Structural and magnetic phase diagram of CeFeAsO1-xFx and

its relationship to high-temperature superconductivity

Jun Zhao1, Q. Huang2, Clarina de la Cruz1,3, Shiliang Li1, J. W. Lynn2, Y. Chen2,4, M. A.
Green2,4, G. F. Chen5, G. Li5, Z. Li5, J. L. Luo5, N. L. Wang5, and Pengcheng Dai1,3,*

Department of Physics and Astronomy, The University of Tennessee, Knoxville,

Tennessee 37996-1200, USA

2

NIST Center for Neutron Research, National Institute of Standards and Technology,

Gaithersburg, Maryland 20899-6012 USA

3

Neutron Scattering Science Division, Oak Ridge National Laboratory, Oak Ridge,

Tennessee 37831, USA

4

Department of Materials Science and Engineering, University of Maryland, College

Park, Maryland 20742-6393 USA

5

Beijing National Laboratory for Condensed Matter Physics, Institute of Physics,

Chinese Academy of Sciences, Beijing 100080, China

*

To whom correspondence should be addressed. E-mail: daip@ornl.gov

Abstract
We use neutron scattering to study the structural and magnetic phase transitions in the
iron pnictides CeFeAsO1-xFx as the system is tuned from a semimetal to a high-transitiontemperature (high-Tc) superconductor through Fluorine (F) doping x. In the undoped
state, CeFeAsO develops a structural lattice distortion followed by a stripe like
commensurate antiferromagnetic order with decreasing temperature. With increasing
1

Fluorine doping, the structural phase transition decreases gradually while the
antiferromagnetic order is suppressed before the appearance of superconductivity,
resulting an electronic phase diagram remarkably similar to that of the high-Tc copper
oxides. Comparison of the structural evolution of CeFeAsO1-xFx with other Fe-based
superconductors reveals that the effective electronic band width decreases systematically
for materials with higher Tc. The results suggest that electron correlation effects are
important for the mechanism of high-Tc superconductivity in these Fe pnictides.

A determination of the structural and magnetic phase transitions in doped

transition metal oxides is essential for understanding their electronic properties. For
high-transition-temperature (high-Tc) copper oxides, the parent compounds are
antiferromagnetic (AFM) Mott insulators1. When mobile electrons or holes are doped
into the parent compounds, the static long-range AFM order is rapidly suppressed and
optimal superconductivity emerges after a complete suppression of the static AFM
order2,3. Much like copper oxide superconductors, high-Tc superconductivity in the
recently discovered rare-earth Fe-based oxide systems RFeAsO (R, rare-earth metal) and
(Ba1-xKx)Fe2As2 are also derived from either electron4-8 or hole9,10 doping of their
semimetal parent compounds. Although the parent compound LaFeAsO also exhibits
long range static AFM order that is suppressed upon electron doping to induce
superconductivity11-13, there has been no systematic measurement to establish the doping
evolution of the AFM order and its relationship to superconductivity. A determination of
the structural, magnetic, and superconductivity phase diagram in one of the RFeAsO
systems will allow a direct comparison with the phase diagram of high-Tc copper oxides.

Such a comparison is important because it might reveal whether the physics of high-Tc
superconductivity in the Fe-based materials is fundamentally related to that of the high-Tc
copper oxides14-16.

