Synthesis, characterization and optoelectronic properties of chemically stable (CH3)3SPbI3?xBrx and (CH3)3SPbI3?xClx (x?=?0, 1, 2, 3) perovskites

Article abstract of DOI:10.1016/j.poly.2017.11.030

We report on the novel series of (CH₃)₃SPbI_3?xBr_x and (CH₃)₃SPbI_3?xCl_x (x = 0, 1, 2, 3) perovskite compounds. X-ray diffraction analysis shows that the solid solutions of (CH₃)₃SPbI_3?xBr_x (x = 0, 1, 2, 3) and (CH₃)₃SPbI_3?xCl_x (x = 0, 1, 2) crystallize in hexagonal symmetry (space group P6₃mc, No. 186) with 1D columns of face-sharing [PbX₆] octahedra. (CH₃)₃SPbCl₃ crystallizes in orthorhombic symmetry (space group Pnma, No. 62) forming 3D network of vertex- and face-sharing [PbCl₆] octahedra. Optical and vibrational properties were investigated using UV–vis reflectance, photoluminescence and Raman spectroscopy at room temperature. The compounds show high chemical stability in ambient air at temperatures up to 80 °C and under solar simulator, in contrast to the hygroscopic CH₃NH₃PbI₃ or CH(NH₂)₂PbI₃ that are commonly used in perovskite solar cells. First principles theoretical ab initio and efficient semiempirical extended Hückel calculations were performed to evaluate the energy band gap values, whose results are in good agreement with the experimentally determined values.

Full text of DOI:10.1016/j.poly.2017.11.030

Accepted Manuscript
Synthesis, characterization and optoelectronic properties of chemically stable
(CH3)3SPbI3-xBrx and (CH3)3SPbI3-xClx (x = 0, 1, 2, 3) perovskites
Andreas Kaltzoglou, Mohamed M. Elsenety, Ioannis Koutselas, Athanassios G.
Kontos, Kyriakos Papadopoulos, Vassilis Psycharis, Catherine P. Raptopoulou,
Dorothea Perganti, Thomas Stergiopoulos, Polycarpos Falaras
PII:
S0277-5387(17)30758-1
https://doi.org/10.1016/j.poly.2017.11.030
POLY 12932
DOI:
Reference:
Polyhedron
To appear in:
Received Date:
Accepted Date:
30 September 2017
24 November 2017
Please cite this article as: A. Kaltzoglou, M.M. Elsenety, I. Koutselas, A.G. Kontos, K. Papadopoulos, V. Psycharis,
C.P. Raptopoulou, D. Perganti, T. Stergiopoulos, P. Falaras, Synthesis, characterization and optoelectronic
properties of chemically stable (CH3)3SPbI3-xBrx and (CH3)3SPbI3-xClx (x = 0, 1, 2, 3) perovskites, Polyhedron
(2017), doi: https://doi.org/10.1016/j.poly.2017.11.030
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Synthesis, characterization and optoelectronic properties of chemically
stable (CH3)3SPbI3-xBrx and (CH3)3SPbI3-xClx (x = 0, 1, 2, 3) perovskites
Andreas Kaltzogloua,*, Mohamed M. Elsenetya, Ioannis Koutselasb, Athanassios G. Kontosa, Kyriakos
Papadopoulosa, Vassilis Psycharisa, Catherine P. Raptopouloua, Dorothea Pergantia, Thomas
Stergiopoulosc,*, Polycarpos Falarasa,*
a Institute of Nanoscience and Nanotechnology, National Centre for Scientific Research “Demokritos”,
15341, Agia Paraskevi Attikis, Athens, Greece
b Materials Science Department, School of Natural Sciences, University of Patras, 26504, Greece
c Laboratory of Physical Chemistry, Department of Chemistry, Aristotle University of Thessaloniki,
54124, Thessaloniki, Greece
e-mail addresses: a.kaltzoglou@inn.demokritos.gr, stergt@chem.auth.gr, p.falaras@inn.demokritos.gr
Telephone number (P. Falaras): +30-210-6503644
Fax number: +30-210-6511766
1

Abstract
We report on the novel series of (CH3)3SPbI3-xBrx and (CH3)3SPbI3-xClx (x = 0, 1, 2, 3) perovskite
compounds. X-ray diffraction analysis shows that the solid solutions of (CH3)3SPbI3-xBrx (x = 0, 1, 2, 3)
and (CH3)3SPbI3-xClx (x = 0, 1, 2) crystallize in hexagonal symmetry (space group P63mc, No. 186)
with 1D columns of face-sharing [PbX6] octahedra. (CH3)3SPbCl3 crystallizes in orthorhombic
symmetry (space group Pnma, No. 62) forming 3D network of vertex- and face-sharing [PbCl6]
octahedra. Optical and vibrational properties were investigated using UV-Vis reflectance,
photoluminescence and Raman spectroscopy at room temperature. The compounds show high chemical
stability in ambient air at temperatures up to 80 °C and under solar simulator, in contrast to the
hygroscopic CH3NH3PbI3 or CH(NH2)2PbI3 that are commonly used in perovskite solar cells. First
principles theoretical ab-initio and efficient semiempirical extended Hückel calculations were
performed to evaluate the energy band gap values, whose results are in good agreement with the
experimentally determined values.
Keywords: sulfonium cation; perovskite; luminescence; ab initio calculations; Extended Hückel
calculations
2

