AN1628/D Understanding Power Transistors Breakdown Parameters Prepared by: Michaël Bairanzade http://onsemi.com APPLICATION NOTE TUNNELING EFFECT When the electrical field approaches 106 V/cm in Silicon, a significant current begins to flow by means of the band to band tunneling process. To get such a high field, the junction must have relatively high doping concentration on both side of the p−n junction. As a matter of fact, the field may be so high that it creates enough force on a covalently bounded electron to free it. This creates two carriers, a hole and an electron to contribute to the current. From an energy−band point of view, in this breakdown process an electron makes a transition from the valence band to the conduction band without the interaction of any other particles. The breakdown mechanism for silicon device with breakdown voltage lower than 4Eg/q comes from the tunneling effect. For junctions having breakdown voltage in excess of 6Eg/q, the mechanism is generated by avalanche multiplication. At voltages between 4Eg/q and 6Eg/q, the breakdown is a combination of the two effects. Since the energy band gap in silicon decreases with temperature, the breakdown voltage due to tunneling effect in these semiconductors has a negative temperature coefficient. This temperature effect can be used to differentiate the tunneling effect from the avalanche mechanism which has a positive temperature coefficient. One must point out that the tunneling effect is dominant for zener voltage in the 5 V to 6 V range. This mechanism is not present for high voltage junctions since the doping concentration is too low to generate the tunneling effect. INTRODUCTION Among the electrical parameters of a Bipolar power transistor, the breakdown related ones are the most critical to measure. As a matter of fact, as the breakdown voltage can be pretty high, the instantaneous power dissipated during the test must be accurately controlled to avoid a local hot spot on the chip. On the other hand, the breakdown mechanisms are prone to high frequency oscillations and care must be observed to accurately measure this parameter. This application note gives a definition of the breakdown parameters and the associated physics of semiconductor devices. A chapter describes the test techniques and methods recommended to perform this kind of measures. Since most of the high voltage bipolar transistors are NPN type, all the analysis carried out in this note will be referenced to NPN devices. The PNP mechanisms are essentially the same, with reverse polarities of all biases. BREAKDOWN MECHANISM Let us first consider the Collector to Base junction. Like a p−n diode, the avalanche process limits the collector−base voltage the transistor can sustain. When the voltage is large enough, the n−side depletion region will reach the n+ contact diffusion and, if the voltage increases further, the contact will supply electron to the p−n junction. At this point, the Collector−Base junction is essentially shorted and the current is limited by the external resistances. Basically, three mechanisms control the breakdown: thermal instability, tunneling effect and avalanche multiplication. AVALANCHE MULTIPLICATION This mechanism, also named impact ionization, is the most important one in junction breakdown since the avalanche breakdown voltage imposes an upper limit on the reverse bias of the collector voltage. The value at which breakdown occurs depends on the structure of the junction and the dopant concentration used to manufacture the transistor. Both the structure and the epitaxy being controlled, the breakdown voltage of a given semiconductor is reasonably predictable at the design stage. THERMAL INSTABILITY The breakdown for thermal instability is the result of large increase of leakage current when the junction temperature increases. Germanium based devices are more sensitive to this mechanism compare to the Silicon. The process snowball as the leakage current increases, yielding higher junction temperature which can lead to the destruction of the chip if it extends above the melting point of the semiconductor material. Semiconductor Components Industries, LLC, 2003 August, 2003 − Rev. 2 1 Publication Order Number: AN1628/D AN1628/D E gained from the field E by the moving electron between collisions is: In this process, free carriers can gain enough energy from the field between collisions to break covalent bonds in the lattice when they collide with it. Consequently, every carrier interacting with the lattice creates two additional carriers and the mechanism snowballs as three carriers can also participate to the collisions. This leads