Supramolecular assemblies from ditopic ligands and transition metal salts

Article abstract of DOI:10.1039/b003436n

Tetradentate salicylaldimine ligands of the H2salen-type bearing ort/io-A'-morpholinomethyl substituents function as ditopic ligands, bonding to Ni" or Cu" with the [N2O2]2" donor set of the salen unit and a sulfate or two nitrate anions with the protonat

Full text of DOI:10.1039/b003436n

View Online / Journal Homepage / Table of Contents for this issue
Supramolecular assemblies from ditopic ligands and transition
metal salts†
Hamish A. Miller, Norman Laing, Simon Parsons, Andrew Parkin, Peter A. Tasker* and
David J. White
Department of Chemistry, The University of Edinburgh, Joseph Black Building,
Kings Buildings, West Mains Road, Edinburgh, UK EH9 3JJ. E-mail: p.a.tasker@ed.ac.uk
Received 28th April 2000, Accepted 7th June 2000
First published as an Advance Article on the web 20th September 2000
Tetradentate salicylaldimine ligands of the H2salen-type bearing ortho-N-morpholinomethyl substituents function
as ditopic ligands, bonding to NiII or CuII with the [N2O2]2Ϫ donor set of the salen unit and a sulfate or two nitrate
anions with the protonated morpholine units. The binding of the metal salt by the zwitterionic form of the ligand
provides a novel approach to the transport of metal sulfates in metal recovery processes. A comparison of the solid
state structures of a “free” ligand with a series of nickel() complexes demonstrates that the metal ion templates the
ligand system, orientating the pendant morpholinium groups to form electrostatic and bifurcated hydrogen bonds in
the sulfate complex to the dianion creating a neutral 1:1:1 [LM2ϩX2Ϫ] complex suitable for extraction into water
immiscible solvents. Other binding modes involving bridging of metal complex units by anion binding to the pendant
morpholine groups suggest that these ditopic ligands could also be used to assemble unusual three-dimensional
arrays of metal complexes in the solid state.
acid than that present in the pregnant leach solution. This pH
Introduction
gradient provides the driving force across the circuit. In the final
Many man-made processes to separate metal ions and recover
stage electrolysis (or “electrowinning”) is used to obtain the
metals from aqueous solution depend on ion exchange pro-
pure metal.
cesses1 in which the desired cation Mnϩ replaces a different
An excellent overall materials balance is obtained because the
metal cation (often Naϩ) or protons from an anionic group in a
acid consumed in leaching the oxide is replaced in the extrac-
complexing agent. Such agents may be deployed on a solid sup-
tion step and the organic extractant, HL, is regenerated by
port in processes using ion exchange resins or in a water immis-
stripping the metal with acid generated in the electrowinning
cible liquid in solvent extraction, eqn. (1). While such processes
process. This feature has underpinned the remarkable growth
in the last decade of solvent extraction technology, which now
nHL ϩ Mnϩ
MLn ϩ nHϩ
(1)
accounts for more than 20% of copper production worldwide.
Whilst ion-exchange extractants are well suited to circuits for
processing oxidic and related “transition” ores they cannot
provide a good materials balance for hydrometallurgical treat-
ment of sulfidic ores or for the recovery of metals from many
secondary sources or effluents such as acid mine drainage
streams. Pressure or bacterial leaching of sulfides does not con-
sume acid (Fig. 2). Consequently, if the conventional types of
acidic reagents are used in the extraction process, sulfuric acid
would build up at the front end of the circuit, requiring neutral-
isation which will generate a salt waste which will have to be
removed to allow recycling of the leachate. This problem could
be overcome using reagents, which would transport the desired
metal as its sulfate salt across a leach/solvent extraction/
electrowin circuit as in Fig. 2. Neutralisation of acid is now only
required to control the pH of the electrolyte at the back end of
the circuit.
Transport of metal salts as in eqn. (2) requires reagents with
an affinity for both the metal cation and its attendant anion.
The systematic development of anion complexing agents3 has
been the subject of much work recently. Selective binding of
anions often involves a suitable arrangement of hydrogen bond
donors such as amides, sulfonamides or (thio)ureas which can
be either neutral or cationic.4 Positively charged anion recep-
tors usually contain protonated nitrogen atoms or metal ions.
A number of researchers have exploited these developments
to obtain organic ligands designed to bind a combination of a
metal cation and its attendant anion,5 either by incorporating
separate cation or anion binding sites into a single ligand or by
are well suited to a number of flowsheets used to recover metals
from both primary and secondary sources, there are situations
when a much more efficient approach is to transport metal
salts. Such an approach requires complexing agents which
accommodate both the metal cation and its attendant anion(s).
This paper addresses the design of ditopic ligands, which can
reversibly transfer metal salts from aqueous feed solutions as in
eqn. (2).
L ϩ Mnϩ ϩ XnϪ
MLX
(2)
A “proton-swing” controlled loading and stripping of
metals, as in eqn. (1), is ideal for the recovery of metals from
oxidic ores using the flowsheet outlined in Fig. 1 for the
recovery of copper from “heap leach/solvent extraction/
electrowinning” operations.2 The crude metal sulfate solution
obtained by initial acid leaching of the ore is contacted with
an organic phase containing the lipophilic ligand, selectively
extracting the metal ion and releasing two protons into the
aqueous phase. A pure metal sulfate solution can then be
obtained by stripping the organic phase with stronger sulfuric
† Based on the presentation given at Dalton Discussion No. 3, 9–11th
September 2000, University of Bologna, Italy.
Electronic supplementary information (ESI) available: rotatable 3-D
crystal structure diagram in CHIME format. See http://www.rsc.org/
suppdata/dt/b0/b003436n/
DOI: 10.1039/b003436n
J. Chem. Soc., Dalton Trans., 2000, 3773–3782
This journal is © The Royal Society of Chemistry 2000
3773

View Online
Fig. 1 The flowsheet and materials balance for recovery of copper from oxidic tailings using sulfuric acid heap leaching, solvent extraction with
phenolic oxime reagents and electrowinning in a conventional tankhouse.
Fig. 2 A proposed flowsheet and materials balance for recovery of metals from sulfidic ores using pressure leaching, solvent extraction with metal
salt transportation and electrowinnning.
arranging for the anion to interact directly with the metal centre
via a vacant co-ordination site.6
Of particular interest to us are those molecules or mixtures
of molecules capable of transporting metal ions and their
attendant anions in liquid/liquid extractions or across mem-
branes. Beer et al. have developed a tripodal tris(amidobenzo-
15-crown-5) ditopic ligand to extract technetium as the
[Na][TcO4] ion pair.7 Kavallieratos et al. enhanced the extrac-
tion of CsNO3 by tetrabenzo-24-crown-8 by the addition of
benzene-1,3,5-tricarboxamide which formed hydrogen bonds to
the nitrate ion.8 Similarly, Unguro and co-workers have used
mixtures of anion and cation binders to transport CsCl and
KCl across membranes. These researchers also covalently
linked anion and cation binding moieties into a single molecule
capable of metal salt transport.9
Scheme 1 Representation of ditopic ligands for metal sulfates with
This paper describes some relatively simple ditopic ligands
diacid/dibasic sites to enable the hydrometallurgical unit operations of
concentration and separation.
derived from H2salen, which have been designed10 to effect the
transport of metal sulfates in a circuit of the type shown in Fig.
2. A key design feature of these ligands is that the metal salt is
bound in a zwitterionic form of the ligand (Scheme 1). This will
permit sequential stripping of the cation and anion as shown,
recycling the ligand and generating a concentrated metal sulfate
solution for electrowinning and ammonium sulfate as a saleable
product. The prototype ligand (Scheme 2) contains an [N2O2]2Ϫ
donor set and two pendant tertiary amine groups (morphol-
ines) designed to capture the two phenolic protons liberated
upon metal binding, thus forming a dicationic binding site for
the anion(s). Selection of different bridging groups (R), alkyl
substituents (RЈ) on the phenol and pendant amine function-
alities allows fine-tuning of the strength and selectivity of com-
plex formation and solubility and phase engagement properties
of the extractant.
