Superatom Fusion and the Nature of Quantum ConfinementAnouck M. Champsaur,† Taylor J. Hochuli,† Daniel W. Paley,†,‡ Colin Nuckolls,*,†

and Michael L. Steigerwald*,†

†Department of Chemistry and ‡Columbia Nano Initiative, Columbia University, New York, New York 10027 United States

*S Supporting Information

ABSTRACT: Quantum confinement endows colloidal semiconductingnanoparticles with many fascinating and useful properties, yet a criticallimitation has been the lack of atomic precision in their size and shape. Wedemonstrate the emergence of quantum confined behavior for the firsttime in atomically defined Co6Se8(PEt3)6 superatoms by dimerizing[Co6Se8] units through direct fusion. To accomplish this dimerization, weinstall a reactive carbene on the [Co6Se8] core to create a latent fusionsite. Then we transform the reactive carbene intermediate into a materialwith an expanded core, [Co12Se16], that exhibits electronic and optical properties distinct from the parent monomer. Thechemical transformation presented herein allows for precise synthetic control over the ligands and size of these clusters. We showby cyclic voltammetry, infrared spectroscopy, single crystal X-ray diffraction, and density functional theory calculations that theresulting fused [Co12Se16] material exhibits strong electronic coupling and electron delocalization. We observe a bandgapreduction upon expanding the cluster core, suggesting that we have isolated a new intermediate in route to extended solids.These results are further corroborated with electronic structure calculations of a monomer, fused dimer, trimer, and tetramerspecies. These reactions will allow for the synthesis of extended highly delocalized wires, sheets, and cages.

KEYWORDS: Nanoscale atoms, nanoscale building blocks, superatoms, quantum confinement

In this manuscript, we describe an unprecedented reactionsequence that fuses metal chalcogenide clusters to yield

electronically strongly coupled dimers. The electronic andoptical properties of the fused dimer are drastically differentfrom the parent monomer, whereas a similarly sized bridgeddimer material exhibits the same (additive) electronic andoptical properties as the monomer. A combination of electronicstructure data and calculations indicates that the strongcoupling between the dimer lobes can be interpreted as theonset of bandlike behavior. Furthermore, density functionaltheory (DFT) calculations of a fused trimer and tetramer revealdecreasing highest occupied molecular orbital/lowest unoccu-pied molecular orbital (HOMO−LUMO) gaps with additionalfusion. We thus observe the manifestation of quantum sizeeffects in that the bandgap decreases as the size of thenanostructures increases upon fusion. If realized experimentally,additional fusion steps may lead to the formation of anextended solid.Cooperative physical phenomena observed in the fused

dimer are reminiscent of those in quantum dot (QD) solids,which are formed from epitaxially connected semiconductorQD building blocks.1−3 While electronically coupled QDlattices can be constructed from molecularly linked QDs, amajor obstacle is long-range structural disorder.4−6 Orientedattachment of QDs increases interdot coupling and leads toelectron delocalization in the resulting highly ordered QDsuperlattices.3 Here, similarly, directly connected [Co6Se8]units (fused dimer) exhibit greater electron delocalizationrelative to molecularly connected [Co6Se8] units (bridgeddimer).

The metal chalcogenide clusters fundamental to our studyare members of a larger class of clusters with the [M6E8]stoichiometry (M = metal, E = chalcogen) that are buildingblocks for extended materials.7−14 Some such clusters areconstituent fragments of traditional, inorganic materials withunusual properties such as superconductivity and ferromagnet-ism (e.g., Chevrel and Zintl phases),15−21 whereas in othercases clusters with this core stoichiometry have been convertedinto extended solids in which the [M6E8] building block issignificantly distorted.22−24 The [Co6Se8] clusters belong to thelatter family, and we recently found that we could modify theligand environment around this core to include designed,stoichiometrically and stereochemically precise combinations ofphosphine and CO groups (Figure 1A).25 We have shown thatthe CO ligand is preferentially labile in these clusters, and it canbe replaced photochemically with two-electron donors such asisonitriles and phosphines to yield, for example, the diisonitrile-bridged dimer shown in Figure 1B (BD). Structurally, we viewthese functionalized clusters as superatoms26−30 because wehave been able to combine them to form dimers, trimers, andpolymers through simple substitution reactions of the COgroups in which the [M6E8] unit remains intact.

