Scholarly article on topic 'Double Pancake Bonds: Pushing the Limits of Strong π–π Stacking Interactions'

Double Pancake Bonds: Pushing the Limits of Strong π–π Stacking Interactions Academic research paper on "Chemical sciences"

0
0
Share paper
Academic journal
J. Am. Chem. Soc.
OECD Field of science

Academic research paper on topic "Double Pancake Bonds: Pushing the Limits of Strong π–π Stacking Interactions"

TACS

JOURNAL Of THE AMERICAN CHEMICAL SOCIETY

Subscriber access provided by Aston University Library & Information Services

Article

Double pancake bonds: pushing the limits of strong #-# stacking interactions

Zhong-hua Cui, Hans Lischka, Habtamu Z. Beneberu, and Miklos Kertesz

J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/ja505624y • Publication Date (Web): 24 Aug 2014 Downloaded from http://pubs.acs.org on August 26, 2014

Just Accepted

"Just Accepted" manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides "Just Accepted" as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. "Just Accepted" manuscripts appear in full in PDF format accompanied by an HTML abstract. "Just Accepted" manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). "Just Accepted" is an optional service offered to authors. Therefore, the "Just Accepted" Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the "Just Accepted" Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these "Just Accepted" manuscripts.

ACS Publications

High quality. High impact. Journal of the American Chemical Society is published by the American Chemical

Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

4 Double pancake bonds: pushing the limits of strong n-n stacking interactions

7 Zhong-hua Cui,a Hans Lischka, c Habtamu Z. Beneberua' and Miklos Kertesz*a

9 aDepartment of Chemistry, Georgetown University, 37th & O Streets, NW, Washington, DC 20057-1227, USA.

10 bDepartment of Chemistry and Biochemistry, Texas Tech University, Lubbock, TX 79409-1061,USA.

11 c Institute for Theoretical Chemistry, University of Vienna, A-1090 Vienna, Austria.

12 dDepartment of Chemistry, University of the District of Columbia, Washington, DC 20008,USA.

16 ABSTRACT: The concept of a double-bonded pancake bonding mechanism is introduced to 18

19 explain the extremely short n-n stacking contacts in dimers of dithiatriazines. While ordinary

21 single pancake bonds occur between radicals and already display significantly shorter

24 interatomic distances in comparison to van der Waals (vdW) contacts, the double-bonded

26 pancake dimer is based on diradicaloid or antiaromatic molecules and exhibits even shorter and

28 stronger intermolecular bonds that breach into the range of extremely stretched single bonds in

31 terms of bond distances and binding energies. These properties give rise to promising

33 possibilities in the design of new materials with high electrical conductivity and for the field of

38 correlation theory combining multireference (non-dynamical) effects and dispersion (dynamical)

40 contributions in a balanced way providing accurate interaction energies and distributions of

45 mechanisms operate in several analogous non-radical molecular systems to form double-bonded

47 n-stacking pancake dimers.

52 1. INTRODUCTION

55 n-stacking in radical dimers, some of which are illustrated in Chart 1 in the form of the

57 constituent monomers, is responsible for the formation of a very interesting class of chemical

spintronics. The analysis of the double pancake bond is based on cutting edge electron

unpaired spins. It is also shown that the present examples do not stand isolated but that similar

10 11 12

20 21 22

compounds which display favored packing geometries as described by the maximum multicenter overlap principle between neighboring molecules.1 This preferred orientation is primarily due to the energy lowering of the singly occupied molecular orbital (SOMO) of the radical as it overlaps with its neighbor. This SOMO-SOMO stabilization can be rationalized by the simple molecular orbital (MO) diagram shown in Chart 2a in which the bonding highest occupied molecular orbital (HOMO) is doubly occupied. E.g. for the prototypical phenalenyl (PHY, 1) dimer 12, the efficient n-n overlap provides the driving force for the stabilization of the dimer which is responsible for contact distances significantly shorter and interaction energies larger than for typical van der Waals interactions.3, 4 ^ ^ 7 8 The term "pancake bonding" has been suggested for this type of bonding.9,10 A major motivation in making and understanding these n-

stacking pancake interactions originates in the quest to make new molecular materials with high

electrical conductivity and for spintronics. A crucial condition for the suitability of n-stacked molecules for such purposes is a strong overlap and thus a strong interaction and a short intermolecular bonding distance between the stacked subsystems. High electric conductivities have been achieved for systems with various derivatives of 1, 2 and other n-stacking materials.12 However, there is a strong need to push the limits of pancake interactions to even shorter distances and stronger interactions in order to offer new opportunities for materials design.

.71-/-

N—C S—N

S'^ 1% \ N N 871 \ C-H

N—c'^ S—N

12 3 4

Chart 1. Molecules discussed: phenalenyl radical (1), tetracyanoethylene radical anion (2), 1,2,4,6-thiatriazine radical (3) and 1,3,2,4,6-dithiatriazine (4). 1-3 form single-bonded pancake dimers, 4 forms double-bonded pancake dimers.

