Colloidal lithography (CL) is a promising method for large-area fabrication of nanohole and nanodot arrays with applications in optical biosensing, separations, and magnetic data storage. However, reducing the diameter of the polystyrene sphere mask by plasma etching unavoidably increases their coefficient of variation (CV) and deforms their shape, thereby limiting the pitch-to-hole-diameter ratio of the resulting nanohole array to less than 3:1 and the minimum hole size to 200 nm with a 10% or better CV. We show that tilt-rotate evaporation colloidal lithography (TRE-CL) breaks the trade-off between hole diameter and polydispersity by leveraging glancing angle evaporation, not plasma etching, to adjust the hole size. TRE-CL allows pitch-to-hole-diameter ratios as high as 7:1 and nanohole diameters down to 60 nm while maintaining a nearly constant CV below 10% and hole circularity above 91%. We transfer these hole arrays into ultrathin Si3N4 films to form nearly-monodisperse microsieves for separation applications. Furthermore, we extend TRE-CL to fabricate adhesion-layer-free plasmonic Au nanodot arrays down to 70 nm in diameter with 10% CV.

Fabricating nearly-monodisperse, periodic nanostructures over large areas is one of the grand challenges of nanotechnology.1–4 Colloidal and block copolymer lithography (CL and BCPL, respectively) are promising scalable, large-area nanofabrication techniques. CL utilizes the thermodynamically favored hexagonal close-packed arrangement of nearly monodisperse polystyrene spheres (PSs) to create large-area nanoscale shadow masks.5 Plasma etching transforms these close-packed PS assemblies into nonclose-packed sphere arrays.6 Metal film evaporation on the nonclose-packed sphere arrays creates a hole mask or nanohole array (NHA). Unfortunately, there are limitations in the fabrication of NHAs by CL. Plasma etching of the PSs can deform their shape;7 this deformation affects the final dimensions of the NHA. If the spheres are etched to less than half their original diameter, plasma-induced deformation of the spheres severely decreases the circularity and uniformity of the holes in the array.8–10 Plasma etch-induced deformation of the spheres ultimately limits the final hole diameter to 100 nm or larger.8 Etching PSs with initial diameters below 200 nm is a possible route to overcome this limitation but is challenging due to the rapid plasma-induced deformation of sub-200 nm spheres. Both the air- and substrate-facing sides of the spheres are deformed during the plasma etch step. The air-side of the spheres loses its sphericity and increases its surface roughness, while the substrate-facing side of the spheres is flattened at its contact point with the substrate. Flattening increases with cumulative plasma etch time. Many studies have attempted to optimize the plasma etching conditions to mitigate deformation of PSs but, in general, flattening of the substrate-facing side of the sphere still occurs.7 Furthermore, flattened PSs are often challenging to remove postmetal evaporation. The hole size then becomes limited by flattening with sub-200 nm spheres providing a minimum hole diameter of 100 nm.7 

BCPL is another promising method for fabricating large-area NHAs. BCPL utilizes different polymer blocks joined by covalent bonds.11,12 The polymers spontaneously self-assemble to create an ordered morphology on the surface of the substrate.1 The molecular weight of the polymer chain determines the sizes of the domains, which can be anywhere from 10 to 100 nm.1,11 BCPL features are determined by the interfacial energy between the different polymers and the entropic energy from chain stretching.13,14 Since the polymer chain length determines the domain size, the fabrication of larger holes requires longer polymer chains. If the polymers are too long, ordering is too slow and does not reach equilibrium, resulting in defects in the assembly.14 In practice, the hole size for BCPL is limited to below 50 nm.12,14

While there has been extensive research on improving the average diameter and defect density in CL and BCPL-derived NHAs, little work has been done in reducing the polydispersity of the hole diameters in the NHAs, especially in the sub-200 nm diameter range.5,10,15–20 This paper introduces a new approach to CL that generates NHAs featuring 50–200 nm diameter holes with a sub-10% coefficient of variation (CV), i.e., standard deviation (σ) / mean diameter. The CV of the nanohole diameter fabricated by current CL methods increases as the PS is etched because the diameter decreases while σ, defined by the original PS colloid, remains constant. In practice, this limits the pitch-to-hole-diameter ratio in CL to less than 3:1. Our approach removes this limitation by fixing the diameter of the PS mask and tuning the NHA diameter by varying the glancing angle of evaporation onto a rotating substrate. This remarkably allows us to achieve sub-200 nm holes and a pitch-to-hole-diameter ratio as high as 7:1 all while keeping the CV nearly constant. Both BCPL and CL methods cannot fabricate NHAs with 50–200 nm holes while maintaining near monodispersity of the hole diameter and the pitch of the array.

Tilt-rotate evaporation colloid lithography (TRE-CL) combines the self-assembly and plasma etching processes used in CL with metal evaporation performed on a tilted substrate rotating in a plane tangent to the substrate axes, i.e., tilt-rotate evaporation (TRE). This creates an NHA with hole diameters below 200 nm, CV < 10%, and hole circularity (o), i.e., the ratio of the shortest diameter to the longest diameter of the hole, as high as 98% (see Fig. S1 in the supplementary material).63 Unlike glancing angle deposition (GLAD),21–25 TRE-CL does not operate at ultralow glancing angles, ≤15°, nor rely on shadowing effects to create nanostructures. For TRE-CL, the NHA pitch remains fixed to the original sphere diameter while the glancing angle of evaporation determines the final hole diameter. Angled evaporation has been applied to CL in the past, but the focus was on generating new geometries.26–34 The use of TRE to reduce the hole diameter and the polydispersity of the holes has been overlooked.