to a sudden multiplication of carriers in the space−charge region when the electrical field becomes large enough to trigger the avalanche. Let us consider an electron traveling in the space−charge region of a reverse biased pn junction. The electron travels, on average, a distance L (the mean free path) before losing energy by interacting with an atom in the lattice. The energy L E q E. dx (1) 0 Assuming the electron has accumulated enough energy from the field prior colliding with an atom, the bond between the atom core and one of the bound electron can be broken during the collision. This creates three resulting carriers which are free to leave the region of the collision. A simplified schematic representation is given in Figure 1. REVERSE BIAS Note: drawing is not scaled FREE ELECTRON N+ P− X E N− EXTRA ELECTRON EXTRA HOLE ELECTRIC FIELD COLLISION COLLECTOR−BASE SPACE CHARGE REGION Figure 1. Simplified Representation of the Avalanche Process in a Junction Consequently, the avalanche is confined to the central area of the space charge region where the field is sizable. This is represented in Figure 2, the central region being identified by x1. Near the edge of the space charge region, the electric field is low and practically no carriers can gain enough energy from the field to create a hole−electron pair before they lose their kinetic energy in a collision with the lattice. x dx xp xa xc xn no x1 nf no + n1 p2 E x1 E (b) Carriers pair are created in the region dx at x, by electrons flowing from the left and hole flow− ing from the right (a) Ionization occurs in the high field area x1 of the space charge region Figure 2. Ionization Process Location for a Non Punch Trough Structure http://onsemi.com 2 AN1628/D Let us consider the number of carriers created by avalanche in a small volume of width dx located within x1 at x (see Figure 2b). Let no = density of electrons entering x1 from the left at xa.The avalanche phenomenon will increase this density between xa and xc in a way that the electrons entering the volume Adx at x from the left have a density no + n1. The probability that electrons create electron−hole pairs when travelling through dx is given by the product of the ionization coefficient n, times the length dx. Since the electrons gain energy more rapidly when the field is higher, the ionization coefficient is a function of the electric field, hence of the position in the semiconductor. In the same manner, the density of holes will increase as a consequence of the avalanche, and the p factor represent the ionization coefficient for the holes. Let nf be the density of electrons that reaches xc: n no n1 n2 f The density of ionizing collisions at x is proportional to n* which represent the density of excited electrons arriving at x with enough energy to create the electron−hole pair. The density n*, in turn, is the total electron density n times the probability that an electron has not collided in a distance d necessary to get the right energy: n* n exp d L d E1 qE (2) (3) xc dx M xa with (9) 2n6 Vr: the reverse voltage applied across the In the common Base configuration, the breakdown due to the impact ionization (avalanching) gives a well defined voltage as depicted in Figure 3a, the effect of the current coming from the emitter having little effect on the breakdown. (5) xa As the integral in the denominator of equation 5 approaches unity, the M factor increases to infinity. Therefore, the avalanche will occur when: xc dx 1 n Bv: the breakdown voltage. xc dx junction 1 1 1 1 Vr Bv (4) The ratio of the density of electrons nf leaving the space−charge region to the density no entering is called the multiplication factor M: n M nf o (8) Since the ionization coefficient depends strongly on the electric field, the multiplication factor also increases rapidly with field. Consequently, a small increase in field, as the voltage approaches breakdown, causes a sharp increase in current as observed on a curve tracer. Not only does the ionization coefficient vary with field, hence with the position, in the space charge region, but the width of the space charge region varies also with voltage. Therefore, the evaluation of M with equation 5 is difficult and a good approximation can be derived from the simplified equation 9: Assuming that n = p, we can rewrite equation 3 as follows: n no n f f (7) The length d may be derived from equation 1 by letting E1 being the minimum energy necessary to create the ionization process, and E being the average field that accelerates the electron: where n2: density of electrons created between x and xc. Since electrons and holes are created in pair, n2 = p2 , we may