[7439-97-6]  · Hg  · Mercury(0)  · (MW 200.59)

(photosensitizer for C-H bond cleavage reactions;1-6 solvent for metals to give amalgams which can act as selective reducing agents;8-19 heterogeneous catalyst poison, especially for Group 10 metal catalysts;22 electrode material in electrosynthesis24)

Physical Data: mp -38.87 °C; bp 356.6 °C; d 13.6 g cm-3.

Solubility: sparingly sol organic solvents and water.

Form Supplied in: silvery liquid; widely available; electronic grade (foreign metals <=1 ppm) and ACS grade (99.9995%) are available, normally making purification unnecessary.

Handling, Storage, and Precautions: as a volatile heavy metal, care must be taken to prevent long term inhalation of the vapor. Carrying out reactions in a fume hood, cleaning up spills, and assuring that the air handling system of the laboratory is operating normally is generally sufficient.

Mercury Photosensitization.

This allows the dehydrodimerization of a variety of volatile organic molecules on a 1 to 50 g scale without a solvent in a simple apparatus which can be put together from standard laboratory photochemical equipment.1-5 A 254 nm low-pressure mercury lamp is used, together with quartz glassware. The reactive triplet excited state of mercury is usually designated Hg*. The reaction happens in the vapor phase at rt and pressure, but no special precautions have to be taken; the normal vapor pressure of the substrate is usually enough to replenish the vapor above the substrate when reaction has taken place. Mercury is supplied in the form of a small mercury drop in the substrate. Normal laboratory glassware can contain sufficient adsorbed mercury so that Hg photosensitized reactions can occur. To obtain a valid control experiment in the absence of mercury, it is sometimes necessary to anneal the glassware in a glassblower's oven. This implies that care needs to be taken in interpreting the results of what are ostensibly normal or nonmercury photosensitized reactions where special care to remove trace mercury may not have been taken.

The liquid phase is unreactive and this has important effects on the selectivity of the reaction. Dehydrodimerization or functionalization always leads to the condensation of the product, which protects it from further reaction and so the substrate only undergoes one homolysis even if there are several C-H bonds of similar strength. To take a simple example, cyclohexane dimerizes to bicyclohexyl with no further oligomerization. The origin of the nonreactivity of the liquid phase is probably that the very narrow atomic absorption line for vapor-phase mercury is broadened and shifted when the mercury is dissolved. Perfect matching of the emitter and absorber in the vapor phase, where both are vapor-phase mercury atoms, leads to efficient energy transfer in the vapor; energy transfer is probably poor for dissolved Hg. The use of a high-pressure mercury lamp (or even allowing the temperature of the lamp to become abnormally high) causes the selectivity of the reaction to change sharply for the worse, presumably because absorption can now take place in the liquid phase and so condensation no longer protects the condensate from further reaction.

The absorption coefficient for mercury vapor is so large that all the light is absorbed within a few micrometers of the inside surface of the reactor and so the bulk of the vapor is protected from irradiation. This means that normal organic photochemistry is severely suppressed; essentially all the energy is absorbed by the mercury atoms, leaving none to excite the much more weakly absorbing organic species. Absorption of the 254 nm line is responsible for the chemistry described here, and so the excited state responsible is the 3P1 state of mercury. The quartz glassware is transparent to the 254 nm line, allowing for efficient transmission to the reaction zone.

A useful modification for certain substrates is the use of a reactive atmosphere. In this case, the temperature is typically adjusted so that the vapor pressure of the substrate is 100 Torr. This is necessary so that the reactive gas has a substantial partial pressure (P) in the reaction zone; in the case where the substrate has a P(substrate) of 100 Torr, the P(reactive gas) would be 660 Torr. The two most useful reactive gases are H2 (Hg*/H2 conditions)4 and NH3 (Hg*/NH3 conditions).5 The reason for moving to these reactive gases is that the standard conditions, with reflux under N2 (Hg* conditions), completely fail when the Hg* attacks a functional group in the substrate. This typically happens for substrates having multiple bonds or N lone pairs. Under Hg*/H2 or Hg*/NH3 conditions, H atoms take over from Hg* as the active species, and a much wider range of substrates are reactive. For example, NEt3 fails to react at all under Hg* conditions, presumably because Hg* attacks the N lone pair, and energy transfer leads to thermal excitation of the substrate but not to productive chemistry. Under Hg*/H2 conditions the compound undergoes dehydrodimerization at the C-H bond a to N. Arenes seem to work best under Hg* conditions, probably because H atom addition leads to undesired and unselective partial saturation of the aromatic ring. In practice there is little point in attempting to predict what will happen with a given substrate because it is easy to try Hg* conditions first and then move to Hg*/H2 conditions if the results are unsatisfactory.

