ORNL-TM-3259.txt

 

CFNTRAL RESEARCH LIBRARY ORNL.-TM.3259
TR cy52

3 445k 0383070 9

ENGINEERING DEVELOPMENT STUDIES
FOR MOLTEN-SALT BREEDER REACTOR
PROCESSING NO. 9

L. E. McNeese

OAK RIDGE NATIONAL LABORATORY
CENTRAL RESEARCH LIBRARY
DOCUMENT COLLECTION

LIBRARY LOAN COPY
DO NOT TRANSFER TO ANOTHER PERSON

If you wish someone else to see this
document, send in name with document
and the library will arrange a loan.

 

2,
=

OAK RIDGE NATIO LABORATORY

OPERATED BY UNION CARBIDE CORPORATION = FOR THE U.S. ATOMIC ENERGY COMMISSION

 

 
ORNL-TM-3259

Contract No. W-TL05-eng-26

CHEMICAL TECHNOLOGY DIVISION

ENGINEERING DEVELOPMENT STUDIES FOR MOLTEN-SALT
BREEDER REACTOR PROCESSING NO. 9

L. E. McNeese

DECEMBER 1972

OAK RIDGE NATIONAL LABORATORY
Oak Ridge, Tennessee 37830
operated by
UNION CARBIDE CORPORATION
for the
U. S. ATOMIC ENERGY COMMISSION

MARTIN MARIETTA ENERG

T

3 445k 0383070 9
ii

Reports previously issued in this series are as follows:

ORNL-TM-3053
ORNL-TM-3137
ORNL-TM-3138
ORNL-TM-3139
ORNL-TM-3140
ORNL-TM-~3141
ORNL-TM-3257
ORNL-TM-3258

Period
Period
Period
Period
Period
Period
Period

Period

ending
ending
ending
ending
ending
ending
ending

ending

December 1968
March 1969
June 1969
September 1969
December 1969
March 1970
June 1970
September 1970
iii

CONTENTS

SUWARIES o » » . . . . ° . . . ° . . . . . . . . . . . . . . . . -Vii

1.

2.

INTRODUCTION . @ > L n » . o o ° . . - . » . » . - . * L] L . °

FLOWSHEET ANALYSIS: ISOLATION OF PROTACTINIUM BY OXIDE
PR.ECIPITATION » L . B . s . - . » o . ? o - . L * 8 L

2.1 TIsolation of Protactinium by Oxide Precipitation, and
Recovery of Uranium Daughters by Fluorination. . . . . . .

2.2 1Isolation of Protactinium by Oxide Precipitation Without
the Use of Fluorination for Recovering 233U Produced by
Decay of 233Pa . v i e e e e e e e e e e e e e e e e e

FLOWSHEET ANALYSIS: REFERENCE PROCESSING PLANT FLOWSHEET BASED
ON FLUORINATION, REDUCTIVE EXTRACTION, AND THE METAL TRANSFER
PROCESS ® - . o ° @ s . ® © . . . o > . - » L] ¢ . . . » » ; . "

FLOWSHEET ANALYSIS: IMPORTANCE OF URANIUM INVENTORY IN AN MSBR
PROCESSING PIJANT - o . o - - - . k] - e . > L b a * . » . s - . -

FLOWSHEET ANALYSIS: REMOVAL OF RARE-EARTH FISSION PRODUCTS
FROM LiCl IN THE METAL TRANSFER PROCESS . . « ¢« & & + o & o o

FROZEN-WALL FLUORINATOR DEVELOPMENT: EXPERIMENTS ON INDUCTION
HEATING IN A CONTINUOUS FLUORINATOR SIMULATION. . « « « « o« « &
6.1 Experimental Procedure . . . « ¢ + o 5 s o s e v o4 s o4 s
6.2 Experimental Results . . s o o o o o o s o &+ s o o s o o

PREDICTED CORROSION RATES IN CONTINUOUS FLUORINATORS EMPLOYING
FROZEN—'WALL PROTECT ION . . . . . s . . . . o n ° . . ° ° . . »

7.1 Data on the Rate of Corrosion of Nickel in Gaseous
Fluorine o . & . - o - . » - . - » 5 - - » - a a - » v o .

7.2 Predicted Corrosion RatesS. . « « o o o o s s s 6 s & s o
PREDICTED PERFORMANCE QOF CONTINUOUS FLUORINATORS. . + « « o + &

MEASUREMENT OF AXTAL DISPERSION COEFFICIENTS AND GAS HOLDUP IN
OPEN BUBBLE COLUMNS ¢ &+ & « ¢ o o o o s 5 s o 3 s o o & o o« »
9.1 Previous Studies on Axial Dispersion . . « + + &+ « &+ + o+
9.2 Equipment and Experimental Procedure . . . + « « « « « o
9.3 Experimental Data on Axial Dispersion. . . . . . + « . &
9.4 Experimental Data on Gas Holdup. . « « « « « + & s + & o o

9.5 C(Correlation of Data on Gas Holdup. . + « « + ¢ & « « + &

17

22

24

27

27
30

. 33

33
37

41

52

52
54

. 55

72
74

9.6 Compilation of Data on Axial Mixing. . . . . . . « « +. . 106
10.

11.

12.

15.

1h.

iv

CONTENTS (continued )

Page
9.7 Correlation of Data on Axial Dispersion. . . . . . . . . . 106
9.8 Discussion of Data on Axial Mixing and Gas Holdup. . 126
9.8.1 Flow Regimes and Effect of Gas Superficial Velocity . 126
9.8.2 Effect of Column Diameter . . . c e e e ... 12
9.8.3 Effect of Viscosity of the L1qu1d Phase c e v e v .. 1hs
9.8.4 Effect of Surface Tension of the Liquid Phase . . . . 146
9.8.5 Effect of Gas Inlet Orifice Diameter. . . . . 146
9.8.6 Effect of Number of Orifices in Gas Dlstrlbutor 148
9.9 Future WOork. « « + « ¢ + & v v v v v v v e e e e e 148
SEMICONT INUOUS REDUCTIVE EXTRACTION EXPERIMENTS IN A MILD-STEEL
FACILITY. « « & « + & & o = o o v 4 o o 4 4« v v s o« v o o« « o llg
10.1 Run UTR-6 « + v v v v v v v vt v e e e e e e e e 150
10.2 Preparation for Run UTR-T; Installation of Molybdenum Draft
Tube in Treatment Vessel. . . . + « « « & « + « . « « . . . 154
10.3 Preparation for Mass Transfer Run ZTR-1; Production of 9TZr
by Irradiation of 90Zr. . . . . . . . . . . . . . ... 158
10.4 1Inspection of Salt Feed-and-Catch Tank, and Equipment
Maintenance .« . « « « « ¢ 4 4 e v e e e e 4 e e e e e e 159
DEVELOPMENT OF THE METAL TRANSFER PROCESS: OPERATION OF EXPERI-
MENT MTE=2. « « + v o o o o o o o o o o o o o o o o o v o o o« + . 167
11.1 Experimental Procedure. . . « . . « . . . 168
11.2 Mathematical Analysis of Transfer Rate. . . . . . . . . 169
11.35 Experimental Results. . « . « « « « + & o v & « & o « 17h
11.3.1 Rates of Transfer of Neodymium and Lanthanum. 174
DEVELOPMENT QF THE METAL TRANSFER PROCESS: DESIGN OF EXPERIMENT
MIE=3 +© & & & & o o o o o o s + o s o o + o o e e e 4 e e e o« . . 196
12.1 Mathematical Analysis of Metal Transfer Experiment MTE-3. 197
12.2 Preliminary Design of Metal Transfer Experiment MITE-3 . 202
DEVELOPMENT OF MECHANICALLY AGITATED SALT-METAL CONTACTORS. . . . 205
13.1 Hydrodynamic Studies. . « . « « .« « .+ « ¢ . . . . . 205
15.2 Survey of Literature Relative to Mechanically Agitated,
Nondispersing Salt-Metal Contactors « - « « « « « « « .+ . « 208
HYDRODYNAMICS OF PACKED-COLUMN OPERATION WITH HIGH-DENSITY FLUIDS 216

14.1 Equipment and Experimental Technique. . . . . . . . . . . .

217
15.

6.

17.

CONTENTS (continued)

].)-".2 ResultS- . - . * . - . - . . - . . . .
14.3 Prediction of Flobding Rates and Dispersed-Phase Holdup
in Packed Columns. e e e e . . .

ANALYSIS OF MULTICOMPONENT MASS TRANSFER BETWEEN MOLTEN SALTS
AND LIQUID BISMUTH DURING COUNTERCURRENT FLOW IN PACKED
COLUmS - ». . * . . - & ® - - . - - . 2 * L . * . - - * &

15.1 Literature Review.

15.2 Mathematical Analysis.

15.3 Calculational Procedure. . « « « « « + o o« o« o« o«
STUDY OF THE PURIFICATION OF SALT BY CONTINUOUS METHODS.

16.1 Removal of Oxide from Salt by Countercurrent Contact with

an :[']_-_FI-I']'.2 GaS Stream- . . . . . » . * . . - . . . » . . .

16.2 Removal of Oxide from Column . . « « « + « v o« & + & « &
16.3 Iron Fluoride Reduction RUnms . « « « « o s+ o o o o o o

16.4 Calculated Values for the Mass Transfer Coefficient and
the Reaction Rate Constant During the Reduction of Iron
Fluoride '

REFERENCES .

Page
219

233

238

238
240
2hé

251

252
254
256

256

261
vii

SUMMARIES

FLOWSHEET ANALYSIS: ISOLATION OF PROTACTINIUM BY OXIDE PRECIPITATION

Two flowsheets that employ oxide precipitation for protactinium
removal are described, and the effects of several operating parameters
on the performance of the flowsheets have been investigated. In the
first flowsheet, protactinium is selectively precipitated from MSBR
fuel salt on a 3-day cycle. The resulting oxide and a small amount of
fuel salt associated with it are hydrofluorinated in the presence of a
secondary fluoride salt that is circulated through a fluorinator and
a protactinium decay tank. A small fraction of the salt leaving the
decay tank is returned to the primary reactor circuit to compensate
for salt that is transferred to the decay tank along with the oxide.
The uranium is removed from 10%Z of the fuel salt leaving the precipi-
tator by fluorination or oxide precipitation, and rare earths are
removed from the resulting salt by the metal transfer process. The
purified salt leaving the metal transfer process is combined with the
uranium removed earlier, and the resulting stream is returned to the
reactor. A protactinium removal time of 5 days can be realized if
607 of the protactinium is separated from the salt in the precipitator,
provided the fuel salt transfer rate to the decay tank is as low as
10 to 20 moles/day. For the same protactinium removal time, a pro-
tactinium removal efficiency of 80% would be required in the precipitator
if the fuel salt transfer rate to the decay tank were as large as 3000

moles/day. The uranium inventory in the decay tank would be negligible.

In the second flowsheet, a fluorinator is not used for removal of
uranium from the protactinium decay tank. Fuel salt is withdrawn from
the reactor on a 3-day cycle and combined with a salt stream that is
withdrawn from the protactinium decay tank. Part of the protactinium
in the resulting salt stream is removed by precipitation, and the pre-
cipitate and associated salt are hydrofluorinated in the presence of
processed fuel carrier salt leaving the metal transfer process. The

resulting salt stream then passes through a decay tank, from which it
viii

is fed to the protactinium precipitator in order to return the uranium
to the reactor. Operation of the flowsheet is highly dependent on the
fraction of the protactinium removed in the precipitator and on the

amount of fuel salt that accompanies the oxide precipitate.

FLOWSHEET ANALYSIS: REFERENCE PROCESSING PLANT FLOWSHEET
BASED ON FLUORINATION, REDUCTIVE EXTRACTION,
AND THE METAL TRANSFER PROCESS

Operating conditions that will constitute the reference fluori-
nation--reductive extraction--metal transfer flowsheet were selected,
and additional calculations were performed to indicate the operating
characteristics of the flowsheet. ZEssentially complete extraction of
the protactinium is achieved with a 10-day processing cycle, a five-
stage protactinium extractor, a lithium reductant addition rate of
200 equiv/day, and a uranium removal efficiency of 99%Z in the primary
fluorinator. Rare earths are extracted from the fuel salt with removal
times ranging from 16 to 50 days in a three-stage extractor. A three-
stage extractor is also used for the selective transfer of the rare
earths from the bismuth-plus-thorium phase and the extracted rare earths
to a LiCl stream. The various waste salt streams produced by the pro-
cessing system are combined into a single stream having the composition
76.3-12.3-9.8-0.64 mole 7 LiF—ThFa—BeFZ-ZrF4, 0.864 mole 7 trivalent
rare-earth fluorides, and 0.114 mole % divalent rare-earth fluorides.

The waste salt would be discarded from the processing system at the

rate of 70 ft3 every 220 days.

FLOWSHEET ANALYSIS: IMPORTANCE OF URANIUM
INVENTORY IN AN MSBR PROCESSING PLANT

The MSBR processing flowsheets considered to date have resulted
in uranium inventories in the processing plant that are quite low,
usually less than 17 of the inventory in the reactor. Since several
potential processing flowsheets may result in uranium inventories as
large as 10% of the reactor inventory, the importance of increases

in this inventory was examined. It was found that increasing the
ix

processing plant inventory from O to 10% would increase the fuel cycle
cost by only 0.03 mill/kWhr and would increase the system doubling
time from 22 to 24.2 years. It was concluded that, while there are
incentives for maintaining a low inventory, inventory values of 5 %o

10% would not rule out an otherwise attractive processing system.

FLOWSHEET ANALYSIS: REMOVAL OF RARE-EARTH FISSION PRODUCTS
FROM LiCl IN THE METAL TRANSFER PROCESS

Calculations were made to determine the effect of varying the con-
centration of lithium in the bismuth solution used for removing the
trivalent rare earths from the LiCl in the metal transfer process; the
reactor breeding ratio was found to decrease only slightly (from about
1.063 to about 1.060) as the lithium concentration in the bismuth was
decreased from 5 at. % to approximately 1.67 at. 4. Calculations were
also carried out which indicate that a single-stage extractor has essen-
tially the same removal efficiency for the divalent rare earths in the
reference flowsheet as a two-stage contactor; thus the use of a single-

stage contactor was adopted.

FROZEN-WALL FLUORINATOR DEVELOPMENT: EXPERIMENTS ON INDUCTION
HEATING IN A CONTINUOUS FLUORINATOR SIMULATION

An experiment to demonstrate protection against corrosion by the
use of a layer of frozen salt in a continuous fluorinator requires a
corrosion-resistant heat source to be placed in the molten salt. High-
frequency induction heating appears to be an acceptable heating method,
and equipment has been installed for studying this method in a simu-
lated fluorinator that uses a 31 wt Z HNO3 solution in place of molten
salt. Experimental results on heat generation rates in the acid, in
the pipe surrounding the acid column, and in the induction coil are

presented for the first eight runs.
PREDICTED CORROSION RATES IN CONTINUOUS FLUORINATORS
EMPLOYING FROZEN-WALL PROTECTION

Nickel is the preferred material of construction for fluorinators
in MSBR processing plants since it exhibits greater resistance to attack
by gaseous fluorine than other candidate materials. This resistance
is due to the formation of a tightly adherent film of Nin, and it is
proposed that a layer of frozen salt be used to prevent removal of
the NiF, film via dissolution in the molten fluoride mixture that flows

2
through the fluorinator. However, it is expected that the NiF, film

will be removed periodically as the result of deviations from ihe

desired mode of operation, and an analysis was carried out for estimating
the resulting corrosion rate under such conditions. It was found that,
if the NiF2 film were destroyed 52 times per year, the average yearly
corrosion rates at 450°C would be 2.9 mils and 0.97 mil for types 200

and 201 nickel respectively. It appears that either material will

show satisfactory corrosion resistance if the NiF2 film is destroyed

less frequently than once per week.

PREDICTED PERFORMANCE OF CONTINUOUS FLUORINATORS

Previous data on the extent of removal of uranium from a molten
fluoride salt in a l-in.-diam, open-column fluorinator and recently
obtained data on axial dispersion in open bubble columns were used to
develop a mathematical model for predicting the performance of con-
tinuous fluorinators having diameters ranging from 6 to 12 in. The
results of the analysis are encouraging since they suggest that single
fluorination vessels of moderate size will suffice for removing uranium
from MSBR fuel salt prior to the isolation of protactinium. The ref-
erence MSBR processing flowsheet requires fluorination of fuel salt
at the rate of 170 ft3/day and a uranium removal efficiency of 99%;
the present analysis indicates that an 8-in.-diam fluorinator having

a height of 17.8 ft will meet these requirements.
x1i

MEASUREMENT OF AXTIAL DISPERSION COEFFICIENTS
AND GAS HOLDUP IN OPEN BUBBLE COLUMNS

Measurements of gas holdup and axial dispersion were made in open
bubble columns having diameters of 1, 1.5, 2, 3, and 6 in. for a range
of operating conditions. The effects of changes in the viscosity and
surface tension of the liquid, the superficial gas velocity, the gas
inlet-orifice size, and the number of gas inlets were determined.
These data, as well as data obtained previously, were used to develop
correlations for predicting gas holdup and axial dispersion in open-
column, gas—liquid contactors such as continuous fluorinators in which
a molten fluoride salt is countercurrently contacted with a gaseous
mixture of fluorine and UF,.

6

SEMICONTINUOUS REDUCTIVE EXTRACTION EXPERIMENTS
IN A MILD-STEEL FACILITY

We have continued to operate a facility in which semicontinuous
reductive extraction experiments can be carried out in a mild-steel
system. We are presently studying the mass transfer performance of
an 0.82-in.-ID, 24-in.-long column packed with 1/4-in. molybdenum
Raschig rings. Several experiments were carried out previously in
which a salt stream containing UF4 was countercurrently contacted
with bismuth containing reductant over a range of operating conditions.
In order to measure mass transfer rates in the column under closely
controlled conditions and under conditions where the controlling
resistance is not in the salt phase (as was the case in previous exper-
iments), preparations were begun for experiments in which the rate
of exchange of zirconium isotopes will be measured between salt and
bismuth phases otherwise at chemical equilibrium. Techniques for the
production and charging of 97Zr (half-1life, 16.8 hr) to the salt
were developed, and about 7 mCi of 97Zr was added to the salt in the
feed tank. The first experiment using the 97Zr tracer was interrupted
by a leak in the salt exit line from the feed tank. Because damage

to the feed tank and Calrod heaters on the vessel made salvage of the
xii

tank impractical, a new vessel was fabricated and installed. Examin-
ation of a specimen from the original vessel revealed that, although
some graphitization of the steel had occurred, no evidence of embrittle-

ment was present.,

DEVELOPMENT OF THE METAL TRANSFER PROCESS:
OPERATION OF EXPERIMENT MTE-2

The second engineering experiment (MTE-2) for development of the
metal transfer process was completed. This experiment was performed
at 650°C in a 6-in.-diam carbon steel vessel that was divided into two
compartments interconnected at the bottom by a pool of thorium-saturated
molten bismuth. One compartment contained MSBR fuel carrier salt (72-
16-12 mole % LiF—Ber-ThF4) to which were added 7 mCi of 147Nd and
sufficient LaF3 to produce a concentration of 0.3 mole %Z. The second
compartment contained LiCl, a 35 at. % Li-Bi solution (in a cup), and
a pump for circulating the LiCl through the cup at the rate of about
25 cmB/min. Gas~1ift sparge tubes were used to disperse droplets of
bismuth in the salt phase and thereby improve contact of the phases.
During a 3-month operating period, in which a total of 563 liters of
LiCl was circulated through the cup containing the Li-Bi solution,
more than 857 of the lanthanum and more than 507 of the neodymium were
removed from the fluoride salt. No measurable accumulation of thorium
in the Li-Bi solution (<10 ppm) was noted during this period. The
observed values for the distribution coefficients for lanthanum, neo-
dymium, thorium, and radium during the experiment were in general
agreement with the expected values. From 70 to 100% of the quantities
of the rare earths charged to the system could be accounted for through-
out the experiment. A much greater decrease was observed in the con-
centration of lithium in the Li-Bi solution than was expected; the
reason for this discrepancy has not been determined. Eight days
before the end of the experiment, 1 vol % of fuel carrier salt was
added to the LiCl in order to study the effect of contamination of
the LiCl with fluoride salt. All of the objectives of the experiment

were achieved.
xiii

DEVELOPMENT OF THE METAL TRANSFER PROCESS:
DESIGN OF EXPERIMENT MTE-3

Design of the third engineering experiment for development of the
metal transfer process has been initiated. This experiment (MTE-3) will
use salt and bismuth flow rates that are 1% of the estimated flow rates
required for processing a 1000-MW(e) reactor. Mechanical agitators will
be used for promoting mass transfer between the salt and metal phases in
the experiment. A mathematical analysis was carried out in order to
select approximate equipment sizes and to determine operating conditions
for the system. The experiment will use about 35 liters of MSBR fuel
carrier salt, 6 liters of Th-Bi solution, 6 liters of LiCl, and about
5 liters of Li-Bi solution having an initial lithium content of about
5 at. %Z. The salt-metal contactor will be a 10~in.-diam, two-compartmented

vessel having a mechanical agitator in each compartment.

DEVELOPMENT OF MECHANICALLY AGITATED SALT-METAL CONTACTORS

A program was initiated for the development of mechanically agitated
salt-metal contactors as an alternative to packed columns presently under
consideration for MSBR processing systems. This type of contactor is of
particular interest for the metal transfer process since designs can be
envisioned in which the bismuth phase would be a near-isothermal, inter-
nally recirculated, captive phase. It is believed that such designs will
be less dependent on the technology for molybdenum fabrication than would
a counterpart system based on packed columns. Preliminary tests on the
hydrodynamics of mechanically agitated salt-metal contactors were carried
out using mercury and water. Initially, tests were made using an agitator
that was operated at the water-mercury interface in a manner designed to
disperse the mercury in the water. However, results of these tests led
us to conclude that the contactor should operate under conditions that
minimize dispersion of the mercury. The Lewis contactor appears to have
the greatest potential for achieving effective mass transfer rates with
minimum dispersion of the phases. In this contactor, an agitator, located

well away from the interface, is present in each phase. Each agitator is
Xiv

operated in a manner such that the phases are mixed as vigorously as
possible without actually dispersing one in the other. Information in
the literature on mass transfer rates in Lewis~type contactors was
reviewed. It was concluded that the mass transfer rate correlation
developed by Lewis may be applicable to salt-bismuth systems, and that
adequate mass transfer rates for MSBR processing applications should

be obtained.

HYDRODYNAMICS OF PACKED-COLUMN OPERATION
WITH HIGH-DENSITY FLUIDS

Studies of the hydrodynamics of packed column operation were con-
tinued, using fluids with high densities and a large density difference.
Data were obtained in a 2-in.-diam, 24-in.-long column that was packed
with 3/8-in. Teflon Raschig rings for determining the dependence of
dispersed-phase holdup, pressure drop, and flooding on the viscosity
of the continuous phase. An improved relationship was developed for
predicting packed-column performance during the countercurrent flow
of molten salt and bismuth. The effects of wetting of the packing
by the metal phase on metal phase holdup, flooding, and pressure drop
were also evaluated in a 2-in.-diam, 24-in.-long column packed with 3/8-
in. copper Raschig rings that were wetted by the mercury. The inter-
facial area between the aqueous and mercury phases was decreased sub-
stantially when the packing was wetted, and the column throughput at

flooding was about 40% greater than with nonwetted packing.

ANALYSIS OF MULTICOMPONENT MASS TRANSFER BETWEEN MOLTEN SALTS AND
LIQUID BISMUTH DURING COUNTERCURRENT FLOW IN PACKED COLUMNS

The transfer of materials between a molten salt and liquid bismuth
results in a condition where the fluxes of the transferring ions are
dependent on both concentration gradients and electric potential gra-
dients. This greatly complicates the mass transfer process and makes the
design of continuous reductive extraction columns difficult. A math-

ematical analysis of mass transfer during reductive extraction processes
XV

was carried out to facilitate interpretation of results from present and
proposed experiments in packed columns and as an aid in using these data
for the design of larger reductive extraction systems. A calculational
procedure was developed for solving the resulting relations with as many
as ten transferring materials. Provision was made for calculating rates
of mass transfer between solvent and electrolyte phases for a range of
operating conditions. In future work, particular attention will be paid
to the influence of the electric field on the rate of mass transfer and
to the differences that result from the case where mass transfer rates

are assumed to be dependent only on concentration gradients.

STUDY OF THE PURIFICATION OF SALT BY CONTINUOUS METHODS

Salt purification studies using 66-34 mole 7% LiF—BeF2 were terminated
because of leaks that resulted in the loss of about half of the 1l4-liter
salt charge. The composition of the remaining salt was adjusted to the
approximate composition of the proposed MSBR fuel salt (72-16-12 mole %
LiF—Ber-ThFA). The newly prepared salt was then countercurrently con-

tacted with a H,--10% HF mixture in the packed column in order to remove

oxide from the zalt. Although a considerable quantity of oxide was
removed from the salt, a significant quantity still remained in the
column. In the two flooding runs and one iron fluoride reduction run
that were carfied out during this report period, the pressure drop
across the column increased sufficiently to make operation of the system
difficult., The column was then filled with molten salt, and an HF-H2
stream was allowed to contact the static salt charge for a period of

18 hr in order to remove the oxide from the column. After this opera-
tion had been determined to be successful, eight additional iron flu-
oride reduction runs were completed. Operation of the system was smooth
in each case, and the pressure drop across the column remained low. How-
ever, the results of iron analyses of the salt samples from the runs
were inconsistent. This inconsistency was probably due to the low iron

concentration in the system, although sample contamination was suspected

in some cases.
1. TINTRODUCTION

A molten-gsalt breeder reactor (MSBR) will be fueled with a molten

fluoride mixture that will circulate through the blanket and core

regions of the reactor and through the primary heat exchangers. We

are developing processing methods for use in a close-coupled facility

for removing fission products, corrosion products, and fissile materials

from the molten fluoride mixture.

Several operations associated with MSBR processing are under study.

The remaining parts of this report discuss:

1.

the description and analysis of a flowsheet for isolating

protactinium from MSBR fuel salt by oxide precipitation,

the description of the reference flowsheet for processing
MSBR fuel salt by the fluorination--reductive extraction--

metal transfer process,

an analysis of the importance of the uranium inventory

in a processing plant,

the results of calculations related to the removal of
rare earths from molten LiCl in the metal transfer

process,

experiments conducted in a simulated continuous fluorina-

tor for studying induction heating in molten salt,

predictions of the rate of corrosion of the nickel
vessel in continuous fluorinators employing frozen-

wall corrosion protection,

predictions of the extent of removal of uranium in con-

tinuous fluorinators,

measurement of axial dispersion coefficients and gas
holdup in open bubble columns and the development of

correlations for predicting these quantities,

experiments conducted in a mild-steel reductive

extraction facility, to increase our understanding
10.

11.

12.

1%.

14,

15.

of the rate at which uranium is extracted from molten

salt into bismuth in a packed column,

operation of experiment MTE-2 for demonstrating the
metal transfer process for the removal of rare earths

from MSBR fuel carrier salt,

design of experiment MTE-3 for studying operation of
the metal transfer process using salt and bismuth
flow rates that are 1% of those expected for processing

a 1000-MW (e ) MSBR,

development of mechanically agitated salt-metal con-

tactors,

studies of flooding, dispersed-phase holdup, and
pressure drop during countercurrent flow of liquids
having a large difference in densities in packed

columns,

analysis of multicomponent mass transfer between molten
salts and liquid bismuth during countercurrent flow

in packed columns, and

studies of the purification of salt by continuous

methods.

This work was carried out in the Chemical Technology Division during the

period October through December 1970.
2. FLOWSHEET ANALYSIS: ISOLATION OF PROTACTINIUM
BY OXIDE PRECIPITATION

M. J. Bell L. E. McNeese

Ross, Bamberger, and Baesl have shown that protactinium can be pre-
cipitated selectively as Pa205 from MSBR fuel salt by the addition of
oxide to salt containing Pa>t, and that Pa4+ can be readily oxidized to
Pa5+ by hydrofluorination. Mailen2 has measured the solubility of
PaZO5 in MSBR fuel salt that is saturated with UO2 at temperatures
between 550 and 650°C. Also, Bamberger and Baes3 have found that uranium
oxide can be precipitated from protactinium-free fuel salt as a U02—Th02
solid solution in which the concentration of UO2 at equilibrium is
dependent on the concentration of UF4 in the salt. Bell and McNeese4
have used the equilibrium data of Bamberger and Baes to calculate the
performance of a countercurrent multistage uranium oxide precipitator
and have found that greater than 997 of the uranium can be removed from
fuel salt as a U02-Th02 solid solution that contains less than 107 ThO2
by using only a few equilibrium stages in which the salt and oxide are
countercurrently contacted. These results indicate that oxide precipita-
tion may be an attractive alternative process to fluorination-reductive
extraction for isolating protactinium and removing uranium from the fuel
salt of an MSBR. Two flowsheets that employ oxide precipitation are
described in the remainder of this section, and the effects of several

operating parameters on the performance of the flowsheets are discussed.

2.1 Isolation of Protactinium by Oxide Precipitation, and
Recovery of Uranium Daughters by Fluorination

Figure 1 presents a flowsheet and typical operating parameters for
a process which employs oxide precipitation to isolate protactinium from
MSBR fuel salt and fluorination to recover uranium produced by decay of
the protactinium. Fuel salt is withdrawn from the reactor on a 3-day
cycle, and protactinium is selectively removed by precipitation as Pa205.
The precipitate and a small amount of salt associated with it are hydro-

fluorinated in the presence of a secondary salt that is circulated through
e

 
 

 

     

8.3x104
mole /day

275
mole/day U

  
 

REACTOR
2250 MW (th)
1680 ft3

    
 
   
  

 

 

  

 

 

 

RECOMBINER

 

 

 

 

 

X pof,=8.4x10°

 

 

ORNL DWG 70-14086R1

  

 

 

 

 

URANIUM mdz-:d-” U METAL RARE
REMOVAL |22 0] TRANSFER |—e EARTHS
999, SYSTEM Sr,Ba,Zr,U
UFg

 

'

 

 

 

 

 

 

 

 

 

1
a
8.3x10° mole/day POT . S «
REC'P',A OR1 5.2 molesday OXIDE dOA
60% p— — = o""
. e EFFICIENCY | 200 mole/day SALT n:}f
XpaF;20xI0 0.66 mole/day o=
XyF,=0.0033  UFy T
Hz' HF

    
 

200 mole/day
0.4 mole/day PaF4

a
o Pa DECAY TANK
3 395 moles PoF, &
x [™] 10.0moles UF,
g 0.19x10* moles

= -
3 XyF,=56xI0
F2

 

0.19x10® mole/day ;

Fig. 1. Flowsheet for Isolating Protactinium by Fluorination-Oxide
Precipitation and Removing Rare Earths by the Metal Transfer Process.

T
a fluorinator and a protactinium decay tank. A small fraction of the
salt leaving the decay tank is returned to the primary reactor circuit

to compensate for salt accompanying the oxide precipitate.

The main salt stream exiting from the precipitator vessel contains
most of the fission products and uranium, plus 5 to 40%Z of the pro-
tactinium in the salt leaving the reactor. Ten percent of this stream
is processed to reccover a large fraction of the uranium, and rare earths
are removed from the resulting salt by the metal transfer process. Puri-
fied salt leaving the metal transfer process is combined with the recov-
ered uranium and then returned to the reactor. Removal of the uranium

can be accomplished either by fluorination or by oxide precipitation.

A mathematical analysis of the protactinium isolation portion of
the flowsheet was carried out using the nomenclature shown in Fig. 2.

The following material balance relations can be written for protactinium:

FO = (F + FS)*PEFF:CP1 s (1)
CP2 = CP1* (1 ~ PEFF) , (2)
(F + A VR):CPR = P + F-(CP2 . (3)
A*(VR*CPR + VT-CPT) = P . (4)
and
FO + FS:CP2 = (FS + X+VT)*CPT . (5)
where

FO = flow rate of oxide leaving precipitator, moles/day,
FS = flow rate of salt accompanying oxide leaving precipitator,
moles/day,
F = flow rate of salt leaving the reactor, moles/day,
PEFF = protactinium removal efficiency in precipitator,
r = 233pa decay constant, day ™1,
VR = volume of salt in reactor, moles,
VT = volume of salt in protactinium decay tank, moles,

CP = concentration of protactinium in salt at point denoted by

suffix (defined below), mole fraction.
ORNL DWG 72-13525Ri

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

F
cu2 UFse
Reactor cCP2
v S
Precipitator Mydro- ° Pa Decay
Efficiency - —7 ™ fluorinator }— g S | g Tank
F Cul = PEFF Fo FPS o FPS Volume =VT
CUR CPI F5 cu3 3 cu4
cP2
HF Fa
FPS
CUT
CPT
FS
cuT
CPT

Fig. 2. Nomenclature Used in Mathematical Analysis of Flowsheet for
Protactinium Isolation by Fluorination--Oxide Precipitation.
The suffixes 1, 2, 3, 4, R, and T refer to the following locations in
the flowsheet: 1, entering precipitator; 2, entering hydrofluorinator;
3, leaving hydrofluorinator; 4, entering protactinium decay tank; R,
leaving reactor; and T, leaving Pa decay tank. In the analysis, the
quantities F, PEFF, P, VR, VI, and A are known and the ratio FS/FQ is
specified. The quantities FO, FS, CPl1, CP2, CPR, and CPT are to be
determined. It should be noted that Eq. (4) implies that a negligible
fraction of the total protactinium inventory is in the metal transfer
system and assumes no removal of protactinium except by radioactive
decay. However, as much as 5% of the protactinium inventory may be
present in the metal transfer system in an actual processing plant.

The equations to be solved are nonlinear because of the manner in which
the unknown quantity FS enters the relations. The algorithm developed
to solve Eqs. (1)-(5) involved assuming a value for FO, which fixes the
value for FS and also reduces Eqs. (1)-(4) to a set of four linear
algebraic equations that can be readily solved for the four unknown
concentrations. Equation (5) is then used to improve the estimated
value for FO, and the procedure is repeated until satisfactory con-
vergence in the value of FO is obtained. After the concentrations of
protactinium throughout the system and the quantities FO and FS have
been determined, the concentrations of uranium throughout the system
can be determined by the use of the following analogous set of material

balance relations:

(F + FS)*CU1l = F*CUR + FS-CUT ’ (6)
FS.CUl + (FPS - FS)-CUT = FPS.CU3 s (7)
Cu4 = (1 - H)-CU3 s (8)
and
FPS.CU4 + A-VT+CPT = FPS-CUT , (9)
where
CU = concentration of uranium in salt at point denoted by suffix
defined above, mole fraction,
H = uranium removal efficiency in fluorinator,
FPS = salt flow rate from hydrofluorinator, moles/day.
The quantities H and CUR are assumed to be fixed quantities. In making
the analysis, it was assumed that the volumes of all vessels except the
protactinium decay tank were negligible, and that the flow rates of the

salt streams entering and leaving the fluorinator were equal.

The effects of several parameters on the performance of the flow-
sheet were calculated. As shown in Fig. 3, the protactinium removal
time depends on the precipitator efficiency and the rate at which fuel
salt is transferred to the protactinium decay tank along with the Pa205.
A protactinium removal time of about 5 days can be realized if 60% of
the protactinium is removed from the salt in the precipitator, provided
the salt transfer rate to the protactinium decay tank is as low as 10 to
20 moles/day (a salt-to-oxide flow rate ratio of 2 to 4). If the salt-
to-oxide flow rate ratio were as high as 600, a precipitator efficiency
of about 80% would be required in order to obtain the same protactinium
removal time. The uranium inventory in the decay tank depends on the
efficiency of the fluorinator in the protactinium isolation loop and on
the amount of fuel salt that is transferred to the protactinium decay
tank along with the precipitate, as shown in Fig. 4. The uranium inven-
tory in the decay tank will be only a small fraction of the uranium

inventory in the reactor, and the associated inventory charge will be

less than 0.001 mill/kWhr for a wide range of operating conditions.

2,2 Isolation of Protactinium by Oxide Precipitation Without the Use
of Fluorination for Recovering 233y Produced by Decay of 233p,

As shown in Fig. 5, the isolation of protactinium by oxide precipi-

tation can also be carried out without the use of a fluorinator for

233P

recovering uranium that is produced by the decay of the a. In this

flowsheat, fuel salt is withdrawn from the reactor on a 3-day cycle and
combined with a salt stream that is withdrawn from the protactinium decay
tank. Part of the protactinium in the resulting salt stream is removed
by precipitation as PaZOS' The P3205 precipitate is hydrofluorinated

in the presence of salt exiting from the metal transfer system, and the
resulting salt stream then passes through a decay tank, where part of

233 233U.

the Pa decays to The salt stream leaving the protactinium
Pa REMOVAL TIME(days)

ORNL DWG 70-14098

 

 

 

 

IS I I T T 1 T
Pa CYCLE TIME:3 DAYS
FLUORINATOR EFFICIENCY:95°%%
RATIO OF SALT
FLOW RATE TO
OXIDE FLOW RATE
2 - -
600
9 -
400
200
40
6 -
0
3 i 1 L | 1 L 1 <
20 30 40 50 60 70 80 90 100

Pa PRECIPITATOR EFFICIENCY (%)

Fig. 3. Effects of Protactinium Precipitator Efficiency and Rate of
Transfer of Salt with Precipitate on Protactinium Removal Time for

Fluorination--Oxide Precipitation Flowsheet.
FISSILE U INVENTORY IN Pa DECAY TANK

(% OF REACTOR FISSILE INVENTORY)

ORNL DWG 70-14099R!

 

 
 
   
 
    

 

 

 

T T T ! | | 0.0008
0.26 |-
a 90% PRECIPITATOR EFFICIENCY -
0.24 | DAY SALT RESIDENCE TIME 0.0007
X poF, <0.0022
REACTOR FISSILE INVENTORY=1460 Kg
10% INVENTORY CHARGE
0.22 |-
FLUORINATOR
EFFICIENCY
— 0.0006
0.20
0.18
~-10.0005
0.16 1 | | [ i I 1
0 100 200 300 400 500 600 700 800

RATIO OF SALT TRANSFER RATE TO OXIDE FLOW RATE
(MOLES SALT /MOLE OXIDE)

Fig. 4. Effects of Rate of Transfer of Salt-plus-Precipitate and
Fluorinator Efficiency on Uranium Inventory in Protactinium Decay Tank
for Fluorination--Oxide Precipitation Flowsheet.

hr)

INVENTORY CHARGE (mill/ kW

01
 

 

 

 

 

 

 

 

 

56 tt3/day

 

 

RECONSTITUTION

 

 

 

‘273 mole/day U
5500 mole/day SALT (4 t3)

 

 

ORNL DWG 70-14087R1

 

 

 

RARE
—a EARTHS
Sr, Bo, Zr, U

 

 

 

 

 

 

 

 

 

 

Pa DECAY TANK
150 ft3
U inventory=11.4 kg

 

 

= -8
Xy, *250x10
X PqF, *0.0020

 

2.75
REACTOR 60f3/d URANIUM METAL
2250 MW(th | 60f7/day | RemovAL |-mole/day L) o ANSFER
1680 ft3 99% EFFY SYSTEM
] -6 XpgF,=8.4x107¢
XPaF,20x10 7 s ¢13/day
XyF, =0.0033
4
szOs
Pa 90 mole/day HYDRO ~
PRECIPITATOR ey
6% EFFY FUEL SALT FLUORINATOR
96% 3600 mole/day
Hz-HF
60 ft3/day
Fig. 5. Flowsheet for Isolating Protactinium by Oxide Precipitation

and Removing Rare Earths by the Metal Transfer Process.

11
12

decay tank is combined with the salt entering the precipitator in order

233

to return U to the reactor without including a large quantity of
255Pa.

A mathematical analysis was carried out for the protactinium iso-
lation portion of the flowsheet, using the nomenclature shown in Fig.
6. The concentrations of protactinium at various points in the flow-

sheet are defined by the following material balance relations:

(FPS + F):CPl = F-CPR + FPS-CPT , (10)
FPS-CP3 = FO + FS-CP2 , (11)
P = A-(VR‘CPR + VI-CPT) , (12)
FPS*CP3 = (FPS + A-VT)-CPT , (13)
CP2 = (1 — PEFF)-CPl , (14)
and
(FPS + F):CPl = FO + (FPS + F)-CP2 . (15)

The quantities F, FPS, VR, VI, P, and A are known; the ratio FS/FO is
specified; and the values of CPl, CP2, CP3, CPR, CPT, FO, and FS are to
be determined. The equations were linearized by assuming a value for
FO, and Eqs. (10)-(14) were solved for the five unknown concentrations.
Equation (15) was then used to improve the estimate of FO. As in the
analysis for the earlier flowsheet, Eq. (12) implies that the protac-
tinium inventory in the metal transfer system is a negligible fraction
of the protactinium inventory in the total system. Changes in the

flow rates of the salt streams through the precipitator and the hydro-
fluorinator as the result of precipitation and dissolution of oxide

were neglected.

The concentrations of uranium in the system are defined by the

following material balance relations:
(FPS + F):CUL = F-CUR + FPS-CUT , (16)

FPS-CU3 = FS-CUl + (FPS — FS)-CUL , (17)
ORNL DWG 72-13524

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

< 1
Uranium Removal
and
F Metal Transfer|
Systems FPS-FS
cu4
CP4=0
Reactor
cuz
cpP2
—
wl
Precipitator Hydro - Pa Decay
Efficiency __F:_o_.. fluorinator Tank
= PEFF — »
> FPS Volume sVT

FS cu3

cuz f cP3

cP2
HF

CPT
CuUuT

 

 

 

Fig. 6. Nomenclature Used in Mathematical Analysis of Flowsheet for

Isolating Protactinium Without the Use of Fluorination for Recovering
2330 Produced by Decay of 255Pa.
14

cut = (1 —H)-CUL , (18)

and

FPS:CU3 + A*VI'CPT = FPS-CUT , (19)

where
H = uranium removal efficiency in the step prior to the metal
transfer system,

and the other quantities are as defined previously.

The mathematical model describing operation of the protactinium
isolation system was used to investigate the effects of several operating
parameters on the performance of the flowsheet. The variation of protac-
tinium removal time with changes in the precipitator efficiency and the
fraction of the salt from the metal transfer system which is fed to the
protactinium isolation system are shown in Fig. 7 for a protactinium
processing cycle time of 3 days, a rare-earth processing cycle time of
50 days, a decay tank volume of 150 fti, and a salt-to-oxide molar ratio
of 40 in the stream leaving the precipitator. Since the rate of flow of
salt between the decay tank and the precipitator must be relatively
large in order to limit the uranium inventory in the decay tank, high
precipitator efficiencies are required with this flowsheet. A precipi-
tator efficiency of about 96% would be necessary in order to obtain a
protactinium removal time of 5 days. The uranium inventory in the pro-
tactinium decay tank is relatively sensitive to the amount of salt
accompanying the oxide in the stream leaving the precipitator, as shown
in Fig. 8. For a salt-to-oxide flow rate ratio of 50 (corresponding to
a salt flow rate of ~ 5000 moles/day), the uranium inventory will be
about 1% of the reactor fissile inventory and the associated inventory
charge will be about 0.003 mill/kWhr. However, if separation of salt
from the oxide is difficult and salt-to-oxide flow rate ratios of the
order of several hundred are required (corresponding to a salt flow rate
of about 5 x 1Obr moles/day), the uranium inventory in the protactinium
decay tank will be about 5% of the reactor fissile inventory and the

associated inventory charge will be about 0©.015 mill/kWhr.
Pao REMOVAL TIME (days)

ORNL DWG 70-14097

 

 

 

 

50 I T T I [ T |
% OF SALT FROM METAL
: TRANSFER SYSTEM TO
Pa CY .
o CYCLE TIME:3DAYS o TANK
RARE EARTH CYCLE TIME:30DAYS
40 |~ RATIO OF SALT TRANSFER RATE B
TO OXIDE FLOW RATE:40 100%
90%
5%
\o% _
. N
20 - —
10+ -T
} ] ] i | e | ]
20 30 40 50 60 70 80 90 100

Pa PRECIPITATOR EFFICIENCY (%)

Fig. 7. Effects of Protactinium Precipitator Efficiency and the
Fraction of Salt Fed from the Metal Transfer System to the Protactinium
Removal System on Protactinium Removal Time.
®22U Inventory in Pa Decay Tank (% of Reactor Fissile Inventory)

ORNL DWG 72-13528

 

 

 

 

 

T T T T I T
Po Processing Cycle: 3 days ' 0.020
70 Rore Eorth Processing Cycle: 30 days
r Pa Precipitator Efficiency : 99 %
Reactor Fissile inventory : 6000 moles
60 Inventory Charge: [0 percent per annum
' , 3
Decay Tank Volume: 150 ft 40015
Percent of Salt from Metal
50 Transfer System to
Decay Tank
—~0.010
—~ 0.005
0 1 1 ] ] ] 1 1 0
o 50 100 150 200 250 300 350 400

Fig. 8.

Ratio of Salt Tronsfer Rate to Oxide Flow Rote

(Moles Salt / Mole Pa,0,)

Transfer System to Decay Tank.

Effects on Uranium Inventory of Salt-—to—Pa2
in Stream Leaving Precipitator and of the Percent of Salt

Q0. Molar Ratio

gent from Metal

Inventory Charge (mill /kWhr)

21
17

5. FLOWSHEET ANALYSIS: REFERENCE PROCESSING PLANT FLOWSHEET
BASED ON FLUORINATION, REDUCTIVE EXTRACTION,
AND THE METAL TRANSFER PROCESS

L. E. McNeese

Previously, we described an improved method for removing rare earths
from the fuel salt of a single-fluid MSBR and presented calculated results
5,6

on the system performance for a range of operating conditions. More

T

recently we described an improved method for removing protactinium from
fuel salt. This method is based on the use of fluorination for removing
uranium and reductive extraction for removing protactinium. The isoclated

253Pa is held for decay to 253

U in a secondary molten fluoride salt. We
also devised a method for combining the various wastes streams produced
during the isolation of protactinium and during the removal of rare

T

earths into a single stream. During this report period, we combined
these three processing methods into a single flowsheet and adopted a
set of operating conditions that constitute the reference fluorination--

reductive extraction--metal transfer flowsheet.

The reference flowsheet is shown in Fig. 9. For a 1000-MW(e ) MSBR,_

fuel salt (71.7-16-12-0.33 mole % LiF-BeF —ThFh_UFM) is removed from the

reactor at the rate of 0.88 gpm, which reiresents a 10-day processing
cycle. The salt passes through a delay vessel, which results in an
average decay period of 30 min, before being contacted with fluorine at
the rate of 19.3 liters (STP)/min in a continuous fluorinator (employing
frozen-wall corrosion protection) to remove 99% of the uranium from the
salt stream. The fluorine feed rate is equivalent to 150% of that
required for conversion of the UFH to UF6. The fluoride salt stream
leaving the fluorinator is contacted with a 28.3-liter/min hydrogen
stream in order to remove dissolved fluorine from the salt and to reduce
the valence of the residual uranium from 5+ to 4+. The salt stream is
then contacted countercurrently with a 0.072-gpm bismuth stream containing
0.011 equiv of reductant per gram-mole of bismuth in a salt-metal contac-
tor (which is equivalent to five theoretical stages) in oraer to extract

the protactinium and uranium from the salt. The bismuth stream leaving
 

18

ORNL DWG 72-13529

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3

 

Fiuorinator

 

 

 

Salt 4+ Fission Products
To Waste

Be F,
HeHF Th F,
b1 ] :
Salt | UF, UF, J )
Purification Reduction Absorber -:
i
! .
Hy :
|
k l .
UFs-F, Mo e m e - o =5 :. ______ .
1
| |
Reactor| 1 I
| 1
! |
Hy-HF 1 . 1
"___l Li |
! : _y__
- Soomesny [
D.lfly Fluorinator parger r -@ ; é
Vessel ' LiCl : |
1 [ H,-HFl : :
: :
F' H. T ' l —{ E E :
Hydro- Pa P
fiuerinator UFe=Fa Hg-HF Decay Podeeemee Loo® |
Tank T 1
T ! !
I _J
HF | Fluorinator—» Sparger : !
I
1
1 |
? T I i
F, H ! !
* ' | HF-H.I ' ! |
t e !
. Waste Hydro - 1 \
®) Holdup fluorinater 1 I
I
T I
s
? t.a !
HF 1
I
1
1
I
|
!
|
|
|
'

b

Fig. 9. Reference Flowsheet for Processing MSBR Fuel Salt via Fluo-
rination, Reductive Extraction, and Metal Transfer.
19

the salt-metal contactor is subsequently contacted with an 11.8-1iter/
min HF stream in the presence of a 0.78-gpm secondary fluoride salt
stream in order to transfer the extracted materials (protactinium,
uranium, and zirconium) to the secondary salt. After leaving the
hydrofluorinator, the secondary salt passes through a fluorinator,
where 90% of the uranium is removed as UF6 by contact with a 1.56-liter/
min fluorine stream. (The assumed value for the fluorine utilization
during this step is 1T7%.) The secondary salt is then contacted with

a 2.8-liter/min hydrogen stream in order to remove dissolved fluorine
and to reduce the valence of the residual uranium from 5+ to L+.
Finally, the salt passes through the protactinium decay tank, where

2
53Pa present in the reactor and processing plant is held

253,

most of the
for decay to . The 0.78-gpm secondary salt stream that is fed to

the hydrofluorinator is taken from the protactinium decay tank.

