When a nominally flat surface is bombarded with a broad ion beam at oblique incidence, nanoscale ripples often develop on the surface. For high angles of incidence, surfaces typically develop into a terraced form at the late stages of their time evolution. In the present work, this process is exploited to prevent unwanted smoothing of ordered terraced substrates during the deposition of thin films. A Si surface prepatterned with a 500 nm pitch binary grating structure was bombarded at oblique incidence by a low energy Xe+ ion beam to establish an ordered terraced topography. Subsequently, Si/SiO2 bilayers were deposited on the surface, and further oblique incidence Xe+ bombardment was performed following the deposition of each Si layer to re-establish the ordered terraced topography. Self-organized processes, such as in the present work, that only require exposure of a surface to a plasma or ion source have the potential to provide a simple and inexpensive route for fabricating large-area nanostructured surfaces. The presented procedure has potential applications in the fabrication of multilayer blazed gratings for use in the extreme ultraviolet or soft x-ray regimes.

Low energy ion beam sputtering is an essential part of many erosion and deposition techniques in thin film and surface processing. In ion beam sputter deposition (IBSD), a broad beam of ions is accelerated toward a target of the material to be deposited. The incident ions eject the target atoms through momentum transfer and the sputtered atoms travel through a reduced pressure ambient to deposit on a substrate, forming a thin film. The ability to independently adjust the ion species, ion energy, ion fluence, and process gases in IBSD allows for precise control of film density, thickness, and stoichiometry.1,2 Substrates and thin films can also be bombarded by an independently generated flux of ions. This prior, concurrent, or subsequent bombardment allows for further control over morphology, density, internal stress, crystallinity, and chemical composition of the deposited film.

Aside from the structural modifications due to atomic recoils and the removal of material by sputtering, surface erosion by ion bombardment often results in a pronounced topography. With appropriate geometry, ion beam energy, process gas, and materials, well-ordered patterns form on bombarded surfaces.3–18 Self-organized processes only require exposure of a surface to a plasma or operating ion source and have the potential to provide a simple and inexpensive route for fabricating large-area nanostructured surfaces. One example of self-organized patterning is the development of nanoscale ripples when a nominally flat solid surface is bombarded with a low energy obliquely incident ion beam.19–24 In general, off-normal incidence ion beam erosion of surfaces results in the development of a quasi-periodic height modulation in the surface. Parallel mode ripples, in which the surface has an approximately sinusoidal profile along the projected ion direction, have been observed on a variety of crystalline and amorphous materials for a range of ion species and typically at angles of incidence greater than 60°. If ion bombardment continues for long enough, the ripples often become more disordered and the ripple amplitude saturates. Another common late stage feature is the surface developing a disordered terraced form along the ion beam projection,25–30 as shown in Fig. 1(a) for an initially nominally flat Si surface bombarded with Xe+ ions. Tracing along the surface profile, the slope is nearly equal to a constant positive value for a long spatial interval until the slope changes abruptly to a constant negative value for a long interval. At some point, the slope changes rapidly back again to a nearly constant positive value. The height profile continues flipping between the two selected slopes in this way. Experimentally, the magnitudes of the positive and negative slopes often differ, leading to an asymmetric sawtooth surface profile. This selection of slopes is apparent in Fig. 1(b).

FIG. 1.

AFM image of nominally flat Si sample following 1250 eV Xe+ ion bombardment at 70° off sample normal with inset profile (a). A plot of the distribution of angles (b) demonstrates the selection of the positive slope and the negative slope.

FIG. 1.

AFM image of nominally flat Si sample following 1250 eV Xe+ ion bombardment at 70° off sample normal with inset profile (a). A plot of the distribution of angles (b) demonstrates the selection of the positive slope and the negative slope.

