March - April 1999
eMail the Editor
Microvia Decision Time is Nearing: A Report on 8 Approaches
By Harvey Miller
Contributing Editor
New winds are sweeping the Interconnection World, driven by density and performance requirements and, by CSPs and BGAs. We've all heard the names—microvias and build-up substrates. And before they are spent, the face of the electronics interconnection infrastructure will be transformed.
They were gathered in the "Microvia Pavillion" at NEPCON West '99: eight companies offering different paths into the HDI substrate future, for the edification of OEM attendees. (Most were rehearsing for a larger performance to a different—mostly PC board industry audience—three weeks later at PC EXPO.)
The Tally:
5 laser drilling companies: Electro Scientific Industries, Excellon, Hitachi Digital Graphics, Lumonics, Pluritec.
2 plasma drilling companies: Dyconex, Plasma Etch
1 exponent of photovias: IBM Micro-electronics.
Below are descriptions of their competitive approaches and their credentials. (After hearing the respective pitches, I was ready to buy one of each offering:)To understand the differences better, consider the critical element common to all microvia technologies: Build-up substrates (BUS) on a conventional PC board core.
These approaches are not simple drop-ins to PC board fabrication. For optimum price/performance, a new "build-up" glass, fabric-free structure, 1 to 4 layers in most cases, must be added above and below the conventional multilayer PC board core.1 (This provides stability, ground-voltage planes and wiring).
These BUS's call forth new materials, new resists, new plating chemistries and new processes—not topics for this discussion, except for reference to microvia formation-friendly materials, such as aramid fiber laminates and resin-coated copper (RCC).
|
| Chick-scale packaging? Newly-hatched chick shows how small the board and electronics are inside IBM's minidrive. (Photo courtesy of IBM) |
IBM Yasu produced Photovias in volume in 19912 for Personal Computer Notebooks. Both Multek Austin and IBM Endicott are producing them in large volumes this year for IBM ThinkPads. Pictured is the IBM one inch, 340 Mbyte microdrive with its supporting electronics on a compact Flash card. Four different company divisions in three locations made it possible: Yamamato, Japan, and San Jose, Calif., for the drive, Burlington Vt., for the chips and, Yasu, Japan, for the Surface Laminar Circuit» (SLC) board and assembly. Generalized SLC fabrication steps include:
- Layers of 1.5 mil photosensitive dielectric, wet or dry film, applied to top and bottom of a conventional 2-sided board.
- Imaging to define photovias.
- Through-hole drilling, cleaning and surface adhesion enhancement.
- Full panel electroless plating, photoprocessed, etched.
IBM's latest SLC licensee is Vertex in Taiwan. For more information, contact Irv Memis at IBM Endicott.
Market Application Note for Photovia Technology
Photovias are especially appropriate for large volumes of small boards. An example is Ibiden's success with cellular phone boards. Another market is high-density, rigid BGA substrates. These substrates have holes all over and the standardized design volumes easily absorb the cost of special tooling.Plasma Etch
Two differing proponents on opposing sides of one alcove managed to split my personality in two as I listened to their convincing arguments for different plasma spins. Common plasma elements:- They are multi-purpose machines, useful for cleaning, Teflon¬ activation and polyimide film through-hole drilling, as well as blind microvia formation.
- They remove resin only, not copper. Thus, windows must be etched or laser-abladed.
- They process entire panel microvias simultaneously from both sides, and multiple chambers may be paralleled using one control.
- The most "plasma-friendly" build-up substrates are RCC and polyimide film.
- They are environmentally friendly, replacing solvents and permanganate.
Dyconex3
The company has been a proponent of isotropic plasma etch since they developed it in 1992 as Dycostrate», and the process and hole profile has been improved. Dyconex was number 1 in preserving low resistivity through temperature cycling in an ITRI microvia test program. This is attributed to plating quality, enhanced by hole profile. Dyconex has turned isotropic necessity into a virtue—thinning down outside copper, micro-etching "bounce pad" copper —all at the same time, then etching away the overhang and eliminating "bottlenecks" in the hole profile. (Thanks, Peter Virsik)Plasma Etch4
Their directional etch development is another approach to eliminating the overhang problem—without the need for another etch step. Plasma Etch's microvia technology grew out of work done with Hadco Phoenix. They are shipping microvia boards to Hewlett-Packard using RCC as the build-up substrate. (Thanks, Tim Apol)The Plasma Upgrade Path
Changing gases to reduce etch time, modifying reactive ion etch parameters, technology partnerships are potential sources of product improvement.Market Application Note for Plasma Microvia Technology
Plasma-formed through-vias are especially appropriate for flexible IC package substrates in BGAs and CSPs. Standard volumes can be very large, with many holes required.Laser Microvias5
Lasers are beginning to dominate microvia formation technology. Reasons that lasers will become the mainstream technology include:- The laser beam simulates an electronic drill bit, but has a critical advantage—as drill bits become tinier, cost and breakage skyrocket and speed plummets. For lasers, smaller=faster and cheaper.
- Laser beam diameters offer choices from 1- to 10-mil holes, with energy choices for one- or two-level blind vias, and the same flexible programmable via placement capability that makes mechanical drilling so attractive—but lasers extend that to 3 dimensions!
