Abstract

Aromaticity is typically regarded as an intrinsic property of a molecule, correlated with electron delocalization, stability, and other properties. Small variations in the molecular geometry usually result in small changes in aromaticity, in line with Hammond’s postulate. For example, introducing bond-length alternation in benzene and square cyclobutadiene by modulating the geometry along the Kekulé vibration gradually decreases the magnitude of their ring currents, making them less aromatic and less antiaromatic, respectively. A sign change in the ring current, corresponding to a reversal of aromaticity, typically requires a gross perturbation such as electronic excitation, addition or removal of two electrons, or a dramatic change in the molecular geometry. Here, we use multireference calculations to show how movement along the Kekulé vibration, which controls bond-length alternation, induces a sudden reversal in the ring current of cyclo[16]carbon, C16. This reversal occurs when the two orthogonal π systems of C16 sustain opposing currents. These results are rationalized by a Hückel model which includes bond-length alternation, and which is combined with a minimal model accounting for orbital contributions to the ring current. Finally, we successfully describe the electronic structure of C16 with a “divide-and-conquer” approach suitable for execution on a quantum computer.
1. Introduction
Aromaticity is one of the most debated concepts in chemistry, and its definition has undergone several revisions over the last few decades.1−5 Today, the most commonly used criterion for aromaticity is magnetic,2 equating the presence of a diatropic or paratropic ring current in an applied magnetic field with aromaticity or antiaromaticity, respectively. In molecules with several π systems, local and/or global currents may be present simultaneously,6,7 reinforcing or opposing each other. Aromaticity is a defining characteristic of a molecule, linked with reactivity, stability, HOMO–LUMO8 and singlet–triplet gaps,9 feasibility for singlet fission,10 diradical character,11 wave function coherence,12 and other properties.
Cyclo[N]carbons are all-carbon rings with two orthogonal π systems, usually with N electrons in each.13 In an analogy to annulenes, cyclocarbons with N = 4n + 2 carbon atoms can be classed as doubly aromatic (with both π systems sustaining a diatropic ring current), while N = 4n cyclocarbons are expected to be doubly antiaromatic.14,15 Many cyclocarbons have been found in the gas phase, but only C10,16 C14,16 C16,17 and C1818 have been structurally characterized using scanning probe microscopy.
A recent on-surface investigation of C16 revealed the presence of strong bond-length alternation (BLA) and confirmed that its ground state is doubly antiaromatic.17 Here, we investigate the variation of aromaticity with geometry in the two lowest singlets (S0 and S1), the lowest triplet (T1) and quintet (Q1) state of C16. Surprisingly, we find that the total ring current in the S1 and T1 states can be reversed from aromatic to antiaromatic by movement along the Kekulé vibration (∼2300 cm–1), i.e., by changing the amount of bond-length alternation.19 These aromaticity reversals require a relatively small amount of energy, in contrast with previous reports which require a change in the electronic state,20−23 molecular charge24−26 or composition,27,28 or involve a high-lying transition state29,30 or a highly strained geometry.31 To our knowledge, the only comparable low-energy aromaticity reversal involves a conformational equilibrium between Hückel antiaromatic and Möbius aromatic conformers in a hexaphyrin derivative.32,33
The simplicity and unique electronic structure of cyclocarbons makes them an interesting testing ground for theoretical methods.13 Quantum algorithms, which strongly benefit from execution on a quantum computer, are a promising avenue for the further development of electronic structure methods. However, their execution on current quantum devices is limited by noise and coherence time.34,35 Here, we exploit the orthogonality between the two π systems of C16 to effectively increase the active space of the quantum unitary coupled clusters singles doubles (q-UCCSD) method,36 which we solve variationally.37 This divide-and-conquer approach enables us to obtain highly accurate results that would otherwise not be feasible to compute using near-term quantum devices.
1.1. Ipsocentric Approach
Before discussing the electronic structure and ring current in C16, we briefly summarize the rules for evaluating orbital contributions to magnetic properties in the ipsocentric approach, as developed by Steiner and Fowler:38−40
-
1.The current density J induced by a magnetic field (which, when integrated over the ring, gives the ring current strength, usually measured in nA/T) can be expressed as a sum of spin-allowed transitions from occupied (ψs) to unoccupied (ψt) orbitals. The contribution of each transition (Jst) can be written as a sum of a diatropic term (JDIAst), which is associated with aromaticity, and a paratropic term (JPARAst) associated with antiaromaticity:
The diatropic current, which opposes the externally applied field, is taken to be positive (JDIA > 0), while the paratropic current, which reinforces the external field, is taken to be negative (JPARA < 0).
1 -
2The diatropic (aromatic) contribution of a transition from ψs to ψt to the current density (JDIAst) is determined by the orbital energy difference Δεst and the translational matrix element MTst, which reflects the orbital coupling under the linear momentum operator p̂ (represented by T in Figure 1):39

2 In general, MTst will be large for spatially similar orbitals differing in the number of nodal planes by one,38 corresponding (in planar monocyclic molecules) to a change in the angular momentum k by one. For example, the purely diatropic ring current in benzene can wholly be attributed to translational transitions from k = 1 to k = 2 (Figure 1a left).40
-
3The contribution of an orbital pair to the paratropic current (JPARAst) is also modulated by Δεst and depends on the magnitude of the rotational matrix element MRst, which couples the orbitals under the angular momentum operator l̂ (represented by R in Figure 1):39

3 For planar monocyclic molecules without BLA, MRst will approach unity for orbital pairs related by a rotation, such as A2 and B2 in D8h cyclooctatetraene (COT, Figure 1b; A and B correspond to sine and cosine density patterns). Such systems will have open-shell character21,41,42 and very large paratropic currents. Introducing BLA in planar COT (Figure 1c) lifts the degeneracy between A2 and B2, producing a closed-shell singlet, lowering the total energy and decreasing antiaromaticity.
