Article

pubs.acs.org/acssensors

Enhanced Room Temperature Oxygen Sensing Properties of LaOCl−

SnO2 Hollow Spheres by UV Light Illumination
Ya Xiong,†,‡ Wenbo Lu,‡ Degong Ding,‡ Lei Zhu,‡ Xiaofang Li,‡ Cuicui Ling,*,‡
and Qingzhong Xue*,†,‡
†
State Key Laboratory of Heavy Oil Processing and ‡College of Science, China University of Petroleum, Qingdao 266580, Shandong
P. R. China
*
S Supporting Information
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
Downloaded via NATL YANG MING CHIAO TUNG UNIV on July 10, 2023 at 08:52:52 (UTC).

ABSTRACT: In this paper, a facile and elegant Green Chemistry method for the synthesis of SnO2
based hollow spheres has been investigated. The influences of doping, crystallite morphology, and
operating condition on the O2 sensing performances of SnO2 based hollow-sphere sensors were
comprehensively studied. It was indicated that, compared with undoped SnO2, 10 at. % LaOCl-doped
SnO2 possessed better O2 sensing characteristics owing to an increase of specific surface area and oxygen
vacancy defect caused by LaOCl dopant. More importantly, it was found that O2 sensing properties of
the 10 at. % LaOCl−SnO2 sensor were significantly improved by ultraviolet light illumination, which
was suited for room-temperature O2 sensing applications. Besides, this sensor also had a better
selectivity to O2 with respect to H2, CH4, NH3, and CO2. The remarkable increase of O2 sensing
properties by UV light radiation can be explained in two ways. On one hand, UV light illumination
promotes the generation of electron−hole pairs and oxygen adsorption, giving rise to high O2 response.
On the other hand, UV light activates desorption of oxygen adsorbates when exposed to pure N2,
contributing to rapid response/recovery speed. The results demonstrate a promising approach for room-
temperature O2 detection.
KEYWORDS: LaOCl-doped SnO2, hollow spheres, oxygen sensing, room temperature detection, UV light radiation

O xygen gas sensors are playing a more and more important

role in our daily life. Their typical applications include
the control of air-fuel ratio in automobile engines to enhance
combustion efficiency, optimizing industrial boilers, medical
ventilation assistance, and so on. Therefore, different types of
sensors including electrochemical,1,2 optical,3 and resistive
sensors4 have been developed to monitor oxygen. However,
the structures of electrochemical sensors are too complicated
while optical sensors are bulky and expensive. Resistive sensors
based on metal oxide semiconductor (MOS) offer advantages
in easy fabrication, miniaturization, simple operation, low
production cost, and so forth. Among various types of MOS
materials, TiO2,5 ZnO,6−10 and SnO24 are the most widely used
to detect oxygen (Table 1) and other gases. However, these
sensors usually have high working temperatures of 150−500
°C, which could lead to declining sensor stability, fast aging,
increased power consumption, and complicated circuitry
required to maintain accurate temperature control. Even
worse, high-temperature operation limits the sensor’s utility
in plant respiration and certain medical and food packaging
industries where oxygen measurements have to be made at low
temperature. Therefore, determining how to develop high
performance and reliable O2 gas sensors that can function at
room temperature (RT) is of vital significance. Ahmed et al.11
reported oxygen gas detection at RT using Mn-ZnO nanorods
fabricated by the microwave-hydrothermal method. The results
revealed that the Mn-ZnO nanorod-based sensor had
significantly better sensing performance than undoped ZnO
nanorod-based sensor due to high specific surface area.
However, only three different oxygen contents (5 ppm, 10
ppm, 15 ppm) have been studied. Hu et al.12 synthesized
SrTiO3 to monitor oxygen at near-ambient temperature (40
°C) using the physical high-energy ball milling technique.
Nevertheless, this perovskite material was highly insulating,
which made the reading of electrical signal much more
complicated. Neri et al.13 developed the Pt−In2O3 based
oxygen sensor for RT operation via a nonaqueous sol−gel
method. The sensor showed a response value of about 1.95
R O2
(S = R N2
) to 20% oxygen. Regretfully, the sensor exhibited
slow kinetics with response/recovery time of 1080/2100 s, a
problem that hitherto persists with the low temperature
operation of MOS sensors. It is worth mentioning that slow
recovery speed may cause much inconvenience when the
sensors are required to continuously detect target gas.
Improving gas-sensing performance with low power
consumption is one of the main concerns for gas sensor
applications. With this consideration in mind, the surface
photoexcitation technique is an effective method to study the
changes of conductance (resistance) of the sensor caused by
light illumination. In recent years, there have been a few reports
on oxygen gas sensors based on ultraviolet (UV) light activated
Received: March 3, 2017
Accepted: May 9, 2017
Published: May 9, 2017

