Modifying a composite based on silica molecular sieve and a Ru(II)-based probe with Fe3O4 particles: Construction and oxygen sensing performance

Article abstract of DOI:10.1016/j.ica.2016.02.061

This paper reported a composite based on silica molecular sieve MCM-41 and a Ru(II)-based probe which was further functionalized with magnetic Fe₃O₄ so that site-specific guiding could be achieved. A core–shell structure was applied in this composite, with Fe₃O₄ as core and MCM-41 as shell, respectively. By means of electron microscope images, XRD analysis, IR spectra, N₂ adsorption/desorption measurement and thermal degradation analysis, this composite was analyzed and confirmed. Emission monitoring of this composite under various O₂ concentrations suggested that its emission was quenchable by O₂ through a dynamic mechanism with good stability. Sensitivity of 11.5 and short response time of 10 s were obtained with a linear working plot.

Full text of DOI:10.1016/j.ica.2016.02.061

Inorganica Chimica Acta 446 (2016) 24–31
Contents lists available at ScienceDirect
Inorganica Chimica Acta
journal homepage: www.elsevier.com/locate/ica
Modifying a composite based on silica molecular sieve and a
Ru(II)-based probe with Fe3O4 particles: Construction and
oxygen sensing performance
⇑
Chen Lin, Sun Shuai, Li Jia, Fang Zhigang
School of Chemical Engineering, University of Science and Technology Liaoning, Anshan 114051, China
a r t i c l e i n f o
a b s t r a c t
Article history:
This paper reported a composite based on silica molecular sieve MCM-41 and a Ru(II)-based probe which
was further functionalized with magnetic Fe3O4 so that site-specific guiding could be achieved. A core–
shell structure was applied in this composite, with Fe3O4 as core and MCM-41 as shell, respectively. By
means of electron microscope images, XRD analysis, IR spectra, N2 adsorption/desorption measurement
and thermal degradation analysis, this composite was analyzed and confirmed. Emission monitoring of
this composite under various O2 concentrations suggested that its emission was quenchable by O2
through a dynamic mechanism with good stability. Sensitivity of 11.5 and short response time of 10 s
were obtained with a linear working plot.
Received 31 January 2016
Received in revised form 29 February 2016
Accepted 29 February 2016
Available online 5 March 2016
Keywords:
Fe3O4 core
O2 sensing
Ru(II) complex
Linear response
Stern–Volmer plot
Ó 2016 Elsevier B.V. All rights reserved.
1. Introduction
probe. Among the numerous candidates for supporting matrix, a
silica molecular sieve MCM-41 has been frequently suggested in
Being a life-supporting gas, molecular O2 has been considered
important. Its quantification is consequently highly focused in var-
ious fields such as chemical industry, environmental protection
and food processing/preservation [1,2]. As a novel quantification
method, optical sensing has been highly proposed among the
numerous candidates owing to its advantages of instant response,
simple operational procedure, requiring no sophisticated appara-
tus and low cost. In addition, optical signals are free of electromag-
netic interference, which makes long-distance on-line monitoring
and in-field detection possible [3–6]. After comparing performance
between sensing systems based on pure materials and those based
on organic–inorganic composite materials, composite sensing sys-
tems are found better since they gather virtues of each component
and well preserve them, resulting in various feature combinations
to meet practical application [7–12].
virtue of its highly ordered tunnels which satisfy criteria for an
ideal supporting matrix well [11,12].
While, the organic component in a composite material is usu-
ally applied as sensing probe owing to its good optoelectronic fea-
tures [10–17]. Aiming at complete and fast sensing with analyte,
long lifetime and broad distribution of excited electrons are
expected. It appears that luminescent metal complexes are promis-
ing ones. Their emission is originated from triplet metal-to-ligand-
charge-transfer (MLCT) excited state and generally has a long-lived
lifetime of microseconds. Generally, excited electrons are localized
on conjugation chain of ligands and thus can be readily spread, sat-
isfying above demands.
To properly combine organic and inorganic components and
preserve their features, a number of hybrid structures have been
proposed [11–18]. Core–shell structure has been considered
promising owing to its simple construction route, excellent holding
and preservation of components [15–18]. There is only one issue
that fails to be counted for practical applications, which is
site-specific guiding. Guided by above consideration, we intend
to modify an optical sensing system with magnetic Fe3O4 so that
site-specific oxygen sensing can be achieved. Its design strategy
and detailed construction route are shown as Scheme 1 (see
Supporting information for a detailed explanation on design
strategy of Ru-Phen@MCM-41@Fe3O4).
In this case, inorganic component of a composite system usually
serves as supporting matrix owing to its good mechanical strength
and stability. To realize desired performance, some criteria should
be met by such supporting matrix, including high diffusion coeffi-
cient, uniform microenvironment and compatibility with sensing
⇑
Corresponding author. Tel.: +86 0412 5929637.
E-mail address: anshan_chenlin1@163.com (L. Chen).
http://dx.doi.org/10.1016/j.ica.2016.02.061
0020-1693/Ó 2016 Elsevier B.V. All rights reserved.

