The incredible shrinking chip
The incredible shrinking chip

This free course is available to start right now. Review the full course description and key learning outcomes and create an account and enrol if you want a free statement of participation.

Free course

The incredible shrinking chip

1 The incredible shrinking chip

Two Scottish computer engineers with little or no physics knowledge set out to make a semiconductor transistor. This was 50 years ago, and their efforts gained them the Nobel Prize. The versatility of that transistor is now at the heart of the electronics industry. Millions of transistor switches are shrunk down into the microprocessors that are found in computers, mobile phones and almost everything else electrical.

The first transistor took years to plan and make; today more are made every day than there are people on the earth. In the following video clips, volunteers struggle to make a transistor using crude technology. There is also an explanation of the remarkable scientific advances that have now made the chip ubiquitous. The two volunteers discover the past while the presenter looks to the future.

The video clips include an interview with Gordon Moore, a founder of Intel, about the law of computing that bears his name. It's a rule of thumb that says processor power doubles every eighteen months. The video footage shows how the industry has managed to keep Moore's Law going for several decades. It explains the physics that make chips work, and how the same physics will eventually trip up the industry.

It is predicted that silicon will run out of steam in the near future, when the fundamental physical limit will be reached and a radical new technology will be needed.

The final part of the video returns to Silicon Glen and demonstrates how well the transistors built by the two volunteers actually work.

Activity

0 hours 40 minutes

As you watch the following video clips, make notes and list the issues described that affect the lifespan of the microchip.

 

Click to view the video extract (Part 1, 10 minutes).

Download this video clip.Video player: The incredible shrinking chip - part one
Skip transcript: The incredible shrinking chip - part one

