Chemosphere 249 (2020) 126160

Contents lists available at ScienceDirect

Chemosphere
journal homepage: www.elsevier.com/locate/chemosphere

Effect of UV radiation on the structure of graphene oxide in water and

its impact on cytotoxicity and As(III) adsorption
Waldo Roberto Gallegos-Pe rez a, Ana Cecilia Reynosa-Martínez a, Claudia Soto-Ortiz a,
Mayra Ange 
lica Alvarez-Lemus b  nica García Montalvo d,
, Joaquín Barroso-Flores c, d, Vero
 pez-Honorato a, *
Eddie Lo
a n y de Estudios Avanzados del IPN, Unidad Saltillo, Av. Industria Metalúrgica 1062, Parque Industrial, Ramos Arizpe, Coahuila,
Centro de Investigacio
25900, Mexico
b
Universidad Juarez Auto
noma de Tabasco, Av. Universidad s/n, Magisterial, Villahermosa, Tabasco, 86040, Mexico
c n en Química Sustentable UAEM-UNAM, Carretera Toluca-Atlacomulco Km 14.5, Unidad San Cayetano, Toluca, Estado de
Centro Conjunto de Investigacio
M exico, 50200, Mexico
d noma de M
Instituto de Química. Universidad Nacional Auto exico, Circuito Exterior, Ciudad Universitaria, CD, MX, 04510, Mexico

h i g h l i g h t s g r a p h i c a l a b s t r a c t

Low ultraviolet irradiance induced

important structural changes in gra-
phene oxide.

Exposure time was the most impor-
tant on graphene oxide reduction by
UV light.

UV irradiance reduced 24% As(III)
adsorption capacity of graphene
oxide.

Graphene oxide cytotoxicity can
change with UV irradiance and time.

a r t i c l e i n f o a b s t r a c t

Article history: Graphene oxide (GO) is widely used in different applications, however once released into the environ-
Received 4 November 2019 ment it can change its structure and affect the transport of important contaminants such as arsenic. In
Received in revised form this work we show that UV radiation, even in the range of 28-74 mW/cm2 of irradiance up to 120 h of
4 February 2020
exposure, can induce important changes in the structure of graphene oxide, by eliminating -OH and C]O
Accepted 8 February 2020
Available online 10 February 2020
functional groups. This reduction affected the stability of graphene oxide in water by decreasing its zeta
potential from -41 to -37 mV at pH¼7 with the increase of the exposure time. Our results showed that
Handling Editor: Prof. X. Cao after 24 and 120 h of UV exposure, As(III) adsorption capacity decreased from 5 mg/g to 4.7 and 3.8 mg/g,
respectively, suggesting a lower capacity to transport contaminants with time. Computer modelling
Keywords: showed that even a degraded GO structure can have an interaction energy of 223.84 kJ/mol with H3AsO3.
Graphene oxide Furthermore, we observed that the cytotoxicity of graphene oxide changed after being irradiated at
Ultraviolet radiation 74 mW/cm2 for 120 h, showing 20% more cell viability compared to as-produced GO. Our results stress the
Arsenic adsorption importance of considering the microstructural and compositional changes that GO undergoes even
Reduction
under low irradiance and short periods, when studying its fate and behavior in the environment and
Cytotoxicity
possible applications in water treatment.
© 2020 Elsevier Ltd. All rights reserved.

* Corresponding author.
 pez-Honorato).
E-mail address: [email protected] (E. Lo

https://doi.org/10.1016/j.chemosphere.2020.126160
0045-6535/© 2020 Elsevier Ltd. All rights reserved.
2 W.R. Gallegos-Perez et al. / Chemosphere 249 (2020) 126160