-
4
From eqs 2 and 3, we can deduce that very few orbitals around the HOMO and LUMO will meaningfully contribute to the ring current. In aromatic annulenes, it is generally sufficient to consider only translational transitions with Δk = 1, i.e., HOMO → LUMO. In antiaromatic annulenes, rotational transitions with Δk = 0 (A2 ↔ B2 in Figure 1b,c) are responsible for most of the ring current, with weak contributions from occupied orbitals below the HOMO.40 We also note that this rule is valid only when the magnetic response is global, i.e., when the frontier orbitals are delocalized along the entire ring, which is valid up to ∼C30 in the case of even-numbered cyclocarbons.20
Figure 1.
Translational (green, labeled T) and rotational (purple, labeled R) transitions in the lowest singlet (S0, left) and triplet (T1, right) state for (a) benzene, (b) planar COT with no BLA, and (c) planar COT with BLA. Full and dashed lines show relatively larger and smaller contributions to the ring current, respectively.
The ipsocentric approach provides an orbital-based rationalization of Baird’s rule, which predicts an aromaticity reversal in the lowest triplet (T1) state of annulenes compared to the S0 ground state.43 In benzene (Figure 1a right), promoting one electron to either A2 or B2 produces a very strong paratropic current, resulting in an aromaticity reversal in the T1 (and S1) excited states. On the other hand, in planar COT (Figure 1b,c) flipping an electron renders the rotational A2 ↔ B2 transitions forbidden, leading to a Baird aromatic T1 state. Finally, the evaluation of aromaticity in the S1 (and other excited states) is less straightforward due to a combination of diatropic and paratropic contributions, in agreement with the experimental observation that Baird’s rule is weaker for S1 than for T1.44
In the ground state (S0) of C16, the configuration of both π systems is similar to that in planar COT (Figure 1b,c). Flipping a single electron in C16 will result in a T1 state with mixed aromaticity, where one π system will be Baird aromatic and the other will remain antiaromatic (norcorrole is a similar example45). Flipping an electron in both π systems of C16 will result in a doubly Baird aromatic quintet (Q1) state, as shown by Fowler.46 We also investigate the S1 state in which an electron moves from one π system to the other, resulting in a pair of oppositely charged doublets.
2. Results and Discussion
To understand how the electronic structure and magnetic properties of C16 change with BLA (Figure 2a), we use three different approaches: (1) a tight-binding Hückel-Heilbronner model (HHM), described in the next section; (2) the complete active space self-consistent field (CASSCF) method; and (3) density functional theory (DFT). In CASSCF calculations, all orbitals in a predefined active space are optimized, and the wave function can be written as a linear combination of many configurations, all of which contribute to the final magnetic properties. In contrast, the DFT wave function is only a single determinant, with strictly defined orbital occupations.
Figure 2.

(a) BLA-inducing Kekule vibration in C16, with bond lengths d and nearest-neighbor couplings β denoted. Frost-Musulin diagrams (assuming β″ = β′) for C16 in the absence (top) and presence (bottom) of BLA, with the dominant paratropic contributions in the ground state shown in purple. Offset between the out-of-plane (blue) and in-plane (orange) π systems is denoted by γ. (b) Possible translational (T, green) and rotational (R, purple) transitions associated with k = 4–5 orbitals in low-lying electronic states of C16.
2.1. Hückel–Heilbronner model
To gain more qualitative insight in the electronic structure of C16, we employ a semiempirical HHM,47 which extends the simple Hückel model by replacing the nearest neighbor interaction energy β with an alternating pattern of β(1 + δ) and β(1 – δ), as shown in Figure 2a:
![]() |
4 |
At δ = 0, the HHM reduces to a simple Hückel model. At 0 < δ < 1, it describes a system in which the interaction between shorter bonds is given by β(1 + δ), while longer bonds are coupled by β(1 – δ), which was shown to be a good approximation by Stanger.48 Previously, the HHM has been used to describe how frontier orbital energies of annulenes19 and azaboraheterocycles49 change with BLA.
C16 has two π systems, with out-of-plane orbitals (interaction energy β″, total energy E″π) offset by γ relative to in-plane orbitals (interaction energy β’, total energy E′π). Using HHM, we can find the π orbital energies of C16 by solving (4) twice, allowing for β″ ≠ β′, as shown in Figure 2b. The total energy EHHMTOT can be obtained by adding the sigma contribution aσδ2 (approximated as a parabola with the minimum at δ = 0),47 to the energies of the two π systems (E″π and Eπ′):
| 5 |
In C16 without BLA, the frontier out-of-plane orbitals (A4” and B4”; Figure 2b top) are energy degenerate and slightly (γ = 0.1 eV) lower in energy than their in-plane counterparts (A4′ and B4′), resulting in an unstable double open shell singlet,17 in analogy to D4h cyclobutadiene50 or D8h COT.21,23,42
Introducing BLA lowers the energies of A4″ and A4′ by 2β″δ and 2β′’δ (and raises the energies of B4″ and B4′ by the same value), resulting in a closed-shell doubly antiaromatic configuration. More generally, the HHM predicts that the introduction of BLA lowers the π energy of both 4n and 4n + 2 systems, with the effect being smaller in 4n + 2 than 4n systems with equal n and increasing with n (details in SI). Therefore, a π-conjugated molecule will have no BLA if its σ contribution outweighs the π contribution, which usually occurs in 4n + 2 molecules with small n (e.g., benzene), in agreement with Shaik’s interpretation based on valence bond theory.51 A more rigorous discussion on the distortivity of π-conjugated systems can be found in refs (52, 53, and 54).
We extend the HHM to excited states by promoting electrons to unoccupied orbitals and adding a state-specific offset Exc to the energy of the newly occupied orbitals, in order to account for orbital relaxation and exchange coupling; in state-averaged fashion, aσ, β″, and β′ are kept equal in all states (and at all values of BLA).
Two limitations of the HHM are that it is valid only at small BLA (δ ≈ 0) and that it only accounts for the changes in orbital energies, meaning that it cannot identify a possible switch from the global to a local current.
