Energy and Technology Review

multipass, isotopic stripping towers. As the water is alternately heated in the presence of hydrogen sulfide, a relatively small portion (about 20%) is enriched in deuterium, the remainder becoming slightly depleted. This process is practical from natural deuterium concentration 0.015% deuterium) as electrolysis vacuum distillation, 99+% pur ty required for nuclear reactors. both a cheap and reliable source of electrical power as compared with any technology - nuclear, fossil-fueled, or unconventional - yet proposed. This potential strongly suggests the need to reevaluate heavy water reactor technology. It also provides impetus for the present L1X studies. The U.S. and all mankind may benefit substantially from an accelerated introduction of this economical, reliable, and safe technology for nuclear power generation. sperm orientation. A cost analysis of six potential corrective schemes is now under way. Further Studies, Once we are able to identify and isolate abnormal sperm, we will be looking to answer a number of questions about the effects of energy-related chemical agents on man. Are sperm abnormalities temporary, disappearing once exposure to the mutagen is ended? Or are they permanent? Can sperm abnormalities caused by mutagenic compounds become hereditary traits in subsequent generations? Because the flow microfluorometer allows cell analysis at extremely high rates with statistical precision and sensitivity, it will be an important tool for finding answers to these and other questions about mutagenesis.

reliability for central-station power generation that is, as Hans Bethe recently noted, unrivaled by any other central-station technology, either fossil-fueled or nuclear. These reactors use natural uranium rather than enriched uranium as the nuclear fuel and thus are not tied to enriched-uranium production facilities. Their "special material" requirement -heavy (or deuteriumenriched) water -is a much more readily available commodity, in which a large and multisource interna tional market exists. This is viewed as a significant advantage by many foreign buyers unwilling to commit themselves to any major enriched-uranium supplier (such as the U.S. or ;he Soviet Union) over the 20-to 30-year lifetime of a light water reactor.
The two chief drawbacks to heavy water reactors, until recently, have been the sizeable startup costs associated with their large hea.y water inventory (about 2 kg per kWc of plant capacity in firstgeneration systems) ana their seeming inappropriatehess as breeders. At the cur.ent U.S. price of SI25-l35/kg for heavy water, tht capital costs for the initial inventory are indeed huge, being comparable to those of the associated :learn turbogenerator system. But Canada, whose annual heavy water production is now some eight times greater than ours, estimates its present production costs in the range S7S-100/kg. This reduction, coupled with continuing reactor design improvements to reduce the heavy water inventory, may give Canadian heavy watei reactors a substantial economic edge over Undeveloped light water reactors. Also, it is possible for heavy water reactors to function as true "thermal breeders," producing more fuel than they burn. Their breeding ratios will be somewhat less than those of fast breeders (which use suprathermal neutrons), but they can achieve 0.9-1.1 running on the thorium-233 U cycle, compared with an estimated 1.05-1.25 for fast breeders. Such thermal-breeder heavy water reactors might be introduced without any essentially new technological developments.
It is now considered probable that world heavy water requirements by 1985 will approach 9G00 tonnes annually, due to Canadian-derived heavy water reactor technology alone. U.S. requirements, for nomeactor uses, are projected at some 1000 tonnes annually for the period 1980-85.
• lsotopicnlly selective single-step photodissociation, which capitalizes on differences in the optical absorptivity o! normal and deuterated species to effect K~ge isoiopic selectiviiiss in photochemical decomposition reactuns.
• fsotopically selective two-step photolysis, which couples isotopically selective vibrational excitation of the desired species with radiative raising of the energy of the excited molecules to near their dissociation edge, from where isotopically selective chemical decomposition occurs.
• Isotopically selective photodesorption, which uses low-energy photons to induce isotopically selective desorption of molecules from transparent surfaces.
• Metallic chromatographic separation, by which deuterium is separated from protium by displacement chromatography in a column packed with an intermetallic compound of, for example, a rare earth and a transition metal.
Theoretically, each of these processes -Ascribed below -offers the potential for a considerable savings in heavy water production costs. Our plan is to continue investigating each until a clear choice can be made of which process to pursue-through pilot-plant construction.
Unless otherwise staled, all percentages in the following discussion, such as deuterium's natural concentration in hydrogen (0.015%), are expressed as mole or atomic percentages.
Isotopically Selective Single-Step Photodissociilion. Isotopically selective photodissociation of simple hydrogen-bearing compounds such as formaldehyde (WHO) offers the prospect of single-step deuterium enrichment from natural concentration up t» jbout 10%. Initial low-level enrichment is the costliest step in the present gaseous separation (GS) process, due to 450 tonnes per year, which soon exceeded our domestic needs. One reason was our impending national decision to emphasize light water over heavy water reactors for civilian power production. This emphasis grew out of our relatively advanced naval reactor technology and what was then a superabundance of U.S. uranium-enrichment capability. As capacity continued to outstrip national needs and the production plants aged, they were phased out in stages. Current production -about 270 tonnes annually -is roughly in balance with our present domestic demand for heavy water.
Meanwhile, another line of civilhn nuclear reactor development was being pursued north of the border. Canada, denied access after the ,.jr to U.S. uranium-enrichment technology, determined to develop natural-uranium power plants using heavy water as the moderator and coolant. Also, there may have been reluctance to commit the huge sums needed for building and operating gaseous-diffusion uranium-enrichment plants. The Canadian program received substantial impetus from the advent of the GS process and, in the I960's, they launched an ambitious program of heavy water production to support their power-reactor technology, the CANDU (Canadian deuterium-uranium) system. Indeed, the U.S.' chief customer for surplus and stockpiled heavy water during the late 1%0's and early 1970's was the Canadian nuclear-power-reactor complex, which bought extensively for the purpose of commissioning CANDU reactors for their internal use and foreign sales.
Thai 20 to 30% of the cost of electrical power from CANDU power plants is due to capital charges for their large heavy water inventory has stimulated the Canadians to seek improvements in heavy water production technology. The bulk of this c.fort has gone into the GS process, but Canada is also supporting the world's largest program to explore other processes. Intensive efforts to commission large heavy water production plants have now culminated in a production capability of about 1S50 tonnes annually. Accelerating demand is so great that Canada plans to produce some 35 000 tonnes by 1990 and over 90 000 tonnes by the end of the century.
Their heavy water plants initially were located in the more northerly parts of the nation, where the Canadians cleverly exploited a little-appreciated feature of isotopic geochemistry. Air masses containing water evaporated from oceans in the tropics and advected to the northerly latitudes by planetary circulation become successively drier due to precipitation. At the same time, the fraction remaining in the atmosphere becomes enriched in deuterium to as much as 0.023% at high latitudes. The corollary, of course, is relatively "depleted" rain over most of the US. Thus the early Canadian plants could operate with feedstocks as much as 40% richer in deuterium than those, for example, in the southern U.S. This difference is significant because most of the costs for heavy water production by the GS process are associated with the multiple passes of large volumes of feedstock needed to achieve an enrichment approaching 1% deuterium. the several steps of reprocessing large volumes of water by Yeung and Moore, 1 who separated HCHO from an feedstock associated with this technology (see the equimolar mixture of HCHO and DCDO. The scientific historical overview). The chief problem in the GS feasibility of deuterium separation by selective process is relatively poor single-step isotopic selectivity. photodecomposition of HCDO, the naturally occurring By contrast, the isotopically selective photodissocia-deuterium-bearing