Final Report Summary - ASPIS (Advanced Surface Protection for Improved Reliability PCB Systems)
Executive Summary:
The three year Aspis ‘Research for SMEs’ project has been conducted to address problems found with the nickel gold (ENIG) solderable finishes that are used by the electronics industry to enable components to be soldered to printed circuit boards (PCBs). ENIG has several key advantages over other types of PCB solderable finishes and these are highly significant, making ENIG the preferred choice for many applications. However, while ENIG coatings have a good reputation for excellent solderability, they are also prone to a number of well known reliability problems, such as ‘black pad’, whose formation mechanisms were still poorly understood. When ENIG-related problems occur, they can affect an entire product design or batch and the problems are often only identified after assembly, during which expensive components have been soldered to the PCBs.
The ASPIS project has focused on the development of new chemical processes, as well as methods for detecting ENIG related issues. Via four key work packages, a multi-faceted approach to ENIG reliability problems has enabled significant progress to be made on all fronts. The ASPIS project has carried out work to reduce the likelihood of ENIG-related problems occurring by improving the coating deposition technologies. Through the specific work packages on metal plating development, the ASPIS project developed new processes that will enable thinner coatings to be used, thus reducing the cost of using ENIG while offering reliability enhancements. Another key goal was to develop methods for identifying potential problems related to the use of ENIG on PCBs and a non-destructive screening method has been demonstrated that identify problems on boards before components are soldered. One way to prevent ENIG-related problems is to avoid them altogether by using different coating technologies. With work carried out by partners ITRI and Leicester University the project successfully developed alternative deposition methods from both aqueous solution and ionic liquids. Several new gold deposition processes were successfully developed.
The knowledge outputs of the ASPIS project will help to reduce the number of ENIG-related problems reaching production and also accelerate investigations and appropriate response times. This will reduce associated costs and increase customer confidence in the ENIG process. The development of an assembly line tool for identifying problematic PCBs will reduce lost production costs and lower the risk of field-failures and their potential consequences. The improved and alternative coating technologies developed are now available for further development and will enable European PCB fabricators and their customers to produce more competitive products via both increased process and assembly yields and to provide higher levels of quality in finished products.
The work carried out in the project has been widely publicised via technical papers, articles in trade journals, conference presentations and posters, as well as via an active website and industry briefing events.
Project Context and Objectives:
The three year ASPIS project, (Advanced Surface Protection for Improved Reliability PCB Systems) undertook a multi-faceted approach for developing novel and improved reliability nickel-gold solderable finishes used in the assembly of electronic products. Aspis was a ‘Research for SME Associations’ project with twelve partners including four key research organisations (RTDs) who undertook much of the research. These were ITRI Ltd (UK), the University of Leicester (UK), TNO (The Netherlands) and the Center for Physical Sciences and Technology (Lithuania). The other partners represented a wide cross section of organisations from the PCB sector covering the whole of the industry supply chain. These organisations were Atotech GmbH (Germany), Graphic Plc (UK), Merlin Circuits Technology Ltd (UK), Somacis S.p.a (Italy) and Global Interconnect Services (The Netherlands), the European Institute of Printed Circuits (The Netherlands) and the Institute of Circuit Technology (UK).
The ASPIS project had a specific focus on nickel-gold (ENIG) solderable finishes for PCBs, and its key objectives were to develop new, more reliable materials and processes in order to address the key issues such as ‘black pad’ that have been a cause of concern for many years to both PCB fabricators and end users. One of the key factors impacting the overall reliability of electronic assemblies is the quality of the solder joints that connect components to a circuit board and a critical factor in determining this reliability is the solderable finish that is applied to the board. The standard choice of solderable finish for high reliability, high value electronics is nickel-gold. This utilises an undercoat of electroless nickel which is deposited on to the copper of the PCB and on top of which is a thin coating of immersion plated gold. Globally, the market share of ENIG in the PCB industry as a whole is estimated at 15% by surface area of PCB manufactured. However, in terms of the value of the resulting products, the market share is considerably higher. For the EU’s electronics manufacturers, which are increasingly producing high value and high reliability products, the importance of ENIG is much greater. ENIG has several key advantages over other types of PCB solderable finishes. These are highly significant as, despite the higher price of ENIG over other alternatives, it is the preferred choice of finish for many applications. A key advantage of ENIG is that it offers excellent solderability and can retain this during prolonged storage prior to soldering assembly and during the multiple soldering operations encountered when assembling complex, high value electronics. While it was predicted that ENIG’s market share would increase significantly after implementation of the Restriction of certain Hazardous Substances Directive (RoHS) in 2006, the expected increase has not materialised due to a combination of cost, the emergence of some lead-free HASL alternatives and because of a fear of ENIG-specific problems which can lead to disastrous yields and product failures if they occur. Another advantage of ENIG finishes is that they have excellent planarity and this is crucial when assembling and soldering small fine-pitched surface mount components. These are often of high value and placement accuracy is critical for achieving high assembly yields. Unlike other coating technologies, which are applied solely to facilitate storage and solderability during soldering assembly, ENIG is a functional coating that utilises the solid-state properties of the nickel to stabilise the soldered interface, in effect, acting as a barrier layer to prevent the growth of tin-copper intermetallic compounds, which can embrittle the soldered joints during the service-life of the product. This greatly enhances the reliability of ENIG-coated PCB assemblies, especially during prolonged use at elevated temperatures.
However, while ENIG coatings have a good reputation for excellent solderability, there are a number of technical and economic factors which can cause problems for PCB fabricators and their customers. The most widely recognised of these is referred to as ‘black pad’ which is thought to be attributable to excessive corrosion of the nickel coating during the subsequent immersion gold deposition process. Despite having been identified more than 10 years ago, the mechanisms that cause the effect (and the contributing factors) were still poorly understood prior to the work carried out in this project. Moreover, no design rules, non-destructive testing methods, process optimisation/mitigation steps or viable alternatives existed. Other problematic failure mechanisms include insufficient gold coating thickness, gold coating porosity, nickel migration, excessive phosphorus levels in the nickel layer, poor plating quality, excessive coating stress in the nickel, and general solderability degradation. When ENIG-related problems occur they usually affect an entire product design or batch and the problems are often only identified after assembly, during which expensive components have been soldered to the PCBs. Many ENIG-related problems are only identified later via field-failures and when this occurs, large sums of money are spent identifying and resolving the problem, usually by several members of the electronics supply chain. In some circumstances, especially if liabilities and compensation claims are pursued, the total cost of an ENIG-related problem could be orders of magnitude higher than the cost of the PCBs supplied by the fabricator. It is probable that any ENIG user who has been using the finish for a substantial length of time will have experienced some of these problems.
The ASPIS project aimed to help both the European PCB fabrication and broader electronics industries by developing new chemical processes, as well as methods for predicting, avoiding and detecting ENIG related issues. Prior to the Aspis project,, the failure mechanisms were not fully understood and, therefore, the factors which could contribute to them were not all known. This prevented fabricators and assemblers from properly optimising their processes and users from predicting or even identifying which PCBs might be affected or by which type of mode products have failed. The ASPIS project also had the overall objective of strengthening the European electronics industry supply chain, from material suppliers to end producers, by helping ENIG users and by offering enhanced coating technologies with similar solderability and reliability benefits. It was important to support ENIG users because PCB fabricators’ customers will continue to specify it for their products, regardless of whether new alternatives emerge. This was typically due to familiarity, the lack of historical data of field reliability and because purchase specifications exist which cannot be changed without requalification; a long and expensive process. For these people, the fundamental knowledge developed in ASPIS will help them to identify and avoid design and process-related factors which can increase the risk of failures occurring. As ENIG will continue to be used on existing boards, which obviously cannot retroactively be subject to these precautions, the knowledge outputs of the Aspis research will also enable engineers to identify key areas to investigate in order to validate the quality of the ENIG coated boards. Another key objective of the ASPIS project was to help reduce the likelihood of ENIG-related problems occurring by improving the coating deposition technologies. Unlike previous developments, which have largely been undertaken empirically, the fundamental mechanistic knowledge developed was used to intelligently target specific improvements. For example, one important area to be addressed was the thickness of the gold deposited. Sufficient gold is required to provide a continuous non-porous film but, as this thickness increases, so does the cost and there can also be a negative impact on solder joint reliability. Through specific work packages focussing on plating development the ASPIS project researched new processes that enabled thinner coatings to be used, thus reducing the cost of using ENIG while offering the potential for longer term reliability enhancements. Another key goal was to develop methods for identifying potential problems related to the use of ENIG on PCBs and two key approaches were investigated. The first was to use laboratory methods for identifying which type o failure mechanism caused a particular problem. The second method was to develop a non-destructive screening method that can be used at the PCB fabricator or electronics assembler in order to identify problems on boards before components are soldered onto them. Therefore, another key aim of the ASPIS project was to build a prototype instrument for demonstration and validation purposes. This was successfully achieved and performance evaluation trails were carried out by RTD performer TNO, who developed the equipment, within the production facility of SME partner Graphic.
One way to prevent ENIG-related problems is to avoid them altogether by using different coating technologies, either via different deposition methods or by the use of alternative materials or structures. Research carried out in specific work packages of the ASPIS project investigated and developed alternative deposition methods from aqueous solution and ionic liquids. Most of the methods for depositing different metals from aqueous solutions are well established and are widely used for other applications. The novelty is applying them to this high reliability, functional, solderable coating technology. Plating from ionic liquids is a more recent development and is not yet used in the PCB fabrication industry. However, it has been shown that gold deposited from ionic liquids exhibited improved properties which may, for example, enable thinner coatings to be used.
The current lack of in-depth understanding of the nickel-gold process means that no design rules exist for avoiding problems such as ‘black pad’, nor are there standards in place that can be used to minimise the risks. The most widely applied industry standard, the IPC 6010 series, merely specifies a minimum coating thickness for the nickel and gold layers of 3.00 μm and 0.05 μm respectively for coatings on rigid boards. The knowledge, data and experiences gained and accumulated in ASPIS will be used to improve and expand the industry standards and purchasing specification requirements. Improved ENIG or alternative coating technologies will, likewise, require applicable standards in order to get industry acceptance.
The new information and methodologies being developed by the ASPIS project will help reduce the number of ENIG-related problems reaching production and also accelerate investigations and appropriate response times. This will reduce associated costs and increase customer confidence in the ENIG process. The development of an assembly line tool for identifying problematic PCBs will reduce lost production costs and lower the risk of field-failures and their potential consequences. The development of improved and alternative coatings technologies will also enable European PCB fabricators and their customers to produce more competitive products via both increased process and assembly yields at subsequent higher levels of quality in finished products. Reducing the quantities of nickel and gold used in the solderable coatings by improvement or replacement will also reduce the overall cost of the PCB finish and also help the European electronics industry to adopt more sustainable approaches to materials use.
Project Results:
Introduction
As detailed above, the Aspis project had a number of key objectives aimed at understanding the causes of reliability problems in ENIG finishes. These were undertaken via a four pronged approach and comprised an investigation of the basic black pad mechanisms, formulation of new aqueous and ionic liquid basic deposition chemistries and the evaluation of non-destructive test methodologies and equipment. These were addressed through specific work packages led by each of the individual RTD providers. The scientific and technical foreground is thus reported specific to each of these key individual work packages.
Work Package 2 - Fundamental Research on ENIG Mechanisms
Overview
Prior to the Aspis project the causes of black pad and other related reliability problems in ENIG solderable finishes were poorly understood and the fundamental mechanisms of black pad formation had not been fully elucidated. Therefore, a key to eliminating these problems and improving reliability was gaining a full understanding of the mechanisms and this work was undertaken by the partner ‘LIOC’, in Vilnius, Lithuania, in Work Package 2. The principal scientific objectives of this work package were as follows:
• Identification of mechanisms influencing each failure mode via experimental determination.
• Determination and quantification of key factors and relationships for failure modes identified.
• Establishment of rules for PCB design and layout.
• Recommendations for improved international standards and purchasing specifications.
Five specific tasks were predicted in this work package with the following short characteristics of the planned assignments in each of them.
Task 2.1. A laboratory based study to investigate the fundamental mechanisms and key influencing factors which can lead to ENIG problem.
The following studies were undertaken with the aim of determining the key factors and conditions, which favoured enhancement of the corrosion rate of Ni-P coatings:
• deposition and detailed characterization of Ni-P and coatings, including the following parameters: phosphorus content, thickness, porosity, internal strength
• corrosion behaviour studies of Ni-P coatings in supporting solutions
• characterization of gold layers including the following parameters: thickness, porosity and quality
• corrosion behaviour of Ni-P coatings in gold deposition solutions: a structural characterization of Ni-P samples after removal of the gold layer, including phosphorus distribution on the surface.
Task 2.2. Experimental investigation of the various phenomena and, particularly, the relationship between the plating process, coating quality, PCB design and other associated factors that have been linked with the failure mechanism.
The following investigations were planned: analysis of failures due to mechanically weak areas in Ni-P/solder bonding, which might be caused by:
• Ni-P coating characteristics: the presence of Ni surface morphology defects (mud cracks, major/minor spikes, etc.), excessive coating stress, excessive coating porosity, excessive P levels, excessive levels of inclusions (S, etc.)
• or/and Au coating characteristics such as insufficient coating thickness, excessive porosity, poor plating quality
• or/and the peculiarities of Au deposition process, i.e. micro-galvanic cell formation, non-uniform Au layer plating at sharp nodular boundaries, concentration cell formation on PCB due to varying in pad size, etc.
Task 2.3. Small-scale plating operations in laboratory-scale facilities with the aim to investigate a wider range of conditions and factors related to the “black pad” problem.
Electroless nickel and immersion gold coating deposition from continuously corrected solutions was undertaken with the aim of determining the influence of plating solution on the quality of deposited coatings by controlling the coating thickness, phosphorus content, porosity, deposition rate and surface morphology.
Task 2.4. Characterization of an affected PCB by identifying microstructural and chemical features that occur in problematic ENIG coatings.
The aim of the planned investigation was to establish the potential reasons for the “black pad” issues found in PCB samples presented by different suppliers. The analysis of PCB samples with and without soldered components was performed with the principal aim of establishing the root causes of very poor soldering, the presence of voids on SMT pads, or, in general, to determine the reasons and nature of the defects in PCB samples.
Task 2.5. To establish industrial design guidelines and to improve international standards.
Modified experiments to confirm the predictions of the “black pad” failure mechanisms related to inadequate copper surface preparation, deviations from plating chemistry formulation, variations in electroless nickel surface morphology and defective structure, etc. were carried out with the aim of establishing industrial design guidelines related to the “black pad” problem appearance in PCB production. Having in mind the copper pre-treatment and immersion gold solution composition relationship with the “black pad” phenomena, the aim of the investigations was to determine copper surface etching parameters and copper ion concentration limits in immersion gold solutions, which may cause the “black pad” failure.
Progress and Results
Literature analysis: According to the existing literature, the key influencing factors causing the “black pad” phenomena could be divided into the three groups. The first one related failures to EN characteristics (excessive stress or amount of inclusions, porosity, phosphorus level), the second one to gold coating characteristics (insufficient or excessive thickness, excessive porosity, poor plating quality) and the last one to peculiarities in the IG process (micro-galvanic cell formation due to non-uniform gold plating, concentration cell formation because of varying in the pad size or electric current flow in PCB during IG process). It can be stated, however, that the majority of publications were non-systematic works based on assumptions without real experimental evidence. It is widely accepted that the corrosion properties of EN deposits are governed mainly by their phosphorus content and the corresponding structural and mechanical state, however, the nature of corrosion behaviour of a nickel matrix doped with phosphorus atoms remains unclear. It was stated, as well, that the difficulties of reproducing black pads using pure chemicals was a barrier to the understanding of the phenomenon.
Task 2.1. Electroless nickel coatings deposited with phosphorus contents ranging between 3.5 and 11.0 weight percent exhibited nodular morphology. Corrosion behaviour studies of these coatings in citrate media revealed the fact that corrosion current density values (icorr) were related to P content: Electroless nickel (EN) samples with a higher P content exhibited lower icorr. It was assumed that the higher corrosion resistance of amorphous EN was due to their homogeneous structure, the absence of apparent grain boundaries and, possibly, dislocations, kink sites and other surface defects. For samples with higher P contents, local pitting corrosion sites (spikes) with very small diameters (< 1 mm) and high length-to-diameter ratios were observed, while corrosion of the samples with lower P levels occurred along the grain boundaries and this type of damage was more severe. An attempt was made to modify EN plating solutions by adding organic surfactants with the aim of facilitating hydrogen recombination, however, it can be stated that the presence of sodium lauryl sulfate in the EN bath had no influence on the EN layer morphology. It was concluded, on the basis of the results obtained, that micro porosity, roughness and inhomogeneity, e.g. due to internal stress within the coating, were all strongly dependent on the pre-treatment of the substrate.
Investigations of the immersion gold (IG) process revealed that the gold deposition rate was determined by solution pH, composition and surface reactivity of EN. A more reactive Ni surface (lower P content) led to a higher plating rate in the IG process and an increase in the amount of corrosion damage. The presence of copper in the corroded ENIG area was detected using appropriate analytical techniques. The assumption was made that the porosity of EN coatings might be the source of copper contamination and, therefore, one of the factors which influenced the quality of the IG coating and possibility of the black pad occurrence.
The following mechanism of IG layer formation has been proposed. As EN plating has a nodular structure, there are boundaries and crevices between the nodules. If a boundary or crevice is too deep and thus the supply of Au atoms to the crevice is slowed down, the Au concentration in the crevice will be different from that in the plating bath. Consequently, a galvanic cell will be set up between the crevice and the surface, resulting in heavy corrosion in the crevice. Therefore, the corrosion converts the dense, amorphous EN into a porous, micro-crystallized structure into which the Au atoms have penetrated. It can be stated that corrosion of the Ni surface is due to its activity in the immersion gold process. High IG bath pH value (pH 6), together with a high concentration of citrate (0.4 mol l-1) and the thickness of the IG layer exceeding ~ 80 nm are all parameters which favoured black pad formation. Because of the formation of a voided layer, the interfacial bonding of solder to the ENIG plating may be, by nature, weak even if the EN plating does not suffer from hyperactive corrosion.
Task 2.2. One of the intrinsic defects of EN coatings important in the corrosion process is coating porosity. Hydrogen evolution, which occurs concurrently with nickel deposition, or a rough substrate surface, may be the reasons causing pore formation. An attempt has been made to elaborate a reliable and simple method for the quantitative evaluation of the porosity of EN, IG and ENIG coatings, which is based on the voltammetric measurements in KOH solution. The studies performed disclosed significant differences in the cyclic voltammetric curves of normal EN samples and those of defective ones, indicating that cyclic voltammetry of EN and ENIG samples in KOH solution may be a useful tool for the identification of defective coatings. It must be emphasized that the source of copper contamination and the role of copper on black pad formation had not been disclosed in previous works. The method for evaluation of ENIG porosity was elaborated and it was shown in this study that EN coating porosity may be the source of copper contamination. As the “black pad” phenomena is related to excessive corrosion of EN coatings during the subsequent IG deposition process, the nature of the phenomena must be related to the specific properties of the EN coating and the IG solution characteristics. Therefore, simulated experiments to confirm the predictions of the “black pad” failure mechanisms related to inadequate copper surface preparation, deviations from plating chemistry formulation, excessive amounts of Cu and Ni in IG solutions were carried out. The behaviour of standard EN coatings in IG solution containing Cu ions was compared with the behaviour of modified EN coatings and electroplated Ni-P alloys. EN coatings deposited under experimental conditions simulating inadequate substrate preparation indicated that insufficient degreasing of the Cu surface before EN plating was detrimental to the quality of the EN and ENIG coatings. Macro-structural defects such as pores and cavities were formed on the EN surface in IG solution, leading to direct contact of the substrate (Cu) with the IG solution components. The presence of inclusions on the substrate prevents uniform distribution of the catalytic sites (Pd particles) on the surface, what may lead to the growth of deposits with higher internal stresses. Such deviations from the EN deposition parameters with increased concentration of organic stabilizers or increased pH values of EN deposition solution (and, consequently, reduced P content) led to modification of the boundaries between the nodules, which may be detrimental for the further contact of such samples with IG solution. If accumulated, Ni ions in the IG plating solution do not affect ENIG coatings quality, whereas the presence of Cu causes realization of the black pad effect.
The detailed analysis of the influence of copper revealed, that when Cu, which has no electric contact with EN coating, is immersed into an IG plating solution, formation of a yellowish black or grey deposit took place on the Cu surface and the IG solution was contaminated with Cu ions. IG coatings deposited on EN substrates from solution contaminated with Cu ions were highly porous and loosely adhering to EN substrate. Such coatings cannot protect the EN surface from oxidation and could eventually lead to “black pad” failure of solder joints. When Cu, which has electric contact with EN coating, is immersed into IG plating solution, formation of the IG coating took place on both the Cu and Ni surfaces, because the process is governed by the potential of the Ni. Though spoilage was not readily observed, it would result in delayed darkening of the gold plated copper samples because, in the absence of a barrier layer, Cu atoms will diffuse through the Au and oxidize on the surface, leading eventually to “black pad” failure of solder joints.
Task 2.3. EN and IG coating deposition from continuously corrected solutions was performed with the aim of determining the influence of plating solution exploitation peculiarities on the quality of deposited coatings. The control of coating thickness, P content, porosity, deposition rate and surface morphology was performed. The results of the small-scale operations with EN and IG deposition from the continuously corrected solutions indicated the serious influence of cysteine, which was used as the EN plating solution stabilizer, on the deposition rate and amount of phosphorus in the EN coatings and, at the same time, on its activity during the IG process. The additional introduction of the stabilizer resulted in formation of organic compound inclusions in the EN coating, which were located at the grain boundaries and may be the reason for the appearance of weak boundaries. On the basis of the results obtained it can be stated, that in order to deposit EN coatings with the required P content (10-11 wt.%), it is obligatory to maintain the concentration of the cysteine in the plating solution lower than 0.01 mmol/l.
Before starting the IG plating experiments with continuously corrected solutions it was necessary to determine the influence of the concentration of citric acid in the IG solution on the rates of EN deposition and on the accumulation of excessive Ni in the plating solution, as well on the efficiency of Au deposition. In general, all ENIG coatings deposited from IG solutions without any correction secured the required thickness (＞70 nm) of the Au layer. Cyclic voltammograms revealed formation of rather uniform and compact immersion gold coatings with porosities not exceeding 5 %. The quality of the coating did not deteriorate during the exploitation of the IG plating solution. According to the stoichiometry of the IG deposition reaction, the consumption of the Au should be 750 mg, while only 560 mg was deposited. Consequently, the efficiency of the Au deposition was 0.75 what implied excessive Ni dissolution in the IG solution.
Task 2.4. Characterization of affected PCBs with identification of the microstructural and chemical features that occurred in problematic ENIG coatings can help with the understanding of the mechanism of defect generation. The analysis of defective PCB samples may be useful also for verification of the correctness of the conclusions drawn on the basis of experiments with model systems. In the work carried out in this project, PCB samples provided by different suppliers were investigated with the aim of determining the nature of the defects arising in the samples. The surface morphology analysis of the un-soldered pads of one of the samples disclosed the fact that the surface of the substrate Cu was not pre-treated in a proper manner before EN deposition, which led to the generation of structural defects such as cracks and opening pores in the EN layer and resulted in a poor quality ENIG coating. The next investigated case with the PCB sample confirmed observations that Au layers deposited on EN substrates from IG solutions contaminated with Cu ions were highly porous and loosely adhering to the EN coating, which, in addition, undergoes serious corrosion damages. Therefore, deposition of IG from solutions contaminated with Cu ions is one of the main reasons of the “black pad” defect occurrence.
High porosity of IG layer and the presence of the intermediate layer between Au and Ni-P, which was enriched in oxygen, were the main reasons for the black pad issue in the third PCB sample examined. Since the oxidized EN surface could not be wetted with the solder alloy, the uniform and complete wetting of Ni-P surface did not occur, therefore, the strength of the subsequent solder joints was not sufficient to secure good quality soldering. In summary, the experiments undertaken on samples of problematic PCBs have confirmed the validity of the conclusions drawn on the basis of the investigations with model systems performed during the implementation of this project.
Task 2.5. The results of the studies carried out in this project indicate that the “black pad” phenomenon is related to the specific structure of the EN coatings and some peculiarities of the IG deposition process. It was established that inadequate base metal (Cu) surface preparation resulted in deposition of EN layers with nodular morphology and serious structural defects, such as pores and cavities, which eventually led to a direct contact between the substrate (Cu) and the IG solution components. In addition, it was established, that Au layers deposited in IG solutions contaminated with Cu ions, were highly porous and loosely adhering to the EN coatings. Consequently, such Au deposits cannot protect EN surfaces from oxidation.
In order to establish industrial design guidelines related to the appearance of the “black pad” problem in PCB production, it was of a great interest and importance to conduct investigations with real PCB samples. Having in mind the Cu pre-treatment and IG solution composition relationship with the “black pad” phenomena, the aim of investigations performed was to determine Cu surface etching parameters and Cu ion concentration limits in IG solutions, which may cause the “black pad” failure.