L. Chen et al. / Inorganica Chimica Acta 446 (2016) 24–31
25
Scheme 1. Design strategy and detailed construction route of Ru-Phen@MCM-41@Fe3O4.
and a JEOL JEM-2010 transmission electron microscope, respec-
tively. Mesoporous analysis was carried out on a Nova l000 ana-
lyzer with Barrett–Joyner–Halenda (BJH) model. Diffusion
coefficients of N2 and O2 in our composite were determined by a
H-Sorb 2600T Gas analyzer. For sensing performance evaluation,
Ru-Phen@MCM-41@Fe3O4 powder was used directly and placed
in a gas chamber. Pure N2 and pure O2 were mixed with different
concentrations via gas flow controls and directly poured into a
gas chamber. Sensing performance was discussed based on Ru-
Phen@MCM-41@Fe3O4 steady emission intensity quenching.
2. Experimental
2.1. General information for reagents and equipments
Starting chemicals for synthesis are summarized below. Com-
mon compounds, such as 2,20-bipyridine (bpy, AR), 1,10-phenan-
throline (Phen, AR), 4-hydroxybenzaldehyde, tetraethoxysilane
(TEOS, AR), 3-(triethoxysilyl) propylisocyanate (TESPIC, AR), odium
dodecyl sulfate (SDS, AR) and cetyltrimethylammonium bromide
(CTAB, AR), were commercially supplied by Yongxing Chemicals
and Reagents Company (Hebei, China) and used with no further
purifications. Inorganic reagents and organic solvents, such as
RuCl3ꢀnH2O (AR), NH4AC, FeCl3 (AR), HAc, concentrated ammonia,
concentrated HCl, CH2Cl2, CHCl3, n-hexane, dimethyl formamide
(DMF), ethanol, glycol, were commercially supplied by Yongjia
Chemicals and Reagents Company (Hebei, China). Organic solvents
were purified through standard procedures before use. Solvent
water was deionized. Starting compound Phen-O was obtained fol-
lowing a literature method [11,12].
2.2. Synthesis of Phen-Si and Ru(bpy)2Cl2
Silane modified ligand Phen-Si was synthesized according to a
literature procedure [11,12]. Firstly, Phen-O (20 mmol), 4-hydrox-
ybenzaldehyde (20 mmol), NH4AC (15.4 g) and HAc (35 mL) were
mixed together and heated at 100 °C. 8 h later, this solution was
poured into cold water (200 mL) and extracted with CH2Cl2
(25 mL ꢂ 3). The obtained crude product was recrystallized in
EtOH/H2O (V:V = 1:1), giving Phen-OH as yellow bulk. Yield 79%.
1H NMR (CDCl3, 300 MHz) d [ppm]: 11.60 (s, 1H), 9.12 (m, 2H),
8.98 (m, 1H), 8.59 (m, 1H), 7.71 (m, 4H), 7.50 (m, 3H). 13C NMR
(CDCl3), d (ppm): 110.4, 114.7, 120.1, 121.5, 123.6, 124.8, 127.5,
131.1, 134.7, 136.9, 144.3, 150.4, 152.9, 157.5. MS m/z: [m+1]+ calc.
for C19H12N4O, 312.1; found, 313.2.
Equipments for characterization are summarized below. 1H
NMR, IR and MS spectra were recorded from a Varian INOVA 300
spectrometer,
a Bruker Vertex 70 FTIR spectrometer (400–
4000 cmꢁ1, KBr pellet technique) and a Agilent 1100 MS series/
AXIMA CFR MALDI/TOF MS spectrometer, respectively. A Vario
Element analyzer was used to finish elemental analysis. Emission
spectra and lifetime were obtained from
a
Hitachi F-4500
Then Phen-OH (10 mmol) was dispersed in TESPIC (20 mL) and
exposed to ultrasonic bath for 25 min. This mixture was heated at
80 °C under N2 protection for 3 days and then poured into cold
n-hexane (0 °C, 200 mL). The resulting crude product was recrys-
tallized in ethanol to give Phen-Si as pale yellow powder. Yield
21%. 1H NMR (DMSO-d6, 300 MHz) d [ppm]: 0.51–0.54 (t, 2H),
1.14–1.17 (t, 9H), 1.51–1.54 (m, 2H), 3.01–3.04 (m, 2H),
3.75–3.78 (q, 6H), 7.38 (d, 2H), 7.83–7.86 (m, 2H), 8.27–8.29 (d,
2H), 8.95 (dd, 2H), 9.09 (dd, 2H), 11.75 (NH). 13C NMR (CDCl3),
fluorescence spectrophotometer and a two-channel TEKTRONIX
TDS-3052 oscilloscope excited by 355 nm light (third-harmonic-
generator pump, Nd:YAG laser), respectively. XRD measurement
was finished on
a Rigaku D/Max-Ra X-ray diffractometer
(k = 1.5418 Å). Magnetic response was analyzed on a MPM5-XL-5
superconducting quantum interference device. Thermal degrada-
tion analysis was recorded on a Perkin-Elmer thermal analyzer.
Sample morphology was taken from a Hitachi S-4800 microscope

26
L. Chen et al. / Inorganica Chimica Acta 446 (2016) 24–31
d (ppm): 11.4, 14.7, 22.0, 41.3, 52.6, 114.8, 120.1, 121.4, 122.2, 123.7,
124.8, 127.7, 128.5, 130.1, 133.6, 142.5, 150.1, 151.7, 152.9, 158.5.
