Catalysis with Palladium(I) Dimers

Abstract Dinuclear PdI complexes have found widespread applications as diverse catalysts for a multitude of transformations. Initially their ability to function as pre‐catalysts for low‐coordinated Pd0 species was harnessed in cross‐coupling. Such PdI dimers are inherently labile and relatively sensitive to oxygen. In recent years, more stable dinuclear PdI−PdI frameworks, which feature bench‐stability and robustness towards nucleophiles as well as recoverability in reactions, were explored and shown to trigger privileged reactivities via dinuclear catalysis. This includes the predictable and substrate‐independent, selective C−C and C−heteroatom bond formations of poly(pseudo)halogenated arenes as well as couplings of arenes with relatively weak nucleophiles, which would not engage in Pd0/PdII catalysis. This Minireview highlights the use of dinuclear PdI complexes as both pre‐catalysts for the formation of highly active Pd0 and PdII−H species as well as direct dinuclear catalysts. Focus is set on the mechanistic intricacies, the speciation and the impacts on reactivity.


Introduction
Although the use of multiple metal centers to accomplish selective and efficient catalytic transformations is ubiquitous in naturese nzymes,m ost of homogeneous metal catalysis relies on the use of solely one metal center and only few multinuclear reactivity modes have been established. [1] Despite the advantages of two or more metals working together to facilitate reactivity that might not be possible using either of the metals on their own, harnessing this synergistic power is not trivial due to the necessity of controlling nuclearity and speciation in solution. While nature uses protein backbones to control structure and speciation, in homogeneous metal catalysis similar controlling components are often absent as usually much smaller ligands are utilized. As ar esult, possibilities arise for different metal species to aggregate or dissociate,hence forming several potential active species with different nuclearity,l igation, and even oxidation state.I n particular,owing to the effects of different additives,ligands, and the reaction conditions in general, relying on the distinct formation of just one active species to utilize its reactivity is challenging.H ence an understanding of the underlying processes that determine the fate of the different metal species and thus reactivity is of utmost importance.
In this context, palladium in its even oxidation states is omnipresent in homogeneous catalysis.B yc ontrast, odd oxidation state palladium is comparatively scarce in catalysis. Due to their unpaired electrons and the potential formation of arather strong PdÀPd bond (usually about 25 kcal mol À1 ) [2] these oxidation states tend to form dinuclear complexes.A wide variety of Pd I dimers has been synthesized and has been known for decades.However,the utilization of dinuclear Pd I in catalysis has only emerged in the early 2000s.Inthe search for enhancing the activity of Pd complexes used in catalytic applications,d inuclear Pd I complexes were explored as precatalysts that can release highly active Pd 0 in situ. Tr uly dinuclear catalysis has been reported later in 2013 when our group provided unambiguous support for the direct involvement of aPd I À Pd I framework. With the advent of dinuclear Pd I catalysis not only their unique reactivity and selectivity but also their straightforward application as robust, efficient, and reusable catalysts has been explored.
In this Minireview we (i)present the synthesis and properties of dinuclear Pd I complexes relevant in catalysis,(ii)summarize their application as pre-catalysts and the implications of their in situ formation, and (iii)showcase their utility in dinuclear catalysis.

Synthesis and Properties of Pd I Dimers
Pd I dimers are connected by aPd À Pd bond, forming the diamagnetic dimeric [Pd I À Pd I ] 2+ core unit while the single d 9 Pd I center possesses an unpaired electron. Pd I dimers can be classified on the basis of their ligand structure into supported or unsupported dimers,depending on whether abridging ligand is present or not. In unsupported dimers both Pd centers adopt as quare planar geometry in which one coordination site is occupied by the other Pd atom forming asingle Pd À Pd s-bond without the aid of any additional bridging ligands.In supported dimers the bonding between the two Pd I centers is more extensive due to the influence of additional bridging ligands.M ost commonly,P d I dimers are supported, that is, single-atom-(e.g. halides), allyl-, or arene-bridged, as illustrated in Figure 1( top). Furthermore,h ybrids between the conceptual types of Pd I dimers are feasible.F or ad etailed discussion of the structural diversity of Pd I dimers we refer our readers to previous comprehensive review articles. [3] Pd I dimers are typically synthesized by (i)oxidation of Pd 0 , [4] (ii)comproportionation of Pd 0 and Pd II ,o r( iii)reduction of Pd II (Figure 1). In addition to those direct approaches, Pd I dimers can be accessed via ligand exchange from preformed Pd I dimers. [5] Thef irst synthesis of m-halide-supported dimers [Pd I (m-X)(P t Bu) 3 ] 2 1 (X = Br) and 2 (X = I) was accomplished by oxidation of Pd 0 (P t Bu 3 ) 2 with highly activated electrophiles, Dinuclear Pd I complexes have found widespread applications as diverse catalysts for am ultitude of transformations.Initially their ability to function as pre-catalysts for low-coordinated Pd 0 species was harnessed in cross-coupling.SuchPd I dimers are inherently labile and relatively sensitive to oxygen. In recent years,m ore stable dinuclear Pd I À Pd I frameworks,w hichfeature bench-stability and robustness towards nucleophiles as well as recoverability in reactions,w ere explored and shown to trigger privileged reactivities via dinuclear catalysis.T his includes the predictable and substrate-independent, selective C À Ca nd C À heteroatom bond formations of poly-(pseudo)halogenated arenes as well as couplings of arenes with relatively weak nucleophiles,whichwould not engage in Pd 0 /Pd II catalysis. This Minireview highlights the use of dinuclear Pd I complexes as both pre-catalysts for the formation of highly active Pd 0 and Pd II À Hspecies as well as direct dinuclear catalysts.F ocus is set on the mechanistic intricacies,the speciation and the impacts on reactivity.
