N-Aryl-9,10-phenanthreneimines as Scaffolds for Exploring Noncovalent Interactions: A Structural and Computational Study

Article abstract of DOI:10.1002/ejoc.201700884

A series of 10-[(4-halo-2,6-diisopropylphenyl)imino]phenanthren-9-ones and derivatives of the phenanthrene-9,10-dione ligand have been synthesised and structurally characterised to explore two types of noncovalent interactions, namely the influence of the steric bulk upon the resulting C–H···π and π-stacking interactions and halogen bonding. Selected noncovalent interactions have additionally been analysed by DFT and AIM techniques. No halogen bonding has been observed in these systems, but X lone pair···π, C–H···O=C and C–H···π interactions are the prevalent ones in the halogenated systems. Removal of the steric bulk in N-(2,4,6-trimethylphenyl)-9,10-iminophenanthrenequinone affords different noncovalent interactions, but the C–H···O=C hydrogen bonds are observed. Surprisingly, in N-(2,6-dimethylphenyl)-9,10-iminophenanthrenequinone and N-(phenyl)-9,10-iminophenanthrenequinone these C–H···O=C hydrogen bonds are not observed. However, they are observed in the related 2,6-di-tert-butylphenanthrene-9,10-dione. The π-interactions in dimers extracted from the crystal structures have been analysed by DFT and AIM. Spectroscopic investigations are also presented and these show only small perturbations to the O=C–C=N fragment.

Full text of DOI:10.1002/ejoc.201700884

A Journal of
Accepted Article
Title: N-Aryl-9,10-phenanthreneimines as scaffolds for exploring non-
covalent interactions: A structural and computational study
Authors: David Farrell, Samuel Kingston, Dmitri Tungulin, Stefano
Nuzzo, Brendan Twamley, James A Platts, and Robert J.
Baker
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Eur. J. Org. Chem. 10.1002/ejoc.201700884
Link to VoR: http://dx.doi.org/10.1002/ejoc.201700884
Supported by

10.1002/ejoc.201700884
European Journal of Organic Chemistry
N-Aryl-9,10-phenanthreneimines as scaffolds for exploring
non-covalent interactions: A structural and computational
study
David Farrell,a Samuel J. Kingston,a Dmitry Tungulin,a Stefano Nuzzo,a Brendan Twamley,a
James A. Plattsb and Robert J. Baker a*
a School of Chemistry, University of Dublin, Trinity College, Dublin 2, Ireland
b School of Chemistry, Main Building, Cardiff University, Park Place, Cardiff, CF10 3AT, U.K.
Table of content graphic
N-Aryl-9,10-phenanthreneimines that feature a halogen on the aryl group show lone pair…
interactions, whilst non-halogenated substituents show - stacking.
This article is protected by copyright. All rights reserved.
1

10.1002/ejoc.201700884
European Journal of Organic Chemistry
ABSTRACT. A series of 10-((4-halo-2,6-diisopropylphenyl)imino)phenanthren-9-ones and
derivatives of the phenanthrene-9,10-dione ligand have been synthesised and structurally
characterised to explore two types of non-covalent interactions, namely the influence of the steric
bulk upon the resulting C—H... and -stacking interactions and halogen bonding. Selected non-
covalent interactions have additionally been analysed by DFT and AIM techniques. No halogen
bonding has been observed in these systems, but X lone pair…, C—H…O=C and C—H...
interactions are the prevalent ones in the halogenated systems. Removal of the steric bulk in N-
(2,4,6-trimethylphenyl)-9,10-iminophenanthrenequinone
affords
different
non-covalent
interactions, but the C—H...O=C hydrogen bonds are observed. Surprisingly, in N-(2,6-
dimethylphenyl)-9,10-iminophenanthrenequinone
and
N-(phenyl)-9,10-
iminophenanthrenequinone these C—H...O=C hydrogen bonds are not observed. However, they
are observed in the related 2,6-di-tert-butylphenanthrene-9,10-dione. The -interactions in
dimers extracted from the crystal structures have been analysed by DFT and AIM. Spectroscopic
investigations are also presented and these show only small perturbations to the O=C—C=N
fragment.
Introduction
A number of devices in materials chemistry, such as for applications in organic
semiconductors,[1] spintronics,[2] and optoelectronics,[3] rely on -conjugated organic molecules
and polymers and they have generated intensive interest. However to control the important
electronic properties the structures of these organic molecules, and in particular the packing,
must be well understood. Non-covalent supramolecular interactions therefore become important
to study as these can be used to ‘tune’ the packing, although given the weak nature of these
This article is protected by copyright. All rights reserved.
2