Transcript: The incredible shrinking chip - part one

ANGELA LAMONT:
When the Forth Rail bridge was completed in 1890, it was called the eighth wonder of the world, a potent symbol of Scottish engineering excellence. Well, in the Victorian age the challenge was to make things big, now the challenge is to make things smaller. This is the latest pride of Scotland, the microchip, and millions of them are made here every year. This is the story of the science that makes the microchip tick, and the engineering that keeps it shrinking.
It’s easy to appreciate an achievement like this, it’s a bit harder with a microchip, the miracle is there’s not much to see. The bridge has millions of rivets in it, the chip has millions of microscopic electronic switches, and both were fabricated in Scotland. The bridge is made from steel, this is made from silicon. So the microchip, the thing that makes computers compute, is basically just humble sand, but modified by some very clever science.
To get inside the microchip and explore the science I’ve come to Scotland to set a challenge. Can two Scots who make a living building computers repeat the Nobel Prize winning experiment (from 1947) that started the era of the microchip?
ANGELA LAMONT:
Hi Alastair, John. So what do you actually do?
ALASTAIR HUNTER:
Well, we’re responsible for putting these PCs together.
ANGELA LAMONT:
So you know what’s inside the cream case, which is the only bit we ever see, isn’t it really?
ALASTAIR HUNTER:
That’s right, yes. The case really houses a number of components inside, which are all required for building the machine. The main part would be the motherboard, which is housed a the back there, that really keeps everything together. This on here is the main part, that’s the central processor, the CPU. That’s responsible for doing all the number crunching.
ANGELA LAMONT:
So that’s the brain really?
ALASTAIR HUNTER:
That’s the brain of the machine, that’s where it all happens.
ANGELA LAMONT:
Have you ever seen inside one of these?
JOHN WORTH:
Well actually yeah. A wee while ago we had a chip that was actually broken. To get through the ceramic I actually smashed it with a hammer. (laughter)
ANGELA LAMONT:
Subtle!
JOHN WORTH:
By the time I’d actually got through to the chip there was nothing left pretty much.
ANGELA LAMONT:
You’d just mangled it had you?
JOHN WORTH:
That’s right, yeah.
ANGELA LAMONT:
We’re going to ask you, not to build an entire microprocessor, ‘cos you might be there some time, but a little bit of one. How do you feel about that?
ALASTAIR HUNTER:
We’re using these components every day, so it’ll be interesting to see what goes on inside, yeah.
JOHN:
Yeah, definitely.
ANGELA LAMONT:
This is the Scottish Microelectronics Centre at Edinburgh University.
ANGELA LAMONT:
Hi Alastair, John. Are you ready to build that transistor then?
ALASTAIR HUNTER:
Let’s do it.
ANGELA LAMONT:
Come on then.
It’s here that Dr Les Haworth is going to guide our volunteers as they try and make the key component of a microchip.
What John and Alastair are going to build is an electronic switch called a transistor. Now microchips are made up of millions of tiny transistor switches, which are either on or off, one or zero. Inside a computer these ones and zeros are used to represent numbers or letters or colours, or pretty much anything you like. The thing is, to be any good at processing information the switch has to be able to change from on to off very quickly indeed. And that’s exactly what a transistor was designed to do.
The transistor started life in 1947 at the research lab of the Bell Telephone company in the USA. After ten years of effort William Shockley, John Bardeen and Walter Brattain invented the Holy Grail of electronics, a device that would amplify signals and switch very quickly. They called it the transistor. It was revolutionary and deserved a Nobel Prize because it was made out of solid material.
LES HAWORTH:
The significance was that for the first time electronic manipulation could be done within the solid state, the solid state is a term that many people from years ago might remember being applied to the transistorised material. Up until then valves had been used, but after that solid state devices began to be introduced. So that was the key defining moment of modern micro electronics technology.
What we're trying to do in this experiment is to replicate the original point contact transistor.
ANGELA LAMONT:
While John and Alastair take a trip back in time, I’ll be looking forward to see how the transistor has shrunk so much that tens of millions of them are now squeezed onto a silicon chip like this.
(actuality from Intel animation): ‘We are flying over the brain of the computer … the processor. As complex as a great city, but no bigger than a finger nail’
ANGELA LAMONT:
The relentless march of increasing computer power seems endless. As transistors get smaller, computers get faster … and transistors do keep on shrinking. But how long can it continue? The future of the transistor comes later. First a little history of very small things.
The electronic switches in early computers were called vacuum valves, which were unreliable, power hungry and bulky. The transistor would solve all these problems. No vacuum tube was needed, instead all the action took place inside a lump of material called a semiconductor.
LES HAWORTH:
What we need for this is a piece of semiconductor, and this semiconductor might be silicon, but it could also be germanium.
So we're going to start off with a a flat wafer of silicon. We're going to have a contact to the back of that, then we're going to bring down into contact with that two metal contacts two gold metal contacts which are going to contact the er the silicon surface, and the transistor action is going to take place between the two contacts.
ANGELA LAMONT:
In our transistor a sheet of semiconductor, silicon, will complete a circuit between two gold plates pressed onto the surface. The flow of electrical current through this circuit is switched off and on by essentially changing the silicon from a conductor to an insulator. That amazing transformation is controlled by a current flow through a third wire connected to the back of the silicon.
LES HAWORTH:
The transistor is basically er a switch in its very simplest form. The two contacts on the surface of the silicon pass the majority of the current, and that the amount of current which flows between those two contact is controlled by the current which we can inject at the back contact, so we have a small current in an ideal transistor which controls a much larger current flowing between the two contacts, and that's the principle of the transistor.
ALASTAIR HUNTER:
How is it that we make electrical contact to the semiconductor?
LES HAWORTH:
The metal contact made to the semiconductor material is very easy. What we are going to do is use some silver epoxy. So we’re going to take some glue which is loaded with silver particles, and we’re going to make contact to a gold plated base plate.
ANGELA LAMONT:
If you want to build a transistor you soon find out about the importance of semiconductor materials like silicon.
LES HAWORTH:
Electronics depends absolutely on semiconductors, without semiconductors there would be no computers there'd be no mobile phones, and there'd by no TV technology as we know it today.
ANGELA LAMONT:
So what is a semiconductor? Well, lets start with a conductor. That lets electricity flow through it. The most obvious example being a metal wire. An insulator like plastic is the exact opposite, it stops the flow of electricity.
But instead of having this all or nothing behaviour, conductor or insulator, imagine a material where you can choose exactly how you want the electricity to flow through it. Now that’s a semiconductor.
End transcript: The incredible shrinking chip - part one
Copy this transcript to the clipboard
Print this transcript
The incredible shrinking chip - part one
Interactive feature not available in single page view (see it in standard view).

