Isolable acetylene complexes of copper and silver

Copper and silver play important roles in acetylene transformations but isolable molecules with acetylene bonded to Cu(i) and Ag(i) ions are scarce. This report describes the stabilization of π-acetylene complexes of such metal ions supported by fluorinated and non-fluorinated, pyrazole-based chelators. These Cu(i) and Ag(i) complexes were formed readily in solutions under an atmosphere of excess acetylene and the appropriate ligand supported metal precursor, and could be isolated as crystalline solids, enabling complete characterization using multiple tools including X-ray crystallography. Molecules that display κ2-or κ3-ligand coordination modes and trigonal planar or tetrahedral metal centers have been observed. Different trends in coordination shifts of the acetylenic carbon resonance were revealed by 13C NMR spectroscopy for the Cu(i) and Ag(i) complexes. The reduction in acetylene 
<svg xmlns="http://www.w3.org/2000/svg" version="1.0" width="13.454545pt" height="16.000000pt" viewBox="0 0 13.454545 16.000000" preserveAspectRatio="xMidYMid meet"><metadata>
Created by potrace 1.16, written by Peter Selinger 2001-2019
</metadata><g transform="translate(1.000000,15.000000) scale(0.015909,-0.015909)" fill="currentColor" stroke="none"><path d="M160 680 l0 -40 200 0 200 0 0 40 0 40 -200 0 -200 0 0 -40z M80 520 l0 -40 40 0 40 0 0 -40 0 -40 40 0 40 0 0 -200 0 -200 40 0 40 0 0 40 0 40 40 0 40 0 0 40 0 40 40 0 40 0 0 40 0 40 40 0 40 0 0 40 0 40 40 0 40 0 0 120 0 120 -80 0 -80 0 0 -40 0 -40 40 0 40 0 0 -80 0 -80 -40 0 -40 0 0 -40 0 -40 -40 0 -40 0 0 -40 0 -40 -40 0 -40 0 0 160 0 160 -40 0 -40 0 0 40 0 40 -80 0 -80 0 0 -40z"/></g></svg>
 C 
<svg xmlns="http://www.w3.org/2000/svg" version="1.0" width="23.636364pt" height="16.000000pt" viewBox="0 0 23.636364 16.000000" preserveAspectRatio="xMidYMid meet"><metadata>
Created by potrace 1.16, written by Peter Selinger 2001-2019
</metadata><g transform="translate(1.000000,15.000000) scale(0.015909,-0.015909)" fill="currentColor" stroke="none"><path d="M80 600 l0 -40 600 0 600 0 0 40 0 40 -600 0 -600 0 0 -40z M80 440 l0 -40 600 0 600 0 0 40 0 40 -600 0 -600 0 0 -40z M80 280 l0 -40 600 0 600 0 0 40 0 40 -600 0 -600 0 0 -40z"/></g></svg>
 C due to metal ion coordination is relatively large for copper adducts. Computational tools were also used to quantitatively understand in detail the bonding situation in these species. It is found that the interaction between the transition metal fragment and the acetylene ligand is significantly stronger in the copper complexes, which is consistent with the experimental findings. The CC distance of these copper and silver acetylene complexes resulting from routine X-ray models suffers due to incomplete deconvolution of thermal smearing and anisotropy of the electron density in acetylene, and is shorter than expected. A method to estimate the CC distance of these metal complexes based on their experimental CC is also presented.


Introduction
Acetylene (C 2 H 2 ) is a useful building block in organic and industrial chemistry. 1 It is usually obtained from coal via a process involving calcium carbide (which is different from the petroleum-based, other important C2-feedstock, ethylene). 1c,2 However, compared to ethylene, the applications involving acetylene are somewhat challenging due to its re and explosion risks, especially under high-pressure conditions and in puried form. 1c Furthermore, additional care must be taken when certain metals such as copper and silver are involved because they are known to form explosive acetylides and carbides with acetylene. 1a,3 Nevertheless, transition metals, including copper and silver, have been utilized successfully in many acetylene transformations. 1a,b,4 Selective semihydrogenation of acetylene in ethylene-rich gas streams to produce ethylene is one such application with great industrial importance, as it serves as an effective method to remove acetylene impurities in ethylene feedstocks. Silver-modied palladium is the most commonly used catalyst for this purpose. 5 Various other silver and copper containing materials and copper complexes are also known to facilitate this process. 5a,6 Silver mediated addition 7 and carboxylation 8 reactions of acetylene and use in acetylene sensing 9 have been reported. Copper and/or copper salts also play diverse roles in acetylene chemistry as in the ethynylation (e.g., in the 1,4-butynedione synthesis), hydrochlorination, carbonylation, cross-couplings, and azide-alkyne cycloaddition reactions, as well as vinylacetylene and cuprene synthesis. 1,10 Acetylene has also been separated very effectively from CO 2 using copper containing materials. 11 The metal carbide formations noted above could be considered as "C-H activation" processes. 12 Although limitations must be considered, the advancements stated herein show that copper and silver play an integral role in the acetylene chemistry.
