Diazopyrazolones as weak solvent extractants for copper from ammonia leach solutions

Article abstract of DOI:10.1039/b101780m

Potentially bidentate 4-diazopyrazol-5-one ligands carrying a range of substituents can readily be prepared by coupling arene diazonium salts on to the appropriate pyrazolones. Hydrocarbon-soluble versions are shown to be suitable reagents for the recovery of copper by solvent extraction from ammoniacal leach solutions, and potentially have greater resistance to chemical degradation than the β-diketones which have been used. Cobalt(III), nickel(II), copper(II) and zinc(II) complexes with 4-(4-tert-butylphenylazo)-3-methyl-1-phenyl-5-pyrazolone (HL) have been characterised by X-ray crystallography. The significant deviations from planar coordination geometry which are observed in [CuL₂] and [CuL′₂], where HL′ is 3-methyl-1-phenyl-4-phenylazo-5-pyrazolone, arise from interligand repulsion, and account for the relatively weak complexation and ease of stripping of copper from this class of extractant. The cobalt(III) complex [CoL₃]·3MeOH and the high spin nickel(II) complex [NiL₂(MeOH)₂]·2MeOH both have approximately octahedral geometries, but show significantly different bite distances in the chelate rings. The zinc complex [ZnL₂] has a pseudo-tetrahedral structure.

Full text of DOI:10.1039/b101780m

View Article Online / Journal Homepage / Table of Contents for this issue
Diazopyrazolones as weak solvent extractants for copper from
ammonia leach solutions†
Lucy C. Emeleus,a Domenico C. Cupertino,b Steven G. Harris,a Susan Owens,b Simon Parsons,a
Ronald M. Swart,b Peter A. Tasker*a and David J. Whitea
a Department of Chemistry, The University of Edinburgh, West Mains Rd., Edinburgh,
UK EH9 3JJ
b AVECIA, Hexagon House, Blackley, Manchester, UK M9 8ZS
Received 23rd February 2001, Accepted 23rd February 2001
First published as an Advance Article on the web 26th March 2001
Potentially bidentate 4-diazopyrazol-5-one ligands carrying a range of substituents can readily be prepared by
coupling arene diazonium salts on to the appropriate pyrazolones. Hydrocarbon-soluble versions are shown to be
suitable reagents for the recovery of copper by solvent extraction from ammoniacal leach solutions, and potentially
have greater resistance to chemical degradation than the β-diketones which have been used. Cobalt(), nickel(),
copper() and zinc() complexes with 4-(4-tert-butylphenylazo)-3-methyl-1-phenyl-5-pyrazolone (HL) have been
characterised by X-ray crystallography. The significant deviations from planar coordination geometry which are
observed in [CuL2] and [CuLЈ2], where HLЈ is 3-methyl-1-phenyl-4-phenylazo-5-pyrazolone, arise from interligand
repulsion, and account for the relatively weak complexation and ease of stripping of copper from this class of
extractant. The cobalt() complex [CoL3]ؒ3MeOH and the high spin nickel() complex [NiL2(MeOH)2]ؒ2MeOH
both have approximately octahedral geometries, but show significantly different bite distances in the chelate rings.
The zinc complex [ZnL2] has a pseudo-tetrahedral structure.
Introduction
The economic and environmental benefits of recovering copper
by hydrometallurgical, rather than smelting, processes have led
to considerable research and development1–3 of novel leach and
extraction processes. A process based on ammoniacal leaching
of chalocite-based ores under mild conditions4–6 generates two
copper streams; an enriched ore, covellite, which can be treated
in a conventional smelter, and a concentrated aqueous solution
(ca. 5 M, pH 8.5–10) containing ammine complexes [eqn. )1)].
Solvent extraction with a weakly acidic reagent, as in eqn. (2),
regenerates the leachant which is recycled. Stripping the copper
from the loaded organic solution with acid (eqn. 3), followed by
electrowinning the copper (eqn. 4), regenerates the extractant
without overall consumption of acid. The recycling of reagents
between the unit operations leads to a remarkable materials
balance with the overall process being represented by eqn. (5).
Fig. 1 Structures of commercial metal extractants and acylpyr-
azolones.
