See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/364257648

Pd nanoparticle-mediated acetone sensing performance improvement of

SnO2 substrate: A combined DFT and experimental study

Article in Current Applied Physics · October 2022

DOI: 10.1016/j.cap.2022.10.003

CITATIONS READS

6 285

6 authors, including:

Arunangshu Biswas Oishika Pradhan

Dr. Reddy’s Institute of Life Sciences Hyderabad International Institute of Information Technology, Hyderabad
11 PUBLICATIONS 72 CITATIONS 1 PUBLICATION 6 CITATIONS

SEE PROFILE SEE PROFILE

Anand Thati Debdyuti Mukherjee

International Institute of Information Technology, Hyderabad New York University
2 PUBLICATIONS 42 CITATIONS 26 PUBLICATIONS 740 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Arunangshu Biswas on 18 October 2022.

The user has requested enhancement of the downloaded file.

 Current Applied Physics 44 (2022) 131–143

Contents lists available at ScienceDirect

Current Applied Physics

journal homepage: www.elsevier.com/locate/cap

Pd nanoparticle-mediated acetone sensing performance improvement of

SnO2 substrate: A combined DFT and experimental study
Arunangshu Biswas a, Oishika Pradhan a, Anand Thati b, Debdyuti Mukherjee c, Tapan K. Sau a, *,
Shubhajit Roy Chowdhury d
a
Center for Computational Natural Sciences and Bioinformatics, International Institute of Information Technology-Hyderabad, Gachibowli, Hyderabad, 500032, India
b
Center for VLSI and Embedded Systems Technology, International Institute of Information Technology-Hyderabad, Gachibowli, Hyderabad, 500032, India
c
Institute of Chemistry, The Hebrew University of Jerusalem GivatRam, Edmon J Safra Campus, Jerusalem, 9190401, Israel
d
School of Computing and Electrical Engineering, Indian Institute of Technology, Mandi, Himachal Pradesh, India

A R T I C L E I N F O A B S T R A C T

Keywords: The article presents a combined theoretical and experimental study attempting to show how Pd nanoparticles
Sensor (NPs) loading onto SnO2 substrate improves the acetone gas sensing performance. Pristine nanostructured SnO2
Acetone and Pd nanoparticles (Pd NPs) loaded SnO2 substrates have been prepared, characterized, and their acetone
Ethanol
sensing performances have been measured. Experimental measurements have shown that Pd NP loading onto
SnO2
Pd nanoparticles
SnO2 suppresses the interfering effects of ethanol, water vapors, etc., and enhances the acetone sensor response,
DFT reversibility, response/recovery speeds, and signal-to-noise ratio. Various parameters like the adsorption energy,
HOMO–LUMO energy gap, charge distribution, polarizability change, electrophilicity index, global hardness,
etc., of several model systems, have been computed by using DFT. The computed parameters have been corre­
lated with the conductivity, local reactivity, sensor response and selectivity, response/recovery times, etc., of the
systems to understand the molecular-level effects of the Pd NP loading onto the SnO2 on the gas sensing process.

1. Introduction like good sensitivity, selectivity, and shorter response and recovery
times. The choice of sensing substrates plays a critical role in deciding
Detecting and monitoring various gases and volatile organic com­ the performance of the sensor, along with their cost, robustness, ease of
pounds (VOCs) find applications in the fields of safety and security, use, etc. It has led researchers to explore a variety of sensor materials for
environmental and industrial monitoring, agriculture and automotive gas-sensing devices [3–13]. Semiconducting metal oxides (SMOs) are
industry, and medical diagnosis and monitoring [1,2]. Although various among the favorite classes of materials that show the chemoresistance
sophisticated analytical tools and techniques are available for the ac­ property upon gas adsorptions [3–10]. For example, SnO2 is an n-type
curate determination of low concentrations of gases and VOCs, there are SMO widely used for detecting several reducing and oxidizing gases and
several disadvantages associated with such sophisticated instruments. VOCs [14–17]. SMO materials also have other favorable attributes like
These include their high cost, space and infrastructure requirement, low cost, ease of synthesis and fabrication, high thermal and chemical
complicated and time-consuming sample preparation steps, etc. Further, stability, reliability, and modest efficiency with excellent prospects of
these instruments are not favorable for routine and point-of-care use. miniaturization. However, the SMO-based sensors work by the basic
Thus, there is a great demand for various kinds of gas sensors for both principles of the conductance change of the substrate due to the
predictive and preventive measures. Therefore, designing and devel­ adsorption/surface reactions/desorption of oxygen and target gas mol­
oping gas sensors with improved performance have emerged as a ecules. Therefore, SMO-based sensors also have several drawbacks, such
prominent area of research in recent times. as low target selectivity, long recovery/response times, weak response,
The requirement of the performance parameters of a gas sensor high operating temperature, etc., which need improvement.
varies depending on its use, the target and interfering gases, etc. How­ Various strategies have been explored to overcome such limitations
ever, a real-time high-performance gas sensor should have basic features and improve the performance of SMO-based sensing. For example,

* Corresponding author.
E-mail address: tapan.sau@iiit.ac.in (T.K. Sau).

https://doi.org/10.1016/j.cap.2022.10.003
Received 18 May 2022; Received in revised form 21 September 2022; Accepted 3 October 2022
Available online 8 October 2022
1567-1739/© 2022 Korean Physical Society. Published by Elsevier B.V. All rights reserved.
A. Biswas et al. Current Applied Physics 44 (2022) 131–143

researchers have explored complex metal oxides, composite systems, interactions [50–62]. Such information is essential for designing and
surface engineering, etc. [9,12,15,18–21] Since SMO-based sensing in­ developing efficient future gas sensors. We have employed DFT simu­
volves surface adsorptions/reactions, substrate’s surface states, addi­ lations to systematically investigate the atomic and molecular level in­
tives, etc. play critical roles in the sensor performance. It has led to two fluence of the Pd NP loading on the acetone sensing behavior of SnO2.
effective strategies for improving sensor performance: We have computed parameters like the adsorption and bandgap en­
nanomaterials-based gas sensors and SMO-surface modifications by ergies, charge distribution, polarizability change, electrophilicity index,
noble metals. Nanomaterials have several features like a large global hardness, etc., for various ligand gas (viz., acetone, ethanol, ox­
surface-to-volume ratio, higher surface-active sites, etc., offering ygen, moisture, and CO2) interactions on the pristine SnO2 as well as
favorable substrate surface-gas interactions and superior sensor perfor­ Pd-loaded SnO2 systems. The computational study reveals the molecular
mance [22–24]. On the other hand, SMO-surface modifications with aspects of the mechanisms of acetone sensor performance improvement
noble metal sensitizers like Pd, Pt, Au, Ag, etc., provide the catalytic and due to the Pd NP loading onto nanostructured SnO2 material.
spillover activities towards the substrate-gas interactions [14,24–28].
It is to be noted that despite the long history of using SMO-based 2. Experimental
sensors, mainly a trial-and-error approach has been used to select ma­
terials for various gas sensing. The exact mechanisms of noble metal 2.1. Chemicals and reagents
sensitizers in improving gas sensing performance are yet to be under­
stood. It has been proposed that noble metals activate the target gas and Stannous chloride (SnCl2.2H2O), glacial acetic acid (CH3COOH),
catalyze its oxidation on the SMO surface. For example, Pd can serve as concentrated aqueous ammonia, and polyethylene glycol-200 (PEG-
the initial adsorption site for oxygen adsorption via the formation of PdO 200) were purchased from LOBA Chemie. Potassium tetra­
[29–31]. PdO being a p-type SMO favor an efficient electron withdrawal chloropalladate (II) (K2PdCl4) and sodium borohydride (NaBH4) were
from an n-type SMO like SnO2 particles boosting the sensor response purchased from Sigma-Aldrich. Millipore water was used for all
[30–32]. Pd loading has also been reported to suppress the humidity or experiments.
water vapor poisoning effect on the SnO2 surface [17,33]. On the other
hand, we hardly understand how one can enhance an SMO substrate’s 2.2. Preparation of nanostructured SnO2
sensor response/selectivity towards a given target gas over the other
structurally and/or chemically similar gas(es). Since the SMO-based Among various synthesis strategies such as solvothermal, hydro­
sensing mechanism involves the oxidation of target gases by the thermal, spray pyrolysis, sol-gel, etc., employed for the preparation of
surface-adsorbed oxygen, the presence of more than one reducing gas SMOs, the sol-gel method is one of the most favored synthesis methods
poses a challenge to their SMO-based detection due to the interference. in terms of the ease and practicality of large-scale preparations. Here, for
For example, let us consider the acetone sensing by an SMO substrate in the synthesis of nanostructured SnO2 particles, we employed the facile
the presence of moisture (H2O), ethanol (CH3CH2OH), etc. Ethanol and sol-gel method. In a typical synthesis, a measured amount of solid
moisture are reducing in nature, like acetone. Furthermore, acetone and SnCl2.2H2O was added to 50 mL of hot Millipore water (at ~75 ◦ C) to
ethanol are structurally and chemically similar. Resolving the problem make a 10 mM solution. The mixture was stirred with the help of a
of target gas selectivity requires a deeper knowledge of the interactions magnetic stirrer for about 20 min. Then aq. ammonia was added drop­
of the sensor substrates with oxygen and other ligand gas molecules. wise until white precipitation of Sn(OH)2 formed. Slowly the tempera­
Therefore, the sensor performance improvement requires an under­ ture was brought down to ~60 ◦ C, and the reaction solution was left
standing of the microstructural sensor activation mechanisms, which, in undisturbed for 30 min to settle down the precipitate. The precipitate
turn, requires electronic and structural properties level information on was filtered and washed with Millipore water and diluted ammonium
the interactions of the sensing substrate and ligand gas(es). hydroxide. The residue was transferred into a beaker by adding 20 mL of
This article reports a combined theoretical and experimental study glacial acetic acid. Then 2 mL of polyethylene glycol 200 was added to
exploring the acetone sensing performance improvement of a nano­ the beaker and mixed with the help of a glass rod. The solution was then
structured SnO2 substrate by Pd nanoparticles (Pd NPs) loading. Acetone heated at 180 ◦ C with occasional stirring with the help of a glass rod to
sensors find uses in early diagnosis, health and environmental moni­ remove acetic acid completely. Next, the solution was left to cool down
toring, as well as industrial safety [34–41]. Acetone is commonly used in at room temperature. A light brown-colored sol appeared slowly. The sol
laboratories and chemical industries. However, it is a volatile organic transformed into a brown-yellow colored gel within 2–4 h. The gel was
liquid and toxic substance [42,43]. Exposure to acetone can cause transferred into a silica crucible and was calcined in a Muffle Furnace.
various health problems like eye/skin/respiratory tract irritation, First, it was ramped at 2 ◦ C min− 1 to 350 ◦ C and then heated for 2 h at
headache, fatigue, and nausea and can even damage our central nervous this temperature. Finally, the temperature was raised to 700 ◦ C and left
system at as low as 173 ppm acetone [43–45]. The presence of acetone in at this temperature for 30 min to obtain the desired product [63].
exhaled breath is an important chemical biomarker correlated with
several physiological and pathological processes in the human body. 2.3. Preparation of Pd NPs loaded SnO2
Breath-acetone is the end product of ketone body metabolism. It has
been linked to diabetes, physical activities and diet, acute kidney injury, For the formation and deposition of Pd NPs on the nanostructured
colorectal cancer, lung cancer, etc. [2,46,47] Recently, breath SnO2 in aqueous dispersion, we have employed room-temperature
VOC-based diagnosing and monitoring of human metabolism and health borohydride (NaBH4)-reduction of Pd(II) ions in K2PdCl4. This simple
conditions have gained momentum [2,46,48,49]. Breath-VOCs offer method allows rapid preparation of Pd NP-loaded SnO2 particles
noninvasive and inexpensive real-time sampling and analysis, avoiding avoiding high-temperature modifications of the NPs. Pd NPs formed and
inconveniences like interventional time-consuming sampling tech­ deposited directly on the nanostructured SnO2 particles. In solution,
niques, sample volume, dilution issues, etc. However, the detection of PdCl2−
4 ions can bind to the surface-hydroxyls of tin oxide particles and
breath-acetone among a plethora of volatile organic compounds (VOCs), are then reduced by the BH−4 ions forming Pd NPs on the surface of the
moisture, CO2, ethanol, etc., present in the exhaled breath poses several SnO2 particles. In a typical synthesis, 30 mg of SnO2 was first added to
challenges. 10 mL of Millipore water and was dispersed by magnetic stirring for 30
In recent times, several theoretical approaches, especially the density min. The required amount of 0.5 mM K2PdCl4 salt was added to the
functional theory (DFT), have emerged as powerful methods for dispersion. The solution color turned yellowish at this stage. Next, 10 mL
revealing the atomic and molecular level structural and electronic of freshly prepared ice-cold 3 mM NaBH4 solution was added to the
changes arising from the sensor substrates and the target/non-target gas reaction mixture. The reaction mixture was stirred for another 30 min.

