Photoproducts of carminic acid formed by a composite from Manihot dulcis waste

Article abstract of DOI:10.1016/j.foodchem.2014.10.085

Carbon-TiO₂ composites were obtained from carbonised Manihot dulcis waste and TiO₂ using glycerol as an additive and thermally treating the composites at 800 °C. Furthermore, carbon was obtained from manihot to study the adsorption,

Full text of DOI:10.1016/j.foodchem.2014.10.085

Food Chemistry 173 (2015) 725–732
Contents lists available at ScienceDirect
Food Chemistry
journal homepage: www.elsevier.com/locate/foodchem
Photoproducts of carminic acid formed by a composite from Manihot
dulcis waste
a,b
c
a,
⇑
Cynthia M. Antonio-Cisneros , Martín M. Dávila-Jiménez , María P. Elizalde-González
Esmeralda García-Díaz
a
,
a
Centro de Química, Instituto de Ciencias, Universidad Autónoma de Puebla, Ciudad Universitaria, Edif. 103H, Puebla, Pue. 72570, Mexico
Instituto de Agroingeniería, Universidad del Papaloapan, Campus Loma Bonita, Oaxaca, Mexico
Facultad de Ciencias Químicas, Universidad Autónoma de Puebla, Ciudad Universitaria, Edif. 105, Puebla, Pue. 72570, Mexico
b
c
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 29 July 2014
Received in revised form 6 October 2014
Accepted 16 October 2014
Available online 22 October 2014
Carbon–TiO2 composites were obtained from carbonised Manihot dulcis waste and TiO2 using glycerol as
an additive and thermally treating the composites at 800 °C. Furthermore, carbon was obtained from
manihot to study the adsorption, desorption and photocatalysis of carminic acid on these materials. Car-
minic acid, a natural dye extracted from cochineal insects, is a pollutant produced by the food industry
and handicrafts. Its photocatalysis was observed under different atmospheres, and kinetic curves were
measured by both UV–Vis and HPLC for comparison, yielding interesting differences. The composite
was capable of decomposing approximately 50% of the carminic acid under various conditions. The reac-
tion was monitored by UV–Vis spectroscopy and LC-ESI-(Qq)-TOF-MS-DAD, enabling the identification of
Keywords:
Natural red 4
Carminic acid
Photolysis
Photocatalysis
Product identification
Manihot waste
some intermediate species. The deleterious compound anthracene-9,10-dione was detected both in N
and air atmospheres.
2
Ó 2014 Elsevier Ltd. All rights reserved.
2
TiO /carbon composite
1
. Introduction
Fernandez-Lahore, 2007). In contrast to adsorption, heterogeneous
photocatalysis can be regarded as a water purification process that
can achieve mineralisation by producing, at times, toxic intermedi-
ates during degradation. Incidentally, authors reported observing a
higher toxicity after the biodegradation of CA than before the
microbiological treatment of effluent (Arroyo-Figueroa et al.,
2011). The fact that hydroxylated anthraquinones have presented
significant toxicity is important because these compounds can be
photoproducts of CA (Brack et al., 2003). This fact makes the iden-
tification of the compounds produced during the degradation of CA
necessary.
In the food industry, CA is considered not only a colorant that
resists degradation over time but also one of the most light-stable,
heat-stable and oxidation-resistant dyes among all natural dye-
stuffs. However, Jørgensen and Skibsted (1991) found that the
photolability of CA increased with deprotonation and was
enhanced with irradiation at 254 nm. Recently, Gosetti,
Chiuminatto, Mazzucco, Mastroianni, and Marego (2015) studied
the effect of sun irradiation on CA in aqueous solution and in six-
teen different beverages to mimic the action of sunlight during
Carminic acid (CA) is a natural dye extracted from cochineal
insects that can yield shades of red. The Food Standards Agency
in Europe and the United States Food and Drug Administration
(
FDA) have not banned its use in food. CA is also used in histology,
as well as the coloring of pharmaceuticals, cosmetics, plastics and
fabrics. Particularly in Oaxaca (Mexico), CA is also the most impor-
tant insect-derived dye used in the production of rugs and handi-
crafts. During the fabrication of both food and hand-woven
carpets, significant amounts of CA are released in the food industry
and artisanal carpeting effluents. Fortunately, CA showed negative
genotoxicity in a somatic mutation and recombination test of Dro-
sophila melanogaster (Sarikaya, Selvi, & Erkoç, 2012).
If we consider the problem of removing CA by the most com-
mon technique used for dyes, adsorption (Rangabhashiyam, Anu,
&
using activated carbon. In recent years, the application of adsor-
bents in this respect have focused on the recovery and purification
of CA (Bibi, Galvis, Grasselli, & Fernández-Lahore, 2012; Cabrera &
Selvaraju, 2013), the only reports of such removal are early ones
2
transport. In the presence of TiO , authors degraded CA using
UVA (366 nm) irradiation and demonstrated the adsorption of
the dye onto the photocatalyst (Baran, Makowski, & Wardas,
2008). Furthermore, with the addition of H O , UV irradiation
2 2
⇑
Corresponding author.
E-mail address: maria.elizalde.uap.mx@gmail.com (M.P. Elizalde-González).
http://dx.doi.org/10.1016/j.foodchem.2014.10.085
308-8146/Ó 2014 Elsevier Ltd. All rights reserved.
0

7
26
C.M. Antonio-Cisneros et al. / Food Chemistry 173 (2015) 725–732
caused a 60% decoloration of CA under optimised conditions
Körbahti & Rauf, 2009). In spectroscopic studies, CA also under-
2.3. Characterisation of the composite
(
went degradation in its solid state by exposure to visible light
under anoxic conditions (Koperska, Łojewski, & Łojewska, 2011).
