AEII RESEARCH AND DEVELOPMENT REPORT Re?jif Specia Wi I iy 3 Y45k 035147 2 VISCOSITY MEASUREMENTS ON MOLTEN FLUORIDE MIXTURES S. |. Cohen T. N. Jones i BEBlflSSlHED By AuTee H.n"r 0Oy: ? //t— -'(p / il AP di. .. - é.z 1R 0D nm (o) VA N (e AT DOCUMENT COLLECTION LIBRARY LOAN COPY Do IIOT THMISFEFI TG LR PERSBH 3 lf you R someone else to see lhis ' document, send in name with dflwmm‘l and the library will arrange a Iunn .5A ors OAK RIDGE NATIONAL LABORATORY OPERATED BY UNION CARBIDE NUCLEAR COMPANY A Division of Union Carbide and Carbon Corporation POST OFFICE BOX X * OAK RIDGE, TENNESSEE wtit ot 2270 This document consists of 50 pages. Copy &~ 208 copies. Series A, Contract No. W-T4O5-eng-26 Aircraft Reactor Engineering Division VISCOSITY MEASUREMENTS ON MOLTEN FLUORIDE MIXTURES S. I. Cohen T. N. Jones DATE ISSUED JUN = 8 1957 OAK RIDGE NATIONAL LABORATORY Operated by UNION CARBIDE NUCLEAR COMPANY A Division of Union Carbide and Carbon Corporation Post Office Box X Oak Ridge, Tennessee ARIETTA ENERGY SYSTEMS LIBRARIES y - IR 3 445k 0351470 e b | -ii- ORNL-2278 / - \ ' Reactors-Special Features ' of Aircraft Reactors INTERNAL DISTRIBUTION 1. C. E. Center 51. E. J. ??éeding 2. Biology Library 52. E. R, Dytko 3. Health Physics Library 53. R. Ji Gray L-5. Central Research Library 54, D."P, Gregory §. Reactor Experimental 55. . E. Mott Engineering Library 564 G. L. Muller T-11. ffifigratory'Records Department 57. R. D. Peak 12. Labowatory Records, ORNL R.C. gf58. W. B. Cottrell 13. A. M. inberg ~# 59. 8. J. Cromer 14, L. ;?%§§1§t (x-25) 4 60. J. H. DeVan 15. J. P. Murr 16, J. A. Swartout 17. E. H. Taylor % 18. E. D. Shipley 19, M. L. Nelson 20. 8. C. Lind 2l. F. L. Culler 22. A. H. Snell 23. A. Hollaender 2k, M. T. Kelley ; 25. J. H. Frye, Jr. & 26. K. Z. Morgan s 27. C. P. Keim (Y-12) & 61. W. K. Ergen K 62. A. P. Fraas S 63. W. T. Furgerson ! 64. A. G. Grindell 65. W. H. Jordan 66. F. F. Blankenship €7. W. R. Grimes 68. F. Kertesz 69. H. W. Savage T0. W. D. Manly 71. E. P. Blizard T2. R. E. MacPherson 73. E. R. Mann 28. R. 8. Livingston kN 7h. L. A. Mann 29. T. A, Lincoln 7 Ny, 75. W. B. McDonald 30. A. S. Householdey By 76. R. V. Meghreblian 31. C. S. Harrill " TT. A. M. Perry 32. C. E, Winters “. 78. D. H. Platus 33. D. W. Cardwell *79. D. L. Platus 34. E. M. King / - 80, R. D. Schultheiss 35. A. J. Miller/ 81.:D. B. Trauger 36. R. A. Charpie 82. M. M. Yarosh 37. J. A. Lane/ ' 83. W. F. Boudreau 38. M. J. Skirnner 8k. R. B. Lindauer 39. S. I. Colen 85. D. C. Hamilton LO. W. R. Gawmbill 86. R. E. Moore 41, N. D. Greene 87. G. J. Nessle 42, H. W. Hoffman 88. R. F. Newton 43, T. N. Jones 89. L. G. Overholser L, J. J. Keyes 90. R. E. Thoma 4. A. I, Krakoviak 91. G. M. Watson 46. F. E. Lynch 92, C. J. Barton 47. L. D. Palmer 93, C. M. Copenhaver 8. W. D. Powers oL, ORNL - Y-12 Technical Library, b9, W. J. Stelzman Document Reference -Section 50. J. L. Wantland o . X ‘ . AP ¥ ! 3 o - & . . - g (‘? wtdid - % EXTERNAL DISTRIBUTION 95. Air Technical Intelligence Center 0500, AP troject Cffice, TFort Worth ’ . Albuquerque Operations Office . Argonne National Laboratory A L. Armed Forces Special Weapons Project, Sandia R, Armed Forces Special Weapons Project, Washing \Assistant AF Plant Representative, Downey gsistant Secretary of the Air Force, R&Dfi 105-110. gmic Energy Commission, Washington 111, Njcs International 112. Rle Memorial Institute 113-11k, Nplant (WAPD) 115. & Aeronautics 116. Bureau oMgAeronautics (Code 2&) 117. Bureau of ¥ 118. Chicago Ope¥s 119. Chicago Patel§ Group fif 120. Chief of Nava¥;Research 121. Convair-GeneralDynamics Cérporation 122. Curtiss-Wright s%rporatlon 123. Engineer ResearcAyand Development Laboratories 12k-127. General Electric Cjmpary (ANPD) 128. General Nuclear Engfiwfering Corporation 129. Glenn L. Martin Comfigny 130. Hartford Area Offlceifiu 131-132. Headquarters, Air. Forfil 133, TIdaho Operations.Offic’b 134, Knolls Atomic Pfifier Labératory 135. Lockland Area Office W 136, Los