1. Introduction 2010,Hou et al., 2015; Li and Bubeck, 2013). A solution of H2SO4

(95e98%, Jalmek)/H3PO4 (85.8%, J.T. Baker) with a 9:1 relation
Graphene oxide (GO) is a 2D nanomaterial with a network of (180:20 ml) was first prepared. During constant agitation a mixture
carbon atoms with sp (Pendolino and Armata, 2017) and sp (Chen of graphite flakes/KMnO4 (99%, Sigma Aldrich) in a relation 1:6
et al., 2015) hybridization and oxygen functional groups such as (1.5:9 g) was added. The resulting mixture was mixed under con-
hydroxyl, carboxyl, and epoxy, which makes it very stable in water. stant agitation for 15 min and then heat up to 50  C for 24 h. After
GO has a wide range of applications in areas such as electronics, 24 h, the mixture was slowly cooled to 2  C and 1.5 ml of H2O2 (30%,
energy storage devices and as an adsorbent material for water Jalmek) was added dropwise. The mixture was then taken to pH 1
purification, among others (Zhu et al., 2010; Pendolino and Armata, by the addition of deionized water (2  106 U1cm1, Jalmek). The
2017; Chen et al., 2015; Soni and Shukla, 2019; Zhang et al., 2019; resulting solution was centrifuged (Premiere, XC-2450 series) at
Xie et al., 2019; Nemati et al., 2018). Due to its continuous wide- 3500 rpm for 15 min until the separation of the supernatant. The
spread use, it is important to understand how graphene oxide be- precipitate was then washed with deionized water, HCl (36.5e38%,
haves once released into the environment, particularly in water Jalmek) and ethanol (99.5%, Jalmek) in consecutive sequence twice,
environments where it has been observed that UV irradiation can and then coagulated with diethyl ether (99%, Jalmek). This last
induce a reduction and breakage of this material (Konkena and solvent was removed by heating the solution at 40  C. GO was then
Vasudevan, 2015). For example, Mohandoss et al. observed exfoliated in ethanol using an ultrasonic bath (Branson, 3800) for
though X-ray photoelectron spectroscopy, a decrease in the con- 1 h. Finally, the GO was dried at 80  C for 12 h and ground in an
centration of oxygenated functional groups, related to the detach- agate mortar and sieve through a 100 mesh.
ment of hydroxyl (OH), carboxylic acid (COOH) and epoxide (C-O-C)
groups in the form of CO2. Furthermore, Hou et al. observed similar 2.2. Photoreduction
behavior by monitoring the CO2 formation by means of dissolved
organic carbon (DOC). The release of CO2 was related to a decrease The GO photoreduction was performed in 25 ml flasks with
in carbon content and an increase of defects, yielding a reduction in water recirculation at 20  C. For this purpose, 0.0125 g of GO and
particle size (Chen et al., 2015; Soni and Shukla, 2019; Zhang et al., 10 ml of deionized water were added per flask, at pH 7. These flasks
2019). were placed in a black box with 3 UV 7.2 W lamps (Tecno Lite,
Furthermore, since it is proposed that the cytotoxicity of GO is F8T5BLB), with a wavelength of 368 nm. The intensity of the UV
affected by its degree of oxidation (Liu et al., 2011) and that GO by radiation was measured with a photodiode (OPHIR, PD-300 series)
itself can work as adsorbent material via hydrogen bonds, and at the bottom of the flasks, giving values of irradiance of 28, 37, 54,
electrostatic and anion-p interactions (Chen and Chen, 2016), it is and 74 mW/cm (Pendolino and Armata, 2017). The material was
of paramount importance to understand how these structural kept under constant magnetic stirring during 24 (UV1), 48 (UV2),
changes induced by UV irradiation might affect its cytotoxicity and 72 (UV3), 96 (UV4) and 120 h (UV5) of exposure.
possible transport of contaminants. For example, it has been
observed that GO can increase the phytotoxicity of As(V) in wheat, 2.3. Arsenic adsorption
but at the same time it can increase or decrease the toxicity of
As(III), apparently depending on the as-produced GO structure A standard solution of As(III) (1000 mg/L As(III) in 2% hydro-
(Mohandoss et al., 2017; Hou et al., 2015; Li and Bubeck, 2013; chloric acid, Sigma-Aldrich) was used to prepare a solution with
Zhang et al., 2016). However, it is important to mention that in 25 mg/L of As, from which 10 ml were taken and placed in a flask
these previous studies, pristine GO was used without considering with 0.0125 g of GO. Once the pH was adjusted to 7 the flasks were
the structural changes that GO undergoes once released into the placed in the black box described previously, including one covered
environment (Li et al., 2019). Since As is an element that affects with aluminum foil to avoid UV irradiance on it as reference. The
millions of people worldwide, and whose chronical exposure is adsorption experiment was performed at 24, 48, 72, 96 and 120 h
associated with skin, lung, liver and kidney cancer, among others under constant magnetic stirring. Finally, all the solutions were
(Choong et al., 2007), it is import to study how the reduction and centrifuged for 15 min at 3500 rpm and then filtered with a 0.45
breakage of GO under UV irradiation affects the GO-As interaction, and 0.2 mm polyethersulfone (PES) membrane (Whatman). Two
something that to the authors’ knowledge has not been studied. drops of HNO3 (66.3%, J.T. Beaker) were added to preserve the so-
In this work, we studied the effect of UV irradiance (28-74 mW/ lution, according to the ASTM D2972-15 Norm for subsequent As
cm2) up to 120 h of exposure, on the microstructure and compo- quantification.
sition of GO by combining experimental and modelling work. We
observed that GO does undergo a reduction by eliminating pri- 2.4. Characterization
marily -OH and COOH functional groups with an important change
in microstructure. This affected the capacity of GO to adsorb As(III) The solid was analyzed with a Fourier Transform-Infrared
by decreasing its adsorption capacity from 5.02 down to 3.8 mg/g Spectroscopy (FT-IR) on a PerkinElmer Frontier ATR-FTIR/NIR
after 120 h of irradiance. Our results stress that even after low with a CPU32 Main software. Raman Spectroscopy was per-
irradiance and short periods (GO changing even after 24 h), GO formed in a RENISHAW inVia Microscope using a laser excitation
undergoes important structural and compositional transformations wavelength of 514 nm. X-Ray Photoelectron Spectroscopy was used
that need to be taken into consideration when studying the fate and on a PHI VersaProbe II with a 2  108 mTorr vacuum chamber,
behavior of this nanomaterial in the environment and any possible aluminum anode as X-ray monochromatic source and radiation
application where GO be under UV irradiation. energy of 1486.6 eV. The analysis range was from 1400 to 0 eV.
High-resolution spectra of the C 1s signal was obtained for each of
2. Experimental the samples. The high-resolution spectra were acquired with a step
energy of 11.75 eV. The software used to do the deconvolution was
2.1. Graphene oxide synthesis CasaXPS (version 2.3.19PR1.0). Zeta Potential was measured in a
Malvern Zetasizer Nano Z ZEN2600. As concentration in solution
GO was prepared from the oxidation of graphite flakes (Sigma- was quantified by plasma atomic emission spectrometry (ICP,
Aldrich) using the improved Hummers’ method (Marcano et al., PERKIN ELMER optima 8300 model, using an applied power of
 W.R. Gallegos-Perez et al. / Chemosphere 249 (2020) 126160 3

1300 W with a plasma gas flow rate of 15 L/min, nebulization gas CO2). Production of formazan was quantified by using a microplate
flow rate of 0.55 L/min and wavelengths of 188.9, 193.7, 197.2 and reader (Epoch, Bioteck), monitoring the difference in absorbance at
228.8 nm). Transmission electron microscopy analysis was per- 492 and 690 nm. All experiments were performed in triplicate. The
formed in a FEI-Talos F200S with four in-column Super-X EDS de- effect of each treatment was expressed as a percentage of cell
tectors with a beam current of 300 pA and a collection time of proliferation relative to untreated control cells, differences be-
10 min. tween control and treated groups are shown as a result of a two-
way ANOVA statistics.
2.5. Cell culture

