There is a common belief that elemental fluorine, even when diluted, is a very dangerous material, extremely reactive, and hence completely unselective. We should be scared even when we only think of it! We can find many statements in the literature reflecting this sentiment: “Highly hazardous,” “use with great care,” “only for well-trained professionals,” “but fluorine – watch it,” “highly toxic,” and so on. (We don’t wish to mention the names of those scientists who otherwise are doing great chemistry, but such statements can be found in some of the most respected publications). Usually such statements come from people who never worked with diluted fluorine and in many cases as an “explanation” why they will work with other reagents frequently inferior to F2/N2 mixtures. It should be remembered that dilute fluorine is less dangerous and easier to work with than chlorine, for example, as if released it may spread to much shorter distances, its pressure regulators on the cylinders are more robust than those of the chlorine, and it is considerably less toxic than Cl2, to say nothing about Br2(!). Thus, for example, in order to feel “Irreversible or other serious, long-lasting adverse health effects or an impaired ability to escape” the “Acute Exposure Guideline Levels (AEGL)” required for fluorine is 20 ppm for 10 min, while that for chlorine is only 2.8 ppm and for bromine it is enough to have 0.55 ppm! For that period of time .
Obviously, as with all corrosive materials, the reactions should be performed in a well-ventilated area. Prediluted fluorine (10–20% in N2) is commercially available and when working with it the only thing the operator should take care of is attaching a simple soda-lime trap at the outlet of the reactor. If, however, the cheaper pure (>95%) F2 is preferred, one can dilute it on the spot to whatever degree desired, by using a simple vacuum line. A detailed description for such setup had appeared in the past , but we present it here again (Fig. 1). The reactions themselves are carried out in regular glassware. If elementary precautions are taken, work with F2 is simple and we have had no bad experience working with it. What is more, to the best of our knowledge no fatal accidents have been reported with it in the last 50 years.
In this review, we will mention some of the uses of F2 in organic chemistry, but will concentrate on specific mono-fluorinations on both Csp2 and Csp3 centers. Such kinds of reactions were rarely mentioned during the last quarter of the twentieth century and the beginning of the present one.
Briefly Mentioned Uses of F2 in This Review
Elemental fluorine was used to cure fluorine-containing polymers in order to remove any undesired “end groups” , or to fluorinate the surface of polymers in order to enhance certain characteristics [4, 5]. It was also used in fluorination of heteroatoms such as sulfur to prepare stable starting material for liquid crystals (1)  or nitrogen to prepare high energetic materials such as HNFX (2)  (Fig. 2).
Fluorine is also used to prepare several stable NF reagents used for certain electrophilic fluorination processes. The most popular of these reagents are selectfluor (3) and N-fluorobenzenesulfonimide (NFSI) (4)  (Fig. 3).
Another branch of fluorine chemistry which employs F2 very extensively, but again not the main subject of this review, is the perfluorination of various families of hydrogen containing compounds carried by the late Lagow being the principal investigator. The mechanism of practically all of these procedures is of radical nature. Thus, Lagow and others presented perfluorination of alkanes [9, 10], organometallics , sugars , amines , ethers [14, 15], and more (Fig. 4).
Electrophilic Substitution with Fluorine
While organic chemistry is very rich in nucleophilic substitution such as SN1, SN2, and alike, practically no electrophilic substitutions (SE) on sp3 carbon are known. Fluorine is almost the only reagent which can fulfill the requirements for such a reaction. There are two types of these substitutions, one based on leaving groups different from hydrogen and the other on hydrogen substitutions of specific C–H bonds.
Already in 1981, it has been shown that various iodo and bromo adamantanes could react with F2 to produce the corresponding fluoro derivative . As we will see below, in all electrophilic reactions by fluorine the F2 molecule has to be encouraged to become somewhat polarized and there should be a readily available acceptor for the leaving fluoride anion. Similarly, electrophilic fluorine can replace the metal in C–Hg  or C–Sn substrates to form the desired C–F bond and in the case of tin even to form the important C–18F bond for use with positron emitting tomography – PET [18, 19] (Fig. 5).