In each case, the crude product is collected by removing the volatile starting material by rotary evaporation after the reaction is over. The main limitation is that the substrate be volatile, but compounds with up to 16 nonhydrogen atoms have been successfully dimerized. The reaction proceeds under reduced pressure if this is useful to vaporize the substrate. Another useful modification is to use steam distillation to bring the substrate into the vapor phase; in this variant, water is added to the substrate and the mixture is refluxed. The weakest C-H (or X-H) bond in the molecule is homolyzed and the resulting C-centered (or, in general, X-centered) radicals recombine. That part of the radical pool which disproportionates instead of recombining does not in general lead to lower chemical yield because the H atoms present add to the alkene disproportionation product and re-form the initial radical. Quantum yields of 0.04-0.8 are usual and the majority of substrates have values in the range 0.2-0.6. Chemical yields are good to excellent (40-98%). Conversions depend on photolysis time.

The great advantage of the method is that it allows a number of difficult synthetic transformations to be carried out in one step. The synthesis of a few simple compounds that are otherwise very difficult to make is shown below. In the diamine synthesis, Hg*/NH3 conditions gave the best results (eq 1).

A variant, hydrodimerization of alkenes, takes place under Hg*/H2 conditions (eq 2). The H atoms add to the terminal carbon of the alkene to give the intermediate radical shown.

Another useful feature is the facility with which two different substrates cross dimerize (eq 3).

In suitable cases, the volatility or polarity differences among the three allow easy separation by distillation or chromatography (eq 4).

The intermediate radicals can undergo rearrangement in special cases, as in the case of hexenyl radicals which cyclize. Trapping the intermediate radicals with CO, SO2, and O2 has proved possible, giving aldehydes, ketones, sulfonic acids, and hydroperoxides.6

Since methane has strong C-H bonds, it only reacts well under Hg*/NH3 conditions to give CH2=NH as product.5a Arenes do not undergo cleavage of the strong aryl C-H bonds but benzylic C-H bonds of side chain alkyl groups can be cleaved under Hg* conditions; neither Hg*/H2 nor Hg*/NH3 conditions seem to be useful for arenes, however, probably because H atoms readily add to arene rings to give a complex mixture of products.


Ultrasonically dispersed mercury reduces a,a-dibromo ketones to an intermediate that is believed to be a mercurated 2-oxyallyl species which gives a 4-methylene-1,3-dioxolane with acetone.7


This is a traditional and well-established application of Hg0. Metallic mercury readily forms amalgams with most metals but Na/Hg, Al/Hg, and Zn/Hg are the most useful in organic chemistry. The mercury serves both to keep the surface of the metal clean (because inorganic salts adhere poorly to the amalgam) and to dilute the active metal (and so moderate its thermodynamic reducing potential), which can improve the selectivity of the reduction.

Sodium Amalgam.

Sodium Amalgam is a liquid up to 1%, semisolid at 1.2%, and a pulverizable solid at higher concentrations, except in a narrow range around 40% Na, where the material is a low melting (<30 °C) solid. These materials can be made from elemental Na and Hg (caution: much heat is evolved) and analyzed by titration with acid.8 Na/Hg is useful for the reduction of a,b-unsaturated carboxylic acids to the saturated forms, and for the Emde degradation of a quaternary amine (eq 5).9 The reduction of aldonolactones to aldoses with Na/Hg is a key transformation in sugar chemistry.10 Oximes are readily reduced to amines.11

Aluminum Amalgam.

Aluminum Amalgam, readily prepared12 by treating base-etched elemental aluminum with aqueous mercury(II) salts, is a useful replacement for Na/Hg when the compound to be reduced is base sensitive. Diethyl oxaloacetate can be reduced to diethyl malate (70-80% yield) (eq 6) and aryl ketones can be reduced to the corresponding pinacols (30-60% yield) (eq 7) in this way.12 Desulfurization of disulfides is also possible.13

Dienes undergo what is effectively a 1,4-addition of H2 to give the monoenes, and cumulenes undergo a 1,2-reduction.14 The C-S bond in a,b-unsaturated phenyl sulfones can be hydrogenolyzed stereospecifically to give the alkene in excellent yield.15 Net hydrogenolysis of a P=C bond is involved in the sequence shown in eq 8, in which an acyl halide is converted to a keto ester.16a

In a recent synthesis of mannostatin A, King and Ganem have shown how the N-O bond of a cyclic acyl-nitroso compound can be hydrogenolyzed by Al/Hg (eq 9).16b

Zinc Amalgam.