Taking into account the influence of base metal Cu etching on such parameters of the EN coatings as porosity, surface roughness and morphological features, it can be stated that the optimal etching depth is ~1.5 µm. A Cu etching depth of 0.5 m is insufficient to disclose the inner structure of the Cu, and at the same time for a high quality, non-porous EN coating production. Meanwhile, a higher than 2.0 µm etching depth causes a significant increase in the roughness of the EN coating, which might be the reason for the enhanced EN surface activity in the IG process.
In order to avoid deposition of EN coatings with a structure favorable to the excessive surface activity in IG solution, it is recommended that plating solutions are operated with pH values lower than 5.5 while the optimal is close to ~ 4.8. The concentration of stabilizer (cysteine) is another important parameter, which determines the quality of EN coatings. In order to deposit EN layers with the required P contents (10-11 wt.%), it is obligatory to maintain the concentration of the cysteine in the plating solution to lower than 0.01 mmol/l. Higher concentrations of stabilizer result in formation of EN layers with lower concentrations of P (5.7 wt. %) and increased surface activities for IG layer formation.
An IG plating solution containing 0.4 mol/l citric acid is an optimal one, as the highest Au deposition rate, the highest efficiency of the process and the lowest amount of excessive Ni dissolution in the solution were observed. IG coatings deposited from the solution, containing an admixture of copper appeared to be highly porous and loosely adhering to the underlying EN substrate mainly because of Cu incorporation into IG layer. The detrimental effect of Cu ions in IG plating solution, influencing Au layer quality, excessive thickness and aggressiveness towards EN coating, manifests itself when the Cu ion concentration exceeds 3 mg l-1. Under the mentioned conditions the IG process induces formation of serious structural defects, such as pores and cavities in the EN coating, which provide direct contact between the substrate (Cu) and the components of the IG solution and lead to further contamination of it with Cu ions.
Taking in account the differences in the pad areas in real PCBs samples, it is important to note that Au layer formation on the smaller pads was more intensive and corrosion damage of the EN layer more severe and these differences increased significantly in the case of a Cu ion presence in the IG solution.
A methodology for investigation of ENIG-related failures was proposed, which included three main steps: determination of the EN layer porosity (opening holes), measurement of the IG layer thickness and composition. Voltammetric measurements in KOH solution provide a suitable method for EN porosity and IG layer composition evaluation. The latter parameter may also be obtained by XPS analysis. Meanwhile, the IG layer thickness can be measured using an SEM and ThinFilmId software, which is a part of the Helios NanoLab dual beam workstation with an Oxford Instruments energy dispersive spectrometer.
Work Package 3 - ENIG Screening Tool
Overview
The overall aim of WP3 was the development of a non-destructive screening method which was able to identify ENIG related problems on printed circuit boards before the components are being placed and soldered. To reach this the following objective were defined:
• Evaluation of various potential non-destructive techniques for detecting ENIG problems
• Selection of a preferred technique from those assessed
• Development of a patented prototype instrument for validation and demonstration purposes
• Identification of a manufacturer for commercial model
This work package was led by TNO. To accomplish the objectives of this work package the following tasks were defined:
Task 3.1: A pre-study wouldl be conducted to investigate potentially suitable measuring principles and methods. This took ongoing input from WP2 and considered all ENIG failure modes including properties, underlying mechanisms and factors contributing to initiation. New approaches would also be determined and evaluated based on optical, electrical, chemical and mechanical measuring techniques or combinations of these.
Task 3.2: To conclude a specification of inspection criteria and system requirements such as measuring capabilities and requirements that would assure process flow compatibility.
Task 3.3: Selection of viable measurement techniques based on technical and cost implications. Development of a quality assurance plan which contained the definition of the inspection strategy concerning where in the supply chain and process flows measurement could optimally be effected and on what basis.
Task 3.4: This core task involved extensive laboratory testing of ENIG-based board structures with all variants relating to modes of deposition, methodologies of deposition, solution compositions, board layouts etc with selected measurement techniques. Board input from WPs 4 and 5 into this work package wouldbe extensive.
Task 3.5: For validation and demonstration purposes, a prototype instrument would be built. The technical validation, which included the determination of the repeatability and reproducibility of the instrument would be studied by conducting a Gauge R&R measurement systems analysis study both at TNO and at partner SME sites.
Task 3.6: The economical feasibility will be assessed by conducting a Life Cycle Cost Analysis(LCCA). This was considered important in order to enable an assessment to be made and to determine whether or not it was possible to use it as a standard test method and to identify and involve manufacturers for a commercial model.
Task 3.7: Identification of potential manufacturers for a commercial model post IPR clarification – design, specification and prototype build activity.
Results
During the first phase of this work package the work was concentrated on learning about possible black pad appearances and potential inspection techniques. This study was done based on a literature search and the latest knowledge on materials characterisation techniques. An inventory of currently available inspection techniques, which was for this study focussed on both destructive and non-destructive techniques, resulted in the deliverable 3.1. In a next step the potential inspection techniques; capacitive measurements, magnetic measurements and confocal microscopy, were studied and tested more in-depth. To learn more about the black pad phenomenon and especially in what forms it can appear, a literature study was conducted based on more than 30 papers dealing with either black pad or electroless nickel immersion gold plating defects in general. Next electroless nickel immersion gold plating defects were analysed to study appearances and learn about their detectability. Boards for these analyses were provided by ASPIS SME partners; Graphic plc, Merlin Circuits Technology and Global Interconnection Services.
From this initial work it was found that capacitive measuring principles appeared to have the most potential (based on what is currently known and understood about black pad defects). A plate capacitor was created by using printed circuit board pads as bottom electrodes together with a top electrode (+ dielectric material) which functioned as the probe of the sensor tool. An experimental research setup was built to analyse all kinds of different ENIG boards and ENIG plated samples received from ASPIS RTD partner ITRI. Significant differences in electrical capacitance were demonstrated based on samples which were made with different plating parameter settings and the use of statistical techniques. A description of the concept and the specification of the different building blocks resulted in deliverable 3.2. This deliverable also contained an evaluation of potential inspection strategies. Since the underlying mechanisms of black pad failures were still not fully understood at this stage and the known indicators and failure modes were not exclusively related to this particular defect, it was decided to develop an analysis tool that could help in understanding the defects based on manufacturing results ie a tool that could analyse boards coming from different stages of the plating process. The aim of this was to systematically build up knowledge and evidence to correlate indicators and failure modes to the black pad or other possible surfaced related defects and to detect early possible drifts or shifts in the materials, plating process or products. The SME partners of the project preferred the development of a semi-automatic benchtop version of the tool, a tool which typically could be used in a QA department. Deliverable 3.3 was a built version of such a tool and this deliverable was demonstrated on film during the mid-term meeting in Brussels. Both the use of capacitive measurements to inspect printed circuit board coatings and the anticipated integral approach, appeared to be novel and these have been registered by the inventers for a possible patent application.
In the second phase of the project the built benchtop tool was further modified and optimised based on extensive test work conducted on the test boards which came from other work packages and the boards produced by the SME partners of the project. After finalising and validating a major modification, the whole setup was shipped to SME partner Graphic plc. for a field test in a printed circuit board manufacturing environment. In the last part of the project the work was mainly focussed on understanding how the encountered variations in outcome was built up, especially the spread over time. Cost modelling work was also conducted to investigate the economic side of the different inspection approaches, concluding the TNO inputs to this work package. The technical and economic evaluations resulted in deliverable 3.4; Technical & Economic Validation (LCCA), which assessed whether or not it would be possible to use the device as a standard test method and to identify and involve manufacturers for the potential development of a commercial model that could be sold to PCB fabricators and assemblers.
Modifications
After working with the inspection setup for a certain period, it became clear that the probe head alignment procedure, the probe head design and the placement of a separate dielectric layer underneath the probing pin had to be improved in order to improve the overall performance of the testing results.
The probe head positioning and alignment were initially performed without any means. It appeared to be rather difficult to position and align the head onto mid-sized SMT solder pads. Since it was also the plan to aim for use on even smaller solder pads, it was decided to add a digital microscope camera (10 – 200 x, 2.0 M pixel) to the system which presented a magnified live view on the computer screen.
The probe head was redesigned to make it possible to place the counter-electrode closer to the probing electrode which allowed the probing of smaller pads. The modified design can probe pads of approximately 1 x 0.5 mm when using an anode probe diameter of 0.5 mm. The other reason to redesign the probing head was to change the approaching angle of the contra probing pin to a vertical movement. The vertical movement improved the repeatability and prevented pin sliding, which could cause damage to the solder pads. Spring properties were added to the counter electrode by selecting the material and designing the probe in such a way that the placement force would make the probe function in the elastic regime.
In an earlier phase of the project, different materials including paper, polyimide (PI), polyethylene naphthalate (PEN) and polycarbonate (PC) were tested as dielectric foil. However, these materials all had rather similar low dielectric constants and all appeared to be difficult to attach underneath the probing pin in a reproducible manner. Hafnium oxide (HfO2) and an acrylated urethane were investigated as alternatives. Hafnium oxide has an interesting dielectric constant of around 25 and can be applied by using CVD (chemical vapour deposition). Hafnium oxide was tested by depositing the material underneath copper probe pins. These tests were not successful because the accurately machined copper surfaces of the probe pins were attacked during one of the process steps. Tests conducted with the acrylated urethane were more successful. This material was used as a UV adhesive which could be accurately machined after curing until for instance a layer thickness of 20 μm was achieved. During the rest of the project, acrylated urethane was used as dielectric material underneath the probe pins. Repeatable result were achieved with thicknesses of 50, 42.5 and 30 μm.
The modified measuring setup was tested by measuring samples produced by ITRI (ITRI S5 coupon) and comparing the results with results measured with the earlier setup. This was done by using a multi-vari analysis and hypothesis testing (F-tests) process, which proved that the new system variation during repeated testing could meet the performance of the old setup. Stability tests over time were conducted with different probing pins having a 50, 42.5 or 30 μm thick dielectric layers. Capacitance measurements were done during the 1st measurement and after 100, 300 and 1000 times repositioning on the pad. The results were analysed by using Multi-vari analysis. The surface of the probe with a 42.5 μm thick dielectric material was also inspected by using Scanning Electron Microscope (SEM). This sample showed no visible signs of major degradation after 1000 measurements.
Within pad variation
To investigate possible within pad variations, test coupon S5 was systematically measured on different locations. Sample S5 which, was received from RTD partner ITRI, was plated with the ITRI solution plus an acetic acid buffer formulation (pH 6.0) for 12 minutes at 93.8°C. Capacitance measurements were done in a 7 x 3 matrix with a 2 mm pitch in all directions. The measured capacitance values were graphically (height plots) and statistically correlated with the amount of nickel, phosphorus and gold at the surface. The elements were measured by using Scanning Electron Microscope (SEM) with Energy Dispersive X-ray (EDX) and an acceleration voltage of 15 kV. After removal of one outlier in the EDX results, the average capacitance value was 3.6E-13 F. The within pad variation appeared to have a range of 3.4E-14 F. The measured nickel content indicated a probable to very probable negative linear correlation with the measured capacitance values. The Pearson correlation coefficient (r) was between -0.7 and -0.9. The gold content indicated a probable to very probable positive linear correlation with the measured capacitance values. The Pearson correlation coefficient (r) was between 0.8 and 0.9. The phosphorus content appeared to have no correlation with the measured capacitance values (r<0.5).
Positioning sensitivity
To investigate the probe positioning sensitivity, the S5 test coupons received from ITRI were analysed when repositioning the measuring probe around one nominal centre position. Measurements were conducted on nine positions in a 3 x 3 matrix with a pitch of 0.1 mm (n = 5). The results were analysed by a height plot and ANOVA (Analysis of Variance). The average capacitance was 3.6E-13 F with a spread of 1.1E-14 which is a spread of approximately 3 % around the average value. The height plot revealed, and ANOVA proved, that there was also an orientation effect. The variation within rows appeared to be significant while the variation between rows appeared to be insignificant. These tests were repeated on various other test boards during the course of the project.
Skip plating detection
An electroless nickel immersion gold board received from SME partner Graphic plc was analysed within work package 6 and diagnosed with skip plating. Skip plating can be described as the situation in which gold was plated directly onto the copper layer due to the absence of a nickel layer. The cause of this failure mode is often an incomplete or missing activation step. The visual appearance of skip plated pads can often be black as well. Within work package 3, this board was systematically measured to detect any deviations due to issues underneath the gold layer. It was found that footprints which exhibit skip plating might display up to 5E-13 F in capacitance variation.
Field tests at SME partner Graphic plc
From the field testing conducted at SME partner Graphic plc it was found that the capacitance values contained both a cyclic behaviour and drifts. Especially, the zero measurements revealed very clearly this unwanted behaviour. The inspection setup was installed in a climate controlled QA room with the air conditioning unit set at 20°C and 50 %RH. Data logged by the air conditioning units showed: 20.06°C (19.5 – 20.6°C), 46.4 %RH (43 – 50 %RH). An extended test was run during the night to determine whether or not the capacitance values would stabilise during the night time. This test also confirmed the cyclic behaviour and the drift over time. Next, a temperature and humidity sensor was installed on the printed circuit board table of the test setup. Correlating the outcome of these measurements with the capacitance values, it was found that the capacitance values were following the changes in the ambient humidity. This indicated that when measuring in the femto-farad capacitance range, there might be a need to operate under a more stringent controlled humidity conditions.
Work Package 4: Aqueous Plating
At the inception of the project there were two main prongs to Work Package 4. These were to try to develop improved solderable coating systems based upon the existing Electroless Nickel – Immersion Gold process (ENIG), and to develop alternatives with the same functionality, but that did not suffer from the same level of severe defects which can afflict this otherwise excellent finish.
Earlier work by various researchers and the investigations conducted as part of WP2 in the ASPIS project had highlighted a number of situations that could cause or aggravate conditions such as black pad. Armed with this information, potential changes to the materials and processes were discussed and research strategies were formulated.
The direction of the research performed in WP4 was heavily influenced, and aided, by the industrial partners, who each provided significant advice and inputs. Based upon their own knowledge, the conclusions drawn in WP2 and by the early results obtained in WP4, the industrial partners advised that they strongly recommended that the investigation should concentrate on the development of improved ENIG technologies. The reasoning behind this decision was sound. ENIG is a dual-purpose coating in that the nickel remains in the joint structure after soldering and acts as a barrier layer. The combination of good solderability and low inter-diffusion rates of copper and tin was found to be unique amongst common metals when the previous historic literature was studied. The decision to abandon further work on alternative materials and processes was further vindicated by economics. The industrial partners strongly felt that the costs associated with implementing alternative materials and processes would prevent high take-up rates. The work on alternative materials and processes was therefore abandoned, though some success was achieved using a screen printed gel method for depositing gold coatings, it being thought that this could avoid certain detrimental electrochemical effects associated with conventional immersion gold plating. A report summarising some of the alternative materials and process findings and considerations was issued as Deliverable D4.2.
Modified ENIG Considerations
The features of the failure mechanisms, in particular Black Pad, were considered and it was hypothesised how they could be avoided or suppressed. One mechanism in particular, the excessive corrosion of nickel during the Immersion Gold process, was considered. However, corrosion of the nickel is a necessary part of the IG deposition. Modifying the Immersion Gold process to reduce the risk of excessive corrosion occurring is likely to generate even more defects, as the activity is reduced and would also slow the process down.
The normal cellular morphology of the EN deposits lends itself to differential electrochemical effects due to the deep interstices between neighbouring growth cells. They also increase the risk of a through-coating pore persisting. It was therefore proposed that modifying the morphology and microstructure of the Electroless Nickel deposits could have a beneficial effect if it could be achieved. This then formed the basis of the research performed in the second half of the ASPIS project.
Plating Facilities
The first task that was tackled was to assemble and build the equipment needed to perform the proposed plating experiments. Early stage development work was performed using laboratory glassware, but this is unsuitable for scaling up to plate large numbers of even moderately-sized printed circuit boards. Methods of agitation and temperature control were tested and their performance evaluated.
When using laboratory glassware the method of heating is critical to avoid temperature differences in the plating solution. Rather than heating the glassware directly, it is better to use a bain-marie configuration. Various liquids were used as heat transfer mediums to try to achieve better temperature control and eliminate hot and cold spots, including various oils and glycerol. Whilst these did improve temperature control, there was a high risk of contamination of the plating solution and causing it to collapse. Eventually it was decided that water would have to be used, even though typical plating temperatures would approach 100oC.
It was found that mechanical agitation was not sufficient to achieve consistent results across a sample due to incomplete mixing of the plating solution and the possibility of hot spots arising. Air agitation was found to give far better results and it was shown that this method had the additional benefit of enhancing the bath stability.
A much larger prototyping facility was constructed, based around a 20 litre polypropylene tank. Air agitation was provided by PVC piping, though the original pipes had to be replaced by a more temperature-resistant grade, when it was found that they warped at temperatures above 60oC. Figure 1 shows a schematic of the tank used for the Electroless Nickel deposition part of the work.
Baseline Solutions
Standard Electroless Nickel (EN) and Immersion Gold (IG) processes were required for benchmarking purposes and to serve as starting points for investigations into improving the existing technology. The initial EN solution was identified as the “ITRI” solution and was based upon a nickel sulphate – sodium hypophosphite chemistry. The IG solution was a potassium gold cyanide-based solution which was also used for the investigative work in WP2.
Both of these solutions were used widely in the development and validation of the prototype plating facilities, for evaluations of the pre- and post-plating treatments and selection of the copper substrate preparation methods. A large number of boards using variations of these coatings were produced and used to develop evaluation methods for WP3 (screening tool) and WP6 (validation) in the project.
Figure 1. The electroless nickel tank used for development work
Substrates and Preparation
It was inconceivable that any other substrate material other than copper would be used. On PCBs, the nickel would be deposited onto electrolytically formed copper surfaces. The use of PCB samples for early stage development work was undesirable because of the size requirements and cost, so high purity copper discs were used as a substrate. Early experiments on brass and copper coupons confirmed that the coating morphology was dependent on the condition of the copper substrate surface and that the cold-rolled coupon surfaces produced very different nickel morphologies to electrolytic copper ones. An electrolytic copper process was therefore added.
It is common practice and, indeed necessary, to etch the electrolytic copper and to activate it for Electroless Nickel deposition to occur. Various etches were evaluated to determine their effect on the copper and determine whether this could in turn could have an effect on the character of the nickel deposits. The different etches did produce effects which were deemed important. An etch based upon ammonium persulphate was deemed the most suitable for the project work.
The nucleation of the nickel deposits is an important factor in the growth of the coating and its eventual morphology. The nucleation of the EN coating requires activation and so this process was studied to determine whether this could be used to influence the deposit characteristics. Activation is achieved using solutions containing palladium chloride; no other viable activating agent being known. Changing the concentration of palladium chloride had little effect on the character of the resulting nickel deposits unless the changes to the activating solution were very large. In such circumstances, the activation treatment had a detrimental effect on the deposit and it was realised that this was not a route by which a more desirable nickel morphology could be achieved.
Another way to activate the copper surface is to do so galvanically. This method would be unsuitable for an industrial PCB plating process but can be performed in the laboratory. This was done and the resulting nickel deposit was found to have a smooth, planar morphology with no cellular growth. This indicated that such growth was possible, however, a way to suppress the cellular growth would need to be found.
Growth Modification
One way to modify the growth morphology is to change the nucleation behaviour of the nickel and to then control the points at which depositing atoms can attach themselves to the growth interfaces. Substances which might absorb onto the low energy nucleation sites on the copper substrate and/or which could reside at some of the growth interfaces could cause a change in morphology. Such substances are generally known in the electroplating industry but have, to our knowledge, never been used to modify growth in Electroless Nickel systems. The substances are known as “Brighteners” and are categorised as “Class I” and “Class II”. Whilst they can be effective singly, when a Class I brightener is used alongside a Class II one, the synergistic effect can be far more effective.
“ITRI” Solution Stability
During development work it became apparent that the “ITRI” solution was very stable to changes. This was first noticed when attempts were made to change the phosphorous content of the nickel deposits and also when additions were made to simulate an out-of-control bath chemistry. This stability extended to additions of brighteners and it was deemed that the “ITRI” solution was unsuitable for modification. An alternative Electroless Nickel chemistry was therefore adopted.
The “XL” Formulation
The second EN chemistry was identified as “XL”. Its basic formulation and key operating parameters are shown in Table 1.
NiSO4.H2O 40.9 g•l-1
NaH2PO2 40.9 g•l-1
Glycolic Acid 14-20 ml•l-1
Lactic Acid 12-16 ml•l-1
Acetic Acid 20-25 ml•l-1
Lead as metal 3.5 g•l-1
pH (preferred) 5.0
Temperature 90 – 95oC
Table 1. The formulation and operating conditions of the “XL” solution
With the XL solution, it was possible to modify the chemistry and plating conditions. Various packages consisting of different combinations of Class I and Class II brighteners were added to the basic formulation and the effects assessed. It was found that saccharin was the best Class I brightener. The best results were obtained with a Class II Brightener of either propargyl alcohol or phthalimide. Details of the recommended brightener packages are shown in Table 2.
Package 1 Package 2
Class I 2 x 10-2M Saccharin 2 x 10-2M Saccharin
Class II 1 x 10-3M Propargyl Alcohol 1 x 10-3M Phthalimide
Table 2. The Brightener packages found to be most effective at modifying the EN deposit morphology
Conventional Coating Morphologies
The brighteners added to the “XL” formulation reduced or suppressed the development of the cellular structures that formed during conventional Electroless Nickel deposition. The conventional deposits consist of neighbouring columnar structures, each with a rounded top. This forms a valley between the cell and its neighbours and these have been shown to sometimes contain deep crevices that can penetrate through the entire thickness of the nickel coating. This results in coating porosity which degrades solderability and can also allow copper migration; a secondary failure mechanism. The cellular microstructure can also affect the nickel corrosion that occurs during the immersion gold process. The structures could allow concentration cell effects to arise and the intercellular composition of the nickel-phosphorous coatings can vary, thus meaning the centre of the cells have different resistance to the corrosion mechanism to the intercellular regions.
Figure 2. A conventional ENIG coatings showing deep interstices in the intercellular regions
Modified Electroless Nickel Morphology
The use of brighteners changes the morphology of the Electroless Nickel deposits. In some cases, especially in thin coatings, the effect is to reduce the cellular character of the deposits. In such cases it is possible to make out features that appear to resemble cellular growth, but the intercellular regions do not contain deep valleys or interstices. Thicker coatings can entirely suppress the cellular growth leaving no trace of it on the outer surface. The modified EN deposits can be seen to consist of very fine equiaxed grains which have been determined to be amorphous.
Figure 3. Modified EN coatings showing residual cellular features in a 6 m thick coating
The benefits of this microstructure are that there are no deep crevices or interstices which should reduce the risk of differential electrochemical behaviour in different regions across the surface of the nickel during Immersion Gold deposition. Also, in the absence of these deep penetrating boundaries there is a much reduced risk of there being through-coating pores and so failures associated with such defects are reduced too. Coatings have been produced that do contain pores, but these are very obvious to detect and so the issue would be picked up quickly in a production environment.
The lack of intercellular boundaries may have additional benefits. The minimum EN thickness in an ENIG solderable coating, as specified by the IPC, is 3 m. In part, this is derived from the minimum growth required to ensure that neighbouring growth cells in conventional EN converge and meet, thereby closing any intercellular pores. Using the modified EN process, as described here, that level of growth is not required. It may therefore be possible to use a thinner nickel coating, thereby saving materials costs and reducing the plating time.
WP3 and WP6 Support
As part of the WP4 work programme, a large number of samples were produced to assist the activities of WP3 and WP6. These samples encompassed all of the different stages of the work performed in WP4, including the examples of modified EN using the brightener packages detailed above. In all cases, many different variables were represented in the samples, including coating thickness, phosphorus content and other operator determined variables. Unfortunately, the development of the latest generation of modified boards occurred towards the end of the project as a whole and the long-term tests such as storage life retention and thermo-mechanical reliability have not yet been completed at the time of the preparation of this report. These activities arel continuing at ITRI after the official conclusion of the project, as it is believed that the benefits of the modified EN coatings, which can be produced with minimal changes to existing facilities, may make them attractive to PCB fabricators.
Work Package 5: Ionic Liquid Plating
Introduction
The performance of printed circuit board (PCB) solderable finishes is one of the principal factors affecting the reliability of most electronic devices. These surfaces act as a reliable, solderable surface for forming the connections to the various components that need to be attached to the PCB. The most popular choice for high value electronics is the electroless nickel immersion gold (ENIG) surface coating. This is because the ENIG finish offers excellent solderability that is maintained during prolonged storage times. However, there is one recognised problem that sometimes occurs with ENIG coatings and this is known as “black pad”. Despite being identified more than ten years ago the mechanisms that cause this phenomenon were still poorly understood prior to the work undertaken in the Aspis project. Black pad manifests itself due to excessive corrosion of the nickel coating during the immersion gold process. The main problem with this is that the surface of the coating is still pristine and identification of defective boards is not possible by visual inspection. When the boards are assembled with electronic components, the presence of black pad can result in a subsequent brittle fracture of the resulting solder joints and thus devices failure. The ASPIS consortium was formed to investigate the mechanism of black pad formation, development of a testing procedure to identify black pad and finding alternative solution chemistries for ENIG deposition such that black pad could be eliminated.