10.1002/ejoc.201700884
European Journal of Organic Chemistry
interactions,[4] this is currently somewhat of an empirical process. Of the multitude of non-
covalent interactions reported, aromatic interactions such as C-H… interactions[5] or -stacking
are important, especially in -conjugated organic molecules. -stacking is a much used
misnomer and studies to understand the origin of substitution effects have been ongoing for the
past two decades. The popular model proposed by Hunter and Sanders,[6] namely -polarisation,
has now been called into question and the model popularised by Wheeler and Houk[7] is
consistent with most of the literature. This involves the direct interactions between the C-H/X
dipoles of substituted aromatic interactions, which may be attractive or repulsive. A number of
recent reviews summarise the current thinking,[8] but it is clear that more experimental data is
required. Other non-covalent methods are becoming increasingly recognised include halogen
bonding,[9] particularly in supramolecular chemistry and crystal engineering.[10] In some cases
these can be more favorable than, and thus compete with (termed synthon crossover[11]), or can
cooperate with other non-covalent interactions such as hydrogen bonding,[12] lone pair...[13] or
-stacking.[14] The ability to distinguish between close-packing arrangements and true halogen
bonding interactions was first proposed in 1989[15] whereby type I and type II interactions were
defined. More recently, the emergence of a quasi-type I and type II have been noted and these
individual interactions can be delineated by inspection of the angles such that 0o|1-2|15o are
type I; 15o|1-2|30o are quasi-type I/type II and 30o|1-2| are type II (where  refers to the
two C-X…X angles). Moreover it was found that type II are most favoured with iodine whilst
type I are most favored with Cl.[16]
There are few reported structures containing the phenanthere moeity so in this contribution we
use it as a platform that contains a plethora of potential interactions viz. ‘- stacking’ via the
This article is protected by copyright. All rights reserved.
3

10.1002/ejoc.201700884
European Journal of Organic Chemistry
Wheeler and Houk mechanism, C—H...; C—X lone pair..., C—X...H—C and C—X...X—C
where we can also control the steric demand of the ligand systems. We specifically look at two
types of interactions: (1) those arising from the introduction of a halogen, and (2) those due to
steric demands of the ligands. We find that the dominant halogen interaction is a lone pair...
type. DFT and AIM analysis compares the strength of these interactions, which are surprisingly
strong (ca. 45 kJ mol-1). Changing the steric bulk of the ligand also changes the nature of the
non-covalent interactions and these -interactions are also characterised computationally.
Results and Discussion
Synthesis and Structural characterisation.
The phenanthrene imines (1-7) were synthesised according to Scheme 1. The substituted aniline
was added to a solution of phenanthrene-9,10-dione along with a catalytic amount of formic acid.
This method did not work for aniline, so an alternative method via the Ph3As=O catalysed
reaction of phenyl isocyanate and phenanthrene-9,10-dione was used.[17] Workup and
purification by column chromatography afforded the desired compounds. Similarly the tert-butyl
substituted phenanthrene-9,10-dione (8) was prepared via a literature method and the three
isomers separated by careful column chromatography; we could not grow crystals suitable for X-
ray diffraction for the other two isomers so these are not included in our study. Finally we
attempted the reaction with the very bulky 2,4,6-tritertbutylaniline, but no reaction occurred,
even after prolonged reflux. Crystals of all were obtained by the slow evaporation of chloroform,
Bu2O (5) or hexane (7) solutions. The discussion will firstly focus on the molecular structures,
followed by the packing arrangements and the description of the non-covalent interactions
present.
This article is protected by copyright. All rights reserved.
4

10.1002/ejoc.201700884
European Journal of Organic Chemistry
Scheme 1. Synthesis of compounds described in this work.
This article is protected by copyright. All rights reserved.
5