for example,H CX 3 , N-halosuccinimides or CX 4 (X = Br or I). [6] However,the desired Pd I dimers could only be isolated in moderate yields and stoichiometric amounts of side-products were generated.
Another strategy is the comproportionation of L n Pd 0 and Pd II X 2 (X = halide). This strategy was,f or example,u sed in the synthesis of bromide-bridged dimers 1, 5,a nd 9 from either PdBr 2 or (cod)PdBr 2 , [7] but also for Pd I iodo dimers 2, 3, and 4 from PdI 2 . [8] Usually the PdL 2 complexes are preformed, but in case of dimer 2 it can also be formed in situ from commercially available Pd 2 (dba) 3 . [8a] Furthermore,Pd I dimers can be synthesized via reduction from corresponding Pd II precursors,s uch as Pd II X 2 or preformed mono-or dimeric Pd II complexes.R eductive approaches were employed in the synthesis of Pd I dimers 1, 6,and 12 with the aid of basic alkoxide solution, [9] as well as in the synthesis of dimers 10, 11,a nd 13 where AgBF 4 , phosphine ligand, and Et 3 Ns erved as key reagents and reductants. [10] While attempts to synthesize the N-heterocyclic carbene ligated dimer 6 by oxidation or comproportionation failed, the reduction from the corresponding Pd II dimer [(IPr)Pd II I 2 ] 2 succeeded. [9b] In contrast, the corresponding bromide-bridged dimer remained elusive in an analogous synthesis attempt. Theu nderlying reason is not well understood;h owever, reduction mechanisms to Pd I can involve the formation of inorganic by-products. [9a] Moreover,t here is ac omplex interplay of stabilizing ligands and anions to maintain the (+ 1) oxidation state in the Pd I À Pd I core.For instance,the impact of the bridging halide is  nicely illustrated in the formation of iodide-bridged dimer 7. While the Pd I iodo dimer 7 was formed in high yield, the corresponding bromide-bridged Pd I dimer was not obtained under identical reaction conditions.I nstead, its conformational isomer 8 possessing bridging arene units and ancillary bromide ligands is formed. [7d] Theprecise effects of steric bulk and the donor-acceptor properties of the ancillary and bridging ligands on the structure,s tability,a nd eventually the reactivity of Pd I dimers are not yet understood and subject to ongoing research.

Pd I Dimers as Pre-Catalysts for Pd 0
Although more than 50 different Pd I dimers have been synthesized to date and characterized since the 1970s,o nly few were successfully implemented as pre-catalysts in the ever-growing field of metal-catalyzed cross-coupling.
Owing to its oxygen sensitivity,d imer 1 can only be handled for avery brief period in open laboratory conditions as asolid (e.g.for rapid weighing). In solution, it is essentially immediately deactivated in the presence of oxygen. In contrast, the iodinated dimer 2 [7a, 19] is completely stable in air as as olid and overall very robust. In fact, dimer 2 was unreactive under many of the conditions used for 1,w hich might explain why,until the year 2013, there was only asingle report of its application in catalysis which related to carbonylations of aryl halides. [20] With the aim to improve the liberation of the catalytically active Pd 0 species from Pd I dimers,subsequent efforts focused on developing even more labile variants.I nt his context, Barder,[10a] Vilar,[7d] Spokoyny, [5] as well as Shaughnessy and Colacot [9c] independently established the use of Pd I dimers 9, 10,a nd 12 as labile pre-catalysts in Suzuki cross-coupling,i n amination, as well as in the a-arylation of carbonyl derivatives.