small quantities of other transition metals are transferred to the
pregnant leach solution (pls).7 As a consequence, and in marked
contrast to the well established acid leach process,8 the solvent
extractant used to recover the copper is not required to show
very high selectivity for copper over iron or high “strength”
because the extraction equilibrium [eqn. (2)] is effectively
buffered by [NH4]ϩ/NH3. Henkel’s LIX54 reagent, 1-heptyl-3-
phenyl-1,3-propanedione (mixed isomer heptyl substituent,
Fig. 1), provides “strength” and selectivity well suited to
ammoniacal recovery processes. However, a potential problem
with organic reagents which contain reactive carbonyl groups is
their propensity to form the corresponding imines in contact
with the high levels of NH3 present in the pregnant leach solu-
tion. Apart from the potential loss of the active reagent from
the system, these imines may also contribute to poor phase
disengagement which will in turn lead to high entrainment of
the aqueous phase in the loaded organic. For these reasons
other systems are now being considered.9,10
Acylpyrazolones (Fig. 1) also form neutral β-diketonate-type
complexes suitable for extraction of a range of metals into
organic solvents11–16 and have been considered17 as alternatives
to LIX54 for use in ammoniacal leach circuits. Although they
are stronger copper extractants than LIX54, the low solubility
of their metal complexes limits18 their usefulness. The struc-
turally related diazopyrazolones (Scheme 1) have not previously
been investigated as copper extractants, but possess very well
Cu2S(s) ϩ 2 NH3(aq) ϩ 2[NH4]ϩ(aq) ϩ ¹ O2(g) →
¯
²
Chalcocite
CuS(s) ϩ [Cu(NH3)4]2ϩ(aq) ϩ H2O(l) (1)
Covellite
[Cu(NH3)4]2ϩ(aq) ϩ 2HL(org) →
[CuL2](org) ϩ 2[NH4]ϩ(aq) ϩ 2NH3(aq) (2)
[CuL2](org) ϩ H2SO4(aq) → CuSO4(aq) ϩ 2LH(org) (3)
CuSO4(aq) ϩ H2O(e) → Cu(s) ϩ H2SO4(aq) ϩ ¹O2(g) (4)
¯
²
Cu2S → Cu ϩ CuS
(5)
The pH of the lixiviant and the leach conditions employed
are such that very little of the iron present in the ores and only
† Electronic supplementary information (ESI) available: characteris-
ation data for compounds 2–7 and complexes 13–15, bite angles/
distances in the complexes and deviations from planarity of 1, 2 in the
complexes. See http://www.rsc.org/suppdata/dt/b1/b101780m/
DOI: 10.1039/b101780m
J. Chem. Soc., Dalton Trans., 2001, 1239–1245
This journal is © The Royal Society of Chemistry 2001
1239

View Article Online
phenyl-5-pyrazolone (3.48 g, 20 mmol) in acetone (40 ml) and
water (40 ml) at 0–5 ЊC. Addition of sodium acetate (10 g, 120
mmol) yielded a bright orange suspension, which was stirred
for 2 h at 0–5 ЊC, collected by filtration, washed with water and
recrystallised (methanol–water, 20 : 1) to yield 1 (5.34 g, 66%)
as orange needles; mp 138–139 ЊC (Found C, 71.8; H, 6.75; N,
17.0. C20H22N4O requires C, 71.8; H, 6.6; N, 16.8%). IR (cmϪ1
,
KBr disc): ν 3430m (NH), 1654s (C᎐O), 1551s (C᎐N). 1H NMR
᎐
᎐
(CDCl3, 200 MHz): δ 1.32 (s, 9H, C(CH3)3), 2.36 (s, 3H, CH3),
7.15–7.24 (tt, Jortho 7.41, Jmeta 1.87, 1H, p-H of Ph), 7.34–7.47
(m, Jortho 6.34 Hz, 4H (ArH–N᎐N) and 2H (m-H of Ph)), 7.91–
᎐
7.98 (dd, Jortho 7.58, Jmeta 2.11 Hz, 2H, o-H of Ph), 13.62 (br s,
1H, NH). 13C NMR (CDCl3, 63 MHz): δ 11.71 (CH3), 31.20
(C(CH3)3), 34.52 (C(CH3)3), 115.51 (aryl), 118.48 (aryl), 125.00
(aryl), 126.47 (aryl), 127.88 (aryl), 128.81 (aryl), 137.97 (aryl),
138.59 (aryl), 148.46 (C᎐N–NH), 149.19 ([CH ]C᎐N, ring),
᎐
᎐
3
157.75 (C᎐O). EIMS: m/z 334 (70.0, Mϩ), 319 (100.0,
᎐
M Ϫ CH3), 132 (14.1, (CH3)3CPh), 91 (20.0, PhN), 77 (34.9%,
Ph).
Ligands 2–7 were prepared similarly. Yields and characteris-
Scheme 1 Formation of complexes of diazopyrazolones.
defined synthetic chemistry17 due to their extensive use in the
dyestuffs industry.19 They have been shown20,21 to form neutral
ation data are provided in ESI supplementary material.
2Ϫ
2 : 1 complexes with copper (Scheme 1) with an N2O2 donor
Complexes
set analogous to that in the commercially successful phenolic
oxime extractants (Fig. 1) used to recover copper from acidic
leach solutions.8 In this paper we report the coordination chem-
istry and extraction properties of a series of diazopyrazolones
with a range of first row transition metals.
[Cu(1 ؊ H)2] 8. Copper() nitrate trihydrate (0.242 g,
1 mmol) in methanol (15 ml) was added to 1 (0.668 g, 2 mmol)
in methanol (40 ml) containing potassium hydroxide (0.112 g,
2.00 mmol). The resulting dark brown solution was stirred
for 0.5 h at room temperature to afford a dark black/brown
precipitate, which was recrystallised (methanol–acetone, 3 : 1)
to yield dark red crystals of [Cu(1 Ϫ H)2] (0.528 g, 72%); mp
238–240 ЊC (Found C, 65.5; H, 5.9; N, 15.4. C40H42CuN8O2
requires C, 65.8; H, 5.8; N, 15.3%). IR (cmϪ1, KBr disc): ν 1561s
Experimental
Instrumentation
IR spectra were obtained on a Perkin-Elmer Paragon 1000 FT-
IR spectrometer as potassium bromide discs or as liquid thin
(C᎐N), 1414m and 1367m (C(CH ) ). FABMS: m/z 858 (28.9,
᎐
3
3
Cu2L3), 794 (29.2, Cu2L2), 730 (13.2, CuL2), 461 (32.0, Cu2L),
397 (41.1, CuL), 334 (38.5, L), 133 (100.0, (CH3)3CPh), 63
(89.9%, Cu). µeff = 1.8 µB (Evans).