 

Click to view the video extract (Part 2, 7 minutes).

Download this video clip.Video player: The incredible shrinking chip - part two
Skip transcript: The incredible shrinking chip - part two

Transcript: The incredible shrinking chip - part two

ANGELA LAMONT:
To make their transistor John and Alastair are using a thin wafer of silicon as their semiconducting material. Now it’s almost pure silicon but not quite. And that ‘not quite’ is very important, because by a slight, but precise, modification it’s been engineered to give exactly the right electrical properties. How? Well, lets take a closer look.
Electricity is the flow of subatomic particles called electrons. Now you can’t see these electrons, but imagine they’re like cars cruising down a main road. To increase the flow you could add more cars. Or you could remove some of the cars, so that the ones that are left move more quickly.
In semiconductors you’re able to control very precisely how electricity flows by adding or subtracting electrons from the material. Now this ability to engineer electrical behaviour is what makes semiconductors so useful and this crucial process is called doping.
RICHARD TURTON:
You intentionally introduce some impurities into the material, and the level of doping that's required is something like about one atom, one impurity atom in every million atoms of the the host material.
The interesting point is that you can actually control the level of conductivity in a semi-conductor by varying the amounts of impurities that are present in the material.
ANGELA LAMONT:
Our wafer of silicon has already been doped to increase the electrical conductivity. Now it has to be wired into a circuit that’ll make up the transistor. Two metal plates with connecting wires will complete the circuit.
LES HAWORTH:
The idea is to put the two plates together with the insulator in between them, which is going to separate the contacts by this very small 50 micron distance. And then put one corner of each into these holes here and do this screw up, so that it’ll hold them firmly. And because of the conical screws the tighter you do the screw the more firmly it’ll clamp them together, so you’ll get a really good nice tight contact. The first thing we’ll have to do …
ANGELA LAMONT:
John and Alastair are about to learn the hard way that building a transistor this way isn’t easy. It’s a very fiddly job getting the right pressure of contact onto the silicon surface. The inventors knew they had to think of something better.
LES HAWORTH:
They realised that a much better solution would be to incorporate the whole of the transistor action within the silicon rather than rely on the surface effect and metal contacts on the surface, and they then went on to develop that technology, quite quickly in fact after the original experiment.
ANGELA LAMONT:
In the beginning there was just one transistor … now there are too many to count.
This kind of set-up was never going to be a suitable method for mass production as our volunteers are finding out. They’d be happy to get one transistor to work never mind a million!
(Actuality of frustration from the boys)
Fortunately the inventors soon realised they didn’t need clumsy metal pressure contacts. Instead, why not make a transistor by doping tiny regions inside the semiconductor to control the flow of electricity? The precision of this doping technique meant transistors could be shrunk down dramatically. But there was a limit to the size, a protective case and connecting wires were still needed.
Transistors are crucial components in a microchip but they’re not the only ones. You need connecting wires, capacitors, and other bits and pieces to make them into a useful circuit. So the next amazing revolution in the story of the microchip was to shrink all these components down onto one piece of semiconductor, an integrated circuit.
RICHARD TURTON:
Well, the biggest step really in going from the individual transistor to where we are now, was the invention of the integrated circuit. Someone had the idea to put two or three transistors on to the same piece of of silicon, and actually put all the connections and everything else into that, on to that piece of silicon, so that they were actually creating a circuit, without any need for soldering or anything else.
(actuality ex Intel animation) “The processor has many layers. Together these layers form millions of transistors …”
ANGELA LAMONT:
In 1971 the new integrated circuit technology was used to make the first microprocessor at a fledgling company called Intel.
GORDON MOORE:
When Intel first introduced the microprocessor, it actually was the result of a programme we had to develop chips for a calculator for a Japanese calculator company. Some of our engineers saw that by making a general purpose computer on a chip, not only would it do the calculator functions but could be used for a variety of control applications and so forth. The whole field of computing has just been dramatically changed because of the er invention of the micro processor and its developments.
ANGELA LAMONT:
The first microprocessor contained just over two thousand transistors, now millions more are squeezed into each new generation of chip. With more transistors inside, computers have become ever more powerful, and so has the industry that builds them.
Now silicon valley is just part of a global engineering revolution.
The heart of the Scotland’s electronics industry is ‘Silicon Glen’ which stretches in between Edinburgh and Glasgow. Here at this Motorola Fabrication plant they make the microchips you might find in your car.
These days silicon chips control just about everything from the engine to the windows. But although chips can do very different things, they’re all made in pretty much the same way.
RICHARD TURTON:
What actually happens when they construct these things, is they they take a wafer which is something like about nowadays twenty centimetres across. Er so imagine a disc of of silicon, and that's partitioned so they make lots of these little circuits next to each other.
LES HAWORTH:
So it’s actually this contact here that’s not working. So what I’m going to do now is twist this round, to try and get these aligned. If I turn this screw we should be able to twist it …
ANGELA LAMONT:
Transistor technology has certainly come a long way since the early days
(actuality of Les)
Oh dear…
End transcript: The incredible shrinking chip - part two
Copy this transcript to the clipboard
Print this transcript
The incredible shrinking chip - part two
Interactive feature not available in single page view (see it in standard view).