Alamos Sgientlflc Labbgatory 137. Marquardt Ajrcraft Company A 138. National Aflvisory Committeegfor Aeronautics, Cleweland 139, National Adv1sory Committee 'f 140-142, Naval Ai¥ Development and Matey 143-1kk, Naval Research Laboratory (1 &dpy to L. Richards) 145, New Yqfik Operations Office S 146, i American Aviation, Inc. (Missile Development Division) 147. Nuclfar Development Corporation Rg Special Weapons Center 148, ce of the Chief of Naval Operayions (OP~361) 149, ent Branch, Washington 150. BRtterson-Moos 151-155. @Pratt and Whitney Aircraft Division ox Project) (1 copy %o #S. M. Kepelner) f. San Francisco Operations Office Sandia Corporation School of Aviaition Medicine Sylvania Electriec Products, Inc. 160. Technical Research Group 161. USAF Headquarters & “462. USAF Project Rand RORY 163. U. S. Naval Radiological Defense Laboratory ‘. 16L4. University of California Radiation Laboratory;“leermore 165-182. Wright Air Development Center (WCOSI-3) 183-207. Technical Information Service Extension, Oak Rldge 208. Division of Research and Development, AEC, ORO ", s, * 2 TABLE OF CONTENTS ABSTRACT .« v vv e es s enenananeennennnnnnesnenannasscenannns errene 1 INTRODUCTION . ........ e eraaean. et et e re e, 2 EXPERTMENTAL APPARATUS AND TECHNTQUES .o evvrvneeeencnnnneresnonnsons The DrybDOX +.ceveescoresacocrcosscsoonnsonsoorssoenssssovosasonsscss The Fm‘mce 0 A S PP RS e R NP b YD st R T RE DO R0 LT OE NS E RS rODY S SDO G -~ W\ The Capillary Viscomeler .c.ceescvooconnes P -~ The Brookfield Viscometer ....... ceoscsccsnoenss sececeseiesnescosease 1O INSTRUMENT CALIBRATION ...-6ecceovseococooecsecnsnssaancscansnsens cassess 19 EXPERTMENTAL RESULTS ....c00v000. tsccsssessacesotnansens cresecsenscsss 29 CORRELATIONS ..... crsenone cecavan tsecsceseranona G sconasuntasast e as . 34 RHERMCES * & & % @ & v 9T *® % % 0 3 B B KB B8 F A OO0 0 & P R GO T DL R R OO H SN EF O 9 o 9 & 3 & 0 O 45 ABSTRACT This report is a summary of the experimental viscosity program on fused fluoride mixtures which has been carried out in support of the ANP effort at the Oak Ridge National Laboratory. The experimental techniques which have been de- veloped are described, data on the visceosity of 36 mixtures are tabulated, and several correlations involving these data are discussed. IRTRODUCTION The value of fused fluorides as coolants and fuels for high tempera- ture, circulating-fuel reactors was recognized early in the CORNL-ANP program. However, the scarcity of known and reliable data on the thermal properties of these flulids seriously hampered the initial design efforts. To remedy this situation, a program was outlined to obtein experimental data on the viscosity, enthalpy and heat capacity, density, and thermal conductivity of the fused fluorides and their mixtures. This report is a summary of the portion of this prdgr&m concerned with the measurement of the viscosity of the fused fluoride salts. The experimenter is confronted with a number of difficulties in the measurement of the viscosity of these fluids: (1) The high melting points and wide extent of the liquid state of the fluoride salts establishes a temperature region of experimental interest between SOOOC and l,OOOOC. (2) All of the common materisls available for the construction of the experimental apparatus corrode in the molten fluoride salts. In the presence of even small amounts of oxygen or water vapor, the corrosion rates are greatly accelerated. (3) Since these salts in the molten state combine strongly with both oxygen and water vapor, the purification of the salt is difficult. Further, to insure & continuing purity, and hence avoid large amounts of contaminating corrosion products, it is necessary to maintain the molten salts in an inert atmosphere. () Some of the salts, such as beryllium fluoride, present prob- lems in handling due to their toxicity. (5) The fluid viscosity in the temperature range of intérest is low, varying from approximately 2 to 15 