The Physics of Lights
Light emitting diodes (LEDs) are a huge part of the world we live in. But how exactly do they work? Well, today we’re going to have a deep look into just that.
Light bulbs
Rather than jumping straight into LEDs, we’ll begin by understanding the traditional light bulb or “lamp” as most people are somewhat familiar with this. The light bulb was invented by Thomas Edison in 1879. It consists of a piece of glass that contains a socket that holds a looped wire.
Obviously, this doesn’t just produce light on its own. We apply a voltage (V) across the wire, which acts like a force to push the electric charge carriers (electrons) along the wire — this is known as a current (I). For low temperatures, as you increase the potential difference (voltage), a higher current will flow. This proportionality is known as Ohm’s law. However, as you increase the potential difference, eventually the current will be so high that there is a noticeable effect of the electrons colliding with the metal ions in the wire. In other words, the wire is becoming a worse conductor — its resistance is increasing. Thus, a limit of proportional has been exceeded and current and potential difference no longer increase linearly. This can be seen in the following current-voltage (IV) graph.
As you keep increasing the potential difference, more and more electrons collide with the metal ions leading to lots of wasted energy. This energy is first experienced as heat, and if you keep increasing the potential difference it becomes “red hot”: you can see the energy being given off as light. Eventually, the wire will become “white hot”, and you get even more light. So that’s all a light bulb is really… just a poorly conducting wire.
This design was all well and good for quite some time as it served its purpose: give off light. However, at its fundamental core, a light bulb is not a light generating device. It is a heat generating device, which when hot enough, gives off some light. In fact, when you put energy into a traditional light bulb, only 5% of the energy goes to light! The remaining 95% all goes to heat. It doesn’t take a genius to see that is incredibly inefficient both in terms of energy and money. Sure, for a single light bulb the energy loss is fairly negligible, but when you think that this has been used across the world for the last 142 years… that’s A LOT of wasted energy.
Intrinsic semiconductors
To solve this problem, the scientist Nick Holonyak Jr. invented the light emitting diode (LED) in 1962. The concept is all based on materials known as semiconductors. Think of a simple cable. Inside, you have a copper wire which conducts the electricity: it is a great conductor. On the outside, you have a layer of rubber which encapsulates the current: it is a great insulator. Semiconductors, as you might expect, can be either a conductor or an insulator depending on the situation. The most common type of semiconducting material is silicon. You can use silicon to make transistors which are like switches that can be either on (conducting ) which is like a one, or off (insulating) like a zero. And all of a sudden you’re working in binary and you have a computer! But that’s for another day. However, what I’m trying to drive at is that to understand how LEDs work, you have to know how semiconductors work. So let’s look at that.
So, let’s think about atoms. Atoms are composed of a nucleus at the centre and are orbited by electrons. One of the great discoveries of quantum theory (by Neils Bohr) was that electrons can only occupy certain discrete energy states or “levels” in an atom. Electrons cannot be anywhere/at any random energy level, they cannot only be at specific ones.
However, this is just a single atom. In daily life, you rarely find a single lonely atom, but rather, you get molecules and more complex structures. Well a semiconductor is crystal material, which means the atoms are basically formed in a perfect lattice: regular and repeated rows and columns of the same atom. Now, when these atoms are all next to each other, the energy levels of adjacent atoms overlap. But they do not simply superpose one another, as a result of what is known as the Pauli Exclusion Principle. This quantum mechanical principle states that two or more identical fermions cannot occupy the same quantum state within a quantum system simultaneously (wikipedia). Thus, the electron energy levels actually “smear” over a certain range, meaning you get several discrete “bands”. Within these bands, the energy levels are continuous so electrons are allowed to be there, but they are not allowed outside of the bands (the gap between bands is called the “band gap”).
Now if you think of electrons like marbles in a bowl, they will all want to go to the lowest point possible. Well the energy levels work the same way: all the electrons want to be in the lowest band, which is known as the valence band. The band above is called the conduction band. So at this point, all the electrons will be in the valence band (the valence band is said to be full), and you have an “intrinsic” semiconductor. As the band is full, when we apply a current there is no room for the electrons to move and so you get virtually no current.
Ok, so we haven’t really got a light emitting diode yet, we’ve just got a band full of static charges. We change this by putting some energy into the system by raising the temperature. If an electron gains sufficient energy, it will jump across the band gap into the conducting band.
Not only is there now a free electron in the conduction band, but there is a hole left behind in the valence band. Thus, if a current is now applied, not only does the electron in the conduction band cause a current, but the electrons in the valence band now have room to move and we get a current. Rather than thinking of all the negatively charged electrons in the valence band moving, you can just think of the hole as a positive charge which is moving in the opposite direction. And as you might image, if you increase the temperature further, more electrons jump the gap and you can get more current.
