Nanopore/pillar formation on a Ge substrate can be induced by ion irradiation, which activates the ion beam sputtering and self-organization of point defects. Considering that the size and morphology of nanostructures are dependent on damage production, the irradiation parameters significantly affect nanostructuring. Here, the projected range of incident ions was selected as a parameter to be investigated. The projected range was modified by adding an Au buffer layer on the surface of the substrate, enabling the ions to stop in a shallower layer. The experimental results showed that the deposited Au layer affected the size and morphology of the nanostructures produced by ion irradiation. As a unique morphology, network-like structures were observed on the Au-deposited substrates. These structures were larger than ordinary porous structures.

The use of ion beam technologies to implant ions into a substrate material is indispensable in the manufacture of semiconductors.1–3 Such implanted ions can activate the electronic conductivity of semiconductors, which can be used as components in electronic devices. In addition, ion beam technologies can be employed in the nanofabrication process because of their nanoscale precision.4–9 Ion beam techniques can be employed in the fabrication of nanoholes,5 nanowires,7 and even photonic crystals.6 Thus, these techniques are valuable for the surface nanofabrication process; thus, they have been widely studied.

Surface structures with sizes ranging from several micrometers to nanometers can be fabricated via ion beam irradiation. Certain nanostructuring processes are induced by sputtering and surface diffusion,10–12 whereas others are induced by the migration and aggregation of point defects.13,14 Typically, an ion with a keV-order energy dissipates most of its energy using nuclear stopping power, resulting in a collision cascade.15,16 In a collision cascade, an ion collides with target atoms depositing kinetic energy. When the deposited energy is sufficient for displacements, Frenkel pairs are produced. Some of these displaced atoms in near-surface layers recoil back to a surface, where they overcome the surface binding energy and leave. This phenomenon is called ion beam sputtering,17,18 and it can lead to the formation of nanoholes or trenches on the solid surface. Such irregularities in the surface morphology are major factors in the formation of nanoripples, which are formed when the ion incidence angle is oblique.19–23 This type of nanostructuring is demonstrated by ion beam sputtering and the relaxation process via atomic diffusion.24,25 Ion beam sputtering is possible in insulators, semiconductors, and even metals; thus, it can be used for various materials. Hence, this nanostructuring is also possible in Ge, resulting in the formation of nanodots and nanoripples. Actually, Ziberi et al. reported nanostructuring induced by the 2 keV-Xe+ irradiation on Ge, revealing that the morphology of nanostructure depends on the incident angle. The morphology was significantly sensitive to the change in the incident angle, and an increase of 5° caused the transition from nanodots to nanoripples.26 

Frenkel pairs that are not involved in sputtering will migrate and aggregate. Although a considerable amount of Frenkel pairs can be lost by sinks or recombinations, certain vacancies can form voids. As the irradiation proceeds, these voids will be exposed on the surface and turn the surface morphology into an abundantly porous morphology. This is the process of nanostructure formation by the self-organization of point defects. It requires the growth of dominant voids; therefore, the lifetime of vacancies should be sufficiently long such that they can migrate and coalesce prior to annihilation. Thus, materials, such as Ge27,28 and GaSb,29,30 with long vacancy lifetimes, are suitable as substrates. In the previous research by Romano et al.,27 the change in the structure size of Ge was investigated. The authors conducted the 300 keV-Ge+ implantation and revealed that the structure size increases with an increasing fluence. The research by Rudawski et al. is also notable,31 because it enhanced the potential of the nanostructure, indicating that nanostructuring can be applied to the fabrication of Li ion battery anodes. Furthermore, it has been reported that the morphology of nanostructures on Ge substrates is dependent on the projected ion range because it determines when voids appear on the surface.32 The deeper the voids, the longer it takes for them to appear on the surface. Thus, the morphology can be complicated, with significant void coalescence. During the experiment, the projected range can be changed by modifying the ion energy or applying a buffer layer onto the surface. Here, the projected range was adjusted via Au deposition on the substrates because the energy dependence of the morphology of such porous structures is already known from a previous study.32 When sufficiently thick, the deposited Au films can function as a buffer layer for subsequent ion irradiation, and the projected range in Ge can decrease. Alkhaldi et al. revealed that a SiO2 cap suppressed nanostructuring in Ge, particularly at a low temperature.33 In addition, they observed that the depth of nucleation of the pores deepened when the cap was applied to the surface. This implies that the film deposited on the surface can affect the nanostructuring process in Ge, thereby affecting the morphology of the nanostructure. Compared with previous research, this study investigated the morphological changes in nanostructures fabricated through lower-energy ion irradiation. The ion range reduces at a low incident energy; thus, the cap-induced decrease in the ion range is expected to be critical at a lower incident energy, resulting in a considerably shallow range of damage. This can affect the nanostructuring process significantly, resulting in the formation of a nanostructure barely formed by ordinary ion irradiation. It is expected that such a unique fabrication process will enhance the choice of morphology, thereby enhancing the potential of the nanostructure. Based on the obtained results, the effect of the change in the ion range is discussed herein.

