10-Valence-Electron C≡O and the 14-VE C≡Pt: Two Triple-Bonded Isoelectronic Families Differing by a dδ4 Ring Authors: Michiko Atsumi, Pekka Pyykkö Published: 2023 ABSTRACT 10-VE A≡A′ Diatomics, such as N≡N, C≡O, etc., have a strong σ2π4 triple bond plus a lone pair at each end. In our studies on 14-VE A≡B systems, such as C≡Pt, we find a similar bonding system plus a (5dδ)4 ring. Here, the A atom belongs to groups 13–17 and the B atom to groups 7–11. Also the BB′ combinations, triatomics, such as PtCO or DsCO or uranyl, and longer chains, such as AuCN and [NC-Au-CN]−, are discussed. The δ ring directly contributes to nuclear quadrupole coupling constants, and DFT calculations using the BH and H or mPW1K functionals reproduce the experimental trends of the NQCC. Introduction While developing a set of additive covalent triple-bond radii, r3,1,2 the same covalent radii r(A) were used in fitting the main-group A≡A′ and the A≡B bond lengths, where B is a transition metal. The primary bond lengths came from experimental or computed closed-shell systems with main-group elements and with transition metals as long as the molecular-orbital inspection and the actual bond length fitted the picture. The same covalent radii r(A) were used in fitting the main-group A≡A′ and the A≡B bond lengths for both single, double, and triple bonds, r1, r2, and r3, respectively. In our ongoing calculations, we kept finding two different sets of diatomic species. One was the well-known 10VE family of diatomic systems, such as N2, CO, CN–, NO+, etc. These are known to have a nominal σ2π4 triple bond combined with a lone electron pair at each end. In all, we then have 10 valence electrons (10VE)3 (it should be noticed that, as an alternative to this triple-bond picture, in older literature, a C=O double bond was thought to exist, and also other resonance hybrids have been mentioned, see Pauling4 or Long and Walsh5). The other families, such as CPt, X1Σ had four more electrons, always occupying a δ4 ring, typically the penultimate, HOMO – 1. With this, we mean the occupied (dxy)2(dx2–y2)2 combination, in an alternative form , or (dδ)4. The novelty is that we have an occupied orbital which is formally a nonbonding core orbital but has the size and orbital energy of a typical valence orbital. That observation is not new, and some previous examples are shown in Table 1. We are, however, not aware of any previous articles concentrating on this feature. Table 1Examples on δ4 Systemsyearspeciesref1998CAu+Barysz,8 Pyykkö91999NIrRam102003SiPtBarysz112004CPtPatzschke122015SiNiSchoendorff131985[XAuX]−Bowmaker141985[NC-Au-CN]−Bowmaker14 Examples on δ4 Systems Starting from these 10-VE or 14-VE families of species, we made some further contacts:1It also is possible to have the δ4 ring without any triple bonds to the metal.2In multiple bonds between two B atoms, bond orders beyond 3 can be reached, see Cotton et al.63One way to “see” the δ4 electrons is to measure nuclear quadrupole coupling constants. While many DFT functionals have problems in reproducing the electric field gradient (EFG), q, at transition-metals,7 we have now found two functionals that work. We also report some results for species, not involving the δ4. 1It also is possible to have the δ4 ring without any triple bonds to the metal. It also is possible to have the δ4 ring without any triple bonds to the metal. 