EOS/ESD Association Journal EOS/ESD Association, Inc. Inside this issue The Year’s Most Significant ESD Papers The EOS/ESD Association presents the year’s most significant ESD papers presented at the most important conferences that cover ESD. These papers are what we consider to be best-in-class. Introduction EOS/ESD Association, Inc. reviews and evaluates papers in the specific field of Electrostatic Discharge (ESD) that are presented at the world’s foremost conferences. This compilation comprises our selections. December 2025 Volume 4 All Papers are Reprinted with Permission © 2025 *EOS/ESD Association, Inc., **IEEE Influence of TVS Properties and Printed Circuit Board Design on System Level ESD Robustness for USB-C SuperSpeed Data Lines* In-Situ ESD Current Sensing in a Pick and Place Machine* Measurement and Evaluation of Ionizers* Latch-up Over-Current Protection Design with Power Restarted Function** ESD Robustness and Failure Analysis of Cascode GaN MIS-HEMTs Under Component-Level Test** Coupling Path Analysis of Data Center SSD Storage Systems based on Visualization Technique** Efficient Approach for System-level ESD Simulation Including Secondary Discharge** Power-Rail ESD Clamp Circuit With Modified Inverter Structure for Monolithic GaN-Based Integrated Circuits
2024 German ESD Forum Measurement and Evaluation of Ionizers Abstract: Charged plate monitors (CPMs) are used to qualify and verify ionizers. With the increasing processing of highly sensitive components, ionizer requirements are also increasing. This work investigates various ionizer parameters utilizing different CPMs. It demonstrates that today’s CPMs are not universally suitable for assessing the evolving requirements of ionizers. Citation W. Stadler, R. Pfeifle and A. Vassighi, "Measurement and Evaluation of Ionizers," 2025 47th Annual EOS/ESD Symposium (EOS/ESD), Riverside, CA, USA, 2025, pp. 1-10, doi: 10.23919/EOS/ESD65588.2025.11224284. Measurement and Evaluation of Ionizers | IEEE Conference Publication | IEEE Xplore © 2024 EOS/ESD Association, Inc. Reprinted from 2024 EOS/ESD Symposium. 2024 EOS/ESD Symposium Influence of TVS Properties and Printed Circuit Board Design on System Level ESD Robustness for USB-C SuperSpeed Data Lines Abstract: A USB-C SuperSpeed application is reproduced on custom-made printed circuit boards. The influence of the properties of discrete TVS devices and the board layout on the peak voltage at the input inside the IC and the maximum IC current is investigated by means of (VF-)TLP measurements. SEED simulations are compared with the experimental results. Guidelines for an optimal system level ESD protection for the SuperSpeed data lines of this application are presented. Citation S. Holland, N. Lotfi, M. Pilaski, B. Laue and S. Seider, "Influence of TVS Properties and Printed Circuit Board Design on System Level ESD Robustness for USB-C SuperSpeed data lines," 2024 46th Annual EOS/ESD Symposium (EOS/ESD), Reno, NV, USA, 2024, pp. 1-10, doi: 10.23919/EOS/ESD61719.2024.10702133. Influence of TVS Properties and Printed Circuit Board Design on System Level ESD Robustness for USBC SuperSpeed Data Lines | IEEE Conference Publication | IEEE Xplore © 2024 EOS/ESD Association, Inc. Reprinted from 2024 EOS/ESD Symposium. Awarded 2024 Symposium Outstanding Paper - Device In-Situ ESD Current Sensing in a Pick and Place Machine Abstract: This paper presents a method to measure current discharge directly in an assembly machine. Pre-tests with a multi-purpose current monitor are performed to show the general principle of the method. Furthermore, a novel current sensor is introduced, which can be installed directly into a manufacturing machine. Results from laboratory experiments and measurements in a state-of-the-art assembly machine show the capability of the method to improve and compare assembly machines as well as help to establish the link between voltages and currents.
Citation E. Jirutková, H. Wolf and H. Gieser, "In-situ ESD current sensing in a pick and place machine," 2024 46th Annual EOS/ESD Symposium (EOS/ESD), Reno, NV, USA, 2024, pp. 1-7, doi: 10.23919/EOS/ESD61719.2024.10702165. In-Situ ESD Current Sensing in a Pick and Place Machine | IEEE Conference Publication | IEEE Xplore © 2024 EOS/ESD Association, Inc. Reprinted from 2024 EOS/ESD Symposium. Awarded 2024 Symposium Outstanding Paper – Manufacturing, Best Paper, and Best Student Paper 2025 IEEE Symposium on Electromagnetic Compatibility & Signal/Power Integrity (EMC+SIPI) Coupling Path Analysis of Data Center SSD Storage Systems based on Visualization Technique Abstract: Electrostatic discharge (ESD) is a major source of electromagnetic interference, capable of causing damage, malfunctions, or disruptions in electronic devices. As a result, ESD immunity testing is a critical component of electromagnetic compatibility (EMC) standards. In this study, the radiation emitted from the ESD gun body over 2 GHz is characterized and modeled using Huygens' principle. The equivalent field source was validated across three different environments, demonstrating accuracy with errors of less than 10 dB. A coupling path visualization technique was then employed to identify critical coupling paths, providing guidance for the strategic placement of absorbers. Simulation results showed that applying absorbers to the identified critical areas could reduce coupling by up to 20 dB. Citation X. Su et al., "Coupling Path Analysis of Data Center Ssd Storage Systems Based on Visualization Technique," 2025 IEEE International Symposium on Electromagnetic Compatibility, Signal & Power Integrity (EMC+SIPI), Raleigh, NC, USA, 2025, pp. 184-189, doi: 10.1109/EMCSIPI52291.2025.11170206. Coupling Path Analysis of Data Center SSD Storage Systems based on Visualization Technique | IEEE Conference Publication | IEEE Xplore © 2025 IEEE. Reprinted, with permission, from 2025 IEEE Symposium on Electromagnetic Compatibility & Signal/Power Integrity. 2025 International Symposium on Physical & Failure Analysis of Integrated Circuits (IPFA) ESD Robustness and Failure Analysis of Cascode GaN MIS-HEMTs Under Component-Level Test Abstract: This study investigates the electrostatic discharge (ESD) robustness of cascode GaN MISHEMTs (metal-insulator-semiconductor high electron mobility transistors) at the component level. The experimental results revealed unstable tolerance to Human Body Model (HBM) ESD stress in the Gate-toSource zapping configuration, indicating a potential design vulnerability. The study analyzed the correlation between transmission line pulse (TLP), HBM, and ESD gun tests to assess consistency across different standards. Failure analysis was conducted to identify the locations in the G-S path that are most susceptible to ESD stress.