132
A. Biswas et al. Current Applied Physics 44 (2022) 131–143

The solution color changed to grey, indicating the formation of Pd0 NPs. 3. Results and discussion
Then the reaction solution was left undisturbed for 2 h. The reaction
product was then filtered, and the residue was washed with Millipore 3.1. Phase structure and composition analysis
water. Finally, the residue was dried at room temperature. Ma et al.
reported that the Pd loading reached a saturation level at ~0.7 mol% 3.1.1. Powder XRD and XPS studies
precursor concentration for almost similar-sized SnO2 particles [64,65]. Fig. 1 shows typical powder XRD patterns of the synthesized samples
We present here data for two Pd concentrations such that one is slightly in the range of 2θ = 20◦ –80◦ . The experimental diffraction peaks can be
below, and the other is slightly above this saturation point, namely, 0.5 indexed to SnO2 (JCPDS 88–0287), Pd (JCPDS 46–1043), and PdO
mol% and 1.0 mol% Pd concentrations (henceforth l-Pd/SnO2 and (JCPDS 41–1107), suggesting the formation of tetragonal SnO2 phase
h-Pd/SnO2, respectively). and the presence of Pd0 and PdO [68–70]. The small line shifting in the
XRD pattern of the prepared samples with respect to the standard could
2.4. Characterizations be due to the micro-strains present in the nanostructured particles.
Particularly, the broader diffraction peaks with low intensity of Pd and
The synthesized particles were characterized by using powder X-ray PdO can be attributed to their small crystallite sizes. A few peak profiles
diffraction (XRD, Bruker, D8 Advance, Cu kα radiation, λ = 1.5406 Å), show asymmetric shapes. It could be due to a distribution of lattice
Raman scattering spectroscopy (EZRaman-N-785, Enwave Optronics, parameters arising from the defects and disorders in the lattice sites, a
Inc.), scanning electron microscopy (SEM, Carl Zeiss Ultra 55) equipped chemical variation, and/or a non-stoichiometric compound with a
with energy-dispersive X-ray spectroscopy (EDS), X-ray photoelectron particular phase width [71]. The mean crystallite sizes (D) were esti­
spectroscopy (XPS) (PHI 5000 VersaProbe III), UV–Visible spectropho­ mated to be ~9.9 nm and ~9.8 nm, respectively, for SnO2 and Pd/PdO,
tometry (PerkinElmer Lambda-35), and Photoluminescence (PL) spec­ by using the Debye-Scherrer formula, D = 0.94 λ/β cos θ (where D is the
troscopy (Test Right Portable Spectroscopy Solutions). The BET surface diameter of the particle, λ is the source X-ray wavelength (Cu Kα) =
area measurements were carried out using Quantachrome Novawin 0.15406 nm, β is the full width (in radian) at half maximum of the
(Nova Station A) instrument. diffraction peak at diffraction angle 2θ).
We know that XPS provides near-surface chemical compositions of
2.5. Sensor fabrication and gas sensing measurements the samples. XPS confirmed the presence of Sn and O elements in the
SnO2 sample and Sn, O, and Pd elements in the Pd nanoparticle-loaded
The construction of the sensor element was similar to that reported in SnO2 sample. Fig. 2 shows survey as well as high-resolution XPS of Sn, O,
ref. [66]. A sketch of the electric circuit diagram used for the sensors’ and Pd elements. Sn 3d XPS (2(b)) shows Sn 3d5/2 and Sn 3d3/2 peaks at
signal measurement is shown in SI Scheme S1. The sensor was designed 486.88 eV and 495.25 eV, respectively, with a spitting energy separation
on a ceramic tube with preinstalled gold electrodes on it and platinum
wires attached to it. The ceramic tube was coated with 20 mg of the
substrate materials under test and placed in the testing circuit. A
nichrome wire was used as a heating element that generates ~200 ◦ C at
5V/0.32A. The heating offers enhanced sensitivity and performance to
the sensor elements via inter-particle sintering and enhanced trans­
ducing and surface catalysis [67]. It is worth noting that 200 ◦ C falls at
the lower end of the operating temperature range used for most acetone
sensing experiments. The sensor element was placed in a 200 mL volume
box where the sensing gas was injected. To study the sensor recovery,
once the sensor response reached the stable state for a given period, the
test box was opened to re-expose the sensor element to fresh air when its
resistance returned to the base value. The sensor was tested with target
gas like acetone and interfering gas like ethanol in the 1–100 ppm
concentrations range. The relative humidity (RH) was 40–50% while
performing the experiments. Testing was also performed for up to 90%
moisture to check its interfering effects in the acetone sensing. All
measurements were repeated at least three times.

2.6. Computational methods

In the present work, DFT calculations were performed with the

Gaussian 09 software package. The molecules/clusters were optimized
using the B3LYP/GEN level of theory (Becke, 3-parameter, Lee-Yang-
Parr). The atoms H, C, and O were optimized using the B3LYP/
6–311++G(d,p) basis set, while the heavy atoms like Sn and Pd were
optimized B3LYP/LanL2DZ basis set. The core electrons were modeled
using the LanL2DZ electron core potential (ECP). All the structures were
optimized using the tight optimization criterion, which implies that the
RMS force criterion is set to 10− 5 Hartree/Bohr. The B3LYP functional
yields reasonable results for small clusters and is used to predict energy
gap values. Optimized geometries were used to calculate parameters
such as adsorption energies, bandgap energies, Mulliken charges, etc.
We have taken Pd6 nanocluster that has a symmetrical structure. How­ Fig. 1. Typical powder XRD patterns of the synthesized SnO2 and Pd nano­
ever, ligands were made to interact at different atom sites of Pd6 particle loaded SnO2 samples. Vertical lines show the reference XRD patterns of
nanocluster, and the results were average of all the positions. All the SnO2 (JCPDS 41–1445), Pd (JCPDS JCPDS 87–0638) and PdO
structures were generated using the GaussView program. (JCPDS 41–1107).

133
A. Biswas et al. Current Applied Physics 44 (2022) 131–143

Fig. 2. (a) XPS survey spectra of SnO2 and Pd nanoparticle-loaded SnO2 samples. High-resolution XPS of (b) Sn 3d, (c) Pd 3d, (d) Pd 3d deconvoluted XPS peaks, and
(e) O 1s (with deconvoluted peaks) of the Pd nanoparticle loaded SnO2 samples.

of 8.37 eV which is assigned to Sn4+ in SnO2 [69]. Fig. 2(c) shows the 3.1.2. UV–Vis and photoluminescence (PL) studies
high-resolution Pd 3d XPS with spin-orbit splitting peaks of Pd 3d5/2 and The UV–visible absorption and photoluminescence spectra collected
Pd 3d3/2 at 335.63 eV and 340.75 eV, respectively. A Pd 3d3/2 peak at from SnO2 and Pd nanoparticles loaded SnO2 samples are shown in
340.0–340.9 eV (3d5/2 peak at 334.8–335.6 eV) is attributed to Fig. 3. Asymmetric absorption occurs mainly in the UV region which can
(metallic) Pd0. The peaks shift to higher energy with an increase in the be deconvoluted into two bands (inset Fig. 3(a)) peaking at 295 nm
oxidation state of palladium [72]. For example, Pd2+ species is charac­ (4.20 eV) and 371 nm (3.34 eV). The latter band is close to the bulk SnO2
terized by the Pd 3d3/2 peak at 341.6–342.6 eV (3d5/2 peak at band gap energy of 3.6 eV (444.4 nm). The higher energy absorption
336.4–337.3 eV) whereas Pd4+ species is characterized by the Pd 3d3/2 could arise from the larger band gap of smaller SnO2 nanostructures,
peak at 343.2–344.2 eV (3d5/2 peak at 338.0–338.9 eV) [73]. From the since the quantum confinement effect in the semiconductor nanocrystals
multipeak fitting of the Pd 3d3/2 peak in Fig. 2(d), one can say that both shifts the band gap to higher energy when the particle size decreases.
(metallic) Pd0 and Pdδ+ (0 < δ ≤ 2) species are present in the Pd The presence of Pd dampens the absorbance and shifts the absorption
nanoparticle-loaded sample. It is in agreement with our other experi­ maximum to UV region further due to the Pd nanoparticles’ own ab­
mental and computational studies. Similarly, the O 1s XPS (Fig. 2(e)) sorption in the UV region. The photoluminescence emission spectra
signal can be deconvoluted into two peaks at 530.88 eV and 532.13 eV. (Fig. 3(b)) of SnO2 nanoparticles were collected at 365 nm and 470 nm
The 532.13 eV peak can be assigned to the surface adsorbed oxygen excitations. The former excitation energy employed is slightly higher
species (O−x ) while the 530.88 eV peak is attributed to the oxygen ions in than the bulk band gap energy and the latter is lower than it. The PL
the crystal lattice (Olattice). It shows the presence of surface-adsorbed exhibits maximal emissions at 395 nm (3.14 eV) and 494 nm (2.51 eV),
oxygen species (O−x ), in addition to lattice oxygen (Olattice) ions in the respectively. The former emission is very close to the bulk band gap of
samples [72,74]. SnO2 which can be associated with a band-to-band recombination pro­
cess. The lower energy emission can be attributed to the excitation of the

134
A. Biswas et al. Current Applied Physics 44 (2022) 131–143

Fig. 3. (a) UV–Vis absorption and (b) photoluminescence spectra (with 365 and 470 nm excitations) collected from SnO2 and Pd nanoparticle loaded SnO2 samples.

electrons from valence band to –OH group related defect energy states,
oxygen vacancies, etc. in the SnO2 nanoparticles. The lower PL intensity
in the presence of Pd nanoparticles could be due to the heavy-atom ef­
fect PL quenching. The optical absorption and photoluminescence
spectra were similar to that previously reported in the literature
[75–78].

