It is surprising how little work has been done with the fluorine isotopes which have been known for such a long period of time. Dodgen and L i b b y( 6 0) appear to have been the first investigators to use F18 in chemical exchange studies. They observed a heterogeneous reaction involving metal fluorides which were formed on the wall of the vessel.
Dodgen and L i b b y( 6 0) were concerned with halogen exchange in the gas phase between hydrogen halides and the corresponding halogen, i.e.,
H X * + X2 = H X + X2* .
Exchange, using F1 8, was not observed at room temperature. Fluorine and H F pressures were about 35 cm each. Heating at 200° for 1 hr or more in a brass vessel did result in exchange of atoms between the two materials;
however, a large part of the reaction was due to a heterogeneous reaction involving metal fluorides formed on the wall of the vessel. Efforts to photo-chemically produce the exchange with the light from a mercury arc failed.
This behavior was in contrast with chlorine and hydrogen chloride where rapid homogeneous exchange occurred in the gas phase at room tempera
ture in the dark.
Since previous workers ( 3 6>1 2 5) had shown that HBr and BR2 undergo rapid gas-phase exchange at room temperature and since HI and I2 do the same thing( 1 2 4 ), Dodgen and Libby( 6 0) concluded that the intermediate complex HF3, was not formed. This failure to form a complex was attri
buted to a lack of higher electron orbitals in the fluorine atom, or if the complex was formed, the hydrogen atom in the complex prevented inter
change of the fluorine atoms.
Fluorine exchange between hydrogen fluoride and the halogen fluo
rides has been studied by Rogers and K a t z( 1 7 6 ). Compounds such as chlorine trifluoride are thought to participate in ionic equilibria of the type
CIF3 + H F = C1F2+ + H F2" . fluorine atoms completely in 10 min at room temperature with liquid
B r F 3 , CIF3, B r F 5 , IF5, and SbFs. The exchange reactions between HF*
and gaseous CIF3, B r F 3 , BrFs, and IF7 were complete in 3 min at room temperature. No gas phase exchange was found between HF* and S F 6
F L U O R I N E R A D I O C H E M I S T R Y A N D R A D I A T I O N C H E M I S T R Y 269
or CCI2F2. Chlorine trifluoride failed to exchange with F2 in the gas phase.
Some vapor-solid reactions were studied. These are summarized in Table X.
T A B L E X
Vapor-Solid Exchange Reactions at 27° (1 7 6>
H F * - N a F Rapid and complete
CIF3 - N a H F2* Slow but measurable CIF3 - N a F * Slow but measurable B r F5 - N a H F2* S l o w but measurable
F2 - N a F * N o measurable reaction
The authors pointed out that all liquid-phase results were consistent with a common-ion exchange mechanism. They were not able to answer the mechanism question for the rapid gas-phase reactions; however, they felt that intermediate complexes were responsible. One intermediate complex postulated was
H
F CI: + H F > F CI F
1 /
/ I / I
F F F F
Since carbon and sulfur are completely saturated in the compounds CCI2F2 and SF6, the complexes HF-CCI2F2 and HF-SF6 would have very large activation energies and slow exchange rates.
The gas-phase isotopic exchange reactions of fluorine with labeled CIF3, BrFs, and IF7 have been studied by Bernstein and Katz( 2 6 ). The activities of both the fluorine and halogen fluoride were measured. The reaction vessels employed were copper, aluminum, and nickel. Practically no exchange was observed at room temperature; however, equilibrium could be attained at 300°. Measurable rates were observed between 120°
and 250°.
The thermodynamics of the reaction
CIF3 = C1F + F2
are known in this temperature range, and the authors studied the dis
sociation pressures of IF7 and BrFs. They concluded that reversible dis
sociation would account for the exchange observed with CIF3 and IF7.
Bromine trifluoride was, however, found to be very stable with less than 0.1% dissociation at 400°; consequently, dissociation could not account for the exchange. An associative mechanism was suggested as an alter
native. By assuming that the rate of exchange between fluorine and either
270 JOHN A. WETHINGTON, JR.
CIF3 or IF7 was controlled by the rate of dissociation of the halogen halide, the authors used the method of M c K a y( 1 3 0) to derive an expression for the rate of appearance of activity in the fluorine.
The exchange of the three halogen fluorides with AIF3, NiF2, CuF2, and CaF2 was also studied briefly. Exchange was observed, but it was not fast enough to account for the rate of exchange between the halogen fluo
rides and fluorine.
Further studies with halogen fluorides and fluorine were conducted by Adams et in 1954. It was definitely proved that all fluorine atoms in CIF3, BrFs, and IF7 are exchangeable. The work was done in a 2-liter nickel reaction vessel at three temperatures in the range of 181° to 257°.
A tremendous amount of kinetic data was obtained for the reactions
F2 + CIF3* = F2* + CIF3, F2 + B r F5* = F2* + B r F5, F2 + I F7* = F2* + I F7.
