Revisiting terephthalic acid and coumarin as probes for photoluminescent determination of hydroxyl radical formation rate in heterogeneous photocatalysis

Article abstract of DOI:10.1016/j.apcata.2020.117566

We investigated in detail the applicability of two radical scavengers – coumarin (COUM) and terephthalic acid (TA) – as probe compounds in a widely used photoluminescent (PL) determination of OH^? radical rate formation of photocatalysts. The study reveals that precautions must be taken when using either of the two compounds as neither allows a direct determination of the absolute OH^? radical formation rate. Chromatographic analyses of irradiated COUM and TA solutions revealed that both probe compounds reacted via more than one pathway, out of which, only one pathway in each case yielded the measured photoluminescent 7-hydroxycoumarin (7-OHC) and 2-hydroxyterephthalic acid (TAOH), respectively. Applicability of both probes was also tested on three model TiO₂-based catalysts (anatase TiO₂ nanorods (TNR), anatase TiO₂ nanoparticles (TNP), amorphous TiO₂ nanorods (a-TNR)). Regardless of the probe compound used, the order of relative OH^? radical formation rates determined for these catalysts was the same (TNR < TNP < a-TNR) and in good correlation with the order of bisphenol A (BPA) degradation rates. This demonstrates that (i) formation of OH^? radicals is the predominating criterion when probing a photocatalyst's activity (but not the only one) and that (ii) TA and COUM potentially enable a relative evaluation of photocatalytic materials, despite the numerous shortcomings of these probes. However, in order to allow for general inter-laboratory comparisons the photoluminescence method should clearly be standardized with precisely defined experimental parameters.

Full text of DOI:10.1016/j.apcata.2020.117566

Journal Pre-proof
Revision of terephthalic acid and coumarin as probes for
photoluminescent determination of hydroxyl radical formation rate in
heterogeneous photocatalysis
ˇ
Gregor Zerjav (Conceptualization) (Investigation) (Validation)
(Writing - original draft) (Visualization), Alen Albreht
(Conceptualization) (Investigation) (Validation) (Writing - original
draft) (Visualization), Irena Vovk (Writing - review and editing)
(Funding acquisition), Albin Pintar (Writing - review and editing)
(Funding acquisition)
PII:
S0926-860X(20)30159-9
DOI:
https://doi.org/10.1016/j.apcata.2020.117566
APCATA 117566
Reference:
To appear in:
Applied Catalysis A, General
Received Date:
Revised Date:
Accepted Date:
18 December 2019
6 February 2020
13 April 2020
ˇ
Please cite this article as: Zerjav G, Albreht A, Vovk I, Pintar A, Revision of terephthalic acid
and coumarin as probes for photoluminescent determination of hydroxyl radical formation rate
in heterogeneous photocatalysis, Applied Catalysis A, General (2020),
doi: https://doi.org/10.1016/j.apcata.2020.117566

This is a PDF file of an article that has undergone enhancements after acceptance, such as
the addition of a cover page and metadata, and formatting for readability, but it is not yet the
definitive version of record. This version will undergo additional copyediting, typesetting and
review before it is published in its final form, but we are providing this version to give early
visibility of the article. 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.
© 2020 Published by Elsevier.

Revision of terephthalic acid and coumarin as probes for photoluminescent
determination of hydroxyl radical formation rate in heterogeneous photocatalysis
Gregor Žerjava,§, Alen Albrehtb,§, Irena Vovkb, Albin Pintara
aDepartment of Inorganic Chemistry and Technology, Laboratory for Environmental Sciences
and Engineering, National Institute of Chemistry, Hajdrihova 19, SI-1001 Ljubljana, Slovenia
bDepartment of Food Chemistry, National Institute of Chemistry, Hajdrihova 19, SI-1001
Ljubljana, Slovenia
§Both authors contributed equally to this work
GRAPHICAL ABSTRACT
1

HIGHLIGHTS

OH radical scavengers coumarin and terephthalic acid potentially enable a relative
performance comparison of photocatalysts.

An absolute determination of 푂퐻∙ radicals is not possible due to formation of multiple
photocatalytic products.


