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

Download details:

IP Address: 158.130.80.10

This content was downloaded on 26/05/2016 at 13:38

Please note that terms and conditions apply.

Computer numerical control (CNC) lithography: light-motion synchronized UV-LED lithography

for 3D microfabrication

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

2016 J. Micromech. Microeng. 26 035003

(http://iopscience.iop.org/0960-1317/26/3/035003)

Home Search Collections Journals About Contact us My IOPscience

1 © 2016 IOP Publishing Ltd Printed in the UK

I. Introduction

Recent microfabrication technology has advanced not only in patterning resolution but also in system functionality and process efficiency. Today’s fine 3D microfabrication has been utilized for various applications ranging from small sensors to microsystems [1–4]. While several 3D microfabrication methods have been introduced such as laser lithography [5, 6], stereolithography methods, and 3D printing [7, 8], the relatively high system and process cost of such approaches

has been a barrier to much broader usage of such fabrication processes. Laser or 3D printing enables the fabrication of 3D microstructures; however their serial fabrication nature may result in increased time and cost.

UV lithography has been widely exploited as a batch-compatible and cost-effective process for micromachining. However, traditionally, it has been utilized to produce two-dimensional (2D) or relatively simple high aspect ratio structures (often it is called 2.5D structures). Several alterna-tive UV lithography schemes to expand the design capability

Journal of Micromechanics and Microengineering

Computer numerical control (CNC) lithography: light-motion synchronized UV-LED lithography for 3D microfabrication

Jungkwun Kim1, Yong-Kyu Yoon2 and Mark G Allen1

1 Department of Electrical and Systems Engineering, University of Pennsylvania, Philadelphia, PA 19104, USA2 Department of Electrical and Computer Engineering, University of Florida, Gainesville, FL 32611, USA

E-mail: mallen@seas.upenn.edu

Received 16 October 2015, revised 15 December 2015Accepted for publication 29 December 2015Published 27 January 2016

AbstractThis paper presents a computer-numerical-controlled ultraviolet light-emitting diode (CNC UV-LED) lithography scheme for three-dimensional (3D) microfabrication. The CNC lithography scheme utilizes sequential multi-angled UV light exposures along with a synchronized switchable UV light source to create arbitrary 3D light traces, which are transferred into the photosensitive resist. The system comprises a switchable, movable UV-LED array as a light source, a motorized tilt-rotational sample holder, and a computer-control unit. System operation is such that the tilt-rotational sample holder moves in a pre-programmed routine, and the UV-LED is illuminated only at desired positions of the sample holder during the desired time period, enabling the formation of complex 3D microstructures. This facilitates easy fabrication of complex 3D structures, which otherwise would have required multiple manual exposure steps as in the previous multidirectional 3D UV lithography approach. Since it is batch processed, processing time is far less than that of the 3D printing approach at the expense of some reduction in the degree of achievable 3D structure complexity. In order to produce uniform light intensity from the arrayed LED light source, the UV-LED array stage has been kept rotating during exposure. UV-LED 3D fabrication capability was demonstrated through a plurality of complex structures such as V-shaped micropillars, micropanels, a micro-‘hi’ structure, a micro-‘cat’s claw,’ a micro-‘horn,’ a micro-‘calla lily,’ a micro-‘cowboy’s hat,’ and a micro-‘table napkin’ array.

Keywords: computer numerical control (CNC) lithography, UV-LED lithography, light-motion synchronized, 3D microfabrication

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

J Kim et al

Printed in the UK

035003

JMMIEZ

© 2016 IOP Publishing Ltd

2016

26

J. Micromech. Microeng.

JMM

0960-1317

10.1088/0960-1317/26/3/035003

Paper

3

Journal of Micromechanics and Microengineering

IOP

0960-1317/16/035003+10$33.00

doi:10.1088/0960-1317/26/3/035003J. Micromech. Microeng. 26 (2016) 035003 (10pp)

