F1" EXCHANGE BETWEEN FLUOROCARBONS AND
SOME FLUORINE-CONTAINING COMPOUNDS
THEODORE A. GENS
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
The successful performance of these exchange studies was made
possible by the guidance and aid of Dr. John A. Wethington, Jr. In
addition, he frequently applied his skill in laboratory technique to
overcome some of the difficult problems which otherwise might not have
been solved. The burdens introduced by the ORINS Fellowship arrange-
ment were cheerfully accepted by Dr. Wethington.
Dr. A. R. Brosi of the Chemistry Division of the Oak Ridge
National Laboratory (ORNL) watched over the problems of a radiochemical
nature, and his aid in such matters was invaluable. In addition, Dr.
Brosi readily considered and advised on all the problems which arose.
His careful, scientific approach often led to early solution of these
Dr. E. R. Van Artsdalen, who shared in the supervision of the
early part of this work, was very helpful in planning this research
program and in making arrangements for the design and construction of
the apparatus and instruments.
The personnel of the Fluorine Research Center at Reed Labora-
tory, University of Florida, very generously supplied samples of rare
fluorine-containing compounds. Most of these compounds were pro-
duced by the Electrochemical Process.
The competent aid in the laboratory by D. E. LaValle of the
Analytical Chemistry Division of ORNL made possible the survey with HF
and various inorganic fluorides.
Proton bombardments were planned and supervised in the ORNL
86-inch cyclotron by Dr. J. L. Need of the Applied Nuclear Physics
Group of the Electronuclear Research Division.
MeasuremenU of surface areas were made by the B.E.T. method by
P. K. Melroy of the Special Analytical Services Department of the
Oak Ridge Gaseous Diffusion Plant.
Colorimetric and pyrohydrolysis quantitative analyses of
fluoride ion were performed by the Special Analyses Group of the
Analytical Chemistry Division of ORNL.
Mass spectrometer analyses of fluorocarbons were performed by
J. C. Horton, G. Howell and F. Jones of the Mass Spectrometer Depart-
ment of the Oak Ridge Gaseous Diffusion Plant.
Routine spectrographic analyses were performed by the Spectro-
graphic Group of the Analytical Chemistry Division of ORNL. A
modified Harvey method was employed by M. Murray to analyze for silicon.
In the early search for understanding of the experimental data,
some complicated calculations were coded by E. C. Long for the ORACLE.
Miss B. J. Osborne performed the necessary operations.
Inorganic fluorides were prepared by B. J. Sturm under the
direction of L. G. Overholser of the Chemistry Division of ORNL.
The Instrumentation and Controls Division and the instruments
section of the Chemistry Division and the glass and metal-working shops
at ORNL made essential contributions to this work.
This work was supported by the Oak Ridge Graduate Fellowship
Program of the Oak Ridge Institute of Nuclear Studies. Work done at
the University of Florida was supported by the Office of Naval Research.
The greater part of this work was performed at the Oak Ridge National
Laboratory, which is operated for the U. S. Atomic Energy Commission
by the Union Carbide Nuclear Company.
TABLE OF CONTENTS
LIST OF TABLES ... ....... ....... ......................... vi
LIST OF ILLUSTRATIONS ....................................... vii
I. INTRODUCTION ..........................................
II. APPARATUS .............................................. 7
III. PREPARATION OF FLUORIDE SALTS CONTAINING F18 ........... 26
IV. CALCULATIONS ...................................... 34
V. CHEMICAL AND RADIOCHEMICAL PURITY AMD ERROR ............ 43
VI. GENERAL EXCHANGE STUDIES .............................. 54
High Temperature Survey
Low Temperature Survey
Survey of Hydrogen Fluoride and Various Inorganic
Fluorides for Indications of Exchange with SF6 and
VII. KINETICS OF THE EXCHANGE REACTIONS ............,........ 83
VIII. DISCUSSION OF RESULTS ................................. 91
Fluorocarbon Exchange Mechanisms
Suggestions for Future Work
IX. CONCLUSIONS ................................... ...... 110
BIBLIOGRAPHY ........... .......... .................... 113
LIST OF TABLES
1. Six Methods of Preparing Radioactive F8 ................ 29
2. Error ............................................. 50
3. Data from Studies Having Experimental Difficulties or
Negative Results ....................................... 60
4. Exchange Studies between HF and Some Fluorine-Containing
Compounds .............................................. 79
5. Exchange Studies between Some Inorganic Fluorides
and Fluorocarbons at 5000 ............................... 82
6. Kinetics of Exchange at Constant Temperature ............ 83
7. Heats of Reaction from Survey Studies ................... 92
LIST OF ILLUSTRATIONS
1. Exchange System, Sketch ............................... 8
2. Exchange System with Inconel Reactor Tube ............. 9
3. Constant Temperature Exchanger ..................... 13
4. Exchange System with Pyrex Constant Temperature
Exchanger ............................... ....... 14
5. Solenoid-Activated Circulation Pump ................... 15
6. Exchange System Plus Auxiliary Glass System ........... 17
7. All Nickel System for Preparing Inorganic Fluorides,
Sketch ............................... ........ ..... 19
8. All Nickel System for Preparing Inorganic Fluorides ... 20
9. Inorganic Fluoride Exchange Tube ...................... 21
10. Counting and Temperature Control, Sketch ............. 23
11. Instruments ................. ........................ 24
12. Example of Survey Experimental Data ............. 39
13. Gamma Spectrometer Plot of F18 Radiations ............. 47
14. Pressure-Temperature Behavior of C F above CsF ...... 51
15. Fraction of Exchange Occurring with C3F6 over the
Alkali Fluorides .............................. .... 55
16. Fraction of Exchange Occurring with C3F6 over NaF;
Three Runs ........................................... 57
LIST OF ILLUSTRATIONS Continued
17. Fraction of Exchange Occurring with C4F10 over CsF
and F .................. ... .... .. ....... ..... ........... 58
18. Fraction of Exchange Occurring with C3F6 and C F10
over CsF .. .. .......... .............................. 59
19. Fraction of Exchange Occurring with C3F6 over the
Alkali Fluorides at Lower Temperature ................ 63
20. Fraction of Exchange Occurring with CLF10 over the
Alkali Fluorides at Lower Temperature ................ 64
21. Fraction of Exchange Occurring with (C2F5)20 over
the Alkali Fluorides ................................. 65
22. Fraction of Exchange Occurring with CF4 over RhF
and CsF ....... ................ ....... .. 66
23. Fraction of Exchange Occurring with C3F6 over CsF
by Three Techniques .................................... 67
24. Pressure-Temperature Behavior of CF4 and CF3H over
CsF .............. .... ...................... .. ...... .. 68
25. Pressure-Temperature Behavior of SiF4 Alone and Above
Alkali Fluorides ..................................... 70
26. Results obtained by Passing SIF4 over Alkali Fluorides. 71
27. Pressure-Temperature Behavior of SF6 Alone and Over
the Alkali Fluorides ................................ 73
LIST OF ILLUSTRATIONS Continued
28. Low Temperature Summary Plot .......................... 75
29. Pressure-Temperature Behavior of Argon and Five Gaseous
Fluorine Compounds over the Alkali Fluorides .......... 76
30. Kinetic Studies ................................... 84
31. Kinetic Studies of Exchange Occurring with C3?6
over CsF ............................................. 86
32. Rate Data Taken from Fig. 31 .......................... 88
33. Effect of Duration of CsF Preparation ................. 89
34. Fraction of Exchange Occurring with C3F6 over CsF ..... 94
35. Number of Lattice Layers That Would Be Required to
Account for Values of N Observed in Study AcII ........ 96
36. Second Order Effects with C?6 ........................ 105
The existence of the 112 minute fluorine eighteen isotope was
reported by Snell in 1937. Several studies have been made using this
radioactive isotope to follow exchange reactions. Since F18 is a
positron emitter, the 0.511-Mev. annihilation gamma radiation provides
an excellent means of detecting this isotope. Practically all work up
to the present has been in homogeneous, usually gas phase, systems.
Dodgen and Libby made an early study of rates of exchange in the gas
phase between F2 and HF. They found that little exchange occurred
below 2000 C 2 Above this temperature exchange occurred, but this ex-
change was probably catalyzed by metal fluorides on the wall of the
vessel. Nonexchange was explained by the absence of unfilled electron
levels in the-higher shells of fluorine. Such unfilled levels were
thought to allow formation of intermediates of the general formula HX3
with other halogens and the corresponding hydrogen halides. Rapid
exchange has been observed in these systems.
Katz and co-workers have made several studies using F18
Rogers and Katz studied exchange between HF and several interhalogen
compounds .3 They found that the room temperature exchange reactions
between liquid HF and several liquid interhalogen compounds were
- 1 -
essentially complete in ten minutes while the same reactions in the gas
phase were essentially complete in three minutes. These results made
it appear feasible that an ionic mechanism could be operative in the
liquid phase and that a mechanism postulating an intermediate complex
could explain the gas phase reactions. Essentially zero exchange was
observed between HF and SF6 or CC12P2 and between CIF3 and F2.
Bernstein and Katz studied the gas phase exchange between
interhalogen compounds and fluorine and found essentially no exchange
below 100 4 Above 100 a measurable rate of exchange was observed,
from which it was possible to propose mechanisms for the exchange re-
action. Small but observable exchange between interhalogen compounds
and several metallic fluorides was reported.
