Chemo-Mechanical Joint Detection with Both Dynamic and Static

Jul 9, 2012 - Both the dynamic and the static cantilevers are employed to ... into a Wheatstone-bridge, and the close-up view of the sensing cantileve...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/ac

Chemo-Mechanical Joint Detection with Both Dynamic and Static Microcantilevers for Interhomologue Molecular Identification Haitao Yu,† Tiantian Yang,†,‡ Ying Chen,† Pengcheng Xu,† Dong-Weon Lee,§ and Xinxin Li*,† †

State Key Lab of Transducer Technology and Science and Technology on Micro-system Lab, Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Sciences, Shanghai 200050, China ‡ School of Digital Media, Jiangnan University, Wuxi, Jiangsu 214122, China § School of Mechanical Systems Engineering, Chonnam National University, Gwang-Ju, 500757, Korea S Supporting Information *

ABSTRACT: The study presents a novel chemo-mechanical joint-sensing method to distinguish a certain molecule from its homologous chemicals, using both a resonant cantilever for gravimetric sensing and a static cantilever for surface-stress sensing. Homologous amines of trimethylamine (TMA, Me3N), dimethylamine (DMA, Me2NH), and monomethylamine (MMA, MeNH2) are herein used as model objects for investigation. The molecular identification is based on experimental characterizations on both molecule adsorbing capability (by the resonant cantilever) and intermolecular lateral interaction (by the static cantilever). The intensities of the two sets of sensing signals are expected to be in opposite sequence with each other, due to the complementary relationship among the interhomologue molecule structures, i.e., a molecule containing a greater number of methyl substituents must possess a fewer number of nonsubstituted hydrogens. On the basis of the proposed idea, ppm-level vapors of the three amines are sequentially detected by a resonant microcantilever to characterize the molecular adsorption speed and another static cantilever to characterize the intermolecular lateral attraction induced surface stress. From the experiment, a pair of opposite sequence in sensing-signal amplitude has indeed been obtained that verifies the proposed joint-sensing method. In addition, the two sensing signals both show a linear relationship with chemical concentration (at low-concentration range). Further comparison between the two sensing results can help to build a model to identify the molecule among a series of its homologous chemicals by eliminating the influence from concentration. Since a complementary relationship among homologous molecule structures widely exists, the dual-sensing method is promising in on-the-spot rapid molecular identification among homologous chemicals.

I

effect can be characterized into an electric or electronic signal like electric voltage, photoelectronic current, resonant frequency, value change in piezoresistance or piezoelectricity, and so on. Among the recently developed microsensors, we would like to mention the promising chemo-mechanical microcantilevers that can translate the molecule recognition into micro/nano mechanical effects.2 According to the sensing scheme, the microcantilever sensors can be categorized into dynamic and static styles.3,4 The former is a resonant cantilever that detects the added mass of the adsorbed (or absorbed) molecules by recording the shift of the resonance frequency,5 while the latter detects the targeted molecules by measuring the adsorption-induced nanomechanical surface stress.6−10 When the adsorbed mass, Δm, is much smaller than effective mass of the resonant cantilever itself, meff, the frequency-shift signal can be considered being proportional to Δm, i.e.,

nterhomologous identification is of importance, especially when there is a significant difference in properties among the chemical homologues. For example, methanol features much severer toxicity and harmfulness to human beings than ethanol and, thus, needs to be clearly distinguished in many places. Accurate identification among homologous chemicals normally needs either time-consuming chemical analyzing methods or expensive/bulky instruments like gas chromatograph. By now, there has seldom been a portable and low-cost sensing method for on-the-spot rapid identification among chemical homologues, especially in trace-concentration levels. Therefore, exploring new microsensor techniques for this purpose is highly demanded. Generally, there are two interfaces in most of the bio/ chemical molecule microsensors.1 At the first interface, specific molecule reaction or binding is translated into a physical phenomenon or effect, such as an optical or photonic one, a mass or mechanical one, and a thermal or calorimetric one, as well as, an electrochemical one. In most cases, molecule recognition work is implemented at the first interface. Then, at the second interface that may be named as a signal transduction interface, the specific reaction induced physical phenomenon or © 2012 American Chemical Society