form of formaldehyde, was recently tion process ( Fig. 1) allows the relatively large differ-demonstrated at this Laboratory. We used light from ences in optical absorption between normal and a helium-cadmium laser at 325.03 nm to irradiate deuterated species to effect large isotopic selec-HCHO-HCDO mixtures at room temperature and tivities by adsorbing single photons at specific wave-530-Pa total pressure. Analysis of the resulting lengths in the range (for HCHO) of 300 to 355 nm. hydrogen gas showed a net fifteenfold deuterium The photoactivated molecule then dissociates into enrichment. Irradiation of formaldehyde with natural CO and H2 gas. Such optical selection of deuterium-deuterium concentration resulted in >ydrogen gas bearing species can be made to occur with an containing 0.5% HD, a suitable level for subsequent isotopic selectivity ratio approaching 10000 (ten-enrichment by other inexpensive, proved technologies thousandfold better than the normal occurrence of such as vacuum distillation. I part in 6000). Separation of the selected species The maximum expected isotopic selectivity ratio in takes place at near unit quantum efficiency by the these experiments, based on the relative optical automatic dissociation process into readily removable absorptions of the two isotopic species at 325.03 nm, HD gas. was about 30. Using this laser at 325.03 nm gave us Isotopically selective photodecomposition of a convenient and stable light source for verifying the formaldehyde was first experimentally demonstrated basic feasibility of the formaldehyde process, but this wavelength is not optimal. The use of frequencytunable lasers in this spectral region should permit the selection of *brmaldehyde-photcdecomposilion fre quencies for which the optical isotonic selectivity ratio exceeds 1000, based on absorption spectral measurements by previous workers. Three areas of this process remain to be investigated. First, selectivity ratios of 1000 or mote from tunable laser sources must be demonstrated, to avoid wasting relatively expensive uv photons on absorption in nondeuterated formaldehyde. One possible approach here is to select a suitable frequency and ".<ne a suitable laser to it, using frequency 'election with ctalons and other laser-tuning elem'nts. A second approach is to put nondeuterated formaldehyde (HCHO) into the laser optical cavity so that lasing can occur only at frequencies at which the HCHO is transparent. The transparency bands in HCHO can then
Isotopicalry .elective single-step photodissociation process for deuterium enrichment from natural concentration (0.015 mole%) to about 10 mote%. This is a continuous flow process using deuterium-depleted formaldehyde to <trip deuterium from a i>edstxeam of natural methane. Laser light of the proper war?:cngth excites the deuterium-betting formaldehyde mctecuns, which dissociate into CO and HD gas. The HD is leadily removable. A small amount (about 0.1%) of makeup forrmidehyde is required. Feefi^tream costs are practically nil because methane depleted of its nahi -d deuterium is suitable for any usual purpose.
be selected at which HCDO is highly absorbing. Both schemes have advantages and drawbacks, but the latter superior with respect to frequency stability presently seems preferable.
The second area requiring investigation is demonstration (1) that formaldehyde can rapidly exchange deuterium with a hydrogencous feedstream such as water, methane, or hydrogen gas, and (2) that, afterwards, the formaldehyde can be readily and cheaply stripped from the feedstrcam. Deuteriumdepleted formaldehyde gas will have to be isotopically regenerated: otherwise. U.S. heavy water production will be tied to formaldehyde-synthesis plants and thus limited to only several hundred tonnes per year. Exploitation of the statistical thermodynamics of the formaidchydc-fcedstream deuterium exchange may en able a factor-of-2 chemical isotonic selection, thereby permitting formaldehyde feedstock with a 0.039! deuterium content to be supplied to the photolytic process. This exchange, and the subsequent efficient stripping of foimaldehydc from the feedstrcam, must be demonstrated and quantitatively evaluated.
The third area is development of tunable uv light sources with overall electricity-to-ligbt energy efficiencies of at least 17c. The present efficiency of tunable dye lasers is about 03%, so substantial progress is needed. From a large-scale industrial perspective, the most attractive scheme is an efficient visible or uv gas laser used directly or to pump a tunable uv dye laser. Substantial research and development need to be directed to these possibilities. Recent progress by other workers in the rare-gas/fluoride laser field, demon strating -V> or better pumping efficiency in the near-uv region, is especially encouraging with respect to pumping the required lasers.
Isotopically Selective Two-Step Photolysis. The photodissociation edges of vibrational^ excited molecules are shifted considerably toward the red relative to those of vibrational ground-state molecules. Indeed, the redward shift in dissociation energy may be a substantial multiple of the total vibrational excitation. This characteristic, plus the large isotopic shifts for vibrational transitions in hydrogeneous molecules corresponding to M-H bond excitation (M being the remainder of the molecule), naturally suggest a two-step photolytic separation process (see rig. 2). The first step is isotopically selective vibrational excitation of the desired species, by lasers or incoherent sources. The second step is radiative raising of the excited molecule to near its dissociation edge: that is, to a predissociation level if the subsequent excitation transfer is minimal, or well beyond the edge if the excitation transfer poses problems.
Lasers are not mandatory for either step inasmuch as relatively broadband radiation (such as filtered blackbody radiation' may be used to effect both levels of excitation. However, to minimize the complexity of the reaction kinetics, lasers may be preferred because of their uniquely high spectral brightness. Tue most obvious application would be to the infrared-excitation step, due to the low quantum energy involved plus the relatively high pumpinft efficiency (greater than 10%) of lasers in this spectral region. Filtered flashlamp or metal-arc sources maybe the most appropriate for the near-uv photodissoctttion step.
We have obtained isotcpic selectivity ratios greater than 1000 in model system calculations for two molecules: hydrogen bromide ?.ni formaldehyde. Using lasers and filtered incoherent sources, we now intend to reduce these model systems to practice and to determine the actual isotopic-separation performance parameters for design anrl pilot-plant construction.
Isotopically Selective Photodesorpu'on. Physical adsorption cf gases with low boiling points occurs at temperatures around 100 K with binding energies of, at most, 8 kJ/mole. This suggests using low-energy photons to induce isotopically selective desorption of molecules from some transparent surface (see Fig. 3). Even 10 000-nm photors, corresponding to low-energy vibrations, carry energies aggregating to a kJ/mole of photons, more than adequate to induce photodesorption. The absorption of a molecule on a surface induces changes in its electric dipole moment, activating previously radiatively nonabsorptive vibrational modes. Thus, by irradiating the system at an adsorption-induced infrared absorption frequency, only the adsorbed molecules containing the hydrogen isotope of interest nui absorb the radiation and be desorbed. This ensures a very "clean" isotopic separation, as very little energy transfers between the isotopic species. Also, vibrational modes of different isotopic species have rat!,er large differences in theii frequencies: for example, H2 absorbs at 2400 nm while HD absorbs at 2750 nm; methane's "breathing mode" for CH4 is 3430 nm, while for CH3D it is 4535 nm. Thus one may use filtered, efficiently generated, blackbody light sources for inducing isotopically selective desorption.
Our experiments on isotopically selective photodesorption thus far look favorable. A Dewar flask, constructed with two sets of double sapphire windows sealed into it 180° apart, held a piece of specially conditioned, porous Vycor glass in a copper block cooled from a liquid-nitrogen reservoir. The irradiation source was a quart/.-iodide movie-camera spotlight shining through a narrow-bandpass infrared filter, with a transmission half-height width from 2733 to 2884 nm and better than 75% transmission at 2778 nm. This source was calibrated with a thermopile power meter. Deuterium analyses were accomplished in a glow-discharge sample cell. We observed relative intensities of the Buhner j3-emission spectra of H and D at 486.13 and 486.00 nm, respectively, with a high-resolution spectrophotometer. This means of analysis yields only ratios of total deuterium to total hydrogen atoms and does not directly provide the ratios of H 2 to HD or HD to D 2 .
For a gas sample adsorbed onto cold Vycor with an initial D/(H+D) ratio of 21% -that is, 63% H,,