Boggs, Van Artsdalen, and Brosi found no exchange in the gas
phase between HF and fluorinated methanes at 500 over one hour.5
Although no systematic studies have been made of heterogeneous
exchange between HF and metallic fluorides, such exchange has been
observed in several cases.'33
Consideration of other systems in which similar exchange
studies could be performed points out the wide applicability and the
advantages of the fluorocarbon-inorganic fluoride system. Exchange
studies between hydrocarbons and inorganic hydrides, using deuterium
and tritium, are feasible but limited to the few stable inorganic
hydrides. Many exchange studies between deuterium and hydrocarbons on
the surface of various catalysts, such as the cracking catalysts, have
been made by Taylor and co-workers and others.6,7 These studies, which
were designed primarily to look into surface effects, did not involve
investigation of isotopic distribution within the crystal. Thus, they
differed fundamentally from the studies reported in this manuscript.
However, a series of papers by Wright and Weller describe a more
analogous system.8 Wright and Weller studied isomerization and hydro-
genation of unsaturated hydrocarbons over BaH2 and Cal2. They attri-
buted the catalytic activity to production of dual metal-metal hydride
sites by removal of hydrogen during evacuation at 200 to 300. Ex-
change between hydrogen and deuterium was also found to be catalyzed
by BaH2 and Cal2. Deuterium over Cal2, carefully evacuated at 200,
approached isotopic equilibrium with the hydride within a few hours.
Diffusion of hydride ions between the interior and the surface was
advanced as the explanation of these results. If any reaction mechanism
as rapid as the one which will be discussed with the fluorocarbon-
alkali fluoride system was operative here, its effect was not apparent
at 2000 with the technique used.
Several heterogeneous exchange reactions between completely9
and partially9'10O11 halogenated hydrocarbons and inorganic halides
have been studied. Blau and Willard observed rapid exchange of
chlorine atoms at room temperature between CC14 or partially chlorin-
ated hydrocarbons and AC13 9 Observations made by Kistiakowsky and
Van Wazer0 in the exchange of CH3Br with BaBr2 and AlBr3 may have
resulted from mechanisms similar to those apparently operative in the
fluorocarbon-alkali fluoride system. They observed activation energies
of 12 kcal mole-1 with BaBr2 and 4.6 kcal mole-1 with AlBr While
there are innumerable inorganic chlorides, bromides, and iodides which
may be used in exchange studies, fluorocarbons are the only fully
halogenated hydrocarbons in which such great numbers of compounds are
stable. A great many partially chlorinated, brominated, or iodinated
hydrocarbons, while stable, do not have the high vapor pressure of
fluorocarbons and cannot be conveniently studied in the gas phase.
Many such systems have been studied in the liquid phase.2 )l2'1314
In none of these studies have observations resembling those discussed
in this manuscript been made.
Winter found that exchange of oxygen gas with some inorganic
oxides involved subsurface lattice ions, and he was able to correlate
his exchange results with the semiconducting properties of the oxide.15
Kolthoff and O'Brien found that quite rapid exchange occurred between
Br2 gas and solid AgBr. The mechanisms involved in such inorganic
systems may resemble those involved in exchange between fluorocarbons
and alkali fluorides.
Fluorocarbons have great thermal stability and resistance to
chemical reaction.17 The dissociation energy of CF4 is reported as
-1 18 -1 a19
130 kcal mole compared with 102 kcal mole for CH." These
properties make fluorocarbons valuable in many unique applications,
but also make it impossible to apply synthesis or degradation re-
actions similar to those normally applied in organic chemistry. To
obtain significant reaction with fluorocarbons, hot tube reactions,
often under high pressure, are frequently resorted to. In hot tube
reactions the nature of the surface usually affects the reaction. It
is common practice to pack the hot tube with an inert material which
supplies a surface for reaction and possibly produces other unknown
Many reactions with fluorocarbons which are very feasible
thermodynamically cannot be carried out, even at high temperatures.
There exists great need for catalysts capable of lowering the acti-
vation energy sufficiently to allow smooth low temperature reactions
producing few products in high yield, both for research use and for
successful commercial application of the unique properties of fluoro-
carbons. Most of the science and technology of fluorocarbons has
developed over the last few years. Starting with a small sample of
fluorocarbon supplied by J. H. Simons in 1941, a group at Columbia
University developed the solid and liquid fluorocarbons needed in the
Manhattan project O' At the same time a different method of fluoro-
carbon production was being developed at Johns Hopkins University.
Since the field is so new, there is very little published work to
serve as a guide in searching for materials to catalyze fluorocarbon
reactions. A study of the exchange of fluorine between fluorocarbons
and inorganic fluorides using F18 offered the possibility of gathering
much information concerning the effect of various salts in exchange
reactions. It was thought that if salts were found which exchanged
with fluorocarbons, information thus gathered would be helpful in the
search for catalysts for use in other fluorocarbon reactions.
Photographs are used to show the instruments and the two vacuum
racks containing the apparatus used in this work. In addition, sketches
are included to aid in interpreting the photographs.
Figure 1 is a sketch showing the essential parts of the system
used to study the exchange of fluorocarbons, SIF4, and SF6 with alkali
fluorides. Figure 2 is a photograph of this apparatus. The multiple
unit clam-type furnace has been removed to show the Inconel reactor
tube. Dimensions of the Inconel portion of the reaction tube were:
length, 13.0"; O.D., 0.375"; thickness, 0.025". A Chromel-Alumel
thermocouple was soldered to the middle of the reactor tube with high
melting silver solder. Copper cooling coils through which water was
circulated were soft soldered on the portions of the reactor tube which
extended out of each end of the furnace. These cooling coils prevented
the melting of the soft solder seal on the sleeves directly beneath
the coils. The sleeves connected the Inconel tube to Housekeeper seals,
and thus to the rest of the system, which was made of Pyrex. Inconel
metal was most satisfactory as a reactor tube material because of its
resistance to reaction at high temperature and its unusually poor con-
ductance properties, which made it simpler to protect the Housekeeper
seals. Fans, one of which is visible in Fig. 2, were also used in
I Furnace I
I Furnace I
A Capillary Manometer
B Heavy Lead Shield
C Scintillation Counter
D Solenoid Activated
Fig.1. EXCHANGE SYSTEM, SKETCH
Fig. 2. Exchange System with Inconel Reactor Tube.
- 10 -
cooling the exit and entrance. A split stainless steel sleeve of 1.250"
O.D., 0.250" thickness, and 7.0" length surrounded the Inconel reactor
tube inside the 8.0" long heating coils in the furnace. The large heat
capacity of this steel sleeve prevented rapid temperature fluctuations.
The Chromel-Alumel thermocouple seen in Fig. 2 below the Inconel tube
was inserted into a hole drilled half-way through the length of the
stainless steel sleeve. This thermocouple activated the temperature
controller, as outlined in Fig. 10. This system was built to be used
at temperatures as high as 10000, and at times was actually used at
temperatures of nearly 7000.
The spiral in Fig. 2 was of 4 mm. O.D. Pyrex. It provided the
flexibility needed to open the system at the reactor tube. The stop-
cock just above the spiral was added before the kinetic studies were
made. A cold finger was located beneath the spiral. As the gas entered
the cold finger it was filtered through a type D sintered Pyrex frit.
This frit and a similar frit above the pump on the left side of Fig. 2
isolated the counter from the reactor tube. A well type Nal scintil-
lation counter beneath the lead shield counted the circulating gas.
The solenoid-activated pump just above the counter, and the flowmeter
above the pump, are described below. The capillary mercury manometer,
part of which is visible on the right of Fig. 2, was connected to the
gas system by a capillary Pyrex line which entered above the flowmeter.
Beside this manometer line was the exit to the auxiliary system. At
the bottom of the exchange system was the tap through which radioactive
- 11 -
gas was withdrawn and returned in the counter calibration procedure
which is described under Calculations.
Figure 3 shows a drawing of the Pyrex constant temperature
exchanger used in the kinetic studies. This exchanger was placed in the
position occupied by the Inconel reactor tube in Fig. 2. Figure 4 shows
the exchanger in position. The vapor bath and furnace have been lowered
to show the coil in which the entering gas was preheated. This 4 mm.
0.D. preheater coil was about one yard in length and held 10 cc., nearly
one-tenth of the volume of the system. A Pt, Pt-10% Rh thermocouple,
which was withdrawn for photographing in Fig. 4, was inserted through
the condenser for the experiments, as shown in Fig. 3. The temperature
of the vapor or ice bath was measured from the thermocouple by a Rubicon
slide-wire potentiometer with which the E.M.F. was estimated to the
nearest 0.0001 mv. A 10-inch Pyrex tube with a 50/50 ground glass joint
(the same size tube and joint as in the vapor bath in Fig. 3), wound
with sheet asbestos and Nichrome wire, served as a furnace when it re-
placed the vapor bath. The temperature controller was connected to this
The circulation pump and flowmeter are shown in Fig. 5. Similar
pumps are described in the literature.23,24 Other satisfactory solu-
tions to the problem of obtaining circulation in similar systems have
been reported.25,26 This pump circulated the gas with very small hold-
up through a complete cycle in about a minute. A small fan behind the
pump kept the solenoids cool. A ground glass piston filled with soft
- 12 -
Legend for Fig. 3
A. Gas Exit
B. Gas Entrance
C. CsF Salt
D. Porous "D" Pyrex Frit
E. 12/30 Ground Glass Joint
G. Water Cooled Condenser
H. 50/50 Ground Glass Joints
I. Vapor Bath
J. Liquid Used for Vapor Bath
- 13 -
Fig. 3. CONSTANT TEMPERATURE EXCHANGER
Fig. 4. Exchange System with Pyrex Constant Temperature Exchanger.
Fig. 5. Solenoid-Activated Circulation Pump.
- 16 -
iron operated inside the solenoids. Current to the solenoids was
turned on and off at the desired rate by an ordinary laboratory stirring
motor equipped with a small cam which operated a microswitch. Several
of these pieces of equipment were operated from a small control box be-
hind the lead shield in Fig. 6. Above the pump a Fisher-Porter
variable-area flowmeter was mounted with Apiezon W wax. Between the
flowmeter and pump was a type D sintered Pyrex frit.