3.1.3. Raman spectroscopy study

Raman spectroscopy is one of the most useful vibrational techniques
for the structural and morphological characterizations of bulk as well as
nanoscale metal oxides and various surface-adsorbed oxygen species
[30,71,79–81]. Metal-oxygen vibrations often occur below ~800 cm− 1,
which falls under the fingerprint region of the IR spectra. According to
the literature, single-crystal SnO2 (rutile structure with symmetry D144h )
shows Raman bands corresponding to four Raman-active modes: B1g
(rotation of the oxygen atoms around the c-axis) in the 100–200 cm− 1
region, Eg (translational mode involving the vibration of the oxygen in
the oxygen plane) at 476 cm− 1, A1g (symmetric O–Sn–O stretching) at
638 cm− 1, and B2g (asymmetric O–Sn–O stretching) at 782 cm− 1 [82].
IR-active modes A2u and Eu of the SnO2 single-crystal give rise to IR
bands peaking at 477 cm− 1 (A2u TO, Sn–O stretching), 618 cm− 1 (Eu TO,
Sn–O stretching), 705 cm− 1 (A2u LO, Sn–O stretching), and 770 cm− 1 (Eu
LO, Sn–O stretching).
Fig. 4 shows the recorded Raman spectra of the powder samples of
the pristine tin oxides and Pd NP-loaded tin oxides. Raman bands
peaking at 122, 462, 616, and 764 cm− 1 are in good agreement with the
four characteristic Raman-active modes, B1g, Eg, A1g, and B2g, respec­
tively observed for rutile SnO2 [71,80,82–84]. The redshift of the Raman
modes with respect to the single-crystal bands could be attributed to the
size-induced effects in nanostructured systems due to quantum
confinement, surface stress, and/or surface disorder in nanomaterials
[85]. A few peaks appeared due to the nanoscale dimensions of the
particles, their clustering, and/or surface defects. For example, the
appearance of the (Raman forbidden) IR-active A2u and Eu modes of
SnO2 can be attributed to the nanoscale dimensions of the SnO2 parti­
cles. Peaks at 470 (shoulder) and 712 cm− 1 could be assigned to the
IR-active A2u mode and at 608, and 770 cm− 1 to Eu mode of SnO2 [71,
82–84]. As reported by Drabeski et al., the band at 712 cm− 1 is Raman
forbidden but arises from the disorder activation of the A2u IR-active Fig. 4. Raman spectra (top: 100–800 cm− 1 and bottom: 800–1650 cm− 1 range)
mode [86]. It arises due to the interaction of the surface of tin oxide of the synthesized pristine and Pd NP loaded tin oxides powder samples. Spectra
with adsorbed oxygen. This peak intensity deceased upon Pd NP loading were offset for clarity. The vertical lines are given to guide the eyes for
onto SnO2. The Raman band in the 548 cm− 1 region could be assigned to some peaks.
the surface phonon modes arising from nanometer-sized SnO2 particles
[82,87]. Dieguez et al. attributed the appearance of similar bands to a arise due to the oxygen vacancy in the nanoscale SnO2 particles.
sum of contributions from different morphologies of the SnO2 particles The presence of various surface-adsorbed oxygen species was
[87]. The band peaking at 430 cm− 1 could arise from deeper internal recorded by the Raman study. Surface-adsorbed oxygen species are
oxygen vacancies (OV) [86]. According to Mathew et al., the symmetric essential components in the SMO-based chemoresistive sensors. The
Sn–O stretching Raman bands appear redshifted due to the weakening of surface adsorbed dioxygen species has a small dipole moment and high
the bond when some oxygen atoms are not available to bind to neigh­ polarizability. Therefore, Raman spectroscopy appears to be an excellent
boring Sn atoms [82]. Thus, the Raman peaks at 556 and 582 cm− 1 could tool for their characterizations [81]. Furthermore, Pd metal-oxygen
bonds arising from the surface adsorption of oxygen gas onto Pd NPs

135
A. Biswas et al. Current Applied Physics 44 (2022) 131–143

also show vibrations below the 800 cm− 1 range. As shown in Fig. 4, 3.2. Acetone sensing performance: Experimental study
many peaks in the 240 cm− 1 – 495 cm− 1 region could be assigned to
different Pd–O vibrations [88]. Surface superoxide species (O−2 ) adsor­ The sensor responses of the prepared SnO2 and Pd/SnO2 sensor
bed on the various sites are observed to have O–O stretching vibration substrates were experimentally measured by comparing the sensor sig­
frequencies in the 1010–1180 cm− 1 range [81,89,90]. The O–O vibra­ nals before (in the air) and after the exposure to the target gas (including
tion of peroxide adspecies (O2−2 ) appears in the range 640–970 cm
− 1
air). Exposure of an n-type SMO (e.g., SnO2) surface to the air results in
[89,90]. The bands above ~1525 cm− 1 have been assigned to molecular oxygen adsorption. Raman study has confirmed the presence of chem­
O2 species. On the other hand, the bands above ~1300 cm− 1 have been isorbed oxygen species on the substrate surface (Fig. 4). Oxygen
2 (0 < δ < 1)
attributed to dioxygen adspecies of intermediate forms, Oδ− adsorption causes the withdrawal of carrier electrons from the con­
[89,91]. Bands peaking at 1012, 1036, 1070, 1096, and 1164 cm− 1 can duction band of the SMO. It results in a decrease in the carrier concen­
be ascribed to the stretching frequencies of superoxo (O−2 )-like surface tration and thus an increase in electrical resistance of the substrate.
species [90,92,93]. Raman peaks at 652, 680, 726, 822, and 880 cm− 1 When the SMO sensor is exposed to a reducing target gas, the surface
correspond to the O–O stretching vibrational frequencies of different adsorbed oxygen species reacts with it and gets oxidized (e.g., acetone is
peroxo (O2−2 ) adspecies on Pd NP surfaces [92–94]. Higher Raman in­ oxidized to CO2 and H2O). The process releases the electrons to the
tensity of several surface oxygen species on the Pd/SnO2 substrate than conduction band of SMO, resulting in an increase in the conductance (or
the pristine SnO2 shows enhanced oxygen adsorptions after loading of a decrease in the resistance) of the SMO substrate. Since the resistance
Pd NPs, which can produce improved sensor response. value decreased upon exposure to the reducing gas, the sensor response,
S, was calculated using the relation: S (%) = (Ra − Rg) × 100/Rg, where
3.1.4. SEM-EDS and BET surface area measurement studies Ra is the sensor resistance in the air (original signal or base resistance)
Fig. 5 shows typical SEM images of the prepared SnO2 and Pd/SnO2 and Rg is the resistance in the presence of the target gas along with air.
samples. The roughness of the surface due to the nanoparticulate forms Fig. 7 shows the typical sensor response (S) of the pristine SnO2 and
of the materials can be seen in all samples. The nanoscale roughness of Pd/SnO2 sensors in the range of 1–100 ppm acetone vapor. It can be seen
the sensor materials results in a more accessible surface area and active that the sensor response increased with increasing acetone vapor con­
sites for the adsorption and diffusion of the target gas molecules, which centration. The increase in the sensor response with the increasing
augment the sensitivity and response rate of the sensors. The size of concentration of the target gas can be attributed to the enhanced surface
particles on the surface ranged from ~15 nm to ~25 nm. Pd NPs are reactions due to the increased surface coverage of the gas molecules,
hardly distinguishable from the SnO2 background. The lack of contrast gradually reaching a saturation. First, let us check the developed sen­
between Pd and SnO2 phases could be due to minor differences in the sor’s sensitivity to a gas (as defined by the ‘Rg/Ra’ ratio) and the signal-
atomic numbers of 46Pd and 50Sn [17]. However, to demonstrate the to-noise ratio (S/N) ratio. It is known that when the sensitivity to a gas
existence of Pd, elemental composition analysis and mapping using (Rg/Ra) is less than 0.8, the detection of the gas is possible. The ‘Rg/Ra’
electron energy-dispersive x-ray spectroscopy (EDS) were carried out. ratio value was slightly below 0.8 even at the concentration of 1 ppm
EDS mapping shed light on the dispersion of Pd NPs on the surface of acetone for the prepared Pd/SnO2 sensor substrate. Further, the devel­
SnO2. Fig. 6 shows a typical EDS elemental composition analysis and oped sensors showed a high S/N ratio. The S/N was calculated as fol­
mapping of Sn, O, and Pd on the sample surfaces. The elemental map­ lows: S/N = [(Average sensor response of the sample) – (Average sensor
ping showed a nearly homogeneous distribution of Pd NPs on the sample response of the blank)]/(Standard Deviation of the blank). Taking
surface. A scrutiny of the distribution of the elements in the sample pristine SnO2 as the blank, we calculated S/N ≈ 17 and ≈ 20 for 1 ppm
clearly shows that Sn and O are relatively evenly distributed across the acetone with l-Pd/SnO2 and h-Pd/SnO2 sensors substrates, respectively.
sample, while Pd shows discrete distribution. It suggests that Pd forms The effects of Pd NP loading become clear from comparing the sensor
particles on the SnO2 background. It is interesting to compare the atom% responses of the pristine SnO2 and Pd/SnO2 sensor substrates. Pd NP
compositions of the elements obtained from the EDS study (SI Table T1). loading onto SnO2 improved both the sensor response and the acetone
For the pristine SnO2 particles, the experimentally determined O:Sn selectivity. The pristine SnO2 sensor showed a marginally higher sensor
atom% ratio was 2.2, close to that of the stoichiometric SnO2 compound. response for ethanol than acetone. It is consistent with the slightly
However, the O:Sn atom% ratio of l-Pd/SnO2 ≫ h-Pd/SnO2 ≈ (pristine) higher calculated Ead value for ethanol binding to the pristine SnO2.
SnO2 substrates. The higher oxygen content in l-Pd/SnO2 material However, Pd NP loading (i.e., Pd/SnO2 substrate) significantly increased
samples could arise from the enhanced oxygen adsorption from the the acetone sensor response with respect to ethanol, unlike the pristine
ambiance. It is corroborated by the Raman spectroscopic studies. The SnO2 case (Fig. 7(a)). Pd/SnO2 sensor showed a more than two times
specific surface areas of the samples were determined by using the higher response than the pristine SnO2 sensor. However, a doubling of
Brunauer-Emmett-Teller (BET) method. Both SnO2 and Pd nanoparticles the Pd content produced only a minor increase in the sensor response,
loaded SnO2 samples showed similar BET surface area of 73.34 m2/g and which could be due to a diminished spillover effect due to a greater
69.06 m2/g, respectively. coverage of the SnO2 surface by Pd NPs. The slope of the plot of the
sensor response (S%) vs. gas concentration also gives a measure of the
sensor’s sensitivity. Pd/SnO2 sensors exhibited an almost linear and

Fig. 5. SEM images of the (a) SnO2 and (b) Pd/SnO2 samples.

