Synthesis and photophysics of core-substituted naphthalene diimides: Fluorophores for single molecule applications

Article abstract of DOI:10.1002/asia.200900215

The synthesis and photophysics of two new aminopropenyl naphthalene diimide (SANDI) dyes are reported. A general and convenient method for the synthesis of the precursor mono-, di-, and tetrabrominated 1,4,5,8-naphthalene tetracarboxylic dianhydrides is described. The two coresubstituted SANDIs exhibit many of the photophysical properties required for fluorescence labeling applications including high photostability and high fluorescence quantum yields (>0.5) in the visible region of the spectrum. The emission wavelength is sensitive to the number of substituents on the NDI core, and the fluorescence decay times are in the range of -8-12 ns for both compounds in the solvents investigated. Preliminary fluorescence emission data from single molecules of the compounds embedded in poly(methyl methacrylate) films are also reported and show that single molecules have very low yields of photobleaching, particularly the di-substituted system. Furthermore, only a small proportion (<10%) of the single molecules studied display fluorescence intermittencies or "blinks" in their photon trajectory. The compounds appear to be excellent candidates for applications at the single molecule level, for example, as FRET labels.

Full text of DOI:10.1002/asia.200900215

FULL PAPERS
DOI: 10.1002/asia.200900215
Synthesis and Photophysics of Core-Substituted Naphthalene Diimides:
Fluorophores for Single Molecule Applications
Toby D. M. Bell,*[a] Sheryll Yap,[a] Chintan H. Jani,[b] Sheshanath V. Bhosale,[b]
Johan Hofkens,[c] Frans C. De Schryver,[c] Steven J. Langford,*[b] and
Kenneth P. Ghiggino*[a]
1542
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2009, 4, 1542 – 1550

Abstract: The synthesis and photophy-
sics of two new aminopropenyl naph-
thalene diimide (SANDI) dyes are re-
emission wavelength is sensitive to the
number of substituents on the NDI
core, and the fluorescence decay times
are in the range of ~8–12 ns for both
compounds in the solvents investigated.
Preliminary fluorescence emission data
from single molecules of the com-
pounds embedded in poly(methyl
methacrylate) films are also reported
and show that single molecules have
very low yields of photobleaching, par-
ticularly the di-substituted system. Fur-
ported.
A general and convenient
method for the synthesis of the precur-
sor mono-, di-, and tetrabrominated
1,4,5,8-naphthalene tetracarboxylic dia-
nhydrides is described. The two core-
substituted SANDIs exhibit many of
the photophysical properties required
for fluorescence labeling applications
including high photostability and high
fluorescence quantum yields (>0.5) in
the visible region of the spectrum. The
thermore, only a small proportion
(<10%) of the single molecules stud-
ied display fluorescence intermittencies
or “blinks” in their photon trajectory.
The compounds appear to be excellent
candidates for applications at the single
molecule level, for example, as FRET
labels.
Keywords: core-substitution
energy transfer fluorescence
·
·
·
naphthalene diimide · single mole-
cule fluorescence
Introduction
to pH or polarity), required for reporting on the behavior of
the substrate. There is an increasing need for new and im-
proved fluorophores capable of meeting the demands of the
ever-widening range of applications, particularly at the mo-
lecular level where many labels routinely used in ensemble
measurements are not suitable. Characteristics of a good
fluorescence label include a large absorption cross-section
(molar absorption coefficient, er) and high fluorescence
quantum yield, (Ff) in the visible or near IR spectral regions
for easy detection at low labeling levels. Fluorescence labels
for SM applications also need to be highly photostable and
thus able to withstand multiple emission/excitation cycles.
Ideally, at least 106 photons (cycles) should be emitted prior
to irreversible photobleaching of the molecule. Relatively
long fluorescence lifetimes (on the nanosecond scale) are
another useful property for SM labels to enable the dynam-
ics of processes in this time regime to be monitored. Fluo-
rescence intermittencies, or “blinking” arising from excur-
sions to non-fluorescent states or reversible quenching pro-
cesses, is considered an undesirable trait for fluorophores
used as SM labels. However, a favorable characteristic of
fluorophores is sensitivity of the emission spectrum to
changes in molecular structure of the core chromophore.
Thus, by altering substituents one can effectively tune the
emission of a molecule to accommodate various applications
such as suitability for fluorescence energy transfer experi-
ments.
There has been a rapid expansion in the use of single mole-
cule (SM) fluorescence techniques to address questions in
an increasing range of research areas, particularly in bio-
physics.[1–3] The attraction of SM studies is that information
regarding individual molecules is not lost arising from en-
semble-averaging, thus providing a detailed understanding
of the underlying distribution of behaviors that contribute
to the bulk sample properties.
