Accepted Manuscript

Performance and stability of low-cost dye-sensitized solar cell based crude and pre-
concentrated anthocyanins: Combined experimental and DFT/TDDFT study

Phrompak Chaiamornnugool, Sarawut Tontapha, Ratchanee Phatchana, Nattawat

Ratchapolthavisin, Somdej Kanokmedhakul, Wichien Sang-aroon, Vittaya
Amornkitbamrung
PII: S0022-2860(16)30770-0
DOI: 10.1016/j.molstruc.2016.07.086
Reference: MOLSTR 22789

To appear in: Journal of Molecular Structure

Received Date: 19 January 2016

Revised Date: 18 July 2016
Accepted Date: 20 July 2016

Please cite this article as: P. Chaiamornnugool, S. Tontapha, R. Phatchana, N. Ratchapolthavisin,

S. Kanokmedhakul, W. Sang-aroon, V. Amornkitbamrung, Performance and stability of low-cost dye-
sensitized solar cell based crude and pre-concentrated anthocyanins: Combined experimental and DFT/
TDDFT study, Journal of Molecular Structure (2016), doi: 10.1016/j.molstruc.2016.07.086.

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.
 1
ACCEPTED MANUSCRIPT

Performance and stability of low-cost dye-sensitized solar

cell based crude and pre-concentrated anthocyanins:
Combined experimental and DFT/TDDFT study
Phrompak Chaiamornnugool1,6, Sarawut Tontapha1, Ratchanee Phatchana2, Nattawat
Ratchapolthavisin3,5,6, Somdej Kanokmedhakul4, Wichien Sang−aroon2* , Vittaya
Amornkitbamrung3,5,6*

PT
1
Materials Science and Nanotechnology Program, Faculty of Science, KhonKaen University, KhonKaen 40002, Thailand
2
Department of Chemistry, Faculty of Engineering, Rajamangala University of Technology Isan, KhonKaen Campus, KhonKaen 40000,
Thailand

RI
3
Integrated Nanotechnology Research Center, Department of Physics, Faculty of Science, KhonKaen University, KhonKaen 40002,
Thailand
4
Department of Chemistry, Faculty of Science, Faculty of Science, KhonKaen University, KhonKaen 40002, Thailand

SC
5
Thailand Center of Excellence in Physics, CHE, Ministry of Education, Bangkok 10400, Thailand
6
Nanotec-KKU Center of Excellence on Advanced Nanomaterials for Energy Production and Storage, Khon Kaen 40002, Thailand

U
* Corresponding author: Tel.:+6643 336370; fax: +6643 336381 (W. Sang-aroon), Tel.:+6643 203359; fax: +6643 203359 (V.
Amornkitbamrung)
AN
E-mail addresses: wichien.sa@rmuti.ac.th (W. Sang−aroon),vittaya@kku.ac.th (V. Amornkitbamrung)

Supplementary material is available via the internet at

M

ABSTRACT
D

The low cost DSSCs utilized by crude and pre-concentrated anthocyanins extracted
from six anthocyanin−rich samples including mangosteen pericarp, roselle, red cabbage, Thai
TE

berry, black rice and blue pea were fabricated. Their photo−to−current conversion
efficiencies and stability were examined. Pre-concentrated extracts were obtained by solid
EP

phase extraction (SPE) using C18 cartridge. The results obviously showed that all

pre−concentrated extracts performed on photovoltaic performances in DSSCs better than

C

crude extracts except for mangosteen pericarp. The DSSC sensitized by pre-concentrated
AC

anthocyanin from roselle and red cabbage showed maximum current efficiency η=0.71%

while DSSC sensitized by crude anthocyanin from mangosteen pericarp reached maximum
efficiency η=0.97%. In addition, pre−concentrated extract based cells possess more stability

than those of crude extract based cells. This indicates that pre−concentration of anthocyanin

via SPE method is very effective for DSSCs based on good photovoltaic performance and
stability. The DFT/TDDFT calaculations of electronic and photoelectrochemical properties of

1
 2
ACCEPTED MANUSCRIPT
the major anthocyanins found in the samples are employed to support the experimental
results.

Keywords: Crude and pre-concentrated anthocyanins; Solid phase extraction; Dye−sensitized

solar cell; Photovoltaic performance; DFT/TDDFT

1. Introduction

PT
Since first discovered by Grätzel el al. [1] in 1991, dye−sensitized solar cell (DSSC) has

RI
attracted much attention as a new promising solar to electric convertor because of its low

SC
production cost, easy fabrication, more environmental friendliness compared to conventional

silicon solar cell. The DSSC is composed of nanocrystalline porous semiconductor

U
electrode−adsorbed dye, a counter electrode and an electrolyte. Dye sensitizer plays a crucial
AN
role in absorbing sunlight and transforming solar energy into electric energy. Numerous

organic and metal complexed compounds have been synthesized and utilized as molecular
M

sensitizers in DSSCs. By far, DSSC utilized ruthenium (Ru) complex compounds as

D

molecular sensitizer has been reached 11−12% overall efficiency [2−5]. Even though the
TE

DSSCs−based ruthenium complexes have provided a relatively high efficiency, but they are

not yet suitable to be produced in large scale based on several disadvantages. For example,
EP

the noble metals serving as sensitizer are limited in amount and led to their high production

cost. In addition, complicated synthetic method and toxicity due to incomplete degradation
C

are considered to be the main drawbacks. To overcome this problem, organic dyes, especially
AC

natural pigments derived from different parts of various plants such as flower, fruit, leaf and

wood are promising alternative ways to reduce production cost of solar cells and they have

been widely studied and used to replace the ruthenium dye [6−26].

