• Military

• WMD Menu                           

• Introduction
• Systems
• Facilities
• Agencies
• Operations
• Countries
• Hot Documents
• News
• Reports
• Policy
• Budget
• Congress
• Links

• Intelligence
• Homeland Security
• Space

Nuclear Effects Testing - Introduction

ANFO is inexpensive, relatively safe, and cost effective in terms of energy output the very reasons that the mining industry uses it in ever increasing amount. It is readily available; basically 'AN' is commercial fertilizer made by industry in millions of pounds per day quantities, and the 'FO' is ordinary #2 diesel oil, easily available throughout the country. Also, the ANFO many have hydrodynamic advantages over the block built TNT charges then being used on mayor military hardware tests. Perhaps detonation front and blast anomalies would be minimized in the ANFO charges as they were shown to be in the gas balloon simulation technique because of the greater homogeneity of the grilled explosive material compared to the discontinuities encountered between TNT blocks.

There were many questions, answers to which were not immediately. available, nor indeed, were they to be found at all later in the literature. Will ANFO detonate reliably with predictable blast output in large unconfined piles? Mining industry experience was limited to ANFO charges heavily confined in relatively small diameter bore holes and most loads were primed with dynamite placed every ten feet along the column length to sustain detonation. Such confinement and multiple and time sequenced detonation points would never do for ANFO charges designed. to produce ideal airblast fields. For one, charge confinement was deemed inappropriate; the confining case could produce fragments that could jeopardize test targets directly, and depending on the size of the fragments, they could produce airblast anomalies through the bow waves associated with the fast moving fragments. For another, single point initiation of the charges (as used for hemispherical. and spherical single charges) was considered necessary to assure a constantly advancing symmetrical detonation front so that the blast field could be predicted reliably.

Would large ANFO charges in the hundreds of tons size be safe - chemically and thermally stable over a period of time, say a month or would them be some reactions taking place generating heal or chemical products which would lead to auto detonation? The Texas City explosion in 1947 was remembered in which. two shiploads 6f ammonium nitrate fertilizer caught fire for some undetermined reason and detonated with disastrous results hundreds killed, thousands injured, and millions of dollars worth of damage.

Would the fuel oil settle out of the ANFO mixture during the days or weeks it may take to construct the charge and fire it? The mining industry seldom, if ever, faced this problem; they would mix the AN and FO on site just before filling the bore holes, or would use premixed ANFO freshly delivered from a nearby manufacturer.

Assuming detonation of a large charge of ANFO, say 500 tons, would its blast output be similar to that of a 500 ton TNT charge? It was realized that the density of bulk ANFO was considerably less than that of TNT and thus, its detonation velocity end pressure would be less. This would mean that the peak airblast output of an ANFO charge should be less than for TNT, but at the pressure levels of vast concern to target studies, about 2000 to 3000 psi and less, would the ANFO blast be adequate? This could not be answered initially.

These questions and answers, pros and cons, doubts and concerns and hopes were batted back and forth by Petes and Sadwin, and a little later, with their colleagues. It was felt that before a real case could be made to the Navy and DNA for support to investigate the merits of the ANFO concept, some basic information on the detonability of unconfined ANFO would be required.

The Air Ground Explosions Division of NSWC was heavily engaged in experimental field work with military high explosives. It was a relatively simple, although unorthodox, matter to introduce several ANFO shots between authorized work at its remote Stump Neck, Maryland Airblast Facility. J. F. Pittman conducted these bootleg tests in mid August 1966 with 8-lb, 20 lb,and 64-1b charges of ANFO contained in thin plastic paint buckets and garbage cans which were considered lo be essentially non confining.

The test results were encouraging. The charges detonated reliably; repetitive shots with the same weight charges produced the same pressure distance curves in the range of pressures measured, from 5 to 100 psi. But the outputs of the different weight charges, as evaluated in terms of TNT equivalence, were different. The 8 lb charges had a TNT equivalence of about 0.47, the 20 lb charges 0.51, and the 64 lb charges about 0.75. This increasing output (or TNT equivalence) with increasing charge weight was interpreted to mean that the critical size the minimum diameter required to attain a steady state detonation velocity through the explosive was greater than that realized in the 'small charges used, hence, the full explosive energy of the charges was not realized. Of course, the 64 ib charges may have ,attained full output, but this could not be established from the data. The hope was that, in fact, full output was not attained, that larger ANFO charges would produce blast output more nearly approaching that of TNT.

