The fabrication of microstructured polymer optics enables a multitude of new options in the design of technical optics. However, challenges arise along the varying process chains of mold insert fabrication, integration into molding tools, replication by means of injection compression molding and metrology. In order to study the effects, diffractive optical elements (DOE) and microlens arrays (MLA) are fabricated using two different process chains. DOEs are fabricated using a laser direct writing (LDW) based mold insert fabrication. The MLA mold insert is produced using ultra-precision milling (UP-milling). Both optical parts are replicated using injection compression molding. The occurring effects are discussed and the results show, that with complete process control high quality microstructured polymer optical parts can be produced and characterized.

Introduction

Polymer optics gain increasing popularity in optical applications, competing with traditional glass lenses. Due to the technological advantages of polymer optics regarding manufacturability, complex free-form optics as well as microstructured optics can be fabricated at a significantly lower price as glass optics [1]. Examples for microstructured polymer optics are Fresnel lenses, diffractive optical elements (DOE) and microlens arrays (MLA). However, to measure up to with advanced, traditional glass optics, high quality polymer optics are necessary. Therefore high process control of the fabrication chain is mandatory, especially when microstructured polymer optics are to be produced, in order to meet the requirements of applications such as illumination, imaging, measurement systems and sensors.

A main advantage of polymer optics is a fast and cheap replication of the parts by means of injection molding and hot embossing. Injection compression molding, a process variation of injection molding employing a compression stamper to create a homogeneous pressure distribution, has proofed to result in high quality optical lenses [2]. Nonetheless, controlling this process with all its parameters in combination with tool design and integration of micro structured stampers is mandatory to achieve this high quality also for technical optics with microstructures.

While the fabrication of microstructured polymer optics remains complex and comes with several challenges, also metrology presents a crucial aspect in the process as discussed in this work.

Focusing on those aspects, two different approaches for the fabrication of micro-structured mold inserts are described and evaluated. On the one hand, a DOE is fabricated, using laser direct writing (LDW) and an electroplating process. On the other hand a MLA is produced using ultra-precision milling (UP-milling).The main advantages and disadvantages as well as the resulting process chains are explained. Both elements are replicated using injection compression molding and characterized using different measurement systems.

Challenges

  • (1)

    The fabrication of the mold insert is the first important step in the process. The choice of the microstructuring technology is crucial. The size of the microstructures and the area to structure are opposing factors. As a general rule, it can be said, that structuring technologies for structure sizes in the low micro- and submicrometer range tend to struggle when large areas need to be structured, since processing time increases rapidly. Long-processing time is undesirable due to economic reasons as well as inaccuracies due to temperature drifts during the process [3]. This is true in particular when nonbinary structures like microlenses, continuous diffractive structures, and Fresnel lenses need to be produced. Examples for these technologies are LIGA, nanoimprint, and e-beam writing. On the other hand, structuring technologies that can fabricate larger areas in the millimeter, centimeter, or even meter range tend to be limited in the structure size to several micrometers or more. Examples are ultra-precision machining, electrodischarge machining and LDW. The choices of available technologies are further reduced, when 3D- and freeform substrates need to be structured.

  • (2)

    The replication of micro- and nano‐structures by means of injection and injection compression molding is complex and influenced by a multitude of factors [4]. In contradiction to the intuitive conclusion, the replication of microstructures is significantly more difficult compared to nanostructures, in particular for structures with high aspect ratio, due to the limitations of mold flow into small structures compared to surface effects [5]. To obtain complete filling of microstructures, the process needs to be optimized regarding the molding parameters, which result in most cases in small process windows. On the other hand, nanostructures can be replicated comparably easy, since the surface effects are only limited by the polymer's ability to replicate the surface morphology.

  • (3)

    Another challenge in the fabrication of microstructured polymer optics is the metrology to evaluate the quality of the mold inserts as well as the replicated components. The combination of properties like transparent materials, microstructures, and in some cases large areas and steep angles are challenging for most of the measurement systems [6]. Only the combination of different systems allows the full evaluation and characterization of the components.

Methods

Mold Inserts Fabrication.

