Photothermal catalytic properties of layered titanium chalcogenide nanomaterials

Article abstract of DOI:10.1039/c9dt03798e

The search for novel photocatalysts that make use of almost the entire solar spectrum remains an ongoing task to achieve high efficiency in energy conversion. While titanium chalcogenides offer a variety of phase compositions with different photophysical properties, their photocatalytic performance in pollutant degradation has not been investigated to date. In contrast to the model photocatalyst titanium dioxide, titanium chalcogenides possess small band gaps which make them eligible to absorb light in the visible range up to the near-infrared region, thus making them interesting candidates for photocatalysis. Herein titanium chalcogenide-based photocatalysts are synthesized by the chemical vapor transport (CVT) method and studied for their photocatalytic activity towards pollutant degradation. A series of titanium chalcogenides TiX_n (X = S, n = 1-3; X = Se, n = 2; X = Te, n = 1) have been characterized by a variety of physico-chemical methods. Due to the expected non-stoichiometry of some titanium sulfides, they offer a large number of defect states which make them interesting candidates for photocatalysis. Thus, these titanium-chalcogenides were systematically studied for the photocatalytic degradation of pollutants using methyl orange dye as the test pollutant under simulated sunlight. Particularly TiS and TiS₃ show high photocatalytic and thermocatalytic activity, outperforming the activity of titanium dioxide (pure anatase). By controlling the ratios of titanium and chalcogen elements and the specific reaction conditions, a variety of titanium chalcogenides with different compositions and phases showing a high photocatalytic activity can be accessed. Furthermore, it is found that the formation of a titanium dioxide passivation layer during photocatalysis results in a titanium oxide/titanium sulfide heterostructure. This allows further enhancement of the photocatalytic and thermocatalytic activity compared to the bare Ti-chalcogenides.

Full text of DOI:10.1039/c9dt03798e

Dalton
Transactions
View Article Online
View Journal
PAPER
Photothermal catalytic properties of layered
titanium chalcogenide nanomaterials†‡
Cite this: DOI: 10.1039/c9dt03798e
Sherif Okeil, §a Sandeep Yadav,§a Michael Bruns,b Alexander Zintler,c
a
Leopoldo Molina-Lunac and Jörg J. Schneider
*
The search for novel photocatalysts that make use of almost the entire solar spectrum remains an
ongoing task to achieve high efficiency in energy conversion. While titanium chalcogenides offer a variety
of phase compositions with different photophysical properties, their photocatalytic performance in pollu-
tant degradation has not been investigated to date. In contrast to the model photocatalyst titanium
dioxide, titanium chalcogenides possess small band gaps which make them eligible to absorb light in the
visible range up to the near-infrared region, thus making them interesting candidates for photocatalysis.
Herein titanium chalcogenide-based photocatalysts are synthesized by the chemical vapor transport
(CVT) method and studied for their photocatalytic activity towards pollutant degradation. A series of tita-
nium chalcogenides TiXn (X = S, n = 1–3; X = Se, n = 2; X = Te, n = 1) have been characterized by a variety
of physico-chemical methods. Due to the expected non-stoichiometry of some titanium sulfides, they
offer a large number of defect states which make them interesting candidates for photocatalysis. Thus,
these titanium-chalcogenides were systematically studied for the photocatalytic degradation of pollutants
using methyl orange dye as the test pollutant under simulated sunlight. Particularly TiS and TiS3 show high
photocatalytic and thermocatalytic activity, outperforming the activity of titanium dioxide (pure anatase).
By controlling the ratios of titanium and chalcogen elements and the specific reaction conditions, a
variety of titanium chalcogenides with different compositions and phases showing a high photocatalytic
activity can be accessed. Furthermore, it is found that the formation of a titanium dioxide passivation layer
during photocatalysis results in a titanium oxide/titanium sulfide heterostructure. This allows further
enhancement of the photocatalytic and thermocatalytic activity compared to the bare Ti-chalcogenides.
Received 25th September 2019,
Accepted 2nd December 2019
DOI: 10.1039/c9dt03798e
rsc.li/dalton
abundance, nontoxicity and cost-effectiveness.2,3 The main
drawback of titanium oxide is that it is mainly active under
Introduction
The use of solar energy as a clean energy source for water puri- ultraviolet light due to its wide band gap (3.2 eV for anatase
fication and detoxification through photocatalytic removal of and 3.0 eV for the rutile phase)3 and thus is not making use of
organic pollutants has been recognized as an important the whole spectrum of sunlight which consists of 4% UV-A
research area over the last few decades. A lot of effort has gone radiation (320–400 nm, 3.88–3.09 eV), 53% visible light
into the search for potential materials that are able to utilize (400–800 nm, 3.09–1.55 eV) and 43% IR radiation (>800 nm,
most of the solar spectrum for an efficient photocatalysis <1.55 eV).4,5 Several approaches have been used in order to
process.1 Titanium dioxide is a model photocatalyst that has modify the band gap of titanium dioxide to make use of
been intensively investigated for its activity due to its stability, visible light, such as metal doping, non-metal doping, modifi-
cation with metal nanoparticles, modification with visible-
light-absorbing dyes, inducing defects and its inclusion in
aEduard-Zintl-Institut für Anorganische und Physikalische Chemie,
heterostructures.1,3,6 Particularly the so-called black titanium
Technische Universität Darmstadt, Alarich-Weiss-Straße 12, 64287 Darmstadt,
dioxide has become a hot topic since its first synthesis by
Germany. E-mail: joerg.schneider@ac.chemie.tu-darmstadt.de
bInstitut für Angewandte Materialien (IAM-ESS), Karlsruher Institut für Technologie,
Chen et al.7 who relied on the fact that the hydrogenation of
Hermann-von-Helmholtz-Platz 1, B 321, D-76344 Eggenstein-Leopoldshafen, Germany pure white titanium dioxide results in the formation of a dis-
cFachbereich Material- und Geowissenschaften, Technische Universität Darmstadt,
ordered shell on the titanium dioxide particles leading to its
black appearance. This led to an efficient reduction of the
band gap to 1.54 eV which extended the absorption into the
64287 Darmstadt, Germany
†Dedicated to Prof. Manfred Scheer on the occasion of his 65th birthday.
‡Electronic supplementary information (ESI) available. See DOI: 10.1039/
visible and infrared region of the sunlight spectrum resulting
in a significant enhancement of the photocatalytic activity.
C9DT03798E
§Both authors contributed equally to this work.
This journal is © The Royal Society of Chemistry 2019
Dalton Trans.

View Article Online
Paper
Dalton Transactions
Due to the efficiency of titanium-based materials in photocata- titanium chalcogenides for photocatalysis has been under-
lysis and the intense research into their modification to taken. Our work could be useful to shed light on new poten-
achieve even better photocatalytic behavior, it appears highly tially active photocatalyst materials based on titanium. These
interesting to study other non-oxygen containing titanium could then even be employed for the design of different hetero-
chalcogenide-based materials and to assess their photo- structures, which could allow further improvement in their
catalytic activity. This could pave the way for finding new and photocatalytic activity as shown in this study in which a partial
efficient photocatalysts operating under visible light. To the surface oxidation of titanium sulfides leads to the formation
best of our knowledge, the heavier titanium chalcogenides of titanium oxide/titanium sulfide heterostructures, which is
where the chalcogen elements are S, Se and Te have not been accompanied by a further increase in the photocatalytic
reported for their photocatalytic activity in pollutant degra- activity. A previous study by Ranjith et al.27 showed that the
dation to date. Theoretical band gap calculations of different sulfidation of TiO2 nanofibers leads to the formation of a thin
titanium chalcogenides determine a narrow band gap, for TiS3 shell, thus forming a heterostructure that promotes
8–10
example, Eg < 1.2 eV for TiS3
and Eg < 2 eV for titanium carrier separation and decreases the exciton recombination
dichalcogenides (TiS2, TiSe2 and TiTe2),11 which shows that rate resulting in an enhanced photocatalytic activity. Whereas
these materials could have a potential for visible light or even the sulfidation process of titanium dioxide is difficult due to
IR-driven photocatalysis.
the oxophilic nature of titanium, the surface oxidation of tita-
Moreover, it is known that two-dimensional nanostructures nium chalcogenides would provide a simple approach to
lead to better charge separation resulting in higher photo- obtain titanium oxide/titanium chalcogenide heterostructures
catalytic activity.12 Consequently, two dimensional titanium for improved photocatalysis. As shown in the following study,
dioxide sheets have been studied for their photocatalytic appli- such heterostructures could be readily formed during their use
cations.13 The improved charge separation efficiency of two- in subsequent photocatalytic cycles, which results in an
dimensional materials, compared to the respective bulk improvement in the photocatalytic performance.
