new measurements of the particle size distribution … · new measurements of the particle size...

NEW MEASUREMENTS OF THE PARTICLE SIZE DISTRIBUTION OF APOLLO 11 LUNAR SOIL10084. D. S. McKay' B. L. Cooper 2, and L. M. Riofrio2 'NASA Johnson Space Center(davidsmckay@nasa coy), 2Oceaneering Space Systems.

Introduction: We have initiated a major new programto determine the grain size distribution of nearly alllunar soils collected in the Apollo program. Followingthe return of Apollo soil and core samples, a number ofinvestigators including our own group performed grainsize distribution studies and published the results [1-11]. Nearly all of these studies were done by sievingthe samples, usually with a working fluid such asFreonTM or water. We have measured the particle sizedistribution of lunar soil 10084,2005 in water, using aMicrotraCTM laser diffraction instrument. Details of ourown sieving technique and protocol (also used in [11]).

are given in [4].While sieving usually produces accurate andreproducible results, it has disadvantages. It is verylabor intensive and requires hours to days to performproperly. Even using automated sieve shakingdevices, four or five days may be needed to sieve eachsample, although multiple sieve stacks increasesproductivity. Second, sieving is subject to loss ofgrains through handling and weighing operations, andthese losses are concentrated in the finest grain sizes.Loss from handling becomes a more acute problemwhen smaller amounts of material are used. While wewere able to quantitatively sieve into 6 or 8 sizefractions using starting soil masses as low as 50mg,attrition and handling problems limit the practicality ofsieving smaller amounts. Third, sieving below 10 or20im is not practical because of the problems of grainloss, and smaller grains sticking to coarser grains.Sieving is completely impractical below about 5-10ini Consequently, sieving gives no information onthe size distribution below -10 kim, which includesthe important submicrometer and nanoparticle sizeranges. Finally, sieving creates a limited number ofsize bins and may therefore miss fine structure of thedistribution which would be revealed by other methodsthat produce many smaller size bins.Because of the known complexity of lunar surface soil-forming processes, particles in the finer size range maybe created and modified by processes totally differentfrom the processes that create coarser particles, andsuch differences may leave a record in the grain sizedistribution. The only data in this finer size range isfrom Coulter counter or optical/electron microscopetechniques [4, 12]. Knowledge of the grain sizedistribution of material finer than about 10 im isnecessary to help evaluate possible exposure healthhazards to humans at a lunar location. Nanometer sizeparticles may cause health issues primarily because oftheir small size, regardless of their composition. These

effects may be particularly pronounced for fresh lunarsoil because the grain surfaces may have beenactivated by UV, solar flare radiation, shock effects,solar and galactic particle etching and structuraleffects, and reactive vapor coatings from sputteringand impacts.

Alternative methods: A number of alternativemethods for determining grain size distribution haveexisted for many years (NBS Special Paper 260-85,1983). Major techniques include those based onoptical and electron microscope imaging, volumedisplacement and electrical sensing flow throughdetectors, light scattering including laser light, laserDoppler techniques based on Brownian motiondetection, sieving, impactor, and sedimentationmethods. Our new data is based upon laser lightscattering.

Instrument: For grain sizes down to about 1micrometer, we use an advanced laser light scatteringinstrument (Microtrac TM). This instrument uses aproprietary modified Mie scattering algorithm toaccount for the irregular shapes of the particles andtheir non-transparent nature. This instrument waschosen because of the demonstrated reproducabilityand accuracy of the method, the relative ease of use,and the ability to analyze a large number of samples ina relatively short time. It is now feasible to analyzemultiple splits from the same sample to allow for anevaluation of the homogeneity of the initial materialand the variability caused by any size fractionationduring sampling and splitting. Such analysis is notpractical when using sieving because of the time andlabor required for each analysis.

Method: The method consists of adding a smallamount of soil or dust to a working fluid, usually eitherwater or isopropyl alcohol, and introducing this fluidinto a circulating system containing the analysis cellilluminated by lasers and surrounded by detectors atknown geometries. Care is taken to disperse the sampleand eliminate clumping. The data are accumulated ina few minutes and are then analyzed, reduced, anddisplayed in a variety of formats. Between runs, thesystem can be flushed and evaluated for a particle-freeinitial condition.For sizes below -4 im, and well down into thenanometer scale we are using another laser-basedinstrument (NanotracTM) that detects Brownian motionof the finest grains. These results are not reportedhere.

https://ntrs.nasa.gov/search.jsp?R=20090010196 2018-06-15T19:38:23+00:00Z

Percent Bin

NISTMicrotrac AReported10 0.33 pm 0368 pm -+10%25 0.57 pin 0.530 trn -08°%50

1 0.98 pm 0.959 pm

75 1.52 Vin 1.683 pm -+10%90 2.19 pin 2.808trn -+22%Table 1. Comparison of our result against NISTStandard Reference Material 5RM1978.