Studied TiO2 based catalysts produce 푂퐻∙ mainly through reduction of O2 in solution.
Photoluminescence and chromatographic analytical platforms are both suitable for the
determination of hydroxylation products.
Abstract
We investigated in detail the applicability of two radical scavengers – coumarin (COUM) and
terephthalic acid (TA) – as probe compounds in a widely used photoluminescent (PL)
determination of 푂퐻∙ radical rate formation of photocatalysts. The study reveals that
precautions must be taken when using either of the two compounds as neither allows a direct
determination of the absolute 푂퐻∙ radical formation rate. Chromatographic analyses of
irradiated COUM and TA solutions revealed that both probe compounds reacted via more
than one pathway, out of which, only one pathway in each case yielded the measured
photoluminescent 7-hydroxycoumarin (7-OHC) and 2-hydroxyterephthalic acid (TAOH),
respectively. Applicability of both probes was also tested on three model TiO2-based catalysts
(anatase TiO2 nanorods (TNR), anatase TiO2 nanoparticles (TNP), amorphous TiO2 nanorods
(a-TNR)). Regardless of the probe compound used, the order of relative OH∙ radical formation
rates determined for these catalysts was the same (TNR ˃ TNP ˃ a-TNR) and in good
2

correlation with the order of bisphenol A (BPA) degradation rates. This demonstrates that (i)
formation of OH∙ radicals is the predominating criterion when probing a photocatalyst’s
activity (but not the only one) and that (ii) TA and COUM potentially enable a relative
evaluation of photocatalytic materials, despite the numerous shortcomings of these probes.
However, in order to allow for general inter-laboratory comparisons the photoluminescence
method should clearly be standardized with precisely defined experimental parameters.
Keywords: coumarin; terephthalic acid; photoluminescence; photocatalysis; heterogeneous
photocatalysis; hydroxyl radicals
1. Introduction
In heterogeneous photocatalysis a catalyst is illuminated with light to initiate/accelerate a
chemical reaction. Catalysts are mostly semiconductors which have an electronic structure
consisting of a filled valence band (VB) and an empty conduction band (CB). The band gap
energy (퐸퐵퐺) represents the energy gap between VB and CB. If the energy of the incident
light (ℎ휐) is higher or equivalent to the catalyst’s 퐸퐵퐺, an electron (e-) from filled VB is
transferred to CB. Due to the transfer of the e- to CB, a vacancy called hole (h+) occurs in the
VB. Generated e--h+ pair can be involved in two different reactions. Photo excited e--h+ pairs
can recombine and produce heat, which is an unwanted reaction in most applications of
heterogeneous photocatalysis (waste water treatment, toxin removal, air treatment, etc.). On
the other hand, the generated e- and h+ can react with electron donor/acceptors and participate
in different redox reactions and produce reactive oxygen species (ROS). The non-selective
and highly reactive 푂퐻∙ radical has been considered to be the major active species responsible
3

for photocatalytic oxidation reactions [1-4]. The generation of 푂퐻∙ radicals can occur via two
pathways. The first is the reaction between VB h+ and hydroxyl ions on the catalyst surface
and the second is the indirect pathway via 푂2∙− formed as a result of interaction between
conduction-band electrons and liquid-dissolved oxygen. The identification and quantification
of OH∙ radical is an important parameter, but it is not the only one, when evaluating the
activity rate of a photocatalyst [5]. A direct detection of 푂퐻∙ radical is hindered due to its
short life time and non-selective reactivity. Several indirect detection methods were
developed based on the following techniques, e.g., electron paramagnetic resonance (EPR) [6-
9], UV/Vis absorption spectroscopy [10], fluorescence [4,11], and others. Common to all
methods is that a probe compound is used which upon reaction with 푂퐻∙ radical forms a
measurable signal. The concentration of produced 푂퐻∙ radicals is by common practice simply
calculated via stoichiometric ratio of generated products. The benefits of spectroscopic
techniques for product quantitation are low costs, rapid analysis time and reproducibility.
TA and COUM are considered suitable to act as 푂퐻∙ radical scavengers as they form products
which can be measured selectively by photoluminescence [4,12-18]. They are especially
suitable for photocatalytic reaction systems as they do not absorb light at the excitation
wavelength of photocatalysts. COUM can in theory form different hydroxycoumarins with
푂퐻∙ (3-, 4-, 5-, 6-, 7- and 8-hydroxycoumarin) of which only 7-hydroxycoumarin (7-OHC) is
photoluminescent. Zhang et al. [19] and Nagarajan et al. [20] assumed that 6.1% of total 푂퐻∙
radicals were captured as 7-OHC when 0.1 mM COUM solution was used. On the other hand,
Louit et al. [21] showed that upon illumination of COUM solution (0.1 mM) mostly 5-
hydroxycoumarin is formed which indicates that reaction’s regioselectivity is a function of
experimental conditions. In case of TA, it is widely presumed that only photoluminescent 2-
hydroxyterephthalic acid (TAOH) is formed upon reaction with 푂퐻∙ radical [22,23]. To
calculate the yield of 푂퐻∙ radicals formation from the TAOH concentration, Nakabayashi et
4