material, can be ascribed to a lower bulk diffusion of any
photogenerated charge carriers which allow them to reach the
surface much faster, thus reducing the risk of exciton bulk
Materials and methods
Synthesis of titanium chalcogenides
recombination.14,15 An important characteristic of transition
metal chalcogenides is that they often possess a bulk structure
that is composed of sandwiched layered components that can Titanium powder (abcr 99.5%, ∼325 mesh), sulfur powder
be individually exfoliated to form isolated two-dimensional (Sigma-Aldrich 99.998%), selenium granules (EVOCHEM
materials.2,16
99.999%, 1–5 mm) and tellurium powder (abcr 99.99%,
Belt like titanium trisulfide morphologies have been syn- ∼325 mesh) were used as received. Different titanium chalco-
thesized and have shown effective performance in energy genides were synthesized according to general procedures
storage,17–19 transistors20,21 and photoelectrocatalysis for using chemical vapor transport (CVT) following previously
hydrogen generation.22,23 Titanium trisulfide nanobelts have published methods:22,28–30 titanium and the respective chalco-
also been used as sacrificial templates for the production of gen powder with the required ratio were mixed and ground
different TiS2−x nanobelt morphologies preserving the overall together in a mortar. The well mixed powder was then used to
two-dimensional structure.24 All this makes it interesting to produce a pellet using a manual hydraulic press under 10 tons
investigate titanium chalcogenides (S, Se, Te) with a lamellar load for 10 min. The resulting pellet containing titanium and
structure for photocatalysis.
the respective chalcogen was then inserted in one end of a
In non-stoichiometric materials where vacancies and quartz ampule together with about 5 mg of iodine which was
defects exist, typically the photocatalytic activity is increased as used as the transport medium and the ampule was sealed
the vacancies can act as effective surface electron trapping under a vacuum of 10−3 mbar. This ampule was placed in an
sites, thus further enhancing the necessary charge separation oven and heated slowly to the desired temperature which was
for efficient performance.25 Furthermore, vacancies and held for a certain period of time and then allowed to cool
defects produce intermediate states in the band gap of the slowly to room temperature. After cooling to room tempera-
material, thus extending its absorption edge to longer wave- ture, the titanium chalcogenide either formed on the other
lengths.26 Particularly for titanium disulfide it is well-known end of the ampule or at the same position of the initial pellet.
that on the one hand an exact stoichiometry is hard to control, For the preparation of different titanium sulfides, different
and on the other hand such a non-stoichiometric composition molar ratios of the element powder were used: for titanium tri-
possesses interesting properties.24 Therefore, a study of the sulfide a molar ratio of Ti : S of 1 : 3, for titanium disulfide a
photocatalytic behavior of non-stoichiometric titanium sul- molar ratio of 1 : 2 and for titanium monosulfide a molar ratio
fides seems justified in order to test the importance of of 1 : 1.
vacancies and defects for the photocatalysis of such titanium-
For the preparation of titanium trisulfide 0.33 g of titanium
chalcogenide-based photocatalysts that differ chemically and were mixed with 0.67 g of sulfur powder to form the pellet.
structurally from the well-investigated titanium dioxide. Thus, The ampule was heated up to 550 °C within 5 h and held at
herein the first in depth investigation of the use of different that temperature for 48 h before cooling down. In the case of
Dalton Trans.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Dalton Transactions
Paper
titanium disulfide 0.43 g of titanium was mixed with 0.57 g of system was employed during analysis, using electrons of 8 eV
sulfur to produce the pellet. The ampule was then heated up energy and low-energy argon ions to prevent any localized
to 750 °C within 7 h and held at that temperature for 48 h charge build-up. The charge shift correction was employed
before cooling. For titanium monosulfide a molar ratio of 1 : 1 using a C 1s electron binding energy of 285.0 eV.
was used corresponding to 0.60 g of titanium and 0.40 g of
The synthesized titanium chalcogenides were characterized
sulfur which were mixed for the preparation of the pellet. For using X-ray diffraction (XRD) which was performed on a
low temperature titanium monosulfide (lt-TiS) the ampule was Rigaku Miniflex 600@40 kV 15 mA diffractometer using Cu
heated up to 450 °C within 4 h and this temperature was held Kα1 radiation (λ = 1.541 Å). Transmission electron microscopy
for 48 h before cooling down. For high temperature titanium (TEM) was performed using JEOL ARM-200F operated at 200
monosulfide (ht-TiS) a temperature of 900 °C was achieved in kV. TEM samples were prepared from the powdered samples
9 h and maintained for 6 h only before cooling down to room obtained through grinding followed by ultrasonically disper-
temperature.
sing about 20 mg of the powder in about 10 mL ethanol for
For the preparation of titanium selenide and titanium tell- 15 hours. The obtained dispersion was then drop cast on a
uride 0.23 g of titanium were mixed with 0.77 g of selenium holey carbon grid. Images and selected area electron diffrac-
powder and 0.16 g of titanium were mixed with 0.84 g of tellur- tion (SAED) patterns were recorded on a Gatan US1000. Energy
ium to form the pellet, respectively. The ampules containing dispersive X-ray spectra (TEM-EDX) were recorded on a JEOL
the titanium–selenium and titanium–tellurium mixtures were JED-2300 60 mm2 detector.
heated to 800 °C over a 8 h period and 750 °C over a 7 h period
and maintained for 72 h, respectively.
Raman spectra of the different titanium chalcogenides were
recorded on a Horiba Jobin-Yvon LabRAM HR800 micro-
Raman spectrometer using a 632.8 nm HeNe laser with a 50×
LWD objective (NA = 0.5) and 20 s acquisition time.
Post-treatment of titanium trisulfide to introduce sulfur vacancies
Titanium trisulfide was used as a starting material to obtain
UV-VIS-NIR diffuse reflectance spectroscopy (UV-VIS-NIR
several non-stoichiometric titanium sulfides through the DRS) of the synthesized titanium chalcogenides was per-
removal of sulfur under argon atmosphere at 450 °C for formed with an integrating sphere on a PerkinElmer Lambda
different time intervals. Titanium trisulfide powder was placed 900 UV-VIS-NIR spectrophotometer using the Kubelka–Munk
in a quartz boat inside a quartz tube and heated in a tube function to obtain the corresponding absorption spectra.
furnace (Horst GmbH tube furnace, 500 mm length and
Photocatalytic activity measurement
50 mm inner diameter with a NiCr–Ni temperature sensor and
HT MC1 microprocessor temperature regulator) to 450 °C
under a flow of 250 sccm argon. The temperature was held for
20 min, 40 min and 60 min to produce three different samples
TiS3−x/20 min, TiS3−x/40 min and TiS3−x/60 min with different
sulfur contents. The removed sulfur could be clearly observed
as yellow deposition at the cold end of the quartz tube which
gave evidence of the efficient removal of sulfur from titanium
trisulfide in these short time periods.
The photocatalytic activity of the different titanium chalco-
genides was tested using methyl orange (MO) (ACS, Reag. Ph
Eur, Merck) as a model pollutant. The details of the performed
experiments can be found elsewhere.31 In short, the prepared
test solution consisted of 0.02 mg mL−1 MO, 0.001 M HCl and
0.035% (w/w) hydrogen peroxide. The test solution was mixed
together with 1 mg of the photocatalyst to be tested in a quartz
test tube and was left for 2 hours in the dark to reach adsorp-
tion equilibrium. After that the test solution was illuminated
with a lamp simulating sunlight (Exo Terra Solar Glo®
mercury vapor bulb, 150 W) at a 10 cm distance. At predefined
time intervals samples were drawn from the test solution and
analyzed by UV-VIS to determine the amount of the remaining
MO. The change in the absorption maximum of the MO solu-
tion at 503 nm was monitored and the ratio C/C0 was plotted
against the photocatalysis time, where C/C0 denotes the ratio
of the concentration of MO at the sampling time to the initial
concentration of MO when the photocatalysis experiment
started. The photocatalytic activity of the different synthesized
titanium chalcogenides was compared to commercial anatase
TiO2 (EMPROVE® ESSENTIAL Ph Eur, BP, USP, JP, E 171, Merck).
Characterization
Scanning electron microscopy of the synthesized titanium
chalcogenides was performed on a Philips XL-30 FEG coupled
with an energy-dispersive X-ray (EDX) analyzer for element
analysis using an electron beam at 20 kV as accelerating
voltage. Thermogravimetric analysis (TGA) measurements of
the different titanium sulfides for the determination of their
bulk composition were performed using
a TG209F1-Iris
(Netzsch) thermal analyzer, employing aluminum crucibles.