10.1002/ejoc.201700884
European Journal of Organic Chemistry
1 suffers from whole molecule disorder and we refrain from discussing the metric parameters;
the structure is shown in Figure S1. The molecular structures of 2, 3, 6, 7 and 8 are
representative of the library of compounds and shown in Figures 1-5 whilst 4 and 5 are shown in
Figures S3-S4 respectively. In the series 2-7 the metric parameters of the phenanthrene backbone
do not change and the C—C, C=N and C=O bond lengths are also essentially identical (Table 1).
The most noticeable difference is the planarity of the phenanthrene backbones which are
disrupted in 2 and 6 and to a much lesser extent in 3; while 4, 5 and 7 are flat. 2 has two
molecules in the asymmetric unit; whilst the bond lengths in 2a and 2b are identical, remarkably
the planarity of the phenanthrene backbones differ. In molecule 2a the torsion angles C(7)—
(C8)—C(9)—C(10) = 6.6(4)o and O(1)—C(2)—C(15)—N(16) = 16.2(4)o whilst in molecule 2b
the torsion angles C(35)—(C36)—C(37)—C(38) = 11.5(4)o and O(29)—C(30)—C(43)—N(44)
= 29.6(4)o. A survey of the Cambridge database shows that the median of the C-C-C-C torsion
angle = 2.999o and of the X-C-C-X = 2.882o (X = any atom); deviations are generally due to
steric effects (Figure S5). 7 has two molecules in the unit cell, one of which is disordered, but the
metric parameters are identical. In addition the C=O and C=N are significantly out of the plane
of the phenanthrene unit.
This article is protected by copyright. All rights reserved.
6

10.1002/ejoc.201700884
European Journal of Organic Chemistry
Table 1. Selected structural parameters in the structures of 2-8 (in Å and o).
b
b
C=O
C=N
C-C
C-C-C-C
N-C-C-O
C=N oopa
C=O
oopa
1.216(3)
1.220(3)
1.216(2)
1.208(5)
1.218(3)
1.218(3)
1.215(7)
1.215(2)
1.220(2)
1.282(3)
1.275(3)
1.279(2)
1.280(5)
1.280(3)
1.278(4)
1.273(8)
-
1.526(3)
1.520(3)
1.519(2)
1.519(6)
1.520(2)
1.517(4)
1.506(9)
1.541(2)
6.6(4)
11.5(4)
5.2(2)
2.3(6)
2.1(3)
9.11(4)
3.3(9)
7.1(2)
16.2(4)
29.6(4)
15.2(2)
8.8(6)
-0.429
-1.013
-0.320
-0.179
-0.193
-0.546
-0.170
-
0.375
-
-
2a
2b
3
0.375
0.226
0.142
0.123
0.787
0.464
0
70
82
89
-
112
85
79
-
4
8.6(3)
5
29.1(4)
13.7(10)
6.3(2)
-
-
6
7ac
8
-
-
-
-
a o.o.p = distance out of the plane of the phenanthrene backbone (Å)
b intermolecular contacts  is the angle between the plane of the ring, centroid and X,[13d]
whilst  = C-X…centroid angle.
c = There are two molecules in the asymmetric unit of which one is disordered. Only
dimensions from the ordered molecule (7a) are reported.
This article is protected by copyright. All rights reserved.
7

10.1002/ejoc.201700884
European Journal of Organic Chemistry
a
b
Figure 1. Molecular structure of 2 with atomic displacement parameters shown at 50%
probability. The two molecules are labelled as 2a and 2b in the text. Hydrogen atoms omitted for
clarity.
This article is protected by copyright. All rights reserved.
8

10.1002/ejoc.201700884
European Journal of Organic Chemistry
Figure 2. Molecular structure of 3 with atomic displacement parameters shown at 50%
probability.
This article is protected by copyright. All rights reserved.
9

10.1002/ejoc.201700884
European Journal of Organic Chemistry
Figure 3. Molecular structure of 6 with thermal displacement shown at 50% probability.
Figure 4. Major (7a: 82%) and minor (7b: 18%) disordered moiety of 7 with thermal
displacement shown at 50% probability. Hydrogen atoms omitted for clarity.
This article is protected by copyright. All rights reserved.
10

10.1002/ejoc.201700884
European Journal of Organic Chemistry
Figure 5. Molecular structure of 8 with thermal displacement shown at 50% probability.
In the structure of 8, the C(1)—C(2) bond length of 1.541(2) Å and the C=O bond lengths of
1.220(2) and 1.215(2) are typical for phenanthrene-9,10-diones, e.g. 1.52(3) to 1.524(5) Å for the
C—C and 1.207(6)-1.25(2) Å for the C=O in the different polymorphs of phenanthrene-9,10-
dione.[18] There is a relatively small deviation from planarity (C(7)—C(8)—C(9)—C(10) =
o
7.07(2) ); steric bulk at the C(7) and C(10) positions invokes a bigger bending e.g. 39.9o in
2,4,5,7-tetramethyl-9,10-phenanthroquinone.[19]
This article is protected by copyright. All rights reserved.
11