Single crystalline Ge substrates mirror polished at a (001) plane to where Au deposition and ion irradiation were performed were used. First, a fine coater (JEOL JFC-1200) was used to deposit Au films on the substrates. The deposition was performed under low vacuum conditions with no gas flow. The purity of the target Au was 99.99%, which equaled that of the Ge substrates. Ion irradiation was performed using a focused ion beam (FIB; FEI Quanta 3D 200i), comprising Ga+ accelerated up to 5 and 30 keV. Note that Ga is the only ion source that is available, although nanostructuring in Ge is possible with various ion species.27,32 The fluence and beam current were 5 × 1019–2 × 1021 ions/m2 and 0.5 nA, respectively. The incident angle to the surface normal was set within a range of 0°–60°, and the dwell time was in the range of 1.99–42.9 μs. Using these parameters, the irradiation yielded 8.2–180 s per irradiation session. The total scanned area was 10.8 × 12.5 μm. All the irradiated samples were observed using a scanning electron microscope (Hitachi SU8020) in the SE detection mode and captured from the normal direction to the surface. The accelerating voltage and current were 5 kV and ∼20 μA, respectively. The scanning electron microscope was also used to perform energy-dispersive x-ray spectroscopy (EDX) measurements with an x-ray detector (Horiba X-Max80 011). The mechanism of the nanostructuring is subsequently described in the section titled Discussion, using data from the Stopping and Range of Ion in Matter (SRIM) simulation.34 The SRIM calculations were performed with “Detailed Calculation with full Damage Cascades,” and the displacement threshold energies for Ge and Au were suggested by SRIM as 15 and 25 eV, respectively.35 The elliptic Fourier analysis (EFA)36,37 was employed to evaluate the morphology of nanostructures fabricated by 5 keV ion irradiation. This was conducted using the Python package of “ktch-0.3.1.”38 To visualize the EFA results, the Fourier coefficients were processed through principal component analysis (PCA), using the Python library of “scikit-learn1.2.”39 To estimate the diameter of nanoholes, some of them were segmented through artificial intelligence in OVSeg40 with the model, “Segment Anything,” provided by Meta AI.41 Afterward, the segmented nanoholes were detected using the Python library of OpenCV,42 enabling the measurements of the size of every bounding rectangle. The nanohole diameter was the average of the horizontal and vertical sides of the bounding rectangle.

The first experiment investigated nanostructuring induced by a 5 keV normal ion incidence on Au-deposited Ge substrates. The fluences were 1 × 1021 or 2 × 1021 ions/m2, and the deposition time was in the range of 0–40 s. Figure 1 shows the scanning electron microscopy (SEM) images of the unirradiated surfaces. The unirradiated Au surface was mostly smooth, although slight roughness with certain droplets was observed at a longer deposition time.

FIG. 1.

SEM images of the unirradiated surface. The deposited surfaces were mostly smooth, although slight roughness was observed at a longer deposition time. Droplets were also formed at a longer deposition time (b)–(d).

FIG. 1.

SEM images of the unirradiated surface. The deposited surfaces were mostly smooth, although slight roughness was observed at a longer deposition time. Droplets were also formed at a longer deposition time (b)–(d).

Close modal

Figure 2 shows the resulting nanostructured samples. A fluence of 1 × 1021 ions/m2 produced abundant nanoholes on all the irradiated surfaces, and the pattern became disordered with an increase in the deposition time, resulting in surface irregularities. Contrarily, when the fluence was 2 × 1021 ions/m2, the nanoholes were as uniform as those on the undeposited substrates. Even at a deposition of 40 s, the nanoholes were similar to those obtained using undeposited substrates.

FIG. 2.

Scanning electron microscopic images of nanoholes fabricated with a 5 keV normal ion incidence. The deposition times and fluences were 0–40 s and 1 × 1021–2 × 1021 ions/m2, respectively. At a fluence of 1 × 1021 ions/m2, the pattern became disordered at a longer deposition time (a)–(e). This tendency was not observed at a fluence of 2 × 1021 ions/m2, producing uniform nanoholes on all the samples (f)–(j).

FIG. 2.

Scanning electron microscopic images of nanoholes fabricated with a 5 keV normal ion incidence. The deposition times and fluences were 0–40 s and 1 × 1021–2 × 1021 ions/m2, respectively. At a fluence of 1 × 1021 ions/m2, the pattern became disordered at a longer deposition time (a)–(e). This tendency was not observed at a fluence of 2 × 1021 ions/m2, producing uniform nanoholes on all the samples (f)–(j).