By far, it has been reported that the conversion efficiencies of 1.60 and 1.86 obtained

from roselle and pomegranate [27], 2.3 from monascus [28] and 1.17 and 2.63% from

2
 3
ACCEPTED MANUSCRIPT
mangosteen pericarp [29,30], respectively. From reviews, it is reported that almost of

extracted pigment contain anthocyanins (Figure 1) in different derivatives that show

significant role enhancing DSSC efficiencies according to the active carbonyl (−C=O) and

hydroxyl (−OH) groups which can generate and transfer electron during absorption of light

[31]. Recently, Mozaffari et al. investigated the photovoltaic characteristics of DSSCs

PT
sensitized by crude and purified extracts from Siahkooti fruit as sensitizers [32]. It has been

RI
found that efficiency of DSSCs using purified and crude extracts of the fruit were 0.32% and

SC
0.27%, respectively, indicating that purified extract performs on photovoltaic performance

better than crude extract. On the same hand, it is interesting to note that concentration of the

U
dyes plays a crucial role on photovoltaic performance of DSSCs.
AN
The aim of this work was to study and optimize photovoltaic performance and stability
M

as a function of the concentration and adsorption characteristics of the dyes on TiO2 thin

films. Solid phase extraction (SPE) using C18 cartridge was applied to pre-concentrate
D

anthocyanin from the crude extracts. To this end, dye adsorption behavior on TiO2 electrode
TE

were obtained and analyzed to better understand the relations between concentrations of the
EP

dye contain in extracted solutions and photovoltaic performances as well as stabilities their

DSSCs. The DFT/TDDFT calculations were performed to support the experimental results.
C

2. Experimentalmethod
AC

2.1 Preparation of natural pigments and characterization

Extraction and pre−concentrationof anthocyanin from the six samples were performed

based on the following procedure [33]. The pre-concentration method is shown in Scheme 1.

Fresh samples were cleaned with de−ionized water, air−dried and crushed into powder in a

3
 4
ACCEPTED MANUSCRIPT
mortar. 50 gof powdered sample was put into a beaker, 300 mL of acidified methanol

(0.01%v/v) was added and the mixture was kept overnight, stored at 4°C and dark condition.

The obtained solution was filtered out and centrifuged to get rid of any solid residues;

thereafter a pure and clear pigment solution was obtained. The solvent was removed using

PT
vacuum rotary evaporator at 40°C to get more concentrated crude extract. Further

purification of clear solution was performed using SPE to achieve more concentrated

RI
anthocyanin. A small drop of crude extract was dropped into the C18 cartridge. 5 mL of

SC
acidified de−ionized water was dropped to elute soluble sugar followed by 5 mL of ethyl

acetate was used for elution of undesirable polyphenolic compounds. The remaining

U
AN
anthocyanin in the cartridge was then eluted by acidified methanol. The pigment solutions

obtained before and after pre-concentration were stored and refrigerated at 4 °C before
M

UV−Vis characterization and used as sensitizer in fabrication of DSSCs. The absorption

D

spectra of pigment solutions both before and after pre-concentration were measured by
TE

UV−VIS spectrophotometer (model: T80+, PG Instruments Ltd). Structure of some isolated

anothocynins were characterized by NMR (Varian Mercury Plus 400) and also comparison
EP

with the spectral data reported from literature.

C

2.2 Fabrication of DSSCs

AC

A fluorine−doped SnO2 conductive glass (FTO) sheet resistance (15 Ω/cm2; XinYan,

China) was used as the current collector. The FTO plate was cleaned with acetone and

distilled water to remove impurities. The cleaned FTO glass surface was coated with

nanocrystalline TiO2 (P25, Degussa) paste using screen−prints method working area 0.25cm2.

The FTO/TiO2 electrodes were then sintered at 500 °Cfor 1 hr. The cooled down electrode

4
 5
ACCEPTED MANUSCRIPT
was immersed in the dye extract solution for 24 hrs. The TiO2/dye electrode was rinsed with

de−ionized water to remove impurities and then with methanol to remove trapped water from

the initial rinsing. The counter electrode was prepared by coating platinum catalyst onto FTO

glass. The 5 mM hexachloroplatinic acid (H2PtCl6)(Aldrich) solution in isopropanol was

PT
spread on the FTO glass surface. The Pt−coated FTO electrode was the heated in a furnace at

500 °C for 30 min. The two electrodes were clipped together to make a sandwich type cell.

RI
The electrolyte was composed of 0.1 MLiI (Fluka), 0.05 M I2 (Riedel−deHaën) solution in

SC
acetonitrile (Merck) and was injected into the space of electrodes then the space was filled

U
through the capillary action. AN
2.3 Dye adsoption measurement
M

The experiment for adsorption behavior of dye on TiO2 electrode were carried out by

immersing the TiO2 cell with an active area 4 cm2 in a small chamber filled with a solution of
D

crude or pre-concentrated anthocyanin [34]. The adsorption efficiency at time t was

TE

calcualated from Eq. 1

EP

 A − At 
% adsorption =  0  x100 (1)
 A0 
C

where A0 and At are the initial and any time absorbances determined at maximum
AC

absorption wavelenge of each solution using a UV-Vis spectrophotometer after filtration

through a 0.0015µm memebrane filter.

2.4Measurement of DSSC efficiency and stability

5
 6
ACCEPTED MANUSCRIPT
The solar to electric conversion efficiency (the photocurrent−voltage; J−V curve) was

measured using Peccell® model PEC−L11 solar simulator class A under 1.5 AM and

100mW/cm2 condition. Based on the J−V curve, the fill factor (FF) is defined as:

FF =
( J max x Vmax ) (2)
( J SC x VOC )

PT
RI
where Jmax and Vmax are the photocurrent and photovoltage, respectively for maximum power

output (Pmax) while JSC and VOC correspond to the short−circuit photocurrent and open−circuit

SC
photovoltage, respectively. The overall energy conversion efficiency (η) is defined as:

η=
( J SC x VOC
U x FF )
(3)
AN
Pin

where Pin is the power of incident light. The overall efficiency for all fabricated DSSCs was
M

recorded every 48 hrswithin 816 hrs for their stability tests.

D
TE

Incident photon−to−current efficiencies (IPCE) of DSSC under short−circuit conditions

were determined with a 150W Xe lamp (Oriel, USA) fitted with a monochromator
EP

(Cornerstone TM 130 1/8 m, USA) to provide a monochromatic light source.

