Graf Bernadotte, meinen Damen und Herrn. That is about as far as I can venture into the German language.

I am very bad at languages and so I feel very sorry for interpreters. And hence I will stick fairly closely to my script. This will also help me perhaps to keep closely to the time schedule.

In the time at my disposal it would be impossible to cover in any detail the large field which comes under the title of particle accelerators. I shall limit myself to giving a general outline of those principles which have led to significant advances

and also to a somewhat more detailed discussion of the early work about which I can speak with more authority because I was then engaged on several of the basic types of accelerator. Hence I shall not show you pretty slides of modern giant accelerators,

nor make any significant reference to the nuclear knowledge obtained by the use of accelerators. The urge to produce particle accelerators arose out of Rutherford's work on the artificial transmutation of one element into another.

In 1919 he used natural alpha particles from radioactive substances to bombard nitrogen and he found that swift protons were emitted occasionally. Slide one shows the well-known picture taken by Blackett and Lees of such an event as revealed by a Wilson Cloud Chamber.

You will see up near the top of the picture a case of where one of the many alpha particles shown has struck a nitrogen nucleus

and a proton comes out towards the left but there is no sign of the alpha particle coming out and at that time it could be deduced that no other particle did come out because the neutron was not known. Now this is a type of disintegration which we call the alpha proton type

and for more than a dozen years this was the only type of nuclear disintegration known. It was obviously desirable that there should be available for disintegration experiments streams of high energy particles in large numbers and of various sorts.

Indeed this was mentioned in Rutherford's anniversary address to the Royal Society in 1927. In theory such particles could be produced by applying a few million volts to an evacuated tube.

The currents involved would be small. For example a microampere of helium ions corresponds in number to the total alpha particle emission from about 100 grams of radium. However the difficulties appeared to be very great.

By the standards of that time such voltages were expensive to produce and no one had succeeded in making a vacuum tube capable of withstanding more than about one tenth of the voltage believed to be necessary.

Hence people began to think about possible methods of producing fast particles by indirect or trick methods which would avoid the use of high voltages. There are in this audience young people whose career lies ahead of them and whose field of work is as yet undetermined.

It may interest some of them to hear how one gets involved in special scientific fields and so I will say in a few words what happened to me. I went to Cambridge as a research student in 1927 and after a few months in the nursery as it was called where we learnt some new techniques,

Rutherford called me to his room to discuss my line of research. On being asked if I had any suggestions to make I put forward the idea of accelerating

electrons by letting them move round many times in a circular electric field as in a modern betatron. He suggested a modification which appeared more practicable than my scheme and this is shown in slide two.

You see an evacuated tube there with a coil wound round the middle of it and the tungsten filament to give some electrons. This arrangement as Rutherford pointed out at the time was really just a modification of the

arrangement used by J.J. Thompson in his work on the electroless ring discharge in gases. A high frequency current from a spark discharge was sent through the coil shown there. This gave a rapidly changing magnetic flux and hence an alternating electric field was produced in the tube.

It was hoped that the alternating magnetic field together with a steady field produced by an electromagnet might produce suitable conditions for the acceleration of the electrons. But no trace of evidence was found for the presence of fast electrons.

Calculations were carried out which showed that the magnetic field present was of the wrong type to give radial stability to the motion of the electrons. The field increased from the center out to the coil.

The calculation showed that radial stability would be produced by a field decreasing inversely with the radial distance provided that a suitable high frequency radial electric field was also present.

Experiments along these lines were not successful because the arrangements were too crude and also because no provision was made for stability in the axial direction. Now these experiments were really quite nice ones to start with.

They had one very great advantage as far as I was concerned in that before starting on the work one had not got to read perhaps several hundreds of papers which had gone before. In fact at the time I didn't know of a single paper to be read on the subject and that is a great advantage.

I did discover later on that there was one paper I might have read. It happened to be a patent specification. When the failure of this method became evident another method was suggested to Rutherford.

It was the method utilized in what is now called the linear accelerator. The principle is shown in slide three and it is that the particles acquire their energy as a result of receiving a large number of successive pushes.

