Parts Verification
Test Methods to Identify COUNTERFEIT MLCCS
Electrical and physical characteristics play a role in high-accuracy detection. by YUNG-HSIAO CHUNG,1 CHENG-HSUN LEE,1 LIWEI XU,1 YUQIAN HU,1 ZONGXUAN WANG1 and STEPHEN E. SADDOW2

Most multilayer ceramic capacitors (MLCC) have no marking and cannot easily be distinguished from their package, which gives unscrupulous vendors opportunities for fraud. Here, the authors introduce several test methods for MLCC compliance verification, namely 1) the effect of DC bias on capacitance, 2) capacitance temperature characteristics, 3) high-voltage testing of DCW (dielectric withstand voltage) and IR (insulation resistance), 4) cross-section (dielectric layer and terminal comparison for flex types), and 5) electron microscopy (EDS) material analysis to match with known good device chemical composition.

One important step must be performed before testing Class II capacitors. This is referred to as the “capacitor precondition test.” The standard way to do this, according to Murata, is to perform a heat treatment at 150+0/-10°C for 1 hr., then let the part sit for 24±2 hr. at room temperature, then measure its electrical characteristics.

The reason to perform this precondition test is due to the characteristics of BaTiO3, a typical metal-oxide dielectric used in MLCCs and the base material of Class II MLCCs (FIGURE 1). A decay in dielectric permittivity has been observed with these formulations, whereby the molecular structure of BaTiO3 changed over time. The initial galvanic molecular structure displayed gradually transitions to a chaotic couple structure. The chaotic couple structure of the dielectric molecules has a lower ability to store charge than the galvanic molecular structure, thus causing the capacitance value to decrease. In general, this phenomenon is referred to as the aging process.

This process can be reversed via de-aging using a heat treatment. Provided the temperature of the material is above its Curie point (for BaTiO3 ~125oC), and since most de-aging procedures are 150oC for 1 hr. with a 24-hr. pause, the material can recover. Once the BaTiO3 reaches its Curie point, the molecular structure converts back to the chaotic molecular state, and the device is “reset.”

Capacitance Measurement

Several factors may affect capacitance measurements: test signal level, frequency, and device impedance. TABLE 1 is the industry standard for measuring Class I/ II capacitors.

4-Terminal-pair (4TP) measurement method (impedance). When using an auto-balanced bridge capacitance meter, the most common measurement technique is the 4-terminal-pair (4TP) measurement method.2 In these measurements the Hc and Hp terminals are shorted together, while the Lc and Lp terminals are also shorted (FIGURE 3).

The Hp and Hc terminals are often referred to as the CMH (capacitance meter high) terminal, and the Lp and Lc terminals are commonly referred to as the CML (capacitance meter low) terminals. Some residual inductance and resistance are in the cables, along with parasitic capacitance between the cables, or between the device-under-test (DUT) and ground. When the measurement is performed, parasitic compensation and calibration must be performed to eliminate these parasitic elements. Otherwise the accuracy of the measurement will be greatly reduced.