MS m/z: [m+1]+ calc. for C29H33N5O5Si, 559.2; found, 560.4.
Ru(bpy)2Cl2 was obtained using a literature procedure [11,12].
Below chemicals were added into a flask and heated at 110 °C for
a whole day under N2 atmosphere, including RuCl3ꢀnH2O (5 mmol),
bpy (11 mmol) and redistilled DMF (35 mL). Solvent was then
extracted by rotary evaporation. Solid residue was dispersed in
acetone (70 mL). The remaining solid sample was mixed with etha-
nol (100 mL) and water (100 mL) and heated at 85 °C for 5 h under
N2 atmosphere. Later, anhydrous LiCl (100 g) was added under stir-
ring. Ethanol was extracted by rotary evaporation. The remaining
solution was cooled in refrigerator for 10 h. Crude product was
recrystallized in mixed solvent ethanol:water (V:V = 1:1) and dried
in vacuum at 120 °C for 2 days. MS m/z: [m]+ calc. for C20H16N4-
RuCl2, 484.0; found, 484.4.
to ꢃ260 nm. Its surface is obviously smoothed, but aggregation
between SiO2@Fe3O4 particles is still obvious, which means that
this thin SiO2 layer is effective on modifying particle surface but
limited in decreasing magnetic attraction. After MCM-41 construc-
tion and probe loading procedures, diameter of Ru-Phen@MCM-
41@Fe3O4 is finally increased to 370 nm, with smooth surface
and nearly monodispersal. Its TEM image shown in Fig. 1 indicates
a clear core–shell structure in it. It is thus confirmed that MCM-41
layer has successfully blocked magnetic aggregation between these
particles, showing a good dispersal. MCM-41 layer thickness is
determined as ꢃ55 nm which is slightly smaller than literature
values [17,18]. We assume that these short MCM-41 tunnels may
favor oxygen sensing by decreasing the number of ‘‘dead-sites”
which are inaccessible to O2 diffusion, giving improved sensitivity
and fast response.
3.2. Superamagnetic feature of Ru-Phen@MCM-41@Fe3O4
2.3. Construction of SiO2@Fe3O4
In our target structure, magnetic core is designed for site-speci-
fic aggregation. Magnetic response of Ru-Phen@MCM-41@Fe3O4 is
thus analyzed to evaluate its potential for site-specific aggregation.
That of Fe3O4 core is shown in Fig. 2 for comparison. It is observed
that our as-synthesized Fe3O4 core follows superamagnetic behav-
ior, showing no hysteresis. Such superamagnetic nature enables it
either to be aggregated to a specific-site in the presence of a mag-
net or to be highly dispersed in the absence of external magnetic
field. Saturate magnetization value of our Fe3O4 core is measured
as 67.1 emu gꢁ1. This value is slightly lower than literature values
which can be explained by the small size of our Fe3O4 particles
[17,18]. Literatures have suggested that Fe3O4 particles larger than
30 nm are supposed to follow magnetic behavior instead of supera-
magnetic one [19]. In this work, however, superamagnetic behav-
ior is still observed from our as-synthesized Fe3O4 particles even
though their diameter is as wide as 250 nm. Considering the bulges
and graves on their surface, we assume that each visible Fe3O4 par-
ticle is actually composed of sub-particles smaller than 30 nm. As
for Ru-Phen@MCM-41@Fe3O4, silica encapsulation and MCM-41
growth decrease its saturate magnetization value to 51.3 emu gꢁ1
with its superamagnetic behavior well preserved. This decreased
superamagnetic behavior is still strong enough for site-specific
aggregation [17,18].
Precursor for magnetic supporting matrix SiO2@Fe3O4 was con-
structed following below procedure [17,18]. Glycol (15 mL), SDS
(0.2 g), NaAc (1.5 g) and FeCl3ꢀ6H2O (0.5 g) were mixed together
and stirred at room temperature for 30 min. This mixture was
sealed into a Teflon reaction kettle and heated at 200 °C for 12 h.
The resulting solid sample was washed with deionized water, re-
dispersed in ethanol (40 mL) and exposed to ultrasonic bath for
30 min. During this time, deionized water (40 mL), TEOS (0.5 g)
and NH3ꢀH2O (2 mL) were slowly added. This mixture was allowed
to react at room temperature for 6 h, giving SiO2@Fe3O4.
2.4. Construction of Ru-Phen@MCM-41@Fe3O4 and a reference sample
Our site-specific oxygen sensing composite (Ru-Phen@MCM-
41@Fe3O4) was constructed following below procedure. First,
MCM-41 was grew onto SiO2@Fe3O4 surface with TEOS and
Phen-Si as silica source. SiO2@Fe3O4 (0.5 g), TEOS (1.5 g), Phen-Si
(0.12 g), CTAB (0.5 g), deionized water (150 mL) and NH3ꢀH2O
(3 mL) were mixed together and allowed to react at room temper-
ature for 6 h. The resulting solid product was collected and stirred
in ethanol (200 mL) and concentrated HCl (10 mL) for 2 days to
remove template reagent CTAB. Then this product was mixed with
Ru(bpy)2Cl2 (0.5 g, excess) and ethanol (50 mL), and allowed to
react at 80 °C for 8 h. The final solid product was collected, washed
with ethanol and dried in vacuum to give Ru-Phen@MCM-
41@Fe3O4 as pale dark powder. Yield (0.4 g). Elemental analysis
for Ru-Phen@MCM-41@Fe3O4, found: C, 6.28, H, 1.22, N, 1.07%.
