Orthogonal Nanoparticle Catalysis with Organogermanes

Abstract Although nanoparticles are widely used as catalysts, little is known about their potential ability to trigger privileged transformations as compared to homogeneous molecular or bulk heterogeneous catalysts. We herein demonstrate (and rationalize) that nanoparticles display orthogonal reactivity to molecular catalysts in the cross‐coupling of aryl halides with aryl germanes. While the aryl germanes are unreactive in LnPd0/LnPdII catalysis and allow selective functionalization of established coupling partners in their presence, they display superior reactivity under Pd nanoparticle conditions, outcompeting established coupling partners (such as ArBPin and ArBMIDA) and allowing air‐tolerant, base‐free, and orthogonal access to valuable and challenging biaryl motifs. As opposed to the notoriously unstable polyfluoroaryl‐ and 2‐pyridylboronic acids, the corresponding germanes are highly stable and readily coupled. Our mechanistic and computational studies provide unambiguous support of nanoparticle catalysis and suggest that owing to the electron richness of aryl germanes, they preferentially react by electrophilic aromatic substitution, and in turn are preferentially activated by the more electrophilic nanoparticles.


Introduction
Over the past decade the nanotechnology industry has surged forward to reach ag lobal market of greater than one trillion US dollars in 2018, [1] with diverse applications ranging from materials for solar cells,p hotonics,c osmetics or biomedical applications,s uch as drug delivery,t issue engineering,a nd cancer therapy,t oc atalysis. [2][3][4] While nanoparticle catalysts are generally more reactive than their bulk metal counterparts because of their greater surface area, they frequently need more forcing reaction conditions than homogeneous molecular metal catalysts to trigger the same transformations. [5] However, leaching from molecular catalysts can cause the release of nanoparticles,and consequently their involvement as potentially "true" catalytic species in established catalytic transformations has also been subject to intense debates. [5][6][7] As opposed to homogeneous molecular catalysis,l ower loadings in metal are frequently required under nanoparticle catalysis conditions. [2,5,8] This characteristic, paired with the low cost of preparation and the absence of sensitive ligands in nanoparticles,h as led to immense interest academically as well as their implementation in largescale industrial processes. [9] However,d espite the many publications on nanoparticle catalysis,t od ate,t here is no precedence of unambiguously unique and orthogonal reactivity of nanoparticles compared to homogeneous molecular or heterogeneous bulk catalysts in organic transformations.S uch insights would be of utmost importance as there is ah igh demand for innovative and orthogonal synthetic strategies.E specially,amodular and straightforward access to richly functionalized biaryl motifs is in considerable demand, owing to their widespread abundance in drugs,m aterials,o rp rivileged catalysts. [10] In this context, as trategy that operates in an orthogonal fashion to the widely employed Pd-catalyzed cross-coupling technology [11] would offer an additional dimension for structural diversification as well as potential to overcome existing synthetic challenges through orthogonal synthetic approaches.
Since the advent of metal-catalyzed cross-coupling technology more than 40 years ago,the field has grown to be everincreasingly enabling owing to numerous efforts to push the frontiers of catalyst development and mechanistic understanding;y et the transmetalation step to date largely still relies on the original set of reagents. [11] TheS uzuki-Miyaura cross-coupling of organoboron reagents with aryl halides is most widely and ubiquitously used among organic,medicinal, and materials chemists in academia and industry, [12] as the established alternatives can be associated with basicity, instability (organomagnesium and -zinc reagents), toxicity (organotin), or lower reactivity.Despite its relative mildness, broad scope,and high reactivity,this popular coupling class is not free of challenges,h owever.T hese include,f or example, the occasional instability of boronic acids,w hich is particularly pronounced in the case of 2-pyridyl-and multifluoroarylboronic acids and further aggravated by the presence of (and need for) base. [13] Ingenious masking strategies [14,15] or elegantly more reactive systems that make use of aryl diazonium salts as acceptors [16] have been developed to balance the relative kinetics of deactivation versus productive cross-coupling in these cases. [17] Some toxicity concerns in conjunction with organoboron compounds and their deriva-tives have recently also been reported, [18] which may create an eed for alternative approaches in certain applications ( Figure 1).

