Ordered Iron Oxide Nanotube Arrays of Controlled Geometry and Tunable
Magnetism by Atomic Layer Deposition
Julien Bachmann,*
,†
Jing Jing,
†
Mato Knez,
†
Sven Barth,
‡
Hao Shen,
‡
Sanjay Mathur,
‡
Ulrich Go¨sele,
†
and Kornelius Nielsch*
,†
Max Planck Institute of Microstructure Physics, Am Weinberg 2, 06120 Halle, Germany, and Leibniz Institute of
New Materials, CVD DiVision, Im Stadtwald, Geb. 43A, D-66041 Saarbruecken, Germany
Received April 18, 2007; E-mail: bachmann@mpi-halle.de; knielsch@mpi-halle.de
The interest in arrays of pseudo-one-dimensional ferromagnetic
nanoobjects stems from their potential application to high-density
data storage.
1
This communication reports on the fabrication of iron
oxide nanotubes by atomic layer deposition (ALD) and the
opportunities thus afforded in the magnetic realm.
Among the different geometric types of objects, tubes offer an
additional degree of freedom in their design as compared to wires,
in that not only the length and diameter can be varied but also the
thickness. Changes in thickness are expected to strongly affect the
mechanism of magnetization reversal and thereby the overall
magnetic responsesin particular, the remnant magnetization.
2
The
physics of ferromagnetic nanotubes, however, have remained ill-
defined to date for lack of systems amenable to systematic
experimental variations in the geometric parameters and study of
the corresponding variations in the magnetic behavior.
3
Common
experimental limitations include difficult synthesis of magnetic tubes
of sub-100 nm diameter,
2b
granularity of the tube structure,
3,4
and
inhomogeneities in thickness and/or diameter.
5
In contrast to this,
atomic layer deposition (ALD) of magnetic materials can be applied
to ordered porous substrates to yield arrays of smooth tubes with
a geometry that is (a) tightly controlled and (b) widely tunable.
6
Our focus on iron oxides results from their unique position among
the magnetic materials in terms of abundant availability and
biocompatibility.
Water and the homoleptic dinuclear iron(III) tert-butoxide
complex, Fe
2
(O
t
Bu)
6
,
7
were used for ALD of Fe
2
O
3
. When the
Fe
2
(O
t
Bu)
6
precursor was heated to 100 °C, a temperature window
was found between 130 and 170 °C in which the solid deposits in
amounts proportional to the number of reaction cycles. At lower
temperatures, deposition does not occur as judged from the absence
of color. The onset of Fe
2
(O
t
Bu)
6
thermal decomposition sets the
upper limit of ALD.
7
When the process is carried out with a self-
ordered porous anodic alumina membrane
8
as the substrate, the
internal walls are covered conformally with a smooth layer of Fe
2
O
3
,
yielding arrays of tubes of aspect ratios up to 100, the growth rate
being 0.26((0.04) Å cycle
-1
, as displayed in Figures 1 and S9.
The results are unaffected by preliminary and/or subsequent ALD
of ZrO
2
or TiO
2
, which can be used to facilitate the preparation of
samples for electron microscopy (see Supporting Information).
X-ray photoelectron spectroscopy (XPS) confirms the identity of
the deposited material to be Fe
2
O
3
, albeit with a high carbon content
(Figure S6). The amount of C decreases with depth, which could
be due to residual superficial tert-butoxide ligands unhydrolyzed
by their first exposure to water. Selected-area electron diffraction
(SAED) indicates that the tubes are nanocrystalline, with only very
short-range order. Relative to the ALD processes reported to date
for iron metal or its oxides, based either on an iron(II) amidinate
and H
2
or on an iron(III) acetylacetonate and O
2
or O
3
,
9
the novel
method simultaneously improves the growth rate, the smoothness,
and the aspect ratios accessed vastly. Its low deposition temperature
also makes it attractive for nanostructuring applications based on
selected biotemplates.
10
Reduction of the Fe
2
O
3
nanotubes in the Al
2
O
3
matrix by 5%
H
2
/95% Ar at 400 °C converts Fe
2
O
3
to Fe
3
O
4
,
11
as substantiated
by XPS (Figure S6). The transformation is accompanied by a color
change from golden or coppery brown to black, commensurate with
the decrease in band gap. The reduced tubes are air-sensitive and
were kept under inert atmosphere by depositing a macroscopic layer
of polystyrene onto the membranes. This protection allows for their
convenient handling in air without appreciable degradation over a
period of days to weeks.