write: dn ( )(n n ) n n p o p f 1 dx (6) xa http://onsemi.com 3 AN1628/D Ic COMMON EMITTER CONFIGURATION COMMON BASE CONFIGURATION Ic BVCBO (a) Under breakdown, the emitter current is not affected BVCEO (b) At low impact ionization rate, the base current starts to increase, yielding an emitter current increase and, eventually, a collector current runaway Figure 3. Common Base and Common Emitter Breakdown Typical Curves device voltage capability with a simultaneous internal constriction of current. Figure 4 shows a typical Ic/Vc characteristic of a transistor under second breakdown. The avalanche breakdown (also named first breakdown) occurs when the voltage applied across the junction reaches the value specific for this device as discussed above. As the voltage is increased, the second avalanche occurs. For the common Emitter configuration, the breakdown is not as sharply reflected in the transistor output characteristics as depicted in Figure 3b, and occurs at a lower voltage BVCEO. In this configuration, as soon as the impact ionization process starts, the secondary holes are injected into the base, yielding a base current which, in turns, leads to an increase in the emitter current and a current runaway may eventually occurs as the mechanism snowballs. SECOND BREAKDOWN The use of power transistor is limited by the second breakdown phenomenon which is an abrupt decrease of the SECOND BREAKDOWN Ic AVALANCHE Vr Figure 4. Typical Second Breakdown Characteristic significantly reduced, the chip is at high temperature and the semiconductor is in the intrinsic mode near the breakdown spot location. If the current keeps increasing, the silicon melts and the transistor is permanently destroyed. The initialization of this mechanism is essentially caused by the temperature effect. After a time delay following the power pulse P = Ic*BVCEO applied to the transistor, the device goes into the second avalanche area. During the time elapsed from avalanche to second breakdown, the junction is unstable and may be rapidly downgraded. In particular, the resistance of the breakdown spot becomes extremely low. During the third part of the curve, when the voltage is BASIC POWER TRANSISTOR GEOMETRY The simplified cross section of a power transistor given in Figure 5 illustrates the main areas involved in the http://onsemi.com 4 AN1628/D emitter stripes diffused into the base P−. Such technology is mandatory to get the expected static and dynamic characteristics requested in today’s applications: the photo given in Figure 6 illustrates this concept. breakdown process. When the voltage rises across the transistor, the electrical field is sustained by the N−epitaxy, the depletion region increasing accordingly. One must point out that, in reality, a modern transistor is not built with a single junction as depicted in Figure 5, but with several OHMIC CONTACT EMITTER BASE ÎÎÎ ÎÎÎ P− N+ ÎÎÎÎÎÎÎ ÎÎÎÎ ÎÎÎÎÎÎÎ ÎÎÎÎ SiO2 N− N+ COLLECTOR−BASE DEPLETION REGION COLLECTOR SUBSTRATE Figure 5. Simplified Cross Section of a Bipolar Power Transistor BASE EMITTER Figure 6. Photo of a Typical Power Transistor Using the PLANAR Termination On top of that, the electrical field must not collapse at the edge of the chip, where the base diffusion no longer exists. To prevent this mode of failure, the transistors are designed with a junction termination suited for the kind of device under development. A junction termination example is given in Figure 7 http://onsemi.com 5 AN1628/D SIPOS SiO2 GUARD RING N+ ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ EMITTER −>N+ BASE −>P− COLLECTOR −>N− SUBSTRATE −>N++ Typical PLANAR Junction Termination CHIP END MOAT SiO2 GLASS ÎÎÎÎ ÎÎÎÎÎÎÎÎÎÎÎÎ ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ EMITTER −>N+ BASE −>P− COLLECTOR −>N− SUBSTRATE −>N++ Typical MOAT Junction Termination Figure 7. Typical Junction Terminations for a High Voltage Bipolar Transistor destruction, occurs if the operating point moves outside the specified limits. At the time of printing this document, the PLANAR process is the most widely used technology to manufacture power transistors. However, the MOAT termination is still used for very high voltage device with BVCBO of 1500 V and above. FBSOA The FBSOA curves are specified to define the maximum safe operating area for a given device. This parameter depends on the chip geometry and the process used to built the transistor. Particularly, the emitter fingers perimeter is extremely important since it highly influences the current spread across the silicon. As a matter of fact, at turn on, the current starts to flow from the periphery of the emitter finger, as depicted in