1
films. H and 13C NMR spectra were run on Bruker WP200,
AC250 and DPX300 spectrometers and NOE experiments on
a Bruker WH360 spectrometer. Electron impact (EI) mass
spectra were obtained either on a Finnigan MAT4600 quad-
rupole spectrometer or on a Kratos MS50TC spectrometer and
fast atom bombardment (FAB) mass spectra on a Kratos
MS50TC spectrometer in acetonitrile/3-nitrobenzyl alcohol
matrices. Inductively coupled plasma atomic emission spectro-
scopy (ICP-AES) analysis was performed on a Thermo Jarrell
Ash IRIS ICP-AES spectrometer. Magnetic susceptibilities
were measured for the paramagnetic complexes 8, 11, 12, 13,
and 14 using the Evans NMR method22,23 or a Johnson Matthey
magnetic susceptibility balance and molar susceptibilities were
corrected for diamagnetism using Pascal’s constants.24
[Zn(1 ؊ H)2] 9. Zinc() acetate dihydrate (0.079 g, 0.36
mmol) was added to 1 (0.701 g, 2.1 mmol) in methanol (20 ml)
containing triethylamine (1 ml, 7.17 mmol). The resulting pale
yellow solution was stirred for 0.5 h at room temperature to
yield pale yellow crystals of [Zn(1 Ϫ H)2] (0.193 g, 75%); mp
233–235 ЊC (Found C, 64.1; H, 5.9; N, 15.2; C40H42N8O2-
ZnؒH2O requires C, 64.0; H, 5.9; N, 14.9%). IR (cmϪ1, KBr
1
disc): ν 1549s (C᎐N), 1413s and 1365m (C(CH ) ). H NMR
᎐
3
3
(CDCl3, 200 MHz): δ 1.27 (s, 18H, 2 C(CH3)3), 2.54 (s, 6H,
2 CH ), 7.15–7.47 (m, 8H, ArH–N᎐N and 6H(Ph)), 7.47–7.99
᎐
3
(m, 4H, o-H of Ph). FABMS: m/z 731 (100.0, ZnL2), 334
(42.7%, L).
Solvent and reagent pretreatment
[Co(1 ؊ H)3]ؒ3MeOH 10. Cobalt() acetate tetrahydrate
(0.087 g, 0.35 mmol) was added to 1 (0.701 g, 2.10 mmol) in
methanol (20 ml). The resulting dark orange solution was
stirred for 0.5 h at room temperature to yield red crystals of
[Co(1 Ϫ H)3]ؒ3MeOH (0.241 g, 60%); mp 212–213 ЊC (Found
C, 67.7; H, 6.1; N, 15.4. C60H63CoN12O3 requires C, 68.0; H, 6.0;
Substituted anilines and pyrazolones were commercially
available (Acros or Aldrich) with the exception of 1,3-dimethyl-
5-pyrazolone and 1-tert-butyl-3-methyl-5-pyrazolone, which
were prepared by the method of Butler and DeWald.25 Solvents
used for analytical purposes (NMR, MS) were of spectroscopic
grade. Deuteriated dimethyl sulfoxide for use in NOE experi-
ments was dried over 4 Å molecular sieves under nitrogen. All
other reagents and solvents were used as received.
N, 15.9%). IR (cmϪ1, KBr disc): ν 1556s (C᎐N), 756s (ArH). 1H
᎐
NMR (CDCl3, 200 MHz): δ 0.95 (s, 9H, C(CH3)3), 1.17 (s, 9H,
C(CH3)3), 1.45 (s, 9H, C(CH3)3), 1.99 (s, 3H, CH3), 2.33 (s, 3H,
CH3), 2.59 (s, 3H, CH3); 6.37–6.42 (d, 2H), 6.86 (s, 4H),
6.98–7.33 (m, 18H), 7.45–7.53 (t, 2H), 8.04–8.09 (d, 1H)
(all ArH). FABMS: m/z 725 (100.0, CoL2), 525 (8.6, CoL ϩ
(CH3)3CPh), 133 (9.0%, (CH3)3CPh).
Synthesis of the ligands
Compounds 1–7 were prepared as described for 1 by diazotis-
ation of the reagent grade anilines and coupling13 with the
appropriate pyrazolones.
[Ni(1 ؊ H)2(MeOH)2]ؒ2MeOH 11. Nickel() acetate tetra-
hydrate (0.087 g, 0.35 mmol) was added to 1 (0.701 g, 2.10
mmol) in methanol (20 ml). The resulting dark orange solution
was stirred for 0.5 h at room temperature to yield red crystals of
[Ni(1 Ϫ H)2(MeOH)2]ؒ2MeOH (0.168 g, 61%); mp 245–247 ЊC
(Found C, 64.4; H, 6.2; N, 14.7. C40H42N8NiO2ؒH2O requires C,
4-(4-tert-Butylphenylazo)-3-methyl-1-phenyl-5-pyrazolone 1.