1. Introduction
Organohalide metal perovskite solar devices have attracted much attention in the photovoltaic
community, due to their efficiency in converting solar energy into electricity, which has reached values
over 21% in the last five years [1]. Lead halide perovskites adopt an APbX3 structure, where A is a
+
+
monovalent organic cation (e.g. methylammonium (CH3NH3 , MA+), formamidinium (CH(NH2)2 ,
FA+)) or an inorganic cation (e.g. K+, Rb+, Cs+), and X = Cl−, Br−, I−. They also form solid solutions by
changing the A-site or X-site ions so that their physical properties, such as band gap and charge carrier
diffusion length, are precisely tuned [2] [3]. Nevertheless, the chemical instability of APbX3 (A =
+
+
CH3NH3 , HC(NH2)2 , X = Br-, I-, Cl-) perovskites due to moisture, light and heat has long been the
main restriction for the industrialization of the corresponding devices [4].
Our recent study on (CH3)3SPbI3 has shown that this material is very stable in air [5]. In this context,
we investigate here the role of different halogen atoms in (CH3)3SPbX3 (X = Cl, Br, I). Also, the
luminescence properties are being reported, as they convey information about the excitonic nature of
these semiconductors, as well as for possible exciton engineering involving tuning the organic
component in order to match the luminescence properties of the inorganic network. Finally, theoretical
calculations are reported for relating observational data to models, while in particular the extended
Hückel calculations demonstrate a simple and efficient method for more than qualitative prediction of
the perovskites’ properties. A comparison is also made between various ab initio models, some of
which are the first successful application ab-initio models explaining the band gap origin of perovskites
based on face-sharing octahedra structures.
2. Experimental work
The reagents PbI2, PbBr2, PbCl2, (CH3)2S, methyl chloroformate as well as the solvents diethyl ether,
acetone, dimethylformamide (DMF), dimethyl sulfoxide (DMSO), benzyl bromide and
dichloromethane were purchased from Sigma Aldrich with purity >99.9% and were used without
further purification. Firstly, (CH3)3SX (X = Cl, Br, I) were synthesized. For (CH3)3SI, equimolar
amounts of CH3I and (CH3)2S were mixed in a round-bottomed flask and reacted at room temperature
for 12 hours. White crystals of (CH3)3SI were formed and washed twice with diethyl ether [6]. For
3

(CH3)3SBr, a solution of benzyl bromide (2.38 mL, 20 mmol) and DMSO (10 mL) contained in a
round-bottomed flask equipped with a condenser was heated at 80 oC for 24 hours. When the reaction
mixture was cooled to room temperature, trimethylsulfonium bromide (CH3)3SBr precipitated, which
was then thoroughly washed with acetone and dried with P2O5 under vacuum [7]. (CH3)3SCl was
synthesized by mixing methyl chloroformate (0.025 mol, 2.37 g) with dimethyl sulfide (0.05 mol, 3.1
g) in a pressure bottle at 80 oC overnight [8]. The produced white crystals were washed with diethyl
ether and dried with P2O5 under vacuum. Due to its hygroscopic nature, (CH3)3SCl was thereafter
stored in an Argon-filled glove box with O2 and H2O levels typically below 1 ppm.
Two different methods were incorporated for the synthesis of (CH3)3SPbI3-xBrx and (CH3)3SPbI3-xClx
perovskites. In the solid-state synthesis, equimolar amounts of PbX2 and (CH3)3SX were ground
together and loaded in a silica tube, which was flame-sealed under vacuum, and then heated in a
furnace at 150 °C for 72 hours. In the solution synthesis, equimolar amounts of the above-mentioned
reagents were dissolved in DMF, and heated at 80 °C for three hours. The resulting yellowish products
were washed with dichloromethane and dried in air.
X-ray powder diffraction (XRPD) analysis was performed for all perovskites using a Siemens D-500
diffractometer, that operates in Bragg-Brentano geometry with Cu Kα1 (λ = 1.5406 Å) and Cu Kα2 (λ =
1.5444 Å) radiation. Data were collected over the angular range 5° ≤ 2θ ≤ 80° counting for 3 seconds at
each step of 0.02° in detector position. Rietveld refinements were performed using the FULLPROF
software [9]. A needle-shaped single crystal of (CH3)3SPbCl3 (dimensions 0.02  0.03  0.3 mm) that
was collected after slow evaporation of the solvent was cooled at -103 °C and diffraction measurements
were collected on a Rigaku R-AXIS SPIDER Image Plate diffractometer using graphite-
monochromated Mo Kα radiation. Data collection (ω-scans) and processing (cell refinement, data
reduction and absorption correction) were performed using the CrystalClear program package [10]. The
structure was solved by direct methods using SHELXS-97 [11] and refined by full-matrix least-squares
techniques on F2 with SHELXL2014/6 [12]. Formula: C3H9SPbCl3, Formula weight = 390.7, Z = 8,
space group: Pnma, No. 62, a = 16.2359(5) Å, b = 13.2044(5) Å, c = 8.7900(5) Å, α = β = γ = 90°, V =
1884.45(11) Å3, ρcalcd. = 2.754 g cm–3, Absorption coeff., μ = 18.89 mm–1, F(000) = 1408, 3.054° ≤ θ ≤
27.000°, No. of independent reflections = 17446, No. of data/restraints/parameters = 2143/6/121,
Goodness-of-fit on F2 = 1.05, Final R indices [I > 2σ(I)] = 2.82%, Final R indices [all data] = 4.22%.
The two (CH3)3S units are symmetrically independent, but as they are close to symmetry elements, they
are distributed with half occupancies on both pairs of disordered sites and thus they were refined using
the Part -1 command. The hydrogen atoms for these groups were introduced at calculated positions. All
4