al. [22] considered the radiation chemical yield G. The value of G is defined by the number of
the molecules produced in aqueous solution with the radiation chemical reaction under the
radiation of 100 eV. The reported G values of 푂퐻∙ radical in pure water and TAOH in 0.1
mM TA solution are 2.2 [23] and 0.99 [24], respectively. Since 푂퐻∙ radical is the only
reactant needed for generation of TAOH, it was calculated that 45% of 푂퐻∙ radicals were
detected as TAOH (0.99/2.2 × 100=45). Gonzales et al. [15] recommend a yield of 35% for
pH above 9, decreasing monotonically to 30.5% at pH 2 (YTAOH=30 + 0.43 × pH).
Charbouillot et al. [17] showed that the yield increased from 14% at pH=4 to 23% at pH=10.
Little is known about the formation of side products and intermediates in these photocatalytic
systems and how these compounds might competitively react with 푂퐻∙ radicals. Both
phenomena present a potential interference and a source of error in the determination of 푂퐻∙
radical generation rates of photocatalysts, established through photoluminescent
determination of 7-OHC and TAOH, respectively.
In the present work, we set out to thoroughly assess COUM and TA as probe compounds for
the absolute and relative determination of 푂퐻∙ radical generation rate in heterogeneous
photocatalytic systems by means of photoluminescence. To demonstrate our key findings,
both probes were eventually used for the evaluation of 푂퐻∙ radical generation rate of different
model TiO2 based photocatalysts (TNR, TNP, a-TNR), which were previously used for
degradation of water-dissolved endocrine disrupting compound bisphenol A (BPA) [25,26].
Selectivity of the reaction between the probe compound and a 푂퐻∙ radical and quantitation of
reaction products were determined by high-performance liquid chromatography (HPLC), gas
chromatography (GC), and spectrofluorimetry.
2. Experimental
5

2.1. Photoluminescence probe method
The formation of 푂퐻∙ radicals was measured by using photoluminescence probe method with
terephthalic acid (TA, Alfa Aesar, Haverhill, MA, USA, 98+%, p.a) or coumarin (COUM,
Alfa Aesar, Haverhill, MA, USA, 98%, p.a) as probe compounds. Photocatalyst (10 mg) was
suspended in 50 mL of an aqueous solution containing 6 mM TA and 20 mM NaOH (Merck),
due to its low solubility in water, or in 50 mL of aqueous solution containing 1.4 mM COUM.
The suspension was stirred in the dark for 20 min before illumination with a UVA Hg lamp
(150 W, maximum at λ=365 nm). Aqueous-phase samples were withdrawn during the UV-
light illumination in different time intervals and immediately filtered through a 0.2 nm
membrane filter (regenerated cellulose). The samples were then analyzed by recording the
photoluminescence signal of the generated TAOH or 7-OHC using a Perkin Elmer UV/Vis
photoluminescence spectrophotometer (Waltham, MA, USA, model LS 55). The scanning
speed was 200 nm min-1 and the wavelength of the excitation light was set to 315 nm for 7-
OHC or 338 nm for TAOH. The excitation and the emission slits were both set to the width of
5 nm for 7-OHC or 10 nm for TAOH. PL calibration curves for 7-OHC or TAOH, illustrated
in Figs. S1a and S1b, were obtained by measuring PL curves of solutions with different
amounts of 7-OHC (Aldrich, Steinheim, Germany, 99%, p.a) or TAOH (Aldrich, Steinheim,
Germany, 97%, p.a.) and plotting the PL intensity measured at λ=450 nm for7-OHC or at
λ=425 nm for TAOH. The concentration of O2 in the solutions during N2 purging was
measured with a digital multiparameter meter (WTW, model MultiLine 3420) equipped with
an optical IDS dissolved oxygen sensor (WTW, model FDO 925).
2.2. HPLC analysis
6

The HPLC-PDA-MS system consisted of Accela 1250 UHPLC system (ThermoFisher
Scientific, Hemel Hempstead, UK) coupled to LTQ Velos system (ThermoFisher Scientific).
The separations were carried out on a Gemini C6-Phenyl column (150 mm × 4.6 mm i.d., 3
µm) (Phenomenex, Torrance, CA). Mobile phase consisted of A) 5% formic acid and B)
methanol:acetonitrile (80:20, v/v) and the following gradient was applied: 0–15 min (5–50%
B); 15–17 min (50% B); 17–18 min (50–70% B); 18–20 min (70% B); 20–21 min (70–5% B);
21–25 min (5% B). All solvents were of LC-MS grade (Merck). Flow rate was set to 0.6
mL/min, the column temperature was set to 45°C, and 25 µL of sample was injected in each
analysis. Data was acquired at 275 nm and a UV-Vis scan was performed between 200–400
nm. Analytes were ionized by a heated electrospray (HESI) probe in positive ion mode. MS
parameters were as follows: Tcapillary=350°C; Tsource=325°C; sheath gas = 60 a.e. (arbitrary
units); auxiliary gas = 10 a.e.; sweep gas = 3 a.e.; source voltage = 3 kV and MS scan range
m/z=120–400.
Quantification of TA, TAOH, 4-hydroxybenzoic acid (HBA), COUM, 6-OHC, and 7-OHC
was carried out by a 4 point external standard calibration method. For the determination of
COUM and TA, the reaction mixture was diluted with water 100-fold prior to HPLC analysis.
Repeatability of the photocatalytic reaction, when COUM was used as the probe compound,
was determined by carrying out the reaction on two different days and an average relative
standard deviation (RSD) of chromatographic peak areas corresponding to
monohydroxycoumarins was calculated. Average intraday and interday precisions of HPLC
analysis were determined by injecting the reaction mixture (t=60 min) consecutively 4-times
from the same HPLC vial on two consecutive days and an average RSD of chromatographic
peak areas corresponding to monohydroxycoumarins was calculated.
2.3. GC analysis
7

