The chemical compositions of AlN nanowires with different Mn concentrations were measured by EDS. The at.% of Mn (x) is found to be 0.0, 0.4, 0.9, 1.3 and 2.3 in the samples, respectively. The XRD pattens of AlN nanowires doped with diverse concentrations of Mn ions (AlN: xMn) are presented in Fig. 1 (a). For all samples, the diffraction peaks in patterns can be indexed as wurtzite-structured AlN (PDF card No. 08-0262). Under the resolution of XRD, no peak of impurity phase is detected. As show in Fig. 1 (b), with the increase of doping concentration, the (1 0 0), (0 0 2), and (1 0 1) peak positions shift towards lower angles, indicating lattice constant broadening due to the substitution of Al3+ (0.054 nm) by larger Mn2+ (0.067 nm).
Raman scattering is a powerful tool for detecting dopant incorporation, lattice defects and disorders. The Raman spectra of the as-prepared AlN:Mn nanowires are shown in Fig. 2. For AlN: 0.0%Mn nanowires, four Raman scattering peaks locating at 251, 612, 659 and 671 cm− 1 are assigned to E2(low), A1(transverse optical (TO)), E2(high) and E1(TO) modes of the wurtzite AlN, respectively. A low-intensity and wide peak around 900 cm− 1 is caused by the A1(longitudinal optical (LO)) and E1(LO) modes. With increasing of the Mn concentration in nanowires, the A1 (TO), E2 (high) and E1(TO) peaks gradually broaden and move to lower frequencies due to increasing of defect concentration and disorders resulting from incorporation of Mn2+ ions. The results can further confirm that the Al ions are substituted by Mn ions in AlN nanowires.
The XPS measurement is used to determine the valence state of Mn in the AlN nanowires. The XPS spectrum of AlN: 0.9%Mn nanowires are performed in Fig. 3. The survey scan XPS spectrum in Fig. 3 (a) shows strong peaks for aluminum and nitrogen as expected, except for the C 1s peak from the reference and O 1s peak from absorbed gaseous. As shown in Fig. 3 (b) and (c), the Al 2p peak centered at 73.4 eV and N 1s peak centered at 396.8 eV are related to binding energy of aluminum and nitrogen in AlN, respectively. The Mn 2p XPS spectrum (Fig. 3(d)) displays two major peaks with binding energy values at 642.2 and 654.5 eV, corresponding to the Mn 2p3/2 and Mn 2p1/2 peaks, respectively. Between them, a peak at 647.6 eV corresponding to the shake-up satellite suggests the presence of Mn2+. The results of XRD, EDS, Raman and XPS confirm the successful incorporation of Mn2+ ions into AlN nanowires.
The morphology of the as-synthesized AlN:0.9%Mn nanowires was performed using SEM, as shown in Fig. 4 (a), (b) and (c), revealing that the product consists of high-density nanowires with very high aspect ratio. The diameter of the nanowires ranges from 30 to 200 nm and the length extends to several micrometers. The chemical composition of nanowires is checked by the EDS analysis, as shown in Fig. 4(d). The composition analysis is clear that the contents of Mn in AlN nanowires are about 0.9%.
A typical TEM image of an AlN:0.9%Mn nanowire is given in Fig. 5(a), which clearly shows that the diameter of the nanowire is about 50 nm, which is the same as that observed by SEM. Figure 5(b) shows the HRTEM image recorded from the edge of the nanowire. The adjacent lattice spacing is about 0.25 nm, which is consistent with the (001) facet distance of hexagonal AlN, indicating that the nanowire grows along [001] direction. In addition to this, many dark spots can be seen in the lattice, which are attributed to local lattice distortions and defects caused by the larger Mn2+ ions replacing Al3+ in the nanowires.