The LDW process performed in this work uses a scanning beam interference lithography method with a 405 nm wavelength of the laser. Thereby a photoresist-coated curved glass substrate is structured. The fabrication of DOEs by means of LDW comes with several advantages, which makes the technology a suitable choice for the production of a microstructured mold insert with structure sizes as low as 1–3 μm. Areas with several micro- to centimeters in lateral dimensions even on curved substrates can be structured [7]. Furthermore, it is possible to create continuous microstructures to increase the efficacy of DOEs compared to binary DOEs. Compared to alternative fabrication methods like ultra-precision turning, LDW has some advantages like reduced edge roundness of the microstructures, no wear as well as the possibility to produce varying structures sizes and shapes in the same process. Employing this method, a curved master substrate is produced with a diameter of 25 mm and continuous diffractive microstructures with varying structure sizes down to 5 μm in lateral dimension and a structure height of 1.6 μm. A profile segment of the diffractive structure on the glass master is shown in Fig. 1.

However, the LDW process also comes with drawbacks. Depending on the laser system, the structures are often limited to rotational symmetric areas. Furthermore, a subsequent electroplating process is mandatory to create a mold insert made of metal, since LDW can only be used for the master fabrication in photolithographic resin. Successful transfer of the master structures into a solid mold insert by means of electroplating requires the positioning of the master substrate on a holder, including mounting and centering structures for the subsequent integration into a molding tool. This fact in combination with uneven master surface and holder leads to an uneven, cauliflower like growth of material (Fig. 2(a)).

However, addressing those challenges correctly, the casting accuracy can be extremely high and even very small micro- and nano‐structures can be transferred into a solid mold insert (Figs. 2(b) and 3). Since the master substrate had a curved surface, the minimum thickness of the electroplated Nickel had to be 3 mm. For the integration into a molding tool, the mold insert has to be centered in regard to the microstructures. All of these processes have to be performed very accurately to keep the tolerance chain as small as possible.

The fabrication of a MLA mold insert can be performed by means of UP-milling. The MLA fabricated for the evaluation has a structured area of 13 × 15 mm2 with a total of 12,000 microlenses. Each of the lenses has a radius of 1 mm. Since diamond cutting-tools are used, nonferrous materials are necessary to reduce the wear of the diamond. Therefore, a mold insert made of steel is coated with nickel-phosphor. A major advantage of the UP-milling process is a direct structuring of the mold insert without the necessity of a subsequent electroplating process. Further reasons to use UP-milling for the fabrication of MLA mold inserts are the resulting optical surface quality with Ra <10 nm, less limitations in the substrate shape as well as a reasonable processing time. The achievable structure size is limited by the available diamond tools to a couple of micrometers.

To achieve best results, it is important to evaluate the best milling parameters, reduce processing time as much as possible and to keep the process conditions stable throughout the milling process. Especially when large areas are structured, tool wear becomes a significant quality factor. Material breakouts at the cutting-edge of the diamond tool affect the milling result and lead to unwanted marks and scratches. Once a material breakout occurs, the marks will be transferred in every following microlens. An example of a defective diamond tool and the resulting microlens surface is shown in Fig. 4.

However, when the process conditions are controlled accurately and the milling parameters are optimized, high quality optical mold inserts can be produced (Fig. 5). Processing time of the MLA mold insert was about 4.5 h, with about 1 s for each microlens.

For the milling of the mold insert, a diamond tool with exactly 1 mm was used and every microlens was fabricated by a single immersion of the tool with no additional spiral movement. Therefore, the shape of the milling tool exactly represents the resulting microlenses. The depth of cut was set to 9 μm at a feed rate of 2.5 mm/min. The resulting surface quality of the microlenses was Ra = 4 nm at a form deviation < 1 μm.

Injection Compression Molding.

Injection compression molding proofed to be a suitable method to replicate microstructured polymer optics fast and accurately [8]. In comparison to a regular injection molding process, injection compression molding results in a more homogeneous density distribution in the polymer optical part. Therefore, stress-induced birefringence can be reduced. Furthermore, form accuracy of the overall form as well as of the microstructures is higher. Challenges start in the stage of the mold design, whereas runner system, gate design, and compression movements need to be considered [4]. These have significant influences on the filling behavior of the mold cavity and therefore on the resulting part quality. Possible problems that may occur are flow lines, streaks, incomplete filling, or weld lines (Fig. 6).

