Structure and solvent-induced tuning of laser property and photostability of a boradiazaindacene (BODIPY) dye

Article abstract of DOI:10.1016/j.jphotochem.2017.09.033

Boradiazaindacene (BODIPY) class of laser dyes are highly efficient but degrade rapidly in alcohol solution during lasing action, progressively reducing output power. Photodegradation of these dyes is mainly due to reaction with the in situ generated singlet oxygen (¹O₂). With the aim to increase the lasing lifetimes of these dyes, we have designed and synthesized a new congener of the widely used BODIPY dye (PM567) by substitution at its B-centre. The new dye was highly fluorescent, and showed comparable lasing efficiency, but better photostability relative to PM567 in both polar (ethanol) and non-polar (1,4-dioxane) solvents, when excited by 2nd harmonic (532 nm) of a pulsed Nd-YAG laser. More interestingly, the lasing efficiency and photostability of both the dyes were much better in 1,4-dioxane than in ethanol. The relative photostabilities of the dyes were rationalized by absorption spectroscopic analyses of their triplet state properties in the respective solvents in the presence of a ¹O₂ generator or quencher as an additive, pulse radiolysis studies and quantum chemical calculations.

Full text of DOI:10.1016/j.jphotochem.2017.09.033

Accepted Manuscript
Title: Structure and solvent-induced tuning of laser property
and photostability of a boradiazaindacene (BODIPY) dye
Authors: Monika Gupta, Soumyaditya Mula, Tapan K.
Ghanty, D.B. Naik, Alok K. Ray, Anubha Sharma, Subrata
Chattopadhyay
PII:
S1010-6030(17)30735-9
DOI:
Reference:
http://dx.doi.org/10.1016/j.jphotochem.2017.09.033
JPC 10878
To appear in:
Journal of Photochemistry and Photobiology A: Chemistry
Received date:
Revised date:
Accepted date:
16-6-2017
30-8-2017
11-9-2017
Please cite this article as: Monika Gupta, Soumyaditya Mula, Tapan
K.Ghanty, D.B.Naik, Alok K.Ray, Anubha Sharma, Subrata Chattopadhyay,
Structure and solvent-induced tuning of laser property and photostability of a
boradiazaindacene (BODIPY) dye, Journal of Photochemistry and Photobiology A:
Chemistryhttp://dx.doi.org/10.1016/j.jphotochem.2017.09.033
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Structure and solvent-induced tuning of laser property and photostability
of a boradiazaindacene (BODIPY) dye
a,e
b,
c,e
d,e
a,e
Monika Gupta, † Soumyaditya Mula, † Tapan K. Ghanty, D. B. Naik, Alok K. Ray,
b,e
b,e
Anubha Sharma and Subrata Chattopadhyay *
aLaser and Plasma Technology Division, Bhabha Atomic Research Centre, Mumbai-400085, India.
bBio-Organic Division, Bhabha Atomic Research Centre, Mumbai-400085, India.
cTheoretical Chemistry Section, Bhabha Atomic Research Centre, Mumbai-400085, India.
dRadiation and Photochemistry Division, Bhabha Atomic Research Centre, Mumbai-400085, India.
eHomi Bhabha National Institute, Anushakti Nagar, Mumbai-400094, India.
E-mail: schatt@barc.gov.in (SC)
Graphical Abstract:
Enhancement of photostability of BODIPY dye based on changes in its structure in
conjunction with solvent polarity.
1

Highlights:


Highly efficient B-substituted BODIPY laser dye was synthesized.
The BODIPY dye was 34 times more photostable in non-polar 1,4-dioxane
solvents than in polar ethanol.

High photostability of the BODIPY dye in non-polar solvent is due to its lower
1
reactivity with O
2
.


Rationalization of the results by theoretical calculations.
Besides alkyl substitution at the B-atom, use of less polar 1,4-dioxane medium
in combination, can increase the photostability of the BODIPY dye immensely
without compromising its laser property.

The photostability of the new BODIPY dye in 1,4-dioxane is comparable to
that of Rhodamine 6G in ethanol.
ABSTRACT. Boradiazaindacene (BODIPY) class of laser dyes are highly efficient but
degrade rapidly in alcohol solution during lasing action, progressively reducing output power.
Photodegradation of these dyes is mainly due to reaction with the in situ generated singlet
1
oxygen ( O
2
). With the aim to increase the lasing lifetimes of these dyes, we have designed
and synthesized a new congener of the widely used BODIPY dye (PM567) by substitution at
its B-centre. The new dye was highly fluorescent, and showed comparable lasing efficiency,
but better photostability relative to PM567 in both polar (ethanol) and non-polar (1,4-dioxane)
solvents, when excited by 2nd harmonic (532 nm) of a pulsed Nd-YAG laser. More
interestingly, the lasing efficiency and photostability of both the dyes were much better in 1,4-
dioxane than in ethanol. The relative photostabilities of the dyes were rationalized by
absorption spectroscopic analyses of their triplet state properties in the respective solvents in
1
the presence of a O
2
generator or quencher as an additive, pulse radiolysis studies and
quantum chemical calculations.
2

1
. Introduction
Boradiazaindacene [1-6] (4,4-difluoro-4-bora-3a,4a-diaza-s-indacene, BODIPY) dyes
are extensively used as efficient active media for solid as well as liquid dye lasers due to their
high fluorescence yields, intense absorption profiles in the green spectral region and low
triplet-triplet extinction coefficients over the fluorescence region [7-13]. However, their
application in high-average-power liquid dye lasers remains limited due to the rapid
photochemical degradation of the dyes. For example, after irradiation at 532 nm for 4 h under
non-lasing conditions, the lasing efficiency of the commercially available BODIPY dye,
PM567 (1a) was reduced by 36% of its initial value [12]. The degradation is primarily
1
mediated by their reaction with singlet oxygen ( O
2
), generated in situ during photochemical
excitation of these dyes [11,14,15]. Different strategies to enhance the laser photostability of
these dyes such as the use of deoxygenated dye solutions [16] and/or addition of high
1
concentrations of O
2
quenchers (TIN770, TBP or DABCO) [14,17,18], are inconvenient for
large-scale dye laser operation. Moreover, their laser efficiency may also reduce due to
fluorescence quenching. The photophysics and photostability of dye solutions generally
depend on the dye structure as well as nature of the chosen solvent, because of the specific
and non-specific solute-solvent interactions. Change of chemical structures of the BODIPY
3