A similar procedure was carried out except that no Phen-Si was
used in this run, giving a reference sample, so that performance
comparison between this reference sample and Ru-Phen@MCM-
41@Fe3O4 could be performed.
3.3. XRD pattern of Ru-Phen@MCM-41@Fe3O4
The Fe3O4 core in Ru-Phen@MCM-41@Fe3O4 is further analyzed
and confirmed by its wide angle XRD (WAXRD) pattern, as shown
in Fig. 3A. That of our as-synthesized Fe3O4 particles is shown for
comparison. Six well-resolved diffraction peaks are observed for
our as-synthesized Fe3O4 particles, which are indexed as (220),
(311), (400), (422), (551), (440), respectively. After consulting
literature reports, we come to a conclusion that Fe3O4 core has
been successfully constructed [17,18]. Diameter of these Fe3O4 par-
ticles is calculated as 14.78 nm using Scherrer equation, which
confirms our hypothesis that each visible Fe3O4 particle is actually
composed of sub-particles smaller than 30 nm [17,18]. As for
Ru-Phen@MCM-41@Fe3O4, these six diffraction peaks are still
observed, which means that the Fe3O4 core has been well pre-
served after a sires procedures of silica encapsulation, MCM-41
growth and probe loading. On the other hand, their diffraction
intensity is slightly decreased, which should be explained by the
fact that above modification procedures actually decrease regular-
ity of Fe3O4 core.
3. Results and discussion
3.1. Morphology of Ru-Phen@MCM-41@Fe3O4
Scanning electron microscope (SEM) and transmission electron
microscope (TEM) images of Ru-Phen@MCM-41@Fe3O4 are shown
in Fig. 1 to get a visual understanding on its morphology. SEM
images of Fe3O4 core and SiO2@Fe3O4 are shown in Fig. 1 as well
for comparison. Owing to its magnetic nature, Fe3O4 particles are
intensively aggregated with bad dispersal. These particles are gen-
erally spherical ones with mean diameter of ꢃ250 nm which is
slightly smaller than literature values [17,18]. Their surface, how-
ever, is rather rough with multiple bulges and graves. As for
SiO2@Fe3O4, silica encapsulation procedure increases its diameter
For a tentative investigation on the mesoporous tunnels on
Ru-Phen@MCM-41@Fe3O4 surface, its small angle XRD (SAXRD)
pattern is measured and shown as Fig. 3B. That of a reference sam-
ple is shown for comparison. The reference sample exhibits a sharp

L. Chen et al. / Inorganica Chimica Acta 446 (2016) 24–31
27
Fig. 1. SEM (a) and TEM (b) images of Ru-Phen@MCM-41@Fe3O4, along with SEM images of Fe3O4 core (c) and SiO2@Fe3O4 (d).
Fig. 3A. WAXRD patterns of Ru-Phen@MCM-41@Fe3O4 and Fe3O4 core.
Fig. 2. Magnetic response of Ru-Phen@MCM-41@Fe3O4 and Fe3O4 core.
Bragg reflection peak and two obvious shoulder peaks indexed as
(100), (110) and (200) which are quite similar to those of stan-
dard MCM-41 samples [11,12]. As for Ru-Phen@MCM-41@Fe3O4,
there are three similar peaks, suggesting that there are highly
ordered hexagonal tunnels on Ru-Phen@MCM-41@Fe3O4 surface.
On the other hand, these peaks become wider and weaker than
those of the reference sample. It seems that the existence of silane
coupling ligand and its Ru(II) complex in Ru-Phen@MCM-
41@Fe3O4 slightly decreases regularity of this mesoporous
structure.
3.4. N2 adsorption/desorption measurement of Ru-Phen@MCM-
41@Fe3O4
To further confirm the mesoporous tunnels on Ru-Phen@MCM-
41@Fe3O4 surface, its N2 adsorption/desorption isotherms are
Fig. 3B. SAXRD patterns of Ru-Phen@MCM-41@Fe3O4 and a reference sample.

28
L. Chen et al. / Inorganica Chimica Acta 446 (2016) 24–31
Si–O bonds. These characteristic bands are consistent with silica
shell and MCM-41 layer on this reference sample. As for Phen-Si,
plane blending and stretching vibrations from Si–O bonds
(583 cmꢁ1, 802 cmꢁ1 and 1637 cmꢁ1) are observed from its IR
spectrum, without vibration of Si–O–Si framework. The sharp
bands peaking at 1543 cmꢁ1 and 1700 cmꢁ1 belong to vibrations
of –Si–O–C group. A group of bands around 2973 cmꢁ1 are assigned
to stretching vibrations of –(–CH2–)3– group [11,12]. Above bands
are consistent with Phen-Si molecular structure. As for the IR spec-
trum of Ru-Phen@MCM-41@Fe3O4, characteristic bands from Si–
O–Si framework and Si–O bonds (458 cmꢁ1, 583 cmꢁ1, 802 cmꢁ1
and 1637 cmꢁ1) are clearly observed. Stretching vibration of
–(–CH2–)3– group (2931 cmꢁ1) exhibits blue shift tendency com-
pared to that of Phen-Si, which should be attributed to the covalent
grafting with silica backbone. Vibrations of –Si–O–C group
(1543 cmꢁ1 and 1700 cmꢁ1) are completely missing, owing to
Phen-Si hydrolysis. Taking above result into account, it is safe to
say that our Ru(II)-based probe has been covalently immobilized
in Ru-Phen@MCM-41@Fe3O4 through our silane coupling ligand,
serving as sensing probe.