In this paper crystal structure determinations of nickel()
complexes of ligand 1 and of a “free” ligand 2 (Scheme 3)
are used to establish the mode of anion binding and to account
for the assembly of extractable metal salt complexes. The impli-
3774
J. Chem. Soc., Dalton Trans., 2000, 3773–3782

View Online
emission spectroscopy (ICP-EAS) analysis was performed on a
Thermo Jarrell Ash IRIS ICP-EAS spectrometer.
Solvent and reagent pre-treatment
Unless stated to the contrary commercial grade chemicals
obtained from Aldrich or Acros companies were used without
further purification.
Ligand synthesis
The 5-alkyl-2-hydroxybenzaldehydes and ethoxymorpholino-
methane were prepared by the methods described by Levin and
co-workers12 and Fenton and co-workers.13 The salicylalde-
hydes substituted with pendant morpholino groups were
prepared by an adaptation of the method of Fenton and co-
workers13 as follows.
5-tert-Butyl-2-hydroxy-3-(morpholinomethyl)benzaldehyde. A
mixture of 5-tert-butylbenzaldehyde (14.8 g, 0.1 mol) and
ethoxymorpholinomethane (15.95 g, 0.11 mol) in acetonitrile
(200 cm3) was heated to reflux under a dinitrogen atmosphere
for 72 h. After cooling to room temperature the solvent was
removed in vacuo to yield a viscous light green oil. The product
was dissolved in CH2Cl2 (150 cm3) and extracted with water
(3 × 60 cm3). The organic fraction was evaporated in vacuo and
dried under vacuum (27.23 g, 98%). Found: C, 69.07; H, 8.54;
N, 5.14. Calc. for C16H23NO3: C, 69.29; H, 8.36; N, 5.05%).
δH (CDCl3, 200 MHz) 1.28 (s, 9 H, C(CH3)3), 2.55 (t, J 4.6, 4 H,
NCH2CH2O), 3.69 (s, 2 H, CH2N), 3.75 (t, J 4.6, 4 H, NCH2-
CH2O), 7.37 (d, 1 H, J 2.6, Ar H), 7.59 (d, 1 H, J 2.6 Hz, Ar H)
and 10.24 (s, 1 H, ArCHO). δC (CDCl3) 31 (CH3), 34 (C(CH3)3),
53 (NCH2CH2O), 59 (CCH2N), 67 (NCH2CH2O), 88 (Ar C),
126 (Ar C), 131 (Ar CH), 133 (Ar CH), 159 (Ar C), 157 (Ar C)
and 193 (ArCHO). MS (FAB, thioglycerol): m/z 278 (MHϩ,
17%).
Scheme 2 Sequestration of a divalent transition metal sulfate.
2-Hydroxy-3-(morpholinomethyl)-5-nonylbenzaldehyde.
A
mixture of 5-nonylbenzaldehyde (28.4 g, 0.1 mol) and ethoxy-
morpholinomethane (15.95 g, 0.11 mol) in acetonitrile (200
cm3) was heated to reflux under a dinitrogen atmosphere for 96
h. After cooling to room temperature the solvent was removed
in vacuo to yield a viscous light green oil. The product was
dissolved in CH2Cl2 (150 cm3) and extracted with water (3 × 60
cm3). The organic fraction was evaporated in vacuo and dried
under vacuum (34.6 g, 99%). The product was used without
further purification (Found: C, 71.95; H, 9.39; N, 4.00. Calc. for
C21H33NO3: C, 72.59; H, 9.57; N, 4.03%). δH (CDCl3, 200 MHz)
0.42–1.70 (m, 19 H, C9H19 mixed isomer chain), 2.54 (m, 4 H,
NCH2CH2O), 3.68 (s, 2 H, CH2N), 3.74 (m, 4 H, NCH2CH2O),
7.35 (d, 1 H, J 2.3, Ar H), 7.51 (d, 1 H, J 2.3 Hz, Ar H) and
10.62 (s, 1 H, ArCHO). δC (CDCl3) 8–52 (C9 mixed isomer
chain), 53 (NCH2CH2O), 59 (CCH2N), 67 (NCH2CH2O), 121
(Ar C), 123 (Ar C), 127 (Ar CH), 135 (Ar CH), 139 (Ar C), 158
(Ar C) and 193 (ArCHO). EIMS: m/z 347 (Mϩ, 26%).
Scheme 3 (ia) MgOMe, MeOH–toluene, reflux 3 h; (ib) paraform-
aldehyde, distillation; (ii) ethoxymorpholinomethane, MeCN, N2, reflux
66 h; (iii) diamine, Et2O–EtOH (1:1), r.t. 24 h.
cations of using a templating role of the metal and the strong
anion binding sites to assemble tailor made three-dimensional
arrays of metal complex units are also considered briefly.
Ligand 1. 5-tert-Butyl-2-hydroxy-3-(morpholinomethyl)benz-
aldehyde (6 g, 21.7 mmol) was dissolved in diethyl ether (60
cm3) and added to a solution of ethane-1,2-diamine (0.636 g,
10.6 mmol) in ethanol (60 cm3). The resulting yellow solution
was stirred overnight then concentrated in vacuo to give a
yellow oil which on trituration in hexane at Ϫ78 ЊC gave a waxy
yellow solid. This was washed with hexane (15 cm3) and diethyl
ether (15 cm3) and dried in vacuo (5.8 g, 95%), mp 155–158 ЊC
(Found: C, 70.60; H, 9.06; N, 9.67. Calc. for C17H25N2O2: C,
70.56; H, 8.71; N, 9.68%). λmax/nm (CH2Cl2) 235, 262, 329 and
413. δH (CDCl3, 200 MHz) 1.28 (s, 9 H, C(CH3)3), 2.50 (t, J 4.6,
4 H, NCH2CH2O), 3.58 (s, 2 H, NCH2CH2N), 3.71 (t, J 4.6, 4
H, NCH2CH2O), 3.90 (s, 2 H, Ar CH2N), 7.14 (d, 1 H, J 2.5,
Experimental
Instrumentation
Nuclear magnetic resonance spectra were obtained on Bruker
AC 200 and AC 250 instruments, FAB mass spectra on a
Kratos MS 50 machine, EI mass spectra on a Kratos Profile
instrument, FTIR spectra on a Perkin-Elmer Paragon 1000
spectrometer as dichloromethane films on NaCl plates or as
KBr discs and electronic absorption spectra on a Unicam UV2
spectrometer. Magnetic susceptibly measurements were carried
out on a Johnson Mathey magnetic susceptibility balance
and molar susceptibilities were corrected for diamagnetism
using Pascal’s constants.11 Inductively coupled plasma atomic
Ar H), 7.37 (d, 1 H, J 2.5 Hz, Ar H), 8.37 (s, 1 H, N᎐CH) and
᎐
13.23 (s, br, 1 H, OH). δC (CDCl3) 31 (CH3), 34 (C(CH3)3), 53
(NCH2CH2O), 57 (CCH2N), 60 (NCH2CH2N), 67 (NCH2-
J. Chem. Soc., Dalton Trans., 2000, 3773–3782
3775

View Online
CH2O), 118 (Ar C), 124 (Ar C), 127 (Ar CH), 131 (Ar CH), 141
(Ar C), 157 (Ar C) and 167 (CHN). MS (FAB, thioglycerol):
m/z 579 (MHϩ, 62%).
N6NiO13: C, 50.11; H, 6.88; N, 10.31%). χm = 2.11 × 10Ϫ10 cm3
molϪ1, µeff = 7.0 × 10Ϫ4 µB. δH (CDCl3, 200 MHz) 1.24 (m, 9 H,
C(CH3)3), 3.21 (br, 4 H, NCH2CH2O), 3.45 (s, 2 H, NCH2-
CH2N), 3.91 (br, 4 H, NCH2CH2O), 4.23 (s, 2 H, Ar CH2N),
7.22 (d, 1 H, J 2.1, Ar H), 7.49 (d, 1 H, J 2.1 Hz, Ar H) and 7.65
Ligand 2. This was prepared similarly from 5-tert-butyl-2-
hydroxy-3-(morpholinomethyl)benzaldehyde and trans-cyclo-
hexane-1,2-diamine. The yellow product was recrystallised
from hexane, collected by filtration and air-dried (1.080 g,
74%), mp 103–106 ЊC (Found: C, 72.99; H, 9.64; N, 7.94. Calc.
for C38H56N4O4ؒ¹C6H14: C, 72.85; H, 9.33; N, 8.29%). λmax/nm
(s, 1 H, N᎐CH). λmax/nm (CH2Cl2) 327, 339, 415 and 460 (sh).
