Z. Guo, et al.
JournalofMagnetismandMagneticMaterials490(2019)165473
the [TMX] host would be maintained and intact. Consequently, a deep
understanding on this regulation will be of great help for the material
design in ATM2X2 system. However, tuning the A vacancies in ATM2X2
trends to be unattainable in arsenides due to the request of valence
balance from the negative [TMAs]- [17,18]. For TM chalcogenides with
electrically neutral [TMX], the deficient A was reported to reduce the
structural stability in Ni-based system, leading to the collapse of
ThCr2Si2-type structure [19]. These issues hinder us to thoroughly un-
derstand the regulation of intermediated metal A on the structure and
related properties of ATM2X2. In order to solve this problem, a system
where the content of intermediate metal A can be easily controlled
without inducing any TM vacancy or structural instability is required.
Layered cobalt selenide KCo2Se2 with ThCr2Si2-type structure was
reported to show A-type ferromagnetism [6]. Wide investigations have
been done on KCo2Se2 and its sister compounds ACo2Ch2 (A = Rb, Cs,
Tl and Ch = Se, S) by chemical doping on A, Co and Ch sites, leading to
a series of novel magnetic phase transitions [20–25]. However, the
effort to adjust the A vacancies in these layered Co selenides has not
been reported. In this work, we report the intrinsic feature of tunable K
vacancies in nominal K1−xCo2Se2 (0 ≤ x ≤ 0.3). It is indicated that
both the ThCr2Si2-type structure and intact [CoSe] layers can be well
maintained when the composition of intermediate metal is adjusted in
this system. With the K vacancies increased, the lattice parameter a is
shortened while c is elongated, resulting in an increased tetragonality
(c/a). Concerning physical properties, K vacancies significantly in-
creases electric resistivity of K1−xCo2Se2 by reducing the transferred
electrons. More interestingly, in spite of the shortened Co-Co distance in
[CoSe] functional layers, inducing the K vacancies leads to weaker
magnetic interaction with Curie temperature (TC) of K1−xCo2Se2 de-
creased from 80 to 52 K. This suggests that the ferromagnetic interac-
tion in K1−xCo2Se2 is mainly governed by the charge transfer, rather
than the structural parameters. First-principle calculation is also carried
out to explore the mechanism of these effects from K vacancies.
stoichiometric KCo2Se2 structure was used as the starting model to re-
fine the K1−xCo2Se2 samples with x = 0, 0.1, 0.2 and 0.3. The mor-
phology of samples was investigated via scanning electron microscopy
(SEM, Hitachi S-4800) and transmission electron microscopy (together
with the selected area electron diffraction (SAED)) (TEM, JEOL JEM-
2100F). The component analysis of Co, Se and K was made by energy
dispersive X-ray spectroscopy (EDS). The result for each sample was
obtained based on the average of 7–10 sets of data. The valence state of
Co was investigated by X-ray absorption fine structure (XAFS) which
was collected on the 1W1B beamline at Beijing Synchrotron Radiation
Facility (BSRF). The samples were recorded at the Co K-edge
(E0 = 7709 eV) at room temperature. The magnetization susceptibility
was measured by using a vibrating sample magnetometer (VSM,
Quantum Design). The electrical resistivity was measured with the cold-
pressed samples under a uniaxial stress of 600 kg·cm−2 via a typical
four-probe method on physical property measurement system (PPMS).
2.3. First-principles calculation
Electronic structure calculations were performed using the CASTEP
program with plane-wave pseudopotential method [27]. Generalized
gradient approximation (GGA) in the form of the Perdew-Burke-Ern-
zerhof was chosen to solve the exchange-correlation potentials. The
ultrasoft pseudopotential with a plane-wave energy cutoff of 330 eV
and a Monkhorst Pack k-point separation of 0.04 Å−1 in the reciprocal
space were used for all the calculations [28]. The self-consistent field
was set as 10−6 eV/atom. A 2 × 2 × 1 supercell with Co vacancy sites
was built to simulate the K0.75Co2Se2 with 25% K vacancies.