Close modal

The average diameters of the nanoholes were estimated by image processing, as described in the section titled Experimental Procedure. Figure 3 shows these diameters. At a deposition time of 0 s, the diameters were 26.1 and 24.9 nm at fluences of 1 × 1021 and 2 × 1021 ions/m2, respectively. At a deposition time of 40 s, the values were 44.1 and 42.1 nm, respectively. Although slight dispersion was observed, the average diameter was ∼40 nm under all the experimental conditions.

FIG. 3.

Average diameter of nanoholes shown in Fig. 2 (∼40 nm under all experimental conditions).

FIG. 3.

Average diameter of nanoholes shown in Fig. 2 (∼40 nm under all experimental conditions).

Close modal

Figure 4 shows nanostructuring induced by a 5 keV oblique ion incidence on Au-deposited Ge substrates at an incident angle of 20°. The fluences were 1 × 1021 or 2 × 1021 ions/m2, and the deposition time range was 0–40 s. Although these scanning electron microscopy (SEM) images exhibited nanoholes on the surface, the nanoholes were more oriented against the incident direction, forming obscure ripple patterns. The nanoholes produced at a fluence of 1 × 1021 ions/m2 became nonuniform and disordered, with an increase in the deposition time. Contrarily, as shown in Fig. 2, the nanoholes produced at a fluence of 2 × 1021 ions/m2 were uniform and well-ordered, even at a deposition time of 40 s. The size of these nanoholes could not be estimated using the aforementioned method because of the insufficient precision of the segmentation task. Compared with those in Fig. 2, these nanoholes exhibited an elongated shape, and this may have led to incomplete segmentation.

FIG. 4.

SEM images of nanoholes formed with a 5 keV oblique ion incidence at an incident angle of 20°. The fluences were 1 × 1021 or 2 × 1021 ions/m2, and the deposition time range was 0–40 s. The nanoholes were more disordered at a lower fluence (a)–(e).

FIG. 4.

SEM images of nanoholes formed with a 5 keV oblique ion incidence at an incident angle of 20°. The fluences were 1 × 1021 or 2 × 1021 ions/m2, and the deposition time range was 0–40 s. The nanoholes were more disordered at a lower fluence (a)–(e).

Close modal

The next experiment investigated nanostructuring induced by a 30 keV-Ga+ irradiation on Ge substrates deposited for 0–40 s (Fig. 5). The fluence and angle of incidence were 5 × 1019 to 1 × 1021 ions/m2 and 0°, respectively. The undeposited surface exhibited a porous structure after irradiation at a fluence of 5 × 1019 ions/m2. When the fluence was increased to 1 × 1020 ions/m2, these pores enlarged. At a fluence of 5 × 1020 ions/m2, the surface morphology changed again to form a complicated pattern that comprised deformed sidewalls of pores. This morphology was stable up to a fluence of 1 × 1021 ions/m2. At a deposition time of 10 s, a network-like structure was formed at a fluence of 5 × 1019 ions/m2. Although the structure size was practically equal to that of an undeposited surface at the same fluence, the morphology differed. This network structure became enlarged at a fluence of 1 × 1020 ions/m2. A similar result was obtained at a deposition time of 20 s. Notably, the network structure was larger than that formed on the 10 s deposited substrate. At deposition times of 30 or 40 s, nanoholes and nanodots were observed at a fluence of 5 × 1019 ions/m2. Although a portion of these nanodots could have originated from droplets on the unirradiated surface, they were larger than the droplets. This implies that the nanodots in the SEM images can be formed from the deposited Au, which turns into nanoparticles. The morphology was obscure at a fluence of 1 × 1020 ions/m2, although it indicated surface roughness. Throughout all the deposition times, the morphology was alike at or above a fluence of 5 × 1020 ions/m2. This suggested that the Au effect was negligible at or above a fluence of 5 × 1020 ions/m2. According to the EDX measurements, the Au concentration in the irradiated cases shown in Fig. 5 was ∼1 at. % or less. Considering that an EDX measurement cannot quantify an element of ∼1 at. %, the amount of Au here could not be quantified. However, the intensity peaks indicating the presence of Au were detected, implying that a small amount of Au remained on the surface [Fig. S1(a) in the supplementary material]. Note that the intensity value itself depends on the surface morphology, and it is unreliable to compare the intensity value of every case.

FIG. 5.

SEM images of nanostructures formed with a normal ion incidence on the Au-deposited Ge substrates. The deposition times and fluences were 0–40 s and 5 × 1019–1 × 1021 ions/m2, respectively. The surface morphology formed on the deposited samples differed from that on undeposited samples (a)–(j). The effect of the Au deposition was significant at or below a fluence of 1 × 1020 ions/m2.