To date, the advancements in photocatalysis for the degradation
of target substances are undeniable, due to the development of
semiconductor materials. The photocatalytic removal of dyestuffs
from water has been reviewed (Ahmed, Rasul, Martens, Brown, &
Hashib, 2011; Akpan & Hameed, 2009), but the photocatalytic deg-
The morphology and surface composition of the composite
were determined using an Oxford Instruments INCAX-act Energy
350 EDS spectrometer coupled to a JEOL JSM 6610 LV scanning
electron microscope (SEM). The BET surface area was measured
by nitrogen adsorption at 77 K using an Autosorb-1 from Quanta-
chrome Instruments (Boynton Beach, FL) after out-gassing at
573 K for 12 h. The specific surface area was calculated using the
Brunauer–Emmett–Teller (BET) method, and the pore size distribu-
tion was estimated by the Barrett–Joyner–Halenda method (BJH).
The extent of micro- and mesoporosity was calculated using the
t-method. The specific surface area of the raw residues was deter-
mined by the adsorption of methylene blue using a molecular area
2
radation of CA on TiO /carbon materials has not yet been
addressed.
Combinations of TiO
the field of photocatalysis (Leary & Westwood, 2011). In recent
years, the application of these carbon–TiO materials in photoca-
2
and carbon are common composites in
2
2
talysis has gained widespread attention because (i) the solids are
easily recovered after use in a wide range of applications (Foo &
of 1.35 nm in the vertical orientation (Antonio-Cisneros, 2010).
The density of the dry materials was determined in triplicate using
tarred vessels. The point of zero charge (pHpzc) was measured by
methodology reported elsewhere (Bourikas, Vakros, Kordulis, &
Lycourghiotis, 2003).
Hameed, 2010) and (ii) synergistic effects occur between TiO
and carbon (Matos, Chovelon, Cordero, & Ferronato, 2009).
2
Only a few studies have been devoted to the preparation of
TiO /carbon materials using lignocellulosic residues, such as bam-
boo leaves (Huang et al., 2011), pine sawdust (Asiltürk & Sß ener,
012), maize corncob (Sabinas-Hernández, 2011), canola hull
Mahmoodi, Arami, & Zhang, 2011), corn straw powder (Chen
2
2.4. Adsorption and desorption experiments
2
(
Adsorption tests were performed at 25 °C in thermostatted
batch experiments. The solid materials, raw manihot residues, car-
bon, and TiO /carbon composite with a particle size of 0.25 mm
2
et al., 2012), and manihot residues (Antonio Cisneros, 2010) as car-
bon precursors.
Our goal was to measure the sensitivity of aqueous solutions of
CA to UV light under various conditions as well as the photolability
of CA on TiO immobilised on carbon obtained from dried stem res-
2
were dried for 24 h at 75 °C. Then, 100 mg of each material were
weighed into polycarbonate cylindrical cells with lids and con-
tacted at 25 °C with 3 mL of CA solutions with concentrations in
À1
idues of Manihot dulcis (sweet cassava). The present study focused
the range of 100–3000 mg L . The pH was not adjusted and was
on the LC–MS identification of intermediates produced by the irra-
measured to be 5.7 in deionised water. The solution obtained after
24 h, resulting from adsorption equilibrium, was separated from
the exhausted adsorbent, centrifuged (12,000 rpm) and then ana-
lysed with a Hach spectrophotometer at 495 nm (Hach Company,
Loveland, CO). The adsorbed amount a was determined by measur-
diation of CA in the presence of a TiO
2
/C800 composite rather than
on the assessment of the mineralisation of CA.
2
. Experimental
ing the final concentration C
formula a = (C )V/m, where C
À C
CA, V is the volume added (in litres), and m is the weighed mass
f
of CA in the solution and applying the
0
f
0
is the initial concentration of
2.1. Chemicals
(
in grams).
Carminic acid (C.I. 75470), also called crimson lake, cochineal,
Desorption measurements were carried out as follows. A satu-
natural red 4, or E120, was purchased from Sigma Aldrich (St.
Louis, MO; CAS 1260-17-9). The purity of CA, determined by HPLC
was 94%. Glycerol (87% GR) was purchased from Merck (Darms-
tadt, Germany). Titanium oxide aeroxide P25, batch number
rated sample was separated from the solution by filtration. Then,
it was dried at 90 °C for 12 h. Afterwards, the dried sample was
weighed and placed in contact with 3 mL of deionized water in
polycarbonate cylindrical cells at 25 °C for 24 h. The desorbed
amount ades was determined by measuring the concentration of
CA in the solution produced by desorption Cdes and applying the
formula ades = (CdesÁV)/m, where V is the volume added (in L) and
m is the weighed mass (in grams). To determine the equilibrium
1
036070711 (Degussa, Parsippany, NJ) was used for immobilisa-
tion on manihot carbon and was also used for comparative pur-
poses. The natural pH value of the CA solutions was 5.7.
2
2.2. Immobilization of TiO on carbon
2
concentrations, calibration curves with r = 0.999 were obtained
Raw residues of M. dulcis stems without the rind or pith
0.45 mm sieved particles) dried at 70 °C for 24 h were used as
the carbon precursor (Antonio Cisneros & Elizalde-González,
010). The TiO /carbon composite was fabricated as follows. A
glycerol suspension (10 mL) of manihot particles (10 g) was pre-
pared by mixing with a TiO aqueous slurry (10 mL). The resultant
for both the adsorption and desorption experiments.
