Analysis of transistor damage mechanism and protection measures under lightning voltage

Abstract

In order to study the problem of damage that caused by lightning overvoltage to the transistor, we designed the experiments of combined wave impact on the transistor under different conditions by analyzing the theory of the secondary breakdown of the transistor caused by the lightning electromagnetic pulse. These cases are carried out under the condition of no protection, transient protection diode (TVS) parallel protection, and TVS tube parallel protection in the actual base amplifier circuit. It is concluded that transistor can cause it to further deteriorate under the condition of lightning overvoltage impact, and the situation of the damage can be divided into two categories, transient suppression diode in parallel with a transistor that is struck by a lightning voltage can effectively protect the transistor, which can obviously improve the resistance of the transistor. In practical amplifying circuit, applied voltage reduces the tolerance level of transistors, however, the tolerance level of transistors with transient suppression diodes in parallel is still better than the tolerance level without protection, which confirmed the conjecture that transient suppression diodes can be applied to actual circuits to protect transistors in parallel. In this paper, the theoretical analysis and experimental research on the damage of the lightning voltage on the transistor, the tolerance level of the transistor to the lightning voltage and the protection effect of the TVS tube on the transistor are carried out, which has certain reference value in the actual lightning protection of the transistor.

Keywords

Crystal triode lightning overvoltage secondary breakdown transient suppression diode

1. Introduction

Lightning overvoltage is mainly generated between clouds and clouds, clouds and land, clouds and the sky, which is due to the accumulation of different polar electric charge in different parts of the thundercloud. The electric field generated between different polar electric charges causes the thundercloud to form two charge centers with different polarities but the same amount of charge in the vertical direction. The amount of electric charge between different thunderclouds is also different. The lightning overvoltage can be generated as high as 100 million volts, and the duration is very short, so it is also called lightning shock wave. The lightning electromagnetic pulse has a wide frequency range, high energy density, strong interference and damage [1–3]. The influences of lightning wave on buildings have various aspects [4–6]. Especially for the fast-developing electronics and telecommunications industries in modern society, the direct and indirect economic losses and social impacts caused by lightning hazards seriously hinder the development of electronics and telecommunications industries. Although the transistor device has slow switching speed and large power consumption, it has small size, low power consumption, and high reliability, which can be widely applied to fields such as broadcasting, television, radar, and automatic control devices. However, with the widespread use of transistors in electronics, communications, and other industries, lightning damage caused by transistor overvoltage is difficult to avoid, and there are more and more trends.

At present, there are not many studies on the damage of transistors under the action of lightning voltage at home and abroad. Some scholars have analyzed and tested the electromagnetic pulse damage of transistors [7–12]. Yang [13] creatively pointed out that the most sensitive to electrostatic electromagnetic pulse is the current collecting junction. Not only that, he also mentioned that the thermal secondary breakdown leads to the damage and failure of the triode. Xi [14] found that the damaged part of the transistor is related to the electromagnetic pulse intensity, and the damage energy of the transistor subjected to strong electromagnetic pulse is not constant. Xi [15] further pointed out that the damage of the external circuit components to the transistor is due to the secondary fault of the current mode of the transistor. Chai [16] concluded that changing the magnitude of the injected voltage will cause the transistor to burn out. Hwang et al. [17] based on experiments show that semiconductor devices in the chip will have different degrees of damage at low radiation intensity (11.9 kV/m), and when the radiation intensity is high (greater than 12.6 kV/m), it will also cause chip interconnects and Bonding thermal stress damage. Zhao [18] studied the influence of electromagnetic pulses injected from different poles on the damaged parts of the transistor, and found that the transistor is more vulnerable to damage from the base.

In order to study the damage of the lightning voltage on the transistor and how to effectively protect the electromagnetic pulse of the transistor and find a simple method to realize the transistor protection, we will analyze the above problems theoretically and design experiments to verify that the protection is effective.