© 2017 American Chemical Society 679 DOI: 10.1021/acssensors.7b00129

ACS Sens. 2017, 2, 679−686
ACS Sensors
Table 1. Comparison of the Performance of O2 Sensors Based on MOS Nanostructured Materials
material synthesis route T (°C)
Nb−TiO2 sol−gel 500
ZnO electrochemical deposition 500
5 mol %P−TiO2 sol−gel 116
Mn−ZnO microhydrothermal RT
Pt−In2O3 sol−gel 200
1 wt %Pt−In2O3 sol−gel RT
10 atom %LaOCl−SnO2 hydrothermal RT
a
Denotes a value not explicitly stated in the study, but approximated from a graphical plot; NA denotes no specific data are available.
MOS.16,17 For example, Wang et al.16 studied how UV light
illumination affected the oxygen sensing behaviors of individual
ZnO nanowire transistors. In his later work, Wang et al.17
reported a route to achieve quick oxygen response in individual
β-Ga 2 O 3 nanowire transistors by UV light activation.
Illuminating these sensors with UV light is a feasible alternative
to activate chemical reactions on metal oxide surface without
the necessity of heating. With regard to SnO2 based sensor, so
far some literature is available on improving its sensing
response toward LPG18 and NO219 by UV photostimulation.
Nevertheless, surprisingly little work has been reported on
enhanced oxygen gas sensing behaviors under light illumina-
tion, which is desperately demanded to fundamentally under-
stand the gas sensing mechanism and expand the potential
applications of SnO2 based sensors.
Previous studies proved that better response could be
achieved by a hollow porous structure based sensor because of
enhancement in active surface area and better gas diffusion.20,21
Herein, undoped and 10 at. % LaOCl-doped SnO2 hollow
spheres (HSs) were prepared by employing newly made carbon
microspheres as templates, and sensors based on these HSs
were fabricated. The reason we choose LaOCl dopant is that La
compounds are excellent additives to MOS owing to their
particular 4f−5d and 4f−4f electronic transition.22 Moreover,
doping LaOCl can increase the amount of defect oxygen
vacancies on the surface of SnO2.23 The influences of operating
condition (RT, 100 °C, blue/UV light illumination), crystallite
morphology (hollow spheres, solid nanoparticles), and relative
humidity on the oxygen sensing performance of the SnO2 based
sensors were systematically studied. In addition, the selectivity
of the 10 at. % LaOCl−SnO2 HSs based sensor toward O2, H2,
CH4, NH3, and CO2 was also investigated.
■ EXPERIMENTAL SECTION
Chemicals. All chemicals including tin(II) chloride dehydrate
(SnCl2·2H2O, Sigma-Aldrich Co., UK), lanthanum chloride (LaCl3,
Sigma-Aldrich Co., UK), glucose (Sinopharm, China), dimethylforma-
mide (DMF, Sigma-Aldrich Co., UK), and ethanol (Sinopharm,
China) were analytical grade reagents and used as received without any
further purifications.
Synthesis and Characterization of SnO2 Based HSs. Carbon
microspheres were prepared as described elsewhere.24 In a typical
procedure, glucose (8 g) was dissolved in deionized water (40 mL)
and then the solution was sealed in a Teflon-lined stainless steel
autoclave and maintained at 180 °C for 12 h. Black products were
obtained after five cycles of centrifugation/washing/redispersion with
deionized water and ethanol, separately, and then oven-dried at 80 °C
overnight.
Synthesis of SnO2 Based HSs.24 The typical synthesis process was
as follows: 1.2 g of SnCl2 with LaCl3 with different atom ratios of La to
Sn (0 and 10 at. %) was dissolved in 20 mL of DMF under stirring. 1.5
g of oven-dried carbon microspheres was uniformly dispersed in 40
mL of DMF by ultrasonication, and then the SnCl2 based solution was
680
Article
response/C (ppm) tres/trec (s) ref.
2.99/20 60/90 5
6.2a/80 000 NA 6
29.6/100 35/20 14
3.8a/15 150/90 11
63.3/200 000 NA 15
1.95/200 000 1080/2100 13
2.25/250 182/1315 Current work
added into the 40 mL of DMF at a speed of 5 s per drop. After
ultrasonicating for about 30 min, a small amount of distilled H2O (e.g.,
2 mL) was added to the solution dropwise to ensure hydrolyzation.
After continuous ultrasonication for 1.5 h, the mixed solution was aged
under ambient conditions for 2 days. The resulting precursor was
collected by centrifugation and washed five times with distilled water
and ethanol, respectively, then oven-dried at 80 °C in vacuum
overnight. SnO2 based HSs were finally obtained after the as-prepared
precursor was calcined at 450 °C for 2 h in a muffle furnace.
Characterization. X-ray diffraction (XRD) patterns were recorded
on a Philips X’Pert diffractometer from 2θ = 10−80° using Cu Kα
radiation (λ = 1.5406 Å). The Brunauer−Emmett−Teller (BET)
specific surface area was measured on a Quantachrome NOVA 2000e
sorption analyzer. Energy-dispersive spectroscopy (EDS) was taken on
a JEOL JEM-2010F. Transmission electron microscope (TEM)
observations were performed with a JEM-2100 UHR microscope. X-
ray photoelectron spectroscopy (XPS) studies were analyzed with a
PHI5000 Versa Probe employing Al Kα radiation (ULVAC-PHI,
Japan). Bonding energy was calibrated with reference to C 1s peak
(285.0 eV).
Fabrication and Gas Sensing Measurement of SnO2-Based
Sensors. The SnO2-based HSs were mixed with the solvent prepared
by dissolving 0.2 g ethyl cellulose into 2.5 mL anhydrous terpineol,
which was then subsequently printed onto commercial alumina
substrate equipped with Pt-interdigitated electrodes and Pt heat
element. Eventually, the SnO2-based films were sintered at 450 °C for
2 h in air to evaporate the organic solvent and improve the mechanical
bond of the deposited samples.
Four LED bulbs which can operate independently are installed in
front of the four SnO2 based sensors (Figure 1). Different bulbs with
Figure 1. Scheme diagram of SnO2 based sensor array structure and
test chamber.
wavelengths of 380 nm (UV light) and 460 nm (blue light) are used.
The distance between the bulb and the sensor is fixed at 10 mm and
the light is focused directly onto the sensor surface. Light intensity at
the position of SnO2 based sensor is 30 mW/cm2.
Light illumination and electrical resistance measurements are
performed inside the glass chamber. Sensor elements are tested with
a commercial gas detection apparatus (Huachuang Ruike Science and
DOI: 10.1021/acssensors.7b00129
ACS Sens. 2017, 2, 679−686
ACS Sensors
Technology Co. Ltd., Hubei Province, China). The measurement