In between these pushes which they receive as they travel from one of the cylinders to the next as shown in the diagram, they pass through conducting cylinders and are thus unaffected by any changes made in the potentials of the cylinders. As shown in the diagram the odd numbered cylinders are connected together and so are the even numbered ones.

The two sets of cylinders are connected to the output of a high frequency generator. If the lengths of successive cylinders increase in the correct way it is possible always

to have an accelerating field present as the group of particles move between successive cylinders. Experiments on this carried out in 1928 at Cambridge failed for two reasons. The high frequency voltage was generated by a crude spark gap arrangement and

at the same time very little was known about the focusing of charged particles. Indeed the ends of the cylinders were covered with gauze in order to ensure a field free space inside them. And this effectively removed the natural focusing action which occurs in the gap between two cylinders with a potential between them.

Late in 1928 an important paper by Wiederer appeared in the archive for electro technique. In it he described experiments to verify the basic principle of accelerating

a particle in the circular electrical field produced by a changing magnetic flux. He was able to follow the electrons one and a half times around the circle. This was not the first publication on the Betatron.

The principle had been described by Slepion in a patent taken out in the United States in 1922. Physicists do not normally read these patent specifications. Wiederer's paper also described some experiments in which ions were given two accelerations in a linear accelerator arrangement.

He had developed further the idea of the linear accelerator first put forward by Ising in a Swedish journal in 1924. It is interesting to note at this early date a failure in communications about which we hear so much at the present time.

Although Cambridge at the time was regarded by many as the world center for atomic physics as it was then called, no one there and not even Rutherford himself appeared to have heard of Ising's paper.

The next slide shows a diagram of the method suggested by Ising. His proposal involved the breakdown of a spark gap and the application of the impulsive potential produced to various cylinders through suitable delay lines.

It is interesting to note that he included in his diagram the gauzes at the ends of the cylinders, which were included four years later independently and wrongly in the experiments at Cambridge. Wiederer's paper was important in another respect.

It was responsible for directing Lawrence's attention to the problem of accelerating particles to high energies. At a time when he was looking around for a new line of research, he came upon Wiederer's paper quite accidentally. He said that he knew very little German but was able to understand what the paper was about by studying the diagrams.

The result was that by 1931 he and Sloan were able to report the production of mercury ions of 1.26 million electron volts

using an accelerating potential of only 42,000 volts in a linear accelerator arrangement as shown in the next slide. You can see the series of cylinders up at the top. The rest of the diagram is in connection with the measurement of the energies.

The limit imposed by the electrical capacity of the cylinders in this method then became evident and so Lawrence was led to invent the cyclotron which in effect uses the same pair of cylinders over and over again. Fortunately, the disintegration of elements by artificially accelerated particles did not have

to await the development of the indirect methods which had been suggested. In 1928 Gamow and independently Condon and Gurney applied the then new wave mechanics

to account for the details of the emission of alpha particles from radioactive substances. It explained the statistical character of the emission and the well-known relation between the energy of the alpha particles and the half-life of the radioactive substance.

Cupcroft in Cambridge saw that the theory could be applied in reverse to the penetration of charged particles into the nuclei of atoms. Calculations showed that protons of quite moderate energies had a reasonable chance of penetrating into the interior of nuclei.

If they did so, one might expect that in a proportion of the cases disintegrations might follow immediately. Estimates indicated that a current of 100 microamperes of protons accelerated by a

few hundred thousand volts should produce an ample number of disintegrations for easy observation. The next slide shows Cupcroft and Gamow evidently well pleased with the result of these discussions.

Cupcroft on the left, Gamow on the right. Cupcroft showed these results to Rutherford and it was decided to test the theory by the use of fast protons accelerated directly by the application of a high voltage. It was further decided that Cupcroft would abandon his experiments on the deposition of metallic vapours onto surfaces cooled to

low temperatures and that I would abandon work on indirect methods and that we should work jointly on the new project. When planning to use the direct method employing a high voltage, sometimes referred to as the brute

force method, the first matter to be settled is the type of high voltage generator to be employed. The next slide shows the possibilities at that time to be as shown.

