| || || |
| || |
| || || |
From the 1946 book, Dawn Over Zero.
* * *
ON February 17, 1939, a meeting of the American Physical Society was held at Columbia University. Anyone not a nuclear physicist who by chance might have wandered into Lecture Room 301 of the Pupin Physics Laboratory at Broadway and 119th Street that day would have found the proceedings very dull indeed.
Very few even among the physicists present were more than vaguely aware of the enormous social implications of the words being spoken or the symbols being written in white chalk on the blackboard.
Yet these words and symbols changed the course of history. Through them the atomic bomb, and atomic power for the benefit of man, became a reality.
The participants included a large number of the pioneer explorers of the nucleus of the atom who later played a major part in the development of the atomic bomb. Among them were Drs. Bohr, Fermi, I. I. Rabi, Wheeler, Dunning, and scores of others who subsequently became leaders at Oak Ridge, Tennessee; Hanford, Washington; Los Alamos, New Mexico, and the many other centers of the Atomic Bomb Project.
No mention of an atomic bomb was made at that meeting, but a number of those present, as I learned later, were already thinking, and worrying, about it.
The phrases that kept recurring through the discussion like a refrain were “chain reaction,” “fast neutrons,” “slow neutrons,” “intermediate neutrons.” On the blackboard the recurring symbols “U.238” and “U.235,” scribbled by Dr. Bohr, were interspersed among graphs and mathematical formulae.
They represented the basic fundamental secret of the atomic bomb, written so plainly that even a non-physicist like myself could read them.
The meaning of the symbols is very simple. Uranium as found in nature consists of three types (isotopes), all inextricably mixed together. All three have ninety-two protons in their nucleus, but each has a different number of neutrons.
By far the most abundant type is uranium of atomic weight 238, which has 146 neutrons in its nucleus; it constitutes 99.3 per cent of all the uranium in the world.
The second most common form is uranium of atomic weight 235, the nucleus of which contains 143 neutrons. It represents only 0.7 per cent of the uranium supply. The third type, uranium of atomic weight 234, contains 142 neutrons in its nucleus and represents only 0.008 per cent of all the uranium in the earth’s crust.
They are designated by the symbols U.238, U.235, and U.234.
In the three weeks that had elapsed since the Washington conference Dr. Bohr, in collaboration with young Dr. Wheeler of Princeton, had developed a theory of the fission process. It was this theory that he was expounding for the first time at the Columbia meeting.
In order to explain some puzzling phenomena observed in the various laboratories where experiments on uranium fission were carried out in the last week of January 1939, Drs. Bohr and Wheeler had worked out a hypothesis that fitted the facts and provided guideposts for future investigations. It was only a prediction at the time, since no means were available to check it by experiment.
But a little more than a year later, in March 1940, the prediction was confirmed by experiments carried out at the University of Minnesota, Columbia, and the research laboratories of the General Electric Company.
Neutrons coming out of atomic nuclei with very high energies, which make them travel at speeds of thousands of miles per second, can be slowed down, through elastic collisions with atoms of light elements, to speeds as low as one mile a second.
| || || |
| || |
| || || |
|Dr. Ernest Lawrence works on his early 37 inch cyclotron, mid-1930’s.|
An elastic collision takes place when a moving body collides with another body of nearly equal weight, as happens when one billiard ball strikes another. In such a case the moving body imparts some of its energy to the object it collides with, so that it is slowed down. When a moving object strikes a much heavier object, as, for example, when a rubber ball is thrown against a rock, an inelastic collision takes place, the ball bouncing back with almost the same energy it had to start with.
To slow down neutrons, they are, therefore, made to collide with elements containing a large number of hydrogen atoms, which have an atomic mass nearly equal the mass of the neutron, or with other elements of low atomic mass. Such slowing-down elements for neutrons, known as “moderators,” are:
• water, two thirds of the atoms of which consist of hydrogen;
• “heavy water,” in which the hydrogen has an atomic mass of two, with one proton and one neutron in its nucleus;
• paraffin, which contains a large number of hydrogen atoms; and
• such light elements as beryllium and carbon.
To serve as a neutron moderator an element or compound must not absorb too many neutrons.
