Surfactant-Mediated Resistance to Surface Oxidation in MnO Nanostructures.

The intrinsic physical properties of nanostructures of metals and their oxides are altered when they are prone to surface oxidation in ambient atmosphere. To overcome this limitation, novel synthesis methodologies are required. In this study, solid octahedral shapes of MnO limit the inward oxygen diffusion compared to that of the MnO-nanoparticle-assembled octahedra. In addition to morphology control, which restricts the thickness of the Mn3O4 surface layer, the binding chemistry of the surfactants plays an essential role. For example, the Mn3O4 surface layer is 0.4 nm thinner with trioctylphosphine oxide than with trioctylamine as the surfactant. The nanostructures were prepared by varying the surfactants, surfactant-to-precursor molar ratio, accelerating agent, and reaction heating rate. The surface oxidation of MnO nano-octahedra was probed by Rietveld analysis of X-ray diffraction patterns and X-ray photoelectron spectroscopy and characterized by magnetic measurements, as the presence of ferrimagnetic Mn3O4 shell on the antiferromagnetic MnO core provides an exchange coupling at the core-shell interface. Thicker the Mn3O4 shell, higher is the exchange-biased hysteresis loop shift.

Introduction

Transition metals and their oxides with different shapes and sizes have been a subject of intense interest among researchers owing to their enormous size and morphology-dependent potential applications in sensors,[1] magnetic resonance imaging,[2] catalysis,[3−5] and high-performance permanent magnets.[6] Because of the high fraction of surface atoms, their nanoparticles (NPs) are especially prone to surface oxidation in ambient atmosphere. The uncontrolled inward diffusion of oxygen atoms results in biphasic core–shell NPs with the intact core, shelled by the oxide phase of a metal with a higher oxidation state.[7,8] This surface oxidation, however, compromises the material properties.[9] The 3d transition metal NPs of Co, Fe, and Ni are the most affected ones, whose ferromagnetic (FM) moments decrease after exposure to air. In fact, these metal NPs are more prone to oxidation when prepared from the thermal decomposition of precursors with oxygen-containing anions, such as acetates, acetyl acetonates, carbonyls, and so on.[10] On the other hand, reductive decomposition of noncarbonyl organometallic complexes, such as [Co(η3-C8H13)(η4-C8H12)] or [Fe{N[Si(CH3)3]2}2], produces NPs with thinner surface-oxide layers.[11] In addition, oxidation from air contact during the workup of the synthesized products can be avoided if washing is performed in N2/Ar atmosphere.[12] Surface oxidation of NPs is also inhibited by creating a shell of carbon or other metal ions.[13,14] This study shows that an appropriate choice of surfactants and octahedron-shaped nanostructures can minimize the surface oxidation.

Recent advances in chemical methods such as hot injection, thermal decomposition, and coprecipitation paved the way to produce NPs with a narrow size distribution.[15,16] Surfactants such as trioctylphosphine oxide (TOPO), trioctylamine (TOAm), oleyl amine (OAm), and oleic acid (OA) play a major role in determining the morphologies of these nanocrystals.[17,18] Their shape and size can be adjusted by changing the precursor-to-surfactant ratio, solvents, temperature, and heating rates.[19] The crystal facets and the energetically favored crystalline directions also change with the synthesis methodologies.[20] Octahedron-shaped metal oxides are particularly interesting because of their active {111} surface atoms and highly energetic edges and corners. Furthermore, these octahedral nanostructures have a range of applications; for example, those of Co3O4 and SnO2 are employed in gas sensing;[21] SnO2 in high-performance Li-ion batteries;[22] and MnO octahedra in cataluminescence sensing,[23] supercapacitors,[24] and photodecomposition.[25]

