Unsymmetrically substituted triazacyclohexanes and their chromium complexes

Article abstract of DOI:10.1016/j.poly.2010.01.008

The unsymmetrically N-substituted N,N′-Ar₂-N″-R-1,3,5-triazacyclohexanes 1-4 (Ar = ortho- or para-fluorophenyl, R = n- or iso-propyl) can be obtained in good yields from a one-step condensation reaction with excess amine. Solid state structures of 1-4 resemble closely those of their triaryl-substituted analogues. The condensation reaction to 4 was looked at by detailed NMR investigations and revealed that amine/aniline exchange is occurring in solutions containing free aniline even at ambient conditions setting up an equilibrium between all possible symmetrical and unsymmetrical triazacylcohexanes. Selective crystallisation of 4 from the solution drives the reaction to high yields of 4. Complexes 1-4 react readily with CrCl₃ or CrCl₃(THF)₃ to form the corresponding CrCl₃ complexes. The complexes are insoluble in non-polar solvents and decompose under decomplexation in coordinating solvents.

Full text of DOI:10.1016/j.poly.2010.01.008

Polyhedron 29 (2010) 1399–1404
Contents lists available at ScienceDirect
Polyhedron
journal homepage: www.elsevier.com/locate/poly
Unsymmetrically substituted triazacyclohexanes and their chromium complexes
d,
Saïda Latreche a, Abdelhamid Mousser b, Gabriele Kociok-Köhn c, Randolf D. Köhn c, , Frank Schaper
*
*
a Département de chimie, Université Tébessa, 12000 Tébessa, Algeria
b Département de chimie, Université Montouri Constantine, 25000 Constantine, Algeria
c Department of Chemistry, University of Bath, Bath BA2 7AY, UK
d Département de chimie, Université de Montréal, Montréal, Québec, Canada H3C 3J7
a r t i c l e i n f o
a b s t r a c t
The unsymmetrically N-substituted N,N0-Ar2-N00-R-1,3,5-triazacyclohexanes 1–4 (Ar = ortho- or para-
fluorophenyl, R = n- or iso-propyl) can be obtained in good yields from a one-step condensation reaction
with excess amine. Solid state structures of 1–4 resemble closely those of their triaryl-substituted ana-
logues. The condensation reaction to 4 was looked at by detailed NMR investigations and revealed that
amine/aniline exchange is occurring in solutions containing free aniline even at ambient conditions set-
ting up an equilibrium between all possible symmetrical and unsymmetrical triazacylcohexanes. Selec-
tive crystallisation of 4 from the solution drives the reaction to high yields of 4. Complexes 1–4 react
readily with CrCl3 or CrCl3(THF)3 to form the corresponding CrCl3 complexes. The complexes are insoluble
in non-polar solvents and decompose under decomplexation in coordinating solvents.
Article history:
Received 1 November 2009
Accepted 6 January 2010
Available online 28 January 2010
Keywords:
Triazacyclohexane
Chromium complexes
Chromium(III)
Selective crystallisation
Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction
major products, seemingly indicating a higher reactivity of anilines
in TAC formation. Analytically pure, unsymmetrically substituted
1,3,5-Triazacyclohexanes (TAC) are tripodal ligands, which tend
to adopt a fac-coordination in octahedral metal complexes. In this
they resemble their triazacyclononane analogues, but are synthet-
ically more easily accessible by simple condensation from primary
amines and formaldehyde. With the exception of recent examples
of Ph3TAC-complexes of W(0) [1], their metal complexes have so
far been restricted to triazacyclohexanes (TAC) ligands carrying
three aliphatic substituents. We have previously investigated
ligands can be isolated, however, in greater than 50% yield from
the product mixtures obtained by reactions of a 2:1 mixture of pri-
mary amine:aniline with 1 equiv. of formaldehyde per equivalent
of NH2 group (1 and 3 in Scheme 1). When the amount of formal-
dehyde is limited to the amount required for the desired product
(about 0.5 equiv. per NH2 group) the yield could be raised to 73%
and 91% for 2 and 4, respectively. The bisarylalkyl substituted tri-
azacyclohexanes crystallise from the reaction mixture either dur-
ing the reaction or from a cooled hexane solution and can
therefore be obtained analytically pure. Ortho- and para-fluoro-
substituted anilines were employed to utilise the high sensitivity
and resolution of 19F NMR for the analysis product mixtures and
complexes [8].
Crystals of 4, suitable for an X-ray diffraction study, were ob-
tained by slow diffusion of hexane into a dichloromethane solution
of the ligand (Fig. 1). Preliminary results of the crystal structure
determinations of 1–3 have been reported independently [9–11]
and we will focus here on similarities and differences between
ligands 1–4. All four triazacylcohexane ligands adopt the diaxial-
equatorial chair conformation typically observed in most
aryl-substituted and all mono-fluorophenyl-substituted triazacy-
clohexanes in the solid state [12]. The alkyl substituent in 1–4
occupies the equatorial position, indicating higher 1,4-interactions
for alkyl than for aryl substituents. The Cipso–Cipso distances of
3.22–3.30 Å between the phenyl-substituents (Table 1) are notably
the coordination of aryl-substituted TACs (Ar
= ortho-C6H4F,
para-C6H4F, para-anisyl, and para-tolyl) to CrCl3 as part of the
investigations into the coordination chemistry of N-substituted
1,3,5-triazacyclohexanes undertaken in the Köhn group [2–6] and
found that the purple complexes formed upon those reactions
are extremely sensitive to moisture, insoluble in non-coordinating
solvents and rapidly dissolve in coordinating solvents under
decomplexation of the TAC ligand [7]. We thus decided to investi-
gate the chemistry of unsymmetrically substituted TAC ligands,
carrying aliphatic as well as aromatic substituents.