Δf ≈ −0.5Δm(f0 /meff )

(1)

Received: April 27, 2012 Accepted: July 9, 2012 Published: July 9, 2012 6679

dx.doi.org/10.1021/ac3011022 | Anal. Chem. 2012, 84, 6679−6685

Analytical Chemistry

Article

Figure 1. (a) SEM image of the integrated resonant microcantilever. (b) SEM image of the piezoresistive static microcantilever sensor, which is configured into a Wheatstone-bridge, and the close-up view of the sensing cantilever (with Au coated at top surface). (c) Schematic to illustrate the joint detection to amine vapors using both the resonant and the static microcantilevers.

compressive) surface stress at the cantilever surface.14 The surface stress causing cantilever bending movement can be selfsensed via an embedded piezoresistor in the cantilever for electric signal readout. The surface stress, σS, induced cantilever bending curvature can be expressed by Stoney’s equation15

where f 0 is the initial resonance-frequency before mass loading.11 In other words, the sensing signal is proportional to the number of the adsorbed molecule. The molecule adsorption speed onto a certain surface area of the cantilever can be monotonously determined by the adsorbing capability of the adsorbate molecules to the receptor molecules that were premodified at the cantilever surface.12,13 Therefore, by experimentally characterizing adsorption speed, the dynamic cantilever can discriminate the adsorption-capability sequence among a series of homologous molecules. As for the static cantilever, the adsorption specificity only influences the molecular surface coverage that directly depends on the adsorbed molecule number. Although a high surface coverage is helpful for building the lateral effect among the absorbed molecule chains, the finally generated surface-stress intensity at the cantilever surface is dominant by the lateral (parallel to the surface) interaction force between the adsorbed adjacent molecules instead of the vertical-to-surface adsorption specificity. The lateral intermolecular effects can be induced by one or several following factors: intermolecular hydrogen bonding interaction, electrostatic dipole−dipole or cation− anion interaction, stero-disparity effect, van der Waals force, etc.14 Steric disparity mainly occurs for the cases of the biggersize biomolecules, and van der Waals force merely helps to regulate the adsorbed molecules into an ordered state instead of generating strong lateral interaction force. Then, the strong cation−anion interaction only occurs when the detection is in a solution environment, where ionic dissociation may take place. In vapor detection, the electrostatic effect is mainly expressed as the dipole−dipole repulsion due to the adsorption induced charge redistribution. In the case of physical adsorption, the charge redistribution in the molecule is not severe, and the electrostatic interaction is quite weak. For chemical vapor molecules, it is the hydrogen bond based intermolecular effect that can be quite strong to induce large attractive (or repulsive) force, and the lateral force in turn generates tensile (or

6(1 − υ) 1 = σS r Et 2

(2)

where ν and E are Poisson’s ratio and Young’s modulus of cantilever material and t is cantilever thickness. Along with the sequence in intermolecular lateral interaction capability among the adsorbed homologues, obviously, the molecule with stronger interaction capability will generate larger surface stress, thereby outputting higher nanomechanical sensing signal. The hydrogen bond has been widely used in sensing of gaseous chemical molecules.16,17 The reason lies in that hydrogen bonds or hydrogen-bond-like interactions are weaker than those based on an ionic bond or covalent bond. For chemical vapor detection, a good balance between adsorption rate and desorption rate is generally needed. If ionic bond or covalent bond is used after detection, the sensor would be hard to rapidly recover from the strong molecular binding state, which makes it difficult for detection reproducibility and repeatability. At the other extreme, pure physical adsorption (i.e., van der Waals force based physical adsorption) is still not a good sensing mechanism, as the specificity is too low to provide enough sensing selectivity. Hence, the moderate hydrogen bond adsorption (or hydrogen bond based physical adsorption) is suitable for gaseous chemical sensing, as it provides both adsorption specificity and good adsorption/ desorption balance for on-the-spot rapid detection. As is addressed above, the dynamic microcantilever response can distinguish the sequence in hydrogen-bond adsorption capability among a series of homologous molecules, while the static cantilever can discriminate among the homologues by 6680

