1. Introduction
The first organomercury(II) compound (organomercurial) was reported by E. Frankland (1850), who synthesized dimethylmercury, Me2Hg, by the action of methyl iodide on mercury metal under sunlight irradiation. The number of various organomercurials that were next synthesized for pharmacological purposes is very large, but their role in chemotherapy has now been completely superceded; also their use as fungicides is on the decline, owing to the toxicity to human beings and animals. But numerous organomercurials found increasing applications as useful reagents for the synthesis of many other organometallics via methal exchange reactions, and they are still attractive as synthetic intermediates, which are usually readily available, accommodate essentially all functional groups, and possess remarkable thermal and chemical stability to air, water, dilute acids and bases. These valuable features allow synthetic reactions employing organomercurials to be run under a wide variety of reaction conditions. The major disadvantage of these compounds is the high toxicity of more volatile mercurials. In fact, all heterocyclic(lactamic) organomercurials prepared by us during our research works (1980-2000) were nonvolatile and either slightly soluble or practically insoluble in common solvents, hence they were notably less hazardous than e.g. many aliphatic organomercurials. But due to their very limited solubility in boiling common solvents, they often could not be purified satisfactorily or not at all. This made difficult (or even impossible) to analyze them reliably as to make proper structural assignments. Therefore, their possible structures were drawn from subsequent, well known and effective, chemical reactions (usually iodo-demercuration and/or bromo-demercuration reactions) followed by chemical and spectral analysis of the resulting products (i.e. monoiodo and/or monobromo derivatives of the parent aromatics), often also produced in the other known routes, for the sake of a better comparison.
There is a large number of various methods to prepare unsymmetric aromatic organomercurials, ArHgX, but the direct mercuration of aromatic systems is evidently the most simple of them and very often used. Its proper and effective application strongly depends on: (i) the relative reactivity of the reacted aromatic system to the electrophilic attack by a mercuric salt; (ii) the relative electrophilicity of a mercuric salt applied, e.g. HgCl2 <Hg(OCOCH3)2<Hg(OCOCF3)2 <Hg(ClO4)2; (iii) the applied reaction conditions: the use of a proper solvent, temperature and reaction time. This is ordinary electrophilic aromatic substitution and takes place by the arenium ion mechanism to form the unsymmetric aromatic mercurials, e.g. ArHgCl < ArHgOCOCH3 <ArHgOCOCF3 <ArHgClO4 (just in this order are growing their relative reactivities to the subsequent actions of various electrophilic reagents).
There are known several, more or less effective in each particular case, symmetrization methods to convert unsymmetric compounds ArHgX to the corresponding symmetric ones, Ar2Hg, which are usually more reactive than ArHgX. It must be borne in mind that the following equilibrium should be displaced far to the right to cause effective symmetrization:
Practically, this may be attained either (i) by removal of the mercuric salt HgX2 by strong complexation with e.g. sodium or potassium iodide, sodium thiosulfate, ammonia or EDTA, potassium cyanide, potassium thiocyanate, etc. or (ii) by reduction of HgX2 with hydrazine, sodium stannite, by electrolysis, or otherwise. Hence, there is no single general procedure applicable for various types of ArHgX to be effectively symmetrized; certain types of ArHgX compounds are symmetrized with great ease and high yields of Ar2Hg, while others are resistant to particular symmetrization methods, and the desirable Ar2Hg can only be obtained by change of the applied procedure(s).
In 1979 when we started our preliminary experiments, the direct mercuration with mercuric salts of some important "model" lactamic heterocycles, e.g. uracil, theobromine and theophylline, was not reported in the former literature. The attempted mercuration reactions of these compounds at their C5 or C8 positions, respectively, resulted only in the formation of their insoluble N-Hg salts 1 - 3 (named also 1:1 or 1:2 mercuric "complexes"), which were precipitated out in full from the mercurating reaction mixtures – this impeded to obtain their true organomercurials, i.e. their mercurated derivatives with the mercury atom joined to the organic residues via C5 or C8 carbon atoms.
.
In contrast, heterocycles devoid of any acidic N-H group(s), e.g. caffeine (see Section 2), 1,3-dimethyluracil (see Section 6), 2,4-dimethoxypyrimidine (see Section 7), or 2,3-diphenyl-5-methoxybenzo[b]furan (see Section 5) were readily directly mercurated to form the respective unsymmetric organomercurials, ArHgX, which were often symmetrized to form Ar2Hg. Next, the mercurials of the two types were always iodo-demercurated and/or bromo-demercurated to form the corresponding iodides or bromides, ArI or ArBr, which after their purification were analyzed and studied with 1H NMR spectra, and otherwise, to confirm chemical structures of the starting mercurials.
Generally, it is known that mercury(II) ion, consistently with its pronounced electrophilic character, can effectively mercurate a large number of aromatics at their carbon atom(s) to form their true organomercurials, if they are inherently devoid of any acidic S-H, O-H, N-H, Se-H, sometimes P-H groupings. Otherwise, Hg(II) reveals a higher tendency to combine with those electronegative heteroatoms bearing lone pair(s) of electrons, forming thus some seemingly "mercurated", sparingly soluble final products or intermediates – which are not true organomercurials, since the mercury atom therein is joined to the organic residue not via carbon atom but via the respective heteroatom. Those containing the O-Hg bond are the least stable. Compounds with the S-Hg linkage are formed very readily and exhibit particular stability. The strength of the N-Hg bond varies between wide limits, but mostly it does not exceed that of the S-Hg bond. Organic compounds containing O-Hg, N-Hg, and sometimeseven S-Hg bonds are often the first intermediates formed during the direct mercurations of oxygen-, nitrogen-, and some sulfur-containing compounds, and transform more or less readily into compounds mercurated at the carbon atom (true organomercurials), when process is carried out under more vigorous conditions, e.g. at a higher temperature or at a lower pH. For example, thioanisole can be C-mercurated in 36.6% yield by means of Hg(OCOCH3)2 on a steam bath, giving thus 4-(acetoxymercurio)thioanisole; for more such examples see Ref. 23, p. 93.
In the course of our systematic, many-years¢ studies on various aromatic (carbo- and predominantly heterocyclic) true organomercurials and their reactions, we successfully synthesized novel organomercurials derived from theobromine and theophylline (Sections 9 and 10), uracil (Section 8), and also some other mercurials, though mostly in indirect ways. We also isolated (in ca 28% yield) the supposed-to-be 1,8-bis(acetoxydimercurio)theobromine, which seemingly represents the first stable organic derivative of mercury(I), and next we studied its chemical properties (Section 11). We also discovered several novel halo- and cyano-demercuration reactions; in our opinion, our novel fluoro-, chloro-, and cyano-demercuration reactions are particularly interesting and useful. We also extended, improved and better substantiated the old Willgerodt`s method (1897), which enables to synthesize diaryliodonium chlorides by reacting (dichloroiodo)arenes with symmetric aromatic organomercurials, in stirred hot aqueous suspensions (Section 14). For more informations on our own published achievements [1-16] see the following text.
2. Early Results: 8-Substituted Caffeine Mercurials and Their Demercurating Reactions
(see Refs. 1, 2, 11 and 12)
Previously, Rosenthaler [43] had mercurated caffeine (4a) to afford 8-(acetoxymercurio)caffeine (4b) in 90% crude yield, on boiling 4a for 48 hours with an aq. Hg(OCOCH3)2 solution acidified with CH3CO2H. Next, he symmetrized 4b with a boiling aq. KI solution for 30 minutes, to give 8,8¢ -mercuriobis(caffeine) (5) in 50% yield; he did not carry out any further reactions with 4b and 5. In our laboratory, we improved the above synthesis of 4b, by adding a few drops of conc. aq. HClO4, a recommended [23, 24, 26-28] catalyst, to the mercurating reaction mixture, which shortened considerably the mercuration time to only four hours, with preserving the same 90% crude yield; we also established that the mercurating solutions may be prepared in situ by dissolving the freshly prepared yellow HgO in hot aq. CH3CO2H. We improved as well the crude yield of symmetrization of 4b from 50% to 83%, by applying a hot (90° C) aq. Na2S2O3 solution, instead of aq. KI. We also metathesized 4b into sparingly soluble 8-(chloromercurio)caffeine (83% crude yield) by adding excess aq. NaCl solution to a boiling solution of 4b in water [1, 11]. On heating with an aq. Na2S solution the black precipitate of HgS proved the presence of mercury in the three afore-said caffeine mercurials; their chemical structures were confirmed by their 1H NMR solution spectra as compared with that of caffeine [1, 11], as well as by subsequent halo- and cyano-demercurationreactions discussed below; see also Ref. 11.
