This content has been downloaded from IOPscience. Please scroll down to see the full text.

Download details:

IP Address: 134.151.40.2

This content was downloaded on 15/01/2014 at 09:34

Please note that terms and conditions apply.

Rapid fabrication of nano-structured quartz stamps

View the table of contents for this issue, or go to the journal homepage for more

2013 Nanotechnology 24 055304

(http://iopscience.iop.org/0957-4484/24/5/055304)

Home Search Collections Journals About Contact us My IOPscience

IOP PUBLISHING NANOTECHNOLOGY

Nanotechnology 24 (2013) 055304 (10pp) doi:10.1088/0957-4484/24/5/055304

Rapid fabrication of nano-structuredquartz stamps

Yindar Chuo1, Clint Landrock1, Badr Omrane1, Donna Hohertz2,Sasan V Grayli1, Karen Kavanagh2 and Bozena Kaminska1

1 Department of Engineering Science, Simon Fraser University, 8888 University Drive, Burnaby,BC, V5A 1S6, Canada2 Department of Physics, Simon Fraser University, 8888 University Drive, Burnaby, BC, V5A 1S6,Canada

E-mail: yychuo@sfu.ca

Received 20 October 2012, in final form 8 December 2012Published 16 January 2013Online at stacks.iop.org/Nano/24/055304

AbstractReplication of surface nano-structures from a master stamp allows large-area volumeproduction that is otherwise cost prohibitive through conventional direct-write methods, suchas electron beam lithography and focused ion beam milling. However, the creation of a masterstamp containing sophisticated patterns still requires significant time on such direct-writetools. We demonstrate a method for reducing this tool time by patterning smallernano-structures, and then enlarging them to the desired size through isotropic etching. We cancreate circular structures of any arrangement and size, down to the patterning limits of thedirect-write tools. Subsequent metal mask deposition, lift-off, and anisotropic etchingtransforms the circular patterns to out-of-plane pillar structures for the final stamp. A 1 cm2

area filled with a pattern of 200 nm diameter nano-holes spaced 520 nm apart, requires only21 h to complete using our process, compared to 75 h using conventional fabrication. Wedemonstrate the utility and practicality of the quartz stamps through polymer embossing andreplication. Embossed polymer nano-hole arrays are coated with a Cr/Au (5/100 nm) film tocreate surface plasmon resonance structures. Extraordinary optical transmission spectra fromthe metallized arrays show the expected spectral features when compared to focused ion beammilled structures.

S Online supplementary data available from stacks.iop.org/Nano/24/055304/mmedia

(Some figures may appear in colour only in the online journal)

1. Introduction

Engineered surface nano-structures have exciting applicationsin many technologies, such as bio-chemical sensing [1–3],lighting [4, 5], beam shaping [6], colour displays [7],photovoltaics [8, 9], thermal energy [10], robotics andbiomimetics [11], among many others. However, conventionalscanning methods of direct-write fabrication of these arrangedsurface nano-structures are often too costly and timeconsuming for practical, realistic, applications outside ofthe laboratory environment. An application such as abio-chemical sensing lab-on-a-chip would require production

of surface nano-structure elements over relatively large areasand volumes, at significantly lower costs than currentlyfeasible [9].

Several techniques for fabricating large-area, large-volume, surface nano-structures have attracted much attentionover recent years. The classes of techniques includereplication, self-assembly, and flood-pattern direct-writing.Replication techniques, such as nano-imprinting [12–14],nano-embossing [15], various forms of casting [16], contactprinting [17], and phase shift lithography [18], all require amaster stamp or template. Fabrication of the master stamp, if itis to contain patterns of several geometries and arrangements,

10957-4484/13/055304+10$33.00 c© 2013 IOP Publishing Ltd Printed in the UK & the USA

Nanotechnology 24 (2013) 055304 Y Chuo et al

Figure 1. Conceptual comparison of scan versus flood-patterned nano-structures in an integrated device. Versatile, multiple-period andmultiple-geometry arrays, created by scanning direct-write tools such as EBL or FIB (left). Limited, single period continuous arrays,created by flood-pattern direct-write tools such as LIL (right).

is costly and time consuming. The stamp fabrication islimited by the throughput of the direct-write tools, such aselectron beam lithography (EBL) or focused ion beam (FIB).Self-assembly techniques, such as colloidal-based surfacenano-structures [19, 20] are often prone to propagationerrors during assembly, and limited to a single periodicrepeat pattern. Flood-patterning techniques, such as laserinterference lithography (LIL) are efficient and accurate [21,22]; however, the geometry and layout of the nano-structuresare dictated by the laser interference patterns attainable,and are again limited to a single periodic pattern. Figure1 illustrates the difference between nano-structure patternsgenerated by scanning versus flood-pattern direct-write tools.EBL or FIB scanning-type direct-write tools can createnano-structure arrays with various geometries arranged withdifferent spacing (periods) and orientations, while LILflood-pattern direct-write tool can create only a singlecontinuous pattern with a single period and geometry.

In bio-chemical sensing chips based on surface plasmonresonance (SPR) nano-hole arrays [9], it is useful to havemultiple arrays of nano-hole structures (e.g. groups spanning200 µm × 200 µm) with different hole-to-hole spacing(periodicities) in each array. The combined surface area for allarrays on such sensor chips can be relatively large (e.g. 1 cm2).Large-volume low-cost production of such sensor chips wouldbe extremely impractical through any method other thanreplication. Improving the throughput for fabricating masterstamps using scanning methods of direct-write tools (e.g. EBLor FIB) for circular types of nano-structures for subsequentreplication would be of great value to the engineering of suchsurface nano-structure based devices.