dx.doi.org/10.1021/ac3011022 | Anal. Chem. 2012, 84, 6679−6685

Analytical Chemistry

Article

mesoporous silica is added into the solution and stirred at 70 °C for 3 days. After the solution is filtered, the solid is washed with deionized water for several times. Under vacuum, the −COOH functionalized mesoporous silica (CMS) solid product is dried at 60 °C. Then, 0.01 g of as-prepared mesoporous sensing material is added into 1 mL of deionized water (under ultrasonic) to form a crude suspension. About 0.1 μL of suspension is loaded onto the cantilever top surface using a commercial micromanipulator (Eppendorf, model: Patch Man NP2), with the aid of a microscope (Leica, model: DM4000). The microcantilever is dried in an oven at 60 °C for about 2 h. Loaded with the as-functionalized sensing material, the cantilever is ready for specific detection of amine vapor via acid-to-base reaction. Self-Sensing Static Microcantilever and Surface SelfAssembly of Sensing Layer. The design and fabrication details of the lab-made static microcantilever used in this research has been reported previously.21,22 The cantilever dimensions are 90 μm in lengh, 21 μm in width, and 1 μm in thickness. The self-sensing cantilever with a silicon piezoresistor is integrated for surface-stress signal output. For depression of the noise from an environmental temperature change and the interference from air flow, a Wheatstone-bridge configuration is built by interconnecting the sensing cantilever (with gold thinfilm coated at front-side, while the backside SiO2 is exposed for modification) with a referential piezoresistive cantilever (with identical structure but no coating) and two other fixed resistors (at the chip frame). The two piezoresistors and the two silicon resistors are with the same pattern and impurity doping parameters. The top-view configuration of the static cantilever sensor chip is shown in the SEM image of Figure 1b. Prior to the sensing-terminals modified, the cantilever sensor chip is sequentially cleaned with Piranha solution, deionized water, acetone, and absolute ethanol for several times each. (Caution! Piranha solution reacts violently with organic compounds and should be handled with extreme care.) FAS-17 (heptadecafluorodecyltrimethoxysilane, Gelest Inc.) monolayer is first self-assembled on all of the SiO2 surface of the chip as a hydrophobic coating layer, by immersing the cantilevers into 0.2 mM FAS17/ethanol solution (with 1% water added as catalyst). Then, the cantilevers are rinsed with absolute ethanol for several times and activated at 120 °C in air for 1 h. Selfassembled only on the SiO2 surface, the FAS-17 molecules do not contaminate the Au surface at the top surface of the sensing cantilever. The FAS-17 self-assembly on SiO2 is for eliminating the noise from ambient humidity and for resisting against modification of the sensing groups on insensitive regions, with the technical details described in the previous publication.16 After the Piranha cleaning procedure, 3-mercaptopropionic acid (3-MPA, purchased from Aldrich) sensing monolayer is selfassembled on the gold surface by immersing the sensor into 6 mM 3-MPA/ethanol solution for 24 h. After being rinsed with absolute ethanol and blow-dried with pure nitrogen, the surface-stress sensitive cantilever is ready for detection of amine vapor. The 3-MPA self-assembly on gold surface can be also referred to in the previous publication.14 Preparation of ppm-Level Methylamine Homologous Vapors. A detection chamber (made of Pyrex glass) of 40 L volume is setup in our lab. A small electric fan and a miniheater are fixed inside the chamber for heating and temperature stability. We dropwise inject a constant volume of liquid methylamine homologous sample onto the heater, which is stirred by the fan, to obtain homogeneous vapor and the

