Diastereomeric Δ- 1,4,7,10-tetrakis((R)-2-hydroxy-2- phenylethyl)1,4,7,10-tetraazacyclododecane and its alkali-metal complex ions. A potentiometric titration, nuclear magnetic resonance, and molecular orbital study

Article abstract of DOI:10.1021/ja973367l

¹³C NMR studies are consistent with Δ-1,4,7,10-tetrakis((R)-2- hydroxy-2-phenylethyl)-l,4,7,10tetraazacyclododecane (ΔR-thpec 12) and its eight-coordinate alkali-metal complexes, Δ[M(R-thpec 12)]⁺, existing predominantly as single sq

Full text of DOI:10.1021/ja973367l

2862
J. Am. Chem. Soc. 1998, 120, 2862-2869
Diastereomeric ∆-1,4,7,10-Tetrakis((R)-2-hydroxy-2-phenylethyl)-
1,4,7,10-tetraazacyclododecane and Its Alkali-Metal Complex Ions. A
Potentiometric Titration, Nuclear Magnetic Resonance, and Molecular
Orbital Study
Sonya L. Whitbread,† Peter Valente,‡ Mark A. Buntine,† Philip Clements,†
Stephen F. Lincoln,*,† and Kevin P. Wainwright*,‡
Contribution from the Department of Chemistry, UniVersity of Adelaide, Adelaide, SA 5005, Australia, and
Department of Chemistry, The Flinders UniVersity of South Australia, GPO Box 2100,
Adelaide 5001, Australia
ReceiVed September 25, 1997
Abstract: 13C NMR studies are consistent with ∆-1,4,7,10-tetrakis((R)-2-hydroxy-2-phenylethyl)-1,4,7,10-
tetraazacyclododecane (∆R-thpec12) and its eight-coordinate alkali-metal complexes, ∆[M(R-thpec12)]+, existing
predominantly as single square antiprismatic ∆ diastereomers in dimethylformamide. Molecular orbital
calculations show the parallel square oxygen and nitrogen planes to delineate the square antiprismatic structure
of ∆R-thpec12 and ∆[M(R-thpec12)]+ where the basket defined by the four phenyl groups of ∆R-thpec12
becomes increasingly shallow as the M+ radius increases from Na+ to Cs+ and the four oxygens move further
apart. An intramolecular exchange process, involving double inversion of all four nitrogen centers, occurs in
∆R-thpec12 for which k(298.2 K) ) 46300 ( 1800 s-1, ∆Hq ) 40.8 ( 0.4 kJ mol-1, and ∆Sq ) -18.8 ( 1.7
J K-1 mol-1. In [M(R-thpec12)]+ this process is characterized by k(298.2 K) ) 233 ( 2, 98 ( 1, 4900 (
100, 33500 ( 1000, and 34500 ( 1100 s-1, ∆Hq ) 34.6 ( 0.3, 46.1 ( 0.2, 42.7 ( 0.3, 39.1 ( 0.3, and 38.5
( 0.3 kJ mol-1, and ∆Sq ) -83.5 ( 1.1, -52.2 ( 0.7, -31.1 ( 1.2, -27.2 ( 1.2, and -28.9 ( 1.3 J K-1
mol-1, respectively, when M+ ) Li+, Na+, K+, Rb+, and Cs+. For intermolecular ligand exchange on ∆[M(R-
thpec12)]+, decomplexation is characterized by kd(298.2 K) ) 396 ( 3, 156 ( 3, and 152000 ( 6000 s-1
,
∆Hd ) 46.0 ( 0.3, 62.3 ( 0.5, and 69.8 ( 0.5 kJ mol-1, and ∆Sd ) -40.9 ( 1.0, 6.0 ( 1.9, and 88.4 (
2.1 J K-1 mol-1, respectively, when M+ ) Li+, Na+, and K+. The stability constant, K, of ∆[M(R-thpec12)]+
varies as M+ changes in the sequence Li+ (3.13 ( 0.05), Na+ (4.25 ( 0.05), K+ (4.10 ( 0.05), Rb+ (3.57
( 0.05), Cs+ (3.47 ( 0.05), and Ag+ (8.14 ( 0.03), where the figures in parentheses are log(K/(dm3 mol-1))
determined in dimethylformamide by potentiometric titration at 298.2 K and I ) 0.05 mol dm-3 (NEt4ClO4).
q
q
Introduction
The attachment of a coordinating pendant arm to each
nitrogen of 1,4,7,10-tetraazacyclododecane results in ligands
which form chiral eight-coordinate alkali-metal complex ions.
Thus, achiral thec12 and tmec12 (Figure 1) form Λ and ∆
enantiomeric complex ions in which chirality arises from the
left- and right-handedness of the arrangement of the pendant
arms in the enantiomeric square antiprismatic structures.1-3
Chiral 1,4,7,10-tetrakis((S)-2-hydroxypropyl)-1,4,7,10-tetraaza-
cyclododecane, S-thpc12 (Figure 1), can in principle form two
diastereomeric square antiprismatic eight-coordinate complex
ions where the methyl groups all occupy either equatorial or
axial positions, the ∆ and Λ diastereomers, respectively, but
only the latter was detected in solution.4 Similarly, only the
ΛS-thpc12 free ligand diastereomer, which also possesses a
square antiprismatic arrangement of oxygens and nitrogens, was
† University of Adelaide.
Figure 1. Ligand structures.
‡ The Flinders University of South Australia.
(1) Stephens, A. K. W.; Lincoln, S. F. J. Chem. Soc., Dalton Trans.
1993, 2123-2126.
detected in solution. Such homochirality is unusual in macro-
cyclic ligands and alkali-metal complex ions, and we have now
(2) Whitbread, S. L.; Politis, S.; Stephens, A. K. W.; Lucas, J. B.; Dhillon,
R.; Lincoln, S. F.; Wainwright, K. P. J. Chem. Soc., Dalton Trans. 1996,
1379-1384.