The classic use of Zinc Amalgam is the Clemmensen reduction of ArCOR to ArCH2R.17 Variants of this method have proved successful for specific substrates.18 The nitroalkene closure shown in eq 10 is a more recent application of Zn/Hg.19

Ultrasound20 and Rieke21 methods are increasingly being used as an alternative to Hg amalgamation for activating metals, a trend encouraged by disposal problems of mercury-contaminated wastes.

Catalyst Poison.

Mercury selectively poisons heterogeneous catalysts, particularly of the platinum group metals (PGM). This can be useful when a homogeneous PGM catalyst decomposes with time to give the free metal; in such a case, Hg0 can suppress the heterogeneous component of the reaction.22 This can improve selectivity or give mechanistic information about which products are attributable to which pathway.

Potential Route to Organomercury Compounds from Hg0.

Organomercury compounds are synthetically accessible23 from metallic mercury by a number of routes, including reaction of elemental mercury with alkenes in acid medium and with acyl and alkyl halides (under thermal or photochemical conditions). Organic synthetic applications of this chemistry seem to be very rare, however.

Electrolysis at Mercury Cathodes.

The high overvoltage of a mercury surface in several electrochemical processes is often used to advantage; for example, proton reduction to H2 is kinetically disfavored relative to electron transfer to an organic substrate. An example of an organic electrochemical application is provided by reduction of a number of alkyl halides, RX, to the radical, R&bdot;, which dimerizes to R2, disproportionates to RH and the corresponding alkene, and also leads to the formation of R2Hg.24

1. Brown, S. H.; Crabtree, R. H. TL 1987, 28, 5599.
2. Ferguson, R. R.; Boojamra, C. G.; Brown, S. H.; Crabtree, R. H. H 1989, 28, 121.
3. (a) Brown, S. H.; Crabtree, R. H. JACS 1989, 111, 2935. (b) Brown, S. H.; Crabtree, R. H. JACS 1989, 111, 2946.
4. Muedas, C. A.; Ferguson, R. R.; Brown, S. H.; Crabtree, R. H. JACS 1991, 113, 2233.
5. (a) Michos, D.; Sassano, C. A.; Krajnik, P.; Crabtree, R. H. AG(E) 1993, 32, 1491. (b) Krajnik, P.; Ferguson, R. R.; Crabtree, R. H. NJC 1993, 17, 559. (c) Krajnik, P.; Michos, D.; Crabtree, R. H. NJC 1993, 17, 805.
6. Ferguson, R. R.; Crabtree, R. H. JOC 1991, 56, 5503.
7. Fry, A. J.; Ginsburg, G. S.; Parente, R. A. CC 1978, 1040.
8. FF 1967, 1, 1033.
9. Emde, H. LA 1912, 391, 88.
10. Fischer, E. CB 1890, 23, 930.
11. Hochstein, F. A.; Wright, G. F. JACS 1949, 71, 2257.
12. (a) Wislicenus, H.; Kaufmann, L. CB 1895, 28, 1323. (b) Newman, M. S. JOC 1961, 26, 582.
13. Johnson, J. R.; Buchanan, J. B. JACS 1953, 75, 2103.
14. Kuhn, R.; Fischer, H. CB 1961, 94, 3060.
15. (a) Pascali, V.; Umani-Ronchi, A. CC 1973, 351. (b) Mukaiyama, T.; Narasaka, K.; Maekawa, K.; Furusato, M. BCJ 1971, 44, 2285.
16. (a) Cooke, M. P., Jr. JOC 1982, 47, 4963. (b) King, S. B.; Ganem, B. JACS 1991, 113, 5089.
17. Staschewski, D. AG 1959, 71, 726.
18. (a) Schwarz, R.; Hering, H. OSC 1963, 4, 203. (b) Caesar, D. OSC 1963, 4, 695.
19. Yamada, F.; Makita, Y.; Suzuki, T.; Somei, M. CPB 1985, 33, 2162.
20. (a) Erdik, E. T 1987, 43, 2203. (b) Kitazume, T.; Ishikawa, N. CL 1981, 1679. (c) Han, B.-H.; Boudjouk, P. JOC 1982, 47, 5030.
21. Rieke, R. D.; Li, P. T-J.; Burns, T. P.; Uhm, S. T. JOC 1981, 46, 4323.
22. Anton, D. R.; Crabtree, R. H. OM 1983, 2, 855.
23. Wardell, J. L. In Comprehensive Organometallic Chemistry; Wilkinson, G., Ed.; Pergamon: Oxford, 1982; Vol. 2, Chapter 17.
24. Mbarak, M. S.; Peters, D. G. JOC 1982, 47, 3397 and references cited therein.

Robert H. Crabtree

Yale University, New Haven, CT, USA

Copyright 1995-2000 by John Wiley & Sons, Ltd. All rights reserved.