Citation C. -W. Su, M. -D. Ker, J. -S. Wu, C. -Y. Yang and E. Y. Chang, "ESD Robustness and Failure Analysis of Cascode GaN MIS-HEMTs Under Component-Level Test," 2025 IEEE 32nd International Symposium on the Physical and Failure Analysis of Integrated Circuits (IPFA), Bayan Lepas, Malaysia, 2025, pp. 1-5, doi: 10.1109/IPFA65338.2025.11256502. ESD Robustness and Failure Analysis of Cascode GaN MIS-HEMTs Under Component-Level Test | IEEE Conference Publication | IEEE Xplore © 2025 IEEE. Reprinted, with permission, from 2025 International Symposium on Physical & Failure Analysis of Integrated Circuits. 2025 International Symposium on Reliability Physics (IRPS) Latch-up Over-Current Protection Design with Power Restarted Function Abstract: Latch-up is a reliability issue in CMOS integrated circuits caused by the parasitic siliconcontrolled rectifier (SCR) structure. When it was triggered on, it creates a low-impedance path between the power (VDD) and ground (Vss) to burn out the silicon chip. In this work, an over-current protection design is proposed to detect latch-up occurrence, deactivate it by turning off the power, and reinitialize the circuit after a programmable delay. This design aims to mitigate latch-up hardware failure and restore the normal circuit operation through timely power cycling. The programmable delay, auto-restart function, and independent control circuit are designed to enhance both latch-up immunity and application flexibility. Additionally, the latch-up over-current protection circuit is also designed to resist false triggering caused by inrush current. Through experimental results, the over-current protection circuit can promptly shut down the power supply upon detecting an over-current event, effectively preventing irreversible damage to the chip from latch-up incidents. Citation S. -C. Huang and M. -D. Ker, "Latch-Up Over-Current Protection Design with Power Restarted Function," 2025 IEEE International Reliability Physics Symposium (IRPS), Monterey, CA, USA, 2025, pp. 1-7, doi: 10.1109/IRPS48204.2025.10983844. Latch-up Over-Current Protection Design with Power Restarted Function | IEEE Conference Publication | IEEE Xplore © 2025 IEEE. Reprinted, with permission, from 2025 International Symposium on Reliability Physics. 2025 International Symposium on Electromagnetic Compatibility – EMC Europe Efficient Approach for System-level ESD Simulation Including Secondary Discharge Abstract: This paper presents a methodology for performing efficient system-level electrostatic discharge (ESD) simulations, with an emphasis on analyzing secondary discharges. The proposed method models air discharge by combining circuit simulations and fluid-based plasma solver to simulate secondary discharge and plasma phenomena. The impedance model of the device under test (DUT) is extracted in the form of S-parameters for application in circuit simulations, and the ESD injection voltage is analyzed in the time domain. An ESD discharge is modelled by incorporating a non-linear arc resistance model, and the ESD injection voltage is extracted through transient simulations. By applying a circuit model of each component in the DUT, extracted through the PEEC solver, and the ESD injection voltage obtained
through the fluid-based plasma solver, a systemlevel circuit simulation is performed. The effect of ESD on the entire system is analyzed to identify the cause of malfunction in the DUT. This research is expected to improve the understanding of air discharge phenomena and contribute to efficient system-level ESD vulnerability analysis and countermeasures. Citation S. Jo et al., "Efficient Approach for System-level ESD Simulation Including Secondary Discharge," 2025 International Symposium on Electromagnetic Compatibility – EMC Europe, Paris, France, 2025, pp. 4044, doi: 10.1109/EMCEurope61644.2025.11176389. Efficient Approach for System-level ESD Simulation Including Secondary Discharge | IEEE Conference Publication | IEEE Xplore © 2025 IEEE. Reprinted, with permission, from 2025 International Symposium on Electromagnetic Compatibility – EMC Europe. 2025 European Symposium on Reliability of Electron Devices Failure Physics and Analysis – ESREF Power-Rail ESD Clamp Circuit With Modified Inverter Structure for Monolithic GaN-Based Integrated Circuits Abstract: A modified design of RC-inverter power-rail ESD clamp circuit realized in a 0.5-μm GaN-on-Si technology is proposed in this article. This study carefully verified the proposed design by SPICE simulation and test chip measurements. The experimental results show that the DC standby leakage under 6 V is 6.9 μA, and there is no transient leakage current during normal power-on conditions. Regarding the ESD robustness, the positive and negative TLP failure currents are 4.8 and -5.8 A, respectively. The positive and negative HBM levels are 4500 and -5500 V, respectively. All of the experiments were conducted at the standalone bare-die level without any load. Therefore, the proposed design can be integrated into GaN ICs to enhance the whole-chip ESD robustness. Citation Chao-Yang Ke, Ming-Dou Ker. Power-Rail ESD Clamp Circuit With Modified Inverter Structure for Monolithic GaN-Based Integrated Circuits. ESREF 2025 : 36th European Symposium on Reliability of Electron Devices, Failure Physics and Analysis, 2025, Université de Bordeaux, ADERA, Oct 2025, Bordeaux, France. Power-Rail ESD Clamp Circuit With Modified Inverter Structure for Monolithic GaN-Based Integrated Circuits | ESREF Conference Publication | HAL © 2025 ESREF. Reprinted, with permission, from 2025 European Symposium on Reliability of Electron Devices Failure Physics and Analysis.