2. Theoretical analysis

A transistor, also called a semiconductor transistor, is a semiconductor device. The transistor has the function of amplifying the signal times by adjusting the magnitude of the current, and the essence is to amplify the amplitude of the signal without changing other properties of the signal. The basic function of the transistor is to amplify the current and also owns the function of a switch. When it is saturated, the voltage between the emitter and the collector is very small, which is equivalent to the switch being turned on. When it is turned off, the current between the emitter and the collector is extremely small, which is equivalent to the switch being turned off. In some special cases, the transistor can also achieve the functions of current expansion and substitution.

In recent years, there have been studies on the effects of electromagnetic pulses on transistors at home and abroad. Shockley [19] presented the theory for a new form of transistor. This transistor is of the “field-effect” type in which the conductivity of a layer of semiconductor is modulated by a transverse electric field. It has been found that transistor damage is mainly caused by the destruction of the PN junction, and secondary breakdown is the main cause of transistor failure or damage. Since the second breakdown phenomenon discovered by two researchers, Trinton and Simmons, in 1957, scholars have been studying this phenomenon and proposed various theories to explain the mechanism of secondary breakdown. Some scholars try to analyze the causes of heat, while others consider it from the perspective of electric current, and these two mechanisms are generally referred to as thermally unstable secondary breakdown and avalanche injection type secondary breakdown. Most theories show that secondary breakdown is related to hot spots. When the energy applied to the power tube exceeds the critical value, the internal parts of the device are heated due to local concentration of the electric current, a “hot spot” appears, and the power dissipation is also increased. If the transistor does not dissipate heat in time, it will cause positive feedback of current, which further increase the current and form a hot spot, and finally a vicious circle causes damage to the device. The second breakdown may occur in various operating states of the transistor. The mechanism can be divided into two cases. Firstly, as shown in Fig. 1, when the transistor is off, e-b is positively biased secondary breakdown. Secondly, as shown in Fig. 2, when the transistor is off, e-b is reverse biased secondary breakdown.

Fig. 1.

Positively biased secondary breakdown.

Fig. 2.

Reverse biased secondary breakdown.

When the transistor is biased into the working area, a transverse electric field is generated in the base region due to the lateral current and lateral resistance of the base, so-called base self-biasing effect. When the transistor is biased into the working area, a transverse electric field is generated in the base region due to the lateral current and lateral resistance of the base, so-called base self-biasing effect. At this time, the current from the emitter is concentrated in the narrow region at the edge of the emitter because of the action of the transverse electric field. When this current flows through the b-c junction space charge region, effective heat is generated, forming a local hot spot. If the current does not stop, this hot spot begins a vicious circle, and makes the local current density higher, causing a secondary breakdown and eventually burning. The narrower the width of the base region, the higher the applied voltage V _ce, and the more severe the hot spot.

When the e-b junction is reverse biased, the direction of the lateral electric field caused by the current flowing through the transistor is opposite to that of the positive bias, so that the current flowing into the emitter is concentrated to the middle of the emitter. Due to the higher concentration of local currents formed during reverse bias, the reverse bias secondary breakdown critical energy is much lower than the positive bias. When the e-b junction is in the reverse bias, the occurrence of the secondary breakdown is related to the current, voltage and pulse action time, because the base transverse electric field is related to the cutoff voltage V _be and the series resistance R _be.

Impacting the transistor with electromagnetic pulses of different intensities, the part where the damage occurs is the part generated by the hot spot changes. The low-intensity electromagnetic pulse will cause the n-n+ under the emitter to be damaged first, while the high-intensity electromagnetic pulse will cause the base to be damaged near the side of the emitter. The difference in the location where the hot spot is generated will also cause the damage energy of the transistor to be divided into a constant value region and a rising region. When the damage energy of the transistor is in the rising region, increasing the intensity of the electromagnetic pulse is also difficult to affect the damage time of the transistor.