dyes has significant effect on their photostabilities [13]. We showed earlier that the PM567
analogues, synthesized by substitution at the (i) meso position, 1b; (ii) boron centre, 2a;
and/or (iii) both the sites, 2b were markedly more photostable than 1a [11,12]. Earlier,
solvent-dependent change in the photostability of BODIPY dye 1a has been reported even
under air saturated conditions [19]. More recently, our group has designed photostable
BODIPYs based on changes in their structures as well as solvent polarities [20-22]. To this
end, new BODIPYs with an N-ethyl carbazole at the meso position and Et or Bn groups at the
2
- and 6-positions viz. 3a and 3b have shown good photostabilities in heptane solvent, albeit
with reduced lasing efficiencies [20]. In a related study, the BODIPYs 4a-c, possessing 2,6-
dibenzyl groups showed improved photostabilities in 1,4-dioxane solvent, irrespective of the
meso-substituents (Me, Ph or H) [21]. The chemical structures of the BODIPY dyes 1-4 are
shown in Figure 1. However, syntheses of the BODIPYs 3 and 4 required multiple steps that
may impede their practical laser applications.
Earlier, we showed that the B-substituted dye 2a has a significantly enhanced
o
photostability than its BF
2
-precursor, 1a in polar ethanol (ε = 24.5 @ 25 C) solvent [12].
Mechanistic investigation via a transition state (TS) model suggested higher activation energy
1
for the reaction between O
2
with 2a than that of 1a. This made dye 2a less vulnerable to
oxidation by singlet oxygen than 1a. The TS model also suggested that the activation energy
1
for the reaction between O
2
with 2a can be increased further by lowering the solvent polarity
which may make the B-substituted dye more robust against oxidation. Thus, the aim of the
present study was to evaluate the lasing efficiency and photostability of the B-substituted dye
in a less polar solvent. In general, for laser studies, solvents such as water, methanol, ethanol
and dioxane are best suited for their high specific heat capacity, low refractive index (RI)
temperature coefficient, low viscosity, low inflammability, high thermal diffusibility etc.
0
Amongst these, dioxane is the least polar (ε = 2.25 @ 25 C) solvent. But unfortunately,
4

solubility of the dye 2a in 1,4-dioxane was not enough to perform lasing experiments. Hence,
we synthesized a new B-substituted BODIPY dye 5 (Scheme 1) and examined its
photophysical and lasing properties in polar ethanol with respect to its precursor dye 1a. Then
we evaluated the photophysical and lasing properties of 1a and 5 in 1,4-dioxane solvent to
determine the relative utility of these dye-solvent combinations for high average-power dye
lasers.
Figure 1. Chemical structures of PM567 and some BODIPYs developed by our group.
2
. Materials and methods
2
.1 Experimental Details
All the chemicals and spectroscopic grade solvents were purchased from Aldrich,
Merck or Sigma and used without any further purification. The synthesis of the dye 5a was
carried out as shown in Scheme 1, and the experimental details are provided as the Supporting
o
information. The spectrophotometry was carried out at 25 C with a Jasco V-550 UV-vis
spectrophotometer and Jasco FP 6500 spectrofluorometer respectively. The fluorescence
spectra were corrected.
2
.2 Photophysical Studies
5

-6
The absorption and emission spectra of the dyes (~10 M) in various solvents were
measured using a 1 cm quartz cuvette. The fluorescence quantum yields (fl) of the dye 5
relative to that of the reference dye 1a [11] and the molar extinction coefficients (max) were
determined in ethanol and 1,4-dioxane. The following equation was used to determine the
relative fluorescence quantum yield:
2
S X X S
(X) = (A /A )(F /F )(n
X
/n
S
) (S)
where A is the absorbance at the excitation wavelength, F is the area under the corrected
emission curve and n is the refractive index of the solvents (at 25 °C) used in the
measurements. The subscripts S and X represent standard and unknown, respectively. The
excited state (S ) lifetimes of 1a and 5 in ethanol and 1,4-dioxane were determined by the
1
time-resolved fluorescence measurements, carried out using an LED based time-correlated
single-photon-counting (TCSPC) spectrometer. The fluorescence decays were measured with
a 490 nm LED (1 MHz) excitation source and a TBX4 detection module coupled with a
special Hamamatsu PMT. The instrument response function was 1.2 ns at FWHM. Following
deconvolution analysis of the fluorescence decays, the time resolution of the present setup
was ~50 ps. All the measurements were carried out at ambient room temperature (298 ± 1 K)
using a microprocessor based temperature controller.
2
.3 Lasing studies
The lasing studies of the dyes 1a and 5 in ethanol and 1,4-dioxane were carried out
using a constructed narrow-band dye laser set-up, transversely pumped by the second
harmonic (at 532 nm) output of a Q-switched pulsed Nd:YAG laser at a repetition rate of 10
Hz with ~6 mJ pulse energy and 5-7 ns FWHM pulses. A schematic of the dye laser set up is
shown in Figure S1a. The dye laser was set up in grazing incidence grating (GIG)
configuration (with a grating of 2400 lines/mm) with 25 X 4-prisms pre-expander. The pump
and dye laser powers were measured by the same power meter (OPHIR).
6