Herein, we demonstrate the formation of solid nano-octahedra as well as the clustering of MnO NPs into octahedra using different surfactants, precursor-to-surfactant ratios, and reaction rates. MnO is considered as the representative oxide material because manganese is prone to transit from +2 to its higher oxidation states. The surface oxidation of these nanostructures is probed by X-ray photoelectron spectroscopy (XPS) and Rietveld analysis of X-ray diffraction (XRD) patterns and further characterized by magnetic exchange coupling. The solid nano-octahedra are least oxidized compared to that of the NP-clustered octahedral nanostructures. The ferrimagnetic (FiM) Mn3O4 surface layers over the antiferromagnetic (AFM) MnO core increase the magnetic moment and generate exchange anisotropy due to interface spin coupling between the MnO core and the Mn3O4 shell. Exchange coupling refers to the shift of the hysteresis loop along the magnetic-field axis in materials with AFM-FiM/FM interfaces when field-cooled through Néel temperature (TN) of the AFM phase such that the FiM/FM Curie temperature (TC) is greater than TN.[8]

Results and Discussion

Structural and Morphological Analysis

All of the samples crystallize in the face-centered cubic crystal structure of MnO, and the lattice parameter of a = 0.444 nm calculated from the XRD patterns in Figure a is consistent with JCPDS card no. 07-0230. There is no noticeable secondary phase observed in the XRD patterns. The crystallite sizes obtained using the Scherrer equation are 27.6, 27.3, and 28.1 nm for M-1 (prepared with TOPO and benzyl ether), M-2 (prepared with OA and TOAm), and M-3 (prepared with TOAm), respectively. Rietveld refinements of the XRD patterns show that best fits are obtained with a minor Mn3O4 phase in addition to MnO (Figure b–d). The fitted Mn3O4 phases are 3, 5, and 3.7% in M-1, M-2, and M-3, respectively (Table ). Field emission scanning electron microscopy (FESEM) images and the corresponding histograms for M-1, M-2, and M-3 are illustrated in Figure . Figure shows the morphological evolution of the NPs depending on the nature of surfactant, surfactant-to-precursor molar ratio, accelerating agent, and reaction heating rate. Both M-1 and M-3 are octahedron shaped, whereas M-2 has spherical (diameter ∼20 nm) NPs self-assembled into octahedral clusters. Transmission electron microscopy (TEM) images show that both M-1 and M-3 have a side length of ∼80 nm, whereas that of M-2 is ∼160 nm (Figure ). The solid octahedra of the representative M-1 are highly crystalline with well-defined lattice fringes of MnO (111) plane having a d-spacing of 2.5 Å (Figure a), which matches well with the XRD reflections. The fast Fourier transform (FFT) patterns recorded from the edges of M-1 octahedra also show the (111) and (220) reflections of MnO (Figure a inset). A 2D projection of a regular octahedron is shown in Figure b with the corresponding TEM images. The self-assembly of NPs into M-2 octahedra is clearly visible in the inset of Figure c. In spite of the presence of smaller NPs of M-2 compared with the larger solid octahedra of M-1 and M-3, the XRD crystallite sizes are similar, as the 20 nm NPs of M-2 can be considered to be bulklike with sharp XRD peaks.[26] Although the nanostructures appear to be porous, the surface area is extremely small, with magnitudes of 10.5, 3.0, and 4.5 m2/g for M-1, M-2, and M-3, respectively (Figure S1 in the Supporting Information).

Figure 1

(a) XRD patterns of MnO nanostructures under different synthesis conditions. XRD–Rietveld analyzed patterns of (b) M-1, (c) M-2, and (d) M-3, where diff represents the difference plot between observed and calculated patterns, Obs denotes the observed pattern, Calc is the calculated pattern, and Bckgr represents the background plot.