Data obtained by R. W. Van Dolah in Bureau of Mines tests investigating the sensitivity of ANFO showed that for 1500 lb charges, the TNT equivalence of ANFO was 1.0 for pressures up to 10 psi. Larger charges than, those fired in the bootleg program would have to be fired to check this out. But larger charges could not be bootlegged; they would have to be tested under some authorized, funded, and planned program. The Navy and DNA were logical places to seek the necessary support; both had compelling reasons for seeking nuclear weapon blast and shock simulation techniques.

The Navy had but recently (1965) fielded at Kahoolawe, Hawaii, Operation SAILOR HAT, a three event airblast program in which 500 tons of TNT were fired on each event. Fully operational combatant ships and naval structures, weapons, and equipment located on a floating platform (a converted light cruiser hull) were subjected to long duration blast waves with amplitudes up to about 10 psi. These tests were designed to establish the vulnerability/ survivability of these navy ships and items and to provide guidelines for hardening in a nuclear weapon blast environment. Much valuable information was obtained on the tests, information not available through analysis or other simulation techniques. The Navy looked forward to doing more such tests, but the costs were staggering; it was estimated that the cost of each 500 ton charge was about $1,000,000. The Navy was in the market for. a cheaper explosion simulation source.

DNA, as a Department of Defense agency, had similar and broader reasons for seeking nuclear weapons blast and shock simulation techniques; the use of nuclear weapons and devices in this atmosphere was prohibited by the first test moratorium of October 1958 and by its successor, the Nuclear Test Ban Treaty of 1963. Yet, the Department of Defense had the continuing requirement to obtain nuclear weapons effects data on military equipment and targets. Techniques to simulate airblast and ground shock of nuclear weapon proportions were obvious alternatives. DNA supported and encouraged the pursuit of many such alternatives.

Initially, interest was in the response of targets primarily to airblast. In 1959, DNA initiated a point program with DRES (Defence Research Establishment, Suffield) Canada, to develop TNT as the explosion source for long duration airblast waves. This program continued, expanded, and accelerated DRES's earlier efforts investigating the properties of TNT in various forms for field applications. Using 12x12x4 inch blocks of TNT weighing 33 lbs each, charges were constructed on the ground in a hemispherical shape with single point initiation occurring at the ground in the center of the equatorial plane. This development culminated in 1964 on Operation SNOWBALL when a 500 ton hemispherical charge was detonated successfully. Additional 500 ton TNT hemispheres were fired on Operation SAILOR HAT in 1965.

As time went on, the interests of the military services extended to subsurface facilities, structures, and targets; ground motion as well as airblast became a matter of concern to the military scientific community. Analyses and progressively larger scaled experimentation showed that a spherical charge built on and tangent to the ground would produce airblast, ground shock, and craterting energies in the .same relative proportions as a surface burst nuclear weapon. This relationship of effects was deemed important so as to best satisfy, on a single shot, the requirements of the military scientists and analysts interested in blast, craters, and ground shock. Block built charges were now stacked in this geometry. A number of such 500 ton charges were detonated on military test operations, namely PRAIRE FLAT (1968), DIAL PACK (1970), and MIXED COMPANY (1972).

As useful as the TNT block built charges were they presented problems. For one, as the Vietnam War continued, TNT was becoming available in ever shorter supply; the World War II surplus was about exhausted and the TNT manufacturing plants were nearing the end of their productive lives. For another, the cost of processing the TNT into 33 lb blocks and placing the charge in the field ready for test was high up to $1,000,000 for a 500 ton charge.

A third problem surfaced early in the use of these large block built charges the airblast front was plagued with large, unpredictable anomalies. These manifest themselves generally as ahead running spikes, jets, and protuberances on the main shock front. Some of these anomalies extended 1,000 ft from the explosion source and perturbed the pressure field within a 30° sector measured from the origin. Three such major anomalies, not uncommon on some of the tests, could adversely influence 25% of the area in which targets were located and thus invalidate expensive and important target response studies.

