An LED torch takes two cells. Put them in the right way and it shines. Take them out, turn them round and put them back, and the torch stays completely dark. The cells are fine; they light the torch again the moment they are turned back. A lamp with a filament would have shone either way. The LED in the torch is a p-n junction, and it lets current through in one direction only.
Lowering the barrier
Connect a battery across a junction with the p side to its positive terminal and the n side to its negative terminal. This is forward bias. The battery's voltage \(V\) across the junction opposes the built-in step \(V_0\). The barrier falls to \(V_0 - V\), and the depletion layer becomes thinner.
Recall the balance of lesson 3. The diffusion current depends very strongly on the height of the barrier, because only carriers with enough energy can climb it. Lower the hill a little, and many more carriers get over. The drift current is unchanged. So the diffusion current now wins, and a net current flows from p to n. It grows very rapidly as \(V\) approaches \(V_0\).
Try \(V = 0.4\), \(0.6\) and \(0.7\,\text{V}\) in the sim. The current is about 0.01 mA, then 0.55 mA, then 4 mA. Below about 0.6 V the current is too small to notice on the graph. Above about 0.7 V it shoots up. The barrier is nearly gone, and a large current of electrons and holes streams across the junction.
A silicon junction is forward biased with \(0.5\,\text{V}\). What happens to the barrier and to the depletion layer?
Raising the barrier
Now turn the battery round: p side to the negative terminal, n side to the positive. This is reverse bias. The applied voltage adds to the built-in step, so the barrier rises to \(V_0 + V\). The positive terminal pulls electrons away from the junction, and the negative terminal pulls holes away. More ions are uncovered on each side, so the depletion layer widens.
Almost no majority carrier can climb the higher barrier, so the diffusion current stops. What remains is the drift of minority carriers, swept across by the field. There are very few of them, so the current is tiny, about a microampere or less. It is called the reverse saturation current. It hardly changes with the reverse voltage, because the field already sweeps across every minority carrier that arrives.
The minority carriers are made by heat, so the reverse current grows quickly with temperature. It roughly doubles for every 10 K rise. Heating also lowers the barrier a little, by about 2 mV for each kelvin. With more carriers about, a smaller barrier is enough to balance diffusion. At a very large reverse voltage the junction finally breaks down, and the reverse current rises sharply. The next chapter takes up this breakdown and the Zener diode that uses it.
The reverse voltage across a silicon junction is raised from \(2\,\text{V}\) to \(4\,\text{V}\), well below breakdown. What happens to the reverse current?
A reverse-biased diode warms from \(300\,\text{K}\) to \(320\,\text{K}\). About how does its reverse current change?
The diode and its curve
A p-n junction with a lead on each side is a junction diode. The lead to the p side is the anode and the lead to the n side is the cathode. Its symbol is a triangle pointing at a bar. The triangle points the way conventional current passes easily, from anode to cathode. The bar marks the cathode, the n side.
The graph of current against voltage in the sim is the diode's characteristic. In forward bias the current stays small until the knee voltage, and then climbs almost straight up. The knee is about 0.7 V for silicon and 0.3 V for germanium, the same as the barrier. In reverse bias the current is only microamperes, so the two halves of the graph need different scales. The forward half is in milliamperes and the reverse half in microamperes.
Beyond the knee a small change of voltage gives a large change of current. The ratio of the two is the dynamic resistance \(r_d\). It is the inverse of the slope of the curve at that point, not \(V/I\). Read two points from the sim: 11.0 mA at 0.75 V and 30.0 mA at 0.80 V. Then \(r_d = 0.05/0.019 = 2.6\,\Omega\). In reverse bias a change of many volts barely changes the current, so \(r_d\) there is millions of ohms.
A silicon diode carries \(5\,\text{mA}\) at \(0.70\,\text{V}\) and \(9\,\text{mA}\) at \(0.72\,\text{V}\). Find its dynamic resistance in this range, in ohms.
Two simple models
A real characteristic is a curve, which is awkward in a circuit. Two simple models stand in for it. The ideal diode is a perfect switch. In forward bias it conducts with no voltage across it, like a closed switch. In reverse bias it passes no current at all, like an open switch.
The 0.7 V model is closer to a real silicon diode. It passes no current until the forward voltage reaches 0.7 V. Once it conducts, it keeps 0.7 V across itself, whatever the current. For germanium the drop is 0.3 V. In reverse bias it is an open switch, like the ideal diode.
A way to solve any diode circuit
Guess the state of each diode, on or off. Replace an "on" diode by its drop and an "off" diode by a gap. Solve the circuit with Ohm's law, \(V = IR\) for each resistor. Then check the guess. An "on" diode must carry current from anode to cathode. An "off" diode must have less than its drop across it in the forward direction. If a check fails, change that guess and solve again.
A \(9\,\text{V}\) battery, a silicon diode and a \(1\,\text{k}\Omega\) resistor are in one loop, with the diode forward biased. Find the current and the voltage across the resistor with the 0.7 V model, and compare the ideal model. What changes when the diode is turned round?
The diode is forward biased, so guess "on". It keeps \(0.7\,\text{V}\) across itself.
The resistor has \(9 - 0.7 = 8.3\,\text{V}\) across it, and \(I = 8.3 / 1000\,\text{A}\).
\(I = 8.3\,\text{mA}\), and the resistor has \(8.3\,\text{V}\). The ideal model gives \(9/1000 = 9.0\,\text{mA}\) and \(9.0\,\text{V}\).
Reversed, the diode is "off": no current flows, and the resistor has no voltage across it. So the whole \(9\,\text{V}\) appears across the diode, in the reverse direction. The two models differ by about 8 % here; they agree better when the battery's voltage is much larger than 0.7 V.
Two silicon diodes, both forward biased, are in series with a \(180\,\Omega\) resistor across a \(5\,\text{V}\) battery. Use the 0.7 V model. Find the current, in mA.
A germanium diode and a silicon diode are connected side by side, both forward biased, in series with a \(1\,\text{k}\Omega\) resistor and a \(6\,\text{V}\) battery. Use drops of \(0.3\) and \(0.7\,\text{V}\). Which diode carries the current?
Now the torch of the opening makes sense. With the cells the right way, the LED is forward biased, and a current of some milliamperes lights it. Turned round, the LED is reverse biased. Only the reverse saturation current flows, a microampere or less, far too little to light it. The cells are not being used up either: they push almost no current round the loop.
In Calcutta in the late 1890s, Jagadish Chandra Bose was studying waves of a few millimetres wavelength. He needed a detector for them. He pressed a fine metal point against a crystal of galena, a lead sulphide mineral. The contact let current through much more easily in one direction than in the other, so it turned the waves into a steady signal. He was granted a United States patent for the detector in 1904, one of the first patents for a semiconductor device.
- Adding the applied voltage to the barrier in forward bias.Forward bias opposes the built-in step: the barrier becomes \(V_0 - V\). In reverse bias it becomes \(V_0 + V\).
- Finding the dynamic resistance as \(V/I\) at one point.\(r_d = \Delta V/\Delta I\) uses a small change between two nearby points on the curve.
- Letting both of two parallel diodes conduct.Diodes side by side share one voltage. The one with the smaller drop turns on and holds the voltage below the other's drop.
- Expecting the reverse current to grow with the reverse voltage.Below breakdown it is set by the minority carriers, and hardly depends on the voltage.