Magnetic structure and properties of Cu3(OH)4SO 4 made of triple chains of spins s=1/2

Article abstract of DOI:10.1016/S0022-4596(02)00081-6

Cu₃(OH)₄SO₄, obtained by hydrothermal synthesis from copper sulfate and soda in aqueous medium, is isostructural with the corresponding antlerite mineral, orthorhombic, space group Pnma (62), with a=8.289(1) b=6.079(1) and c=12.057(1)A, V=607.5(2) A³, Z=4. Its crystalline structure has been refined from X-ray single crystal and powder neutron diffraction data at room temperature. It consists of copper (II) triple chains, running in the b-axis direction and connected to each other by sulfate groups. The magnetic structure, solved from powder neutron diffraction data at 1.4K below the transition at 5K evidenced by susceptibility and specific measurements, reveals that, inside a triple chain, the magnetic moments of the copper ions (μ_B=0.88(5) at 1.4K) belonging to outer chains are oriented along the c-axis of the nuclear cell, with ferromagnetic order inside a chain and antiferromagnetic order between the two outer chains. No long-range magnetic order is obtained along the central chain with an idle spin behavior.

Full text of DOI:10.1016/S0022-4596(02)00081-6

Journal of Solid State Chemistry 170 (2003) 255–264
Magneticstructure and properties of Cu (OH)4SO4 made of
3
triple chains of spins s=1/2
S. Vilminot,a,* M. Richard-Plouet,a G. Andre,b D. Swierczynski,a M. Guillot,a
F. Bouree-Vigneron,b and M. Drillona
a
´
Groupe des Materiaux Inorganiques, IPCMS, UMR 7504 CNRS-ULP, 23 rue du Loess, 67037 Strasbourg Cedex, France
b
´
Laboratoire Leon Brillouin, CEA-CNRS, CEA-Saclay, 91191 Gif-sur-Yvette Cedex, France
Received 18 June 2002; received in revised form 12 September 2002; accepted 17 September 2002
Abstract
Cu3(OH)4SO4, obtained by hydrothermal synthesis from copper sulfate and soda in aqueous medium, is isostructural with the
corresponding antlerite mineral, orthorhombic, space group Pnma (62), with a=8.289(1) b=6.079(1) and c=12.057(1) A,
V=607.5(2) A3, Z=4. Its crystalline structure has been refined from X-ray single crystal and powder neutron diffraction data at
room temperature. It consists of copper (II) triple chains, running in the b-axis direction and connected to each other by sulfate
groups. The magnetic structure, solved from powder neutron diffraction data at 1.4 K below the transition at 5 K evidenced by
susceptibility and specific measurements, reveals that, inside a triple chain, the magnetic moments of the copper ions (mB=0.88(5) at
1.4 K) belonging to outer chains are oriented along the c-axis of the nuclear cell, with ferromagnetic order inside a chain and
antiferromagnetic order between the two outer chains. No long-range magnetic order is obtained along the central chain with an idle
spin behavior.
r 2002 Elsevier Science (USA). All rights reserved.
Keywords: Antlerite; Copper hydroxysulfate; Powder neutron diffraction; Magnetic structure; Low-dimensional magnetism; Idle spin
1. Introduction
for natural samples and nothing was known about the
magneticproperties, stimulating a more complete study
of syntheticantlerite. Let us note that the presence of
triple chains of copper (II) (spin-1/2), whose interest is
currently under scope since they may show in some
specific cases an energy gap in the spin excitation
spectrum [4], made such a system very exciting from a
magneticpoint of view.
Hydrothermal synthesis is presently used in our group
in Strasbourg to prepare transition metal (Ni, Co, Fe)
silicates with clay structure such as 1:1 and 2:1
phyllosilicates of transition metals. These compounds
exhibit a ferromagneticbehavior at low temperature
[1,2]. Starting from aminopropyltriethoxysilane instead
of tetraethoxysilane, a family of hybrid organic–
inorganicmaterials has been isolated and their spectro-
scopic, thermal, optical and magnetic properties are
under study [3]. In the course of this study, attempts to
prepare copper phyllosilicates always failed as far as,
like for the other transition metals, the corresponding
acetate was used. Starting from copper sulfate, small
amount of green single crystals was evidenced as copper
hydroxysulfate, Cu3(OH)4SO4, corresponding to the
antlerite mineral. If structural data were already
reported for this compound, they were mainly obtained
2. Experimental
2.1. Synthesis
Antlerite has been prepared by hydrothermal treat-
ment of an aqueous suspension obtained from copper
sulfate pentahydrate, CuSO4.5H2O and soda, NaOH
in molar proportions Cu/Na/H2O=1/1.333/2222. The
reaction was performed under autogenous pressure in
bombs of 125 cm3, with a ca. 1/3 filling ratio. The
precursor suspension was stirred for 1 h before sealing in
the reactor that was heated at 1701C for 6 days. At the
*Corresponding author. Fax: +33-3-88-10-72-47.
E-mail address: vilminot@ipcms.u-strasbg.fr (S. Vilminot).
0022-4596/03/$ - see front matter r 2002 Elsevier Science (USA). All rights reserved.
