Wire Bonding & Soldering on ENEPIG, ENEP Surface Finishes with Pure Pd-Layers

Estimated reading time: 5 minutes

Editor's Note: This article originally appeared in the January 2013 issue of The PCB Magazine.

Abstract

As a surface finish, electroless nickel/electroless palladium/immersion gold (ENEPIG) has received increased attention for both packaging/IC-substrate and PWB applications. With a lower gold thickness than conventional electroless nickel/immersion gold (ENIG) the ENEPIG finish offers the potential for higher reliability, better performance and reduced cost [1,2].

This paper shows the benefits of using a pure palladium layer in the ENEPIG and ENEP (electroless nickel, electroless palladium) surface finishes in terms of physical properties and gold wire bonding test results.

Introduction

The ENEPIG surface finish originated in the mid-1990s as a modification of the conventional ENIG finish. During development of ENEPIG, it was recognized that the addition of a palladium (Pd) layer between the nickel and gold enabled both gold and aluminum wire bonding operations, in addition to the normal soldering application. In addition, the Pd layer was found to limit the corrosion of the nickel by an overly aggressive immersion gold process. An electrolytic nickel/gold finish was typically the process of record (POR) for such wire bonding needs.

Comparison of Properties of Pure Palladium vs Palladium-Phosphorus (PdP) Deposits

One subtle difference in the ENEPIG processes available in the market pertains to the deposition of electroless palladium. The electroless palladium layer in ENEPIG can be deposited either as a palladium-phosphorous alloy (PdP) or as “pure” palladium. The deposition mechanism may be similar, because both can be deposited in an autocatalytic (electroless) manner. However, the physical properties of the two deposits are quite unique, resulting in differences for the assembly steps of soldering and wire bonding.

Hardness of Electroless Deposited Palladium

One key difference between the two types of palladium layers relates to the hardness of PdP and pure Pd deposits. Increasing the phosphorus content also increases the hardness of the palladium deposits, as shown in Figure 1.

Figure 1: Hardness comparison of palladium-phosphorus and pure palladium autocatalytic deposits.

The hardness of autocatalytically deposited pure Pd is 250 HV, whereas the hardness of PdP (with 4-6% phosphorus content) is approximately twice that value. The lower hardness of pure Pd is regarded as one explanation for the better wire bonding performance of ENEPIG with pure Pd in comparison to ENEPIG with PdP.

Internal Stress in Deposited Pd Layer

The value of internal stress is an indicator of the amount of mechanical energy captured within the layer after the electroless deposition. The Pd crystal structure and the type of electroless deposition influence this value. Lower internal stress is clearly shown for pure Pd. The reason for this difference is presumed to be the different crystal structures of pure Pd and PdP.

Table 1: Comparison of internal stress of PdP and pure Pd deposits.

Topography of Electroless Palladium

When comparing the surfaces of pure Pd and PdP depositions, some difference in the topography is apparent. As shown in Figures 2 and 3, the PdP surface shows an even and smooth topography within the individual grains, whereas pure Pd exhibits a form of nano-roughness. The larger grains reflect the known structure of the underlying nickel layer.

Figure 2: PdP deposition (0.15 µm) over nickel, showing a relatively even and smooth surface.

Figure 3: Pure Pd deposition (0.15 µm) over nickel, showing some nano-roughness on the surface.

Crystal Structure

As illustrated in Figures 4 and 5, cross-sections show that the crystal structure of PdP is amorphous, whereas pure Pd is characterized by a fine crystalline structure.

Figure 4: PdP deposition (0.30 µm) shows an amorphous structure.

Figure 5: Pure Pd deposition (0.15 µm) on nickel shows a fine crystalline structure.

Test Conditions for Gold Wire Bond Investigation

The following wire bond test conditions were used for the further wire bond investigations:

Figure 6: Test Layout for wire bonding.

Table 2: Wire bonding and sample parameters.

Figure 7: Comparison of gold wire bond pull test results for ENEPIG (with PdP) vs. ENEPIG (with pure Pd) with varying thickness of Ni, Pd, and Au.

Figure 8: Comparison of gold wire bond failure mode results for ENEPIG (with PdP) vs. ENEPIG (with pure Pd) with varying thickness of Ni, Pd, and Au.

To assess the wire bond performance of ENEPIG finishes with pure Pd in comparison to PdP, investigations were conducted with varying thicknesses of gold, palladium and nickel. As shown in Figures 7 and 8, almost no difference exists between the two finishes in terms of either wire pull force or failure mode for samples with a thicker gold deposit (0.15 µm). However, in the case of lower gold thickness (0.04 µm) the ENEPIG finish with pure Pd exhibits significantly greater pull strength results and a higher incidence of the preferred wire bond failure mode. It is theorized that reducing the gold thickness increases the effect of the palladium hardness on the wire bonding process. Furthermore, it is assumed that a softer Pd layer is beneficial for the wire bonding process. As known from electrolytic deposited Ni/Au (i.e., “soft” gold), the hardness does have a significant influence on gold wire bonding. Conversely, electrolytic deposited hard gold is not used for wire bonding in the market. As such, ENEPIG with pure Pd can operate with a wider operating window for gold wire bonding, but more importantly, it can operate with lower gold thickness and still achieve similar results.

Copper Wire Bonding Capability of ENEP Surface Finish

With respect to the ENEP surface finish, the use of pure Pd does provide a further significant benefit. Recent investigations have shown that copper wire bonding is possible for IC substrate and PWB applications when performed on ENEP surface finishes having a pure Pd layer. For semiconductor applications, copper wire bonding on pure Pd ENEP is already established [4, 5, 6].

Figure 9: Typical copper wire wedge bond.

Summary

These investigations show that using electroless pure Pd depositions (without co-deposited phosphorus) can enhance the performance of ENEPIG surface finish. In the case of ENEPIG, the use of pure Pd widens the process window for gold wire bonding and, as demonstrated, allows a reduction in the gold thickness, thus enabling an increase in yield on the assembly side as well as a possible cost reduction. In addition the ENEP surface finish with pure Pd is offering the associated cost reduction by avoiding the expenses for the gold bath and ENEP with pure Pd enables next generation interconnection techniques, namely copper wire bonding.

References:

1. Johal, K., Roberts, H., and Lamprecht, S., “Electroless Nickel/Electroless Palladium/Immersion Gold Process For Multi-Purpose Assembly Technology,” SMTA International Conference Proceedings, 2004.2. Roberts, H., Lamprecht, S., Ramos, G., and Sebald, C., “Effect of Process Variations on Solder Joint Reliability for Nickel-based Surface Finishes,” SMTA Pan Pacific Microelectronics Symposium Proceedings, 2008.3. Roberts, H., Lamprecht, S., and Sebald, C., “Alternative Nickel-based Surface Finishes for IC Substrate Applications in a Pb-free Environment,” IMAPS International Conference and Exhibition on Device Packaging Proceedings, 2008.4. Horst Clauberg, Asaf Hashmonai, Tom Thieme, Jamin Ling, and Bob Chylak, “Nickel-Palladium Bond Pads for Copper and Gold Wire Bonding.” 5. Bob Chylak, Jamin Ling, Horst Clauberg, and Tom Thieme, “Next Generation Nickel-Based Bond Pads Enable Copper Wire Bonding.” 6. Horst Clauberg, Petra Backus, and Bob Chylak, “Nickel-Palladium Bond Pads for Copper Wire Bonding.”