(
2.5. Photolysis and photocatalytic experiments
2
2
For the irradiation experiments performed on the composite
and TiO (Degussa P25), a 15-mL temperature-controlled quartz
2
2
homogeneous suspension was maintained in a combustion boat at
room temperature for 24 h. The temperature regime of the car-
bonisation consisted of two heating ramps, viz. from room temper-
reactor from Ace Glass Inc. (Vineland, NJ) was used for the exper-
1
iments involving irradiation at 254 nm. A 2 -inch Pen-Ray light
8
source (5.5 W) from Ultra-Violet Products (Upland, CA) was
immersed into the reactor in a quartz jacket. The reactor was
equipped with a water jacket to filter out the IR irradiation of the
lamp and to control the temperature of the solution. A flow of chro-
matographic-grade dry air or nitrogen was allowed to bubble into
the bulk solution. The entire assembly was enclosed in a dark
chamber, and the lamp was switched on after 30 min to establish
adsorption equilibrium. In each experiment, 10 mg of the compos-
À1
ature to 50 °C at 4 °C min , then 1 h at 50 °C and from 50 °C to
À1
8
00 °C at 4 °C min . Then, the residues were maintained at the
maximum temperature for 4 h. The oven was then allowed to cool
to room temperature. A horizontal tubular furnace (Carbolite Fur-
naces Ltd, Hope Valley, UK) with a quartz reactor was used. A car-
2
bon sample not containing TiO , denoted C800, was used for
comparison and was obtained by carbonising raw manihot parti-
cles. The same horizontal tubular furnace and heating program
were used for the C800 sample as for the composite preparation.
À1
ite were added to 10 mL of a solution containing 300 mg L of CA.
Aliquots were collected at regular time intervals during irradiation

C.M. Antonio-Cisneros et al. / Food Chemistry 173 (2015) 725–732
727
and were centrifuged for 45 min before analysis. Samples of all of
the CA solutions subjected to photocatalytic testing with the com-
posite and TiO were analysed using a Hach DR 5000 UV–Vis spec-
trophotometer over the wavelength range of 200–800 nm.
3.2. Adsorption of carminic acid
2
Fig. 1 illustrates the isotherms describing the adsorption of CA
onto TiO , the raw manihot residues and the carbon sample C800
after 24 h of contact. The values of the Freundlich constant (k
2
F
)
2.6. Analysis of the photocatalysis products
À1
À1 1/n
were 19.7 and 4.0 mg g /(mg L
)
2
for TiO and carbon C800,
respectively. According to the Freundlich equation, the energy of
adsorption decreases with the occupation of active sites. Values
of 1/n less than one indicate an adsorbent with high energetic het-
The collected aliquots were analyzed by liquid chromatography
using an LC–MS instrument (Series 1260 chromatograph coupled
with an ESI-(Qq)-TOF-MS 6520 detector; Agilent Technologies,
Santa Clara, CA). Separation was carried out at 25 °C using a Puro-
2
erogeneity. Both the semiconductor TiO and carbon C800 exhib-
ited similar heterogeneity factors of 0.23 and 0.26, respectively.
The corresponding saturation capacity was calculated by the
non-linear form of the Langmuir equation, and the respective
spher StarRP-18e (5
Darmstadt, Germany). Gradient elution at 1 mL min with a
l
m, 150 Â 4.6 mm) column from Merck
–1
(
methanol/water + 0.1% HCOOH (20–70%) mobile phase between 0
and 15 min and 90% methanol up to 20 min was performed. The
methanol was supplied by Burdick–Jackson (Muskegon, MI), and
the HCOOH was obtained from Aldrich (St. Louis, MO). The injec-
À1
À1
2
results were 70 mg g for TiO and 40 mg g for carbon C800.
Our materials were saturated with lower amounts of CA than those
required to saturate a weak anion exchange resin (Cabrera &
Fernandez-Lahore, 2007) and polymeric particles (Bibi et al.,
tion volume was 100 lL.
2
012). In contrast, the extent of CA adsorbed by the TiO
bon was higher than that adsorbed by glass powder (Atun & Hisali,
003). The adsorption isotherms in Fig. 1 demonstrate that both
2
and car-
ESI-TOF-MS was performed in negative ionisation mode
À1
(
ESI(À)) with nitrogen as the drying gas at 11 L min while apply-
2
ing a TOF fragmentor voltage of 175 V, a capillary voltage of 3500 V
over an m/z range of 50–500, a gas temperature of 350 °C, and a
nebuliser pressure of 60 psi. The concentration decay of CA was
the semiconductor TiO
dye. Moreover, the desorption of 4% (0.3
2
and the carbon separately retained the
À2
lmol m ) CA from car-
bon is feasible with water (see inset in Fig. 1). This property allows
CA to concentrate on the surface of the composite by adsorption
2
obtained from a calibration curve with r = 0.9957 and a limit of
confidence of 95%. The calibration curve was determined over a
and allows for the further decomposition of desorbed CA by TiO
under irradiation.
2
À1
linear concentration range from 1 to 100 mg L of CA. Solutions
above the concentration of the calibration range were diluted.
2
The strong adsorption of CA onto TiO can be explained by the
same mechanism as that exhibited by anionic azo dyes, and
depending on the position and number of ionised hydroxyl groups,
mono- and bidentate adsorption complexes can be formed (Bao
et al., 2013). Adsorption occurs as a result of interactions between
3
. Results and discussion
3.1. Description of the TiO /C800 composite
2
the negatively charged CA (pKa,1 = 1.6 ± 0.2) and the TiOH+
semi-
conductor sites at a pH of 5.7, which is slightly lower than the
pHpzc of 6.25 of titanium dioxide (Behnajady, Yavari,
2
The TiO
higher than the densities of the corresponding precursors, i.e.,
the raw residue and TiO . The composite’s surface (pHpzc = 3.5)
was remarkably acidic compared with that of the carbonaceous
support (pHpzc = 11.4).
