R

TO

a

ARRA

1

msgnrpiagpmaiai

oawOmhbt

A

0d

Nuclear Engineering and Design 251 (2012) 69– 77

Contents lists available at SciVerse ScienceDirect

Nuclear Engineering and Design

j ourna l ho me page: www.elsev ier .com/ locate /nucengdes

ecycling irradiated nuclear graphite—A greener path forward

.D. Burchell ∗, P.J. Pappano1

ak Ridge National Laboratory, Oak Ridge, TN 37831, USA

r t i c l e i n f o

rticle history:eceived 23 January 2011eceived in revised form 24 October 2011

a b s t r a c t

Here we report the successful recycle of irradiated graphite to fabricate new nuclear graphite usingconventional manufacturing processes (albeit on a bench scale). Radiological concerns such as the con-

14

ccepted 24 October 2011tainment of contamination in industrial scale manufacturing plants, or the release of C, were notconsidered. Moreover, a study of the annealing kinetics was conducted to elucidate the extent of prop-erty recovery over a representative temperature range. The goal of the preliminary work reported herewas to determine if nuclear graphite, produced through the normal graphite fabrication process, butusing crushed, previously irradiated nuclear graphite could be manufactured with sufficient mechanicalintegrity to warrant further investigation.

. Introduction

Deep Burn reactors, a high temperature reactor that can trans-ute spent light-water reactor fuel, can handle a variety of fuel

ources and be of several reactor designs, but all require nuclearraphite as the moderator. Over the life of the reactor a significantumber of nuclear graphite components will be replaced, creatingadioactive materials management issues, including storage, trans-ortation, and burial, all with associated costs and environmental

mplications. One potential solution to the irradiated graphite man-gement issue is to reuse/recycle the graphite. Reuse of irradiatedraphite could be as straightforward as shuffling the graphite com-onents within the reactor to maximize the useful life of thataterial. The recycle option would use expended graphite blocks

s the raw material for making fresh components. This latter options essentially the production of new graphite components through

true recycle process. Here we report the first experimental studynto the viability of recycling nuclear graphite.

Preliminary investigations (Peter et al., 2010) into the viabilityf recycling irradiated nuclear graphite were separated into threereas: (i) a study of the viability of recycling non-irradiated graphiteas carried out by GrafTech International (GTI), under contract toak Ridge National Laboratory (ORNL), who carried out a para-etric study, with parameters of grinding, mixing, forming, and

eat treatment. The product and properties were then evaluatedy ORNL; (ii) a series of experiments was carried out to determinehe annealing kinetics of irradiated graphite for an annealing range

∗ Corresponding author. Tel.: +1 865 576 8595.E-mail address: burchelltd@ornl.gov (T.D. Burchell).

1 Current address: Office of Fusion Energy Sciences, SC-24.2, 1000 Independenceve. SW, Washington, DC 2085-1290, USA.

029-5493/$ – see front matter © 2011 Elsevier B.V. All rights reserved.oi:10.1016/j.nucengdes.2011.10.068

© 2011 Elsevier B.V. All rights reserved.

appropriate to the recycle process. The crystal annealing kineticswas elucidated by measurement of electrical resistivity and spec-imen dimensions; and (iii) a series of nuclear graphite specimens,similar to that studied by GTI, but which had been irradiated in theHigh Flux Isotope Reactor, were recycled.

Results suggest that, within the narrow parameter range stud-ied, the materials could be formed with a level of density, strength,and thermal conductivity suggesting that the recycling processis viable. It is noted that the irradiated materials used in thisstudy were in a moderate range of irradiation dose associatedwith graphite densification, and that recycling would likely includegraphite irradiated to a higher irradiation dose.

2. Experimental

2.1. Unirradiated graphite recycle

The graphite recycle process, on an industrial scale, would beidentical to the flow diagram shown in Fig. 1, except that thecoke filler would be replaced with particles of ground irradiatedgraphite.

The steps shown in Fig. 1 were followed as closely as possi-ble, though some equipment was not available, specifically a pitchimpregnation autoclave. The purpose of this Deep Burn task was toinvestigate whether ground irradiated graphite can be used as thefiller for making new, or recycled, graphite. Consequently, we firstidentified the appropriate particle size and particle size to pitchratio, and then formed structurally sound graphite. The graphite

was then characterized, measuring compressive strength, thermalconductivity (TC), and coefficient of thermal expansion (CTE).

As a prelude to working with irradiated graphites, the laboratoryscale graphite fabrication process line was optimized to the extent

70 T.D. Burchell, P.J. Pappano / Nuclear Engineering and Design 251 (2012) 69– 77

RAW PETROLEUM OR PIT CH COKE

CALCINED COK E

BLENDED PARTICLES

COAL TAR BINDER PI TCH

GREEN ARTI FAC T

BAKED ARTI FAC T

GRAP HITE

CALC INED AT 1300 °C

CRUSHED, GROUND, AND BLE NDED

MIXED

COOLED

EXTRUDED, MOLD ED, OR ISOSTATICALLY PRESS ED

BAK ED AT 800 -100 0°C

IMPREGN ATE D TO DENSIFY (PETROLEUM PITCH)

REBAKE D & REIMPREGNATE D

ARTIFACT

GRAPH ITIZE D 2500-2800 °C

NUCLEA R GRAPHIT E

PURIFIED

Replac e with Irradiated graphite

pwtp

2

msCfgwgdtn

Fig. 1. Process flow diagram for the production of nuclear graphite.

ossible using non-irradiated material. A large portion of this workas dedicated to minimize graphite dust evolution and containing

he graphite making constituents in sealed environments, such asortable glove bags and hoods.