136
A. Biswas et al. Current Applied Physics 44 (2022) 131–143

Fig. 6. A typical SEM-EDS elemental distribution on the surface of a synthesized Pd/SnO2 sample: (a) Sn (La1) in green, (b) O (Ka1) in red, and (c) Pd (La1) in yellow.
(For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

acetone, the threshold diabetic breath-acetone concentration, the

developed Pd-loaded sensors yielded a ~40% sensor response (with a
<2 s response time and ~8 s recovery time). Fig. 7(b) shows the plot of
the Pd/SnO2 sensor response (S%) vs. log C (where C is the target gas
concentration in ppm) in the 1–100 ppm acetone concentration range.
The plots show the distinct improvement of the sensor response due to
Pd NP loading onto SnO2. These characteristics of the sensors are
promising from the application point of view.
Improving the selectivity of gas sensing is one of the most important
goals of chemoresistive SMO-based sensing. Ethanol, water vapor, etc.,
being reducing in nature (like acetone) interfere and throw a challenge
in the acetone determination. Experimental measurements showed that
the Pd/SnO2 sensor had a distinctly higher sensor response for acetone
than ethanol, with a selectivity factor, k = Sacetone/Sethanol > 2.14. A
clear distinction between acetone and ethanol sensing can be discerned
from their plots of the sensor response (S%) vs. log C (where C is the
target gas concentration in ppm). Unlike the pristine SnO2 sensor, the
plot for acetone sensing on Pd/SnO2 yields distinct straight lines with
widely different slopes from ethanol sensing (Fig. 7(b) and SI Fig. S1).
Thus, one can see that Pd NP loading onto SnO2 improves the acetone
selectivity over ethanol. The calibration graphs involving the sensor
response vs. corresponding gas concentrations can be used to detect
acetone selectively in the presence of ethanol. Water vapor is another
common interfering agent in acetone sensing since it is one of the most
common constituents of the atmosphere and exhaled breath. While the
relative humidity of the ambiance varies in the range of 20–60%, it is
often ~90% in exhaled breath [47]. It is, therefore, essential to inves­
tigate its interfering effects on acetone sensing. Experimentally, there
was practically no change in the Pd/SnO2 sensor’s conductivity in the
presence of moisture (up to 90% relative humidity) (SI Fig. S2). It was
consistent with the DFT predictions that showed that Pd NC loading onto
SnO2 could suppress the interfering effects of H2O.
A clear improvement in the sensor performance (like the response,
recovery, etc.) due to Pd NP loading can be seen in Fig. 8. It shows
typical plots of the time-dependent signal changes of Pd/SnO2 vs. SnO2
sensors exposed alternately to the air and target gas acetone or ethanol.
The ‘pristine SnO2’ (and Pd/SnO2 for ethanol) yielded irregular-shaped
response and recovery curves. On the contrary, the acetone gas test
signal in the Pd/SnO2 sensor shows an almost instantaneous rise upon
exposure. The signal stays stable under constant concentration condi­
Fig. 7. Typical (a) sensor responses (S%) of the SnO2 and Pd/SnO2 sensors in tions and then drops abruptly and returns to the baseline level when the
the range of 1–100 ppm of acetone or ethanol gas and (b) S% vs. log (con­ sensor is exposed to the air. The acetone response–recovery repeatability
centration of acetone/ppm) and possible linear fits.
cycles of the Pd/SnO2 sensors also corroborate the above findings (SI
Fig. S3). One can see consistent output for a given exposure, and there
sharp increase in the sensor response up to 10 ppm acetone, followed by was hardly any sign of the decay of the response amplitude or aging of
a slower growth (Fig. 7(a)). In this context, it should be noted that the the performance over the repeated test cycles on the Pd/SnO2 sensor.
acetone concentration in the exhaled breath varies in the range of The above observations are consistent with the theoretically calculated
~0.1–10 ppm [95], and a diabetic patient shows more than 1.8 ppm high Ead values and slow recovery on the pristine SnO2 systems, which
acetone in the exhaled breath. Therefore, ~1–10 ppm appears to be an improve substantially upon Pd NC loading. Fig. 9 captures the
important concentration range in breath-acetone sensing. At 1.8 ppm improvement in the acetone response and recovery speeds upon Pd NP

137
A. Biswas et al. Current Applied Physics 44 (2022) 131–143

systems. Since oxygen interaction plays a vital role in SMO-based

sensing, the interaction of oxygen molecules with the pristine and Pd-
loaded sensor substrates has been studied. The effects of the other
gases like water vapor (H2O), carbon dioxide (CO2), and ethanol vapor
(CH3CH2OH) have also been studied computationally. H2O vapor and
CO2 are two common constituents of the atmosphere and exhaled
breath. CH3CH2OH vapor is also often present in the exhaled breath and
is reducing in nature like acetone and H2O. Therefore, studies of the
interactions of these gases are essential to understanding and improving
the selectivity of the SMO-based chemoresistive sensors.
It is well-accepted that the noble metals cause electronic and
chemical sensitizations by providing interfacial electronic redistribu­
tion, more active sites for oxygen and target gas adsorptions, and the
spillover effect to the sensor substrate. The spillover zone lies in the
vicinity of noble metal nanoparticles, which are also the preferred sites
for co-adsorption of reducing gas and oxygen. It suggests that the
‘effective working motifs’ involved in the sensing process of the nano­
structured substrates are tiny nanoclusters made of the support mate­
Fig. 8. Dynamic response–recovery performances of the SnO2 and h-Pd/SnO2 rial/noble metal/gas adsorbates, especially from the electronic and
sensors in the 1–100 ppm range acetone/ethanol concentrations (response
structural properties points of view. Therefore, both the support and the
curves correspond to increasing concentrations from left to right as shown on
adsorbed gas can be considered as the ligands of the noble metal
the top).
nanoclusters [96]. For our DFT calculations, we have taken Pd6 nano­
cluster (Pd NC) as a model of the Pd NP, Pd6–SnO2 nanocluster as the
model of the Pd NP-loaded SnO2 systems, and Pd6–SnO2-L nanoclusters
as the ‘working motifs’ of the sensor substrates (where L = ligand gas).
Pd6 nanocluster has a symmetrical octahedral structure. The DFT study
was employed to compare the electronic and structural properties of the
ligand (L) gases like acetone (Act), ethanol (Eth), moisture (H2O), oxy­
gen (O2), and carbon dioxide (CO2) with the substrates (S) like the
pristine SnO2, Pd6, and Pd6–SnO2.

3.3.1. Adsorption energy and structure

In SMO-based gas sensing, the interaction and adsorption of the gas
molecules with the sensor materials are the critical steps in the sensing
process. Therefore, we have calculated and compared the adsorption
energies of the ligand gas molecules with SnO2, Pd6, and SnO2–Pd6
systems. The adsorption energy (Ead) was defined as follows: Ead = [E(S-
L) – E(S) – E(L)], where E(i) represents the energy of species, i, and S-L
stands for the complex of the substrate S and the ligand L. Table 1 lists
the Ead values. For various Pd atom locations, average Ead values have
been considered. According to the definition, negative adsorption en­
ergy (Ead) indicates an exothermic and favorable adsorption process.
It is known that the adsorbed oxygen causes electron withdrawal
Fig. 9. Typical sensor response and recovery times (measured in triplicate) at from the conduction band of the substrate, resulting in a decrease in the
various acetone concentrations on the SnO2 and h-Pd/SnO2 sensors.
conductance (or an increase in the resistance) of an n-type SMO and the
reaction of the adsorbed oxygen with the reducing target gases releases
loading onto SnO2. The response and recovery times were calculated as the electrons back. Thus, oxygen adsorption plays a crucial role in the
the time required for the signal to undergo a change of 90% in adsorp­ SMO-based chemoresistive sensor response. From Table 1, one can see
tion and desorption of the target gas, respectively. Unlike the pristine that the presence of Pd nanoclusters makes O2 adsorption highly
SnO2, the ranges of acetone response and recovery times were very favorable, which is evident from their large negative Ead values
narrow, mostly <10 s in the Pd NPs loaded SnO2 substrates. For compared to the pristine SnO2 and the distorted structures of O2-
example, experimental response times as short as <2 s and ~4 s and adsorbed complexes (SI Fig. S4). It is further inferred from the analysis
recovery times ~7 s and ~4 s were determined respectively for 2 ppm of the O–O and Pd–O bond distances. It was found (Table 2) that the
and 100 ppm acetones on the h-Pd/SnO2 sensor substrates. Ultrafast Pd–O bond lengths were in the 2.023 Å – 2.044 Å range, like the Pd–O
response and recovery are beneficial for both the reusability of the chemical bonds in palladium oxide solids, implying the chemisorption of
sensor and the reliable determination of acetone in real-time the O2 molecule on the substrate [97]. The O–O bond in the adsorbed O2
applications. species elongates significantly in the presence of SnO2–Pd6 (compared to
1.205 Å in free O2). The O–O bond distance increases in the following
3.3. DFT study order: SnO2–O2 < Pd6–O2 < SnO2–Pd6–O2. It is to be noted that the O–O
bond length falls in the typical superoxo range [98]. The bond elonga­
DFT is a powerful method in studying various interactions and tion suggests that the SnO2–Pd6 substrate activates the adsorbed O2
atomic and molecular levels electronic and structural changes. We have more than the pristine SnO2 [99]. The DFT calculations and structural
employed DFT simulations to explore how the Pd NP loading onto the information were in agreement with the Raman spectroscopic and
SnO2 system affects the acetone sensing process. DFT calculations were experimental sensing results. It suggests that the Pd NC loading will
used to analyze the electronic and structural changes arising from the Pd produce a better sensor response than the pristine SnO2.
nanocluster loading as well as various ligand gas binding onto the SnO2 Now, let us consider the adsorption energies of the reducing gases,

138
A. Biswas et al. Current Applied Physics 44 (2022) 131–143

Table 1
Calculated adsorption energy (Ead), HOMO-LUMO energies, HOMO-LUMO energy gap (Eg), recovery time (tr), electrophilicity index (ω), and polarizability change
(Δα) of various systems.
System Ead (eV) EHOMO (eV) ELUMO (eV) Eg (eV) ω (eV) Δα tra (s)

CH3COCH3 (Act) – − 7.053 − 0.783 6.270 2.448 – –

C2H5OH (Eth) – − 7.636 − 0.339 7.297 2.179 – –
O2 – − 5.345 − 7.271 1.926 20.660 - -
H2O – − 8.788 − 0.634 8.154 2.722 - -
CO2 – − 10.503 − 0.537 9.966 3.057 - -
Pd6 – − 5.100 − 3.811 1.289 15.401 - -
SnO2 – − 8.539 − 3.790 4.749 8.002 - -
SnO2-Act − 0.846 − 7.228 − 2.661 4.567 5.353 5.161 17411
SnO2-Eth − 0.908 − 7.367 − 2.513 4.854 5.028 3.122 192518
Pd6-Act − 0.368 − 4.708 − 3.442 1.266 13.117 18.934 -
Pd6-Eth − 0.333 − 4.748 − 3.478 1.269 13.331 8.837 -
SnO2–Pd6 − 3.007 − 5.514 − 4.233 1.282 18.527 30.300 -
SnO2–Pd6-Act − 0.542 − 5.084 − 3.756 1.328 14.711 17.474 0.133
SnO2–Pd6-Eth − 0.544 − 5.209 − 3.863 1.346 15.286 10.985 0.144
Pd6–O2 − 1.991 − 5.393 − 4.422 0.971 24.803 11.652 -
SnO2–O2 − 0.216 − 8.088 − 5.936 2.152 22.848 13.768 -
SnO2–Pd6–O2 − 1.567 − 5.886 − 4.700 1.186 23.622 14.286 -
Pd6–H2O − 0.275 − 4.752 − 2.466 2.286 5.698 7.516 -
SnO2–H2O − 0.758 − 7.515 − 2.713 4.802 5.446 3.081 -
SnO2–Pd6–H2O − 0.557 − 5.298 − 3.953 1.345 15.907 7.048 -
Pd6–CO2 − 0.071 − 5.030 − 3.753 1.278 15.090 5.634 -
SnO2–CO2 − 0.161 − 8.285 − 3.539 4.746 7.364 2.925 -
SnO2–Pd6–CO2 − 0.139 − 5.447 − 4.161 1.286 17.946 5.581 -
a
Calculated taking attempt frequency, ν0 = 1010 s− 1
and kT = 0.0258 eV (T = 300 K).