The fuel carrier salt stream leaving the protactinium extraction
column is essentially free of uranium, protactinium, and zirconium;
however, its rare earth concentration is about the same as that in the
reactor. This stream is fed countercurrent to a 12.%-gpm stream of
bismuth containing lithium and therium (0.0l11 equiv of reductant per
mole of bismuth) in a three-stage contactor in order to extract frac-
tions of the rare earths from the salt. The bismuth stream leaving
the extractor contains the rare earths, thorium, and lithium; this
stream is countercurrently contacted with a 33.%-gpm LiCl stream in
a three-stage salt-metal contactor. Because of highly favorable dis-
tribution ratios, the rare earths, along with a negligible quantity
of thorium, transfer to the LiCl. The trivalent rare earths are
removed from the LiCl in a single-stage contactor by contact with an
8.1-gpm recirculating bismuth stream that contains 5 at. % lithium.
The net flow rate of the Li-Bi solution through the contactor is 21.5
liters/day. Two percent of the LiCl stream exiting from the trivalent
rare-earth contactor is fed to a two-stage salt-metal contactor, where
the divalent rare earths are removed by contact with a 0.67-gpm recir-
culating bismuth stream that contains 50 at. % lithium. The net flow
of the Li-Bi solution through the contactor is 2.2 liters/day. All of

the extraction operations are carried out at 640°C.
20

The bismuth stream exiting from the divalent rare-earth contactor is
fed to the hydrofluorinator below the protactinium extractor in order to
add lithium to the protactinium decay tank at the rate of about 50 moles/
day. With this addition, the composition of the salt in the decay tank
is 79.6-17.7-2.12 mole % LiF -ThF) -Z1F), 0.367 mole % divalent rare-earth
fluorides, and 0.247 mole % of trivalent rare-earth fluorides. The salt
has a liquidus temperature of about 600°C; at this temperature, the rare-
earth fluoride concentration is well within solubility limits. Salt is
removed from the protactinium decay tank at the average rate of (0.094
fti/day to eliminate the fluorides of lithium, thorium, zirconium, and
the rare earths that accumulate in the secondary fluoride salt. About
0.12% of the fuel carrier salt leaving the rare-earth extractor is dis-
carded as a means of removing LiF that is added durine the extraction of
protactinium and the rare earths. The discards of fuel carrier salt and
secondary fluoride salt from the protactinium decay tank are made on a

220-day cycle. During each cycle, 20.7 ft5

of the secondary fluoride
salt and 44.2 ft3 of the processed fuel carrier salt are transferred to
the waste-salt holdup tank. The Li-Bi stream leaving the trivalent rare-

earth stripper is hydrofluorinated in the presence of the resulting salt

mixture during the 220-day period that is allowed for decay of the 255Pa

in the waste salt. The salt in the waste holdup tank would be fluorinated
253 '

in order to recover the

tion of the waste salt would be 76.3-12.%-9.8-0.64 mole % LiF -ThF, -BeF .-

U before the salt is discarded. The composi-

ZrF , 0.864 mole % trivalent rare-earth fluorides, and 0.114 mole %
divalent rare-earth fluorides. Although the liquidus temperature of the
salt is near 550°C, the salt temperature would have to be maintained at
about 600°C to prevent precipitation of trivalent rare-earth fluorides.
Waste salt would be discarded from the processing system at the rate of

70 ft5 every 220 days.

The 0.072-gpm bismuth stream leaving the hydrofluorinator below
the protactinium extractor is combined with reductant (200 moles of
7Li per day ), and the resulting stream is effectively returned to the
protactinium extractor. The stream is actually combined with the 12.3%-

gpm bismuth stream that circulates through the rare-earth removal system,
21

and an equal quantity of material is removed and fed to the protactinium
extractor; this mode of operation allows for the removal of materials
that would tend to accumulate in the otherwise captive bismuth phase in
the rare-earth removal system. A small quantity of bismuth (2.2 liters/
7

day ) is combined with sufficient 'Li to produce a stream containing 50

at. % lithium for return to the divalent rare-earth extractor.

The processed fuel carrier salt remaining after discard of 0.12% of
the salt stream exiting from the rare-earth extractor is combined with
sufficient quantities of BeF and ThF)_L (47.8 and 47 moles/day, respec-
tively) to make up for the discard of these materials in the waste salt
and burnup of thorium in the reactor. The resulting fuel carrier salt
is then fed to the fuel reconstitution step, which is carried out in a
vessel having two compartments. To the first compartment are fed the
63-37 mole % UF6—F2 stream from the fluorinators (at the flow rate of
20.9 liters/min), the processed fuel carrier salt (at the rate of 0.88
gpm), and a 1.7-gpm fuel salt stream that is recycled from the second
compartment. The rate at which fuel salt is recycled from the second
compartment is such that the quantity of UFM in the recycled fuel salt
is sufficient to give an average uranium fluoride valence of 4.5 (an
equimolar mixture of UFu and UF5)~in the resulting 2.58-gpm salt stream
leaving the first compartment. The 2.58-gpm salt stream leaving the
first compartment is contacted with a 12.8-liter/min hydrogen stream
in the second compartment in order to reduce the UF5 to UFM' A hydrogen
utilization of 50% is assumed during this operation. Fuel salt is with-
drawn from the second compartment for return to the reactor at the rate

of 0.88 gpm, and the remaining salt is recycled to the first compartment.

Before being returned to the reactor, the fuel salt is contacted
with a 12.8-liter/min hydrogen stream to effect reduction of 1% of the
UFM to UFB- The salt is also contacted with nickel wool as a means of

removing traces of bismuth from the salt before its return to the

reactor.

The protactinium removal time obtained with the reference flowsheet
is 10 days; the rare-earth removal times range from 16 to 50 days. The

calculated value for the breeding ratio is about 1.065.
22

L. FLOWSHEET ANALYSIS: IMPORTANCE OF URANIUM INVENTORY
IN AN MSBR PROCESSING PLANT

M. J. Bell L. E. McNeese

The MSBR processing flowsheets considered in the past have uni-
formly resulted in very low uranium inventories in the processing
plant, that is, inventories below 1% of the uranium inventory in the
reactor in most cases. Since several potential processing systems
might result in uranium inventories as large as 5 to 10% of the reactor
inventory, we have investigated the importance of increases in the
uranium inventory. The major effects of an increased uranium inventory
are: (1) an increase in inventory charges on fissile material, and
(2) an increase in the reactor doubling time. The variation of each
of these quantities with processing plant uranium inventory is shown

233

in Fig. 10. The fissile inventory (which includes the Pa in the

processing plant ) was assumed to be 1504 kg for the reference reactor

Z
253y was taken to be $14/g, and

and processing plant, the value of
the capital charge rate was assumed to be 10% per year. The calcu-
lated system doubling time for the limiting case of a zero uranium
inventory in the processing plant is 22 years. It is seen that a
processing plant uranium inventory of 5% of the system fissile inven-
tory would increase the fuel cycle cost by 0.015 mill/kWhr and would
increase the system doubling time from 22 to 23.1 years. A uranium
inventory of 10% of the system fissile inventory would result in a
fuel cycle cost increase of 0.03 mill/kWhr and an increase in doubling
time from 22 to 24.2 years. Thus, while there is incentive for main-
taining a low uranium inventory in the processing plant, it does not
appear that a uranium inventory as high as 5 to 10% of the system

fissile inventory would rule out an otherwise attractive processing

system.
INCREMENTAL INVENTORY CHARGE (mill/kWhr)

ORNL DWG 71-2860

 

 

 

 

0'025 T ]’ T l T r v ' v I v ' ¥ l T
0.020 |-
0.015
— ———
0.010
SYSTEM FISSILE INVENTORY ... 1504 kg
VALUE OF 233y ... $14/9
0.005 - CAPITAL CHARGE RATE........... 10% /year
o 1 4 ] L ] ) ] . 1 3 ] g ] 3
0 f 2 3 4 5 6 7
% OF SYSTEM FISSILE INVENTORY
Fig. 10. Effects of 255U Inventory in Processing Plant on System

Doubling Time

and Inventory Charges.

25

24

23

22

21

20

DOUBLING TIME (years)

¢c
2L

5. FLOWSHEET ANALYSIS: REMOVAL OF RARE-EARTH FISSION PRODUCTS
FROM LiCl IN THE METAL TRANSFER PROCESS

M. J. Bell L. E. McNeese

A flowsheet has been described previously8’9 for removing rare-
earth fission products from MSBR fuel salt using the metal transfer
process, which employs contact of LiCl with Li-Bi solutions for
removing the rare earths and other fission products. 1In the reference
flowsheet, the trivalent rare earths are removed by contacting LiCl
at the rate of 33.4 gpm in a single equilibrium stage with an 8.1-
gpm recirculating bismuth stream having a lithium concentration of
S5 at. %. Bismuth containing extracted rare earths is withdrawn at
the rate of 5.7 gal/day, and an equivalent amount of Li-Bi solution

is added.

Early data indicated that mutual solubility problems might be
encountered between thorium and trivalent rare earths in bismuth having
a lithium concentration as high as 5 at. %. Although this has been
found not to be the case,10 we have made calculations to determine the
effect that varying the lithium concentration in the lithium-bismuth
alloy has on the thorium concentration in the metal and on the
reactor performance. The results are presented in Fig. 11, which
shows the effect of increasing the flow rate of the Li-Bi withdrawal
stream while holding the amount of added reductant constant. The
thorium concentration in the metal is reduced from 420 ppm at the
reference withdrawal rate of 1000 moles/day to 1LO ppm at the discard
rate of 3000 moles/day. The effect on reactor performance is slight;
the breeding ratio decreases from about 1.063 to about 1.060. It
would be possible to compensate, in part, for this loss in breeding
ratio by operating the trivalent rare-earth stripper as a once-through

batch contactor having more than one stage.

In the reference flowsheet, the divalent rare-earth fission prod-
ucts are removed from the LiCl by contacting 2% of the LiCl leaving
the trivalent rare-earth stripper with a 50 at. % Li-Bi solution at

the flow rate of 0.56 gal/day in a two-stage contactor. To accommodate
MOLE FRACTION Th in Li-Bi SOLUTION (x 10%)

400

ORNL DWG 7i-2859R1

MOLE FRACTION Li IN METAL
0.05 0.033 0.025 0.020 0.0167

 

 

 

 

Li-Bi FLOW RATE (moles/day)

Fig. 11. Effect of Lithium-Bismuth Discard Rate on
tration in Discard Stream and on MSBR Performance.

T T T T T 0.064
50 MOLES/DAY Li METAL IN Bi
o 640°C 10063
— —— e - 0.062
p
- -1 0.061
- -1 0.060
" L . A A A \ 0.059
1000 2000 3000

Thorium Concen-

BREEDING GAIN (BREEDING RATIO-1.0)

Gc
26

the high heat generation rates expected in the resulting Li-Bi solution,
it is advantageous to operate the contactor as a continuous column
through which the Li-Bi solution is recycled at a relatively high flow
rate. A large fraction of the fission product decay heat could be
dissipated by placing a decay tank in the recycle stream. However,

this system would result in a contactor having only a single equilibrium
stage. A calculation was made of the system performance for a flowsheet
in which the divalent rare-earth contactor consisted of only one equilib-
rium stage. Removal times for the divalent rare earths were increased
only slightly over those obtained with a two-stage column; the resulting
decrease in the breeding gain of the reactor was negligible. Therefore,
the mode of operation described above has been adopted as part of the

reference MSBR processing flowsheet.
27

6. FROZEN-WALL FLUORINATOR DEVELOPMENT: EXPERIMENTS ON
INDUCTION HEATING IN A CONTINUOUS FLUORINATOR SIMULATION

J. R. Hightower, Jr. C. P. Tung

An experiment to demonstrate protection against corrosion by the
use of a layer of frozem salt in a continuous fluorinator requires the
use of a corrosion-resistant heat source in the molten salt. High-
frequency induction heating has been proposed as the source of heat, and
the estimated performance11 of a frozen-wall fluorinator having an induc-
tion coil embedded in the frozen salt near the fluorinator wall has
indicated that this may be an acceptable heating method. However, there
are uncertainties associated with the effect of bubbles that will be
present in the molten salt on the heat generation rate and in the amount
of heat that will be generated in the metal walls of the fluorinator.
Equipment has been installedl2 for studying heat generation in a simu-
lated frozen-wall fluorinator by induction methods. In the simulation,
a 31 wt Z HNO3

molten salts is used to simulate molten salt in the fluorinator vessel.

solution with electrical properties similar to those of

Results of the first eight experiments with the fluorinator simulation

are described in the remainder of this section.

6.1 Experimental Procedure

A simplified flow diagram for the continuous fluorinator simulation
is shown in Fig. 12; the system is shown in greater detail in Fig. 13.
The acid drain tank was filled with 28.4 kg of 31 wt % HNO3 that was used
during the initial operating period and also during the first eight runms
that were carried out. In each run, the acid recirculation system was
filled with acid by pressurizing the drain tank to about 6 psig with
valve V-7 open (see Fig. 13). When the liquid level in the column (as
indicated by the level in the sight glass) reached the level of the top
of the jacketed pipe, valve V-7 was closed and the pressure in the drain
tank was reduced to 1 atm. The acid recirculation pump and the cooling

water were turned on, and all flow rates were adjusted to the desired
ORNL DWG 70-8964 RI

28

AIR

COOLING
WATER

PUMP

 

 

HEAT
EXCHANGER

COCGLING
WATER

JACKETED
PIPE

SN

S S e N AAIRIA RN R SAINSEN

Il

\N\\\NN\\NNN\\\\N\\N\NNNN\NNNN\\\NN\\\N\HN\KNNN\\\NN\\\\\\\NI\\

 

      

bDbDDDDDDDbDDDDDDDD\

=
/

 

 

 

 

 

 

 