Theinsitu release of the catalytically active species from the pre-catalyst and its speciation influence the reactivity and selectivity in chemical transformations.I nt his context, the crucial question with regard to Pd I dimers as pre-catalysts was to identify the actual speciation, especially its nuclearity,t o explain the distinct and frequently enhanced reactivity as compared to established Pd 0 or Pd II pre-catalysts.
In 2012, our group examined the nature of active species derived from [Pd(m-Br)(P t Bu 3 )] 2 (1)inSuzuki cross-couplings of aryl bromides and chlorides,i ts potential activation mechanism, as well as the origins of limitations in scope. [12] We confirmed that 1 solely serves as reservoir for the release of Pd 0 (P t Bu 3 ) [12] but is otherwise not reactive itself in these transformations.T he high activity seen is due to ad ifferent activation mechanism of the complex as compared to,f or example,L 2 Pd 0 (L = P t Bu 3 ), where dissociation of aphosphine ligand needs to occur which is associated with as ignificant energy penalty. [21] By contrast, the Pd I dimer 1 is most likely activated in an ucleophile-assisted fragmentation. [22] The direct in situ disproportionation (which had frequently been assumed in the literature prior to that) was ruled out on the basis of computational studies.
In Suzuki cross-couplings,the activation of the Pd I dimer (to Pd 0 )w as found to be triggered by hydroxide or the mixture of boronic acid/KF,t hat is,f requently employed reagents and additives in Suzuki-Miyaura-type cross-couplings. [22] By contrast, the corresponding iodide-bridged Pd I dimer,[Pd(m-I)(P t Bu 3 )] 2 (2), is not activated by these species. Consequently,P d I dimer 2 does not catalyze Suzuki crosscouplings with boronic acid/KF conditions.H owever,o ur group showed that the in situ release of Pd 0 (P t Bu 3 )from these Pd I dimers is dependent on the adequate choice of additive, [22] and have established the minimum nucleophilicity necessary (N scale) [23] to do the activation in each case ( Figure 3). Using oxygen-or nitrogen-centered nucleophiles with N ! 16 also activated Pd I dimer 2 and with such bases,S uzuki coupling was then possible. [22] Within our ongoing research program regarding the application of aryl germanes [24] in synthesis and catalysis we found DABCO (N = 18.80) [25] to be asuitable nucleophile for the activation of Pd I dimer 2 and subsequent functionalization of aryl tetrafluorothianthrenium salts ( Figure 3). [26] Thus, am ild and selective method was developed which allows the direct conversion of non-activated aromatic C À Hs ites into aryl germanes via the tetrafluorothianthrenium salt as key intermediate. [26b] It is worth noting that we observed that those pre-catalysts,w hich had previously been reported to efficiently transform thianthrenium salts,d id not give rise to efficient CÀGeEt 3 bond formation, nor did [Pd 0 (P t Bu 3 )] 2 .

In situ Formation of Pd I Dimers and its Implications
Aryl bromides appear especially privileged in reactions with Pd I dimers 1 or 2.Adetailed study of the Suzuki coupling of aryl chlorides with dimer 1 indicated that, although the coupling product forms relatively rapidly,aplateau in conversion is reached and hardly any further conversion to product takes place thereafter. [12] Thed ata indicated that ligand exchange from Pd 0 (P t Bu 3 )toform Pd 0 (P t Bu 3 ) 2 and Pdblack (under precipitation of Pd) occurs over time.O nce Pd 0 (P t Bu 3 ) 2 is solely remaining,reactivity is low.Onthe other hand, the complete conversions in cross-couplings of aryl bromides with dimer 1 might potentially originate from an in situ regeneration of 1.F or example,H artwig identified an unusual autocatalytic behavior in the oxidative addition to ArBr by P t Bu 3 -derived Pd 0 catalysts in which, among other species,the formation of Pd I dimer 1 was detected. However, in this study the autocatalytic behavior was ultimately ascribed to (P t Bu 3 ) 2 Pd II (H)(Br) being formed. [27] As part of our investigations of cross-couplings in air with Pd I dimer 2 (discussed in Section 3.4) we discovered that Pd I dimers can also be formed by aerobic oxidation of Pd 0 species. [28] When Pd 0 (P t Bu 3 ) 2 was treated with n Bu 4 NI and PhMgCl (2 equiv relative to Pd) in the presence of oxygen, the corresponding Pd I dimer 2 was formed along with biphenyl. These data indicated that any Pd 0 species released in the reaction mixture during catalysis can, in principle,b e transformed back to Pd I under suitable conditions (Figure 4).
