ORNL-1554 4 ISCWSS*

i*4-

CENTfiALRfiSPAprprTBfi

MARTIN MARIETTA ENERGY SYSTEMS LIBRARIES

B*: j£fr_J&£4»fc*»

ISCWSS* 3 imst 03^5Eb n

fiSt

HOMOGENEOUS REACTOR P|$if:CT

QUARTERLY PROGRESS REPORT

FOR PERIOD ENDING MARCH 31, 1953

ORNL-1554

Progress

4

i" \

c&

Oak Ridge National laboratoryOPERATED BY

Carbide and Carbon Chemicals companyA DIVISION OF UNION CARBIDE AND CARBON CORPORATION

QH3POST OFFICE BOX P

OAK RIDGE. TENNESSEE

'^^SSSKSfisESSfflWiR!*

ORNL-1554

This document consists of 148 pages.

Copy O of 207 copies. Series A.

Contract No. W-7405-eng-26

HOMOGENEOUS REACTOR PROJECT

QUARTERLY PROGRESS REPORT

For Period Ending March 31, 1953

Project Director — J. A. SwartoutEngineering - C. E. WintersChemistry - S. C. Lind, C. H. SecoyCorrosion — E. G. Bohlmann

Chemical Technology - F. L. Culler, F. R. Bruce

Compiled by W. E. Thompson

DATE ISSUED

JUN •> 5 1953

OAK RIDGE NATIONAL LABORATORY

Operated by

CARBIDE AND CARBON CHEMICALS COMPANY

A Division of Union Carbide and Carbon Corporation

Post Office Box P

Oak Ridge, Tennessee

MARTIN MARIETTA ENERGY SYSTEMSLIBRARIES

3 ^Sb D34T55b D

mmmmmmmmmm

tNL-1554

Progress

INTERNAL DISTRIBUTION

tant)

1. CJ Center

2. Bi ky Library3. Heal B̂Physics Library4. Metal IHy Library

5-6. React or^HcperimentalE:nginee^^kg Library

7-8. Central Rl^Barch Librar•y

9-18. Central Fi^^^19. R. E. Aven ^^^20. S. E. Beall ^^21. w. B. Berggren^HL22. E. P. Blizard ^^^23. E. G. Bohlmann ^^^24. G. E. Boyd ^^25. W. M. Breazeale (Con^ftIt

26. R. C. Briant ^27. R. B. Briggs '28. F. R. Bruce

29. A. D. Callihan

30. D. W. Cardwell

31. G. H. Cartledge32. G. H. Clewett

33. T. E. Cole

34. D. D. Cowen a

35. F. L. Culler M36. J. S. Culver Jt37. L. B. Emlet (Y-12) M38. J. L. English ^m39. D. E. Ferguson ^m40. W. R. Gall m

' 41. J. P. Gill M42. R. J. Goldst^Bf43. C. B. Grahan^y44. J. C. Q,r\e,^g45. W. D. Hai^F46. C. S. n^rm47. p. N. JMnbenreich

48. C. GUîsig49. A. llaender

50. A. S. HouseJMEder51. W. B. Hum^^K-25)52. G. R. Jej^mns (C&CCC, South

Char l£on)53. G. H.jÊnks54. C. PJKeim55. M. HKelley56. EiiV^ * King57. rs. Kitzes

58. J^^M* E. Larson

59.J^V3 * C. Lind

6m'r. s. LivingstonR. N. Lyon

^Ê • W. L. Marshall

"3. H. F. McDuffie

64. J. R. McWherter

65. E. C. Miller

66. K. Z. Morgan

67. E. J. Murphy68. L. R. Quarles69. P. M. Reyling70. W. L. Ross

71. J. P. Sanders

72. H. C. Savage13. C. H. Secoy

C. L. Segaser. E. D. Shipley

76%i^Hf * H. Snell

77. SP iewak

78. WmT * Susano

79. j^ Swartout

80. E. kTaylor81. W. fchompson82. M. Tol

83. S. VlsiT^84. W. P. Wamr

85. A. M. Wei^^g86. T. A. We ItMl

87. E. P. WignenBConsultant)

94.

95.

96.

97-107.

108.

109-113.

114.

115.

116-118.

119.

120-121.

122-127.

128.

129.

130.

131.

132.

133-136.

137.

138.

139-141.

142-145.

146.

147-154.

155.

156-159.

160-161.

162.

163.

164.

165.

166.

167.

168-169.

170-171.

172.

173.

mm*m*mm**mm

91^WP. C. Zmola9WMG. C. Williams

J. C. Wilson

to Phillips Petroleum Company)

Jtics, Clevelandcs, Washington

ions Office

s Office, Augustaford

nse Laboratory

ion Laboratory, BerkeleyLaboratory, Livermore

VI

eports pr^H|ously issued injflnis series are as follows:

ORNL-527 ^Kate Issujp December 28, 1949ORNL-630 ^ktiod EaKng February 28, 1950ORNL-730 Fl^fcibi^Ky Report - Date Issued, July 6, 1950ORNL-826 PerT^Kiding August 31, 1950ORNL-925 Peri^fciding November 30, 1950ORNL-990 PeriijBr^B^ng February 28, 1951ORNL-1057 PedK EiS^May 15, 1951ORNL-1121 PejPod EndMûgust 15, 1951ORNL-1221 Pjviod Ending^Mvember 15, 1951ORNL-1280 MrioA Ending ffl^t 15, 1952ORNL-1318 JBferiod Ending JuI^f^K. 1952ORNL-1424 jpperiod Ending Octo^^kl, 1952ORNL-1478 ^F Period Ending January^^^l953

*mwmmmm

SUMMARY

PART I

HOMOGENEOUS REACTOR EXPERIMENT

Reactor operation was resumed inFebruary, and on February 24 a successful run was made at design conditions of power (1000 kw), temperature(230°C), and pressure (1000 psi). Withthe completion of this run, theoriginal objectives for building theHRE were successfully fulfilled.Reactor stability and response toreactivity adjustments were excellent.Decomposition gas (2H2 + 02) at the1000-kw level was measured as approximately 10 scfm, which corresponds to aG value of 1.4. The reappearance ofchloride during the power run and anapparent increase in corrosion ratemade it necessary to reprocess thefuel charge to remove the chloridecontaminant by precipitation as silverchloride.

A spare electric boiler was installedfor pressurization of the fuel, andthe control circuits were modified

after a review of the operationalsafety of the reactor. Operation wasresumed at the end of the quarter.

PART II

BOILING REACTOR STUDIES

Graphs have been developed to permitrapid estimation of the plutoniumproduction rates and the conditionsrequired for criticality in single-region homogeneous boiling reactors.The UO.SO.-D.O solution system and theU03-D20 slurry system are presented inseparate sets of curves.

An apparatus has been completed formeasuring the effect of supersaturation

with hydrogen on the nucleation ofbubbles by fission fragments in superheated water. Design has been completed for the equipment for determining the minimum oxygen requirementfor stabilization of the U02S04-H20-stainless steel system at temperaturesof 250°C and below.

Electrically heated volume-boi1ingexperiments in a 6-in.-square systemhave been conducted. In these experiments, depths of up to 3 ft and specificpowers of up to 12 kw/liter were used.The results appear to corroborate apreviously derived relationship between specific power and solutiondensity.

A 6-ft tank has been erected for

the study of vapor release from largersystems. In this tank, air is bubbledthrough the liquid to simulate boiling.In preliminary tests, density decreases up to 0.3 have been obtained.The Babcock and Wilcox Co. has been

approached for assistance in developingvapor release devices.

PART III

GENERAL HOMOGENEOUS REACTOR STUDIES

Intermediate-Scale Homogeneous

Reactor Design

The Intermediate-Scale HomogeneousReactor is conceived of as a 50-megawatt,

two-region converter. Uranyl sulfatesolution containing approximately 5 gof uranium per liter of D20 solutionis pumped through the core. Surrounding the core is a blanket of thoriumoxide slurry or thorium oxide pelletsin D20. The core is 4 ft in diameter,and it is separated from the blanket by

a 1/8- to 1/4-in.-thick stainless steelshell. The blanket is 2 ft thick, andit is contained in an 8-ft-dia pressurevessel. Both the core and the blanket

operate at a pressure of 1000 psiaand a temperature of 250°C.

The present concept of the ISHR wasdescribed in some detail in the

previous quarterly report; design dataand a simplified flow sheet werepresented. Flow sheet and componentdesign studies are in progress in aneffort to refine the design. Many ofthese studies are not sufficientlyadvanced to be reported at this time;so the design remains essentially asdescribed previously. Results of themore advanced studies are reported.

Engineering Development

A pear-shaped, slug flow type ofcore has been selected for ISHR becauseof relative ease of fabrication,favorable flow distribution, and lowfluid pressure losses. Small-scaletests are essentially complete andfull-scale tests are planned.

The installation of an 8-ft,rotational flow type of core in the48,000-gpm water-circulating facilityis essentially complete. A limitednumber of tests will be made to confirmthe predicted pressure loss in thiscore before a full-scale, slug-flow,ISHR core is installed in the circu

lating system.Small-scale tests of gas separators

that use 70°F water have shown thefollowing: (1) either vanes or avolute may be used to establish astable vortex in a pipe, and eithermay be used to recover a reasonablefraction of the rotational energy atthe outlet; (2) pressure losses in

these systems are relatively low; (3)gas separation is adequate if a smallvortex is used, and efficient separationis difficult to attain if a largevortex is used. Further developmentalwork is being done that will lead tothe operation of a full-scale experiment.

Preliminary operation of the small-scale, catalytic-recombiner testingfacility, at low and high pressures,has resulted in confirmation of previousdata and an indication of increased

reactivity at elevated pressures.Further tests will be conducted to

investigate alternate recombiners andto study the effects of pressure, gasconcentration, deuterium vs. hydrogen,iodine, and traces of uranyl sulfate.Sufficient information has been accumu

lated to permit the design of a"packed-bed" recombiner, and there isconsiderable assurance that operation"will be successful.

A gas circulator which utilizes aspecial pump impeller and a water-lubricated canned-rotor motor appearsto be feasible for use in the re-

combiner system. It can be designedand fabricated by manufacturers whoare now supplying canned-rotor pumps.

Fatigue tests and stress analysisof a pulsafeeder pump diaphragm indicate, tentatively, that the diaphragmswill operate indefinitely if no dirtis allowed to enter the diaphragmchamber. Further stress and metal

lurgical analyses are needed beforethis indication can be considered as

conclusive.

A Stellite piston and cylinder pumpwas tested in low-temperature, low-pressure fuel, and the results were

favorable. If a satisfactory seal can

be developed for the drive shaft ofthis pump, it will be a promisingalternate for the pulsafeeder pump.

Discussion of tentative ORNL heatexchanger designs with prospectivemanufacturers has emphasized a numberof problems which must be solved. Noinsurmountable problems are evident.

The installation of a 4000-gpmstainless steel loop has been started,and a completion date of August 1 hasbeen predicted.

Completion of the fabrication, inORNL shops, of two small pumps isexpected March 23, 1953. Delivery offive small pumps, which were orderedfrom Allis-Chalmers, is expected to

start in June 1953.A drawing of the proposed in-pile

loop, which utilizes a copper catalyst,is presented. Completion of development of the first loop is anticipatedin August 1953.

A U03 slurry has been circulated atroom temperature through a simulated"let-down" valve and exchanger. Somesettling difficulties caused by inadequate circulation were evident in asimulated dump tank. Further developmental work has been deferred becauseof the urgency of other work.

Corrosion

A loop which operated approximately2100 hr at 250°C with uranyl sulfatesolution containing 300 g of uraniumper liter has been sectioned. Thestraight sections of pipe were in goodcondition, and there was no pitting.The areas of localized attack wereconfined to regions of high turbulence.

A loop designed to test the effecton corrosion of gas bubbles in thesolution and an all-titanium loop havebeen completed and put into operation.

No precipitation of U03 from diluteuranyl sulfate solutions (5 to 15 gof uranium per liter) at 250 C has beenobserved when 25 eq. % excess sulfuricacid is present. The added sulfuricacid does not materially increase cor

rosion at concentration levels of 5and 15 g of uranium per liter in thecase of sulfate, but it substantiallyincreases the corrosion at a uranyl

sulfate concentration level of 40 g ofuranium per liter. Similar resultshave been obtained with the additionof HF to U02F2 solutions containing5 g of uranium per liter at 250 C;higher-concentration solutions willalso be investigated.

Previously observed hydrolyticprecipitation of appreciable percentages of CuS04 in U02S04 solutionshas been confirmed. Again, theaddition of excess sulfuric acidprevented such precipitation.

Additional results on pin and couponspecimens under avariety of conditionsare summarized. Tentative explanations

for some of the anomalous resultspreviously obtained are discussed, butit is obvious that some factors are

still uncontrolled.

Some thorium dioxide pellets wereexposed to a uranyl sulfate solution(5 g of uranium per liter) that was at250°C and flowing at about 3 fps; noserious attack occurred.

Static test results on corrosion ofsynthetic gems and spinels, as well asresults on crevice and stress corrosion

of type 347 stainless steel, arereported. No stress corrosion wasobtained, but definite signs ofcrevice corrosion were observed under

some conditions. Tests on titanium in

IX

«P?SS«S!«ltf*«(ss!9isgp'Sigi

uranyl sulfate solutions at 95°C indi

cated the absence of stress and crevicecorrosion.

Design and planning for an in-pileloop have continued. Construction ofmockups of some of the loop componentsfor out-of-pile tests has been started.

Metallurgy

Several groups of corrosion testpins have been prepared for the dynamiccorrosion loop testing program and arenow being tested. The pins include awide variety of base metals and wereprepared by various metallurgicaltechniques, including heat treating,surface conditioning, mechanicalworking, and joining.

In the experimental welding ofheavy plates of stainless and othersteels, sound crack-free welds havebeen produced by using a combinationof heliarc welding for the initialpasses and coated electrode weldingfor the filler passes. Samples of thewelds and base metal are now beingsubjected to physical and corrosiontests. This work will continue, anda larger variety of base metals willbe tested.

The investigation of the propertiesof titanium that are pertinent tohomogeneous reactor application continues. One of the more promisingaspects of the work thus far is thatit is difficult to effect any appreciable increase of hydrogen (and consequent temperature-sensitive brittle-ness) in commercial-purity titaniumunder simulated service conditions orunder even more drastic conditions,such as cathodic treatment in acidsolution at ambient temperature. Thiswork is continuing, and various modifi

cations of surface preparation, coldworking, and chemical treatment of thesamples are being investigated.Additional work is in progress to testthe possibility of decomposing supersaturated solutions of hydrogen intitanium under reactor radiation.Impact specimens have been prepared byarc melting and working selectedhigh-purity titanium sponge in aneffort to obtain amaterial for testingwhich is initially less susceptible tolow-temperature notch embrittlementthan is commereial- purity titanium.If this work is successfu1, it will notbe necessary to use the far moreexpensive crystal-bar titanium whichis definitely known to have favorablelow-temperature impact behavior.

Aqueous Solution and Radiation

Chemistry

Experiments on the effect of reactorradiation on the static corrosion oftype 347 stainless steel indicate thatradiation does increase the corrosionrate when the partial pressure ofoxygen is high and that it probablyhas some effect when the oxygen contentis low. The initial corrosion rate(for several weeks) appeals to belinear with the square root of time.The dependence of the cqrrosion rateon fission density remains uncertain.

Studies of the stability of thepertechnetate ion, TcO

4 '

of interest as a possible corrosioninhibitor, show that under reactorradiation or gamma radiation, in theabsence of hydrogen, either there isno reduction or there exists a steady-state level of reduced technetium thatis below the limits of detection. Inthe presence of hydrogen, a brown,

/h i ch is

colloidal material, presumably TcO ,is produced, with an initial yield ofapproximately 2.4 eq. of technetiumper 100 ev absorbed.

Continued studies on the mechanism

of the hydrogen-oxygen recombinationreaction in solutions indicate that

even though the uncomplexed copperion is very active cata 1ytically,there exists some other species whichis more active.

An indirect method for measuringthe solubility of hydrogen in uranylsulfate solutions by means of anapplication of hydrogen-oxygen recombination kinetics has been de

veloped. Preliminary tests substantiatedthe feasibility of the method, andapparatus specifically designed forthis purpose is being built.

Investigations of phase equilibriain the system U03-S03-H20 have continued, and some observations in theuranyl dichromate-water system and insystems involving CuS04, U02S04, H2S04,and H20 are reported. Work on thesolubility of U03 in dilute uranylsulfate solutions at 175°C has beencompleted. The mole ratio, U-to-SO ,drops very sharply below unity aturanium concentrations less than

approximately 0.01 molar. Hydrolyticprecipitation of both copper anduranium occurs in low concentration

solutions of these cations at 250°Cunless specified amounts of free acidare added. The required amounts ofacid for selected concentrations ofcopper and uranyl ion are reported.

A single experiment to determinethe distribution of fission-productactivity between the heavy uranium-rich and the light water-rich phasesin the miscibility gap of the uranyl

sulfate-water system indicates thatthe activity is largely associatedwith the uranium-rich phase. Thissuggests the possible utilization ofuranyl sulfate as a scavenger fordecontamination.

Additional vapor-pressure data foraqueous uranyl sulfate solutions havebeen obtained. This work will be com

pleted with measurements at oneadditional concentration, and aterminal report will be written.

Experiments are in progress todetermine the feasibility of additionsof measured amounts of sulfur dioxide

and oxygen to the in-pile loop toeffect a condition equivalent toadding sulfuric acid. Results to dateindicate that the uranyl ion is notreduced by sulfur dioxide and that theoxidation of the sulfur dioxide is

complete and fairly rapid under theconditions of the experiment.

Slurry Chemistry

Work has continued on the investigation of the preparation of pureanhydrous U03, on a kilogram scale, bythe thermal decomposition of ammoniumuranyl carbonate. The oxides producedby this method yielded less than 10ppm of soluble uranium on hydration toU03'H20 in water at 250°C at a slurryconcentration of 250 g of uraniumper liter.

A study of the temperature-crystalstructure relationship of UO "H-0 hasbeen carried out. The transition

temperature for the conversion of rods

to platelets was found to occur ataround 200°C. A summary of U03-H20crystallography is included.

The preparation of U02C03 slurries,by the reaction of C02 with U03, and

the properties of the s lurries have beeninvestigated. Reactions at roomtemperature have been found to producebetter slurries than those carried outat 200 to 250°C. It was found that

a partial pressure of less than 500 psiof C02 would be sufficient to chemicallystabilize a pure U02C03 slurry at 250°C.

Studies on thorium blanket systems

have been started.

An experiment on the effect ofreactor irradiation on a slurry ofenriched U03*H20 platelets showed ahigh rate of gas production (G250oc =0.4) and a high steady-state pressure.Considerable degradation of theparticles to fragments and finesoccurred. As much as 72% of the totalsolids (300 mg of uranium as oxide)plated on the' interior surfaces of theradiation bomb.

Physical Studies of Slurries

Slurries of the U03*H20-H20 systemhave been found to be stable while

circulating at temperatures up to200°C. Above 200°C, a change incrystal form causes plating of thesolids on the walls of parts of thesystem and the loss of uranium fromthe circulating material. Efforts arebeing directed toward finding bothchemical and mechanical means of pre

venting the plating.Thorium oxide slurries were found,

in general, to be too abrasive forcirculation in stainless steel.

Progress has been made in findingsofter forms of thorium oxide and

other suitable forms of thorium.

A simplified method of evaluatingpiutoniurn -producing single-corereactors has been developed and applied

xii

»»mp»l|lWHI»HWIIipHMHffll||Ull» mmmwmw*wm

to boiling reactors containing U02S04-D20 solution and U03-D20 slurries.

Data have been obtained from which

power removal from some types ofboiling reactors can be estimated.

Chemical Processing

Chemical processing studies havebeen devoted to developing a methodfor removing fission products and corrosion products from the fuel of anaqueous, homogeneous, thermal, breederreactor and to assisting in thedevelopment of a suitable thoriumblanket material.

Two general methods for processingdilute uranyl sulfate solution fuelhave been evaluated. Complete decontamination of the fuel by solventextraction was found to be an un

satisfactory method of processingbecause of high cost and high losses;however, adsorption appears to be aneconomically attractive method forremoving impurities. The cost of theadsorption method of processing may beas low as 0.3 to 0.5 mil per kwh ofelectricity produced. Both inorganicadsorbents and organic ion-exchangeresins are being considered for thistype of process.

In studies conducted by the VitroCorp. for the HRP, calcium fluoridehas been found to be a highly effectivematerial for removing rare earthfission products from a dilute uranylsulfate solution. Recent studies here

have shown that this removal of rare

earths is due to a metathesis in whichrare earths replace calcium in thesolid phase. Therefore the feasibilityof using calcium fluoride depends onthe solubility of calcium sulfate in

the fuel solution at 250°C. Thissolubility is being determined.

Thorium oxide pellets are stable inboth water and 0.02 U U02SO„ solutionat 250°C. These pellets are subjectto abrasion against one another;however, the rate of abrasion dependson the smoothness of the pellet sur

face, and the abrasion is negligiblefor pellets having no obvious surfaceimperfec tions.

A satisfactory rate of dissolutionwas obtained for unirradiated thoriumoxide pellets in 60% nitric acidcontaining 0.1M fluoride as a catalyst.

xm

CONTENTS

SUMMARY

PART I. HOMOGENEOUS REACTOR EXPERIMENT

STATUS OF THE HRE

Reactor OperationExperiment at 200 Kilowatts and Reduced PressureHigh Power OperationHRE Off-Gas Pump

PART II. BOILING REACTOR RESEARCH

BOILING REACTOR RESEARCH

Criticality Requirements and Plutonium Production ofBare, Homogeneous Reactor

Nucleation by Fission FragmentsStability of Boiling U02S04 Solution in Stainless SteelPower Removal from Boiling ReactorsTeapot Mockup

PART III. GENERAL HOMOGENEOUS REACTOR STUDIES

INTERMEDIATE-SCALE HOMOGENEOUS REACTOR DESIGN

Internal Recombination CatalystEvaporation of D20 in the Gas SeparatorCooling-Tower Reactor

ENGINEERING DEVELOPMENT

Core and Gas Separator DevelopmentISHR core

Eight-foot rotational-flow modelPipe-line gas separators

Recombiner DevelopmentGas Circulator for ISHR Recombiner SystemPulsafeeder Pump TestsStellite Piston Pump TestsTest Loop for 4000-gpm PumpISHR Heat ExchangerSmall Canned-Rotor Pump DevelopmentConversion of the HRE Mockup to Operate with Uranium Oxide Slurry

CORROSION

Pump Loop Operation and MaintenanceLoop Test Results

PAGE

vii

3

3

4

7

14

19

19

20

20

22

27

31

31

31

32

34

34

34

36

36

39

40

41

41

41

41

42

42

43

43

49

xv

Small-Scale, Dynamic Corrosion Test Program

In-Pile LoopStatic Corrosion Studies

METALLURGY

Service and Miscellaneous Work

Radiation Damage Studies .....Welding of Type 347 Stainless SteelBrittle Fracture of Titanium

Titanium for Valve Trim

Properties of TitaniumZirconium-Titanium Alloys

AQUEOUS SOLUTION AND RADIATION CHEMISTRYStatic Corrosion Tests with Radiation

The Mechanism of the Hydrogen-Oxygen RecombinationReaction in Solutions

Solubility of Hydrogen at High Temperatures and Pressuresin Solutions of Uranyl Sulfate

Description of Filtration Apparatus Used for AqueousSolubility Determinations Above 100 C

Solubility of U03 in Aqueous U02S04 at 175°C-j Phase Diagram for the U03-S03-H20 System at 25°C

Solubility of Uranium Trioxide in Uranyl FluorideSolutions at 25 and 100°C

Equilibration Times for Uranium Trioxide in Uranyl SulfateSolutions

Exploratory Phase Investigations of Systems InvolvingCuS04, U02S04, H2S04, and H20 at Uranium ConcentrationsBelow 30 g per Liter

Comment on the System Uranyl Dichromate-WaterAn Apparatus for Phase-Stability Studies and Corrosion

Tests on 2-ml Solution Volumes

Extension of Two-Liquid Phase Studies of U02S04Oxidation of Sulfur Dioxide in Uranyl Sulfate Solution . .Chemistry of Corrosion

_4 Vapor Pressures of Aqueous Uranyl Sulfate Solutions . . . .

SLURRY CHEMISTRY

Uranium Trioxide ChemistryChemistry of U02C03Thorium ChemistryUranyl Oxide Slurry Irradiations

PHYSICAL STUDIES OF SLURRIES

Thorium Slurries

xvi

65

66

69

80

80

81

82

83

83

83

86

87

87

89

96

99

101

101

101

101

102

105

106

108

108

109

109

110

110

114

118

118

120

120

ssSBaagsW!**,**"*"-1*'

Uranium Slurry LoopsCritical Experiments

CHEMICAL PROCESSING

Fuel ProcessingThorium Oxide Pellet Studies . . . .Dissolution of Thorium Oxide Pellets

122

124

125

125

129

130

XVll

Part I

HOMOGENEOUS REACTOR EXPERIMENT

HRP QUARTERLY PROGRESS REPORT

STATUS OF THE HRE

S. E. Beall, Section ChiefS. Visner, Experimental Physics Group Leader

J. J. Hairston J. W. Hill, Jr. T. H. ThomasOperations Group Leaders

D. I. Eissenberg W. C. Keller S. I. KaplanV. K. Pare' J. L. Redford P. M. Wood

Reactor EngineersT. E. Haynes G. H. JohnstoneA. L. Johnson W. B. Krick

Technicians

REACTOR OPERATION uninterrupted reactor operation could..... r _, . . continue if another failure of theAt the beginning of this report ..,,., j. , , u n - v • inside boiler occurred,period, the Homogeneous neactor experi

ment was out of service for repairs While these repairs were in progress,and decontamination made necessary a routine fuel analysis indicated theby a fuel spill from a differential- presence of sufficient chloride ionpressure cell in the oxygen-gasconcentration (80 ppm at the operatingmeasuring apparatus. A completely new uranium concentration of 40 g/kg)system for fuel oxygenation was com- to cause excessive corrosion of thepleted and tested early in January. type 347 stainless steel piping inThe reactor was restarted on January 5 the fuel system. Plans were made toand operated at 125°C, 1000 psi, and remove the chloride by an electrolytic50 kw for 16 hours. Experiments method, which involved collection ofconducted during this time provided the chloride on a silver electrodethe information that the reactor system mounted in an opening at the f ue 1-was leak tight and that the reactor sampling device. However, afterwas self-stabi lizing; that is, the several days of experimentation innegative temperature coefficient was which it was demonstrated that insuf-sufficiently effective to overcome ficient electrode surface was availablesmall suddenly imposed reactivity in the limited space of the samplechanges and to level the reactor power cavity, a method of precipitation withwithout resorting to the standard silver sulfate was adopted. To ac-control mechanisms. Failure of the complish removal of the chloride byelectric boiler which provides 1500-lb precipitation as AgCl, it was neces-steam for the fuel pressurizer made sary to install a 12-liter, steam-discontinuance of operation necessary. jacketed, reaction vessel and aThe boiler is located in the reflector sintered stainless steel fiIter betweencompartment of the shield and is near the inner and outer fuel storagethe location of the recent fuel leak. tanks so that the highly active fuelIts confined position and high radi- could be processed without removalation level (2 r/hr) made the repair from the reactor system. The fueljob a difficult one. Although sue- was treated in 6-liter batches withcessful repair was accomplished, it was sufficient silver sulfate to precipitatedecided to spend additional shutdown the chloride and was boiled for 20 mintime to install a spare electric boiler to coagulate the precipitate. Afteroutside the reactor shield so that cooling to room temperature, the


batches were filtered and transferredto the dump tanks inside the reactorshield. The source of the chloridewas a liquid-level manometer, whichcontained Merriam red oil, as a fluid,and was connected to the outer storagetanks. This manometer was used toobtain information for fuel inventories, and on one occasion approximately 50 ml of oil was accidentallydischarged into the fuel tanks. Subsequent experimentation showed that theoil (65% mono-chloronapthalene) could beexpected to decompose at 200°C in thepresence of the uranyl sulfate fuelsolution and that the 50 ml was a sufficient quantity to account for abouttwice the amount of chloride found inthe fuel.

In the course of the shutdown andin preparation for operation at higherpower levels, the reactor instrumentation and the control circuitry werereviewed'1' for operational safetyaccording to the reactor characteristicsknown at that time, which includedexperimental data on the temperaturecoefficient of reactivity and theeffectiveness of rods and reflector,as well as information concerning themechanical, chemical, and hydraulicbehavior of the system. The mainproblems arose from the effectivenessof the safety rod and reflector havingbeen shown to be much less and thetemperature coefficient considerablylarger than originally calculated.Thus, a decrease in temperature from250 to 160°C is sufficient to overridethe effectiveness of the reflector.Feasibility studies of the HRE haveindicated, however, that the reactorcan withstand, without the aid ofexternal safety systems, large increases in reactivity because of thelarge negative temperature coefficientof reactivity. A step increase inreactivity of 2% or the introductionof a linear rise in reactivity of

^ ,, i?- Visner. Nuclear Safety of the HRE, ORNLCF-53-2-76 (Feb. 10, 1953).

0,05 %/sec would lead to a power surgewith a peak of about 105 kw and apeak pressure of less than that required to burst the core. In a reactorsystem completely tested and shielded,this should lead to no difficulties.The fundamental philosophy was adopted,however, that because of the absence,at the time, of experimental confirmation of the predictions on thekinetic behavior of the HRE and becauseof the possibility that unforeseenoccurrences might lead to instabilities,the reactor power should be continuallyunder the control of the operatorso that unplanned power excursionscould be prevented.

Several revisions were made in thereactor interlock control system andin the operating procedure to minimizethe occurrence of reactivity changesgreater than the capacity of theavailable safety system. In essence,the revisions were made to protectthe reactor from the possibilities ofexcessive cooling and excessive concentrations of the fuel in the core.

On February 16, operation was resumed by circulating the fuel at 200°Cand at a concentration of 20 g ofuranium per kilogram of solution. Itwas determined by chemical analysisthat no additional chloride had appeared after a 12-hr period. Thereactor was operated at low power againon February 18 and 19, and on February20 the power level was increased to 50kw.

EXPERIMENT AT 200 KILOWATTSAND REDUCED PRESSURE

On February 21, the reactor powerwas increased to 200 kw at a temperature of 140°C and a pressure of 500psi. The main purpose of operation atthis level was to obtain informationwhile the reactor was at low power andlow pressure on its stability in thepresence of gas production at a volumerate corresponding to 1000-kw operationat high pressure. The volume rate ofgas production is proportional to the

excess pressure, that is, the difference between the total pressure andthe vapor pressure of the solution.At 500-psig total pressure, whichcorresponded to 20% of the full powervolume gas rate, oscillations in powerwere observed that had an amplitudeof 12% and a frequency of 0.08/secthat led to reactor periods as short as15 sec with the power remaining constant, as shown in Fig. 1. The totalpressure was then reduced to 330 psig,which yielded a volume gas rate ofapproximately 30% of that correspondingto full power operation. The amplitudeof the oscillations then approached15% and reactor periods as short as12 sec were observed.

The power oscillations were directlyassociated with the operation of thelet-down valve. At design conditions,this valve automatically controls thedischarge of excess liquid and gasfrom the core to the low pressuresystem to maintain a constant levelof fuel in the pressurizer above thecore. Under the conditions of thetest, however, the let-down valvecycled between closed and partly openpositions with the same frequency asthat of the observed power oscillations. When the valve closed, thegas vortex apparently started growinginside the reactor core and thusdecreased the reactivity and thereactor power. When the valve opened,the gas volume in the core decreasedand thus increased the reactivity andpower.

(2) H. T. Williams, Photoneutrons in the HRE,Omi CF-51-12-122 (Dec. 19, 1951).

PERIOD ENDING MARCH 31, 1953

An analysis of the power behaviorwas made by representing the power bya constant term and the first two

terms of a Fourier Series:

P = P0 +APj cos»t- AP2 cos 2 otwith

(1)

a). = 0.46 ,

P0 = 200 kw ,APj = 20 kw ,

AP2 = 7.5 kw .The kinetic equations describing

the sys tern are

tP = S(t)-/3-6 f (P-P0) dA P

+E pi(2)ind

a.P. = -P. + B-P , (3)

r = prompt neutron generation time,0.8 x 10"5 sec,

B. - fraction of neutrons which aredelayed, associated with meanlife, t.,<2>

S(t) = externally applied reactivitychange,

b = ratio of the temperature coefficient of reactivity to theheat capacity of the system,4.6* 10"6 fe ,,/megawatf sec,

P = reactor power, kw,P. - fission power induced by the

delayed neutrons of the igr oup.

Substituting Eq,integrating yields

1 into Eq. 3 and

B P +Hir 0

fit&lcos [ait - tan" x ccrr .]

iTT 2^-2Gi'T

/3.APcos [2oit - tan'1 2axr.] , (4)

n! 1 + A0i2T2t/r.

where the transient term containing the factor e ' l has been neglected. Then,substituting Eqs. 1 and 4 in Eq. 2 gives

DWG.48558A

IfrtJP#fWf%vwf|%^|nwftf|f#|IWIlW^ i»^f(f^m

9 =45 P.M. 9:40 P.M. 9:35 P.M.

FEBRUARY 21, 19539:30 P.M.

Fig. 1. HRE Operation at 200 kw, 400 psi, and 140°C on February 21, 1953.

9:25 P.M.

53"0

oG>•50HW50r

50OO59WC/3O)

50W13O50H


b AP. b AP2sin a>t sin 2cot

a* 2o>.

BAPj cos ait - BAP2 cos 2o* - to> bpx sin cot + 2ra> AP2 sin 2*>t+ •

-.2

P0 + APj coso»t - AP2 cos 2ait

[art - tan"1 oar ]- AP, > ,~T~cos-J 1 + Oi.2T?

SB--cos [2ait - tan"1 awj • (5)

+ 4ai?T*

Further simplification is obtained , • ,i. • _u -ii .. „ i-^^mc in and adequacy of the concrete shieldby neglecting the oscillatory terms in a,lu "* ' «i.^„j

, , • .. j k •„;„«, t-fc„ around the reactor. The events relatedthe denominator and combining the «m»u u „„„,„„ aT.„r f .u <,om» to the attainment of full power aretrigonometric functions of the same . , , * „k„ i Jl „„•„;„«,

r Tk fr,ol r»c„lt aftpr summarized here from the log entriesfrequency. lhe final result, alter .4 v • , i recorded during the experiment. Asubstituting numerical values, is, in recoroea auring i>

. B „. plot of the power level, core inletunits of per cent fcf//, ^ ^^ temper atures ( and steam8-(t) - 0.052 cos (a>t - 0.03) temperature is shown in Fig. 2.

- 0.019 cos (2a>t + 0.11) . (6) Approach to 1000-kw PowerThe amplitude of the reactivity February 23, 1953

change resulting from the variation 7^ pM Reactor critical at 50 kw ,in gas volume in the core is approxi- ' 200°C, and 1000 psig.mately 0.05% k f , or a total change g Q5 power' increased by withdrawingin reactivity of t).l% every 6 seconds. gteam tQ the condenSer.

A heat balance was made on the g 4g Reactor power leveled off insystem, and it was confirmed that the fche vicinity 0f 200 kw.reactor was operating at a power of ^ following observations were208 kw. The primary means of heat

i . , r made.removal was the discharge oi steamr ..l • Uôfr ..a.napr thrniish Monitrons: All readings on low scale,from the main heat excnanger tnrougn

a throttle valve to the turbine con- Stack activity: Four times breathingtolerance. It is expected that there is

denser. a further dilution of perhaps a thousandfold as the gas discharges from the

HIGH POWER OPERATIONr L unr Heat balance: 220-kw reactor power.

High power operation of the HrUi wasattained at 1:08 AM, February 24, with Gas recombiners: 2.4 cfm STP of H2 andthe steam output driving a turbo- 2"

<- - „.,;«- Wii-hin thf next 30 Shield: Fast-neutron radiation abovegenerator unit. Within the next 3U a^lerance outside the instrument thim-min, the reactor power was raised to bles. Additional concrete blocks stacked1000 kw at an electrical output of about these locations reduced the radi-iv *• , rr. , rj,, . ation to a safe value.approximately 160 kw. The importantphysical criteria of successful opera- Stability A^tuê ^powerôs^l^tion were met with respect to reactor as 5Q sec were observed.stability in the presence of gasproduction and a power density of 10:30 PM Power increased by increasing20 kw/liter, air activity in the stack, the heat withdrawal from thehandling of the decomposition gases, heat exchanger.


9 10 11

FEB.23, 1953, PM

DWG 18707A

1 2 3

FEB. 24,1953, AM

Fig. 2. Initial Operation of the HRE at Full Power.

10:50 Power leveled off at approximately 500 kw.

The following observations weremade.

Monitrons: All on high scale. The ventilator monitron located at roof of build

ing was reading off-scale. The highreadings of monitrons were attributed tothe high background in the buildingbecause of radiation from the charcoalbeds which adsorb the fission gases.The beds are located about 100 ft eastof the reactor building. The radiationlevel at the fence east of reactor building, 30 ft from absorbers, was 5 r/hr.

Heat balance: Reactor power level (500kw) confirmed.

Gas recombiners: 5.2 scfm of gas.

Shield: Radiation through shield appearedbelow tolerance.

Stack activity: Ten times breathingtolerance.

Stability: Amplitude of power oscillations 3%. Reactor periods as shortas 40 sec were observed.

11:40 PM An experiment was performedto test the stability of thereactor to changes in heatwithdrawal rate. The steamremoval valve was closed firstand the power was observedto decay, as shown in Fig. 3,from about 500 to 200 kw.The steam valve was thenopened, within 2 sec, to itsprevious position. The difference between the heatwithdrawal rate of 500 kwand the power production of200 kw is equivalent tointroducing reactivity intothe reactor at the rate of0.17% fe ,,/sec. As expected,the reactor power rose to

its new equilibrium valuewith a relaxation time ofapproximately 90 sec.

February 2412:08 AM Power increased by increasing

heat withdrawal. The reactor

was controlled during theprevious 3 1/2 hr solely bypower withdrawal, without theother reactor controls, suchas rods, reflector level, andconcentration, being used.

12:18 Reactor power leveled offat approximately 800 kw.

The following observations weremade.

Heat balance: Reactor power level established at 810 kw.

Stack activity: About 3 times breathingtolerance.

Radiation levels: 8.7 r/hr at fencenear charcoal bed; 10 mr/hr in the control room.


Stability: Oscillation amplitude 8%.Reactor periods as short as 15 sec wereobserved.

12:30 Decision made that sufficient steam was beinggenerated to direct thereactor output steam to theturbine.

12:40 Turbine being warmed up withplant steam. Power removalfrom reactor was decreasedin order to have the fullcondenser capacity availablefor warming up the turbine.

1:02 Turbine synchronized withTVA network. The generatorwas then motoring on the TVAline with plant steam turnedoff.

1:05 Reactor power increased bywithdrawing steam to thecondenser (Fig. 4).

1:08 Power level of 500 kw reached.The steam was directed toturbine and the governorcontrol opened. An electrical output of 70 kw wasimmediately attained.

1:12 Reactor power at 600 kw.Reactor power was varied bychanging governor setting.

