Gold Cyanide Molecular Wires

• Gold cyanide molecular wires are one-dimensional coordination motifs, forming AuCN chains via cyanide bridging between gold electrodes.
• They display conductance modulation driven by destructive quantum interference in series and enhanced by aurophilic coupling in parallel.
• Advanced DFT and NEGF methods reveal that contact geometry and orbital hybridization critically control electron transport in these systems.

Gold cyanide molecular wires are electrically active junction motifs in which cyanide-coordinated gold units participate directly in the transport pathway. In the literature considered here, the term covers two distinct but related cases: one-dimensional coordination chains assembled from AuCN repeat units between Au electrodes, and CN-terminated organometallic backbones contacted to Au through nitrile N atoms. The first realizes in situ formation of single and parallel $(\mathrm{AuCN})_n$ chains under ambient STM break-junction conditions, with conductance governed by destructive quantum interference and aurophilic coupling; the second isolates the Au–CN contact as an interfacial motif in a rigid Fe-based molecular wire and shows that weak N–Au coordination can suppress transport and stability (Luna et al., 2 Aug 2025, Schwarz et al., 2015).

1. Chemical identity and structural motif

In the AuCN-chain realization, the basic structural unit is the linear AuCN fragment, and repetition produces the bonding motif Au–C$\equiv$N–Au along the wire. The related $[\mathrm{Au(CN)}_2]^-$ complex contains a two-coordinate, closed-shell Au(I) center that forms $\sigma$ bonds via Au $6s/5d_{z^2}$ hybridization and $\pi$-backbonding into the C$\equiv$N $\pi^\*$ manifold. DFT-optimized junction geometries yield characteristic bond distances of 2.06 Å for Au–C, 1.17 Å for C$\equiv$N, and 1.97 Å for N–Au. The neutral AuCN unit measures approximately 3.2 Å, and approximately 5.2 Å when the AuCN fragment includes bonding to an apex Au atom of the electrode; these dimensions are consistent with the chain-length increments observed experimentally (Luna et al., 2 Aug 2025).

A distinct usage appears in rigid dinuclear organometallic wires of the form X(PP)$_2$FeC$\equiv$0Fe(PP)$\equiv$1X, with PP = Et$\equiv$2PCH$\equiv$3CH$\equiv$4PEt$\equiv$5 and X = CN, NCS, NCSe, C$\equiv$6SnMe$\equiv$7, or C$\equiv$8SnMe$\equiv$9. There, the backbone consists of two low-spin Fe centers connected by sp-hybridized carbon chains and supports a strongly conjugated, delocalized electron system spanning both metal centers. For the CN-terminated member, the terminal –CN group binds coordinatively through the nitrile N to Au. This system is chemically relevant to gold–cyanide contacts, but it is structurally different from the in situ assembled AuCN coordination chains because the transport backbone is the rigid [FeC$[\mathrm{Au(CN)}_2]^-$0Fe] unit rather than a chain of AuCN repeats (Schwarz et al., 2015).

2. In situ formation of AuCN chain junctions

Ambient STM break-junction experiments assemble AuCN wires by repeatedly contacting and separating an Au tip and Au substrate under constant bias, typically 500 mV, after drop-casting a 1 mM aqueous solution of K[Au(CN)$[\mathrm{Au(CN)}_2]^-$1] on the Au substrate and gently drying it at approximately 45 °C. Dissociation of the salt enables binding of $[\mathrm{Au(CN)}_2]^-$2 and AuCN fragments to Au electrodes through the cyanide lone pair on N, giving Au–N anchoring. Computed terminal N–Au bond-breaking energies are 1.1 eV, 1.5 eV, and 1.6 eV for $[\mathrm{Au(CN)}_2]^-$3–3, respectively. These values exceed typical donor–acceptor linkers such as amines, thioethers, and pyridyl, and are comparable to Au–Au metallic bond strengths of approximately 1.5 eV, supporting extraction of Au atoms from electrodes and step-wise in situ chain growth (Luna et al., 2 Aug 2025).