Finally, it is impossible not to recall that our many attempts to synthesize either 8-(trichloroacetoxymercurio)caffeine or the more thermostable 8-(chloromercurio)caffeine by a one-pot method suitable for preparing the thermostable phenylmercury(II) chloride were without success: the latter benzene mercurial was obtained in 70% yield when yellow HgO and CCl3CO2H were stirred in thiophene-free benzene for ca one hour at 65-70° C; for the explanation see Ref. 11 as well as Section 3.
2.1. Halo-demercuration Reactions of Caffeine Mercurials [1, 2, 11 and 12]
Gomberg [44] had reported the failure to iodinate caffeine (4a)withdiiodine in chloroform; later on, 8-iodocaffeine (4c)was obtained in ca 40% yield by heating caffeine (4a)with diiodine in a sealed tube at 150° C [45]. In our laboratory, at first we iodo-demercurated caffeine mercurials 4b and 5 by applying the widely used [17, 23-28] procedure; on heating these mercurials with hot (80° C) aq. KI3 solutions for 30 minutes, we obtained 4c in the same 95% yield (after purification) from the both substrates 4b and 5 [1, 2]. Similarly, the sparingly soluble 8-(chloromercurio)caffeine (4, X = HgCl) was iodo-demercurated to give pure 4c in 65% yield [11]. The same pure 4c was also obtained by us [1, 2] in 90% yield from the both substrates 4b and 5, by heating them with a hot (80° C) aq. solution of the freshly sublimated ICN [Caution. ICN is strongly toxic]; in fact, it is an effective, though less convenient and unsafe, iodo-demercurating method.
Previously, caffeine (4a) was directly brominated to give 8-bromocaffeine (4d) in high yields [45, 46]. In our laboratory, the both mercurials 4b and 5 were effectively bromo-demercurated, with applying the widely-used [17, 23-28] procedure; the reactions underwent in hot (80° C) aq. KBr3 solutions prior adjusted to pH 7, for 30 minutes, to give 4d (purified) in the same 83% yields from the both substrates 4b and 5. We also bromo-demercurated 4b and 5 (at 60° C, for 2 hours) with an aq. slurry of a labile complex (or adduct) of unknown composition, (KBr×BrCN)x [Caution. It is strongly toxic], obtained by adding Br2 to a cooled saturated aq. KCN solution [47]; this gave compound 4d (purified) in 85% yields from the both substrates 4b and 5. In fact, it is an effective, though less convenient and unsafe,novelbromo-demercurating method.
The direct chlorination of caffeine (4a) to 8-chlorocaffeine (4e) had been reported as early as 1850; fairly high yields of 4e (ca 80%) were later reported in the literature [45, 48]. In our laboratory, we obtained the purified compound 4e in 27-90% yields, but only from the more reactive symmetric mercurial 5 [mercurial 4b did not react under the same reaction conditions]. At first, we reacted 5 with neat liquid S2Cl2 or SCl2 at room temperature for 4 hours to afford 4e in 80% or 70% yields, respectively; though the yield of the latter reaction with SCl2 was lower, nevertheless the crude 4e thus obtained was easier in its purification. When mercurial 5 was similarly suspended in neat boiling SO2Cl2 for 2 hours, then compound 4e (after its isolation and purification) was obtained in only 27% yield. These novel chloro-demercuration methods were later on applied by us also with other mercurials (Section 13.3)). Finally, it is necessary to admit that we failed to obtain compound 4e on passing gaseous Cl2 through solutions or suspensions of mercurials 4b or 5 in water, aq. KCl solutions, formamide, dimethyl sulfoxide, benzene, etc. The same was true for the action of chlorine in statu nascendi (generated in reaction: 6HCl + 2KClO3 ® 6Cl + 2KCl + 3H2O) upon 4b or 5 in aqueous media. For some more reactive organomercurials these chloro-demercurating reactions were successful [17, 23, 24, 26, 28]; see also Section 3.
8-fluorocaffeine (4f) was notreported in the literature before 1981/1982 [1, 2]; until then very few organomercurials have been fluoro-demercurated, but only with difluorine diluted with nitrogen gas [26]. In our laboratory, we suspended the more reactive mercurial 5 [mercurial 4b did not react under the same reaction conditions] in liquefied, neat SF4 (b.p. -40.4° C. Caution. It is very strongly toxic) at -70° C (solid CO2 + acetone) for 16 hours. After evaporating SF4,crude 4f was purified to give 30% yield (after 8 hours only 15%); an increase in the reaction temperature (in an autoclave) considerably diminished the given yield. This novel and interesting fluoro-demercuration reaction was later on applied in our laboratory with other symmetric mercurials; see Ref. 12, where we summarized and commented all our fluoro-demercuration experiments; see also Section 13.4.
2.2. Cyano-demercuration Reaction of Caffeine Mercurials [1, 2]
Previously, 8-bromocaffeine (4d) in the presence of KCN in 80% ethanol yielded caffeine-8-carboxamide, which by treatment with POCl3 yielded 8-cyanocaffeine (4g) [45, 49]; it is worth mentioning that ICN, when reacted at 110° C with dimethylmercury, yielded mercuric iodide and methyl isocyanide [50]. In our laboratory, we reacted upon the more reactive, symmetric mercurial 5 with the freshly purified BrCN [Caution. It is very strongly toxic] in water, at 60° C for 2 hours (at higher temperatures a decomposition of 5 was observed, whereas at 40° C the reaction did not proceed); after purification of crude 4g, we obtained pure 4g in 50% yield; the same reaction with unsymmetric mercurial 4b gave pure compound 4g in only 25% yield. It was surprising that by using aq. KBr solutions instead of pure water for dissolving BrCN, the same compound 4g was obtained, instead of the expected 4d (vide supra). By using aq. KCN solutions instead of water for dissolving BrCN, the yield of this cyano-demercuration reaction was evidently lowered. We applied this interesting but unsafe reaction also for more reactive "model" benzene mercurials (see Section 3). It is also necessary to admit that by acting upon 4b or 5, under widely varied conditions, with (CN)2 or ClCN solutions we failed to obtain compound 4g; the latter reagent decomposed vigorously both 4b and 5 in their solutions or suspensions.
2.3. Unsuccessful Reactions with Caffeine Mercurials, and Conclusions [1, 2]
The preparative demercurating methods discussed above show the usefulness of organomercurials in organic synthesis. There are, however, noticeable differences in the reactivity of various organomercurials. It has been mentioned several times in the literature [23-28, 34] that Ar2Hg compounds are usually more reactive as compared with ArHgX. It is seen that unsymmetric mercurial 4b does not react with S2Cl2, SCl2, SO2Cl2, and SF4 under the same experimental conditions as does symmetric 5. The cyano-demercuration reaction of symmetric mercurial 5 furnished twice as much of 4g as compared with unsymmetric mercurial 4b. Only the iodo- and bromo-demercuration reactions furnished the same yields of 4c and 4d from both 4b and 5. We have also established experimentally that both 4b and 5 did not undergo several well-known [23-28, 34] demercuration reactions with HNO3, (CN)2, (SCN)2, Cl2 or chlorine in statu nascendi, SOCl2, and aryldiazonium cations, which were successful with some other, more reactive, organomercurials. Thus, it seems to us that the both caffeine mercurials 4b and 5 are noticeably less reactive in a number of demercurating reactions as compared with some corresponding organomercurials, ArHgX or Ar2Hg, in which Ar represent e.g. o-nitrophenyl, thienyl or furyl rests [23-28, 34]; cf. particularly our results presented in Sections 3 and 4.
