A rationally designed fluorescence chemosensor for on-site monitoring of carbon monoxide in air

Article abstract of DOI:10.1021/ol5012949

A fluorescence chemsensor for carbon monoxide (CO), based on transformation of weakly fluorescent iodide to strong fluorescent amino product upon reacting with CO, shows abilities of quantitative measurement of CO in air at a level of 50-1000 ppm and real-time and on-site monitoring for CO flammation/explosion.

Full text of DOI:10.1021/ol5012949

Letter
pubs.acs.org/OrgLett
A Rationally Designed Fluorescence Chemosensor for On-Site
Monitoring of Carbon Monoxide in Air
Tengfei Yan,† Jin Chen,† Song Wu,*,† Zhiqiang Mao,‡ and Zhihong Liu*,‡
†School of Pharmaceutical Sciences and ‡Key Laboratory of Analytical Chemistry for Biology and Medicine (Ministry of Education),
College of Chemistry and Molecular Sciences, Wuhan University, Wuhan 430072, China
S
* Supporting Information
ABSTRACT: A fluorescence chemsensor for carbon monoxide (CO), based
on transformation of weakly fluorescent iodide to strong fluorescent amino
product upon reacting with CO, shows abilities of quantitative measurement of
CO in air at a level of 50−1000 ppm and real-time and on-site monitoring for
CO flammation/explosion.
As a highly sensitive and potential on-site detection method,
arbon monoxide is called a silent killer because it is
Ccolorless, odorless, and tasteless but toxic, and partic- a fluorescence-based sensor could be a competent alternative
for CO sensing. So far there has been only one report, recently
contributed by Chang’s group, on a palladium-mediated
fluorescent chemosensor for CO imaging in live cells.6 Herein,
we report a substitution-based chemosensor with much easier
synthesis and greater turn-on response as compared to the
existing one, which is the first fluorescence sensing system
suitable for CO detection in ambient conditions and can be
applicable in on-site warning of CO poisoning or explosion.
The design of sensors for a certain target depends on the
nature of the target. Apart from its reliable metal-coordianation
reactivity, typical reactions of CO such as carbonylation and
hydrocarboxylation always need strong reaction conditions
such as high temperature and/or high pressure.7 This may
explain why development of a CO fluorescence probe working
under ambient conditions still remains a great challenge.
Recently, Grushin et al. reported their elegant work on the
azidocarbonylation of aryl iodides by CO with sodium azide in
the presence of a palladium catalyst under very mild
conditions.8 The product, aroyl azide, may further undergo a
Curtius rearrangement and hydrolysis to form an amino
product in water. This gives us a hint of the possibility to design
a fluorometric CO sensing system by simple modification of the
iodide substrate. Since an iodine atom attached to an aryl ring
tends to quench fluorescence due to the heavy atom effect,
replacement of the iodide by an amino group in the reaction
may result in an enhancement of fluorescence. On the basis of
this speculation, we designed a naphthalene iodide with a
sulfonic group appended peripherally (compound 2 in Scheme
ularly, it is hard to sense. CO commonly originates from
incomplete combustion of carbon-based fuels. Exposure to high
concentrations of CO even for a short time may result in acute
CO poisoning that severely harms heart and central nervous
system and even leads to death. It is reported that more than
40,000 people seek medical attention and approximately 5000
people have died from carbon monoxide poisoning each year in
the United States.1 It is also recognized that a high
concentration of CO (over 1%) in afterdamp formed in a gas
explosion causes major casualties in mining accidents that affect
developing countries.2 Therefore, CO sensing is of great
importance to public health and safety and to social welfare.
Currently the most widely used sensing techniques for CO
are based on electrochemistry (EC)3 and metal oxide
semiconductors (MOSs).4 The EC technology uses metal
wire electrodes to react with diffused CO in an electrolyte
accompanied by the change of potential, which is proportional
to the concentration of CO. Most MOS methods utilize the
oxidation of CO to carbon dioxide (CO2) at high temperature
(200−400 °C) on surfaces of metal oxides with preadsorbed
oxygen species, thereby donating electrons and leading to a
conductance change detectable via resistrometric means.
