“ Copper-in-Charcoal ” Revisited: Delineating the Nature of the Copper Species and Its Role in Catalysis

: “ Copper-in-charcoal ” has been shown to be a versatile catalytic source of supported copper for a variety of important synthetic transformations, as well as in other ﬁ elds such as energy. We herein report the characterization of this material and the implications that its preparation has on catalysis, thus providing a greater understanding of the scope and limitations of this catalyst system.

O ver the past few decades, a range of catalysts have been heterogenized onto solid supports (e.g., SiO 2 , Al 2 O 3 , mol sieves, etc.), and activated carbon is one of the most commonly employed. 1 These heterogeneous catalysts have received much attention due to their notable increase in reactivity and recyclability when compared to traditional transition-metalbased catalysts. 2 However, research in this area has been largely focused on applying the heterogeneous catalyst systems to a variety of applications rather than characterizing and improving them, probably due in part to the expertise in the research groups reporting the work lying toward organic synthesis rather than materials characterization. 3 The Lipshutz group, has pioneered the use of metal-based catalysts on carbon supports, for example, Ni/C, 4 Ni/Cg, 5 Cu/ C, 6 and mixed metal systems, 8 for a variety of reactions. In particular, they have used copper nanoparticles supported within the pores of activated carbon to act as a source of copper(I) or (II) for a range of heterogeneously catalyzed reactions. Oxidized copper nanoparticles on activated carbon have also been used in CuAAC reactions. 9 Several different protocols have been reported for the preparation of Cu/C catalysts. For example, Mehandjiev reported a procedure in which the final step involved the copper-in-charcoal material being calcined at 503 K for 2 h, 10a a similar procedure was reported by Liu with heating at 523 K 10b and Wey reported the heating of the copper-in-charcoal material at >700 K before the material was analyzed: 10c all three groups report X-ray powder diffraction data showing the presence of CuO with lesser amounts of Cu 2 O in the charcoal matrix. The material prepared and used by the Lipshutz group involved the combination of activated carbon (Darco-KB) and an aqueous solution of copper(II) nitrate using sonification, distillation of water followed by azeotropic drying with toluene to afford the catalyst (Scheme 1). 6a,b Transmission electron microscopy (TEM) was also used to record 2D images of the carbon material, from which the group proposed a relatively even distribution of the copper nanoparticles within the charcoal matrix; 6b lastly, quantitative inductively coupled plasma atomic emission spectrometry (ICP-AES) was used to determine the loading of copper within the charcoal, which was found to be 0.344 mmol Cu/g catalyst, as well as the extent of copper bleed into solution during the reactions. 6b Importantly, we noted that leaching of copper(II) from the carbon matrix was reported during azide formation, 6d and under flow conditions. 7 Of particular note, Lipshutz was able to show that high yields in a series of copper assisted Huisgen [3 + 2] cycloaddition reactions could be obtained (Scheme 2). They also showed that the copper-in-charcoal material was particularly robust and were able to use a wide range of solvents using high temperature microwave conditions, with no solubility issues. The catalysts could also be reused (for at least three catalytic cycles) without loss of activity. 6a Our multidisciplinary group is interested in the use of heterogeneous catalysts for a variety of applications, and we have previously investigated the role of copper(II) compounds as precatalysts in copper alkyne azide cycloaddition (CuAAC) reactions. 11 In those reactions, the copper(II) species act as a catalyst in Glaser coupling reactions and is thereby reduced to copper(I) with concomitant formation of copper(I) acetylides which are subsequently involved in catalysis of the CuAAC reaction. 11a,c,12 A dicopper(II)-substituted γ-keggin silcotungstate has been used in Glaser coupling reactions 13 and is also reduced to act as a precatalyst in CuAAC reactions. 14 Copper(I) acetylides can also be useful for the A 3 multicomponent synthesis of a range of tertiary amines, 15,16 and they are also involved in the ligand accelerated Stephens−Castro coupling of terminal alkynes and aryl iodides 17a and Sonagashira−Hagihara reactions. 17b We decided to use the versatile supported copper system and initiated the synthesis of the copper-in-charcoal material following the Lipshutz protocol, outlined in Scheme 1. The preparation went as expected, and we obtained a black powder, which we analyzed using X-ray powder diffraction, in order to confirm its identity before use ( Figure 1). Previous reports 6a,d assumed, as a result of earlier work, 10 that the Cu/C catalyst was a mixture of CuO/Cu 2 O bound to the surface of the carbon support: instead we found that the mineral Gerhardite (Cu 2 (OH) 3 NO 3 ) had been formed as a distinct material, not bound to the support. We then became aware that the copperin-charcoal material was commercially available, 18 and we decided to analyze that material in order to compare it with our initial material. The two materials were found to match each other; both contained a quantity of copper(II) hydroxynitrate, with a small amount of an unidentified impurity: the impurity was not present when we prepared Gerhardite on Norit A SUPRA as the carbon support, or by omitting the carbon from Lipshutz's procedure for "Cu/C" formation (Scheme 3) In order to prove that this impurity did not arise from the copper nitrate starting material, a series of reactions were carried out to prepare the copper-in-charcoal (Darco-KB) with a higher loading of copper, where the copper nitrate loading was increased to 2×, 5×, and 10× from the original reaction. The X-ray powder diffraction patterns of the solid support materials prepared in this way are shown in Figure 2.