2. Results and discussion
2.1. Ligand syntheses
Reactions of formaldehyde with equimolar mixtures of primary
amines and anilines tend to give triaryl-substituted TACs as the
shorter than the optimal p-stacking distance expected for a sand-
wiched or parallel-displaced benzene dimer [13,14]. Consequently,
* Corresponding authors.
E-mail addresses: r.d.kohn@bath.ac.uk (R.D. Köhn), Frank.Schaper@umontreal.ca
the aromatic rings are angled away from each other with angles
(F. Schaper).
0277-5387/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.poly.2010.01.008

1400
S. Latreche et al. / Polyhedron 29 (2010) 1399–1404
(Cipso–Cipso 3.56 and Cpara–Cpara 6.08 Å). DFT is known to underesti-
Ar
mate long-range van der Waals interactions and introduction of
Grimme’s semiempirical van der Waals correction [15] to the
DFT calculation optimises to a structure with much closer aryl
groups (Cipso–Cipso 3.16 and Cpara–Cpara 4.52 Å) similar to the
observed contacts. Thus, an attractive van der Waals interaction
between the aryl groups is responsible for the stability of the con-
formation found in 1–4.
N
RNH2 + 2 ArNH2 + 3 H2CO
R
R
Ar
N
N
- 3 H2O
1 4
-
CrCl3/toluene or
CrCl3(THF)3/CH2Cl2
Ar
1 5
,
: Ar = ortho -C6H4F, R = n-Pr
N
Ar
2, 6:
R = iso-Pr
N
N
2.2. Selectivity in ligand synthesis
3 7
,
: Ar = para- C6H4F, R = n-Pr
Cr
Cl
4 8
,
:
R = iso-Pr
Cl
Cl
5-8
The high yield for 1–4 from reaction mixtures containing very
different aniline to amine ratios (1:2 rather than 2:1) was surpris-
ing given that mixtures of two different amines tend to result in
roughly statistical mixtures of unsymmetrically substituted TACs
[16–20]. Therefore we investigated the formation of 4 in more de-
tail. When all volatiles are removed from the reaction solution, the
NMR spectrum showed the oily residue to contain mostly 4, but
also significant amounts of free aniline, all other combinations of
aryl and iPr substituted triazacyclohexanes and some products as-
signed to intermediates containing ring-opened, iminium cations
(vide supra).
Scheme 1.
Addition of p-fluoroaniline to the NMR sample resulted in a sig-
nificant change of the product distribution favouring the more ani-
line rich products with release of free iPrNH2 within minutes. A
new equilibrium distribution is reached after a few days. The same
redistribution can be observed when p-fluoroaniline is added to a
CDCl3 solution of iPr3TAC. Fast initial change is leading to equilib-
rium distribution after about one week. The rate of redistribution is
influenced by the acidity of the solution. While the equilibration is
expected to follow complex kinetics the decrease of the free aniline
concentration follows approximately a first order exponential rate
law and can be used as an estimate of the rate. A wet or even acid
containing sample shows much faster reaction than a solution con-
taining base (NEt3). Commercial wet aniline (a few% water) had a
half life toward equilibrium of about 10 min. Addition of 50 mol%
NEt3 per amine slowed down the reaction to about 130 min half
life, while the equilibrium was reached by the time the first NMR
spectrum was taken upon addition of 4 mol% triflic acid or extra
water. Dried aniline (MgSO4) showed much slower reaction (half
life of 500 min, details in Supporting material). This fast exchange
by the addition of aniline was unexpected as previous unpublished
results in the Köhn group showed amine exchange with triazacylo-
hexanes taking place at 100 °C or more.
Fig. 1. Crystal structure of ligand 4. Hydrogen atoms and the second disordered site
for C23 omitted for clarity. Thermal ellipsoids are drawn at the 50% probability
level.
between their mean planes of 33–44° and Cpara–Cpara distances ap-
prox. 2 Å longer than the distance of the respective ipso-carbon
atoms. Despite their V-like conformation, the rings remain surpris-
ingly coplanar, with tilt-angles of less than 10° (Table 1). In
general, TAC 1–4 show identical structures with insignificant
differences in bond lengths and with close resemblance to the
structures of their tris(monofluorophenyl)-substituted analogues.
