Suppression of superconductivity by V-doping and possible magnetic order in Sr2VO3FeAs

Suppression of superconductivity by V-doping
and possible magnetic order in Sr2VO3FeAs
Marcus Tegel,1 Tanja Schmid,1 Tobias Stu¨rzer,1 Masamitsu
Egawa,1 Yixi Su,2 Anatoliy Senyshyn,3 and Dirk Johrendt1, ∗
1Department Chemie, Ludwig-Maximilians-Universita¨t Mu¨nchen,
Butenandtstraße 5-13 (Haus D), 81377 Mu¨nchen, Germany
2Ju¨lich Centre for Neutron Science, IFF, Forschungszentrum Ju¨lich,
Outstation at FRM II, Lichtenbergstraße 1, 85747 Garching, Germany
3Institute for Materials Science, Darmstadt University of Technology, 64287 Darmstadt, Germany
(Dated: September 8, 2010)
Superconductivity at 33 K in Sr2VO3FeAs is completely suppressed by small amounts of V-doping
in Sr2VO3(Fe0.93(±0.01)V0.07(±0.01))As. The crystal structures and exact stoichiometries are deter-
mined by combined neutron- and x-ray powder diffraction. Sr2VO3FeAs is shown to be very sensitive
to Fe/V mixing, which interferes with or even suppresses superconductivity. This inhomogeneity
may be intrinsic and explains scattered reports regarding Tc and reduced superconducting phase
fractions in Sr2VO3FeAs. Neutron diffraction data collected at 4 K indicates incommensurate mag-
netic ordering of the V-sublattice with a propagation vector q≈ (0,0,0.306). This suggests strongly
correlated vanadium, which does not contribute significantly to the Fermi surface of Sr2VO3FeAs.
PACS numbers: 74.70.Xa, 74.62.Dh, 74.62.En, 61.05.fm, 61.05.C-
The discovery of iron pnictide superconductors1–3 has
opened a new chapter in superconductor research. Enor-
mous progress has already been made with respect to
the physics of these materials, and it becomes increas-
ingly accepted that the weak magnetism inside the iron
layers plays a decisive role in superconductivity,4 even
though the specific relationship to the pairing mecha-
nism is still unclear. This weak magnetism appears as
a spin-density-wave (SDW) in all iron based materials
with simple PbFCl-5,6 and ThCr2Si2-related structures,7
and is intimately connected with nesting of cylinder-
shaped Fermi surfaces by a wave vector q = (pi, pi).8 How-
ever, no SDW of that or similar kind has been observed
in the more complex arsenides Sr2MO3FeAs (M = Sc,
Cr, V)9–11 where the isoelectronic (FeAs)1− layers are
separated by larger perowskite-like (Sr2MO3)1+ blocks.
Among them, only the V-compound Sr2VO3FeAs is su-
perconducting up to 37 K11 and it has been controver-
sially argued whether in this case the V-atoms signifi-
cantly contribute to the Fermi surface12 or not13. If this
is not the case, the topology of the Fermi surface turns
out to have the same essential features as in the other
FeAs superconductors, otherwise one has to assume an-
other mechanism in the case of Sr2VO3FeAs. The latter
seems to be improbable with respect to the Tc, which is
strikingly similar to other iron arsenide superconductors.
If we assume for the moment that the superconducting
mechanism in Sr2VO3FeAs is the same as in the other
iron arsenides, the absence of any SDW anomaly may
suggest that the FeAs layer is intrinsically doped. Vana-
dium can easily adopt oxidation states between V1+ and
V4+ and would thus be able to supply or to accept elec-
trons from the FeAs layers. Indeed, a recent x-ray ab-
sorption study indicates the presence of V3+ and V4+ in
a Sr2VO3FeAs sample.14 But even in this case, it remains
confusing that also Sr2ScO3FeAs, where the scandium
valence is fixed to Sc3+, shows neither a SDW nor any
other magnetic effect.9
However, thorough investigations of Sr2VO3FeAs are
hampered by the poor quality of the samples, which al-
ways contain significant amounts of the ternary vana-
dium oxides Sr2(VO4), and/or Sr3V2O7−δ.11,14,15 How-
ever, quantitative phase fractions are not specified by the
authors. This might be dangerous in the present case, be-
cause these impurities contain V4+ and exhibit at least
weak paramagnetism.
