Conformational analysis and crystal structure of {[1-(3-chloro-4- fluorobenzoyl)-4-fluoropiperidin-4yl]methyl}[(5-methylpyridin-2-yl)methyl]amine, fumaric acid salt

Article abstract of DOI:10.1016/j.saa.2005.01.002

{[1-(3-Chloro-4-fluorobenzoyl)-4-fluoropiperidin-4yl]methyl} [(5-methylpyridin-2-yl)methyl]amine, fumaric acid salt (C₂₀H ₂₂ClF₂N₃O, C₄H₄O ₄) (1) was synthesized and characterized by the complete ¹H, ¹³C and ¹⁹F NMR analyses. The conformation of the piperidin ring, in the solution state, was particularly studied from the coupling constants determined by recording a double-quantum filtered COSY experiment in phase-sensitive mode. ¹H NMR line-shape analysis was used, at temperatures varying between -5 and +60 °C, to determine the enthalpy of activation of the rotational barrier around the CN bond. Compound 1 crystallizes in the triclinic space group P1 with a = 8.517(3) A, b = 12.384(2) A, c = 12.472(3) A, α = 70.88(2)°, β = 82.04(2)°, γ = 83.58(2)°. The results strongly indicate that the solid and solution conformations are similar. Thermal stability and phases transitions were investigated by thermal gravimetric analysis (TGA) and differential scanning calorimetry (DSC). Furthermore polymorphism screening was studied from recrystallization of 1 performed in seven solvents and by slurry conversion in water. The X-ray powder diffraction (XRPD) and differential scanning calorimetry results suggested that 1 crystallizes into one crystalline form which melts at 157 °C (ΔH = 132 J g^-1).

Full text of DOI:10.1016/j.saa.2005.01.002

Spectrochimica Acta Part A 62 (2005) 353–363
Conformational analysis and crystal structure of
{[1-(3-chloro-4-fluorobenzoyl)-4-fluoropiperidin-4yl]methyl}
[(5-methylpyridin-2-yl)methyl]amine, fumaric acid salt
J.P. Ribeta, ∗, R. Penaa, J.L. Maurela, C. Belinb, M. Tillardb,
B. Vachera, B. Bonnauda, F. Colpaerta
a
Institut de Recherche Pierre Fabre, 17 avenue Jean Moulin, 81106 Castres, France
b
Laboratoire des Agre´gats Mole´culaires et Mate´riaux Inorganiques, UMR 5072 CC15, Universite´ de Montpellier II,
Sciences et Techniques du Languedoc, 2 Place Euge`ne Bataillon, 34095 Montpellier Ce´dex 5, France
Received 15 November 2004; accepted 12 January 2005
Abstract
{[1-(3-Chloro-4-fluorobenzoyl)-4-fluoropiperidin-4yl]methyl}[(5-methylpyridin-2-yl)methyl]amine, fumaric acid salt (C20H22ClF2N3O,
C4H4O4) (1) was synthesized and characterized by the complete 1H, 13C and 19F NMR analyses. The conformation of the piperidin ring, in
the solution state, was particularly studied from the coupling constants determined by recording a double-quantum filtered COSY experiment
in phase-sensitive mode. 1H NMR line-shape analysis was used, at temperatures varying between −5 and +60 ◦C, to determine the enthalpy
¯
˚
of activation of the rotational barrier around ◦the C N bond. Compound 1 crystallizes in the triclinic space group P1 with a = 8.517(3) A,
◦
◦
˚
˚
b = 12.384(2) A, c = 12.472(3) A, α = 70.88(2) , β = 82.04(2) , γ = 83.58(2) .
The results strongly indicate that the solid and solution conformations are similar. Thermal stability and phases transitions were investigated
by thermal gravimetric analysis (TGA) and differential scanning calorimetry (DSC). Furthermore polymorphism screening was studied from
recrystallization of 1 performed in seven solvents and by slurry conversion in water. The X-ray powder diffraction (XRPD) and differential
scanning calorimetry results suggested that 1 crystallizes into one crystalline form which melts at 157 ◦C (ꢀH = 132 J g−1).
© 2005 Elsevier B.V. All rights reserved.
Keywords: Conformational analysis; Crystal structure; Thermal analysis; Polymorphism study; 5-HT1A receptor agonist
1. Introduction
the understanding of the relationship between the confor-
mation of the pharmacophore in the solution state and the
agonist activity at 5-HT1A receptors. Dynamic aspects are
studied by two-dimensional spectrometry (NOESY) because
the conformation of the amide bond is an important subject,
not only in pure chemistry but also in medicinal chemistry.
Rate constants for the rotational barrier are measured over
The present study was undertaken to determine the
conformation in solution and the structure in the solid
state of {[1-(3-chloro-4-fluorobenzoyl)-4-fluoropiperidin-
4yl]methyl}[(5-methylpyridin-2-yl)methyl]amine, fumaric
acid salt (C20H22ClF2N3O, C4H4O4; 1, Fig. 1) a novel 5-
HT1A receptor agonist. This molecule is characterized by
a high affinity for 5-HT1A receptors and a high selectivity
versus ␣1 and D2 [1,2]. The 5-HT1A subtype of serotonin
(5-HT) receptors are implicated in many psychiatric and neu-
rological disorders [3,4]. NMR studies are performed to help
1
a range of temperatures using H NMR line-shape analysis
from which the enthalpy of activation is obtained. These ob-
servations provide evidence of the usefulness of NMR spec-
trometry to establish the active conformation of the ligand.
