- Structural analysis of the prepared configurations on different substrates:
Many attempts were employed in transform the deposited amorphous germanium to poly:Ge via metal induced crystallization process using some metals such as aluminum [12]. Metal induced crystallization occurs when a layer of metal (low melting point) such as Al covered by a semiconductor (SC) layer (high melting point) such as Si and Ge. When the two layers annealed at a temperature that is lower than the crystallization temperature of the semiconductor, the metal layer begins to melt, which causes the semiconducting layer starting to dissolve and diffuse into the metal layer [13] and the dissolved (SC) atoms nucleate. These nuclei conglomerate on each other forming islands of poly:(SC) as shown in the schematic diagram in Fig. 6. Here, we used two different configurations of multilayers on different substrates. Al and Sn are involved to initiate the MIC process and be incorporated in Si and Ge networks.
Figure 1a shows the FESEM micrographs of (Al/Si/Sn/Ge/Sn) and (Al/Ge/Sn/Ge/Sn) multilayers configurations on different substrates. It is observed that the nucleation starts laterally in the sample on the FTO substrate but in the other structure on p-Si, these nuclei start to conglomerate vertically forming hills like shapes of both poly:Si and poly:Ge doped with Sn [14]. The conglomeration increases in the sample deposited on p-Si due to the highly doped base substrate that estimate the aluminum to rapidly diffuse in the Si layer and accelerates the rate of MIC process [43]. It is seen that the rate of crystallization in Al/Ge/Sn/Ge/Sn configuration is increased than the other configurations that involves both Si and Ge. This is because Si has a higher crystallization temperature than Ge, so the rate of crystallization in Ge is higher [15]. The films began to form a continuous Poly:Ge crystallites doped with Sn. It is also observed that the configuration that is deposited on p-Si has smaller grain boundaries than the other on FTO due to the effect of p-Si highly doped substrate that accelerates the diffusion of Al atoms and increases the crystallization of Ge doped Sn as discussed above. These smaller grain boundaries areas increase the charge carrier mobility and the structure becomes suitable for optoelectronic applications [16].
Energy dispersion X-ray analysis asserts the existence of Al, Si, Ge and Sn elements that forms the junctions. It showed also the appearance of Oxygen atoms as a result of the annealing temperature with low vacuum conditions of the value 4×10− 2 mbar. This caused a growth of GeO2 crystallites that give the sample good emission properties along the visible region of spectra as we will discuss later in XRD and Photoluminescence analyses. XRD analysis is carried out to investigate the growth of germanium and silicon nanocrystals and GeSn crystallites as shown in Fig. 1b which shows the existence of GeSn peaks with different orientations in different planes: (111) at 2θ = 27.5° [17], (220) at 2θ = 62.5° [18] and (004) at 2θ = 64° [18]. Ge (220) peak is observed at 2θ = 46° (JCPDS card No. 65–0333) in the sample of the second configuration and disappeared in the first configuration as the second configuration is rich with germanium [17]. Ge (311) (JCPDS card No. 65–0333) is detected at 2θ = 56° for both configurations [17]. Si (111) appears at 2θ = 28.5° (JCPDS card no. 01-079-0613) in the first configuration due to the occurrence of MIC forming poly: Si [4]. There are other three peaks at 2θ = 34°, 38° and 44° corresponding to SnO2 (101), SnO2 (200) (JCPDS Card No 41-1445) and GeO2 (201) JCPDS (card no. 43-1016), respectively due to the annealing at slightly low level of vacuum that causes partially oxidation of Ge and Sn [19, 20]. On the other hand, the configurations deposited