FIG. 5.

SEM images of nanostructures formed with a normal ion incidence on the Au-deposited Ge substrates. The deposition times and fluences were 0–40 s and 5 × 1019–1 × 1021 ions/m2, respectively. The surface morphology formed on the deposited samples differed from that on undeposited samples (a)–(j). The effect of the Au deposition was significant at or below a fluence of 1 × 1020 ions/m2.

Close modal

Figure 6 shows nanostructuring induced by an oblique ion incidence on a Ge substrate within a deposition time range of 0–40 s. The fluence and incident angle were 5 × 1019 to 1 × 1021 ions/m2 and 60°, respectively. The pillars were formed at a fluence of 5 × 1019 ions/m2, although their size and shape were dependent on the deposition time. At a deposition time of 10 s, the lengths of the pillars were longer than those without Au deposition. At a deposition time of 20 s, the pillars were wider than those of 0 and 10 s. At deposition times of 30 or 40 s, the patterns became obscure and practically no pillars were observed. This tendency was also observed at a fluence of 1 × 1020 ions/m2, although these pillars were more elongated and deformed. The aforementioned pillars became disordered and elongated at a fluence above 1 × 1020 ions/m2, forming complicated patterns. These complicated patterns were observed on all samples at a fluence above 1 × 1020 ions/m2. Thus, the effect of the deposited film appeared negligible when the fluence was at or above 5 × 1020 ions/m2, even in the case of oblique ion incidence. The Au concentration could not be quantified in the case shown in Fig. 6 because, as previously described, the amount of Au was insufficient to be detected. However, the intensity peaks indicated remaining Au, even after irradiation. Thus, a small amount of Au could have been present on the nanostructured surface [Fig. S1(b) in the supplementary material].

FIG. 6.

SEM images of nanostructures formed with an oblique ion incidence on Au-deposited Ge substrates. The fluence and incident angle were 5 × 1019–1 × 1021 ions/m2 and 60°, respectively. The deposition time range was 0–40 s. The pillars were obtained at a fluence of 5 × 1019 ions/m2 (a)–(e), although their size and morphology were dependent on the deposition time. The morphology became more complicated at a fluence of 5 × 1020 ions/m2, and it was unaffected by an increase in the fluence (k)–(l).

FIG. 6.

SEM images of nanostructures formed with an oblique ion incidence on Au-deposited Ge substrates. The fluence and incident angle were 5 × 1019–1 × 1021 ions/m2 and 60°, respectively. The deposition time range was 0–40 s. The pillars were obtained at a fluence of 5 × 1019 ions/m2 (a)–(e), although their size and morphology were dependent on the deposition time. The morphology became more complicated at a fluence of 5 × 1020 ions/m2, and it was unaffected by an increase in the fluence (k)–(l).

Close modal

Figures 2 and 4 show nanostructuring with lower-energy ion irradiation on the Au-deposited substrates. As reported by Wei et al.,43 the Ge nanostructures fabricated by 5 keV-Ga+ could be explained using the damped Kuramoto–Sivashinsky model,44–46 suggesting that the mechanism of nanostructuring essentially lies in ion beam sputtering. Our previous study showed that the morphology of these nanoholes was insensitive to fluence.47 Even at a fluence of 1 × 1022 ions/m2, the morphology remained the same as that observed at 1 × 1021 ions/m2. Thus, it was expected that the morphologies shown in Figs. 2 and 4 did not significantly change with an increase in the fluence. However, the irregularities observed in the nanostructured region decreased with an increase in the fluence. At a fluence of 1 × 1021 ions/m2, the number of irregularities increased with an increase in the deposition time. Contrarily, these irregularities were not observed at 2 × 1021 ions/m2, even at a deposition time of 40 s. This implies that the Au films protected the Ge surface from ion beam sputtering, leading to incomplete nanostructuring. The SRIM simulation estimated the sputtering yield of Ge to be less than 0.1 under all the irradiation conditions, implying that the sputtering of Ge atoms hardly occurred at the early stage of irradiation [Fig. S2(b) in the supplementary material]. The sputtered Au atoms naturally increase with an increase in fluence; thus, the films were likely to be removed at a higher fluence. Thus, at a higher fluence, the films can be sufficiently removed, producing well-ordered nanoholes without irregularities. Furthermore, the preferential sputtering such as that described by Bradley and Shipman can occur in the binary system, in which the sputtering yield of each element differs.48 In this case, the element sputtered more easily can be removed faster than the other, forming the regular pattern on the surface. This leads to the formation of a well-ordered morphology like nanodots on GaSb substrates.10 In contrast, the results in the present study showed the Au deposition led to the formation of irregularities on the surface; hence, the preferential sputtering seems negligible here.