10.1002/ejoc.201700884
European Journal of Organic Chemistry
Packing and intermolecular interactions in 2-4.
We first focus on the halogen series. In the structures of 2-4 there are no C—X...O=C or X...X
interactions present, but some identical interactions are observed (Figure 6). In 2 there are a
number of C—H...O interactions[20] between H(5) and H(6) (hydrogens are in the calculated
positions in all structures; dC…O = 3.293(3) and 3.279(3) Å) and C—H...N interactions involving
N(44) and H(7) and H(10) (dC…N = 3.633(11) and 3.618(8) Å) (Figure 6a); if 2 is recrystallized
from DMF, the same structure is formed, indicating that the crystal packing interactions are
reproducible in different solvents. In contrast, in 3 and 4 the C—H...O are from H(10) and H(11)
(3: dC…O = 3.123(4) and 3.081(4) Å; 4: dC…O = 3.108(5) and 3.117(6) Å) and no C—H...N
interactions are present (Figure 6c for 3; 4 is identical). A C—H... interaction from a
phenanthrene H to the X-substituted aryl group (dC—H…cent = 2.564 Å) is present in 2 (Figure 6a),
whilst in 3 and 4 there is also longer one from an isopropyl hydrogen to the phenanthrene (3: dC—
= 2.576 and 3.160 Å; 4: dC—H…cent = 2.552 and 3.126 Å; Figure 6b). In 2a, there are
H…cent
phenanthrene groups not involved in a l.p.... interactions but are close to another Cl atom –
inspection of the angles show that this is a type I interaction. There is a short contact of 3.238 Å
between the C—Cl and the centroid of an adjacent phenanthrene ring which has the larger
deviation from planarity (2b) and shown in Figure 6b. This lone pair... interaction is of a semi-
localized type according to the scheme developed by Shishkin,[21] and within the sum of van der
Waals radii (C...Cl = 3.45 Å). In 3 and 4 this lone pair... interaction (3: 3.396(8) Å; 4: 3.569(7)
Å; Figure 8d) is a delocalized interaction[21] where and inspection of the  and  angles shows
that the heavier halogens move close to the centre of the ring (Table 1). Both are within the van
der Waals radii (3.55 Å for C...Br and 3.68 Å for C...I). However, a recent high level theoretical
This article is protected by copyright. All rights reserved.
12

10.1002/ejoc.201700884
European Journal of Organic Chemistry
study has shown that in R—Br... interactions with benzene (R = H, HCC and NC), it is only
the distance that influences the strength of the interactions and not the angles.[22] The closest
Br...Br and I...I distances are 4.370(6) and 4.125(6) Å respectively and thus outside the sum of the
van der Waals radii. The second form of weak hydrogen bonds are X...H—Csp2 (2: dC…Cl
=
3.781(3) Å; 3: dC…Br = 3. 800(4) Å) from a phenanthrene hydrogen in 2 and an isopropyl
hydrogen in 3; in 4 there are no C—H...I contacts within the sum of the van der Waals radii.
There are no obvious - stacking interactions in 2-4, presumably due to the increased steric
demand of the halogen substituents.
Figure 6. (a) View of 2 highlighting the non-covalent interactions; (b) packing of 2 as viewed
along the crystallographic c-axis; (c) view of 3 highlighting the non-covalent interactions; (d)
packing of 3 as viewed along the crystallographic b-axis. Colour code: Cl – green; Br – brown;
only hydrogen atoms involved in hydrogen bonds are shown, others are omitted for clarity.
This article is protected by copyright. All rights reserved.
13