C

3. Computational method
AC

3.1 DFT calculation

The optimization of the ground−state geometries and calculation of the first singlet

electronic excited state of the main anthocyanins extract from the six samples were carried

out by DFTand TDDFT methods, respectively. Beck’s three parameter hybrid exchange

function [35], Lee−Yang−Parr gradient−corrected correlation (B3LYP functional) [36] and

6
 7
ACCEPTED MANUSCRIPT
6−31+G(d,p) basis set were employed in both DFT and TDDFT methods. The electronic

absorption spectra were computed using TDDFT at the same level by single-point calculation

of the optimized ground state B3LYP/6−31+G(d,p) geometries. The solvent effects were

evaluated using the conductor−like polarize continuum model (C−PCM)[37−39]. All of

PT
calculations were performed with Gaussian 09 program package [40]. The electronic

structure of HOMOs and LUMOs of the pigments were plotted using Gaussview version 5.09

RI
[41].

SC
3.2 Electron injection force (∆Ginject)

U
The free energy change of electron injection process from excited state dye to conduction
AN
band edge of TiO2 can be evaluated as [42,43]:
M

∆ G inject = EOX
dye*
− ECB
TiO2
(4)
D

dye* TiO2
EOX is the excited state oxidation potential of the dyes and ECB is the conduction band
TE

dye*
edge of TiO2. EOX can be estimated with this following equation:
EP

= EOX −0 + ωr
− E 0dye
dye* dye
EOX (5)
C

ωr is the Coulombic stabilization term and can be negligible as mentioned in the previous
AC

dye*
work [44], so EOX dye
is approximately estimated to E OX − E 0dye
−0
TiO
. The experimental ECB 2
= −4.0 eV

used in this work [45]. DFT/TDDFT method was employed to study electron injection

mechanisms of the dye−TiO2 system. However, Preat et al. concluded that the ∆Ginject

computed from relaxed and unrelaxed paths is comparable [46]. Thus, in this work, ∆Ginject

dye*
and EOX were computed according to Eq. 4 and 6, respectively.

7
 8
ACCEPTED MANUSCRIPT
dye *
E OX = E OX
dye
− E 0dye
−0
(6)

dye
Based on the Koopman theorem [47], E HOMO is estimated as the ground state oxidation

dye
potential ( E OX ).

3.3 Deprotonation energy

PT
Deprotonation energy (∆Edep) of the dyes were calculated in B3LYP/6−31+G(d,p) level

RI
under the vacuum condition using the following equation:

SC
A→ A− + H+

U
∆ E dep = ( E A − + E H + ) − E A (7)
AN
A and A− correspond to the neutral and anionic forms of the dyes, EA and EA− are total

energy of neutral and anionic forms, respectively.

M
D

4. Results and discussion

TE

4.1 Experimental results

EP

4.1.1 UV−Vis absorption spectra

C

The dye solutions obtained before and after pre-concentration process called crude and
AC

pre-concentrated extracts were shown in Figure S1 of the Supplementary material (SM). The

colors of crude extracts of mangosteen pericarp, Thai berry and blue pea are transparent

yellow while of the remains are red. Absorption spectra of crude and pre−concentrated

extracts of all samples determined in acidified methanol at pH=1 are shown in Figure 2. The

visible spectra of the extracts show absorption bands in the region of 500 to 600 nm

8
 9
ACCEPTED MANUSCRIPT
corresponding to flavonoid and phenolic compounds which are anthocyanin. Recent

investigations indicated that anthocyanins are responsible for attractive colors of the fruits

with absorption maxima near 520 nm [48]. The most common anthocyanin found in flowers

and fruits are Pelargonidin (orange), Cyanidin (orange−red), Delphinidin (blue−red),

PT
petunidin (blue−red) and Malvinidin (blue−red) [49]. The maximum absorption wavelengths

(λmax) and also the major anthocyanin found in the extracts are summarized in Table 1.

RI
The molar absorptivity of pre−concentrated extracts of red cabbage, black rice and blue

SC
pea is higher than those of their crude extracts. The absorption spectrum of crude and pre-

U
concentrated extracts of Thai berry is superimposed in the region 550−700 nm but the
AN
absorptivity of pre-concentrated extracts at 400−550 is lesser than those of crude extracts. As

seen in Table 1, expected anthocyanins in the six samples are referred to the results of the
M

previous works. The pre−concentrated extract from black rice was successful to be purified by
D

gel column chromatography and structural characterization by NMR spectroscopy. The

TE

results confirm that the anthocyanin found in black rice is Cyanidin−3−O−glucoside (H NMR

data was shown in the SM) which is in agreement to the previous report [50,51]. The main
EP

anthocyanin found in blue pea extract is ternatin (or Dp in its aglycone form) which Cy and

Dp are mostly found in roselle extract [52]. The major anthocyanin found in red cabbage is
C
AC

Cy [53]. Cy and Pg are the main anthocyanins found in mangosteen pericarp [29,30,54,55].

The major anthocyanin found in Thai berry is Cy as observed in the previous report [56].

4.1.3 Photovoltaic properties

The performance of the fabricated DSSCs using crude and pre-concentrated extracts as

photosensitizers was evaluated by short circuit current density (JSC), open circuit voltage

9
 10
ACCEPTED MANUSCRIPT
(VOC), maximum power (Pmax), fill factor (FF) and energy conversion efficiency (η) and the

corresponding values are tabulated in Table 2. The J−V characteristics of the fabricated

DSSCs are shown in Figure 3. It is obviously seen that JSC of the four DSSCs sensitized

pre−concentrated extracts is increased while VOC is decreased in regards to those sensitized

PT
by crude extracts. Magnitude of JSC of DSSCs sensitized by pre−concentrated extracts is in

increasing order: mangosteen pericarp (1.57 mA/cm2) < blue pea (2.27 mA/cm2) < roselle

RI
(2.91 mA/cm2) < black rice (3.04 mA/cm2) < red cabbage (3.84 mA/cm2) < Thai berry (3.92

SC
mA/cm2) and of those sensitized by crude extracts is in this following order: blue pea (0.61

mA/cm2) < black rice (1.08 mA/cm2) < red cabbage (1.40 mA/cm2) < Thai berry (1.52

U
mA/cm2) < roselle (3.09 mA/cm2) < mangosteen pericarp (3.63 mA/cm2). It is shown that VOC
AN
of all DSSCs sensitized by crude and pre−concentrated extracts vary within a narrow range