Maybe more significant and fundamental process is the hydrogen substitution at tertiary Csp3 centers. One of the first encounters with this subject was when 5α-androstane-3,17-diacetate (5) was reacted with elemental fluorine. The product proved to be the 9α-fluoro derivative (6)  (Fig. 6). The fluorine attached itself to a completely deactivated site which has no chance to be involved in any reaction by any other way! (Breslow’s “remote control oxidation” which was beautiful on paper, but hardly practical, could lead to somewhat similar results: ). If needed, subsequent dehydrofluorination could be easily performed providing a double bond 7, which in its turn could serve as an entry to numerous further reactions. This is an unprecedented procedure which has opened a whole new field in chemistry. More elaborated discussion on steroids is found further below.
The full retention of configuration, the absence of any rearrangements, and the unique attack on tertiary carbon–hydrogen bonds were powerful clues to the reaction mechanism. A radical reaction was discredited due to several reasons among them the selectivity of the reaction. Thus, for example, the androstanediol-diacetate (5) has more than 30 C–H bonds and yet the main product resulted from the attack on only one of them forming compound 6. The low reaction temperatures, the polar solvent which contained somewhat acidic hydrogen as in CHCl3, which served as a fluoride acceptor, and the fact that radical chain inhibitors such as dinitrobenzene had no effect on the reaction, were all against radical mechanism. After several experiments, it was concluded that this procedure is probably the finest example of reactions involving a nonclassical carbonium ion namely two electrons three center mechanism (Fig. 7). The electrophilic fluorine does not attack any carbon or hydrogen, but it is attracted by sites with the highest electron density such as tertiary carbon–hydrogen bonds . Later, this mechanism was strongly supported by certain theoretical studies .
The statement that the electron densities of tertiary C–H bonds are higher than those of secondary or primary ones really means that the hybridization of such bonds is higher in p orbitals than the other bonds. It turns out that electrophilic fluorine is very sensitive even to small electronic differences. Such difference is present in the case of 4-methyl-t-butylcyclohexane (8) having two tertiary C–H bonds, one near the stronger donating t-Bu group and the other near the methyl. Indeed, the reaction with F2 produced 9 in 60% yield despite the bulkiness of the t-Bu moiety while the 4-fluoro isomer 10 was obtained in only 10% yield (Fig. 8) .
Such effect was felt also when an equimolar mixture of trans cyclohexanol derivative 10a and 11 or the cis mixture 12 and 13 were treated with F2. The more electron rich C4–H bonds geminal to the t-Bu were substituted faster than the corresponding hydrogens geminal to the methyl group in 10 and 12 forming first the fluorinated derivatives 14 and 15 in 65–80% yield. Only when most 11 and 13 were consumed, the hydrogens α to the methyl groups reacted forming 16 and 17. It should be stressed that when the cis isomer (12) was reacted, only the 4-fluoro-cis-4-methylcyclohexyl p-nitrobenzoate (13) was formed in 65% yield. The lack of any isomerization in these reactions clearly indicates a nonradical mechanism. There are several easy ways for dehydrofluorination by either using acids such as BF3⋅OEt2 or bases such as MeMgBr to produce the appropriate olefins which are very difficult to obtain from the starting materials 10–13. This set of reactions is presented in Fig. 9 [24, 25].
Many cyclic compounds were examined, and as a rule, the most remote hydrogen from an electron withdrawing group was the one which was substituted. In the case of two or more tertiary hydrogens, their hybridization was determined by the PRDDO program . Thus, for example, menthyl p-nitrobenzoate (18) has three tertiary hydrogens at C-1, C-4, and C-8. It is easy to predict that the hydrogen at C-1 will be the least activated toward an electrophilic displacement. Its hybridization is sp2.7, considerably lower than the hybridization of C4–H and C8–H, both having the value of sp2.9. In agreement with the calculations, the experimental results show that no electrophilic mono-fluorination took place at C-1 but rather at C-8 producing the 8-fluoromenthyl p-nitrobenzoate (19) in 60% yield. This compound was accompanied by a small fraction of about 10% yield of 1,8-difluoromenthyl p-nitrobenzoate (20). The fact that no fluorine was found at C-4 could be understood since the fluorine has to approach the reacting center with at least one molecule of chloroform. The C-4 position is both the most hindered and the nearest to the oxygen atom whose nonbonding electrons will oppose any approach of the chlorinated solvent and possibly of the fluorine itself (Fig. 10).