Furthermore, the restricted rotation around the C N bond
in solution could explain the appearance of conformational
polymorphism in the solid state. Polymorphism is the abil-
ity for a compound to exist in more than one crystal form
∗
Corresponding author. Tel.: +33 5 63 71 42 79; fax: +33 5 63 71 42 99.
E-mail address: jean.paul.ribet@pierre-fabre.com (J.P. Ribet).
1386-1425/$ – see front matter © 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.saa.2005.01.002

354
J.P. Ribet et al. / Spectrochimica Acta Part A 62 (2005) 353–363
Fig. 1. Structural formula of 1 the {[1-(3-chloro-4-fluorobenzoyl)-4-fluoropiperidin-4yl]methyl}[(5-methylpyridin-2-yl)methyl]amine, fumaric acid salt.
with different unit cell parameters and conformational poly-
morphism occurs when different conformational isomers of
a compound crystallize as distinct polymorphs. Differences
in solubility between crystal forms can lead to difference in
bio-availability of solid dosage forms if the bio-availability
is dissolution limited. Thus, a polymorph screening is per-
formed by crystallization and slurry processes in various sol-
vents. The solid state property of the samples is studied by
differential scanning calorimetry (DSC), thermal gravimetric
analysis (TGA) and X-ray powder diffraction (XRPD). The
data presented herein, indicate that 1 exists in only one crystal
form. The structure of this polymorphic form is determined
with the help of crystallographic data.
directly stirred with [(5-methylpyridin-2-yl)methyl]amine in
basic and reductive media in methanol. This, was obtained
using a mixture of 1,4-diazabicyclo[2,2,2]octane (DABCO)
and sodium cyanoborohydride. This process gave directly the
{[1-(3-chloro-4-fluorobenzoyl)-4-fluoropiperidin-4yl]meth-
yl}[(5-methylpyridin-2-yl)methyl]amine with a 87% yield.
This oil was treated in ethyl acetate with fumaric acid alco-
holic solution to give a crystalline salt. The stoichiometry
of the salt was 1/1 molar and the molecular weight was
509.95 g mol−1. Chemical purity of sample was superior to
99.5% (w/w) evaluated by HPLC and internal normalization.
2.2. Nuclear magnetic resonance spectrometry (NMR)
All spectra were recorded non-spinning on a Bruker
Avance 400 spectrometer operating at the proton nominal fre-
quency of 400 MHz equipped with a 5 mm inverse multinu-
clear gradient probe-head. The spectrometer frequency was
400.13 MHz for 1H. Variable-temperature experiments were
made using the BVT 3300 and the BCU-05 accessories of the
instrument. Temperatureprobewascalibratedbystudyingthe
temperature dependence in the chemical shifts of the methy-
leneandhydroxylresonancesinan80%ethyleneglycol–20%
dimethylsulfoxide sample tube (Bruker Ref Z10129, 7493
A). The confidence interval was 1 ◦C.
2. Experimental
2.1. Synthesis
In a previous technique, illustrated by WO 98/22459 [5]
the title compound was obtained in nine steps. First, the 3-
chloro-4-fluoro-benzoylchloride was obtained starting from
the corresponding benzoic acid and thionyl chloride, and was
condensed with 4,4-ethylenedioxy-piperidine to give 8-(3-
chloro-4-fluorobenzoyl)-1,4-dioxa-8-azaspiro[4,5]decane.
Then, the 1-(3-chloro-4-fluorobenzoyl)piperidin-4-one was
obtained in acidic media, and a Darzens condensation was
carried out to give 6-(3-chloro-4-fluorobenzoyl)-2-cyano-1-
oxa-6-azaspiro[2.5]octane which was opened with hydroflu-
oric acid–pyridine complex to a cyanohydrine: [1-(chloro-
4-fluorobenzoyl)-4-fluoropiperidin-4-yl](cyano)methanol.
This cyanohydrine intermediate was the starting material of
our new reaction. Previously, this cyanohydrine was trans-
formed in [1-(chloro-4-fluorobenzoyl)-4-fluoropiperidin-
4-yl]methanol by reduction with sodium borohydride, and
then, activated with paratoluene sulfonyl chloride, con-
densed on potassium phtalimide to give the {[1-(chloro-4-
fluorobenzoyl)-4-fluoropiperidin-4-yl]methyl}amine, after
opening the phtalimide by ethanolamine. This {[1-(chloro-
4-fluorobenzoyl)-4-fluoropiperidin-4-yl]methyl}amine was
condensed on 5-methylpyridine-2-carbaldehyde to afford,
after sodium borohydride reduction, the desired compound.
