Improving Quality Control Efficiency in Electronic Industry with Smarter Sample Preparation Method

When the cross-section must be conducted via a specific area or level in the sample, targeted micro-sectioning for microscope inspection of samples is needed. This is most frequently done in the electronics sector for failure analysis and quality control, but also for research and production control purposes.

Since the level of the feature(s) of interest differs from sample to sample, typically only one sample can be done at a time so the targeted micro-sectioning is performed by hand. If the samples surfaces are sufficiently flat or the features to be hit are on the similar plane, it is also possible to use an advanced micro-sectioning system which is able to handle many samples in high volume. As an instance, Buehler PWB-Met system is engineered to serve this specific purpose for quality control of through-plated holes.

A very common preparation process is to manually grind with SiC papers with consecutive finer grits visually examining the sample with a portable microscope or pocket magnifier to see when the level of interest has been attained. Then the sample is polished by two or three steps of alumina powders of different particles size. The outline of these approaches is listed in Table 1.

Table 1. Conventional Preparation Method.

.
Cold-Mounting Accessories
25 mm, 30 mm, 32 mm dia. or home-made mounting cup
Cold-Mounting Consumables
Acrylic, Polyester or Quick-cure epoxy

 

Grinding and Polishing Equipment
Variable speed manual single or twin grinder/polisher
Surface/Abrasive Base
(rpm)
Time per Step
(Min)
*Plain backed SiC papers
120 (P120), 180 (P180), 240 (P280), 320 (P3400), 600 (P1200), 800 (P1500), 1200 (P2500)
*Sometimes only 3 steps
150-300 3
MicroCloth
MicroPolish Alumina Powders 1 μm, 0.3 μm, 0.05 μm
*Sometimes only 1 step
150-200 4
Total Time per sample 15-30

 

Metaserv 250 variable speed Twin Grinder-Polisher

Figure 1. Metaserv 250 variable speed Twin Grinder-Polisher (L) and typical microsection (R) in the electronic industries.

These conventional techniques have been used for a long time, but they still have several limitations:

  1. Low sample throughput
  2. Inconsistency in results because of difference behavior of operators
  3. Cross-contamination because of the dry-out of the alumina on the polishing cloth
  4. Smear which may hide features; for example, micro-cracks, intermetallics, lagers or delamination
  5. Relief problem from excessive polishing time and the use of soft cloth
  6. Inconvenience of switch of SiC papers between steps

Making one sample at a time frequently ends up with preparation becoming a bottleneck. In quality control, production may wait for results which mean loss of money. Additionally, it means limited capacity, irregular sample quality causing false measurements and interpretations, and number of samples not adequate to assure a correct statistical result.

It is thus obvious that many users are keen on ways to enhance the current preparation technique, accelerate the procedure, improve quality, and free up operators from tedious work. This SumMet article thus describes an enhanced technique which is moderately simple and logical to take targeted micro-sectioning a step further.

This improved technique will be illustrated on three common types of samples from electronic industry: PWB, Chip Scale Package (CSP), and wire bonding connections.

CSP, PWB & Wire bonding connections

Figure 2. Common electronic samples: CSP, PWB & Wire bonding connections.

The improved technique follows a moderately direct way of preparation: establish the “known distance” which is between the bottom of the mounted sample and the end level of interest. Then, just start the manual grinding until it reaches a level in the vicinity to the level of the interest, then followed by automated polishing by using steps through which material removal is regulated by the parameters set on the machine. The schematic drawing of this method is seen in Figure 3.

Preparation levels of the sample.

Figure 3. Preparation levels of the sample.

Although the initial step is still manual meaning only one sample at a time, the rest of the process is finished with multiple samples (using single force) automatically. Also, the initial step is often mostly fast, while the rest is the time-consuming part of the preparation. In general, benefits of automated preparation are:

  1. Total processing time per sample will be decreased
  2. Sample quality and uniformity of the results will be enhanced extensively
  3. The result is not impacted by the operator’s experience
  4. The preparation capacity is significantly increased

Case 1 Solderball on the CSP

Preparation Procedures

  1. For these samples, the length of the first manual step relies on the judgment of the operator. If the target is extremely far from the bottom of the mounted sample, either a coarser 180 (P180) or 240 (P280) grit CarbiMet 2 SiC paper will be used. If the “known distance” is comparatively short, 320 (P400) grit CarbiMet 2 SiC paper could be an alternative. The normal grinding time is about 2-3 minutes. The schematic drawing and overview of the sample are illustrated in Figures 4 and 5. When this level is attained, there will be a distant mark of the “next row” on the existing grinding plane which specifies that the target has been very close to being revealed. (See Figure 6).

preparation levels of CSP-BGA

Figure 4. Schematic drawing of preparation levels of CSP-BGA.

