^legibility notice

8/14/2019 ^Legibility Notice

1/151

A major purpose of the Techni-cal Information Center is to providethe broadest dissemination possi-ble of information contained inDO E 's Research and DevelopmentReports to business, industry, theacademic community, and federal,state and local governments.

Although a small portion of thisreport is not reproducible, it isbeing made available to expeditethe availability of information on theresearch discussed herein.


2/151

BNL52204DE90 000886

O P T I M I Z A T I O N O F T H E F A C T O R S T H A T A C C E L E R A T E L E A C H I N GT O P I C A L R E P O R T

M . F u h r m a n n , R .F . P i e t r z a k , E .M . F r a n z ,J . H . H e i s e r I I I , a n d P . C o lo m b o

M a r c h 1 9 8 9

N U C L E A R W A S T E R E S E A R C H G R O U PD E P A R T M E N T O F N U C L E A R E N E R G YB R O O K H A V E N N A T I O N A L L A B O R A T O R Y

A S S O C I A T E D U N I V E R S I T I E S , I N C .U P T O N , L O N G I S L A N D , N E W Y O R K 1 1 9 7 3P r e p a r e d f o r t h e

U N I T E U S T A T E S D E P A R T M E N T O F E N E R G YN A T IO N A L L O W - L E V E L W A S T E M A N A G E M E N T P R O G R A M ;,., g ,,,,,U N D E R C O N T R A C T N O . D E - A C 0 2 - 7 6 C H 0 0 0 1 6 f i $ n S i t R

v f


3/151

DI SCLAI MERThis report was prepared as an account of work sponsored by an ag ency of the UnitedStates Government. Neither the United States Government nor any agency thereof,nor any of their employees, nor any of their contractors, subcontractors, or theiremployees, makes any warranty, express or implied, or assumes any legal liability orresponsibility for the accuracy, completeness, or usefulness of any information,apparatus, product, or process disclosed, or represents that its use would not infringeprivately ow ned right s. Reference herein to any specific comm ercial product, process,or service by trade name, trademark, manufacturer, or otherwise, does not nec essarilyconstitute or imply its endorsement, recomm endation, or favoring by the United StatesGovernment or any agency, contractor or subcontractor thereof. The views andopinions of authors expressed herein do not necessarily state or reflect those of theUnited State s Governm ent or any agenc y, contractor or subcontractor thereof.

Printed in the United States of AmericaAvailable fromNational Technical Information ServiceU.S. Department of Commerce5285 Port Royal RoadSpringfield, VA 22161NTIS price codes:Printed Copy: A08; Microfiche Copy: A01


4/151

TABLE OF CONTENTS

Page

EXE CU TIVE SUM MA RY xvii1. INTROD UCTION 12. MA T E RIA L S A N D ME T H O D S 3

2.1 W aste Form Types 72.1.1 Portland I Ce me nt Containing Sodium Sulfate as Simulated W aste 72.1.2 Portland I Ce me nt Containing Incinera tor Ash as Simulated W aste 72.1.3 Vinyl Ester-Styren e Containing Sodium Sulfate as Simulated W aste 72.1.4 Bitume n Containing Sodium Te trab ora te as Simulated W aste 8

2.2 Pre para tion of Samples 82.3 Leach ing Tests 92.4 Analytical M ethod s 9

2.4.1 Rad iochem ical Analysis 102.4.2 Elem ental Le ach ate Analysis 102.4.3 Alkalinity M easure me nts 102.4.4 pH M easurem ents 102.4.5 SEM /EDS 10

3. LEACH ING MECHANISMS AND MOD ELS 113.1 Introduc tion 113.2 M ethods of Presenting Leaching Da ta as a Function of Time 113.2.1 Tabu lar and Graph ical M ethods 113.2.2 Em pirical Equ ations 133.2.3 M athema tical Solutions to Mass Tra nspo rt Equ ations 133.2.3.1 Analytical So lutions for Diffusion 143.2.3.2 Co ncen tration De pen den t Diffusion 163.2.3.3 Skin Effects of Sam ple Surface 173.2.3.4 Diffusion + Re actio n (K d) 173.2.3.5 Anom alous Tran sport 183.2.4 Num erical Solutions to the Tran sport Equ ations 193.2.5 Te m per atur e Effects and the Arrh enius Equ ation 203.3 Application of Modeling Techniques 213.3.1 M odeling Releases from Cem ent 213.3.1.1 Diffusion Coefficient (D e) Calculated from the Semi-Infinite Model . . . . 21

-iii-


5/151

TABLE OF CONTENTS - continuedPage

3.3.1.2 Th e Finite Cylinder Mod el 223.3.1.3 Effects of Leach ing Con ditions 23

3.3.2 Vinyl Este r-Sty rene + Sodium Suifate 303.3.3 Bitumen 31

3.4 Summary 314. PORTLAND CEMENT CONTAINING SODIUM SULFATE AS SIMULATED

WASTE 354.1 Introduc tion 354.2 M odeling and Mechanisms of Leaching 364.3 Single Factors that Acc elerate Leaching 424.3.1 Tem peratu re 424.3.1.1 Portland Cem ent Pa ste 434.3.1.2 Portland Cem ent Plus 5 wt% Sodium Suifate 454.3.2 Specime n Size 494.3.3 Volume of the Leachant 49

4.4 Com bined Acc eleration Fac tors for Portland Cem ent Con taining SodiumSuifate 504.4.1 Cs-137 Re sults 504.4.2 Sr-85 Res ults 604.5 Conclusions 67

5. PORT LAND CEM ENT CONTAINING INCINERATOR ASH 695.1 Introduction 695.2 M odeling and Mechanisms of Leaching 695.3 Single Factors that Acc elerate Leaching 76

5.3.1 Tem peratu re 765.3.2 Size 795.3.3 Th e Volum e of Leac hant 805.4 Com bined Acceleration Factors 805.4.1 Cs-137 Results 815.4.2 Sr-85 Re sults 87

-iv-


6/151

TABLE OF CONTENTS - continuedPage

5.5 Con clusions 956. ASPH ALT (BITUMEN ) . . 97

6.1 Introduc tion 976.2 Effect of Single Facto rs on the Leaching of the Bitumen 976.3 W aste Loading 1006.4 Mechanisms of Leaching 1066.5 Modeling 1146.6 Con clusions 115

VINYL ESTER -STYR ENE CONTA INING SODIUM SULFA TE SALT 1177.1 Introdu ction 1177.2 M odeling and M echanisms of Leaching 1177.3 Single Factors that Acc elerate Leaching 121

7.3.1 Te m pe ratu re 1217.3.2 Size 1227.3.3 Vo lum e of Lea cha nt 1227.4 Com bined Acce leration Factors 1237.5 Con clusions 129

CON CLUS IONS 1318.1 Leaching Cem ent Containing Sodium Sulfate 1318.2 Portland Cem ent Containing Incinerator Ash 1318.3 Bitumen Containing Sodium Te trabo rate 1328.4 Vinyl Este r-Sty rene Co ntaining Sodium Sulfate 132

-V -


7/151

LIST OF FIGURESPage

Figure 3.1 A compa rison of characteristic leach curves: Fickian, Sigmoid, Two-staged.After Crank [15] 19Figure 3.2 Cs-137 CF L versus t ^ 2 (in days 1 /2) from cement containing sodiumsulfate leached in deionized water at 20C 22Figure 3.3 CF L of Sr-85 and Cs-137 versus time leached from cem ent containing 5wt% sodium sulfate in deionized water at 20C. Th e leachant was changeddaily. Re leases we re modeled using the finite cylinder mode l 23Figure 3.4 A comp arison of some basic leaching techniq ues: Static, Semidynamic, and

Flow tests 24Figure 3.5 CF L of Sr-85 leached from cem ent containing 5 wt% sodium sulfate indeionized water at 20C. A comparison of the effect of change in theleachan t replacem ent schedule is shown 25Figure 3.6 Cs-137 CF L versus time for cem ent containing 15 wt% incinera tor ash at20C. Th e solid lines represen t the finite cylinder model prediction w ithD e = 6 . 0 x l 0 ' 8 cm 2/s for the early results of 0 to 20 days and D e = 1 . 0 x l 0 ' 9for th e long-term results 26Figure 3.7 CF L of Sr-85 versus time from ceme nt containing 5 wt% sodium sulfateleached in deionized water at 20C. A comp arison of results was madebetween results for leachant volume to sample surface area ratios of 10/1,30/1 and 50/1 27Figure 3.8 Log (C FL x V/S) versus log t for Cs-137 from cem ent containing ion-exchange resin. Six sizes of cylindrical waste forms are co mp ared aftercorrection with V/S 28Figure 3.9 Arrh enius plot of Log D e versus 1/T for Cs-137 and Sr-85 leached fromnea t cem ent in deionized wa ter 29Figure 3.10 CF L versus time for Co-57 from vinyl ester-styrene containing 20 wt%sodium sulfate leached in deionized water at 20C. Th e solid linerep res ents the finite cylinder calculation 30Figure 3.11 CF L versus time for Cs-137 leached from bitum en containing 0, 20, 30,and 40 wt% sodium tetrab orate in deionized water at 20C 31

- V I -


8/151

LIST OF FIGURES - continuedPage

Figure 4.1 Averages of triplicate data for three types of leach tests: AN S 16.1, semi-dyriamic with daily Ieachant replacements, and flow-through tests. The sedata are plotted against the square root of time which produces a linearplot up to C FL =0 .2 if diffusion is the leaching mechanism . : . 36Figure 4.2 Th e finite cylinder mo del fails to follow the Cs-137 relea se da ta for thesemidynamic ANS 16.1 test (1), but does accurately model the data fromthe flow test (2) 38Figure 4.3 Experim ental data and model results for Cs-137 from experimen ts with dailyleachant replacements and with the ANS 16.1 schedule of leachant

replacement 39Figure 4.4 Triplicate baseline data for Sr-85 are plotted against the squa re root of time.Th ere is a change in leach rate after the first week of an AN S 16.1 test . . . 40Figu re 4.5 M odeling Sr-85 relea ses requ ired two diffusion coefficients, on e for theearly portion of the experim ent and on e for long-term leaching. For theexperiment with daily leachant replacements, the model fits with onediffusion coefficient 40

