Insight into the Excitation-Dependent Fluorescence of Carbon Dots

Article abstract of DOI:10.1002/cphc.201801061

High quantum yield, photoluminescence tunability, and sensitivity to the environment are a few distinct trademarks that make carbon nanodots (CDs) interesting for fundamental research, with potential to replace the prevalent inorganic semiconductor quantum dots. Currently, application and fundamental understanding of CDs are constrained because it is difficult to make a quantitative comparison among different types of CDs simply because their photoluminescence properties are directly linked to their size distribution, the surface functionalization, the carbon core structures (graphitic or amorphous) and the number of defects. Herein, we report a facile one-step synthesis of mono-dispersed and highly fluorescent nanometre size CDs from a ‘family’ of glucose-based sugars. These CDs are stable in aqueous solutions with photoluminescence in the visible range. Our results show several common features in the family of CDs synthesized in that the fluorescence, in the visible region, is due to a weak absorption in the 300–400 nm from a heterogeneous population of fluorophores. Fluorescence quenching experiments suggest the existence of not only surface-exposed fluorophores but more importantly solvent inaccessible fluorophores present within the core of CDs. Interestingly, time-resolved fluorescence anisotropy experiments directly suggest that a fast exchange of excitation energy occurs that results in a homo-FRET based depolarization within 150 ps of excitation.

Full text of DOI:10.1002/cphc.201801061

Accepted Article
Title: Insight into Excitation Dependent Fluorescence of Carbon Dots
Authors: Divya Sasi, Satya Narayan, Sri Rama Koti Ainavarapu, and
Deepa Khushalani
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: ChemPhysChem 10.1002/cphc.201801061
Link to VoR: http://dx.doi.org/10.1002/cphc.201801061
A Journal of
www.chemphyschem.org

10.1002/cphc.201801061
ChemPhysChem
ARTICLE
Insight into Excitation Dependent Fluorescence of Carbon Dots
S.Divya, Satya Narayan, Sri Rama Koti Ainavarapu* and Deepa Khushalani*
Abstract: High quantum yield, photoluminescence tunability, and
sensitivity to the environment are the few distinct trademarks that
make carbon nanodots (CDs) interesting for fundamental research
with potential to replace the prevalent inorganic semiconductor
quantum dots. Currently, application and fundamental understanding
of CDs are constrained because it is difficult to make a quantitative
comparison among different types of CDs simply because their PL
properties are directly linked to their size distribution, the surface
functionalization, the carbon core structures (graphitic or amorphous)
and the number of defects. Herein, we report a facile one-step
synthesis of mono-dispersed and highly fluorescent nanometre size
CDs from a ‘family’ of glucose-based sugars. These CDs are stable
in aqueous solutions with photoluminescence in the visible range. Our
results show several common features in the family of CDs
synthesized in that the fluorescence, in the visible region, is due to a
weak absorption in the 300-400 nm from a heterogeneous population
of fluorophores. Fluorescence quenching experiments suggest the
existence of not only surface-exposed fluorophores but more
importantly solvent inaccessible fluorophores present within the core
of CDs. Interestingly, time-resolved fluorescence anisotropy
experiments directly suggest that a fast exchange of excitation energy
occurs that results in a homo-FRET based depolarization within 150ps
of excitation.
understood and in fact to date there still is a lack of consensus as
to the origin of this emission. This aspect has been addressed by
various groups by analyzing, for example, the role of surface
groups on emission, size variation, co-excitation of the CDs in
different emissive states, evaluating the charge transfer dynamics
in presence of surrounding molecules/ions, or assessing the pH
dependence of fluorescence of these CDs. However, despite this
large effort, it is apparent that a cohesive and detailed conceptual
understanding of the exact origin of fluorescence and more
importantly its correlation with the CD chemical structure is still
lacking.[4-11] Majority
of
the
studies
indicate
that
photoluminescence (PL) is mainly a surface phenomenon and the
properties are due to a variety of defective functional groups
present solely on the surface. The carbon core is being
conceptualized as either an amorphous or a graphitic structure
and in principle devoid of any PL properties.[10]
One of the major difficulties in collating data from different
published reports has been the large heterogeneity in the
synthesis of the CDs. The precursors used are highly diverse in
their structures ranging from well-defined molecules such as
ascorbic acid to complex configurations such as grass or even
hair fiber.[12] These have then been exposed to a myriad of
conditions, involving large pH ranges and or additives/solvents
(that may or may not be inert in terms of affecting the PL of the
ensuing structures). Moreover, despite considerable research
efforts that have been expended to produce CDs with controlled
dimensions and surface properties (using a variety of techniques
including chemical and thermal oxidation, laser ablation,
microwave processing, template methods and deoxygenation of
natural carbon sources) there is still a lack of a unifying theory on
PL.[13] It is, in fact, difficult to make a quantitative comparison
among different types of CDs simply because their PL properties
are directly linked to their size distribution, the surface
functionalization, the carbon core structures (graphitic or
amorphous) and the number of defects. These in turn are highly
sensitive to the precursors and the synthesis conditions used.