2.2. Minimal Ipsocentric Model for C16
To provide a direct relation between aromaticity and orbital energies, we employ a minimal ipsocentric model in which matrix elements in eqs 2 and 3 are replaced by orbital-independent parameters, and only a few transitions are counted. In this approach, the total ring current of C16 is calculated as the sum of its out-of-plane and in-plane components, which are given by
| 6 |
| 7 |
In eqs 6 and 7, cDIAst has a constant positive value for all transitions with Δk = 1 (e.g., A4” → A5”), while cPARAst has a constant positive value for all Δk = 0 transitions; both are zero for all other transitions. The third parameter in the model, c′, relates the in-plane and the out-of-plane matrix elements (more generally, it relates the currents produced in different π systems by an orbital pair with the same Δεst). cDIAst, cPARAst, and c′ are assumed to be independent of BLA and electronic state; for a specific molecule, they reflect how much current is induced by a single transition between orbitals separated by one unit of energy.
Within this minimal model, only orbitals with k = 3–5 contribute to the ring current of C16 (as only Δk = 0 and Δk = 1 transitions are counted). For simplicity, we will focus on transitions involving k = 4 and k = 5 orbitals shown in Figure 2c, as the diatropic contribution of k = 3 → k = 4 transitions changes in the same manner as the contribution from k = 4 → k = 5 transitions.
In principle, the three parameters cDIAst, cPARAst, and c′ (or two, in the case of an annulene) can be obtained by only three (or two) calculations. In the next section, we use a more sophisticated approach based on CASSCF ring currents and HHM orbitals but note that similar values of cDIAst, cPARAst, and c′ can be obtained by only three DFT calculations (see SI).
2.3. CASSCF and HMM results
We now investigate the variation of the ring current with BLA in the S0, Q1, T1, and S1 states of C16 at a series of 11 D8h geometries with different values of BLA. The considered geometries are interpolated between the ground-state minimum previously found by NEVPT217 (BLA = 11.4 pm, ring radius r = 3.33 Å) and the minimum-energy D16h geometry at the same level of theory (BLA = 0 pm, r = 3.32 Å); results for geometries extended to BLA up to 16 pm, which roughly corresponds to the CASSCF minimum at r = 3.32 Å, are given in Figure 8.
Figure 8.
Comparison between (a) energies and (b) ring current strengths (b) predicted by CASSCF (large symbols) and the HHM (dotted lines). Circles correspond to the S0 state, hollow triangles correspond to S1, full triangles correspond to T1, and pentagons correspond to Q1. Aromatic and antiaromatic ring currents are shown in green and purple, respectively.
CASSCF calculations include 12 electrons in 12 orbitals with k = 3–5 in the active space, capturing all Δk = 0 and Δk = 1 transitions. These CASSCF(12,12) energies (ECAS) and ring curents (JCAS), calculated from nucleus independent chemical shifts (NICS(2)zz) are used to obtain the optimal parameters for the HHM and the minimal ipsocentric model (EHHM, JHHM). In both cases, J is expressed relative to the benzene ring current (Jref = 12 nA/T).
In the following section, different configurations are named according to the occupancies of their four k = 4 orbitals17 (Figure 3), e.g., S0 at large BLA (Figure 2a bottom) can be written as |20 20>. For each electronic state, the magnetic couplings in the most relevant configuration are described. The effect of dynamic correlation is evaluated by comparing the CASSCF and QD-NEVPT255 wave function composition, as well as by density functional theory (DFT) calculations (ωB97XD/def2-TZVP, Figure S1 in SI).
Figure 3.

Most important configurations in the investigated electronic states of C16.
2.4. S0 State
In the ground state, the dominant configuration at virtually all nonzero BLA values is the doubly antiaromatic |20 20> (Figure 4a,d). In this configuration, A4 orbitals in both π systems are doubly occupied and their B4 counterparts are doubly unoccupied, leading to A4″→ B4″ and A4′ → B4′ rotational transitions (Figure 4b).
Figure 4.
Aromaticity of C16 in the ground (S0) state. (a) Wave function composition at different BLA, as calculated by QD-NEVPT2 (top) and CASSCF (bottom). (b) Translational (green) and rotational (purple) transitions in k = 4–5 orbitals at large and zero BLA in the |20 20> configuration. Full and dashed lines show larger and smaller contributions to the ring current. (c) Total ring current and relative energy at different BLA calculated by CASSCF (circles) and the HHM (dashed lines). (d) In-plane (J′) and out-of-plane (J″) contributions to the total ring current.
Decreasing BLA reduces the energy difference between A4 and B4, which rapidly increases antiaromaticity and electronic energy (Figure 4c; we define Erel = 0 for BLA = 11.4 pm). Despite a strong multireference character at low BLA (Figure 4a), HHM reproduces the CASSCF results well (Figure 4c), with low mean absolute errors for energy MAEE = 0.05 eV and ring current (MAEJ = 1.0).
There is a complete absence of the doubly aromatic |22 00> configuration, even at zero BLA, which is a consequence of a small separation γ between the two π systems.17 This is in contrast with single-reference methods and HHM, which incorrectly predict a |2200> ground state for the BLA = 0 configuration.
2.5. Q1 State
The major configuration in the lowest quintet state is 5|11 11>, in which the four k = 4 orbitals are singly occupied by four same-spin electrons. This wave function composition remains largely unaffected by changes in BLA (Figure 5a). No rotational transitions are allowed, leading to double aromaticity (Figure 5b,d).46
Figure 5.
Aromaticity of C16 in the lowest quintet (Q1) state. (a) Wave function composition at different BLA, as calculated by QD-NEVPT2 (top) and CASSCF (bottom). (b) Translational (green) and rotational (purple) transitions in k = 4–5 orbitals at large and zero BLA in the |11 11> configuration. Full and dashed lines show larger and smaller contributions to the ring current. (c) Total ring current and relative energy at different BLA calculated by CASSCF (circles) and the HHM (dashed lines). (d) In-plane (J′) and out-of-plane (J″) contributions to the total ring current.
The Q1 ring current does not change much (∼10%) with BLA, as the increase in the energy of the A4 orbitals is offset by a decrease in B4 energies. At BLA = 0, the energy is minimized and the ring current is maximized (Figure 5c), which is a complete reversal of the result obtained for the doubly antiaromatic S0 (cf. Figures 4c and 5c). A very similar result is obtained with DFT (Figure S1b).
