Polymorph crystal packing effects on charge transfer emission in the solid state

Article abstract of DOI:10.1039/c5sc01151e

Condensation of 1,8-naphthalic anhydride with N,N-(dimethylamino)aniline produced the donor-acceptor compound DMIM, which crystallised from a chloroform-diethyl ether mixture to afford two different coloured crystal polymorphs. Crystals for one polymorph are small and green, whereas the other crystals are orange and needle-like. X-ray crystal structures for both polymorphs were determined. The donor N,N-dimethylaniline and acceptor naphthalimide groups are twisted with respect to each other; the degree of twist is marginally different for the two structures. The orange crystal polymorph crystallises in the monoclinic space group C2/c and contains two slightly different molecular conformers in the unit cell (calculated density is 1.410 g cm^-3). The green crystal polymorph crystallises in the triclinic space group P1 and contains only one type of molecule in the unit cell (calculated density is 1.401 g cm^-3). The crystal packing motifs for the two polymorphs are subtly different, explaining the small variance in the observed densities. Very weak room temperature emission was observed for DMIM in a CHCl₃ solution, but crystals deposited on a glass slide glowed when irradiated at 488 nm using a fluorescence microscope. Disparate solid-state emission spectra and lifetimes for the two polymorphic crystal forms are observed for the dyad. The emission is assigned to charge recombination fluorescence from a charge transfer state. This journal is

Full text of DOI:10.1039/c5sc01151e

Chemical
Science
View Article Online
View Journal | View Issue
EDGE ARTICLE
Polymorph crystal packing effects on charge
transfer emission in the solid state†
Cite this: Chem. Sci., 2015, 6, 3525
a
a
Xiaoyan He, Andrew C. Benniston, Hanna Saarenpaa,b Helge Lemmetyinen,b
*
¨¨
Nikolia V. Tkachenko and Ulrich Baischac
b
*
Condensation of 1,8-naphthalic anhydride with N,N-(dimethylamino)aniline produced the donor–acceptor
compound DMIM, which crystallised from a chloroform–diethyl ether mixture to afford two different
coloured crystal polymorphs. Crystals for one polymorph are small and green, whereas the other crystals
are orange and needle-like. X-ray crystal structures for both polymorphs were determined. The donor
N,N-dimethylaniline and acceptor naphthalimide groups are twisted with respect to each other; the
degree of twist is marginally different for the two structures. The orange crystal polymorph crystallises in
the monoclinic space group C2/c and contains two slightly different molecular conformers in the unit
cell (calculated density is 1.410 g cmꢀ3). The green crystal polymorph crystallises in the triclinic space
group P1 and contains only one type of molecule in the unit cell (calculated density is 1.401 g cmꢀ3). The
ꢀ
crystal packing motifs for the two polymorphs are subtly different, explaining the small variance in the
observed densities. Very weak room temperature emission was observed for DMIM in a CHCl3 solution,
but crystals deposited on a glass slide glowed when irradiated at 488 nm using a fluorescence
microscope. Disparate solid-state emission spectra and lifetimes for the two polymorphic crystal forms
are observed for the dyad. The emission is assigned to charge recombination fluorescence from a
charge transfer state.
Received 31st March 2015
Accepted 18th April 2015
DOI: 10.1039/c5sc01151e
www.rsc.org/chemicalscience
A highly topical eld is that of solid state emitters, and the
coined phenomenon of aggregation-induced emission
enhancement observed for specially packed molecules.6 For
Introduction
Polymorphism in molecular crystals relates to the dissimilar
spatial and orientation arrangement of identical molecules
within the unit cell.1 Several factors such as temperature,
solvent mixtures, pH, speed of crystallisation, seeding can affect
the structure of the nal polymorphic crystal.2 The inevitable
disparity in crystal packing associated with two polymorphic
crystals regularly results in, for example, subtle differences in
their colour, morphology and solubility.3 Crystals oen display
diversity in their bulk properties (e.g., dielectric, magnetic,
mechanical behavior).4 However, the interest in polymorphic
forms was usually conned to pharmaceutical and inorganic
ionic or metallic compounds. More recently research has
focussed on optically active or metal–organic hybrid crystals,
and how polymorphism affects their properties.5
solid state emitters the study of polymorphism is proving highly
rewarding, especially considering that structural changes in
crystal morphology results in subtle optical effects.7 For
example, polymorphic crystals for a dinuclear rhenium complex
show very different absorption and emission spectra.8 The
perturbation of the photophysical properties of the complex is
attributed to alterations in the local organisation of the
molecular dipoles. The co-crystal strategy is also nding appeal