3. "Model" Monosubstituted Benzene Mercurials and Their Demercurating Reactions
(See Refs. 7, 11 and 12)
The survey of methods applied so far for the halo-demercuration of various organomercurials [17, 24-28, 34] reveals that dihalogens do decompose them in two stages: (1) R2Hg + X2 ® RHgX + RX, followed (with an excess of halogens) by (2) RHgX + X2 ®R-X + HgX2; it has been pointed out [23] that symmetric R2Hg reacts more readily than RHgX. In 1870 Dreher and Otto [51] reacted cold solutions of diphenylmercury, Ph2Hg (in ethanol or better in CS2), with equimolar amounts of I2 or Br2, and they obtained the respective halogenomercurio- or halogeno-benzenes, whereas with an excess of the halogens the formers were changed into the respective halogenobenzenes and mercuric salts (with necessary heating). Dry dichlorine gives a vigorous reaction with Ph2Hg yielding PhHgCl, PhCl and HgCl2; when Cl2 is bubbled through a hot aq. suspension of Ph2Hg or (better) through its hot solution in CS2, at first PhHgCl, and then more and more PhCl and HgCl2 are obtained. The German authors [51] have also remarked that PhHgCl seems to be less reactive towards the action of Cl2 than the corresponding PhHgBr and PhHgI towards the action of Br2 and I2. A concentrated solution of hypochlorous acid acts like free Cl2, forming PhHgCl and PhCl from Ph2Hg. An aqueous solution of PhHgOCOCH3 heated with an excess of I2 gives PhI, HgI2 and CH3CO2H [51]. Later on, the iodo- and the bromo-demercuration reactions have vastly been improved [24-28, 34, 52], e.g. by using KI3 or KBr3 solutions, mostly aqueous (cf. Ref. 2), but also with applying a large variety of solvents, e.g. alcohols, acetonitrile, DMF, DMSO, pyridine, dioxane or their mixtures with water. Even permercurated arenes are readily cleaved in this way [52] by NaI3 in DMF at room temperature for 3-14 days, by KBr3 (where 3 is to be subscripted) aq. methanolic solutions at room temperature for 1-24 hours, or by Cl2 in DMF at room temperature for 3-8 hours, giving thus the respective periodo-, perbromo-, and perchloro-arenes in good yields. It is also necessary to recall that on a prolonged heating Ph2Hg with TeCl2, mercuric telluride and PhCl are formed unexpectedly [23, 53]. Iodine monochloride reacts according to the reactions [23, 28]: R2Hg + 2ICl ® 2RI + HgCl2 and RHgX + ICl ® RI + HgXCl; the same diiodocamphor is obtained when mercurated camphor reacts with I2 or with ICl or with IBr in benzene [17]. ICN acts preferably as a iodo-demercuratingagent, whereas BrCN may act both as a bromo-demercurating agent as well as a cyano-demercurating agent depending on the reaction conditions (vide infra, as well as Section 2.2). ClCN has been reported to give no action at all with some organomercurials [17]; cf. however Section 2.2. It is of interest to mention that ICN and Me2Hg in ethereal solution give MeCN at 50° C, and HgI2 and methyl isonitrile at 110° C [17, 23, 50]. It has also been reported [23] that BrCN does not cleave the C-Hg bond, but only replaces e.g. the acetoxy group by bromine in a -acetoxymercurio-b -methoxy-b -phenylethane. Pseudohalogens, X2 = (CNS)2 and (CN)2, do react with some symmetric organomercurials, R2Hg, giving the respective RX and RHgX compounds; we failed, however, to replace mercury atoms in the less reactive caffeine mercurials on acting upon them with (CNS)2 and (CN)2 (see Ref. 2 as well as Section 2.3). The reaction with ClN3 undergoes similarly, viz. R2Hg + ClN3 ® RHgN3 + RCl, and it was applied as well with Ph2Hg (R = C6H5) [17, 24-28, 34].
In our subsequent paper [7] of the series, we have reported several halo- and cyano-demercurating procedures, applied with the more reactive "model" benzene mercurials 6a, 6b and 7; cf. our results discussed in Section 2.
Diphenylmercury, Ph2Hg (7), dissolved in ethanol gave smoothly only PhI (6e) in 72% yield (purified product) on adding pure ICN and then refluxing the mixture for 3 hours. On refluxing for 3 hours a mixture of 7 with pure BrCNin the benzene solvent, only PhCN (6i) was furnished in 86% yield – whereas when 7 was refluxed for 3 hours in an ethanolic solution containing (BrCN× KBr)x, a complex of unknown structure [47], only PhBr (6f) was attained in 79% yield. The reaction of 7 with ClCN, carried out under widely-varied experimental conditions, gave some composite mixtures, which were not studied closer. By acting upon 7 with large excesses of neat liquid either S2Cl2 or SO2Cl2 [SCl2 was not studied], left overnight at room temperature, we obtained the purified PhCl (6g) in 78% or 47% yields, respectively. The former reaction was also carried out at first at -10° C, and next it was left overnight at room temperature: we obtained PhCl (6g) in 35% yield together with 30% of isolated PhHgCl (which throws some light on its mechanism). The second reaction was considerably accelerated by a Friedel-Craft catalyst, viz. AlCl3. Both the symmetric mercurial 7 as well as the two unsymmetric ones, 6a and 6b, did react with an excess of neat S2Cl2 (under the same experimental conditions) to afford PhCl (6g) in 78%, 50% and 69% yields, respectively; previously (Section 2), it was supposed that this new chloro-demercurating procedure was characteristic only of the symmetric organomercurials.
The novel fluoro-demercurating procedure, applicable only for the symmetric organomercurial 7, was reported in our another paper [12]; cf. also Section 2.1. On acting upon 7 with a large excess of neat liquid SF4, for 8 hours at –(60-70° C), it was possible to obtain PhF (6h; purified product) in 58% yield; its structure was confirmed by microanalyses and 1H NMR spectrum. We submitted there the following remark (footnote in p. 26): "The calculated yields [for ArF] correspond to "ideal" reactions: Ar2Hg + SF4® HgF2 + 2ArF (soluble in organic solvents; often volatile), whereas the "real" reactions probably are terminated either in full or in part at the intermediate stage yielding ArF + ArHgF (sparingly soluble in organic solvents; nonvolatile). Since the aim of our investigation [12] was to obtain the possible highest yields of ArF, therefore we did not scrutinize the nonextractable residues composed of mercuric salts heavily contaminated by some sulfur-containing side-products and, probably, by the less-reactive (under the given low-temperature conditions) fluoromercurials, ArHgF." In fact, all our fluoro-demercurating procedures, applicable only for Ar2Hg mercurials, resulted in moderate yields for the purified ArF, viz. 28-58% [12].
In our next work [11] we attempted to mercurate benzene and caffeine by means of mercury(II) trichloroacetate; we submitted therein a detailed review on the former attempts for preparing this thermally unstable mercuric salt. We have come to the conclusion that it is reasonable to prepare this mercuric salt in situ to obtain appropriate mercurating reaction mixtures. When we suspended yellow HgO in thiophen-free benzene containing Cl3CCO2H (a 1:1 molar proportion of Cl3CCO2H to HgO seems to be optimal) and the reaction mixture thus obtained was heated for ca one hour at 65-70° C, then we unexpectadly isolated PhHgCl (6d) in 70% yield (purified product), instead of the thermolabile mercurial 6c. This method was quite unsuitable for the mercuration of less reactive caffeine. Next, mercurial 6d was readily iodo-demercurated to give 6e (purified product) in 59% yield, with a hot (80° C) aq. KI3 solution, for 30 minutes.
4. "Model" 2-Substituted Furan and Thiophene Mercurials and Their Demercurating Reactions
(See Ref. 5)
In order to extend the scope of the afore-said
(Sections 2 and 3) novel chloro-demercuration procedure
[the action of neat, liquid S2Cl2 on
mostly symmetric organomercurials, yielding effectively the
corresponding chloro derivatives, ArCl; see however Section 3], we synthesized
at first (by the known methods) 2,2¢
-difurylmercury (8) and 2,2¢
-dithienylmercury (9). We have expected in advance that they
would be more reactive than those symmetric organomercurials discussed
in Sections 2 and 3. However, in our attempted chloro-demercuration reactions
we had to use only the freshly-redistilled
S2Cl2 to remove the accompanying
SCl2; the latter is known [54] to equilibrate as follows:
2SCl2 
S2Cl2 + Cl2, which is very
inconvenient due to the known high sensitivity of furan and thiophene towards
the action of free dichlorine; see also our paper [5] where various
chlorinating procedures applicable for furan and thiophene were briefly
reviewed, with the relevant references.
