Magnetic structure of Ni(DCOO)2(D2O)2

Article abstract of DOI:10.1021/ic102008a

Ni(HCOO)₂(H₂O)₂ is a structurally simple coordination polymer showing interesting magnetic phase transitions at low temperature (<16K). Previously published studies of these phase transitions have yielded inconsistent results, questioning the correctness of the published magnetic structure. Here heat capacity and magnetic susceptibility of a fully, a partly and a non-deuterated sample were measured, and they all exhibit magnetic phase transitions around 3 and 15 K. Neutron powder diffraction data was collected on the fully deuterated sample at various temperatures between 1.5 and 25 K. A magnetic model was refined against the neutron diffraction data using a spin system composed of two canted antiferromagnetic sublattices. The magnetic moments of the two sublattices show different magnitude, 1.7 μ_B and 1.3 μ_B, and the temperature dependence of the magnetic sublattices is quite different. One of the sublattices shows the expected temperature behavior of an antiferromagnetic compound whereas the other sublattice follows a Brillouin like function with a slowly increasing magnetization below the Neel temperature.

Full text of DOI:10.1021/ic102008a

Inorg. Chem. 2011, 50, 1441–1446 1441
DOI: 10.1021/ic102008a
Magnetic Structure of Ni(DCOO)2(D2O)2
Mads R. V. Jørgensen, Mogens Christensen, Mette S. Schmøkel, and Bo B. Iversen*
Centre for Materials Crystallography, Department of Chemistry and iNANO, Aarhus University,
˚
DK-8000 Arhus C, Denmark
Received October 1, 2010
Ni(HCOO)2(H2O)2 is a structurally simple coordination polymer showing interesting magnetic phase transitions at low
temperature (<16K). Previously published studies of these phase transitions have yielded inconsistent results,
questioning the correctness of the published magnetic structure. Here heat capacity and magnetic susceptibility of a
fully, a partly and a non-deuterated sample were measured, and they all exhibit magnetic phase transitions around
3 and 15 K. Neutron powder diffraction data was collected on the fully deuterated sample at various temperatures
between 1.5 and 25 K. A magnetic model was refined against the neutron diffraction data using a spin system
composed of two canted antiferromagnetic sublattices. The magnetic moments of the two sublattices show different
magnitude, 1.7 μB and 1.3 μB, and the temperature dependence of the magnetic sublattices is quite different. One of
the sublattices shows the expected temperature behavior of an antiferromagnetic compound whereas the other
ꢀ
sublattice follows a Brillouin like function with a slowly increasing magnetization below the Neel temperature.
Introduction
Especially Mn(HCOO)2(H2O)2 has been examined since sin-
gle crystals suitable for neutron scattering experiments can be
grown. However, despite these detailed structural studies
inconsistencies still exist, and the understanding of the mag-
netic properties is incomplete. As an example the magnitude of
the magnetic moments in the manganese system has been
modeled by Radhakrishna et al. using single crystal polarized
neutron diffraction data.5 They found two quite different ions
with the magnitudes of the magnetic moments being 0.38(2) μB
and 1.73(2) μB for the Mn1 and Mn2 site, respectively. On
the other hand, magnetization measurements lead to an over-
all effective magnetic moment of 5.840(2) μB per ion, which is
in good agreement with the free high-spin state of a Mn2þ ion.
This picture was supported by charge density modeling of
X-ray diffraction data.6
The crystal structure of Ni(HCOO)2(H2O)2 is shown in
Figure 1, and it belongs to the monoclinic space group P21/c.