Measurement and Evaluation of Ionizers Wolfgang Stadler (1), Rainer Pfeifle (2), Arman Vassighi (3) (1) Gärtner & Stadler ESD Consulting GmbH & Co. KG, Rudolf-Diesel-Str. 14, 85521 Riemerling, Germany mobile: +49 160 8935308, e-mail: esd@mail-stadler.de (2) Wolfgang Warmbier GmbH & Co. KG, Untere Gießwiesen 21, 78247 Hilzingen, Germany (3) Intel Corp., 5000 W. Chandler Blvd, Chandler, AZ 85226, United States of America Abstract – Charged plate monitors (CPMs) are used to qualify and verify ionizers. With the increasing processing of highly sensitive components, ionizer requirements are also increasing. This work investigates various ionizer parameters utilizing different CPMs. It demonstrates that today's CPMs are not universally suitable for assessing the evolving requirements of ionizers. I. Introduction Ionization is often used to reduce charges on processrelevant insulators, ungrounded conductors, or ESDsensitive (ESDS) items. Ionizers of different types are used for this purpose; the underlying technology is often based on a corona discharge in a high electric field. Ionizers with corona discharge technology can be roughly divided into four categories, depending on the high voltage used: direct current high voltage (“DC steady state”), pulsed direct current high voltage (“pulsed DC”), alternating current high voltage (“AC”) and high-frequency alternating current high voltage (“HF-AC”). Today, all ionizers are qualified according to the test methods described in the industry standards ANSI/ESD STM3.1 [1] or IEC 61340-4-7 [2]. In this process, a Charged Plate Monitor (CPM) determines the charge decay time and the offset voltage at various defined test points at different distances and spatial orientations to the ionizer. The current ESD control program standards ANSI/ESD S20.20 [3] and IEC 613405-1 [4] specify that the magnitude of the offset voltage must be less than 35 V. The discharge time (in this text referred to as charge decay time or decay time) depends on the application [3,4] or must be less than 20 s [4]. The industry standards ANSI/ESD STM3.1 [1] and IEC 61340-4-7 [2] are among the oldest ESD standards, having been applied relatively unchanged for over 30 years. However, the technology of ionizers and their applications has undergone fundamental changes during this time. This has inevitably led to various problems that will be briefly discussed in this article: 1. The CDM sensitivity of components can be lower than the offset voltage of ionizers defined in the ESD control program standards. In so-called “dieto-die processes,” the CDM sensitivity of the interfaces between the chiplets can be as low as 5 V and below, meaning the offset voltage must also not exceed this value. Can today's CPMs accurately measure offset voltages below 10 V? 2. Joshua Yoo et al. have shown [5] that today's CPMs are, at best, suitable for DC ionizers (“DC Steady State”) or DC ionizers with a very low pulse frequency (“pulsed DC”). The CPMs do not respond quickly enough to follow the changing voltage of HF-AC or DC ionizers with a rapid pulse frequency, resulting in an output peak voltage that is significantly too low. 3. The ESD standards define charge decay times from +1000 V to +100 V and from -1000 V to -100 V. In practical applications, both the initial and final voltages may deviate from these values. Ideally, the stop voltage for measuring charge decay time should consistently be the maximum permitted offset voltage. 4. The charge decay and the (time-dependent) offset voltage in the application are crucial where the charge is to be dissipated. These investigations should be conducted as part of an ESD risk assessment during the acceptance of the process. In most cases, the 150 mm × 150 mm plate electrodes (ion collecting plates) of the CPMs used for qualification are too large for this purpose. Furthermore, the plate should be approximately the same size as the ESDS item to replicate what the ESDS item “sees” in the application.
After providing a brief overview of the experimental setup and the devices used in Chapter II, Chapter III summarizes the measurement results. The following results have been selected for this work: • Measurement of small offset voltages (III.A), • Measurement of AC signals from ionizers (III.B), • Measurement of the charge decay time under different conditions (III.C), • Influence of plate size and the orientation of the plate electrode (III.D). Chapter IV summarizes several ideas for future work in the standardization committees. II. Experimental Set-up The measurements were conducted in two different laboratories. Various CPMs (voltage followers and field mill types) were utilized for these measurements. The main characteristics of the CPMs used are summarized in Table 1. For some CPMs, the plate size deviated from the 150 mm × 150 mm values specified in the standards for product qualification [1,2] (VF5, all hand-held devices HHx) or could be substituted with a plate electrode of smaller dimensions (VF1, VF2, VF3). Only units with an analog output could be used in some tests because the start and end voltages were manually evaluated using a data acquisition program. A contact