J Kim et al

2

have been introduced including multidirectional UV lithog-raphy [9, 10], diffuser lithography [11], and lithography combined with a timed-development-and-thermal-reflow process [12]. Although those fabrication processes have proven to be useful in 3D microfabrication, they commonly have used a conventional mercury-vapor bulb based UV light source, which poses several disadvantages. Since the mercury-vapor bulb requires a preheating time, the lithog-raphy system keeps the light bulb ‘on’ during the operation and a mechanical shutter is used for temporal modulation, which may not be a power efficient approach. Light inten-sity adjustment or modulation may not be easily achievable. Also, the mercury-vapor bulb needs high power to operate and requires many optical parts such as a reflector, mono-chromic mirrors, and multiple collimating lenses. Therefore, the conventional mercury-vapor bulb based UV lithog-raphy system is typically bulky with a large power supply and multi-part optics, and its purchase and maintenance is expensive because of parts cost, the requirement for the pre-cise alignment and assembly of the system, and the short lifetime of a mercury-vapor bulb.

Recently, there have been efforts to create a more acces-sible and affordable 3D microfabrication system by replacing the mercury-vapor bulb based UV light source with UV-LEDs [13–15]. UV-LEDs with emission peaks ranging from 360–405 nm have been utilized as alternative light sources for UV lithography. An array of 100 UV-LEDs with 405 nm peak wavelength was introduced as a UV lithography light source and submicron features of 200 nm over 4 inch Si wafers were demonstrated [13]. In this system, a diffuser was employed at the UV-LED array to provide uniform UV light intensity. However, employment of a diffuser sacrifices both light intensity and collimation. This could be problematic for fab-ricating thick and high-aspect-ratio microstructures in thick resist systems. Subsequently, a commercial UV-LED torch was utilized as a UV lithography light source. The UV-LED torch was placed on top of a glass photomask for UV expo-sure. This method demonstrated a 7 μm pattern feature for surface acoustic wave (SAW) device fabrication. However, the method suffered from non-uniform UV-LED illumination, which caused irregular pattern transfer in the fabrication trials [14]. Table 1 compares the performance of UV-LEDs with a mercury lamp for several key parameters of interest for UV light source designers. It can be seen that the UV-LED shows performance very comparable to the conventional light source.

Building on the previous work of the UV-LED lithography [15, 16], the efficacy of a 100-element array of modified

UV-LEDs as a light source for conventional and advanced (3D) UV lithography is assessed. The modification consisted of suppressing the high flare angle light of each UV-LED by means of an opaque layer. The intensity distribution and attenuation as a function of the distance from the array are measured. The performance of the UV-LED source is also compared to the performance of a conventional mercury-lamp-based lithography system.

Exploiting the advantages of the UV-LED array as a UV light source, this paper introduces the concept of computer-numerical-controlled ultraviolet light-emitting diode (CNC UV-LED) lithography for 3D microfabrication as shown in figure  1. CNC UV-LED adopts the tilt-rotational sample holder from multidirectional UV lithography and synchro-nizes the LED illumination with the motion of the sample holder. Unlike conventional additive approaches such as ste-reolithography, CNC-lithography allows the formation of fine features, has no layering feature artifacts, and is a batch pro-cess, albeit at the expense of the ability to create completely arbitrary patterns.

II. System description

II.A. UV light source: UV LED array

A 5 mm-through-hole LED (RL5-UV0315-400, Superbrightleds Inc.) was used for the UV-LED exposure system. The UV-LEDs were assembled into a 10 × 10 array to form the UV light source. Ten LEDs were connected in parallel forming a row, and ten rows were again connected in parallel. Since each LED has a rated current of 30 mA, the maximum current of the UV-LED array was set at 3 A. A double-layer printed circuit board (PCB) was used for the substrate; the positive electrodes were placed on one side and the negative electrodes on the other side. Figure  2(a) shows the fabricated circuit board. Female head connectors were placed on the PCB for ease of access/replacement of the LEDs. The bulb of each LED used a poly-olefin tube to prevent light propagation through the sidewall of the bulb. A housing for the array was fabricated by 3D printing (Makerbot Replicator 2x) using acrylonitrile butadiene styrene (ABS). The completed UV light source is shown in figure 2(b).

