Radiation shielding compositesusing thermoplastic polymersmouldable at low temperature

John McMillan

Department of Physics and Astronomy, The University of SheffieldSheffield, South Yorkshire, S3 7RH, Great Britain

j.e.mcmillan@sheffield.ac.uk

August 15, 2019

Abstract

The formulation of a low temperature thermoplastic mouldable radiation shielding composite is given. Thematerial is based on a low temperature melting polymer filled with lead shot. It can be easily moulded toshape after immersing in hot water, or by using a hot air gun. When cooled it is a rigid self supporting mass.Alternative formulations to provide low background or low toxicity gamma shielding or neutron shieldingare also considered.

I. Introduction

Many medical and experimental proceduresbenefit from the availability of radiation shield-ing that can be moulded to shape without us-ing metal cast at high temperature. Formula-tions for shielding putties based on clays andthermoset resins [1, 2, 3, 4] have been publishedbut these have the disadvantage that the ma-terial is not reusable. A commercial productknown as “EnviroClay” based on bismuth heldin some binding clay is available from Radia-tion Products Design Inc. [5] but this remainssoft and pliable at room temperature.

A thermoplastic shielding material was intro-duced by Maruyama et al. [6] which consistedof lead shot embedded in a matrix of a toolingwax (M. Argüeso & Co., “Rigidax”). A similarcomposition, described by Haskell [7], was athermoplastic dental wax (Moyco Industries“impression compound”) loaded with bismuthpowder, chosen for its low toxicity. This lastwork suggests that “a synthetic thermoplasticcould be substituted for the preferred hydrocar-

bon wax blend” but, presumably, at that time,no suitable material was available.

The current work describes the use of poly-caprolactone as a binder to produce novel ther-moplastic shielding compounds for gamma,X-ray and neutron applications. The impetuswas to improve shielding in low backgroundparticle-astrophysics experiments. Such ex-periments are typically operated deep under-ground within shielding castles constructedfrom lead blocks and with an inner layer ofultra-pure copper blocks. Inevitably, there areports and voids in the shields where pipeworkand cables pass through the castles. Further,the castles are customarily fabricated fromavailable or stock-sized blocks or ingots whichrarely match the experimental apparatus accu-rately.

There is a clear requirement for a shieldingmaterial that will fill the spaces and voids inthe castle as precisely as possible. Either cut-ting or casting metal to shape are possible butboth are problematic, especially in an under-ground laboratory. Another approach is to use

1

arX

iv:1

908.

0519

7v1

[ph

ysic

s.in

s-de

t] 1

4 A

ug 2

019

lead wool or shot held in bags or sacks, butthese have handling problems and the potentialrelease of finely divided material. Incorporat-ing the shielding metal in a polymer to providea mouldable thermoplastic offers a simple sys-tem of containment combined with flexibilityof use and re-usability.

In low background experiments, a key re-quirement is the absence of radioactive impuri-ties in the shielding, consequently the formu-lations based on waxes [6, 7] were not usable.These waxes contain high proportions of min-eral fillers which would be expected to containunacceptable levels of uranium and thorium.EnviroClay [5] was similarly rejected.

It is considered that the composites de-scribed here will have more general applica-bility than the application for which they wereoriginally produced.

II. Polycaprolactone

Polycaprolactone is a rigid crystalline polymer,having the unit formula (C6H10O2)n, whichis remarkable for its low melting point of 58–60◦C. When molten it easily wets uneven orgreasy surfaces and is consequently much usedas a hot-melt adhesive. It is soluble in a rangeof common solvents, is intrinsically non-toxicand is bio-degradable.

A range of thermoplastic polycaprolactonesare manufactured by Ingevity [8] as powders orpellets under the CAPA tradename. These werepreviously sold under the Perstorp and Solvaybrand names. Various grades are available withmolecular weights from 10,000 to 80,000. Smallquantities of polycaprolactone (CAPA6800) inpellet form are marketed for experimental andhobbyist purposes under the tradename “Poly-morph” in the UK and this was the inspirationfor the present study. Similar material is mar-keted under the names “Shapelock”, "Polydoh","Friendly Plastic" and others.

Initial investigations of samples of Perstorppolycaprolactone using a low-background Ger-manium detector indicate that its radiopurityis high, comparable with other pure polymermaterials. Concentrations of uranium, thorium

and potassium were all measured as less than30ppb by weight.

III. Shielding for X-rays and

gamma-rays

To provide maximum shielding for X-rays andgamma-rays the thermoplastic matrix shouldbe loaded as highly as possible with densemetal. For the present project fine lead shotwas chosen for cheapness, but copper could beused to produce a low background composite,while bismuth would produce a low toxicitycomposite, but at higher cost. Tungsten gran-ules could also be considered and this wouldoffer a method of producing very dense shieldswithout the difficulty of machining tungstenmetal.