responding to the hydrogen-bond based lateral intermolecular effect. If only with one type of cantilever, it is really hard to achieve the interhomologous identification. The reason is simple. A higher sensing signal can originate from either a relatively higher chemical concentration of lower-sensitivity molecules or a lower concentration of higher-sensitivity molecules. Fortunately, this molecule sorting task would possibly be well done, when the two microcantilevers joined work together. The key is that, for molecule identification from a series of homologues, the sensing-signal intensities from the two cantilevers are expected to be both proportional to the concentration of the gas. More importantly, the two cantilever sensing signals are in opposite sequence with each other. The reason of the opposite sequence in the two sensing signals lies in that, between two homologous molecules, one molecule which exhibits relatively stronger adsorption capability (i.e., higher signal from the dynamic cantilever) may, as a trade-off, feature weaker lateral intermolecular effect (i.e., lower signal from the static cantilever). Take amine homologues as examples, which are monomethylamine (MMA), dimethylamine (DMA), and trimethylamine (TMA). For a certain methylamine molecule, when it contains a relatively larger number of methyl groups, its central nitrogen will exhibit relatively weaker binding-capacity to the outer hydrogen of the methyl group. Accordingly, the methyl group with more freedom will exhibit stronger hydrogen-bond adsorption capability. On the other hand, more methyl groups mean a fewer number of nonsubstituted hydrogen in the molecule, thereby, leading to a weakened lateral hydrogen-bonding capability between adjacent molecules. Making full use of the pair of inversely ordered properties of the homologous molecule, we explore a novel joint-sensing scheme to identify a certain molecule from its homologues. Both the dynamic and the static cantilevers are employed to compose a higher-order chemical sensing device18,19 for joint sensing to both adsorption capability and adjacent-molecule interaction, respectively.

■

EXPERIMENTAL SECTION Premodified Resonant Microcantilever for Detection of Specific Adsorption. The gravimetric detection scheme for adsorbed mass and the design and fabrication of the labmade integrated resonant cantilever, as well as the interfacecircuit for signal readout, are all detailed in the previous publication.20 The geometric dimensions of the cantilever are 200 μm in length, 100 μm in width, and 3 μm in thickness. As is shown in Figure 1a, an electro-thermal microheater for resonance excitation and a piezoresistive Wheatstone-bridge for signal pickup are both integrated in the silicon microcantilever. A phase-lock-loop (PLL) based interface circuit is used to form a close-loop resonator microsystem for high-performance resonance maintenance. At the sensing location near the cantilever free end, the synthesized nanostructure of mesoporous silica is loaded to enhance the specific surface area for effective adsorbance of amine molecules.5 Before the mesoporous silica loaded onto the cantilever, specific-sensing functionalization of the amine sensor is implemented by the direct in-wall, self-assembling −COOH group to the mesoporous sensing film. First, 1 mL of CES (carbomethoxysilanetriol, sodium salt, 25% (vol/vol) in water, Gelest Inc.) is added into 30 mL of deionized water. Then, concentrated HCl solution (12M) is dropwise added to adjust the pH value to 4.5. Thereafter, 0.2 g of SBA-15 6681

dx.doi.org/10.1021/ac3011022 | Anal. Chem. 2012, 84, 6679−6685

Analytical Chemistry

Article

Figure 2. Room-temperature sensing results from the resonant cantilever. (a) Frequency-shift signals to the three kinds of amine vapors of the same concentration of 4 ppm. (b) The slope signals deduced from the results in (a). (c) At ppm-level concentration range of TMA, the slope signals show a proportional relationship to the vapor concentration. (d) The repeated responses of the resonant cantilever to 4 ppm TMA.

concentration can be worked out by Clapeyron equation.23 In this simple way, low concentration vapors of ppm-level can be generated for sensing experiments. Sensing to Homologous Amine Vapors by Both the Resonant and the Static Cantilevers. The sensing experiments are performed at room temperature. Both the mass sensitive resonant microcantilever and the surface-stress sensitive static cantilever are sequentially introduced to the ultralow concentration vapors of monomethylamine (MMA), dimethylamine (DMA), and trimethylamine (TMA). For each kind of vapor, step-increased concentrations of 2 ppm, 4 ppm, 6 ppm, and 10 ppm are sequentially generated for detection. After detection to the previous kind of vapor, the cantilevers need to be adequately recovered and, then, used again for detection of the next kind of vapor. In other words, the sensing signals for all the three kinds of vapors are from identical cantilevers.

vR ≈ −0.5(f0 /meff )· d(Δm)/dt

(3)