Many spectroscopic and microscopic techniques for study-
ing molecular properties and dynamics are based on detect-
ing luminescence from the substrate. One approach to
studying materials which are not intrinsically fluorescent is
to attach a suitable fluorescent moiety to the substrate of in-
terest. Such fluorescent labels or “tags” offer the sensitivity
and environment-based modulation effects (e.g., in response
[a] Dr. T. D. M. Bell, S. Yap, Prof. K. P. Ghiggino
School of Chemistry and Bio21 Institute
The University of Melbourne
Parkville, Victoria, 3010 (Australia)
Fax : (+61)3-9348-1595
E-mail: tbell@unimelb.edu.au
[b] C. H. Jani, Dr. S. V. Bhosale, Prof. S. J. Langford
School of Chemistry
Monash University
Clayton, Victoria, 3800 (Australia)
[c] Prof. J. Hofkens, Prof. F. C. De Schryver
Department of Chemistry
Examples of compounds that fulfill these requirements
and are currently used as fluorescence labels include laser
dyes such as rhodamines, cyanines, oxazines, and imide de-
rivatives of rigid polynuclear aromatic hydrocarbons, such as
perylene. The latter class have proven particularly suited to
Katholieke Universiteit Leuven
200F Celestijnenlaan, 3001 Heverlee (Belgium)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/asia.200900215.
Chem. Asian J. 2009, 4, 1542 – 1550
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemasianj.org
1543

T. D. M. Bell, S. J. Langford, K. P. Ghiggino et al.
FULL PAPERS
SM spectroscopy owing to their high fluorescence quantum
yields (often>0.9) and photostability.[4,5] Unlike the larger
imide substituted rylene dyes (perylene, terrylene, and qua-
terrylene), unsubstituted naphthalene diimides (NDIs) have
low Ff values and emission is usually confined to the UV
region of the spectrum,[6,7] making them unsuitable as fluo-
rescence labels. The introduction of amino substituents,
however, to the ortho position of the imide functionality re-
sults in their transformation into colored, highly fluorescent
dyes[8–12] and makes them potentially useful fluorescence
label candidates. It is with this versatility in properties that
NDIs have found popularity in the latter half of the
20th century[13] arising, in part, from the pioneering work of
Vollmann in the early 1930s.[14]
The realization that NDIs can act as useful components
for the creation of supramolecular functional materi-
als[13,15–17] has transpired as a result of their desirable elec-
tronic and spectroscopic properties over pyromellitic di-
imides. Their enhanced solubility properties over perylene
diimide dyes also can assist synthetic manipulation and ex-
pands the range of potential applications. Vollmannꢁs syn-
thesis of 2,6-diarylamino core-substituted NDIs from 2,6-di-
chloro-1,4,5,8-naphthalenetetracarboxylic dianhydride which
itself was derived in four steps from pyrene, is still used as a
source of precursors.[13] His description of their fluorescent
properties as dull and non-fluorescent,[14] may well be the
reason why this class of core-substituted NDIs have not re-
ceived much attention until recently.[8–12,13,16] Herein, we
report on the synthesis and photophysical characterization
of two new substituted aminopropenyl naphthalene diimide
(SANDI) derivatives 1 and 2 (Scheme 1) which show poten-
tial as functionalizable fluorescence labels which absorb and
emit in the visible part of the spectrum.
in the molecules imparts further possibilities for synthetic
modification (including polymer formation) and increases
their potential for use in new applications.
Results and Discussion
Compounds 1 and 2 were prepared in good yield by the re-
action of allylamine with 2-bromo- and 2,6-dibromo-1,4,5,8-
naphthalenetetracarboxylic dianhydride (NDA), respective-
ly, following literature procedures (Scheme 2).[8] The impor-
tance of the per-halogenated core substituted NDAs such as
3 and 4, has recently lead to two reported syntheses involv-
ing molecular bromine[18] or dibromoisocyanuric acid
(DBI)[9,19] in oleum as the brominating agent. In our hands,
as in others,[9] the reaction involving molecular bromine
could not be reproduced in high yield and DBI is not readily
available in large quantity. We have recently discovered that
halogenation can be easily achieved using sodium bromide
in oleum (Scheme 2) and that pathways to the 2-bromo 4,
2,6-dibromo 3, and 2,3,6,7-tetrabromo-1,4,5,8-naphthalene
Scheme 1. Molecular structures of mono-allyl SANDI
1 and di-allyl
SANDI 2 used in this work. The terms mono- and di-allyl refer to core
substitution only.
We also report some preliminary SM data which indicate
that these molecules are sufficiently bright and photostable
to meet the exacting requirements of many SM experimen-
tal applications. Of particular note, is the low occurrence of
fluorescence blinking observed from SMs of 1 and, especial-
Scheme 2. The NaBr method leads to highly core-functionalized naphtha-
lene diimides in high yield from cheap and readily available starting ma-
terials.
ly,
2
when embedded in poly(methyl methacrylate)
(PMMA) films. The presence of the propenyl functionality
1544
www.chemasianj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2009, 4, 1542 – 1550

Fluorophores for Single Molecule Applications
tetracarboxylic acid dianhydrides 5 (Scheme 3) are achieva-
ble by varying the reaction conditions (temperature, time,
concentration) and stoichiometry of NaBr used.
Scheme 3. The NaBr method leads to the tetrabromo derivative in high
yield on multi-gram scale.