voltmeter, which can measure the voltage at conductive or dissipative objects, was used for comparison. For measuring various tabletop blower ionizers outside of process equipment, test setups similar to those specified in ANSI/ESD STM3.1 [1] or IEC 61340-4-7 [2] were used (see Fig. 1). Figure 1: Measurement set-up for benchtop ionizers with a small ion collecting plate (for the voltage-follower types VF1–VF3) similar to [1] and [2]. Table 1: Compilation of the parameters of the CPMs and contact voltmeters used from the data sheets. Name Type1 Ion collecting plate (mm) Capacitance2 (pF) Voltage range (V) Vstart (V) Vend (V) DC Accuracy Bandwidth (Hz) VF1 VF 150 × 150 25 × 25 20 20 0 – 1100 programmable, 1-V steps 0.1 % full scale (1.1 V) resolution 0.1 V 80 VF2 VF 150 × 150 25 × 25 20 20 0 – 1100 programmable, 1-V steps 0.1 % full scale (1.1 V) resolution 1 V 10 VF3 VF 150 × 150 25 × 25 20 20 0 – 1020 programmable, 1-V steps 0.1 % full scale (1.1 V) resolution 1 V 80 VF4 VF 150 × 150 20 0 – 1160 250, 500, 1000 35, 50, 100 resolution 1 V 50 VF4m3 VF 150 × 150 20 0 – 100 – – resolution 1 V 1000 VF5 VF 60 × 60 48 0 – 265 250 100 resolution 1 V 50 FM1 FM 150 × 150 20 0 – 4000 500 – 1000 0 – 500 resolution 1 V 1 FM2 FM 150 × 150 20 0 – 5000 1000 – 5000 0 – 1000 resolution 1 V 1 HH1 FM 82.5 × 82.5 20 0 – 2000 (0 – 20000) 1000 10, 50, 100 5 % value resolution 1 V 5 HH2 FM 82.5 × 82.5 20 0 – 2000 (0 – 20000) 1000 10, 50, 100 5 % value resolution 1 V 5 HH3 FM 75 × 150 20 0 – 4000 1000 100 resolution 1 V 1 HH4 VF 60 × 16 12 0 – 265 250 100 resolution 1 V 50 CVM CVM – – 0 – 2000 – – 1 % full scale (20 V) resolution 1 V 1000 1 VF = voltage follower, FM = field mill, CVM = contact voltmeter, HH = hand-held device 2 Capacitance of the ion collecting plate when installed 3 Modified version of VF4 for measuring offset voltages below 100 V
III. Results A. Small Offset Voltages In the first step, a DC voltage source with an accuracy of better than 0.1% of the set voltage was used to apply voltages ranging from 0 to 20 V to the plate electrode, and the offset voltage of the CPMs was measured according to the operating instructions. The measurement duration for the offset voltage was either set to 60 s in the CPM or the value was read out after 60 s, as indicated by the graphical display of the values. If possible, the CPMs were zeroed before the measurement according to the operating instructions with the help of a grounded plate of at least 1.0 m × 1.0 m in size. The voltage follower-type CPMs, VF1-VF4, were measured using plate electrodes of 150 mm × 150 mm. The contact voltmeter measured the voltage on a metal plate measuring approximately 25 mm × 25 mm. Examples of the deviations of the measured values from the set values are shown in Fig. 2. Figure 2: Deviation of the voltage at the plate electrode of various CPMs from the DC voltage applied. The contact voltmeter closely agrees with the applied voltage over the measured range. However, based on the datasheet specification, a much greater deviation would have been expected. All voltage follower CPMs and the FM1 field mill CPM are also relatively accurate over the entire measurement range; up to an applied voltage of 15 V, the deviation is a maximum of 1 V. For a 5-V voltage, this results in a deviation of 20%. The two handheld devices in Fig. 2 exhibit deviations of 2 V to 4 V (40% to 60% at 5 V) from the applied voltages below 10 V and are, therefore, only suitable for rough estimates within this measuring range. VF1 outputs the offset voltage as the average over the selected measuring range and the difference between the maximum and minimum measured voltage values (peak-to-peak voltage). Surprisingly, the peak-to-peak voltage was always in the 10-V range. To investigate this phenomenon more closely, the plate electrode of VF1 was hard-grounded, and the offset voltage was measured again. Fig. 3 shows the result of this experiment. A moving average was calculated to reduce the number of data points. Figure 3: Measured voltage signal of the CPM VF1 with a grounded plate electrode. The CPM measures a peak-to-peak voltage of +10.4 V within the 5-s measurement period. Due to the averaging, the displayed peak values in Fig. 3 are lower (+3.4 V and -3.0 V); however, the measured voltage values still deviate considerably from the expected value of a grounded plate (0 V). The average offset voltage is -0.5 V. With these strongly fluctuating measured voltages at the plate electrode, reliable measurements of AC or pulsed DC ionizing units are impossible, at least not with small voltages, even if the measured offset voltage accurately reflects the actual value. The analog outputs built into the CPMs offer an additional measurement option. The outputs internally convert the voltage at the plate electrode into an analog output signal, typically at a ratio of 1:100 to 1:200, which can be read via an A/D converter or an oscilloscope. Fig. 4 shows an example of the analog output signal of the CPM VF1, recorded with a voltage probe on a 350 MHz oscilloscope. Figure 4: Voltage signal of the analog output of VF1 with a grounded plate electrode, recorded by a 350-MHz oscilloscope.