(3) Stephens, A. K. W.; Dhillon, R. S.; Madbak, S. E.; Whitbread, S.
L.; Lincoln, S. F. Inorg. Chem. 1996, 35, 2019-2024.
(4) Dhillon, R. S.; Madbak, S. E.; Ciccone, F. G.; Buntine, M. A.;
Lincoln, S. F.; Wainwright, K. P. J. Am. Chem. Soc. 1997, 119, 6126-
6134.
S0002-7863(97)03367-2 CCC: $15.00 © 1998 American Chemical Society
Published on Web 03/11/1998

∆R-thpec12 and Its Alkali-Metal Complex Ions
J. Am. Chem. Soc., Vol. 120, No. 12, 1998 2863
69.8 (4C), 64.9 (4C), 51.7 (8C). Alternatively, when (R)-styrene oxide
(5.06 g, 42.1 mmol, Aldrich) was added to a solution of 1,4,7,10-
tetrazacyclododecane (1.77 g, 10.3 mmol) in dry ethanol (15 cm3) and
refluxed for 72 h, ∆R-thpec12 precipitated as colorless prisms. After
cooling, ∆R-thpec12 was filtered off, washed once with cold ethanol
(5 cm3), and dried in vacuo. Yield: 2.5 g (37%). 1,4,7,10-Tetraza-
cyclododecane (cyclen) was prepared by a literature method.7 The
sources of the alkali-metal, silver, and tetraethylammonium perchlorates
were as previously described.2,3 KCF3SO3 and its Rb+ and Cs+
analogues were prepared by reacting the stoichiometric amounts of the
metal carbonates (BDH) and CF3SO3H (Fluka) in water and twice
recrystallizing the product from water. All salts were vacuum-dried
at 353-363 K for 48 h, and were stored over P2O5 under vacuum.
(CAUTION: Anhydrous perchlorate salts are powerful oxidants and
should be handled with care.)
Dimethylformamide was the chosen solvent because ∆R-thpec12 and
its complex ions were insufficiently soluble in other common solvents
for our studies. It was purified and dried by literature methods,8 and
stored over Linde 3 Å molecular sieves under nitrogen. Its water
content was below the Karl Fischer detection level of ∼50 ppm.
Dimethylformamide-d7 (99.5% 2H) and chloroform-d (99.8% 2H,
Aldrich) were used as received. Solutions of ∆R-thpec12 and
anhydrous metal perchlorates (or triflates in the cases of K+, Rb+, and
Cs+ for 13C NMR studies as the corresponding perchlorates were
insufficiently soluble to give reasonable resonance intensities) were
7
prepared under dry nitrogen in a glovebox. For the Li experiments,
dimethylformamide solutions were degassed and sealed under vacuum
in 5-mm NMR tubes coaxially mounted in 10-mm NMR tubes
containing either acetone-d6 or D2O to provide the lock signal. For
13C NMR studies, dimethylformamide-d7 and chloroform-d solutions
of ∆R-thpec12 alone or with the appropriate alkali-metal salt were
transferred to tightly stoppered 5-mm NMR tubes. The stabilities of
∆[M(R-thpec12)]+ were sufficiently high for [∆[M(R-thpec12)]+] and
free [M+] and [∆R-thpec12] in the solutions used in the NMR studies
to be those arising from the stoichiometric complexation of M+ by
∆R-thpec12.
Figure 2. Exchange pathways. The ∆- and ΛR-thpec12 diastereomers
are shown on the left-hand side of the figure, and the ∆- and Λ[M(R-
thpec12)]+ diastereomers are shown on the right-hand side of the figure.
extended our studies to ∆-1,4,7,10-tetrakis(R)-2-hydroxy-2-
phenylethyl)-1,4,7,10-tetraazacyclododecane (∆R-thpec12 in
Figure 1) to further explore the effect of chiral pendant arms
on the overall chirality of macrocyclic ligands and their alkali-
metal complex ions. Our solution 13C NMR studies and
molecular orbital calculations are consistent with ∆R-thpec12
existing predominantly as a single ∆ diastereomer where the
four nitrogens and the four oxygens are at the corners of the
opposed parallel square faces of a square antiprism and exchange
of the macrocyclic ring -CH2- between two different environ-
ments occurs through double inversion of all four nitrogen
centers (Figure 2A,B). The same phenomenon is observed for
the ∆[M(R-thpec12)]+ alkali-metal complex ions (Figure 2D,E).
Thus, ∆R-thpec12 possesses a preformed cavity and may be
viewed as a preorganized, but more flexible, complexing agent
akin to the macrobicyclic cryptands, C221 and C222 (Figure
1), which possess relatively rigid preformed cavities.5,6
Stability constants, K, were determined by triplicated potentiometric
titrations using methods similar to those described in the literature.9,10
1
7Li and 13C (broad-band H decoupled) were run at 116.59 and 75.47
MHz, respectively, on a Bruker CXP-300 spectrometer. In the 7Li
experiments 1000-6000 transients were collected in a 8192 database
over a 1000-Hz spectral width, and for the 13C experiments 6000
transients were accumulated in a 8192 data point base over a 3000-Hz
spectral width for each solution prior to Fourier transformation.