Reaction of NaBr (0.8 equivalents) with 1,4,5,8-naphthale-
netetracarboxylic dianhydride in oleum (20% free SO3) for
24 h at 1108C leads to the formation of the monobromo
NDA 4 in a moderate 25% yield (Scheme 2). The reaction
conditions reflect a compromise between actual yield and
ease of separation from starting dianhydride or the dibromo
NDA by-product 3 which occurs on further exposure to
NaBr or at higher temperature. Reaction of 4 with allyla-
mine in DMF at elevated temperature yields the diimide 1
in 70% yield after chromatography. Increasing the amount
of NaBr (2.2 equivalents) under similar conditions leads to
the clean formation of the dibromo NDA 3 as the major
product.[20] Reaction of 3 with allylamine in DMF yields the
diimide 2 in 70% yield after chromatography. It is worth
noting that a core trisubstituted NDI has been recently pre-
pared from the corresponding dichloro-NDI using a 1,2-dia-
mime,[10] however, none of this type of product was seen
under the conditions employed here. The efficiency of the
reaction to form the dibromo derivative manifests itself
from the reduced reactivity associated with further bromina-
tion. More forcing conditions and higher stoichiometry of
NaBr (4.4 equiv) does lead to the tetrabromo-NDA 5 in
88% yield (Scheme 3).[21] Hence by this method, a range of
core-substituted NDAs are possible.
Figure 1. a) Crystal structure of 2,3,6,7-tetrabromo-1,4,5,8-naphthalene
tetracarboxylic dianhydride 5 from DMSO solvent. b) A representation
of a portion of the crystal lattice of 5.DMSO. NDA=naphthalene dia-
nhydride.
Ensemble Photophysical Properties of 1 and 2
Figure 2 shows the steady-state absorption and emission
spectra of both 1 (upper panel) and 2 (lower panel) in tolu-
ene, ethanol, and acetonitrile. The absorption maxima in the
near-UV spectral region are typical of unsubstituted NDIs
and for such systems, this is the lowest energy transition
Crystals of 5 suitable for X-ray crystallography were
grown by vapor diffusion of water into a DMSO solution of
5.[22] The crystal structure (Figure 1) shows a twisted aromat-
ic core, which is unusual compared to other NDI struc-
tures.[6,23,24] The twist is imparted by the steric bulk of the
ortho bromines.
The torsion of the bromine atoms and the oxygen atoms
can be calculated as deviation from the mean plane of the
naphthalene ring in the molecule.[25] While Br(1) (Figure 1a)
sits in plane with the naphthalene core (deviation: 0.1 ꢂ),
Br(2) is out of the plane by 2.1 ꢂ. The carbonyl oxygens ad-
jacent to the bromines that twist out of the plane (O(3)) are
themselves out of the plane by 2.05 ꢂ arising from steric
strain. Furthermore, the anhydride oxygens are twisted out
of the plane substantially with a deviation of 4.6 ꢂ. The
molecules exhibit p-stacking interactions in the crystal lat-
tice (Figure 1b), with intermolecular separations of 3.5 ꢂ
between aromatic planes within the NDA, leading to chan-
nels filled by ordered DMSO solvent molecules.
Figure 2. Normalized steady state absorption and corrected emission
spectra of 1 (panel a) and 2 (panel b) in toluene (light grey, dot), acetoni-
trile (grey, dash), and ethanol (black, line). lex =450 and 500 nm, for
emission spectra of 1 and 2, respectively.
Chem. Asian J. 2009, 4, 1542 – 1550
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemasianj.org
1545

T. D. M. Bell, S. J. Langford, K. P. Ghiggino et al.
FULL PAPERS
ꢀ
(S0 S1). For compounds 1 and 2, however, a new, additional,
decay function convolved with an instrument response func-
tion using a least squares fitting routine based on the Mar-
quardt algorithm. Satisfactory fits to the data (assessed by
the reduced c2 goodness-of-fit parameter and a random dis-
tribution of the weighted residuals) were obtained for both
SANDI compounds in all solvents. Fluorescence decay
times followed the same trend for both SANDIs: the longest
decay time for each compound was recorded in acetonitrile,
the shortest in ethanol and decay times are longer for 2 than
for 1 in all solvents (Table 1).
The fluorescence lifetimes of both compounds are rela-
tively long: for 1 ~8<tflu <11 ns and 2 ~9<tflu <12 ns with
some dependence on solvent. These values compare favora-
bly to those of most commonly used dyes, including pery-
lene and terrylene imides. Longer lifetimes are desirable for
a number of reasons, the primary advantages being ease of
detection and greater confidence in the interpretation of
small differences owing to changes in experimental condi-
tions.
lower energy transition is present in the visible region of
their absorption spectra with maxima in toluene at 513 and
601 nm, respectively. Similar transitions have also been ob-
served in other NDI compounds with electron donating side
groups.[8,9] This new transition is polarized in the opposite
direction (along the short axis) to the near-UV transition
which is polarized along the long axis of the molecules and
remains mostly unaffected with respect to unsubstituted
NDI. Calculations by Wꢃrthner and co-workers[8] on related
SANDI compounds also showing a similar new transition in-
dicate that it has some charge transfer character. The
strength of the new transition is assisted by intramolecular
hydrogen bonding between the H atom attached to the
amino N atom and the nearby carbonyl O atom. This hydro-
gen bond then forms part of a 6-membered ring and helps
confer planar geometry on the atoms involved. Maximal
overlap of the lone pair on the N atom with the conjugated
pi-system of the core is realized when the dihedral angle of
ꢀ
the core N bond is planar (08 or 1808).