The fundamental chemistry such as structures and bonding of p-acetylene complexes of copper and silver are of signicant interest because they provide useful information for the design and development of processes for separation, 13 activation, and utilization of this important C2-feedstock chemical. 1 However, despite over a 100 year history of coinage metal (Cu, Ag, Au) chemistry of acetylene, 3b,14 and the current importance, 1a 2 ]Cu(C 2 H 2 ) (4), 18 and four silver complexes [HB(3,5-(CF 3 ) 2 Pz) 3 ]Ag(C 2 H 2 ) (5), 19 [Ag(C 2 H 2 ) 3 ][Al(OC(CF 3 ) 3 ) 4 20 4 ] Ag(C 2 H 2 ) (8) 20 containing terminal M(h 2 -HC^CH) bonds (Fig. 1, M ¼ Cu, Ag). It is also noteworthy that these few isolable species differ in terms of charge, coordination number and/or supporting ligands, and therefore are of limited use for comparisons. Even the gas-phase studies of Cu and Ag acetylene species are quite limited. 21 This scarcity is perhaps due to challenges such as facile loss of coordinated acetylene, metal acetylide and carbide formation, and the potential safety hazards associated with this work. 3 Considering the importance of copper and silver in acetylene chemistry, we set out to uncover and characterize a group of molecules suitable for detailed comparisons and analysis. Herein we report the successful stabilization of several p-acetylene complexes of copper(I) and silver(I) and their spectroscopic features and X-ray crystal structures (Fig. 2). Furthermore, in this work, we demonstrate the utility of bis-and tris(pyrazolyl)borate ligands, [Ph 2 B(3-(CF 3 )Pz) 2 ] À , [HB(3,5-(CF 3 ) 2 Pz) 3 ] À , and [HB(3-(CF 3 ),5-(Ph)Pz) 3 ] À to stabilize neutral, and bis(pyrazolyl)methane H 2 C(3,5-(CH 3 ) 2 Pz) 2 to isolate cationic, copper and silver acetylene complexes. A complete, comparative analysis of the bonding situation of these metalacetylene complexes using density functional theory (DFT) calculations is also presented.
Copper and silver complexes 9-15 are thermally stable solids at room temperature under an acetylene atmosphere. They can be handled, even in air, for brief periods (e.g., to prepare NMR samples) without signs of decomposition. Solid samples of 10-12 show some acetylene loss under nitrogen aer several hours (Table 1) but lose acetylene rapidly and completely under reduced pressure. They all however retain the intact scorpionate ligands even aer the acetylene loss, as evident from the NMR data. In fact, except in 10, the original acetylene complexes can be regenerated by exposing acetylene-free solids to C 2 H 2 gas in solution. Compound 10 forms a somewhat insoluble solid (presumably a polymeric material generated as observed with {[PhB(3-(CF 3 )Pz) 3 ]Ag} N ) 26 with the loss of C 2 H 2 , impeding the reverse, acetylene xing process. The tris(pyrazolyl)borate complexes 13, 14 and 5 are notably stable copper and silver acetylene complexes under a variety of conditions. The 1 H NMR data of 9 and 11-15 taken immediately aer preparing solutions in CDCl 3 show the expected products without signs of decomposition or C 2 H 2 loss (while compound 10 indicates some C 2 H 2 loss). Additional details on the stability of copper and silver acetylene complex pairs in the solid form and solution (CDCl 3 ) at room temperature are presented in Table 1 (and ESI †). Table 2 shows available, albeit limited, key 1 H and 13 C NMR data and C^C stretch of structurally characterized copper and silver