1. Introduction

Diluted magnetic semiconductors (DMSs) are anticipated to play an important role in multi-functional electronic devices as a promising material due to the possibility involving charge and spin degrees of freedom in a single substance [1].Since Dietl et al. [2] theoretically predicted that ZnO doped with “magnetic” atoms like Co, Mn, or Fe possess room temperature ferromagnetic (RTFM) behavior, Mn-doped ZnO has led to an intense research interest because of several substantial advantages of Mn ions. It possess the highest magnetic moment along the 3d series and a possible fully occupied majority 3d band, which results in a stable fully spin polarized state. Though the experiment results on RTFM are still under dispute [3–8], Mn-doped ZnO has been widely studied as an important material in different fields. Peng et al. [9] reported that Mn-doped ZnO thin ðlms grown on Pt and Si substrate show unipolar and bipolar resistive switching (RS) behaviors respectively. Z. Yang et al. [10] observed both positive and negative large magneto resistance (MR) in Zn_1-xMn_xO (x<0.35) epitaxial ðlms. But it has never been used as a LPE material to our knowledge.

In this work we will first report a significant infrared LPE observed in Mn-doped ZnO film based on DMS/SiO₂/Si structure. This will add a completely new functionality to Mn-doped ZnO and make it a potential candidate for versatile materials. In addition we make a detailed study of the LPV response to different wavelength and optical power laser. The results show that the position sensitivity of LPV has a dual dependence on both the laser wavelength and optical power. A complete theoretical analysis is presented. This will help modulating LPV in similar LPE materials.

2. Experimental details

The Zn_0.99Mn_0.01O films were grown on n-type Si (1 1 1) substrates by co-sputtering ZnO ceramic (99.99%) and Mn metallic (99.99%) targets. A 0.6 Pa Ar deposition pressure was maintained in a high vacuum system better than 6.0 × 10⁻⁵ Pa prior to deposition. The deposition rate, determined by stylus profile meter on thick calibration sample, was 0.35 Å/s.

The substrates used in our experiment were covered with a native SiO₂ layer about 1.2 nm measured by transmission electron microscopy (TEM). The thickness of the wafer is around 0.3 mm and the resistivity is in the range of 50–80 Ω cm at room temperature. All samples were scanned spatially with lasers focused on a roughly 50 µm diameter spot at the film surface without any spurious illumination (e.g. background light). Measurement details are similar with our recently published papers on LPE [11,12].

3. Results and discussion

Scanning electron microscopy (SEM) images of Zn_0.99Mn_0.01O and pure ZnO films are displayed in Fig. 1 .The morphologies are quite similar, implying doped Mn element is incorporated in ZnO lattice. This result is quite consistent with structure identification of Mn-doped ZnO by different methods in previous literatures [13,14], including Raman spectra (RS), X-ray diffraction (XRD), transmission electron microscope (TEM) and so on. All reports agreed no second phase formed when Mn concentration is below 7at. %, indicating the substitute of Zn by Mn. Zhang et al. [15] confirmed this conclusion using atom location by channelling-enhanced microanalysis (ALCHEMI).

 

Fig. 1 SEM images of Zn_0.99Mn_0.01O film (a) and pure ZnO film (b).

Download Full Size | PDF

The operation mechanism of LPE in ZnO doped materials based structures has been interpreted in detail [12]. As the high transmittance of ZnO doped films to incident lasers is important. Here we give a comparison result between transmission spectra of Zn_0.99Mn_0.01O and pure ZnO film both corrected for a glass substrate at room temperature, as shown in Fig. 2 . The incorporation of Mn element induced a little shift of absorption edge toward higher energy side and a slight decrease of transmittance compared to pure ZnO film, in agreement with result in previous literature [16].But still it is a good transparent film in Vis-IR region.

 

Fig. 2 Optical transmission spectra of Zn_0.99Mn_0.01O and pure ZnO film.

Download Full Size | PDF

To investigate the dependence of the LPV response to lasers of different power and wavelength, LPV measurements are designed in two different modes defined as CW/CF (constant wavelength/constant frequency) mode and CP (constant power) mode.

Figure 3(a) shows the LPVs as a function of laser spot position in Mn-doped ZnO film in CW/CF mode under a 780 nm laser illumination. The output optical power ranges from 0 to 10 mW through an optical attenuation. All results present a perfect linear characteristic of LPV versus laser spot position and the obtained position sensitivities range from 4.7mV/mm to a saturated value of 46.9mV/mm. The dependence of position sensitivity on optical power is displayed in Fig. 3(b). It is clear that a threshold power value exists. The position sensitivity is proportional to the laser power as below the threshold value and then slowly becomes saturated as the power exceeds the threshold value.

 

Fig. 3 (a) LPVs as a function of laser position in Mn-doped ZnO film under 780nm laser illumination with different optical power (b) Position sensitivity dependence on optical power at λ = 780nm.

Download Full Size | PDF

In CP mode we choose four typical wavelength lasers from visible to infrared, including 532nm, 780nm, 810nm and 980 nm. As the obtained result in CW/CF mode we designed the LPV measurements of CP mode in two separate ranges. One group is with a 0.1mW power below all threshold values (called as BTV range) while the other is with a 6mW power above all threshold ones (called as ATV range).The measurement results are shown as Fig. 4(a) and 4(c). Noticeable improvement is the observation of significant LPE in infrared region, which is of prime importance for the application of LPE in infrared position sensitive detectors.