The HHM closely reproduces the electronic energies obtained by CASSSCF (Figure 5c; MAEE = 0.12 eV). The ring current fit is of similar accuracy to S0 (MAEJ = 0.9), but it is limited by the inability of the minimal ipsocentric model to describe variation not driven by changes in orbital energies.
2.6. T1 State
In a naïve single-reference picture, the lowest lying triplet state (T1) might be obtained by changing the spin of a single electron in the S0 |20 20> configuration. For example, flipping the spin of an electron in the out-of-plane π system results in the |11 20> configuration, which is predicted by both CASSCF and NEVPT2 to have the largest contribution to the multireference wave function (Figure 6a).
Figure 6.
Aromaticity of C16 in the lowest triplet (T1) state. (a) Wave function composition at different BLA, as calculated by QD-NEVPT2 (top) and CASSCF (bottom). (b) Translational (green) and rotational (purple) transitions in k = 4–5 orbitals at large and zero BLA in the |11 20> configuration. Full and dashed lines show larger and smaller contributions to the ring current. (c) Total ring current and relative energy at different BLA calculated by CASSCF (circles) and the HHM (dashed lines). (d) In-plane (J′) and out-of-plane (J″) contributions to the total ring current.
In |11 20>, the out-of-plane |11> π system produces a diatropic current, which does not depend much on the BLA (analogously to Q1; Figure 6b left), while the paratropic current produced by the in-plane |20> π system strongly increases with decreasing BLA (analogously to S0; Figure 6b). At this point, very small changes in energy (e.g., 70 meV in either direction, which corresponds to the expectation value of the value of position along the Kekulé vibration at 0 K) result in drastic changes in the ring current (from +0.9 to −2.3 relative to benzene). Unless a measurement with femtosecond resolution can be made, we will therefore observe an average paratropic current (around −1.4 relative to benzene); however, this result illustrates the extreme variation in the magnetic susceptibility with small changes in energy and geometry.
The HHM qualitatively recovers the variation of energy with BLA, and reproduces the ring current changes with relatively high accuracy (Figure 6c; MAEE = 0.09 eV; MAEJ = 0.3). The contribution of other notable configurations, such as |20 11> (in which an in-plane electron was flipped), is analogous to |11 20>, with a combination of paratropic and diatropic contributions. This is demonstrated by DFT, which predicts similar variation J with BLA for both cases (Figure S1c), in qualitative agreement with CASSCF.
2.7. S1 State
Unlike all previously considered states, the first excited singlet (S1) is antisymmetric with respect to the reflection of the molecular plane. This means it consists of configurations with an odd number of electrons in both π systems and can be described as a pair of doublets, with, e.g., 17 in-plane and 15 out-of-plane, or 15 in-plane and 17 out-of-plane electrons.
At nonzero BLA, both CASSCF and NEVPT2 predict |21 10>, which has 17 out-of-plane and 15 in-plane electrons, as the dominant configuration (Figure 7a). In |21 10>, a combination of rotational and translational transitions is present both π systems (Figure 7b), resulting in mixed aromaticity (Figure 7d). At large BLA (>9.1 pm), translational contributions are stronger, resulting in an overall diatropic current. At smaller (<9.1 pm) BLA, the energy difference between A4 and B4 orbitals sufficiently shrinks to induce a reversal of the ring current (Figure 7c).
Figure 7.
Aromaticity of C16 in the first excited singlet (S1) state. (a) Wave function composition at different BLA, as calculated by QD-NEVPT2 (top) and CASSCF (bottom). (b) Translational (green) and rotational (purple) transitions in k = 4–5 orbitals at large and zero BLA in the |21 10> configuration. Full and dashed lines show larger and smaller contributions to the ring current. (c) Total ring current and relative energy at different BLA calculated by CASSCF (circles) and the HHM (dashed lines). (d) In-plane (J′) and out-of-plane (J″) contributions to the total ring current.
The similarity in the variation of energy and ring current in the T1 (Figure 6) and S1 (Figure 7) states can be rationalized by noting that their energies are approximately related by ES1 = ET1 + γ – EEX, where γ is the offset between the in-plane and out-of-plane orbitals and EEX is the exchange interaction. Analogously, ring currents in T1 and S1 are similar because the number of diatropic and paratropic transitions in both is equal (cf. Figures 6b and 7b). At BLA = 9.1, the ring current in S1 is slightly more sensitive to small changes in energy, with a 70 meV variation in either direction changing J from +0.8 to −3.5.
The HHM fit to CASSCF energies and ring current is comparable to that of other states (MAEE = 0.06 eV; MAEJ = 0.6), indicating that the HHM and the minimal ipsocentric approach can be extended to excited states.
2.8. Qualitative Modeling
The state-averaged HHM is successful in describing the variation of energy with BLA in the four lowest-lying electronic states of C16 (Figure 8a), with MAEE = 0.09 eV. The HHM orbital separation γ is 0.2 eV, in good agreement with the γ = 0.1 eV value previously found by DFT.17 We also obtain a ∼38% stronger Hückel coupling between the out-of-plane (β” = 6.1 eV) than the in-plane (β′ = 4.4 eV) orbitals, indicating stronger overlap in the out-of-plane π system.
The frontier orbital ipsocentric approach accurately reproduces the aromaticity switching predicted by CASSCF, remaining remarkably robust through a large range of ring current strengths (Figure 8b), with a mean relative error of 11.9%. Its success validates the approximation that variation in the magnetic response can be recovered purely through the change in frontier orbital energies. The coupling constants we obtain are cDIA = 6.05 Jref/eV and cPARA = 8.05 Jref/eV, providing a simple, direct link between orbital energy and aromaticity through eqs 6 and 7. In agreement with previous work,20,46,56 our minimal ipsocentric approach finds a weaker magnetic coupling in the in-plane π system (c′ = 0.55) relative to the out-of-plane system. Further details on the HHM and the minimal ipsocentric approach are given in the SI.
