approximately 33-fold higher sensitivity in detection for alde-
hyde than that of ketone gas of the same mass was evident. In
our hands, the smallest aldehyde gas, formaldehyde, could
also react with SIL 1 and be detected by QCM (Fig. S2, ESIw).
Encouraged as we were by the aldehyde/ketone results, we
next turned our attention to the SIL detection of amine gases, as
another important class of VOCs (footnote 2, ESIw), and
investigated the SIL 2 to test, based on the transimination
reaction, its effectiveness in amine gas sensing. Similar to the
imine formation, this transimination reaction is one of the most
fundamental and ubiquitous reactions in organic chemistry and
recently has attracted much attention in, for example, the
folding study of supramolecular assemblies (footnote 5, ESIw).
Initially, we investigated the direct transimination reaction of
propylamine as a model amine gas with only the SIL 2 thin-
coated on the quartz chip and found that, albeit detectable at
low concentration (28.5 ppb), the QCM response was minimal
(B0.5 Hz) and seemly reversible in its signal (Fig. 3A and Table
S1, ESIw). Realizing in literature that Lewis acids notably
facilitated the transimination reactions in conventional
molecular solvents (footnote 5, ESIw), we decided to examine
the same methodology for use in ionic liquid and randomly
selected seven metal triflates available in our laboratory, and
carried out screening to identify the most effective Lewis acid to
accelerate the transimination reaction of SIL 2 with propyl-
amine. To our delight, under our experimental condition
(28.5 ppb propylamine gas and 1 mol% metal triflate dissolved
in SIL 2), Sc(OTf)3 was found to be the most powerful Lewis
acid to catalyze the transimination reaction to produce the
largest and irreversible QCM response (20 Hz) (Fig. 3A and
Table S1, ESIw). The reaction course of this Sc(OTf)3-catalyzed
transimination can be explained on the basis of Scheme S3
(ESIw). The result in Fig. 3A also clearly indicates that, without
SIL 2, only Sc(OTf)3 (1 mol%) dissolved in a non-functionalized
[b-3C-im][NTf2] ionic liquid was totally inert to the amine gas,
further proving that the frequency drop in the continuous flow
QCM measurement was not due to the nonspecific dissolution
of amine gas in SIL 2. Fig. 3B details the quantitative study of
SIL 2 reaction with propylamine gas (0–120 ppb). In our
hands, this Sc(OTf)3-catalyzed, transimination-based SIL
platform is highly selective to amine gas: at DF = À1.0 Hz,
the sensitivity of detection was approximately 2.5 ppb for
propylamine. Most remarkably, our reaction-based SIL-on-chip
system is greatly sensitive to the detection of the smallest molecular
weight amine gas, ammonia (Fig. S3A, ESIw): approximately
3.9 ppb at DF = À1.0 Hz (Fig. S3B, ESIw).
In summary, we demonstrate here the successful use of SILs
thin-coated on QCM chips for effective detection of specific
gases. Our chemoselective gas sensing method by SILs is straight-
forward and avoids many problems (surface area, pore size and
shape, thermal stability, sieving effect, diffusivity, and complex
adsorption behaviors) associated when using porous solid
materials as adsorbents for gas adsorption and separation.10
Furthermore, this SIL approach is cost-effective that the QCM
chip can be readily regenerated by washing away the used SIL
and replacing the new one. Our SIL platform is chemoselective
with fast gas diffusion in ILs, readily applicable to very low
molecular weight gases such as ammonia and, most significantly,
insensitive to moisture. It is also noted that this SIL platform
need not be performed in a temperature-controlled device; that
is, the direct measurements on bench at ambient temperatures
sufficed. There are, to our knowledge, no reports in literature
based upon specific chemical reactions demonstrating chemo-
selective gas sensing in ionic liquid measured by QCM (footnotes 6
and 7, ESIw). This research is our first proof-of-concept
inspection of the promising use of a QCM-based sensor for
reaction-directed detection of gas samples, which is part of a
program aimed at studying environmental air quality control
and examining defective goods. Without doubt, more work is
needed. Lastly, the concept of reaction diversity, likely along
with the prefiltering strategy, can be applied and incorporated
in this research to ensure that a set of candidate SILs on QCM
chips will ultimately represent an ‘‘electronic nose’’.
This work was supported by a multi-year Grant-in-Aid
from the ANT Technology (Taipei, Taiwan) and in part
funded by a three-year grant from the National Science
Council (Taiwan, ROC). We thank reviewers for constructive
comments and valuable suggestions.
Notes and references
1 (a) S. Sowmiah, V. Srinivasadesikan, M.-C. Tseng and Y.-H. Chu,
Molecules, 2009, 14, 3780; (b) N. V. Plechkova and K. R. Seddon,
Chem. Soc. Rev., 2008, 37, 123.
2 S.-G. Lee, Chem. Commun., 2006, 1049.
3 A. Amann, P. Spanel and D. Smith, Mini-Rev. Med. Chem., 2007,
7, 115.
4 R. A. Potyrailo and V. M. Mirsky, Chem. Rev., 2008, 108, 770.
5 B. Timmer, W. Olthuis and A. van den Berg, Sens. Actuators, B,
2005, 107, 666.
Fig. 3 (A) Chemoselective detection of propylamine gas (28.5 ppb)
using 9 MHz QCM thin-coated with SIL 2 or [b-3C-im][NTf2]
(10 nL, 909 nm thickness). The sensorgrams of propylamine gas
detection by quartz chips coated with SIL 2 only and [b-3C-im][NTf2]
containing 1 mol% catalyst were shifted 2 and 5 Hz, respectively, for
clarity. (B) A plot of DF (Hz) vs. gas concentration (0, 17.1, 34.2, 51.3,
85.5 and 119.7 ppb).
6 H.-C. Kan, M.-C. Tseng and Y.-H. Chu, Tetrahedron, 2007, 63, 1644.
7 A 9 MHz QCM gas sensing device (the piezo sensor system
8 M. A. Cooper and V. T. Singleton, J. Mol. Recognit., 2007, 20, 154.
9 A. F. Sowinski and G. M. Whitesides, J. Org. Chem., 1979, 44,
2369.
10 J.-R. Li, R. J. Kuppler and H.-C. Zhou, Chem. Soc. Rev., 2009, 38,
1477.
ꢀc
This journal is The Royal Society of Chemistry 2010
Chem. Commun., 2010, 46, 2983–2985 | 2985