Figure 4.6 Da ta from th ree types of leach tests: AN S 16.1, semidynamic with dailyleachant replacements, and flow-through tests. These d ata are plottedagainst the square root of time which produces a linear plot if diffusion isthe leaching mechanism 41Figure 4.7 CF L of Cs-137 from portland type I cem ent paste specimens leached at

20, 30, 40, 50, and 70C 43Figure 4.8 Arrh enius plot of the diffusion coefficients from portland cem ent pastecontain ing Cs-137 and Sr-85, leached at 20, 30, 40, 50, and 70C 44Figure 4.9 CF L for Sr-85 from portland cem ent paste containing radioactive tracers attem pera tures ranging from 20C to 70C 45Figure 4.10 CF L for Cs-137 vs time from p ortland I cem ent containing 5 wt% sodiumsulfate at 20, 40, 50, and 60C 46Figure 4.11 Arrh enius plot of Cs-137 leached from cem ent containing 5 wt% sodium sulfatcshowing diffusion coefficients as a function of the reciprocal temperature inKelvins. Experim ents were conducted at 20, 40, 50, and 60C 47

- V I I -


9/151


Figure 4.12 Sr-85 cumulative fraction leached vs time from portland I cem ent contain-ing 5 wt% sodium sulfate at 20, 40, and 50C leached in deionized water . . 48Figu re 4.13 A rrh eni us plot of Sr-85 showing diffusion coefficients as a function of th ereciprocal tem per atur e in Kelvins 48Figure 4.14 Cs-137 cumulative fraction leached vs time from portland I cem ent containing5 wt% sodium sulfate at waste form volume to surface area (V/S) ratios of 0.42,0.85, and 1.85. Samples we re leached in deionized water at 20C 49Figure 4.15 Curves are calculated from the finite cylinder model, using selected

diffusion coefficients from Table 4.1 to show ranges of leaching obtainedund er different test conditions. Also shown are data points for t hebaseline leaching experim ent run at 20C in 1.3 liters of water 52Figure 4.16 D ata from th e two experim ents that gave the greatest amo unt of accelera-tion for Cs-137 are compared to the baseline data. Each point representsthe average value of triplicate specimens 53Figure 4.17 Ave rage cumu lative fraction release curves for Cs-137 show that leachingat 60C is substantially lower tha n at 50C during a semidynamic leach test . . 54Figure 4.18 Leach ing of Cs-137 is lower at 60C than it is at 50C in static leach testsof cem ent/sulfate waste forms 55Figure 4.19 Leaching of sodium from cem ent/sulfate specimens is lower at 60C than at50C in static leach tests 55Figure 4.20 Leaching of potassium is lower at 60C than at 50C in static leach tests . . . 56Figure 4.21 Da ta for Cs-137 from cement/sulfate specimens during static and semi-dynamic leach tests at 50 and 60C. T he solid lines ar e calculated from

the finite cylinder model 57Figure 4.22 Com parison of Cs-137 releases from small cem ent/sulfate specimens(2.5 x 2.5 cm) leached at 20C and at 50C in different leachan t volumes.Cs-137 releases are relatively insensitive to leachant volumes 58Figure 4.23 Da ta from a leach test using 2.5 x 2.5 cm cem ent/sulfate specimens at 50Cin 3 liters of water with daily leachant replacem ent. Th e finite cylindermodel may be overestimating releases after CF L= 0.6 5 59

-V1U-


10/151


Figure 4.24 Cs-137 results from the optimized acce lerated leach test for cemen t/sulfatespecimens are com pared to results from th e baseline test. Th e accelera-tion factor is approximately 18 59Figure 4.25 Da ta from two sets of portland cem ent specimens containing 5 wt%sodium sulfate. Th e optimized accelerated test is indicated by the filledsquares and has an acceleration factor of about 17 com pared to th ebaseline data. Th e finite cylinder model result is shown as the linethrough the optimized test data 62Figure 4.26 Relea ses of Sr-85 from cement/sulfate waste forms during two semidynamicexperiments run at 60C but having different leac hant volumes. Th e finitecylinder model was used to gen erate the two curves shown 63Figure 4.27 Sr-85 releases from cement/sulfate waste forms during static experim ents at50 and 60 G Th e solid line is the result of the finite cylinder mode l forthe first 6 days of leaching at 60C 64Figu re 4.28 Cs-137 and Sr-85 leaching from static expe rime nts that are CC>2-free or

exposed to air 65Figure 4.29 Calcium cum ulative fraction leached at 20 and 50C in static leach testsopen to air and CC^-fre e. Specimens are portland ceme nt containing5 wt% sodium sulfate 66Figure 4.30 Alkalinity in leac hate at 20 and 50C in static tests op en to air and CO2-

free 67Figure 5.1 Cum ulative fraction leached of Cs-137 plotted against the squ are root oftime. If diffusion is th e leaching mechanism th en th e plot should be linearto abou t 0.20 cum ulative fraction relea se. After that, it should slowly

curve down 70Figu re 5.2 Leach ing of Cs-137 from cem ent/ash waste forms. Th e solid lines arecalculated from th e finite cylinder mod el 71Figure 5.3 Com parison of Cs-137 releases from cemen t waste forms containingincinerator ash with specimens containing sodium sulfate. The curve is themodeled data for the cement/sulfate specimens 72

-IX-


11/151

LIST OF FIGURES - continuedPageFigure 5.4 Long-term comparison of Cs-137 releases from cem ent waste formscontaining incinerator ash with specimens containing sodium suifate 72

Figure 5.5 Cumulative fraction leached of Sr-85 is plotted against the square root oftime for triplicate specimens of portland cem ent containing incinera tor ash . . 73Figure 5.6 Th e finite cylinder model does not adequately model the Sr-85 data fromcem ent/ash waste forms 74Figure 5.7 Cs-137 cumulative fraction leached vs time from portland I cem entcontaining 15 wt% incinerator ash at 20", 40, 50, and 60C in deionizedwater 76Figure 5.8 Averaged Cs-137 cumulative fraction releases from cem ent/ash specimensshown for the first nine sampling intervals. W hile the ord er of relea sefrom the 20, 40, and 50C specimens remain unchanged, the releasesfrom the 60C specimen were anom alous for the last two points 77Figure 5.9 An Arrh enius plot of the diffusion coefficient for cem ent/ash specimensleached a t 20, 40, 50, and 60C . T he re is no statistically significant dif-ference in D e caused by temp erature under these test conditions 77Figure 5.10 Sr-85 cumulative fraction leached vs time from portland I cem ent contain-ing 15 wt% incinerator ash at 20, 40, and 50C in deionized wate r 78Figu re 5.11 An A rrhe nius plot of the diffusion coefficient D e showing that Sr-85leaching at 60C is lower than expected , 79Figu re 5.12 Size effect on relea ses of Cs-137 for cem ent/ash wa ste forms. Size isexpressed as the ratio of waste form volume to surface area (V/S), withthe smaller waste form having the lower ratio 80Figure 5.13 Re leases of Cs-137 from portland cem ent containing 15 wt% inc ineratorash. Static and semidynamic leach tests we re run at 60C. Th e finite

cylinder mo del diverges from the data after 12 days 83Figure 5.14 Altho ugh leaching from a static test at 50 or 60C is faster than at 20C,the CF L values taken for tests at 50 and 60C are almost identical 84Figure 5.15 T he combined effect of small size, elevated temp eratu re, increased volumeof leachant, and increased frequency of replacement gave results that werenot greater than just small size and elevated tem pera ture. Specimens we reportland cem ent containing 15 wt% incinerator ash 84

-x-


12/151


Figu re 5.16 Ar rhenius plot for Cs-137 releases from cement/a* h specimens at 20, 40,50, and 60C. Changes in the volume of leachant m ade no difference toleaching at 60C, but at 50C smaller volumes (or no leachant replace-me nts) could reduce leaching 85Figure 5.17 Rele ases of Cs-137 from triplicate specimens of portland cemen t containingincinerator ash are shown for the optimized accelerated test and for thebaseline. Th e acceleration factor is approximately 6 86Figure 5.18 Relea ses of Cs-137 from accelerated tests are plotted against releases fromthe baseline test. Th e results produc e a linear plot in two segm ents. Th ebreak is caused by the change in sampling interval during th e baseline test . . 86Figure 5.19 Sr-85 releases from cem ent/ash specimens during static tests run at 50 and60C 88Figure 5.20 Rele ases of Sr-85 during static and semidynamic tests at 60C. Th e finitecylinder model fits the data during the daily leachant replacement intervals,but ove restim ates it tor longer intervals 89Figure 5.21 Ar rhenius plot of Sr-85 diffusion coefficients for cem ent/ash specim ens

showing that increasing the leachant volume to 6.5 liters increased releasesat 60C 90Figure 5.22 Da ta from the accelerated test and from the baseline test are plottedtogeth er. T he acceleration factor is approximately 17 91

5.23 Triplicate data sets from the optimized accelerated test and the baselinetest. O ne specimen is modeled with the finite cylinder model but no ne fitthe model very well 91igure 5.24 Plotting the CF L data from the baseline experime nt against the CF L datafrom the accelerated test, a linear plot should result if the leachingmechanism has not changed. Th e break in slope is caused by a chan ge insampling intervals during the baseline test, but each segmen t is linear 92igure 5.25 Sr-85 relea ses from cemen t/ash waste forms w ere no different for speci-mens exposed to atmospheric carbon dioxide than they were for specimensfrom which carbo n dioxide was excluded 93

Figure 5.26 Dissolved strontiu m released from cemen t/ash waste forms in static leachtests shows no effect of carbo nation at 20* or 50C 94-xi-


13/151


Fig ure 5.27 Dissolved calcium release d from cem ent/ash waste forms in static leachtests shows no effect of carbonation, although the 20C CO2-free experi-me nt had slightly higher calcium conce ntrations 94Figure 6.1 Cs-137 cumulative fraction leached vs time from bitume n containing40 wt% sodium tetraborate at 20, 40, and 50C leached in deionizedwater 98Figure 6.2 Leach ing of cobalt is enh ance d in a leachant containing 100 ppm ED TA .The cesium releases are slightly suppressed by the presence of additionalsodium 99Figure 6.3 Averag e values of Cs-137 cumulative fraction leached from experim ents

with daily replacem ents of leachan t and with less frequent replace me nts . . . 100Figure 6.4 Re lease of Cs-137 from bitumen is related to the am ount of salt loading . . 101Figure 6.5 Relea ses of Cs-137 from specimens with various waste loadings during the

early portion of the leaching experim ent 102Figure 6.6 Wh ile leaching of Sr-85 from specimens containing 40 wt% salt was com-parable to Cs-137, releases from the 20 wt% and 30 wt% specimens waslower 104