To get a tentative understanding on probe loading content in
Ru-Phen@MCM-41@Fe3O4, its thermogravimetry analysis (TGA)
and derivative thermogravimetry (DTG) plots are shown in Fig. 6.
It is observed that Ru-Phen@MCM-41@Fe3O4 is quite stable below
200 °C, showing no obvious weight loss. It is thus confirmed that
this composite is thermally stable enough for normal applications.
Upon even higher temperatures, there are two major weight loss
regions, ranging from 200 °C to 340 °C and from 407 °C to 522 °C,
respectively. The first weight loss region is responsible for
11.4 wt% weight loss with multiple endothermic peaks around
240 °C. Combined with TGA curve of a similar Ru(II) complex
[Ru(bpy)2(Phen-OH)]Cl2 (Fig. S1, Supporting information), this
region is attributed to the thermal release of Ru(II)-probe. In other
words, probe loading content in Ru-Phen@MCM-41@Fe3O4 is as
high as 11.4 wt%. This value is slightly higher than literature
values, given our short MCM-41 tunnels [17,18]. It seems that
these short tunnels are more efficient on loading probe than long
ones do. The last weight loss region causes 6.9 wt% weight loss
with an endothermic peak of 441 °C. We attribute this region to
the thermal degradation of organosilicate framework in
Ru-Phen@MCM-41@Fe3O4.
Fig. 4. N2 adsorption/desorption isotherms of Ru-Phen@MCM-41@Fe3O4 and a
reference sample.
shown in Fig. 4. Those of a reference sample are shown for compar-
ison. It is observed that the two samples have nearly identical iso-
therms which are quite similar to those of standard MCM-41
samples [11,12]. These type IV isotherms suggest that hexangular
tunnels have been successfully constructed on Ru-Phen@MCM-
41@Fe3O4 surface and well preserved after loading probe mole-
cules. Mesoporous parameters, including pore diameter, surface
area and pore volume, are determined as 2.05 nm, 317.18 m2 gꢁ1
and 0.206 cm3 gꢁ1
, respectively, for Ru-Phen@MCM-41@Fe3O4.
Corresponding values of the reference sample containing no probe
are 2.74 nm, 713.64 m2 gꢁ1 and 0.389 cm3 gꢁ1, respectively. It is
obvious that mesoporous parameters are all decreased in
Ru-Phen@MCM-41@Fe3O4 after loading probe molecules, which
tentatively confirms that our Ru(II)-based probe has been success-
fully immobilized in MCM-41 tunnels.
3.5. IR spectrum and thermogravimetry analysis of Ru-Phen@MCM-
41@Fe3O4
Aiming at a confirmation on the covalent immobilization of
probe in Ru-Phen@MCM-41@Fe3O4, IR spectra of Phen-Si and
Ru-Phen@MCM-41@Fe3O4 are shown in Fig. 5. That of a reference
sample is shown for comparison. The reference sample containing
no probe demonstrates only a few characteristic bands peaking at
458 cmꢁ1, 583 cmꢁ1, 802 cmꢁ1 and 1637 cmꢁ1, respectively. The
first one is attributed to vibration of Si–O–Si framework. The latter
three ones belong to plane blending and stretching vibrations from
3.6. Performance evaluation of site-specific O2 sensing
3.6.1. Site-specific aggregation
Site-specific aggregation of Ru-Phen@MCM-41@Fe3O4 should
be firstly evaluated since it is our primary objective. In virtue of
Fig. 5. IR spectra of Ru-Phen@MCM-41@Fe3O4, Phen-Si and a reference sample.
Fig. 6. TGA and DTG analysis of Ru-Phen@MCM-41@Fe3O4.

L. Chen et al. / Inorganica Chimica Acta 446 (2016) 24–31
29
Fig. 7. Photos of Ru-Phen@MCM-41@Fe3O4 turbid liquid in ethanol ((a) no magnet, no UV; (b) no magnet, UV 365 nm; (c) magnet, no UV; (d) magnet, UV 365 nm).
its hydrophilic silica shell, this composite can be readily dispersed
in aqueous solution when there is no external magnetic field, as
shown in Fig. 7. Upon photoexcitation of 365 nm, uniform red
emission is observed from the whole turbid liquid, which is consis-
tent with MLCT emission of Ru(II) complexes [11,12]. On the other
hand, owing to its superamagnetic nature, Ru-Phen@MCM-
41@Fe3O4 can be readily aggregated to a specific site by a magnet.
Correspondingly, its red emission is concentrated at this site. We
come to a conclusion that our Ru(II)-based probe has been uni-
formly immobilized and well preserved in MCM-41 matrix, which
is positive for linear oxygen sensing. It is thus safe to say that our
primary objective of site-specific aggregation has been achieved.
To evaluate its sensing performance, maximum sensitivity is
defined as the value of I0/I100, according to a literature quotation.
