INSTITUTE OF PHYSICS PUBLISHING

JOURNAL OF MICROMECHANICS AND MICROENGINEERING

doi:10.1088/0960-1317/16/10/012

J. Micromech. Microeng. 16 (2006) 2000–2005

A high volume precision compression molding process of glass diffractive optics by use of a micromachined fused silica wafer mold and low Tg optical glass A Y Yi1, Y Chen1, F Klocke2, G Pongs2, A Demmer2, D Grewell3 and A Benatar1 1

Department of Industrial, Welding and Systems Engineering, The Ohio State University, 210 Baker Systems Building, 1971 Neil Ave, Columbus, OH 43210, USA 2 Fraunhofer Institute for Production Technology (IPT), Aachen, Germany 3 Department of Agricultural and Biosystems Engineering, Iowa State University, Ames, IA, USA E-mail: [email protected]

Received 13 April 2006, in final form 30 June 2006 Published 25 August 2006 Online at stacks.iop.org/JMM/16/2000 Abstract Recent advances in compression molding of glass optical elements for mass production offer the potential of extending this technology to elements with micro and nano scale features. In this research, glass diffractive optical elements (DOEs) with lateral features in the order of 10 µm and vertical height of 330 nm were fabricated using a fused silica glass mold and a special low Tg (glass transition temperature) glass material K-PG325. Molded DOEs were studied using an atomic force microscope (AFM) and scanning electron microscope (SEM) to evaluate the glass molding process capability. Optical testing of the molded DOEs was a further demonstration of the effectiveness of the molding process for high volume micro and diffractive optical component fabrication. The combination of two high-precision, high-volume processes, i.e., semiconductor batch process for optical mold making and glass molding for DOE replication, is an effective alternative manufacturing method for high-quality, low-cost optical components. The reported experiment is a detailed illustration of the glass molding process capability. With further process optimization a robust manufacturing process can be developed for mass production of diffractive and micro glass optical elements. (Some figures in this article are in colour only in the electronic version)

1. Introduction Growing demands on high-speed data, voice and video signal transmissions depend on low-cost and reliable components to convert optical signals to electrical signals or vice versa. Microlenses and diffractive optical elements (DOEs) and mirrors are increasingly playing an important role in these 0960-1317/06/102000+06$30.00

applications [1, 2]. Currently, these optical components are either fabricated using cleanroom micromachining technology or a direct ultraprecision machining process; therefore, the manufacturing cost remains very high. The method discussed in this paper is different from many traditional fabrication processes [1–3], where instead of direct manufacturing, a replication method based on the compression glass molding

© 2006 IOP Publishing Ltd Printed in the UK

2000

A high volume precision compression molding process of glass diffractive optics

process is presented [3, 4]. In previous publications, an experimental study and numerical simulation of the compression glass molding process have been described [3, 4]. It was shown that various sizes of diffraction limited aspherical glass lenses can be molded using the compression molding technology. However, very limited attempts have been made toward making micro optical and diffractive optical elements where glass molding maybe an excellent choice. Among the pioneering works that were carried out in this field is the use of a micromachined silicon wafer in structuring borosilicate glass substrates with surface tension on the glass surface, by Merz and his colleagues [5]. In this process, the microlens surface had no direct contact with micro cavities on the silicon wafer. One of the difficult issues related to DOE fabrication is the manufacturing cost. Although direct fabrication using photolithography can be used to mass produce these elements, the cost remains high due to the complex procedures involved in the mask making and the following lithography process. Furthermore, there are concerns about the chemicals that are used in photolithography. For cleanroom micromachining processes, there are limitations on optical material selection due to strict requirements for machinability and cleanliness. Injection molding of plastic optical elements allows a substantial reduction in the manufacturing cost, but plastic optical elements lack in performance compared to glass optical elements. Meanwhile, diffraction-limited optical elements have been successfully molded by the compression molding process using common optical glasses, thus making it an attractive alternative manufacturing method for large-volume, high-quality and low-cost production of optical elements. These glass optical elements are finding more and more applications in opto-electro-mechanical systems [6]. Currently, glass components with micro/nano scale features have to be manufactured using aggressive etching processes [7, 8] or machined using mechanical machining processes which are not well suited for high-volume production [9]. In another case, glass chips for the micro total analysis system (µTAS) are often used in pharmaceutical research where this molding process can drastically reduce the production cost while maintaining high micro and nano scale fidelity of the glass components [10]. Furthermore, this process can be part of a manufacturing process for microelectronic devices where micro and diffractive optical elements can be directly fabricated and integrated into an electronic circuit thus lowering the overall manufacturing cost. In summary, the goals of this study were to use standard lithography techniques to produce a glass diffractive optical element, which was then used as a mold for compression molding of low glass transition temperature glass. The diffractive pattern used in this study was designed to project a simple square pattern when illuminated with a collimated beam, such as a laser pointer. The basic setup for beam shaping with diffractive elements and a possible application in precision welding or hot embossing is shown in figure 1 [11]. The design of the diffractive optical element used here was thoroughly investigated by one of the authors [11]. A general description of methods for designing and fabricating diffractive optical elements can be found in [11, 12].