PII: S 0 0 2 2 - 4 5 9 6 ( 0 2 ) 0 0 0 8 1 - 6

256
S. Vilminot et al. / Journal of Solid State Chemistry 170 (2003) 255–264
end of the reaction, antlerite was obtained in the form
of needle-like green single crystals. Attempts to pro-
mote crystal growth by introducing seeds in the
suspension before hydrothermal treatment were unsuc-
cessful.
3. Results and discussion
3.1. Structural characterization
3.1.1. X-ray powder diffraction
For neutron powder diffraction measurements, the
same synthesis was used starting from copper sulfate
pentahydrate, CuSO4 5H2O, NaOD solution and heavy
water D2O in order to substitute deuterium for
hydrogen that exhibits a high incoherent diffusion
factor resulting in an increase of the background. Using
the same thermal treatment, a powder was obtained
instead of single crystals. The powder was filtered and
acetone washed. Such washing induces the formation of
a small amount of Na2SO4.
X-ray powder diffraction pattern of the title com-
pound confirms the formation of antlerite, Cu3(OH)4
SO4, in agreement with JCPDS file 07-407.
3.1.2. Thermal analysis
Thermal analysis reveals a weight loss within two
steps (Fig. 1): the first one, between 4651C and 4901C,
corresponds to a water loss according to the following
reaction:
Cu3ðOHÞ4SO4-2CuO þ CuSO4þ2H2O
2.2. Characterization
the second one, between 7301C and 7901C, corresponds
to the decomposition of copper sulfate according to the
following reaction:
Thermal analysis was performed with a TG-DTA92
Setaram apparatus, at a 3.51C/min heating rate, under
air. Infrared spectra were recorded on a ATI Mattson
spectrometer using the KBr pellet technique.
CuSO4-CuO þ SO3:
The observed weight losses do not differ by more than
1% from the calculated ones. X-ray diffraction patterns
recorded at different steps of the thermal evolution
confirm the above reactions.
X-ray powder diffraction patterns were recorded by
means of a D5000 Siemens diffractometer (CuKa1,
1.5406 A). For single-crystal data collection, a single
set of 180 frames, 20 s/frame, f scan of 11/frame (each
frame measured twice), crystal detector distan-
ce=30 mm, y=01, k= 01 was collected on a Nonius
Kappa CCD diffractometer at the Service Commun de
Rayons X, Universite Louis Pasteur, Strasbourg.
Structure refinement was done starting from the results
of Hawthorne et al. [5] using SHELXS-86.
The neutron diffraction experiments were performed
at the Laboratoire Leon Brillouin (CEA Saclay) using
the 3T2 and G4.1 diffractometers. The high-resolution
powder diffractometer 3T2 (l=1.2252 A, 61o2yo1261)
was used for the refinement of the nuclear structure at
300 K and the multi-detector (800 cells) G4.1
(l=2.4266 A) for the determination of the magnetic
structure and the thermal evolution study of the low-
temperature patterns. Therefore, 11 diffraction patterns
were recorded in the 2y range 101–89.91, at different
temperatures between 1.4 and 6 K. The powder sample
was set in a cylindrical vanadium can and held in a
liquid helium cryostat. Nuclear and magnetic structures
were refined using the FULLPROF [6] program. The
nuclear scattering lengths (bCu=0.7718 ꢀ 10ꢁ12 cm,
bS=0.2847 ꢀ 10ꢁ12 cm, bO=0.5803 ꢀ 10ꢁ12 cm, bD=
0.6671 ꢀ 10ꢁ12 cm and bH=ꢁ0.3739 ꢀ 10ꢁ12 cm) and
copper magnetic form factors were those included in
this program.
3.1.3. Infrared spectroscopy
For deuteriated and hydroxylated compounds, the
infrared spectrum (Fig. 2) reveals the presence of four
band families, around 3500, 1100, 800 and 600 cmꢁ1
.
Whereas the intensity and position of the 1100 and
600 cmꢁ1 bands are not affected by deuteriation, this is
no more the case for the other ones that can be therefore
attributed to vibrations involving OH groups. Two
sharp bands are observed at around 3573(s) and
3485(s) cmꢁ1 and are attributed to OH stretching
vibrations. For D partially exchanged sample, they are
shifted to 2637(s) and 2576(s) cmꢁ1, respectively. How-
ever, the bands related to OH vibrations are still present
in relation with a non-complete substitution due to the
presence of H2O coming from copper sulfate pentahy-
drate. The presence of two bands with similar intensities
lead to the conclusion that two kinds of OH groups in
equal number are involved. This aspect will be further
discussed from neutron diffraction data. Numerous
vibrations appear between 880 and 740 cmꢁ1, at
873(m), 852(m), 785(sh) and 754(s) cmꢁ1. Owing to the
strong decrease of their intensities after H/D exchange,
they are also attributed to vibrations involving OH
groups. The n3 and n4 vibrations of the sulfate group
appear as broad and strong bands around 1100 and
600 cmꢁ1, respectively. Due to the distortion from
regular tetrahedral symmetry, the degenerate vibrations
split into three bands, at 1153(m), 1109(s), 1070(s) cmꢁ1
and 634(s), 613(s) and 602(s) cmꢁ1 for n3 and n4;
respectively. From point group considerations [7], the
Magneticsuespctibility measurements were per-
formed in the range 2–300 K by means of a Quantum
Design MPMS-XL SQUID magnetometer. Specific heat
measurements were carried out with a home-made
equipment using a quasi-adiabaticmethod.

