The increasing demand for both the element and for substances that can only be made by its use has stimulated technical improvements in the large scale production of fluorine by electrochemical means. Simmons et a/.*682) described corrosion studies of materials for anode assemblies and concluded that copper alloys are satisfactory. A high capacity long life fluorine cell is described by Vavalides et #/.*745) and the layout and operation of a 25 lb per hr fluorine plant is described by Dykstra, Thomp
son, and Paris*1 9 2). The chemical preparation of the element, except by use of substances which themselves must be made by its use, is still unac
complished. Seel and Detmer*6 5 0) prepared it at 200 to 250° from IF7.
They complex the IF7 with either AsFs or SbFs in the compounds IF7ASF5 or IF73SbF5 and heat these with KF to form the element. IF5 is formed which combines with the K F as KIF6.
The reaction between fluorine and hydrogen is a very vigorous one, and the combination results in flames of high temperature. Wilson et a/.*7 9 2)
measured the temperature of such flames and reported for the 1 to 1 mole ratio at atmospheric pressure the flame temperature of 4300 ± 150°K.
They state that this agrees with a calculated value. Grosse and Firsh-baum*2 8 2) found, that using the pure gases (the fluorine must be freed from HF) and in the absence of transition metals particularly copper, iron, and nickel, they could premix the gases before the flame. They reported a burning velocity higher than for any other flame. A rough extrapolation would give a 1 to 1 mole ratio a burning velocity of 10,000 cm per sec at room temperature of the entering gases. Stokes and Grosse*7 1 9) obtained flame temperatures of 3950°K at 1 atm, 4400 at 10 atm, and 4890°K at
100 atm for the burning of HCN in a mixture of oxygen and fluorine.
Altman and Farber*9) questioned the calculated value of 4300°K for the F2 — H2 flame reported by Wilson et #/.*7 9 2) because of the use of the value of 63 kcal per mole dissociation energy of fluorine. As this value is now considered too high, they show that the calculated value should be 3956, if a more reasonable value of 37 kcal per mole is used. Durie*1 8 9) studied the spectra of flames of fluorine with a number of fuels such as hydrogen and hydrocarbons.
Some of the physical properties of fluorine as reported in volume one of this series are in error due to errors in early measurements or to calcu
lations based either on incomplete measurements or assumptions. For example, White, Hu, and Johnston*7 7 9) measured the density of liquid fluorine from 66 to 80°K and found it to fit the equation: d = 1.5127 + 0.00635 (85.02° — T) gm per ml. This is in agreement with the measure
ments of Kilner, Randolph, and Gillespie*3 9 6) and also those of Elverum and Doescher*2 0 2) from 65 to 85°K but in considerable disagreement with measurements reported by Kanda*3 8 5). At —196° Kanda's value is 1.13 but the later value is 1.54 ± 0.02 gm per ml. Jarry and Miller*3 6 5) also measured the density of liquid fluorine from 67 to 103°K and found, in agreement with White, Hu, and Johnson, that it fits the equation of d = 1.907 - 2.201 x 10-3T - 2.948 x 10"5T2 gm per ml with an uncertainty of ± 0 . 1 % in density and ± 0.05° in temperature. White, Hu, and Johnson*7 7 9) found the boiling point to be 85.02°K and the surface tension to fit the equation yi (d — d') = 1.276. This is in agreement with similar measurements of Elverum and Doescher*2 0 2), who also gave the viscosity of the liquid, which they found fits the equation, rj = 2.43 x 10~4 exp 196/T poise. The physical chemical properties of fluorine have been summarized by Wicke and Franck*7 8 1). Values of a first order tran
sition temperature 45.6°K, the freezing point 53.54°K, heat of transition of 173.90 ± 0.04 cal per mole, heat of fusion of 121.98 ± 0.5 cal per mole, and heat of vaporization at 84.71°K and 738 mm of 1563.98 ± 3 cal per mole were determined by Hu, White, and Johnson*3 4 6). They also gave the
80 J. H. SIMONS
normal boiling point at 85.02 ± 0.02°K and the vapor pressure by means of the equation logioPmm = 7.08718 - 357.258/r - 1.3155 x 101 3/!T8. The entropy in the ideal gas state at 85.02°K is 39.58 ± 0.16 e.u. from measurements and 39.56 ± 0.01 and 39.62 ± 0.02 e.u. as calculated from spectroscopic data; and the heat capacity and thermodynamic functions were given from 15 to 85° for F2 liquid. The free energy function of fluorine considered as an ideal monatomic gas was given from 25°C to 8000°K by Kolsky, Gilmer, and Gilles*4 1 5 ). Wicke and Franck*7 8 1) found the polarizability of fluorine to be 3.22 cm3 per mole and the dielectric con
stant of the liquid at the boiling point to be 1.43. The thermodynamic properties of the fluorine atom and molecule and of hydrogen fluoride to 5000°K have been calculated by Cole, Farber, and Elverum*1 3 4).