10.1002/ejoc.201700884
European Journal of Organic Chemistry
DFT and AIM calculations on lone pair… interactions.
In order to ascertain whether the lone pair... interaction is simply due to crystal packing, we
utilized a DFT and AIM approach. Dimers exhibiting a range of intermolecular interactions were
extracted from the crystal structure of 4, and counterpoise-corrected binding energies calculated
with M06-2X with both Def2TZVP and Def2SVP basis sets. The dimers are bound by -43.9 and
-43.6 kJ mol-1, respectively. We also included a dispersion-corrected functional (B97D), which
gives a binding energy of -58 kJ mol-1. Given the excellent agreement between the basis sets, the
computationally smaller Def2SVP basis can be used to give qualitative data for comparison
between our small library of compounds (the dimers can be seen in Figures S28-S36). While
dimers may not entirely reproduce the overall crystal environment, this approach allows us to
interrogate individual types of interaction, and also to compare observed and hypothetical
interaction modes on an equal footing.
The isostructural dimer of 3 containing X... contacts are bound by -45.5 mol-1. A hypothetical
dimer of 2, constructed by replacement of Br by Cl followed by optimization of the position of
Cl, is bound by -42.2 kJ mol-1 and so almost identical to those for 3 and 4. In contrast, the real
dimers of 2 extracted from the crystal structure are much more weakly bound: the dimer
containing Cl... contacts is bound by -22.3 kJ mol-1, while that containing Cl...Cl contact is
predicted to be unbound (binding energy = +2.1 kJ mol-1 after counterpoise correction), in
agreement with the assignment of a type I interaction. It is important to note that this is the total
energy of the interactions between the dimers, and not of individual components, but lone
pair...interactions, in cooperation with other non-covalent interactions are energetically
significant; this has been noted in previous examples.[8a],[23]
This article is protected by copyright. All rights reserved.
14

10.1002/ejoc.201700884
European Journal of Organic Chemistry
The molecular electrostatic potentials have been calculated for the monomeric units in 2-4.
Figure 7 shows the results for 4, but all are similar (Figures S6 and S7 show the plots for 2 and
3). There is a -hole on the iodine atom but is small; for 2 and 3 this is not observed and this is
likely due to the differences in electronegativity of the halogen atom.[24] A small -hole, defined
as a positive region, on the central phenanthrene ring (3 = 2.2 x 10-2 a.u.; 4 = 2.9 x 10-2 a.u.) is
evident, corresponding to the lone pair... interaction; in the plot of 2 this is absent and
corroborates the differences in binding energies. The aromatic hydrogens are slightly +, possibly
due to the polarizing effect of the carbonyl function, and could explain the prevalence of the C—
H...O hydrogen bonds.[25]
Figure 7. Molecular Electrostatic Potential of 4 (numerical values are x 10-2 a.u; colour: blue =
6.0 x 10-2 to red = -6.0 x 10-2 a.u.).
This article is protected by copyright. All rights reserved.
15

10.1002/ejoc.201700884
European Journal of Organic Chemistry
AIM analysis showed multiple bond paths corresponding to intermolecular interactions, as
shown in Figure 8 and Table 2. Two X... bond paths are present, along with one C—H...X path.
A further three ... contacts and one C—H... are predicted, along with two C—H...O contacts.
The value of the electron density at the bond critical point, bcp, is widely used as an estimate of
the strength of contacts: largest values are observed for C—H...O, followed by X... with C—H...
and ... contacts predicted to be the weakest. The literature is replete with computational studies
of lone pair... interactions, and the electron density at the bcp is similar to that calculated for a
variety of electron rich or electron poor arenes and lone pairs originating from oxygen or
nitrogen donors.[26]
Table 2. Calculated electron density (in a.u.) at bcps in the dimers extracted from 2-4.
X...
C—H...
...
Compound
C—H...O
C—H...X
X...X
2 hypothetical 0.008,
0.006,
0.007
0.006
0.006
0.005 x2,
0.006
0.010
2 X...X
0.008
2 X...
0.010
0.005
0.009
0.007
0.003,
0.004 x2
0.006 x2
0.008,
0.009
0.009,
0.010
0.007 x2
0.006 x2
0.005 x2,
0.006
3
4
0.004
0.004,
0.005 x2
This article is protected by copyright. All rights reserved.
16

10.1002/ejoc.201700884
European Journal of Organic Chemistry
Figure 8. AIM properties of 3 displaying intermolecular bond paths as dotted lines
Packing and intermolecular interactions in 1, 5-8.
Compounds 1, 5, 6 and 7 allows us to probe the relative stability of the non-covalent interactions
when the steric demands of the ortho-groups are changed, and perhaps comment on the potential
of -interactions. The packing of 1 is shown in Figure S2; there are only C—H... interactions
present between a phenanthrene hydrogen and the centroid of the aryl ring (dC…cent = 2.570 Å) in
this structure. The packing of 5 (Figure 9a) has no C—H... interactions and two C—H...O
interactions (dC…O = 3.370(3) and 3.493(3) Å), which forms chains. There are also ‘- stacking’
between two phenanthrene rings (dcent…cent = 3.622 Å) that stack in a head to tail fashion.
However in 6 (Figure 9b), where only one methyl group in the para-position has been removed,
the C—H...O hydrogen bond is no longer present. Instead, the prevalent interaction in 6 is ‘-
This article is protected by copyright. All rights reserved.
17

