Oxidation does not (always) kill reactivity of transition metals: Solution-phase conversion of nanoscale transition metal oxides to phosphides and sulfides

Article abstract of DOI:10.1021/ja106397b

Unexpected reactivity on the part of oxide nanoparticles that enables their transformation into phosphides or sulfides by solution-phase reaction with trioctylphosphine (TOP) or sulfur, respectively, at temperatures of ≥370 °C is reported. Impressively, s

Full text of DOI:10.1021/ja106397b

Published on Web 10/21/2010
Oxidation Does Not (Always) Kill Reactivity of Transition Metals:
Solution-Phase Conversion of Nanoscale Transition Metal Oxides to
Phosphides and Sulfides
Elayaraja Muthuswamy and Stephanie L. Brock*
Department of Chemistry, Wayne State UniVersity, Detroit, Michigan 48202, United States
Received July 19, 2010; E-mail: sbrock@chem.wayne.edu
Abstract: Unexpected reactivity on the part of oxide nanoparticles
that enables their transformation into phosphides or sulfides by
solution-phase reaction with trioctylphosphine (TOP) or sulfur,
respectively, at temperatures of e370 °C is reported. Impres-
sively, single-phase phosphide products are produced, in some
cases with controlled anisotropy and narrow polydispersity. The
generality of the approach is demonstrated for Ni, Fe, and Co,
and while manganese oxides are not sufficiently reactive toward
TOP to form phosphides, they do yield MnS upon reaction with
sulfur. The reactivity can be attributed to the small size of the
precursor particles, since attempts to convert bulk oxides or even
particles with sizes approaching 50 nm were unsuccessful.
Overall, the use of oxide nanoparticles, which are easily accessed
via reaction of inexpensive salts with air, in lieu of organometallic
reagents (e.g., metal carbonyls), which may or may not be
transformed into metal nanoparticles, greatly simplifies the
production of nanoscale phosphides and sulfides. The precursor
nanoparticles can easily be produced in large quantities and
stored in the solid state without concern that “oxidation” will limit
their reactivity.
2
Figure 1. (a, c) PXRD patterns of (a) NiO nanoparticles and (c) Ni P
nanorods with comparisons to their reference patterns (powder diffraction
files) from the ICDD database. (b, d) TEM images of (b) 3-5 nm NiO
nanoparticles and (d) 17.9 ( 2.6 nm × 4.0 ( 0.3 nm Ni
2
P nanorods.
Transition metal phosphides exhibit a rich and varied range of
physical properties resulting in their exploitation in diverse ap-
plications such as thermoelectrics and catalysis. The promise of
octylether/oleylamine solvent/surfactant system at 250 °C for 2-3
h under a flow of air. The oxide nanoparticles were isolated by
precipitation with excess ethanol followed by centrifugation. The
isolated nanoparticles were characterized by powder X-ray diffrac-
tion (PXRD) and transmission electron microscopy (TEM) before
they were subjected to reaction with TOP.
1
improved properties has led to an emphasis on the preparation of
2
these materials in nanoscale forms. Reported strategies for the
synthesis of transition metal phosphide nanomaterials include
3
solvothermal reactions, decomposition of organometallic reagents
In the case of Ni, the PXRD pattern was found to be a perfect
match to the reference pattern of NiO (Figure 1a). No additional
peaks were observed, suggesting that few or no crystalline impurities
were present. Moreover, the room-temperature magnetic suscepti-
bility was field-independent, confirming the absence of unreacted
Ni (Figure S1 in the Supporting Information). TEM images revealed
the formation of nanoparticles with roughly spherical morphology
and sizes in the 3-5 nm range (Figure 1b). This is consistent with
the size obtained by Scherrer analysis of the PXRD pattern (4.4
nm), suggesting that the particles were fully crystalline. The free-
flowing dry NiO nanoparticles were then combined with 10 mL of
octylether and 2 mL of oleylamine, degassed under vacuum at 100
°C for 15-20 min, and then heated to 300 °C under argon. TOP
(15 mL) was injected into the system, and then the temperature
was raised to 350 °C. The system was left at 350 °C for 2-3 h,
during which time the color of the solution turned from the
characteristic brown of NiO to black. Isolation of the final product
was carried out by precipitation with excess ethanol followed by
centrifugation.
4
in the presence of phosphines, decomposition of single-source
5
precursors, high-temperature hydrogen reduction of nanoscale
6
7
phosphates, and sonochemical synthesis. Schaak and co-workers
described the use of trioctylphosphine (TOP) as a general precursor
8
for converting transition metal nanoparticles into phosphides, an
approach we have exploited for the rational synthesis of phase-
pure phosphides of Fe and consequential elucidation of their
9
intrinsic magnetic properties. A limitation of this approach is the
8b
reported lack of reactivity of oxides toward TOP, because surface
oxidation is endemic among most transition metal nanoparticles.
Here we show not only that surface oxidation is no impediment to
phosphide formation but also that even completely oxidized particles
can be converted to phosphides using TOP. Moreover, this reaction
is capable of generating phase-pure products in solution at tem-
peratures of e370 °C, and in some cases, the particles are highly
uniform and even adopt anisotropic morphologies. The conditions
that must be met for successful conversion and the extension of
this method to sulfides will be highlighted.
Metal oxide nanoparticles (NiO, Fe
3
O
4
, CoO, Mn
3
O
4
) were
The PXRD pattern of the product of NiO with TOP (Figure 1c)
prepared by decomposition of metal acetylacetonate salts in an
2
suggested complete conversion to Ni P. There was no evidence of
10.1021/ja106397b  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 15849–15851 9 15849