range of resistance is 100 Ω to 1 GΩ with an error less than 5%. Gas
concentration is controlled by four mass flow controllers. An inlet and
outlet port allows a gas supply to be introduced to continuously purge
the apparatus at a total gas flow rate of 2 L/min. The moisture level of
gas is controlled by mixing dry and wet air and the relative humidity
(RH) in the test chamber is measured via a sensor integrated in the
base plate. The O2 response of the sensor is defined as S = R0/R, while
for reducing gases such as H2, CH4, NH3, and CO2, it is defined as S =
R/R0, where R0 is the resistance of the material when exposed to a
certain concentration of target gas and R is the resistance of the
material in N2 or air background atmosphere. The response time (tres)
is defined as the time needed for the sensor electrical resistance to
change from R to R+90% (R0 − R) when shifted from N2 to oxygen
while recovery time (trec) represents the time required for the sensor
resistance to change from R0 to R0 −90% (R0 − R) when switched
back to N2 from oxygen.
■ RESULTS AND DISCUSSION
Characterization of the Samples. XRD analyses of the
pure SnO2 and 10 at. % LaOCl−SnO2 samples (Figure 2) show
Figure 2. X-ray diffraction patterns of (a) pure SnO2, (b) 10 at. %
LaOCl−SnO2 HSs (inset shows the corresponding X-ray diffraction
patterns of pure SnO2 and 10 at. % LaOCl−SnO2 HSs at 2θ = 25−
29°).
diffraction peaks that are characteristic of the tetragonal rutile
structure of SnO2 (JCPDS 41−1445). The mean crystallite
sizes of pure SnO2 and 10 at. % LaOCl−SnO2 deduced from
the Scherrer equation are 34.64 and 30.94 nm, respectively.
XRD pattern of the 10 at. % LaOCl−SnO2 sample mainly
shows strong reflections of SnO2. However, the presence of
small and well-dispersed tetragonal crystalline LaOCl nano-
particles is confirmed by the emergence of weak (110) and
(103) reflections centered at 2θ = 30.46° and 45.44°. The inset
shows the shift in the position of the major diffraction peak of
LaOCl−SnO2 toward lower angle comparing that of pure
SnO2, which suggests that LaOCl is strongly incorporated into
the SnO2 crystal lattice. The Brunaeur-Emmet-Teller (BET)
surface areas and pore sizes of pure SnO2 and 10 at. % LaOCl−
SnO2 HSs are shown in Table 2. It is found that LaOCl doping
can effectively increase surface area and decrease pore size of
SnO2 based HSs.
TEM images of pure SnO2 and 10 at. % LaOCl−SnO2 HSs
are shown in Figure 3. It is clear that the shells are porous with
a thickness of about 20 nm and consist of massive SnO2 based
nanoparticles. Most of the SnO2 HSs have a rough morphology
with a diameter of about 200 nm and the decrease of grain size
is significant when doping LaOCl. The EDS image in the inset
of Figure 3b clearly reveals the existence of La and Cl in the 10
681
Article
Table 2. Average BET Surface Areas and Pore Sizes of Pure
SnO2 and 10 at. % LaOCl−SnO2 HSs
material BET surface area (m2/g) pore size (nm)
Pure SnO2 34.98 7.084
10 at.% LaOCl−SnO2 43.97 6.677
at. % LaOCl−SnO2. The left of C element may be due to the
TEM carbon-coated Cu grid used to support the sample.
Since the chemical characteristics of the surface of MOS are
decisive on the sensing properties, the chemical composition of
different films was investigated using XPS. Figure 4a shows
typical XPS survey spectra recorded on pure SnO2 HSs and 10
at. % LaOCl−SnO2 HSs. One pronounced feature in the
spectrum recorded on 10 at. % LaOCl−SnO2 HSs is the
presence of La 3d core level. XPS spectrum of La element is
shown in Figure 4b. Figure 4c,d displays the XPS spectra of
oxygen for pure SnO2 HSs and 10 at. % LaOCl−SnO2 HSs,
respectively. The spectrum of O 1s can be fitted into nearly
three peaks. The peak at 529.5 eV in the O 1s spectrum
corresponds to crystal lattice (OL) in the SnO2 phase while the
medium binding energy of 531.1 eV is associated with oxygen
vacancies (OV) within the matrix of SnO2.25 The peak located
at 532.2 eV can be assigned to adsorbed and dissociated oxygen
species (OC). From Table 3, we know the concentration of OV:
10 at. % LaOCl−SnO2 > pure SnO2.
O2 Sensing Properties. Figure 5 shows electrical resistance
and response changes of 10 at. % LaOCl−SnO2 sensor to 1000
ppm of O2 without light illumination (RT and 100 °C) and
illuminated by blue and UV lights at RT with the same light
density of 30 mW/cm2. It is worth mentioning that here the
condition of no light illumination actually means no additional
light but natural light. As can be seen in Figure 5a,b, the
response of the sensor to O2 at RT without light illumination is
insignificant and the sensor can hardly recover to its baseline
resistance within the allotted time. However, when the sensor is
illuminated under blue or UV light, obvious improvements in
response time and especially recovery time are observed. In
addition, the baseline resistance of the sensor is lower under
UV light compared with that under blue light. Not only the
sensor response but also the response/recovery speed is higher
when exposed to shorter light. In order to compare the sensing
characteristics of 10 at. % LaOCl−SnO2 sensor under light
illumination and heat activation, the same sensor is heated at
100 °C and tested with the same O2 concentration under no
light illumination. Compared with the sensor operating at 100
°C, the sensor illuminated by blue or UV light has a better
response.
To sum up, 380 nm UV light is the best choice to get the
optimal sensor response while maintaining applausive re-
sponse/recovery time. Therefore, RT UV light activation was
chosen as the operation condition in all further experiments.
Figure 6 shows the response/recovery characteristic of pure
SnO2 and 10 at. % LaOCl−SnO2 sensors to 250 ppm of O2
illuminated by 380 nm UV light. It is found that LaOCl doping
not only increases the sensor response but also accelerates the
response/recovery speed. For example, the response of 10 at. %
LaOCl−SnO2 to 250 ppm of O2 is 2.25 with response/recovery
time of 161/1003 s, while that of pure SnO2 is only 1.14 with
response/recovery time of 182/1315 s.
The normalized resistance and response changes of pure
SnO2 and 10 at. % LaOCl−SnO2 sensors with time under
different O2 gas concentrations is shown in Figure 7.
DOI: 10.1021/acssensors.7b00129
ACS Sens. 2017, 2, 679−686
ACS Sensors Article