Figure 6 covers the same area seen in Fig. 4 and also shows
some of the manifold and gas handling system.. A gas storage bulb was
attached to the manifold. On the right was the mercury diffusion pump
and the vacuum line leading to the Welch pump beneath the vacuum.rack.
In Fig. 7 are sketched the essential parts of the system which
was used to produce salts containing F18 by exchange with HF. This
same system was used for exchange studies between gaseous fluorine
compounds and BF, and between gaseous fluorine compounds and NiF2,
CuF2 PdF2, and CsF. Figure 8 shows a photograph of the same area.
The soda lime trap to prevent HF from entering the pump is also visible
on the right. The fluorothene tube, in which the products from the
LITR bombardment were placed for exchange with liquid HF, is attached
below the filter in Fig. 8. The Booth-Cromer pressure gage27 allowed
measurement of the vapor pressure of the HF. The diffusion pump, which
is on the left side of Fig. 8, was not used when HF vapor was present.
The nickel tube in which the inorganic fluorides were exchanged
with HF and the nut by which it was attached are shown in Fig. 9. A
Fig. 6. Exchange System Plus Auxiliary Glass System.
~J1 ; ;
- 18 -
Legend for Fig. 7
A. BF Container
B. Threaded Nut for Attachihg Inorganic Fluoride Exchange Tube
C. Thermocouple and Potentiometer
F. Threaded Nut for Attaching Tube Containing Reactor Products
G. Booth-Cromer Pressure Gage
PUMP -< -0-
Fig.7. ALL NICKEL SYSTEM FOR PREPARING INORGANIC FLUORIDES, SKETCH
Fig. 8. All Nickel System for Preparing Inorganic Fluorides.
Xqqmmqpiim -i i
J I i
Fig. 9. Inorganic Fluoride Exchange Tube.
- 22 -
copper washer was used to insure a vacuum tight seal. An inch deep
thermocouple well in the bottom of the tube is not visible.
No reaction of HF was observed to occur on the exposed parts of
this nickel system, except that NiF2 was formed in the exchange tube at
elevated temperatures. The Fulton bellows type valve had inert Teflon
A sketch of the counting and temperature control apparatus is
shown in Fig. 10. The impulses from the photomultiplier tube and small
preamplifier were fed into the linear count rate meter. The counting
rate was recorded on the Brown chart recorder, along with the temper-
ature, which was measured by the Chromel-Alumel thermocouple on the
Inconel reactor tube. A similar thermocouple activated the temperature
controller which controlled the current to the furnace. The temper-
ature controller was not used during the kinetic studies, nor was tem-
perature recorded on the Brown recorder. For these studies the tem-
perature of the vapor baths was determined more accurately with the
Figure 11 shows some of the instruments used in this work. The
bottom instrument on the left is an Atomic Instruments Model 312 high
voltage supply. It supplied high voltage to the photomultiplier tube.
The ORNL Counter Amplifier is directly above. The dial on this instru-
ment gave the percentage of input impulses that were amplified. At the
counting rates used in this work, this percentage remained at essential-
ly 100. The output from the amplifier went to the ORNL Linear Count
EI Z I
A NaI Scintillation Counter with Photomultiplier and Preamplifier
B Amplifier and High Voltage Supply
C Linear Counting Meter
D Chart Recorder
F Temperature Control
Fig.10. COUNTING AND TEMPERATURE CONTROL, SKETCH
--4 A I
.-- S ~;
Fig. 11. Instruments.
- 25 -
Rate Meter, which is the top instrument on the left. This instrument
was designed and built for this work by Edward Fairstein and co-workers.
The ORNL Specification No. is 136, and the Drawing No. is Q-1511. Vari-
ous multiplier and range settings allowed measurement of counting rates
up to 2 x 10 cpm. The counting rate was read from the face of the
instrument and also recorded on the two point, 10 mv. Brown Chart Re-
corder. Temperature of the reactor tube in the survey experiments was
simultaneously printed on the recorder. The temperature range which
the recorder could follow was from 0 to approximately 10000.
At the lower right of Fig. 11 is the Minneapolis Honeywell
Electroline Temperature Controller, which was capable of seeking and
holding any temperature up to 12000. The pointer at the top of the
controller face reads the actual temperature at the Chromel-Alumel
thermocouple. In this photograph, it is reading room temperature,
The Harshaw standard NaI scintillation counter crystal was not
visible because of the lead shielding. This crystal was of the well
type and was mounted in a moisture proof can. The crystal was 1-3/4"
in diameter and 2" in depth and had a well of 3/4" in diameter and
1-1/2" in depth. When the crystal was canned, this well was large
enough to hold the Pyrex bulb of 5 cc. volume. An RCA-5819 photo-
multiplier tube was attached to the NaI crystal.
Another instrument not shown was the General Electric ionization
vacuum gage used to measure pressure between 10 and 103 mm. of Hg.
vacuum gage used to measure pressure between 10 and l0~ mm. of Hg.
PREPARATION OF FLUORIDE SALTS
Fluorine eighteen can be prepared by using slow neutrons in the
reactions Li (n,t) He ; 0 (t,n) F18. Because this is a two step re-
action in which F8 is produced from tritons which in turn are pro-
duced from slow neutrons, high levels of F18 activity cannot be made by
this method. The most readily available source of slow neutrons was
the ORNL Graphite Reactor,28 which produced a slow neutron flux of about
12 -2 -1 29 I)30
1012 neutrons cm2 sec-1 29 The Low Intensity Test Reactor (LITR)3
13 -2 -1
with a slow neutron flux of above 103 neutrons cm-2 sec1 was more
frequently used in this work. Alkali fluoride salts could be prepared
by using LITR bombardments with between 105 and 106 counts per minute
(cpm) of F18 at the beginning of the exchange experiment. Salts with
about 10 times as much F18 activity were prepared by the reaction
F19 (p,pn) F18 in the ORNL 86-inch cyclotron.
The advantage of the higher F18 activity levels produced by the
cyclotron bombardments was counterbalanced by the disadvantage of a
more uncertain cyclotron operating schedule. Therefore, early exchange
work was all based on the LITR and was of necessity carried out at
higher temperature to obtain easily measurable exchange.
The LITR was chosen over the slightly more convenient Graphite
- 26 -
- 27 -
Reactor because of its higher slow neutron flux. In addition experi-
mental results indicated that the distribution of neutron energies was
more favorable in the LITR. An experiment was performed to test this
possibility. Some cobalt-aluminum alloy was included as a slow neutron
monitor31 in the capsules along with the charges in which F8 was to be
produced. After, neutron bombardment, counting rate measurements were
made of both the F18 present and the amount of Co60 produced in the
monitor. This same procedure was followed in both reactors, keeping
all concentrations, bombardment times, and F18 isolation steps (de-
scribed below) as identical as possible. It was found that 43.5 times
as much F8 was produced in the LITR as in the Graphite Reactor, but
that the slow neutron flux was only 10.5 times as great in the LITR.
The fast neutron reaction F19 (n,2n) F18 in the LITR was found experi-
mentally to produce very little F18 and could not account for the
advantage observed for the LITR over the Graphite Reactor. This con-
clusion agrees in general with similar experimental results of Boggs.5
All the work in the high temperature exchange studies used the LITR
as a source of F The main disadvantage of this method, besides the
fact that the F18 activity produced was low, was that nuclei such as
Na which are activated by thermal neutrons, were always present in
small concentrations. This presented problems both in purification and
It was found in attempts to prepare PdF2 containing F18 directly
in the LITR by the reaction F19 (n,2n) F18 that large thermal neutron
- 28 -
cross sections of trace impurities and of palladium made this method
impractical. Volatile radioactive trace impurities were found to con-
taminate the fluorocarbon gas stream during experiments. These
impurities could not be removed easily before the experiments. Radi-
ation originating from palladium created problems in shielding and
handling. This method of preparing salts was abandoned.
Table 1 lists some of the methods that have been used to pre-
pare F The two nuclear reactions used in this work are included
with references to work in which others have used the same reactions.
This table is not a complete summary. The selections were made from
considerations of variety or practicality. References 32 and 33 give
a more complete survey of the many methods by which F18 has been pro-
In the last reaction listed in Table 1, it can be seen that
different techniques have been used to mix the necessary nuclei, Li6
and 016 in the charge which was bombarded by slow neutrons. Consider-
ation must be given to the short range of the triton particle.41 It is
preferable to have the oxygen and lithium in the same molecule, but the
formation of gaseous products or water interfered with the use of such
salts in the present work.
To remove the F18 after neutron bombardment, HF was condensed on
the products, mixed thoroughly, and distilled off through the filter
(Figs. 7 and 8). Lithium salts which contain oxygen react with HF to
produce gases or water, or both. Hydrogen fluoride has a higher vapor
- 29 -
SIX METHODS OF PREPARING
Nuclear Reaction Target Radiation Source Comments
F19 (n, 2n) F18 10-40 mg. NH F;5 LITR LRPa
liquid HF;2 Neutrons from the
HF, KHF2 cyclotron reaction
Li6 (d, n) Be7
F19 (, n) F18 KHF2 34 Betatron 48 or 84- HRPb
F19 (p, pn) F18 10 mg NaF;5 RNL 86-inch HRP, great
AIF 35 cyclotron quantities
016 (, pn) F18 H20;36 Pb037 UCRL 60-inch Erratic
F19 (d, t) F18 NaF38 Cavendish Threshold
cyclotron at 6 Mev.
