1 of 6
Figure 1: Sumillimeter Mechanical Fabrication Performance.
2 of 6
Figure 2: High aspect- ratio 5 micrometer lines and spaces at 200 micrometer thickness fabricated from nickel using the core iHARM processes. Courtesy: HT MicroAnalytical.
3 of 6
Figure 3: Metal flexure realised in HAR with a nickel alloy with 10 micrometer minimum flexural thickness at a structural height of 200 micrometer. Courtesy: HT MicroAnalytical.
4 of 6
Figure 4: Wafer level sealed iHARM devices. This micrograph shows various views of a 2 mm x 2 mm inertial switch die with hermetic electrical vias. Courtesy: HT MicroAnalytical.
5 of 6
Todd Christenson is Chairman, CTO and Co-Founder of HT MicroAnalytical, Inc., a New Mexico company incorporated in 2003 to design and manufacture metal MEMS. Previously he was on staff at Sandia National Laboratories for eight years working on micro fabricated electro and opto mechanical devices. He received a Ph. D. degree in Electrical Engineering from the University of Wisconsin in 1995 where he worked on semiconductor devices and integrated micro sensors at the Wisconsin Center for Applied Microelectronics. His R&D focus has been on the design and micro fabrication of integrated physical sensors and micro actuators. He has authored 125 publications and is inventor of 45 patents.
6 of 6
Roger Grace is President of Roger Grace Associates of Naples Florida, a marketing consulting firm that he founded in 1982, specialising in the commercialisation of MEMS. His firm provides business development, custom market research, market strategy and integrated marketing communications services to high tech clients worldwide. He has published over 20 articles in industry publications, organised and chaired over 50 MEMS technical sessions and conferences and is frequently quoted in the technical and business press as a MEMS industry guru. He was a visiting lecturer in the School of Engineering at the University of California Berkeley from 1990 to 2003. He holds BSEE and MSEE (as a Raytheon Company Fellow) degrees from Northeastern University where he was awarded the “Engineering Alumni Engineer of the Year Award” in 2004.
Todd Christenson, Ph.D., President / CTO, HT MicroAnalytical Inc., Roger H. Grace, President, Roger Grace Associates
As opposed to bulk substrate etching such as in the modern DRAM cell which uses trench etching to form greater electrical capacitance per unit chip area, and trench etched silicon micro mechanical devices such as those used for present day integrated silicon oscillators, batch wafer processes capable of constructing features well above and beyond the surface of the chip have not been available to the precision micro fabricator. And just as restrictive from a micro mechanical engineering viewpoint, wafer scale processes have largely been reserved for semiconductor- based materials that are not suited for a variety of important electromechanical applications. As a last impediment to realising commercially viable micro manufactured 3D mechanical componentry, processes capable of forming these devices in high volume with commensurate economy of scale have similarly been out of reach until recently. In order to open the path to greater capability for 3D integrated micro mechanics, a new combination of processes have been integrated together to surmount the aforementioned shortfalls. Termed Integrated High Aspect Ratio Metal (iHARM), this group of processes provides a batch fabrication approach for several hundred micron tall structures with metal materials packaged at the wafer level all while substantially maintaining tolerances at the sub millimeter scale. In addition to the expansion of fabrication tools available to the micro mechanical engineer, this process provides for a number of initially unforeseen benefits that support the integration of new materials particularly well suited to meet design and fabrication challenges at the micro scale.
What is the problem?
There exists a dimensional fabrication realm, which is not addressed well by serial-based machining methods and yet is also not readily accessible with batch micro fabrication techniques. In terms of dimensions, this region of mechanical fabrication lies between micrometers and millimeters. In terms of processes it lies between what is commonly practiced using micro lithography and the operations one finds in a miniature precision machine shop or ‘miniature shop’. In terms of industries this manufacturing void spans the gap between integrated circuits and mechanical watchmaking.
Challenges presented by this dimensional scale can be illustrated by considering the fabrication of a 300 micrometer cube. Two categories of micro machining approaches are apparent: these categories include subtractive machining in which one begins with a raw material and removes what is not needed and additive machining whereby the part is built up by a patterned deposition process.
Yet another distinction within each of these machining approaches is whether the process takes place serially, one part after the other, or whether a number of parts are fabricated simultaneously in batch.
Creating the 300 μm cube from a serial subtractive process such as machine cutting can provide the ultimate in material choice while also requiring lengthy machine times, high tooling costs with limited throughput and typical tolerances at this dimensional scale near 30 parts per thousandth (30 ppt). Single point diamond turning can improve tolerances to a few parts per thousandth although with a reduced material base and an even further reduction in throughput. Batch subtractive processing such as that obtained with photochemical etching provides tremendous throughput although with a restricted set of materials which can be etched and dimensional fidelity and minimum tolerances at best for this dimensional regime near 50 to 100 ppt.