Makeup hydrogen bromide
Deuteriumdepleted water Isotopically selective photodesorption b l batch process that effects a veiy "clean" separation of deuterium from light hydrogen (protium). The chamber is cooled to liquid-nitrogen temperature, and natural hydrogen is pumped across porous transparent panels (e.g-, pretreated Vycor glass) irradiated by infrared radiation of the proper wavelength. Bodi deuterium and protium are adsorbed on the panels, > -it the deuterium is promptly and selectively desorbed (evaporated). The process may be continued until one monolayer of H2 gas is adsorbed everywhere on the panels. Then the chamber must be wanned and flushed in preparation for a new cycle.
33% HD, and 4% D2 -initial data indicate about an eightfold deuterium enrichment in the irradiationdesorbed gas with a quantum efficiency of roughly 80%. This high efficiency, meaning that 80 molecules of HD are desorbed from the Vycor and released into the surrounding ' .s envelope for every 100 photons absorbed, is not well understood mechanis tically but is extremely encouraging. It is much higher than needed to stimulate commercial interest in this process.
Present experiments are directed toward demon strating comparable isotopic enrichments and quantum efficiencies in samples with much lower initial deu terium contents. These experiments necessarily use mass spectroscopic analysis, which yields H2/HD/D2 atios. Some decrease in efficiency is expected, due to the greater likelihood of collisional deexcitation in samples with lower initial enrichments. However, the separation process itself, in principle, is independent of the relative isotopic ratios, because there is no overlap of the infrared absorption bands of H. and HD. Thus, higher concentrations of H2 cannot interfere with radiation absorption by HD and will not cause substantial collisional deactivation of radiatively ex cited HD at moderate pressures.
The apparatus for this process consists of a temperature-insulated column containing thin sheets of porous Vycor glass whose adsorption surface is 250 m /g. Over this column flows hydrogen gas previously cooled to liquid-nitrogen temperature (78 K). Through infrared-transparent windows, the Vycor sheets are irradiated with 2778-nm light, the frequency of the HD stretch. By using Iieliun.. for instance, as a continuous carrier gas with a pulsed flow of hydrogen, one effects a photon-mediated separation in which the HD is selectively desorbed faster than the H2, thus exiting the system before the hydrogen. Present cost estimates fur such a system are necessarily crude but strong'y suggests modest capital investment and low operating costs (hydrogen, cooling, and irradiation). Hydrogen costs can be considered negligible because hydrogen depleted of its deuterium can be readily used for other purposes such as ammonia production. Cooling costs likewise are low. Thus the total operating costs seem likely to be small. The modest investment and low operating costs, plus ready availability of prepurified hydrogen on a large scale (about 1800 tonnes of heavy water equivalent annually from U.S. ammo plants alone), all indicate commercial feasibility.
Problems of considerable technical import need to be solved. For example, the nature and magnitude of energy transfer from photoexcitation to vibrational and translational motion of the hydrogen molecule (e.g., V-*T transfer) are not well characterized. Indeed, it has been speculated that a major fraction ot the 43 kJ/mole of the photodesorbed molecule.; may end up as either vibrational or translation^ energy of the molecules leaving the absorption surface.
We need to optimize infrared-light sources for photodesorption. The choice is between filtered blackbody radiation and various laser sources. By using a reflective filter for a blackbody source, which would pass only the desired frequency and reflect the rest back onto the source, one might dramatically increase the efficiency of a blackbody radiation generator as a narrow-band frequency source. However, as noted above, this might not be necessary.

Metallic Chromatographic Separation of Deuterium.
It was experimentally demonstrated some 1S years ago 'hat deuterium can be separated from protium by displacement chromatography in a chromatographic column packed with palladium at room temperature. Palladium adsorbs hydrogen to form palladium hydride, protium forming a stronger hydride bond than deuterium. Separation is effected when a hydrogen-deuterium mixture is passed through the column. Initial deuterium concentrations of less than 2% have been quantitatively separated from protium by this technique. Palladium's hi^ cost has made this process economically impractical for large-scale heavy water production.