Table 2 where ν0 stands for the attempt frequency of the gas molecule, k is for
Pd–O and O–O bonds distances (in Å) for the adsorbed O2 species. the Boltzmann’s constant, and T is the Kelvin temperature of the system.
If the attempt frequencies of the ligand molecules are assumed to be in
System D (O–O) (Å) D (Pd—OL) (Å)
(in O2) (OL in O2) the same order of magnitude, the equation shows that a longer recovery
time arises from a higher (negative) Ead value. One can find that the
O2 1.205 –
SnO2–O2 1.219 –
presence of Pd NC considerably reduces the Ead values of acetone- and
Pd6–O2 1.319 2.038 ethanol-bound systems (Table 1) compared to the pristine SnO2 case.
SnO2–Pd6–O2 1.326 2.044 Therefore, the Pd NC loading onto SnO2 improves the recovery speed
compared to the pristine SnO2. As mentioned above, the response/re­
covery rates are also related to the reactivity properties of the target
namely, moisture (H2O), acetone, and ethanol. The adsorption energies
gases on the substrates. A higher reactivity of the ligand gas favors a
decrease in the following order: SnO2-Eth > SnO2-Act > SnO2–H2O >
faster sensor response/recovery, which was mentioned above in the
SnO2–Pd6–H2O ≈ SnO2–Pd6-Eth ≈ SnO2–Pd6-Act. It shows that the
context of the adsorbed O2 species and will be further discussed later.
adsorption is significantly more exothermic on the pristine SnO2 than
that in the presence of Pd NC. In general, higher exothermic adsorption
3.3.2. HOMO-LUMO gap (Eg) and polarizability change
means better selectivity and rapid response [50]. At the same time, the
The frontier molecular orbitals (FMOs), namely, the highest occu­
stronger adsorption also suggests a longer recovery time on the pristine
pied molecular orbitals (HOMOs) and the lowest unoccupied molecular
sensor substrate. The low Ead values or a weaker binding of the ligands in
orbitals (LUMOs) and their energy gaps (Eg = ELUMO − EHOMO) are very
the presence of Pd NC favor the faster recovery of the sensor [100]. The
useful for predicting and interpreting the molecular level interactions,
lowering of H2O Ead values in the presence of the Pd NC implies its roles
qualitative comparison of the chemical stabilities and reactivities of
in decreasing the H2O sensitivity of the pristine SnO2. It is consistent
clusters and molecules, and the electrical conductivity of the clusters
with the earlier report claiming that Pd loading suppressed the water
[55,56]. Table 1 lists the HOMO-LUMO energies of relevant systems. In
vapor poisoning of the SnO2 surface [33]. Similarly, from a comparison
SMO-based gas sensing, the sensor response is associated with the
of the corresponding Ead values, one can find that the presence of the Pd
electrical conductivity change of the sensor substrate resulting from the
NC also reduces the interference effects of CO2. However, SnO2–Pd6-Act,
ligand-binding events. A smaller energy gap (Eg) of the clusters offers an
SnO2–Pd6-Eth, and SnO2–Pd6–H2O have similar Ead values. Therefore,
easier transfer of electrons from the valence band to the conduction
from the adsorption energy point of view, the SnO2–Pd6 substrate shows
band. We know that the intrinsic conductivity (σ) of a semiconductor
almost similar selectivity towards these reducing gases. Later, we shall
system is mainly controlled by Eg/kT ratio and related to Eg as follows
see that the Pd loading influences various electronic properties of the
[101]:
systems, which favor acetone selectivity.
( )
We have seen above that the adsorption/desorption process and the − Eg
σ ∝exp
ease of target gas reactions play important roles in deciding the times or 2kT
speeds of the sensor response and recovery. Faster response and recov­
ery of the sensor (i.e., shorter response/recovery times) are desired for where Eg, k, and T are the semiconductor HOMO-LUMO band gap,
practical gas sensors. The recovery time, tr, for gas desorption from the Boltzmann’s constant, and thermodynamic temperature, respectively. It
sensor materials has been related to the adsorption energy, Ead, as fol­ shows that a smaller Eg value gives rise to higher electrical conductivity,
lows [56]: σ . In other words, the σ and hence the sensor’s sensitivity will increase if
( ) the Eg value decreases upon the Pd NC and the ligand gas binding onto
1 − Ead SnO2 [102].
tr = .exp
ν0 kT First, let us consider the effects of ligand binding onto the pristine

139
A. Biswas et al. Current Applied Physics 44 (2022) 131–143

SnO2. We can find that only O2 adsorption produces a significant energy of the acetone and ethanol ligands bound systems and lowers the
decrease in the Eg values, whereas acetone adsorption onto pristine SnO2 LUMO energy of the corresponding O2-bound systems (Table 1 and
results in a slight decrease in Eg value. However, the Eg values either Fig. S5). The SnO2–O2 system can behave as an oxidant to SnO2–Pd6-Act
slightly increased or barely changed in the cases of other ligand bind­ since the SnO2–O2 system has a lower LUMO energy than SnO2–Pd6-Act
ings, although these ligands showed favorable exothermic adsorptions. HOMO energy. A clear measure of the electron donation-acceptance
It implies that the pristine SnO2 is not favorably sensitive to the detec­ tendency of the systems can be obtained from the comparison of the
tion of these gas molecules (except the acetone adsorption), and the global electrophilicity index, ω, (defined as ω = μ2/2η, where μ is the
conductivity change is the result of the O2 adsorption and desorption electronic chemical potential and η is the chemical hardness as described
processes on the pristine SnO2. above) values [103,104]. A stronger electron-acceptor tendency is
On the other hand, Pd NC loading caused a substantial drop in the Eg characterized by a larger electrophilicity index (ω) value.
values compared to the pristine SnO2 (SI Fig. S5). Comparing the Eg In these sensor systems, the chemisorbed O2 species (S–O2) are the
values of the pristine SnO2 and SnO2–Pd6 systems shows that the Pd electron-acceptors, and the other S-L species are electron donors. A
loading resulted in approx. 70% decrease in the Eg values. Therefore, Pd comparison shows that the electron-acceptor systems, S–O2, have high
NC loading can produce a significant increase in the electrical conduc­ electrophilicity index values than the electron donors, S-L. The higher
tivity and hence the sensor response of the sensor substrate. It is electron affinity of oxygen plays a crucial role compared to acetone,
important to note that the Eg values increased upon Act and Eth binding ethanol, CO2, and water molecules. When the SMO is exposed to a
to SnO2–Pd6, and the increase was more in the case of ethanol. It sug­ reducing target gas (viz., ethanol, acetone), the surface-adsorbed oxygen
gests that Pd loading will produce a higher sensor response for acetone species will oxidize the reducing gas or be replaced by competitive
than ethanol, consistent with the experimental observations. adsorption, which will release the captured electrons to the conduction
The DFT calculated dispersion of isodensities of frontier orbitals of band of the SMO. It leads to an increase in the conductance or decrease
all species (SI Fig. S6) shows two different patterns of electronic distri­ in the resistance of the sensor substrate. It can be noted that Eth system
butions. The HOMO and LUMO were mainly distributed over the sub­ has a lower ω and the pristine SnO2-Eth system has a higher Ead energy
strate, SnO2–Pd6, when there was a weak interaction and small change and a lower ω than the pristine SnO2-Act system (Table 1). It is consistent
in the HOMO-LUMO energy gap. On the other hand, a uniform distri­ with our experimental observations that ethanol at lower concentrations
bution over the entire complex and notable change in the orbital density showed a marginally higher sensor response than acetone on the pristine
were observed in the cases where there were stronger interactions like SnO2 (see Fig. 7). The effects of the Pd NC loading can be assessed by
O2. Therefore, the Pd NC loading onto SnO2 substrate provides sites for comparing the electrophilicity index values of SnO2 and SnO2–Pd6 sys­
co-adsorption of target gas and O2, which is favorable for their reactions. tems (Table 1). One can see that the electron-acceptor tendency in­
The HOMO− LUMO energies have also been correlated with the creases upon Pd NC loading, which is mainly determined by the
chemical stability and reactivity of various chemical systems [103,104]. electrophilicity of the Pd6 NC. O2 binding further increases the elec­
The knowledge of the chemical reactivities of the ligand-bound clusters trophilicity while the other ligands decrease it. The decrease in the
is valuable in the context of the response/recovery times, operating electrophilicity index due to the binding of the ligands onto SnO2–Pd6
temperature, etc. Parameters like the chemical hardness (η) and Mul­ followed the order: Act > Eth > H2O > CO2. It is to be noted that the
liken electronegativity (χ), which have been defined as η = (ELUMO – presence of the Pd NC caused a reversal in the electrophilicity of the
EHOMO)/2 = Eg/2 and χ = − (ELUMO + EHOMO)/2 (= − μ, electronic ethanol and acetone systems with respect to their pristine systems. It
chemical potential), are useful reactivity descriptor parameters [103, suggests that the electron donation tendency of acetone increases due to
104]. One can see that while the Mulliken electronegativity represents the Pd loading onto SnO2. Thus, the Pd loading onto the SnO2 substrate
the average valence electron energy, the chemical hardness (η) is a results in acetone selectivity over ethanol and water vapor.
measure of the breadth of the HOMO− LUMO energy difference. In other The changes in the total polarizability of the ligand complexes were
words, the chemical hardness (η) represents the resistance to the elec­ computed. The polarizability, α, of a species was calculated as follows: α
tron distribution change and hence the chemical stability and reactivity = (αXX + αYY + αZZ)/3, where αXX, αYY, and αZZ are polarizability com­
of a system. Thus, the clusters with smaller Eg values correspond to ponents along the X, Y, and Z axes, respectively. The change in total
lower chemical hardness (i.e., conversely, higher global softness). In polarizability, Δα, was defined as: Δα = [α(S-L) – α(S) – α(L)], where α(i)
other words, smaller Eg value clusters are softer where electron density stands for the polarizability (α) of species, i, and S = SnO2, Pd6 or
changes easily, showing higher chemical reactivity and vice versa. SnO2–Pd6, and L = Act, Eth, O2, and H2O). Δα values are given in
It can be seen that SnO2 with a high Eg value is chemically relatively Table 1. It can be seen that the Pd NC loading affected the polarizability
stable. In fact, the higher chemical stability of the sensor building block of the substrate clusters to different extents depending on the system.
is essential for designing robust SMO-based sensors. On the other hand, The polarizability change was significant in the cases of O2 adsorption
Pd6 NC, as expected, is highly reactive or chemically less stable. How­ onto the substrate and acetone or ethanol binding. It should be noted
ever, as discussed above, in the context of conductivity change, the that the Pd NC loading caused a considerable difference in the polariz­
comparison of the SnO2 and SnO2-L systems showed hardly any change ability of the acetone and ethanol systems. Since softer molecular sys­
in the Eg values upon ligand binding to SnO2 (except the SnO2–O2 and tems where electron density changes more easily could show higher
SnO2-Act cases). A comparison of the Eg values of SnO2–Pd6-L and SnO2- chemical reactivity, the higher polarizability could be one of the factors
L systems showed that the SnO2–Pd6-L systems have relatively low Eg of the higher acetone selectivity of Pd NC-loaded SnO2 substrates over
values. Based on the chemical hardness (η) values, it can also be ethanol compared to the pristine SnO2 substrates. Selectivity is a crucial
concluded that the presence of the Pd NC loading is favorable for the factor in sensing, especially in the presence of gases with similar phys­
higher chemical reactivities of the corresponding systems, which can icochemical properties to the target gas.
significantly improve the sensor performance like the recovery time,
operation temperature, etc. It also implies that Pd NC loading onto SnO2 3.3.3. Mulliken charges
substrate improves the acetone sensing performance compared to the Let us recall that the conductivity change arising from the bandgap
pristine SnO2. (Eg) change due to the ligand-binding was not favorable for sensing in
In these sensing reactions, reducing gases are the electron-donors, the case of most ligands. The electronic charge transfer between the
while the chemisorbed O2 species are the electron-acceptors. There­ substrate and the adsorbed molecules plays a critical role in determining
fore, the energy gap between the HOMO of the ligand adsorbed system the conductivity and the performance of the chemoresistive sensors
and the LUMO of the O2-adsorbed system can shed light on this charge [105]. We have employed the Mulliken population analysis to evaluate
donation process. It can be seen that the Pd NC loading raises the HOMO the direction and magnitude of the electronic charge transfer due to the