Figure 6.7 Re leases of cobalt are much lower than for Cs-137 and Sr-85, but aresystematic with loading up to 400 days 104Figure 6.8 Value s of p H of leachates are typically around 9 at 20G Da ta is forbitumen containing 30 wt% sodium tetrabora te 105Figure 6.9 Schem atic showing the mechanism of leaching of a soluble salt incor-pora ted in bitumen 107Figure 6.10 An unleached surface of a bitumen specimen containing 40 wt% sodiumtetra bo rate . Outlines of salt grains can be seen 108Figure 6.11 After leaching, the surface of a 40 wt% loaded bitumen specimen ischaracterized by swelling and burst blisters where the saturated saltsolution broke through the bitumen 109Figure 6.12 At a magnification of 1000 times, a single blister with num erou s holesillustrates how satu rate d salt solution is leached 109

-xii-


14/151


15/151


Figure 7.5 Cs-137 releases from VES containing 40 wt% salt during an experim entwith daily replacem ents of the leachant. Th e finite cylinder model fits th edata accurately 120Figure 7.6 Ave rage releases of Cs-137 at 20, 40 and 50C during the first 11 days ofleaching for VE S waste forms containing 40 wt% sodium sulfate 121Figu re 7.7 A rrhe niu s plot of diffusion coefficients for V ES leach ed at 20, 40 and50C. Th ese specimens contained 40 wt% sodium sulfate 122Figure 7.8 Ave rage releases of Cs-137 from three different sized specimens of VE Scontaining 40 wt% sodium sulfate 123Figure 7.9 Re leases of Cs-137 from VES/sodium sulfate at 60C from static and semi-dynamic tests are indistinguishable from each oth er 127Figure 7.10 Leaching of Cs-137 in semidynamic tests using 6.5 liters of wa ter at eachinterval showed relatively uniform behavior and a definite temperatureeffect 128Figure 7.11 Th e finite cylinder model fits data from expe riments run at 60 and 20C,

during the daily replacem ents of leacha nt. W hen th e intervals of replace-me nt are longer, the model overestimates releases 129

-xiv-


16/151


17/151

EXECUTIVE SUMMARY

The prediction of long-term teachability of low-level radioactive waste forms is anessential elem ent of disposal-site performance assessment. This report describes experimentsand modeling techniques used to develop an accelerated leach test that meets this need.

The acceleration in leaching rates caused by the combinations of two or more factorswe re experimentally determ ined. The se factors we re identified earlier as being able toindividually accelerate leaching. They ar e: elevated te m pera ture, th e size of th e waste form, theratio of the volume of leachant to the surface area of the waste form, and the frequency ofreplacement of the leachant.

The solidification agents employed were ones that are currently used to treat low-levelradioactive wastes, namely portland type I cem ent, bitumen , and vinyl ester-styrene . Th esimulated wastes, sodium sulfate, sodium tetraborate, and incinerator ash, are simplifiedreprese ntatives of typical low-level waste streams. Experim ents determ ined the leachingbehavior of the radionuclides of cesium (Cs-137), strontium (Sr-85), and cobalt (Co-60 or Co-57) from sev eral different form ulations of solidification agents and waste types. Leachingresults were based upon radiochemical and elemental analyses of aliquots of the leachate, andon its total alkalinity and pH at various times during the experim ent (up t o 120 days). Solidphase analyses were carried out by Scanning/Electron Microscopy and Euergy DispersiveSpectroscopy on the waste forms before and after some leaching experiments.

Temperatures up to 50C accelerated leaching from portland cement containing 5 wt%sodium sulfate, but above this temperature the releases of some elements declined due tochanges in the structur e of th e cem ent. At 50C, the leaching of Cs-137 was accelerated byabou t a factor of 18. Acc elerated leaching of Sr-85 required larger volumes of leacha te andgave an acceleration factor of 17. Th e release of Sr-85 can be mod eled if leaching is notinfluenced by secondary reactions such as carbonation, and larger volumes of water are requiredto prevent these reactions.

-XVH-


18/151

The accelerated leaching conditions were optimized for portland cement containing15 wt% of incinerator ash. Th e greatest accelerating factor for Cs-137 occurred at 50C in asemidynamic tes t with 1.3 liters of water, giving an accelera tion fac tor of 6. T he finite cylindermodel does not adequately describe long-term leaching of Cs-137 from this type of waste form.Temperatures up to 50C accelerated the release of Sr-85 with an optimum acceleration factorof 17; at 60C, the rate decreased, although an increased volume of leachant brought the rateup slightly. At 50C the presenc e of atmosph eric carbon dioxide had no effect on the leachingof radionuclides. Th e leaching of Sr-85 can only be m odeled for th e first five days of th eexperim ent with the finite cylinder mod el. Therea fter, releases are overestimated . As with Cs-137, a model is needed with an adsorption term that is a function of time.

Th e leaching of bitumen wastes cannot be accelerated uniformly. Elec tron microscopyconfirmed the hypothesis in the literature on th e physical mechanism of leaching. Th e complexrelationships of leaching from bitumen waste forms are discussed.

Leaching from vinyl ester-styrene containing sodium sulfate was accelerated by tempera-ture of 60C and by using 6.5 liters of water at each re placem ent interval. This com bination offactors increased the leaching rate by a factor of 17. Leaching of this waste form can bemodeled by diffusion.

A variety of mathem atical mo dels w ere review ed, including mode ls for diffusion from asemi-infinite m edium, and from a finite medium . Th e results of our ex perime nts we re m odeledusing diffusion from a finite cylinder to see whether the mechanism of leaching was consistentthroughout the leaching cycle and could be compared with results of a standard (baseline)semidynamic leach test.

Accelerated leaching can be performed for cement and vinyl ester-styrene waste forms ina way that does not change the mechanisms of leaching and gives results that can be modeled.A significant reduction in the time required to show consistent leaching properties for a majorfraction leached was demonstrated.

This work will lead to the development of an accelerated leach test that can be used forcement-based waste forms, and for thenmosetting polymers.

-xviu-


19/151

1. INTRODUCTION

This topical report presents the results from experimental and modeling efforts to providea basis for the development of an accelerated leach test to predict long-term ieachabilities oflow-level wa ste forms. Th e results of an expe rimental investigation of com bined factors thataccelerated leaching are discussed. Th e investigation is based on work, detailed in a previoustopical report [1], that discussed changes of leach rates caused by individual factors such astemperature, the size of the waste form, and the chemical composition and volume of theleachant.

The approach taken to study combined leach-rate acceleration factors requires a series ofxperiments using two or more of the factors that previously were identified as having a positive

ect on leach rates. Th e results of these experiments are then modeled to determ ine if theeaching mechanism is the same as that observed in earlier "baseline" experim ents. In turn, theodels can be used to determine if leaching is being suppressed by some experimental artifact.

The solidification agents used are ones that are currently used to treat low-level waste

Th e simulated wastes that are solidified (sodium sulfate, sodium te trab ora te and

This two-tiered approach, using experimental results for modeling and then optimizing

em ent and on vinyl ester-styrene. Th ere is no model for bitumen waste forms.

Leaching of radionuclides from disposed low-level radioactive waste is the first event in aenc e that must be understood to produce reasonable risk assessments. For wastes treated

- 1 -


20/151

with a solidification agent, chemical reactions between the solidification agent and componentsof the waste ren der some radionuclides insoluble and, therefo re, immobile. O the r eleme nts arenonrea ctive and may be highly mobile. To ade quately perform radiological assessments, thereliable prediction of the long-term leaching behavior of waste forms depends highly on theintegrity of th e waste form in its disposal environm ents. Leaching studies are a key elem ent inunderstan ding these processes (and others) that lead to releases. This knowledge, in turn, isimpo rtant to developing means of preventing releases. M oreover, leaching may be an im portantdegradation process for many of the structural materials considered for engineered disposalunits. Studies of mechanisms that transport wa ter through these barrier materials (andtransport radionuclides out) are closely related to studies of leaching mechanisms.

Determining the mechanisms of leaching is the first step in defining what mathematicalmodels (if any) can describe leaching in a way that can be projected beyond the available data.A variety of leaching mechanisms have been proposed for different types of waste forms, butvery few have been incorporated in to com puter program s. Fortun ately, the mathem atics of dif-fusion have been so reduced that com puter programs can be written relatively easily. Diffusionthrough a porou s medium is the most common mechanism for low-level waste forms. W he reoth er m echanisms have been identified, or wh ere othe r processes affect diffusion, differentcom puter programs are required. Th e scope of the accelerated leach test project was not todevelop com puter program s to model leaching. A thorou gh understan ding of leachingmechanisms will lead to more sophisticated modeling and a better ability to provide long-termpredictions for risk assessments.

-2 -


21/151

2. MATERIALS AND METHODS

This section summarizes the materials used for fabricating the waste form, the leach testme thods, and the analytical me thods. Th e solidification agents, portland type I cem ent,bitumen, and vinyl ester-styrene copolymer were primarily selected to represent a cross-sectionof material properties. Th e waste streams chosen are sodium sulfate wastes genera ted atBoiling Water Reactors (BWRs), boric acid waste generated at Pressurized Water Reactors(PWRs) and incinerator ash, which is becoming a major waste stream as advanced volume-reduction processes become common.

Th e choice of solidification agents was based on two criteria: (i) that they we re either inuse or being considered for use in low-level waste management, and (ii) that they cover a rangeof materials prop erties. By providing a detailed knowledge of the leaching bevahior of severaltypes of materials, e.g., hydraulic cement, thermoplastic binders, and thermosetting polymers, itis anticipated that the results would apply in general to solidification agents that may bedeveloped in the future.

The following test specimens were investigated:

Solidification age nts. Solidification age nt with radioactive trace rs. Solidification age nts contain ing simulated waste. Solidification agents containing simulated waste with radioactive tracers.