Colacot and his team found that the combination of phosphine ligand and Pd II Br 2 gives rise to the formation of Pd I dimer 1,w hich was ultimately harnessed in ap rocess for the synthesis of 1 on larger scale. [9a, 29c] In ac ollaborative experimental and computational study,w ee xamined the mechanistic intricacies of this unusual reduction from Pd II to Pd I (as opposed to Pd 0 ). It was found that unusual side species may also form in the process (such as (BrP t Bu 3 )(Br) and [Pd 2 II Br 6 ]-[(BrP t Bu 3 )(Br)] 2 ), which in turn are greatly influenced by the ligand stoichiometry,a dditives,o rs olvent. Thei mplications for the catalytic performance were also illustrated. [9a] In 2014, the group of Hazari observed the formation of Pd I dimers by in situ comproportionation of (IPr)Pd 0 and (IPr)Pd II (Cl)(allyl) complexes. [29b] It was demonstrated that by modulating the steric demand of substituents on the allyl ligand an improved catalytic performance in cross-couplings can be obtained. Increased steric bulk on the ligand was found to accelerate the release of Pd 0 and decrease the tendencyto (re-)dimerize,o verall favoring the in situ liberation of catalytically competent Pd 0 .T hese studies demonstrate the multitude of in situ activation and deactivation pathways for Pd (pre-)catalysts via readily occurring redox processes and hence reinforce the challenges associated with such interchangeable Pd speciation in defining the catalytically active species in Pd catalysis.
In 2012, our group reported on the impact of additives on the in situ formation of Pd I dimers and the resulting rateaccelerating or inhibiting effects in catalysis. [29a] We uncovered that upon addition of aC uo raAg salt-common additives in, for example,S onogashira or Suzuki crosscoupling reactions-to Pd 0 (P t Bu 3 ) 2 aP d I dimer and the corresponding Cu I or Ag I cubanes are formed. We showed that with CuBr 2 the highly reactive [Pd(m-Br)(P t Bu 3 )] 2 1 is generated from Pd 0 (P t Bu 3 ) 2 ,w hich in turn can liberate Pd 0 (P t Bu 3 )m ore readily in situ and consequently accelerate  . In situ liberationo fcatalytically active monoligated palladium from Pd 0 L 2 or Pd I dimer as pre-catalysts (top) [12,21] and N-scale quantification for dimer activation (bottom). [22] Asuitable nucleophile is required for activation, which is present as areagent or additive and should not function well as stabilizing bridge.

Angewandte
Chemie diverse transformations.I ns tark contrast, the corresponding cross-couplings with CuI are often slowed or even completely inhibited. We found that upon addition of CuI, [Pd(m-I)(P t Bu 3 )] 2 2 is formed in ar edox process. [29a] However,i n comparison to Pd I bromo dimer 1,[Pd(m-I)(P t Bu 3 )] 2 2 displays high stability in solution towards diverse nucleophiles and is thus less competent to release Pd 0 (P t Bu 3 )asactive species in CÀCc ouplings,w hich in turn leads to decreased activity. Moreover,wefound that Pd I dimers 1 or 2 are competent to react directly with alkynes,a nd their in situ formation under typical Sonogashira coupling conditions (i.e., Cu salt/Pd 0 )will hence lead to an unproductive consumption of alkyne as competing process. [29a] This side-process is especially pronounced in the Sonogashira coupling of aryl chlorides and delivers an explanation of the previously observed inhibiting effects of Cu or Ag additives in Sonogashira couplings. [30]

Mechanistic Support for Dinuclear Reactivity
Thec atalytic role of Pd I dimers was solely ascribed to being pre-catalysts until the year 2013, when our group disclosed that Pd I dimers could also react directly with an aryl halide.Inacombined experimental and computational study we provided unambiguous support of this and provided am olecular mechanistic picture. [31] We discovered that Pd I dimer 1 can undergo ah alide exchange reaction with aryl iodides to form aryl bromides by formally exchanging the mbridging bromide for iodide ( Figure 5). 31 PNMR revealed the formation of different dimers,g enerated upon exchange of one or both of the bridges from bromide to iodide,b ut no species related to Pd 0 ,P d II ,o rf ree phosphine ligand were detected in the stoichiometric process.M oreover,P d 0 was shown to be ineffective to trigger this halogen exchange,and an independently synthesized Pd II complex did not lead to formation of the product via reductive elimination, overall excluding the involvement of Pd 0 or Pd II .Apotential homolytic cleavage of the dimer into Pd I radicals was found unlikely because the computed activation barrier is roughly 8kcal mol À1 higher than the alternative direct oxidative addition at dinuclear Pd I .T his is also in line with the fact that no paramagnetic species were observed in the course of the reaction. Additionally,w hen radical initiators such as AIBN/Bu 3 SnH were employed, no product formation was observed. Likewise,t he addition of H-atom donors such as 1,4-cyclohexadiene did not significantly influence the reaction outcome. [31] Kinetic studies furthermore corroborate dinuclear reactivity as first-order dependence in Pd I dimer was observed. [32] Based on these observations as well as computational studies,adinuclear mechanism was proposed. Several modes of activation of the aryl halide