In this paper, we report systematic neutron scattering studies of structural and

magnetic phase transitions in the Fe pnictides CeFeAsO1-xFx as the system is tuned from a
semimetal to a high-Tc superconductor through F doping x. We find that CeFeAsO
undergoes a structural lattice distortion from tetragonal to orthorhombic structure near
155 K followed by a commensurate AFM ordering on the Fe sublattice below ~140 K as
shown in Figs. 1 and 2, similar to that of LaFeAsO (ref. 11). While the structure phase
transition temperature decreases gradually with increasing F doping, the AFM ordering
temperature and static Fe ordered moment reduce rapidly and vanish before the
emergence of superconductivity, resulting an electron phase diagram similar to that of the
high-Tc copper oxides (Fig. 1D). Our detailed analysis of the low temperature
CeFeAsO1-xFx structures reveal that F doping does not change the Fe-As distance but
reduces the Ce-As distance and Fe-As-Fe angles (Fig. 3). The results suggest that the
main effect of F doping is to transfer electrons from the Ce-O/F layers to the As-Fe-As
block (Fig. 3A). Comparison of the structural evolution of CeFeAsO1-xFx with other rareearth Fe pnictides8,11,17,18 and (Ba1-xKx)Fe2As2 (refs. 10, 19) suggests that the effective
electronic band widths (W) in these materials are controlled by the Fe-As-Fe angle,
revealing a systematic trend that the parent compounds with smaller band widths tend to
have higher Tc upon doping (Fig. 4). The results suggest that although parent compounds

of iron pnictides are semimetals, electron correlation effects are very important to their
properties14-16.
We use neutron diffraction to study the structural and magnetic phase transitions
in polycrystalline nonsuperconducting CeFeAsO1-xFx with x = 0, 0.02, 0.04, 0.06 (as
confirmed by measurements using a commercial SQUID) and superconducting
CeFeAsO1-xFx with x = 0.16 (with the onset Tc = 35 K as determined by the susceptibility
measurement using a SQUID) using the method described in Ref. 6. Our experiments are
carried out on the BT-1 high resolution powder diffractometer and BT-7 thermal tripleaxis spectrometer at the NIST Center for Neutron Research, Gaithersburg, Maryland.
Some measurements were also performed on the HB-3 thermal triple-axis spectrometer at
the High Flux Isotope Reactor, Oak Ridge National Laboratory.

In previous work, it was found that LaFeAsO undergoes a structural distortion

below 155 K, changing the symmetry from tetragonal (space group P4/nmm) to
monoclinic (space group P112/n)11 or orthorhombic (space group Cmma)20, and followed
by a long range commensurate AFM order with a stripe like spin structure below ~137 K
(ref. 11). For convenience in comparing the low temperature nuclear and magnetic
structures, we use orthorhombic Cmma space group to describe the low temperature
structural data in this paper. Since CeFeAsO1-xFx has rare earth Ce which carries a local
magnetic moment6 and therefore different from the nonmagnetic La in LaFeAsO1-xFx (ref.
11), we first need to determine whether this material has the same lattice distortion and
magnetic structure as those of LaFeAsO1-xFx. Our high-resolution neutron powder
diffraction measurements on BT-1 confirm that the lattice symmetry of CeFeAsO also

displays the tetragonal to orthorhombic transition below ~158 K (Figs. 1D and 2A),
where the (2,2,0)T peak in the tetragonal phase is split into (0,4,0)O and (4,0,0)O peaks in
the orthorhombic phase (inset in Fig. 2A).

To see if the Fe spins in CeFeAsO exhibit the same magnetic order as that of
LaFeAsO (ref. 11), we carried out measurements on BT-7. The Ce moments order
magnetically below ~4 K (ref. 6 and Fig. 2E), we took data at 40 K to avoid any possible
induced-moment influence of Ce on the intensities of the Fe magnetic peaks (Fig. 1C).
Comparison of Fig. 1C with the same scan at 170 K (see Supporting Online Material) and
with Fig. 3c in ref. 11 for LaFeAsO immediately reveals that the Fe magnetic unit cell in
CeFeAsO can be indexed as 2aN 2bN cN, where aN, bN, and cN are nuclear lattice
parameters of the unit cell (see Table 1a). This indicates that CeFeAsO has the same
stripe-type in-plane Fe magnetic structure as that of LaFeAsO, but the c-axis nearestneighbor spins are parallel in CeFeAsO rather than anti-parallel as in LaFeAsO. Hence
there is no need to double unit cell along the c-axis (Fig. 1A), and an excellent fit to the
data is achieved using the magnetic and nuclear unit cells in Figs. 1A and 1B as shown by
the solid red line of Fig. 1C. The ordered iron moment is 0.8(1) B at 40 K, where
numbers in parentheses indicate uncertainty in the last decimal place and B denotes Bohr
magneton. The magnitude of the Fe moment in CeFeAsO is about twice that of the Fe
ordered moment in LaFeAsO (ref. 11). We also determined the Ce magnetic structure
using data collected at 1.7 K (see Supporting Online Material) and found a strong
coupling between the Fe and Ce moment below 20 K (Figs. 2E-2G). Our determined Ce
and Fe magnetic structures are shown in Figs. 1A and 1B. The lack of the c-axis unit