centipoises. These difficulties impose severe requirements on the measuring equipment by limiting the choice of techniques, construction materials, and instrumentation end by complicating the experimental manipuletion of the apparatus. A number of techniques hasve been investigated by the experimental groups at the Oak‘Ridge National Leboratory in developing suitable instruments for measuring the viscosities of fused salts. Knox and KErteszl and later Redmond2 3 used a modified form of the Brookfield rotational viscometer. El‘cvbiasslF developed a gravity flow capillary efflux viscometer. Knox and KErtesz5 oper- and Cisar ated a capillary device in which fluid flow was accomplished by gas pressure. A "falling-ball" viscometer using & radioactive cobslt-plated Pyrex bead was developed by Redmond and Kaplans. After review of this early work on the measurement of the viscosity of the fluoride salts, it was decided that the most expedient route to results of the desired sccuracy was through further refinements of the apparatus and techniques for the modified Brookfield and capillary efflux viscometers. Dur- ing this same period of apparatus development, salt preparation'techniques were greatly advanced by the Materisls Chemistry Division. Thus, the resultis reported are in genersl for high purity salt mixtures. The general techniques for fused salt viscosity messurements have been developed sufficiently to place these measurements in the category of routine. g _— " The detalled discussion of these techniques and the apparastus modifications is contained in the section of this report on experimental equipment and procedures. Some 36 mixtures have been studied at ORNL*. These have included binary, ternary, and quaternary mixtures of the following fluorides: NeF, LiF, KF, RWF, BeF,, ZrF, UF,, UF,, and TuF,. 3) A modest program of viscosity measurement on fused fluorides will be continued. As new mixtures are formulated and found to be of sufficient interest, viscosity data will be supplied to the various design groups studying fused fluoride reactors. In addition, a program to investigate fused salt systems other than fluorides is planned. At present this plan includes studies on chlorides, and mixtures of chlorides and fluorides. Sonme data7 have already been obtained in the chloride system. Other sys- tems will be studied as they become of interest. In addition to the ORNL program, the Mound Laboratory is involved in an extensive investigation of the viscosity of fluorides comprised mostly of mixtures of the alkali fluorides with BeF2 and UF#' This work is being carried out by B. C. Blanke using a rotational viscometer developed at the Mound Laboratory. o R — -5~ EXPERIMENTAL APPARATUS AND TECHNIQUES The experimental apparatus and techniques used in the measurement of the viscosity of the fused fluoride mixtures are described in the paragraphs which follow. This equipment includes the two viscometers (capillary efflux and rotationsl), the viscometer furnace, and the dry- box. The Drybox. Because of the necessity for handling these salts in an inert atmosphere, all measurements were made In a drybox under an argon blanket. This box was constructed from 1/8-inch mild steel with the top and front consisting completely of Plexiglas windows. The box was 50 inches high, 30 inches wide, and 24 inches deep. Long rubber gloves were attached to Plexiglas rings located on the front at shoulder height to ensble convenlient and comfbrtable manipulation of the equipment located within the box. A photograph of the box is shown in Figure 1. One side of the drybox was constructed as a door to provide access to the inside. This door was constructed of L/h-indb aluminum to reduce its weight and was sealed against a live rubber tubular gasket by KHUfVISE H-200 clamps located around its periphery. Shelves and a bracket to support the Brookfield viscometer were welded to the inside wells and a cooling coil made from L/h~inch copper