Solution temperature was controlled to within (0.3 K using a Bruker
B-VT 1000 temperature controller. The Fourier-transformed spectra
were subjected to complete line shape analysis11 on a VAX 11-780
7
computer to obtain rate data. The temperature dependent Li and 13C
line widths and chemical shifts employed in the complete line shape
analysis were obtained by extrapolation from low temperatures where
no exchange-induced modification occurred. Molecular orbital calcula-
tions were carried out using Gaussian 94 with the LanL2DZ basis set12
on a Silicon Graphics Power Challenge and a Silicon Graphics Indigo2
workstation. These calculations incorporated all electrons for H, C,
Experimental Section
∆-1,4,7,10-Tetrakis((R)-2-hydroxy-2-phenylethyl)-1,4,7,10-tetraaza-
cyclododecane (∆R-thpec12) was prepared by adding (R)-styrene oxide
(5.06 g, 42.1 mmol, Aldrich) to a solution of 1,4,7,10-tetrazacyclodo-
decane (1.77 g, 10.3 mmol) in dry dimethylformamide (15 cm3), and
the solution was refluxed for 5 h. The solvent was removed under
reduced pressure. The resulting oil was dissolved in the minimum
amount of boiling 95% ethanol (ca. 20 cm3) and the solution cooled to
room temperature. The white crystalline product which precipitated
was filtered off and washed with cold ethanol (3 × 5 cm3). Yield:
(4.1 g, 61%). Mp: 194-196 °C. Anal. Calcd for C40H52N4O4: C,
73.59; H, 8.03; N, 8.58. Found: C, 73.8; H, 8.2; N, 8.7. 13C NMR
(CDCl3, 295 K): δ 142.2 (4C), 128.2 (8C), 127.1 (4C), 125.9 (8C),
(7) Richman, J. E.; Atkins, T. J. J. Am. Chem. Soc. 1974, 96, 2268-
2270.
(8) Perrin, D. D.; Armarego, W. L. F.; Perrin, D. R. Purification of
Laboratory Chemicals, 2nd ed.; Pergamon: Oxford, 1980.
(9) Cox, B. G.; Schneider, H; Stroka, J. J. J. Am. Chem. Soc. 1978, 100,
4746-4749.
(10) Cox, B. G.; Garcia-Rosas, J.; Schneider, H. J. Am. Chem. Soc. 1981,
103, 1384-1389.
(11) Lincoln, S. F. Prog. React. Kinet. 1977, 9, 1-91.
(12) Gaussian 94, Revision D. 3: Frisch, M. J.; Trucks, G. W.; Schlegel,
H. B.; Gill, P. M. W.; Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.;
Keith, T.; Petersson, G. A.; Montgomery, J. A.; Raghavachari, K.;
Al-Laham, M. A.; Zakrzewski, V. G.; Ortiz, J. V.; Foresman, J. B.;
Cioslowski, J.; Stefanov, B. B.; Nanayakkara, A.; Challacombe, M.; Peng,
C. Y.; Ayala, P. Y.; Chen, W.; Wong, M. W.; Andres, J. L.; Replogle, E.
S.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Binkley, J. S.; Defrees, D. J.;
Baker, J.; Stewart, J. P.; Head-Gordon, M.; Gonzalez, C.; Pople, J. A.,
Gaussian, Inc., Pittsburgh, PA, 1995.
(5) (a) Lehn, J.-M.; Sauvage, J. P. J. Am. Chem. Soc. 1975, 97, 6700-
6707. (b) Lehn, J.-M. Acc. Chem. Res. 1978, 11, 49-57.
(6) Lehn, J.-M. J. Inclusion Phenom. 1988, 6, 351-397.

2864 J. Am. Chem. Soc., Vol. 120, No. 12, 1998
Whitbread et al.
Table 1. Complex Ion Stabilities in Dimethylformamide at 298.2
K and I ) 0.05 mol dm-3 (NEt4ClO4)
log(K/(dm3 mol-1))
M+
Li+
)
M+
Na+
)
M+
K+
)
M+
Rb+
)
M+
Cs+
)
M+
)
complex
Ag+
∆[M(R-thpec12)]+ a 3.13
4.25
3.76
5.68
3.37
7.93
6.17
4.10
3.63
3.62
1.59
6.66
7.98
3.57
3.56
2.73
1.39
5.35
6.78
3.47
8.14
Λ[M(S-thpc12)]+ b
[M(tmec12)]+ c
[M(thec12)]+ d
[M(C221)]+ e
3.24
3.61
2.99
3.58
3.41 11.3
2.28 13.73
1.23 11.16
3.61 12.41
2.16 10.07
[M(C222)]+ e
a This work. Errors of (0.05 and (0.03 apply to the alkali-metal
and Ag+ complex ions, respectively, and represent 1 standard deviation.
b Reference 4. c Reference 3. d Reference 2. e Reference 5.
Figure 4. Temperature variations of τ for the ∆R-thpec12/∆[M(R-
thpec12)]+ systems in dimethylformamide. (a) Diastereomer exchange
of ∆[Na(R-thpec12)]+, 5000τ. (b) Intermolecular ∆R-thpec12 exchange
on ∆[Na(R-thpec12)]+, 1500τc. (c) Diastereomer exchange of ∆[Li-
(R-thpec12)]+, 100τ. (d) Intermolecular Li+ exchange on ∆[Li(R-
thpec12)]+, 50τc. Data for the solutions in which the equilibrium [Li+]
and [∆Li(R-thpec12)+] were, respectively, 0.0071 and 0.0130, 0.0125
and 0.0110, and 0.0132 and 0.0069 mol dm-3 are represented by circles,
diamonds, and squares, respectively. (e) Intermolecular ∆R-thpec12
exchange on ∆[Li(R-thpec12)]+, 20τc. (f) Diastereomer exchange of
∆[K(R-thpec12)]+, τ. (g) Intermolecular ∆R-thpec12 exchange of
∆[K(R-thpec12)]+, τc. (h) Diastereomer exchange of ∆R-thpec12 in
chloroform-d, τ/10. (i) Diastereomer exchange of ∆[Rb(R-thpec12)]+,
τ/50. (j) Diastereomer exchange of ∆[Cs(R-thpec12)]+, τ/150. (k)
Diastereomer exchange of ∆R-thpec12, τ/300. The concentrations of
∆R-thpec12 alone in (h) and (k), its alkali-metal complex ion alone in
(a), (c), (f), (i), and (j), and each of ∆R-thpec12 and its alkali metal
complex ion in (b), (d), and (e) were 0.10 mol dm-3. The solid lines
represent the best fits of the combined data for each group of solutions
to either eq 1 or its analogue.
