This key idea was not sufficient by itself. Before a workable design could be developed a scheme was needed for generating efficient compression using this energy flux, as was a means for igniting the fuel once it was compressed. Ulam's idea was to use the neutron flux or the hydrodynamic shock wave of the expanding bomb core to achieve compression. Working with Ulam, Teller added additional refinements to this insight during the month of February. Teller's principal contribution during this period was realizing that the thermal radiation flux from the primary was a more promising means of generating the necessary implosive forces. On 9 March 1951, Ulam and Teller jointly wrote a report, _On Heterocatalytic Detonations I. Hydrodynamic Lenses and Radiation Mirrors_, that summarized these ideas.

From this point on Teller increasingly began to claim exclusive credit for the breakthrough, and eventually came to deny that Ulam had made any original or significant contribution.

Later in March Teller added an important additional element to the radiation implosion scheme. Adapting Ulam's idea to use staged implosion to trigger a fission reaction, Teller suggested placing a fissile mass in the center of the fusion fuel. The convergent shock wave would compress this to supercriticality upon arriving at the center, making it act as a "spark plug" to ignite the fusion reaction. This idea is perhaps not strictly necessary, the convergent shock wave will generate very high temperatures in the center any way and might suffice to initiate fusion as it does in modern laboratory inertial confinement fusion experiments.

Since the continuing compression on the fusion fuel would act to confine the fission spark plug, this final combined design concept was termed the "equilibrium thermonuclear". Teller wrote this idea up in a report on 4 April, 1951.

It was only in April 1951 that the necessary physical principles were in hand to allow the development and testing of an actual hydrogen bomb to go forward. More computations were required to design the device than for any other project in human history up to this point (made possible by the recent invention of the programmable computer). The elapsed time from this point until the detonation of the Mike device was less than 19 months, an achievement as remarkable in its own way as the Manhattan Project.

In April 1951 experiments with fusion reactions and atomic bombs were already being prepared by the US as part of the Greenhouse test series, including a test of the idea of fusion boosting. The Greenhouse George test in particular provided a valuable opportunity to evaluate the Teller-Ulam ideas by allowing the observation of radiation effects in heating and compressing (although not imploding) an external mass of fusion fuel.

Since there are several known designs for incorporating fusion reactions into weapons we come to a question that is largely a matter of definition: Which design qualifies as a *true* hydrogen bomb? I will not try to debate this issue here (see Section 11: Questions and Answers), instead I am including descriptions of all of the significant tests that lead to the development and deployment of early thermonuclear weapons.

The tests are listed in chronological order. Each is followed by a brief discussion of its significance to weapons development.

Greenhouse George
Detonated 5/9/51 at 0930 (local time) on a 200 ft tower on Ebireru/Ruby island at Eniwetok atoll.
Total yield: 225 kt
George was a test of a pure fission bomb, and the highest yield bomb tested up to that time. The bomb was a cylindrical implosion U-235 bomb, perhaps based on a design by physicist George Gamow. An experiment called the Cylinder device was piggy-backed on George to test the ignition of a thermonuclear reaction. The cylindrical implosion design allowed the fusion experiment to be heated directly by the pit without the shielding effects of a high explosive layer, and avoided disruption by expanding detonation gases. A deuterium-tritium mixture external to the large fission core was ignited by the thermal flux, and produced detectable fusion neutrons. This was the first ignition of a thermonuclear reaction by an atomic bomb. Rhodes, in _Dark Sun_, estimates the fusion energy yield as 25 kt, although the mass of fuel given ("less than an ounce") is at a factor of 12 too small for this. Other sources simply give the fusion yield as "small". This approach provided no prospect for development into a high yield thermonuclear weapon. This test fortuitously provided useful data for evaluating the Teller-Ulam design which had been devised two months prior.

Greenhouse Item
Detonated 5/25/51 at 0617 (local time) on a 300 ft tower on Engebi/Janet island at Eniwetok atoll.
Total yield: 45.5 kt
First test of a boosted fission device. A deuterium-tritium mixture in the U-235 bomb core boosted fission yield by 100% over its expected unboosted yield. This innovation was eventually incorporated into most or all strategic weapons, but the fusion yield was negligible and overall yield was still limited by the capabilities of fission designs.

