Multiple lamination cycles add cumulative thermal stress to the innerlayer bonds.

Industry requirements to meet the demands of lead-free applications have challenged many areas of PCB fabrication and design. These demands have initiated a rapid influx of new dielectric materials, which, despite increased thermal stress, must remain effectively bonded to copper innerlayers. Delamination failures in advanced multilayer and sequential-build products, caused by the conversion to lead-free soldering, have also driven the need for improved oxide replacement technology.

Such products must resist higher peak temperatures, some 30°C greater than eutectic applications. Based on statistically designed experiments, a new process has been developed that delivers a more resilient copper conversion coating. The process has shown improved bonding performance and greater stability at higher temperatures, thus improving its capability for more thermal excursions (assembly-reflow cycles at >260°C) without failure. This article focuses on the product development criteria and performance evaluation required to develop a process that is suitable for this challenging environment. The testing and qualification will include data on peel strength, solder float and IR-reflow testing.

Over the past decade, traditional alkaline black oxide (or brown oxide) bonding processes needed replacement due to a range of technical and environmental issues as well as heightened cost pressures. The acidic peroxide-sulfuric (oxide-replacement) processes quickly found favor due to lower cost and more environmentally friendly chemistry. These peroxide-sulfuric technologies, which use a primary organic additive to enhance the surface texturing, were shown to deliver excellent bond integrity in the pre-RoHS assembly setting.

However, it comes as no surprise that continuing advances in PCB fabrication are now pushing the performance requirements of these processes as well. These include:

• Advanced board constructions / sequential lamination / microvia capable build-up processes
• New halogen free, high Tg and filled phenolic cured dielectric materials
• Demanding, high temperature lead-free assembly conditions
  

The Challenges

The current peroxide-sulfuric oxide replacement process work based on a “controlled etch” mechanism that typically removes 45 to 75 microinches (1.1 to 1.9 μm) of copper. Innerlayers are first cleaned and then conditioned in an alkaline medium, rinsed, and then are processed in the peroxide-sulfuric micro-etching bath. The copper saturation capacity of this bath is limited and has historically been on the order of 18 to 22 g/L. Over the past three years, this technology has improved, allowing a higher solution capacity of 50 g/L in high-volume production environments. This capacity increase has been largely the result of improved copper solubility coupled with greater process stability.

The requirements for the high-yield production of controlled impedance and fine line innerlayers also call for further reductions in the copper etch levels. This low etch approach greatly benefits high frequency signal integrity by significantly reducing the “skin-effect” caused by a deep etch profile. Reduced etching is often at odds with improved bond performance. The need to improve bonding performance to meet the demands of new resin systems and RoHS assembly temperatures, with minimal removal in the order of 35 to 45 microinches (0.9 to 1.1 μm) of copper, has become a benchmark for a “best in class” process.

Within this challenge, higher values for peel strength and increased T260 and T288 (time-to-delamination) values are continually sought, with the process undergoing continuous review in order to deliver even greater performance. This improvement is needed to meet the requirements associated with the sequential lamination cycles applied in some build-up processes, and to withstand the extreme thermal stresses incurred in lead-free assembly and re-work.

The Oxide Replacement Process

The peroxide-sulfuric oxide replacement process is controlled by a primary organic material that modifies the etching mechanism to provide a consistent, highly textured, and characteristic surface structure. Not only does the primary organic additive accelerate the etch and modify the copper surface, but it also chemically combines with the copper to produce the characteristic organo-copper coating. This organic coating is based mainly on cuprous (Cu+) oxide but also contains some cupric (Cu+2) oxide. These give the conversion coating its characteristic red-brown color. The schematic formation of the alternative oxide coating is shown in Figure 1.

It is well understood that a cuprous-rich oxide coating typically delivers higher bond strength in the prepreg to copper layer. Cuprous oxide is more thermally stable than cupric oxide. In addition, a cuprous rich oxide has a higher chemical resistance to PTH chemistries, which can give rise to “pink ring” through the attack and separation of the oxide planes around the periphery of the holes. This chemical resistance is critical not only for the metallization of high aspect ratio through-holes, but more importantly for the landing pads on blind microvias. It is possible to post treat the oxide conversion coating using an alkaline process to remove the less resistant cupric oxide. Such post treatment can increase the peel strength of the coating by 10 to 15%, but it does not improve the resistance to thermal delamination.

The elemental analysis is shown for both a standard and a post-treated coating. Figure 2 shows the actual XPS spectra of the as-produced coating and post-treated coating (following a proprietary post dip). As would be expected, the post-treated layers have a lighter more-reddish appearance, indicative of the increased cuprous oxide content. This coating composition is also outlined in Table 1.