Primary mononuclear cells were isolated using density-gradient 2.6. Computer modelling
separation with the reagent Histopaque-1077 (Sigma-Aldrich,
USA), according to the manufacturer’s instructions. All experiments H3AsO3, H2AsO 3 and HAsO3
2
were independently adsorbed
were set up with mononuclear cells extracted from normal in- over degraded graphene oxide and their geometries were opti-
dividuals (n¼5). Cells were cultured with RPMI 1640 medium plus mized at the LC-uPBE/6-31G (d,p) (Fig. S5) level of theory with the
10% of Fetal Bovine Serum, 10 mM Penicillin Streptomycin, and use of the Gaussian 09 suite of programs (Frisch, 2009) The inter-
10 mM of L-glutamine (Thermofisher Scientific, USA) in a 37  C, 5% action energies were also calculated with the NBODel procedure
CO2 humidified incubator. Cell viability was evaluated by the XTT with the NBO3.1Weinhold and Glendening, 1996 program as pro-
(sodium 2,3,-bis(2-methoxy-4-nitro-5-sulfophenyl)-5-[(phenyl- vided within the aforementioned suite, and the obtained values are
amino)-carbonyl]-2H tetrazolium, Roche™) assay. Briefly, cells collected in Table 3. The NBODel procedure deletes all orbital in-
were seeded in a 96-well plate at a density of 5  103 cells/well, and teractions between both species and the concomitant raise in en-
the compounds were incorporated into the medium at a concen- ergy is ascribed to the interaction between them.
tration of GO of 10, 30 and 50 mg/ml (GO after UV irradiation was
filtered and dried to prepare these solutions in order to ensure the Table 3
same amount of mass regardless of the effect of UV irradiation on Interaction Energies [kJ/mol] calculated at the LC-uPBE/6-31G (d,p) level of theory.
the structure of the material). The compounds employed for the
Functional group As (III)
treatments were freshly prepared in Dimethyl sulfoxide (DMSO);
H3AsO3 H2AsO HAsO2-
control cells were incubated only with DMSO. After 48 h, the me- 3 3

dium was aspirated and cells were washed with PBS. Cells were Graphene 16.32 1064.99 1603.93
cultured in a mix of 60 ml of DMEM/F12 medium without FBS plus Epoxide 1508.62 1380.63 3.01
Hydroxyl 1583.64 1603.93 1585.65
40 ml of XTT-solution for 2 h in a humidified incubator (at 37  C, 5%

Table 1
Percentage of functional groups in GO and irradiated GO (UV3 and UV5) at 37 mW/cm2 and 74 mW/cm (Pendolino and Armata, 2017).

Type of bond GO 72 h (UV3), 37 mW/cm2 72 h (UV3) 74 mW/cm2 120 h (UV5) 37 mW/cm2 120 h (UV5) 74 mW/cm2

% of bond

C¼C 12.0 32.6 6.3 12.8 8.4

CeC 31.0 30.5 34.7 53.0 47.3
CeOH 24.5 8.3 30.1 7.2 12.5
CeOeC 21.2 20.8 23.9 20.8 28.0
C¼O 9.8 5.0 4.0 5.6 3.2
COOH 1.7 2.8 0 0.7 0.6

Table 2
Oxygenated functional groups lost after ultraviolet radiation.

Light source l (nm) Power (W) Irradiance Time (h) Lost functional group

UV lamp 254 15 e 58 CeOH (Ji et al., 2013)

UV lamp 365 12.5 e 48 OeCeO (Ding et al., 2011)
CeOH
Solar-light e e 42e373 W/m2 16 C¼O (Mohandoss et al., 2017)
CeOH
2
Xenon arc lamp 290e700 1 kW 0.065 W/cm 10 CeOH (Hou et al., 2015)
C¼O
HOeC]O
Hg vapor lamp e 400 17 W/cm2 48 CeOH (Konkena and Vasudevan, 2015)
C¼O
UV lamp (this work) 368 7.2 37 mW/cm2 72 CeOH
120 CeOH
CeOeC
C¼O
HOeC]O
74mW/cm2 72 C¼O
HOeC]O
120 CeOH
C¼O
HOeC]O
4 W.R. Gallegos-Perez et al. / Chemosphere 249 (2020) 126160

3. Results and discussions OH, C-O-C, O-C]O, COOH groups decreased 17.2, 0.6, 4.2, 1%,
respectively. It should be noted that compared to 72 h of irradiation,
3.1. Changes in composition the CeC bonds showed an increment of 22.5% (see Table 1), and the
concentration of C]C bonds decrease from 32.6 to 12.8% from 72 to
Fig. 1 shows two examples of the FTIR spectra of GO and GO after 120 h. These differences could be related to the bond energy
UV irradiation with different irradiance during 72 (UV3) and 120 h needed to detach a functional group from the C structure (Mulyana
(UV5). In these spectra is possible to identify the signal of the et al., 2014). For example, 360 kJ/mol are required to break a C-O
stretching mode -OH at 3219 cm1, deformation mode of C]O at bond corresponding to the CeOH functional group, whereas 370
1724 cm1, stretching mode of C]C at 1619 cm1, deformation and 680 kJ/mol are needed for a C-C and C]C bond, respectively
mode of C-H at 1372 cm1 and flexion mode of C-O at 1036 cm1, (Glockler, 1958). Therefore, CeOH are transformed faster than the
particularly for the reference material, GO (Konkena and C-C and C]C bonds (Glockler, 1958). After 120 h of irradiation, due
Vasudevan, 2012). As the UV irradiance increased from 28 to to the small concentration of oxygenated functional groups, carbon
74 mW/cm (Pendolino and Armata, 2017), the intensities of these bonds appear to be altered.
bands changed, most of them decreased, giving the first sing of a As can be seen in Fig. 3 and Table 1, as the UV radiation increase
structural transformation. For example, the most notorious to 74 mW/cm2 for 72 h, the C]C decreased 5.7% and the C-C, C-OH,
changes, in both UV3 and UV5, are the disappearance of the -OH C-O-C bonds increased 3.8, 6.6, and 2.7% respectively, compared to
band and considerable reduction of the C]O band at 74 mW/cm the as-produced GO. However, the C]C bonds decrease 26.3% and
(Pendolino and Armata, 2017). Conversely, the signals assigned to the CeC bonds increased 4.3% as the irradiance increased from 37
C]C and CeH bonds increased in relation to the other bands. to 74 mW/cm (Pendolino and Armata, 2017). Conversely, the C]O
Similar behavior is observed for GO irradiated for 24 and 96 h (see and COOH bonds decreased by 5.7 and 1.7%, respectively compared
Figs. S1 and S2 in supporting information). Fig. 1, although quali- to as-produced GO. It should be noted that C-O-C bonds had an
tatively, shows that irradiance plays an important role in GO increase of 3.1% as the UV irradiance increased during the first 72 h
reduction. of irradiation. Similarly, as the irradiation time increased to 120 h
In order to obtain a more quantitative analysis of the effect with 74 mW/cm (Pendolino and Armata, 2017), the C]C bonds
induced by UV irradiation, XPS analysis was performed. Fig. 2 decreased 3.6% and the C-C bonds increased 16.4% compared to the
shows the XPS spectra for carbon (1s) in the as-produced GO and as-produced GO, but with 4.4% and 5.7% less than the value
GO irradiated with 37 mW/cm2 during 72 and 120 h. The as- measured for 120 h with lower irradiance (37 mW/cm2). Similar to
produced GO showed the presence of C]C, C-C bonds (284.4 and the previous cases, the C-OH and C]O functional groups decreased
285.2 eV) attributed to the carbon structure with sp (Pendolino and 11.9, and 6.6%, respectively, compared to pristine GO. Overall, it
Armata, 2017) and sp (Chen et al., 2015) hybridization, respectively. appears that longer periods under UV irradiation reduced the
Additionally, hydroxyl (CeOH 286.4 eV), epoxide (Ce-O-eC oxygen-bearing functional groups, as the C/O ratio changed from
287.1 eV), carbonyl (C]O 288 eV), and carboxylic acid (COOH 1.4 of pristine GO up to 2.2 and 2.7 for 120 h of irradiation. However,
289.2 eV) functional groups (Li and Bubeck, 2013; Liu et al., 2011; as time increased up to 120 h, the C-O-C concentration seems to
Chen and Chen, 2016; Zhang et al., 2016), with a C/O ratio of 1.35, increase up to 7.2%, compared to the results obtained for 72 h of
were observed similar to those previously identified by FTIR in irradiation. The elimination of C-OH, C]O and COOH functional
Fig. 1. The percentage of each type of bond is shown in Table 1. As groups with an increase of C-C and C]C bonds, has also been
the sample got irradiated for 72 h with 37 mW/m (Pendolino and previously observed by Hou et al., even though the irradiance they
Armata, 2017), the C]C and COOH bonds increased 20.6 and used was 65 000 mW/cm (Pendolino and Armata, 2017), 106 times
1.2%, respectively. Conversely, the C-OH, C-O-C, C]O bonds higher than the one used in this study (Hou et al., 2015; Li and
decreased 16.1, 0.4 and 4.7%, respectively (Table 1), as previously Bubeck, 2013), see Table 2.
suggested by FTIR in Fig. 1. The mechanism for GO reduction by UV radiation has not been
Similarly, after 120 h of irradiation with 37 mW/cm (Pendolino elucidated, however, some authors agree that the first step is water
and Armata, 2017), the C-C bonds increased 22%, whereas the C- ionization and the formations of the radicals HO
and H
, besides

