A study of the formation and application of new chiral water soluble lanthanoid benzoates using real-time infrared spectroscopy

Article abstract of DOI:10.1016/j.poly.2010.02.024

A range of water soluble lanthanoid benzoate complexes of composition [Ln(Bz)₃(H₂O)_n] (Ln = La, Gd, Ho and Yb; Bz = 3,5-bis((R)-2,3-dihydroxypropoxy)benzoate and 3,4,5-tris((R)-2,3-dihydroxypropoxy)benzoate) have been prepared by reaction of lanthanoid bicarbonates with three equivalents of the corresponding optically active benzoic acid in water. Application of [Ln(Bz)₃(H₂O)_n] as asymmetric catalysts for epoxide ring opening reactions has been investigated using styrene oxide, showing complete conversion after 20 h, albeit with no significant enantiomeric excess observed. The formation of the lanthanoid complexes and subsequent catalytic conversion of styrene oxide to phenylethane-1,2-diol were monitored using real-time infrared (RTIR) spectroscopy, yielding information about reaction pathways and intermediates.

Full text of DOI:10.1016/j.poly.2010.02.024

Polyhedron 29 (2010) 1764–1770
Contents lists available at ScienceDirect
Polyhedron
journal homepage: www.elsevier.com/locate/poly
A study of the formation and application of new chiral water soluble lanthanoid
benzoates using real-time infrared spectroscopy
Philip C. Andrews a, Benjamin H. Fraser a,b, William J. Gee a, David Giera a,1, Peter C. Junk a,
,
*
Massimiliano Massi a,2, Kellie L. Tuck a
a School of Chemistry, Monash University, Clayton, Victoria 3800, Australia
b The Scripps Research Institute, 10550 North Torrey Pines Road La Jolla, California 92037, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
A range of water soluble lanthanoid benzoate complexes of composition [Ln(Bz)3(H2O)n] (Ln = La, Gd, Ho
and Yb; Bz = 3,5-bis((R)-2,3-dihydroxypropoxy)benzoate and 3,4,5-tris((R)-2,3-dihydroxypropoxy)ben-
zoate) have been prepared by reaction of lanthanoid bicarbonates with three equivalents of the corre-
sponding optically active benzoic acid in water. Application of [Ln(Bz)3(H2O)n] as asymmetric catalysts
for epoxide ring opening reactions has been investigated using styrene oxide, showing complete conver-
sion after 20 h, albeit with no significant enantiomeric excess observed. The formation of the lanthanoid
complexes and subsequent catalytic conversion of styrene oxide to phenylethane-1,2-diol were moni-
tored using real-time infrared (RTIR) spectroscopy, yielding information about reaction pathways and
intermediates.
Received 28 October 2009
Accepted 15 February 2010
Available online 18 February 2010
Keywords:
Lanthanoid
Rare earth
Carboxylates
In situ IR
Ó 2010 Elsevier Ltd. All rights reserved.
Ring-opening
1. Introduction
gands
(e.g.
PyBOX = 2,6-Bis[4-phenyl-2-oxazolinyl]pyridine)
which limit application to non-polar solvents [4]. In searching for
alternative complexes which meet the ‘sustainability’ objectives
of being non-corrosive, of low toxicity and able to be utilized in
an aqueous environment [5], we recently described the first effi-
cient and highly effective use of substituted lanthanoid benzoate
complexes in the kinetic separation of the pure diastereomer of
trans-limonene oxide through selective ring opening of the cis-iso-
mer of a racemic mixture in water [6]. Since lanthanoid benzoates
are generally highly insoluble in water the benzoic acids need to be
modified with polar groups to support aqueous solubility while
bulky groups can prevent polymeric aggregation. Our previous li-
gands, amide functionalized benzoates, while imparting water sol-
ubility were achiral in nature. In advancing this chemistry, we now
describe the synthesis of two new optically active benzoate ligands
incorporating chiral diol side arms, and their use in the formation
of a series of tris-substituted benzoate complexes of La, Gd, Ho and
Yb. Both complex formation and their application as ring opening
catalysts have been studied using real-time infrared (RTIR) spec-
troscopy; an in situ analytical technique which in recent years
has been successfully applied in the study of reaction kinetics [7]
and in elucidating reaction mechanisms [8].
In recent years the metal organic complexes of the lanthanoid
metals have come to be identified and studied as useful and highly
adaptable catalysts in a variety of important organic transforma-
tions including, for instance, hydroelemention, Michael addition,
asymmetric aldol and hetero Diels–Alder reactions [1]. This has
been largely predicated on properties intrinsic to the metal cat-
ions; high Lewis acidity, variations in charge, size and coordination
numbers, as well as the low toxicity and relatively high stability of
the complexes. While many chiral complexes have been prepared
[2], the lability of the ligands and strong oxophilicity of metal cat-
ions has limited their application to use in anhydrous solvents
since water has a tendency to close down the catalysts, either
through restricting coordination of organic substrates at the metal
center or through hydrolysis. The development of complexes
which can be used as water soluble catalysts [1b] has centered al-
most exclusively on the triflates, which are achiral and require the
addition of chiral auxiliaries to support enantioselective reactions.
In terms of epoxide opening reactions lanthanoid complexes are
typically used to facilitate regioselectivity as opposed to enantiose-
lectivity [3]. The few attempts made have focused on lipophilic li-
2. Experimental
* Corresponding author. Tel.: +61 3 9905 4570; fax: +61 3 9905 4597.
E-mail address: peter.junk@sci.monash.edu.au (P.C. Junk).
on exchange from Institut für Anorganische Chemie, Universität Leipzig, 04103
2.1. General procedures
1
Leipzig, Germany.
2
Chemicals were obtained from Sigma–Aldrich and were used as
received. Methyl 3,4,5-trihydroxybenzoate was synthesized
Present address: Department of Applied Chemistry, Curtin University, Western
Australia 6845, Australia.