 

Click to view the video extract (Part 3, 11 minutes).

Download this video clip.Video player: The incredible shrinking chip - part three
Skip transcript: The incredible shrinking chip - part three

Transcript: The incredible shrinking chip - part three

ANGELA LAMONT:
More from the boys later, meanwhile back to the future to find out about the struggle to make chips today.
Aha, you’ve found me, stripped of my ‘presenter’ makeup and cocooning myself in this rather lovely all over ‘bunny suit’. When it comes to chips you’ve got to keep it clean. The dirtiest things about the while process is the people. And even a tiny flake of skin can become huge problem on the microscopic scale of a chip.
RICHARD TURTON:
One of the most important things in in manufacturing, any micro electronic device, is cleanliness, and really the reason for that, is because, these things are so tiny.
ANDREW BEAUMONT:
Over a space of one or two centimetres your electrical signal has to go through or can go through many millions of transistors, and each one of those transistors is crucial for correct functioning of the chip.
You could fit around about two hundred transistor active areas into a hairs breadth.
ANGELA LAMONT:
Right, so it's pretty important to keep your hair on then?
ANDREW BEAUMONT:
Absolutely critical.
ANDREW BEAUMONT:
This is the gowning area for a wafer fab. These suits have to go on over our every day clothes to keep er particulates and shedding materials getting into the atmosphere and potentially contaminating the products that we're making.
ANGELA LAMONT:
This is one of the cleanest places on earth, cleaner than an operating theatre…
ANDREW BEAUMONT:
… it's more than a thousand times cleaner.
Initially it takes a little bit of getting used to but after a couple of weeks of working in there er it's pretty comfortable.
They initially look very much the same but you get used to peoples size and particular shape, the way they walk. Apart from that each suit has got the persons name on the front of it so if you don’t recognise them from the walk then you can just have a quick look at their name.
ANGELA LAMONT:
John’s finding that dust can get in the way of making a good contact to silicon.
(actuality from Les)
I’m sure it’ll work eventually.
Come and have a look in here. This is one of the key areas. I mean, one of the first things you notice is the colour. It’s essentially a photographic process going on in here, so they use yellow light so it doesn’t interfere with it. It’s called photolithography and that’s where they put a pattern of all the major features they want on the microchip down onto the silicon wafer.
ANDREW BEAUMONT:
Basically you project a light through a mask on to a photo sensitive material, and that mask basically weighs out the pattern of er the transistors and the interconnecting circuitry.
The photolith process defines the sizes of those devices, so it is absolutely the cornerstone of shrinking to smaller and smaller geometries.
ANGELA LAMONT:
Once the pattern’s on the silicon wafer this gas plasma etches away at the areas you don’t want, creating channels in the silicon.
This is where the doping process takes place by implanting them with phosphorous or boron, and it’s the precision of this process that means our transistors can keep on getting smaller and smaller.
But can it continue? We’ll explore the limits later, first the good news.
There’s progress in the lab.
LES HAWORTH:
Voila! That’s a diode characteristic as we would expect from a single contact. You can see the voltage at this polarity, we have no current flowing, and then we have some current flowing at this polarity. That’s exactly the characteristic that we want, so we’re winning so far.
Now we’ve got it set up so we can look at the other contact and see whether we’ve got a good contact there, so if we push the measure button again … and see what happens.
ANGELA LAMONT:
Oops!
LES HAWORTH:
OK, and this isn’t looking so good, because we’ve got a straight line along the bottom. We’ve got no current. So it rather looks like we’ve not got a contact there at all. Let’s do an autoscale just to check on that … and that’s just noise again, so we’ve not got a good contact.
ANGELA LAMONT:
Getting two good contacts to the silicon surface is the problem. Hopefully it’s just a matter of time.
LES HAWORTH:
No luck yet. More of the same is called for.
ANGELA LAMONT:
After all it took ten years of research to make the first transistor, but since then the pace of change has been beyond anything we’ve seen before.
In 1957 when they came to build a new road bridge over the Forth new engineering techniques and materials meant it was completed in less than half the time of the old rail bridge, with only a tenth of the workforce. But progress in the microchip dwarfs that achievement.
