Photophysical and Photoacoustic Properties of Quadrupolar Borondifluoride Curcuminoid Dyes

Article abstract of DOI:10.1002/chem.201704423

The synthesis and characterization of a series of donor-π-acceptor-π-donor (D-A-D) curcuminoid molecules is presented herein that incorporates π-extended aryl and electron-donating amino terminal functionalization. Computational evaluation shows these molecules possess quadrupolar character with the lowest energy transitions displaying high molar extinction coefficients with broad tunability through manipulation of terminal donating groups. Consistent with their quadrupolar nature, these molecules show varying degrees of solvatochromic behavior in both their absorption and emission spectra, which has been analyzed by Lippert–Mataga and Kamlet–Taft analysis. Photophysical and photoacoustic (PA) properties of these molecules have been investigated by the optical photoacoustic z-scan (OPAZ) method. Selected curcuminoid molecules display nonlinear behavior at a high laser fluence through excited state absorption that translates to the production of an enhanced photoacoustic emission. A relative comparison of “molar PA emission” is also presented with the crystal violet linear optical absorbing/linear PA emitting system being utilized as a standard reference material for OPAZ experiments. Furthermore, PA tomography experiments are presented to illustrate the enhanced PA contrast obtainable via an excited state absorption.

Full text of DOI:10.1002/chem.201704423

A Journal of
Accepted Article
Title: The Photophysical and Photoacoustic Properties of Quadrupolar
Borondifluoride Curcuminoid Dyes
Authors: Stephanie Bellinger, Maryam Hatamimoslehabadi, Seema
Bag, Farha Mithila, Jeffrey La, Mathieu Frennete, Samir
Laoui, David Szalda, Chandra Yelleswarapu, and Jonathan
Rochford
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: Chem. Eur. J. 10.1002/chem.201704423
Link to VoR: http://dx.doi.org/10.1002/chem.201704423
Supported by

10.1002/chem.201704423
Chemistry - A European Journal
FULL PAPER
The Photophysical and Photoacoustic Properties of Quadrupolar
Borondifluoride Curcuminoid Dyes
Stephanie Bellinger, [a] Maryam Hatamimoslehabadi, [b] Seema Bag, [a] Farha Mathila, [a] Jeffrey La, [b]
Mathieu Frenette, [a] Samir Laoui, [b] David J. Szalda, [c] Chandra Yelleswarapu, [b]* Jonathan
Rochford[a]*
The synthesis and characterization of a series of donor--
acceptor--donor (D-A-D) curcuminoid molecules are here
presented incorporating -extended aryl and electron-donating
amino terminal functionalization. Computational evaluation shows
these molecules possess quadrupolar character with the lowest
energy transitions displaying high molar extinction coefficients
with broad tunability through manipulation of terminal donating
groups. Consistent with their quadrupolar nature, these molecules
show varying degrees of solvatochromic behavior in both their
absorption and emission spectra which has been analyzed by
Lippert-Mataga and Kamlet-Taft analysis. Photophysical and
photoacoustic (PA) properties of these molecules have been
investigated by the optical photoacoustic z-scan (OPAZ) method,
which found select curcuminoid molecules display nonlinear
behavior at a high laser fluence through excited state absorption
that translates to the production of an enhanced photoacoustic
emission. A relative comparison of ‘molar PA emission’ has been
presented with the crystal violet linear optical absorbing/linear PA
emitting (LA) system being utilized as a standard reference
material for OPAZ experiments. Furthermore, PA tomography
experiments are also presented to illustrate the enhanced PA
contrast obtainable via an excited state absorption.
and herbal remedy due to its antioxidant, anti-inflammatory, and
anti-tumor properties.[1] In addition to its medicinal applications,
the characteristic photophysical properties of the curcumin
molecule have made it an attractive protagonist for a range of
optical and medical applications such as nonlinear refraction,[2]
dye-sensitized solar cells,[3] two-photon absorption imaging,[4]
two-photon fluorescence,[5] photothermal therapy,[6] anti-cancer
agent,[7] two-photon fluorescence imaging and photodynamic
therapy.[8] Our interest in studying the curcuminoid class of dyes
is motivated by the emerging techniques of photoacoustic (PA)
imaging[9] and microscopy.[10] The PA effect, first demonstrated by
Alexander Graham Bell in 1880, is manifest by the photoinduced
propagation of an acoustic wave upon the thermoelastic
expansion of the absorbing material in its surrounding medium.[11]
A major advantage of PA imaging over alternative imaging
modalities is its capability of providing high contrast images at
enhanced penetration depths in living tissue (~5 cm). Thus PA
(light-in/acoustic-out) imaging provides the best of both worlds;
the high contrast of fluorescence (light-in/light-out) imaging with
the depth penetration of ultrasound (acoustic-in/acoustic-out)
imaging.[12] The prospective for high contrast deep tissue imaging
has inspired us to research the effective production of a PA signal
by molecular photoacoustic contrast agents (MPACs). Small
molecules are an attractive option as contrast agents as they offer
NIR tunable absorption properties and desirable size relative to
currently used nanomaterials.[13] Large nanomaterials (100 – 200
nm) are required for in vivo imaging and leveraging the enhanced
permeability and retention (EPR) effect.[14] But large
nanomaterials have undesirable pharmacokinetics as they cannot
be cleared by the kidneys because the physiological pore size of
the glomerular capillary wall in the kidney has a cutoff of 5 nm. [15]
Instead, they are taken by the mononuclear phagocyte system
and induce tissue toxicity in liver and spleen.[16] On the other hand
small molecule dyes have good tissue penetration, biodistribution,
fast clearance and low toxicity.[17] Molecular requirements for the
production of a photoacoustic signal are the strong absorption of
light as quantified by the molar extinction coefficient () and
efficient excited state dissipation of energy through non-radiative
pathways. Recently it has been shown that MPACs displaying
nonlinear absorption are capable of producing amplified
photoacoustic signals relative to their linear absorbing
counterparts.[18] Quadrupolar dyes have been shown to have
large optical cross sections and the potential for reverse saturable
absorption, two important characteristics in developing MPACs.
With this in mind, a series of dyes, inspired by the naturally
occurring curcumin chromophore, have been synthesized through
the incorporation of -extended aryl and tertiary arylamine
Introduction
Curcumin is a naturally occurring dye found in the turmeric root
(curcuma longa) which has been used for centuries as a spice
[a]
[b]
[c]
Dr. S. Bellinger,, Dr. S. Bagg, F. Mathila, Dr. M. Frenette, Prof. Dr. J.
Rochford
Department of Chemistry
University of Massachusetts Boston
100 Morrissey Boulevard, Boston, Massachusetts 02125
E-mail: Jonathan.Rochford@umb.edu
Dr. M. Hatamimoslehabadi, J. La, Dr. S. Laoui, Prof. Dr. C.
Yelleswarapu
Department of Physics
University of Massachusetts Boston
100 Morrissey Boulevard, Boston, Massachusetts 02125
Email: Chandra.Yelleswarapu@umb.edu
Dr. D. Szalda
Department of Natural Science
Baruch College, CUNY
New York, New York 1001
Supporting information for this article is given via a link at the end of
the document.((Please delete this text if not appropriate))
This article is protected by copyright. All rights reserved.