WP5 focused on the design of ENIG coatings obtained partially or wholly from alternative solvents based on ionic liquids. Ionic liquids are liquids that are composed exclusively of ions and they possess many unusual properties such as high solubilities of metal salts including metal oxides. Electrolytic, electroless and immersion deposits from ionic liquids can have considerably different properties when compared to their aqueous-based counterparts.
In order to structure the research in this work package, a series of tasks were followed:
• Development of electroless nickel (EN) and immersion gold (IG) formulations from ionic liquids
• Fundamental characterisation of metal/metal interfaces
• Evaluation of new ENIG coatings for the PCB industry using standard solderability/wetting balance evaluations.
The investigations of electroless nickel coatings contains two essential steps; the generation of a palladium activated Ni surface, which is then subsequently followed by the EN deposition proper.
Development of Electroless Ni Process from Ionic Liquids
Bare copper is not in itself sufficient for electroless nickel plating, a highly adsorbing surface such as Pd is required to catalyse the reaction between the reducing agent and the metal salt during the initial stages of the deposition process. In the conventional aqueous process, this is obtained by immersing the copper-clad printed circuit board material into a bath containing a Pd salt. Small particles of Pd are then immersion deposited onto the surface of the copper, which act as seeds in the electroless Ni process.
Methods of replicating this approach were attempted from ionic liquids instead of aqueous media. A study of immersion coatings of Pd onto copper sheet from ILs was undertaken using PdCl2 as the Pd source in a 0.1 M solution in Ethaline 200 for 20 minutes at 50ᵒC. Pd was deposited as a grey metallic coating with good overall coverage of the Cu surface. EDAX analysis showed the presence of Pd on the surface of the Cu, with Cu as the majority element of 87.1 wt% Cu compared to 12.9 wt% Pd, showing that the Pd coating was very thin. AFM analysis confirmed that this coating was approximately 250 nm thick.
In addition to using an analogous Pd process from ionic liquids, an alternative Ni activation process was also developed. This took advantage of the fact that the standard reduction potentials for Cu and Ni are switched when compared to the aqueous process. In Ethaline 200, Ni is more noble than Cu and, as a consequence, an immersion coating of Ni on Cu is obtained when a Cu PCB token is immersed in a 0.1 M solution of NiCl2 in Ethaline 200 at temperatures above 100ᵒC. Quartz crystal microbalance analysis showed that there was a net loss in mass on the surface as the Cu was oxidised to Cu+, whereas Ni metal was reduced from Ni2+. As a consequence, there was a net loss of mass from the surface. SEM analysis showed that the Ni deposited as small particles onto the Cu surface. Both of these Pd and Ni activated samples then successfully had EN deposited onto them from an aqueous EN bath. They behaved very much like those of the standard Pd deposit with comparable deposition rates and compositions.
The next logical step from this point was the development of an electroless nickel bath from an ionic liquid. The use of hypophosphite as the reducing agent was key to the final properties of the coating. When hypophosphite was used as the reducing agent for Ni, a substantial amount of phosphorous was incorporated into the deposit causing a change in the structure and behaviour of the deposit. Above P levels of 5 wt%, the structure became amorphous and with increasing P concentrations the deposit became more corrosion resistant.
When NiCl2 and Na2PH2O2 were heated in the ionic liquid Ethaline 200, a spontaneous, particulate deposit started to form, which EDAX analysis showed to be Ni. The turbidity of the Ni solution had been caused by the reduction of the Ni(II) ions in situ by the hypophosphite reducing agent causing the formation of black Ni metal particles. For any electroless process it is essential that reduction of the metal occurs only at an activated surface, otherwise an extremely large amount of metal waste occurs. In this case, the additional aspect of the IL inhibiting the Ni plating produced even greater obstacles to overcome. Various attempts were made to stabilise the plating bath; the addition of Pb2+ ions to the plating solution, reduction of the concentration of reducing agent and Ni salt, plating at lower temperature, addition of water to the plating solution and the addition of an acid source to change the pH of the liquid. Unfortunately, none of this made an impact, these solutions either did not plate onto an activated Cu surface, the solution destabilised, or both. Full details of the experimental work and efforts regarding the development of electroless Ni coatings onto Cu from ionic liquids have been documented in Deliverable 5.3: Catalytic Deposition Methodology.
It was at this stage, with the given inherent problems of developing a suitable electroless Ni plating solution and the success that had been demonstrated when studying immersion gold deposition onto aqueous electroless Ni surfaces that it was agreed with the Aspis partners that the focus of the ongoing efforts should be on the study and development of an immersion Au process from ionic liquids.
Immersion Au deposition from Ionic Liquids onto nickel
The conventional aqueous IG process used acidic media and potassium gold dicyanide as the Au source. Initial efforts were focused on IG coatings from the IL Ethaline 200 onto aqueous EN with the Au salts AuCl, AuCN and KAu(CN)2. The coatings from these three systems were substantially varied. Visually, AuCl produced a dark brown poorly adherent coating, AuCN gave a bright gold coloured finish and KAu(CN)2 gave a slightly duller Au coloured coating. Surface SEM analysis of these three samples showed that two of them produced coatings with very little surface structure, whereas 1 showed a rough particulate surface. This same effect was seen with AFM images recorded across a 2 µm2 footprint. One sample had a surface area difference of 180%, whereas both the other two had surface area differences of 1.9% and 2.65% respectively. The marked difference between these samples showed the dramatic influence that the Au salt used can have on the resulting Au deposit morphology.
Cross section analysis of samples showed that a uniform Au layer was formed on top of the Ni/P layer and, contrary to the conventional aqueous process, there was very limited attack down Ni/P nodule boundaries. The first of the three samples examined showed a dendritic, particulate nature.
Removal of the Au with a cyanide Au stripping solution revealed information regarding the impact on the Ni/P substrate of the IL IG process. In all cases no evidence of the mud cracking associated with black pad was seen. A standard, uniform, nodular EN surface was seen across the whole sample.
Solder wetting balance measurements on two of the three samples showed that the wetting time for the IL samples was both faster than a standard aqueous sample and that they were wetted more reliably. This is a key measure of the solderability of these surface finishes and provides evidence that these coatings could perform as well as, if not better, than conventional aqueous coatings.
Use of Additives in Immersion Au Process from Ionic Liquids
The use of additives in the ionic liquid IG bath was also investigated. The purpose of these additives was to affect the electrochemical behaviour of Ni in the selected ionic liquids. The addition of ethylenediamine to an ionic liquid causes the rate of electrochemical dissolution of Ni to increase dramatically. In addition, a quartz crystal microbalance (QCM) study showed that the addition of 10 mM ethylenediamine to one of the IG bath formulations caused the rate of gold coating to increase dramatically as well. (This will be referred to as solution/sample four below).
IG coatings from sample 4 produced a bright lustrous coating. SEM analysis showed that the coating was uniform across the EN surface with very little defects. AFM analysis also showed that the coating had low surface area difference. Stripping of the Au surface revealed that the underlying Ni/P nodules had been levelled as indicated by a reduction in the value of Ra with respect to the original virgin Ni/P surface. This is extremely exciting as the formation of black pad is thought to originate along nodule boundaries. These results confirmed that the Ni was consumed preferentially on top of the Ni nodules, potentially eliminating the mechanism by which black pad forms.
AuCl is not stable with respect to ethylenediamine and as such an alternative additive is required that does not oxidise/reduce AuCl and that activates the Ni surface. Iodine was found to do this effectively depending on concentration. A number of different concentrations of I2 were added to an AuCl solution, ranging from 0.125 mM to 5 mM and ENIG samples were produced from these respective solutions. It was found that an evolution in morphology was observed with increasing I2 concentration. At low concentrations of I2 a similar morphology to that of an earlier sample was observed, where it was rough and particulate in nature. As the concentration increased, the surface roughness decreased and particle sizes were smaller until between concentrations of 0.5 – 2 mM I2 a smooth uniform deposit was produced. Above 2 mM I2 an alternative morphology was observed where the Au was deposited as small nodules and some of the EN surface was visible when viewed at high magnification on an SEM. A fifth key formulation was developed of 5 mM AuCl and 1 mM I2 in Ethaline 200.
Ionic liquids have potential as a replacement for the aqueous immersion Au bath in ENIG coating production. A thorough investigation has been undertaken of the various factors that affect the resulting immersion Au deposit morphology and the resulting impact on the underlying electroless Ni layer.
An evolution in morphology was observed with increasing cyanide content in the Au salt used in deposition. Dramatically improved deposits were obtained when an at least one molar equivalent of cyanide to gold was used. In addition, the influence of additives also proved to be vitally important. A number of amine containing compounds were found to improve the otherwise sluggish kinetics of Ni dissolution in ionic liquids leading to the use of ethylendiamine, which produced quicker deposition rates as well as improved ENIG coating morphology. I2 was found to be an effective additive into AuCl solutions, activating the Ni/P substrate while stabilising the AuCl solution in Ethaline 200.
Work Package 6: Validation
The overall aim of work package 6 concerned the validation of improved electroless nickel immersion gold (ENIG) Printed Circuit Board (PCB) coatings and possible alternative coatings for high-end electronic applications. To reach this the following objective were defined:
• Verify the compatibility of improved and alternative coating technologies with soldering assembly methods and practices by assessing yields in lead-free reflow and related soldering operations
• Establish effective storage life and solderability retention characteristics
• Determine long-term reliability by evaluation in normal and hostile operating environments using thermal cycling and related appropriate accelerated testing conditions.
• Validate performance in production trials at partner facilities
During the first phase of this work package the work concentrated on validating the solderability and solderability retention of various printed circuit board finishes applied by the project partners. Other work during this period concerned the definition, building and evaluation of a reliability test setup based on the combination of thermal cycling and vibration testing. During the second phase of the project, newly developed alternative electroless nickel immersion gold finishes were characterised and tested on solderability and the reliability aspects like intermetallic growth and high temperature humidity storage. Failure analysis, to study the black pad failure mode and black pad related failure modes, was a significant part of the work and a continuous activity throughout the whole project phase. All the printed circuit boards for these failure analyses were provided by the Aspis project’s SME printed circuit board manufacturers.
Solderability
The first series of solderability assessments were conducted on as-plated and (accelerated) aged Printed Circuit Board (PCB) samples, which were fabricated at SME project partner Merlin Circuits Technology. The board designs were according to the Atotech Test Board specification ATB v1.1. Boards with standard ENIG (57 nm Au), thin gold (39 nm Au), thick gold (79 nm Au) and ITRI ENIG (70 nm Au) were tested regarding solderability with a MUST-2+ wetting balance (globule method). Solderability testing was done before and after aging at 50°C and 50% RH for 24 hours. The standard ENIG, thin gold and thick gold boards showed unanticipated poor wetting results (large wetting angles and incomplete pad coverage). The measured wetting curves showed unacceptable wetting times and wetting forces, and a large variation between the samples. The ITRI ENIG samples based on a formulation developed within work package 4, showed proper wetting results (wetting angles < 90°, full pad coverage). Nevertheless, 6 out of the 10 wetting curves had wetting times longer than 2 s (up to 12 s). The second series of solderability investigations was conducted on samples from work package 4 (aqueous plated) and work package 5 (ionic liquid plated). The solderabilities of both aqueous and ionic liquid plated samples were found to be good. Aging at 50°C and 50% RH were found to have no noticeable effect on the solderability of samples even after exposures of up to 28 days.
Furthermore, various methods were applied to ENIG-plated circuit boards to determine if it was possible to:
1) Identify problematic coatings immediately after fabrication
2) Produce accelerated ageing effects whereby the useable storage life of ENIG-coated PCBs could be predicted.
As has been mentioned above, coatings developed in WP4 and WP5 of the ASPIS project were found to have experienced no noticeable deterioration in solderability after 28 days of storage at 50°C and 50%RH. Other potential artificial ageing conditions were investigated. These included:
1) Steam Ageing
2) Nitric acid vapour exposure
3) Hydrochloric acid vapour exposure
4) Polysulfide solution treatment
5) Hydrogen sulphide exposure
Conditions 2 to 5 were found to be useful for identifying coating defects, in particular coating porosity and deep intercellular interstices in conventional ENIG coatings. Hydrogen sulphide was especially effective at identifying sub-standard Immersion Gold coatings which rapidly discoloured in regions in which there was insufficient gold. The easiest and most effective test for identifying isolated pores in the gold coating was hydrochloric acid exposure for 2 to 5 minutes. This caused coloured crystals of nickel chloride, copper chloride, or both, to grow on the surface of the affected coating. The polysulfide treatment was the best test for assessing porosity of all types in the nickel coating. However, all four methods were found to be ineffective at accelerating solderability degeneration in a manner that could be used to predict the storage life of good ENIG coatings. Deterioration was either instantaneous, especially for coatings containing defects, or non-existent when high performance gold coatings were exposed.
Steam ageing was found to cause a gradual degradation in the solderability of ENIG coatings as the exposure time increased. The rate of deterioration was very variable. This is almost certainly due to variability in the quality of the ENIG coatings being tested. Standard steam ageing test methods specify exposure times of up to 8 hours. PCBs with thin gold coatings were found to have very short exposure resilience, almost certainly due to the patchiness of some coatings allowing large areas of the underlying nickel to passivate. Gold coatings that had isolated porosity were found to be more resilient. The pad areas directly below the pads deteriorated, but this usually left a large enough area on each pad that could be wet by molten solder. High quality gold coatings had the best solderability retention, but there was still a noticeable deterioration in all coatings after 8 hours of steam ageing. The later modified EN coatings in which the normal cellular microstructure was suppressed were found to have excellent solderability retention during steam ageing treatment.
Reliability
A combined thermal cycling and vibration test with in situ monitoring facilities was developed with the aim to detect solder joint degradation based on the electrical resistance. The development of an evaluation protocol led to the specification of analysis techniques and equipment to be used. Test specifications were based on the selected high-end electronics applications areas ie automotive and aerospace. Vibration tests were carried out to evaluate both the setup and the test board design (ATB v1.1). Test results were obtain from soldered 0603 chip resistors, 0805 chip resistors, Mini-MELF resistors, MELF resistors, 1206 chip resistors and Mini-MELF resistors. These component types were subjected to random vibrations with a maximum frequency of 2 kHz, while the daisy-chain circuit resistances were measured.
During the second phase of the project a new series of Atotech test boards was subjected to high temperature and humidity tests and high temperature aging tests. The Atotech test boards according to specification ATB v1.1 were manufactured at printed circuit board manufacturer and SME project partner Merlin Circuits Technology. These boards were manufactured and shipped to RTD project partner ITRI without solderable solder finishes. At ITRI, the test boards with copper were metallised with the newly developed Standard ITRI and XL electroless nickel immersion gold platings. These plating types were developed by ITRI within work package 4. The electroless nickel / immersion gold plated boards were visually inspected regarding plating quality and for possible plating defects by using optical microscopy.
The thickness of the standard ITRI and XL electroless nickel immersion / gold platings were measured by using X-Ray Fluorescence (XRF). The electroless nickel plating of the XL boards, which were shipped to TNO for further testing, had an average thickness of 5.883 µm with a standard deviation of 1.074 µm (n=13). The immersion gold plating of these boards had an average thickness of 0.095 µm with a standard deviation of 0.022 µm. The standard ITRI plated boards shipped to TNO had an average electroless nickel thickness of 6.710 µm with a standard deviation of 0.812 µm (n=19). The immersion gold plating of these boards had an average thickness of 0.084 µm with a standard deviation of 0.013 µm. The industry standard “IPC-4552 Amendment 1 Specification for Electroless Nickel / Immersion Gold (ENIG) Plating for Printed Circuit Boards” specifies 3 to 6 μm [118.1 to 236.2 μin] for electroless nickel and min. 0.05 μm [1.97 μin] for immersion gold. With respect to this specification, only the XL electroless nickel thickness was slightly above specification. A thinner nickel layer might, regarding the black pad failure mode, probably be more critical, since these layers could become porous and possibly induce hyper-corrosion during immersion gold plating. A thicker electroless nickel layer might, on the other hand, contain higher stresses inside the layer, which could subsequently result in cracking.
In order to further characterise the electroless nickel immersion / gold platings of both the standard ITRI and XL test boards, capacitance measurements were conducted using the inspection tool setup developed within work package 3. The outputs of these capacitance measurements were both graphically and statistically analysed and correlated together with the thickness measurements. A Multi-vari analyses was used to investigate the variation within samples and between samples / over time. These analyses revealed already quite a spread within samples and over time. Correlating the capacitances with the XRF measured layer thicknesses showed both graphically and statically that there was no correlation between the two parameters (r<0.5).
Temperature-humidity tests were conducted on bare boards according the JEDEC 22 A110 standard. Both boards with standard ITRI and XL electroless nickel immersion / gold plating were subjected to 85°C and 85% relative humidity conditions. Samples were taken at t = 0, 250, 500, 750 and 1000 hours. The test was used as an accelerated high temperature humidity storage test. No degradation or defects were found during visual inspections done by using optical microscopy. Furthermore, samples were analysed using Scanning Electron Microscopy (SEM) - Energy Dispersive X-ray (EDX). Elemental analysis was used to investigate whether or not corrosion products were present at the top surface of the electroless nickel immersion / gold layers. Since nickel and not gold is expected to oxidise, oxygen nickel ratios were calculated and compared between the different samples. The EDX element analysis showed, as expected, that nickel, phosphorus and gold were the main elements on the standard ITRI plated test boards, this next to a small amount of oxygen. Also, the standard ITRI sample taken at 1000 hours contained the same elements. The oxygen nickel ratio on the standard ITRI samples increased from 0.012 at t = 0 hours to 0.025 at t = 1000 hours, which meant an increase of 1.3 %. The EDX element analysis on the XL plated test boards showed next to nickel, phosphorus, gold and oxygen, also a certain amount of copper. EDX analyses at different acceleration voltages confirmed that the copper was present at the top surface of the plated layers. Further analysis on sample 36B from the XL series revealed that Cu could be particularly found in the pin holes which were present in the plating layer. The XL sample taken at t = 1000 hours showed the same elements as found on the t = 0 sample. The oxygen nickel ratio on the XL plated samples increased from 0.020 at t = 0 hours to 0.036 at t = 1000 hours, which meant an increase of 1.6 %, and indicated an increase in oxygen nickel ratio comparable to the standard ITRI plated boards.
High temperature aging tests according to the JEDEC Standard 22-A103C Condition B were conducted to study the growth and possible degradation of intermetallic interfaces. Samples were prepared by reflow soldering SnAgCu spheres to the Ball Grid Array (BGA) pads of both Standard ITRI and XL plated test boards. The boards were soldered using a Kester TSF Tacky Flux (type ROL0) and a heating plate at 260°C. The TAL (Time above Liquidus) was 60 seconds. The visually inspected joints appeared to be free from soldering defects and showed proper wetting towards both the Standard ITRI and XL finishes. After that the soldered test boards were aged at 150°C, samples were taken at t = 0, 20, 100, 250 and 500 hours. Cross-sectioning was performed to study the solder joints and interfaces by using a light microscope and Scanning Electron Microscopy (SEM) - Energy Dispersive X-ray (EDX). Some joints exhibited small cracks in the solder surface due to shrinkage during cooling down. A number of boards also showed voids at the interface at a rather low and acceptable voiding rate. On nearly all samples, the intermetallic layer was present over the whole length of the interface. Only a number of voids seems to have locally prevented wetting and subsequently the formation of an intermetallic layer. Voiding did occur on both the Standard ITRI and the XL plated test board. The intermetallic layers on both board types showed an intermetallic layer that consisted of a thin uniform layer with a scallop structure on top. Visual inspections could not detect any irregularities in intermetallic growth or the introduction of defects. Thickness measurements (n = 20) were conducted to measure the uniform intermetallic layer thickness (tu), the scallop intermetallic layer thickness (ts) and calculated the effective intermetallic layer thickness (teff). The outcomes of these measurements were both graphically and statistically analysed. An ANOVA (Analysis of Variance) revealed that there were significant differences (α = 0.05) in both the uniform intermetallic layer thickness (tu) over time and the scallop intermetallic layer thickness (ts) over time. Furthermore, there was no evidence for intermetallic growth differences between the Standard ITRI and XL plated boards.
The robustness of the XL plating processes was investigated by attempting to deliberately induce the black pad failure mode. Attempts were made to induce hyper corrosion by systematically reducing the thickness of the electroless nickel layer. This was with the aim of creating a porous nickel layer, layers which tend to be more prone to hyper-corrosion during immersion gold plating. Soldering tests showed that even the boards with the lowest nickel layer thicknesses (1 µm), did not result in black pad failures. A series of solder joints did show deviations in wetting behaviour, but no de-wetting or wetting behaviour that could be linked to black pad. Cross-sections of these thin nickel layer joints did not show the typical corrosion spikes, mud cracks or a phosphorus enriched black band along the interface.
Failure analysis
To study the black pad failure mode and black pad related failure modes, failure analyses were conducted on both bare and assembled printed circuit boards. The boards analysed within work package 6 were provided by Aspis SME companies Merlin Circuit Technology, Graphic plc, Global Interconnection Services and Liad.
The electroless nickel / immersion gold plated printed circuit boards provided by the Dutch SME company Liad failed in the field on electrical failures and sometimes had a low drop impact resistance. Failure analysis was conducted using microscopic analyses as well as Scanning Electron Microscopy (SEM) - Energy Dispersive X-ray (EDX). Bare boards, cross-sections of both failed and intact joints and the fractured surfaces were analysed. A Ball Grid Array (BGA) on the board appeared to be the critical location exhibiting cracks at the board side. The fracture occurred between Ni-P and the intermetallic layer, which made it possible to analyse the phosphorus content at the surface, which appeared to be around 15 wt.%. Cross-sections showed the typical corrosion spikes at the topside of the nickel layer and at some joints the phosphorus-enriched black band as well. Furthermore, copper was again detected at the surface of pads on a bare board. Still not understood are the presence of aluminium and silicon at some locations.
The failed electroless nickel / immersion gold plated printed circuit boards provided by Graphic plc contained at some footprints a number of black pads. Scanning Electron Microscopy (SEM) - Energy Dispersive X-ray (EDX) was used to study both the morphology and composition of black, greyish and normal bright looking solder pads. The EDX elemental analyses proved that nickel was missing underneath the black and greyish pads. The strongly deviating morphology at the black pads appeared to be a result of copper being directly attacked during the plating processes. In both the literature and the industry this failure mode is called skipped plating.
The electroless nickel / immersion gold plated printed circuit boards provided by Global Interconnection Services gave problems during pull tests conducted after the wire bonding operation. Examination with Scanning Electron Microscopy (SEM) showed that the electroless nickel / immersion gold plating exhibited severe cracks at the top surface along the grain boundaries. Cross-sectioning showed that these cracks were already present in the nickel layer before plating the gold layer. On certain locations, the cracks were reaching to the copper layer. These severe discontinuations in the nickel layer might probably have influenced the ultrasonic energy distribution during the wire bonding process.
The first electroless nickel / immersion gold plated printed circuit board provided by Merlin Circuit Technology came from the field and exhibited pads with non- / de-wetted areas. Cross-sectioning together with light microscopy and Scanning Electron Microscopy (SEM) - Energy Dispersive X-ray (EDX) were used to systematically analyse properly, half and non- / de-wetted areas. These conditions were all present on the same printed circuit assembly, sometimes even on neighbouring pads. This analysis revealed that non- / de-wetting could be related to the copper condition underneath the nickel layer. A height difference in the copper layer morphology caused stresses and a height difference in the nickel layer. This particular location determined the boundary to where the pad could be wetted. The remaining part of the pad had the typical black corroded nickel appearance.
The second electroless nickel / immersion gold plated printed circuit board provided by Merlin Circuits Technology came from the printed circuit board manufacturing process. This board was half stripped regarding the immersion gold layer. The stripped pads were black / greyish coloured. Cross-sectioning together with light microscopy and Scanning Electron Microscopy (SEM) - Energy Dispersive X-ray (EDX) were used to systematically analyse a SO (Small Outline) footprint which was half stripped. These analyses revealed that the results of excessive nickel corrosion were present underneath the immersion gold layer. The nickel layer itself had a thickness which exceeded the common maximum industry specification of 6 µm. Cracking together with excessive nickel corrosion was found right at the corner of a pad, a location with probably a concentration of the highest stresses.
The third electroless nickel / immersion gold plated printed circuit board provided by Merlin Circuits Technology came from the field with an accompanying failure report. Since for this board a de-wetting issue with a defective rate of 44 % and an Au layer thickness above specification were mentioned in the problem statement, it was decided to: 1) analyse the pad metallisation thicknesses by using XRF, 2) conduct soldering experiments using different times above liquidus (TAL) and 3) conduct SEM analyses on cross-sectioned solder joints.
