2.1 Palladium

The first observation came unexpectedly when Pd NPs were generated from the combination of palladium acetate (Pd(OAc)₂) and sodium borohydride (NaBH₄) in aqueous TPGS-750-M at room temperature (14). The resulting heterogeneous aqueous mixture could be used to great advantage, converting a variety of unsymmetrically disubstituted alkynes to the corresponding Z-alkenes, typically with >99:1 Z:E selectivity (Figure 3). Part of the success observed in these net Lindlar-like reductions is the facility with which the reaction mixture could be recycled, including the water, the surfactant therein, and the Pd catalyst. Thus, an ‘in-flask’ extraction with an ethereal solvent (for example methyl tert-butyl ether (MTBE)) afforded the product, leaving behind all other ingredients ready for introduction of the same or an alternative educt, along with fresh reductant. The overall reliance on organic solvents, therefore, was shown to be at least an order of magnitude lower than amounts often required with similar reactions in the literature (15). Moreover, these differences are to be realised prior to recycling of the aqueous mixtures.

A similar ‘nano-to-nano’ effect was seen with new NPs derived from the reduction of iron(III) chloride (FeCl₃), where either the naturally occurring content of Pd within FeCl₃ or by externally doping with Pd(OAc)₂ at the ppm level sufficed to arrive at active catalysts useful for important cross-couplings (Figure 4) (16). That is, given the threshold presence of ca. 350 ppm Pd, NP formation upon treatment of the mixture (i.e., FeCl₃ + 350 ppm Pd) with methylmagnesium chloride (MeMgCl) in tetrahydrofuran (THF) at ambient temperatures affords the desired NP catalysts. When prepared in the presence of SPhos, the resulting NPs mediate Suzuki-Miyaura couplings in aqueous nanomicelles between room temperature and 45ºC. Extensive analyses of this isolable powder, including a cryo-TEM experiment conducted on the reaction medium (TPGS-750-M + water) again confirmed the aggregation of nanomicelles together with the solid iron nanoparticles containing ppm levels of palladium (here designated Fe/ppm Pd NPs).

This newly developed NP platform, consisting of mostly Mg, Cl and THF, with Fe accounting for only ca. 2.5% of the mix, can be altered as a function of the ligand added prior to their preparation. Changing the recipe from inclusion of SPhos to XPhos now allows for efficient ‘nano-to-nano’ catalysis of Sonogashira couplings in the same recyclable aqueous mixtures (Figure 5) (17). Recycling is smoothly orchestrated using very limited amounts of a single and recyclable, organic solvent for ’in flask’ extraction.

These NPs derived from FeCl₃ can also be prepared in the absence of any ligand, otherwise required for Pd-catalysed cross-couplings. In this case, the ‘nano-to-nano’ effect can be used to effect reductions of aromatic and heteroaromatic nitro groups in water at room temperature (18, 19). The stoichiometric reductant is NaBH₄, although as recently updated at Novartis in Basel, Switzerland, the addition of potassium chloride (KCl) or the use of fresh potassium borohydride (KBH₄) appears to be the hydride source of choice (20). Only 80 ppm of palladium (as Pd(OAc)₂) is required for these NP reductions, again implicating a palladium hydride species (Figure 6). The mild reaction conditions account for the tolerance of many functional groups. Importantly, use of hydrogen gas in place of a borohydride is totally incompatible with this catalyst and in fact, is detrimental to the overall reduction.

2.2 Nickel

Notwithstanding the reported success of the catalyst (Fe/ppm Pd NPs + BH₄^–) applied to nitro group reductions (above) (18, 19), doping of this NP platform with metals other than Pd offers either additional benefits to existing processes, or potentially new opportunities to lower both base and precious metal usage to ppm levels. Part of the incentive to further investigate along these lines is that the amount of residual metal(s) in the desired products has been found to be below tolerance levels as established by the US Food and Drug Administration (FDA). That is, going into these reactions with ppm amounts of transition metal catalysts, rather than the more typical 1–5 mol% range (10,000–50,000 ppm), leads to acceptable ppm levels of metal impurities that obviate additional time and cost for their removal, a very common occurrence under traditional conditions in organic solvents. In work soon to appear (21), doping these Fe-based NPs with additional ppm levels of a Ni(II) salt affords a reagent that has been found, likewise, to efficiently reduce nitro group-containing aromatic or heteroaromatic compounds, but often at a far greater rate under otherwise identical conditions of concentration, time and temperature. Figure 7 shows a few comparison cases.

2.3 Copper

Another metal used extensively in organic chemistry is copper and hence, doping the iron nanoparticles with ppm levels with a Cu salt, rather than Pd or Ni (Fe/ppm Cu NPs) might form a potentially useful reagent, presumably benefiting from the same ‘nano-to-nano’ effect. Indeed, related NPs are formed using Cu(I) admixed with the same FeCl₃, followed by standard treatment with MeMgCl in THF (vide supra). Akin to observations with other NPs in this series, the catalyst can be generated and used in situ or isolated for use at a later date. The first type of catalysis examined has been click chemistry between a terminal alkyne and an azide; when performed in aqueous TPGS-750-M at room temperature, cycloadditions to the anticipated triazoles take place quite readily (22). In addition to representative examples shown in Figure 8, products formed upon recycling of the aqueous reaction mixture are suggestive that the catalyst does not lose its activity when handled under an inert atmosphere to prevent autoxidation to the otherwise inactive Cu(II) form.