Fig. 1. ATR-FTIR spectra of a) as-produced GO and irradiated GO UV3 (72 h) and b) as-produced GO and irradiated GO UV5 (120 h) at 28, 37, 54, 74 mW/cm2 and at pH¼7.
 W.R. Gallegos-Perez et al. / Chemosphere 249 (2020) 126160 5

Fig. 2. HR-XPS spectra for a) GO and irradiated GO, b) UV3 (72 h) and c) UV5 (120 h), at 37 mW/cm2 and at pH¼7.

Fig. 3. HR-XPS spectra for irradiated GO a) UV3 (72 h) and b) UV5 (120 h), at 74 mW/cm2 and at pH¼7.

the formation of solvated electrons (e) (Mohandoss et al., 2017; et al., 2013). The detachment of C-OH groups frequently results in
Hou et al., 2015; Ji et al., 2013; Gengler et al., 2013;Ganguly et al., the formation of new C]C bonds (Silva et al., 2017).
2011; Nyangiwe et al., 2015). The radical H
and solvated e act Furthermore, it has been observed that graphene can also be
as reducing agents, especially in groups such as CeOH and C-O-C, oxidized under prolonged exposure to UV radiation and form new
on the basal plane, which are very reactive and are easily detached C-O-C and C-OH functional groups (Hu et al., 2015) by the action of
in the earlier exposure to UV radiation (Mohandoss et al., 2017; Ji the radical HO
(Ji et al., 2013). This reduction route could explain
6 W.R. Gallegos-Perez et al. / Chemosphere 249 (2020) 126160