S. Vilminot et al. / Journal of Solid State Chemistry 170 (2003) 255–264
257
20
0
0
-20
-40
Cu3(OH)4 SO4
2CuO + CuSO4
-20
-40
-60
3CuO
0
200
400
600
800
1000
T (˚C)
Fig. 1. DTA–TGA traces of antlerite recorded under air with a 3.51C/min heating rate.
band at around 551(s) cmꢁ1 for the D exchanged sample
is attributed to a vibration involving OD groups.
160
140
120
100
80
Cu3(OH)4SO4
3.1.4. X-ray single-crystal structure
The structural findings from the single-crystal X-ray
diffraction study are very similar to those previously
reported for natural antlerite [5]. The lattice parameters,
atomicpositions, interatomicdistances and bond angles
are given in Tables 1–3 where OHi atoms refer to oxygen
atoms of the hydroxyl groups.
669
785
987
3573
3485
873
1153
1109
513
602
852
754
1070
60
613
634
The structure may be viewed as a ribbon running
along the b-axis of three connected chains, each one
consisting of edge sharing Cu(O,OH)6 octahedra
(Fig. 3). Such chains are connected to each other by
SO24ꢁ anions as depicted in Fig. 4. A similar structure
was also evidenced in the case of szenicsite, Cu3(OH)4
MoO4 [8]. Whereas SO4 tetrahedra are nearly regular
with a mean S–O bond length of 1.481 A, the octahedra
around both crystallographically independent metal
ions are strongly distorted, due to the Jahn–Teller
distortion usually observed for octahedrally coordinated
Copper(II). The equatorial bonds are quite similar
(1.977–2.005 A, 0.7% maximum deviation from the
mean value) around Cu1 ions, while the apical bonds
differ significantly (2.306–2.552 A, 5% deviation). The
reverse situation is observed for Cu2 where the
deviations are stronger (1.925–2.043 A, 3% deviation)
in the equatorial plane compared to the apical direction
(2.361–2.428 A, 1.4% deviation). Considering in more
detail the triple chains (Fig. 3), differences appear
between the Cu–Cu distances. Whereas all Cu1–Cu1
distances are all the same, it is no more the case for Cu2–
Cu2 ones which show alternating distances (3.00 and
3.07 A) along the b-axis. This is related to the fact that
SO4 ions share corners with oxygen belonging to
octahedra involving 3.00 A Cu–Cu distances. On the
other hand, copper–copper distances between chains are
40
4000
3000
2000
1000
0
Wavenumber (cm-1)
70
60
50
40
30
20
10
0
Cu3(OHD)4SO4
748
792
842
879
823
684
652
634
3573
3479
507
551
2637
2576
983
1147
612
590
OD
OH
1064
1109
4000
3000
2000
1000
0
Wavenumber (cm-1)
Fig. 2. Infrared spectra of H and D antlerites.
real symmetry is probably C2v for sulfate group.
Moreover, a line appears at 987(w) cmꢁ1 attributed to
n1 vibration, normally IR inactive for Td symmetry but
becoming IR active in C2v symmetry. After H/D
exchange, any significant shift is not observed and the
intensities are of same magnitude. The presence of a new

258
S. Vilminot et al. / Journal of Solid State Chemistry 170 (2003) 255–264
larger (3.18 and 3.27 A). For more details on the
structure description, see Ref. [5].
disposition of copper ions appears to be in between one-
dimensional (1D) and two-dimensional (2D) systems.
Therefore, unusual magneticproperties oculd be ex-
pected in relation with the change of dimensionality.
3.1.5. Neutron powder diffraction at room temperature
The results of the structure refinement from neutron
powder diffraction measurements at 300 K are summar-
ized in Tables 4 and 5. They allow to evidence a high D
content in the sample, around 94%. The O–D bond
lengths (Table 3) do not significantly differ from each
other with a mean value of 0.960 A. In all cases, the
next-nearest oxygen atom is O3, but the D?O3
distances are quite long, 2.473, 2.481 and 2.183 A for
D1, D2 and D3, respectively. Therefore, it seems that no
hydrogen bonding is involved in the case of antlerite. If
it was the case, one would expect a more important
distortion of the SO4 tetrahedron, in particular with a
much longer S–O3 bond length. However, the differ-
ences between the D?O3 distances can explain the
infrared results, with two kinds of OH groups in equal
number. The observed and calculated diffraction pat-
terns are shown in Fig. 5.
3.2. Magnetic properties
3.2.1. Magnetic susceptibility
Magnetic susceptibility of the title compound
recorded at H=100 Oe follows above 40 K a Curie-
Weiss law w ¼ C=ðT ꢁ yÞ ¼ 1:28=ðT ꢁ 1:69Þ emu/molꢁ1
(Fig. 6), where C=Ng2mB2 S(S+1)/3k is the Curie con-
stant, assuming a spin only value with S=12 for
copper(II). The positive y value is characteristic of a
ferromagneticbehavior in the high-temperature regime.
