With the combination of an easily driven MOS gate and low conduction loss, the IGBT is the device of choice for high current and high voltage applications. Now with the latest generation of PT IGBTs, the tradeoff between switching and conduction losses is balanced so that IGBTs encroach upon the high frequency, high efficiency domain of power MOSFETs. In fact, the industry trend is for IGBTs to replace power MOSFETs in switch mode power supply (SMPS) applications if the voltage is above about 300 Volts.

This trend is made possible by a significant improvement in switching speed with the latest generation of PT IGBTs while retaining the low conduction loss that is characteristic of IGBTs. In most cases, circuit designers who use these latest technology IGBTs can significantly reduce costs with little, if any sacrifice in efficiency. This paper gives an overview of PT IGBT technology, compares features and benefits with power MOSFETs, and shows performance improvements in various high frequency, high voltage SMPS applications.

Punch through IGBT structure
A PT IGBT is basically an N-channel power MOSFET constructed on a p-type substrate [1], as illustrated by the generic IGBT cross section in Figure 1.

Figure 1 PT IGBT Cross Section

Equivalent circuit
In an IGBT, the reverse current through the MOSFET intrinsic body diode is blocked. This leads to a simple equivalent circuit model for an IGBT, which is simply a diode in series with the drain of an N-channel power MOSFET, as shown in Figure 2. The blocking capability of this series diode is typically only about 15 to 30 Volts. A separate diode must be connected anti-parallel to the IGBT if reverse current flow or protection from reverse voltage is required as in a bridge circuit, which will be discussed.

Figure 2 Simple IGBT Equivalent Circuit

Looking at Figure 2, it would seem that the on-state voltage across the IGBT would be one diode voltage drop higher than for the N-channel power MOSFET by itself. It is true in fact that the on-state voltage across an IGBT is always at least one diode voltage drop. However, compared to a conventional high voltage power MOSFET of the same die size and similar voltage rating, an IGBT has significantly lower on-state voltage.

The reason for this is that a MOSFET is a majority carrier only device, meaning that only electrons flow; current is unipolar. In an IGBT, the flow of electrons through the channel and drift region draws positive carriers, called holes, into the drift region toward the emitter. Therefore current flow in an IGBT is composed of both electrons and holes; current is bipolar.

This injection of holes (minority carriers) significantly reduces the effective resistance to current flow in the drift region. Stated otherwise, hole injection increases the conductivity, or the conductivity is modulated. The resulting reduction in on-state voltage is the main advantage of IGBTs over conventional high voltage power MOSFETs.

IGBTs trade lower on-state voltage for slower switching speed at turn-off. The reason for this is during turn-off the electron flow can be stopped rather abruptly by reducing the gate-emitter voltage below the gate threshold voltage, just as with a power MOSFET. However, holes are left in the drift and body regions, and there is no terminal connection to facilitate removing them. The only ways to remove these minority carriers are by sweep-out, which depends on the voltage across the device, and by internal recombination. A tail current exists until recombination is complete, and this tail current has historically been the major drawback of IGBTs. Because of the high injection efficiency of the p+ layer, an n+ buffer layer (shown in Figure 1) is used to control the transconductance (gain) of the device by limiting the number of holes that are injected into the drift region in the first place. Since minority carrier lifetime in the buffer layer is much lower than in the drift region, the buffer layer also absorbs trapped holes during turn-off.

In addition to the n+ buffer layer, the tail current in a PT IGBT is controlled by limiting the amount of time that a minority carrier dwells before being recombined. This is called minority carrier lifetime control. An electron irradiation process during manufacturing creates recombination sites throughout the silicon, which greatly reduces minority carrier lifetime and hence the tail current. Holes are quickly recombined, even with no voltage across the device as in soft switching. The tradeoffs with lifetime control are an increase in on-state voltage, called V_CE(on), and slightly higher leakage current at high temperature. Switching speed
Switching speed
Turn-on characteristics of an IGBT are very similar to a power MOSFET. Turn-off differs though because of the tail current. Thus the turn-off switching energy in a hard switched clamped inductive circuit gives an indication of the switching speed and tail current characteristic of an IGBT. Figure 3 depicts the tradeoff between turn-off switching energy E_off and V_CE(on).

Figure 3 Turn-Off Energy vs. V_CE(on)

Within any technology, in order to reduce conduction loss, switching energy must increase, and vice versa. Only technology improvements can yield both lower conduction and lower switching losses. Figure 3 shows two technology curves, the dashed curve depicting the performance of previous generation IGBTs, and the solid curve depicting the performance and tradeoff point of the latest generation of PT IGBTs, such as the Power MOS 7 �'IGBTs from Microsemi. By utilizing the latest PT IGBT technology, switching energy has been reduced by 30% to 50% without significantly increasing V_CE(on), resulting in a high performance IGBT optimized for high voltage SMPS applications.

Usable frequency versus current
A useful way to compare the performance of different devices is to graph usable frequency versus current. What makes this so valuable is that it incorporates not only conduction loss, but also switching loss as well as thermal resistance. It also makes it easy to compare the performance of different types of devices, like IGBTs and power MOSFETs.