Figure 1: Material characteristics inside MLCC devices
FIGURE 1. Material characteristics inside MLCC devices. The initial electric dipole orientation (a) explains how capacitors store charge, with galvanic molecular structure, while (b) shows the chaotic couple structure. The rotation of the dipoles is how capacitors lose charge storage over time. (c) Structure when temperature is above the Curie point. (Source: Murata)
Figure 2: MLCC physical characterization
FIGURE 2. MLCC physical characterization. (a) Device cross-section location. (b) Photograph of device examined under SEM. (c) Cross-section SEM micrograph (50000X) of capacitor dielectric (BaTiO3) before preconditioning. (d) Cross section SEM micrograph (50000X) of capacitor dielectric (BaTiO3) after preconditioning.
Table 1. Industry Standard Tests for Class I/II Capacitors1
Table 1: Industry Standard Tests for Class I/II Capacitors
Figure 3: Measurement of MLCC capacitance using the 4TP measurement method
FIGURE 3. Measurement of MLCC capacitance using the 4TP measurement method. (Source: Agilent/Keysight)
Figure 4: Open (a) / short (b) correction is used to calculate residual impedance and stray admittance of the test fixture
FIGURE 4. Open (a) / short (b) correction is used to calculate residual impedance and stray admittance of the test fixture. (Source: Agilent/Keysight)
Figure 1: Material characteristics inside MLCC devices
FIGURE 1. Material characteristics inside MLCC devices. The initial electric dipole orientation (a) explains how capacitors store charge, with galvanic molecular structure, while (b) shows the chaotic couple structure. The rotation of the dipoles is how capacitors lose charge storage over time. (c) Structure when temperature is above the Curie point. (Source: Murata)
Table 1. Industry Standard Tests for Class I/II Capacitors1
Table 1: Industry Standard Tests for Class I/II Capacitors
Figure 2: MLCC physical characterization
FIGURE 2. MLCC physical characterization. (a) Device cross-section location. (b) Photograph of device examined under SEM. (c) Cross-section SEM micrograph (50000X) of capacitor dielectric (BaTiO3) before preconditioning. (d) Cross section SEM micrograph (50000X) of capacitor dielectric (BaTiO3) after preconditioning.
Figure 3: Measurement of MLCC capacitance using the 4TP measurement method
FIGURE 3. Measurement of MLCC capacitance using the 4TP measurement method. (Source: Agilent/Keysight)
Figure 4: Open (a) / short (b) correction is used to calculate residual impedance and stray admittance of the test fixture
FIGURE 4. Open (a) / short (b) correction is used to calculate residual impedance and stray admittance of the test fixture. (Source: Agilent/Keysight)

Compensation and calibration. Four types of compensation and calibration steps are usually performed: open correction, short correction, cable length calibration, and load correction. The open/short correction (FIGURE 4) is used to compensate for stray admittance and residual impedance due to the test fixture.

With these two corrections, then we can extract Z of the text fixture.

Cable length calibration formula
Cable length calibration improves bridge balance stability at high frequencies, and it compensates for any phase drift induced from cable length and high frequency. Perform phase compensation before open/short cable compensation to achieve the best calibration condition for the test.

Load correction is usually performed if the testing frequency is greater than 5MHz. Since the MLCC test frequencies are all below 1MHz, we do not need to discuss this further here.

Test frequency. Test frequency can be a useful means to detect counterfeit MLCCs, especially in the common case where type Class II capacitors are substituted for Class I capacitors. The frequency test characteristics of Class I capacitors are very stable since the capacitance does not change with frequency. On the other hand, Class II capacitors display a well-known drop in capacitance at high frequencies. Thus, it is easy to determine if the MLCC is Class I or Class II by simply doing a frequency sweep. FIGURE 5 and 6 compare Class I and II capacitance during a frequency sweep from 20Hz to 10MHz:

Figure 5: Class-I MLCC capacitance measurement
FIGURE 5. Class-I MLCC capacitance measurement. C0805C101F1GACTU capacitance change vs. frequency (20Hz to 10MHz).
Figure 6: Class-II MLCC capacitance measurement
FIGURE 6. Class-II MLCC capacitance measurement. C0805C152KDRACAUTO capacitance change vs. frequency (20Hz to 10MHz).

Frequency sweep in these two samples:

Sample 1: C0805C101F1GACTU
Rated: 100pF, 100V, C0G (class I) 1% tolerance
Sample 2: C0805C152KDRACAUTO
Rated: 1500pF,1000V, X7R (class II) 10% tolerance.

Frequency sweep 20Hz to 10MHz, E4990A Keysight Impedance Analyzer; test fixture: Keysight 16034G.

Test level. The test level will change the measured capacitance, especially for general Class II capacitors. FIGURE 7 shows data for a general Class II capacitor from Murata.