10.1002/ejoc.201700884
European Journal of Organic Chemistry
stacking’ between two phenanthrene rings (dcent…cent = 3.929 Å), longer than that found in 5.
There are also weak C—H...N hydrogen bonds (dC…N = 3.606(14) Å), comparable to that
observed in 2a. 7 (Figure 9c) has a different packing where C—H… interactions dominate
whereby hydrogens (H12 and H5) on the phenanthrene backbone and the phenyl ring of the
disordered imine are involved and extend along the crystallographic b-axis. ‘- stacking’
between two phenanthrene rings (dcent…cent = 3.803 Å) and C—H…O interactions (dC…O
=
3.218(8) Å) are present, but the latter involves hydrogens on C(13) and not the same as that
observed in 2-4. Therefore as the steric bulk is reduced around the ortho-positions from iPr to Me
to H the propensity for ‘- stacking’ increases, as might be expected. Surprisingly however
there is a difference between the mesityl and xylyl substituents, and we are not sure of an
explanation for this.
This article is protected by copyright. All rights reserved.
18

10.1002/ejoc.201700884
European Journal of Organic Chemistry
Figure 9. Packing of (a) 5 and (b) 6 viewed along the crystallographic b-axis; (c) 7 viewed along
the crystallographic c-axis; (d) 8 viewed along the crystallographic a-axis.
In order to further analyse the influence of the steric bulk of the ligand, the structure of 8
(Figure 9d) is useful as the steric bulk is closer to the seemingly principal C—H...O interactions.
8 is built up of chains of phenanthrene-9,10-diones hydrogen bonded {dC…O = 3.243(2) and
3.392(2) Å} along the c-axis and a head-to-tail ‘- stacking’ between the two central rings
(dcent…cent = 3.759 Å) and in the outer rings (dcent…cent = 3.772 Å), along with C—H...
interactions.
This article is protected by copyright. All rights reserved.
19

10.1002/ejoc.201700884
European Journal of Organic Chemistry
DFT and AIM calculations on 5-8.
In order to examine the -interactions in these species we have used the same computational
methodology as described above (M06-2x/Def2SVP level) and in our previous work.[27] The
binding energies of the dimers extracted from the crystal structures are -52.94 kJ mol-1for 5, -
48.54 kJ mol-1for 6, -36.53 kJ mol-1 for 7 and -74.45 kJ mol-1for 8. A possible explanation for the
interactions comes from the molecular electrostatic potentials shown in Figure 10 for 5 and 8 (6
and 7 are similar and are shown in Figure S8 and S9). From this analysis the hydrogen atoms on
the phenanthrene are polarised due to the C=O and C=N (for 5-7) electron withdrawing groups,
but as the prevalent interaction is a head to tail stacking, it follows that the positively polarised
hydrogens are preferentially interacting with the negatively polarised carbonyl function. This is
consistent with the direct interactions proposed by the Wheeler and Houk mechanism for non-
covalent -interactions.
This article is protected by copyright. All rights reserved.
20

10.1002/ejoc.201700884
European Journal of Organic Chemistry
Figure 10. Molecular Electrostatic Potential of 5 (left; colour: blue = 6.7 x 10-2 to red = -6. 7 x
10-2 a.u.) and 8 (right; colour: blue = 8.0 x 10-2 to red = -8.0 x 10-2 a.u.).
AIM analysis of 5-8 showed multiple bond paths corresponding to intermolecular interactions, as
shown in Figure 11 and Table 3. There are a number of ... contacts, C—H... and C—H...O
contacts in this series. Examination of the electron density at the bond critical point, bcp, shows
the trend that the ... contacts are stronger than C—H...Moreover 7 appears to report the
weakest binding, in line with the binding energies; 8 on the other hand has the most stabilisation
from ... contacts.
This article is protected by copyright. All rights reserved.
21

10.1002/ejoc.201700884
European Journal of Organic Chemistry
Figure 11. AIM properties of 5 (left) and 8 (right) displaying intermolecular bond paths as
dotted lines
Table 3. Calculated electron density (in a.u.) at bcps in the dimers extracted from 5-8.
O...
C—H...
0.006 x 2
0.005 x 2
0.002 x 2
...
Compound
C—H...O
0.007 x 2
0.006 x 3
0.006 x 2
0.005
5
6
0.004 x 2
0.006 x 2
0.007 x 2
0.004 x 4
0.007 x 2
7
8
0.004 x 2 0.002 x 4
Hirshfeld analysis
To extend our study on the different packing arrangements we have used Hirshfeld analysis.[28]
The results of the fingerprint analysis for the lone pair... interactions in 2-4 are shown in Figure
This article is protected by copyright. All rights reserved.
22