hydrogen atoms were refined isotropically with the calculated ones, as riding on their respective
bonded atoms. All non-hydrogen atoms were refined anisotropically.
Micro-Raman spectra were measured for (CH3)3SPbI3-xBrx and (CH3)3SPbI3-xClx (x = 0, 1, 2, 3) in the
30 − 3200 cm-1 range in backscattering configuration on a dispersive Renishaw in-Via Reflex
spectrometer with 250 mm focal length equipped with a high sensitivity, deep depletion CCD detector.
Excitation was performed by a near infrared (NIR) solid state laser emitting at a wavelength of 785 nm.
The laser beam was focused on the sample’s surface using a ×50 long focal distance objective lens of a
Leica DMLM microscope providing a power density of approximately 1 mW μm-2. Raman scattered
signal was filtered by a 30 cm-1 cut-off dielectric filter and analyzed with a 1200 lines/mm diffraction
grating. Measurements were carried out with circular polarization of the incident laser beam on
perovskite crystals aligned at 45° relative to the direction of the spectrometer slit in order to record
unpolarized Raman spectra. Spectral deconvolution was carried out by linear least squares fitting of the
Raman peaks to a mixture of Lorentzian and Gaussian line shapes.
UV-Vis spectra were recorded for the perovskites as well as their precursors (CH3)3SX (X = Cl, Br, I)
on a Hitachi U-3010 spectrophotometer equipped with an integrating sphere. The materials were
measured as 1 mm thick pellets in diffuse reflectance (R) mode and the band gap was determined using
the Kubelka-Munk equation: F(R) = (1-R)2 / 2R. Photoluminescence (PL) spectra were obtained from
solid pressed pellets, placed on quartz plates in a reflecting configuration, mounted on a Hitachi F-2500
FL spectrophotometer, employing a xenon 150 W lamp and a R928 photomultiplier. The excitation and
detection slits were set at 5 nm and the accelerating voltage was set to 700 V, while spectra were
acquired at 300nm/sec. Photoluminescence excitation was measured and corrected using the
instrument’s correction factor. Synchronous scans as well as 3D scans were employed to determine
particular details of the samples’ properties.
The photochemical stability experiments were performed under 1 sun (1000 W m-2) illumination from a
300W-Xe source (Oriel) in combination with AM 1.5G optical filter. The surface features of the
perovskites were analyzed on a Scanning Electron Microscopy Quanta 200 SEM instrument by
accelerating the electron beam at 25 kV. The chemical composition of the samples was estimated using
an Energy Dispersive X-ray (EDAX) spectroscopy detector.
Extended Hückel (EHT) calculations for all perovskites were made with an in-house code based on
Slater type orbitals centered on all atoms, composed of single zeta s and p orbitals. First-principles
calculations have been performed using CASTEP [13] (Cambridge-Sequential-Total-Energy-Package)
code by (LAD/CA-PZ, GGA/PBE, and GGA/PBESOL) functions. A plane-wave energy cutoff off 280
5

and 340 eV and Monkhorst-pack [14]. Ab initio DFT calculations, using the Accelrys materials studio
package, have been carried out to predict the optimized lattice parameters, band structure, density of
states (DOS) and partial density of states (PDOS) for (CH3)3SPbI2Br.
3. Results and Discussion
3.1 Structural analysis
(CH3)3SPbI3-xBrx perovskites were obtained either from solid-state or solution synthesis in high
crystallinity materials with hexagonal symmetry, same as in the previously reported room-temperature
modification of (CH3)3SPbI3 (space group P63mc, No. 186) [5]. No extra peaks of unreacted precursors
or other impurities were observed by XRPD analysis in all samples (Figure 1). A gradual shift of the
diffraction peaks to higher angles occurs with increasing Br content due to the decrease in lattice
parameters. Structural analysis was performed for (CH3)3SPbBr3 with the Rietveld method (Figure S1,
Table S1) using initially the structural model of (CH3)3SPbI3 [5] without further refinement of the
trimethylsulfonium group. All thermal displacement values were set to 0.03 Å2. The distorted, face-
sharing [PbBr6] octahedra form 1D chains along the c axis (Figure 2), and the trimethylsulfonium
cations occupy interstitial sites between the octahedra with their 3-fold rotation axis parallel to the
crystallographic c axis.
Figure 1. XRPD patterns for (CH3)3SPbI3-xBrx (x = 0, 1, 2, 3) in comparison with the theoretical pattern
of hexagonal (CH3)3SPbI3.
6