1:16 Electrical load of reactorand building switched tothe generator output so thatthe building electrical requirement o£ about 70 kw wassupplied by the reactor.

1:25 Turbine, carrying the building load, resynchronizedwith TVA.

1:45 Reactor power at 1000 kw;electrical output 160 kw.Excess electrical power beingdelivered to ORNL network.Figure 5 shows the reactorpower and temperature behavior at that time.

2:05 Power decreased by turningdown governor control.

2:30 Fuel solution diluted withcondensate and the reactor

shutdown.

250

UJ

or

<a:UJ

a 200

DWG. 18554 A

/- CORE OUTLET TEMPERATURE

//

"^COF?E INLET TEMPERATURE

11:55 P.M. 11:50 P.M. 11=45 P.M. 11:40 P.M.

FEBRUARY 23,195311:35 P.M. 11 =30 P.M.

Fig. 3. HRE Operation at 1000 psi and 200°C on February 23, 1953.

s50Tf

OG>SOHin58r«:

1393OO33WC«

sow13o58H

o° 250

or3

<

UJQ.

200

DWG.48557A

REACTOR STEAM TO TURBINE 1:08 A.M.CLOSED VALVE 323

CLOSED STEAM REMOVAL

VALVE 323-750kw

CORE OUTLET TEMPERATUF?EX

^—^

^CORE INLET TEMPERATURE

1H5 A.M. MO A.M. 1-05 A.M.

FEBRUARY 24, 1953

1--00 A.M.

Fig. 4. HRE Operation at 1000 psi and 200°C on February 24, 1953.

M58M

©o

w2OM

zo

s•90nSB

CO

oo

3

cc

Ill

A ™ Wh

V*— SHUTDOWN BY DILUTION ANDSTOPPING OF HEAT EXTRACTION

~

250

/- CORE OUTLET TEMPERATURE

200

\—CORE INLET TEMPERATURE

2 =05 A.M. 2-00 A.M. 1=55 A.M.

FEBRUARY 24, 1953

1=50 A.M. 1-45 A.M.

Fig. 5. HRE Operation at 1000 kw, 1000 psi, and 200°C on February 24, 1953.

SB5013

©G>50Hm

50©©50W

t»

58m13©58H

Reactor Off-Gas and Recombiner

System. The rate of dissociation ofthe water of the fuel solution while

the reactor was at high power wasdetermined by measuring the rate ofheat removal from the gas recombinersat steady state, since this is ameasure of the recombination rate of

the hydrogen and oxygen. The resultsobtained during the initial high powerrun on February 23 and 24 are shownin Fig. 6, where a straight line wasdrawn through the data. From thisinformation, the value of 1.4 wasobtained for the G value, that is, thenumber of water molecules dissociated

per 100 ev.The charcoal beds for adsorbing the

fission—product gases appeared tofunction satisfactorily, as evidencedby the fact that there was no indication of excessive activity in thestack to which the beds discharge.

Analyses of fuel samples removedduring the 1000-kw run indicated that


the chloride ion had reappeared to theextent of approximately 30 ppm andthat the corrosion rate of the system,based on the nickel concentration in

the fuel, had increased from 0.3 to1.0 mpy to 3 to 8 mpy. The highercorrosion rate might be attributed tothe chloride content of the fuel, toradiation effects, or to the erosiveaction of gas bubbles circulating atthe high power level. It was considered necessary to eliminate thechloride, which was assumed to havecome from outer-dump-tank manometeroil. The fuel was retreated with

silver sulfate, as described previously, and the chloride content wasagain reduced to 1 ppm.

While the fuel processing was inprogress, activity was discovered inthe condensate on the steam side of

the main fuel heat exchanger, and itwas thought that a fuel leak haddeveloped in one of the heat exchangertubes. However, subsequent testing

DWG. 18552

d<°

isF" 8

<m 7o

+

x

- 5

o

Q

o

Q_

IT)<ID

^ ©T -v 230°C

/^W^ =220°C

_^*^ 1 — 244 C

T =

T=212°C(210°C§^

ls^

T = 206°(^^=210)°C

.—= 204°C

23

0 100 200 300 400 500 600 700 800 900 1000

REACTOR POWER (kw)

Fig. 6. Gas Production in the HRE-on the First Run at Full Power, February

24, 1953.

13


of the exchanger proved this assumptionto be erroneous. Further investi

gation indicated that a quantity ofhighly active liquid from the wastestorage tank had leaked back throughthe blow-down valve of the heat ex

changer and had contaminated the steamcondensate system. The activity didnot reappear when reactor operationwas resumed.

The purpose of the period ofoperation beginning March 16 was todetermine the cause of the sudden

increase in corrosion indicated duringthe 200- to 1000-kw run. The plan wasto run the reactor at zero power forsufficient time to determine the rate

of corrosion and chloride content in

the absence of radiation and then

to add the radiation variable byincreasing the reactor power first to10 kw and then to 1000 kw. Over a

period of 5 days of continuous operation, the corrosion rate, as indicatedby the concentration of nickel ionin the fuel, was found to be approximately 8 mpy, irrespective of thereactor power level. The cause forthis apparent increase in corrosionrate as compared with that experiencedprior to the appearance of the chlorideion is being studied.

In the course of this operatingperiod, the reactor was started froma subcritical condition at 200 C byconcentrating the fuel to a valuesufficient to make the reactor supercritical on a 10- to 20-sec period.At this point, concentration wasstopped and the reactor level permitted to rise until the slight excessreactivity was overcome by the temperature coefficient because sufficient

heat had been generated to raise theoperating temperature. The reactorleveled itself at approximately 10 kwand continued to operate steadilywithout a mechanical control havingbeen touched. The power level was thenincreased in the usual manner, thatis, by withdrawing steam, to thenormal level of 1000 kw. During this

14

•UMiBSHIffWiBWfflWW mmm*mmmmim

run, it was proved that the source ofthe oscillations described earlier in

this report was, as suspected, thegas-liquid let-down valve, and by increasing the output of the pulsafeederfeed pump, the valve was made tothrottle instead of operating with anintermittent open-close action. Thischange reduced the amplitude of theoscillations to 2% and made operationat higher power levels possible. Fora period of 1 hr the reactor level wasmaintained at 1200 to 1500 kw and195-kw electrical output withoutdifficulty. This period of operationprovided good evidence that a homogeneous system of the HRE type can beoperated at power densities of atleast 30 kw/liter without difficultyfrom nuclear instabilities.

HRE OFF-GAS PUMP

R. H. Chapman

A leak in the fuel gas condenser inthe low-pressure fuel system in the HREpermits radioactive gases to escapeif the low-pressure system is operatedat greater than atmospheric pressure.At the present time, subatmosphericpressure is maintained by a mechanicalvacuum pump in the line between thecharcoal beds and the stack. Use ofthis pump is not satisfactory becausethe resistance of the charcoal beds

to the flow of gas is high. Thepressure in the beds must be reducedfar below atmospheric pressure toachieve the desired pressure in thefuel system. Since the capacity ofthe charcoal to adsorb fission-productgases is greatly reduced at lowpressure, it is doubtful whether theHRE could operate under existingconditions for long periods at highpower without discharging prohibitiveamounts of radioactive gas.

A gas pump has been designed to beinstalled between the low-pressurefuel system and the charcoal beds toovercome the present difficulty.Because the specific activity of the

gas is high, the pump is required tobe absolutely leak proof. The designflow rate is approximately 0.5 cfm, andthe specified suction and dischargepressures are 10 and 45 psia, re-pectively. No commercial pumps werefound which could meet all the specifications, and therefore a reciprocatingpump with a mercury piston and a gasseal was chosen as the basis for

design.The pump assembly consists of two

6-in,-dia cylinders and a 3-gpm rotarygear pump, connected as shown in Fig,7. The lower chamber of each cylindercontains oil and mercury; the upperchamber is the gas-pumping cylinder.As oil is forced into the lower chamber

by the gear pump, mercury is displacedinto the upper chamber to compress thegas to a pressure high enough for itto flow into the discharge line.Mercury-sealed, ball check valves areused in both the suction and dischargelines. A high, theoretical, volumetricefficiency is achieved by allowing themercury to flow out through the discharge valve before reversing theflow. This also provides a supply ofmercury to both valves for sealing.Flow is reversed by reversing thedirection of rotation of the gearpump on a signal from probes in thedischarge valve chamber. A safetyprobe is provided in each suctionchamber to reverse the pump if thereis leakage past the check. A probe


is also provided in each cylinder toreverse the pump if the mercury levelgets too low.

The valves, as shown by Fig. 7,employ 3/4-in.-dia balls which floatpartially submerged in mercury. Inthe suction valve, the mercury actsas a liquid spring to keep the ballseated. The gas entering must bubblethrough a very small head of mercurybefore it enters the pump chamber.On the discharge stroke, the mercury,in addition to keeping the suctioncheck seated, acts as a gas seal. Thedischarge check will trap a smallamount of mercury for a gas seal onthe downstream side of the ball at

the start of each suction stroke.

The pump will be installed on aconcrete pad adjoining the stack atthe southeast corner of Building 7500.Concrete blocks and lead bricks are

to be stacked around the installation,as needed, for shielding. Traps areprovided on both the upstream anddownstream sides of the pump for useif all safety measures fail. Eachtrap has sufficient volume to holdthe entire amount of mercury and oilfrom both cylinders. The traps, whichwill be in the existing off-gas line,will be buried under several feet of

earth.

The pump is presently being assembled. Upon completion, extensivetesting is planned to demonstrate thereliability of the pump.

15

<0: :_''.'.v-' •':'• -. "*.' •\A:'-f;

<7. •• '

.«•'•• '

'.{7 .V•'•}••• C

GAS OUTLET,

16

)»WHWB«HwnWBlEPI»

T=?^ :SS

SUCTION-CHECKASSEMBLY WITH MERCURYSEAL7

i <1 . ? . 1

UNCLASSIFIEDDWG 19809

GAS INLET MERCURY-FILL PLUG' DRAIN PLUG'

PUMP DETAIL

Fig. 7. HRE Off-Gas Pump.

Part II



R. N. Lyon, Section Chief

CRITICALITY REQUIREMENTS AND PLUTONIUMPRODUCTION OF BARE, HOMOGENEOUS

REACTORS

P. R. Kasten

Criticality conditions and plutoniumproduction rates in single-regionhomogeneous reactors have been calculated as a function of fuel enrichment,moderator ratio, and vapor fraction at250°C for two fuel systems: (1)U02S04-D20 solution and (2) U03-D20slurry. The geometry, temperature,

and vapor fraction parameters can becombined into one term, the ratio ofgeometric buckling to the square ofthe fluid density, {B/p)2, which isonly a function of moderator ratio andenrichment. The results are shown in

Fig. 8, in which (B/p)2 is given as afunction of moderator ratio, /3, forvarious fuel enrichments.

With the aid of Fig. 8, the plutoniumproduction rate can be calculated as afunction of B from

2Xio-3

-3

77— U02S04 - D20

DWG. 19531

uu3-u2u

, y = FUEL ENRICHMEN T

5,. . -•—

/t

\

<

- c

>

o /•y /D / /

-->^ s

d.

-4

s / /

°-/-r

\

\\

^

^\

\

" ±?T t̂~~~ •' /' ~*^v. \ \M-

1\l/ \ V\

I „ 1 sy

1 ' "/ / \S~Si.

1 / 9/ /07-yc

<00

\\ \ \

b 1 '1 '

3

V \\1 I \\

> \\[ I 1 0/

~^

JV-5

! II

~J 0/ V

\\\

•5-

UJO

10

10

10

10 10^ 2 5 10°

MODERATOR RATIO, /3 (D„0 molecules /U atom)

10

Fig. 8. Ratio of Geometric Buckling to the square of the Fluid Density vs.Moderator Ratio for Homogeneous Reactors.

19


PPR

where

PPR -

2.5 (1 - P)

' *<:?)* -0^0

1 - a5.138 x 10 -3

(1)

plutonium production rate,atoms/fission,

tq = effective Fermi age at fluiddensity pQ,

a. = fuel enrichment,p - resonance escape probability.

Figure 9 gives B and (B/p)2, forseveral enrichments, as a function ofplutonium production rate. In thiscase, B is essentially the same forboth fuel systems.

The neutron losses in sulfur, asexpected, have a marked effect on theplutonium production rate. Forexample, a 15-ft-dia reactor with 0%vapor fraction and 1% enriched fuelgives a plutonium production rate of1.18 for the U03 fuel, as comparedwith 1.00 for the U02-S04 fuel.

The calculations are accurate towithin a 3% error caused by temperaturevariations in the range 250 ± 25°C.The values of p(1) and F used in thecalculations are given in Figs. 10 and11,(2) which present the resonanceescape probability and fraction offluid occupied by moderator as functionsof B. The calculations and resultsare being reported by Kasten.(3)

NUCLEATION BY FISSION FRAGMENTS

R. J. Goldstein

The apparatus for determining theeffect of supersaturation with hydrogenon the nucleation of bubbles by fissionfragments in slightly superheated

'Values obtained from P. N. Haubenreich andR. B. Briggs, private communication.

(2)M. Tobias and P. N. Haubenreich, Critical

Calculations for Half-Filled ISHR Sphere, ORNLCF-52-5-230 (May 16. 1952).

(3)P. R. Kasten, Critical i ty Requirements and

Plutonium Production of Bare Homogeneous Reactors.ORNL CF-53-3-108 (in press).

20

iTiWraPflfflllUMMH

aqueous solutions of enriched uranylsul fate has been completed, and measurements will be undertaken in the nearfuture.

Preliminary experiments with theequipment and with small glass ampouleshave been performed on relatively gas-free (preboiled) samples. These experiments indicated that fissionfragments will form vapor nuclei atapproximately 130.7 and 135.3°C fortotal pressures of 746 and 1000 mm Hg,respectively, in an aqueous uranylsulfate solution (~27 g of U235 per kgof solution). In addition, finelydrawn ampoules containing water and anaqueous solution of a boron compoundhave been superheated more than 100°Cat atmospheric pressure in the absenceof radiation. These samples will beused to study the effect of variousradiations on nucleation in an effortto understand the mechanisms involved.

Some qualitative observations onthe superheating of water and aqueoussolutions in the absence of radiationhave also been made in the course ofthis study:

1. In general, the smaller theinner diameter of the ampoule, thegreater is the attainable superheat.

2. As also found by other experimenters,'4) exposing the ampoule tothe atmosphere for some time after itis drawn and before it is filled withliquid seems to lower the maximumsuperheat.

3. Ampoules in which the liquidis nucleated and then allowed to condense, without allowing air to contactthe inner heated surface, will ebullateat much lower temperatures on subsequent reheating.

STABILITY OF BOILING U02S04 SOLUTIONIN STAINLESS STEEL

C. G. Lawson

The motivation and approach for thestudy of the stability of boiling

(4)'F. B. Kenrick, C. S. Gilbert, and K. L.

Wismer, J. Phys. Chem. 28, 1297 (1924).

3

o"

or

or

IorUJoO2

9X10

00

<or

inzUJo

\

z

o

00

0.6


0.8 0.9 1.0 1.1

PLUTONIUM PRODUCTION RATE (atoms/fission)

DW6. 19532

Fig. 9. Moderator Ratio and Ratio of Geometric Buckling to the Square of theFluid Density as Functions of the Plutonium Production Rate.

21


E

O

oro

£

104 DWG. 19533

5

2

103

5 / /

2

I02uo2s °4-D20 ~*~y/

5

?

10

^**s

U03-D20^

J***^

>^-UO3(N03)2- o2o —U02S04-H20

5

2 -

1

U 4 0.5 0(5 o-?I

0.8

/>, RESONANCE ESCAPE PROBABILITY

Fig. 10. Resonance Escape Probability for Uranium-Aqueous Systems.

uranyl sulfate solution in stainlesssteel were described, in the lastquarterly report.(5) Equipment designwas completed and construction wasstarted during this quarter.

POWER REMOVAL FROM BOILING REACTORS

W. S. Brown M. RichardsonH. A. MacColl P. C. Zmola

Power removal from boiling, aqueous,homogeneous systems depends in large

measure upon the size of the unit. Forsmall units (on the order of 2 ft inheight), the predominant mechanism forenergy removal is natural vapor risethrough the bulk of the liquid. Inlarge units (10 or 20 ft in height),natural convection, or circulation ofthe liquid-vapor mixture resulting fromdensity differences in the core, issuperposed on the natural vapor risemechanism; for some arrangements,natural convection appears to be thepredominant mechanism for energyremoval. In addition, the existenceof a free surface, particularly for

(5)

Jan. 1

22

S. N. Lyon et al., HRP Quar. Prog1953, ORNL-1478, p. 11.

Rep .


1.00

0.90

0.80

•

< c• "^

•

• H20-U02S04 SOLUTION, VOLUME PER GRAM-HOF H20 CONSTANT, DENSITY TAKEN AT 30°C.DATA OF LANSING AND NODERER.

D20-U02S04 SOLUTION, VOLUME PER GRAM-M

dOLE

A OLE OF

O

U02S04 CONSTANT, DENSITY TAKEN AT 30° TO 100°CD20-U02S04SOLUTION, VOLUME PER GRAM-MOLE OF

o

D20 CONSTANT, DENSITY TAKEN AT 50"C

A H20-U02S04 SOLUTION, VOLUME PER GRAM-MOLE OF

•

32 34 36 3810 12 14 16 18 20 22 24 26 28 30

103//3

Fig. 11. Calculation of F vs. 103//3 by Various Methods.

large units, may limit the power as aresult of unsatisfactory vapor disengagement (splashing) at the topsurface. These problems have beendiscussed to some extent in previousquarterly reports.

At present, effort is being directedtoward obtaining information on1. the specific power attainable in

small electrical-resistance-

heated volume boiling systems as afunction of the mean density decrease and density distribution,

2. the influence of natural circu

lation and gas disengagement atthe free surface in a mockup of alarge-scale system in which air isbubbled into a6-ft-dia tank filled

with water,3. the need for controlled vapor-

separation devices in large-sizeunits.

Volume Boiling in Small Systems.

The apparatus has a test section 4 fthigh and 6 in. square in which diluteacid and salt solutions are boiled by

eleetrical-res istance heating of theliquid. The electrodes form two sidesof the test section; the other sides

.are of Lucite. Vapor leaving the topis condensed and returned to the test

sec tion.

The mean fluid density is determinedfrom measurements of the boilingliquid level. A vertical densitydistribution was obtained by measuringthe attenuation of a collimated gammabeam through the fluid. A lauritsenelectroscope was used as the detector.The source was iridium which had been

irradiated in the ORNL graphitereactor. The mounting of both thesource and the detector permitted atraverse of the entire height of thetest section. Glass-enclosed thermo

couples were used to measure thevertical temperature distribution inthe fluid.

The apparatus has been operated overa range of boiling liquid depths offrom 2 to 36 in. and at total powerlevels of from 1 to 50 kw, which

23


resulted in specific power levels offrom 0.3 to 12 kw/liter. The highestspecific power corresponds to thelowest liquid levels. For liquidlevels greater than 15 in. , the maximumspecific power attained was 2.75kw/liter. A recent change in the powersupply will make it possible to operateat a total power of 100 kw.

Photographs of volume boiling inthe test section at 0.65 and 2.05

kw/liter are shown in Figs. 12 and 13.The sharp horizontal line indicatesthe nonboiling liquid level. Representative curves of the vapor fractionas a function of the vertical heightin the boiling liquid, y/Y, aregiven in Fig. 14 for specific powersof 1.2 and 2.75 kw/liter. The verticaltemperature profile is shown in Fig. 15.

An expression for the attainablespecific power as a function of themean density decrease and liquid heightin a natural vapor rise system, obtained for a simple analytical model,has been reported previously '. *• 6 ^

P \ /l\2/3

0.04 1

where

P

"avj

PiY

fe

h6

A

h

In Fig.

hf^ h(A\i/3

P ft /

(1)

specific power, kw/liter,average density of reactorcore, lb/ft3,density of liquid, lb/ft3,height of boiling liquid, ft,latent heat, Btu/lb,specific volume change inevaporation, ft3/lb,heat transfer coefficient to

bubble, Btu/hr-ft2-°F,liquid superheat, °F,drag coefficient of bubble, di-mensionless.

16, experimental results arecompared with values predicted by Eq. 1

(6)R. V. Bailey, E. N. Lawson, H. A. MacColl,

and P. C. Zmola, HRP Quar. Prog. Rep. Oct. 1,1952. ORNL-1424, p. 15. Equation 6 as given inOHNL-1424 should have a numerical coefficient of0.04 instead of 0.01, and Figs. 6 and 7 should bechanged accordingly.

24

Fig. 12. Atmospheric-Pressure Volume

Boiling at 0.65 Kilowatts per Liter.

Total power 86 kw

Initial liquid height 19 in.Boiling height 22.5 in.

Average vapor fraction 0.15Average vapor velocity aboveliquid level 0.9 fps

Fig. 13. Atmospheric-Pressure VolumeBoiling at 2105 Kilowatts per Liter.

Total power 34 kwInitial liquid height 19 in.

Boiling height 28 in.Average vapor fraction 0.32

Average vapor velocity aboveliquid level 3.5 fps


06

04

02

UNCLASSIFIED0WGI8692A

1 1 1 1 --° '

//

rrzr

1.2 kw/l if 2.75 kw/l

ELECTROLYTE. NoCI

1 1

NITIAL LIQUID LEVEL, 17(/2

1 1

in.

i

0.2 0.4 0.6 0.8

VAPOR FRACTION (BY VOLUME)

I 0

Fig. 14. Vertical Density Distribution in Atmospheric-Pressure Volume

Boil ing.

UNCLASSIFIEDDWG. IB7I0A

IB —*••—•— 1 ! 1

t.6 i?r

POWER DENSITY-2.1 kw/l

1.4Xf

i i

1.2

1 t

1.0

0.8

BOILING LEVEL

<i 'i D

v-O

0.6ska

\Ao\ •

0.4 \

A no

0.2

1 1

3

1

0 1.0 20 3.0

TEMPERATURE DIFFERENCE, Tug-TVAI, (°C)

Fig. 15. Temperature Distributionin Atmospheric-Pressure Volume Boiling.

25


-(kw/l)

UNCLASSIFIEDDWG. 16884B

50

20

10

5

2

1

_V 1 1 1

A A \

Mil

150,000

I 1 1 IIIIL

,-y—= 7500%.\ ° >

\. °°x^ o

-

oT\flo

o

« \

0 I 8^ôso0 X:

o EXPERIMEN

THEORETIC

P = 0.04

, , 1

TAL DAT

AL EOUA

H1 1 1 1

A \^TION: s.

0.5 1.0 5.0

HEIGHT OF BOILING LIQUID, Y (ft)

Fig. 16. Comparison of Theoretical

and Experimental Values for Atmospheric-Pressure Volume Boiling.

At the time the equation was evolved,it was believed that values of 150,000and 7,500 Btu/hr-ft2 representedreasonable upper and lower limitsof the parameter hfeA//fl, Thesewere based on published values ofhb,(7) A,(7> and /n.<8> In comparingthe actual boiling conditions with theanalytical model, there is an inclination to give little credence to thetheoretical basis of the equation. Inthe actual system, the identity of thebubbles is lost even at moderatespecific power. However, it appearsthat Eq. 1 can be used with someconfidence in evaluating power removalfor systems in which natural vaporrise predominates, provided an experimental value of nfcA//n obtained from asimilar system is employed. In thiscase, the parameter rather than theindividual factors would have significance.

It is planned to extend the presentwork to a higher pressure and to determine the splashing limits, that is,

(7)M. Jakob, Heat Transfer, Vol. 1, Chap. 29,

p. 620-635, Wiley, New York, 1949.(8)'L. Hopf, Handbuch der Physik, Vol. 7, p. 168,

Springer, Berlin, 1927.

26

the conditions under which appreciablequantities of liquid are ejected fromthe violently agitated surface.

Mockup of Large-scale Systems. A6-ft-dia tank 12 ft high and open atthe top has been set up so that it canbe filled with water to various depthsand so that air can be bubbled throughcanvas sheeting on the tank bottom tosimulate boiling. The objectives ofthe experiment were outlined in thelast quarterly report.'9' Some preliminary runs have been made. Valuesof mean density decrease of up to 0.3were attained. Some circulation wasevident, but it was not clear whether

the circulation was characteristic ofa unit of this size or simply peculiarto this system.

Controlled Vapor Separation in

Large-scale Units. It appears thatcirculation in large-size units willpermit a noticeable specific powerincrease for a given mean densitydecrease, as compared with the specificpower attainable by natural vapor riseonly if a large fraction of theentrained vapor can be removed atevery pass of the fluid mixture nearthe top surface of the reactor. Atpresent, little information is availableon the behavior of steam separationdevices in the range of 1iquid-to-vaporratios that are necessary for reactorapplication. The Babcock and WilcoxCo. was visited for a discussion ofthe possibility of obtaining theirassistance in determining whethercentrifugal separators are usablein a homogeneous boiling-reactorsystem.(10) A preliminary study indicated that there is an economic

advantage to a boiling reactor thatoperates with circulation and separatorsas compared with a boiling reactordependent only on natural vapor rise.

(9)P. C. Zmola, W. S. Brown, H. A. MacColl, and

M. Richardson, HRP Quar. Prog. Rep. Jan. 1, 1953.ORNL-1478, p. 16-17.

( °'R. N. Lyon and P. C. Zmola, CentrifugalVapor Separators at Babcock and Wilcox-A TripReport, ORNL CF-53-2-67 (Feb. 7, 1953).

TEAPOT MOCKUP

J. R. McWherter R. L. Cauble

The 150-psi core tank for the Teapotmockup has been constructed, and thedesigns for a recombiner, an aftercondenser, and a supporting frameworkhave been completed. Construction ofthese components has been delayed, however, pending a decision on the mostpractical condenser-recombiner arrangement.

The mockup of the core-condenser-recombiner system, as currently designed and partly constructed, is shownin Fig. 17. Many of the tests on thissystem will be meaningless, however,if liquid blanketing or iodine poisoning reduces or obliterates the ef


fectiveness of the catalytic recombinerin operation as a reactor.

Several other arrangements are beingconsidered, one of which is a succession of partial condensers withrecombiners in the limited space between each condenser. One such arrangement, which will be tested during thecoming quarter, is a unit designed forflame-recombination tests, which iffavorable, will indicate considerablerevision of the present mockup design.It is believed that steam dilution upto the coil just below any burningregion wil 1 prevent serious flashbacks.A similar system with catalytic recombiners between condenser stages isalso being designed for study.

27

28

CATALYTIC RECOMBINER -

MAIN CONDENSER- _

STEAM-WATERSEPARATOR

STEAM DISTRIBUTOR

UNCLASSIFIEDDWG. 19534

COOLING WATER

AFTER CONDENSER

LEVEL CONTROL

150-lb STEAM

Fig. 17. Mockup of Core-Condenser-Becombiner Systen

•*®&Hm6*»*>mMi

Part III

GENERAL HOMOGENEOUS REACTOR STUDIES

INTERMEDIATE-SCALE HOMOGENEOUS REACTOR DESIGN

R. B. Briggs, Section Chief

R. H. ChapmanW. R. Gall

C. L. SegaserF. C. Zapp

D. K. TaylorP. N. Haubenreich

M. Tobias

R. E. Aven

L. Cooper

W. TerryJ. P. Sanders

INTERNAL RECOMBINATION CATALYST

It has been shown that cupric ionsin a fuel solution effectively catalyzethe recombination of hydrogen andoxygen. The use of copper as aninternal recombination catalyst in theISHR has obvious advantages. Sufficient copper would prevent formationof gas bubbles in the core, eliminateany explosion hazard, and make a gasseparator and external recombinersystem unnecessary during normaloperation. Disadvantages of copper inthe system include parasitic absorptionof neutrons, possible acceleratedcorrosion, and the rapid decrease inefficiency in effecting recombinationas the temperature is reduced.

In order to estimate the effectof copper in the ISHR, calculationshave been made to determine the amountof copper required and its effect oncritical concentration and conversionratio. If a homogeneous catalyst isused, the maximum concentration ofhydrogen would occur at the coreoutlet. The hydrogen concentration atthis point has been calculated as afunction of copper concentration by amethod which takes into account thechanging temperature of the solutionas it moves around the circulatingsystem. Figure 18 shows the totalpressure required to keep the gas insolution at 250°C, if a l-to-2 ratioof oxygen to hydrogen is assumed.Also shown is the mole fraction ofhydrogen and oxygen (in stoichiometricratio) in the vapor phase at thecorresponding pressure and 250 C. Inorder that the mixture be nonexplosive,this fraction must certainly be lessthan 0.3 and probably less than 0.2.Because the effectiveness of copper

is quite sensitive to temperature,the results are dependent upon theassumed distribution of temperature inthe core. Uncertainty as to theactual temperature distributionaccounts for the band in the figure.Nuclear effects were calculated fortwo copper concentrations: 0.2 and0.4 mole/liter. Results given beloware for the two-region converter witha 1/4-in.-thick stainless steel coretank and a slurry blanket containing1000 g of thorium per liter.

NUCLEAR EFFECTS OF COPPER ADDITION

IN THE ISHR

Copper concentration,

mole/liter

Critical uranium

concentration,

g/liter

Conversion ratio,

U23,235

t II233atoms oi U per

atom of U

0.2 0.4

4.79 4.93 5.50

0.79 0.75 0.72

The effect of copper additions onthe conversion ratio has not beendetermined for the 1/8-in.-thick coretank and for the higher concentrationsof thorium in the blanket reported inprevious quarterly reports. A qualitative examination of the conditionsof interest indicates that the percentage change in conversion ratioshould be nearly constant for a givencopper concentration.

EVAPORATION OF D20 IN THE GAS SEPARATOR

In operating the ISHR, it isnecessary to evaporate D20 from thefuel solution in the circulatingsystem to change the fuel concentration, to provide D20 for feeding

31


800

32

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

CUPRIC ION CONCENTRATION (mole/liter)

Fig. 18. Effect of Copper on TotalPressure and on Vapor-Phase Composition.

into the bearings of the pumps, and tooperate valves or other mechanisms inthe system. During normal operationthere is a continuous supply of about9 gpm of D20 from the recombiner.A high-pressure evaporator has beenshown on the flow sheets to supply1.5 gpm of D20 when the reactor isoperating at very low power. Thisevaporator presents problems inminimizing holdup and in corrosion andstability of the fuel solution. Theevaporator is operated at a temperatureof at least 285°C, and the oxygen israpidly stripped from the boilingsolution.

A study has been made to determinewhether an equal quantity of D20 mightbe obtained by subcooling the gaswhich is recirculated through therecombiner system so that D20 would be


evaporated into the gas in the gasseparator. The necessary heat wouldbe supplied to the fuel solution bysupplying steam to the main heatexchanger.

The amount of D20 that would beevaporated in the gas separator atlow power levels (no gas production)was calculated for various temperatures. The calculations were basedon the following assumptions:

1. The gas is saturated with D20at the entering temperature.

2. The change in the temperatureand quantity of the liquid streamin passing through the gas separatoris negligible.

3. The flowofgas in the separatoris 310 cfm for each calculation.

Two methods of calculation wereemployed which gave comparable results.Each method involved the calculationof film coefficients based on empiricalequations developed for wetted-wallcolumns. In method I of the calculations, the individual heat and masstransfer coefficients were calculatedseparately, the total mass and heattransferred were determined, and theresultant temperature and amount ofD20 evaporated were found by trialand error. In method II, coefficientsinvolving an enthalpy driving forcewere calculated from the same empiricalrelationships, and the resultantconditions were determined. Theresults of these calculations areshown in Table 1. It appears that asufficient quantity of D20 can beevaporated from the gas separator attemperatures above 150°C.

COOLING-TOWER REACTOR

In considering the evaporation ofD20 in the gas separator, the possibility was suggested that all the heatmight be transferred from the fuelsolution by evaporation of D20 into arecirculated gas stream. A preliminarystudy has been made of a 50-megawattreactor in which the heat is removedfrom the fuel solution in a cooling


TABLE 1. CALCULATED CONDITIONS OF OPERATION OF GAS SEPARATOR

CASE I CASE II CASE III CASE IV

Temperature of liquid, C

Entering gas temperature, C

Pound moles of gas entering per hour

Rate constants:

k , lb-moles/hr*ft2'atmh, Btu/hr*ft2,°F

K', Btu/hr'ft2'A enthalpy

Method I

Exit gas temperature, C

D20 evaporated, lb-moles/mingpm

Method II

Exit gas temperature, C

D20 evaporated, lb-moles/mingpm

UO2S04 -D20

37,(40 lb/min

482°F

(5000gpm)

COOLING TOWER

ri_L •5

2 50

230

1894

0.268

64.6

286

232.9

0.714

1.95

233.6

0.874

2.39

250

200

2014

0.242

69.6

301

207.8

1.340

3.67

210.0

1.721

4.71

4,480 lb OF 02 PER min

1.106 lb OF D20 PER lb OF 02455° F

I f

o 4,480 lb OF 02 PER min CONDENSER

0.299 lb OF D20 PER lb OF 02

390° F

2700 cfm

U02S04-020

37,140 lb/min

4I4°F

U02S04-D20

33,520 lb/min

416°F3,620 lb OF D20 PER min

390°F

(460 gpm)

200

150

2252

0.181

70.6

327

155.8

0.431

1.18

155.9

0.446

1.22

150

100

2553

0.179

76.4

369

104.9

0.120

0.33

104.4

0.107

0.29

DWG. 19536

SATURATED STEAM

2,630 lb/min

215 psia

388°F

2,630 lb OF H20 PER min

150°F

CONDENSATE STORAGE

Fig. 19. Flow Sheet for Cooling-Tower Reactor.

33


tower. The solution is cooled byevaporation of part of the moderatoras it passes countercurrently toan oxygen stream in a packed or slattedtower. The gas is cooled, and part ofthe vapor is condensed in a heatexchanger producing steam. Such asystem should reduce corrosion problemsin the heat exchanger and eliminatethe gas separator, D20 evaporator, andhigh-pressure storage tank.

It was found that the requiredoxygen flow and tower height increaserapidly with system pressure because,primarily, increased pressure reducesthe amount of moisture associated withthe gas at a given temperature. Onthe other hand, gas bubbles in thecore make higher pressures necessaryfor reducing their volume. Use of aninternal recombination catalyst would

prevent excessive bubble formation andpermit operation at lower pressures.

Figure 19 represents a systemoperated at 700 psia. At this pressure,gas bubbles would occupy 7.8% of thecore volume unless an internal catalystwere used. Figure 20 shows a possiblearrangement in which the core isinside the vessel containing thepacking, that is, stacked 3-in.Raschig rings. Designed for theconditions in Fig. 19, the bed is7 ft wide, 6 ft deep, and about 30 ftlong. The condenser, as shown, has2200 seven-eighths in. tubes 25 ftlong. Solution holdup is about 3300liters. In addition, approximately1000 kg of D20 is held up as vapor orin the condensate return line. Thisholdup compares favorably with the4000-liter holdup of solution in thepresent ISHR.

ENGINEERING DEVELOPMENT

C. B. Graham, Section Chief

CORE AND GAS SEPARATOR DEVELOPMENT

J. A. Hafford I. SpiewakL. B. U R. H. Wilson

ISHR Core. The best flow arrangement for a high power - densitycore appears to be slug flow, with anequal temperature rise in each streamline. In addition to orderly heatremoval, this type of core has otherimportant advantages: low pressuredrop, good gas removal, and a lowturbulence level.

Slug flow is obtained in the coreproposed for the ISHR by means ofscreens mounted in the entrancesection, as shown in Fig. 21. Flowtests on an 18-in. plastic modelindicate that this arrangement willbe quite satisfactory.(^ Based on

C. B. Graham et al. , HRP Quar. Prog.Jan. 1, 1953, ORNL-1478, p. 45-60.

34

Rep.

the experimental data, the pressuredrop through the ISHR core at ratedconditions is estimated to be 1.92 psi.

The screen specifications have beenreported in an ORNL memorandum(2)that also includes a calculation ofthe maximum tensile stress caused inthe screens by the flow; the calculatedvalue was 5900 psi. It is believedthat this is a safe value, especiallysince the calculation was quiteconservat ive.

Flow tests on a full-scale ISHRcore are planned for later this yearto confirm the design. These testswill be followed by pressure tests tooptimize the mechanical design of thecore tank with respect to neutronabsorption.

(2)L. B. Lesem, Sp eci fi c at ion s and Stress

Calculations for ISHR Screens, ORNL CF.53-2.93(Feb. 19. '"-»<

on s fo i1953).

FUEL CIRCULATINGPUMP

FUEL COLLECTIONHEADER

GAS DISTRIBUTION

HEADER

ENTRAINMENT SEPARATOR

FEEDWATER INLET-

Fig. 20. Cooling-Tower Reactor.

STEAM OUTLET

CONDENSER

13cn98M

oo

w2DM

2!

90Oac

w

V0

cn

CO


12-in. SCHEDULE 20 PIPE-

DWG. 19538

48-in.-ID SPHERE

2V2 MESH, 0.120-in.-DIA.WIRE SCREEN

>Z% MESH, 0.092-in.-DIA.WIRE SCREENS90-deg INCLUDED-ANGLECONICAL DIFFUSER-

12-in. SCHEDULE 20 PIPE

Fig. 21. ISHR Core.

Eight-Foot Rotational-Flow Model.The construction of the 8-ft rotational-flow model described inORNL-1280<3> has been completed.Operation is expected to begin duringthe next quarter.

(3)C. B. Graham et al., HRP Quar. P

Mar. 15, 1952, ORNL-1280, p. 151.rog. Rep.

36

Pipe-Line Gas Separators. The useof straight-through flow in the ISHRcore necessitates development of agas separator for the external circulating system. It is desired toproduce high separating efficiencywithout there being a great sacrificein pressure drop and liquid holdup.

The system believed to meet theserequirements best is a pipe-line


separator that uses a volute or system. However, if a layout requiresstationary vanes to give the liquid a two 90-deg elbows, volutes may be usedrotating component. The rotation without there being any sacrifice incentrifuges gas into a central void holdup.from which it is removed. Following The relatively high pressurethe gas removal section, most of the losses in a volute used to impartremaining rotational energy can be rotation to the fluid are caused byrecovered with either vanes or a the increase in rotational velocityvolute# as the fluid moves radially inward,

Operation of a 5-in. pipe-line as shown in Fig. 22. A modifiedseparator was described in previous volute has been built that eliminatesquarterly reports. (4' !) This model this radial velocity, as indicatedhas been tested with both vanes and in Fig. 23. Fluid enters the pipevolutes to give rotation and pressure from a rectangular inlet, which is onrecovery. The following conclusions a chord of the pipe, and is deflectedwere drawn from the investigation: axially by a helical vane. As a

1. Both vanes and volutes provide result, there is little radial movementa gas-separating efficiency of close of liquid and minimum pressure loss,to 100%, provided the gas takeoff is Preliminary tests of this modifiedproperly designed. volute indicate that gas separation is

2. Either vanes or volutes may be excellent and that the pressureused for pressure recovery with about losses are only slightly higher thanequal success. The estimated ef- those for an equivalent set of vanes,ficiency is about 70%. Although development of vanes and

3 The over-all pressure drop volutes appears to be well under way,across the separator can be limited the ISHR gas separator cannot beto two inlet velocity heads, or 5.4 specified until a suitable gas takeoffpsi, at an inlet velocity of 20 fps. is developed The gas takeoff can beThe over-all pressure drop will be located at either end of the separator,lower when vanes are used to produce and its size and shape will determinerotation than when a volute is used. void size and gas separating ef-

4. The lengthof separator required ficiency. In principle, it shouldk «- 3 *- ~ r\ ninP diameters be sized so that all the gas comes oilis about o too pipe diameters, _, • j • „j;^^^„c without liquid entramment. independing on design conditions. wiuuuui- h

5 The gas void produced by a practice, some liquid must be with-volute is more stable than that drawn with the gas as a means ofproduced by vanes, since the wakes controlling the size of the vortexproduced at the trailing edge of the If a small centrally locatevanes adversely affect stability. takeoff is used, a small void withThis is particularly true when the high gas -separat ing efficiency isvoid diameter is more than half the obtained. Applied to the ISHR system,pipe diameter. this takeoff would be large enough to

6 Vanes are much easier to remove gas formed in the core but tooconstruct than volutes, particularly small to allow dilution toa non-if the latter must withstand 1000 psi. explosive mixture in the void of the

7. For most layouts vanes require gas separator,less holdup of fuel solution, since If a large void that will alsothey fit directly into an existing serve as a pressurizer is desired,pipe in the high-pressure circulating some sort of annular takeoff is

required to maintain a void largeT7\ — enough to accommodate liquid expansion. lC-^^£r,lLal;-6H2RP Quar- Pr°S- *'" under all operating conditions.Oct. 1, 1952, ORNL-1424, p.