Potential Impact:
Introduction
The potential impact of the work carried out in the Aspis project varies between work packages and is detailed for each one below. It is clear that the fundamental mechanistic work undertaken by LIOC has, for the first time, revealed the key causes of black pad and related reliability issues in ENIG coatings. The use of this knowledge will enable European PCB producers to adjust their processes to avoid conditions likely to cause the problem. The demonstration of a non-destructive test methodology and equipment also offers potential for the commercialisation of a device that could be used by PCB fabricators to confirm that the printed circuit boards they are producing are free of the interfacial integrity problems that are a symptom of potential reliability problems. The formulation of a range of novel gold deposition chemistries based on the use of ionic liquids that can give enhanced performance over conventional approaches, also offers the potential for a new commercially viable alternative to the existing aqueous based chemistries. The work to develop alternative aqueous chemistries with enhanced performance proved to be more difficult and confirmed that, as had been found by existing suppliers of such products, that the development of a new aqueous ENIG processes that didn’t exhibit reliability problems was extremely difficult and at the end of the project further work was still needed in this area. At the time of writing this work was showing some very positive initial results and further performance testing was continuing. It also has to be reported that in the five year period from when the Aspis project proposal was first developed to the recent completion, there has been a move by the PCB industry away from the use of ENIG finishes to other new types of solderable coatings which do not exhibit the typical ENIG problems. This was largely due to the fact that the industry’s chemistry suppliers were unable to successfully solve the black pad issue.
Impact WP2: Fundamental Research on ENIG Mechanisms (LIOC)
The specificity of the “black pad” problem is that the incidence of it is very sporadic and its occurrence is very unpredictable, making it a very difficult problem to address. However, the major advances in understanding achieved as a result of the fundamental research work carried out on the mechanisms of black pad formation in ENIG coating (WP 2) promise to have major implications for both the PCB manufacturing industry and the project’s RTD providers.
One of the most important outcomes of the Aspis work was the procuration and extension of knowledge in the area of hyper-corrosion of EN coatings in IG deposition solutions during Au layer formation. The results of the fundamental research into ENIG failure mechanisms, which was carried out during the implementation of ASPIS project objectives, indicated that the “black pad” phenomenon was related to the specific structure of the EN coatings and some peculiarities associated with the IG deposition process. It has been established that inadequate preparation of the base metal (Cu) surface results in deposition of EN layers with nodular morphology containing serious structural defects, such as pores and cavities, which further leads to a direct contact between the substrate (Cu) and the IG solution. The source of Cu contamination and the role of Cu in the formation of black pad has been idenitifed and disclosed for the first time. It has also been demonstrated that Au layers, deposited in IG solutions contaminated with Cu ions, are highly porous and loosely adhering to the underlying EN coatings. Consequently, such Au deposits cannot adequately protect EN surfaces from oxidation and could eventually lead to “black pad” failure of solder joints. Fundamental knowledge developed in WP 2 of the Aspis project has enabled to negated the widespread point of view that it is impossible to reproduce the “black pad” effect in the laboratory applying pure chemicals. Another key development that will be of use to the PCB industry was the elaboration of a new method for determining the porosity of EN and IG coatings. The results obtained will also help to enable and guide the focus of more profound research works in this field in the future.
The fundamental mechanistic knowledge gained in WP 2 was also useful to, and applied by, the project’s RTD partners in prosecution of their tasks in creating the ENIG screening tool and also for developing alternative ENIG coating technologies.
The knowledge data and experience gained and accumulated in WP 2 also enabled the establishment of industrial design guidelines for the implementation of processes that are able to minimise or avoid the likelihood of black pad formation. Specifically, these guidelines provide advice which regulates the base metal copper preparation process before EN deposition, including the etching depth, EN and IG deposition process parameters, including the IG solution tolerance to the presence of copper ions. The latter findings will help manufacturers to minimize the risk of ENIG-related failure problems occurring in their printed circuit boards. A methodology for investigation of ENIG related problems has been elaborated, which includes the use of a voltammetric examination of the porosity of EN and IG layers, measurement of IG layer composition and thickness, chemical analysis of the boundary layer between IG and EN. A suite of techniques suitable for the mentioned analysis has been developed and proposed, including SEM and TEM with EDS application, as well XPS with ion sputtering capability. The implementation of both structural and analytical examination approaches provides the possibility to proceed with the examination of defective PCB samples directly in a PCB’s manufacturer’s site, and not just in an external RTD institution. The implementation of this methodology and analysis of the results should lead to superior PCB production processes, elimination or at least a significant reduction in the occurrence of black pad, as well as enhanced reliability performance in assembled printed circuit boards, increased confidence in the use of ENIG as a high performance solderable finish and reduced field failures in service. The methods developed will also accelerate the outcomes of ongoing fundamental investigations in this area and will enable the process chemical suppliers to confidently develop improved performance coatings.
The major output from the work performed in WP2 was a methodology for the minimization or avoidance of the risk related to ENIG reliability problems such as black pad. As PCB reliability problems are of a great importance and actuality to all of the PCB manufacturing industry, the outcomes of the project will lead to a reduction of “black pad” cases in the electronics industry, as well to superior production yields, higher quality and better reliability performance of its products. The deeper and enhanced understanding of the principal causes of ENIG failure will enable the project’s SME partners to implement design rules in order to lower the risk and, in general, to avoid the appearance of the “black pad” phenomenon in their manufacturing practices.
The particular work was done with characterization of affected PCB’s supplied by project’s SME PCB fabricator partners, which enabled information to be gained regarding the nature of the problematic ENIG coatings, as well a comparison of the advantages and disadvantages of different ENIG deposition technologies and the resolution of some problems experienced by the PCB provides. The results from the analysis of problematic PCB samples have validated the fundamental knowledge developed in WP2, and will allow for RTD performers to intelligently target improvements in any future related work that they undertake.
One of the additional outcomes from the work package 2 activities is the gained understanding of how complicated the problem of “black pad” actually is and how many different factors can influence it. The results obtained in this work will be exploited for the future fundamental studies related to ENIG problem. The potential areas of the future fundamental work in ENIG-related problems are the following: deposition of alternative EN coatings with nodular-less structure, the search of new copper activation technologies before EN deposition, ie that exclude the use of an expensive Pd salt application step, new technologies of thinner and, at the same time, less porous gold layer deposition. The latter two technologies could lead to a significant reduction in materials costs, but at the same time they are also required to offer longer term reliability enhancements. A competitive advantage will be received by reduction of resource use as well by using more environmentally friendly alternative solutions. The technologies involved in applying these alternatives would also give the European electronics industry a new route for the manufacture of the new generations of PCBs.
The socio-economic impact of the knowledge generated in work package 2 is related with the outcomes which are of a great significance to the European PCB industry. The ability to enhance the reliability of its products will improve the international competitiveness of European PCB fabricators, thereby helping to ensure the successful future of a strategically important industrial sector in the face of intense international competition. This in turn will help to maintain employment in both the European PCB industry and with its suppliers.
A key aspect of the ASPIS project has also been to deliver the results of the project and the expected outcomes to a broader number of final users. Dissemination activities of the results gained in work package 2 to date have been as follows: presentations at the International Conferences Chemistry 2011 and Chemistry 2013 in Vilnius, Lithuania, at the 63rd Annual Meeting of The International Society of Electrochemistry, Prague, 2012, the EIPC Winter Conference in Berlin, the EUROCORR Conference in Estoril, Portugal, 2013. Summaries of the key findings and results obtained were published in two Circuit World papers during 2012 and 2013.
Impact WP3: ENIG Screening Tool (TNO)
The most important achievement of work package 3 has been the development and realisation of a non-destructive, flexible and easy to use measuring concept for bare printed circuit boards. This is a truly novel concept which should subsequently be patentable, as initially required by the consortium.
The developed tool is based on the use of the capacitive plate sensor principle. This measuring technique can be used for different purposes such as 1) characterisation at a research and development stage, 2) inspections as acceptance sampling on product (bare board) level during printed circuit board manufacturing and 3) a process control solution which monitors critical to quality indicators during flow production.
During the ASPIS project the tool proved to be very valuable to systematically build up knowledge and evidence to correlate indicators and failure modes to black pad or other possible surface defects. Furthermore, the tool has been extensively used to study the effects of process modifications and to detect possible drifts or shifts in both materials and processes.
Detection and control of variation
A key piece of newly developed knowledge concerns the ability to detect the presence of an immersion gold layer and the ability to correlate the amount of gold in the top layers with capacitance characteristics. Also, the ability to distinguish between layers made with different plating parameter settings, also adds to the new knowledge developed in this work package. Regarding printed circuit boards, crucial aspects were learned about within pad variations, between pad variations and product to product variations. Furthermore, a lot was learned about the new tool’s specific capabilities and system variations (repeatability and reproducibility).
Novel probing head
An important part of the test setup concerned the novel probing head, which proved to be reproducible and still has the potential for further miniaturisation. Another important feature of the head is the fact that it was developed to minimise the risk of damaging the printed circuit board. Furthermore, the development of a procedure to produce accurate probing pins with dielectric material attached underneath the tip was crucial for the performance of the new measuring system.
Cost modelling
The development of a dedicated cost modelling tool to study economic aspects of different inspection strategies is an important non-technical achievement of the ASPIS project. The model concerns a parametric tool which enables the evaluation of different scenarios before adding an extra inspection step to the printed circuit board manufacturing process. Within the project, the inspection step was defined as a sub-process.
Further development and commercialisation
Since a measuring concept has been developed instead of a specific tool, there will be different possibilities for the further development and commercialisation stages of the setup. A standalone bench-top setup is most likely to be the best variant to first further develop the measuring method and generate more knowledge regarding the capacitive behaviour of stacked plated finishes. Since the SME partners in the project are all printed circuit board manufacturers, a company like Atotech would more likely be a future company to develop, for instance, an integrated process control solution which coupled chemistry with equipment. Atotech were an LE partner in the ASPIS project and are suppliers of plating solutions on an global basis.
Impact WP 4: Aqueous Plating (ITRI)
The majority of the work undertaken in WP4 was concentrated on one of the possible routes envisaged at the start of the project; modification of the existing ENIG technology to produce improved coating performance. The decision to do this was made on the advice of the industrial partners after work had been underway in all areas for some time. The reasoning was that it was apparent that manufacturers adopting alternative materials and processes would likely have to spend a significant sum of money and that this would deter them from making the switch. However, the preliminary work on immersion gold deposition from a screen-printed gel was promising and may have some applications outside of the electronics industry. The work undertaken has shed light on the importance of certain process parameters and given a better understanding of how these might influence productivity and product quality.
The identification of the morphology of the Electroless Nickel as the key vehicle for improving the quality of ENIG coatings required that the failure mechanisms be considered and ways to avoid these from occurring be identified. This approach has not, to our knowledge, been proposed previously nor attempted. The use of brightener packages in Electroless Nickel is novel, but builds upon knowledge already used to electroplate nickel. This may make it easier for acceptance in industry and this approach means minimal changes would be required, thus lowering implementation expenditure. The coating morphology is such that the minimum nickel thickness required may be reduced, thus saving materials costs and plating times.
Unfortunately, identification and development of the modified EN coatings has taken longer than originally envisaged in the project plan. The biggest contributor to this was staff turnover as the original key researcher for the work package left ITRI and it took several months for his replacement to be recruited due to the unique requirements of the position. At the official end of the project the overall impact of the work package has been diminished by the slip in timescales. This is because key long-term data could not be completed by the end of the project. This work is continuing at ITRI past the official end date, so that it will be generated. This will be circulated to the other project partners once complete and some of this data has already been provided as an annexe the work package 4 deliverables. The data that has been generated indicates that the general performance of the modified EN coatings is better than that of laboratory and industry-produced conventional ENIG coatings.
WP4 contributed the bulk of test samples used in WP3 and WP6. Both these work packages have investigated and refined various types of test methods and evaluation tools to better characterise ENIG coatings, gauge their quality and highlight defects and coating performance.
It was said on various occasions by different people that WP4 was probably the hardest of the work packages in the ASPIS project, in part because aqueous ENIG technology is mature. The author believes that the work conducting in WP4 has developed a novel improvement that, if the long-term reliability and storage data proves acceptable, will prove interesting to the PCB industry.
Impact WP5: Ionic Liquid Plating (UoL)
Introduction
In electronics assembly, the reliability of the electrical connection between printed circuit boards (PCB) and components is paramount to efficient manufacturing processes and good longevity of the resulting products. One of the methods used to achieve this is the incorporation of a protective surface finish onto the copper tracks on PCBs. The most common of these is the electroless nickel immersion gold (ENIG) finish which is popular due to its high reliability, good planarity, solderability and long storage life.
Despite its popularity, ENIG finishes still have a number of issues that are a cause of major concern, the most important of which is “black pad”. Black pad describes the process of oxidation of the underlying electroless Ni (EN) layer during the immersion Au (IG) process, leading to poor solderability and solder joints that are unreliable and prone to brittle failure. The biggest issue with black pad is its difficulty in identification, since it is not readily visible and yet can be the cause of expensive failures in subsequently assembled equipment. Due to the disrupted surface being covered in a layer of gold, it is very hard to recognise this problem when it occurs, leading to the boards entering use and causing the failure of a large number of electronic items.
Ionic Liquids and the ASPIS Project
The ASPIS project was undertaken to identify the root mechanism for ENIG reliability problems, the key one being so called ‘black pad formation’, to find alternative formulations for elimination of ‘black pad’ and to design and develop a testing methodology to enable identification of PCB displaying black pad before they were used and assembled with components.
WP5 focused on the development of alternative solvent methodologies, namely the use of ionic liquids, for use in the design of an IG plating process which removed the need for acidic media and/or poisonous cyanide salts. The galvanic corrosion of EN nodule boundaries by protons in acidic media has long been associated with black pad formation. Also, the use of traditional cyanide salts was deemed to be increasingly undesirable and any new solution that eliminated their use would be of great commercial interest due to the reduction in processing costs, elimination of hazardous materials, less environmental and waste treatment costs, enhanced operator safety etc.
Ionic liquids have a number of unusual properties that make them of interest in metal processing. They are composed of alkyl ammonium halide salts and small polar organic molecules. As such, they have a very high chloride ion concentration (c.a. 6 molar), as well as containing polar uncharged compounds which makes them excellent for solvating a wide variety of metal salts including oxides. At the same time ionic liquids themselves are relatively benign and are conventionally made from widely available and sustainable materials.
Neutral Solutions for Immersion Gold Deposition
Four new gold deposition solutions have been developed as part of this work that will be of great interest to printed circuit board fabricators. The basic formulations of these solutions were as follows:
1 5 mM AuCN in Ethaline 200
2 5 mM KAu(CN)2 in Ethaline 200
3 5 mM KAu(CN)2, 10 mM ethylenediamine in Ethaline 200
4 5 mM AuCl, 1 mM I2 in Ethaline 200
A key advantage of each of these solutions is that they avoid the nodule boundary corrosion problems common with conventional aqueous methodologies for ENIG plating. In the case of the first two of these solutions, 1 and 2 above, both of these formulations produce bright, lustrous coatings consistent with those of the conventional aqueous process from neutral media, something that is not possible using aqueous media. These coatings have also been demonstrated to wet solder faster and more reliably than conventional commercial ENIG plating systems. They thus provide the basis for a completely new approach to gold deposition in ENIG coatings and one that could be further developed towards commercialisation by Aspis partner Atotech or other European chemistry suppliers to the PCB industry (see the Aspis business and exploitation plan for more details).
Formulation number 3 follows an extremely fast deposition profile while producing bright lustrous coatings due to the ability of ethylenediamine to accelerate the nickel dissolution kinetics. In addition, the centre of the EN nodules dissolves faster than that of the nodule boundaries eliminating the main mechanism of black pad formation producing highly reproducible coatings with excellent solderability properties. This is backed up by solder wetting data which, as with 1 and 2, shows fast, reproducible, solder wetting rates.
Cyanide-Free Coatings from Ionic Liquids
currently no cyanide-free formulations are available commercially for the immersion Au part of the ENIG process. However, formulation 4 above was developed using gold chloride as the Au source and I2 to activate the Ni surface to provide uniform, bright and quickly depositing Au coatings. This was also done from an acid-free formulation potentially eliminating one of the main drivers of blackpad.
Real World Potential
All of the formulations and protocols that have been developed could be developed for further use as direct drop in replacements for the conventional aqueous based coating methodologies. The newly developed coatings perform as well as, if not better than, standard systems, while also either reducing or eliminating the main mechanism of black pad formation. As such, these novel processes would have a low barrier to entry to commercial adoption once further development, scale up and validation in a production environment has been completed. In addition, the ability to remove cyanide from the Au plating bath serves as a highly attractive driver towards commercial adoption which could potentially lead to widespread use as an immersion gold formulation in many ENIG plating processes. Following the successful development of these highly novel chemistries and the completion of the Aspis project, further collaborative development work is needed to further optimise the formulations, to fully assess the reliability that they provide in real world applications and to consider their production on a larger scale. Such work will need to be undertaken in concert with a suitable supplier of such coatings to the PCB industry. Atotech would be an obvious partner as they were active in the Aspis project and also because they are an existing established supplier to the PCB and metal finishing industries not just in Europe but on a global basis. At the time of writing of this report, further collaboration and development opportunities are currently being investigated.
Impact WP 6: Validation (TNO / ITRI)
At the start of the project there was no common understanding about the black pad failure mechanism and there were no procedures available to analyse and confirm black pad or black pad related failures. The failure analyses conducted within WP 6 were done on printed circuit boards provided by the SME partners of the project (Merlin Circuits Technology, Graphic plc, Somacis and Global Interconnection Services) and one Dutch electronics assembly company called Liad. The boards were coming both from the field as assembled boards and as bare boards from the printed circuit board manufacturing process.
An important result of the conducted failure analyses was the correlation which was found between the morphology of the copper conductor and certain failures on ENIG and solder joint level. In one case, even the extent of solder wetting within a pad could be directly related to a high difference in the copper layer underneath. The remaining non- / de-wetted part of the solder pad showed a black nickel surface. Also, WP 4 trials comparing nickel deposits on cold-rolled copper and electrolytic-plated copper substrates showed that the morphology and characteristics were strongly dependent on the nature of the underlying copper. Discussions with printed circuit board manufacturers revealed that there were currently no industrial standards or guidelines regarding the copper surface condition for printed circuit boards entering the ENIG plating process. The knowledge gained about this correlation could be further investigated with a Design of Experiments (DoE) to quantify limits and tolerances. Such an outcome could then be implemented in an industrial standard or an industrial guideline.
Another important result concerns knowledge about a potential black pad indicator. This result is based on the failure analyses conducted within this work package and a result of the failure mechanisms work package, WP 2. It has been repeatedly found that black pad and black pad related failure sites exhibit traces of copper at the top surface of defective electroless nickel immersion gold layers. One hypothesis is that, in case of hyper-corrosion during immersion gold plating, the resulting cracks between the electroless nickel grain boundaries will allow copper to diffuse from the base conductor material to the top surface. On bare printed circuit board level, the copper will then be present at top of the electroless layer, underneath the immersion gold layer. This failure mode can be used as a possible indicator for black pad or black pad related failures. During the project a procedure was developed to use Scanning Electron Microscopy (SEM) with Energy Dispersive X-ray (EDX) to determine whether or not copper traces are really present at the top surface. By measuring the same printed circuit pad location with different acceleration voltages, it can be determined how the copper content changes as a function of the acceleration voltage. The copper is present underneath the surface in case the copper content decreases while increasing the acceleration voltage. This practical, non-destructive procedure has been extensively used within the project to get a first indication during a failure analysis investigation. Since such copper traces can also be a result of a contaminated plating bath, it should be noted that this can only be used as a possibility indicator. The detection of copper traces should always be further investigated to confirm black pad or a black pad related defect.
The soldering tests within this work package showed that printed circuit test boards made with both the Standard ITRI and the XL plating process provide proper solderable surfaces. Soldering tests conducted with a lead-free SnAgCu alloy showed promising wetting behaviour and defect free solder joints. Cross-sectioning revealed that nearly all investigated samples exhibited complete intermetallic layers at the interfaces. Reliability testing has also proved the potential stability of these interconnects. Both Standard ITRI and XL test boards were subjected to a high temperature aging test at 150°C. Samples taken at 0, 20, 100, 200 and 500 hours did not show any irregularities in intermetallic growth or the introduction of defects. Statistically, there appeared to be no significant difference in intermetallic growth between the Standard ITRI and the XL plated test boards, which regarding this aspect, confirms the level of the newly developed plating processes.
Another important result regarding the XL plating processes was the outcome of the experiments to deliberately induce the black pad failure. Hyper-corrosion was tried to induce by systematically reducing the thickness of the electroless nickel layer. This with the aim to create a porous nickel layer, layers which tend to be more prone to hyper-corrosion during immersion gold plating. Soldering tests showed that even the boards with the lowest nickel layer thicknesses (1 µm), did not result in black pad failures. Cross-sectioning of these thin layers did not even show the typical corrosion spikes, mud cracks or black bands along the interface. These test results indicate the robustness of the newly developed plating processes. These and the above described work package 6 achievements form should form valuable inputs for the further development and market introduction of the newly developed plating processes.
Summary and Conclusions
The Aspis project was undertaken to address a very important and intractable problem known as ‘black pad’ that impacted the reliability of printed circuit boards assembled using nickel-gold (ENIG) solderable finishes. These problems had been known for many years and, despite the efforts of both the board manufacturers and their plating chemistry suppliers, little progress had been made in either understanding the basic mechanisms causing the problem, or in identifying new processes that didn’t exhibit the problems. The Aspis project adopted a four pronged approach to addressing these reliability issues and this included; a detailed study of the mechanisms, the formulation of two new and distinctly different plating chemistries and the assessment of new non-destructive test methodologies. By developing an understanding of the mechanisms that caused these reliability problems, it was anticipated that the knowledge could be used to develop new aqueous plating processes. In addition, the use of ionic liquid based plating chemistries was felt to offer the possibility to use novel materials and deposition conditions that would avoid the possibility of initiating ‘black pad’ and related reliability problems. Finally, because ‘black pad’ is not visible or detectable without destructive testing, the project investigated the possibility of developing a non-destructive test method that could be used both by PCB fabricators and assemblers to give confidence that there were no hidden, latent reliability problems. Based on extensive experience, each of these four approaches was known to be extremely difficult to successfully complete and thus there was a high element of risk inherent within the overall project plans.
The mechanistic work carried out in work package 2 by LIOC has been extremely successful. Extensive, complex and detailed studies have, for the first time, enabled the mechanisms of ‘black pad’ to be properly elucidated. This in turn has enabled the chemistry formulations and process conditions leading to its formation to be identified. This knowledge will enable process chemistry suppliers to take these factors into account when formulating new chemical processes for the deposition of nickel gold solderable finishes. From a scientific perspective this work package has generated much new knowledge and its significance and importance cannot be overstated.
Attempts to develop new ‘black pad’-free ENIG systems have been made by the key industrial process chemistry suppliers over a number of years prior to the Aspis project, but without success. The work package to develop new aqueous-based processes was, therefore, known to be extremely challenging, with a high risk and low potential for success. However, the availability of the mechanistic information from work package 2 enabled the aqueous process development work in work package 5 to be focussed on novel formulations and conditions that were likely to produce reliable coatings. Whilst some progress has been made, the results at the end of the project confirmed that a substantial amount of additional work would be required before any of the formulations investigated could be considered for a move towards commercialisation.
Non-destructive testing of circuit boards for the presence of ‘black pad’ has been demonstrated and a bench scale prototype constructed and evaluated in a production environment. This has never been possible before and the capacitative method developed by TNO in work package 4 has the potential, with further development, to become a valuable piece of test equipment that could be used on a routine production basis by PCB manufacturers to confirm that their production is ‘black pad’-free. It would also be useful to board assemblers who could employ it as a quality control tool for the incoming inspection of circuit boards prior to soldering. Further work will be needed to take this technology forward and the partners will continue to explore the possibilities following the end of the Aspis project.
The development of new metal deposition processes from ionic liquids was always considered to be a more long term solution to the ENIG black pad problems. However, the University of Leicester has developed a range of gold deposition formulations which avoid the use of materials and process conditions that can lead to black pad formation. Using information from the mechanistic studies undertaken by LIOC, it was soon realised that the key to avoiding ‘black pad’ would be the development of new gold deposition processes. These have been successfully developed and evaluated on the laboratory scale and evaluated by the PCB fabricator partners. There is considerable interest in taking these chemistries forward to commercialisation and the possibilities will be further explored with the Aspis partners and other potential exploiters.
In final summary, the Aspis project has been successful in achieving most of its objectives in each of the four key work packages. Progress on the development of new aqueous chemistries has not been as significant as was hoped for, but it has given clear directions for future work. The demonstration of a non-destructive test method and equipment was highly innovative and offers potential for the production of a commercial unit. Finally, and perhaps the key achievements of the Aspis project have been the mechanistic studies and the development of ionic liquid-based gold deposition processes which are both world firsts and represent highly innovative scientific information and new technologies that should help to eliminate the occurrence of black pad in printed circuit boards.