From the Curie–Weiss expression, we can deduce a g
value of 2.13 as expected. The slight increase of the wT
product from room temperature down to 40 K (Fig. 7)
confirms the presence of ferromagnetic correlations
while the maximum of w at 6.5 K together with the low
temperature wT variation agrees with the presence of
competing antiferromagnetic interactions. Finally, at
very low temperature, a small downturn of w is observed
at T=5 K (inset of Fig. 6) which points to a transition
towards a three-dimensional (3D) long-range order.
Such a manifestation is quite unusual in antiferromagic
systems at the 1D/3D transition, and so specific heat
measurements have been performed to confirm that
result.
Copper octahedra are connected to each other giving
rise to ribbons running in the b-axis direction. Such
Table 1
CCD X-ray data collection and structure refinement of synthetic
antlerite
a (A)
b (A)
8.289(1)
6.079(1)
Crystal size (mm3)
Total reflexions
12.057(1) Unique reflections
0.10 ꢀ 0.15 ꢀ 0.22
2870
962
c (A)
V(A3)
607.5(2)
Pnma
660
Unique jIoj43sðIoÞ 919
3.3.2. Specific heat
Space group
F(000)
Rint
0.0362
0.0335
0.0446
1.05
The temperature dependence of the specific heat,
illustrated in Fig. 8, shows a characteristic l-type
anomaly at T=5.3 K, indicating a phase transition
towards a 3D state. After deducing the phonon
contribution, namely a T3 contribution in the low-
temperature range, and taking into account a Tꢁ2 tail
R
wR
Dcalc. (g/cm3) 3.877
m (mmꢁ1
Z
)
10.73
4
Goodness-of-fit
P
P
R ¼ ðjFo ꢁ jFcÞ= jFoj;
wR
ꢀ
ꢁ2
P
wðjFojꢁjFc =w ¼ 1=ðs2ðFÞ þ 0:0032F2Þ:
Table 2
Final atomicparameters for syntheticantlerite from X-ray data
a
a
a
a
a
a
a
x/a
y/b
z/c
Uiso
U11
U22
U33
U12
U13
U23
Cu1
Cu2
S
0.00472(4)
0.28997(4)
0.13039(8)
0.2616(3)
0.1984(4)
0.0313(2)
0.2807(3)
0.7012(3)
0.0464(3)
0.362(6)
0.25
0.00129(3)
0.12592(3)
0.36422(6)
0.2829(2)
0.4775(2)
0.3483(2)
0.0250(2)
0.7782(2)
0.1017(2)
0.978(5)
93(2)
95(2)
78(3)
131(3)
109(3)
91(4)
100(4)
111(3)
119(4)
101(3)
115(3)
74(4)
0
ꢁ34(2)
ꢁ22(1)
ꢁ2(3)
39(9)
0
12(1)
0.00288(4)
0.25
0.25
ꢁ3(1)
0
0
O1
O2
124(11)
147(11)
131(7)
96(10)
110(10)
98(7)
127(11)
234(14)
127(9)
159(12)
165(12)
118(10)
210(13)
217(13)
151(8)
147(12)
99(11)
198(10)
96(11)
91(11)
85(10)
0
0
0.25
0
ꢁ28(7)
0
ꢁ61(10)
22(7)
0
ꢁ21(8)
0
O3
0.0480(3)
0.25
0.25
OH1
OH2
OH3
H1
137(11)
132(11)
154(11)
ꢁ6(9)
19(9)
9(8)
0
0
3(5)
0.5067(2)
0.25
0.25
ꢁ10(5)
713(197)
1700b
1712(320)
H2
H3
0.274b
0.768b
0.512(11)
0.032(15)
0.664(10)
OHi atoms refer to oxygen atoms of the hydroxyl groups.
a Uij=Uij ꢀ 104 (A2), Uiso (A2) refers to results of the isotropicrefinement.
b Fixed during refinement.