140
A. Biswas et al. Current Applied Physics 44 (2022) 131–143

various ligand binding. It can be seen that ligand O2 acts as the charge Table 3
acceptor, and the other ligands act as the charge donors (Table 3). The Calculated Mulliken charge transfer from the substrate (S) to ligand (L).
charge withdrawal due to the oxygen adsorption is vital in the System (S-L) Charge (e) transferred to Ligand (L)
SMO-based chemoresistive gas sensors. The Mulliken charge analysis
SnO2–O2 − 0.021
showed that O2 adsorption onto the pristine SnO2 caused only a small SnO2-Act 0.198
charge separation. The presence of Pd NC caused large charge separa­ SnO2-Eth 0.141
tions between the substrate and the adsorbed O2. Since the reaction of SnO2–H2O 0.097
the adsorbed oxygen species with the target gas releases back the SnO2–CO2 0.049
Pd6–O2 − 0.283
negative charge to the substrate, the higher charge separation gives rise Pd6–H2O 0.073
to a greater change in the electrical conductivity and hence a higher Pd6–CO2 0.035
sensor response. Pd NC loading enhanced the charge separation between Pd6-Act 0.125
the donor ligands and the loaded substrates. A scrutiny of the charge Pd6-Eth 0.108
SnO2–Pd6 0.443
distributions of the SnO2–Pd6-L systems showed that Pd loading caused
SnO2–Pd6–O2 − 0.284
higher charge separations in acetone and ethanol systems compared to SnO2–Pd6-Act 0.234
the other donor ligands, namely, H2O and CO2 systems. Thus, one can SnO2–Pd6-Eth 0.231
conclude that the sensor performance improvement involves several SnO2–Pd6–H2O 0.135
factors like the Ead, Eg, reactivity, charge separation, etc. SnO2–Pd6–CO2 0.061

4. Conclusions interests or personal relationships that could have appeared to influence

the work reported in this paper.
A combined theoretical and experimental study explored the
improvement in the acetone sensing performance of the SMO, SnO2,
Acknowledgments
upon Pd NP loading and molecular level factors that could influence the
sensing performance. For experimental studies, nanostructured SnO2
The authors acknowledge the financial support of DST and CSIR,
and Pd loaded SnO2 sensor substrates have been prepared, and their
Govt. of India, and IIIT-Hyderabad. We gratefully acknowledge the help
acetone sensing performances have been investigated. The sensor sub­
of Dr. A. Semparithi, Mr. Yash Chaurasia, Mr. Supriyo Bose, and Ms.
strates have been prepared from commonly available chemical reagents
ShriVidhatri in computational studies. We would also like to thank Mr.
using simple chemical synthesis methods and have been characterized
Mahesh K Uppada, Dr. Saikat Sen, and Dr. Srinivas Oruganti (Dr. Red­
using various techniques like XRD, XPS, Photoluminescence, Raman
dy’s Institute of Life Sciences, Hyderabad) for their supports in UV–Vis
spectroscopy, BET surface area measurement, etc. Pd NPs loading onto
and Fluorescence Spectroscopic measurements; Dr. Joydip Joardar
nanostructured SnO2 significantly improved the acetone sensor perfor­
(Materials Characterization and Testing, International Advanced
mance, such as the sensor response, response/recovery speeds, sensi­
Research Center for Powder Metallurgy & New Materials (ARCI),
tivity, and signal-to-noise ratio. Pd loading also suppressed the
Hyderabad) for helping in the XRD characterization; Mr. Avinash Cha­
interference effects of ethanol, moisture, etc. DFT computations have
van (Department of Metallurgical and Materials Engineering, IIT-
been performed to understand the electronic and structural properties
Kharagpur) for his help in BET Surface Area measurements, and XPS
level effects of the Pd NC loading onto SnO2 on the acetone sensing
Lab Central Research Facility (IIT-Kharagpur) for XPS characterizations.
process. Adsorption energy (Ead), HOMO-LUMO gap energy (Eg), global
hardness (η), electrophilicity index, Mulliken charge transfer, polariz­
Appendix A. Supplementary data
ability change, etc., were calculated and compared for some model
systems. The model systems involved ligand gases like O2, CH3COCH3,
Supplementary data to this article can be found online at https://doi.
C2H5OH, H2O, and CO2, and their interactions with pristine SnO2 as well
org/10.1016/j.cap.2022.10.003.
as Pd nanocluster loaded SnO2. The calculated parameters have been
correlated with various sensor performance parameters like the sensor
response, speeds of response/recovery, selectivity, etc. The results of the References
DFT calculations showed that the sensor performance improvement
[1] J. Dai, O. Ogbeide, N. Macadam, Q. Sun, W. Yu, Y. Li, B.-L. Su, T. Hasan,
could be linked to multiple factors like the Pd nanocluster loading, X. Huang, W. Huang, Printed gas sensors, Chem. Soc. Rev. 49 (6) (2020)
electronic redistribution, O2 adsorption, electron donor-acceptor prop­ 1756–1789.
erty change, polarizability change, and so on. DFT calculations provide [2] Y.Y. Broza, X. Zhou, M. Yuan, D. Qu, Y. Zheng, R. Vishinkin, M. Khatib, W. Wu,
H. Haick, Disease detection with molecular biomarkers: from chemistry of body
important insights into the origins of acetone sensing performance fluids to nature-inspired chemical sensors, Chem. Rev. 119 (22) (2019)
enhancement due to Pd NPs loading onto SnO2 materials and underscore 11761–11817.
the importance of the electronic properties level knowledge of such [3] W. Ge, S. Jiao, Z. Chang, X. He, Y. Li, Ultrafast response and high selectivity
toward acetone vapor using hierarchical structured TiO2 nanosheets, ACS Appl.
nanostructured sensor materials in order to design and develop cheaper Mater. Interfaces 12 (11) (2020) 13200–13207.
and miniaturized gas sensors with improved performance. [4] Z. El Khalidi, B. Hartiti, M. Siadat, E. Comini, H.M.M.M. Arachchige, S. Fadili,
P. Thevenin, Acetone sensor based on Ni doped ZnO nanostructues: growth and
sensing capability, J. Mater. Sci. Mater. Electron. 30 (8) (2019) 7681–7690.
Author contributions [5] G. Li, Z. Cheng, Q. Xiang, L. Yan, X. Wang, J. Xu, Bimetal PdAu decorated SnO2
nanosheets based gas sensor with temperature-dependent dual selectivity for
A.B., O.P., A.T., & D.M.: Methodology, Investigation, Validation, detecting formaldehyde and acetone, Sens. Actuators, B 283 (2019) 590–601.
[6] X. Liu, K. Zhao, X. Sun, X. Duan, C. Zhang, X. Xu, Electrochemical sensor to
Formal analysis, Manuscript Writing, Review. T.K.S.: Conceptualization,
environmental pollutant of acetone based on Pd-loaded on mesoporous In2O3
Methodology, Resources, Manuscript Writing, Review & Editing, Su­ architecture, Sens. Actuators, B 290 (2019) 217–225.
pervision, Project administration, Funding acquisition. S.R.C.: Concep­ [7] S. Liu, S. Gao, Z. Wang, T. Fei, T. Zhang, Oxygen vacancy modulation of
tualization, Methodology, Resources, Review & Editing, Supervision, commercial SnO2 by an organometallic chemistry-assisted strategy for boosting
acetone sensing performances, Sens. Actuators, B 290 (2019) 493–502.
Funding acquisition. [8] R.K. Sonker, B.C. Yadav, V. Gupta, M. Tomar, Fabrication and characterization of
ZnO-TiO2-PANI (ZTP) micro/nanoballs for the detection of flammable and toxic
Declaration of competing interest gases, J. Hazard Mater. 370 (2019) 126–137.
[9] Y. Zhang, Q. Yang, X. Yang, Y. Deng, One-step synthesis of in-situ N-doped
ordered mesoporous titania for enhanced gas sensing performance, Microporous
The authors declare that they have no known competing financial Mesoporous Mater. 270 (2018) 75–81.