II.B. Tilt-rotational sample holder

The tilt-rotational sample holder comprises a 3 inch (7.6 cm)-circular plate equipped with stepper motors for tilt and rotational movement. National Electrical Manufacturer Association (NEMA) 23 stepper motors were employed with micro stepping function for smooth tilt and rotational motions. A counterweight was added to the shaft of the rotational stepper motor making the circular plate balanced even at a tilt position. In this way, the tilt motor can hold the weight of the circular plate with the rota-tion motor without providing a large torque. The circular plate was painted with non-glossy black color to prevent parasitic reflections. All frames were made from 0.25 inch (6.35 cm) thick aluminum plates.

Table 1. Comparison between UV-LED and mercury lamp lithography system

System UV LED Mercury lamp

Bulb life time ~100 000 h ~2000 hBulb cost ~$1/LED ~$300Typical light intensity 8–30 mW cm−2 5–40 mW cm−2

Light collimation ~15° (air) ~7° (air)~7° (photoresist) ~3° (photoresist)

J. Micromech. Microeng. 26 (2016) 035003

J Kim et al

3

II.C. Integrated system

The integrated CNC lithography system is presented in figure 3. A commercial motor drive used as a control unit (EZ4AXIS, Allmotion Inc.) was employed to operate the on-off state of the UV light source and the tilt-rotational sample holder. The control unit also rotates the UV light source using a direct cur-rent (DC) motor with 3D printed housing. The control unit can operate each part independently by communicating with the computer for the CNC lithography program. Connectors and frames for the UV-LED array circuit and the DC motor were 3D printed (Makerbot Replicator 2x) using ABS.

III. UV-LED characterization

Wavelength and intensity of the UV light are important parameters in UV lithography, particularly influencing pat-tern quality. Typically, a conventional mercury-vapor bulb radiates a broadband spectrum of UV wavelengths ranging approximately from 200 nm to 430 nm, and the corresponding

light intensity per each wavelength is varied. During expo-sure in a conventional system, the different wavelengths with various light intensities often cause parasitic patterns and scattering. Therefore, a single peak narrow bandwidth UV wave is preferred for UV lithography. Often, commer-cial optical high-pass or band-pass filters are used to sort out the undesirable wavelengths from the broadband UV light. However, this adds one more component to the UV light path, lowering the UV intensity as well as the light quality from possible diffraction and reflection at the inter-faces of the additional optical part. Also, it adds extra cost. An LED is designed to radiate at a single wavelength. The wavelength of the proposed UV-LED has been measured with spectroscopy (RSpec-Explorer, Arbor Scientific) and compared with that of the mercury-vapor lamp (MA-6, KarlSuss) as shown in figure 4(a). The spectrum of the LED shows a single peak at 400 nm while the mercury-vapor lamp of the mask aligner shows multiple peaks at 365 nm (i-line), 405 nm (h-line), and 436 nm (g-line) as expected. Since longer-wavelength UV h-line (405 nm) exposure has been

Figure 1. Schematic diagram of CNC-lithography.

Figure 2. UV-LED system overview: (a) fabricated circuit board with female head connectors, (b) completed UV-LED array light source. The 10 × 10 LED array extends over an area of 5 × 5 cm2.

J. Micromech. Microeng. 26 (2016) 035003

J Kim et al

4

Figure 3. CNC-Lithography system.

Figure 4. UV-LED light characteristics: (a) light intensity distribution as a function of wavelength of a UV-LED and conventional mercury lamp, (b) light intensity as a function of applied current.

J. Micromech. Microeng. 26 (2016) 035003

J Kim et al

5

previously utilized for thick and high-aspect-ratio microstruc-ture formation [17, 18], LEDs with peak emission at 400 nm were selected so as to mimic the performance of these SU-8 processes.

The intensity of the UV-LED is measured using a UV inten-sity meter (400 nm, AB-M). A single LED was set at 5 mm above the intensity meter and connected to a current source. The DC was gradually increased from 20 mA to 100 mA, and the corresponding UV intensity was measured in mW cm−2 as shown in figure  4(b). The voltage during the measure-ment varied between 3 V and 3.4 V. While a minimum applied current of 20 mA showed an intensity of 13 mW cm−2, the intensity linearly increased until the applied cur rent reached approximately 70 mA, after which nonlinearities were observed. The maximum intensity was measured in this experiment was 50 mW cm−2 at 100 mA. In most of the sub-sequent exposures in this paper, an intensity of 23 mW cm−2 was used, as the rated current was 30 mA; this intensity was sufficient for thick and high-aspect-ratio SU-8 fabrication.