some the changes on C]O and -OH functional groups detected by (Pendolino and Armata, 2017), lower stability is achieved, partic-
XPS, where C-C bonds concentration always increases with irradi- ularly at pH¼5. These changes in zeta potential agree with the re-
ance and time, but the concentration of -OH and -CO, change sults obtained by FTIR and XPS, suggesting the loss of functional
considerably depending on time and irradiance. For example, at groups, particularly C-OH and C]O, which confer its hydrophilic
72 h of irradiation with 74 mW/cm2 the concentration of -OH property to GO and allows it to remain in a stable suspension (Dai
functional groups increased from 24 to 31%, however as time et al., 2015). Furthermore, apparently, regardless of the clear
increased to 120 h, its concentration dropped to 12%. It appears that changes in composition detected by FTIR and XPS, irradiated GO
at short times and low irradiance in the range of 37 mW/cm maintains a certain degree of stability even after 120 h of
(Pendolino and Armata, 2017), GO is reduced However, a prolonged irradiation.
exposure or higher irradiance, allows an increase on oxygen con- Our results suggest that as the composition of the material
tent in GO as an intermediate step on the decomposition and change, its stability in suspension is also affected, therefore, even
reduction of GO, probably due to the formation of a higher con- though low irradiance in water might not mean a full trans-
centration of hydroxyl radicals (Ji et al., 2013) This can be seen in formation into reduced graphene oxide, it confers important
the increase of C-OH bonds detected by XPS at 72 h but the final structural changes that with time affect its composition and
reduction on C-OH concentration at 120 h. It is important to therefore its surface charge. Consequently, as time progresses and
mention that in our experiments it is apparent that C-O-C con- functional groups are loss, GO will become less and less hydro-
centration remained almost the same with even a slight increase in philic, thus favoring the formation of agglomerates and promoting
concentration as irradiance and time increased to 120 h and 72 mW/ the material to resurface. As this happens a higher level of irradi-
cm2 (from 21% for GO to 28%). This could be related to epoxy ring- ance will be received thus accelerating this reduction and
opening/closing reactions induced by OH groups generated by UV decomposition.
irradiation (Hatakeyama et al., 2017).
These changes in composition are expected to have a strong
impact on the stability of this material in suspension. The mea- 3.2. Microstructural changes
surement of the zeta potential (z) is a factor that describes this
stability, which according to the literature, values of z higher than The structural changes previously discussed were visible also by
30 or -30 mV correspond to stable graphene oxide suspensions TEM. Fig. 5a shows the TEM micrographs of as-produced GO, con-
(Hunter, 2013). Fig. 4 shows the effect of pH, irradiance and time on firming the formation of a single layer structure of approximately
z. It is observed that at any given pH the as-produced GO is stable in 25 mm in length with soft edges and distinctive wrinkles, particu-
water since all the pH tested had a zeta potential below -30 mV. larly at the edge (Fig. S3). This tortuosity is distinctive to GO due to
However, at pH¼3 it is observed that the z of most of the irradiated the double carbon vacancy (C2) defects which generates non-
GO crosses the stability line with values higher than -30 mV, while regular rings commonly 5-8-5 member rings, but also Stone-
the as-produced GO remains stable. Similarly, it is observed that at Wales defects formed by a pair of adjacent 5-7 member rings
pH values higher than 3, the stability of the irradiated materials increasing tension in carbon structure (Ji et al., 2013; Gengler et al.,
increases with pH. This effect is related with the concentration of 2013), besides the formation of carbons with sp (Chen et al., 2015)
ionizable species such as -OH and COOH, which are protonated hybridization results in the tortuosity of the carbon structure.
under acid conditions and deprotonated under alkaline pH Overall, most of the structure of as-produced GO is amorphous as
(Konkena and Vasudevan, 2012), since the carboxylic groups have can be seen in Fig. S3.
two pK values of 4.3 and 6.6, whereas the hydroxyl groups have a However, as the sample got irradiated with 74 mW/cm2 for 120 h,
pK value of 9.8 (Konkena and Vasudevan, 2012). Overall, the GO changed considerably as it can be seen in Figs. 5b and S4, where
structure of GO does not change, however its agglomeration is the distinctive single-layer structure of GO was no longer observed.
promoted under acid conditions (Konkena and Vasudevan, 2012; Instead, a large number of irregular particles with rough edges
Jiang et al., 2016; Chowdhury et al., 2015). between 20 nm and 8 mm in length were detected. At higher
By increasing the irradiance of GO from 37 to 74 mW/cm magnification, it was possible to identify that some of the particles
were in fact agglomerates of smaller particles of around 100 nm

Fig. 4. Zeta Potential of GO and GO irradiated for 72 h (UV3) and 120 h (UV5) at 37, 74 mW/cm2 with pH from 2 to 11 at room temperature.
 W.R. Gallegos-Perez et al. / Chemosphere 249 (2020) 126160 7

Fig. 5. TEM micrographs of a) a layer of as-produced GO and irradiated GO with 74 mW/cm2 120 h, b) and c) bright and dark field image showing the presence of crystalline
structures within GO agglomerates. d) And e) high magnification micrograph showing the presence of ordered structures within GO particles. And.

(Fig. 5b) and that the irradiated graphene oxide layers were formation of smaller particles with sharp edges (Hatakeyama et al.,
severely defective at the edges and on the plane (Fig. 5d) as a results 2017,Mattevi et al., 2009).
of the decomposition process induced by UV irradiation (Ren et al.,
2018). These new structures showed clear evidence of higher 3.3. As(III) adsorption tests
crystallinity as can be observed in the dark field images (Fig. 5c)
where the bright sections correspond to crystalline structures with Previous studies have shown that As(III) adsorbs onto GO
similar orientation (Lo pez-Honorato et al., 2008). This increase in
through hydrogen bonds, the lone pairs of electrons from the
crystallinity was also evident in higher resolution TEM where epoxide group and p electron density of the sp (Pendolino and
graphene domains were now visible, albeit still in a random s-Arriagada and Toro-Labbe
Armata, 2017) structure (Corte , 2015;
arrangement (Fig. 5e). These results suggest that graphene oxide Reynosa-Martínez et al., 2020), resulting in a stronger As-GO
suffered partial reduction but a considerable degradation and interaction with GO with a higher degree of oxidation. Therefore,
8 W.R. Gallegos-Perez et al. / Chemosphere 249 (2020) 126160

Fig. 6. As(III) adsorption capacity of irradiated GO as function of exposure time to UV

radiation. Initial GO adsorption capacity was 5.02 mg/g As(III) at pH¼7.
Fig. 7. Optimized geometry for degraded graphene oxide complexes with H3AsO3.