Another fact that is hardly noticed so far concerns the
true stoichiometry and homogeneity of the Sr2MO3FeAs
compounds. The ionic radii of Fe2+, V2+/3+, and Cr3+
are similar, and in face of synthesis temperatures above
1000oC, their mixing is easily conceivable. Indeed, we
have recently found by neutron diffraction, that the
chromium compound Sr2CrO3FeAs is not stoichiomet-
ric, but intrinsically Cr-doped in the iron layer.16 We
suggest that this mixing probably poisons superconduc-
tivity. Such mixing of iron and vanadium can also occur
in Sr2VO3FeAs, and already small amounts of V in the
Fe layer may seriously affect the electronic and magnetic
properties.
In order to shed some light on this puzzling material,
we have optimized the synthesis of Sr2VO3FeAs in order
to minimize the amounts of impurity phases, which al-
lowed us to perform combined x-ray and neutron powder
scattering investigations. In this paper, we report on the
exact stoichiometry, magnetism and superconductivity of
differently prepared Sr2VO3FeAs-samples. We show that
the superconducting phase is almost ideally stoichiomet-
ric, whereas small V-doping, which can easily be achieved
in the Fe-layer, suppresses superconductivity. We present
susceptibilty data and discuss hints to magnetic ordering
of the V-sublattice from neutron scattering data.
Sr2VO3FeAs was synthesized by heating mixtures of
ar
X
iv
:1
00
8.
26
87
v2
[
co
nd
-m
at.
su
pr
-co
n]
7
Se
p 2
01
0

2strontium, vanadium, iron (III) oxide and arsenic oxide
in a molar ratio of 20:11:5:5. Two separate batches of 1
gram were prepared in alumina crucibles sealed in silica
ampoules under an atmosphere of purified argon. Each
mixture was heated to 1323 K at a rate of 60 K/h, kept at
this temperature for 60 h and cooled down to room tem-
perature. The products were homogenized in an agate
mortar, pressed into pellets and sintered at 1323 K for
60 h. The latter step was performed twice. The batches
were then united, reground, pressed into pellets of 6 mm
in diameter and sintered together at 1323 K for 68 h.
The obtained black crystalline product Sr2VO3FeAs is
stable in air. Sr2VO3(Fe0.93V0.07)As was synthesized by
heating mixtures of strontium, vanadium, iron (III) ox-
ide, arsenic oxide and vanadium (V) oxide in a ratio of
100:54:20:25:3 in two separate batches accordingly.
Resistivity measurements of the undoped sample show
a rather broad superconducting transition at 33 K
(Fig. 1). Superconductivity is verified by the zero-
field-cooled field-cooled measurements using a Quan-
tum Design MPMS-XL5 SQUID magnetometer, how-
ever, the estimated superconducting volume fraction is
only 20%. No superconductivity is detected in the V-
doped sample. The susceptibilities of V-doped and un-
doped Sr2VO3FeAs measured at 1000 Oe and hysteresis
loops at different temperatures are depicted in Fig. 2.
Both samples exhibit Curie-Weiss-like paramagnetic be-
havior between 160 and 390 K. Anomalies appear at
≈ 150 K, ≈ 70 K and ≈ 50 K in the stoichiometric sam-
ple, and similar also in the V-doped sample. Almost
the same behavior has recently been reported in Ref.14,
where the authors suggest possible magnetic transitions
of the Sr2VO3-layers and the absence of a structural tran-
sition, both in agreement with our neutron scattering re-
sults (see below).
FIG. 1. Resistivity of Sr2VO3FeAs. Inset: Zero-field-cooled
/ field-cooled measurements (20 Oe).