The ratio of lead to binder was chosen toensure that the composite would have the max-imum strength available consistent with themaximum density, which would be obtainedwhen all voids between the lead are filled withpolymer. Assuming that the lead shot con-sisted of perfect identical spheres, which wererandomly stacked, the highest packing fractionobtainable is 0.64 [9]. The proportional mass ofpolymer required to bind the lead is given by:

mpoly

mPb=

(1 − 0.64)ρpoly

0.64ρPb= 0.56 · 1.1

11.3= 0.054

Maruyama et al. [10] point out that by usinga mixture of shot sizes, even higher packingfractions may be attained. Specifically, with amixture of two sizes having a ratio of diam-eters of 0.22 and a mixing ratio of between1:2–1.4 larger to smaller pellets, a packing frac-tion of 0.72 was obtained. More recent workon packing fractions of particles with differentdistributions of sizes was reviewed by Brouw-ers [11]. Such distributions were not furtherinvestigated in the present study.

Lead based thermoplastic shielding compos-ite was prepared using Perstorp CAPA6506, apolycaprolactone with an approximate molecu-lar weight of 50000 produced as powder witha 600µm grain size. It would be difficult to

2

ensure a homogenous sample if pelleted poly-caprolactone were used.

The lead shot was obtained from Calder In-dustrial Materials Ltd. [12] and had an averagesize of 1.3mm, though it was clear that therewas a distribution both in size and shape. Thepacking fraction of this material was measuredas 0.63, only marginally less than the theoreti-cal prediction for uniform spheres [9].

Initial 100g samples were prepared by weigh-ing out the lead shot and polycaprolactoneand intimately mixing them by simply stirring.They were then heated in a shallow stainlesssteel vessel (known in the UK as a “balti dish”)on a hot plate. Production quantities of thecomposite, up to 1kg, were subsequently madein larger stainless steel pans. It was foundadvantageous to include 50ml or so of waterper 100g of lead to prevent the temperatureexceeding 100◦C, as the polycaprolactone de-composes at about 200◦C and smokes badlyif overheated. The water also helped conductthe heat uniformly to the composite mass and,to an extent, prevented the composite fromadhering to the pan.

A certain amount of stirring or kneading thematerial was required to ensure that the leadand polymer were adequately mixed. At thisstage the composite was a viscous toffee-likemass. It could be removed from the dish witha spoon or gloved hand and any water remain-ing on it would run off. The warm compositewas somewhat sticky and when moulding thematerial, it was found that the mould shouldeither be greased or lined with polythene ormylar sheet. On cooling, it was found almostimpossible to remove the composite from themould if this precaution was not taken.

When the composite is at the boiling point ofwater, extreme care should be taken as gettingthe hot material on bare skin would be likelyto cause a painful scald. However, when it hadcooled to 60◦C it was found that it could beeasily worked to shape by hand and remainedworkable for a few minutes.

The cooled material was a tough resinoussolid with a measured density, ρcomp = 6.7,which was slightly less than the predicted den-

Figure 1: Test sample of lead composite.

sity of 7.4. Presumably a small fraction of voidsremain. A small (66g) sample moulded intoa round cornered cylinder can be seen in fig-ure 1. At the surface, all the lead particles arecovered with a film of polymer which shouldassist in minimizing contact with toxic metal.Surprisingly, the composite was found to beelectrically non-conducting; each particle oflead being insulated from its neighbours byfilms of polymer.

The composite could be remelted by simplyreturning it to the pan and reheating. Someexperiments were made on moulding the com-posite using a hot air gun. This was possiblebut care had to be taken since the compositehad rather poor thermal conductivity and itwas found easy to overheat the material at thesurface before the core had melted.

Simple tests of the gamma shielding capabil-ity of the composite using weak sources anda geiger counter indicated that its propertieswere exactly that which would be expectedfrom reduced density lead.

3

IV. Shielding for neutrons

Shielding of particle-astrophysics experimentsagainst environmental neutrons typically re-quires hydrocarbon material of at least 10cmthickness in order to ensure that the externalneutrons are thermalised and captured. In thecase of detector systems larger than a cubic me-tre, the requirement for neutron shielding ma-terial is of order some tonnes, therefore thereis the overriding need for low cost material[13]. As before, there is also the requirementfor high radiopurity of low background mate-rials. Similar considerations are also applicableto other experimental disciplines where bulkneutron shielding is required.

Sakurai [3] and Okuno [4] both describethe use of thermoset resins loaded with ei-ther isotopically enriched lithium or boroncompounds to capture thermal neutrons. Forparticle-astrophysics applications, the difficultyof re-use of such material makes the useof thermosets unattractive while the cost oflarge quantities of these materials remains pro-hibitive.

Considerable use has been made of thickslabs of polypropylene or high densitypolyethylene, which are widely available com-mercially but remain relatively expensive. Afar cheaper option is to use raw polyethyleneor polypropylene pellets held in fabric bags orsacks or contained within wooden shuttering[13]. The thickness of material has to be in-creased to allow for the packing fraction butthis is rarely problematic.