on n- Si and p-Si show highly oriented poly-GeSn crystallites at (400) direction of 2θ = 62°. This is due to the effect of single crystalline silicon base substrates as a seed layer that has one direction of crystal growth [21]. It is observed that the intensity of the GeSn peak in the samples deposited on p-Si is higher than the other one deposited on n-Si. This behavior can be attributed to the diffusion of Al atoms away from highly doped p-Si substrate and diffuse in Ge and Si layers which leads to an increase of the rate of MIC. There is a very weak peak of GeSn (004) for the samples on p-Si at 2θ = 64°. XRD analyses show the growth of Ge with a lattice constant = 5.506 Å and 5.57 Å at different orientations for Al/Si/Sn/Ge/Sn and Al/Ge/Sn/Ge/Sn prepared on FTO respectively. In both configurations that prepared on FTO, it is observed that there is incorporation of Sn inside Ge network that causes the growth of GeSn crystallites with different orientations for both configurations [6, 17]. As a result of interstitial incorporation, the lattice constants of the different orientation growth of Ge increased to positive tensile strained for GeSn up to 5.96 Å for the growth direction of plane (400) since the lattice constant of Sn = 6.49 Å is higher than the germanium [22]. It causes a tensile force inside Ge network which increased the lattice constant as mentioned [17]. On the other hand, it is observed the preferred high orientation of the growth of the strained GeSn is associated with an increase of the lattice constant up to 6.06 Å (See supplementary Table S1), which refer to higher incorporation of Sn inside Ge network. This mostly leads to a one direction crystal growth, which increases the electron mobility than the growth in many orientations [23]. The higher strain in the Ge network modulates the valley of the valence band to be aligned with the bottom of conduction band yields modulation of the indirect band gap to direct one [5, 8]. Figure 1c shows the Raman spectra of the fabricated heterostructures. It is observed that for the Al/Si/Sn/Ge/Sn on p-Si device, there is a peak at 500–510 cm− 1 due to the existence of aluminum and tin induced crystallization of amorphous silicon to form poly:Si [24]. The peak at 520 cm− 1 is due to the crystalline p-Si wafer [24]. SiGe is observed at around 390–400 cm− 1 as a result of the alloying between Si and Ge at the high annealing temperature [24]. An appearance of Poly:Ge and GeSn modes are detected at 300 cm− 1 and 270 cm− 1 respectively, which may be due to the transform of the amorphous germanium layer to polycrystalline one via tin induced crystallization and incorporation of Sn metal inside Ge network [25]. The figure also shows two peaks at 180 cm− 1 and 120 cm− 1 that are due to Sn-Sn and SnO respectively [26]. For the heterostructures of Al/Ge/Sn/Ge/Sn on p-Si and n-Si, a peak of NC: Si appeared at 520 cm− 1, which is due to the n type or p type silicon substrates. There are other modes appear at 300 cm− 1 and 270 cm− 1 that are corresponding to poly:Ge and GeSn. Other modes at 180 cm− 1 and 120 cm− 1 refer to Sn and SnO as observed in XRD analysis. Raman confocal microscope images of the fabricated heterostructures are used to investigate the effect of different substrates on the crystallization rate. For the heterostructure of Al/Si/Sn/Ge/Sn on p-Si, the formation of islands of the doped crystallites for both Si and Ge with different colors is observed which may be due to the metal induced crystallization of the deposited amorphous Si and Ge via aluminum and tin. For Al/Ge/Sn/Ge/Sn, the formed clusters are due to Ge nanocrystals as discussed in FESEM images (See supplementary figure S2).
- Direct transition investigation using diffuse reflectance spectroscopy, PL measurements at different cooling temperatures.