In recent years, the potential applications of AlN-based DMS in spintronics have attracted great interest. To explore the magnetism of AlN:Mn nanowires, the magnetic measurements of AlN: Mn nanowires with different Mn concentrations are carried out at room temperature using VSM. Figure 6 shows the corresponding magnetic hysteresis (M-H) curves for 0.0, 0.4, 0.9, 1.3 and 2.3 at% Mn doped AlN nanowires. All AlN:Mn samples exhibit well-defined hysteresis loops and ferromagnetic behaviors. The saturation magnetization (Ms) and coercivity values (Hc) of AlN:xMn nanowires are shown in Table 1. According to the results of XRD, Raman, XPS and HRTEM, it can be confirmed that the Mn2+ ions are successfully doped into AlN nanowires without changing wurtzite structure, and excluded the possibility of other impurities. Thus, the observed room temperature ferromagnetism is the intrinsic properties of AlN:Mn nanowires.
Table 1
saturation magnetization (Ms) and coercivity values (Hc) of AlN:xMn nanowires.
Mn ratio (at. %)
|
Ms (emu/g)
|
Hc (Oe)
|
0.0
|
0.015
|
35.4
|
0.4
|
0.032
|
40.0
|
0.9
|
0.045
|
48.9
|
1.3
|
0.058
|
77.0
|
2.3
|
0.050
|
83.4
|
Up to now, the ferromagnetic origin of DMSs is still confusing. There are two most prevalent mechanisms of ferromagnetism in DMS: the carrier-mediated model and the bound magnetic polaron (BMP) model [28, 29]. Because AlN has a high resistance at room temperature, the carrier-mediated model is not suitable for explaining the ferromagnetic properties of the AlN:Mn nanowires and the BMP model is the preferred mechanism for AlN: Mn nanowires [30]. Based on BMP model, the magnetism for semiconductors is very sensitive to the defects such as dopants, anion or cation vacancies and interstitials. Mn has the 3d54s2 valence configuration. According to Hund’s rule, there are five unpaired 3d electrons with same spin direction, which may contribute to magnetic ordering. However, the possibility of ferromagnetic coupling is very small when the concentration of Mn2+ is low in the doped AlN nanowires. In previous literatures, the doping of large functional ions in AlN can easily lead to high Al or N vacancies, because of large size mismatch [31, 32]. Also, the presence of high Al or N vacancies in the AlN:Mn nanowires are revealed by HRTEM, which trap electrons and mediate the magnetic spin of Mn2+, leads to the delocalized magnetic moments and long-range coupling of the BMPs. Therefore, the ferromagnetism of AlN:Mn nanowires may be due to BMPs produced by the interactions between the trapped electrons in Al or N vacancies and the magnetic spins of Mn2+ ions.
On increase Mn dopant ion concentration, the Hc increase gradually, which is related to the increase of Al or N vacancies in the nanowires. The increase in Hc values with increase doping concentration is similar to previous report for Sc doped AlN nanowires [33]. Meanwhile, it is noticed that the Ms is enhanced with the increase of Mn content (approximately up to 1.3%), and then decreases. This decrease in Ms is due to the fact that higher dopant ion concentrations result in shorter distances between the doped atoms, thus strengthening the antiferromagnetic interactions at the expense of ferromagnetic ordering [29, 34, 35]. Therefore, the Mn contents in AlN nanowires may exist a critical value. Below this value, Ms increases with increasing Mn concentration, and above this value, Ms decreases. In brief, the magnetic properties of AlN: Mn nanowires can be tuned by controlling the doping concentration of Mn.