Results and Discussion
We aspired to widen the conceptual coupling repertoire and focused on organogermanium compounds.P romisingly, no toxicity has been associated with this compound class, [19] and our stability tests of ap entafluoroaryl germane (Ar-GeEt 3 ,A r = C 6 F 5 )i ndicated that as opposed to the corresponding boronic acid, which has al ifetime of milliseconds, [13,16] ArGeEt 3 remains completely stable even upon subjection to acid (HCl) or base (NaOH, KF) for 2h at 90 8 8C ( Figure 2). [20] Similarly the 2-pyridyl derivative proved to be fairly stable under basic conditions but was sensitive to acid.

Mechanistic Tests for Potential Reactivity with Pd II and Mechanistic Support of Nanoparticle Reactivity
Thef ew reported Pd-catalyzed cross-coupling reactions involving organogermanium compounds ascribed relatively low reactivity to the latter as compared to the established coupling partners,a nd coupling attempts exclusively applied basic and relatively harsh conditions without any detailed mechanistic interrogation. [21][22][23][24] We envisioned that adetailed investigation of the fundamental aspects of the coupling process involving organogermanium compounds might likely offer inspiration. In this context, we initially probed the potential of defined homogeneous Pd 0 /Pd II coupling cycles and synthesized av ariety of Pd II complexes of the nature L n Pd II (X)(Ar), in which "X" was ahalide or hydroxide.Upon subjection of phenyltriethylgermanium to the well-established mono-, bis-, and bidentate phosphine-coordinatedPd II complexes 2-7 at room temperature or 80 8 8C, we saw no indication of transmetalation taking place,r egardless of the coordinated halide (I, Br, F) or hydroxide,t he employed solvent (THF,DMF,toluene), or additive (TBAF,CsF,KOH, or K 2 CO 3 )( Figure 3B). In particular,P d II ÀFc omplexes are usually privileged intermediates that typically undergo direct transmetalation with the established cross-coupling partners ArSiR 3 ,A rSnR 3 ,o rA rB(OH) 2 without the need for additives. [12,25] Indeed, while the organogermane remained untouched ( Figure 3), our comparative studies showed that organoboron reagents react with the same Pd II ÀFc omplex within seconds,a nd organosilane and organotin reagents within an hour at room temperature (see the Supporting Information, Table S1). These data clearly reinforce that typical Pd 0 /Pd II reactivity modes are not readily amenable to organogermanium species.
Interestingly,although Pd II -iodide complex 7 did not give rise to any transformation of the organogermane,u pon addition of one equivalent of AgBF 4 at room temperature, the cross-coupled product 8 was generated in 18 %y ield ( Figure 3C). We next tested whether Pd II complex 7 in conjunction with AgBF 4 could also trigger the catalytic conversion of organogermanes.U sing 2.5 mol %l oading of 7 with AgBF 4 indeed gave efficient catalytic transformation of 1-iodo-4-(trifluoromethyl)benzene with 4-fluorophenyl triethylgermane 1 to yield biaryl 8.Given the rather labile nature of the Pd II complex 7 as well as the visible metal precipitation upon AgBF 4 addition, we speculated that palladium nanoparticles might be generated under these conditions and hence might also be involved in the coupling process.W e therefore next adopted conditions that are known to generate nanoparticles and used Pd 2 dba 3 along with AgBF 4 (Figure 3C). [26] This resulted in efficient coupling of 1-iodo-4-(trifluoromethyl)benzene with organogermane 1 under these conditions,and we isolated biaryl 8 in 89 %yield after 16 hat 80 8 8C.
Further support of nanoparticle catalysis was gained through the following experiments and observations:1 )Our  analysis of the mixture through TEM imaging revealed the presence of spherical particles of approximately 5nmd iameter. EDX composition analysis showed the presence of both Pd and Ag in these nanoparticles.
2)The addition of mercury to this successful catalytic reaction resulted in complete inhibition;there was no significant product formation ( 5%; Figure 3C). These results are in accord with trapping and deactivation of the active nanoparticles. [27] Moreover,3)when we monitored the formation of 4-methyl-4'-(trifluoromethyl)-1,1'-biphenyl as well as the consumption of the starting materials over time,weobserved abrief induction period (of 90 min). This induction period was found to significantly prolong to 27 hi nt he presence of added ligand (5 mol % PPh 3 ). These observations are common indicators that the true active species is phosphine-free and is formed during the initial induction phase.I nt his context, our further experimentation revealed that the induction period and hence the formation of active species is independent of the aryl germane (see the Supporting Information for additional details).

