PN Junction: Starting from the PN Junction

 

The PN junction is the foundation of semiconductors, and doping is the soul of semiconductors. First, let's clarify a few points:

 

1. P-type and N-type semiconductors: Intrinsic semiconductors are doped with trivalent elements. According to the principle of chemical bond stability learned in high school, there will be "holes" that are easy to conduct electricity. Therefore, here, holes are "majority carriers"; similarly, doping with pentavalent elements results in electrons being the "majority carriers", and the doping type is N (negative) type.

 

2. Carriers: Conductive media, divided into majority and minority carriers, the concept is very important and will be referenced later.

 

3. "Holes" carry positive charge, while electrons carry negative charge, but the doped semiconductor itself is electrically neutral.

 

4. P+ and N+ indicate heavy doping; P- and N- indicate light doping.

 

The principle of the PN junction is shown in the figure below, where the diffusion of holes and electrons forms a depletion layer, and the direction of the electric field in the depletion layer is as shown in the figure:

(1) Diode

 

Forward bias of the PN junction: Applying forward voltage to the PN junction, as shown in the figure below:

At this time, the majority carriers "holes" in the P region move towards the N region under the action of the electric field, while the majority carriers electrons in the N region move in the opposite direction, causing the depletion layer to narrow to the point of disappearance. Forward conduction is okay, and it can also be understood as the external electric field overcoming the electric field within the depletion layer to achieve conduction. This voltage is generally 0.7V or 0.3V. This is the principle of forward conduction of the diode.

 

Reverse bias of the PN junction: Applying reverse voltage to the PN junction, as shown in the figure below:

During reverse bias, the majority carriers move under the action of the electric field, causing the PN junction to widen, and current cannot pass through, resulting in reverse cutoff; this is the principle of reverse cutoff of the diode. However, at this time, minority carriers move under the influence of the internal and external electric fields, and the direction of the electric field in the depletion layer makes it easier for minority carriers to pass through the PN junction, forming leakage current.

 

An important conclusion can be drawn, highlighting: During reverse bias, majority carriers are cut off, while minority carriers can pass through very easily, even more easily than majority carriers passing through the PN junction during forward bias.

 

(2) Transistor

 

As mentioned above, during reverse bias of the PN junction, minority carriers can easily pass through, forming current. Under normal circumstances, the number of minority carriers is very small, and the reverse current can be ignored.

 

Now we control this reverse current by injecting minority carriers into the N region. How to inject? By adding another P region below the N region and making the newly added PN junction forward biased, as shown below:

In the figure above, the emitter junction is forward biased, and a large number of holes enter the base region. In the base region, they still act as minority carriers. At this time, as mentioned earlier, these injected minority carriers can easily pass through the reverse-biased PN junction—the collector junction—to reach the collector electrode, forming the collector current Ic.

 

The condition for the transistor to amplify is "emitter junction forward biased, collector junction reverse biased," which is very easy to understand, as shown in the previous characteristic curve of the transistor.

This involves the issue of the saturation region. When the transistor operates in the saturation region, Vce is very small. Some say the condition for the saturation region is that the emitter junction is forward biased and the collector junction is also forward biased, which can easily lead to misunderstanding; it is fine for the emitter junction to be forward biased, but the collector junction has not reached forward bias conduction. If the collector junction is forward biased, it would be no different from having two diodes together.

 

The forward bias voltage at the collector junction impedes the drift of minority carriers in the base region towards the collector electrode. The stronger the forward bias, the more challenging it becomes for minority carriers to move towards the collector, resulting in a smaller Ic. Consequently, the Ic in the saturation state is less than the βIb in the amplification state. At this point, the transistor demonstrates a very low junction resistance, referred to as saturation conduction.

 

(3) MOSFET

 

The structure principle of the MOSFET: Taking N-MOS as an example, a: P-type semiconductor as the substrate; b: two N-type regions diffused on top; c: covered with SiO2 insulating layer; two holes are etched on the N region, and then by metallizing, three electrodes are formed in the insulating layer and the two holes: G (gate), D (drain), S (source).