Fig. 4. Metallic chromatographic separation is also a batch process but can be made almost continuous in large-scale applications.
It depends for its selectivity on the stronger affinity of protium than deuterium for certain intermetallic compounds such as LaNi$. As shown here: (a) a column packed with the intermetallic compound is filled with an inert gas (e.g., nitrogen); the gas mixture to be separated -natural hydrogen -Is forced in from the bottom until -Jie column is 25 to 50% filled with hydrogen, (b) Then a pusher gas such as deuterium-depleted protium is introduced, pushing the slug of hydrogen up the column. Deuterium, with its weaker affinity for the Litermetallic compound, travels faster than protium and gathers on the leading edge, (c) Finally the hydrogen slug is forced out of the column, with the deuterium-rich leading edge (the leading 1%) emerging first. The remaining hydrogen is then pushed out die column top, and a new cycle begins.
However, workers at Phillips Laboratories recently discovered a group of compounds of the general formula AB 5 -A being a rare earth and B a transition metal -that absorb and desorb up to six hydrogen atoms per molecule of intermetallic compound at room temperature and hydrogen pressures of about 10 to 100 kPa.* The absorption and desorption reactions are rapid and quantitative. If one of these relatively inexpensive compounds can be substituted for palladium, the chromatographic method of hydrogen isotope separation becomes practical.
The method of choice ( Fig. 4) is displacement chromatography rather than the more familiar elution chromatography used in analytical work. In displacement chromatography, the column is filled with an inert gas -one not strongly absorbed by the column -and the gas mixture to be separated is forced in from the bottom until the column is ibout one-quarter to one-half filled with the mixture. Then the mixture is pushed on through the column by a different gas of equal or greater absorptivity (one that moves no more rapidly through the column than the mixture it is pushing or displacing). In hydrogen isotope separation, deuterium-depleted protium can serve as the pusher gas, and the inert gas might then be nitrogen, which would be readily available if the deuterium production unit were operated in conjunction with an ammonia plant.
In a large-scale separation plant, multiple columns would be operated together so that, while some columns were being filled with the hydrogen-deuterium mixture, others would be undergoing column purge. In this manner, batch-process disadvantages could be minimized. Because deuterium comes off the column first, there is no need to carefully flow the mixture completely through the column. After the first 1% or so of the mixture emerges, that remaining in the column can be purged rapidly and the column set up for the next cycle.
Calculations based on LaNi 5 , one of the better hydrogen-binding intermetallic compounds, suggest substantial economies for chromatographic separation procedures using materials of its general properties. U.S. rare earth mining and purifying facilities, now running well below capacity, are capable without expansion of producing the amounts of rare earths needed for large-scale intermetallic-hydride chromat ographic deuterium separation plants, /ery large quantities of prepurified hydrogen feedstock are avail able at virtually no cost from ammonia plants, as noted above.
The chromatographic method is expected to have substantial cost advantages over the water-hydrogen sulfide method now used for heavy water production because it is a single-temperature, single-step process requiring only the comparatively miniscule amounts of energy needed for compressing the hydrogen gas to column-flow conditions.
We have done preliminary experiments to measure the hydrogen and deuterium pressures in equilibrium with various AB 5 hydrides at different temperatures. In addition to LaNij, these tests to date have included many other rare earth-transition metal combinations in varying proportions. Most show deuterium pressures about 20% higher than hydrogen; some exhibit deuterium pressures as much as twice as high. Even 10% pressure ratios are adequate for purposes of chromatographic separation. A 100% ratio implies very small, highly efficient columns. Separation factors are now being measured for the most promising intermetallic compounds.
Experimental separation of deuterium from hydrogen feedstocks of varying isotopic enrichments is currently being demonstrated with early versions of such columns. These efforts will soon culminate in attempts to demonstrate hundredfold single-step enrichment using natural hydrogen as the feedstock. Single-step quantitative separation of deuterium from enriched hydrogen feedstock (5% deuterium) has already been demonstrated using large column loadings and a fast feed rate.
Much work remains to develop a practical deuterium production process. There are other intermetallic compounds of the same general type as LaNi 5 that may have similar hydride-formation properties but considerably cheaper constituents. A systematic study of these compounds is underway to detennine isotopic separation factors, speed of equilibrium attainment, equilibrium pressure ranges, compound stability and useful lifetimes, and costs of compound preparation.

Summary and Conclusions
We believe the above to be promising approaches to substantially reducing the production costs for heavy water. When one couples these cost reductions with existing and anticipated improvements in heavy water reactor technology, it appears that the present large fraction of power-production costs due to the heavy water inventory -some 20 to 30% -can be reduced to less than 10%. If this happens, heavy water reactors will apparently become both a cheap and reliable source of electrical power as compared with any technology -nuclear, fossil-fueled, or unconventional -yet proposed.
This potential strongly suggests the need to reevaluate heavy water reactor technology. It also provides impetus for the present L1X studies. The U.S. and all mankind may benefit substantially from an accelerated introduction of this economical, reliable, and safe technology for nuclear power generation.

SAFETY LIMITS OF HALF-MASK CARTRIDGE RESPIRATORS FOR ORGANIC SOLVENT VAPORS
Our recent studies of the effective service life (safety limits) for typical half-mask cartridge respirators have shown these devices to be unsuitable for certain organic vapors -methanol, methylamine, vinyl chloride, and dkhloromethane -because the effective service life is too short. For these vapors we recommend other forms of protection, such as air-supplied respirators. The experimentally determined service life for many vapors is shorter -sometimes significantly shorter -than predicted by adsorption theory.

Respiratory protective devices have been used for decades to protect workers from toxic gases, vapors, and particulate matter in their working environments.
A common device is the half-mask respirator worn over the nose, mouth, and chin, whose major components are a removable cartridge (filter), through which air is inhaled, and a valve, through which exhaled air is expelled. The usual filter materia) is activated carbon. Such masks are regularly used for potentially hazardous operations associated with normal laboratory routines at LLL and elsewhere.
Recently we began studies of the effective canister service life of these devices for several common organic vapors under a variety of working conditions (ambient vapor concentration, humidity, work rate, etc.). Our tests addressed adsorption characteristics of the cartridges, not possible leakage through mask components. Airstreams containing known vapor concentrations were continuously forced through standard canisters. A flame-ionization detector measured the vapor concentration downstream of the canister. We considered the canister's safety limit to be exceeded when the downstream concentration reached 10%. Thereafter, the concentration climbs rapidly to about 50% before leveling off to a slower rate of increase.
Experimental results '"ere compared in all cases with values calculated from adsorption theory. With four common vapors -methanol, methylamine, dichloromethane, and vinyl chloride -we found the service life to be less than 30 minutes, too short for TESTING EQUIPMENT Concurrent with the studies described in this article, LLL has designed and built much of the cartridge and canister testing equipment for the Testing and Certification Laboratory (TCL) of the National Institute of Occupational Safety and Health. TCL is charged with testing and approving all respiratory protective devices before they are issued to the general public. Testing involves preparation, control, and analysis of contaminant gas mixtures in air. Typical test gases are carbon monoxide, sulfur dioxide, nitrogen dioxide, ammonia, methyl amine, hydrogen chloride, chlorine, and carbon tetrachloride.