However, when the capacitance is greater than 10µF, the impedance is too low to keep the voltage at the right voltage level. In this case,

When the capacitance is greater than 10µF, the impedance is too low to keep the voltage at the right voltage level

To keep the AC signal in the 1Vrms level, the meter must have the ability to source the current to:

To keep the AC signal in the 1Vrms level

Also, the meter needs to function in auto-level control (ALC) mode, which will increase the output voltage on the test meter to adjust for the voltage divider. FIGURE 8 shows the measurement result without the ALC function: Voltage across the DUT was 181mV with a set voltage of 1.0Vrms. If the ALC function is enabled, the instrument will automatically raise the source voltage to achieve the desired 1.0Vrms across the DUT. FIGURE 9 shows a measurement of the same 10µF capacitor using the Keysight E4980 LCR meter with the ALC feature set to ON.

TABLE 2. Class I and II Capacitor Capacitance Change vs. Frequency
Table 2: Class I and II Capacitor Capacitance Change vs. Frequency
Figure 7: AC voltage characteristics from Murata Class II
FIGURE 7. AC voltage characteristics from Murata Class II (GRM188D70J106MA73).

In some cases, we noticed general Class II capacitors were used to replace automotive MLCCs. Both are Class II. Therefore, the frequency method or temperature method is not able to detect the counterfeit part. However, the AC characteristic can be used in this case.

The automotive-grade MLCC capacitance is more stable vs. AC voltage variation. FIGURE 10 shows the AC voltage characteristics of the GCM32EL8EH106KA07 Class II capacitor. Compared with the general MLCC, GRM188D70J106MA73 capacitance change (loss) vs. AC voltage was 27.9% at 0.01Vrms (Figure 10), while the GCM32EL8EH106KA07 and GCM32EC71H106KA03 experienced only a 4.2% loss (FIGURE 11 and 12).

Insulation Resistance and Leakage Current

Insulation resistance is one of the important parameters used to identify counterfeit MLCCs. The different MLCCs have different insulation resistance, which depends on the application. From experience, a common method to counterfeit MLCCs is to place low specification chips into high specification packages, and then claim it as a high specification part. For some applications, the MLCC must have a higher insulation resistance. If the user chooses the counterfeit MLCC, the device/circuit performance may initially seem fine, but over time leakage current and breakdown voltage will degrade, adversely affecting circuit performance and possibly leading to device/circuit failure. This is particularly the case where low insulation resistance affects the operation of circuits intended to be isolated. Unexpected high leakage currents can eventually lead to deterioration of the insulation by heating or direct current electrolysis. Consequently, knowing how to measure MLCC insulation resistance is one of the important methods to identify counterfeit MLCCs (FIGURE 13).

Figure 8: Capacitance measured without ALC
FIGURE 8. Capacitance measured without ALC. The testing level set to 1.0V; however, the voltage monitor shows only using 181.864mV.
Figure 9: Capacitance measured with ALC on
FIGURE 9. Capacitance measured with ALC on. The testing level set to 1.0V, and the voltage monitor also showing using 999.787mV.

The insulation resistance values for MLCCs are usually very large, generally in the mega-ohms (MΩ) range. In terms of the RC time constant, the product is typically in the ohms-farads (ΩF) range or larger. For example, if the MLCC capacitance is 10µF and the minimal insulation resistance is 500ΩF, the insulation resistance equals 500ΩF/10µF or 50GΩ. This value cannot be measured by a conventional ohmmeter, as those instruments are only accurate up to ~1GΩ. We thus need to measure the insulation resistance using the electrometer/high-resistance meter and follow the procedure outlined in MIL-STD-202-302.3