An extended and detailed study of anomalies was initiated by a working group of American, British, and Canadian scientists under the auspices of TTCP (The Technical Coordinating Program) soon after J. M. Dewey of DRES, in 1965, reported the occurrence of serious blast front perturbations on operation SNOWBALL. The report of the working group, published in 1970, found that anomalies were, in fact, characteristic to explosions of solid high explosives; evidence was found for jets, spikes, and perturbations from charge sizes ranging from gram weights up to the 500 ton charges of immediate interest. They concluded that one of the major reasons for the anomalies arises from the very unstable nature of the detonation process as it progresses from explosive grain to explosive grain on a microscale; this is particularly true for cast TNT with its relatively large and irregular granular structure. In block built TNT charges, the instabilities are accentuated on a larger scale because of the significant reduction in detonation velocity as it progresses across the somewhat irregular interfaces between blocks.

The TTCP working group suggested ways to reduce the number and severity of the anomalies that were explosive dependent but offered little hope for eliminating them so long as the 12x12x4 inch block built construction was used. (All the anomalies identified by the TTCP working group are not explosive oriented; some arise along paths of ground surface discontinuities, e.g., roads and trenches, and occur regardless of the high explosive used.) Significantly, they recommended the study of other explosive materials for nuclear weapon blast simulation detonable gases, slurries, and ANFO.

ANFO for Airblast Tests

Early in November 1966, Petes and Sadwin disclosed their thoughts on the use of ANFO as a nuclear weapon blast simulant explosive to J. Kelso, DNA, and Y. Park, NAVSEC (Naval Ships Engineering Center, then Naval Ships Systems Command). Many discussions were held. All the information in hand on ANFO was presented. and discussed; the merits of ANFO its documented low cost, about five tents per pound delivered to any continental test site; the controlled and repeatable detonability of unconfined ANFO as demonstrated in Pittman's tests; the ready availability of ANFO on the commercial market from dozens of manufacturers; and, the safety and ease of handling as evidenced by its increased acceptance and use (approaching one million tons per year) by the mining and quarrying industries. The unresolved questions and doubts originally and subsequently raised by Petes and Sadwin were reviewed also, so that the technical and financial risks involved in exploring ANFO as a suitable blast simulation source could be put in perspective. Would large charges, 500 tons, detonate reliably? Would self heating of such large charges be a hazard? Would the known hygroscopicity of ANFO preclude its use for Navy purposes in a sea or near sea environment? Would the fuel oil settle out of the ANFO? Questions and more questions which had no ready answers. The apparent merits of ANFO and the enthusiastic and persistent (and, perhaps, overstated) salesmanship of the NSWC personnel outweighed the doubts: Kelso and Park agreed to fund a small experimental study to determine the feasibility of using ANFO. An official proposal was submitted by NSWC to NAVSEC in March 1967; reprogrammed DNA funds were provided via NAVSEC in December 1967.

The principal objectives of this new task were to determine 'the blast yield for hemispherical ANFO charges weighing up to 4000 1bs and, to demonstrate, on this scale, the predictability of the airblast field. Depending on the results of this task, a follow on program would be recommended with the goal of eventually constructing and testing up to 500 ton charges.

At the very start of this project, the fundamental decision was made to use commercially developed and available material, such as Gulf Oil Corporation Spen C N 1 premixed ANFO and N IV prilled AN. The project specifically avoided the attractive research task of developing a new AN based explosive or even exploring the. many AN based explosives described in the literature and patent disclosures. Such an investigation would be time consuming, costly, and probably would have been counter productive by ' compromising the demonstrated virtues of commercial ANFO low cost, safety, and ready availability. It was recognized that additives such as aluminum and TNT would increase the blast output of AN based mixtures but the greater sensitivity of these mixtures militated against advocating their pursuit and use. Water as an additive to make an AN slurry or gel was also discussed as an alternative but the idea was dismissed from further consideration because of the obvious requirement for a case to contain the mixture; casing, particularly heavy casing, could result in deleterious fragment effects on the blast field and the test targets.