The dye solutions of 1a and 5 (O. D. = 7 at 532 nm, 1 cm cell) in ethanol and 1,4-
dioxane were used to determine the narrow band lasing efficiency. The tuning curves of the
dye solutions were obtained by scanning the wavelength of dye laser through the gain profiles
of the dyes, and measuring the dye laser wavelength by a wavemeter, average pump and dye
laser powers with a power meter. For determining the pump laser threshold (L ), the output
T
power of dye lasers, while tuning the dye laser wavelength at the peak of the respective gain
curves, were measured as a function of input pump power at 532 nm.
2
.4 Photostability studies
The dye solution (2 mL) in a dye laser cuvette was exposed to pump energy of ~4 mJ
at 532 nm. The concentration of the dye solution was chosen such that the pump beam was
totally absorbed within the dye solution in the cuvette during the excitation period (4 h). The
solution was constantly stirred with a teflon-coated magnetic stirrer to avoid local heating.
The number of photodegraded dye molecules in the exposed volume of the dye solutions was
estimated from the absorbances (at the corresponding λabs) before and after photoexposure for
a set period of time. The reflection loss of the pump beam on incident surfaces of the dye cell
was taken into consideration while calculating the absorbed cumulative pump photons.
2
.5 Electrochemical studies
Cyclic voltammetry (CV) study was carried out at 25 °C in deoxygenated CH
2
Cl
2
-3
containing TBAP (0.1 M), and a solute concentration of 1-5 × 10 M. The redox potentials
were standardized with ferrocene (Fc) as the internal reference, and converted to SCE
+
assuming that E1/2 (Fc/Fc ) = +0.38 V SCE. The error in half-wave potentials was ±10 mV.
-1
All waves were monoelectronic. CV scans were recorded at a scan rate of 50 mVs . All the
electrochemical experiments were carried out using Potentiostat Autolab 100. A glassy carbon
7

electrode, cleaned by ultra-sonication and dried by dry air flow, was used as the working
electrode.
1
2
.6 Studies of relative O
2
generation of the dyes
A solution of 1,3-diphenylisobenzofuran (DPBF) (50 μM) and dye 1a or 5 (5 μM
each) in aerated ethanol (3 mL) taken in a 10 mm quartz cell was irradiated with a filtered
2
visible light source (estimated intensity of 20 W/m and wavelength >495 nm) at room
temperature for 15 min. The 40 W lamp was operated using a voltage stabilized power supply
to obtain steady power. The absorbances of the solutions were measured at λabs of DPBF (412
1
nm) at an interval of 2 min. The relative O
2
generation capacities of dyes 1a and 5 in ethanol
were determined by measuring the time-dependent reduction of the DPBF absorbance.
Solutions of DPBF (negative control) and of the dyes 1a and 5 (positive controls) were also
1
irradiated. Similar experiments were carried out in 1,4-dioxane to measure the relative O
2
generation capacities of the dyes 1a and 5 in 1,4-dioxane.
2
.7 Pulse radiolysis
The triplet states studies of the dyes were carried out using the nanosecond pulse
radiolysis technique [23]. The dose absorbed per pulse was determined using an aerated 0.01
-
3
-
-4
2 -1
mol dm KSCN solution, taking (G.) of (SCN)
dose was kept at 20 Gray per pulse. The dye solutions (1  10 mol dm ) were prepared in
spectroscopy grade 1,4-dioxane/benzene (9:1) and purged with high purity N to remove
dissolved oxygen. The pulse radiolysis experiments were carried out using 4,4,-
2
as 2.59  10 m J at 475 nm [24]. The
-
4
-3
2
-3
-3
dimethylbiphenyl (DMBP) (5  10 mol dm ) as the triplet sensitizer. Formation of the dye
triplets and their extinction coefficients were determined by energy transfer from the DMBP
triplet to the dyes. For studying the reactions of the dyes at triplet state with O
2
, the
-3
-3
experiments were carried using a high concentration of the dye (3  10 mol dm ) without
any triplet sensitizer. The decays of the triplet states of the dyes were monitored at their
8

respective λmax. Concentration of oxygen in solutions was adjusted by mixing appropriate
volumes of nitrogen purged dye solutions in 1,4-dioxane/benzene (9:1) with aerated dye
solutions in 1,4-dioxane/benzene (9:1).
1
2
.8 Studies of O
2
reactivity with dyes
An aerated ethanol solution (3 mL) of methylene blue (MB) (10 μM) and dye 1a or 5
2
(
each, 10 μM) in 10 mm quartz cells was irradiated at 7.5 W/m of filtered light source of
wavelength > 590 nm at room temperature for 1.5 h. A 590 nm cut off filter was used during
irradiation to ensure that the light was absorbed by the photosensitizer, MB only. Absorption
at λabs of dye 1a (518 nm) and λabs of dye 5 (515 nm) were measured at various fixed intervals
1
(
15 min.). Relative reactivities of the dyes 1a and 5 with O
2
in ethanol were determined by
measuring the reduction in absorbance of the respective dyes at their λabs over time.
Irradiation of MB solution in the absence of dyes as a negative control and solution of dye 1a
and dye 5 as positive controls were also carried out. Similar experiments were carried out in
1
1
,4-dioxane/ethanol (1:4) to measure relative O
2
generation capacities of dye 1a and 5 in 1,4-
dioxane/ethanol (1:4). The experiments could not be done in 1,4-dioxane due to poor
solubility of MB.
2
.9 Theoretical calculations
The minimum energy structures in the ground (S ) state of all BODIPY dyes (5) was
0
established by applying a correlated hybrid density function (B3LYP), using 6-31G* basis
sets for all the atoms. The Quasi Newton-Raphson based algorithm was used to carry out
geometry optimization for each of these dyes with various possible conformers as the initial
structures. The minimum energy structures of transition states (TS) and products for the
1
reaction of O
2
and the dyes were calculated by applying the GAMESS suite of ab initio
electronic structure program. TS structure was confirmed by Hessian calculation having only
one imaginary frequency.
9

3
. Results and discussion
3
.1 Design and synthesis
For the present work, we designed the new dye
5 with a propargyl alcohol
appendage at the B-centre. The appendage was chosen to keep a balance between the
hydrophobicity and hydrophilicity of the dye to ensure its solubility in both ethanol and
1
1
,4-dioxane for the lasing studies. Moreover, the electronic excitation energy of O
2
is
comparable with the energy associated with the high frequency vibrations of C-H and
O-H bonds. This energy matching may lead to an electronic-to-vibrational energy
1
1
transfer between O
2
and the dye to increase the O quenching rate (kΔQ) [25]. Hence it
2
was hypothesized that the extra C-H and especially the two O-H bonds of the dye
5
1
Δ
1
may assist physical quenching of O
2
and reduce the lifetime (

) of O
2
in both the
Δ
1
solvents. This appeared to be particularly crucial in 1,4-dioxane, as