Aqueous sodium nitrite (2 M, 10 ml) was added to a stirred
solution of 4-tert-butylaniline (2.98 g, 20 mmol) in water (20
ml) containing concentrated hydrochloric acid (4 ml, 40 mmol)
at 0–5 ЊC. After 1 h this was added to a solution of 3-methyl-1-
1240
J. Chem. Soc., Dalton Trans., 2001, 1239–1245

View Article Online
Table 1 Crystallographic data for complexes 8–12
8
9
10a
11
12
Formula
M
C40H42N8CuO2
730.36
C40H42N8O2Zn
732.19
C60H63CoN12O3ؒ3MeOH
115.28
C42H42N8NiO2ؒ2MeOH
853.69
C32H26N8CuO2
618.15
Crystal system
Space group
Monoclinic
P21/n
Triclinic
P1
Triclinic
P1
Triclinic
P1
Triclinic twin
P1
¯
¯
¯
¯
a/Å
b/Å
c/Å
α/Њ
11.603(3)
23.586(4)
13.649(2)
12.280(5)
13.615(5)
14.172(5)
113.74(2)
95.13(2)
113.635(14)
1896.2(12)
2
12.965(6)
15.604(6)
15.584(6)
75.83(3)
85.67(3)
86.50(4)
3045(2)
2
10.5849(12)
14.8955(17)
15.1532(17)
76.735(6)
79.025(6)
77.539(7)
2245.6(4)
2
8.307(4)
11.311(6)
16.415(8)
75.87(4)
75.79(4)
71.58(4)
1395.1(12)
2
β/Њ
97.24(2)
γ/Њ
U/Å3
3705.5(13)
4
1.197
5435
Z
µ/mmϪ1
1.256
6746
0.342
10736
1.065
7969
1.485
3946
No. of unique data
No. data with [F > σ(F)]
R1
wR2
4000
0.0614
0.1732
5713
0.0342
0.1002
5472
0.0846
0.2201
6601
0.0389
0.1042
3098
0.0504
0.1753
a This unit cell may be transformed into a pseudo C-centred monoclinic cell by the matrix (011/01 Ϫ 1/ Ϫ 100), but Rint for 2/m Laue symmetry was 64%.
64.6; H, 6.0; N, 15.1%). IR (cmϪ1, KBr disc): ν 1581s (C᎐N),
The relationship between the two orientation matrices
identified after the initial reflection search can be expressed by
the matrix30 shown. Several twinning models were considered,
᎐
755s (ArH). FABMS: m/z 725 (100.0, NiL2), 393 (16.9, NiL),
334 (74.2, L), 257 (8.1, L Ϫ Ph), 133 (66.2, (CH3)3CPh), 91
(71.1, PhN), 57 (85.9%, C(CH3)3); µeff = 2.7 µB (Evans).
Complexes 12–15 were prepared as described for 12 by
reaction of two equivalents of the appropriate diazopyrazolone
ligand with one equivalent of copper() acetate monohydrate in
methanol.
0.01 Ϫ0.414 0.50










0
Ϫ1
0
2.03 Ϫ0.815 0.01
including splitting relections with k = 5n and l = 2n, but
with only slight improvement. In all cases the twin scale factor
was less than 10%. It is possible that the off-diagonal terms
in this matrix are far enough from rational numbers that the
scan path taken during data collection meant that the twinning
did not affect the intensity measurements in a consistent
manner. The best agreement indices were obtained when all
reflections whose transformed indices differed from integral
values by less than 0.15 were omitted (603 data in all). R1 under
these conditions was 5.04%. Other refinement details are given
in Table 1.
[Cu(2 ؊ H)2] 12. Copper() acetate monohydrate (0.100 g,
0.5 mmol) was added to 2 (0.278 g, 1 mmol) in methanol
(25 ml). The resulting red solution afforded a red/brown
powder, [Cu(2 Ϫ H)2]ؒ2MeOH (0.277 g, 90%). FABMS: m/z
682 ([Cu(2 Ϫ H)2]ؒ2MeOH). This was recrystallised by dif-
fusion of water into a DMF solution of the complex to yield
red crystals, [Cu(2 Ϫ H)2], mp 220–222 ЊC (Found C, 61.8; H,
4.1; N, 18.1. C32H26CuN8O2 requires C, 62.2; H, 4.25; N,
18.1%). IR (cmϪ1, KBr disc): ν 1564s (C᎐N), 1394w and 1364m
᎐
(C(CH3)3). FABMS: m/z 1300 (7.0, Cu3L4), 1022 (4.0, Cu3L3),
745 (13.8, Cu3L2), 681 (14.8, Cu2L2), 618 (55.6, CuL2), 340
(17.9, CuL), 77 (40.5%, Ph). µeff = 1.80 µB (Balance).
Yields and characterisation data for 13–15 are provided as
ESI.
CCDC reference numbers 158372–158376.
See http://www.rsc.org/suppdata/dt/b1/b101780m/ for crys-
tallographic data in CIF or other electronic format.
Crystallography
Results and discussion
In all cases data were collected at 220 K on a Stoe Stadi-4
diffractometer equipped with an Oxford Cryosystems low-
temperature device, using Cu-Kα radiation for 8, 9, 11 and 12,
and Mo-Kα radiation for 10. Structures 8 and 12 were solved
by direct methods (SHELXTL or SIR 92)26,27 and 9–11 by
Patterson methods (DIRDIF).28 All were completed by iter-
ative cycles of least-squares refinement against F2 and Fourier
difference syntheses (SHELXTL). H atoms were idealised,
those belonging to CH3 groups being first located in a difference
synthesis. In all cases all non-H atoms were modelled with
anisotropic displacement parameters and final refinement
statistics are presented in Table 1.
