A Glance into LIGA's Prospering Future

by Dr. Joachim Schulz

Topics Covered

Abstract
Introduction
An Attempt to Describe the Unique Features of LIGA
LIGA Over the Years in View of Market Needs
     LIGA and Molding
     Access to Synchrotrons
     LIGA Standards
     LIGA Masks
Microworks' Approach
Conclusions

Abstract

Over the past decades, many applications have been proposed as commercially viable for LIGA (Lithographie, Galvanoformung, Abformung (Lithography, Electroplating, and Molding)). However, most of them were pushed into the market via silicon MEMS or even classic microdrilling and micromilling. To date, not a single commercial x-ray LIGA effort has prospered and support for academic LIGA centers is fading. However, the authors believe that the current demand for LIGA applications, including x-ray optical elements such as LIGA-lenses, precision apertures etc., is the advent of a second dawn for LIGA.

This paper will give a more strategic than technical overview of the past and future of LIGA.

Introduction

In the 1990's, LIGA was a pacemaker technology for MEMS, used for many smart systems such as acceleration sensors or optical devices integrated with electronic read-out circuits. This first LIGA hype resulted in many different showpieces that proved in an impressive way that the technology is very versatile. Even micropumps were made using LIGA, though this kind of application does not really utilize any of the unique features of LIGA. When judged from their dates of academic publication, several of these LIGA systems preceded similar systems from competing technologies. Yet LIGA's breakthrough on the market still hasn't happened. Why not?

The most apparent reasons are that other technologies such as silicon micromechanics were pushed by big companies, and that classic technologies such as EDM, milling, drilling and the like were pushed by many small and medium-sized companies to achieve ever more challenging fine features. So at a first glance it looks like not enough engineering effort was put into the advancement of LIGA's capabilities.

Public institutions might also share the blame: None of the LIGA centers was ever forced to earn its money on the free market. Yet another reason might be that LIGA is simply too complicated, time consuming and expensive. And finally, while other technologies have advanced quite a bit and have reached sufficient complexity to accommodate more functionality in micropatterned structures, LIGA is still severely restricted to essentially two dimensions on a substrate with limited options for several mask levels and only a handful of materials.

All of the above arguments bear some truth. Nevertheless, microworks is convinced that LIGA's technological advancements in recent years and the needs of the market are merging so much that there will be niches for LIGA. One longstanding niche is microgears. Another one is currently being stimulated by demand from synchrotrons for x-ray optical elements. Other examples could be listed.

LIGA structure in nickel, 1600µm high, aspect ratio of 160.

LIGA structure in SU8 used to focus x-rays. The narrowest parts are 2 µm wide and extend several hundred µm down.

LIGA structure in gold as an absorption grating for x-rays. The aspect ratio of these structures is near 50

An Attempt to Describe the Unique Features of LIGA

The four strong points that were used to promote LIGA in the 1990's are quickly listed:

1. High-aspect ratio
2. High sidewall quality / low roughness
3. Extreme precision
4. Complete 2-D freedom

In the beginning, these four points were sufficient to convince relevant players in technology screening, from governments to larger companies like IBM and Bosch, to look into LIGA. The drive for more miniaturization was apparent; in silicon, however. only bulk micromachining with its severe restrictions was available.

And nearly as quickly, one could list how these strong points tumbled over the years. High-aspect ratios of, say, 20 at completely free 2D geometry can now be easily fabricated by DRIE in silicon; this technology can be tuned to high sidewall quality as well, and, if need be, cryoetching can be applied.

Extreme precision has always been claimed for LIGA, but a methodology to prove it has only recently been developed. The results so far are not conclusive.

Then there is UV-LIGA, which exploits the potential of SU8 resists to fabricate nearly anything that fits the following requirements: +/-2 µm precision, less than 800 µm in height, and made of nickel or nickel-phosphorous as a material. This technology was excellently placed on the market by mimotec in Switzerland, apparently leaving little room for x-ray LIGA. On the other hand, it's success clearly shows that LIGA is a competitive technology. One positive aspect for x-ray proponents: exposure to x-rays makes the process chain more expensive, but only by 30 to 50%.

