Regrowth of Ge with different degrees of damage under thermal and athermal treatment

Abstract

In this report, the recrystallization of pre-damaged Ge samples is extensively investigated under steady-state thermal annealing and ultrafast thermal spike-assisted annealing generated by high-energy ions. The (100) single-crystal Ge samples were pre-damaged using 100 keV Ar ion implantation. Three sets of pre-damaged Ge samples with sub-threshold (set A), threshold (set B) and above-threshold (set C) doses of amorphization, as estimated by Rutherford backscattering spectrometry in channeling mode (RBS/C), were suitably selected. Cross-sectional transmission electron microscopy (XTEM) images show distributed damaged pockets surrounded by crystalline material in the case of the as-damaged set A sample and completely damaged layer in the set C sample. These samples were used to study the regrowth of damage by (i) vacuum annealing at temperatures ranging from 373 K to 873 K for 30 minutes each and (ii) 100 MeV Ag ion irradiation-assisted annealing at four different temperatures: 100 K, 300 K, 373 K and 473 K. After 100 MeV Ag ion irradiation, set A samples have undergone complete recrystallization at 473 K. Similar recrystallization, but with lower magnitude, is also observed in the set B sample with increase in temperature. In set C samples, interestingly, nanowire formation was observed instead of recrystallization after irradiation at 100 K and 300 K, but recrystallization is observed at high-temperature irradiation, though it is much lower than those of set A and set B samples. The Arrhenius plot of the recrystallized fraction reveals a reduced activation energy of recrystallization by a substantial factor due to thermal spike-assisted recrystallization.

---

Introduction

Germanium is a very significant material used for various applications in numerous fields of nanotechnology,^1,2 and it is important for high-mobility nanodevice applications.^3,4 The band gap of Ge is suitable for photo-absorption at communication wavelength,^5,6 making it attractive for the fabrication of high-quality photodetectors.^7,8 Further applications of Ge include nanoscale transistors,⁹ high-efficiency anodes for lithium ion batteries^10,11 and high-performance electronic devices.¹² For device technologies, doping of a semiconductor is required. But the energetic ions, in turn, induce amorphization or damage in the lattice, which is required to be annealed.¹³ Low-energy ion interactions are dominated by elastic processes or nuclear stopping, resulting in the ballistic atomic displacements of substrate atoms.¹⁴ Hence, damage creation is attributed to energy transfer to the atomic structure, which results in displacement of target atoms from their lattice sites. Ion implantation and recovery processes have been studied extensively by various groups in the past decade.^15–17 Ion beam-induced recrystallization has also been studied in Si,¹⁸ and the possible role of electronic energy loss on recrystallization in Si was reported. On the other hand, besides of numerous applications there are few relevant reports on the study of damage formation in Ge due to low- and high-energy ion implantation, e.g. 56 keV B ions,¹⁹ 185 MeV Au ions²⁰ and 20 keV B ions.²¹ The main difference between Ge and Si is that Ge becomes porous after a critical dose.^22,23 There is a lack of extensive study on ion implantation-induced defect formation and its thermal annealing in Ge. Decoster et al.²⁴ investigated the damage evolution due to implantation and its recovery via thermal annealing. Crystallization from a disordered structure is eventually a diffusive process, and it is strongly temperature-dependent. Hence, thermal annealing is the traditional method for elimination of implantation-induced defects and for activation of implanted impurity. However, in the case of Ge-based devices, high-temperature annealing is not advisable.²⁵ Nowadays, solid-phase epitaxial growth,²⁶ electron, flash lamp,²⁷ laser²⁸ and swift heavy-ion (SHI)^29,30 irradiation are the most widely used methods for recrystallization of a-Ge due to their uniformity in depth distribution. The evolution of pre-damaged Ge and recrystallization behavior due to the effect of energetic ions at room temperature has been reviewed earlier by our group.²⁹ Such interactions are governed by inelastic processes where electronic stopping dominates, which results in the excitation and ionization of Ge atoms. In the case of irradiation with SHI, the major part of energy is deposited to the loosely bound electronic subsystem and then transferred to the atomic subsystem through electron–phonon coupling.³¹ The elastic processes with dominating nuclear stopping are negligible in this regime. So, the annealing of pre-existing damage is expected.

In this work, we prepared three sets of Ge samples in which we deliberately introduced a defined degree of damage, such that the first set (A) with damage of ∼0.25 displacements per atom (dpa) contains isolated amorphous pockets surrounded by crystalline material, the second set (B) with damage of ∼0.5 dpa contains interconnected amorphous zones, and the third set (C) with damage of ∼7 dpa contains a fully amorphous layer. Displacements per atom are estimated on the basis of theoretical formulation followed in TRIM simulation.³² The recrystallization dynamics of these pre-damaged Ge samples is reported here under ultrafast thermal spike-assisted annealing produced by 100 MeV Ag ion irradiation at variable temperature and compared with steady-state thermal annealing at temperatures up to 873 K. The recrystallization is characterized by RBS/C and micro-Raman spectroscopy and is supported by X-TEM.

