- Email: [email protected]
Thermally controlled growth of surface nanostructures on ion-modified AIII-BV semiconductor crystals
Accepted Manuscript Title: Thermally controlled growth of surface nanostructures on ion-modified AIII-BV semiconductor crystals Authors: Elzbieta Trynkiewicz, Benedykt R. Jany, Dominik Wrana, Franciszek Krok PII: DOI: Reference:
S0169-4332(17)32598-9 http://dx.doi.org/10.1016/j.apsusc.2017.08.240 APSUSC 37077
To appear in:
APSUSC
Received date: Accepted date:
18-7-2017 31-8-2017
Please cite this article as: Elzbieta Trynkiewicz, Benedykt R.Jany, Dominik Wrana, Franciszek Krok, Thermally controlled growth of surface nanostructures on ion-modified AIII-BV semiconductor crystals, Applied Surface Sciencehttp://dx.doi.org/10.1016/j.apsusc.2017.08.240 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Thermally controlled growth of surface nanostructures on ion-modified AIII-BV semiconductor crystals Elzbieta Trynkiewicz†, Benedykt R. Jany, Dominik Wrana, Franciszek Krok Marian Smoluchowski Institute of Physics, Jagiellonian University, Lojasiewicza 11, 30-348 Krakow, Poland
†
Corresponding author: [email protected]
Highlights
A sample temperature influence on the pattern formation, under Ar+ ion beam sputtering, is extensively elaborated. Thermal surface diffusion of mobile adspecies acts substantial contribution for evolution of pillar-like structures. Formation of stable self-masking clusters, and effectively their further growth, is temperature dependent. Redeposition is proposed to be self-driven mechanism when higher and higher structures are formed.
The primary motivation for our systematic study is to provide a comprehensive overview of the role of sample temperature on the pattern evolution of several AIII-BV semiconductor crystal (001) surfaces (i.e., InSb, InP, InAs, GaSb) in terms of their response to low-energy Ar+ ion irradiation conditions. The surface morphology and the chemical diversity of such ion-modified binary materials has been characterized by means of scanning electron microscopy (SEM). In general, all surface textures following ion irradiation exhibit transitional behavior from small islands, via vertically oriented 3D nanostructures, to smoothened surface when the sample temperature is increased. This result reinforces our conviction that the mass redistribution of adatoms along the surface plays a vital role during the formation and growth process of surface nanostructures. We would like to emphasize that this paper addresses in detail for the first time the topic of the growth kinetics of the nanostructures with regard to thermal surface diffusion, while simultaneously offering some possible approaches to supplementing previous studies and therein gaining a new insight into this complex issue. The experimental results are discussed with reference to models of the pillars growth, abutting on preferential sputtering, the self-sustained etch masking effect and the redeposition process recently proposed to elucidate the observed nanostructuring mechanism. Keywords: ion-induced surface nanopatterning , nanostructures, ion irradiation, thermal surface diffusion, SEM, AIII-BV semiconductors
1. Introduction Special attention has been drawn of late towards the synthesis of nanostructures such as nanodots, nanotubes and nanowires on AIII-BV semiconductors [1-4]. In particular, the fabrication of regular vertically aligned semiconductor nanowire arrays with a precisely defined height and width is of enormous scientific and technological interest. Semiconductor nanorods with a high aspect ratio 1
may be cut, connected and integrated into functional circuits offering new design possibilities including ideal elements for field-effect transistors [5], light-emitting diodes [6], interconnections with larger scale miniaturization devices [7], as well as antireflective or hydrophobic coatings [8]. A significant step forward within miniaturization technology was achieved through ion-beam sputtering methods. Specifically, the wide interest in regular pattern formation on semiconductor surfaces encompasses those obtained by moderate energy (~few to a dozen keV) ion sputtering using noble gases [9-12]. For many years an ion beam was commonly used for the surface modifications on account of its capacity to produce a formation of one- (1D), two- (2D) or threedimensional (3D) nanostructures on the surface with freely controllable size, simply by using appropriately selected experimental settings. Typically, most of these early experiments covered the evolution of periodic wave-like surface modulations, namely ripples, during irradiation with oblique ion beam incidence [13, 14], and ordered dots generated after low-energy normal bombardment [15, 16]. However, those dots can be alternatively obtained by simultaneous sample rotation with offnormal sputtering [17, 18]. All the aforementioned surface features have been already produced on one- [13] and two-component [19-21] semiconductors, metals [22, 23], and many other materials [24-26]. Several works have also reported arrays of vertical columns (nanowires) or cone-shape structures (pillars) formed on AIII-BV component semiconductors after high fluence irradiation [9, 2729]. A general methodology of development periodic structures on large-scale relies on the spontaneous self-assembly phenomenon. The first theoretical model, proposed by Bradley and Harper (BH) [30], integrates the approach of Sigmund’s model of sputtering [31, 32] and HerringMullins’ classical thermal surface diffusion theory [33, 34]. Starting from a quantitative description of the sputtering process, as well as a functional form of deposited energy transferred by an implanted ion, the BH model has considered that the local sputtering yield depends on the local curvature of the surface [35]. Accordingly, higher total energy is deposited in the regions with positive curvature (valleys) than in negative ones (hillocks), amplifying faster the sputtering of the troughs than the regions of the crest, successively resulting in an increased surface roughness. On the other hand, diffusion processes of mobile surface adspecies [36] lead to a smoothening of the irradiated surface, as opposed to its roughening. As a consequence, these mechanisms involve surface perturbations that can induce the formation of ripple pattern during ion sputtering. Until now, several non-linear extensions of the BH model [37-41] have been suggested to gain a better understanding of pattern formation. Unfortunately, the formation of 3D structures on binary compounds possesses a high complexity level since additional effects therein appear that should be considered. More recently, S. Le Roy and colleagues [42, 43] have introduced a generally accepted phenomenological theory of the behavior of ion-induced binary materials running out beyond the usual BH model. This model assumes that the growth of nanostructures proceeds in three stages. The general concept is based on the preferential sputtering of multi-component materials [44, 45], which is attributed primarily to the difference in the partial sputtering yields of particular elements in the target. For binary materials, like AIII-BV semiconductors, heretofore it was observed the component BV usually sputtered more efficiently on leaving an AIII-enriched surface [46, 47]. Hence, the changes of chemical composition in the near-surface layer, arising from a depletion of one (BV) constituent, initiate phase separation and the agglomeration into clusters. These small islands enriched with metallic component act as a local mask (shield) against the sputtering at the early growth stages. This 2
strong masking effect and mass diffusion in the surface layer initiates a vertical nanowires elongation with prolonged sputtering. Qualitatively this theory is rather attractive, but certain shortcomings are also noticeable. In the work of J. H. Wu and R. S. Goldman [48], a more universal microscopic model has been provided. The authors take into consideration surface kinetics in the process of nanostructure growth, including sputtering and new mechanism – the redeposition of sputtered atoms – the one missing in the previous model. Nevertheless, this theory still does not take into account the effect of mass redistribution on an irradiated surface during ion beam erosion. By dint of the development of the above-mentioned philosophical inquiry purely theoretical works are enjoying a renaissance right now. In particular, the remarkable contribution has been introduced by Bradley-Shipman’s (BS) extensions [49-51] and Norris work [52], as envisaged on the basis of Shenoy, Chan and Chason [53] initially composed theory. Without going into details, the BS model explains the nature of ion-induced binary material surface features in terms of coupling between height perturbations typically induced by variations in the local sputtering yield (taken from the extended BH model), and composition modulations primarily triggered by a preferential sputtering process, within the thin surface layer. However, if we want to keep this progress in broadening our knowledge, we still need some stimulation in that direction through new experimental support. In recent decades the extensive subject literature has provided an exhaustive review of the influence of those experimental parameters that affect the morphology, wavelength or feature size, including ion energy, incidence angle, fluence (equivalent to the time of sputtering), flux, sample temperature, and so on (see for example, Ref. [11, 12] and references therein). Among them, the substrate temperature seems to be an important factor that has an impact on the formation of various kinds of surface patterns. So far, there are only a few experimental reports [9, 54-57] that have concentrated on the influence of the target temperature on pattern development for AIII-BV semiconductor surfaces, however, these measurements were performed at a narrow range of sample temperature and do not provide a decent explanation. The main aim of the present work is to experimentally characterize the growth dynamics of nanostructure patterns on several AIII-BV semiconductor materials typically employed in the nanoelectronics industry, namely InSb(001), InP(001), InAs(001) and GaSb(001), irradiated with 3 keV Ar+ ion beam in normal geometry at a constant ion flux and fluence, with only variable sample temperatures. We have advanced the morphological evolution of fabricated nanostructures in conjunction with their chemical modifications. The experimental results exhibit the formation of surface features diversity within the considered temperature range, including clusters/islands, pillars/nanorods, and in certain circumstances smoothened surface as well. Our experimental findings suggest the conditions for the formation of a stable self-masking cluster, under ion-beam sputtering, are temperature dependent.