᎐
νmax/cmϪ1 1366vs and 1398vs (NO3). MS (FAB, thioglycerol):
m/z 699 (MHϩ, 4%). Dissolving the material in dichloro-
methane and layering with hexane produced single crystals
suitable for X-ray diffraction.
¯
²
(CH2Cl2) 234, 261, 328 and 417. δH (CDCl3, 200 MHz) 1.23 (s, 9
H, C(CH3)3), 1.66 (m, br, 4 H, cyclohexane CH2), 1.86 (m, br, 4
H, cyclohexane CH2), 2.48 (t, J 4.5, 4 H, NCH2CH2O), 3.30 (m,
1 H, cyclohexane NCH), 3.70 (t, J 4.5, 4 H, NCH2CH2O), 3.54
(s, 2 H, Ar CH2N), 7.06 (d, 1 H, J 2.4, Ar H), 7.33 (d, 1 H, J 2.4
[Cu(1)(SO4)], MXn ؍ CuSO4ؒ5H2O. Recrystallisation from
EtOH–ether gave a dark brown crystalline material form-
ulated as [Cu(1)(SO4)]ؒ5H2O (0.111 g, 43%), mp 268–270 ЊC
(Found: C, 51.96; H, 7.05; N, 6.33. Calc. for C34H60CuN4O13S:
C, 52.44; H, 7.77; N, 6.80%). λmax/nm (CH2Cl2) 371 and 557.
νmax/cmϪ1 1119vs (SO4). MS (FAB, thioglycerol): m/z 734
(MHϩ, 30%).
Hz, Ar H), 8.27 (s, 1 H, N᎐CH) and 13.5 (s, br, 1 H, OH).
᎐
δC (CDCl3) 24.07 (cyclohexane C), 31.23 (CH3), 31.43 (cyclo-
hexane C), 33.70 (C(CH3)3), 53.08 (CCH2N), 53.57 (NCH2-
CH2O), 56.65 (cyclohexane C), 59.58 (cyclohexane C), 66.84
(NCH2CH2O), 117.6 (Ar C), 124.0 (Ar C), 127 (Ar CH), 131
(Ar CH), 141 (Ar C), 157 (Ar C) and 165 (CHN). MS (FAB,
thioglycerol): m/z 633 (MHϩ, 42%).
[Ni(2)(SO4)], MXn ؍ NiSO4ؒ6H2O. Recrystallisation from
diethyl ether gave an orange microcrystalline material form-
ulated as [Ni(2)(SO4)]ؒ5H2O which decomposed at 270–271 ЊC
(0.210 g, 79%) (Found: C, 51.83; H, 7.50; N, 6.23. Calc. for
C38H66N4NiO13S: C, 51.95; H, 7.52; N, 6.38%). δH (CDCl3, 200
MHz) 1.26 (s, 9 H, C(CH3)3), 1.36 (br, 2 H, cyclohexane CH),
1.99 (br, 1 H, cyclohexane CH), 2.50 (br, 1 H, cyclohexane CH),
3.04 (br, 1 H, cyclohexane CH), 2.5–3.5 (br, 4 H, NCH2CH2O),
3.7–4.5 (br, 4 H, NCH2CH2O), 4.20 (m, 2 H, Ar CH2N), 7.19 (d,
1 H, J 2.4, Ar H), 7.26 (d, 1 H, J 2.4 Hz, Ar H) and 7.50 (s, 1 H,
Ligand 3. This was similarly prepared from 2-hydroxy-3-
(morpholinomethyl)-5-nonylbenzaldehyde and benzene-1,2-
diamine. The crude product was isolated as a viscous oil,
dissolved in chloroform (200 cm3) and extracted with water
(3 × 100 cm3). The chloroform solution of 3 was isolated, evap-
orated to dryness and the product dried in vacuo to yield a
highly viscous yellow oil (3.07 g, 82%). The product was used
without further purification in solvent extraction experiments
(Found: C, 67.19; H, 8.25; N, 6.90. Calc. for C48H70N4O4ؒ
CHCl3: C, 66.39; H, 8.07; N, 6.32%). δH (CDCl3, 200 MHz)
0.5–1.69 (m, 19 H, C9H19 mixed isomer chain), 2.53 (m, 4 H,
NCH2CH2O), 3.72 (m, 6 H, OCH2CH2N ϩ Ar CH2N), 6.77 (m,
1 H, Ar H), 7.07 (m, 1 H, Ar H), 7.24–7.37 (m, 2 H, Ar H), 8.65
N᎐CH). λmax/nm (CH2Cl2) 329, 345, 417 and 460 (sh). νmax/cmϪ1
᎐
1120vs (SO4). MS (FAB, thioglycerol): m/z 787 (MHϩ, 20%).
Diffusion of diethyl ether vapours into a saturated dichloro-
methane solution produced red plate crystals suitable for X-ray
diffraction.
[Ni(1 ؊ 2H)]. This was prepared by using Ni(CH3CO2)2ؒ
4H2O as the source of nickel in the procedure outlined above.
Recrystallisation of the crude product from diethyl ether gave
an orange-red microcrystalline material formulated as [Ni(1 Ϫ
2H)]ؒ2H2O (0.167 g, 73%), mp 235–240 ЊC (Found: C, 61.00;
H, 7.38; N, 8.19. Calc. for C34H52N4NiO4: C, 60.77; H, 7.75;
N, 8.34%). χm = 2.47 × 10Ϫ10 cm3 molϪ1, µeff = 7.6 × 10Ϫ4 µB.
δH (CDCl3, 200 MHz) 1.24 (m, 9 H, C(CH3)3), 2.53 (t, 4 H,
J 4.4, NCH2CH2O), 3.35 (s, 2 H, NCH2CH2N), 3.61 (s, 2 H,
Ar CH2N), 3.73 (t, 4 H, J 4.4, NCH2CH2O), 6.89 (d, 1 H, J 2.6,
(s, 1 H, CH᎐N) and 13.50 (br, 1 H, OH). δ (CDCl ) 10–63 (C
᎐
C
3
9
mixed isomer chain), 52.8 (NCH2CH2O), 66.60 (NCH2CH2O),
66.70 (CCH2N), 110.4 (Ar C), 114.6 (Ar C), 115.5 (Ar C), 118.1
(Ar C), 118.5 (Ar CH), 119.0 (Ar CH), 119.6 (Ar CH), 122.2
(Ar CH), 123.5 (Ar C), 127.2 (Ar CH), 127.7 (Ar CH), 153.6
(Ar C) and 157.3 (CHN). MS (FAB, thioglycerol): m/z 768
(MHϩ, 90%).
Preparation of complexes of metal salts
A solution of ligand (0.3 mM) in methanol (20 cm3) was stirred
together with a solution of the appropriate metal salt MXn
(1 mol equivalent) in methanol (20 cm3) overnight. Colour
changes due to complex formation were instantaneous. After
removal of the solvent in vacuo the products were recrystallised
as indicated and air-dried.
Ar H), 7.37 (d, 1 H, J 2.6 Hz, Ar H) and 7.46 (s, 1 H, N᎐CH).
᎐
λmax/nm (CH2Cl2) 324, 346, 417 and 460 (sh). MS (FAB,
thioglycerol): m/z 636 (MHϩ, 50%). Diffusing hexane vapours
into an ethyl acetate solution of the complex grew single
crystals suitable for X-ray diffraction.
Solvent extraction studies
[Ni(1)(SO4)], MXn ؍ NiSO4ؒ6H2O. Recrystallisation from
MeOH–water 3:1 gave a red microcrystalline material form-
ulated as [Ni(1)(SO4)]ؒ6H2O (0.59 g, 93.6%), mp 235–240 ЊC
(Found: C, 48.86; H, 7.37; N, 6.53. Calc. for C34H62N4NiO14S:
A 0.01 M chloroform solution of ligand 3 (10 cm3) was intim-
ately mixed with a 1 M aqueous solution of CuSO4 (10 cm3) at
room temperature for 24 h to allow equilibration by which time
the solution had turned dark brown. The mixture was allowed
to separate and a sample (1 cm3) of the chloroform solution was
removed for ICP-EAS analysis to determine the copper and
sulfur content.