2. Rapid fabrication of master stamp

2.1. Methodology and process steps for fabrication

We demonstrate a novel method for rapid fabrication ofnano-structured master stamps on quartz. In this process, wecan achieve any circular type of pattern of any dimensions,down to the patterning limits of the direct-write tools. Thecapacity of the direct-write tool limits the maximum size ofthe pattern area. The fabrication process uses scanning-typetools, such as EBL or FIB, to create circular dot patternsthat are much smaller than the final desired. Subsequentenlargement through isotropic etching, a relatively low-costprocess, produces the full dimensions of the structures. Assuch, the patterning tool time is drastically shorter thanotherwise required, leading to significant cost savings andhigher throughput. Subsequent metal mask deposition, lift-off,and anisotropic etch process steps are applied to transform thecircular patterns to out-of-plane pillar structures for the finalstamp.

Reduction of the initial patterning tool time as a methodto improve throughput was first demonstrated several decadesago with quantum lithography for the fabrication of integratedcircuit masks [23–25]. In quantum lithography, pre-patternednano-sized metal tiles are placed on the target substrate.A lithography resist coats the tiles, and a direct-write tool(e.g. EBL) patterns small holes, thereby marking desiredtiles. The small holes create openings in the overlaying resistlayer, allowing certain tiles to be selectively etched away.As a result, a large pattern area is defined without havingto scan its full dimensions. While the benefits are similar, in

2

Nanotechnology 24 (2013) 055304 Y Chuo et al

Figure 2. Schematic diagrams of the process flow of the rapid fabrication method for a master nano-structured quartz stamp: (a) initialsmall structures (e.g. 50 nm holes) are patterned onto the writing layer (PMMA); (b) the pattern is transferred and enlarged (e.g. 250 nmholes) onto an intermediate transfer layer (Cr) by wet etching; (c) the bottom lift-off layer (Au) is undercut by wet etching; (d) the writinglayer is removed; (e) a positive mask layer (+Cr) is deposited; (f) the backplane material (Au and first Cr layers) is lifted off; (g) quartzpillars are created by RIE; (h) remaining Cr etch mask is removed.

contrast, our method does not require pre-patterned tiles onthe target substrate. Our patterns are limited to geometriesthat can be created by isotropic enlargement of small circularholes, or collective arrangements of such patterns. Complexarrangements of small holes can generate elliptical, triangular,star patterns, and much more.

Figure 2 illustrates the steps in our rapid fabricationmethod. The process is demonstrated on a quartz substratedue to its durable mechanical properties and its opticaltransparency for UV-based replication processes; however,the same methodology can be applied to a variety of othersubstrates, such as Si, sapphire, and polymer resins.

The quartz substrate is prepared with a thin layer ofAu lift-off layer in contact with the bulk substrate, followedby a thin layer of Cr pattern transfer layer, and finally aresist writing layer. The process begins with patterning smallnano-hole structures on the writing layer (figure 2(a)). Thewriting layer is a thin material layer, such as an electron beamsensitive positive polymer resist polymethyl-methacrylate(PMMA), for use with EBL or FIB direct-write tools. Weselect the resist thickness such that it is close to or smallerthan the diameter of the desired holes, while still being able tomaintain an even flat film coating. It is typically more difficultand time consuming to pattern high-aspect ratio holes;generally, the thinner the resist layer, the shorter the requireddose is to expose an area. Given an operating electron beamcondition (energy and aperture), the direct-write tool can beoptimized to achieve the fastest dot-exposure performance;however, it is the reduction of the patterning area that dictatesthe overall time savings (and cost) in this first step of the rapidfabrication process.

Once the initial small holes are patterned in the resistlayer, the Cr pattern transfer layer is exposed to wet etchantthrough the openings (figure 2(b)). The Cr is selectivelyand isotropically etched to translate the pattern from thewriting layer to the pattern transfer layer, registering largerholes centred at the same location as the small holes above.Experimentally determined etch conditions (e.g. duration,temperature, concentration) control the final diameter of theholes. This pattern transfer enlargement step creates larger

nano-structures without requiring the lengthy time to scan thetotal surface area.

Note that through-holes in the EBL resist layer as smallas 20 nm will give effective Cr etching beneath [22]. It isalso quite possible that even smaller holes will still allowcontrollable etching. We find that due to the common spot sizeof EBL beams, hole sizes anywhere from 20 to 80 nm are mostconvenient, and quickest to write. The time and cost savingsare most notable when the final hole (and subsequently pillar)size is enlarged to greater than 150 nm (particularly suitablefor SPR nano-hole applications).

Next, we enlarge the Au lift-off layer by selective wetetching to undercut the Cr pattern transfer layer (figure 2(c)).This undercut ensures a separation between subsequentmaterials deposited into the cavity from the backplane above,similar to that of a more conventional polymer bi-layerresist lift-off process [14]. Solvent etching removes thewriting layer resist, revealing the enlarged hole structurebeneath (figure 2(d)). It is worth noting that while it ispossible to use different material combinations as the writing,pattern transfer, and lift-off layers, it is necessary to ensurethe selected materials have compatible selective etchants toenable the removal of layer materials without affecting theothers.

At this point in the process, a nano-hole template of fullyenlarged cavities is complete; it can be used as-is in replicationtechniques such as casting [16] and contact printing [17].However, the nano-hole template can be further transformedinto a more versatile and reliable replication stamp master ofpillar-type structures on the quartz substrate.