As a result of their relatively long fluorescence decays
and spectral characteristics, 1 and 2 are well suited for use
as labels in Fçrster resonant energy transfer (FRET) appli-
cations. The critical distance, Ro, for through space energy
transfer is defined as the separation of the donor and ac-
ceptor at which energy transfer occurs with 50% efficiency
and is given by Equation (1),
It can be readily seen in Figure 2 how the addition of the
second alkylamino substituent leads to a red shift in the ab-
sorption and emission maxima, highlighting the spectral tun-
ability of these compounds by means of the number and
placement of the substituents. On the other hand, only small
spectral shifts are observed with changes of solvent. The
emission maximum of 1 red shifts only 8 nm (260 cmꢀ1)
from 551 nm in toluene to 559 nm in acetonitrile and a fur-
ther 11 nm (345 cmꢀ1) to 570 nm in ethanol. The emission
maximum of 2 is at 632 nm in toluene, 634 nm in acetoni-
trile, and 638 nm in ethanol (a total of 150 cmꢀ1 red shift
compared to 605 cmꢀ1 for 1). It is likely that the increased
symmetry of 2 balances the charge redistribution with re-
spect to 1 leading to the weaker solvent dependent red shift
observed. Similar, although more pronounced, solvatochro-
mic behavior was seen for the SANDI compounds of
Wꢃrthner.[8,9]
À
Á1
=
Ro ¼ 0:2108 h4Ff k2J
6
ð1Þ
where, h is the refractive index of the medium, Ff the fluo-
rescence quantum yield of the donor in the absence of the
acceptor, k2 the orientation factor for the dipole–dipole in-
teraction, and J is the normalized spectral overlap integral.
Assuming a value of 2/3 for k2 (for a random orientation of
the donor and acceptor transition dipoles), the critical Fçr-
ster distance for the mono-allyl and di-allyl pair is 41 ꢂ im-
plying that FRET processes should be readily observable
for donor–acceptor separations in the 2–7 nm range.
Compounds 1 and 2 were found to have values for er in
toluene of 11000mꢀ1 cmꢀ1 and 13100mꢀ1 cmꢀ1, respectively.
These values lie at the lower end of the range generally ex-
hibited by fluorescent labels and are similar to those report-
ed for related SANDI compounds.[8,9] Both SANDI deriva-
tives display favorably high fluorescence efficiencies in all
three solvents, with Ff values typically>0.5 (Table 1). Fluo-
rescence decay profiles were also recorded in toluene, etha-
nol, and acetonitrile using the time correlated single photon
counting method and fluorescence decay times (tflu) were
obtained by fitting the decay profiles with an exponential
Single Molecule Photophysical Properties of 1 and 2
Single molecules of both the SANDI compounds are bright
and readily imaged by confocal fluorescence microscopy.
Some example fluorescence trajectories of single molecules
of 1 and 2 embedded in PMMA films and studied under ni-
trogen are shown in Figures 3 and 4, respectively. Single
molecules of both SANDIs emit with high count rates typi-
cally in the range of (5–10)ꢄ103
counts/second under the experi-
Table 1. Selected ensemble photophysical data for 1 and 2 in various solvents.
mental conditions employed.
As comparison, perylene
monoimide under similar condi-
tions displayed count rates
about 2–3 times higher.[4]
SANDI
Solvent
lmax(abs) [nm]
er [Mꢀ1 cmꢀ1
]
lmax(em) [nm]
Ff
tflu [ns]
a
1
1
1
2
2
2
toluene
EtOH
MeCN
toluene
EtOH
MeCN
513
511
511
601
596
595
11000
–
–
13100
–
–
551
570
559
632
638
634
0.51
0.44
0.65
0.62
0.60
0.56
9.5
7.7
11.2
11.3
9.5
Trajectories of
1 typically
11.8
lasted of the order of some sec-
1546
www.chemasianj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2009, 4, 1542 – 1550

Fluorophores for Single Molecule Applications
the application of a fluorophore as a SM label as the mole-
cules will be ꢅoffꢁ for considerable periods. These SM meas-
urements were carried out under a nitrogen atmosphere and
even when the data are binned in short 1 ms intervals, there
are no obvious triplet blinks implying that these compounds
have a very low yield of triplet formation. Fluorescence in-
terruptions on a timescale of 100s of ms to seconds, associat-
ed with reversible quenching processes, are also often ob-
served in many single molecule fluorescence studies and can
be a severe hindrance to emitter performance. Long blinks
such as those in the trajectory of the SM in Figure 4c were
seen in approximately a third of the trajectories recorded
from SMs of 1 and in less than 10% of the trajectories of
SMs of 2. Fluctuations in emission intensity, such as are evi-
dent in the trajectory of 1 shown in Figure 3b, were almost
entirely absent for SMs of 2. Indeed, most SMs of 2 showed
steady, single level emission with no intermittencies before
photobleaching in one step—essentially ideal fluorophore
behavior for SM labeling applications.