A DFT calculation on the structure of 4 reproduces the geometry
well except for a larger repulsion between the aryl groups
Similarly, isolated (>95 wt.% pure) 4 showed significant redistri-
bution with 1.4 equiv. p-fluoro-aniline within hours and reached
equilibrium within two days containing Ar3TAC, Ar2iPrTAC and Ar-
iPr2TAC in a ratio of 1:33:8 besides free isopropyl amine, aniline
and the side products/intermediates discussed below. Thus more
isopropyl rich triazacyclohexanes are formed in the mixture upon
addition of the aniline (remaining aniline fragments are found in
Table 1
Comparison of structural parameters (distances (Å), angles (°)) for TAC 1–4 and their symmetrically substituted analogues.a
1
2
3
4
(o-FPh3)TACb
(p-FPh3)TACb
N1–C11
N1–C2
1.422(3)–1.425(3)
1.437(3)–1.478(3)
3.22
4.83
33
1.413(8)–1.420(8)
1.449(3)–1.46(2)
3.25
5.04
39
1.421(3)–1.422(4)
1.451(4)–1.474(3)
3.30
5.33
44
1.420(2)
1.452(2)
3.30
5.16
40
1.412–1.417
1.443–1.479
3.28
6.05
58
1.424
1.442–1.476
3.38
5.03
40
C11–C11A
C14–C14A
Ph–Phc
Tilt angled
4
9
7
1
10
2
Crystal structures of TACs 1–3 were subject of preliminary communications.
a
Numeration as shown in Fig. 1 for TAC 4.
See Ref. [12].
Angle between the mean planes (C11–C16, F7) and (C11A–C16A, F7A).
Average deviation of C12–C16–C16A and C16–C12–C12A from 90°.
b
c
d

S. Latreche et al. / Polyhedron 29 (2010) 1399–1404
1401
iPr
N
Ar
N
Ar
N
Ar
N
K1 = 2
K2 = 0.3
iPr iPr
K3 = 0.01
N
N
N
N
N
N
N
N
iPr
iPr
+ 3 ArNH2
iPr
Ar
Ar
+ 3 iPrNH2
Ar
+ 2 ArNH2
+ iPrNH2
+ ArNH2
+ 2 iPrNH2
Ar = para-C6H4F
Scheme 2. Approximate equilibrium constants for the differently substituted triazacyclohexanes.
the side products) which therefore acts foremost as a catalyst to
reach the equilibrium composition.
The final equilibrium concentrations allow an approximate
determination of equilibrium constants for the possible mixed tri-
azacyclohexanes in CDCl3 as shown in Scheme 2.
A
B
7.42d and 7.07d
150.2
7.57d and 7.36d
H
H
H
H
N
N
+367
N
N
Thus, 4 (or generally aniline versus isopropylamine) is not fa-
voured by thermodynamics in this series. Indeed, directly observed
NMR spectra of the EtOH/water reaction solution in the ratio of
ArNH2:iPrNH2:CH2O of 1:2:2 showed Ar3TAC, Ar2iPrTAC (4), AriPr2
TAC and iPr3TAC (Ar = p-fluoroaniline) in the ratio of 0:3:34:24, be-
sides excess amine, aniline and some side products, before the final
product 4 started to crystallise from the solution. This corresponds
to similar equilibrium constants of K1 = 1.0 and K2 = 0.1 in ethanol/
water. Crystallisation from the reaction solution and vacuum re-
moval of excess isopropyl amine shifted the equilibrium towards
4 and resulted in a product highly enriched in 4 in high yield. Once
all free aniline is used up or removed, solutions of 4 in CDCl3 are
stable.
3.65s
60.0
3.28sep
63.3
1.19d
23.8
F
-123.8
Scheme 4. Characteristic NMR signals for terminal groups in ring-opened inter-
mediates (1H with, hetero nuclei without multiplicity). Assignments confirmed by
HMBC experiments.
while the TAC concentrations change towards their equilibrium.
Other major side products were found to be aminals, mostly
(ArNH)2CH2 and small amounts of ArNHCH2NHiPr. The 1H NMR
resonances of the former (see Supplementary material) had been
confirmed as the major product of the direct reaction of formalde-
hyde with 2 equiv. of p-fluoroaniline [21].
The observed acid catalysis led us to propose the mechanism
shown in Scheme 3 for the exchange of N-substituents in triazacy-
clohexanes for the first aniline. An initial DFT exploration at the
BP86/SVP level indicates much easier ring-opening after proton-
ation with barriers of less than 20 kcal/mol.