Ivy Mike
Detonated 11/1/52 at 0714:59.4 +/- 0.2 sec (local time) at ground level on Elugelab/Flora island at Enewetak atoll.
Total yield: 10.4 megatons.
This was the first test of the Teller-Ulam (or Ulam-Teller) configuration. The Mike device used liquid deuterium as the fusion fuel. It was a massive laboratory apparatus installed on Elugelab Island in the Enewetak Atoll consisting of a cylinder about 20 feet high (more exactly 243.625 inches or 6.19 m), 6 ft 8 in wide, and weighing 164,000 lb (including attached diagnostic instruments); also said to weigh 140,000 lb without "the cryogenic unit" (this may mean the casing by itself). It was housed in an open hanger-like structure 88 ft x 46 ft, and 61 ft high, where assembly started in September of 1952.

The Mike device consisted of a massive steel cylinder with rounded ends, a TX-5 implosion bomb at one end acted as the primary, and a giant stainless steel dewar (thermos) flask holding several hundred liters of liquid deuterium surrounded by a massive natural uranium pusher/tamper constituted the secondary fusion stage (know as the "Sausage").

The welded steel casing was lined with a layer of lead. A layer of polyethylene several centimeters thick was attached to the lead with copper nails. This layer of plastic generated plasma pressure during the implosion.

The Sausage consisted of a triple-walled stainless steel dewar. The inner most wall contained the liquid deuterium. Between this wall and the middle wall was a vacuum to prevent heat conduction. Between the middle wall and the outer wall was another vacuum, and a liquid nitrogen-cooled thermal radiation shield made of copper.

To reduce thermal radiation leakage even further, the uranium pusher (which was oxidized to a purple-black color, making it an excellent thermal radiator) was lined with gold leaf.

Down the axis of the dewar, suspended in the liquid deuterium was a plutonium rod that acted as the "spark plug" to ignite the fusion reaction once the compression shock wave arrived at the center. It did not run the entire length of the dewar, but was supported at each end by axial columns. The spark plug was a boosted fission device, it was hollow and was charged with a few grams of tritium/deuterium gas (which of course liquified once the dewar was charged with liquid deuterium).

The Mike device had a conservative design. The external casing was made of steel and was extraordinarily thick (usually described as "a foot thick", but more likely 10 inches to be consistent with the weight) to maximize the confinement of the radiation induced pressure inside. The interior diameter was thus about 60 inches. A very wide radiation channel was provided around the secondary stage to minimize thermal gradients, and to make success less dependent on sophisticated analysis. Due to the low density of liquid deuterium, and the necessity of thermal insulation, the secondary itself was quite voluminous which, when combined with the wide channel between the secondary and the casing led to the 80 inch diameter. The massive casing accounted for most of Mike's weight (about 85%).

The TX-5 device was an experimental version of the implosion system that was also deployed as the Mk-5 fission bomb. It used a 92 point ignition system, that is, 92 detonators and explosive lenses were used to make the spherical imploding shock wave. This allows the formation of the implosion shock wave with a thinner layer of explosive than earlier designs. The TX-5 was designed to use different fission pits to allow variable yields. The highest reported yield for a TX-5 test was Greenhouse Easy at 47 kt on 20 April 1951, with a 2700 lb device. The smaller mass compared with earlier designs kept the temperature higher and allowed thermal radiation to escape more quickly from the primary, thus enhancing the radiation implosion process. If the Easy configuration was used in Mike, then the secondary fusion/primary yield ratio was 50/1. The deployed Mk-5 had an external diameter of 43.75 inches, the TX-5 would have been substantially smaller since it lacked the Mk-5 bomb casing.

Three fuels were considered for Mike: liquid deuterium, deuterated ammonia (ND3), and lithium deuteride. The reason for choosing liquid deuterium for this test was primarily due to two factors: the physics was simpler to study and analyze, and extensive studies had already been conducted over the previous decade on pure deuterium fuel. The desirability of lithium-6 deuteride as a fuel was known, but sufficient Li-6 could not be produced in time to make the November 1952 target date (in fact construction of the first lithium enrichment plant had just begun at the time of the test).

Liquid deuterium produces energy through four reactions:

1. D + T -> He-4 + n + 17.588 MeV
2. D + D -> He-3 + n + 3.268 MeV
3. D + D -> T + p + 4.03 MeV
4. He-3 + D -> He-4 + p + 18.34 MeV

For Mike to function successfully, densities and temperatures in the secondary sufficient to ignite reactions 2 and 3 were required. This requires densities hundreds of times normal, and temperatures in the tens of millions of degrees K (say, 75 g/cm^3 and 3x10^7 K).

Since the reaction cross section of 1 is some 100 times higher than the combined value of 2 and 3 the tritium is burned as fast as it is produced, contributing most of the energy early in the reaction. Reaction 4, on the other hand, requires temperatures exceeding 200 million K before its cross section becomes large enough to contribute significantly. Whether sufficient temperatures are reached and quantities of He-3 are produced to make 4 a major contributor depends on the combustion efficiency (percentage of fuel burned).