While we have only focused on the effect of BLA, aromaticity reversals may occur with any distortion that causes a stronger response in the orbital energies than in their spatial overlap. For example, bond-angle alternation (BAA) could be described by adding a βBAA(1 ± δBAA) term to the in-plane π system. In cases where no significant open-shell character is present, the minimal ipsocentric approach could employ orbital energies calculated by DFT, making the calculation of coupling constants that relate aromaticity and frontier orbital energy very simple.
The addition of dynamic correlation does not seem to have a significant effect on the ring currents: DFT also predicts an aromaticity reversal in T1 (Figure S1c), and perturbatively including coupling with unoccupied orbitals outside the active space (QD-NEVPT2) does not significantly change the composition (Figures 3a–6a) of the CASSCF wave function. This is in agreement with previous work comparing the performance of HF and DFT, which concluded that the computed current depends more strongly on the used geometry than on the level of theory.57,58
2.9. Divide-and-Conquer Approach
Due to their orthogonality, the two π systems in cyclocarbons are usually considered separately.20,46,56 This suggests that one could approximate a 12,12 active space, which is used throughout this work, with a combination of in-plane and out-of-plane 6,6 active spaces. Here, we compute the energy of the two 6,6 subspaces (E′qUCCSD and E″qUCCSD) using the q-UCCSD* method, which includes all single and double excitations between all spin–orbitals in the active space (further details are given in the SI). The total energy is calculated according to
| 8 |
where Einactive(12,12) is the complete active space configuration interaction inactive energy, which can be obtained efficiently without solving the entire active space. It should be noted that in each of the two q-UCCSD(6,6) calculations, the correlated π-system feels the static orbitals of the noncorrelated π-system, i.e., the two π-system systems remain coupled at a mean-field level.
In the S0 state, this divide-and-conquer q-UCCSD approach works well for a wide range of nonzero BLA (Figure 9, MAE = 0.16 eV relative to CASSCF), producing results very similar to those of canonical CCSD and CCSD(T) methods. Around zero BLA, both q-UCCSD and canonical coupled clusters are limited by the poor quality of the underlying HF determinant, although q-UCCSD appears to be more robust in case of BLA ≈ 1 pm. In these cases, the performance of q-UCCSD may be further improved by starting from a better reference (e.g., CASSCF) or by using a flavor which includes orbital optimization.59 q-UCCSD calculations for the T1 and Q1 states are of similar quality to S0 (see the SI), illustrating a way for applying the q-UCCSD method to systems with many strongly correlated electrons (other possible use cases may be a complex with two weakly coupled transition metals, or a dye with two chromophores connected with sigma bonds).
Figure 9.

Energies of the S0 state C16 at different amounts of BLA calculated using CASSCF (gray circles), CCSD (blue hollow triangles), CCSD(T) (green hollow triangles), and divide-and-conquer q-UCCSD (red rectangles), all obtained with the cc-pVDZ basis set. Note the two breaks on the y axis.
3. Conclusions
Using CASSCF, qualitative modeling, and DFT, we have demonstrated that the ring current in C16 can be reversed by small changes in BLA. This reversal of aromaticity (according to the ring-current criterion2) is unique as it is not associated by any notable change in the electronic structure, but only with a small change in energy. It occurs in electronic states displaying mixed aromaticity, which can be characterized by the presence of both diatropic and paratropic currents (Table 1).
Table 1. Ring Current of C16 in Low-Living Electronic States at Different Amounts of Bond-Length Alternation.
|
net ring current |
|||
|---|---|---|---|
| state | electronic structure | BLA ≫ 0 | BLA ≈ 0 |
| S0 | doubly antiaromatic | paratropic | paratropic |
| Q1 | doubly Baird aromatic | diatropic | diatropic |
| T1 | Baird aromatic + antiaromatic | diatropic | paratropic |
| S1 | doubly mixed | diatropic | paratropic |
The Hückel–Heilbronner tight-binding model captures the essence of the electronic structure of C16, qualitatively reproducing the variation of energy obtained by much more complex methods (CASSCF and range-separated DFT). Magnetic couplings between the C16 orbitals are successfully described by using a minimal ipsocentric model. This model provides a direct link between aromaticity and orbital energy, thus offering a simple avenue for rationalizing commonly observed correlations between aromaticity and various molecular properties.8,9,10,11
Aromaticity reversals may also be interpreted in terms of antiaromaticity relief.30,60 The doubly aromatic Q1 state adopts a BLA = 0 geometry, which maximizes its aromaticity, while the doubly antiaromatic S0 minimizes its antiaromaticity by increasing BLA as much as the σ system allows. The mixed aromatic T1 and S1 states have approximately nonaromatic D8h minima, revealing their aromatic and antiaromatic nature by movement along one and the other directions of the Kekulé vibration.
In any system with more than one π pathway, the ring currents produced by individual circuits may be opposed, leading to a possibility of aromaticity reversals of the type discussed here. More generally, we can expect the behavior of such “mixed aromatic” systems to be very sensitive to changes in geometry.
We also demonstrate that the applicability of quantum algorithms such as q-UCCSD can be improved by partitioning the active space and calculating the correlation energy of each subspace separately. In the case of C16, this divide-and-conquer approach is particularly successful, approaching the accuracy of fully self-consistent multireference calculations. Finally, as quantum devices have direct access to the wave function, it is straightforward to compute the expectation value of any observable. This suggests that the ipsocentric approach could be combined with virtually any quantum ansatz.
Acknowledgments
We thank David Tew, Shantanu Mishra, and Florian Albrecht for discussions. This work was supported by the ERC (grant 885606 ARO-MAT and grant 951519 MolDAM) and the European Commision (MSCA project 101064401 ElDelPath). Computational resources were provided by the Cirrus UK National Tier-2 HPC Service at EPCC (http://www.cirrus.ac.uk) funded by the University of Edin-burgh and EPSRC (EP/P020267/1), as well as the Ministry of Education, Youth and Sports of the Czech Republic through the e-INFRA CZ (ID: 90140).
Supporting Information Available
The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacs.3c10207.
Molecular geometries, DFT results for the S0, T1, and Q1 states, and MATLAB code used for the HHM (PDF)
Author Contributions
§ These authors contributed equally.