complexes 1-8 and those of the newly synthesized complexes 9-15. A copper complex Cu 4 (m-[3,5-(CF 3 ) 2 Pz]) 4 (m-HC^CH) 2 containing a m 2 -h 2 ,h 2 -(HC^CH) (which is a bridging acetylene) 10a has also been included for comparisons. The 1 H NMR spectra of copper(I) complexes in general show a larger downeld shi (shi towards the typical alkene region) of acetylenic proton signal from the free acetylene resonance, whereas the silver analog shows only a smaller congruent shi. For example, the room temperature 1 H NMR spectrum of 9 in CDCl 3 exhibited the acetylenic proton resonance at d 4.22 ppm which is a signicant downeld shi relative to the corresponding signal of the free acetylene (d 1.91 ppm). Furthermore, the acetylenic protons of cationic 11 in (CD 3 ) 2 CO were observed at d 5.14 ppm. This also indicates that the acetylene on [{H 2 C(3,5-(CH 3 ) 2 Pz) 2 }Cu] + does not get displaced by acetone. The corresponding resonance of silver complexes 10 and 12 was observed at d 2.13 and 2.25 ppm, respectively, closer to the free acetylene peak position. The 13 C NMR resonances of the copper(I) and silver(I) coordinated acetylene carbons are interesting in the sense that they show shis in opposite directions from that of the free acetylene carbon signal (Table 2). For example, 13 C NMR resonance of the acetylenic carbons of free acetylene, copper complex 9 and silver complex 10 has been observed at d 72.0, 78.7, and 70.9 ppm, respectively. Note that there are other d-block metal-acetylene complexes with comparatively larger shis in acetylenic proton and carbon signals. 27 For example, (Ph 3 P) 2 Ni(HC^CH) 28 complex involving the signicantly better backbonding Ni(0) displays its proton and carbon signals for the nickel-bound h 2 -(HC^CH) in 1 H and 13 C NMR spectra at d 6.41, 122 ppm, respectively. The Raman and IR data of the h 2 -(HC^CH) copper(I) complexes show a reduction of C^C stretch by over >100 cm À1 with an average of reduction of 160 cm À1 relative to that of the free acetylene stretch observed at 1974 cm À1 . 29 This implies a weakening of the C^C bond due to s/p-interaction between copper(I) and acetylene (both components reduce the CC bond order) in terms of the Dewar-Chatt-Duncanson picture. 30 However, the reduction in wavenumber is not as high as that observed with Cu 4 (m-[3,5-(CF 3 ) 2 Pz]) 4 (m-HC^CH) 2 containing bridging acetylenes, which is understandable. Furthermore, ligand effects on n C^C are also apparent from some related complexes in which weakly donating ligand support on copper(I) produces molecules that display relatively higher HC^CH stretch, e.g., 9 vs. 4 or 14 vs. 13. Compared to Cu(I), the effect of Ag(I) on h 2 -(HC^CH) is relatively small as evident from a signicantly smaller reduction (average 60 cm À1 reduction from the corresponding stretch of the free C 2 H 2 ). This is   in agreement with silver(I) being a weaker s-bonding and p-backbonding metal ion compared to copper(I) atom (e.g., d 10 / d 10 s 1 electron affinities of Cu(I) and Ag(I) ions are 7.72 and 7.57 eV, in terms of energy released, respectively, and d 10 / d 9 p 1 promotional energies of Cu(I) and Ag(I) are 8.25 and 9.94 eV, respectively). 31 A much more detailed analysis of metal-acetylene bonding using DFT is also given below. Unfortunately, the background uorescence and acetylene loss prevented the observation of the n C^C band of some silver complexes reported in this manuscript.