The atomic weight of fluorine was determined by chemical means by Scott and Ware*6 4 4) to be 18.999 on the chemical scale of atomic weights.
It is in good agreement with the atomic weight determining by mass spectro-metric methods and given on the physical scale of atomic weights (1.000275 times the chemical scale) of 19.004444 ± 0.000022 by Ogata and Matsuda*5 1 4) and 19.0044429 ± 0.0000020 by Scolman, Quisenberry, and Nier*6 4 3).
These figures show that the two scales should be made identical by the choice of the same value of the atomic weight of fluorine on both scales as has been suggested by Dole*1 7 8).
The dissociation energy of the fluorine molecule has been the subject of considerable controversy. At the time the manuscripts were prepared for volume one of the series most estimates of the value exceeded 60 kcal per mole, as is seen in "Fluorine Chemistry" Vol. 1, p. 327, Table IV, and went as high as 81. In a study of this problem Evans, Warhurst, and Whittle*2 1 2) critically examined existing data and theories, questioned the generally accepted value of 63.3 kcal per mole of that time, showed that some data in the literature lead to calculated values as low as 8.6, and selected 37 ± 8 kcal per mole as the most probable correct value. Potter*5 5 5) also discussed this problem at about the same time giving con
sideration to available spectroscopic data. He also pointed out the great range of estimates and concluded that the value should be about 1.5 e.u. or about 35 kcal per mole. Since that time there has been many types of measurements from which this value has been determined or calculated and additional theoretical studies. A chronological review of some of them will show the many ramifications of the problem and the not inconsiderable importance of arriving at as precisely correct value as possible. In 1950, Caunt and Barrow*117) determined the heat of solution in water of thallous fluoride and from it estimated Z)(F2) at 298° to be <45 kcal per mole.
By considering the vibrational frequency of the atoms in the fluorine molecule Butkov and Rozenbaum*1 0 1) theoretically calculated Z)(F2) =
72.0 kcal per mole. Schumacker*6 3 7) calculated it to be 33.3 ± 1 kcal by considering the dissociation energies of C1F and CI2, and to be 30 + 8 kcal by considering the dissociation energies of F2O and 02. From thermal conductivity Frank and Wicke*2 4 1) in 1951 estimated DF2 to be not less than 45 kcal. By directly measuring the dissociation pressures as a function of temperature Doescher*1 7 7), in the range 759 to 115°K, found AH for the reaction F2(g) -> 2F(g> to be 37.7 ± 0.4 kcal per mole. In 1952, Wicke*7 8 0) by measuring the thermal conductivity of gaseous fluorine up to 500°
in nickel tubes and using nickel wires, concluded that the value for
Z)(F2> cannot be less than 40 and probably not more than 45 kcal per mole.