The authors declare no competing financial interest.
Supplementary Material
References
- Merino G.; Solà M.; Fernández I.; Foroutan-Nejad C.; Lazzeretti P.; Frenking G.; Anderson H. L.; Sundholm D.; Cossío F. P.; Petrukhina M. A.; et al. Aromaticity: Quo vadis. Chem. Sci. 2023, 14, 5569–5576. 10.1039/D2SC04998H. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gershoni-Poranne R.; Stanger A. Magnetic criteria of aromaticity. Chem. Soc. Rev. 2015, 44, 6597–6615. 10.1039/C5CS00114E. [DOI] [PubMed] [Google Scholar]
- Schleyer P. v. R. Introduction: Aromaticity. Chem. Rev. 2001, 101, 1115–1118. 10.1021/cr0103221. [DOI] [PubMed] [Google Scholar]
- Martín N.; Scott L. T. Challenges in aromaticity: 150 years after Kekulé’s benzene. Chem. Soc. Rev. 2015, 44, 6397–6400. 10.1039/C5CS90085A. [DOI] [PubMed] [Google Scholar]
- Merino G.; Solà M. Celebrating the 150th anniversary of the kekulé benzene structure. Phys. Chem. Chem. Phys. 2016, 18, 11587–11588. 10.1039/C6CP90088G. [DOI] [PubMed] [Google Scholar]
- Bradley D.; Jirásek M.; Anderson H. L.; Peeks M. D. Disentangling global and local ring currents. Chem. Sci. 2023, 14, 1762–1768. 10.1039/D2SC05923A. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Stanger A.; Monaco G.; Zanasi R. NICS-XY-scan predictions of local, semi-global, and global ring currents in annulated pentalene and s-indacene cores compared to first-principles current density maps. ChemPhysChem 2020, 21, 65–82. 10.1002/cphc.201900952. [DOI] [PubMed] [Google Scholar]
- Gershoni-Poranne R.; Rahalkar A. P.; Stanger A. The predictive power of aromaticity: Quantitative correlation between aromaticity and ionization potentials and HOMO–LUMO gaps in oligomers of benzene, pyrrole, furan, and thiophene. Phys. Chem. Chem. Phys. 2018, 20, 14808–14817. 10.1039/C8CP02162G. [DOI] [PubMed] [Google Scholar]
- Frederickson C. K.; Rose B. D.; Haley M. M. Explorations of the indenofluorenes and expanded quinoidal analogues. Acc. Chem. Res. 2017, 50, 977–987. 10.1021/acs.accounts.7b00004. [DOI] [PubMed] [Google Scholar]
- Stanger A. Singlet fission and aromaticity. J. Phys. Chem. A 2022, 126, 8049–8057. 10.1021/acs.jpca.2c04146. [DOI] [PubMed] [Google Scholar]
- Xin S.; Han Y.; Fan W.; Wang X.; Ni Y.; Wu J. Enhanced aromaticity and open-shell diradical character in the dianions of 9-fluorenylidene-substituted expanded radialenes. Angew. Chem., Int. Ed. 2022, 61, e202209448 10.1002/anie.202209448. [DOI] [PubMed] [Google Scholar]
- Jirásek M.; Anderson H. L.; Peeks M. D. From macrocycles to quantum rings: Does aromaticity have a size limit?. Acc. Chem. Res. 2021, 54, 3241–3251. 10.1021/acs.accounts.1c00323. [DOI] [PubMed] [Google Scholar]
- Anderson H. L.; Patrick C. W.; Scriven L. M.; Woltering S. L. A short history of cyclocarbons. Bull. Chem. Soc. Jpn. 2021, 94, 798–811. 10.1246/bcsj.20200345. [DOI] [Google Scholar]
- Baryshnikov G. V.; Valiev R. R.; Nasibullin R. T.; Sundholm D.; Kurten T.; Ågren H. Aromaticity of even-number cyclo[n]carbons (n = 6–100). J. Phys. Chem. A 2020, 124, 10849–10855. 10.1021/acs.jpca.0c09692. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Baryshnikov G. V.; Valiev R. R.; Kuklin A. V.; Sundholm D.; Ågren H. Cyclo[18]carbon: Insight into electronic structure, aromaticity, and surface coupling. J. Phys. Chem. Lett. 2019, 10, 6701–6705. 10.1021/acs.jpclett.9b02815. [DOI] [PubMed] [Google Scholar]
- Sun L.; Zheng W.; Gao W.; Kang F.; Zhao M.; Xu W.. Aromatic annular carbon allotropes: Cumulenic cyclo[10]carbon and Peierls-transition-intermediate cyclo[14]carbon. Research Square, 2023, DOI: 10.21203/rs.3.rs-2616838/v2. [DOI]
- Gao Y.; Albrecht F.; Rončević I.; Ettedgui I.; Kumar P.; Scriven L. M.; Christensen K. E.; Mishra S.; Righetti L.; Rossmannek M.; Tavernelli I.; Anderson H. L.; Gross L.. On-surface synthesis of a doubly anti-aromatic carbon allotrope Nature 2023 10.1038/s41586-023-06566-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kaiser K.; Scriven L. M.; Schulz F.; Gawel P.; Gross L.; Anderson H. L. An sp-hybridized molecular carbon allotrope, cyclo[18]carbon. Science 2019, 365, 1299–1301. 10.1126/science.aay1914. [DOI] [PubMed] [Google Scholar]
- Bean D. E.; Fowler P. W. Effect on ring current of the kekulé vibration in aromatic and antiaromatic rings. J. Phys. Chem. A 2011, 115, 13649–13656. 10.1021/jp2077424. [DOI] [PubMed] [Google Scholar]
- Baryshnikov G. V.; Valiev R. R.; Valiulina L. I.; Kurtsevich A. E.; Kurtén T.; Sundholm D.; Pittelkow M.; Zhang J.; Ågren H. Odd-number cyclo[n]carbons sustaining alternating aromaticity. J. Phys. Chem. A 2022, 126, 2445–2452. 10.1021/acs.jpca.1c08507. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Karadakov P. B. Aromaticity and antiaromaticity in the low-lying electronic states of cyclooctatetraene. J. Phys. Chem. A 2008, 112, 12707–12713. 10.1021/jp8067365. [DOI] [PubMed] [Google Scholar]