C, 48.53; H, 7.43; N, 6.66%). χm = 1.14 × 10Ϫ9 cm3 molϪ1, µeff
=
1.7 × 10Ϫ3 µB. δH (CDCl3, 200 MHz) 1.27 (m, 9 H, C(CH3)3),
2.6–3.3 (br, 4 H, NCH2CH2O), 3.49 (s, 2 H, NCH2CH2N), 3.7–
4.2 (br, 4 H, NCH2CH2O), 4.3 (s, 2 H, Ar CH2N), 7.21 (d, 1 H,
J 2.5, Ar H), 7.29 (d, 1 H, J 2.5 Hz, Ar H) and 7.67 (s, 1 H,
Collection and reduction of X-ray data
N᎐CH). MS (FAB, thioglycerol): m/z 733 (MHϩ, 10%). λmax/nm
᎐
(CH2Cl2) 328, 339, 418 and 460 (sh). νmax/cmϪ1 1119vs (SO4).
Crystals suitable for X-ray diffraction were obtained by evapor-
ation of a saturated methanol solution of the complex.
Data were collected at 220 K on a Stoe 4-circle diffractometer
using Mo-Kα radiation (λ = 0.71073 Å). The structures were
solved by direct methods (SHELXS 97)14 and refined on F2 by
full matrix least squares (SHELXL 97).15 The quaternary
ammonium hydrogen atoms in the structures of [Ni(1)(SO4)],
[Ni(1)(NO3)2] and [Ni(2)(SO4)] were located on the difference
map and fully refined. All other hydrogen atoms were placed at
calculated positions and refined using riding. All non-H atoms
[Ni(1)(NO3)2], MXn ؍ Ni(NO3)2ؒ6H2O. Recrystallisation
from diethyl ether gave an orange-red microcrystalline powder
formulated as [Ni(1)(NO3)2]ؒ3H2O (0.253 g, 91%), mp 200–
204 ЊC (Found: C, 49.79; H, 6.35; N, 10.76. Calc. for C34H56-
3776
J. Chem. Soc., Dalton Trans., 2000, 3773–3782

View Online
Table 1 Crystallographic data
2
[Ni(1 Ϫ 2H)]
[Ni(1)(SO4)]
[Ni(1)(NO3)2]
[Ni(2)(SO4)]
Empirical formula
M
C41H63N4O4
675.95
C34H48N4NiO4
635.47
C37.75H65N4NiO11.75
518.14
S
C35H52Cl2N6NiO10
846.44
C80H116ClN8Ni2O16S2
2052.75
Crystallographic system
Space group
Monoclinic
C2/c
Triclinic
P1
Monoclinic
P2/c
Monoclinic
P21/c
Triclinic
P1
¯
¯
a/Å
b/Å
c/Å
24.786(3)
12.9082(15)
26.542(4)
91.715(18)
8488(2)
7.3052(5)
15.1256(14)
15.6704(14)
87.161(5)
1631.9(2)
2
18.9754(4)
12.8666(3)
19.7197(4)
118.571(1)
4228.2(2)
4
16.5599(16)
14.2742(14)
16.7199(16)
91.773(7)
3950.3(7)
4
11.108(5)
14.319(7)
16.230(11)
107.95(3)
2368(2)
β/Њ
U/Å3
Z
8
1
µ(Mo-Kα)/mmϪ1
0.531
0.637
0.571
0.689
0.456
Independent reflections (Rint
No. parameters
R
)
7339 (0.0593)
434
0.0720
6557 (0.1733)
385
0.0623
8643 (0.07)
434
0.0502
6972 (0.0664)
506
0.0519
6928 (0.1161)
470
0.1119
wR2
0.2059
0.1362
0.1457
0.1189
0.3271
Largest difference peak
0.237 and Ϫ0.209
0.494 and Ϫ0.438
1.234 and Ϫ0.669
0.277 and Ϫ0.282
0.577 and Ϫ0.442
and hole/e ÅϪ3
were modelled with anisotropic displacement parameters and
final refinement statistics are presented in Table 1.
CCDC reference number 186/2020.
See http://www.rsc.org/suppdata/dt/b0/b003436n/ for crystal-
lographic files in .cif format.
Results and discussion
Synthesis of new Schiff base ligands
The Schiff base ligands were readily prepared by a four-step
convergent synthesis (Scheme 3) from a substituted salicylalde-
hyde, paraformaldehyde, morpholine and a diamine of choice.
The preferred method of preparation of the substituted sali-
cylaldehydes was based upon the industrial process developed
by Levin and co-workers12 involving magnesium-mediated
formylation using magnesium methoxide and paraformalde-
hyde. Subsequent substitution in the remaining ortho position
with the morpholinomethyl group was achieved via a Mannich
reaction13 under non-aqueous aprotic conditions. The Mannich
base ethoxymorpholinomethane was prepared13 from morpho-
line and paraformaldehyde in ethanol with potassium carbon-
ate and purified by distillation. The Schiff base condensations
of the substituted salicylaldehydes with appropriate diamines
were undertaken at room temperature in non-aqueous solvents.
Each ligand has been characterised by 1H and 13C NMR
spectroscopy, FAB mass spectrometry and elemental analysis.
Fig. 3 An ORTEP16 plot (30% probability ellipsoids) of ligand 2
showing the pseudo C2 rotation axis relating the halves of the ligand A
and B. Selected bond lengths (Å); C(22A)–N(2A), 1.460(4), C(21A)–
N(2A), 1.271(4), C(22B)–N(2B), 1.460(4), C(21B)–N(2B), 1.271(4),
C(21A)–C(2A), 1.452(4), C(21B)–C(2B), 1.454(4), C(1A)–O(1A),
1.359(3), C(1B)–O(1B), 1.364(4). Intramolecular distances between
nitrogen and oxygen atoms (Å); N(2A) ؒ ؒ ؒ O(1A) 2.635 and
N(2B) ؒ ؒ ؒ O(1B) 2.604.
O(1A) and O(1B) are separated by 6.53 Å. The dihedral angle
of 95.3Њ between the benzene rings of 2 suggests a predis-
position for tetrahedral co-ordination geometry. However, the
distance between the phenolic oxygen donors is large compared
with that needed to define adjacent sites at a metal ion. The 1,2-
cyclohexane bridge constrains the two pendant morpholine
groups to be closer together (N(62A) ؒ ؒ ؒ N(62B) 11.77 Å) than
in the ethane bridged ligand 4.
Neither of these ligands in their metal-free or “at rest” state
has a geometry preorganised for either cation or anion binding.
To bring the morpholine rings into close proximity to encap-
sulate an anion requires a major change to the geometry of the
ligand backbone. The extent to which incorporation of a metal
cation into the salen donor set “templates” the creation of an
anion binding site is considered below. The preorganisation of
ligands for anion binding by incorporation of metals into the
system has previously been reported.19 The novelty of the
ligands 1–3 is the concomitant creation of an electrostatic
binding site for the anion (see below).
“Free” ligand structures
Single crystals of ligand 2 suitable for structure determination
were grown by evaporation of a CH2Cl2 solution. The molecule
(Fig. 3) adopts a trans configuration with a pseudo twofold
rotation axis bisecting the cyclohexane unit. The trans configur-
ation is similar to that found17 in the solid state structures of a
number of similar H2salen ligands. The tertiary butyl groups
are disordered with about two sites of approximately equal
occupancy.
During the course of this work, Shanmuga et al.18 published
the structure of ligand 4, which is similar to 1, having two
pendant morpholine groups, but with the tertiary butyl groups
replaced by methyls. The molecule adopts a fully extended
conformation with an N–C–C–N torsion angle of 180Њ in the
central ethane bridge and a distance of 8.19 Å between the
phenolic oxygen atoms. The morpholine rings are also well
separated, with a morpholine N ؒ ؒ ؒ N distance of 16.22 Å.
