Abstract
Mn2+ ion doping is used as an electron paramagnetic resonance (EPR) probe to investigate the influence of low-coordination structural defects such as step edges at the surface of terraced (001) MgO nanoparticles on the electronic properties. Beside the well-known hyperfine sextet of Mn2+ ions in the cubic crystal field of MgO, an additional EPR feature with a striking nonmonotonous temperature dependent shift of the -factor is observed in terraced nanoparticles in the temperature range from 4 K to room temperature. By linking the difference in the temperature dependence of the Mn2+ sextet intensity in cubic and terraced nanoparticles with the possible s-d exchange shift and enhanced Zeeman splitting we conclude that the novel EPR feature originates from the loosely trapped charge-compensating carriers at the abundant structural defects at the surface of terraced nanoparticles due to their exchange interaction with neighboring Mn2+ ions.
1. Introduction
Iconic wide-gap oxide MgO has been used for many applications connected with the spin states of electrons, such as spintronics and superconductivity, mostly for buffer tunneling layers in nanoscale. On the other hand, magnetic Mn2+ ions, which play a decisive role in spintronics of diluted magnetic semiconductors (DMS), can be easily doped into MgO and its nanostructures. Pure MgO forms an ionic crystal with cubic NaCl structure consisting of Mg2+ and . In contrast, the chemical bond in a free MgO molecule is much more covalent than in the bulk of the crystal and the effective charges on the ions are Mg1+ and O1−. Thus, in the bulk of MgO crystal the electron affinity is practically eliminated by the Madelung potential. Meanwhile, (001) MgO surface can have structural defects such as step edges, kinks, and corners that can possess various electronic properties interesting for applications such as sensors and catalysts. Having the bandgap of 7.7 eV, completely filled valence bands, and empty conduction bands, MgO would transfer electrons/holes only along the low-coordination structural surface defects, mostly step edges. It could happen because the Madelung potential decreases rapidly as the Mg and O coordination decreases, and for the Mg ion at the edge, corner, and other low coordinated (LC) sites the electron affinity becomes comparable with the electron affinity for a free ion. The calculated relaxed electron affinities of 3-coordinated Mg sites of 1-2 eV demonstrate that they can serve as electron traps [1, 2]. A trapped electron is almost entirely localized at the terminating Mg ion, which thus becomes similar to Mg1+. The same is true for trapped holes and for O1− ions. One would expect that foreign ions like Mn2+ if close enough to the LC sites would also need charge compensation like Mn2+ + e− to produce an electron-Mn2+ complex.
For MgO nanoparticles the surface structural defects depend on the method of synthesis [3] that may also produce bulk defects. Recently we reported that terraced MgO nanoparticles are distinguished by peculiar bulk electronic defect states leading to luminescence with a double-band light emission at 260 nm and 490 nm [3]. A further attempt to investigate electronic states in terraced MgO nanoparticles with the electron paramagnetic resonance (EPR) technique at the X-band by comparing them with cubic MgO nanoparticles led us to unexpected results. In fact, as far as the Inductively Coupled Plasma analysis indicated that all the nanoparticles always contain unintentionally doped Mn2+ impurities at several ppm level [3], a pronounced sextet signal from those impurities was only expected to be identical for cubic and terraced nanoparticles.
Using Mn2+ as an EPR probe has been well known [4–6]. On the other hand, numerous papers report optical detection that demonstrates how a single Mn2+ ion influences the DMS exciton through the sp-d exchange field in a carrier-Mn complex in semiconductor single quantum dots [7–9]. The shift in the band edge energy leads to a great -factor enhancement [10–12]. If LC defects trapped a charge carrier (as far as such a charge transfer was predicted for Mg and O surface ions) [1, 2] then an Mn2+ ion would also influence the nearby electron through the s-d exchange field.
Indeed, for both cubic and terraced nanoparticles we found the well-known hyperfine sextet of Mn2+ ions, usually for (001) MgO, yet for terraced ones we observed an additional broad feature with -factor of 2.23 at room temperature (RT). Curiously, this feature moves downfield considerably when the temperature lowers to 4 K. Such lability makes it difficult to assign this feature to an isolated paramagnetic defect. The -factor has nonmonotonous temperature dependence that may be interpreted as a shift from the electron free value that may be caused by the exchange field of nearby Mn2+ ions. Meanwhile, a low-field feature has been known in GaAs: Mn2+ and was explained as the signal from the complex Mn2+ (d5) plus the weakly coupled delocalized hole [13]. Note that an excited paramagnetic system in a triplet state with the exchange already showed a striking and unexplained nonmonotonous shift (similar to what we observed) proportional to a quantity differing only slightly from the susceptibility [14].
Here we report the results of our EPR measurements on terraced and cubic MgO nanoparticles and correlate the temperature behavior of the low-field EPR feature with the temperature behavior of the EPR signal from Mn2+ ions. By this correlation we conclude that the new EPR feature of terraced MgO nanoparticles may originate from trapped electrons/holes that were predicted to exist along the edges and ledges of the terraces being LC structural defects [1, 2]. The electrons may exchange/interact with nearby Mn2+ ions which provide an effective field that acts on the loosely trapped electrons as the EPR shows. Thus the puzzling EPR feature has been qualitatively explained and its existence can be considered as supporting the presence of loosely bound charge carriers along the edges of terraces that may find an application in nanoelectronics, based on the wide-gap oxide.