Figure 8 The rbb’ increases as we move toward the center of the emitter finger and the Base−Emitter junction is not fully forward biased during the first microseconds. Consequently, the current density is maximum at the edge of the junction and the emitter becomes prone to local hot spots. DYNAMIC BREAKDOWN CHARACTERISTICS Although the static breakdown characteristics are valuable to define the performances of a semiconductor device, they do not provide all the informations needed to safely use a power transistor. Leaving aside the switching parameters, the dynamic behaviour of a power transistor is bounded by three main parameters: FBSOA: Forward Bias Safe Operating Area RBSOA: Reverse Bias Safe Operating Area BVsus: Breakdown Sustaining The designer of a given application must carefully check that, under no circumstances, the transistor will operate out of the curves provided in the data sheet: severe damage, or +Vbb FORWARD BIAS EMITTER CARRIERS INJECTED IN B/E BASE ÎÎÎÎ ÎÎÎÎÎÎÎ ÎÎÎÎ ÎÎÎÎÎÎÎ ÎÎÎÎ ÎÎÎÎ COLLECTOR rbb’ COLLECTOR CURRENT CARRIERS FLOW HOT SPOT MAY BE GENERATED HERE Figure 8. Turn on Mechanism for a Bipolar Power Transistor http://onsemi.com 6 AN1628/D The data provided for the FBSOA curve are based on a maximum junction temperature of +150°C and is bounded by four points as shown in Figure 9: BVCEO: maximum allowable voltage Icm: maximum allowable collector current Second breakdown Pm: maximum power dissipation Pw: pulse width Since this parameter is temperature dependant, for a constant power, the FBSOA depends upon the pulse width. When the pulse is very short, below 1s, the curves can be extended above the BVCEO limit for transistors specifically designed for switching applications. The typical FBSOA test circuit is given in Figure 10 Cares must be observed to avoid overheating the silicon during the test. Of particular importance are the short pulses which have a high energy content since such pulses can generate a local hot spot within the silicon chip, even though the case temperature stays at ambient temperature. Since the FBSOA test can damage the transistor, one must not re−use devices exposed to a characterization procedure. On the other hand, using conventional power supplies to bias the device under test is not recommended since they do not have the capability to accurately control either the pulse width or the current amplitude. It is preferable to use commercial equipment designed for that purpose, or to built a specific test jig, with appropriate dynamic performances, to evaluate a limited number of devices. I Pw = 10s Pw = 1ms Imax PULSED Imax DC Pw = DC Pmax 2nd BV V BVCEO Figure 9. Typical FBSOA Curve DUT VCB SENSE Ie Vcb CONTROL Figure 10. Typical Forward Bias Safe Operating Area Test Circuit RBSOA The RBSOA curves are used to define the maximum current/voltage a transistor can sustain under Base−Emitter reverse bias. As depicted by the typical test circuit given in Figure 11, this parameter is essentially an analysis of the breakdown, under dynamic condition, when the device is loaded by an inductor. http://onsemi.com 7 AN1628/D +Vcc Lc Ic I Imax IB1 Vbeoff = −5V Vce Vbeoff CLAMP BVCEO BVCBO V Figure 11. Typical Reverse Bias Safe Operating Area Test Circuit time than the collector current fall time (all of the 1/2L*I2 energy must be dissipated) under avalanche condition. Extreme care must be observed when performing such a test since the transistor is very rapidly destroyed if the current is not carefully limited by the external circuit. As a standard rule, the test is carried out by slowly increasing the voltage clamp until the transistor goes into avalanche, the current being limited by the pulse width applied during the charging time of the inductor. The RBSOA test is sometime named sustaining breakdown (BVCEsus) when it is performed unclamped, or inductive switching when the collector voltage is clamped below BVCES. The test is normally performed with a clamp connected across emitter−collector to limit the voltage within the BVCES/BVCEX of the device under test. In this case, the transistor is not avalanched and most of the energy stored into the inductor is dissipated in the clamp circuit. As a matter of fact, when a voltage clamp is used, the transistor sustains the current−voltage stress only during its collector current fall time. The same parameter can be evaluated without providing a voltage clamp: in this case, the transistor will be forced in the avalanche region (assuming L*(dI/dt) > BVCES) and all the energy accumulated in the inductor will be dissipated into the device under test. This condition forces much more stress into the silicon as the current will flow during a longer http://onsemi.com 8 AN1628/D HOT SPOT SILICON CRACK Figure 12. (a) Microphotography of a Power Transistor Damaged by a RBSOA TOP SIDE HOT SPOT WAFER BOTTOM SIDE Figure 12.