The selection of ANFO and AN was made with sane trepidation: safety was of, paramount concern. After all, single quantities of up, to 500 tons of ANFO were envisioned for test purposes; an accidental explosion was unthinkable. Neither industry nor the military had experience with such large quantities. Industry uses ANFO loaded into relatively small diameter (1" to 12") columns set in a pattern of bore holes optimized for rock break up.. Any one hole could hold up to about 1,000 lbs of ANFO. The military has used AM in relatively small; one man, deployable cratering and demolition charges (AN 86:6x, dinitrotoluene 7.6%, non explosive. ingredients 5 8x), and in mixtures with TNT for use as the explosive fill for ammunition. Amatol 80 ZO contained .80x AN and 201; TNT; amatol 50=50 had 50% AN and 50% TNT (It is interesting to note that gust as a driving force for suggesting AN F as a nuclear blast simulation source was the shortage of TNT in 1960 70, the amatols were introduced as a military explosive during World War I in order to reduce the demand for TNT which was then in short supply.)

Although AN is not classified under U.S. Department of. Transportation regulations as an explosive, but rather as an oxidizer, AN can detonate and has detonated with catastrophic results. The most recent large accident was the Texas City, Texas explosion on 16 April 1947. Two freighter loaded with commercial fertilizer grade AN detonated at the pier resulting in 454 deaths, 150 missing, 3,000 injured, and damage estimated at $50,000,000. A fire in one of the ships has been attributed as the cause of the resulting explosion.

Twenty six years prior to the Texas City explosion another large AN accident took place in what, in retrospect, appears like bizarre circumstances. The Badesche Company manufactured AN based fertilizer at its, plant at Oppau, Germany. Large masses of the material were stored outdoors where it was subject to the ravages of weather. The AN would cake because of its hygroscopicity and it could freeze. It had been the standard practice to break up the caked AN with explosives. On 29 September 1921, this procedure was used with disastrous results. An estimated 4,500 tons of AN detonated. More than 1,000 persons were killed; about 76% of .the houses in Oppau were leveled or made uninhabitable, and crater 400 feet in diameter and 90 feet deep was formed. The blast was felt in Munich, 175 miles away. The bizarre feature of this episode is that AN was considered to be so safe that explosives could be used to break it up; even though as early as 1867 the properties of AN as an explosive ingredient and indeed, as an explosive, were recognized in a Swedish patent issued to Ohlsson and Norbein. On second thought, perhaps it is not bizarre; even today industry finds it necessary to provide the following warning with the product, "We also stress, that dynamite or any other explosive must not be used to break up caked ammonite nitrate."

The selected ANFO was available in 50 1b bags ready mixed in the stoichiometric proportions of 94 to 6, by weight, of AN Lo FO respectively; prilled AN, which could be mixed with #2 diesel fuel oil at a test site, was available in bulk quantities and in 50 1b bags. This prilled AN was developed in the 1940's specifically for use as the base material for ANFO explosives. Untreated AN is highly hygroscopic leading to caking and dissolution in humid or wet atmospheres. In fertilizer grade AN, the grills are normally coated with a. diatomaceous earth which inhibits water absorption. In AN designed for ANFO applications, the diatomaceous earth is replaced with a surfactant which, in addition to inhibiting water absorption, permits uniform fuel oil absorption by the grill, resulting in an intimate and homogeneous ANFO mixture.

Prill size or AN bulk density is a parameter which influences ANFO sensitivity and explosive output; the higher the grill density, the lower the sensitivity and the output. The use of high density prills such as used in agricultural grade AN, results in an uneven distribution of fuel oil with most of the oil being concentrated on the surface of the grill rather than being uniformly distributed throughout the grill. This oil rich surface and the oil poor inner grill material upset the stoichiometric balance of the ANFO on a grill scale; this leads to low output. The AN used in both the premixed and field mixed ANFO had a bulk density of 0.88 g/cm3.

This type of prill with a surfactant and a bulk density of about 0.88 g/cm3 was found by industry to be favorable for combining with fuel oil in stoichiometric proportions: This results in a balanced reaction with ail the' reactants being consumed. The reaction 3NH4NO3 + CH2 7H2O + CO2 + 3N2 calls for 5.65% fuel oil. Analyses and experiments indicated that the optimum , or near optimum explosive output of ANFO is . obtained for a fuel oil content of from 5% to 7%.. This leeway in oil content is fortunate because in practice it is difficult to maintain a precise 5.65% oil content without strict quality control. In fact, it is common practice to overfuel, i.e., approach the 7% limit, because the ANFO output is affected less by overfueling than underfueling. This overfueling turns out to be an advantage for large charge simulation preparation; it compensates to some extent for the evaporation losses of FO that occur when the ambient temperature is high.