Close modal

Concurrent with the experimental work, theoretical models have been developed. In the generally accepted models, two regimes are distinguished: linear and nonlinear. Based on Sigmund’s theory of sputtering of amorphous targets,31 the Bradley–Harper model explains the origin of ripple formation as a consequence of a surface instability caused by competition between the destabilizing effect of curvature-dependent sputtering and smoothing (surface diffusion) processes.32 Within this linear regime, the resulting ripple orientation (parallel or perpendicular) depends on the angle of incidence, and the wavelength of the semi-periodic ripple pattern is constant. To account for the longer time behavior, the linear model has been extended to include higher order linear and nonlinear effects with different physical origins. In agreement with experimental observations, the nonlinear regime models yield ripples whose amplitude saturates and develops into a terraced state. This terraced state is also known to coarsen over time. Hauffe suggested that the terraced structures coarsen as a result of sputtering by reflected ions.25 This explanation is in disagreement with more recent work by Pearson and Bradley, which predicts that terraces emerge as a result of a cubic nonlinearity in the equation of motion,33 and that coarsening occurs in the absence of ion reflection and re-impingement. Continuing in this vein, Harrison et al. have shown that the surface of a terrace is not completely flat between the regions where the surface slope changes rapidly but instead has a small amplitude ripple topography.34 

As a practical application of slope selection under oblique incidence ion bombardment, Harrison and Bradley (HB) have introduced and analyzed a two-stage procedure for producing high efficiency blazed diffraction gratings by taking advantage of the slope selection as the terraced profile forms.35 Like other optical diffraction gratings, blazed gratings have a constant line spacing. However, the lines in a blazed grating have a triangular (sawtooth) cross section with a characteristic blaze angle chosen so that the diffracted beam and the reflected beam are in the same direction. This allows for high efficiency at a chosen diffraction order and wavelength of light. In the first stage of HB’s procedure, conventional lithography is used to produce a prepatterned substrate with a periodic height modulation in one direction.36–38 Traditional lithographic techniques for the production of periodic structures are well developed but usually do not produce the regions of constant slope and sharp slope transitions required for a blazed grating. This being the case, the second stage consists of bombarding the prepattern with a broad ion beam at a high angle of incidence. While bombarding a flat surface at oblique incidence can produce a terraced surface, the surface does not evolve into a state that is nearly periodic. However, establishing a periodic structure lithographically and then ion etching has been demonstrated to overcome the deficiencies of each technique, producing a suitable blazed grating.39–43 HB’s simulations suggest that while the wavelength of the initial profile is important, the development of the terraced topography is largely insensitive to the detailed shape. This implies that the starting profile can be noisy and contain minor defects without affecting the end result.

In applications where higher efficiency is required from a blazed grating, a multilayer high reflector can be deposited on the grating profile to increase the reflectivity, making a multilayer blazed grating (MBG). Although the starting blazed grating profile may be excellent, the deposition of the multilayer causes the apexes of the terrace profile to round and the grating profile to degrade with each subsequent layer.44–47 

To improve the efficiency of a MBG consisting of alternating layers of materials A and B, HB proposed an extension to the two-stage procedure described above.48 First, the two-stage procedure is applied to establish the initial blazed grating profile in material A. Then, as the multilayer is deposited, each new layer of material A is bombarded with the same obliquely incident ion beam used to establish the initial profile. A layer of material B would then be deposited before the process is repeated again with material A, and so forth. Since conditions identical to those used to establish the initial profile are used to tune up the profile during deposition, this method has the potential to maintain a nearly perfect grating profile for an arbitrary number of deposited layers. Furthermore, since the period can be selected through a lithographic process and the blaze angle can be selected by the ion beam conditions, the proposed method could allow for flexibility in MBG designs.