Table 1

XRD–Rietveld Refinement Parametersa

sample [space group]	lattice parameters (Å), angles (deg)	atomic positions (x, y, z)	occupation number	goodness of fit (reduced χ2)	weighted profile (Rwp) (%)
M-1
97% MnO, [Fm3̅m]	a = b = c = 4.4437 ±0.0001 Å, α = β = γ = 90°	Mn1 (0, 0, 0)	Mn1 = 1	1.281	1.75
		O2 (0.5, 0.5, 0.5)	O2 = 1
3% Mn3O4, [I41/amd]	a = b = 5.7880 ± 0.0001 Å, c = 9.393 ± 0.0002 Å, α = β = γ = 90°	Mn1 (0, 0.75, 0.125)	Mn1 = 1
		Mn2 (0, 0.5, 0.5)	Mn2 = 1
		O3 (0, 0.0321, 0.2598)	O3 = 1.0
M-2
95% MnO, [Fm3̅m]	a = b = c = 4.4357 ± 0.0002 Å, α = β = γ = 90°	Mn1 (0, 0, 0)	Mn1 = 1	1.020	3.5
		O2 (0.5, 0.5, 0.5)	O2 = 1
5% Mn3O4, [I41/amd]	a = b = 5.7880 ± 0.0001 Å, c = 9.393 ± 0.0001 Å, α = β = γ = 90°	Mn1 (0, 0.75, 0.125)	Mn1 = 1
		Mn2 (0, 0.5, 0.5)	Mn2 = 1
		O3 (0, 0.0321, 0.2598)	O3 = 1.0
M-3
96.3% MnO, [Fm3̅m]	a = b = c = 4.4499 ± 0.0001 Å, α = β = γ = 90°	Mn1 (0, 0, 0)	Mn1 = 1	1.058	3.84
		O2 (0.5, 0.5, 0.5)	O2 = 1
3.7% Mn3O4,[I41/amd]	a = b = 5.7880 ± 0.0002 Å, c = 9.393 ± 0.0002 Å, α = β = γ = 90°	Mn1 (0, 0.75, 0.125)	Mn1 = 1
		Mn2 (0, 0.5, 0.5)	Mn2 = 1
		O3 (0, 0.0321, 0.2598)	O3 = 1.0

Here, , , , , N = number of data points, Iobs = observed intensity, Ical = calculated intensity, and P = number of parameters.

Figure 2

(a) FESEM image of M-1 and (b) the corresponding histogram. (c) FESEM image of M-2 and the corresponding histograms for (d) NP diameters and (e) the side length of the assembled octahedra. (f) FESEM image of M-3 and (g) the corresponding histogram.

Figure 3

Schematic illustration of MnO nanostructures under different synthesis conditions (Mn(acac)2: Mn-acetylacetonate; P: precursor; HR: heating rate).

Figure 4

(a) TEM image of M-1. (Upper inset) HRTEM image of the selected portion of an octahedron and (lower inset) FFT pattern obtained from the same octahedron. (b) Different 2D TEM projections for M-1 octahedra. (c) TEM image of M-2. (Insets) Different projections of the nano-octahedra. (d) TEM image of M-3. (Inset) Enlarged view of one nano-octahedron.

Formation Mechanism

By varying the surfactants, such as TOPO, TOAm, OAm, and OA, and their combination with 1,2-dodecanediol (DDD) as a cosurfactant, the morphology of the nanostructures could be controlled in the single-pot reaction. All of the reactions were performed at 290 °C because the boiling points of all of these surfactants are above 300 °C. Among the surfactants, TOPO and TOAm contain 8C in a single chain, whereas OAm and OA have 18C with a double bond. Surfactants with 8C, such as octyl amine or octanoic acid, could not be used for comparing the morphologies because they have boiling points below 250 °C. When only TOPO is used, the molar ratio of TOPO to Mn(acac)2 precursor of 2.5:1 is found to be sufficient to obtain the solid octahedral shapes (Figure a). When TOPO is reduced to 0.5 from 2.5, NP clusters are obtained (Figure S2), which highlights the requirement of an optimum concentration of the surfactant to create octahedral shapes. The nature of the surfactant also plays a critical role in maintaining the 2.5:1 ratio with Mn(acac)2, and TOAm only results in spherical NPs. To produce MnO octahedra, a higher amount of TOAm is needed such that the TOAm/Mn(acac)2 molar ratio is ≥7. When a long-chain fatty acid, such as OA, is added to the branched TOAm such that the molar ratio of TOAm/OA is 7:2, an NP-assembled octahedron is obtained as in M-2 (Figure c). In the absence of branched surfactants, such as TOPO or TOAm, clusters of spherical NPs are obtained (not shown). This clearly substantiates the importance of branched surfactants in creating octahedral shapes of metal oxides. Our results are in direct contrast to those of Fontaíña-Troitiño et al., who showed that OA plays a key role in stabilizing the octahedral shape, as OA selectively binds the {111} facets.[27] However, Lu et al. showed that OA selectively binds the {100} facets and produces cube-shaped particles.[28] The different functional groups on TOPO, TOAm, OA, OAm, and DDD have different binding energies at the NP surface, to control their morphology.[29,30]