were explored computationally ( Figure 5, bottom) including (i)oxidative addition, (ii)activated exchange at the dinuclear platform, and (iii)an ipsosubstitution where acationic dimer acts as aL ewis acid. The oxidative addition mechanism was found to be clearly favored by almost 30 kcal mol À1 over the alternatives.I nterestingly, while the calculated transition state suggests bond activation to occur primarily at one Pd center of the Pd I dimer,a fter oxidative addition aPd II dimer was computationally obtained, which suggests that overall aP d I ÀPd I /Pd II ÀPd II oxidative addition occurs,w hich involves both Pd centers (instead of aPd I /Pd III mechanism). Pleasingly,the halide exchange reaction was also feasible using only catalytic amounts of Pd I dimer 2 in the presence of an excess of n Bu 4 NBr as an ucleophile ( Figure 6). The proposed catalytic cycle involves an initial one-or two-fold nucleophilic exchange of the m-bridging nucleophiles of the Pd I dimer prior to oxidative addition of the aryl halide. [31][32] Thei ntermediately formed Pd II dimers then undergo reductive elimination to yield the product and re-form aPd I dimer. This new concept relies on the employed coupling partner (nucleophile) to overall be able to stabilize the dinuclear Pd I framework. Notably,h owever,w ithin this dinuclear catalysis concept the oxidative addition and transmetalation elementary steps are reversed, which circumvents the formation of potentially unreactive Pd II by-products during transmetalation (see Section 3.4). Moreover,s ince the nucleophilic exchange takes place at oxidation state (I) [rather than the usual (II) as in Pd 0 /Pd II catalysis] different driving forces for the exchange process may overall exist, which in turn may enable privileged reactivities. Figure 5. Dinuclear Pd I -mediated halide exchange-experimentala nd computational support for dinuclearc atalysis. [31,32] Angewandte Chemie Minireviews

Distinct Bond-Formations Harnessing Dinuclear Catalysis
Following the initial observation of dinuclear catalysis, our group examined the wider synthetic impact of this concept. Our developments focussed on using the iodidebridged Pd I dimer 2 as ac atalyst because this complex is completely air-stable as asolid and can be stored on the bench without special precautions.T he corresponding coupling partner was employed as asalt.
We developed the direct catalytic trifluoromethylthiolation (CÀSCF 3 coupling) of aw ide range of aryl iodides and bromides using the bench stable salt NMe 4 SCF 3 . [8a,33] In this context, the Pd I concept complemented existing Pd 0 -o rN i 0based strategies that were limited in compatible aryl halide. [34] Thef irst direct catalytic C À SeCF 3 bond formation of aryl iodides and bromides was also accomplished using the corresponding NMe 4 SeCF 3 salt. [8a] In both cases nucleophile exchange at the dinuclear Pd I scaffold was confirmed in stoichiometric studies and the novel SCF 3 -and SeCF 3 -bridged Pd I dimers were isolated and characterized (Figure 7). Both complexes were also completely air-stable. 31 Pa nd 19 FNMR monitoring of their stoichiometric reaction with aryl iodide shows the step-wise transfer of the trifluoromethylated chalcogenide nucleophiles from the Pd I dimers to form the product. This is in line with the computed free-energy pathway that suggests feasible activation barriers for the direct reactivity and an overall exergonic thermodynamic driving force.U sing catalytic amounts of Pd I dimer 2 along with chalcogenide nucleophile in the form of soluble tetramethylammonium salts Me 4 NSCF 3 and Me 4 NSeCF 3 ensured formation of the active dimers in situ and enabled catalytic turn-over. Thef ormed Pd I dimers were found to be very robust and their recovery after the reaction was feasible by simple open-atmosphere column chromatography and even allowed for their reuse in further reactions without significant loss in activity. [8a, 33a] Following our initial work on the electron-deficient SCF 3 / SeCF 3 nucleophiles in Pd I catalysis our group reported the thiolation and selenolation of aryl iodides and bromides using sodium thiolates and selenolates as electron-rich nucleophilic coupling partners (Figure 8). [35] In this case the dinuclear Pd I catalysis concept provided advantages as compared to established Pd 0 /Pd II chemistry.I np articular,c atalyst poisoning by the electron-rich nucleophiles via formation of deactivated Pd II -ate off-cycle intermediates-a recognized problem in Pd 0 catalysis-was not encountered. Moreover, the developed Pd I protocols allowed for an exclusive and apriori predictable site-selectivity as only aryl iodides and bromides were reactive in the presence of other potentially reactive functional groups such as CÀCl and CÀOTfg roups. ForP d 0 -based strategies,s ite selectivity is often not general, but substrate-specific and sometimes unpredictable,w hereas for Pd I it is dependent solely on the leaving group (see additional discussion below). Additionally,p reviously no catalytic methods existed for the direct selenolation of aryl halides.The remarkable stability and robustness of Pd I dimer 2 allowed for several recovery cycles of the catalyst by simple column chromatography maintaining its catalytic activity.