Several serial additive processes can now approach the 300 μm dimensional regime albeit with substantial restrictions in throughput, materials, and/or tolerances. Becoming ever more popular, fused deposition modeling (FDM), however, cannot reasonably address the 300 μm cube and typical rapid prototyping stereo lithography has minimum dimensions of at best 50 μm. Micro stereo lithography (SLA) has demonstrated resolution of a few micrometers with specific photopolymers yet still resides in a modeling or prototyping category. Selective laser sintering (SLS) processes provide exciting new material categories within the 3D printing category although again with tolerances greater than 100ppt at this scale and not able to feasibly address a 300 μm cube. However, multi- layer thin film electro deposition provides tolerances of 5-10 μm (~20ppt) and provides expanded metal material choices. Throughput in this category is dependent upon the number of tooling centers although it can produce many parts simultaneously. The accompanying micro fabrication process comparison table (figure 1) further highlights the tradeoffs important to designers of complex 3D micro miniature componentry.
To demonstrate where even more significant fabrication problems could occur, the 300 μm cube is now scaled to a part with lateral dimensions of 20 μm. The part is now a 20 μm wide by 300 μm tall wall or a 20 μm x 20 μm by 300 μm tall pillar. A review of the above processes will show that for this dimensional part scale there are really no good alternatives particularly in high volume and high rate production.
Why iHARM Fabrication?
The aforementioned part containing hundreds of micrometers of feature height with relatively small lateral features of tens of micrometers and below has been termed to have ‘high aspect-ratio’ (HAR). From a design standpoint, a more apt measure of HAR is ‘run-out’ or the amount of lateral variation achievable per unit height. Numerous micro scale devices have a need for accurate HAR features, which is an innate capability of iHARM, such as mechanical flexures, fluidic channels, and microwave waveguides. Additionally, iHARM provides for the implementation of this HAR feature capacity with an extensive metal material base while providing simultaneous fabrication of a multitude of parts using batch wafer fabrication as in IC fabrication thus providing enhanced item-to-item repeatability, high throughput and low unit cost.
How does iHARM fabrication work?
Belonging to the category of additive micro fabrication, the iHARM process follows at its core a “LIGA” based approach whereby photolithography (LI) is used to create a mould and electrodeposition (G) is a primary means of filling. A moulding replication step (A) is not used, but the resulting filled mould is then planarised. To realise iHARM, this basic process is variously augmented with patterned sacrificial layers, post processed coatings, and wafer scale packaging (WSP).
The first step in the process, photolithography, can be accomplished using UV light from a conventional photomask aligner or a direct write laser can also be used to expose and pattern the mould material. The requirements for the mould material is that it have the photochemical resolution needed which for a number of UV photoresists is of the order of 10 nanometers, be resilient to the mould filling step, be able to withstand planarisation operations, and ultimately be removable in the context of batch wafer processing. To achieve high aspect ratio patterns, the photoresist also needs to have enough absorption length at the exposure wavelength to accommodate the structural height needed.
Once the mould is created the material used to fill it is most often electrolytically deposited and as such a conductive layer is needed as a seed layer to effect a through mould deposition process.
Other mould filling processes that are possible include slurry casting, thermal spray, hot embossing and hot forging. The latter methods often require an intermediate metal mould, which can provide for a number of high temperature forming processes. Given the non-uniform nature of these filling processes, a planarisation step is necessary to create a flat surface which is a crucially important procedure required to maintain vertical dimensional tolerances, support the formation of multiple layers, and provide a means to integrate packaging on the wafer scale. This process is often carried out with diamond lapping or ‘nano-grinding’, a procedure which can maintain the edge definition of metal features embedded in plastic without the use of active chemical additives. The final core process step is the removal of the mould material typically by liquid chemical dissolution or plasma etching.
To demonstrate the capability of a high aspect ratio process, figure 2 reveals what can be done at the limits of high aspect-ratio micro fabrication wherein highly collimated x-rays generated via synchrotron radiation from a charged particle storage ring are used to create a mould for electrodeposition. The pattern shown in figure 2 is used as part of a ‘drop-in’ test die used on all fabrication runs to monitor not only classic line and space critical dimensions as in a planar process but also to measure run-out. Measuring run-out above 5:1 aspect ratio with submillimeter dimensions is challenging and as a result several on-chip processing aids have been developed to monitor this feature including micro scale versions of a ‘go’ – ‘no-go’ gauge. Additional layering may be obtained prior to the removal of the mould with direct electrodeposition, or after the mould is removed via metal based wafer level bonding processes. Combined with patterned sacrificial layers, a partially released, yet substrate captive integrated mechanism can be created.