By comparison, the rigid 1,2-cyclohexane bridging group of
compound 2 constrains the ligand to a configuration better
predisposed to metal complexation by bringing the [N2O2]
donor set into closer proximity. The torsion angle N(2A)–
C(22A)–C(22B)–N(2B) is 63.1Њ and the phenolic oxygens
Formation of metal salt complexes
Nickel() was selected for study of the binding of metal salts by
the ditopic ligands 1 and 2 on the grounds that it readily forms
low spin diamagnetic planar complexes with H2salen-type
ligands and that these can conveniently be studied by 1H NMR
J. Chem. Soc., Dalton Trans., 2000, 3773–3782
3777

View Online
Table 3 Shift in ligand proton resonances upon complexation (∆δ
(ppm), 200 MHz, CDCl3)
Table 2 Electronic absorption spectra of ligands 1 and 2 and their
complexes of NiII and CuII
Compound
λmax/nm
Compound
N᎐CH
NCH2CH2N
Ϫ0.23
Ϫ0.09
Ϫ0.13
—
Ar CH2N
Ϫ0.29
ϩ0.4
ϩ0.33
ϩ0.66
᎐
1
2
235, 262, 329, 413
234, 261, 328, 417
324, 346, 417, 460 (sh)
328, 339, 418, 460 (sh)
327, 339, 415, 460 (sh)
329, 345, 417, 460 (sh)
371, 557
[Ni(1 Ϫ 2H)]
[Ni(1)(SO4)]
[Ni(1)(NO3)2]
[Ni(2)(SO4)]
Ϫ0.91
Ϫ0.70
Ϫ0.72
Ϫ0.77
[Ni(1 Ϫ 2H)]
[Ni(1)(SO4)]
[Ni(1)(NO3)2]
[Ni(2)(SO4)]
[Cu(1)(SO4)]
co-ordination which was confirmed by X-ray crystallography
(see below). The NMR spectra showed no evidence of para-
magnetic line broadening or shifts suggesting the square planar
geometry existing in the solid state is retained in CDCl3 solu-
tion and that there is no strong interaction with the anions
in axial co-ordination sites. Metallation with nickel shifts
signals of protons close to the co-ordination site to lower
frequency relative to those of the “free” ligand and the
phenolic proton signal is lost consistent with metal incorpor-
ation into an [N2O2]2Ϫ donor set. Data for the azomethine
group and bridging methylene groups are shown in Table 3.
A comparison of the spectra of [Ni(1 Ϫ 2H)] with those of
the salt complexes [Ni(1)(SO4)], [Ni(1)(NO3)2] and [Ni(2)(SO4)]
in regions associated with the pendant morpholine groups
provides evidence for the protonation of the tertiary nitrogens
and their bonding to the attendant anions in solution. The
Ar CH2N methylene signal of [Ni(1 Ϫ 2H)] is shifted upfield
relative to the “free” ligand to a similar extent to those of
methylene groups linking the azomethine nitrogen atoms (Table
3). In contrast, in the salt complexes protonation of the tertiary
amine nitrogens shifts the Ar CH2N methylene signals down-
field (Table 3). Also, unlike [Ni(1 Ϫ 2H)], for the salt complexes
both the morpholine methylene resonances are broadened and
can only just be seen above the baseline. Association of the
anions with the morpholinium rings appears to reduce ring
flexibility and lowers their symmetry. Protonation of the mor-
pholine nitrogen also reduces the overall symmetry in addition
to slowing the rate of ring conformation inversion.
spectroscopy. The formation of the nickel() sulfate complexes
occurs almost immediately upon mixing in alcoholic solution.
Elemental analysis and FAB mass spectrometry confirm the
formation of 1:1:1 nickel:ligand:sulfate assemblies in the
solid state. When nickel() nitrate is used the dinitrato complex
[Ni(1)(NO3)2] is obtained. Copper() sulfate complexes are also
readily prepared in a similar manner. In this paper we include
only the one copper complex, [Cu(1)(SO4)], the characteristic of
which relates to solvent extraction data (see below). Character-
istic absorbances for the sulfate and nitrate anions were
observed in the IR spectra of these complexes.
The basicity of the counter ion is important in determining
the outcome of the complexation reactions. When nickel()
acetate is used as the metal source there is no evidence for the
presence of the attendant anion in the product. The formation
of [Ni(1 Ϫ 2H)] follows the norm for H2salen-type ligands, pre-
sumably because acetate effectively competes with the morphol-
ine nitrogen atoms for the protons released by the phenolic
groups.20 The ability to prepare neutral complexes of either a
metal dication or its salts is very unusual and could have
important consequences in developing protocols to deal with
the loading and stripping in solvent extraction processes (see
below).
Electronic absorption spectroscopy
The electronic absorption spectra of ligands 1 and 2 consist of a
set of three intense bands at 235, 262 and 329 nm involving
π → π* transitions and a low intensity feature in the 410–420
nm region involving n → π* excitation (Table 2).21 The colour
changes due to metal complexation follow those expected for
H2salen type complexes of these metals, i.e. red-orange for NiII
and dark brown for CuII. Metallation with nickel results in two
overlapping features assigned to two different types of transi-
tions. A broad band in the region 300–360 nm involves intra-
ligand π → π* transitions and a new set of absorbances in the
380–460 nm region with charge transfer character. A weak
feature assigned to d–d transitions is present in the 500–600 nm
region.22 Changes in the spectra upon copper complexation are
less dramatic. An intense mixture of π → π* and n → π*
ligand based transitions occurs at around 370 nm. A weak
asymmetric broad band in the visible region around 550 nm can
be assigned to d–d transitions.23,24
A comparison of the spectrum of the “nickel only” complex
with those of the nickel salt complexes indicates that the effect
of anion binding upon the electronic absorption spectra of
these complexes is minimal. This suggests that the cation and
anion binding sites are well separated in these assemblies in
solution and that the anions do not interact strongly with the
metal-based chromophore. This observation is supported by
NMR and solid state structural studies (see below).
Metal salt extraction experiments
Some preliminary solvent extraction experiments have been
carried out to establish proof-of-concept of transporting metal
sulfates in water immiscible solvents. For these experiments
copper sulfate was used as the kinetics of phase transfer of
nickel salts was found to be very slow in comparison and much
longer equilibration times are needed. Close to 100% extraction
of copper and sulfate into a chloroform solution of lipophilic
ligand 3 was observed on contacting an aqueous CuSO4 solu-
tion (1 M) at pH 3.8.24 Analysis of the loaded organic solution
indicates that copper sulfate has been accommodated as dis-
crete 1:1:1 copper:ligand:sulfate packages. Development of
the stripping protocols for copper and sulfate have required
tuning down the “strength” of the metal binding site to facili-
tate the removal of copper prior to recovery of the sulfate as in
Scheme 1. The development of lipophilic ligands with selec-
tivities and strengths to affect recovery of specific metals will be
described in forthcoming papers. In the section below solid
state structures of nickel complexes are used to define modes of
anion binding and consider the co-operativity of Ni2ϩ and
anion complexation.
X-Ray crystallography
The crystal structure of [Ni(1 Ϫ 2H)] confirms the absence of
any associated anions (Fig. 4). The approximately planar
[Ni(salen)] components stack “back to front” about the inver-
sion centre with azomethine H to adjacent phenyl ring distances
of approximately 3.8 Å and a weak C–H ؒ ؒ ؒ N intermolecular
hydrogen bond between one of the tertiary amine nitrogens and
a bridging methylene hydrogen (H(22C) ؒ ؒ ؒ N(62A) 2.51 Å with
1H NMR spectra of the nickel(II) complexes
The 1H NMR spectra of the nickel() complexes were recorded
at 200 MHz in CDCl3 at room temperature. Complexes of
ligand 1 were shown to be diamagnetic with no solid state room
temperature magnetic moment, consistent with square planar
3778
J. Chem. Soc., Dalton Trans., 2000, 3773–3782

View Online
Table 4 Bond lengths (Å) and angles (Њ) in the co-ordination spheres of [Ni(1 Ϫ 2H)], [Ni(1)(SO4)] and [Ni(1)(NO3)2]
[Ni(1 Ϫ 2H)]
[Ni(1)(SO4)]
[Ni(1)(NO3)2]
Part Aa
Part Bb
Part A
Part B
1.858(2)
Part A
Part B
Ni(1)–N(2)
Ni(1)–O(1)
1.834(3)
1.849(2)
1.838(3)
1.844(3)
1.846(2)
1.848(2)
1.852(3)
1.859(3)
1.852(3)
1.855(2)
1.854(2)
O(1)–Ni(1)–N(2)
O(1)–Ni(1)–O(1)Јb
N(2)–Ni(1)–N(2)
O(1)–Ni(1)–N(2)Ј
N(2)–Ni(1)–O(1)Ј
Σ cis-angles
94.77(13)
84.92(11)
86.22(14)
175.46(12)
175.75(12)
360.3(5)
12.3
94.42(13)
93.67(8)
85.90(7)
85.97(9)
177.20(8)
178.86(9)
360.1(3)
38.3
94.51(8)
—
—
—
—
—
—
94.16(12)
86.38(11)
85.73(13)
173.36(14)
174.40(13)
360.7(5)
42.4
94.38(12)
—
—
—
—
—
—
—
—
—
—
—
—
N(2)–C(22)–C(22)Ј–
N(2)Ј torsion
Inclinations (Њ) between the least squares plane of the N2O2 donor set and the N(2)–C(2)–C(3)–C(22)–O(1) chelate rings
4.0 9.1 16.9 1.1 Ϫ10.6 7.6
Displacement of morpholine ring atoms (Å) from least squares plane of the N2O2 donor set
N(62)
O(65)
0.806
2.286
1.672
3.825
1.559
4.168
Ϫ1.250
Ϫ4.065
0.138
2.091
2.121
4.225
a Parts A and B refer to the chemically equivalent halves of the ligand (see figures for definitions of atom labels). b The prime refers to the chemically
equivalent atom in part B of the ligand.