.2. Irradiated graphite recycle

Available irradiated and non-irradiated nuclear graphite speci-ens were reviewed for possible Deep Burn use. A previous GEN-IV

tudy irradiated and tested 32 grade NBG-10 (supplied by SGLarbon GmbH) specimens, leaving approximately 35 g of materialor this study (Burchell et al., 2006). The pedigree of this nuclearraphite along with the irradiation conditions of the previous studyas ideal for this recycle work. There was also a 40 g quantity of

raphite bend bars (50 mm × 6 mm × 3 mm) that had not been irra-iated (control material); these bars were used to setup and qualifyhe graphite recycle fabrication line. It is important to note that sig-ificant resources and time for this recycle work dealt with the safe

Fig. 2. Image of nuclear graphite bend bar specimens (50 mm × 6 mm × 3 mm) thatwere used as filler to make recycle graphite.

handling of the highly dispersible radioactive graphite powdersproduced.

Fig. 2 shows an image of the bend bars used for this study. Thebars were ground to a specific particle size, mixed with coal-tarpitch, formed into a cylinder, and then heat-treated. Based on pub-licly available data the maximum ground particle size was set at0.850 mm. This particle size roughly corresponds to a −20 mesh USStandard sieve size. The grinding technique reduced the bar intoa powder that would fit through a No. 20 sieve (mesh openingsof 0.850 mm). As part of this study several methods for grindingthe irradiated bars were investigated. The metrics used for evalua-tion of the effectiveness of the process were the quality of resultingpowder, and the ability to contain the radioactive contaminationproduced.

It was determined that each bar needed to be fragmented, andthen charged to a coffee grinder. The initial fragmentation was per-formed using a standard food processer. The bars were placed in theprocessor and the chopping action of the blade reduced the bar intoabout 10 smaller fragments. The fragments were then sent to thegrinder and −20 mesh material was made with one pass throughthe unit. The yield of −20 mesh graphite was about 90% after twopasses through the grinder. The process flow for reducing the bendbars into usable feedstock for the recycle graphite is shown in Fig. 3.The −20 mesh graphite was then mixed with powdered coal-tarpitch. The pitch was made into a −20 mesh powder via the coffeegrinder as well. Various percentages of graphite to pitch were triedin order to determine the maximum amount of graphite that couldbe used per forming run.

The graphite–pitch mixture was then poured into a graphite dieand pressed to various forces in order to determine what level offorce was needed to produce structurally sound recycle graphite.

One complicating factor that had to be considered during setupof the recycle graphite fabrication line was minimization and con-tainment of radioactive dust during grinding and forming. This wasthe primary reason un-irradiated samples were first tested. It wasdecided that to be as safe as possible, that the process line shouldbe fully enclosed in a portable glove bag. That is, all of the grind-ing, sizing, and die filling should occur in one glove bag. As thefinal pressing could not take place in the glove bag, plastic bagswere placed around the die during pressing. Fig. 4 shows an imageof the graphite die being loaded with the graphite–pitch mixture

while in the glove bag. ORNL safety and radiological safety per-sonnel reviewed all of the steps involved in the recycle graphiteprocess line. After all necessary approvals were received; the

T.D. Burchell, P.J. Pappano / Nuclear Engineering and Design 251 (2012) 69– 77 71

rs into

pw

tistaooagow

Fee

Fig. 3. Process flow for reducing graphite bend ba

rocess line was moved into a contamination/radiation zone forork on irradiated specimens.

The irradiated specimens were grouped according to irradiationemperature. Table 1 shows a list of the specimen groups. Approx-mately 5–7 g of bend bars was used to make one recycled graphitepecimen. Based on the un-irradiated bend bar work it was foundhat 100% ground bend bar could be combined with pitch. That is nodditional filler, such as synthetic graphite, was needed. The ratiof ground bars to pitch was held at 80:20, respectively. However,ne of the samples, L05-L66 was divided into two sub-samples, (a)

nd (b). One of the sub-samples was combined with a syntheticraphite (KS-15 provided by Timcal (http://www.timcal.com)) inrder to determine if the properties of the final recycled graphiteere affected by the presence of this synthetic graphite.

ig. 4. Image of graphite die being filled with ground bend bar–pitch mixture, whilenclosed in a portable glove bag for personnel safety and minimization of dustscape.

−20 mesh powder suitable for recycle feedstock.

Another variable that was investigated (that could affect thefinal properties of the recycle graphite) was annealing the bend barsbefore or after grinding (or not at all). Accordingly, some of the barswere annealed to 1800 ◦C under vacuum, while others were groundinto −20 mesh powder and then annealed (Fig. 5).