for tuning the emission properties of solid state materials.9 An
especially noticeable feature of the molecular systems studied
to date is the spatial localisation of the excited state. That is, any
alteration in charge distribution is over a limited distance set by
the closest contact separation of, for example, molecular part-
ners in dimers. One pertinent question to ask is over what
distance can charge be made to migrate in a single molecular
system in the crystalline state? Recently we demonstrated
energy transfer in a crystal using a quaterthiophene-Bodipy
aMolecular Photonics Laboratory, School of Chemistry, Newcastle University,
Newcastle upon Tyne, NE1 7RU, UK
bDepartment of Chemistry and Bioengineering, Tampere University of Technology,
molecular dyad.10 Contributions by the dipole–dipole Forster
¨
and Dexter-type dual electron exchange mechanisms are
possible within the single molecular entity. In the search for a
basic molecular system to exhibit unequivocal uni-directional
charge transfer (CT) our attention turned to basic donor–
acceptor systems. There was precedent that this approach could
Tampere, Finland
cDepartment of Chemistry, University of Malta, Msida, MSD 2080, Malta
† Electronic supplementary information (ESI) available: X-ray data, temperature
dependent uorescence spectra, FLIM data, uorescence microscope images
and calculations. CCDC 1003612 and 1003613. For ESI and crystallographic
data in CIF or other electronic format see DOI: 10.1039/c5sc01151e
This journal is © The Royal Society of Chemistry 2015
Chem. Sci., 2015, 6, 3525–3532 | 3525

View Article Online
Chemical Science
Edge Article
reward results. Very early work by Kozankiewicz11 identied long- 124.14, 123.43, 113.19, 77.68, 77.36, 77.04, 40.93 ppm. FT-MS + p
lived emission in crystalline bimolecular charge-transfer NSI: m/z calcd for [M + H]+ ¼ 317.1285; fnd ¼ 317.1288. Elemental
complexes. Dual uorescence and intramolecular CT within analysis calcd (%) for C20H16N2O2: C 75.93, H 5.10, N 8.85; fnd C
crystalline 4-(diisopropylamino)benzonitrile is also known.12 The 75.10, H 5.12, N 8.83. IR 1698, 1661 cmꢀ1 (n C]O); 1608, 1584,
basic donor–acceptor dyad we identied was DMIM (Scheme 1), 1521 cmꢀ1 (nC]C of aromatic ring). Melting point: 331–332 ꢁC.
which incorporates the N,N-dimethylaniline donor in close
Steady state emission spectra were collected using a Hitachi
proximity to a naphthalimide acceptor. Light activation was F-4500 spectrometer. Crystals for each polymorph were carefully
envisaged to promote CT along the molecular axis to generate an sandwiched between two clean glass slides and aligned in the
excited state that would collapse back to the ground state with spectrometer and the output signal optimised. Spectra were
coupled emission. This so called charge recombination uores- collected and averaged using the available spectrometer so-
cence is a well-known phenomenon for donor–acceptor mole- ware. The background emission spectrum from the glass slide
cules in solution and has a detailed theoretical basis.13 In fact, the was used for subtraction of scattered light and spurious uo-
molecular system displayed distinct long-wavelength emission rescence. For temperature dependence studies samples of the
only in the crystalline state. Luminescence was not associated crystals were ground with dry KBr and pressed into a thin disc
with localised emission from the individual organic components. which was placed in a thermostated sample holder connected to
Furthermore, the dyad crystallised in two polymorphs which a thermocouple. The disc was heated to set temperatures and
afforded very different emission spectra and lifetimes.
le for ca. 20 min to equilibrate before recording uorescence
spectra. Fluorescence lifetime microscope MicroTime-200
(PicoQuant GmBH) was used to acquire uorescence lifetime
images and measure emission decays.
Experimental
Cyclic voltammetry experiments were performed using a fully
automated HCH Instruments Electrochemical Analyzer and a
three electrode set-up consisting of a glassy carbon working
electrode, a platinum wire counter electrode and an Ag/AgCl
reference electrode. All studies were performed in deoxygenated
DCM containing TBATFB (0.2 M) as background electrolyte.
Redox potentials were reproducible to within ꢂ15 mV.
All chemicals were purchased from Aldrich Chemical Co. and
were used as received unless stated otherwise. DMF was dried
1
over 4A molecular sieves. H- and 13C-NMR spectra were recor-
ded with a JEOL 400 MHz spectrometer. Residual protiated
solvent was used as reference for chemical shi of 1H- and
13C-NMR spectra.
X-ray crystallographic data for O-DMIM and G-DMIM were
collected on an Oxford Diffraction Gemini A Ultra diffractom-
Preparation of DMIM
˚
eter at 150 K using Cu Ka radiation (l ¼ 1.54184 A). Empirical
A round-bottomed ask was charged with 1,8-naphthalic anhy-
dride (793 mg, 4.00 mmol), N,N-dimethyl-p-phenylenediamine
(562 mg, 4.00 mmol), DMF (10 mL) and molecular sieves (1.00 g).