LiF and Al20 5,34
LiCO39 LiNO 33940
a. Low radiochemical purity
b. High radiochemical purity
- 30 -
pressure than water and slow distillation leaves the water behind. How-
ever, to avoid the possibility of entrainment of water it was preferable
not to use the salts which reacted to give water. Experimental procedure
was also simpler if there were no gaseous products to remove. Because of
the small particle size of less than one micron, it was found that Linde
B Alumina worked satisfactorily when mixed with LiF. High purity of
this alumina was also a very desirable property. Analysis of a sample
of this alumina put through a dry run with HF showed some fluoride ion
to be present. It was thought that this result was caused by strongly
adsorbed HF rather than by reaction. Nevertheless, some HF residue was
always discarded in the distillation procedure to insure that no
moisture was distilled off.
At least a 2:1 molar excess of this HF was condensed onto the
fluoride salt in the fluoride exchange tube in Fig. 9 and the tube was
warmed slowly to 700, with the exit valve closed. After allowing about
fifteen minutes for achievement of isotopic equilibrium, the HF was
pumped off through the soda-lime trap. Pumping was continued while the
fluoride salt was heated, usually at 4000. With the alkali fluorides,
increasing difficulty in decomposing the bifluorides was encountered in
progressing through the group from LIF to CsF. With CsF a temperature
between 5500 and 6000 was maintained over at least one-half hour, after
which the CsF was found to have sublimed to the cooler part of the re-
actor tube near the top. Previous work has indicated that even after
this procedure some HF may have been present in the CsF.42 Flushing
- 31 -
with argon at high temperature was required to remove the last of the HF.
Alkali fluoride salts were prepared by cyclotron bombardment for
the low temperature exchange studies and the kinetic studies. Between
0.3 and 0.5 gram of dry alkali fluoride was packed into cylindrical
tubes made of 2S aluminum. These tubes were of 0.250" O.D., 0.0045"
thickness, and 2.75" length. The ends of the tube were flattened,
folded over once, and crimped tightly shut in a vise. The tubes were
flattened and placed in the cyclotron target head behind a 2S aluminum
window of 0.012" thickness. Cooling water was circulated at 40 pounds
per square inch gage pressure at a sufficient rate to allow only a one
degree temperature rise in the cooling water during bombardment. The
protons at the target had an energy of 21 Mev. The beam current was
approximately 80 microamperes. It was possible by this method to pre-
pare one curie of 1 in 15 min. A general description of the cyclotron
is available, as is a discussion of radioisotope production rates44
and a description of target heads 5 quite similar to those used in pro-
The aluminum tubes were opened behind a lead barricade by
quickly ripping them into two pieces with long handled pliers. The
alkali fluorides were washed from the two pieces into a beaker of water.
Alkali fluoride carrier was added, and the solutions were filtered
through Whatman No. 50 filter paper. The water was removed by evapo-
ration on a hot plate. The hot, dry alkali fluorides were placed in
hot weighing bottles to prevent adsorption of atmospheric moisture.
- 32 -
This precaution was not required for LIF and NaF. The weighing bottles
were stored in a desiccator. The weight of a portion of alkali fluoride
removed for an experiment was determined by difference. The portion re-
moved was quickly placed in the exchange system and evacuated.
From this point on the fluoride salts prepared by either method
were handled in the same manner. Evacuation was continued as the salts
were heated. For the survey experiments the salts were heated to approxi-
mately two-thirds of their melting points. The CsF was heated to 4000
for the kinetic studies. These temperatures were maintained until the
pressure as read on the ionization vacuum gage was reduced to 10 mm.
Dry argon was circulated while the alkali fluoride was kept at high
temperature. If any radioactivity was observed in the argon, it was
pumped out and the process was repeated until no more activity was de-
tected. The system was then thoroughly evacuated and brought to the
temperature desired for the start of the exchange study. In the kinetic
studies, the argon was not monitored by circulation through the counter,
but was admitted and pumped off the CsF three times to insure a clean
Besides the alkali fluorides, CdF2, CrF2, PdF2, MgF2, NiF2 and
CuF2 were prepared by exchange with HF. Palladium difluoride showed
very little tendency to exchange with HF. By use of the HF with high
specific activity described in the survey with HF, it was possible to
prepare a gram of PdF2 with over 10 cpm of F Many fluorides have
been found to be quite soluble in HF.46 This technique probably can be
used successfully with a great many inorganic fluorides to incorporate
F The technique of direct bombardment by protons also appears to be
applicable to many inorganic fluorides.
The amount of exchange in all experiments described as survey
or kinetic studies was calculated by the following equation:47
N = Net fraction exchanged = gas total (1)
18 1 18
F total = sum of F counting rates, gas and solid phases,
Fgas = weight of fluorine, gas phase,
Ftotal = sum of weights of fluorine, gas and solid phases.
When N equaled one, the F8 atoms were randomly distributed
among all fluorine atoms in the two phases.
The following experimental values, which varied, were used to
solve this equation:
Temperature of the Inconel reaction tube, C = Ti,
Pressure of fluorocarbon, mm. of Hg = Pi
Initial counting rate of the alkali fluoride, cpm = (AFCR)o,
Counting rate of the background, cpm = (Bgd)i,
Recorded counting rate of fluorocarbon in the
counting chamber, including (Bgd)i, cpm = (RCR)I,
Counting rate of total gas when counted at the
same geometry as the salt, cpm = (CRTG)i,
- 34 -
- 35 -
Factor to correct (RCR)i-(Bgd)i to (CRTG)i = (CF)i.
All counting rates (RCR)i, (AFCR)o and (CRTG)i were corrected
for decay of F by the exponential law48 R/R = e4t where\ =
0.693/tl/2 and tl/2 = 112 minutes.
Other values, constant within a particular run, needed to solve
Equation 1 were:
Pressure of gas in the exchange system at 2980 K,
mm. of Hg = P298,
Volume of the exchange system, cc. = Vs
Volume of the counting chamber, cc. = Vcc,
Weight of fluorine in one mole of gas, g. = WFgas
Weight of fluorine in the salt, g. = Fsalt'
298 K = T.
The terms in Equation 1 became:
Fgas, g. = (WF gas)(moles of gas)
= (WF gas)(P298Vs/RT), by the ideal gas law.
Total, g. = Fgas + Fsalt
= (WF gas)(P298Vs/RT) + Fsalt
F8gasP cpm. = (CRTG)i = (CF)i (RCR)i (Bgd) .
18 18 18
F total' cpm* = F gas + F alkali fluoride
= (AFCR) corrected for decay, since all F l
was initially in the alkali fluoride.
- 36 -
N= (CF)i IRCR)i (Bgd)i] / (AFCR)0 (2)
(WFgas) (P298Vs/RT)/(WFgas )(P298s/RT) + Fsalt
Radioactive C3F6 was counted as a liquid in an external well
type scintillation counter very similar to the one in the gas line to
give the quantity (CRTG)I. This radioactive C3F6 was placed in the ex-
change system and (CF)i was determined from the observed values of (RCR),
- (Bgd)i. For correct evaluation of N, it was necessary that (CF)i be
determined very similarly to (AFCR) Therefore, (AFCR) was determined
by counting aqueous solutions of the alkali fluoride in the same ex-
ternal counter at the same geometry, at approximately the same counting
rates, and applying counting loss corrections. The empty reactor tube
of Fig. 2, or the empty exchanger of Figs. 3 and 4, was at room tempera-
ture when (CF)i was determined. When the reactor tube or exchanger was
warmed, (CF)I decreased and Pi increased.
Consideration of (CF)i shows that it is a function of the ratio
of the total number of moles of gas present in the system to the number
of moles of gas in the counting chamber. While the reactor tube or
exchanger was at room temperature, this ratio, which will be called
ns/ncc, was equal to the ratio Vs/Vcc, by the ideal gas law. Thus the
following evaluation of (CF)i, as a function of Pi as Ti changed, could
(CF)298 = k (ns/ncc) = k (V/Vc) (3)
(F8cr dr )EC
where k = 18 18 rEC
(F cr dr GLC1
- 37 -
and (F cr 18dr )EC = the ratio of counting rate to disintegration rate
in the external counter,
crand (Fr GLC = the same ratio in the gas line counter.
(CF)i = k(ns/ncc)Ti = k(ns/k Pi) (4)
where ns = P298 Vs/RT (5)
and k = Vc/RT (6)
since the counter chamber temperature stayed approximately at T, and the
only systems considered were those in which ns stayed constant. Dividing
5 by 6 gave
ns/k = V298 cc7)
Substituting (7) in (4) gave
(CF)i = (k V/Vcc)(298/i (8)
Substituting (3) in (8) gave
(CF)I = (CF)298 (P298/P (9)
Thus, (CF)I could be calculated very simply from the experimentally de-
termined (CF)298. From known values of V and Vs, Equation 3 gave k
from the experimental value of (CF)298. The first calibration gave a
value of 25.27 for (CF)298 and k was calculated to be 1.38. In a
similar calibration in the kinetics chapter, where Vs and (CF)298 were
both different and small adjustments had been made in the count rate
meter, k was calculated to be 1.31. The term k did not enter directly
into calculations, but this agreement showed that the values for k, Vs,
Vc and (CF)298 were all approximately correct.
- 38 -
For the high temperature experiments, the volume of the system,
86.3 cc., was less than in the calibration experiment where a cold
finger had been added. By Equation 3, where Vc = 5,
(CF)298 = 1.38 x 86- = 23.82.
This value of (CF)298 was used for the high temperature studies and the
first few low temperature studies.
It was not necessary to have (CF)298 in Equation 9 if it was
known at some other temperature, since:
(CF)2 = (CF)298(P298/P2),
(CF)1 = (CF)298(P298/P1),
(CF2/CF1) = (P1/P2),
(CF)2 = (CF)1 (PI/P2). (10)
Equation 10 was applied in the kinetic studies.