0.27−0.51 V and 0.27−0.46 V, respectively. Decreasing of VOC from crude to

M

pre−concentrated systems is found for all cells except for roselle system. VOCs of the two
D

mangosteen systems are not much different. Fill factors of the six DSSCs sensitized by crude
TE

and pre-concentrated extracts are found within the range 0.36−0.60 and 0.37−0.56,

respectively. According to the photovoltaic characteristics, most of DSSCs sensitized by pre-

EP

concentrated extracts have more photoelectric conversion efficiency (%η) in comparison to

C

those of sensitized by crude extracts except for the DSSCs sensitized by mangosteen pericarp
AC

extracts which its crude extract shows higher efficiency. The ascending order of the

magnitude of efficiencies of the DSSCs utilized pre−concentrated extracts are roselle (0.71%)

∼ red cabbage (0.71%) > mangosteen pericarp (0.54%) ∼ black rice (0.53%) >Thai berry

(0.47%) > blue pea (0.33%) while of utilized crude extracts are mangosteen pericarp (0.97%)

> roselle (0.48%) > red cabbage (0.39%) > Thai berry (0.23%) and blue pea (0.13%),

respectively.

10
 11
ACCEPTED MANUSCRIPT
The incident photo conversion efficiency (IPCE) of all fabricated DSSCs was shown in

Figure 4. All IPCEs of DSSCs utilized by crude and pre-concentrated extracts of all six

samples agrees to their photovoltaic characteristics and UV−Vis spectra. The IPCE plots of

DSSCs sensitized by red cabbage, black rice, Thai berry and blue pea extracts are in

agreement with their UV−Vis spectra in which high conversion efficiency is found in the

PT
wavelength range 500−700 nm. The IPCEs of these four DSSCs sensitized by their

RI
pre−concentrated extracts are higher than those of their crude extracts corresponding to

higher photovoltaic characteristics as listed in Table 2. Flat IPCE curve observed in the

SC
wavelength range 500−700 nm of the crude blue pea extract, agree to its least overall

U
efficiency. AN
4.1.3 Stability of DSSCs
M

The stability of these crude and pre-concentrated extracts based DSSCs was

systematically examined with perfect sealing to avoid leaking and contacting with moisture.
D

Figure 5 represents the plot of decreasing of normalized efficiency versus times recorded
TE

within 816 hrs. It is found that most of DSSCs sensitized by pre−concentrated extracts have
EP

decreased their efficiency up to 40%. In contrast, DSSCs sensitized by crude extracts degrade

completely such as red cabbage and blue pea while by black rice and mangosteen pericarp
C

degrade close to 90%. To gain better understanding, therefore it would be meaningful to

AC

investigate and compare the adsorption affinity of the crude and pre−concentrated dyes on

TiO2 electrode to the photovoltaic performance and stability of DSSCs. The adsorption

experiments of the dyes from roselle and black rice are shown in Figure 6. With increasing

time, the amounts of adsorbed dyes of both roselle and black rice gradually increased. It is

obviously shown that the adsorption affinities of pre−concentrated dyes were much greater

than that of the crude dyes. The less amount of crude dye adsorption is influenced by the

11
 12
ACCEPTED MANUSCRIPT
other components as shown in the concentration decay curves. Therefore, it is confirmed that

photovoltaic performance as well as stability of pre−concentrated dyes are higher than those

of crude dyes indicating by a large amount of dyes adsorbed on TiO2 electrode. In case of

mangosteen pericarp, both J−V and IPCE characteristics show that crude dye performs on

photovoltaic efficiency better than pre−concentrated dye but however, lesser in stability. This

PT
may be due to a number of photoactive dyes contains in crude extract i.g. rutin, α-mangotin

RI
[29] and others which can adsorb and sensitize effectively on the dye−TiO2 system. When

those photoactive dyes are getting rid and larger number of anthocyanins obtained after the

SC
pre-concentration process, the efficiency of DSSC is reduced and quite comparable to other

U
systems. Based on the results, it can be said that the pre-concentration process of anthocyanin
AN
using SPE can reach high anthocyanin content by reducing undesirable contaminants such as

phenolics, proteins, polysaccharides and other flavonoids [32].

M
D

4.2 Computational results

TE

The DFT/TDDFT method was employed in this work to differentiate how these
EP

anthocyanin dyes perform photovoltaic performance in DSSC in electronic scale. The

molecular structures of all anthocyanins found in the samples; Cyanidin (Cy), Delphinidin
C

(Dp), Pn (Peonidin) and Pelagonidin (Pg) (See Table 1) are shown in Figure 7. The aglycone
AC

forms of these molecules were used instead of glycone forms to reduce computational cost.

Their geometries and electronic structures were calculated by DFT and TDDFT in both vacuo

and solvent media.

4.2.1 Ground and excited state oxidation potential and electron injection force

12
 13
ACCEPTED MANUSCRIPT
dye dye*
Ground ( E OX ) and excited ( EOX ) state oxidation potentials and electron injection force

( ∆Ginject ) derived from orbital energies are listed in Table 3. The E OX

dye
of all dyes computed in

vacuo are in this decreasing order: Pg > Cy ∼ Pn > Dp and in methanol are: Pg ∼ Pn > Cy >

dye*
Dp. E OX of all dyes computed in vacuo and methanol are decreasing order: Pg > Cy > Pn >

PT
dye*
Dp and Pn > Pg > Cy > Dp, respectively. E OX of single−point gas−phase geometries

calculation are a bit difference to the geometry optimization in methanol. The most oxidizing

RI
species in both ground and excited states is Dp.

SC
∆Ginject states the spontaneousness of electron injection process from excited state dye to

U
conduction band edge of TiO2. The magnitude of ∆Ginject as term of spontaneousness of the
AN
dyes is in decreasing order: Dp > Cy > Pg > Pn. The spontaneous charge transfer process

from the dye excited state to conduction band of TiO2 requires LUMO energy level of the dye
M

upper than conduction band of TiO2 (−4.0 eV) while HOMO energy lower than reduction

potential energy of the I−/I3− electrolyte (−4.80 eV) requires for the spontaneous charge
D
TE

regeneration. The energy level diagram of the HOMO and LUMO of the dyes based on

optimized gas−phase and methanol−phase geometries and ECB of TiO2 and redox potential
EP

energy of the electrolyte I−/I3− is depicted in Figure 8. The LUMO energy levels of all dyes

computed based on optimized methanol−phase geometries are upper the ECB of TiO2 while
C

their HOMO energy levels are lower than the redox potential energy of electrolyte. Thus, it
AC

can be said that all of the dyes possesses a positive response to charge transfer and charge

regeneration related to photooxidation process.