Figure 1. Schematic of the DOE welding setup.

2. Experimental procedures 2.1. Mold design and fabrication The diffractive mold was fabricated on a 100 mm diameter wafer that was 1 mm thick, with a quarter wavelength flatness. To simplify the mold making process, a two-level design was used. Standard lithography techniques were used to etch the fused silica to the desired depth. The photoresist was SPR2207.0 and it was deposited with a thickness of 10 µm in order to assure that the photoresist withstood the etching procedure. In the aligner, a dummy (silicon) wafer was placed under the sample to prevent reflected UV from bleeding through the clear fused silica wafer. The wafer with photoresist was exposed to UV radiation for 3 s at a density of 15 mW cm−2. The wafer was then developed for 2 min and water rinsed followed by a spin dry. An LAM 490 unit (LAM and Associates) was used for reactive ion etching of the wafer. The pressure during etching was 2.6 Torr with an RF power of 600 W. The gas composition during etching was 120, 40 and 45 cm3 min−1 for He, CHF3 and CF4, respectively, with a gap of 0.38 mm between the wafer and electrode. The etch rate was measured as a function of time and least-squares regression was used to fit the data to a quadratic equation, which was then used to determine the proper etch time for the desired depth [11]. 2.2. Glass molding experiment The DOE molding experiments discussed in this paper were performed on a Toshiba GMP 211 V machine at the Fraunhofer Institute for Production Technology (IPT). The detailed descriptions of the press and glass molding process can be found elsewhere [3, 4, 13]. The focus of this paper is to study the implementation of the glass molding process to DOE fabrication. The fused silica glass mold was placed between two glassy carbon wafers (manufactured by Tokai Carbon Co. Ltd, Tokyo, Japan) as shown in figure 2. The glassy carbon wafers were optically polished and served as support for the fused silica wafer and the second (top) mold surface. The shaded component in the figure is the low Tg optical glass disc (Tg = 285 ◦ C). K-PG325 is a low Tg glass developed by Sumita Co. (Saitama, Japan) for compression molding applications. The refractive index of the glass is 1.506 70. The thermal expansion coefficient (TEC) of this glass is 17.3 × 10−6 ◦ C−1, which is not uncommon for optical glasses but larger than some optical 2001

Fused silica mold

Upper platen

1000

250

800

200 600 150 400

100 Temperature (°C)

200

Load (N)