2.3. Complex syntheses
Additional support for this mechanism is gained from interme-
diates or side products observed during the formation of 4. A char-
acteristic pair of doublets was found above 7 ppm in 1H NMR
spectra, which were correlated by HMBC experiments to two far
down-field shifted signals in 13C and 15N NMR spectra (150 and
367 ppm, respectively; Scheme 4, Supplementary material). Long
range HMBC correlation experiments link these signals to an iso-
propyl and a methylene group and they were consequently as-
signed to the iminium species A in Scheme 4. The remaining
NMR signals are likely to be similar to the dominant TAC signals
and could not be identified. A second pair of doublets, present in
lower concentrations, was tentatively assigned to iminium species
B carrying a fluorophenyl group. These ring-opened intermediates
are observed in small concentrations throughout the reactions
Additionof 1or 2to CrCl3(THF)3 indichloromethane atroomtem-
perature or of 3 or 4 to anhydrous CrCl3 in hot toluene at 110 °C
yielded the corresponding purple coloured CrCl3 complexes 5–8 in
good to moderate yields. The isolated complexes gave correct ele-
mental analyses, with the exception of 6 which may contain a CH2Cl2
solvent molecule. Analyses of the ether phase used to separate the
unreacted ligand from the formed complex showed only the pres-
ence of unchanged ligand, indicating that TAC isomerisation does
not take place under the conditions employed here. As previously
observed for their (Ar3TAC)CrCl3 analogues, complexes 5–8 proved
to be insoluble in non-coordinating solvents, such as toluene, dichlo-
romethane and ether, even at elevated temperatures, which pre-
vented further purification of 6 and characterizations of 5–8 by 19F
+ H+
N
N
N
N
H
N
N
H
N
N
N
N
+ ArNH2
–H+
N
N
N
N
N
+ H+
NHAr
–H+
N
NHAr
H
N
Ar
–iPrNH2
Scheme 3. Proposed mechanism for the exchange via ring-opened intermediates.

1402
S. Latreche et al. / Polyhedron 29 (2010) 1399–1404
Table 2
redistribution of N-substituents catalysed by free aniline and high
volatility of the propylamines during work-up under vacuum. Un-
der homogeneous conditions alkyl substituents are favoured ther-
modynamically over aromatic substituents in triazacyclohexanes.
The new ligands readily form CrCl3 complexes, which are, however,
insoluble in unpolar solvents and unstable against decomplexation
in polar solvents. Thus, aniline based triazacyclohexanes are much
less suitable for coordination chemistry than their trialkyl
analogues.
Observed and selected calculated IR frequencies in cmꢀ1 for 4 and 8 (intensity, 15N
isotope shift for 8).
Exp of 4
Calc. of 4
Exp of 8
Calc. of 8
517
530
561w
505(28)
527(30)
547(2)
522s
527s
542w
519(17, 5)
529(26, 0)
552(36, 2)
573(2)
582w
619
643w
700w
581(4)
554
573w
639
568(1, 4)
585(7, 4)
327(1, 0)
704(4, 0)
708(4, 0)
621(27)
640(37)
694(5)
697(23)
700(5)
713w
4. Experimental
727
818s
720(8)
813(131)
746w
768w
748(42, 8)
772(8, 7)
792(7, 0)
828(79, 0)
All manipulations of air- and moisture-sensitive compounds
were carried out under an atmosphere of nitrogen using standard
Schlenk-line or glove box techniques. Solvents were dried accord-
ing to standard methods and collected by distillation or dried by
passage through activated aluminum oxide and de-oxygenated
by repeated extraction with nitrogen. CrCl3(THF)3 was prepared
according to literature methods [23]. 1H, 13C, 19F and 15N NMR
spectra were recorded on Bruker MXR-300, ARX-400, Acance-400
or -500 spectrometers and referenced to residual solvent (CHCl3:
d 7.26, CDCl3: d 77.00), to internal TMS when added or external
CFCl3 (0 ppm for 19F) or MeNO2 (381.5 ppm rel. to liq. ammonia).
15N NMR shifts were obtained be 1H–15N HMBC using a pulsed
field gradient for long range coupling of 7 Hz. Elemental analyses
were performed by the Laboratoire d0Analyse Elémentaire (Univer-
sité de Montréal). High resolution electrospray mass spectra were
obtained on a Bruker TOF instrument in dichlormethane solution.
Isotope pattern match the assigned ion and only the monoisotopic
signal is stated.
826s
852w
877
823(36)
866(52)
836s
Fig. 2. DFT calculated structure of 8.
4.1. 1-Pr-3,5-bis(o-C6H4F)-1,3,5-triazacyclohexane (1)
n-Propylamine (3.95 mL, 48 mmol) and o-fluoroaniline (2.32
mL, 24 mmol) were dissolved in ethanol (20 mL). An aqueous solu-
tion of formaldehyde in water (37%, 5.4 mL, 72 mmol) was added
under stirring. The reaction mixture was kept at room temperature
for 2 days. The solution was concentrated to half its volume and
left for another day, after which the remaining solvent was evapo-
rated. Recrystallisation from hexane at ꢀ20 °C yielded 2.6 g (68%)
of colorless microcrystals. Anal. Calc. for C18H21N3F2 (317.38): C,
68.12; H, 6.67; N, 13.24. Found: C, 67.51; H, 6.41; N, 12.93%. 1H
NMR (CDCl3, 300 MHz): d 7.28–6.85 (m, 8H, Ph), 4.73 (s, 2H,
–N(Ar)CH2N(Ar)–), 4.29 (s, 4H, –N(Pr)CH2N(Ar)–), 2.61 (t, 2H,
NCH2Et), 1.54–1.42 (m, 2H, NCH2CH2Me), 0.86 (t, 3H, NCH2CH2Me).