Hence it is has become apparent that in order to elucidate the PL
aspects in a detailed manner, it is imperative that the composition
and structure of the CDs have to be carefully controlled and this
is crucial for understanding their complex luminescence
mechanism.
Introduction
Since their inception, carbon dots (CDs) have emerged as an
intriguing class of photoactive nanomaterials attracting
considerable attention. Despite the absence of classically defined
quantum confinement, their fluorescence properties are
impressive as they are known to display excitation dependent
emission irrespective of their sizes. As such, they are being
promoted as the next generation of luminescent materials simply
because of their properties such as high luminescence, chemical
solubility, facile surface modification, biocompatibility, high
resistance to photobleaching and low cost. Many of the
aforementioned aspects are in fact superior to conventional
fluorescent organic dyes and luminescent inorganic quantum dots,
and hence CDs are being actively pursued in a wide variety of
applications. [1-3]
To address these issues, presented here is a systematic study
whereby a carefully controlled range of CDs have been
synthesized in a facile manner without interference of additives.
The synthesis is highly versatile employing a range of six
individual precursors which belong to the same ‘family’ of sugars.
The only alteration has been their molecular weight and
number/type of functional groups. The structures are shown in
figure 1 below and it can be seen that they range from a
monomeric unit to a well-defined polymer consisting of specifically
a seven-membered ring of glucose (β-cyclodextrin). Importantly,
Despite this immense interest, it is intriguing however to note that
the exact cause and nature of their excitation-dependent
fluorescence, which actually violates the Kasha-Vavilov’s rule of
excitation independent emission luminescence, is still not fully
Department of Chemical Sciences, Tata Institute of Fundamental
Research, Homi Bhabha Road, Colaba, Mumbai 400 005, India.
khushalani@tifr.res.in; koti@tifr.res.in
Supporting information for this article is given via a link at the end of the
document.
This article is protected by copyright. All rights reserved.

10.1002/cphc.201801061
ChemPhysChem
ARTICLE
Cyano-4-hydroxycinnamic acid) are in fact assisting the
desorption of CDs during acquisition. As such the MALDI data
represents in principle not a fragmentation pattern but instead it is
a representation of particle weight heterogeneity. The pattern in
the mass spectrum of Glucose CDs displayed a m/z range of 800-
1200 which was analogous to the ranges for the other CDs that
were synthesized. Similar to TEM, results from MALDI also
exhibited an intriguing trend among all the products in that
similarly spaced mass differences were observed irrespective of
the precursor used, indicating that the entire set of CDs
synthesized had a large similarity in their final molecular structure.
Dramatic differences were not observed and in fact the highest
m/z obtained from all the different CDs are compiled in table 1 and
these were found to be comparable and centered at a value of ca.
1150. Moreover, the pattern in all the six CDs synthesized showed
common mass differences of peaks whose values were at 14, 17,
28, 32, 44 m/z corresponding to the mass difference of a methyl
(-CH3) or a hydroxyl (-OH) group, aldehyde (-CHO), methoxy (-
OCH3) and carboxylic group (-COOH) respectively. The only
major difference observed was that in the data of CDs
synthesized from β-Cyclodextrin-SH, a peak observed at m/z
value of 47 which could be ascribed to methanethiolate (-CH3S)
whereas in β-Cyclodextrin-NH2, the MALDI pattern had the most
number of new peaks and specifically one that was observed at
∆m/z value of 17 which could be ascribed due to the loss of amino
(-NH2) group.