Figure 1 plots the hole diameter and CV from NHAs made from CL, BCPL, and TRE-CL methods. Papers with NHAs of different diameters made each way are plotted against their CV. Each approach is grouped by color to show the different size regimes it can fabricate. CL and BCPL are limited in their ability to fabricate nearly monodisperse nanostructures. The BCPL method excels at manufacturing masks with hole diameters below 50 nm, while CL can only achieve sub-10% CV for hole diameters above 200 nm. Neither method can create hole masks with diameters in the 50–200 nm range while maintaining a CV below 10%. Overcoming the limitations of these approaches while keeping scalability will allow for the low-cost fabrication of periodic nanostructures with improved biosensing and separations capabilities.45,46 TRE-CL supports a low CV with high circularity, 0.91 < o < 0.98, while optimizing nanostructures for a size regime that proves difficult with current low-cost, scalable methods.

FIG. 1.

Hole diameter and the corresponding CV from 13 papers on CL—right region—and BCPL—left region (Refs. 1,35–44). Hole diameter and CV for the TRE-CL method reported herein are plotted in the center region. TRE-CL covers an area (center) unachievable by CL and BCPL methods. BCPL can fabricate <50 nm holes, and CL can achieve larger holes, i.e., >200 nm, with a low CV, i.e., <10%. TRE-CL covers the region between 50 and 200 nm diameter holes while maintaining a CV below 10%.

FIG. 1.

Hole diameter and the corresponding CV from 13 papers on CL—right region—and BCPL—left region (Refs. 1,35–44). Hole diameter and CV for the TRE-CL method reported herein are plotted in the center region. TRE-CL covers an area (center) unachievable by CL and BCPL methods. BCPL can fabricate <50 nm holes, and CL can achieve larger holes, i.e., >200 nm, with a low CV, i.e., <10%. TRE-CL covers the region between 50 and 200 nm diameter holes while maintaining a CV below 10%.

Close modal

Si wafers diced to 25 × 20 mm2 were used as the substrates for the NHAs. We cleaned each chip using a modified RCA method and then rinsed and stored them in type 1 18.2 MΩ cm water.47 Once ready to spin coat, we dried the substrates in N2 and centered them on the vacuum chuck of a Laurell spin coater. A 100 μl suspension of 425 ± 10 nm diameter PSs at 10 wt. % (Microparticles GmbH) was pipetted onto the surface of the substrate. Substrates were immediately spin-coated at 1750 rpm for 60 s to create hexagonal close-packed monolayers of the spheres. We etched the close-packed PS monolayers in an Oxford Instruments inductive coupled plasma (ICP)-100 instrument. The etching time was selected to reduce the diameter of the PSs by 20%, creating a nonclose-packed hexagonal array with a spacing of 50 nm between each sphere. The recipe combines ICP, and capacitance coupled plasma (CCP) using an N2/O2 plasma (P = 60 mTorr, 60 SCCM N2, 6 SCCM O2). The ICP power was 300 W, and the CCP power was 10 W. We confirmed the height, diameter, and flattening of the spheres with electron microscopy.

We mounted our substrates to a custom-built rotating stage held 4 cm off-center from the covered evaporation boat inside our vacuum chamber. An off-center position was only used due to geometry constraints within our vacuum chamber and was considered when calculating the glancing angle (α). Before metal evaporation began, the modified rotating stage was tilted to the desired glancing angle (α). To ensure a pointlike evaporation source and avoid spitting from Cr sublimation, we used a covered thermal evaporation source (S30A-.010 W boat, R.D. Mathis). A 20 nm film of Cr was evaporated at 2 Å/s while rotating the stage. After depositing the Cr film, we sonicated the sample in toluene for 10 min to remove the PSs and dried the substrate with an N2 gun.

All electron micrographs were collected with a cold-field emission Hitachi S4500 scanning electron microscope (SEM). We characterized the hole and dot diameters from the electron micrographs using the “Analyze Particles” function in imagej, finding the circularity and diameter for at least 1000 holes on each array. We calculated the CV of the holes and dots by dividing their standard deviation (σ) by their average diameter. Figure 1 analyzed comparative data from other papers with a minimum hole count (sample size) of 81.

The same colloidal lithography approach detailed above was used but on 25 × 20 mm2 Si chips coated with 50 nm of low-pressure chemical vapor deposition (LPCVD) Si3N4. Holes in the Cr film were transferred to the Si3N4 layer using an Oxford ICP 100. Oxygen plasma was used in the ICP 100 to remove any possible remaining PSs. A CCP etch using an N2/CF4 plasma (P = 10 mTorr, 5 SCCM N2, 35 SCCM CF4) with a power of 20 W was performed to transfer the holes in Cr to Si3N4. To remove Cr the wafer was placed in a Cr etchant [5 g ammonium cerium (IV) nitrate, 1.2 g acetic acid, 25 ml H2O] for 10 min. After the wafer was rinsed, it was cleaned with oxygen plasma in the ICP 100 for 5 min. 30 nm of Au was then deposited in a custom-built thermal evaporator. Scotch tape was then placed on the wafer and pulled off to remove excess Au on top of Si3N4. The wafer was then placed in the ICP 100, and Si3N4 was removed using the same etch chemistry to transfer the holes into the Si3N4 layer. Dot arrays were imaged in the SEM to confirm the size, pitch, and circularity.