The immobilisation of the TiO
characteristics of the support prepared using manihot residues, as
demonstrated by measuring the N adsorption–desorption iso-
therms. According to the IUPAC classification system, the carbon
sample C800 displayed a hybrid isotherm (figure not shown here)
of type I–II, which is typical for materials with a considerable
degree of microporosity. This nitrogen adsorption isotherm con-
2
/C800 composite was a black powder. Its density was
&
2
Modirshahla, 2014). Although other reported values for the pKa,1
of CA exist, 2.81 (Jørgensen & Skibsted, 1991; Rasimas, Berglund,
& Blanchard, 1996) and 3.13 (Reyes-Salas et al., 2011), all of the
corresponding studies support the fact that CA exists in an anionic
form in a solution of deionised water without pH adjustment. The
same explanation applied to the adsorption of CA on C800, in
which pH < pHpzc once again because the carbon exhibits basic
type H (Valix, Cheung, & McKay, 2004) character with a pHpzc of
11.4. However, the adsorption isotherms in Fig. 1 show that the
2
on carbon affected the textural
2
trasted significantly with the isotherm exhibited by the bare TiO
2
before immobilisation, whereas the isotherm of the composite
occupied an intermediate position. As expected, the presence of
TiO2
0
raw
25
C800
50
crystalline TiO
2
and its agglomerates drastically reduced the spe-
2
À1
cific surface area of the composite (200 m g ) relative to that of
2
À1
2
1
0
the carbonaceous support (434 m g ), mainly due to the obstruc-
tion of the carbon micropores. This reduction was confirmed by a
decrease in the micropore volume from 61% to 28%.
1
00
8
0
The surface composition obtained from the EDS spectrum of the
composite indicated the presence of titanium, carbon and the
intrinsic elements of the plant manihot. An SEM micrograph of
the composite demonstrated that isolated TiO nano-agglomerates
2
were attached to the surface of the carbon. This finding suggests
that adsorption and photocatalysis phenomena may occur simulta-
60
4
2
0
0
0
neously due to the low loading of TiO
2
onto the carbon. In contrast,
in the studies of Mahmoodi et al. (2011), it was observed that TiO
2
particles entirely covered the surface of sodium alginate, and that
composite was used in the investigation of the adsorption and deg-
radation of the anthraquinonic dye acid green 25. By using glyc-
erol, Sreethawong, Ngamsinlapasathian, and Yoshikawa (2012)
0
1000
, mg L
2000
-1
C
equil
Fig. 1. Adsorption and desorption (inset) isotherms of CA from an aqueous solution
on TiO raw manihot residues, and the carbonaceous support C800 of the
composite.
observed the formation of uniform TiO
of clusters.
2
nanoparticles in the form
2
,

7
28
C.M. Antonio-Cisneros et al. / Food Chemistry 173 (2015) 725–732
carbon C800 adsorbs less CA than does TiO
surface area is nearly eight times larger than that of TiO
ble explanation is the degree of microporosity of the carbon. The
2
, although its specific
AC initial solution
Photolysis in air
Photocatalysis with:
Composite, air (15 C)
Composite, N2
2
. A possi-
2
molecular size of CA (
x = 4.57 nm , Antonio-Cisneros, 2010) makes
o
it difficult for it to penetrate the micropores, which contributes to
6
1% of the specific surface area of C800.
Finally, we aimed to demonstrate the relative adsorption affin-
Composite, air
ities of the semiconductor TiO
2
, carbon, and the composite upon
TiO , air
2
À1
the adsorption of CA at moderate concentrations (300 mg L ) in
the short period planned for the photolysis and photocatalytic
experiments. When a CA solution was measured photometrically
after 90 min of contact with the three different materials, the band
at 450–550 nm in Fig. 2A indicates that adsorption occurred on the
3
2
1
00
00
00
A
B
2
Degussa TiO to a greater extent than on the carbon sample C800
(
Fig. 2B) and the composite (Fig. 2C) under the same conditions.
0
This finding corroborates the results presented in Fig. 1. Due to
the basicity (pHpzc = 11.4) of the carbon, a bathochromic effect on
the Vis-band of the CA solutions was observed (Fig. 2B).
3
2
00
00
3
.3. Photodegradation of CA under UV-lamp irradiation in the presence
100
0
of a TiO /C800 composite
2
Fig. 2D–F depicts the UV–Vis spectra of the irradiated CA solu-
tions at 254 nm after 90 min of contact with the composite under
different temperature and environmental conditions. As shown,
0
20
40
60
80
100
Time (min)
Fig. 3. Concentration decay of CA subjected to photolysis and photocatalysis at
0 °C under different atmospheres in the presence of the composite and TiO . The
2
the prepared TiO /C800 composite exhibited a low decomposition
5
2
rate toward the dye in air at 15 °C (Fig. 2D). At 50 °C, CA decom-
posed readily and to an even greater extent in air (Fig. 2E) than
in nitrogen (Fig. 2F), due to the presence of oxygen. These spectra
allow for the monitoring of the photolytic and photocatalytic
decomposition rate of CA under the conditions studied, as shown
in Fig. 3A. The photolysis and photocatalysis kinetic curves exhib-
ited pseudo-first order behaviour, and the parameters were calcu-
lated by non-linear regression. Table 1 shows the values of the
rate constants, which indicate that the photocatalysis of the CA
was favoured at the applied temperature, as expected. For compar-
ison, a photocatalytic experiment using the bare semiconductor
kinetic curves were measured by UV–Vis at 432 nm (A) and by LC/QTOF-MS (B).
Table 1
Kinetic rate constants of the pseudo-first order photodegradation of CA measured by
UV–Vis and LC-ESI/QTOF-MS.