U U U U000 0D vy Oy Y g O U 0O U
~~~~~~~~~~~~~~~~~~~ i S e 0 2 AP W N 3
2 mul‘ flflflflflflflflflflflflflflflflflflflfl LTt T LTt T C TS TTTS
ns )
25
a3
o
o

 

NITRIC ACID

9

INDUCTION
COIL

 

 

Simplified Flow Diagram of a Fluorinator Simulation in

 

ig. 12.
Which Heat Is Generated by Induction Heating.

COOLING
WATER
29

ORNL DWG 70-14719

OFF—-GAS

P |

 

 

 

 

  
  
  

     
    
   

 

 

 

 

s
FI-4
V-4 GLASS
COLUMN | HEAT
INDUCTION EXCHANGER
COOLING
WATER JACKETED < Y
T PIPE
INDUCTION § P
COIL
< h
LS 9
10
« X
Fl-2 Y o

 

 

 

 

 

 

 

COOLING
WATER
TO

JACKETED
PIPE

 

 

..n.-J.n-A-y—-- - o

 

 

 

DRAIN
TANK

Fig. 13. Flow Diagram of Fluorinator Simulation and Nitric Acid
Recirculation System.
350

values. The radio frequency (rf) generator was then started, and the
current to the induction coil was adjusted to a value between 100 and
160 A (rms value). The flow rates and the coil current were maintained
at constant levels by manual adjustment until the temperatures around
the recirculation system became steady. Steady-state temperatures were
usually achieved within 1 hr of operation with constant values for the
flow rates and coil current. After steady-state conditions had been
established, the run was completed by recording the operating conditions.
The rf power and the acid recirculation pump were then turned off, and

the acid was drained from the recirculation system by opening valve V-7,

6.2 Experimental Results

Eight runs have been carried out with the continuous fluorinator
simulation to determine heat generation rates in the acid, in the pipe
surrounding the acid column, and in the induction coil. Heat generation
rates were calculated from the steady~state increase in the temperature
of the acid as it passed through the column and in the temperature of
the cooling water as it passed through the coil and the jacketed pipe.
The experimental results for the first eight runs with the system are
summarized in Table 1. During these runs, the induction coil consisted
of 17 smaller coils arranged with alternate coils wound in opposite
directions. Each coil section consisted of about 6.5 turns of 1/4-in.-
diam Monel tubing that had been wound on a 5.6-in.-diam mandrel. The
total length of the 1l7-section coil was 5 ft. The frequency of the rf

current was 412 kHz during each of the runs.

The heat generation rates in acid in the 5-ft-long column section,
in the pipe representing the fluorinator wall, and in the induction
coil were proportional to the square of the total coil current as expected.
The ratios of the heat generation rates (in watts) to the square of the
coil current (rms value, in amperes) were designated as effective resist-
ances that will be used subsequently in circuit calculations for design
of the rf generator. The average effective resistance for the acid was
0.0161 & at average acid temperatures of 25 to 29°C and 0.0177 Q for

average acid temperatures of 46 to 51°C. The difference between these
2]

Table 1. Summary of Results from Experiments on Induction Heating in a
Simulated Continuous Fluorinator

 

 

 

 

Total Average
Run Coil Acid Heat Generation Rate Equivalent Resistance
No. Current Temp. (W) ()
(CFS-) (A) (°C) Acid Pipe Coil Acid Pipe Coil
1 24.1 242 140
2 130 24.6 409 167 0.0242 0.0099
3 150 26.5 316 279 1442 0.0140 0.0124 0.0641
4 100 25.6 141 116 0.0141 0.0116
150 28,1 378 263 0.0168 0.0117
150 29.1 383 309 0.0170 0.0137
5 160 26.7 448 308 0.0175 0.0120
6 150 46.5 393 282 0.0175 0.0125
7 150 28.9 377 273 1356 0.0167 0.0121 0.0585
120 26.6 235 178 870 0.0163 0.0126 0.060
8 150 51.4 402 275 1442 0.0178 0.0122 0.0641

 

values may be within experimental error. The effective resistance of

the acid should have increased by about 257Z as the temperature was
increased from the lower to the higher value; however, the observed
increase was only about 10%. The smaller variation in effective resis-
tance with changes in temperature may be due to the use of short coil
sections since the effective resistance of the acid should be proportional
to the specific conductance for an infinitely long coil. The average
effective resistance for the stainless steel pipe surrounding the acid

was 0.0123 Q; the average effective resistance of the induction coil

was 0.0617 Q.
52

The fraction of the total heat generation that occurred in the acid
can be found by dividing the resistance of the acid by the sum of the
resistances of the acid, pipe, and coil. About 197 of the total heat
generatjon occurred in the acid during these runs. A 5-ft-long fluori-
nator having a 4.7-in.-diam molten zone would require generation of about
8.3 kW of heat in the molten salt.11 If an efficiency of 197 were

obtained for heating the molten salt, a 43.7-kW generator would be required.

The calculated values for the resistances of the acid and coil under
the experimental conditions are 0.135 and 0.017 & respectively. The dis-
crepancies between the calculated and measured values probably result
from the use of a coil composed of a number of short sections and the

use of equations that were derived for infinitely long coils.

We are presently evaluating alternative coil designs with a view

toward improving the heating efficiency.
55

T. PREDICTED CORROSION RATES IN CONTINUOUS FLUORINATORS
EMPLOYING FROZEN-WALL PROTECTION

J. R. Hightower, Jr.

Nickel is the preferred material of construction for fluorinators
in MSBR processing plants since it exhibits greater resistance to attack
by gaseous fluorine than other candidate materials. This resistance is

due to the formation of a tightly adherent film of NiF,_ (a corrosion

product ) through which additional fluorine must diffusi in order to
cause further corrosion of the nickel. The rate of attack is greatly
reduced, although not to zero, once the protective film of NiF2 is
formed on the exposed nickel surface. 1In the proposed continuous
fluorinators, a'layer of frozen salt will be formed on the fluorinator

wall in order to prevent the protective NiF,. film from being dissolved

2
by the molten salt. 1In the present analysis, no credit is taken for
resistance to diffusion of fluorine which may be afforded by the layer

of frozen salt. Since the protective NiF, film will likely be destroyed

2
several times during operation of a fluorinator, an analysis has been
carried out for estimating the corrosion rate under conditions such

that the NiF_ film is destroyed periodically.

2

-1 Data on the Rate of Corrosion of
Nickel in Gaseous Fluorine
Data relative to the corrosion of Ni-200 and Ni-201 in gaseous
fluorine at a pressure of 1 atm were collected from the literature.
A summary of this information is given in Tables 2 and 3. It has been

15

shown - that the reaction of high-purity nickel with fluorine initially
follows a parabolic rate law at temperatures of 300 to 600°C but that,
after a period of time, the reaction rate decreases and follows a third-
or higher-order rate law. The literature-derived corrosion-rate data
were used to calculate rate constants for the reaction of fluorine with
nickel in the temperature range of 360 to TO0°C under the assumption
that the reaction rate follows a parabolic rate law at all times. The
assumption of a parabolic rate law correctly represents the initial

reaction rate but should yield high estimates of the corrosion rate

for long periods of exposure.
3L

Table 2. Calculated Parabolic Reaction Rate Constants for
Ni-200 Exposed to Gaseous Fluorine

 

 

Depth of Exposure Calculated
- Parabolic Rate Constant

Temp. Penetration Time 1/2 a
°c) (mils) (hr) (mils/hr ) Reference
362 0.0048 120 0.000438 [14]1 (5)
400 0.00681 33.3 0.00118 [13}

400 0.00701 50 0.000991 [15]

475 0.0187 240 0.00242P [16]

475 0.0375 480 0.00342P [16]

475 0.0750 960 0.00484b [16]

500 0.0132 33.3 0.00229 [13]

500 0.0144 5 0.00643 [17] (27)
500 1.97 240 0.1272 [17] (25)
500 0.0754 50 0.01066 [171 (23)
500 0.0206 150 0.00168 [17] (13)
500 0.043 150 0.00352 [17] (13)
500 0.0243 9 0.0081 [14]

535 0.181 120 0.0165 [17] (18)
550 0.00299 6 0.00122 (141 ()
550 0.192 120 0.0175 [14]  (5)
600 0.0616 33.3 0.01067 [13]

600 0.00684 5 0.00306 [171  (27)
600 0.0657 96 0.00671 (17}  (20)
600 0.315 96 0.0322 [171  (20)
600 0.00404 8 0.00143 [(17)  (24)
600 0.0931 8 0.0329 [17] (24)
600 0.753 50 0.1065 (171  (23)
600 0.00887 13 0.00246 [18]

600 0.0736 28 0.0139 {18]

600 0.0417 28 0.00788 [18]

600 0.0287 77 0.00327 (18]

600 0.0171 77 0.00195 rig]

600 0.134 93 0.0139 [18]

600 0.0744 93 0.00772 [18]

600 0.0416 124 0.00374 (18]

600 0.0307 124 0.00276 [18]

600 0.162 132 0.0141 (18]

600 0.0936 132 0.00815 [18]

600 0.0690 195 0.00494 [18]

600 0.0628 195 0.0045 [18]

600 0.104 243 0.00666 (18]

700 0.103 8 0.0364 (17]  (24)
700 0.0274 8 0.00969 [17]1 (24
700 3.28 240 0.212 [171 (25
700 0.727 5 0.325 (171 (@27

 

a . . :
The numbers shown in brackets are primary references; the numbers shown in parentheses
are reference numbers in the primary references.

b
Measured in F2—N2 (50-50%) mixture; rate constant corrected to 100% Fz.
35

 

 

Table 3. Calculated Parabolic Corrosion Rate Constants
for Ni-201 Exposed to Gaseous Fluorine
Calculated Parabolic
Depth of Exposure

Temp. Penetration Time Rate Coniignt

(°c) (mils) (hr) (mils/hr™/ <) Reference”
. 380 0.00085 5 0.000380 [14] (5)
500 0.00959 5 0.00429 [17] (27)

500 0.00899 6 0.00367 [14]

550 0.0115 5 0.00514 [14] (5)

600 0.00259 5 0.00116 [17]1 (5)

600 0.0471 28 0.00891 [18]

600 0.0283 28 0.00534 [18]

600 0.0936 7T 0.01067 [18]

600 0.235 77 0.0268 [18]

600 0.109 93 0.0113 [18]

600 0.111 93 0.0115 [18]

600 0.142 132 0.012k4 [18]

600 0.160 132 0.0139 [18]

700 0.0595 5 0.0266 [(17] (27)

 

a . .
The numbers shown in brackets are primary references; the numbers shown
in parentheses are rjference numbers inm-the primary references.

of nickel exposed to gaseous fluorine is described by the following

relat

where

For the assumption of a parabolic rate law, the extent of corrosion

ion:

0.
it 1l

W
1

d =kVt 3

depth of nickel attacked by F2, mils,

(20)

time of exposure of nickel metal to fluorine after a zero

film thickness, hr,

parabolic rate constant, mil hr

1/2_
36

The rate constant values calculated from the literature data are shown
in Tables 2 and 3, along with information on the length of exposure of
the nickel specimens to fluorine, the extent of attack on the specimens,

and the references from which data were obtained.

It was assumed that the temperature dependence of the parabolic
rate constants would be of the Arrhenius type; thus the calculated

values were fitted to the following equation:

In k = B + A/T , (21)

where
T = temperature, °K,
A,B = constants,

with the criterion that the best fit occurred when the quantity

]2
Z[ln k — 1n kobs

was minimized. This criterion places more importance on reaction rate
constant values resulting from low corrosion rates than does the usual
criterion that the quantity

)2

E:(k - kobs

have a minimum value. This procedure was used since it was believed
that experimental errors were likely to yield corrosion rates that

were too high rather than rates that were too low.

The resulting equations for the variation of the parabolic rate

constants for Ni-200 and Ni-201 are as follows:

In k = 0.3773 — 3961/T , (22)

I

and

In k = 4.308% — 7836/T , (23)
57

where

-1
parabolic rate constant, mil hr /2,

o
1t

T = temperature, °K.

The variation of these constants with temperature is shown in Figs. 14
and 15.

7.2 Predicted Corrosion Rates

If it is assumed that the protective NiF2 film is removed n times
per year at equal time intervals, the extent of corrosion experienced

during a l-year period is given by the expression:
d = n kV/8760/n (2k)

where

d

I

corrosion rate, mils/year,

n = number of times NiF, film is destroyed annually.

2

Figure 16 shows the variation of the average corrosion rate with fre-

quency of destruction of the NiF, fium at 450°C, the approximate wall

temperature that will be used inea frozen-wall fluorinator. If the
NiF2 film were destroyed 52 times annually, the average corrosion
rates at this temperature would be 2.9 mils/year and 0.97 mil/year
for Ni-200 and Ni-201 respectively. It appears that Ni-201 is more
resistant to corrosion by fluorine than Ni-200. However, either of
these materials will show satisfactory corrosion resistance if the

NiF2 film is destroyed less frequently than once per week.
33

ORNL DWG 72-353|

Temperature (°C )

 

 
  
 

 

 

 

100 €50 00 5850 800 480 400 380
e T T Ty Ty T T T T r [y T T T T I -
- -
L .
- . -
*
o L . i
: —
a ' ]
—~
w 0.0 - -
< ~
EF 2 ]
- - -
£ - : 1
2 [~ . j
- [~ E
S .
: b * —ef
c .
o .
© . .
o 0001 1 o -
< - -
x ~ .
o - -
= L . -
o
a » -
2
9 L -
a
0000t | Variation of Rate Constant with Temperature for Corrosion fi
E of Ni-200 in Fluorine 3
1 ] 1. 1 I 1
0.0000i 1.00 1o .20 .30 1.40 1.80 .60 170
1000 / T°K

Fig. 1l4. Corrosion Rate Constant for Ni-200 as a Function of
Temperature.
59

ORNL DWG 72-13530

TEMPERATURE (°C)

 

 

 

 

 

100 0 s00 $80 so0 480 400 380
1o ‘
- TrT LI DL L | T 17T 1
F T T 1 | [ 3
- vl
0.1 L ~
" i
- o
- -1
0.0i - —
-~ N -
-
= -
~ | -
E —
x
o
=
o 0.00I - “1
° C -
- - .y
c - i
o N -
o [ i
° b -
e
o b -1
@
O - -—
)
L
O o.0001 | 7
o - 3
a — -~
- w—
| -
- e
- —
= —
! | | ] l 1
000001, 56 110 120 1.80 i.40 150 .60 i.70
1000 / T°K

Fig. 15. Corrosion Rate Constant for Ni-201 as a Function of
Temperature.
ORNL DWG 72-13527

 

o

, (mils /year)

Lo

g% Corrosion Rate
o

Avera

o
w

[ T T TTTTI i P T T TTTT

 

F0% T TTT]
W —

o1

| 1111

 

 

] 1 ] 111 I {
[] 10 30

I
80 100

n, times NiF, film is destroyed annually

Fig. 16.
Corrosion Rates of Ni-200 and Ni-201 at 450°C.

 

Effect of Frequency of Destruction of NiF. Film on the

2
L1

8. PREDICTED PERFORMANCE OF CONTINUOUS FLUORINATORS

L. E. McNeese J. S. Watson
T. O. Rogers

Most of the flowsheetslg_22 considered to date for processing MSBR
fuel salt require fluorination of molten salt for removal of uranium at
one or more points., These applications include:

(1) removal of trace quantities of uranium from relatively

small salt streams prior to discard,

(2) removal of uranium from a captive salt volume in which 233Pa

is accumulated and held for decay to 233U

>
(3) removal of most of the uranium from relatively large fuel
salt streams prior to isolation of protactinium and removal
of rare earths, and
(4) nearly quantitative removal of uranium from a salt stream
containing 233Pa in order to produce isotopically pure 233U.
Not all of these applications require continuous fluorinators; in fact,
the use of batch fluorinators results in definite advantages in certain
cases. However, as the quantities of salt and uranium to be handled
increase, the use of continuous fluorinators becomes mandatory in order
to avoid undesirably large inventory charges on uranium and molten salt
as well as the detrimental increase in reactor doubling time that is

associated with an increased fissile inventory.

We previously estimated23 the performance of continuous fluorinators
by assuming that the rate of removal of uranium from the salt is first
order with respect to the concentration of uranium in the salt. If the
transfer of uranium in the salt by axial dispersion and by convection
is taken into account, the concentration of uranium in the salt is
defined by the following relation:

2
D-é~% - v-%% -
dx

kC =0 (25)
Lo

where
D = axial dispersion coefficient, cmzlsec,
C = concentration of uranium in salt, moles/cm3,
X = position in column measured from top of column, cm,
V = superficial salt velocity, cm/sec,
k = reaction rate constant, sec—l.

The terms in Eq. (25) represent the transfer of uranium in the salt by
axial dispersion, the transfer of uranium in the salt by convection, and
the removal of uranium from the salt by reaction with fluorine respec-
tively. The assumption of the first-order reaction does not imply a
particular rate-limiting reaction mechanism; however, it is consistent
with the assumption that the rate-limiting step is diffusion of uranium
in the salt to the gas-liquid interface. In this case, the first-order
expression would imply that the concentration of uranium in the salt at
the interface is negligible in comparison with the uranium concentration

in the salt at points a short distance from the interface.

The boundary conditions chosen for use with Eq. (25) assume that
the diffusive flux across the fluorinator boundaries is negligible: at
X = 0 (top of fluorinator),

dc

_ v
X = -5 [Cfeed = Co4l ’ (26)

X=0+
and at X = L (bottom of fluorinator),

dc

=0 . (27)
dX " xar,
where
Cfeed = concentration of uranium in salt fed to the fluorinator,
CO+ = concentration of uranium in salt at top of the fluorinator.

Note that C0+ is not equal to Cfeed since there is a discontinuity in
uranium concentration in the salt at the top of the fluorinator where the

salt enters.
¥

Solution of Eq. (25) with the stated boundary conditions yields the
following expression for the ratio of the uranium concentration in salt

leaving the fluorinator to the concentration in the feed salt:

 

feed
1
J__LE__ML + 1/2 E[‘V 1;)"' + n = 1/2] ng__fl ]/2 "é[l/g +,/1;)—l + 'r]]
e — — e s
1 + 4 1
J1+ by J1 + by (28)
where

c(L)

concentration of uranium in salt leaving fluorinator,

n__EB

===,
Vv
VL

€ =7,

L = length of fluorinator, cm.

Application of Eq. (28) to the design and evaluation of continuous
fluorinators requires values for the rate constant k and the axial
dispersion coefficient D. When we made the earlier estimates of
fluorinator performance,go only limited data were available for the
axial dispersion coefficient; these data resulted from studies with
air and water in 1.5-, 2-, and %-in.-ID columns. At that time it was
assumed that the axial dispersion coefficient was represented by the

following relation:

D =5.22,/G , (29)

where
. . . . 2
D = axial dispersion coefficient, cm /sec,

G = gas flow rate at top of fluorinator, cm3/sec.

The rate constant, k, was evaluated from experimental data obtained with
a l-in.-diam open-column, continuous fluorinator.20 In correcting the
data for the effect of axial dispersion, results obtained with the 1l.5-
in.-diam column were used and no correction was made for the differences
in the physical properties of molten salt and water. Since that time,

additional data on axial dispersion in open bubble columns have been
Ll

obtained in 1-, 1.5-, 2-, 3-, and 6-in.-diam columns using widely varying
gas flow rates and aqueous solutions having a range of physical properties.
One method for correlating the data yields the following relations: for

low gas flow rates (bubble flow),

N n

Npe = 18.0 Ny Npr Mgy ’ (30)
and at high gas flow rates (slug flow),
0.4 0.11 -0.38
NPe = 0.46 NRe NAr NSu s (31)
where
N, =2 = peclet numb
pe = p = Peclet number,
_ pdV _
NRe . Reynolds number,
d3 2
N, =%L2 8 - Archimedes number ,
Ar 2
Y .
NSu = Q%E_z Suratman number,
u
d = column diameter,
V = superficial gas velocity,
p = density of liquid,
p = viscosity of liquid,
o = surface tension of liquid,
g = acceleration of gravity,
n = number of gas inlets in disperser.

The transition from bubble to slug flow occurs at the point represented
by the following relation:
1.14 -0.635 n0.099

_ 4 .
Ne_ = 4.81 x 107" N, _ Ng. . (32)

These relations for the axial dispersion coefficient differ somewhat from
those developed most recently (see Sect. 9); however, the axial dispersion

coefficient values predicted by the two sets of relations are in good
b5

agreement. It is not believed that the use of Egs. (30)-(32) introduces
significant error in the calculated performance data for continuous flu-

orinators given later in this section.

The reaction rate constant, k, was reevaluated from data from a 1-
in.-diam continuous fluorinator operated at 525°C with an inlet uranium
concentration of 0.35 mole %23 by using the mathematical model represented
by Egqs. (25)-(27) and the data on axial dispersion represented by Egs.
(30)-(32). The results, summarized in Table 4, show no trend with salt
or fluorine flow rate; that is, the values for k are seen to be reasonably

constant.

Table 4., Summary of Data for Evaluation of Fluorination Reaction
Rate Constant from Data Obtained at 525°C in a
l-in.~diam Continuous Fluorinator

 

 

 

F
. . 2
Salt Super?lClal c(L) Flow Rate D k

Velocity C T3 2 _1
(cm/sec) feed (cm™/sec) (cm™/sec) (sec . )
0.0625 0.0257 6.8 17.6 0.00805
0.0445 0.0096 5.0 14.6 0.01033
0.0225 0.00457 3.82 10.6 0.00886
Avg 0.00908

 

The performance of large open-column, continuous fluorinators (6, 8,
10, and 12 in. in diameter) was estimated from Eq. (28) using the pre-
viously discussed estimate of k and the correlations for predicting the
axial dispersion coefficient, D. The required fluorinator heights are
shown in Figs. 17-20 for fractional uranium removal values of 0.9, 0.95,
0.99, and 0.999. The uranium concentration in the inlet salt was assumed
to be 0.0033 mole fraction in each case, and the fluorine flow rate was
assumed to be 150% of the stoichiometric requirement. These results are
encouraging since they suggest that single fluorination vessels of moderate

size will suffice for removing uranium from MSBR fuel salt prior to the
ORNL OWG 72-13526

 

  
     
 
  
   

 

 

 

100 1 1T T TTTTT T T T T IT1T1T1 T T T T T 1T
70 Uronium Removael Efficiency 90% =
sot- Inlet Uranium Concentration  0.0033 mole froction ~
- Fluorine Feed Rote 150 % of stoichiometric =

wii —
10.0}p- 3
7-0"" —

> S0 —

* —

< Fluorinator

=4 - . -

° Diometer

z - 6 in, —

Y

o

e 10} -

- - —

3 o ]

LW osp ~
o.sh— L
o.1 ] L1ttt | I I 1131t 1 L1111

i0 30 850 100 300 500 1000 3000 5000 10,000

Salt Flow Rate (ft’/day)

Fig. 17. Variation of Calculated Fluorinator Height with Salt Flow
Rate and Fluorinator Diameter for a Uranium Removal Efficiency of 90%.
FLUORINATOR HEIGHT {ft)

by

URNL DWG 71-10788

 

        
 

 

 

 

IO_ { | IIIIII[ I T lTlIll] | 1 F v 1171
: Uranium Remcval Efficiency 95 % :
- Inlet Uranium Concentration 0.0033 mole fraction —
- Fluorine Feed Rate 150 % of Stoichiometric —
- mamy
10— ]
: Fluorinator -
| Diameter _
I i Lt o1yl 1 L1 1 111
10 100 1000 KPOC

SALT FLOW RATE(#7Yday)

Fig. 18. Variation of Calculated Fluorinator Height with Salt Flow
Rate and Fluorinator Diameter for a Uranium Removal Efficiency of 95%.
FLUORINATOR HEIGHT (ft)

ORNL DWG 7i-10793

 

100,

 

 

 

 

T T T T T T 1 7T [ | T T T T T lTl T T T 1 11

Uranium Removal Efficiency 99% ;

70 Inlet Uranium Concentration 0.0033 mole fraction 7 —

 Fluorine Feed Rote 150% of stoichiometric N

50+ T

30 -

FLUORI

IO_— _j

F -

' _

| .

5 ]

3+ O ]

' 1 i L i L .1 11 l 1 1 i 1 1411 I —1_ 1 1 r i 1 4.1

10 100 1000 10000

SALT FLOW RATE (ft3/day)

Fig. 19. Variation of Calculated Fluorinator Height with Salt Flow
Rate and Fluorinator Diameter for a Uranium Removal Efficiency of 99%.

81
FLUORINATOR HEIGHT (ft)

ORNL DWG 72-13509R]|

 

 

 
   
    
 

 

 

00 T T T T T T T T T T T T T T T T T T 1]
80| ' —
60 — —
50+ _
FLUORINATOR
30 DIAMETER ]
k ng.
8 in.
10in.

20 12 in, -
10 —
sl ]

- —
6 —
8 URANIUM REMOVAL EFFICIENCY - 999% —

INLET URANIUM CONCENTRATION - 0.0033 mole fraction

— FLUORINE FEED RATE -150 % OF STOICHIOMETRIC “
3 ]
2+ _

I 1 1 Lo | i Lo o1l 1 1 I L1 11t
10 100 1000 10000

Fig.

Rate and

SALT FLOW RATE (ft3day)

20. Variation of Calculated Fluorinator Height with Salt Flow
Fluorinator Diameter for a Uranium Removal Efficiency of 99.9%.

t
50

isolation of protactinium by reductive extraction. The reference flow-
sheet for isolating protactinium by fluorination--reductive extraction2
requires fluorination of fuel salt at the rate of 170 ft3/day, which is
equivalent to a 10-day processing cycle. A 6-in.-diam fluorination having
a height of 10.2 ft will be required for a uranium removal efficiency

of 95%Z; an 8-in.-diam fluorinator having a height of 17.8 ft will be

required for a uranium removal efficiency of 99Z%.

Fluorinators having a high uranium removal efficiency are required
in the production of high-purity 233U because incomplete removal of
uranium from a salt stream containing 233Pa would result in contamination
of the 233U with other uranium isotopes. Therefore, fluorination of salt
streams having flow rates of 550 to 1700 ft3/day with uranium removal
efficiencies as high as 99.9% may be required. As shown in Fig. 20, a
column diameter of 10 in. and heights of 42.5 to 60 ft would be required
if a single, open-column, continuous fluorinator were used. In this case,
the fluorinator would be divided into several open—-column fluorinators
operating in series. If two columns were used, the required heights of
each column would be less than half the height required for a single
column since there would be no axial dispersion across the fluorinator
inlets and outlets. The required uranium removal efficiency for each
column would be 96.8%; and, as shown in Fig. 21, column heights of 17 to
28.3 ft would be required for a 10-in.-diam fluorinator. The use of
three columns, each with a 90Z uranium removal efficiency, would reduce
the total column height even further. Column heights of 7.8 to 17.2 ft

would be required for a 1l0-in.-diam fluorinator in this case.
FLUORINATOR HEIGHT

100
80

60

50 t+

40
30

20

w S0 @

~N

ORNL DWG 72-13508

 

T T ¥ 1

— FLUORINATOR
DIAMETER

" 6in.

8 in.

" 10in.

- 12 in. ‘\

vl

 

1 I U rrrvy 1 L L L f T T v 0

URANIUM REMOVAL EFFICIENCY-96.8%
INLET URANIUM CONCENTRATION-0.0033 MOLE FRACTION
FLUORINE FEED RATE-150% OF STOICHIOMETRIC

L i1l 1 { 1t 11 el 1 L1 11

-

1 1

L 1111

1

 

 

10

Fig. 21. Variation of Calculated Fluorinator Height with Salt Flow

Rate and Fluorinator D

100 1000
SALT FLOW RATE ( ft3/day)

{ameter for a Uranium Removal Efficiency of 96.8%.

16
52

9. MEASUREMENT OF AXIAL DISPERSION COEFFICIENTS AND
GAS HOLDUP IN OPEN BUBBLE COLUMNS

J. §. Watson L. E. McNeese

Axial dispersion is important in the design and performance of
continuous fluorinators to be used in processing MSBR fuel salt. Since
molten salt saturated with fluorine is corrosive, the fluorinators will
be simple, open vessels having a protective layer of frozen salt on all
exposed metal surfaces. In such systems the rising gas bubbles may
cause appreciable axial dispersion throughout the salt. For the past
few years, we have been involved in a program for measuring axial dis-
persion resulting from the flow of air through liquids in open bubble
columns. The objectives of this program are to evaluate the effect of
axial dispersion on fluorinator performance and to account for this

effect in the design of fluorinators.

9.1 Previous Studies on Axial Dispersion

Initial studies on axial dispersion in open bubble columns were
carried out by Bautista and McNeese,25 who studied axial dispersion
during the countercurrent flow of air and water in a 2-in.-ID, 72-in.-
long column. Two regions of operation were observed. The first of
these consisted of a "bubble flow" region at low gas flow rates in which
the air moved up the column as individual bubbles and coalescence was
minimal. The second consisted of a "slug flow" region at higher gas
flow rates in which the air coalesced rapidly into bubbles having
diameters equal to the column diameter. A plot of the logarithm of
the dispersion coefficient vs the logarithm of the gas flow rate was
linear in both regions. However, the slope of the line representing
data in the slug flow region was higher than that for data in the bubble

flow region. The transition between the two regions was well defined.

The same column and equipment were used by A. M. Sheikh and J. D.

26
Dearth, of the MIT Practice School, for investigating the effects of
the viscosity and surface tension of the liquid. The dispersion coef-

ficient was found to decrease in the bubble flow region as the viscosity
53

of the liquid was increased from 1 cP to 15 cP by the addition of
glycerol to the water; little effect was noted in the slug flow region.
An increase in the dispersion coefficient was observed as the surface
tension of the liquid was decreased by the addition of n-butanol to

the water.

21 of

The equipment was also used by A. A. Jeje and C. R. Bozzuto,
the MIT Practice School, who investigated the effects of gas inlet
diameter and column diameter on axial dispersion and obtained data on
gas holdup in bubble columns. In the slug flow region, the dispersion
coefficient appeared to be proportional to the square root of the
volumetric gas flow rate, but was independent of column diameter. 1In
the bubble flow region, the dispersion coefficient was dependent only
on the volumetric gas flow rate in the case of columns having diameters
of 2 in. or larger. Dispersion coefficient data obtained with a 1.5-
in.~diam column deviated from this condition. At low gas flow rates,
the gas holdup was linearly dependent on the superficial gas velocity
but independent of column diameter. At superficial velocities above
the transition from bubble to slug flow, the gas holdup data for the
various column diameters diverged; the holdup was greatest for the

smallest column diameter.

All of the dispersion coefficient data obtained by the above
investigators resulted from measurements of the steady-state axial
distribution of a cupric nitrate tracer that was continuously injected
into the bottom of the column near the water exit. The studies indi-
cated that the axial dispersion coefficient is independent of both
axial position in the column and water superficial velocity in the
range of interest. The steady-state experimental technique had two
principal disadvantages: (1) the measurements were time-consuming
since about 2 hr was required for the column to attain steady state;
and (2) at high gas flow rates, air was entrained with water withdrawn
from the column for determination of the tracer concentration, and
the resulting error in the dispersion coefficient data was unacceptably
high. 1In order to circumvent these problems, Bautista28 developed a

transient technique for obtaining data on axial dispersion. In this
5L

technique there was no net flow of water through the column; however,
data obtained with the steady-state technique indicated that the water
flow rate did not affect the axial dispersion coefficient at the water
flow rates of interest. A small amount of electrolyte tracer (KCl)
was quickly injected into the top of the column and the concentration
of the tracer was measured continuously at a point near the bottom of
the column by use of a conductivity probe. The resulting data were in
agreement with earlier data obtained with the steady-state technique;
on the other hand, data obtained with the transient technique showed
minimal scatter even at high gas flow rates and could be obtained in
less than 10% of the time required for the steady-state technique.
During this report period, additional studies were carried out using
the transient technique in order to determine the effects of changes
in column diameter, gas inlet design, and physical properties of the

liquid phase on axial dispersion and gas holdup.

9.2 Equipment and Experimental Procedure

The equipment and experimental procedure used in the present studies
have been described previously-28 The equipment consisted of an open
bubble column, a means for injecting KCl tracer solution at the top of
the column, a conductivity probe located at an intermediate axial point
along the column for determining the KCl concentration in the aqueous
solution at the point, an electronics system and a recorder for recording
the output from the conductivity probe, an air supply and metering sys-
tem for feeding air at a known flow rate to a gas disperser located
in the base of the column, and a manometer for obtaining data on gas
holdup in the column. Eight-foot-long Plexiglas columns with inside
diameters of 1.0, 1.5, 2, 3, and 6 in. were used. Gas distributor
plates having different numbers and sizes of orifice openings were
installed at the bottom of the column in order to determine the effect
of gas inlet design on axial dispersion and gas holdup. The air flow
rate was measured at the top of the column. A soap bubble buret was
employed for flow rates below 15 cm5/sec; a wet-test meter was used for
higher flow rates. The solutions used in the study consisted of

demineralized water or mixtures of demineralized water and glycerin or
55

n-butanol. The aqueous solutions were prepared in a tank and pumped to
the column. Demineralized water could be introduced at the top of the

column to facilitate cleaning of the column between runs.

After the column had been filled with liquid having the desired
physical properties, a sufficient volume of 2.k N KCl tracer solution
(5 to 15 cmj, depending on column size) was added to the liquid in the
column in order to obtain a recordable reading from the conductivity
probe. The air flow rate was then adjusted to the desired value, and
a second volume of tracer was quickly injected at the top of the column.
Subsequently, the response of the conductivity probe was recorded until
the tracer was uniformly dispersed throughout the column. The height of
the gas-liquid mixture in the column and the height of the liquid with
no gas flow were measured. Samples of the liquid were then taken for
surface tension and viscosity measurements. The viscosity of the
liquid was determined with a Ubbelohde viscometer, while surface tension
measurements were made using the capillary rise method. Visual observa-
tions of the gas and liquid in the column were made during the course of

the experiments.

9.3 Experimental Data on Axial Dispersion

Experimental data on axial dispersion in open bubble columns were
obtained during this report period in a series of four separate studies.
The first study, made by a group of students at the University of
Tennessee, was carried out with a l-in.-ID, 8-ft-long column. The gas
disperser at the bottom of the column consisted of a single inlet having
an inside diameter of 4.3 mm. The studies were carried out with demin-
eralized water, and the superficial gas velocity was varied from 0.156
to 76.6 cm/sec. The data obtained during this study are summarized in
Table 5. The axial dispersion coefficient values obtained in the l-in.-
diam column fall below the values obtained previously at the same super-

ficial gas velocity in columns having diameters of 1.5 and 2 in.

The second study was carried out by J. C. Bronfenbrenner, L. J.
Marquez, and J. F. Mayer, of the MIT Practice School, who determined

the effects on axial dispersion caused by changes in column diameter,
56

Table 5. Summary of Data on Axial Dispersion Obtained in a 1.0-in.-ID
Open Bubble Column Containing Demineralized Water at 25°C

3

Tracer injection volume: ~ 5 cm
Relative probe positiona: 0.825

Gas inlet: one orifice, 4.3 mm in diameter

 

 

Gas Flow Superficial Dispersion
Run Rateb Gas Velocity Coefficient
No (em/sec ) (cm/sec) (em®/sec)
1 ok.6 18.7 82.3
2 387.9 T6.6 319.0
3 387.9 76.6 261.8
L 231.1 65 .k 255.2
5 378.4 Th. T 182.3
6 283.8 56.0 319.0
T 189.2 37.3 170.2
8 ok.6 18.7 63.8
9 141.9 28.0 170.2
10 236.5 46.1 232.0
11 52.0 10.3 52.0
12 QL. 6 18.7 92.8
13 146.6 28.9 128.3
14 118.2 23.% 102.6
15 70.9 14.0 61.1
16 165.6 22.7 146.6
17 189.2 37.3 213%.8
18 236.5 Lo.7 185.5
19 321.7 63.5 285.1
20 227.0 4.8 181.1
21 5.91 1.166 20.8
22 5.51 1.088 18.3
23 5.12 1.011 17.1
24 4. 73 0.933 22.5
25 4.33 0.855 17.1
26 3.94 0.778 15.5
27 3.55 0.700 14.1
28 0.79 0.156 5.84
29 6.33 1.25 14.8
30 5.17 1.02 2.6
51 3-95 0.78 19.7
32 2.37 0.468 1L.8
3% 0.79 0.156 7.04
3h 6.7k 1.33 22.h
35 5.12 1.01 17.8
36 3.56 0.702 1k.5
37 1.98 0.390 10.7
38 0.79 0.156 6.42

 

a . .

Ratio of distance of probe from surface of gas-liquid mixture to total
height of gas-liquid mixture.

Measured under conditions at top of column.
5T

viscosity of the liquid phase, and superficial gas velocity. Thle columns
used in this study consisted of 1.5-, 2-, and 6-in.-ID Lucite tubes, each
having a length of 8 ft. A conductivity cell was inserted in the columns
at a height of 70.5 cm from the bottom of the column. The liquid in the
column consisted of mixtures of distilled water and glycerin in which the
glycerir concentrations were O, 25, and 65 wt %. The physical properties
of these solutions are summarized in Table 6. Air entered the column
through a single orifice at the bottom of the column; the orifice ID was
0.04 in. for the two smaller columns and 0.4 in. for the 6-in.-diam
column. The superficial gas velocity was varied from 0.26 to 40 cm/sec
in the smaller columns and 0.27 to 9.5 cm/sec in the 6-in.-diam column.
Data on axial dispersion obtained during the second study are summarized
in Tables 7-9. The results obtained with water in a 1.5-in.-diam column
are in good agreement with those obtained previously;28 the axial disper-
sion coefficient shows little change as the viscosity of the liquid is
increased from 0.9 cP to 1.8 cP. Similarly, there is little difference
in the axial dispersion coefficient values obtained in a 2-in.-~diam
column with a liquid having a viscosity of 0.9 cP and those obtained

with a liquid having a viscosity of 1.8 cP. Dispersion coefficient
values obtained with a liquid having a viscosity of 12.1 c¢P are about

50 to 70% of those obtained with liquids having viscosities of 0.9 and
1.8 ¢cP. Essentially no difference was observed in the axial disper-

sion coefficient values obtained in a 6-in.-diam column for liquids

having viscosities of 0.9, 1.8, and 12.1 cP.

Table 6. Physical Properties of Water-Glycerin Solutions
Used During Second Study of Axial Dispersion
and Gas Holdup in Bubble Columns

 

 

Glycerin Surface
Concentration Viscosity Density Tension
(wt %) (cP) (g/cm>) (dynes/cm)
O 0.89 0.997 T3
25 1.8 1.05 T2

65 12.1 1.16 67.9

 
58

Table 7. Summary of Data on Axial Dispersion Obtained in a
1.5-in.-ID Column During Second Study

Relative probe position®: 0.69

Gas inlet: one orifice, 1.0 mm in diameter

 

 

Gas Flow Superficial Glycerin Axial Dispersion
Run Rate Gas Velocity Conc. Coefficient
No. (em>/sec) (cm/sec ) (wt %) (cm®/sec )
1 187.0 16.4 0 138.5
2 97.8 8.58 0 85.6
3 140.2 12.3 0 114.8
L 112.8 9.89 0 93.L4
5 19.6 1.72 0 31.6
6 4L0%.6 5.4 0 4h82.2
7 7.87 0.69 0 20.8
8 403.6 35,4 0 380.5
9 3.1 2.99 0 39.0
10 58.2 5.11 0 65.0
11 10.4 0.912 0 24.8
12 255.4 22.4 0 177.0
13 118.6 10.4 0 149.0
1k 191.5 16.8 0 136.0
15 155.0 13.6 0 95.2
16 118.6 10.4 25 88.9
17 18.9 1.66 25 2.9
18 372.8 32.7 25 813.9
19 566.6 Lo.7 25 655.9
20 372.8 32.7 25 155.1
21 46.3 4.06 25 Li.6
22 118.6 10.4 25 88.9
2% 2L9. 7 21.9 25 115.3%
2L 149.4 13.1 25 131.4
25 118.6 10.4 25 56.2
26 75.5 6.62 25 €2.2

 

a
Ratio of distance of probe from surface of gas-liquid mixture to total
height of gas-liquid mixture.

b
Measured under conditions at top of column.
29

Table 8. Summary of Data on Axial Dispersion Obtained in a
2.0-in.-ID Column During Second Study

a
Relative probe position”: 0.69
Gas inlet: one orifice, 1.0 mm in diameter

 

 

Gas Flow Superficial Glycerin Axial Dispersion
Run Rate Gas Velocity Conc. Coefficient
No. (em?/sec ) (cm/sec ) (wt %) (cm®/sec )

1 81.1 4.0 0 51.0
2 h11.4 20.3 0 206.0
3 482.4 23.8 0 172.7
L 291.9 4.k 0 144.9
RS 11.4 0.56 0 22.3
6 78.6 3.88 0 58.7
7 120.8 5.96 0 67.5
8 117.2 5.78 25 72.2
9 1rr.3 8.75 25 97.8
10 210.8 10.4 25 119.7
11 377.0 18.6 25 178.3
12 553.3 27.3 25 L27.0
13 56.5 2.79 25 45.3
14 81.3 L.01 25 52.1
15 10.7 0.53 25 26.5
16 : 8.31 0.41 25 26.5
17 15.2 0.75 25 31.2
18 2L .5 1.21 25 34.5
19 L. 8 2.21 25 48.2
20 18.4 0.91 65 27.6
21 9.93 0.49 65 21.h
22 6.28 0.31 65 18.9
23 25.1 1.24 65 27.8
24 L7.6 2.35 65 zh.,1
25 60.8 3.0 65 27.2
26 83.3 4.11 65 L. L
27 116.9 5.77 65 51.0
28 184.4 9.1 65 75-3
29 208.8 10.3 65 79.2
30 326.3% 16.1 65 105.3
31 504.7 29 65 1hl.2
32 758.0 37.4 65 213.6

 

#Ratio of distance of probe from surface of gas-liquid mixture to total
height of gas-liquid mixture.

bMeasured under conditions at top of column.
60

Table 9. Summary of Data on Axial Dispersion Obtained in a
6.0-in.-ID Column During Second Study

Relative probe position?: 0.69
Gas inlet: one orifice, 10 mm in diameter

 

 

Gas Flow Superficial Glycerin Axial Dispersion

Run Rate Gas Velocity Conc. Coefficient

No. (cmj/sec) (cm/sec) (wt %) (cm?/sec )
1 1530 8.39 0 207.0
2 561.8 2.08 0 178.8
3 145.9 0.8 0 152.1
L 9k.8 0.52 0 156.6
5 ok .8 0.52 0 160.7
6 271.8 1.49 0 183.2
T 394.0 2.16 0 161.3
8 1299 T7.12 0 370.8
9 113.1 0.62 0 173.4
10 698.6 3.83 0 156.9
11 195.2 1.07 0 123.1
12 60.2 0.33% 0 146.0
13 286.4 1.57 0 146.2
14 1665 9.1% 0 280.6
15 899.3 4.93 0 245.8
16 479.8 2.63 0 141.0
17 1372 7.52 25 231.5
18 1757 9.63 25 265.9
19 L1757 9.63 25 303.0
20 hh.3 2.6 25 169.5
21 923.0 5.06 25 262.0
22 1572 8.62 25 229.3
23 58.4 0.32 25 124.7
24 372.1 2.0h 25 187.7
25 217.1 1.19 25 180.5
26 5436 2.98 25 184.0
27 707.8 3.88 25 210.8
28 5h3.6 2.98 25 181.1
29 1094 6.0 25 215.9
30 220.7 1.21 65 152.8
31 295.5 1.62 65 148.6
32 Lg.2 0.27 65 161.8
35 T727.8 5-99 65 181.1
3k 521.7 2.86 65 179.6
35 361.2 1.98 65 145.6
36 1096 6.01 65 215.6
61

Table 9. (continued)

Relative probe position®: 0.69
Gas inlet: one orifice, 10 mm in diameter

 

 

Gas Flow Superficial Glycerin Axial Dispersion
Run Rate Gas Velocity Conc. Coefficient
No. (cm?/sec ) (cm/sec) (wt %) (cm?/sec )
37 1532 8.h 65 o5k .2
38 93.0 0.51 65 157.2
39 136.8 0.75 65 172.6

 

a
Ratio of distance of probe from surface of gas-liquid mixture to total
height of gas-liquid mixture.

b
Measured under conditions at top of column.

The third study carried out during this period was made by A. K.
Padia, G. T. Marion, and R. H. McCue, of the MIT Practice School, who
studied the effects of changes in the number and size of gas inlet
orifices, column diameter, superficial air velocity, and viscosity
and surface tension of the liquid phase on the axial dispersion coef-
ficient and gas holdup. The ranges of the independent parameters that
were varied in this study are summarized in Table 10. Column diameters
of 1.5, 2, and 3 in. were used with both single and multiple orifices
ranging in size from O.4 to 6.4 mm. The viscosity of the liquid phase
was varied from 0.9 to 11.% c¢cP, and the surface tension of the liquid
was varied from 27 to 7O dynes/cm. The superficial gas velocity was
varied from 0.0318 to 20 cm/sec in 12 to 17 increments for each value
of colurn diameter, gas distributor design, and property of the liquid

phase. Data obtained during the third study are summarized in Tables
62

Table 10. Ranges of Parameters During Third Study
of Axial Dispersion in Open Bubble Columns

 

 

Parameter Values Used
Column Diameter, in. 1.5, 2, 3
Number of Orifices in Gas Inlet 1, 5, 19, 37
Gas Inlet Orifice Diameter, mm 0.4, 1, 2, 4, 6.4
Surface Tension of Liquid, dynes/cm 27, 45, TO
Viscosity of Liquid, cP 0.9, 2.05, 10.7, 11.3
Superficial Gas Velocity, cm/sec 0.0318 to 20

 

11-19. The variation of the axial dispersion coefficient with changes
in the superficial gas velocity and orifice diameter for a 2-in.-ID
column for which the gas distributor consisted of five orifices is in
general agreement with that obtained previously for a 2-in.-diam

column operated with a single gas inlet. The variation in the dispersion
coefficient in the slug flow region with changes in the diameter of the
gas inlet orifices is not believed to be significant. However, the
differences observed in the bubble flow region are probably meaningful.
Data of the same type, obtained with a gas disperser consisting of 37
orifices, show even less deviation from previous values obtained with

a single gas inlet. For a column diameter of 2 in. and gas distributor
orifice diameters of 1 mm, only slight differences in the dispersion
coefficient are noted in the slug flow region; however, in the bubble
flow region, a progressive increase in axial dispersion coefficient is
observed as the number of orifices is increased. The variation of the
axial dispersion coefficient with changes in the viscosity of the liquid

phase and the superficial gas velocity for column diameters of 1.5 and
63

Table 11. Summary of Data on Axial Dispersion Obtained
in a 1.5-in.-ID Column Containing
Water During Third Study

5

s s - ‘
Tracer injection volume: /~ 5 cm
, P -
Relative probe position : 0.79

Gas inlet: one orifice, 1 mm ID

 

 

Gas Flow Superficial Axial Dispersion

Run Ratel Gas Velocity Coefficient
No. (cm5/sec) (cm/sec ) (cm®/sec )

1 11.9 1.0k45 25.8

2 2.42 0.212 12.9

3 5.14 0.451 18.9

L 8.20 0.719 2L )

5 15.2 1.33 32.9

6 19.2 1.68 31.6

7 25.2 2.21 39.5

8 Ly, 0 3.86 50.0

9 Th.6 6.54 69.4

10 101.8 8.93 85.9

11 2Lh2.8 2l.3 242.0

 

®Ratio of distance of probe from surface of gas-liquid mixture to total
height of gas-liquid mixture.

b ‘o
Measured under conditions at top of column.
6L

Table 12. Summary of Data on Axial Dispersion in a
l1.5-in.~ID Column Containing Aqueous
Isopropanol During Third Study

3

Tracer injection volume: ~ 5 cm
Relative probe position®: 0.79
Surface tension of liquid: L45.3% dynes/cm

Gas inlet: one orifice, 0.638 cm ID

 

 

Gas Flow Superficial Axial Dispersion

Run Rate Gas Velocity Coefficient
No. (em’/sec) (cm/sec ) (cm®/sec )
1 11.9 1.045 20.9

2 8.09 0.710 23%.3%

3 6.10 0.535 20.5

L 0.87 0.252 17.3

5 1.06 0.0930 14.6

6 16.6 1.456 23.1

7 17.7 1.55 26.7

8 29.h 2.58 33.7

9 49.1 L.31 Lo.6

10 66.0 5.79 50.0

11 BL4.9 T .45 58.0

 

%Ratio of distance of probe from surface of gas-liquid mixture to total
height of gas-liquid mixture.

Measured under conditions at top of column.
65

Table 15. Summary of Data on Axial Dispersion in a 1.5-in.-ID
Column Containing Aqueous Isobutanol During Third Study

3

Tracer injection volume: -~ 5 cm
Relative probe positiona: 0.79
Surface tension of liquid: 27.3 dynes/cm

Gas inlet: one orifice, 0.638 cm ID

 

 

Gas Flow Superficial Axial Dispersion

Run RateP Gas Velocity Coefficient
No. (cm’/sec ) (cm/sec ) (cm?/sec )

1 11.6 1.02 21.9

2 1.52 0.133 15.0

3 0.804 0.0705 11.9

L 2.4k 0.214 15.8

5 5.63 0.LokL 20.6

6 8.70 0.763% 22.2

7 15.6 1.37 22.3

8 23.6 2.07 27.9

9 40.6 3.56 36.2

10 73.0 6.4 51.5

11 98.0 8.6 65.7

 

fRatio of distance of probe from surface of gas-liquid mixture to total
height of gas-liquid mixture.

bMeasured under conditions at top of column.
66

Table 14. Summary of Data on Axial Dispersion Obtained
in a 1.5-in.-ID Column Containing a Water-Glycerin
Mixture During Third Study

3

Tracer injection volume: ~ 5 cm
s .4
Relative probe position : 0.79

Gas inlet: one orifice, 0.638 cm ID

 

 

Gas Flow Superficial Glycerin | Axial Dispersion
Run Rate Gas Velocity Conc. Coefficient
No. (em?/sec ) (cm/sec ) (wt %) (cm?/sec)
1 0.727 0.0638 25 T.65
2 2.13 0.187 25 12.5
3 3.80 0.333% 25 24.6
4 6.02 0.528 25 26.4
5 10.2 0.895 25 28.4
6 h.2 1.25 25 36.4
T 18.5 1.625 25 36.4
8 29.8 2.61 25 k3.5
9 65.9 5.78 25 60.6
10 92.0 8.07 25 81.0
11 2.4k 0.21%4 65 10.6
12 1.01 0.0888 65 7.45
13 3.57 0.313 65 12.3
14 5.13 0.45 65 15.2
15 T.42 0.651 65 21.3