C O M M U N I C A T I O N S
3 4
Figure 2. PXRD patterns of (a) Fe O nanoparticles, (b) CoO nanoparticles,
(
c) FeP nanoparticles, and (d) CoP nanoparticles, all compared with their
reference patterns. * indicates peaks from the sample holder.
residual oxide or any other nickel phosphide phase. Additionally,
relative to the other peaks in the pattern, the (002) reflection was
found to be sharp and the (210) reflection broad, suggesting that
Figure 3. PXRD patterns of (a) MnS nanoparticles, (b) FeS
2
nanoparticles,
(
c) Co nanoparticles, and (d) Ni nanoparticles, all compared to their
9
S
8
9 8
S
reference patterns. * indicates peaks due to minor phases [Ni
Fe S in (c)].
3 2
S in (a) and
4c
the particles have a preferred direction of growth along the c axis.
7
8
The formation of rods was confirmed by TEM analysis (Figure 1d
and Figure S2), and electron-dispersive spectroscopy (EDS) data
were in agreement with the expected Ni/P ratios (Figure S2). The
formation of nanorods indicates that the reaction of NiO with TOP
is not topotactic but probably arises from dissolution of the oxide,
with the anisotropy being exhibited as a consequence of a high
constant concentration of Ni obtained from NiO and the hexagonal
S8b). These observations lead us to conclude that the transformation
of oxide to phosphide is related to solubility or reactivity of the
oxide nanoparticles, each of which is augmented at small particle
sizes.
In principle, transition metal sulfides should be even easier to
prepare from oxides. Indeed, the oxide-to-sulfide nanoparticle
10
symmetry of the crystal structure. This new method of producing
nanorods does not require continuous controlled injection of
transformation has been reported for conversion of Nd
2 3
O nano-
4
b,c,10
b
precursors
or the use of multiple ligands with different
particles into phase-pure NdS using a mixture of boron and sulfur
2
1
1
binding strengths to promote the formation of rods/wires.
Moreover, it is possible to adjust the aspect ratio of the nanorods
by varying the NiO concentration. Thus, increasing concentration
13
powders at 450 °C in vacuum-sealed tubes. Accordingly, we tested
the applicability of our oxide conversion method to the formation
of transition metal sulfides. The PXRD pattern of the product of
NiO nanoparticles with sulfur in the presence of oleylamine and
octylether at 300-350 °C (Figure 3a) indicates the successful
3
-fold led to a corresponding increase in aspect ratio (Figure S3).
Reactivity toward TOP is not limited to NiO. However,
3 4
conversion of Fe O nanoparticles (4-7 nm; Figure 2a and
transformation into Ni
also successfully transformed under similar reaction conditions,
producing phase-pure Co (Figure 3b) and a mixed-phase sample
of FeS (major) and Fe (Figure 3c), respectively. Intriguingly,
9 8 3 2
S (major) and Ni S . Co and Fe oxides were
Figure S4a) and CoO nanoparticles (8-12 nm; Figure 2b and
Figure S5a) using TOP over 4 h at 350 °C resulted in a mixture
of phases (Figure S6). Like Ni, the metal (M)-rich M P phases
2
were favored, but these were accompanied by significant
quantities of the MP phases. In an attempt to prepare phase-
pure products, the heating time was extended to 24 h, and in
the case of Co, the temperature was raised to 370 °C. Phase-
pure MP was obtained in both cases, as indicated by PXRD
9 8
S
2
7 8
S
although MnO was not reactive toward TOP, it also combined with
sulfur to form single-phase MnS (Figure 3d).
These results clearly demonstrate that oxidation does not preclude
nanoparticle conversion to phosphides, provided the samples are
small enough to dissolve (ca. 10 nm) and that the metal oxides are
reactive (Fe, Co, Ni). Moreover, the low rate of reactant introduction
and the high local concentration can enable shape anisotropy to be
accessed and controlled. Oxides are even more reactive toward
sulfur, yielding sulfides even in cases where the corresponding
phosphides cannot be produced (e.g., with Mn). Overall, the use
of oxide nanoparticles, which are easily accessed via reaction of
inexpensive salts with air, in lieu of organometallic reagents (e.g.,
metal carbonyls), which may or may not be transformed into metal
nanoparticles, greatly simplifies the production of nanoscale phos-
phides and sulfides. The precursor nanoparticles can be produced
in large quantities and stored in the solid state without concern
that “oxidation” will limit their reactivity.
(
Figure 2c,d). The formation of the less-metal-rich phase at
longer times and higher temperatures is consistent with our
9
x y
previous work on the Fe-to-Fe P nanoparticle transformation.
TEM analyses of the final products indicated that the FeP
nanoparticles were present as clumps with diameters of up to
5
0 nm (Figure S4), whereas CoP nanoparticles formed as discrete
particles with size comparable to that of the CoO precursor
(
Figure S5).
Intriguingly, oxide nanoparticle conversion does not seem to be
a universal process. Attempts to convert Mn
a phosphide phase by reaction with TOP were unsuccessful. While
TOP reduced Mn nanoparticles to MnO, there was no evidence
3 4
O nanoparticles into
3
O
4
of phosphide formation (Figure S7). The greater barrier for
phosphide generation with Mn has been attributed to its low
1
2
electronegativity relative to later transition metals.
Acknowledgment. This work was supported by the National
Science Foundation (DMR-331769). We thank A. Dixit and G.
Lawes for the magnetic data.
In order to test the effect of size on the conversion of oxides to
phosphides, control reactions were carried out by treating a sample
of bulk NiO with TOP. The PXRD pattern of the product clearly
indicated that the bulk sample remained principally oxide (Figure
S8a). However, when a polydisperse sample of NiO nanoparticles
with sizes of e50 nm was treated with TOP for 24 h at 385 °C,
Supporting Information Available: Experimental procedures for
synthesis and characterization; additional PXRD patterns, TEM images,
and EDS and magnetic data. This material is available free of charge
via the Internet at http://pubs.acs.org.
2
the product was found to be a mixture of NiO and Ni P (Figure
1
5850 J. AM. CHEM. SOC. 9 VOL. 132, NO. 45, 2010