6 Therefore, is there a possibility to create even stronger pancake bonds and more attractive

8 interactions? In fact, the answer is yes as the dimer orbital diagrams, shown in Chart 2b or 2c

11 demonstrate. They are based on a monomer with a triplet or a singlet diradicaloid ground state

13 with a low lying-triplet state in the latter case in combination with antiaromatic electron count

16 (8n) and antiaromatic character13. If such a situation exists, a double pancake bonding could arise

18 as a four electron/multicenter (4e/mc) bonding interaction with a formal bond order pMO of 2

20 since two bonding orbitals are doubly occupied. In general we compute the bond order pMO at the

23 MO diagram level as

25 Pmo = ^(Nbind - Nanti), (1)

27 where Nbind is the number of electrons in the bonding orbitals and Nanti is the number of electrons

30 in antibonding orbitals.

32 LUMO L1JMO

34 LUMO ,SOMO' SOMO' \ SOMO'

35 \j_ J ;/LUMO<W

36 n^v , homov;*^ -4K\HOMO'/„>-4-

37 SOMO SOMO SOMO SOMO SOMO X^V SOMO

39 (a) H0M0 (b) HOMO (c) XMO

42 Chart 2. (a) Molecular orbital diagram for single pancake-bonded dimers. (b) Molecular

45 monomer. (c) Molecular orbital diagram for double pancake-bonded dimer based on a

47 singlet diradicaloid ground state of the monomer with a low HOMO-LUMO gap. The

49 formal bond order according to eq (1) is 1 for 12, 22, and 32 (single-bonded pancake) and 2

51 for 42 (double-bonded pancake).

55 The (4e/mc) bond would lead to a significant improvement in terms of the interaction

57 strength in contrast to the formal bond order of 1 for 12 as indicated by Chart 2. Antiaromatic

orbital diagram for double pancake-bonded dimers based on a triplet ground state of the

4 compounds are usually characterized by low stability which makes the search for appropriate

6 candidates for the double pancake bond difficult. Nevertheless, appropriate compounds should

8 exist. A good example is given in the Cambridge Structural Database, CSD14 in the form of the

11 phenyl derivative15 and the 4-chlorophenyl derivative16 of 1,3,2,4,6-dithiatriazine, 4, where the

13 phenyl and 4-chlorophenyl has been replaced by H. It will be compared to an analogous stable 7

16 n-electrons radical, 1,2,4,6-thiatriazine17, 3, that forms a traditional 2e/mc bonded pancake bond.

18 1,3,2,4,6-dithiatriazine (4) is a neutral molecule with 8n-electrons which forms a very short

21 pancake bonded dimer according to its crystal structure.15

24 SOMO' -\ SOMO'

'''-»uiviu "'-^yi^Q^ ',V7IV1V7 , /'' | . .

j >4" ii0M0'^i=tt , i

somo somo somo

iomo IT lr I

31 S(l2,22,32) S(42) T(42) Q(42)

32 Chart 3. MO diagrams for the dimers of various states of 12, 22, 32 and 42. States are

34 designated as S for singlet, T for triplet, and Q for quintet.

38 Can the 1,3,2,4,6-dithiatriazine dimer really be viewed as an example for the double

41 pancake bond and, if so, what can we learn from it for the construction of other and possibly

43 better examples? To answer this question from a theoretical point of view in a thorough way one

45 has to go well beyond the simple MO schemes presented so far. It is crucial to understand the

48 subtle interplay of two kinds of electron correlation effects which make these n-stacking

50 interactions so challenging to understand and design. On one side there is the quasi-degeneracy

53 of the HOMO and LUMO calling for multireference methods for an adequate description

55 whereas on the other side dynamical correlation effects are essential for the description of vdW

interactions.

4 It is the purpose of this contribution to resolve the question of the energetic feasibility of

6 a double pancake bond using the high-level multi reference average quadratic coupled cluster,

8 MR-AQCC, theory . This level of theory provides an excellent approach to the simultaneous

11 treatment of static and dynamic electron correlation. It has been successfully used previously in

13 interpreting the bonding characteristics of the phenalenyl dimer and the TCNE anion dimer ,

15 two prototypical examples of pancake bonding. The multi-reference starting point assures that

18 the multiradical character is included in the theory from the outset, and the approximate coupled

20 cluster level assures that the millions of configurations necessary for the dispersion interaction

20 21 22

22 are well accounted for.

27 2. METHODS AND MODELS

32 Complete active space self-consistent field (CASSCF) and multi-reference averaged coupled

34 cluster MR-AQCC18/6-311++G(2d,2p)24 calculations including full geometry optimizations were

37 carried out on the n dimers 32 and the 42. The electronic state configurations of these two n

39 dimers with C2v symmetry are illustrated in Figure 1. The CASSCF(2,2) (32) and CASSCF(4,4)

41 (42) calculations have been performed using the bonding and antibonding orbitals of the SOMOs

44 in 3 and 4 as the active orbital space for 32 and 42 n dimers, respectively. Molecular orbitals

46 (MOs) created by the CASSCF method were used in the MR-AQCC calculations including

2.1. Computational details

gradients with the same active orbital spaces as used in the CASSCF calculations. The total

10 11 12

20 21 22

HOMO' (bj) HOMO (a,) LUMO (b2) LUMO' (a2)

