Cone array formation on Si surfaces by low-energy He plasma irradiation with magnetron sputtering pre-deposited Ta Open Access

Since the pioneering discovery¹ of the cones on the surface of ion-sputtered metals in a glow discharge experiment in 1942, considerable attention has been devoted to the fabrication and analysis of cone array covered surfaces. These cone-structured materials, particularly on silicon (Si) substrates, exhibit distinct properties that deviate from those of the base material, owing to their unique geometric configurations. As a result, they hold great promise for a wide range of applications, including, but not limited to, solar cells,^2,3 surface-enhanced Raman spectroscopy,^4,5 and field emitters.^6,7 Among the various fabrication methods,^8,9 ion beam bombardment (IB) with impurity deposition^6,10,11 has emerged as a prospective approach notable for its efficiency, cost-effectiveness, and the absence of toxicity, rendering it a viable and practical method.

Recently, a new ion beam method, characterized by low-energy (<100 eV) and high-flux (∼10²² m⁻² s⁻¹) helium (He) plasma irradiation, has extended the application of ion beam nano-processing to the He–Si scenario.^12–14 In this method, samples are negatively biased to accelerate incident He ions through the electric field in the sheath, imparting them with energies equal to the difference between the plasma potential and the applied negative bias, typically remaining below 100 eV. Dense cone arrays have been successfully generated by various researchers utilizing deliberate impurity co-deposition.^13,14 However, the co-deposition approach presents challenges in achieving precise control over the quantity of metal impurities deposited onto the Si substrate. Moreover, the growth mechanism governing these cone arrays remains elusive.

In this paper, we conduct research on the effects of He plasma irradiation on a Si (100) substrate with prior tantalum (Ta) deposition using magnetron sputtering. In contrast to the flat surface observed in Ref. 11 following argon (Ar) beam irradiation with Ta co-deposition, our work reveals the formation of cone arrays, underscoring the unique patterning capability of He plasma on the Si surface. Meanwhile, a systematic study is performed investigating the growth dynamics of these cones through a series of experiments conducted at varying fluences. Additionally, a detailed examination of the cone structure is conducted. The findings are employed to elucidate the mechanism underlying cone formation.

The irradiation experiments are conducted in the linear plasma device, CLIPS (Compact LInear Plasma-Surface interaction device), at the University of Science and Technology of China. The steady-state He plasma is generated by a Philips ionization gauge (PIG) discharge using a LaB₆ cathode that is heated by a direct current. The base pressure in the vacuum chamber is 5 × 10⁻⁵ Pa, and the working pressure during plasma irradiation is maintained at 1.5 Pa. Figure 1(a) depicts a schematic diagram of the experimental setup. To ensure that all the clamps and ceramics in the holder are covered by Ta, a specially designed sample holder system is employed, as shown in Fig. 1(b). The Si samples are placed downstream perpendicular to the magnetic field line. The magnetic field strength is approximately 0.01 T, which suppresses the radial diffusion of the plasma and results in a high-density plasma. The typical electron density n_e is in the order of 10¹⁸ m⁻³, and the electron temperature of the center of the plasma column is T_e ∼ 8.5 eV. The He ion flux to the Si samples is measured using an electrostatic probe located around 10 mm upstream from the sample, and the He flux is fixed at 1.7 × 10²² m⁻² s⁻¹ in this experiment. The samples are electrically biased in the He plasma with a potential of +5.0 V, and the incident ion energy, E_i, which is determined by the potential difference between the sample and the plasma space potential, is controlled by changing the biasing voltage. The He ion incident direction is approximately normal to the sample surface due to acceleration in the sheath. The Si samples are mounted on a water-cooling stage, and the surface temperature, T_s, is measured using a K-type thermocouple. To ensure a good thermal connection, the head of the thermocouple presses against the back of the sample.