There's the induction coil which had a long and honourable place in physical laboratories. The impulse generator as used by Brash and Langer. The Tesla coil or resonant transformer as used at the Carnegie Institute by Bright, Chew, and Dahl. Or power transformers as used by Lauritzen in the United States.

Or a transformer and rectifier or electrostatic machines. Most of these methods produce voltages which vary with the time and thus when used to accelerate ions they give a beam containing ions with a wide range of energies.

Electrostatic machines tend to give a high voltage, low current output and this is just what is required for nuclear work. Furthermore they usually give a steady voltage and are thus capable of producing a stream of ions all of identical energy.

The built type of electrostatic machine was introduced by Van de Graaff in 1929 and has been a valuable tool in nuclear research. Its main use has been in the range of from about 1 million to about 20 million volts. For lower voltages the circular type developed at Grenoble by Felice has been found to be very suitable.

The Van de Graaff machine works on very simple principles which are well known and I shall not discuss them here. In 1928 the most suitable method of producing a high steady voltage of a few hundred kilovolts with an output

of a few milliamperes appeared to be a power transformer with its output rectified and smoothed by a suitable capacitor. This was the method chosen for the experiments in Cambridge.

Some development work had first to be done on the construction of high voltage continuously evacuated rectifiers and on the problems which arose when they had to be used in series. Fortunately the problems were greatly simplified by the use of a modification of a circuit introduced by Schenkel in 1919.

Using this it was possible to multiply the output of a transformer any even number of times and to produce at the same time a steady output voltage. It enabled the rectifiers to be placed one above the other to form a large glass pillar

which could be evacuated by a diffusion pump placed at earth potential as shown in the next slide. You can see the tall column of glass cylinders near the middle of the picture here and this is the accelerating tube over there.

Now this arrangement also ensured that the reverse voltages across the various rectifiers were equalized automatically. The apparatus is known as a voltage multiplier and was used in 1932 in the first disintegration experiments using artificially accelerated particles.

Voltages of up to nearly 800 kilovolts could be generated with this apparatus. The next slide shows a Wilson Chamber photograph of some of the disintegrations produced. You can see a lot of alpha particles coming out from the center of the picture.

These are produced by the bombardment of lithium with some fast protons. And you can see that there are a large number of these particles which were actually emitted in a time of something like one fiftieth of a second.

Indeed this lithium reaction is so easy to observe that the experiments could have been performed ten years earlier. For this about 20,000 volts would have sufficed but no one would have dared to do it at that time. Because anybody seen attempting to do this would have been more or less laughed at because he would

have been an ignoramus quite unconversant with the nature of the electric fields near the nucleus of an atom. There's something to be said all the same for trying what I sometimes call as a fool's experiment.

Provided that it can be done quickly, easily and quietly in a room which can be kept locked so that no one will know about the experiment if it is not successful. The Cambridge voltage multiplier apparatus was capable of disintegrating some

of the light elements and of producing artificial radioactivity in others. These results gave great encouragement to those working on the indirect methods for these were the only ones likely to provide particles sufficiently energetic to disintegrate the heavier elements.

There seemed to be no doubt that a gateway to large unexplored area was open to anyone who had available a beam of high energy particles. At this stage I'll return to the indirect methods and consider the major problems involved and the limits to the energy attainable in each case.

The betatron or magnetic induction accelerator already mentioned is of use only for the acceleration of electrons. The reason for this is that the particle acquires only a small addition to its energy during each revolution. And so the particle must travel around the circular electric field perhaps a million times while the flux increases from zero to its maximum value.

Electrons by reason of their small mass and high velocity can do this but positive ions with their much larger masses will make a much smaller number of revolutions and acquire only a small energy in the time.

As the electrons have to travel perhaps a million times around the circle it is essential that these should be not merely an equilibrium orbit for the electrons but that it should also be a stable one. Otherwise any slight disturbance would cause the particles eventually to move off the equilibrium orbit and be lost.

This problem of orbital stability is fundamental to all machines in which particles during acceleration must remain near to a definite path. The problem can be split into two separate parts, stability in the direction of the

magnetic field or axial stability and stability in the plane of the orbit or radial stability. The first of these arose in Lawrence's work on the cyclotron. The next slide shows the sort of magnetic field that you get near the edge of the pole pieces.