(b) Cross Section of the Damaged Chip (given Fig. 12a) Showing how the Silicon Melts. POWER DEVICE BEHAVIOUR The reverse bias characteristics of a power transistor are split into two families: the leakage current and the breakdown. Generally speaking, the leakage current tests do not generate high stress into the chip, unless the transistor under test has a voltage capability well below the voltage bias. As a matter of fact, since the bias voltage is defined prior being forced across the junction, there is no snap back in the characteristic and the test circuit remains stable. On top of that, leakage currents are in the micro ampere range, for modern silicon power transistors, and the power generated during the test cannot damage the chip. As an example, let us assume an ICEO test performed at 400 V, the maximum current being 100 A, yielding 40mW in the junction. This value is well within the power handling capability of the transistor and cannot damage the device. However, although they provide reliable informations, the leakage tests characterize neither the breakdown voltage nor the avalanche mode of a bipolar transistor. These behaviour are analyzed by means of the safe operating area tests as discussed above. When the transistor is tested in BVCEO, the current/voltage characteristic exhibit a snap back effect when the Collector/Emitter voltage jumps from the breakdown to the test point point defined at a higher current. This phenomenon, illustrated in Figure 13, is prone to heavy uncontrolled oscillation due to the negative impedance which is developed from point A to point B. http://onsemi.com 9 AN1628/D Ic PROGRAMMED TEST CURRENT 1mA 100A B Oscillations yield uncontrolled operating biases A Vce BREAKDOWN AT Ic = 1µ A Figure 13. Typical BVCEO Characteristic Figure 14 Under no condition can the transistor be exposed to a voltage−current above the maximum rating provided in the data sheets. Such behaviour can be observed with a curve tracer by limiting the power dissipated into the transistor. The power transistors are characterized, among others parameters, by a set of breakdown curves as depicted in I V BVCES BVCEO V 1k 220 BVCEO BVCER BVCES BVCBO 100 10 Rbe Bv = f(Rbe) Figure 14. Typical Breakdown Curves (BVEBO not shown) variable. The problem can be cured by using anti oscillation networks right across the transistor electrodes (see Figure 15), but the device can be damaged or destroyed if the current rises above the maximum value allowable for a given chip. Because the current source used to perform the breakdown test has a high voltage compliance (in fact, the open load voltage must be higher than the maximum breakdown voltage one expects to characterize), it is not easy to make it oscillation free when the load is very http://onsemi.com 10 AN1628/D L T POWER SENSE POWER R SENSE SENSE SENSE POWER POWER The transformer T1 provides an AC feedback to the current source and is adjusted for a given set of test conditions. Typically, the primary inductance is 100 H with Np/ Ns=1. The values of R and L are adjusted for a given transistor and test conditions. Typical values are R=470 and L=100 H. All the network must be located as close as possible to the transistor under test. Kelvin contacts are mandatory. Shielding cables are recommended. Figure 15. Typical Anti Oscillation Networks Since one cannot define accurately the parasitic inductances and stray capacitances existing in the test circuit, it is preferable not to use a capacitive feedback across the transistor. The inductor L and the transformer T can be built around a toroid using a high permeability ferrite material. A typical application is given in Figure 16 It is also possible to use a capacitor, connected across Collector−Base of the transistor under test, to generate a local AC feedback, although this is not recommended. Keeping in mind that the current source associated with the transistor is a high gain transconductance amplifier, any LC network connected across it generates high frequency oscillations that will make the BVCEO test impossible. POWER SENSE POWER Toroid: 16x20 Ferrite: N27, B50, A4A The transformer may require a larger core to accommodate the primary and secondary turns. Using a high permeability material limits the number of turns needed to get the expected inductance. Figure 16. Typical Inductor Design becomes non negligible and are charged up to BVCEO when the current source is connected to the device. Since