The samples were heated at a rate of 10 K min−1 till 800 °C
under oxygen atmosphere and this temperature was held for
one hour to ensure complete conversion to stoichiometric tita-
nium dioxide.
XPS characterization of the titanium chalcogenide powder
samples was performed using a K-Alpha XPS spectrometer
(ThermoFischer Scientific, East Grinstead, UK) equipped with
a microfocused, monochromated Al Kα X-ray source (400 µm
Results and discussion
Materials synthesis and characterization
spot size). Data acquisition and processing was done using the Transition-metal dichalcogenides (MX2) and transition-metal
Thermo Avantage software. The K-Alpha charge compensation trichalcogenides (MX3) are known to possess a sheet like
This journal is © The Royal Society of Chemistry 2019
Dalton Trans.

View Article Online
Paper
Dalton Transactions
arrangement composed of the layers of chalcogen–metal–chal- insertion of titanium atoms into the interlayer gaps (Ti1+xS2 or
cogen (X–M–X) which are held together by van der Waals Ti1+xS3). In the case of titanium sulfides the non-stoichiometry
forces. This makes it easy to cleave the crystal along the hori- resulting from the synthesis is mainly due to titanium intersti-
zontal axis perpendicular to the crystal symmetry axis to tials which has been proved by X-ray and density measure-
obtain few-layered or even single layered flakes by breaking the ments on different titanium sulfide crystals indicating the
van der Waals forces acting between these layers. The tita- presence of different interphases between stoichiometric TiS2
nium-based transition-metal dichalcogenides TiS2, TiSe2 and (CdI2-type structure), Ti2S3 (Fig. 1c) and TiS with its non-
TiTe2 have a hexagonal CdI2-type structure where one layer is layered NiAs-type structure (Fig. 1d).38 Phase diagrams of the
formed of a transition-metal plane sandwiched between two titanium-sulfur system also reflect the complexity of the Ti/S
planes of chalcogen atoms (Fig. 1a).24,32,33 Titanium trisulfide phase regime with various phases and polytypes which makes
(TiS3) which belongs to the layered transition-metal trichalco- it hard to define the exact boundary lines between these
genides has a monoclinic ZrSe3-type structure having the phases and polytypes.39
space group P21/m where two of the three sulfur atoms per
Fig. 2 shows the products obtained from the solid state
unit formula contain a S–S bond, resulting in disulfide anion reaction of titanium and sulfur powder at 550 °C to form TiS3
species (S22−) forming a triagonal base parallel to the crystal nanobelts.20–22,40 While the powder material shows the pres-
a-axis and pointing towards the interlayer gap where weak van ence of a mixture of different sized TiS3 nanobelts together
der Waals interactions hold the layers together (Fig. 1b).34,35 In with some sheet-like structures and nanobelt fragments (Fig. 2
the case of titanium trisulfide, trigonal prismatic TiS6 units and Fig. S1a and b‡), the TiS3 belts that are formed at the end
are individually connected.20,24 This is in contrast to individ- of the tube have a width between 3 and 12 µm and are usually
ual TiS6 octahedra in the case of titanium disulfide. The several hundreds of µm long with a thickness ranging from
structure of titanium trisulfide can thus be formulated as 50 nm to 200 nm showing the presence of belt type structures
Ti4+S2−(S22−) indicating that the oxidation state of titanium is in the range of several nanometers up to micrometers in size
4+ as in titanium disulfide, however with one sulfide and one (Fig. 2a and b). For the ease of discussion all the TiS3 belts
disulfide anion.36 Non-stoichiometry in these titanium chalco- formed at the end of the ampule will be termed “TiS3 nano-
genides is very likely and well-known in the literature.37 The belts”. In the region where the powder-like material is formed,
deviation from the stoichiometry in titanium sulfides can be a mixture of different long and short nanobelt structures or
either due to the presence of sulfur vacancies (TiS2−x or TiS3−x
)
even nanobelt fragments exists beside some sheet-like struc-
or due to the presence of Ti interstitials formed through the tures (Fig. 2c and d). Additional heat treatment of the TiS3
powder material, after removal from the ampule, at 450 °C
under argon flow in a tube furnace allows the removal of some
elemental sulfur but does not change the morphology of the
TiS3 nanobelts and fragments found (Fig. S2‡). The reaction of
titanium and sulfur powder at 750 °C (ratio 1 : 2) yields a shiny
Fig. 1 Representation of the crystal structure and unit cell of (a) hexag-
onal TiS2 (CdI2-type, space group: P3¯m1), (b) monoclinic TiS3 (space
group: P21/m), (c) monoclinic Ti2S3 (space group: Cc), where the red
arrows show the positions of interstitial titanium between the “TiS2
layers” and (d) hexagonal TiS (space group: P63/mmc) where the inter-
layer gaps are filled with titanium resulting in a non-layered NiAs-type
structure.
Fig. 2 SEM images at different magnifications of (a and b) TiS3 crystals
attached to the ampule inner wall showing the formation of TiS3 nano-
belts and (c and d) TiS3 powder showing the presence of TiS3 nanobelts
of different sizes together with sheet-like structures.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Dalton Transactions
Paper
dark greenish powder of composition TiS2 (Fig. S1c and d‡) stoichiometric TiS3. For TiS2 64.84% S and 35.16% Ti (S/Ti
which reveals the presence of micrometer-sized prism-shaped ratio of 1.84) were obtained (Fig. S4c‡) which are close to the
crystals that show a lamellar structure upon further magnifi- values for the stoichiometric TiS2 and agree well with reports
cation (Fig. 3a and b). In the case of TiS two different synthesis that indicate that TiS2 usually does not form with an ideal stoi-
temperatures (450 °C and 900 °C) have been investigated while chiometry of 2 : 1 (S/Ti).24,37 For the non-stoichiometric tita-
keeping the stoichiometric ratio of the elements Ti and S at nium sulfides 56.81% S and 43.19% Ti (S/Ti ratio of 1.32) for
1 : 1. In both cases, a greyish to black powder is obtained lt-TiS and 61.64% S and 38.36% Ti (S/Ti ratio of 1.61) for ht-
(Fig. S1e and f‡ for lt-TiS) where the powder of the lt-TiS is dull TiS were obtained (Fig. S4d and e‡). Thermogravimetric (TG)
while that of ht-TiS is shiny similar to TiS2. SEM of both analysis leads to similar results, thus confirming the obtained
powders reveals the presence of layered prism structures for stoichiometries, independently. For the crystalline TiS3 a S/Ti
ht-TiS similar to the crystals of TiS2 in which the layers are ratio of 2.85, while for the TiS3 powder a ratio of 3.29 were
electron transparent (Fig. 3c and d) while for lt-TiS crumpled obtained, indicating the presence of some excess sulfur in the
sheet-like structures with thickness of a few nm are observed case of the TiS3 powder. For TiS2 a S/Ti of 1.94 was found
(Fig. 3e and f). The other titanium chalcogenides, TiSe2 and while for the non-stoichiometric titanium monosulfides (lt-TiS
TiTe, synthesized in this work also show layered crystals in the and ht-TiS), the ratios of 1.23 and 1.57 were obtained, respect-
SEM ranging from a few micrometers to several hundred ively. The EDX for the selenium-based titanium chalcogenide
micrometers in size (Fig. S3‡).
TiSe2, 67.06% Se and 32.94% Ti is in accord with the 1 : 2 stoi-
EDX analysis gives a deeper insight into the bulk compo- chiometry. However, for the tellurium-based titanium chalco-
sition of the synthesized nanomaterials.41,42 XPS analysis genide, the EDX determined atomic percentages of 47.58% Te
yields mainly information from the material surface with an and 52.42% Ti, thus suggesting the formation of TiTe in 1 : 1
information depth of 1 to 5 nm which thus supplements this stoichiometry (Fig. S5a and b‡). EDX analysis of the obtained
information nicely.43 For TiS3 crystal compositions of 74.52% TiS3 powder thermally treated for different time periods at
S and 25.48% Ti (S/Ti ratio of 2.92), while for the TiS3 powder 450 °C under argon (TiS3−x/20 min, TiS3−x/40 min and TiS3−x
/
75.44% S and 24.56% Ti (S/Ti ratio of 3.07) were obtained by 60 min) indicates a stepwise removal of sulfur with increasing
EDX analysis (Fig. S4a and b‡) indicating the formation of treatment time. The S/Ti ratio is 3.07 for the TiS3 powder and
decreases to 2.16 for TiS3−x/20 min, then to 1.59 for TiS3−x
/
40 min and finally reaches 1.23 for TiS3−x/60 min (Fig. S6‡).
EDX spectra do indicate the presence of trace amounts of
oxygen. A potential oxygen source could come from the sample
handling process. A reliable quantification of oxygen is not
possible due to the severe overlap of the O K peak with the Ti
L excitation line.