The global yield, starting from the cyanohydrine was about
20%, calculated on a 2.5 kg size batch. A new one-pot
reaction replaces favorably this succession of five steps
synthesis (FR 2820743–A1) [6]. The cyanohydrine was
For the structural analysis by 1H NMR and 13C NMR the
compound was dissolved in dimethylsulfoxide-d6 (Eurisotop
D310-B, batch Q 2801) at approximately 17 mM concen-
tration and the temperature of the probe was maintained to
25 ◦C. The chemical exchange was studied by analysis of the
line shapes at temperatures varying between −5 and +60 ◦C
in methyl alcohol-d4 (isotopic enrichment 99.80%, Euriso-
top D324-B, batch L2831) with 0.03% of tetramethylsilane
used as internal standard. The chemical shifts, coupling con-
stants and line widths of the resonances were measured at low
temperatures before dynamic broadening. The measurements
were carried out for solution of the title compound having
concentration of ca. 14 mM. The line-shape analysis of the
1H NMR spectra and the simulation of the fluorine-2 split-
ting pattern were performed by means of a computer program
WINDNMR, Version 7.1.6, 2002 [7]. To optimize the exper-
imental conditions of the NOESY and of the GDQF-COSY
experiments, the sample was dissolved in D2O (Eurisotop,
D-214 EP, batch IO 481) at approximately 20 mM concen-
tration. For these experiments the solution was carefully fil-
tered through an external filter tip (45 ␮m, Polylabo 91943)

J.P. Ribet et al. / Spectrochimica Acta Part A 62 (2005) 353–363
355
in a dry 5 mm sample tube (New Era Enterprises NE-HP5-9”)
then the solution was subjected to five freeze–pump–thraw
cycles before being sealed under nitrogen. Chemical shifts
δ are given in parts per million (ppm) relative to T.S.P. ((3-
trimethylsilyl)-propionic acid-d4 sodium salt) obtained from
Merck (Art. 8652) used as internal standard and the coupling
constants in Hz. NOESY studies were acquired in the phase-
sensitive mode, using the noesytp program with a mixing
time set to 400 ms and 32 scans per time increment.
The double-quantum filtered GDQF-COSY experiment
was acquired in the phase-sensitive mode using time-
proportional phase increment (TPPI) with a spectral width of
903.18 Hz in both dimensions (cosygpmftp program), 1 × 2K
data points before and 2K × 2K after zero-filling, resulting in
a digital resolution of 0.44 Hz per point in the ω1 and ω2 di-
mensions. Processing was carried out using sine-bell squared
functions shifted by π/2.
and isotropic thermal displacement were refined. Electron
density residuals in the final Fourier difference map were
+0.49 and −0.28 e A−3. The ellipsoid plot was drawn using
˚
ORTEP-3 for windows [11] which is a MS-windows version
of the current release of ORTEP-III. All computations were
carried out on a Pentium 4 computer.
The powder diffraction measurements were undertaken
using a Philips analytical Xpert X-ray diffractometer
equipped with a copper tube, a hybrid monochromator (a
parabolic multilayer mirror and a two-crystal monochroma-
tor) and the Xcelerator detector. Rietveld analyses were per-
formed using the program LHPM-Rietica [12].
2.4. Differential scanning calorimetry
Differential scanning calorimetry was performed with a
DSC Q100 (TA Instrument) equipped with Universal Anal-
ysis 2000 software. The calibration (temperature and cell
constant) was done using an indium standard (SPIN, ref.
LGC2601). Assays were performed at rate of 5 ◦C min−1
from 0 to 170 ◦C in hermetic aluminium pans (TA instru-
ments, pan: 900793.901, lid: 900794.901) under nitrogen
purge at 50 mL min−1. Samples, about 5 mg, were weighed
with a Sartorius ultramicrobalance MP 8.
19F NMR spectra were recorded on the title compound
at 188.2 MHz on a Bruker AC200 spectrometer, at ambient
temperature, in methyl alcohol-d4 (Euriso-top, D324 B, batch
IO491) solution. Proton-decoupled and proton-coupled spec-
tra were recorded. Chemical shifts δ are given in parts per
million (ppm) relative to trifluoro-acetic acid in benzene-d6.
2.3. X-ray diffraction
2.5. Thermogravimetry analysis
The crystallized compound was obtained as follows:
1 g of 1 was dissolved in warm isopropanol (40 mL) and
allowed to stand, in an open flask, at room temperature for
8 days. The resulting crystals were filtered and dried under
reduced pressure. Enraf-Nonius CAD-4 diffractometer was
used for X-ray diffraction data collection at T = 293 K using
TGA and DTG curves were obtained using a TA High
Resolution 2950 thermogravimeter analyzer equipped with
the Universal Analysis 2000 software (TA Instruments) and
a nitrogen purge at 60 mL min−1. The sample was weighed
in an open aluminium crucible with a cross section area of
0.327 cm2 and deposited in a platinum pan. The sample size
was about 5–9 mg in order to cover the whole of the cru-
cible section uniformly. The thermobalance was calibrated
for temperature with the melting of tin. The magnitude and
linearity of the balance response were checked with standard
milligram masses.
˚
graphite monochromated Mo K␣ radiation, λ = 0.71069 A
and ω–θ scans. Three standard reflections were checked
every 100 reflections without any intensity decay. Cell
parameters were refined from 25 automatically centered
reflections θ = 7.6–11.82◦. A total of 3195 reflections were
measured up to θmax = 25◦ within the four octants hkl,
hkl, hkl, hkl. The structure was solved by direct methods
using the program SHELXS97 [8] and refinement was
carried out using SHELXL97 [9]. The absorption effects
(µ = 0.21 mm−1, transmission factors varying from 0.92
to 0.99) were corrected using the numerical procedure
provided by SHELX-76 program [10]. Intensities of equiv-
alent reflections were merged (Rint = 0.0336) into 3005
(from which 1635 with I > 2σ(I)) independent reflections
used for refinements. The final refinement was carried
out to R = 0.0504, ωR(F2) = 0.1509, weighting scheme
The non-isothermal TGA was recorded at 5 ◦C min−1
from room temperature to 450 ◦C.