- Karadakov P. B.; Di M.; Cooper D. L. Excited-state aromaticity reversals in möbius annulenes. J. Phys. Chem. A 2020, 124, 9611–9616. 10.1021/acs.jpca.0c08594. [DOI] [PubMed] [Google Scholar]
- Karadakov P. B.; Preston N. Aromaticity reversals and their effect on bonding in the low-lying electronic states of cyclooctatetraene. Phys. Chem. Chem. Phys. 2021, 23, 24750–24756. 10.1039/D1CP04394C. [DOI] [PubMed] [Google Scholar]
- Rickhaus M.; Jirasek M.; Tejerina L.; Gotfredsen H.; Peeks M. D.; Haver R.; Jiang H.-W.; Claridge T. D. W.; Anderson H. L. Global aromaticity at the nanoscale. Nat. Chem. 2020, 12, 236–241. 10.1038/s41557-019-0398-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Peeks M. D.; Jirasek M.; Claridge T. D. W.; Anderson H. L. Global aromaticity and antiaromaticity in porphyrin nanoring anions. Angew. Chem., Int. Ed. 2019, 58, 15717–15720. 10.1002/anie.201909032. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sokolov A. Y.; Magers D. B.; Wu J. I.; Allen W. D.; Schleyer P. V. R.; Schaefer H. F. III. Free cyclooctatetraene dianion: Planarity, aromaticity, and theoretical challenges. J. Chem. Theory Comput. 2013, 9, 4436–4443. 10.1021/ct400642y. [DOI] [PubMed] [Google Scholar]
- Jiang Y.; Wu Y.; Deng J.; Wang Z. Antiaromaticity–aromaticity transition of cyclo[16]carbon upon metal encapsulation. Phys. Chem. Chem. Phys. 2021, 23, 8817–8824. 10.1039/D0CP06256A. [DOI] [PubMed] [Google Scholar]
- Li S.; Sun Y.; Li X.; Smaga O.; Koniarz S.; Pawlicki M.; Chmielewski P. J. Oxidative insertion of amines into conjugated macrocycles: Transformation of antiaromatic norcorrole into aromatic azacorrole. Chem. Commun. 2023, 59, 3739–3742. 10.1039/D2CC06648C. [DOI] [PubMed] [Google Scholar]
- Gibson C. M.; Havenith R. W. A.; Fowler P. W.; Jenneskens L. W. Planar homotropenylium cation: A transition state with reversed aromaticity. J. Org. Chem. 2015, 80, 1395–1401. 10.1021/jo502179k. [DOI] [PubMed] [Google Scholar]
- Slanina T.; Ayub R.; Toldo J.; Sundell J.; Rabten W.; Nicaso M.; Alabugin I.; Fdez; Galván I.; Gupta A. K.; Lindh R.; et al. Impact of excited-state antiaromaticity relief in a fundamental benzene photoreaction leading to substituted bicyclo[3.1.0]hexenes. J. Am. Chem. Soc. 2020, 142, 10942–10954. 10.1021/jacs.9b13769. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Saha P. K.; Mallick A.; Turley A. T.; Bismillah A. N.; Danos A.; Monkman A. P.; Avestro A.-J.; Yufit D. S.; McGonigal P. R. Rupturing aromaticity by periphery overcrowding. Nat. Chem. 2023, 15, 516–525. 10.1038/s41557-023-01149-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sankar J.; Mori S.; Saito S.; Rath H.; Suzuki M.; Inokuma Y.; Shinokubo H.; Suk Kim K.; Yoon Z. S.; Shin J.-Y.; et al. Unambiguous identification of möbius aromaticity for meso-aryl-substituted [28]hexaphyrins(1.1.1.1.1.1). J. Am. Chem. Soc. 2008, 130, 13568–13579. 10.1021/ja801983d. [DOI] [PubMed] [Google Scholar]
- Torrent-Sucarrat M.; Navarro S.; Cossío F. P.; Anglada J. M.; Luis J. M. Relevance of the DFT method to study expanded porphyrins with different topologies. J. Comput. Chem. 2017, 38, 2819–2828. 10.1002/jcc.25074. [DOI] [PubMed] [Google Scholar]
- Bauer B.; Bravyi S.; Motta M.; Chan G. K.-L. Quantum algorithms for quantum chemistry and quantum materials science. Chem. Rev. 2020, 120, 12685–12717. 10.1021/acs.chemrev.9b00829. [DOI] [PubMed] [Google Scholar]
- Li W.; Huang Z.; Cao C.; Huang Y.; Shuai Z.; Sun X.; Sun J.; Yuan X.; Lv D. Toward practical quantum embedding simulation of realistic chemical systems on near-term quantum computers. Chem. Sci. 2022, 13, 8953–8962. 10.1039/D2SC01492K. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Barkoutsos P. K.; Gonthier J. F.; Sokolov I.; Moll N.; Salis G.; Fuhrer A.; Ganzhorn M.; Egger D. J.; Troyer M.; Mezzacapo A.; et al. Quantum algorithms for electronic structure calculations: Particle-hole hamiltonian and optimized wave-function expansions. Phys. Rev. A 2018, 98, 022322 10.1103/PhysRevA.98.022322. [DOI] [Google Scholar]
- Peruzzo A.; McClean J.; Shadbolt P.; Yung M.-H.; Zhou X.-Q.; Love P. J.; Aspuru-Guzik A.; O’Brien J. L. A variational eigenvalue solver on a photonic quantum processor. Nat. Commun. 2014, 5, 4213. 10.1038/ncomms5213. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Steiner E.; Fowler P. W. On the orbital analysis of magnetic properties. Phys. Chem. Chem. Phys. 2004, 6, 261–272. 10.1039/B312289C. [DOI] [Google Scholar]
- Steiner E.; Fowler P. W. Four- and two-electron rules for diatropic and paratropic ring currents in monocyclic π systems. Chem. Commun. 2001, 2220–2221. 10.1039/b104847n. [DOI] [PubMed] [Google Scholar]