10.1002/chem.201704423
Chemistry - A European Journal
FULL PAPER
electron donating substituents at the terminal position of the -
diketonate borondifluoride backbone (Figure 1).
Figure 2. X-ray crystallographic ortep diagram of 5∙CHCl3. The solvated CHCl3
structure has been removed for clarity.
Crystal structure analysis has previously been reported for the
free acid and copper bound bis-anthracenyl curcuminoid
analogues allowing for comparison.[20] The O-B-O bond angle for
5 of 110.4(4)° is significantly larger than that of the Cu compound
(93.25°) consistent with the smaller atomic radius of the
coordinating boron. The terminal anthracenyl groups in 5 are
twisted out of plane of the central ring in a symmetric manner by
35.5(1)° and 33.9(1)°, making the -system of 5 more planar and
delocalized with respect to the copper complex (dihedral =
45.84°) . The B(11)-O(11) bond length is 1.481(5) Å and B(11)-
O(16) is 1.492(6) Å representing a decrease relative to the
reported copper complex and is attributed to the increased
electronegativity of the boron difluoride moiety.
Figure 1. Structures of -extended (2-6) and electron donating (7-10) series of
curcuminoid molecules as well as the reference curcuminBF2 (1) molecule.
Electronic Absorption Properties and TDDFT Calculations
The predominate form of naturally occurring curcumin, (1E,6E)-
1,7-bis(4-hydroxy-3-methoxyphenyl)-1,6-heptadiene-3,5-dione,
is a -diketone that exists in both its keto and enol tautomeric
forms in solution. Coordination of the BF2 moiety with the
deprotonated diketonate anion provides conformational stability
as well as an electron withdrawing effect that results in a
bathochromic shift of its absorption maximum from 420 to 495 nm
and a two-fold increase in its molar extinction coefficient () from
4.28 x 104 M-1 cm-1 to 8.85 x 104 M-1 cm-1 in toluene for the free
acid curcumin and curcuminBF2 derivative, respectively (Figure
S1).[8b] The UV-Vis electronic absorption spectra recorded in
toluene (dielectric constant ϵ = 2.38) at room temperature are
presented relative to the naturally derived curcuminBF2 (1)
reference molecule in Figure 3 for the aryl -extended (2-6) and
amino electron donating (7-10) series. Incorporation of aryl
substituents of increasing conjugation results in a bathochromic
shift (phenyl < biphenyl < naphthalenyl < anthracenyl  pyrenyl)
ranging from 449 nm for 2 to 574 nm for 6. This is consistent with
a destabilization of the HOMO level combined with a slight
stabilization of the LUMO level due to the expansion of the
contributing  and * orbitals, respectively, as confirmed by
voltammetry and computational analysis (vide infra). Molar
extinction coefficients are consistent across the aryl series 2 – 6
relative to 1 with the bis-biphenyl curcuminBF2 dye 3 displaying
the strongest light absorption of this series with  = 1.09 x 105 M-
1 cm-1 at 450 nm. The bis-anthracenyl curcuminBF2 dye 5 exhibits
the most red shifted absorption maximum of the aryl series with a
Results and Discussion
Synthesis and Characterization
Two
series
of
quadrupolar
donor--acceptor--donor
difluoroboron curcuminoid dyes have been synthesized in order
to establish a structure-property relationship with respect to both
their photophysical and photoacoustic properties. Extended aryl
-conjugated and tertiary amine donating substituents have been
introduced at the terminal positions of the difluoroborn -
diketonate backbone following a previously established synthetic
protocol.[19]
Briefly,
the
intermediate
acetylacetone
borondifluoride intermediate was first isolated in bulk (~1 g) from
acetylacetone and boron trifluoride diethyl etherate. Selected
aldehydes were then coupled at the terminal methyl positions of
the acetylacetone boron difluoride intermediate through
a
modified Knoevenagel condensation reaction in refluxing toluene
to yield compounds 1-10 (Figure 1). In all cases crude product
precipitated from the reaction mixture upon cooling to room
temperature. Analytically pure solids were subsequently obtained
in respectable yield upon recrystiallization from dichloromethane
1
and hexane. Compound structure and purity were verified by H
NMR and 13C NMR (when possible due to solubility) spectroscopy
as well as MALDI-TOF high resolution mass spectroscopy. Single
crystals suitable for x-ray diffractometry were isolated in the case
of the bis-anthracenyl curcuminBF2 compound 5 (Figure 2).

max of 577 nm of 7.94 x 104 M-1 cm-1 molar extinction coefficient.
This article is protected by copyright. All rights reserved.

10.1002/chem.201704423
Chemistry - A European Journal
FULL PAPER
nm albeit with a slight reduction in at 1.17 x 105 M-1 cm-1 relative
to 7 and 9.
1
2
3
4
5
6
10
To better understand the origin of electronic transitions in their
UV-Vis electronic absorption spectra single-point energy time-
dependent density functional theory (TDDFT) calculations were
carried out following prior DFT optimization and frequency
analysis of all compounds 1-10 (Figure 4). A full collection of
TDDFT spectra is provided in the supporting information (Figures
S3 – S12). Suffice to say the -extended aryl series 2 - 6 show
satisfactory quantitative agreement with their experimental UV-
Vis spectra by TDDFT using the B3LYP/6-311g(d,p)
functional/basis-set combination. However, the electron rich
compounds 1 and the tertiary amine electron donating series 7 –
10 only exhibited satisfactory agreement with experiment when
using the CAM-B3LYP/6-311g(d,p) functional/basis set
combination with the added cis(d) keyword to incorporate
configurational interaction of ground and excited electronic states
in the TDDFT calculation. This is most likely due to the greater
extent of charge-transfer character involved in their lowest energy
 transition due to contribution from the N-donor lone electron
5
0
400
500
600
700
Wavelength (nm)
20
15
10
5
1
7
8
9
10
-2
-3
-4
0
400
500
600
700
2.41 eV
( f = 2.22)
2.84 eV
( f = 1.44)
3.24 eV
( f = 1.83)
2.39 eV
( f = 0.97)
2.96 eV
( f = 2.36)
Wavelength (nm)
-5
-6
Figure 3. UV-Vis absorption spectra for 1 – 6 (top) and 1, 7 – 10 (bottom) plotted
as wavelength versus molar extinction coefficient. Note the larger y-axis scale
for the amino curcumin series of dyes.
2
3
4
5
6
Also of note is that the increase in conjugation associated with
progression from the anthracenyl (5) to pyrenyl (6, max = 574 nm)
substituents does not result in a significant spectral shift, however,
it does provide for an increased light harvesting capacity with 6
exhibiting a molar extinction coefficient of 8.91 x 104 M-1.
Introduction of electron donating tertiary amine substituents at the
terminal ends of the -conjugated bis-vinyl acacBF2 backbone in
compounds 6 - 10 results in a strong bathochromic shift of
absorption maxima relative to 1. This is consistent with a
significant destabilization of the HOMO level due to the strong
donating influence of the tertiary amine substituents (relative to
the more electronegative O-donors of 1), again confirmed by
voltammetry and computational analysis (vide infra). Most
impressive is how increasing electron donor strength results in a
significant increase in the lowest energy absorption band molar
extinction coefficient. For example, relative to the basic
curcuminoid structure of 2, having simple phenyl substituents with
a maximum  of 7.38 x 104 M-1 cm-1 at 449 nm, the bis-p-
dimethylaminophenyl cucuminoid 7 displays an  of 1.63 x 105 M-
1 cm-1 at 572 nm and the bis-carbazole curcuminoid 9 displays an
 of 1.73 x 105 M-1 cm-1 at 540 nm, all recorded in toluene (Figure
3). Compound 8 with terminal triphenylamine electron donating
groups exhibits the most red shifted absorption maximum at 578
-2
-3
-4
-5
-6
1.84 eV
( f = 1.90)
1.95 eV
( f = 1.91)
1.78 eV
( f = 1.72)
2.07 eV
( f = 1.78)
2.24 eV
( f = 1.65)
1
7
8
9
10
Figure 4. Frontier molecular orbital energy levels (eV), HOMO-LUMO surfaces
(isofactor = 0.04), and oscillator strengths (f ) as calculated by DFT and TDDFT
analysis. Electron occupancy is removed for clarity so HOMO and LUMO levels
are highlighted in blue and red, respectively. DFT and TDDFT analysis of the
aryl series 2 – 6 (top) was performed with the B3LYP/6-311g(d,p)/toluene PCM
functional/basis set/solvent model. DFT analysis of 1 and the tertiary amine
series 7 – 10 (bottom) was performed with the CAM-B3LYP/6-311g(d,p)/toluene
PCM functional/basis set/solvent model with the configurational interaction
keyword cis(d) added for TDDFT analyses.
This article is protected by copyright. All rights reserved.