13C NMR (CDCl3, 75 MHz): d 155.6 (d, J = 245 Hz, 2C, CArF), 137.5 (d,
J = 9 Hz, 2C, –NAr), 124.5 (d, J = 4 Hz, 2C, Ar), 122.7 (d, J = 8 Hz, 2C,
Ar), 120.4 (d, J = 3 Hz, 2C, Ar), 115.8 (d, J = 21 Hz, 4C, Ar), 71.2 (d,
J = 4 Hz, 2C, –N(Pr)CH2N(Ar)–), 69.5 (t, J = 3 Hz, 1C, –N(Ar)CH2
N(Ar)–), 54.1 (s, 1C, NCH2Et), 20.9 (s, 1C, NCH2CH2Me), 11.9 (s,
1C, NCH2CH2Me). 19F NMR (CDCl3, 282 MHz): d ꢀ125.1 (m).
NMRorUV–Visspectroscopy. Solubilisationinpolarsolventssuchas
THF or acetonitrile led to decomplexation of the TAC ligand and, de-
spite repeated attempts, only crystals of the TAC ligand or of
CrCl3(THF)3 were obtained in recrystallisation experiments. How-
ever, the limited solubility in dry dichloromethane allowed analysis
of 8 by ESI–MS and showed a weak signal for an ammonium adduct
with the correct high resolution mass and isotope pattern. IR spectra
of both the ligand 4 and the complex 8 showed significant differ-
ences in the region just above 500 cmꢀ1 were rocking vibrations of
the triazacyclohexanes are observed in analogy to those observed
for trioxane at 469 and 521 cmꢀ1 [22]. This assignment can be con-
firmed by DFT calculated IR spectra where significant N involvement
is indicated by the calculated 15N isotope shifts as shown in Table 2.
Such vibrations also involve N–Cr stretching and are therefore an
indication of ligand bonding.
Therefore the structure and bonding of 8 was further investi-
gated by DFT methods which were found to reproduce crystallo-
graphic Cr-ligand distances within 2 pm for trialkylsubstituted
CrCl3 complexes. Fig. 2 shows the calculated structure of 8. The
Cr–N(aniline) bonds (2.18 Å) are significantly longer than the Cr–
NiPr bond (2.15 Å) also when compared to calculated iPr3TACCrCl3
(2.14 Å). This result is supported by the experimental observation
that aniline based triazacyclohexanes undergo ligand substitution
with donor solvent while all-alkyl substituted complexes are inert.
4.2. 1-iPr-3,5-bis(o-C6H4F)-1,3,5-triazacyclohexane (2)
Isopropylamine (4.10 mL, 48 mmol) and ortho-fluoroaniline
(2.27 mL, 24 mmol) were dissolved in ethanol (20 mL). An aqueous
solution of formaldehyde in water (37%, 2.7 mL, 36 mmol) was
added under stirring. The reaction mixture was kept at room tem-
perature for 8 h. The solution was concentrated to half its volume
and left for another day, after which the remaining solvent was
evaporated. Recrystallisation from hexane at ꢀ20 °C yielded
2.78 g (73%) of colorless microcrystals. Anal. Calc. for C18H21N3F2:
C, 68.12; H, 6.67; N, 13.24. Found: C, 68.83; H, 6.20; N, 13.37%.
1H NMR (CDCl3, 300 MHz): d 7.30–6.83 (m, 8H, Ar), 4.70 (s, 2H,
–N(Ar)CH2N(Ar)–), 4.37 (s, 4H, –N(Pr)CH2N(Ar)–), 3.12 (sep.,
3. Conclusions
Unsymmetrical TAC ligands with one alkyl and two aryl substit-
uents can be prepared in good yields employing an excess of the
alkyl amine. Detailed NMR studies have shown that the high yield
is due to favourable crystallisation behaviour coupled with fast

S. Latreche et al. / Polyhedron 29 (2010) 1399–1404
1403
J = 6 Hz, 1H, –CHMe2), 1.07 (d, J = 6 Hz, 6H, –CHMe2). 13C NMR
4.5. {Pr(o-C6H4F)2TAC}CrCl3 (5)
(CDCl3, 75 MHz): d 155.6 (d, J = 245 Hz, 2C, CArF), 137.5 (d, J = 9
Hz, 2C, CArN), 124.4 (d, J = 4 Hz, 2C, Ar), 122.7 (d, J = 8 Hz, 2C, Ar),
120.6 (d, J = 3 Hz, 2C, Ar), 115.8 (d, J = 21 Hz, 2C, Ar), 69.9 (t,
J = 4 Hz, 1C, –N(Ar)CH2N(Ar)–), 68.4 (d, J = 3 Hz, 2C, –N(Pr)CH2-
N(Ar)–), 49.0 (s, 1C, –CHMe2), 20.4 (s, 2C, –CHMe2). 19F NMR
(CDCl3, 282 MHz): d ꢀ125.1 (m).