That hybridization is a dominant factor for a successful electrophilic substitution on a saturated center can be also demonstrated by the pivalic ester of (2-methylcyclopropyl) methanol (21). This compound has a tertiary hydrogen separated by four bonds from the oxygen atom, usually not a prohibitive distance for electrophilic substitution. When, however, 21 was reacted with fluorine, even at concentrations as high as 20% F2 in N2, no selective electrophilic substitution took place and only slow deterioration caused by a traces of fluorine radicals, was observed. This result is in accordance with the fact that the hybridization of the tertiary hydrogens in the cyclopropyl ring is sp2.1, far too low for a successful electrophilic attack (Fig. 11) .
The above electrophilic substitution reaction is not confined only to simple cyclic compounds but has also been extended to the aliphatic field, to bicyclic compounds, and various steroids including bile acids. Thus, passing fluorine through a cold (−78 °C) 1:1 CFCl3:CHCl3 solution of 3-methylnonane (22) with its 22 hydrogens, resulted in substitution of only one of them to produce 3-methyl-3-fluorononane (23) in 65% yield. Oxygenated functions are tolerated as shown by 3,7-dimethyloctyl p-nitrobenzoate (24). Only one monofluorinated compound was obtained and identified as 3,7-dimethyl-7-fluoro-octyl p-nitrobenzoate (25). Quite impressively, despite the two tertiary hydrogens the fluorine substitute only the hydrogen more remote from the oxygenated function (Fig. 12) .
As stated above, bicyclic compounds were also subjected to this electrophilic substitution, sometimes with highly interesting results. Adamantane (26) served as a starting point in this subfield and indeed dilute fluorine reacts with it to form 1-fluoroadamantane (27) in 90% yield. As all above reactions proceed in attacking the electrons of a C–H bond, they did not involve classical carbonium ions and therefore rearrangements or strain releasing reactions of strained bicyclic derivatives were not observed. For example, the hybridization of the hydrogen at C2 in nopyl acetate (28) is favorable for electrophilic substitution compared to the other tertiary hydrogens in the strained part of the molecule (sp3.2 vs. sp2.6), so the mono-fluoro derivative, 29, was formed in a respectable 40% yield (Fig. 13).
A most interesting reaction took place when bicyclo[2.2.2]octane (30) was reacted with F2. Two major components were readily isolated. It was easy to show that the main compound was the expected 1-fluorobicyclo[2.2.2]octane (31, 50% yield), while the second derivative proved to be 2-fluorobicyclo[3.2.1]octane (32, 40% yield). This is the first and only example of a rearrangement process associated with this electrophilic substitution. This result offers an excellent addition proof for the nonclassical carbonium ion mechanism, which requires eight electrons around each carbon. The fluorine atom enhances the stability of the transition state A compared to the known parent hypercoordinated carbocation. Eventually, the charged A ejects a hydrogen bonded to either C1, C2 or C6 to form 31 and 32 respectively  (Fig. 14).
Probably there is no field more important for the implementation of this reaction than steroids. The reaction of F2 with androstanol-3,17-diacetate shown in Fig. 6 has already been mentioned. The tertiary hydrogen at 9 which is farthest away from the electron withdrawing groups (the two oxygens at 3 and 17) was the one with the highest hybridization and, as a result, the one which was substituted by fluorine. Manipulating the steroidal skeleton provides the opportunity for fluorine to substitute any tertiary hydrogen save 8β-H, which is shielded by the 18 and 19 angular methyls. 3β-Acetoxy-pregn-5-en-20-one (33) can serve as an example. Obviously, the double bond must be protected. Since the dibromo derivative proved to be unsuitable for reactions with elemental fluorine, the double bond was chlorinated to form 3β-acetoxy-5α,6β-dichloropregnan-20-one (34). The deactivating chlorine atoms left only the 14α-hydrogen to be susceptible toward electrophilic attack and 3β-acetoxy-5α,6β-dichloro-14α-fluoro-pregnan-20-one was obtained (35). Treatment of 35 with Zn/EtOH gave 3β-acetoxy-pregn-5-en-14α-fluoro-20-one (36) in 65% overall yield. Basic dehydrofluorination resulted in a double bond at 14 forming 3β-acetoxy-pregn-5,14-dien-20-one (37), providing an excellent entry into synthetic cardenolides. Having more than one double bond as is the case in stigmasterol acetate (38) does not change the picture since both olefinic functions can be protected by chlorination. The fluorination of the resulting 5α,6β,22,23-tetrachloro-24-ethyl-3β-ol stigmasterol acetate took place only at the 14 position and 39 was formed in 50% yield (Fig. 15).