2. Experimental
MgO nanoparticles were produced in an oxyhydrogen flame and were distinguished by intensive terraced structure (say, of type I, Figure 1(b)) as described earlier in [3]. Type II nanoparticles (Figure 1(a)) were produced by burning Mg in dry air and were distinguished by exclusively cubic shape of nanoparticles (well-known MgO smoke). The powder samples were measured in a standard Bruker EPR spectrometer setup equipped with a helium flow cooling system with temperature stabilization. The measurements were performed in a sweeping mode in X-band of about 9.4 GHz.
(a)
(b)
3. Results and Discussion
Both types of nanoparticles contained similar amount of Mn2+ impurities of several tens ppm revealed by EPR as the characteristic sextet of hyperfine splitting (HFS) of Mn2+ ions in the cubic crystal field (Figure 2) with -factors (of 1.88; 1.93; 1.98; 2.03; 2.09; 2.14). The sextet is known to have been widely used as a marker along with the famous DPPH for calibration of -factors in EPR experiments. However, only for terraced samples of type I there appeared an additional broad EPR signal (temperature dependent linewidths of ca. 100–200 Oe) at low field with at RT that was not observed for type II samples. Moreover, the additional low-field EPR feature demonstrated a complex behavior with the temperature (Figure 3). The enhanced -factor extracted from graphs in Figure 3 is shown in Figure 4 and is pronouncedly temperature dependent when -factor changed up to at 4 K (Figure 4). In parallel, one can observe a non-Curie temperature dependence of the HFS sextet intensity for both types of the samples (Figure 2); however, the specific temperature dependence was quite different for type I and type II samples as presented in Figures 2(b) and 2(d), correspondingly.
(a)
(b)
(c)
(d)
(a)
(b)
(c)
(d)
A non-Curie behavior seems quite natural for the Mn2+ ion sextet because of the interplay between the thermal and radiofrequency energy population of the energy levels [15]. However, the difference between Figures 2(b) and 2(d) suggests that some Mn2+ ions in type I samples participate in the resonance differently just because they may lie in the vicinity of the LC structural defects that may modify the charge state of the surrounding oxygen ion. Therefore, as we mentioned before, there is a need in charge compensation, so similar to Mg1+ ion on the surface an Mn1+ ion could be created. However, if all ions were such, the Mn1+ ions (being a d6 non-Kramers ion) would be nearly EPR silent. Thus instead one may rather expect a number of complexes electron -Mn2+, such that the electron may be donated to a nearby LC defect and be loosely bound. Analogous charge compensation by a hole was discussed in [8]. It is intriguing to notice that in GaAs: Mn the bound hole produced a broad low-field EPR feature at along with the Mn2+ signal [13]. Additionally, a flip-flop in two resonant forms Fe3+–O2−↔ Fe2+–O1− (it could be realized in the Mn2+–O1− state jumping into the Mn3+–O2− state and back) was reported in [16] as the main reason for EPR feature (the radiation dose dependent) at in feldspars.
We suggest a simple semiquantitative analysis of our experimental findings based on the analogy with DMS. A loosely trapped electron brought to an LC defect is in the field of a nearby Mn2+ ion which magnetic moment relaxes relatively slow. The local field induced by the exchange between the electron and the ion influences the carrier [12]. The local field from the magnetic ion makes magnetization leading to a shift in a resonance field of the nearby loosely trapped electron, thus producing a temperature dependent -factor. It is similar to what happens when an Mn2+ ion acts on an exciton in a quantum dot with a so-called giant Zeeman effect [10–12].
We can write down the equation for the effective -factor of the electron that explicitly expresses it through magnetization and the susceptibility in analogy with [12]:where is the loosely trapped electron -factor (this is quite different from the DMS where instead of in (1) there is , the negative -factor of the band electron) and is the Mn2+-factor, is the exchange integral, positive for ferromagnetic interaction between the electron and Mn2+ ion, is the Bohr magneton, and is the static magnetic susceptibility.
Instead of , in (1) we deal with the EPR susceptibility which is proportional to the EPR response of the Mn2+ ions residing at the LC defects and which can be obtained by subtracting the appropriately rescaled EPR susceptibility of cubic samples (calculated in the usual way from the linewidth and the intensity of the signal in arbitrary units) taken from Figure 2(b), from the EPR susceptibility of terraced samples (taken from Figure 2(d)): where is a constant. Finally, the -factor is the sum of the free electron value and the shift: where is another constant. Both and can be obtained from comparing with the experimental dependence in Figure 4. The line with the fitting parameters and is rather close to the experimental data. Finally, it also qualitatively reproduces the otherwise puzzling nonmonotonous behavior of .
4. Conclusion
In conclusion, we presented unusual EPR data of a nanostructured material, based on MgO terraced nanoparticles, that demonstrated a possibility of manipulating the electronic properties of the wide-gap oxide by magnetic Mn2+ ions similar to the well-established manipulation with DMS quantum dots. An encouraging example of such manipulation is room temperature magnetism found in ZnO:Mn2+ [17].
Competing Interests
The authors declare that they have no competing interests.
Acknowledgments
This work was done by financial support from the Global Frontier Center for Multiscale Energy Systems (2011-0031561) supported by the Korean Ministry of Science and Technology. The authors thank Korea University and Sunhee Kim from KBSI for the EPR measurements.