10.1002/chem.201704423
Chemistry - A European Journal
FULL PAPER
pairs. Regardless of the quantitative agreement between
experiment and theory however, all levels of theory employed
were consistent with the qualitative nature of the electronic
transitions observed. The lowest energy absorption band for all
curcuminoids (1-10) is consistently assigned to a single S0S1
electronic transition of high oscillator strength (f ) primarily derived
from a HOMOLUMO electronic transition. For the aryl series 2
- 6 DFT geometry optimization and frequency analysis (B3LYP/6-
311g(d,p)/toluene PCM) show a consistently planar, i.e. complete
-conjugation, configuration in all ground state optimized
structures for 1, 3, 4 and 6. Compounds 2 and 5 are exceptions
due to the steric constraints of their terminal aryl substituents. In
2, the terminal phenyl rings of the biphenyl substituents are
predicted to rotate out of plane from the planar
phenylcurcuminBF2 core unit with a dihedral angle of 35.5, while
in 5 the bulky anthracenyl substituents are predicted to rotate out
of plane by 38.2 and 38.9 from the bis-vinyl acacBF2 core. The
predicted non-planar DFT optimized ground state configuration of
5 is consistent within 5° of the experimentally determined value
from crystal structure analysis (Figure 2), which exhibits both
terminal anthracenyl groups out of plane with dihedral angles of
35.51 and 33.86. The non-planar configuration of 5 results in a
reduction of symmetry relative to other molecules in the series,
which is reflected in the decreased  and calculated oscillator
strength for 5 (f = 0.80) vs. 6 (f = 1.82). Optimized ground state
calculations for compounds bearing the amino substituents, 7-10,
show consistently planar configurations extending to the tertiary
amine. Having two terminal, electron donating, tertiary amine
substituents in their conjugated -systems, the theoretical HOMO
energies of 7-10 are destabilized relative to 1 due to the lower
electronegativity of the N vs O donor atoms. In contrast, the
LUMO energies of dyes 7-10 are less influenced from the tertiary
amine substituents. This is consistent with the observed red shift
in their HOMOLUMO transition based absorption maximum and
the donor-acceptor nature of the lowest energy transition. The
calculated electronic S0S1 transition for all molecules is
assigned primarily to the HOMOLUMO transition (97-100%
contribution). A charge-transfer shift in calculated electron density
is observed for this electronic transition from delocalization across
the entire -backbone of the molecule in the HOMO to a more
compacted * orbital involving the bis-vinyl acacBF2 in the LUMO
with only partial contribution from the terminal substituents; hence
their quadrupolar donor--acceptor--donor description. It is also
noteworthy that this electronic transition shows the presence of a
nodal plane in the LUMO level across the vinyl arms of the
molecule that would allow for a twisted intramolecular charge-
transfer (TICT) excited state configuration.
1
2
4
5
7
8
9
10
50 A
1.5
1.0
0.5
0.0
-0.5
-1.0
-1.5
-2.0
Potential (V) vs. Fc/Fc+
Figure 5. An overlay of square wave voltammetry and cyclic voltammetry
(greyscale) for a selection of borondifluoride curcuminoid dyes. Experimental
conditions: 0.1 M Bu4NPF6 acetonitrile electrolyte, 3 mm diameter glassy carbon
working electrode, Pt wire counter electrode.
For the aryl series, 2 - 6, deviations in their electrochemical
properties are most evident in their oxidation potentials which is
consistent with significant electron density of their HOMO levels
located at the terminal aromatic substituents. With the terminal
substituents in 2 – 6 being planar aromatic hydrocarbon ring
systems all oxidation potentials are irreversible and relatively
positive (+1.2 to +1.5 V vs Fc+/0) with the exception of the bis-
anthracenyl system 5, possibly due to steric induced distortion
from planarity as discussed above. Consistent with DFT
predictions little change is observed between LUMO reduction
potentials centered at the core of the molecule. The tertiary amine
series show significant shifts at both their HOMO oxidation and
LUMO reduction potentials. This is indicative that the electron
donating strength of the substituent is not only destabilizing the
HOMO level but influencing the LUMO level, albeit to a lesser
degree. Consistent with DFT predicted electron donating
strengths, both 7 and 10 show the most cathodic HOMO oxidation
potentials at +0.38 V and +0.28 V vs. Fc+/0, respectively.
Accordingly, 7 and 10 also show the most cathodic LUMO
reduction potentials at -1.27 V and -1.49 V vs. Fc+/0, respectively.
However with destabilization of the HOMO level being dominant
they each show narrow electrochemical band-gaps (Ebg) of 1.65
V and 1.77 V, respectively. HOMO-LUMO bandgaps have been
determined experimentally through electrochemical (Ebg) and
optical (E0-0) investigations as well as theoretically through DFT
calculations (EDFT) for comparison. Ebg and E0-0 are expected to
be correlated in characterizing the HOMO-LUMO energy gap,
however the differing electronic states and molecular
environmental factors (e.g. coulombic charge and electrolyte)
between the two techniques preclude exact agreement. The
tabulated electrochemical data is shown in Table 1 including
spectroscopically and computationally determined bandgaps for
Electrochemical Properties
The electrochemical properties of compounds 1 - 10 were
determined via cyclic and square wave voltammetry in 0.1 M
Bu4NPF6 acetonitrile electrolyte using a glassy carbon working
electrode and are reported versus the ferricenium/ferrocene
(Fc+/0) pseudo reference potential (Figure 5).
This article is protected by copyright. All rights reserved.

10.1002/chem.201704423
Chemistry - A European Journal
FULL PAPER
comparison. All three methods show good agreement in trend
with an exception being the electrochemically determined value
for 6. The pyrenyl substituent of 6 is known to undergo -stacking
aggregation at elevated concentration which is likely to have an
influence on its observed electrochemical potentials recorded at
an electrode surface relative to dilute optical solution.[21]
substituents show enhanced fluorescence quantum yields of 0.77
and 0.76 relative to the remaining series, which is attributed to
1
2
1.0
3
4
0.8
0.6
0.4
0.2
0.0
5
6
Table 1. Electrochemical potentials for compounds 1 - 10 reported vs. the
ferricenium/ferrocene redox in 0.1
M Bu4NPF6 acetonitrile electrolyte.
Spectroscopic and computational bandgaps are included for comparison.
Eox (V)
Ered (V)
Ebg b (eV)
E0-0 (eV) c
EDFT (eV) d
1
2
3
4
5
6
7
8
9
10
0.76a
1.49a
1.45a
1.26a
0.89a
1.40a
0.38a
0.56
-1.15a
-0.98a
-1.03a
-0.95a
-0.92a
-1.19a
-1.27a
-1.13a
-1.19a
-1.49a
1.91
2.47
2.48
2.21
1.81
2.59
1.65
1.69
1.85
1.77
2.44
2.73
2.52
2.42
2.07
2.07
2.09
2.06
2.24
2.31
2.24
3.24
2.96
2.84
2.39
2.41
1.94
1.78
2.07
1.84
400
500
600
700
800
Wavelength (nm)
1
7
8
9
1.0
10
0.8
0.6
0.4
0.2
0.0
0.66
0.28a
400
500
600
700
800
Wavelength (nm)
a irreversible b electrochemical bandgap b optical bandgap determined in room
temperature toluene c theoretical bandgap
Figure 6. Fluorescence emission spectra of 1 – 6 (top) and 1, 7 – 10 (bottom)
recorded in toluene at room temperature.
Fluorescence Emission Properties
enhanced radiative rate constants for both compounds at 5.80 x
108 s-1 and 4.49 x 108 s-1, respectively. On the other hand, the aryl
series exhibit a lower range of fluorescence quantum yields with
a minimum of 0.23 exhibited by 6 attributed to its large
nonradiative rate constant of 7.54 x 108 s-1 (Table 2).
All curcuminoids 1 - 10 are strongly emissive in toluene with
radiative S1S0 fluorescence quantum yields (fl) ranging from
0.23 for 6 to 0.77 for 7. Fluorescence lifetimes and quantum yields
were insensitive to dissolved oxygen confirming the spin-allowed
S1S0 transition multiplicity. Similar to observed trends in their
ground state S0S1 electronic absorption properties,
a
Solvatochromic Studies
bathochromic shift in singlet emission maxima (em) is observed
with increased conjugation and donating capacity of the terminal
substituents. Deviation from this trend is observed for 5, where
the emission is more red shifted and with less defined vibrational
structure (fwhm = 1,979 cm-1) in comparison to 6 (fwhm = 2,368
cm-1). This observation can possibly be attributed to a decent in
symmetry in the structure of 5 vs. 6 due to the steric bulk of the
terminal anthracenyl groups as confirmed by x-ray analysis.
Emission spectra of the amine series (7 - 10) exhibit narrower
fwhm and exhibit maxima over a narrower range than the aryl
substituted series. A collection of all emission spectra recorded in
toluene is provided in figure 6. Time-resolved fluorescence
measurements in toluene show singlet emission lifetimes (1)
ranging between 0.93 – 2.00 ns across both series of compounds.
Interestingly compounds 7 and 9 both bearing tertiary amine
A fully comprehensive study of the photophysical properties of 1
- 10 has been conducted in solvents of ranging dielectric
constants (Table S1); toluene (ϵ = 2.38), dichloromethane (ϵ =
8.93), methanol (ϵ = 32.70), and acetonitrile (ϵ = 37.50). The
twisted intramolecular charge-transfer (TICT) nature of the S0S1
electronic transition makes these quadrupolar curcuminoid dyes
particularly susceptible to solvatochromic effects. Increasing
polarity of the solvent medium stabilizes the charge distribution in
both the ground and excited states, however, a much larger
influence is observed on the thermally equilibrated excited (thexi)
states of 1 - 10 due to the rate of solvent reorganization (ps)
relative to 1(ns). A significant decrease of fl is observed for all
compounds in solvents of greater polarity with the exception being
the bis-biphenyl compound 3, which exhibits a slight increase in
This article is protected by copyright. All rights reserved.