As expected, the different copper loadings did not show much variation on the quantity of the impurity that formed. However, the reaction using the 2× loading was carried out at a lower final temperature and did not result in the formation of Gerhardite: the impurity phase was the only copper material present. The lack of Gerhardite proved to be an advantage, giving a clearer XRD pattern of the impurity, which allowed a better library search of that material. A perfect match was made with the mineral Libethenite (Cu 2 (PO 4 )OH), a phosphoruscontaining species, which explains why the copper loading did not affect the quantity of the impurity formed. A sample of Libethenite was then prepared independently from the support material for XRD analysis 19 and gave an even clearer match to the impurities present in the Darco-KB copper-in-charcoal support materials.
We believe Libethenite is formed due to the nature of the carbon support; that is, Darco-KB is formed through the chemical activation of wood using the phosphoric acid process. 20 The residual phosphate on the carbon could interact with the copper nitrate during the reaction to prepare the copper-in-charcoal support material. Libethenite is not observed using Norit A SUPRA as this carbon material is a steam-activated carbon of natural origin. 21 In order to delineate the role of Libethenite and Gerhardite in the catalytic processes reported by Lipshutz for "Cu/C", we tested our copper containing materials in the CuAAC reaction ( Table 1). Treatment of phenyl acetylene and benzyl azide under microwave conditions employing a 10 mol % catalyst loading generally afforded good to excellent yields of a single regioisomer of the triazole. 22 Gerhardite, without the charcoal support (entry 5), performed equally as well as "Cu/C" and was reduced to copper(I) to produced the required polymeric copper(I) acetylide precatalyst, which was easily removed at the end of the reaction by filtration. 23 Under the conditions reported in Table 1, Et 3 N was not required for the reactions to proceed, as noted by Lipshutz for "Cu/C". 6a,24 Importantly,  Libethenite was found to be inactive in the CuAAC reaction (entry 6). For reasons which are unclear at present, the commercial "Cu/C" performed poorly in this reaction, perhaps due to degradation of the catalyst, although XRD analysis did show good correlation between commercial and our prepared material. 25 Having noted previously that materials such as Cu 2 (OH) 3 OAc react with alkynes to initially form copper(I) acetylides, 11d we were interested in examining the supported materials after the CuAAC reaction. While there was some conversion shown from Gerhardite to copper(I) phenylacetylide during the reaction, it was not a full conversion so a series of time controlled reactions were performed to try to drive the conversion of Gerhardite to copper(I) phenylacetylide. 26 Using the conditions shown in Table 1 the triazole click reaction was catalyzed using a mixture of Gerhardite and Libethenite without the solid support material present. We believed that this would help to remove some of the background noise from the XRD patterns making interpretation easier. The results are shown in Figure 3. The XRD patterns from 0 h to 10 days show the gradual formation of the copper(I) phenylacetylide and the corresponding gradual disappearance of Gerhardite as it is consumed. The Libethenite remains present in all XRD patterns, which shows that it takes no part in the formation of the copper(I) species and hence cannot take part in catalysis of triazole formation.
These observations led us to question why we did not observe CuO or Cu 2 O, as had been previously assumed 6 and reported independently by Mehandjiev, Liu, and Wey. 9 The detailed procedures used by the research groups reveal a crucial difference: in the Mehandjiev, Liu, and Wey reports, the final step in the experimental procedure involved the copper-incharcoal material being calcined at either 503 K, 10a 523 K, 10b or >700 K 10c before the material was analyzed by X-ray powder diffraction. Under the Lipshutz conditions, that final step was not carried out, and as a result, the inorganic salts were not decomposed but remained in the form of Gerhardite and Libethenite.
In summary, we have shown that the procedure 6 used to prepare copper-in-charcoal does not afford CuO or Cu 2 O bound to the charcoal matrix as previously assumed. However, Gerhardite is easily prepared and can be stored for long periods without decomposition, as also can, for example, the copper(I) acetylide polymer precatalysts. Hence, the copper species using Cu/C, that is responsible for reactions such as the CuAAC process, has been identified as Gerhardite (Cu 2 (OH) 3 NO 3 ), which acts as a precatalyst and is reduced under the reaction conditions to a copper(I) acetylide in order to carry out the triazole forming step. Additionally, we observed an impurity phase in the materials prepared using the Darco-KB activated carbon. This impurity was identified as Libethenite, Cu 2 (PO 4 )-OH, and was shown to be inactive in the CuAAC reaction. We are currently investigating the application of Gerhardite to other catalyed processes and early indications show that both Gerhardite and Libethenite can catalyze Ullmann aryl ether synthesis. A full investigation will be published in due course. We anticipate that understanding of the catalyst involved in this heterogeneous system will inspire further studies and advancements in this area.

Notes
The authors declare no competing financial interest.

■ ACKNOWLEDGMENTS
This work was supported by Loughborough University (PhD studentship to E.C.S.). B.R.B. thanks the Research Councils U.K. for a RCUK Fellowship and the EPSRC for an undergraduate summer research bursary to RB.