Figure 1. Illustration of the bonding and antibonding combinations of the two and four SOMOs for 32 (a) and 42 (b), respectively. D is the short intermolecular sulfur-sulfur contact, DS-S.

space of configuration state functions (CSFs) was constructed by applying single and double excitations from valence orbitals to all virtual orbitals for all reference CSFs and imposing generalized interacting space restrictions. 25 The 1s core orbital of C, N and S atoms and 2s and 2p orbital of S atom were frozen throughout all MR-AQCC calculations (additionally, eight low-lying occupied orbitals were frozen in 42). The analysis of the radical character of the complexes was performed by analyzing (i) the natural orbitals (NOs) of the one-particle MR-AQCC density matrix and (ii) the effectively unpaired density using the non-linear formula of Head-Gordon.32 Atomic values are based on a Mulliken analysis for the unpaired density. The COLUMBUS suite of programs was used for the MR-AQCC and CASSCF computations.26 In addition to the single state CASSCF(4,4) approach for the singlet and quintet states, state averaged CASSCF calculations have been performed on the triplet state dominated by two main configurations: Oi=i^ai2bi1a21b20i and O2=l...bi2ai1b21a20l. Density functional theory (DFT) was used to supplement the MR-AQCC calculations for three candidate molecules that are promising for double pancake bonding. Substitution effects were assessed by DFT calculations.

Figures 2 and S1 compare the crystallographic data for the two experimentally observed derivatives of the double bonded pancake dimer 42. The phenyl and chlorophenyl substitution has little effect on the geometry of the dithiatriazine core validating the use of 42 as a good model for these experimentally observed systems.

10 11 12

20 21 22

Figure 2. Structures of two substituted dithiatriazine (HCN3S2)2 n dimers indicate close

similarity of their structures. These dimers were excised from their respective crystal structures: the phenyl derivative15 (52) and the 4-chlorophenyl derivative16 (62) are derivatives of 1,3,2,4,6-dithiatriazine, 4.

For the cationic dimers the Coulomb energy was estimated by using the following formula based on the approximate Qi atomic charges. We use Qi values based on electrostatic potentials (ESP) following the Hu-Lu-Yang charge fitting method (HLY scheme)27 in equation

Ecu =£ QiQj /R,j - c

in which C is taken as the reference Coulomb energy at D = 10.0 A.

The interaction energy En (D) of the dimer with intermolecular separation D between the monomers is computed at the MR-AQCC level as the energy of the complex with reference to the energy at a separation of D=10.0 A where the overlap is sufficiently close to zero: Eint (D) = ETotal(D) - ETotal( 10.0 A). (3)

Further computational details are given in the SI section.

6 2.2. Approximate separation of the interaction energy: vdW and pancake bonding

8 components

11 The separation of the vdW and the attractive SOMO-SOMO interaction is essential for

13 the analysis of the interaction energy, Eint (D). It is written as the sum of the specific pancake n-

15 n bonding SOMO-SOMO interaction (Esomo-somo) and the van der Waals (Evdw) term 3 6 7

18 Eint (D) — esomo-somo (D) + Evd.W (D) . (4)

21 The vdW term includes dispersion, Pauli repulsion, and electrostatic interactions. Evdw is

23 approximated by the interaction energy En

Evdw{D) ~ E»nst(D) . (5)

respectively (equation (1)). The singlet states of 12, 22 and 32 all possess a bond orderp of 1. The

29 computed for the high-spin (HS) state taken at the same distance D since in this case bonding

31 and antibonding interactions derived from the SOMO orbitals approximately cancel and pmo=0,

33 (equation (1)).3, 7 19

36 According to Chart 3, for 42 the singlet (S), triplet (T) and quintet (Q) states contain

38 double, single and no pancake bond character with formal bonders pMO equal to 2, 1 and 0,

43 SOMO-SOMO interaction term for both the single and double pancake bond, respectively, is

45 then approximated as follows:

47 LS HS

48 esomo-somo (D) — ELnSt(D) - EHnSt(D). (6)

51 LS labels the respective low spin component which is only singlet for 12, 22 and 32 but

53 can be singlet or triplet for 42. The high-spin component is triplet and in the former case and

55 quintet in the latter. A version of this approximation restricted to LS = singlet HS = triplet has

58 been used by Mota et al. for the analysis of the interaction energy of pancake bonded dimers of

10 11 12

20 21 22

1 and 2 and has been recently validated for both of these systems within the context of the MR-AQCC level of theory. One result, relevant for this study, was that the vdW term becomes repulsive at the short contacts typical for pancake bonds. According to this analysis, the pancake contacts shorter than the typical vdW distance result from the large negative (bonding) Esomo-somo pancake interaction.

3. RESULTS AND DISCUSSION

3.1. Computed interaction energies and its components of 32 and 42. Total energy minimization of the dimer structures at MR-AQCC level followed by rigid scans as a function of the shortest sulfur- • • sulfur distance (Ds-s) were performed for the n-stacking pancake dimers 32 and 42. The resulting interaction energies are presented in Figure 3a and 3b, while the derived Esomo-somo pancake bonding energy terms based on equation (6) are shown in Figure 3c.