(27.0

s) of O
2
in 1,4-dioxane is almost 3 times to that (9.7
s) in ethanol [25,26].
o
Scheme 1. Synthesis of dye
5
. a) DHP, PPTS, CH
2
Cl
2
, 25 C, 12 h. b) EtMgBr, THF,
o
o
5
0 C, 2 h, 1a, 24 h. c) PTS, MeOH/H
2
O (9:1), 25 C, 10 h.
For the synthesis (Scheme 1), the hydroxyl group of propargyl alcohol (
tetrahydropyranylated by a pyridinium para-toluenesulphonate (PPTS)-catalyzed
reaction with 2,3-dihydropyran (DHP) to obtain . The Grignard reagent, prepared
from was subsequently coupled with the dye 1a to obtain the dye [27]. Finally, its
6) was
7
7
8
1
0

depyranylation with para-toluenesulphonic acid (PTS) in aqueous MeOH furnished the
dye
5
.
3.2 Photophysical properties
The photophysical properties of the dyes 1a and
5 in ethanol and 1,4-dioxane
o
were evaluated at room temperature (25 C). The measured photophysical parameters
(
longest wavelength absorption maximum (λabs), emission maximum (λem),
fluorescence quantum yield (Φfl), maximum molar absorptivity (εmax) and fluorescence
lifetime ( )) and the calculated Stokes shift ( ), radiative (k ) and non-radiative (knr
decay rates of the dye relative to those of dye 1a in both the solvents are presented in
Table 1. The normalized (to keep the spectral heights similar for the individual)
absorption and emission spectra of dyes 1a and in ethanol (Figure 2) and 1,4-dioxane
Figure S7) were similar, each showing a strong S0→S1 absorption band at about 520
nm, typical of the BODIPY dyes [11,12,28-31]. This indicated that the energy of the
S1 states of 1a and are similar in these solvents. Both the dyes were highly
fluorescent (Φfl > 0.78). The λabs and λem of were blue shifted by 3 nm than that of 1a
in both the solvents. Also, the λem values of 1a and were red shifted by 2 nm in 1,4-
dioxane than in ethanol. Thus, the Stokes shifts of the dyes were marginally more in

F

r
)
5
5
(
5
5
5
1,4-dioxane than in ethanol.
According to Strickler-Berg equation, the absorption coefficient (max) and the
radiative rate (k
present studies, while the εmax values for the dyes 1a and
changing the solvent from EtOH to 1,4-dioxane, their respective k
r
) are approximately proportional to each other. However, in the
were found to decrease on
values increased.
5
r
This may be due to the specific interaction (hydrogen bonding) of solvent (ethanol)
with the dye molecules. Similar observations in protic and aprotic solvents with dye 1a
and other BODIPY dyes have been reported previously [32-34].
1
1

Table 1. Photophysical properties of the BODIPY dyes 1a and 5 in ethanol and 1,4-dioxane
o
solvent at 25 C.
[a]
[b]
max
[c]
em
[d]
fl
[e]
Dye Solvent
λ
abs
ε
λ

Φ
F
k
r
k
nr
4
-1
-1
8
-1
8 -1
-
1
[
nm]
[10 M cm ] [nm]
(10 s ) (10 s )
[
cm ]
[ns]
[
f]
5
5
5
5
18.0 7.1
18.0 6.2
15.0 8.6
15.0 6.7
533.0 543.3
535.0 613.4
530.0 549.6
532.0 620.5
0.84 6.48
0.89 6.52
0.78 6.77
0.89 7.18
1.30
1.36
1.15
1.24
0.25
0.17
0.32
0.15
Ethanol
1
5
a
1
,4-dioxane
Ethanol
,4-dioxane
1
[
a] Error: ± 0.2 nm. [b] Extinction coefficients at respective λmax. [c] Error: ± 0.3 nm. [d]
Error: ± 0.04. The fluorescence of the dyes 1a and 5 were measured at the excitation
wavelength 505 nm. The fluorescence quantum yield of 5 is relative to that of 1a, reported
earlier [11]. [e] Error: ± 0.05 ns. [f] This data is taken from ref. [11].
1
2

Dye 1a
Dye 5
1.0
0.8
0.6
0.4
0.2
0.0
4
50
500
550
600
650

(nm)
Figure 2. Normalized absorption (bold lines) and emission (dotted lines) spectra of the dyes
1
a (blue)and 5 (black) in ethanol.
The respective εmax values of dye
5 in the chosen solvents were higher than that
of dye 1a. Interestingly, the Φfl values of both the dyes were similar in 1,4-dioxane,
and more than those in ethanol. These data clearly revealed that substitution of the F
groups at the B-atom with the propargyl alcohol appendage had little effect on the
optical properties of the BODIPY chromophore [12]. Taken together, the
-
photophysical properties of the dyes 1a and
5
in the chosen solvents appeared
nd
favorable for their use as green wavelength (CVL- 511 nm or 2 Harmonic of
Nd:YAG – 532 nm)-pumped laser dyes. More importantly, superior lasing
performances of 1a and
5 were anticipated in 1,4-dioxane because of their better
optical properties in it.
3
.3 Lasing characteristics
For the laser experiments, the dye concentration needs to be optimized so that the
pump beam (80-90% intensity) will be almost uniformly absorbed within the dye solution in
the dye cell. Additionally, the spatial profile of the dye laser output should be approximately
circular i. e., divergence of the dye laser beam is similar in both horizontal and vertical
1
3