The mixture (dark brown) was stirred at 130ꢁ for 4 hours. Then
the mixture was cooled to RT and during the cooling process a
yellow/green solid precipitated. Excess DCM was added to
dissolve the solid and the molecular sieves were ltered off to
afford a dark brown solution. The crude product precipitated out
again aer removal of DCM on a rotory evaporator. The DMF
solution was ltered to afford a yellow/green solid, which was
washed with a small amount of diethyl ether before drying under
vacuum. The crude product was puried by recrystallisation from
CHCl3 and Et2O through a vapour diffusion method to afford an
orange solid (1.00 g, 79% yield). 1H NMR (400 MHz, CDCl3):
d ¼ 8.64–8.65 (d, J ¼ 7.2 Hz, 2H; H of Napth), 8.24–8.26 (d, J ¼ 8.0
Hz, 2H; H of Napth), 7.76–7.80 (t, J ¼ 8.0 Hz, 2H; H of Napth),
7.15–7.17 (d, J ¼ 9.2 Hz, 2H; H of Ar), 6.84–6.86 (d, J ¼ 9.2 Hz, 2H;
H of Ar), 3.02 (s, 6H; H of CH3). 13C NMR (101 MHz, CDCl3): d ¼
165.11, 150.80, 134.33, 132.05, 131.85, 129.15, 128.85, 127.29,
absorption correction using spherical harmonics, implemented
in SCALE3 ABSPACK14 scaling algorithm were applied. Struc-
tures were solved by direct methods and rened on all unique F2
values, with anisotropic non-H atoms or as constrained riding
isotropic H atoms. Programs were CrysAlisPro15 for data
collection, integration, and absorption corrections as well as
OLEX2 (ref. 16) or SHELXTL14 for structure solution, renement,
and graphics. Full details about crystallographic experimental
information is provided as supplementary material.
Computational calculations were performed using a 32 bit
version of Gaussian09 (ref. 17) on a quadruple-core Intel Xeon
system with 4 GB RAM. The calculations were run in parallel,
fully utilising the multi-core processor. Energy minimisation
calculations were monitored using Molden and run in parallel
with frequency calculations to ensure optimised geometries
represented local minima. Time-dependent density functional
Theory (TD-DFT) calculations to simulate absorption spectra
were carried out using B3LYP and the 6-31G+(3df) basis set
(nstates ¼ 16) and an IEFPCM solvent model. The simulated
absorption spectrum was read with GView and the peak half-
width adjusted to match the observed spectrum.
Results and discussion
Synthesis and structure
The procedure for preparation of the molecular unit, DMIM, is
Scheme
1 Preparation of the N,N-dimethylanilinonaphthalimide
DMIM.
shown in Scheme 1. Heating a sample of 1,8-napthalic
3526 | Chem. Sci., 2015, 6, 3525–3532
This journal is © The Royal Society of Chemistry 2015

View Article Online
Edge Article
Chemical Science
Table 1 Selected bond lengths (A˚) and angles (ꢁ) for O-DMIM
anhydride 1 and p-(dimethylamino)aniline 2 in dry DMF affor-
ded a yellow solid which was puried by recrystallization. All
standard analyses of the sample (e.g., NMR spectroscopy, mass
spectrometry, combustion analysis) were consistent with the
proposed structure.
Slow vapour diffusion of Et2O into a sample of DMIM dis-
solved in CHCl3 produced rstly green crystals (G-DMIM).
Leaving the sample over several days to allow for slow solvent
evaporation afforded needle-like orange crystals (O-DMIM).
Observation of different coloured crystals is commonly associ-
ated with polymorphs. The green crystals likely represent the
kinetically favoured product. The single-crystals were subjected
to X-ray diffraction analysis, and no solvent molecules were
identied in the crystal lattice. The orange crystal polymorph
crystallises in the monoclinic space group C2/c and contains
two slightly different molecular conformers in the unit cell
(calculated density is 1.410 g cmꢀ3). The green crystal poly-
a
˚
Bond length /A
Bond anglesa/ꢁ
N(1)–C(1)
N(1)–C(3)
N(2)–C(9)
N(3)–C(21) 1.4448(18)
N(3)–C(22) 1.376(2)
N(4)–C(26) 1.4054(13)
O(1)–C(9)
O(3)–C(26) 1.2139(14)
N(1)–C(2)
N(2)–C(6)
N(2)–C(20) 1.4065(15)
N(3)–C(21A) 1.4449(18)
N(4)–C(25) 1.452(2)
N(4)–C(26A) 1.4055(13)
O(2)–C(20) 1.2152(14)
1.4433(18)
1.3750(16)
1.4034(15)
C(1)–N(1)–C(2)
C(1)–N(1)–C(3)
C(2)–N(1)–C(3)
C(6)–N(2)–C(9)
C(6)–N(2)–C(20)
C(9)–N(2)–C(20)
C(21)–N(3)–C(21A)
C(21)–N(3)–C(22)
C(21A)–N(3)–C(22)
C(25)–N(4)–C(26)
C(25)–N(4)–C(26A)
C(26)–N(4)–C(26A)
C(20)–N(2)–C(6)–C(5)
119.22(11)
120.29(12)
120.34(11)
117.39(9)
117.89(9)
124.71(10)
119.06(18)
120.47(9)
120.47(9)
117.53(7)
117.53(7)
124.94(14)
69.14(14)
1.2154(14)
1.4433(19)
1.4460(15)
C(26)–N(4)–C(25)–C(24) ꢀ84.85(9)
a
Standard deviations in bracket.