Overview of the sample

Figure 5. Overview of the sample. A known distance is shown on a CSP-BGA; best stopping level at the level of 0.7 mm from the edge.

  1. The 600 (P1200) grit CarbiMet 2 is usually used in semi-automated grinding with multi-samples. Load four semi-finished samples as shown in Figure 7 in the single-force holder of the Metaserv 250/Vector grinder/polisher and begin the grinding cycle. With the EcoMet/AutoMet 250/300, its six single-force fingers can prepare six samples at the same time by the way. Extended grinding time can cause loss of target. Here, operating time is based on the individual sample. The surface finishing is seen in Figure 8. The removal of materials is established by grinding time and can easily be repeated for the following batch of samples because the parameters (for example - pressure, speed) are totally controlled by the machine.
  2. Perform the automated polishing with 3 µm MetaDi Diamond Suspension (see Figure 9) on the VerduTex (or TriDent) cloth. Time is about three minutes.

a vague contour of the solderball

Figure 6. The target is almost revealed by 320 (P400) grit CarbiMet 2 SiC paper, a vague contour of the solderball is shown underneath the current grinding plane.

Four semi-finished samples in single-force holder

Figure 7. Four semi-finished samples in single-force holder (L); automated grinding by Metaserv 250/Vector grinder/polisher (R).

Decent scratches pattern

Figure 8. Decent scratches pattern from Grit 600/P1200 CarbiMet 2 SiC paper; 50x.

VerduTex 3 µm finishing

Figure 9. VerduTex 3 µm finishing: 100x(L); 200(R).

  1. Perform the final polish with MasterPrep on a ChemoMet cloth. Polishing time should not be over two minutes (see Figure 10). The intermetallic layer is illustrated in Figures 11 and 12; in addition the full sample preparation technique and parameters are listed in Table 2.

Final Polishing with MasterPrep

Figure 10. Final Polishing with MasterPrep on ChemoMet (L) Final Polishing Result; 200x.

Intermetallic whiskers at the solder joint

Figure 11. Intermetallic whiskers at the solder joint; 1000x.

SEM Micrograph of intermetallic whiskers

Figure 12. SEM Micrograph of intermetallic whiskers.

Table 2. Preparation Method for CSP.

.
Cold Mounting Consumables & Accessories
EpoThin Epoxy & 25 mm SamplKup
Polishing Equipment
Metaserv 250/Vector with Single Force Holder: 4 x 25 mm & 10” platen

 

Surface/Abrasive Head/Base
(rpm)
Head Direction Force
(Lb)
Time per Step
(Min)
Time per Sample*
(Min)
**Time per Sample (Min)
320(P400) grit CarbiMet 2 150 NIL NIL 3 3 3
600(P1200) grit CarbiMet 2 60/150 Comp 3 1.5 0.4 0.3
VerduTex 3 μm Metadi Diamond suspension 60/150 Comp 5 3 0.8 0.5
ChemoMet 0.05 μm MasterPrep polishing suspensions 60/150 Comp 2 2 0.5 0.3
Total Time 10 >5 >4

*Vector single-force holder allows for auto preparation flexibility up to 4 samples
**AutoMet 250/300 grinder/polisher allows for auto preparation flexibility up to 6 samples

Case 2 Tested PWB with PbSn solder

Preparation Procedures

  1. The time to reach the area close to the target based on the decision of the operator. In this case, the target is the second row of the solder which is approximately 3 mm away from the edge. One step of 320 (P400) grit CarbiMet 2 SiC paper is sufficient to get rid of the first row of solder and bring the target plane to the second row of solder. Manual grinding time is 2-3 minutes. The target plan is illustrated in Figures 13 and 14. After this step, the target plane has been closed to the edge of the second row of solder. The vague mark of the “next row” can also be seen from the existing sectioned plane (refer Figure 15).

preparation levels of PWB.

Figure 13. Schematic drawing of preparation levels of PWB.

A known distance (~3.3 mm)

Figure 14. A known distance (~3.3 mm) is shown on a PWB coupon.

Current grinding plane

Figure 15. Current grinding plane nearly touched the edge of second solder row (L); an unclear mark of second-row contour is shown (L).