directions. If the solution is too dilute, then majority of the pump beam will mostly escape
interaction with the dye solution. On the other hand, if it is too concentrated, then the pump
beam will be absorbed by the dyes at the glass surface only. In that case, due to local heating
of the dye solution, scattering of the laser beam will be increased, decreasing the laser output.
Presently, the absorption and emission spectra of 1a and 5 showed substantial overlap,
implying the existence of some homo-FRET that may lead to concentration-dependent non-
radiative decay, reducing the fluoroscence quantum and laser efficiency. In the present
studies, the optimum concentration of 1a and 5 for the narrow-band lasing studies were first
determined by the broad band lasing experiments. For this, the concentration-dependent broad
band lasing efficiencies () of the respective dyes in ethanol solution were first determined,
keeping other parameters, including cavity alignment unchanged. The concentration-
dependent lasing efficiencies of the dyes followed an expected pattern, initially increasing
with the dye concentration to reach a maximum before decreasing again (Figure S1b). The
narrow band lasing studies were subsequently carried out at the concentrations where the
respective dyes showed maximum  values. For each dye, the same concentration was used
for both ethanol and 1,4-dioxane solutions.
The narrow band lasing characteristics of the BODIPY dyes 1a and
5 in ethanol
and 1,4-dioxane are shown in Table 2 and Figure 3. The narrow band lasing profiles of
the dyes in both the solvents followed an expected pattern, showing a maximum
efficiency at particular wavelengths (λ
observed photophysical properties, the laser tuning ranges and λ
were similar in both ethanol and 1,4-dioxane. The maximum lasing efficiency values
η) of the dyes 1a (15.3% in ethanol, 17.1% in 1,4-dioxane) and (11.8% in ethanol,
5.3% in 1,4-dioxane) at the respective λ (Table 2) revealed insignificant reduction in
the laser efficiency due to substitution at the B-centre. The threshold pump energy (L
L
), characteristic of the dyes. Consistent with the
L
values of 1a and
5
(
5
1
L
T
)
1
4

values (Table 2, column 6) of 1a and
However, the L values of were lesser than that of 1a in the respective solvents.
Despite this, both the dyes showed higher narrow band dye lasing (DL) efficiencies in
,4-dioxane than in ethanol. This is in accordance with their larger radiative and
smaller non-radiative decay rates in 1,4-dioxane (Table 1). The excited-state
absorption (ESA, S –S and T -T processes) may also contribute to this, although the
5 were larger in 1,4-dioxane than in ethanol.
T
5
1
1
n
1
n
ESA cross-sections for the BODIPY derivatives are very low [35-37].
Table 2. Lasing properties of the BODIPY dyes 1a and 5 in ethanol and 1,4-dioxane. Pump
nd
laser: 2 harmonic of Nd:YAG at 532 nm.
[
a]
Tuning range [b]
λ
L
T
L [mJ] GSA
Dye Solvent
-18
2
[
nm]
[nm]
[10 cm ] at λ
L
[
%]
5
5
5
5
53.8
55.8
54.0
54.0
541-583
15.3
17.1
11.8
15.3
0.29
0.61
0.28
0.49
1.34
Ethanol
1
a
540-584
541-581
540-580
1.36
1.45
1.06
1
,4-dioxane
Ethanol
,4-dioxane
5
1
[
a] Error: ± 0.2 nm. [b] Error: ± 1.0 %
1
5

dye 5 / 1, 4-dioxane
dye 5 / ethanol
dye 1a / 1, 4-dioxane
dye 1a / ethanol
1
1
6
2
8
4
0
5
40
555
570
585
(nm)
Figure 3. Narrow band dye laser (DL) efficiency of 1a and 5 in ethanol and 1,4-dioxane,
determined by pumping with 532 nm radiation of a Q-switched pulsed (10 Hz) Nd:YAG laser.
Next, the ground-state absorption (GSA) of both the dyes in the chosen solvents at the
respective 
L
values were determined by absorption spectroscopy. The GSA cross-section
1.06) of dye 5 in 1,4-dioxane was much lower than that (1.45) in ethanol, as well as those
1.34 and 1.36 respectively) of dye 1a in these solvents. Thus, the loss of laser photons with
(
(
dye 5 would be less in 1,4-dioxane due to lower GSA. Nevertheless, its DL efficiency was
slightly less than that of 1a. This may be because various other factors such as effective
stimulated emission cross-section, the cavity losses and dynamic changes caused by time-
dependent saturation of the gain contribute to the net gain, hence, the DL efficiency.
3
.4 Photostability
The photostabilities (Φpd-1) of the dyes 1a and 5 were studied in air-equilibrated
ethanol and 1,4-dioxane under non-lasing conditions by measuring the photodegradation
quantum yields (Φpd). Photo-cleavage of the dyes would take place at the olefinic site to
produce smaller molecules that do not absorb visible light. Hence, no change in the max
1
6

values and/ or additional peaks were seen in the absorption spectra of the photo-exposed dyes
11]. But, the photo-degradation was evident from the reduced peak heights of their longest
wavelength absorption bands (S →S ).
[
0
1
Table 3. Photostabilities of the dyes 1a and 5 in ethanol and 1,4-dioxane.
[
a]
-1
pd dye 5
-1
[a]
pdDABCO
-1
Φ
pd
Φ
Φ
-
/
Φ
Φ
-
pd dioxane
/
Φ
Φ
pd DABCO
/
Dye
Solvent
Ethanol
-
1
-1
-1
pd dye 1a
pd ethanol
Φ
pd
6.7 x 10-4
.1 x 10-5
3.2 x 10-4
.4 x 10-6
1.5 x 10-4 4.5
1.5 x 10-5 2.1
1.4 x 10-4 2.3
8.1 x 10-6 1.2
1
5
a
3
-
21.6
-
1
,4-dioxane
Ethanol
,4-dioxane
a] Error: ± 1 %.
Our results revealed that dye
Table 3, column 4) which is comparable with the photostability of 2a in ethanol [12].
was further increased (3.3 fold)
compared to 1a (Table 3, column 4) and was comparable to that of the established
2.1
3.3
9
34.0
1
[
(
5
was 2.1 fold more photostable than 1a in ethanol
But in 1,4-dioxane, the photostability of dye
5
-
6
photostable laser dye, rhodamine 6G (Φpd = 1.8 x 10 ) [38] in ethanol. Most
importantly the dye solution in dioxane was more photostable than that of our earlier
5
developed dye 2a in ethanol solution. Overall, the present data clearly established that
substitutions of the F- atoms with the propargyl alcohol moieties at the B centre
improved the photostability of the BODIPY, irrespective of the solvent used. We also
showed that the solvent nature had a prominent role in the photostability of the
BODIPY dyes. The dyes 1a and
5
were 21.6 and 34.0 fold more photostable in 1,4-
1
dioxane than in ethanol (Table 3, column 5). The results are surprising, given that O
2
1
7