Contact Gary O. Nelson (Ext. 3923) for further information on this article.
safe use of a half-mask respirator. For ten of the vapors studied, the actual service lives are shorter than those calculated from theory.

Adsorption Characteristics
The activated carbon in all commercial organicvapor cartridges is a granular solid usually manufactured by heating coconut shells or petroleum. Subsequent heating with steam produces a highly porous surface that collects the vapors of interest. The carbon's adsorptive capacity (total amount of pore filling) is not affected by the breathing rate through the cartridge. The pores fill faster at faster breathing rates, of course, but only to levels determined by other factors: ambient concentration, relative humidity, and gas or vapor properties. Figure 6 shows that the service life is inversely proportional to the inhalation rate -the pores filling faster at heavier work rates -and that the time to equilibrium is a function of the vapor lype. Relative humidity and the ambient concentration are held constant in this figure. Pore filling varies widely for different gases or vapors. The more volatile solvents with lower boiling points are more difficult to adsorb than those of higher molecular weights. The degree of adsorption is a complicated function of the nature of the carbon and the vapor and cannot be assigned to a single physical property. Table 1  Typical examples of actual and calculated service Life times are given in Table 2. Note that the more volatile materials show the widest deviation from theoretical values. _ This is caused in part by competition with water for the available adsorption sites. Adsorption theory does not now adequately correct for the presence of moisture in the air.
The practical consequences are obvious. For example, the safety limit for methanol, widely used as an industrial antifreeze and fuel for stoves and soldering torches, is about three minutes compared with the theoretical eight or rune minutes. Other  From this table, also, one can identify which vapors penetrate the cartridge rapidly. Methanol is die chief culprit. Our studies, at 1000 ppm, represent a concentration that could easily result from a methanol spill in a closed area on a warm day. (The complete volatilization of just one cup -1/5 litre -of methanol in a garage or storage shed could produce this concentration.) Other vapors whose safety limits are less than 30 minutes are methylaminc, dichloromethane, and vinyl chloride. Along with its derivatives, vinyl chloride is used in refrigeration, plastics (such as PVC pipe), and hair sprays. It is also a suspected carcinogen. For such quickly penetrating vapors, we recommend other forms of protection, such as air-supplied respirators.

Summary
Respirator cartridge service life is a complex function of the breathing rate, ambient concentration,

IS
and relative humidity as well as the nature of the adsorbed vapor and characteristics of the activated carbon. Theoretical calculations using the adsorption isotherm, Mecklenburg, and modified Wheeler equations can deviate sharply from observed values, especially for volatile materials at elevated relative humidity. Our testing program has established practical safety limits for half-mask respirators for several organic vapors and proved conclusively that activated carbon does not perform acceptably for certain vapors, either in theory or in practice.

Flow microfluorometry has become a key tool in biomedical research because it allows celi analysis at extremely high rates (100 000 eels per minute) with statistical precision and sensitivity. One application of the LLL tricolor flow microfluoromr.ter is to detect variability in the DNA content mong individual mollusk and mammalian sperm. DNA variability should be a sensitive indicator of mutagenic events in these cells. Ultimately, we hope to identify and assess the effects on man and his environment of the chemical byproducts from various energy-production streams.
Fundamental to many biomedical problems is an understanding of the molecular processes that underlie critical functions of the living cell. Of special concern is the danger of chemical damage to important biological molecules -to DNA, the chemical basis of heredity, or to the large proteins that carry out the cellular iife processes. Cancer production or mutation are dramatic possible outcomes of changes in DNA. We know little about the consequences of subtle changes in cellular function and control processes, but some obvious possibilities ate embryonic or fetal disturbances and reduced life expectancy.

BJcolor Flow Microfluoromeler
A key tool in our cell studies is the flow microfluorometer.* Analysis of cell populations by high-speed flow methods has become widespread in recent yin.
In the bicolor flow microfluoromeler at LLL. cell suspensions, previously stained with an appropriate fluorescent dye, are introduced into a laminar-sheath flow of distilled water or saline solution Conner Barton I.. GteJhill text. 3S&0) for further information on this article.

Key Words: respirators; masks; respirators -testing; respirators -service life.
in a vacuum system (see Fig. 9

). The cells move at constant velocity (about 10 m/s) in a narrow stream (10-iim diam) through an elliptical beam of exciting light. The light source is an argon-ion laser operated at 488 nm with a light-stabilized beam power of 1 W. The laser's elliptical shape is produced by a pair of crossed cylindrical lenses with 25-and 1.6-cm focal lengths. The beam is oriented with its major axis horizontal; cells flow vertically through its center.
When excited by the laser beam, each cell emits fluorescent light whose duration is determined by cell velocity and beam dimension. Typical pulse widths are a few microseconds; lypical pulse rates are 1000/s. These light pulses are collected by a 20X microscope objective whose optic al axis is perpendicular to both the laser beam an'j sample stream. After passing through an optical filter that suppresses the 488-nm light scattered from the laser, the fluorescent light pulses illuminnte the photocathode of a photomultiplier, which converts them into electrical pulses. (A second photomultiplier is available for simultaneously measuring fluorescence in a different spectra! region.) Each electrical pulse is integrated and amplified; the magnitude of the resulting pulse is digitized and stored in the memory of a multichannel pulse-height analyzer. The memory then contains a frequency distribution of the number of cells as a funcUon of fluorescence; this data en be displayed, recorded on magnetic tape, and transmitted to a computer for processing. In one minute, the flow microfluorometer can measure and record the fluorescence of 100000 cells. ;wo studies, we are working with cells or cellular parts that we want to isolate for further analysis (e.g., chemical). In the sperm and cancer studies, we are looking for unknowns -i.e., for parameters of abnormal celts that will enable the flow system to ioentify and sort out these abenants. Our study of sperm DNA content illustrates the potential and difficulties of flow microfluorometry.