In this study, commercial P-type (100) Si samples (10 × 10 × 1 mm³) with an electrical resistivity of ∼0.001 Ω cm are used as targets for the investigation. Before processing, the Si samples are cleaned with acetone, ethanol, and de-ionized water, successively. The natural oxide layer on the surface of the commercial Si wafers does not affect the results since the plasma promptly sputtered the oxide layer at the beginning of plasma irradiation.^14,15 During the irradiation process, the Ta cover floats in He plasma with much less potential than the sputter threshold energy, which is more than 100 eV.¹⁶ This ensures that no metal atoms will be sputtered from the fixture around the sample and deposited onto the sample during the He plasma irradiation process. The magnetron sputtering method is used to pre-deposit Ta on the surface of the Si samples. The deposition process is controlled by adjusting the deposition rate (0.014 nm/s in this case) of the magnetron sputtering device and deposition time, which leads to a sub-monolayer with an average thickness of approximately 0.4 nm. The deposited Ta thickness is chosen due to the threshold of metal atoms for nanostructure formation, which is approximately 10¹⁵/m².^17–19 The He plasma irradiation conditions are as follows: E_i is 75 eV, T_s is 800 K, and the He ion fluence, $Φ He$⁠, is in the range of 1 × 10²⁵–1 × 10²⁶ m⁻².

After plasma irradiation, the surface morphology analysis is carried out using a Hitachi SU8220 Field Emission Scanning Electron Microscope (FE-SEM). The structure of the grown cone is characterized by high-resolution transmission electron microscopy (HRTEM) using a JEM-2100PLUS Electron Microscope. The TEM sample is fabricated by focus ion beam (FIB) milling.

Figures 2(a)–2(c) illustrate the top-view FE-SEM images of the Si surfaces, depicting the pristine surface, the surface with Ta pre-deposition, and the surface post-irradiation by He plasma with Ta pre-deposition under conditions of E_i = 75 eV, T_s = 800 K, and Φ_He = 5 × 10²⁵ m⁻². Figures 2(d)–2(f) present the corresponding cross-sectional images. It is evident that the impact of He plasma irradiation on the Si surface in the presence of Ta leads to the formation of cone arrays. The irradiated surface is covered by densely distributed cones, exhibiting smooth sides and sharp tips. The typical separation distances of the cones are ∼90–320 nm. The fast-Fourier transform (FFT) analysis shown in the inset in Fig. 2 shows no distinct peaks, indicating that the cones are arranged irregularly. The cones extend perpendicular to the surface, oriented in the direction of incident He ions, which are accelerated by the electric field within the sheath.

Upon closer examination of the sample cross section under the electron microscope, we observe the appearance of cavities beneath some of the cones, arranged nearly vertically downward, as indicated in the inset in Fig. 2(f). These cavities correspond to He bubbles²⁰ generated during the high-fluence implantation of He ions into the Si lattice. It is worth noting that similar cone arrays have been produced by He plasma irradiation in conjunction with molybdenum (Mo) co-deposition,^13,14 albeit without the presence of these linearly aligned He bubbles. This discrepancy can be attributed to the employment of a higher surface temperature and He irradiation fluence in our study. This phenomenon may be related to the following processes. Different from the case that vacancies and interstitial atoms are directly formed by knock-on processes in energetic He ion beam bombardment, the He atoms can diffuse deeper into the lattice through interstitial sites in low-energy and high-flux He plasma irradiation. The high-flux (10²² m⁻² s⁻¹) plasma used in our study, the elevated surface temperature, and the final high He fluence (10²⁵ m⁻²) facilitate this process by abundant implanted He ions. Meanwhile, the implanted He atoms in the lattice tend to move to the tetrahedral interstitial site within the Si lattice.^21,22 Consequently, the He atoms tend to accumulate on the (110) plane during the diffusion process of He into deeper depth. Nevertheless, further experiments and simulations are needed to reveal the underlying mechanism.

Ever since Ozaydin's groundbreaking work,²³ both foundational experimental investigations^13,14,17,18 and theoretical studies²⁴ have established the critical role of impurities in surface morphology evolution during ion beam irradiation, ultimately leading to the creation of nanodots and, under specific conditions, cone arrays at elevated temperatures. As reported by Song et al.,¹¹ when subjected to 1500 eV Ar⁺ ion beam irradiation at 873 K in the presence of Ta co-deposition, the Si surface remains flat and devoid of features. This outcome sharply contrasts with our results, where cone arrays formed through He plasma irradiation. The divergence in the surface morphology arises from two distinct factors.

First, the sputtering yield of Ta when exposed to 75 eV He plasma irradiation is nearly negligible,¹⁶ falling below that of the Si sputtering yield. Consequently, pre-deposited Ta serves as a protective barrier for the Si substrate during He plasma irradiation. Conversely, in the 1500 eV Ar beam irradiation scenario in Ref. 11, the sputtering yield of Ta slightly exceeds that of the Si substrate. These differing preferential sputtering processes are accountable for the distinct surface transformations.