2In multiple bonds between two B atoms, bond orders beyond 3 can be reached, see Cotton et al.6 In multiple bonds between two B atoms, bond orders beyond 3 can be reached, see Cotton et al.6 3One way to “see” the δ4 electrons is to measure nuclear quadrupole coupling constants. While many DFT functionals have problems in reproducing the electric field gradient (EFG), q, at transition-metals,7 we have now found two functionals that work. We also report some results for species, not involving the δ4. One way to “see” the δ4 electrons is to measure nuclear quadrupole coupling constants. While many DFT functionals have problems in reproducing the electric field gradient (EFG), q, at transition-metals,7 we have now found two functionals that work. We also report some results for species, not involving the δ4. Methods We use in the present calculations DFT and ZORA at either the scalar relativistic (SR) or spin–orbit (SO) level. This will provide reasonable accuracy and easy interpretation of chemical bonding questions. The calculations were carried out by using the ADF software package15 with the Perdew–Burke–Ernzerhof (PBE) functional during the structural calculations. The triple-ζ basis sets with two polarization functions (TZ2P) were used for all elements, treated at the nonfrozen core level. We have added into Tables 2, 3, 4, 5, and 6 the results of some earlier calculations and a comparison with the suma of our triple-bond covalent radii r3.1,2 Table 2Calculated and Experimental Bond Lengths, Re, for AB Diatomicsa Re/pm ABPWexpcalcr3δBRhd170.36b169.2168.516179HBPt–181b 179.901183H-1 180.9c BAu192.1b 192.271196H-2 191.6c 190.617 H-1CNi160.5b162.7318159.601161H-1CRud160.9b160.5485(2)19160.119163H 160.4220 CPd172.2b172.2171.61172H-1 172.3c CPt168.3c167.721172.512170H-1 167.8b 167.451 H-1CAu+178.9b 176.68,9183H-2 178.8c CDs172.2b 172.712172 172.6c NIr160.8c160.68281(35)10160.910161H-1SiNi201.2b203.165(24)22208.613203HSiPt208.2b206.29(2)21011212H-1 208.7c SiAu+219.3b 218.88225H-2 219.4c PIr198.3b199.28e,23 201H 199.9c TlAu269b 273H-1 266.6c BiP229.5b229.61520(80)24 229 230.2c BiAu258.6b 258H-1 263c 258H-2,H-3aδ gives the energetic placement of the dδ4 MO with respect to the HOMO (H). “PW”: Present work. r3: From triple-bond covalent radii.1,2bSR.cSO.d12VE species.eR0. Calculated and Experimental Bond Lengths, Re, for AB Diatomicsa δ gives the energetic placement of the dδ4 MO with respect to the HOMO (H). “PW”: Present work. r3: From triple-bond covalent radii.1,2 SR. SO. 12VE species. R0. Table 3Calculated and Experimental Bond Lengths, Re for BB′ Diatomicsa Re/pm BB′PWexpcalcr3δCr2159.5b167.88156.228206HCrU185.4b 188.327221H 187.6c MoU202.8b 202.127231H 205.9c WU208.5b 208.027233H 217.5c ThU248.2b 254H 279.3c HThPt237.3b2542925011246H 237.7c ThAu+271.6b 2631259H-1 262.4c aδ gives the energetic placement of the dδ4 MO with respect to the HOMO (H). PW: Present work. r3: From triple-bond covalent radii.bSR.cSO. Calculated and Experimental Bond Lengths, Re for BB′ Diatomicsa δ gives the energetic placement of the dδ4 MO with respect to the HOMO (H). PW: Present work. r3: From triple-bond covalent radii. SR. SO. Table 4Calculated and Experimental Bond Lengths, Re, for Linear Polyatomic Speciesa Re/pm speciesbondPWexpcalcr3δAuCNcAu–C190.6191.5145 183H-2 C–N116.9115.5645 114 [NCAuCN]−1bAu–C198.8198.4d198.843183H-2 19946 [HAuCN]−1Au–C207.4b 202.547183H-2 202.1c [FAuF]−Au–F199.9b 196.343176H-2 199.3c [ClAuCl]−Au–Cl230.2b228228.343216H-2 229.6c 228.348 [BrAuBr]−Au–Br242.2b240239.443233H-2 241.5c 240.848 [IAuI]−Au–I259.6b253256.143248H-2 259.2c 256.948 [AtAuAt]−Au–At268.6b 263.943261H-3 270.2c PtCOPt–C176.1b176.04649177.630170HPtCPtPt–C173.8c 173.530170H-2[AuCAu]2+Au–C182.6b 173.530183H-2 182.4c NUIrN–U173.9b 172.232172H-1 179.8c U–Ir217b 218.432225H-1 222.5c OUPt2+O–U172.7b 171.450171H-1 174.3c U–Pt228.6b 221.450228 229c aThe [XAuX]− experimental values for X = Cl–I are taken from.44 In the last column, the δ gives the energetic placement of the dδ4 MO with respect to the HOMO (H). Slight variations of the order may occur as a function of the method. PW: Present work. r3: From triple-bond covalent radii. “Calc.” refers to earlier calculations.bSR.cSO.dIn Nd[Au(CN)2]3·3H2O.51 Calculated and Experimental Bond Lengths, Re, for Linear Polyatomic Speciesa The [XAuX]− experimental values for X = Cl–I are taken from.44 In the last column, the δ gives the energetic placement of the dδ4 MO with respect to the HOMO (H). Slight variations