Figure 1: Structure of the different CD precursors.
specifically, only one –OH moiety of β-cyclodextrin has been
altered to either a -SH or -NH2 such that direct insight can be
gained into effect of the functionality (-OH vs -SH vs -NH2) on the
optical properties. A trivial synthetic protocol has been employed
to form CDs that uses only water as the solvent under microwave
pyrolysis without the addition of any other acid, base,
reducing/oxidizing agent or even capping agent. Subsequently,
spectroscopic studies have been performed to evaluate the
photo-physics of the synthesized CDs and inferences have been
obtained in regard to how the structure of the precursor dictates
the final composition of the CDs and hence their optical
properties. Based on the experimental results we assume that the
PL origin mainly arise not only from the chemical groups on the
surface electronically coupled to the carbon core but also due to
coupling between the intrinsic defects in the core and the surface
states forming molecular fluorophores.
Results and Discussion
The as-synthesized colloidal suspensions of CDs obtained were
characterized within at least 24 hrs. Irrespective of this aspect,
however it should be noted that the CDs were stable for up-to two
months as the data obtained within this period was always
reproducible. Photoluminescence spectra were recorded before
and after dialysis with no visible differences. Hence our protocol
proves to be facile and purification free method for synthesizing
carbon dots. Figure 2a displays a TEM image of Glucose CDs. It
is observed that the particles are spherical in nature with a narrow
size distribution centered at ca. 4 nm and table 1 shows the
average size determined for all the precursors used (data collated
from figure S3). It can be surmised that shape, size and size
distribution are correspondingly similar irrespective of the
precursors used for CD synthesis. As the CDs had
monodispersed nanometer size distribution, in order to get insight
into the chemical structure and molecular weight of the CDs,
MALDI analysis was performed, figure 2b (additional data in figure
S4). For these MALDI measurements, the incident laser
wavelength used was 337 nm, therefore it is expected that the
CDs have minimal direct absorption and in fact the matrix (α-
Figure 2: (a) TEM image and particle size distribution histogram of the carbon
dots from Glucose. (b) MALDI-TOF MS of CDs synthesised from Glucose.
For further structural insight, X-ray diffraction studies were
performed on the lyophilized samples figure S5. The
diffractogram displays a single peak at ca. 3.8 Å (similar to the
reports published by other groups)[14,15] and as it is larger than the
interlayer spacing for the conventional (002) planes of bulk
graphite (3.34 Å), it can therefore be deduced that the CD
framework consists of a turbostratic carbon structure having a
disordered graphitic plane alignment. In the FTIR spectra, figure
S6, there were practically negligible differences observed
between the six CDs synthesized and all the spectra consistently
displayed peaks at ca. 3300, 2900 and 1600 cm−1 which could be
assigned to the O-H, C-H, C=C symmetric stretch of the carbon
skeleton.
In order to evaluate the optical properties of the CDs, UV-Vis
absorption spectra were acquired, figure 3. Again, it can be
observed that the CDs prepared from different carbon sources
This article is protected by copyright. All rights reserved.

10.1002/cphc.201801061
ChemPhysChem
ARTICLE
Mol. Mass
of Precursor (g/mol)
Highest m/z
(g/mol)
Average size
(nm)
PL Emission Range
(nm)
Sample
180
342
360
1142
1184
1168
Glucose CD
Sucrose CD
Maltose CD
4.3 ± 0.8
4.2 ± 1.2
4.2 ± 1.1
430-555
433-553
439-553
1135
1151
1134
1191
1168
1196
β-Cyclodextrin CD
β-Cyclodextrin-SH CD
β-Cyclodextrin-NH2 CD
4.0 ± 0.9
4.4 ± 0.9
3.8 ± 0.9
432-556
432-549
435-541
Table 1: Size and fluorescence properties of CDs
gave similar absorption features. It needs to be clearly stated that
the as-synthesized product, after microwave synthesis, was
diluted 5 times (consistently for each of the six CDs) and then the
absorption spectrum was acquired. A strong band at 220 nm was
observed and it is known to originate from the →* transition of
the aromatic carbons that are normally hypothesized to be
present in the core.[16] The broad intense peak at 280 nm can be
assigned to the typical absorption of an n→* transition which is
again due to the structural motif and has been observed in CDs
prepared from a variety of other precursors as well.[17] Other than
these two strong features, the rest of the spectrum can be
considered to consist of a long absorptive tail extending up to
800nm (see inset). The only difference between the data of the
six CDs was not in the position of the absorption features but in
fact the intensity of these bands. The β-Cyclodextrin-NH2 based
CDs gave the highest intensity of the bands while glucose-based
CDs gave the least intense spectrum.