C O M M U N I C A T I O N S
(
(
7) Sweet, J. D.; Casadonte, D. J., Jr. Ultrason. Sonochem. 2001, 8, 97–101.
8) (a) Henkes, A. E.; Vasquez, Y.; Schaak, R. E. J. Am. Chem. Soc. 2007,
129, 1896–1897. (b) Henkes, A. E.; Schaak, R. E. Chem. Mater. 2007, 19,
4234–4242.
References
(
1) Hullinger, F. Struct. Bonding (Berlin) 1968, 4, 83–229.
(
2) (a) Brock, S. L.; Perera, S. C.; Stamm, K. L. Chem.sEur. J. 2004, 10,
3
364–3371. (b) Brock, S. L.; Senevirathne, K. J. Solid State Chem. 2008,
(9) Muthuswamy, E.; Kharel, P. R.; Lawes, G.; Brock, S. L. ACS Nano 2009,
1
81, 1552–1559.
3, 2383–2393.
(
(
3) Luo, F.; Su, H.-L.; Song, W.; Wang, Z.-M.; Yan, Z.-G.; Yan, C.-H. J.
Mater. Chem. 2004, 14, 111–115.
(10) Chen, Y.; She, H.; Luo, X.; Yue, G.-H.; Peng, D.-L. J. Cryst. Growth 2009,
311, 1229–1233.
4) (a) Perera, S. C.; Fodor, P. S.; Tsoi, G. M.; Wenger, L. E.; Brock, S. L.
Chem. Mater. 2003, 15, 4034–4038. (b) Park, J.; Koo, B.; Yoon, K. Y.;
Hwang, Y.; Kang, M.; Park, J.-G.; Hyeon, T. J. Am. Chem. Soc. 2005,
(
(
(
11) Gregg, K. A.; Perera, S. C.; Lawes, G.; Shinozaki, S.; Brock, S. L. Chem.
Mater. 2006, 18, 879–886.
12) Gopalakrishnan, J.; Pandey, S.; Rangan, K. K. Chem. Mater. 1997, 9, 2113–
1
27, 8433–8440. (c) Park, J.; Koo, B.; Hwang, Y.; Bae, C.; An, K.; Park,
2
116.
J.-G.; Park, H.; Hyeon, T. Angew. Chem., Int. Ed. 2004, 43, 2282–2285.
13) Wu, L.-M.; Sharma, R.; Seo, D.-K. Inorg. Chem. 2003, 42, 5798–5800.
(5) Kelly, A. T.; Rusakova, I.; Ould-Ely, T.; Hofmann, C.; L u¨ ttge, A.; Whitmire,
K. H. Nano Lett. 2007, 7, 2920–2925.
(
6) Stamm, K. L.; Garno, J. C.; Liu, G.-y.; Brock, S. L. J. Am. Chem. Soc.
2
003, 125, 4038–4039.
JA106397B
J. AM. CHEM. SOC. 9 VOL. 132, NO. 45, 2010 15851