10.1002/chem.201704423
Chemistry - A European Journal
FULL PAPER
Table 2. UV-Vis absorption and emission data recorded in toluene including fluorescence lifetimes, radiative and non-radiative rate constants.
λabs
(nm)

λem
(nm)

(ns)
kr
knr
(108 s−1
)
Stokes shift
(cm-1)
L-M slope a
(R2)

fl
(104 M-1 cm-1)
(108 s−1
)
1
2
3
4
5
6
7
8
9
10
495
449
480
479
577
574
572
578
540
517
8.85
522
465
506
532
644
616
612
628
570
555
0.42
0.29
0.32
0.53
0.28
0.23
0.77
0.49
0.76
0.33
1.36
1.15
0.93
1.69
1.82
1.03
1.32
1.73
1.69
2.00
3.07
2.50
3.45
3.12
1.56
2.22
5.80
2.85
4.49
1.63
4.29
6.21
7.28
2.79
3.95
7.54
1.75
2.92
1.43
3.36
1,077
720
7,538 (0.7865)
1,762 (0.9997)
6,974 (0.3417)
6,236 (0.9421)
11,295 (0.9775)
5,451 (0.8156)
4,909 (0.9946)
9,646 (0.9960)
6,394 (0.9889)
3,344 (0.9905)
7.38
10.48
8.13
1,109
1,576
2,353
1,214
1,188
1,377
944
7.94
8.91
16.30
11.71
17.30
13.95
1,292
a The Lippert-Mataga slope (L-M slope) and R2 were calculated via linear fit regression of experimentally obtained Stokes shifts in toluene, dichloromethane, and
acetonitrile.
fl from 0.32 to 0.37 on going from toluene to acetonitrile. For the
majority of dyes, a bathochromic shift of emission spectra upon
increasing solvent dielectric is accompanied by a significant
reduction in fl associated with a large increase in the
nonradiative rate constant (knr). This is most likely due to
stabilization of a TICT excited state associated with molecular
rotation at the nodal plane of the vinyl bond. Indeed preliminary
molecular rotor studies on 1 have confirmed this hypothesis
(Figure S2).
Protic solvents such as methanol are best evaluated separate
from non-protic solvents, as such the experimentally determined
 in methanol has been plotted but has not been included in the
linear fit calculation to derive the Lippert-Mataga slope. In fact,
previous literature has reported on the degradation of
borondifluoride--diketonate structures in protic solvents.[24]
A
similar degradation was here observed that prevented accurate
quantitative analysis in methanol of , fl, or photoacoustic studies
in this solvent. However, reliable qualitative spectra were easily
obtained allowing abs and em to be accurately determined.
Complimentary to Lippert-Mataga analysis, Kamlet-Taft analysis
was also conducted allowing for a more in-depth evaluation of
specific solute-solvent interactions. Details of the Kamlet-Taft
analysis are provided in the supporting information. Briefly, it can
be concluded that solvatochromic effects on emission are
primarily impacted by solvent polarity and polarization (*) and
hydrogen bond acceptor basicity), although to varying degrees
throughout the series (Table S2). Susceptibility to  is consistent
with deviation from linear behavior observed for methanol
(=0.66) in the Lippert-Mataga evaluation as well as a previously
proposed degradation pathway.[24c]
This excited state solvatochromism has been evaluated using the
Lippert-Mataga equation which provides
a framework for
discerning the difference between ground and excited state dipole
moments (e-g) from the Stokes shift dependence upon solvent
polarizability.[22] Details of the Lippert-Mataga analysis is provided
in the supplementary information. All Lippert-Mitaga slopes
(Figure S13, Table 2) are positive consistent with a positive
solvatochromism. The data shows a good linearity for most
compounds (R2 > 0.8) except for 1 and 3; the former due to non-
negligible solvent interactions (likely H-bonding) with the latter
deviating most from linear behavior indicating
a lesser
susceptibility to different solvent polarities (f). Radiative (kr) and
non-radiative (kn ) rate constants for 1 - 10 (Table 2) are within
Photoacoustic Properties
r
one order of magnitude from each other throughout the series in
varying polarity environments, consistent with previously reported
values for difluorboron curcuminoid molecules.[23] Due to the
stabilizing effect on charge distribution within the dye molecules,
solvents with a large 훥푓 allow greater access to nonradiative
relaxation pathways and thus exhibit a decrease in fl as well as
fluorescence lifetime (1). Indeed a reduction in fl is observed as
a function of increasing polarizability (Table S1). Notably, knr
recorded in acetonitrile (ϵ = 37.50) is predominantly larger than in
toluene or dichloromethane as expected of a more stabilized
emitting state consistent with the energy gap law.
A photoacoustic signal is produced through the conversion of light
energy to sound energy by an absorbing material. The intensity of
a generated photoacoustic (PA) signal can be described by Eq. 1
PA = 휀g퐶gΓ I Φnr + 휀e퐶eΓ 퐼2 … + 휀n퐶n훤퐼n+1 (1)
where g is the ground state molar extinction coefficient at the
incident wavelength, Cg is the ground state concentration of dye
molecules,  is the Grüneisen coefficient (a constant that
quantifies a medium’s ability to conduct sound), I is the incident
photon fluence, nr is the quantum yield for nonradiative decay
This article is protected by copyright. All rights reserved.

10.1002/chem.201704423
Chemistry - A European Journal
FULL PAPER
(equivalent to 1-fl), e is the first excited state molar extinction
coefficient at the incident wavelength and Ce is the concentration
of excited state dye molecules.[25] Following Kasha’s rule[18], nr
should equal unity for any excited state relaxation process. Thus,
nr is only relevant for the initial linear term in Eq. 1. It is also
important to note that the concentration of excited state dye
molecules will vary with time and thus the laser pulse width,
relative to the lifetime of the excited state chromophore, is also an
0.55
0.50
0.45
0.40
0.35
0.30
0.25
0.20
Crystal violet
Cy3
1
4
5
6
7
8
9
10
important consideration when optimizing
a nonlinear PA
amplification. For molecules undergoing only linear absorption
processes, i.e. single photon absorption, the first linear term need
only be considered. As such, for straight forward linear absorption,
the PA emission varies linearly with the absorption cross section
of the sample (gCg), nr and the photon flux (I). For nonlinear
absorption processes, e.g. excited state absorption, the PA signal
enhancement is generated by a quadratic dependence on .[26]
There are three distinct types of PA emitters that can be
categorized based upon the correlation between their optical
absorption and PA emission characteristics as defined by Eq. 1:
(i) Linear absorber (LA)  linear PA emitter ( << laser); An
ultrashort excited state lifetime precludes nonlinear absorption.
Rapidly propagating S0↔S1 absorption-relaxation occurs within a
single laser pulse resulting in a linear dependence of the PA
signal amplitude on laser fluence. (ii) Saturable absorber (SA) 
weak PA emitter (e < g,  > laser). Saturable absorbers exhibit
excited state lifetimes longer than the laser pulse width. With
negligible excited state absorption, ground-state photobleaching
occurs with increasing photon transmission at increasing laser
fluence. (iii) Reverse saturable absorber (RSA)  nonlinear PA
emitter (e/g >1,  > laser). Non-zero absorption cross sections of
both the ground and excited states, as well as an excited state
lifetime longer than the pulse width, results in a nonlinear increase
in absorption and PA response at increasing laser intensity.
Ground-state photobleaching occurs but does not impede
excited-state absorption.
The optical photoacoustic z-scan (OPAZ) method is ideal for the
evaluation of fluence dependent nonlinear optical properties such
as those displayed by RSAs. A full description of the z-scan
apparatus has been previously reported.[18] Briefly, in the z-scan
method a sample is situated on a stage translatable along the
optical axis of a focused laser beam which allows for the
simultaneous monitoring of optical and acoustic responses at
varied laser fluence.[26a, 27] By adjusting the power of the input
laser, a samples optical absorption and corresponding PA
emission properties can be evaluated for a wide range of fluences.
The optical and photoacoustic responses of select compounds (1,
4 - 10) with g > 0 at the 532 nm laser excitation wavelength in
acetonitrile solvent have been evaluated using the OPAZ method.
Crystal violet has been included as a recently established LA
reference material for PA studies (532(CH3CN) = 4.56 x 104 M-1 cm-
1, fl(CH3CN) = 0).[18] The cyanine dye Cy3 has also been included
due to its efficient absorption at 532 nm (532(CH3CN) = 5.35 x 104
M-1 cm-1,fl(CH3CN) = 0.16) as well as the popularity of this class of
dye in PA imaging.
50
100
150
200
250
300
350
400
Laser Power Density (mJ cm-2)
Crystal violet
Cy3
1
4
5
6
7
8
9
0.20
0.16
0.12
0.08
0.04
10
50
100
150
200
250
300
350
400
Laser Power Density (mJ cm-2)
Figure 7. Optical (top) and photoacoustic (bottom) z-scan for molecules 1, 4 –
10, Cy3 and crystal violet reference compounds.
The crystal violet reference dye exhibits LA behavior up to a laser
fluence of 200 mJ cm-2 after which point the sample slowly
becomes saturated and exhibits weak SA behavior (Figure 7).
This LA optical z-scan behavior of crystal violet is correlated with
the positive linear slope of its ‘PA emission vs laser fluence’
response, at least until ca. 200 mJ cm-2 after which point the PA
emission tapers off due to sample saturation at higher laser
fluence. Of the curcuminoid dyes, the pyrenyl and julolidine
systems 6 and 10 exhibit distinct SA behavior from the onset of
their optical z-scan response consistent with a declining slope in
their ‘PA emission vs. laser fluence’ plots (Figure 7). It is the Cy3
cyanine dye however which exhibits prototypical SA behavior with
the weakest PA response compared to all of the curcuminoid dyes
here studied. Curcuminoids 5, 8 and 9 exhibit what appears to be
LA type behavior in their optical z-scan spectra. However, as the
latter dyes exhibit nanosecond lifetimes of their S1 states it is likely
that ground state bleaching, i.e. a SA response, is compensated
for by weak excited state absorption ultimately giving rise to an
overall LA response. Deviation from LA behavior to give the
desired RSA behavior is observed for curcuminoids 1, 4 and 7.
The onset of RSA behavior in the optical z-scan experiment is
This article is protected by copyright. All rights reserved.