Analysis of Sperm DNA Content
Flow microfluorometric analysis of sperm is central to our goal of measuring the variability of DNA content among mammalian testicular (spermatogone) cells and sperm, so that from these measurements we can identify and assets the effects of energy-related agents on man. We want to leam if any chemicals from the various new energy-production processes (such as in situ coal gasification or geothermal resource recovery) change the DNA content of sperm and thus are mutagenic. For example, is H2S, a foul-smelling byproduct of geothermal wells, or ethylene dibromide (CH2BrCH2Br) mutagenic? The latter chemical compound is used a.; a lead substitute in gasoline, as a soil sterilizer, and as a grain fumigant, particularly for wheat. In this last application, it adheres to the stored grain and so is present in many processed foods. We know that high concentrations of the compound cause sperm abnormalities in cattle and sheep, reducing their fertility. The question is whether ethylene dibromide is mutagenic in the low concentrations likely to be ingested by man.

Variability in DNA content should be a sensitive indicator of mutagenic events in spermatogenic cells.
We are now running tests with known mutagens to see how they affect the DNA content of mollusk and mammalian sperm. Our objective is to define the parameters of an abnormal sperm so that we can program the microfluorometer's computer to recognize and sort out these aberrant cells. The differences in DNA content between normal and abnormal sperm, however, are likely to be very small, requiring very high resolution in microfluorometric analysis for detection.
Effects of Sperm Geometry. We compared the fluorescence frequency distributions of acrifiavinestained sperm from three untreated mollusk and six untreated mammalian species/ Because sperm are usually homogeneous in appearance, haploid in CNA content, and lacking in DNA synthesis, they should generate narrow, unimodal DNA distributions with flow-system analysis. As shown in Fig. 11, spherical or cylindrical sperm from the three mollusk species -oyster, octopus, and abalone -produced the expected narrow, symmetric distributions. However, the stained populations of flat sperm from the six mammalian species -hamster, mouse, rabbit, stallion, boar, and bull -produced skewed distributions, with a peak representing lower fluorescence values and a lateral extension to higher \alues (Fig. 12).
Sperm of the mammalian species thus far examined have a common basic morphology. The sperm's profile view varies little from one to another. The head is considerably flattened; in the bull, for example, it is only 0.5 urn thick. The most striking differences involve the plane view of the head; in the bull, boar, rabbit, and stallion, the head is roughly oval and more or les: bilaterally symmetric (see Fig. 13). In the hamster and m jse, the head is hooked and asymmetric.
In  These data suggested that a homogeneous sperm population can yield a broad range of fluorescence values depending on the spatial relationship of each cell to the excitation source and detector. To test this explanation, we devised a simple, two-dimensional mathematical model. We assumed a flat sperm with a rectangular cross section and a width-to-thickness ratio of i0:i. We used the same geometry as in the microfluorometer, i.e., the axes of flow, f'" citing light, and detector were orthogonal. Cell orientation about the flow axis was random. We also assumed cells immersed in water, uniform distribution and excitation of stain, no self-absorption of fluorescent light, and a detector light-collection angle of 30°. For light emitted at angles greater than the critical angle, we accounted for multiple internal reflection (light piping) . Light emitted at angles [est. .:.-.-. the critical angle is transmitted through the interface between the cell and the medium and is refracted away from the normal -i.e., toward the plane of the head. We ignored internal reflection of the reflected component of this light. Based on geometrical optics, our model also ignored possible physical optical effects due to the sperm head's thickness approximating the wavelength of the excitin" light.
Computation of the distribution of fluo. escent light emission as a function of angle showed more light bent toward the plane of the head than normal to its flat surface, resulting in more light collection by the detector when it views the sperm head edgewise. Combining the effects of this angular light distribution with those of the random <-ol| orientation nboul the flow axis and the detector light-collection geometry, we obtained a theoretical fluorescence distribution qualitatively similar to our experimental skewed distribution. That is, the theoretical distribution had a peak due to heads oriented with their flat surfaces toward the detector and a lateral extension toward higher fluorescence values due to heads oriented with their edges toward the detector. On the whole, therefore, the model appeared to be a reasonable approximation of reality and supported our explanation of the skewed distribution.
Further confirmation of this interpretation comes from two other experiments. First, we found that the degree of asymmetry of the DNA distributions was reduced by raising the refractive index of the suspending fluid closer to that of the sperm. And second, we have sorted sperm from various parts of the asymmetric distribution and found that each of these subpopulations regenerates the complete distribution when put back through the How microfluorometer.
If the flow microfluorometer is to be a useful tool in mutagen testing with sperm, the resulting DNA spectra must be well characterized to clearly mark deviations from normalcy. The coefficient of variation for abnormal sperm is likely very small -perhaps only a few percent -and could be masked if the coefficient of variation of the DNA spectra in general is 5% or more. As was shown in Fie. II, the abalone sperm heads yielded a narrow, symmetric fluorescence distribution with a 2% coefficient of variation, suggesting their potential use in mutagen testing. Our recent studies with abalone sperm has reduced this coefficient to an even more promising 1.25%. The asymmetry of the fl-'Mnprm distributions, on the other hand, currently restricts the utility of flow-system analysis of mammalian sperm for studying mutagenesis. It appears, however, that this asymmetry can be eliminated by either controlling sperm orientation with planar flow conditions or sensing and then accounting for sperm orientation. A cost analysis of six potential corrective schemes is now under way.
Further Studies, Once we are able to identify and isolate abnormal sperm, we will be looking to answer a number of questions about the effects of energy-related chemical agents on man. Are sperm abnormalities temporary, disappearing once exposure to the mutagen is ended? Or are they permanent? Can sperm abnormalities caused by mutagenic compounds become hereditary traits in subsequent generations? Because the flow microfluorometer allows cell analysis at extremely high rates with statistical precision and sensitivity, it will be an important tool for finding answers to these and other questions about mutagenesis.

cc NUCLEAR CHEMICAL MINING FOR COPPER RECOVERY -
To meet projected domestic demand for copper, deeper ore deposits will have to be exploited and a technology developed for processing these ores. We have investigated combining nuclear technology and in situ ieaching to recover copper ores from deposits too deep to be mined economically by conventional techniques. The chemical mining concept developed at LLL involves emplacing nuclear explosives below the wnte> table in a deep-lying copper deposit, detonating these explosives to rubblize the rock, and then passing oxygen and water through the rubblized zone to (each out the copper as copper sulfate. The copper sulfate is then brought to the surface where liquid ion exchange and electrowinning are used to remove the copper.
Our technical evaluations of this chemical mining process have been favorable. In 1973, we teamed with the Kennecott Copper Corporation and ERDA to consider the feasibility of commercializing the process. The resulting joint study affirmed the technical feasibility of nuclear L-hernical mining and identified the process uncertainties that must be resolved before its commercial feasibility can be assured.