The samples containing radioactive tracers were used to investigate the leaching behaviorof ndion uclid es of cesium, strontium and cobalt from th e matrices. Th e other samples we reused to study the leaching of the components of the solidification agent and nonradioactiveelements from the waste.

-3 -


22/151

Table 2.1 summarizes the types of waste forms containing simulated wastes that wereselected for study. Table 2.2 lists their compositions, corresponding to th e abb reviated summaryin Table 2.1.

Table 2.1Summary of Types of Waste Forms

Portland I Cement5 wt% Sodium Sulfate

Portland I Cement15 wt% Incinerator Ash

Bitumen20 wt% Sodium Tetraborate30 wt% Sodium Tetraborate40 wt% Sodium Tetraborate

Vinyl Ester-Stvrene20 wt% Sodium Sulfate30 wt% Sodium Sulfate40 wt% Sodium Sulfate

-4-


23/151

Table 2.2Composition (weight %) for Each of the Waste Form TypesContaining Simulated Waste

Portland Type I Cement/Sodium SulfateActivity

Vinyl Ester-Styrene/Sodium SulfateActivityComponent

Cement PowderSodium Sulfate(anhydrous)WaterCo-57Cs-137Sr-85

Weight %65530---

Portland I Cement/Incinerator i

ComponentCement PowderIncinerator AshWaterCo-57Cs-137Sr-85

Weight %601525---

(uCi)

-

6.36.312.5

Activityf/iCi)

10010020 0

Vinyl Ester-Styrene/Sodium Sulfate

Component

Vinyl Ester-Styrene MonomerSodium Sulfate(anhydrous)WaterCatalystPromoterCo-57Cs-137Sr-85

Weight %

77.320.00.62.0.08-

ActivityfeCi}_----16.516.525.5

ComponentVinyl Ester-Styrene MonomerSodium Sulfate(anhydrous)WaterCatalystPromoterCo-57Cs-137Sr-85

Weight %54.739.31.81.50.06

-Vinyl Ester-Styrene/Sodium Sulfate

ComponentVinyl Ester-Styrene MonomerSodium Sulfate(anhydrous)WaterCatalystPromoterCo-57Cs-137Sr-85

Weight %67.630.00.61.70.07--

Bitumen/Sodium Tetraborate

Component

BitumenSodium

Weight %

60.040.0Tetraborate (anhydrous)Co-57Cs-137Sr-85---

iifca

---

1 00100200

Activity(ud)_

--16.516.525.5

ActivityiixCi)m-91911 82

-5 -


24/151

Table 2.2 (continued)

Composition (weight %) for Each of the Waste Form TypesContaining Simulated Waste

Bitumen/Sodium TetraborateActivity

ComponentBitumenSodiumTetraborateCo-57Cs-137Sr-85

Weight %70.030.0(anhydrous)-

JteCi]-393978

Bitumen/Sodium TetraborateActivity

ComponentBitumenSodiumTetraborateCo-57Cs-137Sr-85

Weight %80.020.0(anhydrous)

- 1 001 00200

-6 -


25/151

2.1 Waste Form Types

Studies were made to select the composites of solidification agent/waste material that arechemically compa tible. Form ulations of the waste form were optimized to m aintain physicalintegrity when immersed in water during the leaching experiments.

2.1.1 Portland I Cem ent Containing Sodium Sulfate as Simulated W aste. Sodium sulfatewaste is a product of ion exchange resin regen eration at BW R po wer stations. It is usuallycon centra ted by evap oration to approximately 22 wt% solids con tent. Fu rther , evap oration ofthis con cen trate to dryness results in sodium sulfate deca hydrate (G laub er's salt). Glau ber's saltdehydrates to anhydrous sodium sulfate at 32C.

Up to approximately 45 wt% sodium sulfate can be solidified with cement. However,aste forms containing more than approximately 8 wt% sodium sulfate were unstable when

immersed in deionized water, disintegrating before com pletion of a 90-day imm ersion test. Aormulation of 5 wt% sodium sulfate, 30 wt% water, and 65 wt% cement was selected for

2.1.2 Portland I Cem ent Containing Incinerator Ash as Simulated W aste. Up to 50 wt%

ever, on curing specimens swell developing large voids. A formulation of 15 wt% ash, 25

Incinerator ash was obtained from the waste incinerator of the Tennessee ValleyTh e major com ponen ts of incinerator ash are typically uncom busted ca rbon,

oxide (AI2O 3), ferric oxide (Fe2C3), and silicon dioxide (SiC^ )- T he ash was passed

-7 -


26/151

2.1.3 Vinyl Ester-S tvrene C ontaining Sodium Sulfate as Simulated W aste. Mixtures of upto 60 wt% dry sodium suifate with vinyl ester-styrene monomer polymerized satisfactorily topro du ce waste forms with hard surfaces. T he leaching of sodium sulfate from solidified wasteforms was used to select a composition which maximized the amount of incorporated wastewhile minimizing leaching. Th e amo unt of sodium that was leached increase d rapidly at wasteloadings above 40 wt%; therefore, this was the maximum loading used.

2.1.4 Bitumen Containing Sodium Te trab ora te as Simulated W aste. Boric acid wastegen erate d at a typical PWR Plant contains approximately 12 wt% boric acid in aqu eou ssolution. Solidification of the waste with bitum en requires tha t the wate r is eva pora ted off.Because boric acid dehydrates at 160C, the loss of water causes foaming of the bitumenmixture. A satisfactory pretre atm ent is to neutralize the boric acid waste stream with sodiumhydroxide to pH 9.3: sodium tetrabora te is the predominant product [3]. Evaporation of thesolution to dryness results in the hydrated crystalline salt, sodium tetraborate decahydrate(borax). Th e borax must be further dried at 200C to produce th e anhydrous tetrab orate .

A solid waste form can be produced by mixing the anhydrous sodium tetraborate withmo lten bitum en. W e used waste loadings of 20 wt% , 30 wt% and 40 wt% of sodiumtetraborate in bitumen.

2.2 Prepa ration of Samples

Cylindrical samples were prepared for leaching experiments with approximate dimensionsof 4.8 crn diamete r and 6.4 cm length (V /S=0 .84). Based on past expe rience , this size isconvenient for laboratory studies. Additional samples were prepared for size-scale studies,including smaller cylinders, approximately 2.5 cm diameter by 2.5 cm height (V/S=0.42), andlarger cylinders approximately 10 cm diameter by 13 cm height (V/S=1.85).

Radioactive samples prepared for leach testing had the radioactive tracers Co-57, Cs-137,and Sr-85 incorpora ted into the waste. Th e low energy gamma ray em itted by Co-57

-8-


27/151

= 270 days) is better suited to the automated Nal counting system than is Co-60, and alsoworks well with the intrinsic Ge counting system.

2.3 Leaching TestsThe ANS 16.1 Leach Test is a semi-dynamic leaching test in which the leachate is

replaced periodically after intervals of static leaching [2]. Specimens are placed into th eleachant solution in such a way that all the external surface area is directly exposed to thesolution. Specim ens are leached in individual containe rs containing a ratio of 10 cm betw eenthe volume of the leachant and the external geometric surface area of the specimen unlessotherw ise specified. Specimens are usually tested in triplicate to dete rm ine th e variation inleaching. Th e results are expressed as incremental fraction rele ase, as cumu lative fractionrelease, or as a release rate, to facilitate alternative methods of treating the data.

Th e leac hant is typically distilled wate r with a con duc tance of less than 5/xmhos/cm. T hesampling interval was modified to give more frequent intervals than specified in the ANS 16.1Method and, in some cases, to extend the duration of the test beyond the 90-day standard.

Our study differed from the ANS 16.1 test procedure by changes in the leachantreplace m ent intervals. Extrem es ranged from daily replacem ents to static tests with noreplac em ents. O the r variations included experimen ts run at elevated tem pera tures in anenvironmental chamber (Forma Scientific) with strict temperature controls.

2.4 Analytical Methods

Leachates are analyzed for a variety of materials, depending on the composition of thesolidification agent and the simulated waste. Cem ent leachates are subjected to the m ostanalysis since reactions within the matrix cause significant differences in leaching amongdifferent ele m ent s. Specific analytical methods are given below.

-9-


28/151

2.4.1 Radiochem ical Analysis. Th e radiochemical com pone nt of the leacha te, such asOs-137, Sr-85 and Co-60 (or Co-57) is analyzed by gamma-ray spectroscopy using an intrinsicgermanium detector or a sodium iodide detector in accordance with the methods described inASTM D3648-78 [4] and ASTM D3649-78 [5].

2.4.2 Elem ental Leac hate Analysis. Analysis of leachates for non-radioactive elem entsis conducted with standard methods such as ASTM E663 [6] and those in Analytical Methodsfor Atomic Absorption Spectrophotometry, revised January 1982, Perkin-Elmer Corporation,Norwalk, CT, [7J.

2.4.3 Alkalinity Mea surem ents. Th e total alkalinity of leachates is me asured by titrationto the phenoiphthalein end point according to Method Number 403 from Standard Methods forthe E xamination of W ater and W aste W ater. 15th edition, 1980 [8].

2.4.4 pH M easurements. Th e pH of leachates is measured using ASTM D1293 [9] witha combination probe.

2.4.5 SE M /ED S. W aste forms are analyzed before and after leaching by Scann-ing/Electron Microscopy (SEM) to observe any changes in morphology, and by EnergyDispersive Spectroscopy (EDS) to determine the elemental ratios in profile and on the surfacesof the waste form. Th e methods used are discussed by Goldstein an d Yakowitz in PracticalScanning Electron Microscopy, Electron and Ion Microprobe Analysis. (1975) [10].

-10-


29/151

3. LEACHING MECHANISMS AND MODELS

3.1 Introduction

Standard leach tests provide no assurance that the optimal conditions for leaching areused. M oreo ver, while many tests assume that diffusion is the op erativ e mechan ism of leaching,typically no effort is made to determ ine the actual mechanism. Th e goal of an accelerated testis to measure the maximum leach rates for a waste form material in the shortest time, byconducting the test under th e most favorable conditions. Th e effect of many leach testcondit:ons we re measu red and repo rted earlier [1]. Determ ination of the optimum cond itionsfor t: developm ent of an accelerated test now requires the application of specific models topredict radionuclide releases from specific waste form materials under the leaching conditionswhich we re used. Some modeling concepts were reviewed earlier by Dou gherty and C olombo[11]. In this section, the application and implications of mathematical modeling will beexamined to find the most favorable leaching conditions for an accelerated leach test.