Figures 2(a) and 2(b) sketch the metal evaporation setups for CL and TRE-CL, respectively. The sample stage is rotated about its axis in both arrangements, but in CL, it is normal to the evaporation flux. In contrast, in TRE-CL, the stage is tilted so that the evaporation flux lands on the surface at a glancing angle (α). The results of normal incident evaporation versus TRE on nonclose-packed PS arrays can be seen in Movie S1 and Movie S2 in the supplementary material.63 In short, normal incident evaporation used in CL results in a nearly direct transfer of the diameter of the PS to the diameter of the NHA [Fig. 2(c)]. Therefore, σ for the diameter of the etched PSs matches the original σ of the colloidal PSs. Typical σ for PSs is 10% of the diameter.48 While σ is a commonly reported value, CV is a better metric of the NHA since the performance of the array is determined by both the average hole diameter and the size distribution across the substrate. Plasma etching reduces the size of the PSs to the final desired hole diameter but simultaneously increases the CV of the holes due to the reduction of the mean diameter of the spheres. Plasma etching can also impart damage to the PSs, as discussed in the Introduction. Therefore, when using CL to fabricate <200 nm holes, the original size distribution of the PSs coupled with the long etch times results in a CV of the final NHA well above 10%, (Fig. 1).

FIG. 2.

Sketches of the metal evaporation setups for (a) CL and (b) TRE-CL methods. Diagrams of the geometry for (c) normal incidence and (d) TRE with a glancing angle of (α) around one PS. The resulting hole diameters are (d) for CL and (dTRE) for TRE-CL. Graphs of the calculated CV vs. the diameter of the hole mask fabricated by (e) traditional CL and (f) TRE-CL methods.

FIG. 2.

Sketches of the metal evaporation setups for (a) CL and (b) TRE-CL methods. Diagrams of the geometry for (c) normal incidence and (d) TRE with a glancing angle of (α) around one PS. The resulting hole diameters are (d) for CL and (dTRE) for TRE-CL. Graphs of the calculated CV vs. the diameter of the hole mask fabricated by (e) traditional CL and (f) TRE-CL methods.

Close modal

TRE-CL circumvents these limitations by leveraging TRE to adjust the final hole diameter, thereby avoiding long plasma etching times, maximizing the size of the PSs, and minimizing plasma damage. Plasma etching is only used to convert the PS monolayer from a close-packed array to a nonclose-packed array. We reduced the initial PS diameter by ∼20% for all TRE-CL experiments. We choose 20% because it allows for enough separation between spheres to prevent shadowing of evaporation flux from first-neighbors and minimizes flattening of the spheres at the substrate surface. We then leverage TRE to provide precise control over the hole mask diameter by varying the glancing angle, α. Figure 2(d) illustrates how the combination of the sample stage rotating along its axis and an evaporation flux at a glancing angle allows the metal to deposit under the PSs, thereby reducing the hole diameter from d to dTRE, where dTRE = c − 2τ. c is the chord of the PS, τ=h/tan(α), and h is the height of the chord. Decreasing α allows the evaporation flux increased access under the PS, further reducing dTRE. A minimum value for α exists, avoiding first-neighbor PSs shadowing the incoming flux.

TRE-CL also improves the CV of the NHA diameter compared to the standard CL method. Even if we assume negligible damage during plasma etching, the CV of the PSs increases when their diameter decreases [Fig. 2(e)]. This increase in CV is unavoidable and limits the performance of CL fabricated NHAs. Below, we demonstrate through simple geometry the ability of TRE-CL to shrink the NHA diameter while holding CV constant [Fig. 2(f)].

Starting with dTRE = c − 2τ and evaluating c, τ, and h and in terms of α and d, we develop Eq. (1) [see SI for the derivation of Eq. (1)],

dTRE(α)=dtan(α/2).
(1)

Equation (1) demonstrates how decreasing the glancing angle (α) of evaporation forms a hole with a diameter (dTRE) that is smaller than the PS diameter (d) after etching because tan(α/2) is less than 1 for all possible glancing angles of evaporation, (e.g., 0° < α < 90°). To better understand how TRE improves the CV of the NHA, we imagine two PSs, one with diameter (d) and the other with (d + σ), where (σ) represents the original standard deviation of the colloidal PSs. Substituting (d + σ) for (d) in Eq. (1) gives the resulting hole diameter (dTRE + σTRE) generated from TRE at a glancing angle (α) of evaporation on a PS with a diameter (d + σ), shown below in Eq. (2),

(dTRE+σTRE)(α)=(d+σ)tan(α/2).
(2)

Subtracting Eq. (1) from Eq. (2) directly relates σTRE and σ giving Eq. (3),

σTRE(α)=σtan(α/2),
(3)

where σ is a constant value given by the original size distribution of the colloidal PSs. Therefore, Eq. (3) shows that σTRE is a function of α and always less than σ because, as previously noted, tan(α/2) is <1 for all glancing angles (α). As α → 0°, σTRE → 0, but shadowing from adjacent PS limits the glancing angle (α) to more than 0°. Finally, to determine the CV for TRE-CL, we divide Eq. (3) by Eq. (1) and cancel tan(α/2), giving us the constant

CVTRE=σ/d.
(4)

This illustrates the power of TRE-CL to overcome the limitations of conventional CL methods and fabricate NHAs with low CV values that are independent of the final hole diameter (dTRE).