Phosphorothioates ( À SP(=O)(OR) 2 )w ere also demonstrated to be suitable coupling partners in Pd I -mediated bond formation, which enabled the first C À SP( = O)(OR) 2 coupling of aryl halides (Figure 8). By contrast, traditional Pd 0 /Pd II catalysis failed to deliver the phosphorothioate cross-coupling product as ac onsequence of ap rohibitively high activation barrier for the reductive elimination from the key Pd II intermediate and al ack of driving force. [36] Stoichiometric experiments with Pd I dimer 2 and 31 PNMR analysis clearly indicated the formation of an ew Pd I dimer,d emonstrating the competence of phosphorothioates to stabilize the Pd I scaffold and the altered driving force for exchange at Pd I . Furthermore,computational analysis predicted aclear driving force for the direct reactivity of phosphorothioate-bridged Pd I Figure 6. First dinuclear Pd I catalysis (top) and proposed mechanism (bottom). [31,32] Figure 7. Application of dinuclear Pd I catalysis in trifluoromethylthiolations and -selenolations:isolation of catalytically competent SCF 3 -and SeCF 3 -bridgedP d I dimers (top) and scope of the catalytic transformations (bottom). [8a, 33a] Angewandte Chemie dimers,w hich was confirmed in experiments.Anumber of aryl iodides were efficiently phosphorothioated using convenient tetramethylammonium phosphorothioate salts and the air-stable Pd I dimer 2 as catalyst. Theprotocol is general, operationally simple and tolerates structural diversity on the phosphorothioate backbone.A ryl esters as well as axially chiral phosphorothioates were coupled with high yields.T he chiral information at phosphorus was retained in the coupling process (Figure 8). [36] Al ong-standing challenge in Pd 0 catalysis is the siteselective functionalization of poly(pseudo)halogenated arenes (bearing,f or example,C ÀBr,C ÀCl, CÀOTfi nt he same molecule). Thes electivity in standard Pd 0 /Pd II catalysis is dependent on the catalyst/ligand, solvent, additives,a nd reaction conditions, [8b,37] but most importantly,a lso the substrate itself. [38] Subtle steric and electronic effects can vastly impact the selectivity.Consequently,selective coupling conditions identified for agiven substrate may no longer give rise to selective coupling with the next. Fori nstance,4bromophenyl triflate underwent selective Kumada coupling at C À Br with PdCl 2 (P(o-tol) 3 ) 2 ,b ut when two methyl groups were introduced ortho to C À Br amixture of couplings at C À Br and C À OTfw as observed under otherwise identical reaction conditions (Figure 9, top). [28,37i] With Pd I dimer 2 af ully apriori predictable arylation and alkylation protocol was achieved using organomagnesium [22] or organozinc species as appropriate nucleophilic cross-coupling partners. [28,39] Themethod enables the functionalization at C À Br sites in the presence of other reactive functional groups,i ncluding C À OTfa nd/or C À Cl, within seconds to af ew minutes at room temperature in air. Thep rotocol also allows for larger-scale applications.Both yield and selectivity were found to be fully substrate-independent and even highly sterically demanding ortho-adamantyl C À Br sites were selectively functionalized while leaving C À OTf and C À Cl sites untouched. [40] Alkylation of aryl (pseudo)halides has been an especially challenging area of development. Challenges include b-hydride eliminations as side processes or metal-halogen exchange.F ew catalysts have been developed that efficiently overcame these challenges.O ur tests of these state-of-the-art alkylation catalysts gave no selectivity when challenged with poly-(pseudo)halogenated arenes however. As such, the Pd I -based CÀCbond formations are asignificant advance in being fully predictable in selectivity,air-tolerant, and extremely efficient (reaction times of seconds to afew minutes).