At first, we reacted upon solid finely-powdered mercurials 8 or 9 as well as upon 2-(chloromercurio)furan or 2-(chloromercurio)thiophene with an excess, as previously (see Sections 2 and 3), of neat liquid S2Cl2 with no solvent, over a wide temperature range from -70° C up to room temperature. Even at -70° C the reactions were extremely vigorous and could hardly be controlled. Colorless oils, in nearly quantitative yields, were isolated from the reaction mixtures, which were composite mixtures of several highly chlorinated compounds, in part also of open-ring structures, with only small admixtures of 2-monochloro and 2,5-dichloro derivatives of furan or thiophene, respectively. After several attempts we achieved a proper and effective method of chloro-demercurating mercurials 8 and 9, but only using carbon disulfideas aninert diluent – the reactions were completed after two days either at room temperature with more reactive 8 or, with less reactive 9, on boiling under a reflux condenser; with still less reactive 2-(chloromercurio)furan or 2-(chloromercurio)thiophene the reactions did not proceed under the same experimental conditions. From the reaction mixtures we isolated either pure 2-chlorofuran (10a) in 60% yield or pure 2-chlorothiophene (11a) in 70% yield, with no trace of higher-boiling 2,5-dichlorinated admixtures. Later on, we considerably simplified the syntheses of compounds 10a and 11a just by mixing together pure furan or thiophene with HgCl2 prior dissolved in an excess of the freshly-redistilled S2Cl2, followed by keeping the mixtures overnight at room temperature. Most likely, the reactions proceeded via some 2-mercurio intermediates, since no 2-chloro derivatives were formed in the absence of HgCl2. From the reaction mixtures we isolated pure compounds 10a or 11a in 50% or 60% yields, respectively; see Ref. 5 for experimental details.
The various methods for the direct iodination of furan and thiophene (and also other heterocycles) were very extensively reviewed in Ref. 55; they were also briefly reviewed in our paper [5]. We reacted symmetric mercurials 8 and 9 with the well-known [23, 26, 28, 34] iodo-demercurating agent, viz. an aq. KI3 solution, for 30 minutes at 80° C, which gave pure 10b or 11b in 60% or 65% yields, respectively. These yields are considerably higher than those previously reported, when the same iodo-demercuration method was applied either with 2-(chloromercurio)furan (32% yield was reported [56]) or 2-(chloromercurio)thiophene (a low yield was reported [57]). Our iodo-demercuration reactions confirmed the structures of the starting mercurials 8 and 9 as well as their higher reactivities as compared with the respective 2-chloromercurio derivatives.
In our paper [5] we submitted the following general remark (footnote in p. 445): "It is rather a common procedure that after completing several direct mercuration reactions with more electrophilic (than HgCl2) mercuric acetate or trifluoroacetate, the less soluble, but the least reactive, chloromercurio derivatives are precipitated out with aq. NaCl or CaCl2 solutions, in order to increase the isolated yields of desirable mercurials. In our opinion, it is often more advantageous to collect the respective symmetric mercurials (usually less soluble as well), which may be obtained by subsequent addition – if possible – of symmetrizing agents (e.g. KI, Na2S2O3, etc.) directly to the said reaction mixtures, and then to complete the symmetrization reaction in order to obtain the resulting, more reactive Ar-Hg-Ar compounds". In fact, we applied this approach inpractice in our works [6] and [9]; see Sections 7 and 8 for details.
5. 6-Substituted 2,3-Diphenyl-5-methoxybenzo[b]furan Mercurials and Their Demercurating Reactions
(see Ref. 10)
Egyptian chemists [58] had synthesized a number of derivatives of heterocyclic compound 12a, and later they studied their biological activities. They established that in various reactions obeying the SE-type mechanism, the respective derivatives of 12a were predominantly substituted at C6 carbon atom. However, bromination of 12a with Br2 in CCl4 gave 4,6-dibromo derivative of 12a. In co-operation with the Egyptian chemists we decided to synthesize in our laboratory two mercurials 12b and 13, and next to study their bromo- and iodo-demercuration reactions.
The direct mercuration of 12a with Hg(OCOCH3)2 in boiling ethanol, for 2 hours, gave unsymmetric mercurial 12b in ca 70% crude yield, which next was readily symmetrized by a hot (90° C) KI/EtOH/H2O solution, for 2 hours, to afford symmetric mercurial 13 in 60% crude yield. Further proof of the structures of 12b and 13 was given by their routine bromo- and iodo-demercuration reactions, carried out for ca 30 minutes in hot (80° C) aq. KBr3 or KI3 solutions, respectively. These two reactions furnished the respective 6-bromo or 6-iodo derivatives of 12a, i.e. pure compounds 12c and 12d, in 70% yield, the same for the two mercurials 12b and 13, and the same for the two different reactions.
The above syntheses of the mercurials 12b and 13 may open up an easy way for preparing many other 6-substituted derivatives of 12a (e.g. 12, X = F, Cl, CN) by the well-checked by us demercurating procedures explained in Sections 2-4, or otherwise. It would also be possible to prepare readily from 12b and/or 13 many metallo- and metalloido-organic derivatives of compound 12a, by using known [23, 26, 28, 34] methods applicable for ArHgX and/or Ar2Hg mercurials.
6. 5-Substituted 1,3-Dimethyluracil and 1-Methyluracil Mercurials and Their Demercurating Reactions
(see Refs. 6, 8 and 12)
Uracil [Pyrimidine-2,4(1H,3H)-dione] hardly fits to be directly C-mercurated, since it forms at once a sparingly soluble 1:1 mercury complex (N-Hg salt; formula 1 in Section 1) [59]; see however Ref. 60, which will be discussed in Section 8. Contrariwise, 1,3-dimethyluracil (14a) can be readily directly C-mercurated in its position 5. We applied the routine [23, 24, 26, 28, 34] mercurating procedures by reacting 14a either with a boiling aq. Hg(OCOCH3)2 solution containing CH3CO2H, for 4 hours (which gave the purified mercurial 14b in 52% yield), or with a boiling buffered C2H5OH/H2O solution of HgCl2, for 4 hours, which gave the purified mercurial 14c in 64% yield. The same 14c was also obtained in 45% yield on adding NaCl to a hot aq. solution of 14b. Subsequently, we prepared symmetric pure mercurial 15 by reacting 14b with a hot boiling aq. KI solution, for 0.5 hour, in 89% yield, after recrystallization from water.
Subsequently, continuing our previous studies related in Sections 2 - 4, we carried out the following chemical experiments on halo-demercuration of the 1,3- dimethyluracil mercurials 14b and/or 15, viz.
i) on reacting 14b or 15 with hot aq. KI3 solutions, we obtained known pure 5-iodo-1,3-dimethyluracil (14d) in 88% or 83% yields, respectively;
ii) on reacting 14b or 15 with (adjusted in advance to pH = 7) hot (80° C) aq. KBr3 solutions, we obtained known pure 5-bromo-1,3-dimethyluracil (14e) in the same 82% yield;
iii) on reacting only dry 15 with a large excess of pure liquid S2Cl2, we obtained known pure 5-chloro-1,3-dimethyluracil (14f) in 74% yield;
iv) on reacting only dry 15 with a large excess of neat liqufied SF4for 48 hours at -(60-70oC), and following workup, explained in Ref. 12, pure compound 14g was afforded in 30.3% yield [12].
So far, there is no easy method of removing N-methyl group(s) from the uracil ring system. Therefore, there is (so far) no easy way of transforming compounds 14d, 14e and 14f into the respective non-methylated analogues. See, however, Section 7, where such the demethylation was possible for 5-substituted 2,4-dimethoxypyrimidines.
Uracil has two N-H groups, which differ chemically and otherwise, e.g. at 25° C [61]: pKa (1-N-H) = 9.43; pK¢a (3-N-H) ca. 13.2 (estimate). In order to check whether or not the 1-N-H group alone must be blocked to accomplish a successful direct mercuration in position 5, we synthesized 1-methyl-uracil (16a), and we subsequently carried out its routine mercuration with a boiling, slightly acidified with CH3CO2H, aq. Hg(OCOCH3)2 solution, for 3.5 hours, which resulted in the formation of mercurial 16b (purified) in 49% yield; by its bromo-demercuration, similarly to that of 14b, we obtained 5-bromo-1-methyluracil (purified) in 78% yield. In fact, also other 1-N-substituted uracils, e.g. uridine [62] and 2¢ -deoxyuridine [63] were readily mercurated in their positions 5 with buffered aq. Hg(OCOCH3)2 solutions.