25,31 Here weshow that a CO group can be replaced with a carbene, forminga reactive intermediate species that allows us to crack open the[Co6Se8] core, and this crack leads to the fusion of two

Received: May 4, 2018Revised: May 28, 2018Published: June 7, 2018

Letter

pubs.acs.org/NanoLettCite This: Nano Lett. 2018, 18, 4564−4569

© 2018 American Chemical Society 4564 DOI: 10.1021/acs.nanolett.8b01824Nano Lett. 2018, 18, 4564−4569

Dow

nloa

ded

via

CO

LU

MB

IA U

NIV

on

Janu

ary

4, 2

019

at 2

0:12

:18

(UT

C).

Se

e ht

tps:

//pub

s.ac

s.or

g/sh

arin

ggui

delin

es f

or o

ptio

ns o

n ho

w to

legi

timat

ely

shar

e pu

blis

hed

artic

les.

superatoms, yielding the fused dimer, Co12Se16(PEt3)10 (FD,Figure 1C). We conclude from electrochemistry, opticalspectroscopy, density functional theory calculations, and X-ray crystallography that the lobes in these dumbbell shapeddimers are strongly coupled to each other, so much that thebonding between the two is best viewed as covalent. Thisexpansion of the cluster core is a step in the growth of the solidand exhibits the essentials of quantum confinement in itselectronic structure.Results and Discussion. Our strategy to fuse superatom

cores was to synthesize an intermediate with a labile ligand. Tothis end, we investigated whether a reactive species could begenerated in situ that would displace the CO ligand under theproper conditions and leave the superatom activated for furthertransformations (Figure 2A). The monocarbonyl precursor

[Co6Se8(CO)(PEt3)5]n[PF6]n (n = 0 or 1)25 and

(trimethylsilyl)diazomethane (TMSD) react when irradiatedwith visible light to give a carbene-terminated cluster,[Co6Se8(C(H)SiMe3)(PEt3)5]

n[PF6]n (Figure 2A). Irradiationserves the dual purpose of ejecting the carbonyl from thecluster and generating the reactive carbene, :C(H)(TMS). Thecarbene-cluster adduct is stable and isolable but the carbeneremains reactive. Details for the 1H- and 31P NMR used to

monitor the reaction and characterize the materials arecontained in the Supporting Information. Figure 2B displaysthe structure from single crystal X-ray diffraction (SCXRD) of[Co6Se8(C(H)SiMe3)(PEt3)5][PF6] (1).The carbene-terminated cluster (1) shows a surprising and

unprecedented SeCCo binding motif (inset in Figure 2B):the carbenoid carbon has inserted into one of the CoSebonds in the superatom, and thus has cracked the [Co6Se8]shell. The inorganic core of 1 is significantly distorted relativeto that of the parent, all phosphine cluster [Co6Se8(PEt3)6]

+

(monomer, M).32 The “cracked” CoSe distance is signifi-cantly longer than the vicinal CoSe distances in 1 (2.439(5)Å vs 2.32−2.42 Å, inset in Figure 2B), and it is longer thanCoSe bonds in the all-phosphine core (2. 439(5) Å vs 2.32−2.36 Å).32 As a result of this distortion, the carbon-boundcobalt atom is pushed into the plane of its four adjacentselenium atoms. This leads to a compression of thecorresponding antipodal CoCo distance to 3.7 Å (comparedto the other two antipodal CoCo distances of 4.1 Å) andthus significant shortening of cis CoCo distances to2.670(2), 2.689(2), 2.754(1), and 2.772(2) Å. In M, the cisCoCo distances are 2.90−2.95 Å. The insertion of thecarbene into one of the CoSe bonds not only breaks thatparticular bond but it also affects the entire [Co6Se8] core, asdemonstrated by this modulation of the CoCo distances.In order to gain more insight into 1, we studied models of its

constituent parts using quantum chemical calculations (seeSupporting Information for details). One question is the natureof the coordinatively unsaturated superatom, [(Et3P)5Co6Se8]

+,an intermediate that would form were [(Et3P)5Co6Se8(CO)]

+

to lose its CO. We first studied the conceptually simplerelectrically neutral system (Et3P)5Co6Se8.

