Cu Seed Composition and Thickness

PCB panel

Introduction

Seed copper is deposited during the PCB manufacturing process, often over other Cu layers, a process that requires accuracy in the thickness of this thin layer. Measurement of the seed layer thickness for quality assurance is available using Energy Dispersive X-ray Fluorescence (EDXRF), which overcomes the challenge of differentiating seed copper from the Cu layer beneath.

XRF Analysis

• XRF results of sample surface
• The following elements were detected: Cu, Al, Si, S, Fe, Ni, Cu
• Further analysis examples:
  
  • Light element detection
  • What other elements should be expected?

Acquisition time: 30 sec

Wet etch

• Etch performed by dipping the sample (for one minute) in a solution containing:
• H₂0₂ (3%)
• Citric acid
• NaCl

Measurement objective

To measure the thickness of the top, electroless-deposited Cu seed layer at several sample locations.

Electroless Cu seeding (1 µm)

• ABF, Ajinmoto Build-up Film, is a carbon-based epoxy film developed by Ajinmoto Co.
• FR-4, flame retardant class 4, is a type of laminated substrate material used in the manufacture of printed circuit boards (PCBs).

Top Cu layer isolation

• The top Cu layer was removed from one of the samples using wet etch.
• XRF comparison was made between etch and non-etch samples.
• Cu Lα comes only from the top Cu layer.

Cu Lα saturation test

• For the top Cu layer, 1 µm is much below the saturation depth.
• Saturation in XRF is a specific depth of the layer whereby intensity no longer changes beyond this layer thickness.
• Thickness changes correlate to changes in the current.
• XRF measurement was performed at the same location using varying currents in order to determine the ideal linear section of the intensity vs. current curve.

Cu thickness sample comparison

• XRF scan was performed over six points (1 mm step) on three different samples.
• Cu Lα peak intensity was compared.
• The average intensity of Samples B and C were used as reference with thickness of 1 µm.
• Sample A has significantly thinner top Cu layer than Samples B and C.
• The top Cu layer of Samples B and C has uniform thickness, unlike Sample A.

Cu intensity

Intensity vs. current

Across-wafer variation

Cu layer thickness [µm]

Cu seed thickness measurement by direct and indirect methods

• Cu layer thickness is measured by an automated algorithm using two methods.
• Direct method (0–0.7 µm) measuring the Cu Lα peak and calculating Cu layer thickness using linear regression
• Indirect method (0.7–1.4 µm) measuring the Si Kα peak and calculating Cu layer thickness using exponential regression
• Cu Lα peak intensity is saturated due to its low energy (0.93 keV), so the Cu layer thickness is measured indirectly using the Si from the ABF layer.

• Cu Lα and Si Lα were clearly detected by the EDXRF.
• Cu upper layer thickness can be monitored by using both peaks independently – Cu Lα or Si Kα peaks.
• Acquisition time: 12 sec

Regression curves

• Linear regression for the direct method
• Exponential regression for the indirect method

Roughness

• 3D scanner was used for measuring roughness by measuring surface height variations.
• Height variation up to 2 µm

3D scan imaging

2D microscope image

Summary

• XRF is very sensitive to thickness variation up to 0.1 nm. Many elements can be analyzed, qualitatively and quantitatively.
• Top Cu layer thickness can be measured using Cu Lα peak intensity changes.
• For accurate results, standards are required.
• The 3D scanner can monitor surface roughness.