37

L?feis£&^fti&fflvff£Ww£eW!ti^M*i¥a^3=;i


38


Fig. 22. Volute.

Although several annular takeoffs havebeen tested, none have proved satisfactory, since gas removal efficiencyhas been below 80%. Several newdesigns will be tested during thenext quarter.


3O

HELICAL VANE

Fig. 23. Helical Volute.

*wmmmmmm

RECOMBINER DEVELOPMENT

J. A. Ransohoff I. SpiewakS. R. West

Tentative specifications have beenprepared for a high- pressure ISHRrecombiner. The recombiner consistsof platinum deposited on aluminapacking arranged in a fixed bed2 1/2 in. deep and containing 1.75 ftof packing. Since it is desired tooperate with a safe, nonexplosivegas, the off-gas from the gas separatoris diluted to 2 vol % D2 before beingfed to the recombiner. The pressuredrop through the bed under theseoperating conditions is estimated tobe 0.25 psi.

The recombiner design is based onthe assumption that deuterium andoxygen diffusion to the catalystcontrols the reaction rate. This hadbeen suggested by two MIT PracticeSchool groups*5,6) for the flow range

* *T. W. Costikyan, C. B. Hanford, and D. L.Johnson, The Catalytic Recombination of Hydrogenand Oxygen, KT-130 (June 2, 1952).

ORIFICE VALVE

H cx*—

-txy-

-M-

STEAM

RUPTURE DISK-

LEGEND :

(?) PRESSURE GAGE(j) THERMOCOUPLE(d?) DIFFERENTIAL PRESSURE GAGE


used in the design (630,000 scfh ofgas per ft of catalyst). The theoretical efficiency of recombination inthe bed is 99.99%; however, practicalconsiderations .will probably reducethis.

Since the existing data on catalyticrecombinerswere obtained at atmosphericpressure in clean beds, furtherdevelopment is required to increaseconfidence in the proposed design. Arecombiner test unit (Fig. 24) hasbeen built that is capable of handlingup to 2 scfm of hydrogen. The effectsof the following variables will bestudied:

1. pressure,2. hydrogen and oxygen concentration,3. deuterium behavior as compared

with that of hydrogen,4. traces of iodine,5. traces of entrained uranyl sulfate.

In addition to being used forevaluating the packed bed, the test

( J. V. Gaven, R. B. Bacastow, and A. C.Herrington, Catalytic Recombin ation of H% and02-II, KT-134 (Aug. 29, 1952).

-oo-AIR

A® ®

RECOMBINER

<dP>

a

UNCLASSIFIED

DWG. 19571

ANALYZER

~®SILICA

GEL

-CxJ-

Fig. 24. Flow Diagram of Catalytic Recombiner Test Unit.

39


unit will be used to evaluate the

feasibility of a tube - and - she 11recombiner, which utilizes stainlesssteel tubes with a platinized, porous,stainless steel surface. If this

design is feasible, the heat of thereaction would be rapidly removed fromthe surface where it is generated,and, as a result, less diluent gaswould be required.

Some preliminary runs have beenmade in the test unit with the use of

a packed-bed catalyst. Pressures havebeen varied from 25 to 75 psia, andhydrogen concentrations from 2 to 7%have been used. The low-pressureresults thus far obtained appear to bein reasonable agreement with the MITdata referred to above. Raising thepressure increases recombinationefficiency, and therefore design ofhigh-pressure recombiners is, apparently, justified on the basis ofdata obtained at atmospheric pressure.

GAS CIRCULATOR FOR ISHR RECOMBINER

SYSTEM

W. L. Ross

Recent calculations' ' have shownthat the ISHR will produce at 48-megawatts power approximately 1.0lb-mole of deuterium per minute,which corresponds to a G value of1.67. Experimental data indicatethat this assumed G value is con

servative. Studies made at Battelle ôn explosion limits of hydrogen-oxygen-water systems indicate that agas-vapor mixture which has a hydrogenpartial pressure of 2% of the totalpressure will probably be nonexplosiveat the ISHR design pressure and temperature. Consequently, there must be agas-vapor circulation rate of about50 lb-moles/min to render the dis-

(7)J. P. Sanders, Gas P roduction in the ISHR,

ORNL CF- 53-1-110 (Jan. 6, 1953).(8)

H. A. Pray, C. E. Schweickert, and E. F.Stephan, Explo sion Limits of the Hydrogen-Oxygen-Hater System at Elevated Temperatures, BMI-705(Not. 1, 1951).

40

•iipiiiiiiiiiwmiiriniiiiiiiii

sociation gases nonexplosive and tomaintain a safe recombiner system.

The gas-vapor mixture is withdrawnfrom the circulated reactor fuel by agas separator and it passes through anentrainment gas scrubber which removesentrained moisture. From the scrubber,the mixture is delivered to the

catalytic recombiner followed by agas cooler-condenser. A gas circulatoror blower then forces the gas-vapormixture into the gas separator whichagain picks up dissociated gases anddelivers them through the recombinersys tern.

Preliminary consideration of arecombiner system which employs asmall void gas separator, a gasscrubber, a packed-bed type of catalyticrecombiner, a condenser which condensesa small portion of the vapor, andappropriate piping indicates thatthe pressure drop required to circulate50 lb-moles/min of gas mixture is inthe range of 2.5 to 5 psi. A moreaccurate calculation of the pressuredrop cannot be made at the presenttime because none of the systemcomponents have been designed.

The service requirements of the gascirculator needed to provide sufficientgas-vapor flow under ISHR designconditions can be stated on a tentative

and approximate basis, as follows:

Capacity 1250 lb/min (~470 cfm)

Developed head 135 ft at 2.5 psi to 270ft at 5 psi

Suction pressure 990 psia

Gas temperature 480 F

Gas density 2.7 lb/ft

Fluid horsepower 5 hp at 2.5 psi to 10 hpat 5 psi

Gas Steam and oxygen withradioactive contamination

Leakage No leakage to atmospherepermissible

There should be no problem infabricating a circular impeller tomeet the requirements because the

impeller can be quite similar to theusual pump impeller. The degree ofcorrosion attack should be smallcompared with that for a fuel-mediumpump. A semistandard canned-rot or-pump motor can probably be used fordriving the impeller. The motor willbe mounted vertically below theimpeller, and condensed D20 from thegas system will flow through themotor; therefore, water-lubricatedbearings can be used in the motor.A water-gas interface wouId be providedby an overflow below the impeller.Preliminary discussions of thisarrangement with prospective manufacturers indicate that a blowerdesigned for operation in this mannerwill be relatively easy to develop.

PULSAFEEDER PUMP TESTS

P. N. Stevens C. D. Zerby

Diaphragm Endurance Tests. The

diaphragm endurance test' * in whichcompressed air is used to move thediaphragm was continued until tenmillion cycles were accumulated. Thediaphragm was tested periodically forleaks, and none were detected throughout the duration of the test. Visual

examination of the diaphragm showed noobvious cracks; however, some smallpits appeared near the outer edge.The Metallography Section is studyingthe pits and looking for incipientfatigue failure in the diaphragm.

Diaphragm Stress Analysis. The

present, calculated value for maximumstress in the diaphragm is 20,000 psiand the experimental value is 24,000psi. Further work must be done toestablish confidence in the results

obtained both mathematically and withstrain-gage experiments. Tests todate have been run on a diaphragm0.019 in. thick. The next step willbe to study a series of thickerdiaphragms.

(9)J. R. McWherter et al. , HRP Quar. Prog. Rep.

Nov. 15. 1951, ORNL-1221, p. 18.


STELLITE PISTON PUMP TEST

J. S. Culver

The test' ' of the pump with aStellite No. 3 piston and cylinderhas operated a total of 500,000 cyclesvertically on distilled water, 500,000cycles horizontally on distilledwater, and 450,000 cycles on uranylsulfate solution containing 40 g ofuranium per liter at room temperature.Visual observation and photographsshow no evidence of galling of eitherthe piston or the cylinder. Rubbingsurfaces are being polished to amirror finish.

There was no visible evidence ofcorrosion after any of the runs;however, a brown, gummy materialappeared in the system after operationon uranyl sulfate. This material hasnot as yet been identified, but it isthought to be merely an accumulationof oil from the air supply. The uranylsulfate solution is perfectly clear.The total time of operation on uranylsulfate solution was 504 hours.

TEST LOOP FOR 4000-gpm PUMP

J. S. Culver

Construction of the loop ' fortesting the 4000-gpm pump has beenstarted. Design of the electricalsupply for the pump is 50% complete,and material for this work will be

ordered immediately. The constructioncan proceed to completion with littledelay, and the proposed completiondate is August 1, 1953.

ISHR HEAT EXCHANGER

W. L. Ross

L. F. Goode

Graham

R. B. BriggsW. R. Gall

C. B

Recent discussions with prospectivefabricators of the ISHR heat exchangershave emphasized the major problemsinvolved in obtaining a suitableexchanger.

41


Tube Sheets. Considerable diffi

culty has been experienced in obtaininglarge, high-quality, type 347 stainlesssteel forgings or plate for tube-sheetfabrication. The use of type 304stainless steel or stainless cladsteel for the sheets and other heavyparts may present a possible solutionto this difficulty.

Tube and Tube Joints. A small

percentage of the stainless steeltubing used, both seamless and welded,has been found to be faulty by heatexchanger manufacturers. Specialprecautions should therefore be takento test the tubing adequately. Sinceno inspection procedures have beenestablished, techniques will have tobe developed to cull the faultytubing. Tube-to-tube-sheet weldingtechniques require further developmentto eliminate nonuni formi ty. Thesolution to this problem may tie inwith the development of an automaticwelding machine. The corrosionresistance of rolled or flared tubejoints must be investigated.

SMALL CANNED-ROTOR PUMP DEVELOPMENT

Allis-Chalmers 5-gpm Pump (W. L.Ross). As stated in the previousreport, ' ^ some pump parts are subjectto corrosion by fuel solution. Therefore, five Allis-Chalmers 5-gpmpumps, similar to the one that hasbeen operated satisfactorily withwater, have been ordered for generalutility purposes in corrosion programsand in the development of an in-pileloop. Fabrication should not takelong because some of the criticalconstruction materials (stainlesssteel) were supplied to the manufacturer. The five pumps will beconstructed so that the pump casingcan be closed by either a metal gasketor a seal weld. Otherwise, the pumpsare the same as the model which has

been tested.ORNL Small Pump (C. D. Zerby,

T. H. Mauney, C. B. Graham). Two,

42

small, canned-rotor pumps (describedin the previous quarterly report, >are being fabricated for test purposes.One of the pumps will have a rotorwhich will be machined from a solidrod of type 410 stainless steel; theother will have a standard rotor whichwill be canned in type 347 stainlesssteel. The discharge from one pumpwill have the normal tangential outletfrom the volute; the other will havean axial outlet. The axial dischargeis to be used on the pump for thein-pile loop described in the sectionof this report on "Corrosion.

Development of a stator from astandard 1/2-hp motor to operate thepump is now in progress. Cooling ofthe stator laminations and coil willbe attempted by running water coolingtubes through the cooling holes in thestator iron; an alternate method ofair cooling will also be investigated.

CONVERSION OF THE HRE MOCKUP TO OPERATEWITH URANIUM OXIDE SLURRY

C. D. Zerby W. L. Ross

Prior to converting the HRE mockupfor operation with uranium oxideslurry, a series of component testsof various parts of the system wereplanned. The first component testconsisted of pumping slurry througha low-pressure loop with a 1-gpm Lapppulsafeeder pump; the results of thistest were reported in the previousquarterly progress report. Subsequent alterations to the low-pressureloop converted it for high-pressureoperation, and it now includes alet-down heat exchanger and let-downvalve to mockup the let-down systemof the HRE.

Considerable difficulty was encountered with the slurry settling inthe suction line to the pulsafeederpump during shutdown; if the suctionline was made small, the line wouldplug; if the line was made large, theslurry would not resuspend uponstartup because of low velocities.

This problem has not been solved, butit is believed that continuous circu

lation to the dump tanks through thesuction line of the pulsafeeder pumpwhen it is not in operation may bethe solution.

Samples of the mixture beingcirculated were taken periodicallyduring several days of intermittentoperation. A chemical analysis of thesamples indicated a variation ofapproximately 40% in concentration at2 00 and at 50 g of uranium per literof mixture. This variation in concen

tration is attributed to the poormixing conditions in the loop dumptank. It is apparent that a recirculation pump loop on the dump tank isnecessary to maintain a consistentconcentration in the tank and, therefore, in the system.


Helium was injected into the topof the let-down heat exchanger duringone run to observe whether unstableflow conditions would be created inthe let-down system which would giverise to oscillations in pressure inthe system. At room temperature,800-psig pressure, a solution concentration of 150gof ur anium per li ter ,and a flow rate of 1. 146 gpm from thepulsafeeder pump, a ga s-to-1 i qu i dvolume ratio of 7. 076 was achievedwith no observable oscillations.

The let-down valve plug and seatwere removed and observed for wear.After 95 hr of operation at concentrations of from 50 to 300gof uraniumper liter, only the very slightestindications of the flow pattern wereobservable on the valve plug. Workhas stopped on this program because oftime demands of other developmentalwork.

CORROSION

E. G. Bohlmann, Section Chief

PUMP LOOP OPERATION AND MAINTENANCE

H. C. Savage R. A. LorenzD. Schwartz

Loop Status. Ten dynamic corrosiontest loops have been in almost continuous operation during the pastquarter. Major emphasis has been onoperation with solutions of uranylsulfate under various conditions;only one loop has been in operationwith uranyl fluoride solution. Dynamiccorrosion tests were conducted, aspreviously described, ' on variousmetals and alloys in the form ofround-pin and flat-coupon test specimens.

Loop F was dismantled and cut openfor examination of the interior of the

(l) H. C. Savage it al., HRP Quar. Prog. Rep.Mar. 15, 1952, ORNL-1280, p. 46.

pipe. In view of the operating historyof this loop (~2100 hr with uranylsulfate solutions containing 100 to300 g of uranium per liter at 250 C),the interior surfaces of the 1 1/2-in.type 347 stainless steel pipe were ingood condition. The straight sectionsof pipe were coated with a smooth,black film which is very difficultto remove, and there is no evidence ofpitting. The welds in the straightsections cf pipe were also in goodcondition (Fig. 25).

The areas of pitting and corrosionwere confined to a few isolated areas

where there was high turbulence. Asusual in test loops circulatingsolutions of uranyl sulfate of theseconcentrations and temperature, thepump impeller and impeller housingwere severely pitted and corroded.

43


FLOW WELD-

UNCLASSIFIED

PHOTO 20069

!iJKia«Bt&**':"^

Fig. 25. Straight Pipe Section, Loop F.

In addition, a section of pipe, up toa distance of about 5 in. downstream

from a flow restrictor, was pitted andcorroded (Fig. 26), as were the buttwelds of three, forged, type 3 47stainless steel tees downstream from

the change in direction of flow(Fig. 27). The bulk flow rate in theloop was about 17 fps. In particular,heavy, localized attack was found atthe junction of the 1 1/2-in. pipe andthe small 1/4-in. by-pass pipe used todirect solution flow to the pin andcoupon test specimens. No seriouscorrosion was found in the straightsections of the smaller pipes that aremore than 1 or 2 in. beyond the takeoff point. The bulk velocity in the1/4-in. pipes was about 45 fps. Theentire loop has been cross sectionedand mounted on a plywood board; it isavailable for inspection.

44

One new loop, which replaces thedismantled loop F described above, wascompleted and put in operation duringthe past quarter. This loop was madewith 1 1/2-in., schedule 80, type 347stainless steel pipe throughout. Allwelds were made with the inert-gasheliarc method and type 347 stainlesssteel welding rod. The pressurizerwas designed to serve as a gas separator.It contains a tangential inlet andstraightening vanes at the outlet(Fig. 28). The two sections designedto contain sample units were placed sothat one set of sample units could beexposed to a two-phase (gas and liquid)solution and the other to the solution

returned from the gas separator. Loopoperation is as follows: The systemis filled with liquid to a point abovethe top sample line entrance to thepressurizer. Gas, usually oxygen, is


UNCLASSIFIED

PHOTO 20071

Fig. 26. Orifice Plate Section, Loop F.

then introduced to a pressure ofabout 200 psig. This is done tosupply the oxygen necessary forsolution stability and to maintainan excess pressure to prevent boilingof the liquid. With the liquidcirculating, some of the excess oxygen(and steam) from the top of thepressurizer is pumped back into theloop ahead of the pressurizer througha 1/4-in. tube connected to a venturi.An orifice plate in the line ahead ofthe pressurizer provides pressure forthe flow through the top sample unit.The circulating solution passes througha second unit after most of the excess

gas has been removed. However, thissolution still contains gas in excessof that required for saturation.

The purpose of this arrangement isto evaluate the effect on corrosion

and/or erosion of uranyl sulfatesolution containing oxygen concen

trations above saturation. Based on

the calculated theoretical values for

the solubility of oxygen in water, thesolution leaving the gas separatingpressurizer contains about 30% moreoxygen than that required for saturation, whereas the solution enteringthe pressurizer and passing throughthe top sample holder contains threeto seven times the amount required forsaturation. These values can be

varied somewhat by changing the liquidlevel in the pressurizer and the sizeof the orifice plates used in thecirculating line. However, it has notbeen possible to obtain perfectseparati on.

An initial run on this loop is nowin progress with a solution of uranylsulfate containing 300 g of uraniumper liter at 250 C. The calculatedsolubility of oxygen (based on theHenry's law constant for water) at

45


WESTINGHOUSE MODEL100A PUMP^

46

,WELD WELD,

WELD

Fig. 27. Forged Tee, Loop F.

PRESSURIZER AND

GAS SEPARATOR.

Fig. 28. Loop F with Gas Separator.

UNCLASSIFIED

PHOTO 20070

UNCLASSIFIED

DWG. 19556

operating conditions (250°C and 1050psig) is approximately 1250 cm (STP)per liter. Analyses of gas sampleswithdrawn from the loop when it wasoperating at the above conditionsgave the following results:

WITHDRAWAL POINT

Entering pressurizer

Leaving pressurizer

Through top sample unit

OXYGEN CONTENT

(cm /liter)

4500

1800

7600

Thus it can be seen that all the

solution contains excess gas, butthe amount of gas obtained through thetop sample unit is approximately fourtimes that leaving the pressurizer.Since this run is still under way, nocorrosion results have been obtained.

Loop G, an all-titanium loop, hasbeen completed with the exception ofone piece of 1/4-in. tubing for theby-pass line. This tubing has beenshipped by the vendor. The system hasbeen pressure tested at 1500 psig andis ready for operation as soon as thetubing is received and installed.The only section exposed to thecirculating liquid in this system thatis not titanium is the stainless steel

pump stator diaphragm and rotor can.However, the solution temperature inthis region is probably less than125°C. All other sections, includingthe rotor stub shaft, impeller,impeller housing, and sample valves,are of titanium. In addition, thesystem contains a number of titaniumwelds of varying complexity. Thus themechanical feasibility of fabricatingsuch a system of titanium is animportant consideration that must beconsidered along with the study of thecorrosion resistance of the titanium

and its effect on solution stability.Two new circulating loops, identi

fied as D and M, have been underconstruction during the past quarter.Loop D is designed for small volume(~4 liters) operation so that specialsolutions or uranyl sulfate solutions


with additives of limited availabilitymay be used for dynamic corrosiontests.

Loop M is a type 347 stainlesssteel system and differs from testloops previously described' ' in thelocation of the holders for corrosion

test specimens (Fig. 29). The eightspecimen holders' ' provided givetwo to four times the capacity of thepresent systems. In addition, theliquid flow rate through each specimenholder has been increased by theelimination of the small (1/4 to3/8 in.) by-pass lines previouslyused. Thus corrosion rates at wider

ranges of flow rates may be studiedunder identical environmental con

ditions .

Westinghouse Model 100A Pumps. TheWestinghouse model 100A pumps used inthe dynamic corrosion test loopscontinue to give good service. Nopump failures occurred in ten pumpsthat were in almost continuous operationduring the quarter. However, one pumpwas removed from service after approximately 11,000 hr of operation.A routine Dy-Chek of the Inconeldiaphragm gave indications of a smallleak that could not be found by visualinspection. A new stainless steeldiaphragm will be installed beforethis pump is reused. Six pumps withstainless steel diaphragms haveaccumulated an operating time ofapproximately 14,000 hr with nofailures.

Corrosion of the front hubs of

stainless steel impellers, along withthe resulting loss of pump thrustbalance, continues to be the majorsource of concern. When the uranylsulfate concentrations used are above

100 g of uranium per liter and theoperating temperature is 250 C,frequent pump thrust balance inspectionsand adjustments are necessary. Roughly300 to 500 hr is required beforeexcessive corrosion of the front hub

(2)Ibid., p. 41.

47


WESTINGHOUSE MODEL

IOOA PUMPN

.X.

UNCLASSIFIED

DWG 19557

FILLER CAP •

SOLUTION-SAMPLING

VALVE

Fig. 29. Dynamic corrosion Test Loop M.

results in the loss of thrust balanceunder the above conditions. A solution

concentration of 40 g of uranium perliter and a temperature of 250 C mayrequire 2000 to 3000 hr to produceexcessive corrosion of the impellerhub, whereas at 15 g of uranium perliter and 250°C somewhat more than4000 hr are required. In one loop,more than 7500 hr of operation at250°C with uranyl sulfate solutioncontaining 5 g of uranium per literhave been accumulated without excess

corrosion of the impeller hub. Ofcourse, the above-mentioned times

48

will vary with individual pumps,depending on the initial clearancebetween the impeller hub and sealrings (usually 0.010 to 0.020 in.)and the distance the rotor must move

forward before contacting the frontthrust bearing.

The problem of corrosion of theimpeller hub is being corrected by theuse of a titanium impeller or byinstallation of titanium bushings onthe front hubs of all stainless steel

impellers after corrosion has occurredand by the use of either titanium ortantalum seal rings. To date, no

- - ?iw».-i?s'*Sie^w^»!^B(t;^^^%s,

appreciable corrosion of titaniumimpellers has been observed at uranylsulfate concentrations of up to 300 gof uranium per liter and temperaturesof up to 2-50°C.

Heat Exchanger Tests. Operation ofthe heat exchanger test assembly'3îs continuing. To date, approximately2000 hr of operating time with uranylsulfate solutions (5 and 40 g ofuranium per liter) at 250°C has beenaccumulated on two heat exchangertubes (approximately 1000 hr on eachtube). Flow through the tubes hasbeen approximately 18 fps.

Inspection of the assembly showedthat all parts exposed to uranylsulfate solution, including the inletand exit welds, became covered with asmooth, black film. No pitting type ofcorrosion was observed. The outside

of the tube (steam side) was coveredwith a brownish translucent film.

Distilled water was used in the steam

j acke t.Reproducibility of the heat transfer

measurements has not been good. Thepulse-welded thermocouples used tomeasure the temperature drop acrossthe exchanger have given some erraticreadings. He 1 ia re - welded junctionswill be used in an effort to improvetemperature measurements. Severalexperiments are being planned fordetermining the over-all heat transfercoefficient as accurately as theexisting apparatus will permit.

LOOP TEST RESULTS

J. C. Griess R. E. Wacker

The objective of the loop testprogram continues to be that ofobtaining a better understanding ofthe nature of the corrosion processesoccurring in uranyl sulfate (anduranyl fluoride)-steel systems. Inthisregard, studies have been initiated

(3)E. G. Bohlmann et al. , HRP Quar. Prog. Rep.Jan. 1, 1953, ORNL-1478, p. 62-63.


which should ultimately give information on the effect of such variables

as temperature, velocity, acidity, andtime on the corrosion rate of several

alloys. Some of the preliminary dataare given here. Long-term testing ofpromising systems has been continued,and the results obtained from several

such systems are given in the followingsections. Preliminary data on thestability of thorium dioxide pelletsin a dilute, circulating, uranylsulfate solution are presented.

Solution Stability and CorrosionDamage in Dilute Uranyl Solutions.

The fact that dibute uranyl sulfateand fluoride solutions (5 to 15 g ofuranium per liter) undergo substantialhydrolysis in the loops at 250°C hasbeen reported previously.(3) Allsuch systems investigated to date haveshown that 10 to 25% of the totaluranium precipitated as hydrateduranium trioxide, with a correspondingincrease in the acidity of the system.

It has been demonstrated that the

addition of 25 eq. % of the corresponding acid to the uranyl saltsolution at the start of a run com

pletely stabilizes the solutions,allows no precipitation of the uraniumat 2 50 C, and, at the same time, doesnot greatly increase corrosion.Table 2 shows the weight losses ofcorrosion specimens in runs with andwithout added acid.

One loop, run L-13, has beenoperated for approximately 10 00 hrwith a starting solution containing0.02 M U02S04 and 0.05 M CuS04 andhaving a pH of 3.35. The firstsolution sample removed from the loopafter it had reached a temperature of250 C had a pH of 1.8, a copperconcentration of 0.027 M, and theoriginal uranium concentration. Theseconcentrations remained constant for

the remainder of the run, except for aslight increase in nickel concentration. At the conclusion of the

run, a green precipitate with theformula 3CuO•S03•2H20 was found in

49


TABLE 2. EFFECT OF COPPER SULFATE AND ADDED ACID ON THECORROSION RATE OF SEVERAL STAINLESS STEELS AT 250°C

TIME

(hr)

SALT

SOLUTIONS

TYPE OF STAINLESS STEEL

RUN NO. 347 321 304 304 ELC 316 309 SCb

CORROSION RATE (mpy)

B-24 787 0.02 MU02S04 0.54 0.32 0.35 1.00 0.43

and 0.48 0.37 0.37 0.46 0.48

0.005 MH2S04 0.46

0.46

0.32

0.39

L-12 1017 0.02 MU02S04 0.39 0.27 0.32 0.29 0.25

0.25 0.25 0.32 0.30 0.40

0.32

L-13 1017 0.02 M U02S04 0.59 0.28 1.90 0.47 1.3

and 0.40 0.37 0.44 0.42 0.54

0.05 MCuS04 0.42

J-20 1102 0.02 MU02F2 0.02 0.15 0.06 0.03 0.19 0.14

0.15 0.34 0.05 0.05 0.19 0.17

0.09 0.08

0.12 0.06

0.08

J-21 1031 0.02 « U02F2 0.18 ] .0 0.13 0.12 0.18 0.26

and 0.18 0.13 0.21 0.17 0.18 0.30

0.01 M HF 0.20

0.12

0.08

0.18

0.20

the loop. The following final concentrations were calculated from the

analytical data: 0.02 M U02S04,0.027 M CuS04, and 0.015 M H2S04.The weight loss data, given in Table 2,for the corrosion specimens in theloop show that even with the relativelylarge amount of additional sulfuricacid the corrosion damage was notmuch more than that occurring withoutadded acid.

General Corrosion Rates. In the

last quarterly report, ' a listing ofall data obtained from pin-typecorrosion specimens was included.Additional information collected

during the past quarter is given inTable 3. As in the previous report,the corrosion rates are reported inmils of penetration per year (mpy).

50

For runs that contained more than two

pins of a given alloy, only theminimum, maximum, and average ratesare listed. As will be discussed in

a later section, the corrosion ratesvary with time; and when one run iscompared with another, it should beremembered that only runs of similarduration are directly comparable. Allcorrosion rates are based on the

actual amount of metal oxidized, sinceall specimens were defilmed before thefinal weighing.

In addition to the pins describedin Table 3, some pins containingwelds in the center have been exposedto loop solutions. To date, only alimited number of such pins has beenexamined, and the duration of thetests has been short. However, in

TABLE 3. RESULTS OF PIN CORROSION TESTS

TEST CONDITIONS

PIN MATERIALNO. OF

PINS


Uranium Temperature Time Flow RateRUN NO.Solution Concentration

(g/DCO (hr)

Additions (fps± 10%) Minimum Average Maximum

K-12 uo2so, 5 250 2100 200 psi 02 11 Zirconium-2.5% tin„(.>

K-ll and uo2so4 5 250 2915 200 psi 02 11 Zirconium-2.5% tinj(a)

K-12 Zirconium-5% tin j(o)

Zirconium, crystal bar 1 ( •)

K-10 U02S04 5 250 4328 200 psi 02 11 Titanium (RC-70)3(M 0 0.01 0.02

throughK-12

K-5 uo.so. 5 250 6334 200 psi 02 11 Type 347 stainless steel 3 0.02 0.03 0.05

throughK-12

2 4 Type 304 stainless steel 2 0.04 0.05 0.05

Type 309 SCb stainless steel 2 0.03 0.04 0.04

Type 316 stainless steel 2 0.03 0.04 0.05


Worthite steel 1 0.11

Carpenter-20 steel 2 0.13 0.14 0.14

L-12 uo.so. 5 250 1017 200 psi 02 17 Type 347 stainless steel 3 0.25 0.3. 0.392 * Type 304 stainless steel

Type 309 SCb stainless steelType 316 stainless steelType 321 stainless steelWorthite steel

2

2

2

2

3

0.32

0.25

0.29

0.25

1.7

0.32

0.33

0.30

0.26

1.8

0.32

0.40

0.30

0.27

1.9

L-13 uo2so, 5 250 1022 200 psi 02, 19 Type 347 stainless steel 3 0.40 0.48 0.59

0.-05 MCuSO/c) Type 304 stainless steel 2 0.44 1.2 1.9




Worthite steel 1 1.4

B-25 uo2so, 5 250 1038 200 psi 02, 11 Type 347 stainless steel 2 0.74 0.76 0.77

0.005 U H2S04 Type 304 stainless steel 1 0.56

Type 321 stainless steel 1 0.53

Zirconium-2.5% tin2(.)

B-24 UOjSO^ 5 250 787 200 psi 02, 14 Type 347 stainless steel 6 0.46 0.71 1.6

0.005 « H2SO, Type 347 stainless steel, cast 2 0.63 0.71 0.78


Type 304 stainless steel, cast 2 0.67 0.94 1.2



Type 316 stainless steel,cast 2 0.46 0.73 1.0

Type 316 stainless steel, cast 2 0.69 0.85 1.0


Worthite steel, cast 2 2.7 2.8 2.9

FA-20 steel, cast 2 1.5 1.5 1.5

Type 304 M stainless steel'*" 2 0.37 0.38 0.39

13M50M

oD

2

SOOs

cn

K3 TABLE 3. (continued)

TEST CONDITIONS

PUN NO.Solution

Uranium

Concentration

(g/D

Temperature

CO

Time

(hr) Add itionsFlow Rate

(fps ± 10%)PIN MATERIAL

NO. OF

PINS

Minimum Average Maximum

J-20 U02F2 5 250 1102 200 psi 02 11 Type 347 stainless steelType 304 stainless steel

Type 304 ELC stainless steelType 309 SCb stainless steelType 316 stainless steelType 321 stainless steel

Type 310 S stainless steelType 304 M stainless steel*''*Type 310 SM stainless steelU)Type 430 stainless steelType 443 stainless steelTitanium (RC-70)

5

2

2

2

2

2

2

2

2

2

2

3</>

0.02

0.06

0.03

0.14

0.19

0.15

0.09

0.05

0.12

0.40

0.05

0

0.09

0.07

0.04

0.16

0.19

0.25

0.15

0.06

0.15

0.49

0.31

0

0.15

0.08

0.05

0.17

0.19

0.34

0.20

0.06

0.17

0.57

0.56

0

J-21 U02F2 5 25 1031 200 psi 02, 11 Type 347 stainless steel 5 0.08 0.15 0.200.01 M HF Type 304 stainless steel

Type 304 ELC stainless steelType 309 SCb stainless steel

Type 316 stainless steelType 321 stainless steelType 310 S stainless steel

Type 304 M stainless steel(d*Type 310 SM stainless steel1'"Type 430 stainless steel

Titanium (RC-70)

Titanium (Ti-150A)

Titanium (Ti-IOOA)

2

2

2

2

2

2

2

2

2

3'«>

0. 13

0.12

0.26

0.18

0.13

0.25

0.18

0.38

0.64

0

0.17

0.15

0.28

0.18

0.57

0.28

0.19

0.39

0.64

0.04

0.07

0.07

0.21

0.17

0.30

0.18

1.0

0.31

0.20

0.40

0.64

0.10

E-8 uo2so, 15 250 2548 200 psi 02 12 to 16 Type 347 stainless steelType 347 stainless steel, castType 304 stainless steelType 304 stainless steel, castType 304 ELC stainless steelType 304 ELC stainless steel, castType 309 SCb stainless steelType 309 SCb stainless steel, castType 321 stainless steelZirconium, crystal barZirconium, arc-melted spongeZirconium-2. 5% tin

Zirconium-5% tin

7

2

3

2

4

2

2

2

22(a)

2<*>2<«>2<«>

0.08

0.09

0.26

0.28

0.13

0.23

0.41

0.67

0.20

0.39

0.24

0.40

0.32

0.25

0.24

0.48

0.68

0.25

0.72

0.38

0.67

0.35

0.32

0.25

0.54

0.68

0.29

H-10(i) uo2so4 40 250 351 200 psi 02 17 Type 347 stainless steelType 321 stainless steel

3

3

4.2

3.0

4.3

3.3

4.5

4.4

H-ll and uo2so4 40 250 1364 200 psi 02 10 Type 347 stainless steel 3 0.85 2.1 3.3Type 321 stainless steel 3 0.50 0.88 1.4

H-10

through

H-12

uo2so, 40 250 1715 200 psi 02 10 Type 347 stainless steel 4 0.94 1.4 2.0Type 321 stainless steel 4 0.58 0.71 0.91

EC9313

OG>SOHWS3r

SO©

sow

soPI

osoH

tnCO

RUN NO.Solution

1-2 uo„so.

1-1 uo2so4

1-3

A-31

A-32

A-31 and

A-32

A-33

uo,so.

U02SO,

Uranium

Concentration

(g/D

40

40

40

40

40

40

40

TABLE 3. (continued)

TEST CONDITIONS

Temperature

CO

250

250

250

250

250

250

250

Time

(hr)

90 200 psi 02,0.04 U H2S04

599 200 psi 02,0.04 M H„SO,

1036

117

234

351

97

200 psi 02,0.04 U H2S04

psi v*2

added during

run

100 psi 0ed duU.I)

0,

100 psi 02, 02added duringrun"."

100 psi 02, 02added duringrun(t.M

100 psi 02, 02added duringr„n'*."

Flow Rate

(fps ± 10%)

12 to 14

20

15 to 17

16

16

16

16

PIN MATERIAL

Type 347 stainless steelType 304 stainless steelType 304 ELC stainless steelType 309 SCb stainless steelType 321 stainless steelType 329 stainless steelExperimental alloy, R-3280'Titanium (RC-70)

Titanium (Ti-75A)

Titanium (Ti-IOOA)Titanium (Ti-150A)

Zirconium-2.5% tin

Type 347 stainless steelType 304 stainless steelType 309 SCb stainless steelType 316 stainless steelType 321 stainless steel

Type 347 stainless steelType 304 stainless steelType 304 ELC stainless steelType 309 SCb stainless steelType 316 stainless steelType 316 ELC stainless steelType 321 stainless steelTitanium (RC-70)

Titanium (Ti-75A)Titanium (Ti-IOOA)

Zirconium-2.5% tin

Type 347 stainless steelType 304 ELC stainless steelType 329 stainless steel


Type 347 stainless steelType 304 ELC stainless steelType 329 stainless steelExperimental alloy, R-328Zirconium-2.5% tin

Type 347 stainless steelType 304 ELC stainless steelType 321 stainless steelTitanium (RC-70)

Zirconium—2.5% tin

NO. OF

PINS

4

2

2

2

3

2

2

2

1

1

12<o)

4

2

2

2

2


Minimum Average Maximum

99

120

70

140

69

95

99

0

14

25

25

37

14

100

130

100

190

76

110

110

0

0

0

0

16

25

73

48

31

110

140

130

230

88

130

120

0

18

25

120

58

48

6 9.5 11 14

2 11 15 19

6 7.3 21 32

2 66 88 110

2 31 41 50

2 20 21 21

5 8.0 9.2 11 •B2 0 0 0 m2 0 0 0 50

2 0 0 0 M

2<»» OD

2 53 57 60W

D

2 68 68 68

2 71 78 84

M2 25 26 26 22 27 28 29 «2 16 46 76

21 11 >2 25 27 29 SO2 23 24 25 o2 11 15 18 SB2<«>

W

2 46 48 49 H"

2 46 48 49

2 11 23 35I-"

2 0 0.09 0.172U> VO

w

W

cn

RUN NO.Solution

A-33

A-34

A-34

A-35

A-35

A-36 U02SO4

A-36

A-37

A-37

Uranium

Concentration

(g/D

40

40

40

40

40

40

40

40

TABLE 3. (continued)

TEST CONDITIONS

TemperatureCO

250

250

250

250

250

250

250

250

250

Time

(hr)

97

Additions

added duringrun'*."

