CCL Delamination and CAF Failure: Humidity, Thermal Cycling and Drilling Parameter Root Cause

Overview #

The two failure modes that generate the most field returns in CCL-based PCB assemblies — delamination and conductive anodic filament (CAF) formation — are almost never caused by the laminate material alone. In our supplier qualification program, we consistently find that buyers who specify FR-4 Tg 150°C and consider the job done are the same buyers who call us six months later with blistering failures at reflow. The root causes are almost always a combination of incoming material moisture content, drilling parameter deviation, and thermal excursion — and all three are measurable before the board ever reaches the assembly line. Understanding the threshold values for each failure mode, and knowing which test to run at which stage, is what separates a qualification program that holds at production volume from one that passes samples and fails batches.

CCL Delamination: Causes, Thresholds, and Detection #

Delamination in copper-clad laminates occurs when the bond between the resin matrix and the glass fabric — or between laminate layers in a multilayer stack — fails under thermal or mechanical stress. The failure is irreversible. Once the interface separates, no rework recovers the board.

Primary cause: moisture absorption before or during lamination

The most common trigger we see in Chinese CCL supply chains is moisture ingress during storage or transit. FR-4 laminates absorb moisture at the resin-glass interface, and when that moisture vaporizes during reflow — typically at 260°C peak in lead-free profiles — the vapor pressure exceeds the interlaminar bond strength. The critical threshold is a moisture content above 0.2% by weight (per IEC 61189-2 test conditions), which is routinely exceeded when panels are stored in uncontrolled warehouse environments above 60% relative humidity for more than 72 hours.

The measurable parameter buyers should request on every incoming lot is Time to Delamination (T-288 or T-300), tested per IPC-TM-650 Method 2.4.24.1. T-288 measures the time in seconds before delamination occurs at 288°C in a DSC instrument. For standard FR-4 Tg 150°C material, a T-288 value below 5 minutes is a rejection threshold in our incoming inspection program. High-Tg FR-4 (Tg ≥ 170°C) should hold T-288 ≥ 15 minutes. If a supplier cannot provide T-288 data on the COA — not just Tg — treat that as a qualification red flag.

Secondary cause: resin content deviation

Resin content in prepreg directly controls the void-filling capacity during lamination pressing. The nominal resin content for standard 7628 glass weave prepreg is 43 ± 3%. When resin content drops below 40%, incomplete fill at the glass-resin interface creates micro-voids that act as delamination initiation sites under thermal cycling. We have seen Chinese prepreg suppliers ship material at 38–39% resin content while the COA reads 43% — the discrepancy is only caught by incoming resin content testing per IPC-4101 specification sheets.

Comparison: Delamination Risk by Material Grade and Condition

The Tg deviation row is not a specification — it is what we observe across incoming lots from mid-tier Chinese CCL suppliers. The ±10°C deviation on halogen-free grades reflects the fact that halogen-free resin systems are more sensitive to cure cycle variation, and Chinese compounders vary cure parameters more than their datasheets suggest.

Most Western buyers do not realize that GB/T 4722 — the Chinese national standard governing CCL electrical and mechanical properties — permits Tg measurement by either DSC or TMA, and the two methods can produce results that differ by 10–15°C on the same material. A supplier reporting Tg 150°C by TMA may be delivering material that measures 138°C by DSC. Specify the test method on your purchase order, not just the value.

For buyers sourcing PCB & Electronic Substrates from China, the pre-bake step before assembly is not optional — it is a process control that compensates for the moisture exposure that is almost certain to have occurred somewhere in the supply chain between the CCL mill and your SMT line.

CAF Formation: Drilling Parameters, Glass Weave, and Humidity Thresholds #

Conductive anodic filament (CAF) failure is the migration of copper ions along the glass fiber-resin interface between adjacent drilled holes or between a hole and a conductor, forming a conductive bridge that causes insulation resistance degradation or short circuit. CAF is a latent failure — it does not appear at incoming inspection or even at initial functional test. It develops under bias and humidity over weeks or months in the field.

The drilling parameter that most procurement teams ignore

The parameter that drives CAF susceptibility more than any other is drill bit condition at the time of drilling — specifically, the degree of glass fiber pullout and resin smear at the hole wall. When a worn or incorrectly fed drill bit tears rather than cuts the glass fibers, it creates a continuous channel along the fiber-resin interface. That channel is the CAF migration path.