X-Ray powder diffraction patterns at room tempera-
ture were recorded using a STOE STADI P diffractome-
ter (Cu Kα1, λ = 0.154056 nm). Neutron powder diffrac-
tion patterns at 300 K and 4 K were recorded at the
high resolution powder diffractometer SPODI at FRM
II (Garching, Germany) with incident wavelengths of
0.155 nm and 0.146 nm, respectively. Rietveld refine-
FIG. 2. (Color online) Molar susceptibility of
Sr2VO3FeAs (blue triangles) and Sr2VO3(Fe0.93V0.07)As (red
circles) at 1 kOe. Insets: Hysteresis loops of
Sr2VO3FeAs (blue) and Sr2VO3(Fe0.93V0.07)As (red).
ments were performed with the TOPAS package17 using
the fundamental parameter approach as reflection pro-
files. Vanadium was used as a sample container for the
neutron measurements and had to be included in the re-
finements. In order to describe small peak half width
and shape anisotropy effects of the samples, the approach
of Le Bail and Jouanneaux18 was implemented into the
TOPAS program and the according parameters were al-
lowed to refine freely. Preferred orientation of the crys-
tallites were described using March Dollase or spherical
harmonics functions. The Fe:V ratio at both the iron and
the vanadium site was determined by refining the occu-
pancy of the neutron and/or x-ray powder data, oxygen
deficiency was ruled out by refining the occupancy of all
oxygen sites.
The crystallographic data of Sr2VO3FeAs and
Sr2VO3(Fe0.93V0.07)As are compiled in Table I, the pow-
der patterns and Rietveld fits are depicted in Fig. 3.
The amounts of impurity phases were determined by
quantitative Rietveld analysis. The undoped sample con-
sists of Sr2VO3FeAs (89.0 wt%), Sr3V2O7−x (7.8 wt%),
FeAs (2.9 wt%) and traces of SrO (0.3 wt%), the doped
sample consists of Sr2VO3(Fe0.93V0.07)As (87.3 wt%)
and Sr3V2O7−x (12.7 wt%) as determined by neutron
diffraction. However, the doped sample also shows small
amounts of a further, unidentified impurity phase, which
could not be included in the refinement. The lattice pa-
rameters change only slightly on V doping (a is shortened
by 1.2 pm, c is unchanged), but the refinement of the Fe
site (neutron data) displays a mixed occupancy of Fe and
V in a ratio of 93±1% Fe : 7±1% V in the doped sample.
Since the X-ray data gives full occupancy of the 2a-site
(see Table I), we can rule out vacancies at the Fe site and
the Fe/V mixing is unambiguous. In contrast to this, the
superconducting undoped sample is almost exactly stoi-
chiometric regarding the V, Fe and O occupancies.
The neutron powder pattern of undoped, supercon-
ducting Sr2VO3FeAs at 4 K shows weak additional peaks

3FIG. 3. (Color online) X-Ray (top) and neu-
tron (bottom) powder patterns of Sr2VO3FeAs and
Sr2VO3(Fe0.93V0.07)As with Rietveld refinements. Inset:
Crystal structure of Sr2VO3FeAs.
at the Q-values 0.525, 1.49 and possibly 0.28 A˚−1, marked
by arrows in Fig. 4. These can be indexed as satel-
lites of the (00l) reflections according to Q(001) + ∆Q,
Q(004)−∆Q and Q(001)−∆Q with ∆Q ≈ 0.123 A˚−1, re-
spectively. This suggests incommensurate, possibly heli-
cal magnetic ordering along the c-axis with a propaga-
tion vector q ≈ (0,0,0.307) r.l.u.. Since low-temperature
57Fe-Mo¨ssbauer-data show no signal splitting,14 we ex-
pect ordering of the V-sublattice. This supports the idea
of highly correlated vanadium in Sr2VO3FeAs, where
vanadium d-states are removed from the Fermi level by
the magnetic exchange splitting. Recent angle resolved
photoemission experiments19 are in agreement with this
model likewise. Even though our data strongly suggests
the existence magnetic ordering, we consider them as pre-
liminary. Further experiments in the low-Q region by
polarization analysis are required for a precise determi-
nation of the magnetic structure and of the temperature
dependence of the magnetic order parameter.