S. Vilminot et al. / Journal of Solid State Chemistry 170 (2003) 255–264
259
Table 3
Interatomicdistances (A ) and angles (1) for X-ray refinement (upper part) and involving D atoms from powder neutron diffraction data at 300 K
(lower part)
S–O1
S–O2
1.464(3)
1.478(3)
1.490(2) ꢀ 2
1.481
O1–S–O2
O1–S–O3
109.6(3)
108.9(3) ꢀ 2
109.2(3) ꢀ 2
111.0(2)
96.8(1)
/O–S–OS
109.5
S–O3
S–/OS
O2–S–O3
O3–S–O3
Cu1–OH3
Cu1–OH3
Cu1–OH1
Cu1–O2
/Cu1–OS
Cu2–OH2
Cu2–OH1
Cu2–OH3
Cu2–O3
Cu2–O2
Cu2–O1
/Cu2–OS
1.977(2) ꢀ 2
2.005(2) ꢀ 2
2.306(3)
2.552(3)
2.137
OH3–Cu1–OH3
OH3–Cu1–OH3
OH3–Cu1–OH3
OH3–Cu1–OH3
OH3–Cu1–OH1
OH2–Cu2–OH1
OH2–Cu2–OH3
OH2–Cu2–O3
OH2–Cu2–O2
OH2–Cu2–O1
OH1–Cu2–OH3
OH1–Cu2–O3
OH3–Cu1–OH1
OH3–Cu1–O2
OH3–Cu1–O2
OH1–Cu1–O2
75.8(1) ꢀ 2
81.4(2) ꢀ 2
96.4(2) ꢀ 2
167(1)
102.2(1)
80.4(1) ꢀ 2
177(4) ꢀ 2
106.9(2) ꢀ 2
178(5)
1.925(2)
1.935(2)
2.041(3)
2.043(2)
2.361(2)
2.428(2)
2.122
OH1–Cu2–O2
OH1–Cu2–O1
OH3–Cu2–O3
OH3–Cu2–O2
OH3–Cu2–O1
O3–Cu2–O2
91.76(8)
90.3(2)
174(2)
95.8(2)
88.8(1)
86.2(1)
85.0(1)
91.9(1)
99.4(2)
83.9(1)
176(3)
91.69(9)
83.9(1)
91.8(1)
O3–Cu2–O1
O2–Cu2–O1
Cu1–Cu1
Cu2–Cu2
Cu2–Cu2
Cu1–Cu2
Cu1–Cu2
3.0407(5) ꢀ 2
3.0042(10)
3.0748(10)
Cu1–OH3–Cu1
Cu2–O1–Cu2
Cu2–OH1–Cu2
Cu2–O2–Cu2
Cu2–OH2–Cu2
99.6(1)
Cu1–OH1–Cu2
Cu1–OH3–Cu2
Cu1–O2–Cu2
Cu1–OH3–Cu2
96.7(1) ꢀ 2
103.6(1)
83.31(9)
108.9(1)
76.44(6)
101.9(1)
81.26(7)
106.0(1)
3.1790(12) ꢀ 2
3.2684(12) ꢀ 2
D1–OD1
D2–OD2
D3–OD3
0.955(5)
0.962(5)
0.963(3)
D1?O3
D2?O3
D3?O3
2.474(3) ꢀ 2
2.481(3) ꢀ 2
2.183(3)
OD1-D1?O3
OD2-D2?O3
OD3-D3?O3
131.4(7)
133.4(7)
138.8(7)
2
1
3.04
3.27
2
3.18
3.07
3.00
3.00
3.18
3.27
3.18
3.18
3.27
3.07
3.27
b
c
a
Fig. 3. View of a triple chain of copper octahedra (lower part) running
in the b-axis direction. Bond lengths inside the triple chain (upper
part). Cu1 atoms and octahedra are shaded.
a
Fig. 4. Projection of the antlerite structure on the ac plane.

260
S. Vilminot et al. / Journal of Solid State Chemistry 170 (2003) 255–264
Table 4
Crystallographic data and refinement characteristics at 300 and 1.4 K from neutron powder data for deuteriated compound
300 K
1.4 K
300 K
1.4 K
System
OrthorhombicNumber of parameters
Pnma (62)
58
18
Space group
a (A)
b (A)
Rp (%)
Rwp (%)
Rexp (%)
RB (%)
RF (%)
6.72
6.32
4.48
2.87
1.99
1.99
10.4
11.5
1.59
2.63
2.28
8.2574(2)
6.0529(1)
12.0099(2)
1.2252
8.2259(7)
6.0466(5)
11.9792(10)
2.4266
c (A)
Wavelength (A)
2y range and step
Reflections used/total
6–125.71 0.051
689/1033
10–89.91 0.11
76/247
Goodness-of-fit (%)
Rmagnetic(%)
5.04
Table 5
Atomic coordinates and isotropic thermal motion values at 300 K from neutron powder data for deuteriated compound (upper part). Position of the
Cu2 magneticatoms and their magneticmoment from 1.4 K data refinement (lower part)
Atom
x/a
y/b
z/c
Occupancy
B (A2)
Cu1
Cu2
S
0.0048(4)
0.2900(2)
0.1307(7)
0.2611(4)
0.1997(4)
0.0316(2)
0.2808(4)
0.7024(4)
0.0467(2)
0.3633(5)
0.3633(5)
0.2964(5)
0.2964(5)
0.5118(3)
0.5118(3)
0.25
0.0012(2)
0.1262(1)
0.3635(5)
0.2822(3)
0.4764(3)
0.3482(2)
0.0248(3)
0.7779(2)
0.1024(2)
0.9696(3)
0.9696(3)
0.7691(3)
0.7691(3)
0.6720(2)
0.6720(2)
4.0
8.0
0.81(3)
0.73(2)
0.83(9)
1.08(5)
1.28(5)
1.07(4)
0.80(5)
0.85(5)
0.71(3)
2.15(7)
2.15(7)
2.08(6)
2.08(6)
2.07(4)
2.07(4)
0.0033(3)
0.25
0.25
4.0
4.0
O1
O2
0.25
4.0
8.0
O3
0.0479(3)
0.25
0.25
OH1
OH2
OH3
H1
4.0
4.0
0.5062(4)
0.25
0.25
8.0
0.26(2)
3.74(2)
0.31(2)
3.69(2)
0.38(3)
7.62(2)
D1
H2
0.25
0.25
D2
H3
0.0125(4)
0.0125(4)
D3
Atom
x/a
y/b
z/c
Mx=My
Mz (mB)
Cu21
Cu22
Cu23
Cu24
Cu25
Cu26
Cu27
Cu28
0.2900
0.2100
0.0033
ꢁ0.0033
0.5033
0.1262
0.6262
0
0
0
0
0
0
0
0
ꢁ0.88(5)
ꢁ0.88(5)
0.88(5)
ꢁ0.2900
0.7900
ꢁ0.1262
0.3738
0.4967
0.88(5)
0.88(5)
ꢁ0.2900
0.7900
0.2900
ꢁ0.0033
0.0033
0.4967
ꢁ0.1262
0.3738
0.1262
0.88(5)
ꢁ0.88(5)
ꢁ0.88(5)