Figure 7 (a) shows the PL and PL excitation (PLE) spectra of the AlN:0.9% nanowires. Under excitation at 266 nm, the AlN:0.9%Mn nanowires show an intense red emission band with a maximum at 597 nm, which corresponds to intra-atomic transition of Mn2+ from the excited state 4T1(4G) to the ground state 6A1(6S). As indicated in the inset of Fig. 7 (a), this bright red emission can be clearly seen by the naked eye under UV lamp (254 nm) irradiation. The full width at half-maximum of the emission band is about 39 nm, indicating luminescence of AlN:Mn nanowires have a high color purity. The PLE spectrum, monitored with 597 nm, shows a wide excitation peak with a maximum at 266 nm, which is associated with the defects of nitrogen substituted by O2− in AlN lattice. Figure 7(b) shows the high resolution PLE spectrum in the wavelength range between 350 and 550 nm. In this range, the PLE spectrum consists of several weak peaks at 400, 455, 489, 520, and 548 nm corresponding to the spin and parity forbidden d-d transitions of Mn2+ from the ground state 6A1(6S) to the excited state 4E(4D), 4T2(4D), (4A1(4G), 4E(4G)), 4T2(4G) and 4T1(4G) [17]. The valency of Mn ions in AlN remains controversial, because Mn4+ also exhibits red emission in the 560–670 nm range. Also, the ionic radius of Mn4+ (0.053 nm) is closer to Al3+ (0.054 nm) than Mn2+ (0.067 nm), suggesting Mn4+ is more easily doped into AlN lattice. However, the Mn4+ ions usually produce line emissions. Therefore, the successful doping of Mn2+ in AlN nanowires is also evident from the luminescence spectrum.
In order to study the effect of the Mn content on the optical properties, PL measurements were performed for AlN nanowires with Mn doping concentration ranging from 0.0 to 2.3%, as shown in Fig. 8. With increase of Mn2+ content, the emission intensity increases rapidly, reaches the maximum at 0.9%, and then decreases again at higher doping levels. The deterioration of the emission signal is usually attributed to concentration quenching caused by nonradiative energy transfer between adjacent doped Mn2+ ions.
The PL decay curves of AlN:xMn nanowires monitored at 597 nm and excited at 266 nm are displayed in Fig. 9. The decay curves are well fitted by using the following double-exponential relationship:
where I means the luminescence intensity, I1 and I2 represent constants, t stands for the time, τ1 and τ2 are decay times for exponential components. The detailed fitting parameters and lifetime values of AlN:xMn nanowires calculated by Eq. (1) are listed in Table 2. As with other Mn2+ doped materials, the lifetime is in the range of milliseconds [36]. It is well known that the PL intensity is related to the lifetime of the luminescent center and the radiation velocity [37]. For the same luminescent center, the radiation velocity can be supposed to constant (the amount of radiation photons in a specific period). Therefore, a stronger luminescence intensity corresponds to a longer fluorescence lifetime.
Table 2
Fitting parameters of different decay curves for AlN:xMn nanowires (with x = 0.0%, 0.4%, 0.9%, 1.3% and 2.3%).
Mn ratio (at. %)
|
I1
|
τ1(ms)
|
I2
|
τ2(ms)
|
t(ms)
|
0.0
|
10172.64
|
0.41
|
2314.08
|
2.61
|
1.71
|
0.4
|
9989.51
|
0.48
|
2873.16
|
2.77
|
1.91
|
0.9
|
8134.32
|
0.64
|
4815.46
|
2.76
|
2.16
|
1.3
|
9603.02
|
0.40
|
2765.95
|
2.74
|
1.96
|
2.3
|
10126.27
|
0.39
|
2150.14
|
2.66
|
1.73
|
The thermal stability of light-emitting materials is a key index to evaluate the potential application of the final devices, because it has an impact on the different characteristics of the device. Thus, it is necessary to elucidate the thermal stability of light-emitting materials according to the relationship between intensity and temperature. Figure 10 (a) and (b) presents the temperature dependent PL emission spectra and corresponding relative integrated intensity trend for AlN:0.9%Mn nanowires, respectively. It is important to note that the emission intensity of Mn2+ transition, initially increase slightly (until the temperature reached 353 K), and then gradually drops. When the temperature is below the 413 K, the emission intensity is still higher than initial value at 293 K. In general, an increasing of temperature can increase the population of higher vibration levels, the density of phonons and the probability of non-radiative transfer (energy migration to defects), leading to the emission intensity gradually drops. Here, the AlN:Mn nanowires exhibit an abnormal thermal quenching behavior, which should be related to the traps in AlN band gap. The detail of reason will be discussed later.