−0.5(f 0/meff) is a constant and, thus, vR can be further expressed as

vR ∝ M dn/dt

(4)

where n is adsorbed molecule number and M is molecular weight. Therefore, vR can be used to represent the molecule adsorbing speed of dn/dt. On the basis of Langmuir adsorption theory,24,25 the surface adsorption speed can be expressed as dn/dt ∝ k1(1 − θ )cN

(5)

where the adsorption speed constant of k1 represents the adsorbing capability, θ is surface coverage, c is vapor concentration, and N is the whole cite number for molecule adsorption. At the early stage of the adsorption process (except for the initial few seconds of exposure25), θ ≪ 1 and N can be considered as a constant. Thus, we have

■

dn/dt ∝ k1c

RESULTS AND DISCUSSION Sensing Results of the Resonant Microcantilever. The resonant microcantilever is used to sequentially detect the three sorts of homologous amine vapors. Figure 2a shows the experimentally obtained frequency shift (Δf) to monomethylamine (MMA), dimethylamine (DMA), and trimethylamine (TMA), respectively. The concentrations of the three vapors are all 4 ppm. As is shown in Figure 2b, the response is further transformed into adsorbing speed signal by calculating the time derivative, i.e., the slope signal of vR = d(Δf)/dt. According to eq 1, the slope signal of vR can be expressed as

(6)

Then, eq 4 can be rewritten as

vR = kR c

(7)

where the constant kR ∝ M·k1. Hence, for a fixed concentration of c, the slope signal vR is proportional to kR and determined by the molecular weight and the molecule adsorbing capability. As is shown in Figure 2b, the slope signals for the three amine vapors are in the order of Me3N > Me2NH > MeNH2. On the basis of sensing experiments, the apparent adsorbing speed (characterized by the slope signal) for various trimethyl6682

dx.doi.org/10.1021/ac3011022 | Anal. Chem. 2012, 84, 6679−6685

Analytical Chemistry

Article

Figure 3. (a) Measured static cantilever sensor response to the three kinds of amine vapors with identical concentration of 4 ppm at room temperature. (b) Proportional relationship between the static cantilever signal and the ppm-level concentration of TMA. (c) Repeated responses of the static cantilever sensor to 4 ppm TMA. (d) Response of the bare static cantilever (i.e., without gold functionalization) to 4 ppm TMA.

attraction intensity, the surface stress has a proportional relationship with the concentration of the vapor, i.e.,

amine (TMA) concentration of 2 ppm, 4 ppm, 6 ppm, and 10 ppm is obtained and shown in Figure 2c. Obviously, at low concentration range of amine vapor, the adsorbing speed signal has a proportional relationship with the concentration. The signal repeatability of the resonant cantilever sensor is experimentally evaluated. The three-cycle continual detection signal to 4 ppm TMA is recorded in Figure 2d. On the basis of the frequency-shift data for TMA, DMA, and MMA, the corresponding slope signals can be further obtained. The repeatability error for the TMA detecting slope signal is obtained and denoted in Figure 2c, and the results for DMA and MMA are shown in Figure S1 of the Supporting Information. The repeatability error for the three vapors is generally smaller than 14%. It needs to be noted that all the above-mentioned results are obtained using one resonant cantilever. Sensing Results of the Static Microcantilever. The sensing results of the static cantilever sensor to the three amine homologues are shown in Figure 3a, where the three vapors are all in 4 ppm. Tensile surface stress is always measured for the three amines. More importantly, great disparity in the amplitude of surface stress (σS) is obtained. The order for the generated surface-stress signal is Me3N < Me2NH < MeNH2 that is opposite to the order for the adsorbing speed detected by the resonant microcantilever. Then, the cantilever is used to sequentially detect various TMA concentrations of 2 ppm, 4 ppm, 6 ppm, and 10 ppm, with the results shown in Figure 3b. As the characterization of intermolecule lateral

vS = k Sc

(8)