ꢀ
morph crystallises in the triclinic space group P1 and contains
only one type of molecule in the unit cell (calculated density is
1.401 g cmꢀ3). The molecular structure for O-DMIM is shown in
Fig. 1 and selected bond lengths and angles are collected in
Table 1. There are two slightly different DMIM molecules within
the unit cell. The subtle difference between the two structures is
related to the dihedral angle (q) between planes created using
the naphthalimide and N,N-dimethylaniline units. In molecule
A this angle is 69.41ꢁ whereas the angle is 84.85ꢁ for structure B
(Fig. 1). Only one type of molecule is observed in the crystal
structure for G-DMIM (see ESI†) and q ¼ 74.94ꢁ, which is
evidently similar to q measured in structure A for O-DMIM. The
intramolecular distances from the amino nitrogen (N1, N3) to
the centroid of the naphthalimide (red dots) are both around
between the central naphthalimide C11 atoms is either 7.81 or
˚
9.86 A (Fig. 2B). Intermolecular interactions are therefore clearly
present only in O-DMIM in form of p–p-stacking whereas in
G-DMIM no such interaction was observed; only very weak
C(H)–p interactions with a nearest distance of not less than
˚
3.5 A were measured between aromatic CH donors and both the
naphthalimide and aniline centroids.18 More differences
between the two polymorphs become evident especially when
analysing the planarity of the naphthalimide groups. Ideally,
torsion angles between the keto groups and the aromatic
carbon atoms should be zero due to the delocalised p electron
˚
6.2 A.
The differentiation between the two polymorphs is best
viewed in their crystal packing diagrams (Fig. 2). Whereas
molecules in O-DMIM are arranged in a way so that selected
naphthalimides face each other (distances between the two
˚
naphthalimide units ¼ 3.73–3.75 A), molecules in G-DMIM do
not show this form of stacking. The corresponding distance
Fig. 2 Selected crystal packing diagrams and intermolecular distances
Fig. 1 Molecular structure for O-DMIM showing the two related in A˚ for O-DMIM (A) and G-DMIM (B). Note: the 9.92 A˚ distance for
molecules in the unit cell. Note: red dots represent the centroids structure B is the closest distance where the two naphthalimide units
discussed in the main text.
are parallel.
This journal is © The Royal Society of Chemistry 2015
Chem. Sci., 2015, 6, 3525–3532 | 3527

View Article Online
Chemical Science
Edge Article
system. However, in G-DMIM the torsion angle O1–C1–C2–C3 is
3.01ꢁ whereas in O-DMIM the corresponding torsion angles are
1.81ꢁ and 5.52ꢁ. No H-bond interactions are observed in both
˚
˚
polymorphs in the range D/A 2.5 A–3.2 A. Intermolecular
interactions above this range are considered to be very weak,
and thus were not considered to be of sufficient inuence for
this study.19
Cyclic voltammetry
The cyclic voltammogram recorded for DMIM (Fig. 3) in dry
DCM (0.2 M TBATFB) contained upon oxidative scanning a
quasi-reversible wave at E1/2 ¼ +1.1 V (80 mV) vs. Ag/AgCl. This
wave is associated with one electron oxidation of the dimethy-
lamino group. Upon reductive scanning a quasi-reversible wave
was observed at E1/2 ¼ ꢀ1.19 (120 mV) vs. Ag/AgCl, and is one-
electron addition to the naphthalimide group. The energy
difference (DE) between the two waves is 2.29 V (18 470 cmꢀ1).
Fig. 4 Representation of the HOMO and LUMO for DMIM calculated
using DFT (B3LYP) and a 6-311G+(3df) basis set in a cyclohexane
continuum solvent model.
observed in the oscillator strengths (q ¼ 69.43ꢁ, lCT ¼ 594 nm,
f ¼ 0.0011; q ¼ 71.47ꢁ, lCT ¼ 560 nm, f ¼ 0.0009). Evidently in
solution where q can vary because of rotation at the aryl–aryl
connector bond the CT transition should become more
discernible. The rst major contribution to the absorption
prole was calculated at lmax ¼ 341 nm (f ¼ 0.29) and is
assigned to a localised p–p* transition on the naphthalimide
subunit.