The structure contains two distinct Ni ions sites (Ni1 and
Ni2) in special crystallographic positions, which are linked by
formate groups, and Ni2 is furthermore connected to water
molecules, see Figure 1a. The Ni1 ions are coordinated to six
formate oxygen atoms forming a slightly distorted octahedron,
thereby creating layers in the bc planes, see Figure 1b-c. The
ions within these layers are interconnected with formate
bridges. The Ni2 ions are coordinated to two formate oxygen
Traditionally studies of magnetic properties of materials
tend to focus either on inorganic ionic solids (e.g., oxides) or
on molecular complexes. In the past decade studies of hybrid
materials in the form of extended network structures contain-
ing metal centers linked by organic linkers (coordination
polymers) have received enormous attention.1 However,
because of the structural complexity of coordination polymers
such studies rarely attempt experimental determination of
the microscopic magnetic structure, and they primarily focus
on measurements of macroscopic magnetic properties.2,3
One exemption is the transition metal formate dihydrate,
M(HCOO)2(H2O)2, which is a comparatively simple coordi-
nation polymer. The crystal structure of the isostructural
M(HCOO)2(H2O)2 compounds has been known since the
1960s, and many studies have focused on their magnetism.4
*To whom correspondence should be addressed. E-mail: bo@chem.au.dk.
Fax: (þ45) 8619 6199.
(1) (a) Chen, B. L.; Xiang, S. C.; Qian, G. D. Acc. Chem. Res. 2010, 43,
1115. (b) Farah, O. K.; Hupp, J. T. Acc. Chem. Res. 2010, 43, 1166. (c) Qui, S. L.;
Zhu, G. S. Coord. Chem. Rev. 2009, 253, 2891. (d) Wang, Z. Q.; Cohen, S. M.
Chem. Soc. Rev. 2009, 38, 1315. (e) Lee, J.; Farah, O. K.; Roberts, J.; Scheidt,
K. A.; Nguyen, S. T.; Hupp, J. T. Chem. Soc. Rev. 2009, 38, 1450. (f) Li, J. R.;
Kuppler, R. J.; Zhou, H. C. Chem. Soc. Rev. 2009, 38, 1477. (g) Robin, A. Y.;
Fromm, K. M. Coord. Chem. Rev. 2006, 250, 2127. (h) Kitagawa, S.; Kitaura, R.;
Noro, S. I. Angew. Chem., Int. Ed. 2004, 43, 2334.
(2) Poulsen, R. D.; Bentien, A.; Chevalier, M.; Iversen, B. B. J. Am. Chem.
Soc. 2005, 50, 9156.
(3) Clausen, H. F.; Overgaard, J.; Chen, Y.-S.; Iversen, B. B. J. Am. Chem.
Soc. 2008, 64, 7988.
(4) (a) Osaki, K.; Nakai, Y.; Watanabe, T. J. Phys. Soc. Jpn. 1964, 19, 717.
(b) Osaki, K.; Nakai, Y.; Watanabe, T. J. Phys. Soc. Jpn. 1963, 18, 919.
(5) Radhakrishna, P.; Gillon, B.; Chevrier, G. J. Phys.: Condens. Matter
1993, 5, 6447.
(6) Poulsen, R. D.; Jørgensen, M. R. V.; Overgaard, J.; Larsen, F. K.;
Morgenroth, W. G.; Graber, T.; Chen, Y.-S.; Iversen, B. B. Chem.;Eur. J.
2007, 13, 9775.
r
2011 American Chemical Society
Published on Web 01/19/2011
pubs.acs.org/IC

1442 Inorganic Chemistry, Vol. 50, No. 4, 2011
Jørgensen et al.
Figure 1. (a) Structure of Ni(HCOO)2(H2O)2 at 100 K with thermal ellipsoids shown at 50% probability level. (b) The structure viewed along the
monoclinic b-axis. The H5B O2 hydrogen bonds are shown using dotted bonds. (c) Unit cell showing only Ni atoms and schematic “formate”
connections.