¯
¯ ¯ ¯
2.6. Polymorph screening
Recrystallization of 1 was performed in the following
solvents: ethyl acetate, acetonitrile, methanol, ethanol, iso-
propanol, tetrahydrofuran, acetone. Sample was suspended
into 5 mL solvent and was stirred in an oil bath at 60 ◦C. Sol-
vent was added until dissolution. The solutions were cooled
at room temperature and filtered when 1 precipitated oth-
erwise solutions were cooled at +4 ◦C until crystallization
(24 h). Powders were analyzed by X-ray powder diffraction
and DSC.
Slurry conversion studies were realized in water. A satu-
rate solution was prepared with 100 mg of 1 and 0.75 mL of
water in vials closed with a screw stopple. The suspension
wasstirred16 hat45 ◦C, 150 rpm, withaNewBrunswickSci-
ω = 1/[σ2(F02) + (0.0706P)2]
where
P = (F02 + 2Fc2)/3,
S = 1.00. Extinction coefficient was refined to 0.006(2).
Hydrogen atoms attached to carbon are treated as riding,
following the SHELXL97 HFIX/AFIX instructions; they
were given an isotropic displacement parameter equal to
1.2 times the Ueq of the parent C atom. The hydrogen
atoms attached to nitrogen and oxygen were positioned
from a Fourier difference map, and both their positions

356
J.P. Ribet et al. / Spectrochimica Acta Part A 62 (2005) 353–363
Table 1
1H (400 MHz) and 13C (100 MHz) NMR spectrum of 1
Carbon
δ (13C) (CD3SOCD3), 25 ◦C
δ (1H) (CD3SOCD3), 25 ◦C
δ (1H) (CD3OD), 25 ◦C
δ (1H) (D2O), 25 ◦C
1
2
3
4
5
6
7
157.3 (d, 249.10)
119.7 (d, 17.95)
129.2
133.8 (d, 3.95)
127.7 (d, 7.85)
117.0 (d, 21.30)
166.6
–
–
–
–
–
–
7.68 (dd, 7.20, 1.95)
–
7.45 (ddd, 8.60, 5.0, 1.95)
7.50 (dd, 8.60, 8.60)
–
7.61
–
7.41
7.35
–
7.59
–
7.38
7.35
–
9 or 13
43.0
ax 3.15–2.95 (m)
eq 3.50–3.30 (m)
ax 3.30–3.10
eq 3.70–3.55
ax 3.30–3.20
eq 3.75–3.65
10 or 12
32.5
ax 1.75–1.60 (m)
eq 1.90–1.80 (m)
ax 1.80–1.70
eq 2.05–1.90
ax 1.85–1.70
eq 2.05–1.90
11
94.5 (d, 171.65)
31.9
–
–
–
12 or 10
ax 1.85–1.75 (m)
eq 2.05–1.90 (m)
ax 1.90–1.80
eq 2.20–2.02
ax 1.90–1.80
eq 2.20–2.10
13 or 9
37.5
ax 3.30–3.20 (m)
eq 4.40–4.20 (m)
ax 3.50–3.30
eq 4.60–4.40
ax 3.50–3.40
eq 4.45–4.35
14
16
17
19
20
21
22
23
55.0
53.5
2.85 (d, 19.75)
3.90 (s)
–
8.37 (dd, 1.50, 0.85)
–
7.60 (dd, 8.0, 1.50)
7.35 (d, 8.0)
2.30 (s)
3.15
4.20
–
8.43
–
7.68
7.37
2.35
6.70
–
3.43
4.40
–
8.45
–
7.78
7.44
2.35
6.64
–
155.0
148.9
131.5
137.0
121.5
17.5
CH-fumaric acid
CO-fumaric acid
134.3
166.5
6.60 (s)
–
Chemical shifts (δ, ppm) and coupling constants (J, Hz in parentheses).
entific stirrer. Afterwards the suspension was filtered, dried
at 40 ◦C under vacuum and analyzed by DSC.
After crystallization and slurry conversion processes the
stoichiometry of samples was checked by HPLC.
first order multiplet corresponding to its coupling with H-6
(3J5–6 = 8.60 Hz), H-3 (4J5–3 = 1.95 Hz) and the fluorine nu-
cleus F-1 (4J5–F1 = 5.0 Hz). Finally H-3 showed a doublet of
doublets at δ 7.68 (4J3–5 = 1.95 Hz, 4J3–F1 = 7.20 Hz).
The aliphatic region (between 1.5 and 4.5 ppm) showed
a large doublet at δ 2.85 owing to the coupling of the
two protons H-14 with the vicinal fluorine nucleus F-2
(3J14–F2 = 19.75 Hz). The H-16 protons appeared downfield
(δ 3.90) as singlet. The piperidin ring nuclei H-9eq and H-13eq
were strongly deshielded comparatively to H-9ax and H-13ax
due to the magnetic anisotropic effect of the carbonyl group.