In the experiments on uranium it had been observed that both slow neutrons and fast neutrons produced fission, whereas neutrons of intermediate speeds were rather inefficient in producing a split. On the basis of these observations Drs. Bohr and Wheeler arrived at the conclusion that the fast-neutron and slow-neutron processes were due to two different isotopes of uranium.
The fast neutron, they theorized, produced fission in U.238, whereas the slow neutron produced the splitting of U.235.
The inefficiency of the neutrons of intermediate energies in producing fission was explained on the grounds that neutrons in that range of energies were the only ones that could be properly digested by the “cannibal” in the uranium, which, they predicted, existed only in the U.238.
In other words, the energy level in the U.238 nucleus resonated in tune with the energy level of the intermediate neutrons, so that the U.238 nucleus could swallow one of them without suffering an attack of indigestion severe enough to make it burst.
In such an event—namely, the capture of a neutron of intermediate speed by the U.238 atom without the production of a split—the nucleus of the U.238 is increased in weight, as it now contains 147 neutrons instead of 146. In other words, the U.238 atom is transformed into a new isotope of uranium, of atomic weight 239 (U.239).
It was further believed, on the basis of empirical observations, that the probability of such a neutron being captured by U.238 was much greater than the probability of U.238 being split by a fast neutron. This was because it was found that a large percentage of the high-speed neutrons, the only ones that could cause fission in U.238, lost energy by inelastic collisions with the heavy U.238 nuclei without entering them.
This meant that a large proportion of the fast neutrons emitted in the fission process would become intermediate neutrons of the energy range that could be swallowed by the U.238 without producing fission.
The importance of this finding, from the immediate military point of view, could hardly be overestimated. For it revealed that the U.238 would eat up so many neutrons as a result of their inelastic collisions that not enough fast neutrons would be left to maintain a chain reaction.
This ruled out completely the possibility of using U.238 in an atomic bomb. The cannibal in the uranium turned out in reality to be the instrument of a benevolent destiny, for U.238 was the only type of uranium readily available in large amounts for almost immediate use.
It made certain that the Nazis would not have atomic bombs in time to begin their war on civilization, which, as everyone sensed at the time, was just round the corner.
| || || |
| || |
| || || |
|(left to right) Niels Bohr, Werner Heisenberg, and Wolfgang Pauli meet in the mid-1930s. Bohr fled German occupied Denmark, Pauli fled Austria leaving Heisenberg to research nuclear fission for the Nazis.|
But the threat of the Nazis developing an atomic bomb was only postponed, not eliminated. For it was soon realized that U.235 could be split not only by slow neutrons but also by fast neutrons.
More significant still, the theory gave strong grounds for believing that enough free neutrons would be liberated in both cases to maintain a chain reaction with fast neutrons as well as with slow ones.
This meant that U.235, if a method could be found for separating it in large quantities from the 99.3 per cent of U.238, could serve both as material for an atomic bomb of enormous destructive power, by bringing about an uncontrolled chain reaction with fast neutrons; and as a substance for utilizing atomic energy as a fuel two million times more powerful than gasoline, through a controlled chain reaction with slow neutrons.
The greatest obstacle that stood in the way at the time was that no method for concentrating U.235 in large amounts was known and the problems for developing such a method were of such magnitude as to make it obvious that it would involve unprecedented expenditures of time and money.
Considering what they stood to gain by it, it appeared certain that the Nazis would get busy at once on an all-out effort to develop such a method.
But what about the democracies? Could they be made to realize the danger?
Would they, who were not even producing tanks and guns and airplanes in adequate amounts, stand ready to venture millions, possibly hundreds of millions, on a weapon based on mere theory, the correctness of which no one could prove? One might as well expect Congress to appropriate millions for a perpetual-motion device.
But Dr. Fermi, who had been in this country only a few weeks, just refused to “know better.” Immediately after the Columbia meeting he approached his chief, Dean George B. Pegram of Columbia, with plans to bring the matter at once before the proper Army and Navy authorities.
With the aid of introductions by Dean Pegram, Dr. Fermi set off shortly thereafter for Washington.
