This paper elaborates on problematic of MLCC capacitors cracks – literature survey and practical experiments to develop methodology to induce electrode-to-electrode cracks without deterioration of the capacitor’s immediate electrical parameters. In the next step these capacitors are subjected to thermal vacuum and high temperature life test to evaluate its impact to space flight operating conditions.
The paper was presented by Tomas Zednicek, EPCI European Passive Components Institute, Lanskroun, Czech Republic at the 3rd PCNS 7-10th September 2021, Milano, Italy as paper No.2.2.
Cracks in MLCC ceramic capacitor are, unfortunately, a well know phenomena that can depend to several factors. It is believed to reduce the reliability of the capacitor leading to catastrophic failure like short circuit.
When cracked capacitors are found in space projects the usual practice is to replace the defective parts and/or to solve the root cause of the problem as for example by modifying the assembly parameters to reduce the thermo-mechanical stress during assembly of the capacitor. However, in many cases, it is not possible to solve the issue and projects have to take risks by flying with potential defectives capacitors.
There are number of publications and works dedicated to MLCC cracks issues and evaluation of methods to reveal it by non-destructive methods. The first step of our assessment focused on available literature study.
CRACKS INDUCING METHODS
The literature survey on MLCC capacitors cracks inducing methods can be summarised as follows:
- High-temperature manufacturing processes and materials with different CTEs that are used in MLCCs result in significant built-in mechanical stresses. Deviations from the optimal conditions and some anomalies in the manufacturing process might form hidden defects and reduce the strength of the parts.
- Additional thermo-mechanical stresses associated with soldering and post-soldering handling of the assemblies can increase stresses to the level sufficient for cracking. These cracks might not affect performance of the system initially but result in failures with time of operation.
Teverovsky in its NASA paper proposed to assess the effect of cracking by pressing the surface with a Vickers indenter. Results showed that cracking does not affect capacitance, which is consistent with the results of flex testing when in many cases capacitance recovers when the stress is removed. The parts electrically passed this test, just dissipation factor showed some sensitivity and increases after cracking by 10 to 50%, but remains within the specified limits.
METHODS TO REVEAL MLCC CAPACITOR CRACKS
Commonly used non-destructive techniques to detect cracks include:
- external visual optical examination under magnifying glass / optical microscope
- electrical characterisation
- ultrasonic inspection / SAM
- Xray 2D/3D and tomography / CT scan
Despite substantial efforts that have been made to develop methods for revealing cracks in MLCCs after soldering, none is universal and can detect defective parts reliably. Suitability of crack detection by methods also depends to the size of MLCC capacitors. More methods are suitable for larger MLCC capacitors above 1812 size, only few can be used for 1210 case sizes and smaller.
Experimental testing was performed on 1812 case size X7R MLCC readily available part with high CV at this case 22uF 25V parts from leading MLCC manufacturer date code 2030.
THERMO-MECHANICAL SHOCK TEST
The first experiment aims to maximize the thermo shock stress. The test sequence consists of visual inspection pre and post test, basic electrical parameters characterization including third harmonic measurement THD and the thermal shock cycle. The thermal shock followed three cycles of dip into hot 245C solder followed by immersion into a liquid nitrogen -195C. Dwell time between the dips were less than 3s.
- Sample Size: 10 units 1812 MLCC 22uF 25V X7R
- Optical visual inspection under microscope: no microcracks observed under optical microscope
- Electrical Parameters: All pre and post Capacitance, DF and IR parameters were within specification without significant degradation.
- THD Third Harmonic Measurement: without significant degradation.
The test sequence performed on ten samples did not show any signs of surface cracks observable under optical microscope. There was also no deterioration of electrical parameters neither THD voltage after the test.
CRACKS INDUCED BY MECHANICAL PIN STRESS
The second method tried to simulate extreme flex cracks by replicating of Teverovsky’s method with defined mechanical pin stress force to the center of capacitor body.
Sharp mechanical pin was applied from top side to the center of the MLCC body. The pin strength was set to the maximum device capability force: 510±10 N. The parts were visually inspected under optical microscope and electrically characterized including THD before and after the test.
- Sample Size: new set of 15 units 1812 MLCC 22uF 25V X7R
- Optical visual inspection under microscope: Surface cracks with prolonging microcracks visible on the surface of the MLCC body – see Figure 1. Some units crack appeared mildly; others showed remarkable level of mechanical damage.
- Electrical Parameters & THD: Some units resulted in short circuit after the test, others remained within its electrical parameter specification.
Some MLCC samples were considerably mechanically damaged, nevertheless there is not necessary a direct link between the level of mechanical damage and electrical deterioration / short circuit.
Mechanical stress exposure parameters modification
The following mechanical pin stress parameters were modified for detail investigation on its impact to initiation of cracks.
- Pin stress force varied from 350 to 900N (mechanical equipment was also adjusted to be able to apply even higher stress level)
- Pin shape was changed from low radius (sharp) to high radius (blunt)
- THD was continuously measured during the force applied and values at MAX force recorded.
New set of 10 samples were used, samples No. 1-5 were investigated after sharp pin exposure and changing force, the other samples 6-10 with blunt tool pin shape and higher force ratio.
- Low radius sharp pin tool causes a surface damage with microcracks even at lower applied force unlike the blunt tool that even with quite high force 900N did not induced visible surface cracks.
- THD increased during the force being applied in the case of sharp pin tool and returned to original numbers after the force was released – this is in line with Teverovsky results using Vickers hardness measurement tool that referred increase in DF when force is applied and its return afterwards. THD did not increased during the high force applied in the case of blunt pin tool. (that also visually did not damage the MLCC body surface).