any change in the structure and composition of GO is expected to

have an impact on As(III) adsorption and transport. Fig. 6 shows the of eOH functional groups, which have the strongest interaction
effect of UV irradiance and time on As(III) adsorption at pH 7. with As. However, as irradiation and time increase, the remaining
Originally, the as-produced GO had an adsorption capacity of GO structure, in particular, the constant presence of epoxy groups
5.02 mg/g As(III) at pH 7. However, after 24 h of irradiation, the and even the newly form sp (Chen et al., 2015) structure appear to
As(III) adsorption dropped between 4.7 and 4.8 mg/g for all values maintain the interaction with As, therefore the constant presence of
of irradiance tested (Fig. 6). Furthermore, our results show that As epoxy groups and the new interaction with sp (Chen et al., 2015)
adsorption of irradiated GO was almost similar until 96 h. It was sites, might be responsible for the small variation in adsorption
after 120 h of irradiance that adsorption capacity reached values capacity up to 96 h of irradiation. This can be seen in Fig. 7 (and
between 3.8 and 4.1 mg/g. It appears that time plays a more Fig. S5) and Table 4 where a simulated degraded GO with a hole in
important role in As(III) adsorption under UV irradiation. its structure and the presence of sp (Chen et al., 2015) hybridiza-
It has been previously been suggested that UV irradiation can tion, reached interaction energies of 223.84 kJ/mol for the neutral
result in the degradation of GO by induced photolysis through the H3AsO3. This suggests that even during the disordering of the
formation of holes in the basal plane and the concomitant forma- structure, GO-As interaction can be maintained to some extent, but
tion of carbon atoms with sp (Chen et al., 2015) hybridization as lower than the original GO structure.
functional groups are reduced by UV irradiation (Hatakeyama et al., Our results show that the complex reduction and decomposi-
2017; Lopez-Honorato et al., 2008). This reduction process could be tion process of GO under UV-irradiation can reduce the capacity of
done through the oxidation of pristine sp (Pendolino and Armata, transport of contaminants. As GO change in composition and
2017) carbon regions with hydroxyl groups by its reaction with structure, other heavy metal ion or pollutants might even have
highly reactive hydroxyl radicals and the formation of epoxy groups higher adsorption capacity since electrically charged molecules
as an intermediate step in the reduction of the hydroxyl functional have a stronger interaction with also the carbon structure of GO as
groups (Ren et al., 2018). Therefore, the observed reduction on can be seen in Table 3 for H2AsO 3 . Further work would be required
adsorption capacity after 96 h of irradiance could be attributed to in order to fully elucidate the effect of GO decomposition on the
the loss of the oxygenated functional groups such as C-OH and the transport of contaminants, for example under irradiances as high as
changes in the carbon structure with sp (Pendolino and Armata, 50 W/m and with secondary ions (carbonates, phosphates and
2017) hybridization, since they reduced from 24.5% of C-OH and sulfates) such as those found in semi-desertic areas and natural
12% of C]C, in as-produced GO, to 12.5 and 8.4, respectively in GO waters (Pinedo et al., 2006).
irradiated with 74 mW/cm2 for 120 h.
These variations in functional groups could explain the varia- 3.4. Cytotoxicity
tions seen in adsorption capacity. Table 3 shows the interaction
energies for graphene (sp (Pendolino and Armata, 2017) carbon), As a result of UV irradiation the structure of GO underwent
epoxy and hydroxyl functional groups with the different As(III)
species calculated for pristine GO. In the modelling work only the
H3AsO3, H2AsO 2
3 and HAsO3 species were taken into account, since Table 4
AsO33 can only be found at pH  13.5 and at extreme reducing NBODel Interaction Energies [kJ/mol] of degraded GO
and As(III).
conditions, close to 800 mV, conditions beyond those found in
natural waters (Larios et al., 2012; Herath et al., 2016). As(III) Energy (kJ/mol)
Graphene has a negligible interaction with arsenic (H3AsO3 H3AsO3 223.84
since we worked at pH 7). The reduction in As(III) capacity of UV H2AsO
3 592.32
irradiated GO during the first 72 h could be related to the reduction HAsO2-
3 73.34
 W.R. Gallegos-Perez et al. / Chemosphere 249 (2020) 126160 9

4. Conclusions

Our results suggest that graphene oxide suffers from a partial

reduction and decomposition under UV irradiation and that this
process is strongly dependent on exposure time and irradiance.
During short time and low irradiance (72 h at 37 mW/cm2), we
observed by XPS and FTIR that GO is reduced, losing primarily -OH
and C]O functional groups. However, regardless of time or irra-
diance, the epoxide group appears to maintain its concentration, as
confirmed by XPS measurements. Furthermore, these variations in
surface composition were also reflected in the capability of GO to
adsorb As, particularly As(III). During the first 24 h of irradiation,
GO reduced its adsorption capacity by 6.4%, reaching a minimum of
3.8 mg/g (24% reduction) after 120 h of irradiation. TEM, XPS and
computer models suggest that the presence of the C-O-C group and
a structure with sp (Chen et al., 2015) hybridization can maintain to
some extent the GO-As(III) interaction, reaching interaction en-
ergies of 223.84 kJ/mol with the neutral H3AsO3, a value higher
than the sp (Pendolino and Armata, 2017) structure found in gra-
phene, which only had an interaction energy of 16.32 kJ/mol. These
changes in composition and crystallinity also were reflected in the
Fig. 8. Cytotoxicity of as-produced GO and irradiated GO at 37 and 74 mW/cm2 during cytotoxicity of the material, where a clear reduction in cytotoxicity
72 h (UV3) and 120 h (UV5) in mononuclear cells. was observed after 120 h of irradiation at 74 mW/cm (Pendolino and
Armata, 2017). Our results suggest that the degradation of GO
induced by UV irradiation needs to be considered when studying
the true toxicity and behavior of this material in the environment
several modifications, therefore is expected that its behavior in
since the use of as-produced material as it is commonly done could
nature, especially its hazards, will change too. Fig. 8 shows the
provide erroneous information.
dose-dependent effect of as-produced and irradiated GO on cell
viability, where monocytes were used as test cells. At 10 mg/ml of
Declaration of competing interest
as-produced GO the cell viability decreased by 20%, compared with
the control test, whereas all irradiated GO materials decreased by
There are no conflicts to declare.
~10%. This difference in cytotoxicity reduced as concentration
increased up to 50 mg/ml, where as-produced and irradiated GO
CRediT authorship contribution statement
induced almost the same effect on cell viability (60%). However, a
clear difference in behavior was observed for the samples irradiated
Waldo Roberto Gallegos-Pe rez: Conceptualization, Methodol-
with 74 mW/cm2 after 5 days of irradiation, where cell viability
ogy, Investigation, Formal analysis, Writing - original draft. Ana
reached approximately 80%.
Cecilia Reynosa-Martínez: Validation, Formal analysis, Writing -
Previous reports have suggested that the cytotoxicity of GO is
original draft, Writing - review & editing. Claudia Soto-Ortiz:
strongly depended on particle size and oxygen content since the
Methodology, Investigation, Formal analysis. Mayra Ange lica
oxygenated functional groups facilitate the interaction between the