2. Experimental details Samples, of 5×5 mm2, were cut from commercially available epi-ready n-doped InSb(001), InP(001), InAs(001), and GaSb(001) wafers (purchased from Kelpin Crystal and MTI Corporations). Prior to being introduced into the ultra-high vacuum (UHV) the chamber samples were cleaned in an
3
ultrasonic bath with isopropanol and subsequently were rinsed sequentially in isopropanol and pure ethanol, and finally dried in a flow of air. Experiments were performed in a UHV chamber with a base pressure of ~1×10-10 mbar. The system was equipped with a standard broad-beam sputtering ion gun with a spot size of a few centimeters in diameter. During the irradiation the noble gas pressure in the UHV chamber was maintained at ~4×10-6 mbar. All samples were sputtered with a 3 keV Ar+ ion beam impinging perpendicular onto the sample surface. The targets were exposed to an average ion beam flux of J=2.98×1013 ion/s·cm2 (J~6.3×1013 ion/s·cm2 for InSb). The typical ion fluence was ϕ=2.15×1017 ion/cm2 (ϕ~4.5×1017 ion/cm2 for InSb), equivalent to an exposure time of 2 hours. The ion irradiation was carried out over a wide substrate temperature (Tsub) range spanning from -129°C to 150°C, as measured using a thermocouple mounted on the sample holder. Sample heating was carried out by passing an electrical current through the sample heating stage. For cooling, liquid nitrogen (or nitrogen gas) flowed through the cooling circuit attached to the sample stage and next, the specimen was resistively heated to the desired temperature. This configuration was realized to ensure precise temperature control. After sputtering the sample was immediately returned to room temperature to enable further analysis. To characterize the resulting surface morphology and chemical composition, the modified samples were investigated ex situ using a field-emission scanning electron microscope (SEM) FEI Quanta 3D FEG. SEM topographical images were collected for both the top and side views (0° and 52°, consequently, with respect to the surface normal) with an electron beam accelerated to 5 kV. SEM micrographs were later analyzed by use of an open-source ImageJ/Fiji [58] software package. The pattern evolution was studied by analyzing the surface structure density and the parameters of shape, i.e., the height, base size and aspect ratio (i.e., the height to size ratio) of the structures. For each sample the average height, base size, and thereby the aspect ratio of the features, were determined from around 100-200 structures. In turn, the average density was designated on the basis of the analysis of three regions with the same area, sufficiently large to retain correct statistics (i.e., around 300-500 structures).
3. Results 3.1 Characterization of surface topography 3.1.1 Indium antimonide (InSb)
Figure 1 depicts the compilation of tilted SEM images of surface morphology after exposing InSb(001) surfaces with 3 keV Ar+ ions at different sample temperatures, ranging from -129°C to 100°C. In the insets are shown the respective images of the magnified single structures. Irradiation at temperature of -129°C results in a surface morphology dominated by small islands of 47±4 nm high and 59±5 nm wide, respectively, surrounded by tiny clusters with a radius of less than ~5 nm. All the islands appear to be oval-shaped without any peculiar facets. The surface irradiated at -82°C reveals a severe longitudinal expansion of the islands into the surface, extending 159±20 nm in width and 95±10 nm in height. The first sign of pillar-like structures was observed at 47°C. Then, the nanostructures become higher monotonously with a decrease in their density, when 4
increasing the sample temperature, up to -21°C, where the mean height of the nanostructures and their width of base were attained at around 282±48 nm and 203±30 nm, respectively. Nonetheless, the character of the body and the envelope of such heterogeneously distributed pillars at reduced temperatures is rather non-symmetric. Continually increasing the temperature, up to 33°C, causes a slight reduction in the mean structures size. Under this sputtering conditions, the pillars (which are created at depressions) possess a height of around 162±11 nm, and a base size of 61±4 nm. Ultimately, the surface morphology developed at 100°C displays a featureless and rather smooth surface with very weakly visible pits and trenches (i.e., overlapping holes). At higher temperatures the arisen pattern seems to be relatively more uniform concerning the symmetry of the obtained features and their local distribution, since the surface contains structures with a regular shape, and a more homogeneous distribution in terms of local ordering when compared with those that occurred at lower temperatures. 3.1.2 Indium phosphide (InP)
Figure 2 illustrates a series of SEM images of the ion-induced surface topography of InP(001) samples. The temperatures were chosen in the range from -50°C to 150°C. A magnified representations of the surface nanostructures are shown in the insets. On a surface irradiated at -50°C, the random texture of densely packed fine drop-shaped features with a radius of less than ~5 nm is quite evident. After irradiation at Tsub = -25°C (not shown), it appears that initially the emergence of small clusters on the surface is due to the limited effective diffusion length, which aggregate into the bigger drop-shaped islands of a diameter of ~30 nm surrounded by small dots with a diameter of less than ~15 nm. A closer look to the topography developed at 0°C shows an intriguing finding, in the form of small holes ~10 nm deep, surrounded by almost ~20 nm clusters formed more preferably at the holes’ edges. At 10°C the high density of weakly pronounced clusters with a drop-like appearance re-appears, quite similar in terms of their size to the pattern obtained at -50°C. Thereafter, for a continually increasing sample temperature, distinct columns 106±7 nm high and 45±3 nm wide were formed at RT. Their size continuously increases reaching a maximum value (~570 nm in high and ~190 nm in width) at a temperature of 100°C. Then, the decreasing of structures size is continued as the sample temperature approaches higher values, but their shape mimics rather massive clusters or pillars with quite a wide base length. The more intensive SEM contrast of the pillars heads indicates their different chemical composition in relation to the pillar body, which will be discussed in detail later. A detailed analysis of the surface topography demonstrates the formation of nanorods in the depressions once again (exactly as in the case of an InSb surface irradiated at a slightly raised temperature) and the presence of a smoothened bare surface (between structures) at just above 40°C. We have found that these depressions become more pronounced at higher temperatures. Further, between 22°C and 75°C one can observe the emergence of hexagonal- and square-shaped pillar heads and peculiar facets of the pillar body in the investigated temperature range indicating their different internal structure. To sum up, under the low sample temperature regime the surface morphology depicts a rough profile which gradually evolves from small dots to well-developed pillars at higher temperatures. It
5
can be also seen that there is quite a large temperature range (over 100°C), in which capped nanorods are formed. 3.1.3. Indium arsenide (InAs)
Figure 3 shows the SEM images for nanostructures developed on InAs(001) surfaces under irradiation with Ar+ ion beam accelerated to 3 keV. In this case, the sample temperature was varied in the range from -45°C to 50°C. On the surface, after bombardment performed at -45°C, small densely-packed ~28 nm dots appear. At a slightly higher temperature of -25°C, the surface morphology is dominated by small round clusters with a diameter of ~20 nm. Then, a further increase in sample temperature affects the decrease of the structures’ density, with a monotonous increase of their vertical size, up to RT. The irradiation at RT leads to the development of diverse features, including clusters, tiny pits, and sparsely distributed small pillars ~80 nm high and ~45 nm wide. The cones predominantly grow in the depressions of the rather smooth substrate surface, similarly as in the case of InSb(001) and InP(001) surfaces. The surface morphology is with small “empty” pits and holes filled with small pillars or seeds that may act as a starting point for the pillars growth for prolonged sputtering. Ultimately, at 50°C, the surface morphology starts to be entirely smooth and no sign of any distinct features can be found. 3.1.4 Gallium antimonide (GaSb)
Figure 4 visualizes a collection of typical SEM morphologies of GaSb(001) surfaces developed after irradiation with an Ar+ ion beam of energy E = 3 keV and a temperature ranging from -95°C to 100°C. It is clearly seen that a regular network of 3D nanostructures is formed over quite a wide temperature range. In general, the successive increase in the sample temperature generates gentle growth in the surface features size, but then their height maintains almost the same order of about ~100 nm in a range of between -42°C and 40°C. The presence of structures with a cone-like body with truncatedcone shape are maintained from -95°C to -25°C, while their shape transforms to a columnar at higher temperatures, till 100°C when the pattern shows an array of nanoclusters of a radius of around ~13 nm. Furthermore, we noted that the degree of symmetry of 3D nanostructures becomes higher with increasing sample temperature, since from -95°C to -42°C fewer and fewer deformed (asymmetric) structures are formed, whereas just at -25°C rather symmetric capped pillars with a round apex have been identified. 3.1.5. Comparison of the results
To give a more quantitative analysis of the impact of the sample temperature on the development of the ion-beam induced morphology of AIII-BV crystal surfaces, several characteristics are measured directly from the SEM micrographs. In Figure 5, the average nanostructures height (a) and their density (b) has been plotted versus the sample temperature. In general, on all irradiated samples the features are formed with different average dimensions and density values at comparable parameters for the ion beam in respect to the sample temperature as well as the particular material type. We noted the correlation between the structures’ size (Fig. 5a) and their density (Fig. 5b); the smaller the size or height of the structures, the higher their density, which is 6
strictly associated with the lowering diffusion length of surface adspecies under decreasing temperature. We have found also that the indium-based samples present the same trends, since the curves resemble bell-shaped profiles, with maximum values corresponding to distinct temperatures (see Fig. 5a). In fact, it can be seen that these dependences are rather complex – no correlation (ambiguous relationship) between the position of the peaks of any of the studied crystal has actually been spotted. Furthermore, the experimental data enable us to draw a morphological diagram of the developed nanostructures (Fig. 5c) that appear on the irradiated AIII-BV surfaces in relation to the sample temperature. We assume that the value of H/S = 1.0 corresponds to the droplet-like structures. Thus, three kinds of surface modifications can be found. The high aspect ratio of 3D structures was defined for H/S>1.0 (light gray), whereas for 0
3.2 Chemical composition analysis Hereafter, the chemical composition of the ion-beam modified surface morphology of AIII-BV crystals was examined with SEM-BSE analysis. In general, SEM imaging is due to the recording of two types of signals, such as SE (secondary electrons) and BSE (backscatter electrons). Their emission yield depends on the local beam incidence angle following the local surface curvature. In contrast, the BSE are high energetic electrons of the primary SEM beam (electrons energy ~ beam energy) resulting from quasi-elastic, high-angle collisions in the sample. Their emission depends on the average sample atomic number (Z). For this reason, the SE mode focuses only on sample topography imaging, whereas BSE imaging provides relevant information on both the sample topography and sample composition and the chemical diversity of the particular objects [59]. Fig. 6(a, b) presents simultaneously the recorded SE and BSE images of ion-beam modified InP(001) surface upon irradiating with a 3 keV Ar+ ion beam carried out at Tsub = 75°C. Fig. 6(b) illustrates different grey tones which made it possible to differentiate particular masses. Areas of brighter contrast can be identified as greatly In-rich zones due to the higher atomic number of indium (In) in comparison to phosphorus (P), or even just pure stoichiometric InP. The variations in the composition of the other parts of the pillars are very noticeable; the body of individual pillars includes a slightly In-rich phase when compared to the surface and the outer walls. Furthermore, most of the heads are essentially composed of In content. In general, all kinds of semiconductors reveal alteration of the compositional distribution in favor of the AIII component at different temperature windows, presumed to have occurred following the loss of BV by preferential sputtering [60-62]. These primeval qualitative studies are coherent with the recently published work by Paramanik et al., [27] where the close cross-sectional TEM-EDX mapping and quantitative chemical analysis of individual pillar also reveals a significant enrichment in marked favor of the metallic constituent.