Experimental

In order to understand and quantify the damage recovery and ionization effects in a better way, different pre-damaged states were introduced in Ge by means of low-energy ion irradiation with 100 keV Ar⁺. The high nuclear stopping power S_n (0.6 keV nm⁻¹)³² is responsible for the displacement damage production. Different fractional damage levels were introduced using different doses, with peak disorder at a depth of ∼80 ± 10 nm. Three sets of samples with damage of 0.25 dpa, 0.5 dpa and 7 dpa at the damage peak were chosen. These samples with specific degrees of damage correspond to Ar ion fluences of 4 × 10¹³ (set A), 8 × 10¹³ (set B) and 1 × 10¹⁵ (set C) ions cm⁻², respectively. The evolution of crystalline damage resulting from Ar ion irradiation was studied as a function of ion dose by Rutherford backscattering spectrometry in channeling mode (RBS/C). The RBS/C experiments were carried out at the Inter University Accelerator Centre (IUAC) in Delhi, India, using 2 MeV He ions. RBS/C results demonstrate that the channeled yield of the sample irradiated at 8 × 10¹³ ions cm⁻² (set B) just touches the random spectrum. This indicates irradiation at the threshold dose of amorphization. The depth distribution of defects from the RBS/C spectra was computed using the simulation code DICADA,³³ which is based on the discontinuous model of dechanneling. The transmission electron microscopy (TEM), using a FEI TF30, S-TWIN microscope operating at 300 kV equipped with a GATAN Orius CCD camera, was used to investigate the state of damage in these samples.

The damage recovery in these pre-disordered states was studied by sequential ionization assistance at variable temperature and over a range of ion doses. Hence, the three sets of samples were irradiated with 100 MeV Ag ions using a 15 UD Pelletron Accelerator at IUAC Delhi, at temperatures ranging from 100 K to 473 K. Such experimental conditions provide a controlled investigation to evaluate the ionization effects separately without introducing considerable displacement damage due to elastic collisions. Moreover, these selected sets of damaged samples were annealed in vacuum with a base pressure of 1 × 10⁻⁶ mbar for 30 minutes at temperatures ranging from 373 K to 873 K to study recrystallization under steady-state thermal annealing. The rate of increase of temperature was kept at 5 °C min⁻¹. Nanowire formation was observed after irradiation of set C samples at room temperature.²⁹ The effect of post-Ag ion irradiation annealing is also studied in these samples, with annealing at up to 873 K in a vacuum environment.

The evolution of the crystallized fraction resulting from thermal annealing and athermal annealing was studied as a function of irradiation dose and irradiation temperature using RBS/C. For detailed quantitative analysis of the change in damage fraction, simulation of the RBS/C spectra was performed using DICADA.³³ Micro-Raman spectroscopy of the samples was carried out using the Renishaw inVia Raman spectrometer for 514.9 nm wavelength laser with spot size of 1–2 μm.

Results

The ion beam-induced amorphization and subsequent damage recovery under thermal as well as athermal (ion beam-assisted) annealing processes in Ge (100) were analyzed with the help of RBS/C, Raman and high-resolution TEM analysis. Cross-sectional TEM (XTEM) images recorded for the as-damaged set ‘A’ sample, which were prepared using 100 keV Ar with the fluence of 4 × 10¹³ ions cm⁻², show isolated distributed damaged pockets surrounded by crystalline material (see Fig. 1(a)) up to a thickness of ∼110 ± 10 nm from the surface, as shown in inset I. The average size of damage pockets was estimated and found to be 4–5 nm in size. Some of them are pointed out in this image. These damaged regions are surrounded by nanocrystalline particles with much larger sizes than the damaged pockets. Inset II in Fig. 1(a) is the fast Fourier transform (FFT) pattern of the as-damaged set A sample collected from the region highlighted by a red rectangle in the image. It shows diffused rings along with presence of some crystal spots. Fig. 1(b) is the TEM image of the as-damaged set C sample prepared with the fluence of 1 × 10¹⁵ ions cm⁻² using 100 keV Ar ions , showing a uniform amorphous layer of thickness ∼170 ± 10 nm consisting of nanocrystallites. Hence, different disorder profiles of ∼0.25, 0.5 and 7 DPA were produced using 100 keV Ar ions in the corresponding set of samples such that the samples consist of isolated damaged regions, joining damaged regions, and completely amorphous layer, respectively.

Fig. 1 XTEM images of damage creation in 100 keV Ar ion-irradiated: (a) set A sample consisting of isolated damage pockets; inset I shows the image of the entire affected layer; inset II shows the FFT pattern of as-damaged Ge taken from the region marked with a red rectangle; (b) set C sample.

Fig. 2 shows the cross-sectional HR-TEM image of set B sample annealed at 873 K. One can see clearly from this lattice image that substantial defect annihilation has happened after thermal treatment. The HRTEM image is collected from the near-surface region, and it shows uniform morphology with very few defects and single crystalline nature. However, a few residual defects, like dislocation loops in the pre-damaged layer, are present after annealing, as shown in the low-magnification XTEM image in inset I of Fig. 2. FFT pattern (shown in the inset II) taken from the region highlighted with rectangle in this HRTEM image shows a distinct spot pattern, which again confirms the crystalline quality of the layer. The residual defects lie at a depth of ∼40 nm from the surface. Moreover, the material is recrystallized above and below the layer containing these defects. From the contrast of the low-magnification TEM image in inset I, the presence of an implanted layer is clear.