16 11.4 1.0 65 21.8
17 15.7 1.38 65 27.%
18 19.8 L.74 65 27.7
19 33.1 2.9 65 3.7
20 66.0 5.79 65 55.6

 

#Ratio of distance of probe from surface of gas-liquid mixture to total
height of gas-~liquid mixture.

b
Measured under conditions at top of column.
67

Table 15. Summary of Data on Axial Dispersion Obtained in a
2.0-in.-ID Column Containing Water During Third Study

5

Tracer injection volume: ~ 5 cm
a
Relative probe position : 0.79

Gas inlet: one orifice, 1 mm ID

 

 

Gas Flow Superficial Axial Dispersion

Run RateP Gas Velocity Coefficient
No. (cmd/sec) (cm/sec ) (cm?/sec )

1 0.896 0.04kL2 19.6

2 0.255 0.0126 28.1

3 L. 2k 0.209 28.9

L 5.98 0.295 31.1

5 8.21 0.405 37.6

6 10.4 0.515 40.5

T 12.5 0.618 L3.4

8 16.2 0.798 45.5

9 19.2 0.95 L8.2

10 230.8 1.52 52.7

11 59.2 2.92 60.8

12 92.8 L.58 T2.6

13 145.9 T.2 81.5

 

4Ratio of distance of probe from surface of gas-liquid mixture to total
height of gas-liquid mixture.

bMeasured under conditions at top of column.
Table lt.

Summary of Data on Axial Dispersion Obtained in

Number of orifices:
Tracer injection volume:
Relative probe position?:

a -in.-ID Column Containing Water During Third Study

~

L
cm-’
Runs 1-12

= 0.76T

Runs 135-68 — 0.730

 

 

Gas Flow Superficial Gas Inlet  Axial Dispersion Gas Flow Superficial Gas Inlet  Axial Dispersion
Run Rate Gas Velocity Orifice ID Coefficient Run Rateb Gas Velocity Orifice 1D Coefficient
No. (cm”/sec) (cm/sec) (mm ) (cm/sec ) No. (cm’/sec) (cm/sec ) (mm ) (cm</sec )
1 B0 0. 58 G.h Lo.6 35 L.76 0.235 2.0 35.8
o 0.026 o0kt O.4 15.4 36 10.2 0.503% 2.0 30.0
3 2.50 0.123, 0.4 16.4 37 20.3 1.0 2.0 k4.0
Y L .90 0.242 0.4 22.8 38 5.13 0.253 4.0 27.1
g 10.9 0.537 0.4 33,1 39 10.3 0.509 4,0 4i.2
L 15.2 0.749 0.4 354 Lo 31.2 1.54 L.o 41.8
v 19.4 0.96 0.4 36.5 Li 84.5 L.17 L.o 65.5
8 344 1.7 Ok 1.4 42 1.90 0.0938 L.o 30.8
9 65.% 4.en 0.% 57.8 43 10.4 0.512 4.0 28.3
10 97.1 L.to 0.4 60.8 L 45.8 2.26 4.0 S5L.0
11 141.9 .00 0.k 2.6 Lg 111.5 5.5 4.0 69.5
12 171.9 §.46 O.h B5.6 46 450 22.2 4.0 326.0
13 13.9 N. 688 1.0 7.2 47 282 13 9 4.0 120.0
14 1.55 0.0767 1.0 31.7 L8 19.7 0.97 4.0 33,7
15 3.20 0.158 1.0 31.5 Lg 2.45 c.121 4.0 26.4
16 5.2 0.26 1.0 4L .8 50 26.% 1.3 k.0 41.8
17 7.03 0.3L7 1.0 47.0 51 14.6 0.72 4.0 36.8
18 9.50 0.459 1.0 37.9 52 7.56 0.373 k.0 32.5
19 11.9 . 5,80 1.0 35.5 53 1.54 0.076 4.0 25.5
20 18.6 0.2 1.0 %9.0 54 1.0k 0.0512 2.0 20.8
21 50.8 1.52 1.0 55.1 55 2.53 0.125 2.0 41.0
22 5.8 et 1.0 57.5 56 4.09 0.202 2.0 45.6
23 86.4 ooy 1.0 6%.5 57 6.16 0.304 2.0 37.0
2k 113.5% D) 1.0 92.4 58 8.21 0.405 2.0 26.5
25 147.8 T.20 1.0 111.1 59 "10.6 0.525 2.0 k9.2
26 280 13.8 1.0 137.0 60 13.2 0.65 2.0 57.0
27 88.2 Yoz 2.0 T1.6 61 15.4 0.76 2.0 62.7
28 HOBT 0.264 2.0 5.8 62 18.7 0.923 2.0 52.2
29 16.2 0.80 2.0 23.8 63 30.9 1.525 2.0 k6.8
30 5007 2.5 2.0 57k 6L 63.8 3.15 2.0 59.5
31 185.4 .1y 2.0 139.0 65 89.2 b.bh 2.0 69.5
32 te 21.6 2.0 200.0 66 117.8 5.81 2.0 75.2
33 Z.23 O.11 el 26.8 67 1445 T.13 2.0 96.5
34 4.0l 0.108 2.0 30.5 68 330 16. 2.0 87.2

 

 

 

®Ratio of distance of probe from surface of gas-liquid mixture to total height

b
Measured under conditions at top of column.

of gas-liquid mixture.

89
69

Table 17. Summary of Data on Axial Dispersion Obtained in a
2.0-in.-ID Column Containing Water During Third Study

Gas inlet: 19 orifices, 1 mm ID

5

Tracer injection volume: ~ 5 cm

Relative probe position®: 0.808

 

 

Gas Flow - Superficial Axial Dispersion
Run Rateb Gas Velocity Coefficient
No. (em?/sec) (cm/sec) (cm?/sec)
1 1.02 0.0505 26.4
2 2.31 0.11k 36.7
3 4,22 0.208 38.5
L 6.10 0.301 597
5 8.19 0.L40k 4O.1
6 10.8 0.535 hh.5
T 12.8 0.632 46.5
8 Lh.7 0.726 ht.0
9 18.6 0.919 4.9
10 30.0 1.48 L2.8
11 61.2 3.02 54,4
12 85.9 L.2h 62.5
13 115.7 5.71 80.4
14 149.6 7-38 12k.0
15 280.0 13.8 155.0

 

aRatio of distance of probe from surface of gas-liquid mixture to total
height of gas-liquid mixture.

bMeasured under conditions at top of column.
70

Table 18. Summary of Data on Axial vispersion Obtained in a 2.0-in.-ID
Column Containing Wa:er During Third Study

Gas 1nlet: 37 orifices

Tracer injection volume: ~5 cm3

Relative probe nositiona: 0.738

 

 

Gas Flow Superficial Gas Inlet Axial Dispersion
Run Rate® Gas Velocity Orifice ID Coefficient
No. (cm3/sec) (cm/sec) (mm) {(cm*/sec)
1 1.06 0.0525 1.0 27.3
2 2.41 0.119 1.0 33.3
3 3.46 0.171 1.0 40.4
4 5.86 0.289 1.0 37.8
5 7.96 0.393 1.0 33.7
6 10.7 0.529 1.0 34,2
7 13.9 0.687 1.0 45.0
8 15.8 0.78 1.0 48.0
9 19.2 0.945 1.0 48.0
10 30.8 1.52 1.0 44,4
11 60.2 2.97 1.0 58.2
12 88.4 4.36 1.0 121.0
13 113.5 5.6 1.0 95.0
14 149.0 7.35 1.0 101.0
15 259 12.8 1.0 136.0
1 503 24.8 1.0 214.0
17 746 36.8 1.0 228.0
18 2.27 0.112 2.0 36.7
19 7.62 0.376 2.0 34.7
20 12.6 0.67 2.0 39.3
21 1.42 0.0703 2.0 20.5
22 5.31 J.262 2.0 35.5
23 9.69 0.478 2.0 33.2
24 17.7 0.875 2.0 38.0
25 27.8 1.37 2.0 49.5
26 44.8 2.21 2.0 61.0
27 101 5.0 2.0 91.8
28 72.6 3.58 2.0 80.8
29 3.59 3.182 2.0 33.9
30 148 7.28 2.0 100.0
31 590 26.12 2.0 269.0
32 14.9 0.737 4.0 33.4
33 1.25 J.0618 4.0 28,2
34 3.00 J.148 4.0 33.1
35 5.09 G.z51 4.0 31.9
36 6.71 0.331 4.0 30.0
37 8.17 0.403 4.0 33.7
38 11.8 0.585 4.0 32.4
39 20.5 i.01 4.0 51.3
490 30.2 1.49 4.0 50.0
41 53.9 2.66 4.0 59.7
42 73.4 3.62 4.0 61.0
43 94,4 4.66 4.0 72.2
44 118 5.8 4.0 79.5
45 150 7.43 4.0 77.5
46 537 26.5 4.0 164.0

 

a . . L .
Ratio of distance of probe from surface of gas-liquid mixture to total height
of gas-liquid mixture.

b
Measured under conditions at top of colunn.
71

Table 19. Summary of Data on Axial Dispersion Obtained in a 3.0-in.-ID
Column Containing a Water-Glycerin Mixture During Third Study

Gas Inlet: one orifice, 0.638 cm ID

Tracer injection volume: Runs 1-14, v 4.5 cm3
Runs 15-27, ~ 1.5 em

Relative probe positiona: 0.79

 

 

Gas Flow Superficial Glycerin Axial Dispersion
Run Rateb Gas Velocity Concentration Coefficient
No (cm3/sec) (cm/sec) (wt %) (cml/sec)
1 36.9 0.81 25 50.0
2 1.60 0.0352 25 28.1
3 3.50 0.0768 25 43,2
4 5.56 0.122 25 38.2
5 8.94 0.196 25 48.9
6 13.0 0.285 25 51.8
7 19.4 0.425 25 45.0
8 25.2 0.553 25 48.7
9 46.7 1.025 25 46.2
10 84.8 1.86 25 64.0
11 130.0 2.85 25 66.4
12 258 5.66 25 89.1
13 377 8.26 25 92.2
14 1.47 0.0323 25 30.2
15 1.45 0.0318 65 27.0
16 3.51 0.0769 65 29.1
17 4.92 0.108 65 42.4
18 7.07 0.155 65 56.0
19 10.6 ‘ 0.232 65 41.4
20 15.4 0.338 65 47 .4
21 19.3 0.424 65 45.0
22 30.8 0.675 65 52,8
23 76.2 1.67 65 47 .4
24 57.0 1.25 65 56.3
25 104 2.28 65 57.2
26 135 2.96 65 64.3
27 245 5.37 65 86.7

 

8Ratio of distance of probe from surface of gas-liquid mixture to total
height of gas-liquid mixture.

bMeasured under conditions at top of column.
T2

3 in. shows that axial dispersion coefficient values obtained with the
3-in. column are higher than those obtained with the l1.5-in. column at

a given superficial gas velocity, particularly in the case of bubble
flow. 1In general, there is a progressive decrease in the value of the
axial dispersion coefficient as the viscosity of the liquid is increased.
The axial dispersion coefficient is observed to decrease as the surface

tension is decreased in 2 1.5-in.-diam column.

The final study carried out during this report period was made to
determine the effect of using a side gas inlet, which consisted of a
1-in.-ID tube attached to the side of the column at an angle of 45° with
respect to the column axis. This type of gas inlet is envisioned for
use in continuous fluorinators having frozen-wall corrosion protection
as a means for also protecting the inlet gas nozzle from corrosion.29
Demineralized water was used during the study, which was carried out
with 2 3-in.-ID column. Data on axial dispersion obtained during this
study are summarized in Table 20. Comparison of the data with other
data on axial dispersion in a 3-in.-diam column for which the gas distrib-
utor consisted of a single, 0.432-cm-diam orifice shows very little dif-

ference between the two sets of data.

9.4 Experimental Data on Gas Holdup

Experimental data on gas holdup in open bubble columns were obtained
during this report period in the first, second, and third studies. The
ranges of operating parameters used during these two studies are summa-
rized in Table 21. Column diameters of 1.0, 1.5, 2, 3, and 6 in. were
used with gas inlet dispersers that consisted of both single and multiple
orifices having diameters that ranged from O.4 to 6.4 mm. The viscosity
of the liquid phase was varied from 0.9 to 12.1 ¢P, and the surface ten-
sion of the liquid was varied from 27 to 73 dynes/cm. The superficial
gas velocity was varied from 0.013% to 50 cm/sec. Data obtained during

the studies are summarized in Tables 22-3.4.
>

Table 20. Summary of Data on Axial Dispersion Obtained in a 3-in.-ID
Column Having a 1l-in.-ID Side Inlet Joined to the Column at an
Angle of 45° with Respect to the Column Axis

5

Tracer injection volume: ~ 5 cm

-Relative probe positiona: 0.79

 

 

Gas Flow Superficial Dispersion

Run RateP Gas Velocity Coefficient

No. (em?/sec ) (cm/sec) (cm?/sec )
1 852.8 18.7 111.9
2 706.8 15.5 159.2
3 579.2 12.7 137.2
L L83.4 10.6 132.6
5 367.6 8.06 114.9
6 277 T 6.09 91.1
T 118.8 b1k 67.3
8 98.0 2.15 4.3
9 192.0 .21 T0.7
10 162.3 3.56 Th .k
11 141.8 3,11 65.0
12 124.0 2.72 68. 4
13 oL.4 2.07 53T
14 89.4 1.96 58.0
15 72.0 1.58 49.0
16 67.9 1.49 59.1
17 4O.4 0.887 56.8
18 20.0 0.438 h7.3
19 26.2 0.57h 67.8
20 18.0 0.395 37.6
21 13.5 0.296 39.9
22 9.62 0.211 L7.1
03 5.88 0.129 31.8
ol 2.79 0.0612 28.1
25 97.1 2.13 60.7
26 141.8 3.11 65.%
27 9l.2 2.00 54.9
28 21.9 0.481 42.0
29 218.9 4.80 T1.8
30 5.65 0.124 29.5

 

%Ratio of distance of probe from surface of gas-liquid mixture to total
height of gas-liquid mixture.

bMeasured under conditions at top of column.
Th

Table 21. Ranges of Parameters During Studies of
Gas Holdup in Open Bubble Columns

 

 

Parameter Values Used
Column Diameter, in. 1.0, 1.5, 2, 3, 6
Number of Orifices in Gas Inlet 1, 5, 19, 37
Gas Inlet Orifice Diameter, mm 0.4, 1, 2, 4, 6.4
Surface Tension of Liquid, 27, 45, 70, 73
dynes/cm
Viscosity of Liquid, cP 0.9, 1.8, 2.05, 10.7, 11.3, 12.1

Gas Superficial Velocity, cm/sec 0.013 to 49.7

 

9.5 (Correlation of Data on Gas Holdup

The data on gas holdup obtained in the second and third studies of
this report period and in studies by Bautista28 were correlated by the
method of least squares. These data, a total of 349 holdup determina-
tions, cover column diameters from 1.5 to 6 in. and include a range of
values for the superficial gas velocity and physical properties of the
liquid. The results obtained for the l-in.-diam column in the first
study were not used in developing the correlation since these values are
believed to be of a lower quality than the remaining ones. It was found

that gas holdup could be represented by the relation

Vv

h - 8 — (33)
1.468 Vg + 0.407IY gd

 

where
h = fraction of column volume occupied by gas,
V_ = superficial gas velocity, cm/sec,
g = acceleration of gravity, cm/secg,

d = column diameter, cm.
5

Table 22. Summary of Data on Gas Holdup Obtained in a 1.0-in.-ID
Open Bubble Column Containing Demineralized Water at 25°C

Gas inlet: one orifice, 4.3 mm in diameter

 

 

Gas Flow Superficial Bubble Rise

Run Rate? Gas Velocity Gas Velocity

No. (em?/sec ) (cm/sec ) Holdup (cm/sec )
1 gl . 6 18.7 0.492 38.0
2 387.9 76.6 0.719 106.5
3 387.9 76.6 0.719 106.5
L 331.1 65 .4 0.707 92.5
5 378.4 Th. 7 0.719 103.9
6 28%.8 56.0 0.713 78.5
T 189.2 27.3 0.641 58.2
8 9.6 18.7 0.513 36.5
9 141.9 28.0 0.6 he. T
10 236.5 6.1 0.686 67.2
11 52.0 10.3 0.309 33,3
12 9Lk.6 18.7 0,450 h1.6
13 146.6 28.9 0.529 54.6
1k 118.2 23.3% 0.492 WL
15 70.9 1L.0 0.372 37,6
16 165.6 32.7 0.568 57.6
L7 189.2 37.3 0.575 64.9
18 236.5 L6.7 0.587 79.6
19 321.7 63.5 0.642 98.9
20 227.0 Ly .8 0.584 6.7
21 5.91 1.166 0.0658 7.7
22 5.51 1.088 0.0598 8.2
23 5.12 1.011 0.0523 19.3
2k .73 0.933 0.0516 18.1
o5 .23 0.855 0.0486 17.6
26 3.9k 0.778 0.0412 18.9
27 3.55 0.700 0.0387 18.1
28 0.79 0.156 0.0098 15.9
29 6.33 1.25 0.0645 19.4
30 5.17 1.02 0.0387 26.4
31 3-95 0.78 0.0451 17.%
32 2.37 0.468 0.0254 18.4
33 0.79 0.156 0.00645 2h.2
3L 6.7h 1.33 0.068 19.6
35 5.12 1.01 0.056 18.0
36 3.56 0.702 0.0L5 15.6
37 1.98 0.390 0.023 17.0
38 0.79 0.156 0.015 10.L

 

®Measured under conditions at top of column.
T6

Table 23. Summary of Data on Gas Holdup Obtained in a 1.5-in.-ID
Column During Second Study

 

 

Gas Flow Superficial Glycerin Bubble
Run Rated Gas Velocity Concentration Gas Rise Velocity
No. (em3/sec) (cm/sec) (wt %) Holdup (cm/sec)
1 187.0 16.4 0 0.328 50.0
2 97.8 8.58 0 0.211 40.7
3 140,2 12.3 0 0.272 45.2
4 112.8 9.89 0 0.233 42.4
5 19.6 1.72 0 0.052 33.1
6 403.6 35.4 0 0.516 68.6
7 7.87 0.69 0 0.022 31.4
8 403.6 35.4 0 0.523 67.7
9 34,1 2.99 0 0.09 33.2
10 58.2 5.11 0 0.14 36.5
11 10.4 0.912 0 0.03 30.4
12 255.4 22.4 0 0.394 56.9
13 118.6 10.4 0 0.242 43.0
14 191.5 16.8 0 0.333 50.5
15 155.0 13.6 0 0.291 46.7
16 118.6 10.4 25 0.307 33.9
17 18.9 1.66 25 0.056 29.6
18 372.8 32.7 25 0.598 54.7
19 566.6 49.7 25 0.675 73.6
20 372.8 32.7 25 0.598 54.7
21 46.3 4.06 25 0.149 27.2
22 118.6 10.4 25 0.307 33.9
23 249.7 21.9 25 0.514 42.6
24 149.4 13.1 25 0.281 46.6
25 118.6 10.4 25 0.262 39.7
26 75.5 6.62 25 0.161 41.1

 

*Measured under conditions at top of column.
1T

Table 24. Summary of Data on Gas Holdup Obtained in a
2.0-in.~-ID Column During Second Study

 

 

Gas Flow Superficial Glycerin Bubble Rise

Run Rate? Gas Velocity Conc. Gas Velocity

No. (cmB/sec) (cm/sec) (wt %) Holdup (cm/sec)
1 81.1 4.0 0 0.108 37.0
2 h11.4 20.3 0 0.359 56.5
3 482.4 23.8 0 0.378 63.0
L 291.9 4.k 0 0.239 60.3%
5 11.4 0.56 0 0.012 hé.T
6 78.6 3.88 0 0.098 39.6
T 120.8 5.96 0 0.136 43,8
8 117.2 5.78 25 0.159 36.4
9 1773 8.75 25 0.207 ho.3
10 210.8 10.4 25 0.233 L. 6
11 377.0 18.6 25 0.35 5%.1
12 553.3% 27-3 25 0.466 58.6
13 56.5 2.79 25 0.075 37.2
14 81.% 4.01 25 0.106 37.8
15 10.7 0.53% 25 0.009 58.9
16 8.31 0.41 25 0.009 L5.6
17 15.2 0.75 25 0.02 37.5
18 ol .5 1.21 25 0.035 3.6
19 .8 2.21 25 0.063 35.1
20 18.4 0.91 65 0.021 43,3
21 9.93 0.49 65 0.012 40.8
22 6.28 0.31 65 0.009 3L .4
23 25.1 1.24 65 0.035 5.0
2l L7.6 2.35 65 0.054 43,5
25 60.8 3.0 65 0.073 h1.1
26 83.3 4.11 65 0.095 43,3
27 116.9 5.7T 65 0.124 h6.5
28 184 .4 9.1 65 0.18 50.6
29 208.8 10.3 65 0.199 51.8
30 226.3 16.1 65 0.254 63.L
31 50L.7 2.9 65 0.325 76.6
32 758.0 37 L 65 0.k 03.5

 

8Measured under conditions at top of column.
T8

Table 25. Summary of Data on Gas Holdup Obtained in a 6.0-in.-ID
Column During Second Study

 

 

Gas Flow Superficial Glycerin Bubble
Run Rated Gas Velocity Concentration Gas Rise Velocity
No. (cm3/sec) (cm/sec) (wt 7%) Holdup (em/sec)
1 1530 8.39 0 0.146 57.5
2 561.8 3.08 0 0.071 43.4
3 145.9 0.8 0 0.021 38.1
4 94.8 0.52 0 0.016 32.5
5 94.8 0.52 0 0.016 32.5
6 271.8 1.49 0 0.036 41.4
7 394.0 2.16 0 0.054 40.0
8 1299 7.12 0 0.128 55.6
9 113.1 0.62 0 0.016 38.8
10 698.6 3.83 0 0.081 47 .3
11 195.2 1.07 0 0.028 38.2
12 60.2 0.33 0 0.009 36.7
13 286.4 1.57 0 0.034 46.2
14 1665 9.13 0 0.156 58.5
15 899.3 4.93 0 0.089 55.4
16 479.8 2.63 25 0.056 47.0
17 1372 7.52 25 0.16 47.0
18 1757 9.63 25 0.217 44,4
19 1757 9.63 25 0.217 44,4
20 474.3 2.6 25 0.076 34.2
21 923.0 5.06 25 0.126 40.2
22 1572 8.62 25 0.196 44,0
23 58.4 0.32 25 0.008 40.0
24 372.1 2.04 25 0.06 34.0
25 217.1 1.19 25 0.032 37.2
26 543.6 2.98 25 0.074 40.3
27 707.8 3.88 25 0.09 43.1
28 543.6 2.98 25 0.074 40.3
29 1094 6.0 25 0.118 50.8
30 220.7 1.21 65 0.028 43.2
31 295.5 1.62 65 0.038 42.6
32 727.8 3.99 65 : 0.081 49.3
33 521.7 2.86 65 0.066 43.3
34 361.2 1.98 65 0.052 38.1
35 1096 6.01 65 0.105 57.2
36 1532 8.4 65 0.13 64.6
37 93.0 0.51 65 0.013 39.2
38 136.8 0.75 65 0.017 44,1

 

#Measured under conditions at top of column.
9

Table 26. Summary of Data on Gas Holdup Obtained in a 1.5-in.-ID
Column Containing Water During Third Study

Gas inlet: one orifice, 1 mm ID

 

 

Gas Flow Superficial Bubble Rise

Run Rate? Gas Velocity Gas Velocity

No. (em?/sec) (cm/sec) Holdup (cm/sec )
1 2.42 0.212 0.00862 oh.6
2 8.20 0.719 0.0308 23.3
3 15.2 1.33 0.0424 31.4
L 19.2 1.68 0.0531 31.6
5 25.2 2.21 0.0675 32.7
6 WL, o 3.86 0.121 31.9
T Th.6 6.5k 0.183 35.7
8 101.8 8.93 0.250 357
9 2h2.8 21.3 0.427 49.9

 

a
Measured under conditions at top of column.

Table 27. Summary of Data on Gas Holdup in a 1.5-in.-ID Column
Containing Aqueous Isopropanol During Third Study

Surface tension of liquid: L45.3 dynes/cm

Gas inlet: one orifice, 0.638 cm ID

 

 

Gas Flow Superficial Bubble Rise

Run Rate? Gas Velocity Gas Velocity

No (cm?/sec ) (em/sec) Holdup fem/sec)
1 11.9 1.045 0.034L 30,4
2 8.09 0.710 0.0247 28.7
3 6.10 0.535 0.019 28.2
4 2.87 0.252 0.00976 25.8
5 1.06 0.0930 0.00423 22.0
6 16.6 1.456 0.048T7 29.9
7 17.7 1.55 0.0567 27.3
8 29.4 2.58 0.089 29.0
9 9.1 h.31 0.149 28.9
10 66.0 5.79 0.18% 21.6
11 84.9 745 0.167 .6

 

8Measured under conditions at top of column.
80

Table 28. Summary of Data on Gas Holdup in a 1.5-in.-ID Column
Containing Aqueous Isobutanol During Third Study

Surface tension of liquid: 27.3% dynes/cm
Gas inlet: one orifice, 0.638 cm ID

 

 

Gas Flow Superficial Bubble Rise

Run Rate?® Gas Velocity Gas Velocity
No. (emd/sec) (cm/sec) Holdup (cm/sec)

1 11.6 1.02 0.0328 31.1

2 1.52 0.133 0.00463 28.7

3 0.80L 0.0705 0.00278 25.4

b4 2.44 0.214 0.0088 24.3

5 5.6% 0.494 0.017k 28.4

6 8.70 . 0.76% 0.0276 27.6

7 15.6 1.37 0.052 26.3

8 23.6 2.07 0.081 25.6

9 40.6 3.56 0.141 25.2
10 73.0 6.4 0.247 25.9
11 98.0 8.6 0.38 22.6

 

a
Measured under conditions at top of column.

Table 29. Summary of Data on Gas Holdup Obtained in a 1.5-in.-ID
Column Containing a Water-Glycerin Mixture During Third Study

Gas inlet: one orifice, 0.638 cm ID

 

 

Gas Flow Superficial Glycerin Bubble Rise

Run Rate? Gas Velocity Conc. Gas Velocity
No. (em3/sec) (cm/sec) (wt %) Holdup (cm/sec)

1 2.13 0.187 25 0.00793 23.6

2 3.80 0.333% 25 0.0126 26.4

3 6.02 0.528 25 0.0194 27.2

L 10.2 0.895 25 0.0298 30.0

5 4.2 1.25 25 0.0387 20,3

6 18.5 1.625 25 0.052 31.3

7 29.8 2.61 25 0.0732 35,7

g 65.9 5.78 25 0.171 33,8

3 92.0 8.07 25 0.20k 39.6
1o 3.57 0.313 65 0.0101 31.0
11 5.13 0.45 65 0.016 28.1
12 7.4p 0.651 65 0.0216 30.1
13 11.4 1.0 65 0.0335 29.9
1L 15.7 1.%8 65 0.0L435 31.7
15 19.8 1.7k 65 0.0554 31.4
16 33.1 2.9 65 0.0925 31.4
17 66.0 5.79 65 0.227 25.5

 

a Lo
Measured under conditions at top of column.
81

 

 

Table 30. Summary of Data on Gas Holdup Obtained in a 2.0-in.-ID
Column Containing Water During Third Study
Gas inlet: one orifice, 1 mm ID
Gas Flow Superficial Bubble Rise
Run Rate? Gas Velocity Gas Velocity
No. (em?/sec) (em/sec) Holdup (cm/sec)
1 0.896 0.04L2 0.00225 19.6
2 2.55 04126 0.00L5 28.0
3 h.2k 0.209 0.00625 33,0
L 5.98 0.295 0.00936 31.5
5 8.21 0.405 0.012 33.8
6 10.4 0.515 0.0146 35.3
7 12.5 0.618 0.0186 33,0
8 16.2 0.798 0.0216 36.9
9 19.2 0.95 0.0252 7.7
10 30.8 1.52 0.0L437 34.8
11 59.2 2.92 0.07 h1.7
12 92.8 L.58 0.107 h2.8
13 145.9 7.2 0.152 hy.L

 

®Measured under conditions at top of column.
Table 1. Summary of Data on Gas Holdup Obtained in a Z-in.-ID Column Containing Water During Third Study

 

 

 

Number of orifices:
Gas Flow superficial pubple 1 G Fl Superticial Bubble
Rate 2 Gas Gas Inlet Rise a;ategw Gas Gas Inlet Rise
Run . Velocity Orifice ID Gas Velocity Run Velocity Orifice ID Gas Velocity
No. {em /sec) (cm/sec ) (mm ) Holdup  (cm/sec) No. (cm5/sec) (cm/sec ) (mm ) Holdup (cm/sec)
1 .80 0. %65 0.4 0.0142 27.1 35 h.76 0.235 2.0 0.0129 18.2
;2 0.926 0. 0Ly 0.k 0.00283 16.1 36 10.2 0.503 2.0 0.0237 21.2
5 2.50 0.123y O.h4 0.00k(3 26.1 37 20.3 1.0 2.0 0.0302 33.1
L L.9o 0,02 O.4 0.00988 2h .y 38 5.13% 0.253 4.0 0.0075 33,7
g 10.9 0.54%7¢ Ok 0.0186 28.9 39 10.3 0.509 L.o 0.0219 25.2
6 1,.2 0.k 0.k 0.0276 7.1 ~ Lo 31.2 1.54 4.0 0.0376 41.0
o 19.4 0. 96 0.h 0.03544 27.9 41 8h.5 L1t 4.0 0.098 4o.6
8 344 1.7 0.4 0.0632 26.9 42 1.90 0.0938 L.o 0.0033%5 28.0
9 65.3 3.22 Ouh 0.101 31.9 43 10.4 0.512 4.0 0.0159 32.2
10 97.1 ] 0.4 0.127 7.7 LYy Ls.8 2.26 L.0o 0.0713 39.7
11 141.9 ‘.00 0.4 0.157 4h.6 4s 111.5 5.5 4.0 0.125 k.o
12 171.5 846 0.k 0.176 48.1 L6 450 2.2 L.o 0.339 65.5
15 13.9 0.688 1.0 0.0205 33.6 47 282 13.9 4.0 0.24k7 56.3
1h 1.5 0.0767 1.0 0.00264 29.1 43 1y. f 0.9 4.0 0.026% 56.9
1. 3.20 0.158 1.0 0.0048¢ 32.6 49 2.45 c.121 4.0 0.00418 25.9
1 L 0.26 1.0 0.00615 42.3 50 26.3 1.3 4.0 0.0346 37-6
17 T.0% O. 547 1.0 0.00835 L1.6 51 14.6 0.72 4.0 0.0188 38.%
18 3,30 0459 1.0 0.011hL 40.3 52 7.56 0.373 L.o 0.0368 10.1
19 11.9 0.589 1.0 0.014 Lol 53 1.04 0.0-12 2.0 0.00176 29.1
20 1§.6 0.92 1.0 0.025% 36.4 54 2.53 0.125 2.0 0.00Uk 28.4
21 50.8 1.52 1.0 0.0405 37-5 55 L.09 0.202 2.0 0.0057 35.4
22 59. & 2.95 1.0 0.0738 40.0 56 6.16 0.304 2.0 0.0105 28.9
2% - 86.5 .oy 1.0 0.10 k2.7 57 8.21 0.405 2.0 0.018 22.5
24 113.3 5.59 1.0 0.124 4.1 58 10.6 0.525 2.0 0.0219 24.0
25 147.8 T.29 1.0 0.127 57.4 59 13.2 0.65 2.0 0.0175 37.1
26 280 15.8 1.0 0.228 £0.5 €0 15.4 0.76 2.0 0.0145 52.4
27 88.5 Lh.35 2.0 0. 107 LO.T 61 18.7 0.923 2.0 0.024T 37-4
28 5.37 0.265 2.0 0.011% 23.5 62 30.9 1.525 2.0 0.0k413 26.9
29 16.2 0.80 2.0 0.0543 23.% 63 63.8 3.15 2.0 0.0745 L2.3
50 50.7 2.4 2.0 0. 061 L1.0 o 89.2 i 2.0 0.0982 4h.8
31 185.4 9.1% 2.0 0.18 50.8 65 117.8 5.81 2.0 0.127 Ls.7
32 458 21.6 2.0 0. 529 65.7 66 1hh .5 7.13 2.0 0.15% hA. 0
53 2.2% 0.11 20 0.0052 21.2 67 330 16.5% 2.0 0.26 €2.7
34 k.ol U198 2.0 0.0082 2h.1

 

 

e

 

a
Measured under conditions at top of column.
83

Table 32. Summary of Data on Gas Holdup Obtained in a 2.0-in.-~ID
Column During Third Study

Gas inlet: 19 orifices, 1 mm ID

 

 

Gas Flow Superficial Bubble
Run Rate? Gas Velocity Gas Rise Velocity
No. (cm3/seC) (cm/sec) Holdup (cm/sec)
1 1.02 | 0.0505 0.00219 23.1
2 2,31 0.114 0.00392 29.1
3 4.22 0.208 0.0061 34.1
4 6.10 0.301 0.0087 34.6
5 8.19 0.404 0.0117 34,5
6 10.8 0.535 0.0156 34.3
7 12.8 0.632 0.0173 36.5
8 14.7 0.726 0.021 34.6
9 18.6 0.919 0.0253 36.3
10 30.0 1.48 0.0353 41.9
11 61.2 3.02 0.0748 40.4
12 85.9 4.24 0.11 38.5
13 115.7 5.71 0.133 42,9
14 149.6 7.38 0.158 46.7
15 280 13.8 0.249 55.4

 

dMeasured under conditions at top of column.
8l

Table 33. Summary of Data on Gas Holdup Obtained in a 2.0-in.-ID Column
Containing Water During Third Study

Gas inlet: 37 orifices

 

Gas Flow Superficial Gas Inlet ' Bubble

 

Run Rate? Gas Velocity Orifice ID Gas Rise Velocity
No. (cm3/sec) (cm/sec) (mm) Holdup (cm/sec)
1 1.06 0.0525 1.0 0.00307 17.1
2 2.41 0.119 1.0 0.00438 27.2
3 3.46 0.171 1.0 0.00788 21.7
4 5.86 0.289 1.0 0.00918 31.5
5 7.96 0.393 1.0 0.0118 33.3
6 10.7 0.529 1.0 0.0144 36.7
7 13.9 0.687 1.0 0.0195 35.2
8 15.8 0.78 1.0 0.0282 27.7
9 19.2 0.945 1.0 0.0288 32.8
10 30.8 1.52 1.0 0.0474 32.1
11 60.2 2.97 1.0 0.0895 33.2
12 88.4 4.36 1.0 0.097 44.9
13 113.5 5.6 1.0 0.144 38.9
14 149.0 7.35 1.0 0.174 42.2
15 259 12.8 1.0 0.233 54.9
16 503 24.8 1.0 0.325 76.3
17 746 36.8 1.0 0.395 93.2
18 2.27 0.112 2.0 0.0039 28.7
19 7.62 0.376 2.0 0.0112 33.6
20 13.6 0.67 2.0 0.0185 36.2
21 1.42 0.0703 2.0 0.00175 40.2
22 5.31 0.262 2.0 0.00825 31.8
23 9.69 0.478 2.0 0.0148 32.3
24 17.7 0.875 2.0 0.024 36.5
25 27.8 1.37 2.0 0.0372 36.8
26 44.8 2.21 2.0 0.0592 37.3
27 101 5.0 2.0 0.12 41.7
28 72.6 3.58 2.0 0.0895 40.0
29 3.69 0.182 2.0 0.0065 28.0
30 148 7.28 2.0 0.16 45.5
31 590 29.12 2.0 0.385 75.6
32 14.9 0.737 4.0 0.0218 33.8
33 1.25 0.0618 4.0 0.0026 23.8
34 3.00 0.148 4.0 0.00488 30.3
35 5.09 0.251 4.0 0.00574 43.7
36 6.71 0.331 4.0 0.00795 41.6
37 8.17 0.403 4.0 0.015 26.9
38 11.8 0.585 4.0 0.0185 31.6
39 20.5 1.01 4.0 0.0282 35.8
40 30.2 1.49 4.0 0.0424 35.1
41 53.9 2.66 4.0 0.071 37.5
42 73.4 3.62 4.0 0.0876 41.3
43 94.4 4.66 4.0 0.109 42.8
44 118 5.8 4.0 0.131 44.3
45 150 7.43 4.0 0.155 47.9
46 537 26.5 4.0 0.34 77.9

 

a .
Measured under conditions at top of celumn.
85

Table 34. Summary of Data on Gas Holdup Obtained in a 3.0-in.-ID
Column Containing a Water—Glycerin Mixture During Third Study

Gas inlet: one orifice, 0.638 cm ID

 

 

Gas Flow Superficial Glycerin Bubble
Run Ratead Gas Velocity Concentration Gas Rise Velocity
No. (cm3/sec) (cm/sec) (wt 72) Holdup (cm/sec)
1 36.9 0.81 25 0.0252 32.1
2 1.60 0.0352 25 0.000928 37.9
3 3.50 0.0768 25 0.0045 17.1
4 5.56 0.122 25 0.004 30.5
5 8.94 0.196 25 0.0056 35.0
6 13.0 0.285 25 0.01 28.5
7 19.4 0.425 25 0.0107 39.7
8 25.2 0.553 25 0.0162 34,1
9 46.7 1.025 25 0.0295 34.7
10 84.8 1.86 25 0.0492 37.8
11 130.0 2.85 25 0.0765 37.3
12 258 5.66 25 0.113 50.1
13 377 8.26 25 0.171 48.3
14 1.47 0.0323 25 0.00141 22.9
15 3.51 0.0769 65 0.00188 40.9
16 4.92 0.108 65 0.00327 33.0
17 7.07 0.155 65 0.00468 33.1
18 10.6 0.232 65 0.00658 35.3
19 15.4 0.338 65 0.00889 38.0
20 19.3 0.424 65 0.0117 36.2
21 30.8 0.675 65 0.0199 33.9
22 76.2 1.67 65 0.043 38.8
23 57.0 1.25 65 0.0316 39.6
24 104 2.28 65 0.0628 36.3
25 135 2,96 65 0.0701 42.2
26 245 5.37 65 0.114 47.1

 

aMeasured under conditions at top of column.
86

A comparison of the experimental data with values predicted by the corre-
lation is shown in Fig. 22. Equation (33) predicts that gas holdup
should be essentially proportional to superficial gas velocity at low
gas flow rates and that the dependence should decrease as the gas flow
rate is increased. A limiting value for gas holdup of about 0.68 is

predicted.

The following expression for the bubble rise velocity, Vg/h, can be

obtained from Eq. (33):
% \Ved
= = 1.468 vg + 0.4071 gdc . (34)

This relation predicts that the bubble rise velocity should be essen-
tially constant at low gas flow rates and should depend on the square
root of the column diameter. The bubble rise velocity should increase
at a rate that is essentially proportional to the superficial gas
velocity at high gas flow rates and should show little dependence on
the column diameter. The physical properties of the liquid are of
negligible importance in determining either gas holdup or bubble rise

velocity.

Comparisons of the predicted and experimentally determined values
for gas holdup and bubble rise velocity are shown in Figs. 23-40. The
measured and the predicted values are seen to be in good agreement.

Most of the predicted values lie within 15% of the measured values.
Figures 23-32 show the variations of gas holdup and bubble rise velocity
resulting from changes in the viscosity of the liquid phase and the
superficial gas velocity for columns having diameters of 1.0, 1.5, 2,

%, and 6 in. In the bubble flow regime, the bubble rise velocity is
essentially constant at about 25 to 40 cm/sec, and the gas holdup
increases almost linearly as the superficial gas velocity is increased.
Increases in the column diameter result in small increases in the bubble
rise velocity and small decreases in the gas holdup. In the slug flow
regime, the bubble rise velocity increases with increases in the super-
ficial gas velocity and the gas holdup is not linearly dependent on

the superficial gas velocity.
87

 

 

COMMON LOGARITHM OF MEASURED (Vg /h)

 

 

e | ] | ] |
1.4 15 1.6 0.7 1.8 19 2.0
COMMON LOGARITHM OF PREDICTED (Vg/h)

 

Fig. 22. Comparison of Experimentally-Determined and Calculated
Data on Gas Holdup in Open Bubble Columns Having Diameters Ranging from
1.5 to 6 in.
83

ORNL DWG 72-11742

 

1T T T 11

T

o.

T T TTT]

GAS HOLDUP

 

  

T T T 1771717171 T |

  
   

PREDICTED VALUES

Lit1i1al

1

 

 

0.0l 5
:
0.00I i 1113111l 1 L3 1 13119l 1 L 1.1 1111
o.1 1.0 10 100
SUPERFICIAL GAS VELOCITY (cm /sec)
Fig. 25. Variation of Gas Holdup with Changes in Superficial Gas

Velocity in a 1.0-in.-ID Bubble Column.
BUBBLE RISE VELOCITY (cm/sec)

ORNL DWG 72-11743

 

 

 

 

 

 

140 1 T T 1T T T TT] T T T
120 -
2 pts.
v
100 - —
80 - —
60 -
40 |- =
PREDICTED VALUES
o O
20 o oy —d
o o © 860
©
0 L 1 1111 1 1 i P 11 1al 1 1 11 1 111
0.1 1.0 10 100

SUPERFICIAL GAS VELOCITY(cm /sec)

Fig. 24. Variation of Bubble Rise Velocity with Changes in Super-
ficial Gas Velocity in a 1.0-in.-ID Bubble Column.

€3
90

ORNL DOWG 72-13507

 

LERLLIL

»
I

GAS HOLDUP

 

T T VrrrTg f T rrreg | T 1T 7T 7T

—
—
-

   

PREDICTED
VALUES

 

L 11l

)

1

1

VISCOSITY (¢P)
- 0.89

O 208
A 1.3

Ll

1

L1 1]l

 

L 11l 1 L 1

L 111l ] 1 1

 

0.001
0

Fig.

Velocity

25.

1.0 10
SUPERFICIAL GAS VELOCITY (cm/sec)

100

Variation of Gas Holdup with Changes in Superficial Gas

and Viscosity of Liquid in a 1.5-in.-ID Bubble Column.
BUBBLE RISE VELOCITY (ecm/sec)

ORNL DWG 72-13506

 

140 T T T T T T 17717 1 T T

T UTTg I 1 v rrrT

VISCOSITY (cP)

  
 

120 e 0.89 -
e 205
a L3
IOO = -t
80 | 4
PREDICTED VALUES
60 - ey \Zp't - \fi

40 | , 1
® .. .
A & e %M
® ®

 

 

 

 

—— A
e
20 e -
0 1 1 1 1 11 11 1 i 1 1 1 1 2l 1 1 1 A 14 i 1
0.1 I 10 100

SUPERFICIAL GAS VELOCITY (cm /sec)

Fig. 26. Variation of Bubble Rise Velocity with Changes in Super-
ficial Gas Velocity and Viscosity of Liquid in a 1.5-in.-ID Bubble
Column.
92

ORNL DWG 72-13519

 

 

 

 

 

|.0_ T T T !riiI] 1 T T T T 1117 T T T rriT)]

e A -y

= . m -

[ /‘3 4

0
- . EP -
PREDICTED VALUES fi/
B .

0.1 : =

l —

: -
o

3 -
o

T VISCOSITY -

g (cP)
© . 0.89 7
a 2.05
0.9 D .3 —
0.001 1 1 1411l L 11t 2114l 1 1 11 1 13}
0.1 1O 10.0 100.0

SUPERFICIAL GAS VELOCITY (cm/sec)

Fig. 27. Variation of Gas Holdup with Changes in Superficial Gas
Velocity and Viscosity of Liquid in a 2.0-in.-ID Bubble Column.
BUBBLE RISE VELOCITY (cm/sec)

ORNL DWG 72-13518

 

140 T 7 T T T T, T T T T T T T T T Y

-

VISCOSITY (cP)

120 |- * 0.89 4
a 1.8
- @ 2.1 -
100 | i
. | o
80 |- i
e

  

 

 

 

 

40 o 4
e ? e * o a. * D °
’ * PREDICTED VALUES
20} | | :
0 L1 | | 1 11 1a1 11 1 11 L1 1 11 B 1 1 L1 11 l 4.3 111
0.04 0.05 007 ol 02 05 Q7 |10 20 130 50 70 10 20 40

SUPERFICIAL GAS VELOCITY (cm /sec)

Flg 28. Variation of Bubble Rise Velocity with Changes in Super-
ficial Gas Velocity and Viscosity of the Liquid in a 2.0-in.-ID Bubble
Column.

¢6
GAS HOLDUP

9L

ORNL DWG 72-1352I|

 

       

 

 

 

1.0 T T 0T YT T T Illill[ T Y T Uiy T Y T T rrT7T
[ VISCOSITY (¢P) 3
- e 0.89
- A 2.05 -
- g 10.7 -
0.4} -
- -
— PREDICTED VALUES 1
0.0l |- —
n ]
- -
a
0.00! i £ 1t il 1t 4 vl i 1t e et by 11
0.01 0.1 1.0 10.0 100.0

SUPERFICIAL GAS VELOCITY (cm/sec)

Fig. 29. Variation of Gas Holdup with Changes in Superficial Gas
Velocity and Viscosity of Liquid in a 3.0-in.-ID Bubble Column.
BUBBLE RISE VELOCITY (¢m/sec)

140

130

120

11O

90

80

70

60

50

40

30

20

95

ORNL DWG 72-13517

 

s

VISCOSITY (cP)

®
a
o

T rrIrty 1 v rTrTrrirng ! 1 TV rrrry v T 1T 7T 7177177

0.89
2.05
10.7

    

PREDICTED VALUES ~

05

 

 

 

1111l 1 L Lt 11l i L4 431911 L L.l 1 1111

 

Fig. 30.

0.l 1.0 10.0 100.0
SUPERFICIAL GAS VELOCITY (cm/sec)

Variation of Bubble Rise Velocity with Changes in Super-

ficial Gas Velocity and Viscosity of Liquid in a 3.0-in.-ID Bubble
Column.
96

ORNL DWG 72-13516

 

 
   
 

 

 

 

1.0 T T T T T T7TT] T T T T rTTITg 1 — T T I

- 3

0.1 - -

a - ]

=

g VISCOSITY (cP) -

T — . 0.89 -
o A 1.8

- - B 12.1 -

PREDICTED VALUES
0.0l .
0.00!| i L4 111l 1 11 1133l 1 1 1 1 1111
0.1 1.0 10.0 100.0

SUPERFICIAL GAS VELOCITY (cm/sec)

Fig. 51. Variation of Gas Holdup with Changes in Superficial Gas
Velocity and Viscosity of Liquid in a 6.0-in.-ID Bubble Column.
(cm/sec)

BUBBLE RISE VELOCITY

70

50

40

20

ORNL DWG. 72-13515

 

a
i PREDICTED VALUES / -
*

° VISCOSITY (cp)
. 0.89 -
A 1.8
a 12.1

 

 

1 i ) i 1 1 1 1

 

0.1 0.2 0.5 1.0 2 5 10 20 50 100

SUPERFICIAL GAS VELOCITY (cm/sec)

Fig. %2. Variation of Bubble Rise Velocity with Changes in Super-
ficial Gas Velocity and Viscosity of Liquid in a 6.0-in.-ID Bubble
Column.

L6
GAS HOLDUP

g8

ORNL DWG 72-13522

 

     

 

 

 

|.0:‘ 1 1 FTETITTg I T Flllll] ; ! ! VITTETT I T T TTT1

:SURFACE TENSION ]
- (dynes/cm)

[ . 7200 A _ )

» L 453 -
& 273
0.1 - 4
I PREDICTED VALUES i
0.0l .:
0.00| 1 1 ¢ttt 1 Lt 1istal 1 1 4 11l 1 L L L 111y
0.0l 0.1 | 10 |00

SUPERFICIAL GAS VELOCITY (cm/sec)

Fig. 33. Variation of Gas Holdup with Changes in Superficial Gas
Velocity and Surface Tension of Liquid in a 1.5-in.-ID Bubble Column.
BUBBLE RISE VELOCITY (cm/sec)

ORNL DWG 72-13514

 

   

 

 

 

 

80 T T 1] T T T T 7 T T 1] T T T 1T 171t [ I T T
SURFACE TENSION
(dynes/cm)
. T2
701- O 453 .
A 723 o*
60+ —
50 —
a0 -
o
PREDICTED o *
VALUES
30 A 20 —
A D A A A
a .
Q A
20} 4
jol—1 11 Ll 1 | Lt 1t ] i L1 11l ] i |
0.05 0.1 0.2 0.5 1.0 2.0 50 10 20 50

SUPERFICIAL GAS VELOCITY (cm/sec)

Fig. 34. Variation of Bubble Rise Velocity with Changes in Super-
ficial Gas Velocity and Surface Tension of Liquid in a 1.5-in.-ID Bubble
Column.

66
GAS HOLDUP

100

ORNL DWG 72-13523RI

 

|0 T ]|]||||] T ]]][1[[| T I‘Illllll T r 17T . 7T71T07

 

 

 

 

C  ORIFICE DIAMETER -
t {mm) .
N + 0.4 i
» a |0 , i

O 20 PREDICTED VALUES R0
-

© 40 g°
Ol | -
r_ .
0.0l - .
. i
0.001 i L 111l 1 i 4 1ok 1 1 1 i L4 1.1
0.01 0.1 t.0 10.0 100.0

SUPERFICIAL GAS VELOCITY (cm/sec)

Fig. 35. Variation of Gas Holdup with Changes in Superficial Gas
Velocity and Diameter of Orifices in Gas Distributor in a 2.0-in.-ID
Bubble Column Filled with Water. The gas distributor consisted of five
orifices of the diameter indicated.
BUBBLE RISE VELOCITY (cm/sec)

101

ORNL DWG. 72-13513

 

80 g ! T T T T T
ORIFICE DIAMETER
(mm)
. 0.4
70} A |.0 4
® 2.0
+ 4.0
60 -
a
50+ o
A A .
40} 4 a + "a -
u+ » + Q
‘. A VALUES

 

f
‘\‘

 

 

 

. ® ® .
s ‘g a s s
20} - = _
»
10 ] 1 i L 1 i l L
0.04 0.08 0.2 04 08 20 40 8.0 20 40

SUPERFICIAL GAS VELOCITY (cm/sec)

Fig. %6. Variation of Bubble Rise Velocity with Changes in Super-
ficial Gas Velocity and Diameter of Crifices in Gas Distributor in a 2.0-
in.-ID Bubble Column Filled with Water. The gas distributor consisted of
five orifices of the diameter indicated.
102

ORNL DWG 72-135I12R!I

 

     
  
   

 

 

 

1.0 T T T TTTTT T T. T 7T TTT] T T T T T TT71]
— -
|-
0.l i :
— PREDICTED VALUES ]
o i -
D — -
Q
- [~ -
o
I -
wn
g — -
o
0.0l = —
0.00| 1 1 L1 1 1 1 11| i 1 L1
0.05 0.1 03 0.5 1.0 3 5 10 30 50

SUPERFICIAL GAS VELOCITY (cm/ sec)

Fig. 5(. Variation of Gas Holdup with Superficial Gas Velocity in
a 2.0-in.-ID Bubble Column Filled with Water. The gas distributor con-
sisted of 19 orifices, each having a diameter of 1.0 mm.
BUBBLE RISE VELOCITY (cm/sec)

ORNL DWG 72-1351I

 

TOr

60 -
50 -
40 |-

30 |

 

  

PREDICTED VALUES

i L 1 1 1 L 1 |

 

 

0.1 0.2 0.5 | 2 5 10 20
SUPERFICIAL GAS VELOCITY (cm/sec)

Fig. 38. Variation of Bubble Rise Velocity with Superficial Gas

Velocity in a 2.0-in.-ID Bubble Column Filled with Water. The gas
distributor consisted of 19 orifices, each having a diameter of 1.0

min «

50

¢01
10k

| ORNL DWG 73-Ti7
0 T T T TTTTT T ™ T T TTTTT T =T T T

0.7

 

e
—
—

1

0.5

C.3-

 
  

PREDICTED

Ol VALUES

 

0.07

L1 111t

0.05

]

0.03

0.02 ORIFICE DIAMETER (mm)
. 1.0
A 2.0

B 4.0

GAS HOLDUP

0.0l

0.0C7
0.005

11 1 111

i

0.0C3

0002

 

 

0.001 1t 111 g1 1 L L1111 ] 1 1 (htr]
0.05 ol 02 03 05 1.0 20 30 50 10 20 30 50

SUPERFICIAL GAS VELOCITY (cm/sec)

 

Fig. 39. Variation of Gas Holdup with Changes in Superficial Gas
Velocity and Diameter of Orifices in Gas Distributor in a 2.0-in.-ID
Bubble Column. The gas distributor consisted of 37 orifices of the
diameter indicated.
BUBBLE RISE VELOCITY {(cm /sec)

ORNL DWG 72-135I10

 

    

140 T T T T T T T T
. ORIFICE DIAMETER
120 | (mm) _
e |.0
A 20
100 - g 4.0
80

PREDICTED VALUES

50T

 

 

 

o I 1 l 1 ] 1 1 ]
0.05 0.1 0.2 0.5 I 2 5 10 20 50

SUPERFICIAL GAS VELOCITY (cm/ sec)

 

Fig. 40O. vVariation of Bubble Rise Velocity with Changes in Super-
ficial Gas Velocity and Diameter of Orifices in Gas Distributor in a
5.0-in.-ID Bubble Column. The gas distributor consisted of 37 orifices
having the indicated diameter.
106

The variations of gas holdup and bubble rise veloéity with changes
in the surface tension of the liquid and the superficial gas velocity
are shown in Figs. 33 and 34 for a column diameter of 1.5 in. Decreases
in the surface tension tend to increase the gas holdup, to decrease the
bubble rise velocity, and to cause the transition from bubble flow to
slug flow to occur at higher superficial gas velocities; on the other

hand, the importance of changes in these parameters is very slight.

Data on gas holdup and bubble rise velocity obtained in a 2-in.-diam
column with multiple gas inlets having diameters of 0.4, 1.0, 2.0, and
4.0 mm are shown in Figs. 35-40. Within the accuracy of the data, there
is no dependence of gas holdup or bubble rise velocity on the number or
diameter of the orifices in the gas distributor except in the case of
the smallest orifice diameter studied (O.4 mm). The bubble rise velocity
observed with the O.4-mm-diam orifice is appreciably lower, and the gas

holdup is higher, than for the larger orifice diameters.

9.6 Compilation of Data on Axial Mixing

A compilation was made of the axial dispersion data that have been
accumulated in experiments made thus far. This was done in preparation
for developing a general correlation for predicting axial dispersion in
open bubble columns. The information, which consists of about 420
measurements of the axial dispersion coefficient in columns with diameters
ranging from 1.0 to 6.0 in., was divided into 34 data sets (as shown in
Table 35). The results obtained in the individual runs with column
diameters of 1.0, 1.5, 2.0, 3.0, and 6.0 in. are summarized in Tables 36,

37, 38, 39, and 40, respectively.

9.7 Correlation of Data on Axial Dispersion

The axial dispersion data were analyzed using dimensional analysis
and were fitted by the method of least squares to the resulting power-

law expression. The power-law expression is of the form
107

Table 33. Description of Data Sets for Data on Axial Dispersion in Open Bubble Columns

 

Data Column Orifice Number Fluid Surface Number

 

Set Diameter Diameter of Viscosity Tension Density  of Data
No. (in.?} {cm) Orifices (cP) {(dynes/cm) {g/cm3) Points Reference
1 1.5 0.432 1 0.894 73.0 0.997 15 a
2 2.0 0.432 1 0.894 73.0 0.997 7 a
3 6.0 0.432 1 0.894% 73.0 0.997 16 a
4 2.0 0.1 1 0.894 73.0 0.997 13 b
5 2.0 0.04 5 0.89% 73.0 0.997 12 b
b 2.0 0.1 5 0.89% 713.0 0.997 14 b
7 2.0 0.2 5 0.894 73.0 0.997 26 b
8 2.0 0.4 5 0.894 73.0 0.997 16 b
9 2.0 0.1 19 0.894 73.0 0.997 15 b
10 2.0 0.1 37 0.894 73.0 0.997 17 b
11 2.0 0.2 37 0.894 73.0 0.997 14 b
12 2.0 0.4 37 0.894 73.0 0.997 15 b
13 1.5 0.432 1 0.894 73.0 0.997 11 b
14 1.5 0.432 1 1.07 45,3 0.995 11 b
15 1.5 0.432 1 1.09 27.3 0.996 11 b
16 1.5 0.432 1 2.05 71.5 1.04 10 b
17 1.5 0.432 1 11.3 70.0 1.15 10 b
18 3.0 0.432 1 2.05 71.5 1.04 14 b
19 3.0 0.432 1 10.7 70.0 1.15 13 b
20 2.0 0.432 1 1.8 71.8 1.05 12 a
21 2.0 0.432 1 12.1 67.9 1.16 13 a
22 1.5 0.432 1 1.8 71.8 1.05 11 a
23 h.0 0.432 1 1.8 71.8 1.05 13 a
24 6.0 0.432 1 12.1 67.9 1.16 10 a
25 2.0 0.102 1 0.894 37.6 0.997 6 23
26 2.0 0.102 1 0.89% 53.2 0.997 4 23
27 2.0 0.102 1 0.894 73.0 0.997 5 23
'8 2.0 0.102 1 15.0 65.2 1.15 5 23
29 2.0 (G.102 1 0.89% 67.4 0.997 1 23
30 2.0 0.102 1 0.894 73.0 0.997 21 22
31 1.5 0.432 1 0.894 73.0 0.997 7 25
52 2.0 0.432 1 0.89% 73.0 0.997 13 75
33 3.0 0.432 1 0.894 73.0 0.997 9 25
1L 1.0 0.432 1 0.8%4 73.0 0.997 38 C

 

 

dhata obtained during second study.
I
Pata obtained during third study.

"bata obtained during first study.
Table 36. Experimentally Determined Values® for the Axial Dispersion
Coefficient in a 1.0-in.~-ID Open Bubble Column Filled with Water

Daita set No. 34

 

 

Superficial Axial Dispersion Superficial Axial Dispersion

Run Gas Velocity Coefficient Run Gas Velocity Coefficient
No (em/sec) (cm?/sec) No. (cm/sec) (cm?/sec)

1 18.7 82.3 20 44.8 181.1

2 76.6 319.0 21 1.166 20.8

3 76.6 261.8 22 1.088 18.3

4 65.4 255.2 23 1.011 17.1

5 74.7 182.3 24 0.933 22.5

6 56.0 319.0 25 0.855 17.1

7 37.3 170.2 26 0.778 15.5

8 18.7 63.8 27 0.700 14.1

9 28.0 170.2 28 0.156 5.84
10 46.1 232.0 29 1.25 14.8
11 10.3 52.0 30 1.02 24.6
12 18.7 92.8 31 0.78 19.7
13 28.9 128.3 32 0.468 14.8
14 23.3 102.6 33 0.156 7.04
15 14.0 61.1 34 1.33 22.4
16 32,7 146.6 35 1.01 | 17.8
17 37.3 213.8 36 0.702 14.5
18 46.7 185.5 37 0.390 10.7
19 63.5 285.1 38 0.156 6.42

 

 

 

a
Data obtained during first study.

801
Table 37. Experimentally Determined Values for the Axial Dispersion Coefficient
in a 1.5-in.-ID Open Bubble Column

 

 

 

Gas Properties of Liquid Axial
Data Superficial Disperser Surface Dispersion
Set Run Gas Velocity Diameter? Density Viscosity Tension Coefficient
No. No. (cm/sec) (cm) (g/cm3) (cP) (dynes/cm) (cm?/sec) Ref.
1 1 16.4 0.635 0.997 0.894 73.0 138.5 b
1 2 8.58 0.635 0.997 0.894 73.0 85.6 b
1 3 12.3 0.635 0.997 0.894 73.0 114.8 b
1 4 9.89 0.635 0.997 0.894 73.0 93.4 b
1 5 1.72 0.635 0.997 0.894 73.0 31.6 b
1 6 35.4 0.635 0.997 0.894 73.0 482.2 b
1 7 0.69 0.635 0.997 0.894 73.0 20.8 b
1 3 35.4 0.635 0.997 0.894 73.0 380.5 b
1 9 2.99 0.635 0.997 0.894 73.0 39.0 b
1 10 5.11 0.635 0.997 0.894 73.0 ' 65.0 b
1 11 0.912 0.635 0.997 0.894 73.0 24,8 b
1 12 22.4 0.635 0.997 0.89% 73.0 177.0 b
1 13 10.4 0.635 0.997 0.894 73.0 149.0 b
1 14 16.8 0.635 0.997 0.894 73.0 136.0 b
1 15 13.6 0.635 0.997 0.894 73.0 95.2 b
13 1 1.045 0.635 0.997 0.89 72.0 25.8 c
13 2 0.212 0.635 0.997 0.89 72.0 12.9 c
13 3 0.451 0.635 0.997 0.89 72.0 18.9 c
13 4 0.719 0.635 0.997 0.89 72.0 24.4 c
13 5 1.33 0.635 0.997 0.89 72.0 32.9 c
13 6 1.68 0.635 0.997 0.89 72.0 31.6 c
13 7 2.21 0.635 0.997 0.89 72.0 39.5 c
13 8 3.86 0.635 0.997 0.89 72.0 50.0 ¢
13 9 6.54 0.635 0.997 0.89 72.0 69.4 c
13 10 8.93 0.635 0.997 0.89 72.0 85.9 c

601
Table 37 (continued)

 

 

 

Gas Properties of Liquid Axial

Data Superficial Disperser Surface Dispersion

Set Run Gas Velocity Diameter? Density Viscosity Tension Coefficient

No. No. (cm/sec) (cm) (g/cm3) (cP) (dynes/cm) (cmzlsec) Ref.
13 11 21.3 0.635 0.997 0.89 72.0 242 c
14 1 1.045 0.635 0.995 1.07 45.3 20.9 c
14 2 0.710 0.635 0.995 1.07 45.3 23.3 c
14 3 0.535 0.635 0.995 1.07 45.3 20.5 c
14 4 0.252 0.635 0.995 1.07 45.3 17.3 c
14 5 0.0930 0.635 0.995 1.07 45.3 14.6 c
14 6 1.456 0.635 0.995 1.07 45.3 23.1 c
14 7 1.55 0.635 0.995 1.07 45.3 26.7 c
14 8 2.58 0.635 0.995 1.07 45.3 33.7 c
14 9 4,31 0.635 0.995 1.07 45.3 46.6 c
14 10 5.79 0.635 0.995 1.07 45.3 50.0 c
14 11 7.45 0.635 0.995 1.07 45.3 58.0 c
15 1 1.02 0.635 0.996 1.09 27.3 21.9 c
15 2 0.133 0.635 0.996 1.09 27.3 15.0 c
15 3 0.0705 0.635 0.996 1.09 27.3 11.9 c
15 4 0.214 0.635 0.996 1.09 27.3 15.8 c
15 5 0.494 0.635 0.996 1.09 27.3 20.6 c
15 6 0.763 0.635 0.996 1.09 27.3 22,2 c
15 7 1.37 0.635 0.996 1.09 27.3 22.3 c
15 8 2.07 0.635 0.996 1.09 27.3 27.9 c
15 9 3.5 0.635 0.996 1.09 27.3 36.2 c
15 10 6.4 0.635 0.996 1.09 27.3 51.5 c
15 11 8.6 0.635 0.996 1.09 27.3 65.7 c
16 1 0.0638 0.635 1.04 2.05 71.5 7.65 c
16 2 0.187 0.635 1.04 2.05 71.5 12.5 c
16 3 0.333 0.635 1.04 2.05 71.5 24,6 c

11
 

 

 

Table 37 (continued)
Gas Properties of Liquid Axial

Data Superficial Disperser Surface Dispersion

Set Run Gas Velocity Diameter? Density Viscosity Tension Coefficient

No. No. (cm/sec) (cm) (g/cm3) (cP) (dynes/cm) (cm2/sec) Ref.
16 4 0.528 0.635 1.04 2,05 71.5 26.4 c
16 5 0.895 0.635 1.04 2.05 71.5 28.4 c
16 6 1.25 0.635 1.04 2.05 71.5 36.4 c
16 7 1.625 0.635 1.04 2.05 71.5 36.4 c
16 8 2.61 0.635 1.04 2.05 71.5 43.5 c
16 9 5.78 0.635 1.04 2.05 71.5 60.6 c
16 10 8.07 0.635 1.04 2.05 71.5 81.0 C
17 1 0.214 0.635 1.15 11.3 70.0 10.6 c
17 2 0.0888 0.635 1.15 11.3 70.0 7.45 c
17 3 0.313 0.635 1.15 11.3 70.0 12.3 c
17 4 0.45 0.635 1.15 11.3 70.0 15.2 c
17 5 0.651 0.635 1.15 11.3 70.0 21.3 c
17 6 1.0 0.635 1.15 11.3 70.0 21.8 c
17 7 1.38 0.635 1.15 11.3 70.0 27.3 C
17 8 1.74 0.635 1.15 11.3 70.0 27.7 c
17 9 2.9 0.635 1.15 11.3 70.0 34.7 c
17 10 5.79 0.635 1.15 11.3 70.0 55.6 c
22 1 10.4 0.635 1.05 1.8 71.8 88.9 b
22 2 1.66 0.635 1.05 1.8 71.8 24.9 b
22 3 32,7 0.635 1.05 1.8 71.8 814 b
22 4 49.7 0.635 1.05 1.8 71.8 656 b
22 5 - 32.7 0.635 1.05 1.8 71.8 155 b
22 6 4.06 0.635 1.05 1.8 71.8 44,6 b
22 7 10.4 0.635 1.05 1.8 71.8 88.9 b
22 8 21.9 0.635 1.05 1.8 71.8 115 b
22 9 13.1 0.635 1.05 1.8 71.8 131 b

TTT
Table 37 (continued)

 

 

 

Gas Properties of Liquid Axial

Data Superficial Disperser Surface Dispersion

Set Run Gas Velocity Diameterd Density Viscosity Tension Coefficient

No. No. (em/sec) (cm) (g/cm3) (cP) (dynes/cm) (em2/sec) Ref.
22 10 10.4 0.635 1.05 1.8 71.8 56.2 b
22 11 6.62 0.635 1.05 1.8 71.8 62.2 b
31 1 0.631 0.102 0.997 0.89 73.0 18.3 25
31 2 7.88 0.102 0.997 0.89 73.0 74.2 25
31 3 3.11 0.102 0.997 0.89 73.0 39.4 25
31 4 25.3 0.102 0.997 0.89 73.0 194 25
31 5 20.7 0.102 0.997 0.89 73.0 154 25
31 6 13.3 0.102 0.997 0.89 73.0 100.3 25
31 7 11.2 0.102 0.997 0.89 73.0 97.6 25

cll

 

%cas disperser consisted of a single orifice having the indicated inside diameter.
bData obtained during second study.

“Data obtained during third study.
Table 38.

Experimentally Determined Values for the Axial
Dispersion Coefficient in a 2.0-in.-ID Open Bubble Column

 

 

 

 

Superficial Gas Disperser Design Properties of Liquid Axial

Data Gas Number Inlet Surface Dispersion

Set Run Velocity of Diameter Density Viscosity Tension Coefficient

No. No. (cm/sec) Inlets (cm) (g/cm?) (cP) (dynes/cm) (cme/sec) Ref.
2 1 4.0 1 0.432 0.997 0.894 73.0 51.0 a
2 2 20.3% 1 0.432 0.997 0.894 7%.0 206.0 a
2 3 2%.8 1 0.432 0.997 0.894 73%.0 172.7 a
2 h 4.4 1 0.432 0.997 0.894 73.0 14h.9 a
2 5 0.56 1 0.432 0.997 0.894 73.0 22.3 a
2 6 3.88 1 0.432 0.997 0.894 73.0 58.7 a
2 7 5.96 1 0.432 0.997 0.894 73.0 67.5 a
L 1 0.0442 1 0.1 0.997 0.894 72.0 19.6 b
L 2 0.126 1 0.1 0.997 0.894 72.0 28.1 b
L 3 0.209 1 0.1 0.997 0.894 72.0 28.9 b
b L 0.295 1 0.1 0.997 0.894 72.0 31.1 b
L 5 0.405 1 0.1 0.997 0.894 72.0 37.6 b
b 6 0.515 1 0.1 0.997 0.894 72.0 40.5 b
L T 0.618 1 0.1 0.997 0.894 72.0 43,4 b
L 8 0.798 1 0.1 0.997 0.894 72.0 45.5 b
L 9 0.95 1 0.1 0.997 0.894 72.0 418.2 b
b 10 1.52 1 0.1 0.997 0.894 72.0 52.7 b
b 11 2.92 1 0.1 0.997 0.894 72.0 60.8 b
b 12 4.58 1 0.1 0.997 0.894 72.0 72.6 b
L 13 7.2 L 0.1 0.997 0.894 72.0 81.5 b
5 1 0.385 5 0.04 0.997 0.894 7%.0 40.6 b
5 2 0.045T 5 0.0k 0.997 0.894 7%.0 15.4 b
5 3 0.124 5 0.04 0.997 0.894 73.0 16.4 b
5 L 0.242 5 0.0h 0.997 0.894 73.0 22.8 b
5 5 0.537 5 0.0k 0.997 0.894 73.0 33,1 b

¢11
Table 38 (continued)

 

 

 

 

Superficial Gas Disperser Design Properties of Liquid Axial

Data Gas Number Inlet Surface Dispersion

Set Run Velocity of Diameter Density Viscosity Tension Coefficient

No. No. (cm/sec) Inlets (cm) (g/cm”) (cP) (dynes/cm ) (cme/sec) Ref.
5 6 0.749 5 0.0k 0.997 0.894 73.0 35.4 b
5 7 0.96 5 0.0k 0.997 0.894 73.0 36.4 b
5 8 1.7 5 0.0k 0.997 0.894 73.0 41.5 b
5 9 3.22 5 0.04 0.997 0.894 7%.0 57.8 b
5 10 L. 79 5 0.0k 0.997 0.894 73.0 60.8 b
5 11 7.00 5 0.0k 0.997 0.894 7%.0 72.6 b
5 12 8.46 5 0.04 0.997 0.894 7%.0 83%.6 b
6 1 0.688 5 0.1 0.997 0.894 7%.0 37.2 b
6 2 0.0767 5 0.1 0.997 0.894 73%.0 31.7 b
6 3 0.158 5 0.1 0.997 0.894 73.0 31.5 b
6 L 0.26 5 0.1 0.997 0.894 73.0 4.8 b
6 5 0.347 5 0.1 0.997 0.894 73.0 47.0 b
6 6 0.459 5 0.1 0.997 0.894 73.0 37.9 b
6 7 0.589 5 0.1 0.997 0.894 73.0 35,5 b
6 8 0.92 5 0.1 0.997 0.804 73.0 39.0 b
6 9 1.52 5 0.1 0.997 0.894 73.0 55.1 b
6 10 2.95 5 0.1 0.997 0.894 73.0 57.5 b
6 11 h.o7 5 0.1 0.997 0.894 7%.0 63.5 b
6 12 5.59 5 0.1 0.997 0.894 73.0 92.4 b
6 13 7.29 5 0.1 0.997 0.894 73.0 111.1 b
6 14 13.8 5 0.1 0.997 0.89%4 73.0 137.0 b
T 1 h.35 5 0.2 0.997 0.894 75.0 71.6 b