10.1002/chem.201704423
Chemistry - A European Journal
FULL PAPER
observed at ca. 120 - 150 mJ cm-2 laser fluence. This corresponds
nicely to the onset of an enhanced PA emission indicated by a
deviation from a linear dependence on  (a function of the z
position in the OPAZ set up) Most significant is the relative
increase in the PA signal of 4 which is indicative of a significantly
increased capacity for excited state absorption at the 532 nm
probing wavelength for this dye. The excellent RSA behavior of 4
is consistent with it having the highest fluorescence lifetime in
acetonitrile of 3.71 ns. Quenching experiments confirm the
sequential nature of the two-photon absorption process.
150
100
50
slope (104 V/M)
crystal violet
Cy3
0.69
0.51
1
4
5
6
7
8
9
10
0.37
0.07
0.79
0.81
0.78
1.06
1.19
1.17
To probe the efficacy of all dyes as potential PAI and PAM
contrast agents a concentration dependence study has been
conducted at both low (15 mJ cm-2) and high (360 mJ cm-2) laser
fluence to determine the molar PA emission of all dyes, here
reported using the observed PA vs. concentration linear
regression slopes in units of V M-1 (Figure 8). It must be kept in
mind that the molar PA emission values reported are not absolute
but relative and are strongly dependent upon all experimental
conditions; most notably the laser fluence, pulse width and
wavelength, as well as the solvent and sample path length. At the
low laser power of 15 mJ cm-2 all dyes are restricted to a linear
PA emission due to the low probability of excited state absorption.
It is also noteworthy that the laser fluence of 15 mJ cm-2 at 532
nm falls below the maximum permissible exposure limits defined
by the American National Standards Institute (ANSI). The
phenomenon of nonlinear PA emission which currently requires a
high laser fluence is therefore not applicable to living tissue but
has potential application for in vitro applications using PAM.[28]
Under low laser fluence conditions the variable PA emission from
all samples is dictated by the product of the sample absorption
cross section (gCg) and quantum yield for nonradiation decay
(nr). Thus, from a molecular spectroscopy perspective variations
in PA emission can be easily explained by following trends in both
the S0S1 molar extinction coefficient at the excitation
wavelength and the quantum yield for competing fluorescence
decay (fl). Essentially, the linear correlation between molar PA
emission and concentration at low laser fluence is consistent with
Beer’s law for LA optical behavior, albeit slightly modified due to
competing fluorescence decay.
laser fluence = 15 mJ cm-2
0
0.00
0.02
0.04
0.06
0.08
Concentration (mM)
500
400
300
200
100
0
slope (104 V/M)
1.86
crystal violet
Cy3
0.84
1
4
5
6
7
8
9
10
1.34
0.61
2.27
1.93
2.89
3.35
5.21
2.69
laser fluence = 360 mJ cm-2
0.00
0.02
0.04
0.06 0.08
Concentration (mM)
Figure 8. Molar photoacoustic emission of all dyes measured at low (15
mJ cm-2, top) and high (360 mJ cm-2, bottom) laser fluence. Note the larger
y-axis scale used for presenting the high laser fluence data.
Ironically, at least when absorption cross section is not a variable
during equal O.D. measurements (Figure 7), it is this same
property which imparts 4 with an amplified PA emission under
high laser fluence conditions due to the higher probability of its
long-lived excited state to participate in a second photon
absorption. Comparing low vs. high laser fluence molar PA
measurements we can take advantage of the LA crystal violet dye,
capable of quantitative PA emission at both 15 mJ cm-2 and 360
mJ cm-2, a 2.7-fold increase in PA emission (1.86 x 104 V M-1/0.69
x 104 V M-1) is observed with the order of magnitude increase in
laser fluence. A 8.7-fold increase in molar PA emission is
exhibited by 4 comparing low and high fluence experiments (0.61
x 104 V M-1/0.07 x 10-4 V M-1) thus confirming its excited-state
absorption enhancement in molar PA emission relative to crystal
violet. It appears that although 4 is capable of enhanced molar PA
emission at high laser fluence it still suffers from a relatively weak
ground state absorption cross section at 532 nm as the molar PA
emission of 4, under RSA experimental conditions, still falls short
of the LA and SA characteristic dyes due to their greater S0S1
molar extinction coefficients at the excitation wavelength (Figure
8, bottom). Of the curcuminoid dyes, the naturally derived
The anthracenyl (5, 0.79 x 104 V M-1), pyrenyl (6, 0.81 x 104 V M-
1), dimethylaminophenyl (7, 0.78 x 104 V M-1), triphenylamine (8,
1.06 x 104 V M-1), carbazole (9, 1.19 x 104 V M-1) and julolidene
(10, 1.17 x 104 V M-1) curcuminoids each outperform the reference
dyes crystal violet (0.69 x 104 V M-1) and Cy3 (0.51 x 104 V M-1)
most likely due to their strong S0S1 molar extinction coefficients
at 532 nm (Figure 3). The naturally derived curcuminBF2 dye (1,
0.37 x 104 V M-1) exhibits a slightly weaker molar PA emission at
low laser fluence relative to both the crystal violet and Cy3
reference dyes. The weakest molar PA emission of the series is
exhibited by the napthacenyl terminated curcuminoid 4 (0.07 x 104
V M-1). This weak molar PA emission from 4 observed at low laser
fluence is attributed to its relatively long lived excited state in
acetonitrile of 3.71 ns which gives rise to SA behavior.
This article is protected by copyright. All rights reserved.

10.1002/chem.201704423
Chemistry - A European Journal
FULL PAPER
curcuminBF2 dye 1 exhibits a 3.6-fold enhancement of its molar
PA emission at high laser fluence due to weak nonlinear
enhancement, allowing it to overtake the SA Cy3 dye. The
dimethylaminophenyl curcuminoid 7 exhibits a strong molar PA
emission at both low (0.78 x 104 V M-1) and high (2.89 x 104 V M-
1) laser fluence. Of the entire series the bis-carbazole curcuminoid
9 exhibits the strongest molar PA emission at high laser fluence
(5.21 x 104 V M-1). This is attributed to 9 possessing the largest
S0S1 molar extinction coefficient of all dyes here studied at 532
nm but also due to residual RSA behavior as eluded to earlier in
the discussion of PAZ-scan data. The observation of a 4.4-fold
increase in molar PA emission for 9 confirms this residual RSA
behavior relative to the LA crystal violet dye (2.7-fold increase)
upon comparison of low vs high laser fluence ‘PA emission vs.
concentration’ slopes.
At low laser fluence there is little difference in contrast between
each sample with curcuminBF2 (1), the bis-pyrenyl (6) and bis-
triphenylamine (8) curcuminoids performing equally as well as the
excellent crystal violet LA dye due to their high molar extinction
coefficients at 532 nm. At the higher laser fluence of 150 mJ cm-2
RSA behavior is not yet strongly evident, however SA behavior of
Cy3 and 6 is evident due to their relatively weaker PA contrast. At
the highest applied laser fluence of 450 mJ cm-2 the nonlinear PA
enhancement of the bis-naphthalenyl (4) dye really comes to the
fore exhibiting a strong contrast with 1 following close behind. A
quantitative plot of normalized PA emission extracted from the
PAT image clearly illustrates the dominance of 1 and 4 and high
laser fluence (Figure 9). The remaining tertiary amine
functionalized curcuminoids 7-10 each also exhibit an excellent
contrast with 8 performing on par with the naturally derived
curcuminoid 1.
Conclusions
Two series of donor--acceptor--donor quadrupolar curcuminoid
molecules have been successfully synthesized through the
incorporation of aryl and tertiary amino substituents in order to
investigate the effect of -extension and increased electron
donating strength on their photophysical and photoacoustic
properties.
UV-Vis
absorption,
fluorescence emission,
electrochemical and computational investigations all consistently
correlate red shifted absorption maxima in both series of
curcuminoids with increased -conjugation and/or electron
donating strength. Time resolved fluorescence emission studies
have found all molecules to possess S1 excited state lifetimes in
the 1-2 ns range, thus providing the potential to display RSA
behavior with a 3 ns pulse width Nd:YAG laser at 532 nm. Optical
and photoacoustic z-scan (OPAZ-scan) experiments conducted
have determined that the bis-naphthalenyl (4), bis-p-
dimethylaminophenyl (7), as well as the naturally derived
curcuminBF2 (1) display RSA behavior at laser fluences > 120 mJ
cm-2. The strongest enhancement of PA signal due to this excited
state absorption phenomenon is displayed by 4. Trends in the
OPAZ-scan experiments were also borne out in PAT imaging
experiments, notably with the bis-triphenylamine curcuminoid (8)
outperforming the LA crystal violet reference dye indicative of
weak RSA behavior not evident in OPAZ-scan studies.
Importantly, the ‘molar photoacoustic response’ has been
reported which allows for a truly critical evaluation of the potential
of these dyes towards application in PA imaging and microscopy
relative to the LA crystal violet 532 nm reference dye. At a low
laser fluence of 15 mJ cm-2, within ANSI maximum permissible
exposure limits, the bis-carbazole curcuminod (9) exhibits the
strongest molar PA emission at 1.19 x 104 V M-1. Upon ramping
to a higher laser fluence of 360 mJ cm-2, an 8.7-fold increase in
molar PA emission is exhibited by 4 relative to the 2.7-fold
increase exhibited by the LA crystal violet dye, thus confirming its
nonlinear enhancement in molar PA emission. Of the entire series
of dyes studied, included crystal violet and Cy3, the bis-carbazole
curcuminoid 9 exhibits the strongest molar PA emission at high
laser fluence (5.21 x 104 V M-1). This is attributed to 9 possessing
the largest S0S1 molar extinction coefficient of all dyes here
studied at 532 nm but also due to a weak but contributing RSA
behavior. Importantly, with respect to the future design of MPACs,
15 mJ cm-2
150 mJ cm-2
450 mJ cm-2
4
8
1
7
10
9
CV
6
5
Cy3
Figure. 9. (top) PAT images recorded using three various laser fluence at λexc
= 532 nm. The x-axis color scale represents the normalized acoustic intensity.
(bottom) A plot of the relative PA emission intensities extracted from each PAT
image illustrating the nonlinear PA amplification for RSA dyes. (note: CV =
crystal violet).
Photoacoustic tomography (PAT) was also conducted on
curcuminoids 1, 4-7, alongside a solvent blank and reference
dyes crystal violet and Cy3 (Figure 9). The same image was
recorded at low (15 mJ cm-2) and high (150 mJ cm-2 and 450 mJ
cm-2) laser fluences where each sample was prepared with an
equal optical density of 0.15 in 1 mm borosilicate capillary tubes.
This article is protected by copyright. All rights reserved.