7. 5-Substituted 2,4-Dimethoxypyrimidine Mercurials and Their Demercurating Reactions as well as an Indirect Method for Preparing 5-Halogeno-substituted Uracils
(see Refs. 6, 9 and 12)
2,4-Dimethoxypyrimidine (17) was readily mercurated with a boiling aq. Hg(OCOCH3)2 solution acidified with CH3CO2H, for 2 hours, and this hot solution containing 18 was applied at once in subsequent reactions (vide infra) [only a small isolated sample of 18 was recrystallized from ethanol, and its 1H NMR spectrum was run for identification purposes]. By adding an aq. NaCl solution to the afore-said hot solution containing 8, the metathesized, insoluble mercurial 20 was precipitated out and collected to give 48% crude yield. By adding dropwise a nearly saturated aq. KI solution to the afore-said hot solution containing 18, and refluxing for 0.5 hour, the symmetrized mercurial 21 (recrystallized from water) was isolated in 79% yield. Similarly, an appropriate amount of I2 was added to the boiling afore-said solution containing 18, which resulted (after cooling) in the isolation of iodo derivative 19a (purified) in 58% yield; alternatively, the same compound 19a (purified) was obtained from solid organomercurials 21 by its iodo-demercuration with a saturated hot (60° C) aq. KI3 solution, for 45 minutes, which resulted in 52% yield of 19a. Bromo-demercuration of 18 was attained by adding a neutralized aq. KBr3 solution to the afore-said solution containing 18, and next the reaction was carried out at 60-80° C for 45 minutes – isolation of bromo derivative 19b (purified) resulted in 51% yield; also solid organomercurial 21 was similarly bromo-demercurated to give 19b (purified) in 55% yield. Solid organomercurial 21 was chloro-demercurated with excess of the freshly redistilled liquid S2Cl2 (to avoid the undesirable presence of SCl2; see Section 4 for the explanation). The reaction underwent for 8 hours at 20° C, and the reaction mixture was left overnight – after its two-step workup, pure 5-chloro-2,4-dimethoxypyrimidine (19c) was afforded in 49% yield. On reacting symmetric mercurial 21 with a large excess of neat liqufied SF4 at -(60-70) for 48 hours and the following workup explained in Ref. 12, pure compound 19d was obtained.
5-Halogeno-2,4-dimethoxypyrimidines 19a, 19b, 19c, and 19d, were readily demethylated by adding them to a 10% aq. hydrochloric acid (used in a large excess), and by evaporating the solutions to the dryness on a boiling water bath; the reactions were accompanied with a vigorous foaming. The residues were recrystallized to give pure 5-halogenosubstituted uracils 22a, 22b, 22c, and 22d in 69%, 83%, 69% and 81% [12] yields, respectively. The title method is named to be indirect, because compound 17 is prepared from uracil, converted at first by POCl3 into 2,4-dichloropyrimidine, which next is reacted upon with CH3ONa, yielding finally compound 17 [64]. In reverse, also the direct demethylation of 2,4-dimethoxypyrimidine (17) was easily accomplished on heating 17 with hot hydrochloric acid [65].
8. 5-Substituted Uracil Mercurials and Their Demercurating Reactions
(see Refs. 4, 6, 9, 12 and 14)
In Section 6 we reported that 1-methyluracil (16a) may be readily C-mercurated to give 5-(acetoxymercurio)-1-methyluracil (16b) which next was bromo-demercurated to give 5-bromo-1-methyluracil (16c). There is (so far) no easy and effective method for removing N-methyl group(s) in the uracil ring system, hence it is not possible to obtain e.g. 5-bromouracil (26b) from compound 16c. However, it is known [66] that 1-acetyluracil (23) is readily deacetylated even by cold water to form the initial uracil. We made numerous attempts to mercurate 23 in various anhydrous solvents with the less reactive mercurating agents, viz. HgCl2 or Hg(OCOCH3)2. On prolonged heatings only the sparingly soluble N-Hg salt of uracil (formula 1 in Section 1) was precipitated out in excellent yields, probably with a negligible C-mercuration in position 5. Finally, we effectively C5-mercurated 23 in boiling anhydrous acetonitrile, for 10 hours, but only with the strongly electrophilic Hg(OCOCF3)2. We did not isolate the intermediate mercurial 24, but its yellowish, clear hot solutions were applied as such in the following reactions (vide infra); when the same mercuration was carried out, quite similarly, in CD3CN, then the presence of the soluble mercurial 24 was confirmed with 1H NMR spectrum. If we did prolong for more than ten hours the above mercuration reaction, a white solid was precipitated out in still increasing amounts, and it represented the N-Hg salt 1, but substituted with the HgOCOCF3 group in its position 5 (which was confirmed chemically, spectrally, and partly analytically). The desirable 5,5¢ -mercuriobis(uracil) (25) was obtained in 50% crude yield by adding excess KI dissolved in dry acetonitrile to the afore-said yellowish solution of 24. This mixture was refluxed for 4 hours, cooled, and filtered. The collected white precipitate was heated for 30 minutes at 60° C with a saturated aq. KI solution [which also split off the 1-N-acetyl groups] and after cooling, the collected white precipitate was washed well with boiling water, dried, and analyzed. Its structure was confirmed by the following routine iodo-demercuration reaction with a hot aq. KI3 solution to give known compound 26a (purified) in 93% yield; see Ref. 4 for details. Alternatively, the same iodo-demercurating reaction was carried out but with using the afore-said yellowish boiling solution of 24, which was treated at first with diiodine and next the collected white precipitate was heated for 30 minutes, at 600C, with an aq. KI solution to give finally 26a (purified) in 60% yield.
The respective routine bromo-demercurationreactions with hot aq. KBr3 solutions (adjusted to pH = 7) of either solid 25 or 24 (in its CH3CN solution) were reported elsewhere [9], and gave known purified 5-bromouracil (26b) in 84% and 77% yields, respectively. The above results confirm a somewhat lesser reactivity of unsymmetric mercurial 24 as compared with symmetric mercurial 25.
The chloro-demercuration of symmetric mercurial 25 was reported in our paper [4]. Solid 25 was slowly added to the freshly–redistilled S2Cl2 (used in excess), and this was left overnight at room temperature. The collected precipitate was washed with dry CH3CN, recrystallized from ethanol to give known purified 5-chlorouracil (26c) in 65% yield.
The novel fluoro-demercuration procedure, submitted in Ref. 12, gave 5-fluorouracil (26d) in 27.1% yield, on acting upon symmetric mercurial 25 with a large excess of neat liquid SF4 for 48 hours at ca -600C, and the following workup; see Section 8 for another method of preparing compound 26d.
Note. The electron-donating groups, viz. the methyl or methoxy groups in caffeine, 2,3-diphenyl-5-methoxybenzo[b]furan, 1,3-dimethyluracil and 1-methyluracil, and 2,4-dimethoxypyrimidine do increase more or less an electron density in the substituted parent heterocyclic systems, facilitating thus their direct mercuration even with less electrophilic mercuric salts, viz. mercuric acetate, sometimes also with HgCl2 in its buffered aq. solutions. In contrast, any electron-withdrawing N-acyl groups (e.g. the acetyl, trichloroacetyl or trifluoroacetyl groups), which were introduced by us into the uracil, theophilline or theobromine parent ring systems, considerably render difficult their direct mercurations with Hg(OCOCH3)2, and do eliminate at all the use of HgCl2. It is why we had to use the strongly electrophilicHg(OCOCF3)2, usually prior prepared in situ, to mercurate effectively N-acylated uracil (Section 8), theophilline (Section 9), and theobromine (Section 10) in anhydrous solvents to obtain possibly highest yields of the desired C-substituted unsymmetric mercurials.
Visser et al. [60] have succeded to synthesize microquantities of radioactive 5-X-uracils (X = 211At or 131I) on acting upon 45 m mol of uracil, dissolved in 1 ml of 0.2 M aq. H2SO4, with HgSO4 (40 m mol), for 3 hours at room temperature, followed with NaCl (90 m mol). Without isolating the intermediate 5-(chloromercurio)uracil, they added subsequently 0.9 equivalent of 211At/I2 or 131I/I2, which resulted in very good radiochemicalyields of final radioactive products. Similar approaches were used, with varying the reaction times and temperatures, e.g. for imidazole, thyrosine, phenylalanine, etc. We scaled up [14] the aforementioned procedure thousandfold (i.e. to the milimolar scale) as to directly C5-mercurate uracil as well as, for the sake of comparison, 2-thiouracil and theobromine. Only uracilgave 5-(chloromercurio)uracil in 85% crude yield; its following routine iodo-demercurationwith a hot (80° C) aq. KI3 solution led to 5-iodouracil in ca 61% crude yield. All our attempts to symmetrize 5-(chloromercurio)uracil by means of hot aq. KI, Na2S2O3 or KSCN solutions as well as by a methanolic solution of hydrazine were hardly successful. Therefore, the symmetric mercurial 25 should be prepared in the way explained above. Theobromine reacted as above [60] but in the milimolar scale, furnished 1-N-(chloromercurio)theobromine in ca 77% crude yield, which being iodo-demercurated gave the initial theobromine in ca 72% yield, with no detectable amount of the expected 8-iodotheobromine; see also Section 12. Similarly, 2-thiouracil reacted as above [60] but in the milimolar scale, formed nearly quantitatively an insoluble raw mercurial, which after its demercuration with an aq. KI solution, yielded solely the initial 2-thiouracil with no detectable amount of any C-mercurated product. We concluded [14] the above experiments as follows: "the direct mercuration procedure offered in Ref. 60 should always be tried in the future, since it is relatively simple and less hazardous than the other ones (cf. Section 11), though it is less general than it has been expected and wanted".