Exploration of Synthetic Potential
In light of these results,that is,the lack of reactivity of the organogermane with homogeneous Pd II complexes,b ut high reactivity under [Pd] nanoparticle catalysis,w ea nticipated that there could be significant potential towards maximizing diversity in cross-coupling and hence set out to explore the potential of catalytic cross-coupling with organogermanes in greater detail.
Pleasingly,when we applied these nanoparticle conditions to both electron-rich and electron-poor aryl iodides with avariety of aryl germanes and stirred the mixture overnight, we obtained excellent yields of the corresponding biaryl products (8-18;F igure 4). Alkyl, ester,m ethoxy,o rf luorinated groups were well tolerated, and all electronic combinations of biaryl (i.e., electron-rich/-rich, electron-poor/electron-rich, or electron-poor/-poor) could be prepared. Notably, the coupling was not affected by oxygen or moisture as the same coupling results were obtained under inert conditions or when the reaction was run in an open flask (see 36, 37 in Figure 4).

Exploration of C À Iv ersus C À Br/C À Cl Chemoselectivity
Another pertinent challenge in the cross-coupling arena is site-selective bond formation. [28,29] Chemoselective coupling strategies are of widespread interest as they provide access to densely functionalized biaryl motifs and enable the rapid creation of diversely substituted compound libraries.F or poly(pseudo)halogenated arenes,t ypical L n Pd 0 /L n Pd II -based coupling protocols generally suffer from low predictability of the favored coupling site and ap ronounced substrate specificity.B yu tilizing Pd I dimers or cationic Pd trimers, predictable site-selective functionalizations were recently achieved with basic Grignard or organozinc reagents. [28,29] Pleasingly,o ur mild and base-free conditions involving organogermanes allowed for C À Is elective functionalization with multiply halogenated substrates,b earing bromide and chloride as well as the pseudo-halogen OTff unctionalities (29-32;F igure 4).

Exploration of Practicability
With the exquisite synthetic potential of nanoparticlecatalyzed couplings of aryl germanes showcased, we next assessed practical and sustainability features of the reaction for its wider applicability.W hile 2.5 mol %o ft he Pd source was employed in the above experiments,o ur tests indicated that the transformation also proceeds at asignificantly lower loading of [Pd] and is scalable:Using 0.1 mol %ofPd 2 (dba) 3 , biaryl product 9 was efficiently prepared on as cale of about 1gand 96 %yield. Aside from catalyst loading and scalability, the reaction medium and time will also influence wider applications,e specially in an industrial context. To this end, ac loser examination of the required reaction time revealed that much shorter times are sufficient and alternative solvents can be utilized. Only asmall amount of DMF was found to be necessary for the formation of the active nanoparticle.A s such, pre-stirring of catalytic amounts of Pd 2 dba 3 with (potentially sacrificial) iodobenzene (2.5 and 5mol %, respectively), AgBF 4 (1.5 equiv) with little DMF (2.5 equiv) for 40 min at 80 8 8Cw as found to be sufficient and then allowed the rapid coupling of an aryl germane with an aryl iodide within 1hin dioxane in good yields (43-48;F igure 5).