7
In our case, direct quantitative evaluation, however, is not possible since BSE images would permit only a relative mass identification. Even though the analysis is only indicative, the presented BSE and SE micrographs helped us to better understand qualitatively the structure of the single pillars. To date, it is believed that the metal constituent is continuously supplied by surface diffusion in the surface layer to create stable clusters, something that also strongly correlates with our investigations. For instance, our results are compatible with other studies, as derived from SAED (Selected Area Electron Diffraction) [14], as well as more specific our previous studies [28]. Both studies uncovered the core-shell structure of the individual pillars fabricated on InSb(001) subjected to Ga+ sputtering. The body regions of the single structure are a crystalline structure for the bulk, with an amorphous shell part for the head and the outer walls. Furthermore, the amorphous part of the pillars is greatly enriched in the almost pure metallic component and implanted gallium, whereas the core is composed purely and simply of crystalline InSb.
4. Discussion The presented results clearly show that in the process of the ion-beam induced modification of AIII-BV surfaces, for fixed ion-beam fluence, the developed surface topography depends on the sample temperature. With increasing temperature, the gradual change from a small dots pattern to well-developed pillars array has been noticed, what correlates with the increasing importance of the thermal surface diffusion. Moreover, the most relevant changes at various temperatures have been observed in relation to the degree of the symmetry of the obtained 3D structures and the surface roughness. Even though at low sample temperatures somewhat higher structures can be generated, but with an asymmetric body shape, while those formed at higher temperatures are much more symmetrical. This result may indicate that the diffusion of adatoms just after attachment to the developed structure is still likely (in order to minimize the surface energy), whereas at low temperatures their movement can be greatly reduced or even detained. In all the analyzed In-based materials, especially when well-developed pillars have been formed, the substantial irregularity in their height and base size is due to the different stage of their growth arising from the different times at which the formation of new stable clusters was initiated [28]. Furthermore, at higher temperatures the vast majority of the features are located at an interior of depressions, while at reduced temperatures this effect has not been observed - the base of 3D structures is actually on the same level as the substrate surface. This reveals that the depressions are not a result of the enhanced erosion in the periphery of the well-build cones as has been claimed by Kramczynski and Gnaser [29], but that their origin is mainly temperature induced. To discuss the mechanism of the formation of ion-induced patterns we now consider the mean effective diffusion length in relation to the sample temperature, which is given in Fig. 7(a). The surface density of nanostructures (D), determined directly from SEM images, is strictly associated with the effective diffusion radius of adspecies (RD) during the surface pattern formation, therefore 1 1 . 2 √𝐷
we can estimate RD ≅ ∙
In general, for In-based semiconductors the diffusion length tends to increase with the substrate temperature, something we can correlate with the progress of thermal surface diffusion, according to the Arrhenius law. 8
Conceptually, in order to shed more light on the formation of the developed patterns, one may analyze step by step the processes that are active during the evolution of observed nanostructures, and which can be also expressed in the view of the given temperature-dependent regimes. In the first step of the structures’ evolution (the initial stage of the irradiation), the deviations of the surface composition from its original state in favor of the metallic constituent are triggered by the preferential sputtering of the BV component. Afterwards, the minimization of the total Gibbs free energy provokes the agglomeration of an excess metallic component until stable, self-shielding clusters (islands) are formed. The behavior of the nucleation in the very-early stage is characteristic for the diffusive coalescence process. In principle, the formation of these clusters can be possible within a suitable range of sample temperatures, since at excessively low and moderate temperatures the aggregation of adatoms comes from relatively small areas, enabling the rather quick formation of the stable clusters. We suggest that the formation and also further development of the nanostructures are very sensitive to the efficiency of the attachment/detachment of coming/outgoing adatoms which principally increase with sample temperature. Simply, at too low temperatures the limited diffusion length is enough only for the creation of a high density of tiny structures, but at suitable (somewhat higher) temperatures, the stable clusters (which still may be produced rather quickly) can be further developed into gradually higher vertical nanostructures due to the increasing diffusion radius of mobile adspecies. However, above a certain critical temperature, the pillar’s size begins to decrease and the extremely smooth surface can, in some cases, be stimulated. We believe that the increasing efficiency of the detachment of adatoms from the growing clusters is the main factor that hinder the formation of stable islands, which results consequently in the slower development of the nanostructures. Thus, the sample temperature affects the creation of stable self-masking clusters. We have also checked that the different rates of nanostructures’ growth have not been influenced by the change in the total sputtering rate as the height of the sputtered crater is temperature-independent. As can be seen in Fig. 7(b), at certain temperatures, the height of the structures is noticeably higher than the estimated diffusion radius of the adatoms. What is more interesting, we have also found that for the same sample temperatures (marked by a dotted circle), the structures’ height is almost 2 times higher than the depth of the sputtered crater, suggesting that up-hill surface diffusion is not sufficient and there should be another mechanism which supplies material in the process of the structures’ growth. We propose that this process can be associated with the redeposition of sputtered adspecies. When higher and higher structures are formed, the redeposition increasingly (non-linearly) drives the growth rate of the structures. This stems from the character of the angular distribution of sputtered atoms that can be "captured" by the high structures; the higher the structures, the more sputtered atoms can be caught. Finally, we would like to emphasize that the GaSb samples behave in a little different manner, producing more uniform structures in terms of their size distribution and local ordering in the whole range of considered temperatures. These results show that stable clusters are created rather at the same stage of ion irradiation, independently of the sample temperature, thereby the GaSb sample conditions for the formation of self-shielding clusters are very weakly sensitive to the target temperature.
5. Summary and conclusion 9
In conclusion, this paper reports on the influence of the mass surface diffusion on several semiconductor binary materials, whereby the pattern after exposing to an argon ion beam varies in relation to the sample temperature. Depending on the sample temperature under ion beam irradiation, particular AIII-BV surfaces show the formation of three types of distinct modifications in the surface topography. We have observed the transitional behavior between these kinds of patterns starting from small islands via high-aspect ratio pillars to the featureless surface, when the sample temperature increased. This result confirms that the thermal surface diffusion supplies the redistributed material in the process of nanostructure growth. But, in fact, the process of the creation of the initial, stable clusters, leading finally to the nanostructures development, is temperature-dependent. Furthermore, under suitable conditions the growth kinetics can be additionally strengthened by the redeposition process. In our work we would like to emphasize that the surface features are strongly subjected to the applied sample temperature under ion irradiation, since the size of the structures can be tuned in the range from a few to ~600 nm, just by selecting the optimal target temperature. Not only the size, but also the shape, lateral arrangement and density of the fabricated nanostructures can be directly governed via sample temperature using similar ion beam parameters.
Acknowledgements The authors acknowledge the financial support of the Polish National Science Center, grant no. DEC-2012/07/B/ST5/00906. Part of the research was carried out with equipment purchased with financial support from the European Regional Development Fund within the framework of the Polish Innovation Economy Operational Program (Contract No. POIG.02.01.00-12-023/08).
References [1] E. A. Chekhovich, M. N. Makhonin, A. I. Tartakovskii, A. Yacoby , H. Bluhm, K. C. Nowack, L. M. K. Vandersypen, Nature Materials 12 (2003) 494. [2] W. Han, S. Fan, Q. Li, Y. Hu, Science 277 (1997) 128. [3] L. J. Lauhon, M. S. Gudiksen, D. Wang, C. M. Lieber, Nature 420 (2002) 57. [4] J. Zhang, M. Brehm, M. Grydlikac, O. G. Schmidtac, Chem. Soc. Rev. 44 (2015) 26. [5] M.-H. Ham, J.-H. Choi, W. Hwang, C. Park, W.-Y. Lee, J.-M. Myoung, Nanotechnology 17 (2006) 2203. [6] M. S. Gudiksen, L. J. Lauhon, J. Wang, D. C. Smith, C. M. Lieber, Nature 415 (2002) 617. [7] G. S. Snider, R. S. Williams, Nanotechnology 18 (2007) 035204. [8] K. Askar, B. M. Phillips, Y. Fang, B. Choi, N. Gozubenli, P. Jiang, B. Jiang, Colloids Surf. A: Physicochem. Eng. Asp. 439 (2013) 84. [9] M. P. Seah, S. J. Spencer, P. J. Cumpson, J. E. Johnstone, Appl. Surf. Sci. 144–145 (1999) 151. [10] V. S. Smentkowski, Prog. Surf. Sci. 64 (2000) 1-58. [11] M. A. Makeev, R. Cuerno, A.-L. Barabási, Nucl. Instrum. Meth. Phys. Res. B 197 (2002) 185. [12] W. L. Chan, E. Chason, J. Appl. Phys. 101 (2007) 121301. [13] A. Keller, S. Facsko, Materials 3 (2010) 4811. [14] M. Kang, J. H. Wu, W. Ye, Y. Jiang, E. A. Robb, C. Chen, R. S. Goldman, Appl. Phys. Lett. 104 (2014) 052103. [15] S. Facsko, T. Dekorsy, C. Koerdt, C. Trappe, H. Kurz, A. Vogt, H. L. Hartnagel, Science 285 (1999) 1551. [16] S. Facsko, H. Kurz, T. Dekorsy, Phys. Rev. B 63 (2001) 165329. [17] F. Frost, A. Schindler, F. Bigl, Phys. Rev. Lett. 85 (2000) 4116-4119. [18] B. Ziberi, F. Frost, M. Tartz, H. Neumann, B. Rauschenbach, Thin Solid Films 459 (2004) 106. [19] S. Facsko, T. Bobek, H. Kurz, T. Dekorsy, S. Kyrsta, R. Cremer, Appl. Phys. Lett. 80 (2002) 130.