Working principle: Generally, the substrate and source are shorted together. When Vds is applied with positive voltage and Vgs=0, the PN junction is reverse biased, and there is no current. When Vgs is applied with positive voltage, negative charges are induced above the P substrate, which is opposite in polarity to the majority carriers (holes) of the P substrate, forming an inversion layer, connecting the N-type regions of the drain and source to form a conductive channel. When Vgs is relatively small, the negative charges neutralize with the holes, and conduction is still not possible. When Vgs exceeds the threshold for conduction, the induced negative charges connect the N-type regions to form an N-channel, starting conduction. As Vgs continues to increase, the channel expands, resistance decreases, and thus current increases.

To improve device performance, various structures such as VMOS and UMOS have emerged, but the basic principle is the same.

 

(4) IGBT

 

IGBT is a composite device of MOS and BJT. How it is composed will be explained below. From the structural perspective, the IGBT structure is very similar to that of power MOS, with a P+ injection layer added to the back.

The conductive path of the IGBT is as follows:

As shown in the figure above, the P well and the N-drift region form a reverse-biased PN junction, thus producing the JFET effect, as shown in the figure below.

Thus, in the above IGBT structure, the resistance in the direction of electron flow can be represented as shown in the figure below, which is clear when combined with the description above.

To reduce the above resistance and improve the gate area utilization, trench gate IGBTs have become mainstream, with the effect shown in the figure below.

In addition, to enhance the voltage resistance of the IGBT and reduce tail current, considerable effort has been made in the N-drift region and back-end processes (thinning and injection).

 

The efforts in the N-region include the following types:

1. PT: Using high-concentration P+ single crystal silicon as the starting material, a layer of heavily doped N-type buffer layer (N+ buffer layer) is first grown, and then a lightly doped N-type epitaxial layer is deposited as the drift region of the IGBT. After that, a P-base and N+ source are formed on the surface of the N-type epitaxial layer as the cell, and finally, the P-type substrate is thinned as needed.

 

2. NPT: Using lightly doped N-type zone melted single crystal silicon as the starting material, first create the unit cells on the front side of the silicon surface and protect them with a passivation layer, then thin the silicon wafer to an appropriate thickness. Finally, inject boron into the back of the thinned silicon wafer to form a P+ collector.

 

3. FS: Using lightly doped N-type zone melted single crystal silicon as the starting material, first create the unit cells on the front side of the silicon surface and protect them with a passivation layer. After thinning the silicon wafer, first inject phosphorus into the back of the silicon wafer to form an N+ cutoff layer, and finally inject boron to form a P+ collector.

 

What are the differences between transistors, MOSFETs, and IGBTs?Why is it said that IGBT is a device composed of BJT and MOSFET?

 

 

To clarify the relationship between IGBT, BJT, and MOSFET, one must have a general understanding of the internal structures and working principles of these three.

 

BJT

 

Bipolar junction transistor, commonly known as a transistor. The internal structure (taking PNP type BJT as an example) is shown in the figure below.

Bipolar means that there are two types of charge carriers, holes and electrons, participating in conduction within the device. Since BJT is called a bipolar junction transistor, it must have both holes and charge carriers inside. Understanding the movement of these two types of charge carriers is key to understanding the working principle of BJT.

 

Since the hole concentration in the P region of e (emitter) is greater than that in the N region of b (base), hole diffusion will occur, meaning holes will diffuse from the P region to the N region. Similarly, the electron concentration in the P region of e (emitter) is lower than that in the N region of b (base), so electrons will also diffuse from the N region to the P region.

 

This movement will ultimately create an electric field at the emitter junction pointing from the N region to the P region, known as the built-in electric field. This electric field will prevent holes in the P region from continuing to diffuse into the N region. If we apply a positive bias voltage (p positive, n negative) at the emitter junction to weaken the effect of the built-in electric field, holes can continue to diffuse into the N region.

 

Some of the holes that diffuse into the N region will recombine with the majority carriers—electrons in the N region, while another part will drift to the collector under reverse bias (p negative, n positive) conditions, forming the collector current.

 

It is worth noting that the electrons in the N region will not be insufficient after recombining with holes from the P region, because the b electrode (base) will continuously provide electrons to ensure that the above process can continue. Understanding this part is very helpful for understanding the relationship between IGBT and BJT later.

 

MOSFET

 

Metal-Oxide-Semiconductor Field-Effect Transistor, abbreviated as MOSFET. The internal structure (taking N-MOSFET as an example) is shown in the figure below.