3 3 3
atoms and four water ligands, also in a slightly distorted
octahedral geometry. The water ligands in the equatorial
plane of the coordination sphere of Ni2 provide only electro-
static interactions between the Ni2 centers andthusefficiently
disconnect the Ni2 ions within these layers that interpenetrate
the layers formed by the Ni1 ions. The two different layers are
connectedby formatebridges betweenNi1 and Ni2 creating a
three-dimensional (3D) network. In this network the Ni1 ions
are connected through covalent bonds while the Ni2 ions are
virtually isolated from each other. In the crystal structure
there are four reasonably strong hydrogen bonds formed by
the water hydrogen atoms. Three of these four hydrogen
bonds are from the water in the Ni2 layer to the Ni1 layer
while the last (connecting H5B to O2) is within one Ni2 layer.
The latter hydrogen bond thus leads to a more “direct”
interaction between neighboring Ni2 ions, Figure 1b.
The magnetic ordering of Ni(HCOO)2(H2O)2 was reported
to consist of an antiferromagnetic ordering of Ni1 moments
(μ1) at TN = 15.5 K while the moments on Ni2 (μ2) behave
almost paramagnetic until sublattice antiferromagnetic
ordering at T = 3 K.7,8 In the manganese analogue, Pierce
and Friedberg noted, that the magnetic ordering in the Mn1
sublattice necessarily has to involve a finite ordering of the
Mn2 sublattice. This ordering, however, is small and thus the
Brillouin function is small. This leads to moments that
behave nearly as in a paramagnetic system.9 Kageyama
et al. has recently reported a molecular-field model that also
shows a similar small ordering of the Ni2 sublattice at the
10
ꢀ
overall Neel temperature.
Ni(HCOO)2(H2O)2 has been studied with a broad range
of methods such as magnetization and magnetic suscepti-
bility,8,10 specific heat capacity,7,11 and neutron powder
diffraction.12 More recently, the compound was studied by
proton NMR by Zenmyo et al.13 There are discrepancies
between results obtained with different methods. Magnetic
susceptibility measurements and neutron diffraction suggest
μ
eff = 3.14 μB and μ1 = μ2 ∼ 2.0(2) μB, respectively. The spin
model presented by Zenmyo et al. based on NMR measure-
ments suggests μ1 = 2.38 μB and μ2 = 0.38 μB, and this is
in agreement with the weak ferrimagnetism observed by
Kageyama et al.10 The NMR model also seems to explain
the compensation point observed by the same authors.10 The
moments reported by Zenmyo et al. raise the question why
μ(Ni1) > μ(Ni2) whenμ(Mn2) > μ(Mn1) inthe isostructural
manganese analogue. If this is the case then the chemical
(9) Pierce, R. D.; Friedberg, S. A. Phys. Rev. 1968, 165, 680.
(10) Kageyama, H.; Khomskii, D. I.; Levitin, R. Z.; Vasil’ev, A. N. Phys.
Rev. B 2003, 67, 224422.
(11) Takeda, K.; Kawasaki, K. J. Phys. Soc. Jpn. 1971, 31, 1026.
(12) Burlet, P.; Burlet, P.; Rossatmignod, J.; Decombarieu, A.; Bedin, E.
Phys. Status Solidi B 1975, 71, 675.
(7) Pierce, R. D.; Friedberg, S. A. Phys. Rev. B 1971, 3, 934.
(8) Hoy, G. R.; Barros, S. D. S.; Barros, F. D. S.; Friedberg, S. A. J. Appl.
Phys. 1965, 36, 936.
(13) Zenmyo, K.; Kubo, H.; Tokita, M.; Yamagata, K. J. Phys. Soc. Jpn.
2006, 75, 104704.

Article
Inorganic Chemistry, Vol. 50, No. 4, 2011 1443
bonding in the coordination polymer must play a central role
in determining the magnetic properties.