Computer calculations
The ground-state structure for DMIM was calculated using DFT
(B3LYP) and a 6-311G+(3df) basis set in a cyclohexane solvent
continuum. As illustrated in Fig. 4 the HOMO clearly resides on
the N,N-dimethylaniline group while the LUMO is localised on
the naphthalimide subunit; the HOMO–LUMO energy gap is
2.72 eV (21 938 cmꢀ1). One point to note is the near perfect
orthogonality of the two subunits for the calculated structure
(q ¼ 89.98ꢁ). At rst glance the HOMO to LUMO electronic
transition would appear to represent an intramolecular charge
transfer process, and we were interested to see if this in turn was
a contributing factor to the dyad's electronic absorption spec-
trum. Thus, using time-dependent density functional theory
(TD-DFT) the electronic absorption spectrum for DMIM was
calculated in a cyclohexane solvent bath. A partial breakdown of
the calculated electronic transitions and the accompanying
orbital contributions is given in ESI.† In summary, the identi-
ed HOMO to LUMO transition is located at lCT ¼ 551 nm,
however the oscillator strength (f) is zero. This is not too
surprising considering the orthogonal nature of the two
contributing molecular orbitals. In fact, by setting the dihedral
angles (q) to those observed for the X-ray determined structures,
and repeating the TD-DFT calculations, a slight increase is
Spectroscopy
The room temperature electronic absorption spectrum for a
cyclohexane solution of DMIM is shown in Fig. 5. In light of the
TD-DFT calculations the series of sharp bands below 350 nm are
mainly assigned to localised p–p* electronic transitions on the
naphthalimide group. The uorescence of DMIM in solution
was very weak with a quantum yield close to 0.02% and a
maximum at 366 nm (see ESI†). By contrast, a broad uores-
cence prole, centred at lEM ¼ 605 nm, was seen for crystalline
O-DMIM (Fig. 5). In comparison, the uorescence spectrum for
G-DMIM is blue shied with lEM ¼ 549 nm. Fluorescence
Fig. 5 Room temperature absorption spectrum for DMIM in dilute
cyclohexane and the expansion of the region around 400 nm. The
Fig. 3 Cyclic voltammogram for DMIM in dry DCM (0.2 M TBATFB) at a fluorescence spectrum recorded for crystalline O-DMIM (red) and
glassy carbon working electrode vs. Ag/AgCl. Scan rate ¼ 50 mV sꢀ1
.
G-DMIM (green).
3528 | Chem. Sci., 2015, 6, 3525–3532
This journal is © The Royal Society of Chemistry 2015

View Article Online
Edge Article
Chemical Science
microscope images were also collected for the crystals deposited
on a glass slide (see ESI†), and displayed clear sharp uorescent
blocks corresponding to clusters of crystals. Importantly, both
proles cannot be assigned to the structured emission seen in
solution at around lEM ¼ 375 nm for previously reported
derivatives.20 The observed uorescence is unique to the crys-
talline samples and does not originate from spurious impuri-
ties. Re-puried and recrystallized samples afforded the same
results.
Emission quantum yields for the crystals were estimated
using
a front-face illumination method, assuming that
all excitation light (at 370 nm) was absorbed by the sample,
and comparing integral emission spectra with that of a
high concentration of (4-(dicyanomethylene)-2-methyl-6-(p-
dimethylaminostyryl)-4H-pyran) standard in ethanol in a 1 mm
cuvette and measured in an identical front-face arrangement.
The measured quantum yields are 1% and 3.5% for G-DMIM
and O-DMIM, respectively.
Fig. 6 Plots to demonstrate the dependency of total fluorescence
intensity versus 1/T ( ¼ G-DMIM; ¼ O-DMIM). Lines are drawn to
help show the deviation points. Insert depicts fluorescence spectra for
DMIM with increasing temperature.
It might appear that strong intermolecular interactions are
responsible for new emission bands of the crystalline samples.
Hence, the same intermolecular interactions may affect
absorption spectra. Although the absorption spectra of crystal-
line samples could not be measured directly, uorescence
excitation spectra were recorded and provide some information
on perturbations of the absorption spectra. The measurements
were carried out at two monitoring wavelengths of 600 and
660 nm, and indicated appearance of a new absorption bands in
the visible part of the spectrum with maxima at roughly 535 and
560 nm for G-DMIM and O-DMIM, respectively (see ESI†). It is
worth noting that these maxima are in fairly good agreement
with calculated lCT values from TD-DFT for intramolecular
charge transfer.
Fluorescence spectra were also collected for both polymorphs
dispersed in a dried KBr disc over a modest temperature range
(293 K–568 K) and below the melting point of DMIM. For both
cases uorescence decreases with an increase in temperature
and is accompanied by a slight blue-shi in the emission
maximum.‡ There is, however, a striking difference in the uo-
rescence temperature dependence for the two polymorphs, that
is easily observed in basic ln(total intensity) vs. 1/T plots (Fig. 6).
For G-DMIM the uorescence temperature dependence at lower
temperatures (<400 K) is far more pronounced than in the O-
DMIM case. There is a clear thermally activated process which
controls the uorescence output for both crystals. Such an idea
is in line with previously reported photoluminescence temper-
ature dependence of single crystals.21 An inexion point is also
seen in the plots for both the crystals (G-DMIM ꢃ417 K, O-DMIM
ꢃ476 K). The rst temperature must correspond to the poly-
morphic transition of G-DMIM to O-DMIM,22 conrming the
green crystal is the kinetic product. Since a polymorph transition
temperature is also seen for O-DMIM then presumably a third
polymorphic structure also exists but can be only accessed at
high temperatures. A high activation barrier is presumably the
reason why the third polymorph is not observed at room
temperature crystallisation.
prole and image for O-DMIM are shown in Fig. 7. The uo-
rescence decay prole was best t to a tri-exponential model
(sobs ¼ A1 exp(ꢀt/s1) + A2 exp(ꢀt/s2) + A3 exp(ꢀt/s3)). Data
collected on different crystals, orientations and from multiple
areas of a crystal could be analysed in an identical manner.