Fig. 4 An ORTEP plot (50% probability ellipsoids, as in all structures
shown) of the molecular structure of [Ni(1 Ϫ 2H)].
Fig. 5 An ORTEP plot of the molecular structure of [Ni(1)(SO4)].
4) is very similar to that in [Ni(1 Ϫ 2H)] and other26 [Ni(salen)]
type complexes and only small variations in chelate bite angles
accompany the incorporation of the anion in the morpholine
binding site (see below). The protons formally transferred from
the phenolic oxygens to the morpholine nitrogen atoms were
C(22B)–H(22C) ؒ ؒ ؒ N(62A) 158.4Њ). The planar NiN2O2 unit
has bond lengths and angles typical26 for [Ni(salen)] type
complexes (Table 4). The sum of chelating angles around the
nickel is 360.3(5)Њ.
The co-ordination of Ni2ϩ to the [N2O2]2Ϫ donor set has a
major effect on the conformation of the ligand and, in particu-
lar, the disposition of the pendant morpholine groups. The
bridging ethane-1,2-diamine torsion angle is reduced from 180Њ
in 4 to 12.3Њ and the benzene rings are rotated through almost
180Њ to bring the morpholine groups into proximity; the mor-
pholine N ؒ ؒ ؒ N separation in [Ni(1 Ϫ 2H)] is 5.17 Å, while the
“free” ligand 4 has a morpholine N ؒ ؒ ؒ N distance of 16.22 Å.
The structure determination confirms that the planar nickel()
centre templates the ligand to create a potential anion-binding
cavity between the two protonatable morpholine rings. Incor-
poration of a dianion into this dicationic cavity will generate
a charge-neutral entity with a relatively lipophilic exterior,
increasing the overall solubility of the package in non-polar
solvents, an important feature for solvent extraction purposes.
The two morpholine groups are also likely to hydrogen-bond
strongly to any incorporated anion.
detected in the difference map (largest peak is 1.234 e ÅϪ3
,
deepest hole is Ϫ0.669 e ÅϪ3). The sulfate anion is bound in the
resulting dipositive cavity by a combination of electrostatic
interactions and two bifurcated N–H ؒ ؒ ؒ O hydrogen bonds
(Table 5). The amine protons are orientated inwards, towards
the cavity occupied by the sulfate, giving an N ؒ ؒ ؒ N separation
of 5.16 Å.
The shortest nickel–sulfate contact distance (Ni ؒ ؒ ؒ O 5.44 Å)
is too large for there to be any strong interaction between the
anion and the Lewis acidic metal centre. Such an arrangement
contrasts with the uranylsalen anion receptor of Reinhoudt and
Ϫ
co-workers28 which has been shown to bind H2PO4 via a
combination of two N-H ؒ ؒ ؒ O (non-charge assisted) hydrogen
bonds and a single co-ordinate bond between a phosphate
oxygen and the uranyl metal centre.
In [Ni(1)(SO4)] the incorporation of the sulfate between the
morpholinium groups is associated with changes in the con-
formation of the ligand framework. The morpholine rings are
located on opposite sides of the NiN2O2 plane and the ligand
backbone experiences a slight twist with an increase in both the
bridging ethane-1,2-diamine torsion angle and dihedral angle
between the benzene rings (Table 4). The nickel co-ordinate
The structure determination of [Ni(1)(SO4)] (Fig. 5) demon-
strates how a combination of preorganisation of the ligand by
nickel() for anion binding and transformation of the ligand
into its zwitterionic form are key to the strong metal salt bind-
ing. The geometry of the metal co-ordination sphere (see Table
J. Chem. Soc., Dalton Trans., 2000, 3773–3782
3779

View Online
Table 5 Hydrogen bonded distances (Å) and angles (Њ) around the
sulfate anion in [Ni(1)(SO4)]
N ؒ ؒ ؒ O
N–H ؒ ؒ ؒ O
H ؒ ؒ ؒ O
O1 ؒ ؒ ؒ H(62A)–N(62A)
O2 ؒ ؒ ؒ H(62B)–N(62B)
O3 ؒ ؒ ؒ H(62A)–N(62A)
O3 ؒ ؒ ؒ H(62B)–N(62B)
2.870(4)
2.923(3)
2.993(4)
2.915(4)
166.2
154.1
128.8
132.1
1.958
2.322
2.058
2.209
Note: the N–H distances were fixed at 0.91 Å. Hydrogen bonds were
assigned using the PLATON program27 to interactions of the oxygen
and nitrogen atoms that were <3.12 Å.
Table 6 Hydrogen bonded distances (Å) and angles (Њ) around the
nitrate anions in [Ni(1)(NO3)2]
N ؒ ؒ ؒ O
N–H ؒ ؒ ؒ O
H ؒ ؒ ؒ O
Fig. 7 An ORTEP plot of the molecular structure of [Ni(2)(SO4)].
Selected bond lengths (Å) and angles (Њ); Ni(1)–N(15) 1.824(8), Ni(1)–
N(8) 1.859(7), Ni(1)–O(22) 1.832(6) and Ni(1)–O(1) 1.854(6); O(22)–
Ni(1)–N(15) 94.9(3), O(1)–Ni(1)–N(8) 94.6(3), O(22)–Ni(1)–O(1)
85.2(3) and N(15)–Ni(1)–N(8) 86.2(3).
O(1S1) ؒ ؒ ؒ H(62A)–N(62A)
O(1S2) ؒ ؒ ؒ H(62A)–N(62A)
O(2S1) ؒ ؒ ؒ H(62B)–N(62B)
O(2S3) ؒ ؒ ؒ H(62B)–N(62B)
2.906
2.938
3.306
2.760
137.30
160.24
129.00
171.38
2.116
2.005
2.619
1.811
Details as in Table 5.
suited to sulfate binding because there are only very minor
changes to the nickel geometry when the sulfate ion is incorpor-
ated (compare [Ni(1 Ϫ 2H)] and [Ni(1)(SO4)] in Table 4).
Such co-operativity of cation/anion binding will have an impor-
tant consequence on the selectivity of recovery of metal salts
by solvent extraction. The geometric interplay between the cat-
ion and anion binding sites will also influence the external
structure of the package. More favourable distribution coef-
ficients for extraction of metal salts into the hydrocarbon
solvents preferred for industrial use are likely to be observed
when the polar cation and anion components in the overall
charge-neutral complex are shielded from the solvent by hydro-
phobic substituents. Work is in hand to adapt the prototype
ligands described here to achieve this.
Extended supramolecular assemblies
Fig. 6 An ORTEP plot of the molecular structure of [Ni(1)(NO3)2].
