A New Method for Non-Destructive Characterization of Through Holes in Printed Circuit Boards Open Access

Plated through holes (PTHs) are used in the design and production of printed circuit boards (PCBs) to electrically interconnect traces on different layers and provide features for the attachment of surface mount, ball grid array, or compliant pin components. PCBs require robust PTHs that can withstand the thermal stresses of solder reflow processing during electronic card assembly and the thermal stresses associated with power cycling in the field. However, the laminate materials used in PCBs have a coefficient of thermal expansion (CTE) much greater than that of copper plating, especially above the laminate glass transition temperature (T_g), resulting in significant stress on the PTHs during thermal cycling [1]. Excessive stress can cause circumferential barrel cracks in the PTHs, resulting in open circuits and electrical failures. High aspect ratio PCBs are especially susceptible to this failure mechanism. PTH cracking is accelerated by defects or irregularities in the through hole resulting from drilling, desmear, or plating processes that create local stress concentrations [2]–[5].

Currently, cross-sectioning is the most commonly used method to evaluate plating quality. However, detection of plating abnormalities with cross-sectioning is a time consuming and destructive process. Half of the PTH is lost to grinding and only one vertical plane can be analyzed at a time. Computer tomography (CT) x-ray techniques have also been used to evaluate PTH defects. However, this requires expensive equipment and the images produced can be difficult to interpret [6]. To overcome the limitations of cross-sectioning and CT x-ray, a novel, non-destructive method of evaluating through hole quality and reliability was developed.

This paper describes a method to create castings of through holes and demonstrates the method's effectiveness in replicating plated and non-plated through hole surfaces for reliability assessment and prediction.

PTH arrays were excised from a PCB and placed in standard 1.25 inch diameter mounting cups. Isokote 1000 silicone spray (Isokote, Canton, GA) was applied to the cups and the samples to improve removal of the elastomeric castings. Sylgard 184 (Dow Corning, Midland, MI), a two-part silicone elastomer, in a 10:1 base to curing agent ratio was used to create silicone castings. Sylgard 184 was poured into the mounting cups, submerging the PCB sections. The mounting cups were then placed in a vacuum chamber for approximately 15 minutes to remove air bubbles and draw the Sylgard 184 into the through holes. The samples were then cured at 100 °C for one hour. The castings were excised by removing the excess cured silicone elastomer with a razor blade and gently pulling the casting from the PTHs. An illustration of this method is shown in Fig. 1.

Optical microscopy of silicone castings and the corresponding epoxy-mounted, cross-sectioned through holes was performed using a Nikon Eclipse LV150N. Scanning electron microscopy of silicone castings and the corresponding dry cross-sectioned through holes was performed using a Hitachi S-4700 field emission-SEM at an accelerating voltage of 10 kV. Samples for SEM analysis were coated with chromium.

Castings of 27 current induced thermal cycling (CITC) coupons, 13 from lots prior to a drilling improvement and 14 from lots after the improvement, were created. The coupons were subjected to CITC testing according to IPC 2.6.26 [7]. The surface roughness of the castings was measured using a Keyence VK-9700 3D laser scanning microscope. The center region of a representative PTH from each coupon was analyzed at 20×. Using the VK Analyzer software, the surface roughness of the castings was measured with the multi-line roughness feature. In this experiment, the roughness of 25 lines along the long axis of each PTH was measured. Both R_a (arithmetic mean roughness) and R_z (cross-point average roughness) for each line were measured according to JIS B 0601-2001, recorded, and averaged [8].

To demonstrate feasibility of the casting method, known defective 6.19 mm thick impact connector PCB samples and impact connector current induced thermal cycling (CITC) coupons with 0.52 mm diameter as-drilled holes and 0.44 mm diameter finished PTHs (~14:1 aspect ratio) were procured for casting analysis. PCB samples 5.63 mm thick with 0.3 mm as-drilled and 0.25 mm diameter PTHs (~22:1 aspect ratio) were also evaluated.

After removal of the castings from the PCB samples, optical and scanning electron microscopy showed that the castings captured the microtopography of the plating. Figs. 2 and 3 show representative images comparing the castings to the cross-sectioned PTHs from which they were made. The images demonstrate that the castings accurately replicated plating features such as nodules, cracks, and internal planes even for PTHs with aspect ratios as high as 22:1.

Other PTH attributes can also be evaluated using the castings, including PTH diameter along the length of the barrel and back drill parameters such as depth, concentricity, and presence of resultant stubs. The images in Fig. 4 demonstrate how the castings can be used to evaluate back drilling. This type of evaluation can be performed for multiple PTHs across an array simultaneously.

In addition to plating evaluations, castings of through holes prior to plating can be created, as shown in Fig. 5. Without the plating, the as-drilled and/or post-desmear hole wall can be assessed. The castings replicate the glass fiber bundle protrusion from the resin and resin surface roughness. Excessive fiber bundle protrusion into the hole wall causes plating ridges and valleys that act as stress concentrators during thermal cycling. Therefore, evaluating non-plated hole walls can help to assess PTH reliability and pinpoint causes of low reliability.