The resulting one-dimensional $[\mathrm{Au(CN)}_2]^-$4 chains are linear, and elongation occurs in discrete steps of approximately 4.5–5 Å per appended repeat unit. Minimum-energy Au$[\mathrm{Au(CN)}_2]^-$5–$[\mathrm{Au(CN)}_2]^-$6–Au$[\mathrm{Au(CN)}_2]^-$7 geometries give electrode separations of 7.1 Å, 12.3 Å, and 17.3 Å for $[\mathrm{Au(CN)}_2]^-$8, 2, and 3, respectively, matching the experimentally observed increment. The same assembly process also yields junctions in which two or three adjacent chains bridge the gap in parallel. In those geometries, edge Au(I) atoms in neighboring chains approach to approximately 2.92 Å, which lies within the reported aurophilic regime of approximately 2.8–3.5 Å, and form inter-chain Au(I)–Au(I) $[\mathrm{Au(CN)}_2]^-$9–$\sigma$0 $\sigma$1 bonds. These interwire Au$\sigma$2Au contacts occur across the HG, MG, and LG junction families and are central to the transport response (Luna et al., 2 Aug 2025).

3. Conductance regimes and experimental signatures

Transport through the AuCN-chain junctions is characterized from thousands of conductance–displacement traces analyzed with log-binned histograms, linear-binned histograms, and two-dimensional conductance-versus-displacement histograms. Au point-contact plateaus appear at integer $\sigma$3, with $\sigma$4, followed by molecular plateaus after rupture of the metallic contact. The conductance signatures are largely bias independent below approximately 500 mV, and the reported data are typically taken at 500 mV with at least 6000 traces per sample. Eight histogram peaks group into high-, medium-, and low-conductance regimes, denoted HG, MG, and LG. The longest plateau in each regime, H1, M1, and L1, is assigned to a single fully stretched chain with $\sigma$5, 2, and 3 AuCN repeats, respectively (Luna et al., 2 Aug 2025).

Assignment	Conductance	Parallel enhancement in same regime
H1, $\sigma$6	$\sigma$7	H2/H1 $\sigma$8; H3/H1 $\sigma$9
M1, $6s/5d_{z^2}$0	$6s/5d_{z^2}$1	M2/M1 $6s/5d_{z^2}$2; M3/M1 $6s/5d_{z^2}$3
L1, $6s/5d_{z^2}$4	$6s/5d_{z^2}$5	L2/L1 $6s/5d_{z^2}$6

Plateau lengths further support the assignments. HG features show short molecular plateaus of approximately 0.2 nm, whereas LG features show long plateaus of approximately 1.1 nm. After accounting for Au snapback of approximately 0.5–0.8 nm, these correspond to molecular lengths of approximately 0.7–1.0 nm for HG and approximately 1.6–1.9 nm for LG, consistent with $6s/5d_{z^2}$7 and $6s/5d_{z^2}$8 chain lengths. Two-dimensional histograms show length increments of approximately 4.5 Å between HG, MG, and LG, consistent with appending AuCN units in series. The original $6s/5d_{z^2}$9 can also bind without rearrangement and exhibits conductance similar to the $\pi$0 case (Luna et al., 2 Aug 2025).

The secondary and tertiary peaks in each regime arise from two or three adjacent wires in parallel rather than from simple histogram broadening. Their enhancement factors are superlinear relative to the corresponding single-chain conductance and therefore deviate from simple integer addition expected for non-interacting parallel channels. This experimental distinction is one of the defining empirical signatures of metallophilicity in the AuCN-chain junction family (Luna et al., 2 Aug 2025).