Photoelectron spectroscopic characterization
Titanium trisulfide TiS3. XPS analysis for the different tita-
nium sulfides could give more detailed information about
their surface structure and composition which is extremely
important to understand their photocatalytic activity. The Ti
2p high-resolution core level spectra of the freshly prepared
TiS3 nanobelts display two main peaks at 455.9 eV and 461.9
eV which can be assigned to Ti4+ Ti 2p3/2 and Ti 2p1/2, respect-
ively, with a spin–orbit doublet splitting of 6 eV which rep-
resents Ti associated with S.44 Upon deconvolution two further
small peaks at 458.8 eV and 464.6 eV could be fitted corres-
ponding to the Ti 2p3/2 and Ti 2p1/2 of TiOx, respectively45 indi-
cating the presence of additional surface oxidation as the crys-
tals were kept under ambient atmosphere before the XPS
measurements (Fig. 4a). Long term stability experiments (two
years) under ambient atmosphere before the initial XPS
measurements reveal, besides the Ti 2p3/2 and Ti 2p1/2 peaks
indicating Ti4+ of Ti–S at 456.5 eV and 462.5 eV, two further
prominent peaks at 459.5 eV and 465.4 eV corresponding to
the Ti 2p3/2 and Ti 2p1/2 of TiOx, respectively (Fig. 4a). This
reveals that even at longer periods of time, surface oxidation
proceeds at a rather slow rate, indicating good stability of the
synthesized TiS3. In contrast to this observation, XPS analysis
Fig. 3 SEM image of (a and b) TiS2 crystals, (c and d) ht-TiS and (e and
f) lt-TiS at different magnifications showing the different morphologies
and the layered composition of these nanomaterials.
This journal is © The Royal Society of Chemistry 2019
Dalton Trans.

View Article Online
Paper
Dalton Transactions
Fig. 4 High-resolution XPS showing Ti 2p and S 2p core level signals of (a and b) TiS3 powder and TiS3 nanobelts (freshly prepared and after two
years of storage under ambient conditions) and (c and d) TiS3 powder treated at 450 °C under argon for 20 min, 40 min and 60 min.
of the Ti 2p core level spectrum for the TiS3 powder formed
XPS analysis of the TiS3 powder heated under argon at
during the same synthesis and which is also stored for two 450 °C for 20 min, 40 min and 60 min (Fig. 4c and d) shows
years under ambient atmosphere, however shows intense the presence of disulfide and sulfide anions which can be
surface oxidation compared to the sample composed of large attributed to the high-resolution S 2p core level spectra which
crystals. This is due to the higher surface area exposed in the upon deconvolution show the corresponding characteristic
powder (Fig. 4a). The presence of oxygen related to TiOx can doublets, thus indicating that after heat treatment TiS3 is still
also be found in the O 1s high-resolution core level deconvo- present (Fig. 4d). In a study by Hawkins et al.24 heat treatment
luted spectra at 530.6 eV together with oxygen related to the of TiS3 in a vacuum sealed vial using elemental titanium
surface-bound OvC–O species at 532.3 eV as well as signals powder as a sink for the released sulfur at 450 °C for 24–144 h
corresponding to aliphatic alcohol and ether species at 533.5 resulted in a transformation of TiS3 into non-stoichiometric
eV stemming from adventitious carbon formed due to sample TiS2−x. Calculation of the S22−/S2− ratio in these spectra shows
exposure to ambient atmosphere (Fig. S7a, ESI‡).45,46 The a decrease in the S22−/S2− ratio to about 1.6 in comparison
amount of surface oxidation can be clearly seen from the with 1.8 in the untreated powder indicating the formation of
deconvoluted O 1s spectra through the ratio of the peaks sulfur vacancies. This might be a first indication that qualifies
related to TiOx and the adventitious carbon related OvC–O that material for further photocatalytic studies. This finding is
species thus showing that the freshly prepared crystals have also supported by an overall decrease in the S/Ti ratio deduced
the lowest degree of surface oxidation followed by the two-year- from the S 2p and Ti 2p peaks of the corresponding XPS survey
old nanobelt crystals and finally the two-year-old TiS3 powder. spectra ranging from 2.5 for the untreated TiS3 powder to 1.8
A comparison of the S 2p spectra of these TiS3 samples shows for TiS3−x/20 min, to 1.5 for TiS3−x/40 min and to 1.35 for
no difference where the S 2p can be deconvoluted to two pairs TiS3−x/60 min (Fig. S8,‡ XPS survey spectra). This trend is in
of peaks at 161.5 eV and 162.7 eV corresponding to the S 2p3/2 line with the EDX analysis (see above) for the heat treated TiS3
and S 2p1/2 of sulfide S2− and 162.7 eV and 163.9 eV represent- powders. Comparing the high-resolution Ti 2p core level
ing the S 2p3/2 and S 2p1/2 of disulfide anion S22− (Fig. 4b). The spectra of these heat treated TiS3 samples, they all show the
ratio of S22−/S2− is close to 2 : 1 which is typical of titanium tri- doublets corresponding to the Ti4+ characteristic for the Ti–S
sulfide which consists of one S22− unit and one S2− unit.47
bond (Ti4+(S)) and TiOx indicating the presence of different
Dalton Trans.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Dalton Transactions
Paper
degrees of surface oxidation dependent on the heat treatment binding energy doublet could not be assigned to a specific
time (Fig. 4c). By determination of the ratio of TiOx/Ti4+(S) as sulfur species.44,48 It can be speculated that it could corres-
1.80, 2.35, 2.85 and 3.27 for the untreated TiS3 powder and for pond to some surface-bound thiol species.49
the heat treated TiS3 samples TiS3−x/20 min, TiS3−x/40 min
Titanium monosulfide TiS. Titanium monosulfides, lt-TiS
and TiS3−x/60 min, respectively, a systematic increase in and ht-TiS, reveal, similar to the TiS2 and TiS3 samples, two
surface oxidation is observed. The same trend can be observed doublets in the Ti core level spectra, one at 456.2 eV and 462.2
in the deconvoluted high-resolution O 1s core level spectra eV with a splitting of 6 eV assigned to Ti 2p3/2 and the Ti 2p1/2
(Fig. S7b, ESI‡).
of Ti4+ and associated with the Ti–S bond, and the other at
Titanium disulfide, TiS2. Titanium disulfide, TiS2, shows a 458.9 eV and 464.7 eV with a splitting of 5.8 eV corresponding
high-resolution Ti 2p core level spectrum similar to that of to the Ti 2p3/2 and Ti 2p1/2 peaks of TiOx (Fig. 5c). Ht-TiS syn-
TiS3 (Fig. 5a), in which two doublets are also observed. One thesized at 900 °C shows a diminished degree of surface oxi-
doublet is at 456.3 eV and 462.4 eV with a doublet splitting of dation with a TiOx/Ti4+(S) ratio of 0.47 compared to lt-TiS syn-
6.1 eV, thus corresponding to the Ti 2p3/2 and Ti 2p1/2 of Ti4+ thesized at 450 °C with a TiOx/Ti4+(S) ratio of 1.98. The two-
associated with the Ti–S bond, respectively and the other year-old lt-TiS shows complete surface oxidation as no Ti4+(S)
doublet is at 459.0 eV and 464.7 eV with a doublet splitting of can be found anymore. The same can be seen in the corres-
5.7 eV corresponding to the Ti 2p3/2 and Ti 2p1/2 of TiOx, ponding high-resolution O 1s core level spectra (Fig. S7d, ESI‡)
respectively, the latter due to surface oxidation during sample where the O 1s (TiOx)/O 1s (OvC–O) ratio is 1.91, 3.58 and
storage (see also O 1s core level spectrum in Fig. S7c‡).
5.09 for ht-TiS, lt-TiS and the two-year-old lt-TiS, respectively,
The S 2p core level spectrum of TiS2 (Fig. 5b) differs signifi- indicating the increasing degree of surface oxidation as well.
cantly from that of TiS3. No doublet signal corresponding to The high-resolution S 2p core level spectra of TiS (Fig. 5d)
the disulfide anions S22− is observed, but instead another high show a similar pattern to that observed for TiS2 with a decon-
binding energy doublet at 163.8 eV and 165.0 eV corres- volution into two doublets, one at 161.2 eV and 162.4 eV
ponding to S 2p3/2 and S 2p1/2 with a doublet splitting of 1.2 corresponding to the S 2p3/2 and S 2p1/2 peaks of sulfide S2−
.
eV can be found besides the characteristic S 2p doublet of The second doublet at 163.5 eV and 164.7 eV corresponds to
sulfide S2− found at 160.9 eV and 162.2 eV for S 2p3/2 and S the S 2p3/2 and S 2p1/2 peaks of the yet unidentified sulfur
2p1/2, respectively. In accord with the literature, the high species. Thus, the XPS data reveal that the titanium monosul-
Fig. 5 High-resolution XPS showing the Ti 2p and S 2p core level spectra of (a and b) TiS2 and (c and d) low temperature TiS (freshly prepared and
after two years of aging) and high temperature TiS.