As a 10 by 10 UV-LED array consists of 100 discrete LEDs, the light intensity adjacent to the top of LED bulbs and the space between LED bulbs shows high contrast and may result in non-uniform micropatterning over the substrate. A diffuser could be utilized to make the light intensity uniform

[11], but the system may lose light collimation, limiting its ability to fabricate tall and high-aspect-ratio microstructures. In this proposed UV-LED system, the light source is rotated to minimize the non-uniformity of the light intensity without losing light collimation. To measure light intensity distribu-tion, the UV light was projected onto a white screen, and the image was captured on the opposite side of the screen using an optical camera (Cannon, T3i). The camera was set to capture the image continuously for 8 s for both no-rotation and rotation images. The rotation speed of the UV-LED was set at 60 rpm. Figure 5(a) shows the captured images of the no-rotation (left) and the rotation (right) of the LED array. While the contrast between dark and bright areas was clearly distinguished in the image of no-LED rotation, the contrast in the image of LED rotation was markedly reduced. The intensities of the two captured images in figure  5(a) were plotted using software (imageJ, v. 1.45s) as shown in figure 5(b). The data for the light intensity were collected on the diagonal line (solid red line, 80 mm scanning length) as shown in figure 5(a, left). The bold red line shows light intensity from the LED rotation image while the blue line represents one from no rotation. In the plot, the center of the light source was set as an origin, and the dis-tance to the right was set as the positive distance and the left as the negative. If the effective area of UV exposure is ranged

Figure 5. Light intensity distribution: (a) optical photographs of intensity distribution without and with rotation, (b) intensity distribution plot.

J. Micromech. Microeng. 26 (2016) 035003

J Kim et al

6

from −30 mm to 30 mm, the lowest and the highest intensity in the ‘no-rotation’ graph show 51% and 97%, respectively, while 91% and 97% for the ‘rotation’ graph. These data show that in the ‘no-rotation’ case, the applied UV exposure energy can vary as much as a factor of two, potentially resulting in non-uniform micropatterning. Relative percentage contrast over the scanned line was expressed in equation (1):

=−

×I I

IContrast 100%max min

ave (1)

where I represents intensity (%). Since the averages for ‘no-rotation’ and ‘rotation’ data were obtained as 89.9% and 95.4%, respectively, contrast for ‘no-rotation’ was 51% while that from the rotation was 6%. Therefore, the light intensity with LED rotation has shown significant improvement in uniformity over the UV exposure area. Due to the improved uniformity of the UV exposure intensity during rotation as compared with no rotation, rotation of the UV source during

exposure was adopted as the default state, and was utilized for the remaining 3D microstructure demonstrations in this paper.

IV. Computer numerical controlled (CNC) UV lithography

Since CNC UV lithography utilizes adjustable inclined UV-exposure with synchronized light illumination state, the optical path through the photomask is varied and, therefore, enables the fabrication of various 3D microstructures. A sche-matic diagram of inclined UV lithography is described as shown in figure 6. UV light with an inclined angle from the LED passes through air, photomask glass, and photoresist. Since the refractive index of the photoresist is typically higher than that of air and glass, the final refractive angle at the photo-resist is smaller than the incident angle in air. The relationship between the incident angles (δ) in air, the refractive angle (θ) at the photoresist can be expressed by Snell’s law:

Figure 6. Schematic diagram of inclined UV-LED exposure through a photoresist.

Figure 7. Fully automated program for a micro-‘V’ pillar.

J. Micromech. Microeng. 26 (2016) 035003

J Kim et al

7

( )θ θ= − n nsin sin /a1

p p a (2)

( )θ θ θ= °− = °− − n n90 90 sin sin /p a1

a a p (3)

where na and np are the refractive indices of air (1) and the photoresist (1.67 in SU-8), respectively. θp becomes a struc-turally inclined angle after microfabrication. Due to the refraction at the interface between the two different media, the maximum structural inclined angle can be calculated as 53.2° at the incident angle δ = 90° from equation (2). Also, a theoretical range of the structural width of the microfabricated structures resulting from the refraction at the air/photoresist interface can be obtained. The structural width l is expressed as the relationship between the photomask length d and the refractive angle at the photoresist:

( )θ= −l d cos 90 p (4)

Since the theoretical limit of the structurally inclined angle (θp) is 53.2°, the range of l is from 0.8d to d.