cell doubling in the Fe magnetic structure is likely due to magnetic interactions of rare
earth Ce ions, which doubles the nuclear unit cell along the c-axis (Fig. 1A).

Having shown that the lattice distortion and Fe magnetic unit cells are rather
similar between CeFeAsO and LaFeAsO, it is important to determine the evolution of the
lattice and magnetic structures with increasing F doping as superconductivity is induced.
If the stripe-type AFM order in CeFeAsO and LaFeAsO is a spin-density-wave (SDW)
instability arising from a nested Fermi surface21, electron doping will change the electron
and hole pocket sizes, but may23 or may not24 induce incommensurate SDW order. For
pure metallic Cr (ref. 22), where the SDW order has a long wavelength incommensurate
magnetic structure, electron/hole doping quickly locks the SDW to commensurate
antiferromagnetism with an ordered moment that is doping independent22. Figure 2
summarizes the structural and magnetic phase transition temperatures for CeFeAsO1-xFx
with x = 0, 0.02, 0.04, 0.06. Inspection of Figs. 2A-2D and their insets immediately
reveals that the onset lattice distortion temperature (seen as the initial drop in (2,2,0)T
peak intensity) and the magnitude of the lattice distortion (the low temperature splitting
of the (0,4,0)O and (4,0,0)O peaks) both decrease gradually with increasing x (Fig. 1D).
On the other hand, the wavevector positions and coherence-length limits of the (1,0,2)M
magnetic peaks [Q = 1.838(1), 1.833(1), 1.837(1), and 1.831(3) -1; and = 140(6),
137(8), 134(11), and 140(30) for x = 0, 0.02, 0.04, 0.06, respectively (see inset of Fig.
2G)] are doping independent, and indicate no observable commensurate to
incommensurate phase transition. The integrated intensity of the (1,0,2)M magnetic peak
decreases rapidly with increasing x and essentially vanishes near x = 0.06 (inset in Fig.

1D). The corresponding Nel temperatures for TN(Fe) and TN(Ce) are determined by
measuring the temperature dependence of the (1,0,2)M magnetic reflection (Figs. 2E-2H).
The resulting structural and magnetic phase diagram of CeFeAsO1-xFx together with
superconducting transition temperatures6 is summarized in Fig. 1D.

Figure 3 summarizes the impact of F-doping on the crystal structure of

CeFeAsO1-xFx obtained from our detailed refinement analysis of the BT-1 data. The
undoped CeFeAsO has an orthorhombic low-temperature structure with c > a > b (Fig.
3A). Doping fluorine gradually suppresses both the a (the long Fe-Fe nearest-neighbor
distance) and c axes lattice constants while leaving the b-axis (the short Fe-Fe nearestneighbor distance) unchanged (Fig. 3B). The reduction in the c-axis lattice constant is
achieved via a large reduction of the Ce-As distance, while the Ce-O/F and As-Fe-As
block distances actually increase with increasing F-doping (Figs. 3C and 3E). This
suggests that the effect of F-doping is to bring the Ce-O/F charge transfer layer closer to
the superconducting As-Fe-As block, and thereby facilitating electron charge transfer
(Fig. 3A). Since the Fe-As distance (2.405 ) is essentially doping independent (Fig.
3E), the strong hybridization between the Fe 3d and the As 4p orbitals25 is not affected by
electron-doping. On the other hand, if we assume that the Fe-Fe nearest-neighbor (J1)
and next-nearest-neighbor exchange couplings (J2) are mediated through the electron FeAs-Fe hopping and controlled by the Fe-As-Fe angles26, Figure 3D suggests that J2 and
one of the nearest-neighbor exchange constants (J1) decrease with increasing F-doping
while the other J1 remains unchanged.