While the oscilloscope showed only a maximum noise of 200 mV when the voltage probe was connected directly to the grounded plate electrode, the noise of the analog output was two orders of magnitude greater. A peak-to-peak voltage of 26 V was measured at the analog output of the CPM using the oscilloscope, which is once again significantly higher than the signal directly output from the CPM on the display. The average value is about 1 V, again in the expected range. Table 2 compares the average values (offset voltage) and the peak-to-peak voltages of some selected CPMs. Table 2: Offset and peak-to-peak voltages read out at the analog output of the CPMs/CVM with a 350-MHz oscilloscope. CPM Offset Voltage (V) Peak-to-Peak Voltage (V) VF1 1.0 26 VF2 0.4 22 VF3 0.8 20 VF4 0.1 5.5 FM1 0.5 6.5 HH1 -0.5 45 HH2 0.5 16 CVM 0.1 20 With all CPMs (see Table 2), the noise level is too high to ensure, by measurement, that the maximum offset (peak) voltage seen by an ESDS item is below 5 V. Even for checking compliance with the limit for the offset voltage of ±35 V [3][4], the noise is still too high for AC or pulsed DC ionizers. The contact voltmeter also shows significant noise. Obviously, the analog output signals of the CPMs and the analog output of the contact voltmeter are not optimized for small voltages. A possible solution is presented in the following chapter. B. Small AC Signals From the bandwidth specifications in Table 1, it is evident that the CPMs can still resolve only the contact voltmeter and, with restrictions, the voltage-follower types VF1, VF3, VF4, and VF5 can be considered for measuring AC ionizers. None of the devices examined here is suitable for HF-AC ionizers operating at frequencies up to several kilohertz. The analog output of the devices is fast enough, but for measurements in the range of less than 10 V, the noise on the analog output is too large (see Table 2). Reading the voltage on the plate electrode requires a very high input impedance of the device connected to the plate electrode. This prevents charge from flowing from the plate electrode through the readout electronics. This applies to both types of CPM and the contact voltmeter. It also clarifies that connecting an oscilloscope to the plate electrode is impossible. The 20-pF capacitance of the plate electrode is discharged via a 1MΩ input with a time constant of = × = 1 MΩ × 20 pF = 20 µs. This is too short to record fast AC signals with their full, unaltered amplitude. A suitably large resistor in front of the oscilloscope input, or an attenuator, reduces the signal strength to a value below the oscilloscope's noise level. An alternative option is to utilize an operational amplifier (op-amp). The high-impedance input of the op-amp is connected to the plate electrode, and the op-amp operates as a voltage follower. The oscilloscope connects to the output of the op-amp circuit. The op-amp can be selected for the application based on the electrical parameters in the data sheet. If the voltage to be measured can be limited to 15 V, many op-amps with input impedances of 1012 Ω and higher are available. A simple example is the LMC6041 type, which has an input impedance of 1013 Ω and an operating voltage of 4.5 V to 15.5 V. This configuration can measure voltages of up to 15 V. An example of measuring a grounded plate electrode with this op-amp is shown in Fig. 5. Figure 5: Voltage signal of the output of the op-amp LMC6401 connected to a grounded plate electrode, recorded by a 350-MHz oscilloscope. Despite the simple experimental laboratory setup, the electronic noise was only 270 mV, which is nearly two orders of magnitude lower than the noise from the analog output of VF1. The offset voltage was measured at -6.5 mV. When a DC voltage source is used to apply 2.0 V to the plate electrode, exactly 2.0 V and a peakto-peak voltage of only 70 mV are measured at the output of the op-amp. Under the same conditions, FM1 measures an offset voltage of 2.2 V and a peak-to-peak voltage of approximately 5 V. If a 15-V measurement range is insufficient, other opamps can be used. As a second example, VF4 was modified to measure offset voltages up to 100 V (VF4m). With the modified VF4m, the peak-to-peak voltage of
a grounded plate electrode was 1200 mV, considerably lower than the 5.5 V of the unmodified VF4. Obviously, there is a compromise between measurement range and signal noise; the larger the offset voltage that the CPM can measure, the higher the signal noise. The circuit using the op-amp and VF4m was then measured with an AC voltage of 8 V applied to the plate electrode by an AC voltage generator at different frequencies. Fig. 6 shows the amplitudes of the AC signal recorded at the op-amp output and the analog output of VF4m as a function of frequency. Figure 6: Amplitude of an AC signal from an AC voltage generator measured using an op-amp on a plate electrode (150 mm × 150 mm) and VF4m at various AC frequencies. The measured amplitude only slightly decreased up to a frequency of about 1 kHz. Op-amps and circuits covering an even higher frequency range can undoubtedly be found. In [5], it was shown that DC ionizing units can also exhibit significant peak-to-peak voltages. In the study by [5], peak-to-peak voltages of up to 40 V at a frequency of approximately 5 kHz were measured using an oscilloscope connected to a plate electrode at a distance of 30 cm from an ionizer. Using the op-amp connected to the plate electrode, this behavior was not observed in any DC ionizers examined here, not even when the plate electrode was placed only 10 cm from the ionizers. This is a significant finding for using the DC ionizers examined in this study. It can be assumed that the ESD-sensitive component at a minimum distance of 10 cm from the ionizer does not see any harmful high-frequency voltages that could lead to ESD damage if it contacts a conductive object. The expected high peakto-peak voltages of AC ionizers [5] could not be measured with this setup because of the voltage limitation of the op-amp. The op-amp was also connected to small metal plates used in process equipment, such as those described in ANSI/ESD SP3.4 [6]. This ensures that the CPM accurately observes the charging and discharging that an ESDS of a similar size also experiences during the process. Additionally, no harmful peak-to-peak voltages were detected for the DC ionizers examined here. C. Decay Time Under Different Conditions The measurement methods for charge decay time defined in the international ESD standards for different types of ionizers aim to compare various ionizers under well-defined conditions. However, these qualification tests provide virtually no information regarding whether the ionizer will fulfill the requirements of a specific application, such as in an automated process. In a process, the voltage present before ionization and the maximum allowable voltage in the process may differ from the initial (“start”) voltage (±1000 V) and the final (“end”) voltage (±100 V) defined in the qualification test. Therefore, adjusting the start and end voltages for an acceptance test that assesses the ESD risk in a process may be necessary. For example, if ungrounded conductors in the process pose a risk, then ANSI/ESD S20.20 [3] and IEC 61340-5-1 [4] require that the charge on these conductors does not exceed 35 V. Therefore, an end voltage of 35 V should be chosen for the decay test. The initial charging of the ESDS item during the process often varies from the preset ±1000 V, so the start voltage should also be adjusted to the application-related value. Lower start values prevent over-dimensioning of the ionizers in the process; higher start values must guarantee that the charge reduction to a final value occurs within the charge reduction time specified by the process. Fig. 7 shows the charge decay time from an adjustable start voltage V0 (1000 V to 5500 V) to an end voltage of 100 V. The measurements were carried out using the CPM FM1; the ionizer was operated at two blower speeds, designated as “low” and “high”. Figure 7: Charge decay time of an ionizer at fan speeds “high” and “low” measured with FM1 from different start voltages V0 to an end voltage of 100 V. Calculations using linear (LIN) and logarithmic (LN) dependence are included for comparison; see text.