10 11 12

20 21 22

70 60 50 40 30 20 10 0

-10 J 40.0 -i

32 (S) 32 (T)

Ds-s (A) J i A.....I.....I

0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0

J -40.0 -

I -80.0 H

^-120.0 -

-160.0

Ds-S (A)

j-1-1-L-pJ-1—*

5.0 5.5 6.0

42 (Eint Eint) 42 (Eint ~ Eint) 32 (Eint ~ Eint)

Figure 3. Potential energy scans for (a) the singlet and triplet states of 32, and (b) for the singlet, triplet and quintet states of 42. The SOMO-SOMO interaction energies are represented in (c) and are defined in the insert according to equation (6). Computations refer to C2v symmetry using an MR-AQCC/6-311 ++G(2d,2p) level of theory.

The fully optimized geometries and the respective interaction energies are discussed first; key data are collected in Table 1. The singlet minimum of the single pancake bonded 32 shows an interaction energy of -7.0 kcal/mol at DS-S = 2.870 A. This distance is much shorter than the vdW distance of 3.60 A29 and is clearly indicative of pancake bonding. The experimentally

10 11 12

20 21 22

observed contact for the dimers of the diphenyl substituted 3 in the crystal is 2.677 A (CSD refcode CUVTAO)17. The agreement between computation and experiment is good, given the missing steric repulsions due to the phenyls in the model compound used and due to intermolecular interactions in the crystal also not included in the calculations. The interaction energy for the triplet state of 32 at the equilibrium geometry of the singlet (Ds-s = 2.870 A) is repulsive with +11.8 kcal/mol, a value which is used to approximate Evdw at this distance according to equation (5). The SOMO-SOMO binding energy of 32 is -18.8 kcal/mol at the equilibrium geometry computed from equation (6) and represents a significant attraction. On the other hand, the modest attraction of -1.8 kcal/mol at the minimum distance Ds-s of 4.0 A for the triplet state of 32 corresponds well to what is expected of pure vdW interactions both in terms of the location and depth of the minimum.

Table 1. Computeda interaction energies, Eint and its components, E vdw and Esomo-somo.

Speciesb Dc (A) F uint (kcal/mol) FvdW (kcal/mol) FSOMO-SOMO (kcal/mol)

12 (S) 3.104c -11.5 5.7 -17.2

12 (T) 3.676d -3.3 -3.3 0.0

12 (T) 3.104c 5.7 5.7 0.0

22 (S) 2.735c -10.1 13.0 -23.1

22 (T) 3.820d -2.7 -2.7 0.0

22 (T) 2.735c 13.0 13.0 0.0

32 (S) 2.870d -7.0 11.8 -18.8

32 (T) 4.0e -1.8 -1.8 0.0

32 (T) 2.870d 11.8 11.8 0.0

42 (S) 2.571d -27.7 62.5 -90.2

42 (T) 3.6e -2.9 -0.4 -2.5

42 (Q) 2.571d 62.5 62.5 0.0

42 (Q) 3.6e -0.4 -0.4 0.0

42 (Q) 4.1e -1.8 -1.8 0.0

a MR-AQCC/6-311 ++G(2d,2p) level of theory. Data for 12 are from ref. 19and for 22 from ref. 3. b S, T, and Q stand for singlet, triplet and quintet states, respectively. c D represents C-C contacts for 12 and 22 and S-S contacts for the rest of the dimers. d Optimized geometry of the singlet (S) dimer.

e Minimum on the rigid D scan using the singlet ground state geometry of the dimer.

4 Turning to the double pancake case of 42, the singlet minimum shows a much larger

6 interaction energy of -27.7 kcal/mol at a remarkably short contact distance of Ds-s = 2.571 A

8 (Table 1). This distance is considerably shorter (by 0.3 A) than in the radical dimer 32 discussed

11 above and by more than 1 A shorter than the vdW distance of 3.60 A29. The experimentally

13 observed Ds-s distance for the dimers of the phenyl substituted 4 in the crystal is 2.529 A

15 (average of two values from CSD, refcode DESSID)15. The value in the isostructural 417

18 chlorophenyl derivative dimer is 2.522 A (average of two values from CSD, refcode PAFLAJ)16

20 still much longer than the typical single S-S bond of about 2.04 A.16 The agreement between

23 computation and experiment is very good and again the differences are largely attributable to

25 steric repulsions due to the two phenyls (not present in the computations) and to intermolecular

27 interactions in the crystal. The vdW interaction energy computed from equation (5) using the

32 large positive value indicates that the SOMO-SOMO interaction for the singlet with such a short

34 distance must more than overcome this repulsive term. The development of the SOMO-SOMO

37 pancake bonding energy according to equation (6) with the intermolecular distance Ds-s is shown

39 in Figure 3c. At the equilibrium geometry of the singlet of 42 it reaches the amazingly large

41 attractive value of -90.2 kcal/mol. For comparison, according to Figure 3b and Table 1 the

44 modest attraction of -1.8 kcal/mol at the minimum D=4.1 A of the quintet corresponds well to

46 what is expected of purely vdW interactions.