- Steiner E.; Fowler P. W. Patterns of ring currents in conjugated molecules: A few-electron model based on orbital contributions. J. Phys. Chem. A 2001, 105, 9553–9562. 10.1021/jp011955m. [DOI] [Google Scholar]
- Wenthold P. G.; Hrovat D. A.; Borden W. T.; Lineberger W. C. Transition-state spectroscopy of cyclooctatetraene. Science 1996, 272, 1456–1459. 10.1126/science.272.5267.1456. [DOI] [PubMed] [Google Scholar]
- Karadakov P. B.; Gerratt J.; Cooper D. L.; Raimondi M. The electronic structure of cyclooctatetraene and the modern valence-bond understanding of antiaromaticity. J. Phys. Chem. 1995, 99, 10186–10195. 10.1021/j100025a020. [DOI] [Google Scholar]
- Baird N. C. Quantum organic photochemistry. II. Resonance and aromaticity in the lowest 3.Pi.Pi.* state of cyclic hydrocarbons. J. Am. Chem. Soc. 1972, 94, 4941–4948. 10.1021/ja00769a025. [DOI] [Google Scholar]
- Rosenberg M.; Dahlstrand C.; Kilså K.; Ottosson H. Excited state aromaticity and antiaromaticity: Opportunities for photophysical and photochemical rationalizations. Chem. Rev. 2014, 114, 5379–5425. 10.1021/cr300471v. [DOI] [PubMed] [Google Scholar]
- Karadakov P. B. Norcorrole: Aromaticity and antiaromaticity in contest. Org. Lett. 2020, 22, 8676–8680. 10.1021/acs.orglett.0c03254. [DOI] [PubMed] [Google Scholar]
- Bean D. E.; Fowler P. W.; Soncini A. Double aromaticity and ring currents in open-shell systems. Chem. Phys. Lett. 2009, 483, 193–197. 10.1016/j.cplett.2009.10.089. [DOI] [Google Scholar]
- Heilbronner E. Why do some molecules have symmetry different from that expected?. J. Chem. Educ. 1989, 66, 471–478. 10.1021/ed066p471. [DOI] [Google Scholar]
- Stanger A. The different aromatic characters of some localized benzene derivatives. J. Phys. Chem. A 2008, 112, 12849–12854. 10.1021/jp801634x. [DOI] [PubMed] [Google Scholar]
- Soncini A.; Domene C.; Engelberts J. J.; Fowler P. W.; Rassat A.; van Lenthe J. H.; Havenith R. W. A.; Jenneskens L. W. A unified orbital model of delocalised and localised currents in monocycles, from annulenes to azabora-heterocycles. Chem.—Eur. J. 2005, 11, 1257–1266. 10.1002/chem.200400678. [DOI] [PubMed] [Google Scholar]
- Monino E.; Boggio-Pasqua M.; Scemama A.; Jacquemin D.; Loos P.-F. Reference energies for cyclobutadiene: Automerization and excited states. J. Phys. Chem. A 2022, 126, 4664–4679. 10.1021/acs.jpca.2c02480. [DOI] [PubMed] [Google Scholar]
- Shaik S.; Shurki A.; Danovich D.; Hiberty P. C. A different story of π-delocalizationthe distortivity of π-electrons and its chemical manifestations. Chem. Rev. 2001, 101, 1501–1540. 10.1021/cr990363l. [DOI] [PubMed] [Google Scholar]
- Binsch G.; Heilbronner E.; Murrell J. N. The theory of double bond fixation in conjugated hydrocarbons. Mol. Phys. 1966, 11, 305–320. 10.1080/00268976600101141. [DOI] [Google Scholar]
- Fowler P. W.Symmetry aspects of distortivity in π systems. In Adv. Quantum chem., Vol. 44; Academic Press, 2003; pp 219–237 10.1016/S0065-3276(03)44014-8. [DOI] [Google Scholar]
- Fowler P. W.; Rassat A. Symmetry and distortive π-electrons in two- and three-dimensional conjugated systems. Phys. Chem. Chem. Phys. 2002, 4, 1105–1113. 10.1039/b109712c. [DOI] [Google Scholar]
- Angeli C.; Borini S.; Cestari M.; Cimiraglia R. A quasidegenerate formulation of the second order n-electron valence state perturbation theory approach. J. Chem. Phys. 2004, 121, 4043–4049. 10.1063/1.1778711. [DOI] [PubMed] [Google Scholar]
- Fowler P. W.; Mizoguchi N.; Bean D. E.; Havenith R. W. A. Double aromaticity and ring currents in all-carbon rings. Chem.—Eur. J. 2009, 15, 6964–6972. 10.1002/chem.200900322. [DOI] [PubMed] [Google Scholar]
- Soncini A.; Teale A. M.; Helgaker T.; De Proft F.; Tozer D. J. Maps of current density using density-functional methods. J. Chem. Phys. 2008, 129, 074101 10.1063/1.2969104. [DOI] [PubMed] [Google Scholar]
- Havenith R. W. A.; Meijer A. J. H. M.; Irving B. J.; Fowler P. W. Comparison of ring currents evaluated consistently at density functional and Hartree–Fock levels. Mol. Phys. 2009, 107, 2591–2600. 10.1080/00268970903449396. [DOI] [Google Scholar]
- Sokolov I. O.; Barkoutsos P. K.; Ollitrault P. J.; Greenberg D.; Rice J.; Pistoia M.; Tavernelli I. Quantum orbital-optimized unitary coupled cluster methods in the strongly correlated regime: Can quantum algorithms outperform their classical equivalents?. J. Chem. Phys. 2020, 152, 124107. 10.1063/1.5141835. [DOI] [PubMed] [Google Scholar]
- Zhang R.; Ellern A.; Winter A. H. Anti-aromaticity relief as an approach to stabilize free radicals. Angew. Chem., Int. Ed. 2021, 60, 25082–25088. 10.1002/anie.202110870. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.