3. Experiment analysis

3.1. Transistor breakdown experiment analysis

In order to prevent the polarity of the triode from causing the error of the test, we use four different types of triodes, 9013, 1815, 9015 and 3906. Among them, 9013 and 1815 are NPN type transistors, and 9015 and 3906 are PNP type transistors. The combined wave generator (1.2/50 μs, 8/20 μs combined wave) is used to simulate the lightning impact experiment on the b-e and c-e ends of each triode, and we install an oscilloscope at both ends of the experimental triode to show changes in the impact. Since the power of the transistor is small, the impact voltage is set to start from 0.1 kV, and each time 0.1 kV is added until the transistor is completely damaged. Record the characteristics of the transistor after impact and its parameter changes at each impact, and collect its voltage and current waveforms. The experiment circuit diagram is shown in Fig. 3.

Fig. 3.

Damage experiment diagram of the transistor under lightning protection without any measures.

Based on the performance and principle of the crystal triode, transient suppression diode and lightning combined wave generator, we built a comparative test and place the triode in a 12 V amplifier circuit. The base resistor of the triode has a resistance of 10 kΩ, the collector resistor R _c has a resistance of 1.5 kΩ, and the emitter is grounded. As in Figure 1, the combined wave generator is used to test the forward and reverse directions of the two ends of the transistors b-e and c-e, and the voltage and current changes at both ends are observed by the oscilloscope, as shown in Fig. 4.

Fig. 4.

Damage to the transistor in the amplifier circuit due to lightning strike.

The impulse voltage value of each experimental combination wave generator is increased by 0.1 kV based on the voltage value of the previous experiment, and repeating the above steps until the transistor is completely damaged. The specific performance of the triode’s complete damage is the violent explosion and the bright white light. There are visible obvious cracks on the transistor and the measured resistance between the two poles of the triode drops to 0 Ω. The experiment results of the transistor being damaged in the unprotected state are shown in Table 1:

Table 1

Impulse voltage values when four transistors are damaged in the state of unprotected device

Type	SS9013(NPN)	2NC1815(NPN)	SS9015(PNP)	2N3906(PNP)
Ube (Maximum voltage value)/kV	1.9 kV	1.4 kV	1.3 kV	1.6 kV
Ueb (Maximum voltage value)/kV	1.3 kV	0.6 kV	1.7 kV	1.7 kV
Uce (Maximum voltage value)/kV	1.2 kV	2.8 kV	1.2 kV	1.3 kV
Uec (Maximum voltage value)/kV	0.9 kV	2.7 kV	1.2 kV	1.1 kV

Figures 5–7 show that some representative waveforms of an unprotected transistor during the impact process. Figure 5 is a voltage waveform diagram when a voltage of 0.1 kV is applied across the transistor. It is apparent that the waveform conforms to the parameters of the 1.2/50 μs waveform. Figure 6 is a voltage waveform diagram when the voltage across the transistor is applied to 0.5 kV, and the transistor at this time is still working properly, the voltage waveform is similar to that of Fig. 5. It can be clearly seen in Fig. 7 that the voltage waveform is no longer a normal lightning waveform, because the waveform at this time is the voltage waveform when the transistor is blown up.

Fig. 5.

Voltage waveform when a 0.1 kV voltage is applied to an unprotected transistor.

Fig. 6.

Voltage waveform when a 0.5 kV voltage is applied to an unprotected transistor.

Fig. 7.

Voltage waveform when the unprotected transistor is blown up.

When the transistor is placed in the amplifier circuit to test the lightning strike, the amplifier circuit operates normally, but as long as a voltage is applied across the transistor, the transistor will burst immediately. Even if a voltage of 0.1 kV is applied, the transistor will burst, which is the result of several sets of experiments.