33 We found thatwh i l e the coord ina t i v e l y s a tu r a t ed supe r a tom ,(Et3P)5Co6Se8(CO) is a closed shell, singlet species (as is theall-phosphine monomer), the open shell, triplet state of(Et3P)5Co6Se8 is lower in energy than the closed shell singlet.Although we had not anticipated this result, on reflection it isnot entirely surprising. It is reminiscent of situations intraditional coordination chemistry and organometallic chem-istry where strong-field ligands promote low-spin states whenthey coordinate to previously high-spin subsystems. Apparently,the hypothetical naked superatom, Co6Se8, were it to exist,would be a high-spin entity.We can learn more about the hypothetical intermediate,

(Et3P)5Co6Se8, by looking at the two singly occupied orbitalsthat characterize the triplet (Figure S1). One of these orbitals islocalized on the coordinatively unsaturated Co atom, and theother is distributed among this Co atom and its Se neighbors.Although this triplet state is certainly quite complicated, thesimplest interpretation is that it shows a broken CoSe bond.Thus , i f and when the CO d i s soc i a t e s f r om(Et3P)5Co6Se8(CO) a CoSe bond cleaves.Whether or not (Et3P)5Co6Se8 (or its cation) actually occurs

during the formation of 1, the electronic structure of the formerhints at the structure of the latter: if (Et3P)5Co6Se8 is a triplet,then it is primed to form two covalent bonds to a reactionpartner such as a triplet carbene. In simple terms, 1 has acarbene which has inserted into one of the Co−Se bonds of theoriginal cluster. There is one fewer CoSe bond in 1 than inM, but in compensation there is a new CoC and a new SeC covalent bond. We will use this analysis when we discussbelow the nature of the fused dimer.

Figure 1. (A) CO and phosphine substituted [Co6Se8] clusters. (B)BD with a diisonitrile linker {(PEt3)5Co6Se8}−L−{Co6Se8(PEt3)5} (L= CNC6H4NC). (C) FD, [Co12Se16(PEt3)10].

Figure 2. (A) Reaction scheme to form the carbene-cluster adduct[Co6Se8(C(H)SiMe3)(PEt3)5][PF6] (1) from the reaction of amonocarbonyl precursor with (trimethylsilyl)diazomethane. (B)Molecular structure from SCXRD of carbene-terminated intermediate[Co6Se8(C(H)SiMe3)(PEt3)5][PF6] (1). (C) Reaction scheme toform the fused dimer (FD) dication, [Co12Se16(PEt3)10][PF6]2 fromthe heating of 1. (D) Molecular structure from SCXRD of the dicationof FD, [Co12Se16(PEt3)10][PF6]2. PF6

− counterions and hydrogenatoms in (B,D) (except C(H)SiMe3) have been omitted to clarify theview. Thermal ellipsoids are set at 20% probability level in (B) and50% probability level in (D).

Nano Letters Letter

DOI: 10.1021/acs.nanolett.8b01824Nano Lett. 2018, 18, 4564−4569

4565

These CoC and SeC bonds in 1 are still quite reactive:the dication of FD [Co12Se16(PEt3)10][PF6]2 forms in 66%isolated yield when 1 is heated gently (5 h, 70 °C) (Figure2C,D). We can also prepare the neutral dimer,Co12Se16(PEt3)10, and the 1+ dimer, [Co12Se16(PEt3)10][PF6].Details of the syntheses of these species as well as theirstructures from SCXRD are in the Supporting Information.Table S1 contains the important structural metrics for thesedimers in different oxidation states. The high yield andspecificity of the reactions to make FD are magnified whencomparing to that of the closest analog that has been made todate, [Co12S16(PEt3)10][TNCQ]2. This analog was obtainedfortuitously from a mixture in low yield and was not able to befully characterized.34 This type of cluster fusion has also beenobserved for other cluster core types such as the cubic [Fe4S4]clusters to form fused dimers and tetramers.35

By comparing the electrochemical behaviors of three[Co6Se8]-based systems, a Co6Se8(PEt3)6 monomer (M), thearyl-diisonitrile bridged dimer {(PEt3)5Co6Se8}−L−{Co6Se8(PEt3)5}] (BD, Figure 1B),25 and the fused dimer[Co12Se16(PEt3)10] (FD, Figure 1C), we conclude that there isstrong electronic communication between the two superatomlobes in FD.Figure 3A displays the cyclic voltammograms (CVs) of the