of the order may occur as a function of the method. PW: Present work. r3: From triple-bond covalent radii. “Calc.” refers to earlier calculations. SR. SO. In Nd[Au(CN)2]3·3H2O.51 Table 5Vibrational Frequencies, ωe (cm–1), for Diatomics ωe/pmspeciesPWexpcalcBRhc939b BPt–844a 842b BAu658a 70417,d71017BAu663b CNi959875.15518 CRuc1156b110019 CPd885a 881b CPt1059b1051.13(2)21109612CAu+827a 8738 826b CDs1115b 114712NIr1166b1126.1757(28)10116110 1195(38)54,e SiNi499a467.432245813 497b SiPt537a549.0(3)53111 527b SiAu+427.2a 4598 426.6b PIr624a57023 591b TlAu135a 139b BiAu168a157.755 ThPt248a2212921111 245b ThAu+143a 18811aSR.bSO.c12VE species.dr0.eIn noble-gas matrices. Vibrational Frequencies, ωe (cm–1), for Diatomics SR. SO. 12VE species. r0. In noble-gas matrices. Table 6Vibrational Frequencies for Polyatomic Speciesb,c ωe/cm–1speciesmodePWexpcalcAuCNν1(σ)2161 2276,45 218156 ν2(π)33532045313″, 28556 ν3(σ)47448045476″, 47256[Au(CN)2–]C–Nσu21302158a,51 C–N σg2147216651 Au–C σu403474, 49151 Au–C πu420 Au–C σg427 πg290 πu80 AuCl2–σg300.63295731043 σu320.935057330″ πu110.511657103″AuBr2–σg19320957194″ σu23825457238″ πu74.6775769″AuI2–σg14215857143″ σu20021057194″ πu55.8635752″aIn solid Nd[Au(CN)2]3·3H2O.bSR.c12VE species. Vibrational Frequencies for Polyatomic Speciesb,c In solid Nd[Au(CN)2]3·3H2O. SR. 12VE species. Results Diatomics Table 2 gives the bond lengths of the A≡B diatomics, specified in Figure 1, mostly for 14VE species. Schematic placement of the main-group elements (A) and transition elements (B). The 10VE A≡A′ and 14VE A≡B combinations are indicated. l.p. = lone pairs. These AB Re values in Table 2 are close to the sum of triple-bond covalent radii. To the contrary, the BB′ ones in Table 3 can be much shorter, reflecting a bond order higher than three. Quadruple bonding in ground or excited states of certain AB diatomics20,25 and BB systems6 has also been discussed. Liu et al.26 consider the donation and its direction in certain multiply bonded AA′ diatomics. Cheung et al.16 also discuss 4-fold bonding in RhB. The MU species (M = Cr–W) of Ruipérez et al.27 have the δ ring as the HOMO, see Table 3. Note that at the scalar relativistic level, these are 12-VE species with a (1π)4(1σ)2(2σ)2(1δ)4X1Σ ground state, see ref (27)Figure 1. Polyatomic Chains If group 11 (Cu, Ag, Au, and Rg) doubles as a halogen and group 10 (Ni, Pd, Pt, and Ds) as a chalcogen, it is easy to take the step from OCO to PtCO or DsCO.12 With metals at both ends, we have [Au=C=Au]2+, or the analogues CPt, CPt2, and CPt32– to CO, CO2, and CO32–, respectively.30 Of the species in Table 4, the triatomic PtCPt and [AuCAu]2+ have two δ4 rings, both a g and a u one, separated from each other by the carbon atom. Similarly, the uranyl isoelectronic series31 would yield NUN and the NUIr of Gagliardi and Pyykkö.32 The NUO+ predicted by Pyykkö et al.31 was later made by Heinemann and Schwarz33 in the gas phase and by Zhou and Andrews34 in Ne matrices. The OUIr+ was prepared by Santos et al.35 For reviews on uranyl analog complexes, see Wei et al.,36 or Maria and Marçalo.37 A closer analysis of U≡A multiple bonding was given by Motta and Autschbach.38 With 5d metals at both ends, linear species like PtThIr– were found by Hrobárik et al.39 This species had a δ4 ring at the Pt end, and some evidence for Th–Ir δ bonding at the Ir end. Two of the σ bonds were characterized as “a sausage inside a tube”. The isoelectronic series of [ClAuCl]− can be continued at least to the high-pressure compound Li5AuP2.40 The energetic location of the δ4 ring in the [XAuX]− series, X = F–At, is discussed in the last column of Table 4. Pauling41 considered a δ bond in Re2Cl82–, an assigned quadruple-bond case. For further examples, see the book.6 He also found two triple U≡O bonds in uranyl. Possible members of the uranyl isoelectronic series down to [CUC]2– have been discussed.31 Most of the World’s gold production42 is based on