Microvia Quality and Reliability in HDI Boards
A microvia is a hole that is formed by a laser and is smaller than the drill bits that conventional processes use. It is the structure that makes high density interconnect possible, and it is also the structure that fails in ways that a conventional through hole does not.
The failure modes are known, which is why reliability for a microvia design can be engineered. What cannot be engineered is a stack that was never characterised, so the fabrication data for a high density board has to state the geometry, the dielectric and the target pad with the same precision as the impedance data of a high speed board.
How a Microvia Is Formed
The hole is opened by a laser, usually a carbon dioxide or an ultraviolet source, and the energy is chosen so that it penetrates the dielectric and stops at the copper below. The copper acts as the stop layer, and the process window is defined by the difference in absorption between the resin and the metal.
The result is a small, tapered hole with a clean floor, and the quality of that floor decides everything that follows. Residue left on the target pad prevents the plating from bonding to the lower layer, and the electrical connection then depends on a bond that was never formed.
Target Pad and Registration
The target pad is the copper land at the bottom of the via, and it has to be larger than the via opening so that the laser always lands on metal. Registration between the laser and the pad is a fabrication capability, and the process window sets how much the layers may shift before the via misses.
A via that lands partly off the pad still forms a connection, but the connection is through a partial contact that concentrates current and stress at the edge of the overlap. This is the defect that shows up as an intermittent open after thermal cycling rather than at final test.
Aspect Ratio and Depth
The aspect ratio of a microvia, taken as depth divided by diameter, determines how well the plating chemistry reaches the bottom. A deep narrow via is harder to plate evenly than a shallow wide one, and the industry keeps the figure below a limit that the chemistry can support reliably.
Where a design needs more depth, the answer is usually a stacked or a staggered structure rather than a taller single via, since each additional level is a separate process with its own reliability data.
Plating Thickness and the Barrel Wall
The plating thickness in a microvia matters more than in a through hole because the barrel is short and the entry is narrow. A thin wall concentrates the current density during thermal cycling and cracks earlier, so copper thickness at the bottom corner is one of the parameters worth measuring.
Measurement is by cross section, because a surface inspection cannot see the wall. Coupons are sectioned and photographed, and the images are the evidence that the process is inside its window for that stack up.
Resin Recession and the Dimple
The resin around a microvia can recede during cure or during subsequent lamination, leaving the copper standing proud. The result is a dimple at the via top, which affects how the next layer bonds and how the surface behaves under a pad.
Resin recession is measured in the same cross section that provides the plating data, and it is one of the reasons a stack up with many sequential laminations needs a coupon for every level. A stack that passes at level two says nothing about level five.
Thermal Cycling and the Weak Point
Microvias fail under thermal cycling at the corner where the barrel meets the target pad, because that is where the geometry changes and the stress concentrates. The failure begins as a crack that grows with each cycle, and it becomes an open circuit only after the crack has crossed the wall.
This is why the qualification test for a high density product is a thermal cycling test on a daisy chained coupon, with resistance measured continuously. Intermittent opens that appear during the test are the early evidence, and they are missed by a test that only measures before and after.
Stack Up Choices
A single level of microvias is a well characterised structure with a long reliability record. Stacked vias, where one microvia sits directly on another, concentrate stress at the interface and require a more capable process, so they are used only where the routing demands them.
Staggered vias, offset from one level to the next, distribute the stress through the dielectric and are more forgiving. The choice is a layout decision with reliability consequences, and it should be made with the fabricator rather than assumed by the designer.
Inspection and Yield
Optical inspection cannot see inside the stack, and electrical test cannot detect a via that is connected by a partial contact until the connection fails. Cross sectioning on coupons and on a sample of production panels is the practical check.
Where a stack is new, the sample rate should be high for the first builds and then reduced once the process has demonstrated capability. This mirrors the approach used for the parameters listed in manufacturing tolerances, where the sampling is adjusted to the demonstrated stability.
Design Rules That Protect Reliability
Keeping the microvia in the outer layers, avoiding a via directly under a pad that will be soldered, and leaving copper around each via are rules that reduce both fabrication risk and assembly risk. They cost routing space, which is why they are often challenged and rarely wrong.
The stack up should also be documented with the dielectric materials and their glass transition temperatures, because reliability under thermal cycling depends on the resin as much as on the copper. A design released without that data cannot be assessed later.
Process Control and Verification
On a design of this kind, plating thickness is the item that decides how the rest of the board is arranged. Running a first article through the same checks as the production panel confirms that the two agree, and that comparison is the cheapest form of process control available at prototype stage. Keeping a sample from the panel turns a dispute into a measurement, because the same coupon can be re-examined by both parties without rebuilding the batch.
Related reading: our fabrication notes, board quality and design release notes cover the same ground.
A first article check confirms that the process and the drawing agree on the points listed above, and that the coupon data supports the values used in the design.
Where a measurement falls outside the expected window, the sample is retained so that the cause can be established before the balance of the batch is released.
A first article check confirms that the process and the drawing agree on the points listed above, and that the coupon data supports the values used in the design.
Where a measurement falls outside the expected window, the sample is retained so that the cause can be established before the balance of the batch is released.
FAQ
Can a microvia be used in a via in pad structure? It can, provided the via is filled and capped, since an open microvia under a pad drains paste exactly as a through via does.
How many levels of microvia are reasonable? As many as the product requires and the fabricator can characterise, with each additional level adding a separate coupon and reliability dataset.
Is a stacked microvia less reliable than a staggered one? Generally yes, because the stress concentrates at the interface between the two barrels rather than spreading through the dielectric.
What proves a microvia is good? A cross section that shows full plating to the target pad, combined with continuous resistance monitoring through thermal cycling on a coupon from the same build.