3.2 M ethods of Presen ting Leaching Da ta as a Function of Tim e

It should be noted that proper consideration of radioactive decay has been assumed forall of the following discussion.

3.2.1 Tabu lar and Graphical M ethods. Several ways of graphically presenting experime n-tal data on the leaching of radionuclides from waste form solids have been used by manylaboratories worldwide. The International Atomic Energy Agency (IAE A) and the Am ericanNuclear Society (ANS) proposed specific leach tests and methods of expressing the results.Table 3.1 lists ways to present leaching results as a function of time.

-11-


30/151

Table 3.1Graphical presentations of leaching data as a function of time.

Type

ABCDE

Amount Leached

CFL(V/S)CFLCFLlog CFLIFL

Time Function

ttl/ 2tlogtA(tV2)

a n = am ount leached in leaching interval nAo = initial am ount of material being leachedCFL = Ea n/Ao cumulative fraction leachedCFL (V/S) = cumulative fraction leached normalized to specimen sizeIFL = a n/Ao incremental fraction leachedt = total leaching timeV = sample volume, cmS = sample geom etric surface area, cm 2

The IAEA recommends presenting the results as a plot or in tabular form as shown inTab le 3.1, Types (A and B ). Both of these forms of prese ntation imply the use of a specificmathem atical modeling mechanism for leaching (discussed late r). Th e sample size "normalizingfactor" V/S is op erativ e only if the implied leaching mechanism is bulk diffusion from a semi-infinite m edium an d all its limitations apply. Also, th e implication tha t CF L is a linear functionof t 1 / 2 only works if diffusion is the op era tive mechanism.

Presentations of the leaching data according to Types C and D in Table 3.1 were used inthis report to avoid the inference of a specific leaching mechanism until it is demonstrated as

-12-


31/151

being operative. Th e plots of log CF L vs log t are especially convenient forms of presentingdata that span several orders of magnitude.

Finally, presentation Type E of Table 3.1 was recommended as a means of determiningthe diffusion coefficient, since the data points are not coupled [12].

Correlation diagrams in which CFL from one experiment is plotted against the CFL of areference experiment at identical points in elapsed time have proven useful in demonstratingthat changes in the leaching mechanism were not induced by changes in the leaching conditions[1J. If the data from an experim ent lie on a straight line, the n th e two sets of data linearlycorr elate . Presumably, data taken a t different tem peratu res will be linearly correlated as long asthe m echanism of leaching is unchanged. The slope of the line, for linearly correlated data,provides a relative measure of the leachability compared to that of the reference experiment.

3.2.2 Em pirical Equations. Empirical and semi-empirical equa tions for fitting data we reused extensively to reduce large amounts of experimen tal data to a few param eters. Thesemethods are quite accurate within the bounds of the experimental data but useless forextrapolation beyond the measured limits. Neve rtheless, these equations are convenient m eansof data reduction, where mathematical solutions to the appropriate transport equations are notreadily available.

Simple exponential equations were used by Godbee and Joy for their data on theleaching of asphalt-sludge leaching [13]. Rece ntly, Cote e t al. [14] used a sem i-empiricalequation to evaluate the leaching of heavy metals (As, Cd, Cr, and Pb) from cement-basedwaste forms. Th e calculated param eters were used to determine the relative merit of the wasteforms, not to predict the effects of long-term leaching.

3.2.3 M athema tical Solutions to Mass Tran sport Equation s. Mod els based on masstransport theory that have been validated with experimental data are an excellent means ofestimating the am ounts of material released by solidified w aste. Th e mathem atical theory oftran spo rt by diffusion from solids is based on Fick's hypothesis that the diffusion rate is

-13-


32/151

proportional to the concentration gradient [15,16]. The fundamental partial differentialequation for diffusion is:

= D e V 2 C (1)dtwhere,

V 2 = Laplacian Op eratorin one dimension,

C = conce ntration of species,t = time,D e = effective diffusion coefficient in a porou s medium .

Mathematical solutions to the transport equation (1) have been applied to the leaching ofradionuclides for waste solids [17].

3.2.3.1 Analytical Solutions for Diffusion. G en era l solution s to th e diffusion equ ation(1) can be obtained , provided that the diffusion coefficient is con stant. T he constan t coefficientassumes that there are no significant changes with time in the physical or chemical structure ofthe waste form during the leaching process and no dependence on the concentration of thediffusing spe cies. In most systems the diffusion coefficient depen ds on the con centra tion of thediffusing substance in the solid. How ever for dilute solutions, the dep end enc e on th e con-centration of the diffusing species is slight and the diffusion coefficient can be assumed to beconstant.

Th e Semi-Infinite M edium Mo del. Th e exact form of the solution to the mass transportequa tion of diffusion dep end s on th e initial and boundary conditions of the prob lem. A simplecase is that of a semi-infinite solid with a constant diffusion coefficient where the cumulativefraction leached is predicted as:

-14-


33/151

Ean S rD - t= 2 _f_AQ V [ x

where,E a n = the total amoun t of radioactive material released in all leaching periods up to

time, t,AQ = the initial am ount of radioactive material,V = the volum e of the waste form,S = the surface area of the waste form,D e = th e effective diffusion coefficient in a porous medium .

he semi-infinite medium solution has been used extensively to demonstrate that the operatingechanism for leaching is diffusion where graphical presentation of CFL vs t1>/2 is linear whenFL is less than a bo ut 0.20. Diffusion coefficients th en can be calculated from th e slope of

In practice, it is necessary to add an intercep t to equa tion (2) since extrapolation to

Finite Medium M odel. Finite samples that release radioactive ma terials by a diffusion1 / 2 up to a CF L of about 0.20. Exact

r finite shape s we re described by Crank [IS]. Since waste forms and samples for

Th e solution of the mass transport equation for a cylinder that is uniform and

oo coEan 32 e - t D e ^ + ( 2 n- l) V / 4 /2 ]

2(/?)2(2n-l) 2(/?m )

-15-


34/151

where,r = the cylinder radius, cm,/ = th e cylinder half-height, cm, and/9 m = the positive roots of the zero-orde r Bessel function, Jo(r/0m) = 0 cm .

The convergence of equation (3) is slow but computer programs were developed to determinethe cumulative fraction leached. Th e ANS 16.1 leach test standard provides a tabular methodof calculating De from leaching data for cylindrical waste forms [2]:

D e = G d 2/t (4)where,

G = dimensionless time factor [2]d = cylinder diam eter

and is listed for 1/d ratios of 0.3 to 5.0 and for a C FL of 0.20 to 0.99. T he dimension 1 is thehalf-height of the cylinder.

The most useful application of the finite cylinder equation (3) would be to predictleaching at longer times using optimal values of D e , especially with a graphical presentation ofthe data and the comp uted leach curve. Deviation of the predicted curve (using a constantdiffusion coefficient) from the data would most likely be due to physical or chemical changes inthe waste form or changes in the leaching conditions.

3.2.3.2 Co ncen tration De pen den t Diffusion. In most leaching environ me nts thediffusion coefficient depends on the concentration of the diffusing species and the ionic strengthof solution. T he diffusion coefficient can be written in the form:

D e(C) = DO { l - f (Q } (5)

-16-


35/151

where,D e(C ) = the diffusion coefficient at con cen tration , C,

= the diffusion coefficient at infinite dilution or t= 0 ,C = the conc entration of the species,f(C) = function of C.

How ever, in dilute solutions, the depen den ce is slight or negligible. To deve lop an acceleratedleach test, the accelerating factors of leachant volume to sample surface area, and the frequencyof leachant replacement will be chosen to avoid such concentration effects on the diffusioncoefficient.

The effect of changes in concentration is of interest in predicting how waste forms leachin the disposal environm ent. Som e insight can be gained by static test results, wh ere theconcentration buildup of all the leached species is continuous.

3.2.3.3 Skin Effects of the Sam ple's Surface. Many solids have a surface skin that canbe observed visually and which exhibits properties that differ from those of the bulk material.Evidence of a surface layer was reported for cement based-waste forms by Fuhrmann andColombo [19]. The effects are expected to be secondary to bulk diffusion, but of appreciablesignificance as an explanation for observed changes in the diffusion coefficients for Sr-85 andCs-137. In a private communication betwee n H . Go db ee and R. Pietrzak, a model was usedthat incorporated the concept of rapid sorption followed by desorption at the surface skin toeffectively model leaching data from cement waste forms.

3.2.3.4 Diffusion + Reac tion (K^V Mass trans por t of th e diffusing radionuc lides fromsolidified w aste can be limited by local instan tane ous equilibrium. Th e simplest case is wh erethe adsorbed species (S) is proportional to the mobile species (C):

S = K dC (6)

-17-


36/151

The solutions for the semi-infinite media and the finite cylinder are identical to equations (2)and (3) respectively, where:

D e = Df(6/r2)[l/(l+ pjty*)] (7)

where,S = constrictivity,r^ = tortuosity,Pa = apparent density,e. = porosity,Df = diffusion coefficient in free liquid,Kj = sorption coefficient, mL/g.

Models for diffusion and kinetically controlled reactions, interface resistance, irreversiblereactions, dissolution, desorption, and moving boundaries were reported for numerous practicalapplications. Several such models were reviewed in an earlier report [11].

3.2.3.5 Anomalous Transport. The diffusion of radionuclides from several types ofsolidified waste forms (such as with bitumen containing soluble waste) cannot be describedadequately by diffusion transport equations with a constant boundary. Penetration of water intothe waste form eventually causes swelling and the characteristic leach curve shows time- andwaste loading-dependent stages. Empirical methods such as polynomial curve fits can quantifythe radionuclide leach characteristics for relative comparison but not for long-term predictions.

Qualitative differences in the shapes of the leach curve for CFL versus time have beenobserved frequently. This difference is illustrated in Figure 3.1, adapted from Crank [15]. Thecurve labeled "Fickian" represents a monotonically increasing curve expected for pure diffusion.The "Sigmoid" and "Two Stage" curves in Figure 3.1 are typical of leach curves for solidsexhibiting a time lag in the diffusion curve, due to the effects of slow penetration by water intothe waste form. Two or more parameters are necessary to describe the interacting diffusionand relaxation effects inherent in the advancing boundary and solubilization of the waste. No

-18


37/151


38/151

Recently, Nimnual reported the numerical modeling of Sr-85 from cement where the partialdifferential equation for transport by diffusion included a desorption term [21]:

_dC

= D eV2C + KP (8)at

where,K = a constant,P = the desorption rate for Sr-85.