In a previous work on the phase diagram of oxygen deficient RFeAsO1- (ref. 7), it
was found that systematically replacing R from La, to Ce, Pr, Nd, and Sm in RFeAsO1-
resulted a gradual decrease in the a-axis lattice parameters and increase in Tc. If Tc for
different Fe-based superconductors is indeed correlated to their structural properties, one
would expect to find a systematic trend between Tc and the Fe-As-Fe bond angles, since
the exchange couplings (J1 and J2) and therefore the electronic band width W are directly
related to the Fe-As-Fe bond angles14,26 (Fig. 4A). Figures 4B and 4C plot the Fe-As(P)Fe angles and Fe-Fe/Fe-As(P) distances versus maximum Tc for different Fe-based rareearth oxypnictides8,10,11,17,18,27,28 and Ba1-xKxFe2As2 (ref. 17) superconductors. While the
Fe-Fe/Fe-As(P) distances may not have a clear trend amongst different Fe-based
superconductors, it is remarkable that the maximum Tc appears to be directly related to
the Fe-As(P)-Fe angles for a variety of materials (Fig. 4C). This suggests that the most
effective way to increase Tc in Fe-based superconductors is to decrease the Fe-As(P)-Fe
bond angles, and thereby reducing the electronic band width W.

In summary, we have mapped out the structural and magnetic phase transitions of
CeFeAsO1-xFx and found that the Fe static AFM order vanishes before the appearance of
superconductivity29. The phase diagram of CeFeAsO1-xFx is therefore remarkably similar
to that of the high-Tc copper oxides1-3. In addition to suppressing the static
antiferromagnetism and inducing superconductivity, F doping also reduces the long-axis
of the orthorhombic structure in the undoped CeFeAsO and decreases the Fe-As-Fe bond
angles. Comparison of structural parameters of various Fe-based superconductors reveals
that the Fe-As(P)-Fe bond angle, and therefore the electronic band width W, decreases

systematically for superconductors with increasing Tcs. If we assume that the effective
Coulomb interaction U is comparable amongst different classes of Fe-based
superconductors14, these results suggest that the controlling parameter U/W increases for
samples with higher Tcs. This means that electron correlations effects, or Mott Physics,
becomes increasingly important and should be taken into account as we consider a
mechanism for high-Tc superconductivity in these Fe-based materials.

References and Notes:

1. P. A. Lee, N. Nagaosa, and X.-G. Wen, Rev. Mod. Phys. 78, 17 (2006).
2. R. J. Birgeneau, C. Stock, J. M. Tranquada, and K. Yamada, J. Phys. Soc. Jpn. 75,
111003 (2006).
3. J. M. Tranquada, in Handbook of High-Temperature Superconductivity, edited by
J. R. Schrieffer and J. S. Brooks (Springer, New York, 2007), p. 257.
4. Kamihara, Y., Watanabe, T, Hirano, M. & Hosono, H. J. Am. Chem. Soc. 130,
3296 (2008).
5. X. H. Chen, T. Wu, G. Wu, R. H. Liu, H. Chen, and D. F. Fang, Nature 453, 761
(2008).
6. G. F. Chen et al., Phys. Rev. Lett. (in the press); Preprint at
http://arxiv.org/abs/0803.3790v2 (2008).
7. Zhi-An Ren et al., Preprint at http://arxiv.org/abs/0804.2582 (2008).
8. Jan-Willem G. Bos et al., Preprint at http://arxiv.org/abs/0806.0926v1 (2008).
9. Wen, H. H., Mu G., Fang, L., Yang, H. & Zhu, X. Y. E, Euro. Phys. Lett. 82,
17009 (2008).