Figure 2. Crystal structure of the 2 x 2 x 2 unit cell (CH3)3SPbBr3 across the c axis. Thick blue lines
denote the unit cell edges.
The (CH3)3SPbI3-xClx (x = 1, 2, 3) perovskites were also synthesized by solid-state and solution
methods. No extra peaks of unreacted precursors or other impurities were observed by XRPD analysis
for these samples, too. For x = 1, 2, Bragg peaks of hexagonal symmetry (space group P63mc, No. 186)
are recorded with a gradual shift to higher angles with increasing Cl content (Figure 3).
Figure 3. XRPD patterns for (CH3)3SPbI3-xClx in comparison with the theoretical patterns of hexagonal
(CH3)3SPbI3 and orthorhombic (CH3)3SPbCl3.
In contrast, according to single-crystal X-ray diffraction analysis (CH3)3SPbCl3 crystallizes in
orthorhombic symmetry (Figure 4, Tables S2-S4). For a better description of the structure we have to
consider as a formula unit the double of the above (which corresponds to the content of the asymmetric
unit), i.e. [(CH3)3S]2Pb2Cl6. The two (CH3)3S units are symmetrically independent, but as they are close
to symmetry elements, they are distributed with half occupancies on both pairs of disordered sites. That
7

consisting of S1, C1, C2, C3 atoms are close to the mirror plane (x,-0.25,z) and those of S2, C3, C4 and
C5 atoms are close to the center of symmetry (0.5,0,1). Units of two face-sharing [Pb2Cl6] octahedra
(through Cl1, Cl2, Cl3 atoms) connect to adjacent six units (three above and three below the plane of
the central unit) through common vertices (through Cl5 and pairs of symmetry equivalent Cl4 atoms).
The three dimensional arrangement of pairs of face-sharing octahedral in the compound under study
resembles that of SrMnO3 structure (a 4H four layer hexagonal perovskite, with space group P63/mmc
(No. 194) [15,16,17]) and the one to one correspondence of the different structure units of both
structures renders them isostructural. In the present structure the trimethylsulfonium group, the Pb and
Cl atoms have replaced the Sr atoms, Mn and oxygen atoms, respectively.
The difference in structure type among (CH3)3SPbCl3 and the rest of the perovskites under
investigation is attributed to the large decrease in crystal density with decreasing halogen size. The
actual density of the orthorhombic (CH3)3SPbCl3 is 2.754 g cm-3 (from single-crystal XRD data),
whereas for the hexagonal (CH3)3SPbI3, (CH3)3SPbBr3 and (CH3)3SPbICl2 perovskites, these values
equal to 3.49, 3.12 and 2.71 g cm-3, respectively (from XRPD data and Rietveld refinement). A
hypothetic 1D-hexagonal (CH3)3SPbCl3 would allow for very larger voids among the inorganic
columns and the organic cations, leading to even lower crystal density which is thermodynamically
unstable. Therefore, (CH3)3SPbCl3 has higher structural similarity with ((CH3)3S)4Pb3Br10 [28] rather
than with its heavier (CH3)3SPbX3 analogues. It is noteworthy also that (CH3)3SPbCl3 is
thermodynamically stable in a large temperature range (-103 °C to 25 °C) and retains its orthorhombic
symmetry, in contrast to (CH3)3SPbI3 that undergoes structural phase transitions below -5 °C [5].
Figure 4. Crystal structure of (CH3)3SPbCl3 with partially labeled atoms. H atoms are omitted for
8

clarity. Thick blue lines denote the unit cell edges.
3.2 Vibrational and optoelectronic properties
The micro-Raman spectra for the novel (CH3)3SPbI3-xBrx and (CH3)3SPbI3-xClx perovskite materials
(Figure 5) show the vibrational modes of the organic moiety (CH3)3S+ at ca. 3000, 2915, 1410, 1045,
730, 655, 280 cm-1 with a very small red shift (1 – 4 cm-1) with increasing Br or Cl content. On the
other hand, the vibrational modes of the inorganic framework appear at low frequencies (30 – 200
cm-1). The strongest mode for (CH3)3SPbI3 is the Pb-I stretching found at 100ꢀcm−1 [5] and it shifts to
higher frequencies in the series of (CH3)3SPbI3-xBrx and (CH3)3SPbI3-xClx due to the decrease of the
reduced mass of the Pb-Br and Pb-Cl bonds, respectively. This behavior is in accordance with similar
systems such as CH3NH3PbX3 (X = Cl, Br, I) [18], where the spectra of the single halide samples
present less and narrower low frequency Raman bands in comparison to the mixed halide samples
which have mixed Pb-X bond character. Nevertheless, the spectra of the CH3NH3PbI2X samples are
dominated by the Pb-I vibration at about 100 cm-1 suggesting possible segregation of the CH3NH3PbI3
phase in nanodomains.
Figure 5. Raman spectra of (CH3)3SPbI3-xBrx and (CH3)3SPbI3-xClx (x = 0, 1, 2, 3). The inset zooms in
the 30 – 250 cm-1 region.
Electronic properties of (CH3)3SPbI3-xBrx and (CH3)3SPbI3-xClx were investigated using UV-Vis and
photoluminescence spectroscopy at room temperature. The diffuse absorption spectra acquired from
transforming the diffuse reflectance spectra Kubelka-Munk transforms of (CH3)3SPbI3-xBrx and
(CH3)3SPbI3-xClx are presented in Figure 6. It is possible to find the energy band gap value (Eg) by
9