has a significantly higher lifetime in 1,4-dioxane than in ethanol. Interestingly, the
effect of solvent polarity on stability enhancement was more predominant with dye
5
compared to dye 1a
.
1
Addition of the O
2
quencher, DABCO improved photostabilities of the dyes 1a
1
and
5
in both the solvents (Table 3, column 6), confirming that the generated O
2
is
responsible for the dye degradation. It was noticed that photostability enhancement by
the same molar concentration of DABCO (50 times to dye) was almost half in case of
5
than with 1a in both the solvents (Table 3, column 7). Also, the effect of DABCO
was more pronounced in ethanol than in 1,4-dioxane for both the dyes. Indeed DABCO
has an insignificant effect on the photostability of
5
in 1,4-dioxane. Taken together,
1
1
these data clearly indicated that O
2
generation capacity and/or reactivity with O
2
of
(
i) dye 5 are less than that of dye 1a and (ii) both the dyes are less in 1,4-dioxane than
in ethanol. These aspects were further investigated by a series of experiments as
described below.
3
.4 Cyclic voltammetry (CV)
The CV technique is conveniently used to assess the electrochemical properties
such as the number of electrons per molecule oxidized or reduced, the diffusion
coefficient, redox potential etc. of an analyte in solution. It involves a time-dependent
increase of the working electrode potential at a specific rate to reach a set value,
followed by reversing the sweep to the initial potential. In the present experiments, the
potentials were scanned in the window of 0 to 1.6 V, the current was measured
continuously during the potential scan, and plotted against the applied potentials to
generate the current-voltage curves as shown in Figure 4. The relative oxidation
potentials (Eox) of the dyes 1a and
5 and the effect of solvent polarity on the Eox values
were investigated by recording their CVs in highly polar acetonitrile (ACN) and less
1
8

polar dichloromethane (DCM). Each of the dyes showed a reversible peak in the
anodic portion of the CVs (Figure 4), which was assigned to one electron oxidation of
the BODIPY unit [39]. Substitution at the B-centre had a negative effect on the
oxidation potential. Thus, compared to 1a, dye
and 60 mV in DCM and ACN, respectively (Table 4). Thus, the trend in oxidation
probability of dyes in the chosen solvents was 1a , which is opposite to the
5 was more easily oxidizable by 70 mV
<
5
observed rate of their photodegradation (Φpd values, Table 3).
Dye 1a / DCM
Dye 1a / ACN
0.90
0.45
0.00
Dye 5 / DCM
Dye 5 / ACN
0.0
0.4
0.8
1.2
1.6
potential / V vs Ag/AgCl
Figure 4. Cyclic voltammograms of the BODIPY dyes 1a and 5 in DCM and ACN at room
temperature.
Table 4. Electrochemical properties of 1a and 5 in DCM and ACN.[a]
Eox [V]
ox
E , DCM- Eox, ACN [mV]
Dyes
Solvent
1
a
DCM
ACN
1.00
0.98
20
1
9

5
DCM
ACN
0.93
0.92
10
-3
[
a] The Eox values were determined by CV at a solute concentration of 1-5 × 10 M in
deoxygenated solvents containing TBAP (0.1 M), at 25 °C. The Eox values were
standardized with ferrocene (Fc) as the internal reference and converted to SCE assuming
that E1/2 (Fc/Fc+) = 0.38 V SCE. The error in half-wave potentials was ± 10 mV. All
waves were monoelectronic.
Nevertheless, decrease in the solvent polarity had a positive effect on the Eox
values of the dyes. The dyes 1a and
5 were less oxidizable in DCM than in ACN by 20
mV and 10 mV, respectively (Table 4). Thus, both the dyes would be more resistant to
oxidation in non-polar solvents than in polar ethanol, as observed experimentally. But
the CV results could not explain the higher photostability of dye
5 than dye 1a, and the
much higher stability of both the dyes in 1,4-dioxane. Hence, we measured the relative
1O
generation rates of both the dyes as well as their reactivity with O
1
, and
2
2
rationalized the results by triplet state studies using pulse radiolysis and theoretical
1
calculations of the ground and excited states of the dyes as well as of O
2
. These are
sequentially discussed below.
0.9
0.6
0.3
0.0
a
b
3
60
420
480
540

(nm)
2
0

Figure 5. Change in absorption spectra of DPBF (50 µM) in 1,4-dioxane during dye-
sensitized photooxidation in presence of dye 1a (5 µM): (a) 0 min; (b) 15 min. A stabilized 40
W tungsten lamp, with a cutoff filter allowing exposure >495 nm was used for photo-
irradiation.
1
3
.5 Relative O
2
generation by 1a and 5
1
The O
2
generation capacities of the dyes 1a and 5 were determined by
monitoring the dye-sensitized photooxidation of 1,3-diphenylisobenzofuran (DPBF)
40-42]. For this, the ethanol and 1,4-dioxane solutions of DPBF and 1a or were
irradiated over a time period of 900 s, using a stablized tungsten lamp with a
[
5
wavelength cut off filter >495 nm. Under these conditions, only the BODIPY dyes
1
were excited to generate O
2
, which would primarily react with DPBF because the
1
reaction rate of O
2
is significantly higher with DPBF than the BODIPY dyes [43,44],
and also, DPBF was taken at 50 times higher molar concentration in comparison to the
BODIPY dyes. Therefore, the absorption bands of the dyes 1a and remained
5
unchanged, but the absorption peaks of DPBF decreased gradually (Figure 5 and
Figures S9-S11). The plots of the time-dependent DPBF consumption in the presence
of the photoexcited dyes 1a and
5 in ethanol and 1,4-dioxane solutions are shown in
Figure 6. The relative rates of DPBF consumption, obtained from the slopes of the
1
plots corresponded to the efficiency of O
2
generation by the dyes. It was seen that
1
both the dyes generated O
2
in ethanol and 1,4-dioxane significantly. The higher rate of
1
consumption of DPBF in 1,4-dioxane may be due to longer life of O
2
in this solvent.
1
Possibly, the reactivity of O
2
with both the dyes is much less in 1,4-dioxane.
1
Hence, when DABCO is added in dioxane, it merely quenches the unreacted O
2
without showing much effect on the photostabilities of the dyes, especially 5. But in
1
ethanol, the reactivity of O
2
with the dyes, especially 1a is more. Hence, in this
2
1