The surfactants or their mixtures with different binding affinities to the NP surface and selectivity toward certain crystal facets provide control over the size and morphology of the nanostructures. The growth of the nano-octahedra is largely due to the minimization of the facets with high surface energy. Branched-chain hydrocarbon surfactants such as TOPO and TOAm promote limiting surface tension and stabilize the low-surface-energy facets of NPs in contrast to linear-chain surfactants, such as OA.[31] The larger nano-octahedra of M-1 and M-3 could be the result of nucleation followed by aggregation when the surfactant molecules peel off from the low-energy facets of the NP surface.[30] To minimize the surface energy, the majorly unprotected high-energy facets will facilitate this aggregation even further. When OA is added alongside TOAm, the high-energy facets are also protected and thereby 20 nm NPs are stabilized in M-2. With a relatively higher concentration of TOAm, the NPs self-assemble in octahedral shapes.

Accelerating agents such as DDD are crucial in preparing faceted nanostructures.[32] In all of the reactions of M-1 to M-3, the molar ratio of DDD to Mn(acac)2 is maintained at 2.5:1 and in the absence of DDD spherical NPs are obtained (Figure S3). Moreover, to obtain octahedron-shaped structures of M-1, the reaction mixture was heat-treated from room temperature to 250 °C at 4 °C/min and further up to 290 °C at 8 °C/min. To identify the effect of reaction rate on the final morphology, when the reaction is ramped up to 290 °C at 10 °C/min, spherical NP clusters are produced (Figure S4). Low reaction rates benefit the controlled nucleation of the crystal lattice, subsequent fast random attachment of the nuclei, and intraparticle ripening, that is, the diffusion of reactants along the surface of the nucleus to alter the shape temporally.[33] Low reaction rates will, however, increase the size of the octahedra, which is energetically more favored.[6] In the formation of octahedral shapes, surface energies of the different facets create an instability, which is minimized by the slow dissolution of the high-energy facets via intraparticle diffusion. This leads to the growth of low-energy facets with an increase in particle size, in this case up to ∼80 nm side length of the M-1 and M-3 octahedra.

Surface Oxidation

The samples were synthesized in N2 atmosphere. Washing, drying, and storage were performed under ambient atmosphere for similar durations, to compare the surface oxidation of the nanostructures. The characterization of the samples was performed simultaneously or in close time intervals. Moreover, XRD data were collected at different intervals (1, 10, and 30 days) and the patterns were observed to be similar according to JCPDS card no. 07-0230, without any noticeable Mn3O4 reflection. The samples were also allowed to withstand additional 30 days before performing the XPS and magnetic measurements. To investigate the effect of solid ∼80 nm MnO octahedra (M-1) compared to that of the NP-assembled octahedra (M-2) on surface oxidation, the XPS spectra were analyzed with best fit to different oxidation states of manganese (Figure a,b). In M-2, each of the self-assembled NPs has their surface accessible for air oxidation in contrast to only the outer surface of the solid M-1 octahedra. Second, the Mn2+ ions of MnO can be oxidized to the immediately higher oxidation state of Mn3+ in the form of cubic α/γ-Mn2O3 or the normal spinel structure of Mn3O4. Taking a cue from the Rietveld refinements of the XRD patterns, the XPS peaks are best fitted from the Mn 2p3/2 level with the binding energies of 640.4–640.7 eV for Mn2+ in MnO/Mn3O4 (1) and 642.3–642.8 eV for Mn3+ in Mn3O4 (2).[34,35] As XPS is a surface-sensitive technique, the information is obtained only up to a few atomic layers below the surface and therefore the formation of the Mn3O4 phase at the MnO surface is well substantiated. The extra peak at 644.8 eV is due to the distinctive satellite peak of Mn (3). Limited oxygen diffusion through the solid octahedral nanostructures of M-1 results in a thinner Mn3O4 surface layer compared to that of the self-assembled NPs of M-2, where oxygen diffusion is random. This has been verified from the Mn3+/Mn2+ ratio of 0.6 in M-1 compared to 1.5 in M-2.