In this paper, we validate experimentally the HB procedure and its extension. In our process, Xe+ ions bombarded a prepatterned Si surface at oblique incidence to form an ordered terraced topography with a period selected by the prepattern. Following the establishment of the terraced surface, Si/SiO2 bilayers were deposited onto the terraced surface. Further oblique incidence Xe+ bombardment after the deposition of each Si layer re-established the terraced topography without significant degradation of the ordered surface. Self-organized processes, such as in the present work, that only require exposure of a surface to a plasma or operating ion source have the potential to provide a simple and inexpensive route for fabricating large-area nanostructured surfaces. The presented procedure has potential applications in the fabrication of multilayer blazed gratings for use in the extreme ultraviolet or soft x-ray regimes.

Owing to the availability of high purity samples and mature processing techniques, Si was selected as the prepattern material and as the material for deposition with subsequent bombardment. Pre-patterned samples were commercially available 1 cm2 Si gratings with a square profile, a 500 nm period, and a nominally 100 nm feature height. As a spacer material, SiO2 was selected out of convenience. With the introduction of an O2 atmosphere, SiO2 could be deposited from the same target as the Si and provide enough contrast for subsequent cross sectional imaging. Xe+ was selected as the ion species based on the work of Teichmann et al.,22 who demonstrated the formation of ripples with a terraced profile on Si surfaces.

The experiment was performed in a Veeco Spector IBSD system using a modified geometry as described in Fig. 2. The deposition chamber was pumped to a base pressure of 5×107 Torr and operated at a working pressure of approximately 2×104 Torr depending on the ion sources and process gases in use. For the deposition of Si and SiO2, a 16 cm gridded radio frequency ion source incident on a 35 cm Si target was used. The deposition ion source was operated with Ar as the process gas, an ion energy of 500 eV, and an integrated beam current of 100 mA during Si deposition. During SiO2 deposition, the ion source energy and current were changed to 800 eV and 450 mA, respectively, and an O2 flow of 40 sccm was introduced in front of the target. For sample cleaning and oblique incidence ion bombardment, a 12 cm gridded radio frequency assist ion source was used with O2 and Xe process gases, respectively. The O2 plasma cleaning was performed at oblique incidence with an ion energy of 100 eV and an integrated beam current of 100 mA. The oblique incidence Xe+ bombardment was performed with an ion energy of 1500 eV and an integrated beam current of 100 mA. Although Mo grids were used in both sources, Rutherford backscattering measurements of sputtered samples indicated less than 1 ppm of Mo contaminants. To prevent charge buildup, both ion sources were operated with radio frequency neutralizers as external electron sources. Samples were mounted 30 cm from the assist ion source along its central axis and 24 cm from and nearly coplanar to the Si target. The sample was oriented such that the assist ion beam was incident at 75° from the sample normal, and the sample orientation was not changed during or between ion bombardment and deposition steps. To reduce co-deposition from the sample fixture during the bombardment, the fixture was oriented such that the sample surface did not have “line of sight” to other surfaces exposed to the assist ion beam. Si shielding placed on the fixture and Si deposition prior to the experiment further reduced contamination from the fixture and other surfaces in the chamber. No active cooling was present on the sample fixture and the experiment was conducted with a chamber temperature of 60°C.

FIG. 2.

(a) Overview of the deposition system. A 12 cm gridded RF ion source is used for oblique incidence bombardment of the sample with Xe+ ions at 75° off the sample normal. In a separate deposition step, a 16 cm gridded RF ion source sputters a Si target with Ar+ ions to deposit Si or SiO2 (with the introduction of an O2 atmosphere) onto the sample at near normal incidence. (b) Cross section diagram of the terraced surface and sputtering geometry. The subscripts u and d denote local geometry of the upwind and downwind facets, respectively. The slope angle of the surface α measured relative to the global sample normal, and θ is the local Xe+ angle of incidence.

FIG. 2.

(a) Overview of the deposition system. A 12 cm gridded RF ion source is used for oblique incidence bombardment of the sample with Xe+ ions at 75° off the sample normal. In a separate deposition step, a 16 cm gridded RF ion source sputters a Si target with Ar+ ions to deposit Si or SiO2 (with the introduction of an O2 atmosphere) onto the sample at near normal incidence. (b) Cross section diagram of the terraced surface and sputtering geometry. The subscripts u and d denote local geometry of the upwind and downwind facets, respectively. The slope angle of the surface α measured relative to the global sample normal, and θ is the local Xe+ angle of incidence.