Figure 4: (a & b) Excitation dependent photoluminescence spectra of
Glucose CDs. (c & d) Emission spectra of CDs excited at 360nm.
observed at 440 nm when the sample was excited with 360 nm,
(figure 4 b, c & d). This behavior was observed for all the other 5
CDs (additional data in figure S7). Table 1 gives the tunable PL
range of CDs from different precursors and the fluorescence peak
position was found to vary in the range from 430-555 nm when
the excitation wavelength was varied from 300-500 nm. Generally,
such variable emission has been attributed to “defect states” with
a broad energy distribution on either graphitized core of CDs or
the oxygenated groups that are commonly present (such as
carbonyl, carboxyl acid, ether etc.).[18] It was observed for our
samples, that as the number of monomer units were increased in
the precursor, initially the emission intensity of the corresponding
CDs decreased i.e., from glucose to sucrose to maltose but the
emission was augmented dramatically when the monomer unit
increased to seven. Polymeric nature of β-cyclodextrin therefore
had a beneficial effect on the PL intensity. Correspondingly, when
the functional groups were varied, it was observed that replacing
a hydroxyl group with either a thiol (-SH) or amino(-NH2) group,
the emission was significantly suppressed.
Figure 3: Absorbance spectra of the different carbon dots. Inset figure
shows the magnified image of the absorbance from 360 nm to 800.
Further insight was obtained using photoluminescence
measurements. Figure 4a shows the normalized steady state PL
data for glucose CDs while being irradiated with a variety of
excitation wavelengths ranging from 300 to 500 nm. Tuneable
emission was observed with the most intense emission being
This article is protected by copyright. All rights reserved.

10.1002/cphc.201801061
ChemPhysChem
ARTICLE
It needs to be highlighted that all the CDs exhibited an absence
of fluorescence emission when excited at the maximum of the UV
absorption band at 280nm. Such behavior is quite contrary to that
of semiconductor quantum dots but has been
of these surface fluorophores, the buried species (which are not
being as effectively quenched by acrylamide) are populating the
emission spectrum and are now being more clearly observed.
Thus, according to the above results CDs may be considered to
have at least two populations of fluorophore species
characterized by different fluorescence maxima and a differing
accessibility to acrylamide.
To test, whether the fluorescence quenching is static or dynamic,
we performed time-resolved fluorescence experiments.
Normalized time resolved fluorescence curves and corresponding
Stern-Volmer plots of CD-water dispersions are shown in figure
6a (additional data in figure S11, excitation wavelength used was
366 nm and emission was at 440 nm). All the fluorescence decay
curves were fitted to multi-exponentials and mean fluorescence
commonly observed for CDs and also for graphene and graphene
oxide nanoparticles. [19-21] It can therefore be postulated that there
is a prevalence of two types of chromophores, non-emitting and
those emitting in the visible range. The absorption coefficient of
the carbon core is higher than that of the fluorophore species so
that the relative absorption of the core is visible while that of the
fluorophore appears to be suppressed. Figure S8 represents the
excitation spectra of the CDs at emission wavelength of 440 nm.
It is clearly visible that the excitation spectra have negligible
resemblance to the UV-Vis absorption spectra. Hence, it can be
surmised that the majority of the structure of the CDs consist of a
turbostratic graphitic/amorphous carbon structure within which
are embedded fluorophores. This large skeletal framework is not-
fluorescent while there exists within it a heterogeneous population
of fluorophores with excitation dependent emission spectra
depending on their chemical structures.
To gain further insight into location and accessibility of these
fluorophores, fluorescence quenching experiments were
performed using two well-known quenchers: acrylamide and
potassium iodide (KI). The Stern–Volmer plots in figure 5
demonstrate that both quenchers reduce the PL intensity in a
concentration dependent manner and as the plots are
predominantly linear it suggests that the quenching mechanism is
either static or dynamic, but not both. Acrylamide is normally
expected to quench both the ‘‘exposed’’ and ‘‘buried’’
fluorophores (as has been shown e.g. for buried tryptophan
residues in proteins22), while iodide being charged, very large in
size, and highly hydrated, is limited normally to quenching surface
fluorophores. But here, KI seems to quench the PL to a larger
extent (2-5 times) than acrylamide (1.5 - 2.5 times) for all the CDs
except for β-cyclodextrin CD for which acrylamide appears to be
a better quencher. The quenching constants determined are
detailed in table 2 (individual PL spectra with quenchers KI and
acrylamide are shown in figures S9 and S10).