Synchronizing the UV light illumination and the move-ment of the tilt-rotational holder facilitates automated UV

lithography by forming lithographically definable microstruc-tures without manual operation. For example, a ‘V’ shape micro-pillar could not previously be fabricated without manu-ally switching the UV-light source to an appropriate sample holder position. Figure 7 shows a test operation of synchroni-zation functions. The first UV exposure (Step 1) is performed at an inclined angle of 45° without rotation. Then the UV-LED is turned off (Step 2) while the sample holder tilts to an inclined angle of −45°. The second UV exposure (Step 3) follows to complete the process. The corresponding scanning electron microscopy (SEM) image of the fabricated structures is shown in figure 7-right. Note that the resultant structural angle differs from the inclined angle of 45° in accordance with equation (3).

V. Results

Various test microstructures have been presented including high-aspect-ratio micropillars, tilt-rotated microstructures, and 3D exotic microstructures using different UV exposure

Figure 8. High-aspect- ratio microstructures using backside UV exposure: (a) a pillar array, (b) a rectangular slab array.

Figure 9. 3D microstructures from a static CNC exposure scheme.

J. Micromech. Microeng. 26 (2016) 035003

J Kim et al

8

schemes: backside UV exposure, static CNC UV exposure, and dynamic CNC UV exposure.

Figure 8 shows high-aspect-ratio microstructures using a backside UV exposure scheme. The backside UV exposure utilized a glass photomask as a photoresist coated substrate: there is no gap between the photomask and the photoresist,

and, therefore, optical diffraction associated with the gap is suppressed. A micropillar array has been fabricated as shown in figure 8(a). 100 μm-circular holes were utilized as a photo-mask, and a UV exposure energy of approximately 22 J cm−2 was applied. The height of the fabricated pillars was measured as 820 μm. Figure  8(b) shows an array of rectangular-slab

Figure 10. Fabrication results: (a) micro-‘horn’ (b) micro-‘cat’s claw,’ (c) micro-‘hi,’ (d) micro-‘calla lily,’ (e) micro-‘cowboy’s cat,’ (f) micro-‘table napkin’.

J. Micromech. Microeng. 26 (2016) 035003

J Kim et al

9

microstructures. The dimension of the rectangular photo pat-tern was 80 μm by 500 µm, and a UV exposure energy of approximately 32 J cm−2 was applied. The height of the fab-ricated microstructures was measured as 1150 μm. The UV exposure doses for both microstructures are comparable to those of the mercury lamp based system [12]. Overall, the pro-posed CNC-lithography system has successfully fabricated the high-aspect-ratio microstructures in a batch process with reasonable quality.

A static CNC UV exposure scheme is composed of mul-tiple discrete UV exposures at different tilt-rotational positions in a fully automated mode. This type of UV exposure was only possible with manual operation in the previous work [12]. Figure 9 shows test microstructures fabricated from the static CNC UV exposure scheme. Various photomask pat-terns including ‘crossbar,’ ‘star,’ ‘rectangular,’ and ‘triangle’ shapes were utilized on the same photomask/substrate. 200 μm thick SU-8 was coated on the substrate and baked at 95 °C for 3 h. Two steps of UV exposure were programmed. The first exposure was performed at ° °0 and 0tilt rotation, in which the UV light is perpendicular to the substrate and followed by the second exposure where the tilt position of the sample holder was moved to 65°. A UV exposure energy of 12 J was applied for both exposures. The sample was post-baked at 95 °C for 50 min and developed in propylene glycol methyl ether acetate (PGMEA) solution for 20 min to complete the fabrication process. More than 150 microstructures as a batch were successfully fabricated as shown in figure  9-center. Some fabricated microstructures are highlighted in enlarged views of the SEM images at each corner of figure 9. The struc-turally inclined angle from the first exposure was measured as 90° (vertical) and the second exposure as 32°.