And the arrows on it show the direction of the forces acting on the particle going into the plane of the screen there. And you can see that these forces have got components tending to bring the particles back to the central plane of the apparatus.

If a particle wanders from the axial direction it is obvious that it will be brought back again. The magnetic field must be curved outwards as shown and this entails a magnetic field which decreases in the radial direction. Mathematically it means that if the magnetic field is proportional to one over R to the N then N must be greater than naught.

N equal to naught would correspond to a uniform field. The condition for stability in the radial direction is that if the particle starts to move on an orbit of slightly larger radius the magnetic field must be strong enough to cause this particle to move in a circle of smaller radius than that of the new orbit.

Hence as the centrifugal force falls off inversely with the radial distance the magnetic field must fall off less rapidly. And thus we see that if the magnetic field is proportional to one over R to the N then N is less than one.

Thus for stability in both directions this index, this field index N must lie somewhere between naught and one. The full mathematical theory of this was given in 1941 by Kurst and Serber who also dealt with the damping out of such oscillations as might occur.

An understanding of the results enabled Kurst to construct the first practical betatron which is shown in the diagram on the next slide. Now I'll not go into the details of that, I just want you to notice the dimensions. The width of that apparatus is something like half a meter and yet with this very simple apparatus he was able to produce

2.3 million volt electrons in sufficient numbers to produce a gamma ray intensity equivalent to those that emitted by one gram of radium. The largest betatron constructed gives 300 million electron volt electrons and was designed by Kurst and others in 1950.

It may well be the largest to be built for at this size many practical problems become severe. More important still a fundamental limitation arises.

An electron in its orbit is being accelerated all the time to the center and so we get a continuous radiation of electromagnetic energy. This loss is proportional to the fourth power of the energy of the particle and so a point is rapidly reached when the loss of energy per revolution by radiation becomes as great as the energy per revolution given to the particle by the betatron action.

Fortunately the electron synchrotron by reason of the greater radius of the electron orbit in it and the much greater energy added per revolution can give electrons a much greater energy.

Return now to the linear accelerator. In dealing with the betatron I spoke about the need for orbital stability of the electrons. We need also a similar stability in a linear accelerator so that particles cannot wander too far from the axis of the cylinders.

Hence it is perhaps more natural to talk about the focusing of the particle. Indeed this terminology has been extended to cover both linear accelerators and orbital accelerators. If there were no grids at the end of the cylinder some focusing occurs naturally as can be seen in the next slide.

These dotted lines represent the lines of electric force. The full lines represent the path of the particle. As the particle enters into this gap you can see that there is a force acting down this way along the direction of the lines of electric force.

And the particle gets accelerated up. Then there's a defocusing force here due to this line of electric force tending to pull the particle away from the axis. But the particle is then moving faster. This force acts for a shorter time and the overall effect is a focusing one.

Now the details of this were worked out by Rose and Wilson in 1938 for the cyclotron. The same theory applies to the linear accelerator. It's equivalent optically to a large number of convex lenses one after the other.

Now the next slide shows the type of result that we get arising from this focusing action. This is the axis of the accelerator and you have a particle starting off this way. If there are no focusing action it will go off like that. With this electrostatic focusing action we get a path like this for the particle.

One of increasing amplitude and increasing wavelength. And we see initially there's quite a helpful action in keeping the particle near the axis. But ultimately the electrostatic focusing is of no use. The particle is going too fast for the electrical lenses to have any significant effect.

Now unfortunately I should have said perhaps that the electrical focusing can be increased by having a varying potential.

One which decreases as the particles cross the gap. And in that case we get a stronger lens but unfortunately we get phase D focusing in the linear accelerator. When Lorenz realized the limitations of the linear accelerator he devised the method of the cyclotron.

And the first publication was by Lorenz and Adelafson in 1930. It verified the basic principle of the cyclotron. Next slide. Here we have really two hollow electrodes in the shape of D's between which an alternating voltage is applied.