Second, the dissimilar behavior of noble gas ion species in these two cases also contributes to the contrast in cone formation. Implanted He atoms tend to aggregate into He bubbles within the Si lattice,²⁰ while Ar atoms have a propensity to depart from the Si substrate.²⁵ The presence of He bubbles induces lattice distortion, heightening the stress in the surface layer,²⁶ thus promoting the surface evolution toward cone formation by stress-induced surface diffusion.²⁷ 

Figures 3(a)–3(e) provide additional insights into the growth dynamics of the cones. The cross-sectional view of these five irradiated Si samples, each subjected to varying fluences, distinctly depicts cone array growth in length and width, which is clearly dependent on the irradiation fluence or, in other words, the plasma irradiation time. Upon a close examination of the overall cross sections of the samples, in Figs. 3(a)–3(c) at lower irradiation fluence levels, the substrate and interior of the cones seem to be smooth. However, as the irradiation fluence is 5 × 10²⁵ m⁻², the cavities become apparent beneath the cones, as indicated by the arrows in Figs. 3(d) and 3(e). Notably, the voids are present even at the base of the cones even at the highest fluence level of 1 × 10²⁶ m⁻² in our experiments, as demonstrated in an enlarged micrograph in Fig. 3(f). These imply that the He atoms accumulated in Si have a potential evolution process with the increase in the irradiation fluence. In low fluence conditions, the He atoms may exist mostly as clusters, at scales smaller than the resolution of electron microscopy.

The growth kinetic of the cones is explored in Fig. 4. The average lengths of the cones, measured from the cross-sectional images in Fig. 3, are plotted against the square root of the irradiation time in Fig. 4(a). The associated uncertainties represent the standard error (SE) of the measured length, reflecting variations in defining the cone tip and substrate surface. Notably, the cone lengths are well characterized by t^1/2 dependence, as demonstrated by the linear fit. The width of the cone base is also measured and plotted in Fig. 4(b). These two figures indicate that the cones are growing in all directions, including the length and the width, during He plasma irradiation. The concordance between the measured values and the fit suggests that the cone growth process is primarily governed by diffusion. During He plasma irradiation, a certain amount of Si atoms are sputtered by He ions impinging or ejected by the He bubbles from their original lattice position and then migrating toward the surface, forming adatoms lying on the surface. At the surface temperature in this research, the diffusion of Si adatoms becomes significant and is involved in the formation of cone arrays.

Assuming a simple one-dimensional growth law along the length direction of the cones, d = (2Dt)^1/2, arising from Fick's law, this linear relationship corresponds to a diffusion-driven growth process characterized by an effective diffusion coefficient of D = 6.49 ± 0.83 × 10⁻¹⁶ m⁻² s⁻¹. Furthermore, it is worth noting that the point where the straight-line fit intersects with the t^1/2 axis is not zero [12.74 s^1/2 in Fig. 4(a) and 13.51 s^1/2 in Fig. 4(b)], indicating the presence of an incubation fluence (Φ_He = 2.76–3.10 × 10²⁴ m⁻²) that may precede the cone growth. This incubation period is related to the early stage during the evolution from single He atoms, He clusters to He nanobubbles within the near surface. During this period, the aggregation form of He mostly existed in the form of clusters, and the evolution of surface morphology is also in an incubation stage, without the formation of nanostructures.

This rate-limited finding suggests that Si adatom migration toward the cone tips is likely the mechanism driving cone growth, a hypothesis supported by the transmission electron microscope (TEM) observations detailed in Sec. III C.