FIG. 4. (Color online) Neutron powder patterns of
Sr2VO3FeAs at 4 K (top) and 300 K (bottom). Magnetic
reflections are marked arrows.
In summary, we have shown that Sr2VO3FeAs is sen-
sitive to Fe/V mixing in the FeAs layer, which is detri-
mental to superconductivity. Small V-doping of 7%, un-
ambiguously detected by neutron diffraction, suppresses
superconductivity completely, while the superconducting
phase is nearly stoichiometric. We suggest that even
smaller Fe/V inhomogeneities are intrinsic in this ma-
terial, and may be responsible for the scattered crit-
ical temperatures and superconducting phase fractions
reported in the literature. Small but significant addi-
tional reflections emerge in the neutron powder pattern
of superconducting Sr2VO3FeAs at 4 K. A preliminary
analyis indicates incommensurable, possibly helical mag-
netic ordering of the V-moments with a propagation vec-
tor q≈ (0,0,0.306). This is in agreement with strongly
correlated vanadium, which does not significantly con-
tribute to the Fermi surface. Thus, Sr2VO3FeAs fits to
the other iron pnictide superconductors and represents no
new paradigm, although the absence of iron magnetism
and possible self doping effects remain open questions.
ACKNOWLEDGMENTS
This work was financially supported by the German
Research Foundation (DFG), Project No. JO257/6-1
∗ johrendt@lmu.de
1 Y. Kamihara, T. Watanabe, M. Hirano, and H. Hosono, J.
Am. Chem. Soc. 130, 3296 (2008).
2 M. Rotter, M. Tegel, and D. Johrendt, Phys. Rev. Lett.

4TABLE I. Crystallographic data of Sr2VO3FeAs and Sr2VO3(Fe0.93V0.07)As at 300K. Space group P 4nmm (o1).
Sr2VO3FeAs Sr2VO3(Fe0.93V0.07)As
diffractometer STOE (x-ray) SPODI (neutron) STOE (x-ray) SPODI (neutron)
wave length (nm) 0.154 0.155 0.154 0.146
a (pm)a 394.58(1) 393.5(1) 393.38(1) 393.1(1)
c (pm)a 1573.1(1) 1570(1) 1573.1(1) 1573(1)
V (nm3)a 0.2449(1) 0.243(1) 0.2434(1) 0.243(1)
Z 2 2 2 2
data points 10650 2955 10650 3029
reflectionsb 130 196 130 238
total variables 101 95 94 113
d range 0.941 − 15.735 0.797 − 15.700 0.939 − 15.717 0.749 − 15.719
RP , wRP 0.0147, 0.0196 0.0317, 0.0412 0.0187, 0.0264 0.0321, 0.0405
RBragg 0.0050 0.0101 0.0044 0.0231
Atomic parameters:
Sr1 [2c (0, 12 , z)] z = 0.8100(1) z = 0.8096(1) z = 0.8096(2) z = 0.8092(2)
Uiso = 131(5) Uiso = 39(5) Uiso = 171(6) Uiso = 49(7)
Sr2 [2c (0, 12 , z)] z = 0.5859(1) z = 0.5858(2) z = 0.5862(2) z = 0.5852(2)
Uiso = 220(7) Uiso = 96(6) Uiso = 239(9) Uiso = 82(7)
V/Fe [2c (0, 12 , z)] z = 0.3080(3) z = 0.306(2) z = 0.3086(3) z = 0.309(3)