Contact Clyde J. Sixmore /Ext. 8801) for further information on this article.
To meet the nation's copper needs, lower-grade deposits haw '..een progressively expjnited and massive mining n.ettiods have been developed -for example, the introduction of large earth-moving equipment and improvements in ore processing. Although these deposits and methods appear adequate to meet demand for the next 10 to 15 years, new sources will be needed thereafter. Deeper or even lower-grade copper deposits must then be utilized, and a technology must be developed specifically to process these ores economically. Because of the long lead times to bring a new technology online, new mining techniques must be developed now to meet the projected copper requirements.
Chemical mining is a likely candidate for exploiting deep copper deposits. Over the past five years, we have worked extensively on one particular chemical process: nuclear rubblization with subsequent in situ leaching of copper ores. More recently, we joined with the Kennecott Copper Corporation and ERDA to evaluate the commercial potential of this process. In general, although the process was judged technically feasible, we found several uncertainties that must be resolved before commercialization is economically advisable. We identified the most critical technical issues, however, and outlined an experimental program for investigating them.

COPPER SUPPLY AND DEMAND
Copper is one of several essential metals -crucial to industrial processes -in which the U.S. demand each year exceeds domestic production. Demand is growing and expected to continue so.H The annual shortfall presently is met by imports, but these have been beset in recent years by severe economic and political problems including nationalization of U.S.-owned production facilities. The international climate, plus widely fluctuating copper prices, keen competition for expansion capital, and higher plant and operating costs, have discouraged copper exploration and development.
The U.S. Bureau of Mines, in 1970, estimated our total domestic copper resources to be about 125 million tonnes. Of this, based on 1968 domestic copper prices, 75 million tonnes were considered economically minable. U.S. needs for the period 1968-2000 were projected at 70 to 90 million tonnes, depending on the copper-demand growth rate. The world demand forecast for the same period, exclusive of the U.S., was 230 to 365 million tonnts compared with economically minable reserves of 200 million tonnes (1968 international copper prices). Imports cannot be expected to meet U.S. shortages, over the long range, without driving copper prices to exhorbitant levels.
Thus, concludes the Bureau of Mines study, discoveries of new n-erves and the development of new technologies for economical copper recovery from presently submarginal resources appear necessary to meet forecasted U.S. and worldwide demands for primary copper.

Nuclear Chemical Mining
The LLL concept for nuclear chemical mining calls for copper minerals to be rapidly dissolved in a rubblized ore column by operating the column at elevated temperatures and relatively high oxygen concentrations. First, a large rubble column is created well below the water table. Either conventional mining methods or nuclear explosives can be used, but the conventional approaches are estimated to be more expensive. Once created, the rubble column, or chimney, is then filled with water either artificially or by natural inflow.
When the chimney is full, oxygen is introduced into the bottom at pressures slightly above hydrostatic pressure. While the gas bubbles rise through the chimney, part of the oxygen dissolves and becomes available as an oxidant for the leaching process, as shown in Fig. 14. The rising undissolved oxygen provides a lifting force that induces a circulation cell within the chimney. As the liquids circulate within this cell, they carry the dissolved oxygen to all parts of the chimney. This dissolved oxygen oxidizes the primary sulfide minerals to produce sulfuric acid and heat, which lowers the pH and raises the temperature. The net result is a relatively rapid dissolution of the copper minerals.
Leach solution is continuously pumped to the surface where it is split into two streams. One stream is diverted to the process plant for copper removal. The other is combined with barren electrolyte solution from the plant and reintroduced into the chimney along with added oxygen.
The solution diverted to the process plant is passed through a heat exchanger to lower its temperature to about 50°C. Afterward, a solvent extraction process strips the copper from the cooled leach solution. Electrowinning* is then used to recover copper from the electrolyte generated in the final stage of the solvent extraction process. Although other extraction methods may be feasible, solvent extraction and subsequent electrowinning appear the best.