  1. Load six semi-finished samples (see Figure 16) in the single-force holder of the AutoMet/EcoMet 250 grinder/polisher for automated grinding. Consumable to be used is 600(P1200) grit CarbiMet 2. Here, the likely stopping level should be very close to the center of the solder.

Automated grinding by AutoMet/EcoMet

Figure 16. Automated grinding by AutoMet/EcoMet 250 (L) Proper scratches from Grit 600/P1200 CarbiMet 2 SiC paper; 50x.

  1. When the 600(P1200) grit CarbiMet 2 step is completed, the subsequent few polishing steps involve removing the deformed material on the surface and uncovering the real microstructure. Perform the automated polishing with 9 µm MetaDi Diamond Suspension (see Figure 17) on the TriDent cloth. This hard cloth typically gives better flatness on very soft materials polishing because of the low resilience. Polishing time is about three minutes.

Coarse polishing scratches

Figure 17. Coarse polishing scratches from TriDent 9 µm polishing; 50x.

TriDent 3 µm Diamond Finishing; 50x

Figure 18. TriDent 3 µm Diamond Finishing; 50x.

  1. Soft and smear material (for example Sn-based solder or copper plating) typically requires diamond polishing steps to eliminate the preparation deformation. In this case, the 3 µm polishing is the first polishing step (see Figure 18). The conventional alumina powders polishing tend to create a lustrous surface, and smeared material sometimes unfortunately covers the materials defect (for instance - pores or crack). Diamond abrasive is thus considered as the best cutting alternative because of the sharp cutting edge to reveal the real microstructure. If certain miniature defects cannot be exposed after 3 µm polishing, an optional TriDent 1 µm diamond polishing step would be a solution.
  2. Perform the final polish in general with MasterPrep on a ChemoMet cloth; however, the polishing time should not be more than 1.5 minutes. The finished microstructure is illustrated in Figure 19; a plating crack defect is exposed after the full sample preparation (refer Figure 20). The developed technique with test parameters is illustrated in Table 3.

Final polishing

Figure 19. Final polishing by 0.05 µm Al2O3 suspension; 200x.

A plating crack on the copper plating

Figure 20. A plating crack on the copper plating; 500x.

Table 3. Preparation Method for PWB with PbSn solder.

.
Cold Mounting Consumables & Accessories
EpoThin Epoxy & 25 mm SamplKup
Polishing Equipment
AutoMet/EcoMet 250 with Single Force Holder: 6 x 25 mm & 10” platen

 

Surface/Abrasive Head/Base
(rpm)
Head Direction Force
(Lb)
Time per Step
(Min)
Time per Sample
(Min)
320(P400) grit CarbiMet 2 150 NIL NIL 3 3
600(P1200) grit CarbiMet 2 60/150 Comp 3 3 0.5
TriDent 9 μm Metadi Diamond suspension 60/150 Comp 3 3 0.5
*TriDent 3 μm Metadi Diamond suspension 60/150 Comp 3 3 0.5
ChemoMet 0.05 μm MasterPrep polishing suspensions 60/150 Comp 1.5 2 ~0.3
Total Time 14 ~5

*An optional TriDent 1 μm is required if the 3 μm polishing result is not fully satisfied.

Case 3 Copper wire bonding connection

Preparation Procedures

  1. Since the “known distance” is only 0.3 mm in this instance (see Figure 21 and 22), the grit size of the SiC paper for this primary manual can be fine. The ordinary SiC paper is 600(P1200) grit CarbiMet 2 which gives acceptable damages on a silicon wafer or the copper bump. With the coarse P240 or P180 SiC paper, the delicate silicon wafer will crack in general. The best level of this initial step should just contact the edge of the copper bump or reveal the 40-60% area of the copper bump (see Figure 23). The processing time is about 30 seconds to 1 minute; another closer view of this grinding position is illustrated in Figure 24.

preparation levels of wire bonding connection.

Figure 21. Schematic drawing of preparation levels of wire bonding connection.

wire-bonding connection

Figure 22. A known distance (~0.3 mm) is shown on a wire-bonding connection.

  1. Load six semi-finished samples in the single-force holder of the AutoMet/EcoMet 250 grinder/polisher for automated polishing. With the high pressure, the 3 µm diamond suspension on the TriDent cloth is able to eradicate all the earlier preparation damages and acquire a moderately good polishing result (see Figure 25 and 26).
  2. Blend 98 ml of MasterMet with 1 ml hydrogen peroxide and 1 ml ammonium hydroxide as the specialized final polishing fluid on the ChemoMet cloth. MasterMet itself has the chemical polishing effect on non-ferrous materials while the addition of the other chemicals would properly improve the whole chemical polishing process. The polishing result is illustrated in Figures 27 to 30. The full sample preparation technique and parameters are listed in Table 3.