À2
À1
Process
System conditions
k
1
 10 (min
)
measured by
UV–Vis
1.1
LC–MS
Photolysis
In air at 50 °C
2.4
–
2.4
4.2
52.7
Photocatalysis With composite in air at 15 °C 0.2
With composite in N at 50 °C 0.7
With composite in air at 50 °C 1.0
TiO
posite in air was higher than that in the presence of N
highest rate constant was obtained in the experiment with TiO
which was not surprising. However, the advantage of using the
2
is also presented. The degradation rate exhibited by the com-
2
2
, and the
2
,
With TiO
2
in air at 50 °C
2.6
5
au
A
D
t0
t0
9
0 min
90 min
t0
B
C
E
F
t0
90 min
90 min
t0
t0
90 min
90 min
200 300 400 500 600 700 200 300 400 500 600 700 800
Wavelength (nm)
À1
Fig. 2. UV–Vis spectroscopic monitoring of a solution of CA (300 mg L ) during adsorption (A–C) on TiO
2
(A), carbon C800 (B) and the composite (C) and subjected to
(F).
photocatalysis (D–F) for 90 min in the presence of the composite at 15 °C in air (D), at 50 °C in air (E) and at 50 °C in N
2

C.M. Antonio-Cisneros et al. / Food Chemistry 173 (2015) 725–732
729
composite over titanium dioxide nanoparticles is the ability to
recover the solid material, in which case, the solution to be ana-
anthraquinone moiety (with kmax,1 = 250 nm and kmax,2 = 325 nm),
with the hydroxyl groups acting as auxochromes, which in conju-
lysed is free of suspended matter. The TiO
aggregates on the carbon surface, and the content of the photocat-
alytic material in the composite (2 mg of pure TiO in 10 mg of
composite) was much lower than in the experiment with titanium
dioxide (10 mg of pure TiO ). Nevertheless, the composite was
capable of decomposing approximately 50% of the CA in 90 min
under different conditions.
The progress of the photocatalytic reaction under different con-
ditions was also monitored by HPLC, as shown in Fig. 3B. The first-
order rate constants calculated from the HPLC concentration data
2
nanoparticles formed
gation with the p electrons of the anthraquinone core might
increase kmax to a wavelength similar to that of CA. During photo-
catalysis, this phenomenon produces a decrease in the band at
495 nm measured by UV–Vis spectroscopy, which does not corre-
spond to CA itself but to a mixture of degradation products.
Instead, the concentration fate of CA observed by HPLC by means
of a single peak at a defined retention time is unaffected by the
contribution of the products. As indicated by the comparison of
Fig. 3A and B, in analysing the treated solution, the decrease in
the concentration of CA is actually faster when using HPLC than
UV–Vis analysis. This discrepancy explains why the rate constants
for the kinetics of the decomposition of the dye itself assessed by
HPLC are higher than those determined by UV–Vis spectroscopy.
At least nine new peaks of considerable intensity eluted prior to
the appearance of the dye peak in the chromatogram (Fig. 4) of the
aliquot analysed during the photocatalytic experiment. The peaks
that are numbered in the figure correspond to the intermediate
products obtained using the composite, and they exhibit a defined
isotopic pattern. Table 2 presents the retention times, exact masses
2
2
1
are significantly different from the values of k obtained by UV–
Vis spectroscopy (see Table 1). Although UV–Vis spectroscopy
measured the signal of the compounds in the bulk, HPLC separated
the coloured decomposition products of CA from the dye peak and
allowed for the quantification of the individual compounds of
interest using a calibration curve. Although no correlation between
the adsorption of anthraquinonic dyes and their photodegradation
rate was found in the work of Baran et al. (2008), our experiments
demonstrated that the photodegradation rate constant of CA was
much higher with TiO
2
(see the last column in Table 1), which
of the molecular ions (m/z), error values (Dm), and proposed struc-
adsorbed CA to a greater extent than the composite did (compare
Fig. 2A and C). Another unforeseen result was the fact that after
tures for 13 of the products detected in Fig. 4. The chromatogram
labelled ‘‘initial CA’’ in Fig. 4 shows that peaks 9 and 12 existed
in low amounts in the starting CA reagent.
9
0 min of treatment, the CA itself decomposed almost completely
in the presence of both TiO and the composite in air and in nitro-
2
The accurate identification of the aforementioned structures
indicates that under the applied photocatalytic conditions
(90 min of irradiation at 254 nm, under air at 50 °C), the TiO /
2
C800 composite tended to not completely decompose the dye mol-
ecule. The products listed in Table 2 indicate that the breakage of
the CA molecule yields mainly the glucopyranosyl-dioxoanthra-
cene moiety with different degrees of hydroxylation, with and
without the original carboxylic group. Only four compounds (peaks
gen (compare Fig. 3B with A). This finding could only be made with
the help of chromatography and will be discussed in the next
section.
3.4. Photodegradation intermediates of CA detected under UV
irradiation by LC/QTOF-MS
1
–4) corresponded to the aglycone moiety of the CA structure. The
For the identification of uncoloured and coloured intermediates,
we analysed aliquots by LC-ESI/QTOF-MS during the photocatalytic
experiments with the composite. Fig. 4 shows three selected chro-
matograms. The structures of the photodegradation products of CA
are presented in Table 2. The intermediates mainly conserved the
peaks marked u1 and u2 in Fig. 4 correspond to unknown products
whose structures could not be assigned, although well-defined
mass spectra were obtained, and these peaks corresponded to
masses of 255.2349 and 463.0906, respectively. The binuclear
compound (peak 4) eluting at 3.86 min was only detected in the
aliquot collected in an oxygen-rich atmosphere. The faster decom-
position of CA in the presence of air compared to that in nitrogen
0
0
.2
.1
0
1.2
0.6
0
9
CA
in N
(
see Fig. 3B) could be responsible for this result. Correspondingly,
intermediates 10–12 are absent in the chromatogram of the 90-
min aliquot shown in Fig. 4 because they transformed faster in
air into compounds free of the sugar moiety, such as those repre-
sented by peaks 1–4. Thus, the intensities of peaks 1–3 are higher
in the chromatogram of the sample collected from the experiment
performed in air.