0277-5387/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.poly.2010.02.024

P.C. Andrews et al. / Polyhedron 29 (2010) 1764–1770
1765
according to a known literature procedure [9]. 1H and 13C NMR
spectra were recorded on a Bruker AM 300 spectrometer and Var-
ian Unity Nova 500 spectrometer. Chemical shifts were recorded
on the d scale and referenced to the solvent. Solid-state IR spectra
were recorded on a Bruker Equinox 55 Infrared Spectrometer fitted
with a Specac Diamond ATR source. Solution RTIR scanning mea-
surements were recorded using a Mettler Toledo ReactIR 10 spec-
trometer fitted with a DiComp probe connected to an MCT detector
with a K6 Conduit and scanning in the region of 4000–650 wave-
numbers at 8 wavenumber resolution. Water content was deter-
mined by TGA using Setaram TG92 with a heating rate of 10 K/
min under argon in a range from 50 to 200 °C. Melting points were
measured on a Stuart Scientific Melting Point Apparatus SMP3 in
an open capillary and are uncorrected. Optical rotation was mea-
sured using a PolAAr 2001 automatic polarimeter at the sodium
D line (587 nm) at 25 °C. Elemental analyses were performed by
Campbell Microanalytical Laboratory, Department of Chemistry,
University of Otago, Dunedin, New Zealand. Lanthanoid metal
analyses were determined by first digesting the accurately
weighed samples with conc. H2SO4 and HNO3. Then the residual
solids were dissolved in minimal water and made into a standard-
ized 100 mL solution followed by complexometric titrations with
standardized 0.0100 M Na2H2EDTA solution with Xylenol Orange
indicator and hexamethylenetetramine buffer. Solubility measure-
ments were obtained by incremental addition of analyte to a
known volume of deionised water each hour until no further disso-
lution was observed. Results shown represent an average of three
repetitions.
1105, 1053, 1030, 1009, 974, 936, 894, 833, 795, 765, 670,
641 cmꢀ1. MS (ESI): m/z = 419.2 [MNa]+.
2.4. Synthesis of methyl 3,4,5-tris(((S)-2,2-dimethyl-1,3-dioxolan-4-
yl)methoxy)benzoate (2b)
The methyl benzoate 2b was obtained using a similar procedure
as in 4.1.2 from 1 and methyl 3,4,5-trihydroxybenzoate (1.61 g,
8.70 mmol). Yield: 1.1 g (23%). M.p. 77.5–78.5 °C. [a]
25 = +8.7°.
D
Anal. Calc. for C26H38O11: C, 59.3; H, 7.3. Found: C, 59.4; H, 7.4%.
1H NMR (500 MHz, CDCl3): d = 1.37 (s, 3H, CH3), 1.39 (s, 6H, CH3),
1.41 (s, 3H, CH3), 1.46 (s, 6H, CH3), 3.89 (s, 3H, OCH3), 3.91–4.03
(m, 6H, 1B-H, 3A-H, 4B-H), 4.10–4.19 (m, 6H, 1A-H, 3B-H, 6A-H),
4.42 (p, 1H, J = 6.0 Hz, 5-H), 4.48 (pd, 2H, J1 = 1.0 Hz J2 = 5.5 Hz, 2-
H), 7.31 (s, 2H, ArH). 13C NMR (50 MHz, DMSO): d 26.2 (2C), 27.4
(2C), 53.1, 66.4, 67.1, 70.6, 74.2, 74.6, 74.9, 109.0, 109.4, 109.7,
125.65, 142.2, 152.7, 166.6. IR (Diamond ATR):
m 2987, 2937,
1717, 1591, 1503, 1429, 1371, 1330, 1248, 1204, 1121, 1079,
1046, 1023, 1002, 978, 938, 895, 838, 763 cmꢀ1. MS (ESI): m/
z = 549.2 [MNa]+.
2.5. Synthesis of 3,5-bis(((S)-2,2-dimethyl-1,3-dioxolan-4-
yl)methoxy)benzoic acid (3a)
A solution of methyl benzoate 2a (1.67 g, 4.2 mmol) in 1:2 mix-
ture of THF and H2O was adjusted to pH 12 using aqueous sodium
hydroxide (0.5 M). This solution was refluxed for 12 h, after which
time the THF was removed and the aqueous remainder adjusted to
pH 2 and extracted with diethyl ether (10 ꢁ 350 mL). The com-
bined organics were combined, dried (MgSO4) and evaporated to
dryness yielding a white powder. Yield: 1.46 g (91%). M.p. 99–
2.2. Synthesis of (R)-(2,2-dimethyl-1,3-dioxolan-4-yl)methyl 4-
methylbenzenesulfonate (1)
101 °C. [a]
25 = +2.7°. Anal. Calc. for C19H26O8: C, 59.7; H, 6.8.
D
Synthesized according to a literature procedure [9]. Yield: 5.6 g
Found C 59.7, H 6.8%. 1H NMR (500 MHz, CDCl3): d = 1.34 (s, 6H,
CH3), 1.40 (s, 6H, CH3), 3.82–3.84 (dd, 2H, J1 = 8.5 Hz J2 = 6.0 Hz,
3A-H), 3.89–.93 (ddd, 2H, J1 = 11.5 Hz J2 = 7.5 Hz J3 = 2.0 Hz, 1B-H),
3.99–4.12 (ddd, 2H, J1 = 9.0 Hz J2 = 5.5 Hz J3 = 2.0 Hz, 3B-H), 4.09–
4.12 (dd, 2H, J1 = 8.5 Hz J2 = 7.0 Hz, 1A-H), 4.42 (p, 2H, J = 6.0 Hz,
2-H), 6.68 (s, 1H, ArH), 7.20 (s, 2H, ArH). 13C NMR (50 MHz, CDCl3):
d = 26.3, 27.4, 66.5, 67.0, 74.5, 106.6, 108.7, 109.8, 133.8, 160.4,
(94%). [
a]
25 = ꢀ5.1° (lit. ꢀ4.8°) [9]. 1H NMR (300 MHz, CDCl3):
D
d = 1.27 (s, 6H, CH3), 2.47 (s, 3H, Ar-CH3), 3.66 (dd, J1 = 8.5 Hz
J2 = 5.5 Hz, 1H, 1A-H), 3.99 (m, 2H, 3B-H, 1B-H), 4.13 (dd,
J1 = 7.0 Hz J2 = 3.5 Hz, 1H, 3B-H), 4.26 (m, 1H, 2H), 7.54 (d,
J = 8.0 Hz, 2H, ArH), 7.82 (d, J = 8.0 Hz, 2H, ArH). 13C NMR
(50 MHz, CDCl3): d = 21.8, 25.3, 26.8, 66.3, 69.6, 73.0, 110.2,
128.2, 130.0, 132.8, 145.2. IR (Diamond ATR):
m
3056, 2997, 2895,
167.7. IR (Diamond ATR): m 2990, 2937, 2878, 1685, 1601, 1446,
1595, 1486, 1373, 1356, 1341, 1266, 1213, 1191, 1168, 1098,
1084, 967, 934, 981, 820, 735, 664 cmꢀ1. MS (ESI): m/z = 309.0
[MNa]+.