30th November 2013
List of Websites:
Web site address: www.aspis-pcb.eu
Coordinator name and email address. Dr Martin Goosey, martingoosey@aol.com
The three year Aspis ‘Research for SMEs’ project has been conducted to address problems found with the nickel gold (ENIG) solderable finishes that are used by the electronics industry to enable components to be soldered to printed circuit boards (PCBs). ENIG has several key advantages over other types of PCB solderable finishes and these are highly significant, making ENIG the preferred choice for many applications. However, while ENIG coatings have a good reputation for excellent solderability, they are also prone to a number of well known reliability problems, such as ‘black pad’, whose formation mechanisms were still poorly understood. When ENIG-related problems occur, they can affect an entire product design or batch and the problems are often only identified after assembly, during which expensive components have been soldered to the PCBs.
The ASPIS project has focused on the development of new chemical processes, as well as methods for detecting ENIG related issues. Via four key work packages, a multi-faceted approach to ENIG reliability problems has enabled significant progress to be made on all fronts. The ASPIS project has carried out work to reduce the likelihood of ENIG-related problems occurring by improving the coating deposition technologies. Through the specific work packages on metal plating development, the ASPIS project developed new processes that will enable thinner coatings to be used, thus reducing the cost of using ENIG while offering reliability enhancements. Another key goal was to develop methods for identifying potential problems related to the use of ENIG on PCBs and a non-destructive screening method has been demonstrated that identify problems on boards before components are soldered. One way to prevent ENIG-related problems is to avoid them altogether by using different coating technologies. With work carried out by partners ITRI and Leicester University the project successfully developed alternative deposition methods from both aqueous solution and ionic liquids. Several new gold deposition processes were successfully developed.
The knowledge outputs of the ASPIS project will help to reduce the number of ENIG-related problems reaching production and also accelerate investigations and appropriate response times. This will reduce associated costs and increase customer confidence in the ENIG process. The development of an assembly line tool for identifying problematic PCBs will reduce lost production costs and lower the risk of field-failures and their potential consequences. The improved and alternative coating technologies developed are now available for further development and will enable European PCB fabricators and their customers to produce more competitive products via both increased process and assembly yields and to provide higher levels of quality in finished products.
The work carried out in the project has been widely publicised via technical papers, articles in trade journals, conference presentations and posters, as well as via an active website and industry briefing events.
Project Context and Objectives:
The three year ASPIS project, (Advanced Surface Protection for Improved Reliability PCB Systems) undertook a multi-faceted approach for developing novel and improved reliability nickel-gold solderable finishes used in the assembly of electronic products. Aspis was a ‘Research for SME Associations’ project with twelve partners including four key research organisations (RTDs) who undertook much of the research. These were ITRI Ltd (UK), the University of Leicester (UK), TNO (The Netherlands) and the Center for Physical Sciences and Technology (Lithuania). The other partners represented a wide cross section of organisations from the PCB sector covering the whole of the industry supply chain. These organisations were Atotech GmbH (Germany), Graphic Plc (UK), Merlin Circuits Technology Ltd (UK), Somacis S.p.a (Italy) and Global Interconnect Services (The Netherlands), the European Institute of Printed Circuits (The Netherlands) and the Institute of Circuit Technology (UK).
The ASPIS project had a specific focus on nickel-gold (ENIG) solderable finishes for PCBs, and its key objectives were to develop new, more reliable materials and processes in order to address the key issues such as ‘black pad’ that have been a cause of concern for many years to both PCB fabricators and end users. One of the key factors impacting the overall reliability of electronic assemblies is the quality of the solder joints that connect components to a circuit board and a critical factor in determining this reliability is the solderable finish that is applied to the board. The standard choice of solderable finish for high reliability, high value electronics is nickel-gold. This utilises an undercoat of electroless nickel which is deposited on to the copper of the PCB and on top of which is a thin coating of immersion plated gold. Globally, the market share of ENIG in the PCB industry as a whole is estimated at 15% by surface area of PCB manufactured. However, in terms of the value of the resulting products, the market share is considerably higher. For the EU’s electronics manufacturers, which are increasingly producing high value and high reliability products, the importance of ENIG is much greater. ENIG has several key advantages over other types of PCB solderable finishes. These are highly significant as, despite the higher price of ENIG over other alternatives, it is the preferred choice of finish for many applications. A key advantage of ENIG is that it offers excellent solderability and can retain this during prolonged storage prior to soldering assembly and during the multiple soldering operations encountered when assembling complex, high value electronics. While it was predicted that ENIG’s market share would increase significantly after implementation of the Restriction of certain Hazardous Substances Directive (RoHS) in 2006, the expected increase has not materialised due to a combination of cost, the emergence of some lead-free HASL alternatives and because of a fear of ENIG-specific problems which can lead to disastrous yields and product failures if they occur. Another advantage of ENIG finishes is that they have excellent planarity and this is crucial when assembling and soldering small fine-pitched surface mount components. These are often of high value and placement accuracy is critical for achieving high assembly yields. Unlike other coating technologies, which are applied solely to facilitate storage and solderability during soldering assembly, ENIG is a functional coating that utilises the solid-state properties of the nickel to stabilise the soldered interface, in effect, acting as a barrier layer to prevent the growth of tin-copper intermetallic compounds, which can embrittle the soldered joints during the service-life of the product. This greatly enhances the reliability of ENIG-coated PCB assemblies, especially during prolonged use at elevated temperatures.
However, while ENIG coatings have a good reputation for excellent solderability, there are a number of technical and economic factors which can cause problems for PCB fabricators and their customers. The most widely recognised of these is referred to as ‘black pad’ which is thought to be attributable to excessive corrosion of the nickel coating during the subsequent immersion gold deposition process. Despite having been identified more than 10 years ago, the mechanisms that cause the effect (and the contributing factors) were still poorly understood prior to the work carried out in this project. Moreover, no design rules, non-destructive testing methods, process optimisation/mitigation steps or viable alternatives existed. Other problematic failure mechanisms include insufficient gold coating thickness, gold coating porosity, nickel migration, excessive phosphorus levels in the nickel layer, poor plating quality, excessive coating stress in the nickel, and general solderability degradation. When ENIG-related problems occur they usually affect an entire product design or batch and the problems are often only identified after assembly, during which expensive components have been soldered to the PCBs. Many ENIG-related problems are only identified later via field-failures and when this occurs, large sums of money are spent identifying and resolving the problem, usually by several members of the electronics supply chain. In some circumstances, especially if liabilities and compensation claims are pursued, the total cost of an ENIG-related problem could be orders of magnitude higher than the cost of the PCBs supplied by the fabricator. It is probable that any ENIG user who has been using the finish for a substantial length of time will have experienced some of these problems.
The ASPIS project aimed to help both the European PCB fabrication and broader electronics industries by developing new chemical processes, as well as methods for predicting, avoiding and detecting ENIG related issues. Prior to the Aspis project,, the failure mechanisms were not fully understood and, therefore, the factors which could contribute to them were not all known. This prevented fabricators and assemblers from properly optimising their processes and users from predicting or even identifying which PCBs might be affected or by which type of mode products have failed. The ASPIS project also had the overall objective of strengthening the European electronics industry supply chain, from material suppliers to end producers, by helping ENIG users and by offering enhanced coating technologies with similar solderability and reliability benefits. It was important to support ENIG users because PCB fabricators’ customers will continue to specify it for their products, regardless of whether new alternatives emerge. This was typically due to familiarity, the lack of historical data of field reliability and because purchase specifications exist which cannot be changed without requalification; a long and expensive process. For these people, the fundamental knowledge developed in ASPIS will help them to identify and avoid design and process-related factors which can increase the risk of failures occurring. As ENIG will continue to be used on existing boards, which obviously cannot retroactively be subject to these precautions, the knowledge outputs of the Aspis research will also enable engineers to identify key areas to investigate in order to validate the quality of the ENIG coated boards. Another key objective of the ASPIS project was to help reduce the likelihood of ENIG-related problems occurring by improving the coating deposition technologies. Unlike previous developments, which have largely been undertaken empirically, the fundamental mechanistic knowledge developed was used to intelligently target specific improvements. For example, one important area to be addressed was the thickness of the gold deposited. Sufficient gold is required to provide a continuous non-porous film but, as this thickness increases, so does the cost and there can also be a negative impact on solder joint reliability. Through specific work packages focussing on plating development the ASPIS project researched new processes that enabled thinner coatings to be used, thus reducing the cost of using ENIG while offering the potential for longer term reliability enhancements. Another key goal was to develop methods for identifying potential problems related to the use of ENIG on PCBs and two key approaches were investigated. The first was to use laboratory methods for identifying which type o failure mechanism caused a particular problem. The second method was to develop a non-destructive screening method that can be used at the PCB fabricator or electronics assembler in order to identify problems on boards before components are soldered onto them. Therefore, another key aim of the ASPIS project was to build a prototype instrument for demonstration and validation purposes. This was successfully achieved and performance evaluation trails were carried out by RTD performer TNO, who developed the equipment, within the production facility of SME partner Graphic.
One way to prevent ENIG-related problems is to avoid them altogether by using different coating technologies, either via different deposition methods or by the use of alternative materials or structures. Research carried out in specific work packages of the ASPIS project investigated and developed alternative deposition methods from aqueous solution and ionic liquids. Most of the methods for depositing different metals from aqueous solutions are well established and are widely used for other applications. The novelty is applying them to this high reliability, functional, solderable coating technology. Plating from ionic liquids is a more recent development and is not yet used in the PCB fabrication industry. However, it has been shown that gold deposited from ionic liquids exhibited improved properties which may, for example, enable thinner coatings to be used.
The current lack of in-depth understanding of the nickel-gold process means that no design rules exist for avoiding problems such as ‘black pad’, nor are there standards in place that can be used to minimise the risks. The most widely applied industry standard, the IPC 6010 series, merely specifies a minimum coating thickness for the nickel and gold layers of 3.00 μm and 0.05 μm respectively for coatings on rigid boards. The knowledge, data and experiences gained and accumulated in ASPIS will be used to improve and expand the industry standards and purchasing specification requirements. Improved ENIG or alternative coating technologies will, likewise, require applicable standards in order to get industry acceptance.
The new information and methodologies being developed by the ASPIS project will help reduce the number of ENIG-related problems reaching production and also accelerate investigations and appropriate response times. This will reduce associated costs and increase customer confidence in the ENIG process. The development of an assembly line tool for identifying problematic PCBs will reduce lost production costs and lower the risk of field-failures and their potential consequences. The development of improved and alternative coatings technologies will also enable European PCB fabricators and their customers to produce more competitive products via both increased process and assembly yields at subsequent higher levels of quality in finished products. Reducing the quantities of nickel and gold used in the solderable coatings by improvement or replacement will also reduce the overall cost of the PCB finish and also help the European electronics industry to adopt more sustainable approaches to materials use.
Project Results:
Introduction
As detailed above, the Aspis project had a number of key objectives aimed at understanding the causes of reliability problems in ENIG finishes. These were undertaken via a four pronged approach and comprised an investigation of the basic black pad mechanisms, formulation of new aqueous and ionic liquid basic deposition chemistries and the evaluation of non-destructive test methodologies and equipment. These were addressed through specific work packages led by each of the individual RTD providers. The scientific and technical foreground is thus reported specific to each of these key individual work packages.
Work Package 2 - Fundamental Research on ENIG Mechanisms
Overview
Prior to the Aspis project the causes of black pad and other related reliability problems in ENIG solderable finishes were poorly understood and the fundamental mechanisms of black pad formation had not been fully elucidated. Therefore, a key to eliminating these problems and improving reliability was gaining a full understanding of the mechanisms and this work was undertaken by the partner ‘LIOC’, in Vilnius, Lithuania, in Work Package 2. The principal scientific objectives of this work package were as follows:
• Identification of mechanisms influencing each failure mode via experimental determination.
• Determination and quantification of key factors and relationships for failure modes identified.
• Establishment of rules for PCB design and layout.
• Recommendations for improved international standards and purchasing specifications.
Five specific tasks were predicted in this work package with the following short characteristics of the planned assignments in each of them.
Task 2.1. A laboratory based study to investigate the fundamental mechanisms and key influencing factors which can lead to ENIG problem.
The following studies were undertaken with the aim of determining the key factors and conditions, which favoured enhancement of the corrosion rate of Ni-P coatings:
• deposition and detailed characterization of Ni-P and coatings, including the following parameters: phosphorus content, thickness, porosity, internal strength
• corrosion behaviour studies of Ni-P coatings in supporting solutions
• characterization of gold layers including the following parameters: thickness, porosity and quality
• corrosion behaviour of Ni-P coatings in gold deposition solutions: a structural characterization of Ni-P samples after removal of the gold layer, including phosphorus distribution on the surface.
Task 2.2. Experimental investigation of the various phenomena and, particularly, the relationship between the plating process, coating quality, PCB design and other associated factors that have been linked with the failure mechanism.
The following investigations were planned: analysis of failures due to mechanically weak areas in Ni-P/solder bonding, which might be caused by:
• Ni-P coating characteristics: the presence of Ni surface morphology defects (mud cracks, major/minor spikes, etc.), excessive coating stress, excessive coating porosity, excessive P levels, excessive levels of inclusions (S, etc.)
• or/and Au coating characteristics such as insufficient coating thickness, excessive porosity, poor plating quality
• or/and the peculiarities of Au deposition process, i.e. micro-galvanic cell formation, non-uniform Au layer plating at sharp nodular boundaries, concentration cell formation on PCB due to varying in pad size, etc.
Task 2.3. Small-scale plating operations in laboratory-scale facilities with the aim to investigate a wider range of conditions and factors related to the “black pad” problem.
Electroless nickel and immersion gold coating deposition from continuously corrected solutions was undertaken with the aim of determining the influence of plating solution on the quality of deposited coatings by controlling the coating thickness, phosphorus content, porosity, deposition rate and surface morphology.
Task 2.4. Characterization of an affected PCB by identifying microstructural and chemical features that occur in problematic ENIG coatings.
The aim of the planned investigation was to establish the potential reasons for the “black pad” issues found in PCB samples presented by different suppliers. The analysis of PCB samples with and without soldered components was performed with the principal aim of establishing the root causes of very poor soldering, the presence of voids on SMT pads, or, in general, to determine the reasons and nature of the defects in PCB samples.
Task 2.5. To establish industrial design guidelines and to improve international standards.
Modified experiments to confirm the predictions of the “black pad” failure mechanisms related to inadequate copper surface preparation, deviations from plating chemistry formulation, variations in electroless nickel surface morphology and defective structure, etc. were carried out with the aim of establishing industrial design guidelines related to the “black pad” problem appearance in PCB production. Having in mind the copper pre-treatment and immersion gold solution composition relationship with the “black pad” phenomena, the aim of the investigations was to determine copper surface etching parameters and copper ion concentration limits in immersion gold solutions, which may cause the “black pad” failure.
Progress and Results
Literature analysis: According to the existing literature, the key influencing factors causing the “black pad” phenomena could be divided into the three groups. The first one related failures to EN characteristics (excessive stress or amount of inclusions, porosity, phosphorus level), the second one to gold coating characteristics (insufficient or excessive thickness, excessive porosity, poor plating quality) and the last one to peculiarities in the IG process (micro-galvanic cell formation due to non-uniform gold plating, concentration cell formation because of varying in the pad size or electric current flow in PCB during IG process). It can be stated, however, that the majority of publications were non-systematic works based on assumptions without real experimental evidence. It is widely accepted that the corrosion properties of EN deposits are governed mainly by their phosphorus content and the corresponding structural and mechanical state, however, the nature of corrosion behaviour of a nickel matrix doped with phosphorus atoms remains unclear. It was stated, as well, that the difficulties of reproducing black pads using pure chemicals was a barrier to the understanding of the phenomenon.
Task 2.1. Electroless nickel coatings deposited with phosphorus contents ranging between 3.5 and 11.0 weight percent exhibited nodular morphology. Corrosion behaviour studies of these coatings in citrate media revealed the fact that corrosion current density values (icorr) were related to P content: Electroless nickel (EN) samples with a higher P content exhibited lower icorr. It was assumed that the higher corrosion resistance of amorphous EN was due to their homogeneous structure, the absence of apparent grain boundaries and, possibly, dislocations, kink sites and other surface defects. For samples with higher P contents, local pitting corrosion sites (spikes) with very small diameters (< 1 mm) and high length-to-diameter ratios were observed, while corrosion of the samples with lower P levels occurred along the grain boundaries and this type of damage was more severe. An attempt was made to modify EN plating solutions by adding organic surfactants with the aim of facilitating hydrogen recombination, however, it can be stated that the presence of sodium lauryl sulfate in the EN bath had no influence on the EN layer morphology. It was concluded, on the basis of the results obtained, that micro porosity, roughness and inhomogeneity, e.g. due to internal stress within the coating, were all strongly dependent on the pre-treatment of the substrate.
Investigations of the immersion gold (IG) process revealed that the gold deposition rate was determined by solution pH, composition and surface reactivity of EN. A more reactive Ni surface (lower P content) led to a higher plating rate in the IG process and an increase in the amount of corrosion damage. The presence of copper in the corroded ENIG area was detected using appropriate analytical techniques. The assumption was made that the porosity of EN coatings might be the source of copper contamination and, therefore, one of the factors which influenced the quality of the IG coating and possibility of the black pad occurrence.
The following mechanism of IG layer formation has been proposed. As EN plating has a nodular structure, there are boundaries and crevices between the nodules. If a boundary or crevice is too deep and thus the supply of Au atoms to the crevice is slowed down, the Au concentration in the crevice will be different from that in the plating bath. Consequently, a galvanic cell will be set up between the crevice and the surface, resulting in heavy corrosion in the crevice. Therefore, the corrosion converts the dense, amorphous EN into a porous, micro-crystallized structure into which the Au atoms have penetrated. It can be stated that corrosion of the Ni surface is due to its activity in the immersion gold process. High IG bath pH value (pH 6), together with a high concentration of citrate (0.4 mol l-1) and the thickness of the IG layer exceeding ~ 80 nm are all parameters which favoured black pad formation. Because of the formation of a voided layer, the interfacial bonding of solder to the ENIG plating may be, by nature, weak even if the EN plating does not suffer from hyperactive corrosion.
Task 2.2. One of the intrinsic defects of EN coatings important in the corrosion process is coating porosity. Hydrogen evolution, which occurs concurrently with nickel deposition, or a rough substrate surface, may be the reasons causing pore formation. An attempt has been made to elaborate a reliable and simple method for the quantitative evaluation of the porosity of EN, IG and ENIG coatings, which is based on the voltammetric measurements in KOH solution. The studies performed disclosed significant differences in the cyclic voltammetric curves of normal EN samples and those of defective ones, indicating that cyclic voltammetry of EN and ENIG samples in KOH solution may be a useful tool for the identification of defective coatings. It must be emphasized that the source of copper contamination and the role of copper on black pad formation had not been disclosed in previous works. The method for evaluation of ENIG porosity was elaborated and it was shown in this study that EN coating porosity may be the source of copper contamination. As the “black pad” phenomena is related to excessive corrosion of EN coatings during the subsequent IG deposition process, the nature of the phenomena must be related to the specific properties of the EN coating and the IG solution characteristics. Therefore, simulated experiments to confirm the predictions of the “black pad” failure mechanisms related to inadequate copper surface preparation, deviations from plating chemistry formulation, excessive amounts of Cu and Ni in IG solutions were carried out. The behaviour of standard EN coatings in IG solution containing Cu ions was compared with the behaviour of modified EN coatings and electroplated Ni-P alloys. EN coatings deposited under experimental conditions simulating inadequate substrate preparation indicated that insufficient degreasing of the Cu surface before EN plating was detrimental to the quality of the EN and ENIG coatings. Macro-structural defects such as pores and cavities were formed on the EN surface in IG solution, leading to direct contact of the substrate (Cu) with the IG solution components. The presence of inclusions on the substrate prevents uniform distribution of the catalytic sites (Pd particles) on the surface, what may lead to the growth of deposits with higher internal stresses. Such deviations from the EN deposition parameters with increased concentration of organic stabilizers or increased pH values of EN deposition solution (and, consequently, reduced P content) led to modification of the boundaries between the nodules, which may be detrimental for the further contact of such samples with IG solution. If accumulated, Ni ions in the IG plating solution do not affect ENIG coatings quality, whereas the presence of Cu causes realization of the black pad effect.
The detailed analysis of the influence of copper revealed, that when Cu, which has no electric contact with EN coating, is immersed into an IG plating solution, formation of a yellowish black or grey deposit took place on the Cu surface and the IG solution was contaminated with Cu ions. IG coatings deposited on EN substrates from solution contaminated with Cu ions were highly porous and loosely adhering to EN substrate. Such coatings cannot protect the EN surface from oxidation and could eventually lead to “black pad” failure of solder joints. When Cu, which has electric contact with EN coating, is immersed into IG plating solution, formation of the IG coating took place on both the Cu and Ni surfaces, because the process is governed by the potential of the Ni. Though spoilage was not readily observed, it would result in delayed darkening of the gold plated copper samples because, in the absence of a barrier layer, Cu atoms will diffuse through the Au and oxidize on the surface, leading eventually to “black pad” failure of solder joints.
Task 2.3. EN and IG coating deposition from continuously corrected solutions was performed with the aim of determining the influence of plating solution exploitation peculiarities on the quality of deposited coatings. The control of coating thickness, P content, porosity, deposition rate and surface morphology was performed. The results of the small-scale operations with EN and IG deposition from the continuously corrected solutions indicated the serious influence of cysteine, which was used as the EN plating solution stabilizer, on the deposition rate and amount of phosphorus in the EN coatings and, at the same time, on its activity during the IG process. The additional introduction of the stabilizer resulted in formation of organic compound inclusions in the EN coating, which were located at the grain boundaries and may be the reason for the appearance of weak boundaries. On the basis of the results obtained it can be stated, that in order to deposit EN coatings with the required P content (10-11 wt.%), it is obligatory to maintain the concentration of the cysteine in the plating solution lower than 0.01 mmol/l.
Before starting the IG plating experiments with continuously corrected solutions it was necessary to determine the influence of the concentration of citric acid in the IG solution on the rates of EN deposition and on the accumulation of excessive Ni in the plating solution, as well on the efficiency of Au deposition. In general, all ENIG coatings deposited from IG solutions without any correction secured the required thickness (＞70 nm) of the Au layer. Cyclic voltammograms revealed formation of rather uniform and compact immersion gold coatings with porosities not exceeding 5 %. The quality of the coating did not deteriorate during the exploitation of the IG plating solution. According to the stoichiometry of the IG deposition reaction, the consumption of the Au should be 750 mg, while only 560 mg was deposited. Consequently, the efficiency of the Au deposition was 0.75 what implied excessive Ni dissolution in the IG solution.
Task 2.4. Characterization of affected PCBs with identification of the microstructural and chemical features that occurred in problematic ENIG coatings can help with the understanding of the mechanism of defect generation. The analysis of defective PCB samples may be useful also for verification of the correctness of the conclusions drawn on the basis of experiments with model systems. In the work carried out in this project, PCB samples provided by different suppliers were investigated with the aim of determining the nature of the defects arising in the samples. The surface morphology analysis of the un-soldered pads of one of the samples disclosed the fact that the surface of the substrate Cu was not pre-treated in a proper manner before EN deposition, which led to the generation of structural defects such as cracks and opening pores in the EN layer and resulted in a poor quality ENIG coating. The next investigated case with the PCB sample confirmed observations that Au layers deposited on EN substrates from IG solutions contaminated with Cu ions were highly porous and loosely adhering to the EN coating, which, in addition, undergoes serious corrosion damages. Therefore, deposition of IG from solutions contaminated with Cu ions is one of the main reasons of the “black pad” defect occurrence.
High porosity of IG layer and the presence of the intermediate layer between Au and Ni-P, which was enriched in oxygen, were the main reasons for the black pad issue in the third PCB sample examined. Since the oxidized EN surface could not be wetted with the solder alloy, the uniform and complete wetting of Ni-P surface did not occur, therefore, the strength of the subsequent solder joints was not sufficient to secure good quality soldering. In summary, the experiments undertaken on samples of problematic PCBs have confirmed the validity of the conclusions drawn on the basis of the investigations with model systems performed during the implementation of this project.
Task 2.5. The results of the studies carried out in this project indicate that the “black pad” phenomenon is related to the specific structure of the EN coatings and some peculiarities of the IG deposition process. It was established that inadequate base metal (Cu) surface preparation resulted in deposition of EN layers with nodular morphology and serious structural defects, such as pores and cavities, which eventually led to a direct contact between the substrate (Cu) and the IG solution components. In addition, it was established, that Au layers deposited in IG solutions contaminated with Cu ions, were highly porous and loosely adhering to the EN coatings. Consequently, such Au deposits cannot protect EN surfaces from oxidation.
In order to establish industrial design guidelines related to the appearance of the “black pad” problem in PCB production, it was of a great interest and importance to conduct investigations with real PCB samples. Having in mind the Cu pre-treatment and IG solution composition relationship with the “black pad” phenomena, the aim of investigations performed was to determine Cu surface etching parameters and Cu ion concentration limits in IG solutions, which may cause the “black pad” failure.
