Particle Analysis of Palanquin and Sedan Events

Page 1

Particle Analysis of Palanquin and This is an informal report intended primarily for internal or limited external distribution. The opinions and conclusions stated are those of the author and may or may not be those of the Laboratory. Work performed under the auspices of the U.S. Department of Energy by the Lawrence Livermore National Laboratory under Contract W-7405-ENG-48.

Page 2

This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency Portions of this document may be illegible in electronic image products. Images are produced from the best available

Page 4

UNIVERSITY OF CALIFORNIA Lawrence Radiation Laboratory PARTICLE ANALYSIS OF PALANQUIN AND SEDAN EVENTS

Page 5

PARTICLE ANALYSIS OR PALANQUIN AMD SEDAN EVENTS Lawrence Radiation Laboratory, Livermore, California Samples of the particulate radioactive clouds produced by the PALANQUIN and SEDAN experiments were systematically separated into particle groups according to size and density. The individual par ticle groups were subjected to optical microscopic examination, x-ray diffraction analysis and high resolution gamma spsctrometric analysis, for the purpose of determining the chemical, elemental, and isotopic composition of the radioactive particulates as a function of particle The particulate samples used in this program are residual parts of close-in aerial filter and fallout collections obtained at the time of detonation for the purpose of obtaining diagnostic information by radio chemical analysis. The aerial filter samples were collected on cellu lose fiber filters, the fallout samples on trays. The procedures used to effect the particle separations are described stepwise in the fol (i) ISOLATION OF PARTICLES. Fallout samples are analyzed as received and require no particular isolation step. The particles collected on cellulose fiber filters may be quantitatively recovered by low temper ature ashing in oxygen gas at low pressure which is excited in an R.F. field and passed over the filter paper. The cellulose is converted to carbon dioxide and water while the particles, which are principally inorganic oxides, are unaffected. Recovery is monitored by pre- and post-ash counting and gamma spectroscopy. Recoveries in excess of 95 percent and with no distinguishable differences in gamma spectra are routinely obtained. A commercial Low Temperature Asher manufactured by Tracerlab, Inc. is used. (2) DISPERSION OF PARTICLES. Upon completion of the ashing process and prior to ail subsequent transfer and analytical operations, the particle sample is dispersed in a liquid medium. The principal problem in this part of the analysis has to do with the state of agglomeration of the

Page 6

particles. As they existed in the original cloud, they were subjected to some degree of physical agglomeration. This is evidenced by the observation that extremely fine particles, too small to have fallen by gravitational sedimentation, are found in close-in fallout collections. Unfortunately, the particle sampling process, whether it be aerial filtration or collection on a fallout tray, produces additional agglom eration to an unknown extent. Therefore, for any given particle col lection it is impossible to determine the state of agglomeration which existed in the original cloud. For past events, the only apparent method of resolving this difficulty is to analyze a sequence of collec tions including fallout, early air and late air samples, and in each case to completely disperse the particles. Then by determining the size distribution with time and relating this to what has fallen out, ve may arrive at a size-related removal rate which includes the effect of agglomeration. Complete dispersion as used here is taken to mean separation in a suspension fluid of those particles which were held together by physical forces but not those which, were chemically fused. It has been found by microscopic examination of the dispersed samples that this is accomplished by ultrasonic agitation at a 1 megacycle (3) FAHTICLS SEPARATION ACCORDING TO DENSITY. A continuous linear density gradient ranging from 2.96 to 2.20 grams/cm? is established in a centrifuge tube by using a continuously varying mixture of tetraoromoethane and propyl alcohol. The particle sample, dispersed in propyl alcohol, is floated on top of the liquid column and the tube is centrifuged for a period sufficiently long that the particles reach equilibrium density levels. Based on the location of visible particle bands, the sample is partitioned into separate density fractions by cutting the tube into sections by means of a Beckman Tube Slicer. Individual fractions are filtered onto a .01 micron pore diameter mem brane filter, weighed, and set aside for subsequent analysis. (4) PAFTICL5 SEPARATION ACCORDING TO SFDIMBFTAfION RATE. The rate at which a particle will traverse a column of fluid is given by Stokes Law which may be expressed as follows: of = a "shape" factor, unity for spheres and close to unity for other particles