The compound Ni(HCOO)2(H2O)2 does not grow suffi-
ciently large single crystals for a single crystal neutron
diffraction experiment. Therefore, the present neutron study
concerns a powder sample, but this leads to complications
because of the relative high hydrogen content of the material,
which causes a high incoherent background. To improve the
signal-to-noise ratio the neutron study was carried out on a
deuterated sample. A total of three samples, a fully, a partly
and a non-deuterated sample, were synthesized. Measure-
ments of the heat capacity and the magnetic susceptibility
were carried out on all three samples to assess the similarity
between the deuterated and the fully hydrogenated com-
pounds.
Experimental Section
Figure 2. Specific heat capacity of the samples with circles (O) for the
hydrogenated sample, plus symbols (þ) for the partly deuterated sample,
and squares (0) for the fully deuterated sample. The lines are guides to the
eye. Insets show peaks close up.
Synthesis. The sample used in both the neutron diffraction
experiments and the measurements of the physical properties
was grown at the interface between a layer of deuterated formic
acid and a layer of Ni-acetate in a mixture of heavy water
(∼100% pure) and ethanol. Hydrogen contamination from the
atmosphere was avoided by carrying out the reaction in an inert
N2 atmosphere. The synthesis produced a powder consisting of
small green crystals. A partly deuterated sample was prepared
using the same approach but using “regular” hydrogenated
formic acid and heavy water (∼100% pure). Finally, a fully
hydrogenated sample was prepared using hydrogenated formic
acid and water. During the synthesis it was observed that the
crystallization of the deuterated compounds was slower than
the synthesis using regular “light” water. Powder X-ray diffrac-
tion confirmed that the prepared compounds were the title
compound and that the samples were phase pure.
detector covering 80° with 400 channels at 0.2° interval, and
successive powder diffraction patterns were collected with 0.1°
steps to improve peak resolution.14 Diffraction patterns at low
Q with good statistics were recorded at 25, 6, and 1.5 K, that is,
above, in-between, and below the two phase transitions. These
˚
patterns were recorded at two different wavelengths of 2.45 A
˚
and 4.2 A using a vertical focusing (002) graphite monochro-
mator. Furthermore, a series of patterns with less precise
statistics were recorded stepping up in temperature in small
intervals. To prevent any hysteresis effects10 the desired tem-
perature was always reached by cooling from above 25 K to the
lowest possible temperature (∼1.5 K) and then reheating to the
desired temperature.
Physical Properties Measurement. The magnetic suscepti-
bility measurements were carried out on small pellets of pressed
powder. A Quantum Design Physical Property Measurement
System (PPMS) at the Department of Chemistry, Aarhus
University, was used in the temperature range from 1.9 K to
room temperature in a field of 2 T to measure the magnetic
susceptibility. The same sample batch was used for the magnetic
susceptibility and the subsequent neutron powder diffraction
experiments. During the measurement of the magnetic suscept-
ibility of the partly deuterated sample there were some problems
with the gain amplification leading to systematically lower
values of the magnetic susceptibility. Consequently, the mag-
netic susceptibility of the partly deuterated sample was scaled
linearly to the magnetic susceptibility of the deuterated sample
using the temperature interval 50 to 250 K. Measurements of
the heat capacity were also performed with the PPMS. Small
pellets of the three samples were mounted on the sample holder
with a small amount of grease. The amount of grease used in the
measurement of the heat capacity of the deuterated sample
was unfortunately erroneous, thus leading to an incorrect
addenda signal subtraction. Therefore the addenda signal, that
is, the mass of grease, has been rescaled.
After the DMC measurements the sample was transferred to
the HRPT instrument and placed in a similar helium cryostat to
record the high-Q region of the powder diffraction pattern. The
HRPT instrument is located at the hot neutron source, and it is
equipped with a banana-shaped detector covering 160° with
1600 channels at 0.1° intervals, and successive powder diffrac-
tion patterns were collected with 0.05° steps to improve peak
resolution.15 Three diffraction patterns were recorded at 1.5, 6,
˚
and 25 K using neutrons with a wavelength of 1.89 A obtained
from a vertically focusing (511) Ge monochromator. Neutron
diffraction data were only measured using the fully deuterated
sample.