With cholesterol (40), an opportunity presented itself to fluorinate the tertiary position at C-17. Trifluoroacetylation of the hydroxyl at 3 and chlorination of the double bond (41) protected these reactive centers and deactivated the 9 and 14 position toward electrophilic attack. The remaining tertiary 17α hydrogen was substituted, and after dechlorination, 17α-fluoro-3β-hydroxy-cholest-5-ene (42) was isolated as a major product in 45% yield. An additional minor derivative (20% yield) found in this reaction proved to be 17α,25-difluoro-3β-hydroxy-cholest-5-ene (43). Unfortunately, dehydrofluorination of the 17α-fluorine atom did not furnish solely the 17,20-double bond, which could serve as an entry to the expensive androstane derivatives, but rather a mixture of both 17,20 and 17,16 olefins (44). The importance of the two chlorine atoms in the examples above could be clearly appreciated, since when they were absent as in 3β-trifluoroacetoxy-5α-cholestanol (45) only 25% of the 17α-fluoro derivative (46) was formed and minor components with fluorine at 9, 14, and other sites could be detected as well (Fig. 16).
The less common 5β and/or 3α-sterols have also been fluorinated. These families of steroids are of biological importance since most bile acids belong to them.
The α-configuration of the acetoxy in 3α-acetoxy-5α-androstane (47) did not provide any surprises, and as with the 3β-ol, the expected 9α-fluoro derivative 48 was obtained in 35% yield along with 30% yield of the 14α-fluoro derivative 49. Changing the site of the electron withdrawing acetate as in 17β-acetoxy-5α-androstane (50) resulted in deactivation of the 14C–H bond, but this time along with the formation of the 9α-fluoro derivative (51) we have observed for the first time also substitution at the 5 position to form 5α-fluoro-17β-acetoxy-androstane (52). Compounds 51 and 52 were obtained in 30% yield each (Fig. 17).
Steric factors also play an important role in the fluorination of the large group of 5β steroids of which the bile acids are important members. 5β-Androstan-3α-acetoxy-17-one (53) can be considered as a typical example. After the reaction with fluorine, two main fractions were isolated. The first proved to be 5β-fluoroandrostan-3α-acetoxy-17-one (54), while the second was identified as the 14α-fluoro derivative 55, in 35 and 25% yield, respectively. Complete absence of any fluorination at C-9, despite the fact that electronically this C–H bond is the most suitable for electrophilic substitution, is remarkable. This phenomenon is observed consistently in all the A/B cis steroids. It seems that ring A, which in the 5β series is perpendicular to the steroidal plane, completely blocks the simultaneous approach of F2 and CHCl3, to the 9α-CH bond (Fig. 18).
In the case of 3α-acetoxy-5β-cholanic acid methyl ester (56), which lacks the electron withdrawing carbonyl at C-17, the 14α-CH bond becomes less deactivated toward electrophilic fluorination. The methyl ester of 3α-acetoxy-14α-fluoro-5β-cholanic acid (57) was thus formed in 35% yield. As with cholesterol and cholestanol derivatives, an additional 10% of the corresponding 17α-fluoro derivative 58, was also detected.