10.1002/chem.201704423
Chemistry - A European Journal
FULL PAPER
9: 1H NMR (CDCl3)  ppm 1.37 (br. s., 6 H) 1.37 - 1.48 (m, 7 H) 4.19 - 4.29
(m, 4 H) 5.86 - 6.03 (m, 2 H) 6.69 (d, J=0.27 Hz, 2 H) 6.64 (d, J=0.14 Hz,
2 H) 7.20 - 7.37 (m, 11 H) 7.40 - 7.56 (m, 5 H) 7.56 - 7.73 (m, 4 H) 8.01 -
8.07 (m, 3 H) 8.09 - 8.18 (m, 5 H). HRMS (ESI-) calcd. for C35H29BF2N2O2
558.2290, found 558.2296
this observation suggests that the magnitude of nonlinear molar
PA enhancement alone is not appropriate and that a strong RSA
behavior combined with a large ground state molar extinction
coefficient is an attractive strategy to provide for a high PA
contrast. Equipped with this better understanding of nonlinear PA
emission enhancement future studies will focus on transitioning
this technology towards the NIR biological window (700 – 900 nm)
using water soluble contrast agents.
10: 1H NMR (CDCl3)  ppm 1.24 - 1.27 (m, 1 H) 1.29 (s, 17 H) 1.64 - 1.68
(m, 1 H) 1.73 - 1.79 (m, 5 H) 2.23 (s, 4 H) 3.33 - 3.44 (m, 6 H) 5.85 (s, 1
H) 6.28 - 6.31 (m, 1 H) 6.33 - 6.36 (m, 1 H) 7.33 (s, 2 H). HRMS (ESI-)
calcd. for C39H49BF2N2O2 626.3855, found 626.3862
Physical Measurements
Experimental Section
NMR spectra were recorded on an Agilent spectrometer operated at
399.80 MHz for 1H , 282 MHz for 19F and 100.54 MHz for 13C nuclei in d-
chloroform or d6-DMSO and the residual solvent signals used as an
internal reference point for reporting the chemical shift ().[30] Mass
spectroscopy was carried on a Thermo Scientific Orbitrap Velos Pro mass
spectrometer. Electrochemical experiments were carried out on a CH
Instruments 620D potentiostat. A standard three electrode cell was used
under an atmosphere of argon with 0.1 M Bu4NPF6 in spectrophotometric
grade acetonitrile as the supporting electrolyte. Glassy carbon (3-mm
diameter) and Pt wire were used as working and counter electrodes
respectively. A non-aqueous reference electrode was used to minimize IR
drop at the solvent interface. This consisted of a Ag wire in the same
supporting electrolyte separated by a vycor frit. All sweep and pulse
voltammetry experiments were calibrated using ferrocene as an internal
pseudo reference due to the relative instability of the reference electrode
employed and all potentials are thus reported in reference to the
ferricenium/ferrocence (Fc+/0) couple. Due to the irreversibility of all
potentials the redox potential E could not be calculated and oxidation or
reduction potentials are reported from square-wave voltammograms (step:
4 mV, amplitude: 10 mV, frequency: 15 s-1) as a best estimate. UV-Vis
absorption spectra were recorded on an Agilent 8456 diode array
spectrophotometer in spectrophotometric grade solvents. Steady state
and time-resolved fluorescence measurements were carried out on a
Photon Technology International Quantamaster 40 & 25 fluorimeter at
room temperature. Quantum yields were calculated by the optically dilute
Materials
Trifluoroboron
benzaldehyde, biphenyl-4-carboxaldehyde, 2-naphthaldehyde,
anthracenecarboxaldehyde, 1-pyrenecarboxaldehyde,
etherate,
4-hydroxy-3-methoxybenzaldehyde,
9-
4-
(dimethylamino)benzaldehyde, 4-(diphenylamino)benzaldehyde, 9-ethyl-
3-carbazolecarboxaldehyde, 1,1,7,7-tetramethyl-9-formyljulolidine, and
acetylacetone were purchased from Sigma Aldrich and used as received.
Spectroscopic grade toluene, dichloromethane and methanol were
purchased from Sigma Aldrich. ACS reagent grade toluene,
dichloromethane and hexane for synthesis were purchased from Pharmco
Aaper. Deuterated solvents d6-dimethylsulfoxide and d-chloroform were
used as received from Sigma Aldrich.
Synthesis
For synthetic ease the reported method for borondifluoride curcuminoids
was amended with the first step to obtain the pure acetylacetone
difluoroboron intermediate (acacBF2) done in bulk prior to performing the
reported Knovenagel condensation29 with select aldehydes (1, 4-hydroxy-
3-methoxybenzaldehyde; 2, benzaldehyde; 3, biphenyl-4-carboxaldehyde;
4,
2-naphthaldehyde;
5,
9-anthracenecarboxaldehyde;
6, 1-
pyrenecarboxaldehyde; 7, 4-(dimethylamino)benzaldehyde; 8, 4-
(diphenylamino)benzaldehyde; 9, 9-ethyl-3-carbazolecarboxaldehyde; 10,
1,1,7,7-tetramethyl-9-formyljulolidine) for curcuminoid formation. All
products are sparingly soluble in toluene and yielded substantial
precipitate upon cooling the reaction mixture. Analytically pure compounds
were obtained following a wash with cold toluene and recrystallization from
dicholormethane:hexane (1:10) consistent with the reported general
method for borondifluoride curcuminoid synthesis. 29 Structures and purity
of final compounds 1-10 were verified by 1H-NMR (13C-NMR where
solubility allowed) and high resolution mass spectroscopy. Compounds 1,
2, 7 and 8 have been previously reported and analytical data matched
those of the literature.29, [29]
technique with fluorescein (aqueous NaOH, pH = 1, λexc = 390 nm, Φref
=
0.925) and rhodamine 6G (neat acetonitrile, λexc = 480 nm, Φref = 0.94) used
as reference standards according to Eq. 2.
2
퐼
휂
s
Φfl = (퐴ref) ( )( ) Φref
(2)
s
퐴
퐼
휂
ref
s
푟푒푓
The subscript “s” refers to the unknown sample and the subscript “ref” to
the reference sample, A is the absorbance at the excitation wavelength, I
is the integrated emission area, and η is the solvent refractive index.
Excitation and emission slits were both set to 5 nm. Fluorescence lifetimes
(τ) were recorded at room temperature at the emission maximum following
LED excitation at 456 nm or 572 nm. The radiative rate constant (kr) and
nonradiative rate constant (knr) were both calculated from τ and Φ by using
Eqs. 3-6.
3: 1H NMR (d6-DMSO) δ ppm 6.68 (s, 1 H) 7.24 - 7.61 (m, 12 H) 7.69 (br
s, 3 H) 7.73 - 7.90 (m, 8 H) 7.99 (d, J=8.24 Hz, 3 H) 8.08 (s, 1 H) 8.14 (s,
2 H). 13C NMR (d6-DMSO) δ ppm 126.88, 127.37,129.12, 130.35 (s, 1 C).
HRMS (ESI-) calcd. for C31H23BF2O2 476.1759, found 476.1765.
1
4: H NMR (CDCl3) δ ppm 5.96 (s, 1 H) 6.76 (d, J=15.66 Hz, 2 H) 7.43 -
1
휏 =
(3)
(4)
(5)
(6)
7.67 (m, 10 H) 7.80 - 7.97 (m, 8 H) 8.28 (d, J=8.24 Hz, 3 H) 8.50 - 8.63 (m,
1 H). HRMS (ESI-) calcd. for C27H19BF2O2 424.1446, found 424.1451
푘
+ 푘
푛푟
r
푘
r
Φ =
푘
+ 푘
푛푟
r
1
5: H NMR (CDCl3) δ ppm 5.99 (s, 1 H) 6.67 (d, J=15.93 Hz, 2 H) 7.48 -
Φ
휏
푘r
=
7.57 (m, 8 H) 8.01 - 8.08 (m, 4 H) 8.30 - 8.35 (m, 4 H) 8.48 (s, 2 H) 8.67 -
8.74 (m, 2 H). HRMS (ESI-) calcd. for C35H23BF2O2 524.1759, found
524.1765.
1−Φ
푘nr
=
휏
6: 1H NMR (CDCl3) δ ppm 3.08 (s, 12 H) 5.89 (s, 1 H) 6.46 (d, J=15.38 Hz,
2 H) 6.69 (d, J=8.79 Hz, 4 H) 7.51 (d, J=9.06 Hz, 6 H) 7.96 (d, J=15.38 Hz,
2 H). HRMS (ESI-) calcd. for C39H23BF2O2 572.1759, found 572.1766.
Computational Details
All calculations were carried out using density functional theory (DFT) with
the B3LYP (2-6) and CAM B3LYP (1,7-10) functional as implemented in
the Gaussian 09 program package and 6-311g (d,p) basis set.[31] The
optimization calculations were carried out using the polarizable continuum
model (PCM) with the dielectric constant of toluene.[32] A vibrational
This article is protected by copyright. All rights reserved.