The Effect of Sequential Build

Increasingly, PCB construction, especially for thin card mobile or PDA-type products, is produced using sequential lamination techniques. Here, each lamination cycle adds additional and cumulative thermal stress to the innerlayer bond within the core construction. Depending on the lamination temperature, (typically 190 to 200°C), this can progressively reduce bond strength prior to final soldering, hence increasing requirements for improved thermal integrity. The negative impact on peel strength caused by sequential bonding cycles is shown in Figure 3.

The discussion of the repeated thermal cycles used in SBU fabrication must include consideration of the added effects of the higher temperatures seen in lead-free assembly. Exposure to multiple thermal cycles can have a cumulative effect on bond integrity.

Impact of Lead-Free Assembly on the Alternative Oxide Bond

The IR reflow soldering cycles used in assembly apply a large amount of stress to both the core and sequential build layers. Although the peak temperatures are targeted at 245 to 250°C, they can clearly overshoot these values, potentially rising to 260°C or more. At this point, or during the subsequent solder-wave application, the boards can fail due to delamination. Large copper areas, especially those carrying very high cluster-densities of small holes, can be especially vulnerable.

Delamination can be triggered by a combination of many factors, and the first step in determining root cause is to answer the key question – where is the delamination occurring? The problem is typically diagnosed by using cross section techniques to determine if the breakdown is between the copper-to-dielectric layer (adhesive failure), or within the dielectric itself (cohesive failure). Examples of these two conditions are shown in Figure 4 and Figure 5, respectively.

Product Improvement Approach

Following the transition to lead-free, the development group had already completed work designed to provide alternative oxide products with improved high temperature performance. An older, first generation, 25 g/L low copper capacity product (LCC) with improved thermal stabilization was used as the benchmark. This LCC prototype product had a demonstrated capability to withstand extended IR lead-free reflow cycles, and was developed specifically for critical high-end applications. For comparison, a 50 g/L high copper capacity (HCC)/low etch product was used to address the requirements of higher signal integrity, improved environmental capability (reduced waste), and lower cost of ownership.

The HCC technology used organic stabilization similar to the LCC product and had an already demonstrated better performance for peel strength and thermal shock resistance in high volume production environments. The defined approach was to further optimize the HCC technology through more exhaustive thermal stability testing, by using extensive L9 and L18 Taguchi methodology.

To test the product improvements against the benchmark, a rigorous testing protocol was developed. It included a test vehicle (Figure 6) that was based on a six-layer board with a rigid FR-4 based four-layer core. The FR-4 material was standard 150°C Tg epoxy. To simulate a sequential build-up product, resin coated copper was added as additional layers (one and six) to the four-layer core. Prior to lamination of the resin coated copper foil, the core layers two and five were electroplated with a standard acid copper process to simulate a buried via construction. This build allowed the examination of the effects of in-house electroplated copper on bond strength.

During the testing process the test vehicles were subjected to high stress thermal excursions of 270 to 280°C in an IR reflow chamber. The failures that occurred could be found in either the central core or within the surface layers. The profile of the IR reflow excursion is shown in Figure 7.

Initial Benchmarking Fesults

Using the IR profile described, copper innerlayers treated with the HCC process were subjected to the reflow cycle. Where older generation technology showed significant color change and pronounced thermal breakdown effect at temperatures over 260°C, the HCC process was able to withstand the peak of 270°C without any pronounced change of color.

In addition, Auger studies were used to determine the effects of temperature on the organo-coating. The Auger tests showed no significant loss (evaluating the elements Carbon and Nitrogen) or oxidation of the coating. Further, significant improvements in stability were seen over the lower soldering temperature range. No perceptible change or evidence of oxidation was experienced. The Auger results are shown in Figure 8.

The improved HCC technology also demonstrated a 6 to 15% improvement in peel strength of the coated layers. These results represent a major step forward in improving the coating’s thermal resistance. Furthermore, at 270°C the HCC technology was able to withstand an average of 1.5 additional cycles before failure due to delamination compared to the first generation processes.

Having achieved very satisfactory improvement with the high copper capacity formulation, the next step was to optimize the formula for thermal performance using the Taguchi approaches previously described.

The applied series of test arrays included a) the primary and secondary inorganic acids (that influence the etching characteristic and solubilize the copper); b) the organic components (that drive the modified etch reaction rate, produce the organo conversion coating, and stabilize the process and the coating); c) the grain refiner (essentially a chloride based species), and finally; d) the oxidizing agent – hydrogen peroxide (also drives the reaction-rate and texture depth).