The presence of another electron withdrawing group as in 3α,6α-diacetoxy-5β-cholanic acid methyl ester (59) increases the deactivation of C-14 rendering the 17α-CH bond a comparatively better candidate for substitution. The two main compounds isolated were 3α,6α-diol diacetate-14α-fluoro-5β-cholanic acid methyl ester (60) in 25% yield and the 17α-fluoro derivative 61 in 15% yield. When the oxygenated functions are more evenly distributed on the steroidal skeleton, as in 3α,12α-diacetoxy-5β-cholanic acid methyl ester (62), all tertiary C–H bonds are more or less equally deactivated and the reaction become very slow, producing eventually all possible tertiary fluoro derivatives (at 5, 14, and 17), albeit in low yields. Adding a stronger deactivating carbonyl group to a diacetoxy bile acid as in 3α,7α-diacetoxy-12-oxo-5β-cholanic acid methyl ester (63), completely deactivate the molecule and even prolonged treatment with a higher than usual F2 concentration, leaves this compound untouched (Fig. 19) .
13C NMR Spectroscopy: a Tool for Configurational Studies
The above syntheses and their results provided an excellent opportunity to outline some relationships between the 13C NMR spectroscopy and stereochemistry of the fluoro-steroids especially when no functional groups are found in the vicinity of the fluorine. It was determined that all α- and β-carbons are deshielded by this halogen. The γ-carbons, however, are divided into two groups. The carbons gauche to the fluorine atom are all shielded by 1–8 ppm, whereas the γ-carbons anti to the halogen are all deshielded by 2–5 ppm. Apart from the α-carbons, which have coupling constants of 170–180 Hz, all β-carbons are also coupled to the fluorine with a 3J(CF) of approximately 20 Hz. The γ-carbons gauche to the fluorine have a very small coupling constant (usually 0–3 Hz), while the anti γ-carbons are split by the fluorine by 6–8 Hz. Both chemical shifts and coupling constants can thus be used for the stereochemical evaluation of various fluoro-steroids .
Reactions of Fluorine with Carbon–Carbon π-Centers
There are thousands of publications in the chemical literature describing the addition of chlorine and bromine across double bonds. The parallel reactions with fluorine do not exceed the number of a person’s fingers. It seems that the major obstacle for such reactions is the weak F–F bond, which can be readily homolitically cleaved to the very reactive and indiscriminating fluorine radicals. A number of indirect methods have been devised for constructing vicinal difluoro compounds in order to circumvent the direct use of the element itself (see for example: [31, 32, 33]). Some early success in adding F2 to unsaturated centers was achieved with perfluoroalkenes. In many cases, however, the dominant products were dimers which obviously originated from radical reactions . This tendency was used for preparing some of the most stable radicals known to organic chemistry .
Working with nonfluorinated alkenes, however, is another matter. After some experiments, a general way for preparing the uncommon 1,2-difluoro compounds directly from F2 and alkenes was achieved . A key factor is working with highly diluted fluorine in nitrogen (about 1–5%) in solvents containing ethanol, which slows the rate of the reaction increasing the absorption of the evolved heat and thus decreasing fluorine radical formation.
Thus, a cold solution of trans-1-acetoxy-3-hexene (64) in a mixture of trichlorofluoromethane, chloroform, and ethanol, produced only one difluoro adduct in 55% yield, identified as threo-1-acetoxy-3,4-difluorohexane (65). This compound results from syn addition of fluorine to the olefin. The same mode of addition was observed when cis-1-acetox-3-hexene (66) was reacted. In this case, only the erythro difluoro derivative 67 was obtained in similar yield. In both cases, the adducts were not contaminated by each other excluding the possibility of any anti addition process. The exclusive syn addition originates from initial nucleophilic attack of the double bond on the positive pole of the somewhat polarized fluorine molecule caused by the polar ethanol. The resulting α-fluorocarbocation is, of course, highly unstable and the tight ion pair A collapses before any rotation around the carbon–carbon bond can take place. This is also supported by the fact that no nucleophilic attack by the EtOH was detected, an attack which should have taken place if a loose ion-pair would have resulted (Fig. 20).
Similar results were obtained when trans- and cis-ethyl cinnamate (68 and 69 respectively) were reacted with fluorine. Again, the trans olefin produced only the threo-ethyl 2,3-difluoro-3-phenyl-propionate (70), while the cis isomer was converted to the erythro difluoro adduct 71, both in good yields.