where vS is the response voltage of the static cantilever and kS is a constant to represent the intensity of the intermolecular effect. The three-cycle continual detection signal to 4 ppm TMA is recorded in Figure 3c. On the basis of the experimental data for TMA, DMA, and MMA (the latter two are shown in Figure S2 of the Supporting Information), the repeatability error is generally less than 9%. All the above experimental results are obtained using the same static microcantilever. As a reference, the response of a bare static cantilever, i.e., without gold functionalization, to 4 ppm TMA is shown in Figure 3d. After the target analyte is injected, no obvious sensing signal can be observed. The result also verifies that, under our detection condition, the Au/SiO2 bimetallic effect (i.e., the thermal mismatch between the two materials of the cantilever) has negligible influence to the output signal of the piezoresistor. Interhomologous Molecule Identification by Division Operation between the Two Signals. On the basis of the sensing experimental results in previous subsections, two conclusions can be made: (1) at low concentration range of ppm-level, the two types of cantilevers both show proportional sensing signal to vapor concentration; (2) when sensing to the three methylamine homologues, the order of the frequencyshift speed signal (i.e., Me3N > Me2NH > MeNH2) is opposite 6683

dx.doi.org/10.1021/ac3011022 | Anal. Chem. 2012, 84, 6679−6685

Analytical Chemistry

Article

Figure 4. (a) Specific-adsorption schematic showing the hydrogen bond between the adsorbed MMA molecules and the prefunctionalized −COOH sensing groups onto the inwall of the porous-silica sensing film. (b) After specific adsorption, the hydrogen-bonding-like lateral intermolecular effects are analyzed in detail for MMA, DMA, and TMA, respectively.

of amine and O of either CO or −OH in the −COOH group. If a methylamine homologous molecule contains larger methyl number, it must contain fewer unsubstituted H numbers. Such a molecule structure will facilitate to decrease the formation probability of lateral hydrogen-bonding between H and the O in the neighboring −COOH groups, thereby weakening the lateral intermolecular attraction. In the molecule of MMA, DMA, or TMA, the free H number is two, one, or null, respectively, and the order of lateral-interaction induced surface-stress strength should be TMA > DMA > MMA. According to this pair of opposite sequence between the two cantilever sensing signals, which reflects the effect of moleculestructure difference on micro/nanomechanical response, interhomologous molecules can be rapidly identified with the joint-detection model illustrated in Figure 5. Using an embedded MCU (micro control unit) in the sensor package, real-time calculation of the ratio between the pair of signals (vR/vS) can be easily implemented. Since the two signals are both proportional to the vapor concentration [see Figures 2c and 3b], the influence from the unknown concentration, c, can

to the order of the surface-stress strength (i.e., Me3N < Me2NH < MeNH2). The frequency-shift speed signal represents the early stage adsorbing speed of amine molecules that reflects the strength of the molecule adsorbing capability. Figure 4a schematically illustrates the adsorption of MMA (MeNH2) molecules onto the resonant cantilever surface via hydrogen bonding with the premodified −COOH sensing terminals. According to the theory about gas basicity of methylamine homologues, the order in gas-phase basicity is Me3N > Me2NH > MeNH2, which can be inferred from the electron-releasing “inductive effect”.26,27 In other words, if the amine molecule contains a larger methyl number, the central N atom will exhibit stronger electronegativity and higher molecule adsorbing speed (dn/dt) at the early adsorption stage (i.e., θ ≪ 1). Therefore, the speed signal order experimentally obtained from the resonant microcantilever indeed reflects the adsorption capability of the three amine molecules. For the static microcantilever that translates adsorption induced surface stress into piezoresistive signal, the vertical interaction between the premodified −COOH terminated molecule and the targeted amine molecule contributes much less to surface-stress generation compared to the lateral interaction between adjacent molecules. The lateral interaction may come from hydrogen-bond effect, steric factor, electrostatic force, or van der Waals force.14,28,29 In this vapor sensing experiment, there is no anionic/cationic lateral interaction occurring. The possible dipole-to-dipole lateral interaction between adjacent molecules is also very weak, just like the other weak effects of steric factor (mainly existing for bigger-size biomolecules) and van der Waals force. Therefore, the lateral hydrogen-bond capability will dominate the generated surface stress. As is schematically illustrated in Figure 4b, herein, the vertical interaction is via hydrogen-bonding adsorption between the central atom N of amine molecule and H of −COOH. Hence, the lateral intermolecular attraction should come from the hydrogen-bond-like interaction between the adsorbed amine molecule and its adjacent molecules with the −COOH groups. The lateral hydrogen-bond attraction can be between H