Fig. 2 XTEM image of set B sample after steady-state thermal annealing under vacuum condition, at temperature 873 K, showing recrystallization in the near-surface region. Inset I is the low-magnification image of this sample, which shows the residual defects at a depth of ∼33 nm after annealing even at such high temperature. Inset II is the FFT pattern collected from the region highlighted with a red rectangle.

Micro-Raman and RBS/C spectra were recorded for set A, B and C samples in as-damaged state as well as after thermal and athermal annealing to complete the study over the entire range of samples. Fig. 3(a) and (b) show the micro-Raman spectra of set B and set C samples, respectively, after thermal treatment. The spectrum of c-Ge wafer is also shown for comparison. The peak at 301 cm⁻¹ in pristine Ge is related to the longitudinal optical (LO) phonon mode of c-Ge. In the as-damaged samples, a broad band centered at around 270 cm⁻¹ is observed, which corresponds to the LO phonon mode of amorphous Ge phase. For set B and C samples, clear indication of recrystallization is observed from the Raman spectra after annealing. The increasing intensity of c-Ge peak (301 cm⁻¹) with temperature indicates an increasing contribution of crystalline phase at the cost of the amorphous phase. In set B samples, recrystallization starts after annealing at 573 K, as the peak corresponding to c-Ge starts to appear (see Fig. 3(a)). Further increase in temperature results in increase in intensity and sharpness of the peak at 301 cm⁻¹, which corresponds to the undamaged (c-Ge) sample. Fig. 3(b) shows the appearance of c-Ge peak at 301 cm⁻¹ at 773 K, which became strong and sharper at 873 K, emphasizing the damage recovery in set C samples with increasing annealing temperature. However, the amorphous component gives rise to a tail in this peak. The penetration depth of the laser used for Raman measurements was found to be ∼20 nm in Ge. This clearly indicates the near-surface recrystallization after annealing. This is also reflected in the HRTEM image of this sample shown in Fig. 2. The re-crystallization results were further corroborated by RBS/C. Thermal agitation-induced recovery is manifested as a reduction in disorder with increasing temperature. The damage recovery behaviour in the pre-damaged samples with D₀ = 0.5 dpa (set B) and 7 dpa (set C), evaluated with the help of simulation code DICADA, are shown in Fig. 4. Up to 473 K, no significant effect on damage profile was observed. In set B samples, though the width of the damage profile starts decreasing after annealing at 573 K, reduction in amplitude of the damaged regions is observed only at 773 K, as shown in Fig. 4(a). This emphasizes that the simple defects in tail regions start annealing at lower temperature. However, the complex defects start annealing only at 773 K. The width of damage profile is reduced from 190 nm to 50 nm, but the peak damage is reduced by 90%. Moreover, the damage peak maxima at ∼40 nm from the surface in Fig. 4(a) indicates some remnant defects in the set B sample, even after annealing at 873 K, which enhanced the backscattered yield of the RBS/C spectrum. This result is supported by the TEM result shown in Fig. 2, where defects as the end product of annealing treatment at 873 K are present at a depth of ∼40 nm below the surface. Fig. 4(b) shows reduction of the damage profile width at 773 K but onset of annealing of the peak damage at 873 K for set C samples. The damage profile width decreased from ∼210 nm to ∼53 nm at 873 K. Since the material is not yet completely recrystallized, asymmetry is present in the LO peak related to c-Ge (301 cm⁻¹) in the Raman spectrum of the set C sample annealed at 873 K, as shown in Fig. 3(b).

Fig. 3 Micro-Raman spectra of thermally annealed (a) set B and (b) set C samples showing regrowth of damage after a particular temperature. For comparison, the spectra for 100 keV Ar ion-irradiated samples (as-damaged) are also shown in both cases.

Fig. 4 Damage profile (extracted from DICADA), as a function of annealing temperature, of the steady-state thermally annealed (a) set B and (b) set C samples. For comparison, the spectra for 100 keV Ar ion-irradiated samples (as-damaged) are also shown in both cases.

In the case of 100 MeV Ag ion irradiation-assisted annealing, RBS/C and Raman spectroscopy results indicate significant recrystallization even at room temperature. The recovery process was prominent in set A samples, but in set B samples recovery was observed across the amorphous-crystalline boundary but not in the central zone of the damage region. However, in set C samples no recrystallization took place after Ag ion irradiation up to the highest dose of 1 × 10¹⁴ ions cm⁻². Instead, the region turned into nanowires after 100 MeV Ag ion irradiation, as we have reported earlier.²⁹ As a consequence of sequential high-temperature Ag ion irradiation at various ion doses, significant damage annealing is observed in the pre-damaged region with the help of detailed analysis of RBS/C and Raman spectra. Fig. 4 shows the micro-Raman spectra of the corresponding set A samples irradiated at three different substrate temperatures along with the undamaged Ge sample. In the as-damaged sample, a broad band is observed at around 270 cm⁻¹, which is related to the LO phonon mode of a-Ge phase, along with a small peak at 301 cm⁻¹, demonstrating the co-existence of both crystalline and amorphous phases. The peak related to the crystalline component increases with increasing Ag ion dose, which indicates increasing recrystallization with irradiation. We also observe that recrystallization is more efficient when samples are irradiated at higher temperature. From Raman spectroscopy results shown in Fig. 5(a), it can be concluded that in set A samples, the recrystallization of defects starts even at 100 K. Considerable damage recovery is observed after Ag ion irradiation at room temperature. Further increasing the substrate temperature to 373 K results in increasing intensity and sharpness of the peak at 301 cm⁻¹, though the amorphous component produces only a tail in this peak, revealing near-complete recrystallization as shown in Fig. 5(b). Moreover, complete recrystallization is observed on irradiating with highest dose at 473 K (Fig. 5(c)). Fig. 6(a) and (b) shows the damage profile of set A samples irradiated at temperatures of 373 K and 473 K, respectively, extracted from RBS/C data using DICADA. The data corresponding to the undamaged Ge sample are also shown for comparison. The profile in Fig. 6(a) shows a reduction in disorder by more than 50% at the damage peak maxima at highest fluence used with irradiation at 373 K. This recovery is more than 95% in the case of irradiation at 473 K, as shown in Fig. 6(b). Hence, these results corroborate the Raman observations. Complete recrystallization is marked in set A samples at the ion dose of 1 × 10¹⁴ ions cm⁻² at 473 K with the help of Raman spectroscopy (Fig. 5(c)) and RBS/C results (Fig. 6(b)).