Despite the achievements made in this field in the past years,
however, these sensing methods are energy costly and
unsuitable for use in extreme surroundings with potential
flammation and explosion. Some other methods, such as
colorimetric methods based on coordination chemistry exhibit
fast response, low limit of detection, and good color contrast,
but their poor color modulation toward different CO
concentrations still remains a problem.5
Received: May 6, 2014
Published: June 12, 2014
© 2014 American Chemical Society
3296
dx.doi.org/10.1021/ol5012949 | Org. Lett. 2014, 16, 3296−3299

Organic Letters
Letter
1). The introduction of the sulfonic group involves two aspects
of consideration. One is to enhance the solubility of the
bond into Pd0 oxidatively, two possible pathways, namely, N3 /
I− exchange followed by insertion of CO or the reverse
sequence, both lead to the formation of the intermediate 3.
Removal of the PdL gives the azidocarbonylation product 4.
Subsequent Curtius rearrangement of 4 by releasing N2 to form
isocyanate 5 and hydrolysis offers the final product 1 with
strong fluorescence. Notably, the quantum yield of the product
in water was as high as 0.767. The more than 100-fold
enhancement of the quantum yield is an outstanding level
among all the fluorescence turn-on chemosensors, which
indicates the rationality of our substitution-based design. Also
notable is that the design is straightforward, allowing a very easy
synthesis.
−
Scheme 1. (A) Interconversion of Strongly Fluorescent
Compound 1 and Weakly Fluorescent Compound 2. (B)
Proposed Mechanism for the Amination of the Iodide
Substrate 2
To seek an ideal catalyst for the substitution-based CO
sensing, we then compared the effect of the ligand of palladium
on the reaction. Commercially available Pd(PPh3)4 and
Pd(dba)2 as well as palladium/xantphos were tested under
the same CO concentration and identical reaction conditions.
The reactivity of the catalysts was compared by measuring the
fluorescence intensities of the systems at a fixed reaction time.
It was found that Pd/xantphos possesses the highest catalytic
activity, while Pd(dba)2 exhibits about two-thirds that of Pd/
xantphos, and Pd(PPh3)4 even lower (Figure S6 in Supporting
Information). The exact reason for the reactivity order of Pd/
xantphos > Pd(dba)2 > Pd(PPh3)4 is yet unclear but quite
consistent with those findings that prove xantphos a superior
ligand for a variety of palladium-catalyzed carbonylation
reactions.9 The amount of the catalyst was also found to
profoundly affect the reaction rate. Using 1000 ppm of CO as a
model gas, as shown in Figure S7 in Supporting Information,
the reaction rate was accelerated upon the increase of Pd/
xantphos and reached the maximum with 0.4 equiv of the
catalyst when NaN3 was largely excessive (250 equiv) We then
studied impact of environmental factors such as temperature
and pH values on fluorescence response to our system that a
real-world CO sensing may encounter. As shown in Figure S8
in Supporting Information, compared to that at ambient
temperature of 10 °C after reaction with 1000 ppm of CO with
the sensing system for 2 h, the fluorescence intensities at 30, 40,
and 50 °C increased by 1.61-, 2.24-, and 2.5-fold, respectively.
From a practical consideration, we tested the fluorescence
response of pH values in a range of 4 to 10. A moderate
influence could be observed by the increasing ratio of 0.5
compared the highest fluorescence intensity at pH = 8 with the
lowest intensity at pH = 4 (Figure S9 in Supporting
Information). The above observations suggest that our CO
sensing system is environment-related.