While the Pd I dimer 2 triggers solely CÀBr functionalization in the presence of C À OTforC À Cl in toluene or THF,in NMP,D MPU,D MI, or DMAc efficient functionalization of aryl triflates and chlorides can also be achieved. [41] Culminated in the modular,s ite-selective diversification of polyhalogenated arenes in the reactivity order Br > OTf > Cl, one-pot double and triple functionalizations sequences were performed in less than 50 minutes under mild reaction conditions.T he coupling procedure was further expanded to the functionalization of (or in the presence of) aryl fluorosulfates (ArÀOSO 2 F), am otif of increasing interest due its properties which are relevant to medicinal chemistry and material science,f or which no site-selective coupling had previously existed. ThePd I dimer 2 enabled the site-selective functionalization of C À Br sites in presence of the OSO 2 F (= OFs) groups as well as using the OFs moiety as asustainable and inexpensive triflate surrogate. [42] Furthermore,o ther CÀCb ond formations,s uch as the Heck reaction of terminal olefins and a-arylations of esters and ketones,w ere also enabled by the air-stable Pd I dimer 2 Figure 8. Heteroatom bond formation via dinuclear catalysisu sing thiolates or selenolates (top) [35] and phosphorothioates( bottom). [36] showing the same exquisite chemoselectivity (I, Br @ Cl, OTf). [44] Despite harsh conditions (100 8 8Ci np resence of base) recovery of the dinuclear Pd I catalyst was possible after Heck reaction. [44a] Bromide-bridged dimer 1 had also been shown to be an efficient pre-catalyst for a-arylations,although specifically tailored conditions were required for different carbonyls. [18a,d-f] In contrast, using the air-stable dimer 2 allowed for am ore general method that enabled easy access to pharmaceutically relevant a-arylated cyclopropyl ketones. [44b] Our group further showed that the outstanding reactivity and selectivity of the Pd I dimer 2 as cross-coupling catalyst for small molecules remains viable as ag eneral strategy to also polymerize valuable macromolecules,that is,polyfluorenes or poly-para-phenylenes. [43] Starting from monomers containing two bromides,i ns itu lithiation and transmetalation to the corresponding organozinc enabled rapid polymerizations with good average molecular weights (M n )a nd polydispersities (PDI) using Pd I dimer 2 as ac atalyst, even in the presence of air. As compared to the single CÀCb ond formations above,t he polymerization is similarly general, selective and rapid (in seconds) for av ariety of different monomers and polymers.T his is as ignificant advance to previous M 0/II polymerization strategies where for each monomer as pecific catalyst needs to be tuned and usually long reaction times and/or elevated temperatures are required to achieve high conversion as well as the exclusion of oxygen is needed. [45]

Applications in Other Transformations
Gooßen and co-workers demonstrated the one-bond olefin migration of allylic esters to make enol esters in moderate E/Z selectivity using [Pd(m-Br)(P t Bu 3 )] 2 1 (Figure 10). [46] This as well as previous work [7a, 27] indicated that a[ Pd II ÀH] is generated in situ from the labile dimer 1 under these conditions.G ooßen also found that since am onophosphine-derived Pd II ÀHs pecies is liberated, the corresponding isomerization can be coupled with metathesis in the same pot. [47] Other typical isomerization catalysts,s uch as (P t Bu 3 ) 2 Pd II (Cl)(H), release phosphine in situ, which in turn would inhibit the metathesis catalyst and therefore make them incompatible ( Figure 10). [48] Parker and co-workers demonstrated the potential of this Pd I -dimer-catalyzed isomerization approach by implementing it as ak ey synthetic step in at otal synthesis towards tesirine and pyrrolobenzodiazepines on ag ram scale. [49] Then ature of the catalytically active Pd species-in particular its nuclearity as the key parameter determining its reactivity-was investigated in apreliminary DFT analysis in the original report [46] and recently re-investigated and -evaluated in at horough DFT analysis in collaboration with the Koley group. [50] Their DFT analysis ruled out that the Pd I dimer itself is sufficiently activated to perform the isomerization of olefins.I nstead, the data indicate that it serves as the source of am ononuclear Pd II ÀHs pecies.T he in situ release is suggested to proceed via cyclopalladation with one of the coordinating trialkyl phosphine ligands and eventually Figure 9. CÀCbond formation by dinuclear Pd I catalysis: site-selective arylations and alkylations (top), [22,28,39,40] modular,sequential diversification (middle), [41,42] and polymerization (bottom). [43] dissociation of the dimeric cyclopalladated Pd species to yield amononuclear and highly reactive L 1 BrPd II ÀH.