X-ray crystal structures of copper and silver acetylene complexes supported by scorpionates
The copper and silver acetylene complexes, [Ph 2 B(3-(CF 3 )Pz) 2 ] Cu(C 2 H 2 ) (9) and [Ph 2 B(3-(CF 3 )Pz) 2 ]Ag(C 2 H 2 ) (10) afforded excellent single crystals and were characterized by X-ray crystallography. Fig. 3 depicts the molecular structures of these molecules. They are three-coordinate, trigonal planar metal complexes with k 2 -bound [Ph 2 B(3-(CF 3 )Pz) 2 ] À ligands. The acetylene ligand coordinates to the metal in a familiar h 2fashion. The M(NN) 2 B core (M ¼ Cu, Ag) adopts a boat conformation. These molecules feature a anking phenyl group above the metal-acetylene moiety with closest M/C(phenyl) separations of 3.01 and 2.88Å in the Cu and Ag complex, respectively. Although these atoms are within the Bondi's van der Waals separation distances of 3.10 and 3.42Å (or 4.15 and 4.30Å proposed byÁlvarez) 32 for Cu/C and Ag/C, 33 any interactions present between the metal and phenyl group do not affect the trigonal planar geometry at the metal (see also the computational section, below).
The molecular structures of the cationic acetylene complexes (12) involving a bis(pyrazolyl) methane ligand are illustrated in Fig. 4. The M(NN) 2 C core of the bis(pyrazolyl)methane ligand in 11 and 12 adopts a at boat conformation. The key difference between bis(pyrazolyl)borate and bis(pyrazolyl)methane backbone shapes is reected in the larger MN/NM separation of the pyrazolyl groups of the latter (see ESI Fig. S43 †).
As evident from the data presented in Table 3, Cu-N and Cu-C distances of 9 and 11 are shorter than the related separations involving silver in 10 and 12, which is expected as Ag is the largest metal of the coinage metal triad. 34 Consequently, the C-Cu-C and N-Cu-N angles are signicantly larger than those parameters involving silver. The acetylene ligands of 9 and 10 are essentially coplanar with the N-M-N plane (M ¼ Cu, Ag; silver complex shows the larger twist angle of 3 but it is still Table 3 Selected bond lengths and angles of three-coordinate copper and silver acetylene complexes and those of several related ethylene complexes for comparison. The CC distance of free acetylene is 1.20286(3)Å based on gas-phase experimental data 36 and 1.193(6)Å from neutron diffraction data on solid acetylene. 37 The CC bond distance (r(spec)) estimated from CC stretch is given in italics for metal acetylene complexes with n C^C data (see Table 2 and eqn. (1)). The CC distance of free ethylene for comparison is 1.3305(10)Å from gas phase data and 1.313Å from X-ray data 38 [ minor). This parallel orientation of NMN and CMC planes is the best for maximizing metal-(h 2 -ligand) backbonding interactions, rather than the orthogonal conformation. 35 However in contrast to 9 and 10, C-M-C and N-M-N planes of 11 and 12 deviate somewhat from co-planarity with the copper and silver adducts showing 8.8 and 11.9 inter-planar twist angles. Crystal packing diagrams indicate that one of the uorine atoms of [BF 4 ] À in 11 sits near Cu at 2.8842(12)Å, while two uorine atoms of two different [SbF 6 ] À counter-ions are closer to the silver center (at 3.364 (4) 2 ] À therefore appears to produce the same net result as the neutral and electron-rich H 2 C(3,5-(CH 3 ) 2 Pz) 2 on the bond distances and angles associated with copper(I) or silver(I).
The molecular structures of [HB(3-(CF 3 ),5-(Ph)Pz) 3 ]Cu(C 2 H 2 ) (14) and [HB(3-(CF 3 ),5-(Ph)Pz) 3 ]Ag(C 2 H 2 ) (15) have also been investigated using X-ray crystallography (see ESI †). Unfortunately, the M(h 2 -HC^CH) moieties of these molecules suffer signicant positional disorder and therefore are not suitable for the analysis of metrical parameters. We have also observed a similar disorder in copper and silver ethylene complexes of the analogous tripodal scorpionates. 23c,39 These molecules possess pockets that allow signicant degrees of free motion for the small p-ligands bonded to the metal sites. Nevertheless, basic structural features and atom connectivities of 14 and 15 are clear and indicate the presence of h 2 -(HC^CH) moieties, consistent with the spectroscopic data, and tetrahedral metal sites.