GC analyses were carried out by using a Focus GC-FID system (ThermoFisher Scientific).
Compounds were separated on an Rtx-50 column (30 m × 0.25 mm i.d., 0.25 µm) from
Restek (Bellefonte, Pennsylvania, United States). Inlet temperature was set at 300°C, He was
used as a carrier gas with a constant flow rate of 1 mL/min, and temperature of the FID
detector was maintained at 320°C. Injection volume was 0.5 μL with a split ratio equal to
1:30. Temperature gradient: 10°C/min from 210–270°C, held 2 min at 270°C. Samples were
concentrated 80-fold prior to GC analysis by removing the solvent (water) under reduced
pressure and reconstituting the solid residue in ethanol. A relative FID response factor of 1
was used for all monohydroxycoumarins since they are isomeric compounds.
Identity of COUM and monohydroxycoumarins was confirmed by analyzing the samples on a
GC Ultra-system (Thermo Electron Corporation) equipped with a DSQ II MS detector
(Thermo Electron Corporation). Same parameters as above were used, but instead of an FID,
MS detector with an ion source temperature set at 200°C was used. Electron ionization (70
eV) ion source in positive mode was used and MS data was acquired in the m/z=50–650
range.
2.4. Preparation of the catalyst and bisphenol A (BPA) photocatalytic oxidation runs
The commercially available anatase TiO2 nanoparticles (DT-51, Crystal Company, TNP) have
an ellipsoid shape with a length of around 30 nm and a diameter of 20 nm and specific surface
area of 85 m2/g. TNP were hydrothermally treated to obtain TiO2 nanorods with a core–shell
structure (a-TNR). The core–shell structure consists of ∼8 nm thick amorphous TiO2 shell
and anatase core. Pure anatase TiO2 nanorods (TNR) were obtained by further calcination of
a-TNR for 2 h at 500°C in air. The obtained nanorods had a length of around 70 nm and
8

thickness of 10 nm. Due to the calcination procedure the specific surface area of a-TNR
decreased from 381 to 102 m2/g. More details about the synthesis procedures are available in
our previous works [25,26], where also details about the BPA photocatalytic oxidation runs,
measurements of total organic carbon (TOC) content of fresh and treated BPA solutions and
about CHNS elemental analysis of fresh and spent catalyst samples are reported.
3. Results and discussion
3.1. Analytical pitfalls when using COUM and TA as probes
One needs to be reassured that methodologies and techniques which are used to probe the
performances of different (new) catalysts are suitable and produce data which can be directly
compared with other materials, especially between different laboratories and research groups.
For this purpose, a series of experiments were carried out in this study to re-evaluate two
well-known radical scavengers (COUM and TA) as probe molecules for the determination of
푂퐻∙ radical formation rate of photocatalysts.
Preliminary control experiments with 1.4 mM COUM or 6 mM TA solutions in absence of
catalysts were performed in order to investigate whether 7-OHC or TAOH were produced
upon UV-light illumination as was reported before [27]. The results presented in Fig. S2 show
that no 7-OHC or TAOH were generated under these conditions.
Further, we executed experiments in which TNR was used as a catalyst and the samples were
collected after each hour of UV-light illumination; the acquired PL spectra are illustrated in
Fig. S3, which depicts TAOH and 7-OHC formation. On the same samples also HPLC and
GC measurements were conducted to identify and determine PL silent products formed during
the photocatalytic process.
9