51 behavior between that of the singlet and quintet with a minimum at 3.6 A and an interaction

53 energy of -2.9 kcal/mol. The triplet state is used as a tool to connect the single- and double-

quintet at the equilibrium geometry of the singlet is repulsive with +62.5 kcal/mol. This very

The triplet interaction energy curve for 42 in Figure 3b shows an intermediate

bonded pancake interactions in 42. The bare SOMO-SOMO pancake bonding energy term for the

4 triplet of 42 (Figure 3c) coincides remarkably well with that of the typical pancake bonded dimer

6 of 32. As has been discussed in connection with Chart 3, in the triplet state of 42 only one

8 pancake bond is left over as compared to the singlet and therefore it agrees well with the singlet

11 32, which also represents one pancake bond.

13 The dissociation limit of the 42 dimer yields three degenerate states (S, T and Q). They

15 arise from the coupling of the two triplet monomers as discussed in the SI section S.IV. There is

18 also a lower energy singlet of the monomer that shows signs of a second order Jahn-Teller

20 symmetry breaking.13

23 3.2. Diradical characters and unpaired density analysis. The extent and character of

25 unpaired density of the complexes was analyzed by the natural orbitals (NOs) of the one-particle

27 MR-AQCC density matrix and the effectively unpaired electron density30,31,32, which provides a

32 effectively unpaired electrons (Nu) is computed with the following formula :

35 N^ZUnKl-ntf (7)

37 where ni refers to the i-th natural orbital occupancy and N to the number of natural orbitals. We

39 have selected the non-linear formula given in Ref. 32 since it reduces the relative contributions

44 and thus highlighting only the truly open-shell contributions of the radical centers.

46 The total number of effectively unpaired electrons, NU, is displayed in Figure 4a as a

51 state of 32. At large separations NU values around 4.28 e are obtained for 42 exceeding the

53 expected 4.0 e because at the MR-AQCC level in addition to the two unpaired electrons on each

56 of the two monomers dynamical correlations provide a slight excess over 4.0 e. As the contact

measure for the separation of an electron pair into different spatial regions. The total number of

of the ni values that are close to 0 or 2 diminishing the contributions from dynamical correlation

function of the S---S contact distance (DS-S) for the different multiplicities of 42 and the singlet

minimum of D=3.6 A indicating that it is located intermediate between the quintet and singlet in

4 distance is reduced, the singlet Nu is dramatically reduced as the electrons start to pair. At the

6 equilibrium D=2.571 A the pairing is still incomplete where NU=0.57 e indicates a remaining

8 limited polyradical33 character. The triplet Nu values of 32 are not shown; they are essentially

11 constant at ~2.4 e, representing the two unpaired electrons plus some contributions from

13 dynamical correlation. At large separations the singlet state of 32 has the same number of

15 unpaired electrons as the triplet, which, however, in the former case is substantially reduced as

18 the two-electron pancake bond is being formed. At the equilibrium geometry (D=2.870 A) the

20 Nu=0.93 e value indicates a still existent significant diradicaloid character, typical for such single

00 3 34

23 pancake bonds. ' The qualitative difference compared to the single-bonded pancake is that the

25 double-bonded pancake bond is substantially shorter, and the multiradical character is reduced.

27 The triplet state of 42 retains a high value of multiradical character of Nu near 3.5 e at its

32 terms of electron pairing.

34 Figure 4b displays the natural orbital occupation numbers (NOONs) for the relevant

37 frontier orbitals of the dimers 32 and 42 shown in Figure 1. The values are nearly equal to 1.0 at

39 large separations as expected. All NOON values are evolving toward 2.0 and 0.0 at a similar

41 pace as D is reduced. However, at the respective equilibrium distance of each molecule these

44 values differ significantly since the Ds-s distance is considerably smaller for the double-bonded

46 42 pancake as compared to the single-bonded 32 pancake.

10 11 12

20 21 22

5.0 4.5 4.0 3.5 3.0 S 2.5 -^ 2.0 1.5 1.0 0.5 0.0

42 (S) 42 (T) 42 (Q) 32 (S)

J_I_I_I_I_I_I_I_I_I_I_I_I_I_|_I_I_I_I_I_I_I_I_I_I_I_I_I_l_

2.0 3.0

4.0 5.0

Ds-S (A)

2.0 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0

42 (LUNO) 42 (LUNO') 42 (HONO') 42 (HONO) 322 (LUNO) 322 (HONO)

—I_i_i | i_i_i_i | i_i_i_i | i_i_i_i | i_i_i_i | i_i_i_i—_|

3.0 4.0 5.0 6.0 7.0 8.0

Ds-S (A)

Figure 4. (a) Total number of effectively unpaired electrons (Nu) of 32 and 42 and (b) occupation numbers of the frontier NOs of the singlet states of 32 and 42 as a function of the separation distance (DS-S).

10 11 12

20 21 22

The unpaired electron densities, shown in Figure 5, indicate this difference also: the radical character of the double-bonded pancake is much smaller as compared to the single-bonded 32 case.

0.003 0.035

32 =0.93 e )

42 (#^=0.57 e)

Figure 5. Effectively unpaired electron density (isovalue 0.002 a.u) and atomic contributions for the singlets of 32 and 42. Nu is the number of effectively unpaired electrons given in parenthesis indicating stronger electron pairing in 42 compared to 32.