The nuclei H-9eq (δ ≈ 3.4) and H-13eq (δ ≈ 4.3) were in dif-
ferentchemicalenvironmentsandthereforetheyhaddifferent
resonance frequencies. Consequently the protons H-9ax and
H-13ax were assigned at δ ≈ 3.0 and δ ≈ 3.3. It was known
that for piperidin derivatives with a sterically fixed chair con-
formation, chemical differences of 0.10–1.0 ppm between the
axial and equatorial nuclei were found, the axial nuclei being
more strongly shielded than the equatorial. However the sig-
nal multiplicity of the protons H10–12ax and H10–12eq was not
resolved from the 1D NMR spectrum owing to the superpo-
sition of their resonances. This problem was outlined by per-
forming a double-quantum filtered COSY in phase-sensitive
mode (Fig. 2). Finally the secondary amine and carboxylic
acid protons showed a broad resonance signal between δ 10.0
and 9.50. The phase-sensitive COSY with a double-quantum
3. Results and discussion
3.1. Assignment of 1H NMR spectrum
The numbering system adopted to designate the atoms
of the molecule (Fig. 1) was used to facilitate the spectral
assignments and did not correspond to the Organic Chemistry
Nomenclature Commission of IUPAC.
Concerning the pyridin ring the signal at δ 8.37 was as-
signed to H-19 (H-␣ pyridin, Table 1). This proton exhib-
ited two long-range couplings, the former with H-21 and
the last with H-22 (4J19–21 = 1.50 Hz and 5J19–22 = 0.85 Hz).
The doublet of doublets at δ 7.60 was characteristic of H-
21 and was in agreement with the di-substituted pyridin ring
4
(3J21–22 = 8.0 Hz and J21–19 = 1.50 Hz). Thus, the nucleus
H-22 was easily assigned at δ 7.35 (d, 3J22–21 = 8.0 Hz).
Concerning the tri-substituted phenyl ring the pseudo-
triplet at δ 7.50 was due to the coupling of H-6 with H-5
and of H-6 with the fluorine nucleus F-1 (3J6–5 = 8.60 Hz and
3J6–F1 = 8.60 Hz). At δ 7.45, H-5 exhibited a characteristic

J.P. Ribet et al. / Spectrochimica Acta Part A 62 (2005) 353–363
357
Fig. 2. Double-quantum filtered H,H-COSY spectrum of 1 in D2O. The cross-peaks are pure antiphase, negative cross-peaks are represented by dashed contours.
Spectral width in F2 and F1 = 903.18 Hz. The region of interest shows the multiplet structure of the cross-peak between H-10ax and H-9eq
.
3JF2-H10eq would be inferior to 8 Hz [13]. From the expanded
multiplet shown in Fig. 2 the geminal coupling constant be-
tween H-10ax and H-10eq was determined equal to 13.50 Hz,
the vicinal coupling constant between H-10ax and H-9eq was
found to 5.50 Hz and finally the diaxial protons H-10ax and
H-9ax showed a typical coupling constant of about 13.50 Hz.
filter was the method of choice for assigning the coupling
network in the 1.50–4.0 ppm region. The correlation peaks
observed in pure absorption phase provided detail coupling
information. The axial orientation for the fluorine F-2 was
corroborated by its coupling constant with H-10ax and H-12ax
(3J = 36.50 Hz). This large value reflected an antiperiplanar
disposition between the fluorine atom and the axial protons.
In addition the coupling constant of fluorine with the equa-
torial protons was found equal to 10.50 Hz. These observa-
tions confirmed the rigid chair conformation of the piperidin
ring and the axial orientation of the fluorine-2. If the fluorine
nucleus was equatorial the coupling constant 3JF2-H10ax and
3.2. Assignment of 13C NMR spectrum
By using heteronuclear shift correlated experiment such as
HMQC and HMBC we identified all aromatic carbons and all
aliphatic carbons from the knowledge of the corresponding

358
J.P. Ribet et al. / Spectrochimica Acta Part A 62 (2005) 353–363
1H NMR chemical shifts (Table 1). The carbon C-1 at δ 157.3
exhibited a large coupling constant with the fluorine nucleus
F-1 (1JC1–F1 = 249.10 Hz). Geminal couplings were found
to be 2JC2–F1 = 17.95 Hz and 2JC6–F1 = 21.30 Hz. Long-range
couplings through more than two bonds were also determined
accurately, 3JC5–F1 = 7.85 Hz and 4JC4–F1 = 3.95 Hz. A value
of 171.65 Hz was found for coupling between C-11 and F-
2 and the geminal coupling constant between C-14 and F-2
gave 2J = 19.75 Hz [14].
protons H-19 and H-21. NOE appeared between H-16, H-22
and between H-16, H-14. Cross-peaks due to chemical ex-
change were often more intense than those due to NOE. For
example the slice at δ 3.45 exhibited a large downward cross-
peak at δ 3.25 corresponding to the exchange position in the
molecule of the two axial protons H-9 and H-13. These two
axial protons were in different chemical environment because
of the restricted rotation around the C7 N8 bond. Therefore,
1H NMR line-shape analysis was used to determine the rota-
tional barrier of this C N bond.
3.3. Assignment of 19F NMR spectrum
3.5. 1H NMR line-shape analysis
Thefluorineresonanceswerecharacterizedbylargechem-
ical shifts. The spectrum exhibited two signals under proton
noise decoupling at δ −34.9 (F-1) and δ −87.3 (F-2). These
valueswereingoodagreementwiththosepreviouslyreported
in the literature [15].