4.2.2 Electronic structure

The optimized structure of all dyes computed in methanol their isopotential plots of

frontier molecular orbitals are shown in Figure 9. All the ground state energy levels (HOMO,

13
 14
ACCEPTED MANUSCRIPT
HOMO−1, HOMO−2) of all dyes are π type. At HOMO−2, the molecular orbital of Dp is

localized in 2−phenyl ring while of others are delocalized over the entire molecule. At

HOMO−1, the molecular orbital of Cy and Dp is similar, which delocalized over the entire

molecules. At the same time, molecular orbital of Pg and Pn is quite similar which molecular

orbital is observed to be denser in benzopyrilium ring A and 2−phenyl ring. At HOMO state,

PT
molecular orbital of all dyes is in similar pattern which occupied whole molecule as π type.

RI
Anchoring groups of Pg and Pn deduced to be proton and methoxy groups as the molecular

orbital are observed to be localized in these groups at HOMO. At excited state LUMO, all the

SC
dyes show similar pattern of molecular orbital occupied in this state. As observed, for

U
anchoring group on 2−phenyl ring, molecular orbital only occupied at OH2 of all molecules.
AN
This indicates that this hydroxyl group is deduced as the most likely anchoring group that can

provide efficient electron transfer from their excited state to conduction band of TiO2.
M

The ability to generate current in DSSCs sensitized by crude and pre−concentrated

D

extracts of anthocyanin−rich samples is photo−induction whereby electron transfer process

TE

takes place and therefore excitation generates the charge separated state. As observed, the

degree of extension of LUMO molecular orbital towards the hydroxyl group that links the
EP

dye to the TiO2 surface of all dyes are not different but at higher state, degree of extension of

LUMO+1 of Pg and Pn is lower compare to Cy and Dp. This suggests that Pg and Pn will
C

shows lower photo to current compare to that of Cy and Dp.

AC

4.2.3 Deprotonation energy

Deprotonation energy is one of a computed parameter used to predict the binding mode

of interaction between anchoring group of the dye and TiO2. However, the geometrical

structures were optimized in the solvent media but deprotonation energy does not include the

solvated cavity model. In the present work, it is defined as the energy difference between the

14
 15
ACCEPTED MANUSCRIPT
optimized neutral and deprotonated molecules (Eq.7). The ascending order of the

deprotonation energy values obtained by that previous work was not the same order

investigated in this work. The lowest values of the deprotonation energy indicate the most

probable that anchors to the TiO2. Single deprotonation energies of the four anthocyanidins

(see Figure 7) are presented in Table 4. H2 has the smallest deprotonation energy whereas H1

PT
has the highest deprotonation energy observed for Cy and Dp molecules. The deprotonation

RI
energies of three hydroxyl groups on benzopyrilium ring are in between of H1 and H2 and

varied slightly. Pg and Pn have free four hydroxyl groups. In contrast, H5 and H6 of Pg and

SC
Pn have comparable deprotonation energies and smaller than those the others. Thus, either

H5 or H6 of Pg and Pn can potentially bind to TiO2. However, double deprotonation energies

U
have been tested for choosing the best anchoring group. It is indicated that double
AN
deprotonation energy is very high, higher than those single deprotonation [57]. This shows
M

that single deprotonation is favorable for anthocyanin to bind with TiO2.

The efficient binding modes of interaction between four anthocyanidins and TiO2 are
D

shown in Figure 10. Cy will bind bi−dentate mode to TiO2. Dp has three hydroxyl groups on
TE

2−phenyl ring but however, tri−dentate mode will be unfavorable. Bi−dentate mode of Cy is
EP

more favorable compared to tri−dentate mode of Dp due to lack of repulsive effect of the

third hydroxyl group H3 and TiO2 surface. In case of Pg and Pn, even though H5 and H6
C

show less deprotonation energy which means that this two hydroxyl groups are potential to
AC

bind with TiO2 as mentioned above, but binding via H6 group is inefficient for electron

injection because no molecular orbital localized on this group at LUMO state. Anyway,

although molecular orbital occupied on H5 is found both ground and excited states, but

direction of electron transfer is generated from benzopyrilium ring to 2−phenyl ring

corresponds to excitation of ground→excited states. Thus, H5 is also not suitable for efficient

electron transfer. H2 of both Pg and Pn molecules will bind to TiO2. The Pg molecule, of

15
 16
ACCEPTED MANUSCRIPT
course, monodentate mode is favorable. The Pn molecule will bind monodentate or bind

bidentate mode when methyl of methoxy group oriented in less repulsive interaction to TiO2.

Finally, it can be summarized into two main points. Firstly, pre−concentration of dye

solution via SPE method enhances overall efficiency in DSSC by getting rid of other

PT
contaminants and rising anthocynin contents. Dye adsorption was performed to confirm this

conclusion. Secondly, Dp molecule performs the best action based on theoretical aspect.

RI
Conversely, DSSCs sensitized by Dp as major pigment (blue pea) shows the lowest

SC
efficiency based on inefficient binding interaction to TiO2. Cy which is major dye in

magosteen pericarp, roselle, black rice, red cabbage and Thai berry performs the best as

U
sensitizer in DSSC.
AN
5. Conclusions
M

The DSSCs utilized crude and pre−concentrated extracts from six anthocyanin−rich

samples were fabricated and their photo−to−current conversion efficiency and stability were
D
TE

examined. Pre−concentrated extracts were obtained by SPE using C18 cartridge. The results

obviously showed that all pre−concentrated extracts performed on photovoltaic performances

EP

in DSSCs better than crude extracts except mangosteen pericarp. In addition, pre-

concentrated extract−based cells show more stable than those of crude extract−based cells.
C

The adsorption experiment indicates that adsorption affinities of pre−concentrated dyes on

AC

TiO2 electrode are much greater than that of the crude dyes enhancing photovoltaic

performance as well as stability of DSSCs. This suggests that pre−concentration via SPE

method is very attractive for DSSCs based on photovoltaic performance and stability. The

DFT/TDDFT computations suggest that Dp will perform the best in DSSC following by Cy,

16
 17
ACCEPTED MANUSCRIPT
Pg and Pn. However, the best dye is Cy confirmed by both experimental and computational

observations.