For instance, for the Keysight B2987A electrometer/high-resistance meter, the capacitance resolution is 0.01fA, with a maximum resistance measurement of 10PΩ. On the other hand, there are two basic ways to measure leakage current: the series method and the parallel method. In the series method an electrometer is placed in series with the capacitor and voltage source (FIGURE 14). For the parallel method, a voltmeter is in parallel with a resistor, and then connected in the series to the capacitor and voltage source (FIGURE 15). The series method measures leakage current for the MLCC. From the MLCC datasheet, we apply the rated voltage to the capacitor for 60 to 120 sec., depending on the capacitance because, while we apply a DC voltage to the capacitor terminals, current will start to charge the capacitor, and, after charging is complete, the current will decrease and then level off (FIGURE 15). From this steady-state current we can identify it as the leakage current. In this measurement, we determine the voltage applied to the capacitor and leakage current passing the capacitor after it is fully charged. Then we can calculate the insulation resistance of MLCC by Ohm’s law, R = V/I.

Figure 10: General rating class II capacitor
FIGURE 10. General rating class II capacitor (GRM188D70J106MA73), capacitance loss (-27.9%) at 0.01Vrms. Data from manufacturer.
Figure 11: Automotive rating class II capacitor
FIGURE 11. Automotive rating class II capacitor (GCM32EL8EH106KA07), capacitance loss (-4.2%) at 0.01Vrms. Data from manufacturer.
Figure 12: Automotive rating class II capacitor (GCM32EC71H106KA03)
FIGURE 12. Automotive rating class II capacitor (GCM32EC71H106KA03), capacitance loss (-4.2%) at 0.01Vrms. Data from manufacturer.

Since the MLCC is made using a real dielectric material with a non-zero loss tangent, which is not a perfect insulator, there will always be leakage current present. Additionally, MLCCs have a different value of insulation resistance because they are composed of different materials or combinations of materials. Therefore, low MLCC insulation resistance or high leakage current has many causes, such as device temperature and moisture, dielectric contamination, oxidation, loss of volatile materials, and material cracking. Insulation resistance measurement is especially helpful to determine the extent to which the insulating properties are affected by deteriorative influences and whether the MLCC is counterfeit or of low quality.

For example, the MLCC part no. C2012JB1A476M125AC, manufactured by TDK, has a nominal capacitance of 47µF. We used the B2987A electrometer/high-resistance meter to measure its insulation resistance by the series method (FIGURE 16). We set the voltage source of the meter to 10V, which is its rated voltage. From the datasheet, we found the voltage application time for the insulation resistance measurement is 60 sec., and the minimum insulation resistance is 2MΩ. Figure 16 shows a stable measurement of 59.8444 MΩ, which is higher than the datasheet spec. Therefore, we can identify the device under test as matching the manufacturer’s specification. On the other hand, in FIGURE 17, we observed the insulation resistance was lower than the specification and kept decreasing as we charged the MLCC for 60 sec. Therefore, we can identify this DUT as failing the insulation resistance measurement and designate it a counterfeit MLCC.

Figure 13: Behavior of MLCC current vs. charging time
FIGURE 13. Behavior of MLCC current vs. charging time. Note three distinct current levels: charge current (peak MLCC current), absorption current (exponential decay due to device RC time constant) and steady-state leakage current.
Figure 14: Series method for insulation resistance measurement
FIGURE 14. Series method for insulation resistance measurement.
Figure 15: Parallel method for insulation resistance measurement
FIGURE 15. Parallel method for insulation resistance measurement.
Dielectric Withstanding Voltage

The dielectric withstanding voltage test assesses the reliability and expected lifetime of an MLCC. Failure during a dielectric withstanding voltage test results in short circuits caused by decreasing insulation resistance and increased current, which will damage other chips on the board. The typical breakdown voltage for an MLCC is much greater than the rated voltage. But a voltage less than the breakdown voltage may permanently damage the insulation and thereby reduces its safety factor. For an MLCC, dielectric withstanding voltage failures lead to internal damage by electrical overstress cracking (FIGURE 18).