The availability of a mouldable compositebased on polymer pellets would mean thatthe containment could be dispensed with. Acomposite material can be produced usingpolypropylene pellets embedded in water ex-tended polyester (WEP) at reasonable cost [13].Unfortunately, this has the disadvantages men-tioned above with respect to thermosets com-bined with problems of manufacture associatedwith the presence of styrene monomer in theresin.

Pure polycaprolactone, being (C6H10O2)n isan intrinsic neutron shielding material, how-

ever, it is not a cost-effective choice, beingapproximately seven times as expensive aspolypropylene pellets. To reduce the cost, acomposite was produced using polypropylenepellets bound together using polycaprolactone.A composite in which all the voids were filledwould have optimal strength, while a compos-ite with minimum binder would result in thelowest cost.

The materials used were CAPA6506 poly-caprolactone powder and Targor Procompolypropylene pellets. These latter wereroughly cylindrical with a diameter of about4.5mm and an average of 3.6mm thick, thoughthere was considerable variation between pel-lets. Experiments were performed to determinethe optimum ratio of pellets to binder. The pro-duction method was similar to the lead version.To produce large panels of this composite, itshould be possible to make a mould in whichthe ingredients are heated by blowing steamthrough the material.

It was found that with low proportions ofCAPA to polypropylene beads, the resultingcomposite was not structurally supporting andeasily disintegrated. Composite produced with25g CAPA to 100g of polypropylene beads ormore was structurally sound, however, it isof comparable price to commercially availablethick slabs of polypropylene. While it is onlymarginally cost-effective, it may have advan-tages in terms of easy fabrication in complexshapes and in re-usability.

While this was not attempted, it is clear thatquantities of boron or lithium compounds, orminerals such as colemanite, could be added tothe composite without greatly affecting its me-chanical properties. Other neutron absorberssuch as compounds of gadolinium, cadmiumor europium could similarly be considered.Polycaprolactone could also be employed asan alternative to the epoxy resin binder usedin “crispy mix” boron-carbide based neutronshielding composites [14], again giving a re-mouldable thermoplastic material.

4

V. Long term stability

While no long term assessments of these com-posites have been performed, a few generalobservations can be made. Polycaprolactonesare biodegradable, but this appears to be byfungal attack in the presence of moisture. Ifthe material is kept dry, it can be expected tolast as long as any other plastic material. Ifpolycaprolactone is exposed to very high doses(500kGy) of gamma radiation the polymer willcross-link [15], which would be expected to ren-der the material stiffer and even more viscousin the molten state and would presumably leadto embrittlement in the solid phase.

VI. Acknowledgements

The author would like to acknowledge the as-sistance of Solvay in providing samples of poly-caprolactone.

References

[1] M. Farshani and F.C. Eichmiller. Metal-polysiloxane shields for radiation therapyof maxillo-facial tumors. Med. Physics,18:273–278, 1992.

[2] F.C. Eichmiller. Device and method forshielding healthy tissue during radiationtherapy. US Patent 5360666, 1994.

[3] Y. Sakurai, A. Sasaki, and T. Kobayashi.Development of neutron shielding ma-terial using metathesis-polymer matrix.Nucl. Instr. Meth. Phys. Res., A522:455–461,2004.

[4] K. Okuno. Neutron shielding materialbased on colemanite and epoxy resin. Rad.Prot. Dosim., 115:258–261, 2005.

[5] Radiation Products Design, Inc., 5215Barthel Industrial Drive, Albertville, MN55301, USA,See also http://www.rpdinc.com.

[6] Y. Maruyama, V.C. Moore, D. Burns, andM.T.J. Hilger. Individualized lung shields

constructed from lead shot embedded inplastic. Radiology, 92:634–635, 1969.

[7] D.A. Haskell. Remoldable thermoplas-tic radiation shield for use during radia-tion therapy. US Patent 5550383, WIPO96/33501, 1996.

[8] Ingevity Corp., 5255 Virginia Ave., NorthCharleston, SC29496, USA. See alsohttps://www.ingevity.com.

[9] H.M. Jaeger and S.R. Nagel. Physics ofthe granular state. Science, 255:1523–1531,1992.

[10] Y. Maruyama, D. Wrede, E. Van Arsdale,J. Sayeg, and E.P. Engels. Comments onshielding by the lead shot method. Radiol-ogy, 102:445, 1972.

[11] H.J.H. Brouwers. Particle-size distributionand packing fraction of geometric randompackings. Phys. Rev., E74:4031309, 2006.

[12] Calder Group, Elder House, Park West,Jupiter Drive, Chester, CH1 4RL, UK,See also https://www.caldergroup.eu.

[13] J.E. McMillan. Neutron shielding for par-ticle astrophysics experiments. ArXivPhysics e-prints physics/0510186.

[14] V.T. Pugh and B.W. Hendy. Crispy mixand Flexy mix – high boron carbide con-tent resin bonded neutron shielding mate-rials. Proc. ICANS VIII, RAL-85-110, p670–675, 1985.

[15] D. Darwis, H. Mitomo, and F. Yoshii.Enzymatic degradation of radiationcrosslinked poly(ε-capro-lactone). Polym.Degrad. Stabil., 62:259–265, 1998.

5