We then investigated the response of this chemosensor
toward varying concentrations of CO as a function of time. It is
known that the adverse physiological effects of CO to human
are associated with both its concentration and the exposure
time. It is generally accepted that 50 ppm is the critical
concentration for CO at which the gas becomes toxic to a
healthy adult with continuous exposure over an 8 h period,
while an exposure to 1600 ppm of CO causes serious
discomforts such as tachycardia and nausea within 20 min
and death in less than 2 h.10 We tested a wide range of CO
concentration, i.e., from 50 ppm to 1% (10,000 ppm), to check
the effective coverage of the sensing system. To simulate the
on-site CO sensing, we simply purged the gas into a 1 L flask
containing the sensing system and gently stirred it at ambient
conditions. It should be noted that the concentration of CO
denotes the in-air content. The concentrations of CO in the
molecule in water so as to facilitate its practical application, and
the other is to improve the photoluminescence property
because the electron-withdrawing sulfonic group together with
the electron-donating amino group will form a pull−push
electronic structure, which increases the polarity of the
fluorophore and thus favors its photophysical properties such
as the fluorescence quantum yield.
Compound 2 was readily synthesized by diazotization of the
sodium 4-aminonaphthalene-1-sulfonate 1 followed by iodiza-
1
tion with total yield of 29.8% and was characterized by H
NMR and 13C NMR, high-resolution MS, and HPLC (Figures
S1−S3 in Supporting Information). As expected, compound 2
showed very weak fluorescence with a quantum yield of 0.0063
at λex = 320 nm in water (see the Supporting Information for
the details of determination). In a preliminary test, a sensing
system was constructed by dissolving compound 2, sodium
azide, and palladium acetate/xantphos (2% equiv) in water with
10% DMSO. In this system, all components give very low or
zero intrinsic fluorescence, producing an extremely low
background. The fluorescence of the mixture dramatically
increased within several minutes after exposure to CO gas. The
fluorescent product was isolated and identified as compound 1
1
by means of H NMR (Figure S4 in Supporting Information),
MS (Figure S5 in Supporting Information), and HPLC (Figure
S3 in Supporting Information), confirming the occurrence of
the amination of 2 in the reaction. The mechanism of the
transformation of the iodine atom into an amino group is
demonstrated in Scheme 1B, which is proposed according to
Grushin’s reaction.8 In this system, the catalyst complex PdL (L
= xantphos) is formed in situ first. After the insertion of Ar−I
3297
dx.doi.org/10.1021/ol5012949 | Org. Lett. 2014, 16, 3296−3299

Organic Letters
Letter
reaction solution, in consideration of the vapor−liquid
equilibrium, were actually much lower than the denoted values.
The signals of the reactions are presented with fluorescence
enhancement times, (F − F0)/F0, in which F0 and F are the
fluorescence intensities of the chemosensor before and after
reacting with CO. As shown in Figure 1A, the fluorescence of
Figure 2. Photographs of sensing solutions containing 50 μM 2 after
exposure to CO in air for 10 min: (A) 0 ppm, (B) 1000 ppm;, (C)
5000 ppm, (D) 1%. All photographs were taken under 320 nm
irradiation. Ambient temperature: 10 °C.
fluorescence spectra are given in Figure S10 in Supporting
Information. The brightness of the emission was concentration-
dependent and distinguishable by eye between the different
concentrations. The fast and unambiguous response of the
sensor toward 1000 ppm of CO suggests the potential
application for CO acute poisoning warning. As for the
concentration of 1%, although it is still far less than the
explosion limit of 12.5%, a ca. 25-fold enhancement of
fluorescence intensity arose in 10 min (Figure S10 in
Supporting Information). Considering the concentration-
dependence of the reaction rate, it is safe to deduce that this
sensor would give even faster and more distinct response to
higher concentrations of CO, suggesting that the system can be
particularly useful to CO flammation/explosion monitoring.
From a perspective of realistic context application, we also tried
to use our sensing system on a solid support. By comparison we
found that paper is a good material. We also optimized the
sensing conditions in a 10% CO environment that is close to
the flamamtion/explosion lower limit (for detail of paper-based
sensing see the Supporting Information). As shown in Figure
S11 in Supporting Information, even after only 2 min of
reaction, the enhancement of fluorescence could be discerned
with the naked eye under irradiation of 254 nm wavelength.