Cyclic or straight-chain olefins with no functional groups can also be easily and selectively fluorinated. Cycloheptene (72) was fluorinated in 40% yield to 1,2-difluorocycloheptane (73). Larger rings react efficiently and stereoselectively as well. A mixture of trans- and cis-cyclododecene (74) in a ratio of trans:cis = 2 was fluorinated to produce two isomers which were separated and identified as dl-(75) and meso-(76) difluorocyclododecane in 60% combined yield in 2:1 ratio. Parallel results were obtained with the straight hydrocarbon 7-tetradecene (77), in which the ratio of the trans to cis isomers was 2.5. Again, two difluoro adducts were obtained, separated, and identified as dl-7,8-difluorotetradecane (78), in 50% yield, and meso-7,8-difluorotetradecane (79) in 20% yield, in agreement with a syn mode addition.
Passing F2 through the solutions of terminal olefins such as 1-octene (80) and 1-dodecene (81) also produced the expected vicinal difluoro adducts 82 and 83 in good yields (Fig. 21) .
Fluorine was added to diarylacetylenes as well. It provided the desired tetrafluoro derivatives in reasonable yields of around 70%. However, other acetylenes did not give satisfactory results (monoaryl acetylenes formed the corresponding tetrafluoro derivatives in yields of 10–20%, while aliphatic acetylenes gave a whole array of fluorinated materials, none of them the desired products) (Fig. 22) .
Fluorine addition to deactivated α,β-unsaturated carbonyls could also be achieved, although at a considerably slower rate. This was evidenced by the large excess of fluorine gas (about three times more than that for isolated double bonds) required to complete the reaction. The flexible, aliphatic α,β-unsaturated carbonyl compound trans-3-nonen-2-one (84), afforded threo-3,4-difluoro-2-nonanone (85) in 50% yield without any detectable amount of the corresponding erythro isomer. The same was true for 2-ethylhexyl acrylate (86) and dodec-3-en-2-one (88), which formed 2,3-difluoropropionic acid 2-ethylhexyl ester (87) and 3,4-difluoro-2-dodecanone (89) in 65 and 50% yield, respectively. Dehydrofluorination of 89 could be achieved by treatment with methanolic NaOH, providing 3-fluoro-3-decen-2-one (90) in 85% yield. In the case of a rigid cyclic conjugated enone, like coumarin (91), the difluoro adduct 92 was obtained in 55% yield. This compound could be quantitatively dehydrofluorinated to 93, simply by absorbing it onto a silica-gel column. The easy elimination of HF is due to the anti-configuration of the H and F atoms resulting from syn fluorine addition. Tetrasubstituted enones which would not normally react with halogens in an ionic mode do react with fluorine. Thus 2-(cyclopentylidene)cyclopentanone (94) produced the difluoro adduct 95 in moderate yield, but naturally no elimination was observed (Fig. 23) [36, 38].
Aryl enones have been a subject to this reaction as well. When the unsubstituted chalcone–benzylideneacetophenone 96a was reacted with 7–10% F2 in N2 in the presence of about 2% EtOH, 2,3-difluoro-1,3-diphenyl-1-propanone (96b) was obtained in 76% yield. The dehydrofluorination of 96b was carried with 1eq NaOH in aqueous methanol forming 2-fluorobenzylideneacetophenone (96c) in 90% yield. Similarly, when 4-nitrobenzylideneacetophenone (97a) was reacted with fluorine, 2,3-difluoro-3-(4-nitrophenyl)-1-phenyl-1-propanone (97b) was formed in 60% yield. This compound was dehydrofluorinated as described above to the corresponding 2-fluoro-3-(4-nitrophenyl)-1-phenylpropenone (97c) in 83% yield. Other aryl enones were also successfully reacted. Fluorination of ethyl 4-nitrocinnamate (98a) produced ethyl 2,3-difluoro-3-(4-nitrophenyl)propionate (98b) in 66% yield, which could be dehydrofluorinated to 2-fluoro-3-(4-nitrophenyl)-2-propenoic acid methyl ester (98c) in 90% yield by treatment with methanolic NaOH. Benzalacetone (99a) and 4-chlorobenzalacetone (100a) were also treated with fluorine, forming 3,4-difluoro-4-phenyl-2-butanone (99b) and 3,4-difluoro-4-(4-chlorophenyl)-2-butanone (100b) in 60 and 62% yield, respectively. Dehydrofluorination of 99b and 100b produced 3-fluoro-4-phenyl-3-buten-2-one (99c) and 4-(4-chlorophenyl)-3-fluoro-3-buten-2-one (100c) in 85 and 90% yield, respectively (Fig. 24) .