We can find that there are two cases of lightning overvoltage damage to the transistor. First, the transistor is completely damaged. The main performance is that the transistor emits an explosion sound and bright white light. The appearance has obvious cracks, the resistance between the electrodes drops to 0 Ω, and the voltage waveform changes. Second, the transistor is not completely damaged, showing a decline in performance parameters, no explosion and white light, and some fluctuations in the voltage/current waveform. Both explosive and white light are manifestations of thermally unstable secondary breakdown.

These four different types of transistors are very resistant to lightning over-voltage and can withstand lightning voltage surges of nearly a few kilovolts or even thousands of volts. Among them, the b-e end of the 1815 transistor and the c-e end have a large difference in lightning withstand voltage, and the difference between the two values can reach more than one thousand volts. Comparing the lightning voltage withstand capability of the four transistors, it can be found that the b-e, e-b, c-e, and e-c terminals have the highest withstand voltages of 9013 (NPN), 9015 (PNP), and 3906 (PNP), and 1815 (NPN), and be endurance is stronger than c-e end tolerance.

3.2. Transistor breakdown protection test analysis

In order to study the transistor breakdown protection measures, we designed the experiment of lightning protection with transient suppression diode parallel protection transistor, and the experiment of diode parallel protection transistor in the lightning impulse basic amplification circuit. We protect the transistor with the simplest protection device and observe whether the protection device can protect the transistor. Then put the transistor into the basic amplifier circuit to verify that the protection device can be applied in the actual circuit. The experimental data of the two experiments were compared to analyze whether the transient suppression diode has a protective effect on the transistor.

We connected a TVS tube in parallel with the tested end of the transistor. Similarly, the combined wave generator is used to perform the impact test in the forward and reverse directions on both ends of b-e and c-e of each triode, and we install an oscilloscope at both ends of the impact, as shown in Fig. 8.

Fig. 8.

Damage test of lightning protection with TVS tube parallel protection transistor.

After the circuit is connected, the transistor impact test under the condition of transient suppression diode protection is performed. The voltage starts from 0.1 kV and increases by 0.1 kV each time until the performance parameter of the transistor decreases. Use a multimeter to measure the resistance between the poles to fall below the original measured resistance, below 400 Ω, even as low as 0 Ω. The test results obtained are shown in Table 2.

Table 2

Impulse voltage values when four transistors are damaged in the parallel protection state of TVS tube

Type	SS9013(NPN)	2NC1815(NPN)	SS9015(PNP)	2N3906(PNP)
Ube (Maximum voltage value)/kV	2.5 kV	6.0 kV	6.0 kV	6.0 kV
Uce (Maximum voltage value)/kV	3.7 kV	6.0 kV	6.0 kV	6.0 kV

The resulting experimental current waveform is shown in Figs 9 and 10.

Fig. 9.

Current waveform diagram of 1.2 kV voltage surge with TVS tube shunt protection transistor.

Fig. 10.

Current waveform diagram of 3.7 kV shock with TVS tube shunt protection.

In the experiment of impact on transistors with diode protection in parallel in the amplifier circuit, which is shown in Fig. 11. We still put the triode in a 12 V amplifier circuit. The base resistor R _b of the transistor is 10 kΩ, the collector resistance R _c is 1.5 kΩ, and the emitter is grounded. The TVS tube is connected in parallel between the two sets to be measured. The combined wave generator is used to impact the two ends of the transistor to observe the change of the voltage waveform on the oscilloscope.

Fig. 11.

Damage experiment of TVS tube parallel protection transistor in lightning impulse basic amplification circuit.

We connect the circuit and repeat the experimental steps of the previous experiment until the performance parameters of the transistor decrease. That is, the resistance between the bases drops below the originally measured resistance, below 400 Ω, and even as low as 0 Ω. The experimental results are shown in Table 3.