Figure 3. Emission intensity trajectories as a function of time for two
single molecules (panels a and b) of 1 in PMMA film under nitrogen. Ex-
citation was at 485 nm.
Example emission spectra from a single molecule of 2 in
PMMA film under nitrogen are shown in Figure 5. Succes-
sive spectra were taken every three seconds thus providing a
Figure 5. Emission spectra from a single molecule of 2 embedded in a
PMMA film under nitrogen. Successive spectra are recorded every three
seconds. Excitation was at 543 nm.
Figure 4. Emission intensity trajectories as a function of time for 3 single
molecules (panels a, b, and c) of 2 in PMMA film under nitrogen. Excita-
tion was at 543 nm.
record of changes in the emission with time. Most molecules
showed very stable emission in terms of spectral characteris-
tics which resemble well the spectra of bulk solutions. Very
little spectral shifting over time was observed for individual
molecules. Most molecules behaved similarly to the one in
Figure 5 showing a constant wavelength of emission maxi-
mum throughout their trajectory. Emission maxima varied
somewhat from molecule to molecule and were distributed
over the range of 580–650 nm, reflecting local heterogenei-
ties in the PMMA film. The small difference in average SM
emission maximum (~615 nm) and the ensemble values (~
630 nm) is attributed to differences between the film and so-
lution environments.
onds before photobleaching occurred. The trajectory shown
in Figure 3a was the longest lived of the data set at 23 s,
while the trajectory in Figure 3b lasting for 8 s is typical of
SMs of 1. Di-allyl SANDI 2, however, shows much greater
photostability with some single molecules emitting for over
a minute such as the SM in Figure 4a. Also apparent in the
photon trajectories is the almost complete lack of interrup-
tions to the fluorescence emission or “blinking”. Visits to
the triplet state are the most prevalent blinking mechanism
and occur on a micro to millisecond time range for most flu-
orophores. Excessive triplet formation can severely restrict
Chem. Asian J. 2009, 4, 1542 – 1550
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemasianj.org
1547

T. D. M. Bell, S. J. Langford, K. P. Ghiggino et al.
FULL PAPERS
The two SANDI molecules synthesized and studied here
show many properties desirable for fluorescence labeling ap-
plications. They have high emission quantum yields (>0.5),
absorption and emission in the visible part of the spectrum,
and nanosecond fluorescence decay times. Most importantly,
they also show excellent photostability. Aerated solutions
(~10ꢀ5 m) of 1 and 2 in toluene were subjected to eight
hours of irradiation from the focused output of a 150 W Xe
arc lamp. Over this time, 1 underwent photobleaching of
only ~6% while 2 showed no observable change in absorp-
tion during irradiation. From the known spectral intensity of
the lamp, this indicates a yield of photobleaching of the
order of 10ꢀ6 for 1 and less than 10ꢀ7 for the di-allyl com-
pound 2.
This very low photobleaching yield is realized at the SM
level for 2 with some molecules undergoing tens of millions
of cycles of excitation and photon emission before photo-
bleaching. For example, approximately 800,000 photons
from the single molecule shown in Figure 4a were detected
by the APD detector before the molecule photobleached.
Taking into account that the total detection efficiency of the
experimental set-up is only about 5% and that half of the
collected photons were directed to the CCD camera, the
actual number of photons that this one single molecule
emitted is of the order of 30 million. The longest lived single
molecule of 2 emitted a steady stream of ~4500 photons per
second for longer than 10 min, implying that more than
100 million cycles of excitation and emission were achieved
before photobleaching occurred.
Although these preliminary SM data sets are small
(<50 molecules) a large variation in emission intensity
(count rate) from molecule to molecule is noted. Not nearly
so much variation in count rate over time (aside from “off
times”) is observed for single molecules. One possible mech-
anism that could lead to extensive variation in emission
(count rate) is rotation of the alkylamino side arms. Calcula-
tions by Wꢃrthner and co-workers[8] of similar SANDI com-
pounds showed that the alkylamino side group can donate
electron density from the lone pair on the nitrogen atom
into the naphthalene core as a result of restricted rotation
through hydrogen bonding. Electron donation into the core
would strengthen the S0–S1 transition resulting in an in-
creased rate of photon absorption and, hence, emission
(count rate).
tron donation to the substituent side-arm orientation, a
range of conformations could lead to a distribution of pho-
tophysical behavior. A more detailed study of the photophy-
sics of these new SANDI compounds as single molecules is
currently in progress and will be reported subsequently.
Conclusions
We have developed a convenient, efficient, and scalable al-
ternate reaction protocol for the synthesis of mono-, di-, and
tetra-brominated core substituted naphthalenetetracarboxyl-
ic dianhydrides. This method will allow easy access to core-
substituted NDIs and hence allow further functionalization
of these molecules which will have an impact on their fluo-
rescent properties for supramolecular and material chemis-
try applications. Studies in these areas, and in particular the
chemistry of 5, are underway.