The measurable threshold: drill bit life should not exceed 3,000 hits for standard FR-4 at 1.0 mm diameter, and feed rate should be maintained at 0.025–0.030 mm/rev. When feed rate drops below 0.020 mm/rev (a common outcome when fabricators try to extend bit life), fiber pullout increases sharply and CAF susceptibility rises. We request drill parameter logs — not just hole quality inspection reports — as part of our fabricator qualification package.

Humidity threshold for CAF initiation

CAF formation requires three conditions simultaneously: a migration path (the drilled channel), ionic contamination (flux residue, process chemistry), and bias voltage under humidity. The critical humidity threshold for CAF initiation in standard FR-4 is 85% RH at 85°C under bias — the standard IPC-TM-650 Method 2.6.25 CAF test condition. The pass threshold for insulation resistance under these conditions is ≥ 100 MΩ after 500 hours. Material that passes at 168 hours but fails at 500 hours is not uncommon in our testing of Chinese CCL — and 168-hour data is what most suppliers provide by default.

Always request 500-hour CAF test data, not 168-hour data. The difference is not academic — it is the difference between a product that passes qualification and one that fails in the field at 18 months.

Hole-to-hole spacing and CAF risk

CAF risk increases non-linearly as hole-to-hole spacing decreases below 0.5 mm. At 0.3 mm spacing — now common in HDI designs — standard FR-4 with typical Chinese drilling practice shows CAF failure rates that are 3–5× higher than at 0.5 mm spacing under the same 85°C/85% RH/bias conditions. If your design calls for hole-to-hole spacing below 0.4 mm, specify CAF-resistant resin systems (typically modified epoxy or cyanate ester blends) and require 500-hour test data at the actual design spacing, not at the standard 1.0 mm test coupon spacing.

In our qualification program, we have seen suppliers pass CAF testing on standard 1.0 mm spacing coupons and then deliver boards with 0.3 mm spacing that fail in the field within 12 months. The test coupon spacing was not representative of the actual design. This is a specification gap that the buyer owns, not the supplier.

Thermal Cycling Failure: Z-Axis Expansion and Via Barrel Cracking #

Thermal cycling failure in CCL-based multilayer boards manifests as via barrel cracking — the copper barrel inside a plated through-hole fractures due to the mismatch between the Z-axis CTE of the laminate and the CTE of the copper plating. The copper CTE is approximately 17 ppm/°C. Standard FR-4 Z-axis CTE is 50–70 ppm/°C below Tg, rising sharply above Tg. The mismatch drives fatigue cracking at the knee of the via barrel over repeated thermal cycles.

The threshold that determines via reliability

The critical parameter is Z-axis CTE, not Tg alone. A material with Tg 170°C but Z-axis CTE of 65 ppm/°C will crack vias faster under thermal cycling than a material with Tg 150°C and Z-axis CTE of 45 ppm/°C, if the operating temperature stays below Tg. Most procurement specifications list Tg and ignore Z-axis CTE entirely. We consider Z-axis CTE ≤ 50 ppm/°C (below Tg) a minimum requirement for any multilayer board that will see more than 500 thermal cycles between -40°C and +125°C.

The test method is IPC-TM-650 Method 2.4.41 (TMA measurement of Z-axis CTE). Request this data on the material COA, not just on the datasheet. Datasheet values are nominal; COA values reflect the actual production lot.

Production failure scenario: via cracking in automotive sensor PCBs

We evaluated a batch of multilayer PCBs for an automotive sensor application — 6-layer boards, 1.6 mm thick, via diameter 0.3 mm, copper plating thickness 20–25 µm. The CCL was specified as FR-4 Tg 150°C from a Chinese tier-2 supplier. Field returns began appearing at 14 months, with intermittent open circuits on inner-layer vias.

Cross-section analysis showed via barrel cracking at the inner-layer interface, consistent with Z-axis CTE fatigue. Incoming lot testing of the CCL revealed Z-axis CTE of 68 ppm/°C — within the supplier’s datasheet range of 60–70 ppm/°C, but the purchase order had not specified a Z-axis CTE limit. The Tg was confirmed at 152°C by DSC, which was within spec. The failure was entirely predictable from the material data — but the buyer had not requested Z-axis CTE on the COA, and the supplier had not volunteered it.