Page 7

Thus for particles of density 2.5 falling through a fluid of density 0.3 and viscosity 2.0, the time required for a particle to traverse a column 32 cm in height is given by: Thus, a particle 100 microns in diameter would be down in l.l4 minutes; one of 10 microns diameter in ll4 minutes and one of 1 micron diameter in 11430 minutes. This strong dependence of time of arrival on particle size permits an accurate separation of a particle sample into size fractions according to time of arrival at the base of the column. The relationship of the sedimentation.rate in the fluid column to that in air is given by: Vfluid fluid air The experimental data presented in this report were measured in a fluid column but are reported in terms of sedimentation rate in still air expressed in feet per hour. This unit is convenient, since for particle densities like 2.5 g/cn?, the Stokes diameter of the particle in microns is equal to the square root of the sedimentation rate (i.e., a particle diameter of 100 microns corresponds to a sedimentation rate of 10,000 feet per hour in air). The dimensions and general characteristics of the columns used to effect the particle separations are shown schematically in Figure 1. In use, the screen-bottom column is capped at the base and filled with a continuously varying mixture of carbon tetrachloride and propyl alco hol such that the fluid at the base of the column contains 3 percent carbon tetrachloride solution, that at the top 2 percent carbon tetra chloride and there is a continuous linear variation in between. The density gradient thus established completely suppresses convecuive mo tion of the column fluid and permits separation of particles down to about one micron in diameter. (For smaller particles, an analogous centrifugal system may be used.) linen the column is filled to a 32 cm height, the particle sample dispersed in propyl alcohol is floated on top of the fluid column, the ground glass taper joint top is put in place, and the stopcock is closed. The base cap is then removed and the column is supported by surface tension at the screen. A beaker containing a few milliliters of 3 percent carbon tetrachloride in propyl alcohol is positioned at the base of the column so that the two liquid surfaces join and the column is then supported by atmospheric pressure. Particles fall through the screen into the receiving beaker which is removed' at regular time intervals and replaced with a fresh collection vessel in order to obtain the separated particle fractions, lie have standardized our column dimensions, fluid compositions, and beaker removal tirr.es

Page 8

(l, 2, 4, 8, ... 4096 minutes) in order to permit comparison of analyt ical data on equivalent particle groups from different particle samples. The individual sedimentation rate fractions are filtered onto mem brane filters, weighed and set aside for subsequent analysis. (5) AZTALY5IS 0? SgRARATRD PARTICLE FRACTIONS. Both the density sepa ration and sedimentation rate separation procedures are carried out in duplicate. This provides an internal consistency check on the quality of the separations since the weight and gross activity distributions on duplicates must agree, and also provides two sets of particle samples. One set is permanently reserved for high-resolution gamma spectroscopy, the other is committed to destructive analysis. The ?AIAZ79JjTI'I event of April 1965 was a vented underground deto nation of a 4 kt device. It was carried cut in a rhyolite which con sists of a fine-grained ground mass in which large crystals of feldspar are embedded. The composition of the original environment was as and some glass. The fine-grained material is principally feldspar with a few percent of magnetite and hematite. The principal constituent of the PAIAl'IQ.UH'I environment was feldspar whose density is 2.56 g/cmr. Three samples of RALAiltUDT particulate debris were analyzed. A fallout sample, Ko, was obtained about a mile from ground zero and was collected over a period from 6 hours prior to detonation to 24 hours after. Radiochemical analysis of a portion of the Ko sample showed it to be the most representative of all PALAIvuUIK collections from the standpoint of fractionation. The atom ratio of the refractory radio nuclide zirconium-95 to the volatile barium-140 was equal within experi mental error to the ratio in which they are known to be formed in fission. An early air collection, AL. was obtained at 11,000 ft i-ISL over two one-minute intervals at 11 and lo minutes after detonation. Radiochemical analysis of a portion of the AL collection indicated that for the amount 140 present, only 15 percent of the expected zirconium-95 was present in the AL sample.