The manufacturer uses the dielectric withstanding voltage test to determine the voltage rating and verify the MLCC can operate at its rated voltage without material degradation. It is also used to assess whether the device can withstand a momentary overvoltage event due to switching spikes or surges. In other words, the dielectric withstanding voltage represents the maximum level of continuous voltage that can be applied across an MLCC. Dielectric withstanding voltage tests vary on the voltage applied or stress condition. According to military5 and manufacturer specifications, the dielectric withstanding voltage for Class I MLCCs is usually three times the rated voltage. For Class II MLCCs, the dielectric withstanding voltage is 2.5 times the rated voltage.

For the dielectric withstanding voltage test we used a Vitrek V73, which is an AC/DC/IR hipot tester. It can provide 5kVAC/DC with a 20mA source current. As mentioned, the dielectric withstanding voltage test measures the MLCC breakdown voltage and confirms the MLCC can safely operate at the manufacturer’s rated voltage. When an MLCC fails the dielectric withstanding voltage test, application of the test voltage will result in a disruptive discharge, such as a flashover, sparkover or breakdown. Additionally, MLCC deterioration due to excessive leakage current may change the device’s electrical parameters or physical characteristics.

For example, from the datasheet for the C2012JB1A476M125AC MLCC tested in the last section, the rated voltage is 10V, has a JB temperature characteristic and is Class II. That means the dielectric withstanding voltage of the device under test is 2.5 x 10V = 25V, and the voltage application time is 1 sec. We observed no breakdown when 25V was applied to the device. The rated voltage of the device under test matches the manufacturer’s specification and passes this test (FIGURE 19).

Figure 16: Insulation resistance measurement of a qualified MLCC using the B2987A electrometer/high resistance meter
FIGURE 16. Insulation resistance measurement of a qualified MLCC using the B2987A electrometer/high resistance meter.
Figure 17: Insulation resistance MLCC failure using the B2987A electrometer/high resistance meter
FIGURE 17. Insulation resistance MLCC failure using the B2987A electrometer/high resistance meter. Note the decay in resistance vs. time.
MLCC DC Bias Effect

MLCCs use dielectric materials, which makes them different from other capacitors, such as electrolytic. Their materials provide a high dielectric constant (Dk) that changes according to environmental factors.

MLCCs are divided into classes based on the dielectric materials used. Two of the most common types of MLCCs used in the industry are Class I, which is temperature compensating, and Class II, which has a high Dk. Class I capacitors tend to have lower capacitance values and are more stable than Class II capacitors.

Class I MLCCs contain a low-loss dielectric and are very stable, as shown in the measured data of FIGURE 20. These measurements were taken at room temperature at a frequency of 1MHz under various DC bias applied voltage ranging from 0V to 40V.

Class II MLCC permittivity depends on the applied electric field. Therefore, with different applied voltage, the MLCC capacitance varies accordingly. The measured data in FIGURE 21 show that the capacitance changed after performing the DC bias sweep. These measurements were taken at room temperature at 1kHz and under various DC bias applied voltage ranging from 0V to 10V.

As shown above for a Class 2 MLCC, as the applied voltage increases, the change of capacitance becomes more significant. For Class I MLCCs, however, different voltage ratings hardly affect how they perform. With this knowledge, we can tell if a part is a legitimate Class I MLCC or not based on its DC bias capacitance profile.

Figure 18: MLCC dielectric withstanding voltage test
FIGURE 18. MLCC dielectric withstanding voltage test. (a) Dielectric breakdown by EOS (electrical overstress). (b, c) SEM morphology of dielectric showing a local stratification phenomenon.4
Figure 19: Dielectric withstanding voltage testing using the Vitrek V73 AC/DC/IR hipot tester
FIGURE 19. Dielectric withstanding voltage testing using the Vitrek V73 AC/DC/IR hipot tester showing the device passed the 2*voltage rating for the Class II MLCC part no. C2012JB1A476M125AC. which is expected for this class of MLCC.
MLCC Temperature Characteristics Testing

As mentioned, Class II MLCCs tend to have a larger capacitance compared to Class I MLCCs. The capacitance value of Class II MLCCs changes greatly with temperature, yet for Class I MLCCs this is not the case. The following research results show how to measure the difference between these two kinds of capacitors based on temperature cycling.