0.2100
0.5033
0.6262
for the magnetic contribution, we deduce the specific
heat of magneticorigin ( Fig. 8) [9]. Clearly, the variation
of Cp in the range 10–30 K agrees with the low-
dimensional character of the system while the peak at
5.3 K confirms the magnetic susceptibility results.
and spin space symmetries of a closed chain. This allows
to rule out the drawback encountered with linear finite
chain. Then, the thermodynamic properties are com-
puted by the usual procedure [12].
In the case under consideration, five different
exchange interactions within the triple chain have been
P
3.2.3. Magnetic analysis
An analysis of the magneticdata has been carried out
considered in the spin hamiltonian H ¼ ꢁ JaSI SJ; in
agreement with structural findings (inset of Fig. 7).
Although the distances between copper(II) ions are in
the range 3–3.27 A, no interaction was neglected,
because of the dependence of Cu–O–Cu angles and
O2ꢁ (or OHꢁ) bridging ligands on the exchange
constants. The fitting routine has been performed for a
fragment of 12 spin-12 —a further step with 15 spins was
too time consuming—giving a realistic picture of the
1
2
by assuming a model of spin Heisenberg triple chain,
which is convenient to describe the above copper(II)
compound in the paramagnetic regime. The numerical
procedure used, initiated by Bonner and Fisher [10] for
1
2
spin linear chains and further extended to more exotic
systems [11], is based on the diagonalization of the
energy matrix by taking into account the geometrical

S. Vilminot et al. / Journal of Solid State Chemistry 170 (2003) 255–264
261
Cu3(OH)4S04 300 K
1600
1400
1200
1000
800
600
400
200
0
-200
-400
0
20
40
60
80
100
120
140
2ꢀ (˚)
42000
36000
Cu3(OH)4S04 1.4 K
700
600
500
400
300
200
100
0
30000
24000
18000
16000
1
2
3
4
5
6
Temperature (K)
6000
0
-6000
-12000
10
18
26
34
42
50
58
66
74
82
90
2ꢀ (˚)
Fig. 5. Upper part: observed (dotted line) and calculated (solid line) profiles of the powder neutron pattern of antlerite obtained on the 3T2
diffractometer at 300 K with position of the Bragg reflections (short vertical lines, first line for antlerite, second line for Na2SO4) and difference
between observed and calculated profiles. Lower part: Observed (o) and calculated (solid line) profiles of the powder neutron pattern of antlerite
obtained on the G4.1 diffractometer at 1.4 K, with just below the difference between observed and calculated profiles. The short vertical lines below
the patterns mark the position of all Bragg reflections (nuclear antlerite, nuclear Na2SO4 and magneticlines). The arrow evidences the (1 0 0)
magneticline. Inset: Thermal evolution of the integrated intensity of the (1 0 0) magneticBragg peak.
thermodynamics of the infinite system. Using the Lande
factor g=2.063 determined from the room temperature
wT value, a very good agreement between theory and
experiment was obtained for the parameters J1=2.8 K,
J2=58.8 K, J3=2.6 K, J4=ꢁ9.5 K and J5=ꢁ4.6 K for
the range 8 o T o 300 K, i.e., for temperatures higher
than the 3D transition (Fig. 7). Clearly, the system may
appear overparametrized, and so several fits have been
performed to test the validity of the results. In fact, the
temperature dependence of the wT product is very
sensitive to small changes of the exchange constants,
and the set of parameters given above seems to be quite

262
S. Vilminot et al. / Journal of Solid State Chemistry 170 (2003) 255–264
0.14
0.13
0.12
0.11
0.10
0.09
140
120
100
80
0.12
0.10
0.08
0.06
0.04
0.02
0.00
2
3
4
5
6
7
8
9 10
60
40
20
0
0
20
40
60
80
100
120
140
160
Temperature (K)
Fig. 6. Magnetic susceptibility and its inverse evolutions versus temperature for antlerite. (Magnetic susceptibility measured under a 100 Oe applied
magneticfield). Inset: low-temperature domain of the w evolution exhibiting the transition.