What is more, with the extension of reaction time, the
fluorescence enhanced accordingly. Our preliminary inves-
tigation results clearly demonstrate our sensing system is
suitable not only for CO flammation/explosion monitoring in
solution but also for solid-support-based sensing.
The selectivity of the chemosensor was also investigated.
Since CO2 is often produced accompanying CO in combustion,
we first studied the interference of CO2. As shown in Figure 3,
no significant fluorescence variation was observed after long-
term exposure (6 h). Other industrial gases such as SO2, NO,
and NO2 also gave the similar results. This is superior to the
sensors employing the coordinating property of CO, because
NO and NO2 at high concentration in some cases may take the
place of the coordination center and interfere with the
sensing.5a,b Also, some commonly used volatile organic
compounds (VOC) including chloroform, dichloromethane,
hexane, methanol, ethanol, acetone, toluene, xylene, and
formaldehyde were tested by separately adding these
compounds to the system directly at a very large excessive
amount over CO (104 equiv), and no significant fluorescence
changes were observed. These results indicated pronounced
selectivity of the substitution-based sensor toward CO.
Figure 1. (A) Relationship between fluorescence enhancement and
reaction time (CO concentration from upper to lower: 1000, 800, 500,
200, 100, and 50 ppm). Inset: with CO concentration 1% (red) and
5000 ppm (blue). Data are presented as average
SD from three
independent measurements. (B) The double logarithmic relationship
between the reaction rate (the slope in panel A) and the concentration
of CO (ppm). Ambient temperature: 10 °C.
the systems linearly increased with reaction time at all
concentrations studied, implying steady rates (r) of the
reactions that can be expressed by the slopes of the curves.
Further, we also looked into the dependence of the reaction
rate on CO concentration. The kinetic equation of the reaction
at a fixed concentration of the iodide 2 can be written as r =
k[CO]a[azide]b (eq 1), where k is the apparent rate constant.
Since the concentration of azide in the reaction was largely
excessive to CO, it can be merged to the rate constant and the
equation is thus derived as r = k′[CO]a (eq 2). The
bilogarithmic diagram of the reaction rate versus CO
concentration showed a linear curve in the range of 50−1000
ppm (Figure 1B), which fit well to eq 2, and the slope of the
curve (ca. 1.73) suggested an approximate second-order
reaction for CO. The existence of a linear fitting also makes
quantitative measurement of CO concentration from 50 to
1000 ppm possible.
To demonstrate the practical applicability of the CO sensor
for on-site warning of CO acute poisoning and explosion, we
further explored the possibility of visual detection. For safety
concerns in the experiments, we tested three CO concen-
trations, i.e., 1000 ppm, 5000 ppm and 1%, which were below
the lower flammation and explosion limit of CO (12.5%). After
several minutes, all of the reaction solutions emitted bright blue
fluorescence that was easily observable with the naked eye
under 320 nm excitation (Figure 2). The corresponding
In summary, we propose a CO fluorescence chemsensor
working in aqueous solution based on the transformation of a
weakly fluorescent iodide to a strongly fluorescent amino
3298
dx.doi.org/10.1021/ol5012949 | Org. Lett. 2014, 16, 3296−3299

Organic Letters
Letter
(b) Yamaura, H.; Tamaki, J.; Moriya, K.; Miura, N.; Yamazoe, N. J.
Electrochem. Soc. 1997, 144, 158−160.
(5) (a) Moragues, M. E.; Esteban, J.; Ros-Lis, J. V.; Martinez-Manez,
R.; Marcos, M. D.; Mareinez, M.; Soto, J.; Sancenon, F. J. Am. Chem.
Soc. 2011, 133, 15762−15772. (b) Esteban, J.; Ros-Lis, J. V.; Martinez-
Manez, R.; Marcos, M. D.; Mareinez, M.; Soto, J.; Sancenon, F. Angew.