The test shown in TABLE 3 on Class I capacitors was performed at 25oC, -55oC and 125oC at a frequency of 1KHz. TABLE 4 helps understand temperature coefficients for Class II MLCCs. The capacitance change rates (around 10% under -55oC) of Class II MLCCs are more obvious than Class I MLCCs. Thus, by looking at the capacitance change rate versus temperature, we can tell if a part belongs to Class I or Class II.

Energy dispersive spectroscopy (EDS) is a chemical micro-analysis technique used with scanning electron microscopy (SEM) and widely applied to research of elemental analysis and material characterization. The fundamental principle is based on the interaction between an excitation source and the specimen. Different elements have their own unique structure, and x-rays are emitted with unique peak energy forming an energy spectrum for each sample. It is a similar, but opposite, principle behind element identification using x-ray fluorescence (XRF). In both cases photon energy is emitted; with EDS they are stimulated using electrons, while in XRF x-rays do the job.

In EDS, the specimen is illuminated with an electron beam and transfers its energy to the atom, which changes the electron state in the atom. When the electron relaxes back to its original state, energy is released in the form of an x-ray photon. EDS can measure the energy and the number of x-ray photons, called counts (cnts.). Thus, we can identify the elemental composition of the specimen, but the sensitivity of the analysis depends on the atomic number of the element and the matrix it resides in.

Cross-section and Metallography

The physical cross-sectioning of an MLCC always provides essential information to aid in understanding electrical testing of the device. It physically shows the device structure and allows easy material characterization of the metal and dielectric components. Cross-sectioning is not a means but a goal. It is a destructive metallographic technique to show the internal structure for material analysis, whereas x-ray inspection only provides information on device geometry.

Metallography was originally used on metal alloys but has since been applied to a variety of materials such as plastics and ceramics. In the IC industry, metallographic techniques are often applied during failure analysis because they can reveal the internal structure of the PCB, joint terminals and electronics inside the component package. In failure testing of MLCCs when cross-sectioning is used, check if the capacitor is soldered properly to the PCB. For example, open cracks may be found in the solder joint(s) but appear during electrical testing as a device failure, resulting in a false positive.

Figure 20: Effect of DC bias on Class I MLCC capacitance for a 270pF, C0G
FIGURE 20. Effect of DC bias on Class I MLCC capacitance for a 270pF, C0G. Note that the capacitance did not change as the DC bias was swept from 0 to 40V, which is expected for this class of MLCC.
Figure 21: Class II MLCC capacitance for a 10µF, X5R
FIGURE 21. Class II MLCC capacitance for a 10µF, X5R. Note that the capacitance changed value as the DC bias was swept from 0 to 10V, which is expected for this class of MLCC.
TABLE 3. Class I MLCC Temperature Test for CL21C682JBFNNNE
Table 3: Class I MLCC Temperature Test for CL21C682JBFNNNE
TABLE 4. Class II MLCC Temperature Characteristics Codes6
Table 4: Class II MLCC Temperature Characteristics Codes

Besides electrical functionality tests, the other testing method to identify counterfeit MLCCs can be binned into two types: structural component and material composition characterization. These methods are more likely to allow observation of the entire capacitor from its external to its internal structure. The choice of tools and equipment is vital when performing material structure analysis. X-ray, x-ray fluorescence spectroscopy (XRF), optical microscopy or SEM are the most common tools used to perform material analysis. However, each has its limitations and, often, complementary methods are used to evaluate an MLCC structure and material composition.