Fig. 5 Micro-Raman spectra obtained from as-amorphized and 100 MeV Ag-irradiated set A samples in the temperature range of 100 K to 473 K. The spectrum of pristine (un-implanted) Ge is also shown for reference. The units of the indicated fluences are in ions cm⁻².

Fig. 6 The damage profile of the ion-assisted, recrystallized set A samples (extracted from DICADA) as a function of irradiation fluence at the temperatures (a) 373 K and (b) 473 K.

Fig. 7(a) and (b) show the Raman spectra of set C samples after irradiation with Ag ions at 373 K and 473 K, as no recovery of defects was observed up to room-temperature irradiation.²⁹ In the as-damaged sample, the presence of the broad band corresponding to a-Ge is observed at around 270 cm⁻¹, and the absence of a peak at 301 cm⁻¹ related to the crystalline phase emphasized the complete amorphization of the Ar ion-irradiated, near-surface layer of Ge. On irradiating with Ag ions at 373 K, the LO mode corresponding to c-Ge starts to appear at ∼301 cm⁻¹ along with the a-Ge peak. The intensity of the c-Ge peak increases with increasing irradiation dose, as shown in Fig. 7(a). This reveals the onset of recrystallization and the co-existence of amorphous and crystalline phases. Fig. 7(b) reveals that increasing irradiation temperature to 473 K results in an increase in the crystal fraction, which further increases with ion dose. At the highest dose of 1 × 10¹⁴ ions cm⁻², the c-Ge LO peak becomes sharper, with a reduction of tail in the lower wavenumber shift side showing further recrystallization of the damaged region. However, an asymmetry is still present in the LO peak related to c-Ge, which signifies the presence of defects in the irradiated region as evident from the RBC/C result.

Fig. 7 Micro-Raman spectra obtained from as-amorphized and 100 MeV Ag-irradiated set C samples at 373 K and 473 K plotted against ion fluence. The spectrum of pristine (un-implanted) Ge is also shown for reference. The units of the indicated fluences are in ions cm⁻².

Similar analysis of set B samples was performed using the RBS/C and Raman spectroscopy results, as shown in Fig. 8. The simulated disorder concentration for all set B samples using software DICADA is plotted as a function of ion dose as shown in Fig. 8(a)–(c) for different temperatures. Fig. 8(a) shows that in these samples, damage recovery is observed in the tail region only; however, there is no damage reduction at the peak maximum with irradiation at 100 K. Here, the FWHM of the damaged region is reduced from 190.4 nm to 158.6 nm after irradiation at 1 × 10¹⁴ ions cm⁻². This reveals that the damaged region present in set B samples was reduced. The reduction took place from both the surface and bulk, highlighting the recrystallization in pre-damaged samples after SHI irradiation. It is found that recrystallization is less effective at the peak-damaged region. This damage recovery increases with ion dose as corroborated by the Raman spectra shown in Fig. 8(d), which show the appearance and strengthening of c-Ge peak with fluence. With further increase in irradiation temperature to 373 K and 473 K, recrystallization of defects increases, as shown in Fig. 8(b), (c), (e) and (f). The width of the damaged profile is reduced from 190.4 nm to 136.6 nm, and the area under the damaged region is reduced by ∼50% after irradiation at 473 K at the ion fluence of 1 × 10¹⁴ ions cm⁻². At these temperatures, the LO peak for c-Ge becomes sharper, with reduction of asymmetry in the tail showing further recrystallization of the damaged region. Moreover, in set B we observed shrinking of amorphous regions from the boundaries and in the central zone as well. However, no recrystallization was observed in vacuum-annealed set B and set C samples up to 473 K, as shown in Fig. 3 and 4. This reveals that the ion-induced thermal spike assists the recrystallization of the damaged layer at much lower temperatures. However, the recrystallization is not complete and this region still constitutes defects as demonstrated by the damage profile and asymmetry in the Raman peak (at 301 cm⁻¹) at 473 K, even at the highest dose of 1 × 10¹⁴ ions cm⁻². The recrystallization is relatively rapid in terms of the Ag ion fluence and irradiation temperature for the less damaged crystals, i.e., those with dpa ∼0.25, and relatively slow for the heavily damaged region, i.e., those with disorder of dpa ∼0.5 and 7, which corresponds to threshold and above-threshold fluences of amorphization in this study.