97 200 psi 0,111

97

94

94

12

12

44

200 psi 0 <"

200 psi 02, 50 ppmCI as KC1

200 psi 02, 50 ppmCI as KC1

200 psi 0„(1»

200 psi 02('>

200 psi O <"

200 psi 0,("

Flow Rate

(fps ± 10%)

100

16

100

15

100

15

100

PIN MATERIAL

Type 347 stainless steel

Type 304 ELC stainless steelType 321 stainless steelTitanium (RC-70)

Zirconium-2.5% tin


Type 304 ELC stainless steelType 321 stainless steelTitanium (RC-70)Zirconium-2.5% tin


Titanium (RC-70)Zirconium-2.5% tin

Type 347 stainless steelType 321 stainless steelTitanium (RC-70)

Zirconium-2.5% tin

Type 347 stainless steelType 321 stainless steel

Titanium (RC-70)

Zirconium-2.5% tin


Type 304 stainless steelType 304 ELC stainless steelType 309 SCb stainless steelType 321 stainless steel



Type 347 stainless steelType 304 stainless steelType 304 ELC stainless steel

Type 309 SCb stainless steelType 321 stainless steel

Type 347 stainless steelType 304 stainless steel

Type 304 ELC stainless steelType 309 SCb stainless steelType 321 stainless steel

NO. OF

PINS

2

2

1

2

2<°>

2

2

2

2

2<°>

2

1

2

2

2<«>

4

2

22U>

3

2

2

2<°>

4

2

2

3

3

4

2

2

3

3

4

2

2

3

3

4

2

2

3

3


Minimum Average Maximu

180

160

0.18

10

5.3

4.7

0

150

120

0. IE

22

27

0

140

130

0

31

68

68

27

86

47

70

67

30

51

59

86

65

67

49

62

91

97

62

110

190

170

160

0.36

19

15

7.4

0

160

140

140

0.27

28

35

0

150

150

0

45

78

80

35

95

49

70

87

39

70

68

87

78

84

71

81

94

120

85

120

190

180

0.53

27

24

10

0

160

150

0.35

39

42

0

160

170

0

62

88

91

45

105

51

70

107

51

81

87

91

100

87

120

97

140

110

120

33SOT3

OG>SOHHSOr-<

sooC5so

O)m

som*0oS3H

cncn

TABLE 3. (continued

TEST CONDITIONS

PIN MATERIALNO. OF

PINS


RUN NO.Solution

Uranium

Concentration

(g/D

Temperature

CO

Time

(hr)Additions

Flow Rate

(fps ± 10%) Minimum Average Maximum

A-38 uo2so4 40 250 98 200 psi 02(" 15 Type 347 stainless steelType 304 stainless steelType 304 ELC stainless steelType 309 SCb stainless steelType 321 stainless steel

4

2

2

3

3

48

52

10

37

15

53

67

47

55

30

60

82

84

66

40

A-38 uo2so4 40 250 98 200 psi 02<" 100 Type 347 stainless steel


4

2

2

3

3

110

110

120

170

140

150

120

120

180

160

160

120

120

190

200

A-39 uo2so4 40 250 209 200 psi 0j«" 15 Type 347 stainless steelType 304 stainless steelType 304 ELC stainless steelType 309 SCb stainless steelType 321 stainless steel

4

2

2

3

3

17

5.6

4.5

12

7

22

15

22

25

16

26

25

39

36

25

A-39 uo2so4 40 250 209 200 psi 02<" 100 Type 347 stainless steelType 304 stainless steelType 304 ELC stainless steelType 309 SCb stainless steelType 321 stainless steel

4

2

2

3

3

140

150

140

160

140

150

160

140

180

190

160

170

140

220

210

A-40 uo2so4 40 250 397 200 psi 02<" 15 Type 347 stainless steel


4

2

2

3

3

2.4

1.5

3.5

2.5

2.3

5.1

1.6

3.9

6.6

5.7

7.4

1.6

4.2

8.9

8.8

A-40 uo2so4 40 250 397 200 psi 02(!l 100 Type 347 stainless steel


4

2

2

3

3

140

170

160

180

140

170

180

200

180

170

180

180

230

190

230

C-15 uo2so4 300 150 997 200 psi 02 16 to 20 Type 347 stainless steelType 347 stainless steel, castType 304 stainless steel

Type 304 stainless steel, castType 304 ELC stainless steelType 304 ELC stainless steel, castType 309 SCb stainless steel

Type 309 SCb stainless steel, castType 321 stainless steelStellite 25

9

2

6

2

4

2

2

2

2

2

2.0

3.4

3.6

4.4

4.5

4.0

0.62

0.56

3.9

5.8

2.8

3.7

4.9

4.9

4.6

4.6

0.65

0.62

4.5

5.9

3.8

3.9

6.4

5.4

4.8

5.2

0.68

0.67

5.0

6.0

13

oD

W2!DIH

21CD

SOnSB

OS

0\ TABLE 3. (continued)

TEST CONDITIONS

RUN NO.Solution

Uranium

ConcentrationTemperature

CO

Time

(hr)Flow Rate

(fps ± 10%)PIN MATERIAL

NO. OF

PINSMinimum

(g/1)Average Maximum

C-16 uo2so4 300 175 308 200 psi 02 16 to22 Type 347 stainless steel

Type 347 stainless steel, castType 304 stainless steelType 304 stainless steel, castType 304 ELC stainless steel

Type 304 ELC stainless steel, castType 309 SCb stainless steelType 309 SCb stainless steel, castType 321 stainless steel

Stellite 25

9

2

6

2

4

2

2

2

2

2

16

19

11

19

10

11

3.2

6.0

14

28

18

20

17

20

18

13

5.1

7.3

15

29

23

21

19

20

25

14

7.0

8.5

16

29

C-17 uo2so4 300 150 100 300 to 500 ppm 02 16 to 20 Type 347 stainless steelType 304 stainless steel

Type 304 ELC stainless steelType 309 SCb stainless steel


8

6

6

7

4

4

6

2.4

5.6

3.6

1.7

5.3

4.3

5.6

3.7

6.5

5.9

2.1

6.4

4.7

6.5

5.3

7.3

7.5

2.6

7.5

5.3

7.3

C-18 uo2so4 300 175 100 300 to 500 ppm 02 16 to 20 Type 347 stainless steelType 304 stainless steelType 304 ELC stainless steelType 309 SCb stainless steel

Type 316 stainless steelType 316 EIC stainless steelType 321 stainless steel

8

6

6

7

4

4

7

16

24

24

9.1

18

20

29

22

29

26

12

21

23

38

25

33

28

15

25

28

43

C-19 uo2so4 300 200 99 300 to 500 ppm 02 16 to 20 Type 347 stainless steelType 304 stainless steelType 304 ELC stainless steelType 309 SCb stainless steelType 316 stainless steelType 316 EIC stainless steel

Type 321 stainless steelStellite 25

8

6

6

6

4

4

6

2

83

64

72

27

68

58

140

100

100

78

120

36

71

61

170

110

110

100

180

43

78

69

220

110

C-20 uo2so4 300 225 101 300 to 500 ppm 02 16 to 20 Type 347 stainless steel

Type 304 stainless steelType 304 ELC stainless steelType 309 SCb stainless steelType 316 stainless steelType 316 ELC stainless steelType 321 stainless steelStellite 25

7

6

6

6

4

4

6

2

100

120

110

66

110

100

110

210

130

140

140

110

130

120

130

220

150

190

190

170

140

140

150

220

C-21 uo2so4 300 250 101 300 to 500 ppm 02

L

16 to 20 Type 347 stainless steelType 304 stainless steelType 304 ELC stainless steelType 309 SCb stainless steelType 316 stainless steelType 316 ELC stainless steelType 321 stainless steel

Stellite 25

8

6

6

6

4

4

6

1

94

140

120

240

210

180

130

240

270

270

350

280

260

150

730

320

370

430

480

320

320

170

3S3"O

oe>soHW

-0S3OOS3Wcncn

«3oS3H

cn-J

Footnotes for Table 3

(a) To date, attempts to defilm these pins have been unsuccessful. The deposit of Fe203 and 0203 on the surface exposed to the solution could not be removed by the electrolyticdefilming technique.

(6) These pins showed dark, purple films after the FeÔ, an<lCr„03 scales had been removed.

(c) After attaining a temperature of 250 C, approximately halfthe copper precipitated as 3CuO'SOj-2H2O and left the solutionwith a pH of 1.8 for the remainder of the run.

* Type 304 M stainless steel is a standard melt of type 304stainless steel which includes small additions of misch metal.

Type 310 SM stainless steel is a standard melt of type 310S stainless steel which includes a small amount of misch metal.

'After removing the Fe.O, and Cr„03 film the titanium pinsshowed very dark, purple films. These films indicated that ifthe run had been continued for a longer period, corrosion damagemight have been apparent from the weight-loss data.

(2)8 All pins in run J-21, except the Ti-IOOA and Ti-150A pins,

had deep-purple films. The Ti-IOOA pin had a somewhat lessintense purple color, and the color of the Ti-150A pin wasintermediate between straw color and purple.

These two pins developed the Fe203 and 0203 deposit but,in addition, each pin was intensively attacked along its longitudinal axis. The attack appeared to be due to an inclusion inthe zirconium.

(i) This run was summarized in the last quarterly report, butit is included here for comparative purposes.

steel.

This is an experimental, low-nickel, ferritic stainless

The oxygen was added through the rear of the pump at arate 5 to 15 STP cm3/min. The purpose in adding the oxygen wasto check the effect of excess gas on bearing wear in the Ti-100Apump, in connection with HRE pump operation.

A titanium impeller was used in this run.

13nsoM

Oo

wzOM

2!O

s>50nas

V0


most cases, these tests, have shownno adverse effects directly attributable to the welds.

A rather extensive investigation ofwelding techniques and methods and ofthe corrosive attack on the welds is

being carried out in conjunction withthe Metallurgy Group, and a morecomplete discussion of corrosiondamage of welds will be given afterthis investigation has been completed.

Corrosion Rate vs. Time and Ve

locity. It has been known for sometime that corrosion rates of stainless

steels in uranyl sulfate solutions at

250 C are not constant but that, ingeneral, the corrosion rates decreaseas the exposure time increases. Thecomparison of the various loop I andloop H runs listed in Table 3 clearlyshows that the longer the duration ofthe run, the lower the corrosion rate.To further emphasize this point, someof the data collected from loop K, inwhich 0.02 MU02S04 was circulated at250 C, are shown in Table 4. 11 c anbe seen that within rather wide limits

the weight losses of all specimens areunchanged from 1000 to 6 00 0 hr ofexposure. Since there were no weight

TABLE 4. WEIGHT LOSSES OF STAINLESS STEEL SPECIMENS EXPOSED TO 0.02 M U02S04

Temperature, 250°CVelocity, 12 to 17 fps

TYPE OF STAINLESS STEEL

TIME

(hr)347 304 316 309 SCb 321

WEIGHT LOSS (mg)*

1001 1. 9 1.3 0. 9 1.3 0.7

1.4 1.2 0.6 1.0 0.6

1.3

1.6

1017 2.3 1.9 1.7 1.5 1.6

1.5 1. 9 1.8 2.4 1.5

1.9

1023 1.5 1.6 2.5 2.5 1.6

2.6 1.6 2.2 2.5 1.4

1.7

1413 2.5 2.3 4.7 1.7

2.2 4.8 3.3 1.8

2005 1.4

1.3

0.9

1.0

2229 3.0

3.2

6334 0.9 1.8 1.2 1.6 1.0

1.0 1.6 1.9 1.3 0.8

*Corrosion rate, mpymg loss x 170

houi

58

losses between 1000 and 6000 hr, it isobvious that during this period thecorrosion rate was zero.

To investigate the corrosion damageof several alloys as a function oftime and velocity, loop A has beenoperated at 250°C with 0.17 MU02S04(40 g of uranium per liter) forvarious periods of time. In thisparticular series of runs, one pin-type sample holder was placed in themain stream of the loop and one in aby-pass line. A tapered coupon holderwas also included in the by-pass.Each pin holder contained 14 stainlesssteel pins, including the followingalloys: four type 3 47, three type

*. sc

• s

3LUTI0N

3LUTI0NVELOCITVELOCIT

Y, 100 fY, 16 f

PS

*

190 mpy

150 100 150 200 250 300 350 400

TIME (hr)

Fig. 30. Weight Loss of Pin-Type

Corrosion Specimens in 0.17 M L02S04.Each point represents the averageweight loss of all the following 14stainless steel pins: 4 type 347, 3type 309 columbium-stabilized, 3 type321, 2 type 304, 2 type 304 ELC.


321, three type 309 SCb, two type 304,and two type 304 ELC. The couponholder containedten type 3 47 stainlesssteel coupons. The solution flowedpast the pins in the main line at arate of 100 fps and in the by-passline at 16 fps. The flow past thecoupons varied from 8. 5 fps at theentrance end to 77 fps at the exitend. At the end of each run, allspecimens were removed, defilmed,and weighed. Then, new specimens ofthe same types were inserted in theloop and the next run started. Theresults of the pin tests are shown inFig. 30 and the results of the coupontests in Fig. 31. The lengths of the

DWG. 19559

56

48

40

32

•5*9i j

Y

""/

98 hr r24

16

' 44hr

8

\i>hr

of |tO 20 30 40 50

VELOCITY (fps)

60 70

Fig. 31. Effect of Time and Velocity

on the Corrosion Damage to Type 347

Stainless Steel in 0.17 M U02S04 at250°C.

59


vertical lines in Fig. 30 representthe maximum spread of the data forfourteen pins, whereas the symbolsrepresent the average of the data forfourteen pins. Since there appearedto be no consistent differences in the

steels tested, it seemed justifiableto plot the data as a unit; however,the individual pin data for these runsare presented in Table 3, runs A-36through A-40.

It is evident from Fig. 30 thatafter a short induction period thecorrosion rate was high and constantin solution flowing at a rate of100 fps, whereas with a flow rate of16 fps, the corrosion rate was approximately zero after the first50 hours. The pins exposed at 100 fpsdeveloped only very thin films,

275

250

DWG .19560

'

225,

1

PER liter, 250°C, 1065 hr /«—» RUN E-8,15g0FU(AS SULFATE) /

PER | tar ?Rn©r o*a.n h, /

175

150

125

100

75

50

u "

* *xj^-y-

-

25X—X

n

10 20 30 40 50 60

VELOCITY (fps)

70 80 90

Fig. 32. Effect of Velocity on the

Corrosion of Type 347 Stainless Steel

in 0.06 M L02S04.

60

whereas those exposed at 16 fps werecovered with the characteristic,heavy, black coatings. In Fig. 31,the coupon data are represented by aseries of curves. To obtain these

curves, the best line was drawnthrough the individual coupon datumfor each run. The ordinate is expressedin weight loss per square centimeter.

Other loops have contained couponsidentical to those used in the above

series of tests. Figure 32 shows aplot of the data from runs E-7 andE-8 in which the uranyl sulfateconcentration was 0.06 M but the

test periods were different by afactor of 2. 5. Figure 33 shows theweight loss of type 347 stainlesssteel coupons as a function of ve-locityof the solution, which contained0.02 M U02S04 with and without addedsulfuric acid or copper sulfate.Figure 34 shows the effect on thecorrosion of type 3 47 stainless steelof adding sulfuric acid to 0. 17 MU02S04 and of varying the solutionflow rate. In Figs. 32, 33, and 34,the weight losses of the individual

3.4

DWG. 19 5 6

1 1 1 1 1 1

3.1 _ l> « RUN B-25, 0.02 M U02S04 ,0.005 M_. ./.-.H2S04,250°C,l038hr /

2.7 _ » « RUN L-13,0.02/tf U02S04,0.05/tf- / -—

CuS04,250°C,t022hr

_ 24Q.

E

Z 2°

z 1-7o

10 17 hr

!

[

1

icno

ou

[ /i/

1.0

0.68 r^^

>• •<

=8—_*-=

0.34

0

Ji

10 20 30 40 50 60 70 80 90

VELOCITY (fps)

90

80

70

60

50

40

30

20

10

0


Corrosion of Type 347 Stainless Steel

in 0.02 M U02S04.

opoo 0W6. 19562

5 B ^—

V*/

2'

1000

5%-

2Mi. N •£- "

100

5t_ / '

2

10

h-l

5 —« « RUN 1-3, 0.17/tf U02S04, 0.04Af H2S04, ZiRUN H-12, 0.17/1/ UOjSO^, 250°C, 1002hr

0°C,1036hr -

' '

1

5 15 25 35 45 55 65 75 85

VELOCITY (fps)


Corrosion of Type 347 Stainless Steelin 0.17 M U02S04.

coupons are plotted on the ordinate,and the lengths of the horizontallines represent the range of solutionvelocity over a given coupon. Thesurface area of the coupons exposedto the solution was 12. 5 cm .

Corrosion Rates as a Function of

Temperature. Until recently, only alimited amount of corrosion data at

temperatures other than 250°C hadbeen collected. During the pastquarter, however, loop C was used toinvestigate the effect of temperatureon corrosion rates. For this study,a 1. 26 M uranyl sulfate solution andan exposure time of 100 hr werearbitrarily chosen. The first runwas made at 150°C, and each runthereafter has been made at a temperature 25 C higher than the previoustemperature. The results of the pintests are shown in Table 3, runs C-17through C-21. Since the experimentis still in progress and the tempera


ture range is being extended beyond2 50 C, no analysis of the data fromthe steel pins will be made at thistime; a complete summary of theexperiment will be included in thenext quarterly report. Similarily,the results of the coupon tests willbe withheld until all the data havebeen collected.

As was indicated previously, acomplication arises in the interpretation of corrosion rates of steel

specimens because of the formation ofprotective oxide films which preventthe corrosion rates under most con

ditions from being constant. However,one alloy, Haynes alloy 25, has beeninvestigated rather extensively, andunder all conditions examined, nofilm of visible thickness has been

observed. As might be expected, runsof different duration have shown that

the corrosion rate is constant. The

results obtained to date from the

data on all tests on this alloy areshown in Fig. 35, where the log of thecorrosion rate is plotted against thereciprocal of the absolute temperature. From the slope of the lineobtained, an activation energy of19,500 calories has been calculated.Of the four points at a temperatureof 250°C (1/7= 1.91 x io- 3), tWowere obtained in 0.06 M U02S04 and theother two in 1.26 JlfU02S04. All theother points were obtained in 1.26 Muranyl sulfate. From Fig. 35 it canbe shown that the corrosion rate of

Haynes alloy 25 is related to theabsolute temperature T by the following equation:

1A ,10.947 - 4260'mpy = 10

Stability of Thorium Dioxide Pellets

in Dilute Uranyl Sulfate Solutions.

At the request of R. B. Briggs, somethorium dioxide pellets were exposedin a circulating uranyl sulfatesolution because of the interest inthe possibility of either thoriumpellets being put into the core or

61


1000

100

o

<norrrro

10

DWG. 19563

TEMPERATURE CO

250 225 200 175 150 1251 I 1 1

1 1 11

o 300 g OF U (AS U02S04) PER literX \

"2- -4' •

1.90 2.00 2.10 2.20 2.30 240

fxio32.50

Fig. 35. Corrosion Rate of Haynes

Alloy 25 in U02S04 Solutions.

the fuel solution being allowed toflow through the blanket. A number ofnearly spherical pellets, which rangedin diameter from 1/16 to 3/8 in. andweighed 75 g, was placed in a cylindermade of stainless steel screen and the

cylinder was inserted in a by-passline of a loop. The flow rate throughthe line was 1 to 2 fps, but sincethe cylinder fit snugly into the line,the flow rate past the pellets wassomewhat greater than the normal flowthrough the line. The pellets wereexposed for 209 hr at 250°C to 0.02 MU02S04 that was 0.005 M in sulfuricacid.

After the pellets had been removedfrom the loop, they were washed in

62

water, rinsed in acetone, and dried at100°C for 2 hours. When reweighed,the pellets showed a weight gain of2 grams. Visual examination revealedthat some residue had been precipitatedon the surfaces of the pellets. Thisprecipitate, along with absorbedsolution in the pellets, probablyaccounts for the observed weight gain.Only one pellet cracked, and it wasone of the smallest pellets. Thebreak seemed to be clean, judgingfrom what appeared to be two piecesof the pellet that were found in thescreen. At present, most of thepellets are being retained so thatthey can be reinserted in a loop in afuture run.

Discussion. The stabilization of

dilute uranyl sulfate or uranylfluoride solutions at 250 C by theaddition of either sulfuric or hydrofluoric acid has been demonstrated.

Although the addition of acid lowersthe room temperature pH of the solutionsconsiderably, corrosion is not significantly increased. This is perhapsnot surprising since, in reality,the addition of acid prior to thebeginning of a run probably produces asolution not greatly different froman unacidified solution that has

undergone partial hydrolysis in aloop operated at 250 C.

A comparison of the corrosion datafrom different solutions has clearlyshown that the corrosiveness of a

solution is not dependent alone on theroom temperature pH of the solution.For example, all runs made with 0.02 MU02S04 solution and 25 mole % acidhave had a pH lower than the pH of a0.17 M U02S04 solution; however, it isevident that the extent of corrosion

was considerably greater in the lattersolutions. In run L-13 in which the

original solution was 0.02 M U02S04containing 0.05 M CuS04, approximatelyone half the copper precipitated,with the formation of 0.015 M H2S04.Even with this large excess of sulfuricacid, the corrosion damage was very

small; however, the effect of thepresence of the cupric ions must beevaluated.

On the other hand, the addition of25 mole % H2S04 to a uranyl sulfatesolution containing 40 g of uraniumper liter greatly increased the extentof corrosion over that observed inthe same solution without acid (seeruns H-10 to H-12 vs. runs I-1 to 1-3,Table 3). With the added acid, thepH of the solution was approximately1.2, which was considerably lower thanthat of the 0.02 M solution withadded acid. Further studies on theeffect of acidity on corrosion byU02S04 solutions of various concentrations are contemplated.

The corrosion resistance of allsteels, and even titanium, to 0.02 MU02F2 solutions with and withoutadded hydrofluoric acid was surprising(see runs J-20 and J-21, Table 3).Although the titanium corrosionspecimens showed little or no weightchange, they had developed films,probably of titanium dioxide, thatwere thicker than those formed in

uranyl sulfate solutions during asimilar exposure. Unfortunately,zirconium and all its alloys testedto date have shown acomplete lackof resistance to all concentrations

of uranyl fluoride at 250°C.

The corrosion data presented inTable 3 are, in most cases, sufficiently clear that discussion ofthem is unnecessary. However, inruns A-31 through A-33 in which oxygenwas added through the rear of thepump, it appeared that the corrosiondamage was greater than that in therunsin which the solution was initiallysaturated with oxygen. When gas wasadded to the system continuously,some gas bubbles that were circulatedwith the solution may have beenresponsible for the increased corrosion. To investigate this effectfurther, a new loop which contains agas entrainer and separator has beenconstructed and will be ready for use


in the near future (cf., previoussection on "Pump Loop Operation andMaintenance," this chapter).

In the previous quarterly report,^3'it was noted that the data at the 40 gof uranium per liter level werediscordant, but no explanation wasoffered. It now appears that threefactors can be used to partly clarifythe data. First, the data representedby the lower curve in Fig. 30 show themisleading impressions that can beconveyed by reporting all corrosiondamage as a constant rate process(at low solution velocities). Althoughthe data in Fig. 30 are for a test ofonly 400-hr duration, the data fromruns H-10 through H-12 in Table 3 aredirectly comparable. In these runs,the weight losses of the specimenswere, within rather wide limits, thesame after 1700 hr as they were after50 hr in the A runs. The data inTable 4 show that the effect of timeis not confined to the 40-g level butis the same at the 5-g level. Similarly, the old loop, loop F, which wasused to circulate 1.26 M U02S04 andU02F2, demonstrated that in areaswhere the velocity was low the corrosion damage, even at the 300-glevel,'was very low after the protectiveoxide film had formed (see section on"Pump Loop Operation and Maintenance*for description of loop).

Second, it can be observed fromTable 3 that, even with the same typeof steel specimens in a run, widevariations in weight losses are bothpossible and probable. The reasonsfor the large differences in differentspecimens machined from the same pieceof steel are not readily apparent, butthe differences may be related toheterogeneities of the steel and/ordifferences in the degree of cold workas a result of slight differences inthe machining methods. Preparationsfor evaluating such factors are beingmade.

Third, there have been indicationsthat the corrosion behavior of the

63


loop itself can influence the extent and/or turbulence in systems con-of attack on the specimens. If some taining 40 g (or more) of uranium perpart of the loop system, such as the pump liter.impeller, undergoes severe attack, the Coupons contained in taperedcorrosion products originating at that holders have been exposed to a numberpoint are rather uniformily spread of different solutions, and Figs. 31throughout the system, and iron and through 34 represent some of thechromium oxides are deposited on the data that have been collected. Althoughspecimens, as well as on the rest of thedata are not always sel f-consi stent,the loop surface exposed to the two generalizations can be made: thesolution. Because of the oxide velocity effect is time dependent, anddeposits, the specimens can possibly the more corrosive the solution, theform protective films faster if the greater the effect of velocity and theloop contains some part which corrodes lower the velocity at which it becomesat a high rate. Hence the data apparent. Experience has shown thatpresented in the previous quarterly when a pronounced velocity effect isreport'3' can be at least partly observed, the coupons that show largereconciled by consideration of the weight losses are covered with veryfollowing factors: time, variation thin films instead of black films,in "identical" specimens, and the Hence, as was the case with pinscorrosion behavior of the system as a tested with high-velocity solution,whole. the corrosion resistance of coupons

The effect of velocity on the seemed closely related to the stabilitycorrosion of several systems has been °f the bulk oxide iilm.demonstrated, if not clearly evaluated. In Fig. 32 it is seen that theFrom the upper curve in Fig. 30, it is velocity effects in dilute uranylevident that if the flow of the sulfate solutions with and withoutsolution past the specimens is suf- acid are not excessive. In thisficiently fast, the corrosion rate, connection, it is interesting to noteafter a short induction period, is that in these systems there was noconstant and high. At present, no damage to impellers, except a slightexplanation can be offered for the damage to the hubs (cf. , section onrelatively low corrosion rate at the "Pump Loop Operation and Maintenance"high velocity used duringthe induction for a discussion of this effect), andperiod, but it appears to be real. that the loop suffered very littleA visual comparison of the pin-type attack elsewhere.corrosion specimens tested with Figure 31 demonstrates how thesolution velocities of 16 and 100 fps shape of the velocity curves variesrevealed that at the lower velocity with time in runs A-3 6 through A-40.the pins developed heavy, black films, There appeared to be a rather uniformwhereas those tested in the higher- pattern in all the runs except A-40.velocity solution had only very thin It is hoped that further investigationfilms. It thus appeared that in the and experience will clarify thehigh-velocity solution the black, inconsistencies noted in run A-40.oxide film could not form on the The corrosion data as a whole, andsurface of the steel and that in the particularly the data represented byabsence of this protective barrier the Fig. 30 and Tables 3 and 4, rathersteel corroded at a very high, constant clearly indicate that the austeniticrate. The same phenomenon could be stainless steels investigated in theresponsible for the usual high cor- wrought condition do not sufficientlyrosion rate of impellers and other differ in their corrosion resistancesparts exposed to high-velocity solution to uranyl sulfate or fluoride solutions

64

at 250 C so that it can be said thatone is attacked less than the others.

It appears that at low solutionvelocities (perhaps up to 20 fps) theaustenitic stainless steels show

extremely low corrosion rates inuranyl sulfate solutions at 250°Cafter a protective film has beenformed.

Although the corrosiond ata presentedin this report are probably moreoptimistic than those reported previously, a number of variables (exclusiveof radiation effects) remains to beinvestigated. Any one of thesevariables could be a critical factor.

Among the variables to be investigatedare the effect of gas bubbles oncorrosion damage, the ability of weldsto resist corrosion, the effect of thepresence of fission products, and theeffect of copper and added acid oncorrosion rates. In addition, variousmetallurgical factors must be investigated. Additional work will alsobe directed toward determining thesuitability of zirconium and titaniumfor container materials. Included in

the titanium investigation will be theoperation of the all-titanium loop.

SMALL-SCALE, DYNAMIC CORROSION

TEST PROGRAM

D. Schwartz G. E. Moore

Progress on the small-scale, dynamiccorrosion test program continues at avery slow pace because of diversion ofthe effort of the personnel involvedto higher priority problems. A relatively simple mechanism having therequired motion to move a slug ofliquid around a toroid was operatedsporadically during the latter part ofthis quarter. No corrosion resultshave, as yet, been obtained from thismechanism.

The Design Group has consideredanother mechanism (described in the

following) which will also give thedesired motion to an assembly oftoroids, and preliminary drawings havebeen made.

PERIOD ENDING MARCH 31, 195 3

Torus Rotator (W. L. DeRieux, J. R.McWherter). A preliminary design hasbeen made of a machine to circulate

hot liquids in a pipe without the useof a conventional pumping unit. Thismachine has been designed on the sameprinciple as that used for circulationapparatus described in NACA PME51D12.(4)

When a torus less than half full of

liquid is rotated about a verticalaxis other than its own, centrifugalforce causes formation of a slug ofliquid which tends to seek the portionof the torus farthest from the center

of rotation, as shown in Fig. 36. Ifthe torus is simultaneously rotated

(4) L. G. Desnon and D. R. Mosher, PreliminaryStudy of Circulation in an Apparatus Suitable forDetermining Corrosive Effects of Hot FlowingLiquids. NACA RME51D12 (June 1951).

UNCLASSIFIED

DWG. 19564

Fig. 36. Liquid Flow in Torus.

65


about its own vertical axis, flow of or completely failed. An explosivethe liquid with respect to the pipe hazard would thus be inherent in thisall will occur, provided thecentrifu- design. Primarily because of thi

gal force on the liquid due to the hazard, the attempt to develop a sys terncrank throw is sufficient to overcome of this type has been abandoned, atfrictional resistance to flow. least for the present.

The proposed machine is shown in The work is now being directedFig. 37. Ten 6-in.-dia tori fabricated toward the development of a systemfrom 3/8-in. pipe are mounted on each that uses CuS04 in solution as theof two diametrically opposed, rotating recombination catalyst. The studiesshafts, which are, in turn, mounted on made to date indicate that, witha large rotating pulley. Rotation of respect to explosion hazards, safethe shafts is accomplished by using a operation of this system can be ac-chain drive that runs on a stationary complished. Corrosion informationsprocket on the central support pedes- obtained with the loop will be affectedtal. Velocities of liquid relative to by the presence of copper ion; but,the wall of from 0 to 25 fps can be even so, it is believed that a measureobtained by using appropriate pulley of the effect of irradiation on cor-speeds and sprocket sizes. A stationary rosion in homogeneous reactor systemsfurnace built around the tori will be can be obtained.capable of maintaining the samples at Loop Development and Design. The250°C. in-pile loop design has been modified

Four openings sealed off with tube to accommodate the addition of copperfittings will be spaced symmetrically sulfate as an internal, homogeneouson each torus to permit insertion of catalyst for the recombination ofsolution and sample pins. hydrogen and oxygen. The major com

ponents of the loop will be acentrifu-IN-PILE LOOP gal pump, a steam pressurizer, a loop-

. p. temperature control system, a "core,"and corrosion sample holders. The

T* Jones loop will operate at250°C and 1000 psi.There will be two corrosion sample

holders in each loop. One sampleAs pointed out previously, one of holder will be held in the center of

the most difficult problems associated the core of the loop in the highestwith the operation of an in-pile test flux region. Arrangements will besystem is that of safely controlling made so that the solution leaving thethe pressures of hydrogen and oxygen core will pass through the holder andformed in the decomposition of water subject the corrosion specimen to theunder irradiation. Originally, it was highest possible short-lived fission-proposed to control these pressures by product concentration. The otherthe recombination of the gases over a sample holder will be placed in thedry catalyst at the pressure of the circulating system so that it will besystem. A venturi type of pump, to- removed from the high flux region,gether with a liquid-gas separator, The two sample holders will be identi-was to be used to effect circulation cal with respect to provisions forof gases over the catalyst. Careful holding the samples. Considerableanalysis of the operation of the effort has been made to pattern theproposed system shows that it would be design of these sample holders afterextremely difficult, if not impossible, the tapered ones used in the criticalto prevent the formation of explosive corrosion-velocity investigations ingas mixtures if the recombiner partly the out-of-pile dynamic corrosion

w

G. H. Jenks

C. B. Graham

R. J. Kedl

66

W.

C. D. Zerby

D.

T. H. Mauney R. A. Rosenberg

s

FURNACE

TORI

STATIONARY SPROCKET

PULLEY

CHAIN

CENTRAL SUPPORTPEDESTAL


MOTOR

Fig. 37. Torus Rotator.


67


studies to provide the best possiblecorrelation between data from the twoprograms. The new sample holder isdesigned to accommodate an anticipatedflow of 50 fps, and it differs in sizefrom the older design. The design isessentially complete, and models willbe fabricated and tested in thedynamic corrosion test loops.

The progress on the centrifugalpump for the loop is described in thesection of this report on "SmallCanned-Rotor Pump Development." Designof the instrumentation to control theloop is nearing completion and orderswill be placed for the necessaryequipment in the near future. Effortis under way to develop operatingmethods and equipment and for handlingthe system after irradiation. Preliminary considerations of this problemhave been made, and detailed designhas been started.

The use of steam pressurization inthe loop has several advantages. First,the main stream can operate with adissolved gas pressure well below thesaturation pressure so that no bubbleformation will take place. The absenceof bubbles reduces the possibility ofexplosion in the system and eliminatesa possible source of corrosion acceleration. Second, the pressure ofhydrogen and oxygen in the pressurizercan be maintained well below the

explosion limit. An increase inpressure from these gases can bedetected, and corrective measures canbe taken before the explosion limit isreached.

Operating Characteristics. The

approximate operating conditions ofthe loop as presently visualized arelisted below:

Volume of liquid in pressurizer 200 mlVolume of gas space in pressurizer 200 ml

Temperature of pressurizer 280°C

Excess oxygen pressure in pressurizer 70 psi

Flow rate through pressurizer 3 ml/secTotal volume in main stream 1000 ml

68

Volume in core

Power density in core (fromfission)

Power density in other portionsof loop (from fission)

Temperature of main stream

Pressure of excess oxygen dissolved in main stream atstartup

Steady-state pressure of dissolved stoichiometric gas inmain stream

300 ml

5 watts/ml

0

250'3C

100 psi

100 psi

The excess oxygen is introduced intothe pressurizer at startup, and theconcentration of the excess oxygen inthe main stream is maintained bycirculating a portion of the streamthrough the pressurizer. The indicatedcirculation rate is more than ample tomaintain a steady-state distributionof oxygen between the pressurizer andthe main stream, even at the highestcorrosion rates which can be foreseen.Of course, the excess oxygen pressurewill decrease as corrosion proceeds,and additional oxygen will have to beintroduced from time to time. Thepressure of the oxygen in the pressurizer will be recorded during operation,and the data will be used as a measureof the over-all corrosion rate, aswell as an indication of when additionaloxygen is needed.

Sufficient CuS04 will be added tothe fuel solution to hold the steady-state pressure of stoichiometric gasat the desired value. It is estimatedthat a copper concentration of 0.01 Mwill result in a steady-state pressureof 100 psi.

The steady-state pressure of stoichiometric gas in the pressurizer will beabout 12 psi under the specifiedoperating conditions. Since the excessoxygen pressure wil1 not exceed 70 psi,the partial pressures of oxygen andhydrogen together will normally notexceed 82 psi, or 8% of the totalpressure. It is probable that in gasmixtures such as those which exist in

the pressurizer, hydrogen and oxygenmust supply more than 18% of the totalpressure before explosions will take


place. (Note: Determinations of the 4. A study was made to determineexplosive limits in these mixtures are whether commercial-grade titaniumbeing made at Battelle.) metal was subject to crevice and

A failure of the copper catalyst stress corrosion attack in uranylwould be followed by an increase in sulfate solutions containing 5, 40,the pressures of hydrogen and oxygen; 100, and 300 g of uranium per liter atsome increase can be tolerated before 95 C.the explosive limit is reached. It is The pertinent data from each ofplanned that corrective measures will these projects are presented in thebe taken when the total pressure of following sections.the system becomes 120 psi greater than Corrosion Resistance of Syntheticsteam pressure. Attainment of the Gems and Spinels. At the request of120-psi value corresponds to an in- R. H. Powell, an investigation wascrease in the stoichiometric gas undertaken to determine the corrosionpressure of 50 psi if no excess oxygen resistance of synthetic titania,has been consumed or to an increase of synthetic sapphire, and magnesium oxide120 psi if all excess oxygen has been spinel in oxygenated uranyl sulfateconsumed. In either case, the partial solutions at 250°C. The purpose ofpressures of hydrogen and oxygen will, the investigation was to find atogether, comprise about 12% of the corrosion-resistant, transparenttotal pressure, and the mixture will material which could be used forbe below the explosive limit. windows in a cell developed for the

determination of uranium concentrations

STATIC CORROSION STUDIES in uranyl sulfate solutions' } by. , adsorption measurements.

ng 1S The specimens for the tests wereStatic corrosion data are summarized supplied by the Guardian Manufacture

here for the following four projects and Supply Co. in the form of small,which were completed during this unpolished stones. The average densityquarter: of the stones was 4.3 g/cm3. The first

1. An examination was made of the set Qf tests was run in uranyl sulfatecorrosion behavior of synthetic gems solutions containing 40 g of uraniumand spinels in oxygenated uranyl per liter at 250°C with an oxygensulfate solutions containing 40 and partial pressure of about 150 psia.300 g of uranium per liter at 250 C. The oxygen partial pressure wasIncluded in this study was the effect produced from the thermal decompositionof synthetic fission-product additions Q£ hydrogen peroxide that was added toon corrosion resistance at 250°C. the test solutions at room temperature.

2. A study was made of the cor- j^ e solutions and specimens wererosion resistance of type 347 stainless contained in type 347 stainless steelsteel and the stability of uranyl autoclaves of 225-ml total capacity,sulfate solutions at 125°C in the Weight losses and calculated corrosionabsence of excess quantities of dis- rates for the specimens are presentedsolved oxygen. Corrosion resistance in ja^\e 5,and solution stability were examined jt was apparent from the data inas a function of uranium concentration. Table 5 that the fused synthetic

3. An investigation was made to titania exhibited the best over-alldetermine the susceptibility of type corrosion resistance; the final cor-347 stainless steel to crevice and rosion rate was 0.3 mpy. The magnesiumstress corrosion as a function oftemperature, uranium concentration (5)w r Da¥enporti Jr. and R. H. Po.ell. HRPtype of solution aeration, and magnitude Quar. Prog. Rep. Mar. 15. 1952. ORNL-1280,of applied stress. P- 74-80.

69


TABLE 5. CORROSION RESISTANCE OF SYNTHETIC GEMS AND SPINEL INURANYL SULFATE SOLUTIONS CONTAINING 40 g OF URANIUM PER LITER

AND PRESSURIZED WITH 150 psia OF OXYGEN AT 250°C

EXPOSURE WEIGHT LOSSES AND CALCULATED CORROSION RATES

TIME Titsin ia Sapphire MgO Spi ne1(days )

mg/cm mpy mg/cm mpy mg/c m mpy

7 -0.05 0.2 -0.50 2.2 -2.13 10.5

14 0.03 < 0.1 0.55 1.3 2.77 6.8

21 0.15 0.2 0.80 1.3 3.38 5.5

28 0. 16 0.2 0.98 1.2 3.70 4.5

35 0.19 0.2 1.06 1.0 4.38 4.3

42 0.26 0.2 1.11 0.9 5.02 4.1

49 0.40 0.3 1.19 0.8 5.50 3.8

56 0.48 0.3 1.22 0.8 6.03 3.7

63 0.50 0.3 1.27 0.7 6.60 3.6

oxide spinel exhibited the poorestcorrosion resistance, with a finalcorrosion rate of 3.6 mpy after 63days. The specimens, except for themagnesium oxide spinel, showed nosigns of corrosion damage. The surfaces of the spinel were somewhatetched and generally roughened intexture.

The same specimens were also subjected to exposure in uranyl sulfatesolutions containing 300 g of uraniumper liter and pressurized with about150 psia of oxygen at 250°C. Thecorrosion data after 63 days areshown in Table 6. The corrosionresistance of the titania was littleaffected by the increase in uraniumconcentration from 40 to 300 g perliter; the steady-state corrosion ratewas similar to that obtained in thesolutions containing 40 g of uraniumper liter, that is, 0.2 mpy. Corrosionof the sapphire and magnesium oxidespinel was greatly increased, however,by the increase in uranium concentration. Corrosion rates for thesematerials were more than four timesgreater than the values obtained in

70

solutions containing 40 g of uraniumper liter.