To evaluate the disorder induced by the Au film, 10 nanoholes in every nanostructured region in Figs. 2 and 4 were randomly selected for the EFA.36,37 The selected images of nanoholes were binarized by contrast, and their outlines were extracted by determining 100 points on the outlines. This provided the periodic functions used for the EFA, reflecting the morphology of nanoholes. Based on these periodic functions, the Fourier series is described as follows:
x ( t ) = a 0 2 + n = 1 N ( a n c o s 2 n π t T + b n s i n 2 n π t T ) ,
(1)
y ( t ) = c 0 2 + n = 1 N ( c n c o s 2 n π t T + d n s i n 2 n π t T ) .
(2)

Thus, the Fourier coefficients of EFA were obtained from each nanohole. These coefficients were processed by PCA. Figures 7(a) and7(b) show the PCA results for the incident angles of 0° and 20°, respectively. t and F in the legend represent the deposition time and fluence, respectively. The gray outlines on the graphs are the imitative images of nanoholes, reconstructed from the Fourier coefficients. PC1 was the most dominant component, followed by PC2 (Fig. S3 in the supplementary material). Thus, the axes were chosen along PC1 and PC2. Although the dispersion of plots in b appeared slightly greater than that of a, both graphs indicated that most of the plots were located around the center. This implied that most of the nanoholes possessed circular or ellipsoid outlines, as indicated by the reconstructed image. This also implied that almost no change in the outlines of individual nanoholes was induced, although certain irregularities were observed at a longer deposition time. Thus, the disordered appearance of the nanostructure appeared to originate from the irregularities in the entire region, not in the outlines of individual nanoholes.

FIG. 7.

PCA results. (a) and (b) correspond to the results for the incident angles of 0° and 20°, respectively. t and F represent the deposition time and fluence, respectively. Most of the plots were located around the centers, corresponding to circular or ellipsoid outlines.

FIG. 7.

PCA results. (a) and (b) correspond to the results for the incident angles of 0° and 20°, respectively. t and F represent the deposition time and fluence, respectively. Most of the plots were located around the centers, corresponding to circular or ellipsoid outlines.

Close modal

For the nanostructuring induced by a 30 keV-Ga+ irradiation, the effect of the deposited Au film was significant at or below a fluence of 5 × 1020 ions/m2. This can be caused by ion beam sputtering, which removes the film from the surface. As previously mentioned, the films were likely to be removed at a higher fluence. To measure the thickness of the Au films, two deposited substrates were cleaved to observe their cross sections (Fig. 8). The average thicknesses were 18.4 and 33.4 nm at deposition times of 80 and 120 s, respectively. This suggested that the maximum thickness of the film in this study was ∼10 nm or less because the maximum deposition time was 40 s.

FIG. 8.

Cross-sectional SEM image of the substrates deposited for 80 (a) and 120 s (b) with thicknesses of 18.4 and 33.4 nm, respectively. Layers comprising Au were observed on top of the substrates.

FIG. 8.

Cross-sectional SEM image of the substrates deposited for 80 (a) and 120 s (b) with thicknesses of 18.4 and 33.4 nm, respectively. Layers comprising Au were observed on top of the substrates.

Close modal

Figure 9 shows the total number and depth profile of vacancies produced using a 30 keV-Ga+, which were estimated through the SRIM simulation. Each plot symbol in b represents the difference in the film thickness. As estimated in Fig. 9(a), a Ga can penetrate through a 10 nm film, producing ∼300 vacancies in Ge. This implies that damage production in the Ge substrate is possible during the initial stage of irradiation. Simultaneously, the erosion on the Au surface proceeded, and the thickness of the Au film decreased. This enabled the penetration of ions through the deposited Au layer, inducing damage in Ge. Thus, the present nanostructuring can begin under a deposited film, resulting in the formation of unique nanostructures.

FIG. 9.

Total number and depth profile of vacancies produced using a 30 keV-Ga+ predicted by the SRIM simulation. The total number of Ge vacancies decreased with an increase in the film thickness (a). The profile peaks also decayed with an increase in the film thickness (b).

FIG. 9.

Total number and depth profile of vacancies produced using a 30 keV-Ga+ predicted by the SRIM simulation. The total number of Ge vacancies decreased with an increase in the film thickness (a). The profile peaks also decayed with an increase in the film thickness (b).

Close modal

The profile peaks shown in Fig. 9(b) decayed with an increase in the film thickness. This implied that a thicker film could absorb most of the ion energy, thereby reducing the damage production. The ion range and total number of Ge vacancies decreased with an increase in the film thickness. The SRIM simulation estimated the ion ranges of 30 keV-Ga+ to be ∼18 and ∼8 nm at film thicknesses of 0 and 10 nm, respectively. This suggests that the buffer effect by a film can result in a decrease of ∼10 nm in the ion range at maximum in this study. If the ion range affected the nanostructuring, then the complicated morphologies, such as the networks shown in Fig. 5, were produced within a shorter ion range.