10. M. Rotter, M. Tegel, and D. Johrendt, Preprint at

http://arxiv.org/abs/0805.4630v1 (2008).
11. C. de la Cruz et al., Nature 453, 899 (2008).
12. S. Kitao et al., Preprint at http://arxiv.org/abs/0805.0041 (2008).
13. J. P. Carlo et al., Preprint at http://arxiv.org/abs/0805.2186v1 (2008).
14. Q. Si and E. Abrahams, Preprint at http://arxiv.org/abs/0804.2480v1 (2008).
15. C. Fang, H. Yao, W. F. Tsai, J. P. Hu, and S. A. Kivelson, Preprint at
http://arxiv.org/abs/0804.3843v1 (2008).
16. K. Haule and G. Kotliar, Preprint at http://arxiv.org/abs/0805.0722v1 (2008).
17. M. Tegel, I. Schellenberg, R. Rttgen, and D. Johrendt, Preprint at
http://arxiv.org/abs/0805.1208 (2008).
18. T. M. QcQueen et al., Preprint at http://arxiv.org/abs/0805.2149 (2008).
19. M. Rotter et al., Preprint at http://arxiv.org/abs/0805.4021v1 (2008).
20. T. Nomura et al., Preprint at http://arxiv.org/abs/0804.3569 (2008).
21. J. Dong et al., EuroPhys. Lett. (in the press); Preprint at
http://arxiv.org/abs/0803.3426v1 (2008).
22. S. Komura, Y. Hamaguchi, and N. Kunitomi, J. Phys. Soc. Jpn. 23, 171 (1967).
23. T. A. Maier and D. J. Scalapino, Preprint at
http://arxiv.org/abs/0805.0316v1 (2008).
24. C. K. Xu, M. Mller, and S. Sachdev, Preprint at
http://arxiv.org/abs/0804.4293v1 (2008).
25. J. S. Wu, P. Phillips, and A. H. Castro Neto, Preprint at
http://arxiv.org/abs/0805.2167v1 (2008).
10

26. T. Yildirim, Preprint at http://arxiv.org/abs/0804.2252v1 (2008).

27. Y. Kamihara et al., J. Am. Chem. Soc. 128, 10012 (2006).
28. Jan-Willem G. Bos et al., Preprint at http://arxiv.org/abs/0806.1450v1 (2008).
29. C. Tarantini et al., Preprint at http://arxiv.org/abs/0805.4445v1 (2008).

Acknowledgements We thank Elbio Dagotto, Adriana Moreo, Randy Fishman, and

Thomas Maier for helpful discussions. We also thank J. L. Zarestky for his help on the
HB-3 measurements. This work is supported by the US National Science Foundation
through DMR-0756568, by the US Department of Energy, Division of Materials Science,
Basic Energy Sciences, through DOE DE-FG02-05ER46202. This work is also
supported in part by the US Department of Energy, Division of Scientific User Facilities,
Basic Energy Sciences. The work at the Institute of Physics, Chinese Academy of
Sciences, is supported by the National Science Foundation of China, the Chinese
Academy of Sciences and the Ministry of Science and Technology of China.

11

Figure 1 Low temperature magnetic structures for Ce and Fe in CeFeAsO and the
structural and magnetic phase diagram of CeFeAsO1-xFx. The data in panel C) were
collected using BT-7 with an incident beam wavelength = 2.36 with pyrolytic
graphite (PG) (0,0,2) as monochromator and PG filter. A) The three dimensional
antiferromagnetic structures of Ce and Fe as determined from our neutron diffraction
12