A second Cr metal layer is deposited across the entiretemplate surface, the portion that falls in the hole cavitiesbecomes the positive Cr etch mask (+Cr) to define thefinal pillar-like structures (figure 2(e)). Here, the metaldeposition must necessarily be a directional depositionprocess (e.g. thermal evaporation) ensuring a separationbetween the newly deposited layer on the top backplanesurface from the base of the cavity in contact with thebulk quartz. Subjecting the template to selective Au etchant(figure 2(f)) removes both the Au and the material above the

3

Nanotechnology 24 (2013) 055304 Y Chuo et al

Figure 3. Centre—SEM image (top view) showing a 60 µm× 60 µm section from a 1 in. × 1 in. quartz nano-pillar stamp fabricated usingour rapid process. The stamp contains nano-structured regions spanning larger than 1 cm2, of a variety of pillar geometries and arrayarrangements all on the same substrate. Higher magnification SEM images (at 45◦ tilt) of the nano-structure regions are labelled and shownin: (a) and (a′) conical nano-pillars arranged in a square lattice; (b) diagonally-oriented, elongated pillars, arranged in a square lattice;(c) linearly-oriented, elongated pillars, arranged in a square lattice; all pillars stand 500 nm tall. The annotated dotted-outlines help indicatethe 3D structure of the pillars.

remaining lift-off layer. Only the +Cr dots from the holecavities remain in direct contact with the quartz surface; theyserve as the final etch mask for the pillar-like structures.

Etching the bulk material of the substrate using ananisotropic reactive ion etch (RIE) (figure 2(g)) createsout-of-plane pillar-like structures. By selecting the thicknessof the +Cr mask, combined with tuning of the RIE recipe,more columnar or more conical pillar structures can begenerated. The etch consistency is very uniform across largeareas. The final step of the process removes any remaining+Cr mask material by wet etching.

2.2. Nano-pillar quartz stamps by rapid fabrication

Figure 3 shows high-magnification scanning electron micro-scope (SEM) images of a small 60 µm × 60 µm exemplarysection from a 1 in.× 1 in. quartz nano-pillar stamp fabricatedusing the rapid process described. The total nano-structuredarea is greater than 1 cm2 and contains several types ofpillar array patterned on the same quartz substrate. Thecentre image in figure 3 shows multiple arrays of variousarrangements, including square, hexagonal, diagonal, rotated,and overlapped, as well as varying periodicities betweenpillars, and varying spacing between arrays. The versatilityof scanning direct-write tools (e.g. EBL or FIB) in creatinga variety of surface nano-structure arrangements over floodpatterning (e.g. LIL) or self-assembly is well illustrated here.

Figure 3(a) shows a higher magnification SEM image ofnano-pillars, at 45◦ tilt, arranged in a basic square lattice array.These particular nano-pillars are conical and circular. Figure3(a′) further shows a higher magnification SEM image of thesame nano-pillars. The pillars stand 500 nm tall, while thebase and the top measure 400 nm and 100 nm respectively.

The slope of the conical nano-pillars is 70◦ from thebase. The sloped conical design provides excellent materialindentation and easier release during replication processessuch as embossing and casting.

Figure 3(b) shows a higher magnification SEM imageof elongated nano-pillars, at 45◦ tilt, orientated diagonallyin a square lattice array, while figure 3(c) shows similarnano-pillars orientated linearly. All pillars on this exemplarstamp stand 500 nm tall. The elongated pillars in variousorientations demonstrate the more advanced pillar shapesthat are achievable; negative replica hole structures can beembossed or cast on polymer surfaces subsequently with thestamp. One example of the application of such elongated orshaped nano-holes is in polarized and enhanced SPR opticaldevices [3].

The surface of the stamp shown in figures 3(a) and (a′)appears rough, and is exaggerated by the way the Aumetal film on the surface of the quartz is formed duringdeposition for SEM imaging. In some cases, the underlyingroughness could create undesirable difficulties in imprintingand de-moulding; however, we have been able to imprintand cast very effectively with all our nano-structured stamps.A silane anti-adhesion release layer is applied to the stampprior to imprinting [16], which may be a compensating factor.To improve the surface morphology and smoothness of thestamps, we are actively pursuing optimizations in the RIErecipe in final etching of the pillars. RIE parameters can beadjusted to achieve smoother surfaces [26].

It is worth noting here that as a fabrication process flowbecomes more elaborate, there is a higher chance of failureor errors during the execution of the various steps. Further,the more pieces of fabrication equipment required, the greaterthe chance of equipment failure and the more difficult it

4

Nanotechnology 24 (2013) 055304 Y Chuo et al

Table 1. Tool time and cost required for patterning exemplary large surfaces with nano-holes. A basic nano-hole array arrangement isconsidered: square lattice of 520 nm in periodicity and final hole diameter of 200 nm. The patterning time/cost is based on the Raith eLineEBL tool operated at 30 µm aperture, 10 keV beam energy, and 200 µm write-field; considering a typical tool usage cost of $100 h−1.

Surface area(cm2)

Time required usingrapid fabrication (h)

Cost required usingrapid fabrication ($)

Time required usingconventional fabrication (h)

Cost required usingconventional fabrication ($)

Difference($)

1 21 2 100 75 7 500 5 4002 42 4 200 150 15 000 10 8005 105 10 500 375 37 500 27 000

is to deal with equipment scheduling and availability. Thisis a common and inevitable fact of multi-step fabricationthat should be considered when comparing and choosingprocesses for manufacturing. The multi-step process flowcan be contrasted with the simplicity of fabrication usingtools such as focused ion beam milling, where the final hardstamp structure is machined directly onto the bulk materialwithout the need for any other process steps; however, aspreviously explained, the milling time is often much toolong for large-area fabrication. With that said, we furtherreport that we are able to achieve excellent output from allof our production runs using our rapid fabrication process.The estimated nano-structure yield on the master stampsis 99.96%, where in our sample analysis we consider anymissing or broken pillars a failed structure.