Table 3

Impact voltage values of four transistors in the amplifying circuit when the TVS tube is in parallel protection state

Type	SS9013(NPN)	2NC1815(NPN)	SS9015(PNP)	2N3906(PNP)
Ube (The transistor explodes)/kV	1.5 kV	2.5 kV	1.5 kV	4.0 kV
Ube (Maximum voltage value)/kV	2.0 kV	2.5 kV	3.5 kV	4.5 kV
Uce (The transistor explodes)/kV	–	1.5 kV	2.5 kV	–
Uce (Maximum voltage value)/kV	6.0 kV	6.0 kV	6.0 kV	5.0 kV

Table 4

Junction resistance before and after testing of four transistors

Resistance value	Resistance type		SS9013(NPN)	2NC1815(NPN)	SS9015(PNP)	2N3906(PNP)
Before/Ω	Emission junction		683	690	697	693
	Collector junction		681	681	693	688
After/Ω	Emission junction	Transistor	683 (0 Ω at 2.5 kV)	683	683	683
		Transistor in the amplifier circuit	683 (0 Ω at 2.0 kV)	683 (0 Ω at 6.0 kV)	683	683
	Collector junction	Transistor	683 (0 Ω at 3.7 kV)	683 (0 Ω at 2.5 kV)	683 (0 Ω at 3.5 kV)	683 (0 Ω at 4.5 kV)
		Transistor in the amplifier circuit	683	683	683	683 (0 Ω at 5.0 kV)

Comparing the data of the above two types of experiments, we can find that the transient suppression diode has a protective effect on the transistor. We deeply analyze the test in 3.2 and test the device parameters before and after the experiment.

We use a transistor tester to display various characteristic curves of a semiconductor device and measure its static parameters. The unit of the Y coordinate is 200 μs per degree, and the unit of the X coordinate is 1 V per degree. Figure 12 and Fig. 13 show that the output characteristics of the transistor before and after the experiment. It can be clearly seen that at the same collector voltage, the collector current of the impact transistor is slightly smaller than the collector current of the transistor before the experiment. However, the overall performance parameters are maintained well, with little change, and still have good output characteristics.

Fig. 12.

Output characteristics of the transistor before the experiment.

Fig. 13.

Output characteristics of the transistor after the experiment.

By observing the experimental data and waveforms, we can find the following characteristics of the experimental results. After being protected by the parallel suppression of the transient suppression diode, the ability of the transistor to withstand lightning overvoltage surge is significantly improved. In addition to the 9013 (NPN) tolerance is not obvious, the other three types of transistors are outstanding, and they can withstand a lightning surge of 6 kV while keeping the performance parameters of the transistor unchanged.

When the transistor is placed in an amplifying circuit, the transistor is more susceptible to damage than when it is impacted alone, because that the applied voltage makes the transistor more vulnerable. Although the withstand voltage capability of the transistor in the amplifying circuit is reduced, it is still better than the case where there is no transient suppression diode in parallel protection. The transistor has many huge explosions and bright white light, but it is not damaged immediately. And it has no cracks in appearance and the resistance value remains the same.

4. Conclusion

Through the experiment of lightning voltage withstand capability under different conditions for two types of four types of transistors 9013, 1815, 9015 and 3906, we can get the following conclusions:

(1) Lightning overvoltage can cause damage to the transistor, and the main cause of damage is thermal unstable secondary breakdown. The damage can be divided into two categories: one is that the transistor is completely damaged and the resistance is reduced to 0 Ω; the other is that the transistor is not completely damaged, and only the performance parameter is degraded.

(2) Parallel connection of the transient suppression diode to the transistor can effectively protect the transistor and significantly improve its withstand lightning voltage impact level.

(3) In the actual amplifying circuit, although the applied voltage reduces the tolerance level of the transistor, the tolerance level of the transistor with the transient suppression diode in parallel protection is still higher than that without the protection. This also confirms our hypothesis that the transient suppression diode can be applied to the actual circuit to protect the transistor in parallel.

Footnotes

Acknowledgements

The work was financially supported by the Postgraduate Research & Practice Innovation Program of Jiangsu Province, the Natural Science Foundation of Jiangsu Province (Grants No. BK 20150903) and the Natural Science Fundamental Research Project of Jiangsu Higher Education Institutions of China (Grants No. 15KJB170010).