Taking into account the influence of base metal Cu etching on such parameters of the EN coatings as porosity, surface roughness and morphological features, it can be stated that the optimal etching depth is ~1.5 µm. A Cu etching depth of 0.5 m is insufficient to disclose the inner structure of the Cu, and at the same time for a high quality, non-porous EN coating production. Meanwhile, a higher than 2.0 µm etching depth causes a significant increase in the roughness of the EN coating, which might be the reason for the enhanced EN surface activity in the IG process.
In order to avoid deposition of EN coatings with a structure favorable to the excessive surface activity in IG solution, it is recommended that plating solutions are operated with pH values lower than 5.5 while the optimal is close to ~ 4.8. The concentration of stabilizer (cysteine) is another important parameter, which determines the quality of EN coatings. In order to deposit EN layers with the required P contents (10-11 wt.%), it is obligatory to maintain the concentration of the cysteine in the plating solution to lower than 0.01 mmol/l. Higher concentrations of stabilizer result in formation of EN layers with lower concentrations of P (5.7 wt. %) and increased surface activities for IG layer formation.
An IG plating solution containing 0.4 mol/l citric acid is an optimal one, as the highest Au deposition rate, the highest efficiency of the process and the lowest amount of excessive Ni dissolution in the solution were observed. IG coatings deposited from the solution, containing an admixture of copper appeared to be highly porous and loosely adhering to the underlying EN substrate mainly because of Cu incorporation into IG layer. The detrimental effect of Cu ions in IG plating solution, influencing Au layer quality, excessive thickness and aggressiveness towards EN coating, manifests itself when the Cu ion concentration exceeds 3 mg l-1. Under the mentioned conditions the IG process induces formation of serious structural defects, such as pores and cavities in the EN coating, which provide direct contact between the substrate (Cu) and the components of the IG solution and lead to further contamination of it with Cu ions.
Taking in account the differences in the pad areas in real PCBs samples, it is important to note that Au layer formation on the smaller pads was more intensive and corrosion damage of the EN layer more severe and these differences increased significantly in the case of a Cu ion presence in the IG solution.
A methodology for investigation of ENIG-related failures was proposed, which included three main steps: determination of the EN layer porosity (opening holes), measurement of the IG layer thickness and composition. Voltammetric measurements in KOH solution provide a suitable method for EN porosity and IG layer composition evaluation. The latter parameter may also be obtained by XPS analysis. Meanwhile, the IG layer thickness can be measured using an SEM and ThinFilmId software, which is a part of the Helios NanoLab dual beam workstation with an Oxford Instruments energy dispersive spectrometer.
Work Package 3 - ENIG Screening Tool
Overview
The overall aim of WP3 was the development of a non-destructive screening method which was able to identify ENIG related problems on printed circuit boards before the components are being placed and soldered. To reach this the following objective were defined:
• Evaluation of various potential non-destructive techniques for detecting ENIG problems
• Selection of a preferred technique from those assessed
• Development of a patented prototype instrument for validation and demonstration purposes
• Identification of a manufacturer for commercial model
This work package was led by TNO. To accomplish the objectives of this work package the following tasks were defined:
Task 3.1: A pre-study wouldl be conducted to investigate potentially suitable measuring principles and methods. This took ongoing input from WP2 and considered all ENIG failure modes including properties, underlying mechanisms and factors contributing to initiation. New approaches would also be determined and evaluated based on optical, electrical, chemical and mechanical measuring techniques or combinations of these.
Task 3.2: To conclude a specification of inspection criteria and system requirements such as measuring capabilities and requirements that would assure process flow compatibility.
Task 3.3: Selection of viable measurement techniques based on technical and cost implications. Development of a quality assurance plan which contained the definition of the inspection strategy concerning where in the supply chain and process flows measurement could optimally be effected and on what basis.
Task 3.4: This core task involved extensive laboratory testing of ENIG-based board structures with all variants relating to modes of deposition, methodologies of deposition, solution compositions, board layouts etc with selected measurement techniques. Board input from WPs 4 and 5 into this work package wouldbe extensive.
Task 3.5: For validation and demonstration purposes, a prototype instrument would be built. The technical validation, which included the determination of the repeatability and reproducibility of the instrument would be studied by conducting a Gauge R&R measurement systems analysis study both at TNO and at partner SME sites.
Task 3.6: The economical feasibility will be assessed by conducting a Life Cycle Cost Analysis(LCCA). This was considered important in order to enable an assessment to be made and to determine whether or not it was possible to use it as a standard test method and to identify and involve manufacturers for a commercial model.
Task 3.7: Identification of potential manufacturers for a commercial model post IPR clarification – design, specification and prototype build activity.
Results
During the first phase of this work package the work was concentrated on learning about possible black pad appearances and potential inspection techniques. This study was done based on a literature search and the latest knowledge on materials characterisation techniques. An inventory of currently available inspection techniques, which was for this study focussed on both destructive and non-destructive techniques, resulted in the deliverable 3.1. In a next step the potential inspection techniques; capacitive measurements, magnetic measurements and confocal microscopy, were studied and tested more in-depth. To learn more about the black pad phenomenon and especially in what forms it can appear, a literature study was conducted based on more than 30 papers dealing with either black pad or electroless nickel immersion gold plating defects in general. Next electroless nickel immersion gold plating defects were analysed to study appearances and learn about their detectability. Boards for these analyses were provided by ASPIS SME partners; Graphic plc, Merlin Circuits Technology and Global Interconnection Services.
From this initial work it was found that capacitive measuring principles appeared to have the most potential (based on what is currently known and understood about black pad defects). A plate capacitor was created by using printed circuit board pads as bottom electrodes together with a top electrode (+ dielectric material) which functioned as the probe of the sensor tool. An experimental research setup was built to analyse all kinds of different ENIG boards and ENIG plated samples received from ASPIS RTD partner ITRI. Significant differences in electrical capacitance were demonstrated based on samples which were made with different plating parameter settings and the use of statistical techniques. A description of the concept and the specification of the different building blocks resulted in deliverable 3.2. This deliverable also contained an evaluation of potential inspection strategies. Since the underlying mechanisms of black pad failures were still not fully understood at this stage and the known indicators and failure modes were not exclusively related to this particular defect, it was decided to develop an analysis tool that could help in understanding the defects based on manufacturing results ie a tool that could analyse boards coming from different stages of the plating process. The aim of this was to systematically build up knowledge and evidence to correlate indicators and failure modes to the black pad or other possible surfaced related defects and to detect early possible drifts or shifts in the materials, plating process or products. The SME partners of the project preferred the development of a semi-automatic benchtop version of the tool, a tool which typically could be used in a QA department. Deliverable 3.3 was a built version of such a tool and this deliverable was demonstrated on film during the mid-term meeting in Brussels. Both the use of capacitive measurements to inspect printed circuit board coatings and the anticipated integral approach, appeared to be novel and these have been registered by the inventers for a possible patent application.
In the second phase of the project the built benchtop tool was further modified and optimised based on extensive test work conducted on the test boards which came from other work packages and the boards produced by the SME partners of the project. After finalising and validating a major modification, the whole setup was shipped to SME partner Graphic plc. for a field test in a printed circuit board manufacturing environment. In the last part of the project the work was mainly focussed on understanding how the encountered variations in outcome was built up, especially the spread over time. Cost modelling work was also conducted to investigate the economic side of the different inspection approaches, concluding the TNO inputs to this work package. The technical and economic evaluations resulted in deliverable 3.4; Technical & Economic Validation (LCCA), which assessed whether or not it would be possible to use the device as a standard test method and to identify and involve manufacturers for the potential development of a commercial model that could be sold to PCB fabricators and assemblers.
Modifications
After working with the inspection setup for a certain period, it became clear that the probe head alignment procedure, the probe head design and the placement of a separate dielectric layer underneath the probing pin had to be improved in order to improve the overall performance of the testing results.
The probe head positioning and alignment were initially performed without any means. It appeared to be rather difficult to position and align the head onto mid-sized SMT solder pads. Since it was also the plan to aim for use on even smaller solder pads, it was decided to add a digital microscope camera (10 – 200 x, 2.0 M pixel) to the system which presented a magnified live view on the computer screen.
The probe head was redesigned to make it possible to place the counter-electrode closer to the probing electrode which allowed the probing of smaller pads. The modified design can probe pads of approximately 1 x 0.5 mm when using an anode probe diameter of 0.5 mm. The other reason to redesign the probing head was to change the approaching angle of the contra probing pin to a vertical movement. The vertical movement improved the repeatability and prevented pin sliding, which could cause damage to the solder pads. Spring properties were added to the counter electrode by selecting the material and designing the probe in such a way that the placement force would make the probe function in the elastic regime.
In an earlier phase of the project, different materials including paper, polyimide (PI), polyethylene naphthalate (PEN) and polycarbonate (PC) were tested as dielectric foil. However, these materials all had rather similar low dielectric constants and all appeared to be difficult to attach underneath the probing pin in a reproducible manner. Hafnium oxide (HfO2) and an acrylated urethane were investigated as alternatives. Hafnium oxide has an interesting dielectric constant of around 25 and can be applied by using CVD (chemical vapour deposition). Hafnium oxide was tested by depositing the material underneath copper probe pins. These tests were not successful because the accurately machined copper surfaces of the probe pins were attacked during one of the process steps. Tests conducted with the acrylated urethane were more successful. This material was used as a UV adhesive which could be accurately machined after curing until for instance a layer thickness of 20 μm was achieved. During the rest of the project, acrylated urethane was used as dielectric material underneath the probe pins. Repeatable result were achieved with thicknesses of 50, 42.5 and 30 μm.
The modified measuring setup was tested by measuring samples produced by ITRI (ITRI S5 coupon) and comparing the results with results measured with the earlier setup. This was done by using a multi-vari analysis and hypothesis testing (F-tests) process, which proved that the new system variation during repeated testing could meet the performance of the old setup. Stability tests over time were conducted with different probing pins having a 50, 42.5 or 30 μm thick dielectric layers. Capacitance measurements were done during the 1st measurement and after 100, 300 and 1000 times repositioning on the pad. The results were analysed by using Multi-vari analysis. The surface of the probe with a 42.5 μm thick dielectric material was also inspected by using Scanning Electron Microscope (SEM). This sample showed no visible signs of major degradation after 1000 measurements.
Within pad variation
To investigate possible within pad variations, test coupon S5 was systematically measured on different locations. Sample S5 which, was received from RTD partner ITRI, was plated with the ITRI solution plus an acetic acid buffer formulation (pH 6.0) for 12 minutes at 93.8°C. Capacitance measurements were done in a 7 x 3 matrix with a 2 mm pitch in all directions. The measured capacitance values were graphically (height plots) and statistically correlated with the amount of nickel, phosphorus and gold at the surface. The elements were measured by using Scanning Electron Microscope (SEM) with Energy Dispersive X-ray (EDX) and an acceleration voltage of 15 kV. After removal of one outlier in the EDX results, the average capacitance value was 3.6E-13 F. The within pad variation appeared to have a range of 3.4E-14 F. The measured nickel content indicated a probable to very probable negative linear correlation with the measured capacitance values. The Pearson correlation coefficient (r) was between -0.7 and -0.9. The gold content indicated a probable to very probable positive linear correlation with the measured capacitance values. The Pearson correlation coefficient (r) was between 0.8 and 0.9. The phosphorus content appeared to have no correlation with the measured capacitance values (r<0.5).
Positioning sensitivity
To investigate the probe positioning sensitivity, the S5 test coupons received from ITRI were analysed when repositioning the measuring probe around one nominal centre position. Measurements were conducted on nine positions in a 3 x 3 matrix with a pitch of 0.1 mm (n = 5). The results were analysed by a height plot and ANOVA (Analysis of Variance). The average capacitance was 3.6E-13 F with a spread of 1.1E-14 which is a spread of approximately 3 % around the average value. The height plot revealed, and ANOVA proved, that there was also an orientation effect. The variation within rows appeared to be significant while the variation between rows appeared to be insignificant. These tests were repeated on various other test boards during the course of the project.
Skip plating detection
An electroless nickel immersion gold board received from SME partner Graphic plc was analysed within work package 6 and diagnosed with skip plating. Skip plating can be described as the situation in which gold was plated directly onto the copper layer due to the absence of a nickel layer. The cause of this failure mode is often an incomplete or missing activation step. The visual appearance of skip plated pads can often be black as well. Within work package 3, this board was systematically measured to detect any deviations due to issues underneath the gold layer. It was found that footprints which exhibit skip plating might display up to 5E-13 F in capacitance variation.
Field tests at SME partner Graphic plc
From the field testing conducted at SME partner Graphic plc it was found that the capacitance values contained both a cyclic behaviour and drifts. Especially, the zero measurements revealed very clearly this unwanted behaviour. The inspection setup was installed in a climate controlled QA room with the air conditioning unit set at 20°C and 50 %RH. Data logged by the air conditioning units showed: 20.06°C (19.5 – 20.6°C), 46.4 %RH (43 – 50 %RH). An extended test was run during the night to determine whether or not the capacitance values would stabilise during the night time. This test also confirmed the cyclic behaviour and the drift over time. Next, a temperature and humidity sensor was installed on the printed circuit board table of the test setup. Correlating the outcome of these measurements with the capacitance values, it was found that the capacitance values were following the changes in the ambient humidity. This indicated that when measuring in the femto-farad capacitance range, there might be a need to operate under a more stringent controlled humidity conditions.
Work Package 4: Aqueous Plating
At the inception of the project there were two main prongs to Work Package 4. These were to try to develop improved solderable coating systems based upon the existing Electroless Nickel – Immersion Gold process (ENIG), and to develop alternatives with the same functionality, but that did not suffer from the same level of severe defects which can afflict this otherwise excellent finish.
Earlier work by various researchers and the investigations conducted as part of WP2 in the ASPIS project had highlighted a number of situations that could cause or aggravate conditions such as black pad. Armed with this information, potential changes to the materials and processes were discussed and research strategies were formulated.
The direction of the research performed in WP4 was heavily influenced, and aided, by the industrial partners, who each provided significant advice and inputs. Based upon their own knowledge, the conclusions drawn in WP2 and by the early results obtained in WP4, the industrial partners advised that they strongly recommended that the investigation should concentrate on the development of improved ENIG technologies. The reasoning behind this decision was sound. ENIG is a dual-purpose coating in that the nickel remains in the joint structure after soldering and acts as a barrier layer. The combination of good solderability and low inter-diffusion rates of copper and tin was found to be unique amongst common metals when the previous historic literature was studied. The decision to abandon further work on alternative materials and processes was further vindicated by economics. The industrial partners strongly felt that the costs associated with implementing alternative materials and processes would prevent high take-up rates. The work on alternative materials and processes was therefore abandoned, though some success was achieved using a screen printed gel method for depositing gold coatings, it being thought that this could avoid certain detrimental electrochemical effects associated with conventional immersion gold plating. A report summarising some of the alternative materials and process findings and considerations was issued as Deliverable D4.2.
Modified ENIG Considerations
The features of the failure mechanisms, in particular Black Pad, were considered and it was hypothesised how they could be avoided or suppressed. One mechanism in particular, the excessive corrosion of nickel during the Immersion Gold process, was considered. However, corrosion of the nickel is a necessary part of the IG deposition. Modifying the Immersion Gold process to reduce the risk of excessive corrosion occurring is likely to generate even more defects, as the activity is reduced and would also slow the process down.
The normal cellular morphology of the EN deposits lends itself to differential electrochemical effects due to the deep interstices between neighbouring growth cells. They also increase the risk of a through-coating pore persisting. It was therefore proposed that modifying the morphology and microstructure of the Electroless Nickel deposits could have a beneficial effect if it could be achieved. This then formed the basis of the research performed in the second half of the ASPIS project.
Plating Facilities
The first task that was tackled was to assemble and build the equipment needed to perform the proposed plating experiments. Early stage development work was performed using laboratory glassware, but this is unsuitable for scaling up to plate large numbers of even moderately-sized printed circuit boards. Methods of agitation and temperature control were tested and their performance evaluated.
When using laboratory glassware the method of heating is critical to avoid temperature differences in the plating solution. Rather than heating the glassware directly, it is better to use a bain-marie configuration. Various liquids were used as heat transfer mediums to try to achieve better temperature control and eliminate hot and cold spots, including various oils and glycerol. Whilst these did improve temperature control, there was a high risk of contamination of the plating solution and causing it to collapse. Eventually it was decided that water would have to be used, even though typical plating temperatures would approach 100oC.
It was found that mechanical agitation was not sufficient to achieve consistent results across a sample due to incomplete mixing of the plating solution and the possibility of hot spots arising. Air agitation was found to give far better results and it was shown that this method had the additional benefit of enhancing the bath stability.
A much larger prototyping facility was constructed, based around a 20 litre polypropylene tank. Air agitation was provided by PVC piping, though the original pipes had to be replaced by a more temperature-resistant grade, when it was found that they warped at temperatures above 60oC. Figure 1 shows a schematic of the tank used for the Electroless Nickel deposition part of the work.
Baseline Solutions
Standard Electroless Nickel (EN) and Immersion Gold (IG) processes were required for benchmarking purposes and to serve as starting points for investigations into improving the existing technology. The initial EN solution was identified as the “ITRI” solution and was based upon a nickel sulphate – sodium hypophosphite chemistry. The IG solution was a potassium gold cyanide-based solution which was also used for the investigative work in WP2.
Both of these solutions were used widely in the development and validation of the prototype plating facilities, for evaluations of the pre- and post-plating treatments and selection of the copper substrate preparation methods. A large number of boards using variations of these coatings were produced and used to develop evaluation methods for WP3 (screening tool) and WP6 (validation) in the project.
Figure 1. The electroless nickel tank used for development work
Substrates and Preparation
It was inconceivable that any other substrate material other than copper would be used. On PCBs, the nickel would be deposited onto electrolytically formed copper surfaces. The use of PCB samples for early stage development work was undesirable because of the size requirements and cost, so high purity copper discs were used as a substrate. Early experiments on brass and copper coupons confirmed that the coating morphology was dependent on the condition of the copper substrate surface and that the cold-rolled coupon surfaces produced very different nickel morphologies to electrolytic copper ones. An electrolytic copper process was therefore added.
It is common practice and, indeed necessary, to etch the electrolytic copper and to activate it for Electroless Nickel deposition to occur. Various etches were evaluated to determine their effect on the copper and determine whether this could in turn could have an effect on the character of the nickel deposits. The different etches did produce effects which were deemed important. An etch based upon ammonium persulphate was deemed the most suitable for the project work.
The nucleation of the nickel deposits is an important factor in the growth of the coating and its eventual morphology. The nucleation of the EN coating requires activation and so this process was studied to determine whether this could be used to influence the deposit characteristics. Activation is achieved using solutions containing palladium chloride; no other viable activating agent being known. Changing the concentration of palladium chloride had little effect on the character of the resulting nickel deposits unless the changes to the activating solution were very large. In such circumstances, the activation treatment had a detrimental effect on the deposit and it was realised that this was not a route by which a more desirable nickel morphology could be achieved.
Another way to activate the copper surface is to do so galvanically. This method would be unsuitable for an industrial PCB plating process but can be performed in the laboratory. This was done and the resulting nickel deposit was found to have a smooth, planar morphology with no cellular growth. This indicated that such growth was possible, however, a way to suppress the cellular growth would need to be found.
Growth Modification
One way to modify the growth morphology is to change the nucleation behaviour of the nickel and to then control the points at which depositing atoms can attach themselves to the growth interfaces. Substances which might absorb onto the low energy nucleation sites on the copper substrate and/or which could reside at some of the growth interfaces could cause a change in morphology. Such substances are generally known in the electroplating industry but have, to our knowledge, never been used to modify growth in Electroless Nickel systems. The substances are known as “Brighteners” and are categorised as “Class I” and “Class II”. Whilst they can be effective singly, when a Class I brightener is used alongside a Class II one, the synergistic effect can be far more effective.
“ITRI” Solution Stability
During development work it became apparent that the “ITRI” solution was very stable to changes. This was first noticed when attempts were made to change the phosphorous content of the nickel deposits and also when additions were made to simulate an out-of-control bath chemistry. This stability extended to additions of brighteners and it was deemed that the “ITRI” solution was unsuitable for modification. An alternative Electroless Nickel chemistry was therefore adopted.
The “XL” Formulation
The second EN chemistry was identified as “XL”. Its basic formulation and key operating parameters are shown in Table 1.
NiSO4.H2O 40.9 g•l-1
NaH2PO2 40.9 g•l-1
Glycolic Acid 14-20 ml•l-1
Lactic Acid 12-16 ml•l-1
Acetic Acid 20-25 ml•l-1
Lead as metal 3.5 g•l-1
pH (preferred) 5.0
Temperature 90 – 95oC
Table 1. The formulation and operating conditions of the “XL” solution
With the XL solution, it was possible to modify the chemistry and plating conditions. Various packages consisting of different combinations of Class I and Class II brighteners were added to the basic formulation and the effects assessed. It was found that saccharin was the best Class I brightener. The best results were obtained with a Class II Brightener of either propargyl alcohol or phthalimide. Details of the recommended brightener packages are shown in Table 2.
Package 1 Package 2
Class I 2 x 10-2M Saccharin 2 x 10-2M Saccharin
Class II 1 x 10-3M Propargyl Alcohol 1 x 10-3M Phthalimide
Table 2. The Brightener packages found to be most effective at modifying the EN deposit morphology
Conventional Coating Morphologies
The brighteners added to the “XL” formulation reduced or suppressed the development of the cellular structures that formed during conventional Electroless Nickel deposition. The conventional deposits consist of neighbouring columnar structures, each with a rounded top. This forms a valley between the cell and its neighbours and these have been shown to sometimes contain deep crevices that can penetrate through the entire thickness of the nickel coating. This results in coating porosity which degrades solderability and can also allow copper migration; a secondary failure mechanism. The cellular microstructure can also affect the nickel corrosion that occurs during the immersion gold process. The structures could allow concentration cell effects to arise and the intercellular composition of the nickel-phosphorous coatings can vary, thus meaning the centre of the cells have different resistance to the corrosion mechanism to the intercellular regions.
Figure 2. A conventional ENIG coatings showing deep interstices in the intercellular regions
Modified Electroless Nickel Morphology
The use of brighteners changes the morphology of the Electroless Nickel deposits. In some cases, especially in thin coatings, the effect is to reduce the cellular character of the deposits. In such cases it is possible to make out features that appear to resemble cellular growth, but the intercellular regions do not contain deep valleys or interstices. Thicker coatings can entirely suppress the cellular growth leaving no trace of it on the outer surface. The modified EN deposits can be seen to consist of very fine equiaxed grains which have been determined to be amorphous.
Figure 3. Modified EN coatings showing residual cellular features in a 6 m thick coating
The benefits of this microstructure are that there are no deep crevices or interstices which should reduce the risk of differential electrochemical behaviour in different regions across the surface of the nickel during Immersion Gold deposition. Also, in the absence of these deep penetrating boundaries there is a much reduced risk of there being through-coating pores and so failures associated with such defects are reduced too. Coatings have been produced that do contain pores, but these are very obvious to detect and so the issue would be picked up quickly in a production environment.
The lack of intercellular boundaries may have additional benefits. The minimum EN thickness in an ENIG solderable coating, as specified by the IPC, is 3 m. In part, this is derived from the minimum growth required to ensure that neighbouring growth cells in conventional EN converge and meet, thereby closing any intercellular pores. Using the modified EN process, as described here, that level of growth is not required. It may therefore be possible to use a thinner nickel coating, thereby saving materials costs and reducing the plating time.
WP3 and WP6 Support
As part of the WP4 work programme, a large number of samples were produced to assist the activities of WP3 and WP6. These samples encompassed all of the different stages of the work performed in WP4, including the examples of modified EN using the brightener packages detailed above. In all cases, many different variables were represented in the samples, including coating thickness, phosphorus content and other operator determined variables. Unfortunately, the development of the latest generation of modified boards occurred towards the end of the project as a whole and the long-term tests such as storage life retention and thermo-mechanical reliability have not yet been completed at the time of the preparation of this report. These activities arel continuing at ITRI after the official conclusion of the project, as it is believed that the benefits of the modified EN coatings, which can be produced with minimal changes to existing facilities, may make them attractive to PCB fabricators.
Work Package 5: Ionic Liquid Plating
Introduction
The performance of printed circuit board (PCB) solderable finishes is one of the principal factors affecting the reliability of most electronic devices. These surfaces act as a reliable, solderable surface for forming the connections to the various components that need to be attached to the PCB. The most popular choice for high value electronics is the electroless nickel immersion gold (ENIG) surface coating. This is because the ENIG finish offers excellent solderability that is maintained during prolonged storage times. However, there is one recognised problem that sometimes occurs with ENIG coatings and this is known as “black pad”. Despite being identified more than ten years ago the mechanisms that cause this phenomenon were still poorly understood prior to the work undertaken in the Aspis project. Black pad manifests itself due to excessive corrosion of the nickel coating during the immersion gold process. The main problem with this is that the surface of the coating is still pristine and identification of defective boards is not possible by visual inspection. When the boards are assembled with electronic components, the presence of black pad can result in a subsequent brittle fracture of the resulting solder joints and thus devices failure. The ASPIS consortium was formed to investigate the mechanism of black pad formation, development of a testing procedure to identify black pad and finding alternative solution chemistries for ENIG deposition such that black pad could be eliminated.