In extension of the simple integer bond order pmo defined in equation (1), we computed a more detailed bond order pno based on the two and four frontier orbital NOONs for 32 and 42, respectively,

pno = (NEBO - NEABO)/2, (8)

where NEBO is the number of electrons in bonding orbitals, NEABO is the number of electrons in the antibonding orbitals based on the natural orbital occupancies for the two frontier orbitals for 32 and four frontier orbitals for 42, respectively. The pno values obtained are 0.695 for 32 and 1.715 for 42. This reflects a major difference in the occupancies, which mirror the much stronger and shorter pancake bonds in the n-stacking dimer 42 compared to that in the radical dimer 32.

3.3. Further examples of double pancake bonded dimers. Based on these insights we have

stacking contacts as expected from double pancake bonding.

6 designed three new double pancake bonded dimers each with an even number of n-electrons.

8 These systems that might exhibit double pancake bonding were obtained by substituting Se for S

11 in 4 and substituting S for CH in 4, respectively, arriving in both cases at isoelectronic 8n-

13 electron rings. A further example is based on the C5H5+ ring that has a triplet ground state with

16 D5h symmetry and exhibits an antiaromatic electron count. Computational results at the MR-

18 AQCC and UB3LYP levels indicate that these systems exhibit very short intermolecular n-

20 21 22

25 3.3.1. Substitution of Se for S in dithiatriazine: Se2N3CH with 8 n-electrons.

28 The optimized geometry of a hypothetical double pancake bonded dimer using UB3LYP/6-

30 311++G(2d,2p) is shown in Figure 6. The UDFT geometry optimization (including UB3LYP

35 strong evidence for double pancake bonding in this dimer. The UDFT methods produced a C2v

37 symmetry optimized geometry for the (Se2N3CH)2 dimer with intermolecular Se-Se distances of

30 2.770 A (UB3LYP) and 2.685 A (UM06-2X). These contact values are strikingly shorter by

O O 36

42 1.030 A than the vdW distance for Se--Se (3.800 A). The rest of the intermolecular distances

44 in the (Se2N3CH)2 n dimer are similar to those of the (S2^CH)2 n dimer, 42. The (Se2^CH)2 n

47 dimer has also a large interaction energy of -27.0 kcal/mol (UB3LYP/6-311++G(2d,2p))

49 resulting from the perfect SOMO-SOMO overlap, indicating that this unique Se-bearing dimer

54 analysis and perhaps synthesis. The total interaction energy is comparable to that in 42: -27.0

and UM06-2X) on the (Se2N3CH)2 n dimer (72) using the 6-311++G(2d,2p) basis set provided

displays strong double pancake bonding and therefore it would be a good candidate for further

10 11 12

20 21 22

indicating the overall strength of the double pancake bonding and its ability to more than overcome vdW (Pauli) repulsion at these extremely short contact n-stacking distances.

Figure 6. Optimized geometry of the Se analogue of 42 with UB3LYP/6-311++G(2d,2p).

3.3.2. Substitution of S+ for CH in dithiatriazine: S3N3+ with 8 n-electrons. Two low-lying local minima of the hypothetical double pancake bonded dimer, (S3N3+)2 ( 82 and 92) were obtained by UB3LYP/6-311++G(2d,2p) and are shown in Figure 7.

Figure 7. Optimized geometry of the S+ substituted analogue of 42, (S3N3+)2. Two configurations are shown with UB3LYP/6-311++G(2d,2p).

The 82 configuration with pecfect overlap was further considered by rigid scan

the corrected (bare) interaction energy the Eint - Ecoul. The minimum of this corrected

6 calculations at the MR-AQCC(4,4)/6-311++G(2d,2p) level starting with the UB3LYP/6-

8 311++G(2d,2p) dimer optimized geometry. As shown in Figure 8, the total energy curve (Eint)

11 indicated that there is a metastable minimum with significant barrier to dissociation into two

13 S3N3+ fragments, which arises from strong cation-cation Coulomb repulsion. We approximated

18 interaction energy of the n dimer is -40.7 kcal/mol at about D = 2.7 A. Furthermore, the SOMO-

20 SOMO interaction of 82 has been investigated by subtracting the interaction of the highest spin

23 state (quintet) from the interaction of the singlet state as shown in Figure 8b. It turns out that the

25 82 possesses a very significant SOMO-SOMO component in the intermolecular interactions,

27 which is very similar to the behavior of 42 providing a tremendeous driving force toward

30 establishing a double bonded n-n stacking pancake.

32 In the 82 case the S---S contact was computed by a very large amount of 0.812 A shorter

37 based approximate point charges. The total interaction energy after subtraction of the Coulomb

39 repulsion is as large as for 42: -49.7 kcal/mol. This value and the short contact indicates that the

42 SOMO-SOMO bonding interaction is very strong and thus further examples of double pancake

44 bonding should be forthcoming.

than the vdW distance after subtracting a large Coulomb repulsion term approximated using ESP

10 11 12

20 21 22

100 80 60 40 20 0 -20 -40 -60 240

-120 -

-180 -

Eint ECoul

Ds-S (A)

Figure 8. The potential energy scans of the singlet and quintet of the (S3N3)2 n dimer with D3h symmetry (82) as a function of the S---S distance (DS_S) computed at the MR-AQCC(4,4)/6-311++G(2d,2p) level.