Figure 3. TEM images of (a) pure SnO2 HSs and (b) 10 at. % LaOCl−SnO2 HSs (insets show the corresponding EDS images).

Figure 4. (a) XPS survey spectra recorded on pure SnO2 and 10 at. % LaOCl−SnO2 HSs, (b) XPS spectrum of La 3d for 10 at. % LaOCl−SnO2
HSs, corresponding peak fittings of O 1s energy region of (c) pure SnO2 HSs and (d) 10 at. % LaOCl−SnO2 HSs.

Table 3. Peak Area Ratios of the Oxygen Species (OL, OV, The 10 at. % LaOCl−SnO2 HSs sensor is also tested to
OC) of Pure SnO2 and 10 at. % LaOCl−SnO2 examine the selectivity against other gases, including H2, CH4,
species position (eV) pure SnO2 10 at.% LaOCl−SnO2
NH3, and CO2 (each gas concentration is 400 ppm; the
response tests of H2, CH4, NH3, and CO2 are performed under
OL 529.5 3.50% 2.20%
air background atmosphere). It is found that the sensor shows
OV 531.1 53.94% 56.18%
high response to O2 but little response to H2, CH4, NH3, and
OC 532.2 42.46% 41.62%
CO2, as can be seen in Figure 8. Thus, it is believed that the 10
at. % LaOCl−SnO2 HSs sensor has good selectivity to O2. The
Apparently, the responses of two sensors increase with large difference in response in our case can be attributed to
increasing gas concentration. Moreover, the response value different sensing mechanisms of O2 and other gases.
and recovery speed of 10 at. % LaOCl−SnO2 sensor are much To substantiate the effect of structure on oxygen sensing
higher than those of pure SnO2 sensor. The inset in Figure 7a performance of the sensors, a comparative study between 10 at.
shows an enlarged part of the measured data at 100 ppm of O2, % LaOCl−SnO2 HSs and 10 at. % LaOCl−SnO2 solid
indicating that the 10 at. % LaOCl−SnO2 sensor responds well nanoparticles was performed, keeping other conditions the
even to as low as 100 ppm of O2 gas. As demonstrated in same. The 10 at. % LaOCl−SnO2 particles prepared by the
Figure 7b, the responses of two sensors increase rapidly with ammonia precipitation method26 have a mean crystallite size
increasing O2 concentration from 100 to 1000 ppm, and then (49.54 nm, calculated from Figure S1, Supporting Information)
almost follow a linear increase with further increasing O2 bigger than that of 10 at. % LaOCl−SnO2 HSs (30.94 nm,
concentration. Figure 2). The BET surface area of the prepared 10 at. %
682 DOI: 10.1021/acssensors.7b00129
ACS Sens. 2017, 2, 679−686
ACS Sensors Article

Figure 5. Electrical resistance change (a) and response change (b) of 10 at. % LaOCl−SnO2 HSs based sensor exposed to 1000 ppm of O2 at
different operating conditions (RT with no light illumination, 100 °C with no light illumination, RT illuminated by blue light, RT illuminated by UV
light).