0 0

Forming force

2.3. Forming test conditions The test DOEs molding conditions were selected based on previous experience. These conditions were found to be effective but they are not necessarily the optimal conditions. The samples were molded under the following conditions. (1) Cycle started and lower mold was moved up to the heating position (2–3 mm from making contact), followed by heating of the mold assembly and glass blank to the molding temperature of 325 ◦ C at a rate of 3.8 ◦ C s−1. (2) Initial forming at 325 ◦ C was performed when the lower mold was pushed upward at a velocity of 0.5 mm min−1. As shown in figure 3, the forming force was kept constant at 1 kN for initial forming for about 1 min while the temperature was maintained at 325 ◦ C. During the initial forming period, the lower mold position was automatically adjusted to maintain a constant forming load using an adaptive control that is incorporated into the glass molding machine design. Adaptive controls are often used in the manufacturing process control when certain variables such as forming force need to be maintained at some values but the machine axes are controlled using position feedback elements [15]. The detailed process conditions are illustrated in figure 4, where the forming load was plotted along with the mold position to show their relationship during a molding cycle.

400

600

0 800

Figure 3. Molding conditions: temperature and load histories.

Position (mm)

glasses (for example, for BK-7 the TEC is 7.5 × 10−6 ◦ C−1 and for fused silica it is only about 0.47 × 10−6 ◦ C−1), but it is still much smaller than that of optical plastics (for example, the TEC of polymethylmethacrylate or PMMA is 50–100 × 10−6 ◦ C−1). This glass is also slightly more brittle than BK-7 and is easier to chip. The schematic in figure 2 illustrates the heating and forming phases of the process. The process setup is similar to that of a hot embossing process in its simplicity [14]. The glass blank used was optically polished on both sides and placed manually on the lower mold. The total molding cycle time was about 12 min, matching the process conditions for macro- and meso-scale optical component fabrication [4]. In a real production setup, the lower mold plate (which was used to support the wafer) can be designed to match a standard wafer. Robotic arms can be integrated for loading and unloading the finished optics. Details of the molding process conditions are explained below.

200

Time (second)

Figure 2. Illustration of a DOE molding process.

2002

1200

300

50

Lower platen

Mold motion

350

32

1200

31

1000

30

800

29 600 28 400

27

Position (mm)

26

Load(N)

25 0

Load (N)

Glassy carbon wafer

Temperature (°C)

Forming force

Load (N)

A Y Yi et al

200

0 100 200 300 400 500 600 700 800 Time (second)

Figure 4. Molding conditions: load and position histories.

(3) Slow cooling of the glass blank and mold assembly was started with the forming load at 500 N and the cooling rate was slowed down to 0.2 ◦ C s−1. Figure 4 shows that the lower mold was continuously being pushed upward to maintain the constant forming load. This is because glass becomes viscoelastic at temperatures around Tg. The slow cooling was extremely important to the final lens quality as was demonstrated in experiments that were published earlier [4, 13]. At the end of the slow cooling stage, the mold and the glass blank were not yet separated. (4) Once the mold and part temperature reached approximately 280 ◦ C, rapid cooling of the glass blank and mold was initiated by increasing the cooling rate to 1.0 ◦ C s−1 to a release temperature of 140 ◦ C. A 200 N forming load was maintained until the glass blank and molds were separated in the middle of the rapid cooling process when the lower mold was pulled downward at about 660 s as shown in figure 4. (5) The finished DOE was then manually removed from the mold assembly and subsequently cooled to room temperature. The DOE was easily separated from the upper mold during each of the experimental runs. No sticking was observed between the glassy carbon wafer surface and the low Tg glass blank. At the end of the cooling period after reaching room temperature, the DOE was also separated from the fused silica molds (lower mold) without operator intervention. This finding has shown that using the cleanroom technology one can easily produce microlens array or diffractive optical element molds and then use them for manufacturing of lowcost, high-fidelity optical components by compression molding. During each experiment, a vacuum was applied at the beginning to ensure that no air remained in the gap between the DOE mold and low Tg glass blank. This was immediately

A high volume precision compression molding process of glass diffractive optics

(a)

(b)

Figure 5. SEM scans of (a) master chrome glass mask and (b) the molded DOE.

followed by nitrogen purge to further remove oxygen residual. Nitrogen was also used to control the cooling rate during the third and fourth stages. The molding force was varied during the molding process as shown in figure 3 as well as during the cooling period to minimize the residual stresses in the optical elements that results from the forming action and temperature change. To ensure complete replication of the diffractive pattern, the glass sample was pressed to produce a displacement of 440 µm after the initial contact between the glass sample and the fused silica mold was made. The lower mold was supported by the post that is controlled by a feedback servo system, which enabled the positions of the lower mold to be monitored and controlled in real time [16].