Figure 3. Temperature variation of the broad-band 1H-decoupled 13
C
NMR spectrum (75.47 MHz) of 0.10 mol dm-3 ∆R-thpec12 in
dimethylformamide-d7. Experimental temperatures and τ values derived
from complete line shape analyses of the coalescing doublet arising
from the macrocyclic ring carbons, a and b, appear to the left and right
of the figure, respectively. The resonances arising from the pendant
arm >NCH2- (c), -CH(C6H5)OH (d), and -CH(C6H5)OH (e) carbons
are indicated by italic letters.
and that variations of their relative importance dominate the
differences in stabilities among the series of alkali-metal
complex ions formed by the four tetraaza macrocyclic pendant
arm ligands in Table 1. The much higher stabilities of ∆[Ag-
(R-thpec12)]+ and its analogues by comparison with their alkali-
metal analogues, arise from the strong affinity of soft acid14
Ag+ for nitrogen donor atoms.15,16
N, and O, and the valence electrons for the alkali-metal ions together
with their effective core potentials.13
These data contrast with those for the cryptates formed
by 4,7,13,16,21-pentaoxa-1,10-diazabicyclo[8.8.5]tricosane,
[M(C221)]+, for which stability also is at a maximum when
M+ ) Na+ (Table 1). The relatively rigid cavity radius of C221
(110 pm)6 more closely approximates the radius of seven-
coordinate Na+ (112 pm)17 than those of the other M+. This
confers the highest stabilities on [Na(C221)]+ while the flex-
ibility of ∆R-thpec12 results in lower selectivities and stabilities
for ∆[M(R-thpec12)]+ where M+ is eight-coordinate. The 140-
pm cavity radius of C222 most closely approximates that of
eight-coordinate K+ (155 pm), and [K(C222)]+ is the most stable
cryptate in the series.
Results and Discussion
Λ[M(R-thpec12)]+ Stability. The stability constant, K )
[∆[M(R-thpec12)]+]/[M+][∆R-thpec12], varies with M+ in the
sequence Li+ < Na+ > K+ > Rb+ > Cs+ (Table 1). (The free
ligand and the complex ion diastereomer identification is
discussed below.) In earlier studies, the stabilities of the closely
related [M(thec12)]+ and [M(tmec12)]+ varied with M+ and
solvent consistent with their stabilities being dominated by (i)
the solvation energy of M+, (ii) the electron-donating power of
the donor atoms of thec12 and tmec12, and (iii) the ability of
thec12 and tmec12 to assume a conformation that optimizes
bonding with M+. It is probable that these three factors are
the major determinants of stability for ∆[M(R-thpec12)]+ also,
(14) (a) Pearson, R. G. J. Am. Chem. Soc. 1963, 85, 3533-3539. (b)
Pearson, R. G. Coord. Chem. ReV. 1990, 100, 403-425.
(15) Cotton, F. A., Wilkinson, G. AdVanced Inorganic Chemistry, 5th
ed.; Interscience: New York, 1988.
(16) Buschmann, H.-J. Inorg. Chim. Acta 1985, 102, 95-98.
(17) Shannon, R. D. Acta Crystallogr., Sect. A: Cryst. Phys., Diffr.,
Theor. Gen. Crystallogr. 1976, A32, 751-767.
(13) (a) Wadt, W. R.; Hay, P. J. J. Chem. Phys. 1985, 82, 284-298. (b)
Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299-310.

∆R-thpec12 and Its Alkali-Metal Complex Ions
J. Am. Chem. Soc., Vol. 120, No. 12, 1998 2865
Table 2. Parametersa for Diastereomer Exchange and Intermolecular Ligand and Li+ Exchange in Dimethylformamide
∆Hq
∆Sq
kd(298.2 K)
(s-1
∆Hd
(kJ mol-1
∆Sd
(J K-1 mol-1
10-5kc(298.2 K)
q
q
k(298.2 K)
(s-1
species
)
(kJ mol-1
)
(J K-1 mol-1
)
)
)
)
(dm3 mol-1 s-1
)
∆R-thpec12
46300 ( 1800b
16400 ( 400d
233 ( 2e
40.8 ( 0.4
33.1 ( 0.2
34.6 ( 0.3
-18.8 ( 1.7
-53.2 ( 1.1
-83.5 ( 1.1
∆R-thpec12c
∆[Li(R-thpec12)]+
396 ( 3f
391 ( 12g
46.0 ( 0.3
42.6 ( 0.5
62.3 ( 0.5
69.8 ( 0.5
-40.9 ( 1.0
-52.3 ( 2.1
6.0 ( 1.9
5.35
5.27
27.7
47.2
∆[Na(R-thpec12)]+
∆[K(R-thpec12)]+
∆[Rb(R-thpec12)]+
∆[Cs(R-thpec12)]+
98 ( 1h
46.1 ( 0.2
42.7 ( 0.3
39.1 ( 0.3
38.5 ( 0.3
-52.2 ( 0.7
-31.1 ( 1.2
-27.2 ( 1.2
-28.9 ( 1.3
156 ( 3i
4900 ( 100j
33500 ( 1000l
34500 ( 1100m
152000 ( 6000k
88.4 ( 2.1
a Errors represent 1 standard deviation. b Rate constant at coalescence temperature shown in parentheses, kcoal ) 241 ( 3 s-1 (228.8 K). c In
d
e
f
CDCl3. kcoal ) 266 ( 3 s-1 (320.1 K). kcoal ) 184 ( 2 s-1 (293.5 K). kcoal ) 290 ( 2 s-1 (293.5 K). g 7Li NMR, kcoal ) 27.5 ( 0.5 s-1 (260.1
K). kcoal ) 276 ( 2 s-1 (314.8 K). kcoal ) 112 ( 2 s-1 (294.5 K). kcoal ) 245 ( 2 s-1 (256.0 K). kcoal ) 139 ( 1 s-1 (240.3 K). kcoal ) 238
h
i
j
k
l
( 2 s-1 (229.9 K). kcoal ) 239 ( 2 s-1 (228.8 K).