data. B) The magnetic unit cells of Ce and Fe. The Fe moments lie in the a-b plane and
form a stripe-type antiferromagnetic structure similar to that of LaFeAsO (ref. 11), while
nearest-neighbor spins along the c-axis are parallel and so there is no need to double the
magnetic cell along the c-axis. C) Observed (crosses) and calculated (solid line) neutron
powder diffraction intensities of CeFeAsO at 40 K using space group Cmma for nuclear
structure and A,B) for magnetic structure. The dashed vertical lines indicate the expected
nuclear Bragg peak positions while the solid vertical lines represent magnetic Bragg peak
positions for the spin structure of the right panel of B). D) The structural and magnetic
phase diagram determined from our neutron measurements on CeFeAsO1-xFx with x = 0,
0.02, 0.04, 0.06, 0.16. The red circles indicate the onset temperature of P4/nmm to
Cmma phase transition. The black squares and green triangles designate the Nel
temperatures of Fe TN(Fe) and Ce TN(Ce), respectively, as determined from neutron
measurements in Figs. 2E-2H. The superconducting transition temperatures for x = 0.08,
0.012, 0.016, 0.20 are from the onset Tc of the resistivity measurements adapted from ref.
6. We note that Tc determined from susceptibility measurements in Fig. 4 has a lower
value. The inset in D) shows the F doping dependence of the Fe moment as determined
from the intensity of the (1,0,2)M magnetic peak at 40 K, where the influence of the Ce
moment on the Fe magnetic Bragg peak intensity can be safely ignored.

13

Figure 2 Structural and magnetic phase transition temperatures as a function of

increasing F doping in CeFeAsO1-xFx. The data in A-D) and E-G) were collected on
BT-1 and BT-7, respectively. The Q-scan for x = 0.06 (inset in E) and H) were carried
out on HB-3 using similar setup as BT-7. The BT-1 diffractometer has a Ge(3,1,1)
monochromator and an incident beam wavelength of = 2.0785 . A-D) Temperature
dependence of the (2,2,0)T (T denotes tetragonal) nuclear reflection indicative of a
structural phase transition11 for various x. The insets show the (2,2,0)T reflection above
14

and below the transition temperatures11. E-H) Temperature dependence of the order
parameter at the magnetic Bragg peak position (1, 0, 2)M as a function of F doping. The
large increase in intensity below 4 K is due to Ce ordering, as confirmed by temperature
dependence of the Ce-only magnetic Bragg peak (0,0,1)M (see Supporting Online
Material). The inset shows the doping dependence of the (1,0,2)M Bragg peak
normalized to the nuclear Bragg peak intensity. The peak positions and widths are
essentially doping independent, suggesting that the AFM order is commensurate at all
doping levels.

15

Figure 3 Low temperature structural evolution of CeFeAsO1-xFx as a function of F

doping obtained from analysis of the BT-1 data. The atomic positions of CeFeAsO1xFx

are shown in Table 1b and the effect of F doping is to expand the Fe-As-Fe block and

to move the Ce-O/F block closer to Fe-As-Fe block, thereby facilitating electron doping
to the superconducting Fe-As-Fe layer. A) schematic diagram defining the Fe-As-Fe
block and illustrating the process of electron doping. B) a, b, c lattice constants of the
orthorhombic unit cell and the two Fe-Fe nearest-neighbor distances as a function of F
doping. C) Ce-O/F and Ce-As distances as a function of F doping. The slight increase in
the Ce-O/F block size is compensated by much larger reduction in the Ce-As distance,
resulting an overall c-axis lattice contraction as shown in B). D) Fe-As-Fe bond angles as
16

defined in the inset versus F doping. While angle 1 hardly changes with doping, angles 2
and 3 decrease substantially with increasing F doping. E) The Fe-As bond distance and
As-Fe-As block size versus F doping. The Fe-As distance is independent of F doping.