After the publication of the discovery of uranium fission, when scientific periodicals in this country, France, and England devoted much space to new developments in this field, there came a complete blackout on further work on uranium in Germany.
This silence was in itself strong confirmation of what our scientists had suspected from the very first.
But even stronger confirmation came soon after, through bits of information passed on by the underground and newly arrived exiles. When pieced together, these bits revealed that a large section of the Kaiser Wilhelm Institute in Berlin had been set aside for intensive research on uranium, and that some two hundred top-ranking German scientists had been ordered to devote all their energies to the problem, which had been given top priority.
The Nazis were at work on an atomic bomb.
At that time, as was learned later, work on the uranium problem and its possible military applications was being pushed vigorously in England. But in this country progress from a military standpoint was at first painfully slow. As Professor Henry De Wolf Smyth points out:
“At that time  American-born nuclear physicists were so unaccustomed to the idea of using their science for military purposes that they hardly realized what needed to be done.”
Consequently, as Professor Smyth states, the early efforts both at restricting publication and at getting Government support were stimulated largely by a small group of foreign-born physicists centering on Dr. Szilard, a native of Hungary, and including Professor Eugene P. Wigner of Princeton, Professor Teller, Professor Victor F. Weisskopf of the University of Rochester, and Professor Fermi.
While it was impossible to carry out large-scale experiments because of the unavailability of sizable amounts of U.235, it was nevertheless realized that even with a submicroscopic sample it could be determined whether U.235 was fissionable by slow neutrons, or fast neutrons, or both. If this was found to be the case, it would serve as proof that the Bohr-Wheeler theory was correct, and that, in turn, would lend strong support to the correctness of other predictions based on the theory.
The job of obtaining samples of pure U.235 was undertaken by Dr. Alfred O. Nier of the University of Minnesota, and Drs. K. H. Kingdon and H. C. Pollock of the General Electric Company’s research laboratories at Schenectady. They built improved models of an instrument known as a mass spectrometer (to be described later), and by March 1940 they had produced the first bits of pure U.235 that ever existed on earth.
Dr. Nier arranged to produce two samples “weighing” one billionth and two billionths of a gram, respectively. Drs. Kingdon and Pollock managed to produce one weighing one hundredth of a millionth of a gram.
At their rate of production it would have taken thousands of years to concentrate one gram, but those infinitesimal amounts, the most precious bits of metal in the world, were enough for the purpose. In a sense they were the first atomic bombs.
The samples were rushed to Columbia University, where they were subjected to bombardment by slow neutrons produced by the cyclotron. Those first historic tests were carried out by Drs. E. T. Booth, Dunning, and A. V. Grosse.
| || || |
| || |
| || || |
|Dr. John Dunning (left) with Fermi and Mitchell at the Columbia University cyclotron.|
The first sample, produced by Dr. Nier, was tested on March 3, 1940; the second, obtained by Drs. Kingdon and Pollock, on March 20; a third, concentrated by Dr. Nier, was tested on April 3.
These tests established definitely that only U.235 could be split by slow neutrons, and that U.238 could be split only by fast neutrons, exactly as Drs Bohr and Wheeler had predicted.
Other tests in other laboratories soon revealed that U.238 eats up too many neutrons that have lost energy through inelastic collisions with it (intermediate neutrons), and therefore would not sustain a chain reaction.
By June 1940, another fact of the utmost importance for later developments became generally known both here and abroad. It was found that U.235 could be split not only by slow neutrons but by fast neutrons as well.
Since slow neutrons do not occur naturally and require large quantities of a light element to serve as a moderator, it would be impossible to employ them in a bomb, as such a bomb would have to be of enormous dimensions. Furthermore, the reaction would be too slow.
On the other hand, fast neutrons, if there were enough of them to sustain a chain reaction, could create an explosion of unprecedented violence in a small amount of U.235.
If there were enough of them to sustain a chain reaction—there was the rub. Nobody knew for certain whether there were, and yet this was the very heart of the problem.
To find out whether a chain reaction with fast neutrons would be self-sustaining, sizable quantities of concentrated U.235 would be necessary, and that would take millions of years to produce by the best methods then known, unless hundreds of millions of dollars were to be spent to build mammoth plants that might or might not work.