The importance of annealing is discussed in Section 3.3. Thethree ways to anneal (or not anneal) the bars prior to recyclegraphite formation are depicted in Fig. 5. Table 1 reports the formof the graphite, bar or powder, that was annealed.

Following forming by compression molding in a die, the “green”recycle graphites were carbonized to 950 ◦C under flowing helium.A heating rate of 5 ◦C/min and 1 h hold time at maximum tempera-ture was used. The carbonized graphites were then heat treated to1800 ◦C under vacuum, using a 20 ◦C/min heating rate and 1 h holdtime. The material was cooled at 20 ◦C/min until 700 ◦C was reach,after which time the furnace power was switched off.

2.3. Property determinations

The recycle graphite samples were all nominally 24.4–38.1 mmin length, and 12.7 mm in diameter. The following tests were per-formed on each sample: coefficient of thermal expansion (CTE),thermal conductivity (TC) by flash diffusivity, and tensile strength(Brazilian disc). The order of testing was also that order, since CTEand TC are non-destructive, but the CTE rig required a 1-inch sam-ple size. Following CTE measurement, the recycle graphite cylinderswere cut, using a diamond saw, to ∼4.8 mm thick discs for theflash diffusivity unit. After TC calculation from the flash test, thediscs were diametrally compressed to measure the tensile strength(Brazilian Disc test) (ASTM, 2011).

CTE testing was performed using an Anter Inc. dilatometer. Thesamples were heated 100–1000 ◦C at a rate of 5 ◦C/min. The unitdwelled for 30 min from 100 to 1000 ◦C at 100 ◦C intervals. A datapoint was taken at each 100 ◦C interval following the 30 min dwelltime.

Thermal conductivity (TC) was calculated from flash diffusivitymeasurements. Approximately 4.8 mm thick by 13 mm diameterdisks were prepared. One face of the disk was subjected to a thermalpulse and the time was measured for this pulse to be registered on

72 T.D. Burchell, P.J. Pappano / Nuclear Engineering and Design 251 (2012) 69– 77

Table 1Bend bars that were combined to make a recycle graphite.

Combined bend barsa Mass (g) Pitch add (g) Tirr (◦C) Mean irr. dose, n (cm−2) [E > 0.1 MeV] Annealed

L13-L43 7.5294 1.75 275 3.36 × 1025 NAL49-L27 5.2036 1.30 300 2.43 × 1025 NAL67-L48 7.3575 1.70 350 3.40 × 1025 NAL12-L21 6.6803 1.50 380 4.85 × 1025 PowderL07-L41 4.7657 1.10 675 5.49 × 1025 NAL05-L66(a) 4.0000 1.00 700 5.28 × 1025 Bar

70

io of g

thTh

t(wcpcat

�

wls

2

thdpM

L05-L66(b) 3.6700 1.15

a L05-L66(b) was also mixed with 0.92 g KS15 graphite to produce a 64:16:20 rat

he opposite face, and hence diffusivity was obtained. The specificeat (at room temperature) was assumed (from the literature) andC was calculated using the diffusivity, density, and the specificeat.

Brazilian disc tests (ASTM, 2011) were conducted to determinehe tensile strength using an electromechanical testing machineInsight 10, MTS Systems Co.) with a load capacity of 10 kNith ±0.26% precision (MTS Systems Co.). All measurements were

onducted at room temperature (∼20 ◦C) in air at a crosshead dis-lacement rate of 0.5 mm/min. A data acquisition rate of 25 Hz washosen. In order to prevent scattering of the irradiated fragmentsfter specimen failure, each specimen was wrapped in a conven-ional sealing wrap (Glad Press’n Seal) during the test.

The fracture stress was evaluated as follows:

frac = 2P

�dt

here �frac is fractured tensile strength, P is maximum appliedoad, d is diameter of the specimen and t is thickness of thepecimen.

.4. Electrical resistivity of annealed irradiated specimens

The effect of annealing was determined by measuring the elec-rical resistivity (ER) of irradiated graphite samples before and after

eat treatment to a given temperature. Electrical resistivity wasetermined in accordance with ASTM Standard C 611–98 (reap-roved 2005), “Standard Test Method for Electrical Resistivity ofanufactured Carbon and Graphite Articles at Room Temperature”

Fig. 5. Three ways to incorporate annealing of bend bars into recy

0 5.28 × 1025 Bar

raphite:KS15:pitch.

(ASTM, 2009). The test setup employed a Keithley 2400 SourceMeter (current supply), and a Keithley 2182 Nanovoltmeter. A pairof probes (knife edges) for applying the electrical current throughthe specimen and a second pair of probes for measuring the poten-tial were mounted in an insulating (Plexiglas) block. The magnitudeof the current was low, and the time that the current was allowedto flow through the specimen was kept short such that the tem-perature rise in the specimen was negligible, and consequentlythe resistance of the specimen was not changed. Multiple resis-tivity measurements were made on each specimen to determinean average value.