There is still another niche for LIGA technology in our opinion: the parallel mass fabrication of tiny structures with high-aspect ratios. These are typically arrayed structures which work together over reasonably large lengths or areas. A good example is the x-ray lenses which were recently developed at KIT (central picture above), where the focusing range of each lens is very small. However, when hundreds of them are placed next to each other, the effect can be well exploited.

What is needed to exploit the niche for potential markets? We will try to answer this in the next section.

LIGA Over the Years in View of Market Needs

An excellent summary of LIGA has been given in a fairly recent book, LIGA and its Applications (Saile et al. 2009)¹. The book clearly identifies LIGA as a powerful tool for making excellent microstructures. In this section, we will try to evaluate these findings from the point of view of a small company, microworks. The main criteria are: the time involved before the first parts are produced; proof of the value proposition through standards; as well as cost factors.

Many aspects of LIGA technology have been addressed over the past decade, predominantly as part of government-funded R&D projects. Their goals have typically been to identify a new use, adapt the LIGA process chain and make first parts. Few projects have focused on addressing technological aspects of the LIGA process. In this respect, microworks is thankful for the German government's support of both FELIG² and INNOLIGA³, whereby FELIG focused on automating the process and INNOLIGA on ensuring the reliability of a resist based on the same basic chemistry as SU8.

LIGA and Molding

In the very first publication on LIGA, Becker et al.⁴ suggested molding as the way to go for mass fabrication. From today's point of view, the technology used to make mold inserts is still unreliable, which makes it hard to guarantee delivery dates to customers. More important, however, is that the outstanding features which keep LIGA attractive are essentially impossible to retain in molding: high-aspect ratios, many features on the substrate and high precision. All of this is particularly true for injection molding, while hot embossing does not produce the expected drastic cost advantage over direct LIGA. On the other hand, the process to use molded structures as a form to electroplate metal has never been successfully demonstrated.

There is another indication that LIGA molding tools are not commercially viable: mimotec SA in Switzerland started out in the late 1990s primarily making mold inserts. Today, these make up less than 10% of their business, whereas 90% is the direct manufacturing of LIGA metal parts.

Access to Synchrotrons

A main bottleneck for x-ray LIGA has virtually disappeared over the years: the availability of synchrotron radiation. Increased availability was a major step towards being able to reliably meet delivery dates.

Some synchrotrons now offer automated sample handling and measurements in protein crystallography at a high degree of standardization, which should be possible for LIGA exposures as well. Since the LIGA technique involves more than just exposures, it unfortunately is still not that easy to transfer a process from one synchrotron to another. The second source issue has therefore still not been completely solved, but sufficient arguments are at hand that this will not pose a problem with respect to potential large orders.

LIGA Standards

Ron Lawes and others have argued at several HARMST conferences over many years that the various LIGA groups need to agree on one design, one process and one set of properties to be measured, possibly by using different methods, making structures and comparing the results of all the groups. Unfortunately, this has never occurred. Perhaps the typical researcher saw no R&D reward in doing the same thing over and over; also, all of the LIGA groups have been in competition with each other.

As a result, if asked today what specifications LIGA can actually achieve, the community has no answer regarding precision, maximum aspect ratios, sidewall roughness, verticality etc. What we do have, however, is a situation in which all of the major effects that produce non-ideal results have been identified, and these effects can usually be minimized by typical engineering approaches.

LIGA Masks

In our opinion, here lie the most severe deficiencies of the last decades: While other batch technologies have moved from 4-inch to 6- and 8- or even 12-inch wafers, LIGA still works with 20x60 mm² titanium membranes (KIT), 30x30 mm² silicon nitride membranes (CAMD and HTmicro), and 83 mm round graphite membranes (CAMD and BESSY). Also, several failed projects attempted to use beryllium for its obvious advantages in x-ray applications⁵.

The main reason for larger areas has been the cost factor for DRIE, UV LIGA and the like. This point never had to be addressed by any of the LIGA centers. However, we argue that the main motivations for developing large-area LIGA will be its compatibility with other technologies and the option to make large devices.