Gilles and Margrave*2 6 8) determined the dissociation pressure of the gas in a copper vessel up to 860°K and concluded the Z)(F2) = 31.5 ± 0.9 kcal per mole from these experiments. By considering other experimenter's results they gave a value of 36 ± 3 as the best value. In 1953, Barrow and Caunt*43) measured the ultraviolet absorption spectra of many alkali halides and by comparing the spectra of the fluorides with those of the other halides calculated the dissociation energy of fluorine. From the average of all comparisons they estimated a value of 37.6 ± 3.5 kcal per mole. By the explosion method Wicke and Friz*7 8 2) arrived at a value of 37.0 ± 2 kcal per mole. In 1954, Wicke and Frank*7 8 1) considered all previous work on this problem and offer a value of 37.4 ± 1 kcal per mole as the best available. The heat of dissociation of fluorine by the effusion method in the temperature range of 500 to 800°K was determined by Wise*795) with a value of 37.6 ± 0.8 reported. Sanderson*6 1 3) criticized the experiments upon which previously determined values were based. Then by a long extrapolation of the properties of other diatomic molecules and of the other halogens suggested a value of 95 kcal per mole. By calculating the dissociation energy of fluorine from the appearance potentials of ions from chlorofluoromethanes, Margrave*4 5 6) in 1956 showed a range of values from —34 to +180. The method is certainly either questionable or not precise. Wray and Hornig*8 0 8) studied this dissociation by means of shock wave techniques and arrive at a value of 31.0 ± 4.3. From a study of the continuous absorption spectrum of the fluorine molecule Rees*5 7 0) in 1957 obtained a value of 37.1 ± 0.85 kcal per mole. In 1958, Stamper and Barrow*707) examined recent measurements and suggested that the best value is, Z>298(F2) = 37.72 ± 0.13 kcal per mole. In 1959, from the absorp
tion spectrum of fluorine in the vacuum ultraviolet Iczkowski and Mar
grave*356) obtained a value of 37.5 and Milne and Gilles*4 7 6) reported a value of 41.3 ± 0.5 kcal per mole from the magnetic deflection of molecular beams. It appears that at this time a value of 37.5 + 0.3 would satisfy most requirements.
The electron affinity of the fluorine atom is related to the dissociation
82 J. H. SIMONS
energy of the molecule. Since 1950, a number of measurements have been made of the electron affinity. In 1951, Bernstein and Metlay*61) evaluated it from measurements previously reported of the ionizing dissociation of fluorine on a hot fluoride coated tungsten filament. A value of 82.2 ± 3.9 kcal per gram atom was obtained which is in agreement with 82.4 calcu
lated on the basis of the dissociation energy being 37.7. Johnston*3 7 5* using elaborate theoretical considerations concluded from spectroscopic data that the electra affinity of fluorine is 73 ± 3 kcal per gram atom and the dissociation energy of the molecule 18 ± 12 kcal per mole. This is in considerable variance with other reported values. Wicke and Friz*7 8 2) gave a value of the electron affinity of 81 kcal per gram atom in 1953. Skinner and Pritchard*6 9 5) used a value of 3.63 ev for the electron affinity of the fluorine atom. Margrave*455) in 1954, calculated the electron affinity of the fluorine atom and indicated a value of 83.2 ± 0.3 kcal per gram atom. Theoretical calculations of Moiseiwitsch*489) give a value of 3.05 ev which is much lower than values experimentally determined. Wicke and Francke*7 8 1) offered the value 81.2 ± 2. The most recent and the most direct measure
ment of the electron affinities of the halogen atoms was reported in 1958 by Bailey*36). He gave a value of 82.1 ± 2 . 1 kcal per gram atom at 0°K.
The ratio of the negative to positive ions emitted by a hot tungsten filament in the vapor of an alkali halide salt is measured in the first method. In a second method the ratio of the negative halide ions emitted from the fila
ment when in the vapor of an interhalogen compound is determined. This second method gives the difference in electron affinities of the two halides.
By comparison the electron affinities of CI, Br, and I are found to be 86.6 ± 2, 80.9 ± 1.5, and 73.3 ± 1 . 7 kcal per gram atom respectively.
Stamper and Barrow*7 0 7) recalculated some previously reported experi
mental values and arrived at a value of 83.4 ± 2.5 at 298.16°K which they compared with the value of Bailey at the same temperature of 83.7 ± 2.1.
Reese, Dibeler, and Franklin*5 7 1) found a minimum value of 3.0 ev for the electron affinity of F2 from the appearance of F2~ in the mass spectra of the ions produced from S 02F2 by electron impact. This seems to be inconsistent with the above determined electron affinities of the fluorine atom as the atom would be expected to have a much larger value than the molecule particularly in view of the low polarizability of the fluorine molecule. On the other hand, if F2~ is to exist it must be more stable than its decomposition products F~ and F and this means an electron affinity of over 2 ev. Ahearn and Hannay*4) also reported F2~ in mass spectrometer studies of the negative ions formed from S F 6 . The ionization potential of molecular fluorine, F2, was found by Herron and Dibeler*3 2 5 b) to be 15.83 ev. This compares with Cl2, 11.64 and Br2, 10.58 ev.