1
2
1
1
6
8
7
12
u2
5
10
13
u1
u1
The masses corresponding to peaks 2, 3 and 5 were also
detected in beverages during the photodegradation process
9
12
3
(Gosetti et al., 2015). The structures proposed by the authors for
initial CA
in air
those masses might derive from the interactions between CA and
the emulsifiers contained in the beverages. This can explain the
different structure proposed by us and in the work of Gosetti
et al. (2015) for the nominal masses 207 (peak 3) and 237 (peak
4
6
8
2
1
3
2
). Since we performed photocatalytic experiments of CA, we pro-
pose oxidation of one hydroxyl group in the glycopyranoside sub-
stituent to yield the nominal mass of 489 (peak 5), however the
exact position (C2, C3 or C4) of the keto group cannot be deter-
mined by LC–MS.
0
5
10
15
time (min)
Interestingly, four peaks exhibiting a nominal mass m/zexper of
475 were detected. In this case, the structures of the pairs 7 & 8
and 9 & 11 were assigned by considering the log P values presented
in the last cells of Table 2. The pair 7 & 8 corresponds to a molecule
in which the glucosyl substituent carries three hydroxyl groups
instead of four as in CA, i.e., the substituent corresponds to a
Fig. 4. TIC obtained from LC–MS analysis of the stock solution (dotted line, right
À1
axis) of CA (300 mg L
) after 90 min of photocatalytic treatment with the
composite under nitrogen (right axis) and air (left axis) at 50 °C. The chromato-
graphic conditions are described in the experimental section. Peaks correspond to
the identified products P1–P13 gathered in Table 2. Chromatograms were shifted
upwards from their baselines for better viewing.

Table 2
Retention time, molecular mass and possible structures identified by LC-ESI-Qq-TOF-MS of metabolites of CA produced after 90 min of irradiation (254 nm) at 50 °C on a TiO
2
/C800 composite. The peak numbering is as in the
chromatograms of Fig. 4.
Proposed structure
Peak no.
(min)
1
2
3
4
t
R
1.97
2.25
2.64
3.86
Formula
m/zexper
m/zcalc
C
15
H
10
O
2
C
15
H
10
O
3
C
14
H
8
O
2
11 14 8
C H O
221.0623
221.0608
6.79
237.0576
237.0557
8.01
207.0461
207.0451
4.83
273.0635
273.0616
6.96
D
m
Detected in
2
N , air
N
2
, air
N
2
, air
Air
Proposed structure
Peak no.
5
6
7
8
9 = dcII
11.73
11
t
R
(min)
10.172
10.35
10.71
11.29
13.86
Formula
m/zexper
m/zcalc
C
22
H
18
O
13
C
19
H
18
O
13
C
22
H
20
O
12
22 20 12
C H O
489.0686
489.0674
2.45
453.0683
453.0674
1.99
475.0896
475.0882
2.95
475.0900
475.0882
3.79
D
m
Detected in
N
2
N
2
, air
N
2
2
N , air
N
2
N
2
Proposed structure
log P
Peak no.
CA
10 and 13
13.52
12
7 and 8
9 and 11
0.93
t
R
(min)
12.20
15.29
14.91
C H O
21 20 10
431.1
431.0983
3.94
À1.38
Formula
m/zexper
m/zcalc
C
22
H
20
O
13
C
21
H
20
O
11
C
21
H
20
O
11
491.0852
491.0831
4.28
447.0937
447.0933
0.89
447.095
447.0933
3.80
D
m
Detected in
–
N
2
2
N , air
N
2

C.M. Antonio-Cisneros et al. / Food Chemistry 173 (2015) 725–732
731
deoxyglucose moiety. This finding indicates that under the applied
conditions, the photocatalytic process removes one hydroxyl group
at this site from the CA molecule, but such removal can occur at
any position of the glucosyl substituent. Indeed, this removal is
probably the reason why two peaks, 7 and 8, exhibited the same
thracene with different degrees of hydroxylation. The degradation
products generated during photocatalysis depend on the atmo-
sphere in which the process is carried out. In the presence of air,
we observed fewer products than in a N atmosphere due to the
2
faster decomposition of CA favoured by an oxygen-rich atmo-
sphere. The deleterious compound anthracene-9,10-dione was
D
m/z of 475. The different position of the hydroxyl substituent in
the deoxyglucosyl moiety could shift the retention times of the
two isomers due to intramolecular bonding. Furthermore, the
hydrophilicity of the aglycone/glycone moieties of biologically
active compounds determines the extent of chromatographic
retention (Davydov, Elizalde Gonzalez, & Kiselev, 1982) and
explains why, in this case, the pair 7 & 8 elutes before the pair 9
detected both in N
2
and in air.
Acknowledgements
This work was supported by the following research projects:
CONACYT CB-2008-01-100439, and INFR-123779-2009. We thank
the financial funding of the university (UAP) for the last project.
CMAC is grateful for the grant (227566) received from CONACYT
– Mexico.
&
11.
The first possible identity of the structure of the compounds
eluting as peaks 9 & 11 is 7-C-glucopyranosyl flavokermesic acid,
with the hydroxyl group in position 8 (C-8-OH) of the anthraqui-
none ring. This derivative of CA has been labelled dcII (from the
insect Dactylopius coccus) and has been detected in historical
objects (Stathopoulou, Valianou, Skaltsounis, Karapanagiotis, &
Magiatis, 2013). Although Peggie, Hulme, McNab, and Quye
References
Ahmed, S., Rasul, M. G., Martens, W. N., Brown, R., & Hashib, M. A. (2011). Advances
in heterogeneous photocatalytic degradation of phenols and dyes in
wastewater: A review. Water Air and Soil Pollution, 215, 3–29.