2.3. Comparison with conventional EBL fabrication

Well-known conventional EBL processes to create nano-pillarstructures on hard substrates (e.g. quartz, Si, etc) include(a) bi-layer positive resist lift-off process [27], and (b) etchresistant negative resist pattern transfer [28]. The height ofthe final pillar structures attainable from method (b) is heavilylimited by the thickness of the EBL patterning resist, where,often at best, the etch resistance selectivity is one-to-oneagainst a hard substrate such as quartz. In contrast, method (a)applies a metal mask that is much more etch resistant. In ourbasic comparison, we consider method (a), the conventionalbi-layer positive resist lift-off process, against our rapidfabrication process.

Table 1 illustrates a comparison of the time and cost ofconventional nano-hole patterning versus our rapid processfor basic square nano-hole arrays with a periodicity of520 nm, and target hole diameters of 200 nm. An EBL(Raith eLine) patterning tool with 30 µm aperture, 10 keVbeam energy, and 200 µm write-field is used for bothscenarios. As the area of nano-structures increases from 1 to5 cm2, the time and cost savings become drastically moremeaningful, with cost savings increasing from thousands ofdollars to tens-of-thousands of dollars. In practical use, thisrapid fabrication process enables quick prototyping of originalstamp masters of any design. A total pattern area of 1 cm2

can be completed within 24 h using the new process ratherthan the three days as required by conventional writing usingthe same tool parameters. Note that even with faster EBLwrite settings, the rapid fabrication process correspondinglyimproves the throughput by reducing the patterning timeneeded (ultimately limited by the machine hardware scanning

speed). The subsequent fabrication process steps describedearlier can be completed within minutes to achieve fullyenlarged nano-hole arrays, and within hours to arrive atout-of-plane nano-pillar stamps.

We further compare the total time required in theprocesses to arrive at pillar structures on the hard substratesbeneath the resist/process layer. In our rapid process, twometal deposition cycles are required (one for the metal processlayers, another for the Cr metal mask). In the conventionalbi-layer lift-off process, two metal deposition cycles are alsorequired (one for the thin metal film on the substrate beneaththe EBL resists, another for the hard metal mask for lift-off).Each of these process steps would add 4–6 h to the overallfabrication time.

We can also compare our process to the conventionalnegative resist direct patterning. Here, we assume the negativeresist is adequately etch resistant, even as effective as a metalmask. In this case, we only require one metal deposition cycle(for the thin metal film required on the substrate beneath theEBL resist).

We now consider all the other processing times forour process: chemical etching, approximately 1 h includinghandling time; RIE pillar etch, approximately 2 h includingsetup and pump-down time; and lift-off, approximately 2 h.Thus, a total of 17 h of additional post-EBL processingtime is required in our process. The conventional bi-layerlift-off process requires almost the same, 15 h (where thethinner initial metal layer beneath EBL resist requires shorterdeposition time). The conventional negative resist patterningrequires only an additional 7 h. A summary of the totalprocessing time is given in table 2.

When considering a sample of 1 cm2 nano-pillarpopulated area, we must add 17 h to the 21 h of EBLpatterning time, totalling 38 h for our process. We must alsoadd 15 h to the conventional bi-layer lift-off method, totalling90 h, and 7 h to the conventional negative resist method,totalling 82 h. It is evident that while the additional processtimes are not negligible, it certainly does not reduce the timeand cost savings percentage by much, particularly when evenlarger pattern areas are considered.

3. Embossing and SPR applications using rapidfabricated stamps

3.1. Polymer embossing with nano-pillar stamps

To validate the utility and practicality of the nano-pillar stamp,a stamp was used to emboss on polyethylene terephthalate

5

Nanotechnology 24 (2013) 055304 Y Chuo et al

Figure 4. SEM images of embossed nano-holes (cavities) from selected small regions (approx. 6 µm× 6 µm) on a medium-grade PETsubstrate embossed with a master quartz stamp containing multiple nano-pillar arrays of multiple designs spanning a total area larger than1 cm2. (a) Square lattice; (b) hexagonal lattice; (c) higher magnification of embossed nano-holes, showing a conical ‘bucket’ shape.

Table 2. Summary of additional and total processing times for our rapid fabrication process, the conventional bi-layer lift-off process, andthe conventional negative resist process.

Additional process steps

Processing time (h)

Rapid fabricationprocess

Conventional Bi-layer lift-offprocess

Conventional negativeresist process

Metal deposition 6+ 6 4+ 6 4RIE pillar etch 2 2 2Liftoff 2 2 0Resist development and/or chemical etching 1 1 1Total (additional) 17 15 7Total (including EBL time for 1 cm2 pattern areaa) 38 90 82Total (including EBL time for 5 cm2 pattern areaa) 122 390 382

a For nano-structures same as that described in table 1.

(PET) substrates. One of the master quartz stamps applied was1 in.× 1 in., and contained multiple arrays of multiple designsthat combined to span more than 1 cm2 in total patternedsurface area. The entire 1 cm2 featured region was transferredconsistently onto PET negatives.