Here I0 stands for emission intensity in the absence of O2 and I100
is that in pure O2, respectively [11,12]. Maximum sensitivity of
Ru-Phen@MCM-41@Fe3O4 is calculated as 11.5 which is much
higher than literature values (ꢃ5) of similar magnetic-mesoporous
sensing composite samples [20–23]. We attribute this good sensi-
tivity to the following factors. (1) There are large coplanar conjuga-
tion planes in our Ru(II) complex which may increase electronic
distribution and lifetime of probe excited electrons. Correspond-
ingly, sensing collision probability between excited probe and ana-
lyte should be increased. (2) O2 molecules are efficiently
transported in MCM-41 tunnels with high diffusion coefficient,
favoring oxygen sensing.
3.6.2. Sensing sensitivity
By recording spectral response of Ru-Phen@MCM-41@Fe3O4
emission under various O2 concentrations ranging from 0% to
100%, its oxygen sensing performance can be evaluated. As shown
in Fig. 8, Ru-Phen@MCM-41@Fe3O4 has a broad emission band
under pure N2 condition peaking at 591 nm with FWHM of
78 nm which is consistent with its red emission shown in Fig. 7.
Here, FWHM means full width at half maximum. No vibronic pro-
gressions are found, suggesting that corresponding emissive center
has a charge transfer character. This finding is consistent with the
MLCT emissive nature of Ru(II) complex. Emission intensity of
Ru-Phen@MCM-41@Fe3O4 gradually decreases with increasing
O2 concentrations with band shape well preserved, which means
that its emissive center does not change. It is hereby confirmed
that Ru-Phen@MCM-41@Fe3O4 emission is quenchable by O2,
making itself an oxygen sensing material.
3.6.3. Dynamic sensing mechanism and working plot
To get a further understanding on the energy transfer between
our probe and O2 molecules and then specify its sensing mecha-
nism, emission decay dynamics of Ru-Phen@MCM-41@Fe3O4
under pure N2 and pure O2 conditions are recorded and shown as
the inset of Fig. 8. It is observed that both emission decay curves
follow single exponential decay pattern. In the absence of O2, its
lifetime is as long as 1.30 ls, indicating the phosphorescent nature
of this emission. This long-lived emissive center is consistent with
triplet MLCT emissive center of Ru(II) complexes, further confirm-
ing that this emission comes from Ru(II)-based probe [11,12]. Its
single exponential decay pattern confirms that these probe mole-
cules are homogeneously immobilized in supporting matrix, which
is consistent with our observation from Fig. 7. In the presence of
O2, Ru-Phen@MCM-41@Fe3O4 emission lifetime is greatly
decreased to 0.06 ls under pure O2 condition and 0.29 ls under
air condition, respectively, suggesting that its emissive center is
quenched by O2 molecules. We thus come to a conclusion that
Ru-Phen@MCM-41@Fe3O4 emissive center is directly quenched
by O2 molecules through collision-induced energy transfer. This
procedure falls in the definition of a dynamic quenching mecha-
nism, as described by Formula 1. Here ‘‘⁄” means an excited state.
ꢄ
ꢄ
3
1
RuðIIÞ-probe þ O2 ! RuðIIÞ-probe þ O2
ð1Þ
Above analysis has confirmed that: (1) all probe molecules are
homogeneously distributed in supporting matrix; (2) probe emis-
sion follows single exponential decay pattern; (3) probe excited
state is quenched by a dynamic energy transfer. In this case, probe
emission intensity variation upon various analyte concentrations
can be expressed by Stern–Volmer equation, as shown by Formula
2 [11,12]. Here I and I0 stand for emission intensity and that in pure
N2 atmosphere, respectively. KSV and [O2] are Stern–Volmer con-
stant and O2 concentration, respectively. According to Formula 2,
an ideal plot of I0/I against [O2] is supposed to be a linear curve with
slope of KSV. The I0/I versus [O2] plot of Ru-Phen@MCM-41@Fe3O4
Fig. 8. Spectral response of Ru-Phen@MCM-41@Fe3O4 emission under various O2
concentrations ranging from 0% to 100% with interval of 10%. Inset: emission decay
dynamics of Ru-Phen@MCM-41@Fe3O4 under pure N2 and pure O2 conditions.

30
L. Chen et al. / Inorganica Chimica Acta 446 (2016) 24–31
is quickly recovered to normal level. After a few such cycles,
maximum emission can always be recovered, indicating that
Ru-Phen@MCM-41@Fe3O4 has a good photostability owing to the
covalent grafting between probe and supporting matrix.