Figure 8. Cross-response of the sensor based on 10 at. % LaOCl−

Figure 6. Electrical resistance changes of pure SnO2 and 10 at. %
LaOCl−SnO2 HSs based sensors exposed to 250 ppm of O2 with UV
light illumination.
LaOCl−SnO2 nanoparticles is 30.30 m2/g, which is smaller
than that of 10 at. % LaOCl−SnO2 HSs (43.97 m2/g). As
shown in Figure 9, the sensor based on 10 at. % LaOCl−SnO2
HSs has not only high response but also short response/
recovery time compared with 10 at. % LaOCl−SnO2 particles.
The enhancement in gas sensing properties lies in the ability of
HSs to facilitate UV light excitation and O2 gas diffusion,
making O2 molecules accessible to both the exterior and
interior surfaces of SnO2 based HSs.
The response variations of the 10 at. % LaOCl−SnO2 HSs
sensor to 1000 ppm of O2 at different relative humidities under
SnO2 HSs sensor to various gases under UV light irradiation.
UV light are also investigated (Figure S2, Supporting
Information). It can be speculated that, when the RH is
below 30%, moisture evidently has a positive influence on the
O2 response, while the response remains almost the same when
RH is higher than 30%.
The stability and reproducibility are also important
parameters for the commercial use of any electronic devices.
In order to check the stability of the 10 at. % LaOCl−SnO2
HSs sensor, its response toward 250 ppm of O2 was measured
for two months, at an interval of 3 days, as shown in Figure 10.
It is found that O2 response of the 10 at. % LaOCl−SnO2
hollow-sphere sensor shows no obvious change during two
months, confirming the good stability of the sensor.

Figure 7. Electrical resistance changes (a) and response changes (b) of pure SnO2 and 10 at. % LaOCl−SnO2 HSs based sensors as a function of O2
concentration with UV light (inset shows dynamic resistance curves of two sensors toward 100 ppm of O2).