If only reactions 1-3 contribute significantly, corresponding to the combustion of 25% of the deuterium fuel or less, then the energy output is 57 kt/kg. If reaction 4 contributes to the maximum extent, the output is 82.4 kt/kg. The maximum temperature generated by an efficient burn reaches 350 million K.

The fission fraction for Mike was quite high - 77%. The total fusion yield was thus 2.4 megatons, which corresponds to the efficient thermonuclear combustion of 29.1 kg of deuterium (172 liters), or the inefficient combustion of 41.6 kg (249 liters). The total fission yield was 7.9 megatons, the fission of 465 kg of uranium. All but some 50 kt of this was due to fast fission of the uranium secondary stage tamper by fusion neutrons, a 3.3 fold boost.

The amount of deuterium actually present in Mike was no more than 1000 liters, which is the amount of liquid deuterium handled by Operation Ivy. In fact, it was probably substantially less than this since excess LD2 was undoubtedly brought along in case leakage or other losses occurred.

Prior to test, Mike's yield was estimated at 1-10 megatons, with a most likely yield of 5 Mt, but with a remote possibility of yields in the range of 50-90 Mt. The principal uncertainties here would have been the efficiency of the fusion burn, and the efficiency with which the tamper captured neutrons. Both of these factors are strongly influenced by the success of the compression process. The fusion efficiency involved novel and complex physics which could not be calculated reliably even if the degree of compression were known. The physics for determining the efficiency of neutron capture on the other hand were well understood and could be calculated if the conditions could be predicted.

The upper limit estimate provides some insight into the mass of the uranium fusion tamper. Presumably the 90 Mt figure was calculated by assuming complete fusion and fission of all materials in the secondary. If 1000 liters of deuterium were burned with complete efficiency, the yield would be 13.9 Mt. Fission must account for 76.1 Mt, corresponding to a uranium tamper mass of 4475 kg. Lower amounts of deuterium would lead to higher tamper estimates (a ratio of 0.82 kg of U for each liter of LD2).

The detonation of Mike completely obliterated Elugelab, leaving an underwater crater a 6240 feet wide and 164 ft deep in the atoll where an island had once been. Mike created a fireball 3 miles wide; the "mushroom" cloud rose to 57,000 ft in 90 seconds, and topped out in 5 minutes at 135,000 ft - the top of the stratosphere- with a stem eight miles across. The cloud eventually spread to 1000 miles wide, with a stem 30 miles across. 80 million tons of soil were lifted into the air by the blast.

TX-16/EC-16
The Mike design was actually converted into a deliverable weapon, demonstrating that lithium deuteride is not essential to making a usable weapon. The weaponized design, designated the TX-16, went into engineering development in June 1952 (5 months before the Ivy Mike test). The design eliminated the cryogenic refrigerator, reduced the weight of the tamper, drastically reduced the dimensions and mass of the casing, used a lighter and less powerful primary, and pared the weight in other areas. The expected yield was reduced to 7 Mt. The device was about 60 inches in diameter, 25 ft long, and weighed 30,000 lb. This weapon design would have been filled with liquid deuterium at a cryogenic filling station before take-off, a reservoir in the weapon held sufficient liquid hydrogen to replace boil-off losses during flight. Components for about five of these bombs were built in late 1953, and had reached deplyment by the time of the Castle tests.

A unit of the TX-16, code named Jughead, was slated for proof test detonation on 22 March 1954 as part of the Castle series, prior to its expected deployment as the EC-16 (Emergency Capability) gravity bomb in May 1954. The excellent results with the solid-fueled Shrimp device in the Castle Bravo test on 1 March(see below) resulted in the cancellation of this test, and then of the entire EC-16 program on 2 April 1954.

Soviet Test: Joe 4/RDS-6s
Detonated: 12 August 1953, on a tower at Semipalatinsk in Kazakhstan
Total yield: 400 kilotons
This was the fifth Soviet trest, and first Soviet test of a weapon with substantial yield enhancement from fusion reactions. This bomb (designated RDS-6s) did not employ the Teller-Ulam configuration, instead it used the "Sloika" design invented by Andrei Sakharov and Vitalii Ginzburg. A sloika is a layered Russian pastry, rather like a napoleon, and has thus been translated as "Layer Cake". The design was first invented in the United States by Edward Teller (who called it "Alarm Clock") but it was not developed into a weapon there.