Percent Certified GlassMicrotrac ABin Standard

10 47-54trn 52.86 pm Within range50 55.5-59.5 pm 56.26 pm Within range90 57 - 70 jim 57.06 pm Within rangeWidth Less than 25 7.86 Within rangeTable 2. Comparison of our result against acertified (traceable to NIST) glass standard 1131.

Discussion: This measurement was done usingdistilled water as the working (carrier) fluid, whichallows us to directly compare our results with those of1, 2, 11], who also used distilled water for their wet

sieving processes on subsamples. The overall shape ofthe distribution is in agreement with that of Basu et al.for particles smaller than 3.47 (p (90i.tm). For thecoarser sizes, our data suggest that there are moreparticles in the 2 p to 2.74 p (150j.tmto 250i.tm) thanrevealed in the Basu et al. analysis.

The difference between our result and those ofBasu may be due to variation between subsamples.The subsample that we used was a few tens ofmilligrams. In a sample of this size, there may be some

variation in the largest size fraction because there maynot be enough particles to achieve a statistically validmeasurement.Conclusion: Direct comparison of sieve data to light-scattering data is difficult and the results must becarefully checked with standard known particle sizesand distribution. Our analyses of NIST or NIST-traceable grain size material shows that the Microtracresults are within acceptable ranges for particle sizesbetween 0.33 pm and 60pm. The correspondence ofthe sieve data from 10084 to the light-scattering data isgood and provides confidence that the two methodsproduce comparable results. These results open thedoor to future automated, rapid, and reproducible grainsize analysis of planetary soils and dust.

References: [1] Duke. MB., et al. (1970) Proc. Apollo 11 LunarSci. Conf., Geochirn. et Cosmochim. Acta Suppl., 1: p. 347; [2] KingJr, E.A., et al. (1971) Proc. 2 Lunar Sci, Conf, 2: p. 737 .746: [3]CalTier, W.D. (1973) The Moon, 6(3-4): p. 250-263: [4] McKay,D.S., et al. (1974) Proc. 5' Lunar Sci. Conf., Suppi. 5. Geochiinicaet Cosinochiinica Acta, 5: p. 887-906: [5] McKay. D.S., et al. (1976)Proc. 7th Lunar & Planet. Sci. Conf., 7: p. 295-313: [6] McKay, D.S.,et al. (1977) Proc. 8 th Lunar Sci. Conf.: p. 2929-2952; [7] McKay,D.S., et al., Grain si:e and the evolution ofLuna 24 soils, in MareCrisiuin: The View from Lana 24. Suppl. 9, Geochimica etCos,nochiozica Acta. 1978, Pergamon Press. p. 125-136; [8] McKay.D.S., et al. (1978) Proc. 9th Lunar & Planet. Sci. Conf.: p. 1913-1932: [9] McKay, D.S., et al. (1980) Proc. 1 i' Lunar & Planet. Sci.Conf.: P. 1531-1550: [10] McKay. D.S., et al.. The Lunar Regolith,in Lunar Sourcebook, G. Fleiken, D. Vaniman, and B.M. French,Editors. 1991, Cambridge University Press: Cambridge. p. 285-356;[11] Basu, A., et al. (2001) Meteoritics and Planetary Science, 36: p.177; [12] Greene, G.M., et al. (1975) Proc. 6th Lunar Sci. Conf., 1:

P . 517-527: [13] Plantz, P.E., Traceabiliti and Standards Applied toMicrotrac Instruments. 2007, Microtrac Inc. (download atinicrotrac.coin).

104) 54) 04)

120

I A,IL

40-___ ___ 74 ___ ___

2:__

—Subsample A-i, Run 1

—Subsample A-2, Run 1

Subsample B-i, Run 1

—Subsample B-2, Run 1

Basu et at.

Duke el al.

- King et at., 10084,80

King et al., 10084, 166

King et at., 10084,167

Sellers et al., in Graf

0.1 1 10 100 1000 10000

Diameter, urn

Figure!. Comparison of particle size distribution results from previous workers and our current research.,Our data were measured using water, which was also used for the Basu data. This allows straightforward comparison; however, becauseof the effect of water on lunar soil, we will re-measure this sample in isopropyl alcohol which we have found to minimize clumpingartifacts.