T 2 0.265 5 0.2 0.997 0.894 73.0 35,8 b
7 3 0.80 5 0.2 0.997 0.894 73.0 23,8 b
7 b 2.5 5 0.2 0.997 0.894 73.0 5Tk b

W1l
Table 38 (continued)

 

 

 

 

Superficial Gas Disperser Design Properties of Liquid Axial

Data Gas Number Inlet Surface Dispersion

Set  Run Velocity of Diameter Density Viscosity Tension Coefficient

No No. (cm/sec) Inlets (cm) (g/cm ) (cP) (dynes/cm ) (cm®/sec ) Ref.
T 5 9.15 5 0.2 0.997 0.894 73.0 139.0 b
T 6 21.6 5 0.2 0.997 0.894 73.0 200.0 b
7 T 0.11 5 0.2 0.997 0.894 7%.0 29.8 b
T 8 0.198 5 0.2 0.997 0.894 73.0 30.5 b
T 9 0.235 > 0.2 0.997 0.894 73.0 35.8 b
T 10 0.50% 5 0.2 0.997 0.894 73.0 30.0 b
7 11 1.0 5 0.2 C.997 0.894 73.0 Lh.o b
8 1 0.253 5 0.4 0.997 0.894 73.0 27.1 b
8 2 0.509 5 O.L 0.997 0.89k 73.0 41.2 b
8 3 1.54 5 0.4 0.997 0.894 73.0 41.8 b
8 L h.17 5 0.4 0.997 0.89L4 73.0 65.5 b
8 5 0.0938 5 0.k 0.997 0.894 73.0 30.8 b
8 6 0.512 5 0.k 0.997 0.894 7%.0 38.3 b
8 7 2.26 5 0.4 0.997 0.894 73.0 54.0 b
8 8 5.5 5 0.4 0.997 0.894 73.0 69.5 b
8 9 22.2 5 0.4 0.997 0.894 73.0 326 b
8 10 13.9 5 0.4 0.997 0.894 73.0 129 b
8 11 0.97 5 0.4 0.997 0.894 73.0 33,7 b
8 12 0.121 5 O.k 0.997 0.894 73.0 26.4 b
8 13 1.3 5 O.L 0.997 0.894 3.0 41.8 b
8 14 0.72 5 0.4 0.997 0.894 73.0 36.8 b
8 15 0.373 5 0.L 0.997 0.894 73.0 32,5 b
8 16 0.076 5 0.4 0.997 0.89L 73.0 25.5 b
T 12 0.0512 5 0.2 0.997 0.894 73.0 20.8 b
T 13 0.125 5 0.2 0.997 0.804 73.0 41.0 b
T 14 0.202 5 0.2 0.997 0.894 7%.0 L45.6 b

S11
Table 38 (continued)

 

 

 

 

Superficial Gas Disperser Design Properties of Liquid Axial
Data Gas Number Inlet Surface Dispersion
Set Run Velocity of Diameter Density Viscosity Tension Coefficient
No. No. (cm/sec ) Inlets (cm) (g/cmB) (cP) (dynes/cm) (cme/sec) Ref.
T 15 0.30kL 5 0.2 0.997 0.894 73.0 37.0 b
7 16 0.405 5 0.2 0.997 0.894 73.0 36.5 b
T 17 0.525 5 0.2 0.997 0.894 3.0 ho.2 b
7 18 0.65 5 0.2 0.997 0.894 73.0 56.0 b
T 19 0.76 5 0.2 0.997 0.89% 73.0 €2.7 b
7 20 0.923 5 0.2 0.997 0.894 73.0 52.2 b
7 21 1.525 5 0.2 0.997 0.894 73.0 46.8 b
T 22 3,15 5 0.2 0.997 0.894 73.0 59.5 b
T 23 L. 5 0.2 0.997 0.894 73.0 69.5 b
T 24 5.81 5 0.2 0.997 0.894 73.0 75.2 b
7 25 T.13 5 0.2 0.997 0.894 73.0 96.5 b
7 26 16.3 5 0.2 0.997 0.894 73.0 87.2 b
9 1 0.0505 19 0.1 0.997 0.89%4 73.0 26.4 b
9 2 0.114 19 0.1 0.997 0.894 73.0 36.7 b
9 3 0.208 19 0.1 0.997 0.894 73.0 38.5 b
9 L 0.301 19 0.1 0.997 0.894 73.0 39.7 b
9 5 0.404 19 0.1 0.997 0.894 73.0 40.1 b
9 6 0.535 19 0.1 0.997 0.894 T73.0 4hh.5 b
9 T 0.632 19 0.1 0.997 0.894 73.0 46.5 b
9 8 0.726 19 0.1 0.997 0.894 73.0 L7.0 b
9 9 0.919 19 0.1 0.997 0.894 73.0 4.9 b
9 10 1.48 19 0.1 0.997 0.89L4 73.0 42.8 b
9 11 3,02 19 0.1 0.997 0.894 73.0 s5h.L b
9 12 4,24 19 0.1 0.997 0.894 73.0 62.5 b
9 13 5.71 19 0.1 0.997 0.894 3.0 80.4 b
9 14 T.38 19 0.1 0.997 0.894 73.0 124 b
9 15 13.8 19 0.1 0.997 0.894 73.0 155 b

o1t
Table 38 (continued)

 

 

 

 

Superficial Gas Dispersor Design Properties of Liquid Axial

Data Gas Number Inlet Surface Dispersion

Set Run Velocity of Diameter Density Viscosity Tension Coefficient

No. No. (cm/sec) Inlets (cm) (g/cmj) (cP) (dynes/cm) (cmg/sec) Ref.
10 1 0.0525 37 0.1 0.997 0.89k 73.0 27.3 b
10 2 0.119 37 0.1 0.997 0.894 73.0 33.3 b
10 3 0.171 37 0.1 0.997 0.89k 73.0 Lo. L b
10 L 0.289 37 0.1 0.997 0.894 73.0 37.8 b
10 5 0.393 37 0.1 0.997 0.894 73.0 33,7 b
10 6 0.529 37 0.1 0.997 0.894 73.0 3h.2 b
10 T 0.687 37 0.1 0.997 0.894 73.0 45.0 b
10 8 0.78 37 0.1 0.997 0.894 73.0 48.0 b
10 9 0.945 37 0.1 0.997 0.894 73.0 48.0 b
10 10 1.52 37 0.1 0.997 0.894 73%.0 L. L b
10 11 2.97 37 0.1 0.997 0.894 73.0 58.2 b
10 12 4.36 37 0.1 0.997 0. 894 7%.0 121 b
10 13 5.6 37 0.1 0.997 0.894 73%.0 95.0 b
10 14 T.35 37 0.1 0.997 0.894 73.0 101 b
10 15 12.8 37 0.1 0.997 0.894 73.0 136 b
10 16 24 .6 37 0.1 0.997 0.894 73.0 214 b
10 17 36.8 37 0.1 0.997 0.894 73.0 228 b
11 1 0.112 37 0.2 0.997 0.894 73.0 36.7 b
11 2 0.376 37 0.2 0.997 0.894 73.0 34,7 b
11 3 0.67 37 0.2 0.997 0.894 73.0 39.3 b
11 L 0.0703% 37 0.2 0.997 0.894 73.0 20.5 b
11 5 0.262 37 0.2 0.997 0.89L 73%.0 35.5 b
11 6 0.478 37 0.2 0.997 0.894 73.0 33,2 b
11 T 0.875 37 0.2 0.997 0. 89k 75-0 38.0 b
11 8 1.37 37 0.2 0.997 0.894 7%.0 hg.5 b
11 9 2.21 37 0.2 0.997 0.894 73.0 61.0 b
11 10 5.0 37 0.2 0.997 0.894 7%.0 91.8 b

L11
Table 38 (continued)

 

 

 

 

. 20

Superficial Gas Disperser Design Properties of Liquid Axial

Data Gas Number Inlet Surface Dispersion
Set Run Velocity of Diameter Density Viscosity Tension Coefficient
No. No. (cm/sec ) Inlets (em) (g/cm’) (cP) (dynes/cm) (cm?/sec) Ref.
11 11 3.58 37 0.2 0.997 0.89k 73.0 80.0 b
11 12 0.182 3T 0.2 0.997 0.894 73.0 33,9 b
11 13 7.28 37 0.2 0.997 0.894 73.0 100 b
11 14 29.1 37 0.2 0.997 0. 89k 73.0 269 b
12 1 0.737 37 0.4 0.997 0.894 73.0 33,4 b
12 2 0.0618 37 0.4 0.997 0.894 73.0 28.2 b
12 3 0.148 37 0.4 0.997 0.89L 73.0 33,1 b
12 L 0.251 37 0.k 0.997 0.894 73.0 31.9 b
12 5 0.331 37 0.4 0.997 0.894 73.0 30.0 b
12 6 0.403 37 0.4 0.997 0.894 73.0 33,7 b
12 T 0.585 37 0.4 0.997 0.894 73.0 32.4 b
12 8 1.01 37 0.4 0.997 0.89L 73.0 51.3 b
12 9 1.49 37 0.4 0.997 0.894 73.0 50.0 b
12 10 2.66 37 0.4 0.997 0.894 73.0 59.7 b
12 11 3.62 37 0.4 0.997 0.894 73.0 61.0 b
12 12 L. 66 37 0.4 0.997 0.894 73.0 72.2 b
12 13 5.8 37 0.4 0.997 0.89k 73%.0 79.5 b
12 14 T.43 37 0.4 0.997 0.894 73.0 77.5 b
12 15 26.5 37 0.4 0.997 0.894 73.0 164 b
20 1 5.78 1 0.432 1.04T 1.8 71.8 T72.2 a

2 8.75 1 0.432 1.047 1.8 71.8 97.8 a
20 3 10.4 1 0.432 1.04T 1.8 71.8 120 a
20 4 18.6 1 0.432 1.047 1.8 71.8 178 a
20 5 27.3% 1 0.432 1.0LT 1.8 71.8 Lot a
20 6 2.79 1 0.432 1.0LT 1.8 71.8 45.3 a
20 7 4.01 1 0.432 1.047T 1.8 71.8 52.1 a

811
Table 38 (continued)

 

 

 

 

Superficial Gas Disperser Design Properties of Liquid Axial

Data Gas Number Inlet Surface Dispersion

Set Run Velocity of Diameter Density Viscosity Tension Coefficient

No. No. (cm/sec) Inlets (cm) (g/cm) (cP) (dynes/cm) (cm?/sec) Ref.
20 8 0.53 1 0.432 1.047 1.8 71.8 26.5 a
20 9 0.41 1 0.432 1.047 1.8 71.8 26.5 a
20 10 0.75 1 0.432 1.047 1.8 71.8 31.2 a
20 11 1.21 1 0.432 1.0LT 1.8 71.8 3L.5 a
20 12 2.21 1 0.432 1.04T 1.8 71.8 L8.2 a
21 1 0.91 1 0.432 1.163 12.1 679 27.6 a
21 2 0.49 1 0.432 1.163 12.1 67.9 21 .4 a
21 3 0.31 1 0.432 1.163 12.1 67-9 18.9 a
21 L 1.24 1 0.432 1.163 12.1 67-9 27.8 a
21 5 2.35 1 0.432 1.163 12.1 67.9 3h.1 a
21 6 3.0 1 0.432 1.163 12.1 67-9 37.2 a
21 7 h.11 1 0.432 1.163 12.1 67.9 i L a
21 8 5.77 1 0.432 1.163 12.1 67.9 51.0 a
21 9 9.1 L 0.43%2 1.163 12.1 67-9 75.3 a
21 10 10.3 1 0.432 1.163 12.1 67-9 79.2 a
21 11 16.1 1 0.432 1.163 12.1 67.9 105 a
21 12 2L.9 1 0.432 1.163 12.1 67-9 144 a
21 13 37.h 1 0.432 1.163 12.1 67.9 21k a
30 1 0.405 1 0.102 0.997 0.894 73.0 29.6 22
30 2 0.493 1 0.102 0.997 0.894 73.0 27.% 22
30 3 0.7h4 1 0.102 0.997 0.894 73.0 29 22
30 L 1.50 1 0.102 0.997 0.89h 73.0 33,3 22
30 5 2.55 1 0.102 0.997 0.894 7%.0 39.1 22
30 6 3.6L 1 0.102 0.997 0.894 73.0 52.0 22
30 7 0.256 1 0.102 0.997 0.894 73.0 26.9 22
30 8 0. 404 1 0.102 0.997 0.894 7%.0 28.4 22
30 9 0.405 1 0.102 0.997 0.894 73.0 28.6 22

611
Table 38 (continued)

 

 

 

 

Superficial Gas Disperser Design Properties of Liquid Axial

Data Gas Number Inlet Surface Dispersion

Set Run Velocity of Diameter Density Viscosity Tension Coefficient

No. No. (cm/sec) Inlets (cm) (g/cm”) (cP) (dynes/cm) (cm?/sec) Ref.
30 10 O. LT 1 0.102 0.997 0.89k 73.0 26.4 22
30 11 0.582 1 0.102 0.997 0.894 73.0 31.4 22
30 12 0.671 1 0.102 0.997 0.894 73.0 28.9 22
30 13 0.784 1 0.102 0.997 0.894 73.0 29.6 o2
30 14 0.918 1 0.102 0.997 0.894 73.0 28.9 22
30 15 1.159 1 0.102 0.997 0.894 73.0 32.1 22
30 16 1.50 1 0.102 0.997 0.89k 73.0 32.6 22
30 17 1.55 1 0.102 0.997 0.894 73.0 35.6 22
30 18 2.26 1 0.102 0.997 0.894 73.0 35 22
30 19 2.91 1 0.102 0.997 C.894 73.0 43,6 22
30 20 3.69 1 0.102 0.997 0.894 73.0 L8 22
30 21 5.28 1 0.102 0.997 0.894 73.0 68.7 22
27 1 0.432 1 0.102 0.997 0.894 73.0 28.5 23
o7 2 3.31 1 0.102 0.997 0.894 T3.0 Lo.T 23
27 3 3.3] 1 0.102 0.997 0.894 73.0 61.7 23
27 L 18.7 1 0.102 0.997 0.894 73%.0 6T.7 23
27 5 10.9 1 0.102 0.997 0.894 73.0 76.8 23
28 1 0.432 1 0.102 1.15 15 65.2 19.1 23
28 2 3.3] 1 0.102 1.15 15 65.2 Lg.7 23
28 3 2.28 1 0.102 1.15 15 65.2 ho.T 23
28 Y 1.77 1 0.102 1.15 15 65.2 28.2 23
28 5 1.25 1 0.102 1.15 15 65.2 23.6 2%
25 1 0.432 1 0.102 0.997 0.894 37.6 38.8 23
25 2 1.25 1 0.102 0.997 0.894 37.6 Ly, 6 23
25 3 2.28 1 0.102 0.997 0.894 37.6 45.6 23
25 L 3.31 1 0.102 0.997 0.89k 37.6 57.1 23

0cl
Table 38 (continued)

 

 

 

 

Superficial Gas Disperser Design Properties of Liquid Axial

Data Gas Number Inlet el Surface Dispersion

Set Run Velocity of Diameter Density Viscosity Tension Coefficient

No. No. (cm/sec) Inlets (cm) (g/cm”) (cP) (dynes/cm) (cm?/sec) Ref.
25 5 0.201 1 0.102 0.997 0.894 37.6 37.9 23
25 6 . L8 1 0.102 0.997 0.894 37.6 89.1 23
26 1 0.201 1 0.102 0.997 0.894 5%.2 37.9 23
26 2 3.31 1 0.102 0.997 0.890L 53,2 55.0 23
26 3 0.432 1 0.102 0.997 0.894 53.2 ho.2 23
26 L L. 48 1 0.102 0.997 0.89k 5%.2 9.k 23
29 1 1.25 1 0.102 0.997 0.89k 647 39.0 23
30 1 10.2 1 0.432 0.997 0.894 73.0 113 25
32 2 15.0 1 0.432 0.997 0.894 73.0 148 25
32 3 18.5 1 0.432 0.997 0.894 73.0 139 25
32 L 23.3 1 0.432 0.997 0.894 73.0 176 25
32 5 21,0 1 0.432 0.997 0.894 7%.0 145 25
32 6 2k L 1 0.432 0.997 0.894 7%.0 282 25
32 T 25.4 1 0.432 0.997 0.894 3.0 192 25
32 8 37,7 1 0.432 0.997 0.894 73.0 265 25
32 9 29.8 1 0.432 0.997 0.894 73.0 258 25
32 10 0.384 1 0.432 0.997 0.894 7%.0 21.4 25
32 11 2.83 1 0.432 0.997 0.894 7%.0 46.0 25
22 12 1.59 1 0.432 0.997 0.894 73.0 33,1 25
22 13 0-380 1 0.432 0.997 0.894 7%.0 31.0 25

 

%Data obtained during second study.

bData obtained during third study.

el
Tabte 39. lxperimentally Determined Values for the Axia! Dispersion Coefficient in a 3.0-in.-ID Open Bubble Column

Gas disperser

: one orifice, 0.432 cm ID

Properties of Liquid

 

 

 

Axial

Data Superficial Surface Dispersion

Set Run Gas Velocity Density Viscosity Tension Coefficient
No. No. (em/sec) (g/cm3) (cP) (dynes/cm) {(cm</sec) Reference
18 1 0.81 1.04 2.05 71.5 50.0 a
18 2 0.0352 1.04 2.05 71.5 28.1 a
18 3 0.0768 1.04 2.05 71.5 43.2 a
18 4 0.122 1.04 2.05 71.5 38.2 a
18 5 0.196 1.04 2.05 71.5 48.9 a
18 6 0.285 1.04 2.05 71.5 51.8 a
18 7 0.425 1.04 2.05 71.5 45.0 a
18 8 0.553 1.04 2.05 71.5 48.7 a
18 9 1.025 1.04 2.05 71.5 46.2 a
18 10 1.86 1.04 2.05 71.5 64.0 a
18 11 2.85 1.04 2.05 71.5 66.4 a
18 12 5.66 1.04 2.05 71.5 89.1 a
18 13 8.26 1.04 2.05 71.5 92.2 a
18 14 0.0323 1.04 2.05 71.5 30.2 a
19 1 0.0318 1.15 10.7 70.0 27.0 a
19 2 0.n0769 1.15 10.7 70.0 29.1 a
19 3 0.108 1.15 10.7 70.0 42.4 a
19 4 0.155 1.15 10.7 70.0 56.0 a
19 5 0.232 1.15 10.7 70.0 4l.4 a
19 6 0.338 1.15 10.7 70.0 47.4 a
19 7 0.424 1.15 10.7 70.0 45.0 a
19 8 0.675 1.15 10.7 70.0 52.8 a
19 9 1.67 1.15 10.7 70.0 47 .4 a
19 10 1.25 1.15 106.7 70.0 56.3 a
19 11 2.28 1.15 10.7 70.0 57.2 a
19 12 2.96 1.15 10.7 70.0 64.3 a
19 13 5.37 1.15 10.7 70.0 86.7 a
33 1 10.7 0.997 0.894 73.0 124 25
33 2 18.3 0.997 0.894 73.0 164 25
33 3 19.3 0.997 0.894 73.0 207 25
33 4 3.70 0.997 0.894 73.0 714 25
33 5 7.32 0.997 0.894 73.0 94.2 25
33 & 0.700 0.997 0.894 73.0 43.0 25
33 7 4.78 0.997 ‘0,894 73.0 78.9 25
33 8 11.6 0.997 0.894 73.0 132 25
33 g 11.1 0.997 0.89¢4 73.0 119 25
33 10 26.5 0.997 0.894 73.0 239 25

 

“Data obtained during third study.

ccl
Table 40. Experimentally Determined values®
in a 6.0-in.-ID Open Bubble Column

Gas disperser:

one orifice, 0.432 cm ID

1

for the Axial Dispersion Coefficient

 

 

 

Properties of Liquid Axial
Data Superficial Surface Dispersion
Set Run Gas Velocity Density Viscosity Tension Coefficient
No. No {(cm/sec) (g/cm3) (cP) (dynes/cm) (cm?/sec)
3 1 8.39 0.997 0.894 73.0 207
3 2 3.08 0.997 0.894 73.0 179
3 3 0.8 0.997 0.894 73.0 152
3 4 0.52 0.997 0.89%4 73.0 157
3 5 0.52 0.997 0.894 73.0 161
3 6 1.49 0.997 0.894 73.0 183
3 7 2.16 0.997 0.894 73.0 161
3 8 7.12 0.997 0.894 73.0 371
3 9 0.62 0.997 0.894 73.0 173
3 10 3.83 0.997 0.894 73.0 157
3 11 1.07 0.997 0.894 73.0 123
3 12 0.33 0.997 0.894 73.0 146
3 13 1.57 0.997 0.89%4 73.0 146
3 14 9.13 0.997 0.89%4 73.0 281
3 15 4.93 0.997 0.894 73.0 246
3 16 2.63 0.997 0.894 73.0 141
23 1 7.52 1.047 1.8 71.8 232
23 2 9.63 1.047 1.8 71.8 266
23 3 9.63 1.047 1.8 71.8 303
23 4 2.6 1.047 1.8 71.8 169.5
23 5 5.06 1.047 1.8 71.8 262
23 6 8.62 1.047 1.8 71.8 229
23 7 0.32 1.047 1.8 71.8 125
23 8 2.04 1.047 1.8 71.8 188
23 9 1.19 1.047 1.8 71.8 180
23 10 2.98 1.047 1.8 71.8 184
23 11 3.88 1.047 1.8 71.8 211
23 12 2.98 1.047 1.8 71.8 181
23 13 6.0 1.047 1.8 71.8 216
24 1 1.21 1.16 12.1 67.9 153
24 2 1.62 1.16 12.1 67.9 149
24 3 0.27 1.16 12.1 67.9 162
24 4 3.99 1.16 12.1 67.9 181
24 5 2.86 1.16 12.1 67.9 180
24 6 1.98 1.16 12.1 67.9 146
24 7 6.01 1.16 12.1 67.9 216
24 8 8.4 1.16 12.1 67.9 254
24 g 0.51 1.16 12.1 67.9 157
24 10 0.75 1.16 12.1 67.9 173

 

“pata obtained during second study.

 
where

Pe

Re

Su

Ar

n

h

124

b c d e f
N, = KN ° Ng ° N, (B/dc) n. (1-h)" , (35)

Peclet number = ££& s

I
0o
l'O
=

Reynolds number

Suratman number

1l
no
.

Archimedes number = ——— ,

column diameter, cm,
superficial gas velocity, cm/sec,
. * ’ » & 2
axial dispersion coefficient, cm /sec,
. . 3
density of liquid, g/cm ,
viscosity of liquid, g/cm-sec,
surface tension of liquid, dynes/cm,
gas distributor diameter, cm,
number of openings in gas distributor,

fraction of column volume occupied by gas, and

a, b, ¢, d, e, £, and K are constants to be determined by the method of

least squares.

The quality of the data from the l-in.-diam column was considered to

be lower than that for the larger-diameter columns; hence only data

obtained with column diameters of 1.5, 2, 3, and 6 in. were used in deter-

mining the final correlation. Because of the observed difference in

behavior in the bubble- and slug-flow regions, the data were divided into

two sets corresponding to the two flow regimes; each data set was then
125

correlated separately using the general power-law expression given by

Eq. (35). The criterion for separating the data was that the manner of
division should result in the best possible fit of the data to the two
resulting power-law expressions. In determining the final expressioms,
all terms shown in Eq. (35) were considered; however, only those terms
were retained for which the uncertainty in the value of the exponent

was less than 30% of the value of the exponent, or which, when neglected,
resulted in a poorer fit of the data as indicated by the F factor. The

F factor is defined as:

T\ 2
2:(1n NPe In NPe)

 

 

F = (36)
c 2
Z (1n N, — In NPe)
where
In NPe = natural logarithm of an experimentally determined value
for the Peclet number,
1n NPe = average of all values for the natural logarithm of the
experimentally determined values of the Peclet number,
In N;e = predicted value for the natural logarithm of the Peclet

number,

and the indicated summations are over all data points considered. The
values for the F factor for Egs. (37) and (38) were 1200 and 125
respectively. Both of these values imply that the agreement between
the experimentally determined values and the predicted values is very
good. The final correlations resulting from the analysis are as

follows:

for bubble flow,

_ 0.82 -0.4L4  -0.09% |
Ny, = 8.30 Npo N, n : (37)
for slug flow,
Pe Re Su ’
126

The expression for bubble flow is based on 190 data points; a comparison
of the experimental and calculated values is shown in Fig. 41. The
expression for slug flow is based on 122 data points; a comparison of
the experimental and calculated values is shown in Fig. L4L2. Seven data
points in the transition region between the bubble and slug flow regions
were excluded from consideration in order to facilitate fitting of the

data.

Comparisons of the predicted and experimentally determined values
for the axial dispersion coefficient are shown in Figs. 43-54 for
columns having diameters of 1, 1.5, 2, 3, and 6 in. 1In general, good
agreement is observed. The predicted values for the 1-in.-diam column
are about twice the experimentally determined values (which were not
used in developing the correlation). and show the greatest divergence
from the experimentally determined values. The points at which the
transition from bubble flow to slug flow occurs are indicated in the
figures. These points are in good agreement with visual observations

on the changes in flow mode.

9.8 Discussion of Data on Axial Mixing and Gas Holdup

The following observations can be made with regard to the general
flow behavior in open bubble columns and to the effects on axial dis-
persion and gas holdup caused by variations in column diameter, viscosity
and surface tension of the liquid, gas inlet orifice diameter, and

number of orifices in the gas inlet.

9.8.1 Flow Regimes and Effect of Gas Superficial Velocity

Three different flow regimes can be identified by visual observation
of the mixing patterns in open bubble columns and by examination of the
associated data on axial mixing and gas holdup. These regimes — bubble
flow, transition flow, and slug flow — are observed successively as the

superficial gas velocity is increased from low to high values.

For superficial gas velocities below about 1 cm/sec, the bubbles
rise as discrete entities and do not occupy the entire cross section of

the colurn. The gas holdup increases linearly with increases in the
COMMON LOGARITHM OF MEASURED PECLET NUMBER

127

 

 

 

 

ORNL DWG 72-13503Rl

-0.4

-0.6f : 7]
-0.8 o ° _

s LY
. o * o
-1.0F * O -
at A8
o". . *
Predicted Values 0. * e o
\ . .' _
...|2"" . ‘ . . * —
o o, * ..
-1.4+ . . .‘ . —
-1.6 o * * —
. . . .
—l 8'_ . ... - —
20F -
-2.2 ! ] | i i I ] |
2.2 -2.0 -1.8 -1.6 -1.4 -1.2 -1.0 -08 -06 -0.4
COMMON LOGARITHM OF PREDICTED PECLET NUMBER
Fig. 41. Comparison of Predicted and Experimentally -Determined

Values for the Peclet Number During Bubble Flow in Columns Having

Diameters Ranging from 1.5 to 6 in.
128

ORNL DWG 72-1350I
0.0

 

-0l —

   

      
 

Predicted Values

COMMOM LOGARITHM OF MEASURED PECLET NUMBER
o
-fl
I

 

 

 

| | ] L
-.4 -1.2 -1.0 -08 -06 -0.4 -0.2 0.0
COMMON LOGARITHM OF PREDICTED PECLET NUMBER ~

 

Fig. L42. Comparison of Predicted and Experimentally-Determined
Values for the Peclet Number During Slug Flow in Columns Having
Diameters Ranging from 1.5 to 6 in.
AXIAL DISPERSION COEFFICIENT (cmzlsec)

ORNL DWG 73-29

 

500 T T T —F T T T 171 T

300

100~

50

  

 

 

 

5 | L 4 L4 i ] 11 1 1 111 1 11§ 11
0.1 0.2 0.3 05 0.7 | 2 3 5 7 10 20 30 50 70

SUPERFICIAL GAS VELOCITY (cm/sec)

Fig. 43. Variation of Axial Dispersion Coefficient with Changes in
Superficial Gas Velocity in a 1.0-in.-ID Bubble Column Filled with
Water. The transition point between the bubble- and slug-flow regimes
is denoted by a vertical arrow.

100

621
DISPERSION COEFFICIENT (cm¥sec)

AXIAL

ORNL DWG. 73-49

 

 

 

 

0%1

1000 T T T TTTTT T T T TTTTI T T T T,

80U VISCOSITY A
= {cP) A -
600K ® 0.9 —
- a 1.8 —
400} 0 1.3 _

o

200} . .

o ' a

A
|00} * 4 —
— 7L -
gol- 275N -
- / 2 pts, —
o 2
60} "’éfi A -
A ,/A
40 = 0.9 cP . —
H A A
- A ,;fi—;‘fi?ifif' H=18cP -
4 = B p=IL3 cP
20 . —
o
A e @
10 10 ] L1 1111 ] 1 L 11111 L 1 1 1 1141
0.1 02 04 06 08 10 20 40 60 80 |0 20 40 60 80 I00
SUPERFICIAL GAS VELOCITY (cm/sec) .
Fig. Uh. Variation of Axial Dispersion Coefficient with Changes in

Superficial Gas Velocity and Viscosity of Liquid in a 1.5-in.-ID Bubble
Column. The transition between the bubble- and slug-flow regimes is
denoted by vertical arrows for the three viscosity values considered.
COEFFICIENT {(cm?Z/sec)

DISPERSION

AXTAL

ORNL DWG. 73-54

1000 T T

800F SURFACE TENSION
— (dynes/cm)

600+ s T2
- A 453

400 o 273

 

[ 111

200} o = 72 dynes/cm —

100
80

T 11711
1111

|

60 A

T
B
1

40} e o

]

20} J0A ]

 

 

10 ] ] ] 1 ] i i ]
05 0.1 0.2 0.5 l.O 2.0 50 10 20 50

 

SUPERFICIAL GAS  VELOCITY (cm /sec)

Fig. 45. Variation of Axial Dispersion Coefficient with Changes in
Superficial Gas Velocity and Surface Tension of Liquid in a 1.5-in.-ID
Bubble Column Having a Single 0.432-cm-ID Gas Inlet.

1¢1
(cmZ/sec)

AXIAL DISPERSION COEFFICIENT

1000,

ORNL DWG. 73-4|

 

900
800
700~
600k~

400~

300+

2001

100
90
80
70

60
SO

FTTI

[

40+

30

 

VISCOSITY (cP)
® 0.9

A
O

1.8
12.1

I I T T T1TTTI I T 1T T 1T U

Lt 111l

-l
O
O
1

 

] 1 | 11 1 1 (11 ] 1 1 111

 

0.

Column.

0.2 03 04 c6 08

Fig. L6.

Variation

3

1
1.0 2.0 30 40 60 80 10 20 30 60 80 100
SUPERFICIAL GAS VELOCITY (cm/sec)

of Axial Dispersion Coefficient with Changes in

Superficial Gas Velocity and Viscosity of Liquid in a 2.0-in.-ID Bubble

The transition points between bubble- and slug-flow regimes are
denoted by vertical arrows.

¢l
(cm2/sec )

COEFFICIENT

AXIAL DISPERSION

ORNL DWG. 73-53

1000 T I [ I I {

 

[

VISCOSITY (cP)

fi

Q

O
[

« 089

A 150

3
O
|

W
o
O

3

o
Q

20

 

0 L 1 L 1 1 1 9 1 fF 4 i

 

 

 

0.04 0.07 0.1 02 0.3 0.5 07 1O 20 30 50 70 10 20

SUPERFICIAL GAS VELOCITY (cm/sec)

Fig. L7. Variation of Axial Dispersion Coefficient with Changes in
Superficial Gas Velocity and Viscosity of Liquid in a 2.0-in.-ID Bubble
Column. The transition points between bubble- and slug-flow regimes are
denoted by vertical arrows.

30

40

et
DISPERSION COEFFICIENT (c¢m?2/sec)

AXIAL

ORNL DWG. 73-52

 

 

    
 

 

 

1000~ j T — — T l | — T T T
800} SURFACE TENSION /////
| (dynes/cm)
600F . 37.6
— o 53.2
400t A 73.0
200
o= 376 dynes/cm
|00 .
801 -
e . —
60 —
— /2 pts. —]
y
40} cJ E] 3 pts. hn
[ o & o & .
O e 0% o
20 ]
10 | ] 1 | 1 ] 1 ] L1 | 1 | ]
Ol 02 04 06 08 10 2.0 40 60 80 10 20 40 60 80

SUPERFICIAL GAS VELOCITY (cm/sec)

Fig. 4L8. Variation of Axial Dispersion Coefficient with Changes in
Superficial Gas Velocity and Surface Tension of Liquid in a 2-in.-ID Bubble
Column.

100

7el
DISPERSION COEFFICIENT (cm%sec)

AXIAL

1000

ORNL DWG 73-31

 

700

8
7

300

200

100

70

50

 

T T T T T T T T T T T T T T
ORIFICE DIAMETER (cm)

. 0.432

A 0.1

 

i 1 1 ] 1 1 1 1 1 i i 1 i i

 

 

10
0.04

Q.07 Q. 0.2 03 05 07 10 20 3.0
SUPERFICIAL GAS VELOCITY (cm/sec)

Fig. 49. Vvariation of Axial Dispersion Coefficient with Changes in

Superficial Gas Velocity and Gas Inlet Diameter in a 2.0-in.-ID Bubbie
Column. The transition point between the bubble- and slug-flow regimes
is denoted by a vertical arrow.

50 7.0 0 20 30

40

T
AXIAL DISPERSION COEFFICIENT (cm2/sec)

1000

ORNL DWG 73-40C

 

900}
800—

TO00—
600

SO0

400

200

100
90
80
70

60
50

40

30

20

 

1o

l T T [ | rol !

ORIFICE DIAMETER
(mm)

©

a
o
©

c.4
1.0
2.0
4.0

1

 

! I Il 1 f | 1l 1 |

L1

[ |1

 

 

004

2 in.

0.08 0l

Fig. 50.
Superficial Gas Velocity and Gas Inlet Diameter for a Column Diameter of
The gas distributor consisted of five orifices of the indicated

diameter.

0.2 04 08 Lo 2.0 4.0 8.0 10 20

SUPERFICIAL  GAS VELOCITY (cm/sec)

Variation of Axial Dispersion Coefficient with Changes in

The transition point between the bubble- and slug-flow regimes
is denoted by a vertical arrow.

40

9¢1
AXIAL DISPERSION COEFFICIENT (cm@sec)

1000

ORNL DWG 73-30

 

 

 

 

 

— T T T T T 1717 T T T T T 1711 | T T T T 7T171]
[ ORIFICE DIAMZTER (mm) ]
[ o 1.0 -
500 A 2.0

= o 4.0 .
300 .
too -
—
50 -
30 .

0 I |1 4 41111 I L0 111l ! L1 1 )]
0.05 oN 0.2 0.3 0.405 1.0 2.0 30 4050 10 20 30 40 50

SUPERFICIAL GAS VELOCITY (cm/sec)
Fig. 51. Variation of Axial Dispersion Coefficient with Changes in

Superficial Gas Velocity and Gas Inlet Orifice Diameter for a Column

Diameter of 2 in.

cated diameter.

The gas inlet consisted of 37 orifices of the indi-

LT
DISPERSION COEFFICIENT {(cm?2/sec)

AXIAL

OQRNL DWS 73~ 50R|

 

 

 

 

1000 T 7 I T S — I T T 1 =
| NUMBER OF —]
8O0 oRIFICES ]
600 . I ]
L A 5 _
400 Y 9
& 37
O
200}- ¢ L
a
‘A
© 5
|00 37 Orifices R © n
80 19 Orifices e B . -
B 5 . _ -
60— / n?‘n ] o
o
i . A ~
B N A TOesC o L . _
40 o - °' - A A& 2 ® ° *
A 0 ' ® 0 5 ® . * —_
5 " ® . lo.. »®
20t O 5 Jrifices T _
| Orifice
0 L L I 1 L1 | 1 L1 |
.04 0.1 0.2 04 1.0 20 40 1O 20 400

SUPERFICIAL GAS VELOCITY (cm/sec)

Fig. 52. Variation of Axial Dispersion Coefficient with Changes in
Superficial Gas Velocity and Number of Orifices in Gas Distributor in a
2-in.-ID Column Filled with Water. The gas inlet orifices had a diameter
of 1 mm. The transition between the bubble- and slug-flow regimes for
the case of a single orifice is denoted by a vertical arrow.
AXIAL DISPERSION COEFFICIENT (cm?/sec)

ORNL DOWG. 73-5IR}

 

500 T T T T ! I T I I
400} VISCOSITY fi
(cP)
300 o 0.9 ]
A 2.05
200 O o107 ]

100

50
40

30

20

 

 

 

J ] i ] J i l i

 

025 05

Fig. 55.

Ol 02 0.5 LO 2.0 50 10 20 25

SUPERFICIAL GAS VELOCITY (cm/sec)

Variation of Axial Dispersion Coefficient with Changes in

Superficial Gas Velocity and Viscosity of Liquid in a Column Having a
Diameter of 3 in.

641
AXIAL DISPERSION COEFFICIENT (cm®sec)

ORNL DWG T73-3

 

1000
F

T00

500

300

200

100

70—

30

20

 

T 1 T T T TTT T I t T T T rTT I i lll//I;’

 

1 i I 1 1 1111 1 P 1t 11l I [ 1 1 1 111

VISCOSITY (cP)
. O.g —
A 1.8

0 2.1 -

 

 

10
0.1

02 03 05 07 1.0 2.0 3.0 50 70 10 20 30 50 70 100

SUPERFICIAL GAS VELOCITY (cm/sec)

Fig. 54. Variation of Axial Dispersion Coefficient with Changes in

Superficial Gas Velocity and Viscosity of Liquid in a 6-in.-ID Bubble

Column.

ORI
141

superficial gas velocity. The bubble rise velocity (for the case of a
negligible liquid superficial velocity) is determined by the ratio of
the superficial gas velocity to the gas holdup and remains essentially
constant at a value in the range of about 25 to 40 cm/sec. This flow
regime is designated as bubble flow. Bubble size and bubble size
distribution in this flow regime are a complex function of gas inlet
orifice diameter, gas flow rate, and properties of the liquid phase.
For air-water systems, bubble rise velocities of 25 to 30 cm/sec have

50,31

been reported in the literature as approximate values.

As the superficial gas velocity is increased, bubble-bubble inter-
actions become important and some coalescence is observed at the top
of the column, resulting in the formation of larger bubbles. This marks
the onset of transition flow, and further increases in the superficial
gas velocity lead to greater coalescence and the formation of bubble
platelets having a size in the horizontal directiofi equal to about 90%
of the column diameter and lengths of about 1 to 2 cm. Further increases
in the superficial gas velocity result in still greater coalescence and
the formation of bullet-like bubbles whose diameters are almost equal
to the column diameter. These large-diameter bubbles, or "slugs", are
initially formed near the top of the column; however, the point of slug
formation gradually moves downward with increases in the superficial
gas velocity. Thus, in the transition region, part of the column may
be characterized by bubble flow and part by slug flow, and the axial
dispersion coefficient begins to increase at a higher rate with increases

in the superficial gas velocity.

In the slug flow regime, bubbles that have a smooth, spherical cap
and a flat trailing edge are present throughout the column. The bubbles
or slugs are usually 2 to 7 cm in length, although even greater lengths
are observed at very high gas velocities. 1In this flow regime,
acceleration of the liquid near the column wall occurs since the liquid
must pass through a narrow, annular channel between the gas-liquid inter-
face and the column wall. Vigorous mixing is produced at the rear of
each slug, and the axial dispersion coefficient increases rapidly with

increases in the superficial gas velocity. 1In slug flow, the dependence
1h2

of gas holdup on superficial gas velocity diverges from the linear

relationship observed at low superficial gas velocities.

The mixing patterns observed in a 6-in.-diam bubble column for
superficial gas velocities greater than 6 cm/sec were different from
those observed for smaller columns. No pronounced transition to slug
flow was observed; instead, very turbulent circulation and mixing
patterns were seen. Three types of bubble motion were noted: small
bubble swirls that circulated downward along the column wall; long,
extended bubbles that spiraled up the column across the entire cross-
section; and a core of bubbles that ascended up the center of the
column. Similar circulation patterns have been observed by Towell et

gl:BE in large-diameter bubble columns.

The variation of the axial dispersion coefficient with changes in
the superficial gas velocity is shown in Fig. 55 for water-filled
columns having diameters of 1, 1.5, 2, 3, and 6 in. The axial dis-
persion coefficient was found to depend on the superficial gas velocity
to the 0.12 and 0.63 powers in bubble and slug flow, respectively,

regardless of the column diameter.

9.8.2 Effect of Column Diameter

 

The variation of the axial dispersion coefficient with changes in
column diameter is shown in Fig. 56 for superficial gas velocities of
0.5 cm/sec (bubble flow) and 25 cm/sec (slug flow). 1In each of the
two flow regimes, the axial dispersion coefficient increases rapidly
with increases in the column diameter. In bubble flow, the axial dis-
persion coefficient shows a 1.5 power dependence on column diameter.

35 5h

Aoyama and Imafuku also reported a 1.5 power dependence of the
axial dispersion coefficient on column diameter. Ohki35 derived semi-
empirical relations which suggest that the dispersion coefficient is
dependent on the column diameter to the 2.0 power in bubble flow. The
power dependence of the dispersion coefficient on column diameter, as
predicted by Egs. (33) and (38), is not constant but has a value of
about 0.64. The semiempirical relations derived by Ohki55 suggest a

power dependence of 1.0.
AXIAL DISPERSION COEFFICENT (cm2/sec)

ORNL DWG. 73-46

 

800
600

  
   

H
3

200

3

 

 

1 S N N | 11 11t ! 11 b1
ol 0.2 c4 06 08 10 2.0 40 60 80 10 20 40 60 80 100
SUPERFICIAL GAS VELOCITY (cm/sec)

 

10

Fig. 55. Variation of Axial Dispersion Coefficient with Changes in
Superficial Gas Velocity and Column Diameter.

41
144

ORNL DWG T73-32
1000 T I T T T T

900 -
800 -
700 -

600 -

 

500 .

400 Vg =25 cm/sedt

300 -

> b -
200 Vg = 0.5 cm/sec

00 -
30
80 -

T

60} —

S50 -

AXIAL DISPERSION COEFFICIENT (cmsec)

40} .

30 -

 

 

10 1 1 1 ] 1 1
0 I.O 2.0 3.0 4.0 5.0 6.0

COLUMN DIAMETER {in))

 

Fig. 56. Variation of Axial Dispersion Coefficient with Changes in
Column Diameter and Superficial Gas Velocity.
145

Although the transition from bubble to slug flow in the present
studies did not occur at a single value of the superficial gas velocity,
certain qualitative conclusions can be inferred. As the column diameter
is increased, the superficial gas velocity at which slug flow is
obtained increases rapidly. Slug flow was obtained in the l.5-, 2-,
and 3-in. columns at superficial gas velocities of 0.005, 1.14, and
5.54 cm/sec, respectively, for bubble columns filled with water; how-
ever, slug flow was not obtained in the 6-in.-diam column at velocities

as high as 9.5 cm/sec.

9.8.3 Effect of Viscosity of the Liquid Phase

The variation of the axial dispersion coefficient with changes in
the viscosity of the liquid and the superficial gas velocity is shown
in Figs. 4k, 46, 47, 53, and 54 for columns having diameters of 1.5,

2, 3, and 6 in. 1In each case, no observable change was noted in the
axial dispersion coefficient as the viscosity of the liquid was
increased from 0.9 cP to about 2 cP. A further increase in viscosity
to about 11 ¢P resulted in a decrease of about 20% in the axial dis-
persion coefficient for the smaller columns (1.5 and 2 in. in diameter),
while the effect was less pronounced for the 3-in.-diam column. How-
ever, increasing the viscosity of the liquid by as much as a factor of
12 produced no observable change in the dispersion coefficient for the
6-in.-diam column. Although the axial dispersion coefficient values
for the more viscous solutions were lower, the observed dependence of
the axial dispersion coefficient on the superficial gas velocity was

the same as that observed for columns filled with water. These obser-
vations are in agreement with the correlations shown in Eqgs. (37) and
(38), which predict that the dispersion coefficient depends on the
viscosity of the liquid phase to the -0.057 and -0.072 powers for bubble

and slug flow respectively.

The variation of gas holddp and bubble rise velocity with changes
in the viscosity of the liquid is shown in Figs. 25-32 for 1.5-, 2-,
3-, and 6-in.-diam columns. An increase in the viscosity of the liquid

results in a slight increase in the gas holdup for the smaller-diameter
146

columns. However, the gas holdup values for the larger columns appear
to be unaffected by changes in viscosity. The relation used for corre-
lating the holdup data [Eq. (33)] shows no dependence of holdup on

viscosity of the liquid phase.

9.8.4 Effect of Surface Tension of the Liquid Phase

 

The effects of changes in the surface tension of the liquid on the
axial dispersion coefficient and the gas holdup in a 1.5-in.-diam column
are shown in Figs. 45 and 33 respectively. A decrease in the surface
tension of the liquid from about 72 dynes/cm to about 27 dynes/cm resulted
in a 25% decrease in the dispersion coefficient during slug flow; how-
ever, essentially no change was observed during bubble flow. A slight
increase in the gas holdup was noted as the surface tension was decreased
for a given value of the superficial gas velocity. The correlations for
the axial dispersion data indicate no dependence of dispersion coeffi-
cient on surface tension during bubble flow and a 0.43-power dependence
during slug flow. The correlation for gas holdup indicates that holdup
is not affected by changes in surface tension. Decreases in the sur-
face tension of the liquid delayed the change from transition flow to
slug flow; that is, the change in flow regime occurred at progressively
higher values of the superficial gas velocity as the surface tension of

the liquid was decreased.

9.8.5 Effect of Gas Inlet Orifice Diameter

 

The variation of the axial dispersion coefficient with a change in
the gas inlet diameter is shown in Fig. 49 for a 2-in.-diam column
filled with water. The gas distributor consisted of a single orifice
having a diameter of either 0.1 or 0.432 cm. Only a minor effect at

low superficial gas velocities was noted.

The effect of the diameter of the gas inlet orifice on the axial
dispersion coefficient in a 2-in.-diam column employing a gas distribu-
tor consisting of five orifices is shown in Fig. 50. The gas inlet ori-
fice diameter was varied from O.4 to 4 mm, which corresponds to a range

of 13 to 127 for the column diameter/orifice diameter ratio. For gas
1h7

distributors containing multiple orifices, the axial dispersion coef-
ficient values in the transition flow regime exhibit an anomalous
behavior, which is probably due to complex bubble-bubble interaction
effects. A similar behavior was observed by two other 3’.nvestJ'.gatc'rs,56’57
who considered it to be a hindered mixing effect caused by bubble

coalescence.

In the slug flow regime, the axial dispersion coefficient values
show no dependence on gas inlet orifice diameter. This is to be
expected since the large bubbles or slugs formed by coalescence of
smaller bubbles are independent of the initial bubble population. 1In
the bubble flow regime, however, the smallest orifice diameter (0.4 mm)
produced axial dispersion coefficient values that were significantly
lower than those produced by the other, larger orifices. The axial dis-
persion coefficient values for the larger orifices were essentially the
same (within the accuracy of the data). The gas holdup values for the
1.0-, 2.0-, and L4.0-mm-diam orifices were identical, while the values
obtained by using a O.L-mm-diam orifice were higher. The higher values
imply a lower bubble rise velocity. Thus, except in the case of very
small orifice diameters, the dispersion coefficient is independent of
orifice diamster,.as is assumed by the dispersion coefficient corre-

lations.

Data were also obtained with gas distributor plates containing 37
orifices. No effect on axial dispersion coefficient was observed
(within the accuracy of the data) when the orifice diameter was varied
from 1 to 4 mm, as shown in Fig. 51. Reith58 found no effect ef orifice
diameter on axial dispersion during slug flow; during bubble flow, the
axial dispersion appeared to be marginally affected, although no quanti-
tative dependence was reported. For column-to-orifice diameter ratios
less than 80, Ohki35 observed a very small influence of orifice diameter
on axial dispersion. However, for higher values of this ratio, a marked

decrease in the dispersion coefficient was observed.
148

9.8.6 Effect of Number of Orifices in Gas Distributor

 

The effect of the number of orifices in the gas distributor on the
axial dispersion coefficient is shown in Fig. 52 for a 2-in.-diam bubble
column filled with water. The orifice diameter was 1 mm in each case.
In the bubble flow regime, the axial dispersion coefficient increased
as the number of orifices in the gas distributor was increased from 1
to 37. 1In the transition flow regime, no clear dependence on number
of orifices is apparent. In the slug flow regime, the axial dispersion
coefficient values are essentially independent of the number of orifices
in the gas distributor. This is to be expected since, for a given super-
ficial gas velocity, the number of orifices will only alter the initial
size and number of bubbles, and the effects of these quantities are
eliminated by coalescence and slug formation. The correlations of the
dispersion coefficient data indicate that the dispersion coefficient is
weakly dependent (0.098 power ) on the number of orifices in the gas
distributor during bubble flow and independent of the number of orifices

during slug flow.

9.9 Future Work

It is believed that sufficient experimental data on gas holdup and
axial dispersion have been obtained; hence no further experimental work
in this area is planned. In the future, efforts will be made to develop
relations that correlate the data on gas holdup satisfactorily and also
extrapolate well to known limiting cases, such as the rate of rise of
individual bubbles in vessels having very large diameters. Attempts
will also be made to correlate the dispersion data obtained during slug
flow without including gas holdup dependence since such a correlation

would probably be much simpler.
1k9

10. SEMICONTINUOUS REDUCTIVE EXTRACTION EXPERIMENTS
IN A MILD-STEEL FACILITY

B. A. Hannaford C. W, Kee
L. E. McNeese

We have continued operation of a facility in which semicontinuous
reductive extraction experiments can be carried out in a mild-steel

39

system. Initial work with the facility was directed toward obtaining
data on the hydrodynamics of the countercurrent flow of molten salt

and bismuth in a 0.82-in.-ID, 24-in.-long column packed with 1/L-in.
molybdenum Raschig rings. We have been able to show that flooding data
obtained with this column are in agreement with predictions from a
correlationbrO based on studies of the countercurrent flow of mercury
and aqueous solutions in packed columns. We have carried out several
experiments for determining the mass transfer performance of the packed
column in which a salt stream containing UFu was countercurrently con-
tacted with bismuth containing reductant over a range of operating
conditions. During the second uranium mass transfer run (UTR-2), 95%
of the uranium was extracted from the salt during a 40-min period in
which the bismuth and salt were countercurrently contacted at flow rates
of 247 and 52 cmj/min respectively.hl Two additional uranium mass
transfer runs (UTR-3 and -4) were carried out under conditions such
that the uranium extraction factor remained high throughout the column.
The fraction of uranium extracted from the salt phase increased from
0.63 to 0.91 as the metal-to-salt flow rate ratio was varied from 0.75
to 2.05. It was found that the rate at which uranium transferred to
the bismuth was controlled by the diffusive resistance in the salt film,
and that the extraction data could be correlated in terms of the height
of an overall transfer unit based on the salt phase. The HTU values
ranged from O.77 to 2.1 ft as the bismuth-to-salt flow rate ratio was
decreased from 2.05 to 0.75.h2 In order to measure mass transfer rates
in the column under more closely controlled conditions and under condi-
tions where the controlling resistance is not necessarily in the salt
phase, preparations were begun for experiments in which the rate of

exchange of zirconium isotopes will be measured between salt and bismuth
150

phases otherwise at chemical equilibrium. In preparation for these
experiments, reductant was removed from most of the bismuth by hydro-
fluorination of the salt and bismuth with a 70-30 mole % HE—HF mixture.
After hydrofluorination, the salt and bismuth were transferred through
the system (during run UTR-5) in order to obtain uniform concentrations
in the salt and bismuth phases throughout the system. There was a sur-
prisingly large variation (155%) in the reported concentrations of

b3

uranium in salt samples removed from the column effluent during the run.

10.1 Run UTR-6

The objectives of run UTR-6 were: (1) to obtain additional informa-
tion relative to the variation in the reported uranium concentrations in
samples taken from the salt stream leaving the extraction column, because
of the large variation (+35%) in these values during run UTR-5; and (2)
to obtain additional hydrodynamic data on the countercurrent flow of salt
and bismuth in the packed column by operating under conditions where
flooding was predicted by the correlation based on countercurrent flow
of mercury and water in packed columns. Following run UTR-5, the salt
and bismuth were returned to the treatment vessel, from which the salt
and bismuth phases were transferred to their respective feed tanks.

After the transfers, the bismuth feed tank contained about 17 liters of
bismuth in which the thorium concentration was about 6 ppm. The salt
feed tank contained about 15 liters of salt having a uranium concentra-
tion (as UFM) of about 3100 ppm. Essentially no extraction of uranium
from the salt during the run would be expected because of the negligible
concentration of reductant in the bismuth. As shown in Fig. 57 the run
was initiated by starting a salt flow through the column; then, a few
minutes later, a bismuth flow was begun. Constant salt and bismuth flow
rates of 125 and 117 cmi/min, respectively, were attained after a short
time, and seven pairs of bismuth and salt samples were withdrawn from
the salt and bismuth streams leaving the column. Subsequently, the
flows of salt and bismuth to the column were stopped, and the freeze
valve below the specific gravity pot was closed by cooling the line to

a temperature below the salt liquidus temperature. The run was then
REMAINING ()

FEED

ORNL DWG.73-48

 

  
 

 

20
IvvivllTIllllrr'trIITTTTlIlIt!TT1rlllirrvlvtrllrillllltr1ll"'llillrv1111l1I1I111IVII[vr1ll1llv'1l1|||lt1|llluvlllllrvtwvl||1I111l|1]11]1||1|lrl]_
+ + + + + o+ o+ + + + + + + 4+ SAMPLE PAIRS TAKEN:
I 2 3 4 5 6 7 8 9 10 11 12 13 14 3
I8
H7 cm3
l BISMUTH /min BISMUTH
————— SALT
S
~
S
14 s
12
209 cm3
0 ~__ l BISMUTH/min
I25cm:5
8 e SALT 7/ min ']
6
4
2
O
0 10 20 30 40 50 60 70 80 90 100 1o (20 130 140
TIME (min)

Fig. 57. Volume of Bismuth and Salt Remaining in Feed Tanks vs Run
Time, Run UTR-6. Volumetric flow rate (ml/min) for each indicated inter-

val was inferred from the slope.

 
 

 

 

L unlluluunnllll lllll|lluu_1un|1ulluu_l_uu IlllllIllllllllluAlllllllllllll_l_lllll 233

o
O

16T
152

resumed with salt and bismuth flow rates of 150 and 209 cmB/min. This
combination of salt and bismuth flow rates was selected in order to
further test the prediction of flooding by the correlation based on
countercurrent flow of mercury and aqueous solutions in packed columns
(see Sect. 1k). Seven additional salt samples were taken during the
remainder of the run for bismuth analyses. Shortly after the salt and
bismuth flows were resumed, the apparent dispersed-phase holdup (as
indicated by column pressure drop) stabilized at 30%. The salt flow
rate was then increased slightly (to 151 cm3/min). This caused the
apparent dispersed-phase holdup to increase until it reached a value
of about 60%, at which point the specific gravity pot began to fill
with bismuth (which clearly indicated that the column was flooded ).
The conditions under which the column flooded agree reasonably well
with the values predicted from the correlation since the ratio of the
observed column throughput to the predicted throughput at flooding was
1.13.

Results of the analyses of the salt and bismuth samples taken
during the first part of the run are shown in Table 4l. Little signif-
icance is attached to the analyses of the bismuth phase; the indicated
increase in the concentration of reduced metals is probably the result
of slight contamination of the samples with salt or may simply represent
analytical errors. The uranium concentration in the salt samples should
have remained constant at a value of 3100 ppm; nevertheless, the data
show considerable scatter. The average of the uranium concentrations
in the flowing salt samples is only 2714 ppm; however, the uranium
content of the salt sample removed from the receiver tank is in agree-
ment with the expected value. The variation of uranium concentrations
in samples removed from the salt stream during this run is similar to
that observed during run UTR-5; on the other hand, a much smaller
variation in reported uranium concentrations was observed during an

b1

earlier, similar run (UTR-1).

Bismuth concentrations in the salt samples taken during the second
half of run UTR-6 are given in Table L42. These concentrations are

relatively low, ranging from 0.8 to 15 ppm. It should be noted that
153

Table 41. Summary of Results for Samples Taken During Run UTR-6

 

Materials in

 

Uranium Conc.

 

Sample Source Bismuth Samples (ppm) in Salt
Li Th U (ppm)
Feed tanks 0.052 <2 <1 3100
0.42 6 1 3200
Flowing stream
1 0,060 9 2.45 2700
2 0.046 15 <1 3200
3 0.046 5 <1 2400
4 0.055 14 <1 3100
5 0.042 4 <1 2300
6 0.035 6 <l 2500
7 0.063 53 <1 2800
Avg. 2714
Receiver tanks 0.043 <2 <8 3100

 
15k

Table 42. Tests for Bismuth Entrainment in Flowing
Salt Samples? Taken in Run UTR-6

Flow rates: 150 ml/min salt, 209 ml/min bismuth

 

Sample Bismuth Concentration Equivalent Bismuth Entrainment”

(Fs-) (ppm ) (vol %)

 

 

 

8 0.8 --

9 6.4 0.00015
10 6.9 0.00017
11 9.4 0.00026
12 15 0.00046
13 11 0.00032
14 8 0.00021

 

A 0.042-in.-diam hole drilled in the side of a standard fritted sample
capsule permitted collection of unfiltered samples.

bCorrected for the average concentration of bismuth reported for blank
sample (2.0 ppm).

some of the bismuth entrained in salt leaving the column may have been
separated from the salt in the specific gravity pot before the salt

stream entered the sampler.