Results and Discussion
Physical Property Measurements. The signatures of the
two phase transitions are easily seen in the heat capacity,
Figure 2. The change in heat capacity at low tempera-
ture ∼3.5 K is assumed to be associated with the complete
magnetic ordering of the Ni2 sublattice. This phase
transition is close to identical for the three samples, while
the second peak differs slightly between the three samples.
The two samples containing heavy water are shifted
toward lower temperature, ∼15.1 K, with respect to the
fully hydrogenated sample, ∼15.4 K, and the shape of the
peak is also different. The fully deuterated and fully
Neutron Powder Diffraction. The neutron diffraction experi-
ments were performed at the quasi-continuous neutron spalla-
tion source (SINQ) at the Paul Sherrer Institute (PSI) in Villigen,
Switzerland. The sample was handled in a He containing glove-
box, where it was ground in a mortar to avoid preferential orienta-
tions of crystallites. The powder was transferred to a 6 mm
vanadium cylinder, which was sealed using a piece of indium
wire. The helium gas in the sample container acts as protecting
atmosphere and heat exchanger. The sample was transferred to
a helium cryostat (Orange ILL type) at the DMC neutron
powder diffractometer located at the cold neutron source. This
instrument is equipped with a banana-shaped position sensitive
(14) Fischer, P.; Keller, L.; Schefer, J.; Kohlbrecher J. Neutron News
2000, 11, 19.
€
(15) Fischer, P.; Frey, G.; Koch, M.; Konnecke, M.; Pomjakushin, V.;
€
Schefer, J.; Thut, R.; Schlumpf, N.; Burge, R.; Greuter, U.; Bondt, S.;
Berruver, E. Physica B 2000, 276-278, 146.

1444 Inorganic Chemistry, Vol. 50, No. 4, 2011
Jørgensen et al.
a
˚
˚
˚
Table 1. Refinement Residual Factors for the HRPT Data at 1.89 A (I), the DMC Data at 4.2 A (II), and the DMC Data at 2.45 A (III)
25 K
II
6 K
II
1.5 K
II
1.5 K NMR model
II
I
III
I
III
I
III
I
III
no. of reflections
no. of parameters
RF
Rwp
Rp
RBragg
RMag
χ2
682
88
42
88
145
88
1369
47
88
47
251
47
1369
68
88
68
293
68
1369
65
88
65
293
65
3.19
10.7
10.0
5.03
N/A
10.1
1.97
5.92
6.41
1.96
N/A
1.71
5.49
5.45
2.15
N/A
3.33
11.8
11.1
5.33
21.1
11.7
1.35
5.63
5.81
1.83
8.55
1.53
5.91
5.82
2.40
13.5
3.02
10.4
9.93
4.70
14.9
11.8
1.24
5.12
5.32
1.34
7.27
1.68
6.17
6.11
2.49
7.16
3.02
10.8
10.3
4.70
51.7
22.4
1.61
8.43
7.60
1.50
78.7
1.82
7.62
7.05
2.79
60.1
P
P
P
P
P
P
a The agreement factors are defined by Rp = |yo,i - yc,i|/ |yo,i|, Rwp = [ wi|yo,i - yc,i|2/ wi|yo,i|2]1/2, χ2 = wi|yo,i - yc,i|2/σi2, RF = |Fo - Fc|/
|Fo|, Rmag is similar to RF but for the magnetic peaks.
P
Table 2. Components and Magnitudes of the Magnetic Moments at 1.5 and 6 K
hydrogenated samples both show sharp peaks, while the
partly deuterated sample shows a broad round peak.