While optical and scanning electron microscopy can provide qualitative information about the plating topography and quality, quantitative data can also be collected from the castings by measuring the surface roughness using a laser scanning microscope. This was demonstrated by evaluating CITC coupons manufactured before and after a drilling improvement. CITC is test used in the PCB industry to assess PTH reliability. The CITC coupons are heated at 3 °C/s by continuously adjusting the current and are cooled to ambient temperature by fans. One heating and cooling phase is considered a cycle [2]. The more cycles a coupon can withstand, the higher the reliability.

The coupons manufactured after the drilling improvement had significantly higher CITC performance with an average of 49.0 ± 6.9 cycles to failure as compared to the before drilling improvement coupons which had an average of 22.9 ± 4.3 cycles to failure. Furthermore, castings of the after drilling improvement coupons appeared to have smoother plating than the before drilling improvement coupons. As shown in the representative laser images in Fig. 6, the coupons manufactured before the drill improvement had more noticeable ridges and nodules in the plating. Measurement of the surface roughness in lines along the length of the castings, as shown in Fig. 7, confirmed that the coupons manufactured before the drilling improvement were significantly rougher than the coupons manufactured after the drilling change. The significance threshold was set at p < .001. A comparison of the before and after drilling improvement coupon average roughness is shown in Fig. 8. The before drilling improvement samples had an average R_a of 1.66 μm ± 0.52 μm and an average R_z of 10.76 μm ± 2.44 μm; the after drilling samples had an average R_a of 0.91 μm ± 0.17 μm and an average R_z of 6.07 μm ± 0.87 μm. The increased variability in the before drilling improvement samples is in part due to the presence of plating nodules of varying size. The results show that the casting method indicated a directional improvement in PTH reliability consistent with the CITC results. A good correlation between the individual coupon roughness and cycles to failure was also demonstrated and is shown Fig. 9. Some variation in this correlation is expected because roughness can vary PTH to PTH within a single coupon and along the length of a PTH. Other factors such as where the drill bit was in its life cycle when a given PTH was drilled could have affected the roughness measurement.

Suboptimal drilling, desmear, and plating processes during PCB manufacturing can cause irregularities in PTH plating that create stress concentrators and accelerate PTH cracking during thermal cycling. Recently, Hiroshima et.al. confirmed that plating roughness has a significant impact on PTH life through computer simulations and showed that concave features on the inner wall of the PTH are especially detrimental [5]. Therefore, assessment of the plating roughness and uniformity is critical to ensure manufacture of reliable PCBs. Further, it would be ideal to conduct this assessment on production level hardware, meaning a non-destructive method is required. The casting method presented in this paper provides a technique to accomplish these goals.

PTH castings accurately replicate microfeatures and roughness measurements of the castings correlate to thermal cycles to fail, demonstrating that the castings can be used to help predict PTH reliability and functional life. Furthermore, because a casting can be made and analyzed in a couple of hours as compared to the days to weeks required for thermal cycling, this technique can be used as a rapid screening tool to detect the presence of pervasive plating irregularities. For example, prior to entering a battery of expensive and time-consuming qualification tests, castings of PTHs can be used to assess likelihood of success. The castings could be made from the coupons going into the test and if the plating appears smooth and uniform, those same coupons could continue to be subjected to the more extensive qualification tests. If necessary, a cleaning step could be incorporated using Dow Corning brand OS fluids to remove any residual silicone elastomer. The casting technique would also be useful in quickly evaluating the impact of process or materials changes and in performing health of the line checks to ensure that the manufacturing process is stable and that PTH reliability is maintained. Failure analysis is another potential application of the casting technique. Castings can provide information about the plating quality that was previously difficult to obtain with cross-sectioning alone which could assist in determination of failure root cause. Further, the castings can be made without damaging the sample, preserving it for other analysis or testing.

To fully utilize the non-destructive nature of the casting technique, equipment could be developed to make castings on production level circuit boards without excising samples. One could envision an apparatus that clamps onto a particular region of interest on the PCB and could be used to introduce the silicone elastomer material and draw a vacuum. Automated software could then be used to measure the roughness along the length of multiple castings.

While the casting technique was demonstrated on PCB PTHs, usage could also be expanded to include assessment of laser or plasma processed holes on laminate, ceramic, silicon, and glass chip carriers and as such, be valuable in reliability analysis and prediction in first level packaging applications.

A method was developed to create castings of PTHs and enable non-destructive qualitative and quantitative measurement of through-hole roughness which is a critical factor in PTH reliability. As evidenced by the optical and SEM images, the castings were successfully removed from even high aspect ratio PTHs and accurately replicated the micro-topographic features. The surface roughness of the castings was measured by laser scanning microscopy and these roughness measurements were shown to correlate to CITC cycles to failure. Therefore, the casting method can be used to aid in prediction of PTH reliability. In addition to plating surface characterization, the casting technique can be used to evaluate other through-hole features such as back drill quality, diameter as a function of barrel length, as well as the effectiveness of drill and desmear processes by creation of castings prior to plating. Thus, this technique can be used during PCB qualification, manufacturing process optimization, process monitoring, and failure analysis to assess through-hole quality and reliability.

The authors would like to thank Jim Bielick and Jennifer Bennett for their guidance in use of the laser scanning microscope.