the [NC-Au-CN]− ion, which has a beautiful δ ring, as seen from Figure 2. This δ energy level occurred as H-2 also in earlier calculations.43 Highest occupied MO/s of a free [NC-Au-CN]− anion. The δ4 ring is the HOMO – 2 one at −3.460 eV. Other Properties Relation of Bond Lengths and the Triple-Bond Covalent Radii Both A≡A′ and A≡B triple bonds with various principal quantum numbers, n, were used to originally fit the additive r3. As seen from Tables 2 and 3, the present R mostly agree with1for triple bonds. The higher bond-orders in Table 3 may be substantially shorter. Clear trends as a function of the group and the row are seen in Figure 3. For the shortness of the n = 2 radii, see Wang et al.52 Note for the n-dependence of the main-group elements A the clear trend 2 ≪ 3 < 4 < 5 < 6, while the transition metals, B, in the groups 9–10 rather have 6 < 5 ≈ 4. The sixth-row elements near gold have a local maximum of the relativistic bond-length contraction.53 Additive triple-bond covalent radii, r3 (in pm), for transition-metal atoms (B) and main-group elements, (A) belonging to the groups 3–11 and 13–17, respectively. n is the number of the row. Vibrational frequencies are another direct connection to experiments. As examples, we can take Table 5 on diatomics or Table 6 on polyatomic chains. Dissociation energies and predissociation of species like CuB, AuB, or AlB are discussed by Merriles and Morse.58 The δ MO is quoted in their Supporting Information. Nuclear quadrupole coupling constants, NQCCs, give a direct access to the electric field gradient, q at the nuclei involved. The four electrons in the δ4 ring make a major contribution. The RuC has been mentioned as an example by Wang et al.,19 who estimate that +1350 MHz of the experimental B0 of +433 MHz of 101RuC come from the δ4 ring. Without this contribution, even the sign would be wrong. Gusmão et al.59 obtained similar q values for RuC.b For nuclear quadrupole coupling constants, BHandH and mPW1K functionals were applied. The BHandH gave earlier good q for HCl and CuCl,60 CdMe2,61 and Cd(SMe)2.62 In early multiple-scattering Xα calculations, Bowmaker et al.14 gave an orbital-based analysis of their q and found for [Au(CN)2–] that about +15.805 of the total q of −9.650 au arise from the H-4 δ4 electrons. They emphasized in their scalar-relativistic discussion of the AuX2– species the q contributions from the Au 6pz orbital. In the linear HAuH– both the Au 5dπ and 5dδ would be “inert” and only the Au 6s and 5dσ participate in bonding, as seen from Figure 4. The AuH2– anion has been seen in matrix spectroscopy63 and photoelectron spectroscopy.64 Highest occupied orbitals of the AuH2– anion. The result resembles Figure 6 of Liu et al.64 In actual fact, one must emphasize that the XAuX– calculations show some unusually strong dependence on the chosen functional.57 Moreover, the Mössbauer measurements are performed on solid samples with so far unknown matrix effects. The gas-phase measurements included in Tables 7 and 8 have a chance to provide cleaner comparisons. The gas-phase microwave quadrupole coupling constants of Okabayashi et al.65 for triatomic AuCN were in complete disagreement with our earlier calculations, while their structural and vibrational AuCN parameters are in adequate agreement both with our more approximate work, and the latest and best studies.56,66 Then, we found that the alternative functionals BHandH and mPW1K gave a semiquantitative agreement between the experimental and calculated B, as seen from Figure 5. Table 7Calculated (SO) and Experimental NQCC, Be = e2qQ/h (MHz), at the SO Level, Assuming the Q Values in Table 8 Be/MHzspeciesNuclPWexpcalcBRh11B–1.29b BPt–11B–3.47 BAu11B–4.41 197Au260.0 CNi61Ni89.2 CRuc101Ru544.7433.19(8)19476.24c,19CPd105Pd370.5 CAu+a197Au322.9 NIr14N–4.69 103Ir1951.71721c,71 SiNi61Ni78.0 SiAu+197Au219.5 PIr103Ir1484.21424c,71 BiN209Bi1101.3905.066(88)24 14N–2.91–2.468(13)24 BiP209Bi1110.0903.03124 TlAu197Au93.5 BiAu209Bi–237.4 197Au151.3 ThPt229Th–3759.1 CrU53Cr–1.59 235U1547.3 MoU95Mo6.87 235U291.4 WU235U6823.2 AuH197Au164.3187.116(99)72 AuD197Au164.3188.119(33)72 AuF197Au–72.6–52.2344(67)73 AuCl197Au2.619.63312(13)74 35Cl–60.1–61.99694(81)74 AuBr197Au25.937.2669(14)74 79Br463.8492.3271(12)74 AuI197Au58.578.273(11)75 127I–1703.9–1707.881(25)75 aAssumed Q/mb from Pyykkö,76 unless otherwise stated: 11B 40.59(10), 14N 20.44(3), 53Cr 150(50), 61Ni 162 mb, 101Ru 457(23), 100Rh 153, 105Pd 660(11), 193Ir 751(9), 197Au 547(16). 