m
Figure 6. Temperature variation of the broad-band 1H-decoupled
Figure 5. Temperature variation of the broad-band 1H-decoupled 13
NMR spectrum (75.47 MHz) of ∆[Na(R-thpec12)]+ (0.10 mol dm-3
C
)
75.47-MHz 13C NMR spectrum of ∆R-thpec12 (0.10 mol dm-3) and
∆[K(R-thpec12)]CF3SO3 (0.10 mol dm-3), respectively, in dimethyl-
formamide-d7. Experimental temperatures and derived τc values appear
to the left and the right of the figure, respectively. At 228.8 K the
resonances are assigned to the macrocyclic ring carbons of ∆R-thpec12,
a* and b*, and ∆[K(R-thpec12)]+, a and b, ∆R-thpec12 >NCH2-, c*,
∆[K(R-thpec12)]+ >NCH2-, c, ∆R-thpec12 -CH(C6H5)OH, d*,
∆[K(R-thpec12)]+ -CH(C6H5)OH, d, ∆R-thpec12 -CH(C6H5)OH, e*,
and ∆[K(R-thpec12)]+ -CH(C6H5)OH, e, for the two downfield
resonances, but the other three resonances are not resolved.
in dimethylformamide-d7. Experimental temperatures and τ values
derived from complete line shape analyses of the coalescing doublet
arising from the macrocyclic ring carbons, a and b, appear to the left
and right of the figure, respectively. The resonances arising from the
pendant arm >NCH2- (c), -CH(C6H5)OH (d), and -CH(C6H5)OH
(e) carbons are indicated by italic letters.
Diastereomer Exchange in ∆R-thpec12. The broad-band
1H-decoupled 13C NMR spectrum of ∆R-thpec12 in dimethyl-
formamide-d7 consists of a set of eight resonances (Figure 3)
under slow exchange conditions on the NMR time scale. This
is consistent with a single diastereomer, identified as ∆R-thpec12
below, being present at detectable concentrations. At 216.3 K,
the macrocyclic ring -CH2- singlet resonances a and b are
observed at 49.65 and 52.46 ppm, the pendant arm >NCH2-
resonance c is observed at 63.9, the -CH(C6H5)OH resonance
d is observed at 69.83 ppm, and the -CH(C6H5)OH resonances
e are observed at 126.70, 127.51, 128.67, and 144.36 ppm. (In
chloroform-d, the analogous frequencies at 216.3 K are 49.62,
51.91, 64.80, 69.60, 126.07, 127.40, 128.47, and 142.03 ppm.)
As the temperature increases, the pendant arm resonances c-e
narrow consistent with a decrease in solution viscosity while
the macrocyclic ring resonances coalesce to a singlet consistent
with an intramolecular exchange of the macrocyclic -CH2-
between two magnetic environments, a and b, entering the fast
exchange regime. Complete line shape analysis11 of the
coalescence of the macrocyclic -CH2- resonances a and b
yields the mean site lifetimes, τ, plotted in Figure 4 and the
rate parameters, derived through eq 1, in Table 2. The
>NCH2-, -CH(C6H5)OH, and -CH(C6H5)OH resonances (c,
d, and e, respectively) are not affected by this exchange process.
k ) 1/τ ) (kBT/h) exp(-∆Hq/RT + ∆Sq/R)
(1)
The intramolecular exchange process shown for the square
antiprismatic conformation of the ∆R-thpec12 diastereomer
shown in Figure 2A,B satisfies the requirements of the 13C NMR

2866 J. Am. Chem. Soc., Vol. 120, No. 12, 1998
Whitbread et al.
Figure 7. View approximately down the C4 axes of the global energy-minimized structures of ∆R-thpec12 (A) and ∆[Na(R-thpec12)]+ (B) and the
K+ and Cs+ analogues (C and D) determined through Gaussian 94 using the LanL2DZ basis set. In the first two structures the oxygens mask the
nitrogens. Hydrogen bonds are shown as broken lines in (A), and bonds to Na+, K+, and Cs+ are not shown in (B), (C), and (D).
spectral temperature variation. This is readily understood if the
square plane delineated by the four nitrogens is considered fixed.
Thus, ∆R-thpec12 has the square plane delineated by the four
hydroxy groups aboVe that delineated by the four nitrogens,
and two environments, a and b, exist for the macrocyclic
-CH2-. Double inversion about all four nitrogens produces
∆R-thpec12 (Figure 2B) in which the square plane delineated
by the four hydroxy groups is below that delineated by the four
nitrogens and exchanges the macrocyclic -CH2- between
environments a and b while the pendant arms exchange between
identical environments, c-e. If exchange between ∆R-thpec12
(Figure 2A,B) and the ΛR-thpec12 (Figure 2C) alternative
diastereomer occurred, the ∆ and Λ diastereomers should each
exhibit eight distinct 13C resonances in the slow exchange
regime. Similar observations were made in chloroform-d
consistent with ∆R-thpec12 being the only diastereomer present
at detectable levels. The intramolecular exchange parameters
appear in Table 2.
at 50.31, 52.32, 65.49, 70.30, 126.67, 127.54, 128.60, and
145.44 ppm, for ∆[Rb(R-thpec12)]+ at 49.98, 52.60, 63.90,
69.79, 126.62, 127.80, 128.68, and 144.33 ppm, and for ∆[Cs-
(R-thpec12)]+ at 49.69, 52.54, 64.02, 69.81, 126.62, 127.80,
128.66, and 144.33 ppm.