What has been achieved is that mask-making has become more realiable in combination with better access to beam time, enabling significantly shorter times for the development of first products. However, two iterations of a new design are still necessary in most cases, and that still takes more than six months.

What needs to be achieved for any future perspective for LIGA is an increase of the mask area with no or virtually no compromise to structural quality.

Microworks' Approach

From the beginning, our mission has been to become a leading LIGA manufacturer. It has never been a strong "business case", because otherwise we would not have been the first such endeavor in Europe. We have looked for other manufacturing opportunities as well, and within a year or so, several projects were acquired. A portfolio analysis may be rather uncommon for a small company like microworks, but it helps us to keep our focus. The picture looked very diverse, in fact way too diverse in 2008. A restructuring of our projects, last but not least enforced by the financial crisis, makes microworks now look quite different.

Figure 2. Portfolio of microworks' projects at the end of 2008. The ellipse size roughly gives the total project value, and the color indicates the technological risks associated with the technological requirements (from low to high risk: dark grey, light grey, green, yellow, orange).

Figure 3. Same as fig. 2 but in mid-2010. The differences to 18 months earlier can be summarized as follows: we agreed to a somewhat higher degree of technological risks on average but cancelled high-risk projects with low benefits, added one key R&D project (in cooperation with KIT) and kept cash cows.

The difference becomes more obvious when we look at another portfolio plot following Wheelright and Clark⁶. We shall skip the cloud picture of the status at the end of 2008 and only show how things look like today:

Figure 4. microworks repositioned itself with respect to the diversity of its projects. A central R&D project, with the support of the State of Baden-Württemberg, is set out to increase mask sizes dramatically. All other projects fit into current platforms or include only minor changes.

The major R&D project focuses on mask-making, with an emphasis on searching for a way to obtain large-area masks. Unfortunately, we cannot report on these aspects today since a patent application is still in progress. But we can say that if we are successful, we will be able to advance to very large areas, say beyond 10 inches, and still keep the option of patterning the masks in an e-beam.

Conclusions

Over the years, several COMS and HARMST conferences have seen excellent contributions from LIGA experts on the advancements of this technology. A commercially successful company has not yet appeared. We have argued in this contribution that the size constraints of high precision, e-beam-written x-ray masks have been a limiting factor not only with respect to costs but also with respect to device size and compatibility with other technologies. Solving these size constraints is the key issue for the future of LIGA technology. We have also argued that the strategic positioning of microworks' projects together with a potential solution to the mask problem may lead to commercially viable x-ray LIGA.

References

1. LIGA and its Applications; V. Saile, U. Wallrabe, O. Tabata, J. G. Korvink (eds.); Wiley, 2009, ISBN: 978-3-527-31698-4
2. FELIG: Modulare Fertigungsstraße für Mikrobauteile über Röntgentiefenlithographie und Galvanik; Project funded by the German Ministry of Education and Research, 2005 to 2009; http://www.imt.kit.edu/downloads/FELIG-Broschuere.pdf (in German)
3. INNOLIGA: Innovative Resist- und Direkt-LIGA-Technologieentwicklung für eine rentable Fertigungsmethode für LIGA-Präzisionsteile für KMUs mit einem stabil arbeitenden Röntgen-Negativresist; Project funded by the German Ministry of Education and Research, 2007 to 2010 under contract number 16SV3522
4. E. W. Becker, W. Ehrfeld, P. Hagmann, A. Maner, and D. Muenchmeyer, "Fabrication of Microstructures with High-Aspect Ratios and Great Structural Heights by Synchrotron Radiation Lithography, Galvanoforming, and Plastic Moulding ( LIGA Process)," Microelectronic Engineering, vol. 4, pp. 35ff, 1986.
5. See the article by Jost Göttert in: LIGA and its Applications; Volker Saile, Ulrike Wallrabe, Osamu Tabata, Jan G. Korvink (eds.); Wiley, 2009, ISBN: 978-3-527-31698-4
6. "Creating project plans to focus product development", S. Wheelright, K. Clark, Harvard Business Review, March-April 1992, pp. 2-14

Copyright AZoNano.com, MANCEF.org