10.2 Preparation for Run UTR-T; Installation of Molybdenum
Draft Tube in Treatment Vessel

After run UTR-6, the salt and bismuth phases were transferred to
the treatment vessel so that reductant could be added to the bismuth
phase and produce the desired uranium and zirconium distribution coef-
ficient values in preparation for the zirconium mass transfer experi-
ments. It was desired that about half of the uranium and zirconium be
reduced into the bismuth phase in order to give a distribution coef-
ficient of about 1 for the zirconium. A perforated mild-steel basket
(shown in Fig. 58) containing 95.3 g of thorium metal was lowered into

the bismuth in the treatment vessel, which was held at a temperature of
 

~ 8 PHOTO 101669

Fig. 58. Molybdenum Draft Tube Used for Agitating Bismuth and Salt
Phases, and Perforated Steel Basket Used for Adding Thorium Metal to the
Bismuth Phase in the Treatment Vessel.

QST
S — = ———— e — e — —

156

600°C. The thorium dissolution rate was low; only 70 g of the metal
had dissolved after 50 hr. In order to improve mixing within the bis-
muth phase and contact between the bismuth and salt phases, a molyb-
denum draft tube (shown in Fig. 58) was installed in the treatment
vessel. As shown in Fig. 59, this tube had a diameter of 1 in. and a
length of & in. A 1.5-in.-diam baffle was placed directly above the
draft tube in order to deflect the discharged bismuth into the salt
phase. The length of the draft tube was chosen such that bismuth from
the bottom of the bismuth pool would be contacted with the salt phase.
After installation of the draft tube, an additional 105 g of thorium
metal was placed in a new reductant addition basket and was lowered
into the bismuth phase. Salt samples taken after 17 hr of operation
of the draft tube at an argon flow rate of 1.5 scfh showed that the
uranium concentration in the salt had decreased from 2800 ppm to 2000
ppm and implied that the thorium dissolution rate was still relatively
low. The temperature of the treatment vessel was increased to 650°C,
and the argon flow rate to the draft tube was increased to 2.5 scfth
(which is about twice the rate that had produced good pumping in earlier
qualitative tests with the draft tube using mercury and water). After
90 hr the perforated basket was removed and weighed; results showed
that about 86 g of thorium had dissolved. The thorium dissolution
rate was only about 1 g/hr, and it apparently was not increased by

use of the draft tube or the higher temperature. Bismuth samples
removed at the end of the thorium addition period were found to be
contaminated with salt, thus preventing a direct measurement of the
uranium and zirconium distribution coefficients. However, a salt sample
contained 1580 ppm of uranium, which indicates a uranium distribution
coefficient of 1.11 and a zirconium distribution coefficient of 1.01,
assuming that the salt and bismuth phases were homogeneous and that
chemical equilibrium had been attained before the salt and bismuth

phases were sampled.

Following equilibration of the salt and bismuth in the treatment
vessel, the phases were transferred to their respective feed tanks and

run UTR-7 was carried out in order to obtain more nearly constant
127

ORNL DWG 73-39

ARGON IN

o~ 3/8-in.-00 TUBING

DEFLECTOR FORMED FROM

 

 

 

 

 

 

 

 

 

 

 

45° CONES, 1.5-in.-diam BASE
T T
o | &
b | |
|
b
1 . . .
. -in.-00, O.lI10-in.-wall TUBING
I
i
; I
1
1y
]
h | |
8l
i |
1
4 |
b
]
Ly
]
iR
: ]
8 o
: | i
i
I
i i
i |
I
|
|
R
|
)
i
ty
1
|y
1
i
I 3 EACH 0.040-in.-diam HOLES
| :/H/
|
¥
1 ) | !
2y | )
18 - 3/32-in.-diam SUPPORT RODS,
2y BOTH ENDS OF TUBE

 

Fig. 59. All-Molybdenum Draft Tube for Circulating Molten Bismuth
from a Pool Approximately 7 in. Deep.
158

concentrations of materials in the salt and bismuth throughout the sys-
tem. During the run, the pressure drop through the column was erratic
and an abnormally high pressure was required in the salt jackleg. No
hydrodynamic data were obtained during the run because of variations in
the flow conditions. Post-run examination of the system indicated that
the bismuth return line between the column and the bismuth catch tank
was slightly restricted and that a small rupture had occurred, causing

the release of about 20 g of bismuth.

10.3 Preparation for Mass Transfer Run ZTR-1; Production of
977y by Irradiation of 96Zr

Plans have been made to carry out a series of mass transfer experi-
ments in which the rate of exchange of zirconium isotopes between salt
and bismuth phases containing natural zirconium will be measured. The
results that are obtained will permit the mass transfer performance of
the packed column to be evaluated. Zirconium-97, which has a 16.8-hr
half-1ife, is thought to be the best zirconium isotope for these experi-
ments since its use would not result in the successive buildup of radio-
activity in the experimental facility. In our initial attempt at pro-
ducing the tracer, 20 mg of 96ZrO2 was sealed in a quartz ampul (Super
Seal), which was then irradiated for 5 hr at a thermal flux of 2 x 1014
neutrons cm_2 sec:_1 in the 0Oak Ridge Research Reactor. The ZrO2 had

the following isotopic analysis: 96Zr, 85.257%; 94Zr, 3.85%; 92Zr, 2.24%;

912r, 1.41%; and 9OZr, 7.25%. After the irradiation period, the result-

ing 97Zr activity was about four times the calculated value. The dis-

b3

crepancy was traced to the use of an outdated value for the thermal-
neutron cross section of 96Zr (0.05 barn) as compared with a more
recent ValuemL of 0.2 barn. After a period of 20 days the 96ZrO2 was
reirradiated for a period of 2 hr in order to produce a calculated ~ Zr
activity of 6.9 mCi. The measured activity after irradiation was about
100% higher than expected but was well within the limits required for
safe handling after an 18-hr decay period, during which the activity

of 3lSi (half-life, 2.62 hr) resulting from activation of the quartz
ampul was allowed to decrease to a low level. The activity of 65-day

5
Zr produced during irradiation of the ZrO2 was calculated to be about
159

o1 95

0.1% of that observed initially for ° Zr; therefore, the cumulative 7~ Zr
activity that will result from several tracer experiments is expected to

be negligible.

The irradiated ZrO2 powder was transferred from the quartz ampul to

a 3/4-in.-diam carbon-steel capsule, which was used for adding the ZrO2
to the salt feed tank. The design of the capsule is shown schematically
in Fig. 58. After the capsule had been lowered into the salt feed tank,
argon was fed through the capsule at the rate of ~ 1 scfh in order to
dissolve the ZrO2 in the molten salt. Samples of the salt were taken
periodically following the addition of tracer and were counted for 972r
activity. These measurements indicated that addition of the tracer to
the 15-liter salt volume was 75% complete within a 2-hr period. The 9TZr
counting rate in a sample taken 10 hr after the addition of tracer to the
salt feed tank was about 200,000 counts/min, which indicates that the
counting precision in the planned tracer experiments should be satisfac-

tory.

When the salt feed tank was pressurized prior to starting experiment
ZTR-1, salt did not begin transferring from the feed tank at the expected
tank pressure. After a salt flow was established, it soon became appar-
ent that a leak existed in the salt exit line at a point near the tank.
Therefore, the experiment was canceled and an attempt was made to trans-
fer the salt to the salt receiver tank via a bypass; however, this
resulted in a second salt leak, which caused the Calrod heaters on the
salt feed-and-catch tank to fail. Results of examination of the tank,
along with details on equipment maintenance, are given in the following

section.

10.4 Inspection of Salt Feed-and-Catch Tank,
and Equipment Maintenance
Inspection of the salt feed-and-catch tank, after cooling to room
temperature, showed that the failures in the transfer lines leaving the
tank were identical, although they had occurred in different lines. 1In
each case, the 3/8-in.-diam mild-steel tubing (initial wall thickness,

0.59 in. ) had failed due to severe air oxidation in a narrow band on
160

the exterior of the tubing adjacent to a compression fitting; of course,
thinning of the tubing caused by the ferrules in the compression fitting

may have also contributed to the failures.

Damage to the protective coating of nickel aluminide on the vessel
exterior and to the Calrod heaters made salvage of the vessel impractical.
(The outside of the vessel was coated initially with 15 to 20 mils of
nickel aluminide in order to decrease the rate of external air oxidation.)
The protective coating appeared to be intact over a significant area of
the tank, but it was either nonadherent or had completely spalled from
the surface in a number of areas where extensive air oxidation had
occurred. A l1.5-in.~diam metal specimen that was removed from the tank
wall with a hole saw was examined by the Metals and Ceramics Division.LLS
The vessel had operated for about 16,000 hr in the temperature range of
600 to 700°C and had been cooled to room temperature three or four times.
The wall of the vessel had been constructed from 12-in.-diam, 3/8-in.-
wall-thickness, seamless ASTM-AlO6 Grade B Steel [Fe--0.3% C (max)--0.29
to 1.06% Mn--0.048% P (max)--0.58% § (max)--0.1% Si (min)].

A cross section of the specimen from the vessel wall was mounted so
that the inside and outside surfaces could be observed. Figure 60 pre-
sents photomicrographs of the inside surface. Intergranular voids that
extend 2 to 3 mils into the metal are apparent in Fig. 60(a). The
marks near the surface in Fig. 60(b) are directional polishing marks and
have no significance. The inside of the vessel appeared to be in good
condition, with little evidence of corrosion. Figure 61 presents photo-
micrographs of the external surface of the vessel. The nickel aluminide
coating and most of the oxide corrosion product spalled from the speci-
men during its preparation. The predominant feature noted in the photo-
micrographs is the absence of fine black precipitates to a depth of
about 10 mils. These precipitates are probably small graphite flakes,
and their absence near the surface is indicative of decarburization,
which would be expected under the highly oxidizing enviromment in which

the vessel had operated.

It was of particular interest to determine whether the steel had

graphitized. Steels normally contain carbon in the form of Fe,C; however,

5
161

 

(b]

Fig. 60. Photomicrographs of the Inner Surface of the Salt Feed-and-
Catch Tank in the (a) As Polished and (b) Etched Condition After Exposure
to Salt at 600 to T700°C for 16,000 hr. Magnification, 100X. Etchant,

o% Nital.
162

PHOTO 1631-73

 

= -
- "
& = ] ¥ - ¥ =
. ‘ =
-
- *
r - - -
A . . :
uld" £ -
» -J ' - F4 - » -
* i & ‘1 - & = + -
- v 0 L
. T A + _-"‘ £
- - = & r
v Av s 78 ~
A - e ot ’
. -? - - .’ % .1‘ - - €
Vg . A ' vt i '
- ‘ = *
- - =~ - -
-
- *a { = % § -
-y g . A e
¢ s T S =
- . 3 . & . - s :‘ ¢
- . '_- ‘,
- Tewme - -
P R ",‘: L= .
. " = . e F .
- - -

Fig. 61. Photomicrographs of the Outer Surface of the Salt Feed-and-
Catch Tanks in the (a) As Polished and (b) Etched Condition After Exposure
to Air at 600 to TO0°C for 16,000 hr. Magnification, 100X. Etchant, 2%
Nital.
 

 

 

163

this carbide will often decompose during prolonged exposure at elevated
temperatures to iron and graphite. Decomposition is influenced strongly
by the chemical composition of the steel, with the presence of Ni, Si,
and C favoring the formation of graphite, Fe and S (if not present as

a sulfide) favoring the stability of Fefic, and Mn having little effect

on the stability of the carbide if present in concentrations below 1%.
Only the nominal composition of the specimen is known. The most influ-
ential elements are carbon and silicon, and both would encourage graphite

formation.

Examination of the microstructure shown in Fig. 62 indicates that
two types of precipitate are present. The long stringers, which are
gray, are typical of Feic' The black precipitate is in the form of
flakes that are typical of the appearance of graphite. The photomicro-
graphs in Figs. 60(a) and 61(a) show that the mixed carbide-graphite
structure is typical of most of the cross section. The graphite flakes
may have been present before the vessel was put into service, or they

may have formed during its operation.

The important effects of graphite formation are reductions in
strength and ductility of the steel. The specimen was not of sufficient
size for strength measurements; however, a test that would indicate the
ductility of the material was made. A small piece (about 1.5 in. long,
3/8 in. thick, and 5/16 in. wide) of the specimen was bent 180° in a
vise without formation of cracks, as shown in Fig. 63. Thus, no evidence

for embrittlement of the material is apparent.

Dimensional measurements by the Inspection Engineering Division on
the salt feed-and-catch tank and on the bismuth feed-and-catch tank
were inconclusive, partly as a result of air oxidation of the reference
points on the vessels and partly because of inaccessibility of the
points. The observed strain, based on comparison with measurements made
on the tanks after an initial stress relief, showed a variation of

approximately + 1%, with the mean value being near zero.

A new salt feed-and-catch tank of the initial design was fabricated

and installed in the system. Thermal insulation was removed from all
 

 

164

Y-104840

 

 

 

Fig. 62. Photomicrograph Showing Microstructure of Steel from Sg-:lt
Feed-and-Catch Tank in the As-Polished Condition After Operation at 600
to TOO°C for 16,000 hr. Magnification, 500X.
165

Y-104649

 

Fig. ©63. Photograph of a Sample of Steel from the Salt Feed-and-
Catch Tank After Being Bent 180° at Room Temperature. The dark spots
are remnants of the original surface oxide. The tank was operated at
600 to T700°C for 16,000 hr. No evidence for embrittlement of the steel
is apparent.
166

transfer lines to allow inspection of the lines, and lines that were
more than moderately oxidized were replaced. The salt transfer line
from the feed tank to the salt jackleg was rerouted to a point 1l in.
higher than in the original design in order to improve control of the
salt feed rate and to prevent backflow of bismuth into the salt feed

tank during column upsets.
167

11. DEVELOPMENT OF THE METAL TRANSFER PROCESS:
OPERATION OF EXPERIMENT MTE-2

E. L. Youngblood L. E. McNeese

It has been found that rare earths distribute selectively into mol-
ten LiCl from bismuth solutions containing rare earths and thorium, and
an improved rare-earth removal process based on this observation has
been devised. Work that will demonstrate all phases of the improved
rare—-earth removal method, which is known as the metal transfer process,

is presently under way.

We previously46 carried out an engineering experiment (MTE-1) for
studying the removal of rare earths from single-fluid MSBR fuel salt by
this process. During this experiment, approximately 507 of the lantha~
num and 25% of the neodymium originally present in the fluoride salt were
removed at about the expected rate. Surprisingly, however, the lanthanum
and neodymium removed from the fluoride salt did not accumulate in the
Li-Bi solution used for removing these materials from LiCl. It is
believed that reaction of impurities in the system with the rare earths

caused this unexpected behavior.

A second engineering experiment (MTE-2) was designed and put into
operation previously.47 This experiment, completed during this report
period, was carried out in a vessel, made of 6-in.-diam carbon-steel
pipe, which was divided into two compartments by a partition that extended
to within 1/2 in. of the bottom of the vessel. The compartments were
interconnected by a 2-in.-deep pool of molten bismuth, saturated with
thorium, which formed a seal between the two compartments. One com-
partment contained MSBR fuel carrier salt (72-16-12 mole % LiF—BeFZ—
ThF4) to which were added 7 mCi of 147Nd and sufficient LaF3 to pro-
duce a lanthanum concentration of 0.3 mole Z. The other compartment
contained LiCl, a cup containing a Li~-Bi solution, and a pump for
recirculating the LiCl through the cup at the rate of about 25 cm3/min.
The concentration of reductant (35 at. % lithium) in the Li-Bi solu-
tion was sufficiently high that essentially all of the lanthanum and

neodymium would be extracted from LiCl in equilibrium with the Li-Bi

solution. In order to obtain mixing in the main bismuth pool, about
168

107 of the metal volume was forced to flow back and forth through the
1/2-in. slot below the partition between the fluoride and chloride com-
partments every 7 min. Gas-lift sparge tubes were used to disperse drop-
lets of bismuth in the salt phase to improve contact between the salt and
bismuth phases in each compartment. Provisions were made for obtaining

filtered samples of the salt and bismuth phases in the experiment.

The remainder of this section describes the experimental procedure
followed in the experiment, gives a mathematical analysis of factors
affecting the rate of transfer of the rare earths, and presents the

results of the experiment.

11.1 Experimental Procedure

After the materials47 had been charged to the experiment, the sys-
tem was allowed to stand for 76 hr at operating temperature (650°C)
before circulation of the LiCl was initiated. The sequence of opera-
tions pursued during the experiment consisted in pumping the LiCl
through the Li-Bi container for a 3-hr period at the rate of 25 cm3/min.
Circulation of the LiCl was then stopped, and the salt and bismuth
phases were allowed to approach chemical equilibrium during a 4-hr
period before filtered samples were taken of the salt and bismuth.

This sequence was repeated on a three-shift-per-day basis during the
first week and a two-shift-per-day basis during the second week. For
the remainder of the three-month operating period, the LiCl was circu-
lated for about 7 hr during the day shift and the salt and bismuth were
allowed to approach chemical equilibrium during the off-shifts and on
week-ends. During the pumping periods, the Th-Bi phase was forced back
and forth between the fluoride and LiCl compartments at the rate of 10%
of the metal volume every 7 min to promote mixing of the bismuth phase.
The gas-1lift sparges were kept in operation almost continuously except

when samples were being withdrawn.

Eight days before the end of the experiment, 1 vol 7 of fluoride
salt (72-16-12 mole % LiF—Ber—ThFa) was added to the LiCl compartment
in order to determine the effect of entrainment of small amounts of

fluoride salt into the LiCl via circulation of the bismuth.
169

The filtered samples taken during the experiment were prepared for
analysis by cutting off the filter section with tubing cutters and
LT
e

cleaning the external surfaces with emery cloth. Th Nd activity in
a sample was determined by direct counting of the 0.5%5-MeV gamma radia-
tion emitted by the sample. The lanthanum concentration in a sample

228
e

was determined by neutron activation. Th Ra activity in the samples

was determined by counting the 0.G-MeV gamma radiation emitted by the

228Ac after a delay of about 24 hr to ensure that the 228Ac was in secu-

lar equilibrium with the 228Ra

*

11.2 Mathematical Analysis of Transfer Rate

A mathematical analysis was carried out to determine the removal
rate for a fluoride salt contaminant which could transfer between the
salt and bismuth phases in experiment MTE-2. In making the analysis,
it was assumed that equilibrium was maintained between the three phases
present in the system: the fluoride salt, bismuth that is saturated
with thorium, and the portion of the LiCl that is in contact with the
Th-Bi phase. Under these conditions, the concentrations of the trans-
ferring material in these phases are related by the following distri-

bution ratios:

X
BT
DF = v 2 (59)
Xrs
X
D, = -, (40)
CS
where
DF = distribution coefficient for the material transferring
between the fluoride salt and the bismuth,
DC = distribution coefficient for the material transferring

between the chloride salt and the bismuth,
XBT = concentration of the transferring material in the main

bismuth pool, mole fraction,
170

~concentration of the transferring material in the fluoride

Xrs

salt, mole fraction,

XCS

concentration of the transferring material in the LiCIl,

mole fraction.

If it is assumed that a portion of the transferring material is removed
from the LiCl during its passage through the cup containing the Li-Bi
solution and that an additional portion is removed from the system by
radioactive decay, a material balance on the transferring material con-
tained within the three phases that are assumed to be in equilibrium

yields the following relation:

& = AL —F X (41)
where
I = inventory of the transferring material in the three phases
that are assumed to be in equilibrium, moles,
t = time, sec,
N = decay constant for the transferring material, sechl,
FCS = rate at which LiCl is circulated through the cup containing
the Li-Bi solution, moles/sec,
f = fraction of the transferring material that is removed from
the LiCl during its contact with the Li-Bi solution,
XCS = concentration of the transferring material in LiCl that is

in equilibrium with the Bi-Th solution, mole fraction.

The inventory of the transferring material in the three phases that are

assumed to be in equilibrium is given by the following relation:

I = MpgXpg + Mpp¥pr *+ MogXeg (k2)

where

Mrs
Mpr

CS

quantity of fluoride salt, moles,

quantity of Bi-Th solution, moles,

=
1l

quantity of LiCl that is in equilibrium with Bi-Th solution,

moles.
171

Substitution of Eqs. (39), (40), and (42) into Eq. (41) yields the

following relation:

F f —

dX CS D
S C
T - M Xpg - ()

D
_F
Mpg + Mpplp + Meg D,

Integration of Eq. (L3) between the limits indicated in Eq. (Li),

XFS,p D t

T = — | A+ dt
FS F
Xps, 1 Mpg * MppPp + Mg D, 0

 

(14 )

5
o

where
XFS,i = concentration of the transferring material in the fluoride
salt at time t = O, when contact of the LiCl with the Li-Bi
solution is initiated, mole fraction,
XFS,p = concentration of the transferring material in the fluoride
salt at time tp, when contact of the LiCl with the Li-Bi

solution is stopped, mole fraction,

tp = length of time that the LiCl is circulated, sec,

yields the following relation for the concentration of the transferring
material in the fluoride salt after the LiCl has been contacted with

the Li-Bi solution for the indicated time period:

D
. e
ES,p _ exp{—| A + ¢ . (45)
Xps, i

 
172

During experiment MIE-2, several cycles of operation were carried
out in which the LiCl was contacted with the Li-Bi solution for a cer-
tain period of time. On conclusion >f this contacting period, no circu-
lation c¢f the LiCl was allowed to occur. The relative concentration of
the transferring material in the fluoride salt after a period in which
the LiCl was circulated is given by Bq. (45), and the relative concen-
tration of the transferring material after a period during which no

circulation occurred is given by the relation:

s | (At ), (46)
XFS,p s
where
XFS,S = concentration of the transferring material in the fluoride
salt at time ts’ mole fraction,
tS = length of time during which no LiCl circulation occurs,

5ec.

Combination of Eqs. (45) and (46) yields the following relation for the
concentration of the transferring material in the fluoride salt after
one cycle of operation in which the LiCl is circulated for part of the

time period and no circulation occurs during the remainder of the

 

period:
xFS
=225 - exp [—K(t + t )] exp (-Kt ), (47)
Xs, i > P b
2
h
where DF
chf -
K = ¢
DF
Mpg + MppPp * Mgg D,

Similarly, the relative concentration of the transferable material in
the fluoride salt after n cycles in which both pumping and equilibra-

tion occur is given by the following expression:
173

XFS sn [ = ] -
- = exp | -A z: (t . + t .) exp ('Kl E: t -) » (AB)
Xps, 1 j=1 83 PJ j=1 PJ

where
xFS,sn = concentration of the transferable material in the fluoride
salt after n cycles in which LiCl circulation and equili-~
bration occur, mole fraction,
tsj = length of time during which no LiCl is circulated during
cycle j, sec,
tpj = length of time during which LiCl is circulated during
cycle j, sec.
It should be noted, however, that the two time summations in Eq. (48)

have the following meanings:

n
t_ = .Z—Zl(tsj tr), (49)
_]__.
)
t = t ., (50)
pn j=1 Pl
where
tn = length of time between the beginning of the experiment and the
end of the nth cycle, sec,
tpn = length of time that LiCl circulation has occurred between the

beginning of the experiment and the end of the nth cycle, sec.
Substitution of Eqs. (49) and (50) in Eq. (48) yields the following rela-
tion for the relative concentration of the transferable material in the

fluoride salt at the end of the nth cycle:

XFS n
—22°20 -~ exp (-A tn) exp (-Kt_ ) . (51)

Xps, 1 pn

This relation indicates that a plot of the quantity §—§L§E exp (htn) Vs
FS,1i

the length of time that the LiCl has been circulated should yield a
straight line having a slope of -K. The quantity f can be calculated

after the value of K has been determined from a plot of this type.
17k

11.3 Experimental Results

11.3.1 Rates of Transfer of Neodymium and Lanthanum

During the experiment, the lanthanum and 147Nd that were added to

the fluoride salt initially behaved as expected. The rates of accumu-
lation of these two rare earths in the Li-Bi solution during the experi-
ment are summarized in Table 43 and shown in Fig. 64. There was essentially
no accumulation of either of the rare earths in the solution prior to
operation of a gas-lift sparge tube in the cup (about 50 liters of LiCl
was circulated through the Li-Bi compartment before operation of the
sparge tube was initiated). After 400 liters of LiCl had been circulated
(about two-thirds through the run), approximately 50% of the lanthanum and
approximately 30Z of the neodymium originally in fluoride salt were found
to be in the Li-Bi solution. During the last one~third of the run the
rare earths continued to accumulate ih the Li-Bi solution, but the rate
of accumulation could not be determined accurately because of a leak that
developed in the lithium-bismuth cup. This leak allowed about 30% of

the solution to flow into the area between the cup and the holder.

The extent of removal of lanthanum and neodymium from the fluoride
salt is summarized in Tables 44 and 45 and shown in Figs. 65 and 66. As
seen in the figures, more than 857 of the lanthanum and more than 50%
of the neodymium had been removed from the fluoride salt at the end of
the experiment. After about 50 liters of LiCl had been circulated through
the cup containing the Li-Bi solution, the concentrations of lanthanum
and neodymium decreased in the manner suggested by Eq. (51). There is
some uncertainty in the f wvalues that correspond to the lines shown in
the figures because of uncertainty in the values for the distribution
coefficients, as will be discussed later. The concentration of thorium
in the Li-Bi solution in which the rare earths were accumulated remained

below the limit of detection throughout the experiment (< 10 ppm).

Data on the distribution of lanthanum and neodymium between the
fluoride salt, the thorium-saturated bismuth, and the LiCl are summarized

in Tables 46 and 47. The distribution coefficient data were generally
Table 43.

Rate of Accumulation of La and Nd in the Li-Bi
Solution During Metal Transfer Experiment MTE-2

 

Volumé of

Volume of

 

Pumping LiCl La Nd Pumping LiCl La Nd
Time Circulated Concentration Concentration® Time Circulated Concentration Concentratiop®
(hr) (liters) (mg/g) (dpm/g) x 107 {hr) (liters) (mg/g) (dpm/g) x 107
0 0 0.0 0.0 108.2 173.8 - 0.149
6 9.6 0.005 0.0 124.2 195.3 4.46 0.180
12 19.6 0.023 0.001 137.6 211.2 4.91 0.200
18 25.1 0.058 -0.003 150.1 226.5 4.80 0.199
24 39.0 0.092 - 156.1 233.4 5.42 0.214
27 44.0 - 0.003 169.1 250.7 6.05 0.202
30 48.8 0.11 0.004 185.3 273.9 5.97 0.225
36 59.0 0.82 0.037 212.6 319.6 6.25 -
44.5 73.2 1.40 0.050 219.2 330.5 6.75 -
47.5 77.9 1.66 0.067 250.5 384.3 7.17 0.290
53.5 87.6 2.26 0.083 273.5 421.0 - 0.345
59.5 97.1 2.70 0.089 302.7 471.2 - 0.568
65.5 107.0 2.67 0.106 309.7 482.8 10.53 0.584
94.2 153.2 3.74 0.131 343.9 538.8 10.31 -
101.3 164.0 4.27 0.129 358.1 563.2 11.17 -

 

 

 

a . . .
Corrected for radiocactive decay during the run.

QLT
PERCENT OF TOTAL IN Li-Bi

80

70

60

4
o

&

W
o

20

ORNL DWG TI-TOR1

METAL TRANSFER EXPERIMENT-MTE-2

 

-~
L. XV
i 7
i XV /008 ]

Y
| ! T
X—~LANTHANUM ! ! ' ' ' [ '
O-NEODYMIUM X

alT

X" x

o O
ooooooo o

~0 ©Oo _

 

 

L x/x&flflp ]
*°

Lx 00X

1 I 1 l 1 l i I 1
i00 200 300 400 500 600

VOLUME OF LiCi PUMPED (LITERS)

Fig. 64. Rates of Accumulation of Lanthanum and Neodymium in the
Li-Bi Solution During Metal Transfer Experiment MTE-2.
Table 44. Concentration of Lanthanum in Fluoride Salt
During Metal Transfer Experiment MTE-2

 

 

Pumping Volume of LiCl La Concentration Pumping Volume of LiCl La Concentration
Time Circulated in Fluoride Salt Time Circulated in Fluoride Salt

(hr) (liters) (mg/g) (hr) (liters) (mg/g)
0 0 6.54 124.2 -195.3 4.14
6 9.6 6.15 137.6 211.2 3.70
12 19.6 5.78 150.1 226.5 3.93
18 29.1 6.26 156.1 233.4 2.96
27 44.0 6.33 169.1 250.7 3.40
30 48.8 6.20 185.3 273.9 3.09
36 59.0 6.72 212.6 319.6 2.04
44 .5 73.2 5.54 219.2 330.5 1.94
47.5 77.9 5.63 250.5 384.3 1.42
53.5 87.6 5.75 273.5 421.0 1.38
59.5 97.1 5.65 309.7 482.8 1.17
65.5 107.0 5.30 343.9 538.8 1.10
94,2 153.2 4.31 358.1 563.2 0.86

101.3 164.0 4.84

 

 

 

LLT
Table 45. Concentration of Neodymium in Fluoride Salt
During Metal Transfer Experiment MTE-2

 

 

Pumping Volume of LiCl Nd Concentration? Pumping Volume of LiCl Nd Concentration?
Time Circulated in Fluoride Salt Time Circulated in Fluoride Salt
(hr) (liters) (dpm/g ) x 10T (hr) (liters) (dpm/g) x 107

0 0 0.508 80.0 130.4 0.422
3 h.7 0.545 g4.2 153.2 0.390

6 9.6 0.L87 101.3 164.0 0.387
9 4.7 0.531 108.2 173.8 0.383
12 19.6 0.474 114.9 183.5 0.358
15 2h.5 0.473 124.2 195.3 0.397
18 29.1 0.458 130.7 202.3 0.329
21 3%.9 0.458 137.6 211.2 0.313
27 Lh.o 0.492 143.5 218.2 0.323
30 L8.8 0.463 150.1 226.5 0.347
33 5h.1 0.450 156.1 233.4 0.289
36 53.0 0.517 162.6 241.2 0.310
hh.s 73.2 0.493 169.1 250.7 0.268
LT.5 T7-9 0.LT77 175.9 258.5 0.300
50.5 82.7 0.471 185.3 27%.9 0.315
53.5 87.6 0.478 250.5 384.3 0.278
56.5 92.3 0.458 27%.5 Lh21.0 0.251
59.5 97.1 0.441 302.7 Lh71.2 0.114
62.5 102.1 0.469 309.7 482.8 0.138
65.5 107.0 0.43L

 

 

QLT

 

a
Corrected for radioactive decay during the run.
179

ORNL. DWG Ti-63R1

METAL TRANSFER EXPERIMENT-MTE-2

 

 

 

 

 

1
I | I [ 1 [ T T T T T

-

—
o o
z
=z —4
<
=
o -
@
=z -
=2
z
!
e -
-
=
o
-
&
z
©
- -
Q
<
x
W

X
¥
0.1 1 l i [ L l L 1 1 ] 1
0 100 200 300 400 500 600

VOLUME OF LiCl PUMPED (LITERS)

Fig. 65. Variation of Lanthanum Inventory in the Fluoride Salt
Surge Tank with Time During Metal Transfer Experiment MTE-Z2.
1.0

0.9

o8

0.7

0.6

0.5

0.4

03

FRACTION OF NEODYMIUM REMAINING (CORRECTED FOR DECAY)

Are T [ T [ T I T I T

l o T
$

;:/‘
o1

7

180

ORNL DWG 7i-64R1

METAL TRANSFER EXPERIMENT-MTE-2

 

 

 

 

1 I L l 1 I 1 I 1

 

0.2

o 100 200 300 400 500

VOLUME OF LiCl PUMPED (LITERS)

Fig. 66. Variation of the Neodymium Inventory in the Fluoride Salt

Surge Tank with Time During Metal Transfer Experiment MTE-2.
Table 46.

181

Variation of Lanthanum Distribution Coefficient
with Time During Metal Transfer Experiment MTE-2

 

 

Volume of Distribution Distribution

Pumping LiCl Coefficient at Coefficient at

Time Circulated Fluoride Salt-- LiCl--Bi~Th

(hr) (1iters) Bi-Th Interface Interface

0 0 0.045 4,54
6 9.6 0.046 2.78
12 19.6 0.042 5.97
18 29.1 0.043 2.04
27 44,0 0.052 -
30 48.8 0.039 3.67
36 59.0 0.034 4.49
44,5 73.2 0.044 9.28
47 .5 77.9 0.038 2,18
53.5 87.6 0.039 1.84
59.5 97.1 0.023 1.17
65.5 107.0 0.034 0.93
94.2 153.2 0.056 2.07
124.2 195.3 0.055 4.17
124.2 195.3 0.078 3.51
137.6 211.2 0.045 1.88
150.1 226.5 0.045 2,00
156.1 233.4 0.055 2.45
169.1 250.7 0.055 0.93
185.3 273.9 0.054 -
212.6 319.6 0.071 7.54
219.2 330.5 0.083 -
250.5 384.3 0.058 3.14
273.5 421.0 0.056 2.69
309.7 482.8 0.046 2.24
343.9 538.8 0.036 2.32
358.1 563.2 0.048 2,27
358.1 563.2 0.051 3.56

 
Table 47.

182

Variation of Neodymium Distribution Coefficient

with Time During Metal Transfer Experiment MTE-2

 

 

Volume of Distribution Distribution
Pumping LiCl Coefficient at Coefficient at
Time Circulated Fluoride Salt-—- LiCl1--Bi-Th
(hr) (liters) Bi-Th Interface Interface
0 0 0.066 2.88
3 4,7 0.055 6.98
6 9.6 0.064 4,94
9 14.7 0.062 4.56
12 19.6 0.062 4.54
15 24.5 0.064 4.83
18 29.1 0.067 4,12
21 33.9 0.078 5.10
27 44,0 0.051 3.02
30 48.8 0.062 3.39
33 54.1 0.063 2.60
36 59.0 0.063 3.72
44.5 73.2 0.052 6.00
47.5 77.9 0.058 5.73
50.5 82.7 0.056 10.59
53.5 87.6 0.059 6.50
56.5 92.3 0.058 4.85
59.5 97.1 0.057 4,17
62.5 102.1 0.054 4,89
65.5 107.0 0.058 3.49
80.0 130.4 0.069 5.96
94.2 153.2 0.063 -
101.3 164.0 0.060 -
108.2 173.8 0.071 4.52
114.9 183.5 0.071 4.42
124.2 195.3 0.063 5.76
130.7 202.3 0.083 ——
137.6 211.2 0.093 —
143.5 218.2 0.085 -
150.1 226.5 0.048 -
156.1 233.4 0.076 ——
162.1 241.2 0.082 —
169.1 250.7 0.092 -
175.9 258.5 0.070 -
185.3 273.9 0.058 ——
185.3 273.9 0.069 -
250.5 384.3 0.052 —_
250.5 384.3 0.038 -

 
183

obtained from samples taken after circulation of the LiCl had been stopped
and the phases had been allowed to approach chemical equilibrium for four
or more hours. The variation of the coefficient for the distribution of
lanthanum between the fluoride salt and the Th-Bi phase during the experi-
ment is shown in Fig. 67. Although there is some variation in the distri-
bution coefficient values, the average value (0.05) is relatively close

to the value of 0.06 that was calculated from previously reported equilib-
rium relations. The variation of the coefficient for the distribution of
lanthanum between the LiCl and the thorium-saturated bismuth during the
run is shown in Fig. 68. The average value (3.1) is somewhat higher than
the value predicted by equilibrium relations reported earlier (i.e., 0.9).
Data on the coefficient for the distribution of neodymium between the
fluoride salt and the thorium-saturated bismuth during the run are shown
in Fig. 69. The average value for the distribution coefficient (0.06)

is in good agreement with the expected value (0.062). The variation of
the coefficient for the distribution of neodymium between the LiCl and

the thorium-saturated bismuth during the experiment is shown in Fig. 70.
The average value for the distribution coefficient (4.8) is slightly
higher than the expected value of 3.5. The uncertainties in the distri-
bution coefficient values for lanthanum and neodymium during the experi-
ment result in some uncertainties of the values for f, the fractions

of the lanthanum or neodymium that are removed from the LiCl during its
passage through the cup containing the Li-Bi solution. The values of f
for lanthanum and neodymium are 0.27 and 0.21, respectively, based on the
average values for the distribution coefficients, and 0.08 and 0.16

based on the predicted values for the distribution coefficients.

The variations of the neodymium and lanthanum inventories in the
system during the experiment are shown in Fig. 71. From 70 to 100Z of
each of the rare earths initially charged to the system could be accounted
for, based on filtered samples. This indicates that essentially all of

the rare earths remained in solution during the experiment.
13k

ORNL DWG Ti-61R2

METAL TRANSFER EXPERIMENT-MTE-2

 

 

 

 

 

0.'0 v —r ¥ "— ¥ l T [ L I ¥
0.09} _
- . -
- 0.08 br ° "‘
2
W 0.07} * -
O
t " 4
w 006 ° . ]
® AVERAGE=0. o
Z 0.05 ERAGE=0.05 -
- ° o0 * -
2 °®
o ® ®
- " o o )
5 o003} -
| . ]
0.02 |- -
- ;
0.0l - .
o . | L 1 ) ] . 1 L ] .
o 100 200 300 400 500 600

VOLUME LiC! PUMPED (LITERS)

Fig. 67. Variation of the Coefficient for the Distribution of
Lanthanum Between the Fluoride Salt and the Thorium-Saturated Bismuth
During Metal Transfer Experiment MIE-2.
185

ORNL DWG T7I-59R1

METAL TRANSFER EXPERIMENT-MTE-2
Io-o T '[ LB I T ' v I L I T

 

90 - -
80 .
7.0 —

60e -

®
° AVERAGE=3. ® -

 

DISTRIBUTION COEFFICIENT
o
O
=
L

 

 

 

o 3 l 2 l 3 l i l 3 | 2
0 100 200 300 400 500 600
VOLUME OF LiC! PUMPED (LITERS)

Fig. 68. Variation of the Coefficient for the Distribution of
Lanthanum Between the LiCl and the Thorium-Saturated Bismuth During
Metal Transfer Experiment MTE-Z.
1866

ORNL DWG 7i-62R!

METAL TRANSFER EXPERIMENT-MTE-2

 

 

 

 

 

 

0.'0 I = T [ r T ™ T T
®
0.09 . )
° ®
008 * _

; . -

Z 007 « °° ® -

o AVERAGE=0.06

u °% '

w006 o -

Q .

o °
0.05 B

S - o J

5

3 004 . .

£ -

v —

@ 0.03 _ {
002 | i
(lOI[ i,

O L 1 . ] L | L i — 1 )
0 00 200 300 400 500 600

VOLUME OF LiCl PUMPED (LITERS)

Fig. 69. Variation of the Coefficient for the Distribution of
Neodymium Between the Fluoride Salt and the Thorium-Saturated Bismuth
During Metal Transfer Experiment MTE-Z2.
DISTRIBUTION COEFFICIENT

1.0 —

100 |-

9.0 |

80|

70 |e

60 |-

50

4.0 |-

 

187

ORNL DWG T7I-60R1
METAL TRANSFER EXPERIMENT-MTE-2

 

] ¥ ] T
®

o @ AVERAGE=4.8

1 A 1

 

 

Neodynium Between the LiCl and the Thorium-Saturated Bismuth During

Fig. T0O.

100 200
VOLUME OF LiCl PUMPED (LITERS)

Variation of the Coefficient for the Distribution of

Metal Transfer Experiment MTE-2.

300
188

ORNL DWG 7I-7IRI

 

 

 

 

120 T T - T T T
o. X
1I00» ,5!:: . -
"d‘.‘ Xe o o .x . X
X .o. > X . ® < X
1 e .x ° X X
80 .ox. .! . o . x _
* % .
@ X X ®
2
> 60 —
a
O
'_
4
s
Z 40 e NEODYMIUM _
F x LANTHANUM
20 .l —y
0O i ] ] l 1 1
0 100 200 300 400 500 600 700

VOLUME OF LiClI PUMPED (LITERS)

Fig. Tl. Variation of the Inventories of Neodymium and Lanthanum
During Metal Transfer Experiment MTE-2.
189

Decrease in Concentration of Lithium in the Li-Bi Solution. — During

 

the experiment, the concentration of lithium in the Li-Bi solution decreased
from an initial value of 1.8 wt %7 to 0.5 wt Z (0.35 to 0.13 mole fraction),
as summarized in Table 48 and shown in Fig. 72. A decrease of about 0.2

wt Z in lithium concentration was expected as the result of extraction of
the rare earths into the Li-Bi solution; however, the observed decrease

was about six times this value. The reason for the decrease has not been
determined at this time; it may have resulted from a slight solubility of
lithium or a lithium-bismuth intermetallic in the LiCl or from transfer of
lithium ions through the LiCl due to the difference in emf between the two
bismuth phases in the system. The decrease in lithium concentration appears
to have occurred only while LiCl was being circulated through the Li-Bi

cup since no decrease was noted during long periods in which no circulation

occurred (e.g., weekends).

Distribution of Radium Between the Salt and Metal Phases. — Radium
was present in the system as a decay product of thorium, which was a con-
stituent of the fluoride salt and of the bismuth phase (thorium-saturated).
Data on the distribution of radium in the system during the experiment are
summarized in Table 49 and shown in Fig. 73. At the end of the 76-hr
equilibration period before circulation of the LiCl was begun, about 607%
of the radium had transferred to the portion of the LiCl that was in con-
tact with the thorium-saturated bismuth; the remaining 407 was present in
the fluoride salt. No radium was present in the Li-Bi solution prior to
circulation of the LiCl, and the radium concentration in the thorium-
saturated bismuth remained below the limit of detection throughout the
run. The radium slowly transferred into the Li-Bi solution as the LiCl
was circulated through the Li-Bi cup. When operation of the gas-lift
sparge tube in the Li-Bi cup was initiated (after about 50 liters of LiCl
had been circulated), the concentration of 228Ra in the Li-Bi solution
increased abruptly and the concentration of radium in the LiCl decreased
simultaneously. After a short period of time, the concentrations of
radium throughout the system appeared to have reached equilibrium values,
with about 40% of the radium in the Li-Bi solution, 407 in the LiCl, and
20% in the fluoride salt. During the remainder of the experiment, the
radium slowly transferred out of the Li-Bi solution as the concentration

of lithium in the solution decreased.
Table 48. Variation of Lithium Concentration in the Li-Bi Solution
During Metal Transfer Experiment MTE-2

 

 

Pumping Volume of LiCl Lithium Conc. Pumping Volume of LiCl Lithium Conc.
Time Circulated in Li-Bi Time Circulated in Li-Bi
(hr) (liters) (wt %) (hr) (liters ) (wt %)

0 0 1.80 137.6 211.2 1.29
6 9.6 1.94 150.1 226.5 1.20
12 19.6 L.76 156.1 233 .4 1.23
18 29.1 1.76 169.1 250.7 1.13
24 39.0 1.81 185.3 273.9 1.08
30 48.8 1.70 212.6 319.6 1.03
i .5 73.2 1.72 219.2 330.5 1.03
47.5 77.9 1.62 250.5 38L4.3 0.95
53-5 87.6 1.52 273.5 421.0 0.89
59.5 97.1 1.63 309.7 482.8 0.83
65.5 107.0 1.56 343.9 538.8 0.80
94 .2 15%.2 1.45 358.1 563%.2 0.76

124.2 195.3 1.35 433.6 690.2 0.5

 

 

 

06T
(wt %)

LITHIUM CONCENTRATION IN Li-Bi

0.5

ORNL DWG 71-65R2

 

 

I I T 1 T |

TST

 

 

 

I ! I 1 |
100 200 300 400 500 600 700
VOLUME OF LiCl PUMPED (liters)

Fig. 72. Variation of Lithium Concentration in the Li-Bi Solution
with Time During Metal Transfer Experiment MIE-Z.
Tabie A47). Variation of the Distribution of Radium with
Time During Metal Transfer Experiment MIE-2

 

 

 

Volume of Radium Cone . Radiumn Radium Volume of Radium Conc. Radium Radium

Pumping LicCl in Fluoride Conc. Conc . Pumping LiCl in Fluoride Conc - Conc.

Time Circulated Salt in LiC1 in Li-Bi Time Circulated Salt in LicCl in Li-Bi

(hr) (liters) (cpm/g )® (cpm/g )® (cpm/g )? (hr) (liters ) (cpm/g )? (cpm/g )? (cpm/g )?
0 0 557 1565 0 143 .5 218.2 166 1028 901
3 .7 ugy 1500 103 150.1 £226.5 152 1107 96U
6 §.6 hs1 1410 216 156.1 233,14 136 1103 885
9 4.7 L57 -- 284 162.6 2hk1.2 154 1010 856
12 19.6 koo 1511 L6 175.9 258.5 97 1241 997
15 2o 348 1511 384 185.3 273%.9 115 -- 907
18 29.1 Loy 1540 508 250.5 384.,3 182.3 898 771
21 3%.9 549 1506 51% 250.5 28,3 -- 1372P --
24 39,0 -- 1361 621 250.5 384 .3 187.4 1846 97T
27 Lh.0 265 1310 617 250.5 z8l.3 - 1309P --
30 L8.8 318 1508 -- 273.5 421.0 63 1356 733
33 54.1 316 1571 692 273.5 L21.0 52 1177 700
36 59.0 275 1029 1100 302.7 471.2 -- 1103 71
%9 65.9 -- 907 1057 309.7 L82.8 -- 1176 628
bh.s 73,2 278 950 1235 %09.7 482.8 -- 1135P --
hT.5 7.9 96 966 1261 34%.9 538.8 -- 1296 684
50.5 2.7 197 1588 1184 343.9 538.8 -- 1362P --
53.5 87.6 188 824 121% 343.9 538.8 -- LL86 665
56.5 2.3 200 916 115k 343.0 538.8 -- 1h19P --
59.5 971 184 701 1151 358.1 56%.2 -- 1311 640
62.5 102.1 226 o7 1117 358.1 563.2 -- 1676P -
65.5 107-0 op 850 1160 358.1 563.2 -- 1463 696
80.0 130. 4 183 10%2 1286 358.1 563%.2 -- 15700 --
k.5 153.2 15% 101y 1197 591.1 619.1 -- 1367 51t
108.2 17%.8 e 106k 108L 351.1 615.1 -- 1318P --
114.9 183. 5 5 1186 1134 413.3 656.0 -- 1549°¢ Ls7
124.2 195.% g2 1220 1097 h13.% 656.0 -- 1739¢ --
130.7 202. % 80 1117 112% §1%.5% 656.0 -- 1424€ --
137.6 211.2 -~ 1069 1029 hio.2 665.2 -- 1378¢ 498

 

261

 

 

a
Counts per minute per gram.
b
Samples taken from Li-Bi alloy container.

“After 1 vol % fluoride salt was added to LiCl.
PERCENT OF TOTAL RADIUM

100

ORNL DWG 71-!'72RlI

 

90 |-

80 |-

 

 

1 I !

A Li-Bi
o LicCl
X FLUORIDE SALT
Negligible amount of Ra in Bi-Th

 

 

 

 

Sy o X
nl\x X X
X X x X %
oL | X i 1 1
0 100 200 300 400

VOLUME OF LiCl PUMPED (liters)

Fig. 73. Variation of the Distribution of Radium with Time in the
Salt and Metal Phases in Metal Transfer Experiment MTE-Z2.

500

€61
194

Addition of Fluoride Salt to the LiCl. — During operation of the
metal transfer process, it is possible that small quantities of fluoride
salt will be transferred to the LiCl by entrainment of salt in the bismuth
that contacts these salt streams. It has been determined previously that
the thorium distribution ratio between the LiCl and the bismuth is quite
sensitive to the concentration of fluoride in the LiCl, and hence that
the thorium--rare-earth separation factor in the process decreases markedly
as the concentration of fluoride in the LiCl is increased. Throughout
the experiment, very little fluoride was transferred to the LiCl, as
indicated in Table 50; a final fluoride concentration of 1174 ppm was
noted. Eight days before the end of the experiment, 1 vol % fluoride
salt (72-16-12 mole % LiFwBer—ThFa) was added to the LiCl in order to
simulate entrainment of fluoride salt in the circulating bismuth. The
concentrations of beryllium, thorium, and fluoride in the LiCl were deter-
mined periodically after the addition of the fuel carrier salt, as shown
in Table 51. During the 93-hr period in which the LiCl was not circulated
through the Li-Bi container, the beryllium concentration remained constant
at the initial value of 490 ppm and the thorium concentration decreased,
as expected, from the initial value of 9480 ppm to a value of 644 ppm
because of transfer of thorium into the Th-Bi solution. When circulation
of the LiCl was resumed, the beryllium concentration in the LiCl began
to decrease, probably because of reduction of Bez+ by the Li-Bi solution.
After 27 hr of LiCl cifculation, a2 beryllium concentration of 135 was
observed in the LiCl. The thorium concentration in the LiCl was 171 ppm

at this time.
195

Table 50. Variation of the Fluoride Concentration in the
LiCl and the Chloride Concentration in the Fluoride
Salt During Metal Transfer Experiment MTE-2

 

 

Volume of Chloride Fluoride
Pumping LiCl Concentration Concentration
Time Circulated in Fluoride Salt in LiCl
(hr) (liters) (ppm) (ppm)
0 0 - 77
9 14.7 175 530
56.5 92.3 121 -
101.3 164.0 - 517
273.5 421.0 392 205

413.3 656.0 — 1174

 

Table 51. Variation of Beryllium, Thorium, and Fluoride
Concentrations in the LiCl After Addition of
1 vol 7 Fuel Carrier Salt

 

 

Time After Pumping

Addition Time Fluoride Beryllium Thorium
(hr) (hr) (wt %) (ppm). (ppm)
0 0 0.982 4902 94802

19.8 0 1.49 490 7200

45.4 0 1.08 500 1100

69.7 0 1.81 320 ——

93 0 2.03 490 644
115.9 5.9 2.34 470 -
163.9 20.3 1.07 120 1400
188.8 27 .4 0.91 - 350

b b b
188.8 27 .4 1.7 135 171

 

aCalculated values based on the amount of fluoride salt added.

bSalt taken from vessel after experiment was concluded.
196

12. DEVELOPMENT OF THE METAL TRANSFER PROCESS:
DESIGN OF EXPERIMENT MTE-3

L. E. McNeese E. L. Nicholson
W. F. Schaffer, Jr. E. L. Youngblood
H. 0. Weeren

It has been found that rare earths distribute selectively into
molten LiCl from bismuth solutions containing rare earths and thorium,
and an improved rare-earth removal process based on this observation
has been devised. We are currently engaged in experiments designed to
demonstrate all phases of the improved rare-earth removal method, which