The pendant arm resonances of ∆[Na(R-thpec12)]+, c-e,
narrow slightly with an increase in temperature consistent with
a decrease in solution viscosity while the macrocyclic ring
resonances a and b coalesce to a singlet consistent with an
intramolecular exchange of the macrocyclic -CH2- between
two magnetic environments entering the fast exchange regime
(Figure 5). Complete line shape analysis of the coalescence of
the macrocyclic -CH2- resonances a and b yields the mean
site lifetimes, τ, plotted in Figure 3 and the rate parameters in
Table 2 which were derived through eq 1. The >NCH2-,
-CH(C6H5)OH, and -CH(C6H5)OH resonances c, d, and e,
respectively, are not affected by this exchange process. The
spectra of the other four ∆[M(R-thpec12)]+ show similar
temperature dependences, but over different temperature ranges,
consistent with the occurrence of exchange processes similar
to that occurring in ∆[Na(R-thpec12)]+.
The diastereomer exchange process proposed for ∆[M(R-
thpec12)]+ (Figure 2D,E) is consistent with the observed 13C
NMR spectral temperature variations. Thus, ∆[M(R-thpec12)]+
has the square plane delineated by the four hydroxy groups
aboVe that delineated by the four nitrogens, and two environ-
ments, a and b, exist for the macrocyclic -CH2- (Figure 2D).
Double inversion about all four nitrogens produces ∆[M(R-
thpec12)]+ (Figure 2E) in which the square plane delineated
by the four hydroxy groups is below that delineated by the four
nitrogens and exchanges the macrocyclic -CH2- between
environments a and b while the pendant arms exchange between
Diastereomer Exchange in ∆[M(R-thpec12)]+. The broad-
1
band H-decoupled 13C NMR spectrum of ∆[Na(R-thpec12)]+
in dimethylformamide-d7 exhibits a temperature variation
(Figure 5) consistent with the occurrence of a diastereomer
exchange process similar to that observed for ∆R-thpec12, but
occurring at a slower rate. Under slow exchange conditions at
216.3 K, the macrocyclic ring -CH2- singlet resonances a and
b are observed at 49.45 and 52.20 ppm, pendant arm >NCH2-
(c) and -CH(C6H5)OH (d) resonances are observed at 63.16
and 69.76 ppm, respectively, and the -CH(C6H5)OH (e)
resonances are observed at 126.55, 127.55, 128.57, and 145.02
ppm. Similar assignments are made for the resonances observed
at 216.3 K for ∆[Li(R-thpec12)]+ at 49.57, 51.31, 61.91, 69.42,
126.44, 127.63, 128.63, and 144.68 ppm, for ∆[K(R-thpec12)]+

∆R-thpec12 and Its Alkali-Metal Complex Ions
J. Am. Chem. Soc., Vol. 120, No. 12, 1998 2867
identical environments, c-e. If exchange between ∆[M(R-
thpec12)]+ (Figure 2D,E) and the alternative diastereomer,
Λ[M(R-thpec12)]+ (Figure 2F), occurred, each diastereomer
should exhibit eight distinct 13C resonances in the slow exchange
regime. The observation of only one set of eight resonances is
consistent with ∆[M(R-thpec12)]+ alone existing at detectable
concentrations.
Nitrogen double inversion in ∆[Li(R-thpec12)]+ and its Na+
and K+ analogues is greatly slowed by comparison with that in
∆R-thpec12 (Table 2) because of the large negative ∆Sq
characterizing ∆[Li(R-thpec12)]+ and the larger ∆Hq and more
negative ∆Sq characterizing its Na+ and K+ analogues, compared
with these parameters for ∆R-thpec12. While ∆Hq and ∆Sq
for ∆[Rb(R-thpec12)]+ and its Cs+ analogue differ substantially
from those for ∆R-thpec12, there is a compensation effect which
causes the nitrogen double inversion rates of these three systems
to be similar. Each nitrogen double inversion requires the
dissociation of a M-N bond, but further mechanistic detail is
unclear.
Intermolecular Exchange of ∆R-thpec12 on ∆[M(R-
thpec12)]+. The parameters (Table 2) for intermolecular ∆R-
thpec12 exchange on ∆[M(R-thpec12)]+ (eq 2) were determined
kd
∆[M(R-thpec12)]+ + *∆R-thpec12 y
\
kz
d
∆[M(*R-thpec12)]+ + ∆R-thpec12 (2)
from the complete line shape analyses of the temperature
dependent coalescences of the >NCH2- 13C resonances of free
and complexed ∆R-thpec12 for ∆[Li(R-thpec12)]+, and its Na+
and K+ analogues (Figure 6). The temperature variation of the
mean lifetime of ∆R-thpec12 in ∆[M(R-thpec12)]+, τc ) Xcτf/
Xf ) 1/kd ) 1/(kcK) (where τf is the mean lifetime of ∆R-thpec12
in the free state, Xc and Xf are the corresponding mole fractions
for the ∆[M(R-thpec12)]+ solutions, kd is the decomplexation
rate constant, kc is the complexation rate constant, and K is the
∆[M(R-thpec12)]+ formation constant) yields the rate parameters
for the decomplexation of ∆[M(R-thpec12)]+ through an equa-
tion analogous to eq 1.
Figure 8. Side views of the global energy minimized structures of
∆R-thpec12 (A) and ∆[Na(R-thpec12)]+ (B) and the K+ and Cs+
analogues (C and D) determined through Gaussian 94 using the
LanL2DZ basis set. In all four structures either some oxygens or some
nitrogens or both are masked by other atoms. Hydrogen bonds are
shown as broken lines in (A), and bonds to Na+, K+, and Cs+ are not
shown in (B), (C), and (D).