Figure 4 shows several SEM images of embossednano-hole (or nano-cavity) structures on medium-grade PET(surface roughness approx. 50 nm). Figures 4(a) and (b) showselected 6 µm × 6 µm regions from the stamped surface ofa single PET substrate. These structures represent negativemould formations of the nano-pillars on the stamp. Figure 4(a)shows a basic nano-hole array arranged in a square lattice.Figure 4(b) shows nano-hole arrays arranged in a hexagonallattice. Figure 4(c) shows a higher magnification SEM imageof two adjacent nano-hole structures. The diameter on the baseof the hole (cavity) in the sample is 150 nm while the diameteron the top is 300 nm.

The sidewalls of the embossed holes and cavities areevidently rougher than the sidewalls of the pillars on themaster stamp. The backplane on the PET surface also hasrelatively rougher texture than the original master quartzstamp. However, by comparison with blank PET substrates,the surfaces show similar roughness, and thus we deduce thatthe stamp does not create any unexpected deformations ordamage to the embossed surfaces.

Using image analysis software (IGOR Pro) we detectthe jagged boundaries that define the edges of the circularembossed holes. We calculate the circularity, which is aquantitative metric user to evaluate the quality of nano-hole

Table 3. Circularity measurements of the nano-pillar structures onquartz, nano-hole structures on embossed PET, and FIB-millednano-holes.

Quartz stamppillar

Embossednano-holes

FIB-millednano-holes

Circularity 1.4± 0.1 6± 2 1.19± 0.03

structures in SPR applications, where the circularity, Circ, isdefined as

Circ =P2

(4πA). (1)

Thus, a circularity of 1 indicates a perfect circle, while thelarger the circularity value, the more the perimeter deviatesfrom being perfectly circular.

We examine the embossed nano-hole structures andcompare this with the circularity of higher quality FIB-millednano-hole structures. We also provide a circularity assessmenton the pillars of the quartz stamps used to create the embossedholes. A summary of results is given in table 3. It is evidentthat the circularity of the embossed nano-holes is not asexcellent as the FIB-milled, however, the circularity of thepillar structures are quite close to the high-quality FIB-milledhole structures.

We note that much of the deviation from highercircularity on the embossed holes is actually due to theoriginal roughness of the PET film grains.

6

Nanotechnology 24 (2013) 055304 Y Chuo et al

Figure 5. Large-area nano-hole array features replicated on high-grade PET from quartz master stamps. Left—A1 coated PET replicationscompared with a Canadian two-dollar coin; images (a)–(d) further show magnified SEM images on various regions of nano-hole arrays atthe indicated PET surface.

Figure 5 further shows large-area replicated nano-holearrays on higher-grade PET using our quartz hard stamp.The magnified SEM images give zoomed-in perspectivesat various regions of one sample. The consistency of thenano-hole arrays demonstrates the quality of the stamp andtransfer. The macroscopic image of the nano-hole arrays onPET pieces further show that the replications are extremelycomplete and repeatable across large surfaces.

3.2. SPR characterization of nano-hole structures onembossed substrates

We selected some basic embossed nano-hole arrays for SPRsensing. While many applications can benefit from the rapidfabrication process for master stamps, SPR sensing is an areaof science well recognized for its use of nano-hole structures,but challenged by costly production and difficult reuse. Assuch, it is particularly suitable for illustrating the utility of therapid fabrication process.

Upon exposure to incident radiation, periodic nano-holearrays on PET coated with an appropriate metal may exhibitan SPR-related phenomenon known as extraordinary opticaltransmission (EOT). Shifts in the EOT spectral peaks indicatechanges in the local refractive index surrounding the nano-structures [3]. An embossed PET surface containing squarelattice nano-hole arrays with four different periodicities,ranging from 500 to 625 nm, were coated with 5 nm of Crfollowed by 100 nm of Au. EOT spectra were collected.

The wavelengths of the EOT peaks are predicted by

λ0 =

√εmεd

εm + εm

a2

i2 + j2(2)

where a is the periodicity of the array, ij are integersdescribing the order of the resonance, and εm and εdare the dielectric functions of the metal and the dielectricrespectively [29, 30]. In reality, the EOT peaks areexpected to be red-shifted from their predicted location

Figure 6. Typical EOT transmission spectra for four differentsquare lattice arrays, periodicities ranging from 500 to 625 nm. Theprominent [1, 0] Au–air resonance shifts to higher wavelength withincreasing periodicity. Inset compares the 550 nm periodicityembossed array with a more accurate FIB-milled array.

because equation (2) does not consider parameters such asdirect scattering contributions [31], hole shapes and holesizes [32–34]. Figure 6 shows experimental, backgroundsubtracted, EOT air spectra from the embossed arrays. Thebackground transmission is the light passing through an areaof plain metal coating the same thickness as the nano-holeregions. For clarity, we plot only every eighth data point onthe spectra. The locations of the prominent [1, 0] air–Aupeaks, marked by black lines, show a direct dependence onthe array periodicity, as predicted by equation (2). The peaksshift to higher wavelength with increasing array periodicity.EOT peaks are a combination of SPR resonance and directlyscattered light, which gives the peaks their characteristicasymmetric Fano-type shape. The direct component falls offas (r/λ)4 and becomes less important at higher wavelengths.

