Page 1
A 57Fe Mössbauer Spectroscopy Study of the 7 K
Superconductor LaFePO
Marcus Tegela, Inga Schellenbergb, Rainer Pöttgenb, and Dirk Johrendta, *
a Department Chemie und Biochemie, Ludwig-Maximilians-Universität
München, Butenandtstrasse 5–13 (Haus D), D-81377 München, Germany
b Institut für Anorganische und Analytische Chemie, Universität Münster,
Corrensstrasse 30, D-48149 Münster, Germany
Reprint requests to D. Johrendt. E-mail: dirk.johrendt@cup.uni-muenchen.de
Z. Naturforsch. 2008, 63b, ###–###; received …………
((Heading: M. Tegel et al. . 57Fe Mössbauer Spectroscopy on LaFePO))
A polycrystalline sample of superconducting LaFePO was prepared in a tin flux at
1123 K. The structure was determined from single crystal data (ZrCuSiAs-type,
P4/nmm, a = 3.9610(1), c = 8.5158(2) Å, Z = 2) and the phase analysis was
performed by the Rietveld method. LaFePO is Pauli-paramagnetic and becomes
superconducting at 7 K after removing the ferromagnetic impurity phase Fe2P from
the sample. 57Fe Mössbauer spectroscopy measurements at 298, 77, 4.2 and 4 K
show single signals at isomer shifts around 0.35 mm/s, subject to weak quadrupole
splitting. At 4 K, a symmetric line broadening appears, resulting from a small
transferred magnetic hyperfine field of 1.15(1) T and accompanied by an angle of
54.7(5)° between Bhf and Vzz, the main component of the electric field gradient
tensor.
Key words: Pnictide Oxide, Iron, Superconductivity, Mössbauer Spectroscopy
Page 2
2
Introduction
The recent discovery of high-TC superconductivity in the fluoride doped
quaternary iron arsenide oxides REFeAs(O1−xFx) (RE = La-Sm) [1] with critical
temperatures (TC) up to 55 K [2] has sparked a tremendous interest in compounds
with the ZrCuSiAs-type structure [3]. The parent compound LaFeAsO does not
become superconducting, but shows a spin density wave (SDW) anomaly at 150 K
and antiferromagnetic ordering below 138 K [4, 5]. When doping the oxide site with
fluoride, the SDW is suppressed and superconductivity occurs at critical
temperatures up to 41 K in the iron arsenide oxide LaFeAs(O1−xFx) [6]. On the other
hand, the iron phosphide oxide LaFePO had previously been reported to become
superconducting at 3.5-4.1 K even in the undoped case [7, 8]. Upon fluoride doping,
the TC had been reported to increase slightly to 6 K. Such large differences in TC
between the phosphide and arsenide oxides do not occur in the isotypic nickel
compounds LaNiPO [9,10] and LaNiAsO [11]. Both compounds are superconductors
around 2-4 K and up to now, their low transition temperatures could not be increased
significantly by doping.
These results emphasize the exceptional position of the iron arsenide oxides. One
key factor for their higher TC obviously is the existence of a SDW in the undoped
phase, which becomes unstable around 150 K, before it locks into an
antiferromagnetic spin ordering. This suggests that spin fluctuations play an
important role in the pairing mechanism, similar as in the high-TC cuprates. It is not
known up to now, if a comparable magnetic anomaly also occurs in superconducting
LaFePO. However, the much lower TC and its insignificant increase upon doping
could imply that the pairing mechanism is different in phosphide and arsenide
oxides.
Recent investigations on the magnetic properties of doped and undoped LaFeAsO
by 57Fe Mössbauer spectroscopy have proved spin ordering in LaFeAsO and its
suppression upon doping [12, 13]. In order to shed light on a potentially different
nature of superconductivity in the corresponding phosphide oxide, we present 57Fe
Mössbauer spectra, magnetic measurements and structural details of LaFePO in this
paper.
Page 3
3
Experimental
Synthesis
A polycrystalline sample of LaFePO was synthesized by heating a mixture of La,
Fe, Fe2O3 and red P in a ratio of 3:1:1:3 (a total of ~650 mg) in a tin flux (2.5 g tin)
[14] under argon atmosphere by using an alumina crucible in a silica ampoule. The
sample was heated at 1123 K (40 K/h) for 7 days and slowly cooled to room
temperature (15 K/h). The tin ingot was dissolved in 6 M HCl at room temperature.
This procedure yielded a black powder consisting of platelet crystals with a metallic
lustre.
X-Ray diffraction
X-ray powder patterns of the LaFePO sample were recorded on a Stoe Stadi-P
diffractometer (Mo-Kα1, Ge(111)-monochromator, λ = 70.93 pm, Si as external
standard). The powder data were analyzed by the Rietveld method using the GSAS
suite [15]. Single crystal data of LaFePO were collected using an Enraf-Nonius κ-
CCD equipped with a rotating anode (Mo-Kα radiation, λ = 71.073 pm). Intensities
were corrected for absorption with SADABS [16] and refined against F2 using
SHELXL97 [17] with anisotropic displacement parameters for all atoms except
oxygen. Starting parameters were taken from isostructural PrFePO [18]. Details
about the crystal structure determination may be obtained from: Fach-
informationszentrum Karlsruhe, D-76344 Eggenstein-Leopoldshafen, Germany,
e-mail: crysdata@fiz-karlsruhe.de, on quoting the registration No. CSD-391428.