3. Results and discussion

3.1. Graphite recycle

Both un-irradiated and irradiated recycle graphites were fabri-cated in this project. Work was initially started on un-irradiatedsamples so that the recycle fabrication process line could be opti-mized without wasting the small available supply of irradiatedNBG-10 material (about 35 g was available). The best method forfragmenting and grinding the graphite was identified—initial frag-mentation in a food processer followed by grinding in a coffeegrinder. After the un-irradiated samples were reduced to the proper

particle size (<0.850 mm) they were mixed with pitch (solely) or amixture of pitch and synthetic graphite. The additional syntheticgraphite, KS-15 supplied by Timcal (http://www.timcal.com), wasthought to be necessary because the ground graphite would have

cle graphite formation. All three were tested in this project.

T.D. Burchell, P.J. Pappano / Nuclear Engineering and Design 251 (2012) 69– 77 73

Table 2Irradiated recycle graphites dimensions and density. Results are for un-impregnatedgraphite.

ID numbera Length (cm) Diameter (cm) Mass (g) Density (g/cm3)

L13-L43 3.689 1.368 7.165 1.32L67-L48 3.419 1.367 6.965 1.39L12-L21 3.417 1.377 6.966 1.37L07-L41 2.552 1.386 5.244 1.36L05-L66(a) 2.216 1.372 4.505 1.37L05-L66(b) 2.304 1.381 4.681 1.36

a L05-L66(b) was also mixed with 0.92 g KS15 graphite to produce a 64:16:20r

mAmg1

ggscufi

trttmutaiidca

cdmni

ctbtdsfgpiv

stpwtdi

0

1

2

3

4

5

6

7

110010009008007006005004003002001000

Ave

rage

CTE

(x10

-6/C

)

L13-L4 3L67-L4 8L12-L2 1L05-L66 (a)L05-L66 (b)L07-L4 1

ger the effect on the post-recycle CTE value. Another intriguing

atio of graphite:KS15:pitch.

inimal compressibility, leading to weak final recycle graphite.ccordingly a suite of 6 recycle graphites made from un-irradiatedaterial was fabricated with varying concentrations of synthetic

raphite. All of the graphites were pressed to 15.6–17.8 kN force at20 ◦C.

The density and appearance of samples with added syntheticraphite was not very different from samples without syntheticraphite added. Therefore it was determined that no additionalynthetic graphite was needed to form structurally sound recy-le graphite. This also maximized the fraction of recycle graphitesed, since the mix did not have to be supplemented with any otherller.

The densities of the recycled graphites (Table 2) were lowerhan that of commercially available nuclear grade graphites. Theeason for the low density of this recycled graphite was thatypical nuclear grade graphite is impregnated with pitch threeimes in order to increase its density. An autoclave and isostatic

older suitable for pitch impregnations were available to then-irradiated samples, but not the irradiated samples, as the con-amination zone lacks this equipment. The results obtained herere therefore for un-impregnated graphite. However, the viabil-ty of re-impregnating recycled graphite is not considered a keyssue and will be addressed at a later time. Moreover, the lowerensity may be attributed to the lab-scale equipment employedompared to production manufacturing equipment, and the lack of

graphitization step for the binder pitch.With demonstration of a recycle graphite fabrication pro-

ess route that produced specimens of acceptable strength andensity, while minimizing dust formation, work on irradiatedaterial began. The equipment was transferred to the contami-

ation/radiation laboratory and work on irradiated specimens wasnitiated.

One difference in the processing of irradiated specimens, asompared to un-irradiated samples, was the lack of need for ini-ial fragmentation. The un-irradiated graphites first needed to beroken, or fragmented, into smaller pieces before being ground inhe coffee grinder. The food processer did not fragment the irra-iated graphites. Instead, the chopping blade imbedded into theample and remained attached to the blade during the run. It wasound that the whole irradiated bend bar could be sent through therinder and an 80% yield of −20 mesh material was observed. Twoasses led to a 90% yield. Also, there was no observed difference

n grinding behavior between bend bars that had been annealedersus bars that had not.

The ground graphite was mixed with pitch and pressed in theame fashion as the un-irradiated samples. The only difference washat the pressing force was reduced to 2500 lbs to prevent theotential rupture of the graphite die. After pressing the graphitesere carbonized and heat-treated. The densities were lower than

hat of the un-irradiated samples because of the lower force useduring pressing. Densities for the recycled irradiated graphite were

n the range 1.32–1.39 g/cm3 (Table 2).

Temperature (C)

Fig. 6. Average CTE values for irradiated recycled graphites.

3.2. Recycled irradiated graphite properties

Fig. 6 reports the average CTE data for all of the specimens from100 to 1000 ◦C at 100 ◦C intervals. There was some difference in CTEat lower temperatures. The CTE values range from 3.7 × 10−6/C to5.2 × 10−6/C at 100 ◦C. Sample L07-L41 had the lowest CTE value,and it was a composite of bars that were irradiated at about 675 ◦C.The highest CTE value at 100 ◦C was sample L67-L48, which wasirradiated at 350 ◦C. At higher temperatures the CTE values con-verge. At 1000 ◦C the CTE values are all about 6 × 10−6/C.