The studied responses were coating color, etch-rate, pull/peel strength after coating, and most importantly, the resistance to delamination using multiple IR reflow cycles with peak temperatures of 270°C and 280°C respectively.

Details of some of the individual component plots, drawn from one of the later L18 arrays employed, are illustrated in Figure 9.

As expected, varying levels of response, and some interactions, were seen with many of the parameters employed. The findings were interesting for several reasons. It was determined that only one organic additive parameter significantly affected the color of the coating, but several of the additives had a significant influence on the delamination resistance. By contrast, the peel strength was mainly impacted by two significant factors.

Interestingly, all the factors that positively influenced the thermal performance also improved the primary peel strength results. These factors also contributed to a good aesthetic coating color and appearance. This can be seen in the similarity of the displayed Taguchi responses that moved largely in unison. This facilitated the completion of an optimized HCC product for the confirmation runs. A summary of the Taguchi optimization is shown in Table 3.

The confirmation testing showed an expected and very positive gain with the achievement of more than 10 IR reflow cycles without failure, based on the aggressive test vehicle and profiles involved. The resulting confirmation runs for delamination, as compared to the first generation technology and HCC standard benchmarks are shown in Figure 10.

Conclusions

The work described in this article has encompassed a long period of design, development and testing, and was carried out over a three-year period. Along the way, a lot of input has been received from the market, in terms of both defining the challenge and applying best practices to meet the growing industry requirement for improved PCB thermal resilience.

Several leading PCB fabricators have made significant progress in selecting improved materials and incorporating better lamination and process procedures, all of which open up the operating window for more demanding products. From this work, it is clear that many factors can significantly and adversely outweigh the alternative oxide contribution in achieving the balance required for bullet-proof performance.

To play its part, this study has been focused solely on finding ways to improve the performance and contribution of the alternative oxide bonding technology. The goal has been to provide a robust process with a wider operating window offering better thermal resilience.

Conclusions from this specific study can be summarized as follows. Delamination failures due to lead-free thermal stress can be caused by many factors, most significantly those involving the dielectric materials and the lamination process. In the broader context, industry opinion is that the alternative oxide process has a smaller impact than other material factors in delamination. Improving the process can, however, make an important contribution, especially for PCBs with large copper-to-dielectric bonding areas such as internal ground planes.

This can be seen with the older generation alternative oxide bonding technology, which cannot consistently meet the increasing industry requirement of highly advanced and performance sensitive designs for IR reflows without blistering, delamination or related electrical failure in lead free applications. The stability of the respective organo-metallic conversion coating diminishes at 260°C, where bonding performance decreases with each successive thermal excursion, leading to potential delamination issues.

The use of proprietary post treatments, which increase the cuprous oxide ratio can provide improvements to peel strength, but do not necessarily solve the thermal resilience challenge. One conclusion that can be drawn from this work is that the industry-standard pull/peel strength measure cannot adequately predict in-process bond strength, or a PCB’s resistance to delamination under thermal stress.

Improved LCC (low copper capacity) and HCC (high copper capacity) products have already demonstrated higher thermal stability than their earlier counterparts, along with greater capability to withstand multiple lead-free reflows at inflated peak IR reflow temperatures of 270 to 280°C.

The laboratory work described in this article establishes that an optimized HCC process can also more effectively meet the 10 plus multiple reflow requirements associated with sequential lamination and RoHS assembly simulation. Product design specifications for a low-etch attribute (1.0 to 1.2 μm) for controlled impedance, fine line integrity and improved high frequency signal integrity have been key factors in this work. Maintaining the required high copper capacity of 50 g/L delivers reduced environmental impact and the lowest cost of ownership. PCD&F

Dr. Abayomi I. Owei is president/principal research scientist at Avo Tech International Inc.; This email address is being protected from spambots. You need JavaScript enabled to view it.. Dr. Jean Rasmussen is R&D manager, PWB metallization, at Enthone; This email address is being protected from spambots. You need JavaScript enabled to view it.. Dr. Axel Dombert is European technical manager, PWB products, at Enthone; This email address is being protected from spambots. You need JavaScript enabled to view it.. Danis Isik is a research scientist, PWB products, at Enthone; This email address is being protected from spambots. You need JavaScript enabled to view it.. David Ormerod is business director, PWB metallization, at Enthone; This email address is being protected from spambots. You need JavaScript enabled to view it..