The inset in figure 6 shows a comparison of thebackground subtracted EOT results for a square lattice,FIB-milled array (periodicity 550 ± 10 nm, 124 ± 4 nm

7

Nanotechnology 24 (2013) 055304 Y Chuo et al

hole diameter on glass), and an embossed array (periodicity550 ± 5 nm, 250 ± 20 nm hole diameter on PET). Spectraare scaled to account for differences in the total numberof nano-holes in each sample. We attribute peaks seen at540 ± 1 nm, 569 ± 2 (embossed) and 547 ± 2 nm (milled)to the [2, 1], [2, 0] Au–PET and [2, 0] Au–glass resonances,respectively. Peaks at 638 ± 2 nm (milled) and 628 ± 1 nm(embossed) are the [1, 0] Au–air resonances. Finally, 701 ±5 nm (milled) and 746 ± 6 nm (embossed) are the [1, 0]Au–glass and [1, 0] Au–PET peaks, respectively. Fabricationinaccuracies and geometry differences caused a slight latticemismatch between milled and embossed arrays, which arelikely the largest factors contributing to the 10 nm differencebetween the locations of the [1, 0] Au–air resonance peaksfor milled versus embossed arrays. Between the [2, 0] Ausubstrate and the [1, 0] Au–air peaks we see a series ofminima attributed to Wood’s anomalies [29]. The full-widthhalf-maximum (FWHM) of the two EOT [1, 0] Au–air peaksare similar: 30 ± 1 nm for embossed versus 27 ± 2 nm formilled.

The magnitudes of the resonance peaks are very differentin the stamped versus milled arrays. The greater surfaceroughness in the stamped Au (50 ± 25 nm) versus milledAu (16± 4 nm) may weaken the EOT resonance by allowingscattering and radiative decay of the SP modes [35]. Further,the nature of the metal film deposition on the embossedarrays is such that it also coats the bottom and sidewalls ofeach indented nano-hole cavity. It has been verified throughcomputer simulation that thicker metal film coatings on thesidewalls correspond to higher attenuation of the EOT light.

It is observed that the attenuation for the embossednano-hole array [1, 0] air–Au peak is 11 times, while thatof the [2, 0] Au substrate peak is nine times compared tothe more accurately FIB-milled arrays. Further discussions onthe influence of the metal film and hole geometry on EOTattenuation based on simulated results are provided in thesupporting information (available at stacks.iop.org/Nano/24/055304/mmedia). The 550 nm periodicity array has a 0.5 ±0.1% transmission of the total incoming light at 630 nm.

The distinct [1, 0] air–Au SPR resonances shownin figure 6 suggest that the quality of the embossed nano-holearrays on PET can in fact generate significant SPR, withthe same characteristic EOT resonance peaks as that ofa FIB-milled array. Although the resonance intensity isapproximately an order of magnitude lower, the resonancepeaks are clearly detectable and identifiable when theappropriate light source and detectors are applied. Giventhe vast selection of electronic hardware available in today’stechnology, it can be expected that even at lower intensity,without further optimization of the polymer surface and metalfilm coating, such embossed nano-hole arrays can potentiallybe practical sensor elements in low-cost SPR-based sensing.

Furthermore, comparisons with other nano-hole arraystructures produced by nano-imprint lithography from siliconmasters made in a conventional manner [36], or by phaseshift lithography from quartz or silicone-based templates(masks) [18, 37] show the holes are comparably circular. TheSPR peaks of our metallized embossed nano-holes display

similar resonance peaks to those demonstrated in the otherreplicated nano-hole arrays [37].

4. Conclusions

We have demonstrated a rapid fabrication process forquartz master stamps, featuring nano-pillar structures thatproduce adequate templates for the reproduction of nano-holestructures through direct replication methods, such asembossing. The resulting replicated nano-structures representgood conformal negative forms of the master structures; thegeometry is sufficiently well-formed to generated distinctEOT resonance peaks such that they can be consideredfor applications such as SPR-based nano-optics. Our rapidprocess creates master stamps at a fraction of the time and cost(demonstrated on quartz) of conventional EBL methods. Inthe engineering of surface nano-structures, there is still a needfor efficient, low-cost, large-area, large-volume fabrication.Many applications require multiple array types and sizes ofnano-hole-like structures all on a single substrate. We havedeveloped a method that meets these challenges, adding to thepool of techniques available to nano-scientists.

5. Experimental section

5.1. Metal films for process layers

The metal films for the lift-off layer and pattern-transfer layerwere prepared using a thermal deposition tool. A 50 nmAu layer was first deposited as the lift-off layer, then a40 nm Cr layer was deposited as the pattern transfer layer.The two metal layers were deposited consecutively withoutbreaking vacuum in a dual source thermal deposition tool.The deposition rate for the Au layer was 0.4 A s−1 while thedeposition rate for the Cr layer was 0.3 A s−1.

5.2. PMMA resist coating

Polymethyl-methacrylate (PMMA) was applied as the writinglayer for EBL patterning according to the fabrication processdescribed. PMMA (MicroChem Inc., 2% solids in Anisole)was spin-coated at 500 rpm for 7 s and then ramped to4000 rpm for 60 s to obtain a 50 nm film layer. A subsequentsoft bake step at 180 ◦C for 10 min was carried out toevaporate the solvent.

5.3. Parameters for direct-write tool

The direct-write tool used to demonstrate the rapid fabricationprocess was electron beam lithography (EBL). However, notethat both a focused ion beam tool (FIB) or an ion beamlithography tool (IBL), as well as EBL, are compatible withthe master stamp fabrication process presented. A beamenergy of 10 keV (higher backscattering for thinner resistfilms) at 30 µm aperture was applied for all cases. Thewrite-field was set to 200 µm for all comparisons.