This design is based on a combination of what Sakharov has called the "First and Second Ideas". The First Idea, developed by Sakharov, calls for using a layer of fusion fuel (deuterium and tritium in his original concept) around a fission primary, with an outermost layer of U-238 acting as a fusion tamper. The U-238 tamper confines the fusion fuel so that the radiation-driven shock wave from the fission core can efficiently compress and heat the fusion fuel to the ignition point, while the low conductivity of the fusion tamper prevents heat loss and at the same time yields addition energy from fast fission by the fusion-generated neutrons. The Second Idea, contributed by Ginzburg used lithium-6 deuteride (with some tritium) as the fusion fuel. Being a solid, this is a convenient material for designing a bomb, and it also produces additional tritium from fission neutrons through the Li-6 + n reaction. This establishes a coupled fission -> fusion -> fission chain reaction in the U-238 tamper, with the fusion fuel acting in effect as a neutron accelerator. Larger bombs can be created by placing additional successive layers of Li-6 D and U around the bomb. The device tested in 1953 probalby had two layers.

A small U-235 fission bomb acted as the trigger (about 40 kt). The total yield was 400 kt, and 15-20% of the energy was released by fusion, and 90% due directly or indirectly to the fusion reaction.

A few weeks before the test it was belatedly realized that despite the sparse population of the area around Semipalatinsk, a serious fallout hazard nonetheless existed for tens of thousands of people. The options were to carry out a mass evacuation or delay the test until an air-dropped system could be arranged, which would take at least six months. Rather than delay the test, a hasty evacuation was conducted. [Note: This implies that the Layer Cake was not available as a usable weapon until after Feb. 1954, a time at which the US had actually deployed the EC-14, a megaton-range lithium deuteride fueled Teller-Ulam design. See the Castle Union test below.]

Castle Bravo
Detonated: 1 March 1954 (0645 local time) on reef 2950 ft off of Nam/Charlie island, Bikini Atoll
Total yield: 15 megatons
The Shrimp device detonated in the Bravo test was the first test of a Teller-Ulam configuration bomb fueled with lithium deuteride. This became the standard design for all subsequent hydrogen bombs (including Soviet designs). Shrimp was a cylinder 179.5 in long, and 53.9 in wide, weighing 23,500 lb. The lithium in Shrimp was enriched to a level of 40% Li-6. The predicted yield of this device was only 6 Mt (range 4-6 Mt), but the production of unexpectedly large amounts of tritium through the fast neutron fission of Li-7 boosted the yield to 250% of the predicted value, making it the largest bomb ever tested by the US (and destroying much of the measuring equipment). The fission yield was 10 Mt, the fusion yield was 5 Mt for a fusion fraction of 33%.

The explosion created a 6000 ft crater, 240 ft deep in the atoll reef. The cloud top rose to 114,000 ft.

The Bravo test created the worst radiological disaster in US history. Due to failure to postpone the test following unfavorable changes in the weather, combined with the unexpectedly high yield, the Marshallese Islanders on Rongerik, Rongelap, Ailinginae, and Utirik atolls were blanketed with the fallout plume. They were evacuated on March 3 but 64 Marshallese received doses of 175 R. In addition, the Japanese fishing vessel Daigo Fukuryu Maru (Fifth Lucky Dragon) was also heavily contaminated, with the 23 crewmen received exposures of 300 R (one later died from complications). The entire Bikini Atoll was contaminated to varying degrees, and many operation Castle personnel were subsequently over-exposed as a result. After this test the exclusion zone around the Castle tests was increased to 570,000 square miles, a circle 850 miles across (for comparison this is equal to about 1% of the entire Earth's land area).

The two stage device Shrimp design was used as the basis for the Mk-21 bomb. The weaponization effort began on 26 March, only three weeks after Bravo. By mid April the military characteristics were defined. On 1 July an expedited schedule for deployment was approved. The use of the final fast fission stage was apparently eliminated. After a number of efforts to reduce the weight, the design seems to have stabilized in mid-July 1955 with a projected yield of 4 megatons (subsequently tested at 4.5 megatons in Redwing Navajo, 95% fusion, 11 July 1956). Quantity production began in December 1955 and ended in July 1956 with 275 units being produced. The Mk 21 weighed about 15,000 lb; it was 12.5 ft long, and 56 in. in diameter. During June-November 1957 it was converted to the Mk 36 design.