(a)

(b)

Figure 6. AFM scans of the (a) fused silica mold (female/negative) and (b) molded low Tg glass DOE (male/positive).

3. Results and discussion

3.1. Molded glass DOE To observe the surface morphology, a molded DOE was sputter coated with Pt–Pd for 3 min under 20 mA current using Emitech K550X coater (Empdirect Products Inc., 704/17033 Butte Creek, Houston, TX 77090). The Pt–Pd coating was needed to prevent the dielectric sample from charging inside the SEM chamber. We selected the Pt–Pd coating process because it is readily available in the SEM lab in our college. Figure 5(a) shows the original chrome mask where the feature size is 10 µm. Figure 5(b) shows the SEM scan of the molded DOE verifying the compression molding process quality. Note that the width of the patterns is exactly 10 µm, which is the design value used for mold fabrication. These two SEM scans were performed using different SEMs at the Ohio State University. The depth of the DOE and surface roughness were also measured quantitatively using an AFM. 3.2. Replication accuracy The molded DOEs were also measured for geometric accuracy and surface finish. Figure 6 shows the molded DOE and

500 400 Height (nm)

At the end of the cooling cycle, molded DOEs were inspected and no visual change of color to the fused silica wafer or the glass samples was observed under high-intensity light. The glass sample thickness was measured and confirmed to be reduced by the amount of compression created by the lower mold. The glass samples and silica mold were cleaned using alcohol before the measurements were taken.

300 200 Mold Sample

100 0 -100 0

5

10 15 20 Distance (um)

25

30

Figure 7. AFM line scans of the mold and molded DOE.

the corresponding mold location measured on a Digital Instrument’s AFM (Veeco NanoMan Dimension 3100 with NanoScope IV SPM Control Station). To ensure a valid comparison, AFM scans were performed on the same area, i.e., figure 6(a) is the negative (female) image of figure 6(b) (positive/male). It is shown in figure 6 that molded DOEs resemble the fused silica mold in nanometer scale with slight rounding of the sharp features. As shown in the AFM line scan of the mold and the molded DOE in figure 7, the side walls of the mold are relatively vertical as a result of the anisotropic etching from the reactive ion etching (RIE) process. Using the AFM line scans, the average of the peak-to-valley distances on the molded DOE is measured at about 316 nm while the same value for the mold is about 398 nm, which gives a total deviation of around 82 nm in the vertical direction. This discrepancy of about 20% may be attributed in part to 2003

A Y Yi et al CCD line scan camera

DOE

Laser

(a)

Rotary table

(b)

Figure 8. Image formation comparison for (a) the fused silica DOE mold and (b) the molded low Tg glass DOE.

(a)

differences in thermal expansion between the mold and blank materials as well as to structural relaxation (structural change due to temperature change as in [17]) of the glass material when it undergoes a temperature change of approximately 300 ◦ C. It is concluded from figure 7 that the lateral dimension on the glass sample matches the fused silica mold to a much higher degree. Specifically, the sample width is 19.34 µm and the original mold width is 18.75 µm; therefore, the width difference is only ∼2.5% between the mold and the sample.

The AFM line scan in figure 7 was also used to measure the surface roughness. The optical surface roughness on the mold is Ra = 24 nm (arithmetic average) and on the molded surface it is Ra = 22 nm, indicating that the compression molding process has a slight smoothing effect on the glass surface. This finding is similar to that of a molded refractive optical component as was demonstrated in previous publications [4, 18]. However, in those investigations, the conventional polished optical molds had much better surface roughness than the etched fused silica surface.