As a final check on the corrosionbehavior of the three materials, theywere exposed in urany1 sulfate solutioncontaining 40 g of uranium per literand synthetic fission -product additions. The fission products wereadded to the original test solutionsas sulfates. Not all the additiveswere completely soluble in the uranylsulfate; however, sufficient quantitieswere added to give the following concentrations, in ppm: Ce, 43; Ba, 5;Zr, 15; Sr, 6; Ru, 9; Te, 4; Mo, 12;and I, 3. Weight losses and corrosionrates are presented in Table 7 for 56

days of operation at 250°C. Theuranyl sulfate solutions were pressurized with about 150 psia of oxygen ina manner similar to that used in thepreviously described tests. No adverseeffect on the corrosion resistance ofthe three materials was observed uponthe addition of simulated fissionproducts to the corrosion environment.The final corrosion rates obtainedafter 56 days were nearly identical tothe values obtained for a similar


TABLE 6. CORROSION RESISTANCE OF SYNTHETIC GEMS AND SPINEL INURANYL SULFATE SOLUTIONS CONTAINING 300 g OF URANIUM PER LITER

AND PRESSURIZED WITH 150 psia OF OXYGEN AT 250°C

EXPOSUREWEIGHT LOSSES AND CALCULATED CORROSION RATES

TIME Ti t an i a Sapphi re MgO Sp ine1

(days ) mg/cm mpy mg/cm mpy mg/cm mpy

7 -0.07 0.3 -1.73 8.5 -9.77 47.9

14 0.07 0.2 2.30 5.7 14.70 36.0

21 0. 16 0.3 2.92 4.8 17.82 29.1

28 0.16 0.2 3.41 4.2 20.23 24.8

35 0.21 0.2 3.68 3.6 22. 00 21.6

42 0.25 0.2 3.98 3.3 23.82 19.4

49 0.27 0.2 4.23 3.0 25. 19 17.6

56 4.54 2.8 26.87 16. 5

63 4.84 2.6 28.45 15.5

TABLE 7. CORROSION RESISTANCE OF SYNTHETIC GEMS AND SPINEL IN OXYGENATEDURANYL SULFATE SOLUTIONS CONTAINING 40 g OF URANIUM PER LITER

AND SIMULATED FISSION-PRODUCT ADDITIONS AT 250°C

EXPOSUREWEIGHT LOSSES AND CALCULATED CORROSION RATES

TIME Titania Sapphi re MgO Spinel

(days) / 2mg/ cm mpy

/ 2mg/ cm mpy

/ 2mg/ cm mpy

7

14

21

28

35

42

49

56

-0.09

0. 12

0.12

0.49

0.51

0.51

0.47

0.53

0.4

0.3

0.2

0.6

0.5

0.4

0.3

0.3

-0.32

0.31

0.55

0.69

0.82

0.93

0.96

1.04

1.6

0.8

0.9

0.8

0.8

0.8

0.7

0.6

-1.02

1.73

2.58

3.25

3.67

4.42

5.05

5.45

5.0

4.3

4.2

4.0

3.6

3.6

3.5

3.4

period in fission-product-free solutions containing 40 g of uranium per

1 iter.

The synthetic fused titania offeredthe best corrosion resistance in allmedia tested. The synthetic sapphireexhibited the second best corrosionbehavior, whereas the magnesium oxide

spinel was the poorest. A summary ofthe data is presented in Table 8.

Corrosion and Solution Stability

Studies in Uranyl Sulfate at 125°C.

An investigation was completed todetermine the corrosion resistance oftype 347 stainless steel as a functionof uranium concentration in uranyl

71


TABLE 8. SUMMARY OF CORROSION DATA FOR SYNTHETIC GEMS AND SPINEL INURANYL SULFATE SOLUTIONS PRESSURIZED WITH 150 psia OF OXYGEN AT 250°C

MATERIALURANIUM CONCENTRATION

(g/liter )FISSION

PRODUCTS

TIME

(days)

CORROSION RATE

(mpy)

Ti tania 40 Absen t 63 0.3

Sapphi re 40 Absen t 63 0.7

MgO Spinel 40 Absen t 63 3.6

Ti tania 40 Present 56 0.3

Sapphi re 40 Pres en t 56 0.6

MgO Spinel 40 Pres ent 56 3.4

Ti tania 300 Absen t 49 0.2

Sa pphi re 300 Absent . 63 2.6

MgO Spinel 300 Absent 63 15.5

sulfate solutions at 125°C. Stabilityof the uranyl sulfate solutions was ofinterest because the tests were operatedin the absence of the excess dissolvedoxygen required for assuring solutionstability at elevated temperatures;the stability was studied by makingtotal uranium analyses. The availableoxygen per test, 27 mg, was the amountdissolved in the test solution at roomtemperature and the amount of oxygenpresent in the air space of the testautoclave. Based upon the solubilityof oxygen in water, the dissolvedoxygen content at 125°C was estimatedto be somewhat less than 25 ppm.

The corrosion tests were run intype 347 stainless steel autoclaves of225-ml capacity which had been usedpreviously for corrosion testing at250°C. Prior to use, the autoclaveswere cleaned with HN03-H202 solutionand then pickled for 20 to 30 min at60°C in 10 vol %HN03, 4 vol %HF. Thetest wafers were machined from type347 stainless steel bar stock andtested in a machined condition.

The uranyl sulfate test solutionscontained uranium concentrations of40.9, 101.4, 193.1, and 300.5 g/liter,

72

'*l«WIWf«W saam «.»it

respectively. The tests were stoppedat weekly intervals for examining andweighing the specimens. The solutionswere analyzed for final total uraniumand dissolved nickel contents. Thecumulative as-removed weight lossesand final e 1ectro1ytica 11y defilmedweight losses are reported in Table 9as a function of time and uraniumconcentration.

The physical appearance of thespecimens differed markedly as theuranium concentration was increasedfrom 40 to 300 g/liter. In all cases,the corrosion films appeared to bethin but the colors ranged from theoriginal lustrous, metallic gray inthe solution containing 40 g of uraniumper liter to a very dark, brown-coloredfilm in the solution containing 300 gof uranium per liter. The final totaluranium and dissolved nickel contentsin the test solutions are shown inTable 10.

The final uranium analyses showed areduction of 6 to 9 mole % in totaluranium content for all solutions.Whether these reductions were due tothe hydrolysis of the uranyl ion toform a hydrated UO precipitate or due


TABLE 9. CORROSION OF TYPE 347 STAINLESS STEEL IN UNOXYGENATED

URANYL SULFATE SOLUTIONS AT 125°C

URANIUM CONCENTRATION (g/liter)

EXPOSURE 40 . 9 101 .4 193.1 300 . 5

TIME

(days)WEIGHT LOSSES AND CALCULATED CORROSION RATES

, 2mg/ cm mpy

, 2mg/cm mpy

, 2mg/cm mpy mg/cm mpy

7 -0.08 0.19 -0.01 0.02 -0.12 0.30 -0.13 0.34

14 0.14 0.20 0.16 0.21 0.15 0.19 0.22 0.29

21 0.17 0.14 0.16 0.14 0.26 0.22 0.40 0.34

28 0.16 0.10 0.15 0.09 0.23 0.15 0.50 0.32

35 0.15 0.08 0.18 0.09 0.19 0.09 0.53 0.27

42 0.17 0.07 0.21 0.09 0.23 0.10 0.63 0.27

49 0.19 0.07 0.27 0.10 0.31 0.11 0.75 0.28

Defilmed 0.16 0.06 0.28 0.10 0.42 0.16 0.75 0.28

TABLE 10. FINAL TOTAL URANIUM AND

DISSOLVED NICKEL CONTENTS IN URANYL

SULFATE TEST SOLUTIONS AT 125°C

URANIUM CONCENTRATION DISSOLVED

(g/li ter ) NICKEL

Initial Final (mg/ml)

40. 9 36. 9 10

101.4 93.1 17

193.1 182.8 54

300. 5 280.5 79

to an actual reduction of the uranylion to U02 and/or U3Og is not known.There were no visual indications ofdeposited UO *jcH O during operation ofthe tests, and the final solution pHmeasurements were steady at values of0.2 to 0.3 unit higher than the initialpH values.

Corrosion attack on the type 347stainless steel specimens increasedwith an increase in total uranium con

centration in the solutions. The

magnitude of attack as a function ofthe uranium concentration is shown in

Fig. 38. At a concentration of 40.9 g

0.3

UNCLASSIFIED

DWG. 19566

50 100 150 200 250

URANIUM CONCENTRATION (g/l)

300

Fig. 38. corrosion Behavior ofType 347 Stainless Steel as Affectedby Uranium Solutions After 49 Days at

125°C

of uranium per liter, the corrosionrate was 0.06 mpy after 49 days; therate increased to 0.28 mpy as theuranium content was increased to 300.5g/liter. Microscopic examination ofthe defilmed specimens disclosed corrosion attack to be generalized andof a uniform nature; no evidences ofpitting attack were observed.

The final dissolved nickel values

were used to calculate corrosion rates

over the surfaces of the type 347stainless steel autoclaves wetted by

73


the uranyl sulfate solutions at 125°C.Including the surface area of thespecimen in each autoclave, the totalimmersed area was approximately 245

cm . A comparison between the de-filmed corrosion rates on the specimens and the corrosion rates calculated

from the dissolved nickel contents

appears in Table 11. This comparisonactually distinguishes between thecorrosion behavior of newly machinedstainless steel specimen surfaces and

the pickled stainless steel surfacesof the autoclave walls. The agreementin values obtained by the two methodswas quite good.

TABLE 11. COMPARISON OF CORROSION RATES

OBTAINED FROM DEFILMED WEIGHT LOSSES

AND FROM DISSOLVED NICKEL CONTENTS

IN URANYL SULFATE SOLUTIONS

AT 125°C AFTER 49 DAYS

INITIALCORROSION RATE (mpy )

URANIUM

CONCENTRATION

(g/liter)

WeightLoss

Method

D issolved

Nickel

Me thod

40.9 0.07 0.03

101.4 0.10 0.05

193.1 0.16 0.17

300. 5 0.28 0.24

Stress and Crevice Corrosion studies

with Type 347 Stainless Steel. Con

siderable effort was expended on aseries of investigations to determinewhether type 347 stainless steel was

subject to crevice and stress corrosion attack under a wide range ofexposure conditions. The susceptibilityto these two types of corrosion attack

was examined as a function of temperature, uranium concentration, type ofsolution aeration or deaeration, andmagnitude of applied stress.

The results of studies made with

the following solutions and conditionsare reported:

1. uranyl sulfate solutions containing5, 40, 100, and 300 g of uranium

74

per liter at 95°C; natural de-

aeration employed;2. uranyl sulfate solutions containing

200 g of uranium per liter at 95°C;helium and oxygen aeration employed;

3. uranyl sulfate solutions containing300 gof uranium per liter at 95°C;natural deaeration, helium aer

ation, and oxygen aeration employed;

4. uranyl sulfate solutions containing300gof uranium per liter at 95°C;pressurized with 1000 psi oxygen;

5. uranyl sulfate solutions containing40gof uranium per liter at 250°C;pressurized with 150 psia oxygen.

Studies wi th Uranyl Sulfate SolutionsContaining 5, hO, 100, and 300 g ofUranium per Liter. The stress specimens used in this investigation were

of a type recommended by F. LaQue ofthe International Nickel Co. for

qualitative examination of the sus

ceptibility of a material to stresscorrosion attack. The specimens, 8 cmin length, were cut from 0.95-cm-dia

type 347 stainless steel welded tubing.One end of each tube was pinched together tightly in a vise; the oppositeend was then bent at an angle of 90

deg to the central axis of the tube.Thus, a graduation in stresses up toand exceeding the elastic limit of the

stainless steel was incorporated intothe specimens. The validity of theLaOue-style specimen for determiningthe susceptibility of type 347 stainless steel to stress corrosion attack

was established initially by exposingspecimens in 42% MgCl solution at152°C. Within 2 hr, severe crackingoccurred at the highly stressed areas.

The stress specimens, along withunpinched and unbent control specimens,were exposed in uranyl sulfate solutionscontaining 5, 40, 100, and 300 g ofuranium per liter, respectively. Thesolutions were essentially oxygen-freeat 95 C, and the test was operated fora period of 79 days. Specimens wereexamined and weighed at weekly intervals. At completion of the tests, the

specimens were e1ectro1ytica 1 1 ytreated for removal of the corrosion

films. The final defilmed corrosion

data are presented in Table 12.There were no marked differences

between the corrosion rates for the

control and stress specimens in theirrespective environments. In fact, thecorrosion behavior of type 347 stainless steel, stressed or otherwise, waslittle affected by an increase inuranium concentration from 5 to 300

g/liter. The specimens were carefullycut in half, longitudinally, and theinterior and the mating surfaces wereexamined microscopically at 100X forindications of stress and/or crevicecorrosion attack. No signs of suchattack were disclosed by the examination. The surfaces were light grayin color, with uniform distribution ofcorrosion damage.

Studies with Uranyl Sulfate Solutions Containing 200 g of Uraniumper Liter. An examination was made ofthe corrosion resistance of type 347stainless steel stress specimens at95°C in uranyl sulfate solutions containing 200 g of uranium per liter.The specimens were of the LaQue styledescribed in the previous section, andthey were prepared from 0.6-cm-diawelded tubing. One siightmodification


was made in the specimen geometry,however. The 90-deg bend was made onthe end which was pinched together sothat the opposite end of the specimenwas entirely unstressed.

Two types of aeration were employedfor the tests. One solution was aer

ated by the introduction of oxygen ata rapid rate; the second solution wasmade oxygen-free by the passage throughit of helium gas. The tests were runfor a period of 95 days, with frequentexamination and weighing of the specimens. At the completion of the run,the specimens were defilmed electro-lytically for measurement of actualweight losses.

The results showed that the type ofaeration used in the solutions had

very little effect on the corrosionbehavior of the type 347 stainlesssteel specimens. The defilmed weightloss on the he 1ium-aerated specimenwas 0.12 mg/cm as compared with a0.13-mg/cm2 loss on the oxygen-aeratedspecimen. These losses were equivalentto a corrosion rate of 0.03 mpy forthe 95-day period.

The specimens were cut in half,longitudinally, and the interior walland the mating surfaces were examinedmicroscopically. The stressed metalareas were completely free of any signs

TABLE 12. DEFILMED CORROSION DATA ON TYPE 347 STAINLESS STEEL

STRESS AND CONTROL SPECIMENS AFTER 79 DAYS IN

DEAERATED URANYL SULFATE SOLUTIONS AT 95°C

INITIAL URANIUM CONCENTRATION

(g/liter)

4. 9

39.8

101.0

300.0

SPECIMEN

TYPE

Stress

Control

Stress

Control

Stress

Control

Stress

Control

WEIGHT LOSS

(mg/cm )

0. 09

0.06

0.07

0.07

0.12

0.09

0. 15

0.23

CORROSION RATE

(mpy)

0.02

0.01

0.02

0.02

0.03

0.02

0.03

0.05

75


of localized corrosion attack or

cracking. The pinched ends were spreadapart, and it was found that theseareas of direct metal-to-metal contact

were equally free from indications ofcrevice corrosion attack.

Studies with Uranyl Sulfate Solutions Containing 300 g of Uraniumper Liter. Stress and crevice corrosion studies were continued with

type 347 stainless steel in uranylsulfate solutions containing 300 g ofuranium per liter at 95°C. Threetypes of solution aeration and deaeration were used for these tests:

natural deaeration, helium aeration,and oxygen aeration. The test specimens were of two types: a modifiedLaQue-style and a type that is referredto as the Heinkel loop test specimen. (6 )

The modified LaQue-style specimensconsisted of various lengths of0.6-cm-dia type 347 stainless steelwelded tubing; one end of each specimenwas pinched together tightly. No bendswere formed in these tubes. The

Heinkel loop test specimens were

F. A. Champion, Corrosion Testing Procedures,p. 146-147, Wiley, New York, 1952.

prepared by forming 0.2-cm-thick by1.9-cm-wide by 12.7-cm-long type 347stainless steel strips into a U shape.The ends of the U were then bolted

together with type 347 stainless steelstuds and nuts to subject the specimento an elastic stress. Unstressed

control specimens for each materialwere tested simultaneously with thestress specimens.

The tests were operated for 76 to82 days, and then the specimens weredefilmed electrolytically and examined

microscopically for signs of stressand crevice corrosion attack. The

final defilmed corrosion data are

summarized in Tables 13 and 14 for the

two types of specimens.

The stressed-tube type of specimenswere bisected longitudinally and examined for indications of accelerated

corrosion attack. As observed in

previous tests, the interior surfaceswere semidull and whitish-gray incolor; there were no signs of localizedcorrosion attack in the stressed

zones, and no contact corrosion attack

was visible on the mating surfaces of

TABLE 13. DEFILMED CORROSION DATA FOR MODIFIED LaQUE-STYLE, TYPE 347

STAINLESS STEEL, STRESS SPECIMENS AFTER 82 DAYS AT 95°C IN URANYL

SULFATE SOLUTIONS CONTAINING 300 g OF URANIUM PER LITER

AERATION SPECIMEN SPECIMEN LENGTH WEIGHT LOSS CORROSION RATE

TYPE TYPE (cm ) (mg/cm ) (mpy)

None Stress 4.0 0.17 0.04



None Control 5.0 0.16 0.04

Helium Stress 4.0 0.16 0.04

Helium Stres s 6.5 0.16 0.04

Helium Stress 9.0 0.11 0.03

Helium Control 5.0 0. 14 0.03

Oxygen Stress 4.0 0.17 0.04



Oxygen Control 5.0 0.14 0.03

76

the pinched ends. Similarly, an examination of stress and contact areas on

the Heinkel loop specimens disclosedno evidences of localized corrosion

a t tack.

Studies with Uranyl Sulfate Solutions Containing 300 g of Uraniumper Liter. LaQue- and Heinkel-typestress specimens of type 347 stainlesssteel were exposed at 95°C in uranylsulfate solution containing 300 g ofuranium per liter, and an oxygenpartial pressure of 1000 psig wasmaintained. Unstressed control

specimens were tested along with eachtype of stress specimen.

The specimens and solution werecontained in a type 347 stainless steel


autoclave which was pressurized with1000 psig of oxygen at room temperature. The test was operated for 76days, and the specimens were periodically inspected. The defilmed corrosion data are presented in Table 15.

The defilmed corrosion rates were

not indicative of accelerated cor

rosion attack due to either stress or

crevice corrosion. The corrosion rates

for the stressed specimens were ofsimilar magnitude to the values obtainedin previous tests of exposure in uranylsulfate solution containing 300 g ofuranium per liter at 95°C.

Microscopic examination was made ofthe interior surfaces of the bisected

tubing; again, corrosion attack was of

TABLE 14. DEFILMED CORROSION DATA FOR HEINKEL LOOP, TYPE 347 STAINLESS STEEL,

STRESS SPECIMENS AFTER 76 DAYS AT 95°C IN URANYL SULFATESOLUTIONS CONTAINING 300 g OF URANIUM PER LITER

AERATION SPECIMEN WEIGHT LOSS CORROSION RATE

TYPE TYPE (mg/cm ) (mpy )

None Stress 0.23 0.05

None Control 0.25 0.06

Helium Stress 0.17 0.04

Helium Control 0. 18 0.04

Oxygen Stress 0.24 0.06

Oxygen Control 0.22 0.05

TABLE 15. DEFILMED CORROSION DATA FOR LaQJUE- AND HEINKEL-TYPE STRESS SPECIMENS

OF TYPE 347 STAINLESS STEEL EXPOSED IN URANYL SULFATE SOLUTION CONTAINING

300 g OF URANIUM PER LITER AND PRESSURIZED WITH 1000 psig OF OXYGEN AT 95°C

SPECIMEN

•"• • • — —

EXPOSURE TIME WEIGHT LOSS CORROSION RATE

TYPE (days) (mg/cm ) (mpy )

LaQue 76 0.13 0.03

LaQue 76 0. 17 0.04

LaQue 76 0.18 0.04

Control 76 0. 15 0.04

Heinkel 76 0.27 0.06

Control 76 0.62 0.15

77


a generalized nature and there wascomplete absence of pitting or crackingin the regions of stress and metal-to-metal contact. The Heinkel-loopspecimens exhibited the usual, slightlyroughened surfaces that are representative of generalized corrosion.The contact faces were stained some

what at junctions with the nuts andstuds used to draw the ends of the U

tubes together, but, otherwise, theywere unchanged.

Studies with Uranyl Sulfate Solutions Containing 40 g of Uranium perLiter. The last series of tests fordetermining the susceptibility of type347 stainless steel to crevice and

stress corrosion was concerned with

exposure at 250°C in oxygenated uranylsulfate solutions containing 40 g ofuranium per liter. In these tests, astress was applied by three-pointloading of beam-type specimen assemblies, and the stress was measuredwith an SR-4 strain gage. Briefly,the test assembly consisted of a0.3-cm-thick specimen centrally mountedon a fulcrum and base plate. A loadwas applied to the specimen by tightening the nuts on the studs positionedin each end of the assembly. Theapplied load, in turn, was measuredwith a strain gage cemented to thecenter of the test specimen. The

values of stress used in this investi

gation represent values for the maximumouter fiber stress on the specimensurface directly above the fulcrum.The applied stresses included in thestudy were 7,000, 15,000, 30,000, and75,000 psi.

The test specimens, 0.3 cm thick by1.6 cm wide by 8.8 cm long, wereannealed in air at 1010°C for 1 hr,air-cooled, and lightly pickled inwarm 10% HN03-4% HF solution for 10 minto remove annealing scale. Unstressedcontrol specimens were treated in asimilar fashion. The tests were run

in type 347 stainless steel autoclaveswith an oxygen partial pressure of 85to 90 psi at 250°C. The autoclaveswere pressurized with 150 psig ofoxygen at room temperature to producethe partial pressure at 250°C. Thedissolved oxygen content at 250°C wasestimated to be 500 ppm.

The tests were run for 48 to 63

days at 250°C,and the specimens wereexamined frequently. During the tests,the specimens developed heavy, dark-colored corrosion films which obscured

corrosion damage effects on the underlying metal surfaces. At the completion of the tests, the specimenswere defilmed electrolytically. Theresults are shown in Table 16.

TABLE 16. DEFILMED CORROSION DATA FOR BEAM-TYPE STRESS SPECIMENS OF

TYPE 347 STAINLESS STEEL EXPOSED IN OXYGENATED URANYL SULFATE

CONTAINING 40 g OF URANIUM PER LITER AT 250°C

APPLIED MAXIMUM STRESS EXPOSURE TIME WEIGHT LOSS CORROSION RATE

(psi ) (days ) (mg/cm ) (mpy )

7,000 48 0.69 0.26

15,000 48 0.69 0.26

30,000 48 0.65 0.24

None 48 0.63 0.24

75,000 63 0.71 0.21

75,000 63 0.77 0.22

None 63 0.69 0.20

78

."•*~-•^ffmn^^^^mm^#^m^s^>^


The corrosion rates for the stressed specimens. A total exposure time ofspecimens were completely independent 86 days was accumulated in the tests,of the magnitude of the applied stress The stress specimens were fabricatedand were in excellent agreement with from commercial-grade titanium metalthe rate values for the unstressed designated as RC-70. This materialcontrol specimens. No signs of ac- was supplied byRem-Cru Titanium, Inc.,celerated corrosion attack resulting in the form of annealed, 0.32-cm-thickfrom stress were found by microscopic sheet. Other components of the stressexamination. Definite indications of assembly, such as the fulcrum, studs,crevice corrosion were observed, how- and nuts, were machined from similarever, on two of the specimen as- material. The base plates were pre-semblies. In both cases, the localized pared from 0.64-cm-thick, annealed,corrosion took place on the specimen sheet TI-75A, supplied by Titaniumarea contacted by the stainless steel Metals Corp.nut used to apply the desired load. The applied stress on the specimensThe maximum pit depth, measured micro- was measured in terms of microinchesscopically, ranged from 3 to 5 mils in of strain with an SR-4, type A-7,these areas. strain gage and a Baldwin SR-4 strain

In summarizing the results of the indicator. The stud-nut combinationcomplete investigation to determine mounted in each end of the stresswhether type 347 stainless steel was assembly was tightened to apply a loadsubject to stress and crevice cor- to the test specimen centrally supportedrosion attack under a wide range of on the fulcrum. The resulting strainexposure conditions, it may be stated was indicated by the gage cementedthat, under the specific test con- directly above the fulcrum. Theditions included in the investigation, 50,000-psi stress obtained was thethere was no evidence of corrosion maximum outer fiber stress on theattack which could be attributed specimen at the point directly abovedirectly to the effects of stress at the fulcrum.temperatures of 95 to 250°C. Also, no At the completion of the 86-dayindications of crevice or contact tests, the stress assemblies were dis-corrosion attack were observed at mantled, and the specimens were ex-temperatures of 95°C, but there were amined and defilmed electrolytically.positive signs of such type of attack The final corrosion data for thein the 250 C tests. control and stress specimens appear in

Stress and Crevice Corrosion Studies Table 17.

with Commerical-Grade Titanium. A Microscopic examination of thestudy was made to determine whether stressed surfaces at frequent intervalscommercial-grade titanium metal was during the tests disclosed no signs ofsubject to stress and/or crevice stress corrosion attack. The finalcorrosion attack in oxygen-free, appearance of the stress specimen wasquiescent uranyl sulfate solutions the same as that of the control speci-containing 5, 40, 100, and 300 g of men; all surfaces were covered withuranium per liter at 95°C. Stresses very thin, adherent, golden-brown-of 50,000 psi were obtained by three- colored films. After the assembliespoint bending of beam-type specimens were dismantled, the areas contactedmounted on rigid base plates. Un- by the fulcrums and nuts were inspectedstressed control specimens were tested closely for indications of contactsimultaneously with the stressed corrosion attack; none were found.

79


TABLE 17. CORROSION OF TITANIUM STRESS AND CONTROL SPECIMENS AFTER86 DAYSAT95°C IN OXYGEN-FREE URANYL SULFATE SOLUTIONS

URANIUM CONCENTRATION

(g/liter)

SPECIMEN

TYPE

WEIGHT LOSS

(mg/cm )

CORROSION RATE

(mpy )

4.9 Stress

Con trol

0.03

0.03

0.01

0.01

39.8 Stress

Control

0.04

0.03

0.02

0.01

101.0 Stress

Con trol

0.05

0.02

0.02

0.01

3 00.0 Stress

Control

0.04

0.05

0.02

0.02

METALLURGY

E. C. Miller, Group LeaderW. J. Leonard T. W. Fulton

W. J. Fretague W. O. HarmsMetallurgy Division

A. R. Olsen, Reactor Experimental Engineering DivisionF. J. Lambert, Y-12

J. C. Wilson R. G. Berggren

Solid State Division

SERVICE AND MISCELLANEOUS WORK

E. C. Miller T. W. Fulton

W. O. Harms

The service work, described inprevious reports, for the Corrosion,Design, Instrument, and EngineeringGroups has continued.

Several groups of corrosion testpins have been prepared for theDynamic Corrosion Group for looptesting. The principal objective ofthese tests is to evaluate the metal

lurgical variables, because they mayinfluence attack under dynamic corrosion test conditions.

Previous work indicated that a

sensitized type 304 stainless steelspecimen holder (which was erroneouslyintroduced into the system in a

80

sensitized condition) underwent

serious intergranu1ar disintegrationin contact with stagnant uranylsulfate solution trapped underneatha type 347 stainless steel band.Since the portion of the sensitizedmaterial which was freely exposed toflowing solution did not appear to beattacked, several specimens of type 304stainless steel have been prepared inthe fully annealed, sensitized,machined, and e1ectropolished conditions for controlled dynamic corrosion tests.

Another group of type 347 stainlesssteel specimens containing variousamounts o f martensite has been preparedto determine the effectof the presenceof martensite on corrosion. Toproduce the martensite, type 347


austenitic stainless steel was cold behavior under dynamic corrosion testworked (bent 90 deg and then straight- conditions of weld metal containingened) at various temperatures - liquid up to 10% ferrite. In the fabricationnitrogen temperature, dry ice temper- of heat exchangers and other equipmentature, room temperature, boiling water from austenitic stainless steels, atemperature, and approximately 400°F. question regarding the amount of cold

Preliminary static corrosion tests work permissible in the fabrication iswere made on type 347 stainless steel raised. It is known that stressspecimens in water that had been corrosion cracking sometimes occursdegassed and to which hydrogen and *n tube sheets, as well as in tubes;oxygen had been added. The type 347 the cracking cannot be reproducedstainless steel specimens were tested consistently in laboratory tests norin the as-machined, the fully annealed, c an tne causes for it be preciselyand the commercially descaled con- defined.ditions. Both sand blasting and a Specimens are being prepared forcommercial salt bath treatment were evaluating and comparing the effects ofused to prepare the descaled specimens. cold working at va'rious temperaturesThe tests indicated that very sub- (from liquid nitrogen temperature tostantial differences in appearance approximately 400°F), particularlyand in measured corrosion attack with respect to the resultant variationsresulted from the differences in the ^n martensite content, the partialinitial conditions of the stainless stress relief and full annealing ofsteel surfaces. This work is being these cold-worked materials, and theextended to include most of the superimposi tion of elastic stresses onsurface conditions which would be tne residual stresses which resultencountered normally or might reason- from cold work. Specimens of fullyably be achieved in the commerical annealed stainless steel tubing arepreparation of austenitic stainless crimped, twisted, and bent at thesteels. Finishing operations can desired temperatures and are subse-produce surfaces which differ ap- quently stress relieved and annealedpreciably from the interior of the or left in the cold-worked condition,metal in chemical composition, degree as desired. The specimens will beof cold work, and the presence and exposed to the environments which willdisposition of the various microscopic probably be encountered in homogeneousphases. The surface variations may reactors and environments which areaccount for variations in corrosion known to be capable of causing stressbehavior which are sometimes left corrosion cracking.unexplained or are attributed to theenvironment. Specimens prepared by RADIATION DAMAGE STUDIESvarious methods are being tested under n r u a d <~>i

j , , . , • • n. u. Berggren A. H. Olsenstatic and under dynamic conditions. nt g pp uiA particular effort will be made totest specimens that have been sue- Impact specimens of type 304cesively machined, fully annealed, stainless steel, type 347 stainlessdescaled, and then elec tropol ished so steel, commercial-purity titaniumthat they have fully austenitic (Ti-75A), and iodide titanium havesurfaces, are free from cold work, been irradiated in the ORNL graphiteand have the same chemical composition reactor, and impact tests will beas that of the interior of the speci- completed shortly. An impact testermen. has been installed in the hot cell

Another group of specimens is being designated for this work, and theprepared to determine the corrosion test work is now in progress.

81


WELDING OF TYPE 347 STAINLESS STEEL

W. J. Leonard

Welds of 1-in. type 347 stainlesssteel plate to type 347 stainlesssteel plate and to a 30% carbon steelplate were made with type 347, lime-coated, stainless steel filler rod andtype 307 (Mn), lime-coated, stainlesssteel filler rod over type 347 stainless steel, bare-rod, heliarc rootpasses. Other welds were made withthe exclusive use of the two 1ime-

coated rods for the root and the

filler passes in the welding of carbonsteel plate to stainless steel plate.The plates were cut into standard testspecimens for tensile and bend tests.

The tensile strengths of the weldsmade of type 347 stainless steelexceeded the yield strengths andtensile strengths of the base metals.The per cent elongation in the weldmetal was usually lower than that(35%) of the type 347 stainless steelbaseplate because of the work-hardeningby peening during the welding process.Specimens usually broke in the basemetal adjacent to the weld metal. Allthese welds were considered as satisfactory. The tensile strengths ofwelds made of carbon steel to stainless

steel were lower than a 11-stainless

steel welds, since the stainless steelbase plate and weld metal were bothstronger than the carbon steel baseplate. The specimens usually brokeoutside the heat-affected zone in thecarbon steel plate, about 1/8 to1/4 in. from the weld-to-carbon steelinterface. The average yield strengthwas 42,000 psi and the tensile strengthwas 78,000 psi. The welds were allconsidered to be satisfactory.

Plates of type 347 stainless steel,2 in. thick, were welded with the useof a type 347 stainless steel bare rodand a type 347, lime-coated, stainlesssteel electrode. The pertinentfeatures of the welding design are a

20-deg included angle, a 1/16 _q" nnn

82

in. root face, and a 3/8-in. rootradius with a U groove. The firstfour passes of the heliarc processwere made with 3/32-in. bare, fillerrod. The remaining passes were madewith 1/8- and 5/32-in., lime-coatedelectrodes. Penetration through thebase of the root was normal. Radio

graphs and liquid-penetrant inspectionmethods showed a normal, sound weld.

Because this design restricted themovement of the welder using theheliarc torch, the second weld wasmade with a design that allowedgreater freedom of manipulation whilethe root passes were being made. AU groove, with an included angle of40 deg and a 1/4-in. radius, was usedin the second weld. The proceduresfor welding were identical to thosefor the first weld, except that3/16-in., coated rod was used for allthe filler passes. Good penetrationand a sound weldment were indicated byliquid-penetrant inspection andradiographs. The volume of weld metalwas larger than that used in the firstweld, but the design gave ample freedomof manipulation for the welder.

This second weld was sectioned intospecimens for side bend tests and formacroscopic, microscopic, and soundnesstests. A standard, guided-bend-testjig was used to bend specimens toangles of between 140 and 165 degrees.All the bend-test specimens were sound.Macroscopic and microscopic examinationalso indicated sound, normal welds.A soundness test, which consisted ofan etch in a saturated ferric chloridesolution to show very small fissuresin the weld metal which might haveescaped detection by the other methods,indicated normal, sound, weld metal.

Specimens of type 347 stainlesssteel weld metal, laid under conditionsof severe enough restraint to causecracking in the weld metal, werereceived from anotherAEC installation.The cracks in the weld metal are beingstudied microscopically and by x-raydiffraction methods to determine the

role of the chemical concentration ofcertain of the elements in type 347stainless steel in the promotion ofcracks and to determine the preferredlocation and mechanism of cracking.

BRITTLE FRACTURE OF TITANIUM

A. R. Olsen

A literature survey has been startedto cover the wide variety of laboratory-scale specimens, procedures, andcriteria for an engineering evaluationof ductile-to-brittle transition oftitanium. It is hoped that carefulscreening of the tests will lead to anexperimental program that will permitcomparison of the present test resultswith data more easily converted intodesign specifications.

TITANIUM FOR VALVE TRIM

The actual tests of valves arereported by the groups doing the work;however, the Metallurgical Group hascontinued to assist in these programs.A study has been made of surfacehardening with nitrogen, and, morerecently, in cooperation with theStatic Corrosion Group, attempts havebeen made to reduce galling of non-hardened surfaces by pretreatmentin oxygen-saturated water at 250 C.

PROPERTIES OF TITANIUM

W. J. Fretague

Commercial Titanium. Twelvefinish-machined impact specimens ofcommercial titanium (Ti-75A, item 24,heat L782), which had been vacuumannealed at 500°C for 1 hr and furnacecooled prior to machining, werevacuum annealed at 600°C for 1 hr andfurnace cooled. Four of the specimenswill be used to establishthe transitiontemperature of annealed titanium.The eight remaining specimens, four ofwhich were abraded lightly with400-mesh emery paper, were submittedto the Dynamic Corrosion Group forloop testing in uranyl sulfate solution


prior to impact testing to study theeffect of cold work on the rate ofhydrogen absorption by Ti-75A. Amore detailed outline of the test waspresented in the previous report.

Since cold working is reportedto have a drastic effect on the rateof hydrogen absorption of titanium,corrosion and impact tests will bemade on a series of severely cold-worked samples. Titanium bar (Ti-75A,item 24, heat L782), 1 in. in diameter,was swaged to 0.2 19-in.-dia rod fromwhich 20 specimens have been machined.Four specimens are to be corrosiontested in uranyl sulfate in a loop.Four specimens will be cathodicallytreated in approximately 1 M H2S04(room temperature) at a currentdensity of approximately 0.2 amp persquare inch. Eight specimens will beimpact tested without further treatmentother than swaging and machining.Four specimens will be kept in reservefor future testing.

Two, 0.032-in.-thick by 1 7/8-in.-wide Ti-75A sheets were fabricated bypack' rolling(1) 1/4-in.-thick Ti-75Aplate at 850°C. The sheets were thensheared to suitable lengths for sheetfatigue specimens and surface groundon both sides to a final thickness of0.027 inch. The 18 specimen blanksprepared were vacuum annealed at600°C for 1 hr and furnace cooled.Four, sheet, fatigue specimens weremachined from annealed specimen blanksand fatigue tested at three differentstress levels. The results obtainedappear in Table 18 and are plotted inFig. 39.

Six of the remaining 14 specimenblanks (3 5/8 in. long by 1 7/8 in.wide by 0.027 in. thick) were heliarcwelded by placing two specimen blankstogether and welding around theperiphery of the assembly. One ofthese assemblies and two, single,

(1)W. J. Fretague, HRP Quar. Prog. Rep. Jan. 1.1953, ORNL-1478, p. 83-88.

^2'd. P. Smith, Hydrogen in Metals, p. 5, 21,24, and 26, Univ. of Chicago Press, Chicago, 1948.

83


TABLE 18. RESULTS OF FATIGUE TESTSON ANNEALED SPECIMENS OF Ti-75A

SPECIMENNO.

STRESS

(psi)NUMBER OF CYCLESBEFORE FAILURE

1 40,500 41 x 106*

2 45,000 16.6 x 104

3 50,000 16.4 x 104

4 50,000 8.0 x 104**

•Specimen did not break; test discontinued.

**By mistake, this specimen was subjected to apreload in the fatigue-testing machine and failedprematurely.

g 48

2 5

CYCLES

UNCLASSIFIEDDWG 19567

Fig. 39. Fatigue Properties ofTi-75A Hot-Rolled and Annealed Sheet.

specimen blanks were prepared forstatic exposure to uranyl sulfatesolution at approximately 250°C.Another welded assembly and two,single, specimen blanks will be givencathodic treatment in sulfuric acid.

A third set of specimens (onewelded assembly and two, single,specimen blanks) was heated at 750°Cfor 4 1/2 hr in a hydrogen atmosphere.This treatment was somewhat drastic,and the specimens warped badly whenmoved from the hot zone to the coldzone of the furnace. When an attemptwas made to straighten one of thewarped specimen blanks by hand, itbroke into several smaller pieces.

84

Obviously, the hydrogen absorbed hadseverely embrittled the material. Inan attempt to salvage them, the threeremaining pieces that were treated inhydrogen were vacuum annealed at 950°Cfor 4 hr and furnace cooled. On theassembly made up of two specimenblanks welded together, the hydrogendiffusing from the inner surfaces ofthe two specimen blanks built up apressure great enough to actuallydeform and separate the inner surfacesof the two welded pieces. In anattempt to salvage them, these vacuum-annealed pieces will be cold rolledvery slightly. If this is successful,they will be treated again in hydrogenat a lower temperature and for ashorter period of time.

Two sheet specimens machined from0. 027-in. -thick, vacuum-annealed,Ti-75A sheet will be irradiated andthen fatigue tested.