The photophysical properties of two new aminopropenyl
substituted derivatives of naphthalene diimide have been
characterized in bulk solution in a range of solvents and
have undergone preliminary assessment at the single mole-
cule level. Both mono-allyl and di-allyl SANDI compounds
show many properties desirable for fluorescence labeling ap-
plications such as absorption and emission in the visible part
of the spectrum, quantum yields greater than 0.5, nanosec-
ond fluorescence decay times, and a high degree of photo-
stability. The absorption and emission properties are also
highly tunable with the addition of the second aminopropen-
yl group leading to a considerable red-shifting of the emis-
sion. The critical Fçrster distance for a randomly oriented
mono-allyl SANDI di-allyl SANDI pair is 4.1 nm suggesting
their suitability as FRET labels capable of reporting on sep-
arations in the 2–7 nm range. Single molecules of these com-
pounds in PMMA film under nitrogen show strong emission
with very few interruptions or “blinks” and good photosta-
bility. This is particularly so for the di-substituted compound
with some single molecules emitting tens of millions of pho-
tons before photobleaching.
Experimental Section
Synthetic Procedures
The extent of electron donation will be strongly depen-
dent on the angle between the naphthalene core and the al-
kylamino side arm with electron donation expected to be
maximal when the core and side arm are in a planar confor-
mation as this maximizes orbital overlap between the pi-
electrons of the core and the nitrogen lone pair. The single
bond linking each side arm to the core would be expected
to allow the side arms a certain degree of relatively facile
rotation. In bulk solution, the equilibrium arrangement
would dominate, however, it is quite possible that single
molecules embedded in a polymer film could adopt a range
of conformations—particularly in a film prepared by rapid
spin-casting. Depending on the extent of sensitivity of elec-
General bromination: Naphthalenetetracarboxylic dianhydride (0.1 mm)
in the presence of oleum (20% free SO3, 15 mL) and NaBr is stirred and
heated in a Carrius tube for 16–24 h. Slow addition of propionic acid
(CAUTION: highly exothermic) precipitates the product as a yellow
solid. For 5: 5 g NDA yields 6.7 g, 62%.
1
Characterization data for 3: m.p.:>3508C; H NMR (400 MHz, DMSO):
d=8.74 ppm (s, 2H, ArH); mass spectrum (ESI) m/z: 426 [M]ꢀ. For 4:
m.p.:>3508C; 1H NMR (400 MHz, DMSO): d=8.73 (d, J=4.1 Hz, 2H,
ArH), 8.68 (s, 1H, ArH), 8.67 ppm (d, J=4.1 Hz, 2H, ArH): mass spec-
trum (ESI) m/z: 347 [M+H]+. For 5: m.p.:>3508C; IR (nujol): n˜ =1787,
1731, 1192, 1174, 1149, 1093, 986, 936 cmꢀ1; UV/Vis (DMF): lmax =269,
363(sh), 377 nm; mass spectrum (ESI) m/z: 583.55 [M]ꢀ.
General imidation: Compound 4 (100 mg) and allyl amine (4.4 equiv)
were stirred in DMF (20 mL) at 1108C overnight. The reaction mixture
was allowed to cool down to room temperature and water (30 mL) was
1548
www.chemasianj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2009, 4, 1542 – 1550

Fluorophores for Single Molecule Applications
added. The resulting highly colored precipitate was filtered and collected
at the pump, dried in a dessicator, and purified using column chromatog-
raphy (20% hexane in DCM). Yield of 1=70%; m.p.: 195–1968C;
1H NMR (300 MHz, CDCl3, see Figure S1 in the Supporting Informa-
tion): d=10.23 (brt, 1H, NH), 8.66 (d, J=7.8 Hz, 1H, ArH), 8.36 (d, J=
7.8, 1H, ArH), 8.1 (s, 1H, ArH), 5.9–6.1 (m, 3H, allyl H), 5.1–5.4 (m,
6H, allyl H), 4.83 (dt, J=5.8, 1.4 Hz, 2H, CH2), 4.77 (dt, J=5.8, 1.4 Hz,
2H, CH2), 4.23 ppm (m, 2H, CH2); 13C NMR (100 MHz, CDCl3, see Fig-
ure S2 in the Supporting Information): d=166.0, 163.1, 162.8, 162.7,
152.5, 132.7, 132.2, 131.7, 131.6, 131.2, 129.6, 127.9, 126.2, 124.9, 123.7,
120.2, 119.6, 118.5, 118.0, 117.7, 45.5, 43.0, 42.5 ppm; MS (MALDI-TOF):
402.1 (100%), (M+H)+; UV/Vis (tol): lmax (e)=350, 368, 515 nm
(11000); Fluorescence (tol): lmax =551 nm. Yield of 2 is 70%; m.p.: 223–
2248C; 1H NMR (300 MHz, CDCl3, see Figure S3 in the Supporting In-
formation): d=9.5 (brt, 1H, NH), 8.16 (s, 2H, ArH), 5.9–6.1 (m, 4H,
allyl H), 5.1–5.5 (m, 8H, allyl H),4.81 (dt, J=5.5, 1.3 Hz, 4H, CH2), 4.1–
4.2 ppm (m, 4H, CH2); 13C NMR (100, CDCl3, see Figure S4 in the Sup-
porting Information): d=165.8, 149.1, 133.4, 132.1, 125.7, 121.2, 118.6,
117.8, 117.6, 117.3, 102.0, 45.4, 42.4, 29.7 ppm; MS (MALDI-TOF): 456.1
(100%), [M]+; UV/Vis (tol): lmax (e)=346, 362, 602 nm (13100); Fluores-
cence (tol): lmax =629 nm.