This journal is © The Royal Society of Chemistry 2019
Dalton Trans.

View Article Online
Paper
Dalton Transactions
fide TiS represents a non-stoichiometric material, most prob-
ably due to a large number of defect sites.
XRD analysis of the TiS3 powder heated at 450 °C under
argon for different time intervals shows some changes.
However, the main reflexes for TiS3 remain unchanged while
the reflexes for elemental sulfur (S8) observed in the as-pre-
pared TiS3 powder disappear and a new broad peak at 25.15°
appears and is assigned to anatase TiO2 (Fig. 7). Relating these
data to the observations obtained from the EDX (Fig. S6‡) and
the XPS (Fig. 4c and d) confirms that some sulfur vacancies
are formed, parallel to a certain degree of surface passivation
due to oxide formation. The significant broadening of the
(001) peak at 25.15° of the formed anatase also indicates the
formation of TiO2 nanocrystallites which agrees with the idea
of filling of surface sulfur vacancies with oxygen from the
atmosphere to which the material is exposed after the anneal-
ing process resulting in the formation of such isolated nano-
crystallites on the surface.
X-ray characterization
For TiS3 nanobelts the XRD shows main reflexes for the (001),
(002), (003), (004), (005) and (006) planes. Additional reflexes
are missing due to the preferential orientation of the TiS3
nanobelts along their c-axis (Fig. 6a). For TiS3 powder the XRD
however shows different crystal orientations (Fig. 6b) indicat-
ing that different nanostructures exist in the powder besides
the nanobelt morphology which has a preferential growth
direction. Additional reflexes are due to traces of elemental
sulfur present (S8 ICDD PDF no. 85-0799). The XRD of TiS2
(Fig. 6c) matches well with the powder diffraction file ICDD
PDF no. 88-2479 for titanium disulfide where the difference
observed in the measured intensities is also due to the prefer-
ential orientation of the hexagonal crystals. In the case of the
phases obtained for lt-TiS and ht-TiS, an unambiguous assign-
ment is difficult due to the presence of a huge number of
different polytypes for different titanium sulfides. An assign-
ment to the non-stoichiometric titanium sulfides Ti2.45S4
(ICDD PDF no. 72-0820) and Ti1.083S2 (ICDD PDF no. 84-0372)
for lt-TiS and ht-TiS, respectively represents the best fit
(Fig. 6d and e).
Transmission electron microscopic characterization
The bright-field TEM images (Fig. 8a1–e1) show that after
extensive sonication exfoliation of the different titanium sul-
fides takes place resulting in very thin and mostly electron-
transparent specimens underlining their layered two-dimen-
sional structure. The lattice fringes in the HRTEM (Fig. 8a2–e2)
and the corresponding SAED patterns with well-defined spots
(Fig. 8a3–e3) reveal the monocrystalline nature of the different
titanium sulfide samples. The spacings of the lattice fringes
obtained for the TiS3 nanobelts are 3.3 Å, 2.4 Å and 1.7 Å
corresponding to the (010), (200) and (020) planes of TiS3,
respectively (Fig. 8a2) and show a preferential growth direction
along their b-axis [010]. For the TiS3 powder also nanobelts are
observed (Fig. 8b1) which in the HRTEM show lattice spacings
of 4.7 Å and 3.3 Å corresponding to the (100) and (010) planes
of TiS3, respectively.
Fig. 7 XRD of the as-prepared TiS3 powder, TiS3−x/20 min, TiS3−x
/
Fig. 6 XRD of (a) TiS3 nanobelts, (b) TiS3 powder (reflexes with asterisk
correspond to elemental sulfur), (c) TiS2, (d) lt-TiS and (e) ht-TiS. The lt-
TiS phase is best fitted with Ti2.45S4 (ICDD PDF no. 72-0820) and ht-TiS
matches Ti1.083S2 (PDF no. 84-0372).
40 min and TiS3−x/60 min obtained by heating of the as-prepared TiS3
powder at 450 °C under argon resulting in the disappearance of
elemental sulfur and the appearance of some anatase TiO2 (ICDD PDF
no. 15-0783 of TiS3 is given for comparison).
Dalton Trans.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Dalton Transactions
Paper
Fig. 8 Bright-field TEM (a1–e1) are overview images. HRTEM with the corresponding d-spacing of the lattice fringes is shown in (a2–e2) and the
indexed SAED is given in (a3–e3) where (a–e) denote: (a) TiS3 nanobelts with SAED along the [001] zone axis, (b) TiS3 powder with SAED along the
[001] zone axis, (c) TiS2 with HRTEM along the [011] and SAED along the [001] zone axis showing a hexagonal structure, (d) lt-TiS with SAED along
the [100] zone axis and (e) ht-TiS with SAED along the [001] zone axis.
This journal is © The Royal Society of Chemistry 2019
Dalton Trans.

View Article Online
Paper
Dalton Transactions
In the case of TiS2, the hexagonal crystal structure is and 63% S was obtained. For lt-TiS a ratio close to 1 : 1 was
formed according to the crystal form, the observed lattice obtained where Ti is 53% and S is 47%. However, for the ht-
angles and on the basis of the SAED (Fig. 8c1–3). Lattice TiS sample, the TEM-EDX unravels a difference in composition
spacing of the TiS2 crystals as determined from the HRTEM between the center and edge position of the hexagonal crystals.
(Fig. 8c2) revealed a spacing of 2.9 Å, 2.6 Å and 2.6 Å corres- The crystal center shows an almost ideal TiS2 stoichiometry
ˉ
ˉ
ponding to the (100), (011) and (111) planes, respectively.
(38% Ti and 62% S) while the edge is highly sulfur deficient
For the titanium monosulfides (lt- and ht-TiS) the SAED with 70% Ti and 30% S.
(Fig. 8d3 and e3) also shows a hexagonal pattern indicating the
similarity to the TiS2 unit cell. Streaking is indicated in the
SAED, which relies on crystallographic stacking faults in lt-TiS
Raman studies
Raman measurements of the different titanium sulfides
(Fig. 8d2 and d3) or short-range ordering, the latter being
observed in the HRTEM and SAED of ht-TiS (Fig. 8e2 and e3).
This might most probably refer to the insertion of Ti in the
van der Waals gap interstitials forming partially filled metal
layers.37,50,51 Interestingly, ht-TiS shows hexagonal symmetry
in the bright-field TEM (Fig. 8e1) while lt-TiS (Fig. 8d1) shows
more elongated structures indicating a preferential growth
direction along the crystallographic b-axis [010] (Fig. 8d2).
Lattice spacing as determined from the HRTEM of lt-TiS
(Fig. 8d2) shows distances of 2.03 Å and 2.89 Å which are in
accord with the (006) and (012) planes of Ti2.45S4, respectively,
while for ht-TiS (Fig. 8e2) a distance of 2.96 Å is obtained
which correlates with the (100) and (010) planes of Ti1.083S2.
TEM-EDX analyses of the TiS3 samples (both nanobelts and
powder) show an ideal stoichiometry with 25% Ti and 75% S
in the case of the nanobelts and 24% Ti and 76% S in the case
of the powder. For TiS2 a near ideal stoichiometry with 37% Ti
further confirmed the identities of the synthesized materials.
For the TiS3 nanobelts and for the TiS3 powder a similar
Raman spectrum was obtained, showing the four A1g modes
characteristic of TiS3 at 175, 298, 368, and 557 cm−1 (Fig. 9).52
In the case of TiS2, the two Raman active modes Eg (in-plane
vibrational mode) and A1g (out-of plane vibrational mode)
characteristic of TiS2 are observed at 222 and 331 cm−1
,
respectively which indicate that the measured position of the
bulk TiS2 sample, as detected by micro-Raman measurement,
contains about two individual layers of TiS2 as the position of
these modes is sensitive and characteristic to the number of
layers.53 Furthermore, a shoulder at about 387 cm−1 is
Fig. 10 UV-VIS-NIR DRS spectra (Kubelka–Munk function) and the
corresponding Tauc plot for the determination of the direct band gap of
TiO2 nanoparticles (EMPROVE® ESSENTIAL, Merck), TiS3 nanobelts, TiS3
powder, TiS2, lt-TiS and ht-TiS.