3.2.5 Te m per atu re Effects and the Arrh enius Eq uation . Th e leaching of radionuclidesfrom waste forms generally can be increased by elevated temperatures, with some notableexception s. Fo r diffusion-controlled leaching the diffusion coefficients exhibit Ar rhe niu s typebehavior. More specifically:

D e = A e " E a / R T (9 )where,

A = pre-e xpone ntial factor,E a = activation energy,R = gas constan t (1.99 cal/mole-K),T = temperature, K

Diffusion processes are expected to have activation energies of about 5-6 kcal/mole as reportedfor cement [11,22,23]. Th e plot of Log D e versus 1/T should be linear with the slope of theline equal to -2.303E a/R. Parallel lines should be observed for each type of material andisotope.

Although many materials exhibit favorable Arrhenius behavior for leaching, somematerials such as bitumen soften and a re self-sealing, and produc e a negative effect. In thesecases, elevated temperature does not act as an accelerating factor.

-20-


39/151

3.3 Application of Mod eling Technique s

This section deals with the applicability of specific models to particular waste formmaterials and leaching conditions to describe and extrapolate leaching data to long-terms.Th ere are m ore detailed discussions in the chap ters, dealing specifically with each type of wasteform, that consider the effects of leaching condition factors and the optimization of combinedleaching factors.

3.3.1 Mo deling Relea ses from Cem ent. Do ugherty et al. rep orted experim ental resultsfrom ah extensive series of tests to investigate the effects of leach-test condition factors [1].Efforts to quantify some significant observations led to the use of the semi-infinite media andthe finite cylinder diffusion models for leaching for cement and cement-solidified wastes.

3.3.1.1 Diffusion Coefficient (D e) Calculated from the Semi-Infinite M odel. M aterialsfor which plots of C FL versus t ' ^ ar e linear can be used to ca lculate D e for a particularleaching experiment by the application of equation (2). The linear least-squares regression slopeof the da ta for CFL less tha n 0.20 will be directly pro por tiona l to th e diffusion coefficient, D e .The diffusion coefficients for cement, cement containing sodium sulfate, and cement containingincinerator ash were determined for each material under all the leaching conditions used duringthis program . An example is shown in Figure 3.2, wh ere C FL for Cs-137 is plotted against t*' .A linear least-squares regression line is shown for th e observed d ata. A slope of 8.663 x 10*2days 1 / / 2 and an intercept of 4.767 x 10*3 was found, with a correlation factor for the leastsqua res line of 0.996. From th e slope of the line, a diffusion coefficient of 4.9 x 10"8 cm 2/s wascalculated. A "data base" of the diffusion coefficients for all experiments made on cement-basedwaste forms cond ucted d uring this program was established in this m anne r. This "D e data base"for cement-based waste forms then was used to determine quantitatively the effects of theleaching condition factors.

-21-


40/151

IV

u

Figure 3.2 Cs-137 CFL versus t 1 / 2 (in days1/2) from cement containing sodium sulfatc leachedin deionized water at 20C.

3.3.1.2 Th e Finite Cylinder M odel. W hen laboratory samples are leached beyond 0.2CF L, the effects of depletion must be considered. Conseq uently, models must be applied thataccount for depletion of the leached species. Th e leaching of Sr-85 and Cs-137 from cem entcontaining 5 wt% sodium sulfate, in deionized water at 20C is shown in Figu re 3.3. Th e solidline is the result for CFL from the finite cylinder model using diffusion coefficients of 6.4 x10*10 cm 2/s for Sr-85, and 4.2 x 1 0 8 cm 2/s for Cs-137. Bulk diffusion accoun ts for the en tireleaching curve in both cases. Particular not e should be made that th e leaching conditionsremained constant (the leachant was changed at daily intervals) and there were no obviouschanges in the waste forms.

-22-


41/151


42/151

in contact with the waste form. However, the flow test is complex to perform and especiallydifficult at temperatures above 20C.

Test Methods

Static

Semidynamic

Flow

-*

r

MCC1

ANS 16.1an d

IAEA

MCC4S

Figure 3.4 A comparison of some basic leaching techniques: Static, Semidynamic, and Flowtests

The best compromise appears to be the semi-dynamic test as exemplified by the IAEA orth e ANS 16.1 test. The se tests are convenient to perform and attempt to maintain dilutesolutions of leachant in contact with the specimen. Because the incremental amounts ofleached substances generally decreases with time, these methods successively increase theintervals between changes of leachant. Even though the leached radionuclide concentrationsare very low, the buildup of other substances brought into solution by the leachant mayadversely affect the leaching process. This effect can be seen in Figure 3.5 which compares theleaching results for Sr-85 from two sets of cement samples containing 5 wt% sodium sulfalcwhen the replaceme nt intervals were different. W hen the leachant was changed daily over aperiod of 35 days, the results could be fully accounted for by a D e = 4.0 x 10'10 cm 2/s .

-24-


43/151

However, changing the sampling schedule from daily to weekly after the first week of leaching(ANS 16.1) resulted in a sharp break in the leaching curve which could only be accounted forby lowering D e from 2.5 x 10' 1 0 to 5.0 x 10" 1 1 cm 2/s, and continuing to calculate the CFLversus time curve by the use of the finite cylinder model. Th e small initial difference betw eenthe two sample sets reflects the inherent problems of comparing samples prepared on twodifferent occasions.

inCDI1_

u

. 0 6

. 0 5

. 0 4

. 0 3

. 0 2

. 0 1Id on Iy AId l ink 2wks a

2 0 3 0T i m e ( d ) 5 0

Figure 3.5 CFL of Sr-85 leached from cem ent contain ing 5 wt% sodium sulfate in deionizedwater at 20C. A comp arison of the effect of chang e in the leachant re place me ntschedule is shown. The triangles represent data from an expe rimen t with dailyleachant replacement. The squares represent data from the ANS 16.1 test.

Cs-137 leached from cement containing 15 wt% incinerato r ash show ed a similar chang e

-25-


44/151

lnUU.u

50 10B 150 200T i me ( d 250 300 350

Figure 3.6 Cs-137 CFL versus time for cement containing 15 wt% incinerator ash at 20C. Thesolid lines represent the finite cylinder model prediction with D e=6.0 x 10 '8 cm^/sfor the early results of 0 to 20 days and D e=1.0 x 10"^ for the long-term results.

The IAEA procedure recommends that the leachant be renewed once per day until itspH d rops below eight for cemen t [24]. The A NS 16.1 test does not consider this factor.Different waste forms and isotopes are expected to have varying sensitivity to the changes inthe leachant replacem ent interval. Consequen tly, all estimates of the effective diffusioncoefficients for this program were made from the data for the daily changes by using the semi-infinite model where the CFL versus t1^2 is linear.

Th e Volume of the Leac hant and Surface Area of the Sample. Since the frequency ofleachant replacement affects the diffusion of radionuclides from waste forms, we investigatedthe effect of the ratio of the volume of the leachant volume to the sample's surface area.Increasing the relative volume would decrease the absolute concentrations of leached materialsin equa l sampling intervals for a semidynamic test. Th e ANS 16.1 stand ard leach test

-26-


45/151

recom mend s a volume to surface are a ratio of 10/1 cm. Th e effect of increased ratios of 30/1and 50/1 are com pared to the results for daily sampling in Figure 3.7. Th e increased ratio ofvolumes to surface area enhanced the leaching of Sr-85, although not as much as by simplychanging the leachant on a daily basis.

inCDIue nu.

. 0 8 -

. 0 6 -

. 0 4

. 0 2

1 0 0T i m e ( d )

Figure 3.7 CFL of Sr-85 versus time from cement containing 5 wt% sodium sulfate leached indeionized water at 20C. A comparison of results was made between results forleachant volume to sample surface area ratios of 10/1, 30/1 and 50/1.

Effect of Sam ple Size. T he ability to correct for size is critical to th e deve lopm ent of aleach test that can be extrap olated to full-scale waste forms. Several studies of sam ple size andradioactive leaching have bee n re porte d for cem ent waste forms [25,26,27]. Th e semi-infinitemodel equation (2) indicates that tl:s effective diffusion coefficients, D e , should be independentof sample size or, specifically, th e ratio of the sample's volume to its surface area (V/S ). Ineach case reported, D e was relatively independent of the sample's scale, as is illustrated for Cs-137 leached from cement containing ion exchange resin [26]. Figure 3.8 shows the norm alized

-27-


46/151

cumulative fraction leached (CFL x V/S) plotted against time on a log-log scale for samplessizes ranging from a V/S ratio of 0.8 to 9.1 . Six different sized specimen s show a band with awidth of about a factor of 2. This spread in thedata is probably caused by the inherentvariability among batches of waste forms. Similar variability was observed for scaling experi-ments as reported earlier in this program [1].

in\*E CU .u7

lO t 0

10T - 1

10T-2

10t-3

A+8V^S-9.10V/S-5.09V^S-2.52V/-S-1.31V/S-0.934

I0T-2 1 D T - 1 lo t o lot iTIME (d) lo t alO t 3

Figure 3.8 Log (CFL x V/S) versus log t for Cs-137 from cement containing ion exchangeresin. Six sizes of cylindrical waste forms are compared after correction with V/S.

Temperature. In leaching experiments with cement specimens, as the temperatureincreases so do theconcentrations of dissolved specimens in the leachate for any given intervalof time. Experim ents at elevated temperatures therefore may require increased volumes ofleachant or frequent replacement. To optimize the effect of increased temperatures onleaching, the leachate must remain as dilute as possible, yetstill contain easily measurable

-28-


47/151

amounts of the radionuclides being studied. Calculations of activation energies from equation(9) must assume that the leaching conditions have been optimized at each temperature. TheArrhenius isotherms (as exemplified by Figure 3.9 for Cs-137 and Sr-85 from neat cement) areexpected to be parallel lines for each waste form material and isotope, since the activationenergies for diffusion are expected to be relatively constant (4 to 6 kcal/mole). Table 3.2givesthe calculated activation energies obtained for Cs-137 and Sr-85 from cement waste forms.Activation energies for Sr-85 are higher than expected for diffusion, indicating that an addi-tional process is retaining the Sr-85. M ore energy is required to leach a given quantity of Sr-85 , giving a higher activation energy.