683 DOI: 10.1021/acssensors.7b00129

ACS Sens. 2017, 2, 679−686
ACS Sensors
Figure 9. Electrical resistance changes of 10 at. % LaOCl−SnO2 HSs
and 10 at. % LaOCl−SnO2 nanoparticle-based sensors to 250 ppm of
O2 with UV light illumination.
Figure 10. Responses of 10 at. % LaOCl−SnO2 HSs based sensor to
250 ppm of O2 repeated with 20 iterations of testing during two
months under UV light illumination.
Sensing Mechanism. Without Light Illumination. It is
known that the chemical potential of oxygen is below the
conduction band of SnO2. As exposed to oxygen at RT, the
majority of free electrons from SnO2 will be trapped by
adsorbed oxygen on the surface with the formation of O−2(ads)
(chemisorbed oxygen ion, eqs 1−2),27 which induces the
formation of depletion layers (Figure 11a).
O2(g) ⇔ O2(ads) (1)
O2(ads) + e− ⇔ O2(ads)− (2)
Without light illumination, the carrier concentration on the
surface of HSs is very low and almost all free electrons are
trapped by adsorbed oxygen; further modulation of the sensors’
resistance by oxygen will be very small. As a result, the sensors
are not sensitive to oxygen. Since the desorption activation
energy is much higher than chemisorption activation energy,
oxygen adsorbates are difficult to desorb from the surface of
SnO2 based films at RT.28 Therefore, an extremely sluggish
recovery is observed at RT without light illumination.
At 100 °C, molecular and/or adsorbed oxygen undergoes
dissociation according to the reactions below (eqs 3−4).11 The
abstraction of electrons from the SnO2 surface occurs with a
relatively high rate, leading to a fast increase of oxygen
adsorption content and great enhancement of sensor response.
In addition, the temperature of 100 °C facilitates desorption of
oxygen adsorbates when exposed to pure N2. Therefore, O2
684
Article
Figure 11. Schematic illustration of the sensing process of the RT
SnO2 HSs based sensor without (a) and with (b) UV light
illumination.
sensing characteristics of the sensor are boosted when
operating at 100 °C compared with those at RT.
O2(ads)− + e− ⇔ 2O(ads)− (3)
−
O2(ads) + 2e ⇔ 2O(ads)− (4)
With Light Illumination. When investigating the effects of
light illumination on the sensing properties of SnO2 based
sensors, it is suggested that gas sensor performances are
affected through the following ways: the generation of
electron−hole pairs29 and the desorption of oxygen adsor-
bates.30
Generation of Electron−Hole Pairs. As known to all, the
required energy for band to band excitation depends on the
band gap energy of semiconductor. SnO2 exhibits a band gap of
3.6 eV. The photon energy is determined by light wavelength
(E = hcλ−1). This means that the photon energies of 380 nm
hc 4.136 × 10−15 × 3 × 1017
( E = hν = λ = 380
eV = 3.3 eV < 3.6 eV )
UV light and 460 nm (E = 2.7 eV < 3.6 eV) blue light are not
sufficient for band to band excitation of SnO2. However,
previous density functional theory (DFT) studies indicated that
oxygen vacancies could form shallow donor levels, allowing
optical excitation of electrons with light energies even below
the band gap energy.31,32
hυ → e− + h+ (5)
−
O2 + e ⇔ O2 (hυ) −
(6)
SnO2 shows an n-type electrical conductivity, meaning that
the material is nonstoichiometric with available oxygen vacancy
defects. Doping of LaOCl further increases the amount of
oxygen vacancy defects on the surface of 10 at. % LaOCl−
SnO2, as proofed by the XPS results above. Therefore, we
affirm that the excitation energy used in our experiments (380
nm, 3.3 eV; 460 nm, 2.7 eV) is sufficient for electron excitation
from valence band to conduction band owing to the oxygen
vacancies (eq 5). In this case, substantial photogenerated
electrons will be trapped by oxygen and photoinduced oxygen
ions (O2−(hυ)) are thus created (eq 6), as illustrated in Figure
11b. The experimental results above have shown that sensor
response S(RT illuminated by UV light) > S(RT illuminated by blue light) >
DOI: 10.1021/acssensors.7b00129
ACS Sens. 2017, 2, 679−686
ACS Sensors
S(100 °C without light illumination) > S(RT without light illumination). The im-
provement of R0/R under UV light radiation compared with
that under other conditions can be explained by the largest
resistance variation (ΔR = R0 − R) due to the largest number of
adsorbed oxygens under UV light activation.
Desorption of Oxygen Adsorbates. There are two
mechanisms for the direct oxygen desorption under UV light,
namely, recombination with a light-generated hole in the
valence band as reflected by eq 7 and direct excitation of a
binding electron from oxygen adsorbates to the conduction
band of SnO2 (eq 8). It is believed that the adsorption energy
of an oxygen atom on a metal oxide is about 1.5 eV33 and
photoinduced oxygen ions O2− (hν) is weakly adsorbed on the
surface of material.34 Hence, eqs 7−8 are possible as the
photons from the 380 nm (3.3 eV) UV light and 460 nm (2.7
eV) blue light are sufficiently energetic. As a result, the