N-type and P-type semiconductors
So, the intrinsic semiconductor that I’ve just spoken about can be described as having no impurities. But what happens if we do add some precisely placed impurity atoms? We’ve still got the same valence band, band gap, conduction band sandwich as in the diagram above, but now we’ll add some atoms with additional electrons roughly at the energy level of the conduction band. As a consequence, these impurity atoms “donate” their additional electrons into the conduction band, and so they are often called “donor” atoms. Now there are free electrons in the conduction band which can move, allowing a current to flow. Note that the valence band has not changed: it is still full and thus current does not flow. This type of conductor is known as an n-type semiconductor — remember this by thinking of n for negative as the electrons are negative.
But you don’t just get an n-type, you also get p-type semiconductors. These are where those impurity atoms don’t have additional electrons in their outer energy level, but rather, they have missing electrons. In other words, they have room to gain electrons. Rather than “donor” atoms, these are often called “acceptor” atoms. And so electrons in the valence band are essentially stolen by the acceptor atoms, leaving holes behind. Like earlier, you can imagine these as positive charges that flow in the opposite direction to electrons. Thus, you can remember p for positive for a p-type material.
LEDs
So what on Earth is the use of these weird types of semiconductors? Well, you have to put an n-type material and p-type material together and you start getting some interesting outcomes. Now it’s not like you’re just coming along and gluing two separate things together. But rather, you have a piece of semiconductor material which is manufactured to have donor atoms on one side and acceptor atoms on the other.
So what happens? Well you may notice that there is an offset in the bands. There are electrons which can flow in the conduction band of the n-type material (right) and holes which can flow in the valence band of the p-type material (left). You can imagine that the electrons want to move left, up the hill, and the holes want to move right up their (upside down) hill. The problem is that unless we do something, they cannot get over that slope and so nothing happens. But as you can see in the diagram above, we don’t just do nothing: we apply a voltage across the materials. The positive terminal is connected to the p-type and negative to the n-type. This has the effect of “pulling down” the hill. Imagine it like a fence. Lots of runners are happily running in a straight line (current) but they get to a fence. If that fence is 6ft tall, that’s almost definitely going to stop them all from moving (like a resistor). However, if a force (voltage) is applied to pull down the fence to say 4ft, then some probably could get over that as the resistance has decreased. Decrease it to 1 ft and almost all will get over. And so you can say that resistance is inversely proportional to current… Ohm’s law. So in the case of these “hills”, when we apply a voltage we are allowing more electrons and holes to flow over the hill.
Now this is where it gets exciting. When the electrons have climbed over the hill, and the holes have flowed under their hill, you have electrons on top of holes. And remember what I said earlier? Electrons always want to be in the lowest energy state possible. Thus, those electrons fall (de-excite) from the conduction band to the valence in a process called recombination (which you can see occurring in the diagram above). But those electrons lose energy if they fall to a lower energy state, and we know that energy cannot be created or destroyed, so where does it go? Well this is the crunch point. Another huge discovery from quantum mechanics (this time by Einstein) was that light is just made up of packets of discrete amounts of energy, known as photos. Each photon carries an amount of energy given by E = hf, where h is Planck’s constant, and f is the frequency of the photon. And so, when an electron de-excites, the energy it loses is emitted as a photon. The frequency and energy of this photon corresponds to the difference between the energy levels that the electron has moved between. This photon is the light we observe from an LED! And without really meaning to, I’ve just explained how you can get different colours of LED: by using different materials with different energy levels, the photons emitted will have different frequencies and thus different colours. So we now have a device that has the primary effect of producing light when electricity flows through it — far more efficient than Edison’s light bulb!
Now linking back to that fence analogy. If you apply the force in the opposite direction, making the fence larger, then you are even more certain that no runners (charge carriers) will make it over. This is a defining property of a semiconductor diode: charge can only flow in one direction. No matter how big you make your voltage/potential difference, if it is in the wrong direction then diode will only resist the current. And you can also imagine that there is a critical point where you apply just enough force to the fence (in the right direction) that the runners can start to get over. Anything below this force and nothing will happen. This critical voltage is also a defining feature of diodes. Both of these features can be seen in the following graph.
This graph also shows that if you are past the critical point (often 0.6 volts) and continue to apply a greater potential difference in the right direction, then more current will flow. What’s the effect of this? A brighter LED!
Originally published at http://thephysicsfootprint.wordpress.com on January 2, 2022.