1.0E-06E:

1.0E-07 =

1.0E-08 E

1.0E-09 E

De (cm*/sec)

1.0E-102 .8 2.9 3.1 3.21/T (K) x1000 3 .3 3 .4 3 .5

Cs-137 Sr-85

Figure 3.9 Arrhenius plot of Log D e versus 1/T for Cs-137 and Sr-S5 leached from neatcement in deionized water.

-29-


48/151

Table 3.2Activation Energies Calculated from Leaching Data

Waste Form

CementCement + 5 wt% N a 2 S O 4Cem ent + 15 wt% AshSimple ions in water

Sr-

14.3 i16.7

4 -

E a

: 4.5: 7.76

(kcal/mole K)

Cs5.2 1.75.4 0.94.9 1.5

4 - 63.3.2 Vinyl Ester-S tyrene + Sodium Sulfate. Th e leaching of radionuclides from vinyl

ester-styrene containing 20, 30, and 40 wt% sodium sulfate follows a diffusion mecha nism. Asdiscussed in Section 7, the data can be well described by the finite cylinder model (Figure 3.10).

. 2 5

IN .i nIoL.U

. 15 -

. 0 5

180 270T i m e ( d ) 3 6 0 4 5 0

Figure 3.10 CFL versus time for Co-57 from vinyl ester-styrene containing 20 wt% sodium sulfateleached in deionize d water at 20C. Th e solid line represen ts the finite cylindercalculation.-30-


49/151

3.3.3 Bitumen . The long-term leaching of Cs-137 from n eat bitumen in deionized waterat 20C can be well described as simple diffusion. How ever, the leaching data for bitum encontaining th e soluble salt sodium tetrab orate is quite different. Figure 3.11 shows the averageCFL of Cs-137 from bitumen containing various amounts of sodium tetraborate leached indistilled water at 20C. Th e leaching curves are highly dep ende nt on th e waste loading andshow a distinct initiation time and sigmoid shape. Th e leaching curves for Cs-137 from bitumenwith sodium tetraborate appear to belong to the case of anomalous non-Fickian leachingbehavior. Th e observed swelling of these waste forms strongly influences the leaching process,as specimens that are the most highly loaded are the first to swell and exhibit the most rapidleaching behavior.

Cumulative Fraction Leached

100

40% Salt

2 00 300 400Time (daya)30% Salt

50 0

20% Salt

6 0 0 7 0 0

0% Sal tFigure 3.11 CF L versus time for Cs-137 leached from b itume n conta ining 0, 20, 30 and 40 wt%sodium tetraborate in deionized water at 20C.

3.4 Summary

Several modeling techniques were tried to reduce the leaching data obtained in thisprogram to characteristic param eters . Th e most successful was the application of bulk diffusion

-31-


50/151

equ ation s (2) and (3) to obta in diffusion coefficients. Th e genera l applicability of thes eequations to waste form types is shown in Table 3.3. Cem ent-based wa ste forms in general,vinyl cster-styrene containing 20 or 30 wt% sodium sulfate, and neat bitumen leaching could bemodeled if the effects of changes in the replacement interval of the leachant were taken intoaccoun t. Diffusion coefficients w ere calculated for all th e experiments that could be successfullymodeled by a simple diffusion mech anism. Th ese coefficients will be used in the followingsections to evaluate the effects of the test acceleration factors.

Efforts are continuing to obtain values of the diffusion coefficients that are optimized forthe best fit of the leaching data throug h the use of the finite cylinder mode l equation (3). Th emagnitude of additional effects such as chemical reactions on diffusion, and the effects of thesurface skin are also being further investigated since they may be significant under acceleratedtest conditions.

For materials which could not be modeled, graphical methods will be employed toevalua te the leach test conditions. Th e principles which generally apply to a diffusion mecha-nism for leaching are expected to be operative but complicated by additional mechanisms, suchas adsorption. Further efforts to model materials which show anomalous leaching characteristicsare being made.

-32-


51/151

Table 3.3Applicability of Mathematical Models to Leaching Results forVarious Waste Forms Leached in Deionized Water at 20C.

Binder

CementCementCementCementCementV ESV ESVESBitumenBitumenBitumenBitumen

Waste

Sodium Sulfate15 wt% Ash25 wt% Ash35 wt% Ash20 wt% Na 2 S O 430 wt% Na 2 S O 440 wt% Na 2 S O 4

20 wt% Na 2 B 4 O 730 wt% N a 2 B 4 O 740 wt% N a 2 B 4 O 7

Co-57

N ON ONON ONODDDDAAA

Sr-85

DDDDDDDDN OAAA

Cs-137

DDDDDDDDDAAA

not observed= Diffusion only, equ ation s (2) and (3) are applicable.= Anom alous, no satisfactory mathem atical model available.

-33-,


52/151

4. PORTLAND CEMENT CONTAINING SODIUM SULFATE AS SIMULATED WASTE

4.1 Introduction

Portland cement is the solidification agent most commonly used for low-level radioactivewaste. It is used eit he r witho ut modification o r with any of a variety of additives and admix-tures intended to improve properties of the waste form. Portland cem ent solidifies by a seriesof hydration reactions primarily involving dicalcium silicate and tricalcium silicate, but alsoinvolving smaller quantities of tricalcium aluminate and tetracalcium aluminoferrite [28,29].Other compounds such as gypsum (CaSO42H2O) and portlandite (Ca(OH)2, a product of thehydration reactions) are involved in the setting and hardening of cem ent [28]. Portland ceme ntsolidifies into a poro us structure; its porosity is propo rtional to the am ount of water m ixedwith the cement powder and strongly influences the properties of the hardened cement, includ-ing its teachability.

Th e various com pone nts of cem ent react with water, with each oth er, and also withsome com ponen ts of waste stream s to form new compoun ds. This process can be beneficial, asn the case of incinerator ash, where some components of the waste join in the cement-forming

reactions. In other cases, such as sodium sulfate, the process can be detrime ntal because the

Chemical reactions between cement and waste components can influence the leachingr different radionuclides. In addition, the physical structu re of cemen t waste forms

ent. Since leaching from a porous medium is partly controlled by the

While the leaching behavior of cement waste forms is more complex than for other

-35-


53/151

4.2 Mo deling and Mechanisms of Leaching

To determine if diffusion is the dominant leaching mechanism for Cs-137 fromcement/sodium sulfate waste forms, the cumulative fraction leached was plotted against thesqua re roo t of time for thr ee types of experime nts: AN S 16.1, daily leachant replacem ent, andflowing leach ant (Fig ure 4.1). This plot will yield a straight line if diffusion is th e leachingmechanism and if no ne of the experimental conditions are limiting releases. This statem ent istrue up to about CFL=0.2 (20% release), then depletion begins to slow down the leach rate.The plot of Cs-137 releases during the ANS 16.1 test is linear until the replacement interval ofthe leachant beco mes greater than one day, when th e data falls below the projected line. Thischange occurs at about 20% release, indicating that: 1) depletion is becoming important, 2)diffusion is being inhibited, or 3) som e process, such as adsorption or precipitation, is removingthe tracer from solution.

1.00

0.80

0.60

0.40

0.20

0.00

Cumulative Fraction Lcachtd (Ca-137)

ANS 16.1Replacement Interval

- B - Daily Sampling Flow Through

10


54/151

If leaching is suppressed by longer intervals of replacement, suppression would be theresult of increases in concen tration of dissolved ions in the leachate . This can be shown bycomparing the data from a leach test with daily or flowing leachant replacement (Figure 4.1)with data from a test with less frequent replacem ent. Th e curves showing Cs-137 releases fromexperim ents using flowing leachant or daily replacem ents are very similar to thos e using theAN S 16.1 replace me nt schedule. Curves for all thre e both experim ents are linear below 25 %release, with the data from the two semidynamic-type tests tending to curve downward afterthant point.

To obtain bette r information, a mo re sophisticated approac h is necessary. The refore , amodel was used that is based on Nestor's approximation for diffusion from a finite cylinder [18].This model takes depletion into account, so comparisons can be made to greater CFL values.However, our computer program for the finite cylinder model has a limited number ofiterations. Consequ ently, it is useful for C FL values less than about 0.95 (9 5% relea se) (se eSection 3).

This model can be used to describe Cs-137 releases during the early part of the semi-dynamic experim ents. How ever, as shown in Figure 4.2 (curve #1 ), the model results aresignificantly g rea ter than th e da ta after 50 days. Since dep letion is included in the calculation,this finding implies that some ou tside factor is inhibiting leaching. By using a seco nd diffusioncoefficient and starting the calculations at time=0 and CFL=0.2, a good fit can be obtained fortimes greate r than 100 days. Th e lowering of the leach rate with time does not app ear to because d by an ongo ing chan ge in the m aterial because a single diffusion coefficient is effective inmo deling the da ta from 100 to 500 days. If the m aterial were progressively changing, it couldnot be m odeled by on e diffusion coefficient for this long period. The refore , the reducedfrequency of leachant replace me nt may cause the drop in release rate . To test this hypothesis,the finite cylinder model was applied to data from a flow experiment, which provided a greatertotal volume of leacha nt than did the semidynamic ANS 16.1 leach test. Th e m odel fit the Howdata very well from the start of the experiment (Figure 4.2, curve # 2 ) to about 75 days. After75 days, some small cracks in the material caused a change in the leach rate that can be sconon the figure. The good modeling fit leads to the conclusion that for Cs-137 (at 20C in 1300

-37-


55/151


56/151


57/151

0.060.05 -


0.00 6(Tlme)*0.5

Figure 4.4 Triplicate baseline data for Sr-85 are the plotted against the squ are root oftime. There is a change in leach rate after the first week of an ANS 16.1 test.

inC OIL.inu.u

. 0 8

. 0 5

. 0 4

. 0 3

. 0 2

. 0 !Id onlyId Iwk 2uks

2 0 3 0T i m e ( d ) 4 0 5 0

Figu re 4.5 Mo deling Sr-85 releases required two diffusion coefficients, on e for the earlyportion of the experiment and one for long-term leaching. Fo r the experimentwith daily leachan t replacem ents, the mo del fits with on e diffusion coefficient.