It was a vicious circle: no fast-neutron chain reaction could be demonstrated without sizable amounts of U.235, but no such amounts could even be dreamed of unless it could be demonstrated with a reasonable degree of certainty that a self-sustaining chain reaction would take place.
But how about using the unseparated U.235 in a natural mixture of uranium? Since fast neutrons would split both U.238 and U.235, why not take just a big block of uranium as found in nature, remove the neutron-absorbing impurities, and let a stray neutron start a chain reaction?
Luckily for the world, this was not possible because of the existence of the “friendly cannibal” in the U.238. Tests revealed that the number of neutrons captured and devoured by him without splitting would be great enough to prevent any chain reaction.
How about slow neutrons? The cannibal does not touch slow neutrons, whereas the probability of U.235 being split by slow neutrons had been found to be greater than the likelihood of its being split by fast irons. These two factors should lead to a vast increase in the number of neutrons available to split more atoms, and the greater the number of atoms split, the greater the number of neutrons born.
If the birth rate of the neutrons available for further fissions is higher than their death rate, then a self-perpetuating chain reaction should take place in the U.235 even without separating it from the U.238.
In other words, with slow neutrons, ordinary uranium, as found in nature, could be used for determining whether enough neutrons were born through fission to maintain a self-perpetuating chain reaction.
If that was found to be so, it would serve as an indication that a chain reaction in U.235 could be maintained also with fast neutrons. For in a natural unseparated mixture of U.238 and U.235 the cannibal in the U.238 would still swallow a great many of the neutrons when they passed through the intermediary speed range, in the course of their being slowed down to low speeds.
On the other hand, in pure U.235, which has no pronounced cannibalistic traits, nearly all the neutrons would be available for further fission and the production of an uncontrolled chain reaction—namely, a nuclear explosion.
The idea of testing the possibilities of a chain reaction with slow neutrons in U.235 unseparated from natural uranium occurred independently to Dr. Szilard, Dr. Fermi, and their associates at Columbia, as well as to scientists in England, shortly after the discovery of uranium fission was announced. It offered enormous obstacles, but it was the only way in which the feasibility of an atomic bomb could be proved by experiment, and without such proof there would be no justification for the enormous expenditures that would be entailed.
| || || |
| || |
| || || |
|(left to right) Ernest Lawrence, Enrico Fermi and Isidor Rabi. Could our scientists control the production of fission.|
As it turned out later, the decision to go all out on the Atomic Bomb Project was made a year before such proof had been obtained; but by that time strong circumstantial evidence was already available, and in the face of the national emergency that was considered enough.
In the words of Mr. Henry L. Stimson, when he was Secretary of War: “The decision to embark on large-scale production at such an early stage was, of course, a gamble, but, as is so necessary in war, a calculated risk was taken and the risk paid off.”
And so at Columbia University, behind a thick veil of self-imposed secrecy, Drs. Szilard and Fermi, working at first independently along different lines, but joining forces later in a team that included Drs. Anderson and Zinn, George Weil and B. Feld, began work on the design of a structure in which, they hoped, a chain reaction with slow neutrons could be made to operate in an unseparated natural mixture of U.235 and U.238.
One of the first obstacles to be overcome was to find a suitable moderator for slowing down the neutrons. To serve as a neutron-moderator a substance has to be:
• of light atomic weight (to produce elastic collisions),
• it must not absorb too many neutrons,
• it must be a readily available substance, and
• it must not be too difficult to handle.
Drs. Szilard and Fermi came to the conclusion that graphite, the soft carbon used in lead pencils, best met all the requirements.
As Professor Smyth points out, it had occurred to a number of physicists that it might be possible to mix uranium with a moderator in such a way that the high-speed neutrons produced by fission, after being ejected from the uranium and before re-encountering other uranium nuclei, would have their speeds reduced below the speeds at which capture by the cannibal in U.238 is highly probable.
But while the general scheme of using a moderator mixed with the uranium was pretty obvious, Drs. Szilard and Fermi worked out a specific manner of using such a moderator, which laid the foundation for the gigantic atomic power plants, or “piles,” later erected at the Hanford Engineer Works near Pasco, Washington.