this source has a finite settling time (i.e., the current does not stabilize to the programmed value in less than a few tens of microseconds!), the current coming from the discharge of the stray capacitance cannot be controlled and may damage the chip within a few micro second. This is particularly true for the BVCEO test as depicted in Figure 17: during the negative going slope, the Collector/Emitter voltage collapses, discharging the One must point out that the design of such an anti−oscillation network very much depends on the test system used to measure the breakdown. Consequently, it’s not easy to modelize an accurate circuit and several attempts are usually necessary to tune the network for a given environment. Another issue with the BVCEO test is the discharge of stray capacitances into the transistor. When the tests are carried out with an automatic instrument, the cables can be several meter in length and the stray capacitance http://onsemi.com 11 AN1628/D stray capacitance by the delta V from A to B, yielding a peak current above the programmed value. Im POWER Ic SENSE Cs1 Cs2 B A V SENSE V POWER Standard coaxial cables have stray capacitance of 100 pF/meter between central wire and shield. Figure 17. Stray Capacitances Effect Assuming a V of 60 V and a total series resistance of 10 , the peak current flowing from the stray capacitances Cs1/Cs2 as depicted in Figure 16, is 6 A. On the other hand, most of the automatic test handler are connected to the tester by coaxial cables up to five meters long. In this case, the total stray capacitance (combining Power and Sense) is 1000 pF, yielding a time constant of: Pw = 5*Cs*R Pw = 5*1000*10−12*10 Pw = 50ns During this amount of time, the transistor must dissipate the energy accumulated in Cs1 & Cs2: Ej = Cs*V2*0.5 Ej = 1000*10−12*60*60*0.5 Ej = 1.8J This value cannot damage the transistor but, if the circuit oscillates, the V can be much higher and, for a high voltage power transistor, the energy can rise to a value above the maximum capability of the junction. As an example, let us assume V = 500 V, then Ej = 125 J. Of course, a power transistor can easily handle such amount of energy, but the same transistor can be rapidly damaged when the energy is generated during the transient phase of the breakdown mechanism. As a matter of fact, during uncontrolled oscillation, most of the energy developed in the circuit will be dissipated in a very small area of the die, yielding a very high current density which, in turn, generates a hot spot in the silicon. Ultimately, the silicon may melt and the device is destroyed. This is a consequence of the multiplication factor M as discussed above. Consequently, it’s recommended to reserve the breakdown test for engineering and characterization purpose, using the leakage current test to perform, incoming inspection. However, the breakdown parameters are fully tested, under several biases and case temperatures, during the development of a new power transistor. Regular samplings also provide statistics, at production level, to make sure the devices are fully within the guaranteed limits published in the data sheet. PARAMETERS DEFINITIONS The off condition parameters are split into two families: the leakage currents and the breakdown voltages. Both are characterized during the development of a transistor and tested during the final test process. The voltage spans from a low 6.00 V for the Bvebo, to a high 1800 V for the Bvcbo, the leakage currents ranging from nano ampere to one milli ampere at room temperature. http://onsemi.com 12 AN1628/D BREAKDOWN VOLTAGES Im BVCEO: reverse Collector to Emitter voltage, with the Base open, under a given collector current bias. Vce Im: Vce: programmed measured Im BVCBO: reverse Collector to Base voltage, with the Emitter open, under a given collector current bias. Im: Vcb: Vbe programmed measured Im BVCER: reverse Collector to Emitter voltage, the Base connected to the Emitter with a low ohm resistor, under a given Collector current bias. Im: Vce: Vce programmed measured Resistor R must be located as close as possible of the device under test. http://onsemi.com 13 R AN1628/D Im BVCES: reverse Collector to Emitter voltage, the Base shorted to the Emitter, under a given Collector current bias. Im: Vce: Vce programmed measured Im BVCEX: reverse Collector to Emitter voltage, with a reverse Base to Emitter bias, under a given Collector current bias. Im: Vce: Vce programmed measured Vbeoff Im BVCEY: reverse Collector to Emitter voltage, with a forward Base to Emitter bias, under a given Collector current bias. Im: Vce: Vce programmed measured Vbeon Im BVEBO: reverse Emitter to