For the measurements in Fig. 7, the plate electrode was charged to the desired start voltage V0 using an external high-voltage source. The voltage at the CPM's analog output was transmitted to a data acquisition program via an analog-digital converter. Then, the time between the end of the charging process and the point at which an end voltage of 100 V was reached was determined. Fig. 8 shows the charge decay time from a start voltage of 1000 V to varying end voltages, Vend. The measurements were conducted using the CPM's integrated timer. Figure 8: Charge decay time of an ionizer at fan speeds “high” and “low” measured with FM1 from a start voltage V0 = 1,000 V to different end voltages Vend. Calculations using linear (LIN) and logarithmic (LN) dependence are included for comparison; see text. Simple physical laws can approximate the curves in Figs. 7 and 8: The ionizer emits an approximately constant ion drift current J. The ion drift current that strikes the plate electrode depends, among other things, on the distance between the plate electrode and the ionizer. This ion drift current leads to a constant charge dissipation rate on a plate electrode with an area A. At the same time, the charges on the plate electrode q(t) attract the ions through the electrostatic field (Coulomb's law); for example, ions that would otherwise not hit the plate electrode in the air stream if there were no electrostatic field from the plate electrode. These two physical effects lead to a change over time in the charges on the plate electrode q(t). They can be described mathematically in a simplified form by the following differential equation: d ( ) d + + 2 ( )=0 (1) Where k2 is a constant that accounts for the ion drift current in air due to the electrostatic field of the plate electrode E. E varies as 1/r² for a point charge, where r is the distance to the plate electrode. However, for larger plate electrodes, the dependence of E on the distance (and, thus, the electrostatic field) may differ [7]. The ion current density can be calculated using the ion velocity v, the ion density c, and the ion charge ion: = ion = 1 =const (2) The voltage V, the capacitance C of the plate electrode, and the charge q are related by the basic equation: ( ) = ( ) (3) Substituting (2) and (3) in (1) results in the differential equation d ( ) d + 1 + 2 ( ) = 0 (4) When the ion drift current dominates, (4) simplifies to d ( ) d + 1 =0 and ( )= 0 − 1 (5) With (5), the charge decay time Δt can be calculated: ∆ = − end 1 (6) Therefore, the charge decay time depends not only on the initial (“start”) and final (“end”) voltage values but also on the capacity of the plate electrode and on the constant k1, which includes, amongst other things, the plate electrode area and the ion velocity. In the case that the Coulomb attraction dominates the charge decay, (4) simplifies to d ( ) d + 2 ( ) = 0 und ( ) = 0 − 2 (7) This results in a charge decay time of ∆ = 1 2 ln 0 end (8) In this case, the charge decay time is independent of the capacity of the plate electrode in the first approximation and is dependent on the distance of the plate electrode from the ionizer through the constant k2. Fig. 8 clearly shows that at a low blower stage (low fan speed), which means a lower ion drift current, the charge reduction for the ionizing unit under examination is dominated by Coulomb attraction. At a higher blower output, more ions will strike the plate due to the drift movement in the blower for geometrical reasons, which is why a linear portion can be seen. The charge decay time also depends on the distance between the plate electrode and the ionizer. Fig. 9 shows a series of tests using a desktop blower ionizer with CPM VF1 at distances ranging from 15 cm to 60 cm. The start voltage was fixed at 1000V, and the end voltage was variable.
Figure 9: Charge decay time measured with VF1 at different distances from the plate electrode to the ionizer A from a start voltage V0 = 1000 V to a variable end voltage Vend. From the data in Fig. 9, the charge decay times from 1000 V to 35 V and from 1000 V to 10 V were set in relation to the “standard charge decay time” from 1000 V to 100 V (Table 3). Table 3: Ratios of the charge decay times from 1000 V to 35 V or from 10 V to the charge decay time from 1000 V to 100 V with different distances of the plate electrode from the ionizer. Distance (cm) Ratio of Charge Decay Time 1000 V →100 V to 1000 V → 35 V 1000 V → 10 V 15 1.31 1.57 30 1.35 1.60 45 1.30 1.60 60 1.28 1.57 If (6) is applied to the start and end voltages, then for the limiting case of charge decay exclusively through the ion drift current of the ionizer, for an end voltage of 35 V, the ratio of the charge decay time to 100 V is 1000 V − 35 V 1000 V − 100 V =1.07 and for an end voltage of 10 V 1000 V − 10 V 1000 V − 100 V =1.10 For the limiting case of charge decay exclusively through Coulomb attraction, the ratio of the charge decay time for Vend = 35 V to the charge decay time for 100 V is calculated according to (8) ln 1000 V 35 V ln 1000 V 100 V =1.46 ⁄ and for Vend = 10 V ln 1000 V 10 V ln 1000 V 100 V =2.0 ⁄ When the drift current of ions dissipates the charge on the plate electrode, it is assumed that the ion current density is always constant regardless of the distance of the object to be discharged from the ionizer. Without the influence of a field, the ion beam broadens with increasing distance from the ionizer, and thus the ion current density decreases. According to (8), the logarithmic relationship implies that the discharge from 1000 V to 100 V takes the same time as the discharge from 100 V to 10 V, or, for instance, from 350 V to 35 V. The dissipation of the charge solely through Coulomb attraction, meaning without ion drift current from the ionizer, can always be calculated as the “worst case”; in practical applications, the charge decay time is shorter due to the added effect of ion drift current. Suppose the maximum charge decay time for a start and end value is known for an application. In that case, the worst-case scenario can be calculated for any other start and end values using the logarithmic relationship. In subsequent experiments, the charge decay times of three different ionizers (A, B, and C) were measured at varying CPMs, with distances ranging from 15 cm to 60 cm. All CPMs analyzed here were fitted with a 150 mm × 150 mm plate electrode. Ionizers B and C had a similar form factor and were measured at two fan speeds: “low” and “high.” Here, “low” corresponds to the minimum adjustable fan speed, and “high” to the maximum adjustable fan speed. The airflow of ionizer B with fan speed “high” is 110 cfm, and that of ionizer C is 130 cfm. Ionizer A had a significantly smaller form factor than B and C and a much smaller airflow of 30 cfm, making it suitable for applications in process plants. It had no adjustable fan speeds. The results of the investigations are shown in Fig. 10. The charge decay times for ionizers B and C are very close; ionizer B tends to be slightly slower than C following the slightly smaller airflow. As expected, ionizer A, with a much lower airflow, is significantly slower. The measured values with the different CPMs show only a very small deviation from the mean value. For ionizers A and B, the deviation of the measured charge decay times from the mean value is less than 7% at a distance of 15 cm between the plate electrode and the ionizer and less than 3% for greater distances. This is an excellent match for the short times, as the mean value of the charge decay times at a distance of 15 cm for ionizers A and B is only approximately 1 s. For ionizer C, the deviation is significantly greater (up to 40% at a distance of 15 cm and 10% at a distance of 60 cm). This is mainly due to the deviation of VF3 from the other CPMs. VF3 appears to have difficulties in accurately measuring very short time intervals.