is known as the metal transfer process. In a previous engineering

46

experiment (MTE-1), we studied the removal of rare earths from single-
fluid MSBR fuel salt by this process. During the experiment, approxi-
mately 50Z of the lanthanum and 257 of the neodymium originally present
in fluoride salt were removed at about the predicted rate. Surprisingly,
however, the lanthanum and neodymium that were removed from the fluoride
salt did not accumulate in the Li-Bi solution used for removing these
materials from the LiCl. Reaction of impurities in the system with the
rare earths is believed to have caused this unexpected behavior. A
second engineering experiment (MTE-2) has been designed and put into
operation (see Sect. 1l). The design of the third engineering experi-
ment is currently under way. The third experiment (MTE-3) will use

salt and bismuth flow rates that are about 1% of the estimated flow
rates required for processing a 1000-MW(e) reactor. In the two previous
experiments, the salt and bismuth phases were only slightly agitated,
resulting in a low rate of transfer of rare earths from the fuel carrier
salt to the Li-Bi solution. In experiment MITE-3, the salt and metal
phases will be mechanically agitated in order to increase the rate of
transfer of materials between the phases. In the remainder of this
section, results from a mathematical analysis carried out for estimating
the performance of the system are presented and the experiment is briefly

described.
197

12.1 Mathematical Analysis of Metal Transfer Experiment MTE-3

A mathematical analysis of experiment MTE-3 was performed in order
to determine the approximate operating conditions for the system and to
aid in setting values for parameters such as flow rates, solution volumes,
etc. In considering the conceptual design for the experiment, it was
concluded that the salt-metal contactor should be of the two-compartmented,
stirred-interface type. As pointed out in the following section of this
report (Sect. 13), such a contactor is of interest because, with it,
adequate mass transfer rates can apparently be achieved without dispersal
of either the salt or bismuth phases. Contact of the salt and metal
streams without dispersion of the phases should considerably diminish
the problem of entrainment of bismuth in the processed fuel carrier salt
and the subsequent transfer of bismuth to the reactor, which is con-
structed of a nickel-base alloy that is subject to damage by metallic
bismuth. The bismuth in a salt-metal contactor of the type being
considered would be a near-isothermal, internally circulated phase,
which is a desirable condition. Also, it is believed that a processing
system employing this type of contactor may be more easily fabricated
from molybdenum and graphite than one employing packed columns. The
major equipment items considered in the mathematical analysis are shown
in Fig. 74. Fuel carrier salt (72-16-12 mole Z% LiFmBer-ThF4) con-
taining rare-earth fluorides would be circulated between the fluoride
salt surge tank and the side of the salt-metal contactor containing
fluoride salt. Lithium fluoride would be circulated between the other
side of the contactor and a vessel containing a Li-Bi solution having
a lithium concentration of about 5 at. %Z. During operation of the
system, the rare earths would be extracted from the fluoride salt and

would accumulate in the Li-Bi solution.

In carrying out the mathematical analysis, it was assumed that the
salt and bismuth phases in the salt-metal contactor remain at equilibrium
at all times. Although this condition will clearly not be met during
operation of the experiment, such an assumption will allow a useful
representation of the salt-metal contactor. The concentrations of a
rare earth in the salt and bismuth phases in the salt-metal contactor

at time t are then related by the following expressions:
198

ORNL DWG. 73-44

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

— F.X o Fe ) X¢ 7
FLUORIDE —->—%» FLUORIDE e o LiCl
AND < A I
e . o
RARE o
EARTHS
Li-Bi
Th-Bi
FLUORIDE SALT SALT - METAL RARE -EARTH
SURGE TANK CONTACTOR ACCUMUL ATION
VESSEL

Fig. 74. Equipment Representation and Nomenclature Used for Mathe-
matical Analysis of Experiment MTE-3.
199

D, =— ’ (52)

where
D, = rare-earth distribution ratio between fluoride salt and
bismuth containing reductant at time t,
X,, = equilibrium concentration of rare earth in bismuth in con-
tactor at time t, mole fraction,
X, = equilibrium concentration of rare earth in fluoride salt
in contactor at time t, mole fraction,

and

D =— ’ (53)

where
D = rare-earth distribution ratio between LiCl and bismuth
containing reductant at time t,
Xx = equilibrium concentration of rare earth in LiCl in con-

tactor at time t, mole fraction.

Combination of Eqs. (52) and (53) yields the following expression:

o

c
Xp = oo X, , (54)

=

which relates the concentration of rare earth in the fluoride salt to
that in the LiCl in the contactor at time t. The rate at which the

rare earth is transferred through the contactor is then given by the

expression
rate = Fc X, (1 — £) . (55)
where
Fc = flow rate of LiCl between the contactor and the vessel con-
taining the Li-Bi solution, moles/sec,
f = fraction of rare earth removed from LiCl during its pas-

sage through the vessel containing the Li-Bi solution.
200

The rate of transfer is also equal to the instantaneous rate at which
rare earths are removed from the flunride salt surge tank, which is

given by the expression

rate = F(x —-xF) , (56)
where
F = flow rate of fluoride salt between the contactor and the
fluoride salt surge tank, moles/sec,
X = concentration of rare earth in fluoride salt surge tank at

time t, mole fraction.

If it is assumed that the rate of accumulation of rare earth in the
salt and metal phases in the salt-metal contactor is negligible, Egs.
(55) and (56) can be equated. Substituting Eq. (54) in the resulting
relation and solving for Xpo the concentration of rare earth in the

fluoride salt in the contactor at time t, yields the following expression:

(57)

 

X
F F D

¢c F
1 + T

D
c

11— £

A material balance on the rare earth in the fluoride salt surge tank

yields the following relation:

dx _ _
-V ac F(x XF) ’ (58)

where

V = volume of fluoride salt in the surge tank, moles.

Combination of Eqs. (57) and (58) yields the following differential

 

equation:
F D
F
X =5 -0 t
dx _ F c
x v F_D_ dt (59)
. l+"F—-—D-"(l—f) 0

o Cc
201

where
X, = initial concentration of rare-earth fluoride in fluoride salt
surge tank, mole fraction.
This equation can be integrated between the indicated limits to yield
the following expression for the concentration of rare-earth fluoride

in the fluoride salt surge tank at time t:

 

t
—=e (60)
0
where
F D
¢ F ,
£ £ -1
_F F D
Vv FC DF
1+ fr-fiz'(l — f)

If it assumed that the flow rates for the fluoride salt and the LiCl are,
in each case, 1% of the expected flow rates required for removing rare
earths from a 1000-MW(e) MSBR (see Sect. 3) and if the concentration of
reductant in the bismuth in the salt-metal contactor corresponds to a
thorium concentration equal to 907 of the solubility of thorium in
bismuth at the operating temperature of 640°C, the following values are

obtained for the parameters in Eq. (60) for the transfer of neodymium:

DF = 0.0564 ,
Dc = 3.5 , and
2251

F

If it assumed that 50% of the neodymium is extracted from the LiCl
as it passes through the vessel containing the Li-Bi solution, the

expression for the quantity k can be written as:

« = 0.168 % : (61)

If it is desired to reduce the concentration of neodymium in the fluoride
salt surge tank by 20%Z during a 24-~hr operating period, the quantity «
must have the value 0.223, which then sets the value for the quantity V/F
202

at 0.753 day. The resulting volume of the fluoride salt in the surge
tank is 36.1 liters. Table 52 summarizes the resulting design parameters
which are dependent on the following assumptions: (1) the salt and
metal phases in the salt-metal contactor remain at equilibrium at all
times, (2) salt and metal flow rates equal to 1% of the expected rates
required for processing a 1000-MW(e) MSBR are used, (3) neodymium is

the transferring rare earth, and (4) the concentration of neodymium
fluoride in the fluoride salt surge tank is to be reduced by 20% during

a 24-hr operating period.

Table 52. Calculated Values of the Design
Parameters for Experiment MTE-3

 

 

Quantity ' Value
DF 0.0564
Dc 3.5
3
33 cm” /min
. 1.25 liters/min
v 36.1 liters

 

12.2 Preliminary Design of Metal Transfer Experiment MTE-3

The planned experiment, shown schematically in Fig. 75, will use
mechanical agitators to promote efficient contact of the salt and metal
phases. Fuel carrier salt containing rare-earth fluorides will be circu-
lated between one side of the salt-metal contactor and a fluoride salt
reservoir. Lithium chloride containing rare-earth chlorides will be cir-
culated between the other side of the salt-metal contactor and a rare-
earth stripper, where the rare earths will be extracted into a Li-Bi

solution.

The experiment will use approximately 35 liters of MSBR fuel carrier
salt, 6 liters of Th-Bi solution, 6 liters of LiCl, and about 5 liters

of Li-Bi solution having an initial lithium content of about 5 at. Z.
ORNL -DWG-71-147-RI

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

AGITATORS
VENT LEVEL
ELECTRODES LEVEL
/ELECTRODES
FLUORIDE
=~ SALT PUMP VENT
i 1 ] 1 ]
L J L ] L J
ARGON
SUPPLY —
ARGON
SUPPLY
33 Cm3/ min
I.25
— Zmi
3 — \\§/</5/< (5 :::;m e Akl
L "'///'5\ _/‘F\-’ A 7]
LiCl
72-16-12 mole % Bi-Th AN [ Li-Bi
LiF-BeFp- ThF,
FLUORIDE SALT- METAL RARE EARTH
SALT CONTACTOR STRIPPER

RESERVOIR

Fig. 75. Flow Diagram for Metal Transfer Experiment MTE-3.

02
204

The system will require three process vessels, each of which will be
made of carbon steel. The largest vessel will be the fluoride salt
reservoir, which will contain approximately 32 liters of fuel carrier
salt. The remaining 3 liters of fluoride salt will be contained in the
salt-metal contactor. The fluoride salt will be recirculated continuously
from the reservoir to the contactor at the rate of about 33 cm3/min by

a pump similar to the one used in the second metal transfer experiment.48
The salt-metal contactor will be a 10-in.-diam, two-compartmented vessel
having a mechanical agitator in each compartment. The agitator will
consist of two paddles mounted on a common shaft with a paddle operating
in each of the salt and metal phases in a manner such that neither of

the phases is dispersed. Each paddle will consist of 4 blades that are
canted at a 45° angle in a manner such that the salt and bismuth will be
forced to flow toward the salt-metal interface in the area around the
agitator shaft. The Th-Bi solution will be captive in the salt-metal
contactor and will form a seal to isolate the fluoride salt from the LiCl.
The Th-Bi solution will be recirculated between the two compartments of
the salt-metal contactor by utilizing the pumping capability of the
agitators. The third vessel, which will be similar in design to one
compartment of the salt-metal contactor, will contain the Li-Bi solution.
The LiCl will be circulated between the salt-metal contactor and the
rare—earth stripper at the rate of about 1.25 liters/min by varying the
gas pressure above the LiCl in the rare-earth stripper. All of the
vessels will be operated at 650°C. The carbon steel used for fabricating
the system will be protected from external air oxidation by the use of

a nickel aluminide coating.
205

13. DEVELOPMENT OF MECHANICALLY AGITATED
SALT-METAL CONTACTORS

H. 0. Weeren L. E. McNeese
J. S. Watson

A program has been initiated for the development of mechanically
agitated salt-metal contactors as an alternative to packed columns pres-
ently under consideration for MSBR processing systems. This type of
contactor is of particular interest for the metal transfer process since
designs can be envisioned in which the bismuth phase would be a near-
isothermal, internally recirculated, captive phase. It is believed that
such designs will require a less highly developed technology for molyb-
denum fabrication than would a counterpart system based on packed columns.
During this report period, we carried out preliminary tests on the hydro-
dynamics of mechanically agitated salt-metal contactors using water and
mercury and reviewed information in the literature relative to mass trans-
fer rates in contactors in which the phases are not dispersed. Results

of work in these areas are summarized in the remainder of this section.

13.1 Hydrodynamic Studies

Studies of the hydrodynamics of mechanically agitated salt-metal
contactors carried out thus far have been concerned primarily with the
selection of a contactor design for experiment MTE-3 (discussed in Sect.
12), which will have salt and bismuth flow rates that are about 1% of
the estimated rates for processing a 1000-MW(e ) MSBR. Several scouting
tests were carried out with water and mercury in 4- to T-in.-diam ves-
sels, and a mockup of the contactor proposed for experiment MTE-3 was
built for additional study with mercury and water. The mockup consists
of an 8-in.-diam vessel having a central partition that extends to
within 1/2 in. of the bottom of the vessel. The first tests with the
mockup were made with a flat, four-bladed paddle located in the mercury-
water interface. The paddle, 1.5 in. in diameter, was located inside
a 3-in.-diam, 3-in.-high shroud containing four O.3-in.-wide, 3-in.-long

baffles. The design of the proposed contactor was chosen to maximize
206

the extent of dispersion of the mercury in the water, thereby maximizing
the interfacial area between the two phases. The mockup of the MTE-3
system was tested with agitator speeds up to 1600 rpm. It was found
that a stable dispersion of very small mercury droplets was frequently
formed under these operating conditions. Also, a dispersion of water
droplets was formed in the mercury at all but the lowest mixer speeds,
and these droplets were pumped from one chamber of the contactor to the
other. Such pumping of molten salt solutions cannot be tolerated in

the metal transfer process since it would lead to mixing of the chloride
and fluoride salts. The tendency of the salt-bismuth system to form
emulsions may be quite different from that of the water-mercury system;
however, it was concluded that the contactor should operate under condi-

tions that minimize the likelihood of emulsion formation.

The contactor design that has the greatest potential for achieving
effective mass transfer rates with minimum dispersion of the phases

ko

appears to be the Lewis cell. This contactor has a paddle in each

phase, located well away from the interface as shown in Fig. T76. Each
agitator is operated in a manner such that the phases are agitated as
vigorously as possible without actually dispersing one phase in the

other. The contactor cell mockup described previously was modified in
order to study the hydrodynamics of this type of contactor. The shroud

was removed, and paddles having a range of diameters were tested at

several positions in baffled cells and also in cells containing no baffles.
A significant amount of entrainment of water via circulation of the mercury
between the compartments was noted at agitator speeds above 180 rpm when
5-in.-diam, four-bladed vertical paddles were used. The agitator speed

at which entrainment began to occur could be increased slightly (to about
220 rpm) when baffles were used in the cell, when the lower paddle was
located within 0.25 in. of the bottom of the cell, or when the depth of

the mercury in the cell was increased by about 0.5 in. When smaller-
diameter agitators were used, the speed at which entrainment was noted
increased; however, the Reynolds number based on the paddle diameter was

almost the same in each case.
207

ORNL DWG 71-9544Rl

 

  
 
 
   
   
 
 

~LiCl
FLUORIDE
LT

>

BETWE

_—PA RTngN
CELL HALVES

Bssmum_

GAP FQR P
BISMUTH FLOW

Fig. 76. Mechanically Agitated Salt-Metal Contactor Proposed for
Use with Metal Transfer Experiment MTE-3.
208

Investigation showed that use of a 3-in.-diam, four-bladed paddle
having blades canted at a 45° angle did not result in entrainment of
water until an agitator speed of 400 rpm was reached. It is believed
that the flow patterns in the water and mercury that are created by the
use of canted blades inhibit the formation of a vortex around the agita-
tor shaft and thereby reduce the extent of dispersion of water in the

mercury phase.

13.2 Survey of Literature Relative to Mechanically Agitated,
Nondispersing Salt-Metal Contactors

Various investigations have been carried out for determining the
rate of mass transfer between aqueous and organic phases in mechanically
agitated contactors in which the phases are not dispersed. Lewisa9 has
investigated mass transfer rates in six solvent-water systems and two
nonaqueous systems. The solvents used were aniline, ethyl acetate, ethyl
acetoacetate, ethyl formate, furfural, and isobutanol; formic acid--benzene
and aniline-cyclohexane were the nonaqueous systems. A sketch of the
apparatus used, with reported dimensions, is given in Fig. 77. The speed
and direction of rotation of the agitators were controlled individually;
most of the data were taken with the agitators contrarotating, but it was
stated that the relative direction of rotation made little difference.
Lewis correlated his data by the expression:
1.65

60 k

1 6

 

- 2
y = 6.76 x 10 Re, + Re2 EI + 1, (62)

1

 

 

where
k1 = individual mass transfer coefficient, cm/sec,

v, = kinematic viscosity (u/p), Cmg/sec,

|

Re = ndg/v, dimensionless,

. = viscosity, poises,

o = density, g/cmj,

d = agitator diameter, cm, and
n = agitator speed, rps.
209

ORNL DWG 73-36

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

LEWIS
300 mA |
n 4in
GORDON STIRRER BARS 0.48cm in DIAMETER
8
SHERWOOD
15.2¢cm
| 12 STIRRER BARS 3in. LONG
OLANDER —t
- 3l:m
6.2c¢cm —
-
12
6in
I |_GRID
PROCHAZKA
. PADDLE DIAMETER 4.5cm
BULICKA  '¢™ =t =
L1
10.7em

 

 

 

Fig. 77. Schematic Diagrams of Equipment Used By Various Investi-
gators for Measuring Mass Transfer Coefficients in Stirred Interface
Contactors.
210

The subscripts 1 and 2 refer to phases 1 and 2, respectively. In a sub-
sequent paper, Lewis50 reported results for the rate of transfer of a
third component between the two phases of several solvent-water systems.
He stated that the relation given by Eq. (62) also correlated the data
on the rate of transfer of the third component.

51

McManamey” correlated Lewis' results and his own data (obtained in

a cell similar to that used by Lewis ) by the expression:

60 k
1 = 0.1 cm_1

 

 

 

v -0.3% )0_9
vy D 1

 

where
D = diffusivity of the transferable material in the phase indicated,
cmg/sec.

52

Mayers correlated Lewis' results and his own data (obtained in a

cell similar to that used by Lewis ) by the expression:

 

k,d M) 1-9 o) -2k 1/2(\;1 5/6
51—: 0.00316 J;} (0.6 + “—1_) (Re1 Reg) 5; . (6h4)

53

Gordon and Sherwood determined individual mass transfer coeffi-
cients in the system isobutanol-water with and without a third solute.
A sketch of their apparatus is given in Fig. T77. The agitators in each

phase were mounted on a common shaft. No correlation was suggested.

L .
Olander5 determined mass transfer coefficient values in several
solvent-water systems (many of which were identical to those used by
Lewis) in the apparatus shown in Fig. T7. The agitators for each phase

were mounted on a common shaft. The suggested correlation was:

1

)o.lm 0.67
1

= 0.046 (%) , (65)

 

where

w = agitator speed, radians/sec.
211

55

Prochazka and Bulicka determined mass transfer coefficient values
in water--ethyl acetate and water-isobutanol systems. A sketch of their
apparatus is shown in Fig. 77. The agitators were controlled separately,
and the agitator blades were canted at an angle of 45°. The suggested

correlation for their mass transfer coefficient data was:

e 1T

(Re, ) — . (66)

P
1 + (—L)
Po

The magnitudes of the individual mass transfer coefficients obtained

=~
o ¥

—
<

 

g

o

 

 

|

by most of the investigators were approximately the same -- from about
0.007 cm/sec for ethyl acetate--water systems to about 0.001 cm/sec for
isobutanol -water systems. The values for the individual mass transfer
coefficients reported by Prochazka and Bulicka, however, were two orders
of magnitude greater -- from about 1.1 cm/sec for ethyl acetate --water
systems to about 0.1 cm/sec for isobutanol-water systems. It was later
found that a decimal point has been misplaced in their literature arti-

56

cle. After the appropriate correction had been made, the reported
results were found to be consistent with the data of other investiga-

tors.

Although the magnitudes of the individual mass transfer coeffi-
cients obtained by different investigators were about the same, the
reported dependence of the mass transfer coefficient on the agitator
speed varied widely. Part of this difference is apparently due to the
different relations used to correlate the data, and part is due to a
different degree of interfacial turbulence created in the various pieces
of equipment at a given agitator speed. The scatter in the reported
data is shown in Figs. 78-80. In Fig. 78 the data from all the investi-
gators are plotted in the manner suggested by Lewis; in Fig. T9 the
data are plotted according to the correlation suggested by QOlander; and

in Fig. 80 the data are plotted as suggested by Prochazka and Bulicka.
ORNL DWG 73-67

 

  
   

1000 T T T TTTT] T T T 7T77H T T T T 7773

700 »

500 B

300 _

zoor- o
Et. Ac.-Water

   
  

>PROCHAZKA and BULICKA

;o4

1 4141

1 1

1

Et. Ac.- OLANDER

2 30 LEWIS
.a:lo:. '_ \

1l

  
  
  
   

 

 

 

20
Water - Isob
TE 3
— PROCHAZKA -
85 and -
— BULICKA -
3 »
2l _
Water -1sob.
' + Bi-LiCl ]
~L .
_ LiCI-Bi \\ Bi-F
; F-8i \\\’ l ’/
05 | 1ot J Lot || o1
1000 3000 5000 10000 30000 50000 100000 300000 1000000

[}
Rcl + Rea r""l

Fig. 78. Correlation of Data on Mass Transfer Coefficients in Stirred-
Interface Contactors in the Manner Suggested by Lewis. The operating con-
ditions that will be obtained at the various interfaces in metal transfer
experiment MTE-3 with an agitator speed of 3 rps are shown on the abscissa.
The notation F-Bi denotes transfer from fluoride salt to a bismuth phase;
LiCl-Bi, transfer from molten LiCl to a bismuth phase; Bi-LiCl, transfer
from a bismuth phase to molten LiCl; and Bi-F, transfer from a bismuth
phase to fluoride salt.
)0.44

213

 

    
 

 

 

 

ORNL DWG 73-66
100 T ! T T T 1 1 1 1 1
T
50
30
20r a
LEWIS: Reg fural \
10}~ PROCHAZKA and BULICKA
= 4750
2L a
5 =2370
™ OLANDER
3r 8 8
2k °
Bi-F
| L L a1 3 | | L 11 I 1
100 300 500 1000 3000 5000 10000 30000 50000
w/v
Fig. 79. Correlation of Data on Mass Transfer Coefficients in

Stirred Interface Contactors in the Manner Suggested by Olander. The
operating conditions that will be obtained at the various interfaces
in metal transfer experiment MTE-3 with an agitator speed of 3 rps are
shown on the abscissa. The notation F-Bi denotes transfer from fluo-
ride salt to a bismuth phase; LiCl-Bi, transfer from molten LiCl to a
bismutl phase; Bi-LiCl, transfer from a bismuth phase to molten LiCl;
and Bi-F, transfer from a bismuth phase to fluoride salt.
21k

ORNL DWG 73-65

 

100 1 ¥ A ] | 1 T T I I T T T T I
701 -fi
50~ -J
3CH Et. Ac.- Water —
zor —

 

   

Et. Ac.-Water

&~ LEWIS, Water-Furfural

 

 

7 .
5 //. -
\Waur- [sob.
L . _
2_ —
Water - Isob.
L _
F Salt-Bi LiCl- B Bi-LiCl Bi-F
100 300 500 1000 3000 5000 10000 30000 50000
w
2§
Fig. 80. Correlation of Data on Mass Transfer Coefficients in

Stirred Interface Contactors in the Manner Suggested by Prochazka and
Bulicka.
interfaces in metal transfer experiment MTE-3 with an agitator speed of
3 rps are shown on the abscissa. The notation F-Bi denotes transfer
from fluoride salt to a bismuth phase; LiCl-Bi, transfer from molten
LiCl to a bismuth phase; Bi-LiCl, transfer from a bismuth phase to
molten LiCl; and Bi-F, transfer from a bismuth phase to fluoride salt.

The operating conditions that will be obtained at the various
215

Examination of the three methods for correlating the data shows
that the Olander correlation has serious shortcomings. This correla-
tion assumes that the individual mass transfer coefficient in one phase
is independent of the level of turbulence in the second phase. How-
ever, data of Lewis and data of Prochazka and Bulicka show that the
individual mass transfer coefficient in one phase is definitely depen-
dent on the level of turbulence in the other phase. The Prochazka-
Bulicka relationship does not seem to adequately correlate data from
different systems; for instance, the data for water-isobutanol and
isobutanol-water do not fall on the same line. Of the three correla-

tions, the Lewis correlation seems to represent the data best.

The operating conditions that will be obtained with an agitator
speed of % rps for the phases to be used in experiment MTE-3 are indi-
cated along the abscissas of the figures for the three correlations.

The values for the Reynolds-number grouping lie outside the range of

the data obtained by Lewis but inside the range of the data obtained by
Prochazka and Bulicka. The Prochazka-Bulicka correlation indicates a
probable value for the individual mass transfer coefficient in the
fluoride salt of about 0.0015 cm/sec; extrapolation of the Lewis corre-
lation indicates a probable value of 0.05 cm/sec. Estimates of the
individual mass transfer coefficients from the MTE-2 experiment indicated
a coefficient of about 0.001 cm/sec under conditions where practically
no agitation oécurred. The value obtained from the Lewis correlation
seems much more likely to reflect operating conditions that will be used

in experiment MTE-3.
216

14. HYDRODYNAMICS OF PACKED-COLUMN OPERATION
WITH HIGH-DENSITY FLUIDS

J. S. Watson L. E. McNeese

The hydrodynamics of packed-column operation with fluids having
high densities and a large density difference is being studied in order
to evaluate and to design countercurrent contactors for use in MSBR
processing systems based on reductive extraction. Mercury and water
are being used to simulate bismuth and molten salt in these studies.
Earlier experiments57 carried out with 1/8- and 1/L-in. solid cylinders
and with 3/16- and 1/4-in. Raschig rings in a 1~in.-ID column demon-
strated that a transition in mode of flow of the dispersed phase occurs
between the packing sizes of 3/16 and 1/4 in. This transition appeared
to be a function of the packing size only and was not related to packing
shape (solid cylinders or Raschig rings ). With the larger packing the
mercury was dispersed into small droplets, which produced a large inter-
facial area. With the smaller packing, the mercury flowed down the
column in continuous channels and resulted in a much smaller interfacial
area. A large interfacial area (and hence large-sized packing) is
desired in order to obtain high mass transfer rates between the salt
and metal phases. A 2-in.-diam column was installed in the experimental
system in order to carry out studies with packing of larger sizes, and
measurements were made of flooding rates, dispersed-phase holdup, and
pressure drop Wéth 1/4-in. solid cylinders and with 3/8- and 1/2-in.
>

Raschig rings. The results of these studies have shown that the dis-

persed-phase holdup and the column throughputs at flooding can be corre-

58

lated on the basis of a constant superficial slip velocity in the

following manner:

Vc Vd
Tx "% " Vs> (67)
1/2 1/2 1/2
Ve, Vo =V o (68)
217

where
Vc = superficial velocity of continuous phase, ft/hr,
Vy = superficial velocity of dispersed phase, ft/hr,
v, = superficial slip velocity, ft/hr, and
X = dispersed-phase holdup.

The subscript f denotes conditions at flooding.

These relations were previously58 extended to cover dispersed-phase
holdup and throughput at flooding with salt-bismuth systems by assuming
that, for a given packing size, the slip velocity was (1) independent of
the viscosity of the continuous phase, (2) proportional to the difference
in the densities of the phases, and (3) proportional to the packing void
fraction. Although the resulting relations predicted flooding rates that
were in excellent agreement with flooding rates measured with molten salt
and bismuth, it was realized that the agreement did not constitute veri-
fication of the assumed effects of the continuous-phase viscosity and the

difference in the densities of the phases.

During this reporting period, data showing the dependence of slip
velocity on continuous-phase viscosity were obtained by a group of

students from the MIT Practice School.59

A 2-in.-diam, 24-in.-long
column packed with 3/8-in. Teflon Raschig rings, which were not wetted
by either the continuous phase or the dispersed phase, was used in the
study. Results of this study and the development of an improved rela-
tion for predicting packed column performance during the countercurrent
flow of molten salt and bismuth are discussed later in this section.

Studies for evaluating the effect of wetting of the packing by the

nominally dispersed phase were also carried out.

14.1 Equipment and Experimental Technique

The experimental facility used in the present study has been

58

described previously. However, the system (shown in Fig. 81) was
modified for the present studies so that water or water-glycerin solu-
tions could be circulated through the column at constant temperature
by installation of a heat exchanger and aqueous-phase surge tank.

During the study in which the effect of the viscosity of the continuous
218

ORNL DWG 72-3243

 

 

 

PACKED
DISENGAGING. 7,
SECTION ///-
' BALL
VALVE

 

 

NN

GLASS COLUMN
2%-in. LONG x t~in. L.D.
24-in. LCNG x 2-in. 1.D.

i

 

 

 

 

 

 

BALL.
VALVE|

 

 

1
I
1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

|

P

! :

.

| —

| =T

J e s
WATER PUMP MERCURY PUMP
(CENTRIFUGAL) (POSITIVE DISPLACEMENT)

Fig. 81. Flow Diagram of Equipment Used for Determining Dispersed-
Phase Holdup, Flooding Rate, and Pressure Drop in a Packed Column During
the Countercurrent Flow of Mercury and Aqueous Solutions.
219

phase on slip velocity was determined, the column consisted of a 2-in.-
ID, 24-in.-long glass tube that was packed with 5/8-in. polyethylene
Raschig rings having a void fraction of 0.66. During subsequent tests
for determining the effect of wetting of the packing by the nominally
dispersed phase, the column was packed with 3/8-in. copper Raschig rings
(void fraction, 0.81) that had been treated with nitric acid in order to
promote wetting of the packing by the mercury. A 6-in.-long section
above the column was packed with O.5-in. Raschig rings in order to dis-
tribute the mercury uniformly over the column cross section. Ball
valves having internal diameters equal to that of the column were pro-
vided above and below the column for use during measurements of dispersed-
phase holdup. The holdup measurements were effected by simultaneously
closing both ball valves and subsequently draining the mercury into a
graduated cylinder. It was found that approximately 35 ml of mercury
remained in the column during the draining operation. The mercury flow
rate was measured during the studies by closing the lower ball valve

and determining the time required for filling a 10-in.-long section of

the column which had been calibrated.

During the flooding measurements, the water flow rate was increased
incrementally with a fixed mercury flow rate until flooding was observed.
Steady-state conditions were normally reached within 30 min after each
adjustment of the aqueous-phase flow rate. The column was considered
to be flooded when it became impossible to maintain fixed flow rates
without a constantly increasing pressure drop. With the part}cular
equipment used in this experiment, the continuous—-phase flow rate and
the pressure drop across the column fluctuated widely under "flooding"
conditions. Mercury could be seen accumulating in the column until the
water flow was essentially blocked. Mercury would then drain from the
column, and the water flow rate would return to the desired value. This

behavior was repeated in a cyclic manner.

14.2 Results

Data were obtained on dispersed-pfiase holdup, column pressure drop,
and flooding, using viscosities of 1, 7.5, and 15 c¢P for the continuous

phase. Tables 5%-55 and Figs. 82-84 present data on holdup and pressure
220

Table 53. Experimentally Determined Values for Dispersed-Phase
Holdup and Pressure Drop During Countercurrent Flow of
Mercury and Water Through a 2-in.-diam Column
Packed with 5/8-in. Teflon Raschig Rings

Continuous-phase viscosity, 1 cP

 

 

 

Superficial Velocity (ft/hr) Dispersed- Pressure
Dispersed Continuous Phase Drop a

Phase Phase Holdup (mm Hg-mm H,0)

0 36.3 -- ~0

0 50.3 -~ 1

0 65.54 -- 1

0 1hk -- 2

0 206 -- 3

0 272 -- L

0 340 -- 6

71 0 0.0506 18

71 67 0.0678 18

71 1hbk 0.0826 16
71 205.7 0.0838 19

71 271 0.0666 24

71 339 0.0801 5
145 0 0.1%6 35
145 14l 0.183 26
145 205.7 0.160 39
145 271 0.17h 50
145 339 0.183 52
188 0 0.183 L9
188 14k 0.206 39
188 206 0.217 52
188 271 0.2Lh7 79
188 339 0.271 102

 

Millimeters of liquid having a density of 12.6 g/cma.

drop for a range of values of the continuous-phase superficial velocity.
The points at which flooding was observed are indicated in the figures.
Table 56 and Fig. 85 give values of holdup and pressure drop obtained
during operation with a continuous-phase viscosity of 1 cP for the case
in which the packing was wet by the dispersed phase. Data on flooding

are presented in Figs. 86-89, where the square root of the dispersed-phase
Table 54.

221

Experimentally Determined Values for Dispersed-Phase
Holdup and Pressure Drop During Countercurrent Flow of
Mercury and a Water-Glycerin Solution Through a

2-in.-diam Column Packed with 3/8-in.

Continuous-phase viscosity, T.5 cP

Teflon Raschig Rings

 

Superficial Velocity {ft/hr)

 

 

Dispersed- Pressure
Dispersed Continuous Phase Drop
Phase Phase Holdup (mm Hg-mm HQO)a
0 13.9 - 1
0 2l .1 -- 1.5
0 33.9 -- 2
0 81.1 = 3
0 136.2 -- L
0 198.4 -- 57
0 259.5 -- - 8-13
0 303 -- 11-14
0 373 ~- 15-18
0 431.5 -- 22-25
0 L84 - 31
0 568 -- 3T
67 0 0.0716 28
67 81.1 0.0900 33
67 136.2 0.0998 4o
67 198.4 0.0839 48
67 172 0.0962 L3
67 218 0.122 61
67 259 0.131 70
67 28k 0.128 60
107 0 0.154 26
107 81.1 0.154 L
107 3%.9 0.148 27
107 136.2 0.168 33.5
107 198.4 0.143 63
107 160 0.138 50
107 150 0.127 38
107 230 0.181 67
12k 0 0.154 40
12k 33.9 0.17h 61
124 81.1 0.181 50
124 136.2 0.192 65
124 172 0.211 72

 

Millimeters of liquid having a density of 12.6 g/cm3.
222

Table 55. Experimentally Determined Values for Dispersed-Phase
Holdup and Pressure Drop During Countercurrent Flow of
Mercury and a Water-Glycerin Solution Through a
2-in.-diam Column Packed with 5/8-in.

Teflon Raschig Rings

Continuous-phase viscosity, 15 cP

 

 

 

Superficial Velocity (ft/hr) Dispersed- Pressure
Dispersed Continuous Phase Drop

Phase Phase Holdup (mm Hg~-mm H20)a

0 22.4 -- ~0

0 51.2 -- 1

0 116.5 -- 9

0 167 -- 12

0 216.2 -- 15

0 270 -- 19

0 330 -- 2l

G 397 -- 28
71 0 0.107 5
71 100 0.109 9.5
71 200 0.141 12
71 220 0.17h 62
s 0 0.131 0
85 22.4 0.157 --
85 51.2 0.143 --
85 110.5 0.158 --
85 167 0.181 --
85 196 0.220 --
85 228 0.25% --
126 0 0.204 31
126 22,4 0.195 3%
126 51.2 0.197 36
126 110.5 0.217 57

 

®Millimeters of liquid having a density of 12.6 g/cm5.
223

ORNL DWG 73-47

 

 

 

 

 

 

' { T [ I I
W -
L 0.3 L
{ o’ -A
{ & - ® .‘. e
1 AT
2 _ A

8 O 02 ‘ .......................... . A.-’/—'
" =
x©
wXx
Q
0 0.1} a -
o T""‘"""‘"‘ ______ -._ —————————— ._ ----- o

0 —+ : t t t : +=

Dispersed Phase Velocity .

8o} {ft/br) A 7
a 8 - 0
o _———_—g--
T eof . !
o —— 145

E | s
W E ; Ao 188
o
8 o
& E
______ ——
—— T T
50 100 150 200 250 300 350
CONTINUQUS PHASE VELOCITY (ft/hr)
Fig. 82. Variation of Dispersed-Phase Holdup and Column Pressure

Drop During the Countercurrent Flow of Mercury and Water in a 2-in.-ID
2h-in.-long Column Packed with 5/8-in. Teflon Raschig Rings.

 

 

 

 

2
ool

ORNL DWG. 73 -35RI

 

 

 

 

 

 

 

 

5
_DJ T T B v 0 ¥ L
O
T
tJ O 3 = =
2 = FLOCDED
IS 1 ruonoes
> ]
oOT
w
n
@
u—l —
a
v
o
0
FLOODED
80| | FLOODED FLOTDED ~
/
7/
60 - 1
3o DISPERSED PHASE
5 N VELOCITY, Vg4
o T | (£1/hr)
W E 40 e o T
T -—--8-——- 67
A o ——e—— 107
Ll X 20 .‘ ....... |24 =
x e -
a —p -
E e
o - = v L 1 1 1
0 50 100 150 200 250 300 350

CONTINUOUS PHASE VELOCITY, V¢ (ft/hr)

Fiz. 83. Variation of Dispersed-Phase Holdup and Column Pressure
Drop During the Countercurrent Flow of Mercury and a Water-Glycerin
Solution in a 2-in.-ID, 24-in.-long Column Packed with 5/8—in. Teflon
Raschig Rings. The viscosity of the continuous (aqueous ) phase was

7.5 cP.
ORNL DWG. 73-37

 

   

 

 

 

 

 

 

a
3 . : ‘ i | ]
J -
% > FLOODED _
- FLOODED
g Q ‘
< o T e FLOODED
I~ 02*.‘..‘. |
5 .
o
uJI e
n o B
xoe OIF """ |
W~
a
»
0
0 } 4 : : % :
DISPERSED PHASE
3 VELOCITY, Vg —
Qo
g (ft/nr)
xrT FLOODED _ _  _ e o
e FLOODED
E 60} b
& -
2. * 'l e—— 85
o I
ST I ol 126
nxT |
Ll 40 ', -
@x e '
G.E |
ODED e
Pl -
- | |
0 50 100 150 200 250 300 350

CONTINUOUS PHASE VELOCITY, Vg (ft/hr)

Fig. 8L4. variation of Dispersed-Phase Holdup and Column Pressure
Drop During the Countercurrent Flow of Mercury and a Water-Glycerin
Solution in a 2-in.-ID, 24-in.-long Column Packed with 3/8-in. Teflon
Raschig Rings. The viscosity of the continuous (aqueous ) phase was 15

cP.
Table 56.

Experimentally Determined Values for Dispersed-Phase Holdup and Pressure
Drop During Countercurrent Flow of Mercury and Water Through a 2-in.-diam

Column Packed with 3/8-in. Copper Raschig Rings

Continuous-phase viscosity,

1

cP

 

Superficial Velocity (ft/hr)

 

Dispersed~-Phase

Pressure Drop

 

Dispersed Phase Continuous Phase Holdup (mm Hg-mm HEO)
113 0 0.108 15
113 Lhk 0.106 L6
113 272 0.113 4O
113 406 0.119 14
315 0 0.201 33
315 RN 0.204 36
315 206 0.194 --
315 339 0.215 41-50
315 Lo6 0.270 75
315 L0 0.308 103
315 380 0.233 58

 

Millimeters of liquid having a density of 12.6 g/cmj_

9¢e
227

ORNL DWG 73-I

 

 

 

 

 

 

! I T T 1 T T 1
a 03 /
o |
3
& a
---------------------- - _
§ 0.1k -
W
= 0 ] 1 1 | L 1 I I
1 T T 1 i I | I
100~ Dispersed Phase Velocity =
(ft/hr)
e enee 13
QS 315
o N
o« I
w B
% '
e 40} a -
@~ .
E tF————————em et e === [ e
& E L .
|
0 1 ] { 1 1 i 1 1

 

 

0 50 0O 150 200 250 300 350 400 450
CONTINUOUS PHASE VELOCITY (ft/hr)

Fig. 85. Variation of Mercury Holdup and Column Pressure Drop During
the Countercurrent Flow of Mercury and Water in a 2-in.-ID, 2L-in.-long
Column Packed with 3/8-in. Copper Raschig Rings. The packing was wet by
the mercury.
172 I/72
(DISPERSED PHASE VELOCITY, V;) (ft/hr)

ORNL DWG 73-2RI

 

 

 

  

    

 

 

 

20 T T I T ! T

18 ® — Not Flooded 7]

o -

CALCULATED FLOODING LINE for
72 172 IV2
'ZF ¢ ¢ o Ve +VYy . 33.6 (1t/hn) B
IO- T
® ® ®

o .
¢ n
4}~ -
2 -
0 5 10 1S 20 25 30 35

Fig. 86.

/2 I/
(CONTINUOUS PHASE VELOCITY, V) (ft/hw) 2

Flooding Data Obtained During the Countercurrent Flow of

Mercury and Water Through a 2-in.-ID, 24-in.-long Column Packed with

%/8-in. Teflon Raschig Rings.

gece
/72
(ft/hr)

i172

(DISPERSED PHASE VELOCITY,Vq)

ORNL DWG. 73-42 RI

 

 

 

 

 

14 T T T T T
2 CALCULATED FLOODING LINE for
° ° o o + Vc|/z+ V‘ln + 29.1 (tt/n)”?
ol ® e oo o o + -
sl ® e o o0 O -
6} -
4} -
® NOT FLOODED -
+ FLOODED
o) i 1 L L 1 ]
o) 5 |10 15 20 25 30 35

/
(CONTINUOUS PHASE VELOCITY ., V)2 (ft/nn)V?

Fig. 87. Flooding Data Obtained During the Countercurrent Flow of
Mercury and a Glycerin-Water Solution Through a 2-in.-ID, 24-in.-long
Column Packed with %/8-in. Teflon Raschig Rings. The viscosity of the
continuous (aqueous ) phase was 7.5 cP.

6cc
ORNL DWG. 73-43

 

    
 

 

 

 

/2 172
(CONTINUOUS PHASE VELOCITY, Vo)  (ft/hn)

Fig. 88. Flooding Data Obtained During the Countercurrent Flow of
Mercury and a Glycerin-Water Solution Through a 2-in.-ID, 24-in.-long
Column Packed with 3/8-in. Teflon Raschig Rings. The viscosity of the
continuous (aqueous ) phase was 15 cP.

~N

=

£

=

N:: 14 r { T Y I T

>

o I12F CALCULATED FLOODING LINE for -

=) . . o 4+ 4+ 4 2 12 172

> Ve +Vy4 =25.95 (ft/hr)

~ 10} .

O r . ° ® o o

S et ¢

w ]

>

w 6 -

2

Y -
i ® NOT FLOODED

8 2 + FLOODED -

v

T o 1 1 1 L | 1

% 0 5 10 15 20 25 30 35

a

oeEe
I/72
(ft/hr)

172

(DISPERSED PHASE VELOCITY, V4 )

ORNL DWG. 73-38

 

20 1 T 1 I L |
CALCULATED FLOODING LINE

p . . e o060 72 1/2 172
for Vo +Vq = 39.75 (ft/hr)

TES

® NOT FLOODED

 

 

 

5k -
+ FLOODED
0 | i 1 1 1 l ]
0 5 10 15 20 25 30 35 40
/2 172
(CONTINUOUS PHASE VELOCITY,V(;)I (ft/hr)

Fig. 89. Flooding Data Obtained During the Countercurrent Flow of
.~ Mercury and Water in a 2-in.-ID, 24-in.-long Column Packed with 3/8-in.
Copper Raschig Rings. The packing was wet by the mercury.
232

superficial velocity is plotted vs the square root of the continuous-
phase superficial velocity. The conditions for which the system was
operated in a nonflooded state, as well as conditions for which flooding
was observed, are indicated. The results obtained during this study
were similar to those observed previously for mercury and water. The
data on dispersed-phase holdup involving nonwetted packing could be
correlated in terms of a constant superficial slip velocity, as shown
in Table 57. The relative standard deviations for the slip velocities
were about #10%; no dependence of slip velocity on the flow rate of
either phase was noted. The variation of slip velocity with the
continuous-phase viscosity is shown in Fig. 90, which indicates that
the superficial slip velocity is proportional to the -0.167 power of
the continuous-phase viscosity. As expected, the dependence is not
large; however, neglect of this effect was significant in the earlier
extrapolation of data from a mercury-water system to a salt-bismuth

system since the continuous-phase viscosity differs by a factor of 12.

The effect of wetting of the packing by the metal phase on dispersed-
phase holdup, flooding, and pressure drop was also evaluated by packing
the column with 3/8-in. copper Raschig rings that had been etched with
nitric acid (to promote wetting of the packing by the mercury). Com-
plete wetting of the packing was obtained. Although the mercury was
saturated with copper, the solubility of copper in mercury is quite low
and no iwportant changes in other physical properties should occur.

After a few hours of operation, solids were observed at the water-mercury
interface below the column; these solids were believed to be copper

oxide that was formed as the result of the reaction of dissolved copper
with oxygen in the water. Periodic additions of nitric acid to the

aqueous phase quickly removed the solids.

The interfacial area between the aqueous and mercury phases was
decreased substantially when the packing was wetted by the mercury, as
noted by comparing Figs. 91 and 92, which show operation with nonwetted
and wetted packing respectively. No dispersion of the mercury was
observed with the wetted packing, and the interfacial area was essentially
equal to the packing surface area. The superficial slip velocity (and

hence the flooding rates) was considerably greater with wetted packing
233

Table 57. Variation of Superficial Slip Velocity with
Continuous-Phase Viscosity and Wetting of the
Packing for a Column Packed with 5/8-in. Raschig Rings

 

 

. . Slip
Continuous- Packing . Standard
Type Phase Void Ve13c1ty, Deviation, D P?riént
of Viscosity Fraction, S g EVl7 ton,
Packing (cP) € (ft/hr) (ft/hr) Vg
Nonwetted 1 0.66 1129 +105 9.3
Nonwetted 7.5 0.66 8Le +88 +10.4
Nonwetted 15 0.66 674 +49 +7.3
Wetted 1 0.82 1583 +268 +16.9

 

than with nonwetted packing, as shown in Table 57. It was not clear
whether the data on metal-phase holdup with wetted packing could be
correlated on the basis of a constant superficial slip velocity. Only
nine metal holdup measurements were made, and a quantitative analysis
of the relation between mercury holdup and the water and mercury super-

ficial velocities was not possible.

14.3 Prediction of Flooding Rates and Dispersed-
Phase Holdup in Packed Columns

After the dependence of slip velocity on the continuous-phase vis-
cosity had been determined, it was possible to reevaluate the dependence
of slip velocity on the difference of the densities of the two phases by
using the two data points afforded'by the meréury—water data and salt-
bismuth data. The salt-bismuth data indicated a superficial slifi velocity
of 388 ft/hr with a continuous-phase viscosity of 12.1 cP, a density
difference of 6.28 g/cm;, and a packing void fraction of 0.84. It was
assumed that the dependence of slip velocity on the difference in den-
sities is a power-type dependence, and the resulting power value was 0.5.

This result is interesting because it is the same as the dependence of
SLIP VELOCITY x10~2 (ft/hr)

ORNL DWG 71-2867

 

20 T l

O
1

  

@
I

SLOPE=—(VISCOSITY EXPONENT)
=—b2 0.167

)]
|

H
'

N
i

 

I | | | 1 1 11 11 | | | 1 1 1 111

 

 

 

0. 0.2 04 06 08 10 2 4 6 8 10 20

CONTINUOUS PHASE VISCOSITY (cP)

Fig. 90. Variation of Superficial Slip Velocity with Continuous-
Phase Viscosity in a 2-in.-diam Column Packed with 5/8-in. Teflon Raschig
Rings.

hee
PHOTO 0009-71

R

s LELS

 

Fig. 91. Countercurrent Flow of Mercury and Water in a 2-in.-diam
Column Packed with 3/8-in. Teflon Raschig Rings. The packing is not
wetted by the metal phase, and the interfacial area is much greater
than the surface area of the packing.
 

Fig. 92. Countercurrent Flow of Mercury and Water in a 2-in.-diam
Column Packed with 5/8—in. Copper Raschig Rings. The packing is wetted
by the metal phase, and the interfacial area is essentially equal to
the surface area of the packing.
237

drop terminal velocity on the difference in densities for conditions
under which inertial forces predominéte and the drag coefficient is
essentially constant. The final relation for predicting the variation
of superficial slip velocity with packing void fraction, the difference

in the densities of the phases, and the continuous-phase viscosity is,

 

 

 

 

then:
-0.167 0.5
Vs = Vs Hg-H_O eE ' L_ih—) (25_49—__- ’ (69)
PP “Ret/ P10 Hg -H,,0
where
VS = superficial slip velocity,
v = slip velocity for mercury-water for the packing size being
s,Hg-HQO
used,
e = packing void fraction,
_ . . I . d _
€Ref void fraction for packing for which Vs,Hg-H 0 was deter

mined, 2

i = viscosity of continuous phase,

My o = viscosity of water at 20°C,
2

Mo = difference in the densities of the phases, and

fipHg-H o= difference in the densities of mercury and water at 20°C.
2

Slip velocity values calculated from Eq. (69) can be used with Eqs. (67)
and (68) for determining both the dispersed-phase holdup and throughputs