Tests of Silver-Free Reactivity and Hypervalent Iodine Reagents
Our mechanistic data indicated that the primary role of silver was as an iodide scavenger. As such, we envisioned that hypervalent iodine compounds might also be effective in the coupling of organogermanium compounds as they are inherently more activated. Indeed, we found that diaryliodonium salts function as complementary electrophiles,e ven for the exceptionally challenging 2-pyridyl substrates (38-42;F igure 5). Both the BF 4 and PF 6 iodonium salts were shown to be effective.N otably,t he coupling of organogermanes with diaryliodonium salts is also effective in the absence of AgBF 4 . Fore xample,w hen a4 -methoxyphenylgermane was reacted with ad iphenyliodonium salt in the presence of Pd 2 (dba) 3 (2.5 mol %) in DMF for 2hat 80 8 8C, 51 %ofbiaryl product 38 was isolated. In this context, we unambiguously confirmed the formation of nanoparticles under silver-free reaction conditions with diaryliodonium salts (see the Supporting Information for further information and TEM images). [30] Theability to conduct these reactions in asilver-free fashion suggests that [Pd] instead of [Ag] is the key active component in the nanoparticles that allow for couplings of aryl germanes.
Are Aryl Germanes Truly Privileged with Nanoparticles?
With the synthetic potential of the nanoparticle-catalyzed coupling of aryl germanes with aryl iodides or hypervalent iodine reagents established, we next assessed whether the aryl germanes are truly privileged in these transformations.T othis end, we subjected established transmetalation agents,t hat is, para-fluorophenylboronic acid and the pinacol boronic ester thereof to the nanoparticle-catalysis conditions and attempted to couple 1-iodo-4-(trifluoromethyl)benzene ( Figure 6A, right). While the corresponding ArGeEt 3 reagent delivered the coupling product 8 in 57 %y ield after 30 min, the other transmetalating agents failed to deliver the coupling product in appreciable amounts and gave 8 in only 2-8 %yield.
As such, there is ap rofound selectivity reversal from traditional molecular Pd 0 /Pd II versus [Pd] nanoparticle conditions.While the aryl germane proved to be the least reactive (= unreactive) under classical molecular L n Pd 0 /L n Pd II crosscoupling conditions as compared to the established crosscoupling partner ( Figure 6A), it becomes the most reactive under nanoparticle conditions.C onsequently,s elective couplings should also have potential. Indeed, the intramolecular competition of silane-and Bpin-substituted aryl germanes in the coupling with 4-iodophenylboronic acid pinacol ester gave exclusive coupling at the CÀGe site (33-35,s ee Figure 4). Moreover,the intermolecular competition between ArGeEt 3 versus ArB(Pin) or ArB(MIDA) in the coupling with iodobenzene also showed orthogonal selectivities (Figure 6B), offering therefore an orthogonal tool for selective C(sp 2 ) À C(sp 2 )c oupling reactions and an additional mode to increase diversity.