For discussion convenience, we define response time as the
time taken by Ru-Phen@MCM-41@Fe3O4 to decrease to 5% of its
emission maximum when surrounding environment is switched
from pure N2 to pure O2 [11,12]. Similarly, recovery time is defined
as the time taken by Ru-Phen@MCM-41@Fe3O4 to increase to 95%
of its emission maximum when surrounding environment is
switched from pure O2 to pure N2. Response time of Ru-
Phen@MCM-41@Fe3O4 is calculated as 10 s which is comparable
to or even shorter than literature values (12–21 s) of similar mag-
netic-mesoporous sensing composite samples [20–23]. The follow-
ing factors should be blamed responsible. (1) The large coplanar
conjugation planes in probe increase electronic distribution and
lifetime of probe excited electrons. By offering more sensing colli-
sion chances, excited probe can be efficiently quenched. (2) Sup-
porting matrix provides highly ordered tunnels with high
diffusion coefficient, guaranteeing efficient oxygen sensing. On
the other hand, recovery time is calculated as 30 s and found much
longer than response time. This phenomenon has been blamed to
diffusion-controlled dynamic response and recovery behavior
[24]. Diffusion coefficients of N2 and O2 in our composite are mea-
sured as 0.61 ꢂ 10ꢁ8 m2/s and 1.72 ꢂ 10ꢁ8 m2/s, respectively. This
result is consistent with our observation that recovery time is
longer than response time. Mills and coworkers have reported that
response time and recovery time can be calculated by Formulas 4
Fig. 9. Stern–Volmer working plot of Ru-Phen@MCM-41@Fe3O4 emission under
various O2 concentrations ranging from 0% to 100% with interval of 10%. Inset:
emission intensity monitoring of Ru-Phen@MCM-41@Fe3O4 upon periodically
changing environment atmosphere between 100% N2 and 100% O2.
is close to a linear one, as shown in Fig. 9. However, a more com-
prehensive model should be proposed to perfectly explain this
sensing plot.
I0=I ¼ 1 þ KSV½O2ꢅ
ð2Þ
Considering that there may be ‘‘dead-sites” in Ru-Phen@MCM-
41@Fe3O4 which are inaccessible to O2 diffusion, it is rational to
assume that there are two or more kinds of sensing sites. Only
one of them is sensitive to and quenchable by O2 molecules, while
the others are not. In this case, sensing contribution of each site
should be taken into account. A modified two-site model is then
proposed to describe this case, as shown by Formula 3. Here f1
and f2 stand for fractional contributions of sensing sites
(f1 + f2 = 1), KSV1 and KSV2 mean Stern–Volmer quenching constants
of these sensing sites, respectively [11,12]. This modified two-site
model can well describe our working plot, as shown in Fig. 9, with
f1 = 0.953, f2 = 0.047, KSV1 = 0.1917 [O2%]ꢁ1 and KSV2 = 0.0022
[O2%]ꢁ1 (R2 = 0.999), respectively. It is observed that KSV2 is greatly
smaller than KSV1 by two orders of magnitude, indicating that site-
2 is nearly immune to O2 molecules. If we set KSV2 as 0, then For-
mula 3 can be further simplified, with f1 = 0.964, f2 = 0.036, and
KSV1 = 0.1800 [O2%]ꢁ1 (R2 = 0.999), respectively. This simplified
model can well fit our working plot. Maximum sensitivity and lin-
earity of working plot are compromised by site-2, even though its
contribution is slim. Site-2 is attributed to the ‘‘dead-sites” in
Ru-Phen@MCM-41@Fe3O4 which are inaccessible to O2 diffusion.
Its slim contribution suggests that most MCM-41 tunnels can
efficiently transport O2 molecules, as anticipated in Section 3.2.
For performance improvement, such ‘‘dead-sites” should be elimi-
nated so that a more linear working plot can be expected.
and 5, which mathematically explains why a recovery time (Trec
)
is always longer than a response time (Tres) [25]. Here, b is sample
thickness, D is diffusion coefficient, KSV is Stern–Volmer constant,
respectively.
!
ꢀ
ꢁ
b2
D
9KSVfPO2ðfinalÞ ꢁ PO2ðinitialÞg
1 þ PO2ðfinalÞKSV
Tres ¼ 3:06
Trec ¼ 3:06
Ln 10 ꢁ
ð4Þ
ð5Þ
!
ꢀ
ꢁ
b2
D
9KSVfPO2ðfinalÞ ꢁ PO2ðinitialÞg
1 þ PO2ðfinalÞKSV
Ln 10 þ
:
3.6.5. Selectivity: interfering effect from other gases
For a tentative evaluation on selectivity of Ru-Phen@MCM-
41@Fe3O4 towards O2, its emission spectra under pure N2, CO2
and a few VOC gases, including benzene, toluene, CHCl3 and CH2Cl2
vapors, are shown in Fig. 10. No obvious interfering effect is
I0
I
1
¼
ð3Þ
f1
f2
þ
1þKSV1pO2
1þKSV2pO2
3.6.4. Response/recovery character and photostability
Aiming at a further confirmation on the dependence between
emission quenching and O2 presence, surrounding environment
is periodically changed between pure O2 and pure N2 when
Ru-Phen@MCM-41@Fe3O4 emission is continuously monitored.
As shown by the inset of Fig. 9, sample emission is clearly depen-
dent on O2 presence. Under pure O2 condition, sample emission is
decreased to minimal level and maintained, showing sensing sig-
nal. When atmosphere is switched to pure N2 condition, emission
Fig. 10. Emission spectra of Ru-Phen@MCM-41@Fe3O4 under the following condi-
tions: N2 (101 kPa), CO2 (101 kPa), benzene (13.33 kPa), toluene (4.89 kPa), CHCl3
(21.2 kPa) and CH2Cl2 (46.5 kPa).