Figure 5. After calculating the ratio value, |vR/vS|, between the sensing signals from the two cantilevers, interhomologue molecular identification can be realized. 6684

dx.doi.org/10.1021/ac3011022 | Anal. Chem. 2012, 84, 6679−6685

Analytical Chemistry

Article

Notes

be eliminated by the division operation between the two signals, i.e., vR/vS = kR/kS. On the basis of the calculated ratio values for the three kinds of vapors and the various concentrations, three sparsely separated horizontal lines are obtained and shown in Figure 5. Thanks to the opposite order between the two signals for the three amine vapors, the great disparity in the ratio value facilitates one to distinguish a certain molecule from its homologues. Without the interference from uncertainty of the vapor concentration, the targeted molecule can be identified by comparing the resultant ratio with the yaxis magnitude of the three horizontal lines, where each line creates a benchmark for one kind of molecule. It is worthy to note that, serving as the benchmark for interhomologues identification, the lines should be precalibrated by sensing experiments. The dual-cantilever joint-detection method is expected to be widely used for molecule identification of many other homologous chemicals, because such a pair of inversely ordered signals between adsorbing capability and lateral interaction capability exists generally in various chemical homologues. The oppositely sequenced signals reflect the complementary relationship among the interhomologue molecule structures, e.g., more substituents corresponding to less nonsubstituted atoms. The technical key lies in finding appropriate sensing groups, which are premodified on the cantilevers, to make great disparity in the oppositely ordered dual signals for reliable molecule identification among the homologues.

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The research is supported by Chinese 973 Program (2011CB309503), NSFC Program (91023046, 61161120322, 61021064, 61102010) and Chinese National Key Technology R&D Program (2012BAK08B05). D.-W. Lee. also thanks Korean WCU project (R32-2009-000-20087-0).

■

■

CONCLUSIONS With both resonant cantilever for gravimetric sensing and static cantilever for surface-stress sensing, a novel chemo-mechanical joint detection method is presented to distinguish a kind of molecule from its chemical homologues. Homologous methylamines are herein taken as examples for study. The molecular identification is based on a pair of opposite order between adsorbing capability of the molecule and intermolecular attraction of adjacent adsorbed molecules. In detail, a relatively greater methyl number in one amine molecule features stronger surface-adsorption capability; meanwhile, the fewer numbers of unsubstituted hydrogen shows weaker lateral interaction capability of the neighboring molecules30 and vice versa. By employing both the resonant cantilever and the static cantilever, we have experimentally verified that the intensities of the two sets of sensing signal are indeed in opposite sequence with each other. By calculating the ratio between the two oppositely ordered signals for the three amine molecules, which reflects the complementary relationship in molecule structure among homologous chemicals, interhomologous molecules can be rapidly identified. Unlike the traditional identification techniques such as gas chromatogram that is bulky, expensive, and time-consuming, this dual-cantilever joint detection method is promising to be widely used for on-thespot rapid identification among homologous chemicals.

■

ASSOCIATED CONTENT

S Supporting Information *

Additional information as described in the text. This material is available free of charge via the Internet at http://pubs.acs.org.