Values for lifetimes and pre-exponential factors were identical
within the accuracy of calculations. The major lifetime (s1) is
1.20 ns and represents 92% of the uorescence decay prole.
The two longer lifetime components contribute to the remain-
ing fraction of the decay. An identical lifetime imaging experi-
ment performed on G-DMIM (see ESI†) resulted in decay prole
that was similarly analysed as a tri-exponential. However, s1 is
reduced signicantly to only 350 ps, but the A1 value is still
comparable to the case for O-DMIM (Table 2).
Fig. 7 Room temperature emission decay curve and instrument
response function recorded for O-DMIM and the least-squares fit to a
tri-exponential (red line). Insert shows a typical fluorescence lifetime
Picosecond uorescence lifetime imaging experiments were
performed on both crystal polymorphs. A representative decay image taken for the crystal sample.
This journal is © The Royal Society of Chemistry 2015
Chem. Sci., 2015, 6, 3525–3532 | 3529

View Article Online
Chemical Science
Edge Article
Considering the rather large difference in the lifetimes the
contribution of each decay component to the steady state
emission cannot be concluded based on relative values of the
pre-exponential factors only. Instead, products siAi must be
compared to deduce the effect of each component on the
steady-state emission spectrum. This gives 53, 38 and 29%
contribution of the fast, middle and slow component to the
emission spectrum of O-DMIM, and 29, 13 and 48% contribu-
tion to the spectrum of G-DMIM. One can suspect that different
components have different origins. To answer this question the
decays were measured in a wide spectrum range with constant
signal collection time and tted globally to obtain so-called
decay component associated spectra (see ESI†). The component
spectra are virtually identical for O-DMIM, which leads to
conclusion that only one type of emissive and spectrally
distinguishable type of molecular assembly is formed. In the
case of G-DMIM the fast component is blue shied relative to
the longer-lived components. Thus, for this crystalline form two
types of emissive arrangements can be expected.
Fig. 8 Fits of the emission spectra for G-DMIM and O-DMIM to a
charge-transfer model. The measured spectra are shown by the solid
lines and fits by the dashed lines.
0.002 eV and SG ¼ 2.47 ꢂ 0.02, and DGO ¼ 2.463 ꢂ 0.002 eV and
SO ¼ 0.38 ꢂ 0.03 for G-DMIM and O-DMIM, respectively.
Qualitatively, the difference in the free energy is apparent from
the different positions of the emission bands. The difference in
electron-vibrational coupling is also expected since the spectra
have different shapes, though the calculations suggest the
difference to be more than six fold. The latter means that the
internal reorganization energy associated with the CS state
relaxation to the ground state is more than six times larger for
G-DMIM than for O-DMIM, being 0.33 and 0.05 eV respectively.
This result is in agreement with earlier suggestion that G-DMIM
is kinetic product whereas O-DMIM has a more thermody-
namically stable structure which requires less reorganization
when switching from the CS state to the ground state. The
difference in the basic potential energy surfaces for the two
polymorphs is illustrated in Fig. 9, using the calculated
Interpretation
In rigid medium the reorganization energy, l, associated with
partial charge separation is expected to be relatively low,
whereas the energy of the charge separated (CS) state, DG, for
DMIM is relatively large according to cyclic voltammetry
measurements, 2.29 eV. Therefore, the charge recombination is
expected to be in the inverted Marcus regime and can follow
radiative decay route in addition to the non-radiative one. In
this case, the emission band shape analysis can be used to
estimate energetic parameters of the CS states, namely the free
energy, DG, the outer sphere reorganization energy, l, the
vibrational energy, Ev, and the electron-vibrational coupling,
S (see ESI†).23 This type of analysis was applied to the emission
spectra for both G-DMIM and O-DMIM. To reduce the number
of t parameters we assumed that the vibrational and outer
sphere reorganization energies are the same for both samples,
since they are composed of identical molecules, the environ-
ment is rigid and the unit cell consists of the same number of
molecules. Though the free energy and the electron-vibrational
coupling can be different, considering that the molecules have
different arrangements in the crystals, the internal reorganiza-
tion energy, which contributes to S, can be different. The results
of spectral tting are presented in Fig. 8.
The ts afforded l ¼ 0.400 ꢂ 0.002 eV and Ev ¼ 0.132 ꢂ 0.002
eV as common parameters for both samples, and DGG ¼ 2.873 ꢂ
Table 2 Emission lifetimes and pre-exponential parameters measured
for the polymorphic crystals
Polymorph
s1,c ns
A1
s2,c ns
A2
s3,c ns
A3
O-DMIMa
G-DMIMb
1.20(0.06)
0.35(0.01)
0.92
0.94
6.4(1.5)
2.2(0.1)
0.06
0.05
31(5)
40(7)
0.02
0.01
Fig. 9 Constructed simplified potential energy surfaces diagram using
determined parameters to show the difference between the emitting
states for the two crystal forms. GS ¼ ground state, CTS ¼ charge
transfer state.
a
Orange crystal polymorph. b Green crystal polymorph. c Error in least-
squares t is given in brackets.