10
[20] M. Kang, J. H. Wu, S. Huang, M. V. Warren, Y. Jiang, E. A. Robb, R. S. Goldman, Appl. Phys. Lett. 101 (2012) 082101. [21] O. El-Atwani, S. A. Norris, K. Ludwig, S. Gonderman, P. Allain, Sci. Rep. 5 (2015) 18207. [22] S. Rusponi, G. Costantini, F. Buatier de Mongeot, C. Boragno, U. Valbusa, Appl. Phys. Lett. 75 (1999) 3318. [23] G. Costantini, S. Rusponi, F. Buatier de Mongeot, C. Boragno, U. Valbusa, J. Phys.: Condens. Matter 13 (2001) 5875. [24] T. M. Mayer, E. Chason, A. J. Howard, J. Appl. Phys. 76 (1994) 1633. [25] S. Habenicht, W. Bolse, H. Feldermann, U. Geyer, H. Hofsäss, K. P. Lieb, F. Roccaforte, Europhys. Lett. 50 (2000) 209. [26] D. Flamm, F. Frost, D. Hirsch, Appl. Surf. Sci. 179 (2001) 95. [27] D. Paramanik, T. Suzuki, N. Ikeda, T. Nagai, C. Van Haesendonck, Physica E 44 (2012) 1644. [28] B. R. Jany, K. Szajna, M. Nikiel, D. Wrana, E. Trynkiewicz, R. Pedrys, F. Krok, Appl. Surf. Sci. 327 (2015) 86. [29] D. Kramczynski, H. Gnaser, Appl. Surf. Sci. 355 (2015) 653. [30] R. M. Bradley, J. M. E. Harper, J. Vac. Sci. Technol. A 6 (1988) 2390. [31] P. Sigmund, Phys. Rev. 184 (1969) 383. [32] P. Sigmund, J. Mater. Sci. 8 (1973) 1545. [33] C. Herring, J. Appl. Phys. 21 (1950) 301. [34] W. W. Mullins, J. Appl. Phys. 28 (1957) 333. [35] E. Chason, T. M. Mayer, B. K. Kellerman, D. T. Mcllroy, A. J. Howard, Phys. Rev. Lett. 72 (1994) 3040. [36] J. Erlebacher, M. J. Aziz, E. Chason, M. B. Sinclair, J. A. Floro, Phys. Rev. Lett 82 (1999) 2330. [37] R. Cuerno and A.-L. Barabási, Phys. Rev. Lett. 74 (1995) 4746. [38] T. C. Kim, C.-M. Ghim, H. J. Kim, D. H. Kim, D. Y. Noh, N. D. Kim, J. W. Chung, J. S. Yang, Y. J. Chang, T. W. Noh, B. Kahng, J.-S. Kim, Phys. Rev. Lett. 92 (2004) 246104-1. [39] M. Castro, R. Cuerno, L. Vazquez, R. Gago, Phys. Rev. Lett. 94 (2005) 016102. [40] J. Muñoz-García, M. Castro, R. Cuerno, Phys. Rev. Lett. 96 (2006) 086101. [41] R. Cuerno, M. Castro, J. Muñoz-García, R. Gago, L. Vázquez, Nucl. Instrum. Meth. Phys. Res. B 269 (2011) 894. [42] S. Le Roy, E. Barthel, N. Brun, A. Lelarge, E. Søndergård, J. Appl. Phys. 106 (2009) 094308. [43] S. Le Roy, E. Søndergård, M. Kildemo, I. S. Nerbø, M. Plapp, Phys. Rev. B 81 (2010) 161401. [44] H. Gades, H. M. Urbassek, Nucl. Instrum. Methods Phys. Res. B 102 (1995) 261. [45] V. I. Zaporozchenko, M. G. Stepanova, Prog. Surf. Sci. 49 (1995) 155. [46] S. K. Tan, A. T. S. Wee, J. Vac. Sci. Technol. B 24 (2006) 1444. [47] F. Krok, Vacuum 83 (2009) 745. [48] J. H. Wu, R. S. Goldman, Appl. Phys. Lett. 100 (2012) 053103. [49] R. M. Bradley, P. D. Shipman, Phys. Rev. Lett. 105 (2010) 145501. [50] P. D. Shipman, R. M. Bradley, Phys. Rev. B 84 (2011) 085420. [51] R. M. Bradley, P. D. Shipman, Appl. Surf. Sci. 258 (2012) 4161. [52] S. A. Norris, J. Appl. Phys. 114 (2013) 204303. [53] V. B. Shenoy, W. L. Chan, E. Chason, Phys. Rev. Lett. 98 (2007) 256101. [54] S. W. MacLaren, J. E. Baker, N. L. Finnegan, C. M. Loxton, J. Vac. Sci. Technol. A 10 (1992) 468. [55] Y. Yuba, S. Hazama, K. Gamo, Nucl. Instrum. Methods Phys. Res. B 206 (2003) 648. [56] F. Frost, B. Ziberi, T. Höche, B. Rauschenbach, Nucl. Instrum. Methods Phys. Res. B 216 (2004) 9. [57] S. R. Saeed, O. P. Sinha, F. Krok, T. Zembok, R. Pedrys, M. Szymonski, Nucl. Instrum. Methods Phys. Res. B 267 (2009) 2752. [58] J. Schindelin et al., Nat. Methods 9 (2012) 676–682. [59] P. Goodhew, J. Humphries, R. Beanland (Eds.), Electron Microscopy and Analysis, third edition, Taylor & Francis, London, 2001, pp. 141-149. [60] M. Nozu, M. Tanemura, F. Okuyama, Surf. Sci. Lett. 304 (1994) L468. [61] M. Bouslama, C. Jardin, M. Ghamnia, Vacuum 46 (1995)143. [62] M. Szymonski, F. Krok, P. Struski, J. J. Kolodziej, B. Such, Prog. Surf. Sci. 74 (2003) 331. Figure caption
11
Fig. 1. Sequence of SEM images of the tilted views (52° off the normal) of modified InSb(001) surfaces bombarded with a 3 keV Ar+ ion beam (J~6.3×1013 ion/s·cm2, ϕ~4.5×1017 ion/cm2) at normal incidence, presented for increasing sample temperatures. The insets in the upper-right corners represent the corresponding single structure with scale bars of 200 nm for each image.
Fig. 2. SEM images of the tilted views of modified InP(001) surfaces bombarded with a 3 keV Ar+ ion beam (J~2.98×1013 ion/s·cm2, ϕ~2.15×1017 ion/cm2) at normal incidence, presented for increasing sample temperatures. The insets in the upper-right corners represent the corresponding single structure with scale bars of 100 nm for each image.
Fig. 3. SEM images of the tilted views of modified InAs(001) surfaces bombarded with a 3 keV Ar+ ion beam (J~2.98×1013 ion/s·cm2, ϕ~2.15×1017 ion/cm2) at normal incidence, presented for increasing sample temperatures. The insets in the upper-right corners represent the corresponding single structure with scale bars of 50 nm for each image.
12
Fig. 4. SEM images of the tilted views of modified GaSb(001) surfaces bombarded with a 3 keV Ar+ ion beam (J~2.98×1013 ion/s·cm2, ϕ~2.15×1017 ion/cm2) at normal incidence, presented for increasing sample temperatures. The insets in the upper-right corners represent the corresponding single structure with scale bars of 25 nm for each image.
Fig. 5. Temperature dependent average (a) height (H), (b) surface density (D), and (c) aspect ratio (H/S) of the nanostructures formed on AIII-BV semiconductor (001) surfaces irradiated with a 3 keV Ar+ ion beam at normal incidence (J~6.3×1013 ion/s·cm2, ϕ~4.5×1017 ion/cm2 for InSb, otherwise J~2.98×1013 ion/s·cm2, ϕ~2.15×1017 ion/cm2). All values were extracted directly from the SEM images. The solid lines on all the graphs are merely guides for the eye.
Fig. 6. (a) Secondary electron (SE), and (b) backscatter electron (BSE) images of the same area of InP surface bombarded at 75°C. All images were collected 30° off-normal. Note that SE and BSE signals were taken simultaneously with the same scan (both, for an energy electron beam of 10 keV). The insets present the individual structures with scale bars of 100 nm.