Figure 10: Charge decay time of different ionizers (A, B, C) at various distances from the plate electrode to the ionizer was measured at a low (top) and high (bottom) fan speed. Ionizer A has only one fan speed. Several factors contribute to the longer charge decay times at greater distances between the plate electrode and the ionizer. On the one hand, the drift velocity of the ions decreases with increasing distance due to frictional losses. As a result, fewer ions hit the plate electrode with increasing distance, and the charge decay time is extended. At a distance of 30 cm from the ionizer, typical air speeds of a tabletop blower ionizer, such as types B and C, are 0.5–2 m/s in the center of the air stream. If the distance between the plate electrode and the ionizer is increased from 30 cm to 60 cm, the air speed at ionizer C decreases by a factor of about 1.5. The measured ratios of the charge decay times, from a 30 cm to a 60 cm distance, are approximately 2 for ionizing units B and C, regardless of the blower performance. In addition to the decrease in air speed, the increase in the range of the jet from 30 cm to 60 cm probably also plays a role. For type C, the ratio of the charge decay times from 30 cm to 60 cm is approximately 3, indicating that the air speed decreases to a greater extent due to the significantly smaller fan. D. Size and Orientation of the Ion Collecting Plate In many practical applications, especially when processing integrated circuits (ICs), the 150 mm × 150 mm plate electrode is much larger than the ESDS from which the charge is to be dissipated. This means that more ions are trapped compared to the application with the large plate electrode, which could lead to a charge decay time that is too fast. Furthermore, the 150 mm × 150 mm plate electrodes are often unsuitable for use inside process equipment due to insufficient space. Therefore, smaller plate electrodes are usually better for acceptance measurements or process evaluation. Fig. 11 illustrates a series of experiments involving several CPMs with varying plate electrode sizes. Figure 11: Charge decay time of an ionizer with two fan speeds (“low”, “high”) at a distance of 15 cm (bottom) and 60 cm (top) from the plate electrode to the ionizer (type C) measured with CPMs with different plate sizes. According to the considerations in Section III.C, there are two limiting cases. If the ion current determines the charge decay, then the charge decay time is inversely proportional to the area of the plate electrode (solid lines in Fig. 11), as shown in equations (1) and (2). If the Coulomb attraction determines the charge decay, then there should be no dependence of the charge decay time on the plate size (dashed lines in Fig. 11). Fig. 11 confirms the previously obtained results that, at least up to a plate electrode size of 225 cm2 (“standard size” 150 mm × 150 mm according to [1][2]) to 68 cm2, the Coulomb attraction dominates the charge decay. For a plate size of 25 mm × 25 mm (as specified in [6]),
the charge decay time for both distances decreases significantly, although the capacitance is also 20 pF. The ratio of the charge decay times for the large to small plate electrode in this experimental setup, using the selected ionizer, is 4.9 (fan “low”) and 4.8 (fan “high”) for a 15-cm distance, and 4.5 and 4.7 for a 60cm distance, respectively. The ratio of the plate electrode areas is 225 cm2 / 6.3 cm2 37, which is much greater than the measured ratio of the charge decay times. In a second ionization device (type A) with a similar experimental setup, a ratio of the charge decay times of about 5 was also measured. This ratio cannot be explained with simple estimates. It is also apparent that the ratio can differ significantly in a different “test setup," for example, in process equipment. Fig. 11 also shows the charge decay times of two CPMs whose plate electrode capacitances differ from the “standard capacitance” of 20 pF. Predicting the charge decay times for these CPMs is even more challenging because the influence of several parameters overlaps. The orientation of the plate electrode to the air jet can also significantly influence the charge decay times. The extreme settings are the orientation of the plate electrode perpendicular to the air jet and parallel to the air jet, as shown in Fig. 12. Figure 12: Schematic representation of the orientation of the plate electrode perpendicular to the air jet of the ionizer (top) and parallel to the air jet of the ionizer (bottom). The geometric center of the plate electrode is aligned with the center of the ionizer. While the alignment is defined in the qualification tests according to [1] and [2] to compare different ionizing units, the alignment for regular compliance verification is less obvious. The standards for periodic verification, ESD TR53 [8] and IEC TS 61340-5-4 [9], require that the plate electrode be positioned at the location where the charge is to be dissipated and that it be aligned perpendicular to the airflow of the ionizing device. This orientation makes sense if the correct functioning of the ionizing device is to be checked and, therefore, compared with an initial condition. But it does not provide any information about whether the object from which the charge is to be neutralized actually sees the ion beam. Small changes in the angle of the air jet can significantly influence the charge decay time of an object parallel to the air jet. These critical changes will likely be missed when using a plate electrode perpendicular to the air jet. For acceptance testing and process evaluation, it is clear that the plate electrode must be oriented parallel to the surface of the object from which charge is to be dissipated, independent of the air jet. Fig. 13 presents examples of the ratio of charge decay times when the plate electrode is positioned parallel or perpendicular to the air jet. A positive percentage indicates that the charge decay time with the plate electrode parallel to the air jet is longer