9. 8-Substituted Theophylline Mercurials and Their Demercurating Reactions
(see Refs. 3 and 4)
When theophylline as well as theobromine and uracil, all having the acidic N-H group, are reacted with Hg(II) salts, then their sparingly soluble N-Hg salts (see formulae 1 – 3 in Section 1) are immediately precipitated out from the mercurating solutions [63, 66], and their effective C-mercuration can hardly be performed; see Section 8, where this topic is discussed. N-Acetyl derivatives of theophylline and theobromine [67], are easily hydrolyzable by the action of water, likewise as does 1-acetyluracil (Section 8). We made many attempts to mercurate 7-acetyltheophylline (27a) in various anhydrous media, e.g. in CH3CO2H/(CH3CO)2O mixtures, with less reactive mercurating agents, viz. HgCl2 or Hg(OCOCH3)2, but all our experiments were without success for the same reasons as those explained for 1-acetyluracil (Section 8). We succeeded to C-mercurate theophylline system in its position 8 (which is somewhat more reative than the same position in theobromine and caffeine), but only with using the strongly electrophilic Hg(OCOCF3)2 dissolved in boiling anhydrous mixtures made of CF3CO2H and (CF3CO)2O. The reaction was completed after 10 hours, the resulting reaction mixture was concentrated under dimished pressure, water was added to the viscous residue, and a white collected precipitate was recrystallized from water to give
8-(trifluoracetoxymercurio)theophylline(29) in 53% yield. We also simplified the above procedure as follows. Theophylline was refluxed with (CF3CO)2O for 2 hours [thus, 7-(trifluoroacetyl)theophylline (27b) was produced there in situ], then a solution of Hg(OCOCF3)2 in CF3CO2H/(CF3CO)2O was added, and the same mercuration reaction was carried out under a reflux condenser for 10 hours; the following workup was the same as previously to afford mercurial 29 in 64% yield. This latter yield was next increased to 70%, when 7-(trichloroacetyl)theophylline (27c) was applied as the starting substrate (Ref. 4; see footnote in p. 386).
Next, we carried out the routine iodo- and bromo-demercuration reactions with mercurial 29 in hot aq. KI3 or KBr3 (adjusted to pH = 7) solutions, at 80° C for 30 minutes, which resulted in preparing the purified compounds 30a and 30b in 95% and 96% yields, respectively. However, our attempts to prepare 8-chlorotheophilline (30c) from dry unsymmetric mercurial 29, and with using pure liquid S2Cl2 or SCl2 as the chloro-demercurating agents, were unsuccessful.
10. 8-Substituted Theobromine Mercurials and Their Demercurating Reactions
(see Refs. 3, 4 and 13)
Since all our experiments on the C-mercuration of theobromine and theophylline, published in [3], were performed quite parallelly and were very close ones to the others, hence there is no need to relate here all our unsuccessful preliminary experiments aimed at the preparation of the desired theobromine mercurial(s). As previously, we succeeded in the mercuration of 1-acetyltheobromine (31a) only with Hg(OCOCF3)2 in boiling anhydrous CF3CO2H/(CF3CO)2O mixtures, for 10 hours. The following workup was practically the same as that for preparing mercurial 29, which gave
8-(trifluoroacetoxymercurio)theobromine (32a) in 56% yield. Later (Ref. 4; see footnote in p. 386), we increased this yield to 70% with using
1-(trichloroacetyl)theobromine (31c) as the starting substrate. As previously, the solid unsymmetric mercurial 32a was next iodo- and bromo-demercurated with hot (80° C) aq. KI3 or aq. KBr3 (pH = 7) solutions, under nearly the same reaction conditions as those related in Section 9, which resulted in the purified 8-iodo- (32b) and 8-bromo-theobromine (32c) obtained in 95% and 92% yields, respectively. Our attempted chloro-demercurating experiments with mercurial 32a acted upon with neat S2Cl2 or SCl2 were quite unsuccessful as well.
Later [13], we prepared pure 8,8¢ -mercuriobis(theobromine) (33) in 88%
yield, though only from 1,8-bis(acetoxydimercurio)theobromine
(34). Using the routine iodo- and bromo-demercuration methods with the
symmetric mercurial 33, we obtained the purified either
8-iodo- (32b) or 8-bromo-theobromine (32c) in 97%
and 92% yields, respectively. However, the contents of our paper [13] will be
discussed more extensively in the following Section 11.
11. Preparation of 1,8-bis(Acetoxydimercurio)theobromine and Its Reactions
(see Ref. 13)
Australian chemists [52] have prepared numerous fully mercurated (permercurated) arenes by the reaction of an excess of moltenHg(OCOCF3)2 with suitable arenes at ca 180-245° C; the reaction temperature depended on the reactivity of a given starting arene. The crude(not analyzed) permercurated arenes were subsequently halo-demercurated to give the corresponding perhalogenated arenes. Therefore, we expected that the same melting chemical procedure would be suitable for the direct permercuration of several fairly stable, N-H acidic, lactamic heterocycles, e.g. theobromine, theophylline, xanthine, hypoxanthine, uracil, etc. For our preliminary study we have chosen theobromine, which mercurials are discussed in Section 10.
In order to accomplish the said melting chemical procedure, we intimately mixed pure theobromine with an excess of molten Hg(OCOCF3)2 (prepared in situ). A vigorous reaction was observed with the evolution of geseous, strongly toxic products as well as tiny droplets of metallic mercury. This is why, in our opinion, the organic derivatives of mercury(I) would readily be formed in the hot reaction mixture.
Note. We submitted in our paper [13] the following assumption (see footnote in p. 30): Mercury in statu nascendi (Hg*), probably formed during the thermal decomposition of Hg(OCOCF3)2, would, at a guess, react readily with intermediates of mercury(II), forming the resulting compounds of mercury(I), viz.
The main problem was to control properly the reaction temperature, since the afore-said reaction mixture thickened continuosly, until it had solidified. Probably, it would be desirable in future to look for some inert solvents, e.g. some highly boiling hydrocarbons or perfluorinated hydrocarbons, in order to control better the reaction. For the time being, we obtained some crude melt, evidently containing tiny droplets of metallic mercury as well as some highly mercurated products, very likely mostly 1,8-bis(trifluoroacetoxydimercurio)theobromine. The latter tentative opinion was further indirectly supported by the subsequent reaction. On boiling the powdered crude melt with glacial acetic acid – where the following metathesis would take place: -Hg-Hg-OCOCF3 + excess CH3COOH ? -Hg-Hg-OCOCH3 + CF3CO2H – a considerably purified yellow product was obtained (after the concentration of the CH3CO2H extract under a diminished pressure) in 28% yield calculated in respect of the starting theobromine. Its analysis is given above. Its 1H NMR spectrum (in CF3COOD) shows the absence of any 8-H low-field proton signal at ca. 8.5 ppm, and the presence of two N-CH3 three-proton signals (at 3.38 and 4.08 ppm) characteristic of the theobromine [3]. At ca 1.8 ppm there is, however, an additional and intense six-proton singlet derived from two, apparently spectrally equivalent, CH3COO groups. Thus, it may be guessed that the two assumed –Hg-Hg-OCOCH3 groups are substituted in mercurial 34 in its 1 and 8 positions. The lack of any 1-NH proton signal is insignificant, since it might have been due to a quick isotopic exchange with the deuterated solvent. Hence, the said substitution in position 1 is better explained by the full absence of any characteristic N-H absorption band over the 3070-3150 cm-1 IR range which, in contrast, is found at 3120 cm-1 in the comparative IR spectrum of theobromine taken also in Nujol.