13
Fig. 7. (a) Mean diffusion radius (RD) in the representation of natural logarithm and (b) height (H) to diffusion length (RD) ratio, plotted as a function of sample temperature (Tsub) during irradiation with a 3 keV Ar+ ion beam with respect to the particular AIII-BV semiconductors.
14
S0169-4332(17)32598-9 http://dx.doi.org/10.1016/j.apsusc.2017.08.240 APSUSC 37077
To appear in:
APSUSC
Received date: Accepted date:
18-7-2017 31-8-2017
Please cite this article as: Elzbieta Trynkiewicz, Benedykt R.Jany, Dominik Wrana, Franciszek Krok, Thermally controlled growth of surface nanostructures on ion-modified AIII-BV semiconductor crystals, Applied Surface Sciencehttp://dx.doi.org/10.1016/j.apsusc.2017.08.240 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Thermally controlled growth of surface nanostructures on ion-modified AIII-BV semiconductor crystals Elzbieta Trynkiewicz†, Benedykt R. Jany, Dominik Wrana, Franciszek Krok Marian Smoluchowski Institute of Physics, Jagiellonian University, Lojasiewicza 11, 30-348 Krakow, Poland
†
Corresponding author: [email protected]
Highlights
A sample temperature influence on the pattern formation, under Ar+ ion beam sputtering, is extensively elaborated. Thermal surface diffusion of mobile adspecies acts substantial contribution for evolution of pillar-like structures. Formation of stable self-masking clusters, and effectively their further growth, is temperature dependent. Redeposition is proposed to be self-driven mechanism when higher and higher structures are formed.
The primary motivation for our systematic study is to provide a comprehensive overview of the role of sample temperature on the pattern evolution of several AIII-BV semiconductor crystal (001) surfaces (i.e., InSb, InP, InAs, GaSb) in terms of their response to low-energy Ar+ ion irradiation conditions. The surface morphology and the chemical diversity of such ion-modified binary materials has been characterized by means of scanning electron microscopy (SEM). In general, all surface textures following ion irradiation exhibit transitional behavior from small islands, via vertically oriented 3D nanostructures, to smoothened surface when the sample temperature is increased. This result reinforces our conviction that the mass redistribution of adatoms along the surface plays a vital role during the formation and growth process of surface nanostructures. We would like to emphasize that this paper addresses in detail for the first time the topic of the growth kinetics of the nanostructures with regard to thermal surface diffusion, while simultaneously offering some possible approaches to supplementing previous studies and therein gaining a new insight into this complex issue. The experimental results are discussed with reference to models of the pillars growth, abutting on preferential sputtering, the self-sustained etch masking effect and the redeposition process recently proposed to elucidate the observed nanostructuring mechanism. Keywords: ion-induced surface nanopatterning , nanostructures, ion irradiation, thermal surface diffusion, SEM, AIII-BV semiconductors
1. Introduction Special attention has been drawn of late towards the synthesis of nanostructures such as nanodots, nanotubes and nanowires on AIII-BV semiconductors [1-4]. In particular, the fabrication of regular vertically aligned semiconductor nanowire arrays with a precisely defined height and width is of enormous scientific and technological interest. Semiconductor nanorods with a high aspect ratio 1
may be cut, connected and integrated into functional circuits offering new design possibilities including ideal elements for field-effect transistors [5], light-emitting diodes [6], interconnections with larger scale miniaturization devices [7], as well as antireflective or hydrophobic coatings [8]. A significant step forward within miniaturization technology was achieved through ion-beam sputtering methods. Specifically, the wide interest in regular pattern formation on semiconductor surfaces encompasses those obtained by moderate energy (~few to a dozen keV) ion sputtering using noble gases [9-12]. For many years an ion beam was commonly used for the surface modifications on account of its capacity to produce a formation of one- (1D), two- (2D) or threedimensional (3D) nanostructures on the surface with freely controllable size, simply by using appropriately selected experimental settings. Typically, most of these early experiments covered the evolution of periodic wave-like surface modulations, namely ripples, during irradiation with oblique ion beam incidence [13, 14], and ordered dots generated after low-energy normal bombardment [15, 16]. However, those dots can be alternatively obtained by simultaneous sample rotation with offnormal sputtering [17, 18]. All the aforementioned surface features have been already produced on one- [13] and two-component [19-21] semiconductors, metals [22, 23], and many other materials [24-26]. Several works have also reported arrays of vertical columns (nanowires) or cone-shape structures (pillars) formed on AIII-BV component semiconductors after high fluence irradiation [9, 2729]. A general methodology of development periodic structures on large-scale relies on the spontaneous self-assembly phenomenon. The first theoretical model, proposed by Bradley and Harper (BH) [30], integrates the approach of Sigmund’s model of sputtering [31, 32] and HerringMullins’ classical thermal surface diffusion theory [33, 34]. Starting from a quantitative description of the sputtering process, as well as a functional form of deposited energy transferred by an implanted ion, the BH model has considered that the local sputtering yield depends on the local curvature of the surface [35]. Accordingly, higher total energy is deposited in the regions with positive curvature (valleys) than in negative ones (hillocks), amplifying faster the sputtering of the troughs than the regions of the crest, successively resulting in an increased surface roughness. On the other hand, diffusion processes of mobile surface adspecies [36] lead to a smoothening of the irradiated surface, as opposed to its roughening. As a consequence, these mechanisms involve surface perturbations that can induce the formation of ripple pattern during ion sputtering. Until now, several non-linear extensions of the BH model [37-41] have been suggested to gain a better understanding of pattern formation. Unfortunately, the formation of 3D structures on binary compounds possesses a high complexity level since additional effects therein appear that should be considered. More recently, S. Le Roy and colleagues [42, 43] have introduced a generally accepted phenomenological theory of the behavior of ion-induced binary materials running out beyond the usual BH model. This model assumes that the growth of nanostructures proceeds in three stages. The general concept is based on the preferential sputtering of multi-component materials [44, 45], which is attributed primarily to the difference in the partial sputtering yields of particular elements in the target. For binary materials, like AIII-BV semiconductors, heretofore it was observed the component BV usually sputtered more efficiently on leaving an AIII-enriched surface [46, 47]. Hence, the changes of chemical composition in the near-surface layer, arising from a depletion of one (BV) constituent, initiate phase separation and the agglomeration into clusters. These small islands enriched with metallic component act as a local mask (shield) against the sputtering at the early growth stages. This 2
strong masking effect and mass diffusion in the surface layer initiates a vertical nanowires elongation with prolonged sputtering. Qualitatively this theory is rather attractive, but certain shortcomings are also noticeable. In the work of J. H. Wu and R. S. Goldman [48], a more universal microscopic model has been provided. The authors take into consideration surface kinetics in the process of nanostructure growth, including sputtering and new mechanism – the redeposition of sputtered atoms – the one missing in the previous model. Nevertheless, this theory still does not take into account the effect of mass redistribution on an irradiated surface during ion beam erosion. By dint of the development of the above-mentioned philosophical inquiry purely theoretical works are enjoying a renaissance right now. In particular, the remarkable contribution has been introduced by Bradley-Shipman’s (BS) extensions [49-51] and Norris work [52], as envisaged on the basis of Shenoy, Chan and Chason [53] initially composed theory. Without going into details, the BS model explains the nature of ion-induced binary material surface features in terms of coupling between height perturbations typically induced by variations in the local sputtering yield (taken from the extended BH model), and composition modulations primarily triggered by a preferential sputtering process, within the thin surface layer. However, if we want to keep this progress in broadening our knowledge, we still need some stimulation in that direction through new experimental support. In recent decades the extensive subject literature has provided an exhaustive review of the influence of those experimental parameters that affect the morphology, wavelength or feature size, including ion energy, incidence angle, fluence (equivalent to the time of sputtering), flux, sample temperature, and so on (see for example, Ref. [11, 12] and references therein). Among them, the substrate temperature seems to be an important factor that has an impact on the formation of various kinds of surface patterns. So far, there are only a few experimental reports [9, 54-57] that have concentrated on the influence of the target temperature on pattern development for AIII-BV semiconductor surfaces, however, these measurements were performed at a narrow range of sample temperature and do not provide a decent explanation. The main aim of the present work is to experimentally characterize the growth dynamics of nanostructure patterns on several AIII-BV semiconductor materials typically employed in the nanoelectronics industry, namely InSb(001), InP(001), InAs(001) and GaSb(001), irradiated with 3 keV Ar+ ion beam in normal geometry at a constant ion flux and fluence, with only variable sample temperatures. We have advanced the morphological evolution of fabricated nanostructures in conjunction with their chemical modifications. The experimental results exhibit the formation of surface features diversity within the considered temperature range, including clusters/islands, pillars/nanorods, and in certain circumstances smoothened surface as well. Our experimental findings suggest the conditions for the formation of a stable self-masking cluster, under ion-beam sputtering, are temperature dependent.