C N bond between carbonyl group and nitrogen atom has
significant double bond character. At −5 ◦C, in methanol-
d4, the rotation around the C7 N8 bond was sufficiently
slow on the NMR time scale to show the peaks of both H-
13eq and H-9eq at respectively, δ 4.55 and 3.65 (Fig. 4). At
higher temperatures the fine structure became less sharp and
the chemical shift difference between H-13eq and H-9eq de-
creased. At +60 ◦C with the more rapid rotation around the
C7 N8 bond, H-13eq and H-9eq became chemically equiv-
alent and the lines coalesce. The enthalpy of activation for
the dynamic process (ꢀH* = 62.0 0.9 kJ mol−1) was de-
duced from the slope of the Eyring plot [18,19]. At +60 ◦C,
H-10ax and H-12ax were also chemically equivalent and ap-
From the splitting pattern of the signal at δ −87.3
(Fig. 3) we deduced the coupling constants through
three bonds (3JF2–H12ax = 36.50 Hz, JF2–H12eq = 10.50 Hz
3
3
and JF2–H14 = 19.75 Hz). These values confirmed the axial
orientation of the fluorine-2 and the rigid chair conformation
of the piperidin ring. From the coupling pattern of the sig-
nal at δ −34.9 the coupling constants between the fluorine-1
and the aromatic protons were confirmed: 3JH6–F1 = 8.60 Hz,
4JH5–F1 = 5.0 Hz and 4JH3–F1 = 7.20 Hz.
peared between δ 1.65 and 1.90 as one well resolved multiplet
∼
(Fig. 5). The multiplet splitting ( 36.50, 13.50, 13.50 and
=
3.4. Nuclear Overhauser exchange spectrometry
5.50 Hz) confirmed the axial configuration of the fluorine-2
at the temperature of +60 ◦C. The enthalpy of activation in
D2O was found to ꢀH* = 83.5 1.3 kJ mol−1 compared with
62.0 0.9 kJ mol−1 in methanol. This could be attributed to
stabilization of the transition state by hydrogen bonding.
The phase-sensitive NOESY spectrum of 1 showed posi-
tive NOEs owing to the small correlation time of the molecule
in D2O at 25 ◦C (ωτc < 1.12). The diagonal peaks had a nega-
tive phase and the cross-peaks due to chemically exchanging
protons had a phase like diagonal [16–17]. The slice at δ 4.40
exhibited two upward signals at δ 3.25 and 3.45 representing a
positive NOE between H-9eq or H-13eq and H-9ax or H-13ax.
Similarly the slice at δ 3.70 showed positive NOE between
H-9eq or H-13eq and H-9ax or H-13ax. The same observation
was made between H-10eq/H-12eq and H-10ax/H-12ax. NOE
cross-peaks were also observed between H-23 and the two
3.6. Single crystal and powder X-ray study
Compound 1 crystallizes in the triclinic space group
¯
˚
˚
P1 with cell parameters a◦= 8.517(3) A, b = 12.384(2) A, c =
◦
◦
˚
12.472(3) A, α = 70.88(2) , β = 82.04(2) and γ = 83.58(2) ,
M = 509.93, V = 1227.8(5) A , dcal = 1.379 Mg m−3. The unit
cell contains two units of formula C24H26F2N3O5Cl. The
asymmetric unit is composed of one C20H22F2N3OCl
molecule and one fumaric acid C4H4O4 molecule. The
atomic positions and equivalent displacement parameters for
atoms are reported in Table 2. Some selected structural pa-
rameters as inter-atomic distances, bond angles and torsion
dihedral angles are indicated in Table 3.
3
˚
The molecule C20H22F2N3OCl is protonated at the ni-
trogen N15 site by a hydrogen atom from a molecule of
fumaric acid. The ORTEP representation of the protonated
molecule C20H23F2N3OCl+ is given in Fig. 6. As expected,
the aromatic chlorofluorobenzoyl cycle is flat with consti-
tuting atoms (Cl, F1 and C1–C7) deviating no more than
˚
0.035 A from the mean atomic plane. With a deviation of
˚
0.09 A from the mean plane (atoms N18, C16, C17, C19–C23),
Fig. 3. Characteristic splitting pattern of the fluorine-2 at 188.2 MHz. Top,
experimental and bottom, calculated 19F NMR spectrum.
the methylpyridin cycle is also very flat. The atoms N8, C9,

J.P. Ribet et al. / Spectrochimica Acta Part A 62 (2005) 353–363
359
Fig. 4. Temperature dependence of the H-13eq and H-9eq proton resonances in methyl alcohol-d4. The observed (a) and the calculated (b) 400.13 MHz spectrum
at several temperatures.
Fig. 5. 1H NMR spectra of 1 (a) in DMSO-d6 at 25 ◦C, (b) in D2O at 25 ◦C, (c) in MeOH-d4 at −5, +25 and +60 ◦C.

360
J.P. Ribet et al. / Spectrochimica Acta Part A 62 (2005) 353–363
Fig. 6. ORTEP view of the {[1-(3-chloro-4-fluorobenzoyl)-4-fluoropiperidin-4yl]methyl}[(5-methylpyridin-2-yl)methyl]amine protonated molecule. Displace-
ment ellipsoids are drawn at the 50% probability level for non-hydrogen atoms.
Fig. 7. Representation of the protonated molecule interacting with fumaric acid molecule via hydrogen bonding (50% probability level ellipsoids). For clarity,
hydrogen atoms attached to carbon are omitted.
Fig. 8. The infinite chains formed by fumaric acid molecules (50% probability level ellipsoids) through hydrogen bonding.

J.P. Ribet et al. / Spectrochimica Acta Part A 62 (2005) 353–363
361
˚
Fig. 9. The experimental X-ray powder pattern (λ = 1.5418 A) of compound 1 (data points) and the calculated diagram from Rietveld refinement (single line).