This card has a microchip inside it. You open it up, it plays you a little tune, it makes you laugh and costs about £2.00, but it contains more processing power than existed in the entire world before 1950. Now, if cars had progressed at the same rate as silicon chips your average family saloon would travel at twice the speed of sound, and cost about 50 pence. This phenomenal progress is summed up by a remarkable prediction for the incredible shrinking chip … Moore’s Law.
GORDON MOORE:
Hi I'm Gordon Moore, er I've spent something over 40 years in the semi conductor industry.
Moore’s Law is a term that got applied to a curve I published in 1965 that projected how complex integrated circuits were going to get over the next decade, that is how many individual transistors and resistors we were gonna put on a single chip. Er taking the very early data I made the bold extrapolation from about sixty components to something like sixty thousand over the 10 year period. Er one of my friends dubbed this 'Moore's Law' and it stuck ever since.
RICHARD TURTON:
Gordon Moore looked at integrated circuits, which at the time had only been around for a few years, and he noticed that the number of devices that people were managing to put on an integrated circuit, was going up roughly by a factor of two every year. I'm sure he certainly didn't expect it to carry on for another thirty odd years.
GORDON MOORE:
I certainly didn't anticipate that it was going to continue this long, I thought er extrapolating for 10 years was really a stretch at the time I did it.
It certainly is a technical challenge to keep things moving this rapidly. We've gone from single transistors to hundreds of millions of transistors that on a chip in a 40 year period.
ANGELA LAMONT:
To achieve that rate of progress the industry has relentlessly stuck to Moore’s ‘law’. That means doubling the number of transistors on a chip about every two years.
Well maybe that doesn’t sound so hard. But let me tell you a story about doubling.
Legend has it that the Emperor of China was so taken with the game of Chess he summoned the inventor to reward him. Now the inventor’s request was simple, all he wanted was one grain of rice on the first square of his chess board, two on the second, four on the third, and so on, doubling for each successive square. Well the Emperor agreed, but he had no idea how hard it would be to keep his promise. One, two, three …
RICHARD TURTON:
By the time you get to the tenth square, you're up to about a thousand grains of rice. By the time you get to the twentieth square, you're up to a million grains of rice, and by the time you reach the last and sixty fourth square, you're up to something which is a nineteen digit number, which is far more than the grains of rice in the whole world.
ANGELA LAMONT:
… 508, 509, 510, 511, 512. Fortunately the Emperor’s problem was mythical. But for the microchip scientists it’s all too real. Keeping up with Moore’s Law means that they have to double the number of transistors on a microchip every 18 months to two years. It’s the equivalent of filling yet another square on the Emperor’s chess board.
RICHARD TURTON:
In 1999 circuits with up to a quarter of a billion transistors were being produced. So, that's now equivalent to being on about the twenty eighth square of the chessboard.
ANGELA LAMONT:
And they’re beginning to feel the pain.
LES HAWORTH:
One of the biggest practical problems is just er keeping the technology going to to make the devices physically smaller and smaller. Patterns on the semi conductor are defined optically, er and depend very much on the wavelength of light which are used
GORDON MOORE:
In order to go smaller we have to get a shorter wavelength, and it's hard to make images smaller than the wavelength of the light you're using. That means we're working well into the ultraviolet now, and we're about at the point where there are no optical materials, no glasses or crystals that are transparent to the light we would like to use. So we're going to have to move from lens systems to all reflective systems we use mirrors instead of lenses to make the optical system. And we have to make them with the precision er that has never been done on an industrial basis before. Er the mirrors will have to have better figures er more precise than the Hubble Telescope has, but they'll have to be made in production quantities.
End transcript: The incredible shrinking chip - part three
Copy this transcript to the clipboard
Print this transcript
The incredible shrinking chip - part three
Interactive feature not available in single page view (see it in standard view).