8

Nanotechnology 24 (2013) 055304 Y Chuo et al

5.4. Wet chemical etchants

Two wet chemical etchants were required in the fabricationprocess; one for the controlled enlargement of the patterntransfer layer, and another for the undercutting of the lift-offlayer. For the Cr pattern transfer layer, we applied a Cr etchant(Transene 1020 AC) that is compatible with PMMA and Au.The etchant was diluted with one part deionized water priorto use, to decrease the etch rate, and to allow better controlof timed enlargement of the nano-hole structures. For the Aulift-off layer, we applied an Au etchant (Transene TFA) that iscompatible with Cr. The Au etchant was diluted with twentyparts deionized water to significantly reduce the etch rate toallow better control of the lift-off layer undercutting geometry.All etching was carried out at 19± 1 ◦C.

5.5. RIE conditions

Directional etching to create nano-pillar (or conical)structures from +Cr masks was carried out with reactiveion etch (RIE). RIE (Sentech Etchlab 200) was operated at6 mTorr with a mixture of gases CF4, SF6, and O2. To vary thepillar geometries (slope, height, width), the RF power rangedfrom 200 to 400 W, while the etch time ranged from 5 to10 min.

5.6. Surface preparation for SEM imaging

Unless otherwise specified, all non-conductive samplesprepared for scanning electron microscopy (SEM) werecoated with 10 nm Cr and 20 nm Au film using a thermaldeposition tool.

5.7. Embossing PET

Polyethylene terephthalate (PET) sheets (120 µm thick,Grafix Plastics Dura-lar) were thermally embossed using aCarver 4386 heated press hydraulic system. The quartz masterstamp is first inserted into the heat press, placed on the bottomplate with the nano-structures facing upwards. After cleaningfor 30 s with acetone, isopropyl alcohol (IPA) rinse, and thenblow-drying with nitrogen air, the PET was placed on topof the quartz stamp inside the press. The PET was softenedby increasing the temperature of the heater press platensto above that of the polymer glass transition temperature,approximately 80 ◦C. The embossing system transferred heatto the PET material from the bottom plate, through the quartzstamp. Prior to placing the polymer on the quartz stamp, thetemperature on the quartz surface was directly measured usinga surface probing thermocouple until the desired temperaturewas reached. A pressure of 7 MPa (1000 psi) was applied for5 min. Once completed, while the embossed PET was still infirm contact with the quartz stamp, the entire assembly wascooled to a temperature below 50 ◦C prior to detaching thepolymer from the stamp.

5.8. Metal film deposition for embossed samples for SPRcharacterization

The nano-pillar stamp embossed PET samples were ultrason-ically cleaned with IPA prior to metal deposition. The metalfilms were prepared using a thermal deposition tool. A 5 nmCr layer was first deposited in contact with the embossed PETsurface, followed by a 100 nm Au layer. The two metal layerswere deposited consecutively without breaking vacuum in adual source thermal deposition tool. The deposition rate forthe Cr layer was 0.2 A s−1 while the deposition rate for theAu layer was 0.5 A s−1.

5.9. EOT spectra collection

Extraordinary optical transmission (EOT) spectra werecollected using a spectrometer (Photon Control SPM-002-C,customized 500–1000 nm range, 50 µm slit and resolution <1.6 nm) coupled through a fibre-optic cable to a Reichertinverted microscope. Surface plasmon resonance (SPR) wasgenerated by exciting the metal-coated embossed nano-holearrays using a halogen white light source (Fibre Lite 180,Dolan Jenner Industries). The spectra obtained were anaverage of 20 acquisitions collected at an exposure of 200 ms.The spectra curves were smoothed using a boxcar algorithmover each five adjacent points. Background spectra werecollected by measuring the transmission through an area ofthe metallized PET substrate without nano-structures.

Acknowledgment

This work was supported by NSERC and MITACS Canada.

References

[1] Brolo A, Nirwan R, Scott J, Gulzar N, Gray B, Romanuik S,Grist S, Kavanagh K, Hohertz D and Gordon R 2010 A newmethod for therapeutic antibody screening SPIE Newsroomdoi:10.1117/2.1201006.002995

[2] Sinton D, Gordon R and Brolo A 2008 Nanohole arrays inmetal films as optofluidics elements: progress and potentialMicrofluid Nanofluid 4 107

[3] Gordon R, Sinton D, Kavanagh K and Brolo A 2008 A newgeneration of sensors based on extraordinary opticaltransmission Acc. Chem. Res. 41 1049

[4] Do Y R, Kim Y C, Song Y W, Cho C O, Jeon H, Lee Y J,Kim S H and Lee Y H 2003 Enhanced light extraction fromorganic light-emitting diodes with 2D SiO2/SiNx photoniccrystals Adv. Mater. 15 1214

[5] Kim Y C and Do Y R 2005 Nanohole-templated organiclight-emitting diodes fabricated using laser-interferinglithography: moth-eye lighting Opt. Express 13 1598

[6] Tokushima M, Kosaka H, Tomita A and Yamada H 2000Lightwave propagation through a 120◦ sharply bentsingle-line-defect photonic crystal waveguide Appl. Phys.Lett. 76 952

[7] Arsenault A, Puzzo D, Manners I and Ozin G 2007Photonic-crystal full-colour displays Nature Photon. 1 468

[8] Fan Z et al 2009 Three-dimensional nanopillar-arrayphotovoltaics on low-cost and flexible substrates NatureMater. 8 648

9

Nanotechnology 24 (2013) 055304 Y Chuo et al

[9] Menezes J W, Ferreira J, Santos M J L, Cescato L andBrolo A G 2010 Large-area fabrication of periodic arrays ofnanoholes in metal films and their application in biosensingand plasmonic-enhanced photovoltaics Adv. Funct. Mater.20 3918