Line at the bottom represents the difference between experimental and calculated diffractograms. Positions of individual reflections are indicated by markers.
Fig. 10. DSC, TG and DTG profiles of 1.

362
J.P. Ribet et al. / Spectrochimica Acta Part A 62 (2005) 353–363
C13, C7 of the tertiary amine, and oxygen atom O1 are lo-
Table 2
Atomic coordinates (×104) and equivalent isotropic displacement parame-
cated in a same plane (mean deviation of 0.072). The F2
fluorine atom is found attached at the axial position to the
cyclohexane-like cycle (N8, C9–C13) with chair conforma-
tion. As represented in Fig. 7, the hydrogen atoms on nitrogen
2
3
˚
ters (A × 10 )
Atom
x
y
z
Ueq
Cl
F1
F2
O1
O2
O3
O4
O5
C1
C2
C3
C4
C5
C6
C7
N8
−6808(2)
−6101(4)
2239(3)
−318(5)
2956(3)
2050(3)
4543(4)
3433(7)
−4824(8)
−4975(6)
−3659(6)
−2212(6)
−2081(7)
−3377(8)
−805(6)
−90(5)
3672(1)
4951(3)
3137(2)
1535(4)
7023(3)
5338(2)
8615(3)
8304(4)
4339(4)
3706(4)
3100(4)
3099(4)
3719(5)
4354(5)
2342(5)
2553(3)
3627(4)
3410(4)
2573(3)
1508(4)
1778(4)
2262(3)
3267(3)
2951(4)
2295(4)
2498(3)
1925(4)
1181(5)
1002(5)
1563(5)
573(5)
10771(1)
12209(3)
7360(2)
11762(3)
3896(3)
4836(3)
2415(3)
1073(3)
11891(4)
11199(4)
10843(4)
11219(4)
11932(4)
12256(5)
10966(5)
9923(3)
8989(4)
7895(4)
7597(4)
8619(4)
9674(4)
6559(4)
5504(3)
4450(4)
4202(4)
4585(3)
4355(4)
3734(5)
3359(5)
3588(5)
3523(5)
4450(4)
4679(3)
551(3)
83(1)
94(1)
58(1)
113(2)
50(1)
49(1)
63(1)
139(2)
62(2)
54(1)
53(1)
52(1)
66(2)
72(2)
65(2)
54(1)
54(1)
45(1)
43(1)
51(1)
60(1)
41(1)
40(1)
51(1)
46(1)
53(1)
63(2)
70(2)
87(2)
75(2)
109(3)
37(1)
36(1)
48(1)
52(1)
N15 are involved in bent medium strength hydrogen bonding
withfumaricacidmolecules(distancesN15 O3 are2.695and
˚
2.786 A). Furthermore, as represented in Figs. 7 and 8, hy-
drogen bonding also occurs between neighbouring fumaric
acid molecules. Through linear hydrogen bonding involv-
˚
ing atoms O2 and O4 separated by 2.578 A, fumaric acid
molecules form infinite chains interacting (N15 O3 hydro-
gen bonds) with the protonated molecule C20H23F2N3OCl+.
Owing to the conjugated action of nitrogen basic centers on
one of the fumaric acid molecules (see above the description
of the N15 O3 hydrogen bond), the linear O2 · · · H24 O4
(angle O2 H24 O4 of 177(5)◦) hydrogen bond is asym-
metric with O4 H24 and O2 H24 distances of 1.22(7) and
C9
−295(6)
−512(5)
794(5)
˚
1.36(7) A. The remaining oxygen atom O5 of fumaric acid
C10
C11
C12
C13
C14
N15
C16
C17
N18
C19
C20
C21
C22
C23
C24
C25
C26
C27
which is not involved in hydrogen bonding displays a larger
atomic displacement parameter. Same feature is also found
fortheterminalmethylgroup(atomC23)ofthemethylpyridin
cycle.
The X-ray powder pattern was initially recorded (standard
acquisition times and 2.5–30◦ theta domain) for identification
purpose and polymorphism checking. The diffractogram is
characteristic of a single phase with all diffraction lines in-
dexing in the triclinic cell. Profile refinement led to refined
994(6)
1220(6)
626(5)
625(5)
679(6)
2197(6)
3468(5)
4840(6)
5013(8)
3672(10)
2253(8)
6588(8)
3168(5)
4815(5)
5020(5)
4258(6)
˚
˚
˚
parameters a = 8.546(7) A, b = 12.402(5) A, c = 12.480(4) A,
α = 70.86(4)◦, β = 82.09(5)◦, γ = 83.58(5)◦. Fairly good Ri-
etveld agreement factors, RBragg = 7.88% and Rp = 9.80%, are
obtained with the atomic positions and displacement parame-
ters from X-ray single crystal structure determination. These
values exclude the presence of any other polymorph in the
sample. The experimental and calculated X-ray powder pat-
5978(4)
5517(3)
9851(4)
8840(4)
1374(5)
Ueq is defined as one-third of the trace of the orthogonalized Uij tensor.
˚
terns (λ = 1.5418 A) are represented in Fig. 9.