As noted above, X-ray crystallographic data on a limited number of copper(I) and silver(I) acetylene complexes are available for comparison. The C^C bond distance of those compounds and the ve uncovered in this work (Table 3 and ESI  Table S19

227(4)Å in Cu 2 (m-[4-Br-3,5-(CF 3 ) 2 Pz]) 2 (C 2 H 2 ) 2 (3). 10a
The C^C bond distance of Cu(I) and Ag(I) bound acetylene complexes is expected to be longer than that of the free acetylene (which is 1.20286(3)Å based on gas-phase experimental data 36 and 1.193(6)Å from neutron diffraction data on solid acetylene), 37 as both the s-donation and p-backdonation interactions between the metal and acetylene causes a reduction in CC bond-order and a lengthening of the C^C bond distance relative to that of the free acetylene. The Raman and IR data (Table 2) also support this expectation. However, most of the metal-bound C^C bond distances of these silver and copper acetylene complexes resulting from X-ray crystallographic studies (Table 3 and ESI Table S19 †) are lower than that of the free ligand. As Krossing, Scherer and co-workers have pointed out, this apparent contradiction is a result of systematic errors associated with the measurement. 20 In small molecules such as acetylene involving multiple covalent bonds between light atoms, libration effects, incomplete deconvolution of thermal smearing and anisotropy of the electron density tend to produce bond distances that are too short from standard X-ray models. 20,40 Such effects can be minimized by collecting data closer to absolute zero temperature and to very high angles (e.g., 2q ¼ 100 ). 41 For example, the collection and analysis of the very high-resolution X-ray diffraction data of [Al(OC(CH 3 )(CF 3 ) 2 ) 4 ] Ag(C 2 H 2 ) (8) (to resolution d ¼ 0.476Å) at 10 K has produced a C^C bond distance of 1.209(1)Å, 20 which is in good agreement with the theoretical model (1.213Å), while the same molecule at d ¼ 0.84Å and 90 K resulted in a length that is 0.063 A shorter at 1.146(4)Å. This also shows the relative impact of core and valence electrons on the X-ray scattering factors (i.e., scattered X-rays at higher angles are relatively less affected by the valence electrons, and therefore produce more precise nuclear or core-electron positions). 42 Although X-ray crystal structures of 9-13 reported here do not reach the resolution level of the specialized work noted above for 8, they are quite respectable (d (resolution) of 0.73 to 0.60Å at 100 K) for standard X-ray crystallography. Indeed, the analysis of the data of 9-13 at lower resolution levels (e.g., using d ¼ 0.84Å, 2q ¼ 50 cutoff) produced relatively shorter C^C bond distances (see ESI, Table S20 †). Minor libration effects are also evident even at 100 K based on the TLS analysis (see ESI †). 43 Overall, due to a combination of factors noted above, acetylene C^C bond distances based solely on routine X-ray crystallography are not suitable for discussions of metal-ligand bonding in most Cu(I) and Ag(I) complexes, and to parse out the metal and supporting ligand effects on the acetylene moiety. Furthermore, some of the C^C bond distance changes as a result of Cu(I) and especially Ag(I) ion coordination are also expected to be small. They are oen overshadowed by the relatively high estimated standard deviations (esds) associated with the measurement and are not signicantly different at the 3s limit of estimated standard deviations. Similar issues have been noted also with ethylene complexes, particularly those involving silver(I). 41,44 It is Table 4 Results of the EDA-NOCV calculations (ZORA-BP86-D3/TZ2P//RI-BP86-D3/def2-TZVPP level, in kcal mol À1 ) on Cu(I)-and Ag(I)-(C 2 H 2 ) complexes using LM and C 2 H 2 as fragments (L ¼ supporting ligand) however, possible to estimate the C^C bond distances of the metal complexes utilizing changes in C^C vibration. As noted below in the computational section, this technique produces a more realistic estimate of the C^C bond distance for copper and silver acetylene complexes. Table 3 also includes structural data on a select group of Cu(I) and Ag(I) h 2 -ethylene complexes. With the availability of the analogous acetylene complexes, it is now possible to make a meaningful comparison between the two families. As expected, and despite the issues noted above with CC bond distances based on routine crystallography, the metal-bound acetylene bond distances are signicantly shorter than the related ethylene bond lengths. The Cu-C and Cu-N bond distances are also shorter in the copper(I) acetylene complexes compared to their ethylene analogs. Interestingly, however, Ag-N and Ag-C distances are essentially the same in the two families. It would be interesting to see if this difference holds true also for a larger dataset.