As a result of the long-time annealing (24 h) at 500 ºC, Al and Sn are interstitially existing in the Si and Ge networks. These metals incorporation inside Ge and Si modulate the optical energy gaps of these IV group indirect transition semiconductors and the direct optical transition occurs [3, 4]. Figure 2(a) I, II, III, IV and V shows the plot of (Khʋ)2 versus hʋ using modified Kubelka-Munk model of the heterostructures of micro-strained poly-germanium doped with Al and Sn on n-Si, p-Si and FTO, respectively. The higher observed absorption for the heterostructures that interfaces p-Si and FTO substrates than the other one on n-Si. This can be explained as mentioned before: In the heterostructure on FTO, the diffusion rate of metals in Ge is higher since the surface area of interaction between Ge and metal layers increases as a result of the change in morphology of the FTO bottom layer as a result of the high annealing temperature near its melting point. In the case of heterostructure on p-Si, the Al atoms doesn't segregate in the p-doped substrate but these atoms of Al layer are rapidly segregating in the lower majority carriers concentration intrinsic Ge layer and thus increase the rate of crystallization and incorporation of Sn increasing GeSn direct allowed transition [27]. A direct transition of GeO for all heterostructures is ranged between 1.15 to 1.85 eV due to different level of oxygen incorporation in the samples [27]. Another direct transition at Ge crystallites of higher level of oxidation is around 2.02 up to 2.73 eV occurs in the structures prepared on FTO and p-Si substrates. This was suggested due to the diffusion of oxygen atoms of FTO layer in Ge layers under the influence of the annealing that causes vacancies in FTO layer and the low level of vacuum during annealing of the rapid diffusion of multilayers that is on p-Si substrate [28, 29]. Direct transition in the strained GeSn crystallites in all samples is ranged between 0.43 up to 0.61 eV (See supplementary Table S2). These direct transitions at different regions of the visible and near infrared spectrum give these heterostructures good chance in optoelectronic devices applications. The bulk germanium has emission spectra at around 1400–1600 nm. The bulk GeSn alloy has emission at NIR 1800–2200 nm [30, 10]. These ranges of emission may be shifted towards the blue region as the effect of oxidation level [31]. Photoluminescence emission measurements of the prepared heterostructures based on the tensile strained Ge nanocrystals show emission peaks at different regions from the violet light to NIR (Fig. 2b). Two GeOx peaks, the violet peak at 402 nm is due to the oxidation of Ge which is resulted from the annealing at low vacuum up to 10− 2 mbar. The oxidation of Ge shifts the germanium emission to the short wavelengths [32]. Another green peak of GeOx is 514 nm. A red peak at 790 nm is due to lower degree of oxidation as shown in Figure (2b I) [33]. Figures (2b II) shows a broad peak at around 1223 nm in NIR region. It is due to Ge doped with Sn forming GeSn strained nanocrystals [34]. It is observed that the 790 nm and 1223 nm peaks disappear; the 402 nm violet peak and the 514 nm green peak are shifted towards the red region for the prepared Al/Ge/Sn/Ge/Sn on n-Si heterostructure. It is due to the lower rate of MIC process than other samples prepared on p-Si and FTO substrates and hence lower incorporation of oxygen and tin in Ge lattice as mentioned in structural measurements. It is observed a higher emission intensity for both Al/Si/Sn/Ge/Sn and Al/Si/Sn/Ge/Sn heterostructures on p-Si and FTO that increases the rate of MIC process and forming NC:Ge and strained GeSn structures. Figure 2c shows the emission of the prepared configurations on different substrates at different cooling temperatures from 280 K to 20 K. The intensity of the peaks increased with the decreasing of the temperature down to 20 K. This confirms the direct transition occurrence and increasing the radiative recombination [35]. When the temperature is decreased, it enhances the radiative nonthermally transitions [36]. The Al/Si/Sn/Ge/Sn on FTO or p-Si show lower intensity than the Al/Ge/Sn/Ge/Sn on FTO since the second one is rich with germanium. The peak at around 1200 nm in NIR region has higher intensity in Al/Si/Sn/Ge/Sn on FTO and p-Si than the Al/Ge/Sn/Ge/Sn on FTO as a result of the higher solubility of Sn in Ge than its solubility in Si [37]. This will increase the incorporation of Sn in Ge lattice away from silicon causing high intense GeSn peak [38]. The heterostructure of Al/Si/Sn/Ge/Sn on FTO has a continuous wide emission spectra from the visible to NIR region than other structure of the same configuration on p-Si that has its emission mostly in NIR region as a result of the various oxidation levels due to the diffusion of oxygen atoms from FTO layer into the above layers as discussed throughout this article.
- I-V characteristics of the fabricated heterostructures junctions under dark conditions
The diode parameters such as ideality factor n, saturation current Is, threshold voltage Vth and the barrier height are calculated using Schottky Eq. (1) [39] and Richardson Eq. (2) [40].