This considerably purified product 34 was used as such in the following reactions with no further purification. Compound 34 was readily iodo- and bromo-demercurated with either a boiling I2 solution in dry CH3CN, for 1.5 hours, or by an aq. KBr3 solution, at 80° C for one hour, to give finally either 8-iodotheobromine (32b) (purified) in 97% yield, or 8-bromotheobromine (32c) (purified) in 92% yield; in the course of both demercurating procedures, the (N1)-Hg-Hg-OCOCH3 groups are exchanged in full by the hydrogen atoms (proto-demercuration), whereas those attached to the C8 atom of 34 are replaced by the respective halogen atoms (halo-demercuration). The same difference was observed in the course of our attempted symmetrization of compound 34 by a hot ethanolic solution of hydrazine (the other symmetrizing agents were either ineffective or gave worse results). Only the C8-substituted –Hg-Hg-OCOCH3 groups were all engaged in the formation of -Hg- bridging present in the formed 8,8¢ -mercuriobis(theobromine) (33), which was obtained, after its purification, in 88% yield [its structure was confirmed analytically, spectrally as well as by its subsequent iodo- and bromo-demercuration reactions; see Section 10], whereas those groups substituted in 1-N position were readily split off in favor of the hydrogen atom. All these reactions are new in the field of organic derivatives of mercury(I) – to the best of our knowledge.
Note. Organic C-derivatives of mercury(I) so far are rare and generally regarded as unstable. For example, controlled electrolysis of 1,6-dibromohexane with a mercury cathode gives the dimer [Me(CH2)5Hg]2 [68].
It is impossible not to mention that the permercurated
arenes prepared by the Australian chemists [52] had been neither purified nor
analyzed, but they were immediately used as such in the subsequent
halo-demercuration reactions, resulting in numerous, purified and analyzed,
perhalogenated arenes. Nobody knows whether the said intermediate
permercurated arenes were, in fact, the derivatives of mercury(II), mercury(I)
or, possibly, were composite mixtures of both; it should be elucidated in
future.
12. Further Studies on Some Heteroaromatic Mercurials
(see Ref. 14)
Using various known methods of the direct C-mercuration it was attempted to mercurate: uracil, thiouracil (35a), S-ethyl-2-thiouracil (35b), theobromine (36a), and 6-O-ethyltheobromine (37). Only uracil was effectively C5-mercurated in ca 85% crude yield giving the raw 5-(chloromercurio)uracil (Section 8), whereas the rest formed (in good yields) either S-Hg bonds (compounds 35a and 35b) or N-Hg bonds (compounds 36a and 37); in the latter its 6-O-ethyl groups were wholly split off under the reaction conditions. All the newly obtained raw mercurials were next iodo-demercurated by the routine methods [23-25, 34, 42]. This furnished from S-ethyl-2-thiouracil mercurial the corresponding C-iodinated product, i.e. S-ethyl-5-iodo-2-thiouracil (35c) in ca 64% crude yield. The other mercurials were mostly proto-demercurated, though with some noticeable exceptions discussed below. The main purpose of our work [14] was to detect whether the afore-said raw mercurials do contain (or not) any detectable ammounts of C-mercurated admixtures, which next would have been transformed into recognizable C-iodinated derivatives formed in the subsequent iodo-demercuration reactions.
It has been shown in Section 7 that 2,4-dimethoxypyrimidine (17) was effectively C5-mercurated on boiling with an aq. Hg(OCOCH3)2 solution acidified with CH3CO2H, for 2 hours. Hence, we did expect that compound 37 would behave similarly to give a true mercurial C-substituted in position 8. Thus, compound 37 was refluxed for 55 hours (the reaction was controlled with TLC) with a Hg(OCOCH3)2 solution in boiling glacial acetic acid. Quite unexpectedly, we obtained a new mercurial 36b, i.e. 1-(acetoxymercurio)theobromine, in 83% yield, which structure was confirmed by chemical tests and its IR and 1H NMR spectrum; no 1H NMR spectral evidence was found for the C8-mercuration. The 6-O-ethyl groups in compound 37 were wholly split off under the reaction conditions, forming momentarily the parent tautomeric theobromine (36a), which immediately reacted with Hg(OCOCH3)2 to give the new mercurial 36b. Our guess was supported as follows: compound 37 was refluxed for 48 hours solely with neat CH3CO2H; after cooling, we isolated theobromine (36a) in 83% yield, which was evidenced chemically and spectrally. We also obtained another new mercurial 36c, i.e.
1-(chloromercurio)theobromine in ca 77% crude yield (Section 8). The both crude mercurials 36b and 36c were refluxed for 2 hours with excess diiodine dissolved in dry CH3CN; after workups this furnished theobromine (36a) in 76% and 72% yields, respectively. No detectable amounts of known 8-iodotheobromine – which would have supported the searched C8-mercuration – were found in the both reaction mixtures furnishing only theobromine.
It is necessary to add that, so far, only one theobromine mercurial in the ratio 2 : 1 Hg was reported [69]; its structure is shown in Section 1. In the same paper [69] also other N-Hg mercurials furnished from theophylline, hypoxanthine, xanthine, guanine and uracil (the latter is a 1 : 1 Hg complex shown in Section 1) were also reported.
S-Ethylthiouracil new mercurial (supposedly a 1 : 1 Hg complex) was synthesized as follows. Compound 35b was dissolved in CH3OH acidified with two drops of conc. aq. HClO4 added. Then the solution of Hg(OCOCH3)2 in methanol was added to the former solution, the mixture was refluxed for 8 hours, and then left overnight. The collected white precipitate was practically insoluble in common solvents, and was obtained in ca 42% crude yield. When this novel crude mercurial was boiled with a conc. aq. KI solution until the whole was clear and slightly yellowish, then after cooling we isolated the recovered compound 35b in 99% yield. But the same crude mercurial being refluxed with an aq. KI3 solution for 30 – 40 minutes, furnished unexpectedly (after cooling) the iodo derivative35c in ca 64% crude yield; it was recrystallized from ethanol and 2-propanol yielding pure 35c.
13. Final Results of Our Halo- and Cyano-demercurating Reactions
(see Refs. 1-15)
13.1. Aromatic Iodides from Aromatic Organomercurials
It is known [23, 26, 28, 34, 42] that aromatic iodides and bromides may be obtained from both symmetric and unsymmetric aromatic organomercurials, by reacting them with either hot aq. KI3 and KBr3 neutralized solutions or with hot solutions of diiodine and dibromine in acetonitrile, DMF, etc. We had repeatedly applied the "classic" iodo-demercuration and bromo-demercuration reactions to elucidate better the chemical structures of organomercurials prepared by us in various ways. Thus, we synthesized (the yields given in brackets) the following aromatic iodides and bromides, viz.
i) 8-iodocaffeine (95%) from both 8-(acetoxymercurio)caffeine and 8,8’-mercuriobis(caffeine) [1, 2];
ii) 8-iodotheophylline (95%) from 8-(trifluoroacetoxymercurio)theophylline, and 8-iodotheobromine (95%) from 8-(trifluoroacetoxymercurio)theobromine [3];
iii) 5-iodouracil (60 or 93%, resp.) from either 1-acetyl-5-(trifluoroacetoxymercurio)uracil or 5,5’-mercuriobis(uracil) [4];
iv) 2-iodofuran (60%) from 2,2’-difurylmercury, and 2-iodothiophene (65%) from 2,2’-dithienylmercury [5];
v) 5-iodo-1,3-dimethyluracil (87.5 or 82.5%, resp.) from either 5-(acetoxymercurio)-1,3-dimethyluracil or 5,5’-mercuriobis(1,3-dimethyluracil) [8];
vi) 5-iodo-2,4-dimethoxypyrimidine (58 or 52%, resp.) from either 5-(acetoxymercurio)-2,4-dimethoxypyrimidine or 5,5’-mercuriobis(2,4-dimethoxypyrimidine); the acid hydrolysis of pure 5-iodo-2,4-dimethoxypyrimidine afforded 5-iodouracil (69%) [9];
vii) 6-iodo-2,3-diphenyl-5-methoxybenzo[b]furan (70%) from both 6-(acetoxymercurio)-2,3-diphenyl-5-methoxybenzo[b]furan and 6,6’-mercuriobis(2,3-diphenyl-5-methoxybenzo[b]furan) [10];
viii) iodobenzene (59%) from (chloromercurio)benzene, and 8-iodocaffeine (65%) from 8-(chloromercurio)caffeine [11];
ix) 8-iodotheobromine (97%) from 1,8-bis(acetoxydimercurio)theobromine; the latter compound was the first stable organic derivative of mercury(I) [13]
x) 5-iodouracil (61% crude yield) from 5-(chloromercurio)uracil, and 2-ethylthio-4-hydroksy-5-iodopyrimidine (64% crude yield) from 2-ethylthio-4-hydroxypyrimidine raw mercurial [14].