To demonstrate the advantage TRE-CL offers over the standard CL method for fabricating sub-200 nm NHA, we used both methods to create a ∼100 nm diameter NHA with a 425 nm pitch. Figure 3 displays electron micrographs of the process and results for TRE-CL (a) and (b) and CL (c) and (d) fabricated 100 nm diameter NHAs. Both methods start with a close-packed array of 425 nm PS. Since TRE-CL uses the glancing angle during TRE, not plasma etching, to adjust the final hole size, we only reduced the PSs diameter by ∼20% during the plasma step. This short etch step yielded the nonclose-packed array of 350 nm PS [Fig. 3(a)]. We found that a 20% reduction in the size of the spheres resulted in a good balance between avoiding shadowing from adjacent spheres during TRE and mitigating plasma damage to the spheres. We then performed TRE evaporation at a glancing angle of 40° on the nonclose-packed 350 nm PSs. Figure 3(b) is an electron micrograph of the resulting NHA with 80 nm holes and a 425 nm pitch. We also attempted to make the ∼100 nm diameter NHA using CL with normal incident evaporation. As illustrated in Fig. 2(c), CL with normal incident evaporation requires the PS to be etched to the final desired hole diameter. Figure 3(c) is an electron micrograph of the PS etched with a combination of plasma and UV ozone to impart minimal etch damage.49 We found that a single plasma etching step left behind only remnants of the PS. Still, despite the gentler sequential etch process, the circularity of the spheres is poor at 76.9%, and the CV is 15.4%, well above 10%. The large shadow of the initial PS diameter is due to plasma-induced damage to the substrate.8 Unfortunately, normal incident evaporation of 20 nm of Cr followed by PS removal by sonication in toluene did not form an NHA. The electron micrograph in Fig. 3(d) shows that the etched PS becomes covered in the Cr mask, trapping the PS underneath. Thinner Cr layers could be attempted to aid in removing the PS, but this would still not overcome the high CV and low circularity of the PS due to etch-induced damage. Figure 3 demonstrates that TRE-CL is superior to the standard CL method for making <100 nm diameter holes. Furthermore, TRE-CL allows facile tuning of the hole diameter by adjusting the glancing angle of evaporation.

FIG. 3.

Electron micrographs of etched PSs (a) and (c) and resulting 100 nm diameter NHAs (b) and (d) fabricated by TRE-CL (top) and CL (bottom). (a) Nonclose-packed array of 350 nm diameter PS after initial short plasma etching of 425 nm PS. (b) 80 nm hole array after TRE of Cr on (a) at a glancing angle of 40°. (c) Original close-packed 425 nm PS etched to ∼100 nm showing significant etch damage. (d) Normal incident evaporation of Cr on (c) results in a failed nanohole array. The PS from (c) could not be released post-Cr evaporation. All scale bars are 500 nm.

FIG. 3.

Electron micrographs of etched PSs (a) and (c) and resulting 100 nm diameter NHAs (b) and (d) fabricated by TRE-CL (top) and CL (bottom). (a) Nonclose-packed array of 350 nm diameter PS after initial short plasma etching of 425 nm PS. (b) 80 nm hole array after TRE of Cr on (a) at a glancing angle of 40°. (c) Original close-packed 425 nm PS etched to ∼100 nm showing significant etch damage. (d) Normal incident evaporation of Cr on (c) results in a failed nanohole array. The PS from (c) could not be released post-Cr evaporation. All scale bars are 500 nm.

Close modal

The trend of dTRE versus glancing angle (α) for various pitches (colored lines) is calculated using Eq. (1) and plotted in Fig. 4(a). We set the pitch and PS diameter (d) equal to the initial unetched PS diameter and the etched (e.g., 20% reduced) sphere diameter, respectively. To confirm that the measured and calculated hole diameters (dTRE) match, we made NHA at multiple glancing angles (α). Electron micrographs of these hole arrays are displayed in Figs. 4(b)4(f). The initial PS diameter was 425 nm, which sets the pitch of these arrays. We first etched the close-packed PS array, reducing their diameters from 425 to 350 nm (e.g., ∼20% reduction) to form a nonclose-packed array. We then evaporate a chromium hard mask on the surface at a given glancing angle (α) while the sample rotates about its tilted axis. We analyzed a minimum of 1000 holes from each NHA in Figs. 4(b)4(f) to determine the average hole diameter of each sample within a 95% confidence level. The analysis process can be seen in Fig. S2 in the supplementary material.63 

FIG. 4.