3530 | Chem. Sci., 2015, 6, 3525–3532
This journal is © The Royal Society of Chemistry 2015

View Article Online
Edge Article
Chemical Science
parameters obtained from the ts. The value of free energy for
O-DMIM is slightly higher than that estimated from electro-
chemical measurements in solution. The difference can be
attributed to the coulombic interaction which increases the CS
state energy, and is not accounted for in electrochemical
measurements. Even higher free energy of G-DMIM may arise
from somewhat different degree of charge separation which
affects the coulombic term directly.
1 J. Bernstein, Polymorphism in Molecular Crystals, Oxford
University Press, 2007; R. J. Davey, N. Blagden, G. D. Potts
and R. Docherty, J. Am. Chem. Soc., 1997, 119, 1767;
A. Mahieu, J.-F. Willart, E. Dudognon, M. D. Eddleston,
`
W. Jones, F. Danede and M. Descamps, J. Pharm. Sci., 2013,
102, 462.
2 Y. V. Seryotkin, T. N. Drebushchak and E. V. Boldyreva, Acta
Crystallogr., Sect. B: Struct. Sci., Cryst. Eng. Mater., 2013, 69,
77; V. V. Krisyuk, I. A. Baidina, I. V. Korolkov,
P. P. Semyannikov, P. A. Stabnikov, S. V. Trubin and
A. E. Turgambaeva, Polyhedron, 2013, 49, 1; G. J. Han,
S. Thirunahari, P. S. Chow and R. B. H. Tan,
CrystEngComm, 2013, 15, 1218; C. S. Towler, R. J. Davey,
R. W. Lancaster and C. J. Price, J. Am. Chem. Soc., 2004,
126, 13347.
Conclusions
The present study visibly demonstrates that the emission
properties for polymorphic crystals of a charge-transfer molec-
ular system can be highly sensitive to packing within the crystal
lattice. The origin of the effect for DMIM appears to be alter-
ations in rate constants for non-radiative decay. A simple
calculation using data collected from the emission prole ts
predicts a ca. 2.6 fold higher value of the non-radiative rate
constant for G-DMIM (see ESI†). By assuming radiative rate
constants are similar for both polymorphs§ the discrepency
from lifetime measurements is around 3.5 fold. The agreement
is good considering all the assumptions used in the interpre-
tation. That a thermally activated process affects uorescence
output for each crystal is indeed consistent with an electron
transfer process. As discussed, the shi in the emission prole
is an environment effect (e.g., alteration in packing) associated
with a perturbation in energy of the emitting state. Prior work
has shown that different packing arrangements within
p-conjugated molecular systems manipulate luminescence,
because of alterations in relative orientations and overlap
between chromophores.24 One key point to note for DMIM,
which is somewhat distinctive to other literature examples, is
the observation of photoinduced electron transfer within the
dyad in the crystalline state. In DMIM this distance is around
´
3 T. M. R. Maria, M. E. S. Eusebio, J. A. Silva, A. J. F. N. Sobral,
˜
C. Cardoso, J. A. Paixao and M. R. Silva, J. Mol. Struct., 2012,
1030, 67; L. Yu, Acc. Chem. Res., 2010, 43, 1257.
´
´
4 N. Maso, H. Beltran, E. Cordoncillo, D. C. Sinclair and
A. R. West, J. Am. Ceram. Soc., 2008, 91, 144; M. C. Menard,
B. L. Drake, G. T. McCandless, K. R. Thomas,
R. D. Hembree, N. Haldolaarachchige, J. F. DiTusa,
D. P. Young and J. Y. Chan, Eur. J. Inorg. Chem., 2011,
3909; J. O. Williams and J. M. Thomas, Trans. Faraday Soc.,
1967, 63, 1720; J. M. Thomas and J. O. Williams, Trans.
Faraday Soc., 1967, 63, 1922.
5 A. Nangia, Acc. Chem. Res., 2008, 41, 595; V. S. S. Kumar,
A. Addlagatta, A. Nangia, W. T. Robinson, C. K. Broder,
R. Mondal, I. R. Evans, J. A. K. Howard and F. H. Allen,
Angew. Chem., Int. Ed., 2002, 41, 3848; S. Varughese, J.
Mater. Chem. C, 2014, 2, 3499.
6 T. He, X. Tao, J. Yang, D. Guo, H. Xia, J. Jia and M. Jiang,
Chem. Commun., 2011, 47, 2907; H. Li, Z. Chi, B. Xu,
X. Zhang, X. Li, S. Liu, Y. Zhang and J. Xu, J. Mater. Chem.,
2011, 21, 3760.
7 T. Mutai, H. Tomoda, T. Ohkawa, Y. Yabe and K. Araki,
Angew. Chem., Int. Ed., 2008, 47, 9522; J. Seo, S. Kim and
S. Y. Park, J. Am. Chem. Soc., 2004, 126, 11154; D. Yan and
D. G. Evans, Mater. Horiz., 2014, 1, 46.
8 E. Q. Procopio, M. Mauro, M. Panigati, D. Donghi,
P. Mercandelli, A. Sironi, G. D'Alfonso and L. De Cola, J.
Am. Chem. Soc., 2010, 132, 14397.