Figure 5

XPS spectra of (a) M-1 and (b) M-2.

The resistance to surface oxidation of M-1 octahedra is further verified from the loop shift in the negative direction of the field axis due to the exchange coupling at the AFM/FiM interface (Figure a,b).[8,36,37] At the MnO/Mn3O4 interface, a torque is created on the Mn3O4 spins (Figure b inset) and an exchange coupling takes place when the samples were cooled in the presence of a 2 T magnetic field from above TC of Mn3O4 to below TN of MnO. Thicker surface-oxidized shells will therefore result in larger loop shifts. The exchange bias equivalents to the loop shifts are observed to be 1358 and 2648 Oe for M-1 and M-2, respectively (Figure a,b). The nearly double loop shift in M-2 clearly verifies a thicker well-defined oxidized Mn3O4 shell in each NP. Furthermore, because there is a collection of NPs in M-2, multiple MnO/Mn3O4 interfaces increase the exchange coupling rate. Because of the surface Mn3O4 layer, M-1 and M-2 demonstrate atypical moments of 1.15 and 1.23μB/Mn, respectively. In fact, the small moment for M-1 is consistent with other reports on MnO,[26] which suggests that the M-1 octahedra allow limited surface oxidation and hence the thin Mn3O4 shell has a minimal effect on the total magnetization. The thickness of the Mn3O4 shell is calculated from the equation of the random-field model of exchange anisotropy: HE = 2(AAFMKAFM)1/2/(MFiMtFiM), where, HE is the magnitude of the hysteresis loop shift (exchange anisotropy field); KAFM = 9 × 103 erg/cm3 and AAFM are the uniaxial anisotropy energy and the exchange stiffness of the MnO core, respectively; and MFiM and tFiM are the magnetization and thickness of the Mn3O4 surface layer, respectively.[38]AAFM is estimated to be 5 × 10–7 erg/cm, and different metal oxides show almost similar AAFM values. With HE = 1358 Oe and MFiM = 9.1 emu/g, the thickness (tFiM) of the Mn3O4 layer of M-1 is found to be 2 nm, and with HE = 2684 Oe and MFiM = 9.7 emu/g, M-2 has a Mn3O4 layer thickness of 3.4 nm. The higher layer thickness of 1.4 nm of M-2 than M-1 reflects the impact of the oxidized surface layer on the magnetic property of M-2.

Figure 6

(a) M–H curves of M-1, M-2, and M-3 at 5 K recorded in a field range of ±4 T after 2 T field cooling. (b) M–H curves in a low-field range showing the corresponding loop shifts. Inset shows the image of the exchange coupling at the AFM/FiM interface and the pinning of the FiM spins by AFM spins when the field (H) is reversed.

Role of Binding Chemistry

Besides nanostructure morphology, surface chemistry is found to play a major role in determining the extent of surface oxidation. Although M-1 and M-3 have nearly identical shapes, their capping agents differ, namely, TOPO for M-1 and TOAm for M-3. The Mn3+/Mn2+ ratio for M-3 is observed to be 0.8 (Figure S5) as opposed to 0.6 for M-1. The thickness of the Mn3O4 surface layer in M-3 is calculated to be 2.4 nm from the exchange bias loop shift of 1665 Oe and the magnetic moment of 0.79μB/Mn (Figure a,b). With similar carbonic tails and flexible conformers, both TOPO and TOAm will have indistinguishable distribution over the octahedron surface. Hereby, we hypothesize that oxygen affinity when TOPO is bound through the oxygen of the O=P moiety on the M-1 octahedron surface is less than that when TOAm is bound through N-groups in M-3. Therefore, chances of aerial oxidation are much higher in M-3. Hence, surface oxidation can be also controlled depending on the binding chemistry between the surfactant and the surface of the nanostructures.