(
2008) did not define the stereochemistry of the glycosidic linkage,
Akpan, U. G., & Hameed, B. H. (2009). Parameters affecting the photocatalytic
2
degradation of dyes using TiO -based photocatalysts: A review. Journal of
Hazardous Materials, 170, 520–529.
Lech and Jarosz (2011) proposed an -glucopyranosyl derivative
a–D
of flavokermesic acid for dcII. The second possibility for the iden-
tity of dcII is the molecule 7-C-glucopyranosyl flavokermesic acid
with the hydroxyl group in position 5 (C-5-OH) of the anthraqui-
none moiety. According to Peggie et al. (2008), this alternative is
unlikely in cochineal insects, but it cannot be neglected as a photo-
product. Finally, C-8-OH eluted very close to but before the CA
peak in Stathopoulou’s work (2013), and we can assume that, in
our case, it appeared as peak 9 and not as peak 11 in Fig. 4. How-
ever, any assignment of the isomers C-8-OH and C-5-OH would be
Antonio-Cisneros, C. M. (2010). DSc. Thesis. México: Universidad Autónoma de
Puebla.
Antonio-Cisneros, C. M., & Elizalde-González, M. P. (2010). Characterization of
Manihot residues and preparation of activated carbon. Biomass and Bioenergy,
3
4, 389–395.
Arroyo-Figueroa, G., Ruiz-Aguilar, G. M. L., López-Martínez, L., González-Sánchez, G.,
Cuevas-Rodríguez, G., & Rodríguez-Vázquez, R. (2011). Treatment of a textile
effluent from dyeing with cochineal extracts using Trametes versicolor fungus.
The Scientific World Journal, 11, 1005–1016.
Asiltürk, M., & Sß ener, S. (2012). TiO
characterization and photocatalytic activities. Chemical Engineering Journal, 180,
54–363.
2
– activated carbon photocatalysts: Preparation,
3
speculation. Compound 12, with m/z = 475 À 44 (CO
2
) = 431, is the
Atun, G., & Hisarli, G. (2003). Adsorption of carminic acid, a dye onto glass powder.
Chemical Engineering Journal, 95, 241–249.
Baran, W., Makowski, A., & Wardas, W. (2008). The effect of UV radiation absorption
of cationic and anionic dye solutions on their photocatalytic degradation in the
presence TiO . Dyes and Pigments, 76, 226–230.
2
Bao, N., Li, Y., Yu, X. H., Niu, J. J., Wu, G. L., & Xu, X. H. (2013). Removal of anionic azo
dye from aqueous solution via an adsorption-photosensitized regeneration
only case of decarboxylation of the anthraquinone core of dcII.
As described in the work of Brack et al. (2003), anthracene mod-
ified its structure under sunlight, yielding polyaromatic quinones
and hydroxyquinones, similar to those reported here for the photo-
degradation of CA. Brack et al. reported organism-dependent toxic-
ity and genotoxicity. In particular, anthracene-9,10-dione (peak 3
in Fig. 4 and Table 2) was reported to have caused 57% inhibition
of the algal reproduction of the unicellular green alga Scenedesmus
vacuolatus (Brack et al., 2003). This result may be important with
respect to the photolysis of CA in wastewater from the food indus-
try. Furthermore, to visualise the effect of incomplete mineralisa-
tion, we followed the formation of intermediates by LC–MS and
detected this deleterious photodegradation compound.
process on a TiO
97–906.
2
surface. Environmental Science and Pollution Research, 20,
8
Behnajady, M. A., Yavari, S., & Modirshahla, N. (2014). Investigation on adsorption
capacity of TiO -P25 nanoparticles in the removal of a mono-azo dye from
2
aqueous solution: A comprehensive isotherm analysis. Chemical Industry &
Chemical Engineering Quaterly, 20, 97–107.
Bibi, N. S., Galvis, L., Grasselli, M., & Fernández-Lahore, M. (2012). Synthesis and
sorption performance of highly specific imprinted particles for the direct
recovery of carminic acid. Process Biochemistry, 47, 1327–1334.
Bourikas, K., Vakros, J., Kordulis, C., & Lycourghiotis, A. (2003). Potentiometric mass
titrations: Experimental and theoretical establishment of a new technique for
determining the point of zero charge (PZC) of metal (Hydr)oxides. Journal of
Physical Chemistry B, 107, 9441–9451.
Brack, W., Altenburger, R., Küster, E., Meissner, B., Wenzel, K. D., & Schüürmann, G.
4
. Conclusions
(
2003). Identification of toxic products of anthracene photomodification in
simulated sunlight. Environmental Toxicology and Chemistry, 22, 2228–2237.
Cabrera, R. B., & Fernandez-Lahore, H. M. (2007). Primary recovery of acid food
colorant. International Journal of Food Science and Technology, 42, 1315–1326.
Chen, Y. X., Luo, B. H., Zhao, B. Y., Lai, Y. J., Wang, H. Z., Gao, M. Y., Zhang, W. X., &
A composite material with photocatalytic activity was obtained
from the plant stems of M. dulcis, a major agricultural residue.
Aggregates of TiO nanoparticles on carbon combine the adsorp-
2
2 2
Chen, K. B. (2012). Method to control the phase of TiO in porous carbon–TiO
tion capability and catalytic activity of both materials, respectively.
The composite particle size allows for better handling in photocat-
alytic systems than that possible using TiO alone. On the other
2
composite. Journal of Inorganic and Organometallic Polymer Materials, 22, 90–96.