Nevertheless, because M-N bond breaking is necessary for the
exchange between the equivalent forms of ∆[M(R-thpec12)+]
to proceed, it is possible that part of this process is similar to
that preceding decomplexation. Some support for this view is
afforded by the observation that lability toward the first process
follows the M+ sequence K+ > Li+ > Na+ which is the same
as that for the second. Intermolecular exchange of ∆R-thpec12
in ∆[Li(R-thpec12)]+ and its Na+ analogue is moderately faster
than their diastereomer exchange processes, and for the K+
analogue it is 30 times faster. The intermolecular ligand
exchange processes on ∆[Rb(R-thpec12)]+ and its Cs+ analogue
were in the fast exchange regime of the 13C NMR time scale at
216.3 K so that broadened environmentally averaged 13C
resonances were seen.
The combination of a smaller ∆Hdq and a substantial negative
q
∆Sd causes kd for ∆[Li(R-thpec12)]+ to be greater than that
for ∆[Na(R-thpec12)]+, while the much greater kd for ∆[K(R-
thpec12)]+ results from larger ∆Hd and ∆Sd values. The
q
q
q
q
nonsystematic variation of kd, ∆Hd , and ∆Sd with change in
M+ probably indicates that changes in the solvation energy of
M+ and the amount of strain induced in the transition state with
a change in the size of M+ make varying contributions to the
values of the decomplexation rate parameters as M+ varies.
While the decomplexation process involves a series of steps as
octadentate ∆R-thpec12 decomplexes M+, the present data do
not permit the assignment of a particular step as rate-determin-
ing.
A 7Li NMR kinetic study of the intermolecular exchange of
Li+ between the solvated state and ∆[Li(R-thpec12)]+ yielded
the mean lifetimes, τ, of Li+ in ∆[Li(R-thpec12)]+ from a
complete line shape analysis of the coalescence of the 7Li
resonances of solvated Li+ and ∆[Li(R-thpec12)]+. These data
are plotted in Figure 4 from which it is seen that τ is independent
of [Li+solvated], consistent with the operation of a dominant
monomolecular exchange mechanism. The kinetic parameters
derived from these τ data (Table 2) are similar to those for ligand
exchange on ∆[Li(R-thpec12)]+ derived from the 13C NMR data.
The second-order complexation constant, kc, is the product
of the stability constant for the encounter complex (where ∆R-
thpec12 resides in the second coordination sphere in contact
with the first coordination sphere of solvated M+) formed at a
rate close to diffusion control and the first-order rate constant
for the subsequent rate-determining complexation step. Its
variation with M+ is 100-fold smaller than that for kd which
indicates that the decomplexation transition state more closely
resembles solvated M+ and free ∆R-thpec12 than it does
∆[M(R-thpec12)]+. The activation parameters for the decom-
plexation of ∆[M(R-thpec12)]+ and the nitrogen double inver-
sion in ∆[M(R-thpec12)]+ are significantly different, consistent
with the two processes following different reaction paths.
The resonance of Li+
appears 0.30 ppm downfield from
solvated
that of ∆[Li(R-thpec12)]+ at 234.1 K, and their coalescence with
an increase in temperature follows the usual pattern.4
Molecular Orbital Calculations. The approximately cubic
global energy-minimized ∆R-thpec12 structure calculated using

2868 J. Am. Chem. Soc., Vol. 120, No. 12, 1998
Whitbread et al.
Table 3. Parameters Derived from Molecular Orbital Calculations Using the Gaussian 94 LanL2DZ Basis Seta
∆[M(R-thpec12)]+
param
∆R-thpec12
M+ ) Li+
M+ ) Na+
M+ ) K+
M+ ) Rb+
M+ ) Cs+
O-O distance (pm)
272
O1-O2 ) 267
O2-O3 ) 262
O3-O4 ) 268
O4-O1 ) 269
N1-N2 ) 301
N2-N3 ) 302
N3-N4 ) 302
N4-N1 ) 304
O1-N1 ) 296
O2-N2 ) 281
O3-N3 ) 296
O4-N4 ) 290
Li-O1 ) 288
Li-O2 ) 239
Li-O3 ) 292
Li-O4 ) 277
Li-N1 ) 232
Li-N2 ) 227
Li-N3 ) 232
Li-N4 ) 236
b
276
370
408
446
N-N distance (pm)
O-N distance (pm)
M-O distance (pm)
M-N distance (pm)
317
292
311
292
254
254
321
302
282
294
325
306
299
314
328
310
318
338
M-O plane distance (pm)
M-N plane distance (pm)
H-O distanced (pm)
163
128
96
106
188
95
79
215
95
48
245
95
c
97
97
176
H-O distancee (pm)
H1-O4 ) 176
H2-O1 ) 172
H3-O2 ) 178
H4-O3 ) 179
h
193f
395f
439f
483f
twist angle φg (deg)
3.3
2.2
5.4
2.3
1.2
angle between phenyl plane
36.1
j
33.9
23.4
18.6
13.5
and O plane (deg)i
a The globalized minimum energies for ∆R-thpec12 and its Li+, Na+, K+, Rb+, and Cs+ complex ions are -2061.6107, -2069.0531, -2061.7770,
-2089.4281, -2085.1408, and -2081.1503 H, respectively, where 1 H ) 2617.13 kJ mol-1 b The oxygen atoms have no common plane. c The
.
nitrogen atoms have no common plane. d The distance between H and O in hydroxy groups. e The distance between the H and O of adjacent
hydroxy groups. f No hydrogen bonding. g Twist angle φ ) 0° for a cubic structure. h No meaningful φ as C4 symmetry is absent. i An angle of 0°
exists when the phenyl and the O planes are parallel. j No meaningful angle as C4 symmetry is absent.
Gaussian 9412 has C4 symmetry (Figures 7A and 8A and Table
3) and shows near superimposition of the two parallel square
planes delineated by four oxygens and four nitrogens, respec-
tively. The hydroxy protons point toward the adjacent hydroxy
oxygen and are at distances where weak hydrogen bonding
exists.18 The (R)-2-hydroxy-2-phenylethyl arms are oriented
in a clockwise direction when viewed down the C4 axis from
the four-oxygen plane so that the phenyl groups project out from
the structure. This makes the two carbons in each macrocyclic
ethylene linkage inequivalent, in broad agreement with the ∆R-
thpec12 structure proposed on the basis of the 13C NMR studies.