The Fe
3
O
4
tubes behave as hard ferromagnets at 300 K, as shown
by SQUID data (Figure 2). At lower temperatures, the magnetic
hystereses widen, an approximate doubling in coercive field being
usually observed between 300 and 5 K (Figure S2). Temperature-
dependent magnetization data (Figure S4) do not show any abrupt
change indicative of a crystallographic phase transition (Verwey
transition),
12
an observation consistent with a glassy structural state.
Coercive fields (H
c
//
) as high as 76.5((1.5) mT (765((15) Oe)
can be obtained at room temperature for the Fe
3
O
4
tube arrays (in
parallel applied field), a value which compares favorably to the
largest reported to date for Fe
3
O
4
zero- and one-dimensional
†
Max Planck Institute.
‡
Leibniz Institute.
Figure 1.
Electron micrographs (scanning, SEM; transmission, TEM) of
the iron oxide tubes. Scale bars: 100 nm. (a) SEM of an array of narrow
tubes (11((4) nm Fe
2
O
3
, green circles) embedded in the alumina template;
contrast enhanced by colorization (original micrograph provided in Figure
S5). (b) TEM of a single thick and short tube (42((4) nm Fe
3
O
4
) isolated
by dissolution of the template; the inset zooms in on the very smooth wall.
(c) SEM of an array of thick ZrO
2
/Fe
2
O
3
/ZrO
2
tubes (12((2)/26((4)/
12((2) nm) embedded in the template: edge view at a crack, with tubes
broken in their length and emerging on the top side of the membrane.
Published on Web 07/14/2007
9554
9
J. AM. CHEM. SOC. 2007,
129
, 9554-9555 10.1021/ja072465w CCC: $37.00 © 2007 American Chemical Society
nanoobjects.
2b,5,13
More importantly, our synthetic approach offers
the unique opportunity to vary the geometry of the tubes at will
and in a tightly controlled fashion. Figure 2 focuses on Fe
3
O
4
tubes
of 50 nm outer diameter and shows the dramatic evolution of their
magnetic properties when the wall thickness d
w
is varied. Soft
ferromagnetism is observed in the limit of very thin tubes, with
vanishing remnant magnetization and coercive field. Increasing the
thickness of the Fe
3
O
4
tubes yields a monotonic improvement of
their magnetic properties up to an optimum situated near d
w
) 13
nm. Further thickness increases are accompanied by receding
remanence and coercivity. While our semiquantitative micromag-
netic (Oommf)
14
simulations of a single tube isolated in vacuum
allow one to expect the initial increase in H
c
//
with d
w
, they do not
reproduce the trend observed for 13 nm < d
w
< 20 nm (Figure
S7). Consequently, we ascribe the latter to the interaction of each
tube with the stray fields produced by the arraysan effective
antiferromagnetic coupling between neighboring tubes, which
reduces H
c
//
as previously demonstrated for the case of nickel
nanowires.
1d,15
Indeed, the stray field produced by the ensembles
of nanotubes at magnetic saturation are significant, on the order of
45 mT for d
w
) 7 nm and 85 mT for d
w
) 15 nm (see Supporting
Information). The experimental magnetic anisotropy of the tube
arrays further corroborates our interpretation, in that applying the
external field H perpendicular to the long axis (z) of the tubes
instead of parallel to it results in large H
c
⊥
values (on the order of
60-70 mT, virtually independent of d
w
, Figure S3). This observa-
tion, which contrasts the situation of isolated tubes (presenting the
easy magnetization axis on z), is a recognized hallmark of strong
interactions between magnetic neighbors.
16
The results of this communication highlight the importance of a
well-controlled chemical synthesis in the context of solid-state
physics.
17
For the preparation of well-defined, tunable, ordered
nanostructures, ALD has proven to be the ideal method. It has
enabled us to observe a strong size dependence in the magnetism
of Fe
3
O
4
nanotubes and identify an optimal thickness. A more
thorough understanding of the phenomenon unveiled here will be
required for engineering future high-density data storage systems.
Our findings warrant detailed analytical investigation of the ALD
reaction (particularly with regard to the C content of the deposited
film), further experimental study on the magnetism of nanotube
arrays and its dependence on geometry, as well as theoretical
modeling of the magnetization reversal in such structures.