Single axis controller

PC

CCD camera

(b)

Figure 9. Optical performance test setup. (a) Schematic of optical test setup and (b) schematic of data acquisition for optical testing. 1.2

Normalized Intensity

3.3. Surface roughness

Rotary table

Hub

Fused silica mold

1 0.8 0.6 0.4 0.2 0 200

DOE

700

1200

1700

Lateral Distance (pixel)

3.4. Image formation The setup shown in figure 1 was used to perform the image formation test. A HP-5517B He–Ne laser (Hewlett Packard, now Agilent Technologies, Inc., 395 Page Mill Road, Palo Alto, CA) was used. The HP-5517B laser can provide a continuous beam of 1 mW peak power at a wavelength of λ = 632.99137 nm. The molded DOE was placed about 25 mm away from the He–Ne laser. The diffraction pattern was projected onto a dark screen that was about 2 m away from the DOE for observation. Figure 8(a) shows the image produced by the DOEs on the fused silica wafer (mold) and figure 8(b) shows the image from the molded glass DOE. The image produced by the molded DOEs closely resembles that of the original fused silica mold. To quantitatively compare the fused silica mold and the glass DOE, a more sophisticated system was devised and the details are explained next. 3.5. Optical performance The experimental setup used to analyze the optical image formation was also similar to the system shown in figure 1. The same laser source (HP-5517B) was also used in this 2004

Figure 10. Optical performance test results.

experiment. Both the fused silica mold and the molded DOE were placed about 350 mm away from the laser. The diffraction patterns were projected onto a Charge Coupled Device (CCD) line scan camera that was mounted 355 mm away from the DOE (Model 1000 line scan camera Entwicklungsb¨uro, Reinholdstraße. 5 D-12051 Berlin, Germany). The 1000 model is equipped with a Sony line scan CCD sensor ILX553. The CCD sensor has 5,150 pixels, each pixel has a sensing area of 7 × 7 µm2. Figure 9(a) is the schematic of the optical testing system. To scan an entire image, the rotary table (Soloist single-axis rotary stage by Aerotech, Inc., Pittsburgh, PA) was operated through a PC to move at a fixed speed. The onboard circuit of the CCD camera system allows a line scan to be acquired at each stoppage at 16 kHz and the results were uploaded to a PC as schematically illustrated in figure 9(b). In figure 10, the diffraction patterns from both the fused silica mold and glass DOE were plotted. For clarity only one line scan (a horizontal scan in figure 8) for each component was used. The dashed line represents the diffraction pattern from the DOE and solid line from the mold. A 15 point moving

A high volume precision compression molding process of glass diffractive optics

average was performed to both scans to remove high frequency noise. It can be seen that the optical performance of the molded DOE matched well to the fused silica mold.

Foundation under grant no 0547311. Any opinions, findings and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.

4. Conclusions Through design, fabrication and measurement of a compression molded low Tg glass DOE, we have demonstrated that compression molding can be a promising process for high volume production of precision DOEs with micro and nano scale features. Geometrical and optical characterizations of the molded DOEs were demonstrated using different measurement techniques. The proposed strategy for DOE fabrication has several inherent advantages over existing technologies. First, it is flexible in that there are several hundred optical glasses available versus only a few optical plastics. Second, the molded DOEs can preserve the optical quality of the mold surfaces; therefore, no post-molding processes are needed, i.e., it is a net shape manufacturing process. Third, the compression molding process maybe integrated with other semiconductor processes due to its low molding temperature. Fourth, this method can be applied to fabricating many different optical components thus making it a very powerful tool to mass produce various optical elements at low cost, which opens the door for enhanced and more affordable optical systems. Finally, the compression molding process can be used to fabricate other components with micro and nano scale features at an affordable cost, e.g., microfluidic devices for biomedical applications. We will further investigate the replication accuracy using optical glass blanks and other mold materials (e.g. glassy carbon or other hard materials) in the future. Other efforts will also include molding of glass optics using more common glass materials. It will include error analysis and process optimization in order to make this process better suited for industrial applications. In addition, microfluidic devices using common glasses will also be molded and evaluated for biomedical applications.