tubing was located on the wall opposite the door. This coil was provided to reduce the temperature rise in the drybox due to prolonged operation of the furnace, The efficiency of this coil could be increased for high heat loads by the use of a 15 cfm blower to establish a forced draft across the tubes. \ _ o L UNCLASSIFIED: T PHOTO 27741 P N . et . . . o . T = 1 Fig. 4. Photograph of Drybox Showing Furnace and Brookfield Viscometer in Operating Position. | -7- A number of precautionary measures were taken during construction of the box to insure gas tightness. Thermocouple leads were brought through the drybox wall in Conax fittings and attached to terminal blocks on both the inside and ocutside surfaces of the wall to eliminate strain on the seal. Electrical leads were introduced through similar special fittings. These fittings, as well as the fittings used for the cooling coil, gas in- let, and gas outlet, were welded in place. The welds were heavily pasinted with Glyptal before the finasl paint was applied to the drybox. The box wes periodically checked for tightness with a Freon lesk detector. A series of {ests was conducted to ascertain fihe optimufi procedure for obtaining argon drybox atmosPheresg. It was found that best results were produced by introducing argon at the bottom of the box and allfiwing it to exit near the top through an oil bubbler. The box was usually purged in this manner overnight with a flow of from 4 to 8 cubic feet per hour of argon before the test mixture was exposed to the drybox stmosphere. The Furnace. It was necessary in carrying out this experimental program to design a tube furnace which would bring a capsule consisting of a 1k-inch section of one-inch I.P.S. pipe filled with a fused fluoride to red heat quickly, hold the temperature constant along the entire length of the capsule, and attain equilibrium rapidly after adjustments in the temperature setting. A disgram of the furnace along with the temperature measurement and control systems ig shown in Figure 2, and a photograph of the furnace is shéwn in Figure 1. The furnace shell was a cylinder 16 inches high and 10 incheg in dia- meter, mounted on casters and equipped with leveling screws which can be THERMOCOUPLE LOCATION UNCLASSIFIED ORNL-LR-DWG 19834 SALT CAPSULE BROWN 4 Cr-Al THERMOCOUPLES MULTIPOINT HEATER ELEMENT—\ RECORDER N 7 N\ z Y - A8 c|x % PN Cr-Al :;: %J Z %%% 2 SIMPLITROL @ | w Z 8/ . Sy . 7'2\2 | RELAY °] AN /\é SN /\_ § 7 . VARIAC "0 v 1 T |y A 7 \24 N 7 7 E \i / N g %8 SIMPLITROL T g , 0 4\g =2 é % e — 7 \ A / “ \-msuumom [ L : 1 Y / Y - CERAMIC BLOCK COPPER LINER s Fig. 2. Diagram of Furnace with Temperature Control and Measurement Systems. _8- pamnitl a Vo operated from several inches above the furnace. It was insulated with semi- circular blocks cut from slabs of JM-3000, an aluminum silicate insulating brick manufactured by Johns-Manville. Two pairs of 3-inch I.D. "clam-shell" resistance heaters made up the furnace element. These two units, having léngths of lO-L/E inches and 3-1/2 inches, were located in the furnace with the short section on top and were controlléd separately (see furnace diagrsm, Figure 2). The lower unit heated the liquid and the upper unit served as a "guard zone" heater. This guard heater was necessary for two reasons., First, the zone just above the liguid surface must be kept at the temperature of the ligquid or the resulting tempersture gradients will cause convection currents giving slightly erroneous readings on the Br@okfiéld viscometer. Second, 1t is essentiel that the efflux cup viscometer be held in a zone at the tempera- ture of the liquid while it is draining or an error is obviously introduced. It was found that two separately controlled elements provided the simplest means of maintaeining the temperature of this zone at the liquid temperature. Temperatures of the two elements were controlled by circuits containing chromel-alumel thermocouples, relays, and "Simplitrols.” Variacs were also inclfided in these circuits to afford more sensitive adjustment. The hot Junctions of the thermocouples were imbedded in the outer face of a heavy copper pipe occupying the annuius between the furnace element and the capsule containing the sample. This copper pipe, having a wall thickness of almost one inch, served as a thermal diffuser to provide very even temperature dis- tribution in the sample. L -10- y Liquid temperatures were measured using four chromel-alumel thermo- couples inserted in wells of 0.08-inch 0.D. thin-walled tubing which were snapped into 0.08-inch square grooves machined longitudinally in the outer wall of the capsule (see Figure 2). The wells were cut to such length that the junctions of two thermocouples were located in the liquid zone and two were in the zone above the liquid. A diagram showing the relative positions of the four thermocouples to the two viscometers in operation may be found in Figure 3. Temperatures measured by these four couples were read on a Brown multipoint recorder. Messurements were made with both the capillary and rotational instru- ments using the same sample of salt during = single heat-up period. About 100 cc of molten salt were required to fill the capsule to sufficient depth to operate the two instruments. A complete viscosity-temperature curve was determined using one of the instruments before measurements were begun with the other. After a satisfactory inert atmosphere had been obtained in the drybox, the furnace controls were set at a temperature 25 to 50 degrees above the liquidus temperature of the salt and the system was allowed to come to equilibrium. The small guantity of graphite present in the salt from the preparation equipment rose to the surface as the salt melted. This graphite was removed by touching the surface of the salt with a cold tamper, asllowing the superficial layer conteining the graphite to freeze on the tamper. The capillary cup was usually used in making the first measurements on a new salt. To establish the entire viscosity-temperature eurve, the température of the melt was incrementally raised and measurements were made v — -11- ORNL-LR-DWG 19835 = \ g \ 18 \ \\ 1.85 E‘JJ \ = 16 \ 1.75 4 \\ 2 \\ o \\ 8 AN ABSOLUTE VISCOSITY \ \ (centipoises) 6 \ \ \ \\ 4 \ \ — KINEMATIC VISCOSITY \ '\\ tistok 2 centistokes] \\ ~— \ 0 100 150 200 250 300 350 400 i Fig. 6. The Density and Viscosity of "HTS". TEMPERATURE (°C) g _2k- manner similar to that deseribed in conjunction with the glycerol solution calibrations. Use of this salt mixture had the additional advantage that it reduced the already small error resulting from the slightly different dimensions of the instruments in the glycerol solutions st room tempera- ture and the fluorides at high temperatures, The other calibrating liquid, s-tetrabromoethane, is a heavy, highly toxic, organic halogen compound having a room tempersture density of 2.964 gm/cc. The viscosity of this liquid varies from about 11.5 centi- poises at 20°C to about 3.8 centipoises at 6500. The viscosity-temperature curves for this liquid may be found in Figure 7. Calibration was carried out in this liquid by containing it in one of the capsules used for the fluorides and supporting this capsule in a constant temperature bath. This liquid was used less frequently than HTS because of its toxicityll. A complete calibration curve was determined for each new cup or spindle before it was put into use. After each set of messurements on a fluoride mixture, two or three recalibration points were made to insure that the in- strument had not been altered during use. Representative calibration curves for the Brookfield and capillary viscometers are shown in Figures 8 and 9, respectively. Above a resding of about 10 centipoises, the ratio of Brook- field reading to viscosity becomes constant at 1-1/2. The changes made in the use of the Brookfield, coupled with the furnace design employing two elements and the use of calibrating liquids of higher density, have virtually eliminated the earlier poor agreement between the two different viscometers. The results with the two instruments now agree o -25- UNCLASSIFIED ORNL-LR-DWG 19838 N\ KINEMATIC VISCOSITY N N N ABSOLUTE VISCOSITY (centipoises) 14 12 10 \éBSOLUTE\HSCOSHY N \\ \\ 20 DENSITY: p4 =2.964 g/cc 10 20 30 40 50 60 TEMPERATURE (°C) Fig. 7. The Viscosity of s-Tetrabromoethane. 