A dynamic CNC exposure scheme utilizes multiple com-binations of continuous and discrete UV exposures at various tilt-rotational positions. In this scheme, the CNC lithography creates not only the tilt-rotational UV traces from the sample holder movement but also controls the applied UV doses and, therefore, the height of the fabricated microstructures from the speed of the sample holder. The height control feature could probably be done in the manual operation mode described in the previous work [12] at the expense of speed of fabrication. Moreover, the height control feature with UV light illumination synchronization for the discrete UV exposure is not possible with the previous work. From those dynamic operations, var-ious 3D microstructures have been successfully fabricated as shown in figure 10. Figure 10(a) shows a micro-‘horn’ struc-ture. The pillar at the center of the horn was first UV-exposed at °0tilt & rotation. Exposure for the circular horn region then fol-lowed. Continuous rotation (15 rpm) at a tilt angle of 55° was applied during exposure. Figure 10(b) shows a micro-‘cat’s claw’ structure. UV exposure was performed at five evenly spaced positions separated by 72° while the tilting angle was fixed at 45°. A micro-‘hi’ structure is shown in figure 10(c). Two combinations of UV exposure schemes were adopted. During the first exposure, tilting angle was varied from 0° to 15°, and then the sample holder was moved to the tilting

position at 59° for the second exposure. Note that the UV light was turned off while the sample holder was moving to 59°. No rotation was used in this fabrication, but the stepper motor was kept powered to hold the sample plate. Figure 10(d) presents a micro-‘calla lily.’ The flowercore region was UV-exposed by varying the tilting from 0° to −40°. The sample holder was moved to −30° tilt position with the light off. The second exposure for the floral leaf region was applied while varying the rotation angle from −30° to 30°. While figures 10(a)–(d) utilized a circular photomask pattern, a rectangular shape photomask was used to fabricate the microstructures as shown in figures 10(e) and (f). A micro ‘cowboy’s hat’ is presented in figure 10(e). The rim of the hat was UV-exposed using a continuous rotation at °40tilt while the head region of the hat utilized the UV exposure scheme, in which two discrete UV exposures were applied at 25° and −25° with 0° rotational angle. Figure 10(f) presents an array of micro-‘table napkin’ structures. In the image, the semi-circular region at the back part was first UV-exposed with a fixed tilt angle of −30° and a varied rotation angle of −45° to 45°. Then the sample holder was moved to 30° tilt angle for the rest of structure fabrica-tion. The rotation angle varied from −45° to 45° while the tilt angle was synchronously moved to 35° at the rotation angle of 0–45°, then back to 30°.

VI. Conclusion

CNC-lithography, the synchronized motion and activation of light source and sample holder during lithographic exposure, has been introduced as a versatile 3D micropatterning process. The system comprises 100 UV-LEDs, a tilt-rotational sample holder, and a control unit. The CNC lithography system is fully automated to fabricate various 3D exotic microstructures. Several test microstructures have been successfully fabricated and demonstrated using the CNC lithography scheme. Tall and high-aspect-ratio microstructures were demonstrated, as the height of the microstructure was 1.15 mm with an aspect ratio of 14. A batch fabrication of 3D microstructures was also demonstrated using a discrete CNC lithography scheme. Exotic 3D microstructures such as new 3D shapes com-pared to the previous work have been successfully fabricated including micro-‘hi,’ micro-‘horn,’ micro-‘calla lily’ and micro-‘cowboy’s hat’ as well as the batch-fabricated array of the micro-‘table napkin’. CNC lithography has great potential to create structures in diverse application areas such as micro-filters, scaffold structures for cell manipulation, and antenna structures for GHz and THz range applications.

Acknowledgments

We thank the Singh Center for Nanotechnology at University of Pennsylvania for facility use and MSMA group members and alumni for valuable discussions and advice. We also thank Mr Richard Shafer at Georgia Institute of Technology for the tilt-rotational sample holder.