Fig. 9 Raman spectra of TiS3 nanobelts, TiS3 powder, TiS2, lt-TiS and
ht-TiS.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Dalton Transactions
Paper
observed and is most probably due to the defects in TiS2.53 vibrational mode relative to the A1g mode when compared to
Two further peaks are observed at 152 and 635 cm−1 and the Raman spectrum of TiS2 which is in accord with non-stoi-
correspond to TiO254 which is formed due to some surface oxi- chiometric titanium sulfides.55
dation of the sample when it is in contact with the atmo-
sphere. The Raman spectra of the non-stoichiometric titanium
sulfides (lt- and ht-TiS) show features that are similar to the
Raman spectrum of TiS2 but typically with some peak shift.
For lt-TiS mainly two peaks at 329 and 395 cm−1 are observed
which are analogous to the A1g mode of TiS2 and correspond
to the vibrations of sulfur atoms perpendicular to the sulfur
layer. An additional shoulder that is observed in TiS2 is due to
the defects obtained by the interlayer insertion of titanium
atoms thus resulting in phonon mode stiffening.53,55 The
peaks found at 153, 512 and 635 cm−1 correspond to anatase
UV-VIS-NIR studies
UV-VIS-NIR diffuse reflectance spectra (Fig. 10) give direct
information about the spectral absorptivity especially in the
visible region important for photocatalysis. The diffuse reflec-
tance spectra were recorded using the Kubelka–Munk function
F(R) which enables to determine the optical absorbance of the
solid material from its reflectance R according to:56
2
ð1 ꢀ RÞ
FðRÞ ¼
2R
54
TiO2 which is also formed due to surface oxidation due to
Besides, the Tauc plot was obtained in order to determine
the direct band gap of the different titanium sulfides. The
Tauc plot for the determination of the direct band gap was cal-
culated from the Kubelka–Munk function according to the fol-
lowing formula:56
contact with the atmosphere. For ht-TiS the similarity to the
TiS2 Raman spectrum is even more obvious, where the peaks
obtained at 228 and 343 cm−1 are analogous to the Eg and A1g
vibrational modes of TiS2 and even a shoulder similar to the
shoulder found in the TiS2 spectrum is observed at about
407 cm−1. Interestingly, there is an increase in the Eg
2
½FðRÞhνꢁ ¼ Aðhν ꢀ EgÞ
Fig. 11 (a) Comparison of the degradation profile of methyl orange using one mg of TiO2 nanoparticles, different titanium chalcogenides and a
ZnS@CNT nanocomposite31 photocatalyst under irradiation of simulated sunlight, (b) comparison of the degradation profile of methyl orange using
5 mg of each of TiO2 nanoparticles, TiS3 nanobelts and ht-TiS, (c) comparison of the degradation profile of methyl orange using one mg of TiS3
powder treated at 450 °C under argon for different time intervals and (d) recyclability of 5 mg TiS3 nanobelts over three cycles.
This journal is © The Royal Society of Chemistry 2019
Dalton Trans.

View Article Online
Paper
Dalton Transactions
where A is a constant, h is the Planck constant, ν is the fre-
quency and Eg is the band gap of the material. The diffuse
reflectance spectrum of TiO2 nanoparticles that were used
in this study for a comparison of the photocatalytic activity
shows only light absorption in the ultraviolet range of the
spectrum resulting in a wide band gap of 3.39 eV as obtained
from the corresponding Tauc plot (Fig. 10). TiS3 nanobelts and
the TiS3 powder showed similar absorption spectra where the
absorption extends to the near infrared region of about
1500 nm which is reflected in the low band gaps of 0.94 eV
and 0.98 eV obtained from the corresponding Tauc plots of
TiS3 nanobelts and TiS3 powder. In the case of TiS2 and the
non-stoichiometric titanium sulfides (lt-TiS and ht-TiS) their
absorption extends over the whole measured UV-VIS-NIR
region and so no band gap determination using the Tauc plot
method was possible (Fig. 10). This however indicates that
these sulfides can make full use of the whole sunlight
spectrum.
Photocatalytic activity of titanium chalcogenides
The studied low band gap materials are suggested to be able to
utilize most of the sunlight where the IR radiation of the sun-
light could activate the low band gap materials through two
possible pathways. One pathway could be indirect as IR radi-
ation is able to activate molecular vibrations and rotations of
the water molecules typically resulting in a rise in the tempera-
ture of the aqueous medium containing the photocatalyst. Due
to the small band gap the resulting thermal energy produced
by molecular vibrations and rotations is sufficient to induce
further electron–hole separations leading to an increase in the
photocatalytic efficiency itself. On the other hand, direct infra-
red excitation should be possible as IR radiation in the range
of 700–3000 nm which has an energy between 1.77 and 0.41
eV is still sufficient to induce charge separation in these low
band gap materials.
IR radiation accounts for 43% of the total solar radiation4
and would be very beneficial if it is efficiently used for photo-
catalysis. The design of materials that make use of IR radiation
in photocatalysis e.g. for water decontamination could show an
efficient way to use solar radiation in photocatalysis and even-
tually result in more efficient photocatalysts compared to TiO2.
Testing the different synthesized titanium chalcogenides for
their photocatalytic performance under simulated sunlight
indeed revealed the high activity of TiS3 (nanobelts and
powder) and also titanium monosulfides (lt- and ht-TiS) where
lt-TiS showed the highest photocatalytic activity (Fig. 11a). On
the other hand, other titanium chalcogenides (stoichiometric
TiS2, TiSe2 and TiTe) synthesized in this work show almost no
activity (Fig. 11a).
It is known that TiS2 and TiSe2 show semimetallic
behavior24,32,57 which results in no effective charge separation
thus explaining the absence or low photocatalytic activity of
these materials. On the other hand, TiS3 is found to be a semi-
conductor with a small band gap of about 1.2 eV8 which
enables good charge separation even at longer wavelengths
extending towards the IR-region (till about 1033 nm). While
Dalton Trans.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Dalton Transactions
Paper
one mg of the tested TiO2 nanoparticles, TiS3 (nanobelts and further differentiation, showing that ht-TiS and TiS3 have
powder) and ht-TiS showed similar profiles for the degradation higher photocatalytic activities than TiO2 (Fig. 11b). The
of methyl orange, increasing the photocatalyst amount gave different active titanium sulfides also show different degra-
Fig. 12 High-resolution XPS spectra showing (a) Ti 2p, (b) S 2p, (c) O 1s and (d) N 1s core level spectra of TiS3 nanobelts (fresh before photocataly-
sis), TiS3 nanobelts after one photocatalytic cycle and TiS3 nanobelts after three photocatalytic cycles. Spectra showing (e) Ti 2p, (f) S 2p, (g) O 1s
and (h) N 1s core level spectra of ht-TiS before and after one photocatalytic cycle.
This journal is © The Royal Society of Chemistry 2019
Dalton Trans.

View Article Online
Paper
Dalton Transactions
dation kinetics when compared to TiO2 where the titanium TiS3 nanobelts proved to be highly efficient photocatalysts,
sulfides reveal a zero order reaction behavior while TiO2 shows even exceeding the photocatalytic activity of TiO2.