Base voltage, with the Collector open, under a given Emitter current bias. Im: Vce: programmed measured http://onsemi.com 14 Veb AN1628/D LEAKAGE CURRENTS Vce ICEO: Collector to Emitter current under reverse Collector to Emitter voltage, Base open. Vce: Iceo: Iceo programmed measured Vcb ICBO: Collector to Base current under reverse Collector to Base voltage, Emitter open. Vcb: Icbo: Icbo programmed measured Vce Rbe ICER: Collector to Emitter current under reverse Collector to Emitter voltage, Base connected to Emitter by a low ohm resistor. Icer Vce: programmed Icer : measured. http://onsemi.com 15 AN1628/D Vce ICES: Collector to Emitter current under reverse Collector to Emitter voltage, Base shorted to Emitter. Vce: Ices: programmed measured Ices ICEX: Collector to Emitter current under reverse Collector to Emitter voltage, with a reverse Base to Emitter bias. Vce: Icex: programmed measured Icex ICEY: Collector to Emitter current under reverse Collector to Emitter voltage, with a forward Base to Emitter bias. Vce: Icey: Vce Vbb programmed measured Vce Vbb Icey Vbb IEBO: Emitter to Base current under reverse Emitter to Base voltage, Collector open. Iebo Vbb: programmed Iebo: measured http://onsemi.com 16 AN1628/D Breakdown Electrical Fields in Semiconductors ANNEXES Symbols E: E1 : Eg : q: n: np: L: M: n : p : Material BIBLIOGRAPHY Jasprit SINGH: “Semiconductor Devices, an introduction’’, McGRAW−HILL International Electrical field minimum energy for ionizing collision Band Gap energy electron charge electron density excess electron density (p region) mean free path avalanche multiplication factor ionization coefficient for electrons ionization coefficient for holes Richard S. MULLER & Theodore I. KAMINS: “Device Electronics for Integrated Circuits’’, second edition, Wiley International S. M. SZE: “Physics of Semiconductor Devices’’, second edition, Wiley International P. LETURCQ & G. REY: “Physique des composants actifs à semicondcuteurs”, Dunod Université Breakdown Electrical Field (V/cm) Bandgap (eV) GaAs 1.43 4*105 Ge 0.664 105 InP 1.34 − Si 1.10 3*105 In0.53 Ga0.47 As 0.80 2*105 C 5.50 107 SiC 2.9 2.5*106 SiO2 9 107 Si3N4 5 107 http://onsemi.com 17 AN1628/D ON Semiconductor and are registered trademarks of Semiconductor Components Industries, LLC (SCILLC). SCILLC reserves the right to make changes without further notice to any products herein. SCILLC makes no warranty, representation or guarantee regarding the suitability of its products for any particular purpose, nor does SCILLC assume any liability arising out of the application or use of any product or circuit, and specifically disclaims any and all liability, including without limitation special, consequential or incidental damages. “Typical” parameters which may be provided in SCILLC data sheets and/or specifications can and do vary in different applications and actual performance may vary over time. All operating parameters, including “Typicals” must be validated for each customer application by customer’s technical experts. SCILLC does not convey any license under its patent rights nor the rights of others. SCILLC products are not designed, intended, or authorized for use as components in systems intended for surgical implant into the body, or other applications intended to support or sustain life, or for any other application in which the failure of the SCILLC product could create a situation where personal injury or death may occur. Should Buyer purchase or use SCILLC products for any such unintended or unauthorized application, Buyer shall indemnify and hold SCILLC and its officers, employees, subsidiaries, affiliates, and distributors harmless against all claims, costs, damages, and expenses, and reasonable attorney fees arising out of, directly or indirectly, any claim of personal injury or death associated with such unintended or unauthorized use, even if such claim alleges that SCILLC was negligent regarding the design or manufacture of the part. SCILLC is an Equal Opportunity/Affirmative Action Employer. PUBLICATION ORDERING INFORMATION Literature Fulfillment: Literature Distribution Center for ON Semiconductor P.O. Box 5163, Denver, Colorado 80217 USA Phone: 303−675−2175 or 800−344−3860 Toll Free USA/Canada Fax: 303−675−2176 or 800−344−3867 Toll Free USA/Canada Email: [email protected] JAPAN: ON Semiconductor, Japan Customer Focus Center 2−9−1 Kamimeguro, Meguro−ku, Tokyo, Japan 153−0051 Phone: 81−3−5773−3850 ON Semiconductor Website: http://onsemi.com For additional information, please contact your local Sales Representative. N. American Technical Support: 800−282−9855 Toll Free USA/Canada http://onsemi.com 18 AN1628/D