three materials. The CV of M has three reversible one-electronoxidations from n = 0 to +3, and two reductions to n = −1 and−2, although these latter reductions occur consecutively with asmall voltage separation between them. The CV of BD alsodisplays three reversible oxidation events at similar potentials inaddition to a reduction event but in contradistinction to thebehavior of M each transition is a two-electron event, as wehave previously shown.25 Therefore, again as we havepreviously shown the redox behavior of BD is essentially thatof its monomeric components (except each event is doubled),and BD behaves electronically as a sum of its component partswith little to no electronic coupling between the two lobes.25 Ashas been reported, one quantitative measure of the extent ofintercluster coupling in dimeric species is the peak separationbetween the 0/+1 and +1/+2 redox couples (ΔE1/2) as itreflects the stability of the intermediate +1 state.36 This value isrepresented by the blue arrow in Figure 3A. In BD, ΔE1/2 isessentially zero, since both one-electron oxidations (0/+1 and+1/+2) occur near the same potential, and no such +1 state canbe isolated.In contrast, the electrochemical behavior of FD is very

different. Its CV has five distinct and reversible one-electronevents from n = 0 to +5 and a reversible reduction event from 0to −1 at −2 V. The significance of the CV of FD is that eachpeak is a one-electron event (determined as with BD bymeasuring the open circuit potential of the neutral, + 1 and +2species), and therefore, the two [Co6Se8] units are notelectrochemically isolated as they are in BD. This is significantbecause although FD is structurally composed of two clustercores, the electronic data shows that the behavior of thematerial diverges drastically from a simple additive behavior ofits component building blocks. Importantly, while we have seenno evidence for the discrete presence of the monocation of BD,we have been able to prepare and isolate the monocation ofFD, [Co12Se16(PEt3)10][PF6] (details in the SupportingInformation). The stability of this species is correlated to themagnitude of ΔE1/2 between the 0/+1 and +1/+2 redoxcouples, which is 0.65 V (compared to 0 V in BD). Putting thisvalue in the context of other [M12E16] metal chalcogenide fused

dimers we find that it is the largest ([Cr12S16], 0.42 V;[Mo12S16], 0.41 V; [Mo12Se16], 0.38 V; [Re12Se16], 0.22V).7,34,37−40 We conclude that intercluster coupling for ourcobalt dimer FD is higher than that of previously reportedspecies. Although all of these dimers are structurally similar,important factors detailed below distinguish FD from theothers.The separation between the −1/0 and 0/+1 potentials,

measured from the peak extrema, has been used to estimateHOMO/LUMO splittings;41,42 one striking conclusion fromthe CV data in Figure 3A is that there is a large reduction of theHOMO/LUMO splitting in FD relative to eitherM or BD (redarrows in Figure 3A). Markedly, while BD is physically as large,if not larger, than FD, it has a HOMO/LUMO gap identical tothat of M. Furthermore, the modulation of this gap is not aresult of a change in the ligand environment: Figure 3A displaysthe CVs of two additional monomeric species, [Co6Se8(CO)-(PEt3)5]

n and [Co6Se8(CNC6H3Me2)(PEt3)5]n, which have

similar HOMO/LUMO splittings as the all-phosphinemonomer. The [Co6Se8] and [Co12Se16] cluster cores arestoichiometrically similar to the solid state compound Co3Se4,which the solid state literature reports is a narrow band gapsemiconductor.43 [Co12Se16(PEt3)10]

n is both easier to oxidizeand easier to reduce than [Co6Se8(PEt3)6]

n, as observed in thedecrease in its first oxidation potential and increase in its

Figure 3. (A) Cycl ic voltammogram of fused dimer[Co12Se16(PEt3)10]

n (FD), aryl-diisonitrile bridged dimer[Co12Se16(CNC6H4NC)(PEt3)10]

n (BD), parent monomer[Co6Se8(PEt3)6]

n (M), [Co6Se8(CNC6H3Me2)(PEt3)5]n and

[Co6Se8(CO)(PEt3)5]n. The red arrow shows the effective bandgap

as estimated from CV and the blue arrow shows the ΔE1/2 for thedimeric species. (B) Near-infrared spectra of FD, BD, and M inmultiple charged states. The data were collected in dichloromethane.Black arrows show the absorption features around 1300 and 1500 nmin the spectra of FD. The spikes observed at ∼1700 nm are due tosolvent absorption bands. (Co6Se8(PEt3)6 is poorly soluble andtherefore its extinction coefficient is lower than expected).