When analyzing the CTE data we must consider two distincteffects (Burchell, 1997): (a) irradiation induced in-crystal effects(displacement damage) which stiffens the basal planes and influ-ences their thermal vibrational behavior, and (b) the irradiationinduced closure of aligned porosity (between the basal planes) thataccommodated the c-axis thermal expansion of the crystals. Ther-mal annealing will remove or modify a fraction of the in-crystaldefects (see subsequent discussion of electrical resistivity effects).Thus altering the lattice stiffness and thermal expansion (reducingthe defect density may restore some aligned porosity). Moreover,mechanical grinding may permanently eliminate a fraction of thealigned (accommodation) porosity and thus alter the expansionbehavior.

CTE measurements are good indicators of changes of a certainclass of pores in the graphite, i.e., those pores aligned perpendicu-lar to the crystals c-axis expansion, and as such yields informationabout the structure of the irradiated graphite (Kelly and Burchell,1994). The structural changes that occur on irradiation are depen-dent upon the irradiation temperature and the neutron dose. Thuswe might expect to see differing effects of the annealing on the CTEdepending upon the irradiation condition. The data in Fig. 6 indi-cate just that. Below ∼1000 ◦C the measured CTEs of the recycledgraphite specimens are markedly different. The CTE values of therecycled graphites are greater than that of the unirradiated (pre-cursor) NBG-10 graphite, thus grinding and annealing (to 1800 ◦C)has not removed all of the irradiation damage, which is known toincrease the CTE (in this dose/temperature regime).

Returning to Fig. 6, the general trend in the CTE data is for thelower irradiation temperatures and/or the higher irradiation dosespecimens to exhibit the largest reduction in CTE at temperaturesin the range 100–600 ◦C (Fig. 6). Evidently, the greater the extentof irradiation induced damage in the precursor materials the big-

observation is that in the annealing study discussed subsequently,annealing to 2000 ◦C alone had little or no effect on the dimensionsof a sample, yet the CTE is evidently very much modified (in the

74 T.D. Burchell, P.J. Pappano / Nuclear Engineering and Design 251 (2012) 69– 77

re

csrdg

cL1TTwbp

gFfirtCtt

ttduidosm

asgLg

3

o

Fig. 7. Mean room temperature TC values for irradiated recycled graphites.

ecycled specimens). The mechanism of thermal annealing and itsffects on CTE clearly requires further elucidation.

Following CTE determination, several ∼4.8 mm thick discs wereut from each sample. The discs were then tested in a flash diffu-ivity unit and TC (at room temperature) was calculated from theesults. Fig. 7 shows the mean TC data, in W/m K, for each of theiscs cut from a given sample. The samples made with 80% groundraphite and 20% pitch ranged in TC from 15 to 23 W/m K.

Interestingly, sample L05-L66(a) exhibited nearly twice theonductivity of L05-L66(b), 20.3 compared to 11.5 W/m K. The05-L66(b) sample was made with a 64:16:20 ratio of graphite:KS-5 synthetic graphite (http://www.timcal.com):pitch, respectively.he added synthetic graphite reduced the TC of the recycle graphite.his is another indication that no additional material is neededhen making the recycled graphite. The irradiated material can

e ground to the appropriate particle size and mixed solely withitch.

The room temperature thermal conductivity of unirradiatedraphite is ∼100–150 W/m K. The mean conductivities reported inig. 7 are significantly lower than this value. This is attributed to twoactors. First, the density of the recycled irradiated graphite artifactss much less than that of unirradiated graphite. Secondly, in ourecycle process the graphite is heat treated only to 1800 ◦C and noto 2500–2800 ◦C which is more typical in graphite manufacturing.onsequently, the defect structures produced in the graphite crys-al lattice on irradiation would not have been fully annealed andhus partially remain in the lattice as phonon scattering centers.

The specimens that were used for TC testing were finally frac-ured using the Brazilian Disc test (ASTM, 2011) to determine theensile strength. Un-irradiated disks were tested first in order toetermine the failure mode of the sample. Observation of the fail-re mode of the un-irradiated disk confirmed that the sample failed

n tension. Strength of the irradiated recycle graphites was thenetermined. As with the un-irradiated graphite, the failure modef the samples was tensile, occurring along the maximum tensiletress axis, located normal to the loading axis. Fig. 8 reports theean tensile stress (MPa) for each specimen group.All of the samples failed in tension at less than 3 MPa. Again,

significant difference was observed in the L05-L66(a) and (b)amples. The L05-L66(a) sample, which was comprised of groundraphite and pitch, had an average strength of 2.16 MPa. The L05-66(b) sample, which was comprised of ground graphite, syntheticraphite, and pitch, had a maximum average strength of 0.76 MPa.

.3. Annealing studies

Thermal annealing was performed on irradiated samplesf graphite NBG-10 to establish what fraction of in-crystal

Fig. 8. Mean tensile strength values for irradiated recycled graphites.

(displacement damage) and ex-crystal (pore generation) damagecan be recovered by thermal annealing. Six fractured NBG-10 bendbars were selected to represent the available spectrum of irra-diation temperatures and doses. The physical dimensions of thespecimens and electrical resistivity were measured, using the 4-point probe method (ASTM, 2009), and recorded in the irradiatedcondition, and following stepwise annealing at 500 ◦C incrementsfrom 500 to 2000 ◦C. All of the irradiated samples examined herehad their major axis aligned parallel to the extrusion axis. The(WG) resistivity for NBG-10 of 8.4 �� m (skin) and 9.1 �� m (core)are slightly less than the un-irradiated resistivity measured here(10.46 �� m).

