1. Introduction
The technology for fabricating micro-optical elements for low cost optical interconnects by micro-jet (ink-jet) printing has been under development for over two years [1,2]. This data-driven method of micro-optics fabrication offers the benefits of low cost, flexibility and in-situ, non-contact processing. These features can be used to advantage in applications where increasing the efficiency of optical power coupling as a value-added step is a goal, and they provide unique capabilities for rapid prototyping and customization of microlens arrays. Here we present our latest results in developing this Optics-Jet technology for applications such as collimation of the outputs of LEDs and diode lasers, as well as for increasing the efficiency of focusing of GRIN lenses [3]. Data on printed microlens optical characteristics will be shown, as well as their performance in collimation and astigmatism reduction of optical sources.
 

Figure 1.  Generation of 50µm droplets at 2 kHz.

2. Microlens Printing Method
In "drop-on-demand" microjet printing a single droplet is ejected from the print head every time its piezoelectric actuator is pulsed, producing droplets with precise volumetric control at rates up to 5 kHz, as illustrated in the photograph of Figure 1. To fabricate a microlens by this method, a print head with 30-60 µm orifice diameter is used to deposit at 165C similarly sized droplets of UV-curing optical material onto the target site which is typically held at 40C. An example of microlenses formed in this way is the array pictured in Figure 2 of 220 µm diameter lenslets printed directly onto one end of a 5 mm diameter GRIN lens for reduction of optical insertion loss. For a given optical material, device orifice size and substrate, the diameter and focal length of a printed lenslet are logarithmic functions of the number of deposited droplets, as seen in the data of Figure 3. The speed of a printed microlens for a given mass of deposited material is a function of the degree of spreading which occurs on the substrate prior to solidification which, in turn, is controlled by the viscosity level of the material and the degree to which it wets the substrate surface. For polymeric optical materials deposited onto a silanized glass substrate, microlens speed remains relatively constant over a wide range of diameters.
 

Figure 2.  10x10 array of 220µm diameter micro-lenses printed onto 5mm GRIN lens.

Figure 3.  Variation of printed hemispherical microlens geometry with number of deposited droplets (open file for readable resolution).

3. Printed Microlens Optical Characteristics and Reproducibility
The optical imaging quality of printed hemispherical microlenses was assessed by measuring the Modulation Transfer Function (MTF). The measurement system, which utilizes a virtual point source of 665 nm to illuminate the convex surface of the lenslet under test and a 100x microscope objective to project the focused light onto a CCD camera, is capable of measuring microlenses with diameters in the 100 µm - 50 mm range. Measured MTF data for a printed 110 µm diameter microlens with speed of f/1.25, given in Figure 4 along with the those for an ideal (diffraction-limited) lenslet of the same speed, show an identical cutoff spatial frequency (1,200 1p/mm). The Strehl Ratio, obtained from the ratio of integrated areas under the measured and theoretical curves, is 0.71, indicating that the printed microlens produces relatively little spherical aberration in this measurement configuration. To determine the reproducibility of printing a microlens, a microscope system is used to measure diameters and focal lengths of individual microlenses within variously configured arrays. Focal length data for a 10 x 10 array of 495 µm diameter microlenses printed onto low-wet-coated glass on 750 µm centers are given in Figure 5. The standard deviations from the average values of focal length and diameter within such arrays are on the order of 1.5% and 1%, respectively, which are within the measurement errors and sufficient for many array optical interconnect applications. Microlens placement accuracy depends primarily on the accuracy of the substrate stages and the distance between print head orifice and substrate surface. At a typical printing distance of 1 mm, our R&D system can place microlenses with an accuracy 2 µm.
 

Figure 4.  Modulation transfer function for an f/1.25 microlens printed with one 50µm drop of UV-curing optical material (solid curve), compared to ideal, diffraction-limited case (dashed curve) at same f/# (open file for readable resolution).

Figure 5. Distribution of focal lengths of 100 each  0.5 µm diameter microlenses printed in 10x10 array on 750 µm centers, showing a typical standard deviation from the average value of 1.5% (open file for readable resolution).

Figure 6.  Geometry for LED array collimation by printed microlenses (open file for readable resolution). 

Figure 7.  Array of 355 µm diameter micro-lenses printed on 375 µm centers, shown in substrate plane (top) and in profile (bottom).

Figure 9.  Variation with droplet spacing of focal lengths and dimensions of hemi-elliptical microlenses of Figure 8 (open file for readable resolution).

References
[1] W.R. Cox, T. Chen, D.W. Ussery, D.J. Hayes, and, R.F. Hoenigman, "Microjet printing of anamorphic microlens arrays," SPIE Proceedings, Vol. 2687, pp.89-98, 1996.

[2] W.R. Cox, D.J. Hayes, T. Chen, H-J Trost, M.E. Grove, R.F. Hoenigman and D.L. MacFarlane, "Low cost optical interconnects by microjet printing," IMAPS International Journal of Microcircuits & Electronic Packaging, Vol. 20, No. 2, pp.89-95, 1997.

[3] V. Baukens, A. Goulet, H. Thienpont, I. Veretennicoff, W.R. Cox and C. Guan, "GRIN-lens based optical interconnection systems for planes of micro-emitters and detectors: microlens arrays improve transmission efficiency," OSA Diffractive Optics and Micro-Optics Topical Meeting, Kailua-Kona, Hawaii, June, 1998.