WP5 focused on the design of ENIG coatings obtained partially or wholly from alternative solvents based on ionic liquids. Ionic liquids are liquids that are composed exclusively of ions and they possess many unusual properties such as high solubilities of metal salts including metal oxides. Electrolytic, electroless and immersion deposits from ionic liquids can have considerably different properties when compared to their aqueous-based counterparts.
In order to structure the research in this work package, a series of tasks were followed:
• Development of electroless nickel (EN) and immersion gold (IG) formulations from ionic liquids
• Fundamental characterisation of metal/metal interfaces
• Evaluation of new ENIG coatings for the PCB industry using standard solderability/wetting balance evaluations.
The investigations of electroless nickel coatings contains two essential steps; the generation of a palladium activated Ni surface, which is then subsequently followed by the EN deposition proper.
Development of Electroless Ni Process from Ionic Liquids
Bare copper is not in itself sufficient for electroless nickel plating, a highly adsorbing surface such as Pd is required to catalyse the reaction between the reducing agent and the metal salt during the initial stages of the deposition process. In the conventional aqueous process, this is obtained by immersing the copper-clad printed circuit board material into a bath containing a Pd salt. Small particles of Pd are then immersion deposited onto the surface of the copper, which act as seeds in the electroless Ni process.
Methods of replicating this approach were attempted from ionic liquids instead of aqueous media. A study of immersion coatings of Pd onto copper sheet from ILs was undertaken using PdCl2 as the Pd source in a 0.1 M solution in Ethaline 200 for 20 minutes at 50ᵒC. Pd was deposited as a grey metallic coating with good overall coverage of the Cu surface. EDAX analysis showed the presence of Pd on the surface of the Cu, with Cu as the majority element of 87.1 wt% Cu compared to 12.9 wt% Pd, showing that the Pd coating was very thin. AFM analysis confirmed that this coating was approximately 250 nm thick.
In addition to using an analogous Pd process from ionic liquids, an alternative Ni activation process was also developed. This took advantage of the fact that the standard reduction potentials for Cu and Ni are switched when compared to the aqueous process. In Ethaline 200, Ni is more noble than Cu and, as a consequence, an immersion coating of Ni on Cu is obtained when a Cu PCB token is immersed in a 0.1 M solution of NiCl2 in Ethaline 200 at temperatures above 100ᵒC. Quartz crystal microbalance analysis showed that there was a net loss in mass on the surface as the Cu was oxidised to Cu+, whereas Ni metal was reduced from Ni2+. As a consequence, there was a net loss of mass from the surface. SEM analysis showed that the Ni deposited as small particles onto the Cu surface. Both of these Pd and Ni activated samples then successfully had EN deposited onto them from an aqueous EN bath. They behaved very much like those of the standard Pd deposit with comparable deposition rates and compositions.
The next logical step from this point was the development of an electroless nickel bath from an ionic liquid. The use of hypophosphite as the reducing agent was key to the final properties of the coating. When hypophosphite was used as the reducing agent for Ni, a substantial amount of phosphorous was incorporated into the deposit causing a change in the structure and behaviour of the deposit. Above P levels of 5 wt%, the structure became amorphous and with increasing P concentrations the deposit became more corrosion resistant.
When NiCl2 and Na2PH2O2 were heated in the ionic liquid Ethaline 200, a spontaneous, particulate deposit started to form, which EDAX analysis showed to be Ni. The turbidity of the Ni solution had been caused by the reduction of the Ni(II) ions in situ by the hypophosphite reducing agent causing the formation of black Ni metal particles. For any electroless process it is essential that reduction of the metal occurs only at an activated surface, otherwise an extremely large amount of metal waste occurs. In this case, the additional aspect of the IL inhibiting the Ni plating produced even greater obstacles to overcome. Various attempts were made to stabilise the plating bath; the addition of Pb2+ ions to the plating solution, reduction of the concentration of reducing agent and Ni salt, plating at lower temperature, addition of water to the plating solution and the addition of an acid source to change the pH of the liquid. Unfortunately, none of this made an impact, these solutions either did not plate onto an activated Cu surface, the solution destabilised, or both. Full details of the experimental work and efforts regarding the development of electroless Ni coatings onto Cu from ionic liquids have been documented in Deliverable 5.3: Catalytic Deposition Methodology.
It was at this stage, with the given inherent problems of developing a suitable electroless Ni plating solution and the success that had been demonstrated when studying immersion gold deposition onto aqueous electroless Ni surfaces that it was agreed with the Aspis partners that the focus of the ongoing efforts should be on the study and development of an immersion Au process from ionic liquids.
Immersion Au deposition from Ionic Liquids onto nickel
The conventional aqueous IG process used acidic media and potassium gold dicyanide as the Au source. Initial efforts were focused on IG coatings from the IL Ethaline 200 onto aqueous EN with the Au salts AuCl, AuCN and KAu(CN)2. The coatings from these three systems were substantially varied. Visually, AuCl produced a dark brown poorly adherent coating, AuCN gave a bright gold coloured finish and KAu(CN)2 gave a slightly duller Au coloured coating. Surface SEM analysis of these three samples showed that two of them produced coatings with very little surface structure, whereas 1 showed a rough particulate surface. This same effect was seen with AFM images recorded across a 2 µm2 footprint. One sample had a surface area difference of 180%, whereas both the other two had surface area differences of 1.9% and 2.65% respectively. The marked difference between these samples showed the dramatic influence that the Au salt used can have on the resulting Au deposit morphology.
Cross section analysis of samples showed that a uniform Au layer was formed on top of the Ni/P layer and, contrary to the conventional aqueous process, there was very limited attack down Ni/P nodule boundaries. The first of the three samples examined showed a dendritic, particulate nature.
Removal of the Au with a cyanide Au stripping solution revealed information regarding the impact on the Ni/P substrate of the IL IG process. In all cases no evidence of the mud cracking associated with black pad was seen. A standard, uniform, nodular EN surface was seen across the whole sample.
Solder wetting balance measurements on two of the three samples showed that the wetting time for the IL samples was both faster than a standard aqueous sample and that they were wetted more reliably. This is a key measure of the solderability of these surface finishes and provides evidence that these coatings could perform as well as, if not better, than conventional aqueous coatings.
Use of Additives in Immersion Au Process from Ionic Liquids
The use of additives in the ionic liquid IG bath was also investigated. The purpose of these additives was to affect the electrochemical behaviour of Ni in the selected ionic liquids. The addition of ethylenediamine to an ionic liquid causes the rate of electrochemical dissolution of Ni to increase dramatically. In addition, a quartz crystal microbalance (QCM) study showed that the addition of 10 mM ethylenediamine to one of the IG bath formulations caused the rate of gold coating to increase dramatically as well. (This will be referred to as solution/sample four below).
IG coatings from sample 4 produced a bright lustrous coating. SEM analysis showed that the coating was uniform across the EN surface with very little defects. AFM analysis also showed that the coating had low surface area difference. Stripping of the Au surface revealed that the underlying Ni/P nodules had been levelled as indicated by a reduction in the value of Ra with respect to the original virgin Ni/P surface. This is extremely exciting as the formation of black pad is thought to originate along nodule boundaries. These results confirmed that the Ni was consumed preferentially on top of the Ni nodules, potentially eliminating the mechanism by which black pad forms.
AuCl is not stable with respect to ethylenediamine and as such an alternative additive is required that does not oxidise/reduce AuCl and that activates the Ni surface. Iodine was found to do this effectively depending on concentration. A number of different concentrations of I2 were added to an AuCl solution, ranging from 0.125 mM to 5 mM and ENIG samples were produced from these respective solutions. It was found that an evolution in morphology was observed with increasing I2 concentration. At low concentrations of I2 a similar morphology to that of an earlier sample was observed, where it was rough and particulate in nature. As the concentration increased, the surface roughness decreased and particle sizes were smaller until between concentrations of 0.5 – 2 mM I2 a smooth uniform deposit was produced. Above 2 mM I2 an alternative morphology was observed where the Au was deposited as small nodules and some of the EN surface was visible when viewed at high magnification on an SEM. A fifth key formulation was developed of 5 mM AuCl and 1 mM I2 in Ethaline 200.
Ionic liquids have potential as a replacement for the aqueous immersion Au bath in ENIG coating production. A thorough investigation has been undertaken of the various factors that affect the resulting immersion Au deposit morphology and the resulting impact on the underlying electroless Ni layer.
An evolution in morphology was observed with increasing cyanide content in the Au salt used in deposition. Dramatically improved deposits were obtained when an at least one molar equivalent of cyanide to gold was used. In addition, the influence of additives also proved to be vitally important. A number of amine containing compounds were found to improve the otherwise sluggish kinetics of Ni dissolution in ionic liquids leading to the use of ethylendiamine, which produced quicker deposition rates as well as improved ENIG coating morphology. I2 was found to be an effective additive into AuCl solutions, activating the Ni/P substrate while stabilising the AuCl solution in Ethaline 200.
Work Package 6: Validation
The overall aim of work package 6 concerned the validation of improved electroless nickel immersion gold (ENIG) Printed Circuit Board (PCB) coatings and possible alternative coatings for high-end electronic applications. To reach this the following objective were defined:
• Verify the compatibility of improved and alternative coating technologies with soldering assembly methods and practices by assessing yields in lead-free reflow and related soldering operations
• Establish effective storage life and solderability retention characteristics
• Determine long-term reliability by evaluation in normal and hostile operating environments using thermal cycling and related appropriate accelerated testing conditions.
• Validate performance in production trials at partner facilities
During the first phase of this work package the work concentrated on validating the solderability and solderability retention of various printed circuit board finishes applied by the project partners. Other work during this period concerned the definition, building and evaluation of a reliability test setup based on the combination of thermal cycling and vibration testing. During the second phase of the project, newly developed alternative electroless nickel immersion gold finishes were characterised and tested on solderability and the reliability aspects like intermetallic growth and high temperature humidity storage. Failure analysis, to study the black pad failure mode and black pad related failure modes, was a significant part of the work and a continuous activity throughout the whole project phase. All the printed circuit boards for these failure analyses were provided by the Aspis project’s SME printed circuit board manufacturers.
Solderability
The first series of solderability assessments were conducted on as-plated and (accelerated) aged Printed Circuit Board (PCB) samples, which were fabricated at SME project partner Merlin Circuits Technology. The board designs were according to the Atotech Test Board specification ATB v1.1. Boards with standard ENIG (57 nm Au), thin gold (39 nm Au), thick gold (79 nm Au) and ITRI ENIG (70 nm Au) were tested regarding solderability with a MUST-2+ wetting balance (globule method). Solderability testing was done before and after aging at 50°C and 50% RH for 24 hours. The standard ENIG, thin gold and thick gold boards showed unanticipated poor wetting results (large wetting angles and incomplete pad coverage). The measured wetting curves showed unacceptable wetting times and wetting forces, and a large variation between the samples. The ITRI ENIG samples based on a formulation developed within work package 4, showed proper wetting results (wetting angles < 90°, full pad coverage). Nevertheless, 6 out of the 10 wetting curves had wetting times longer than 2 s (up to 12 s). The second series of solderability investigations was conducted on samples from work package 4 (aqueous plated) and work package 5 (ionic liquid plated). The solderabilities of both aqueous and ionic liquid plated samples were found to be good. Aging at 50°C and 50% RH were found to have no noticeable effect on the solderability of samples even after exposures of up to 28 days.
Furthermore, various methods were applied to ENIG-plated circuit boards to determine if it was possible to:
1) Identify problematic coatings immediately after fabrication
2) Produce accelerated ageing effects whereby the useable storage life of ENIG-coated PCBs could be predicted.
As has been mentioned above, coatings developed in WP4 and WP5 of the ASPIS project were found to have experienced no noticeable deterioration in solderability after 28 days of storage at 50°C and 50%RH. Other potential artificial ageing conditions were investigated. These included:
1) Steam Ageing
2) Nitric acid vapour exposure
3) Hydrochloric acid vapour exposure
4) Polysulfide solution treatment
5) Hydrogen sulphide exposure
Conditions 2 to 5 were found to be useful for identifying coating defects, in particular coating porosity and deep intercellular interstices in conventional ENIG coatings. Hydrogen sulphide was especially effective at identifying sub-standard Immersion Gold coatings which rapidly discoloured in regions in which there was insufficient gold. The easiest and most effective test for identifying isolated pores in the gold coating was hydrochloric acid exposure for 2 to 5 minutes. This caused coloured crystals of nickel chloride, copper chloride, or both, to grow on the surface of the affected coating. The polysulfide treatment was the best test for assessing porosity of all types in the nickel coating. However, all four methods were found to be ineffective at accelerating solderability degeneration in a manner that could be used to predict the storage life of good ENIG coatings. Deterioration was either instantaneous, especially for coatings containing defects, or non-existent when high performance gold coatings were exposed.
Steam ageing was found to cause a gradual degradation in the solderability of ENIG coatings as the exposure time increased. The rate of deterioration was very variable. This is almost certainly due to variability in the quality of the ENIG coatings being tested. Standard steam ageing test methods specify exposure times of up to 8 hours. PCBs with thin gold coatings were found to have very short exposure resilience, almost certainly due to the patchiness of some coatings allowing large areas of the underlying nickel to passivate. Gold coatings that had isolated porosity were found to be more resilient. The pad areas directly below the pads deteriorated, but this usually left a large enough area on each pad that could be wet by molten solder. High quality gold coatings had the best solderability retention, but there was still a noticeable deterioration in all coatings after 8 hours of steam ageing. The later modified EN coatings in which the normal cellular microstructure was suppressed were found to have excellent solderability retention during steam ageing treatment.
Reliability
A combined thermal cycling and vibration test with in situ monitoring facilities was developed with the aim to detect solder joint degradation based on the electrical resistance. The development of an evaluation protocol led to the specification of analysis techniques and equipment to be used. Test specifications were based on the selected high-end electronics applications areas ie automotive and aerospace. Vibration tests were carried out to evaluate both the setup and the test board design (ATB v1.1). Test results were obtain from soldered 0603 chip resistors, 0805 chip resistors, Mini-MELF resistors, MELF resistors, 1206 chip resistors and Mini-MELF resistors. These component types were subjected to random vibrations with a maximum frequency of 2 kHz, while the daisy-chain circuit resistances were measured.
During the second phase of the project a new series of Atotech test boards was subjected to high temperature and humidity tests and high temperature aging tests. The Atotech test boards according to specification ATB v1.1 were manufactured at printed circuit board manufacturer and SME project partner Merlin Circuits Technology. These boards were manufactured and shipped to RTD project partner ITRI without solderable solder finishes. At ITRI, the test boards with copper were metallised with the newly developed Standard ITRI and XL electroless nickel immersion gold platings. These plating types were developed by ITRI within work package 4. The electroless nickel / immersion gold plated boards were visually inspected regarding plating quality and for possible plating defects by using optical microscopy.
The thickness of the standard ITRI and XL electroless nickel immersion / gold platings were measured by using X-Ray Fluorescence (XRF). The electroless nickel plating of the XL boards, which were shipped to TNO for further testing, had an average thickness of 5.883 µm with a standard deviation of 1.074 µm (n=13). The immersion gold plating of these boards had an average thickness of 0.095 µm with a standard deviation of 0.022 µm. The standard ITRI plated boards shipped to TNO had an average electroless nickel thickness of 6.710 µm with a standard deviation of 0.812 µm (n=19). The immersion gold plating of these boards had an average thickness of 0.084 µm with a standard deviation of 0.013 µm. The industry standard “IPC-4552 Amendment 1 Specification for Electroless Nickel / Immersion Gold (ENIG) Plating for Printed Circuit Boards” specifies 3 to 6 μm [118.1 to 236.2 μin] for electroless nickel and min. 0.05 μm [1.97 μin] for immersion gold. With respect to this specification, only the XL electroless nickel thickness was slightly above specification. A thinner nickel layer might, regarding the black pad failure mode, probably be more critical, since these layers could become porous and possibly induce hyper-corrosion during immersion gold plating. A thicker electroless nickel layer might, on the other hand, contain higher stresses inside the layer, which could subsequently result in cracking.
In order to further characterise the electroless nickel immersion / gold platings of both the standard ITRI and XL test boards, capacitance measurements were conducted using the inspection tool setup developed within work package 3. The outputs of these capacitance measurements were both graphically and statistically analysed and correlated together with the thickness measurements. A Multi-vari analyses was used to investigate the variation within samples and between samples / over time. These analyses revealed already quite a spread within samples and over time. Correlating the capacitances with the XRF measured layer thicknesses showed both graphically and statically that there was no correlation between the two parameters (r<0.5).
Temperature-humidity tests were conducted on bare boards according the JEDEC 22 A110 standard. Both boards with standard ITRI and XL electroless nickel immersion / gold plating were subjected to 85°C and 85% relative humidity conditions. Samples were taken at t = 0, 250, 500, 750 and 1000 hours. The test was used as an accelerated high temperature humidity storage test. No degradation or defects were found during visual inspections done by using optical microscopy. Furthermore, samples were analysed using Scanning Electron Microscopy (SEM) - Energy Dispersive X-ray (EDX). Elemental analysis was used to investigate whether or not corrosion products were present at the top surface of the electroless nickel immersion / gold layers. Since nickel and not gold is expected to oxidise, oxygen nickel ratios were calculated and compared between the different samples. The EDX element analysis showed, as expected, that nickel, phosphorus and gold were the main elements on the standard ITRI plated test boards, this next to a small amount of oxygen. Also, the standard ITRI sample taken at 1000 hours contained the same elements. The oxygen nickel ratio on the standard ITRI samples increased from 0.012 at t = 0 hours to 0.025 at t = 1000 hours, which meant an increase of 1.3 %. The EDX element analysis on the XL plated test boards showed next to nickel, phosphorus, gold and oxygen, also a certain amount of copper. EDX analyses at different acceleration voltages confirmed that the copper was present at the top surface of the plated layers. Further analysis on sample 36B from the XL series revealed that Cu could be particularly found in the pin holes which were present in the plating layer. The XL sample taken at t = 1000 hours showed the same elements as found on the t = 0 sample. The oxygen nickel ratio on the XL plated samples increased from 0.020 at t = 0 hours to 0.036 at t = 1000 hours, which meant an increase of 1.6 %, and indicated an increase in oxygen nickel ratio comparable to the standard ITRI plated boards.
High temperature aging tests according to the JEDEC Standard 22-A103C Condition B were conducted to study the growth and possible degradation of intermetallic interfaces. Samples were prepared by reflow soldering SnAgCu spheres to the Ball Grid Array (BGA) pads of both Standard ITRI and XL plated test boards. The boards were soldered using a Kester TSF Tacky Flux (type ROL0) and a heating plate at 260°C. The TAL (Time above Liquidus) was 60 seconds. The visually inspected joints appeared to be free from soldering defects and showed proper wetting towards both the Standard ITRI and XL finishes. After that the soldered test boards were aged at 150°C, samples were taken at t = 0, 20, 100, 250 and 500 hours. Cross-sectioning was performed to study the solder joints and interfaces by using a light microscope and Scanning Electron Microscopy (SEM) - Energy Dispersive X-ray (EDX). Some joints exhibited small cracks in the solder surface due to shrinkage during cooling down. A number of boards also showed voids at the interface at a rather low and acceptable voiding rate. On nearly all samples, the intermetallic layer was present over the whole length of the interface. Only a number of voids seems to have locally prevented wetting and subsequently the formation of an intermetallic layer. Voiding did occur on both the Standard ITRI and the XL plated test board. The intermetallic layers on both board types showed an intermetallic layer that consisted of a thin uniform layer with a scallop structure on top. Visual inspections could not detect any irregularities in intermetallic growth or the introduction of defects. Thickness measurements (n = 20) were conducted to measure the uniform intermetallic layer thickness (tu), the scallop intermetallic layer thickness (ts) and calculated the effective intermetallic layer thickness (teff). The outcomes of these measurements were both graphically and statistically analysed. An ANOVA (Analysis of Variance) revealed that there were significant differences (α = 0.05) in both the uniform intermetallic layer thickness (tu) over time and the scallop intermetallic layer thickness (ts) over time. Furthermore, there was no evidence for intermetallic growth differences between the Standard ITRI and XL plated boards.
The robustness of the XL plating processes was investigated by attempting to deliberately induce the black pad failure mode. Attempts were made to induce hyper corrosion by systematically reducing the thickness of the electroless nickel layer. This was with the aim of creating a porous nickel layer, layers which tend to be more prone to hyper-corrosion during immersion gold plating. Soldering tests showed that even the boards with the lowest nickel layer thicknesses (1 µm), did not result in black pad failures. A series of solder joints did show deviations in wetting behaviour, but no de-wetting or wetting behaviour that could be linked to black pad. Cross-sections of these thin nickel layer joints did not show the typical corrosion spikes, mud cracks or a phosphorus enriched black band along the interface.
Failure analysis
To study the black pad failure mode and black pad related failure modes, failure analyses were conducted on both bare and assembled printed circuit boards. The boards analysed within work package 6 were provided by Aspis SME companies Merlin Circuit Technology, Graphic plc, Global Interconnection Services and Liad.
The electroless nickel / immersion gold plated printed circuit boards provided by the Dutch SME company Liad failed in the field on electrical failures and sometimes had a low drop impact resistance. Failure analysis was conducted using microscopic analyses as well as Scanning Electron Microscopy (SEM) - Energy Dispersive X-ray (EDX). Bare boards, cross-sections of both failed and intact joints and the fractured surfaces were analysed. A Ball Grid Array (BGA) on the board appeared to be the critical location exhibiting cracks at the board side. The fracture occurred between Ni-P and the intermetallic layer, which made it possible to analyse the phosphorus content at the surface, which appeared to be around 15 wt.%. Cross-sections showed the typical corrosion spikes at the topside of the nickel layer and at some joints the phosphorus-enriched black band as well. Furthermore, copper was again detected at the surface of pads on a bare board. Still not understood are the presence of aluminium and silicon at some locations.
The failed electroless nickel / immersion gold plated printed circuit boards provided by Graphic plc contained at some footprints a number of black pads. Scanning Electron Microscopy (SEM) - Energy Dispersive X-ray (EDX) was used to study both the morphology and composition of black, greyish and normal bright looking solder pads. The EDX elemental analyses proved that nickel was missing underneath the black and greyish pads. The strongly deviating morphology at the black pads appeared to be a result of copper being directly attacked during the plating processes. In both the literature and the industry this failure mode is called skipped plating.
The electroless nickel / immersion gold plated printed circuit boards provided by Global Interconnection Services gave problems during pull tests conducted after the wire bonding operation. Examination with Scanning Electron Microscopy (SEM) showed that the electroless nickel / immersion gold plating exhibited severe cracks at the top surface along the grain boundaries. Cross-sectioning showed that these cracks were already present in the nickel layer before plating the gold layer. On certain locations, the cracks were reaching to the copper layer. These severe discontinuations in the nickel layer might probably have influenced the ultrasonic energy distribution during the wire bonding process.
The first electroless nickel / immersion gold plated printed circuit board provided by Merlin Circuit Technology came from the field and exhibited pads with non- / de-wetted areas. Cross-sectioning together with light microscopy and Scanning Electron Microscopy (SEM) - Energy Dispersive X-ray (EDX) were used to systematically analyse properly, half and non- / de-wetted areas. These conditions were all present on the same printed circuit assembly, sometimes even on neighbouring pads. This analysis revealed that non- / de-wetting could be related to the copper condition underneath the nickel layer. A height difference in the copper layer morphology caused stresses and a height difference in the nickel layer. This particular location determined the boundary to where the pad could be wetted. The remaining part of the pad had the typical black corroded nickel appearance.
The second electroless nickel / immersion gold plated printed circuit board provided by Merlin Circuits Technology came from the printed circuit board manufacturing process. This board was half stripped regarding the immersion gold layer. The stripped pads were black / greyish coloured. Cross-sectioning together with light microscopy and Scanning Electron Microscopy (SEM) - Energy Dispersive X-ray (EDX) were used to systematically analyse a SO (Small Outline) footprint which was half stripped. These analyses revealed that the results of excessive nickel corrosion were present underneath the immersion gold layer. The nickel layer itself had a thickness which exceeded the common maximum industry specification of 6 µm. Cracking together with excessive nickel corrosion was found right at the corner of a pad, a location with probably a concentration of the highest stresses.
The third electroless nickel / immersion gold plated printed circuit board provided by Merlin Circuits Technology came from the field with an accompanying failure report. Since for this board a de-wetting issue with a defective rate of 44 % and an Au layer thickness above specification were mentioned in the problem statement, it was decided to: 1) analyse the pad metallisation thicknesses by using XRF, 2) conduct soldering experiments using different times above liquidus (TAL) and 3) conduct SEM analyses on cross-sectioned solder joints.