Fig. 8 The damage profile and micro-Raman spectra as a function of irradiation dose of the ion-assisted regrown set B samples at the temperature of 100 K in (a) and (d), 373 K in (b) and (e), and 473 K in (c) and (f), respectively.

The recovery from disorder as a function of temperature was calculated for thermally annealed samples. The recrystallization rate was measured at damage peak maximum and plotted for set B and set C samples. Here, the rate of recrystallization is found to increase exponentially, with a lower rate in set C samples as compared to set B. Similarly, the recrystallization rate was calculated from the change in the amount of disorder as a function of ion dose for Ag ion-irradiated samples. From these calculations, one can calculate the epitaxial recrystallization rate between two successive ion doses ϕ_i and ϕ_i+1 for pre-damaged samples at the peak damage z. The recrystallization rate is expressed as³⁴

where D(z, ϕ_i) is the amount of disorder at depth z and dose ϕ_i. Using the above formula, the recrystallization rate was found to be increased rapidly in set A samples, and the growth rate is slower with increase in the amount of initial damage. The activation energy was also calculated from the Arrhenius plot of the recrystallized fraction with the inverse of temperature for thermally annealed set B and C samples, shown in Fig. 9(a) and (b), respectively, and for the Ag-irradiated case in Fig. 9(c). In the case of thermally annealed samples, the activation energy was found to be ∼0.165 eV and ∼0.23 eV for set B and set C samples, respectively. On the other hand, the activation energy calculated in irradiated samples was found to be reduced to ∼0.06 eV and ∼0.05 eV for set B and C samples, respectively. However, the activation energy in set A samples after irradiation was found to be ∼0.07 eV. Here, Raman results show substantial damage recovery as compared to RBS/C. This may be due to the higher sensitivity of RBS/C as compared to Raman spectroscopy for short-range disorder introduced by the presence of a high density of stacking faults and dislocations in the re-crystallized Ge region.

Fig. 9 Arrhenius plot of regrowth rates for (a) thermally annealed set B, (b) thermally annealed set C and (c) 100 MeV Ag-induced crystallization, as a function of the inverse of temperature; (d) variation of stress as a function of irradiation temperature; here the lines join the points to guide eyes.

Besides recovery of a-Ge after Ag ion irradiation, there is remnant stress in the recrystallized Ge material. This is signified by a shift in the Raman peak related to c-Ge towards the lower wavenumbers compared to single-crystal Ge, which is stress-free. It consequently reveals the appearance of tensile stress in the Ge lattice due to microstructural changes during the recrystallization process. The shifts and corresponding stress values are average values resulting from the laser scattering which occurs over the damage, distributed throughout the volume. The magnitude of this tensile stress (σ) is estimated by using the following equation from the in-plane stress model:³⁵ σ (MPa) = −250Δω (cm⁻¹), where Δω = ω_I − ω_o. In this expression, ω_o and ω_I are the Raman shift values related to the c-Ge peak of the stress-free single crystal and recrystallized Ge samples, respectively. For all three sets of samples, the stress was quantified using the above equation, and its variation with irradiation temperature is shown in Fig. 9(d). From Fig. 9(d), it can be concluded that the stress is reduced with increase in temperature, though the rate of reduction is higher for samples with higher initial damage (set C). Furthermore, the stress is not removed completely in set B and set C samples even after irradiation at 473 K. This is due to the presence of significant isolated damaged zones. In set A samples, the remnant stress is much less, which corroborates the RBS/C results showing approximately complete recrystallization.

The XTEM image shown in Fig. 10(a) reveals that after Ag ion irradiation at room temperature with an ion dose of 1 × 10¹⁴ Ag ions cm⁻², the pre-damaged layer in set C sample transformed to nanostructure-like void and nanowires. The Ag ions result in the melting of a-Ge due to thermal spike generation. Consequently, void formation took place in a-Ge material, during resolidification from the melt phase, due to the high density of Ge in the molten phase.³⁶ These voids add up to the surface, and remnant material results in nanowire structures. This phenomenon is explained in more detail in the earlier report.²⁹ The inset in Fig. 10(a) is the SAED pattern showing the amorphous nature of these nanowires. The results presented above clearly establish that in the case of samples damaged at 7 dpa, high-temperature irradiation induces recrystallization but irradiation at temperatures up to room temperature induces nanowire and void formation. To study the effect of post-irradiation annealing in these samples consisting of nanowires, micro-Raman investigations were carried out after thermal annealing. Fig. 10(b) shows the Raman spectra of annealed samples along with c-Ge. These spectra reveal that the nanowires were initially amorphous in nature, and they sustain their amorphous phase after annealing up to a temperature of 673 K, as the Raman spectra showed only a LO phonon peak related to a-Ge at ∼270 ± 2 cm⁻¹. However, the c-Ge peak emerges at the annealing temperature of 773 K, becoming sharper and stronger at 873 K, as shown in the inset of Fig. 10(b). Hence, Raman measurement emphasized that thermal annealing exhibits partial recrystallization of nanowires because of the presence of asymmetry in LO peak corresponding to c-Ge even at 873 K. The pores and voids in Ge are highly stable upon annealing; therefore, recrystallization of nanowires does not account for the alteration of the void structures. This observation also accounts for the onset of recrystallization of a-Ge material at 773 K no matter if it is in bulk form or in nanowire form. In set C sample, two distinct observations are made in the samples irradiated at elevated temperature as compared to irradiation at room temperature: (i) absence of voids and (ii) emergence of crystalline phase.