Nano Letters Letter

DOI: 10.1021/acs.nanolett.8b01824Nano Lett. 2018, 18, 4564−4569

4566

reduction potential. The direct result of this is the shrinking ofthe effective bandgap from ∼1.6 eV in the monomer to ∼0.7 eVin the fused dimer. In the context of the related solid statecompound Co3Se4,

44 this suggests the onset of the evolution ofbulk electronic properties from the molecular regime.We then compared the optical behaviors of the three

[Co6Se8]-based systems and observe drastic differencesbetween the fused species relative to the bridged dimer andmonomer. Figure 3B displays the near-infrared (IR) of all threespecies in their different charged states. M and BD absorptions(all charged states) look quite similar to each other aside froman approximate doubling of extinction coefficient in BD. Thisindicates additive (or noninteracting) monomer units, as wasobserved with cyclic voltammetry. In contrast, the near IR dataof all three fused dimer species (neutral, + 1, + 2) show muchlarger extinction coefficients than those of M and BD.Furthermore, a small absorption is observed in the neutraland singly charged fused dimer at ∼1300 nm, and an intenseand broad absorption is observed in the +2 species at 1525 nm.Therefore, both electrochemistry and optical spectroscopyshow that while BD largely recapitulates the properties of thecoordinatively saturated monomer, FD is qualitatively different:it is easier to oxidize, easier to reduce, and easier to exciteoptically. Were the monomer and dimer clusters to beappreciated as small, albeit distorted, fragments of bulk cobaltselenide, and were quantum confinement to occur in thismaterial, one would expect changes such as these.45

We sought to explain the dramatic change in propertiesbetween the BD and FD, and why the latter behaves, at leastelectrochemically and optically, more like the extended solid.We conclude that upon expanding the superatom core from[Co6Se8] to [Co12Se16] in the dimerization process, theintercluster bonding arrangement implied in Figure 4Adominates. From the crystal structure of FD (Figure 2D) weobserve that the intercluster cobalt−cobalt distance (2.731(1)Å) is significantly shorter than intracluster cobalt−cobaltdistances (2.92−3.03 Å) but still elongated relative to theCoCo distance in metallic Co (2.508(8) Å).46 In addition,the intracluster CoSe bonds at the fusion point (2.409(1) Å)(Se in the fusion point is represented in light pink in Figure4A) are elongated relative to both the other intracluster CoSe (2.32−2.38 Å) and intercluster CoSe bonds (2.3081(9)Å). The length of this elongated intracluster CoSe distancein FD is comparable to that of the analogous CoSe distancein the CoCSe motif of carbene 1 (2.439(5) Å). Weremove this “bond” in the representation of FD in Figure 4A toemphasize the structural reorganization of CoSe bonds ingoing from the monomer to the dimer; within each [Co6Se8]unit in the fused dimer, there is one fewer CoSe covalentbond, that bond having been “moved” to become anintercluster covalent bond.We can rationalize this bonding by again considering the

(Et3P)5Co6Se8 subunit. Because this unit has a triplet groundstate (or at least this state is very low-lying), its two unpairedspins are primed to bond to other triplet-state molecules.Perhaps the simplest candidate for its reaction partner isanother copy of itself: two triplet, coordinatively unsaturatedmonomers combine to form FD, this dimer being held togetherby two Co−Se covalent bonds.While in FD we observe the reorganization of CoSe bonds

as shown in Figure 4A, what we find in the apparentlyanalogous cases of [M12E16] (M = Re, Mo, Cr; E = S, Se) is therepresentation shown in Figure 4B, in which metal−metal

bonds and weak intercluster interactions are dominant. TableS2 contains important bond metrics and atomic distances for[Co12Se16(PEt3)10] compared to these other [M12E16(PEt3)10]dimers. M−M bonds within each [M6] unit are comparable toM−M bonds within the metal whereas intercluster M−Mseparations are larger and nonbonding.46 Furthermore,intercluster M−E bonds are significantly longer than theintracluster M−E bonds. Thus, as shown in Figure 4B, in thedimerization process there are no bond rearrangements withineach core. Instead, the two cluster cores remain the same as inthe monomer and thus the overall bonding in the dimer is bestrepresented by two independent metal-bound octahedraconnected via weakly interacting M−E bonds across the fusionpoint. We have performed DFT calculations on (Me3P)5Mo6S8as a prototype coordinatively unsaturated “Chevrel” monomer(see Supporting Information for details). In this case, theground state is the simple, expected singlet, and the lowest-energy triplet is quite high in energy. Furthermore, the twosingly occupied orbitals that describe this triplet state indicatethat it is a MoMo bond, not a MoS bond, that is cleaved inthis triplet (Figure S2). Therefore, even if this triplet wereavailable, intercluster Mo−S covalent bonding would not befavored. Nevertheless, the Mo, Re, and Cr dimers do form. Weconclude that the intercluster Mo−S bonds in the Mo12S16dimer are donor−acceptor (M−L-type) bonds. Conceptually,the intercluster bonding in Mo12S16(PEt3)10 arises from the