X-ray and visual inspection by optical microscopy are typically nondestructive methods used to observe the sample; however, they are limited in the details they reveal. X-ray analysis can only show the rough structure, and visual inspection is limited to exterior information, such as the package and leads. In both cases they cannot reveal more about the materials used in the construction of the device. Therefore, cross-sectioning provides another tool to explore details of the device more completely and allows access to the internal material composition, such as the dielectric compounds used, which is critical to ascertain if an MLCC is legitimate. Cross-sectioning reveals the material grain structure and internal boundary conditions between the metal layers and the gap spacing. It is therefore an important technique to analyze the structure of MLCCs, with the proviso that it is a destructive test.

Figure 22: Material analysis of class I capacitor - C0402C0G500-470JNP
FIGURE 22. Material analysis of class I capacitor – C0402C0G500-470JNP. Note the dominant peaks are for Au, Ca, Ti, Pd, Sn, etc. Inset: relative count percentage for each element.

Metallography method. Metallographic specimen preparation can be broken down into a few steps: device mounting, sectioning, grinding, polishing, and then etching. Preparation of the sample is vital to preparing a suitable cross-sectioned device. Improper specimen preparation can contaminate the various device components, making elemental chemical analysis impossible.

Mounting the specimen encapsulates the sample in an epoxy, acrylic or polymer compound (FIGURE 23 ). This mechanically fixes the specimen within the compound to be held easily during the grinding and polishing processes and reduces contamination caused by debris migration across the sample. Additionally, the orientation should be considered when mounting the specimen. FIGURE 24 shows the cross-section of an MLCC we want to study and the mounting orientation.

A variety of methods are used to section the specimen, such as hacksawing, diamond blade cutting or a hot flame blade for larger specimens. In the metallography for small specimens, an abrasive or precision cutter is often used. As a result, mechanical and thermal damage cannot be avoided during this type of process, but if done properly it is possible to minimize damage. This allows more precise steps that follow to accurately reveal device material with minimal contamination.

During the grinding process, silicon carbide or alumina grit sandpaper is widely used. Grinding is always performed using water since it helps reduce heat damage and removes impurities during the process. The main purpose during grinding is to remove damage caused from sectioning.

Polishing is the final step needed to finish preparing the specimen for material analysis. With proper polishing, a fine, flat surface without scratches and deformation results. The choices of polishing abrasives are usually diamond, aluminum oxide and silicon dioxide. Polishing cloth is also used when performing gel polishing. Generally, low nap cloth is used for coarse polishing, and medium or higher nap cloth is used for final polishing.

After polishing, a chemical etching step is used to complete specimen preparation. The different metal elements in the part have different resistance levels to chemical solvents. After the etching process, the microstructure of the metals (and ceramic) parts are more obvious. Proper etching during metallography can be used on IC terminals, which are made of alloys. In the MLCC case, we care about the chemical composition of the materials that can only be obtained from part cross-sectioning. Therefore, we don’t need to perform etching in this case.

Figure 23: Specimens mounted in both epoxy and acrylic for cross-section processing
FIGURE 23. Specimens mounted in both epoxy and acrylic for cross-section processing.
Figure 24: MLCC before mounting in epoxy or acrylic compound
FIGURE 24. MLCC before mounting in epoxy or acrylic compound. Red line and arrows show where device is to be sectioned. (Sample dimensions: length = 1.75mm, w = 0.98mm, thickness = 0.95mm).
Figure 25: MLCC prior to sectioning
FIGURE 25. MLCC prior to sectioning. (a) Front view of capacitor under test, (b) top view of capacitor under test. (Sample dimensions: length = 2.04mm, width = 1.21mm, thickness = 0.59mm).
Figure 26: Cross-section view showing proper orientation of sample after cross-sectioning
FIGURE 26. Cross-section view showing proper orientation of sample after cross-sectioning. Note metal alloy package leads (left and right) and capacitor plate layers imbedded in dielectric (white material). (Magnification: 100x).