Calculations for possible NHA dimensions by TRE and corresponding electron micrographs of hole masks at select glancing angles (α) of evaporation. (a) Calculated NHA diameters (dTRE) possible by TRE as a function of glancing angle (α) for various pitches (colored lines) of the array. (b)–(f) Electron micrographs of NHAs fabricated with 425 nm PSs at different glancing angles (α). All PS arrays were initially etched to a diameter of ∼350 nm before TRE was performed. (b) 352 ± 20 nm holes deposited at α = 82°, (c) 242 ± 20 nm holes deposited at α = 73°, (d) 132 ± 11 nm holes deposited at α = 53°, (e) 111 ± 4 nm holes deposited at α = 40°, and (f) 60 ± 5 nm holes deposited at α = 31°. Scale bar in (b)–(f) is 500 nm.

FIG. 4.

Calculations for possible NHA dimensions by TRE and corresponding electron micrographs of hole masks at select glancing angles (α) of evaporation. (a) Calculated NHA diameters (dTRE) possible by TRE as a function of glancing angle (α) for various pitches (colored lines) of the array. (b)–(f) Electron micrographs of NHAs fabricated with 425 nm PSs at different glancing angles (α). All PS arrays were initially etched to a diameter of ∼350 nm before TRE was performed. (b) 352 ± 20 nm holes deposited at α = 82°, (c) 242 ± 20 nm holes deposited at α = 73°, (d) 132 ± 11 nm holes deposited at α = 53°, (e) 111 ± 4 nm holes deposited at α = 40°, and (f) 60 ± 5 nm holes deposited at α = 31°. Scale bar in (b)–(f) is 500 nm.

Close modal

Table I reports the glancing angle (α), predicted hole diameter from Eq. (1), average measured hole diameter, standard deviation (σ), and CV for the samples. As expected, σ decreases with the glancing angle (α) while CV stays nearly constant. The CV increases slightly at 31° due to the flattening of the PSs, limiting the size of the hole and increasing σ. That said, the measured hole diameters in Table I show excellent agreement with the calculated hole diameters in Fig. 4(a) for a pitch of 425 nm and a maximum pitch-to-hole-diameter ratio of 7:1.

TABLE I.

Numerical values of calculated and actual diameter for each glancing angle used along with the standard deviation and CV at each angle.

Glancing angle (°)Calculated diameter (nm)Actual diameter (nm)σ (nm)CV (%)
82 304 352 20 5.7 
73 258 242 20 8.3 
53 174 132 11 8.3 
40 127 111 3.6 
31 97 60 8.3 
Glancing angle (°)Calculated diameter (nm)Actual diameter (nm)σ (nm)CV (%)
82 304 352 20 5.7 
73 258 242 20 8.3 
53 174 132 11 8.3 
40 127 111 3.6 
31 97 60 8.3 

While smaller glancing angles (α) lead to smaller hole diameters, α cannot go below the point where the PS shadow the evaporation flux. The model in Fig. 2 assumes the PSs stay perfectly spherical post etchings, but it was seen experimentally that the plasma etch does induce a slight ellipsoidal shape to the PSs (Fig. S3 in the supplementary material).63 This ellipsoidal shape allows for lower glancing angles to be achieved. Atomic force microscopy (AFM) scans for glancing angles of 73°, 53°, and 40° were also performed to determine what shadowing effects were present (Fig. S3 in the supplementary material).63 The AFM of the NHA made with α = 73° shows lines where the Cr mask is thinner due to the shadowing effects seen from the first nearest neighbors. Where there are no neighboring PSs, there is no variation in the Cr mask thickness. As α is decreased, the shadowing increases, as seen with the wider areas of thinner Cr in the AFM of the NHA made at α = 53°. At even lower angles, shadowing from all neighbors creates a more uniform Cr mask thickness as seen with the AFM of the NHA made at α = 40°. For the 425 nm PS etched to 350 nm, we found the glancing angle (α) needed to be greater than or equal to 74° to avoid shadowing. The angle where shadowing begins depends on both the height and pitch of the PSs at the time of evaporation. As the initial diameter of the PS increases, the minimum glancing angle to avoid shadowing is expected to gradually decrease unless the PS is etched greater than 20%. When TRE was performed at a glancing angle α = 23°, the Cr layer was removed entirely when the PS mask was sonicated off, as shown in Fig. S4 in the supplementary material,63 indicating complete shadowing of the Cr flux.

Importantly, NHAs fabricated by TRE-CL can be transferred to other materials. To demonstrate this, we used reactive ion etching (RIE) to transfer the holes from the Cr layer into a 50 nm thick Si3N4 layer grown on top of the Si substrate (Fig. 5). Si3N4 membranes have shown promise in separation processes and can function as microsieves for wastewater, crude oil, and bacteria filtration in the 50–200 nm hole diameter range.50 

FIG. 5.

TRE-CL formed NHAs transferred into 50 nm Si3N4 films on Si supports. The two hole sizes are created from different glancing angles of evaporation. (a) has an average diameter of 117.44 ± 8.85 nm and (b) has an average diameter of 73.3 ± 6.84 nm. The scale bar is 500 nm.

FIG. 5.

TRE-CL formed NHAs transferred into 50 nm Si3N4 films on Si supports. The two hole sizes are created from different glancing angles of evaporation. (a) has an average diameter of 117.44 ± 8.85 nm and (b) has an average diameter of 73.3 ± 6.84 nm. The scale bar is 500 nm.