Computational analysis of the copper and silver acetylene complexes
Density Functional Theory (DFT) calculations at the relativistic ZORA-BP86-D3/TZ2P//RI-BP86-D3/def2-TZVPP level (see computational details in the ESI †) were carried out to understand the chemical bonding between the scorpionate-M moieties and acetylene in the above-described LM-(C 2 H 2 ) complexes (L ¼ supporting ligand; M ¼ Cu, Ag). To this end, the combination of the Energy Decomposition Analysis (EDA) and the Natural Orbitals for Chemical Valence (NOCV) methods were applied to gain a detailed quantitative insight into the interaction between the LM and C 2 H 2 fragments. From the data in Table 4, it becomes clear that in all cases the main contribution to the interaction between the LM and C 2 H 2 fragments comes from the electrostatic attractions (measured by the DE elstat term), which represents ca. 60% of the total attractive contribution. This indicates that the nature of the LM-acetylene bond is markedly ionic. Despite that, the orbital interactions (measured by the DE orb term) are also signicant as they contribute ca. 35-40% to the total interaction energy. At variance, the interactions coming from dispersion forces are much less important in the description of the bonding (<5%) and can be considered negligible.
The NOCV extension of the EDA method allows us to not only identify but also quantify the main orbital interactions contributing to the total DE orb term. According to the NOCV method, two main donor-acceptor orbital interactions dominate the orbital interactions in these acetylene complexes. On one hand, the s-donation from the doubly-occupied p(C^C) molecular orbital of the acetylene ligand to the empty s atomic orbital of the transition metal (denoted as DE orb (1)) and, on the other hand, the backdonation from a doubly-occupied d atomic orbital of the transition metal to the vacant p*(C^C) molecular orbital of acetylene (denoted as DE orb (2), see Fig. 6 for complex 9). Interestingly, our NOCV calculations indicate that, in all cases, the backdonation from the transition metal fragment is signicantly stronger (ca. twice as strong) than the donation from the acetylene ligand (DE orb (2) > DE orb (1)), regardless of the transition metal and the supporting ligand. In addition, our EDA-NOCV calculations conrm that both orbital interactions are stronger (in particular, the LM/ p*(C^C) backdonation) in the Cu(I)-complexes as compared to their Ag(I)-analogues, which is in agreement with the above-commented weaker sbonding and p-backbonding ability of Ag(I) as compared to copper(I). 31 Despite that, the bonding situation in these acetylene complexes can be safely described in terms of the Dewar-Chatt-Duncanson model involving two donor-acceptor interactions (s-donation from the acetylene ligand and p-backdonation from the transition metal fragment). Note that the acetylene p/p*-orbitals perpendicular to the MC2 plane form only relatively weaker interactions with the transition metal fragment in these scorpionate ligand supported copper and silver complexes.
Interesting trends emerge from a closer inspection of the data gathered in Table 4. First, when comparing the copper complexes with their silver counterparts, it is found that, regardless of the supporting ligand, the interaction between the transition metal fragment and the acetylene ligand is clearly stronger in the corresponding copper complexes (DE int $ 20 kcal mol À1 ). This is consistent with above-commented higher NMR-downeld shis (or redshis of the C^C stretch), with respect to free acetylene, observed experimentally for the copper complexes. According to the data in Table 4, the enhanced interaction in the copper(I) complexes is the result of an enhancement of all the main attractive interactions (DE elstat , DE orb (1) and DE orb (2)) as compared to the corresponding silver(I) complexes. This nding suggests that the observed experimental shis of these mono-acetylene complexes are closely related to the computed interaction energies (as well as their main energy contributors). To our delight, we found that indeed good linear correlations are obtained when plotting these experimental values versus not only the computed total interaction energies (DE int ) but also their main EDA-NOCV contributors (see Fig. 7 for the linear relationships involving the 13 C-NMR shis, Dd). From the data in Fig. 7, there appears to exist a limit dening the observed shi in the 13 C-NMR spectra with respect to free acetylene: while complexes having a LM-(C 2 H 2 ) interaction DE int $ À40 kcal mol À1 lead to a positive (i.e., downeld) shi with respect to free acetylene (Dd > 0 ppm), complexes exhibiting lower LM-(C 2 H 2 ) interaction energies provoke the opposite (i.e., upeld shi) effect (Dd < 0 ppm).