4. Interference, metallophilicity, and the transport mechanism

For AuCN chains in series, the conductance follows the Landauer form

$\pi$1

and the measured length dependence is fitted as

$\pi$2

or equivalently

$\pi$3

with $\pi$4–5 Å per AuCN repeat. The decay constants are $\pi$5 and $\pi$6. DFT and eigenstate analysis attribute this strong suppression of conductance to destructive quantum interference. In the monomer junction, transport near $\pi$7 is governed by a HOMO-like state denoted E1, while Au $\pi$8- and $\pi$9-derived LUMO-like states E2 and E3 lie nearby. Their symmetry relations produce a deep transmission dip at $\equiv$0 and an antiresonance in the HOMO manifold around $\equiv$1 eV. The resulting $\equiv$2 is therefore large and makes $\equiv$3 wires unexpectedly insulating despite the presence of transition-metal centers (Luna et al., 2 Aug 2025).

Parallel AuCN junctions behave differently because neighboring chains interact through Au(I)–Au(I) metallophilic coupling. DFT/NBO analysis shows that a monomer LUMO-like Au $\equiv$4-derived state E2, with approximately 52% Au $\equiv$5 character, becomes a symmetric HOMO-like state E4 in the dimer, with approximately 47% Au $\equiv$6 character. This reordering indicates formation of an inter-chain Au(I)–Au(I) $\equiv$7–$\equiv$8 $\equiv$9 bond. The coupled system lifts degeneracies, shifts antiresonances away from $\pi^\*$0, and increases $\pi^\*$1. Electron-density maps of the dominant LUMO-derived resonances, including HL1/HL2 in H2, ML1/ML2 in M2, and LL1/LL2 in L2, show substantial density between edge Au atoms along the interwire axis at Au$\pi^\*$2Au $\pi^\*$3 Å, consistent with aurophilic bonding (Luna et al., 2 Aug 2025).

A common misconception is that the conductance increase in parallel junctions is simply “two resistors in parallel.” The reported scaling rejects that interpretation. Experiment and DFT agree closely on the enhancement ratios: H2/H1 is approximately 2.6 experimentally and 2.8 in DFT; H3/H1 is approximately 4.1 and 4.2; M2/M1 is approximately 2.7 and 3.2; M3/M1 is approximately 6.4 and 5.2; and L2/L1 is approximately 2.6 and 2.7. The excess over naive integer addition is therefore attributed to quantum reordering of frontier states induced by metallophilic coupling rather than to independent parallel channels (Luna et al., 2 Aug 2025).

5. Landauer–NEGF treatment and computational representations

For the AuCN-chain junctions, the energy-resolved transmission is computed as

$\pi^\*$4

where $\pi^\*$5 are the coupling matrices for molecule–electrode hybridization and $\pi^\*$6 are the retarded and advanced Green’s functions of the junction Hamiltonian including electrode self-energies. The calculations use FHI-aims with PBE-GGA. The electrodes are Au(111) slabs comprising 32 Au atoms in 4 layers with trimer apex tips, and the back three layers are frozen during relaxation. Candidate geometries are scanned in 0.05 Å steps to identify the minimum-energy electrode separation. “Light” basis sets are used for $\pi^\*$7, pseudo-tight for $\pi^\*$8, transmission is computed with AITRANSS, and gas-phase NBO and NAO analyses are performed in Gaussian at PBE/def2-QZVPD. The calculations typically overestimate absolute conductance by approximately 50%, but they reproduce the length dependence and the parallel enhancement ratios (Luna et al., 2 Aug 2025).

Within this framework, metallophilic coupling is represented explicitly through overlap of Au(I) $\pi^\*$9 orbitals in adjacent chains. A minimal two-chain coupling picture is given in which an interchain hopping $\equiv$0 between Au $\equiv$1-derived orbitals produces bonding and antibonding combinations split by approximately $\equiv$2, moving the antiresonance away from $\equiv$3 and increasing $\equiv$4. This minimal picture is consistent with the DFT frontier-state analysis and with the superlinear conductance increases for dimers and trimers (Luna et al., 2 Aug 2025).

