Since solid-state transistors have replaced vacuum tubes, the semiconductor industry has experienced amazing developments that have changed our world. Without many of these improvements, we won’t be able to work remotely and stay connected even during enforced isolation and we won’t be able to enjoy all the other marvels that technology has to offer. One example: the extraordinary gain in processing power that was made possible by the constant efforts of engineers to squeeze more and more transistors per unit area onto silicon wafers. Defined as Moore’s Law, the observation that the density of a transistor can be doubled every 18 months or so, has guided the development of microprocessor generations in the semiconductor industry for more than 50 years. Now, we have reached atomic and physical limits that require new techniques, such as vertical stacking of layers. At the same time, we are also in the midst of another revolution with the development of wide bandgap semiconductors such as silicon carbide (SiC) and gallium nitride (GaN). The new materials have unique physical properties that allow for improved efficiency, increased energy density and safer performance under the most extreme thermal conditions. STMicroelectronics is currently mass-producing STPOWER SiC MOSFETs, helping to drive the adoption of electric vehicles (EVs) and sparking a massive electrification era. It is also envisioned that this will eventually lead to self-driving for sustainable mobility. Another revolution involving high-voltage (ie, above 200 V) silicon power transistors occurred at the turn of the century, when superconducting MOSFETs appeared. Until the end of the 1990s, designers had to accept the “axiom” that for a planar transistor, the merit number (defined as the resistance multiplied by the area of the chip) is proportional to the breakdown voltage (BV) raised to 2.5 . This axiom indicates that the only solution to reach lower values of resistance at a given voltage is to increase the area of the mold. This made it more difficult to use devices with packages with a small outline. Superjunction technology came to the rescue of high voltage MOSFET circuits by bringing the above relationship close to linear. ST named the technology MDmesh and made it part of the STPOWER sub-brand. Principle of Superjunction Transistors The mechanism of action of a superjunction transistor takes advantage of one of Maxwell’s equations, simplified to a one-dimensional case – to move the vertical axis, y. It states that the slope of the electric field along this axis is equal to the charge density r divided by the permittivity e. In symbols, dE/dy = r/e. The other equation relates the voltage V to the electric field component E along y; That is, E = –dV/dy. Stated differently, the voltage V is the integral of E, or in geometric terms, the area under the E curve as a function of y. We can see how it works by comparing the vertical structure of a standard planar MOSFET against its analog of similar size. The superjunction is basically an extension of the p-body of the base transistor inside the vertical drain by realizing a p-type column. In a planar structure (see Fig. 1, left) starting from the surface along the y axis we face the body p, therefore, the slope is positive until we reach the point A. From A to B, we have the drainage. With the opposite polarity, therefore, the tendency to negative is reversed. From B to the substrate, the polarity becomes more negative (n–), so the slope increases. The green area in the graph represents the voltage that can continue in the off state. In the superjunction diagram shown on the right, the addition of the p-type region column changes the electric field distribution. In fact, from C to A, the electric field distribution remains constant (the body and shaft have the same polarity), then the slope is reversed as in the planar structure due to the discharge and the substrate. As a result, the area under the electric field is larger, and therefore the voltage V2 continues. Here, the column has done its magic. Now, at a given voltage, we can reduce the drain resistance and we can reduce the resistance. Figure 1: Planar (left) and superconducting MDmesh (right) The evolution of MOSFETs technology Since their first appearance, MDmesh transistors have been relentlessly improved and refined, and a wide range of power conversion applications continue to benefit from their use. Process techniques for creating vertical shafts have been greatly improved for better manufacturing productivity and machine durability. Depending on the target circuit topology and application, different customized product series are now available. This flexibility and technological flexibility allow system designers to choose from a variety of options. The general purpose M2 series has the best cost/performance in the 400- to 650-V range, there are application specific variants that separately handle PFC, soft-switching LLC and bridge topologies, with a voltage rating extending to 1700V Furthermore, lifetime kill techniques are used Such as infusing platinum ions to improve the performance of the integrated body diode to reduce the trr reverse recovery time, in addition to charging the reverse recovery Qrr as well as improving dV/dt (DM). Series). These features are ideal in bridge circuits and high-power phase-shift circuits. The fast diode version can compete with IGBTs in lower power motor drives, eliminating the need for a common diode. In terms of efficiency, a typical example is represented by a 150W inverter for a refrigerator compressor as shown in Fig. 2. Fig. 2: Efficiency curves of a MDmesh MOSFET fast diode compressor inverter versus IGBT in a DPAK package. Test condition: 0.23 Nm (load) 220 V / 50 Hz (input voltage) It is not surprising that MDmesh transistors are produced everywhere in the billions! In Figure 3, by comparing the features achieved with the latest improved M6 series of resonant transducers, we see how diligent ST designers were in terms of improving the first version of the M2. Figure 3: M2 to M6 – Improvement of gate charge, threshold voltage, and output capacitance in the graphs In Figure 3, from left to right we see that lower gate charge, increased threshold voltage, and linear output capacitance vs. voltage translate into a higher switching frequency, And reduced switching losses, higher efficiency at lighter loads. Basic superjunction technology combined with more advanced process steps has resulted in a high-performance high-voltage MOSFET with a particular focus on key switching parameters such as dI/dt and dV/dt, as the safe operation diagram demonstrates in Figure 4. Thanks to these improvements, the series fit DM6 MDmesh works well with solar inverters, charging stations, and On-Board Chargers (OBC) to name a few. Figure 4: Safe Operating Areas dI/dt vs. dV/dt Application areas ST MDmesh transistors are used in many applications and this allows us to show their advantages in a small but representative selection. One of the larger applications is smartphone adapters. Figure 5 shows the 120W version. Figure 5: MDmesh in a Smartphone Adapter Figure 6 shows how the “tailored” M5 series can improve efficiency in a 1.5 kW PFC at higher power with respect to the “core” M2 series. The two MOSFETs used display similar impedance (37 and 39 MΩ on the resistances for M5 and M2, respectively) and voltage blocking capability (650 V). Figure 6: How the M5 series (in blue) can improve PFC efficiency at higher power Another interesting example is shown in Figure 7: a 3 kW half-bridge LLC circuit for OBC cars comparing the latest DM6 (STWA75N65DM6) against the best competition at Vin = 380 420 V, Vout = 48 V, switching frequency f = 250 Hz to 140 kHz. Figure 7: 3 kW full-bridge LLC – Out-of-Power and Delta Efficiency vs. Pout By Filippo Di Giovanni, Director of Strategic Marketing, Innovation and Principal Programs at ST Microelectronics Please visit the e-book for the full article.
0 Comments