Close modal

Another nanostructure we fabricated via TRE-CL is nanodot arrays (NDAs).51,52 NDAs have garnered significant industrial and academic interest for biosensing, especially when surface lattice resonances (SLR) are present.53,54 The spectra from SLRs have extremely sharp peaks with a full-width half maximum (FWHM) of only a few nanometers allowing for small shifts in the resonance peak to be evident.53 There is a desire to decrease the diameter of the dots in the array to narrow the SLR peak further.55 Diameter is one factor that can enhance the SLR, but to increase sensitivity, the NDA must be tuned for pitch and diameter while keeping the particles nearly monodisperse.55,56 Current, low-cost, scalable methods for creating 50–200 nm diameter NDAs suffer from a diameter with high polydispersity and cannot achieve the pitch necessary, i.e., >400 nm, to achieve SLR in the biological window (e.g., 700–900 nm), which limits their performance as biological sensors.55,57 Furthermore, Cr and Ti adhesion layers are a significant source of damping in current plasmonic NDAs.58–60Figure 6 displays our extended TRE-CL process for fabricating adhesion-layer-free plasmonic NDAs along with electron micrographs of two different diameters NDAs formed. First, the NHAs formed in Cr are transferred into a 50 nm Si3N4 layer using RIE, as demonstrated in Fig. 5. Cr is then removed to expose a hole array in the Si3N4 layer. The sample is then cleaned using a modified RCA method,47 and 30 nm of Au is evaporated onto Si3N4. Due to the weak adhesion between Si3N4 and Au,61 the excess Au layer on top of Si3N4 is removed with tape. The remaining Si3N4 is removed with RIE, leaving a uniform Au dot array with no adhesion layer. We used this technique to create Au dot arrays on silicon with 120 and 77 nm diameter dots [Figs. 5(b) and 5(c)]. Both dot arrays maintained a low CV and high circularity of ∼8.5% and ∼92%, respectively. Significantly, this approach can be extended to fused silica (FS) substrates to facilitate optical transmission. Figure S5 in the supplementary material63 shows 332 nm dots on FS substrates. The surface roughness of nanodots can also be improved via mild substrate heating (e.g., 400 °C for 2 min)62; see Fig. S6 in the supplementary material.63 

FIG. 6.

Fabrication of adhesion-layer-free plasmonic dot arrays via extended TRE-CL approach. (a) schematic of the dot array fabrication process. (b) Electron micrograph of 120 ± 10 nm Au dots on silicon. (c) Electron micrographs of 77 ± 7 nm Au dots on silicon. Adhesion-layer-free Au dot arrays can also be formed on FS substrates; see Fig. S5 in the supplementary material.63 Scale bar in (b) and (c) is 500 nm.

FIG. 6.

Fabrication of adhesion-layer-free plasmonic dot arrays via extended TRE-CL approach. (a) schematic of the dot array fabrication process. (b) Electron micrograph of 120 ± 10 nm Au dots on silicon. (c) Electron micrographs of 77 ± 7 nm Au dots on silicon. Adhesion-layer-free Au dot arrays can also be formed on FS substrates; see Fig. S5 in the supplementary material.63 Scale bar in (b) and (c) is 500 nm.

Close modal

In summary, we developed and tested a new method, TRE-CL, for fabricating NHAs with sub-200 nm diameter holes and CVs < 10%. Current methods for NHA fabrication are either fundamentally unable to reach the 50–200 nm diameter hole regime (e.g., BCPL) or result in holes with CVs > 10%. Furthermore, CL methods require long etch times to shrink the PSs, increasing their CV and degrading their circularity via plasma damage. TRE-CL resolves both problems by fixing the diameter of the PS mask and tuning the NHA diameter by varying the glancing angle of evaporation onto a rotating substrate. Furthermore, we demonstrate that TRE-CL can be extended to form adhesion-layer-free plasmonic NDAs on silicon and fused silica substrates. TRE-CL is a versatile approach that fills a critical need to fabricate 50–200 nm diameter NHA and NDAs over an extensive range of pitches.

The authors acknowledge the Louisiana Board of Regents for their financial support for M.J.C. through a graduate research fellowship. The authors also thank Sergi Lendinez and the LSU Nanofabrication Facility for assistance with collecting AFM scans of the nanohole arrays. The authors also thank Andrey Krayev for his assistance in selecting the most appropriate tip for the AFM scans.

The authors have no conflicts to disclose.

MaCayla J. Caso: Formal analysis (lead); Investigation (lead); Writing – original draft (lead). Michael G. Benton: Supervision (supporting); Writing – review & editing (supporting). Kevin M. McPeak: Conceptualization (lead); Resources (lead); Supervision (lead); Writing – review & editing (lead).