141
A. Biswas et al. Current Applied Physics 44 (2022) 131–143

[10] X. Gao, T. Zhang, An overview: facet-dependent metal oxide semiconductor gas [39] T.D.C. Minh, D.R. Blake, P.R. Galassetti, The clinical potential of exhaled breath
sensors, Sens. Actuators, B 277 (2018) 604–633. analysis for diabetes mellitus, Diabetes Res. Clin. Pract. 97 (2) (2012) 195–205.
[11] X.-Z. Song, F.-F. Sun, S.-T. Dai, X. Lin, K.-M. Sun, X.-F. Wang, Hollow NiFe2O4 [40] S. Ghosal, P. Bhattacharyya, Irreversible n to p transition and corresponding
microspindles derived from Ni/Fe bimetallic MOFs for highly sensitive acetone performance improvement of RGO/TiO2 nanotubes hybrid vapor sensor devices
sensing at low operating temperatures, Inorg. Chem. Front. 5 (5) (2018) by varying electrophoretic deposition time, IEEE Trans. Nanotechnol. 17 (6)
1107–1114. (2018) 1098–1105.
[12] J. Lu, Y. Xie, F. Luo, H. Fu, X. Huang, Y. Liu, H. Liu, Heterostructures of [41] N. Teshima, J. Li, K. Toda, P.K. Dasgupta, Determination of acetone in breath,
mesoporous hollow Zn2SnO4/SnO2 microboxes for high-performance acetone Anal. Chim. Acta 535 (1–2) (2005) 189–199.
sensors, J. Alloys Compd. 844 (2020), 155788. [42] A. Koo, R. Yoo, S.P. Woo, H.-S. Lee, W. Lee, Enhanced acetone-sensing properties
[13] M. Yin, Y. Wang, S. Liu, Synthesis of Fe2O3–ZnWO4 nanocomposites and their of Pt-decorated al-doped ZnO nanoparticles, Sens. Actuators, B 280 (2019)
enhanced acetone sensing performance, J. Alloys Compd. 831 (2020), 154713. 109–119.
[14] X. Kou, N. Xie, F. Chen, T. Wang, L. Guo, C. Wang, Q. Wang, J. Ma, Y. Sun, [43] Q. Jia, H. Ji, Y. Zhang, Y. Chen, X. Sun, Z. Jin, Rapid and selective detection of
H. Zhang, G. Lu, Superior acetone gas sensor based on electrospun SnO2 acetone using hierarchical ZnO gas sensor for hazardous odor markers
nanofibers by Rh doping, Sens. Actuators, B 256 (2018) 861–869. application, J. Hazard Mater. 276 (2014) 262–270.
[15] S. Shao, H. Wu, S. Wang, Q. Hong, R. Koehn, T. Wu, W.-F. Rao, Highly crystalline [44] S.T. Navale, Z.B. Yang, C. Liu, P.J. Cao, V.B. Patil, N.S. Ramgir, R.S. Mane, F.
and ordered nanoporous SnO2 thin films with enhanced acetone sensing property J. Stadler, Enhanced acetone sensing properties of titanium dioxide nanoparticles
at room temperature, J. Mater. Chem. C Mater. Opt. Electron. Devices 3 (41) with a sub-ppm detection limit, Sens. Actuators, B 255 (2018) 1701–1710.
(2015) 10819–10829. [45] H. Zhang, H. Qin, P. Zhang, Y. Chen, J. Hu, Low concentration acetone gas
[16] M. Parthibavarman, B. Renganathan, D. Sastikumar, Development of high sensing properties of 3 Wt% Pd-doped SmCoxFe1-XO3 nanocrystalline powders
sensitivity ethanol gas sensor based on Co-doped SnO2 nanoparticles by under UV light illumination, Sens. Actuators, B 260 (2018) 33–41.
microwave irradiation technique, Curr. Appl. Phys. 13 (7) (2013) 1537–1544. [46] H. Haick, Y.Y. Broza, P. Mochalski, V. Ruzsanyi, A. Amann, Assessment, origin,
[17] S.-W. Tsai, J.-C. Chiou, Improved crystalline structure and H2S sensing and implementation of breath volatile cancer markers, Chem. Soc. Rev. 43 (5)
performance of CuO–Au–SnO2 thin film using SiO2 additive concentration, Sens. (2014) 1423–1449.
Actuators, B 152 (2) (2011) 176–182. [47] M. Righettoni, A. Tricoli, S.E. Pratsinis, Si WO3 sensors for highly selective
[18] J. Yang, J. Liu, Y. Xu, X. Li, J. Wu, Y. Han, Z. Wang, X. Zhang, Enhanced selective detection of acetone for easy diagnosis of diabetes by breath analysis, Anal.
acetone-sensing performance of hierarchical hollow SnO2/α-Fe2O3 microcubes, Chem. 82 (2010) 3581–3587.
J. Mater. Chem. C Mater. Opt. Electron. Devices 7 (38) (2019) 11984–11990. [48] P. Zhu, S. Li, X. Jiang, Q. Wang, F. Fan, M. Yan, Y. Zhang, P. Zhao, J. Yu,
[19] M. Epifani, S. Kaciulis, A. Mezzi, T. Zhang, J. Arbiol, P. Siciliano, A. Landström, Noninvasive and wearable respiration sensor based on organic semiconductor
I. Concina, A. Moumen, E. Comini, C. Xiangfeng, Rhodium as efficient additive for film with strong electron affinity, Anal. Chem. 91 (15) (2019) 10320–10327.
boosting acetone sensing by TiO2 nanocrystals. Beyond the classical view of noble [49] P. Zhu, Y. Wang, P. Ma, S. Li, F. Fan, K. Cui, S. Ge, Y. Zhang, J. Yu, Low-power
metal additives, Sens. Actuators, B 319 (2020), 128338. and high-performance trimethylamine gas sensor based on n-n heterojunction
[20] S. Wei, S. Li, R. Wei, S. Liu, W. Du, Different morphologies of WO3 and their microbelts of perylene diimide/CdS, Anal. Chem. 91 (9) (2019) 5591–5598.
exposed facets-dependent acetone sensing properties, Sens. Actuators, B 329 [50] Z. Zhao, Y. Yong, R. Gao, S. Hu, Q. Zhou, X. Su, Y. Kuang, X. Li, Adsorption,
(2021), 129188. sensing and optical properties of molecules on BC3 monolayer: first-principles
[21] Y.J. Hong, J.-W. Yoon, J.-H. Lee, Y.C. Kang, One-pot synthesis of Pd-loaded SnO2 calculations, Mater. Sci. Eng. B 271 (2021), 115266.
yolk-shell nanostructures for ultraselective methyl benzene sensors, Chemistry 20 [51] X. Wang, Y. Yong, W. Yang, A. Zhang, X. Xie, P. Zhu, Y. Kuang, Adsorption, gas-
(10) (2014) 2737–2741. sensing, and optical properties of molecules on a diazine monolayer: a first-
[22] S. Yang, G. Lei, H. Xu, Z. Lan, Z. Wang, H. Gu, Metal oxide based heterojunctions principles study, ACS Omega 6 (17) (2021) 11418–11426.
for gas sensors: a Review, Nanomaterials 11 (4) (2021) 1026. [52] W. Chen, Y. Gui, T. Li, H. Zeng, L. Xu, Z. Ding, Gas-sensing properties and
[23] B. Saruhan, R. Lontio Fomekong, S. Nahirniak, Review: influences of mechanism of Pd-GaNNTs for air decomposition products in ring main unit, Appl.
semiconductor metal oxide properties on gas sensing characteristics, Front. Sens. Surf. Sci. 531 (2020), 147293.
2 (2021), 657931. [53] Y. Yong, X. Su, Y. Kuang, X. Li, Z. Lu, B40 and M@B40 (M=Li and Ba) fullerenes as
[24] C. Liu, Q. Kuang, Z. Xie, L. Zheng, The effect of noble metal (Au, Pd and Pt) potential molecular sensors for acetone detection: a first-principles study, J. Mol.
nanoparticles on the gas sensing performance of SnO2-based sensors: a case study Liquids 264 (2018) 1–8.
on the 221 high-index faceted SnO2octahedra, CrystEngComm 17 (33) (2015) [54] Y. Yong, X. Su, H. Cui, Q. Zhou, Y. Kuang, X. Li, Two-dimensional tetragonal GaN
6308–6313. as potential molecule sensors for no and NO2 detection: a first-principle study,
[25] S.M. Majhi, P. Rai, Y.T. Yu, Facile approach to synthesize Au@ZnO core-Shell ACS Omega 2 (2017) 8888–8895.
nanoparticles and their application for highly sensitive and selective gas sensors, [55] L.E. Gálvez-González, J.A. Alonso, L.O. Paz-Borbón, A. Posada-Amarillas, H2
ACS Appl. Mater. Interfaces 7 (2015) 9462–9468. adsorption on Cu4-XMX (M = Au, Pt; X = 0–4) clusters: similarities and differences
[26] L. Xiao, S. Shu, S. Liu, A facile synthesis of Pd-doped SnO2 hollow microcubes as predicted by density functional theory, J. Phys. Chem. C 123 (51) (2019)
with enhanced sensing performance, Sens. Actuators, B 221 (2015) 120–126. 30768–30780.
[27] D. Singh, V.S. Kundu, A.S. Maan, Structural, morphological and gas sensing study [56] G. Li, X. Chen, Z. Zhou, F. Wang, H. Yang, J. Yang, B. Xu, B. Yang, D. Liu,
of palladium doped tin oxide nanoparticles synthesized via hydrothermal Theoretical insights into the structural, relative stable, electronic, and gas sensing
technique, J. Mol. Struct. 1100 (2015) 562–569. properties of PbnAun (n = 2–12) clusters: a DFT study, RSC Adv. 7 (72) (2017)
[28] B.-Y. Kim, J.S. Cho, J.-W. Yoon, C.W. Na, C.-S. Lee, J.H. Ahn, Y.C. Kang, J.-H. Lee, 45432–45441.
Extremely sensitive ethanol sensor using Pt-doped SnO2 hollow nanospheres [57] A. Granja-DelRío, J.A. Alonso, M.J. López, Competition between palladium
prepared by Kirkendall diffusion, Sens. Actuators, B 234 (2016) 353–360. clusters and hydrogen to saturate graphene vacancies, J. Phys. Chem. C 121 (20)
[29] K.D. Schierbaum, U.K. Kirner, J.F. Geiger, W. Göpel, Schottky-barrier and (2017) 10843–10850.
conductivity gas sensors based upon Pd/SnO2 and Pt/TiO2, Sens. Actuators, B 4 [58] M. Blanco-Rey, J.I. Juaristi, M. Alducin, M.J. López, J.A. Alonso, Is spillover
(1–2) (1991) 87–94. relevant for hydrogen adsorption and storage in porous carbons doped with
[30] M. Batzill, U. Diebold, The surface and materials science of tin oxide, Prog. Surf. palladium nanoparticles? J. Phys. Chem. C 120 (31) (2016) 17357–17364.
Sci. 79 (2–4) (2005) 47–154. [59] Y. Yong, C. Li, X. Li, T. Li, H. Cui, S. Lv, Ag7Au6 cluster as a potential gas sensor
[31] S. Matsushima, Y. Teraoka, N. Miura, N. Yamazoe, Electronic interaction between for CO, HCN, and NO detection, J. Phys. Chem. C 119 (13) (2015) 7534–7540.
metal additives and tin dioxide in tin dioxide-based gas sensors, Jpn. J. Appl. [60] M. Hu, D.P. Linder, M. Buongiorno Nardelli, A. Striolo, Hydrogen adsorption on
Phys. 27 (10R) (1988) 1798–1802. platinum–gold bimetallic nanoparticles: a density functional theory study,
[32] X. Yang, H. Gao, L. Zhao, T. Wang, P. Sun, F. Liu, G. Lu, Enhanced gas sensing J. Phys. Chem. C 117 (29) (2013) 15050–15060.
properties of monodisperse Zn2SnO4 octahedron functionalized by PdO [61] T.M. Inerbaev, Y. Kawazoe, S. Seal, Theoretical calculations of hydrogen
nanoparticals, Sens. Actuators, B 266 (2018) 302–310. adsorption by SnO2 (110) surface: effect of doping and calcination, J. Appl. Phys.
[33] F. Gyger, A. Sackmann, M. Hübner, P. Bockstaller, D. Gerthsen, H. Lichtenberg, 107 (10) (2010), 104504.
J.-D. Grunwaldt, N. Barsan, U. Weimar, C. Feldmann, Pd@SnO2 and SnO2@Pd [62] I. Swart, F.M.F. de Groot, B.M. Weckhuysen, P. Gruene, G. Meijer, A. Fielicke, H2
Core@shell nanocomposite sensors, Part. Part. Syst. Char. 31 (5) (2014) 591–596. adsorption on 3d transition metal clusters: a combined infrared spectroscopy and
[34] K. Musa-Veloso, S.S. Likhodii, E. Rarama, S. Benoit, Y.-M.C. Liu, D. Chartrand, density functional study, J. Phys. Chem. A 112 (6) (2008) 1139–1149.
R. Curtis, L. Carmant, A. Lortie, F.J.E. Comeau, S.C. Cunnane, Breath acetone [63] H. Köse, A.O. Aydin, H. Akbulut, The effect of temperature on grain size of SnO2
predicts plasma ketone bodies in children with epilepsy on a ketogenic diet, nanoparticles synthesized by sol-gel method, Acta Phys. Pol., A 125 (2) (2014)
Nutrition 22 (1) (2006) 1–8. 345–347.
[35] S.S. Likhodii, K. Musa, S.C. Cunnane, Breath acetone as a measure of systemic [64] N. Ma, K. Suematsu, M. Yuasa, T. Kida, K. Shimanoe, Effect of water vapor on Pd-
ketosis assessed in a rat model of the ketogenic diet, Clin. Chem. 48 (1) (2002) loaded SnO2 nanoparticles gas sensor, ACS Appl. Mater. Interfaces 7 (10) (2015)
115–120. 5863–5869.
[36] T. Toyooka, S. Hiyama, Y. Yamada, A prototype portable breath acetone analyzer [65] N. Ma, K. Suematsu, M. Yuasa, K. Shimanoe, Pd size effect on the gas sensing
for monitoring fat loss, J. Breath Res. 7 (3) (2013), 036005. properties of Pd-loaded SnO2 in humid atmosphere, ACS Appl. Mater. Interfaces 7
[37] Z. Wang, C. Wang, Is breath acetone a biomarker of diabetes? A historical Review (28) (2015) 15618–15625.
on breath acetone measurements, J. Breath Res. 7 (3) (2013), 037109. [66] C. Wang, X. Cui, J. Liu, X. Zhou, X. Cheng, P. Sun, X. Hu, X. Li, J. Zheng, G. Lu,
[38] A. Thati, A. Biswas, S. Roy Chowdhury, T.K. Sau, Breath acetone-based Design of superior ethanol gas sensor based on Al-doped NiO nanorod-flowers,
noninvasive detection of blood glucose levels, Int. J. Smart Sens. Intell. Syst. 8 (2) ACS Sens. 1 (2) (2016) 131–136.
(2015) 1244–1260. [67] N. Yamazoe, New approaches for improving semiconductor gas sensors, Sens.
Actuators, B 5 (1–4) (1991) 7–19.