Usually the background (Bgd)i was substracted from the recorded
counting rate (RCR)i directly as these values were obtained from the
recorder chart paper. Therefore, this new quantity, the net recorded
counting rate, (NRCR)I, and also the value for (CF)I from Equation 9
were introduced into Equation 2. Rearrangement yielded:
N = > (11)
where C = 298 WFgas) P298Vs + RTFsalt]
w(WF as) Vs
Figure 12 shows an example of the experiment data to which
Figure 12 shows an example of the experimental data to which
10 2It~2 30
Fig. 12. Example of Survey Experimental Data.
Equation 11 was applied. The heavy dark line on the left in Fig. 12
shows the temperature of the reaction zone. Times were written on the
left margin with temperature in the adjacent column. For the first
55 minutes a constant temperature of about 480 was maintained. For the
next 40 minutes the temperature was raised at about 5 per minute. The
temperature was then again held constant for 40 minutes at about 2560.
The counting rate record starts on the right of the chart. A full
scale value of 10,000 cpm (10 K in Fig. 12) was sufficient for 62
minutes; 20,000, 67 minutes; 100,000, 98 minutes; 200,000, the re-
mainder. A straight line was drawn for the slightly changing back-
ground. Data for two crude constant temperature studies and one survey
run were obtained from this chart. The survey run is discussed in
connection with Fig. 19 as Run C14a. The constant temperature results
are tabulated-as RbF, C3F6 in Table 6. For the kinetic studies, no
temperature data were recorded as in Fig. 12, and the speed of the re-
corder was increased so that the distance covered in 10 minutes in
Fig. 12 was covered in one minute.
As an example of actual calculations as they were made from
chart data, the (NRCR)i of 7050 at 810 and 59 minutes from Fig. 12 is
substituted along with other experimental data in Equation 11 to give
the following value for N:
N (7983)(7050) = 8.686 x 10-4
(287)(2.259 x 108 )
The experimental values used in calculating C were as follows:
- 40 -
- 41 -
C = (25.27) E114)(283)(91.3) + (1.86 x 107)(.0184 = 7983
The particular value calculated above appears in Fig. 19 at a value of
2.825 for 1000/T, oK.
By use of Equation 11, only the exchange visible to the counter
could be evaluated. Exchange which is invisible to the counter also
occurs. Once an F18 atom enters the gas phase, the probability that it
will re-exchange with a fluorine atom in the alkali fluoride is as great
as is the probability that any other fluorine atom in the gas phase will
exchange with a fluorine atom in the alkali fluoride. The results
obtained by Equation 11 had to be correlated with the mechanistic
picture which developed during this work. The following paragraphs
discuss this correlation and its effect.
As will be seen in the discussion of results, all the evidence
indicated that both a fast and slow exchange reaction occurred. The
fast reaction soon led to a pseudo-equilibrium between the gas phase
and a portion of the alkali fluoride. The calculated magnitude of N
was not an actual measure of the amount of exchange, since exchange con-
tinued after random distribution by the fast reaction without changing
the F8 concentration in the gas. Thus, N should be considered as the
fraction of the alkali fluoride which was in equilibrium with the gas
phase when the fast reaction is being considered. No corrections for
exchange invisible to the counter needed to be made on the results of
Equation 11 as long as the fast reaction was being considered, since
- 42 -
equilibrium concentration of F18 was being measured. Results of the
survey studies will be shown to represent almost entirely the fast ex-
Therefore, N, which could only be defined as net fraction
exchanged early in this work, was redefined after analysis of all
observations as fraction of alkali fluoride in exchange equilibrium
with the gas phase. Use of N was continued with its original meaning
because the necessity of the second definition did not become apparent
until after the kinetic studies were made.
The same treatment of the data was applied with SIF4 and SF6
because of its convenience and because of the desirability of comparing
these results with those of the fluorocarbons.
In the kinetic studies, the determination of the rate of the
slow reaction required a different treatment of the exchange results
calculated by Equation 11. To calculate the rate of exchange, a cor-
rection had to be added to the slow increase in counting rate observed
in the gas to account for the exchange which removed F18 from the gas
phase. This same problem is always faced in kinetic studies involving
isotopic exchange, and the standard treatment of McKay was used. At
first consideration, it appeared that McKay's treatment, which was worked
out for systems in which the tracer was distributed homogeneously
throughout the different phases, could not be applied to gas-solid ex-
change experiments. The reason that it could be used in this work is
explained in connection with proposed mechanisms in the discussion of
CHEMICAL AND RADIOCHEMICAL PURITY AND ERROR
Chemical Purity of Gases and Salts
All fluorocarbons were treated as follows to remove air,
moisture and other impurities. The gases were distilled from storage
bulbs into a liquid nitrogen-cooled trap. This trap was open to the
mechanical and diffusion pumps to remove any traces of air. After the
fluorocarbons passed twice through a tube filled with P205, a Regnault
molecular weight determination5 was made. This procedure established
that all molecular weights were within 1% of the theoretical value.
Perfluoropropene was purchased from Peninsular Chem-Research,
Inc. The 6nly suspected impurity was C3F H. Mass spectrographic
analysis of this CF6 and also of C3F prepared by a similar process
in the laboratory showed an estimated 0.3 mole percent of C F H. No
other impurities were detected. It seemed unlikely that this con-
centration of C F H could have any serious effect on fluorine exchange,
at least at the lower temperatures used. No attempt was made to remove
Perfluorobutane was prepared in the laboratory by the Electro-
chemical Process51 starting with C4H9COOH. Acidic products were re-
moved by passage through cold aqueous base. Preliminary single plate
distillation was performed in an open system, and material boiling near
- 43 -
- 44 -
0 was removed. After removing air and drying with P205 this fraction
was carefully distilled through a column of approximately 30.plates.
The material used in these studies boiled between -2.250 and -1.60.
Infrared analysis of C4F10 prepared by a similar procedure indicated
a mixture of the isomers perfluoroisobutane and perfluoro-n-butane.
Perfluorodiethylether, CF4, CF3H, and C2F6 were all prepared
by a procedure similar to that described for C4F10, except that only a
single plate distillation was required to isolate a pure sample of the
low boiling CF4. Oxygen difluoride was removed from these gases by
passage through a buffered KI solution. Infrared analyses indicated no
impurities in these materials. No impurities were detected in a mass-
spectrographic analysis of the C2F6.
The compounds SIF4 of 99.5% purity and SF6 of 99.0% purity were
used as purchased from the Matheson Company, except that air and other
low boiling materials were removed by the same process described above,
and SF6 was dried by passage through a tube packed with P205. Since
SiF4 reacts on contact with moisture, the drying step was not required.
A single plate vacuum distillation was performed in the vacuum system
with both SF6 and Si4, and the first and last portions of the dis-
tillates were discarded.
Baker Analyzed LIF and Fisher Certified Reagent NaF were used
as purchased. Spectrographic analysis indicated that the LIF contained
as much as 1% calcium. Baker and Adamson ACS Reagent Grade KF was first
dried and then sublimed before use. Rubidium fluoride was obtained
- 45 -
from General Chemical Company. The reported impurities were about 1% K
and traces of the other alkali metals. Cesium fluoride was obtained
from the A. D. Mackay Company. The reported impurities included 0.12%
Rb and traces of the other alkali metals. Spectrographic analysis of
this CsF also showed as much as 0.1% Al and traces of Si. A sample of
the very pure CsF used by Bredig, Bronstein, and Smith was obtained
for the last kinetic study. This CsF also showed traces of Al by
spectrographic analysis. Similarity of experimental results in the
last kinetic study to all previous results indicated that impurities
were not the cause of the unique observations in this work. Upon
heating at 4000 under vacuum, RbF and CsF changed color from pure
white to grey. This discoloration appeared to be as intense when the
very pure CsF was used as when the CsF purchased from the A. D. Mackay
Company was used. Support for a defect-caused mechanism was found in
this discoloration. Analysis of samples of all the alkali fluorides by
the spectrophotometric method5 both before and after exchange with
fluorocarbons showed no change in composition from the theoretical
values. The standard pyrohydrolysis method53 of analysis for fluoride
was found to be unsatisfactory for all the more volatile fluorides
(KF, RbF, and CsF).
Anhydrous nickel fluoride was prepared by hydrofluorination of
Special Reagent Grade Baker and Adamson NiC12 6H20 at 400 to 5000.
Anhydrous cupric fluoride was prepared by hydrofluorination at 250 to
300 of the Baker and Adamson ACS Reagent Grade, partially hydrated
- 46 -
product. Hydrofluorination produced salts which were spectroscopically
pure in any materials which formed volatile fluorides. Anhydrous chro-
mous fluoride was prepared by the thermal decomposition of (NH4)3CrF6
in a stream of hydrogen at 700. The (NH4)3CrF6 was prepared by the
reaction of Fisher Scientific CrF3 3-1/2 H20 with NH4F at 1000. A
mixture of PdF2 and PdF3 was prepared by direct fluorination of spectro-
scopically pure PdC12 at 4000. An attempt was made to convert the mixture
completely into the trifluoride by dissolving in excess BrF3 and dis-
tilling off the Br2 and BrF3. The resulting addition compound
PdF3 BrF3 was found to decompose at 200 under vacuum to yield PdF3,
but X-ray analysis showed that the decomposition was incomplete. Heating
the impure PdF3 to 4000 in the presence of HF, followed by an additional
hour of heating at 4000 under vacuum, yielded a product which gave the
theoretical fluoride analysis for PdF2 by the pyrohydrolysis method.53
In early attempts to prepare alkali fluorides by using LITR bom-
bardments as a source of F it was found that other radiations besides
those of F were present in the alkali fluorides. The filter shown in
Figs. 7 and 8 was added to the apparatus. This filter was packed with
a matting made of fine nickel wire which effectively stopped radioactive
materials from contaminating the product. Investigations of purity
were made with the ORNL 60-channel gamma spectrometer.54 Figure 13
shows a typical plot of counting rate against channel number for pure
Fig.43. GAMMA SPECTROMETER PLOT OF F18 RADIATIONS
- 47 -
- 48 -
F8 radiations. It was deduced from such a plot and the known amount
of Na24 radiation in the reactor irradiated sample that there was less
than one ppm contamination of the product by material from the original
sample after installation of the filter. Background variations did not
allow any closer evaluation of contamination. Many plots similar to
Fig. 13 were made after installation of the filter, and no evidence of
radiochemical contamination appeared.