References

Chen

Zhou

and Wang

, Research on energy coordination between multi-level surge protective devices with the transmission line theory, International Journal of Applied Electromagnetics and Mechanics 48 (2015), 345–356, doi:10.3233/JAE-140121.

Zeng

and Wang

, Simulation calculation and analysis of lightning induced overvoltage on power distribution lines, High Voltage Engineering 2 (2013), 415–422.

Zhou

Chen

and Yang

, Propagation and suppression method of lightning wave in coaxial line, International Journal of Applied Electromagnetics and Mechanics 49 (2015), 315–325, doi:10.3233/JAE-140130.

Yang

Zhang

and Jiang

, Aperture coupling analysis on shielding enclosure excited by lightning electromagnetic pulse, International Journal of Applied Electromagnetics and Mechanics 48 (2015), 399–408, doi:10.3233/JAE-140182.

Boumaiza

and Labed

, Lightning electromagnetic field coupling to overhead power distribution lines, International Journal of Applied Electromagnetics and Mechanics 47 (2015), 1093–1106, doi:10.3233/JAE-140144.

Zhu

Chen

and Zhang

, Coupling between over-head transmission lines and lightning electromagnetic fields, International Journal of Applied Electromagnetics and Mechanics 54 (2017), 553–567, doi:10.3233/JAE-160089.

Connell

R.M.

Shiue

and Hoft

R.G.

, Numerical simulation of avalanche breakdown in PIN diodes and bipolar transistors, Industry Applications Society Annual Meeting 1 (1992), 1061–1065.

Arnborg

, Two-dimensional avalanche simulation of collector-emitter breakdown, IEEE Transactions on Electron Devices 37 (1990), 2009–2101.

Fan

Zhang

and Jia

, HPM damage mechanism on bipolar transistors, High Power Laser and Particle Beams 22 (2010), 1319–1322.

10.

Chen

and Gong

, Influence of circuit during injection of EMP into bipolar junction transistor, High Power Laser and Particle Beams 19 (2007), 1197–1202.

11.

Ren

Chai

Yang

Qiao

and Shi

, The damage effect and mechanism of bipolar transistors induced by injection of electromagnetic pulse from the base, Acta Physica Sinica 62 (2013), 464–469.

12.

Wang

Zhou

and Zhang

, Damage effect of high frequency electromagnetic wave on bipolar transistor, Ship Electronic Engineering 2017(2) 135–139.

13.

Yang

, Experimental Studies on Electromagnetic Pulse Effect of Electrostatic Discharge for the Semiconductor Triode, Hebei Normal University, Shijiazhuang, Hebei, China, 2008, pp. 1–50.

14.

, Research on the Damage Mechanism of the Bipolar Transistor Induced by the Electromagnetic Pulse, Xidian University, Xian, Shanxi, China, 2010, pp. 2–59.

15.

Chai

Ren

Yang

and Wang

, Influence of external component on the damage of the bipolar transistor induced by the electromagnetic pulse, Journal of Semiconductors 7 (2010), 49–52.

16.

Chai

Ren

Yang

and Ma

, The damage effect and mechanism of the bipolar transistor induced by the intense electromagnetic pulse, Acta Physica Sinica 11 (2010), 8118–8124.

17.

Hwang

Hong

and Huh

, Characterization of the susceptibility of integrated circuit with induction caused by high power microwaves, Progress in Electromagnetics Research 81 (2008), 61–72, doi:10.2528/PIER07121704.

18.

Zhao

Zhou

and Liu

, The damage mechanism model and simulation of bipolar junction transistor caused by electromagnetic pulse, Journal of Microwaves S2 (2014), 97–100.

19.

Shockley

, A unipolar “Field-Effect” Transistor, Proceedings of the IRE 40 (1952), 1365–1376 doi:10.1109/JRPROC.1952.273964.