Laser Drilling in HDI: Why Microvias Are the Hard Part

Mechanical drilling stops being practical at a certain via size. A carbide bit below a tenth of a millimetre becomes fragile, wanders in the hole, and cannot be replaced in time without slowing production to a crawl. Below that threshold, holes are formed with a laser instead. Laser drilling of microvias is where that limit becomes a production problem rather than a design exercise.

Laser drilling is the enabling process for high-density interconnect boards, and it is what allows conductors to be connected between an outer layer and the layer immediately beneath without a through hole passing through the entire stack. Removing the through hole frees routing area on every layer it would have crossed, which is the source of the density advantage HDI provides.

The process looks simple: aim a beam at a spot, remove the dielectric, stop on the copper underneath. In production it depends on a tight marriage of material behaviour, laser parameters and process control, and most of the difficulty lies in keeping all three aligned across millions of holes.

How the Hole Is Actually Formed

The dielectric is removed by absorbing laser energy, and the copper beneath acts as a stop layer by reflecting most of that energy. That difference in absorption is what makes the process controllable: the beam removes resin quickly and then loses its ability to continue once it reaches copper.

In practice the stopping behaviour is not absolute. A laser pulse that is too energetic will damage the copper surface, creating a pit that later affects adhesion or contributes to a weak connection. A pulse that is too weak leaves a residue of dielectric at the bottom of the hole, which produces a high-resistance connection that passes continuity test and fails later under load.

Controlling this window requires matching the laser parameters to the specific dielectric. A material change alters the window, which is why substituting a build-up film or a resin-coated copper is a process change rather than a purchasing decision, and why it demands re-qualification rather than notification. This is a core part of any serious quality system for HDI production.

The Parameters That Matter

Energy per pulse determines how much material each shot removes. Pulse width and shape determine how the energy is deposited in time, which affects how much heat spreads into the surrounding material. Pulse repetition rate and the number of shots determine throughput and the cleanliness of the finished hole. Beam positioning accuracy determines where the hole lands relative to the pad beneath.

These parameters interact. Increasing throughput by raising the repetition rate can leave residue; reducing energy to protect the copper can slow the process or require more shots. The optimum is a balance, and it differs between materials, hole sizes and panel constructions.

Beam delivery is the other half of the problem. The laser must be positioned with accuracy comparable to the via diameter itself, which for a hole of a hundred micrometres means the positioning error must be a fraction of that. The system achieves this with a combination of stage movement and beam deflection, and both introduce error that must be calibrated and monitored.

One practical consequence of the laser bottleneck is that panel design influences cost more than designers expect. Placing vias in dense clusters forces the laser to dwell in one area, which affects thermal management of the panel and can slow the process. Distributing vias more evenly across the surface allows the system to run closer to its rated throughput and reduces local heating that can distort thin material.

Finally, the laser drilling step interacts with the rest of the build in ways that reward integrated process thinking. A change in lamination pressure alters the dielectric thickness and therefore the drilling depth; a change in copper surface treatment alters the stopping behaviour. Factories that treat the process as one system, rather than as a set of independent steps, resolve such interactions faster and hold yield more steadily over time.

Registration Determines Yield

A microvia is useful only if it lands on the target pad. Since the pad is barely larger than the hole, the tolerance is small, and it must be held after the panel has been laminated, imaged and processed several times. Registration error accumulates through those steps.

This is why the laser drilling step cannot be considered in isolation. The accuracy of the drill is one contributor; the dimensional stability of the material, the precision of the exposure tooling and the fiducial strategy used to align each layer are others. A factory with an accurate laser but a loose lamination process will still miss targets.

Production control therefore relies on measuring registration on real panels rather than on test coupons. The useful measure is the distribution of landing positions across the panel and across time, not a single best-case value. When that distribution shifts, the cause is usually upstream of the laser, and only a factory that measures the whole chain can identify it quickly.

Plating the Hole After It Is Formed

A drilled microvia is an empty hole with a thin copper base. Making it conductive requires plating the walls and the bottom in one continuous layer, which is more difficult in a small, blind hole than in a through hole because the chemistry must reach the bottom and the current distribution must be uniform.

Incomplete plating produces a connection that is electrically continuous but thermally and mechanically weak. Under temperature cycling the thin section concentrates stress and can crack, producing an intermittent fault that is difficult to diagnose. Because the defect is internal, detection depends on cross-section sampling and on reliability testing rather than on inspection.

Where the via will be stacked, plating is followed by filling and planarisation so the next layer can be built on a flat surface. Voids in the fill are the most common cause of stacked via failure, and they are not visible from outside. Process control on the fill step, supported by periodic destructive sampling, is the only reliable way to know that the process is behaving. A manufacturer’s ability to describe that control is a meaningful capability indicator.

Cleaning between drilling and plating deserves attention because it is easy to underestimate. Residue from the ablation process, dust from handling and contamination from the environment all interfere with plating adhesion inside a small hole. A factory that treats cleaning as a controlled step, with defined chemistry and verified results, produces more consistent vias than one that treats it as a preparatory formality.

Throughput and Factory Economics

Laser drilling is slow compared with mechanical drilling on a per-hole basis, and a dense HDI panel can contain hundreds of thousands of vias. Factory economics are therefore dominated by laser time rather than by lamination, which affects how capacity is planned and how lead times behave.

That also means the choice of via structure has a direct cost consequence. A design that uses staggered vias requires more holes for the same connection, while one that stacks vias uses fewer but demands more process control. The trade between hole count and process difficulty is decided in layout and priced in production.

Designers can help by avoiding unnecessary vias. Consolidating grounds, planning the escape pattern deliberately and using the layer stack economically all reduce hole count without affecting function. On a high-volume board those reductions translate directly into capacity released from the bottleneck process.

Materials and the Process Window

Build-up dielectrics differ in how they absorb laser energy and how cleanly they ablate. A material that produces a clean hole at one parameter set may leave residue at another, and a factory running multiple materials must maintain separate recipes for each.

This has a supply chain consequence. Qualifying an alternative build-up film is not only a matter of verifying its electrical properties; it also requires establishing a laser recipe and confirming that the resulting vias meet the reliability requirements. That work takes time, which is why material substitution in HDI production is a planned activity rather than an emergency measure.

Given how much of the industry depends on a small number of film suppliers, planning that substitution in advance is prudent risk management. Factories that have already qualified a second material for their main builds can absorb a supply disruption that would otherwise stop production.

Inspection of microvias remains an area where the industry relies on sampling rather than full coverage, because there is no non-destructive method that reliably detects a small void or a thin wall. Automated optical inspection can verify that a hole exists and looks circular, but not that it is sound. Buyers should understand this limitation and treat the sampling plan as part of the specification rather than as an internal detail.

What to Verify When Sourcing

When sourcing HDI boards, the useful questions concern the microvia process rather than the equipment brand. What is the smallest via produced in routine production, and what is the achievable aspect ratio? How is bottom-of-hole residue controlled and verified? What proportion of production is sampled destructively, and what do those results show?

Answers should come with data. Cross-section images from routine production, via resistance distributions, and thermal cycling results on representative structures all provide evidence that the process is understood. Statements of capability without records are the least useful form of assurance.

Sourcing HDI boards is therefore an exercise in evaluating process control, and it is best approached with the same care as the design itself. A manufacturer who can discuss process in terms of measurements and limits, rather than in terms of features, is the one likely to deliver consistently when the design is difficult.