Feasibility Study
Under ERDA sponsorship, we joined with Kennecott to consider the feasibility of commercializing our nuclear chemical mining process. This joint study focused specifically on the technical, economic, environmental, safety, and administrative considerations that arise when nuclear explosives are used to rubbiize copper-sulfide ores for in situ leaching. A secondary goal was to provide the data for deciding whether an in situ copper-leaching experiment would be mutually advantageous, and, further, to provide detailed recommendations for such an experiment if deemed profitable.
In general, LLL concentrated on rubblization details, fluid circulation, leaching technology, and hydrology. Kennecott studied processing techniques and economics, and ERDA provided input on environmental and exploration concerns.
The first step was to establish a tepresentative site as the basis for a reference case. We chose a hypothetical porphyry copper-ore body suitable to nuclear rubblization. Next, we set initial values for all process parameters based on available field and laboratory data and theoretical assumptions from extrapolations of commercial or analogous technologies. We then developed an operational plan, including schedules for rubblization, leaching, and process-plant operation. Finally, using these schedules, we refined our original assumptions and determined specific values for the reference case; the resulting dimensions are given in Table 3.
It should be emphasized that these are best-judgment values; those for a real site may differ significantly.
Once we arrived at the details of our reference case, we turned to its evaluation. This analysis had three objectives: (1) to estimate process economics, including capital and operating costs as well as return on investment, (2) to perform sensitivity analyses on process variables, and (3) to evaluate the environmental impact of nuclear chemical mining.
Process Economics. Cost estimates were based on data obtained from engineerinr companies, vendors, ERDA, and available literature. The entire mining and processing operation had an investment cost of $113 million and an annual operating cost of £26 million (1974 dollars). It would produce an annual gross income of $60 million. Further analysis yielded a discounted-cash-flow rate of return (DCF-ROR) on investment 18% for the reference case. Table 4 summarizes these economic result Unfortunately, an 18% ROR is not sufficient to justify immediate commercialization of a iugn-risk technology. Subsequent sensitivity analyses also showed that the ROR could vary from 0 to 23%; some of the process uncertainties must therefore be resolved before commercialization is economically advisable. Sensitivity Analyses. To consider uncertainties in all phases of the nuclear mining -.•• ocess, Kennecott made a series of sensitivity analyses. The purpose was to evaluate the economic effects of changes in the main  independent variables, which were grouped into four categories: • Ore-body-related variables: ore grade, pyrite-tochalcopyrite mole ratio, and the COj content.
• Discretionary variables: yield of explosive, number of chimneys per row, size of pumping system for oxygen injection, and chimney life.
• Process variables: particle-size distribution and mixing behavior of the chimney.
• Economic variables: depreciable capital invest ments, investment expense, operating cost, and copper price.
The interaction of variables was not considered nor were they optimized to give the best results. Consequently, the analyses were qualitative rather than quantitative and only indicated trends.
Kennecott's analyses revealed that changes in most of the listed variables had little economic impact. However, a few of them -variables related to the amount of oxygen required -were extremely important. Certain parameters thus assumed increased significance: the effect of particle size on the pressure drop is crucial, the overall particle-size distribution must be clearly defined, oxygen must be efficiently transferred to all parts of the chimney, and the pyrite-to-chalcopyrite ratio must be known.
Environmental Impact. Finally, the environmental impact of the nuclear mining process was evaluated.
Ground motions, industrial-waste control procedures, radioactivity, plant safety, and plans for a decontamination plant were considered.
We found that ground motion and industrial waste problems can be managed with reasonable effort. Also, with proper controls and procedures, radioactivity is not expected to pose a significant environmental problem. All radioactive waste products will be returned to the chimneys. However, radioactivity will be encountered in all phases of plant operation, and safety procedures must ensure that exposures are below established standards for both workers and the public. Radioactivity levels are expected to be high enough in the liquid-ion-exchange and elcctrowinning plants to require careful study.
A major environmental concern was the possibility of radioactive solutions moving away from the rubblized ore body and into the biosphere. Accordingly, we closely evaluated the hydrologic condition expected around a nuclear chimney. Our analysis indicated that before the fluids can move a significant distance from the chimney, the radioactive components will have decayed to negligible levels, and the surrounding rock will have neutralized the fluids' acidity.
Furthermore, even if a prospective ore body has a hydrology problem -for instance, a large permeable fault in the ore zone -there are remedies. One of the more promising is to drill water wells in appropriate places and, by pumping in water, change the local hydrologic gradient to produce a stationary water zone. This again would restrict the contents of the chimney to the vicinity of the ore body. Costs Tor such remedial steps were included in our economic analysis.
On a more positive note, we should point out that the LLL nuclear mining process also offers environmental pluses. The process eliminates a number of major environmental problems associated with most conventional copper recovery techniques. Open pits, SOj emissions, waste dumps, and tailing ponds, for example, are all avoided.

Study Results
Our joint study affirmed that the LLL concept lor nuclear rubblizalion and IN situ leaching of copper ores is technically feasible. It also identified the most critical technical issues that must still be resolved.
The amount of oxygen needed to stir the liquids in the chimney is a major uncertainly. Process economics are so sensitive to oxygen requirements that this uncertainty must be resolved before further commercialization work is undertaken. Excess oxygen provides the fluid circulation needed to carry oxygen to all parts of the chimney. This circulation is generated when the undissolved oxygen, acting as extra void space, lowers the liquid density and thus creates a buoyancy force that drives a circulation cell within the chimney. In this circulation model, flow resistance or frictional drag encountered in a packed bed, or chimney, is the critical parameter. The key issue in establishing the total oxygen tequirement, therefore, is this resistance, which is a function of effective average particle diameter, rubble porosity, and particle shape factor.
A second uncertainty is the movement of radioactivity through all phases of the copper recovery process. Radioactivity levels in the leach solution and chimney gases have been studied in some detail; we used the results of these studies to plan surface and subsurface mining operations. However, solvent extraction and electrowinning processes need similar study before safety and operating procedures can br defined for the complete recovery operation. Rigorous documentation of radioactivity levels is also needed to expedite the licensing procedure. Licensing, an important phase of commercialization, may take time; it cannot begin until radioactivity data are complete.
The question of how many ore bodies are compatible with LLL's recovery process is another unknown. Although geological evidence suggests many large, deep copper-ore deposits, the information needed to verify this contention resides in company files, is purely speculative, or has not yet been systematically sought. Many prospective sites remain unexplored because there have been no economic incentives or guidelines for exploration at the depths suited to nuclear rubblization. Documentation of these sites is needed to put our recovery process in perspective.
Given favorable resolution of the above uncertainties, IK next slep would be tu verify the leaching rales and conditions for a large rubble column. A demonstration experiment using chemical explosives in a porphyry copper-ore deposit could be a useful intermediate step in the scale-up from laboratory autoclave work to nuclear rubblization. Data on particlc-sizc distribution, frictional drag resistance in a rubble column, and leaching rates for large rubble volumes could be thus obtained with less time and money than in a nuclear cxp'riment.
Only a full-scale experiment, however, will prove the feasibility of nuclear rubblization: nuclear detonation in a copper-ore body at the proper depth. Any full-scale experiment should be conducted according to the conditions given in our reference case. After detonation, the chimney parameters, diameter, and height would be verified first. Before leaching began, rubble zone characteristics would also be measured -for example, particle-size distribution and porosity. Finally, leaching should be continued until enough data are taken to verify the predicted leaching curve.
Some important leaching data, however, should be obtained before extensive field tests are under way and could be compiled from work in small laboratory autoclaves. Almost all of today's leaching data are for San Manuel ore, which severely limits the selection of prospective sites for our nuclear mining process. We need data for other ore types, in which mineralization, ore grade, pyrite-to-chalcopyrite mole ratio, or carbonate minerals differ from San Manuel ore.
The two remaining uncertainties are symmetry of collapse and containment. These phenomena are well understood for single detonations but become serious mknowns when closely spaced multiple explosions are planned. Symmetry of collapse is a problem if the chimney of one explosion is markedly changed because of proximity to an existing chimney. For containment, a problem develops if the presence of a nearby chimney decreases the probability of containing radioactivity. Initial study of these uncertainties can be done theoretically. Eventually, however, both will have to be proved with a full-scale experiment incorporating at least two, and perhaps three, nuclear explosions.
There are, therefore, several issues that must be resolved before the commercial potential of nuclear rabbUzation and in sfftr leaching of copper ores can be assured. The most importsnt technical uncertainly is the oxygen injection requirement. The prime commercial uncertainty is market acceptance and government certification of a copper product that contiins trace amounts of radioactivity. The LLL nuclear mining concept has been shown technically feasible, and our joint study with Kennecott and ERDA has identified the next logical steps in developing this concept.