Fig. 10 (a) The high-resolution TEM image of as-prepared nanowires in a-Ge on c-Ge substrate; inset shows the SAED pattern of the nanowire sample. (b) Micro-Raman spectra of annealed nanowires plotted for the temperature range 373 K to 873 K. Inset shows the expanded Raman spectra showing recrystallization on annealing the nanowires.

Discussion

We have observed that irradiation of c-Ge by 100 MeV Ag produces negligible damage that scales with nuclear energy loss (∼0.1 keV nm⁻¹), whereas very high Se (∼160 × S_n) is insensitive to damage formation even at high irradiation ion fluences of up to 10¹⁴ cm⁻². However, the irradiation of damaged Ge with the same ion leads to remarkable structural modifications. The presented results show that these modifications are also very sensitive to the sample temperature at which 100 MeV Ag ion irradiation is performed. During electronic energy loss, interaction of the incident ion with free and bound electrons in a solid leads to the formation of energetic secondary electrons. These electrons are confined within a narrow cylindrical target zone around the ion path called the ion track. The confinement of these electrons depends on electron diffusion length and thus modifies electron-phonon coupling (g). The coupling term ‘g’ governs the efficiency with which energy deposited in the electronic subsystem is subsequently transferred to the lattice sub-system per unit volume and the time to increase lattice temperature. The two-temperature model can be used to describe the rise in temperature of the lattice. The description of the two-temperature model is based on a set of two coupled heat diffusion equations. One of the equations is for the electronic system, and another is for the phonon system, which allows one to estimate the peak temperature along the ion track in cylindrical geometry.³⁷Here, r is the radial distance from the path of ion or track radius. T_e,a, C_e,a, and K_e,a are temperature, specific heat and thermal conductivity for the electronic and atomic subsystems, respectively. All these parameters and their detailed descriptions are given in our previous report.²⁹ Numerical solution of these coupled differential equations give time evolution of temperature along the ion track, which is plotted in Fig. 11, at zero distance from ion path and at two irradiation temperatures, 300 K and 473 K. The peak track temperature depends on electron-phonon coupling strength, i.e. ‘g’, which is estimated to be 1.2 × 10¹² W K⁻¹ cm⁻³ and 1 × 10¹¹ W K⁻¹ cm⁻³ for amorphous Ge and c-Ge, respectively.²⁹ Consequently, this leads the ion track temperature to reach above melting point in a-Ge, whereas it is less than the melting temperature in c-Ge. This may explain the insensitivity of 100 MeV Ag ion irradiation in c-Ge towards damage formation and extensive damage formation in a-Ge as shown in Fig. 10(a) for set C samples irradiated at 300 K.

Fig. 11 The thermal spike calculations of 100 MeV Ag ions in a-Ge; the time evolution of temperature is presented at the centre of the track (i.e. at r = 0). Inset shows the expanded view of region II.

The results show molten track formation due to thermal spike generation in a-Ge in both cases. This is understandable from the time evolution of temperature in a-Ge, extracted from thermal spike calculations, as shown in Fig. 11. This figure clearly shows that the material remains in liquid phase for a time of 2 picoseconds in both the cases of 300 K and 473 K irradiation. The liquid Ge has a diffusivity of order 10⁻⁸ m² s⁻¹ as reported in simulation studies.³⁸ During resolidification from melt phase, i.e. in region 2 in Fig. 11, void formation took place in a-Ge during room temperature irradiation, due to the high density of Ge in molten phase.³⁶ However, resolidification in this region II results in recrystallization of the amorphous material instead of void formation when the substrate temperature was higher, i.e., 473 K. This may be attributed to the quenching rate, which is given by ΔT/Δt, where ΔT is temperature difference and t is corresponding time. So when the substrate itself is at, say, 473 K, then quenching will be slower due to less temperature difference between the spike and substrate such that (ΔT/Δt)_{473 K} < (ΔT/Δt)_{300 K}. Consequently, the vacancies may get more time to diffuse, which inhibits their agglomeration to form void structures. This scenario can be understood with the help of the inset in Fig. 11, which is the expanded view of region II in Fig. 11. It indicates the slower resolidification rate for 473 K as compared to 300 K and, consequently, recrystallization due to diffusion of vacancies before their combination to form voids.