ID = Is[\({e}^{\frac{VD}{nKT}}-1]\) (1)
(2)
The measured IV characteristics are carried out at room temperature in the range of -3 volt to + 3 volts. We obtained four diode junctions of the prepared heterostructures of the two prepared configurations on p-Si, n-Si and FTO substrates as shown in Fig. 3a. Table S3 summarizes the calculated parameters of the prepared junctions. The figure shows that there is no avalanche effect in the reverse biasing of the prepared junctions and a very small dark current in the order of 10− 4 A. This makes the junctions able to use in rectification applications [5, 41]. The ideality factor n is calculated using Schottky equation from the plot of Ln I versus V as shown in Fig. 3b. The values of the ideality factor of the junctions in low voltage biasing region (Region 1) are around 1 near the value for ideal diode structure. However, in the high voltage biasing region up to ± 3 volt (Region 2), the values of the ideality factors are in range of 6 to 35 that are far from the ideal value. This may be due to the tunneling of conduction electrons inside the cloud of metals that contains vacancies in which electrons recombine [5, 42]. From the values of the ideality factor, we can observe the effect of the substrate as mentioned in Raman analyses. In both configurations on p-Si substrate, Al atoms are segregated away from the highly doped p-Si wafer inside Si or Ge networks that cause the trapping effect at the low or high voltage biasing regions. Since the samples have very low dark leakage current that contributes as a noise current [43], they have good opportunity in photodetection applications. Rectification ratio RR calculations refers that the Al/Ge/Sn/Ge/Sn on (n-Si) substrate has lower dark current and higher rectification ratio near 100 that make this heterostructure is a promising junction in photodetection application. The barrier height values indicate that the samples can be detected in near infrared NIR region since the height of the barrier is in the range of 0.6 to 0.8 eV for all samples [44].
- Fluorescence charge carrier lifetime investigation:
Figure 4a shows the raw FLIM data for Al/Si/Sn/Ge/Sn and Al/Ge/Sn/Ge/Sn on FTO, p-Si and n-Si substrates. Photoluminescence intensity shows a decay curve for all the heterostructures. The FLIM images of the samples agree with the densities of the prepared samples that is investigated in PL spectroscopy measurements. Both PL spectroscopy and FLIM images show the highest intensity of emission is corresponding to Al/Si/Sn/Ge/Sn on p-Si heterostructure and the lowest emission is corresponding to Al/Ge/Sn/Ge/Sn on n-Si heterostructure. Figure 4b shows the phasor plot for the prepared heterostructures. It gives a representation of the intensity decays for the corresponding FLIM image shown in the left panel of the figure. In the phasor plot, each pixel in the FLIM image is represented in a 2D diagram with two coordinates namely, S and G. The two coordinates based on the phase shift (φ) between the transmitted wave and the resulting PL wave and demodulation factor (m) in the lase source. The S and G components are given by the following equation [45, 46]:
(3)
where τ is the lifetime and ω=2πf is the laser modulation angular frequency (80 MHz). From Eq.4, the lifetime
(4)
The phasor plot for the prepared heterostructures showed a cluster of point located at the edge of the phasor plot semicircle, indicating that the prepared samples decay is single exponential decay with an average lifetime ranged between 2.22 up to 2.73 ns as calculated from Eq. 4 (See Fig. 4b)). This charge carrier lifetime of GeSn prepared structures is suitable for diode lasing application compared to B. Julsgaard 2020 [47] and S. De Cesari 2019 [48].
- EL measurements of the prepared white LED of the two configurations.
As EL measurements show that the heterostructures emit different wavelengths along the visible spectra region when we applied 11.6 V on one of these junctions and emitted white light. EL measurements show the emission of the prepared Al/Si/Sn/Ge/Sn configuration at visible and NIR (See Fig. 5a). This configuration has higher emission at NIR region due to the more Sn incorporation in GeSn as discussed before. The second configuration Al/Ge/Sn/Ge/Sn has intense EL emission than the other configuration as it is rich with germanium layers. The prepared configurations can be used as a white light emitting diode at low voltage as shown in Fig. 5b. When we increased the voltage from 15 V up to 45 V, the emission increases and very sharp peaks appear along the spectrum from 300 nm to 800 nm (See Fig. 5c and d). The main peak is at about 590 nm with very high intensity that is the main transition discussed in UV-Vis spectroscopy analysis. It gives these samples the chance to act as a white light laser diode media [49] with very small line broadening. The FWHM is equal 2.168 nm or 1.74 meV.