3.3.3. Perfluoro-cyclopentadiene cation: C5F5+ with 4 n-electrons.

C5H5+ has the right electron count to be a candidate for double pancake bonding. It has been investigated for its triplet ground state.35 We turned to the perfluoro derivative of CsH5+, because we anticipated that the use of a-electron withdrawing groups will faciltitate pancake bond

4 formation. The optimized geometry of the hypothetical double pancake bonded n-dimer,

6 (C5F5+)2 (102) using UB3LYP/6-311+G(d) is shown in Figure 9.

10 11 12

20 21 22

24 Figure 9. Optimized geometry of the hypothetical double pancake bonded dimer, (C5F5 )2.

29 102 exhibits a real local minimum with extremely short n-n stacking C-C distances of

31 2.611 A. The dimer of 102 represents the first five-member ring forming a double pancake

36 Thermodynamically, the IO2 is unfavorable because i) the strong Coulomb repulsion ii) the a

38 dimer is much more favorable by means of cycloaddtion. We suspect that the latter mechanism is

bonded system with two prefectly degenerate SOMO-SOMO interaction terms.

41 a main reason why pancake bonded systems with rings consisting of mostly C(sp2) have not yet

43 been charaterized. Energy minimization of the singlet ground state of the dimer of the perfluoro

45 derivative of C5H5+ (10, C5F5+) showed a well-definded local minimum with overall positive

48 interaction energy due to the large Coulomb repulsion. These three examples indicate that it

50 should be possible to find further systems that exhibit double pancake bonding.

4. CONCLUSIONS

stronger pancake bonding may lead to intermolecular contacts that might breach into the range of

6 It has been established for the first time through high-accuracy quantum mechanical modeling

8 that the n-stacking dimer of 42 can be understood as a double pancake bonded molecular

11 aggregate. This finding enriches the toolkit of chemical interactions in a sensitive area

13 connecting the weak vdW interactions to electron pair chemical bonds. The search for shorter and

18 extremely stretched single bonds38, 39, 40 in terms of bond distance and binding energy. Best

20 candidates to utilize this new double pancake bonding mechanism will be likely found among n-22

23 electron rich molecules with their highest two occupied orbitals being of n-type concomitant

25 with either a singlet ground state with low lying triplet state and diradicaloid character or n-

28 electron rich molecules with a triplet ground state.

35 Supporting Information

37 Computational details, validation and convergency of computational modeling, properties of the

40 monomer of 1,3,2,4,6-dithiatriazine (4), and absolute energies and the Cartesian coordinates for

42 all the optimized structures concerned.

44 AUTHOR INFORMATION

47 Corresponding Author

49 Kertesz@georgetown.edu

52 Notes

54 The authors declare no competing financial interest.

ASSOCIATED CONTENT

3 ACKNOWLEDGMENTS We thank the U. S. National Science Foundation for its support of

6 this research at Georgetown University (grant number CHE-1006702) and at the Texas Tech

8 University (grant number CHE-1213263). MK is member of the Georgetown Institute of Soft

11 Matter. Support was also provided by the Robert A. Welch Foundation under Grant No. D-0005

13 and by the Austrian Science Fund (SFB F41, ViCoM).

15 REFERENCES

20 Soc. 2005, 127, 14616.

25 3 Cui, Z.-h.; Lischka, H.; Mueller, T.; Plasser, F.; Kertesz, M. ChemPhysChem . 2014, 15, 165.

27 4 Novoa, J. J.; Miller, J. S. Acc. Chem. Res. 2007, 40, 189.

1 Devic, T.; Yuan, M.; Adams, J.; Fredrickson, D. C.; Lee, S.; Venkataraman, D. J. Am. Chem.

Haddon, R. C. ChemPhysChem. 2012, 13, 3581.

Casado, J.; Burrezo, P. M.; Ramirez, F. J.; Lopez Navarrete, J. T.; Lapidus, S. H.; Stephens, P.

32 W.; Vo, H-L.; Miller, J. S.; Mota, F.; Novoa, J. J. Angew. Chem. Int. Ed. 2013, 52, 6421.

34 6 Jung, Y.; Head-Gordon, M. Phys. Chem Chem Phys. 2004, 6, 2008.

37 7 Mota, F.; Miller, J. S.; Novoa, J. J. J. Am. Chem. Soc. 2009, 131, 7699.

39 8 Tian, Y.-H.; Kertesz, M. J. Am.. Chem. Soc. 2010, 132, 10648.

41 9 Mulliken, R. S.; Person, W. B. Molecular Complexes; (Wiley Sons.; Chap. 16, 1969).

44 10 Suzuki, S.; Morita, Y.; Fukui, K.; Sato, K.; Shiomi, D.; Takui, T.; Nakasuji, K. J. Am. Chem.

46 Soc. 2006, 128, 2530.

49 11 Raman, K. V.; Kamerbeek, A. M.; Mukherjee, A.; Atodiresei, N.; Sen, T. K.; Lazic, P.;

51 Caciuc, V.; Michel, R.; Stalke, D.; Mandal, S. K.; Blugel, S.; Munzenberg, M.; Moodera, J. S.

53 Nature, 2013, 493, 509.

56 12 Curtis, M. D.; Cao, J.; Kampf, J. W. J. Am. Chem. Soc. 2004, 126, 4318.