Preparation of the ligands
The free ligands 1–7 were readily prepared by coupling the
appropriate pyrazolone and diazonium salt derived from the
appropriate para-substituted aniline. All are highly coloured
(yellow/orange), as might be expected, because diazopyr-
azolones have been used19 extensively for over a century in the
dyestuffs industry. Ligands 1–4 with N-phenyl, N-methyl and
unsubstituted pyrazolone units are solids which are easily
recrystallised and characterised, but their solubilities and those
of their copper() complexes in hydrocarbon solvents were
judged to be too low to carry out effective solvent extraction
measurements. Incorporation of a t-butyl group on the pyr-
azolone and a branched alkyl group in the para position of
the phenylazo group gave ligands 5–7, with much higher hydro-
carbon solubility, but which could only be isolated as viscous
oils.
Crystals of 12 grew as aggregates of blocks, and none of
the crystals studied was single. In the case of the sample from
which data are reported here a random reflection search located
25 reflections. This pattern was completely indexed on the basis
of two orientation matrices.29 Some reflections in the search list
gave integral indices with only one of these matrices, others
could be satisfactorily indexed with both. Data were collected
as described above using one of these matrices; no absorption
correction was applied. Routine refinement converged to
R1 = 6.75% with a difference synthesis extrema of 1.13 and
1H and 13C NMR spectra are consistent with the structures
1–7, and NOESY experiments carried out on 1 indicate that
this exists in the hydrazone form (Fig. 2: A) in deuteriated
chloroform or dimethyl sulfoxide. 15N NMR experiments
carried out by Connor et al.30 and crystal structure deter-
mination30–37 support the formulation of arylazopyrazolones
Ϫ0.74 e ÅϪ3
.
J. Chem. Soc., Dalton Trans., 2001, 1239–1245
1241

View Article Online
Fig. 4 Extraction isotherm for azopyrazolone 7 contacted with 30 g
Ϫ1 copper feed (copper values all g lϪ1).
l
phenolic oxime extractants (Fig. 1), used8 commercially for the
recovery of copper from acid sulfate leach solutions, but are
almost ideal for the Escondida-type circuit. Good loading will
result from the [NH4ϩ]/NH3 buffered feed solution, whilst
almost quantitative stripping is expected when using spent
tankhouse electrolyte (pH 0–0.5).
S-Curves of the type shown in Fig. 3 can also be used9 to
assess the selectivity of metal extraction. The pH values, Cu
₁
₂
4.7, Zn 5.8 and Ni > 7.0, are consistent with  selectivities
Fig. 2 Tautomeric forms of diazopyrazolones.
which follow the Irving–Williams order of complex stabilities,
but with the nickel() complex formation being more dis-
favoured than normal relative to zinc. Reproducibility of
extraction of cobalt was poor and the results were consistent
with extraction of cobalt() and subsequent oxidation to
cobalt() (vide infra). Such oxidation is likely to make strip-
ping difficult on both thermodynamic and kinetic grounds.
However, the potential poisoning of the extractant by cobalt
in an Escondida-type circuit is unlikely to be a problem in
practice because the ore bodies do not contain any significant
levels of cobalt. The strength of the azopyrazolones as copper
extractants from ammoniacal feeds is shown in Fig. 4 which
presents the equilibrium extraction isotherm for a 1.0 M solu-
tion of 7 in Philips Orfom SX7 diluent after contact at 20 ЊC
2Ϫ
with an aqueous copper feed containing Cu 30, SO4 75 and
NH3 45 g lϪ1 at various aqueous to organic ratios (A : O). The
copper contents of the aqueous and organic phases for
each A : O ratio were determined using atomic absorption
spectroscopy.
Fig. 3 S-Curves for the extraction of copper by the ligands 5–7
(R1 = t-Bu and R2 is varied as indicated, see also Scheme 1).
Stripping of the loaded organic phase of 7 with simulated
spent tankhouse electrolyte containing 30 g lϪ1 copper and 180
g lϪ1 sulfuric acid was found to be virtually quantitative. Fig. 5
shows equilibrium isotherms for 7 with McCabe–Thiele38
constructions for the recovery of copper from a 30 g lϪ1 copper
feed and stripping with 30 g lϪ1 copper and 180 g lϪ1 sulfuric
acid. The McCabe–Thiele data were modelled for a process
based on 2 stages of extraction and 2 stages of stripping and
assuming 90 and 95% stage efficiencies in extraction and
stripping respectively. The computed model indicates that
copper recoveries in excess of 93% may be expected, while
producing a regenerated electrolyte containing 45 g lϪ1 copper
from which the copper metal value may then be recovered in a
conventional electrowinning process.
The stability of the solvent extractant to the basic conditions
of the Escondida process is of paramount importance. A pre-
liminary assessment of the robustness of the diazopyrazolone 5
was performed by contacting separate samples of a 0.056 M
solution of 5 in toluene with equal volumes of 34 and 3.4 g lϪ1
(19 and 1 × excess of copper) ammoniacal copper feeds at room
temperature over a period of 23 days. Analysis of both the
aqueous and organic phases using ICP-AES showed that,
within experimental error, the amount of copper extracted did
not decrease over this time period, demonstrating that the
ligand does not undergo chemical degradation and is stable to
prolonged contact with ammoniacal feeds.
as their hydrazone tautomers, both in solution and the solid
state.