3.1.1 Coumarin
Photocatalytic oxidation of COUM in the presence of TNR solid was found to be repeatable;
RSD of the formed product concentrations was 6.6% (n=2). Average intraday and interday
precisions of HPLC analysis were also excellent with RSD=1% (n=4) and RSD=3.2% (n=8),
respectively.
HPLC-UV analysis of irradiated solution of COUM revealed that not only 7-OHC, but at least
11 other oxidation products formed as well (Fig. 1). Theoretically, 3-, 4-, 5-, 6-, 7-, and 8-
monohydroxycoumarin could be produced when a stoichiometry of 1:1 is considered. By
means of HPLC-MS analysis, four species corresponding to monohydroxycoumarins ([M +
H]+=163) were detected. The absence of 3-OHC and 4-OHC and the presence of 6-OHC
(tR=11.1 min) and 7-OHC (tR=12.2 min) was chromatographically confirmed by co-elution
with appropriate standard compounds. Thus, the remaining two species at tR=12.4 min and
tR=13.1 min were tentatively identified as 5-OHC or 8-OHC, respectively.
Dihydroxycoumarin isomer (tR=9.2 min; [M + H]+=179) and a putative COUM C-dimer
(tR=17.6 min; [M + H]+=293) were also detected in the reaction mixture in appreciable
amounts. The formation of the latter by irradiation of COUM in polar solvents was reported
before [28]. Highly polar species (tR=4.5–6.5 min) with MS signals at m/z=181 and m/z=197,
which presumably formed via lactone ring opening of monohydroxycoumarins and
dihydroxycoumarins, were also detected. Since photoluminescence of a reaction mixture, such
as the one studied here, stems mainly from the formation of 7-OHC (98-99%, determined by
HPLC-FLD), a wide array of reaction side-products appreciably decreases the apparent yield
of produced 푂퐻∙ radicals when this yield is determined by photoluminescence measurements
using COUM as a probe.
10

Interestingly, only the amount of dihydroxycoumarin decreased with the photocatalytic
reaction time while other major reaction products showed an expected temporal increase in
concentration; a sample stability experiment (no light, T=20°C) indicated that
dihydroxycoumarin is most probably a reaction intermediate or a side product, which is
converted within the reaction mixture to monohydroxy species since their concentrations
increased under these conditions by up to 9% after 24 h. Products other than
monohydroxycoumarins were not quantified due to the lack of appropriate standards,
however, their contribution to the overall content must not be considered negligible (Fig. 1).
COUM and the products 6-OHC and 7-OHC were quantified in the reaction mixture by
HPLC after each hour of irradiation (external standard method) and the results are
summarized in Table 1. As determined by GC-FID, 7-OHC amounted to approximately 24%
of all four monohydroxycoumarin species formed during the photocatalytic reaction.
Moreover, the results show that less than 2% of the overall COUM was converted into the
photoluminescent 7-OHC (t=180 min).
3.1.2 Terephthalic acid
Terephthalic acid (TA) is another compound which is quite often used as a probe for
photoluminescence detection of 푂퐻∙ radical generation. The advantage of TA is that only 2-
hydroxyterephthalic acid (TAOH) can be envisaged as a hydroxylation product due to the
structural symmetry of TA. However, HPLC analysis of the reaction mixture after UV-light
illumination revealed that 4-hydroxybenzoic acid (HBA, tR=8.3 min) also formed alongside
TAOH (tR=11.6 min) (Fig. 2). The identity of both compounds was confirmed with
corresponding standard compounds. Although the content of this side product was roughly an
order of magnitude lower relative to TAOH, it should not be considered negligible (Table 1).
This means that the generation rate of TAOH cannot be directly extrapolated to the generation
11

rate of 푂퐻∙ radicals and the formation of HBA should be taken into account. The results show
that under the experimental conditions used here, 1.8% of the initial TA was converted into
the photoluminescent TAOH after 180 min of UV light illumination. It should be pointed out
that one of the disadvantages of TA is also its poor solubility in water, which necessitates that
the experiments be performed in low molar NaOH solution. High solution pH (11.8 in our
case) might influence the structural stability of the catalyst or lead to deprotonation of its
surface.
The concentrations of 7-OHC and TAOH determined by photoluminescence measurements
were in good agreement with the results obtained by HPLC analysis, which demonstrates that
both analytical techniques are fit-for-purpose (Table 1). The findings presented here show that
both studied probes undergo distinct reactions with different ROS, where OH∙ radicals are just
one of the (but presumably the predominant) reacting species [4]. The amount of generated
photoluminescent product in comparison to the main reactant (COUM or TA) was in both
cases very low (2% of the initial COUM was converted into the 7-OHC and 1.8% of the
initial TA was converted into TAOH after 180 min of UV light illumination). Hence,
hydroxylation rates of reaction intermediates and side products should be considered low on
account of two orders of magnitude larger concentration levels of TA and COUM; the relative
structures of probes and intermediates are assumed to have a minor effect on the rate of
radical hydroxylation. Nonetheless, we showed that, albeit to a lesser extent, reaction
intermediates and products do react with generated OH∙ radicals which was evidenced by the
detection of dihydroxylated products and other highly polar oxidation products. These
introduce additional bias into the analytical measurements.
So, although the photoluminescence methodology does not enable the absolute determination
of OH∙ radical formation due to (i) low 7-OHC and TAOH yield, (ii) formation of a
substantial number of reaction side products and their competition for OH∙ radicals, (iii)
12