Acknowledgment. We thank L. Zhang for experimental as-
sistance. This work was supported by the Bundesministerium fu¨r
Bildung und Forschung: 03N8701 (MPI), 03X5512 (INM). S.M.
thanks the Saarland State and the Leibniz Association for financial
support. J.B. acknowledges the A. von Humboldt Foundation for a
postdoctoral fellowship (3-SCZ/1122413 STP).
Supporting Information Available: Experimental procedures,
hystereses at 5 K and in perpendicular applied field, temperature-
dependent magnetization, original SEM micrograph for Figure 1a, XPS
data, micromagnetic simulations, stray field calculation, saturation of
the ALD reaction, TEM micrograph of tubes in aspect ratio g100. This
material is available free of charge via the Internet at http://pubs.acs.org.
References
(1) (a) Whitney, T. M.; Jiang, J. S.; Searson, P. C.; Chien, C. L. Science
1993, 261, 1316-1319. (b) Prinz, G. A. Science 1998, 282, 1660-1663.
(c) Thurn-Albrecht, T.; Schotter, J.; Kastle, C. A.; Emley, N.; Shibauchi,
T.; Krusin-Elbaum, L.; Guarini, K.; Black, C. T.; Tuominen, M. T.;
Russell, T. P. Science 2000, 290, 2126-2129. (d) Nielsch, K.; Wehrspohn,
R. B.; Barthel, J.; Kirschner, J.; Go¨sele, U.; Fischer, S. F.; Kronmu¨ller,
H. Appl. Phys. Lett. 2001, 79, 1360-1362.
(2) (a) Escrig, J.; Landeros, P.; Altbir, D.; Vogel, E. E.; Vargas, P. J. Magn.
Magn. Mater. 2007, 308, 233-237. (b) Sui, Y. C.; Skomski, R.; Sorge,
K. D.; Sellmyer, D. J. Appl. Phys. Lett. 2004, 84, 1525-2527.
(3) A notable exception: Nielsch, K.; Castan˜o, F. J.; Matthias, S.; Lee, W.;
Ross, C. A. AdV. Eng. Mater. 2005, 5, 217-221.
(4) Sui, Y. C.; Zhou, J.; Li, X. Z.; Skomski, R.; Sellmyer, D. J. J. Appl.
Phys. 2004, 95, 6741-6743.
(5) Yang, J. B.; Xu, H.; You, S. X.; Zhou, X. D.; Wang, C. S.; Yelon, W. B.;
James, W. J. J. Appl. Phys. 2006, 99, 08Q507.
(6) (a) Puurunen, R. L. Appl. Phys. ReV. 2005, 97, 121301. (b) Knez, M.;
Nielsch, K.; Niinisto¨, L.AdV. Mater. 2007, submitted. (c) Daub, M.; Knez,
M.; Go¨sele, U.; Nielsch, K. J. Appl. Phys. 2007, 101, 09J111.
(7) (a) Mathur, S.; Veith, M.; Sivakov.; V.; Shen, H.; Huch, V.; Hartmann,
U.; Gao, H. B. Chem. Vap. Deposition 2002, 8, 277-283. (b) Spandl, J.;
Kusserow, M.; Bruedgam, I. Z. Anorg. Allg. Chem. 2003, 629, 968-974.
(c) Mathur, S.; Sivakov, V.; Shen, H.; Barth, S.; Cavelius, C.; Nilsson,
A.; Kuhn, P. Thin Solid Films 2006, 502,88-93.
(8) (a) Masuda, H.; Fukuda, K. Science 1995, 266, 1466-1468. (b) Nielsch,
K.; Choi, J.; Schwirn, K.; Wehrspohn, R. B.; Go¨sele, U. Nano Lett. 2002,
2, 677-680.
(9) (a) Lim, B. S.; Rahtu, A.; Gordon, R. G. Nat. Mater. 2003, 2, 749-754.
(b) Lie, M.; Fjellvåg, H.; Kjekshus, A. Thin Solid Films 2005, 488,74-
81. (c) Nilsen, O.; Lie, M.; Foss, S.; Fjellvåg, H.; Kjekshus, A. Appl.
Surf. Sci. 2004, 227,40-47. (d) de Ridder, M.; van de Ven, P. C.; van
Welzenis, R. G.; Brongersma, H. H.; Helfensteyn, S.; Creemers, C.; Van
Der Voort, P.; Baltes, M.; Mathieu, M.; Vansant, E. F. J. Phys. Chem. B
2002, 106, 13146-13153.