Metal complexes
Infrared bands at ca. 1650 and 3400 cmϪ1 associated with the
keto–hydrazone tautomers of the ligands are lost on formation
of the metal complexes (8–15, Scheme 1). The FAB mass
spectra for the complexes all show molecular ions associated
with the neutral ML2 or ML3 complexes. In addition, the
copper complexes 8 and 12 show lower intensity high mass
peaks attributed to Cu2L3 and Cu2L2 for 8 and Cu3L4, Cu3L3
and Cu2L2 for 12. In view of the satisfactory analytical data
obtained for these compounds, it is assumed these species are
produced in the mass spectrometer.
The colour changes associated with complex formation can
be used to monitor the displacement of metals from the ligand
in solvent extraction experiments (vide infra).
Solvent extraction
The strength of extractants that operate on a “pH-swing”
controlled process of the type shown in eqn. (2) can be assessed
by determining their pH values (the pH values associated with
₁
₂

50% theoretical loading of a particular metal). Their pH values
₁
₂

(Fig. 3) of ca. 3.7 are higher than those for the “strong”
1242
J. Chem. Soc., Dalton Trans., 2001, 1239–1245

View Article Online
Fig. 5 McCabe–Thiele constructions for the recovery of copper with 7 (1.0 M in Orfom SX7). All scales in g lϪ1
.
Table 2 Selected bond lengths (Å) in the chelate rings of the
complexes of 1
tautomer (Fig. 2: A) leading to the more delocalised system
shown in Scheme 1.
The fused heterocycle and the conjugation in the chelate ring
restrict the N ؒ ؒ ؒ O bites that can be offered to the various metal
ions. The largest bite distances and angles (Table 1 of ESI
supplementary material), in the zinc complex, are significantly
smaller than those required to approach a tetrahedral geom-
etry, while those for the other metals are significantly greater
than ideal values for orthogonal coordination sites in these
metals. The bite angles in [Co(1 Ϫ H)3] are large because the
covalent radius of low spin Co() is smaller than for the other
metals and in order to achieve the shorter Co–N and Co–O
bonds (Table 1 of ESI) the metal ion is located closer to the
centroid of the chelate ring.
Co
Ni
Cua
Zn
M–O5A
M–O5B
M–N42A
M–N42B
M–X
1.909(4)
1.898(4)
1.962(4)
1.954(5)
1.980(5)b
1.959(4)c
1.335(7)
1.342(7)
1.345(7)
1.269(6)
1.301(6)
1.285(6)
2.0126(13)
2.0164(13)
2.0877(17)
2.0824(17)
2.1017(15)d
2.1269(14)d
1.359(2)
1.358(2)
—
1.930(3)
1.928(3)
1.972(3)
1.974(3)
—
1.9394(14)
1.9483(14)
2.028(2)
2.009(2)
—
—
—
C4A–N41A
C4B–N41B
C4C–N41C
N41A–N42A
N41B–N42B
N41C–N42C
1.344(5)
1.337(5)
—
1.294(4)
1.311(5)
—
1.340(2)
1.347(3)
—
1.297(2)
1.293(2)
—
1.283(2)
1.284(2)
—
The coordination of more than one diazopyrazolone unit
to a metal centre may result in ligand–ligand contacts that
significantly influence their planarity and consequently the
conjugation between the aromatic groups. The pseudo-tetra-
hedral arrangement favoured by divalent zinc allows the
maximum separation between the bulky t-butylphenyl group
and the N-phenyl substituent on neighbouring ligands. As a
consequence the phenyl rings are more nearly coplanar with the
best plane through the pyrazolone and chelate ring (Table 2 of
ESI), and the torsion angles on the diazo bond are close to 180Њ.
The patterns of distortion of the diazopyrazolone ligand are
more complicated in the tris complex [Co(1 Ϫ H)3] and in the
cis-octahedral complex, [Ni(1 Ϫ H)2(MeOH)2], although in all
cases the configuration about the diazo bond remains close to
planar (Table 2 of ESI), presumably because the most stable
form of the six-membered chelate ring is close to planar.
Inter-ligand repulsion terms will have a significant influence
on the stability of the copper complexes. In all structures of
the 1 : 2 complexes with the “strong” phenolic oxime ligands
(Fig. 1) the N2O22Ϫ donor set is planar.39–43 In all but one case41
the copper atom is located on an inversion centre and the
CuN2O2 unit is therefore perfectly planar. Planarity is also
observed in all four-coordinate copper() complexes of
unhindered bidentate salicylaldimine ligands36 (Fig. 7,
Table 3).
a Similar values were found in the copper() complex of 2: Cu–O
1.935(3), Cu–O 1.920(3), Cu–N 2.000(4), and Cu–N 1.994(4) Å.
b,c Values for the third diazopyrazolone donors O5C and N42C.
d Values for the coordinated methanol atoms O1N and O1M.
A series of crystal structure determinations was carried out
in an attempt to define why the azopyrazolones are significantly
weaker complexing agents than the structurally related phenolic
oximes. The copper complexes of both 1 and 2 show small, but
significant, deviations from planarity with the O, N, Cu planes
associated with the bidentate ligands inclined at 22.7 and 18.4Њ
respectively (Fig. 6). The zinc() complex of 1 is pseudo-
tetrahedral with the chelate planes inclined at 80.7Њ, while the
nickel complex is pseudo-octahedral with two cis-methanol
molecules and the two azopyrazolone chelates approximately
mutually perpendicular with the oxygen atoms trans to each
other (Fig. 6). The cobalt complex separates under these con-
ditions as the trivalent pseudo-octahedral complex in which the
three azopyrazolone chelates have a mer, mer arrangement of
their three oxygen and three nitrogen donors. As might be
expected bond lengths in the coordination spheres are shortest
for low spin cobalt(), and longest for high spin nickel(), see
Table 2. In each complex the azopyrazolone nitrogen to metal
bonds are longer than the oxygen to metal bonds.