nonstoichiometric reactions of probe molecules with OH∙ radicals, and (iv) low stability of 7-
OHC (see below), a relative comparison of OH∙ radical generation rate between different
catalytic materials should be feasible when the experiments are conducted under strictly
controlled experimental conditions.
3.2. Applicability of COUM and TA for evaluating the generation rates of 푂퐻∙ radicals of
tested photocatalysts
Czilli et al. [27] and Page et al. [18] presumed that 7-OHC and TAOH undergoes photolytic
degradation. To test the photolytic stability of generated 7-OHC and TAOH under the
employed experimental conditions, 6 μM 7-OHC and 40 μM TAOH solutions were exposed
to UV-light. The results in Fig. 3 show that in the absence of catalyst 7-OHC degraded rapidly
as the concentration decreased from 6 μM to 1.26 μM in 60 min. In contrast to 7-OHC,
TAOH was stable (Fig. 3) with only a minor decrease in concentration (1%) after 60 min of
illumination.
Nonetheless, both probes were further evaluated through photocatalytic reactions using
different TiO2 catalysts (TNP, TNR and a-TNR). The results show that the rate of 7-OHC
generation under UV-light illumination was 10.6×10-3 μM/(min mgcat) when TNR catalyst
was used and 8.8×10-3 μM/(min mgcat) for TNP, if we assume zero-order reaction kinetics
(Fig. 4a, Fig. S4). The lowest 7-OHC generation rate was seen in the case of a-TNR catalyst
with a value of 0.2×10-3 μM/(min mgcat). Similar results were obtained when TA was used as
a probe with TAOH generation rates of 0.0576, 0.192, and 0.355 μM/(min mgcat) for a-TNR,
TNP, and TNR, respectively (Fig. 4b, Fig. S5). Thus, regardless of the type of the PL probe,
the 푂퐻∙ radical generation rate of examined photocatalysts followed the same order: TNR ˃
TNP ˃ a-TNR. This relative performance of the TiO2-based catalysts can be explained by
13

their different morphological, surface and electronic properties as discussed before [25,26]. It
can be concluded that the utilized TA and COUM probe compounds enable relative
evaluation of various photocatalysts as long as the latter don’t vary the reaction selectivity
towards TAOH and 7-OHC to a significant extent.
An endocrine disrupting compound bisphenol A (BPA) was also oxidized under the same
experimental conditions in order to probe its degradation mechanism (Table 2, Fig. S6). It is
well known that crystallinity of a catalyst lowers the rate of unwanted e--h+ recombination
[29], so it is not a surprise that anatase TNR achieved higher BPA degradation and
mineralization rate in comparison to amorphous a-TNR. Higher specific surface area of TNR
material beneficially influenced its photocatalytic activity in comparison to TNP material that
exhibits lower specific surface area. The order for BPA degradation and mineralization rates
of the tested photocatalysts was found to be the same as for the 푂퐻∙ radical generation rate
shown above: TNR ˃ TNP ˃ a-TNR. These results confirm that the major pathway to degrade
water-dissolved pollutant BPA is by means of 푂퐻∙ radicals [1-3], so from an analytical
standpoint, evaluation of a certain photocatalyst’s activity through 푂퐻∙ radical generation
seems suitable.
Although it was not within the main focus of this study, the electronic mechanism of the
photocatalytic reactions was explored a bit further. In separate tests COUM and TA aqueous
solutions were thoroughly sparged with pure N2 to remove water-dissolved oxygen (Fig. S7).
Regardless of the type of the photocatalyst and PL probe compound used, the 푂퐻∙ radical
formation rates (Figs. 4a, 4b, S8 and S9) were significantly lower in comparison to aerobic
experiments. This indicates that the reaction rate of photo-excited h+ and surface adsorbed
OH-/H2O is low. Surface adsorbed 푂퐻∙ cannot react with TA or COUM [27] because its
redox potential is lower (1.5 eV) [30] relative to the redox potential of 푂퐻∙ in the solution (1.9
eV), and the adsorption of COUM or TA onto the catalyst surface is weak. Thus, the 푂퐻∙
14