Each pair of ethylene carbons are equivalent to the other pairs,
and a similar equivalence applies for the pendant arms.
symmetry and have chiral characteristics similar to those of ∆R-
thpec12. However, the O-O, N-N, and O-N distances all
increase with the size of the cation, M+, and the M-O and
M-N plane distances decrease and increase, respectively, as
cation size increases from Na+ to Cs+. It is seen in Figures 7
and 8 that the increase in size of the cation causes the structure
of coordinated ∆R-thpec12 to open out substantially. A measure
of this is gained from the changing angle between the plane of
the phenyl rings and the O plane which decreases markedly as
cation size increases. The twist angle between the O and N
planes is small and shows a nonsystematic variation. The M+
is centrally positioned in the ligand cavity, and the (R)-2-
hydroxyphenylethyl arms are oriented in a clockwise direction
so that the phenyl groups project out from the structure and
render the two carbons in each macrocyclic ethylene linkage
inequivalent as found in the 13C NMR studies discussed above.
The similarity of the ∆[Na(R-thpec12)]+ and ∆R-thpec12
dimensions indicates a near optimum fit of Na+ to the ∆R-
thpec12 cavity. The computed ∆[Li(R-thpec12)]+ structure does
not possess C4 symmetry as is shown by the variation in the
Li-O and Li-N distances (Table 3), consistent with Li+ being
too small for an optimum fit into the ∆R-thpec12 cavity. (If
the solution structure of ∆[Li(R-thpec12)]+ is similar to the
computed structure, it appears that a fluxional motion renders
all of the four (R)-2-hydroxy-2-phenylethyl arms equivalent and
all four macrocyclic ring ethylenic moieties equivalent in the
fast exchange limit of the 13C NMR time scale which is much
faster than the nitrogen double inversion process.) A similar
lack of C4 symmetry was reported for the closely related
computed Λ[Li(S-thpc12)]+ structure.4
Four basic structures with either four, three, two adjacent, or
two diagonally opposed pendant arms on the same side of the
tetraaza plane were selected as minimization starting points. A
range of macrocyclic ring conformations were then superim-
posed on these structures to widen the selection of starting
points. In all cases the energy-minimized structure obtained
was that shown in Figures 7A and 8A, consistent with a global
energy minimum being reached. Analogous starting point
structures were used to test that the ∆[M(R-thpec12)]+ structures
shown in Figure 7 and listed in Table 3 were also global
minimized structures. Selected computed interatomic distances
for ∆R-thpec12 and ∆[M(R-thpec12)]+ appear in Table 3.
The computed distorted cubic structure for ∆[Na(R-thpec12)]+
(Figures 7B and 8B) and its K+ (Figures 7C and 8C) and Rb+
and Cs+ (Figures 7D and 8D) analogues also possess C4
(18) Emsley, J. Chem. Soc. ReV. 1980, 9, 91-124.

∆R-thpec12 and Its Alkali-Metal Complex Ions
J. Am. Chem. Soc., Vol. 120, No. 12, 1998 2869
The computed ∆R-thpec12 structure and the five ∆[M(R-
thpec12)]+ structures possess the same macrocyclic ring con-
formation, which is similar to that observed in the X-ray-
determined crystal structures of their [M(thec12)]+ analogues.19-21
In general terms the computed structures of ∆R-thpec12 and
∆[M(R-thpec12)]+ add plausibility to the interpretation of the
13C NMR data above in terms of single ∆R-thpec12 and ∆[M(R-
thpec12)]+ diastereomers.
thpec12 and ∆[M(R-thpec12)]+ with C4 symmetry (except for
∆[Li(R-thpec12)]+ where Li+ is not centrally located in the
ligand cavity) as global energy-minimized structures, support
these deductions. The structure of the eight-coordinated ∆R-
thpec12 in the computed structures opens out so that the basket-
like structure delineated by the four phenyl groups becomes
shallower and wider as M+ moves away from the tetraaza plane
and closer to the tetraoxa plane in the sequence Na+ through
Cs+. These computed structure variations coincide with an
increase in the rate of exchange between equivalent ∆[M(R-
thpec12)]+ diastereomers in the same sequence, and a decrease
in the stabilities of ∆[M(R-thpec12)]+ as M+ changes from Na+
to Cs+. It is planned to test the molecular receptor character-
istics of the adjustable basket-like structures of ∆R-thpec12 and
∆[M(R-thpec12)]+ in future studies.
Conclusion
∆-1,4,7,10-Tetrakis((R)-2-hydroxy-2-phenylethyl)-1,4,7,10-
tetraazacyclododecane (∆R-thpec12) and its eight-coordinate
alkali-metal complexes, ∆[M(R-thpec12)]+, exist predominantly
as single square antiprismatic ∆ diastereomers (Figure 2) that
undergo exchange between equivalent diastereomeric forms in
dimethylformamide. The alternative ΛR-thpec12 and Λ[M(R-
thpec12)]+ diastereomers appear to be less stable and were not
detected. Molecular orbital calculations, which yield ∆R-
Acknowledgment. Funding of this study by the Australian
Research Council, the University of Adelaide, and the Flinders
University of South Australia is gratefully acknowledged, as
are the computing resources provided by the South Australian
Center for Parallel Computing.
(19) Buøen, S.; Dale, J.; Groth, P.; Krane, J. J. Chem. Soc., Chem.
Commun. 1982, 1172-1174.
(20) Groth, P. Acta Chem. Scand. A 1983, 37, 71-74.
(21) Groth, P. Acta Chem. Scand. A 1983, 37, 283-291.
JA973367L