In addition, the previous study by Böttger et al. explained how the ion range affects nanostructuring.32 According to the authors, the complicated morphology of sponges can be formed with a high ion energy, which embeds the voids deeper in the substrates, delaying their exposure to the surface. This promoted the growth and coalescence of voids, which form the complicated cavity, leading to a complicated morphology. Thus, nanostructuring in an Au-deposited substrate should produce a primitive morphology because the Au film reduces the ion range. Considering these facts, if we speculate that the remaining film delays the void exposure to the surface, a reasonable scenario can be proposed. When the film was sufficiently thin, the ions could penetrate through it, producing voids in the Ge substrate. These voids enlarged, coalesced in the substrate, and turned the surface morphology into a porous morphology when exposed to the surface. The remaining film protected the Ge surface during this process; thus, the surface was not eroded, unless the film was completely removed. Without the erosion of the Ge surface, the exposure of the voids could be delayed. This provides the voids with sufficient time to grow and coalesce to produce the complicated morphology on the surface. Thus, an effect similar to that by a larger ion range can be induced, if the film delays the exposure of the voids. In addition, nanostructuring does not occur when the fluence is insufficient. The morphological change was not considerable [e.g., Figs. 5(i) and 5(j)] when the ratio of the fluence to deposition time was extremely small. In addition, the oblique irradiation on the Au-deposited substrates produced a complicated morphology, comprising elongated pillars. Therefore, the present discussion is relevant for cases with oblique ion incidence, although the Au film can be removed more easily because of a high sputtering yield.

Conversely, the unique morphology cannot be explained solely by the decrease in the ion range. If the ion range is the only factor that produces the unique morphology, a lower incident energy also can produce the same nanostructure. However, 8 keV-Ga+ irradiation could not produce such a nanostructure in the previous work.47 The ion range of 8 keV-Ga+ was estimated to be ∼8 nm, which almost equals the lowest ion range in the present work. Hence, another physical mechanism seems to contribute to the formation of the unique morphology. Here, another scenario is possible, if the barrier layer reported by Alkhaldi et al. can be formed under the Au film.33 The authors observed the layer in which denudes of pores were formed under the deposited film. This effect appeared to be independent of the ion species, fluence, and temperature. In this case, the void nucleated deeper in the substrate, delaying its exposure. Compared with the study by Alkhaldi et al., the present film thickness was significantly smaller. Thus, the film could be removed more easily, and it was uncertain whether such an effect could be considerable or not. At least, the EDX measurements revealed the intensity peaks, which implied the presence of remaining Au, although its amount was too negligible to be quantified. Thus, Au was possibly mixed with Ge by the ion irradiation, and the morphology of the nanostructure could have been affected by this ion beam mixing. Thus, further investigation with high-resolution EDX measurements is required.

This study investigated nanostructuring induced by ion irradiation on Au-deposited Ge substrates. The Au film, which functions as a buffer layer for subsequent irradiation, was deposited on the samples and irradiated using a FIB. The results showed that the deposited Au film affected the size, morphology, and order of nanostructures. At an incident energy of 5 keV, no significant change in the size and morphology of the nanostructures was observed. However, the order of nanostructures changed with an increase in the deposition time, resulting in the formation of irregularities in the entire nanostructured region. These irregularities disappeared at a higher fluence, suggesting that the films protected the Ge surface, leading to incomplete nanostructuring. At an incident energy of 30 keV, the size and morphology of the nanostructures changed. Regarding the structure size, the network-like structures formed at a fluence of 1 × 1020 ions/m2 were larger than the porous structures formed on the undeposited surface. The morphology of those structures differed from that of the ordinary porous structures. The oblique ion irradiation on the Au-deposited substrates produced wide pillars hardly formed on the undeposited substrates. This brief discussion on ion range shows that Au films can delay void exposure, thereby contributing to the formation of a complicated morphology.

See the supplementary material for the detailed experimental and simulation data, which can be helpful for discussion.

The authors thank K. Noshita who belongs to Kyushu University for providing the Python package of “ktch-0.3.1,” which was used for the EFA.

The authors have no conflicts to disclose.

Naoto Oishi: Data curation (equal); Investigation (equal); Writing – original draft (equal); Writing – review & editing (equal). Natsumi Higashide: Data curation (equal); Investigation (equal); Writing – original draft (equal); Writing – review & editing (equal). Noriko Nitta: Data curation (equal); Investigation (equal); Writing – original draft (equal); Writing – review & editing (equal).

The data that support the findings of this study are available within the article.