Figure 4. (A) The dominant bonds and bond rearrangements in[Co12Se16]. The CoCo distance in elemental cobalt is 2.508(8).46

(B) The dominant bonds and bond rearrangements in other [M12E16]dimers (M = Re, Mo, Cr; E = S, Se) Black spheres, metal; pinkspheres, chalcogen; light pink spheres, chalcogen involved inintercluster M−E. Dark black lines represent bonding interactionsand dotted red lines represent weaker donor−acceptor interactions.(C) Calculated HOMO and LUMO energies (PEt3 ligands replacedwith PMe3) as well as HOMO/LUMO splitting (bandgap), in M, FD,fused trimer, and fused tetramer.

Nano Letters Letter

DOI: 10.1021/acs.nanolett.8b01824Nano Lett. 2018, 18, 4564−4569

4567

replacement of MoP donor/acceptor bonds with MoSdonor/acceptor bonds. In contrast, the “intercluster” CoSebonds in Co12Se16(PEt3)10 are covalent (M−X-type) bonds.Therefore, these particular CoSe bonds are now part of thebackbone of [Co12Se16(PEt3)10] and covalently connect eachcluster core to form a fused entity in which the bonds of themonomer have rearranged to form the dimer with newproperties that are more than the sum of two individual[Co6Se8(PEt3)6] units.We thus observe the gradual evolution of bulk electronic

properties upon directed expansion of core nuclearity. Tocorroborate this finding, we calculated the HOMO and LUMOenergies of the monomer and fused dimer (with PEt3 ligandssimplified to PMe3) using DFT in addition to the energies for afused trimer and fused tetramer (Co18Se24(PMe3)14 andCo24Se32(PMe3)18, respectively). The final calculated geo-metries for the fused trimer and fused tetramer are displayedin Figure S3. The difference in bandgap between the fuseddimer and monomer mirrors our experimental finding(calculated difference, 0.93 eV; actual difference, 1 eV). Ourcalculations further show that each additional fused [Co6Se8]core leads to the nanostructure’s bandgap reduction (Figure4C) that begins to taper off at the tetrameric core.Conclusions. We have expanded on the reaction chemistry

of clusters as superatoms. We have shown that coordinativelyunsaturated clusters can have diverse electronic states availableto them, and these varied states can allow varied reactionpathways. While L6Co6Se8 is a ground state singlet, L5Co6Se8 isa ground state triplet. We install a reactive carbene on the[Co6Se8] core to create a latent fusion site and we show bycyclic voltammetry and SCXRD that the resulting fused[Co12Se16] material exhibits strong electronic coupling andelectron delocalization. Furthermore, the bandgap reductionobserved upon expanding the cluster core suggests that we haveisolated a more bulklike intermediate than the [Co6Se8]monomer in going from molecules to an extended solid. Theisolation of [Co12Se16(PEt3)10]

n thus represents a significantadvancement in observing the manifestation of quantumconfinement in an atomically precise nanocluster.We have now shown that we can dimerize [Co6Se8] units in

two distinct ways: either by direct fusion or by using a ditopicbridging ligand. In the former case the two units areelectronically strongly coupled, in the latter case less so. If wedraw a comparison to conjugation in organic molecules andmaterials, this is reminiscent of conjugated versus non-conjugated dienes, for example 1,3-butadiene (FD) and 1,4-pentadiene (BD). However, conjugation in these systemsinvolves the overlap of π-orbitals, whereas from a purelychemical bonding standpoint, σ-conjugation dominates thedelocalization of our systems. We are currently extending ourreaction sequence to prepare higher-order conjugated supera-tom wires. More broadly, we can extend the design principleused to access a dimer to our building blocks that have multipleCO groups25 and apply the carbene/fusion sequence to obtainstrongly coupled materials with programmed dimensionality.These strongly coupled systems are necessary to discovermaterials with new and unusual magnetic and electronicproperties.Methods. All synthetic methods, characterization, X-ray

diffraction, computational details, and crystallographic (CIF)files can be found in the Supporting Information.

■ ASSOCIATED CONTENT*S Supporting InformationThe Supporting Information is available free of charge on theACS Publications website at DOI: 10.1021/acs.nano-lett.8b01824.