Page 9

A late air collection, CL, was obtained at 11,000 ft MSL as a series of six two-minute exposures over the interval 6l to 93 minutes after detonation. Radiochemical analysis for barium-140 was,not per formed but the even more volatile cesium-137 was determined, and based on the amount of this isotope which was present, a maximum of 0.4 per cent' of the expected zirconium-95 was present. PALANQUIN DENSITY SEPARATION. All three samples of PALANQUIN debris were subjected to density separation with similar results. In all cases, a small amount of material settled out on the bottom of the centrifuge tube (density greater than 2.96)5 a sharp band of material appeared at the feldspar density of 2.565 and above the feldspar layer a continuous diffuse layer of particles ranging in density from 2.56 down to about 2.2 g/cn? was observed. The density distribution data obtained on the early air sample are shown in fable 1. Density Frac tion 1 probably represents residual organic material from the collection medium. Density Fraction 2, which showed a small percent of feldspar by x-ray analysis, was almost entirely made up of a feldspar glass showing many voids and inclusions. It will be noted that this fraction carried 59 percent of the activity but only 31 percent of the mass. Gamma spectrometric analysis of this fraction indicated it to be heavily enriched in the refractory species and very depleted in volatiles. Den sity Fraction 3 consisted almost entirely of crystalline feldspar. It contained only 34 percent of the activity but 52 percent of the mass, and it was highly enriched in the volatile species. Density Fraction 4 consisted of the heavy minerals. These were crystalline in nature and gamma spectrometric analysis showed them to be highly enriched in the volatile species. Although all fractions contained a range of particle sizes, the larger particles were to be found mostly in Fractions 2 and 4, -while the finer particles were to be found in Fraction 3PALANQUIN SSD242NTATI0N RATS SEPARATION. In Table 2 the weight data obtained from the separated particle fractions are presented to gether with the parameters which define the part le groups, namely the sedimentation rate interval over which the particle fraction was col lected and the corresponding particle size interval. For purposes of graphical representation, the data were first normalized by converting the absolute weights to fractional weights for each analysis. That is, the individual particle sample weights-for AL, for example, were divid ed by the sum of the weights of all particle samples for the AL analy sis. The normalised weight data were then divided by the difference in the logarithm of V, the sedimentation rate. In order to obtain a dis tribution function (or probability density function), the derived hd log V values were plotted against log V with the results shown in .-jure 2. The weight distribution function for each collection cor responds to the particle population as sampled and not (with the

Page 10

possible exception of K6) to the true PALANQUIN particle population. The gross gamma activity data which are given in Table 3 were similarly treated and the gamma distribution functions are shown in Figure 3. The fact that the weight and gamma distribution functions behave differently indicates that the activity is related"to’the particle mass- in a complex way and that the rate at which mass is being removed from the cloud is different than that at which activity is removed. Some further 'insight into what is occurring is obtained by looking at the gamma specific acti vity for the three particle group sets. These data which are given in Table 4 show several interesting things. First of all, there are two distinct regions of high specific ac tivity. One of these is in the very largest particle group of the K6 run, the other is in the smaller size region of the AL and CL runs. This implies that the activity in the largest particles is much more uniformly distributed through the volume of the particle than is the case with the small particles. The rapidly rising specific activity with decreasing size strongly implies surface adsorption. The specific activity data of the very smallest size group should perhaps be dis counted because of the likelihood that any organic residue from the filter ashing process will show up in this fraction. A second factor which appears is the uniformly lower specific activity of the K6 sample in the smallest particle groups. This is undoubtedly a reflection of contamination of the sample collecting tray with non-relevant particulates during the 30-hour exposure period. * A final interesting feature is the generally good agreement in the specific activity data for AL and CL for individual particle samples in the smaller size ranges, he believe that the sedimentation separation procedures used will produce equivalent particle groups regardless of sample source. In the finer particle region, where internal distribu tion within a particle group should not be changing rapidly, it is to be expected that equivalent samples would show similar compositions. This hypothesis is borne out by the gamma spectrometric …