Alvarez-Lemus: Methodology, Formal analysis, Resources, Writing
GO surface and the cell membrane (Díez-Betriu et al., 2013; Ren
et al., 2018). However, our results suggest that not any change in - original draft. Joaquín Barroso-Flores: Software, Validation,
oxygen concentration would result in a variation in cytotoxicity Formal analysis, Writing - review & editing. Vero  nica García
since samples who clearly showed variations in oxygen concen- Montalvo: Resources, Writing - review & editing. Eddie Lo  pez-
tration (Table 1) maintain its degree of cytotoxicity. It is possible Honorato: Conceptualization, Resources, Writing - review & edit-
that a reduction of particle size and an increase of defects at the ing, Supervision, Project administration, Funding acquisition.
edge of these new GO debris (Fig. 5) could be partially responsible
for the cytotoxicity showed. Similar behavior has also been Acknowledgments
observed before for hydrazine reduced GO, which induced higher
cytotoxicity than GO (Liu et al., 2011), presumably due to the ex- This material is based upon work supported by a grant from the
istence of sharp edges. Furthermore, the reduction of oxygenated Consejo Nacional de Ciencia y Tecnología (CONACYT, project
functional groups is generally concentrated along the plane in GO, number: 247080, Problemas Nacionales). The authors would like to
Li et al., 2018 whereas the formation of new oxygenated functional acknowledge CONACYT for the Ph.D., MSc and BSc grants awarded
groups could happen at these highly defective edges. Akhavan and to A.C. Reynosa, W.R. Gallegos Perez and C. Soto-Ortiz, respectively.
Ghaderi, 2010 Fig. S6 shows that despite the variations in concen-
trations detected by XPS, EDS at high-resolution TEM showed that Appendix A. Supplementary data
the edge of GO debris still contains a relatively uniform distribution
of oxygen. Nevertheless, sample irradiated for 120 h with 74 mW/ Supplementary data to this article can be found online at
cm2 (UV5 74) suggests that the cytotoxicity of GO will decrease https://doi.org/10.1016/j.chemosphere.2020.126160.
after prolonged periods irradiance above 74 mW/cm (Pendolino and
Armata, 2017). Nevertheless, further work is currently underway to References
elucidate the origin of this cytotoxicity as GO is degraded by UV-
irradiation. Akhavan, O., Ghaderi, E., 2010. Toxicity of graphene and graphene oxide nanowalls
against bacteria. ACS Nano 4 (10), 5731e5736.
10 W.R. Gallegos-Perez et al. / Chemosphere 249 (2020) 126160

Chen, X., Chen, B., 2016. Direct observation, molecular structure, and location of applications. Macromol. Res. 21 (3), 290e297.
oxidation debris on graphene oxide nanosheets. Environ. Sci. Technol. 50 (16), Li, R., Guiney, L.M., Chang, C.H., Mansukhani, N.D., Ji, Z., Wang, X., et al., 2018.
8568e8577. Surface oxidation of graphene oxide determines membrane damage, lipid
Chen, M.L., Sun, Y., Huo, C.B., Liu, C., Wang, J.H., 2015. Akaganeite decorated gra- peroxidation, and cytotoxicity in macrophages in a pulmonary toxicity model.
phene oxide composite for arsenic adsorption/removal and its proconcentra- ACS Nano 12 (2), 1390e1402.
tion at ultra-trace level. Chemosphere 130, 52e58. Li, M., Zhu, J., Wang, M., Fang, H., Zhu, G., Wang, Q., 2019. Exposure to graphene
Choong, T.S.Y., Chuah, T.G., Robiah, Y., Gregory Koay, F.L., Azni, I., 2007. Arsenic oxide at environmental concentrations induces thyroid endocrine disruption
toxicity, health hazards and removal techniques from water: an overview. and lipid metabolic disturbance in Xenopus laevis. Chemosphere 236, 124834
Desalination 217 (1e3), 139e166. 1e9.
Chowdhury, I., Mansukhani, N.D., Guiney, L.M., Hersam, M.C., Bouchard, D., 2015. Liu, S., Zeng, T.H., Hofmann, M., Burcombe, E., Wei, J., Jiang, R., et al., 2011. Anti-
Aggregation and stability of reduced graphene oxide: complex roles of divalent bacterial activity of graphite , graphite oxide , graphene oxide , and reduced
cations, pH, and natural organic matter. Environ. Sci. Technol. 49 (18), graphene Oxide : membrane and oxidative stress. ASC Nano 5 (9), 6971e6980.
10886e10893.  pez-Honorato, E., Meadows, P.J., Xiao, P., Marsh, G., Abram, T.J., 2008. Structure
Lo
Corte s-Arriagada, D., Toro-Labbe , A., 2015. Improving As(III) adsorption on gra- and mechanical properties of pyrolytic carbon produced by fluidized bed
phene based surfaces: impact of chemical doping. Phys. Chem. Chem. Phys. 17 chemical vapor deposition. Nucl. Eng. Des. 238 (11), 3121e3128.
(18), 12056e12064. Marcano, D.C., Kosynkin, D.V., Berlin, J.M., Sinitskii, A., Sun, Z., Slesarev, A., et al.,
Dai, J., Wang, G., Ma, L., Wu, C., 2015. Study on the surface energies and dis- 2010. Improved synthesis of graphene oxide. ACS Nano 4 (8), 4806e4814.
persibility of graphene oxide and its derivatives. J. Mater. Sci. 50 (11), Mattevi, C., Eda, G., Agnoli, S., Miller, S., Mkhoyan, K.A., Celik, O., et al., 2009. Evo-
3895e3907. lution of electrical, chemical, and structural properties of transparent and