The concentration dependence of the reaction rate showed a maximum with all of the halogen fluorides; however, it was most pronounced in the BrFs—F2 exchange. The maximum was less pronounced in the CIF3—F2 and IF7—F2 exchanges. The rate was a linear function of fluorine con
centration. The authors suggested that the total rate is the sum of a homo
geneous rate and a heterogeneous rate. Simplification of the rate equations, by assuming that the fluorine was less strongly adsorbed than the halogen fluorides, gave qualitative results which agreed with the experimental facts. The authors concluded that CIF3—F2 and IF7—F2 exchange occurred by a combined heterogeneous mechanism and a homogeneous mechanism involving dissociation of the halogen fluoride. The BrFs—F2 exchange occurred by a heterogeneous mechanism. This paper is really a hallmark of kinetic studies for compounds containing fluorine, and it is worthy of careful study.
Further studies were published by Adams et alS2) in 1955. Nickel(II) fluoride was found to be an active catalyst for the gas-phase exchange reaction
H F * + F2 = H F + F2* .
Data were obtained over a wide range of concentrations at 467°, 497°, and 530°K. By assuming that the heterogeneous catalysis involved Langmuir-Hinshelwood competitive adsorption on the fluoride wall-coating, taking a general rate equation for such a process, and using some of the data to evaluate the unknowns, the authors were able to get excellent agreement between the analytical treatment and their data. This was regarded as
FLUORINE RADIOCHEMISTRY AND RADIATION CHEMISTRY 271
T A B L E X I
F18 Exchange Between Gaseous and Solid Fluorides at Room Temperature (27>
N a F L i F N a H F2 C a F2 C u F2 N i F2 AIF3 + + Slow exchange with bulk solid
+ Exchange only surface layers of solid
— Negligibly slow or n o exchange
The nonexchange of fluorine atoms between H F and the compounds CH2F2, CHF3, CF4, and CF2CI2 after heating for one hour at tempera
tures up to 500° has been reported( 3 2 ). No exchange occurred between CH3F and H F during one hour of treatment at 400°. Higher temperature measurements could not be used because of chemical decomposition.
Exchange reactions between fluorocarbons and various metallic fluorides have been studied. Exchange in the temperature range 30-250°
has been studied by Gens et a l .( 7 9 ). All of the alkali fluorides were found to exchange readily with fluorocarbons; however, the order of reactivity per square meter of surface area was found to be Cs > Rb > K, Na, Li.
proof of a heterogeneous catalysis mechanism. Additional experiments showed that the kinetics of the reactions and surface treatment of the reactor were related.
Complex fluorides have received little or no attention. It has been found( 1 4) that aqueous solutions of fluoride ions, labeled with F1 8, under
went complete exchange with inactive H2SiF6 in an aqueous solution at pH 0.5-1.0 after standing for 5 min at room temperature. Studies( 7 8) have been made which involve exchange of fluorine atoms between alkali fluorides and SiF4. The four alkali fluorides used were LiF*, KF*, RbF*, and CsF*. Manometric measurements showed that RbF—SiF4, CsF—SiF4, and KF—SiF4 underwent compound formation as well as exchange.
The system LiF—S1F4 showed no evidence for compound formation, but large amounts of exchange were observed in all cases. Temperature ranges covered were 25°-500°.
The interaction of gaseous and solid fluorine-containing compounds, based on their earlier work( 2 6>1 7 6 }, have been summarized by Bernstein and Katz( 2 7 ). Their summary table is reproduced as Table XI.
272 J O H N A . W E T H I N G T O N , J R .
Any single salt exchanged more readily with C3F6 than with CF4, C4F10, and (C2Fs)20 although appreciable exchange was observed in all cases.
The experiments were performed by circulating the gas over the salt which was heated at the rate of 5°/min; consequently, no attempt was made to establish the kinetics of the reaction. Sulfur hexafluoride failed to exchange with any of these salts below 300°, and above that temperature, com
plicated side reactions vitiated the results.
Additional study of these exchange reactions was reported in a later paper( 2 2 3 ). All of the alkali fluorides exchanged more readily with fluoro-carbons than did a number of other salts as shown in Table XII.
T A B L E X I I (2 2 3>
Exchange Reactions with C3F6
Exchange observed with N o appreciable exchange
C3F6 below 4 0 0 ° C with C3F6 b e l o w 4 0 0 ° C
Some kinetic studies( 2 2 2) have been made on the reaction between C3F6 and CsF at 0 and 55°C. A large amount of exchange occurred as soon as the gas was placed in contact with the solid. This fast reaction was then followed by a slow zero-order reaction. Extrapolation of this latter data back to zero time showed that a large fraction of exchange occurred on contact. The surface area of the salt was determined by nitrogen adsorp
tion, and the number of fluoride ions in this area on the surface of the crystals was calculated. The amount of exchange observed in this initial fast reaction was equivalent to many times the number of fluoride ions available on the surface. Figure 5 shows a correlation between the fraction exchange observed in this fast reaction and the number of surface layers equivalent to this exchange.