The formation of metal complexes is accompanied by sig-
nificant changes in bond lengths of the diazopyrazolone
component of the chelate ring. The N–N bonds (Table 2) are all
significantly shorter than those found30–37 in six uncomplexed
diazopyrazolones [mean 1.317(5) Å], whilst the C–N (pyr-
azolone ring to azo group) bonds are longer than those in the
free ligands [mean 1.308(4) Å]. These changes are consistent
with the deprotonation and complexation of the hydrazone
In contrast, the azopyrazolone complexes 8 and 12 show
significant “tetrahedral” distortions that can be ascribed to
unfavourable contacts between the azo-phenyl group and the
pyrazolyl-phenyl substituent on neighbouring ligands. The
closest contact between the ligands in 12 is 2.77 Å between
H13A (on the pyrazolone N-phenyl ring) and H48B (on the azo
phenyl ring), and between H50A (of the p-t-butyl group) and
H15B (on the N-phenyl ring). Even though the t-butyl group is
a para substituent on the phenylazo unit and is far removed
J. Chem. Soc., Dalton Trans., 2001, 1239–1245
1243

View Article Online
Fig. 6 Structures of the complexes [Cu(1 Ϫ H)2] 8 , [Zn(1 Ϫ H)2] 9, [Co(1 Ϫ H)3]ؒ3MeOH 10, [Ni(1 Ϫ H)2(MeOH)2]ؒ2MeOH 11, and [Cu(2 Ϫ H)2]
12, showing atom labelling schemes used in the tables. Hydrogens and non-coordinated methanol molecules in 10 and 11 have been omitted for
clarity.
apart and increasing the dihedral angle between the chelate
rings by almost 5Њ.
The structural studies imply that it is the bulk of the pendant
groups in the pyrazolone ligands which underlies the relative
weakness of these as extractants for copper(), making it
difficult for 2 : 1 complexes to assume a planar structure. The
2Ϫ
planar N2O2 donor set preferred by copper() is much more
readily provided by the phenolic oxime ligands which form the
intramolecular hydrogen-bonded pseudomacrocyclic unit
shown in Fig. 7. In contrast to the azopyrazolones, these
interligand contacts favour the presentation of a planar donor
set to the complexed metal ion.
Fig. 7 The pseudomacrocyclic structure of the copper complexes of
phenolic oximes, 16, contrasting with a related salicylaldimine complex,
17.
Conclusion
from the copper atom it has a significant influence on the metal
centre and accounts for the deviation from planarity being
greater in 8 than in 12 (see Table 2 of ESI), prising the ligands
Azopyrazolone ligands of the types described above are good
candidates for extractants to recover copper from ammoniacal
1244
J. Chem. Soc., Dalton Trans., 2001, 1239–1245

View Article Online
12 S. Miyazaki, H. Mukai, S. Umetani, S. Kihara and M. Matsui,
Table 3 Deviations from planaritya of the N2O2 donor sets in the
2Ϫ
copper() complexes of 1 and 2 and related bidentate ligands
Inorg. Chem., 1989, 28, 3014.
13 I. Guiguemde, B. A. Diantouba, D. Lakkis, G. J. Goetz-Grandmont
and J. P. Brunette, Analysis, 1996, 24, 318.
14 S. A. Pai, K. V. Lohithakshan, P. D. Mithapar, S. K. Aggarwal and
H. C. Jain, Radiochim. Acta, 1996, 73, 83.
15 S. Umetani and M. Matsui, Anal. Chem., 1992, 64, 2288.
16 P. Thakur, K. C. Dash, M. P. L. Reddy, R. Luxmi Varma, T. R.
Ramamohan and A. D. Damodaran, Radiochim. Acta, 1996, 75,
11.
17 J. March, Advanced Organic Chemistry, Wiley, New York, 3rd edn.
1985, pp. 804, 805.
18 W. Mickler and E. Uhlemann, Sep. Sci. Technol., 1993, 28, 1913.
19 Kirk-Othmer Encyclopædia of Chemical Technology, eds. M. Grayson
and D. Eckroth, Wiley, New York, 3rd edn., 1978, vol 6,
pp. 819–868.
[Cu(1 Ϫ H)2] [Cu(2 Ϫ H)2] [Cu(16 Ϫ H)2]b [Cu(17 Ϫ H)2]c
Cu
N
N
O
0.014
0.277
0.273
Ϫ0.282
Ϫ0.282
0.013
0.223
0.223
Ϫ0.231
Ϫ0.229
0
0
0
0
0
0
0
0
0
0
O
a Distances/Å from the best plane through the CuN2O2 unit. b Mean
displacement/Å for examples39–43 of phenolic oxime ligands (Fig. 1,
t
R1 = H, R = H, Cl or Bu). c Mean displacement/Å for examples44 of
aniline salicylaldimine ligand 17.
20 F. A. Snavely, W. C. Fernelius and B. P. Block, J. Am. Chem. Soc.,
1957, 79, 1028.
21 F. A. Snavely, B. D. Krecker and C. G. Clark, J. Am. Chem. Soc.,
1959, 81, 2337.
22 D. F. Evans, J. Chem. Soc., 1959, 2003.
23 K. D. Bartle, D. W. Jones and S. Maricic, Croat. Chem. Acta, 1968,
40, 227.