■

REFERENCES

(1) Sepaniak, M. J.; Datskos, P. G.; Lavrik, N. V. Anal. Chem. 2002, 74, 568A−575A. (2) Waggoner, P.; Craighead, H. Lab. Chip 2007, 7, 1238−1255. (3) Johnson, B. N.; Mutharasan, R. Biosens. Bioelectron. 2011, 32, 1− 18. (4) Li, X. X.; Lee, D. W. Meas. Sci. Technol. 2012, 23, 022001. (5) Xu, P. C.; Yu, H. T.; Li, X. X. Anal. Chem. 2011, 83, 3448−3454. (6) Fritz, J. Analyst 2008, 133, 855−863. (7) Ji, H. F.; Finot, E.; Dabestani, R.; Thundat, T.; Browna, G. M.; Britta, P. F. Chem. Commun. 2000, 6, 457−458. (8) Gimzewski, J. K.; Gerber, C.; Meyer, E.; Schlittler, R. R. Chem. Phys. Lett. 1994, 217, 589−594. (9) Hwang, K. S.; Eom, K.; Lee, J. H.; Chun, D. W.; Cha, B. H.; Yoon, D. S.; Kim, T. S.; Park, J. H. Appl. Phys. Lett. 2006, 89, 173905. (10) Bashir, R.; Hilt, J. Z.; Elibol, O.; Gupta, A.; Peppas, N. A. Appl. Phys. Lett. 2002, 81, 3091−3093. (11) Ono, T.; Li, X. X.; Miyashita, H.; Esashi, M. Rev. Sci. Instrum. 2003, 74, 1240−1243. (12) Kummer, A. M.; Burg, T. P.; Hierlemann, A. Anal. Chem. 2006, 78, 279−290. (13) Su, J. J.; Carron, C.; Truax, S.; Demirci, K. S.; Beardslee, L. A.; Brand, O. In Transducers 11, Int. Conf. Solid-State Sens. Actuators, Beijing, China, June 5−9, 2011, IEEE Computer Society: New York, 2011; pp 1420−1423. (14) Yang, T. T.; Li, X. X.; Chen, Y.; Lee, D. W.; Zuo, G. M. Analyst 2011, 136, 5261−5269. (15) Stoney, G. G. Proc. R. Soc. London, Ser. A 1909, 82, 172−175. (16) Zuo, G. M.; Li, X. X.; Zhang, Z. X.; Yang, T. T.; Wang, Y. L.; Cheng, Z. X.; Feng, S. L. Nanotechnology 2007, 18, 255501. (17) Li, P.; Li, X. X.; Zuo, G. M.; Liu, J.; Wang, Y. L.; Liu, M.; Jin, D. Z. Appl. Phys. Lett. 2006, 89, 074104. (18) Hierlemann, A.; Gutierrez-Osuna, R. Chem. Rev. 2008, 108, 563−613. (19) Hagleitner, C.; Hierlemann, A.; Lange, D.; Kummer, A.; Kerness, N.; Brand, O.; Baltes, H. Nature 2001, 414, 293−296. (20) Yu, H. T.; Li, X. X.; Gan, X. H.; Liu, Y. J.; Liu, X.; Xu, P. C.; Li, J. G.; Liu, M. J. Micromech. Microeng. 2009, 19, 045023. (21) Li, P.; Li, X. X. J. Micromech. Microeng. 2006, 16, 2539−2546. (22) Chen, Y.; Xu, P. C; Li, X. X. Nanotechnology 2010, 21, 265501. (23) Jenkins, H. D.B. Chemical Thermodynamics at a Glance: Chapter 26. Clausius-Clapeyron Equation; Blackwell: Oxford, 2008. (24) Atkins, P. W.; De Paula, J. Physical Chemistry, 7th ed.; Oxford University Press: New York, 2001. (25) Berger, R.; Delamarche, E.; Lang, H. P.; Gerber, C.; Gimzewski, J. K.; Meyer, E.; Güntherodt, H.-J. Science 1997, 276, 2021−2024. (26) Aue, D. H.; Webb, H. M.; Bowers, M. T. J. Am. Chem. Soc. 1976, 98, 311. (27) Bouchoux, G. Mass. Spectrom. Rev. 2007, 26, 775. (28) Wu, G.; Ji, H.; Hansen, K.; Thundat, T.; Datar, R.; Cote, R.; Hagan, M. F.; Chakraborty, A. K.; Majumdar, A. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 1560. (29) Godin, M.; Tabard-Cossa, V.; Miyahara, Y.; Monga, T.; Williams, P. J.; Beaulieu, L. Y.; Lennox, R. B.; Grutter, P. Nanotechnology 2010, 21, 075501. (30) Solomons, T.; Fryhle, C. Organic Chemistry, 8th ed.; Wiley: New York, 2003.

AUTHOR INFORMATION

Corresponding Author

*Telephone: (+86) 21-62131794. Fax: (+86) 21-62131744. Email: [email protected]. 6685

dx.doi.org/10.1021/ac3011022 | Anal. Chem. 2012, 84, 6679−6685