Close modal

The surface topography was analyzed by atomic force microscopy (AFM) using a Bioscope Resolve from Bruker operating in ScanAsyst mode. The measurements were performed in air using Si tips with a nominal tip radius of 8 nm. Each sample was analyzed with a scan size of 5×5μm and a resolution of 512×512 pixels at several locations across the sample. AFM scans were consistent across each sample. Select scans were also performed with a scan size of 2×2μm and a resolution of 1024×1024 pixels. The AFM data were analyzed with the GWYDDION scanning probe microscopy software. A second order polynomial background was subtracted from each AFM image. Non-cumulative local slope distribution functions were computed as derivatives along the ion beam projection and plotted as the angle between the local surface normal and the global surface normal [Fig. 2(b)]. Cross sectional measurements were performed with transmission electron microscopy (TEM). The sample was prepared using an in situ focused ion beam liftout technique and imaged at 200 kV in bright-field TEM mode and high-resolution TEM mode.

A modified implementation of the procedure proposed by HB is outlined in Fig. 3 and AFM topographs at several points during the procedure are shown in Fig. 4. Before any bombardment, careful preparation of the chamber, substrate fixture, and templated substrate was performed to minimize metal and hydrocarbon contamination during ion bombardment.49 To prepare the chamber and fixture, approximately 1μm of Si was deposited prior to substrate mounting to mask potential sputter contamination. For substrate preparation, the prepatterned substrates were cleaned chemically prior to mounting and further cleaned for 30 s in situ by an obliquely incident 100 eV O2 plasma etch [Figs. 3(1) and 4(a)]. Then, another 40 nm of Si was deposited, further smoothing the starting prepattern surface and masking potential contamination [Figs. 3(2) and 4(b)]. Next, a 35 s Xe+ bombardment stage established the initial ordered terraced topography, transforming the trapezoidal prepattern into a sawtooth profile along the ion beam projection [Figs. 3(3) and 4(c)]. With the initial profile obtained, a 15 nm spacer layer of SiO2 was deposited [Fig. 3(4)] to provide contrast during later cross sectional imaging. The present method differs from the HB procedure at this point. Due to the relatively high erosion rate of Si (7 nm/s) under the oblique Xe+ bombardment in comparison to the deposition rate of Si (0.5 Å/s), concurrent bombardment and deposition as used in the HB procedure was not possible with our selected ion species, material, and geometry. Instead, separate Si deposition and Xe+ bombardment stages were necessary. In this way, 40 nm of Si was deposited, leaving a Si/SiO2 bilayer on the surface [Fig. 3(5)]. These deposition stages had the effect of removing roughness and smoothing the established sawtooth profile. Then, a subsequent 5 s Xe+ bombardment was performed to sharpen the terraces [Fig. 3(6)]. Finally, the 15 nm SiO2 spacer deposition, the 40 nm Si deposition, and the 5 s sharpening Xe+ bombardment were applied again in order, resulting in a multilayer consisting of alternating Si and SiO2 layers in the vertical direction [Figs. 3(7) and 4(d)].

FIG. 3.

Overview of the modified HB procedure steps. (1) An ordered, periodic prepatterned substrate is prepared. (2) A layer of Si is deposited on the prepattern to ensure a clean environment during subsequent ion bombardment steps. (3) Oblique incidence Xe+ ion bombardment of the coated prepattern is used to generate an ordered sawtooth profile. (4) A contrast layer of SiO2 is deposited. (5) The next layer of Si is deposited. (6) The Si layer from step 5 is bombarded at oblique incidence to regenerate the sawtooth profile. Steps 4–6 constitute a bilayer deposition and etch process, which is repeated in step 7.