The formation of these simple supramolecular metal salt
assemblies led us to envisage the possibility of building up
three-dimensional arrays of metal cations using the cationic
binding sites in conjunction with various bridging anions. The
incorporation of metal ions into hydrogen-bonded arrays has
recently received attention as such systems may have interesting
magnetic, optical and conductivity properties.29 The structure
of [Ni(1)(NO3)2] showed the nitrate anions bound on the out-
side of the morpholine cavity which although undesirable for
solvent extraction purposes demonstrates that anions could
perhaps bridge between two ligands. Such an arrangement is
found in the structure of [Ni(2)(SO4)] (Fig. 7). Two nickel com-
plexes of the 1,2-cyclohexane bridged ligand 2, related by an
inversion centre, are linked into a pseudo macrocycle by two
bridging sulfate anions via a combination of electrostatic inter-
actions and intermolecular N–H ؒ ؒ ؒ O hydrogen bonds (Table
7). The nickel bond lengths and angles are not significantly
different from those of the nickel complexes of ligand 1 despite
the fact that the rigid 1,2-cyclohexane unit fixes the bridging
torsion angle closer to that of the “free” ligand (N(15)–C(14)–
C(9)–N(8) 51.8Њ, “free” ligand 63.1Њ).
bonds are not significantly changed and the square planar
geometry is retained. To accommodate this the ligand experi-
ences a slight splaying with a smaller N(2)–Ni–N(2)Ј angle and
a larger O(1)–Ni–O(1)Ј angle. The accommodation of the anion
between the morpholinium groups is facilitated by the flexibility
of the methylene linkages which allow various conformations
to be adopted as demonstrated by this set of structures.
The solid state structure of [Ni(1)(NO3)2] indicates that both
nitrates are bound to the complex unit (Fig. 6) by a combin-
ation of electrostatic and bifurcated N–H ؒ ؒ ؒ O hydrogen bonds
to the morpholinium amine protons (Table 6). In contrast to the
sulfate complex the nitrate anions are bound outside of the
cavity formed by the morpholine rings, and the quaternary
protons point away from each other. The morpholine rings are
arranged on the same side of the plane described by the NiN2O2
complex in a folded arrangement similar to that in [Ni(1 Ϫ 2H)]
whereas in the sulfate structure they sit on opposite sides with
a pseudo C2 rotation axis. One of the tertiary butyl groups is
disordered about two sites of approximately equal occupancy.
A
comparison of the [Ni(1)(SO4)], [Ni(1)(NO3)2] and
[Ni(1 Ϫ 2H)] structures suggests that incorporation of the
sulfate into the morpholinium binding site has a smaller effect
on the nickel co-ordination geometry than accommodation of
the two nitrates. The N–C–C–N torsion angle of the ethane
bridge in [Ni(1)(NO3)2] is increased, and while the co-ordinate
bonds around the square planar nickel are not significantly
different from the other nickel complexes, the ligand is splayed
with a smaller N(2)–Ni–N(2)Ј angle and a larger O(1)–Ni–O(1)Ј
angle (Table 4). It appears that the templating of the ligand 1
by co-ordination to nickel() generates a system which is well
The pendant morpholine rings are arranged on opposite
sides of the plane of the nickel salen complex creating the same
pseudo C2 symmetry as that of [Ni(1)(SO4)]. The two sulfate
anions are bound to the morpholinium amine protons, one by a
single N–H ؒ ؒ ؒ O hydrogen bond and the other by a bifurcated
N–H ؒ ؒ ؒ O hydrogen bond (Table 7). The sulfates are bound
facially to the morpholine rings so that they can form addi-
tional hydrogen bonds to the morpholine quaternary amine
protons of the opposing ligand. In total each sulfate forms one
single hydrogen bond and one bifurcated hydrogen bond.
3780
J. Chem. Soc., Dalton Trans., 2000, 3773–3782

View Online
Table 7 Hydrogen bonded distances (Å) and angles (Њ) around the
sulfate anion in [Ni(2)(SO4)]
N ؒ ؒ ؒ O
N–H ؒ ؒ ؒ O
H ؒ ؒ ؒ O
O(1)Ј ؒ ؒ ؒ H(2A)–N(2A)
O(4)Ј ؒ ؒ ؒ H(2A)–N(2A)
O(2)Ј ؒ ؒ ؒ H(2B)–N(2B)
3.165
2.609
2.646
123.1
173.0
171.4
2.568
1.693
1.732
Note: the N–H distances were fixed at 0.92 Å. Other details as in
Table 5.
Fig. 9 Proposed mode of hydrogen bond formation between the rigid
polyacids terephthalic acid and (4-phosphonophenyl)phosphonic acid
and quaternary ammonium protons.
Conclusions
We have used relatively simple ditopic ligands related to H2salen
to establish the proof-of-concept of selective transportation of
metal sulfates under conditions which would allow the devel-
opment of novel flowsheets for hydrometallurgical processing
of sulfidic ores. The accommodation of the metal sulfate into a
zwitterionic form of the ditopic ligand ensures that the com-
plexed package is charge neutral permitting transfer into non-
polar water immiscible solvents suitable for solvent extraction
and can effectively be stripped to generate an electrolyte suit-
able for metal recovery. These design features, coupled with the
ability to tailor the selectivity of the metal cation binding site,
would form the basis for the unit operations of “separation”
and “reduction” in an industrial process, generating an electro-
lyte with a suitable purity to effect “reduction”. Commercial
viability will depend on developing ligands which show stability
to hydrolysis and oxidation and good kinetics of metal
transport under operating conditions.
Fig. 8 Supramolecular assemblies of metal salts with ditopic ligands.
Motifs a, b and c are observed in the structures of [Ni(1)(SO4)],
[Ni(1)(NO3)2] and [Ni(2)(SO4)] respectively. Assemblies d and e are
proposed using the rigid polyacids (4-phosphonophenyl)phosphonic
acid and terephthalic acid respectively.
The design of prototype extractants has been based on
principles of supramolecular chemistry which should also
allow assembly of unusual solid state materials.
The observation that anions are able to bridge two ligands
via intermolecular charge-assisted hydrogen bonding presents
us with the opportunity to design two- or three-dimensional
supramolecular assemblies. To develop strategies for building
up arrangements of molecular building blocks requires that
they be rigid with a suitable supramolecular synthons, e.g.
a set of complementary hydrogen bond donors and acceptors,
such that the resulting architectures are predictable and
reproducible.30 Preorganisation of ditopic ligands similar to
1–4 by binding different metals to create molecules with par-
ticular dispositions of the anion binding sites should facilitate
their use in this way. The three types of simple metal salt
assemblies obtained so far are represented schematically as
a–c in Fig. 8. Each pendant morpholinium quaternary amine
group represents a single hydrogen bond donor, which can bind
by a combination of electrostatic and hydrogen bonding to a
range of anions. The electrostatic component is largely non-
directional providing the “glue” for the interaction whereas
the hydrogen bonds provide the directional element, allowing
the linking together of ions in a predictable manner.
Rigid polyanions such as terephthalate and (4-phosphono-
phenyl)phosphonate provide anionic hydrogen bond acceptors
with predictable bonding properties. Terephthalic acid has been
used to link metal complexes into extended supramolecular
tapes and sheets.31,32 A combination of these anionic hydrogen
bond acceptor synthons with the cationic hydrogen bond
donor sites of our complexes would be expected to generate
oligomeric systems, such as e or two or three dimensional
lattices depending on the level of multifunctionality in the
anion and/or cation (Fig. 9). Work is in hand using (4-
phosphonophenyl)phosphonic acid to bridge the metal com-
plexes in dumbbell assemblies (d) and terephthalic acid to
form extended sheets or ribbons (e). Assembly is effected by
addition of the polyacid to preformed neutral metal complexes
such as [Ni(1 Ϫ 2H)], etc.
Acknowledgements
We thank Mr J. Millar and Mr W. Kerr for obtaining NMR
spectra, Mr A. Taylor and Mr H. Mackenzie for mass spectra
and Ms L. Eades for elemental analysis. We gratefully acknow-
ledge the EPSRC for a ROPA award (GRM/33303) and Avecia
(formerly Zeneca Specialties) for funding.
Notes and references
1 D. S. Flett, Hydrometallurgy, 1992, 30, 327.
2 J. Szymanowski, in Hydroxyoximes and Copper Hydrometallurgy,
CRC Press, Boca Raton, FL, 1993, p. 62.