1422, 1371, 1309, 1256, 1235, 1210, 1157, 1053, 1028, 943, 834,
766, 732, 671, 642 cmꢀ1. MS (ESI): m/z = 405.2 [MNa]+.
2.6. Synthesis of 3,4,5-tris(((S)-2,2-dimethyl-1,3-dioxolan-4-
2.3. Synthesis of methyl 3,5-bis(((S)-2,2-dimethyl-1,3-dioxolan-4-
yl)methoxy)benzoic acid (3b)
yl)methoxy)benzoate (2a)
The benzoic acid 3b was synthesized by a method identical to
A
solution of tosylate
1
(10.0 g, 35 mmol), methyl 3,5-
that of 3a: Yield: 1.6 g (76%). M.p. 117.0–119.5 °C. [a]
25 = +6.8°.
D
dihydroxybenzoate (1.46 g, 8.70 mmol) and potassium carbonate
(4.83 g, 35 mmol) in DMF (50 mL) was stirred for five days at
100 °C. The mixture was then allowed to cool, and diluted with
diethyl ether (1.5 L). The solution was washed with water
(10 ꢁ 350 mL) and the organic phase dried (MgSO4) and concen-
trated to dryness under reduced pressure. The crude product was
Anal. Calc. for C25H36O11: C, 59.3; H, 7.3. Found: C, 59.2; H, 7.4%.
1H NMR (500 MHz, CDCl3): d = 1.37 (s, 3H, CH3), 1.39 (s, 6H, CH3),
1.41 (s, 3H, CH3), 1.46 (s, 6H, CH3), 3.91–4.03 (m, 6H, 1B-H, 3A-H,
4B-H), 4.10–4.19 (m, 6H, 1A-H, 3B-H, 6A-H), 4.42 (p, 1H, J = 6.0 Hz,
5-H), 4.48 (qd, 2H, J1 = 1.0 Hz J2 = 5.5 Hz, 2-H), 7.31 (s, 2H, ArH).
13C NMR (50 MHz, DMSO): d = 26.2 (2C), 27.4 (2C), 53.1, 66.4,
67.1, 70.6, 74.2, 74.6, 74.9, 109.0, 109.4, 109.7, 125.65, 142.2,
subjected to column chromatography (3:1 hexane:EtOAc) to yield
25
the title compound. Yield: 2.1 g (96%). M.p. 78–79 °C. [
a]
152.7, 166.6. IR (Diamond ATR): m 2987, 2937, 1717, 1591, 1503,
D
+2.1°. Anal. Calc. for C20H28O8: C, 60.6; H, 7.1. Found: C, 60.7; H,
7.1%. 1H NMR (500 MHz, CDCl3): d = 1.40 (s, 6H, CH3), 1.46 (s, 6H,
CH3), 3.87–3.90 (m, 5H, 3A-H, CH3), 3.94–3.97 (ddd, 2H,
J1 = 9.5 Hz J2 = 5.5 Hz J3 = 2.0 Hz, 3B-H), 4.04–4.07 (ddd, 2H,
J1 = 9.5 Hz J2 = 5.5 Hz J3 = 2.0 Hz, 1A-H), 4.14–4.17 (dd, 2H,
J1 = 8.5 Hz J2 = 6.0 Hz, 1B-H), 4.47 (p, 2H, J = 6.0 Hz, 2-H), 6.69 (s,
1H, ArH), 7.19 (s, 2H, ArH) ppm. 13C NMR (50 MHz, CDCl3):
d = 25.6, 27.0, 52.5, 66.9, 69.4, 74.1, 107.1, 108.4, 110.1, 132.3,
1429, 1371, 1330, 1248, 1204, 1121, 1079, 1046, 1023, 1002,
978, 938, 895, 838, 763 cmꢀ1. MS (ESI): m/z = 549.2 [MNa]+.
2.7. Synthesis of 3,5-bis((R)- 2,3-dihydroxypropoxy)benzoic acid
(4aH)
Benzoic acid 3a (0.50 g, 1.31 mmol) was dissolved in a solution
of H2O and trifluoroacetic acid (1:3, 4 mL) and stirred for 2 h. After
this time solvent was removed under high vacuum and the crude
product dissolved in methanol. The methanol was removed under
159.8, 166.8 ppm. IR (Diamond ATR):
m 2985, 2937, 2881, 1714,
1599, 1442, 1411, 1369, 1354, 1304, 1236, 1211, 1177, 1157,

1766
P.C. Andrews et al. / Polyhedron 29 (2010) 1764–1770
high vacuum and the process repeated a second time to ensure
complete removal of trifluoroacetic acid, yielding the title com-
1408, 1334, 1209, 1099, 1035, 865, 761 cmꢀ1. MS (ESI): m/
z = 759.8 [GdL2]+, 1061.9 [GdL3H]+.
pound as
a
viscous colorless oil. Yield: 0.40 g (ꢂ99%).