-40-


58/151

Releases of Sr-85 are plotted against the square root of time in Figure 4.6 from threetypes of experiments: a semi-dynamic test with daily leachant replacement, a flow-through test,and an A NS 16.1 test with lengthening intervals of leachant replace me nt. Th e daily test data islinear thro ugh out its dura tion, indicating tha t diffusion is the leaching mechanism . Results fromone specimen in the flow-through test are similar to those from the daily experiment and areshown on th e plot, but the oth er specimen is significantly different. This finding is believed t obe du e to mixing problems in the sample cell during the flow-through test. In Figu re 4.6, theplots for the ANS 16.1 test and the flow-through test curve upward after about 0.05 CFL; thischange was caused by samples cracking.

0.140.120.10

0.080.060.040.02

Cumulative Fraction Leached (Sr-85)

0.00 MS

Replacement Interval AN S 16.1- B - Daily Sampling

Flow Through

6 8 10 12 14 16(Time)*0.5

Figu re 4.6 Da ta from th ree types of leach tests: ANS 16 .1, semidynam ic with daily Icachuntreplacements and flow-through tests. These data are plotted against the square rootof time which produces a linear p lot if diffusion is the leaching m echanism .

-41-


59/151

While it is apparent that diffusion is the short-term leaching mechanism for Cs-137 andSr-85 from cement/sodium sulfate waste forms, it is not necessarily correct for long times.Portland cem ent is a reactive m aterial that changes with time. Alterations in porosity, insurface chem istry, and struc ture and in bulk com position influence leaching significantly. Fo rexample, absorption of Cs-137 and Sr-85 on carbonate minerals growing on the surface ofcem ent wastes that are exposed to air has bee n observed [19]. This type of process is notnecessarily imp ortant during leach tests at 20C wh ere efforts are m ade to minimize conce ntra-tions of dissolved species in the leach ate. How ever, with high leacha te conce ntrations, in staticleach tests, or in tests at elevated temperatures, back-reactions could become significant.Consequently, when changing test conditions, care must be taken that no secondary reactionsare introduced that cause an apparent reduction in leach rate.

4.3 Single Factors that Ac celerate Leaching

Three single factors were identified earlier that could increase leach rates in a way thatdo not alter the leaching mechanisms of cement/sodium sulfate waste forms [1]. Th ese a re:

1) increased temperature2) decreased size3) increased replacement frequency of leachant/leachant volume

While no actual mass transport modeling was included in the report, some mechanisticinterpre tations w ere made, based on activation energies and linear correlation plots. Since thenthe finite cylinder model was applied to som e of this data. This work, as well as the earlierfindings, are discussed below as background to new results presented later.

4.3.1 Te m per atur e. Te m per atur e is expected to be an accelerating factor whe n diffusionis the dominant leaching mechanism, and should follow the Arrhenius function unless otherexperimental factors limit releases.

-42-


60/151

4.3.1.1 Portland Cem ent Paste. Elevated tem perature s accelera te leaching of Cs-137from harden ed portland ce me nt paste by as much as a factor of 30 at 70C [30]. At 50C,leaching is accelerated by a factor of 11 (Figure 4.7) as measured by the time required to reacha fraction r eleas e of 18% [30]. An Arrh eniu s plot of the diffusion coefficients calcu lated fromthis data (Figure 4.8), shows that even at 70C, the leaching of Cs-137 fits the Arrheniusfunction. This is indicated by the 70C data being on a line with the data from experimentsrun at oth er tem pera tures. Consequently, leaching of Cs-137 from portland cem ent pastecontaining radioactive tracers can be accelerated at 70C and still maintains the leachingmechanism. Th e average activation energy calculated for Cs-137 is 5.4 kcal/mol, a value whichis consistent with that for diffusion of simple ions in water.

Cumulative Fraction Leached Cs-137

2 0 C 3 0 C

1 0Time (days)- & - 4 0 C SO C 7 0 C

Figu re 4.7 Cum ulative fractions leached (CFL ) of Cs-137 from po rtland type I cemen tpaste specimens leached at 20, 30, 40, 50 and 70C.

-43-


61/151

Da lcm'/c)1.0E-06F

1.0E-07 =

1.0E-08 =

3.1 3.21/T (K ) x 1000 3 .3 3 .4 3 .5

Figure 4.8 Arrhenius plot ofthe diffusion coefficients from portland cement pastecontaining Cs-137 and Sr-85, leached at20, 30, 40, 50, and 70C.

Releases ofSr-85 from specimens ofcement paste are about two orders ofmagnitude lowerthan those ofCs-137, and the effect ofincreasing tem perature isinconsistent (Figure 4.9).Th ese releases a re significantly altere d by secondary reactions, m aking interpre tation difficult.In later experiments, precautions were taken tokeep some ofthese reactions (such ascarbona-tion ) from occurring. Diffusion coefficients calculated from t he early pa rt ofthese Hata wereused togenerate the Arrhenius plot forSr-85 (Figu re 4.8). W ith the exception o the datataken at 50C, the diffusion coefficients fall wh ere exp ected. Also, theCs-137 and Sr-85 plotshave similar slopes, which suggests that diffusion isthe primary release mechanism inbothcases. No Co-60 was observed inany leachate from these specimens. At thepH ofcementwaste forms (about 12), cobalt forms aninsoluble hydroxide that cannot beleached.

-44-


62/151

0.10

0.08 -

Cumulative Fraction Ltached Si-85

0.00 1 0Tim* (days)2 0 C 3 0 C 5 0 C 7 0 C

Figure 4.9 Cumulative fraction leached for Sr-85 from portland cement paste containingradioactive tracers at temperatures ranging from 20C to 70C.

4.3.1.2 Portland Ce me nt Plus 5 wt% Sodium Sulfate. Leaching of portland c em entcontaining sodium sulfate as a simulated waste can be accelerated by temperature but theresults show a significant a mo unt of variability between specim ens (F igure 4.10). Th e scatter inthe da ta is not re lated to the m aterial itself since at 20C there is little variability. Th erefo re, itmust be related to experimental conditions, such as temperature or volume of the leachant.Increases in leach rate of cement/sulfate specimens are not as great, at any given temperature,as they are for cem ent pa ste specimens. This difference could be caused by increases inco nce ntra tion s of dissolved species which would slow diffusion. Altern ately, this redu ction couldbe caused by changes in the structu re of the m aterial, particularly at its surface. W hat isbelieved to be the mineral ettringite (calcium aluminate trisulfate-32 hydrate) formed on thesurface of specimens leached at 50C, extensively altering the pore structure of the cement [1].

-45-


63/151


20 C - B - 40 C "A- 50 C "- 60 C

2 0 4 0 60Time (day*) 100 120

Figure 4.10 Cumulative fraction leached for Cs-137 vs time from portland I cementcontaining 5 wt% sodium sulfate at 20, 40, 50, and 60C.

The Arrhenius plot of Cs-137 leached at 20, 40, 50, and 60C (Figure 4.11) shows thatat elevated temperature the Cs-137 diffusion coefficients are higher than at 20C and the dataat 20, 40 and 50C are linearly correlated. The 60C data fall below the tren d of th e ot he rtemperatures, suggesting that leaching at 60C is inhibited in some way (e.g., plugging of pores,precipitation, saturation effects of leach ant). De pletion doe s not appe ar to cause this chan ge,so it could indicate a change in mechanism at 60C. It is m ore likely that th e reduction inleach rate is caused by a change in structural (e.g., porosity) control on leach rate since thedata for 50C and 60C follow th e same curve shap e. Activation ene rgies we re calculated ateach interval using temp eratu res of 20, 40, and 50C with an average activation energy of 5.6kcal/mole for the first week of leaching (similar to activation energies calculated for leachingfrom cement paste).

-46-


64/151

De (cm'/uc)1.0E-06r

1.0E-07 :

1.0E-082.9 3.21/T (K#) x 1000Figure 4.11 Arrhenius plot of Cs-137 leached from cement containing 5 wt% sodium sulfateshowing diffusion coefficients as a function of the reciprocal temperature inKelvins. Experiments were conducted at 20, 40, 50 and 60C.

Leaching of Sr-85 from cement/sodium sulfate waste forms is also accelerated byng the tem pera ture (Figure 4.12). This figure also shows that after th e first week of the

e day), the leach rate decreases dramatically. De pletion is not th e cause, since the totalless than 2 0% . Also, the re is much less scatter in the da ta during daily leachant

ent intervals. Th e Arrhen ius plot of Sr-85 (Figure 4.13) shows that leaching of Sr-85

This data is calculated from the first week of leaching and, therefore ,daily repla cem ents. Th e diffusion coefficients for Sr-85 are ab out two

rs of magnitude lower than those for Cs-137. This difference would cause less build up ofin the leach ate and may account for its mo re orderly behavior. Of particular im portance

-47-


65/151

0 . 1 0

0.08 -

0.06 -

0.04 -

0.02

0 . 0 0


2 0 4 0 6 0Time (days) 8 0 1 0 0 120

Figure 4.12 Sr-85 cumulative fraction leached vs time from portiand I cement containing 5wt% sodium sulfate at 20, 40, 50C, and 60C leached in deionized water.

D (cm'/iie)1.0E-08r

1.0E-09 ;

1.0E-102 .9 3.1 3.2 3.31/T (K*) x 1000

3 .4 3 .5

Figure 4.13 Arrhenius plot of Sr-85 showing diffusion coefficients as a function of thereciprocal temperature in Kelvins.

-48-


66/151

4.3.2 Specim en Size. T he size of the specimen is also an acceleration factor forent/sodium sulfate waste forms. This finding is expe cted if diffusion is the leaching

echanism since the diffusion coefficient is indep ende nt of sample size. A chang e in the


2 0

V/S 0 .41

6 0Time (days)- V/S 0.84

8 0 100

V/S 1.85

1 20

Figure 4.14 Cs-137 cum ulative fraction leached vs time from por tland I cem ent con taining 5wi% sodium sulfate at waste form volume to surface area (V/S) ratios of 0.41,0.84 and 1.85. Sam ples were leached in deionize d water at 20C.

4.3.3 Volum e of the Leachant. Increased volume or replacem ent frequency of the

itself. The effect of replac

^legibility notice

Documents