2. Experimental details Samples, of 5×5 mm2, were cut from commercially available epi-ready n-doped InSb(001), InP(001), InAs(001), and GaSb(001) wafers (purchased from Kelpin Crystal and MTI Corporations). Prior to being introduced into the ultra-high vacuum (UHV) the chamber samples were cleaned in an
3
ultrasonic bath with isopropanol and subsequently were rinsed sequentially in isopropanol and pure ethanol, and finally dried in a flow of air. Experiments were performed in a UHV chamber with a base pressure of ~1×10-10 mbar. The system was equipped with a standard broad-beam sputtering ion gun with a spot size of a few centimeters in diameter. During the irradiation the noble gas pressure in the UHV chamber was maintained at ~4×10-6 mbar. All samples were sputtered with a 3 keV Ar+ ion beam impinging perpendicular onto the sample surface. The targets were exposed to an average ion beam flux of J=2.98×1013 ion/s·cm2 (J~6.3×1013 ion/s·cm2 for InSb). The typical ion fluence was ϕ=2.15×1017 ion/cm2 (ϕ~4.5×1017 ion/cm2 for InSb), equivalent to an exposure time of 2 hours. The ion irradiation was carried out over a wide substrate temperature (Tsub) range spanning from -129°C to 150°C, as measured using a thermocouple mounted on the sample holder. Sample heating was carried out by passing an electrical current through the sample heating stage. For cooling, liquid nitrogen (or nitrogen gas) flowed through the cooling circuit attached to the sample stage and next, the specimen was resistively heated to the desired temperature. This configuration was realized to ensure precise temperature control. After sputtering the sample was immediately returned to room temperature to enable further analysis. To characterize the resulting surface morphology and chemical composition, the modified samples were investigated ex situ using a field-emission scanning electron microscope (SEM) FEI Quanta 3D FEG. SEM topographical images were collected for both the top and side views (0° and 52°, consequently, with respect to the surface normal) with an electron beam accelerated to 5 kV. SEM micrographs were later analyzed by use of an open-source ImageJ/Fiji [58] software package. The pattern evolution was studied by analyzing the surface structure density and the parameters of shape, i.e., the height, base size and aspect ratio (i.e., the height to size ratio) of the structures. For each sample the average height, base size, and thereby the aspect ratio of the features, were determined from around 100-200 structures. In turn, the average density was designated on the basis of the analysis of three regions with the same area, sufficiently large to retain correct statistics (i.e., around 300-500 structures).
3. Results 3.1 Characterization of surface topography 3.1.1 Indium antimonide (InSb)
Figure 1 depicts the compilation of tilted SEM images of surface morphology after exposing InSb(001) surfaces with 3 keV Ar+ ions at different sample temperatures, ranging from -129°C to 100°C. In the insets are shown the respective images of the magnified single structures. Irradiation at temperature of -129°C results in a surface morphology dominated by small islands of 47±4 nm high and 59±5 nm wide, respectively, surrounded by tiny clusters with a radius of less than ~5 nm. All the islands appear to be oval-shaped without any peculiar facets. The surface irradiated at -82°C reveals a severe longitudinal expansion of the islands into the surface, extending 159±20 nm in width and 95±10 nm in height. The first sign of pillar-like structures was observed at 47°C. Then, the nanostructures become higher monotonously with a decrease in their density, when 4
increasing the sample temperature, up to -21°C, where the mean height of the nanostructures and their width of base were attained at around 282±48 nm and 203±30 nm, respectively. Nonetheless, the character of the body and the envelope of such heterogeneously distributed pillars at reduced temperatures is rather non-symmetric. Continually increasing the temperature, up to 33°C, causes a slight reduction in the mean structures size. Under this sputtering conditions, the pillars (which are created at depressions) possess a height of around 162±11 nm, and a base size of 61±4 nm. Ultimately, the surface morphology developed at 100°C displays a featureless and rather smooth surface with very weakly visible pits and trenches (i.e., overlapping holes). At higher temperatures the arisen pattern seems to be relatively more uniform concerning the symmetry of the obtained features and their local distribution, since the surface contains structures with a regular shape, and a more homogeneous distribution in terms of local ordering when compared with those that occurred at lower temperatures. 3.1.2 Indium phosphide (InP)
Figure 2 illustrates a series of SEM images of the ion-induced surface topography of InP(001) samples. The temperatures were chosen in the range from -50°C to 150°C. A magnified representations of the surface nanostructures are shown in the insets. On a surface irradiated at -50°C, the random texture of densely packed fine drop-shaped features with a radius of less than ~5 nm is quite evident. After irradiation at Tsub = -25°C (not shown), it appears that initially the emergence of small clusters on the surface is due to the limited effective diffusion length, which aggregate into the bigger drop-shaped islands of a diameter of ~30 nm surrounded by small dots with a diameter of less than ~15 nm. A closer look to the topography developed at 0°C shows an intriguing finding, in the form of small holes ~10 nm deep, surrounded by almost ~20 nm clusters formed more preferably at the holes’ edges. At 10°C the high density of weakly pronounced clusters with a drop-like appearance re-appears, quite similar in terms of their size to the pattern obtained at -50°C. Thereafter, for a continually increasing sample temperature, distinct columns 106±7 nm high and 45±3 nm wide were formed at RT. Their size continuously increases reaching a maximum value (~570 nm in high and ~190 nm in width) at a temperature of 100°C. Then, the decreasing of structures size is continued as the sample temperature approaches higher values, but their shape mimics rather massive clusters or pillars with quite a wide base length. The more intensive SEM contrast of the pillars heads indicates their different chemical composition in relation to the pillar body, which will be discussed in detail later. A detailed analysis of the surface topography demonstrates the formation of nanorods in the depressions once again (exactly as in the case of an InSb surface irradiated at a slightly raised temperature) and the presence of a smoothened bare surface (between structures) at just above 40°C. We have found that these depressions become more pronounced at higher temperatures. Further, between 22°C and 75°C one can observe the emergence of hexagonal- and square-shaped pillar heads and peculiar facets of the pillar body in the investigated temperature range indicating their different internal structure. To sum up, under the low sample temperature regime the surface morphology depicts a rough profile which gradually evolves from small dots to well-developed pillars at higher temperatures. It
5
can be also seen that there is quite a large temperature range (over 100°C), in which capped nanorods are formed. 3.1.3. Indium arsenide (InAs)
Figure 3 shows the SEM images for nanostructures developed on InAs(001) surfaces under irradiation with Ar+ ion beam accelerated to 3 keV. In this case, the sample temperature was varied in the range from -45°C to 50°C. On the surface, after bombardment performed at -45°C, small densely-packed ~28 nm dots appear. At a slightly higher temperature of -25°C, the surface morphology is dominated by small round clusters with a diameter of ~20 nm. Then, a further increase in sample temperature affects the decrease of the structures’ density, with a monotonous increase of their vertical size, up to RT. The irradiation at RT leads to the development of diverse features, including clusters, tiny pits, and sparsely distributed small pillars ~80 nm high and ~45 nm wide. The cones predominantly grow in the depressions of the rather smooth substrate surface, similarly as in the case of InSb(001) and InP(001) surfaces. The surface morphology is with small “empty” pits and holes filled with small pillars or seeds that may act as a starting point for the pillars growth for prolonged sputtering. Ultimately, at 50°C, the surface morphology starts to be entirely smooth and no sign of any distinct features can be found. 3.1.4 Gallium antimonide (GaSb)
Figure 4 visualizes a collection of typical SEM morphologies of GaSb(001) surfaces developed after irradiation with an Ar+ ion beam of energy E = 3 keV and a temperature ranging from -95°C to 100°C. It is clearly seen that a regular network of 3D nanostructures is formed over quite a wide temperature range. In general, the successive increase in the sample temperature generates gentle growth in the surface features size, but then their height maintains almost the same order of about ~100 nm in a range of between -42°C and 40°C. The presence of structures with a cone-like body with truncatedcone shape are maintained from -95°C to -25°C, while their shape transforms to a columnar at higher temperatures, till 100°C when the pattern shows an array of nanoclusters of a radius of around ~13 nm. Furthermore, we noted that the degree of symmetry of 3D nanostructures becomes higher with increasing sample temperature, since from -95°C to -42°C fewer and fewer deformed (asymmetric) structures are formed, whereas just at -25°C rather symmetric capped pillars with a round apex have been identified. 3.1.5. Comparison of the results
To give a more quantitative analysis of the impact of the sample temperature on the development of the ion-beam induced morphology of AIII-BV crystal surfaces, several characteristics are measured directly from the SEM micrographs. In Figure 5, the average nanostructures height (a) and their density (b) has been plotted versus the sample temperature. In general, on all irradiated samples the features are formed with different average dimensions and density values at comparable parameters for the ion beam in respect to the sample temperature as well as the particular material type. We noted the correlation between the structures’ size (Fig. 5a) and their density (Fig. 5b); the smaller the size or height of the structures, the higher their density, which is 6
strictly associated with the lowering diffusion length of surface adspecies under decreasing temperature. We have found also that the indium-based samples present the same trends, since the curves resemble bell-shaped profiles, with maximum values corresponding to distinct temperatures (see Fig. 5a). In fact, it can be seen that these dependences are rather complex – no correlation (ambiguous relationship) between the position of the peaks of any of the studied crystal has actually been spotted. Furthermore, the experimental data enable us to draw a morphological diagram of the developed nanostructures (Fig. 5c) that appear on the irradiated AIII-BV surfaces in relation to the sample temperature. We assume that the value of H/S = 1.0 corresponds to the droplet-like structures. Thus, three kinds of surface modifications can be found. The high aspect ratio of 3D structures was defined for H/S>1.0 (light gray), whereas for 0
3.2 Chemical composition analysis Hereafter, the chemical composition of the ion-beam modified surface morphology of AIII-BV crystals was examined with SEM-BSE analysis. In general, SEM imaging is due to the recording of two types of signals, such as SE (secondary electrons) and BSE (backscatter electrons). Their emission yield depends on the local beam incidence angle following the local surface curvature. In contrast, the BSE are high energetic electrons of the primary SEM beam (electrons energy ~ beam energy) resulting from quasi-elastic, high-angle collisions in the sample. Their emission depends on the average sample atomic number (Z). For this reason, the SE mode focuses only on sample topography imaging, whereas BSE imaging provides relevant information on both the sample topography and sample composition and the chemical diversity of the particular objects [59]. Fig. 6(a, b) presents simultaneously the recorded SE and BSE images of ion-beam modified InP(001) surface upon irradiating with a 3 keV Ar+ ion beam carried out at Tsub = 75°C. Fig. 6(b) illustrates different grey tones which made it possible to differentiate particular masses. Areas of brighter contrast can be identified as greatly In-rich zones due to the higher atomic number of indium (In) in comparison to phosphorus (P), or even just pure stoichiometric InP. The variations in the composition of the other parts of the pillars are very noticeable; the body of individual pillars includes a slightly In-rich phase when compared to the surface and the outer walls. Furthermore, most of the heads are essentially composed of In content. In general, all kinds of semiconductors reveal alteration of the compositional distribution in favor of the AIII component at different temperature windows, presumed to have occurred following the loss of BV by preferential sputtering [60-62]. These primeval qualitative studies are coherent with the recently published work by Paramanik et al., [27] where the close cross-sectional TEM-EDX mapping and quantitative chemical analysis of individual pillar also reveals a significant enrichment in marked favor of the metallic constituent.