3.7. Thermal analysis and polymorph screening
The melting property and the thermal stability of 1 were
investigated by DSC and TGA, respectively. DSC profile
showed one only endothermic effect at 157 ◦C which cor-
responded to the melting of sample. The TG and DTG pro-
files (Fig. 10) indicated that the beginning of the mass loss
occurred at the end of melting (1%, w/w at 165 ◦C). Two de-
composition stages occurred after the melting with maximum
kinetics at 183 and 346 ◦C, respectively. In our study condi-
tions of polymorph screening, one only crystalline form of 1
was identified by DSC and XRPD. The melting property of
Table 3
˚
Selected inter-atomic distances (A), bond angles and torsion dihedral angles
(◦)
N15 O3
N15 H15B
N15 H15B O3
2.786(5)
0.81(4)
141(4)
2.695(5)
1.04(5)
159(4)
2.578(4)
1.22(7)
1.36(7)
177(5)
O1 C7 N8
O1 C7 C4
N8 C7 C4
C7 N8 C13
C7 N8 C9
C13 N8 C9
F2 C11 C14
F2 C11 C12
C14 C11 C12
F2 C11 C10
C16 N15 C14
121.0(5)
118.6(5)
120.4(4)
119.9(4)
125.2(4)
114.0(4)
106.3(3)
106.8(4)
110.9(4)
106.1(3)
113.3(4)
a
N15 O3
N15 H15A
N15 H15A O3
O2 O4
O4 H24
O2 H24
a
this form was T = 157 ◦C and ꢀH = 130–140 J g−1
.
O2 H24 O4
O2 C24
O3 C24
C27 O5
C27 O4
a
1.257(5)
1.248(5)
1.190(6)
1.287(6)
Acknowledgements
C13 N8 C7 O1
C27 O4 O2 C24
4.9(9)
100.8(5)
The authors wish to thank Prof. Hans J. Reich (University
of Wisconsin, USA) with regard to computer software and are
Symmetry operation −x, 1 − y, 1 − z.

J.P. Ribet et al. / Spectrochimica Acta Part A 62 (2005) 353–363
363
grateful to Prof. A.D. Bain (McMaster University, Canada)
for advice in chemical exchange line-shape analysis.
[7] H.J. Reich, WINDNMR, J. Chem. Educ. Software, 1996, 3D, 2.
[8] G.M. Sheldrick, SHELXL 97: A Program for Refining Crystal Struc-
tures, University of Go¨ttingen, Germany, 1997.
[9] G.M. Sheldrick, SHELXS: A Program for Crystal Structures Solu-
tion, Vol. 97, University of Go¨ttingen, Germany, 1997.
[10] G.M. Sheldrick, SHELX-76, Program for Crystal Structure Determi-
nation, University of Cambridge, Cambridge, 1976.
[11] L.J. Farrugia, J. Appl. Cryst. 30 (1997) 565.
[12] B.A. Hunter, C.J. Howard, LHPM—Rietica: A Computer Program
for Rietveld Analysis of X-ray and Neutron Powder Diffraction Pat-
terns, version 1.7.7, 1998, ANSTO, Lucas Heights Research Labo-
ratories, Menai, Australia.
References
[1] B. Vacher, B. Bonnaud, P. Funes, N. Jubault, W. Koek, M.B. Assie´,
C. Cosi, J. Med. Chem. 41 (1998) 5070–5083.
[2] B. Vacher, B. Bonnaud, P. Funes, N. Jubault, W. Koek, M.B. Assie´,
C. Cosi, M. Kleven, J. Med. Chem. 42 (1999) 1648–1660.
[3] W. Koek, J.-F. Patoiseau, M.B. Assie´, C. Cosi, M. Kleven, E.
Dupont-Passelaigue, E. Carilla-Durand, J.-P. Valentin, G. John, J.P.
Pauwels, J.-P. Tarayre, F.C. Colpaert, J. Pharmacol. Exp. Ther. 287
(1998) 266–283.
[13] E. Pretsch, T. Clerc, J. Seibl, W. Simon, Tables of Spectral Data
for Structure Determination of Organic Compounds, Springer-Verlag,
1983.
[4] F.C. Colpaert, J.-P. Tarayre, W. Koek, J.P. Pauwels, L. Bardin, X.J.
Xu, Z. Wiesenfeld-Hallin, C. Cosi, E. Carilla-Durand, M.B. Assie´,
B. Vacher, Neuropharmacology 43 (2002) 945–958.
[5] B. Vacher, B. Bonnaud, W. Koek, Pyridin-2-yl-methylamine Deriva-
tives, Method of Preparing and Application as Medicine, WO
98/22459, 1998.
[6] J.L. Maurel, B. Bonnaud, J.P. Ribet, B. Vacher, Proce´de´ et in-
terme´diaires de synthe`se pour la pre´paration de de´rive´s de pyridin-
2-yl-methylamine, FR 2820743–A1, 2001.
[14] H.O. Kalinowski, S. Berger, S. Braun, Carbon-13 NMR Spec-
troscopy, John Wiley and Sons, 1988.
[15] H. Gu¨nther, NMR Spectroscopy, John Wiley and Sons, 1980.
[16] J. Jeener, B.H. Meier, P. Bachmann, R.R. Ernst, J. Chem. Phys. 71
(11) (1979) 4546–4553.
[17] M.H. Levitt, Spin Dynamics, John Wiley and Sons, 2001.
[18] K.C. Brown, R.L. Tyson, J.A. Weil, J. Chem. Educ. 75 (12) (1998)
1632–1635.
[19] A.D. Bain, Prog. Nucl. Magn. Reson. Spectrosc. 43 (2003) 63–103.