Computational Studies on the Origins of Reactivity
To gain insight into the origins of orthogonality,w e undertook computational studies at the CPCM (DMF) B3LYP-D3/Def2TZVPP//B3LYP-D3/Def2SVP level of theory. [31] We initially investigated why organogermanes are not reactive in the transmetalation of defined L n Pd II complexes. Interestingly,wefound that the generally assumed concerted four-centered transmetalation of PhGeMe 3 with [(PPh 3 )Pd II -(X)(Ph)] with X = Fo rIis significantly disfavored (DG°> 40 kcal mol À1 ;F igure 7A), which appears to be due to al ack of driving force to form a[Ge]-halogen bond. Our search for alternative modes of activation revealed that electrophilic aromatic substitution (S E Ar) constitutes alower-energy pathway for transmetalation at Pd II (see TS2, Figure 7A), which is characterized by an activation free energy of DG°= 35.8 kcal mol À1 for X = I. While this barrier is still rather high, these results indicate that organogermanes appear to be more prone to react as nucleophiles and hence should prefer more electrophilic and electron-deficient metal species rather than ligand-coordinated Pd II complexes.
We next set out to study the molecular events under nanoparticle catalysis.B uilding on previous studies on the likely speciation of nanoparticles, [32] we used aphosphine-free Pd trimer as ar epresentative model for the active nanoparticle.W es tudied the likely full catalytic cycle and investigated numerous possibilities,o fw hich the favored pathway is featured in Figure 7B.
In the experimentally observed initiation phase involving Pd, silver,a nd the aryl iodide,apalladium cluster is likely formed and stabilized through oxidative saturation of aryl iodide.O ur computational data suggest that addition of two molecules of PhI to the Pd cluster is highly exergonic and favored over coordination of an aryl germane.T he saturated Pd 3 intermediate is likely activated by Ag + ,forming acationic cluster (Int1)and releasing AgI. Alternatively, Int1 is formed directly with diaryliodonium salts (in the absence of silver salts). Subsequent coordination of aryl germane to Int1 is now energetically favored over ArI. Thek ey C À Ge bond activation then takes place from Int2,and was found to proceed by an S E Ar-type mechanism via TS3.T he activation free energy barrier for the CÀGe bond cleavage is 24.6 kcal mol À1 and as such significantly lower than that of S E Ar-type transmetalation at aP d II complex. [33] From Int3 the release of [GeMe 3 ] + and formation of the biaryl is very facile.
Importantly,a lthough we considered at rimer as the model for nanoparticles,our computational data suggest that these reactivity trends also hold for larger Pd clusters.A s such, the computational studies suggest that bond activation under nanoparticle catalysis is reminiscent of an electrophilic aromatic substitution. Owing to its electron richness the aryl germane is ap rivileged reaction partner with electrondeficient Pd species,w hich is the origin of its superior reactivity with electrophilic Pd nanoparticles and the lack of reaction under L n Pd 0 /L n Pd II catalysis.

Conclusion
We have developed ac hemoselective coupling of aryl iodides (and diaryliodonium salts) with aryl germanes under nanoparticle catalysis in the presence of CÀBr, CÀCl, C À BPin, C À BMIDA, and additional functional groups.T he method is characterized by operational simplicity,a ir tolerance,a nd robustness and can be performed at low Pd loadings.T he aryl germanes were shown to be highly stable. Forexample,apentafluoroarylgermane tolerates strong acids or bases over extended times and at elevated temperature, whereas the corresponding boronic acid has al ifetime of milliseconds only.A ss uch, highly challenging couplings can readily be performed with aryl germanes,i ncluding those involving 2-pyridyl or polyfluoroaryl germanes.M echanistic and computational data are presented that unambiguously demonstrate that while organogermanes are the least reactive functionality under Pd 0 /Pd II homogeneous molecular catalysis as compared to established coupling partners,t hey are the most reactive group under nanoparticle conditions.The origin of this privileged reactivity was found to lie in the electron richness of the aryl germanes,which preferentially react by an electrophilic aromatic substitution type mechanism and as such are preferentially activated by more electrophilic nanoparticles.T hese features in turn allow to position organogermanes as an orthogonal coupling motif to the currently established and omnipresent cross-coupling regimes,a nd showcase truly distinguished reactivity of nanoparticles as compared to homogeneous molecular metal catalysts.  Computational study of the nanoparticle-catalyzed crosscoupling;free energy diagram in kcal mol À1 ,calculated at the CPCM (DMF) B3LYP-D3/Def2TZVPP//B3LYP-D3/Def2SVP level of theory. [33]