Figure 4. Fe-As(P)-Fe bond angles, Fe-Fe, and Fe-As(P) distances for different Febased superconductors. There is a systematic decrease in the Fe-As(P)-Fe bond angle
for Fe-based superconductors with higher Tc, suggesting that electron correlation effects
are important. A) Schematic illustration of what happens to the Fe-As-Fe tetrahedron for
Fe-based superconductors. The smaller Fe-As(P)-Fe angle on the right results in a
narrower electronic band width W, increasing U/W, and thereby raising Tc. B,C)
Dependence of the maximum-Tc on Fe-As(P)-Fe angle and Fe-Fe/Fe-As(P) distance. The
Fe-As(P)-Fe angles and Fe-Fe/Fe-As(P) distances are computed using atomic positions

17

given in Refs. 18, 27 for LaFePO; Ref. 11 for LaFeAsO; present paper for CeFeAsO;
Ref. 10 for BaFe2As2; Ref. 28 for NdFeAsO; and Ref. 8 for TbFeAsO0.9F0.1. Note here
we used the maximum-Tc obtained from susceptibility measurement, which is lower than
that of the resistivity measurement on the same system.

Table 1a. Refined structure parameters of CeFeAsO1-xFx with x = 0 at 175 K and x = 0.16
at 60 K. Space group: P4/nmm.
CeFeAsO, a = 3.99591(5), c=8.6522(1) ; CeFeAsO0.84F0.16, a = 3.98470(3), c =
8.6032(1) .
Atom site

z(x = 0)

B (2)(x = 0) z (x = 0.16)

B (2)(x =

0.16)
Ce

2c

0.1413(3)

0.34(4)

0.1480(4)

0.58(5)

Fe

2b

0.25(4)

0.09(3)

As

2c

0.6546(2)

0.28(3)

0.6565(3)

0.27(4)

2a

0.30(5)

0.50(4)

x = 0, Rp =5.02%, wRp = 6.43%, 2= 1.336;

x = 0.16, Rp =5.94%, wRp = 8.24%, 2 = 2.525.
Table 1b. Refined structure parameters of CeFeAsO1-xFx with x = 0, 0.02, 0.04, 0.06 at
1.4 K. Space group: Cmma. Atomic positions: Ce: 4g (0, , z); Fe: 4b (, 0, ), As: 4g
(0, , z), and O/F: 4a (, 0, 0).
Atom

x=0

x = 0.02

x = 0.04

x = 0.06

a ()

5.66263(4)

5.65865(9)

5.6553(1)

5.6511(1)

b ()

5.63273(4)

5.63155(9)

5.6325(1)

5.6346(1)

c ()

8.64446(7)

8.6382(1)

8.6355(2)

8.6335(1)

0.1402(2)

0.1417(4)

0.1419(4)

0.1420(3)

B (2)

0.36(2)

0.37(6)

0.31(6)

0.46(5)

Fe

B (2)

0.34(2)

0.38(4)

0.30(3)

0.34(3)

As

0.6553(1)

0.6548(3)

0.6555(3)

0.6554(2)

B (2)

0.45(2)

0.50(6)

0.36(5)

0.24(4)

Ce

18

O/F

B (2 )

0.54(2)

0.53(6)

0.64(6)

0.63(5)

Rp (%)
wRp (%)

4.31
5.60
2.192

5.44
6.72
1.258

4.90
6.31
0.966

4.71
6.16
0.9622

Supporting Online Material

Structural and magnetic phase diagram of CeFeAsO1-xFx and

its relationship to high-temperature superconductivity
Jun Zhao, Q. Huang, Clarina de la Cruz, Shiliang Li, J. W. Lynn, Y. Chen, M. A. Green,
G. F. Chen, G. Li, Z. Li, J. L. Luo, N. L. Wang, and Pengcheng Dai

Figure S1 below shows the BT-7 measurements on CeFeAsO at 1.5 K, 40 K, and 160 K.
The (0,0,1)M peak corresponds to a doubling of the c-axis unit cell for the Ce magnetic
order, which disappears above the Ce Nel temperature of ~4 K. Comparison of the 40 K
and the 160 K scans reveals the positions of the Fe magnetic peaks as shown in Fig. 1C.
Our detailed refinements for the determination of the Ce magnetic structure and its
interaction with the iron spins will be published separately.

19