10.1002/ejoc.201700884
European Journal of Organic Chemistry
12. It is clear from this that the dominant interaction is H...H, but C...H (i.e. C—H...) are
significant. X...lone pair interactions are small but clearly observed, particularly in the plot of the
shape index. The C—H... interaction is the largest for 7 according to this analysis.
This article is protected by copyright. All rights reserved.
23

10.1002/ejoc.201700884
European Journal of Organic Chemistry
8
7
6
5
4
3
2B
2A
1
0%
20%
40%
60%
80%
100%
X-X
X-C
X-H
X-O
C-C
C-H
H-H
H-N
H-O
Figure 12. (top) Fingerprint plots of 2b, 3 and 4 displaying the X...C interactions (dnorm plotted at
-0.161 to 1.605), (middle) the shape index (plotted at -1 to +1) and (bottom) the quantitative
analysis of the Hirshfeld surfaces for 1-8.
Further Reactivity Studies
Given that no XB bonding was observed in 2-4 we next tried a number of co-crystallization
experiments to see if we could engineer XB. It is known that asymmetric XB typically goes via
Type II interactions,[9] so we co-crystallized a mixture of 2 and 4; the resulting structure, 9,
(Figure S10) showed a disordered structure with a statistical mixture of chloro and iodo
substituents, but the packing interactions were of the iodo type. Next we tried the reactions of 3
with Br2 and 4 with ICl and I2, for which we only obtained crystals of the latter. The structure
showed that the iodide had been partially lost and the structure (10) was a disordered mixture of
This article is protected by copyright. All rights reserved.
24

10.1002/ejoc.201700884
European Journal of Organic Chemistry
4 and 1 (Figure S11). We are unsure of the mechanism of this transformation, but it could be that
the I2 used contained a small amount of impurities such as HI. Further details can be found in the
ESI.
Spectroscopic characterization.
Infrared and UV-vis spectroscopy were utilized to characterize these compounds, and the
pertinent parameters are recorded in Table 4. The C=O and C=N bond stretching frequencies in
the infrared spectra of crushed single crystals are barely perturbed suggesting that both the
electronic influences of the para-substituent are not particularly strong and the hydrogen bond
interactions discussed above are weak.
Table 4. Spectroscopic properties of phenanthrene imines 1-7, 8 and phenanthrene-9,10-dione.

max (nm) (, M-1 cm-1)
Compound
(C=O)
(cm-1)
1674
(C=N)
(cm-1)
-
-*
-*
n-*
319 (3805)
315 (1488)
314 (2367)
318 (3253)
-
412 (1453)
587 (96)
583 (160)
574 (199)
Phenanthrene-
9,10-dione
1
1679
1677
1681
1589
1592
1592
390 (530)
380 (733)
389 (869)
2
3
This article is protected by copyright. All rights reserved.
25

10.1002/ejoc.201700884
European Journal of Organic Chemistry
1679
1675
1677
1647
1675
1591
1590
1591
1591
-
313 (3557)
313 (2150)
314 (4954)
312 (4497)
325 (2034)
385 (1389)
581 (285)
599 (162)
575 (302)
540 (530)
426 (522)
4
5
6
7
8
378 (800)
388 (1604)
-
-
The UV-vis spectra of the compounds in CH3CN are shown in Figure 13. In the series 1-
8 a broad, weak band at ca. 600 nm appears upon substitution of the aryl ring. This has been
assigned to C=O and C=N n-* transitions, whereby the intensity variations likely due to
deviations from non-planarity of the carbonyl and imino fragments.[29] The molar extinction
coefficients of the bands at 390 nm and 315 nm, the -* transitions within the phenanthrene
ring, increase upon halogen substitution from 1-4. Similarly the small shifts in intensity and
extinction coefficients of 9 compared to the unsubstituted phenanthrene-9,10-dione are likely due
to the structural variability caused by the increasing steric bulk and the deviations from planarity
of the imino and carbonyl groups. Thus the UV-vis data suggest generally small perturbations to
the electronic structure due to the structural variations.
This article is protected by copyright. All rights reserved.
26