lifetimes (m) were measured (for details see experimental section
in Supporting Information). Mean lifetime and bimolecular
quenching constants obtained are given in the table 2. It can be
confirmed from this data that quenching is dynamic in nature. For
all the six types of CDs, after the addition of KI, the lifetime of PL
was observed to decrease as observed in table S1-S6. The
fluorescence decays of CDs required multi-exponential fits
suggesting heterogeneity of fluorescent species. It was observed
that as we sequentially evaluate glucose CDs to β-cyclodextrin
based CDs, the lifetime of the fastest component increased from
0.24 ns (50%) to 0.33 ns (39%) whereas the slowest component
decreased from10.4 ns (1%) to 4.6 ns (6%). Upon variation of the
functional groups, it could be observed that the fastest component
had the shortest lifetime of 0.05 ns (62%) for β-cyclodextrin-NH2
whereas the lifetime of the slowest component remained more or
less the same with an overall decrease in the average lifetime.
Furthermore, the estimated bimolecular quenching constant
(Kq=KSV
/
) from the Stern-Volmer constant (KSV) and the mean
lifetime in the absence of quencher (
), were found to be large
(0.5-2.5 x 1010) suggesting that the quenching by KI was diffusion-
controlled and very efficient.[23] Moreover, the linearity in the
Stern-Volmer quenching plots without any saturation (or
downward curvature) suggested that all the fluorophores, both the
surface and in the core, were efficiently quenched.
The aforementioned observations raise an important question
that how a solvated-iodide (from KI), which is larger than
acrylamide, is able to quench buried fluorophores? How the
buried fluorophores are quenched without physically colliding with
iodide ions? This would only be possible if the excitation energy
from the buried fluorophores is transferred to those at the surface
which are then being quenched efficiently by the iodide. We
further investigated this aspect using time-resolved fluorescence
anisotropy experiments as detailed below.
Anisotropy studies based on time resolved fluorescence
polarization is a powerful tool which provides valuable information
on the size and shape of fluorescing particles, and characteristics
of surrounding environment. This technique relies on the fact that
linearly polarized light preferentially excites those molecules or
particles whose transition dipole moment is parallel to the incident
light field. Because of the rotational diffusion of the excited
particles, their average emission dipole moment acquires a non-
vanishing component perpendicular to the exciting field and the
Figure 5: Steady state quenching of CDs with KI and acrylamide as
quenchers.
A uniform decrease was observed in the PL intensity for the entire
emission range for KI without any shift in the emission maxima. In
contrast, the emission maxima were found to shift to longer
wavelengths at elevated concentrations of acrylamide. The
observed spectral red shift of ca. 5-20 nm in case of acrylamide
quenching presumably results from the selective quenching of
fluorophores more exposed to the solvent. Upon the suppression
This article is protected by copyright. All rights reserved.

10.1002/cphc.201801061
ChemPhysChem
ARTICLE
Time-resolved studies
Steady-state
KSV (M-1)
Sample
0
KSV
Kq
KI
13
7
Acrylamide
(M-1)
(M-1s-1)
(ns)
1.1
21.8 × 109
10.8 × 109
11.8 × 109
5.3 × 109
24
7
Glucose CD
Sucrose CD
Maltose CD
7
4
0.65
1.02
1.33
12
7
4
5
β-Cyclodextrin CD
β-Cyclodextrin-SH CD
β-Cyclodextrin-NH2 CD
13
22
21.7 × 109
25 × 109
18
13
0.83
0.52
8
6
2
11
Table 2: Steady state and time-resolved fluorescence quenching properties of CDs.
same occurs to the polarization direction of fluorescence light.
Fluorescence anisotropy decay curves of CDs system are shown
in the figure 6b (additional data in figure S12). Initial anisotropy
was observed to decrease with time due to the loss of initial
uniformly even though some could be at the core and not
physically accessible to the KI quencher. Moreover, the red-shift
in the fluorescence maximum of CDs upon acrylamide quenching
can now also be elucidated. As acrylamide is relatively less
efficient compared to KI (table 2) by a factor of 2, when the
excitation energy is transferred to the surface fluorophores it is
red-shifted and is shown as emission because acrylamide is less
efficient in quenching this, unlike KI which more effectively
reduces all the emission. As such there is more uniform
quenching throughout the emission spectra.
Conclusions
In summary, from the above data the following aspects can be
surmised:
Figure 6: (a) Time-resolved fluorescence studies on quenching of CDs
with KI. (b) Time-resolved fluorescence anisotropy Glucose-CD system.