Díez-Betriu, X., Alvarez-García, 
S., Botas, C., Alvarez, P., S
anchez-Marcos, J., Prieto, C., conducting chemically derived graphene thin films. Adv. Funct. Mater. 19 (16),
et al., 2013. Raman spectroscopy for the study of reduction mechanisms and 2577e2583.
optimization of conductivity in graphene oxide thin films. J. Mater. Chem. C 1 Mohandoss, M., Gupta, S Sen, Nelleri, A., Pradeep, T., Maliyekkal, S.M., 2017. Solar
(41), 6905e6912. mediated reduction of graphene oxide. RSC Adv. 7 (2), 957e963.
Ding, Y.H., Zhang, P., Zhuo, Q., Ren, H.M., Yang, Z.M., Jiang, Y., 2011. A green approach Mulyana, Y., Uenuma, M., Ishikawa, Y., Uraoka, Y., 2014. Reversible oxidation of
to the synthesis of reduced graphene oxide nanosheets under UV irradiation. graphene through ultraviolet/ozone treatment and its non- thermal reduction
Nanotechnology 22 (21). through ultraviolet irradiation. J. Phys. Chem. C 118, 27372e27381.
Frisch, M.L., 2009. Gaussian 09, Revision A.02. Gaussian, Inc, Wallingford CT. Nemati, F., Zare-Dorabei, R., Hosseini, M., Ganjali, M.R., 2018. Fluorescence turn-on
Ganguly, A., Sharma, S., Papakonstantinou, P., Hamilton, J., 2011. Probing the ther- sensing of thiamine based on Arginine e functionalized graphene quantum
mal deoxygenation of graphene oxide using high resolution in situ X ray based dots (Arg-GQDs): central composite design for process optimization. Sensor.
spectroscopies. J. Phys. Chem. C 115 (34), 17009e17019. Actuator. B Chem. 255, 2078e2085.
Gengler, R.Y.N., Badali, D.S., Zhang, D., Dimos, K., Spyrou, K., Gournis, D., et al., 2013. Nyangiwe, N.N., Khenfouch, M., Thema, F.T., Nukwa, K., Kotsedi, L., Maaza, M., 2015.
Revealing the ultrafast process behind the photoreduction of graphene oxide. Free-Green synthesis and dynamics of reduced graphene sheets via sun light
Nat. Commun. 4 (May), 1e5. irradiation. Graphene 54e61, 04(03).
Glockler, G., 1958. Carboneoxygen bond energies and bond distances. J. Phys. Chem. Pendolino, F., Armata, N., 2017. In: Graphene Oxide in Environmental Remediation
62 (9), 1049e1054. Process. 1 St. Kacprzyk J, Polish Academy of Science, Systems Research Instutute.
Hatakeyama, K., Awaya, K., Koinuma, M., Shimizu, Y., Hakuta, Y., Matsumoto, Y., Springer International Publishing AG, Warsaw, Poland, pp. 1e56.
2017. Production of water-dispersible reduced graphene oxide without stabi- Pinedo, J.L.V., Mireles, F.G., Ríos, C.M., Quirino, L.L.T., Da vila, J.I.R., 2006. Spectral
lizers using liquid-phase photoreduction. Soft Matter 13 (45), 8353e8356. signature of ultraviolet solar irradiance in Zacatecas. Geofisc. Int. 45 (4),
Herath, I., Vithanage, M., Bundschuh, J., Maity, J.P., Bhattacharya, P., 2016. Natural 263e269.
arsenic in global groundwaters: distribution and geochemical triggers for Ren, X., Li, J., Chen, C., Gao, Y., Chen, D., Su, M., et al., 2018. Graphene analogues in
mobilization. Curr Pollut Reports 2 (1), 68e89. aquatic environments and porous media: dispersion, aggregation, deposition
Hou, W.C., Chowdhury, I., Goodwin, D.G., Henderson, W.M., Fairbrother, D.H., and transformation. Environ Sci Nano 5 (6), 1298e1340.
Bouchard, D., et al., 2015. Photochemical transformation of graphene oxide in Reynosa-Martínez, A.C., Navarro-Tovar, G., Gallegos, W.R., Rodríguez-Melendez, H.,
sunlight. Environ. Sci. Technol. 49 (6), 3435e3443. Torres-Cadena, R., Mondrago  n-Solo
rzano, G., et al., 2020. Effect of the degree of
Hu, X., Zhou, M., Zhou, Q., 2015. Ambient water and visible-light irradiation drive oxidation ofgraphene oxide on As(III) adsorption. J. Hazard Mater. 384 (15),
changes in graphene morphology, structure, surface chemistry, aggregation, 1e11. https://doi.org/10.1016/j.jhazmat.2019.121440, 121440, In press.
and toxicity. Environ. Sci. Technol. 49 (6), 3410e3418. Silva, KKH De, Huang, H.-H., Joshi, R.K., Yoshimura, M., 2017. Chemical reduction of
Hunter, R.J., 2013. Zeta Potential in Colloid Science: Principles and Applications, vol. graphene oxide using green reductants. Carbon N Y 119, 190e199.
2. Academic press. Soni, R., Shukla, D.P., 2019. Synthesis of fly ash based zeolite-reduced graphene
Ji, T., Hua, Y., Sun, M., Ma, N., 2013. The mechanism of the reaction of graphite oxide oxide composite and its evaluation as an adsorbent for arsenic removal. Che-
to reduced graphene oxide under ultraviolet irradiation. Carbon N Y 54, mosphere 219, 504e509.
412e418. Weinhold, F., Glendening, E.D., 1996. NBO3.1. Wisconsin.
Jiang, Y., Raliya, R., Fortner, J.D., Biswas, P., 2016. Graphene oxides in water: corre- Xie, X., Ma, X., Guo, L., Fan, Y., Zeng, G., Zhang, M., et al., 2019. Novel magnetic multi-
lating morphology and surface chemistry with aggregation behavior. Environ. templates molecularly imprinted polymer for selective and rapid removal and
Sci. Technol. 50 (13), 6964e6973. detection of alkylphenols in water. Chem. Eng. J. 357, 56e65.
Konkena, B., Vasudevan, S., 2012. Understanding aqueous dispersibility of graphene Zhang, C.Z., Li, T., Yuan, Y., Xu, J., 2016. An efficient and environment-friendly
oxide and reduced graphene oxide through pKa measurements. J. Phys. Chem. method of removing graphene oxide in wastewater and its degradation
Lett. 3 (7), 867e872. mechanisms. Chemosphere 153, 531e540.
Konkena, B., Vasudevan, S., 2015. Engineering a water-dispersible, conducting, Zhang, M., Ma, X., Li, J., Huang, R., Guo, L., Zhang, X., et al., 2019. Enhanced removal
photoreduced graphene oxide. J. Phys. Chem. C 119 (11), 6356e6362. of As(III) and As(V) from aqueous solution using ionic liquid-modified magnetic
Larios, R., Ferna ndez-Martínez, R., Silva, V., Loredo, J., Rucandio, I., 2012. Arsenic graphene oxide. Chemosphere 234, 196e203.
contamination and speciation in surrounding waters of three old cinnabar Zhu, Y., Murali, S., Cai, W., Li, X., Suk, J.W., Potts, J.R., et al., 2010. Graphene and
mines. J. Environ. Monit. 14 (2), 531e542. graphene oxide: synthesis, properties, and applications. Adv. Mater. 22 (35),
Li, H., Bubeck, C., 2013. Photoreduction processes of graphene oxide and related 3906e3924.