a first order reaction behavior (Fig. S9‡). Interestingly, it was
The reusability of the active titanium sulfides was tested
observed for the active photocatalysts that they also lead to sig- and it shows a behavior where with an increasing number of
nificant degradation of the dye in the dark before the actual cycles an increase in the degradation activity was observed
photocatalytic experiment under simulated sunlight started (Fig. 11d and S11‡). EDX of the materials after photocatalysis
when compared to the TiO2 nanoparticles (Fig. 11a and b). (Fig. S12‡) shows an increase in the carbon and oxygen
Even when the catalysts TiS3 and non-stoichiometric titanium content due to the presence of organic decomposition pro-
sulfides (lt- and ht-TiS) were kept at room temperature in the ducts on the surface while the stoichiometry of the investi-
dark for longer times in contact with the dye solution, con- gated titanium sulfides is preserved. XPS of the materials after
siderable activity towards dye degradation could be observed photocatalysis (Fig. 12) mainly shows the presence of an oxi-
(Fig. S10,‡ adsorption effects can be ruled out here). Thus, dized surface even after one photocatalytic cycle which is not
heat is sufficient to result in the desired charge separation surprising due to the presence of dissolved oxygen and hydro-
resulting in a catalytic activity in the dark which is known as gen peroxide which was added in small amounts to the dye
thermocatalysis. This concept was also tested by external solution to serve as an additional oxygen source. However, this
heating while keeping the mixture in the dark. This results in does not degrade the performance of the photocatalyst but
a significant degradation of the dye as shown for TiS3 nano- could be even beneficial with the increasing number of cycles
belts at 45 °C in Fig. 11b. This suggests that the use of IR due to the formation of a titanium oxide/titanium sulfide
radiation by the catalyst can be through two ways: either heterostructure, thus enabling better charge separation by
through direct absorption of IR radiation by the photocatalyst decreasing the electron–hole recombination rate. This specu-
which results in charge separation or through indirect heat lation is strongly supported by the fact that there is an increase
transfer to the catalyst wherein heat is generated through the in the photocatalytic activity when the catalyst is re-used in
absorption of IR radiation by water molecules. This should successive photocatalytic cycles as seen in Fig. 11d and S11.‡
also result in charge separation and thus catalytic activity. Heterostructures are well known to increase photocatalytic
Therefore, these materials represent a photocatalyst and ther- activity due to their ability in the spatial separation of the elec-
mocatalyst at the same time resulting in a photo-thermocataly- trons and holes formed, thus preventing their recombination,
tic synergestic effect. Recently, some composite materials have which leads to the design and specific synthesis of different
been prepared to combine the photocatalytic and thermocata- heterostructures to be used in photocatalysis.27,31,68,69 In the
lytic properties of two different materials to achieve this syner- present case, the titanium oxide/titanium sulfide hetero-
gistic effect.58
structure is even formed spontaneously during its use in waste-
Table 1 gives a comparison of the activity of the active tita- water treatment. Further causes for the increasing activity with
nium chalcogenides of this study with other photocatalysts re-use could be the increase of surface area of the used
reported in the literature31,59–66 in the degradation of MO. material due to exfoliation with time while stirring it in the
Interestingly the herein presented titanium-based catalysts dye solution. Furthermore, organic degradation products can
show a superior activity of all catalysts where only one mg is be detected by XPS on the catalyst surface49,70–72 (Fig. 12b–d
sufficient to degrade MO almost completely within 120 min in and f–h) most probably stemming from the degradation of the
the case of the lt-TiS material. Based on the used ratios the used methyl orange dye.
new titanium sulfide photocatalysts outperform the thus far
reported materials.
Having seen the high activity of the titanium sulfide cata-
lysts it was interesting to check whether it is possible to
Conclusions
increase the catalytic efficiency of TiS3 through the production Titanium chalcogenides based on S, Se and Te show interest-
of sulfur vacancies even further. Sulfur vacancies result in the ing properties due to their narrow band gaps which could be
formation of mid-gap states67 which could be beneficial for beneficial for the design of new and different photo-thermoca-
the photocatalytic process. Sulfur vacancies were introduced talysts. In this work the binary titanium chalcogenides TiS3,
into TiS3 through heating at 450 °C under argon. The resulting TiS2, TiS, TiSe2 and TiTe were synthesized through chemical
material TiS3−x/20 min showed some increase in the photo- vapor transport reaction by employing different ratios of the
catalytic activity followed by a decrease for longer treatment powdered elements as starting materials at different synthesis
times where the TiS3−x/40 min sample has the same activity as temperatures. While the titanium dichalcogenides (TiS2, TiSe2)
the as-prepared TiS3 powder and TiS3−x/60 min has a lower and TiTe show semimetallic behavior which makes them poor
activity than the as-prepared powder (Fig. 11c). Thus, there are candidates for direct use in photocatalysis, they could be still
two possible centers which result in the increased catalytic beneficial for the design of heterostructures. On the other
activity of the titanium sulfides. On the one hand the presence hand, the non-stoichiometric TiS and TiS3 show a high photo-
of Ti3+ centers in non-stoichiometric titanium sulfides and on catalytic activity especially in the presence of small amounts of
the other hand the presence of sulfur vacancies could lead to a hydrogen peroxide. It was shown that Ti3+ present in non-stoi-
high catalytic activity in titanium sulfides. At the same time, chiometric titanium sulfides and a certain extent of sulfur
Dalton Trans.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Dalton Transactions
Paper
vacancies lead to an increase in the photocatalytic activity. The
small band gap of these materials also results in an additional
thermocatalytic activity and the ability to make use of the
whole sunlight spectrum. Thus, this work shows that the non-
stoichiometric titanium sulfides (lt-TiS and ht-TiS) along with
R. D’Agosta, I. J. Ferrer and A. Castellanos-Gomez, Adv.
Electron. Mater., 2015, 1, 1500126.
9 J. Kang, H. Sahin, H. D. Ozaydin, R. T. Senger and
F. M. Peeters, Phys. Rev. B: Condens. Matter Mater. Phys.,
2015, 92, 075413.
titanium trisulfide are very interesting candidates for efficient 10 J. Kang and L. W. Wang, Phys. Chem. Chem. Phys., 2016, 18,
photocatalysis and can easily outperform the activity of tita- 14805–14809.
nium dioxide. A great advantage of these materials is the spon- 11 F. A. Rasmussen and K. S. Thygesen, J. Phys. Chem. C, 2015,
taneous formation of an oxide layer on the surface of active 119, 13169–13183.
titanium sulfides leading to a titanium oxide/titanium sulfide 12 Materials Nanoarchitectonics, ed. K. Ariga and M. Ebara,
heterostructure. It is formed in successive photocatalytic cycles Wiley-VCH, 2018.
and obviously allows better charge separation leading to a 13 L. Wang and T. Sasaki, Chem. Rev., 2014, 114, 9455–9486.
highly enhanced photocatalytic activity. The tuning of the reac- 14 Photocatalysis and Water Purification: From Fundamentals to
tant ratios and the reaction temperatures employed enable the
creation of a wide range of non-stoichiometric titanium sul- 15 Y. Li, Y. L. Li, B. Sa and R. Ahuja, Catal. Sci. Technol., 2017,
fides with varying morphologies and good photocatalytic 7, 545–559.
Recent Applications, ed. P. Pichat, Wiley-VCH, 2013.
activities. Next, further investigations should help to elucidate 16 H. Zhang, ACS Nano, 2015, 9, 9451–9469.
the detailed mechanism and allow a fine tuning of the photo- 17 M. Arsentev, A. Missyul, A. V. Petrov and M. Hammouri,
catalytic performance of titanium sulfides as novel photo-
catalyst materials.
J. Phys. Chem. C, 2017, 121, 15509–15515.
18 N. Tanibata, T. Matsuyama, A. Hayashi and
M. Tatsumisago, J. Power Sources, 2015, 275, 284–287.
19 J. Wu, D. Wang, H. Liu, W.-M. Lau and L.-M. Liu, RSC Adv.,
2015, 5, 21455–21463.
20 A. Lipatov, P. M. Wilson, M. Shekhirev, J. D. Teeter,
R. Netusil and A. Sinitskii, Nanoscale, 2015, 7, 12291–
12296.
Conflicts of interest
There are no conflicts to declare.
21 J. O. Island, M. Barawi, R. Biele, A. Almazán,
J. M. Clamagirand, J. R. Ares, C. Sánchez, H. S. J. Van Der
Zant, J. V. Álvarez, R. D’Agosta, I. J. Ferrer and
A. Castellanos-Gomez, Adv. Mater., 2015, 27, 2595–2601.
22 M. Barawi, E. Flores, I. J. Ferrer, J. R. Ares and C. Sánchez,
J. Mater. Chem. A, 2015, 3, 7959–7965.
23 E. Flores, J. R. Ares, I. J. Ferrer and C. Sánchez, Phys. Status
Solidi RRL, 2016, 10, 802–806.
24 C. Hawkins and L. Whittaker-Brooks, ACS Appl. Nano
Mater., 2018, 1, 851–859.
Acknowledgements
The authors thank S. Heinschke and T. Bies (both at Eduard-
Zintl-Institut für Anorganische und Physikalische Chemie, TU
Darmstadt) for XRD measurements and thermogravimetric
analysis (TGA), respectively. XPS measurements were carried
out at the Karlsruhe Nano Micro Facility at KIT under the
KNMF program.
25 A. Zaleska-Medynska, Metal Oxide-Based Photocatalysis:
Fundamentals and Prospects for Application, Elsevier,
2018.
References
1 Visible-Light-Active Photocatalysis: Nanostructured Catalyst
Design, Mechanisms, and Applications, ed. S. Ghosh, Wiley- 26 M. Long and L. Zheng, Chin. J. Catal., 2017, 38, 617–624.
VCH, 2018.
27 K. S. Ranjith and T. Uyar, J. Mater. Chem. A, 2017, 5, 14206–
2 B. Luo, G. Liu and L. Wang, Nanoscale, 2016, 8, 6904–6920.
14219.
3 R. Asahi, T. Morikawa, H. Irie and T. Ohwaki, Chem. Rev., 28 R. M. A. Lieth and J. C. J. M. Terhell, in Preparation and
2014, 114, 9824–9852.
4 N. Alonso-Vante, Chalcogenide Materials for Solar Energy
Conversion, Springer, 2018.