The heat capacity is very similar to the values reported
in the literature for hydrogenated samples.7,11 The round
peak found for the partly deuterated sample at 15 K is
similar to the previously reported shapes, whereas it is the
sharper peaks of the fully deuterated and fully hydro-
genated samples that differ. We currently have no explana-
tion for this difference. The shift of the peak position
could indicate that the hydrogen bonds in the structure
may be involved in the magnetic ordering as proposed by
Poulsen et al.6
μ1(x)/ μ1(y)/ μ1(z)/
μB μB μB
|μ1|/ μ2(x)/ μ2(y)/ μ2(z)/
μB
|μ2|/
μB
μB
μB
μB
1.5 K 1.2(1) 0.4(3) -1.0(2) 1.7(3) 0.4(1)
6 K
0.8(1) 0.5(3) -1.5(2) 1.7(2) 0a
0a
0a
-1.2(2) 1.3(2)
-0.7(1) 0.7(1)
a These components were removed from the model because of
unreliable refinement of the values.
groups was 100% and that the deuterium occupancies on
the water sites were close to 90%. This structural model
was subsequently used as a starting model in the refine-
ment of the magnetic phases.
From the magnetization measurements the magnetic
susceptibility was calculated and the paramagnetic region
from approximately 50 to 270 K was fitted with the
Curie-Weiss law. Quite similar values for the effective
moment (μeff) were calculated for the three samples,
2.88(1) μB, 2.89(1) μB, and 3.03(1) μB for the fully deut-
erated, the partly deuterated, and the fully hydrogenated
sample, respectively, and the Weiss temperatures (Θ)
were found to be -12.1(2), -12.7(1), and -9.1(5) K.
Nevertheless, the data may indicate a slight difference
between the deuterated and the hydrogenated samples.
In a similar experiment Kageyama et al. reported values
for μeff (3.14 μB) and Θ (-15.5 K).10 The value for the
effective magnetic moment is close to, but exceeds, the
spin only value of a free Ni2þ ion; 2.82 μB. Although
the effective magnetic moment of the hydrogenated sam-
ple is higher and the Weiss temperature lower than the
values obtained from the measurements on the deuterated
samples, it appears that the magnetic ordering in the
deuterated sample is similar to the magnetic ordering in
the hydrogenated sample.
The 1.5 and 6 K phases were modeled by adding a
magnetic phase to the 25 K structural model. The mag-
netic model consisted of two canted antiferromagnetic
sublattices with both sublattices canted in the b direction.
All non-hydrogen atom positions and peak profiles were
refined while the positions and occupancies of the hydro-
gen/deuterium atoms along with all displacement param-
eters were kept at the values obtained from the 25 K data.
The magnetic phase was modeled in space group P1 after
a representational analysis performed in the program
SARAh.18
The general assumption regarding the magnetic order-
ing is that the Ni1 lattice orders at 15.5 K while the Ni2
lattice only shows small ordering and can therefore be
approximated as “paramagnetic” down to ∼3 K. There-
fore, the initial 6 K model did not include μ2. However,
addition of this moment significantly improved the fit,
and it was therefore included in the final model. Refine-
ment details from the final models are shown in Table 1. A
comparison of the two different 6 K models can be found
in the Supporting Information.
Although the refinements in general were stable, the
refinement of the magnetic moment along the b axis was
unstable because of its small value (∼ 0.1-0.2 μB), and it
was therefore omitted from the Ni2 sublattice at both
1.5 and 6 K. Similarly, the Ni2 moment along the a axis
had to be removed from the 6 K model as this also was too
small to be modeled reliably. During the refinements it
was noted that the absolute direction of the moment
could be changed to the directly opposite direction caus-
ing only a very small change in the R-values and no visual
change in the fit to the data. This is presumably due to the
rather weak signal of the magnetic scattering compared
with the nuclear scattering combined with the small
Magnetic Structure. The magnetic and structural models
were simultaneously refined against the two DMC data
sets and the HRPT data set using the Rietveld method
implemented in the FullProf program.16,17 The starting
model was obtained from highly accurate single crystal
synchrotron X-ray diffraction data measured at 100 K
(unpublished results). The structural model was refined
against the three 25 K data sets, and all positions,
isotropic displacement parameters, and peak profiles
were included in the refinement. Furthermore, the occu-
pancies of the hydrogen/deuterium positions were refined. It
was found that the deuterium occupancy on the formate
(16) Rietveld, H. M. Aust. J. Phys. 1988, 41, 113.