[a
]
D
25 = +8.2°. Anal. Calc. for C13H18O8: C, 51.7; H, 6.0. Found: C,
2.12. Synthesis of [Ho(4a)3(H2O)2]ꢃ2H2O (5c)
51.4; H, 6.2%. 1H NMR (500 MHz, D2O): d = 3.69 (dq, 4H,
J1 = 12.0 Hz J2 = 6.5 Hz, 3-H), 4.05 (m, 6H, 1-H, 2-H), 6.79 (s, 1H,
ArH), 7.15 (s, 2H, ArH). 13C NMR (50 MHz, DMSO): d = 63.6, 70.8,
Yield: 95%. M.p. 89.4–93.6 °C. [
a
]
25 = ꢀ4.9°. Anal. Calc. for
D
C39H55HoO26ꢃ4H2O: C, 41.1; H, 5.2; Ho, 11.5. Found: C, 40.7; H,
71.0, 106.8, 108.5, 133.7, 160.8, 167.9. IR (Diamond ATR):
m
3302,
4.9; Ho, 10.8%. IR (Diamond ATR): m 3256, 2942, 1569, 1504,
2941, 1697, 1597, 1447, 1301, 1239, 1174, 1118, 1060, 864, 769,
1411, 1335, 1211, 1104, 1035, 929, 868, 761 cmꢀ1. MS (ESI): m/
z = 766.9 [HoL2]+, 1069.0 [HoL3H]+, 1143.2 [HoL3K.(H2O)2]+,
1067.0 [HoL3-H]ꢀ, 1369.2 [HoL4]ꢀ.
633 cmꢀ1. MS (ESI): m/z = 325.2 [MNa]+.
2.8. Synthesis of 3,4,5-tris((R)- 2,3-dihydroxypropoxy)benzoic acid
(4bH)
2.13. Synthesis of [Yb(4a)3(H2O)3]ꢃ2H2O (5d)
Yield: 93%. M.p. 90.0–96.0 °C. [
a
]
25 = ꢀ5.0°. Anal. Calc. for
D
The trisubstituted 4bH was synthesized as colorless viscous oil
by a method identical to that of 4aH. Yield: 0.38 g (ꢂ99%).
C39H55O26Ybꢃ5H2O: C, 40.4; H, 5.3; Yb, 12.8. Found: C, 40.1; H,
4.8; Yb, 12.8%. 1H NMR (500 MHz, D2O): d = 3.40–3.87 (m, 10H,
1-H, 2-H, 3-H), 4.69 (s, 2H, ArH), 6.06 (s, 1H, ArH). IR (Diamond
[a
]
D
25 = ꢀ1.0°. Anal. Calc. for C16H24O11: C, 49.0; H, 6.2. Found: C,
48.8; H, 6.5%. 1H NMR (500 MHz, D2O): d = 3.70 (dq, 6H,
J1 = 18.5 Hz J2 = 4.5 Hz, 1-H, 4-H), 4.05 (m, 9H, 2-H, 3-H, 5-H, 6-
H), 7.22 (s, 2H, ArH). 13C NMR (50 MHz, DMSO): d = 63.6, 63.7,
70.9, 71.5, 71.7, 75.6, 101.8, 126.5, 142.4, 153.1, 167.8. IR (Diamond
ATR):
m 3256, 2942, 1705, 1567, 1503, 1425, 1334, 1209, 1100,
1035, 929, 876, 761 cmꢀ1. MS (ESI): m/z = 775.7 [YbL2]+, 1077.9
[YbL3H]+.
ATR):
m 3192, 2956, 2928, 2871, 1674, 1589, 1505, 1428, 1328,
1209, 1142, 1104, 1047, 1011, 957, 922, 890, 863, 771, 715 cmꢀ1
.
2.14. Synthesis of [La(4b)3(H2O)] (5e)
MS (ESI): m/z = 415.2 [MNa]+.
Yield: 90%. M.p. 59.8–62.9 °C. [
a
]
D
25 = ꢀ9.5°. Anal. Calc. for
C48H69LaO33ꢃH2O: C, 43.3; H, 5.4; La, 13.1. Found. C, 43.0; H, 5.4;
La, 12.2%. 1H NMR (500 MHz, D2O): d = 3.50–3.64 (m, 6H, 1-H, 4-
H), 3.78–3.96 (m, 9H, 2-H, 3-H, 5-H, 6-H), 7.26 (s, 2H, ArH). 13C
NMR (50 MHz, DMSO): d = 63.3 (2C), 70.6, 71.3 (2C), 75.1, 104.3,
2.9. General synthesis of complexes 5a–5h
Benzoic acid 4aH or 4bH (3 equiv) was dissolved in water and
rapidly stirred as LnHCO3 (0.9 equiv) was added in one portion. A
further 0.2 equivalents of LnHCO3 were added by small increments.
Generation of carbon dioxide bubbles and dissolution of the
formed emulsion were indicative of reaction occurring. The solu-
tion was stirred for 12 h after which time the solution was filtered
and the solvent removed under high vacuum yielding Ln com-
plexes 5a–h with yields of 88–95% (Table 1).
107.8, 125.6, 152.2, 166.9. IR (Diamond ATR):
m 3221, 2951,
1534, 1442, 1391, 1334, 1164, 1105, 1046, 930, 862, 782,
676 cmꢀ1. MS (ESI): m/z = 920.8 [LaL2]+, 1312.8 [LaL3H]+.
2.15. Synthesis of [Gd(4b)3(H2O)2]ꢃ2.5H2O (5f)
Yield: 92%. M.p. 51.4–57.2 °C. [
a
]
25 = ꢀ8.4°. Anal. Calc. for
D
C48H69GdO33ꢃ4.5H2O: C, 40.8; H, 5.6; Gd, 13.7. Found: C, 40.6; H,
2.10. Synthesis of [La(4a)3(H2O)2]ꢃ2H2O (5a)
5.6; Gd, 13.1%. IR (Diamond ATR):
m 3256, 2942, 1553, 1466,
1403, 1164, 1112, 1050, 930, 861, 783, 677 cmꢀ1. MS (ESI): m/
z = 939.6 [GdL2]+, 1330.6 [GdL3H]+.