70 KINEMATIC VISCOSITY (centistokes) VISCOSITY (centipoises) 14 10 UNCLASSIFIED ORNL-LR-DWG 19839 O ORIGINAL CALIBRATION -~ A REGALIBRATIONS ] X ,/ a B ° d}/ A / P // /O o @/. ./ s~ >?<\8770 800 " 20 / [ Tes — - ~— /[ 7 1o TIOT 0 T4 14 — / 2 1 2.0 2.5 3.0 35 4.0 45 5.0 5.5 CALCULATED ROOM TEMPERATURE DENSITY (g/cc) s Fig#10. Viscosity at 700°C vs. Calculated Room Temperature Density. _ 37 b mixtures 12 and 43, are the lightest and densest salts, respectively, that are likely to be encountered in non—BeF2 fluoride mixtures. It is seen that, with the exception of two mixtures, the data fall within + 30 per cent of this average line. Figure 10 also.shows that all of the mixtures containing BeF2 are above the + 30 per cent line and that the viscosities of the BeF2 salts vary as their BeF2 content over the concentration range (15 to 50 mole per cent) studied. The relstive effect of the various alkali fluorides on the viscosity is indicated in Figure 10. Mixtures 81, 82, 86, and T2, vwhich are considerably above the average line, are zirconium-base fuels containing LiF as the prin- cipal alkali fluoride. Salts 87, 95, and 117, which fall somewhat below the average line, are zirconium-base fuels containing rubidium as the alkali flu- oride. The mixtures of intermediate viscosity, suéh as 30, 44, and 70, are similar mixtures conteining NaF. All of téeae trends are substantiated by a correlsation resulting from g8 statistical treatment of the data. This approach was based on & number of empirical equations cited by Hatgchek35 relating the viscosities of mixtures to their compositions. The simplest assumption is that the viscosity of an ideal mixture should be additive. Thus, bo=pX+ o, (1-x) (4) where pu is the viscosity of the mixture, and u, are the viscosities of the Hy components, and x is the mole fraction of one of the two components. Bingham stated that the reciprocal of viscosity, the fluidity (@), was additive. Thus, ¢ =0.x+ 0, (1-x) (5) . a5 U ot -38- b Arrhenius proposed the purely empirical function, log = x log p, + (1 - x) log Mo (6) Kendall and Monroe found for some mixtures of nonassociated liquids that 1 — + (1 - x) uz | (7) [~ WYo= X p L Wik From these empirical equations and the observed data trends, a corre- lation based on equation (%), assuming the viscosities of the fluorides to be approximately additive, was indicated. The viscosity of the mixture was represented by the equation b= up Ty +oppfp + e+ f (8) where K represents the viscosity of pure component A, fA is the mole frac- tion of that component, and so forth, for the eight components. A matrix containing as its elements the mole fractions of the components in all the mixtures studied along with the viscosities of the mixtures at 600°C, 700°C, and 800°C was programmed for the ORACLE, Solutions were obtained at»the three temperstures for the viscosities of the pure components. It must be understood that these "pure component viscosities” are merely empirical rep- resentations for use in this relationship since the pure components are not liquids at these temperatures. Solutions wére also obtainéd at the three temperatures using the same matrix along with the viscosities expressed as log p and as fluidity, or 1/u, as in equations (5) and (6). Compsrison of experimental and calculated viscosities using equations (4), (5), and (6) indicated that the three correlations were of about equal merit. ‘ -39- N However, the degree of correlation was only feir. Viscosities calcu- lated by this correlation for four salts at