•
PL increases when the precursor is changed from glucose
orientation as a result of rotational diffusion or of energy transfer
to other emitters. The distinction between these two depolarizing
factors is crucial in the analysis of anisotropy decays of CDs. Our
results demonstrate that in all studied systems there are fast
anisotropy decays approximately equal to zero level with rotation
correlational time constant ≤ ca. 150 ps in all the cases. In
comparison, the rotations of globular protein molecules of similar
size (4.5 nm) are observed in the range 5-10 ns.[24] In our case,
instead of observing high anisotropy as expected from the rotation
of CDs supposed to be in the 5-10ns range, the polarization
decayed faster within 150ps. The observed very fast
depolarization which is devoid of any molecular reorientation
component suggests that the excitation energy of a fluorophore is
being non-radiatively transferred to neighboring fluorophores that
are oriented differently. So, it can be hypothesized that there exist
donors and acceptors on the CD surface which are in different
orientations and in the case of multiple transfers the information
of initial polarization is lost through a homo-FRET mechanism. [21]
It is possible that the fluorophores which are at the core of the CD
transfer their excitation energy to their neighbours very quickly
and this “appears” on the surface fluorophores during the energy
transfer. As shown by the anisotropy studies, this is very fast and
it also explains how KI is able to quench all the fluorophores
to β-cyclodextrin. This suggests that polymeric materials are
better precursors to form luminescent structures.
•
In terms of functionality, the –OH moiety leads to more
fluorescent CDs then either –NH2 or –SH.
For the final CDs that form, there are many “types” of
•
fluorophores, some are buried and some are surface bound. Both
types emit, albeit at different wavelengths and with different
intensity.
•
As such, it can be stated the CDs structures actually consist
of a framework that has C=C and C=O moieties that are strongly
absorbing but not emitting. Buried within this framework (in the
bulk region as well as the surface) are a variety of chromophore
species that absorb from 300 to 600 nm and each emits ca.
between 430 to 560 nm.
•
Our anisotropy results indicate that the fluorophores are
interacting with each other through energy transfer and hence it
can be deduced that depending on the raw precursor used there
are a variety of fluorophores present on and beneath the surface.
This article is protected by copyright. All rights reserved.

10.1002/cphc.201801061
ChemPhysChem
ARTICLE
ESI probe (positive and negative ion modes). 1H NMR was collected in
DMSO at 25°C on Varian 600 MHz. The UV-Vis absorption spectra were
recorded using double beam PerkinElmer Lambda 750-UV/Vis/NIR
spectrophotometer. FT-IR were recorded on a Nicolet-200-135 while
Powder XRD measurements were recorded using a PAN analytical
X’pertpro diffractometer with monochromatic Cu Kα source (λꢀ=ꢀ1.54056ꢀÅ).
All the steady state photoluminescent (PL) measurements were recorded
using a SPEX Horiba Fluorolog fluorimeter.
Experimental Section
Materials: All precursors were procured with AR grade: D-glucose
(SDFCL), Sucrose (SDFCL), Maltose Monohydrate (Loba Chemie), β-
Cyclodextrin (ACROS Organics). All reagents of analytical reagent grade
or above were used as received without any further purification.
Synthesis of mono-6-thio-β-Cyclodextrin
Time-resolved Fluorescence: Fluorescence lifetime measurements and
anisotropy decay kinetics were recorded by employing CW-passively
mode-locked Nd:YLF laser (Millenia X, Spectra Physics, USA) driven Ti-
sapphire pico-second laser (Spectra Physics, Mountain View, CA). Ti-
sapphire pulses at 732 nm frequency doubled to 366 nm using GWU
frequency doubler by Spectra Physics, USA, were used for exciting CDs.
Fluorescence decay curves were obtained at a repetition rate of 8 MHz
using a micro-channel plate photomultiplier (R2809u, Hamamatsu Corp.)
coupled to the time-correlated single photon counting (TCSPC) setup. The
full width at half maximum (FWHM) of the instrument response function
(IRF) was ~40 ps. Fluorescence emission was measured at 440 nm using
Mono-6-thio-β-Cyclodextrin was synthesised as adopted from the literature.
[25] Briefly, 70 g β-Cyclodextrin was suspended in 500 ml water and 20 ml
aqueous solution of 6.6 g NaOH solution was added dropwise which finally
became transparent. 30 ml of 10.1 g p-toluene sulfonylchloride solution
was added to the previous solution to form a white precipitate. The
precipitate was removed by suction filtration and the filtrate was stored at
4°C. Resulting precipitate was separated and recrystallised. 2 g of the
precipitate and 2 g thiourea was heated at reflux for two days in 100 ml
methanol-water solution. The white solid obtained was added to methanol
and stirred. The solid obtained after filtration was dissolved in 10 wt.%
NaOH solution. The pH of the solution was adjusted to 2 with HCl, then 5
ml of trichloroethylene was added and stirred overnight and finally the
precipitate was recovered by suction filtration and was washed with water.