 

Fig. 4 (a) LPVs as a function of laser position in Mn-doped film with P = 0.1mW laser illumination of different wavelength (b) Position sensitivity dependence on wavelength at P = 0.1mW (c) LPVs as a function of laser position in Mn-doped film with P = 6mW laser illumination of different wavelength (d) Position sensitivity dependence on wavelength at P = 6mW.

Download Full Size | PDF

Another interesting phenomenon is the almost inverse result of the LPV position sensitivity dependence on wavelength in BTV and ATV range. As shown in Fig. 4(b) and 4(d), in BTV range the LPV position sensitivity increases with the increasing wavelength and gets a largest value of 5.61mV/mm with an infrared laser (980 nm) illumination. As a contrast, in ATV range the LPV position sensitivity decreases with the increasing wavelength and the largest gets 62.2mV/mm with a 532nm laser illumination. Therefore this DMS/SiO₂/Si based LPE shows a dual-dependence on both wavelength and optical power.

4. Physical mechanism

To better understand the infrared-sensitive LPE and explain the inverse dependence of position sensitivity on wavelength behind this DMS/SiO₂/Si-based LPE, we proposed the following model.

Due to the high transmittance of the Mn-doped ZnO film, when the illuminations occurred energy was mainly absorbed in Si substrate to generate electron-hole pairs (photon energy of illuminated laser above the Si band gap is required), as shown in Fig. 5(a) . As the Si substrate is thick enough (0.3mm) in our experiment, we assume all transmission photons are absorbed if neglecting the reflection losses. Then the amount n of light excited electrons under illumination of a laser with wavelength λ (or frequency ν) and optical power P can be written as:

 

Fig. 5 (a) Diagram of energy absorbed in Si substrate (b) light-induced electrons all transit to Mn-doped ZnO film (c) light-induced electrons partly transit to Mn-doped ZnO film.

Download Full Size | PDF

In terms of quantum theory, the quantized photon energy is described by hν (h is the Planck’s constant and ν is the laser frequency. Here t represents the transmittance of laser to Mn-doped ZnO film. Then the excited electrons transit to the Mn-doped layer from the Si substrate by the built-in field and move laterally away from the illumination spot. If the lateral distance of the laser spot from each electrode is different, the quantity of the collected electrons at the two contacts is different, which results in the LPV. Ideally, the LPV is proportional to the laser position and the involved sensitivity of LPV can be presented as [12,17,18]

Here l₀ is the electron diffusion length in Mn-doped ZnO layer. L is the half distance between two electrode contacts and k is a proportional coefðcient. N is the effective electrons number that transit to Mn-doped ZnO layer among all light exited ones. Usually we describe the relation between N and n through a coefficient $\xi$ called quantum efficiency, which is a wavelength dependent factor for fixed material and structure. And the relation can be written as:

Then in CW mode the LPV position sensitivity dependence on optical power can be well explained. For monochromatic laser all factors in Eq. (4) are constant except the optical power with a premise that the number N below threshold value. So the position sensitivity shows a proportional relation to the optical power in BTV range. As the optical power gets large enough the position sensitivity gets saturated and changes little in ATV range.

The result in CP mode is a little complex as there are two associated parameters changing. Quantum efficiency $\xi$ is a wavelength dependent factor and can’t be modeled directly. In solar cells technology ${\xi }_{\lambda }$ is obtained through a response spectrum [19,20]. And for fixed material or structure the response spectrum shows unique curve. But in CP mode the position sensitivity dependence on wavelength shows almost inverse result in BTV and ATV range and we ascribed this to different dominant factor. For better understanding we discussed ${\xi }_{\lambda }$ through the recombination coefficient r in detail. The relation can described as:

The recombination coefficient $r_{\lambda }$ is quite different in BTV and ATV range. When the optical power is in BTV range, energy is totally absorbed within a limit depth for lasers of all wavelengths. This limit depth can be neglected compared to electron diffusion length in Si substrate as shown in Fig. 5(b). Thus almost all light-induced electrons transit to Mn-doped ZnO film. In this condition contrast of factor r for different wavelength has little influence in BTV range while the amount of light excited electrons n is dominant. Thus the position sensitivity will increase with the wavelength according to above equation.

In ATV range the light exited electrons all get saturated. Thus the recombination coefficient r must be considered for the energy absorbed region can’t be neglected as shown in Fig. 5(c). As it is well known that the electrons with a larger energy in Si can obtain a longer diffusion length. Then the recombined electrons number with short wavelength illumination is much less compared to long wavelength and results in more electrons left that can transit to the film. Thus the position sensitivity decreases with the increasing wavelength.

5. Conclusion

In conclusion, infrared LPE observed in Mn-doped DMS film is first reported. The LPV shows a dual dependence on both the laser wavelength and optical power and this dual relation is carefully studied. Significant advance is the discussion about the LPV dependence on wavelength with constant optical power below or above threshold value and a complete theoretical analysis is given. This research may inspire and promise more opportunity for the future application of DMS materials and help modulating LPV in similar LPE materials.