A main advantage of polymer optics compared to glass lenses is the possibility to include mounting features into the part. However, the mounting features can increase complexity in the molding process since the more complex geometry of the mold cavity becomes more difficult to fill. The MLA produced in this paper has a thickness of 500 μm, combined with mounting features on each side. The small thickness of the part limits the process window during the molding process, since a slow filling process results in partially filled parts (Fig. 7). Therefore, the compression movement and the timing with the melt injection need to be determined and controlled precisely. If the compression movement starts too late, partially solidified material is compressed which leads to inhomogeneous material distribution.

For the molding of the MLA, an injection compression molding tool with pneumatic frame plate was used. At the beginning of the molding process, the frame plate was moved forward to close the cavity; the structured mold insert however was kept in a backward position, with a resulting gap width of 900 μm. With a 0.7 s delay after the polymer injection, the mold insert was moved forward to perform the compression movement and compress the MLA to the required thickness of 500 μm. Even changes in the initial gap width of 100 μm or 0.1 s in the delay time resulted in incomplete filling of the cavity.

The aspect ratio of the microstructures also influences the molding process significantly. For higher aspect ratios, complete filling of the microstructures becomes more challenging (Fig. 8). Increasing mold temperature, melt temperature as well as high compression forces are necessary to produce high quality DOEs. This can be seen in the example of the molded DOE shown in Fig. 9(a). In comparison, the filling of the MLA structures is significantly easier, even when the structures are 4 μm deep compared to the 1.6 μm DOE structures, due to the advantageous aspect ratio, allowing the material to cover the microstructures more easily.

For the injection compression molding, the tool was designed in a way that the compression movement was performed by the ejector unit of the molding machine. This method allows an accurate positioning of the moveable mold insert during the compression process. At the beginning of the molding process, the moveable mold insert is pushed backward by the injected polymer material. With 0.7 s delay time, the compression movement is started, compressing the part to the required part thickness. This movement was necessary to avoid a free jet behavior during the filling of the cavity.

If the process parameters are optimized and controlled precisely, high quality microstructured polymer components can be produced by means of injection compression molding for both cases of microstructures presented. For the molded DOEs diffractive structures with a height of 1.6 μm could be achieved, corresponding to the original design, with the complete microstructures transferred from the glass master substrate into a molded component.

Metrology.

The measurement of microstructured polymer optics is associated with several challenges. Light-based measurement systems like confocal measurement systems and digital microscopes often struggle to detect a focus point on the surface when the component is transparent. However, laser-based measuring systems as well as tactile measurement setups are able to analyze these transparent materials (see Fig. 11).

Characterization of polymer optics is even more challenging, when the components include microstructures. On the one hand, these microstructures often interact with light, since they fulfill functions required for optical components, leading to unwanted measurement artifacts. An example is the measurement of the diffractive microstructures on the molded DOE using WLI, shown in Fig. 10, which leads to artifacts due to diffraction within the measurement signal.

Due to the small size of the microstructures, measuring them exceeds the capability of a lot of measurement systems. Regular tactile measurement systems are limited in the available tool tip radius to about 2 μm. Furthermore the shape of the needle may affect the measurement. While atomic force microscopy is a tactile measurement method suitable to capture these very small structures, the steep walls of the microstructures often exceed its capabilities. Laser-based methods are able to measure the small microstructures, enabling the characterization of molded microstructured polymer optics. A comparison of the two measurements of diffractive structures is shown in Fig. 11. Measurement (1) was performed with a noncontact laser autofocus system (MLP-3, Mitaka Kohki Co. Ltd., Japan) and (2) was measured with a tactile system (Surfcom 5000DS, Carl Zeiss AG, Germany). It can be seen, that the tactile measurement was not able to capture the shape of the microstructures due to the size of the tool tip. The diffractive structures appear to be much more uneven as they are, according to the laser autofocus measurement at the same position.