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

1.
M.
Park
,
C.
Harrison
,
P. M.
Chaikin
,
R. A.
Register
, and
D. H.
Adamson
,
Science
276
,
1401
(
1997
).
2.
A.
Patoux
,
G.
Agez
,
C.
Girard
,
V.
Paillard
,
P. R.
Wiecha
,
A.
Lecestre
,
F.
Carcenac
,
G.
Larrieu
, and
A.
Arbouet
,
Sci. Rep.
11
,
5620
(
2021
).
3.
M. D.
Bishop
 et al,
Nat. Electron.
3
,
492
(
2020
).
4.
A.
Biswas
,
I. S.
Bayer
,
A. S.
Biris
,
T.
Wang
,
E.
Dervishi
, and
F.
Faupel
,
Adv. Colloid Interface
170
,
2
(
2012
).
5.
R.
McGorty
,
J.
Fung
,
D.
Kaz
, and
V. N.
Manoharan
,
Mater. Today
13
,
34
(
2010
).
6.
H. W.
Deckman
,
J. Vac. Sci. Technol. B
1
,
1109
(
1983
).
7.
A.
Plettl
,
F.
Enderle
,
M.
Saitner
,
A.
Manzke
,
C.
Pfahler
,
S.
Wiedemann
, and
P.
Ziemann
,
Adv. Funct. Mater.
19
,
3279
(
2009
).
8.
C.
Brombacher
 et al,
Nanotechnology
20
,
105304
(
2009
).
9.
C.
Haginoya
,
M.
Ishibashi
, and
K.
Koike
,
Appl. Phys. Lett.
71
,
2934
(
1997
).
10.
G.
Zhang
,
D.
Wang
, and
H.
Möhwald
,
Nano Lett.
7
,
3410
(
2007
).
11.
Y.-C.
Tseng
and
S. B.
Darling
,
Polymers
2
,
470
(
2010
).
12.
C. M.
Bates
,
M. J.
Maher
,
D. W.
Janes
,
C. J.
Ellison
, and
C. G.
Willson
,
Macromolecules
47
,
2
(
2014
).
13.
Y.
Mai
and
A.
Eisenberg
,
Chem. Soc. Rev.
41
,
5969
(
2012
).
14.
F. S.
Bates
and
G. H.
Fredrickson
,
Annu. Rev. Phys. Chem.
41
,
525
(
1990
).
15.
J. C.
Hulteen
and
R. P.
Van Duyne
,
J. Vac. Sci. Technol. A
13
,
1553
(
1995
).
16.
S.
Reynaert
,
P.
Moldenaers
, and
J.
Vermant
,
Langmuir
22
,
4936
(
2006
).
17.
K.
Shinotsuka
,
Y.
Kajita
,
K.
Hongo
, and
Y.
Hatta
,
Langmuir
31
,
11452
(
2015
).
18.
J. M.
Shin
 et al,
Chem. Mater.
31
,
1066
(
2019
).
19.
J. M.
Shin
,
M. P.
Kim
,
H.
Yang
,
K. H.
Ku
,
S. G.
Jang
,
K. H.
Youm
,
G.-R.
Yi
, and
B. J.
Kim
,
Chem. Mater.
27
,
6314
(
2015
).
20.
21.
M. M.
Hawkeye
,
M. T.
Taschuk
, and
M. J.
Brett
,
Glancing Angle Deposition of Thin Films
(
Wiley
,
New York
,
2014
).
22.
M. M.
Hawkeye
and
M. J.
Brett
,
J. Vac. Sci. Technol. A
25
,
1317
(
2007
).
23.
S.
Sánchez
,
L.
Soler
, and
J.
Katuri
,
Angew. Chem. Int. Ed.
54
,
1414
(
2015
).
24.
A. G.
Mark
,
J. G.
Gibbs
,
T.-C.
Lee
, and
P.
Fischer
,
Nat. Mater.
12
,
802
(
2013
).
25.
26.
S.
Syrenova
,
C.
Wadell
, and
C.
Langhammer
,
Nano Lett.
14
,
2655
(
2014
).
27.
Z.
Yan
,
P.
Gu
,
W.
Bao
,
W.
Du
,
Z.
Chen
,
X.
Xia
, and
Z.
Wang
,
Plasmonics
10
,
1159
(
2015
).
28.
Z.
Yan
,
X.
Wen
,
P.
Gu
,
H.
Zhong
,
P.
Zhan
,
Z.
Chen
, and
Z.
Wang
,
Nanotechnology
28
,
475203
(
2017
).
29.
M.
Samadi
,
S.
Vasini
,
S.
Darbari
,
A. A.
Khorshad
,
S. N. S.
Reihani
, and
M. K.
Moravvej-Farshi
,
Opt. Express
27
,
14754
(
2019
).
30.
C. L.
Haynes
and
R. P.
Van Duyne
,
Nano Lett.
3
,
939
(
2003
).
31.
A. V.
Whitney
,
B. D.
Myers
, and
R. P.
Van Duyne
,
Nano Lett.
4
,
1507
(
2004
).
32.
M. C.
Gwinner
,
E.
Koroknay
,
L.
Fu
,
P.
Patoka
,
W.
Kandulski
,
M.
Giersig
, and
H.
Giessen
,
Small
5
,
400
(
2009
).
33.
J.
Pan
,
Z.
Chen
,
Z.-D.
Yan
,
Z.-S.
Cao
,
P.
Zhan
,
N.-B.
Ming
, and
Z.-L.
Wang
,
AIP Adv.
1
,
042114
(
2011
).
34.
C.-J.
Hsu
,
K.-P.
Chiu
,
J.-T.
Lue
, and
K.
Yao Lo