TOOOC differed from the experi- mental values by more than 20 per cent. The average deviation between experimental and calculated viscosities at 70000 for all of the mixtures studied was approximstely 10 per cent. It must be concluded, therefore, that this correlation would not be a safe basis for predicting the viscos- ities of new mixtures. The "pure component viscosities" obtained from the ORACLE for the equation involving linear additivity, the simplest of the three functions tested, are found in Table II. A comparison of the experi- nmental and calculated viscosities at TOOOC based on this function is given in Table III. Another correlation which may be applied to the experimental results involves the relastionship between fluidity and the number of "holes” in a liguid. If holes must be present in a liquid before flow can take place, 36 as has been postulated by‘Eyring , it is reasonable to assume that the fluidity of a liquid will be.proportional to the number of holes. In addi- tion, the essential difference between a solid and a liquid may be regarded as the introduction of holes. If V is the molér volume cf the liquid and V, is the molar volume of the unexpended solid, the difference, (V - V_), is proportional tc the number of holes and, hence, to the fluidity. Since the fluidity is the reciprocal of the coefficient of viscosity, it follows that (9) HEV Y 8 ™ where ¢ 1s a constant. This equation was originally proposed by Batschinaki3‘ and w&s fqud to hold for a large number of nonassocisted liquids. [ -h0- ¢ i TABLE II "PURE COMPONENT VISCOSITIES" BASED ON EQUATION (8) "Pure Component Viscosity" (Centipoises) Component 600°¢ 700°¢ 800°¢ NeF (A) - 1.33319 0.414287 0.495901 LiF (B) 8.58465 5.27885 3.38107 KF (c) 1.03168 0.367723 0.332103 RbF (D) - 2.90961 - 1.43543 - 1.93418 BeF, (E) 27.3054 13.4185 8.10863 ZrF), (F) 15.1988 9.09101 6.54347 UF), (@) 60.9819 31.2966 20.5855 ThF), (B) 16.0516 7.43398 7.46834 TABLE III COMPARTSON OF EXPERIMENTAL VISCOSITIES AT TOOOC AND PREDICTED VALUES BASED ON EQUATION (8) Viscogity at TOOOC, Centipoises Per Cent Experimental Predicted Deviation 9.8 8.89 9% 2.9 2.66 8% 2.75 2.90 5% .1 4,12 0 5.4 5.6k 44, 5.2 k.75 % 8.5 10.30 21% 7.0 6.01 14% 10.25 10.70 4% 5.7 5,89 3% 4.6 b 49 2% 5,2 5.3h 3% 6.0 6.92 15% 4,9 7.80 59% 4,0 4,83 21% 7.0 5.09 27% 7.0 6.15 12% 6.45 5.87 9% 4.65 4,93 6% 4,75 4,69 1% 5.0 5.09 2% k.35 b, 71 8% b1 4.57 11% 3.95 4,12 49, 4.6 k. o6 8% Per cent deviation is defined as (difference/experimental) x 100. “hoo ‘;m:f““ TABLE III (Continued) COMPARISON OF EXPERIMENTAL VISCOSITIES AT 700°C AND PREDICTED VALUES BASED ON EQUATION (8) Viscosity at TOOOC, Centipoises Per Cent Mixture Experimental Predicted Deviation# 100 3.2 3.33 4% 107 3.0 3.37 12% 111 7.1 7.10 0 113 8.4 6.92 18% 114 6.7 6.89 3% 115 5.2 5.99 15% 117 4.8 h.25 13% 118 b.75 4,98 5% 119 5.2 6.17 19% * Per cent deviation is defined as (difference/éxperimental) x 100. e Wt -43- This relationship was applied to the experimental data on fluorides reported here, and the resulis are shown graphically in Figure 11. The term p in the equation is the viscosity at 700°C, V is the molar volume or reciprocal of \§iliUMsity at 700°C, and V, is the molar volume or re- ciprocal of the density at room temperature calculated from equation (3). Experimental values are used for the liquid density where available; cal- culated values are used in the other instances. Although the points shown on the figure exhibit considerable scatter, the hyperbols described by the equation éan be discerned. In contrast to the density-viscosity trend shown in Figure 10, this relationship appears to correlate the data for liquids containing BeF2 as well as for the non—BeF2 mixtures. However, the BeFa—bearing ligquids studied have contain- ed only moderate proportions of this constituent. Calculations using some of the data taken at Mound Laborstory on liquids containing enough BeF2 to sharply increase the viscesityl2 result in