And these two hollow electrodes behave like the first two cylinders in a linear accelerator. And just as if we used them over and over again. Now the particles in the cyclotron, I haven't time to go into the details, start somewhere near the middle and spiral outwards as they acquire energy each time they cross the gap between the two D-shaped electrodes.

Now the same electrostatic focusing occurs in near the center of the cyclotron but it's of no significance as you move outwards.

Lorenz introduced the magnetic focusing out there so we have this combination of electrical and magnetic focusing in action in the cyclotron. Now Lorenz built a number of cyclotrons of increasing size, the largest planned having a pole diameter of 4.6 meters.

And this would produce about 350 million electron volt particles. Now there is a difficulty which arises because he had to use a decreasing magnetic field to produce focusing. But this means that the particles get out of step with the alternating voltage as they move into this region of decreasing magnetic field.

And it meant that to get over this you would have to apply a very large voltage between the D's. It was planned for this large cyclotron that something like a million volts alternating potential would be applied between the D's.

And this would have raised very great technical problems. Fortunately a better solution turned up. In 1945 Macmillan in the United States and Wechsler in Russia independently drew attention to the existence of phase focusing which occurs under suitable conditions.

To understand what this means let us think of a linear accelerator designed to accelerate particles which cross the gap at a certain place on each radio frequency oscillation. Next slide please.

We imagine that the machine has been designed so that if particles cross here then when they go through the cylinders they reach the next gap just at the same part of the oscillation. Let us see what happens if a particle arrives with the right energy but a little bit later up here.

It means that it acquires more energy than the normal particle and so at the next gap it arrives a little bit earlier. It does not get quite as much additional energy as this got but still it gets some additional energy and it arrives still earlier at the next gap as shown over here. So the particle moves back towards this point on the oscillations.

It actually overshoots the mark and oscillates about it. Now this meant that if you had in the cyclotron a group of particles going round crossing the gap between the d's at this point here where there is no voltage that these could continue on round in a circle without acquiring any energy.

But if you decrease the frequency then they would start to come in at another point and acquire some energy, come in here and acquire some energy and they would move back again to this proper place for it.

This meant that by simply reducing the frequency you could expand this bunch of particles which are circulating in a circle. Or what was the same thing, you could change the magnetic field and the circle would expand out to meet the new conditions.

This in a few words is the basis of this idea of phase stability. And it enabled this large planned cyclotron of Lorentz's to be used not as a straight cyclotron but as what we call a synchrocyclotron where the frequency was changed to match the changing frequency of revolution in the weaker parts of the magnetic field.

Now I'm afraid the time is getting on and so I'll have to just make some cuts in what I have written out here. Now in the electron synchrotron we can introduce a great saving in the cost of the machine because we can reduce the volume of the magnetic field.

Instead of having particles spiraling outwards from the center they are all the time kept on a circle of constant radius.

And you can do this if you increase the magnetic field with time. Here again in the electron synchrotron the principle of phase stability works and one can produce very high energy particles in this way. It's rather interesting that the synchrotron works for the very reason that the cyclotron fails to work at the high energies.

You have in the cyclotron two limitations. One the necessity to decrease the magnetic field as you go outwards and the other the fact that the particles acquire additional mass as they speed up and this has the same effect as the decreasing magnetic field.

They get out of step rather quickly. If we just consider these synchrotrons you can also have a proton synchrotron by varying both the frequency and the magnetic field. We want to see how we can put up the energy of the particles produced.

Normally one would simply build a bigger magnet so that you had a larger circle for the particles to move round on. But if you can reduce the gap then it is possible to keep the magnet a reasonable size.

Now it's possible to reduce the gap if you can introduce additional stability into the motion of the particles. And this was the next advance, the next big idea in the history of the development of the subject. It meant that one needed a stronger focusing action than one had got previously.

Now this focusing action, this new focusing action called strong focusing or alternating gradient focusing was first of all suggested by Courant and Livingston and Snyder in 1952. Who found that the same idea had been suggested a little bit earlier in a patent filed by Christophilos in 1950.