,
Appl. Phys. Lett.
98
,
261107
(
2011
).
35.
S.
Han
,
Z.
Hao
,
J.
Wang
, and
Y.
Luo
,
J. Vac. Sci. Technol. B
23
,
1585
(
2005
).
36.
H.
Ni
,
M.
Wang
,
H.
Hao
, and
J.
Zhou
,
Nanotechnology
27
,
225301
(
2016
).
37.
S. H.
Lee
,
K. C.
Bantz
,
N. C.
Lindquist
,
S.-H.
Oh
, and
C. L.
Haynes
,
Langmuir
25
,
13685
(
2009
).
38.
J.
Wang
,
G.
Duan
,
Y.
Li
,
G.
Liu
, and
W.
Cai
,
ACS Appl. Mater. Interface
6
,
9207
(
2014
).
39.
K.
Brassat
,
S.
Ramakrishnan
,
J.
Bürger
,
M.
Hanke
,
M.
Doostdar
,
J. K. N.
Lindner
,
G.
Grundmeier
, and
A.
Keller
,
Langmuir
34
,
14757
(
2018
).
40.
R.
Tiron
 et al,
J. Vac. Sci. Technol. B
29
,
06F206
(
2011
).
41.
H. J.
Park
,
M.-G.
Kang
, and
L. J.
Guo
,
ACS Nano
3
,
2601
(
2009
).
42.
K.
Brassat
,
D.
Kool
,
J.
Bürger
, and
J. K. N.
Lindner
,
Nanoscale
10
,
10005
(
2018
).
43.
S.
Krishnamoorthy
,
R.
Pugin
,
J.
Brugger
,
H.
Heinzelmann
,
A. C.
Hoogerwerf
, and
C.
Hinderling
,
Langmuir
22
,
3450
(
2006
).
44.
M.
Jung
,
S.
Lee
,
Y.
Tae Byun
,
Y.
Min Jhon
,
S.
Ho Kim
,
D.-H.
Woo
, and
S.-i.
Mho
,
Microelectron. J.
39
,
526
(
2008
).
45.
H. D.
Tong
,
H. V.
Jansen
,
V. J.
Gadgil
,
C. G.
Bostan
,
E.
Berenschot
,
C. J. M.
Van Rijn
, and
M.
Elwenspoek
,
Nano Lett.
4
,
283
(
2004
).
46.
S. P.
Adiga
,
C.
Jin
,
L. A.
Curtiss
,
N. A.
Monteiro-Riviere
, and
R. J.
Narayan
,
WIREs. Nanomed. Nanobiotechnol.
1
,
568
(
2009
).
47.
S. K.
Sahari
,
J. C. H.
Sing
, and
K. A.
Hamid
,
Int. J. Electr. Comput. Eng.
3
,
262
(
2009
).
48.
See: https://microparticles.de/en/product-palette#Area15 for Microparticles GmbH, Product Palette.
49.
K. M.
McPeak
 et al,
Nano Lett.
14
,
2934
(
2014
).
50.
C. J. M. v.
Rijn
,
Nano and Micro Engineered Membrane Technology
(
Elsevier
,
New York
,
2004
).
51.
D.-G.
Choi
,
S.
Kim
,
S.-G.
Jang
,
S.-M.
Yang
,
J.-R.
Jeong
, and
S.-C.
Shin
,
Chem. Mater.
16
,
4208
(
2004
).
52.
J.-R.
Jeong
,
S.
Kim
,
S.-H.
Kim
,
J. A. C.
Bland
,
S.-C.
Shin
, and
S.-M.
Yang
,
Small
3
,
1529
(
2007
).
53.
V. G.
Kravets
,
A. V.
Kabashin
,
W. L.
Barnes
, and
A. N.
Grigorenko
,
Chem. Rev.
118
,
5912
(
2018
).
54.
S.
Kasani
,
K.
Curtin
, and
N.
Wu
,
Nanophotonics
8
,
2065
(
2019
).
55.
P.
Offermans
,
M. C.
Schaafsma
,
S. R. K.
Rodriguez
,
Y.
Zhang
,
M.
Crego-Calama
,
S. H.
Brongersma
, and
J.
Gómez Rivas
,
ACS Nano
5
,
5151
(
2011
).
57.
H.
Li
,
X.
Luo
,
C.
Du
,
X.
Chen
, and
Y.
Fu
,
Sens. Actuators, B
134
,
940
(
2008
).
58.
M.
Todeschini
,
A.
Bastos Da Silva Fanta
,
F.
Jensen
,
J. B.
Wagner
, and
A.
Han
,
ACS Appl. Mater. Inter.
9
,
37374
(
2017
).
59.
D. T.
Debu
,
P. K.
Ghosh
,
D.
French
, and
J. B.
Herzog
,
Opt. Mater. Express
7
,
73
(
2017
).
60.
S. J.
Madsen
,
M.
Esfandyarpour
,
M. L.
Brongersma
, and
R.
Sinclair
,
ACS Photonics
4
,
268
(
2017
).
61.
K. M.
McPeak
 et al,
Adv. Mater.
27
,
6244
(
2015
).
62.
C.
Thompson
,
Annu. Rev. Mater. Res.
42
,
399
(
2012
).
63.
See supplementary material at https://www.scitation.org/doi/suppl/10.1116/6.0001874 for supplementary figures, optimization methods, Eq. (1) derivation, and rendered traditional CL and TRE-CL videos.

Supplementary Material