greatly increased scatter. Some of the scatter shown in Figure 11 is probably due to the error limits to which the experimental and predicted date are subject. However, it is safe to assume that most of the divergence encountered with this correlation, as well as with the others described, is due to factors such as the general complexity of the viscosity phenomenon and the wide varia- tion in the physical nature of the fluoride mixtures. ORNL - LR—DW! 19842 042 0.10 0 LO o $ 0.08 H—@ N 'F % 0 E o | ° .8 o ® o O%Oo O o 0.06 5 o i\- o 0.04 f © o NON-BeF, SALTS o 0.02 ® BeF, —BEARING SALTS O 0 2 4 o 8 10 12 VISCOSITY AT 700°C (centipoises) Fig. 11. Viscosity at 700°C vs. (VTOOOC - Vs). ] ~45- A 10O. 11. 3.2. 130 1k, 15, 16. tggfifigl REFERENCES Knox, F. A., and Kbrtesz, F., ANP Quarterly Progress Report for Period Ending September 10, 1951, ORNL-1154, p. 136. Redmond, R. F., ANP Quarterly Progress Report for Period Ending June 10, 1952, ORNL-129}+, Po 12}4‘- Cisar, J. M., ibid. Tobias, M., ANP Quarterly Progress Report for Period Ending December 10, 1951, ORNL -1170, p. 12k. Knox, ¥. A., and Kertesz, F., ANP Quarterly Progress Report for Period Ending September 10, 1952, ORNL-1375, p. 145. Redmond, R. F., and Kaplan, 8. I., "Remarks on the Falling Ball Viscometer,"” ORNL CF 53-1-248, January 1%, 1953. Cohen, 8. I., ANP Quarterly Progress Report for Period Ending December 31, 1956, ORNL-2221, p. 228. Cohen, S. I., and Peele, J. M., "Determination of the Optimum Procedure for Obtaining Inert Atmospheres in Non-Vacuum Dryboxes," ORNL CF 55-10-132, October 26, 1955. "Modern Developments in Fluid Dynsmics," Vol. II, pp. 388-390, Oxford Press, Oxford, (1938). Kirst, W. E., Nagle, W. M., Castner, J. B., "A New Heat Transfer Medium for High Temperatures," Trans. AIChE, 36, p. 371, (1940). Sax, N. Irving, "Handbook of Dangerous Materials," p. 5, Reinhold, New York, (1951). Blanke, B. C., MIM CF 55-11-14%, November 7, 1955. Cohen, S. I., Powers, W. D., and Greene, N. D., "A Physical Property Sumuary for ANP Fluoride Mixtures,"” ORNL-2150, August 23, 1956. Cohen, S. I., ANP Quarterly Progress Report for Period Ending December 10, 1955, ORNL-2012, p. 100, Cohen, S. I., and Jones, T. N., ORNL CF 55-%-32, April 1, 1955. Cohen, S. I., and Jones, T. N., ORNL CF 56-5-33, May 9, 1956. ‘ - 1 | -46- & t‘. REFERENCES (Continued) 17. Cohen, S. I., and Jones, T. N., ORNL CF 55-2-20, February 2, 1955. 18, Cohen, S. I., and Jones, T. N., ORNL CF 56-4-148, April 17, 1956. 19, Cohen, S. I., and Jones, T. N., ORNL CF 55-12-128, December 23, 1955. 20. Cohen, S. I., and Jones, T. N., ORNL CF 55-2-89, February 15, 1955. 21. Cohen, 8. I., and Jones, T. N., ORNL CF 55-3-137, March 16, 1955. 22. Cohen, S. I., and Jones, T. N., ORNL CF 55-9-31, September 6, 1955. 23. Cohen, S. I., and Jones, T. N., ORNL CF 55-3-61, March 8, 1955. 24, Cohen, S. I., and Jones, T. N., ORNL CF 55-5-59, May 16, 1955. 25. Cohen, S. I., and Jones, T. N., ORNL CF 55-5-58, May 16, 1955. 26. Cohen, S. I., and Jones, T. N., ORNL CF 55-7-33, July 7, 1955. 27. Cohen, S. I., and Jones, T. N., ORNL CF 55-11-27, November 4, 1955. 28. Cohen, S. I., and Jones, T. N., ORNL CF 55-8-21, August 1, 1955. 29. Cohen, S. I., and Jones, T. N., ORNL CF 55-11-28, November 8, 1955. 30. Cohen, S. I., and Jones, T. N., ORNL CF 55-12-127, December 23, 1955. 31. Cohen, S. I., and Jones, T. N., ORNL CF 55-8-22, August 1, 1955. 32. Cohen, S. I., and Jones, T. N., Unpublished data. 33. Blanke, B. C., MIM CF 55-10-34, October 24, 1955. 34. Cohen, S. I., and Jones, T. N., "A Summary of Density Measurements on Molten Fluoride Mixtures and a Correlation for Predicting Densities of Fluoride Mixtures," ORNL-1702, July 19, 195k, 35. Hatschek, Emil, "The Viscosity of Liquids,” pp. 135-139, D. Van Nostrand, London, (1928). 36. Glasstone, 8., Laidler, K. J., and Eyring, H., "The Theory of Rate Processes, " p. 481, McGraw-Hill, New York, (194l1). 37. Batschinski, A. J., Z. physik Chem., 84, p. 643, (1913). g *