Acknowledgements
This project is supported by Australian Research Council Discovery Proj-
ect DP0664163. TDMB gratefully acknowledges The University of Mel-
bourne Faculty of Science for the award of a Centenary Research Fellow-
ship. We would like to thank the CRC Smartprint for financial assistance
in the form of a Scholarship to CHJ The authors also thank Dr. Scott
Watkins for the depth profile measurements on polymer film samples,
Dr. Michel Sliwa for technical assistance with the single molecule meas-
urements and we acknowledge the contributions of Matthew Belousoff
and Dr. Craig Forsyth in their help in obtaining crystallographic data.
[1] P. Tinnefeld, M. Sauer, Angew. Chem. 2005, 117, 2698–2728;
Angew. Chem. Int. Ed. 2005, 44, 2642–2671.
[2] W. E. Moerner, J. Phys. Chem. B 2002, 106, 910–927.
[3] F. Kulzer, M. Orrit, Annu. Rev. Phys. Chem. 2004, 55, 585–611.
[4] A. C. Grimsdale, T. Vosch, M. Lor, M. Cotlet, S. Habuchi, J. Hofk-
ens, F. C. De Schryver, K. Mꢃllen, J. Lumin. 2005, 111, 239–253.
[5] F. C. De Schryver, T. Vosch, M. Cotlet, M. Van der Auweraer, K.
Mꢃllen, J. Hofkens, Acc. Chem. Res. 2005, 38, 514–522.
[6] G. Andric, J. F. Boas, A. M. Bond, G. D. Fallon, K. P. Ghiggino, C. F.
Hogan, J. A. Hutchison, M. A. P. Lee, S. J. Langford, J. R. Pilbrow,
G. J. Troup, C. P. Woodward, Aust. J. Chem. 2004, 57, 1011–1019.
[7] M. Cotlet, T. Vosch, S. Habuchi, T. Weil, K. Mꢃllen, J. Hofkens, F.
De Schryver, J. Am. Chem. Soc. 2005, 127, 9760–9768.
[8] F. Wꢃrthner, S. Ahmed, C. Thalacker and T. Debaerdemaeker,
Chem. Eur. J. 2002, 8, 4742–4750.
Single Crystal X-ray Diffraction Experiment
Single crystals of 5 were grown by vapor diffusion of water into a DMSO
solution of 5. X-ray diffraction data was collected on a Nonius Kappa
CCD diffractometer (graphite-monochromated MoKa radiation, l=
0.71073 ꢂ). Structures were solved by direct methods (SHELXS-97) and
refined by full-matrix least-squares calculations on F2 with the SHELX-
TL program package. Non-hydrogen atoms were refined anisotropically.
CCDC 649279. contains the supplementary crystallographic data for this
paper. These data can be obtained free of charge from The Cambridge
Crystallographic Data Centre at www.ccdc.cam.ac.uk/data_request/cif
[9] C. Thalacker, C. Rçger, F. Wꢃrthner, J. Org. Chem. 2006, 71, 8098–
8105.
[10] C. Thalacker, A. Miura, S. De Feyter, F. C. De Schryver, F. Wꢃrth-
ner, Org. Biomol. Chem. 2005, 3, 414–422.
[11] C. Rçger, S. Ahmed, F. Wꢃrthner, Synthesis 2007, 1872–1876.
[12] C. Rçger, S. Ahmed, F. Wꢃrthner, J. Org. Chem. 2007, 72, 8070–
8075.
[13] S. V. Bhosale, C. H. Jani, S. J. Langford, Chem. Soc. Rev. 2008, 37,
331–342.
[14] H. Vollmann, H. Becker, M. Corell, H. Streeck, Liebigs Ann. Chem.
1937, 531, 1–159.
[15] S. Bhosale, A. L. Sisson, P. Talukdar, A. Fꢃrstenberg, N. Banerji, E.
Vauthey, G. Bollot, J. Mareda, C. Rçger, F. Wꢃrthner, N. Sakai, S.
Matile, Science 2006, 313, 84–86.
[16] C. Rçger, M. G. Muller, M. Lysetska, Y. Miloslavina, A. R. Holz-
warth, F. Wꢃrthner, J. Am. Chem. Soc. 2006, 128, 6542.
[17] B. A. Jones, A. Facchetti, T. J. Marks, M. R. Wasielewski, Chem.
Mater. 2007, 19, 2703–2705.
[18] P. Jacquignon, N. P. Buu-Hoi, M. Mangane, Bull. Soc. Chim. Fr.
1964, 10, 2517–2523.
Photophysical Measurements
Absorption spectra were recorded using a Varian Cary 50 spectropho-
tometer and fluorescence spectra were measured using a Varian Cary
Eclipse fluorescence spectrophotometer. Fluorescence spectra were cor-
rected for differences in detection efficiency with emission wavelength.