1.
S. J.
Pearton
,
C. B.
Vartuli
,
J. C.
Zolper
,
C.
Yuan
, and
R. A.
Stall
,
Appl. Phys. Lett.
67
,
1435
1437
(
1995
).
2.
J. L.
Benton
,
S.
Libertino
,
P.
Kringhøj
,
D. J.
Eaglesham
,
J. M.
Poate
, and
S.
Coffa
,
J. Appl. Phys.
82
,
120
125
(
1997
).
3.
U.
Bangert
,
W.
Pierce
,
D. M.
Kepaptsoglou
,
Q.
Ramasse
,
R.
Zan
,
M. H.
Gass
,
J. A.
Van den Berg
,
C. B.
Boothroyd
,
J.
Amani
, and
H.
Hofsäss
,
Nano Lett.
13
,
4902
4907
(
2013
).
4.
J.
Melngailis
,
J. Vac. Sci. Technol. B
5
,
469
495
(
1987
).
5.
J.
Li
,
D.
Stein
,
C.
McMullan
,
D.
Branton
,
M. J.
Aziz
, and
J. A.
Golovchenko
,
Nature
412
,
166
–169 (
2001
).
6.
A.
Chelnokov
,
S.
David
,
K.
Wang
,
F.
Marty
, and
J.-M.
Lourtioz
,
IEEE J. Sel. Top. Quantum Electron.
8
,
919
927
(
2002
).
7.
A. A.
Tseng
,
Small
1
,
924
939
(
2005
).
8.
M. Y.
Ali
,
W.
Hung
, and
F.
Yongqi
,
Int. J. Precis. Eng. Manuf.
11
,
157
170
(
2010
).
9.
W.
Li
,
X.
Zhan
,
X.
Song
,
S.
Si
,
R.
Chen
,
J.
Liu
,
Z.
Wang
,
J.
He
, and
X.
Xiao
,
Small
15
,
1901820
(
2019
).
10.
S.
Facsko
,
T.
Bobek
,
H.
Kurz
,
T.
Dekorsy
,
S.
Kyrsta
, and
R.
Cremer
,
Appl. Phys. Lett.
80
,
130
132
(
2002
).
11.
S.
Hazra
,
T. K.
Chini
,
M. K.
Sanyal
,
J.
Grenzer
, and
U.
Pietsch
,
Phys. Rev. B
70
,
121307
(
2004
).
12.
B.
Ziberi
,
F.
Frost
,
M.
Tartz
,
H.
Neumann
, and
B.
Rauschenbach
,
Appl. Phys. Lett.
92
,
063102
(
2008
).
13.
S. M.
Kluth
,
J. D.
Fitz Gerald
, and
M. C.
Ridgway
,
Appl. Phys. Lett.
86
,
131920
(
2005
).
14.
N. G.
Rudawski
and
K. S.
Jones
,
J. Mater. Res.
28
,
1633
1645
(
2013
).
15.
C. A.
Volkert
and
A. M.
Minor
,
MRS Bull.
32
,
389
399
(
2007
).
16.
Y.
Zhang
and
W. J.
Weber
,
Appl. Phys. Rev.
7
,
041307
(
2020
).
17.
P.
Sigmund
,
Phys. Rev.
184
,
383
416
(
1969
).
18.
V. S.
Smentkowski
,
Prog. Surf. Sci.
64
,
1
58
(
2000
).
19.
G.
Carter
and
V.
Vishnyakov
,
Phys. Rev. B
54
,
17647
17653
(
1996
).
20.
B.
Ziberi
,
F.
Frost
,
T.
Höche
, and
B.
Rauschenbach
,
Phys. Rev. B
72
,
235310
(
2005
).
21.
H. X.
Qian
and
W.
Zhou
,
Mater. Lett.
77
,
113
116
(
2012
).
22.
R.
Cuerno
and
J.-S.
Kim
,
J. Appl. Phys.
128
,
180902
(
2020
).
23.
M. A.
Smirnova
,
V. I.
Bachurin
,
M. E.
Lebedev
,
L. A.
Mazaletsky
,
D. E.
Pukhov
,
A. B.
Churilov
, and
A. S.
Rudy
,
Vacuum
203
,
111238
(
2022
).
24.
P.
Sigmund
,
J. Mater. Sci.
8
,
1545
1553
(
1973
).
25.
R. M.
Bradley
and
J. M. E.
Harper
,
J. Vac. Sci. Technol. A
6
,
2390
2395
(
1988
).
26.
B.
Ziberi
,
F.
Frost
, and
B.
Rauschenbach
,
Appl. Phys. Lett.
88
,
173115
(
2006
).
27.
L.
Romano
,
G.
Impellizzeri
,
M. V.
Tomasello
,
F.
Giannazzo
,
C.
Spinella
, and
M. G.
Grimaldi
,
J. Appl. Phys.
107
,
084314
(
2010
).
28.
B. L.
Darby
,
B. R.
Yates