A different but complementary NEGF–DFT treatment appears in the CN-terminated Fe$\equiv$5 wire study. There, transport is computed with PBE in GPAW, spin polarization is neglected because low-spin Fe is the ground state in the tested ligand field, and a scissor-operator correction is applied to the weakly coupled systems 1–3 to mitigate self-interaction and image-charge errors. The calculated transmission for all molecules shows a HOMO/HOMO$\equiv$6 doublet close to $\equiv$7, delocalized across the [FeC$\equiv$8Fe] backbone. The decisive variable is the contact motif: direct Au–C anchors produce broader transmission resonances shifted toward $\equiv$9, whereas coordinative CN, NCS, and NCSe anchors give narrower peaks farther from $_2$0, suppressing $_2$1 (Schwarz et al., 2015).

6. Comparative contact chemistry, limitations, and design implications

The two studies define a sharp contrast between intrinsic AuCN coordination chains and external molecular backbones attached to Au through nitrile N atoms. In ambient STM break junctions, AuCN repeats can be extracted and assembled into apex-bound chains whose conductance is limited primarily by destructive quantum interference in series and enhanced by aurophilic coupling in parallel. In UHV mechanically controllable break-junction experiments on X(PP)$_2$2FeC$_2$3Fe(PP)$_2$4X, by contrast, the CN-terminated molecule 1 forms highly unstable Au–1–Au junctions with only a weak and infrequent accumulation at $_2$5 at 50 mV, whereas NCS gives $_2$6, NCSe gives $_2$7, Au–C after SnMe$_2$8 extrusion in 4′ gives $_2$9, and the shorter Au–C analogue 5′ gives $\equiv$00 (Schwarz et al., 2015).

The poor behavior of the CN-terminated Fe$\equiv$01 wire is attributed to three identified factors: weak N–Au coordination relative to covalent Au–C, possible steric hindrance from bulky phosphine ligands close to the Fe center, and less effective orbital overlap at the N–Au contact. In Landauer terms, weak coupling yields narrow resonances with small tails at $\equiv$02, hence low $\equiv$03 and low conductance. By contrast, the AuCN-chain study shows that cyanometallate motifs can be structurally robust, that Au(I) centers possess low-lying $\equiv$04 orbitals capable of aurophilic interaction, and that lateral coupling between adjacent chains can mitigate destructive quantum interference by reordering frontier states (Luna et al., 2 Aug 2025, Schwarz et al., 2015).

Several limitations are explicit. The AuCN-chain study examines only $\equiv$05–3 series chains and 2–3 parallel wires under ambient conditions; junction geometry, apex oxidation state, and exact interwire registry can modulate coupling, and the dynamic environment can generate junction-to-junction variability. The CN-terminated Fe$\equiv$06 study could not obtain reproducible single-molecule junctions for compound 1, so there is no full I–V data set, no low-temperature spectra, and no reliable statistics beyond a weak accumulation near $\equiv$07–$\equiv$08 (Luna et al., 2 Aug 2025, Schwarz et al., 2015).

The combined design implications are specific. For AuCN-chain architectures, the reported strategy is to exploit metallophilic centers with accessible $\equiv$09-orbital character, engineer interchain geometries that favor $\equiv$10 overlap at Au$\equiv$11Au separations in the aurophilic range, and use break-junction assembly to control length-to-width ratios. For CN-anchored organometallic wires, the proposed future strategies are to use surfaces or adatom-rich tips that stabilize N–Au bonds, modify the environment around CN to reduce steric congestion, explore isocyanide or other stronger N-based anchors, and integrate multidentate motifs that increase contact area without disrupting delocalization. This suggests that “gold cyanide molecular wire” behavior is not determined by cyanide alone: the operative variable is whether cyanide participates in a self-assembled AuCN coordination chain capable of metallophilic state reordering, or merely serves as a weak terminal anchor to Au (Luna et al., 2 Aug 2025, Schwarz et al., 2015).