In partially damaged Ge, where pockets of amorphous Ge are surrounded by c-Ge, irradiation by 100 MeV Ag ions leads to recrystallization at room temperature for set A sample and for set B sample at higher irradiation temperature. This may be due to the synergic effect of both nuclear energy loss and electronic energy loss processes, where S_n efficiently produces interstitial vacancy pairs at the crystal–amorphous (c–a) interface. Therefore, only those defects generated directly at the c–a interface or near it are available for the recrystallization process. So the Ag ion-induced additional vacancies reach the interface and could help to enhance the regrowth process. The number of excess vacancies created by Ag ions is ∼10¹⁸ cm⁻³ as calculated using SRIM.³² While the vacancies/defects are produced all along the ion track, all of them cannot reach the interface and participate in the recrystallization process. Werner et al.³⁹ investigated the effect of various parameters on self-diffusion in Ge and determined that, under equilibrium conditions, vacancies mediate self-diffusion in Ge. Under thermal equilibrium, the existence of interstitials had not been evidenced. The reason behind this might be related to the higher energy of formation of interstitials as compared to vacancies.⁴⁰ However, interstitials participate extensively in self-diffusion under non-equilibrium condition.⁴¹

Conclusion

Three sets of samples consisting of different degrees of damage introduced by sub-threshold (set A), threshold (set B) and above-threshold (set C) doses of amorphization using 100 keV Ar ions were used here. RBS/C, Raman and XTEM analysis of set A, B and C samples after thermal and athermal treatment showed that enhanced recrystallization of the damaged region takes place after irradiation with 100 MeV Ag ions. However, the rate of relative recovery depends on initial disorder level, annealing temperature and ion dose. Hence, a relatively strong annealing effect is observed for the less-damaged set A samples. Considerable damage recovery is observed after Ag ion irradiation even at 100 K, while at an irradiation temperature of 473 K, the Ar ion-induced damage is almost fully healed, and the ordered atomic structure is confirmed. For the high-disorder sample set B with 0.5 dpa, a higher irradiation temperature is needed to completely repair the pre-existing damage. Substantial recovery is observed under the Ag irradiation when the temperature increases from 100 K to 473 K, with the disorder levels dropping to ∼0.75. Additional Ag irradiation is required to fully heal the damaged crystalline structure. Remarkably different results are observed in set C samples, having completely amorphous layer on c-Ge substrate, as 100 MeV Ag ion irradiation formed nanowires in a-Ge when irradiated at 100 K and 300 K, but there is recrystallization when irradiated at ∼500 K and above.

Acknowledgements

Ms. Sonu Hooda is thankful to the Council of Scientific and Industrial Research (CSIR), India, for financial support through fellowship and Prof. Toulemonde for providing thermal spike code. We would like to thank Dr Fouran Singh and Mr Subodh Gautam for help during Raman spectroscopy measurements. Help received from Dr Parvin Kumar and Mr Kedarmal during the ion irradiation experiment is gratefully acknowledged.