The characterization of microstructured polymer optics becomes even more challenging, when the components are curved or include mounting features. Thus, increased working distance for the measurement equipment is necessary to avoid unwanted contact. This limits especially the AFM characterization of components. Furthermore, due to the curved shape or mounting features, as well as the shape of the microstructures, limitations of measureable angles become obvious for a multitude of measurement technologies. WLI and confocal measurement systems are usually limited to angles up to 30 deg. In these cases, tactile- and laser-based measurement systems are very well suited to characterize components.

Similar to the structuring technologies, measurement systems have to deal with the same conflict concerning measurement resolution and measurement areas. High-resolution characterization of microfeatures usually results in small measurement areas and long processing time. Thus, overall form accuracy has to be evaluated separately from the analysis of the microstructures. Therefore, a combination of multiple measurement systems is advisable.

Conclusions

In this work, two microstructured polymer optical parts were fabricated using two different structuring technologies. Both parts were replicated using injection compression molding. A DOE was produced using laser direct writing-based mold fabrication, and a MLA was fabricated by means of UP-milling. Three major challenges in the fabrication of microstructured polymer optics are the selection of a suitable structuring technology, the replication method, and the metrology for characterization. LDW proofed to be a suitable technology to fabricate continuous diffractive microstructures even on curved substrates, whereas UP-turning would not be able to produce a sufficient edge radius. In return, UP-milling is a suitable method to fabricate accurate microlenses with optical quality, which could not be produced efficiently and with the required precision by LDW. Injection compression molding was shown to be employable to replicate microstructured polymer optics; however, high process control is mandatory to keep the process within a very tight window. Especially, the constant control of the process parameters and the compression movement are mandatory to achieve high quality parts. Finally, to evaluate the quality of the components, appropriate metrology is needed. Only a combination of several measurement systems allows a full characterization of the components. Especially, laser-based measurement systems enable excessive investigations of microstructured features.

Acknowledgment

We would like to thank our project partners Institute for Technical Optics of the University of Stuttgart and the Institute for Laser Technology in Medicine and Measurement Technique in Ulm.

Funding Data

  • Federal Ministry of Economic Affairs (AiF-RP-No. 18556 N).

  • Ministry of Finance and Economics of Baden-Württemberg (innBW IDAK; Funder ID: 10.13039/501100006360).

References

References
1.
Doushkina
,
V.
,
2010
, “
Advantages of Polymer and Hybrid Glass-Polymer Optics
,”
Photonics Spectra
,
44
(
4
), pp.
54
58
.
2.
Wang
,
C. Y.
, and
Wang
,
P. J.
,
2014
, “
Analysis of Optical Properties in Injection-Molded and Compression-Molded Optical Lenses
,”
Appl. Opt.
,
53
(
11
), p.
2523
.
3.
Moriwaki
,
T.
, and
Shamoto
,
E.
,
1998
, “
Analysis of Thermal Deformation of an Ultraprecision Air Spindle System
,”
CIRP Ann
,
47
(
1
), pp.
315
319
.
4.
Roeder
,
M.
,
Schilling
,
P.
,
Hera
,
D.
,
Guenther
,
T.
, and
Zimmermann
,
A.
,
2018
, “
Influences on the Fabrication of Diffractive Optical Elements by Injection Compression Molding
,”
J. Manuf. Mater. Process
,
2
(
1
), p.
5
.10.3390/jmmp2010005
5.
Yao
,
D.
,
2002
, “
Injection Molding High Aspect Ratio Microfeatures
,”
J. Inject. Molding Technol
,
6
(
1
), p.
11
.
6.
Manske
,
E.
,
Jäger
,
G.
,
Hausotte
,
T.
, and
Füßl
,
R.
,
2012
, “
Recent Developments and Challenges of Nanopositioning and Nanomeasuring Technology
,”
Meas. Sci. Technol.
,
23
(
7
), p.
074001
.
7.
Häfner
,
M.
, et al. .,
2009
, “
Laser Direct Writing of High Resolution Structures on Curved Substrates: Evaluation of the Writing Precision
,” Fringe 2009, Springer, Berlin, pp.
1
4
.
8.
Sortino
,
M.
,
Totis
,
G.
, and
Kuljanic
,
E.
,
2014
, “
Comparison of Injection Molding Technologies for the Production of Micro-Optical Devices
,”
Procedia Eng.
,
69
, pp.
1296
1305
.