than that with the plate electrode perpendicular to the air jet. Fig. 13: Relationship between the charge decay times with the plate electrode perpendicular and parallel to the air jet for various selected CPMs. Three of the measured CPMs exhibit only a slight deviation between parallel and perpendicular alignments of the plate electrode to the air flow. It is essential to consider that the measurement arrangement may lead to deviations. The distance of the plate electrode in parallel alignment was taken from the geometric center of the plate electrode to the ionizer, so that part of the plate electrode is closer to the ionizer in parallel alignment than in vertical alignment. This can explain why, for small distances and large plate electrodes, parallel alignment leads to shorter charge decay times (FM1, VF3 with a standard plate of 150 mm × 150 mm). CPMs with a small plate electrode show a significant deviation in charge decay times at small distances between the plate electrode and the ionizer. If the plate
electrode proposed in [6] is used, a deviation of 60% is obtained at a distance of 15 cm between the plate electrode and the ionizer. This is probably because fewer ions hit the plate directly when the plates are parallel and the air velocity is high. At greater distances, the air velocity decreases, and more ions are captured by the Coulomb attraction, including ions that do not directly hit the plate, as in the case of perpendicular alignment. A significant deviation, as here for small plate electrodes and small distances, is critical for a process evaluation. Small plate electrodes, which are more likely to match the size of an ESD-sensitive device, such as an integrated circuit, and small distances from the ionizer to the ESD-sensitive device are common scenarios in automated process equipment. Here, a measurement with a large plate perpendicular to the air jet would result in a dangerous underestimation of the charge decay time. IV. Proposals for Standardization Here, various suggestions are briefly summarized based on the measurement results. The evaluation of ionizers should reveal the real danger to ESD-sensitive components that the ionizer poses in the process. One crucial parameter is the offset voltage of the ionizer, which encompasses the positive and negative peak voltages over a specified period. According to current forecasts (see, for example, [10]), so-called “die-to-die interfaces”, which have a CDM robustness of only 3 V or even below, are expected to be processed in the next few years. To ensure that the offset voltage of ionizing equipment remains below this threshold, a measurement method must accurately measure within a range of less than 5 V and detect highfrequency deflections in this voltage range. It is evident that using today's CPMs over a wide voltage range, necessary for measuring charge decay time according to [1] and [2], comes at the expense of increased noise at low voltages. One solution may therefore be to divide the applications of the CPM into two areas: a CPM that is optimized for measuring the charge decay time over a wide voltage range (for example, from 100 V to 1000 V) and a CPM that is optimized for measuring the offset voltage up to a maximum of 35 V (maximum offset voltage according to [3] and [4]) and can resolve frequencies as used in today's ionizing devices. The required measurement accuracy and bandwidth should be defined in the standards. Suppose ionization is used to neutralize the charge of ESD-sensitive components. In this case, it is essential for in-process acceptance measurements that, on the one hand, the start and end voltage can be adjusted to the actual process conditions, and, on the other hand, that plate electrodes of a size corresponding to the ESD-sensitive component are used. If the ESD-sensitive component is a printed circuit board, the 150 mm × 150 mm plate electrode is likely a good approximation; however, for ICs, much smaller plate electrodes must be used. At least for the regular testing of the ionization as described in [8] and [9], smaller plate electrodes and start and end voltages adapted to the process should be used. In this context, it is irrelevant whether or not there is a correlation between the smaller plate electrode and the plate electrode, as described in [1] and [2]. The limit values from the regular inspection depend on the application and are determined through a process evaluation. V. Acknowledgements The authors would like to sincerely thank Joshua Yoo (Core Insight) and Rita Fung (Cisco) for their thorough review of the paper and invaluable comments. VI. References [1] ANSI/ESD STM3.1-2024, Ionization. [2] IEC 61340-4-7:2017, Electrostatics – Part 4-7: Standard test methods for specific applications – Ionization. [3] ANSI/ESD S20.20-2021, Electrostatic Discharge Control Program for Protection of Electrical and Electronic Parts, Assemblies, and Equipment (Excluding Electrically Initiated Explosive Devices). [4] IEC 61340-5-1:2024, Electrostatics – Part 5-1: Protection of electronic devices from electrostatic phenomena – General requirements. [5] J. Yoo, E. Choi, E. Koo, “CPM Test Limitation Study for AC, Pulsed AC and High Frequency AC Ionizers vs. DC Based Ionizers”, 41st Electrical Overstress/Electrostatic Discharge Symposium, 2019. [6] ANSI/ESD SP3.4-2016, Periodic Verification of Air Ionizer Performance Using a Small Test Fixture. [7] W. Stadler, J. Niemesheim, S. Seidl, R. Gaertner, T. Viheriaekoski, “The Risks of Electric Fields for ESD Sensitive Devices”, 40th Electrical Overstress/Electrostatic Discharge Symposium, 2018. [8] ESD TR53-01-22, Compliance Verification of ESD Control Items. [9] IEC TS 61340-5-4:2021, Electrostatics – Part 5-4, Protection of electronic devices from electrostatic phenomena – Compliance verification. [10] Industry Council on ESD Target Levels, White Paper 2: A Case for Lowering Component Level CDM ESD Specifications and Requirements – Part II: Die-to-Die Interfaces, Nov. 2023.
www.esda.orgRkJQdWJsaXNoZXIy NzI4NjYw