1.4
1.2
1.0
0.8
realistic(within an error of about 10%), even if the
lowest J values are smaller than the temperature range
under consideration.
The magneticstructure in the ordered state has been
solved from the neutron powder diffraction data
collected between 1.4 and 6 K. First, the nuclear
structure has been refined starting from the room
temperature findings. No magneticocntribution is
evidenced down to 5 K, but at lower temperature one
new reflection (arrow on Fig. 5) clearly appears for
2y=16.961 (d=8.2258 A), corresponding to the (1 0 0)
Bragg peak systematically absent in Pnma space group.
Its intensity versus temperature is depicted in the inset of
Fig. 5. Other very small magneticocntributions are
observed, particularly at 2y=41.941, 50.631, 52.531 and
53.431. The magneticstructure at 1.4 K is in agreement
with the results of magnetic susceptibility. It can be
viewed as a copper(II) triple chain, consisting of
ferromagneticouter chains (Cu2 atoms), with magnetic
moments in the c-axis direction (0.88(5) mB) antiparallel
to each other (Fig. 9). As regards the central chain (Cu1
atoms), the best fit gives a zero magneticmoment on
metal ions, while a spin- 12 is expected for copper(II). The
observed and calculated diffraction patterns are shown
in Fig. 5.
J1
J2
J1
data
fit
J5
J5
J5
J3
J3
J4
J4
J1
J2
J1
05
0
100
150
200
250
300
Temperature (K)
Fig. 7. Experimental and fit data of the evolution of the wT product
versus temperature, with a scheme defining the corresponding J
coupling constants.
20
15
10
5
3.3. Discussion
The structural study reported above does not evidence
any significant difference between natural and synthetic
antlerites. A 1.7% volume increase, regularly distributed
on the three axis directions (a=8.289 A for 8.244 A,
b=6.079 A for 6.043 A and c=12.057 A for 11.987 A) in
the case of the synthetic compound could be related to
the presence of small impurity amounts in the former
one. The lack of chemical analysis on the mineral does
not allow to identify their nature. Using D2O and
NaOD, it has been possible to prepare antlerite with a
0
0
10
30
20
Temperature (K)
Fig. 8. Specific heat evolution versus temperature: ’, for experi-
mental points; O, for calculated specific heat of magnetic origin; and
&, lattice contribution.

S. Vilminot et al. / Journal of Solid State Chemistry 170 (2003) 255–264
263
singlet for the triple chain, and accordingly wT decreases
to zero. Finally, both susceptibility and specific heat
measurements reveal that at about 5 K a transition
occurs which stabilizes a long-range 3D order between
triple chains. The mechanism involved may be a classical
quantum interaction through (SO4) units or, more
likely, a dipolar coupling between ribbons which would
be supported by the presence of outer ferromagnetic
chains [13].
c
0
The zero magneticmoment obtained for Cu1 from the
neutron diffraction experiments is at first questionable.
In fact, this may be explained by some misorientation of
the magneticmoments due to the competition between
different interactions and a negligible local anisotropy.
Other magnetic structure refinements have been con-
sidered, involving antiferromagneticorder inside the
Cu1 central chains, but all lead to insignificant values of
the moment, always much lower than standard devia-
tions. Such behavior, so-called ‘‘idle-spin’’ behavior, was
already evidenced for other compounds [14–17].
It is known that the exchange coupling in copper(II)
compounds is closely related to the bridging angle
Cu–O–Cu. Thus, it has been demonstrated [18] for a
series of di-m-hydroxo-bridged copper(II) compounds
that positive J values are observed for a bridge angle f
lower than 97.51, whereas higher f values favor a
negative exchange constant.
a
Fig. 9. Projection on the ac plane showing the orientations of the
magneticmoments.
high D content, around 94%, as evidenced by neutron
data refinement.
The infrared spectrum reveals the presence of two
bands at 3573 and 3485 cmꢁ1 attributed to OH
stretching vibrations of OH groups not involved in
hydrogen bonding. They can be assigned to (H1 + H2)
and H3 according to their relative intensities and to the
corresponding occupation numbers of hydrogen atoms,
namely 4c positions for H1 and H2, 8d position for H3.
Considering the structure refinement from neutron data,
it appears that the O–H/D distances are very similar,
0.955, 0.962 and 0.963 A for H1, H2 and H3,
respectively. Therefore, we have to consider the possible
influence of O3 atoms on the O–H stretching vibration
by varying its force constant. For the sulfate group, if
the distortion from Td symmetry is rather low, the small
observed deviations lower the local symmetry to C2v, as
expected from the observation of three bands for n3 and
n4 as well as from the presence of the n1 vibration band.
The proposed magneticstructure in the ordered state
agrees well with the results of magneticsusecptibility.