7
In our case, direct quantitative evaluation, however, is not possible since BSE images would permit only a relative mass identification. Even though the analysis is only indicative, the presented BSE and SE micrographs helped us to better understand qualitatively the structure of the single pillars. To date, it is believed that the metal constituent is continuously supplied by surface diffusion in the surface layer to create stable clusters, something that also strongly correlates with our investigations. For instance, our results are compatible with other studies, as derived from SAED (Selected Area Electron Diffraction) [14], as well as more specific our previous studies [28]. Both studies uncovered the core-shell structure of the individual pillars fabricated on InSb(001) subjected to Ga+ sputtering. The body regions of the single structure are a crystalline structure for the bulk, with an amorphous shell part for the head and the outer walls. Furthermore, the amorphous part of the pillars is greatly enriched in the almost pure metallic component and implanted gallium, whereas the core is composed purely and simply of crystalline InSb.
4. Discussion The presented results clearly show that in the process of the ion-beam induced modification of AIII-BV surfaces, for fixed ion-beam fluence, the developed surface topography depends on the sample temperature. With increasing temperature, the gradual change from a small dots pattern to well-developed pillars array has been noticed, what correlates with the increasing importance of the thermal surface diffusion. Moreover, the most relevant changes at various temperatures have been observed in relation to the degree of the symmetry of the obtained 3D structures and the surface roughness. Even though at low sample temperatures somewhat higher structures can be generated, but with an asymmetric body shape, while those formed at higher temperatures are much more symmetrical. This result may indicate that the diffusion of adatoms just after attachment to the developed structure is still likely (in order to minimize the surface energy), whereas at low temperatures their movement can be greatly reduced or even detained. In all the analyzed In-based materials, especially when well-developed pillars have been formed, the substantial irregularity in their height and base size is due to the different stage of their growth arising from the different times at which the formation of new stable clusters was initiated [28]. Furthermore, at higher temperatures the vast majority of the features are located at an interior of depressions, while at reduced temperatures this effect has not been observed - the base of 3D structures is actually on the same level as the substrate surface. This reveals that the depressions are not a result of the enhanced erosion in the periphery of the well-build cones as has been claimed by Kramczynski and Gnaser [29], but that their origin is mainly temperature induced. To discuss the mechanism of the formation of ion-induced patterns we now consider the mean effective diffusion length in relation to the sample temperature, which is given in Fig. 7(a). The surface density of nanostructures (D), determined directly from SEM images, is strictly associated with the effective diffusion radius of adspecies (RD) during the surface pattern formation, therefore 1 1 . 2 √𝐷
we can estimate RD ≅ ∙
In general, for In-based semiconductors the diffusion length tends to increase with the substrate temperature, something we can correlate with the progress of thermal surface diffusion, according to the Arrhenius law. 8
Conceptually, in order to shed more light on the formation of the developed patterns, one may analyze step by step the processes that are active during the evolution of observed nanostructures, and which can be also expressed in the view of the given temperature-dependent regimes. In the first step of the structures’ evolution (the initial stage of the irradiation), the deviations of the surface composition from its original state in favor of the metallic constituent are triggered by the preferential sputtering of the BV component. Afterwards, the minimization of the total Gibbs free energy provokes the agglomeration of an excess metallic component until stable, self-shielding clusters (islands) are formed. The behavior of the nucleation in the very-early stage is characteristic for the diffusive coalescence process. In principle, the formation of these clusters can be possible within a suitable range of sample temperatures, since at excessively low and moderate temperatures the aggregation of adatoms comes from relatively small areas, enabling the rather quick formation of the stable clusters. We suggest that the formation and also further development of the nanostructures are very sensitive to the efficiency of the attachment/detachment of coming/outgoing adatoms which principally increase with sample temperature. Simply, at too low temperatures the limited diffusion length is enough only for the creation of a high density of tiny structures, but at suitable (somewhat higher) temperatures, the stable clusters (which still may be produced rather quickly) can be further developed into gradually higher vertical nanostructures due to the increasing diffusion radius of mobile adspecies. However, above a certain critical temperature, the pillar’s size begins to decrease and the extremely smooth surface can, in some cases, be stimulated. We believe that the increasing efficiency of the detachment of adatoms from the growing clusters is the main factor that hinder the formation of stable islands, which results consequently in the slower development of the nanostructures. Thus, the sample temperature affects the creation of stable self-masking clusters. We have also checked that the different rates of nanostructures’ growth have not been influenced by the change in the total sputtering rate as the height of the sputtered crater is temperature-independent. As can be seen in Fig. 7(b), at certain temperatures, the height of the structures is noticeably higher than the estimated diffusion radius of the adatoms. What is more interesting, we have also found that for the same sample temperatures (marked by a dotted circle), the structures’ height is almost 2 times higher than the depth of the sputtered crater, suggesting that up-hill surface diffusion is not sufficient and there should be another mechanism which supplies material in the process of the structures’ growth. We propose that this process can be associated with the redeposition of sputtered adspecies. When higher and higher structures are formed, the redeposition increasingly (non-linearly) drives the growth rate of the structures. This stems from the character of the angular distribution of sputtered atoms that can be "captured" by the high structures; the higher the structures, the more sputtered atoms can be caught. Finally, we would like to emphasize that the GaSb samples behave in a little different manner, producing more uniform structures in terms of their size distribution and local ordering in the whole range of considered temperatures. These results show that stable clusters are created rather at the same stage of ion irradiation, independently of the sample temperature, thereby the GaSb sample conditions for the formation of self-shielding clusters are very weakly sensitive to the target temperature.
5. Summary and conclusion 9
In conclusion, this paper reports on the influence of the mass surface diffusion on several semiconductor binary materials, whereby the pattern after exposing to an argon ion beam varies in relation to the sample temperature. Depending on the sample temperature under ion beam irradiation, particular AIII-BV surfaces show the formation of three types of distinct modifications in the surface topography. We have observed the transitional behavior between these kinds of patterns starting from small islands via high-aspect ratio pillars to the featureless surface, when the sample temperature increased. This result confirms that the thermal surface diffusion supplies the redistributed material in the process of nanostructure growth. But, in fact, the process of the creation of the initial, stable clusters, leading finally to the nanostructures development, is temperature-dependent. Furthermore, under suitable conditions the growth kinetics can be additionally strengthened by the redeposition process. In our work we would like to emphasize that the surface features are strongly subjected to the applied sample temperature under ion irradiation, since the size of the structures can be tuned in the range from a few to ~600 nm, just by selecting the optimal target temperature. Not only the size, but also the shape, lateral arrangement and density of the fabricated nanostructures can be directly governed via sample temperature using similar ion beam parameters.
Acknowledgements The authors acknowledge the financial support of the Polish National Science Center, grant no. DEC-2012/07/B/ST5/00906. Part of the research was carried out with equipment purchased with financial support from the European Regional Development Fund within the framework of the Polish Innovation Economy Operational Program (Contract No. POIG.02.01.00-12-023/08).
References [1] E. A. Chekhovich, M. N. Makhonin, A. I. Tartakovskii, A. Yacoby , H. Bluhm, K. C. Nowack, L. M. K. Vandersypen, Nature Materials 12 (2003) 494. [2] W. Han, S. Fan, Q. Li, Y. Hu, Science 277 (1997) 128. [3] L. J. Lauhon, M. S. Gudiksen, D. Wang, C. M. Lieber, Nature 420 (2002) 57. [4] J. Zhang, M. Brehm, M. Grydlikac, O. G. Schmidtac, Chem. Soc. Rev. 44 (2015) 26. [5] M.-H. Ham, J.-H. Choi, W. Hwang, C. Park, W.-Y. Lee, J.-M. Myoung, Nanotechnology 17 (2006) 2203. [6] M. S. Gudiksen, L. J. Lauhon, J. Wang, D. C. Smith, C. M. Lieber, Nature 415 (2002) 617. [7] G. S. Snider, R. S. Williams, Nanotechnology 18 (2007) 035204. [8] K. Askar, B. M. Phillips, Y. Fang, B. Choi, N. Gozubenli, P. Jiang, B. Jiang, Colloids Surf. A: Physicochem. Eng. Asp. 439 (2013) 84. [9] M. P. Seah, S. J. Spencer, P. J. Cumpson, J. E. Johnstone, Appl. Surf. Sci. 144–145 (1999) 151. [10] V. S. Smentkowski, Prog. Surf. Sci. 64 (2000) 1-58. [11] M. A. Makeev, R. Cuerno, A.-L. Barabási, Nucl. Instrum. Meth. Phys. Res. B 197 (2002) 185. [12] W. L. Chan, E. Chason, J. Appl. Phys. 101 (2007) 121301. [13] A. Keller, S. Facsko, Materials 3 (2010) 4811. [14] M. Kang, J. H. Wu, W. Ye, Y. Jiang, E. A. Robb, C. Chen, R. S. Goldman, Appl. Phys. Lett. 104 (2014) 052103. [15] S. Facsko, T. Dekorsy, C. Koerdt, C. Trappe, H. Kurz, A. Vogt, H. L. Hartnagel, Science 285 (1999) 1551. [16] S. Facsko, H. Kurz, T. Dekorsy, Phys. Rev. B 63 (2001) 165329. [17] F. Frost, A. Schindler, F. Bigl, Phys. Rev. Lett. 85 (2000) 4116-4119. [18] B. Ziberi, F. Frost, M. Tartz, H. Neumann, B. Rauschenbach, Thin Solid Films 459 (2004) 106. [19] S. Facsko, T. Bobek, H. Kurz, T. Dekorsy, S. Kyrsta, R. Cremer, Appl. Phys. Lett. 80 (2002) 130.