Magnetic measurements
Magnetization measurements were performed using a Quantum-Design SQUID
magnetometer (MPMS-XL5) between 1.8 and 300 K. A powdered sample of
LaFePO was placed into a gelatin capsule and fixed in a straw as sample holder.
Zero-field-cooling (Shielding) and field-cooling (Meissner) measurement cycles
were performed at 10 Oe between 1.8 and 12 K in the reciprocating sample option
(RSO) mode. The magnetic susceptibility between 2 and 300 K was measured at 10
kOe.
Page 4
4
57Fe Mössbauer spectroscopy
A 57Co/Rh source was available for the 57Fe Mössbauer spectroscopy
investigations. The LaFePO sample was placed in a thin-walled PVC container at a
thickness of about 10 mg Fe/cm2. The measurements were run in the usual
transmission geometry at 298, 77, 4.2, and 4 K. The source was kept at room
temperature.
Results and Discussion
The ZrCuSiAs-type structure [19] was confirmed by single crystal data. LaFePO
is build up by LaO- and FeP- layers alternating along [001]. Lanthanum is eightfold
coordinated by four oxygen (2.355 Å) and four phosphorous atoms (3.343 Å). The
Fe−P bonds lengths (2.295 Å) are close to the sum of the covalent radii of iron and
phosphorous (2.26 Å, [20]). The FeP4 tetrahedra are slightly flattened along [001], as
it can be recognized by a P−Fe−P angle of 119.3°. This distortion is slightly bigger
than in LaFeAsO with an As−Fe−As angle of 113.7° (at 175 K) [4], but much weaker
than that of the NiP4 tetrahedra in LaNiPO, where the P−Ni−P angle is 126.5° [9].
The Rietveld analysis revealed that the sample was not single phase, small
amounts of Fe2P and FeSn2 could be detected. Such magnetic impurity phases can be
very destructive for susceptibility measurements. Fe2P is ferromagnetic below 266 K
[21] and FeSn2 is antiferromagnetic below 380 K [22], thus even small traces of Fe2P
would strongly affect the magnetic measurements. This is especially important with
respect to the weak Pauli-paramagnetic LaFePO. Indeed, our first susceptibility
measurement of the LaFePO sample showed a strong upturn of χ below 270 K,
which coincides with the Curie point of Fe2P. Also no superconductivity could be
detected at temperatures down to 1.8 K. In order to remove impurities, we separated
ferromagnetic particles by stirring a suspension of the finely grounded LaFePO
sample in liquid N2 with a strong permanent magnet. After this treatment, the Fe2P
impurity was drastically reduced and we could successfully fit the complete X-ray
powder pattern with phase fractions of 96 % LaFePO and 4 % FeSn2, respectively.
However, the magnetic measurement of this purified sample still revealed small
traces of a residual Fe2P impurity, as depicted in Figure 1. But now the sample
Page 5
5
exhibited the typical strong diamagnetic shielding of a superconductor. The complete
shielding- and Meissner-cycle (Figure 2) clearly shows the onset of
superconductivity at 7 K, which is significantly higher than 4.1 K as it was reported
for undoped LaFePO [7, 8]. Obviously, the superconductivity has been completely
suppressed by the ferromagnetic impurity phase, which generates a magnetic field
inside the sample. If the critical field of the superconducting phase is sufficiently
small, this additional field can decrease TC or even bring it to zero. Our
magnetization measurements at 1.8 K showed small critical fields of Hc1 ≈ 75 Oe and
Hc2 ≈ 880 Oe for LaFePO, which correspond approximately to the values given by
Hosono [8].
Our results indicate that ferromagnetic impurities can strongly influence the
superconductivity in LaFePO and presumably also in the LaFeAsO compounds. This
may be one reason for the partially different TC’s of supposedly identical compounds,
which are in most cases far from being single phase. However, in the case of
fluoride doped compounds, the exact amount of fluoride in the structure may be
another important problem.
57Fe Mössbauer spectra of LaFePO recorded at 298, 77, 4.2, and 4 K are presented
in Figure 3 together with transmission integral fits. The corresponding fitting
parameters are listed in Table 2. In agreement with the ZrCuSiAs-type crystal
structure, the spectra were well reproduced with single iron sites at isomer shifts
around 0.3 mm/s, slightly smaller than the isomer shifts observed recently for
LaFeAsO and LaFeAsO0.89F0.11 [12,13]. Due to the non-cubic site symmetry, the
spectra are subject to weak quadrupole splitting.
Similar to LaFeAsO and LaFeAsO0.89F0.11, also for LaFePO we observe a slight
increase of the isomer shift with decreasing temperature. For iron, a smaller isomer
shift is consistent with a higher electron density at the nuclei [23]. The isomer shifts
observed for LaFePO are comparable with other iron phosphides with tetrahedrally
coordinated iron [24, 25].
At 298 and 77 K, we observe no magnetic hyperfine field splitting, clearly
manifesting the absence of magnetic ordering, similar to the fluoride doped
End of preview.