FIG. 3.

Overview of the modified HB procedure steps. (1) An ordered, periodic prepatterned substrate is prepared. (2) A layer of Si is deposited on the prepattern to ensure a clean environment during subsequent ion bombardment steps. (3) Oblique incidence Xe+ ion bombardment of the coated prepattern is used to generate an ordered sawtooth profile. (4) A contrast layer of SiO2 is deposited. (5) The next layer of Si is deposited. (6) The Si layer from step 5 is bombarded at oblique incidence to regenerate the sawtooth profile. Steps 4–6 constitute a bilayer deposition and etch process, which is repeated in step 7.

Close modal
FIG. 4.

AFM images with inset line scans of the O2 cleaned trapezoidal pre-pattern (a) after the initial 40 nm Si deposition (b), after the first Xe+ bombardment stage (c), and after two bilayer deposition/Xe+ bombardment stages (d).

FIG. 4.

AFM images with inset line scans of the O2 cleaned trapezoidal pre-pattern (a) after the initial 40 nm Si deposition (b), after the first Xe+ bombardment stage (c), and after two bilayer deposition/Xe+ bombardment stages (d).

Close modal

The slope angle distributions in Fig. 5(a) show that the initial and final sawtooth profiles have similar selected slopes along the ion beam projection. The width of the upwind and downwind slope angle distributions is also similar for both surfaces, demonstrating that the oblique incidence bombardment selects a similarly narrow range of slopes on both facets. The final distributions are broader than the initial distributions, hinting at additional roughness forming on the facets during the multilayer deposition and bombardment. The smaller peak near 0° in the initial distribution is due to incomplete etching of the trapezoidal prepattern. The downwind corner of each trapezoid was shadowed, leaving a notch near the peak of each tooth. Ideally, we would have used a longer etch time to completely remove this42 but such a periodic defect is useful in demonstrating the procedure’s tendency to converge to a sawtooth profile.

FIG. 5.

(a) Distribution of angles along the ion beam projection for the initial (black line), and the final (red dashed line) sawtooth profiles after two bilayer deposition/Xe+ bombardment steps. The angles on the downwind (8°) and upwind (12°) facets are nearly identical before and after. (b) Transverse and (c) parallel 1D PSDs of AFM scans for the initial terraced surface prior to multilayer deposition and the final terraced surface after multilayer deposition.

FIG. 5.

(a) Distribution of angles along the ion beam projection for the initial (black line), and the final (red dashed line) sawtooth profiles after two bilayer deposition/Xe+ bombardment steps. The angles on the downwind (8°) and upwind (12°) facets are nearly identical before and after. (b) Transverse and (c) parallel 1D PSDs of AFM scans for the initial terraced surface prior to multilayer deposition and the final terraced surface after multilayer deposition.

Close modal

The spectral behavior transverse and parallel to the ion beam projection is described by the power spectral density (PSD) plots in Figs. 5(b) and 5(c). In the transverse case, there is not a significant spatial frequency shift of the shoulder or a large peak indicating pattern formation in the initial or final surface. Across the entire spectral range, however, there is an increase in transverse roughness in the final surface, with an average roughness of 0.8±0.3 nm RMS before multilayer deposition increasing to 1.4±0.3 nm RMS after. This does not mean that the roughness would continue to increase with further bilayer and bombardment stages. Analysis by HB48 suggests that as long as the transverse instability is not too strong compared to the instability leading to terraces in the parallel direction, the proposed procedure can be successfully implemented. More bilayers would be necessary to determine if the transverse roughness will continue to develop at the present conditions or if it is sufficiently weak to not be an issue. Along the ion beam projection, the spikes corresponding to the terrace periods line up well between the initial and final surfaces, indicating that periodicity has been maintained. As with the transverse case, there is an increase in roughness across the entire spectral range in the final surface.