10.1002/chem.201704423
Chemistry - A European Journal
FULL PAPER
frequency analysis coupling with PCM model was carried out in order to
confirm the minimum-energy geometry in solution, followed by Time-
dependent Density Functional Theory (TD-DFT).[33] TD-DFT calculations
for 1, 7-10 included the keyword cis(d) to account for configurational
interactions.
For PAT measurments the output of a frequency doubled Nd:YAG laser is
directed onto a prism which allows the laser beam to focus on the sample
at a 45 angle using a 10 cm confocal lens. The sample is placed in a cell
housing unit which is filled with water for acoustic signal coupling. A 10
MHz water immersion unfocused transducer (Olympus V311-SU is placed
directly above the sample. The sample is mounted on an automated XYZ
translation stage (Thorlabs NRT 100) and moved along the x and y
directions in discrete steps to perform 2D raster scan. The PA signal is
collected by the transducer and then amplified using a pulse amplifier
(which was fed to a Lecroy Wavepro oscilloscope for display and data
collection. The scanning and data collections are controlled by a Labview
routine. By collecting data points along the xy plane, a maximum intensity
projection (MIP) is obtained by taking the absolute value of the Hilbert
transform of the acquired signal via MATLAB. The MIP image is the map
of optical absorption of the sample. The lateral resolution of the PAT
system is dictated by the laser beam size which was estimated to be 70 ±
5 휇m. The dyes were injected into 1 mm tube and then placed parallel in
the Y direction in the bottom of the cell housing unit.
Collection and Reduction of X-Ray Data
A crystal (0.50 x 0.10 x 0.03 mm) of 5∙CHCl3 mounted on the end of a glass
fiber and transferred to Bruker Kappa Apex II diffractometer for the
collection of diffraction data. Diffraction data for indicated triclinic symmetry
consistent with space group P-1. Crystal data and information about the
data collection are provided in Tables S3 – S7 in the supporting information.
Determination and Refinement of the Structure
The structure of 5∙CHCl3 was solved[34] by direct methods. In the least-
squares refinement[34], anisotropic temperature parameters were used for
all the non-hydrogen atoms except for the fluorine and oxygen atoms.
Hydrogen atoms were placed at calculated positions and allowed to "ride"
on the atom to which they were attached except for the hydrogen atoms
on the water of crystallization which were not included. The isotropic
thermal parameter for the hydrogen atoms were determined from the atom
to which they were attached. The data was corrected using the multi-scan
method (SADABS).[35]
Acknowledgements
Funding for this research was provided by NIH grant
U54CA156734. FM thanks the Oracle Education Foundation for
financial support. We thank Dr. E. Fujita of Brookhaven National
Laboratory for the use of the Bruker Kappa Apex II diffractometer
for X-ray data collection.
Optical photoacoustic z-scan (OPAZ) experimental details
For OPAZ-scan measurements a 2.0 mm path length quartz cell was
placed at a 45 angle with respect to the incident laser beam (effective
path length = 2.83 mm). A custom made sample cell housing unit was used
wherein the quartz cell is placed and is filled with water for acoustic signal
transmission. Samples were dissolved in spectroscopic grade acetonitrile
having a linear absorption coefficient ) of 345 m-1 at the laser excitation
Keywords: curcumin • photoacoustic • fluorescence • contrast
agent • imaging
wavelength
532
nm
(optical
density
=
0.3).
A
9:1
acetonitrile:dichloromethane mixture was required for 4 - 10 to prevent
precipitation during data collection. The output of a frequency doubled
Nd:YAG laser (Continuum Minilite II, 532 nm, pulse width ~3 ns) was
focused onto the sample using a 18 cm focal length lens. The sample was
mounted on an automated translation stage (Thorlabs NRT 150) and
moved horizontally along the z direction through the focal point of the beam.
Each z-scan experiment took roughly 7 minutes in total. Samples were
translated along the z-axis over 15 cm in steps of 5 mm. At each position
the photoacoustic and optical signals were recorded by averaging the
response of 20 laser pulses (@ 10 Hz = 2 sec. duration). UV-Vis absorption
spectra were recorded before and after each z-scan experiment to ensure
the stability of the sample. No degradation was observed under these
experimental conditions for the series of dyes presented. Each experiment
was repeated 3 times with freshly prepared sample to check for
reproducibility. The beam waist (x0) at focal plane was estimated to be 70
± 5 m. The energy incident on the sample was controlled by the
combination of a half-wave plate and a linear polarizer. The incident laser
energy before the focusing lens S5 was 65 J. At the focal point the sample
experiences optimum pump intensity, which decreases gradually on either
side of the focus as the sample is translated along the z axis. As the
fluence of incident light changes, the optical transmittance varies
according to the sample’s nonlinear electronic absorption properties. The
PA emission was collected using a 10 MHz, 1 inch focal length water
immersion ultrasonic transducer (Olympus NDT U8517074). Importantly,
the linear response of the optical detector as well as the ultrasound
transducer were verified by measuring the transmittance and the
corresponding generated PA response as a function of crystal violet
concentration in acetonitrile (optical density from 0 to 1 ) at 532 nm. A
References
[1]
a) B. B. Aggarwal, C. Sundaram, N. Malani, H. Ichikawa, Molecular
Targets and Therapeutic Uses of Curcumin in Health and Disease 2007,
595, 1-+; b) R. A. Sharma, A. J. Gescher, W. P. Steward, European
Journal of Cancer 2005, 41, 1955-1968; c) A. Goel, A. B. Kunnumakkara,
B. B. Aggarwal, Biochemical Pharmacology 2008, 75, 787-809; d) K.
Bairwa, J. Grover, M. Kania, S. M. Jachak, Rsc Advances 2014, 4,
13946-13978.
[2]
[3]
F. Henari, S. Cassidy, K. E. Jasim, A. Dakhel, Journal of Nonlinear
Optical Physics & Materials 2013, 22, 1350017.
a) T. Ganesh, J. H. Kim, S. J. Yoon, B.-H. Kil, N. Maldar, J. W. Han, S.-
H. Han, Materials Chemistry and Physics 2010, 123, 62-66; b) G.
Richhariya, A. Kumar, P. Tekasakul, B. Gupta, Renewable and
Sustainable Energy Reviews 2017, 69, 705-718.
[4]
[5]
a) S.-S. Zhou, X. Xue, J.-F. Wang, Y. Dong, B. Jiang, D. Wei, M.-L. Wan,
Y. Jia, Journal of Materials Chemistry 2012, 22, 22774-22780; b) M.
Proehl, U. S. Schubert, W. Weigand, M. Gottschaldt, Coordination
Chemistry Reviews 2016, 307, 32-41.
A. Kumar, L. Li, A. Chaturvedi, J. Brzostowski, J. Chittigori, S. Pierce, L.
A. Samuelson, D. Sandman, J. Kumar, Applied Physics Letters 2012,
100, 203701.
[6]
[7]
W. Wu, J. Shen, P. Banerjee, S. Zhou, Biomaterials 2011, 32, 598-609.
a) B. B. Aggarwal, A. Kumar, A. C. Bharti, Anticancer research 2003, 23,
363-398; b) M. Gyuris, L. Hackler, L. I. Nagy, R. Alföldi, E. Rédei, A.
Marton, T. Vellai, N. Faragó, B. Ózsvári, A. Hetényi, Archiv der
Pharmazie 2017.
correlation better than R2
= 0.99 was obtained for transmittance
measurements when a neutral density filter of OD 1 was placed in front of
the detector (Newport 818 series photodiode sensor with a 3 OD filter; data
not shown) while the PA signal showed excellent linearity.
[8]
a) K. Kamada, T. Namikawa, S. Senatore, C. Matthews, P.-F. Lenne, O.
Maury, C. Andraud, M. Ponce-Vargas, B. Le Guennic, D. Jacquemin, P.
Agbo, D. D. An, S. S. Gauny, X. Liu, R. J. Abergel, F. Fages, A. D'Aleo,
Chemistry (Weinheim an der Bergstrasse, Germany) 2016, 22, 5219-
5232; b) C. F. Chignell, P. Bilski, K. J. Reszka, A. G. Motten, R. H. Sik,
T. A. Dahl, Photochemistry and Photobiology 1994, 59, 295-302.
Photoacoustic tomography
This article is protected by copyright. All rights reserved.