24 E. A. Boudreaux and C. N. Mulay, Theory and Applications of
Molecular Paramagnetism, Wiley, London, 1976.
25 D. E. Butler and H. A. DeWald, J. Org. Chem., 1971, 36, 2542.
26 G. M. Sheldrick, SHELXTL version 5, Siemens Analytical X-ray
Instruments, Madison, WI., 1995.
pregnant leach solutions. Good copper transfer efficiency can
be obtained; they are relatively weak extractants and are very
effiiciently stripped by acidic spent tankhouse electrolyte, whilst
the buffering effect of the ammoniacal feed solution ensures
good loading of copper into the organic phase in the extract
stages. They appear to be more resistant to chemical degrad-
ation than β-diketone ligands. One feature, which may restrict
their commercialisation, is their intense colours and good
dyeing properties. Careful handling on a plant will be needed to
prevent staining.
27 A. Altomare, G. Cascarano, C. Giacovazzo and A. Guagliardi,
J. Appl. Crystallogr., 1993, 26, 343.
28 P. T. Beurskens, G. Beurskens, W. P. Bosman, R. de Gelder,
S. García-Granda, R. O. Gould, R. Israël and J. M. M. Smits,
DIRDIF, Crystallography Laboratory, University of Nijmegen,
1996.
29 A. J. M. Duisenberg, J. Appl. Crystallogr., 1992, 25, 92.
30 J. A. Connor, R. J. Kennedy, H. M. Dawes, M. B. Hursthouse and
N. P. C. Walker, J. Chem. Soc., Perkin Trans. 2, 1990, 203.
31 B. Golinski, G. Reck and L. Kutschabsky, Z. Kristallogr., 1982, 158,
271. The R value for the structure reported in the paper is 13% and it
has therefore not been used in calculating mean structural
parameters referred to in the text.
Acknowledgements
We thank AVECIA for a postgraduate studentship (L. C. E.)
and postdoctoral fellowship (D. J. W.) and AVECIA and the
EPSRC for funding for the ICP-AES equipment.
References
1 Proceedings of The Paul Queneau International Symposium on
Extractive Hydrometallurgy of Copper, Nickel and Cobalt, vol. 1:
Fundamental Aspects, eds. R. G. Reddy and R. N. Weizenbach,
The Minerals, Metals and Materials Society, Warrendale, 1993,
pp. 881–910.
32 A. Whitaker, Acta Crystallgr., Sect. C, 1988, 44, 1587.
33 V. Bertolasi, P. Gilli, V. Ferretti and G. Gilli, Acta Crystallogr., Sect.
B, 1994, 50, 617.
2 C. Bouvier, G. Cote, R. Cierpiszewski and J. Szymanowski, Solvent
Extr. Ion Exch., 1998, 16, 1465.
3 O. Herreros, R. Quiroz, E. Manzano, C. Bou and J. Viñals,
Hydrometallurgy, 1998, 49, 87.
4 W. P. C. Duyvestyn and R. N. Hickman, US Pat., 5176802, 1993.
5 W. P. C. Duyvestyn and E. R. Rogier, US Pat., 4175012, 1979.
6 J. M. Sierakoski, WO Pat., 93/04208, 1993.
7 W. P. C. Duyvestyn, Australasian Institute of Mining and Metallurgy
Publication Series, Australasian Institute of Mining and Metallurgy,
Parkville, Victoria, Australia, 1995, vol. 95, ch. 13(3), pp. 29–42.
8 J. Szymanowski, Hydroxyoximes and Copper Hydrometallurgy, CRC
Press, London, 1993, pp. 67–74.
9 M. J. Virnig and G. A. Korolosky, WO Pat., 97/29215, 1997.
10 R. B. Sudderth and M. J. Virnig, WO Pat., 97/09453, 1997.
11 Y. A. Zolotov and N. M. Kuzmin, Extraction of Metals with
Acylpyrazolones, Nauka, Moscow, 1977; Chem. Abstr., 1978, 89,
B81023X.
34 A. Whitaker, Acta Crystallogr., Sect. C, 1988, 44, 1767.
35 A. Whitaker, J. Crystallogr. Spectrosc. Res., 1991, 21, 363.
36 A. Whitaker, J. Crystallogr. Spectrosc. Res., 1992, 22, 385.
37 L. G. Kuzmina, L. P. Grigoryeva, Y. T. Struchkov, Z. I. Yezhkova,
B. Y. Zaitsev, V. A. Zaitseva and P. P. Pronkin, Khim. Geterotsikl.
Soedin., 1985, 6, 816.
38 W. McCabe and E. Thiele, Ind. Eng. Chem., 1925, 17, 605.
39 M. A. Jarski and E. C. Lingafelter, Acta Crystallogr., 1964, 17,
1109.
40 P. L. Orlioli and E. C. Lingafelter, Acta Crystallogr., 1964, 17,
1113.
41 Z. Kangjing, Z. Chengming, E. Xing and Y. Chengye, Kexue
Tongbao, 1985, 30, 1484.
42 L. Drummond, PhD Thesis, Polytechnic of North London, 1989.
43 A. Zissimos, PhD Thesis, University of North London, 1989.
44 L. Wei, R. M. Stogsdill and E. C. Lingafelter, Acta Crystallogr.,
1964, 17, 1058.
J. Chem. Soc., Dalton Trans., 2001, 1239–1245
1245