radical formation for these catalysts proceeded mainly via reduction of O2. The obtained
results demonstrate that heterogeneous photocatalysis is a complex system where the right
interactions between reduction and oxidation reactions are needed in order to maintain the
electrostatic equilibrium and consequently hinder the recombination of generated e--h+ pairs.
4. Conclusions
Photoluminescence technique using probe compounds to determine the amount of generated
푂퐻∙ radicals can be an effective approach to compare the photocatalytic activity of different
catalysts worldwide, although this is certainly not the only available benchmarking tool. In the
present study we re-evaluated the applicability of commonly used probe compounds within
the field of photocatalysis - COUM and TA. The results show that exact quantitation of
generated 푂퐻∙ radicals through these two probes is prevented by many drawbacks, mainly by
nonspecific and nonstoichiometric reactions with 푂퐻∙ radicals and low stability of measured
reaction products (7-OHC). However, a determination of relative 푂퐻∙ radical formation rates
between different photocatalysts is potentially possible through measurements of
monohydroxylated products of COUM and TA (7-OHC and TAOH). This was evidenced by
the order of 푂퐻∙ radical formation rates of tested TiO2 based catalysts (TNR ˃ TNP ˃ a-
TNR), which was in good correlation with the order of BPA degradation rates. To conclude,
the photoluminescence technique using probe compounds could be a rapid, sensitive and
specific method to compare photocatalytic activities of catalysts, based on their ability to
generate OH∙ radicals, if the experimental conditions (type and power of light source, probe
compound and its concentration, concentration of catalyst, etc.) would be clearly defined in a
standard.
15

CRediT author statement
Gregor Žerjav: conceptualization, investigation, validation,writing-original draft, visualization. Alen
Albreht: conceptualization, investigation, validation, writing-original draft, visualization. Irena Vovk:
writing-review & editing, funding acquisition. Albin Pintar: writing-review & editing, funding
acquisition
Acknowledgements
The authors gratefully acknowledge the Slovenian Research Agency (ARRS) for financial
support through Research Programs No. P2-0150 and P1-0005.
16

Table 1. The concentrations of COUM, 6-OHC, 7-OHC, TA, TAOH and HBA in the reaction
mixtures at different UV-light illumination times and using TNR as a photocatalyst.
Concentration (mM)
0 min
60 min
1.03
120 min
0.83
180 min
0.65
COUM
6-OHC
7-OHC
7-OHC#
TA
1.36
0
4.6×10-3
7.7×10-3
6.4×10-3
5.75
7.2×10-3
10.8×10-3
9.4×10-3
5.59
10.8×10-3
12.9×10-3
11.4×10-3
5.45
0
0
5.78
0
HBA
9.7×10-3
40×10-3
32×10-3
15.6×10-3
66.5×10-3
56×10-3
21.8×10-3
89.6×10-3
72×10-3
TAOH
TAOH#
0
0
#Calculated from PL measurements.
17

Table 2. Carbon content accumulated on the surface of catalysts before (TCfresh) and after
(TCspent) photocatalytic BPA degradation and TOC removal (TOCR) combined with true TOC
conversion (TOCm).
Sample
TOCR
TCfresh
TCspent
TOCm
%
TNP*
55
16
87
0.10
0.10
0.12
1.68
0.43
0.11
42.3
10.7
87
a-TNR**
TNR*
*The amount of catalyst in the BPA degradation experiment was 125 mg/L.
**The amount of catalyst in the BPA degradation experiment was 250 mg/L.
18

80
60
40
20
0
9
12
6
8
7
5
1
10
2
3
4
11
13
-20
0
1
2
3
4
5
6
7
8
9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time (min)
Fig. 1. HPLC analysis of COUM irradiated reaction mixture (t=120 min). Legend: polar
hydroxylated coumarin isomers (1–4), dihydroxycoumarin isomer (5), 6-OHC (6), 7-OHC
(7), 5-OHC / 8-OHC (8,9), unknown (10,11), COUM (12), and coumarin dimer (13).
19

Fig. 2. HPLC analysis of TA irradiated reaction mixture (t=120 min). Legend: HBA (1), TA
(2), TAOH (3).
20

100
80
60
40
20
0
7-OHC
TAOH
Dark phase
UV-light illumination
-20
0
20
40
60
Time (min)
Fig. 3. Stability test of 6 μM 7-OHC and 40 μM TAOH solution in dark and under UV-light
illumination in absence of a photocatalyst.
21

a
7
6
5
4
3
2
1
0
TNP
a-TNR
TNR
TNP+N2
a-TNR+N2
TNR+N2
0
10
20
30
40
50
60
Time (min)
b
50
45
40
35
30
25
20
15
10
5
TNP
a-TNR
TNR
TNP + N2
a-TNR + N2
TNR + N2
0
0
10
20
30
40
50
60
Time (min)
Fig. 4. Concentration of 7-OHC (a) and TAOH (b) generated under UV-light illumination of
various TiO2 catalysts. The lines with empty symbols represent the results of measurements
obtained when the reactor system was purged with N2. In some cases, error bars are smaller
than the symbols.
22