To date, most parts have simply been ‘released’ or are removed from the supporting fabrication substrate and used individually. The full utility of iHARM comes from additionally being able to maintain the micro fabricated mechanisms on the substrate and subsequently package them in place at wafer scale.
What advantages does iHARM provide?
The basic iHARM process produces components that are essentially prismatic, that is, they are projections of a 2D shape. Although additionally, as long as Snell’s law is heeded, slanted single and multiple angled UV exposures can be effected and rounded features can be obtained through post-lithographic thermal contouring of the photopolymer. The high aspect ratio feature of the process provides for immense design freedom for the designer of flexures at the sub-millimeter scale.
For in-plane motions, out-of-plane flexural stiffness is needed which is well supported by high aspect ratio (HAR) micro fabrication. Additionally, nearly any 2D flexure shape can be rendered for in-plane flexural motion which allows the designer to distribute flexural stresses at will and even effect largely stress uniform flexural systems. As an example of what can be done, figure 3 contains a micrograph of a HAR flexure, which is as-fabricated in a flexed position. This component, which consists of a Ni-Fe soft ferromagnetic material, is used within a miniature precision magnetic micro actuator.
Lateral tolerances for iHARM components are typically at one micrometer with repeatability of tenths of micrometers. The lateral dimensional repeatability is particularly important for flexure thicknesses since their stiffness depends times on their thickness squared or often as in a cantilever as their thickness cubed. For example, achieving a respectable 20 ppt repeatability for a 10 micrometer thick cantilever flexure still results in a 6% variation in stiffness which, although not nearly as good as the macroscopic situation, can be an acceptable level in many designs.
Yet another difficulty for flexure design at the sub-millimeter scale is that of material maximum stress. If a cantilever flexure is scaled down equally in all dimensions, in order to obtain the same scaled restoring force for the corresponding scaled deflection requires that the material must support a stress inversely proportional to the reduced scale. This is a result of the spring constant (or spring rate) decreasing linearly with reduction in scale and as a consequence having available the highest strength material is desired. Maintaining high aspect ratio can lessen this effect. With electrodeposition one can also take advantage of high strength nanocrystalline alloys that in addition possess tremendously high fatigue resistance[1]. Such sub 100 nm grain size alloys are readily formed with electro deposition technique and additionally possess high yield strengths of well over 1.5GPa as well as high hardnesses of Rockwell C54, which is in the vicinity of tool steel.
The material base extends yet further as a consequence of the additive approach including soft ferromagnetic metals, bonded rare-earth permanent magnet composites, slurry cast ceramics and moulded thermoset plastics, to name a few. Once the high aspect ratio micro structures have been fabricated and post processed, the fact that they exist with a top layer that is now flat and parallel relative to their underlying substrate makes them directly amenable to packaging. For this step, a top metal layer may be bonded and sealed directly to a seal ring provided within the HAR layer itself. Alternatively, glasses and ceramics which are thermally matched to the substrate material can be provided with a metal bond ring used to seal the HAR micro structures with metal bonding processes such as thermosonic bonding, diffusion bonding, and transient liquid phase bonding [2]. Figure 4 shows such an example of an iHARM device that has been packaged using an alumina substrate layer containing electrical via feedthroughs together with a top cap alumina layer and a HAR layer in between. The result is a WSP millimeter scale intertial switch that was developed jointly with the US Army.
Another quality of iHARM is its ability to support short product design cycles and be used as a prototyping tool while at the same time supporting high rate and high volume production levels. Often, thousands of prototypes are created from design concept to test bench within two weeks. Scaling to high rate production can then occur directly much like an IC production run but with parallel processes that further reduces overall fabrication time.
Summary
The adoption of the iHARM process creates a micro fabrication capability well above and beyond its traditional planar wafer surface resulting in enabling a vast array of opportunities that employ mechanisms at the sub-millimeter scale. Within this newly available dimensional regime, micro sensors, micro actuators and micro structures benefit. A mature material base incorporating unique high strength and fatigue resistant metals supports a number of desired micro mechanical functions which when combined with batch wafer scale packaging provide a particularly effective approach to producing innovative micro devices in high volume and at low unit cost. iHARM provides designers and manufacturers of complex 3-dimensional precision micro parts with a viable alternative to their current serial mode approaches.
References
[1] Boyce, Brad L. and Padilla, Henry A, “Anomolous Fatigue Behavior and Fatigue-Induced Grain Growth in Nanocrystalline Nickel Alloys,” Metallurgical and Materials Transactions A, Vol. 42A, July 2011, pp. 1793-1804.
[2] Welch, Warren C., Chae, Junseok, and Najafi, Khalil, “Transfer of Metal MEMS Packages Using a Wafer-Level Solder Transfer Technique,” IEEE Transactions on Advanced Packaging, Vol. 28, No. 4, November 2005, pp. 643-649.