chemisorbed oxygen ions (O−2(ads)) and photoinduced oxygen
ions O2− (hν) can be easily removed under light activation,
giving rise to rapid recovery. In addition, O2− (O−2(ads) and O2−
(hν)) desorption will be faster when higher light energy (hν) is
provided, which explains the reduced recovery time under UV
light compared with that under blue light.
O2− + h+ → O2 (7)
O2− + hυ ⇔ O2 + e− (8)
As for the effect of water vapor on the response of the 10 at.
% LaOCl−SnO2 HSs based sensor, there are two mechanisms
proposed by Heiland and Kohl35 to explain the increase of
surface conductivity in the presence of water vapor.
H 2O + Snlat + Olat → (HO− − Snlat +) + (Olat H)+ + e−
(9)
H 2O + 2Snlat + Olat → 2(HO− − Snlat +) + VO2+ + 2e−
(10)
where Snlat is denominated as lattice Sn, Olat represents lattice
oxygen, and VO 2+ represents bivalent oxygen vacancy.
According to the two interaction mechanisms, water vapor
offers the necessary condition for the creation of electrons and
oxygen vacancy defects, contributing to higher oxygen
adsorption amount. Therefore, to a certain extent (RH is
below 30%), water adsorption enhances the sensor response.
However, RH higher than 30% leads to a large coverage of
hydroxyl groups on the sensor surface, limiting the adsorption
of oxygen. Hence the sensor response remains unchanged when
RH is beyond 30%. To thoroughly clarify the effect of water
vapor, a systematic investigation will be performed in our next
study.
The effect of LaOCl dopant on the oxygen sensing properties
of the sensor can be mainly demonstrated in two ways. On one
hand, the grain size diminishes and specific surface area
increases when doping LaOCl, which facilitate the adsorption
of oxygen. On the other hand, doping LaOCl increases the
amount of oxygen vacancy defects on the surface of SnO2,
promoting the photogeneration of electrons and the increase of
oxygen adsorption amount.
■
CONCLUSIONS
In summary, sensors based on pure SnO2 and 10 at. % LaOCl−
SnO2 HSs were fabricated using novel carbon microspheres as
templates and tested toward O2 in different operating
conditions. It was demonstrated that the 10 at. % LaOCl−
685
Article
SnO2 HSs based sensor showed applausive room temperature
O2 sensing performances compared with pure SnO2. It was also
found that 10 at. % LaOCl−SnO2 HSs sensor showed higher
response and shorter response/recovery time when illuminated
by UV light. The 10 at. % LaOCl−SnO2 HSs sensor also
exhibited a good selectivity to O2 against H2, CH4, NH3, and
CO2, which is believed to be a promising candidate for the
effective detection of O2 gas at room temperature. Further-
more, an explanation for the UV light activated O2 gas-sensing
mechanism by adsorption−desorption model is proposed. UV
light illumination not only promotes the generation of
electron−hole pairs and oxygen adsorption, but also activates
desorption of oxygen adsorbates when exposure to pure N2,
thus contributing to high response and rapid response/recovery
speed. We believe that our research can surely open up new
opportunities for MOS sensors operating at room temperature.
■
*
ASSOCIATED CONTENT
S Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acssen-
sors.7b00129.
X-ray diffraction patterns of 10 at. % LaOCl−SnO2 solid
nanoparticles; response variations of the 10 at. %
LaOCl−SnO2 exposed to 1000 ppm of O2 at RT
under UV light with relative humidity from 0 to 50%
(PDF)
■ AUTHOR INFORMATION
Corresponding Authors
*E-mail: lingcuicui@upc.edu.cn (C. C. Ling). Tel.: 86-0532-
86981169. Fax: 86-0532-86981169
*E-mail: xueqingzhong@tsinghua.org.cn (Q. Z. Xue).
ORCID
Qingzhong Xue: 0000-0002-2873-8037
Notes
The authors declare no competing financial interest.
■ ACKNOWLEDGMENTS
This work is supported by Natural Science Foundation of
China (11374372, 11604390), Natural Science Foundation of
Shandong Province (ZR2014EMQ006), Taishan Scholar
Foundation (ts20130929), Fundamental Research Funds for
the Central Universities (15CX08009A).
■
REFERENCES
(1) Yamazoe, N.; Miura, N. Development of gas sensors for
environmental protection. IEEE Trans. Compon., Packag., Manuf.
Technol., Part A 1995, 18, 252−256.
(2) Weppner, W. Advanced principles of sensors based on solid state
ionics. Mater. Sci. Eng., B 1992, 15, 48−55.
(3) Amao, Y.; Okura, I. Optical oxygen sensing materials:
chemisorption film of ruthenium (II) polypyridyl complexes attached
to anionic polymer. Sens. Actuators, B 2003, 88, 162−167.
(4) Kolmakov, A.; Zhang, Y.; Cheng, G.; Moskovits, M. Detection of
CO and O2 using tin oxide nanowire sensors. Adv. Mater. 2003, 15,
997−1000.
(5) Sotter, E.; Vilanova, X.; Llobet, E.; Vasiliev, A.; Correig, X. Thick
film titania sensors for detecting traces of oxygen. Sens. Actuators, B
2007, 127, 567−579.
(6) Eriksson, J.; Khranovskyy, V.; Söderlind, F.; Käll, P. O.;
Yakimova, R.; Spetz, A. L. ZnO nanoparticles or ZnO films: a
DOI: 10.1021/acssensors.7b00129
ACS Sens. 2017, 2, 679−686
ACS Sensors Article