In Fig. 13 the very small peak at 1.0 Mev was caused by the
occurrence of two gamma rays of 0.51 Mev being registered by the
spectrometer simultaneously as a single radiation. At first, the peak
at 0.7 Mev was not understood. To insure that this peak was not the
result of radiochemical impurity, further studies were made. It was
found to have a 112 minute half life and therefore was almost certainly
associated with F Since the 180 backscattered Compton radiation
from a 0.511-Mev gamma ray would have an energy of about 0.20 Mev and
would be in coincidence with the primary forward gamma of the annihi-
lation pair, the 0.7-Mev peak could result from this scattering process.
In the case of LITR-produced F18 the activity was always associated
with approximately a gram of salt which could act as a back-scatterer.
When measurements were made with high specific activity F18 produced
on the cyclotron, the 0.7-Mev peak was absent. It was possible to pro-
duce a 0.7-Mev peak by placing the high specific activity source be-
tween a scatterer and the scintillation spectrometer detector. It
was concluded that the fluoride salts containing F8 prepared with this
- 49 -
system were of high radiochemical purity.
Less danger of radiochemical contamination existed in the pro-
duction of F18 with the cyclotron, since the impurities have cross
sections comparable with the cross section of F19 under proton bombard-
ment.55 Thus, radiochemical contamination was approximately proportional
to the concentration of the bombarded nuclei, and by using moderately
pure salts, the specific activity of F8 for a few hours after bombard-
ment was such that other radiations could not be detected. This
technique also produced radioactive isotopes from the cations, but the
half lives were such that their radiations caused no difficulty in the
case of the alkali fluorides. For example, with a sample of LiF in
which the total F18 activity was over 110 cpm after bombardment, a
counting rate of about 105 cpm from the 53 day Be isotope was detected
after the F18 had decayed. Several half-life studies on the radioactive
gas, after exchange with these alkali fluorides, indicated that only
F radiations were present. Half-life determinations were also made on
the gases after exchange with salts prepared by the technique based on
the LITR, and only F18 radiations were observed.
The following table presents the most important sources of
error which entered into the terms in Equation 11. These estimates
were made on the typical example of calculations, which accompanied
Fig. 12. The first error in Table 2 has the greatest effect on N.
This error has a small effect on the precision within a particular run.
- 50 -
Term Limits Source
(AFCR)o t 2% tl/2, over 4 hr.
(NRCR)i + 1% Statistical fluctuation of count
P298' Pi 1% Isochore plots obtained with
capillary Hg manometer
t 1% Temperature uncertainty
Vs + 2% Estimation for loss of
Fsalt 1 Weighing
C t 1% P298 Vs' Fsalt
N t 3% Accumulated error
Most systematic errors were kept small by the technique of
comparing counting rates of salts and gases on the same external counter.
Other potential sources of error, not considered significant in any of
the data presented, are discussed below.
The higher boiling materials C4F10 and (C2F5)20 adsorbed suf-
ficently on the counting chamber walls to cause experimental difficulty.
However, a very slight increase in the counting chamber temperature,
with increasing Ti, completely removed this effect with both materials.
Figure 14 demonstrates that much adsorption occurred on the system walls
270 I I
100 200 300
Fig. 4. PRESSURE-TEMPERATURE BEHAVIOR OF C ABOVE
Fig. t4. PRESSURE-TEMPERATURE BEHAVIOR OF C4FO ABOVE CsF
- 52 -
and on the salt with C4Fi at low reactor chamber temperatures. There-
fore, no data is presented for CF10, and (C2F5)20 below a reactor
chamber temperature of 270 although these materials exchanged readily
at 00, the lowest temperature used.
Heating rate was approximately 50 min.-1. Deliberate variation
of this rate indicated that results were not appreciably affected.
Some background variation almost always occurred during low
temperature runs. Figure 12 presents a typical example. This variation
was not significant.
There was some possibility that all HF was not removed,
especially with CsF, in the high temperature survey work. Agreement
in results from the same systems between low and high temperature
surveys indicates that there was no error from this source.
The rate of gas flow was constant at about one cycle per minute.
Determination of the rate of the slow exchange process was not hindered
by this rate of flow. The fast exchange process was probably delayed
in its approach to equilibrium. No error was introduced in the kinetic
studies, since only equilibrium data for the fast reaction were re-
ported. A small delay must have been involved in following N during
survey runs, but this small delay was common to all runs.
In the kinetic studies, the temperature of the constant tem-
perature bath was determined with a precision of + 0.1 by a Pt,
10% Pt-Rh thermocouple. Calibration by ice baths and steam baths
showed this thermocouple to be within t 0.1. of the true temperature
at these two points. The most favorable conditions were chosen and
the best techniques developed in the survey studies were used to keep
all errors at a minimum.
GENERAL EXCHANGE STUDIES
High Temperature Survey
At first these studies were performed on various fluoride
salts in order to find fluorides capable of exchanging with fluoro-
carbons at moderate temperatures. In exchange studies with C3F6 up
to 5000 and 340 over CrF2 and CuF2, respectively, these fluorides
produced few counts in the gas phase beyond what might be attributed
to statistical variation in background. Sodium fluoride produced
easily detectable counts with C3F6 at 4000. Further experiments in
both the high and low temperature surveys were confined to the alkali
Both C3F6 and C4F10 were passed over CsF, KF, NaF and LiF in
the apparatus shown in Figs. 1 and 2. Temperature was increased at
the rate of 50 per minute. With C F10, the counting rate was found to
be sufficient for accurate calculations only with CsF and KF and at
The results were always found to be almost linear up to large
values of N when the logarithm of N was plotted against reciprocal
temperature. Figure 15 shows such plots for C3F6 on each of the four
fluorides studied. Curvature started between N values of 0.2 and 0.4
- 54 -
- 55 -
1.0 I.1 1.2 1.3 1.4 1.5 1.6 4.7
1000/ T, K
Fig.45. FRACTION OF EXCHANGE OCCURRING WITH
CGF OVER THE ALKALI FLUORIDES
- 56 -
with CsF and increased as the run was continued. Some curvature was
observed with NsF at about the same values of N as with CsF. In Fig.
16, three runs with CF6 and NaF were compared. The fact that the
lowest points fell below the line on all three runs is significant for
reasons which become apparent when results are discussed in connection
with Fig. 36.
Figure 17 shows the behavior of C4F10. Low temperature survey
plots included on the same graph showed general reproducibility. The
CsF curve was to the right of the KF curve, as it was in Fig. 15. In
Fig. 17 the CsF plot continued to be linear past the temperature where
curvature started with C3F6. Values of N were lower at comparative
temperatures with C F10 than with C3F6, as Fig. 18 illustrates. Curva-
ture appeared to accompany large values of N and was not affected by
Wide variations in concentration of either phase did not pro-
duce any noticeable change in results. This is shown by Fig. 16, where
approximately the same weight of NaF was used in all three runs, but
CF6 pressures at room temperature varied from about 200 to 600 mm.
Runs which yielded fragmentary information, because of low
counting rate or counter chamber contamination, were summarized in
Low Temperature Survey
Four fluorocarbon gases, CFg6, C F0, (C2F5)20, and CF4 were
- 57 -
4.0 1.1 1.2 1.3 4.4 1.5
4000 / 7; K
Fig.46. FRACTION OF EXCHANGE OCCURRING WITH C3F
OVER NaF; THREE RUNS
- 58 -
I I I I I I
I I I I I I I I
4.4 4.2 1.3 1.4 1.5 4.6 1.7 1.8
1000 / T0K
Fig.47. FRACTION OF EXCHANGE OCCURRING
WITH C4F oOVER CsF AND KF
- 59 -
I I I
I I II I I I I
1.0 4.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8
1000 / T, K
Fig.18. FRACTION OF EXCHANGE OCCURRING WITH
WITH C3F6 AND C4F40 OVER CsF
- 60 -
DATA FROM STUDIES HAVING EXPERIMENTAL
DIFFICULTIES OR NEGATIVE RESULTS
Run No. and Temperature at
Run No. and
Material Which Exchange Results
Was First Observed
14 No exchange up to 3250. Activity
CuF 7 ?6 in CuF^ was such that less than
C2- 6 10 exchange could not be detected.
Cr19 300 Less than 8% exchange at 4250
cKF(c 2 About 0.05% exchange at 600.
C4a Less than 5% exchange at 6250. Back-
KF-(C )0 ground would not return to normal.
(2 520No evidence of decomposition.
C8a 0.4% exchange at 4500. No decom-
CF-(C) position. Background would not re-
-(25) 2 turn to normal.
iC5a 325 Background would not return to normal.
C6a 3100 0.5% exchange at 4470. Background
LiF-SF6 would not return to normal.
C7c 350o Less than 0.2% exchange at 4500
K.-SFG6 Background would not return to normal.
ClOc 4000 Background would not return to normal.
No exchange up to 3500.
- 61 -
TABLE 3 Continued
un o. Temperature at
Materials Which Exchange Results
a Was First Observed
NF3bSF No exchange up to 2800.
C4b 600 Less than 5% exchange at 6500.
KF-CF10 Decomposition started at 600.
C13a 0.04% exchange at 3000. Background
NaF-C F10 would not return to normal.
NaF-Cb Less than 0.2% exchange at 300.
C16d 2900 Decomposition started at 2900
CsF8a 250-3000 Less than 0.03% exchange at 4000.
S-C2g No exchange up to 500.