The oxidation of C3F6 was studied, and these same salts were used as catalysts. Since all of the oxidation products would react with aqueous alkali, the unreacted olefin could be recovered, and the fraction of olefin converted to products could be plotted as a function of temperature as shown in Fig. 6. These conversion curves could then be extrapolated to a hypothetical temperature for zero per cent conversion which agreed with
F L U O R I N E R A D I O C H E M I S T R Y A N D R A D I A T I O N C H E M I S T R Y 273
N U M B E R OF LATTICE L A Y E R S
FI G . 5. N u m b e r of lattice layers required for exchange between C s F * and C3F6 ( G e n s et
FI G . 6. Conversion as a function of temperature for alkali fluorides (Wethington
et alS™)).
the expected Tammann temperature of each salt. The temperature for zero per cent conversion was in the order CsF < RbF < LiF, K F < NaF as was predicted by the exchange experiments. As is usual in catalytic
FRACTION E X C H A N G E X I 03
274 JOHN A. WETHINGTON, JR.
experiments, selectivity in products formed was found between CsF and NaF, the extreme members of the family, as shown in Table XIII. Sodium fluoride at 215° gave no conversion to products. Complete details of the exchange experiments( 7 7) and the oxidation experiments( 5 1) are available.
T A B L E X I I I
Product Distributions Obtained(223) in the Oxidation of C3F6 Using N a F and C s F as Catalysts
CF3CF2COF 20 CF3CF2COF; CF3COCF3 10
C F3C O C F3 1 CF3COCF2CF3 20
CF3OCF2COF 10 ( C F3) 2 C F C O C F2C F3 5
Intercuts 10 Intercuts 25
Fluorobenzene, saturated with hydrogen fluoride, has been irradiated with fast neutrons to study the chemical effects of the (ny 2n) reaction'1 5\ After irradiation, the organic phase was contacted with an aqueous solu
tion of NaOH and NaF. The aqueous phase contained 0.64 of the total F18 activity. The chemical nature of the radioactive fluorine in the organic phase was, unfortunately, not specified. Retention values of 0.29 for bromobenzine have been obtained by Chien and Willard( 4 9 ). The authors stated that comparison of this value with their value of 0.36 showed that the nature of the halogen made little difference in the nature of the active chemical species resulting from neutron capture.
Sodium fluoride has been used for final purification and decontamina
tion of U F 6 made from irradiated reactor fuel( 4 6 ). It is not known whether the process depends on adsorption or on the formation of true compounds.
Martin and Albers( 1 3 7) have reported the existence of stable compounds of UF6 and NaF. Adams, Sheft, and Katz( 3) condensed U F 6 on NaF*
pellets, formed the complex, and then decomposed it by heating. There was a large amount of exchange in the solid complex but little exchange was found at temperatures where the solid complex had a high dissocia
tion pressure. The equilibrium,
U F6( g ) + 3 N a F ( s ) = U F6- 3 N a F ( s ) ,
has been postulated; however, not enough exchange to account for this
F L U O R I N E R A D I O C H E M I S T R Y A N D R A D I A T I O N C H E M I S T R Y 275
reaction was observed. Undoubtedly, these authors will report more work on this subject in the future.
Frank and Foster( 7 2) have made an extensive study of transport pro
cesses involved in the preparation of aluminum by the electrolysis of a Na3AlF6—A1203 melt. The three isotopes F1 8, Na2 4, and Al26 were em
ployed in this work. A small experimental cell with carbon electrodes was used for the electrolytic experiments, and the anode and cathode com
partments were separated by either a BN or AI2O3 diaphragm. The trans
port parameters were evaluated by radioactivity measurements. Exchange of aluminum ions between the AI2O3 and the Na3AlF6 was complete in the time required to make the measurement. Proof was presented that A1+++, A1++, A1+, 0=, and F~ were not present in the melt; consequently, AI2O3 does not dissolve in NasAlFe to yield A l3+ and 0= ions, but the rapid exchange of the aluminum between the two components showed that the AI2O3 did not form a "sugar in water" type solution in Na3AlF6. The surprising result was that the N a+ ions carried all of the current from anode to cathode. Cathode to anode migration, which was small, sug
gested the transport of AIOF2-. The transference data were not consistent with the transport of AIO2F23-.
The melting of cryolite involves the reaction
N a3A l F6 = 2 N a F + N a A l F4,
and solutions of AI2O3 in Na3AlF6 form oxyfluorides by the reactions
N a3A l F6 + AI2O3 - 3 N a A 1 0 F2
5 N a3A l F6 4- 2AI2O3 = 3 N a3A 1 02F2 + 6 N a A l F4.
By using this information plus the transport data, it was concluded that all of the data could be explained by the equilibria shown below.
2 N a3A l F6 + 2 A l2O s -> 6 N a A 1 0 F2
4 N a F + 2 N a A l F4 N a A 1 02 4- 3 N a A l F4.
They were not able to eliminate possible contributions from the reactions
5 N a3A l F6 4- 2 A 1203 -> 3 N a3A 1 02F2 4- 6 N a A l F4
l O N a F + 5 N a A l F4 3 N a A 1 02 4- 6 N a F .