Acknowledgements

This work was gratefully acknowledged the Nanotechnology Center (NANOTEC),

PT
NSTDA, Ministry of Science and Technology, Thailand, through its program of Center of

Excellence Network and the Integrated Nanotechnology Research Center (INRC), Khon

RI
Kaen University, Thailand, for financial support to PC through the research grant

SC
no.KKU55/12. Department of Chemistry, Faculty of Science, Khon Kaen University was

acknowledged for purifition and NMR facilities. The Faculty of Engineering, Khon Kaen

U
campus and Institute of Research and Development, Rajamangala University of Technology
AN
Isan, are also acknowledged for research facilities and partial financial support.
M

References

[1] B.O’Regan, M.Grätzel, Nature 353 (1991)737–740.

D

[2] Y. Chiba, A. Islam, Y. Watanabe, R. Komiya, N. Koide, L.Y. Han, DyeJpn. J. Appl.
TE

Phys. 45 (2006) 638–640.

[3] R. Buscainoa, C. Baiocchib, C. Baroloa, C. Medanab, M. Grätzelc, Md.K.
EP

Nazeeruddinc, G. Viscardi,Inorg. Chim. Acta361 (2008) 798–805.

[4] A. Furube, R. Katoh, T. Yoshihara, K. Hara, S. Murata, H. Arakawa, M. Tachiya, J.
Phys. Chem. B 108 (2004) 12588-12592.
C

[5] R. Argazzi, N.Y.M. Iha, H. Zabri, F. Odobel, C.A. Bignozzi, Coord. Chem. Rev. 248
AC

(2004) 1299-1316.
[6] N.M. Gómez–Ortíza, I.A. Vázquez–Maldonadoa, A.R. Pérez–Espadasb, G.J. Mena–
Rejónb, J.A. Azamar–Barriosa, G. Oskam, Sol. Energ. Mat. Sol. C. 94 (2010) 40–44.
[7] E. Yamazaki, M. Murayama, N. Nishikawa, N. Hashimoto, M. Shoyama, O. Kurita,
Sol. Energy 81 (2007) 512–516.
[8] A. Dumbravă, I. Enache, C.I. Oprea, A. Georgescu, M.A. Gîrţu, Dig. J. Nanomater.
Bios. 7 (2012) 339–351.

17
 18
ACCEPTED MANUSCRIPT
[9] G. Calogero, G.D. Marco, S. Cazzanti, S. Caramori, R. Argazzi, A.D. Carlo, C.A.
Bignozzi, Int. J. Mol. Sci. 11 (2010) 254–267.
[10] A.R. Hernandez–Martinez, M. Estevez, S. Vargas, F. Quintanilla, R. Rodriguez, Int. J.
Mol. Sci. 12 (2011) 5565–5576.
[11] R. Espinosaa, I. Zumetaa, J.L. Santanac, F. Martı́nez–Luzardoc, B. Gonzáleza, S.
Docteura, E. Vigil, Sol. Energ. Mat. Sol. C. 85 (2005) 359–369.

PT
[12] G.R.A. Kumara, S. Kaneko, M. Okuya, B. Onwona–Agyeman, A. Konno, K.
Tennakone, Sol. Energ. Mat. Sol. C. 90 (2006) 1220–1226.

RI
[13] N. Shahzad, S.M. Shah, S. Munir, A. Hana, U. Jabeen, E. Nosheen, B. Habib, A.U.
Khan, Z. Hassan, M. Siddiq, H. Hussain, J. Electron. Mater. 44 (2015) 1167–1174.

SC
[14] S. Hao, J. Wu, Y. Huang, J. Lin, Natural dyes as photosensitizers for dye–sensitized
solar cell, Sol. Energ. 80 (2006) 209–214.
[15] P.M. Sirimanne, M.K.I. Senevirathna, E.V.A. Premalal, P.K.D.D.P. Pitigala, V.

U
Sivakumar, K. Tennakone, J. Photochem. Photobiol. A 177 (2006) 324–327.
AN
[16] A.S. Polo, N.Y.M. Iha, Sol. Energ. Mat. Sol. C. 90 (2006) 1936–1944.
[17] K. Wongcharee, V. Meeyoo, S. Chavadej, Sol. Energ. Mat. Sol. C. 91 (2007) 566–571.
M

[18] K. Tennakone, A.R. Kumarasinghe, G.R.R.A. Kumara, K.G.U. Wijayantha, P.M.

Sirimanne, J. Photochem. Photobiol. A 108 (1997) 193–195.
D

[19] D. Zhang, S.M. Lanier, J.A. Downing, J.L. Avent, J. Lum, J.L. McHale, J. Photochem.
Photobiol. A 195 (2008) 72–80.
TE

[20] M.S. Roy, P. Balraju, M. Kumar, G.D. Sharma, Sol. Energ. Mat. Sol. C. 92 (2008)
909–913.
EP

[21] J.M.R.C. Fernando, G.K.R. Senadeera, Curr. Sci. 95 (2008) 663–666.

[22] G. Calogero, G. Di Marco, Sol. Energ. Mat. Sol. C. 92 (2008) 1341–1346.
C

[23] Q. Dai, J. Rabani, J. Photochem. Photobiol. A 148 (2002) 17–24.