The hydrogen content of a number oftitanium specimens has been determinedby the vacuum-fusion method. Asummary of the samples analyzed andthe results reported was presented inthe previous report. Table 19 liststhe previous results plus the resultsreported during the current quarter.

It has been reported in the literature > that the room-temperaturesolubility of the t i tan ium-hy drogenphase is less than 0.14 at. %H (0.0029± 0.0003 wt %), whereas it has beenshown that,in titanium-hydrogen alloys,up to at least 4.5 at. % H is solublein alpha titanium at 200°C. Thisdrastic decrease in the solubility ofhydrogen with decreasing temperaturesuggests that the rapid cooling fromabove 200°C of titanium-hydrogen alloycontaining more t han 0. 14 at. %H ma yresult in a supersaturated solidsolution of hydrogen in titanium atroom temperature. Murray and Taylor(4)

(T)CM. Craighead, G. A. Lenning, and R. I

Jaffee, J. Metals 4, 1317 (December 1952).

G. T. Murray and W. E. Taylor, Effect ofNeutron Irradiation on Supersaturated Solid Solu-1953)° '" Copper' ORNL-1323 (Jan. 30

MATERIAL

Ti-75A

Iodide Ti


TABLE 19. HYDROGEN CONTENTS OF CORROSION IMPACT SPECIMENS

TREATED IN VARIOUS ENVIRONMENTS

SPECIMEN

NO.

1-2

2-1

4-1

8-1

3-1

5-1

6-1

7-1

9-3

10-9

11-6

12-3

13-1

14-8

15-9

16-9

19-10

18-6

24-6

26-1

28-8

I-1A1

I-2A1

I-3A

I-4A

I-5A

I-6A

I-7A

I-8A

TREATMENT PRIOR TO IMPACT TESTING

Exposed in U02S04 solution containing 300 g of U perliter at 250°C for 671 hr in loop A(a)

Exposed in UO2SO4 solution containing approximately300 g of U per liter at 250°C for 99.7 hr in loopA(«)

Exposed in UO.SO. solution containing 40 g of U perliter at 250°C for 10 days(fc)

Cathodically treated in approximately 1 MH,S04 at0.1 amp for 8 hr at 70°F

Cathodically treated as above for 168 hr

Vacuum annealed at 500 C for 1 hr and furnace cooled

prior to machining

Arc-melted, swaged, and vacuum annealed at 950 C for1 hr and furnace cooled prior to machining

Arc-melted, swaged, and vacuum annealed at 500 C for1 hr and furnace cooled prior to machining

HYDROGEN CONTENT

(it X)

0.016

0.014

0.010

0.010

0.013

0.013

0.013

0.007

0.010(c)0.012(e)0.011(c) 0.0]2(d>0.014<°

0.015

0.020

0.018

0.020

0.061

0.011

0.012

0.008

0.011

0.008

0.010

0.014

0.011

0.012

0.008

0.017

0.013

(a)See HRP Quar. Prog. Rep. Jan. 1, 1953, ORNL-1478, Table 10, p. 84, for detailed test history.

(fc)See HRP Quar. Prog. Rep. Jan. 1, 1953, ORNL-1478, Table 11, p. 85, for detailed test history.U)

(d)

Brown corrosion product present on surface of sample.

Sample abraded on surface prior to charging into vacuum-fusion apparatus.

85


suggest that the course of the decom- information regarding the introductionposition of the supersaturated solid of molecular hydrogen, but it will notsolution of beryllium in copper can necessarily provide information on thebe followed by in-pile measurement of atomic or ionic hydrogen resultingelectrical resistivity. The similarity from chemical reaction.with regard to the formation of super- Bureau of Mines Titanium. Threesaturated solid solutions of the two arc melts of Bureau of Mines, high-alloy systems, c o pp er - be ry 11 ium and purity, titanium sponge (lot 1056,titanium-hydrogen, suggests that in- R-83A) have been prepared, swaged,pile resistivity measurements be made vacuum annealed at 600°C for 1 hr, andon commercial titanium specimens machined into impact specimens. The(which normally contain from 0.3 to three 11 3/4-in.-long impact specimens0.5 at. % H(3)) irradiated at several obtained from melt B-l will be used totemperatures within the range of establish the transition temperaturetemperature planned for the homogeneous of Bureau of Mines, arc-melted, spongereactor components. Specimens (approx- titanium. Three 11 3/4-in.-longimately l/8-in.-dia rods) will be specimens from melt B-2 and threefurnished to the Solid State Division specimens of the same length from meltfor in-pile resistivity measurements. B-3 will be corrosion tested in uranyl

The problem of the rate of hydrogen SuHate in a looP Prior to iniPactdiffusion in commercial titanium has testlng- 0ne 9-in.-long specimen frombeen considered, and it was found melt B"2 and one 10-m. -long specimenthat there is some evidence in the from melt B"3 were retalned to compareliterature^5) to support the theory me 1l s B_2 and B'3 with melt B"1'that the oxide layer present on the whlch wlU be used to determine thesurface impedes the reaction of transltlon temperature of Bureau ofhydrogen with titanium until the Mlnes tltanlum-diffusion of these oxide layersinto the base metal permits the ZIRCONIUM-TITANIUM ALLOYShydrogen to enter the metal. Sincethe proposed operating temperature , Theuf°Ur " "onium-ti tan ium meltsof the homogeneous reactor (250°C) is desc"bed ln the previous report(')fairly low, rapid diffusion of the have fee,n arc-me0lted' swaged, vacuum

r j i i j i_ j annealed at 600 C for 1 hr. furnacesurface oxide layers would be expected. ' •LUilla»-cTo test this, a sample of commercial cooled. and machined into corrosiontitanium (Ti-75A) shee t, 1/4 in. thick, P^s (0.204 in. in diameter by 1.385was charged into the quartz furnace ln' long) and lmPact specimens. Thetube of the Seivert's apparatus, the pinS> f°Ur from each all°^ wiU besystem was evacuated, and the specimen ^"j"011 tested and then imPactwas heated to 250°C; hydrogen ( generated es e .by the decomposition of titanium , J' °\ Bett"ton, Jr. of the Metal-hydride) was then admitted to the lurgy DlVlslon- ln the course of hissystem. This specimen will be exposed lnvestlgation of the zirconium-indiumin a hydrogen atmosphere at 250°C for system- has tentatively determinedapproximately one month, and it will that the beta Phase in thl s sYstem ha^then be sectioned and examined metal- a face-centered-cubic crystal struc-lographically for evidences of hydrogen ture- Since many face -cen te red-cubi cpenetration. This test will give ^tals arenot subject to the transition

from the ductile to the brittle typeof fracture with decreasing tempera-

>R. W. Dayton et al. , Hydrogen Embr i11 lemen t tUTe' 3n attempt was made to produceof Zirconium, BMI-767, p. 27 (Aug. 22, 1952). an arc-melted z i r con i um- i n d ium alloy

(5)r

86

(approximately 28 at. % indium) and tofabricate the melt into specimens forcorrosion and impact tests. The melthas been prepared (with very littleloss in indium, considering the widedifference in the melting point of thetwo metals), but attempts to swage thealloy were not successful. A metal-lographic sample of the alloy is being


examined to determine whether thealloy is in the all-beta-phase field.

GENERAL

Impact testing of the specimenson hand will be resumed as soon asmodifications to facilitate theoperation of the tests in a hot cellare completed by the Research Shops.

AQUEOUS SOLUTION AND RADIATION CHEMISTRY

STATIC CORROSION TESTS WITH RADIATION

F. H. Sweeton W. C. Yee

Five tests in the LITR and threetests of out-of-pile blanks have beencompleted. In all cases, the bombs,which served as the corrosion specimen,were constructed of type 347 stainlesssteel. The method employed to determinethe corrosion rate, which consisted ofmeasurement of the consumption ofoxygen, was described previously.(x)Table 20 summarizes the conditions foreach of the experiments, and Table 21gives the corrosion results from theout-of-pile experiments, as calculatedfrom the nickel content of the solution

(1)C. H. Secoy et al., HRP Quar. Prog. Rep.Oct. 1, 1952, ORNL-1424, p. 94-96.

at the conclusion of the experiment,based on the assumption that thecorrosion proceeds at a uniform ratethroughout the test.

The corrosion results from the in-pile experiments are shown graphicallyin Figs. 40 through 43 in which thecumulative corrosion isplot ted againstthe square root of time. The corrosionappears to be linear with the squareroot of time for a period, and then itincreases rather sharply. Neither thecause of this break nor the factorsdetermining the time at which itoccurs can be definitely specified onthe basis of the data from the fewexperiments completed. Furthermore,attempts to correlate corrosion rateswith the variables involved are notyet justified.

TABLE 20. SUMMARY OF EXPERIMENTAL CONDITIONS FOR STATIC IN-PILE CORROSION TESTS

EXPERIMENT

NO.SOLUTION

TOTAL URANIUMCONCENTRATION

(g/liter)

FISSION

DENSITY

(kw/liter)

BOMBPRETREATMENT

++Cu

CONCENTRATION(moles/liter)

INITIAL 0,PRESSURE''(psi)

DURATION

OF TEST(days)

H-2 uo2so4

uo2so4

uo2so4

"0/2

U02F2

40 4.5 to 6.0 Cr03 0.01 900 91

H-3 43.9 0 None* 0.01 558 63

H-4 43.9 0 Cr03 0.01 453 63

H-5 40 4.5 to 6.0 None* 0.03 900 63

H-6 40 0 None* 0.02 552 6

H-7 uo2so4 40 6.0 None* 0.01 100 18

H-8 uo2so4 5 0.8 None* 0.01 80 18

H-9 uo2so4 40 0.5 None*

—

0.01 80 40

•Cleaned and etched with 20% HNOj-3% HF solution.

87


TABLE 21. CORROSION RESULTS IN OUT-OF-PILE TESTS

EXPERIMENT NO. SOLUTE

DURATION

OF TEST

(days)

CUMULATIVE

CORROSION

(mils)

CORROSION

RATE*

(mpy)

H-3

H-4

H-6

uo2so4

uo2so4

U02F2

63

63

6

0.021

0. 00 91

0.018

0.12

0.053

1.09

•Based on the assumption of a uniform rate of corrosion.

zo

U)

oor

ccoo

0.9

0.8

0 7

DWG I8765A

! > 1 I ! 1 1 1 1 1 1O U02F2 CONTAINING 40g OF U PER liter 1

PLUS 0.03^/ Cut+-, BOMB NOT PRETREA_ EXPERIMENT H-5 L

1 1

11rED

• U02S04 CONTAINING 40g OF U PER literPLUS 0.014/ Cu++i BOMB PRETREATEDWITH Cr03; EXPERIMENT H-2

0.6

0.5

0.4

0.3

FISSION DENSITY: 4.5 TO 6.0 kw/lit

EXCESS OXYGEN : ~ 900 psi

TEMPERATURE ; 250 TO 260°C

er

i

I•

|

1J

<

i

//

/

/

i •j*

o

\Si

oL

Vt (weeks)to

Fig. 40. Corrosion in an In-PileIrradiation Test of a Type 347 Stainless Steel Bomb Containing UranylSulfate Compared with Corrosion in anIdentical Bomb Containing UranylFluoride. Experiments H-2 and H-5.

88

0.6

0.5

r o.4HO

Oor

cc

0.2

0.1

DWG 18764A

EXPERIMENT H-2; Cr03-PRETREATED BOM BWITH 900 psi OF OXYGEN

EXPERIMENT H-7; UNPRETREATED BOMBWITH 100 psi OF OXYGEN

SOLUTION: U02S04 CONTAINING 40 g OFURANIUM PER LITER

ANIONS: SO,"

FISSION DENSITY: 4.5 TO 6.0 kw PER LITER-

Cu++ CONCENTRATION: 0.01 M

TEMPERATURE'- 250°C

V / (weeks)

Fig. 41. Effect of Excess Oxygen onCorrosion in In-Pile Irradiation Tests

of Type 347 Stainless Steel Bombs Containing Uranyl Sulfate. ExperimentsH-2 and H-7.

0.25

0.20

0.15

0.10

0.05

DW0 1878IA

'EXPERIMENT H-7; U02S04 CONTAINING 40 g OFU PER LITER; FISSION DENSITY, 6 kw PER LITER

•EXPERIMENT H-8; U02S04 CONTAINING 5 g OF UPER LITER; FISSION DENSITY, 0.8 kw PER LITER

"ANION: S04"

EXCESS OXYGEN: ~ 80 psi

Cu++ CONCENTRATION: 0.01 M

TEMPERATURE: 250°C

-BOMB PRETREATMENT: NONE

Y /(weeks)

Fig. 42. Effect of uranium concentration and Fission Density on corrosionin In-Pile irradiation Tests of Type347 Stainless Steel Bombs ContainingUranyl Sulfate. Experiments H-7 andH-8.

MECHANISM OF THE HYDROGEN-OXYGENRECOMBINATION REACTION IN SOLUTIONS

H. F. McDuffie L. F. Woo

Additional effort has been made tolearn what copper-containing speciesis [are] the active catalyst[s] responsible for the homogeneous recombination of hydrogen and oxygen inthe presence of solutions of cupriccompounds at high temperatures.

The significant developments fromthis work include:1. the use of cupric perchlorate and

perchloric acid to estimate therate constant for uncomplexed


o.io

0.05

Fig. 43. Effect of Fission Densityon corrosion in In-Pile irradiationTests of Type 347 Stainless Steel BombsContaining Uranyl Sulfate. Experiments H-7 and H-9.

cupric ion and to ascertain theabsence of complexing by perchlorate,

2. extension of the previously reported data for copper sulfateplus sulfuric acid to higher concentrations of acid,

3. determination of the effect ofuranyl sulfate concentrations ofup to 1 K upon the rate constant,in both the presence and theabsence of added sulfuric acid,

4. the use of rate constant data anddirect observation in the Marshall,semimicro, phase-study apparatusto determine the region of hydro-lytic precipitation of copper anduranium in dilute solutions ofsulfuric acid and uranyl sulfate,

5. evidence that in systems containingsulfate at low concentrations somespecies appears to exist which has

89


more activity, catalytically, thanuncomplexed cupric ion.

All experiments were carried out at250°C with solutions which were 0.0010M with respect to copper at roomtemperature. In this report, all concentrations are expressed as those atroom temperature, but all values ofkCu, calculated according to theformulas set forth in a previousreport,(2> have been corrected for theexpansion in volume and resultingchanges in concentration on going to250°C. It should be noted that allvalues of kCu are based on the solubility of hydrogen in water, as reported in BMI-T-25,<3> and that anychanges in these solubility valuesthat occur while changes are beingmade in the nature of the solutionstested will result in different valuesfor the rate constant. In anothersection of this report, efforts tomeasure these solubilities are discussed .

Activity of Cupric Perchlorate PlusPerchloric Acid. Perhaps the simplesthypothesis for explaining the recombination of hydrogen and oxygen in thepresence of dissolved cupric compoundswould involve a reaction of dissolved

(2),H. F. McDuffie, E. L. Compere

Stone, Chem. Quar. Prog. Rep. Mar0RNL-1285, p. 89-94.

(3 )H- A- Pray, C. E. Schweickert

Mmmck, The Solubility of Hydrogen. OxygenNitrogen, and Helium in Water at Elevated T,tures. BMI-T-25 (May 15, 1950).

and H. H.

31. 1952,

and B. H.

Iempera-

hydrogen with the cupric ion. Accordingly, it was very desirable todetermine the activity of copper underconditions which would tend tominimizeany complexing of the cupric ion byanions of the solution. The knownusefulness of the perchlorate ion forthis purpose at room temperatures fora variety of systems suggested thatcupric perchlorate be tested. At aconcentration of 0.0010 M, precipitation was found to occur at 250°C,but the presence of 0.005 M perchloricacid was found to be sufficient tomaintain phase stability at thistemperature. Recombination experiments were run with perchloric acidconcentrations of 0.005, 0.010, and0.050 M in the presence of 0.0010 Mcupric perchlorate, with the resultsgiven in Table 22.

The constancy of kCu at 5.0 suggeststhat this may be the value for theactivity of uncomplexed cupric ion.The data further indicate that nocomplexing by perchlorate takes placeand that there is no direct participation of hydrogen ions in the reaction(except for the suppression of thehydrolytic precipitation).

All experiments with perchloratehave been limited to short times athigh temperatures to avoid excessivecorrosion of apparatus by the tracesof chloride ion released in the decomposition of perchlorate.

Effect of Sulfuric Acid. Table 23and Fig. 44 present data showing the

TABLE 22. RESULTS OF RECOMBINATION EXPERIMENTS WITH PERCHLORIC ACIDIN THE PRESENCE OF CUPRIC PERCHLORATE

EXPERIMENT NO. EXCESS ACID *cu x 10"3 (moles" l hr" 1 liter' 1)

B-49

B-48

B-151

0.000

0.001

0. 005

3.5,

7.1,

5.0

phase instability

phase instability

B-152 0.010 5.0

B-153 0.050 5.0

90

stei****a«1MEi*t*i'+ni-*a*i^.i«


TABLE 23. HOMOGENEOUS RECOMBINATION OF HYDROGEN AND OXYGEN

IN THE PRESENCE OF ADDED SULFURIC ACID

EXPERIMENT k x 10"3(moles* hr" * liter"1)

H2S04 •J [h.soJ(moles/liter) NO. 2 4

0.00 B- 1 6.7

- 18 6.7

- 73 7.7

- 74 9.8

- 77 9.7

- 85 6.1

- 89 7.3

- 99 8.0

-102 6.0

0.00040 - 47 11.2 0.02

0.0010 - 33 20.6 0.032

-136 6.4

-140 12.7

-144 12.3

0.0020 - 34 19.6 0.045

-137 12.7

-141 12.98

-142 11.8

0.0030 - 35 19.2 0.055

-138 14.1

-139 12.2

-143 27.6

-145 11.0

-146 15.9

0.0050 -130 14.2 0.071

-135 14.2

-147 14.6

0.010 -131 12.6 0.10

-134 12.7

0.020 -109 10.5 0.141

-114 10.6

0.050 -132 7.2 0.223

-133 6.9

effect of adding sulfuric acid tosolutions of 0.0010 M copper sulfate.The complete lack of precision atconcentrations of acid below 0.005 M

was found to be paralleled by theappearance of precipitates when thesesolutions were heated to 250 C in

small glass tubes in the Marshall,

semimicro, phase-study apparatus. Itis considered that a region of hydrolytic precipitation exists, as discussed in a subsequent section of thischapter on "Exploratory Phase Investigations of Systems Involving CuS04,U02S04, H2S04, and H20 at UraniumConcentrations Below 30 g/liter." The

91


22

20

18

to

a>

= 14

.rz

_ 12t

to<u

"5 10 i

\ 8"o 6

4

2

0

DWG. 19207A

1o

1

^1 o

\p0

"^K

/ <i\

1 \ ^

L" REGIC3N 0 - PH ASE NSTA BILITY 8-_

0 0.04 0.08 0.12 0.16 0.20 0.24

CONCENTRATION OF SULFURIC ACID [(moles/liter)^]Fig. 44. Effect of Sulfuric Acid

Concentration on Hydrogen and OxygenRecombination in 10"3 M Cupric SulfateSolution at 250°C

precipitation of copper from solutionat these temperatures had not beenanticipated, since the solubility ofcopper sulfate in water from 100 to180°C, as reported by Seidel,(4) was43 to 45 wt %, with practically notemperature dependence in this region.The suggestion that the precipitationis hydrolytic is supported by theeffectiveness of excess acid in preventing its appearance, which isparalleled in this case by the observations on the hydrolytic precipitationof uranium from dilute uranyl solutions.The high precision obtained at concentrations of acid above 0.005 Msupports the opinion that the experimental technique is not responsiblefor the scatter observed at lower

concentrations. It appears that thereis a maximum in the activity at about0.003 M which is hidden by the hydrolytic precipitation. The possibilitythat the partially hydrolyzed copper

(4)A. Seidel, Solubilities of Inorganic and

Metal Organic Compounds, 3d ed., Vol. 1, p. 497,Von Nostrand, New York, 1940.

92

mmmmmmmmmmmm

precipitate may have heterogeneouscatalytic activity has been considered,but at present the emphasis is onlearning more about the homogeneousactivity.

Effect of Uranyl Sulfate. Table 24and Fig. 45 illustrate the effect ofadding uranyl sulfate of varyingconcentrations to solutions of 0.0010 Mcopper sulfate. As with the sulfuricacid discussed above, a region ofimprecision and phase instability wasnoted up to a uranyl sulfate concen

tration of around 10 g of uranium perliter. Also, the activity was loweredat increasing concentrations of uranylsulfate.

Effect of Uranyl Sulfate Plus

Sulfuric Acid. Table 25 and Fig. 46illustrate the effect ofuranyl sulfateconcentration on the activity of0.0010 M cupric sulfate plus 0.020 Msulfuric acid. It will be noted that

these data begin, in a sense, at aplace on the curve of Fig. 44 wherethe activity is already considerablybelow the maximum and where the

addition of more sulfuric acid further

depresses the activity. Similarly,the additionof uranyl sulfate depressesthe activity. It is also noted thatthe precision in the concentrationregion below 10 g of uranium per literis much better than that observed in

the absence of the sulfuric acid.Figure 47 shows the effect of thecombination of uranyl sulfate plus0.020 M sulfuric acid superimposed onthe effect of sulfuric acid alone.

Plans for Next Quarter. The

relatively low activity obtained inthe presence of perchlorate as theonly anion compared with that obtainedin the presence of small amounts ofsulfate suggests that catalytic speciesexists with higher activity than thatof uncomplexed cupric ion. The presenceof the region of hydrolytic precipitation interferes seriously with thedetermination of the maximum activityof this species and with the determination of exact conditions for achieving

»w«w*easi»*«B*s*»«


TABLE 24. HOMOGENEOUS RECOMBINATION OF HYDROGEN AND OXYGENIN THE PRESENCE OF ADDED URANYL SULFATE

URANYL SULFATE

CONCENTRATION

(g of U per liter)

0.25

1.0

2.25

2.62

4.0

EXPERIMENT

NO. (noli

B- ]- 18- 73- 74- 77- 85- 89- 99-102

- 78- 84- 92-100

- 72- 75- 79- 88- 93

- 80- 87- 94

- 62

- 66

- 81- 86- 96

6.25 - 82- 90- 97

9.0 - 83- 91- 98

10.0 - 71- 76

26.2 - 63- 67

83 - 68

142.5 - 69

205 - 70

262 - 64- 65-106-107-108-119-120

*Cu x 101

-3

liter* J)- 1hr

6.76.77.79.89.76.17.38.06.0

8.510.111.19.9

6.78.36.3

10.16.6

10.26.87.1

88

90

9106

4554

9

88

842

87.

14.

182

8.9

33

6.

6.8

9

4.6

3.3

3.3

2.73.03.24.32.93.83.0

N g of U per liter

0.5

1.0

1.5

1.62

2.0

2.5

3.0

3.16

5.1

8.2

11.9

14.3

16.2

93

TABLE 25. HOMOGENEOUS RECOMBINATION OF HYDROGEN AND OXYGEN IN THE PRESENCE

OF 0.020 M SULFURIC ACID AND ADDED URANYL SULFATE

URANYL SULFATE CONCENTRATION

of U per liter 'g ofUper liteg ot U pe

1.0

2.25

4.0

6.25

9.0

12.25

25.0

100

208

1.0

1.5

2.0

2.5

3.0

3.5

5.0

10.0

14.4

noles/liter

0.0042

0.0095

0.0168

0.026

0.038

0.051

0.105

0.420

0.873

EXPERIMENT

NO.

B-109

-114

-110

-113

-124

-128

-111

-115

-116

-118

-125

-129

-112

-117

-126

-127

-148

-149

-150

*Cu X 10 °

(moles hr liter )

10. 5

10. 6

7. 0

7. 2

8. 0*

6. 1

5. 5

6. 2

6 1

8 9 to 6.6**

7 5*

5.9

7 3

8 1

6 8

7 1

7 2

5 .6

3 .1

•Evidence of excess activity due to contamination of reaction vessel in previous runs.

"Poor experimental data.

TOTAL

SULFATE

(moles/1 iter)

0.021

0.0252

0.0305

0.0378

0.047

0.059

0.072

0.126

0.441

0.894

-J [Sulfate]

0.145

0.158

0.175

0.194

0.217

0.242

0.268

0.355

0.665

0.945

B33•a

oG>99Hw99r

•o99OC399m

99M•oO90H


8 9

CONCENTRATION OF URANIUM SULFATE |(g of Uper liter)"

Fig. 45. Effect of Uranyl Sulfate Concentration on Hydrogen and Oxygen Recombination in 10"3 M Cupric Sulfate Solution at 250°C.

12

10

-c 8 o— i° jL, A—.

o o

0 9 o ^=

DWG. 19209A

3 4 5 6 7 8 9 10 H 12 13 14 15

CONCENTRATION OF U02S04 [(got Uper liter l'^]

Fig. 46. Effect of Uranyl Sulfate Concentration on Hydrogen and Oxygen Re-»- 3combination in 10 M Cupric Sulfate Solution at 250°C.

95


10

DWG. I92I0A

^H2SOt

®

i\i \I \

°fso-,L V"

—•-_

KiS(î+Uo>So

\

"-•^% U4

^"-•^

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

TOTAL CONCENTRATION OF SULFATE jjmoles per liter/sj

Fig. 47. Effect of Sulfuric Acid andUranyl Sulfate on Hydrogen and OxygenRecombination in 10"3 M Cupric Sulfateat 250°C

maximum activity. Hence, it is difficult to formulate a network of com

plexing constants which will adequatelydescribe the various systems studied.However, efforts will be continued toformulate a satisfactory framework,based on experiments in which totalsulfate and total acidity are variedindependently. Perchloric acid willbe used in some cases as a source of

excess acidity which is considered notto contribute any complexing anions,and sodium or lithium hydroxide willbe used for lowering the acidity whileretaining the sulfate concentration.It is contemplated that all solutionswill first be tested for phase stabilityin the Marshall, semimicro, phase-study apparatus, and then recombinationstudies will be carried out on the

stable solutions. As mentioned in the

previous report, efforts will be madeto use this reaction as a means for

determining the relative complexingpower of other cations, such as uranyland zirconyl. It is also of interestthat any satisfactory determination of

96

the mechanism of the recombinationreaction will probably include a determination of the second ionization

constant of sulfuric acid at elevatedtemperatures. Once a clearer idea ofthe active species is obtained, it isplanned to again test the dependenceof the reaction rate upon temperatureand upon the concentration of thecopper catalyst.

SOLUBILITY OF HYDROGEN AT HIGH

TEMPERATURES AND PRESSURES IN SOLUTIONS

OF URANYL SULFATE

C. E. Ronneberg

The full elucidation of the reactions

involved in the recombination of hydrogen and oxygen in solution ascatalyzed by dissolved copper sulfateunder the conditions that prevail inthe aqueous, homogeneous reactors ofinterest (high temperatures andpressures) requires a knowledge of thesolubility of hydrogen in aqueoussolutions of uranyl sulfate. Thisknowledge will be significant inconnection with the operation of theHRE.

The solubility of hydrogen in purewater has been shown to increase withtemperature, above a certain minimumtemperature (about 200°F*3^). A directmethod was used at BMI to determine

the solubility, that is, the amount ofdissolved hydrogen in a known amountof water, for a range of temperaturesand pressures.

The direct determination of the

solubility of hydrogen in solutions ofuranyl sulfate is impossible becausethe uranyl ion is reduced by hydrogenat reactor temperatures, with the consequent precipitation of lower oxidesof uranium.'5' It is hoped to getaround the difficulty mentioned aboveby the indirect, kinetic method outlined below.

(5) H. F. McDuffie et al., HRP Quar. Prog. Rep.Aug. 15, 1951, ORNL-1121, p. 146-148.

Kinetic Method for the Determination

of the Solubility of Hydrogen. Therecombination of hydrogen and oxygenin solution to form water is catalyzedby even very small amounts ofdissolvedcopper sulfate:

H2 + K0a(CuS04)

-> H20

The rate of the reaction can beconveniently followed by observing therate of decrease of the pressure ofthe mixture of gases (principallyhydrogen, oxygen, and steam) above asolution of uranyl sulfate, with addedcopper sulfate, in a suitable bombthat initially contains a known amountof hydrogen and oxygen, measured interms of their partial pressures.Under these conditions, the uranylsulfate is not reduced. The recombi

nation reaction is first order with

respect to hydrogen and copper sulfate.The rate of decrease in the pressure(dP/dt) of the gaseous mixture inequilibrium with the solutions dependsupon the rate of the recombinationreaction and the ratio of the volumeof the gaseous mixture (V„) to thevolume of the solution (V,). Thefollowing equation can be shown toapply:

dP

dt P

. + iV, RT

Cu

Equation la follows from the basickinetic equation describing the rateof disappearance of dissolved hydrogenand the basic assumption that the rateof the recombination reaction isdirectly proportional to the concentrations of dissolved hydrogen andcopper, that is,

d [Hi

' sol

[Cu++] (la)

dt

wh ere k

= [H,] [Cu] kI2J L^"J KCu , (16)

is the specific velocityCu

constant due to copper sulfate. Theconcentration of dissolved hydrogen


can be related to the pressure of thehydrogen in the system by Henry's law,

[H2] =_Ll .a

Equation 16 can be rewritten as

dP„

= PB [Cu] kCudt «

moles dissolved hydrogen

moles hydrogen in system

"PH2 CCu] kCu •FG , (lc)

where Ffl is a geometry factor and isequal to the fraction of the totalhydrogen in the system which is in theliquid phase.

It has been shown that'6)

r<-7L a g a

VG RT VL ' RT + 1where

VL = volume of liquid phase,VG = volume of gas phase,R = gas constant,T = absolute temperature in °K,ex = Henry's law constant for hydrogen.

With the use of the value for FQ givenabove, Eq. lc can be modified, and itbecomes

dPu

IT - P«2 [CU] fecuG a

. + iVL RT

which leads directly to Eq. la above;and PH can be replaced by the total

(6)CAap. i.8, ENGINEERING DESIGN INFORMATION,from Section 4, "Aqueous Homogeneous Systems,"Vol. 2 of the Reactor Handbook, p. 58, OBNLCF-52-12-127 (Feb. 13, 1953).

97


pressure of 2:1 hydrogen-oxygen (obtained by subtracting the pressure ofsteam plus excess oxygen from the totalobserved pressure) for convenience inhandling the data.

Equation la can also be convenientlymodified. By letting

dP

dt

and rewriting,

1

K

The kn for various VG/V, ratiosmeasured experimentally for a certainsolution of uranyl sulfate at suitable

temperatures and pressures.In order to determine the Henry's

law constant for hydrogen, a bomb andapparatus are being built to measurekn, or (-dP/dt) • (1/P), for varioussolutions of uranyl sulfate at variousVG/Vi ratios. It is proposed toevaluate a by trying various ways ofhandling the experimental data. Itwill be noted that Eq. 2 has thegeneral form of a first-degree equationwith two variables, y = mx + 6.Equation 2 can be put in this generalform in a number of ways:

Method 1

Method 2

Method 3

a

RT— + 1

sol

a

RT

a

RT

a

RT

sol

1

P

k sol

+ 1

sol

+ ksol

(2)

1 s

In Method 1, for example, the experimental values for k ,/k„ for" S O 1 ' 7T

various V-/V, ratios would be plottedas indicated in Fig. 48. The relationship should be a straight line. The

98

intercept on the y axis should be 1.00and the slope should be a/RT. Hence,a = experimental slope x (RT), inconsistent units. Future experimentation should indicate which of the

above-suggested procedures is best fortranslating k^ and Vc/V, values intovalues for a.

Test of the Theory. A rough testof the above theory and methods forthe indirect determination of a for

hydrogen has been made by determiningk^ at 250°C in a small bomb (designedfor another purpose) containing ^-anylsulfate solution with 262 g of u.aniumper liter and 0.001 M CuS04. Valuesfor dP/dt and k^ were determined forvolume ratios of 3.02, 1.2, and 0.29.

There was no noticeable reduction

of the uranyl sulfate by hydrogen. Aplot of l/kn vs. VG/V,, made as indicated above, gives a slope of 9.1 andan a of 5740 liters*psi•mole" . Thisresult is sufficiently encouraging towarrant proceeding with the construction of a bomb and furnace

designed expressly to measure akinetically.

Experimental Details. A titaniumbomb with capacity of about 55 ml hasbeen built. It was designed to hpve avery comfortable safety factor at

pressures as high as 2500 psi.An aluminum block furnace with

automatic controls to operate at 250 Cis being built. It is hoped that

10

SLOPEa

RT ~\

1.0 2.0 3.0

Fig. 48. Plot of k ,/k vs. VJV.Sol' 7T G' jG' L

temperature control to 1°C or less canbe obtained.

DESCRIPTION OF FILTRATION APPARATUS

USED FOR AQUEOUS SOLUBILITY

DETERMINATIONS ABOVE 100°C

J. S. Gill E. V. Jones

W. L. Marshall

Experimental data have been obtainedduring the past year on the solubilityof uranium trioxide in aqueous uranylsulfate and uranyl fluoride solutionsabove 100°C.(7) Solubility data alsohave been obtained for the solubilityof fission products in aqueous uranylsulfate and in water by Zemel, Lietzke,Stoughton, Gill, Jones, and Marshall,both cooperatively and independently/ 'One of the experimental methods usedin this work has been the filter-bomb

technique by which a solution isequilibrated and filtered at theequilibration temperature. Theapparatus, designed by Zemel andStoughton^9) and of the type used byBooth, Bidwell, and previous workers,as reported by Booth and Bidwell,(10)was modified by Gill, Jones, andMarshall and is described for the

first time, in detail, to show themodifications.

The purpose of the apparatus is toequilibrate a solution and a solid ata constant temperature and then to

filter the solution at that temperature.This objective is accomplished by incorporating a filter unit in theequilibration apparatus. Filtrationis carried out by establishing a difference in pressure between the vesselcontaining the solution in contactwith the solid and the collection

vessel for the filtrate.

(7)W. L. Marshall et al., HRP Quar. Prog. Rep.,ORNL-1221, p. 99; ORNL-1280, p. 175; ORNL-1318,p. 142.

(8)Data reported in HRP Quar. Prog. Reps. for

1950, 1951, 1952.(9)'HRE Quar. Prog. Rep. Aug. 31, 1950, ORNL-826,

p. 78-79.

(10)H. S. Booth and R. M. Bidwell, J. Am. Chem.Soc. 72, 2567 (1950).


The equilibration bomb is shown inFig. 49a. The titanium filter bomb,which consists of two cells, A, issealed in the middle with a platinumgasket, B, and there is a modified,Monroe, porous-platinum filter "stick"unit, C, between the two cells. The

assembled filter bomb is placed in astainless steel outer jacket, D, whichis sealed in the center by a Teflongasket, E. After a solution and theexcess solid are placed in the lowerhalf of the filter bomb, a titanium-clad soft-iron slug, F, is inserted.This slug, which is used as a magneticmixer, has Teflon caps on its endsto prevent frictional contact ofmetallic surfaces. The magnetic slugis activated by means of a solenoid,G, which is controlled with an automaticon-off switch. The solenoid is

enclosed in glass to prevent contactand consequent short circuiting by theliquid salt. The assembled unit isplaced in a liquid-salt thermostatcomposed of KN03, LiN03, and NaN03,( **}and operated with the required temperature accuracy of ±0.5CC at 250°C.After sufficient time has been allowed

for equilibration, the unit is invertedwithin the salt bath and set in a

cooling unit, H, which consistsessentially of a copper coil throughwhich air is blown. Cooling of thelower half of the unit reduces the

vapor pressure at that end, and thusfiltering of the solution from thesolid takes place because of the vapor-pressure difference. The temperaturereduction in the lower end is of the

order of 6 to 8°C. The solution andthe solid are analyzed for theirspecific components.

Another equilibration apparatus isshown in Fig. 496. In this apparatus,the outer steel jacket is left off thefilter bomb and the bomb is inserted

into an aluminum heating cylinder. Theunit is rocked back and forth through

J. A. Beattie, Rev. Sci. Instr. 2, 458(ID

(1931).

99

oo

A - TITANIUM EQUILIBRATION BOMB

B - PLATINUM GASKET

C - MODIFIED MONROE, POROUS-PLATINUM FILTER "STICK" UNITD - STAINLESS STEEL OUTER JACKET

E - TEFLON GASKET

F - TITANIUM-CLAD SOFT-IRON SLUG

G - SOLENOID

H - COOLING UNIT

0. LIQUID-SALT THERMOSTAT 6. ROCKER-HEATER UNIT

Fig. 49. Filtration Apparatus Used for Aqueous Solubility Studies Above 100 C.

UNCLASSIFIED

DWG. 19257A

s93TS

OG>93HM93r

S3oo93M

03

93W13O93H

an angle of about 120 deg 45 times perminute by means of a series of cams.During equilibration time, both halvesof the cylinders are heated uniformlywith twin heating coils. After equilibration, a locking device incorporatedin the rocker assembly is loosened andthe heater unit is inverted. The lowerend of the unit is then cooled severaldegrees by cutting off the current tothe heating coils in the lower half ofthe unit, and filtration occurs. Tospeed up equilibration with concentrated solutions, a small, long-stemmedtitanium cup was introduced into thebomb to hold one-third to one-half thesolid phase near the top of the solution.

SOLUBILITY OF U03 IN AQUEOUS U02S04AT 175°C

E. V. Jones

Investigations of the solubility ofuranium trioxide in uranyl sulfate solutions at 175°C, which have been reported in part,'12) have been continued in an effort to determine

the behavior of this system in thelow concentration region, that is,below 0.1 If. As previously stated,(12)"The question is whether the equilibrium [mole] ratio, U to S04, dropsbelow 1.00 as concentration is lowered.

If the ratio does fall below 1.00,then the data would best be representedby the system U03-H2S04-H20. On thisbasis, there is no valid reason whythe ratio of U to S04 must remainequal to or greater than 1.00. Also,for a given sulfate concentration, theratio should decrease with temperature

because of increased hydrolysis of theuranyl ion."

Data for the solubility of uraniumtrioxide in uranyl sulfate solutionsat 175°C in the low concentrationregion are given in Table 26 and shownin Fig. 50. The U-to-S04 ratio appears

(12)E. V. Jones, J. S. Gill, W. L. Marshall,and C. H. Secoy, Chem. Quar. Prog. Rep. Sept. 30.1952, ORNL-1432, p. 51.


DWG. (9000A

0.8

0.6

00 o n

0

o

6c/

1

Jo

0.06 0.08

UMfl

0.10

Fig. 50. Solub-ility Ratio of U to

trations at 175°C.

to drop rapidly with decreasing concentration, below approximately 0.01 M.Similar studies are being continued at250°C.

PHASE DIAGRAM FOR THE U03-S03-H20SYSTEM AT 25°C

Work is in progress on the completion of the phase diagram for the

y3~^3U0,-S0,-H20 system at 25°C.

SOLUBILITY OF URANIUM TRIOXIDE IN

URANYL FLUORIDE SOLUTIONS AT

25 and 100°C

Three runs at 25°C and three runs

at 100°C on the solubility of uraniumtrioxide in uranyl fluoride solutionsof varying concentrations have givenvery discordant results. These systemsseem to be quite unstable, especiallyin concentrated uranyl fluoride so

lutions at 25°C. Further investigations are in progress.