                                                                                MOSFET internal structure and symbols.

 

Two N+ regions are created on a P-type semiconductor substrate, one called the source region and the other the drain region. The channel region is the lateral distance between the drain and source. On the surface of the channel region, there is a layer of oxide generated by thermal oxidation as a medium, called the insulating gate. A layer of aluminum is evaporated on the source region, drain region, and insulating gate as the lead electrodes, which are the source (S), drain (D), and gate (G).

 

MOSFET is a voltage-controlled device, and its on/off state is controlled by the gate voltage. Observing the diagram, we find that there are two back-to-back pn junctions between the source S and drain D of the N-MOSFET. When no voltage is applied to the gate-source voltage VGS, regardless of how large or what polarity of voltage is applied between the drain-source voltage VDS, one of the pn junctions will always be in reverse bias, and there is no conductive channel between the drain and source, so the device cannot be turned on.

 

However, if VGS is sufficiently positive, an electric field will be generated in the insulating layer between the gate G and the substrate p, directed from the gate to the substrate. Under the action of this electric field, electrons will accumulate at the surface beneath the gate oxide, forming an N-type thin layer (generally a few nm), connecting the two N+ regions on the left and right, forming a conductive channel, as shown in the yellow area of the figure. When VDS > 0V, the N-MOSFET is turned on, and the device operates.

 

IGBT

 

Structure diagram of IGBT.

                                                                                      IGBT internal structure and symbols.

The yellow area indicates the channel formed when the IGBT is turned on. First, look at the yellow dashed part; does it seem familiar upon closer inspection?

 

This part of the structure and working principle is essentially the same as that of the N-MOSFET. When VGE > 0V and VCE > 0V, a channel will also form on the surface of the IGBT, with electrons starting from the n region, flowing through the channel region, and injecting into the n drift region, which is similar to the drain of the N-MOSFET.

 

The blue dashed part corresponds to the BJT structure, with electrons flowing into the n drift region continuously providing electrons for the n region of the PNP transistor, ensuring the base current of the PNP transistor. We apply a positive bias voltage VCE to make the PNP conduct in the forward direction, allowing the IGBT device to operate normally.

 

This is why it is defined that IGBT is a device composed of BJT and MOSFET.

 

In addition, the red part marked here was not mentioned in the definition because it is actually a npnp parasitic thyristor structure. This structure is undesirable for IGBT because the parasitic thyristor can latch under certain conditions, causing the IGBT to lose gate control capability, making it unable to turn off by itself, which can lead to damage to the IGBT.

 

The relationship between IGBT, BJT, and MOSFET

 

BJT appeared before MOSFET, and MOSFET appeared before IGBT, so we explain the causal relationship among the three from the emergence of the intermediary MOSFET.

 

The emergence of MOSFET can be traced back to the early 1930s. The concept of the field-effect transistor proposed by German scientist Lilienfeld in 1930 attracted the interest of many scientists in the field. Bardeen and Brattain from Bell Labs accidentally invented the bipolar junction transistor (BJT) during an attempt to invent a field-effect tube in 1947.

 

Two years later, Shockley, also from Bell Labs, clarified the working principle of BJT using the minority carrier injection theory and proposed the concept of a practical junction transistor.

 

Developed to the present, MOSFET is mainly used in medium and small power applications such as computer power supplies and household appliances, with advantages such as high gate input impedance, low driving power, strong current cutoff capability, fast switching speed, and low switching loss.

 

As downstream applications develop faster and faster, the current capacity of MOSFET is obviously unable to meet market demand. In order to reduce the on-resistance of the device while retaining the advantages of MOSFET, attempts have been made to reduce the on-resistance by increasing the doping concentration of the MOSFET substrate, but increasing the substrate doping will reduce the device's voltage resistance. This is obviously not an ideal improvement method.

 

However, if a bipolar BJT structure is introduced based on the MOSFET structure, it can not only retain the original advantages of MOSFET but also modulate the conductivity of the n-drift region through the minority carrier injection effect of the BJT structure, thereby effectively reducing the resistivity of the n-drift region and improving the current capacity of the device.

 

After continuous improvements, the IGBT can now cover a voltage range from 600V to 6500V, with applications spanning a series of fields from industrial power supplies, inverters, new energy vehicles, new energy generation to rail transit and the national grid.