Fluorine was successfully added to flavones and chromones which chemically constitute interesting olefins since they are both enone and enol derivatives. These compounds are abundant micronutrients possessing various biological roles. When treated with fluorine the difluoro products were obtained in good yields and then easily dehydrofluorinated to form the corresponding 3-fluoroflavones and 3-fluorochromones (Fig. 25) .
Other additions of F2 across various double bonds are also known. Thus, for example, this halogen was added to 3-chloroenones (101) and diacetoxy cyclopentenes (102) as demonstrated by Sato et al. [40, 41], to bycyclic derivatives, such as 103  and to perfluoro olefins 104  all in reasonable yields of 20–50% and excellent stereospecificity (Fig. 26).
Aromatic fluorinations were carried with some success although not many examples are available. Thus, Sanford demonstrated that aromatics containing both electron withdrawing and donating groups could be fluorinated ortho to the electron donating group in good yield . Heteroaromatic compounds, such as pyrazole derivatives, were fluorinated in anhydrous HF in order to avoid an attack on the heteroatoms (Fig. 27) .
C-60, nanotubes, and similar “aromatic” carbon allotropes were treated with F2 in order to develop materials with special characteristics. Selig described in series of publication the results of the reaction of C-60 with F2 under various conditions. The early hopes that C60F60 would be a wonderful material did not realize since only traces of this material was obtained and it proved to be highly unstable. The main derivative (only MS, not isolated) was found to be C60F40 but the average composition was C60F35-44 [46, 47, 48]. Tohara selectively attached fluorine to the sidewall of the outer shell of double wall nanotubes (DWNTs) without disturbing the double-layered morphology. The reaction was carried out with F2 at 200 °C. He found that the stoichiometry of the fluorinated DWNT is CF0.30 .
Studies from this laboratory described in this review have been continuously supported by the Israel Science Foundation.
American Environmental Group Ltd. AEGL (Acute Exposure Guideline Level) (50), October 2, 2009.Google Scholar
Haiges, R.; Wagner, R.; Boatz, J. A.; Yousufuddin, M.; Etzkorn, M.; Prakash, G. K. S.; Christe, K. O.; Chapman, R. D.; Welker, M. F.; Kreutzberger, C. B. Angew. Chem.2006, 45, 5179.CrossRefGoogle Scholar
Chambers, R. D.; Fox, M. A.; Sanford, G.; Trmcic, J.; Goeta, A. J. Fluorine Chem.2007, 128, 29.CrossRefGoogle Scholar
Storer, R.; Ashton, C. J.; Baxter, A. D.; Han, M. M.; Marr, C. L. P.; Mason, A. M.; Mo, C. L.; Myers, P. L.; Noble, S. A.; Penn, C. R.; Weir, N. G.; Woods, J. M.; Coe, P. L. Nucleosides Nucleotides1999, 18, 203.CrossRefGoogle Scholar
Selig, H.; Lifshitz, C.; Peres, T.; Fischer, J. E.; McGhie, A. R.; Romanow, W. J.; McCauley Jr, J. P.; Smith, A. B., III J. Am. Chem. Soc. 1991, 113, 5475.CrossRefGoogle Scholar
Tuinman, A. A.; Gakh, A. A.; Adcock, J. L.; Compton, R. N. J. Am. Chem. Soc. 1993, 115, 5885.CrossRefGoogle Scholar
Kniaz, K; Fischer, J. E.; Selig, H.; Vaugham, G. B. M.; Romanow, W. J.; Cox, D. M.; Chowdhury, S. K.; McCauley, J. P.; Strongin, R. M.; Smith, A. B., III J. Am. Chem. Soc. 1993, 115, 6060.CrossRefGoogle Scholar
Muramatsu, H.; Kim, Y. A.; Hayashi, T.; Endo, M.; Yonemoto, A.; Arikai H.; Okino, F.; Touhara, H Chem. Commun. 2005, 2002.Google Scholar