am. Chem. Soc., 1948, 70,
1820–1825; (b) S. Z. Vatsadze, M. A. Manaenkova, N. V. Sviriden-
kova, N. V. Zyk, D. P. Krut’ko, A. V. Churakov, M. Yu. Antipin, J.
A. K. Howard and H. Lang, Russ. Chem. Bull., 2006, 55, 1184–1194.
13 L. Wang, J. Sheng, H. Tian and C. Qian, Synth. Commun., 2004,
34, 4265–4272.
14 H. I. El-Subbagh, S. M. Abu-Zaid, M. A. Mahran, F. A. Badria
and A. M. Al-Obaid, J. Med. Chem., 2000, 43, 2915–2921.
15 M. D. Brown, M. J. Cook, G. Desimoni and A. R. Katritzky,
Tetrahedron, 1970, 26, 5281–5287.
16 (a) E. A. Meyer, R. K. Castellano and F. Diederich, Angew. Chem.,
Int. Ed., 2003, 42, 1210–1250; (b) M. Nishio, CrystEngComm,
2004, 6, 130–158.
17 (a) A. Bouchemma, P. H. McCabe and G. A. Sim, J. Chem. Soc.,
Perkin Trans. 2, 1989, 583–587; (b) A. Bouchemma, P. H. McCabe
and G. A. Sim, Acta Crystallogr., Sect. C, 1988, 44, 1469–1472; (c)
A. Bouchemma, P. H. McCabe and G. A. Sim, Acta Crystallogr.,
35 (a) T. H. Dunning, J. Chem. Phys., 1989, 90, 1007–1023; (b) R. A.
Kendall, T. H. Dunning and R. J. Harrison, J. Chem. Phys., 1992,
96, 6796–6806; (c) A. K. Wilson, T. Mourik and T. H. Dunning, J.
Mol. Struct., 1996, 388, 339–349.
36 (a) J. P. Foster and F. Weinhold, J. Am. Chem. Soc., 1980, 102,
7211–7218; (b) A. E. Reed, R. B. Weinstock and F. Weinhold, J.
Chem. Phys., 1985, 83, 735–746.
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008
1934 | New J. Chem., 2008, 32, 1924–1934