The profile labeled “1, 2, 3” in Fig. 6 is the result of the first Xe+ bombardment stage that established the initial sawtooth profile. While the slopes selected are relatively smooth, the notch alluded to previously is present near the peak of each terrace. As seen in profile “4,” it is still present in the SiO2 spacer layer but the deposition of the next Si layer and subsequent etching have removed it from the profile (see profile “5, 6”). Further deposition continues this trend, with a smooth SiO2 layer. After another round of Si deposition and bombardment, a similar defect seems to have returned at the peak of profile “7.” This feature is not present in the adjacent peak. In their analysis of the procedure, HB examined the influence of the chosen grating period and the strength of the smoothing during the deposition stage. It was found that there are bands of acceptable grating periods where even with significant smoothing, the grating profile is recovered during the bombardment stage. Outside of these stable bands, there exists the possibility that the surface reforms into one with twice as many local maxima. It is possible that the new “notch” is the beginning of such a process.

FIG. 6.

Cross sectional TEM image of terraced pre-pattern following two bilayer deposition and Xe+ bombardment steps. The profile labeled “1, 2, 3” corresponds to the result of steps 1, 2, and 3 in Fig. 3. Continuing this scheme, the profile labeled “4” corresponds to step 4, “5, 6” corresponds to steps 5 and 6, and “7” is the result of the final bilayer deposition and bombardment.

FIG. 6.

Cross sectional TEM image of terraced pre-pattern following two bilayer deposition and Xe+ bombardment steps. The profile labeled “1, 2, 3” corresponds to the result of steps 1, 2, and 3 in Fig. 3. Continuing this scheme, the profile labeled “4” corresponds to step 4, “5, 6” corresponds to steps 5 and 6, and “7” is the result of the final bilayer deposition and bombardment.

Close modal

Even in a well-formed terraced state with selected slopes and facet edges with fixed relative positions, there are ripple-like disturbances within the regions of approximately constant slope.34 These ripples propagate laterally along the facets and have an amplitude that saturates with continued bombardment. In the present work, the downwind facet does not deviate significantly from the selected slope as additional layers are deposited, while the upwind facet develops a more pronounced ripple profile. A low amplitude ripple is present on both facets of the surface after step 3 (see Fig. 7) and seems to increase in amplitude on the upwind facet with subsequent deposition and bombardment steps (see Fig. 6 profiles “5, 6” and “7”). It will likely take more deposition and bombardment steps to confirm that this ripple profile would saturate in amplitude.

FIG. 7.

AFM line scan of the initial sawtooth profile following steps 1, 2, and 3. Low amplitude ripples are present in the regions of constant slope.

FIG. 7.

AFM line scan of the initial sawtooth profile following steps 1, 2, and 3. Low amplitude ripples are present in the regions of constant slope.

Close modal

We have implemented and analyzed a method for producing multilayer coated ordered terraced structures whose surface profile does not degrade as additional layers are deposited. A trapezoidal profile Si grating structure was bombarded at oblique incidence by Xe+ ions to yield a sawtooth profile. On this initial profile, a thin layer of SiO2 was deposited, followed by the deposition of a thicker layer of Si. The oblique incidence ion bombardment was then repeated, eroding the thick Si layer until the sawtooth profile was recovered. The deposition of the Si/SiO2 bilayer was then repeated along with another round of oblique incidence Xe+ bombardment, leaving a final surface that very nearly replicated the slopes and period of the initial profile. Although our experimental procedure is not identical to the one prescribed by HB, our work does validate the procedure. It is worth noting that most of the conditions used in our work were selected out of convenience and not with making a useful MBG in mind. That this procedure worked as well as it did, given that the conditions were not optimized, shows that it has considerable promise. Significant work is still required to explore prepattern geometries to see how the transverse instability (if present) evolves with more layers, reduce the roughness on the upwind facets, and select the conditions necessary to generate a useful grating geometry.