comparison of the gas sensing capabilities. Sens. Actuators, B 2009, 137, (27) Barsan, N.; Weimar, U. Conduction model of metal oxide gas
94−102. sensors. J. Electroceram. 2001, 7, 143−167.
(7) Tian, H.; Fan, H.; Li, M.; Ma, L. Zeolitic imidazolate framework (28) Brinzari, V.; Korotcenkov, G.; Golovanov, V. Factors influencing
coated ZnO nanorods as molecular sieving to improve selectivity of the gas sensing characteristics of tin dioxide films deposited by spray
formaldehyde gas sensor. ACS sensors 2016, 1, 243−250. pyrolysis: understanding and possibilities of control. Thin Solid Films
(8) Ma, L.; Fan, H.; Tian, H.; Fang, J.; Qian, X. The n-ZnO/n-In2O3 2001, 391, 167−175.
heterojunction formed by a surface-modification and their potential (29) Fritzsche, H.; Pashmakov, B.; Claflin, B. Reversible changes of
barrier-control in methanal gas sensing. Sens. Actuators, B 2016, 222, the optical and electrical properties of amorphous InOx by
508−516. photoreduction and oxidation. Sol. Energy Mater. Sol. Cells 1994, 32,
(9) Tian, H.; Fan, H.; Dong, G.; Ma, L.; Ma, J. NiO/ZnO p-n 383−393.
heterostructures and gas sensing properties for reduced operating (30) Giberti, A.; Malagù, C.; Guidi, V. WO3 sensing properties
temperature. RSC Adv. 2016, 6, 109091−109098. enhanced by UV illumination: an evidence of surface effect. Sens.
(10) Song, N.; Fan, H.; Tian, H. PVP assisted in situ synthesis of Actuators, B 2012, 165, 59−61.
functionalized graphene/ZnO (FGZnO) nanohybrids with enhanced (31) Pashmakov, B.; Claflin, B.; Fritzsche, H. Photoreduction and
gas-sensing property. J. Mater. Sci. 2015, 50, 2229−2238. oxidation of amorphous indium oxide. Solid State Commun. 1993, 86,
(11) Ahmed, F.; Arshi, N.; Anwar, M. S.; Danish, R.; Koo, B. H. Mn- 619−622.
doped ZnO nanorod gas sensor for oxygen detection. Curr. Appl. Phys. (32) Heiland, G.; Mollwo, E.; Stöckmann, F. Electronic processes in
2013, 13, S64−S68. zinc oxide. Solid State Phys. 1959, 8, 191−323.
(12) Hu, Y.; Tan, O. K.; Pan, J. S.; Yao, X. A new form of nanosized (33) Batzill, M.; Diebold, U. The surface and materials science of tin
SrTiO3 material for near-human-body temperature oxygen sensing oxide. Prog. Surf. Sci. 2005, 79, 47−154.
applications. J. Phys. Chem. B 2004, 108, 11214−11218. (34) Prades, J. D.; Jimenez-Diaz, R.; Manzanares, M.; Hernandez-
(13) Neri, G.; Bonavita, A.; Micali, G.; Rizzo, G.; Galvagno, S.; Ramirez, F.; Cirera, A.; Romano-Rodriguez, A.; Mathur, S.; Morante, J.
Niederberger, M.; Pinna, N. A highly sensitive oxygen sensor operating R. A model for the response towards oxidizing gases of photoactivated
at room temperature based on platinum-doped In2O3 Nanocrystals. sensors based on individual SnO2 nanowires. Phys. Chem. Chem. Phys.
Chem. Commun. 2005, 37, 6032−6034. 2009, 11, 10881−10889.
(14) Han, Z.; Wang, J.; Liao, L.; Pan, H.; Shen, S.; Chen, J. (35) Heiland, G.; Kohl, D. Physical and chemical aspects of oxidic
Phosphorus doped TiO2 as oxygen sensor with low operating semiconductor gas sensors. Chemical Sensor Technology 1988, 1, 15−
temperature and sensing mechanism. Appl. Surf. Sci. 2013, 273, 38.
349−356.
(15) Neri, G.; Bonavita, A.; Micali, G.; Rizzo, G.; Pinna, N.;
Niederberger, M. In2O3 and Pt-In2O3 nanopowders for low temper-
ature oxygen sensors. Sens. Actuators, B 2007, 127, 455−462.
(16) Li, Q. H.; Liang, Y. X.; Wan, Q.; Wang, T. H. Oxygen sensing
characteristics of individual ZnO nanowire transistors. Appl. Phys. Lett.
2004, 85, 6389−6391.
(17) Feng, P.; Xue, X. Y.; Liu, Y. G.; Wan, Q.; Wang, T. H. Achieving
fast oxygen response in individual β-Ga2O3 nanowires by ultraviolet
illumination. Appl. Phys. Lett. 2006, 89, 112114.
(18) Haridas, D.; Chowdhuri, A.; Sreenivas, K.; Gupta, V. Enhanced
room temperature response of SnO2 thin film sensor loaded with Pt
catalyst clusters under UV radiation for LPG. Sens. Actuators, B 2011,
153, 152−157.
(19) Park, S.; An, S.; Mun, Y.; Lee, C. UV-enhanced NO2 gas sensing
properties of SnO2-core/ZnO-shell nanowires at room temperature.
ACS Appl. Mater. Interfaces 2013, 5, 4285−4292.
(20) Sun, X.; Li, Y. Ga2O3 and GaN semiconductor hollow spheres.
Angew. Chem. 2004, 116, 3915−3919.
(21) Manjula, P.; Boppella, R.; Manorama, S. V. A facile and green
approach for the controlled synthesis of porous SnO2 nanospheres:
application as an efficient photocatalyst and an excellent gas sensing
material. ACS Appl. Mater. Interfaces 2012, 4, 6252−6260.
(22) Lucovsky, G.; Phillips, J. C. Defects and defect relaxation at
internal interfaces between high-k transition metal and rare earth
dielectrics and interfacial native oxides in metal oxide semiconductor
(MOS) structures. Thin Solid Films 2005, 486, 200−204.
(23) Xiong, Y.; Xue, Q.; Ling, C.; Lu, W.; Ding, D.; Zhu, L.; Li, X.
Effective CO2 detection based on LaOCl-doped SnO2 nanofibers:
insight into the role of oxygen in carrier gas. Sens. Actuators, B 2017,
241, 725−734.
(24) Sun, X.; Liu, J.; Li, Y. Use of carbonaceous polysaccharide
microspheres as templates for fabricating metal oxide hollow spheres.
Chem. - Eur. J. 2006, 12, 2039−47.
(25) Kawabe, T.; Tabata, K.; Suzuki, E.; Yamaguchi, Y.; Nagasawa, Y.
Electronic states of chemisorbed oxygen species and their mutually
related studies on SnO2 thin film. J. Phys. Chem. B 2001, 105, 4239−
4244.
(26) Fu, C.; Wang, J.; Yang, M.; Su, X.; Xu, J.; Jiang, B. Effect of La
doping on microstructure of SnO2 nanopowders prepared by co-
precipitation method. J. Non-Cryst. Solids 2011, 357, 1172−1176.

686 DOI: 10.1021/acssensors.7b00129

ACS Sens. 2017, 2, 679−686