For gaining further insight into the fine structure and growth mechanism of the cones, the irradiated Si sample at a fluence of 5 × 10²⁵ m⁻² is characterized using HR-TEM. Figure 5 presents the cross-sectional HR-TEM images showing the different regions of the cones. The typical cone structure consisting of an inner and outer part with a clear interface is depicted in Fig. 5(a), which is akin to the redeposition layer previously discussed in Ref. 14 with continuous Mo co-deposition under He plasma irradiation. The cone structure exhibits a conical shell wrapping around a short inner cone. The fine structure of the cone is revealed locally using HR-TEM. As shown in Fig. 5(b), the base of the inner part is connected to the substrate. The crystalline nature of the cone is revealed by the 2-D FFT analysis shown in the inset. This growth of such an inner short cone can be explained by the sputtering erosion model. Figure 5(c) illustrates the interface between the two parts, while Fig. 5(d) presents the crystal structure in the middle of the outer part. The 2-D FFT analysis of Figs. 5(c) and 5(d) indicates different diffraction characteristics in the cone. In particular, the outer conical tip consists of nanocrystals with different crystallographic orientations. The analysis cone has a length of approximately 960 nm, and the outer conical shell is 655 nm long. The growth of such a long conical outer cone may be dominated by adatom diffusion toward the cone tip,^28,29 and the discussion in Ref. 30 suggests that the diffusion of surface atoms can become significant at the temperatures used in this study. This stage is governed by the competition between surface diffusion and sputtering erosion. This could be attributed to the combined effect of the two under the effect of metal impurity-enhanced low-energy and high-flux He plasma etching. Furthermore, the smooth surface of the cone is the result of the sputtering yield reduction during subsequent irradiation.³¹ The previous studies^6,10,14 indicate that the metal or metal silicified cone tip is commonly involved in the cone formation mechanism under plasma or ion beam irradiation, but the tip of the cone in Fig. 5 was buried in the platinum deposited during FIB milling, and, therefore, the distribution of Ta elements on the cone still needs further verification.

Based on the preceding discussion, a plausible mechanism, which is based on preferential sputtering,³¹ the destabilizing influence of He bubbles including stress-induced surface diffusion,²⁷ and adatom diffusion,^{28, 29} is proposed for the genesis of cone arrays on the Si surface under He plasma irradiation. Before irradiation, the pre-deposited Ta forms a discontinuous Ta layer with isolated Ta particles. During He plasma irradiation, Ta particles with a lower sputtering yield act as a protective barrier, shielding the underlying Si substrate from sputtering, while the adjacent Si region experiences erosion. This sustained sputtering erosion process initiates the formation of cone-shaped protrusions. Simultaneously, He atoms implanted into the Si lattice tend to aggregate, resulting in the formation of clusters and bubbles. The presence of these aggregated He species induces surface layer distortion and elevates lattice stress, resulting in stress-induced surface diffusion. With prolonged irradiation, additional Si adatoms accrue due to sputtering erosion and the continuous growth of He bubbles. The diffusion of these Si adatoms at elevated temperatures facilitates the progressive enlargement of the cones through a diffusion-driven process, culminating in the development of double-layer cone structures.

This study delves into the impact of 75 eV He plasma irradiation on Si (100) samples. To introduce Ta impurities onto the Si surface, we employ the magnetron sputtering method, known for its better control and quantifiability when compared to the conventional co-deposition approach. Throughout the irradiation process, the samples are maintained at a surface temperature of 800 K, with the fluence spanning from 1 × 10²⁵ to 1 × 10²⁶ m⁻². In contrast to the flat surface observed in the case of Ar beam irradiation, we observe the formation of cone arrays on the Si surface with the presence of He bubbles. The growth of these cones conforms to Fick's law, indicative of a diffusion process operating within a rate-limited regime. The presence of a double layer in the cones is attributed to a multifaceted growth process. Key mechanisms governing the formation of cone arrays include preferential sputtering, the destabilizing influence of He bubbles, and the migration of adatoms.

This work was supported by the Joint Funds of the National Natural Science Foundation of China (NSFC) (Grant No. U2267208). We acknowledge the USTC Center for Micro and Nanoscale Research and Fabrication for supporting the surface morphology analysis.

The authors have no conflicts to disclose.

Zhe Liu: Investigation (equal); Validation (equal); Writing – original draft (equal); Writing – review & editing (equal). Long Li: Investigation (supporting). Zeshi Gao: Investigation (supporting). Ze Chen: Writing – review & editing (equal). Chao Yin: Writing – review & editing (equal). Shifeng Mao: Funding acquisition (equal); Supervision (equal). Shin Kajita: Conceptualization (equal); Supervision (equal); Writing – review & editing (equal). Noriyasu Ohno: Conceptualization (equal); Supervision (equal); Writing – review & editing (equal). Minyou Ye: Conceptualization (equal); Data curation (equal); Funding acquisition (equal); Project administration (equal); Supervision (equal); Validation (equal); Writing – review & editing (equal).

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