The corrective action was to re-specify the material with Z-axis CTE ≤ 50 ppm/°C and switch to a Chinese tier-1 CCL supplier whose production process controls Z-axis CTE through controlled resin content and cure cycle. The re-qualified boards have now completed 1,000 thermal cycles (-40°C to +125°C) without via failure.

This is the pattern we see repeatedly: the failure is not caused by a defective material — it is caused by a specification that did not include the parameter that actually drives the failure mode. Tg is the parameter buyers know. Z-axis CTE is the parameter that matters.

Practical Guidance for Buyers #

When sourcing CCL or finished PCBs from China, the first specification to request from suppliers is not Tg — it is T-288 (time to delamination at 288°C) and Z-axis CTE, both on the production lot COA, not the datasheet. Most buyers ask for Tg because it is the most visible parameter in marketing materials. T-288 and Z-axis CTE are the parameters that predict actual failure in lead-free assembly and thermal cycling, and they are the ones most likely to deviate between the datasheet and the production lot.

The most common sourcing mistake we see is accepting a single sample COA at qualification and not requiring three consecutive production lot COAs before committing to volume. In our experience, lot-to-lot variation in Tg (±8°C on standard FR-4 from mid-tier Chinese suppliers) and Z-axis CTE (±15 ppm/°C) is large enough to push a borderline-specified material into the failure zone on the third or fourth lot.

Before committing to volume, require: T-288 ≥ 5 minutes (standard FR-4) or ≥ 15 minutes (high-Tg FR-4) per IPC-TM-650 Method 2.4.24.1; Z-axis CTE ≤ 50 ppm/°C below Tg per IPC-TM-650 Method 2.4.41; and 500-hour CAF test data per IPC-TM-650 Method 2.6.25 at the actual design hole-to-hole spacing. If a supplier cannot provide all three on a production lot COA, qualify a different supplier.

For related sealing and thermal interface materials used in PCB assembly environments, see Sealing & Thermal Interface Materials for compatible consumable specifications.

Frequently Asked Questions #

Q1: What is the most important test parameter to specify on a CCL purchase order for lead-free assembly?

A: T-288 (time to delamination at 288°C), not Tg. A T-288 value below 5 minutes on standard FR-4 Tg 150°C material is a rejection threshold — Tg alone does not predict delamination resistance under lead-free reflow conditions.

Q2: How do I select between standard FR-4 and high-Tg FR-4 for a multilayer board that will see 500+ thermal cycles?

A: The selection criterion is Z-axis CTE, not Tg alone. High-Tg FR-4 (Tg ≥ 170°C) typically delivers Z-axis CTE of 45–55 ppm/°C versus 60–70 ppm/°C for standard FR-4 — that difference determines via barrel fatigue life under thermal cycling. Refer to the comparison table above and require Z-axis CTE data per IPC-TM-650 Method 2.4.41 on the production lot COA, not just the datasheet.

Q3: What drilling parameter deviation most commonly causes CAF failure in Chinese PCB fabrication?

A: Feed rate below 0.020 mm/rev, usually caused by fabricators extending drill bit life beyond 3,000 hits. This is where most CAF failures originate — not from the laminate material itself, but from the drilling process. Request drill parameter logs, not just hole quality inspection reports.

Q4: What CAF test documentation should I require before approving a Chinese CCL supplier for a high-reliability application?

A: Require 500-hour insulation resistance data (≥ 100 MΩ pass threshold) per IPC-TM-650 Method 2.6.25 at 85°C/85% RH under bias, tested at the actual design hole-to-hole spacing. Most Chinese suppliers provide 168-hour data by default — that is not sufficient for field reliability validation.

Q5: Does pre-baking CCL panels before SMT assembly actually prevent delamination, or is it just a precaution?

A: It prevents delamination when moisture content has exceeded 0.2% by weight — which, in Chinese supply chains with uncontrolled warehouse storage, is more often the rule than the exception. Pre-bake at 120°C for 2 hours is not a precaution. It is a process control that compensates for a real and measurable incoming material condition.

Published by sinoraw.com Technical Team | Request a sourcing consultation