Singlet Oxygen Quantum Yields ( φD ) in Water using Beetroot Extract and an Array of LEDs

É proposta uma estratégia simples e barata para determinar rendimentos quânticos (φD) de oxigênio singlete (O 2 ) de fotossensibilizadores (PS) em água utilizando extrato de beterraba contendo betacianina (Bc) e um conjunto de diodos emissores de luz (LEDs) para excitação. Bc, um corante catiônico natural, foi obtida por purificação através de cromatografia a partir do extrato de beterraba vermelha (Beta vulgaris), e utilizada como uma sonda para detecção de O 2 . Soluções do Bc e PS foram iluminadas com um arranjo de LEDs adaptado no compartimento de um espectrofotômetro comercial e a diminuição da absorvância de Bc foi seguida em função do tempo. O fotobranqueamento de Bc diminuiu em solução purgada com nitrogênio e aumentou em D 2 O, indicando o envolvimento de O 2 . A constante de velocidade de fotobranqueamento observada (k obs ) foi proporcional à intensidade do LED, à concentração e ao φD do PS. Mantendo a fonte de luz constante pudemos estimar a integral de sobreposição (R) entre a absorção do PS e a emissão do LED para diferentes concentrações de PS. A inclinação da curva de R em função de k obs é o valor da constante de velocidade de fotobranqueamento (k), que foi mostrada ser proporcional a φD. Valores de φD obtidos por este método foram comparados com aqueles obtidos através da medição da emissão no NIR (infravermelho próximo) para uma série de corantes fenotiazínicos.


Introduction
Singlet oxygen ( 1 O 2 ) is the most reactive form of molecular oxygen.O 2 has deleterius effects to human beings, for example, it is involved in skin damage caused by sun exposition. 3,4][7] In PDT, the electronic excitation of photosensitizers (PS) leads to triplets that undergo energy transfer, yielding 1 O 2 .The efficiency of 1 O 2 generation by a specific PS, which can be measured as its 1 O 2 quantum yield (φ D ), is an important property to be evaluated concerning the application of PS in PDT.[10] The most specific method to determine 1 O 2 is to measure its emission at 1270 nm. 12,13][12][13][14][15][16] Perhaps the most commonly used method to quantify the production of 1 O 2 is based in chemical trapping, 12,13,[17][18][19] i.e., 1 O 2 reacts with compounds that absorb in the UV-Vis region, causing the decay of its absorption and/or emission bands and/or the rise of the absorption/emission bands associated with the oxidized species.8][19] Highly selective probes for 1 O 2 have been recently reported. 20,21n this work, we propose a rather simple, convenient and inexpensive chemical trapping method to measure φ D , based on the use of beetroot extract containing betacyanin (Bc) as chemical trap and an array of LEDs, which is adapted in the cuvette compartment of a spectrophotometer, as excitation source.

Chemicals
The PS utilized in the experiments were Methylene Blue (MB), Azure A, Azure B and thionine.They were recrystallized from ethanol.Water was double distilled and de-ionized.

Extraction and purification of Bc
3][24] Extraction was realized by addition of the triturated beetroot (50 g) on 30 mL of milli-Q water under stirring by 90-120 min.Heating should be avoided because Bc is very susceptible to oxidation upon heating.In fact, the color of the extract changes from pink to brown when the solution is heated.While Bc is very sensible to oxidation by 1 O 2 , Bx is not (Figure 1).Therefore, in order to obtain a sensitive probe for 1 O 2 it was necessary to separate Bc from Bx.The separation of Bc from the other components of the extract was carried out by column chromatography using silica gel as stationary phase and water as eluent.[24]

Photochemical set up
All experiments were carried out in a diode-array spectrophotometer (Agilent model 8453) working in kinetic mode and using a quartz cuvette with magnetic stirring.For the irradiation, we have built a photo device using a set of commercial LEDs positioned at 90 o to the analyzer beam of the spectrophotometer (Figure 2).They were assembled in groups of four and placed at the laterals of a perforated cuvette holder.A picture of this device taken under working conditions can be seen in Figure 2B.The LEDs had maximum emission around 640 nm with total nominal power output of 16 mW (2 mW/LED, as measured in the Power meter Fieldmate-Coherent) and were connected in series, powered by an electric source of 3.0 V (DC).The photobleaching experiments were performed using 1 mL of milli-Q water, 10 µL of the Bc extract and a PS with maximum absorption smaller than 1 a.u.Under continuous irradiation, spectra were acquired every three seconds.