Case study. FIGURE 25 is an as-received original part before cross-sectioning and after cross-sectioning (FIGURE 26). Figure 26 also shows the proper orientation of the MLCC during cross-sectioning.

We use one known good device and one unknown sample device to do the comparison. FIGURE 27 and 28 show the terminal on the left side, ceramic body and the intermetallic boundary. FIGURE 29 and 30 show the dimensions of the MLCC and intermetallic boundary. We can see the composition and structure are the same for the known good device and test device, which indicates the unknown device is an authentic, i.e., non-counterfeit, part.

Figure 27: Optical micrograph of a known good device – intermetallic boundary
FIGURE 27. Optical micrograph of a known good device – intermetallic boundary. (Magnification: 400x).
Figure 28: Higher magnification optical micrograph of the unknown sample device – intermetallic boundary
FIGURE 28. Higher magnification optical micrograph of the unknown sample device – intermetallic boundary. (Magnification: 400x).
Conclusions

Counterfeit MLCC identification is challenging. Only a few papers or articles focus on this issue. Even manufacturers are unable to provide an effective way to identify counterfeit parts. We contacted many major manufacturers, and they only provide the service to verify part authenticity via the label on the parts reel, which clearly does not solve the counterfeit MLCC issue.

Our work introduces some counterfeit MLCC case studies and several methods to help identify counterfeit MLCCs. These methods are based on their electrical and physical characteristics. Based on the electrical characteristics of the target device such as high-frequency RF capacitance, we can use a test frequency sweep. Using the MLCC high-voltage rating, we can use the dielectric withstand voltage test and insulation resistance test; for soft terminal MLCCs (vibration-proof) we can use cross-section testing based on metallography methods.

MLCC physical characteristics provide a golden sample that helps identify counterfeit capacitors, especially during physical comparison (EDS and cross-section).

A combination of these test methods successfully identifies 80% to 90% of counterfeit MLCCs. Future research will focus on the capacitor mechanical characteristics, such as bent testing, vibration and mechanical shock, etc., and lifetime testing such as moisture resistance, operational life (at high temp.), and thermal shock (temperature-cycle) to keep counterfeit MLCCs out of supply chains.

Ed.: This article is adapted from a paper originally published at the SMTA/CALCE Symposium on Counterfeit Parts and Materials in August 2020 and is published here with permission of the authors.
References
  1. JIS C 5101-16-1:2009, Fixed Capacitors for Use in Electronic Equipment Part 16-1: Blank Detail Specification: Fixed Metallized Polypropylene Film Dielectric D.C. Capacitors – Assessment Levels E and EZ, February 2009.
  2. L. Callegaro, “Four Terminal-Pair Impedance Comparisons at High Frequency” Proceedings of the IEEE Instrumentation and Measurement Technology Conference, IEEE Cat. No.04CH37510, May 2004.
  3. MIL-STD-202-302, Insulation Resistance, April 2015.
  4. Wei Jiang, YongDa Hu, ShengXiang Bao, Song Lijie; Zheng Yuwei, Yongqiang Cui, and Li Qiang, “Analysis on the Causes of Decline of MLCC Insulation Resistance” Proceedings of the 2015 16th International Conference on Electronic Packaging Technology (ICEPT), August 2015.
  5. MIL-STD-202-301, Dielectric Withstanding Voltage, April 2015.
  6. EIA-RS-198, Ceramic Dielectric Capacitors Classes I, II, III and IV, November 1983.
Yung-Hsiao (Steven) Chung, engineering manager, Cheng-Hsun Lee, Liwei Xu, Yuqian Hu and ZongXuan Wang are with Global-ETS; steven@gets-usa.com. Stephen E. Saddow, Ph.D., is a professor in the Department of Electrical Engineering at the University of South Florida; saddow@usf.edu.