3 For reviews of anion recognition: B. Dietrich, Pure Appl. Chem.,
1993, 65, 1457; C. Seel and J. de Mendoza, in Comprehensive
Supramolecular Chemistry, eds. J. L. Atwood, J. E. D. Davies,
D. D. MacNicol and F. Vogle, Elsevier Science Ltd, Oxford, 1996,
pp. 519–552; A. Bianchi, K. Bowman-James and E. Garcia-España
(Editors), Supramolecular Chemistry of Anions, Wiley-VCH, New
York, 1997; F. P. Schmidtchen and M. Berger, Chem. Rev., 1997, 97,
1609; M. M. G. Antonisse and D. N. Reinhoudt, Chem. Commun.,
1998, 443.
4 B. C. Hamann, N. R. Branda and J. Rebek, Jr., Tetrahedron Lett.,
1993, 34, 6837; T. R. Kelly and M. H. Kim, J. Am. Chem. Soc., 1994,
116, 7072; S. Nishizawa, P. Bühlmann, M. Iwao and Y. Umezawa,
Tetrahedron Lett., 1995, 36, 6483; C. Bazzicalupi, A. Bencini,
A. Bianchi, M. Cecchi, B. Escuder, V. Fusi, E. Garcia-España,
C. Giorgi, S. V. Luis, G. Maccagni, V. Marcelino, P. Paoletti and
B. Valtancoli, J. Am. Chem. Soc., 1999, 121, 6807.
5 For metal salt binding ditopic receptors: U. Olsher, F. Frolow,
N. K. Dalley, J. Weiming, Z.-Y. Yu, J. M. Knobeloch and R. A.
Bartsch, J. Am. Chem. Soc., 1991, 113, 6570; E. A. Arafa, K. I.
Kinnear and J. C. Lockhart, J. Chem. Soc., Chem. Commun., 1992,
61; P. J. Smith, M. V. Reddington and C. S. Wilcox, Tetrahedron
Lett., 1992, 33, 6085; S. S. Flack, J.-L. Chaumette, J. D. Kilburn,
G. J. Langley and M. Webster, J. Chem. Soc., Chem. Commun., 1993,
399; P. B. Savage, S. K. Holmgren and S. H. Gellman, J. Am. Chem.
J. Chem. Soc., Dalton Trans., 2000, 3773–3782
3781

View Online
pKa and solution conditions determines whether the anion or
morpholine nitrogen will bind the liberated protons. P. W. Atkins,
Physical Chemistry 5th Edition, Oxford University Press, Oxford,
1994; D. D. Perrin, Dissociation Constants of Organic Bases in
Aqueous Solutions, Butterworth & Co., London, 1965.
21 S. M. Crawford, Spectrochim. Acta, 1963, 19, 255.
Soc., 1994, 116, 4069; M. T. Reetz, in Comprehensive Supramolecular
Chemistry, eds. J. L. Atwood, J. E. D. Davies, D. D. MacNicol and
F. Vogle, Elsevier Science Ltd, Oxford, 1996, pp. 553–563; N. Pelizzi,
A. Casnati, A. Friggeri and R. Ungaro, J. Chem. Soc., Perkin
Trans. 2, 1998, 1307; P. D. Beer and J. B. Cooper, Chem. Commun.,
1998, 129.
6 P. D. Beer and D. K. Smith, Prog. Inorg. Chem., 1997, 46, 1 and
references therein.
22 S. Di Bella, I. Fragala, I. Ledoux, M. A. Diaz-Garcia and T. J.
Marks, J. Am. Chem. Soc., 1997, 119, 9550.
7 P. D. Beer, P. K. Hopkins and J. D. McKinney, Chem. Commun.,
1999, 1253.
8 K. Kavallieratos, R. A. Sachleben, G. J. Van Berkel and B. A.
Moyer, Chem. Commun., 2000, 187.
9 L. A. J. Chrisstoffels, F. de Jong, D. N. Reinhoudt, S. Sivelli,
L. Gazzola, A. Casnati and R. Ungaro, J. Am. Chem. Soc., 1999,
121, 10142.
10 D. J. White, N. Laing, H. Miller, S. Parsons, S. Coles and P. A.
Tasker, Chem. Commun., 1999, 2077; P. A. Tasker and D. J. White,
Br. Pat., 9907485.8, 1999.
11 E. A. Boudreaux and C. N. Mulay, Theory and Applications of
Molecular Paramagnetism, Wiley, London, 1976.
12 R. Aldred, R. Johnston, D. Levin and J. Neilan, J. Chem. Soc.,
Perkin Trans. 1, 1994, 1823.
13 H. Adams, N. A. Bailey, D. E. Fenton and G. Papageorgiou,
J. Chem. Soc., Dalton Trans., 1995, 1883.
14 SHELXS 97, G. M. Sheldrick, University of Göttingen, 1997.
15 SHELXL 97, G. M. Sheldrick, University of Göttingen, 1997.
16 C. K. Johnson, ORTEP II, Report ORNL-5138, Oak Ridge
National Laboratory, Oak Ridge, TN, 1976.
23 M. M. Bhadhade and D. Srinivas, Inorg. Chem., 1993, 32, 5458.
24 S. Koner, Chem. Commun., 1998, 5, 593.
25 A crystal structure of [Cu(1)(SO4)] has been obtained which shows
the copper ion bound in a square planar geometry to the [N2O2]2Ϫ
donor set with bond lengths and angles typical of copper bound to
H2salen type ligands. The sulfate anion is bound between the
morpholine rings electrostatically and by two N–H ؒ ؒ ؒ O hydrogen
bonds to the quaternary amine protons. The metal salt binding in
this structure is analogous to that described in this paper for
[Ni(1)(SO4)]. The structure of [Cu(1)(SO4)] will be reported in full in
a subsequent publication.
26 A survey of the CCDC reveals 66 square planar nickel() salen
complexes. Ni–N and Ni–O distances average 1.86 and 1.85 Å
respectively. D. A. Fletcher, R. F. McMeeking and D. J. Parkin,
J. Chem. Inf. Comput. Sci., 1996, 36, 746.
27 P. van der Sluis and A. L. Spek, Acta Crystallogr., Sect. A, 1990, 46,
34, 194.
28 D. M. Rudkevich, Z. Brzozka, M. Palys, H. C. Visser, W. Verboom
and D. N. Reinhoudt, Angew. Chem., Int. Ed. Engl., 1994, 33,
467.
17 K. Bernardo, S. Leppard, A. Robert, G. Commenges, F. Dahan and
B. Meunier, Inorg. Chem., 1996, 35, 387.
18 S. Shanmuga, S. Raj, R. Thirumurugan, G. Shanmugam, H.-K.
Fun, J. Manonmani and M. Kandaswamy, Acta Crystallogr., Sect.
C, 1999, 55, 94.
29 D. Braga, F. Grepioni and G. R. Desiraju, Chem. Rev., 1998, 98,
1375; R. Vilar, M. P. Mingos, A. J. P. White and D. J. Williams,
Angew. Chem., Int. Ed., 1998, 37, 1258; M. M. Chowdhry, M. P.
Mingos, A. J. P. White and D. J. Williams, Chem. Commun., 1996,
899; A. D. Burrows, C.-W. Chan, M. M. Chowdry, J. E. McGrady
and D. M. P. Mingos, Chem. Soc. Rev., 1995, 24, 329.
30 C. B. Aakeröy, Acta Crystallogr., Sect. B, 1997, 53, 569.
31 E. Kimura, T. Ikeda, M. Shionoya and M. Shiro, Angew. Chem.,
Int. Ed. Engl., 1995, 34, 663.
19 K. Araki, S. K. Lee, J. Otsuki and M. Seno, Chemistry Lett., 1993,
493; P. D. Beer, Chem. Commun., 1996, 689; M. J. Deetz and
B. D. Smith, Tetrahedron Lett., 1998, 39, 6841.
Ϫ
20 Comparison of the pKa of HSO4 (1.92), HNO3 (3.37) and
CH3CO2H (4.75) with the values for a range of N-alkyl substituted
morpholine derivatives (pKa 6–7) suggests a combination of relative
32 A. D. Burrows, D. M. P. Mingos, A. J. P. White and D. J. Williams,
Chem. Commun., 1996, 97.
3782
J. Chem. Soc., Dalton Trans., 2000, 3773–3782