The experiment was conceived by E.R. and C.S.M. in close collaboration with R.M.B. E.R. implemented the experiment, compiled the data, and analyzed the results. All authors contributed to the discussion of the results and to the writing of the paper.

We would like to thank Matthew Harrison for valuable discussions. We are grateful to the National Science Foundation (NSF) for its support through Grant No. 1508745 from the Division of Materials Research. The authors acknowledge M. Martin Chicoine and François Schiettekatte from the Départment de Physique, Université de Montréal, Canada for the Rutherford backscattering spectrometry testing. We would also like to thank AMO GmbH Aachen for the Si gratings, EAG Laboratories for the excellent TEM images, and the creators/contributors of the free, open source software GWYDDION.

The authors have no conflicts to disclose.

The data that support the findings of this study are available from the corresponding author upon reasonable request.

1.
H. R.
Kaufman
,
J. J.
Cuomo
, and
J. M. E.
Harper
,
J. Vac. Sci. Technol.
21
,
725
(
1982
).
2.
M.
Ohring
,
Materials Science of Thin Films
(
Elsevier Science & Technology
,
Burlington, MA
,
2001
).
3.
S.
Facsko
et al.,
Science
285
,
1551
(
1999
).
4.
F.
Frost
,
A.
Schindler
, and
F.
Bigl
,
Phys. Rev. Lett.
85
,
4116
(
2000
).
5.
D.
Flamm
,
F.
Frost
, and
D.
Hirsch
,
Appl. Surf. Sci.
179
,
95
(
2001
).
6.
F.
Frost
,
B.
Ziberi
,
A.
Schindler
, and
B.
Rauschenbach
,
Appl. Phys. A
91
,
551
(
2008
).
7.
B.
Ziberi
,
M.
Cornejo
,
F.
Frost
, and
B.
Rauschenbach
,
J. Phys.: Condens. Matter
21
,
224003
(
2009
).
8.
R. M.
Bradley
and
P. D.
Shipman
,
Phys. Rev. Lett.
105
,
145501
(
2010
).
9.
J.
Völlner
,
B.
Ziberi
,
F.
Frost
, and
B.
Rauschenbach
,
J. Appl. Phys.
109
,
043501
(
2011
).
10.
F. C.
Motta
,
P. D.
Shipman
, and
R. M.
Bradley
,
J. Phys. D: Appl. Phys.
45
,
122001
(
2012
).
11.
T.
Basu
,
J. R.
Mohanty
, and
T.
Som
,
Appl. Surf. Sci.
258
,
9944
(
2012
).
12.
A.
Metya
,
D.
Ghose
,
S. A.
Mollick
, and
A.
Majumdar
,
J. Appl. Phys.
111
,
074306
(
2012
).
13.
S. A.
Mollick
,
D.
Ghose
,
P. D.
Shipman
, and
R.
Mark Bradley
,
Appl. Phys. Lett.
104
,
043103
(
2014
).
14.
J.
Muñoz-García
et al.,
Mater. Sci. Eng. R
86
,
1
(
2014
).
15.
M.
Engler
,
S.
Macko
,
F.
Frost
, and
T.
Michely
,
Phys. Rev. B
89
,
245412
(
2014
).
16.
X.
Ou
et al.,
Nanoscale
7
,
018928
(
2015
).
17.
K. M.
Loew
and
R. M.
Bradley
,
Phys. Rev. E
100
,
012801
(
2019