The Raman spectrum of fluorine was reported by Andrychuk*15) and the absorption spectrum by Steunenberg and Vogel*7 1 6). The refraction of gaseous fluorine was reported by Franck*2 3 7). Fluorine apparently has the smallest polarizability of all gases except He, Ne, and H2. The compari
son follows because of the theoretical importance of its dissociation energy and other thermodynamic properties and its relationship to the previously controversial dissociation energy of the fluorine molecule. Schumacker, Schmitz, and Brodersen*6 3 8) reported its band spectrum and arrived at 60.2 cal per mole as its dissociation energy. The heat capacity, entropy, free energy, and heat content as a function of temperature from 298.16 to 2000°K have been calculated by Cole and Elverum*1 3 8) from spectro
scopic data. The standard entropy of C1F at one atmosphere for the ideal gas is 52.080 cal per mole per deg. Nielsen and Jones*5 0 5) analyzed its infrared spectrum.
The use of chlorine trifluoride to produce inorganic higher fluorides has been studied by a number of investigators. Rochow and Kukin*5 8 4) obtained C o F3 from CoCl2, NiF2 from NiCl2, and AgF2 from AgCl using C1F3 at 250°. Stein and Vogel*713) used it to make U F6 at 25-75°
and found C1F as the other product. They found nickel, monel, and inconel metals satisfactory for reaction vessels. Brass is corroded and "Teflon"
ignites above 200°. Flames in which CIF3 is the oxidizing agent and hydro
gen, hydrocarbons, and similar fuels the reducing agents were studied spectroscopically by Skirrow and Wolf hard*6 9 6). Seel and Detmer*6 4 9) found that CIF3 forms the complexes CIF3 • AsFs and CIF3 • SbFs with AsFs and SbFs.
Because of the theoretical possibility of a complex between H F and CIF3 and the reported infrared band of such a complex by Pemsler and Smith*5 4 3) having an indicated heat of reaction of 3.5 kcal per mole, this system was studied by Rogers, Speirs, and Panish*5 9 1) and by Nikoloev and Malyukov*5 0 8). A large positive deviation from Raoult's law was found, a eutectic at 44.0% C1F3, 56% HF, melting -110.7°, a slight increase in the conductivity of CIF3 by the addition of HF, but no evidence of a stable complex. The adsorption of CIF3 on porous nickel fluoride was studied by Farrar and Smith*2 1 9).
84 J. H . SIMONS
The physical properties of chlorine trifluoride have been measured by a number of investigators. Banks and Rudge*40) found the density of the liquid to follow the equation rf4* = 1.8853 - 2.942 x lO"3* - 3.79 x 10+ 6£2. Rogers, Thompson, and Speirs*5 9 4) found the dielectric constant to be et = 4.754 — 0.018^ from 0 to 42°. The electric moment has been found to be 0.557 Debye units at microwave frequencies by Magnuson*4 4 6) and 0.65 by a refractivity method by Rogers, Pruett, and Spiers*5 8 7).
Rogers, Panish, and Spiers*5 8 6) determined the magnetic susceptibility at room temperature which they found to be —0.285 x 10~6 per gm and per mole —26.5 x 10~6. The structure of CIF3 was studied by X-ray diffraction by Burbank and Bensey*94) at —120°. They concluded that it has a planar structure with one C1F distance 1.621 A and two at 1.716A.
The F—CI—F bond angle is 86°59'. Smith*6 9 8) studied the structure by microwave spectrum and gave the one CI—F distance as 1.598A, the other two as 1.698A and the angle 87°29'. The Raman and infrared spectra of CIF3 have been studied by Claassen, Weinstock, and Malm*1 2 6) and the thermodynamic properties calculated.
Chloryl fluoride, CIO2F, was prepared by Schmeisser and Eben-hoch*6 1 8) by the reaction of CIO2 and F2 in CCI3F at — 78°. It has been pre
pared by the following two reactions.
2 C 1 02 + F2 + 2 B F3 -> 2 C 1 02F • B F3
C I O 2 F • B F3 + N a F -> N a B F4 + C I O 2 F
Besides the BF3 complex, it also formed CIO2F • PF5 and C102F • AsFs.