Crystal Growth of Materials with Layered Structures, ed.
R. M. A. Lieth, Springer Science+Business Media, B.V,
1977, pp. 141–223.
5 J. Moan, in Radiation at home, outdoors and in the work- 29 H. P. B. Rimmington, A. A. Balchin and B. K. Tanner,
place, ed. D. Brune, R. Hellborg, B. Persson and R. J. Cryst. Growth, 1972, 15, 51–56.
Pääkkönen, Scandinavian Science Publisher, Oslo, 2001, 30 J. O. Island, A. J. Molina-Mendoza, M. Barawi, R. Biele,
pp. 69–85.
E. Flores, J. M. Clamagirand, J. R. Ares, C. Sánchez,
H. S. J. Van Der Zant, R. D’Agosta, I. J. Ferrer and
A. Castellanos-Gomez, 2D Mater., 2017, 4, 022003.
6 G. Liu, L. Wang, H. G. Yang, H. M. Cheng and G. Q. Lu,
J. Mater. Chem., 2010, 20, 831–843.
7 X. Chen, L. Liu, P. Y. Yu and S. S. Mao, Science, 2011, 331, 31 S. Okeil, J. Krausmann, I. Dönges, S. Pfleger, J. Engstler
746–750. and J. J. Schneider, Dalton Trans., 2017, 46, 5189–5201.
8 A. J. Molina-Mendoza, M. Barawi, R. Biele, E. Flores, 32 M. Chhowalla, H. S. Shin, G. Eda, L. J. Li, K. P. Loh and
J. R. Ares, C. Sánchez, G. Rubio-Bollinger, N. Agraït,
H. Zhang, Nat. Chem., 2013, 5, 263–275.
This journal is © The Royal Society of Chemistry 2019
Dalton Trans.

View Article Online
Paper
Dalton Transactions
33 A. V. Kolobov and J. Tominaga, Springer Series in Materials 53 P. C. Sherrell, K. Sharda, C. Grotta, J. Ranalli,
Science, Vol. 239: Two-dimensional Transition-metal
M. S. Sokolikova, F. M. Pesci, P. Palczynski, V. L. Bemmer
and C. Mattevi, ACS Omega, 2018, 3, 8655–8662.
54 T. Ohsaka, F. Izumi and Y. Fujiki, J. Raman Spectrosc., 1978,
7, 321–324.
55 M. Ishii, M. Saeki and I. Kawada, Phys. Status Solidi A,
1984, 124, K109–K112.
56 M. T. Uddin, Y. Nicolas, C. Olivier, W. Jaegermann,
N. Rockstroh, H. Junge and T. Toupance, Phys. Chem.
Chem. Phys., 2017, 19, 19279–19288.
Dichalcogenides,
Springer
International
Publishing,
Switzerland, 2016.
34 J. Dai, M. Li and X. C. Zeng, Wiley Interdiscip. Rev.: Comput.
Mol. Sci., 2016, 6, 211–222.
35 S. K. Srivastava and B. N. Avasthi, J. Mater. Sci., 1992, 27,
3693–3705.
36 R. R. Chianelli and M. B. Dines, Inorg. Chem., 1975, 14,
2417–2421.
37 Crystallography and crystal chemistry of materials with 57 C. Tan and H. Zhang, Chem. Soc. Rev., 2015, 44, 2713–
layered structures, ed. F. Lévy, D. Reidel Publishing
Company, Dordrecht, Holland, 1976.
38 J. Benard and Y. Jeannin, Nonstoichiometric Compounds,
American Chemical Society, 1963, pp. 191–203.
39 J. L. Murray, Bull. Alloy Phase Diagrams, 1986, 7, 156–163.
40 A. S. Pawbake, J. O. Island, E. Flores, J. R. Ares, C. Sanchez,
2731.
58 J. Zhang, S. Meng, X. Ye, C. Ling, S. Zhang, X. Fu and
S. Chen, Appl. Catal., B, 2017, 218, 420–429.
59 F. Ahmadi, M. Rahimi-Nasrabadi, A. Fosooni and
M. Daneshmand, J. Mater. Sci.: Mater. Electron., 2016, 27,
9514–9519.
I. J. Ferrer, S. R. Jadkar, H. S. J. Van Der Zant, 60 V. R. Posa, V. Annavaram, J. R. Koduru, V. R. Ammireddy
A. Castellanos-Gomez and D. J. Late, ACS Appl. Mater.
Interfaces, 2015, 7, 24185–24190.
41 J. I. Goldstein, D. E. Newbury, P. Echlin, D. C. Joy,
and A. R. Somala, Korean J. Chem. Eng., 2016, 33, 456–464.
61 Y. Li, H. Zhao and M. Yang, J. Colloid Interface Sci., 2017,
508, 500–507.
C. E. Lyman, E. Lifshin, L. Sawyer and J. R. Michael, 62 R. Cai, B. Zhang, J. Shi, M. Li and Z. He, ACS Sustainable
Scanning Electron Microscopy and X-ray Microanalysis,
Springer, Boston, MA, 2003.
42 H. S. Wong and N. R. Buenfeld, Cem. Concr. Res., 2006, 36,
1076–1082.
Chem. Eng., 2017, 5, 7690–7699.
63 B. Xu, Y. Li, Y. Gao, S. Liu, D. Lv, S. Zhao, H. Gao, G. Yang,
N. Li and L. Ge, Appl. Catal., B, 2019, 246, 140–148.
64 S. N. Basahel, T. T. Ali, M. Mokhtar and K. Narasimharao,
Nanoscale Res. Lett., 2015, 10, 73.
43 H. Hantsche, Scanning, 1989, 11, 257–280.
44 D. Gonbeau, C. Guimon, G. Pfister-Guillouzo, A. Levasseur, 65 Z. Tong, D. Yang, J. Shi, Y. Nan, Y. Sun and Z. Jiang, ACS
G. Meunier and R. Dormoy, Surf. Sci., 1991, 254, 81–89. Appl. Mater. Interfaces, 2015, 7, 25693–25701.
45 F. Fedorov, M. Vasilkov, A. Lashkov, A. Varezhnikov, 66 C. Chen, W. Cai, M. Long, B. Zhou, Y. Wu, D. Wu and
D. Fuchs, C. Kübel, M. Bruns, M. Sommer and V. Sysoev,
Sci. Rep., 2017, 7, 9732.
46 N. Zydziak, C. Hübner, M. Bruns and C. Barner-Kowollik,
Macromolecules, 2011, 44, 3374–3380.
47 M. E. Fleet, S. L. Harmer, X. Liu and H. W. Nesbitt, Surf.
Sci., 2005, 584, 133–145.
48 A. J. Huckaba, S. Gharibzadeh, M. Ralaiarisoa, C. Roldán-
Y. Feng, ACS Nano, 2010, 4, 6425–6432.
67 Z. Hu, Z. Wu, C. Han, J. He, Z. Ni and W. Chen, Chem. Soc.
Rev., 2018, 47, 3100–3128.
68 Y. Wang, Q. Wang, X. Zhan, F. Wang, M. Safdar and J. He,
Nanoscale, 2013, 5, 8326–8339.
69 J. Fu, J. Yu, C. Jiang and B. Cheng, Adv. Energy Mater.,
2018, 8, 1701503.
Carmona, N. Mohammadian, G. Grancini, Y. Lee, 70 F. Obstals, M. Vorobii, T. Riedel, A. de los Santos Pereira,
P. Amsalem, E. J. Plichta, N. Koch, A. Moshaii and
M. Bruns, S. Singh and C. Rodriguez-Emmenegger,
M. K. Nazeeruddin, Small Methods, 2017, 1, 1700250.
Macromol. Biosci., 2018, 18, 1700359.
49 C. Kleber, M. Bruns, K. Lienkamp, J. Rühe and 71 M. Bruns, C. Barth, P. Brüner, S. Engin, T. Grehl,
M. Asplund, Acta Biomater., 2017, 58, 365–375.
50 R. Moret, M. Huber and R. Comès, Phys. Status Solidi A,
1976, 38, 695–700.
C. Howell, P. Koelsch, P. Mack, P. Nagel, V. Trouillet,
D. Wedlich and R. G. White, Surf. Interface Anal., 2012, 44,
909–913.
51 A. I. Gusev, Phys.-Usp., 2006, 49, 693–718.
52 P. Gard, F. Cruege, C. Sourisseau and O. Gorochov,
J. Raman Spectrosc., 1986, 17, 283–288.
72 E. Blasco, M. Piñol, L. Oriol, B. V. K. J. Schmidt, A. Welle,
V. Trouillet, M. Bruns and C. Barner-Kowollik, Adv. Funct.
Mater., 2013, 23, 4011–4019.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2019