10.1002/ejoc.201700884
European Journal of Organic Chemistry
5000
4500
4000
3500
3000
2500
2000
1500
1000
500
1
2
3
4
5
6
7
0
300
400
500
600
700
800
7000
6000
5000
4000
3000
2000
1000
phenanthrene dione
8
0
300
400
500
600
700
800
wavelength (nm)
Figure 13. Electronic absorption spectra of the compounds 1-8 in CH3CN solution
(concentration ≈ 1 x 10-5 M).
Conclusions
In summary, we have structurally characterised a number of examples based on a phenanthrene
core and shown that lone pair... interactions are the dominant halogen type interaction and a
computational study shows that these are stabilizing and of the same order of magnitude as
interactions from lone pairs originating from oxygen or nitrogen donor atoms. However the most
important interaction appears to be a weak hydrogen bond between the C—H of a phenanthrene
and the carbonyl oxygen. Inspection of the molecular electrostatic potential shows that the
This article is protected by copyright. All rights reserved.
27

10.1002/ejoc.201700884
European Journal of Organic Chemistry
aromatic hydrogen atoms are polarized, probably due to the C=O and C=N function on the
opposite side of the ring, which can account for this interaction. The steric bulk of the imino aryl
group has been evaluated and as the steric requirements are reduced then ‘- stacking’ and
eventually C—H… interactions dominate. We posit that the molecular electrostatic potentials
of these compounds are in line with the Wheeler and Houk mechanism for non-covalent -
interactions.
Experimental
13
1H, and C{1H} spectra were recorded on a Bruker AV400 spectrometer operating at 400.23
MHz and 155.54 MHz respectively, or a Bruker Avance II 600 NMR with a TCI cryoprobe
1
spectrometer operating at 150.92 MHz (13C) and were referenced to the residual H and 13C
resonances of the solvent used. IR spectra were recorded on a Perkin Elmer Spectrum One
spectrometer with attenuated total reflectance (ATR) accessory. Mass spectra were measured on
a MALDI QTOF Premier MS system. UV-vis measurements were made on a Perkin Elmer
Lambda 1050 spectrophotometer, using fused silica cells with a path length of 1 cm. 4-X-2,6-
diisopropylaniline (X = Cl,[30] Br,[31] I[32]), N-(phenyl)-9,10-iminophenanthrenequinone[17] and
2,6-di-tert-butyl-9,10-phenanthrenequinone[33] were prepared according to literature procedures.
All other chemicals and solvents were obtained from commercial sources and used as received.
X-ray crystallography: data were measured on either a Bruker SMART Apex Duo Kappa (1-
4, 8) or a Bruker D8 Quest Eco (5, 6, 7, 9, 10) diffractometer at 100 K using an Oxford Cobra
(Duo) or Cryostream (Eco) with samples mounted on a MiTeGen microloop. Bruker APEX
software[34] was used to collect and reduce data and determine the space group. Absorption
This article is protected by copyright. All rights reserved.
28

10.1002/ejoc.201700884
European Journal of Organic Chemistry
corrections were applied using SADABS 2014.[35] Structures were solved with the XT structure
solution program[36] using Intrinsic Phasing and refined with the XL refinement package[37] using
Least Squares minimisation in Olex2.[38] All non-hydrogen atoms were refined anisotropically.
Hydrogen atoms were assigned to calculated positions using a riding model with appropriately
fixed isotropic thermal parameters. Crystal data, details of data collections and refinement are
given in Table 5 and Table S1 for 9 and 10. Further refinement details are given in the
Supporting Information. CCDC 1504190-1504197 and 1520301-1520302 contains the
supplementary crystallographic data for this paper. These data can be obtained free of charge
from The Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
DFT calculations were carried out in Gaussian09[39] with the meta-hybrid M06-2X
functional[40] recommended for study of non-covalent interactions and def2-SVP basis set.[41] All
calculations of interaction energy used the counterpoise method to account for basis set
superposition error,[42] and positions of hydrogen nuclei normalised to optimal positions by
partial geometry optimisation with all heavy-atom positions fixed. Converged molecular orbitals
were obtained from these calculations and used for topological analysis of the resulting electron
density using the AIMAll package.[43]
General Synthesis of mono substituted iminophenanthrenequinones.
Phenanthrenequinone (1 eq.) and the appropriate aryl amine (ca. 1.2 eq.) were dissolved in
methanol. To this reaction catalytic amounts of formic acid were added and refluxed for 12 h.
The reaction was diluted in hexane and the solvent evaporated. The crude sample was eluted
This article is protected by copyright. All rights reserved.
29