Near Infrared (NIR) emission
Phosphorescence decay curves at 1270 nm were recorded with a time-resolved NIR fluorometer (Edinburgh Analytical Instruments) equipped with a Nd:YAG laser (Continuum Surelite III) for sample excitation at 532 nm.The emitted light was passed through a silicon filter and a monochromator before detection by NIR-PMT (Hamamatsu Co. R5509).Data acquisition was performed by a MSA 300 MHz acquisition board (Becker&Hickl GmbH, Berlin). 16,25,26The experiments were performed at room temperature, in air-saturated aqueous solution.The sample absorbance was adjusted to 0.1 a.u. at 532 nm ([Dye] ca.70 µmol L -1 ).To avoid dimerization at these relatively large dye concentrations sodium dodecyl sulfate (SDS) was added to the solution at final concentrations of 10 mmol L -1 . 26,27In the absence of dimerization the value of singlet oxygen quantum yield (φ D ) for MB is 0.52 in several media (even in naflon films) and measured with different techniques. 12,26-29φ D values were calculated using equation 1: [12][13][14][15][16] (1) where φ D a and I a are, respectively, the quantum yield and initial phosphorescence intensity of 1 O 2 at 1270 nm of the MB species, which was used as standard.I b is the phosphorescence emission intensity of the other cationic phenothiazine derivatives.

f D by trapping with Bc and LED excitation
8][19][20][21] In simplest one, which use monochromatic light sources, the solution absorbance at the irradiation wavelength of the standard and sample is set at the same value, in order to guarantee that the same number of photons is absorbed.During the irradiation 1 O 2 is formed and reacts preferentially with the scavenger, which is in large excess compared with PS, causing a decrease in the scavenger absorption band.The absorbance change monitored as a function of time is typically fitted by a monoexponential function. 18The exponential constant gives the observed photobleaching rate constant (k obs ), which should be proportional to φ D .Strategies based in varying the concentration of trapping agents are also common. 12n this work, another strategy was implemented in order to allow the use of an inexpensive polychromatic light source (LED) to promote irradiation.k obs was obtained using different PS concentrations (Bc being always in large excess) and a relationship between k obs and φ D was found and used, considering several assumptions as described bellow.
The actual value of k obs depends on the quantum yield of O 2 lifetime. 12,13,29Because Bc is photochemically inert, we only have to consider the photochemical reactions that start with PS excitation.Calculation of the exact photon flux absorbed by the PS is not trivial because the LED device exhibits a spectral band of emission.It is necessary to calculate the absorption factor of the sensitizer (1-10 -Abs PS ) integrated on the overall spectral region of the LED emission, which should be expressed in photonic units (PE LED ), and this integral should also be corrected by the fraction of incident light transmitted by the Bc (10 -Abs Bc ), (equation 2). (2) Because the main idea of this work is to provide an easy method to calculate φ D , it is fundamental to simplify this calculation.Considering that measurements will be made at different PS concentrations and compared with a standard, even if the simplified equation proposed bellow has systematic errors, these errors will be the same for samples and standard.The first assumption is related with the Bc absorption.Although Bc does absorb the LED emission, this absorption basically does not change during irradiation, i.e., the LED emission maximum is just in an isosbestic point of the Bc absorption change profile during the photochemical reaction (Figure 3).Also the Bc concentration was kept constant in all experiments.Therefore, we will not consider the Bc absorption in the calculations, assuming that the changes that are due to Bc absorption are the same for samples and standard.
In terms of the PS absorption factor and LED emission, because the PS absorption is not large and does not change during irradiation, we propose to estimate the relative fraction of PS excited by the LED at each PS concentration by calculating the overlap integral (R) between PS absorption (ABS PS ) and LED emission (E LED ) (equation 3), instead of calculating the exact photon flux absorbed by the PS (equation 2). (3) Depending on the acceptor and on the solvent, 1 O 2 lifetime may vary during irradiation.However, in water the total quenching of 1 O 2 by Bc (physical and chemical reactions) is negligible compared to quenching by the solvent (we have always obtained 1 O 2 lifetime of 4 ± 0.5 µs in all experimental conditions).Consequently, the PS absorption and the 1 O 2 lifetime did not change during irradiation.
Considering that the Bc concentration was kept constant and in excess compared with PS, the Bc bleaching (k obs ) should follow pseudo first-order kinetics and can be calculated by mono-exponential fitting.Also the plot of k obs versus R can be approximated by a linear fit.The slope of this curve is the rate constant of the chemical reaction of Bc with singlet oxygen (k), which is theoretically proportional to the quantum yield of 1 O 2 production (φ D ).Note that the other factors that could affect the value of k, i.e., the bimolecular rate constant of the chemical reaction between Bc and 1 O 2 and the 1 O 2 lifetime are unchanged.In these experiments small dye concentrations were used and therefore there was no need to add SDS to avoid dimerization.