(17) Rodriguezcarvajal, J. Phys. B 1993, 192, 55.
(18) Wills, A. S. Phys. B 2000, 276, 680.

Article
Inorganic Chemistry, Vol. 50, No. 4, 2011 1445
˚
Figure 3. Diffraction data from DMC at 4.2 A. From top: 25, 6, 1.5 K,
difference between 6 and 25 K, and difference between 1.5 and 25 K.
Vertical lines show Bragg positions. The inset shows the full diffracto-
gram, and the box indicates the region of interest. The diffraction patterns
are offset for clarity.
Figure 4. Projections of the unit cell showing the magnetic moments.
(a) 1.5 K, projection along b. (b) 1.5 K, projection along c. (c) 6 K, projec-
tion along b. (d) 6 K, projection along c.
deviation from orthorhombic symmetry (β ∼ 97°). The
directions presented here (see Table 2 and Figure 4) are
the ones that gave the lowest agreement factors. The low
angle 1.5, 6, and 25 K data from DMC, along with
the difference between data and models, are shown in
Figure 3. Selected Rietveld fits are included in the Sup-
porting Information.
The magnitude of the moments at 1.5 K were found to
be 1.7(3) μB and 1.3(2) μB for Ni1 and Ni2, respectively.
At 6 K the values were refined to 1.7(2) μB and 0.7(1) μB.
The components of the moments are given in Table 2 and
shown in Figure 4. To compare the present magnetic
model with the NMR model proposed by Zenmyo et al.13
the reported moments from the NMR study were added
at fixed values to the structural model, while the atomic
parameters, unit cell and background were refined. This
model leads to very high R-values (see Table 1), and the
corresponding Rietveld fit is shown in Figure 5.
ꢀ
temperature approaches the Neel temperature (15.5 K).
On the other hand μ2 deviates from this behavior and
shows a Brillouin like behavior resulting in a small and
slowly decreasing moment in a large temperature range.
A similar result has been reported in the literature based
on molecular-field modeling.9,10 The agreement between
that model and the present results is quite good for the
Ni1 sublattice magnetization. When comparing the Ni2
sublattice magnetization it is seen that the molecular-field
model overestimates the μ2 ordering in the tempera-
ture range from ∼4 K to ∼13 K. This could be due to a
too simplified model ignoring the Ni2-Ni2 interaction
or simply a slightly wrong estimate of the Ni1-Ni2
exchange parameter. The temperature dependence of
the Ni2 sublattice magnetization is presumably due to
the magnetic field from the Ni1 layers, the exchange with
the Ni1 layers, and the weak interaction in the Ni2 layer.
This is also reflected in the molecular-field models.
As noted above the model reported by Zenmyo et al.
based on NMR measurements gave a poor fit to the
neutron data (Figure 5). The two models differ not only
on the direction of the moments but also on the magni-
tude. Despite these differences they both describe an
antiferromagnetic ordering in the ac planes with canting
in the b direction. The magnetic moment along the b axis,
because of canting, is quite similar. Compared with the
results obtained from magnetic susceptibility measure-
ments it seems that both NMR and neutron scattering
underestimates the moments. However, it is not straight-
forward to compare these values as the magnetization is
measured in the paramagnetic regime whereas NMR and
neutron measurements are performed in the ordered
state. The moments obtained from the ordered state will
be smaller than the ones obtained from the paramagnetic
state because of temperature fluctuations of the moments.