209Bi 422(3)77 and references there. 229Th 3110(60), 235U 4936(6).b12VE species.cAt v = 0. Calculated (SO) and Experimental NQCC, Be = e2qQ/h (MHz), at the SO Level, Assuming the Q Values in Table 8 Assumed Q/mb from Pyykkö,76 unless otherwise stated: 11B 40.59(10), 14N 20.44(3), 53Cr 150(50), 61Ni 162 mb, 101Ru 457(23), 100Rh 153, 105Pd 660(11), 193Ir 751(9), 197Au 547(16). 209Bi 422(3)77 and references there. 229Th 3110(60), 235U 4936(6). 12VE species. At v = 0. Table 8Calculated (SO) and Experimental NQCC, Be = e2qQ/h (MHz), at the SO Level for Polyatomic Species, Assuming the Q Valuesa Be/MHzspeciesNuclPWexpcalcAuCN197Au–23.2–0.7246(46)65 14N–5.01–4.1522(17)65 AuCl2–197Au–648.3(−)765b, (−)802–862.514 –72114 (−)78576 35Cl–35.6(−)35.214,79 AuBr2–197Au–620.3(−)790(50), (−)79478–69214 79Br248.9 20214AuI2–197Au–590.4(−)718, (−)71978–70914Au(CN)2–197Au–1325.9(−)1263–111814 (−)120678 AuAt2–197Au–488.2 AuD2–197Au–1347.3 HAuCN–197Au–1348 14N–4.80 NUIr14N–1.30 235U–4308.6 193Ir1139.1 OUPt2+17O0.84 235U–1199.1 aAssumed Q/mb from Pyykkö,78 unless otherwise stated: 11B 40.59(10), 14N 20.44(3), 53Cr 150(50), 61Ni 162 mb, 101Ru 457(23), 100Rh 153, 105Pd 660, 193Ir 751(9), 197Au 547(16). 209Bi 422(3)77 and references there. 229Th 3110(60), 235U 4936(6). The Au NQR frequencies of Bowmaker14 have been scaled by a Q-ratio of (547 mb/590 mb) = 0.927.bSee80. Calculated (SO) and Experimental NQCC, Be = e2qQ/h (MHz), at the SO Level for Polyatomic Species, Assuming the Q Valuesa Assumed Q/mb from Pyykkö,78 unless otherwise stated: 11B 40.59(10), 14N 20.44(3), 53Cr 150(50), 61Ni 162 mb, 101Ru 457(23), 100Rh 153, 105Pd 660, 193Ir 751(9), 197Au 547(16). 209Bi 422(3)77 and references there. 229Th 3110(60), 235U 4936(6). The Au NQR frequencies of Bowmaker14 have been scaled by a Q-ratio of (547 mb/590 mb) = 0.927. See80. Comparison between calculated (PW) and experimental B. The data are taken from Tables 7 and 8. The red dots refer to B(Au) and the blue dots refer to other nuclei. As one point of reference, the neutral Au atom in a 5d96s22D5/2 state has a B of −1049.781(11) MHz.67 In other words, one d–1 hole gives a GHz. Many of the red dots in Figure 5 are far smaller than that. Concerning the spin–orbit effects and other relativistic effects on q, at the level of multiplicative correction factors, we refer to Pyykkö and Seth.68 The MNC ← MCN conversion has been calculated by Jana et al.69 Hill et al.70 quote a Be of 2.80 MHz for an unspecified nucleus. They discuss the vibrational dynamics in detail. Wang46 observed photoelectron spectra in [Au(CN)2]−(g) and found in calculations significant covalent character in the Au–C bond. Recent Further Examples on δ4 Rings Kalita et al.82 suggest RhSc as a sextuply bonded diatomic with a Rh → Sc δ HOMO. Likewise, Tzeli and Karapetsas20 find X1Σ+ ground states for RuC, RhB, and PdBe, all three having nonbonding δ4 HOMOs. For systems of the [MUM] type with M = Rh, Ir, see Shen et al.83 and references there. Conclusions The δ4 ring lies energetically and radially in the valence range of the 5d elements discussed despite being formally a filled core orbital.81 Yet it yields mostly no covalent bonding contributions, although the Coulomb attraction to the neighbors must be substantial. In its way, this δ4 ring resembles the σ2 lone pairs. An open question is, what kind of chemistry could one do with the δ4?