Results and Discussion
Before starting the chemical trapping experiments, it was important to test for the stability of the Bc solutions under illumination.The Bc solution was illuminated for several minutes using either a laser source that specifically excites Bc (Nd:YAG laser, 20 mW, 532 nm) or LEDs that do not excite Bc (640 nm, set-up mount, see Experimental section).There was no absorbance change in both control experiments.Therefore, although Bc is susceptible to oxidation upon heating, it is photochemically inert.
When an aqueous solution containing Bc (1.0 a.u. in 540 nm) and MB (ca.10 µmol L -1 ) was irradiated with the LED set up (Figure 2) the Bc absorbance decreased during the 8 min irradiation (Figure 3) and the color of the solution changed from purple to blue.Note that the MB absorption did not change during the experiment, indicating that electron transfer reactions that could lead to MB photobleaching, are not taking place under such experimental conditions. 25,30,31n none of the phenothiazines tested, there was any change in the dye absorption during irradiation.
Decrease in Bc absorbance at 540 nm during the irradiation followed an exponential profile and fitted well to a single exponential function, allowing the calculation of k obs (Figure 4A).The temporal decay of the Bc photobleaching in the presence of MB was also compared in solutions with different oxygen concentrations.The bleaching efficiency increased with the increase in the oxygen concentration suggesting again the presence of type II oxidation processes (Figure 4A).Photobleaching induced by electron transfer reactions are usually accelerated in nitrogen-purged samples, because oxygen can suppress triplet and radical intermediates. 30,31ifetime of 1 O 2 is longer in D 2 O (35-40 µs) than in water. 6,12,13Assuming that no other interference occurs, the steady-state concentration of 1 O 2 would be ca.10 times larger giving consequently ca.10 times faster reaction rate in ca.100% D 2 O.This effect has been used to study the role of 1 O 2 in specific photochemical events by carrying out experiments in water and in D 2 O under the same experimental conditions (Figure 4B).However, it is important to emphasize that we were not using ca.100%D 2 O in our experiments because the reaction mixture is prepared in situ by using Bc solution in water (giving around 90% D 2 O:H 2 O mixture).Even so, we have observed ca. 4 times faster rate indicating that the bleaching of Bc by MB excitation is more efficient in D 2 O than in water, which proves that 1 O 2 is the main species responsible for the bleaching.These observations also indicate that Bc is a specific 1 O 2 scavenger and therefore, a good probe for estimating φ D .
Monoexponential decays of Bc absorbance as a function of time were observed for all dyes and concentrations tested, facilitating the calculation of k obs .k obs was plotted as a function of R, which was calculated for each dye concentration (Figure 5).The slope of these curves, found by least square fitting of the experimental data, allows the calculation of k (Table 1, first column).Assuming the proportionality between k and φ D and using MB as standard it was possible to calculate φ D values of azure A, azure B and thionine (Table 1, second column).It is clear that the calculated values are very similar to those found in the literature or obtained by measuring NIR emission (Table 1, third column).

Conclusions
The chemical trapping protocol based in Bc and LED provides an efficient, inexpensive and simple method to   quantitatively determine 1 O 2 formation by a PS.Although killing organisms and treating diseases with dyes and light is a very old process in humanity, now a days, people have to pay thousands of dollars to be treated by PDT.However, inexpensive PDT protocols can be easily established.Considering that finding new drugs represents an increasingly difficult and expensive task, the discovery of new PS for PDT can be a practical solution, especially for underserved populations.The method here presented may be of relevance in pursuing such solution.

1 O 2
production (φ D ), on the photon flux absorbed by the PS, on the bimolecular rate constant of the chemical reaction between Bc and 1 O 2 , on the concentration of Bc and on the 1

Figure 2 .
Figure 2. (A) Top view of the schematic arrays of LEDs arranged around the spectrophometric cuvette.The LEDs were connected in series and powered by a 3.0 V (DC) source.(B) Picture of the irradiation device at working conditions; the LEDs exhibited maximum emission at ca. 640 nm.

Figure 3 .
Figure 3. Absorption spectra of the Bc and MB (11.2 µmol L -1 ) solution under.LED light illumination.The spectra were taken every 30 seconds from a (time=0) to b (time=8 min).The LED emission is shown in red.

Figure 4 .
Figure 4. Absorbance at 540 nm as function of irradiation time: (A) in water samples purged with nitrogen (), air-saturated (), oxygen (); (B) air saturated solutions in water () and in deuterium oxide ().The data were obtained in solutions similar to those used to obtain data of Figure 3.