˚
6 A for the intramolecular process. In the reaction centre
˚
complex for natural photosynthesis the distance is around 25 A,
but it does rely on the protein blanket providing some stabili-
sation effect. To facilitate a similar charge separation in a crystal
environment will require both the correct molecular system and
its crystal packing. We expect to test such ideas in new dyad
systems, with the intention of engineering well-dened struc-
tures for thin-lm solar cell applications.25
ˇˇ ´
9 D. Yan, A. Delori, G. O. Lloyd, T. Friscic, G. M. Day, W. Jones,
J. Lu, M. Wei, D. G. Evans and X. Duan, Angew. Chem., Int.
Ed., 2011, 50, 12483; D. Yan, H. Yang, Q. Meng, H. Lin and
M. Wei, Adv. Funct. Mater., 2014, 24, 587; G. Fan and
D. Yan, Sci. Rep., 2014, 4, 4933.
Acknowledgements
We thank the Engineering and Physical Sciences Research
Council (EPSRC) for funding for the diffractometer and the
EPSRC sponsored Mass Spectrometry Service at Swansea for
collecting mass spectra.
10 A. C. Benniston, G. Copley, A. Harriman, D. B. Rewinska,
R. W. Harrington and W. Clegg, J. Am. Chem. Soc., 2008,
130, 7174.
11 B. Kozankiewicz, J. Al-Abbas and J. Prochorow, Chem. Phys.
Lett., 1995, 235, 456.
12 S. I. Druzhinin, A. Demeter and K. A. Zachariasse, Chem.
Phys. Lett., 2001, 347, 421.
13 I. R. Gould, D. Noukakis, L. Gomez-Jahn, R. H. Young,
J. L. Goodman and S. Farid, Chem. Phys., 1993, 176, 439.
Notes and references
‡ Part of the blue-shi is consistent with the expected change for a charge transfer
´
uorescence prole with temperature, see J. Cortes, H. Heitle and J. Jortner, J.
Phys. Chem., 1994, 98, 2527.
§ Using the measured quantum yields for both polymorphs the radiative rate
constants (kRAD ¼ fFLU/s) are both around 2.9 ꢄ 107 sꢀ1 in tting with the
hypothesis.
This journal is © The Royal Society of Chemistry 2015
Chem. Sci., 2015, 6, 3525–3532 | 3531

View Article Online
Chemical Science
Edge Article
¨
14 G. M. Sheldrick, Acta Crystallogr., Sect. A: Found. Crystallogr.,
2008, 64, 112.
O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski, and
D. J. Fox, Gaussian 09, Revision D.01, Gaussian, Inc.,
Wallingford CT, 2009.
15 CrysAlisPro, Agilent Technologies, 2010, p. Xcalibur CCD
system.
16 O. V. Dolomanov, L. J. Bourhis, R. J. Gildea, J. A. K. Howard
and H. Puschmann, J. Appl. Crystallogr., 2009, 42, 229.
18 E. A. Meyer, R. K. Castellano and F. Diederich, Angew. Chem.,
Int. Ed., 2003, 42, 1210.
19 T. Steiner, Angew. Chem., Int. Ed., 2002, 41, 48.
17 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, 20 C.-S. Choi, K.-S. Jeon, W.-C. Jeong and K.-H. Lee, Bull. Korean
M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, Chem. Soc., 2010, 31, 1375.
B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, 21 E. Simoni, S. Hubert and M. Genet, J. Lumin., 1994, 62, 139.
X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, 22 H. Unesaki, T. Kato, S. Watase, K. Matsukawa and K. Naka,
J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota,
R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,
O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery Jr,
Inorg. Chem., 2014, 53, 8270; N. Harada, S. Karasawa,
T. Matsumoto and N. Koga, Cryst. Growth Des., 2013, 13,
4705.
J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, 23 A. Kapturkiewicz, J. Herbich, J. Karpiuk and J. Nowacki, J.
E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, Phys. Chem. A, 1997, 101, 2332.
J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, 24 H. Zhang, Z. Zhang, K. Ye, J. Zhang and Y. Wang, Adv. Mater.,
S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam,
M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo,
J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev,
A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski,
2006, 18, 2369; Y. Zhao, H. Gao, Y. Fan, T. Zhou, Z. Su, Y. Liu
and Y. Wang, Adv. Mater., 2009, 21, 3165; M. Sase,
S. Yamaguchi, Y. Sagara, I. Yoshikawa, T. Mutai and
K. Araki, J. Mater. Chem., 2011, 21, 8347.
R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, 25 H. Dong, H. Zhu, Q. Meng, X. Gong and W. Hu, Chem. Soc.
P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels,
Rev., 2012, 41, 1754.
3532 | Chem. Sci., 2015, 6, 3525–3532
This journal is © The Royal Society of Chemistry 2015