Yield: 93%. M.p. 83.0–87.2 °C. [
a
]
25 = ꢀ5.2°. Anal. Calc. for
D
C39H55LaO26ꢃ4H2O: C, 42.0; H, 5.3; La, 10.0. Found: C, 41.5; H,
5.0; La, 9.6%. 1H NMR (500 MHz, D2O): d = 3.42–3.52 (m, 4H, 3-
H), 3.91 (m, 6H, 1-H, 2-H), 6.60 (s, 1H, ArH), 6.96 (s, 2H, ArH). 13C
NMR (50 MHz, DMSO): d = 62.5, 69.3, 70.1, 105.9, 108.3, 144.2,
2.16. Synthesis of [Ho(4b)3(H2O)3] (5g)
Yield: 90%. M.p. 58.8–64.0 °C. [
a
]
25 = ꢀ9.0°. Anal. Calc. for
D
159.4, 182.9. IR (Diamond ATR):
m 3231, 2942, 1531, 1391, 1214,
C48H69HoO33ꢃ3H2O: C, 41.4; H, 5.4; Ho, 15.4. Found: C, 41.1; H,
1096, 1033, 927, 869, 768, 682 cmꢀ1
. MS (ESI): m/z = 740.7
5.6; Ho, 15.0%. IR (Diamond ATR):
1408, 1329, 1301, 1164, 1107, 1044, 930, 859, 784, 763,
675 cmꢀ1 [HoL3-H]ꢀ,
MS (ESI): m/z = 1337.2 1410.9
[HoL3Cl.(H2O)2]ꢀ.
m 3265, 2938, 1570, 1443,
[LaL2]+, 814.8 [LaL2KOH.H2O]+, 1042.6 [LaL3H]+, 1117.0
[LaL3K.H2O]+.
.
2.11. Synthesis of [Gd(4a)3(H2O)4]ꢃH2O (5b)
2.17. Synthesis of [Yb(4b)3(H2O)3] (5h)
Yield: 88%. M.p. 92.8–95.7 °C. [
a
]
25 = ꢀ5.2°. Anal. Calc. for
D
Yield: 94%. M.p. 62.0–66.5 °C. [
a
]
D
25 = ꢀ9.6°. Anal. Calc. for
C39H55GdO26ꢃ5H2O: C, 40.7; H, 5.3; Gd, 11.1. Found: C, 41.3; H,
5.5; Gd, 11.9%. IR (Diamond ATR):
m 3251, 2941, 1569, 1502,
C48H69YbO33ꢃ3H2O: C, 41.1; H, 5.4; Yb, 15.3. Found: C, 40.9; H,
5.4; Yb, 14.6%. 1H NMR (500 MHz, D2O): d = 3.50–3.64 (m, 6H, 1-
H, 4-H), 3.78–3.96 (m, 9H, 2-H, 3-H, 5-H, 6-H), 4.17 (s, 2H, ArH).
Table 1
IR (Diamond ATR):
m 3265, 2943, 1556, 1406, 1333, 1163, 1111,
Summary of yields and solubilities for lanthanoid complexes 5a–h.
1046, 930, 861, 784, 677 cmꢀ1. MS (ESI): m/z = 1029.9 [YbL2KOH.-
H2O]+, 1347.7 [YbL3H]+, 1421.7 [YbL3Kꢃ(H2O)2]+.
Complex
Formula
Yield (%)
Water solubility (mol Lꢀ1
)
5a
5b
5c
5d
5e
5f
[La(4a)3(H2O)2]
[Gd(4a)3(H2O)4]
[Ho(4a)3(H2O)2]
[Yb(4a)3(H2O)3]
[La(4b)3(H2O)]
[Gd(4b)3(H2O)2]
[Ho(4b)3(H2O)3]
[Yb(4b)3(H2O)3]
93
88
95
93
90
92
90
94
0.16
0.05
0.09
0.04
0.05
0.05
0.08
0.09
2.18. General procedure for catalytic testing
Catalytic epoxide ring opening reactions were undertaken using
lanthanoid complexes 5a–h by the following method. To 5 mL of
deionised water were added styrene oxide (1 mL, 8.76 mmol) and
catalyst (0.044 mol) at room temperature with stirring. After 4 h
5g
5h

P.C. Andrews et al. / Polyhedron 29 (2010) 1764–1770
1767
the reaction was found to be complete. Removal of the solvent un-
der vacuum gave the crude phenylethane-1,2-diol which was puri-
fied by recrystallisation.
[Ln(4a,4b)3(H2O)n] + 3CO2 + 3H2O
Ln(HCO3)3
+
4aH, 4bH
3(
)
Ln = La, [La(4a)3(H2O)2], 5a
4a 5b
Gd, [Gd( )3(H2O)4],
Ho, [Ho(4a)3(H2O)2], 5c
Yb, [Yb(4a)3(H2O)3], 5d
La, [La(4b)3(H2O)], 5e
3. Results and discussion
4b
5f
Gd, [Gd( )3(H2O)2],
3.1. Synthesis of poly diol substituted benzoate ligands
To obtain water soluble chiral ligands, 3,4-dihydroxybenzoic
Ho, [Ho(4b)3(H2O)3], 5g
Yb, [Yb(4b)3(H2O)3], 5h
Scheme 2. Reagents and conditions: (a), H2O, room temp. 8 h.
acid (a-resorcylic acid) and 3,4,5-trihydroxybenzoic acid (gallic
acid) were modified to contain two and three 1,2-diol groups
respectively. The introduction of these polar functionalities is con-
venient for two reasons; the use of d-mannitol, a naturally occur-
ring sugar from the pool of chiral substances, is cheap and the
protection and deprotection reactions through the acetonide deriv-
atives are relatively straightforward and efficient. The overall syn-
thetic route to the ligands is summarized in Scheme 1.
Synthesis of the tosylated precursor 1 was achieved in three
steps in an overall yield of 40% following a literature procedure
[10]. After nucleophilic attack from the phenolic groups of either
methyl 3,4,5-trihydroxybenzoate, or methyl 3,5-dihydroxybenzo-
ate to give the protected ligands 2a–b in yields of 96% and 23%
respectively after four days. The low isolated yield of methyl ben-
zoate 2a is attributed to contamination of the reaction by water,
resulting in early cleavage to the benzoate 3a, a desired product.