considering the x-axis intercept of a linear extrapolation of the low energy part of the respective
spectra, however, this is not an axiom. It is easy to see the intercepts for the I, Br and Cl analogues, yet
those found by extrapolation of the mixed halide analogue spectra would have provided quite small Eg
values. In many cases, it is more useful to keep as Eg the energy of the first absorption valley occurring
at energy higher than the excitonic peak. The spectra reveal that the absorption onset energy increases
for (CH3)3SPbI3-xBrx from 2.5 to 3.2 eV with increasing Br content and for (CH3)3SPbI3-xClx from 2.5
to 3.6 eV with increasing Cl content. However, the values of the absorption edge cannot be related to
the Eg of the materials, for two reasons: i) due to the excitonic states present and ii) due to the fact they
contradict the values provided for such face sharing [PbX6] quantum wires found in the literature
[19,20,21,22,23]. For example, the (CH3)3SPbBr3 semiconductor has similar structure of face-sharing
PbBr6 octahedra as in [19,24,25], where in the latter refs. an excitonic peak was observed at 328 nm,
thus, the Eg of such quantum wires should be least 3.8 eV, and cannot be related to the lowest energy
absorption appearing here at 3.2 eV. Therefore, the Eg values are estimated by drawing the tangential
lines and taking the intercept to the x-axis not from the lowest energy bands but from the first high
energy band in each case (see Figure 6a). For the mixed halides, these values are being found from the
position of the first absorbance valley towards high energies; for the pure I-, Br-, Cl-compounds the
two methods yield almost the same Eg. For (CH3)3SPbI3 this estimation leads to Eg of 3.15 eV in
accordance to [5] while for (CH3)3SPbBr3 to 3.68 eV. It is anticipated that the mixed I/Br perovskites
will have energy gap values varying in between as in the other corner-sharing perovskite materials [18].
The as above discussed Eg values are collected in Table 1. Similar conclusions could be drawn for the
spectra of (CH3)3SPbI3-xClx, despite the fact that (CH3)3SPbCl3 has different structural symmetry from
the rest. Again, it can be safely deduced from the optical spectra of the Cl-I mixed variants that once
iodine is incorporated as a mixed ion along with Cl, the optical absorption spectra are not distinctively
different from each other, neither from those based on the I-Br materials since the onset of absorption
starts at almost the same energy values.
10

a)
b)
Figure 6. Diffuse reflectance UV-Vis spectra in Kubelka-Munk units for (a) (CH3)3SPbI3-xBrx (b)
(CH3)3SPbI3-xClx (x = 0, 1, 2, 3).
Table 1. Band gap values for (CH3)3SPbI3-xBrx and (CH3)3SPbI3-xClx (x = 0, 1, 2, 3).
Compound
Eg (eV)
3.15
Compound
Eg (eV)
(CH3)3SPbI3
(CH3)3SPbI2Br
(CH3)3SPbIBr2
(CH3)3SPbBr3
3.29
(CH3)3SPbI2Cl ca. 3.2
(CH3)3SPbICl2 ca. 3.3
3.37
ca. 3.75
(CH3)3SPbCl3
3.75
The PL spectra of the reported materials (CH3)3SPbX3 (X = I, Br, Cl) are presented in Figure 7, yet the
luminescence signal of the various one-dimensional face sharing octahedral wires is not strong
compared to one dimensional corner sharing octahedra-based quantum wires. PL spectroscopy has
been used here with excitation energies above 3.8 eV for all materials, while photoluminescence
excitation spectra (PLE) have been acquired by monitoring the tails of the low and high energy
luminescence peaks. The material for X = I, shows a broad PL centered around 470 nm and a weak
signal centered at 382 nm. For the case X = Br, the same set of peaks appears, where the high energy
peak at 379 nm peak is much stronger than in the case of the I, while for the X = Cl case, the high
energy peak though appears strong and its center has blue shifted to at 364 nm. The broad band
appearing around 470 nm for all three cases of X = I, Br, Cl is due to the organic component, however,
the high energy peaks of the semiconductors are not related in any of the spectra to the pristine organic
component. On the other hand, (CH3)3SCl exhibits PL signal centered at 380 nm, thus, the Cl analogue
of the reported compounds is believed not to exhibit PL signal due to the inorganic network, or more
11