[24] N.J. Cherepy, G.P. Smestad, M. Grätzel, J.Z. Zhang, J. Phys. Chem. B 101 (1997)
AC

9342–9351.
[25] P. Luo, H. Niu, G. Zheng, X. Bai, M. Zhang, W. Wang, Spectrochim. Acta Part A 74
(2009) 936–942.
[26] S. Furukawa, H. Iino, T. Iwamoto, K. Kukita, S. Yamauchi, Thin Solid Films 518
(2009) 526–529.
[27] A.R. Hernández–Martínez, M. Estevez, S. Vargas, F. Quintanilla, R. Rodríguez, 1st
International Congress on Instrumentation and Applied Science 10 (2012) 38–47.

18
 19
ACCEPTED MANUSCRIPT
[28] S. Ito, T.Saitou, H. Imahori, H. Ueharad, N. Hasegawa, Energy Environ. Sci 3 (2010)
905.
[29] H. Zhou, L. Wu, Y. Gao, T. Ma, J. Photochem. Photobiol A 219 (2011) 188–194.
[30] W. Maiaugree, S. Lowpa, M. Towannang, P. Rutphonsan, A. Tangtrakarn,
S.Pimanpang, P. Maiaugree, N. Ratchapolthavisin, W. Sang–aroon, W. Jarernboon, V.
Amornkitbamrung, Scientific report 5:15230 (2015) 1 11.

PT
[31] C.Y. Li, H.W. Kim, S.R. Won, H.K. Min, K.J. Park, M.S. Ahn, H.I. Rhee, J. Agric.
Food Chem 56 (2008) 11413.

RI
[32] S.A. Mozaffari, M. Saeidi, R. Rahmanian, Spectrochim. Acta Part A142 (2015) 226–
231.

SC
[33] Extraction, isolation and purification of anthocyanins. Unit F2.1, Chapter 2:
Anthocyanins, In Current Protocols in Food Analytical Chemistry. S. King, M. Gates &
L. Scalettar (Ed.). John Wiley & Sons, Inc. New York, NY.

U
[34] S. Mu, Y. Long , S.-Z. Kang , J. Mu. Catalysis Communications 11 (2010) 741–744.
AN
[35] A.D. Becke, Density–functional thermochemistry. III. The role of exact exchange, The
J. Chem. Phys., 98 (1993) 5648–5652.
M

[36] C. Lee, W. Yang, R.G. Parr, Phys. Rev. B 37 (1988) 785.

[37] J. Tomasi, B. Mennucci, E. Cances, J. Mol Struc t THEOCHEM, 464 (1999) 211–226.
D

[38] E. Cancès, B. Mennucci, J. Tomasi, J. Chem. Phys. 107 (1997) 3032–3039.

TE

[39] B. Mennucci, J. Tomasi, J. Chem. Phys. 106 (1997) 5151–5158.

[40] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman,
G. Scalmani, V. Barone, B. Mennucci, G.A. Petersson, H. Nakatsuji, M. Caricato, X.
EP

Li, H.P. Hratchian, A.F. Izmaylov, J. Bloino, G. Zheng, J.L. Sonnenberg, M. Hada, M.
Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O.
C

Kitao, H. Nakai, T. Vreven, J.A. Montgomery, Jr, J.E. Peralta, F. Ogliaro, M. Bearpark,
AC

J.J. Heyd, E. Brothers, K.N. Kudin, V.N. Staroverov, R. Kobayashi, J. Normand, K.

Raghavachari, A. Rendell, J.C. Burant, S.S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J.
M. Millam, M. Klene, J.E. Knox, J.B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R.
Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W.
Ochterski, R.L. Martin, K. Morokuma, V.G. Zakrzewski, G.A. Voth, P. Salvador, J.J.
Dannenberg, S. Dapprich, A.D. Daniels, O. Farkas, J.B. Foresman, J.V. Ortiz, J.
Cioslowski, D.J. Fox, Gaussian09, Revision A.02, Gaussian Inc, Wallingford, CT,
2009.

19
 20
ACCEPTED MANUSCRIPT
[41] R. Dennington, T. Keith, J. Millam, GaussView, Semichem Inc., Shawnee Mission,
KS, 2009.
[42] R. Katoh, A. Furube, T. Yoshihara, K. Hara, G. Fujihashi, S. Takano, S. Murata, H.
Arakawa, M. Tachiya, J. Phys. Chem. B 108 (2004) 4818 4822.

[43] W. Sang–aroon, S. Saekow, V. Amornkitbamrung, J. Photochem. Photobiol. A 236

(2012) 35–40.

PT
[44] J.L. Goodman, K.S. Peters, J. Am. Chem. Soc. 108 (1986) 1700 1701.

[45] J.B. Asbury, Y.Q. Wang, E. Hao, H.N. Ghosh, T. Lian, J. Phys. Chem. C, 115 (2011)

RI
3753–3761
[46] J. Preat, C. Michaux, D. Jacquemin, E.A. Perpete, J. Phys. Chem. C 113 (2009) 16821–

SC
16833.
[47] T. Koopmans, Physica 1(1933) 104–113.

U
[48] R. Byamukama, J. Namukobe, M. Jordheim, Ø.M. Andersen, B.T. Kiremire, Plumeria
rubra. Sci. Hortic. 129 (2011) 840–843.
AN
[49] C.C. Trout, J.D. Kubicki, Geochim. Cosmochim. Acta 70 (2006) 44–55.
[50] M.H. Buraidah, L.P. Teo, S.N.F. Yusuf, M.M. Noor, M.Z. Kufian, M.A. Careem, S.R.
M

Majid, R.M. Taha, A.K. Arof, Int. J. Photoenergy, (2011) 1 11.

[51] Z. Hou, P. Qin, Y. Zhang, S. Cui, G. Ren, Food. Res. Int. 50(2013) 691–697.
D

[52] N. Terahara, N. Saito, T. Honda, K. Toki, Y. Osajima, Phytochemistry. 29 (1990) 949–

TE

953.
[53] G.J. McDougall, S. Fyffe, P. Dobson, D. Stewart, Phytochemistry. 68 (2007) 1285–
1294.
EP

[54] S. Jorjong, L. Butkhup, S. Samappito, Food Chem. 181 (2015) 248–255.

[55] A. Chaovanalikit, A. Mingmuang, T. Kitbunluewit, N. Choldumrongkool, J. Sondee, S.
C

Chupratum, S. Int. Food Res. J. 19 (2012) 1047–1053.