However, strict adherence to anhydrous conditions would increase
the isolated yield of 2a if required. A subsequent slow increase of
polarity from 3:1 hexane:ethyl acetate to neat ethyl acetate during
chromatographic workup yielded an additional 53% of product as
the substituted benzoic acid.
completion, evidenced by cessation of CO2 emission. Excess insol-
uble bicarbonate was then removed by filtration prior to workup.
The water solubilities and isolated yields for each of the lanthanoid
complexes is summarized in Table 1.
Considerable time and effort was spent in an attempt to grow
single crystals suitable for X-ray crystallography. Unfortunately
due to the high aqueous solubility of complexes 5a–5h, slow evap-
oration of solutions in alcohols (methanol, ethanol) or water
yielded glassy solids. Layering concentrated alcoholic solutions of
each complex with various organic solvents (e.g. 1-propanol, tert-
butanol, DMF, acetonitrile, diethylether etc.) resulted in eventual
deposition of oily residues within each vial. This result was also ob-
served when hot saturated methanolic solutions of 5a–5h were al-
lowed to slowly cool.
3.3. Spectroscopic analyses
Characterization of the compounds in solution was attempted
by 1H and 13C NMR spectroscopy. However, in the 1H spectra only
the lanthanum and ytterbium complexes gave observable signals
in the region of 2–8 ppm. Paramagnetic up field shifting of the aro-
matic region was observed for ligands in the ytterbium complex,
with the effects less pronounced as the distance from the metal in-
creased. The aromatic resonances were observed for complex 5d at
6.06 and 4.69 ppm, compared with resonances at 6.60 and
6.96 ppm respectively for the corresponding lanthanum complex
5a (Fig. 1, asterisked resonances are those for the protons ortho
to the carboxylate group). In both cases the signals for the aliphatic
diol arms were observed across the region of 3.40–4.00 ppm. The
single aromatic signal for complexes 5e and 5h was observed at
4.17 ppm and 6.22 ppm for the ytterbium and lanthanum com-
plexes respectively, with the aliphatic resonances magnetically
equivalent in the range of 3.50–4.00 ppm. Resonances for the gad-
olinium and holmium complexes were not identifiable in the range
ꢀ300–600 ppm, likely due to obscuration in the baseline by signal
broadening. Attempts to obtain clear and informative 13C NMR
spectra proved unsuccessful for the Yb complexes mainly due to
A dual deprotection performed in basic conditions, followed by
treatment with acid, was found to improve purity of ligands 4aH,
4bH resulting in isolated yields of 91% and 76% respectively. The
ligands were found to be soluble in water with solubilities of
0.46 mol Lꢀ1 for 4aH and 0.38 mol Lꢀ1 for 4bH. Optical purity of
both ligands was monitored throughout the synthetic procedure
and no racemisation was observed.
3.2. Synthesis of lanthanoid tris-benzoate complexes
The tris-substituted lanthanoid benzoate complexes were ob-
tained by reaction of three equiv of 4aH, 4bH with the appropriate
lanthanoid bicarbonate in water, as depicted in Scheme 2.
By using lanthanoid bicarbonate as lanthanoid precursor, the
reactions were high yielding and clean, with only carbon dioxide
and water as by-products. Due to the highly polar nature of the li-
gands and their tendency to retain H2O an exact stoichiometric ra-
tio of reactants is difficult to achieve. As such, the reaction involved
using an initial 0.9 equivalents of the lanthanoid bicarbonate, fol-
lowed by the addition of further 0.2 equivalents in small incre-
ments until the reaction was deemed to have gone to
broadening from paramagnetic Yb3+
.
The water content of each complex was determined by thermo-
gravimetric analysis. The mass losses observed indicated between
R'
R'
R'
R
R
R
R
R
R
OTs
a
b
c
O
O
1
O
O
O
OH
O
OH
O
O
O
O
O
R=
R'=H
O
R=
R'=H
2a
3a
3b
O
O
4aH R= HO
R'=H
HO
HO
O
O
O
O
2b R=R'=
O
R=R'=
O
R=R'=
4bH
HO
Scheme 1. Reagents and conditions: (a), 2a:methyl 3,5-trihydroxybenzoate 2.5eq., 2b:methyl 3,4,5-dihydroxybenzoate 3.5eq., dry DMF, K2CO3, 100 °C, 4d; (b), aq. NaOH,
THF:H2O (1:1), 12 h; (c), TFA, H2O, 2 h.

1768
P.C. Andrews et al. / Polyhedron 29 (2010) 1764–1770
ates by the substituted benzoic acids was studied using RTIR scan-
ning spectroscopy coupled with ConcIRTTM software. This
technique allows for the visualization of starting compounds, re-
agents, intermediates and products present in solution, without
the need for invasive sampling or interruption of the reaction pro-
cess. IR spectra were collected every 15 s in the wavenumber range
of 4000–650 cmꢀ1 at a resolution of 8 cmꢀ1. An iterative compari-
son of accrued spectra by ConcIRTTM allows the appearance or dis-
appearance of the spectrum of any given component to be
displayed. This allows species to be monitored using all observable
vibrations across a broad spectral region, as opposed to following
single isolated bands. To follow the loss of carbon dioxide the
asymmetric absorption for CO2 was monitored in the region
around 2400 cmꢀ1, while the relative changes in the asymmetric
and symmetric stretches, corresponding to the relative concentra-
tions of benzoic/benzoate, were monitored in the region of 1900–
900 cmꢀ1. This resulted in the identification of a reaction pathway
proceeding to the tris-substituted complex via a bis-substituted
intermediate.
At the start of the reaction one equivalent of ytterbium bicar-
bonate was added to a stirring solution of three equivalents of
the ligand 5a in deionised water. Fig. 2 shows that the emission
of CO2 quickly saturates the solution, and thereafter the rate of
evolution slows exponentially until after ca 1 h the reaction is
essentially complete.
