J.B. Amaral et al. / Journal of Solid State Chemistry 241 (2016) 26–29
27
The crystalline structures of samples were investigated by the
X-ray powder diffraction (XRPD) measurements using an Empyr-
ean diffractometer (PANalytical) with Cu-K radiation
¼1.5406 Å) for Cu1ꢀxFexO and Cu1ꢀxNixO samples and a Rigaku
diffractometer with Co-K radiation (
¼1.789 Å) for Al and Zn-
doped CuO samples. All data were collected in the continuous
mode with 2 range from 30° to 95°, step size of 0.02° and a
scanning speed of 0.5° /min. The obtained XRD patterns were
compared with that found in the Joint Committee Powder Dif-
fraction Standards (JCPDS) database and analyzed using the Riet-
veld refinement [11] approach as implemented in the
DBWSTools2.3 free software [12,13]. The pseudo-Voigt function
was selected to provide the best fits of the observed peak profiles.
Table 1
Potential parameters, shell model constants and ionic charges used in present
work.
α
(λ
Short-range interactions
λ
α
Buckingham
CuCore–OShell
A (eV)
1650.0
22764.0
k (eV Å2)
25.0
ρ (Å)
C (eV Å)
0.00
27.88
Cutoffs (Å)
12.00
12.00
0.3280
0.1490
r0 (Å)
1.65
θ
O
Shell–OShell
Harmonic
CuCore–OCore
Shell model parameters
Ion
Y (e)
ꢀ2.2
K (eV Åꢀ2
)
48.0
O2ꢀ
2.2. Computational details
In order to describe the defect disorder produced by the TM-
doping in our CuO samples, we have employed static simulation
techniques based on energy minimization procedures. It worth to
comment that the long-range (Coulombic) and short-range pair
potentials were used to treat the interaction among all the ions of
the solid. For this last interaction, the Harmonic and Buckingham
functions were employed. The Buckingham potential can be de-
scribed by following equation
⎛
⎝
⎞
⎠
rij
Cij
ri6j
⎜
⎜
⎟
⎟
V (rij) = Aij exp
−
ρ
ij
(1)
Where, the first term describes the Pauli short-range repulsion
and the second term the Van der Walls attraction. The ionic po-
larizability was taken into account for the O2ꢀ ions through the
shell model [14]. On the other hand, the interionic interactions
between Cu2þ–O2ꢀ and O2ꢀ–O2ꢀ were extracted initially from
Ref. [15] and then refitted in order to reproduce the structural
properties of Cu1ꢀxTMxO (TM¼Fe3þ, Ni2þ, Zn2þ and Al3þ) sam-
ples studied in this work. The point defects were calculated using
the Mott-Littleton method [16] which consider that the region
close to the defects is partitioned into two spherical regions – one
next to the defect and other to larger distances. In the first case,
the ions are relaxed individually under the defect perturbation
while in the second case, the distance extends theoretically to
infinity and the ionic relaxation is treated by using the approx-
imate Mott-Littleton quasi-continuum procedures. In order to
study the effect of dopant concentration in the CuO matrix, the
partial occupancy mean-field was used. Calculations were per-
formed to investigate the substitution of Cu2þ and the term of
charge compensation was included when needed. Another im-
portant feature is that in the mean-field approximations the sub-
stituted ions are randomly distributed in the CuO host for all
concentrations, that is, there is no interaction between the sub-
stituted ions. In this way, the general utility lattice program (GULP)
[17] was used to predict the structural properties discussed in this
work. The details of the methodology used in the simulations have
been described elsewhere [18–20].
Fig. 1. XRD patterns for Cu1ꢀxNixO samples (x¼0; 0.01; 0.05 and 0.10). The red and
green solid lines mean the calculated pattern and the diference between experi-
mental and simulated patterns, respectively. The blue vertical bars are an indexa-
tion obtained from the JCPDS database (48–1548). (For interpretation of the re-
ferences to color in this figure legend, the reader is referred to the web version of
this article.)
CuO samples with concentration 1%, 5% and 10% are showed in
Figs. 1 and 2, respectively. The XRD analysis show that only the
sample obtained with the lowest TM-doping (x¼0.01) present
single phase with structure isomorphous to the monoclinic CuO
structure (space group: C2/c). However, we observed the appear-
ing of spurious phases of hexagonal ZnO (space group: P63mc) and
cubic NiO (space group: Fm3m) for the samples with higher con-
centration of Ni or Zn (x¼0.05 and 0.10).
Figs. 3 and 4 show the XRD patterns for Cu1ꢀxFexO and
Cu1ꢀxAlxO samples (x¼0, 0.01, 0.05 and 0.10), respectively. One
can note that, independently of doping concentration, all samples
present a single phase with monoclinic CuO type structure (space
group: C2/c). Liu et al. [21] show that CuO nanopowders doped
with Fe synthesized by the co-precipitation method also present
similar behavior [21]. On the other hand, a spurious phase of iron
oxide (Fe2O3) have been observed for Cu1ꢀxFexO samples
(0oxo0.30) grown by the combustion synthesis method [22]. In
this case, the Fe2O3 phase was observed to concentrations higher
than x¼0.25. In the same sense, by using the microwave com-
bustion method, Mohamed et al. [23] synthesized Fe-doped CuO
nanostructures with x¼0, 0.05, 0.1, 0.15, and 0.20. They show that
Fe ions were successfully introduced into the lattice substituting
the Cu ions without affecting the crystal structure of the CuO host
[9]. According the above mentioned experimental results, the
It is important to state that atomistic simulations generally
must start with a set of interatomic potentials which can describe
accurately the studied material. In this sense, the potential para-
meters for CuO oxide obtained by empirical fitting method were
used (see Table 1). The difference between the calculated and
experimental lattice parameters was less than 2%, indicating that
our potential reproduces well the CuO structure.
3. Results
The experimental and calculated XRD patterns for Ni, Zn-doped