Structure and composition were confirmed with NMR and LCMS as shown
in figures S1 and S2.
a
combination of a monochromator and a 400 nm cut-off filter.
Fluorescence intensity decay curves were collected from the sample after
the excitation with the emission polarizer oriented at the magic angle
(54.7o) with respect to the excitation polarizer. To optimize the signal-to-
noise ratio, 10,000 photon counts were collected in the peak channel. For
time-resolved anisotropy measurements, the emission data were collected
at 0° (parallel fluorescence intensity, I∥), and 90° (perpendicular
fluorescence intensity, I⊥) with respect to the excitation polarization. [27,28]
Synthesis of mono-6-amino-deoxy-6-β-Cyclodextrin
Mono-6-amino-deoxy-6-β-Cyclodextrin was synthesised from a procedure
adapted from Tang et al. [26] 22 mmol of β-Cyclodextrin was added to 400
ml pyridine. 4 g p-toluenesulphonyl chloride was added to the mixture and
stirred for 24 h. Solvent was removed, and the precipitate was stirred with
acetone. Subsequent solid (solid I) was washed with acetone and
recrystallized. 3.9 mmol of Solid I was added to 5 g sodium azide (76.9
mmol) along with 500 ml of deionized water and refluxed at 100°C.
Precipitate was separated and 5 ml of 1,1,1,1-tetrachloroethane was
added dropwise and stirred for 0.5 h. Solvent was evaporated to obtain a
solid (solid II) which was recrystallized using hot water. 5 mmol of Solid II
was refluxed with 5.5 mmol triphenylphosphine and 10 ml dimethyl
formamide for 2 h at room temperature. 1 ml deionized water was then
added and heated at 90°C for 3 h. Mixture was then cooled to room
temperature and the final product was precipitated using acetone, washed
and stored. Structure and composition were confirmed with NMR and
LCMS as shown in figures S1 and S2.
Acknowledgements
Authors are thankful to TIFR’s TEM facility (Mr. Lalit Borde) and
XRD facility (Nilesh Kulkarni). Mrs. Geetanjali Dhotre is also
thanked for assistance with MALDI. The work at TIFR was
supported through N-PDF Fellowship, SERB, Govt. of India
(PDF/2016/000051) and DAE Plan funding.
Keywords: Carbon dots, Excitation dependent fluorescence,
Lifetime, Anisotropy, Homo-FRET.
[1] Liu H, Ye T and Mao C, Angew. Chem. Int. Ed., 2007, 46, 6473−6475.
[2] A, Baker SN and Baker GA Angew. Chem. Int. Ed., 2010, 49, 6726−6744
[3] Zuo P, Lu X, Sun Z, Guo Y and He H, Microchimica Acta, 2016, 183(2),
519-542.
Microwave pyrolysis method to prepare carbon dots
Synthesis was done by one step hydrothermal method in microwave oven
at a pressure of 200 psi and a temperature of 175 °C with precursor
concentration of 3.5 mmol in distilled water. The heating was done for 4hrs.
During the synthesis no precipitate was observed. We ultra-centrifuged the
samples at 35000 rpm but there was no distinct sedimentation. To confirm
the purity, the samples were dialysed against pure water using dialysis
membrane (1000 Da) for 24 hours. Each of the six precursors were
processed separately yielding six different products. For XRD
measurements, the synthesized colloidal solutions were lyophilized to
obtain the corresponding CD solid phase.
[4] Li X, Wang H, Shimizu Y, Pyatenko A, Kawaguchi K and Koshizaki N,
Chem. Comm., 2010, 47(3), 932-934.
[5] Dhenadhayalan N, Lin KC, Suresh R and Ramamurthy P, J. Phys. Chem.
C, 2016, 120(2), 1252-1261.
[6] Khan S, Gupta A, Verma NC and Nandi CK, Nano Lett., 2015, 15(12),
8300-8305.
[7] Li H, He X, Kang Z, Huang H, Liu Y, Liu J, Lian S, Tsang CH, Yang X and
Lee ST, Angew. Chem. Int. Ed., 2010, 49(26), 4430-4434.