10
[20] M. Kang, J. H. Wu, S. Huang, M. V. Warren, Y. Jiang, E. A. Robb, R. S. Goldman, Appl. Phys. Lett. 101 (2012) 082101. [21] O. El-Atwani, S. A. Norris, K. Ludwig, S. Gonderman, P. Allain, Sci. Rep. 5 (2015) 18207. [22] S. Rusponi, G. Costantini, F. Buatier de Mongeot, C. Boragno, U. Valbusa, Appl. Phys. Lett. 75 (1999) 3318. [23] G. Costantini, S. Rusponi, F. Buatier de Mongeot, C. Boragno, U. Valbusa, J. Phys.: Condens. Matter 13 (2001) 5875. [24] T. M. Mayer, E. Chason, A. J. Howard, J. Appl. Phys. 76 (1994) 1633. [25] S. Habenicht, W. Bolse, H. Feldermann, U. Geyer, H. Hofsäss, K. P. Lieb, F. Roccaforte, Europhys. Lett. 50 (2000) 209. [26] D. Flamm, F. Frost, D. Hirsch, Appl. Surf. Sci. 179 (2001) 95. [27] D. Paramanik, T. Suzuki, N. Ikeda, T. Nagai, C. Van Haesendonck, Physica E 44 (2012) 1644. [28] B. R. Jany, K. Szajna, M. Nikiel, D. Wrana, E. Trynkiewicz, R. Pedrys, F. Krok, Appl. Surf. Sci. 327 (2015) 86. [29] D. Kramczynski, H. Gnaser, Appl. Surf. Sci. 355 (2015) 653. [30] R. M. Bradley, J. M. E. Harper, J. Vac. Sci. Technol. A 6 (1988) 2390. [31] P. Sigmund, Phys. Rev. 184 (1969) 383. [32] P. Sigmund, J. Mater. Sci. 8 (1973) 1545. [33] C. Herring, J. Appl. Phys. 21 (1950) 301. [34] W. W. Mullins, J. Appl. Phys. 28 (1957) 333. [35] E. Chason, T. M. Mayer, B. K. Kellerman, D. T. Mcllroy, A. J. Howard, Phys. Rev. Lett. 72 (1994) 3040. [36] J. Erlebacher, M. J. Aziz, E. Chason, M. B. Sinclair, J. A. Floro, Phys. Rev. Lett 82 (1999) 2330. [37] R. Cuerno and A.-L. Barabási, Phys. Rev. Lett. 74 (1995) 4746. [38] T. C. Kim, C.-M. Ghim, H. J. Kim, D. H. Kim, D. Y. Noh, N. D. Kim, J. W. Chung, J. S. Yang, Y. J. Chang, T. W. Noh, B. Kahng, J.-S. Kim, Phys. Rev. Lett. 92 (2004) 246104-1. [39] M. Castro, R. Cuerno, L. Vazquez, R. Gago, Phys. Rev. Lett. 94 (2005) 016102. [40] J. Muñoz-García, M. Castro, R. Cuerno, Phys. Rev. Lett. 96 (2006) 086101. [41] R. Cuerno, M. Castro, J. Muñoz-García, R. Gago, L. Vázquez, Nucl. Instrum. Meth. Phys. Res. B 269 (2011) 894. [42] S. Le Roy, E. Barthel, N. Brun, A. Lelarge, E. Søndergård, J. Appl. Phys. 106 (2009) 094308. [43] S. Le Roy, E. Søndergård, M. Kildemo, I. S. Nerbø, M. Plapp, Phys. Rev. B 81 (2010) 161401. [44] H. Gades, H. M. Urbassek, Nucl. Instrum. Methods Phys. Res. B 102 (1995) 261. [45] V. I. Zaporozchenko, M. G. Stepanova, Prog. Surf. Sci. 49 (1995) 155. [46] S. K. Tan, A. T. S. Wee, J. Vac. Sci. Technol. B 24 (2006) 1444. [47] F. Krok, Vacuum 83 (2009) 745. [48] J. H. Wu, R. S. Goldman, Appl. Phys. Lett. 100 (2012) 053103. [49] R. M. Bradley, P. D. Shipman, Phys. Rev. Lett. 105 (2010) 145501. [50] P. D. Shipman, R. M. Bradley, Phys. Rev. B 84 (2011) 085420. [51] R. M. Bradley, P. D. Shipman, Appl. Surf. Sci. 258 (2012) 4161. [52] S. A. Norris, J. Appl. Phys. 114 (2013) 204303. [53] V. B. Shenoy, W. L. Chan, E. Chason, Phys. Rev. Lett. 98 (2007) 256101. [54] S. W. MacLaren, J. E. Baker, N. L. Finnegan, C. M. Loxton, J. Vac. Sci. Technol. A 10 (1992) 468. [55] Y. Yuba, S. Hazama, K. Gamo, Nucl. Instrum. Methods Phys. Res. B 206 (2003) 648. [56] F. Frost, B. Ziberi, T. Höche, B. Rauschenbach, Nucl. Instrum. Methods Phys. Res. B 216 (2004) 9. [57] S. R. Saeed, O. P. Sinha, F. Krok, T. Zembok, R. Pedrys, M. Szymonski, Nucl. Instrum. Methods Phys. Res. B 267 (2009) 2752. [58] J. Schindelin et al., Nat. Methods 9 (2012) 676–682. [59] P. Goodhew, J. Humphries, R. Beanland (Eds.), Electron Microscopy and Analysis, third edition, Taylor & Francis, London, 2001, pp. 141-149. [60] M. Nozu, M. Tanemura, F. Okuyama, Surf. Sci. Lett. 304 (1994) L468. [61] M. Bouslama, C. Jardin, M. Ghamnia, Vacuum 46 (1995)143. [62] M. Szymonski, F. Krok, P. Struski, J. J. Kolodziej, B. Such, Prog. Surf. Sci. 74 (2003) 331. Figure caption
11
Fig. 1. Sequence of SEM images of the tilted views (52° off the normal) of modified InSb(001) surfaces bombarded with a 3 keV Ar+ ion beam (J~6.3×1013 ion/s·cm2, ϕ~4.5×1017 ion/cm2) at normal incidence, presented for increasing sample temperatures. The insets in the upper-right corners represent the corresponding single structure with scale bars of 200 nm for each image.
Fig. 2. SEM images of the tilted views of modified InP(001) surfaces bombarded with a 3 keV Ar+ ion beam (J~2.98×1013 ion/s·cm2, ϕ~2.15×1017 ion/cm2) at normal incidence, presented for increasing sample temperatures. The insets in the upper-right corners represent the corresponding single structure with scale bars of 100 nm for each image.
Fig. 3. SEM images of the tilted views of modified InAs(001) surfaces bombarded with a 3 keV Ar+ ion beam (J~2.98×1013 ion/s·cm2, ϕ~2.15×1017 ion/cm2) at normal incidence, presented for increasing sample temperatures. The insets in the upper-right corners represent the corresponding single structure with scale bars of 50 nm for each image.
12
Fig. 4. SEM images of the tilted views of modified GaSb(001) surfaces bombarded with a 3 keV Ar+ ion beam (J~2.98×1013 ion/s·cm2, ϕ~2.15×1017 ion/cm2) at normal incidence, presented for increasing sample temperatures. The insets in the upper-right corners represent the corresponding single structure with scale bars of 25 nm for each image.
Fig. 5. Temperature dependent average (a) height (H), (b) surface density (D), and (c) aspect ratio (H/S) of the nanostructures formed on AIII-BV semiconductor (001) surfaces irradiated with a 3 keV Ar+ ion beam at normal incidence (J~6.3×1013 ion/s·cm2, ϕ~4.5×1017 ion/cm2 for InSb, otherwise J~2.98×1013 ion/s·cm2, ϕ~2.15×1017 ion/cm2). All values were extracted directly from the SEM images. The solid lines on all the graphs are merely guides for the eye.
Fig. 6. (a) Secondary electron (SE), and (b) backscatter electron (BSE) images of the same area of InP surface bombarded at 75°C. All images were collected 30° off-normal. Note that SE and BSE signals were taken simultaneously with the same scan (both, for an energy electron beam of 10 keV). The insets present the individual structures with scale bars of 100 nm.
13
Fig. 7. (a) Mean diffusion radius (RD) in the representation of natural logarithm and (b) height (H) to diffusion length (RD) ratio, plotted as a function of sample temperature (Tsub) during irradiation with a 3 keV Ar+ ion beam with respect to the particular AIII-BV semiconductors.
14