Methods, experimental section, NMR spectroscopy,electrochemistry, crystallography, computational chem-istry, figures, tables, and additional references (PDF)Dimer neutral data (CIF)Carbene 1plus data (CIF)1plus data (CIF)Dimer 2plus data (CIF)

■ AUTHOR INFORMATIONCorresponding Authors*E-mail: cn37@columbia.edu.*E-mail: mls2064@columbia.edu.ORCIDColin Nuckolls: 0000-0002-0384-5493NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSC.N. thanks Sheldon and Dorothea Buckler for their generoussupport. Support for this research was provided by the Centerfor Precision Assembly of Superstratic and Superatomic Solids,an NSF MRSEC (award number DMR-1420634), and the AirForce Office of Scientific Research (award number FA9550-18-1-0020). Single crystal X-ray diffraction was performed at theShared Materials Characterization Laboratory at ColumbiaUniversity, maintained using funding from Columbia Universityfor which we are grateful. We thank Raul Hernandez Sanchezfor insightful discussions. We thank Kihong Lee for his helpwith near-IR spectroscopy.

■ REFERENCES(1) Boneschanscher, M. P.; Evers, W. H.; Geuchies, J. J.; Altantzis, T.;Goris, B.; Rabouw, F. T.; van Rossum, S. A. P.; van der Zant, H. S. J.;Siebbeles, L. D. A.; Van Tendeloo, G.; Swart, I.; Hilhorst, J.; Petukhov,A. V.; Bals, S.; Vanmaekelbergh, D. Science 2014, 344 (6190), 1377−1380.(2) Evers, W. H.; Goris, B.; Bals, S.; Casavola, M.; de Graaf, J.; vanRoij, R.; Dijkstra, M.; Vanmaekelbergh, D. Nano Lett. 2013, 13 (6),2317−2323.(3) Whitham, K.; Yang, J.; Savitzky, B. H.; Kourkoutis, L. F.; Wise, F.;Hanrath, T. Nat. Mater. 2016, 15 (5), 557.(4) Yang, J.; Wise, F. W. J. Phys. Chem. C 2015, 119 (6), 3338−3347.(5) Koole, R.; Liljeroth, P.; Donega, C. D.; Vanmaekelbergh, D.;Meijerink, A. J. Am. Chem. Soc. 2006, 128 (32), 10436−10441.(6) Kovalenko, M. V.; Scheele, M.; Talapin, D. V. Science 2009, 324(5933), 1417−1420.(7) Zheng, Z. P.; Long, J. R.; Holm, R. H. J. Am. Chem. Soc. 1997,119 (9), 2163−2171.(8) Shores, M. P.; Beauvais, L. G.; Long, J. R. J. Am. Chem. Soc. 1999,121 (4), 775−779.(9) Beauvais, L. G.; Shores, M. P.; Long, J. R. Chem. Mater. 1998, 10(12), 3783.(10) Tulsky, E. G.; Crawford, N. R. M.; Baudron, S. A.; Batail, P.;Long, J. R. J. Am. Chem. Soc. 2003, 125 (50), 15543−15553.(11) Choi, B.; Yu, J.; Paley, D. W.; Trinh, M. T.; Paley, M. V.; Karch,J. M.; Crowther, A. C.; Lee, C. H.; Lalancette, R. A.; Zhu, X. Y.; Kim,P.; Steigerwald, M. L.; Nuckolls, C.; Roy, X. Nano Lett. 2016, 16 (2),1445−1449.