Now the full theory of this method is difficult. There are a few simple examples of it given in most of the books that describe the action. The optical analog where you have a series of concave and convex lenses and you get an overall focusing effect there.

Or sometimes mechanical analogs are given such as the way in which the stability of a pendulum can be increased by vibrating the point of support up and down.

Now this, without going into details, this principle of alternating gradient focusing where in the axial direction you have a series of regions where you have strong focusing followed by strong defocusing and in the radial direction you get the opposite effect.

Now this means that in both the axial and the radial direction one gets very strong focusing, you can reduce the size of vacuum tube, you can reduce the gap in your magnet and you can keep the cost of the magnet and the size of the magnet down. Now this is the arrangement used in the largest machines which we have today such as the

Brookhaven one or the one at CERN or the new one that's just come into operation in Russia. And these give energies of varying amounts from about 30,000 million electron volts up to that sort of value.

Now if we look at, perhaps you would turn to slide number 21. This is a diagram due to Livingston showing the progress of linear accelerators over the years. Here are the DC methods round about 1930. This is a scale of

the energies of the particles plotted logarithmically. This is one million electron volts here. And you see the curves tend to turn over. Then later on we have the development of the electrostatic method, the Van der Graaff machine.

And you can see that curve begins to turn over when you get up to a few million electron volts. Then we have this linear proton accelerator and here we have the cyclotron. We have the betatron up here and the synchrotron up here.

And the proton synchrotron up here which is the type of the machines I've just been mentioning. Now all of these show a kind of upper limit lying roughly along this curve. And it corresponds to a very rapid rate of increase in voltage with time.

These are the years marked here and it's curious how you get very nearly a straight line on this semi-logarithmic plot. Now if we look to the future, see what's going to happen next, you've got to go up to very high energies indeed beyond the limits of that curve.

And it's obvious that one's getting to the stage when one needs a new idea. And the idea, various ideas have been suggested and I just mentioned one that has come to the fore in recent months.

This is a suggestion put forward independently by Wechsler and Budkopf way back in 1956. The basis of it is that if we can hold a cloud of electrons together and inject protons into it so that the number of protons is about 1% of the electrons present,

then these protons will be very firmly attached to the cloud by electrostatic forces. If the cloud is accelerated to high speed, the protons will also be accelerated. Because the mass of the protons is much greater than the electron mass, the energy of individual protons will be much greater than the energy of the individual electrons.

Some of the details of the scheme are shown in the next slide. The problem is to hold the cloud of electrons together because they tend to fly apart with the repulsive forces between the negative charges.

And this is done in this projected method, indeed some experimental work has been done on it already, by getting the cloud in the form of a ring. You have fast moving electrons. This helps because you have the pinch effect acting. If you have two parallel currents flowing in the same direction, they tend to move towards each other.

So you start off with a ring of electrons and the cross section is shown here and here. The ring comes right round like this. And it's held in this ring by a magnetic field which obeys the usual betatron conditions. Now what you do is you have these electrons going round with some few millions of electron volts energy.

And then you shrink this ring down by increasing the magnetic field. And you can shrink it down at the same time by a kind of betatron action, increase still further the energy of the particles as they revolve round.

And you end up with this ring like this. And you have electrons going round rapidly. They're held together and there's about one percent of protons mixed up. Then you accelerate this ring up in that direction by putting it through what is in effect a linear accelerator. You can push it out here by reducing the field due to the current in this coil.

The coil comes round here. That opens this magnetic field out and the ring gets pushed out. And it's accelerated up just as an individual particle might be. And the protons of course go out with it. Now this introduces the possibility of getting protons of energies very much higher

than the energies of the individual electrons that are accelerated in that ring. In fact if you started off with slowly moving electrons, one would get protons of nearly 2,000 times the energy of the electrons in the ring which is accelerated up.

Now preliminary experiments have shown the method to be promising. Even if machines of this type are successful, the energies given to particles are likely to be very small in comparison with that possessed by some of the cosmic rays.

We do not know the mode of origin of these. They may indeed be formed in some vast natural form of magnetic and electric fields in space. Which would surely be entitled to be called a celestial ultra high energy particle accelerator.