possibly energy is being transferred from the inorganic onto the organic network appearing as organic
molecule luminescence. The above described behavior is unique in the perovskite semiconductors. The
small PL signal for the I at 382 nm (3.24 eV), for Br at 379 nm (3.27 eV) could be related to excitonic
states, however, for the case of I this would imply some exciton higher than the observed band gap,
which is discussed in terms of the band structure which shows two band gaps, the direct being slightly
higher than the lowest indirect. In all cases, the signal of the organic moiety falls close to that of the
inorganic lattice absorption and energy transfer effects are probably engaged in these systems.
In order to further resolve spectroscopic details, synchronous luminescence scans have been recorded
and presented in Figure 7 in relation to the PL data. In these spectra, (CH3)3SPbI3 has shown
absorbance related peaks at 380 nm and a broad band centered around 470 nm, while the Cl analogue
shows a strong broad absorption peak around 370 nm. Thus, it appears that the broad band at 470 nm in
the PL and synchronous scans of the precursor organic molecule, is strong in the case of (CH3)3SPbI3,
while it is diminished for the Br and Cl analogues; in fact, it is almost vanished in the Cl case. Thus, we
conclude that the organic component, although it shows a broad structure around 470 nm, its signal is
being progressively weakened as the materials shift from I to Br and then to Cl. Moreover, in the same
progression of materials the higher energy peaks at around 320 nm also change, indicating that the
organic molecule is not on its own responsible for this complex PL signals observed, but it is the
interaction of the organic to the inorganic component that creates these complex signals.
Figure 7. PL (solid lines) and synchronous (dashed lines) spectra for (a) (CH3)3SPbI3, (b) (CH3)3SPbBr3
and (c) (CH3)3SPbCl3 (λexc = 300, 330, 250 nm, respectively).
3.3 Stability and scanning electron microscopy
12

Poor stability of organic–inorganic perovskite materials in humid conditions has hindered the success
of perovskite solar cells in real-life applications [26]. The compounds are stable in air even after
several months and do not dissolve or react with water. Moreover, XRPD analysis of the samples after
exposure to 80 °C for 24 hours in ambient air shows no signs of degradation (Figure S2). Regarding
photochemical stability, thin films of the compounds were deposited from DMF solution on
glass/compact TiO2 substrates (glass/TiO2/perovskite). XRPD analysis of the films after exposure to a
solar simulator for 8 hours in ambient air shows no perovskite degradation (Figure S3).
The morphology of the (CH3)3SPbI3-xBrx and (CH3)3SPbI3-xClx bulk samples was analyzed using
Scanning Electron Microscopy and their chemical composition was estimated by EDAX. (CH3)3SPbBr3
appears as well structured rods (Figure S4), whereas (CH3)3SPbCl3 forms crystallites with almost
spherical shape (Figure S5). Figures S6 and S7 show the EDAX spectra of (CH3)3SPbI3-xBrx and
(CH3)3SPbI3-xClx (x = 1, 2, 3) perovskites, confirming their Pb/X atomic ratio, e.g. for (CH3)3SPbBr3
with atomic percentage Pb 9.45%, Br 28.28% (Pb/Br 1:2.99 atomic ratio). Noteworthy, I/Br and I/Cl
ratios are almost constant on different spots of each mixed-halide sample, which proves that there are
no inhomogeneous domains in the bulk materials.
3.4 EHT calculations
In order to evaluate the band structure and energy band gap of the above reported materials as well as
to understand the interplay of the quantum wires with the surrounding organic moiety, Extended
Hückel calculations have been performed. In particular, given that this type of calculation is dependent
on parameters, the methodology in [17] has been used, where the values of [27] have been utilized with
the exception of the exponent of the halogen p Slater type orbitals which was varied in order to match
the CH3NH3PbX3 structure band gap to the experimental value. Some variation in these parameters
from those reported in [23], is due to the different band gaps that were once used derived from optical
experiments. Any other effort to vary other exponent is futile, added to the fact that such effort is not
bearing any physical meaning since the halogen is the inorganic network’s connecting component.
Therefore, the Eg values for the 3D perovskite structures X = I, Br, Cl that have been fitted are 1.5, 2.3,
2.88 eV with found Slater exponents for the outer p orbitals to be 2.47, 2.321, 1.9275, respectively. It is
stressed that the quantum wires in this work are made out of face-sharing octahedra, where the 3D
perovskites are corner sharing octahedral structures, thus, the parameters are expected to be optimum,
yet give a more than quantitative relation between structure and band gap as well as band dispersion
curves. The EHT calculations have been performed either by keeping the organic and inorganic
13

component, or by only using the inorganic component which had been charged negatively with the
electrons donated by the organic moieties. In all cases, the organic component does not change the
HOMO/LUMO energy offsets nor their dispersion except in the case of I and Br where the introduction
of the organic component introduces a shift to the two indirect gap k point positions. (CH3)3SPbCl3 is
found to have a direct band gap, while all I and Br analogues appear to have an indirect band gap as
well, however, with a slightly larger direct band gap. The calculated band dispersion diagrams can be
found in Figure 8.
a)
b)
14