Nano Letters Letter

DOI: 10.1021/acs.nanolett.8b01824Nano Lett. 2018, 18, 4564−4569

4568

(12) Naumov, N. G.; Virovets, A. V.; Sokolov, M. N.; Artemkina, S.B.; Fedorov, V. E. Angew. Chem., Int. Ed. 1998, 37 (13−14), 1943−1945.(13) Jin, S.; DiSalvo, F. J. Chem. Mater. 2002, 14 (8), 3448−3457.(14) Roy, X.; Lee, C. H.; Crowther, A. C.; Schenck, C. L.; Besara, T.;Lalancette, R. A.; Siegrist, T.; Stephens, P. W.; Brus, L. E.; Kim, P.;Steigerwald, M. L.; Nuckolls, C. Science 2013, 341 (6142), 157−160.(15) Fischer, O. Appl. Phys. 1978, 16 (1), 1−28.(16) Scharfe, S.; Kraus, F.; Stegmaier, S.; Schier, A.; Fassler, T. F.Angew. Chem., Int. Ed. 2011, 50 (16), 3630−3670.(17) Saito, T.; Imoto, H. Bull. Chem. Soc. Jpn. 1996, 69 (9), 2403−2417.(18) Perrin, A.; Perrin, C. C. R. Chim. 2012, 15 (9), 815−836.(19) Toberer, E. S.; May, A. F.; Snyder, G. J. Chem. Mater. 2010, 22(3), 624−634.(20) Chevrel, R.; Hirrien, M.; Sergent, M. Polyhedron 1986, 5 (1−2),87−94.(21) Schafer, H.; Eisenmann, B.; Muller, W. Angew. Chem., Int. Ed.Engl. 1973, 12 (9), 694−712.(22) Steigerwald, M. L. Polyhedron 1994, 13 (8), 1245−1252.(23) Stuczynski, S. M.; Kwon, Y. U.; Steigerwald, M. L. J. Organomet.Chem. 1993, 449 (1−2), 167−172.(24) Steigerwald, M. L.; Siegrist, T.; Stuczynski, S. M. Inorg. Chem.1991, 30 (10), 2256−2257.(25) Champsaur, A. M.; Velian, A.; Paley, D. W.; Choi, B.; Roy, X.;Steigerwald, M. L.; Nuckolls, C. Nano Lett. 2016, 16 (8), 5273−5277.(26) Khanna, S. N.; Jena, P. Phys. Rev. B: Condens. Matter Mater. Phys.1995, 51 (19), 13705−13716.(27) Tomalia, D. A.; Khanna, S. N. Chem. Rev. 2016, 116 (4), 2705−2774.(28) Claridge, S. A.; Castleman, A. W.; Khanna, S. N.; Murray, C. B.;Sen, A.; Weiss, P. S. ACS Nano 2009, 3 (2), 244−255.(29) Castleman, A. W.; Khanna, S. N. J. Phys. Chem. C 2009, 113 (7),2664−2675.(30) Chauhan, V.; Reber, A. C.; Khanna, S. N. J. Am. Chem. Soc.2017, 139 (5), 1871−1877.(31) Champsaur, A. M.; Meziere, C.; Allain, M.; Paley, D. W.;Steigerwald, M. L.; Nuckolls, C.; Batail, P. J. Am. Chem. Soc. 2017, 139(34), 11718−11721.(32) Hong, M. C.; Lei, X. J.; Huang, Z. Y.; Jiang, F. L.; Liu, H. Q.Chinese Sci. Bull. 1992, 37 (21), 1798−1803.(33) We have performed similar analysis for the cationic species andobtained similar results. In [(Et3P)5Co6Se8]+ the lowest energy stateis a quartet with a doublet lying slightly higher in energy. The neutralcase is simpler to describe so we chose that case as our focus(34) Cecconi, F.; Ghilardi, C. A.; Midollini, S.; Orlandini, A. Inorg.Chim. Acta 1993, 214 (1−2), 13−15.(35) Goh, C.; Segal, B. M.; Huang, J. S.; Long, J. R.; Holm, R. H. J.Am. Chem. Soc. 1996, 118 (47), 11844−11853.(36) Ward, M. D. Chem. Soc. Rev. 1995, 24 (2), 121−134.(37) Kamiguchi, S.; Imoto, H.; Saito, T. Chem. Lett. 1996, 25, 555−556.(38) Amari, S.; Imoto, H.; Saito, T. Chem. Lett. 1997, 26, 967−968.(39) Amari, S.; Imoto, H.; Saito, T. J. Chin. Chem. Soc. 1998, 45 (4),445−450.(40) Zheng, Z. P.; Holm, R. H. Inorg. Chem. 1997, 36 (23), 5173−5178.(41) Kucur, E.; Riegler, J.; Urban, G. A.; Nann, T. J. Chem. Phys.2003, 119 (4), 2333−2337.(42) Haram, S. K.; Quinn, B. M.; Bard, A. J. J. Am. Chem. Soc. 2001,123 (36), 8860−8861.(43) Cao, S.; Chen, Y.; Kang, L.; Lin, Z. S.; Fu, W. F. J. Mater. Chem.A 2015, 3 (36), 18711−18717.(44) Garcia-Garcia, F. J.; Larsson, A. K.; Noren, L.; Withers, R. L.Solid State Sci. 2004, 6 (7), 725−733.(45) Brus, L. J. Phys. Chem. 1986, 90 (12), 2555−2560.(46) Wyckoff, R. W. G. Crystal Structures 1; Interscience Publishers,1963.

Nano Letters Letter

DOI: 10.1021/acs.nanolett.8b01824Nano Lett. 2018, 18, 4564−4569

4569