Sustainable routes to amines in recyclable water using ppm Pd catalysis

Reactions are discussed that lead to the valuable amine functional group in aromatics and heteroaromatics, carried out under sustainable conditions in water enabled by micellar catalysis. These new technologies are general and have been successfully applied to intermediates associated with several targets in the pharmaceutical space. Each relies on ppm levels of Pd used either as solubilized catalysts or on the surface of easily formed nanoparticles. These processes also feature opportunities for recycling of the aqueous medium, thereby overall, offering greener, environmentally responsible methods for introducing nitrogen-containing functionality into organic molecules.

Introduction

Good luck finding an FDA-approved drug that doesn’t contain nitrogen! And yet, as important as this element is to the pharmaceutical world, the same could be said about finding a method for making amines that, by any yardstick, would be considered “green.” Thus, while there is no shortage of inroads to amines, with many being sufficiently time-honored that they can today be found in most textbooks, the reality is also that most, if not all, are carried out under conditions that can be fairly described today as “unsustainable.” Thus, virtually all amine-forming reactions continue to be run in waste-generating organic solvents and oftentimes require both heat and reliance on transition metal catalysis which, if not exclusively Fe-based, translates into use of an endangered metal, precious, or otherwise. Our dependence today on petroleum cannot last, nor can the rate at which much of the world’s supply of many elements in the periodic table is being consumed [1, 2, 3, 4], especially without a major commitment to recycling. But, as it turns out, technology for introducing nitrogen into organic molecules using known reactions that is not only very efficient but also respectful of our limited precious resources is already here, and available. In this brief account, some recent developments are discussed, each of which is conducted in recyclable water as the gross reaction medium, rather than in traditional organic solvents. Moreover, to make matters better, an order of magnitude less catalyst, focusing on palladium in particular, is all that is required by each of these new processes.

Most practitioners place C–N coupling reactions under the blanket of “Buchwald-Hartwig aminations” (B–H), Pd-catalyzed processes introduced a quarter of a century ago [5, 6, 7, 8, 9]. From the standpoint of the medicinal chemist, these continue to play an increasingly important role, although the corresponding process chemist especially in the pharma world may not be as enthusiastic. The reason is clear, as a recent review highlighting the 25th year of these important contributions clearly illustrates [10, 11, 12, 13]: the amount of Pd catalyst characteristic of these couplings remains in the 2–10 mol % range (i.e., 20,000–100,000 ppm), which by today’s standards, is not only far too costly at scale, but most assuredly not sustainable and highly waste generating given the conditions typically involved. Very recent improvements in catalyst loadings of palladium look encouraging but still come up short from the green chemistry perspective, as these are mainly performed in nonrecycled THF, rather than the rate-enhancing use of EtOAc [14], and oftentimes with heating to ≥90 °C [15]. On the other hand, Handa et al have recently described considerable progress in terms of both the reagent and reaction medium, having developed new mixed Cu and Pd-containing nanoparticles that effect the same aminations, but in water [16].

And as for the loadings of Pd required, this issue has been addressed; today, in hand, is technology for effecting such couplings in water under mild conditions (rt up to 45 °C), using Pd catalyst loadings that hover around 1000 ppm (i.e., 0.10 mol %) [17], which translates into 20 to 100 times less Pd than previously described [10], before recycling. If this “sounds too good to be true”, well, sometimes the expression is not truly reflective of reality, and this is one of those times. Thus, as described in 2019, a readily available π-allylpalladium catalyst LPd (Figure 1), used in an aqueous nano-micellar environment based on the commercially available surfactant TPGS-750-M [18], effects amination reactions involving a variety of reaction partners in water at 1000 ppm Pd. Most effective are aromatic amines, although secondary cyclic systems such as morpholine and pyrrolidine appear to be amenable. Representative examples are also shown in Figure 1.

One especially noteworthy feature focuses on the low ppm levels of residual Pd found in the products (≤4 ppm in all cases examined). This eliminates the need for further processing given the FDA-imposed limit of 10 ppm Pd/dose/day, which to otherwise achieve would require additional time and cost for processing to remove the residual Pd typically found in excess of this limit. Recycling of the reaction mixture, also illustrated in Figure 1, which saves solvent costs, can be performed four times beyond the initial coupling. Although 500 ppm of catalyst was needed to reach full conversion with each recycle, the total for these five reactions was 3000 ppm; i.e., 1000 + (4 × 500 ppm Pd), which is tantamount to 600 ppm (0.06 mol %) Pd per reaction. Comparison reactions using this level of Pd loading (0.10 mol %; same catalyst) in either pure toluene or 1,4-dioxane did not lead to synthetically useful results.

Traditionally, a reductive amination approach to C–N bond formation using a Pd catalyst, as with aminations via C–N bond formation, requires too much Pd, oftentimes in the form of Pd/C [19, 20, 21, 22]. Moreover, these reactions are normally performed, not surprisingly, in organic solvents; hence, there is much room for improving these unsustainable approaches. To remove waste-generating organic solvents from this important transformation, the “switch” to water [23] containing the same TPGS-750-M-derived nanoreactors leads to an effective process, one that requires only 0.20 mol % Pd/C (i.e., 2000 ppm Pd), and uses the least expensive 1 wt % Pd/C in the presence of inexpensive Et₃SiH (1.2 equiv) as the stoichiometric source of hydrogen on the metal (Figure 2) [24]. Other sources of silane (e.g., PhSiH₃, PMHS, DEMS, etc.) were all far less effective (≤40%). Several types of amines readily participated, including 1° and 2° alkyl, 1° and 2° aryl, and (hetero)cyclic cases are illustrative, as summarized in Figure 2.