It also added to SbFs, S i F 4 , and SO3 in a one to one ratio. Schmeisser and Fink*6 2 0) prepared the compound by the reaction of CIO2 + N2 and AgF2 at room temperature, by the reaction of CIO2 and B r F 3 at 30°, and by the reaction of NO2F and CI2O6 at 0°. They found it to react with HSO3F at - 7 8 ° to form H F and CIO2SO3F and with HC1 at - 110°
to form HF, C102, and Cl2. With anhydrous HNO3 at - 3 0 ° H F and NO2CIO4 are formed. At - 5 ° C102F • AsF5 + N 02 -> C102 + N 02 -AsF6 and at room temperature C102F • AsF5 + NO -> C102 + NOAsF6. It reacted with SbCl5 to form C102F • SbF5, C102, and Cl2. C102F • S F5 melts at 78°. It added to BrF3, and C102F • SbF5 reacted with BrF3 to form C102F • B r F 3 and B r F 3 • SbFs. Figini, Coloccia, and Schu
macher*2 2 3) reported making it simultaneously with perchloryl fluoride at 100-120° by the reaction of C1207 and F2.
Perchloryl fluoride, FCIO3, is a relatively stable compound and less chemically reactive than either chloryl fluoride, FC102, or fluorine perch-lorate, FCIO4. It can be made by the action of fluorine on KCIO3 below 20° as shown by Bode and Klesper*6 9). The same reactants were used by
Sicre and Schumacher*6 7 9) who studied methods of obtaining a pure pro
duct and further purifying it. Its preparation by this method was also discussed by Engelbrecht*2 0 4). Barth-Wehrenalp*47) made it by heating a solution of KCIO4 in HSO3F, and Engelbrecht and Atzwanger*2 0 6) found that it can be made by the electrolysis of NaC104 in HF. They found its melting point to be —146 ± 2°, its boiling point —47.5 ± 0.5°, its heat of vaporization 4960 cal per mole, its vapor pressure to follow the equation, logio Pmm = 7.683 - 1083.8/T(-112 to -44°), the density, d = 2.455
- 0.00329T gm per ml ( - 1 0 0 to -40°), critical temperature 95.9°
critical density 0.637 gm per ml, and critical molar volume 161 ml.
Giauque and Koehler*2 6 5) measured its heat capacity from 15 to 225°K.
They found its heat of fusion to be 916.3 cal per mole at the melting point of 125.41°K and the heat of vaporization 46.19 cal per mole at the boiling point 226.48°K. The vapor pressure followed the equation, logio Pmt cm =
- 1 6 5 2 • 37/T - 8.62625 logio T + 0.00460987 + 28.44780. Because the heat capacity results in an entropy at 226.48°of 2.42 e.u. less than the entropy 62.58 based upon gas molecular data at the same temperature, the authors concluded that the crystal fails to distinguish completely O and F atoms.
A complete failure of this kind would result in an entropy difference of 2.75 e.u. Its heat of formation at 25° was found by Neugebauer and Margrave*5 0 2) to be —5.12 ± 0.68 kcal per mole by measuring the heat of hydrogenation.
By a study of the microwave absorption in perchlorylfluoride Lide and Mann*4 3 2) concluded that the dipole moment must be less than 0.90 Debye units. Madden and Benedict*443) and also Lide and Mann*4 3 3) studied the infrared spectrum. The latter authors concluded that the structure has a central chlorine atom surrounded by, and attached to, the other four atoms. This is different from an earlier assumption from chemical reaction information that the structure was CIO2 • OF or an oxyfluoride.
By the direct fluorination of K, Rb, and Cs chlorides, Asprey et <z/.*21a) obtained the tetrafluorochlorates of the metals KCIF4, R b C l F 4 , and CSCIF4.
These thermally decomposed to give C1F and CIF3. The bromides of the same alkali metals gave tetrafluorobromates, and the alkali iodides gave tetrafluoroiodates. The former gave B r F 3 and the latter IF5 upon thermal decomposition. The thermal stability decreased from cesium to potassium.
The reactivity with water decreased from the chlorate to the iodate.
Ray and Mitra*5 6 7 a) studied the monofluorochlorates and report the following: CuC103F • 5 H20 , ZnC103F • 7 H20 , 3CdC103F • 8 H20 , NiC103F • 7 H20 , C0CIO3F • 7 H20 , C0CIO3F • 6 H20 , CaC103F, SrC103F,
Ray and Mitra*5 6 7 a) studied the monofluorochlorates and report the following: CuC103F • 5 H20 , ZnC103F • 7 H20 , 3CdC103F • 8 H20 , NiC103F • 7 H20 , C0CIO3F • 7 H20 , C0CIO3F • 6 H20 , CaC103F, SrC103F,