Conclusions

In summary, both morphology- and surfactant-dependent control of surface oxidation was demonstrated, taking MnO as the representative system and Mn3O4 as its surface-oxidized phase. The octahedral nanostructures were prepared by varying the nature and amount of surfactant, accelerating agent, and reaction rates. Oxygen diffusion was restricted in the solid octahedra than in the self-assembled NP octahedra. The solid MnO octahedra (M-1) prepared using TOPO and benzyl ether had a 2 nm thick Mn3O4 surface layer, whereas the solid octahedra (M-3) synthesized using TOAm and the NP-assembled octahedra (M-2) prepared with OA and TOAm had Mn3O4 shells of thicknesses 3.4 and 2.4 nm, respectively. XPS analysis revealed the Mn3+/Mn2+ ratios of M-1, M-2, and M-3 to be 0.6, 1.5, and 0.8, respectively. The hysteresis loop shifts arising from the exchange coupling at the MnO/Mn3O4 interface provided additional evidence of surface oxidation control. The loop shifts were 1358, 1665, and 2648 Oe for M-1, M-3, and M-2, respectively, in accordance with the extent of surface oxidation. Our strategy of surfactant-mediated morphology control to prevent surface oxidation can be extended to other metals/oxides.

Experimental Section

Materials

Manganese(II) acetylacetonate (Mn(acac)2, 97%, Aldrich), DDD (90%, Aldrich), TOPO (90%, Aldrich), TOAm (98%, Aldrich), OAm (70%, technical grade, Sigma-Aldrich), OA (90%, technical grade, Sigma-Aldrich), and benzyl ether (98%, Aldrich) were used without further purification.

Materials Synthesis

For the synthesis of all samples, 1.707 mmol (0.4322 g) Mn(acac)2, 4.268 mmol (0.8635 g) DDD, and different combinations of surfactants were mixed in a 25 mL round-bottom flask. The mixture was degassed in N2 atmosphere for 30 min and heated at 250 °C for 5 min in N2 at 4 °C/min. Thereafter, the reaction was heated up to 290 °C at 8 °C/min and maintained at 290 °C for 2 h. Finally, the reaction flask was cooled to room temperature and the product was washed with ethanol and dried at 50 °C overnight in a hot-air oven. The samples differ in the type and ratio of surfactants used, as follows:

M-1: 4.268 mmol (1.65 g) TOPO and 11 mL of benzyl ether,

M-2: 3.413 mmol (1.077 mL) OA and 11.96 mmol (5.23 ml) TOAm,

M-3: 15.363 mmol (6.7175 mL) TOAm.

Methods

XRD measurements were carried out with a Rigaku (MiniFlex II, Japan) powder X-ray diffractometer with Cu Kα radiation of wavelength 1.54059 Å. Rietveld refinements were performed by General Structure Analysis System (GSAS) software (Los Alamos National Laboratory Report, 2004). The GSAS was run under least-squares refinement condition. FESEM images were recorded in a Carl Zeiss SUPRA 55VP field emission scanning electron microscope. TEM images were obtained by UHR-FEG-TEM, JEOL, JEM 2100 F model using a 200 kV electron source. The surface-area measurements were carried out with a Micromeritics Gemini VII surface area analyzer. XPS studies were carried out on the samples mounted on copper stubs with silver paste using Al Kα radiation (1486.6 eV) in a commercial photoelectron spectrometer obtained from VSW Scientific Instruments. The base pressure of the chamber was maintained around 5 × 10–10 mbar. Before XPS measurements, the samples were kept in vacuum to avoid moisture adsorption. The XPS analysis was performed in an Omicron XPS system working at a base pressure of −5 × 10–10 Torr. The Al Kα X-ray source was used for excitation, and the kinetic energy of the photoelectrons was measured by a 250 mm multichannel hemispherical analyzer. Survey scans were acquired with a 100 eV pass energy (1 eV resolution), whereas the core levels were acquired at a 25 eV pass energy with a 0.1 eV resolution at 90% of peak height. The acquired data were background-corrected by the Shirley method, and the peaks were fitted using the Fityk software, with Voigt peaks having 80% Gaussian and 20% Lorentzian components to find the valence states of the elements. The magnetic properties were studied using the Cryogenics, Physical Property Measurements System, with VSM probe, in the temperature range of 5–300 K and applied fields of up to 4 T.