Thes ynthetic value of Pd I -dimer-catalyzed double-bond migration and subsequent reactivity via as econd catalytic transformation was furthermore demonstrated in the joint isomerization/hydroformylationo ff atty acid methyl esters, reported by Vorholt and co-workers in 2017. [51] Although [Pd(m-Br)(P t Bu 3 )] 2 1 is,o wing to its high reactivity,apotent pre-catalyst in olefin isomerizations,i t readily decomposes in solution when exposed to air.
In 2020, our group established [Pd(m-I)(PCy 2 t Bu)] 2 3 as an ew air-stable dimer and extended the portfolio of Pdcatalyzed olefin isomerizations ( Figure 11). [8c] Dimer 3 allows for adouble-bond isomerization for the first time under openflask conditions fully exposed to oxygen with more than 40 substrates in excellent yields and with high E/Z selectivities.T he polar protic solvent MeOH proved to be optimal, which contrasts previous developments that relied on nonpolar solvents.B yc ontrast, the bromide-bridged Pd I dimer 1 is ineffective under the same conditions as it is rapidly deactivated in the presence of oxygen. Conversely,t he previous generation of air-stable Pd I dimers,t hat is, 2,w as not effective in this transformation either,a si ti st oo robust and does not liberate Pd II ÀH. Ther eactions with 3 display aw ide scope,l eaving stereocenters in substrates untouched; amino acid derivatives were isomerized with high E/Z ratio. Furthermore,dimer 3 was also shown to be apotent catalyst for site-selective CÀCbond formations and even superior to 2 for CÀOTfc ouplings that bear ortho substituents.

Summary and Outlook
In this Minireview we have highlighted the use of dinuclear Pd I complexes as catalysts in aw ide range of transformations.T heir applications span from being efficient pre-catalysts for highly active monophosphine Pd 0 or Pd II À H species in cross-coupling and olefin isomerization reactions, respectively,t od irect dinuclear catalysis.W hile the airsensitive and labile bromide-bridged Pd I dimer [Pd I (m-Br)-(P t Bu) 3 ] 2 received most attention in the earlier literature since 2002 for its success in triggering especially aminations or Suzuki cross-couplings of aryl halides as well as a-arylations of carbonyl compounds,t he corresponding air-stable iodidebridged Pd I dimer [Pd I (m-I)(P t Bu) 3 ] 2 remained essentially unexplored in catalysis until adecade later.The latter is much more stable and frequently unreactive under standard crosscoupling conditions,a si td oes not readily liberate Pd 0 .T his feature has been harnessed since 2013 when the feasibility of dinuclear catalysis was established, both mechanistically as well as synthetically.H ere,t he Pd I À Pd I bond stays intact during catalysis,w hich has enabled ar ange of selective transformations that are not amenable to Pd 0 /Pd II catalytic cycles.Especially weaker nucleophiles could be implemented owing to the altered driving forces associated with exchanges at Pd I as opposed to Pd II .The associated air-stability of the Pd I species allowed for its recovery and reuse (with multiple rounds of recycling). Moreover,a so pposed to Pd 0 -based catalysis,e xcellent apriori predictable selectivities for poly-(pseudo)halogenated arenes,featuring aryl bromide,chloride, triflate,a nd/or fluorosulfate functionalities were achieved  Figure 10. Isomerization (and metathesis) strategies with Pd I dimer as pre-catalyst. [46, 48a,c, 49] Angewandte Chemie with [Pd(m-I)(P t Bu 3 )] 2 .N otably,t he CÀCb ond formations were accomplished in seconds to af ew minutes at room temperature under open-flask conditions (tolerating oxygen). Given the exquisite selectivities paired with high reactivity, practicability (air tolerance,recyclability), as well as potential for diverse and privileged catalysis,w ee xpect to see many more powerful applications in the years to come.Key in this area is to control the speciation of the Pd I dimer,which in turn controls the mode of reactivity.Asthis is adelicate interplay of multiple factors,i ncluding ligands and additives,t he area will continue to benefit greatly from fundamental insight to guide future developments and novel dimer generations.