C12b Room Physical adsorption interfered with
RbF-(CF3 )3N temperature calculations.
44-111 5000 Pressure dropped throughout the
KF-C1CF=CF 2 experiment.
39-107 4750 Pressure increase occurred before
LiF-CF SF5 counts were detected.
Pressure increased slowly throughout
- 62 -
passed over the five alkali fluorides in the same apparatus used in the
high temperature survey. Figures 19 through 22 show the results, which
were very similar to those of the high temperature survey. In Fig. 19
the same relationship between different fluorides observed in Fig. 15
occurred again. Two runs with RbF showed good reproducibility. The
LiF prepared with the cyclotron contained some volatile material, possi-
bly a complex fluoride. By preparing the LiF very carefully, results
obtained with C4F10 and (C2F5)20 are regarded as dependable.
In Fig. 20, the general conclusion can be drawn that CsF and
RbF exchange better with C F10 than does KF or LIF. With (C2F5)20,
in Fig. 21, CsF is indicated as being superior to RbF, which in turn is
superior to LiF.
Few points were obtained with CF4, since high temperatures were
avoided and detectable exchange did not occur until at about 3000 with
CsF and at nearly 4000 with RbF. Figure 22 shows similar slopes with
these two salts, but CsF appears to exchange better than RbF. The
slopes of Fig. 22 are greater than with the other fluorocarbons.
Figure 23 compares high and low temperature CsF-C3F6 runs
(AC I and AC II are discussed in the kinetics studies). This figure,
plus a further comparison between high and low temperature runs, dis-
cussed below in connection with Fig. 36, indicates that similar pro-
cesses take place and that the different methods of preparing fluorides
yield similar results.
The curve for CF H in Fig. 24 shows an inflection just under
- 63 -
+ CsF -
+\- + C0b -'
1.4 1.3 1.5 4.7 1.9 2.1 2.3 2.5 2.7 2.9 3.1 3.3
1000 / 7; K
Fig.49. FRACTION OF EXCHANGE OCCURRING WITH C3F6 OVER
THE ALKALI FLUORIDES AT LOWER TEMPERATURE
- 64 -
4 x4 0-3
I I I I I I I I
I I I I I I I I
4.1 1.2 4.3 1.4
1.7 1.8 4.9
Fig. 20. FRACTION OF EXCHANGE OCCURRING WITH C4 F40
OVER THE ALKALI FLUORIDES AT LOWER TEMPERATURE
- 65 -
I I I I I I I I I
4 x 4 -2
4 x 10
1.0 1.1 1.2 1.3 1.4 1.5 1.6 4.7 1.8 1.9 2.0
1000/ T oK
Fig. 24. FRACTION OF EXCHANGE OCCURRING WITH (C2F5)2
OVER THE ALKALI FLUORIDES
I:I I I I I*
- 66 -
4000 / TOK
OF EXCHANGE OCCURRING
OVER RbF AND CsF
4 x 0-4
- 67 -
I I I I
4.0 1.2 1.4 1.62.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6
1000 / 7 K
Fig. 23. FRACTION OF EXCHANGE OCCURRING WITH C F6
OVER CsF BY THREE TECHNIQUES
- 68 -
SI I l I l I I I I
\ 0 D
0 L F-
" t O 'O (D o O 0 D Io 0
- 69 -
3000. As listed in Table 3, radioactivity was first observed in the
gas phase at about this temperature with the CF3H-CsF system. Tri-
fluoromethane is known to have thermal and chemical stability approach-
ing that of CF 456 The CF4-CsF system also produced detectable radio-
activity in the gas phase at 3000. It therefore appears possible that
a similar process was involved in both cases. With CF3H, decomposition
must have occurred to yield fewer molecules in the gaseous state, while
CF4 was not transformed into new materials.
When SIF4 was passed over the alkali fluorides, decreasing slopes
of the isochores were observed. Figure 25 shows that with LiF this slope
did not differ appreciably from that of SiFt alone, but the slope de-
creased at higher temperatures with KF, RbF, and CsF. One run with KF
was followed beyond the inflection point, and the pressure continued to
decrease. The quantity N could no longer be easily determined after a
significant quantity of SIF4 was removed from the gas phase. Therefore,
in Fig. 26, the values of N go only up to the temperatures where devi-
ations in the pressure-temperature curves were observed. In Fig. 25,
it can be seen that various starting pressures were used. The two runs
with KF had nearly equimolar quantities of SiF4 but Run C4c had nearly
twenty times as much KF as did Run C7b. This variation of concentration
appears to have separated the curves from these runs in Fig. 26. The
run with LIF had molar concentrations of SiF4 and salt similar to those
of Run C4c with KF, and the results of these two runs in Fig. 26 are
quite similar. A fluorocarbon run was included for comparison in
*- e---**-- .CsF
100 200 300 400 500
Fig.25. PRESSURE-TEMPERATURE BEHAVIOR OF SiF4 ALONE AND ABOVE
100 200 300 400 5C
Fig. 26. RESULTS OBTAINED BY PASSING SiF4 OVER ALKALI FLUORIDES
- 72 -
Fig. 26. Logarithms of the results with SiF4 plotted against reciprocal
temperature, as shown in Fig. 28, were not linear. No traces of Li2SiF6
could be found by spectroscopic analysis for silicon in the LiF after
the experiment, while KF, RbF, and CsF all showed large amounts of
Exchange studies with SF6 were also carried out in the same
manner. Table 3 shows that radioactivity was detected at temperatures
above 3000 with different alkali fluorides. In this respect, SF6
resembled CF4 and CF3H. Comparison of temperatures at which counts
were detected with SF6 to the isochores in Fig. 27 shows that a pres-
sure decreasing reaction accompanied the radioactivity. Pumping
failed to remove up to as much as three-fourths of the radioactivity
present at the end of these experiments. The pressure reducing re-
action might be 2SF6 ) SF10 + F2. The compound S2F10 has a
normal boiling point of 290 and would be adsorbed by the cool counter
The surface areas of representative samples of alkali fluorides
used in the low temperature survey, as measured by the BET method using
krypton gas, were found to be:
CsF = 0.11 m. g.
RbF = 0.08 m. g.
KF = 0.08 m. g.
NaF = 0.22 m. g.
Li = 1.1 m. g.
I I I
!= RbF ,
500 I I I
200 300 400 500
Fig. 27. PRESSURE-TEMPERATURE BEHAVIOR OF SF6 ALONE AND OVER
THE ALKALI FLUORIDES
- 74 -
Figure 28 is a collection of examples of all the low temperature
Figure 29 shows examples of isochores from five gases. None of
these gases which produced curves varying significantly from that of
argon were used in further exchange studies, with the exception of SiF4.
Survey of Hydrogen Fluoride and Various Inorganic Fluorides
for Indications of Exchange with SF6 and Fluorocarbons
Nonexchange of F18 between HF and fluorinated methanes has been
reported.5 This work used HF of low specific activity; consequently
values of N below about 0.10 could not be detected. Since hydrogen may
be considered the first member of the alkali group, this work was re-
peated with HF of high specific activity.
The cyclotron was used to prepare EF containing F The
aluminum tubes containing KF were broken open and dropped into the
fluorothene tube shown in Fig. 8. This tube was attached to the vacuum
line and evacuated. One to two cc of HF was condensed from the HF
storage vessel into the fluorothene tube by cooling with liquid nitrogen.
The fluorothene tube was warmed to room temperature to allow the HF to
exchange with the KF and pick up F 8. Potassium fluoride was chosen be-
cause it goes into solution with a rapid exothermic reaction, thus
causing random distribution of the F18
A molar excess of HF was used, so that most of the F8 was re-
moved with the HF when it was distilled away. This HF usually had
- 75 -
-t ,* .
- -- (CzF5)20
4000/ T, K
4 2 3
Fig. 28. LOW TEMPERATURE SUMMARY PLOT
400 500 600
PRESSURE-TEMPERATURE BEHAVIOR OFARGON AND FIVE GASEOUS
FLUORINE COMPOUNDS OVER THE ALKALI FLUORIDES
- 77 -
sufficient F18 activity to read one roentgen per hour at contact with a
laboratory monitor on the outside of a 20 cc nickel storage vessel.
A portion of this HF was admitted to the manifold at a pressure
of 200 to 400 mm and condensed into the nickel exchange tube shown in
Fig. 9. A sample of the gas to be studied, also at a pressure of 200
to 400 mm, was admitted to the manifold and then condensed into the ex-
change tube with the HF. The exchange tube was quickly warmed to 500 +
100 with a small furnace. This temperature was maintained for one hour.
The furnace was removed and the reactor tube was cooled with
liquid nitrogen. The exchange tube was opened to the manifold and the
pressure measured. A small residual pressure was often noted, and with
CF6 considerable pressure was observed. Therefore the two U tubes on
the exit side of the manifold were cooled with liquid nitrogen and the
reaction products were distilled into these U tubes while the noncon-
densible gases were pumped off.
Any F18 removed in this manner introduced error into the de-
termination. Since the residual pressure was small above the products
cooled with liquid nitrogen, this error was small, except for the
experiment with C3F6. The reaction products were condensed into a
fluorothene tube packed with moist NaOH pellets. This fluorothene tube
was warmed to room temperature to cause reaction of the BF with the
NaOH pellets. Unreacted gas was distilled off through the filter, con-
densed into a small nickel counting vessel, and its counting rate was
determined. The NaOH pellets and the newly formed NaF were washed from
- 78 -
the fluorothene tube and diluted to 100 ml in a volumetric flask.
Aliquots were removed'and counted. Precautions very similar to those
described under the survey calculations were taken during counting to
No evidence of HF was observed in any of the gases counted when
they were bubbled slowly through an indicator solution. The gas from
Run VII had an unmistakable odor of H2S.