Chem., Int. Ed. 2010, 49, 4934−4937. (c) Itou, M.; Araki, Y.; Ito, O.;
Kido, H. Inorg. Chem. 2006, 45, 6114−6116. (d) Benito-Garagorri, D.;
Puchberger, M.; Mereiter, K.; Kirchner, K. Angew. Chem., Int. Ed. 2008,
47, 9142−9145. (e) Giulino, A.; Gupta, T.; Altman, M.; Lo Schiavo, S.;
Mineo, P. G.; Fragala, I. L.; Evmenenko, G.; Dutta, P.; Van der Boom,
M. E. Chem. Commun. 2008, 25, 2900−2902.
(6) Michel, B. W.; Lippert, A. R.; Chang, C. J. J. Am. Chem. Soc. 2012,
134, 15668−15671.
Figure 3. Fluorescence response of 50 μM compound 2 to 1% CO
after 0.5 h, VOCs (104 equiv), and some other gases (100%) after 6 h:
(1) CO, (2) formaldehyde, (3) methanol, (4) ethanol, (5)
dichloromethane, (6) chloroform, (7) acetone, (8) toluene, (9)
xylene, (10) hexane, (11) carbon dioxide, (12) nitric oxide, (13) nitric
dioxide, (14) sulfur dioxide. Ambient temperature: 10 °C.
(7) (a) Barnard, C.F. J. Organometallics 2008, 27, 5402−5422.
(b) Ojima, I. Chem. Rev. 1988, 88, 1011−1030.
(8) Miloserdov, F. M.; Grushin, V. V. Angew. Chem., Int. Ed. 2012, 51,
3668−3672.
(9) (a) Kamer, P. C. J.; van Leeuwen, P. W. N. M.; Reek, J. N. H. Acc.
Chem. Res. 2001, 34, 895−904. (b) Freixa, Z.; Van Leeuwen, P.W. N.
M. Dalton. Trans. 2003, 1890−1901. (c) Freixa, Z.; Van Leeuwen, P.
W. N. M. Coord. Chem. Rev. 2008, 252, 1755−1786. (d) Martinelli, J.
R.; Watson, D. A.; Freckmann, D. M.M.; Barder, T. E.; Buchwald, S. L.
J. Org. Chem. 2008, 73, 7102−7107.
product. The straightforward substitution-based design allowed
a simple synthesis and mild conditions for the target reaction.
The pronounced fluorescence enhancement and specificity of
the sensor toward CO enabled quantitative CO detection. In
addition, the bright emission arising from a short-time reaction
also made visual detection of CO possible, suggesting the
potential application for on-site CO monitoring. The additional
merit is that the amino product 1 can be recoverable.
Considering the vast availability of iodide compounds and a
variety of fluorophores, our method can be extended to CO
sensing with different purposes, i.e., different emissions for
specific needs. Further investigations persuing CO fluorescence
chemsensors with faster response and higher sensitivity are
underway.
(10) Goldstein, M. J. Emerg. Nurs. 2008, 34, 538−542.
ASSOCIATED CONTENT
* Supporting Information
■
S
Additional information including experimental details, charac-
terization data, and fluorescence measurement. This material is
available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Authors
■
*E-mail: songwu@whu.edu.cn.
*E-mail: zhhliu@whu.edu.cn.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by the National Natural Science
Foundation of China (Nos. 21075094, 21375098), the National
Basic Research of China (973 program, No. 2011CB933600),
and the Program for New Century Excellent Talents in
University (NCET-11-0402).
REFERENCES
■
(1) (a) Hampson, N. B. J. Emerg. Med. 1998, 16, 695−698. (b) Cobb,
N.; Etzel, R. A. J. Am. Med. Assoc. 1991, 266, 659−663.
(2) http://www.wired.com/wiredscience/2010/11/afterdamp-in-
new-zealand/.
(3) Stetter, J. R.; Lin, J. Chem. Rev. 2008, 108, 352−366.
(4) (a) Arbiol, J.; Cirera, A.; Peiro, F.; Comet, A.; Morante, J. R.;
Delgado, J. J.; Calovino, J. J. Appl. Phys. Lett. 2002, 81, 1869−1871.
3299
dx.doi.org/10.1021/ol5012949 | Org. Lett. 2014, 16, 3296−3299