3. Results and discussion
The obtained K1−xCo2Se2 samples are dark microcrystalline solids,
which are unstable to air and moisture. So the samples were kept in the
glove box before they were mounted for characterization. Fig. 1 shows
the PXRD patterns for a series of K1−xCo2Se2 with nominal x = −0.2,
0, 0.1, 0.2, 0.3, 0.4 and 0.5 collected at room temperature. The main
phase of each pattern can be well indexed by the body-centered tetra-
gonal cell with space group I4/mmm (No. 139). For sample with excess
K (x = −0.2) an impurity marked by # is assigned to K2CoSe2 (ICSD
PDF: 79–2149). And for samples with x = 0.4 and 0.5, binary CoSe
(ICSD PDF: 70–2870) was observed, marked by asterisk. It can be seen
that the impurity CoSe also arises in sample with x = 0.3 but with a
very small amount. For samples with x = 0, 0.1 and 0.2, pure phase can
be obtained. It suggests that the K1−xCo2Se2 shows solid solution re-
spect to K content with the boundary around x = 0.3. This solid solu-
tion could be further confirmed by the inset of Fig. 1, where the (0 0 2)
and (2 0 0) peaks gradually shift to lower and higher 2θ angles re-
spectively with the nominal x increased from 0 to 0.3.
Rietveld refinements were carried out on PXRD patterns of samples
with x = 0–0.3 to solve the structure of these tetragonal phases. A ty-
pical refinement of sample x = 0.3 is plotted in Fig. 2a, showing a sa-
tisfactory level for fitting the experimental pattern. All the obtained
structural parameters are listed in Table 1. It needs to be mentioned
that only sample x = 0.3 contains 6.6 at.% CoSe, while the other three
show no diffraction peak from any impurity. The K-site occupations for
different samples show significant deviation from each other, varying
from 0.98(1) (x = 0) to 0.69(1) (x = 0.3). On the contrary, both Co and
Se sites are fully occupied. So it is indicated that the vacancies of K are
tunable in this ThCr2Si2-type compound with maximum content ~30%
(Fig. 2b). The observation of hexagonal CoSe impurity in sample
x = 0.3 could be mainly resulted from the loss of alkali metal during the
sample preparation, which leads to less K content in the reaction. EDS
measurements were used to confirm the variation of K content in these
samples and the measured chemical compositions of sample x = 0, 0.1,
2. Experimental
2.1. Synthesis
The polycrystalline samples K1−xCo2Se2 (with x = −0.2, 0, 0.1,
0.2, 0.3, 0.4 and 0.5) were synthesized by solid state reaction. Powders
of Co (99.8%) and Se (99.95%) obtained from Alfa Aesar were mixed
and ground in an agate mortar, followed by cold-pressing into disks
(1 mm in diameter) with a 200 kg/cm2 uniaxial stress. The disks were
put into the alumina crucibles together with the cut K ingot (97%, from
China National Accord Medicines Corporation) with predetermined
compositions and sealed into quartz tubes under vacuum (10−2 mbar).
The tubes were heated to 473 K with the heating rate of 100 K/h, held
for 12 h and then to 973 K for 48 h to obtain the precursors. After
grinding, the samples were re-heated at 923 K for 72 h to obtain the
homogeneous phase and then cooled down to room temperature by
turning off the furnace. All manipulations for sample preparation were
carried out inside an argon-filled glove box (O2 < 1 ppm) in order to
prevent the oxidation.
2.2. Characterization
Powder X-ray diffraction (PXRD) of the nominal samples
K
1−xCo2Se2 were collected at room temperature on a PANalytical dif-
fractometer (X’Pert PRO MRD) equipped with CuKα radiation
(λ = 1.5148 Å) operation at 40 kV and 40 mA and a diffracted-beam
graphite monochromator in a reflection mode (step = 0.017°2θ, scan
speed = 0.07 s/step for phase define and 0.47 s/step for structure
analysis). Temperature-dependent in situ PXRD was performed at 200
and 100 K by using a Rigaku SmartLab SE instrument (CuKα radiation)
equipped with an Anton Paar nonambient sample stage. Rietveld re-
finements were performed using the FULLPROF package [26]. The
0.2 and 0.3 are
K1.02Co1.96Se2, K0.87Co2Se2, K0.76Co1.98Se2 and
K
0.67Co2.01Se2 respectively, which are comparable to our refinements. A
2