Irrespective of the irradiation condition of the annealed recycledspecimen of this study (low dose swelling or subsequent shrinkagein excess of 1%) annealing at temperatures up to 2000 ◦C causedlittle or no change in the macroscopic dimensions (Burchell et al.,2010). However, this does not specifically reflect the dimensionalchange of the individual crystallites, which may have undergonesome dimensional recovery without affecting the macroscopicdimensions.

The electrical conductivity, �, for a given group of charge carrierscan be written (Issi, 2001):

� = Nq� = q2N�

m

where N is the charge carrier density, q is the electric charge, � isthe carrier mobility, � is the relaxation time and m is the chargecarrier effective mass. The relaxation time (�), is related to the car-rier mean free path, �, and is defined as the time elapsed betweentwo collisions, such that

� = �f �

The inverse (1/�) reflects the probability that a charge carrierwill be scattered, and �f is the charge carrier velocity at the Fermisurface. The carrier mean free path is the distance between twoscattering centers.

In graphite the dominant charge carriers are electrons andfor nuclear graphites the dominant scattering effects are intrinsic(phonon-scattering) at temperatures below ∼300 K, and extrin-sic scattering (lattice defects, crystallite edges) above ∼300 K (Issi,2001). In addition to electrons being scattered by defects, vacanciesmay act as electron traps.

The influence of intrinsic and defect scattering may be separated

using Mathiesson’s rule:

1�total

= 1�i

+ 1�d

T.D. Burchell, P.J. Pappano / Nuclear Engineering and Design 251 (2012) 69– 77 75

0

5

10

15

20

25

30

35

40

25002000150010005000

Annealing temperature, oC

Res

istiv

ity, m

icro

-Ohm

m

L-25 (280 C)L-33 (275 C)L-51 (700 C)L-58 (380 C)L-60 (300 C)L-69 (700 C)Non-irradiated

F1

wsotgsdifatit

twaectdwdtbiabT

atdp1t

ibbttiat

101214161820222426283032

25002000150010005000Ann eali ng Temperature, oC

Ele

ctric

al R

esis

tivity

, μΩ

·m

L-33, 275 C, 2.6E^25 n/m2 [ E>0.1 MeV] L-25, 280 C, 4.9E^25 n/m2 [ E>0.1 MeV]L-60, 300 C, 8.8E^24 n/m2 [ E>0.1 MeV]

Fig. 10. The effect of thermal annealing on the electrical resistivity of irradiatedNBG-10 graphite specimens irradiated at Tirr < 400 ◦C.

101214161820222426283032

25002000150010005000

Annealing Temperature, oC

Elec

tric

al R

esis

tivity

, μΩ

·m

L-58, 380C, 6.7E^25 n/m2 [E>0.1 MeV]

irradiation temperature samples (Fig. 10) and the high tempera-ture irradiated samples (Fig. 12) clearly illustrates that the defectsstructure established during irradiation are distinctly different. For

10

1214

1618

2022

2426

28

25002000150010005000o

Elec

tric

al R

esis

tivity

, μΩ

·m

L-51, 700C, 6.7E^25 n/m2 [E>0.1 MeV] L-69, 700C, 4.7E^25 n/m2 [E>0.1 MeV]

ig. 9. The effect of thermal annealing on the electrical resistivity of irradiated NBG-0 graphite specimens.

here �i is the intrinsic scattering and �d the defect mobility. Forcattering by grain boundaries, �d is approximately independentf temperature. An important parameter for characterizing scat-ering in graphite is the mean free path, �, which in well orderedraphites is effectively controlled by the <a>-direction crystalliteize la (Spain, 1981). When graphite is subject to fast neutron irra-iation carbon atoms are knocked out of their equilibrium positions

n the basal planes and Frenkel pairs (interstitial and vacancy) areormed. These additional scattering sites reduce electron mobilitynd electrons may be trapped by the vacancies (Spain, 1981). Hencehe electrical resistivity (inverse conductivity) is seen to markedlyncrease on irradiation from the un-irradiated value of ∼10 �� mo values >25 �� m (Fig. 9).

Thermal annealing of irradiated graphite will affect the conduc-ivity by (a) reducing the number of electron scattering centershen Frenkel pair recombine and annihilate, (b) when defects are

ble to thermally migrate along the crystal basal plane to the crystaldges and self-annihilate, and (c) when vacancies and interstitialsluster produce sites that more easily and effectively trap or scat-er electrons. Consequently, the effects of thermal annealing willepend upon the nature of the irradiation damage accumulated,hich in turn depend upon the neutron dose and critically, the irra-iation temperature. At irradiation temperatures below ∼400 ◦Che vacancies created in the graphite lattice are essentially immo-ile whereas the interstitial carbon atoms are mobile. Above an

rradiation temperature of ∼400 ◦C the vacancies are mobile and areble to coalesce. Consequently the electrical resistivity annealingehavior of the irradiated graphites will be different at Tirr < 400 ◦C,irr ∼ 400 ◦C, and Tirr > 400 ◦C.