To examine complex formation the asymmetric carbonyl
stretch of the bis-substituted benzoic acid 4aH was monitored dur-
ing the course of the reaction. In the free acid this carbonyl stretch
is observed at 1697 cmꢀ1. On deprotonation and complexation
with the metal it is observed in the final tris-substituted complex
at 1563 cmꢀ1. However, a distinct shoulder at 1557 cmꢀ1 was
apparent in the spectrum prior to the complete conversion of the
benzoic acid to benzoate. It was anticipated that this was indica-
tive of a relatively long-lived and stable intermediate. Trends in
the variation of species present in solution were identified using
the software ConcIRTTM, resulting in identification of three compo-
nents across the reaction lifetime, shown in Fig. 3.
Fig. 1. Comparative NMR spectra of lanthanum complex 5a and ytterbium complex
5d displaying up field shifting of aromatic resonances. Asterisked resonances are
those for the protons ortho to the carboxylate group in both compounds. The inset
highlights the upfield shift of the ortho protons influenced by the paramagnetic
Yb3+
.
one and four molecules of water per complex and these are shown
in Table 1. However, there was a slight discrepancy in the water
content of each complex when compared with the microanalytical
data, which indicated additional molecules of water. This is attrib-
uted to the highly hygroscopic nature of the complexes which were
not stored or handled under completely anhydrous conditions
prior to elemental analysis.
Analyses of each complex by electrospray mass spectrometry
identified charged species consistent with the protonated molecu-
lar ion, [M+H]+, with isotopic patterns consistent with each metal
investigated. Evidence of complex fragmentation yielding the
dominant [MꢀL]+ ion was evident for all complexes with the
exception of the holmium complex 5g. Higher order molecular
fragments were sought to determine if polymerization by the dis-
placement of metal bound water by diol arms of adjacent com-
plexes was seen to occur. No evidence of polynuclear species was
observed for any of the complexes.
Deconvolution of the combined spectrum identified two com-
ponent spectra of the free ligand 4aH and the expected tris-substi-
tuted complex 5d. The data compared well with that obtained for
each of the individual species. The unknown intermediate mirrored
most features of the final tris-substituted complex, however, the
asymmetric carbonyl stretch was shifted to a slightly lower wave-
Solid-state IR analyses of the symmetric and asymmetric car-
boxylate stretches for each complex allows for a prediction of the
coordination mode of each benzoate relative to the metal [11].
= ꢀ16 cmꢀ1) (Fig. 4).
The difference between both stretches (Dm) was typically less than
number (Dm
150 cmꢀ1 strongly suggesting the ligands bind to the metal in a
chelating mode, with values closer to ionic bonding (164–
171 cmꢀ1) were observed for gadolinium and holmium complexes
5b and 5c. These results are summarized in Table 2. In all cases
vibrations assigned to scissoring of water molecules were observed
To investigate this intermediate mass spectrometry was con-
ducted on a sample of the reaction mixture shortly after the max-
imum in CO2 emission was observed. This indicated the presence
in the region of 800–500 cmꢀ1
.
3.4. Solution real-time infrared (RTIR) analysis of complex formation
The reaction profile for the formation of the ytterbium complex
5d as a model for the general displacement of lanthanoid bicarbon-
Table 2
Infrared and structural data for complexes 5a–h.
Compound
m
asym(CO2) (cmꢀ1
)
m
sym(CO2) (cmꢀ1
)
D )
(cmꢀ1
5a
5b
5c
5d
5e
5f
1531
1569
1569
1567
1535
1535
1541
1556
1391
1408
1411
1425
1391
1403
1408
1406
140
161
158
142
144
132
133
150
5g
5h
Fig. 2. Generation of carbon dioxide monitored at 2347 cmꢀ1 during the formation
of ytterbium complex 5d.

P.C. Andrews et al. / Polyhedron 29 (2010) 1764–1770
1769
trends could not be identified. This was due to broadening of the
bands within the regions of interest, and an overall decrease in sig-
nal intensity for the lanthanoid species [Ln(4b)3].
3.5. Styrene oxide ring opening reaction with lanthanoid benzoates
To assess the potential for these complexes to act as water-
based asymmetric catalysts the complexes were tested for their
efficacy in the stereoselective conversion of styrene oxide to 1-
phenylethane-1,2-diol. The reaction, here exemplified by lantha-
num complex 5e was followed by RTIR spectroscopy, the result
of which is shown in Fig. 5. A single carbon-oxygen stretch, located
specifically between 861 and 883 cmꢀ1 for the epoxide and 1016–
1043 cmꢀ1 for the diol, was selected and monitored throughout the
reaction. A catalyst loading of 0.5% of 5e was employed in the reac-
Fig. 3. RTIR reaction profile of formation of ytterbium complex 5d as determined by
monitoring the region of 1900–900 cmꢀ1
.
Catalysis of Styrene oxide
of the heteroleptic bis-substituted anionic species, [Yb(4a)2(H-
CO3)–H]ꢀ. It appears that formation of the bis-substituted species
occurs rapidly in solution, producing the initial spike in dissolved
carbon dioxide at 2347 cmꢀ1, seen in Fig. 2. The rate of ligand con-
sumption, seen in Fig. 3 as a negative gradient, is therefore fastest
during formation of the bis-substituted species. Formation of the
tris-substituted complex only begins once the equilibrium concen-
tration of the bis-substituted complex has reached a maximum. At
this point consumption of the free acid slows coincident with the
slower consumption of the bis-substituted complex and formation
of the final tris-substituted complex. This rate reduction would be
expected solely on grounds of steric hindrance. The slightly lower
100
90
80
70
60
50
Phenylethane-1,2-diol
40
Styrene oxide
30
20
10
0
wavenumber for masym(CO2) in the bis complex can be understood
by the marginally greater electrostatic interactions induced by
the bicarbonate anion.
The optimum time for the reaction to go to completion under
these conditions was determined to be 4 h, with the intermediate
formation peaking at 1 h and 20 min.
Real-time IR analysis of complex formation using ligand 4bH
was attempted in an analogous fashion to 5d, however meaningful
Time (hh:mm)
Fig. 5. RTIR monitoring of the conversion of epoxide (C–O stretches at 861 and
883 cmꢀ1) to diol (C–O stretches at 1016–1043 cmꢀ1) catalyzed by lanthanum
complex 5e.