,
N. G.
Rudawski
,
K. S.
Jones
,
A.
Kontos
, and
R. G.
Elliman
,
Thin Solid Films
519
,
5962
5965
(
2011
).
29.
A.
Perez-Bergquist
,
S.
Zhu
,
K.
Sun
,
X.
Xiang
,
Y.
Zhang
, and
L.
Wang
,
Small
4
,
1119
1124
(
2008
).
30.
D. J.
Manzo
,
C. A.
Bolzan
,
A. M. H.
de Andrade
,
J. R.
Schoffen
, and
R.
Giulian
,
Thin Solid Films
687
,
137447
(
2019
).
31.
N. G.
Rudawski
,
B. L.
Darby
,
B. R.
Yates
,
K. S.
Jones
,
R. G.
Elliman
, and
A. A.
Volinsky
,
Appl. Phys. Lett.
100
,
083111
(
2012
).
32.
R.
Böttger
,
K.-H.
Heinig
,
L.
Bischoff
,
B.
Liedke
, and
S.
Facsko
,
Appl. Phys. A
113
,
53
59
(
2013
).
33.
H. S.
Alkhaldi
,
T. T.
Tran
,
F.
Kremer
, and
J. S.
Williams
,
J. Appl. Phys.
120
,
215706
(
2016
).
34.
J. F.
Ziegler
,
M. D.
Ziegler
, and
J. P.
Biersack
,
Nucl. Instrum. Methods Phys. Res., Sect. B
268
,
1818
1823
(
2010
).
35.
J. F.
Ziegler
,
J. P.
Biersack
, and
M. D.
Ziegler
,
SRIM the Stopping and Range of Ions in Matter
(
SRIM Company
,
2008
).
36.
F. P.
Kuhl
and
C. R.
Giardina
,
Comput. Graphics Image Process.
18
,
236
258
(
1982
).
37.
F. J.
Rohlf
and
J. W.
Archie
,
Syst. Zool.
33
,
302
(
1984
).
38.
See ktch https://doc.ktch.dev/notebooks/outline/elliptic_Fourier_analysis.html#efa for information about the Python package of “ktch-0.3.1.”.
39.
F.
Pedregosa
,
G.
Varoquaux
,
A.
Gramfort
,
V.
Michel
,
B.
Thirion
,
O.
Grisel
,
M.
Blondel
,
P.
Prettenhofer
,
R.
Weiss
,
V.
Dubourg
,
J.
Vanderplas
,
A.
Passos
, D. Cournapeau, M. Bruche, M. Perrot, and É. Duchesnay,
J. Mach. Learn. Res.
12
,
2825
2830
(
2011
).
40.
F.
Liang
,
B.
Wu
,
X.
Dai
,
K.
Li
,
Y.
Zhao
,
H.
Zhang
,
P.
Zhang
,
P.
Vajda
, and
D.
Marculescu
, in
Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
(IEEE, Piscataway, NJ,
2023
), pp.
7061
7070
.
41.
A.
Kirillov
,
E.
Mintun
,
N.
Ravi
,
H.
Mao
,
C.
Rolland
,
L.
Gustafson
,
T.
Xiao
,
S.
Whitehead
,
A. C.
Berg
,
W.
Lo
,
P.
Dollar
, and
R.
Girshick
, in
Proceedings of IEEE/CVF International Conference on Computer Vision (ICCV)
(IEEE, Piscataway, NJ,
2023
), pp.
4015
4026
.
42.
See https://opencv.org/ for information about the Python library of OpenCV.
43.
Q.
Wei
,
X.
Zhou
,
B.
Joshi
,
Y.
Chen
,
K.-D.
Li
,
Q.
Wei
,
K.
Sun
, and
L.
Wang
,
Adv. Mater.
21
,
2865
2869
(
2009
).
44.
Y.
Kuramoto
and
T.
Tsuzuki
,
Prog. Theor. Phys.
55
,
356
(
1976
).
45.
G. I.
Sivashinsky
,
Acta Astron.
4
,
1177
(
1977
).
46.
S.
Facsko
,
T.
Bobek
,
A.
Stahl
,
H.
Kurz
, and
T.
Dekorsy
,
Phys. Rev. B
69
,
153412
(
2004
).
47.
N.
Oishi
,
T.
Yasuoka
,
T.
Kawaharamura
, and
N.
Nitta
,
J. Vac. Sci. Technol. A
41
,
063101
(
2023
).
48.
R. M.
Bradley
and
P. D.
Shipman
,
Appl. Surf. Sci.
258
,
4161
4170
(
2012
).