EQUILIBRATION TIMES FOR URANIUMTRIOXIDE IN URANYL SULFATE AND URANYL

FLUORIDE SOLUTIONS

The results obtained in solubilitystudies of uranium trioxide in uranylsulfate and uranyl fluoride solutionsmake desirable an investigation of

101


TABLE 26. SOLUBILITY OF U03 IN U02S04-H20 SOLUTIONSOF LOW CONCENTRATION AT 175°C

Molarities based on room-temperature densities

TOTAL MOLARITY MOLARITY SOLUBILITY-EXCESS TOTALOF OF MOLARITY MOLE RATIO pH

URANIUM so4 OF U03 U03 to S04 (at 25°C)

0.119 0.100 0.0190 1.190 3.250.1046 0.0978 0.0067 1.069 3.40.0638 0.0568 0.0070 1.070 3.20.0630 0.0573 0.0057 1.100 3.60.0622 0.0573 0.0049 1.0 86 3.70.0475 0.0428 0.0047 1.1100.0458 0.0425 0.0033 1.078 3.50.0437 0.0393 0.0044 1. 112 3.50.0284 0. 0258 0.0026 1.101 3.30.0271 0.0260 0.0011 1.0420.0232 0.0229 0.00033 1.014 3.50.01143 0.01114 0.00029 1.0260.01109 0.01083 0.00026 1.0240.00664 0.00687 -0.00023 0.967 3.70.00584 0.00802 -0.00218 0.7280.00567 0.00612 -0.00045 0.9260.00563 0.00635 -0.00072 0.887 3.50.00546 0.00612 -0.00066 0.8920.00370 0.00347 +0.00023 1.066

0.00312 0.00318 -0.000 06 0. 9810.00239 0.00331 -0.00092 0.722 4.2

equilibration times for these systems.Studies are under way for both thesesystems at 25 and 100°C.

EXPLORATORY PHASE INVESTIGATIONS OF

SYSTEMS INVOLVING CuS04, U02S04,H2S04, and H20 AT URANIUM CONCEN

TRATIONS BELOW 30 g PER LITER

W. L. Marshall

An exploration has been made of thefollowing systems with regard totemperature at which either hydrolyticor true solubility precipitation occurs:

CuS04-H20 ,

U02S04-H20 ,

U02S04-CuS04-H20 ,U02S04-CuS04-H2S04, H20 .

102

The micro phase-study apparatuspreviously described was utilized forthe investigation.(13> About 50 tubescontaining the solutions were preparedand tested during the course of thework.

The purpose of the investigationwas to define boundaries - not toobtain exact solubility or hydrolyticprecipitation temperatures. Accordingly, the temperature was raisedslowly, but not slowly enough todetermine the precise precipitationtemperature. This exact investigationwould have delayed attainment of datato a considerable degree. Also, not

(13)H. W. Wright and W. L. Marshall, HRP Quar.

Prog. Rep. Oct. 1. 1952, ORNL-1424, p. 108.

mmwmmmmm mmmmmmmmw*

much emphasis was placed on data obtained below 200°C, the prime interestbeing in the behavior between 200 and300°C. The actual phase - stabilitytemperatures at equilibrium may be asmuch as 10 to 20°C below the valuesreported. In some systems, namely,U02S04-CuS04-H20 and U02S04-CuS04-H2S04-H20, the solid phase appearedto dissolve rapidly at room temperature after precipitating at the uppertemperature, but it did not dissolverapidly at temperatures just below thesolid-phase appearance temperature.For the other two systems, the precipitations appeared to be irreversible,at least under the experimentalconditions.

There is a remote possibility thatreaction occurred with the pyrex andsilica tubes used in this work. However, some tests were made at 250 C onsolutions which had shown precipitation at temperatures between 200 and240°C. Upon inserting tubes containingthese solutions into the furnace at250°C, precipitation occurred in aslittle as 5 seconds. It does not seem

likely that the solution reacts thisfast with the pyrex or silica.

The System CuS04-H20. Hydrolysisdata for the system CuS04-H20 from 1.0to 1.0 x 10"3 M, as evidenced byprecipitation of a bright blue-greensolid phase, are given in Table 27.The solid phase must be a hydrolysisproduct because it appears in the sametemperature range over such a widechange of concentration. Furtherevidence of hydrolysis is given in therather complete investigation byPosnjak and Tunell,(14'15} from whichthe data in Table 28 were calculated.

The System U02S04-H20. Temperaturesat which hydrolytic precipitation of ayellow solid occurs in the U02S04-H20system are given in Table 29. The

(14)E. Posnjak and G. Tunell, ,4». J. Sci. 18,1 (1929).

(15)G. Tunnel and E. Posnjak, J. Phys. Chem.35, 929-46 (1931).


TABLE 27. HYDROLYTIC PRECIPITATIONOF CuSO.

MOLARITY

OF

CuS04

TEMPERATURE AT WHICHPRECIPITATION FIRST OBSERVED

<°C)

0.00116 115

0.00402 100

0.0161 115

0.0532 111

0.103 126

0.204 125

1.01 142

data indicate that the minimum concen

tration of a UO2SO4-H20 solution mustbe 0.017 M to achieve phase stabilityat 250°C.

The System U02S04-CuS04-H20. In thework with the system U02S04-CuS04-H20,U02S04 is held constant at 0.021 M andthe CuS04 concentration is varied. Itappears from the data given in Table30 that a phase-stable system of CuS04-U02S04-H20 with 0.021 MU02S04 andmore than 0.0021 M CuS04 could not bemaintained at 250 C.

The System U02S04-CuS04-H2S04-H20.There are four conditions of thesystem U02S04-CuS04-H2S04-H20 whichhave been considered. These are theeffect of H2S04 on 0.021 MU02S04 containing 0.02 or 0.10 M CuS04 and theeffect of H2S04 on 0.126 M U02S04 containing 0.02 or 0.10 M CuS04. Thedata are given in Table 31. Since0.126 M UCLSCL stabilizes 0.10 M CuSO

2 4 • ?at 250 C, the same concentration oiU02S04 should stabilize 0.021 MCuS04.No work was done on this latter system.

General Conclusions. The generalconclusions that can be drawn fromthis work are:

1. Copper sulfate hydrolyzes andsubsequently precipitates in thetemperature range 90 to 140 C, atleast at concentrations from 1.0 to1.0 x 10*3 M.

2. At 250°C, the minimum concentration of U02S04 possible without

103


TABLE 28. THE SYSTEM CuO-SO,-H„0

Calculated from data by Posnjak and Tunell(14•15>

CONCENTRATION (moles/1000 g of solution)

Temperature ( C) CuO SO, Mole Ratio of SO-""3 to CuO J

50 0.0063 0.0062 0. 980.016 0.016 1.000.21 0.21 1.000.44 0.45 1.02

100 0.00013 0.00025 1. 920.0010 0.0011 1. 100.0025 0.0025 1.000.048 0.050 1.040. 15 0.15 1.000.46 0.48 1.04

200 0.0025 0.0062 2.480.021 0.037 1.760.30 0.37 1.230.57 0.80 1.40

TABLE 29. HYDROLYTIC PRECIPITATION OF UO SO2 4

MOLARITY OF TEMPERATURE AT WHICH PRECIPITATIONuo2so4 FIRST OBSERVED (°C)

0.00624 22 50.0125 2150.0150 2100.0166 2550.0208 325 to 3550.0312

0.0625No solid phase, two-liquid layer at about 330°CNo solid phase, two-liquid layer at about 315°C

0. 125 No solid phase, two-liquid layer at about 306°C

TABLE 30. HYDROLYTIC PRECIPITATION IN SYSTEM U02S04-CuS04-H20 (0.021 MU02S04)MOLARITY

OF CuSO.

0.00210

0.00368

0.00525

0.0101

104

TEMPERATURE AT WHICH PRECIPITATIONFIRST OBSERVED (°C)

250

255

250

190 to 215

APPROXIMATE COLOR OF

CRYSTALS

Yellow

Dull blue-green and yellow

Dull blue-green

Dull blue-green


TABLE 31. HYDROLYTIC PRECIPITATION IN SYSTEM U02S04-CuS04-H2S04-H20

MOLARITYOF

uo2so4

MOLARITYOF

CuSO.

MOLARITYOF

H,SO„

TEMPERATURE AT WHICHPRECIPITATION FIRST

OBSERVED (°C)COLOR OF PRECIPITATE

0.0210

0.0210

0.0210

0.0210

0.0210

0.0210

0.0210

0.0210

0.020

0.020

0.020

0.020

0.100

0.100

0.100

0.100

0.126 0.100

0.126 0.100

0.126 0.100

0.00514

0.0103

0.0154

0.0206

0.0206

0.0308

0.0412

0.0514

None

0.0103

0.0206

240

236

260

285

196

244

280

No crystals observed, two-liquid layer at about

342°C

240

288

Two-liquid layer at 298 C

None, two-liquid layer at

306°C, raised to 350°Cwith no solid appearing

Questionable yellow

Questionable yellow

Questionable yellowQuestionable yellow

Bright blue-green

Bright blue-green

Bright blue-green

Bright blue-green

Bright blue-green

Bright blue-green

hydrolytic precipitation is approximately 0.017 M. Constituents such asNiS04,(16) other corrosion products,CuS04, and fission products will alterthis value because of the solubilityor hydrolysis effects of additionalcomponents on phase equilibriums.

3. At 0.021 M U02S04, solutionscontaining more than 0.0021 M CuS04will not be phase stable at 250 C.

4. Approximately 0.015 M H2S04will be required to maintain phasestability for the solution which is0.021 M U02S04 and 0.020 M CuS04. Asolution 0.035 M in H2S04 will berequired to maintain phase stabilityfor 0.021 MU02S04 and 0.10 MCuS04.For 0.126 M U02S04 that is 0.02 M inCuS04, no H2S04 will be needed at250°C, but approximately 0.005 M H2S04will be required for 0.126 M U02S04that is 0.10 M in CuS04.

(16)J. S. Gill and C. H. Secoy, private com-municati on.

COMMENT ON THE SYSTEM URANYLDICHROMATE-WATER

W. L. Marshall

In earlier studies, the apparentsolubility of uranyl dichromate in thesupercritical fluid that exists atabove the critical temperature formixtures of U03-Cr03-H20 which aresoluble at room temperature was observed by Gill, Secoy, and Marshall.There was some doubt at that time asto whether uranium actually was dissolved in the fluid. With the adventof the semimicro phase-study apparatus,tube explosions were eliminated, andthus an easy method of investigationof this system was afforded.

In the present work, solutions ofchromium trioxide, from 1.0 to 4.0molar, in water were observed in silicatubes in the semimicro apparatus. Decomposition of dichromic acid torefractory Cr203 occurred between 325and 370°C. No orange-red color was

105


observed in the supercritical fluid;however, vision was obstructed somewhat by the deposit of Cr203 on thetube surfaces. It is concluded thatthe orange-red color previously observed by Gill, Secoy, and Marshallwas not due entirely to a chromiumcompound in the supercritical fluid.

Further investigations were made onstoichiometric U02Cr207 from 0.50 to0.050 molar and on solutions containing20% excess and 20% deficient UO++ in

, 2the same concentration range. In allexperiments the supercritical fluidhad an orange-red color. There werealways deep-red crystals present abovethe critical temperature of the system.

This investigation is of interestbecause B. J. Thamer, R. M. Bidwell,and R. P. Hammond of the los AlamosScientific Laboratory recently showedthat uranium oxide dissolves in thesupercritical fluid of the systemU03-H3P04 at certain compositions ofU03 and H3P04.(17>

AN APPARATUS FOR PHASE-STABILITYSTUDIES AND CORROSION TESTS ON

2-ml SOLUTION VOLUMES

W. L. Marshall

In August and September of 1952,exploratory investigations were madeof aqueous uranyl fluoride solutionscontaining dissolved fluorides andhydroxides of alkali and alkalineearth metals.(18) Similar explorationswere made of sulfate systems. Some ofthe solutions containing 300 g ofuranium per liter which had pH valuesof 5.00 to 6.00 showed phase stabilityup to and above 300°C. However, itwas not possible to make accurateobservations in silica tubes on thehigh pH solutions because of thereaction of silica with the solutionsto form insoluble silicates. As the

(17)B. J. Thamer, R. M. Bidwell, and R. P.

Hammond, private communication.(18)

H. 0. Day et oI., HRP Quar. Prog. Rep.Oct. 1, 1952, ORNL-1424, p. 113.

106

temperature was raised, the formationof precipitates occurred at about220°C in all solutions of pH between5.00 and 6.00. It was necessary tostudy the phase stability and cor-rosiveness of these solutions instainless steel, titanium, or platinum-lined bombs. It was deemed desirableto have a small apparatus in which manysolutions could be tested with relativeease, since the large-scale apparatuscurrently available made this exploratory multitesting procedure impractical. If the apparatus were small andsimple in design, many such unitscould be operated at a minimum ofcost. Such an apparatus has beendes igned.

Description of Apparatus. Aschematic drawing of the apparatus isshown in Fig. 51. The thermostatsystem consists of a 4-liter stainlesssteel beaker containing a liquid saltmixture of KN03, NaN03, and LiN03,(11>which melts at about 120°C and isheld at the required temperature bysuitable heating elements and astirrer. The container tube for holdingthe small bombs is fabricated fromstainless steel and is heliarc weldedat the bottom. The inside dimensionsare 5/8 in. in diameter and 8 in. inlength. After the unit is loaded witha maximum of three bombs, it is attachedto a conventional laboratory motor androtated in the thermostat at 1800 rpm.

The outside dimensions of thestainless steel or titanium bombs are1 3/4 in. in length and slightly under5/8 in. in diameter. The containedvolume is 3.2 milliliters. Teflon isused successfully asagasketing agent.The Teflon ring gasket fits into a groovein the bomb cap, which prevents thering from extruding from the gasketingsurface. Because of the expansioncharacteristics of Teflon, the bombcap can be tightened with hand-pressurealone. The expansion of the Teflonwhen the temperature is raised willseal the bomb. Similarly, upon cooling,only hand-pressure is needed to open

UNCLASSIFIEDDWG 16784A

ALNICO PERMANENTMAGNET-

MAGNETIC ROTOR

CORROSIONSPECIMEN

Fig. 51. Experimental Apparatus forPhase Study and corrosion Experiments.

the bomb. In actual operation, however, asmall wrench is used to tightenthe bombs. Teflon appears to be asatisfactory seal for aqueous solutionsat temperatures up to 325 C, at leastfor this type of gasket.

The magnetic rotor is prepared bysealing a permanently magnetized metalin either a type 347 stainless steelor a titanium capsule. The ends of thecapsule are heliarc welded. No leakageoccurs between the magnetized metaland the external system.


Teflon caps placed on the ends ofthe rotor have been used, in someinstances, to eliminate erosion causedby metallic contact between the rotorand the bomb surface.

Corrosion specimens are fabricatedfrom the metal to be tested. A specimen is screwed into the bomb cap andit rotates with the bomb, which, inturn, rotates with the container tube.

Under operating conditions, acontainer tube is loaded with threebombs, each containing a solution, acorrosion specimen, and a magneticrotor. The unit is lowered into thethermostat and rotated at 1800 rpm.Three sets of two Alnico permanentmagnets are adjusted to fit closely tothe sides of the thermostat adjacentto the respective bombs. These magnetscause the rotors within the bombs torevolve around the inside diameter andprovide mixing. Three container tubescontaining three bombs each can beoperated in one 4-liter thermostatbath.

Proposed Use of Apparatus. It isproposed to use this apparatus bothfor phase-stability studies, in whichcase the corrosion specimen will beomitted, and for corrosion studies.With respect to phase-stabi1ityinvestigations, the solution beingstudied will be run at the requisitetemperature, the bomb will be cooledand opened, and the solution will beexamined. Since cooling and openingof the bomb are completed in 60 sec,during which time there is no mixing,little if any equilibrium reversalshould occur.

With regard to corrosion runs, thecorrosion test specimen will be included in the bomb and microphotographswill be made of the specimen beforeand after a run. The specimen canalso be weighed, and the solution canbe analyzed for corrosion products.

The erosion on the sides of thebomb caused by the revolving rotor isone of the disadvantages of thisapparatus. It is anticipated that the

107


use of titanium bombs, which have notbeen used up to the present time, and,also, the use of Teflon caps on theends of the rotors, will eliminatethis erosion difficulty.

If oxygen pressure within the bombis desired, a calculated amount of 30%hydrogen peroxide can be added to eachbomb before it is sealed. The peroxidedecomposes at temperatures above 100°Cand yields oxygen.

EXTENSION OF TWO-LIQUID PHASE STUDIES

OF UO.SO^2 4

H. W. Wright

An apparatus for more detailedstudies of the miscibility break in thesystem U03-S03-H20 has been designed,and fabrication is being completed.Exploratory experiments performed inquartz tubes indicate that both thefission-product activity and the copperions are largely associated with theheavy, uranium-rich phase. Table 32summarizes the results of one experiment for which a sample of HRE fuelsolution was used.

OXIDATION OF SULFUR DIOXIDE IN URANYLSULFATE SOLUTION

D. W. Sherwood

Present plans for intermittentaddition of acid to the inpile loopinvolve the addition of a mixture of ameasured amount of sulfur dioxide and

oxygen to!react in the presence of theions of the solution to form sulfuricacid. Experiments are in progress todetermine the feasibility of thisprocedure with reference to the speedof the oxidation and the possiblereducing action of the sulfur dioxidewith regard to the uranyl ion insolution. Although the analyticaldata and the experiments are incomplete, present results seem to indicate that there has been no damage tothe solution and that the desiredoxidation is occurring in 1 hr or lessat 200 C. The solution contains 40 gof uranium per liter and small amountsof copper and nickel ions to simulatea uranyl sulfate solution after exposure to loop metal. The desiredgases are added to a quartz tube containing the frozen degassed solution.The sealed tubes are then heated in aresistance-wire-wound aluminum block.

TABLE 32. DISTRIBUTION OF FISSION-PRODUCT ACTIVITY

B~ACTIVITY

ACTIVITY (counts/min /ml) DISTRIBUTION

Urani urn

Phas e

-Rich

H

Water-Rich Phase RATIO

Aliquot 1, L. Aliquot 2, L2 H/L1 H/L2

Gross 2.66 x 109 1.25 x 107 1.58 x 107 213 168

TRE* 2.02 x 10 9 6.72 x 106 9.14 x 106 301 221Nb 5.00 x 10s 2.44 x 103 2.25 x 103 205 222

Ru 6.00 x 104 2.5 x 103 2.23 x 103 24 27

Zr 3.40 x 10s 2.8 x 104 1.15 x 104 12 30

Sr 2.94 x 108 3.86 x 106 4.53 x 106 76 65

Cs 4.22 x 106 6.84 x 104 6.30 x 104 62 67

•Total rare earths.

108

CHEMISTRY OF CORROSION

G. H. Cartledge R. P. Yaffe

The study of the electrode potentialof type 347 stainless steel in 0.1 Fsulfuric acid exposed to air has shownthat the electrode may be made morenoble by the addition of small quantities of certain heavy-metal ions.Experiments with radioactive silvershowed that the ennobling could beobtained by precipitation on onlyminute amounts of silver on the

electrode. The ennobling is oftenlinear with the logarithm of theadded ion concentration and, incertain cases, may amount to as muchas 300 mv. Quantitative data arebeing accumulated in the hope thatthey may make possible a fuller interpretation of the electrochemicalprocesses related to corrosion. Theresults may have significance inconnection with possible impuritiesin reactor solutions.

The pertechnetate inhibition studieshave been continued chiefly alonglines related to the theory of theinhibition process. Details forcertain tests may be found in ORNL-1529.<19>

VAPOR PRESSURES OF AQUEOUS URANYL

SULFATE SOLUTIONS

H. 0. Day, Jr.

Some vapor pressures of uranylsulfate solutions at various concen

trations and temperatures were reported previously.(20> In moving tonew quarters in Building 4500, part ofthe differential pressure apparatuswas accidentally broken, and it wasnecessary to construct an entirely newapparatus, similar to the one previously described; however, severalsmall but important changes were

(19) Oak Ridge National Laboratory Status andProgress Report March 1953, ORNL-1529, p. 22.

(20)H. 0. Day, Jr., HRP Quar. Prog. Rep. July 1.1952, ORNL-1318, p. 138; and H. 0. Day, Jr. andW. L. Marshall, HRP Quar. Prog. Rep. Oct. 1, 1952,ORNL-1424, p. 104-105.


made. The manometer tubes were made

longer so that greater pressure differences could be accurately determined. A constriction of capillarytubing was placed in the U of themanometer to damp, as much as possible,any pressure fluctuations.

The tubing connecting the pressurevessels with the manometer was con

structed of thick-wal1ed, glass,capillary tubing. The thickness ofthe walIs caused temperature variationsin the vapor to be drastically reduced.At the same time, the volume of watervapor was also greatly lowered, andtherefore the corrections which must

be applied in calculating concentrations of uranyl sulfate were reduced.The capillary tubes were wrapped withnichrome wire so that they could beheated to prevent condensation.

The thermostat in which the pressurevessels were located was filled with

dibutylphthalate instead of with thegl ycerol-water mixture used previously.Temperature control thereafter wasmuch better at elevated temperatures.Loss of dibuty1phtha 1 ate was muchsmaller than the glycerol loss atthese temperatures. Dibutylphthai atewas preferred over mineral oil, sinceit is easier to stir at lower temperatures and is easier to remove from

the pressure vessels.

Considerable mechanical difficultywas experienced with the capillarytubing in that breaks resulting fromstrains introduced in heating andcooling the connecting tubing werefrequen t.

Two series of determinations of

vapor pressure have been completedrecently. m-Diphenyl benzene was themanometer fluid. Data for the first

series of measurements are given inTable 33. Data from the last series

are not reported at this time becausethe concentration of the uranyl sulfatehas not yet been determined. Uncertainties in the concentrations as a

result of analysis errors are ±0.0004in molality and ±0.000006g in molefraction.

10 9


TABLE 33. VAPOR PRESSURE LOWERING IN URANYL SULFATE SOLUTIONS

TEMPERATURE MOLALITY OF MOLE FRACTION OF AP

(°C) uo2so4 uo2so4 (mm Hg)

25 3.1722 ± 0.0004 0.054061 ± 0.0000066 2.110

30 3.1723 0.054062 2.764

35 3.1724 0.054064 3.696

40 3.1726 0.054066 4.731

45 3.1727 0.054069 5.886

50 3. 1729 0.0 54072 7.390

60 3.1735 0.0 540 82 10.762

70 3.1743 0.054094 16.166

80 3.1753 0.054110 23.272

90 3.1765 0.054129 32.848

SLURRY CHEMISTRY

F. R. Bruce, Section Chief

URANIUM TRIOXIDE CHEMISTRY

Preparation of U03 (L. E. Morse).In the previous report, preliminaryexperiments indicated that UO, of aquality suitable for use in fluid-fuelreactors which employ slurries containing 250 g of uranium per liter andare operated at 250°C could be prepared by either of the followingmethods:

1. a methathesis process in whichammonium diuranate is precipitatedfrom a uranyl nitrate solution andconverted to an ammonium uranylcarbonate, which is subsequentlydecomposed to anhydrous UO.,

2. an ion-exchange method in whichuo: is adsorbed on a cation-

exchange resin and eluted with(NH4)2C03 solution, and the resulting ammonium uranyl carbonatesolution i s decomposed to anhydrousU03.

The metathesis process has beeninvestigated more thoroughly with thepreparation of kilogram quantities of

110

U03. The several steps in the processare discussed below.

Preparation of Uranyl NitrateSolutions. A stock solution of 2.0 M

U02(N03)2 was prepared and diluted tothe desired concentration for precipitation of (NH4)2U20y. After thesolution was allowed to stand for

several days, a brown precipitatesettled out, which was subsequentlyidentified by spectroscopic analysisto be mainly an iron compound. Inmuch of the previous work, the use offresh uranyl nitrate solutions resultedin iron contamination. Therefore it

may be necessary in the future totreat the uranyl nitrate solutions toremove iron.

Precipitation of Ammonium Diuranate

2 U02(N03)2 + 6 NH40H

> (NH4)2U207 + 4 NH4N03 + 3 H20

Excess NH40H diluted 1:1 with H20 wasadded over a period of 1 hr to astirred solution of 0.5 M U02(N03)2


at 85°C. Completion of the pre- not offer any difficulty. The ammoniumcipitation was indicated when a pH of uranyl carbonate produced by this8 was obtained. method settles and filters rapidly.

The (NH4)2U207 obtained by this Since fNH4 )4U02(C03)3 is somewhatprocedure is a granular precipitate water-soluble, there is a uranium losswhich settles and filters rapidly, and associated with the metathesis step.it may be easily washed. Since the In five batches carried out on aprecipitation is carried out in alkaline kilogram scale, the average loss wasmedia, the iron originally present 2.2% U; however, this uranium is in awill also be carried. In this respect, form suitable for recycling.the process is at a disadvantage _, ...

„„ _ i -.i , . . Decomposition of Ammonium Uranylcompared with a uranyl peroxide pre- n , J viuuji.Carbonate

(NH4)4U02(C03)3cipitation, which affords some sepa^ration from iron.

The (NH4)2U207 may be filtered andwashed free of nitrate with dilute > U03 + 4 NH3 + 3 C02 + 2 H20NH.OHi However, since the nitrate is T, , . . . ,

* ^ . , -„. .,„ , . , . lne decomposition was carried out inpresent mainly as NH.NO,, which is ,.,.,.• j .. ..u * t . an open porcelain dish in a muffledecomposed at the temperature used to r • , > , .a /mu \ tt/-> ir^rs \ . iurnace with the temperature beingdecompose (NH4) 4U0, (CO, ) , , the pre- • , _ r „* Omo^r.i„i+ * a A. K \ ij l raised stepwise from 200 to 350 C overcipitate and the supernatant could be . , c , „ , ,r,„T.~±^ri 4-1 u û - .-l a period oi about 8 hours. Even, thoughcarried through the metathesis step , «->ug"without the additional treatment that the ^composition appeared to bewas required in the present test. ""^ Y comPlete' the bating at

350 C was continued overnight. TheMe thathe si s of (NH u) 2(/20 ? to resulting product was an amorphous

(NH u) 6UO 2(CO 3) 3 oxide of low bulk density.When large amounts of material are

(NH4)2U20? + 6 (NH4)2C03 + 3 H20 handled, the stepwise heating procedure

>2 (NH4)4U02(C03)3 +6 NH40H jf C8pe ^"V-* neC/SSary \° minimi-* 4 2 3 3 4 the possibility oi uneven heating and

. ,. u • • , consequent reduction of UO, by NH,.Alter the precipitate and the super- T , r , , . 3, . |„ ,. ,. j? .• . . In the furnaces employed in this work,natant iron) the previous operation , \. ' . , , '

ii j ^ i l i ..ô^ •, rr, tne top strata oi solids were de-were allowed to cool below 40°C, 15% , , • , r , .rn<>v~o°<, (mh 1 rn aa a a û composed lirst and iormed UO, on theexcess (I\H J„LO, was added, and the , , . . , 3

. ^ 423 undecomposed remainder. If the temper-mixture was stirred until metathesis ' , • , • ,,„„„ „. i ,. ^ / l t i i \ tu ^ ature is too high, ammonia evolvedwas complete (about 1 nr). The most r , . i

re

is m

, • »! ^ ^ j. , , irom the lower layers may reduce theliable test lor complete metathesis • , „, • , . • ,.

. i n , , • , oxide. Ihis reduction was, in fact,^ made by slowly decomposing a sample , , , ,..,. . ,T^ , „n '

f -u a . - oon . onnor a observed to occur when (NH. ) .UO, (CO, ) ,oi the product at 280 to 300 C and , , . „ 4 4 2 3 3l ,. • .. o-rcor- -r u .. i c was placed in a furnace at about 370 Cheating at 375 C for about 15 minutes. . ,, , . . ,,™, , . ii,,, ancl allowed to heat rapidly.Ine resulting product should be deepred in color, and there should be no Preparation of U03'H20green or black material to indicatethe formation of U3Og from the de- 85°Ccomposition of unreacted (NH4)2U207. U03 + H2° } to 2 hr > U(V2H20Failure of the decomposed carbonate todissolve completely in' HC1 would also ^0°rindicate the presence of U,0. in the U0,'2H„0 > UOoxide product and, hence, of unreacted iD nr(NH ) U.O, in the carbonate preparationbeing treated. Experience with ki 1ogram ijq + w q > TTO «H Dbatches indicates that metathesis does 3 2 8 to 16 hr 3 2

111


The anhydrous oxide from the decomposition step was hydrated forabout 12 hr at 85°C. Less than 0.1%of the uranium dissolved in the supernatant. After the supernatant wasfiltered, the resulting hydrate wasdecomposed to the anhydrous oxide byheating at 350°C for 16 hours. Rehydration at 250°C yielded monohydrateplatelets of suitable purity. An oxidepreparation is considered to be satisfactory if, after being hydrated overnight at 250°C, the supernatant contains less than 10 ppm of uranium fora slurry concentration of 250 g ofuranium per liter.

Crystallography of U03*H20 (J. 0.Blomeke). The crystal structure ofU03'H20 has been investigated over thetemperature range of 150 to 250 C, andcrystals have been prepared whichexhibit the same x-ray diffractionpatterns* '' as the a-U03*H20 and/3-UO "H 0 reported by Zachariasen. ( 2'A transition from the low-temperaturealpha rods to the high-temperature

the monohydrate is cooled past thetransition point at a rate of about2°C per minute. Revision to the alphaform occurs when the /3-U03*H20 isground at room temperature in a fluid-energy mill.

The importance of crystal structurestudies in the development of anaqueous slurry fuel is emphasized bythe definite relationship that canexist between the crystal structureand other properties of the solids,such as crystal habit, density, heatcapacity, and hardness. Aside fromthis, a thorough knowledge of thecrystal structure could be used tointerpret any changes that might occurin the characteristics of the slurryduring reactor operation.

The crystal structure of U03*H20has been investigated by Vier,(3)as well as by Zachariasen. ' Vierreported the existence of four differentstructures that are dependent on temperature and on the mode of preparation.These structures can be representedby the following schematic equations:

U04'*H2077 to 185°C

-> a-U03-H20185 to 300°C

-» y-U03-H20H20 -H20

A

U02(N03)2- -> Mallinckrodt UO,77 to 185°C

H20» /3-U03-H20

185 to 300°C

H20-» a-U03'H20

beta platelets occurs at approximately200°C. This transition may not occur,however, if the alpha form is annealedby heatingin water at 150 to 180°C forseveral hours before the temperatureis raised past the transition point.The transition from the beta form back

to the alpha form does not occur when

X-ray diffraction analyses performed by R.D. Ellison of the ORNL Chemistry Division.

(2 )W. H. Zachariasen, Report from July 1 to

December 31. 1946 of Mass Spectroscopy and CrystalStructure Division, p. 20, CP-3774 (Feb. 20, 1947).

112

Vier reported the alpha, beta, andgamma structures to be orthorhombicand the delta structure to be tri-

clinic. Unfortunately, eel 1 dimensionswere not given, and it has been impossible to establish the reproducibility of this work.

Zachariasen reported the celldimensions of two different U03*H20crystals but gave no information concerning the chemical history of hissamples. He indexed both these

(3)D. Vier, Thermal Stability of Certain Oxide

Hydrates in HÔ, A-1277 (May 26, 1944).

structures as orthorhombic and called

them a-U03*H20 and y6-U03-H20, independently of Vier's nomenclature.By use of the cell dimensions given,the positions of all possible lines inthe x-ray diffraction patterns werecalculated, and thus a comparisoncould be made with material preparedin this laboratory.

In the course of the present slurrydevelopment work, three well-definedcrystal habits have been reported toexist at 250°C in water: U03'H20prepared from UO'ncH 0 by heating inwater at 250°C assumes the crystalhabit of hexagonal rods, 1 to 5 /x indiameter by 5 to 30 /U. long; U03'H20prepared at 250°C from well-washedMallinckrodt L03 or from rods whichhave been calcined and rehydratedseveral times assumes a platelet habitof seldom more than 40 to 50 /x acrossthe face; and when Mallinckrodt U03 isheated at 250°C in dilute U02(N03)2solutions, a larger crystal, severalhundred microns on edge, that resemblesa truncated bipyramid is formed. X-ray diffraction studies have shownthat the rods conform to the patternof the a,-U0,'H20, whereas the plateletspossess the same pattern as is foundin the beta form in Zachariasen's work.

The study of a single crystal with theuse of a precision camera indicatedthat the bipyramids also had the betastructure and were probably plateletsin an advanced state of growth.

As suspensions containing 250 g ofuranium per liter, the rods and theplatelets had sufficiently differentslurry properties to justify a studyfor establishing which of the twoforms was more likely to be stableunder the conditions of reactor

operation. Heating of the rods andplatelets in water at 250°C, in theabsence of agitation, for as long as70 hr produced no detectable change inthe structure of either. In a dynamicreactor system, however, the crystalswould be expected to be fractured bothby abrasion and by the fission process.


Hence, specimens of both rods andplatelets were dried, pulverized in afluid-energy mill, and then put backinto water at 250°C. Platelet crystalswere reformed from both samples.Pumping tests carried out in the ReactorExperimental Engineering Division alsoindicated that conversion from rods to

platelets would occur after only a fewhours of turbulent circulation at

250°C.(4) Static radiation testscarried out at 250°C in the ORNL

graphite reactor with slurries of rodsshowed a marked tendency for plateletsto be formed as the rods were fraction

ated by fission recoils.'5' It hasbeen concluded from this evidence that

the platelets exhibiting the samex-ray diffraction pattern as that ofZachariasen's /3-U03*H20 are the stableform of U03*H20 in water at 250°C andthat they are the crystal form mostlikely to be present at this temperature in a reactor fuel.

During the past quarter, furtherevidence was obtained to indicate that

platelets are the stable crystal formof U03-H20 in water at 250°C. Also,the existence of a temperature-crystalstructure relationship was demonstrated. The experiments performedare presented schematically in Fig. 52.Initially, it was found that a transitionfrom /3-U03*H20 to a-U03'H20 occurs asa result of grinding platelets at roomtemperature. Although such a transitionfrom a "frozen-in" high-temperatureform to a 1ow-temperature form is notrare among inorganic salts, it has beenknown to occur much less frequentlywi th hydr ates.

The micronized platelets were thendivided into four parts and heated inwater for 16 hr at 150, 180, 215, and250°C, respectively. The samplesheated at 150 and 180°C retained thea-U03*H20 structure, with the fragments

(4)A. S. Kitzes et al., HRP Quar. Prog. Rep.Jan. 1, 1953, ORNL-1478, p. 110.

(5)F. R. Bruce et al., HRP Quar. Prog. Rep.Mar. 15. 1952. ORNL-1280, p. 107.

113


DWG 49062A

/3-U03-H20 (PLATELETS)

tMICRONIZED AT

ROOM TEMPERATURE

a-U03-H20 (FRAGMENTS)

a-uo3 -h2o(ROD FRAGMENTS)

a-uo3 -H20(ROD FRAGMENTS)

/S-UOj-HgO(PLATELETS)

/3-U03-H20(PLATELETS)

150°C

r 180 °C

ut 215 °c

, 250 °c

°2300 psi

H202

AIR•!

/3-uo3-h2o(PLATES)

a-uo3-H2o(RODS)

yS-U03-H20(PLATES)

Fig. 52. Structure Stability of U03'H 0.

growing in size and assuming a rodlike habit. The samples heated at 215and 250°C were converted to plateletsand they possessed the /3-U03*H20structure. These experiments indicatethat the transition temperature fromthe alpha to the beta form must lie inthe neighborhood of 200°C.

The question of what induces theformation of rods from U04"xH,0 at250 C may be partly answered by theremaining three experiments noted inFig. 52. Samples of micronizedplatelets were heated in water at250 C under atmospheres of oxygen thatwere introduced in two different ways.In one case, the oxygen was introducedas the gas to a pressure of 300 psi at250°C. In the second case, 30% H202was added in a quantity that wouldproduce an oxygen pressure of 300 psiupon decomposition at 250°C. Afterbeing heated for 16 hr, the micronizedrods had been converted to plateletsin the first case, but they had retained

114

their alpha structure and regained therod crystal habit in the second case.As before, the customary conversion toplatelets occurred at 250°C when onlyair was present. These experimentsseem to indicate that the presence ofperoxidic oxygen intimately associatedwith the fragmented crystals promotesthe formation of a-U03'H 0, even attemperatures in excess of the transitionpoint.

It is to be noted from these studies

that only two structures of U03'H20were ever found and that these con

formed to those reported by Zachariasen.The existence of a third, orthorhombicstructure or of a triclinic pattern,as reported by Vier, was not substantiated.

CHEMISTRY OF U02C03

Preparation of U02C03 Slurries(J. P. McBride, W. L. Pattison).Interest in the possible use of U02C03as a slurry fuel has been increased

because, as reported in the previousreport, the reaction of C02 with aslurry of oxide platelets yielded acarbonate that exhibited excellentslurry properties. Investigation ofthis reaction has continued in aneffort to determine the conditions oftemperature and pressure which producethe most desirable product. Reactionsat room temperature and C02 pressuresof 225 to 840 psi were found to yieldmuch better slurries than were obtainedat higher temperatures (200 to 250°C).The addition of 0.005 MNa2C03 to theoxide before reaction at room temperature served to essentially stabilizethe resulting slurry against sedimentation so that only a small fractionof the solid separated in one week.Slurries prepared in the absence ofadded electrolyte, although not freeof sedimentation, showed littlesettling in 24 hr and complete dispersion of the unsedimented materialin the slurry phase in contact withthe separated solid.

Table 34 summarizes the data on thepreparation of U02C03. It is apparentthat the best reaction is obtainedwith very pure oxide platelets of lowsoluble uranium content.

Hydrolytic Decomposition of U02C03Slurries (J. P. McBride, W. L. Pattison).The effects of temperature and ofadding dilute solutions of electrolyteson the hydrolytic decomposition ofU02C03 slurries has been studied.

Effect of Temperature. Although nodetailed study of the effect of temperature on the hydrolytic decompositionof U02C03 slurries has been made, ithas been observed that they are hydro-lized by prolonged boiling, but theprocess is slow. At 250°C, however,in a closed system and in the absenceof C02 pressure, a carbonate slurry isdecomposed almost completely to oxidein 2 to 3 hr (Table 35).

Effect of Electrolytes. Thehydrolytic decomposition of a U02C03slurry at 250°C generally yields anoxide in the form of large rods that


are 10 to 20 ft in length. The factthat the original carbonate is prepared in the form of elongated plateletsthat are much less than 1 fx in averagesize makes the relatively large oxideparticles produced by the decompositiondisappointing. It was of interest,therefore, to investigate the effectof adding dilute solutions of electrolytes to the carbonate slurriesbefore hydrolysis in the hope thatcrystal growth might be inhibited bythe adsorption of the ions on theoxide as it is formed.

The effects of Na2C03, NaOH, Na2Si03,Na3P04, Na4V207, and H3P04 were investigated. Table 35 gives the datafor these experiments. The firstthree promo.ted the growth of largerods and platelets. Sodium phosphateinduced the formation of small, thin,uniform platelets, whereas the vanadateyielded small agglomerates (5 to 10 /j.)composed of small particles of approximately the same size as that of theoriginal carbonate particles. Theeffect of phosphoric acid was similarto that of the sodium salt. In allcases, the presence of the electrolyteapparently caused a rather high dissolution of uranium (500 to 3000 ppm).Since some of the supernatants losttheir yellow color upon being allowedto stand, because of the separationof the yellow oxide, the uranium inthe supernatants was probably presentin a colloidal state.