TAC 1 (41 mg, 0.11 mmol) and CrCl3(THF)3 (35 mg, 0.11 mmol)
were stirred in CH2Cl2 (2 mL) for 20 h. Filtration, repeated wash-
ings with ether and drying in vacuo resulted in 3.0 mg (57%) of a
purple solid. Anal. Calc. for C18H21N3F2CrCl3: C, 45.44; H, 4.45; N,
8.83. Found: C, 45.04; H, 4.3; N, 8.91.
4.3. 1-Pr-3,5-bis(p-C6H4)F-1,3,5-triazacyclohexane (3)
4.6. {iso-Pr(o-C6H4F)2TAC}CrCl3 (6)
n-Propylamine (7.9 mL, 96 mmol) and p-fluoroaniline (4.54 mL,
48 mmol) were dissolved in ethanol (20 mL). An aqueous solution
of formaldehyde in water (37%, 10.8 mL, 144 mmol) was added un-
der stirring. The reaction mixture was kept at room temperature
for 1 days. The resulting precipitate was filtered and dried to yield
the required product. 4.0 g (52%). Anal. Calc. for C18H21N3F2: C,
68.12; H, 6.67; N, 13.24. Found: C, 67.87; H, 6.56; N, 13.06%. 1H
NMR (CDCl3, 300 MHz): d 6.90–6.84 (m, 8H, Ph), 4.63 (s, 2H,
–N(Ar)CH2N(Ar)–), 4.18 (s, 4H, –N(Pr)CH2N(Ar)–), 2.50 (t, J = 7 Hz,
2H, NCH2Et), 1.47 (sex., J = 7 Hz, 2H, NCH2CH2Me), 0.86 (t,
J = 7 Hz, 3H, NCH2CH2Me). 13C NMR (CDCl3, 75 MHZ): d 157.6 (d,
J = 240 Hz, 2C, CArF), 145.8 (s, 2C, CArN), 119.5 (d, J = 8 Hz, 4C, Ar),
115.6 (d, J = 22 Hz, 4C, Ar), 72.0 (s, 2C, –N(Pr)CH2N(Ar)–), 70.3 (s,
1C, –N(Ar)CH2N(Ar)–), 54.1 (s, 1C, NCH2Et), 20.82 (s, 1C, NCH2CH2-
Me), 11.9 (s, 2C, NCH2CH2Me). 19F NMR (CDCl3, 282 MHz): d ꢀ124.6
(m).
TAC 2 (1.8 g, 5.68 mmol) and CrCl3(THF)3 (1.6 g, 4.28 mmol)
were stirred in CH2Cl2 (20 mL) for 20 h. Filtration, repeated wash-
ings with ether and drying in vacuo resulted in 0.98 g (48%) of a
purple solid. Anal. Calc. for C18H21N3F2CrCl3: C, 45.44; H, 4.45; N,
8.83. Found: C, 40.04; H, 4.58; N, 8.54%. (Anal. Calc. for
.
C18H21N3F2CrCl3 CH2Cl2: C, 40.70; H, 4.13; N, 7.49%.)
4.7. {Pr(p-C6H4F)2TAC}CrCl3 (7)
Anhydrous CrCl3 (100 mg, 0.63 mmol) and a catalytic amount of
Zn powder were suspended in toluene (15 mL). A solution of 3
(180 mg, 0.57 mmol) in toluene was added and the reaction mix-
ture was heated to reflux for 12 h. After cooling to room tempera-
ture, the resulting suspension was filtered and the precipitate
washed with ether, to yield after drying 147 mg (54%) of a purple
solid. Anal. Calc. for C18H21N3F2CrCl3: C, 45.44; H, 4.45; N, 8.83.
Found: C, 45.37; H, 4.10; N, 8.33%.
4.4. 1-iPr-3,5-bis(p-C6H4F)-1,3,5-triazacyclohexane (4)
4.4.1. Experiment A
Isopropylamine (4.09 mL, 48 mmol) and p-fluoroaniline
(2.27 mL, 24 mmol) were dissolved in ethanol (20 mL). An aqueous
solution of formaldehyde in water (37%, 2.7 mL, 36 mmol) was
added under stirring. The reaction mixture was kept at room tem-
perature for 8 h. The resulting precipitate was filtered and dried to
yield the required product after recrystallisation from hexane
3.49 g (91%). Anal. Calc. for C18H21N3F2: C, 68.12; H, 6.67; N,
13.24. Found: C, 68.09; H, 6.46; N, 13.10%. 1H NMR (CDCl3, 400
MHZ): d 6.95–6.85 (m, 8H, Ph), 4.58 (s, 2H, –N(Ar)CH2N(Ar)–),
4.24 (s, 4H, –N(iPr)CH2N(Ar)–), 3.00 (sep., J = 7 Hz, 1H, –CHMe2),
1.07 (d, J = 6 Hz, 6H, –CHMe2). 13C NMR (CDCl3, 100 MHz): d
157.6 (d, J = 242 Hz, 2C, CArF), 145.7 (s, 2C, CArN), 119.6 (d,
J = 8 Hz, 4C, Ar), 115.6 (d, J = 22 Hz, 4C, Ar), 70.9 (s, 1C, –N(Ar)CH2-
N(Ar)–), 69.1 (s, 2C, –N(iPr)CH2N(Ar)–), 49.2 (s, 1C, –CHMe2), 18.5
(s, 2C, –CHMe2). 19F NMR (CDCl3, 282 MHz): d ꢀ125.2 (m).