Beginning with the fundamental relationships:

It is f_max1, a percent of switching time limitation that limits the frequency at low current; f_max2, a thermal limitation, limits the frequency otherwise. The f_max1 pulse width limitation rule favors small devices with their small capacitances and shorter delay times. In this case, the estimated time spent switching (the sum of delay, rise, and fall times) is no more than 5% of the total switching period. A different pulse width limit rule may be used, but it does make sense to limit the time spent switching to some percentage of the total switching period so the die has time to cool between switching transients. Above a certain current, the frequency is limited by heat dissipation due to switching and conduction losses (f_max2) rather than a pulse width limitation (f_max1).

Usable frequency
Lower losses as well as lower thermal resistance R_θJC result in higher maximum frequency. In general, a device that is thermally capable of the highest switching frequency is the most efficient device. Figure 4 shows usable frequency versus current curves for three devices: a PT IGBT and two power MOSFETs. All three are the latest technology devices from Advanced Power Technology.

Figure 4 Usable Frequency vs. Current

The APT30GP60B is a 600 Volt Power MOS 7 �'IGBT rated at I_C2 = 49 Amps in a TO-247 package. The APT6038BLL and APT6010B2LL are 600 Volt Power MOS 7 �'MOSFETs rated at ID = 17 and 54 Amps respectively. The APT6038BLL is in a TO-247 package, and the APT6010B2LL is in a T-MAX package (TO-247 without a hole).

The test conditions for Figure 4 are: hard inductive switching, 400 V, T_J = 125 °C, T_C = 75 °C, 50% duty cycle, and 5 Ω �'total gate resistance (the equivalent gate resistance for each part is very small because of the metal gate and is therefore ignored). For each device, a 15 Amp, 600 Volt ultra-fast recovery diode was used as the clamp (freewheeling) diode. The test circuit is representative of boost, buck, or other clamped inductive switching circuit topologies.

The APT30GP60B and the APT6038BLL have the same die size, and the APT6010B2LL is approximately three times larger in area. Device cost depends on die area, so the device with the smallest die size that meets an application's requirements is generally the least expensive choice.

Suppose we want to switch 8 Amps at 200 kHz. By looking at Figure 4, it is clear that the APT6038BLL is the best part since it can switch at a much higher frequency than the other parts. Note that this is because of the f_max1 pulse width limitation, which is what limits the usable frequency of the APT6010B2LL and the APT30GP60B below about 27 and 15 Amps respectively.

Now suppose we want to switch 20 Amps at 200 kHz. Either the APT30GP60B or the APT6010B2LL will work, but the APT30GP60B IGBT is about one third the cost of the APT6010B2LL because of its smaller die size. The APT6038BLL is completely out of the running. Above 37 Amps the IGBT wins, even though its die size is smaller; its junction temperature would be lower than that of the MOSFET at a given frequency. This goes against conventional wisdom, which says that a MOSFET is always more efficient than an IGBT, and that higher efficiency means higher cost.

The curves in Figure 4 warrant a few more comments. First, note that the I_D rating of the APT6038BLL is 17 Amps, but in this application, this part can't handle more than about 10 Amps. Under different conditions though, such as lower duty cycle, it could handle its "rated current". Rated currents do not necessarily tell you how much current a device can handle in an application because they are based on continuous conduction loss (no switching loss) with the case held at a certain temperature. Basically they tell you the relative size and conduction loss of the device. Comparisons
Second, a convenient comparison to note is that the ID rating of the APT6010B2LL (continuous conduction with the case at 25 °C) is similar to the I_C2 rating of the APT30GP60B (continuous conduction with the case at 110 °C), 54 and 49 Amps respectively. These two current ratings are similar, and the performance of each device is also similar. Both are capable of 200 kHz operation at about half their current ratings. So matching up the MOSFET ID rating to the IGBT I_C2 rating is a quick way to make initial comparisons between power MOSFETs and Power MOS 7 �'IGBTs.

Third, an IGBT has higher current density, which equates to lower on-state voltage and enables using a smaller die at the same power level as a high voltage power MOSFET. Due to dramatically increased onresistance with increasing breakdown voltage ratings, power MOSFETs rated at or above about 300 Volts have lower current densities than IGBTs. This is why a PT IGBT with a 600 Volt rating can replace a MOSFET rated at around 400 Volts or higher. The smaller IGBT die size results in a higher thermal resistance than a power MOSFET, but the junction temperature is not higher due to lower losses. Remember, thermal resistance is accounted for in the usable frequency versus current curves.

Finally, different devices are best under different operating conditions. At high frequency and relatively low current, a MOSFET is usually the best choice (or you could use a smaller size PT IGBT). At high current, an IGBT is the best choice because conduction loss increases very modestly with increasing current, whereas the conduction loss of a power MOSFET is proportional to the current squared. At most frequency or current ranges, more than one device type might work well, so there is often more than one right answer. However, the latest generation PT IGBT will usually be the least expensive option. This very important point is the reason behind an emerging trend to replace power MOSFETs with IGBTs in high voltage, high frequency power supplies.

References
[1] J. Dodge, J. Hess; "IGBT Tutorial", Application Note APT0201, Advanced Power Technology [2] J. Dodge; "Parallel Connection of Power Electronic Devices", Application Note APT0202, Advanced Power Technology [3] "Application Characterization of IGBTs", Application Note INT990, International Rectifier

About the author
Jonathan Dodge, P.E. is a Senior Applications Engineer for Microsemi Corp's. Advanced Power Technology group. For more information: www.microsemi.com