c)
Figure 8. Calculated band structures along selected path in the Brillouin Zone for (CH3)3SPbX3: a) X =
I, b) X = Br and c) X = Cl.
The calculated band gaps, without including the organic component, are found at 3.51, 4.14 and 4.24
eV for (CH3)3SPbI3, (CH3)3SPbBr3 and (CH3)3SPbCl3, respectively. The iodide and bromide show
indirect band gap, however the difference from the direct to indirect is much less than the thermal
activation energy, while the chloride is direct band gap semiconductor. Also, in order to test the above
calculations, we repeated the same type of method for ((CH3)3S)4Pb3Br10 appeared in [28] and
[C6H5CH2SC(NH2)2]4Pb3I10 in [29] which are made out of face-sharing octahedra broken or infinite
wires, respectively. Moreover, EHT calculations were performed on the mixed halide analogues,
assuming various configurations for the disordered I/Br or I/Cl atoms. For the first compound which is
the second reported perovskite based on trimethylsulfonium, our calculations with the specific
exponents, as shown above predict a gap of 3.58 eV, i.e. at 346 nm. This is in accordance with the
optical spectra in the same reference as well as with previously reported results [30]. Calculations of
the second compound, using the exponents described above, yielded a band gap of 4.43 eV (279 nm),
while the organic part orbital energies appear within the band gap; these results are in agreement with
the [21] (Figure 6), where the excitonic peak appears at 310 nm and the gap at about 4.3 eV. Also,
calculations for the mixed halide analogue (CH3)3SPbIBr2, for various positions of the Br and I within
the unit cell predict band gap values from 2.9 to 3.5 eV, without assuming any supercell structure.
These results are again in accordance with the optical spectra in this work, in the sense that all EHT
band gap values computed coincide with the second peak of the experimental absorption spectra. It is
predicted that the mixed halides are expected to have varying band gap values and it is natural that the
experimental absorption spectra are so broad.
15

3.5 First-principles DFT calculations
Ab initio DFT calculations on (CH3)3SPbI2Br were carried out to predict the optimized lattice
parameters, band structure, total density of states (TDOS) and partial density of states (PDOS), using
different functions (LAD/CA-PZ, GGA/PBE, GGA/PBESOL) with 240 eV and 380 eV energy cutoffs.
(CH3)3SPbI2Br is a characteristic example of mixed halide perovskite in which the site occupancy
factors are equal to SOF(I) = 2/3 and SOF(Br) = 1/3. We discuss here the electronic properties of
(CH3)3SPbI2Br from GGA/PBESOL with 380 eV energy cutoff (Figure 9).
Figure 9. Total (bottom) and partial (top) density of states for (CH3)3SPbI2Br.
It appears that p-orbitals have higher contribution than s-orbitals at the top of the valence band. The
strong hybridization of orbitals can shift the location of the valence bands to the lower energy area with
wider peaks [31]. The PDOS for lead and halogen atoms indicate a weaker bonding between them. At
the top of the valence band, the main contribution is from iodine and bromine (5p and 4p states,
respectively) with overlapping of Pb 6s states. At the bottom of the conduction band, the main
16

contribution is from X 5s, 6s, and Pb 6p. On the other hand, the organic cation has little contribution to
the top of valence band and bottom of the conduction band (Figure S8). The electronic band structures
(Figure S9) and DOS-PDOS (Figure S10) of hexagonal (CH3)3SPbI2Br using different functions
(LAD/CA-PZ, GGA/PBE, GGA/PBESOL) with 240 eV and 380 eV energy cutoffs have been
evaluated. The computed band gaps from the LAD/CA-PZ, GGA/PBE results are in good agreement
with the experimental Eg, but GGA/PBESOL results are lower than experimental value. Our theoretical
results support the strong relationship between structure properties and energy band gap of different
compounds. Figure S9 shows significant dispersion associated with the face-sharing networks of
[PbX6] octahedra. For example, the obtained values from DFT/ LAD/CA-PZ calculation point to an
indirect band gap of 3.38 eV with the minimum from the K to G symmetry point.
4. Conclusions
This class of sulfonium-containing lead halides has high potential for semiconducting applications due
to its high chemical stability in ambient conditions. There is an unusual effect of the halogen atom, as
(CH3)3SPbI3-xClx (x = 0, 1, 2) form solid solutions with 1D octahedral columns, whereas (CH3)3SPbCl3
forms a 3D network of interconnected octahedra. Although, the electronic band gaps are over 3 eV, and
therefore not applicable as absorbers in solar cells, these perovskites are solution processable and may
be used as hole transporters in solar cells. Moreover, semiempirical extended Hückel calculations are
proved to be efficient as a method for band gap calculation in this class of materials.
Appendix A. Supplementary data
CCDC 1577364 contains the supplementary crystallographic data for (CH3)3SPbCl3. These data can be
obtained free of charge via http://www.ccdc.cam.ac.uk/conts/retrieving.html, or from the Cambridge
Crystallographic Data Centre, 12 Union Road, Cambridge CB2 1EZ, UK; fax: (+44) 1223-336-033; or
e-mail: deposit@ccdc.cam.ac.uk.
Acknowledgements
We acknowledge financial support from European Union’s Horizon 2020 (Marie Curie Innovative
Training Network MAESTRO/764787). M. Elsenety is financially supported by Science Achievement
Scholarship of High Education Ministry of Egypt in cooperation with the Hellenic Ministry of Foreign
Affairs for his PhD Scholarship. D. Perganti and T. Stergiopoulos acknowledge financial support from
the “RESEARCH PROJECTS FOR EXCELLENCE IKY/SIEMENS” Programme.
17