(10) Knez, M.; Kadri, A.; Wege, C.; Go¨sele, U.; Jeske, H.; Nielsch, K. Nano
Lett. 2006, 6, 1172-1177.
(11) (a) Regazzoni, A. E.; Urrutia, G. A.; Blesa, M. A.; Maroto, A. J. G. J.
Inorg. Nucl. Chem. 1981, 43, 1489-1493. (b) Jiao, F.; Jumas, J. C.;
Womes, M.; Chadwick, A. V.; Harrison, A.; Bruce, P. G. J. Am. Chem.
Soc. 2006, 128, 12905-12909.
(12) Verwey, E. J. W. Nature 1939, 44, 327-328.
(13) (a) Zhang, L.-Y.; Xue, D.-S.; Xu, X.-F.; Gui, A.-B. J. Magn. Magn. Mater.
2005, 294,10-15. (b) Kazakova, O.; Erts, D.; Crowley, T. A.; Kulkarni,
J. S.; Holmes, J. D. J. Magn. Magn. Mater. 2006, 286, 171-176. (c)
Abid, M.; Abid, J.-P.; Jannin, S.; Serrano-Guisan, S.; Palaci, I.; Ansermet,
J.-P. J. Phys.: Condens. Matter 2006, 18, 6085-6093. (d) Latham, A.
H.; Wilson, M. J.; Schiffer, P.; Williams, M. E. J. Am. Chem. Soc. 2006,
128, 12632-12633. (e) Wan, J.; Chen, X.; Wang, Z.; Yang, X.; Qian, Y.
J. Cryst. Growth 2005, 276, 571-576. (f) Goya, G. F.; Berquo´, T. S.;
Fonseca, F. C. J. Appl. Phys. 2003, 94, 3520-3528.
(14) Donahue, M. J.; Porter, D. G. Interagency Report NISTIR 6376: NIST,
Gaithersburg, MD, 1999; http://math.nist.gov/oommf.
(15) (a) Hertel, R. J. Appl. Phys. 2001, 90, 5752-5758. (b) Hwang, M.;
Abraham, M. C.; Savas, T. A.; Smith, H. I.; Ram, R. J.; Ross, C. A. J.
Appl. Phys. 2000, 87, 5108-5110. (c) Jang, H.-J.; Eames, P.; Dahlberg,
E. D.; Farhoud, M.; Ross, C. A. Appl. Phys. Lett. 2005, 86, 023102/1-
023102/3. (d) Sorop, T. G.; Untiedt, C.; Luis, F.; Kroll, M.; Rasa, M.; de
Jongh, L. J. Phys. ReV.B2003, 67, 014402/1-014402/8. (e) Martin, J.
I.; Nogues, J.; Liu, Kai; Vicent, J. L.; Schuller, Ivan K. J. Magn. Magn.
Mater. 2003, 256, 449-501.
(16) (a) Nielsch, K.; Castan˜o, F. J.; Ross, C. A.; Krishnan, R. J. Appl. Phys.
2005, 98, 034318. (b) Sorop, T. G.; Nielsch, K.; Go¨ring, P.; Kroll, M.;
Blau, W.; Wehrspohn, R. B.; Go¨sele, U.; de Jongh, L. J. J. Magn. Magn.
Mater. 2004, 272, 1656-1657. (c) Bao, J. C.; Xu, Z.; Hong, H. M.; Ma,
X.; Lu, Z. H. Scr. Mater. 2004, 50,19-23.
(17) This theme has been emphasized previously: Grohol, D.; Nocera, D. G.
J. Am. Chem. Soc. 2002, 124, 2640-2646.
JA072465W
Figure 2.
Magnetic hystereses of Fe
3
O
4
tube arrays (50((5) nm outer
diameter, 105((10) nm center-to-center distance, various wall thicknesses
d
w
) in porous alumina at 300 K, in a magnetic field H applied along the
tubes (z); a paramagnetic contribution due to the template was subtracted,
and the z moment µ
z
was normalized. Inset: properties (blue, coercivity
H
c
//
; red, remanence relative to 1 T) for 2.6((0.7) e d
w
e 18.2((1.9) nm.
COMMUNICATIONS
J. AM. CHEM. SOC.
9
VOL. 129, NO. 31, 2007 9555