References

1. C. Claeys and E. Simoen, Germanium-based technologies: from materials to devices, access online via, Elsevier, 2011.
2. E. Simoen, J. Mitard, G. Hellings, G. Eneman, B. de Jaeger, L. Witters, B. Vincent, R. Loo, A. Delabie, S. Sioncke, M. Caymax and C. Claeys, Mater. Sci. Semicond. Process., 2012, 15(6), 588–600.
3. Y. G. Yoon, T. K. Kim, I.-C. Hwang, H.-S. Lee, B.-W. Hwang, J.-M. Moon, Y.-J. Seo, S. W. Lee, M.-H. Jo and S.-H. Lee, ACS Appl. Mater. Interfaces, 2014, 6(5), 3150–3155.
4. L.-H. Zeng, M.-Z. Wang, H. Hu, B. Nie, Y.-Q. Yu, C.-Y. Wu, L. Wang, J.-G. Hu, C. Xie, F.-X. Liang and L.-B. Luo, ACS Appl. Mater. Interfaces, 2013, 5(19), 9362–9366.
5. C. Yan, N. Singh, H. Cai, C. L. Gan and P. S. Lee, ACS Appl. Mater. Interfaces, 2010, 2(7), 1794–1797.
6. F.-F. Ren, K.-W. Ang, J. Ye, M. Yu, G.-Q. Lo and D.-L. Kwong, Nano Lett., 2011, 11(3), 1289–1293.
7. S. Assefa, F. Xia and Y. A. Vlasov, Nature, 2010, 464(7285), 80–84.
8. R. Jones, S. Thomas, S. Bharatan, R. Thoma, C. Jasper, T. Zirkle, N. Edwards, R. Liu, X. Wang and Q. Xie, presented at the IEDm: international electron devices meeting, 2002, unpublished.
9. M. K. Hudait, M. Clavel, Y. Zhu, P. S. Goley, S. Kundu, D. Maurya and S. Priya, ACS Appl. Mater. Interfaces, 2015, 7(9), 5471–5479.
10. T. Kennedy, E. Mullane, H. Geaney, M. Osiak, C. O'Dwyer and K. M. Ryan, Nano Lett., 2014, 14(2), 716–723.
11. X. Li, Z. Yang, Y. Fu, L. Qiao, D. Li, H. Yue and D. He, ACS Nano, 2015, 9(2), 1858–1867.
12. S.-K. Kang, G. Park, K. Kim, S.-W. Hwang, H. Cheng, J. Shin, S. Chung, M. Kim, L. Yin, J. C. Lee, K.-M. Lee and J. A. Rogers, ACS Appl. Mater. Interfaces, 2015, 7(17), 9297–9305.
13. R. A. Kelly, J. D. Holmes and N. Petkov, Nanoscale, 2014, 6(21), 12890–12897.
14. W. Jiang, H. Wang, I. Kim, I. T. Bae, G. Li, P. Nachimuthu, Z. Zhu, Y. Zhang and W. J. Weber, Phys. Rev. B: Condens. Matter Mater. Phys., 2009, 80(16), 161301.
15. J. F. Gibbons, Proc. IEEE, 1972, 60(9), 1062–1096.
16. J. Nakata, J. Appl. Phys., 1996, 79(2), 682–698.
17. L. Pelaz, L. A. Marqués and J. Barbolla, J. Appl. Phys., 2004, 96(11), 5947–5976.
18. P. K. Sahoo, T. Som, D. Kanjilal and V. N. Kulkarni, Nucl. Instrum. Methods Phys. Res., Sect. B, 2005, 240(1–2), 239–244.
19. D. Sigurd, G. Fladda, L. Eriksson and K. Björkqvist, Radiat. Eff., 1970, 3(2), 145–147.
20. K. Gärtner, J. Jöhrens, T. Steinbach, C. S. Schnohr, M. C. Ridgway and W. Wesch, Phys. Rev. B: Condens. Matter Mater. Phys., 2011, 83(22), 224106.
21. G. Impellizzeri, E. Napolitani, S. Boninelli, G. Fisicaro, M. Cuscuna, R. Milazzo, A. La Magna, G. Fortunato, F. Priolo and V. Privitera, J. Appl. Phys., 2013, 113(11), 113505–113507.
22. G. S. Armatas and M. G. Kanatzidis, Nano Lett., 2010, 10(9), 3330–3336.
23. T. Steinbach, J. Wernecke, P. Kluth, M. C. Ridgway and W. Wesch, Phys. Rev. B: Condens. Matter Mater. Phys., 2011, 84(10), 104108.
24. S. Decoster and A. Vantomme, J. Phys. D: Appl. Phys., 2009, 42(16), 165404.
25. P. N. Okholin, V. I. Glotov, A. N. Nazarov, V. O. Yuchymchuk, V. P. Kladko, S. B. Kryvyi, P. M. Lytvyn, S. I. Tiagulskyi, V. S. Lysenko, M. Shayesteh and R. Duffy, Mater. Sci. Semicond. Process., 2016, 42, 204–209.
26. B. L. Darby, B. R. Yates, N. G. Rudawski, K. S. Jones and A. Kontos, Nucl. Instrum. Methods Phys. Res., Sect. B, 2011, 269(1), 20–22.
27. C. Wündisch, M. Posselt, B. Schmidt, V. Heera, T. Schumann, A. Mücklich, R. Grötzschel, W. Skorupa, T. Clarysse, E. Simoen and H. Hortenbach, Appl. Phys. Lett., 2009, 95(25), 252107.
28. J. Bourgoin and J. Corbett, Radiat. Eff., 1978, 36(3–4), 157–188.
29. S. Hooda, B. Satpati, S. Ojha, T. Kumar, D. Kanjilal and D. Kabiraj, Mater. Res. Express, 2015, 2(4), 045903.
30. T. Som, J. Ghatak, O. P. Sinha, R. Sivakumar and D. Kanjilal, J. Appl. Phys., 2008, 103(12), 123532.
31. M. Toulemonde, C. Dufour and E. Paumier, Phys. Rev. B: Condens. Matter Mater. Phys., 1992, 46(22), 14362–14369.
32. J. F. Ziegler, M. D. Ziegler and J. P. Biersack, Nucl. Instrum. Methods Phys. Res., Sect. B, 2010, 268(11–12), 1818–1823.
33. K. Gärtner, Nucl. Instrum. Methods Phys. Res., Sect. B, 1997, 132(1), 147–158.
34. A. Benyagoub and A. Audren, J. Appl. Phys., 2009, 106(8), 083516.
35. V. Paillard, P. Puech, M. A. Laguna, P. Temple-Boyer, B. Caussat, J. P. Couderc and B. de Mauduit, Appl. Phys. Lett., 1998, 73(12), 1718–1720.
36. M. C. Ridgway, T. Bierschenk, R. Giulian, B. Afra, M. D. Rodriguez, L. L. Araujo, A. P. Byrne, N. Kirby, O. H. Pakarinen, F. Djurabekova, K. Nordlund, M. Schleberger, O. Osmani, N. Medvedev, B. Rethfeld and P. Kluth, Phys. Rev. Lett., 2013, 110(24), 245502.
37. S. A. Khan, S. K. Srivastava and D. K. Avasthi, J. Phys. D: Appl. Phys., 2012, 45(37), 375304.
38. V. Hugouvieux, E. Farhi, M. R. Johnson, F. Juranyi, P. Bourges and W. Kob, Phys. Rev. B: Condens. Matter Mater. Phys., 2007, 75(10), 104208.
39. M. Werner, H. Mehrer and H. D. Hochheimer, Phys. Rev. B: Condens. Matter Mater. Phys., 1985, 32(6), 3930–3937.
40. H. Haesslein, R. Sielemann and C. Zistl, Phys. Rev. Lett., 1998, 80(12), 2626–2629.
41. A. Chroneos and H. Bracht, Appl. Phys. Rev., 2014, 1(1), 011301.

---