Acknowledgments The authors would like to express their appreciation to Fraunhofer Munich for financial support for the glass molding experiments at Fraunhofer IPT through the Prof X2 program (AYY). This work is also supported by the Deutsche Forschungsgemeinschaft (DFG) within the SFB/TR4 ‘Process Chains for the Replication of Complex Optical Elements’. The authors would also like to thank MicroMD at the Ohio State University for assistance in the fused silica wafer mold fabrication. Thanks also goes to Todd Lizotte of Hitachi Via Mechanics (USA) for his help with diffractive optical design. The research is also supported by National Science

References [1] Wang M R and Su H 1998 Laser direct-write gray-level mask and one-step etching for diffractive microlens fabrication Appl. Opt. 37 7568–76 [2] Chase H, Handschy M A, O’Callaghan M J and Supon F W 1995 Improvement of spatial light modulator optical input/output performance using microlens arrays Opt. Lett. 20 1444–6 [3] Yi A Y 2003 Optical Fabrication (New York: Dekker) [4] Yi A Y and Jain A 2005 Compression molding of aspherical glass lenses—a combined experimental and numerical analysis J. Am. Ceram. Soc. 88 579–86 [5] Merz P, Quenzer H J, Bernt H, Wagner B and Zoberbier M 2003 A novel micromachining technology for structuring borosilicate glass substrates Transducers ’03 vol 1 pp 258–61 [6] Klocke F, Pongs G and Heselhaus M 2004 Komplexe optiken aus der presse Photonik pp 71–3 [7] Lin C H, Lee G B, Lin Y H and Chang G L 2001 A fast prototyping process for fabrication of microfluidic systems on soda-lime glass J. Micromech. Microeng. 11 726–32 [8] Stjernstrom M and Roeraade J 1998 Method for fabrication of microfluidic systems in glass J. Micromech. Microeng. 8 33–8 [9] Diepold T and Obermeier E 1996 Smoothing of ultrasonically drilled holes in borosilicate glass by wet chemical etching J. Micromech. Microeng. 6 29–32 [10] Yagyu H, Sugano K, Hayashi S and Tabata O 2005 Micropowder blasting with nanoparticles dispersed polymer mask for rapid prototyping of glass chip J. Micromech. Microeng. 15 1236–41 [11] Grewell D 2005 Modeling of molecular healing for micro-laser welding of plastics with diffractive optical elements as spatial modulators PhD Thesis The Ohio State University, Columbus, OH [12] O’Shea D, Suleski T J, Kathman A D and Prather D W 2004 Diffractive Optics, Design, Fabrication and Test (Bellingham, WA: SPEI Press) pp 57–83 [13] Jain A and Yi A Y 2005 Numerical modeling of viscoelastic stress relaxation during glass lens forming process J. Am. Ceram. Soc. 88 530–5 [14] Yao D, Virupaksha V L and Kim B 2005 Study on squeezing flow during nonisothermal embossing Polym. Eng. Sci. 45 652–60 [15] Tlusty J 1999 Manufacturing Process and Equipment (Reading, MA/London: Addison-Wesley/Longman) p 928 [16] Toshiba 2001 Toshiba Product Information Brochure Toshiba High Precision Machine Tools [17] Soules T F, Busbey R F, Rekhson S M, Markovsky A and Burke M A 1987 Finite element calculation of stresses in glass parts undergoing viscous relaxation J. Am. Ceram. Soc. 70 90–5 [18] Maschmeyer R O, Hujar R M, Carpenter L L, Nicholson B W and Vozenilek E F 1983 Optical performance of a diffraction-limited molded-glass biaspheric lens Appl. Opt. 22 2413–15

2005