Surface finishing

Figure 23. Surface finishing of 600(P1200) grit CarbiMet 2; 100x.

targeted copper bump

Figure 24. Position of the targeted copper bump seen through the mounting resin.

Six semi-finished samples

Figure 25. Six semi-finished samples in single-force holder (L); automated grinding by AutoMet/EcoMet 250 grinder/polisher®.

Automated polishing

Figure 26. Automated polishing (L); TriDent 3 µm Diamond Finishing; 200x (R).

Table 4. Preparation Method for copper wire bonding connection.

.
Cold Mounting Consumables & Accessories
EpoThin Epoxy & 32 mm SamplKup
Polishing Equipment
AutoMet/EcoMet 250 with Single Force Holder: 6 x 32 mm & 10” platen

 

Surface/Abrasive Head/Base
(rpm)
Head Direction Force
(Lb)
Time per Step
(Min)
Time per Sample
(Min)
600(P1200) grit CarbiMet 2 150 NIL NIL 1 1
TriDent 3 μm Metadi Diamond suspension 60/150 Comp 5 5 ~1
98 ml MasterMet 0.06 μm polishing suspension + 1 ml Hydrogen peroxide +1 ml ammonia hydroxide solution 60/150 Comp 2 2 ~0.3
Total Time ~8 ~2.5

 

Final polishing result

Figure 27. Final polishing result; 200x.

Final polishing result

Figure 28. Final polishing result; 500x (L); 1000x (R).

SEM Micrograph of copper bump

Figure 29. SEM Micrograph of copper bump; overview.

SEM Micrograph of copper bump

Figure 30. SEM Micrograph of copper bump; closer view.

Conclusion

Reviewing the techniques from above cases, the primary step unquestionably requires certain practices to accomplish the known distance before developing a full successful recipe. It also requires experience to decide the suitable grit size of the grinding paper for this step based on the type and size of the sample and the distance to the target level.

Basically, if the known distance is over 100 µm; a 320 (P400) grit CarbiMet 2 with 2-3 minutes grinding should be able to speedily bring the grinding level close to the analytical point. If the known distance is less than 60 µm, the 600 (P1200) grit CarbiMet 2 most probably is the secure choice even if its removing rate is comparatively slow. If any known distance is shorter than foregoing range, the targeting has to be done with caution to prevent overgrinding. If the device has certain very fragile materials (for instance GaAs), a finer SiC paper would be an alternative to minimize catastrophic damages. Having a standard check under an optical microscope on the grinding level for every 10 seconds is always a safe exercise.

Following the initial step, the rest of the process is finished automatically. It will need some preliminary work to discover the right parameters, but the fact that the single-force mode allows the operator to stop the places and examine the samples makes this much easier. And once done, the process can be easily be repeated. The result then is time-saving of at least 10 minutes per sample is possible and in addition samples will be more uniform with increased capacity and reduced costs.

Buehler

This information has been sourced, reviewed and adapted from materials provided by Buehler.

For more information on this source, please visit Buehler.

Citations

Please use one of the following formats to cite this article in your essay, paper or report:

  • APA

    Buehler. (2019, January 10). Improving Quality Control Efficiency in Electronic Industry with Smarter Sample Preparation Method. AZoM. Retrieved on November 12, 2019 from https://www.azom.com/article.aspx?ArticleID=15123.

  • MLA

    Buehler. "Improving Quality Control Efficiency in Electronic Industry with Smarter Sample Preparation Method". AZoM. 12 November 2019. <https://www.azom.com/article.aspx?ArticleID=15123>.

  • Chicago

    Buehler. "Improving Quality Control Efficiency in Electronic Industry with Smarter Sample Preparation Method". AZoM. https://www.azom.com/article.aspx?ArticleID=15123. (accessed November 12, 2019).

  • Harvard

    Buehler. 2019. Improving Quality Control Efficiency in Electronic Industry with Smarter Sample Preparation Method. AZoM, viewed 12 November 2019, https://www.azom.com/article.aspx?ArticleID=15123.

Ask A Question

Do you have a question you'd like to ask regarding this article?

Leave your feedback
Submit