References
[1] J.S. Manser, J.A. Christians, P.V. Kamat, Chem. Rev. 116 (2016) 12956.
[2] S.F. Hoefler, G. Trimmel, T. Rath, Monatshefte Für Chemie 148 (2017) 795.
[3] C. Wehrenfennig, G.E. Eperon, M.B. Johnston, H.J. Snaith, L.M. Herz, Adv. Mater. 26 (2014)
1584.
[4 ] J.S. Chen, D.Z. Liu, M.J. Al-Marri, L. Nuuttila, H. Lehtivuori, K.B. Zheng, Science China-
Materials 59 (2016) 719.
[5] A. Kaltzoglou, C.C. Stoumpos, A.G. Kontos, G.K. Manolis, K. Papadopoulos, K.G. Papadokostaki,
V. Psycharis, C.C. Tang, Y.-K. Jung, A. Walsh, M.G. Kanatzidis, P. Falaras, Inorg. Chem. 56 (2017)
6302.
[6] E.J. Corey, M. Chaykovsky, J. Am. Chem. Soc. 87 (1965) 1353.
[7] K.H. Park, M.S. So, Y.W. Kim, Bull. Korean Chem. Soc. 26 (2005) 1491.
[8] B. Byrne, L.M.L. Lawter, Tetrahedron Lett. 27 (1986) 1233.
[9] T. Roisnel, J. Rodriguez-Carvajal, Fullprof, Version Sept. 2012, France.
[10] Rigaku/MSC CrystalClear, R.M.I., The Woodlands, Texas, USA, 2005.
[11] G.M. Sheldrick, Acta Crystallogr. Section A 64 (2008) 112.
[12] G. Sheldrick, Acta Crystallogr. Section C 71 (2015) 3.
[13] S.J. Clark, M.D. Segall, C.J. Pickard, P.J. Hasnip, M.J. Probert, K. Refson, M.C. Payne, Z.
Kristall. 220 (2005) 567.
[14] S. Froyen, Phys. Rev. B 39 (1989) 3168.
[15] R. Sondena, S. Stolen, P. Ravindram, T. Grande, N.L. Allan, Phys. Rev. B 75 (2007) 184105.
[16] P.D. Battle, T.C. Gibb, C.W. Jones, J. Solid State Chem. 74 (1988) 60.
[17] I.B. Koutselas, L. Ducasse, G.C. Papavassiliou, J. Phys.:Condens. Matter 8 (1996) 1217.
[18] R.G. Niemann, A.G. Kontos, D. Palles, E.I. Kamitsos, A. Kaltzoglou, F. Brivio, P. Falaras, P.J.
Cameron, J. Phys. Chem. C 120 (2016) 2509.
[19] G.C. Papavassiliou, G.A. Mousdis, I.B. Koutselas, Synthetic Metals 121 (2001) 1339.
[20] G. Papavassiliou, I. Koutselas, A. Terzis, C.P. Raptopoulou, Mat. Res. Soc. Symp. Proc. 358
(1995) 283.
[21] G.C. Papavassiliou, G.A. Mousdis, C.P. Raptopoulou, A. Terzis, Z. Naturforsch. 54b (1999)
1405.
[22] G.B. Gridunova, E.A. Ziger, V.M. Koshkin, S.V. Lindeman, I.T. Struchkov, V.E. Shklover, Dokl.
Akad. Nuak SSSR 278 (1984) 656.
18

[23] Y.-J. She, S.-P. Zhao, Z.-F. Tian, X.-M. Ren, Inorg. Chem. Comm. 46 (2014) 29.
[24] Z. Tang, A.M. Guloy, J. Am. Chem. Soc. 121 (1999) 452.
[25] H. Krautscheid, F. Vielsack, Z. Anorg. Allg. Chem. 623 (1997) 259.
[26] Q.D. Tai, P. You, H.Q. Sang, Z.K. Liu, C.L. Hu, H.L.W. Chan, F. Yan, Nature Comm. 7 (2016)
11105.
[27] P. Alemany, S. Alvarez, Inorg. Chem. 31 (1992) 4266.
[28] M.D. Smith, B.L. Watson, R.H. Dauskardt, H.I. Karunadasa, Chem. Mater. 29 (2017) 7083.
[29] C.P. Raptopoulou, A. Terzis, G.A. Mousdis, G.C. Papavassiliou, Z. Naturforsch. 57b (2002) 645.
[30] G.C. Papavassiliou, Progr. Inorg. Chem. 25 (1997) 125.
[31] Y. Wang, T. Gould, J.F. Dobson, H.M. Zhang, H.G. Yang, X.D. Yao, H. J. Zhao, Phys. Chem.
Chem. Phys. 16 (2014) 1424.
19

Graphical abstract
The novel (CH3)3SPbI3-xBrx (x = 0, 1, 2, 3) and (CH3)3SPbI3-xClx (x = 0, 1, 2) perovskites form solid
solutions with columns of [PbX6] octahedra, whereas (CH3)3SPbCl3 forms a 3D network of [PbCl6]
octahedra. All compounds are very stable in air and they exhibit luminescence and band gaps >3 eV.
20