[8] Cao L, Anilkumar P, Wang X, Liu JH, Sahu S, Meziani MJ, Myers E and
Sun YP, Can. J. Chem., 2010, 89(2), 104-109.
[9] Wang X, Cao L, Lu F, Meziani MJ, Li H, Qi G, Zhou B, Harruff BA,
Kermarrec F and Sun YP, Chem. Comm., 2009, 25, 3774-3776.
[10] Shamsipur, Mojtaba, Ali Barati, and Sara Karami. Carbon,2017, 124, 429-
472.
Characterization
TEM analysis was done on a LIBRA 120, EFTEM Carl Zeiss. Matrix-
assisted laser desorption/ionization (MALDI) measurements were
performed on a TOF Bruker Daltonics Spectrometer. Low-resolution mass
spectral analyses were carried out using a Shimadzu LCMS-2020 with an
[11] Kong W, Wu H, Ye Z, Li R, Xu T and Zhang B, J. Lumin., 2014, 148, 238-
242.
This article is protected by copyright. All rights reserved.

10.1002/cphc.201801061
ChemPhysChem
ARTICLE
[12] Lim, Shi Ying, Wei Shen, and Zhiqiang Gao. Chem. Soc. Rev.,
2015,44(1),362-381.
[21] Dekaliuk, M. O., Viagin, O., Malyukin, Y. V., and Demchenko, A. P, Phys.
Chem. Chem. Phys., 2014, 16(30), 16075-16084.
[22] Strambini, G. B., and Gonnelli, M., J. Chem. Phys. B, 2010, 114(2), 1089-
1093.
[13] Hu Q, Paau M. C, Zhang Y, Chan W, Gong X, Zhang L and Choi M. M.,
J. Chromatogr. A, 2013, 1304, 234-240.
[14] Zhang, J., Shen, W., Pan, D., Zhang, Z., Fang, Y., and Wu, M., New J.
Chem, 2010, 34(4), 591-593.
[23] J. R. Lakowicz, Principles of Fluorescence Spectroscopy, Springer, New
York, 3rd edn, 2006.
[15] Puvvada, N., Kumar, B. P., Konar, S., Kalita, H., Mandal, M., and Pathak,
A, Sci. Technol. Adv. Mater, 2012, 13(4), 045008.
[24] Guo, X. Q., Castellano, F. N., Li, L., & Lakowicz, J. R., Anal. Chem., 1998,
70(3), 632-637.
[16] Paredes, J. I., Villar-Rodil, S., Martinez-Alonso, A., and Tascon, J. M. D,
Langmuir, 2008, 24(19), 10560-10564.
[25] Wang J, Kong L, Guo Z, Xu J and Liu J, J. Mater. Chem., 2010, 20(25),
5271-5279.
[17] Hu, Q., Paau, M. C., Zhang, Y., Chan, W., Gong, X., Zhang, L., and Choi,
M. M., J. Chromatogr. A, 2013, 1304, 234-240.
[26] Tang W, & Ng S. C, Nat. Protoc., 2008, 3(4), 691-697.
[27] Mukhopadhyay S, Nayak, P. K., Udgaonkar, J. B. and Krishnamoorthy G.,
J. Mol. Biol., 2006,358(4), 935-942.
[18] Deng, Y., Chen, X., Wang, F., Zhang, X., Zhao, D., and Shen, D,
Nanoscale, 2014, 6(17), 10388-10393.
[28] Jha A, Narayan S, Udgaonkar J. B, and Krishnamoorthy G, Biophys. J.,
2012, 103(4), 797-806.
[19] Jaiswal, A., Ghosh, S. S., and Chattopadhyay, A, Chem. Comm., 2012,
48(3), 407-409.
[20] Lin, L. P., Wang, X. X., Lin, S. Q., Zhang, L. H., Lin, C. Q., Li, Z. M., and
Liu, J. M, Spectrochim. Acta A, 2012, 95, 555-561.
This article is protected by copyright. All rights reserved.

10.1002/cphc.201801061
ChemPhysChem
ARTICLE
Entry for the Table of Contents
ARTICLE
S. Divya, Satya Narayan, Sri Rama Koti
Ainavarapu and Deepa Khushalani*
Page No. – Page No.
Insight into Excitation Dependent
Fluorescence of Carbon Dots
•
•
•
Random dispersion of fluorophores
Fast-FRET (<150 ps)
Collisional quenching
This article is protected by copyright. All rights reserved.