4.8. {iso-Pr(p-C6H4F)2TAC}CrCl3 (8)
Analogous to 7, reaction of TAC 4 (0.45 g, 1.4 mmol) with CrCl3
(0.20 g, 1.3 mmol), yielded 8 in 0.20 g (32%). Anal. Calc. for
C18H21N3F2CrCl3: C, 45.44; H, 4.45; N, 8.83. Found: C, 44.97; H,
4.80; N, 8.63%. A sample was suspended in dry dichloromethane,
filtered and analysed by ESI-MS: m/z 548.1157 (calc. for M +
(NC4H12): 548.1140). The source of the ionising ammonium cation
NC4H12 was in the spectrometer and was confirmed by the obser-
vation of the analogous cation derived from a dichloromethane
solution of known (Octyl)3TACCrCl3 [2] at m/z 654.3988 (calc.
654.3989).
4.9. Computational studies
4.4.2. Experiment B
All calculations have been performed with a development ver-
sion of the ORCA electronic structure program version 2.7-00
[24] in combination with the split-valence (SV) and triple-f valence
(TZV) basis sets developed by the Karlsruhe group that were sup-
plemented with the appropriate polarization functions from the
TurboMole library [25,26].1 For the fitting basis sets in the RI-J
treatment, the ‘def2’ fit bases optimized for the SV and TZV basis
sets were used for DFT calculations, the standard integration grids
of the ORCA package were used. The ORCA implementation of zero
order relativistic correction (ZORA) and of a COSMO solvent model
Isopropylamine (3.5 mL, 41 mmol) and p-fluoroaniline (2.26 g
freshly vacuum-transfered, 20.3 mmol) were dissolved in ethanol
(16 mL). An aqueous solution of formaldehyde in water (37%,
2.3 mL, 31 mmol) was added under stirring. The reaction mixture
was kept at room temperature for 15 h. Even cooling to ꢀ20 °C
for 1 day gave no precipitate. The solvent was removed in vacuo
and trapped in liquid nitrogen and contained iPrNH2 and iPr3TAC
(about 2:1 by weight) besides EtOH and water by NMR. The
remaining oil was analysed by NMR (containing 48 mol%
(25 wt.%) ‘‘ArNH2” (13:8:1 mixture of the aniline and two aniline
like compounds with overlapping aromatic 1H signals and 19F sig-
nals at ꢀ126.9, ꢀ126.4 and ꢀ127.3 ppm, respectively), 0.2 mol%
(0.3 wt.%) Ar3TAC, 41 mol% (61 wt.%) Ar2iPrTAC, 13 mol% (4 wt.%)
AriPr2TAC and 1 mol% (1 wt.%) iPr3TAC) and then recrystallised
from hexane to obtain nearly pure 4 (only significant impurity is
ArNH2 at <10%). Addition of free aniline identified the NMR signals
for ArNH2 but also rearranged the product mixture within minutes
to release free iPrNH2 with a loss of iPr3TAC and AriPr2TAC forming
more Ar2iPrTAC and some Ar3TAC.
with infinite
e was used for all calculations. All calculations were
converged to 10ꢀ7 Eh in the total energy (ORCA keyword TightSCF).
Maximum integration grid 7 was used for chromium. Grimme’s
van der Waals corrections were used. Minima and transition states
were confirmed by numerical frequency calculations.
1
Basis sets were obtained from the ftp server of the Karlsruhe quantum chemistry
group, ftp.chemie.uni-karlsruhe.de/pub/basen.

1404
S. Latreche et al. / Polyhedron 29 (2010) 1399–1404
Table 3
Appendix A. Supplementary data
Crystal data and structure refinement parameters for 4.
Formula
C18H21F2N3
CCDC 743924 contains the supplementary crystallographic data
for 4. Copies of this information may be obtained free of charge
via www.ccdc.cam.ac.uk/conts/retrieving.html or from Cambridge
Crystallographic Data Centre (CCDC), 12 Union Road, Cambridge,
CB2, IEZ, UK; Fax: +44(0)1223-336033; e-mail: deposit@ccdc.cam.
ac.uk. Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.poly.2010.01. 008.
Formula weight; Dcalc
Crystal color and shape
Crystal size (mm)
Crystal System
Space Group
Unit cell dimensions
a (Å)
317.38; 1.288
colorless plate
0.50 ꢁ 0.40 ꢁ 0.20
orthorhombic
Pcmn (62)
6.3167(2)
13.4579(4)
b (Å)
c (Å)
19.2516(5)
1636.69(8); 4
1.288
0.093
672
3.03–28.67/99.1%
20031/2172/0.0459
2172/0/116
0.26 and ꢀ0.21
R1 = 0.053, wR2 = 0.123
R1 = 0.064, wR2 = 0.129
0.039(8)
V (Å3); Z
References
Dcalc (g/cm3)
Absorption coefficient
F(0 0 0)
h (°)/completeness
Reflections collected/unique/Rint
l
(mmꢀ1
)
[1] A.I. Kuramshin, E.V. Asafyeva, R.A. Cherkasov, Russ. J. Org. Chem. 41 (2005)
779.
[2] R.D. Köhn, M. Haufe, G. Kociok-Köhn, S. Grimm, P. Wasserscheid, W. Keim,
Angew. Chem., Int. Ed. 39 (2000) 4337.