The annealing behavior of the NBG-10 specimens irradiatedt temperature <400 ◦C is shown in Fig. 10. The room tempera-ure data show that electrical resistivity increases with irradiationose. As the annealing temperature increases the (room tem-erature) resistivity also increases up to annealing temperatures000 ◦C > Tann < 1500 ◦C. Above Tann of 1500 ◦C the electrical resis-ivity drops markedly.

The annealing peak at 1400 ◦C that has previously been observedn low temperature irradiated graphite (Tsuzuki and Arai, 1971) haseen attributed to the changes in defect structure when vacanciesecome increasingly mobile and coalesce to produce effective elec-ron traps, thus reducing the charge carrier density and relaxationime, and increasing the overall electrical resistivity. With increas-

ng annealing temperature the defect structure is further changednd vacancy clusters may diffuse out of the crystal or grow so largehey become ineffective as scattering center or traps. Similarly,

Fig. 11. The effect of thermal annealing on the electrical resistivity of irradiatedNBG-10 graphite specimens irradiated at Tirr ∼ 400 ◦C.

interstitial defects may be removed at higher annealing tempera-tures and hence the electrical resistivity is reduced. Such behaviorhas previously been reported by Kelly et al. (1966). The resistiv-ity annealing behavior for specimen L-58 (Tirr = 380 ◦C) is shown inFig. 11. There is a slight increase in resistivity with temperature toTann ∼ 1400 ◦C followed by a marked reduction.

The resistivity annealing behavior of specimens irradiated at∼700 ◦C is shown in Fig. 12. No annealing peak is seen in thedata, and the electrical resistivity recovers steadily at Tann > Tirr.The different resistivity annealing behavior observed with the low

Annealing Temperature, C

Fig. 12. The effect of thermal annealing on the electrical resistivity of irradiatedNBG-10 graphite specimens irradiated at Tirr > 400 ◦C.

7 Engine

afirifccdlttss

snastrdpt2

4

tgrmTtSb

sa1t

2iLgt(

iiltes

eagost2at

6 T.D. Burchell, P.J. Pappano / Nuclear

ll the specimens examined here annealing at 2000 ◦C was insuf-cient to completely recover the un-irradiated value of electricalesistivity. The extent of recovery was much greater in the spec-mens with higher irradiation temperatures, and was very smallor higher dose–low temperature specimens. Evidently, some in-rystal defects that scatter or trap electrons are still present in therystal lattice. It is noted (see Fig. 1) that the temperature reacheduring the graphitization process for the recycled graphite will

ikely be in the 2500–2800 ◦C range. It is likely that some addi-ional amount of annealing would have taken place. However, dueo furnace limitations the current work was limited to 2000 ◦C. Theame caveat can be made for the annealing effects on dimensionaltability.

By contrasting the effects of annealing of the gross dimen-ional change with the electrical resistivity it is most likely thato properties of interest to the reformation process occurs fornnealing temperatures below 1000 ◦C. While the gross dimen-ional change is unaffected by annealing to 2000 ◦C, the fact thathe electrical resistivity undergoes recovery in the 1000–2000 ◦Cange may be indicative of defect recovery and perhaps crystalimensional change which may ultimately impact the reformationrocess. This bears further study including a more precise inves-igation of crystallite lattice parameter changes and annealing to600 ◦C.

. Summary and conclusion

Irradiated NBG-10 nuclear graphite bend bars were subjected towo major tests for this project: (1) recycle and formation of a newraphite, and (2) determination of their annealing kinetics. In theecycle task the irradiated specimens were ground into powder,ixed with pitch, pressed into a cylinder, and then heat-treated.

he heat-treated samples were then characterized for density, CTE,hermal conductivity measurements, and tensile strength testing.ome of the irradiated graphites were also annealed in either bendar or powder form prior to mixing with pitch.

The CTE for all of the samples increased with increasing mea-urement temperature. The CTE value reached nearly 6 × 10−6/◦C,nd seemed to remain fairly constant at this value from 500 ◦C to000 ◦C. The expansion of the samples followed a non-linear pat-ern, reaching a maximum of 0.6% expansion at 1000 ◦C.

The thermal conductivities of the samples were around0 W/m K. This is consistent with un-impregnated graphite and

s therefore a positive result. The sample with the lowest TC was05-L66(b), which had synthetic graphite mixed in with the groundraphite and pitch. The presence of the synthetic graphite reducedhe thermal conductivity by almost half compared to sister sampleL05-L66(a)) with no synthetic graphite (11.5 vs. 20.3, respectively).

The strengths of the recycle graphites were all under 3 MPa. Thiss consistent with un-impregnated graphite and is therefore a pos-tive result. The samples were tested for failure strength (diametraloading), however the specimens were observed to have failed inension. Again, the addition of synthetic graphite had a negativeffect on the property being tested. The L05-L66(a) sample had atrength that was nearly double that of the L05-L66(b) sample.