1
solvent, besides directly quenching O
2
, DABCO may also prevent its reaction with the
dyes to improve their photostabilities in ethanol. Overall, the CV and DPBF
photooxidation results could not explain the higher photostability of
may be explained as follows.
5 over 1a which
1.0
0.8
0.6
dye 1a/ ethanol
dye 5/ ethanol
dye 1a/ 1,4-dioxane
dye 5/ 1,4-dioxane
0
300
600
900
time (sec)
Figure 6. Time-dependent depletion of DPBF (50 µM) absorbance at λmax (412 nm) under
dye-sensitized photooxidation in the presence of dye 1a and 5 in ethanol and 1,4-dioxane.
3
.6 Pulse radiolysis studies
The triplet state studies of the BODIPY dyes are important as we can get
valuable information about the lasing characteristics and the photostabilities of the
1
dyes. The triplet states decide (i) the extent of O
2
generation by the dyes via energy
transfer to the dissolved O
efficiency due to triplet (T
2
1
in solvents, and also (ii) the reduction in the lasing
)–triplet (T ) absorption loss of the dye laser photons.
n
Presently, the triplet state spectroscopy and the energy transfer processes of the dyes
and were investigated by carrying out their pulse radiolyses in 1,4-
dioxane/benzene (9:1) solutions, using 4,4'-dimethylbiphenyl (DMBP) (E
1
a
5
T
= 65 kcal
-
1
mol ) as a triplet sensitizer. Radiolysis of benzene gives a high yield of the triplet state
2
2

-
1
(
G = 4.2 /100 eV) with energy of 82 kcal mol , but with low lifetime (
 = 3 ns).
Hence, DMBP was used for efficient generation of the dye triplets.
0.015
0.010
0.005
dye 1a
dye 5
6
00
700
(nm)
800

-4
-3
Figure 7. Triplet state absorption spectra of 1a and 5 (1  10 mol dm ) in 1,4-
-3
-3
dioxane/benzene (9:1) mixture in presence of DMBP (5  10 mol dm ).
The triplet state absorption spectra of the dyes 1a and
5 were similar, showing
broad absorption at 600-800 nm with λmax at 710 nm for 1a and at 720 nm for
5
(Figure
7
). The triplet state extinction coefficients (εTmax) at λmax, the rate constants for the
1
energy transfer from DMBP triplet to the dyes (ket) and from the dyes (1a and
5
) to O
2
3
(
k ( Dye

O
2
)) were calculated from the experimental data, and shown in Table 5.
) value for was ~3 fold less than that of 1a. This suggested that the
is less than that of 1a, contrary to our dye-sensitized
DPBF photooxidation results (previous section). Other factors like a higher singlet to
3
The k ( Dye

O
2
5
1O
generation capability of 5
2
triplet intersystem crossing rate (kisc) for dye
anomaly.
5 than dye 1a may be responsible for this
2
3

Table 5. Triplet-state data for the dyes 1a and 5 in 1,4-dioxane/benzene (9:1).
εTmax[a]
k
et [3DMBP→Dye]
k[ Dye →O
3
2
]
Dye
λ
max
9
-1 -1
8
-1 -1
-1
-1
[10 M S ]
[10 M S ]
[
nm]
[M cm ]
1
5
a
710
720
2440
2830
2.0
3.7
8.7
3.2
[
a] Extinction coefficients at the respective λmax.
1
3
.7 Relative reactivity of O
2
with dye 1a and 5
with dyes 1a and 5 in ethanol and 1,4-dioxane/ethanol
1
The reactivity of O
2
(1:4) were examined by measuring the decrease of their absorbances (at respective λabs)
1
on photo-irradiation in the presence of methylene blue (MB), a known efficient O
2
generator under visible light (tungsten lamp) illumination [45,46]. The experiment
could not be done in neat 1,4-dioxane because of the poor solubility of MB. Under
1
these conditions, only MB is excited to generate O
2
. Individual illumination (λ>590
nm) of ethanol solution of MB, 1a or
5
did not induce substantial changes in their
1
respective absorption spectra. This also confirmed that MB is stable towards O
2
. But
illumination of 1a or
5 and MB together in ethanol led to a time-dependent decrease in
the respective dye absorbances, although there was no change in the nature of the
absorption bands of the dye (Figure S12-S14).
As before, the relative rates of degradation of the dyes were obtained from the
1
slopes of the O. D. vs irradiation time plots (Figure 8). The reactivity of O
2
with dye
5
2
4

was less than that with dye 1a in both the media. Typically, a 90 min illumination led
to 50% and 37% degradation of 1a and respectively in pure ethanol. While the
degradations of 1a and in 1,4-dioxane/ethanol (1:4) were 47% and 35% respectively.
5
5
These results confirmed that dye
5
is more photostable than dye 1a due to its lower
1
1
reactivity with O
2
and the activity of O
2
with these dyes is decreased on lowering the
in
solvent polarity. Both these explain the exceptionally high photostability of dye
,4-dioxane.
5
1
dye 1a /ethanol
dye 1a /(1:4) 1,4- dioxane:ethanol
dye 5 /ethanol
1.0
0.8
0.6
dye 5 /(1:4) 1,4- dioxane:ethanol
0
30
60
90
time (min)
Figure 8. Time-dependent changes in the ODs of the dyes at their respective max during
photo-excitation in presence of MB in ethanol and 1,4-dioxane-ethanol (1:4).
3
.8 Theoretical interpretations
To further explain the reactivity of O
1
2
with the dyes in solvents of different
1
polarities, we carried out the theoretical calculations of the reaction of O
2
with the dye
5
. Earlier we characterized the photodegradation products of the BODIPY dyes to
confirm that the C7'-C8 double bond of the BODIPY core is the most vulnerable
2
5

1
1
position for 1,2-cycloaddition of O
2
[11]. Hence the reaction course of dye
5
with O ,
2
1
was theoretically calculated by placing the O
2
in proximity to the C7'-C8 double bond
of
5 (Figure 9, 5a). The transition state (TS) and the product structures were assigned
as 5b and the peroxide 5c respectively (Figure 9).