It was their idea to build a gigantic lattice in which large lumps of natural uranium are embedded in a matrix of graphite as a moderator of the neutrons. This, with some modifications, is the basic design later utilized so successfully in the atomic piles producing atomic bomb material.
Such a pile consists of large blocks of the purest graphite ever made, piled up in a structure forming an enormous cube. Channels spaced at definite intervals in these graphite blocks, which give it the appearance of a giant honeycomb, are filled with lumps of uranium metal.
Enormous volumes of water circulating through the channels carry away the tremendous quantities of heat in which the energy liberated by the split atoms of U.235 manifests itself.
In such a structure, if the proper dimensions are attained, the chain reaction starts in the following manner:
• Some stray neutrons from within or without the pile split some atoms of the U.235 in the mixture.
• The split atoms liberate from one to three neutrons each, some of which are swallowed by the U.238 while others pass through the graphite, which slows them down.
• These slow neutrons, in turn, split other U.235 atoms, which liberate more neutrons, which split more atoms.
The geometrical arrangement of the uranium-graphite lattice is such that for every hundred neutrons that go into splitting U.235 atoms, a little more than a hundred fission-producing neutrons are emitted.
The ratio of the number of fission-producing neutrons in the second generation to the number of fission-producing neutrons in the first generation is known as the multiplication factor, and is designated by the letter K.
If, for example, 100 neutrons that had caused fission in a hundred U.235 atoms produced a brood of new neutrons of which 105 were left to cause fission, a ratio of 105 to 100, then the K factor would have a value of 1.05. If this factor is constant, then the third generation of fission-producing neutrons will be 105 multiplied by 1.05, and so on ad infinitum.
When the K factor is greater than one, the pile will be chain-reacting, as the birth rate will be greater than the death rate. On the other hand, if 100 fission-producing neutrons give birth to only 99, then the K factor will be 0.99—that is, less than one, not enough to maintain a chain reaction.
| || || |
| || |
| || || |
|K factor much greater than one — chain-reaction. Trinity blast.|
When the Columbia group started planning the design of the first experimental pile, only minute amounts of uranium in metallic form were in existence, and no satisfactory method for its large-scale production, particularly in the high degree of purity required, was known.
This was true also of graphite, for while graphite was plentiful, the available product contained too many neutron-absorbing impurities to serve the purpose.
In spite of these seemingly insurmountable obstacles, the Columbia group managed to get together enough crude material to erect their first pile in July 1941. It was a graphite cube of about eight feet on edge, and contained about seven tons of uranium oxide (compound of uranium and oxygen) in iron containers distributed at equal intervals throughout the graphite. Similar structures of somewhat larger size were set up in September and October.
Since the first piles were too small, because of the lack of sufficient material, and the uranium and graphite contained many impurities, they were not expected to produce a chain reaction. Much fundamental information, was gained through them and utilized later in the building of bigger and better piles.
To make up for the shortage of neutrons, and for making comparative studies of the number of neutrons captured by the uranium and the impurities, as well as of the new neutrons produced through fission, the Columbia group placed near the bottom of the uranium-graphite lattice an artificial neutron source, consisting of a combination of radium and beryllium. The neutrons in this combination are emitted from the nuclei of the beryllium when they are struck by the alpha particles that are constantly being ejected from the radium nuclei.
In this type of structure, known as an “exponential pile,” the neutrons are counted at various points throughout the lattice when no uranium is present in the graphite matrix, and the results are then compared with the number of neutrons emitted at the same points after the uranium has been put in place.
The absorption of neutrons by the cannibal in the U.238 would, of course, tend to decrease their numbers, whereas the liberation of neutrons in the splitting of the U.235 atoms would tend to increase them. The problem is to determine which of these two opposing processes predominates.
For more than a year preceding the building of the first exponential pile, physicists at Columbia and at Princeton had been developing highly ingenious techniques for detecting neutrons and for measuring the amounts of their absorption by graphite and uranium oxide. By these methods they were enabled to detect and count not only the total number of neutrons emitted by a given system, but also to screen them in such a way as to count the neutrons according to their energy levels.
In other words, they could distinguish between neutrons of high, intermediate, and low speeds.