J. Micromech. Microeng. 26 (2016) 035003

J Kim et al

10

References

[1] Tang Y, Peterson R L and Najafi K 2013 Technology for fabricating dense 3-D microstructure arrays for biomimetic hair-like sensors 2013 IEEE 26th Int. Conf. Micro Electro Mechanical Systems (MEMS) (Taipei, Taiwan, 20–24 January 2013) pp 355–8

[2] Chretiennot T, Dubuc D and Grenier K 2013 A microwave and microfluidic planar resonator for efficient and accurate complex permittivity characterization of aqueous solutions IEEE Trans. Microw. Theory Tech. 61 972–8

[3] Altuna A et al 2013 SU-8 based microprobes for simultaneous neural depth recording and drug delivery in the brain Lab Chip 13 1422–30

[4] Maeda Y, Terao K, Suzuki T, Shimokawa F and Takao H 2013 A post-complementary metal oxide semiconductor formation process of high-adhesiveness SU-8 structures for reliable fabrication of integrated microelectro mechanical system sensors Japan. J. Appl. Phys. 52 06GL18-1–5

[5] Nielson R, Kaehr B and Shear J B 2009 Microreplication and design of biological architectures using dynamic-mask multiphoton lithography Small 5 120–5

[6] Chen V W et al 2014 Three-dimensional organic microlasers with low lasing thresholds fabricated by multiphoton and UV lithography Opt. Express 22 12316–26

[7] Kang H W, Jeong Y S, Lee S J, Kim K S and Yun W S 2012 Development of a compact micro-stereolithography (MSTL) system using a blu-ray optical pickup unit J. Micromech. Microeng. 22 115021 (9 pp)

[8] Zheng X, Deotte J, Alonso M P, Farquar G R, Weisgraber T H, Gemberling S, Lee H, Fang N and Spadaccini C M 2012 Design and optimization of a light-emitting diode projection micro-stereolithography three-dimensional manufacturing system Rev. Sci. Instrum. 83 125001

[9] Kim J, Ahn H C, Cheng X, Kim K T, Kim G H and Yoon Y K 2010 Fabrication of multiple height microprobes using UV lithography on timed-development-and-thermal-reflowed photoresist Hilton Head Workshop 2010: A Solid-State Sensors, Actuators and Microsystems Workshop (Hilton Head Island, 6–10 June 2010) pp 202–5

[10] Kim J K, Kim C, Allen M G and Yoon Y K 2015 Fabrication of 3D nanostructures by multidirectional UV lithography and predictive structural modelling J. Micromech. Microeng. 25 1–10

[11] Hafeez H, Ryu H Y, An I S, Oh H, Ahn J H and Park J G 2015 Dimensionally controlled complex 3D sub-micron pattern fabrication by single step dual diffuser lithography (DDL) Microelectron. Eng. 143 25–30

[12] Kim J K, Allen M G and Yoon Y K 2011 Computer controlled dynamic mode multidirectional UV lithography for 3-D microfabrication J. Micromech. Microeng. 21 1–14

[13] Bing C Y, Mohanan A A, Saha T, Ramanan R N, Parthiban R and Ramakrishnan N 2014 Microfabrication of surface acoustic wave device using UV LED photolithography technique Microelectron. Eng. 122 9–12

[14] Huntington M D and Odom T W 2011 A portable, benchtop photolithography system based on a solid-state light source Small 7 3144–7

[15] Kim J K, Yoon Y K, Mier T C and Allen M G 2015 CNC-lithography: computer-controlled multidirectional light-motion synchronized lithography The 18th Int. Conf. on Solid-State Sensors, Actuators and Microsystems, Transducer 2015 (Anchorage, USA, 21–25 June 2015) http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=7181184&tag=1

[16] Kim J, Paik S-J, Herrault F and Allen M 2012 UV-LED lithography for 3-D high aspect ratio microstructure patterning Hilton Head Workshop 2012: A Solid-State Sensors, Actuators and Microsystems Workshop (Hilton Head Island, 3–7 June 2012) pp 481–4

[17] Yang R and Wang W 2005 A numerical and experimental study on gap compensation and wavelength selection in UV-lithography of ultra-high aspect ratio SU-8 microstructures Sensors Actuators B 110 279–88

[18] Kim J K, Herrault F, Yu X and Allen M G 2012 Microfabrication of air core toroidal inductor with very high aspect ratio metal-encapsulated polymer vias The 12th Int. Workshop on Micro and Nanotechnology for Power Generation and Energy Conversion Applications (PowerMEMS) (Atlanta, GA, USA, 2–5 December 2012) pp 30–3

J. Micromech. Microeng. 26 (2016) 035003