[3] R.D. Köhn, D. Smith, M.F. Mahon, M. Prinz, S. Mihan, G. Kociok-Köhn, J.
Organomet. Chem. 683 (2003) 200.
[4] M. Haufe, R.D. Köhn, G. Kociok-Köhn, A.C. Filippou, Inorg. Chem. Commun. 1
(1998) 263.
Data/restraints/parameters
Largest difference in peak and hole (e Åꢀ3
)
Final R indices [I > 2
r
(I)]a
R indices (all data)a
[5] R.D. Köhn, G. Kociok-Köhn, Angew. Chem., Int. Ed. 33 (1994) 1877.
[6] R.D. Köhn, Z. Pan, G. Kociok-Köhn, M.F. Mahon, Dalton Trans. (2002) 2344.
[7] S. Latreche, R. Köhn, unpublished results.
Extinction coefficientb
Goodness-of-fit (GOF) on F2
1.109
c
[8] R.D. Köhn, M. Haufe, G. Kociok-Köhn, J. Am. Chem. Soc. 128 (2006) 10682.
[9] S. Latreche, A. Bouchemma, S. Bouacida, H. Mousser, A. Mousser, Acta
Crystallogr., Sect. E 62 (2006) o4960.
[10] S. Latreche, A. Bouchemma, S. Bouacida, M. Bouhenguel, A. Mousser, Acta
Crystallogr., Sect. E E62 (2006) o4676.
P
P
P
P
2
1=2
a
b
c
R1
¼
kFoj ꢀ jFck= jFoj; wR2 ¼ f ½wðF2o ꢀ Fc2Þ ꢂ= ½wðFo2Þꢂg
:
Fꢃc ¼ kFc½1 þ 0:001 ꢄ x ꢄ F2c ꢄ k3= sinð2hÞꢂꢀ1=4 (k: overall sale factor).
P
2
1=2
GOF = S = ½wðF2o ꢀ F2c Þ ꢂ=ðn ꢀ pÞ
(n: number of reflections, p: number of
parameters).
[11] S. Latreche, A. Bouchemma, S. Bouacida, M. Bouhenguel, A. Mousser, Acta
Crystallogr., Sect. E 62 (2006) o4674.
[12] A. Bouchemma, P.H. McCabe, G.A. Sim, J. Chem. Soc., Perkin Trans. 2 (1989)
583.
4.10. X-ray diffraction studies
[13] T. Dinadayalane, J. Leszczynski, Struct. Chem. 20 (2009) 11.
ˇ
[14] E.C. Lee, D. Kim, P. Jurecka, P. Tarakeshwar, P. Hobza, K.S. Kim, J. Phys. Chem. A
Intensity data were collected at 150 K on a Nonius Kappa CCD
diffractometer, using graphite monochromated Mo K radiation
111 (2007) 3446.
a
[15] S.J. Grimme, J. Comb. Chem. 27 (2006) 1787.
(0.71073 Å). Data were processed using the NONIUS Software [27].
Structure solution, followed by full-matrix least squares refine-
ment was performed using the WINGX-1.70, SHELXS and SHELXL suite
of programs throughout [28,29]. All non-hydrogen atoms were
refined anisotropic, hydrogen atoms were refined on calculated
positions using a riding model. Crystal parameters and details of
the data collection, solution and refinement are summarised in
Table 3.
[16] Unpublished results.
[17] R.D. Köhn, Z. Pan, M.F. Mahon, G. Kociok-Köhn, J. Chem. Soc., Dalton Trans.
(2003) 2269.
[18] S. Mihan, D. Lilge, G. Schweier, R. Köhn, G. Seifert, WO2000058369, 2000.
[19] S. Mihan, D. Lilge, G. Schweier, R. Köhn, G. Seifert, US277823, 2004.
[20] A. Flores-Parra, S.A. Sanchez-Ruiz, Heterocycles 51 (1999) 2079.
[21] E.C. Wagner, J. Org. Chem. 2 (1937) 157.
[22] M. Kobayashi, R. Iwamoto, H. Tadokoro, J. Chem. Phys. 44 (1966) 922.
[23] H. Heiss, W. Herwig, J. Org. Chem. 23 (1958) 1404.
[24] F. Neese, ORCA, an ab initio, Density Functional and Semi-empirical Program
Package, University of Bonn, Germany, 2008.
[25] A. Schäfer, H. Horn, R. Ahlrichs, J. Chem. Phys. 97 (1992) 2571.
[26] A. Schäfer, C. Huber, R. Ahlrichs, J. Chem. Phys. 100 (1994) 5829.
[27] DENZO-SCALEPACK: Z. Otwinowski, W. Minor, Methods in Enzymology, in:
C.W. Carter, R.M. Sweet (Eds.), Macromolecular Crystallography, vol. 276,
Academic Press, 1997, p. 307.
Acknowledgements
The authors thank the University of Bath and the Natural Sci-
ences and Engineering Research Council of Canada for funding.
S.L. was recipient of a Ph.D. stipend from the Algerian government.
[28] L.J. Farrugia, J. Appl. Crystallogr. 32 (1999) 837.
[29] G.M. Sheldrick, Acta Crystallogr. A 64 (2008) 112.