In these studies it was found, for example, that the fast neutrons emitted in the process of fission, traveling at speeds of thousands of miles per second, are slowed down to the speed of thermal neutrons, going at only one mile per second, after being made to pass through forty centimeters (sixteen inches) of graphite.
Since neutrons at this speed are the most efficient for producing fission, this showed that the uranium-oxide containers should be placed in the graphite matrix sixteen inches apart. In other words, it provided a blueprint for the geometrical pattern of the lattice.
By these techniques it was also possible to obtain approximations of the total number of neutrons that would escape through the walls of a graphite block of given dimensions, how many would be absorbed by the graphite itself, and how many neutrons of intermediate energies would be captured by the U.238 cannibal.
Since these absorptions vary with the speed of the neutrons, each element absorbing neutrons of only specific energy ranges, measurements also had to be made to determine the various energy levels of the neutrons absorbed in the uranium-graphite system.
To provide a large source of neutrons for these experiments, protons—namely, nuclei of hydrogen atoms—were accelerated by means of a cyclotron and made to impinge upon a beryllium target. The neutrons thus liberated from the beryllium were equivalent to the yield of a beryllium target bombarded by the alpha particles from 3,500 grams (nearly eight pounds) of radium.
This is greater by far than the world’s total supply of radium. At the pre-war price of $25,000 per gram, such a quantity would be worth $87,500,000.
| || || |
| || |
| || || |
|Dr. Ernest Lawrence (right) with a larger late-1930’s cyclotron.|
Most significant of all, by these techniques the Columbia and Princeton physicists (the latter including Drs. Wigner and Wheeler) were able to determine the number of neutrons emitted for each slow neutron entering the U.235. It was found that there was a little cannibal even in the U.235, so that not every neutron that enters it causes fission.
Nevertheless, the studies showed that the decrease was not great enough to make a chain reaction impossible. It largely depended on how big was the appetite of the cannibal in the U.238.
The data obtained gave reason to hope that if he devoured not much more than half of the average number of neutrons disgorged by the U.235, the balance would be great enough to maintain a chain reaction.
The problem was outlined by Professor Fermi in a lecture before the American Institute of Electrical Engineers in New York on January 24, 1940.
“In order that a chain reaction might occur,” Professor Fermi said at that time, “it is obviously necessary not only to have more than one neutron produced for every neutron that is absorbed in the fission, but also to be able to utilize for producing new fissions a large fraction of the neutrons produced; otherwise the loss might be larger than the gain. Assuming that two neutrons are produced in every fission, it is evident that for the chain reaction to take place more than one half of the neutrons produced must be used in new fission processes.”
Actually, Dr. Fermi said, “experiment shows that probably the average number of neutrons emitted is somewhat larger, between two and three.”
Interpretations of the experimental data obtained from the exponential pile, involving many corrections, calculations, and approximations, gave a discouraging outlook. With the impure materials at hand, these data showed, a pile of infinite dimensions—that is, one from which no neutrons leaked away through its sides—would have a neutron birth rate of only 87 per 100. That meant a multiplication factor of only 0.87, not enough to maintain a chain reaction.
It was universally agreed that an increase in the purity of the materials, improvements in the lattice arrangement, and other factors would almost certainly lead to an increase in the multiplication factor.
But even if limitless amounts of pure metallic uranium and the purest of graphite were suddenly made available, no one at that time could state whether a multiplication factor greater than one would be achieved.
There were many other discouraging factors. A number of fires broke out during the experiments, endangering the lives of the experimenters and necessitating fresh starts. In October 1941 a sphere of seventeen kilograms of powdered metal blew up in Dr. Zinn’s hands, and he was severely burned. He was forced to spend three months in a hospital.
But the experiments continued. The scientists began referring to the elusive multiplication factor greater than one as “the Great God K.” If “the Great God K” existed, he had certainly managed to hide himself very successfully. Was he a reality or only a tantalizing myth? Only the atoms knew and so far they had refused to tell.
But unless proof of his existence could be found, there could be no atomic bomb and the vision of atomic power would be but a delusion.
“The sheer cussedness of nature!” Professor Fermi sighed.