Thermal Stability of U02C03 SlurriesUnder COQ Pressure. The chemicalequilibriums associated with theformation of U02C03 by the reaction ofC02 with U03#H20 may be expressed asfollows:

*iH20+ C02^=^C02(aq) + H2C03 , I

Kr

H2C03 ^= H+ + hco: II

K.

U02(0H)2 + 2H+: UO++ + 2H20 III

115

EXPERIMENT

NO.

H42

R5,

R17,

STARTING MATERIAL

UO3.H2O platelets preparedfrom rods by 3 cycles;calcination at 400°C,hydration at 2 50 C

Dilute portion of slurryproduced in R4

Portion of slurry producedin R4,

UO, •H.O platelets (same as

Anhydrous U03 producedfrom U04 calcination:3 hr at 325°C

Enriched U03-H20 plateletsprepared from rods bycalcination-hydra tion

U03'H20 platelets (same asR4!>

U03-H20 platelets (same asR4,)

UOj-hÔ platelets preparedfrom rods by 5 cycles;calcination at 400 C,hydration at 2 50 C

UO3-H20 platelets (same asR9,)

UOj-HjO platelets (same asR9,)

U03-H20 platelets (same asR9;)

U03'H20 rods; preparedfrom pure, micronizedplatelets by hydrationat <200°C

UOj-HjO platelets (50 to100 ppm of solubleuranium)

U03-H20 platelets similarto R15J

U0..H20 rods prepared atY-12; high concentrationof soluble uranium

TABLE 34.

ADDED

ELECTROLYTE

None

0. 004 Jf NajCOj

0. 004 M Na2C03

0.005 MNa2C03

0.005 *f Na2C03

0.005 * Na2C03

0.01 M Na2C03

None

0.005 M Na2003

None

None

None

REACTION

TEMPERATURE

CO

200

250

250

250

250

250

250

25

25

PREPARATION OF UOjCOj

REACTION

TIME

(hr)

48

18

18

18

48

48

40

23

C02 PRESSURE, ATROOM TEMPERATURE*

(psi )

950

500

500

500

500

500

374 psi addedat 250°C

225

400

URANIUM

CONCENTRATION

(g/liter)

121

163

135

224

♦In two experiments with similar gas and solution volumes and C02 pressures, the partial presslurry was found to be 1.4 and 1.5 times the pressure added at room temperature.

CO5--TO-U RATIO

0.97

0. 99

0. 96

0.95

DESCRIPTION OF

RESULTING SLURRY

Pale yellow; large agglomerates; rapid settling; nopacking

Smaller particles than R4^;lower bulk density

Smaller particles than R4jand R42; low settling rate

Almost white slurry with verysmall, uniform, regularlyshaped particles; averageparticle size 0.6 MI 25%settling in 17 hr

Approximately 40% reaction

Similar to R5.

Similar to R5j

High soluble uranium in supernatant, slurry settled slowlybut packed badly

Large yellow particles whichsettled rapidly and packedbadly; even upon beingallowed to stand for only ashort time

Viscous slurry; low bulkdensity; water-white super-na tan t

Very white, slow-settlingslurry; some tendency toagglomerate and becomethixotropic in 24 hr

Excellent slurry; virtuallystable; less than 5% settlingof solids in one week

Slow-settling slurry

Very little settling in 24 hr

Similar to R15j

Incomplete reaction

sure of CO2 at 250°C over a carbonate

as93"0

oG

w90r

§o133rnOiin

90PI"0o90H

TABLE 35. HYDROLYTIC DECOMPOSITION OF UO CO, SLURRIES AT 250°C; EFFECT OF ELECTROLYTES

EXPERI

MENT

NO.

SLURRY

PREPARATION

(see Table 34)

REACTION

TIME

(hr)

ELECTROLYTE CONDITION OF OXIDE

FINAL

URANIUM

CONCENTRATION

IN SUPERNATANT

(ppm )

R56 R5X 20 0.005 M

Na2C03Needles and rods with good slurry

properties500 to 1000

R102 RlOj^ (washed andres lurried)

18 0.005 M

Na2Si03Large rods; considerable agglomeration and packing

3 90

R103 RIO, (washed andres lurried)

18 0.01 M

NaOH

Large rods and fragments; some agglomeration and slight tendencyto pack

540

R104 R101 (washed andres lurried )

18 None Agglomerated and caked rods 274

R105 RIO, (washed andreslurried)

18 0. 004 M

Na2C03Not completely decomposed to rods;

difficult to slurry1080

R10fi RIO, (washed and

reslurried)

18 0.005 M

Na3P04Small, thin, square platelets of

uniform sizes; very easily re-slurried

3270

R107 RIOj (washed andreslurried)

18 0.005 M

Na4V207Very fine oxide resembling theoriginal carbonate; easily re-slurried

1430

R108 RIO, (washed and

reslurried)

64 0.005 M

H3P04Thin square platelets; some

tendency to pack900

R109 R10x (boiledfor 3 hr)

68 None Large platelets; rapid settling 144

Rll2 Rl^ 20 None Rods; nearly complete decomposition 300

RII3 Recarbonated Rll2 48 None Same as Rll, 189

RH4 Recarbonated Rll3 2.5 None Same as Rll2 111

R142 R14x 2.5 None Rods 110

13W50M

Oa

cn

a

a

90o93

so

VICO


K.

uo;+ + hco: uo2co3 + H+ IV

where the K. are the equilibriumconstants.

Since KtK2 is equal to 3 x 10"7 andthe fraction of the uranium convertedto oxide is proportional to the ratio(HC03)/(H ), it is apparent that therate and completion of the reactionwill depend not only on the CO, pressureand the temperature but also on the pHof the original oxide slurry and theconcentration of acid impuritiesreleased from the oxide on reaction^

Experiments indicate that thepartial pressure of C02 required tostabilize pure U02C03 in a slurry at250°C and in the presence of 0.005 NNa2C03 will be less than 500 psi (seeexperiments R91( Table 34; R57 andR12s, Table 36). Data on the thermalstability of U02C03 slurries are presented in Table 36.

THORIUM CHEMISTRY

J. 0. Blomeke L. E. Morse

J. P. McBride W. L. Pattison

A study of thorium compounds suitablefor use in a breeder blanket has beeninitiated. This work has been dividedinto two parts:1. the preparation of pure Th02 and a

study of its slurry properties,2. an investigation of other thorium

compounds and blanket systems whichwould be suitable for use in abreeder.

URANYL OXIDE SLURRY IRRADIATIONS

J. P. McBride W. L. Pattison

The investigation of the effect ofneutron irradiation on the propertiesof uranium oxide slurries has continuedwith another experiment carried outin hole 11 of the ORNL graphitereactor. A slurry of 93% enrichedU03#H20 platelets containing 40 g ofuranium per liter was irradiated in

118

type 347 stainless steel for fourweeks (600 hr of irradiation) atmaximum flux (7 *1011 neutrons/cm2•sec)and at temperatures of 250 to 295°C.

The gas production rate was muchless than that in a sulfate solutionof the same concentration, the valuefor G (molecules of water decomposedper 100 ev of energy absorbed) beingabout 0.4 at 250°C. The calculatedsteady-state pressures during thefirst four days of the run were muchhigher than those in a solution thatindicated a much slower rate of re

combination of the hydrogen and oxygenproduced by the decomposition ofwater. The steady-state gas pressurewas found to decrease, however, as theirradiation progressed; therefore, atthe end of the run, steady-statepressures (including steam) of lessthan 4000 psi at 250°C and less than2500 psi at 295°C were observed. Withthe exception of the values obtainedduring the first three days of irradiation, the gas production and recombination at any one temperaturecould be correlated by using thekinetic expression already developedfor solutions; that is, a plot ofAP/At vs. P yielded a straight line.(6)The values for the gas production ratefound by extrapolating the AP/At plotto steam pressure were found to de

crease by at least a factor of 2 asthe slurry temperature was increasedfrom 250 to 295°C.

The experiment was carried out withonly 10 ml of slurry, or a total of400 mg of uranium, as the oxide. Only30% of the slurry was readily transferred from the radiation bomb. Visualexamination of the bomb interiorshowed the bulk of the material tohave plated out on the walls of thecontainer. It was possible to recoverthis material by adding 1/4-in. steelballs to the bomb, closing it, andshaking it vigorously for several

C. H. Secoy, HRP Quar. Prog. Rep. Mar. 15,1952, ORNL-1280, p. 164.

TABLE 36. STABILITY OF UO,CO, SLURRIES AT 250°C UNDER CO, PRESSURE2""3

EXPERI

MENT

NO.

SLURRY

PREPARATION

(see Table 34)

ELECTROLYTE

CONCENTRATION

C02 PRESSURE,AT ROOM

TEMPERATURE*

(psi )

REACTION

TIME

(hr)CONDITION OF SLURRY

FINAL

pH(degassed)

R57 R5j 0.005 M

Na2C03150 168 No apparent decomposition to oxide;

complete settling within 24 hr

R122 R12j 0.005 M

Na2C03None 64 Went to partial dryness; nearly com

plete decomposition to large rods

(about 20 /i)

6.4

R123 R12j 0.005 M

Na2C0348 64 Same as R12, 6.7

R124 R12j 0.005 M

Na2C03100 88 More than 50% decomposition to rods;

remaining carbonate well dispersed

6.6

R125 R12x 0.005 M 300 88 No apparent decomposition or particle

growth; complete settling within

24 hr

7.4

R162 R16j None 225 164 Decomposed 5.2

R163 R16x None 390 48 Decomposed 6.1

*In two experiments with similar gas and solution volumes and CO2 pressures, the partial pressure of CO2 at 250 C over• carbonate slurry was found to be 1.4 and 1.5 times the pressure added at room temperature.

wsoI—I

oD

M2!DHH

o

3>SOoB

CM

V0

V\


hours. Plating appeared to be muchmore pronounced above the surface ofthe slurry than below, but no quantitative determination was made.

The portion of the slurry that wasreadily transferred showed no impairment of slurry properties. Photomicrographs showed that irradiatedoxide had undergone considerabledegradation to fragments and fines.Only a very small fraction of thematerial appeared in the form ofrecognizable platelets, which comprisedthe bulk of the original material(less than 10 /x average size). In

addition, there were many largeflakes (20 to 40/x in size) of irregularshape; the flakes exhibited no definitecrystalline form. Since the solubleuranium content at the end of the runwas less than 8 ppm, crystal growthshould not have occurred. The flakeswere probably loosened portions of thefilm formed on the interior surfacesof the bomb by the plating out describedabove. In contrast to the slurriespreviously irradiated, the recoveredslurry was brownish in color ratherthan dark green and no reduced uraniumwas found.

PHYSICAL STUDIES OF SLURRIES

R. N. Lyon, Section ChiefA. S. Kitzes P. R. CrowleyR. B. Gallaher W. Q. Hullings

C. A. Gifford

THORIUM SLURRIES

Thorium Oxide Studies. Thorium

oxide prepared by calcination ofthorium oxalate at temperatures above600°C is extremely abrasive to stainless steel. The reported hardnessof this material is approximately 7.5on the Moh scale, while the hardnessesof stainless steels and titanium on

the same scale are 6 to 6.5. Thesedata conform with the results obtained

with the abrasion tester described inORNL-1121,(1) in which the abrasiveaction of thorium oxide was qualitatively determined by pumping slurrythrough a 0.008-in. gap between therim of a rotating wheel and a polishedtest-plate parallel to the axle of thewhee1.

Studies of the possibility ofproducing a thorium oxide with ascratch hardness of less than 6 onthe Moh scale have been initiated. Inthese studies, thorium oxalate has

' 'A. S. Kitzes et al., HRP Quar. Prog. Rep.Aug. 15, 1951. ORNL-1121, p. 156-160.

120

,.«mmm9wm*

been calcined at various temperaturesbetween 300 and 500°C, and the indications, so far, are that oxideprepared by calcining the oxalate at300°Cis less abrasive than the corres

ponding material calcined at 800 C.Calcining at the lower temperaturesproduces a more chemically reactiveoxide which is expected, because ofits possibly porous structure, to beless hard than an unreactive form.

These tests are being continued toobtain more quantitative data.

The particle shapes and sizes weredetermined from electron-microscopephotomicrographs of three types ofthorium oxide. The data, shown in

Table 37, were abstracted from ORNLCF-53-l-240.(2)

Thorium Fluoride Studies. Thorium

fluoride is another possible materialfor use as a circulating-type ofthorium blanket. Thorium fluoride,obtained from Ames Laboratory in a

(2) F. D. McNeer, Electron Micrographs of Th02Samples, ORNL CF-53-1-240 (Jan. 22, 1953).


TABLE 37. ELECTRON MICROGRAPHS OF Th02

PREPARATION OF Th02 PARTICLE SHAPEPARTICLE SIZE (LL)

Aver age Ran ge

High-temperature calcination at Irregular 0.2 0. 02 to 2

800°C from oxalate

Low-temperature calcination at Plate-like 0.5 0. 02 to 2

300°C from oxalate crystals

Low-temperature calcination at Irregular 0.2 0. 02 to 2

300°C from carbonate

freshly prepared slurry, abraded type347 stainless steel in the abrasion

tester at 12 times the rate for

abrasion by oxide calcined at lowtemperature. However, thorium fluoridewhich had been allowed to remain under

water for seven to ten days prior totesting abraded the stainless steelspecimen only slightly in 24 hr andpossible hydration of the fluorideto produce a softer material wasindicated. Therefore thorium fluoride

prepared by precipitation from Th(N03)4with HF may be nonabrasive, since itwill be fully hydrated. No tests havebeen performed with this material, asyet.

Slurries of thorium fluoride have

one disadvantage which may limit theiruse as a breeder blanket. Because ofthe low bulk density of the fluoride,it may not be possible to obtain ashigh a thorium concentration as isdesired for some reactor configurations.

Thorium Carbonate. A slurry ofthorium carbonate, when autoclavedat 250°C for 24 hr, decomposed to giveC02 and ThO 2. Consequently, thismaterial is not being consideredfurther as a breeding material foruse at 250°C.

Thorium Hydroxide Peptized with

Thorium Nitrate. To alleviate theabrasion problem associated with mostthorium compounds, colloidal thoriumhydroxide was prepared by precipitationfrom Th(N03)4 with concentrated

ammonium hydroxide. Normally, themetal hydroxides, if stable, aresofter than the corresponding oxides.In addition, abrasion would probablybe minimized because of the extremelyfine particles in the colloidal solution. To produce the thorium hydroxidesuspension, a slurry was peptizedwith thorium nitrate at 65 C to givea stable mixture that on visual examination was hardly distinguishablefrom a solution. Prior to peptizationwith thorium nitrate, the thoriumhydroxide was washed by decantationto remove most of the ammonium saltsand then e 1 ec t ro di aly zed until essentially free of all extraneous ions.

The stable suspension, the composition of which is still unknown,was concentrated by evaporating waterat 65°C until a product that was veryviscous at room temperature was obtained. The density of this fluidwas approximately 3.4 g/cm , and itcontained about 2300 g of thoriumper liter, calculated as Th02. Additional quantities of this materialare being prepared so that the compound can be identified, the thermalstability at 250°C can be ascertained,and the stability to reactor radiationcan be studied.

Thorium Oxide Circulating Loops.

During this quarter, the loop forcirculating thorium oxide slurriesat a temperature of 150°C was put intooperation. The loop described in

121


ORNL-1121(1) was modified by substituting a Richa rdson-Fri thsen, fluid-bearing, canned-rotor pump for theoriginal, conventional, centrifugalpump. The pump, previously testedfor its operating characteristics withslurries, circulated a thorium oxideslurry continuously and with no apparent difficulty in an auxiliarytesting loop at room temperature for168 hr before it was installed in theloop. Within 24 hr of operation at150°C with a slurry containing 300 gof thorium (as the oxide) per liter,the pump failed. Examination of thedisassembled pump showed that thestator cans of 0. 010-in .-thick , stainless steel had been perforated becauseof abrasion by the oxide. The totaloperating time of the pump with thoriumoxide slurries in various runs was264 hours.

Modified stator cans are beingdesigned to prevent accumulation ofsolids in stagnant corners of thecans. This type of accumulation was asecondary problem in previous tests,since the pump shaft was vertical.The thorium circulating systems willbe reserved in the future for slurriesof particles less hard than stainlesss teel.

Design of the thorium loop foroperation at 250°C and 1000 psi hasbeen temporarily suspended until asuitable slurry material is developed.The over-all design for the loopdescribed in ORNL-1478(3) has beencompleted, and some detailed drawingsare available.

URANIUM SLURRY LOOPS

Loop S. LoopS has been in operationat 250°C and 1000 psi with uraniumoxide slurries for over 2500 hr with

no evidence of corrosion or erosionin the loop. However, solids stilltend to accumulate in the pressurizer,but they no longer plate out on the

(3 )'A. S. Kitzes et al.. HRP Quar. Prog. Rep.

Jan. 1. 1953. ORNL-1478, p. 105.

122

walls of the main stream pipes.Various attempts were made to eliminate the accumulation of solids in the

pressurizer by changing the level ofliquid in the pressurizer and byintroducing auxiliary flow through thepressurizer to create turbulence, butnone of the attempts were successful.Electron micrographs and x-ray diffraction of the material collected in

the pressurizer always showed theoxide to be of the beta or the plateletform.

Alpha-to-Beta Transformation. Toeliminate the possibility that thetransformation of the alpha oxide(rods) to the beta (platelets) formwas the cause for the accumulation ofsolids in the pressurizer, beta oxideobtained from previous runs was circulated, initially, at 250°C. Within24 hr, the solids collected in thepressurizer. The temperature of theloop was then dropped to 150°C, sinceprevious experience with anotherloop( ' had shown that alpha oxidecould be circulated at 150°C withoutdifficulty. Within 24 hr after theloop reached equilibrium at 150°C, thebeta oxide was being converted to thealpha form, as shown by x-ray diffraction patterns.

By cycling the loop temperaturebetween 250 to 150°C and then to 200°C,the temperature dependence o f the oxideform was established. At temperaturesabove 200 C, the beta oxide is apparently the stable form, and below200 C, the alpha oxide is more stable.Results of subsequent runs proved thatthe alpha oxide is the more stable andthat it can be circulated at 150°C for

long periods of time without accumulation of solids in any part of theloop or pressurizer.

Itwasalso found that accumulations

of beta or platelet form which occurat 250 C can be redispersed by lowering

'R. N. Lyon et al., HRP Quar. Prog. Rep.Mar. 15. 1952, ORNL-1280, p. 83.

swtimmw-* •i«io#«fc»a*i«***>™*


the temperature to 150°C, if the ac- accumulate in the pressurizer readilycumulations have not existed for more dispersed the solids and prevented thethan about 24 hours. Older accumu- plating out of the slurry circulatinglations are dispersed slowly, if at at 250°C. The pH of the slurry wasall, by lowering the temperature. 10.6.These findings reinforce the belief Phosphates are common and efficientthat alpha, or rod-type, slurries do de fl occul at ing agents for the dis-not plate out in a ci rcula ting system. persion of solids in water. TheirThe investigation of the reason for mode of action is probably the follow-accumulation of the beta oxide (plate- ing: adsorption of the phosphate ionslet) form at 250°C is being continued. on the surface of the solids imparts a

U02C03 Slurries. McBride has negative charge that causes repulsionreported(s> that U02C03 slurries are of similar particles and therebythermally stable at 250°C. Interest stabilizes the system.in circulating this material at 250°C With the addition of oxygen tois warranted, not only because of the prevent reduction of U+6 to U+4, afterunderstanding it may give of the 100 hr of operation with uniform con-behavior of rod-like particles ci r- centration of solids throughout theculating at this temperature, but loop, evidence of corrosion was found,also because the U02C03 itself may be The slurry particles turned muddyuseful as a slurry fuel. brown in color, and thus there was

The carbonate was prepared by cir- indication of the formation of Fe203;culating beta oxide in the loop at since the supernate became yellow, there150°C under a pressure of 600 to was possibly soluble chromate present.800 psi of C02 for 65 hours. After Analysis of the slurry confirmedbleeding off some of the gas, the the presence of iron and chromium,loop was heated to 250°C and, within Large quantities of iron compounds24hr after equi li brium was es tabl ished, were found with the solids, and chromatesolids accumulated in the pressurizer. ions, rather than uranium compounds,Microscopic examination of these were the cause for the yellow color ofsolids showed them to be predominantly the supernate.of platelet form. Apparently, the This run was stopped and the systemU02C03 decomposed to U03*H20, which was inspected for corrosion damage,then promptly took the beta form and Most of the system was coated with aplated out in the pressurizer. McBride heavy, brown layer resembling Fe203.discusses this reaction in ORNL-1478(55 This layer dissolved only partially inand reports evidence that 0.005 M nitric acid and left a fine, reddishNa2C03 stabilizes U02C03 slurries at precipitate which could be washed250°C. A run is now in progress in from the system. In three parts ofwhich aU02C03 slurry is being prepared the system, smooth, shiny, base metalin the presence of 0.02 M Na2C03. At was exposed. These parts were the150°C, thepH of the slurry is steadily 3/4-in. orifice plate, part of theincreasing from 6. 4 to 7. 0 and a p- pump casing, and the pump impeller,proaching the value of 7.4 reported all of which were exposed to high,by McBride. No other data are availa- surface, shear stresses during oper-ble, as yet. ation. The pump impeller was found to

The System UO 3'H fl-Na flPO ^ The have lost 16 g in weight compared withaddition of 0.5 MNa2HP04 to the loop its weight when originally installed;in which solids were beginning to however, in previous runs, it had

gained about 5 g, presumably in(5)_ . „. . D . building up a thin oxide scale. Thus* F. R. Bruce et al.. HRP Quar. Prog. Rep. Jan. •. • , i j • i_

1. 1953. ORNL-1478, p. 101. the total weight loss during the

123


phosphate run was 21 g, of which 5 gwas probably represented by oxygen inbase metal oxide. Since the atomic

weights of iron, nickel, and chromiumare roughly three and one half timesthat of oxygen, the major portion ofthe weight loss can be accounted foras loss of metal by conversion tooxide before the phosphate run began.

Further investigation of thephosphate system is planned to determinewhether the base metal is attacked

seriously in this medium. In addition,a run will be made with monobasic

sodium phosphate.Other Proposed Additions. Other

additions, such as ferric nitrate,nickel nitrate, and possibly somealuminum compound, will be added touranium slurries to determine the

influence of these additives on the

stabilization of a particular form ofoxide and on the plating-out problem.

In an effort to study the transformation of the various crystalhabits of U03'H20 (alpha and beta) andpossibly to discover the mechanism forthose transformations, beta U03'H20(platelets) was statically autoclavedat 150, 200, and 250°C for as long as168 hours. Microscopic examination ofsamples removed at various times

showed particle growth but no transformation from the beta form to thealpha allomorph. J. 0. Blomeke of theChemical Technology Group, however,confirmed the dynamic test resultswhen he micronized the beta oxide priorto autoclaving at 150, 175, 225, and250°C for 24 hr and found that below200°C the beta form converted to alphaand that no conversion occurred at

temperatures above 200°C.The mechanism and temperature

dependence of this transformation arestill unknown. Various additives,such as Fe+++, Ni++, P04~", are beingadded to calcined U03 prior to hydrationto UO *H20 in an effort to influencethe production of one allomorph inpreference to the other. Preliminaryresults indicate that Ni favors the

124

production of alpha, whereas the betaform is enhanced by the presence ofFe+++.

Loop T. Loop T for circulatinguranium slurries at 250°C and 1000 psiwill be identical to loop S, exceptthat the pressurizer will not be anintegral part of the loop but, rather,it will be isolated and connected to

the circulating stream only by meansof an S-shaped stainless steel tube tominimize formation of a vortex and the

accumulation of solids. The loop isnow being fabricated and assembled,and it should be ready for a shakedown test by May 1953.

In this system, the effect ofsystem design on slurry stability willbe studied at 250°C. Also, the presentstudies on the effect of additives on

the chemical and physical stability ofslurries will be continued.

Glass Loop. The glass loop hasbeen modified to simulate the designof loop T, and it has been operated atabove atmospheric pressures to aid inanticipating the effect of the isolatedpressurizer on the operation of therest of the loop. The loop has beenintermittently operated for 50 hr withno accumulation of solids in the

pressurizer and with no evidence of avortex. On the basis of these results,no difficulty is expected with theisolated pressurizer for loop T.

CRITICAL EXPERIMENTS

The slurry critical experimentsconducted in cooperation with the ORNLCriticality Group should be completedearly in May. For these experiments,an aluminum tank with an agitator anda water reflector is fed from a

continuously circulating loop thatcontains the enriched slurry. Whenenough slurry has been added to makethe system almost critical, the feedflow is stopped and the agitator isslowed. For a 12-in. reactor and for

a constant reflector height, theresults, to date, indicate that for

all concentrations the reactivity increases rapidly to a maximum when theagitator is slowed from any speedabove 200 rpm to 200 rpm. The suddenincrease in reactivity at 200 rpm isprobably due to the peculiar particledistribution at this speed and alsoto a finite reflector layer beingformed on top of the charge as thesolids begin to settle out. In veryfew cases was it necessary to scramthe system because of the sudden in


crease of reactivity at 200 rpm. Atall agitator speeds above 200 rpm, thesystem was unusually stable andpredictable.

At the completion of the tests, thedata will be compared with predictionsbased on calculations for the system

by P. R. Kasten.(6)

(6). Kasten, Prediction of CriticalityBehavior ofaSlurry Upon Se11ling. ORNL CF- 53-1- 294(Jan. 26, 1953).

CHEMICAL PROCESSING

F. R. Bruce, Section ChiefD. E. Ferguson W. E. TomlinR. G. Mansfield W. B. Howerton

0. K. Tallent

FUEL PROCESSING

Chemical processing of the innercore of a thermal breeder is requiredbecause it is necessary to remove

fission products and corrosion productswhich adversely affect the neutroneconomy of the reactor. In addition,continuous extraction of the fissionproducts to maintain the containedradioactivity at a relatively lowlevel is desirable as a means ofdecreasing the possibility of a reactorcatastrophe. The requirements forthis type of fuel processing differwidely from those for the present,heterogeneous reactor fuel. For theaqueous homogeneous reactor, largedecontamination factors for fissionproducts are not necessary, becausethe fuel is returned to the reactoras a solution. Consideration mustbe given only to removing a largefraction of those fission and corrosionproducts which contribute to reactorpoison. A major factor in all developmental work on fuel processing is thatstrong emphasis must be placed oneconomy if the reactor is to producepower at a cost comparable to presentpower costs.

Fission-Product Poisons. In discussing the fission-product poisons ofa reactor, it is convenient to dividethe fission products into threeclasses: the rare gases, the nongaseousfission products with very high crosssections (Sm149, Sm151, Gd*57 , andEu*55 ) and combined yield of 0.82%for U233 fission, and the remainingnongaseous fission products. It isexpected that the rare gases will beremoved from an aqueous solution fuelin the gas separator and that theircontribution to reactor poison will besmall. Therefore, in the chemicalprocessing, only the nongaseous fission products and the corrosionproducts must be considered.

Figure 53 shows the effect of thechemical cycle on the fission-productpoisons in the ISHR if there is continuous processing of the fuel solution. It can be seen that chemicalprocessing cycles of more than tendays will not significantly lower theamount of poison from the very highcross-section fission products.

To obtain the data presented inFig. 53, a cross section of 40 barnswas used for the "other" fission

12 5


100 DWG 179I4A

BASIS: Sm'49, Sm15', etc ; YIELD 0.8%; a =40,000bc

OTHER NONGASEOUS FISSION PRODUCTS; YIELD 160%; <£ =40b~

10 2 5 100

CHEMICAL PROCESSING PERIOD (days)1000

Fig. 53. Poisoning of ISHR When Operated as a Two-Core Thermal Breeder.

products of U233. This value isprobably pessimistic when the veryhigh cross-section fission productsare excluded.

Corrosion Product Poisons. Based

on a generalized corrosion rate of1 mpy, the poison from nickel andmanganese in the fuel solution willbuild up at the rate of 3% per year(Fig. 53). The rate of dissolution of

126

mmmmmmmmm

type 347 stainless steel from the ISHRwill be 27 g per liter of fuel solutionper year. Of this amount, about 2.7 gof Ni and 0.53 g of Mn per liter ofsolution will remain dissolved. Ironand chromium can be neglected becauseof their low solubilities and lowcross sections.

Distribution of Fission-Product

Poison. A summary of the contributions

nijimipimiwwiMi

to reactor poison of the various fission products, by chemical group, ispresented in Table 38. These valuesare based on the ISHR and include allfission products for which the crosssections are known. (**

It can be seen from Table 38 that

the rare earths contribute about 80%

of the nongaseous fission -productpoison. For short irradiation periods,only the halogens, primarily I131,contribute significant amounts ofpoison in addition to that from therare earths; however, owing to theshort half-life of I131, its concentration does not continue to build up,and for irradiation periods of 100 daysor longer, molybdenum, cesium, and thenoble metals are the most significantof the poisons, except for the rareearths. From these observations,it is possible to list qualitativelythe requirements for processing of thefuel solution of a thermal breeder.

The requirements, in the order oftheir importance, are:1. to remove rare earths rapidly,2. to remove cesium, molybdenum, and

the noble metals,3. to remove the corrosion products,

nickel and manganese,4. to remove the halogens.

'K. Way, L. Fano, M. R. Scott, and K. Thew,Nuclear Data, NBS Circular 499 and Suppl. 1, 2,and 3 (Sept. 1, 1950).


Methods of Processing. Two generaltypes of processing have been considered for the fuel solution of a

thermal breeder: complete decontamination of the fuel solution by solventextraction and partial decontaminationof the fuel solution by adsorption.As stated above, complete decontamination of the fuel is not essential.

The removal of 75% of the rare earths

and 25% of the cesium, molybdenum,noble metals, and corrosion productsis sufficient decontamination if the

fuel solution, after processing, issuitable for direct return to the

reactor.

Complete decontamination of thefuel by solvent extraction would beessentially the same as the presentU23S recovery process, with metaldissolution replaced by heavy-waterrecovery. The decontaminated uranylnitrate product from the process wouldhave to be converted to a heavy watersolution of uranyl sulfate for returnto the reactor. The cost of this

process would be at least $5.00 pergram of uranium if the processingwere done at the Idaho Chemical Proces

sing Plant, provided it was operatingat full capacity.

An adsorption process could becarried out by using an organic ion-exchange medium, an insoluble inorganicadsorbent, such as CaF2 (suggested

TABLE 38. DISTRIBUTION OF FISSION-PRODUCT POISON IN ISHR FOR

U233 FISSION PERIOD OF OPERATION AS INDICATED

CHEMICAL GROUPPER CENT OF TOTAL FISSION-PRODUCT CROSS SECTION

20 Days 50 Days 100 Days 300 Days 900 Days

Rare earths 82 81 80 76 67

Halogens 11 10 7 4 3

Noble metals 2 3 4 6 8

Cesium 2 4 6 8 13

Molybdenum 1 1 3 5 7

Cadmium 2 1 1 < 1 < 1

127


by Vitro Corp.) or BaSO., or a combination of adsorbents. A cost of

20 cents per gram of uranium processedseems to be a pessimistic estimatefor the simple adsorption processenvisioned hers. ' 2' However, theuranium adsorbed along with the fission products would have to be recovered at a cost of at least $5.00

per gram. This would increase thecost of an adsorption process to 45cents per gram of uranium if 5% ofthe uranium is retained on the adsorbent

and to 25 cents per gram if 1% of theuranium is adsorbed along with thefission products.

Optimum Processing Cycles. Since the

main product of this thermal breederis power, economy must be a primeobjective in the design of the reactorsystem, if power is to be produced ata competitive cost. From the standpoint of chemical processing, thisrequires minimizing the cost ofprocessing and the losses of fissionablematerial both in processing and byreactor poisons. The results of ananalysis of these requirements, basedon the processing of the ISHR fuelsolution, are summarized in Table 39.

233A value of $100 per gram for U ind

the chemical processing cost given inthe preceding section were used.

These results show that a processcosting $5.00 per gram of uranium isimpractical for a thermal breeder.

(2)D. E. Ferguson, Fuel Processing of a Homo

geneous Thermal Breeder, ORNL CF-53-1-324 (Jan. 30,1953).

Both the cost of processing and thelosses from poisons are unreasonablewith the optimum process cycle. Theresults also demonstrate that a processing cost of 45 cents per gram ofuranium is reasonable. The cost of

processing would then amount to 0.5mil per kwh of power produced, andU233 production lost because of poisonwould be about 3%. If the U233adsorbed along with the fission productsis only 1% of the uranium processed,chemical processing costs would dropto 0.3 mil per kwh, but the optimumprocess cycle and the poison remainthe same as for 5% uranium retention.

Adsorption of Fission Products on

Calcium Fluoride. Many adsorbentmaterials have been tested for removal

of fission products from the fuelsolution of an aqueous, homogeneous,thermal breeder. Of the materials

tested by Vitro Corp., calcium fluoridewas recommended as the most suitable

for this application.(3^ Additionalexperiments at ORNL have shown thatfinely divided CaF has a remarkablecapacity for removing rare earths fromdilute uranyl sulfate solutions; however, it was found that the mechanismof this removal is not adsorption.It appears to be metathesis of thecalcium fluoride which yields insolublerare earth fluorides and soluble

calcium sulfate. Since the solu

bility of calcium sulfate in dilute

( 3 )January Progress Report for Homogeneous

Reactor Process ing. KLX-1609 (Feb. 10, 1953).

TABLE 39. COMPARISON OF METHODS OF FUEL PROCESSING FOR THE ISHR

PROCESSING METHODCOST OF

PROCESSING

($/g of U)

OPTIMUM

CHEMICAL

PROCESSING

CYCLE

(days )

POISON FROM

FISSION AND

CORROSION

PRODUCTS

(%)

COST OF

PROCESSING

(mils per kwh)

Complete decontaminationby solvent extraction

Adsorption (5% U holdup)

Adsorption (1% U holdup)

5

0.45

0.25

70

50

50

4

3

3

5

0.5

0.3

128

uranyl sulfate solution at elevatedtemperatures is not known, it is notyet possible to evaluate the feasibility of using this material forremoving rare earths from the fuelsolution of a thermal breeder. Ifuranyl fluoride solution were usedas fuel for the reactor, this difficulty would be eliminated and the CaFwould serve as an inert filter medium.

Even though the absolute capacityof CaF2 for removing rare earths fromthe reactor fuel solution is undoubtedlythe stoichiometric amount of rareearths, the practical capacity undercontinuous contact conditions in afixed bed was found to depend ontemperature. At 25°C, 1.5 liters ofCaF2 per day would be required forremoving rare earths from the ISHR;at 75°C only 0.75 liter of CaF2 wouldbe required.

The solubility of calcium fluoridein 0.02 M U02S04 solution was found tobe 0.07 mg/ml at 250°C. At roomtemperature the solubility was 0.33mg of CaF2per milliliter. This shouldnot contribute a significant amountof calcium or fluoride to the fuelsolut i on .

Other Inorganic Adsorbents. Several

inorganic adsorbents other than calciumfluoride were found to be effectivein removing fission products fromdilute uranyl sulfate solution. Inbatch tests at room temperature, thefollowing results were obtained bycontacting the adsorbents with 0.02 MU02S04 solution containing mixedfission products:

ADSORBENT

CaF2De Calso

Zeo Dur

Manganese Zeolite

BaS04 (freshly precipitated)

PER CENT OF BETAACTIVITY ADSORBED

77

96

79

72

88

These adsorbent materials will be

investigated further if CaF2 is foundto contribute excessive calcium to the

fuel solution.


THORIUM OXIDE PELLET STUDIES

Autoclave tests indicated thatthorium oxide pellets which had beenfired at 1700°C were stable for aslong as 150 hr in water at 250°C andfor as long as 120 hr in 0.02 MU02S04solution at 250°C. No dissolvedthorium could be detected in eitherthe water or the uranyl sulfate solution at the end of the test period.

The batch of thorium oxide pelletsselected for testing was preparedby the Norton Co. The pellets wereabout 1 cm in diameter, and they hadan average specific gravity of 8.8and a porosity of ],ess than 0.1%. Twogeneral types of pellets could bedistinguished by visual observation:smooth-surfaced, white pellets andtan, rough-surfaced pellets. Spectrographs analysis showed no significantdifferences in the chemical compositions of the two types. The tanpellets were found to be subject toabrasion and chipping when agitatedagainst one another in water.

Autoclave Tests of Thorium Oxide

Pellets. The result of heating thoriumoxide pellets in water or uranylsulfate solution was determined bysealing white or tan pellets insidepyrex tubes with 10 ml of distilledwater or 0.02 M U02S04 solution andheating the tubes in a rocking autoclave at 250°C for periods of time ofup to 150 hours. Two types of testswere carried out: (1) the pelletswere immobilized in the glass tubeby a coil of platinum wire so thatonly the liquid was in motion, (2) bothpellets and liquid were allowed tomove freely back and forth within thetube. The results of these tests aresummarized in Table 40. No significantloss in weight occurred during thetests in which the pellets wereimmobilized. The small weight lossobserved in the other test was dueto abrasion, since analysis of theliquid phase indicated essentially nodissolved thorium. The pellets heated

12 9

TABLE 40. STABILITY OF THORIUM OXIDE PELLETS

IN WATER AND URANYL SULFATE SOLUTION

Autoclave tests at 250 C

PELLET CONDITIONS OF TEST

DURATIONOF TEST

(hr)

CHANGE IN WEIGHTAS PER CENT OF

ORIGINAL WEIGHT

Tan, rough-surfaced

Pellets stationary, water in motion

Pellets and water in motion

150

150

0.02% loss

1. 53% loss

White, smooth- Pellets stationary, water in motion 150 0.06% gainsurfaced

Pellets and water in motion 150 0.08% loss

Pellets and water in motion 150 0.00%

5 g of U per liter U02S04 solution in motion, 120 0.23% gainpellets stationary

5 g of U per liter U0,S04 solution inpellets stationary

motion, 120 0.29% gain

5 g of U per liter UO2S04 solution in motion, 72 0.06% gainpellets stationary

5 g of U per liter U02S04 solution in motion, 72 0.02% gainpellets stationary

in uranyl sulfate solution actuallygained weight as a result of depositionof U03 on their surfaces.

Two additional tests were made with

immobilized pellets in water at 325°Cfor 56 hours. At the end of thesetests the pellets showed less than0.1% weight loss, and no thorium wasfound in the water phase.

DISSOLUTION OF THORIUM OXIDE PELLETS

The thorium oxide pellets werefound to dissolve in 60% HN03 containing 0.1 M fluoride as a catalyst.

130

The dissolution experiments werecarried out on a 10-ml scale in glassequipment. It was found that by usinga 100% excess of thorium oxide and

60% HNO,, a dissolver solution containing only 0.6 M acid could beobtained in 12 hours. This solution

would be a suitable feed for a solvent

extraction process.Since this dissolution study was

made with unirradiated pellets,additional studies must be made with

irradiated pellets before an acceptabledissolution process can be assured.

Documents

ORNL-1554 4 ISCWSS*