Data in Table 4 also indicates that the nature of the supporting ligand also affects the LMÀ(C 2 H 2 ) interaction. Regardless of the involved transition metal, it is found that complexes having a bidentate bis(pyrazolyl)borate/methane supporting ligand (complexes 4, 4-Ag, 9, 10, 11+, 12+) exhibit stronger LM-(C 2 H 2 ) interactions than the analogous systems having a tridentate tris(pyrazolyl)borate ligand (complexes 5, 13, 14, 15). For instance, when comparing bidentate complexes 4 or 4-Ag with their tridentate counterparts 13 or 5, it becomes clear that the weaker interaction computed for the latter complexes nds its origin in the lower electrostatic and orbital (mainly the LM / p*(C^C) backdonation, DE orb (2)) interactions computed for these species. Therefore, it can be concluded that supporting ligands having a lower number of donor sites lead to stronger LM-(C 2 H 2 ) interactions. This is also supported by the calculations on the naked [Cu(C 2 H 2 )] + and [Ag(C 2 H 2 )] + cations, which exhibit the highest DE int values of their corresponding series (see Table 4). Furthermore, it is found that the replacement of bidentate uorinated bis(pyrazolyl)borate ligand (which is anionic) by the analogous non-uorinated and neutral bis(pyrazolyl)methane ligand provokes an almost negligible effect on the LM-(C 2 H 2 ) interaction (e.g., compare 4 and 11+ or 4-Ag and 12+). Table 5 Computed C^C bond lengths and corresponding stretching frequencies in the Cu(I), Ag(I) and Au(I)-scorpionate complexes together with representative group 10 analogues and including group 1 complexes 4-Li and 4-Na. All data have been computed at the RI-BP86/def2-TZVPP level. For comparisons, the computed CC distance of free acetylene is 1. 207Å   We have also examined the C^C bond distances and the C^C stretching frequencies of the copper(I) and silver(I) complexes, computationally. The expected changes to the CC distance are especially useful considering the challenges associated with measuring this parameter precisely noted above. Table 5 shows the computed C^C distances and the corresponding stretching frequencies for the considered Cu(I) and Ag(I)-complexes together with their Au(I)-counterparts and representative group 1 and group 10 analogues.
From the data in Table 5, it becomes evident that, in all cases, the Cu(I)-complexes exhibit longer C^C distances than their corresponding Ag(I)-analogues, which is translated into a higher redshi of the n C^C stretching frequency. This is therefore fully consistent with the experimental ndings and with the higher LM-(C 2 H 2 ) interaction energies computed for the Cu(I)-complexes as compared to their Ag(I)-congeners (see above). This effect is even higher in the corresponding Au(I)complexes which exhibit the longest C^C distances in the entire group 11 series. Not surprisingly, even longer distances (associated with higher redshis, i.e., larger negative D n C^C values) are found when considering the neutral group 10 transition metal as a consequence of a signicantly stronger pbackdonation. In contrast, the analogous group 1 complexes, where the backbonding is minimal, present values rather similar to free acetylene. In addition, data presented in Table 5 show that uorinated substituents on the supporting ligand L have a noticeable effect on n C^C (see for example 4 vs. 4'; 4-Ag vs. 4'-Ag). For this reason, it is not surprising that an excellent correlation was found when plotting the difference in the computed C^C distances vs. the shi in the n C^C stretching mode with respect to free acetylene (correlation coefficient of 0.999, see Fig. 8), including D n C^C ¼ 0 and Dr C^C ¼ 0 for free acetylene. The computed relationship presented in Fig. 8 can be then used to estimate the C^C distances in the real systems (r(spec), where spec ¼ spectroscopic, eqn (1)) and check the reliability of the X-ray derived data by simply adding the experimental C^C distance in acetylene to the calculated distance change (Dr(calc)) using the equation in Fig. 8 and the experimental D n C^C value (Excel le to compute r(spec) from D n C^C is provided in ESI †). A similar method has been utilized successfully by Krossing and co-workers 41 to estimate the C]C bond distances (i.e., to obtain spectroscopically assessed bond distances, r(spec)), of silver ethylene complexes. r(spec) ¼ 1.20286 + Dr(calc) (1) As an example, the estimated C^C bond distances (r(spec)) of 9 and 13 based on the experimental Raman data ( Table 2, D n C^C ¼ À167 and À129 cm À1 , respectively) are 1.236 and 1.228 A, respectively. They are longer than that of the free acetylene, which is more reasonable and expected based on vibrational and computational data. The eqn (1) can also be used to estimate the C^C bond distance of 8 (i.e., using experimentally observed D n C^C ¼ À60 cm À1 to afford r(spec) ¼ 1.214Å), which is very close to the experimental X-ray model value of 1.209(1)Å based on high-resolution data.