Table 4 demonstrates the stepwise development of the calcu-
lations of the exchange experiments with HF. From the pressure of HF
(column 2) and of the gaseous fluorine-containing compound (column 3),
the fraction of total fluorine atoms (column 7) in the fluorocarbon or
SF6 was calculated from the ideal gas law as follows:
and P = kn,
since V = volume of manifold,
and T = room temperature;
(PFC) (Fs/mole FC)
Column 7 = (PF)(Fs/mole FC) + (P H)(Fs/mole HF)
(P ) (Fs/mole FC)
(PFC)(Fs/mole FC) +(PF)
where PFC = pressure of gaseous fluorine-containing compound,
Fs/mole FC = fluorine atoms per mole of gaseous fluorine-con-
PHF = pressure of HF
- 79 -
EXCHANGE STUDIES BETWEEN HF AND
SOME FLUORINE-CONTAINING COMPOUNDS
SP Pressure Counting Rates
Run, Page Pressure of Residual
Compound of HF, mm. Fluorine- Pressure NaF Gaseous
(1) (2) (3) (4) (5) (6)
V, C5, CF3H
VII, C9, SF6
X, C14, C4F10
- 80 -
TABLE 4 Continued
in Gaseous Sum of At 100% o Percent
Compound (5) and (6) Exchange T C Exchange Remarks
(7) (8) (9) (10) (11) (12)
VLR, no H2S
a. Pumped off noncondensibles
b. Very little residual pressure
- 81 -
Fs/mole HF = 1 = fluorine atoms per mole of HF.
The fraction of total fluorine atoms in the fluorocarbon or SF6
was multiplied by the total F18 counting rate (column 8), which was the
sum of the counting rates of the HF (column 5) and of the gaseous
fluorine-containing compound (column 6), to give the counting rate
possible at random distribution, or 100% exchange. The product of
observed counting rate of the gaseous fluorine-containing compound in
column 6 multiplied by 100 and divided by the counting rate correspond-
ing to 100% exchange in column 9, yielded percent exchange (column 11).
Column 10 gives the temperature which was held for one hour in the re-
action tube. The remark PON in column 12 stands for pumped off non-
condensibles. The amount of F18 pumped off was significant only where
PON was noted. VLR means that the residual pressure over the reaction
products at liquid nitrogen temperature was so small that no significant
amount of F8 was pumped off with the noncondensible materials.
After Run X of Table 4, the nickel reactor tube was removed.
The tube was coated with NIF2 with a sufficient counting rate to
possibly account for the small percentages of exchange listed in
Table 4. Studies were then performed by heating C4F10 over NiF2 and
CF4 over CuF2, PdF2 and CsF for one hour. The results are summarized
in Table 5. Calculations were of the type discussed above with HF.
The four inorganic fluorides were prepared by exchange with HF. Sur-
face areas of these fluorides were measured by the BET method using Kr.
As has been discussed above, CsF is very difficult to prepare free of
- 82 -
HF by this method without loss by sublimation. Because of such loss,
not enough CsF remained to allow surface area measurement. The number
0.01 m2 is a reasonable maximum estimate.
EXCHANGE STUDIES BETWEEN SOME INORGANIC FLUORIDES
AND FLUORMOfARBONS AT 5000
Run, Page Materials Surface Area % Exchange
XI, C16 C4F10, NiF2 4.7 m2 19.5
XII, C19 CF4, CuF2 2.3 m2 11.8
XIII, C22 CF4, PdF2 1.2 m 29.1
XIV, C24 CF4, CsF <0.01 m2 2.8
KINETICS OF THE EXCHANGE REACTIONS
The results of the rate studies performed in combination with
the survey studies are listed in Table 6. Figure 30 shows the graphs
from which these data were obtained.
KINETICS OF EXCHANGE AT CONSTANT TEMPERATURE
Materials T, oC Rate, N, min-
NaF, C3F6 22 t 2 5.14 x 106
NaF, C3F6 260 4 6.18 x 10-5
RbF, C3F 6 46 2 6.53 x 10-6
RbF, C36 256 3 8.49 x 10-5
CsF, CFg6 203 t 3 8.25 x 10-5
RbF, (C2F5)20 300 t 5 3.49 x 10-6
CsF, CF4 368 t 3 3.16 x 10-5
In the two systems, NaF, C3F6 and RbF, C3F6, lower and higher constant-
temperature data were taken from the beginning and end, respectively,
of the same survey experiment. Considerable increase in rate with
temperature was evident in these four studies. Table 6 shows quite
similar rates with C3F6 above NaF, RbF, and CsF when the temperatures
were approximately equal. Both (C2F5)20 and CF4 appeared to produce
- 83 -
405 300 5 --
1 00 Z u--
S5,5 4. -
NoF- 3G 3
RbF- C F x 103 256 30
XX S 6 X-
0 10 20 30
A t, min
Fig. 30. KINETIC STUDIES
- 85 -
slower rates than did C3F6 on the same salts at comparable temperatures.
Figure 30 indicates that after a few minutes at constant temperature the
rates became constant.
Careful rate studies were made in the exchange apparatus shown
in Figs. 3 and 4. The procedure was modified slightly from that used in
survey studies. To prevent physical adsorption difficulties, the C3F6
or CF4 was admitted to all of the system except the exchanger. When
the gas was admitted to the exchanger, by opening the stopcocks which
isolated this part of the system, the pressure decreased in the counter
chamber and no adsorption of radioactive fluorocarbon gas occurred.
Simultaneously, a timer was turned on and times were recorded to the
After collecting data for one-half hour or longer, the gas was
condensed in the cold-finger with liquid nitrogen. A vapor bath con-
taining a higher boiling liquid was installed, and the entire procedure
was repeated using the same gas and salt. For Study AC I, the pro-
cedure was repeated for a third time at a still higher temperature.
Figure 31 shows the results for the Run AC Ia at 98.90, AC Ib at 130.50
and AC Ic at 182.50 and for Run AC IIa at 0.00 and AC IIb at 55.4.
AC Ic was continued for 90 minutes with no deviation from linearity
becoming apparent. In AC I, 5.518 x 10-3 mole of CsF was used; in
AC II, 1.007 x 102 mole. Equal amounts of C3F6 were used in the two
The rates obtained from the slopes in Fig. 31 are shown in
40 20 30 40 50 60 70
Fig.34. KINETIC STUDIES OF EXCHANGE OCCURRING WITH
C3Fg OVER CsF
- 87 -
Arrhenius Plots in Fig. 32. The inflection in AC I at a temperature of
about 4250 K corresponds to the expected Tammann temperature57 at about
one-half the melting point. Activation energies calculated from the
slopes of Fig. 32 are 4.0 kcal mole-1 for AC II and 0.8 kcal mole- for
AC I below the Tammann temperature. Above the Tammann temperature, the
activation energy in AC I was estimated as 17 kcal mole-.
In Fig. 31, extrapolations of the linear plots back to zero time
did not intercept the ordinate at zero. Logarithms of the values of N
at the ordinate intercepts are plotted against 1000/T0K in Fig. 23.
This combination plot shows that the survey results were caused almost
entirely by the fast reaction.
One study, AC III, was attempted with CF4 over CsF at 990 and
2090. At 990 an immediate exchange occurred which gave an N value of
2.1 x 10-5. No change in this value was observed over one-half hour.
At 2090 no change in N could be detected. The counting rate was then
too close to the natural background to follow the exchange further.
The measured BET surface area on this salt was 0.11 m. g.
The study AC IV shown in Fig. 33 was undertaken to determine
whether different heating periods in the CsF preparation were respon-
sible for the differences between AC I and AC II shown in Figs. 31 and
32. Two samples of CsF were compared. One was prepared at 4000 for
1/2 hr., the other at 46 for 1-1/2 hr. Each sample of CsF was used
in the same exchange procedure with C3F6 as was used for other kinetic
studies, except that at 20 min. the ice bath was quickly replaced by a
- 88 -
1 x 10"5
2.4 2.6 2.8 3.0
4000 / To
3.4 3.6 3.8
Fig. 32. RATE DATA TAKEN FROM FIG. 31
Fig. 33. EFFECT OF DURATION OF CsF PREPARATION
- 90 -
boiling-water bath, which was subsequently quickly replaced at 40 min.
by a boiling-nitrobenzene bath. The CF6 was not removed between instal-
lation of different constant temperature baths, and therefore the CsF
changed temperature gradually with C3F6 flowing constantly. The sample
of CsF heated for 1-1/2 hr. during preparation produced larger values
of N at the two higher temperatures. This CsF had a BET surface area
of 0.11 m. g. while the CsF heated for only 1/2 hr. during prepa-
ration had 0.13 m. g. Nearly equal concentrations of CsF and C3F6
were used in both parts of the study.
DISCUSSION OF RESULTS
The order of exchange ability was found to be CsF> RbF > KF,
NaF, LiF regardless of which fluorocarbon was being considered. Fair
agreement between high temperature and low temperature surveys as
shown in Figs. 17, 23 and 36 shows that the different techniques used
in preparing the alkali fluorides did not noticeably affect the results.
Reproducibility of results with CF6 and alkali fluorides was not
affected by concentrations, which varied widely. Results could not,
have been induced by F18 radiations, since the specific activity
varied by a factor of 10 with no apparent change in results.
The fluorocarbon survey results, expressed as the logarithm of
N plotted versus reciprocal temperature, were always linear. This
fact suggested that either the Arrhenius AE or Van't Hoff AE was
being measured. The latter is more probable, since reproducible
linearity is unlikely in an Arrhenius plot at a constant heating rate.
Table 7 lists slopes from the survey studies and AE values calculated
from these slopes. These slopes appear to result essentially from ex-
change by a fast reaction mechanism up to the highest temperatures
used in the low temperature survey. Figures 23 and 36 demonstrate
- 91 -