).
18.
R. M.
Bradley
,
Phys. Rev. E
102
,
012807
(
2020
).
19.
P.
Mishra
and
D.
Ghose
,
J. Appl. Phys.
105
,
014304
(
2009
).
20.
M.
Navez
,
D.
Chaperot
, and
C.
Sella
,
C. R. Hebd. Seances Acad. Sci.
254
,
240
(
1962
).
21.
Q.
Wei
,
J.
Lian
,
L. A.
Boatner
,
L. M.
Wang
, and
R. C.
Ewing
,
Phys. Rev. B
80
,
085413
(
2009
).
22.
M.
Teichmann
,
J.
Lorbeer
,
F.
Frost
, and
B.
Rauschenbach
,
Nanoscale Res. Lett.
9
,
1551
(
2014
).
23.
T. K.
Chini
,
D. P.
Datta
, and
S. R.
Bhattacharyya
,
J. Phys.: Condens. Matter
21
,
224004
(
2009
).
24.
B.
Ziberi
,
F.
Frost
,
T.
Höche
, and
B.
Rauschenbach
,
Phys. Rev. B
72
,
235310
(
2005
).
25.
W.
Hauffe
,
Phys. Status Solidi A
35
,
K93
(
1976
).
26.
M.
Teichmann
,
J.
Lorbeer
,
B.
Ziberi
,
F.
Frost
, and
B.
Rauschenbach
,
New J. Phys.
15
,
103029
(
2013
).
27.
T.
Basu
,
D. P.
Datta
, and
T.
Som
,
Nanoscale Res. Lett.
8
,
47
(
2013
).
28.
D.
Chowdhury
and
D.
Ghose
,
Appl. Surf. Sci.
324
,
517
(
2015
).
29.
D. P.
Datta
et al.,
Appl. Surf. Sci.
360
,
131
(
2016
).
30.
P.
Myint
et al.,
Phys. Rev. B
103
,
195423
(
2021
).
31.
P.
Sigmund
,
J. Mater. Sci.
8
,
1545
(
1973
).
32.
R. M.
Bradley
and
J. M. E.
Harper
,
J. Vac. Sci. Technol. A
6
,
2390
(
1988
).
33.
D. A.
Pearson
and
R. M.
Bradley
,
J. Phys.: Condens. Matter
27
,
015010
(
2015
).
34.
M. P.
Harrison
,
D. A.
Pearson
, and
R. M.
Bradley
,
Phys. Rev. E
96
,
032804
(
2017
).
35.
M. P.
Harrison
and
R. M.
Bradley
,
J. Appl. Phys.
121
,
054308
(
2017
).
36.
Q.
Xie
et al.,
J. Alloys Compd.
449
,
261
(
2008
).
37.
C. P.
Fucetola
,
H.
Korre
, and
K. K.
Berggren
,
J. Vac. Sci. Technol. B
27
,
2958
(
2009
).
38.
F.
Koch
et al.,
Microelectron. Eng.
191
,
60
(
2018
).
39.
Y.
Aoyagi
and
S.
Namba
,
J. Appl. Phys.
15
(
4
),
721
(
1976
).
40.
S.
Somekh
and
H. C.
Casey
,
Appl. Opt.
16
,
126
(
1977
).
41.
L. F.
Johnson
,
Appl. Opt.
18
,
2559
(
1979
).
42.
Q.
Liu
,
J.
Wu
, and
M.
Chen
,
Opt. Eng.
52
,
091706
(
2013
).
43.
F.
Siewert
et al.,
J. Synchrotron Radiat.
25
,
91
(
2018
).
44.
D. L.
Voronov
et al.,
Opt. Lett.
35
,
2615
(
2010
).
45.
D. L.
Voronov
et al.,
J. Appl. Phys.
111
,
093521
(
2012
).
46.
D. L.
Voronov
et al.,
Appl. Surf. Sci.
284
,
575
(
2013
).
47.
D. L.
Voronov
,
E. M.
Gullikson
,
F.
Salmassi
,
T.
Warwick
, and
H. A.
Padmore
,
Opt. Lett.
39
,
3157
(
2014
).
48.
M. P.
Harrison
and
R. M.
Bradley
,
J. Appl. Phys.
121
,
225304
(
2017
).
49.
G.
Ozaydin-Ince
and
K. F.
Ludwig Jr.
,
J. Phys.: Condens. Matter
21
,
224008
(
2009
).