Finally, we were curious to analyze the nature of the weak yet noticeable interaction between one of the phenyl groups attached to the boron atom and the transition metal in complexes 9 and 10 (see above). The NCIPLOT 45 method clearly conrms the occurrence of a signicant noncovalent attractive interaction (greenish surface in Fig. 9) between this aryl group and the transition metal. According to the Natural Orbital Bond (NBO) 46 method, this stabilizing noncovalent interaction nds its origin in the donation of electron density from the closest p(C]C) molecular orbital of the phenyl group to the vacant s atomic orbital of the transition metal (associated stabilizing energy, DE(2) ¼ À1.2 and À1.1 kcal mol À1 , for complexes 9 and 10, respectively).

Summary and conclusions
In summary, we have presented the isolation and complete characterization of several new acetylene adducts of Cu(I) and Ag(I) supported by [Ph 2 B(3-(CF 3 )Pz) 2 ] À , [HB(3,5-(CF 3 ) 2 Pz) 3 ] À , [HB(3-(CF 3 ),5-(Ph)Pz) 3 ] À , and bis(pyrazolyl)methane H 2 C(3,5-(CH 3 ) 2 Pz) 2 , as well as details on their 1 H, 13 C, and 19 F NMR spectroscopy, Raman spectroscopy, and X-ray crystallography. According to our DFT calculations, the bonding situation in these complexes can be described in terms of the traditional Dewar-Chatt-Duncanson model involving two donor-acceptor interactions, namely s-donation from the acetylene ligand to the transition metal and p-backdonation from the transition metal fragment to the p*(C^C) molecular orbital of acetylene (the latter being markedly stronger than the former). Interestingly, the copper complexes exhibit a downeld shi for acetylenic carbons in their 13 C NMR spectra and a more notable reduction in n C^C relative to the free acetylene. This can be ascribed to a stronger interaction between the transition metal fragment and the acetylene ligand in the Cu(I)-complexes than that in their Ag(I)-counterparts as conrmed by our EDA-NOCV calculations (DE int $ 20 kcal mol À1 ). Furthermore, it is found that while the replacement of bidentate uorinated Fig. 9 Contour plots of the reduced density gradient isosurfaces (density cutoff of 0.04 a.u.) for complex 9. The green surfaces indicate attractive noncovalent interactions. bis(pyrazolyl)borate ligand by the analogous non-uorinated bis(pyrazolyl)methane ligand on M(I) ions provokes an almost negligible effect on the LM-(C 2 H 2 ) interaction, the related tridentate tris(pyrazolyl)borate supporting ligand weakens the LM-(C 2 H 2 ) interaction, which is reected into less signicant NMR/Raman shis. The C^C distance of these copper and silver acetylene complexes resulting from routine X-ray models suffers due to incomplete deconvolution of thermal smearing and anisotropy of the electron density in acetylene and is shorter than expected. Although it is possible to minimize this issue by collecting X-ray data at near absolute zero and to very high angles, it is not practical for routine work. However, the experimentally observed n C^C values can be utilized to provide C^C bond distances of Cu(I) and Ag(I) complexes that are more realistic. Molecules presented herein represent the largest collection of isolable copper(I) and silver(I) complexes featuring the terminal, h 2 -HC^CH ligand. We believe that the contents of the present work contribute signicantly to the development of acetylene chemistry.

Data availability
The details and data supporting this article have been uploaded as the ESI. † Crystallographic data can be obtained from the CCDC.

Author contributions
Project administration and conceptualization: HVRD; Writingoriginal dra and funding: HVRD and IF; Synthesis and spectroscopic characterizations: AN-P and SGR; X-ray crystallography: HVRD; Computational work: IF; Data analysis, review and editing: all authors.

Conflicts of interest
There are no conicts to declare.