Figure 6 indicates that μ1 has saturated and that μ2 is close
to saturation. The estimated average moment or effective
moment obtained from NMR is 1.4 μB, and it is 1.5 μB
Temperature Dependence of the Magnetic Moments.
The temperature dependence of the magnetic moments
was modeled using data collected on the DMC instru-
ment. For the temperatures in the range 1.5 to 12 K, the
magnetic model was refined against data collected using
˚
˚
both 2.45 A and 4.2 A neutrons, whereas the model in the
temperature range 13 to 15.5 K only was refined against
˚
˚
4.2 A data (2.45 A data was not collected in this tempera-
ture range). In these models only the background, the
unit cell, and the magnetic moments were refined. As
described above, a significant moment on Ni2 was found
at 6 K. Similar improvements of the fits at higher tem-
perature than 6 K was also observed, and the model includ-
ing μ2 was chosen as the most reliable. Adding μ2 to the
model does not change the magnitude of μ1 significantly.
The temperature dependence of the magnetic moment on
the two sites is shown in Figure 6. The moments obtained
˚
˚
at 1.5 and 6 K using all data (DMC, 2.45 A and 4.2 A, and
˚
HRPT at 1.89 A), are also shown in this figure.
As seen from Figure 6, the temperature dependence of
μ1 shows the expected temperature dependence of an
ordering of magnetic spins with a rapid decrease as the

1446 Inorganic Chemistry, Vol. 50, No. 4, 2011
Jørgensen et al.
˚
Figure 5. Rietveld fit of 1.5 K data measured at DMC with 4.2 A neutrons using (left) the present model and (right) the magnetic model proposed by
Zenmyo et al. The red dots show the measurements, the black line shows the model, and the blue line shows the difference. Green bars show predicted Bragg
positions. The upper green bars show predicted Bragg positions for the structure while lower bars show predicted Bragg positions for the magnetic phase.
two canted antiferromagnetic sublattices leading to a weak
ferrimagnet. The magnitude of the magnetic moments at
1.5 K is found to be 1.7(2) μB and 1.3(2) μB for Ni1 and Ni2,
respectively, which is much lower than the free Ni2þ ion
value, and significantly lower than the value obtained from
magnetization measurements. The temperature dependence
of the sublattices is in good agreement with the molecular-
field model suggested in the literature and shows the small
but significant ordering of the Ni2 sublattice at the Neel
temperature of the Ni1 sublattice. The magnetic model pre-
sented here has weak ferrimagnetism created by two canted
antiferromagnetic sublattices. The canting is along the b axis,
and the resulting moments of the two sublattices are opposed.
Acknowledgment. We gratefully acknowledge and
fondly remember the assistance of recently deceased
Marie Agnes Chevallier during synthesis of the materials.
Figure 6. Temperature dependence of the magnitude of the magnetic
moment of Ni1 (open squares) and Ni2 (open circles). The closed squares
and closed circles represent the magnitude of the Ni1 and Ni2 moments
obtained from all available data sets.
Dr. L. Keller and Dr. D. Sheptyakov are thanked for
assistance during the neutron measurements. This work is
based on experiments performed at the Swiss spallation
neutron source SINQ, Paul Scherrer Institute, Villigen,
Switzerland. The work was supported by The Danish
National Research Foundation (Center for Materials
Crystallography), The Danish Strategic Research Coun-
cil (Centre for Energy Materials), and the Danish
Research Council for Nature and Universe (Danscatt).
from the present neutron study. This is approximately
50% of the effective moment obtained from magnetiza-
tion measurements.
Conclusions
On the basis of heat capacity measurements it is concluded
that the magnetic ordering in Ni(DCOO)2(D2O)2 is similar
to that observed in the regular hydrogenated compound,
Ni(HCOO)2(H2O)2. The magnetic structure is composed of
Supporting Information Available: Magnetic susceptibility
data, specific heat capacity data, and Rietveld refinement plots.
This material is available free of charge via the Internet at http://
pubs.acs.org.