The annealing of the bend bars had little or no effect on the prop-rties of the graphite formed therein. Some of the bend bars werennealed in bar form and some in powder form, and then recycledraphite was formed. There was no noticeable difference in CTE, TC,r strength based on an initial annealing of the graphite. This datauggest that an initial anneal may be unnecessary. However, due

o furnace limitations, the maximum annealing temperature was000 ◦C, which was not sufficient to remove all the crystal dam-ge imparted during irradiation. The annealing task also showedhat an anneal temperature of 2000 ◦C may have had little effect on

ering and Design 251 (2012) 69– 77

the graphite’s pore structure, whereas the electrical resistivity wasmarkedly reduced (partially recovered).

For all the specimens examined here annealing at 2000 ◦C wasinsufficient to completely recover the unirradiated value of elec-trical resistivity. The extent of recovery was much greater in thespecimens with higher irradiation temperatures, and was verysmall for higher dose–low temperature specimens. Evidently, somein-crystal defects that scatter or trap electrons are still present inthe crystal lattice.

The recycle graphites made in this study used 100% groundirradiated material or a ratio of 64% ground graphite, 16% syn-thetic graphite, and 20% pitch. The samples made with 100%recycle material, i.e. no synthetic graphite, were superior instrength and thermal conductivity to samples made with a syn-thetic graphite/recycle graphite blend. The addition of syntheticgraphite reduced the strength and thermal conductivity of the finalgraphite by ∼50%. These results indicate that fabrication of struc-turally sound graphite made from 100% recycle material is feasible.The results also suggest that no additional filler material is required.

In all of the recycle graphites the values for the propertiesmeasured were significantly lower than that of the virgin mate-rial produced at the commercial scale by SGL. This is a resultof manufacturing differences in that the virgin nuclear graphiteswere pitch impregnated three times, thus increasing the densityof the final piece. Also, the recycled samples tested here wereformed via compression molding, whereas the virgin materialwas extruded. Neither an extrusion-press nor pitch impregna-tion autoclaves were available in the contamination zone wherethis research was performed. With the available lab-scale equip-ment and the very limited amount of irradiated material, irradiatednuclear graphite could not be properly “recycled” to produceproduction standard nuclear graphite with sufficient mechanicalintegrity. However, the results achieved at the “lab scale” bothfor virgin and irradiated starting materials are consistent with un-impregnated, low-density carbon and graphites. The difference indensity is therefore not considered an issue for this work, thoughit will be investigated at a later date.

Acknowledgements

This work was carried out for the Deep Burn Project of theU.S. Department of Energy, Office of Nuclear Energy Science andTechnology under contract DE-AC05-00OR22725 with Oak RidgeNational Laboratory, managed by UT-Battelle, LLC. Use of the HighFlux Isotope Reactor at the Oak Ridge National Laboratory wassupported by the U.S Department of Energy. This manuscript hasbeen authored by UT-Battelle, LLC, under Contract No. DE-AC05-00OR22725 with the U.S. Department of Energy. The authors wishto acknowledge the assistance of Joe Strizak, Marie Williams, AshliClark, and Kazumi Ozawa with the physical property determina-tions.

References

ASTM D 3967-08, 2011. Standard Test Method for Splitting Tensile Strength of IntactRock Core Specimens. ASTM International, West Conshohocken, PA, USA.

ASTM C 611-98, 2009. Standard Test Method For Electrical Resistivity of Manufac-tured Carbon and Graphite Articles at Room Temperature, vol. 5.05. Ann. Bookof ASTM Standards.

Burchell, T.D., 1997. Radiation effects in graphite and carbon-based materials. MRSBulletin XXII (2), 29–35.

T.D. Burchell, L.L. Snead, A.M. Williams, J.L. Bailey, J.P Strizak, 2006. Final PostIrradiation Examination Data Report for SGL NBG-10 Nuclear Grade Graphite,ORNL/TM/-2006/430, April 30.

Burchell, T.D., Pappano, P.J., Strizak, J.P., 2010. A study of annealing behavior ofneutron irradiated graphite. Carbon 49, 3–10.

Engine

I

K

K

T.D. Burchell, P.J. Pappano / Nuclear

ssi, J.-P., 2001. Electronic conduction. In: Delhaès, P. (Ed.), In: Graphite andPrecursors, World of Carbon, vol. 1. Gordon and Breach Science Publishers,

chap. 3.

elly, B.T., Burchell, T.D., 1994. Structure-related property changes in polycrystallinegraphite under neutron irradiation. Carbon 32 (3), 499–505.

elly, B.T., Martin, W.H., Price, A.M., Dolby, P., Smith, K., 1966. The annealing ofirradiation damage in graphite. J. Nucl. Mater. 2, 195–209.

ering and Design 251 (2012) 69– 77 77

Peter, J., Pappano, Timothy, D., 2010. Burchell preliminary data on process-ing and characterization of recycled irradiated graphite. Carbon 48, 3303–

3305.

Spain, I.L., 1981. Electronic transport properties of graphite, carbons, and relatedmaterials. Chem. Phys. Carbon 16, 119–322.

Tsuzuki, T., Arai, S., 1971. Annealing effects on galvanomagnetic properties of heavilyradiation-damaged graphite. Jpn. J. Appl. Phys. 10, 580.