Potential Impact:
Introduction
The potential impact of the work carried out in the Aspis project varies between work packages and is detailed for each one below. It is clear that the fundamental mechanistic work undertaken by LIOC has, for the first time, revealed the key causes of black pad and related reliability issues in ENIG coatings. The use of this knowledge will enable European PCB producers to adjust their processes to avoid conditions likely to cause the problem. The demonstration of a non-destructive test methodology and equipment also offers potential for the commercialisation of a device that could be used by PCB fabricators to confirm that the printed circuit boards they are producing are free of the interfacial integrity problems that are a symptom of potential reliability problems. The formulation of a range of novel gold deposition chemistries based on the use of ionic liquids that can give enhanced performance over conventional approaches, also offers the potential for a new commercially viable alternative to the existing aqueous based chemistries. The work to develop alternative aqueous chemistries with enhanced performance proved to be more difficult and confirmed that, as had been found by existing suppliers of such products, that the development of a new aqueous ENIG processes that didn’t exhibit reliability problems was extremely difficult and at the end of the project further work was still needed in this area. At the time of writing this work was showing some very positive initial results and further performance testing was continuing. It also has to be reported that in the five year period from when the Aspis project proposal was first developed to the recent completion, there has been a move by the PCB industry away from the use of ENIG finishes to other new types of solderable coatings which do not exhibit the typical ENIG problems. This was largely due to the fact that the industry’s chemistry suppliers were unable to successfully solve the black pad issue.
Impact WP2: Fundamental Research on ENIG Mechanisms (LIOC)
The specificity of the “black pad” problem is that the incidence of it is very sporadic and its occurrence is very unpredictable, making it a very difficult problem to address. However, the major advances in understanding achieved as a result of the fundamental research work carried out on the mechanisms of black pad formation in ENIG coating (WP 2) promise to have major implications for both the PCB manufacturing industry and the project’s RTD providers.
One of the most important outcomes of the Aspis work was the procuration and extension of knowledge in the area of hyper-corrosion of EN coatings in IG deposition solutions during Au layer formation. The results of the fundamental research into ENIG failure mechanisms, which was carried out during the implementation of ASPIS project objectives, indicated that the “black pad” phenomenon was related to the specific structure of the EN coatings and some peculiarities associated with the IG deposition process. It has been established that inadequate preparation of the base metal (Cu) surface results in deposition of EN layers with nodular morphology containing serious structural defects, such as pores and cavities, which further leads to a direct contact between the substrate (Cu) and the IG solution. The source of Cu contamination and the role of Cu in the formation of black pad has been idenitifed and disclosed for the first time. It has also been demonstrated that Au layers, deposited in IG solutions contaminated with Cu ions, are highly porous and loosely adhering to the underlying EN coatings. Consequently, such Au deposits cannot adequately protect EN surfaces from oxidation and could eventually lead to “black pad” failure of solder joints. Fundamental knowledge developed in WP 2 of the Aspis project has enabled to negated the widespread point of view that it is impossible to reproduce the “black pad” effect in the laboratory applying pure chemicals. Another key development that will be of use to the PCB industry was the elaboration of a new method for determining the porosity of EN and IG coatings. The results obtained will also help to enable and guide the focus of more profound research works in this field in the future.
The fundamental mechanistic knowledge gained in WP 2 was also useful to, and applied by, the project’s RTD partners in prosecution of their tasks in creating the ENIG screening tool and also for developing alternative ENIG coating technologies.
The knowledge data and experience gained and accumulated in WP 2 also enabled the establishment of industrial design guidelines for the implementation of processes that are able to minimise or avoid the likelihood of black pad formation. Specifically, these guidelines provide advice which regulates the base metal copper preparation process before EN deposition, including the etching depth, EN and IG deposition process parameters, including the IG solution tolerance to the presence of copper ions. The latter findings will help manufacturers to minimize the risk of ENIG-related failure problems occurring in their printed circuit boards. A methodology for investigation of ENIG related problems has been elaborated, which includes the use of a voltammetric examination of the porosity of EN and IG layers, measurement of IG layer composition and thickness, chemical analysis of the boundary layer between IG and EN. A suite of techniques suitable for the mentioned analysis has been developed and proposed, including SEM and TEM with EDS application, as well XPS with ion sputtering capability. The implementation of both structural and analytical examination approaches provides the possibility to proceed with the examination of defective PCB samples directly in a PCB’s manufacturer’s site, and not just in an external RTD institution. The implementation of this methodology and analysis of the results should lead to superior PCB production processes, elimination or at least a significant reduction in the occurrence of black pad, as well as enhanced reliability performance in assembled printed circuit boards, increased confidence in the use of ENIG as a high performance solderable finish and reduced field failures in service. The methods developed will also accelerate the outcomes of ongoing fundamental investigations in this area and will enable the process chemical suppliers to confidently develop improved performance coatings.
The major output from the work performed in WP2 was a methodology for the minimization or avoidance of the risk related to ENIG reliability problems such as black pad. As PCB reliability problems are of a great importance and actuality to all of the PCB manufacturing industry, the outcomes of the project will lead to a reduction of “black pad” cases in the electronics industry, as well to superior production yields, higher quality and better reliability performance of its products. The deeper and enhanced understanding of the principal causes of ENIG failure will enable the project’s SME partners to implement design rules in order to lower the risk and, in general, to avoid the appearance of the “black pad” phenomenon in their manufacturing practices.
The particular work was done with characterization of affected PCB’s supplied by project’s SME PCB fabricator partners, which enabled information to be gained regarding the nature of the problematic ENIG coatings, as well a comparison of the advantages and disadvantages of different ENIG deposition technologies and the resolution of some problems experienced by the PCB provides. The results from the analysis of problematic PCB samples have validated the fundamental knowledge developed in WP2, and will allow for RTD performers to intelligently target improvements in any future related work that they undertake.
One of the additional outcomes from the work package 2 activities is the gained understanding of how complicated the problem of “black pad” actually is and how many different factors can influence it. The results obtained in this work will be exploited for the future fundamental studies related to ENIG problem. The potential areas of the future fundamental work in ENIG-related problems are the following: deposition of alternative EN coatings with nodular-less structure, the search of new copper activation technologies before EN deposition, ie that exclude the use of an expensive Pd salt application step, new technologies of thinner and, at the same time, less porous gold layer deposition. The latter two technologies could lead to a significant reduction in materials costs, but at the same time they are also required to offer longer term reliability enhancements. A competitive advantage will be received by reduction of resource use as well by using more environmentally friendly alternative solutions. The technologies involved in applying these alternatives would also give the European electronics industry a new route for the manufacture of the new generations of PCBs.
The socio-economic impact of the knowledge generated in work package 2 is related with the outcomes which are of a great significance to the European PCB industry. The ability to enhance the reliability of its products will improve the international competitiveness of European PCB fabricators, thereby helping to ensure the successful future of a strategically important industrial sector in the face of intense international competition. This in turn will help to maintain employment in both the European PCB industry and with its suppliers.
A key aspect of the ASPIS project has also been to deliver the results of the project and the expected outcomes to a broader number of final users. Dissemination activities of the results gained in work package 2 to date have been as follows: presentations at the International Conferences Chemistry 2011 and Chemistry 2013 in Vilnius, Lithuania, at the 63rd Annual Meeting of The International Society of Electrochemistry, Prague, 2012, the EIPC Winter Conference in Berlin, the EUROCORR Conference in Estoril, Portugal, 2013. Summaries of the key findings and results obtained were published in two Circuit World papers during 2012 and 2013.
Impact WP3: ENIG Screening Tool (TNO)
The most important achievement of work package 3 has been the development and realisation of a non-destructive, flexible and easy to use measuring concept for bare printed circuit boards. This is a truly novel concept which should subsequently be patentable, as initially required by the consortium.
The developed tool is based on the use of the capacitive plate sensor principle. This measuring technique can be used for different purposes such as 1) characterisation at a research and development stage, 2) inspections as acceptance sampling on product (bare board) level during printed circuit board manufacturing and 3) a process control solution which monitors critical to quality indicators during flow production.
During the ASPIS project the tool proved to be very valuable to systematically build up knowledge and evidence to correlate indicators and failure modes to black pad or other possible surface defects. Furthermore, the tool has been extensively used to study the effects of process modifications and to detect possible drifts or shifts in both materials and processes.
Detection and control of variation
A key piece of newly developed knowledge concerns the ability to detect the presence of an immersion gold layer and the ability to correlate the amount of gold in the top layers with capacitance characteristics. Also, the ability to distinguish between layers made with different plating parameter settings, also adds to the new knowledge developed in this work package. Regarding printed circuit boards, crucial aspects were learned about within pad variations, between pad variations and product to product variations. Furthermore, a lot was learned about the new tool’s specific capabilities and system variations (repeatability and reproducibility).
Novel probing head
An important part of the test setup concerned the novel probing head, which proved to be reproducible and still has the potential for further miniaturisation. Another important feature of the head is the fact that it was developed to minimise the risk of damaging the printed circuit board. Furthermore, the development of a procedure to produce accurate probing pins with dielectric material attached underneath the tip was crucial for the performance of the new measuring system.
Cost modelling
The development of a dedicated cost modelling tool to study economic aspects of different inspection strategies is an important non-technical achievement of the ASPIS project. The model concerns a parametric tool which enables the evaluation of different scenarios before adding an extra inspection step to the printed circuit board manufacturing process. Within the project, the inspection step was defined as a sub-process.
Further development and commercialisation
Since a measuring concept has been developed instead of a specific tool, there will be different possibilities for the further development and commercialisation stages of the setup. A standalone bench-top setup is most likely to be the best variant to first further develop the measuring method and generate more knowledge regarding the capacitive behaviour of stacked plated finishes. Since the SME partners in the project are all printed circuit board manufacturers, a company like Atotech would more likely be a future company to develop, for instance, an integrated process control solution which coupled chemistry with equipment. Atotech were an LE partner in the ASPIS project and are suppliers of plating solutions on an global basis.
Impact WP 4: Aqueous Plating (ITRI)
The majority of the work undertaken in WP4 was concentrated on one of the possible routes envisaged at the start of the project; modification of the existing ENIG technology to produce improved coating performance. The decision to do this was made on the advice of the industrial partners after work had been underway in all areas for some time. The reasoning was that it was apparent that manufacturers adopting alternative materials and processes would likely have to spend a significant sum of money and that this would deter them from making the switch. However, the preliminary work on immersion gold deposition from a screen-printed gel was promising and may have some applications outside of the electronics industry. The work undertaken has shed light on the importance of certain process parameters and given a better understanding of how these might influence productivity and product quality.
The identification of the morphology of the Electroless Nickel as the key vehicle for improving the quality of ENIG coatings required that the failure mechanisms be considered and ways to avoid these from occurring be identified. This approach has not, to our knowledge, been proposed previously nor attempted. The use of brightener packages in Electroless Nickel is novel, but builds upon knowledge already used to electroplate nickel. This may make it easier for acceptance in industry and this approach means minimal changes would be required, thus lowering implementation expenditure. The coating morphology is such that the minimum nickel thickness required may be reduced, thus saving materials costs and plating times.
Unfortunately, identification and development of the modified EN coatings has taken longer than originally envisaged in the project plan. The biggest contributor to this was staff turnover as the original key researcher for the work package left ITRI and it took several months for his replacement to be recruited due to the unique requirements of the position. At the official end of the project the overall impact of the work package has been diminished by the slip in timescales. This is because key long-term data could not be completed by the end of the project. This work is continuing at ITRI past the official end date, so that it will be generated. This will be circulated to the other project partners once complete and some of this data has already been provided as an annexe the work package 4 deliverables. The data that has been generated indicates that the general performance of the modified EN coatings is better than that of laboratory and industry-produced conventional ENIG coatings.
WP4 contributed the bulk of test samples used in WP3 and WP6. Both these work packages have investigated and refined various types of test methods and evaluation tools to better characterise ENIG coatings, gauge their quality and highlight defects and coating performance.
It was said on various occasions by different people that WP4 was probably the hardest of the work packages in the ASPIS project, in part because aqueous ENIG technology is mature. The author believes that the work conducting in WP4 has developed a novel improvement that, if the long-term reliability and storage data proves acceptable, will prove interesting to the PCB industry.
Impact WP5: Ionic Liquid Plating (UoL)
Introduction
In electronics assembly, the reliability of the electrical connection between printed circuit boards (PCB) and components is paramount to efficient manufacturing processes and good longevity of the resulting products. One of the methods used to achieve this is the incorporation of a protective surface finish onto the copper tracks on PCBs. The most common of these is the electroless nickel immersion gold (ENIG) finish which is popular due to its high reliability, good planarity, solderability and long storage life.
Despite its popularity, ENIG finishes still have a number of issues that are a cause of major concern, the most important of which is “black pad”. Black pad describes the process of oxidation of the underlying electroless Ni (EN) layer during the immersion Au (IG) process, leading to poor solderability and solder joints that are unreliable and prone to brittle failure. The biggest issue with black pad is its difficulty in identification, since it is not readily visible and yet can be the cause of expensive failures in subsequently assembled equipment. Due to the disrupted surface being covered in a layer of gold, it is very hard to recognise this problem when it occurs, leading to the boards entering use and causing the failure of a large number of electronic items.
Ionic Liquids and the ASPIS Project
The ASPIS project was undertaken to identify the root mechanism for ENIG reliability problems, the key one being so called ‘black pad formation’, to find alternative formulations for elimination of ‘black pad’ and to design and develop a testing methodology to enable identification of PCB displaying black pad before they were used and assembled with components.
WP5 focused on the development of alternative solvent methodologies, namely the use of ionic liquids, for use in the design of an IG plating process which removed the need for acidic media and/or poisonous cyanide salts. The galvanic corrosion of EN nodule boundaries by protons in acidic media has long been associated with black pad formation. Also, the use of traditional cyanide salts was deemed to be increasingly undesirable and any new solution that eliminated their use would be of great commercial interest due to the reduction in processing costs, elimination of hazardous materials, less environmental and waste treatment costs, enhanced operator safety etc.
Ionic liquids have a number of unusual properties that make them of interest in metal processing. They are composed of alkyl ammonium halide salts and small polar organic molecules. As such, they have a very high chloride ion concentration (c.a. 6 molar), as well as containing polar uncharged compounds which makes them excellent for solvating a wide variety of metal salts including oxides. At the same time ionic liquids themselves are relatively benign and are conventionally made from widely available and sustainable materials.
Neutral Solutions for Immersion Gold Deposition
Four new gold deposition solutions have been developed as part of this work that will be of great interest to printed circuit board fabricators. The basic formulations of these solutions were as follows:
1 5 mM AuCN in Ethaline 200
2 5 mM KAu(CN)2 in Ethaline 200
3 5 mM KAu(CN)2, 10 mM ethylenediamine in Ethaline 200
4 5 mM AuCl, 1 mM I2 in Ethaline 200
A key advantage of each of these solutions is that they avoid the nodule boundary corrosion problems common with conventional aqueous methodologies for ENIG plating. In the case of the first two of these solutions, 1 and 2 above, both of these formulations produce bright, lustrous coatings consistent with those of the conventional aqueous process from neutral media, something that is not possible using aqueous media. These coatings have also been demonstrated to wet solder faster and more reliably than conventional commercial ENIG plating systems. They thus provide the basis for a completely new approach to gold deposition in ENIG coatings and one that could be further developed towards commercialisation by Aspis partner Atotech or other European chemistry suppliers to the PCB industry (see the Aspis business and exploitation plan for more details).
Formulation number 3 follows an extremely fast deposition profile while producing bright lustrous coatings due to the ability of ethylenediamine to accelerate the nickel dissolution kinetics. In addition, the centre of the EN nodules dissolves faster than that of the nodule boundaries eliminating the main mechanism of black pad formation producing highly reproducible coatings with excellent solderability properties. This is backed up by solder wetting data which, as with 1 and 2, shows fast, reproducible, solder wetting rates.
Cyanide-Free Coatings from Ionic Liquids
currently no cyanide-free formulations are available commercially for the immersion Au part of the ENIG process. However, formulation 4 above was developed using gold chloride as the Au source and I2 to activate the Ni surface to provide uniform, bright and quickly depositing Au coatings. This was also done from an acid-free formulation potentially eliminating one of the main drivers of blackpad.
Real World Potential
All of the formulations and protocols that have been developed could be developed for further use as direct drop in replacements for the conventional aqueous based coating methodologies. The newly developed coatings perform as well as, if not better than, standard systems, while also either reducing or eliminating the main mechanism of black pad formation. As such, these novel processes would have a low barrier to entry to commercial adoption once further development, scale up and validation in a production environment has been completed. In addition, the ability to remove cyanide from the Au plating bath serves as a highly attractive driver towards commercial adoption which could potentially lead to widespread use as an immersion gold formulation in many ENIG plating processes. Following the successful development of these highly novel chemistries and the completion of the Aspis project, further collaborative development work is needed to further optimise the formulations, to fully assess the reliability that they provide in real world applications and to consider their production on a larger scale. Such work will need to be undertaken in concert with a suitable supplier of such coatings to the PCB industry. Atotech would be an obvious partner as they were active in the Aspis project and also because they are an existing established supplier to the PCB and metal finishing industries not just in Europe but on a global basis. At the time of writing of this report, further collaboration and development opportunities are currently being investigated.
Impact WP 6: Validation (TNO / ITRI)
At the start of the project there was no common understanding about the black pad failure mechanism and there were no procedures available to analyse and confirm black pad or black pad related failures. The failure analyses conducted within WP 6 were done on printed circuit boards provided by the SME partners of the project (Merlin Circuits Technology, Graphic plc, Somacis and Global Interconnection Services) and one Dutch electronics assembly company called Liad. The boards were coming both from the field as assembled boards and as bare boards from the printed circuit board manufacturing process.
An important result of the conducted failure analyses was the correlation which was found between the morphology of the copper conductor and certain failures on ENIG and solder joint level. In one case, even the extent of solder wetting within a pad could be directly related to a high difference in the copper layer underneath. The remaining non- / de-wetted part of the solder pad showed a black nickel surface. Also, WP 4 trials comparing nickel deposits on cold-rolled copper and electrolytic-plated copper substrates showed that the morphology and characteristics were strongly dependent on the nature of the underlying copper. Discussions with printed circuit board manufacturers revealed that there were currently no industrial standards or guidelines regarding the copper surface condition for printed circuit boards entering the ENIG plating process. The knowledge gained about this correlation could be further investigated with a Design of Experiments (DoE) to quantify limits and tolerances. Such an outcome could then be implemented in an industrial standard or an industrial guideline.
Another important result concerns knowledge about a potential black pad indicator. This result is based on the failure analyses conducted within this work package and a result of the failure mechanisms work package, WP 2. It has been repeatedly found that black pad and black pad related failure sites exhibit traces of copper at the top surface of defective electroless nickel immersion gold layers. One hypothesis is that, in case of hyper-corrosion during immersion gold plating, the resulting cracks between the electroless nickel grain boundaries will allow copper to diffuse from the base conductor material to the top surface. On bare printed circuit board level, the copper will then be present at top of the electroless layer, underneath the immersion gold layer. This failure mode can be used as a possible indicator for black pad or black pad related failures. During the project a procedure was developed to use Scanning Electron Microscopy (SEM) with Energy Dispersive X-ray (EDX) to determine whether or not copper traces are really present at the top surface. By measuring the same printed circuit pad location with different acceleration voltages, it can be determined how the copper content changes as a function of the acceleration voltage. The copper is present underneath the surface in case the copper content decreases while increasing the acceleration voltage. This practical, non-destructive procedure has been extensively used within the project to get a first indication during a failure analysis investigation. Since such copper traces can also be a result of a contaminated plating bath, it should be noted that this can only be used as a possibility indicator. The detection of copper traces should always be further investigated to confirm black pad or a black pad related defect.
The soldering tests within this work package showed that printed circuit test boards made with both the Standard ITRI and the XL plating process provide proper solderable surfaces. Soldering tests conducted with a lead-free SnAgCu alloy showed promising wetting behaviour and defect free solder joints. Cross-sectioning revealed that nearly all investigated samples exhibited complete intermetallic layers at the interfaces. Reliability testing has also proved the potential stability of these interconnects. Both Standard ITRI and XL test boards were subjected to a high temperature aging test at 150°C. Samples taken at 0, 20, 100, 200 and 500 hours did not show any irregularities in intermetallic growth or the introduction of defects. Statistically, there appeared to be no significant difference in intermetallic growth between the Standard ITRI and the XL plated test boards, which regarding this aspect, confirms the level of the newly developed plating processes.
Another important result regarding the XL plating processes was the outcome of the experiments to deliberately induce the black pad failure. Hyper-corrosion was tried to induce by systematically reducing the thickness of the electroless nickel layer. This with the aim to create a porous nickel layer, layers which tend to be more prone to hyper-corrosion during immersion gold plating. Soldering tests showed that even the boards with the lowest nickel layer thicknesses (1 µm), did not result in black pad failures. Cross-sectioning of these thin layers did not even show the typical corrosion spikes, mud cracks or black bands along the interface. These test results indicate the robustness of the newly developed plating processes. These and the above described work package 6 achievements form should form valuable inputs for the further development and market introduction of the newly developed plating processes.
Summary and Conclusions
The Aspis project was undertaken to address a very important and intractable problem known as ‘black pad’ that impacted the reliability of printed circuit boards assembled using nickel-gold (ENIG) solderable finishes. These problems had been known for many years and, despite the efforts of both the board manufacturers and their plating chemistry suppliers, little progress had been made in either understanding the basic mechanisms causing the problem, or in identifying new processes that didn’t exhibit the problems. The Aspis project adopted a four pronged approach to addressing these reliability issues and this included; a detailed study of the mechanisms, the formulation of two new and distinctly different plating chemistries and the assessment of new non-destructive test methodologies. By developing an understanding of the mechanisms that caused these reliability problems, it was anticipated that the knowledge could be used to develop new aqueous plating processes. In addition, the use of ionic liquid based plating chemistries was felt to offer the possibility to use novel materials and deposition conditions that would avoid the possibility of initiating ‘black pad’ and related reliability problems. Finally, because ‘black pad’ is not visible or detectable without destructive testing, the project investigated the possibility of developing a non-destructive test method that could be used both by PCB fabricators and assemblers to give confidence that there were no hidden, latent reliability problems. Based on extensive experience, each of these four approaches was known to be extremely difficult to successfully complete and thus there was a high element of risk inherent within the overall project plans.
The mechanistic work carried out in work package 2 by LIOC has been extremely successful. Extensive, complex and detailed studies have, for the first time, enabled the mechanisms of ‘black pad’ to be properly elucidated. This in turn has enabled the chemistry formulations and process conditions leading to its formation to be identified. This knowledge will enable process chemistry suppliers to take these factors into account when formulating new chemical processes for the deposition of nickel gold solderable finishes. From a scientific perspective this work package has generated much new knowledge and its significance and importance cannot be overstated.
Attempts to develop new ‘black pad’-free ENIG systems have been made by the key industrial process chemistry suppliers over a number of years prior to the Aspis project, but without success. The work package to develop new aqueous-based processes was, therefore, known to be extremely challenging, with a high risk and low potential for success. However, the availability of the mechanistic information from work package 2 enabled the aqueous process development work in work package 5 to be focussed on novel formulations and conditions that were likely to produce reliable coatings. Whilst some progress has been made, the results at the end of the project confirmed that a substantial amount of additional work would be required before any of the formulations investigated could be considered for a move towards commercialisation.
Non-destructive testing of circuit boards for the presence of ‘black pad’ has been demonstrated and a bench scale prototype constructed and evaluated in a production environment. This has never been possible before and the capacitative method developed by TNO in work package 4 has the potential, with further development, to become a valuable piece of test equipment that could be used on a routine production basis by PCB manufacturers to confirm that their production is ‘black pad’-free. It would also be useful to board assemblers who could employ it as a quality control tool for the incoming inspection of circuit boards prior to soldering. Further work will be needed to take this technology forward and the partners will continue to explore the possibilities following the end of the Aspis project.
The development of new metal deposition processes from ionic liquids was always considered to be a more long term solution to the ENIG black pad problems. However, the University of Leicester has developed a range of gold deposition formulations which avoid the use of materials and process conditions that can lead to black pad formation. Using information from the mechanistic studies undertaken by LIOC, it was soon realised that the key to avoiding ‘black pad’ would be the development of new gold deposition processes. These have been successfully developed and evaluated on the laboratory scale and evaluated by the PCB fabricator partners. There is considerable interest in taking these chemistries forward to commercialisation and the possibilities will be further explored with the Aspis partners and other potential exploiters.
In final summary, the Aspis project has been successful in achieving most of its objectives in each of the four key work packages. Progress on the development of new aqueous chemistries has not been as significant as was hoped for, but it has given clear directions for future work. The demonstration of a non-destructive test method and equipment was highly innovative and offers potential for the production of a commercial unit. Finally, and perhaps the key achievements of the Aspis project have been the mechanistic studies and the development of ionic liquid-based gold deposition processes which are both world firsts and represent highly innovative scientific information and new technologies that should help to eliminate the occurrence of black pad in printed circuit boards.
30th November 2013
List of Websites:
Web site address: www.aspis-pcb.eu
Coordinator name and email address. Dr Martin Goosey, martingoosey@aol.com