 

Click to view the video extract (Part 4, 3 minutes).

Download this video clip.Video player: The incredible shrinking chip - part four
Skip transcript: The incredible shrinking chip - part four

Transcript: The incredible shrinking chip - part four

ANGELA LAMONT:
Even if the many practical problems are solved, the current silicon technology will eventually run out of steam, there is a fundamental limit. It has to do with the electrons that carry electricity around a microchip. As the width of the components shrinks to just a few atoms the electrons start to ‘leak out’ of the incredibly thin channels designed to hold them. It’s as if the cars in the city suddenly started leaving the roads and randomly appearing anywhere …
RICHARD TURTON:
There are numerous other proposals erm. People have, are doing work into looking at molecular computing where individual molecules could be, could function as a single electronic device, or even quantum computing where it could be a single atom or even a single electron, which performs some sort of electronic function. So, the end certainly isn't in sight for Moore’s law.
Er it's certainly a very exciting time to be a physicist.
ANGELA LAMONT:
At the end of a very long day, a new transistor is born …
LES HAWORTH:
Whoopee! Bingo!
ANGELA LAMONT:
Do I smell success here?
LES HAWORTH:
All the hard work has paid off. What we have here is a transistor characteristic. We’ve verified that we have the two contacts aligned on the silicon. We can be sure that they’re both in contact.
ANGELA LAMONT:
Well, well done. How’s it been? Piece of cake, or a bit of a challenge?
JOHN WORTH:
Well I suppose it feels quite good in the end, now that it actually works, we’ve seen it work.
ALASTAIR HUNTER:
I don’t think we’d have appreciated at the start just how difficult it was to get it to work reliably. So that’s been a bit of an eye opener.
ANGELA LAMONT:
How about if I wanted 9 million of them, all exactly the same, so we could make a microprocessor?
ALASTAIR HUNTER:
I think that would be a tough challenge.
ANGELA LAMONT:
(laughs) Well, we’ll let you off this time then, but well done guys.
LES HAWORTH:
I think John and Alastair did a superb job. They came in, completely cold, and by the end of the day they had made a a working transistor.
I think they can be really proud of the achievement that they had.
End transcript: The incredible shrinking chip - part four
Copy this transcript to the clipboard
Print this transcript
The incredible shrinking chip - part four
Interactive feature not available in single page view (see it in standard view).
4RAIL_1

Take your learning further

Making the decision to study can be a big step, which is why you'll want a trusted University. The Open University has 50 years’ experience delivering flexible learning and 170,000 students are studying with us right now. Take a look at all Open University courses.

If you are new to university level study, find out more about the types of qualifications we offer, including our entry level Access courses and Certificates.

Not ready for University study then browse over 900 free courses on OpenLearn and sign up to our newsletter to hear about new free courses as they are released.

Every year, thousands of students decide to study with The Open University. With over 120 qualifications, we’ve got the right course for you.

Request an Open University prospectus