Figure 11 shows the afterglow decay curves of the AlN:0.9%Mn nanowires after irradiation by a 254 nm UV lamp for 3 min. The decay curve can be successfully fitted by Eq. (1), and the fitting results are summarized in Table 3. Apparently, the decay curve consists of a rapid attenuation process at first (τ1) and then a slow decay process (τ2). Due to the presence of the significant slow decay component (τ2), the long afterglow features can last more than 30 min with brightness ≥ 0.32 mcd/m2 in the darkroom. The observed afterglow feature of AlN:Mn nanowires can significantly save power compared with traditional LEDs, which paves the way for further dealing with severe environmental and energy problems.
Table 3
Simulated results for the afterglow curves of AlN:0.9%Mn nanowires
|
I1
|
τ1(s)
|
I2
|
τ2(s)
|
value
|
247.19
|
18.40
|
173.58
|
123.26
|
It is considered that the existence of traps is the main reason for the abnormal thermal quenching behavior and afterglow performance. To further verify the presence of traps, the TL curve of AlN:0.9%Mn nanowires is shown in Fig. 12 (a). The TL curve gives an asymmetric shape and a wide temperature region from 300 to 620 K, suggesting a wide trap distribution. However, because the incorporation of larger-size Mn2+ ions into AlN lattice results in the coexistence and complexity of different defects, it's very difficult to discriminate and assign or calculate exact trap depths. Through calculations and experiments, N vacancies (VN) with multiple charge states is generally considered to be formed shallow energy level at 0.12-1.0 eV below the conduction band (CB), and VAl level located at 1.2–2.7 eV above the valence band (VB) [38, 39]. Oxygen impurities are another essential defects, and easy to incorporate into AlN:Mn2+ lattice for change balance. In general, oxygen impurities are located at N position to form separated oxygen point defects (ON) and complex defects with Al vacancies (ON-VAl). The ON energy level is at around 1.9 eV below the CB, and ON-VAl complex level is at around 1.2 eV above the VB [40, 41]. Figure 12 (b) exhibits a mechanism diagram for the abnormal thermal behavior and afterglow process. The strong excitation band centered at 266 nm (Fig. 7 (a)), is attributed to the excitation from the ON-VAl level to its excited state level (process Ⅰ). After UV light irradiation (process Ⅰ), most of the excitation energy related to the excited carriers (electrons or holes) will be directly transferred to the luminescence center Mn2+ (process Ⅱ), followed by the 4T1(4G)-6A1(6S) emission as the immediate luminescence (process Ⅲ). However, when some electrons transition into the CB and then are captured by VN or ON levels and the holes are captured by VAl or ON-VAl levels, instead of returning to ground state (progress Ⅳ). After removing the UV light source, with the thermal disturbances at proper temperature, electrons and holes trapped by the defect levels will be gradually released, transferred to the luminescence center Mn2+ (process Ⅴ), and then emit as long afterglow (process Ⅲ). For thermal quenching process, as the sample is heated to proper temperature, the “stored” electrons in the defect levels can be excited to the excited level (process Ⅴ) and finally return to ground state. The higher temperature, the more electrons are excited through process Ⅰ and Ⅴ to process Ⅲ. Meanwhile, the higher temperature results in greater energy loss of the electrons in the relaxation process, due to the enhanced nonradiative transition, leading to the decrease of electrons in process Ⅲ from process Ⅱ. Thus, the thermal quenching behavior should be the coupling contribution by the above two effects. When the electrons excited from process Ⅴ to process Ⅲ is higher than that lost in the relaxation process, leading to the increase of emission intensity and the abnormal thermal quenching behavior.