Davydov, V. Ya., Elizalde Gonzalez, M. P., & Kiselev, A. V. (1982). Correlation
between the retention of cardiac glycosides in reversed-phase high-
performance liquid chromatography
a diphenylsilyl stationary phase, the
hand, manihot carbon has a basic pHpzc and a large surface area,
which allows for the adsorption of significant amounts of anionic
compounds such as carminic acid. Moreover, the adsorption is
reversible, and the dye can be partially recovered by desorption,
which can be advantageous in the case of artisanal carpeting efflu-
ents. Desorption is also a step that can play an important role in the
action of the photocatalyst. Intermediate products detected during
photodegradation indicate that the dye can generate toxic by-
products, in contrast to the nature of the dye, which is considered
safe. Using LC/QTOF-MS, 13 photodegradation products were iden-
tified, most of which contained the group glucopyranosyl-dioxoan-
structure of their molecules and their biological activity. Journal of
Chromatography A, 248, 49–62.
Foo, K. Y., & Hameed, B. H. (2010). Decontamination of textile wastewater via TiO /
2
activated carbon composite materials. Advances in Colloid and Interface Science,
159, 130–143.
Gosetti, F., Chiuminatto, U., Mazzucco, E., Mastroianni, R., & Marego, E. (2015).
Ultra-high-performance liquid chromatography/tandem high resolution mass
spectrometry analysis of sixteen red beverages containing carminic acid:
Identification of degradation products by using principal component analysis/
discriminant analysis. Food Chemistry, 167, 454–462.
Huang, D., Miyamoto, Y., Matsumoto, T., Tojo, T., Fan, T., Ding, J., Guo, Q., & Zhang, D.
(
2
2011). Preparation and characterization of high-surface-area TiO /activated
carbon by low-temperature impregnation. Separation and Purification
Technology, 78, 9–15.

7
32
C.M. Antonio-Cisneros et al. / Food Chemistry 173 (2015) 725–732
Jørgensen, K., & Skibsted, L. H. (1991). Light sensitivity of cochineal. Quantum yields
for photodegradation of carmine acid and conjugate bases in aqueous solution.
Food Chemistry, 40, 25–34.
Rangabhashiyam, S., Anu, N., & Selvaraju, N. (2013). Sequestration of dye from
textile industry wastewater using agricultural waste products as adsorbents.
Journal of Environmental Chemical Engineering, 1, 629–641.
Koperska, M., Łojewski, T., & Łojewska, J. (2011). Vibrational spectroscopy to
study degradation of natural dyes. Assessment of oxygen-free cassette for
safe exposition of artefacts. Analytical and Bioanalytical Chemistry, 399,
Rasimas, J. P., Berglund, K. A., & Blanchard, G. J. (1996). A molecular lock-and-key
approach to detecting solution phase self-assembly.
A fluorescence and
absorption study of carminic acid in aqueous glucose solution. Journal of
Physical Chemistry, 100, 7220–7229.
3
271–3283.
Körbahti, B. K.,
&
Rauf, M. A. (2009). Determination of optimum operating
Reyes-Salas, O., Juárez-Espino, M., Manzanilla-Cano, J., Barceló-Quintal, M., Reyes-
conditions of carmine decoloration by UV/H2O2 using response surface
methodology. Journal of Hazardous Materials, 161, 281–286.
Leary, R., & Westwood, A. (2011). Carbonaceous nanomaterials for the enhancement
Salas, A., & Rendón-Osorio, R. (2011). Titrimetric and polarographic
determination of carminic acid and its quantification in Cochineal
(Dactylopius coccus) extracts. Journal of the Mexican Chemical Society, 55, 89–93.
Sarikaya, R., Selvi, M., & Erkoç, F. (2012). Evaluation of potential genotoxicity of five
food dyes using the somatic mutation and recombination test. Chemosphere, 88,
974–979.
of TiO
Lech, K.,
2
photocatalysis. Carbon, 49, 741–772.
& Jarosz, M. (2011). Novel methodology for the extraction and
identification of natural dyestuffs in historical textiles by HPLC-UV-Vis-ESIMS.
Case study: chasubles from the Wawel Cathedral collection. Analytical and
Bioanalytical Chemistry, 399, 3241–3251.
Sabinas-Hernández, S.A. (2011). MSc. Thesis, Universidad Autónoma de Puebla,
México.
Mahmoodi, N. M., Arami, M., & Zhang, J. (2011). Preparation and photocatalytic
activity of immobilized composite photocatalyst (titania nanoparticle/activated
carbon). Journal of Alloys and Compounds, 509, 4754–4764.
Sreethawong, T., Ngamsinlapasathian, S., & Yoshikawa, S. (2012). Glycerol as an
efficient mesopore-controlling agent for synthesis of mesoporous-assembled
TiO
2 2
nanocrystals and their remarkable photocatalytic H production activity.
Matos, J., Chovelon, J. M., Cordero, T.,
&
Ferronato, C. (2009). Influence of
in
-chlorophenol degradation. Open Environmental Engineering Journal, 2,
1–29.
Materials Research Bulletin, 47, 199–205.
surface properties of activated carbon on photocatalytic activity of TiO
4
2
2
Stathopoulou, K., Valianou, L., Skaltsounis, A. L., Karapanagiotis, I., & Magiatis, P.
(2013). Structure elucidation and chromatographic identification of
anthraquinone components of cochineal (Dactylopius coccus) detected in
historical objects. Analytica Chimica Acta, 804, 264–272.
Valix, M., Cheung, W. H., & McKay, G. (2004). Preparation of activated carbon using
low temperature carbonization and physical activation of high ash raw bagasse
for acid dye adsorption. Chemosphere, 56, 493–501.
Peggie, D. A., Hulme, A. N., McNab, H., & Quye, A. (2008). Towards the identification
of characteristic minor components from textiles dyed with weld (Reseda
luteola L.) and those dyed with Mexican cochineal (Dactylopius coccus Costa).
Microchimica Acta, 162, 371–380.