L. Chen et al. / Inorganica Chimica Acta 446 (2016) 24–31
31
observed for CO2. This is because CO2 has a closed-shell structure
Appendix A. Supplementary material
and is not open for energy transfer from excited Ru(II) center
[11,12]. The presence of benzene and toluene vapors, however,
tends to quench Ru-Phen@MCM-41@Fe3O4 emission. It is assumed
that the conjugation chain in these organic molecules may capture
excited electrons and thus quench Ru-Phen@MCM-41@Fe3O4
emission. As for CHCl3 and CH2Cl2 vapors, they may quench
Ru-Phen@MCM-41@Fe3O4 emission by coordinating with excited
Ru(II) center. For performance improvement, selectivity of our
sensing probe should be taken into account.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.ica.2016.02.061.
References
[1] Y. Guo, L. Zhang, S. Zhang, Y. Yang, X. Chen, M. Zhang, Biosens. Bioelectron. 63
(2015) 61.
[2] D. Sarkar, A. Pramanik, S. Jana, P. Karmakar, T.K. Mondal, Sens. Actuators, B 209
(2015) 138.
[3] O.S. Wolfbeis, Anal. Chem. 80 (2008) 4269.
[4] C. Baleizao, S. Nagl, M. Schaferling, M.N. Berberan-Santos, O.S. Wolfbeis, Anal.
Chem. 80 (2008) 6449.
4. Conclusion
[5] L.C. Clark Jr., Trans. Am. Soc. Artif. Intern. Organs 2 (1956) 41.
[6] L.W. Winkler, Ber. Dtsch. Chem. Ges. 21 (1888) 2843.
[7] M.R. Awual, M.M. Hasan, M.A. Khaleque, Sens. Actuators, B 209 (2015) 194.
[8] C.S. Chu, C.Y. Chuang, Sens. Actuators, B 209 (2015) 94.
[9] G. Zhang, J. Chen, S.J. Payne, S.E. Kooi, J.N. Demas, C.L. Fraser, J. Am. Chem. Soc.
129 (2007) 15728.
[10] A. Gulino, S. Giuffrida, P. Mineo, M. Purrazzo, E. Scamporrino, G. Ventimiglia,
M.E. Van der Boom, I. Fragala, J. Phys. Chem. B 110 (2006) 16781.
[11] B. Lei, B. Li, H. Zhang, L. Zhang, W. Li, J. Phys. Chem. C 111 (2007) 11291.
[12] B. Lei, B. Li, H. Zhang, S. Lu, Z. Zheng, W. Li, Y. Wang, Adv. Funct. Mater. 16
(2006) 1883.
As a conclusion, we designed and constructed an oxygen sens-
ing composite with site-specific aggregation feature. Fe3O4 core
coated with silica molecular sieve MCM-41 was used supporting
matrix. A Ru(II)-based probe was covalently immobilized into this
supporting matrix through a silane coupling ligand. This composite
structure was analyzed and discussed by means of SEM, TEM, XRD,
IR, TGA and N2 adsorption/desorption. Photophysical analysis of
this composite under various O2 concentrations suggested that
its emission was quenchable by O2 with good stability through a
dynamic mechanism. Sensitivity of 11.5 and short response time
of 10 s were obtained. This good performance was attributed to
the following factors: (1) the large coplanar conjugation planes in
probe increased electronic distribution and lifetime of probe
excited electrons. By offering more sensing collision chances,
excited probe could be efficiently quenched. (2) supporting matrix
provided the highly ordered tunnels with high diffusion coefficient,
guaranteeing efficient oxygen sensing. On the other hand, linearity
of our working plot was yet to be satisfied. ‘‘Dead-sites” should be
eliminated so that a linear working plot can be expected.
[13] L. Yang, J.K. Feng, A.M. Ren, M. Zhang, Y.G. Ma, X.D. Liu, Eur. J. Inorg. Chem. 10
(2005) 1867.
[14] J. Jia, Y. Tian, Z. Li, Synth. Met. 161 (2011) 1377.
[15] Z. Si, X. Li, X. Li, H. Zhang, J. Organomet. Chem. 694 (2009) 3742.
[16] L. Zhang, S. Yue, B. Li, D. Fan, Inorg. Chim. Acta 384 (2012) 225.
[17] Y. Wang, B. Li, L. Zhang, H. Song, Langmuir 29 (2013) 1273.
[18] Y. Wang, B. Li, L. Zhang, P. Li, L. Wang, J. Zhang, Langmuir 28 (2012) 1657.
[19] L. Zhang, W. Sun, P. Cheng, Molecules 8 (2003) 207.
[20] J. Li, C. Yang, Y. Wu, B. Wang, W. Sun, T. Shao, Inorg. Chim. Acta 441 (2016) 1.
[21] D. Wan, Y. Li, J. Liu, Inorg. Chim. Acta 437 (2015) 120.
[22] P. Wan, L. Wang, Inorg. Chim. Acta 436 (2015) 45.
[23] J. Li, C. Yang, Y. Wu, B. Wang, W. Sun, T. Shao, Inorg. Chim. Acta 442 (2016) 111.
[24] S.M. Kuang, D.G. Cuttell, D.R. McMillin, P.E. Fanwick, R.A. Walton, Inorg. Chem.
41 (2002) 3313.
[25] A. Mills, A. Pepre, Anal. Chem. 69 (1997) 4653.
Acknowledgment
The authors gratefully thank our university for the equipments
and financial support.