Note. We also applied twice another iodo-demercuration method, by reacting two symmetric organomercurials with hot either aqueous [1, 2] or ethanolic [7] solutions of cyanogen iodide, ICN. In this way we synthesized:
xi) 8-iodocaffeine (90%) from 8,8’-mercuriobis(caffeine) [1, 2], and
xii) iodobenzene (72%) from diphenylmercury [7].
13.2. Aromatic Bromides from Aromatic Organomercurials
i) 8-bromocaffeine (83% or 85%, resp.) from either 8-(acetoxymercurio)caffeine or 8,8¢ -mercuriobis(caffeine) [1, 2];
ii) 8-bromotheophylline (96%) from 8-(trifluoroacetoxymercurio)theophylline [3];
iii) 8-bromotheobromine (96%) from 8-(trifluoroacetoxymercurio)theobromine [3, 13];
iv) 5-bromo-1,3-dimethyluracil (82% or 84,5%, resp.) from either 5-(acetoxymercurio)-1,3-dimethyluracil or 5,5¢ -mercuriobis(1,3-dimethyluracil) [8];
v) 5-bromo-2,4-dimethoxypyrimidine (51% or 55%, resp.) from either 5-(acetoxymercurio)-2,4-dimethoxypyrimidine or 5,5¢ -mercuriobis(2,4-dimethoxypyrimidine); the acid hydrolysis of pure 5-bromo-2,4-dimethoxypyrimidine afforded 5-bromouracil (83%) [9];
vi) 5-bromouracil (77% or 84%, resp.) from either 5-(trifluoroacetoxymercurio)uracil or 5,5¢ -mercuriobis(uracil) [9];
vii) 6-bromo-2,3-diphenyl-5-methoxybenzo[b]furan (70%) from both 6-(acetoxymercurio)-2,3-diphenyl-5-methoxybenzo[b]furan or 6,6¢ -mercuriobis(2,3-diphenyl-5-methoxy-benzo[b]furan) [10].
viii) 8-bromotheobromine (92%) from both 8,8`-mercuriobis(theobromine) or 1,8-bis(acetoxydimercurio)theobromine [13].
Note. We also applied twice our novel bromo-demercuration method, with applying a labile complex (or adduct) of an unknown composition, (BrCN× KBr)x, suspended either in hot water [1, 2] or dissolved in boilling ethanol [7]. In this way we obtained:
ix) 8-bromocaffeine (85%) from both 8-(acetoxymercurio)caffeine or 8,8`-mercuriobis(caffeine) [1, 2];
x) bromobenzene (79%) from diphenylmercury [7].
13.3. Aromatic Chlorides from Aromatic Organomercurials
During our studies we discovered three novel chloro-demercurating reagents, viz. S2Cl2, SCl2, or (the least reactive) SO2Cl2. Mostly, they were applied as pure liquids, but for furan and thiophene mercurials, S2Cl2 was necessarily diluted with CS2 (Section 4). Thus, we obtained the following results:
i) 8-chlorocaffeine (90%) from 8,8¢ -mercuriobis(caffeine), using neat liquid S2Cl2, or (70%), using neat liquid SCl2, or (27%) using neat liquid SO2Cl2 [1, 2];
ii) 5-chlorouracil (65%) from 5,5¢ -mercuriobis(uracil), using neat liquid S2Cl2 [4];
iii) 2-chlorofuran (60%) from 2,2¢ -difurylmercury, using S2Cl2 diluted with CS2, or (50%) from a mixture of furan and dry finely-powdered HgCl2 dissolved in excess S2Cl2 [5];
iv) 2-chlorothiophene (70%) from 2,2¢ -dithienylmercury, using S2Cl2 diluted with CS2,or (60%) from a mixture of thiophene and dry finely-powdered HgCl2 dissolved in excess S2Cl2 [5];
v) chlorobenzene (78%) from diphenylmercury, using neat liquid S2Cl2, or (69%) from (trifluoroacetoxymercurio)benzene, using neat liquid S2Cl2, or (50%) from (acetoxymercurio)benzene, using neat liquid S2Cl2,or (47%) from diphenylmercury, using neat liquid SO2Cl2 [5];
vi) 5-chloro-1,3-dimethyluracil (74%) from 5,5¢ -mercuriobis(1,3-dimethyluracil), using neat liquid S2Cl2 [8];
vii) 5-chloro-2,4-dimethoxypyrimidine (49%) from 5,5`-mercuriobis(2,4-dimethoxypyrimidine), using neat liquid S2Cl2; the acid hydrolysis of pure 5-chloro-2,4-dimethoxypyrimidine afforded 5-chlorouracil (69%)[9].
13.4. Aromatic Fluorides from Aromatic Organomercurials
On acting successively (at ca - 600C) upon several symmetricorganomercurials (vide infra) with a great excess of neat liquid SF4(b.p. - 40.40C) we obtained the following monofluorinated products:
i) 8-fluorocaffeine (30%) from 8,8¢ -mercuriobis(caffeine) [1, 2];
ii) fluorobenzene (58%) from diphenylmercury [12];
iii) 5-fluoro-1,3-dimethyluracil (30.3%) from 5,5¢ -mercuriobis(2,4-dimethyluracil) [12];
iv) 5-fluoro-2,4-dimethoxypyrimidine (32.6%) from 5,5¢ -mercuriobis(2,4-dimethoxypyrimidine); the former compound was demethylated to give 5-fluorouracil in 81.3% yield [12].
v) 5-fluorouracil (27.1%) from 5,5¢ -mercuriobis(uracil) [12].
13.5. Aromatic Nitriles from Aromatic Organomercurials
i) 8-cyjanocaffeine (50% or 25%, resp.) from either 8,8¢ -mercuriobis(caffeine) or 8-(acetoxymercurio)caffeine, with freshly purified BrCN in hot (600C) water [1, 2];
ii) benzonitrile (86%) from diphenylmercury, with freshly purified BrCN in boiling dry benzene [7].
14. Improved Syntheses of Some Diaryliodonium Salts from Symmetric Diarylmercurials and (Dichloroiodo)arenes (Willgerodt`s Method)
(see Ref. 16)
Willgerodt [70-72] had reacted cold (or hot [71]) aqueous suspensions of equal masses [it practically means that the mercurials were used in a deficit] of powdered PhICl2 with powdered Ar2Hg (where Ar = phenyl, 2- and 4-tolyl, and 2-naphthyl) to afford the respective diaryliodonium chlorides (yields were not reported); sparingly soluble ArHgCl and other admixtures, e.g. PhIO [73], were hot-filtered off and discarded, viz.
Beringer and Lillien [74] applied the Willgerodt method to obtain three unsymmetric diaryliodonium chlorides. They obtained only 4-acetamidophenyl(phenyl)iodonium chloride (which was precipitated out as its sparingly soluble iodide, isolated in 10% crude yield) by reacting equimolar amounts of Ph2Hg with 4-AcNHC6H4ICl2 in hot water (40-50oC) for 12 hours. We obtained [16] the same iodide, but in 80% crude yield, by reacting equal masses of PhICl2 with symmetric 4,4-mercuriobis(acetanilide) suspended in stirred hot water (40-50oC) for 12 hours. We explained this evident yield’s increase as follows:
Consequently, by reacting equal masses of PhICl2 with symmetric 4,4-mercuriobis(N,N-dimethylaniline), suspended in stirred hot water (40-50oC) for 12 hours, we isolated from the hot filtrate, after its cooling, 4-dimethylaminophenyl(phenyl)iodonium chloride in 60% crude yield. Previously, Beringer and Lillien [74] failed to obtain this iodonium salts, para-substituted with only one NMe2 group; the same failure was also reported by Neiland [75]. A similar iodonium salt bearing the two p,p’-substituted NMe2 groups was synthesized quite otherwise [76]; this synthesis is shown (Scheme 7) in our paper [16].
We also attempted, without effect, to synthesize various
8-(aryliodonio)caffeine halides with using the Willgerodt method. Hence, we
used our short-cut, oxidative method [77] to obtain
8-(4-methoxyphenyliodonio)caffeine bromide (49% crude yield) by acidic
coupling of the prior oxidized 8-iodocaffeine with anisole. It is, in fact,
the first iodine(III) derivative of caffeine, which may open novel
routes for preparing 8-substituted caffeines by its reactions with various
nucleophiles [78].