AC

[56] A.S. Zarena, K.U. Sankar, Food Chem. 130 (2012) 665–670.
[57] P. Ekanayake, M.R.R. Kooh, N.T.R.N. Kumara, A. Lim, M.I. Petra, N.Y. Voo, C.M.
Lim, Chem. Phys. Lett. 585 (2013) 121–127.

20
 1
ACCEPTED MANUSCRIPT
Table 1 Absorption maximum wavelength and major anthocyanins found in each sample

Samples λmax (nm) Anthocyanins

in MeOH/HCl
crude isolated

Mangosteen 525.0 528.0 Cy−3−sophoroside (major)

pericarp Cy−3−glucoside

PT
Pg−3−glucoside
Red cabbage 528.0 526.0 Cy−3− (sinapoyl) diglucoside−5−glucoside (major)
Cy−3−glucoside

RI
Cy−3,5−diglucoside
Black rice 529.0 532.0 Cy−3−glucoside (major)

SC
Pn−3−glucoside
Roselle 531.5 536.5 Cy−3−glucoside (major)
Dp−3−glucoside

U
Thai berry 535.0 538.5 Cy−3−glucoside (major)
AN
Blue pea 541.0 529.5 Ternatin (or Dp as aglycone form) (major)
M
D
TE
C EP
AC

1
 2
ACCEPTED MANUSCRIPT
Table 2 Photovoltaic performances of all fabricated DSSCs

Samples JSC (mA/cm2) VOC (V) FF %η

Pre- Pre- Pre- Pre-
Crude Crude Crude Crude
conc conc conc conc
Mangosteen pericarp 3.63 1.57 0.44 0.41 0.60 0.37 0.97 0.54
Roselle 3.09 2.91 0.27 0.46 0.56 0.53 0.48 0.71

PT
Red cabbage 1.40 3.84 0.51 0.36 0.55 0.50 0.39 0.71
Black rice 1.08 3.04 0.42 0.31 0.59 0.56 0.27 0.53
Thai berry 1.52 3.92 0.42 0.27 0.36 0.45 0.23 0.47

RI
Blue pea 0.61 2.27 0.48 0.28 0.43 0.53 0.13 0.33

U SC
AN
M
D
TE
C EP
AC

2
 3
ACCEPTED MANUSCRIPT
dye
Table 3 The oxidation potential energies for ground state ( E OX ), oxidation potential energies
dye*
for excited state ( E OX ), and free energy change for electron injection ( ∆G inject ) of all dyes
performed at B3LYP/6−31+G(d,p) level
a a a
dye
E OX dye*
E OX ∆G inject
Dyes
Vacuob MeOHc Vacuob MeOHc Vacuob MeOHc
Cy 9.197 6.449 6.719 3.938 2.719 −0.062
Dp 9.170 6.422 6.634 3.934 2.634 −0.066

PT
Pg 9.360 6.503 6.748 3.948 2.748 −0.052
Pn 9.197 6.503 6.663 3.969 2.663 −0.031
a

RI
In eV.
b
Optimized in gas phase.

SC
c
Optimized in methanol .

U
AN
M
D
TE
C EP
AC

3
 4
ACCEPTED MANUSCRIPT
Table 4 Single deprotonation energies of the anthocyanidins performed at
B3LYP/6−31+G(d,p) level (atomic labeling in Figure 6)

Dyes Deprotonation energies (kcal mol-1)

−H1 −H2 −H3 −H4 −H5 −H6
Cy 294.9 279.1 − 283.2 282.3 282.4
Dp 294.3 280.1 288.5 282.6 282.1 282.2
Pg − 284.0 − 283.3 282.4 282.5

PT
Pn − 282.5 − 283.1 282.2 282.4

RI
U SC
AN
M
D
TE
C EP
AC

4
 ACCEPTED MANUSCRIPT
List of Figures of the illustrations

R
OH

PT
HO O

RI
R

U SC
OR
AN
OR
M

Figure 1. Molecular structure of anthocyanins

D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
Scheme1 Schematic representation for the pre-concentration procedure of anthocyanin
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M

Figure 2. Absorption spectra in methanolic HCl solution (pH=1) between 400 to 700 nm.

(See the color version for more clarity)

D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M

Figure 3. Current density−voltage characteristics of fabricated solar cells. (See the color

version for more clarity)

D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M

Figure 4. The plot of IPCE of all fabricated solar cells. (See the color version for more

clarity)
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M

Figure 5. The stabilities of all fabricated DSSCs in term of normalized overall efficiency and

variable times. (See the color version for more clarity)

D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE

Figure 6. Adsorption efficiency of anthocyanins in crude and pre-concentrated extracts on

TiO2 electrode.
C EP
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M

Figure 7. Molecular structures and atomic labeling of the anthocyanidin molecules.

D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP

Figure 8. Energy level diagram showing various electronic energy levels of the dyes versus
C

experimental redox potential (−4.8 eV) and conduction band of TiO2 (−4.0 eV). The left
diagram was obtained from geometry optimization in gas phase whilst the right diagram was
AC

obtained from geometry optimization in methanol phase.

 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
C EP
AC

Figure 9. Isopotential plots of frontier molecular orbital energy levels of all dyes with
isovalue contour 0.02.
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
Figure 10. Effective binding mode interaction between anthocyanin molecule and TiO2 (a)
Cy−TiO2 and (b) Dp−TiO2 prefer bidentate mode (c) Pg prefers monodentate mode and Pn
M

prefers either (d) bidentate or (e) monodentate mode.

D
TE
C EP
AC
 ACCEPTED MANUSCRIPT
HIGHLIGHTS

• Natural dyes extracted from mangoesteen pericarp, roselle, red cabbage, Thai berry,
black rice and blue pea.
• Pre-concentrated extracts obtained by solid phase extraction using C18 cartridge.
• DSSCs were sensitized by crude and pre-concentrated extracts.
• Pre-concentrated extracts perform on photovoltaic performance better than those

PT
crude extracts.
• The best performance cell is DSSCs sensitized by pre-concentrated roselle and red

RI
cabbage extracts (η 0.71%).
• The best dye performs on photovoltaic characteristics is cyanidin.

U SC
AN
M
D
TE
C EP
AC