- "Laser-friendly" materials, glass-free, also offer superior circuit performance—polyimide impregnated aramid, RCC, for example. (The thinner the copper the better—none on the surface is best.)
- Lasers widen the choice of deposited dielectrics, beyond photo-imageable.6
Several factors have slowed the growth of lasers:
- First, they are not a drop-in replacement for mechanical drilling machines.
- Their optimal use requires the employment of different laminate materials and circuitization (semi- and full-additive copper), changes that require OEM design leadership and/or participation—certainly the kind of changes associated with new design cycles.
- Capital cost is high, typically $600K, twice that of a new six-spindle drilling machine.
- Technological obsolescence since 1992 has compounded the typical new technology adoption problems of training, familiarization and compatibility.
The early laser drilling machine market life cycle phase can be measured by placements.7, 8 At the beginning of 1999, the number was about 450, growing at over 80 percent/year.
That compares to about 200,000 mechanical drilling machine spindles. Although the industry is in its infancy, the number of participants is not unduly large.
They come in two varieties, exemplified by the five laser machine participants in the Microvia Pavillion.
Only two, ESI and Lumonics, may be characterized as laser companies. Both have a history of service to electronic components markets. (For most laser- industry companies, medical, cosmetic and industrial applications provide more inviting, more easily defined targets than the PC board industry, with its composite and variable materials.)
The other three, Excellon, Hitachi Digital Graphics and Pluritec, are major drilling machine companies (now joined by ESI, who bought and incorporated Dynamotion).
All the laser drilling companies recognize and are trying to address PC board manufacturer concerns.
PC Manufacturer Concerns
- They all have aggressive roadmaps for increasing a) laser frequency, b) galvo positioning and c) table speeds, offering upgrade options with trade-in and leasing programs.
- They recognize the need for stacked vias, through layer 2 (ground) to layer 3 (signal).
- They offer laser-drilling services and/or support the establishment of service companies.
Electro Scientific Industries9
The placement of 150 machines by ESI since 1992 makes it #1. Its diode-pumped, frequency-multiplied Nd:YAG (UV) technology trades-off size and speed, power and hole quality—25 micron holes, copper ablation. Organics require low power density, so it's a good compromise. Higher power density switch is used to ablate level 2 copper for stacked vias. RCC is a preferred outer material. Model 5200 is current high offering. (Thanks, Ron Romadka)
Lumonics10
Soon to be GSI Lumonics, which will jointwo laser companies. (GSI also leads in positioning galvos.) Reported to have spent $20 million bringing current hybrid 2-laser GS-600 to market—combines its TEA CO2 (IR) providing short, powerful pulses for ablating dielectrics—with diode-pumped YAG for ablating copper. Dual system shares setup and table for higher throughput. They are allied with Sumitomo, whose TEA machine at JVC is used in high camcorder and digital mobile PC board production for drilling resins. (Thanks, Bill Young and Bob Stobaugh)The other three companies have added laser-drilling machines to their mechanical drilling machine lines. Their powerful CO2 lasers come from Coherent or Synrad.
Excellon
This company is the latest arrival among the current laser generation, via liaison with Exitech, a laser company based in the UK (1984). It is a pioneer and market leader in mechanical drilling (1962). The LVD-2001 was introduced in Wiesbaden in September 1998. Seven systems are in beta. It is a dual with YAG, similar in concept to Lumonics GS-600. (Thank you, Henry Martinez)Hitachi Digital Graphics11
Current series is LCO. Drills through Copper up to 3/8 oz, black oxided. Claims 30,000 holes/minute. (Thanks, Jenny Tran)
Pluritec
LaserVia™ drilling system trades-off drilling through surface copper for two-level holes by increasing CO2 energy, oversizing beam and defocusing. Thermount is preferred material. (Thanks, Larry Burgess)References
- Catalogued in many discussions by Happy Holden of TechSearch. An early, far-sighted example was "PCB Build-up Technology," SMT, August 1996. More recent was R. Mahidhara, "Substrate Requirements for Chip-Scale Packaging," Chip Scale Review, January-February, 1999, p. 25.
- PAC/Asia Circuit News, February 1991, page 6.
- P.Virsik and W. Schmidt, "The Hole Story about Microvias," Proc. NEPCON West'99, February 1999.
- T. Apol, "Directional Plasma Etching," Printed Circuit Fabrication, November 1998.
- L. Burgess, "Blind Microvia Technology by Laser," Proc. NEPCON West'99.
- K. Segawa, "Build-up PWB with Laser-Processed Via Holes VIL," Proc. IMAPS, November 1998.
- A. Biernaux, "A Flexible Production Laser System for Blind Via Drilling," Proc. NEPCON West'99.
- J. Murray, "Alternative Holing Methods," Printed Circuit Fabrication, January 1996.
- R. Ramadka and S. Raman, "Laser and Beam Positioner Technology," Proc. NEPCON West'99.
- J. Morrison, "An Update on TEA CO2 Laser Technology," IPC Conference, November 1997.
- R. Maniwa, "Finer Microvia by Laser and Additive Process," Proc. IMAPS, November 1998