Several other base metal (Mn [25], Ni [26], Cu [27]) and noble metal (Ru [28] and Rh [29]) catalysts are used for reductive amination. Comparisons with alternative technologies based on both base metal catalyst (e.g., Fe [30] and Co [31]), or precious metals (e.g., Ir [32] and Au [33]), each used in organic solvents at elevated temperatures (85–120 °C) showcased the far “greener” nature of this chemistry in water. Importantly, other reducible functional groups, such as halides (Br and Cl), as well as nitro groups remain intact. Recycling (twice) was facile, and the associated E Factor (i.e., waste in grams/product isolated in grams) [34] based on solvent usage was a very low (2.6). Opportunities for introducing deuterium by simply using the combination of Et₃SiD/D₂O afforded the desired amine with 100% incorporation of D at the α-site.

Thus, reductive aminations in water using ppm levels of Pd in Pd/C + Et₃SiH are safe, and green.

Use of nitro groups as precursors to the corresponding primary amines offers a particularly valuable strategy in synthesis. Because the NO₂ residue is highly polar and thereby, electron-withdrawing, it is typically considered activating toward Pd-catalyzed couplings, facilitating an initial oxidative insertion by Pd(0) into a carbon-halogen bond. Alternatively, it is a common activating residue for C–N bond formation via S_NAr reactions (that can also now be done in water) [35] using amines as nucleophiles. For these and several additional reasons, the nitro group figures prominently in syntheses of amines. Unfortunately, alternative processes not only tend to be used in organic solvents but oftentimes rely on unsustainable levels of especially precious metals, and high temperatures.

Given their importance en route to amines, together with the accent on utilizing ‘greener’ options, the implication is that not only is attention being paid to minimization of organic solvents and amounts of endangered metals being used, but also the safety aspect: green chemistry is safe chemistry! [36] Many procedures, including several of recent vintage [37, 38, 39, 40, 41, 42, 43, 44, 45, 46] are inherently dangerous, oftentimes involving elevated temperatures and, in using hydrogen gas, being run under pressure. This field already has several very effective processes for achieving the same desired outcomes, but without the environmental offenses and in the absence of concern regarding flammability and/or explosiveness. These include the following:

Starting with just inexpensive, commercial FeCl₃, to which is added 80 ppm Pd(OAc)₂ (0.008 mol %!) in THF at room temperature, introduction of commercial MeMgCl (in THF) leads to new nanoparticles (NPs) that can be used directly or isolated and stored in a bottle (Figure 3a, left; NPs A) [47]. These are sold by Aldrich (catalog #809330) [48] and do not require that any ligand on Pd be added. This powdery material, once added to aqueous TPGS-750-M (2 wt %), is easily dispersed. In the presence of stoichiometric NaBH₄ at room temperature, nitro groups are smoothly reduced in high yields. From the dozens of examples studied, excellent functional group compatibility is to be expected. Some examples are shown in Figure 3a, highlighting this technology. Noteworthy is the opportunity for further manipulation of the newly created amine as part of a multistep sequence within the same reaction pot. Hence, as illustrated in Figure 3b, initial nitro group reduction of 4-bromonitrobenzene, followed by amide bond formation, and finally, Suzuki–Miyaura coupling with a pyridyl 2-MIDA boronate derivative affords final product in 70% overall isolated yield.

“Can you make these nitro group reductions go faster?” Such was the request from a process chemist in big pharma while discussing applications of the newly disclosed Fe/ppm Pd NPs (see Figure 3a, NPs A). Ultimately, it was found that by “diluting” the surface of these NPs that contain low levels of Pd with an “excess” of Ni, thereby preventing formation of less reactive Pd–Pd aggregates, increased reactivity was realized [49]. These modified NPs could be prepared by simply including Ni(NO₃)₂•6H₂O (0.16 mol %) in the typical recipe [47] of FeCl₃, Pd(OAc)₂, and MeMgCl/THF, leading in situ to active NPs similar in appearance (Figure 3a, left; NPs B). On addition of NaBH₄, direct comparisons between both NPs in nitro group reductions showed that while similar yields were obtained, the presence of Ni led to reactions that were ca. four times faster; one example leading to product C illustrating the differences in reaction rates is shown in Figure 4c. Otherwise, functional group compatibility seems unchanged, as suggested by reductions affording the products shown. Control experiments leading to product D established that Ni alone (i.e., in the absence of Pd) led to far slower rates of conversion.

Historically, Fe/HCl was among the first methods used for reductions of aromatic nitro groups (i.e., Bechamp reduction) [50]. The strongly acidic conditions and typically high reaction temperatures are obvious limitations in the presence of sensitive functionality within the same molecule, thus leading to development of a vast array of alternative procedures. Nonetheless, the fundamental properties of iron together with its many “green” characteristics (e.g., natural abundance, minimal cost, low toxicity, etc.) have maintained iron-based catalysts as popular reagents. One that has been found to offer considerable potential derives from commercial processing of iron carbonyl compounds (e.g., Fe(CO)₅) that are converted thermally into highly pure iron powder, CIP [51]. This inexpensive material has been found to serve admirably as a (super)stoichiometric reductant of nitro groups in water, only requiring the presence of NH₄Cl to adjust the pH of the medium [52]. Although good functional group compatibility was observed, it is also noteworthy that overreduction of pyridyl rings, so prevalent in pharmaceuticals, leading to desired products was not observed (Figure 4d). Other examples are also shown, including polyhalogenated products, formed quantitatively. As usual, recycling of the aqueous surfactant mixture presented no issues, although best results are to be realized when an additional 2.5 equivalents of CIP are added (three recycles were reported).

Hence, nitro group reductions can today be done safely in water using various technologies, each of which is environmentally responsible.