All ensemble samples were prepared using spectrophotometric grade sol-
vents used as received from Aldrich and degassed using successive
freeze-pump-thaw cycles. Fluorescence decay profiles were recorded
using the time-correlated single photon counting technique as outlined
elsewhere.[26] The excitation source was an optical parametric oscillator
(OPO, Coherent) which provided ~100 fs FWHM pulses at 100 kHz. The
OPO was pumped by a regeneratively amplified mode-locked Ti:sap-
phire laser (Coherent Mira). Compounds 1 and 2 were excited at 510 nm
and 590 nm, respectively. Emission from the sample was passed through
a polarizer oriented at 54.78 relative to the vertically polarized excitation
light thereby removing any anisotropic effects from the decay. Sample
emission was passed through a single grating monochromator (Jobin
Yvon H-10), and detected with a microchannel plate photomultiplier
tube (Eldy, EM1–132).
[19] F. Chaignon, M. Falkenstrçm, S. Karlsson, E. Blart, F. Odobel, L.
Hammarstrom, Chem. Commun. 2007, 64–66.
[20] Two regioisomers are likely from the dibromination, those placed
antipodally that is, 2,6-positions and the 2,7-dibromo NDA regioiso-
mer. In our work, as with others, we only see evidence for the 2,6-di-
bromo isomer by NMR spectroscopy. If the other isomer forms in
>5% yield, it is indistinguishable by both chromatographic and
NMR spectroscopic techniques. The result can be explained in
terms of buttressing restraints, but we cannot discount the formation
of the other regioisomer conclusively.
Single Molecule Studies
Samples for single molecule analysis were prepared by spin-casting (60 s
at 2000 rpm) dilute solutions (~10ꢀ10 m) of 1 or 2 in toluene containing
~10 mgmLꢀ1 of poly(methyl methacrylate) onto thoroughly cleaned glass
coverslips. Film thicknesses were checked by depth profiling (Veeco
Dektak) and were found to be ~80–100 nm. Single molecules were excit-
ed by 543 nm light at 4 MHz, 1–2 ps FWHM, provided by an OPO (Spec-
tra Physics) pumped by a regeneratively amplified Ti:sapphire laser
(Spectra Physics). The light was directed into an inverted microscope
(Olympus IX70) and focused at the sample through a 1.4 N.A. oil-immer-
sion objective (Zeiss 100x). Fluorescence was collected through the same
objective, split 50:50 and directed onto an avalanche photodiode
(Perkin–Elmer) and through a spectrograph (Acton 150i) onto a liquid
N2 cooled CCD camera (Princeton). Details of the SM detection arrange-
ment are published elsewhere.[27]
[21] The low solubility of this compound and lack of protons within its
structure (C14O6Br4) makes it difficult to characterize the product by
traditional means. However, mass spectrometry gave data consistent
with 5, in particular the diagnostic isotopic pattern expected for ad-
dition of four bromine atoms to the starting anhydride.
[22] Crystal data for 5: M.p.:>3508C; C14Br4O6: M=579.64, Colorless
rectangular prisms, 0.21ꢄ0.14ꢄ0.13 mm3, triclinic, space group P-1,
a=6.4747(13), b=7.5706(15), c=12.434(3) ꢂ, a=91.28(3), b=
103.27(3),
g=114.24(3)8,
V=536.22(26) ꢂ3,
Z=1,
1cald =
2.292 gcmꢀ3, F000 =356, MoKa radiation l=0.71073 ꢂ, T=173(2) K,
2qmax =56.18, 9652 reflections collected, 2554 unique (Rint =0.0501).
Chem. Asian J. 2009, 4, 1542 – 1550
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemasianj.org
1549

T. D. M. Bell, S. J. Langford, K. P. Ghiggino et al.
FULL PAPERS
Final GoF=1.031, R1=0.0276, wR2=0.0602, R indices based on
2240 reflections with I>2s(I) (refinement on F2), 147 parameters, 0
[26] T. D. M. Bell, K. P. Ghiggino, K. A. Jolliffe, M. G. Ranasinghe, S. J.
Langford, M. J. Shephard and M. N. Paddon-Row, J. Phys. Chem. A
2002, 106, 10079–10088.
restraints. Lp and absorption corrections applied, m=7.747 mmꢀ1
.
[23] G. D. Fallon, M. A. P. Lee, S. J. Langford, Acta Crystallogr. E 2003,
59, o328–o329.
[24] G. D. Fallon, M. A. P. Lee, S. J. Langford. P. J. Nichols, Org. Lett.
2004, 6, 655–658.
[27] T. Vosch, M. Cotlet, J. Hofkens, K. Van Biest, M. Lor, K. Weston, P.
Tinnefeld, M. Sauer, L. Latterini, K. Mꢃllen, F. C. De Schryver, J.
Phys. Chem. A 2003, 107, 6920–6931.
Received: June 22, 2009
[25] The average error in these values is 0.04 ꢂ.
Published online: September 16, 2009
1550
www.chemasianj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2009, 4, 1542 – 1550