References
[1]. Q. Xiao, Z. Si, J. Zhang, C. Xiao, X. Tan, J. Hazard. Mater. 150 (2008) 62-67.
[2]. Q. Xiao, L. Ouyang, Chem. Eng. J. 148 (2009) 248-253.
[3]. K.E. O'Shea, C. Cardona, J. Photochem. Photobiol. A 91 (1995) 67-72.
[4]. Y. Nosaka, A.Y. Nosaka, Chem. Rev. 117 (2017) 11302-11336.
[5]. Q. Xiang, J. Yu, P.K. Wong, J. Colloid Interface Sci. 357 (2011) 163-167.
[6]. V. Brezová, D. Dvoranová, A. Staško, Res. Chem. Intermediat. 33 (2007) 251-268.
[7]. V. Brezová, A. Staško, S. Biskupič, A. Blažková, B. Havlínová, J. Phys. Chem. 98 (1994)
8977-8984.
[8]. V. Brezová, A. Šlebodová, A. Staško, Food Chem. 114 (2009) 859-868.
[9]. Z. Barbieriková, D. Dvoranová, M.V. Sofianou, C. Trapalis, V. Brezová, J. Catal. 331
(2015) 39-48.
[10]. O.I. Aruoma, Methods Enzymol. 233 (1994) 57-66.
[11]. K.-I. Ishibashi, A. Fujishima, T. Watanabe, K. Hashimoto, K., Electrochem. Commun. 2
(2000) 207-210.
[12]. A.A. Ahmed, T.A. Kandiel, T. Oekermann, D. Bahnemann, J. Phys. Chem. Lett. 2
(2011) 2461-2465.
[13]. H.B. Yener, M. Yilmaz, Ö. Deliismail, Ş.Ş. Helvaci, Sep. Purif. Technol. 173 (2017) 17-
26.
[14]. C. Buck, N. Skillen, J. Robertson, P.K.J. Robertson, Chin. Chem. Lett. 29 (2018) 773-
777.
[15]. D.H. Gonzales, X.M. Kuang, J.A. Scott, G.O. Rocha, S.E. Paulson, Anal. Lett. 51
(2018) 2488-2497.
[16]. K. Bubacz, E. Kusiak-Nejman, B. Tryba, A.W. Morawski, J. Photochem. Photobiol. A
261 (2013) 7-11.
23

[17]. T. Charbouillot, M. Brigante, G. Mailhot, P.R. Maddigapu, C. Minero, D. Vione, J.
Photochem. Photobiol. A 222 (2011) 70-76.
[18]. S.E. Page, M. Sander, W.A. Arnold, K. McNeil, Environ. Sci. Technol. 46 (2012) 1590-
1597.
[19]. J. Zhang, Y. Nosaka, J. Phys. Chem. C 117 (2013) 1383-1391.
[20]. S. Nagarajan, N. C. Skillen, F. Fina, G. Zhang, C. Randorn, L. A. Lawton, J. T.S. Irvine,
P. K.J. Robertson, J. Photochem. Photobiol. A 334 (2017) 13-19.
[21]. G. Louit, S. Foley, J. Cabillic, H. Coffigny, F. Taran, A. Valleix, J. P. Renault, Pin. S.,
Radiat. Phys. Chem. 72 (2005) 119-124.
[22]. Y. Nakabayashi, Y. Nosaka, Phys. Chem. Chem. Phys. 17 (2015) 30570-30576.
[23]. R. Nakamura, Y. Nakato, J. Am. Chem. Soc. 126 (2004) 1290-1298.
[24]. R.W. Matthews, Radiat. Res. 83 (1980) 27-41.
[25]. G. Žerjav, M.S. Arshad, P. Djinović, I. Junkar, J. Kovač, J. Zavašnik, A. Pintar,
Nanoscale 9 (2017) 4578-4592.
[26]. G. Žerjav, M.S. Arshad, P. Djinović, J. Zavašnik, A. Pintar, Appl. Catal. B 209 (2017)
273-284.
[27]. H. Czili, A. Horvath, Appl. Catal. B 81 (2008) 295-302.
[28]. G. S. Hammond, C. A. Stout, A. A. Lamola, J. Am. Chem. Soc. 86 (1964) 3103-3106.
[29]. M. Monai, T. Montini , P. Fornasiero, Catalysts 7 (2017) doi.org/10.3390/catal7100304.
[30]. D. Lawless, N. Serpone, D. Meisel, J. Phys. Chem. 95 (1991) 5166-5170.
24