Each triple chain exhibits an antiferromagnetic ground
state, but it appears that interactions of opposite signs
concur to the overall behavior. Thus, the magnetic
moments of the outer chains (Cu2) order ferromagne-
tically (J1 and J2 interactions) and are aligned antipar-
allel to each other, while the moments of the central
chain (Cu1) are antiferromagnetically coupled. Further-
more, these individual chains are connected through the
J3 and J4 exchange couplings giving rise to a strongly
correlated ribbon. Note that the large ferromagnetic
interaction J1, which favors a dimerized structure in the
Cu2 chains, explains the variation of the wT product in
the range 40–300 K. At lower temperature, the other
(positive and negative) interactions promote a spin
For the title compound, the pertinent exchange
pathways within copper(II) ribbons call for the follow-
ing remarks:
*
Cu1 ions are bonded to each other through two
hydroxo (OH3) with a bridging angle f=99.61,
favoring an antiferromagneticcoupling ( J5).
Cu2 ions are bonded to each other through an
*
hydroxo and an oxygen atom. Along the Cu2 chain,
we have alternatively (O1+OH1) and (O2+OH2)
bridging units. The f angle values are less than 901
for bridging oxygen atoms and larger than 97.51 for
both hydroxo bridges. The ferromagneticinteractions
J1 and J2 entail that the dominant pathways are
through oxygen atoms, even if the Cu2–O bond
lengths (2.43 and 2.36 A) are significantly larger than
the Cu2–OH ones (1.94 and 1.93 A),
*
Cu1 and Cu2 chains are alternatively bonded through
hydroxo and oxygen bridges. As above, the presence
of two hydroxo bridges implies an antiferromagnetic
coupling (J4) whereas mixed oxygen and hydroxo
bridges favor a ferromagneticcoupling ( J3)
4. Conclusion
Antlerite, Cu3(OH)4SO4, obtained by hydrothermal
synthesis, has been characterized by its structural and
magneticproperties. Triple hcains of edge sharing

264
S. Vilminot et al. / Journal of Solid State Chemistry 170 (2003) 255–264
copper octahedra running in the b-axis direction are
connected to each other through sulfate tetrahedra.
Both neutron diffraction study and magnetic suscept-
ibility fit agree with a model involving the same
interactions between copper ions. The results have also
been related to the Cu–O–Cu angle values. It is to be
noted that such compound may be viewed as a new class
of low-dimensional system, in between 1D and 2D
systems, which may exhibit unusual quantum effects.
Other members of the series are investigated, as for
instance brochantite, Cu4(OH)6SO4.
Data, Version 3.3 Leon-Brillouin Laboratory, C. E. A.-Saclay,
France, 1997.
[7] K. Nakamoto, Infrared and Raman Spectra of Inorganic and
Coordination Compounds, Wiley-Interscience, New York, 1986,
p. 248.
[8] J. Stolz, T. Armbruster, N. Jb. Miner. Mh. 6 (1998) 278–288;
P. C. Burns, Mineral. Mag 62 (1998) 461–469.
[9] R.L. Carlin, Magnetochemistry, Springer, Berlin, Heidelberg,
1986, p. 165.
[10] J.C. Bonner, M.E. Fisher, Phys. Rev. A 135 (1964) 640.
[11] R. Georges, J.J. Borras, E. Coronado, J. Curely, M. Drillon,
Magnetism: molecules to materials, Models and Experiments,
In: J. Miller, M. Drillon, (Eds.), Wiley VCH, Weinheim, 2001,
pp. 1–47.
[12] M. Drillon, E. Coronado, R. Georges, J.C. Gianduzzo, J. Curely,
Phys. Rev. B 40 (1989) 10992–10998.
References
[13] S. Ostrovsky, W. Haase, M. Drillon, P. Panissod, Phys. Rev. B
64 (2001) 134418/1–134418/17.
[1] M. Richard-Plouet, S. Vilminot, J. Mater. Chem. 8 (1998)
131–137.
[14] G. Courbion, R. De Pape, J. Teillet, F. Varret, J. Pannetier,
J. Magn. Magn. Mater. 42 (1984) 217–232.
[2] M. Richard-Plouet, S. Vilminot, Solid State Sci 1 (1999) 381–393.
[3] M. Guillot, M. Richard-Plouet, S. Vilminot, J. Mater. Chem. 12
(2002) 851–857 (doi:10.1039/b110101c)..
[15] M. Guillot, A. Marchand, M. Leblanc, G. Ferey, J. Phys. C 20
(1987) 2405–2414.
[16] M.A.G. Aranda, J.P. Attfield, E. Batchelor, G.P. Shields,
S. Bruque, M. Gabas, Inorg. Chem. 37 (1998) 1329–1335.
[17] M. Guillot, M. El-Ghozzi, D. Avignant, G. Andre, F. Bouree,
A. Cousson, J. Appl. Phys. 191 (2002) 8519–8521.
[18] V.H. Crawford, H. Wayne Richardson, J.R. Wasson, D.J.
Hodgson, W.E. Hatfield, Inorg. Chem. 15 (1976) 2107–2110.
[4] C.K. Majumdar, D.K. Gosh, J. Math. Phys. 10 (1969) 1388.
[5] F.C. Hawthorne, L.A. Groat, R.K. Eby, Can. Mineral. 27 (1989)
205–209.
[6] J. Rodriguez-Carvajal, FULLPROF: Rietveld, Profile Matching
and Integrated Intensity Refinement of X-ray and/or Neutron