1
1O
1
O
2
O
O
2
2
O
O
8
8
8
7
'
7'
7'
N
N
N
N
N
N
B
B
B
OH HO
OH HO
OH HO
5a
5b
5c
1
Figure 9. Mechanism of the reaction of the BODIPY dye 5 with O .
2
5
a
5b
5c
Figure 10. Energy-minimized ground state structures (5a) of the dye 5, and the structures of
1
the transition state (5b) and product (5c) of the reaction between 5 and O
2
.
The ground-state (S ) minimum-energy structure 5a of the BODIPY dye 5 was
0
optimized by DFT calculations (Figure 10). The calculated bond lengths and charge density
values of 5a, 5b and 5c are presented in Table 6. The TS 5b is formed by partial cleavage of
the C7'-C8 bond and formation of the C7'O2 bond as the initial events, as supported by
the calculated bond lengths (Table 6). The C7'−C8 (1.41 Å) and O1−O2 (1.21 Å) bond
2
6

lengths in 5a increased to 1.52 Å and 1.41 Å, respectively, in the TS 5b. The distance
between C7' and O2 (1.50 Å) indicated that a single bond is formed between them. The C8
and O1 atoms were too far apart (2.71 Å) to form a stable bond, but gradually moved closer
for the bond formation (bond length = 1.51 Å), furnishing the peroxide 5c. The charge density
calculations further established the mechanism. Consistent with the proposed mechanism
(Figure 9), the calculated electron density at O1 is more than that at O2 in 5b. After the
formation of the C8─O1 bond, O1 and O2 become electronically similar, as is evident from
the charge densities on the two atoms (Table 6). However, most importantly, the calculated
charge density values suggested that the charge separation in the TS 5b is more than that in
the dye 5a and the peroxide 5c. Thus, the energies of the initial and final products will not
change much with the change in solvent polarity, but the energy of the TS 5b will be very
high in non-polar 1,4-dioxane. The polar solvent, ethanol will stabilize the TS 5b. This will
make the activation energy of the reaction higher in 1,4-dioxane, compared to that in ethanol.
1
Thus, the reaction rate of O
2
with the dye will be much lower in 1,4-dioxane than in ethanol.
This explains the higher photostabilities of the dyes in 1,4-dioxane. This is consistent with our
1
previous theoretical calculation that revealed lower Eact for the reaction between O
2
and 1a
than that with its another B-subtituted analog [12].
Table 6. The DFT-calculated bond lengths and charge density values of 5a, 5b and 5c.
Bond Length (Å)
Dyes C8-O1 C7’-O2
Atomic charge (a.u.)
O1 O2
C7’-C8
O1-O2
5
5
5
a
-
-
1.41
1.52
1.54
1.21
1.41
1.49
0.000
-0.411
-0.298
0.000
-0.254
-0.289
b
c
2.71
1.51
1.50
1.45
2
7

4
. Conclusions
In short, a new BODIPY dye
B-centre was synthesized. The dye
the commercial dye 1a but was photo-chemically more stable than 1a in both polar
5
containing two propargyl alcohol moieties at the
5
showed comparable lasing efficiency as that of
1
(
ethanol) and non-polar (1,4-dioxane) solvents. Both the dyes 1a and
5
had similar O
2
generation capacities, and the oxidation potential (Eox) of
5
was lower than that of 1a
.
Despite these, dye
1O
5 is more photostable than dye 1a due to its lower reactivity with
2
, particularly in the non-polar solvent 1,4-dioxane, as reflected by our MB-induced
photo-degradation results, and rationalized by theoretical calculations. Earlier, we
found that the B-substituted dye 2a was more photostable than the commercial dye 1a
The key finding of the present investigation is that we showed for the first time that
besides alkyl substitution at the B-atom, use of less polar 1,4-dioxane medium in
combination, can increase the photostability of the BODIPY dye immensely without
.
compromising its laser property. The excellent photostability of the dye
dioxane, comparable to that of Rhodamine 6G in ethanol, makes it a potential photo-
stable new BODIPY dye for further exploration. Also, the new dye may be an
efficient photosensitizer for PDT applications due to its high singlet oxygen generation
capacity. Moreover, since the dye shows higher photostability in less polar
5 in 1,4-
5
5
environment, and is highly fluorescent, it may also be useful for fluorescence
microscopy in living cells.
Acknowledgment
M. G. acknowledges the Homi Bhabha National Institute (HBNI) for the award
of research fellowship.
2
8

References
A. Loudet, K. Burgess, BODIPY dyes and their derivatives:ꢀ syntheses and
1
2
3
4
5
6
spectroscopic properties, Chem. Rev. 107 (2007) 4891−4932.
G. Ulrich, R. Ziessel, A. Harriman, The chemistry of fluorescent bodipy dyes:
versatility unsurpassed, Angew. Chem. Int. Ed. 47 (2008) 1184–1201.
N. Boens, V. Leen, W. Dehaen, Fluorescent indicators based on BODIPY, Chem.
Soc. Rev. 41 (2012) 1130−1172.
A. Kamkaew, S.H. Lim, H.B. Lee, L.V. Kiew, L.Y. Chung, K. Burgess, BODIPY
dyes in photodynamic therapy, Chem. Soc. Rev. 42 (2013) 77−88.
H. Lu, J. Mack, Y. Yang, Z. Shen, Structural modification strategies for the rational
design of red/NIR region BODIPYs, Chem. Soc. Rev. 43 (2014) 4778−4823.
A. Bessette, G.S. Hanan, Design, synthesis and photophysical studies of
dipyrromethene-based materials: insights into their applications in organic
photovoltaic devices, Chem. Soc. Rev. 43
T.G. Pavlopoulos, J.H. Boyer, K. Thangaraj, G. Sathyamoorthi, M.P. Shah, M.L.
Soong, Laser dye spectroscopy of some pyrromethene-BF complexes, Appl. Opt.
1 (1992) 7089−7094.
(2014) 3342−3405.
7
8
2
3
S.C. Guggenheimer, J.H. Boyer, K. Thangaraj, M.P. Shah, M.L. Soong, T.G.
Pavlopoulos, Efficient laser action from two cw laser-pumped pyrromethene-BF
complexes, Appl. Opt. 32 (1993) 3942−3943.
2
2
9