10.1002/chem.201704423
Chemistry - A European Journal
FULL PAPER
[9]
a) P. Beard, Interface focus 2011, rsfs20110028; b) L. Li, L. Zhu, C. Ma,
L. Lin, J. Yao, L. Wang, K. Maslov, R. Zhang, W. Chen, J. Shi, Nature
Biomedical Engineering 2017, 1, Art. No. 0071.
[28] Y. Zhang, X. Cai, Y. Wang, C. Zhang, L. Li, S. W. Choi, L. V. Wang, Y.
Xia, Angewandte Chemie 2011, 123, 7497-7501.
[29] L. Zhai, M. Liu, P. Xue, J. Sun, P. Gong, Z. Zhang, J. Sun, R. Lu, Journal
of Materials Chemistry C 2016, 4, 7939-7947.
[10] a) H. F. Zhang, K. Maslov, G. Stoica, L. V. Wang, Nature biotechnology
2006, 24, 848; b) S. Park, J.-C. Vial, K. Kyhm, Opt. Express 2017, 25,
18917-18928.
[30] G. R. Fulmer, A. J. M. Miller, N. H. Sherden, H. E. Gottlieb, A. Nudelman,
B. M. Stoltz, J. E. Bercaw, K. I. Goldberg, Organometallics 2010, 29,
2176-2179.
[11] A. G. Bell, American Journal of Science 1880, 305-324.
[12] S. Mallidi, G. P. Luke, S. Emelianov, Trends in Biotechnology 2011, 29,
213-221.
[31] Frisch M. J. et al., Revision A.1 ed., Gaussian Inc., Wallingford, CT.,
2009.
[13] a) M. L. James, S. S. Gambhir, Physiological Reviews 2012, 92, 897-
965; b) K. Pu, A. J. Shuhendler, J. V. Jokerst, J. Mei, S. S. Gambhir, Z.
Bao, J. Rao, Nat. Nanotechnol. 2014, 9, 233; c) Q. Miao, K. Pu,
Bioconjugate Chem. 2016, 27, 2808-2823; d) Y. Jiang, K. Pu, Small 2017,
13, 1700710-n/a; e) Y. Lyu, X. Zhen, Y. Miao, K. Pu, ACS Nano 2017,
11, 358-367.
[32] J. Tomasi, B. Mennucci, R. Cammi, Chem. Rev. 2005, 105, 2999-3093.
[33] G. Scalmani, M. J. Frisch, B. Mennucci, J. Tomasi, R. Cammi, V. Barone,
J. Chem. Phys. 2006, 124, 94107.
[34] G. M. Sheldrick, 1994, SHELXL Version 5, Siemens Analytical
Instruments, Inc., Madison WI.
[35] Version 2007/2 ed., (Sheldrick, Bruker AXS Inc.).
[14] J. Fang, H. Nakamura, H. Maeda, Advanced Drug Delivery Reviews
2011, 63, 136-151.
[15] a) A. S. Thakor, S. S. Gambhir, CA: A Cancer Journal for Clinicians 2013,
63, 395-418; b) H. S. Choi, W. Liu, P. Misra, E. Tanaka, J. P. Zimmer, B.
I. Ipe, M. G. Bawendi, J. V. Frangioni, Nature biotechnology 2007, 25,
1165-1170.
[16] a) M. A. K. Abdelhalim, B. M. Jarrar, Journal of Nanobiotechnology 2012,
10, 1-9; b) S. Kommareddy, M. Amiji, Journal of pharmaceutical sciences
2007, 96, 397-407.
[17] G. P. Luke, D. Yeager, S. Y. Emelianov, Annals of biomedical
engineering 2012, 40, 422-437.
[18] M. Frenette, M. Hatamimoslehabadi, S. Bellinger-Buckley, S. Laoui, J.
La, S. Bag, S. Mallidi, T. Hasan, B. Bouma, C. Yeleswarapu, J. Rochford,
J. Am. Chem. Soc. 2014, 136, 15853-15856.
[19] K. Liu, J. Chen, J. Chojnacki, S. Zhang, Tetrahedron Lett. 2013, 54,
2070-2073.
[20] N. Aliaga-Alcalde, P. Marques-Gallego, M. Kraaijkamp, C. Herranz-
Lancho, H. den Dulk, H. Gorner, O. Roubeau, S. J. Teat, T.
Weyhermuller, J. Reedijk, Inorganic Chemistry 2010, 49, 9655-9663.
[21] a) N. N. Barashkov, T. V. Sakhno, R. N. Nurmukhametov, O. A. Khakhel,
Uspekhi Khimii 1993, 62, 579-593; b) O. A. Khakhel, R. N.
Nurmukhamedov, T. V. Sakhno, S. A. Serov, N. N. Barashkov, T. M.
Muraveva, Zhurnal Fizicheskoi Khimii 1992, 66, 2639-2645; c) O. A.
Khakhel, R. N. Nurmukhametov, T. V. Sakhno, Chemical Physics Letters
1995, 247, 185-188; d) V. E. Krikunova, S. A. Serov, O. A. Khakhel,
Optics and Spectroscopy 1999, 86, 373-376.
[22] a) E. Lippert, Z. Elektrochem 1957, 61, 962; b) N. Mataga, Y. Kaifu, M.
Koizumi, Bulletin of the Chemical Society of Japan 1956, 29, 465-470; c)
J. R. Lakowicz, Principles of Fluorescence Spectroscopy, Springer US,
2007.
[23] A. Felouat, A. D'Aleo, F. Fages, Journal of Organic Chemistry 2013, 78,
4446-4455.
[24] a) A. D'Aleo, D. Gachet, V. Heresanu, M. Giorgi, F. Fages, Chemistry-a
European Journal 2012, 18, 12764-12772; b) E. Kim, A. Felouat, E.
Zaborova, J.-C. Ribierre, J. W. Wu, S. Senatore, C. Matthews, P.-F.
Lenne, C. Baffert, A. Karapetyan, M. Giorgi, D. Jacquemin, M. Ponce-
Vargas, B. Le Guennic, F. Fages, A. D'Aleo, Organic & Biomolecular
Chemistry 2016, 14, 1311-1324; c) E. Cogne-Laage, J. F. Allemand, O.
Ruel, J. B. Baudin, V. Croquette, M. Blanchard-Desce, L. Jullien,
Chemistry-a European Journal 2004, 10, 1445-1455.
[25] a) M. Frenette, M. Hatamimoslehabadi, S. Bellinger-Buckley, S. Laoui, J.
La, S. Bag, S. Mallidi, T. Hasan, B. Bouma, C. Yelleswarapu, J. Rochford,
J. Am. Chem. Soc. 2014, 136, 15853-15856; b) D. N. Christodoulides, I.
C. Khoo, G. J. Salamo, G. I. Stegeman, E. W. Van Stryland, Adv. Opt.
Photonics 2010, 2, 60-200.
[26] a) C. S. Yelleswarapu, S.-R. Kothapalli, Opt. Express 2010, 18, 9020-
9025; b) R. R. Hung, J. J. Grabowski, Journal of the American Chemical
Society 1992, 114, 351-353.
[27] a) P. Chapple, J. Staromlynska, J. Hermann, T. Mckay, R. McDuff,
Journal of Nonlinear Optical Physics & Materials 1997, 6, 251-293; b) M.
Sheikbahae, A. A. Said, T. H. Wei, D. J. Hagan, E. W. Vanstryland, Ieee
Journal of Quantum Electronics 1990, 26, 760-769.
This article is protected by copyright. All rights reserved.