5 13 Hoffmeyer, R. E.; Chan, W.-T.; Goddard, J. D.; Oakley, R. T. Can. J. Chem. 1988, 66, 2279.

7 14 Cambridge Structural Database, by Cambridge Crystallographic Data Center, www.ccdc.

cam.ac.uk, 2013. (CSD version 5.35, ConQuest version 1.16).

12 15 Boeré, R. T.; French, C. L.; Oakley, R. T.; Cordes, A. W.; Privett, J. A. J.; Craig, S. L.;

14 Graham, J. B. J. Am. Chem. Soc. 1985, 107, 7710.

17 16 Boeré, R. T.; Fait, J.; Larsen, K.; Yip, J. Inorg. Chem. 1992, 31, 1417.

19 17 Hayes, P. J.; Oakley, R. T.; Cordes, A.W.; Pennington, W. T. J. Am. Chem. Soc. 1985, 107,

21 1346.

24 18 Szalay, P. G.; Bartlett, R. J. Chem. Phys. Lett. 1993, 214, 481.

26 19 Cui, Z.-h.; Lischka, H.; Beneberu, H. Z.; Kertesz, M. J. Am. Chem. Soc. 2014, 136, 5539.

28 20 Shepard, R. In Modern Electronic Structure Theory, D. R. Yarkony, Ed., Vol. 1, p 345, (World

31 Scientific: Singapore, 1995).

33 21 Shepard, R.; Lischka, H.; Szalay, P. G.; Kovar, T.; Ernzerhof, M. A. J. Chem.. Phys. 1992, 96,

Roos, B. O. Adv. Chem. Phys. 1987, 69, 399.

35 2085.

38 22 Lischka, H.; Dallos, M.; Shepard, R. Mol. Phys. 2002, 100, 1647.

40 23 (a) Ruedenberg, K.; Cheung, L. M.; Elbert, S. T. Int. J. Quantum Chem. 1979, 16, 1069. (b)

45 24 (a) McLean A. D. J. Chem. Phys. 1980, 72, 5639. (b) Krishnan, R.; Binkley, J. S.; Seeger, R.;

47 Pople, J. A. J. Chem. Phys. 1980, 72, 650.

49 25 Bunge, A. J. Chem. Phys. 1970, 53, 20.

52 26 Lischka, H.; Müller, T.; Szalay, P. G.; Shavitt, I.; Pitzer R. M.; and Shepard, R. COLUMBUS

54 - A Program System for Advanced Multireference Theory Calculations, Wiley Interdisciplinary

Bondi, A. J. Phys. Chem. 1964, 68, 441.

5 Reviews: Computational Molecular Science (WIREs:CMS), W. Allen Ed., 1 (2011), 191-199,

7 Lischka, H.; et al. COLUMBUS, an ab initio electronic structure program. Release 7.0, 2013,

9 www.univie.ac.at/columbus.

12 27 Hu, H.; Lu, Z.; Yang, W. J. Chem. Theory and Comput. 2007, 3, 1004.

14 28 Tian, Y. H.; Kertesz, M. J. Phys. Chem. A 2011, 115, 13942.

19 30 Takatsuka, K.; Fueno, Y.; Yamaguchi, K. Theor. Chim. Acta 1978, 48, 175.

21 31 See also Staroverov, V. N.; Davidson, E. R. Chem. Phys. Lett. 2000, 330, 161.

24 32 (a) Head-Gordon, M. Chem. Phys. Lett. 2003, 372, 508. (b) Head-Gordon, M. Chem. Phys.

26 Lett. 2003, 380, 488.

28 33 Plasser, F.; Pasalic, H.; Gerzabek, M. H.; Libisch, F.; Reiter, R.; Burgdorfer, J.; Müller, T.;

31 Shepard, R.; Lischka, H. Angew. Chem. Int. Ed. 2013, 52, 2581.

33 34 Beneberu, H. Z.; Tian, Y.-H.; Kertesz, M. Phys. Chem Chem Phys. 2012, 14, 10713.

Gogonea, V.; Schleyer, P. v. R.; Schreiner, P. R. Angew. Chem., Int. Ed., 1998, 37, 1945.

38 36 Mantina, M.; Chamberlin, A. C.; Valero, R.; Cramer, C. J.; Truhlar, D. G. J. Phys. Chem. A

40 2009, 113, 5806.

43 37 Tian, Y. H.; Kertesz, M. J. Phys. Chem. A 2011, 115, 13942.

45 38 Choi, C. H.; Kertesz, M. Chem. Comm. 1997, 2199.

47 39 Kaupp, G.; Boy, J. Angew. Chem. Internat Ed. 1997, 36, 48.

50 40 Isea, R. J. Mol. Struct. Theochem. 2001, 540, 131.

10 11 12

20 21 22

TOC graphics

fi Si i?

We report on the amazing properties of a new type of stacking interaction mechanism between n conjugated molecules in the form of a "double pancake bond" which breaks the record for short intermolecular distances and formidable strength of n-n stacking interaction.