Uiso = 167(9) Uiso = 38c Uiso = 179(11) Uiso = 25c
occ. 1.00(1):0.00(1)d occ. 1.00(1):0.00(1)d
Fe/V [2a (0, 0, 0)] Uiso = 134(10) Uiso = 43(5) Uiso = 131(13) Uiso = 40(6)
occ. 0.99(1)d occ. 0.99(1):0.01(1)d occ. 1.00(1)d occ. 0.93(1):0.07(1)d
As [2c (0, 12 , z)] z = 0.0891(2) z = 0.0896(2) z = 0.0900(2) z = 0.0901(2)
Uiso = 113(7) Uiso = 34(6) Uiso = 178(10) Uiso = 52(7)
O1 [4f (0, 0, z)] z = 0.2921(4) z = 0.2933(1) z = 0.2958(5) z = 0.2951(1)
Uiso = 93(19)e Uiso = 56(5) Uiso = 305(31) Uiso = 65(7)
occ. 1.00(1)d occ. 1.00(1)d occ. 1.00(1)d
O2 [2c (0, 12 , z)] z = 0.4310(4) z = 0.4297(1) z = 0.4313(7) z = 0.4308(2)
Uiso = 93(19)e Uiso = 94(8) Uiso = 419(54) Uiso = 131(10)
occ. 1.00(1)d occ. 1.00(1)d occ. 1.00(1)d
a separate refinements without anisotropy paramters
b main phase
c restrained as minimum value
d restrained to a total occupancy 6 1
e Uiso of O1 and O2 were refined together
101, 107006 (2008).
3 X. C. Wang, Q. Q. Liu, Y. X. Lv, W. B. Gao, L. X. Yang,
R. C. Yu, F. Y. Li, and C. Q. Jin, Solid State Commun.
148, 538 (2008).
4 I. Mazin and M. D. Johannes, Nature Phys. 5, 141 (2009).
5 C. de la Cruz, Q. Huang, and J. W. Lynn, Nature 453,
899 (2008).
6 D. R. Parker, M. J. Pitcher, P. J. Baker, I. Franke, T. Lan-
caster, S. J. Blundell, and S. J. Clarke, Chem. Commun.
16, 2189 (2009).
7 M. Rotter, M. Tegel, I. Schellenberg, W. Hermes,
R. Po¨ttgen, and D. Johrendt, Phys. Rev. B 78, 020503(R)
(2008).
8 I. I. Mazin, D. J. Singh, M. D. Johannes, and M. H. Du,
Phys. Rev. Lett. 101, 057003 (2008).
9 H. Ogino, Y. Katsura, S. Horii, K. Kishio, and J. Shi-
moyama, Supercond. Sci. Tech. 22 (2009).
10 M. Tegel, F. Hummel, S. Lackner, I. Schellenberg,
R. Po¨ttgen, and D. Johrendt, Z. Anorg. Allg. Chem. 635,
2242 (2009).
11 X. Y. Zhu, F. Han, G. Mu, P. Cheng, B. Shen, B. Zeng,
and H. H. Wen, Phys. Rev. B 79, 220512 (2009).
12 K. W. Lee and W. E. Pickett, EPL 89, 57008 (2010).
13 I. Mazin, Phys. Rev. B 81, 020507 (2010).
14 G. Cao, Z. Ma, and X. Wang, arXiv:1007.3980 (2010).
15 A. Pal, A. Vajpayee, R. S. Meena, H. Kishan, and V. P. S.
Awana, J. Supercond. Nov. Magn. 22, 619 (2009).
16 M. Tegel, F. Hummel, Y. Su, T. Chatterji, M. Brunelli,
and D. Johrendt, EPL 89, 37006 (2010).
17 A. Coelho, TOPAS-Academic, version 4.1 (Coelho Soft-
ware, Brisbane, 2007).
18 A. Le Bail and A. Jouanneaux, J. Appl. Cryst. 30, 265
(1997).
19 T. Qian, N. Xu, Y.-B. Shi, K. Nakayama, P. Richard,
T. Kawahara, T. Sato, T. Takahashi, M. Neupane, Y.-M.
Xu, et al., arxiv:1008.4905 (2010).