Castle Romeo
Detonated 27 March 1954 on barge in Bikini atoll lagoon near Bravo test site at 0630:00.4 (local time).
Total yield: 11 megatons
The Runt I device (the second in the Castle series) was another solid fueled two stage design. This device was 224.9 in. long, 61.4 in. in diameter, and weighed 39,600 lb. The fuel for Runt was natural lithium deuteride, a major advantage considering the high cost of lithium-6 enrichment. It exceeded its predicted yield by an even larger margin than Bravo, with a most probable yield of 4 Mt out of a 1.5-7 Mt range. This is consistent with the higher proportion of Li-7, compared to Bravo. The fission yield was 7 Mt, for a fusion fraction of 36%.

The Runt I and Runt II devices (seen Castle Yankee below) were design tests for the EC-17 and EC-24 bombs respectively. These two weapons were very similar (externally identical, similar internal configurations, but with different primaries). They were the most powerful weapons ever built by the US, with predicted yields of 15-20 megatons, and were also the largest and heaviest bombs ever deployed by the US The Mk 17/24 (as the deployed versions were eventually designated) was 24 ft. 8 in. long, with a 61.4 in. diameter, and a weight of 41,400-42,000 lb (30,000 lb of this was the 3.5 in. steel casing).

Although the initial work on these weapons dates at least to Feb. 1953, they went into development engineering in Oct. 1953. The EC-17 and EC-24 became the second and third models of hydrogen bomb to enter the US arsenal. From April to September in 1954 EC-17 and EC-24 bombs were stockpiled (5 EC-17, and 10 EC-24). These bombs were removed in October, modified for better safety features and with drogue parachutes for slower fall, and returned to duty as the Mk 17 Mod 0 and Mk 24 Mod 0 in November 1954. These weapons went through two subsequent modifications, and stockpiles reached 200 Mk 17s and 105 Mk 24s during the October 1954 - November 1955 production run. The Mk 24s were retired in Sept-Oct 1956; the Mk 17s were retired between Nov. 1956 and Aug. 1957.

Castle Union
Detonated 26 March 1954 (0610:00.7 local time) on barge in Bikini atoll lagoon off Yurochi Island.
Total yield: 6.9 megatons
This was actually the fourth test in the Bravo series (the third test - Koon - failed when the fusion stage did not ignite). This was the test of the EC-14 Alarm Clock bomb (unrelated to Teller's earlier Alarm Clock concept), which was the first hydrogen bomb actually to enter the US arsenal and the first Teller-Ulam bomb ever to be deployed anywhere. This was a solid fueled two stage device using 95% enriched lithium-6. It also exceeded expectations, the predicted yield was 3-4 Mt (range 1-6 Mt). The fission yield was 5 Mt for a fusion fraction of 28%. The tested device had a length of 151 in., a diameter of 61.4 in,. and weighed 27,700 lb.

The TX-14 Alarm Clock went into development engineering in August 1952, and procurement was approved in mid-September (some 6 weeks before Mike had even been tested). The first EC-14 weapons were produced in Feb. 1954, two months prior to test of the design. The design was simple but had very poor safety features. A total of 5 were deployed, this low figure can probably be attributed to scarcity of Li-6 at the time. Safety could presumably have been improved through retrofitting, but the high cost of these weapons probably led to their rapid retirement. They were removed from the arsenal in October with the deployment of the EC-17. The Mk 14 (its final deployed designation) had a diameter of 61.4 in., a length of 18 ft. 6 in., and weighed 28,954 lbs. After refitting with a drogue parachute its weight increased to 29,851 lbs.

Castle Yankee
Detonated 5 May 1954 (0610:00.1 local time) on barge in Bikini atoll lagoon, above the Union crater.
Total yield: 13.5 megatons
The Runt II device was very similar to Runt I, mostly differing in the design of the primaries. The fuel for Runt II was also natural lithium deuteride. It also exceeded its predicted yield, with a most probable yield of 8 Mt out of a 6-10 Mt range. See Castle Union for a discussion of weapons derived from this test. This device was 225 in. long, 61 in. in diameter, and weighed 39,600 lb. The fission yield was 7 Mt for a fusion fraction of 48%.

Soviet Test No. 19 Test 11/22/55 (No common name) Detonated 11/22/55
Total yield: 1.6 megatons
The first Soviet test of a Teller-Ulam/Sakharov Third Idea bomb. It used radiation implosion to detonate a lithium deuteride fueled capsule. This was the world's first air-dropped fusion bomb test. After this test the Soviet Union used radiation implosion bombs as the basis for their strategic arsenal. Exploded underneath an inversion layer, the refracted shock did unexpected collateral damage, killing three people.

Information provided by: http://www.fas.org