142
A. Biswas et al. Current Applied Physics 44 (2022) 131–143

[68] M.D. Obradović, Z.M. Stanćić, U.Č Lačnjevac, V.V. Radmilović, A. Gavrilović- [86] R.G. Drabeski, J.V. Gunha(Visualization), A. Novatski, G.B. de Souza, S.
Wohlmuther, V.R. Radmilović, S. Lj Gojković, Electrochemical oxidation of M. Tebcherani, E.T. Kubaski, D.T. Dias, Raman and photoacoustic spectroscopies
ethanol on palladium-nickel nanocatalyst in alkaline media, Appl. Catal. B of SnO2 thin films deposited by spin coating technique, Vib. Spectrosc. 109
Environ. 189 (2016) 110–118. (2020), 103094.
[69] P. Duan, Q. Duan, Q. Peng, K. Jin, J. Sun, Design of ultrasensitive gas sensor based [87] A. Dieguez, A. Romano-Rodriguez, J.R. Morante, U. Weimar, M. Schweizer-
on self-assembled Pd-SnO2/rGO porous ternary nanocomposites for ppb-level Berberich, W. Göpel, Morphological analysis of nanocrystalline SnO2 for gas
hydrogen, Sens. Actuators, B 369 (2022), 132280. sensor applications, Actuators B Chem 31 (1–2) (1996) 1–8.
[70] N. Mohri, B. Oschmann, N. Laszczynski, F. Mueller, J. von Zamory, M.N. Tahir, [88] B.A. Banse, B.E. Koel, Interaction of oxygen with Pd(111): high effective O2
S. Passerini, R. Zentel, W. Tremel, Synthesis and characterization of carbon pressure conditions by using nitrogen dioxide, Surf. Sci. 232 (3) (1990) 275–285.
coated sponge-like tin oxide (SnOx) films and their application as electrode [89] J.G.S. Carrettin, A. Corma, Spectroscopic evidence for the supply of reactive
materials in lithium-ion batteries, J. Mater. Chem. 4 (2016) 612–619. oxygen during CO oxidation catalyzed by gold supported on nanocrystalline
[71] K. Vijayarangamuthu, S. Rath, Nanoparticle size, oxidation state, and sensing CeO2, J. Am. Chem. Soc. 127 (10) (2005) 3286–3287.
response of tin oxide nanopowders using Raman spectroscopy, J. Alloys Compd. [90] R.Q. Long, H.L. Wan, In situ confocal microprobe Raman spectroscopy study of
610 (2014) 706–712. CeO2/BaF2 catalyst for the oxidative coupling of methane, J. Chem. Soc., Faraday
[72] N. Chen, D. Deng, Y. Li, X. Liu, X. Xing, X. Xiao, Y. Wang, TiO2 nanoparticles Trans. 93 (2) (1997) 355–358.
functionalized by Pd nanoparticles for gas-sensing application with enhanced [91] R.Q. Long, Y.P. Huang, H.L. Wan, Surface oxygen species over cerium oxide and
butane response performances, Sci. Rep. 7 (2017) 7692. their reactivities with methane and ethane by means of in situ confocal
[73] C. Wang, N. Xu, T.-T. Liu, W. Xu, H. Guo, Y. Li, P. Bai, X.-P. Wu, X.-Q. Gong, microprobe Raman spectroscopy, J. Raman Spectrosc. 28 (1) (1997) 29–32.
X. Liu, S. Mintova, Mechanical pressure-mediated Pd active sites formation in [92] H. Nakatsuji, H. Nakai, Y. Fukunishi, Dipped adcluster model for chemisorptions
NaY zeolite catalysts for indirect oxidative carbonylation of methanol to dimethyl and catalytic reactions on a metal surface: image force correction and
carbonate, J. Catal. 396 (2021) 269–280. applications to Pd–O2 adclusters, J. Chem. Phys. 95 (1) (1991) 640–647.
[74] J.-C. Dupin, D. Gonbeau, P. Vinatier, A. Levasseur, Systematic XPS studies of [93] E. Hasselbrink, H. Hirayama, A. de Meijere, F. Weik, M. Wolf, G. Ertl,
metal oxides, hydroxides and peroxides, Phys. Chem. Chem. Phys. 2 (2000) Photodissociation and photodesorption of O2 adsorbed on Pd(111), Surf. Sci.
1319–1324. 269–270 (1992) 235–246.
[75] L. Mazeina, Y.N. Picard, J.D. Caldwell, E.R. Glaser, S.M. Prokes, Growth and [94] R. Imbihl, J.E. Demuth, Adsorption of oxygen on a Pd(111) surface studied by
photoluminescence properties of vertically aligned SnO2 nanowires, J. Cryst. high resolution electron energy loss spectroscopy (EELS), Surf. Sci. 173 (2–3)
Growth 311 (2009) 3158–3162. (1986) 395–410.
[76] N.C. Horti, M.D. Kamatagi, N.R. Patil, M.N. Wari, S.R. Inamdar, [95] A.A. Rydosz, Negative correlation between blood glucose and acetone measured
Photoluminescence properties of SnO2 nanoparticles: effect of solvents, Optik – in healthy and type 1 diabetes mellitus patient breath, J. Diabetes Sci. Technol. 9
Int. J. Light Electron Optics 169 (2018) 314–320. (4) (2015) 881–884.
[77] B.V. Ramana, A. Das, S. Amirthapandian, S.K. Dhara, A.K. Tyagi, [96] M.C. Valero, P. Raybaud, P. Sautet, Interplay between molecular adsorption and
Photoluminescence of oxygen vacancies and hydroxyl group surface metal–support interaction for small supported metal clusters: CO and C2H4
functionalized SnO2 nanoparticles, Phys. Chem. Chem. Phys. 17 (2015) adsorption on Pd4/γ-Al2O3, J. Catal. 247 (2) (2007) 339–355.
9794–9801. [97] O.C. Gagne, F.C. Hawthorne, Bond-length distributions for ions bonded to
[78] J. Jeong, S.-P. Choi, C.I. Chang, D.C. Shin, J.S. Park, B.-T. Lee, Y.-J. Park, H.- oxygen: results for the transition metals and quantification of the factors
J. Song, Photoluminescence properties of SnO2 thin films grown by thermal CVD, underlying bond-length variation in inorganic solids, IUCrJ 7 (2020) 581–629.
Solid State Commun. 127 (2003) 595–597. [98] R. Huacuja, D.J. Graham, C.M. Fafard, C.-H. Chen, B.M. Foxman, D.E. Herbert,
[79] H. Luo, L.Y. Liang, H.T. Cao, Z.M. Liu, F. Zhuge, Structural, chemical, optical, and G. Alliger, C.M. Thomas, O.V. Ozerov, Reactivity of a Pd(I)-Pd(I) dimer with O2:
electrical evolution of SnO(x) films deposited by reactive Rf magnetron monohapto Pd superoxide and dipalladium peroxide in equilibrium, J. Am.
sputtering, ACS Appl. Mater. Interfaces 4 (10) (2012) 5673–5677. Chem. Soc. 133 (11) (2011) 3820–3823.
[80] E. Fazio, F. Neri, S. Savasta, S. Spadaro, S. Trusso, Surface-enhanced Raman [99] R. Long, K. Mao, X. Ye, W. Yan, Y. Huang, J. Wang, Y. Fu, X. Wang, X. Wu, Y. Xie,
scattering of SnO2 bulk material and colloidal solutions, Phys. Rev. B Condens. Y. Xiong, Surface facet of palladium nanocrystals: a key parameter to the
Matter 85 (19) (2012). activation of molecular oxygen for organic catalysis and cancer treatment, J. Am.
[81] V.V. Pushkarev, V.I. Kovalchuk, J.L. d’Itri, Probing defect sites on the CeO2 Chem. Soc. 135 (8) (2013) 3200–3207.
surface with dioxygen, J. Phys. Chem. B 108 (17) (2004) 5341–5348. [100] Q. Ren, Y.-Q. Cao, D. Arulraj, C. Liu, D. Wu, W.-M. Li, A.-D. Li, Review—resistive-
[82] X. Mathew, J.P. Enriquez, C. Mejía-García, G. Contreras-Puente, M.A. Cortes- type hydrogen sensors based on zinc oxide nanostructures, J. Electrochem. Soc.
Jacome, J.A. Toledo Antonio, J. Hays, A. Punnoose, Structural modifications of 167 (6) (2020), 067528.
SnO2 due to the incorporation of Fe into the lattice, J. Appl. Phys. 100 (2006), [101] S.S. Li, Semiconductor Physical Electronics, Springer, New York, NY, 2006.
073907. [102] Z. Zhao, Y. Yong, Q. Zhou, Y. Kuang, X. Li, Gas-sensing properties of the SiC
[83] S.H. Sun, G.W. Meng, G.X. Zhang, T. Gao, B.Y. Geng, L.D. Zhang, J. Zuo, Raman monolayer and bilayer: a density functional theory study, ACS Omega 5 (21)
scattering study of rutile SnO2 nanobelts synthesized by thermal evaporation of (2020) 12364–12373.
Sn powders, Chem. Phys. Lett. 376 (1–2) (2003) 103–107. [103] R.G. Parr, R.G. Pearson, Absolute hardness: companion parameter to absolute
[84] M. Aliahmad, M. Dehbashi, Ni-doped SnO nanoparticles synthesized by chemical electronegativity, J. Am. Chem. Soc. 105 (26) (1983) 7512–7516.
Co-precipitation method, Iran. J. Energy Environ. (2013) 49–52. [104] B. Konate, S.T. Affi, N. Ziao, DFT study of dimerization sites in imidazo[1,2-a]
[85] M.X. Gu, L.K. Pan, B.K. Tay, C.Q. Sun, Atomistic origin and temperature Pyridinyl-chalcone series, Comput. Chem. (2021) 1–17, 09 (01).
dependence of Raman optical redshift in nanostructures: a broken bond rule, [105] T. Jiang, Q. He, M. Bi, X. Chen, H. Sun, L. Tao, First-principles calculations of
J. Raman Spectrosc. 38 (6) (2007) 780–788. adsorption sensitivity of Au-doped MoS2 gas sensor to main characteristic gases in
oil, J. Mater. Sci. 56 (2021) 13673–13683.

143

View publication stats