Fig. 4. Comparison of the in situ IR spectrum of known tris-substituted complex (bold) with the unknown intermediate species.

1770
P.C. Andrews et al. / Polyhedron 29 (2010) 1764–1770
3,4,5-tris((R)-2,3-dihydroxypropoxybenzoic acid), and their corre-
OH
O
5e
sponding lanthanoid complexes synthesized, [Ln(Bz)3(H2O)n]
(Ln = La, Gd, Ho, Yb), through reaction with the relevant lanthanoid
bicarbonate. An RTIR study indicated that complex formation oc-
curred in 4 h with the di-substituted benzoic acid and in 6 h with
the trisubstituted benzoic acid. Evidence from analysis of the reac-
tion with Yb suggests that the mechanism involves a metastable
bis-carboxylate complex en route to the final tris-substituted com-
plex. The lanthanoid complexes were all successful as Lewis acid
catalysts in the ring opening of styrene oxide, with total conversion
observed after one hour. Unfortunately, despite the chiral nature of
the ligands no enantioselectivity was observed due, we conclude,
to the high lability of the diol moieties at the metal center in an
aqueous medium.
aq.
OH
Scheme 3. Reaction of styrene oxide with 5e.
Table 3
Observed optical rotation for 1-phenylethane-1,2-diol catalyzed by lanthanoid
complexes 5a–h.
Compound
Optical rotation of product (°)
5a
5b
5c
5d
5e
5f
+4.8
+14.2
+3.1
+4.5
ꢀ2.8
ꢀ0.7
+2.1
+4.7
Acknowledgments
5g
5h
We gratefully acknowledge the Australian Research Council and
Monash University for supporting this work.
tion conducted in deionised water at 25 °C. Total conversion of the
epoxide to the diol was observed within 80 min, with a yield of 98%
(Scheme 3). Repeating the reaction in absence of 5e resulted in no
observable conversion to the diol.
References
[1] (a) H. Furuno, S. Onitsuka, J. Inanaga, J. Synth. Org. Chem. Jpn. 65 (2007) 977.
and references therein;
(b) S. Kobayashi, K. Manabe, Acc. Chem. Res. 35 (2002) 209;
(c) M. Shibasaki, N. Yoshikawa, Chem. Rev. 102 (2002) 2187.
[2] (a) J. Inanaga, H. Furuno, T. Hayano, Chem. Rev. 102 (2002) 2211;
(b) H.C. Aspinall, Chem. Rev. 102 (2002) 1807;
(c) L. Di Bari, P. Salvadori, Coord. Chem. Rev. 249 (2005) 2854.
[3] (a) S.Y. Tosaki, R. Tsuji, T. Ohshima, M. Shibasaki, J. Am. Chem. Soc. 127 (2005)
2147;
(b) M. Martin, S. Bezzenine-Lafollée, R. Gil, J. Collin, Tetrahedron: Asymmetry
18 (2007) 2598.
[4] S.E. Schaus, E.N. Jacobsen, Org. Lett. 2 (7) (2000) 1001.
[5] (a) K. Alimohammadi, Y. Sarrafi, M. Tajbakhsh, Monatsh. Chem. 139 (2008)
1037;
(b) C.J. Li, T.H. Chan, Organic Reactions in Aqueous Media, John Wiley and
Sons, New York, 1997.
[6] P.C. Andrews, M. Blair, B.H. Fraser, P.C. Junk, M. Massi, K.L. Tuck, Tetrahedron:
Asymmetry 17 (2006) 2833.
[7] (a) A. Pintar, J. Batista, J. Levec, Analyst 127 (2002) 1535;
(b) K. Zajsek, A. Gorsek, Rev. Chim. 59 (2008) 1308.
[8] I. Poljansek, B. Likozar, M. Krajnc, J. Appl. Polym. Sci. 106 (2007) 878.
[9] Prepared according to the literature procedure: A. Ashraful, T. Yutaka, I.
Hideyuki, Y. Takashi, T. Sadao, Tetrahedron 61 (2005) 1909.
[10] J.S. Bradshaw, P. Huszthy, C.W. McDaniel, C.Y. Zhu, N.K. Dalley, R.M. Izatt, J.
Org. Chem. 55 (1990) 3129.
[11] G.B. Deacon, R.J. Phillips, Coord. Chem. Rev. 33 (1980) 227.
[12] A. Miyashita, H. Yasuda, H. Takaya, K. Toriumi, T. Ito, T. Souchi, R. Noyori, J. Am.
Chem. Soc. 102 (1980) 7932.
Disappointingly analysis of the optical activity of the products, as
seen in Table 3, was indicative of the catalyst not favoring formation
of either enantiomeric form (lit. [a]
25 = +103.9° for S enantiomer
D
[12], [a]
25 = ꢀ37.8° for R enantiomer [13]). This trend was observed
D
regardless of changes in atomic radius as a result of lanthanoid con-
traction. These results are likely due to the high lability of diol coor-
dination at the metal center. Enantioselectivity is dependent on a
relatively long residence time for the diol moieties at the cationic
metal center, thereby inducing selectivity in the epoxide binding
and the subsequent attack of the OHꢀ nucleophile. It is likely that
H2O coordinates more strongly than either the diol or the epoxide
and so the optimum conditions for inducing selectivity are unfortu-
nately not obtainable in a water medium. In fact, we previously ob-
served that in an analogous complex, [La2(DBA)6(H2O)4]1 DBA =
((R)-4-(2,3-dihydroxypropoxy)benzoate) [14], that polymerization
through diol-metal binding occurs but not until the H2O content is
minimized and the solution heated to 60 °C to favor H2O
displacement.
4. Conclusions
[13] H. Sasai, T. Arai, M. Shibasaki, J. Am. Chem. Soc. 116 (1994) 1571.
[14] P.C. Andrews, B.H. Fraser, P.C. Junk, M. Massi, M. Silberstein, CrystEngComm
(2007) 282.
Two new optically active benzoic acid ligands have been syn-
thesized (3,5-bis((R)-2,3-dihydroxypropoxybenzoic acid and