[10] Celanovic I, Jovanovic N and Kassakian J 2008Two-dimensional tungsten photonic crystals as selectivethermal emitters Appl. Phys. Lett. 92 193101

[11] Krahn J, Sameoto D and Menon C 2011 Controllablebiomimetic adhesion using embedded phase changematerial Smart Mater. Struct. 20 015014

[12] Pina-Hernandez C, Kim J S, Guo J L and Fu P F 2007 Highthroughput etch-selective nanoimprinting and stampingbased on fast-thermal-curing poly (dimethylsiloxane)s Adv.Mater. 19 1222

[13] Lee H, Hong S, Yang K and Choi K 2006 Fabrication ofnano-sized resist patterns on flexible plastic film usingthermal curing nano-imprint lithography Microelectron.Eng. 83 323

[14] Pozzato A, Dal Zilio S, Bruschi L, Mistura G andTormen M 2009 Fabrication of substrates with extendednanostructured surface areas for wetting studiesMicroelectron. Eng. 86 1329

[15] Cecchini M, Signori F, Pingue P, Bronco S, Ciardelli F andBeltram F 2008 High-resolution poly(ethyleneterephthalate) (PET) hot embossing at low temperature:thermal, mechanical, and optical analysis of nanopatternedfilms Langmuir 24 12581

[16] Qin D, Xia Y and Whitesides G M 2010 Soft lithography formicro- and nanoscale patterning Nature Prot. 3 491

[17] Loo Y L, Willett R L, Baldwin K W and Rogers J A 2002Additive nanoscale patterning of metal films with a stampand a surface chemistry mediated transfer process:applications in plastic electronics Appl. Phys. Lett. 81 562

[18] Henzie J, Lee M H and Odom T W 2007 Multiscale patterningof plasmonic metamaterials Nature Nanotechnol. 2 549

[19] Nelson E C et al 2011 Epitaxial growth of three-dimensionallyarchitectured optoelectronic devices Nature Mater. 10 676

[20] Chang C H, Tian L, Hesse W R, Gao H, Choi H J, Kim J G,Siddiqui M and Barbastathis G 2011 From two-dimensionalcolloidal self-assembly to three-dimensionalnanolithography Nano Lett. 11 2533

[21] Langner A, Paivanranta B, Terhalle B and Ekinci Y 2012Fabrication of quasiperiodic nanostructures with EUVinterference lithography Nanotechnology 23 105303

[22] Chou T H, Cheng K Y, Hsieh C W and Takaya Y 2012Roll-to-roll fabrication of a low-reflectance transparentconducting oxide film with subwavelength structuresJ. Micromech. Microeng. 22 045009

[23] Fulton T A and Dolan G J 1983 New approach to electronbeam lithography Appl. Phys. Lett. 42 752

[24] Maluf N I and Pease R F W 1991 Quantum lithography J. Vac.Sci. Technol. B 9 2986

[25] Chou S Y, Li W D and Liang X 2009 Quantized patterningusing nanoimprinted blanks Nanotechnology 20 15530

[26] Zeze D A, Forrest R D, Carey J D, Cox D C, Robertson I D,Weiss B L and Silva S R P 2002 Reactive ion etching ofquartz and Pyrex for microelectronic applications J. Appl.Phys. 92 3624

[27] Chen Y, Peng K and Cui Z 2004 A lift-off process for highresolution patterns using PMMA/LOR resist stackMicroelectron. Eng. 73/74 278

[28] O’Faolain L, Kotlyar M V, Tripathi N, Wilson R andKrauss T F 2006 Fabrication of photonic crystals using aspin-coated hydrogen silsesquioxane hard mask J. Vac. Sci.Technol. B 24 336

[29] Ebbesen T W, Lezec H J, Ghaemi H F, Thio T andWolff P A 1998 Extraordinary optical transmission throughsub-wavelength hole arrays Nature 391 667

[30] Thio T, Ghaemi H F, Lezec H J, Wolff P A andEbbesen T W 1999 Surface-plasmon-enhancedtransmission through hole arrays in Cr films J. Opt. Soc.Am. B 16 1743

[31] Genet C, van Exter M P and Woerdman J P 2003 Fano-typeinterpretation of red shifts and red tails in hole arraytransmission spectra Opt. Commun. 225 331

[32] Klein Koerkamp K J, Enoch S, Segerink F B,van Hulst N F and Kuipers L 2004 Strong influence of holeshape on extraordinary transmission through periodic arraysof subwavelength holes Phys. Rev. Lett. 92 183901

[33] van der Molen K L, Klein Koerkamp K J, Enoch S,Segerink F B, van Hulst N F and Kuipers L 2005 Role ofshape and localized resonances in extraordinarytransmission through periodic arrays of subwavelengthholes: experiment and theory Phys. Rev. B 72 045421

[34] van der Molen K L, Segerink F B, van Hulst N F andKuipers L 2004 Influence of hole size on the extraordinarytransmission through subwavelength hole arrays Appl.Phys. Lett. 85 4316

[35] Raether H 1988 Surface Plasmons on Smooth and RoughSurfaces and on Gratings (Berlin: Springer)

[36] Hamoudaa F, Barbillona G, Helda S, Agnusa G, Gogola P,Maroutiana T, Scheuringb S and Bartenliana B 2009Nanoholes by soft UV nanoimprint lithography applied tostudy of membrane proteins Microelectron. Eng. 86 583

[37] Odom T W, Hanwei G, McMahon J M, Henzie J andSchatz G C 2009 Plasmonic superlattices: hierarchialsubwavelength hole arrays Chem. Phys. Lett. 483 187

10