of the continuous and dispersed phases at flooding.
238

15. ANALYSIS OF MULTICOMPONENT MASS TRANSFER BETWEEN MOLTEN SALTS
AND LIQUID BISMUTH DURING COUNTERCURRENT FLOW IN PACKED COLUMNS

C. P. Tung J. S. Watson

Reductive extraction, an important operation in the removal of prot-
actinium and rare earths from MSBR fuel salt, involves the exchange of
metal ions in the salt phase with neutral (reduced) metal atoms in the
bismuth phase. Since no net electric current flows between the salt and
metal phases, the rate at which metal ions are reduced must equal the
rate at which metal atoms are oxidized (taking into consideration differ-
ences in the charges of the ions involved). 1In the bismuth phase, the
fluxes of the transferring atoms are dependent only on concentration
gradients. However, in the salt phase, electric potential gradients are
generated near the salt-metal interface as the result of differences in
the mobilities and/or charges of the various diffusing ions. This results
in a condition where the fluxes of the transferring ions are dependent on
both concentration gradients and electric potential gradients. These
effects greatly complicate the mass transfer process and make the design
of continuous (differential) reductive extraction columns difficult. We
have begun a mathematical analysis of mass transfer during reductive
extraction processes to facilitate interpretation of the results from
present and proposed experiments in packed columns, and as an aid in
using these data for the design of larger reductive extraction systems.
The results from this study will also be applicable to several other sol-
vent extraction operations that involve exchange reactions between
uncharged species in a solvent and ions in an electrolyte. The extrac-
tion of metals from aqueous solutions by use of "liquid ion exchange"

solvents such as amines is a typical example of such an operation.

15.1 Literature Review

It has been recognized for some time60 that the diffusion of ions in
a liquid is inherently different from the diffusion of uncharged species.
The interdiffusion of charged species as a result of concentration grad-
ients produces an electric potential gradient, which, in turn, alters the
ionic fluxes and prevents further deviation from electroneutrality. An

electrical potential gradient in an electrolyte will produce a flux of
259

ion i, which depends on the concentration of the ion in the electrolyte

and on the valence and mobility of the ion, as indicated by the relation

= d
Ji uiZiCi grad ¢ ,

where
Ji =\f1ux of ion i,
u, = mobility of ion i in electrolyte,
;= valence of ion i,
;= concentration of ion i, and
grad ¢ = electrical potential gradient.

The mobility of an ion is related to its diffusion coefficient in the

electrolyte by the Nernst-Einstein equation as follows:

D.F
a = —=
i RT ’
where

Di = diffusion coefficient of ion i in the electrolyte,
F = Faraday constant,
R = gas constant, and
T = absolute temperature.

If a concentration gradient is also present in the electrolyte, an addi-
tional flux, which is described by Fick's First Law, is produced; the
flux resulting from this effect is given by the relation

Jl = -Dl grad Ci

The net flux of ion i will be the sum of these contributions, which is
knowvn as the Nernst-Planck equation:
Z,C F

ii
RT

 

= - d .
Ji Di(grad Ci + grad ¢)

Several studies related to mass transfer during reductive extraction
operations have been reported in the literature. The effects of electric
fields on the rate of diffusion of ions were considered by Schldgl and
Helfferich.6O Copeland, Henderson, and Marchello61 derived an analytical
solution for the rate of transfer of ions through the liquid film sur-

rounding solid ion exchange resin beads for the case in which the
240

exchanging ions have the same charge. Turner and Snowdon62 extended these
results to cover the case in which the exchanging ions have different
valences. Kataoka, Sato, Nishiki, and Ueyama65 studied systems involving
two different coions (e.g., ions having electric charges opposite in sign
to that of the exchanging ions). This study will attempt to extend the
previous studies in three important ways:

1. Countercurrent liquid-liquid columns will be treated
rather than portions of fixed beds.

2. Diffusion resistances may exist in both phases. Previous
studies have treated either resistance to diffusion in
solid ion exchange resins or resistance to diffusion
through a liquid film surrounding the resin. The case
for which significant resistance occurs in each phase
is considerably more complex than either of these.

3. Any number of transferring ions may be considered. The
results will not be limited to the case in which only
one ion exchanges for another ion. The method that will
be developed will be general, and any number of ions can
be treated. However, provisions are being made for only
as many as ten ions in computer programs that are being
developed.

15.2 Mathematical Analysis

The model selected for the mathematical analysis in the present study
is based on a model similar to the Whitman two-film model; that is, an
effectively stagnant film of liquid is assumed to exist on both sides of
the interface, and all concentration gradients and all electric potential
gradients are assumed to lie within these two films. The phases are
assumed to be in chemical equilibrium at the interface. The rates of dif-
fusion of species in the solvent film are controlled by Fick's First Law
since the transferring materials have no electric charge in the solvent
(bismuth). Thus, at a point within a packed column, the flux of compo-

nent i across the solvent film is given by the relation

1= _—c.), (70)

wvhere

i flux of component i across the solvent film,

si diffusion coefficient of component i in solvent phase,
241

68 = thickness of the solvent film,

SiB = concentration of component i in the bulk solvent, and
= concentration of component i in the solvent phase at the
interface.

sil
The rate of transfer of ions in the electrolyte film, however, involves
electrical transference as well as diffusion, and both concentration and

electrical potential gradients are important. For this case, the flux is

given by the Nernst-Planck equation:

Zz.C .F

_ g L diei ]
Tei Dei[gra Cei RT 8¥ad ¢ (71)
where

Z. = valence (electric charge) of component i,

= electric potential in the electrolyte phase,
= diffusion coefficient of transferring component,
Faraday constant,

= concentration of transferring component,

m 0o ™M g e -
i

= gas constant, and
T = absolute temperature.
The subscript i refers to component i, and the subscript e denotes the

electrolyte phase.

Since the films are thin, it is assumed that the rate of accumulation
of materials in the films is negligible in comparison with the flux of
the materials through the films. Thus, the flux of component i is con-

stant across the films so that

Jei - Jsi ’ (12)

The fluxes, however, will vary with axial position in the column; this
effect will be considered subsequently. There is no net transfer of
electric charge between phases (e.g., no net electric current ); this

is expressed by the relation
E T 7 = E T 7 =0 . {5
. . = - . . = ( )

Also, the accumulation of electric charge at any point can be neglected,
2o

so that at all points in the electrolyte film
Zi:cei 2 = 4 & o (T4)

where tlie subscript Y refers to the pontransferring ion. If the trans-
ferring ions have a positive charge (as is the case with reductive extrac-
tion), at least one other ion in the electrolyte phase must have an
opposite (negative) charge. (Ions that are charged oppositely to the
transferring ions are referred to as coions.) If the transferring ions
are negstively charged, the coions would be positively charged. The

rate of transfer of coions between the phases is assumed to be negligible.

It will be convenient to develop two general expressions from Eqs.
(7t) and (74 ) which will be used subsequently. Multiplication of Eq. (71)
by the quantity ZiDe' and summation of the resulting expression over the
i

i transferring ions yields

J. z Z.°C . F

el 1 1 el
Z - Zzi grad C_, Z = grad ¢ . (75)
1 el 1 1

Taking the gradient of Eq. (T4) yields

E;Zi grad C . = — Z, grad C, . (76)
i
For most aqueous solutions, the concentrations of all ions will be
small compared to the concentration of water molecules; hence it is
assumed that no solubility limits are reached at any point in the elec-
trolyte phase. For this case, the coion concentration will vary across
the electrolyte film, but there will be no net transfer of coions. This

condition is described by the relation

JY = 0 = -DeY grad CeY + —RT grad o] , (T?a)
where the subscript Y refers to the coion. Thus, the concentration pro-
file for coion Y is independent of the diffusion coefficient of the coion.
Kataoka, Sato, and Ueyama have solved Eqs. (71), (73), (74), and (77a)

analytically for the case where there are two different cations
2L3

(transferring ions).65 Combination of Eqs. (76) and (77a) yields the

relation
2
E: ZY CY F
d = ———
a z, grad C_. = grad ¢ , (78)
and substitution of Eq. (78) into Eq. (75) yields the final relation which
describes the electric potential gradient in the electrolyte film. The

resulting relation is

 

J . Z
z: el 1
F i Dei
Ef-grad 0= = — - z: - 5 - . (79a)
ZY Y T i ei

For the case in which there are more than two transferring ioms, Eq. (71)

must be integrated numerically for each ion.

If the electrolyte phase is a molten salt, the concentration of
transferring ions in the electrolyte film relative to the total concen-
tration of ions can be high, and the volume of each ion should be taken
into account. For molten salt systems of immediate interest, this effect
can be treated approximately by a simple technique In the case of MSBR
fuel salt, the electrolyte phase will consist of a mixture of the fluo-
rides of Th, Li, Be, U, Pa, Zr, and rare earths. In reductive extraction
operations involving molten LiCl, the electrolyte phase will consist of
a mixture of the chlorides of these materials (except for beryllium) and
of the rare-earth and alkaline-earth fission products. 1In both cases,
only cations will exchange between the bismuth and salt phases. As shown
in Tables 58 and 59, the equivalent volumes of most of the salts in each
individual group are approximately the same (12.1 CmB/equiv for fluorides
and 27.1 cmB/equiv for chlorides ). It has also been found that the equiv-
alent volumes are linearly additive for fluoride mixtures containing the
elements of interest.6u In view of these factors, the concentration of
fluoride (or chloride) ion will be assumed to be constant throughout the
electrolyte film. Thus

CY = constant = l/D , (TTb)
2Ll

Table 58. Empirical Molar and Equivalent Volumes
of Fluorides at 600°Ca

 

 

Molar Volume Equivalent Volume
Material (cm3/mole) (cm3/equiv)
LiF 13.46 13.46
BeF2 23.6 11.8
ThF, 46.6 11.65
UF4 45.5 11.38
ZrF4 47 11.75
YF3 34.6 11.53
LaF3 37.7 12.57
CeF3 36.3 12.1
PrF3 36.6 12.2
SmF3 39.0 13.0
SrF2 30.4 15.2
BaF2 35.8 17.9

 

%pata taken from ref. 64.

Table 59. Molar and Equivalent Volumes
for Chlorides at 650°C

 

 

Molar Volume Equivalent Volume
Material (cm3/mole) (cm3/equiv)
LiCl 28.6 28.6
SrCl2 55.5 27.8
BaCl2 61.5 30.7
YCl3 76.7 25.6
L3013 72.7 24.3
CeCl3 72.5 24.2
ThC1 112.6 28.2

4

 
U5

where v = equivalent volume of salt, cmi/equiv. This assumption will not
be universally valid for all molten salt systems, but it is believed to
be reasonable for cation exchange between molten salt and bismuth since
the anions in this instance will usually be much larger than the cations.
Equations (75), (76), and (77b) may be combined to obtain an expression
that describes the electric potential! gradient for the case where the

electrolyte is a molten salt. The resulting relation is

L
%f grad ¢ = — EET———E;-——— . (79b)
i

In summary, the relations which define the rate at which components

transfer between an aqueous and an organic phase are as follows:

 

D .,
si
Joi = B (Csr'.l CsiB) ’ (70)
Zi Cei F
_ - Ll grad o
Jos Dei[grad C,. ™+ o grad o) , (71)
= e
Joi = Jdg1 7 (72)
E: Jei %4 = E: Tsi zl ’ (73)
i i
and
J . Z,
z: ei i
F D .
RT grad ¢ = — 5 (79a)

i el
D
Z," Cy + Zl: z.“c_,

Similarly, the relations which define the rate at which components transfer
between a molten salt and bismuth phase consist of Eqs. (70), (T71), (72),

(7%),anc the relation
246

J . Z
2: el 1
%f grad ¢ = «-ni-———Egi——— . (79b)
z C .
i el

Equilibrium relations for the distribution of solutes between salt-
metal or aqueous-organic phases can take many forms, and in principle,
almost any type could be used with the calculational procedure being con-
sidered. A mass action type of equilibrium relation was chosen, which

can be expressed as

N

i
Cei Csr Zr
C, (E::) = Q. , (80)
where Q1 is the equilibrium constant and the subscript r refers to a
reference component. Ferris and co-workers65 have measured the equilib-
rium distributions of several materials between molten salts and bismuth.
Their studies have shown that the activity coefficients for the species
in both phases are essentially independent of composition; hence, the
equilibrium quotients Qi are essentially constant. Equation (80) is

believed to be appropriate for several aqueous systems also.

15.3 Calculational Procedure

The rates of transfer of materials between two liquid phases at a
given axial position in a packed column are calculated by a trial-and-
error integration of Eq. (71) across the electrolyte film. When only
two transferring ions are present, the integration can be made analyti-
cally in a manner similar to that shown by Kataoka, Sato, and Ueyama.65
When more than two transferring ions are present, a numerical procedure
such as the one discussed below is recommended. The bulk concentrations
of the transferring materials in both phases, the mass transfer coef-
ficient (Di/fis) for each component in both phases, and the equilibrium
constants [defined by Eq. (80)] are required for the calculations. The

transfer rate for each ion and the concentration of each ion in the

electrolyte at the interface are computed by the following procedure:
2L

A value is assumed for the flux of each transferring
ion through the electrolyte film.

The electrolyte film is divided into several thin incre-
ments. Equations (79b) and (71) are used to calculate
the electric potential gradient and concentration gradi-
ents of all components in the increment adjacent to the
region where the bulk concentrations occur.

The concentrations of the transferring materials at the
boundary between the first and second increments of the
electrolyte film are calculated using the calculated
values for the concentration gradient.

Steps 2 and 3 are repeated until values for the concen-
trations of the transferring materials are established
throughout the electrolyte film.

A value is assumed for the concentration of the reference
component (which can be any one of the transferring

ions ) in the solvent phase at the interface. Equation
(80) and the concentrations of the transferring mate-
rials in the electrolyte phase at the interface (from
step 4) are used to calculate concentrations of all other
transferring materials in the solvent phase at the inter-
face.

The rate of transfer of each ion across the solvent film
is calculated using Eq. (70). The sum of the products
of the transfer rates and the valences of the components
should be zero, as shown in Eq. (73). 1In general, this
will not be the case for the first value assumed for the
concentration of the reference component in the solvent
phase at the interface. The next value for the concen-
tration of the reference component is obtained by using
Newton's method with Egs. (73) and (80), as follows.

For the expression

= - 1
£ 2; Z; Doy (CsiB CsiI) ’ (81)
f should have the value of zero when the proper value of
C is found. The derivative of f with respect to C
srl srL
is: ., C -
£' = 9. 2.0 D, it . (82)

i si Zr CSrI

The next value to be assumed for CS is then given by

r
the relation I

f?
(Cer)n+1 - (Cer)n T F (83)
248

7. Steps 5 and 6 are repeated until the concentration of
the reference component in the solvent phase at the
interface is kgown within the specified convergence
criterion (10°° has been used thus far in this study).

8. The calculated values for tne flux of each transferring
component across the solvent film are compared with the
assumed values for the respective fluxes across the
electrolyte film. If the values are not equal, a new
set of values is assumed for the fluxes of the trans-
ferring materials through the electrolyte film. The new
value for the flux of compoaent i is taken to be the
arithmetic average of the previously assumed value and
the flux of component i across the solvent f£ilm that was
calculated using this value. Steps 1 through 7 are
repeated until the difference between the values for each
material is less than a specified quantity (1% of the
average of the flux values has been used thus far in this
study ).

The calculational procedure ou:lined above was found to corverge
rapidly; however, it is not necessarily the optimum procedure for solving

Egs. (70)-(80).

After a procedure had been developed for calculating the concentra-
tions and fluxes for the transferring components in the electrolyte and
solvent phases at a given axial location in a packed column, it was
necessary to develop a calculationai procedure that could be used through-
out the column. The column was divided axially into a number of incre-
ments; the concentration of each component was assumed to be constant
throughout a given increment. It was assumed that the electrolyte and
solvent phases were in constant volumetric flow throughout the column
and that no dispersion occurred in either phase. The interfacial area
per unit column volume and the values for the film thickness in the
solvent and electrolyte phases were assumed to be constant throughout
the column. Material balances for each phase in a given column incre-

ment then resulted in the following relations:

d Csi(z) Sa
—y " —'Vg' Ji(Z), (84)
d ¢ (z) .
S
—5— =~ 1.(2), (85)
249

where

Csi(Z) = concentration of component i in the solvent at height

Z in the column,

C ,(Z) = concentration of component i in the electrolyte at

ei
height Z in the column,
VS = volumetric flow rate of solvent,
Ve = volumetric flow rate of electrolyte,

S = cross—-sectional area of column, and

a = interfacial area per unit column volume.
In deriving these relations, it was assumed that transfer of material
from the solvent phase to the electrolyte phase resulted in a positive
value for the flux. These relations can be expressed in finite dif-

ference form as follows:

n-1
At 1
- = — =J,.(jAZ 86
c (0 —C_, (L 7_n 1J(J ) , (86)
j=0
n-1
A
c ,(0)— c .(L) = L2y (342) (87)
j=0
where
At = Sal., total interfacial area in column,
j = column increment number, (j = 0, ...n),
n = number of increments in column
= L/AZ,
AZ = height of column increment, and
Jij = flux of component i in increment j.

These relations were used in the following manner for calculating concen-
trations and fluxes for the transferring materials throughout a column:
(1) Values were assumed for the concentration of each component
in the solvent phase in the increment from which the solvent

phase exits the column. Equations (70)-(79) were used for
250

calculating the fluxes of the transferring materials between
the solvent and electrolyte phases in the increment.

(2) Equations (86) and (87) were used for calculating the con-
centrations of the transferring materials in the solvent and
electrolyte phases in the next increment of the column. The
fluxes of the transferring components were again calculated
in this increment, using Egqs. (74)-(79). This procedure
was repeated until concentrations and fluxes for each trans-
ferring material had been calculated in each increment of
the column.

(3) The calculated values for the concentrations of the trans-
ferring components in the solvent phase entering the column
were compared with values that were specified by the operating
conditions. Usually, differences between the calculated
and specified values were observed. When the difference
was greater than desired, a new set of values was assuned
for the concentrations of the transferring materials in the
solvent stream leaving the column. Each new value was cal-
culated by summing the values for the concentrations of
component i in the exit solvent stream with one-half of the
difference between the specified concentration of component
i in the inlet solvent stream and the calculated value
for the concentration of component i in the inlet solvent
stream from the last iteration. This procedure was repeated
until the calculated and the specified values for the con-
centrations of transferring materials in the inlet solvent

stream were in agreement.

In the future, work will be carried out for calculating rates of
mass transfer between solvent and electrolyte phases for a range of
operating conditions. Particular attention will be paid to the influ-
ence of the electric field on the rate of mass transfer, and to the
differences that result from the case where mass transfer rates are

assumed to be dependent only on concentration gradients.
251

16. STUDY OF THE PURIFICATION OF SALT BY CONTINUOUS METHODS

R. B. Lindauer L. E. McNeese

We have previously described equipment for studying the purification
of salt by continuous methods.66 Initial work with this system was
directed at the measurement of the flooding rates in a 1l.25~in.-diam,
7-ft-long column packed with 1/4-in. nickel Raschig rings. Flooding
data were obtained during the countercurrent flow of molten salt (66-34
mole 7% LiF—Ber) and hydrogen or argon.67 The objective of the present
work is to study the continuous reduction of iron fluoride in molten
salt by countercurrent contact of the salt with hydrogen in a packed
column. We reported previously68 on the first two iron fluoride reduc-
tion runs (runs 1 and 2), which were carried out at a temperature of
700°C. Reasonable values for the mass transfer coefficient for the
transfer of iron fluoride from the bulk salt to the gas-salt interface
were obtained. Operation of the column during these runs was erratic,
and the pressure drop across it increased to twice the initial wvalue.
The increased restriction was believed to have resulted from precipi~
tation of BeO on the column packing as the result of an accumulation

of oxide in the system.

During this report period, salt purification studies using 66-34
mole 7% LiF—BeF2 were terminated because of leaks that resulted in
the loss of about half of the l4-liter salt charge. The composition
of the remaining salt was adjusted to the approximate composition of
the proposed MSBR fuel salt (72-16-12 mole % LiF—Ber—ThFA) by addition
of sufficient quantities of salt having the composition 72.6-27.4 mole

%Z LiF-ThF, and LiF powder to produce 17 liters of salt having the com-

4
position 72.0-14.4-13.6 mole 7% LiF—Ber—ThFA. The newly prepared LiF-
Ber—ThF4 salt was then countercurrently contacted with a Hz——IOZ HF

mixture in the column in order to remove oxide from the salt. Although
considerable oxide was removed, the pressure drop across the packed
column was reduced only slightly. This indicated that a significant

quantity of oxide still remained in the column. Two flooding runs and
252

one iron fluoride reduction run were then made. During these runs, the
pressure drop across the column increased to the point where operation
of the system became difficult. The packed column was then filled with
molten salt, and an HF—H2 stream was allowed to contact the static salt
charge for a period of 18 hr in order to remove the oxide from the column.
This operation was successful in reducing the pressure drop across the
column to approximately the value observed after the first two iron
fluoride reduction runs. Eight additional iron fluoride reduction runs
were subsequently completed. During these runs, operation of the system
was smooth, and there was no increase in pressure drop. The results
obtained by analyzing salt samples from the runs for iron were incon-
sistent, probably because of the low iron concentration in the system
although sample contamination was suspected in some cases. These oper-

ations are described in greater detail in the remainder of this section.

16.1 Removal of Oxide from Salt by Countercurrent Contact
with an HF--H2 Gas Stream

We have found that routine measurement of the pressure drop across
the packed column is a useful means for detecting the buildup of mate-
rials such as metallic iron or insoluble oxides in the column. As
shown in Fig. 93, the pressure drop across the column with an argon
flow rate of 5 liters/min increased from 5.6 in. H20 to 12.0 in. HZO
during previously reported68 experiments with the LiF—BeF2 salt
(flooding runs No. 3-10). Since less than 6 g of iron would have been
reduced during this time, the restriction is believed to have been due
mainly to an accumulation of insoluble Be0O on the column packing. The
oxide in the salt could have originated from several sources. Although
the gas purification and supply equipment for the hydrogen and argon
was used throughout all the experiments, the purification traps were not
regenerated before the experimental work was initiated. Instrumentation
for monitoring the moisture content of the gases was installed after 1
month of operation, and an oxygen analyzer was installed 3 months after

operation of the system began. Typical water and oxygen contents of

the unpurified gases were 15 ppm and 2 ppm, respectively, and are much
*UT!t

"¢6 814

/sxa:n ¢ 3o 93eyg Mmo1d uolay ue yiim doig sanssaid uwnion

0l61=-31vQ

COLUMN PRESSURE DROP, in, of H20
- - n ) o H

O 3] o w o w S o o

I | | 1 1 1
g o AFTER FLOODING RUN NO, 3
o
N AFTER FLOODING RUN NO. IO
o
@ AFTER REDUCTION RUN NO. 2
®
Q 0 AFTER COUNTERCURRENT HZ-HF TREATMENT
w
§ AFTER REDUCTION RUN NO. 3 o
o
< AFTER 3-DAY COLUMN CLEANOUT WITH H2-HF
n
i AFTER REDUCTION RUN NO. 4
3 ® AFTER REDUCTION RUN NO. 5
= ® AFTER REDUCTION RUN NO. €
o
§ ® AFTER REDUCTION RUN NO. 10
®
< o AFTER REDUCTION RUN NO. ||
S

i 1 1 i | !

 

 

 

 

Sh—-cL O9MQd TINMO

£Ge
25k

too low to.account for the quantity of oxide (several hundred ppm) that
had accumulated in the salt during the six months of operation. Some
contamination undoubtedly occurred during repair of plugged vent lines
and possibly as the result of insufficient purge rates when the system
was vented to the atmosphere between runs. Recent practice has been

to close the system vent valve between experiments, while maintaining

a sufficient purge rate to keep the system at a few psi of positive

pressure (a small amount of gas leakage occurs through the vent valve).

In order to remove the oxide that had accumulated in the salt, the
salt was contacted with a 10-90 mole 7% HF-—H2 gas stream for 165 min at
average salt and gas flow rates of 103 cm3/min and 5 liters/min respec-
tively. Some difficulty was experienced in maintaining a constant gas
sample flow rate to the analyzer used for determining the concentration
of HZO in the gas stream leaving the column; however, during a period
of steady flow near the end of the run, the analyzer indicated that 100
ppm of oxide was being removed from the salt as it passed through the
column. The HF utilization at this point was 15Z. Treatment of the
salt with the HF—H2
from 12.0 in. H,O0 to 10.3 in. H

2 2
min, which indicated that the column still contained an appreciable

stream reduced the pressure drop across the column

O with an argon flow rate of 5 liters/

quantity of oxide.

16.2 Removal of Oxide from Column

Following the countercurrent contact of the salt with an HZ—HF
gas stream, two flooding runs and one iron fluoride reduction run (No.
3) were carried out. During the second flooding run, it was necessary
to heat the molten salt (downstream from the column) to a much higher
temperature (650°C on the bottom of the filter housing) than usual in
order to maintain the flow of salt through the filter. This could have
been due to the presence of oxide in the salt since the solubility of
BeO in salt increases as the temperature is increased. During the
reduction run, the pressure drop across the column increased, as had

been observed in the two previous runs made with salt having the
255

composition 66-34 mole % LiF—Ber. However, a test with argon after
reduction run No. 3 showed an even higher column pressure drop (see
Fig. 93) than had been observed during the run. An HF-—H2 gas stream
was then passed through a static charge of molten salt in the column

in order to provide more opportunity for the dissolution and hydro-
fluorination of insoluble oxide. The column was filled with salt to
the top of the packing (about 1-1/2 liters) by observing the hydro-
static pressure at the bottom of the column. For the first 6.5 hr,

the column was held at 650°C and a 29-71 mole ¥ HF--H2 gas mixture was
passed through the column at the rate of 8.3 liters/min. The concen-
tration of HF in the stream was higher than planned because of an

error in the size of the capillary used for measuring the HF flow rate
(2.4 liters/min). Only about 1.5% of the HF was utilized at this HF
flow rate, and oxide was removed from the column at the rate of 0.024
mole/hr. Passage of the I-IF—H2 stream through the column was continued
for 6.5 hr, after which it was terminated at the end of the day shift.
When the treatment was resumed the following day, the column was cooled
to 590°C in order to increase the concentration of HF in the salt, and
the HF flow rate was reduced to 750 cm3/min. The oxide removal rate
(as indicated by the water analyzer in the off-gas stream) was about
the same as had been observed previously, and the HF utilization increased
to about 5%Z. This compares favorably with the value obtained during
countercurrent flow of salt and gas through the column (i.e., 15%).

The treatment was resumed on the third day with a column temperature

of 700°C to facilitate the dissolution of any remaining oxide. The
output from the water analyzer decreased to a value below the reference
value, which indicated that the NaF trap in the sample stream had become
saturated and was allowing HF to reach the sample electrode. Since the
rate of oxide removal could no longer be measured, the oxide removal
operation was terminated after the salt had been contacted with the
HF—H2 gas stream for a total of 18 hr. The remainder of the batch of
salt was then passed countercurrent to a 10-90 mole Z HF—H2 gas stream
in the column. On completion of these operations, the pressure drop across

the column had been reduced to 14.8 in. H20° After two additional iron
256

fluoride reduction runs (runs No. 4 and 5), the column pressure drop
had decreased to 9 to 10 in. H20 (probably from dissolution of insoluble
oxide remaining on the packing) and remained in this range for the rest

of the report period.

16.3 Iron Fluoride Reduction Runs

During this report period, nine iron fluoride reduction runs were
carried out with salt having the composition 72.0-14.4-13.6 mole % LiF-
Ber—ThF4. Table 60 summarizes the data from these runs, along with the
data from the two previous reduction runs (Nos. 1 and 2) made with salt
having the composition 66-34 mole % LiF-BeF,. During the nine current
runs, the concentration of iron fluoride in the salt was decreased from
220 ppm to 70 ppm or less. Operation of the equipment was satisfactory
following run 3 (after removal of oxide from the column); however,
analyses of the resulting salt samples for iron produced inconsistent
data. The reported data for four of the last eight runs showed an
increase in the iron fluoride content of the salt. Two of the reported
values exceeded the total iron concentration believed to be possible
(276 ppm). These unusually high analyses may have been caused by con-
tamination of the samples during their removal from the nickel samplers

or by the diminished accuracy of the analytical method as the iron con-

centration is decreased.

16.4 Calculated Values for the Mass Transfer Coefficient and
the Reaction Rate Constant During the Reduction
of Iron Fluoride

The rate of reduction of FeF, by reaction with hydrogen can be

affected by a number of factors, 2One of the objectives of the present
work is to determine the rate-controlling steps involved in the reduction
of FeF2 in a packe§8column, and to evaluate the associated rate constants.
We have previously = carried out mathematical analyses for two limiting

cases:
257

Table 60. Data From Iron Fluoride Reduction Runs

Column Temperature, 700°C

 

 

Gas Flow Analyses of
Rate Salt Flow Filtered Samples Percent of
Run (std. liters/min) Rated (ppm_iron) Batch
No. Date H2 Ar (cm3/min) Feed Product Contacted
1 7/27 20.0 -— 100 425 307 68.5
2 7128 13.5 -- 100 307 228 27.6
3 10/22 16.6 -— 100 220 158 75.3
4 11/3 14.6 ’ - 210 158 373 84.7
5 11/4 18.2 - 161 373 137 68.8
6 11/5 4.5 4.5 106 137 110 81.8
7 11/6 3.9 3.9 142 110 70 81.8
8 11/9 24,0 - 103 70 75 86.5
9A 11/13 4.5 4.5 93 b c
75 55 79.4
9B 11/13 3.5 3.5 136
d b
10 11/17 14.1 -— 105 69 77 81.8
11 11/19 3.1 4.5 117 207 339 81.2
104°

 

#The first two runs used LiF-BeF salt; the remaining runs used LiF—BeFZ—ThF4 salt.

2
bAverage of flowing-stream samples.

“A value of 100% is used in calculating kpa since samples are of flowing-stream tvpe.

dCalculated from flowing-stream samples from previous run.
258

(1) the case in which the rate of reduction is limited by the
rate of transfer of Fer to the gas-salt interface from
the bulk salt, and
(2) the case in which the rate of reduction is limited by the
rate of reaction between H2 and FeF2 dissolved in the salt
at the gas-salt interface.
The associated rate constants can be evaluated by the relations given
below. For the case in which the rate of reduction is limited by the
rate of transfer of FeF2 from the bulk salt to the gas-salt interface,
the product of the mass transfer coefficient k2 and the interfacial
area a (which is not known) is given as:
X

 

 

k,a = XH In Xi , (88)
o
where
kfl = mass transfer coefficient for the transfer of FeFs from the
bulk salt to the salt-gas interface, moles/sec*cm?,
a = gas-salt interfacial area per unit column volume, cm2/cm3,
A = cross-sectional area of the column, cmz,
L = salt flow rate, moles/sec,
H = column height, cm,
Xi = concentration of FeF2 in salt fed to column, mole fraction,
XO = concentration of FeF2 in salt leaving column, mole fraction.

This relation is valid only under conditions such that the concentration

of FeF2 in salt that would be in equilibrium with the HF--H2 mixture

adjacent to the salt being considered (X*) is small in comparison with

the concentration of FeF2 in the bulk salt in the region being considered.

For the case in which the rate of reduction is limited by the rate of

reaction between FeF2 in molten salt and H2 in the gas at the gas-salt

interface, the rate constant is given by the relation

L Xi
k = ————— 1ln — , (89)
s KH HA Py Xo
259

where

-1
reaction rate constant, sec ~,

=
i

Henry's law constant for hydrogen in salt, moles/cm3-atm,

ol

partial pressure of hydrogen in gas, atm,

0
N
[

and the other quantities are as defined above.

Values for the mass transfer coefficient and the reaction rate con-
stant were calculated for the current runs for which meaningful results
could be obtained. These values, along with values from the two runs

reported previously, are summarized in Table 61.

Since the value for X* is negligible compared with the FeF, con-

centration in the bulk of the salt for the runs considered, thezexPression
used for calculating kza should be valid. Also, since the partial
pressure of hydrogen (sz) is nearly constant throughout the column, the
expression given above for calculating kS should be valid. Although

the inlet hydrogen partial pressure was reduced to 0.5 atm in the last
three runs for which rate constants are shown, the variation in the cal-
culated constants is not sufficiently large to determine whether the

rate of iron fluoride reduction is countrolled by the rate of transfer

of FeF, to the gas-salt interface or by the rate of reaction between

2

FeF2 and hydrogen at the interface. This question will be resolved

after additional data are obtained.
Table 61. Calculated Values for the Mass Transfer Coefficient
and the Reaction Rate Constant fer Runs R-1 and R-2

 

 

 

 

 

 

Mass Transfer Controlled Reaction Rate Controlled
FeF, Concentration kya 3 X* Py
Run (mole fraction x 10 ) (moles/sec-cm (mole fraction (se2-1 2 (atm) .
No. Xq X, x 109) x 10%) x 10%4) Inlet Exit Equilibrium’
R-1 5.13 3.70 2.4 0.033 2.4 1.0 0.9987 0.9850
R-2 3.70 2.75 5.4 0.198 5.4 1.0 0.9967 0.9959
R-3 2.65 1.90 1.9 0.035 1.9 1.0 0.9993 0.9974
R-6 1.65 1.33 1.2 0.042 2.5 0.5 0.4995 0.4969
R-7 1.33 0.84 2.9 0.170 5.7 0.5 0.4989 0.4975
R~9A 0.90 .66 1.8 0.032 3.7 0.5 0.4995 0.4978

Cace

 

a
Calculated from the composition of the exit gas.

bCalculated from the FeF2 composition in the salt leaving the column.
10.

11.

12.

13.

4.

15.

261

17. REFERENCES

R. G. Ross, C. E. Bamberger, and C. F. Baes, Jr., MSR Program Semiann.
Progr. Rept. Aug. 31, 1970, ORNL-4622, pp. 92-95.

 

 

J. C. Mailen, MSR Program Semiann. Progr. Rept. Feb. 28, 1971, ORNL-
L6T6, pp. 245-48.

C. E. Bamberger and C. F. Baes, Jr., J. Nucl. Mater. 35, 177 (1970).

M. J. Bell and L. E. McNeese, MSR Program Semiann. Progr. Rept. Aug.
31, 1970, ORNL-4622, pp. 202-8.

L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder
Reactor Processing No. 5, ORNL-TM-3140 (October 1971), pp. 2-15.

 

L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder
Reactor Processing No. 6, ORNL-TM-3141 (December 1971), pp. 4-13.

 

L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder
Reactor Processing No. 8, ORNL-TM-3258 (May 1972), pp. 3-2L.

 

L. E. McNeese, MSR Program Semiann. Progr. Rept. Feb. 28, 1970, ORNL-
4548, pp. 282-88.

M. J. Bell and L. E. McNeese, MSR Program Semiann. Progr. Rept. Aug.
31, 1970, ORNL-4622, pp. 199-202.

 

F. J. Smith and C. T. Thompson, Ibid., pp. 207-8.

L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder
Reactor Processing No. 7, ORNL-TM-3257 (February 1972), pp. 16-29.

L. E. McNeese, Enginéefing Development Studies for Molten-Salt Breeder

 

Reactor Processing No. 8, ORNL-TM-3258 (May 1972), pp. 2h4-30.

 

R. L. Jarry, W. H. Gunther, and J. Fischer, The Mechanism and Kinetics

 

of the Reaction Between Nickel and Fluorine, ANL-6684 (August 1963 ).

W. H. Gunther and M. J. Steindler, Laboratory Investigations in
Support of Fluid-Bed Fluoride Volatility Processes. Part XIV. The
Corrosion of Nickel and Nickel Alloys by Fluorine, Uranium Hexa-
fluoride, and Selected Volatile Fission Product Fluorides at 500°C,
ANL-7241 (December 1966 ).

 

 

 

Chemical Engineering Division Summary Report, July-September, 1958,

ANL-5924 (1958).
16.

17.

18.

19.

20.

21.

22.

23,

2kh.

o5,

26.

27.

28.

29.

30.

31,

32,

262

Chemical Engineering Div. Semiann. Rept., January-June, 1964, ANL-

6900 (August 1964 ).

P. D. Miller and W. E. Berry, A Survey

of Corrosion in the Fluidized-

Bed Volatility Process, BMI-X-362 (November 1965 ).

 

F. T. Miles et al., Progress Report of

the Reactor Science and Engi-

neering Department, BNL-176 (March 1952), pp. 17-18.

 

L. E. McNeese, Engineering Development

Studies for Molten-Salt

 

Breeder Reactor Processing No. 2, ORNL-

 

22-43.

MSR Program Semiann. Progr. Rept. Feb.

TM-3137 (February 1971), pp.

29, 1968, ORNL-4254, pp. 260-63.

H. F. Bauman, ORNL, personal communication, August 1970.

L. E. McNeese, Engineering Development
Reactor Processing No. 5, ORNL-TM-3140

 

L. E. McNeese, Engineering Development
Reactor Processing No. 6, ORNL-TM-3141

 

L. E. McNeese, Engineering Development
Reactor Processing No. 8, ORNL-TM-3258

 

L. E. McNeese, Engineering Development
Reactor Processing No. 4, ORNL-TM-3139

 

L. E. McNeese, Engineering Development
Reactor Processing No. 5, ORNL-TM-3140

 

L. E. McNeese, Engineering Development
Reactor Processing No. 6, ORNL-TM-31L41

 

L. E. McNeese, Engineering Development
Reactor Processing No. 8, ORNL-TM-3258

 

Studies for Molten-Salt Breeder
(October 1971), pp. 15-22.

Studies for Molten-Salt Breeder
(December 1971), pp. 23-39.

Studies for Molten-Salt Breeder
(May 1972), pp- 3-21.

Studies for Molten-Salt Breeder
(August 1971), pp. 38-83.

Studies for Molten-Salt Breeder
(October 1971), pp. 22-30.

Studies for Molten-Salt Breeder
(December 1971), pp. 13-23.

Studies for Molten-Salt Breeder
(May 1972), pp. 31-58.

M. E. Whatley et al., Unit Operations Section Quarterly Progress

 

Report, April-June 1968, ORNL-L365 (February 1970), pp. 33-45.

V. G. Levich, Physiochemical Hydrodynamics, pp. 47-52, translated by

 

Scripta Technica, Inc., Prentice-Hall,

H. P. Meissner, Process and Systems in

Englewood Cliffs, N. J., 1962.

Industrial Chemistry, Chap. 9,

 

Prentice-Hall, Englewood Cliffs, N. J.,

1971.

G. D. Towell, C. P. Strand, and G. H. Ackerman, "Mixing and Mass
Transfer in Large Diameter Bubble Columns," A.I.Ch.E. J., Chem. Eng.

Symp. Ser., No. 10, p. 97 (1965).
35

3.

35

36.

57

58.

29-

L0,

h1.

Lo,

Lz,

L.

L5.
L46.

b

L8.
4g.

263

Y. Aoyama, 0. Katsunori, K. Koide, and H. Kubota, "Liquid Mixing in
Concurrent Bubble Columns,"” J. Chem. Eng. Japan 1(2), 158 (1968).

K. Imafuku, T. Wang, K. Koide, and H. Kubosta, "The Behavior of Sus-
pended Solid Particles in the Bubble Column," J. Chem. Eng. Japan 1

(2), 153 (1968).

Y. Ohki and H. Inoue, "Longitudinal Mixing of the Liquid Phase in
Bubble Columns," Chem. Eng. Sci. 25, 1 (1970).

K. B. Bischoff and J. B. Phillips, "Longitudinal Mixing in Orifice
Plate Gas-Liquid Reactors," Ind. Eng. Chem., Process Design Develop.

5(4), 416 (1966).

S. H. Eissa, M. M. El-Halwagi, and M. A. Saleh, "Axial and Radial
Mixing in a Cocurrent Bubble Column,” Ind. Eng. Chem., Process Design
Develop. 10(1), 31 (1971).

T. Reith, S. Reuken, and B. A. Israel, "Gas Holdup and Axial Mixing
in the Fluid Phase of Bubble Columns," Chem. Eng. Sci. 23, 619 (1968 ).

L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder
Reactor Processing No. 1, ORNL-TM-3053% (November 1970}, pp. 1l-14.

 

 

L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder
Reactor Processing No. 6, ORNL-TM-3141 (December 1971), pp. 73-T9.

 

 

L. E. McNeesé, Engineering Development Studies for Molten-Salt Breeder
Reactor Processing No. T, ORNL-TM—525f_zFebruary 1972), pp. 52-58.

 

 

L. E. McNeese, Engineering Development Studies for Molten-~Salt Breeder
Reactor Processing No. 8, ORNL-TM-3258 (May 1972), pp. 64-89.

 

 

D. T. Goldman and J. R. Roesser, "Chart of the Nuclides," 9th ed.,
1966.

N. E. Holden and F. W. Walker, "Chart of the Nuclides," 10th ed.,
1968.

H. E. McCoy, personal communication to L. E. McNeese (February 1971).

L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder
Reactor Processing No. 7, ORNL-TM-3257 (February 1972), pp. 29-L6.

 

L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder
Reactor Processing No. 8, ORNL-TM-3258 (May 1972), pp. 58-63.

Ibid., p. 61.
J. B. Lewis, Chem. Eng. Sci. 3, 248-59 (1954 ).

J. B. Lewis, Chem. Eng. Sci. 3, 260-78 (1954).
51.

55

Sh.

55

56.

o~

58.

29

60.
61.

62.
63.

6L

65.

66.

67 .

68.

264

W. J. McManamey, Chem. Eng. Sci. 15, 251-54 (1961).
G. R. A. Mayers, Chem. Eng. Sci. 16, 16-75 (1966).

K. F. Gordon and T. K. Sherwood, Chem. Eng. Progr.,Symp. Ser. 50(10),
15-23  (195h4).

D. R. Olander, Nucl. Sci. Eng. 1k, 287-94 (1962).

J. Prochazka and J. Bulicka, Session VA: Transport and Interfacial
Phenomena, Inst. Chem. Proc. Fund. Prague, Czechoslovakia (April

1971).

Jaroslav Prochazka, personal communication to H. O. Weeren (Mar. 2k,

1972).

J. S. Watson and L. E. McNeese, Unit Operations Section Quarterly
Progress Report, July-September 1968, ORNL-4366 (April 1970), pp.
57-98.

 

 

L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder
Reactor Processing No. 5, ORNL-TM-3140 (October 1971), pp. 102-3.

 

 

P. A. Carroad and A. C. Meyerey Flooding Characteristics of a Counter-

 

current Mercury-Water Packed Column, ORNL-MIT-115 (Nov. 18, 1970).

 

R. Schldgl and F. Helfferich, J. Chem. Phys. 26, 5 (1957).

J. P. Copeland, C. L. Henderson, and J. M. Marchello, A.I.Ch.E. J.
13, bho (1967).

J. C. R. Turner and C. B. Snowdon, Chem. Eng. Sci. 23, 1009 (1968 ).

T. Kataoka, M. Sato, T. Nishiki, and K. Ueyama, Chem. Eng. Japan 33,
59 (1966).

S. Cantor, Reactor Chem. Div. Ann. Progr. Rept., Dec. 31, 1965, ORNL-
5913, pp. 27-29.

 

L. M. Ferris, J. C. Mailen, J. J. Lawrance, F. J. Smith, and E. D.
Nogueira, J. Inorg. Nucl. Chem. 32, 2019-35 (1970).

L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder

 

Reactor Processing No. 6, ORNL-TM-3141 (December 197l), pp. 59-72.

 

L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder
Reactor Processing No. 7, ORNL-TM-3257 (February 1972), pp. L46-52.

 

L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder
Reactor Processing No. 8, ORNL-TM-3258 (May 1972), pp. 97-106.
oo~V MLWND

63.
64.

65.

66.
67.

68.

69.

70.

71.

72-73.
74.

265

INTERNAL DISTRIBUTION

S. E. Beall 38-39. M. W. Rosenthal

M. R. Bennett 40. A. D. Ryon

R. E. Blanco 41, W, F. Schaffer, Jr.

R. E. Brooksbank 42. J. H. Shaffer

K. B. Brown 43, M. J. Skinner

W. L. Carter 44, Martha Stewart

H. D. Cochran, Jr. 45, 0. K. Tallent

F. L. Culler 46. D. B. Trauger

D. E. Ferguson 47. W. E. Unger

L. M. Ferris 48. C. D. Watson

W. R. Grimes 49, J. S. Watson

P. A. Haas 50. A. M. Weinberg

B. A. Hannaford 51. W. M. Woods

J. R. Hightower, Jr. 52. R. G. Wymer

C. W. Kee 53. E. L. Youngblood

J. A. Klein 54-55. Central Research Library
R. B. Lindauer 56. Document Reference Section
H. G. MacPherson 57-59. Laboratory Records

L. E, McNeese 60. Laboratory Records, RC
J. P. Nichols 61. Y-12 Document Reference Section
E. L. Nicholson 62. ORNL Patent Office

J. H. Pashley (K-25)

EXTERNAL DISTRIBUTION

J. A. Accairri, Continental 0il Co., Ponca City, Oklahoma 74601

M. J. Bell, Effluent Treatment Systems Branch, USAEC, Washington,
D. C. 20545 Mail Station Beth. 008

Delonde R. DeBoisblanc, Ebasco Services, Inc., 2 Rector Street,
New York, New York 10006

Norton Haberman, USAEC, RDT, Washington, D. C. 20545

T. R. Johnson, Argonne National Laboratory, 9700 S. Cass Avenue,
Argonne, Illinois 60439

R. D. Pierce, Argonne National Laboratory, 9700 S. Cass Avenue,
Argonne, Illinois 60439

J. Roth, Combustion Engineering Inc., Prospect Hill Road, Windsor,
Connecticut 06095

D, D. Sood, Radiochemistry Division, Bhabha Atomic Research Center,
Bombay 85, India

N. Srinivasan, Head, Fuel Reprocessing Division, Bhabha Atomic
Research Center, Trombay, Bombay 74, India

TIC, Oak Ridge, Tennessee

Research and Technical Support Division, AEC, ORO