Inkjet-Printed Silver Nanoparticle Paper Detects Airborne Species

Mar 10, 2014 - Anhui Institute of Optics and Fine Mechanics, Chinese Academy of Sciences, Hefei, Anhui 230031, China. •S Supporting Information...
0 downloads 0 Views 4MB Size
Download PDF
Article pubs.acs.org/ac

Inkjet-Printed Silver Nanoparticle Paper Detects Airborne Species from Crystalline Explosives and Their Ultratrace Residues in Open Environment Jianping Wang,†,‡ Liang Yang,†,‡ Bianhua Liu,‡ Haihe Jiang,§ Renyong Liu,‡ Jingwei Yang,§ Guangmei Han,‡ Qingsong Mei,‡ and Zhongping Zhang*,†,‡ †

Department of Chemistry, University of Science and Technology of China, Hefei, Anhui 230026, China Institute of Intelligent Machines, Chinese Academy of Sciences, Hefei, Anhui 230031, China § Anhui Institute of Optics and Fine Mechanics, Chinese Academy of Sciences, Hefei, Anhui 230031, China ‡

S Supporting Information *

ABSTRACT: An electronic nose can detect highly volatile chemicals in foods, drugs, and environments, but it is still very much a challenge to detect the odors from crystalline compounds (e.g., solid explosives) with a low vapor pressure using the present chemosensing techniques in such way as a dog’s olfactory system can do. Here, we inkjet printed silver nanoparticles (AgNPs) on cellulose paper and established a Raman spectroscopic approach to detect the odors of explosive trinitrotoluene (TNT) crystals and residues in the open environment. The layer-by-layer printed AgNP paper was modified with p-aminobenzenethiol (PABT) for efficiently collecting airborne TNT via a charge-transfer reaction and for greatly enhancing the Raman scattering of PABT by multiple spectral resonances. Thus, a Raman switch concept by the Raman readout of PABT for the detection of TNT was proposed. The AgNPs paper at different sites exhibited a highly uniform sensitivity to TNT due to the layer-by-layer printing, and the sensitive limit could reach 1.6 × 10−17 g/cm2 TNT. Experimentally, upon applying a beam of near-infrared low-energy laser to slightly heat (but not destruct) TNT crystals, the resulting airborne TNT in the open environment was probed at the height of 5 cm, in which the concentration of airborne species was lower than 10 ppt by a theoretical analysis. Similarly, the odors from 1.4 ppm TNT in soil and 7.2, 2.9, and 5.7 ng/cm2 TNT on clothing, leather, and envelope, respectively, were also quickly sensed for 2 s without destoying these inspected objects.

E

for the collection of airborne chemical species.4 Unfortunately, the dogs are expensive to train and tire easily. A sensor similar to the olfactory system of dog heavily depends on addressing three key problems: (1) the method to collect the very few airborne molecules in open environment, (2) the prevention to the dynamic desorption of analyte from the sensing surfaces, and (3) the supersensitive approach to the signal readout. To date, these still remain the unparalleled challenges to the sensing principles and techniques. Trinitrotoluene (TNT) is the most commonly used crystalline explosives with a very low vapor concentration (only ∼4 ppb) at room temperature, and its theoretical vapor concentration even at the height of 1 cm away from the crystals is at least 3−4 orders of magnitude lower. The predominant technique for the explosive screening in airports is ion mobility spectroscopy that can determine the molecular mass of TNT by its mobility in the electric field after vaporization. The limited sensitivity requires the explosive residues to be physically

lectronic noses based on chemical gas sensors have attracted much attention because of their applications and potentials in the detections of foods, drugs, explosives, and environments.1−3 The chemical gas sensors widely use the sensing concept of surface adsorption of airborne species, where the efficiency of adsorption is, therefore, critical to the sensitivity of detection. It is true that the volatile chemicals at low concentrations usually require the sophisticated collecting, transporting, and concentrating devices2 or special surface processes3 to increase their adsorption amounts. These reported gas sensors have achieved the supersensitive detection of gaseous or volatile organic compounds with a relatively high vapor pressure, such as alcohol, hydrogen sulfide, benzene, amine, etc. On the other hand, it is rarely explored to probe the odors emanating from crystalline/solid organic compounds, which is very important in many fields and, in particular, for the detection of explosives in the homeland security against terrorism. The key difficulty is that the vapor pressures of most crystalline compounds are too low to directly detect their odors using the present sensing techniques. The best performance for the detection of odors is a sniffer dog (Canis familiariz), whose olfactory system evolves to be highly effective © 2014 American Chemical Society

Received: October 21, 2013 Accepted: March 9, 2014 Published: March 10, 2014 3338

dx.doi.org/10.1021/ac403409q | Anal. Chem. 2014, 86, 3338−3345

Analytical Chemistry

Article

The shape and size of printing were beforehand set on the computer. The printing was repeated until the desired number of layers on the cellulose paper was obtained. Modification of Raman-Active Molecules on AgNPs Paper. The freshly printed AgNP paper was modified with various Raman-active molecules as signal readout units. Typically, the AgNP paper was soaked into the aqueous solution of 1.0 × 10−7 M p-aminobenzenethiol (PABT) overnight. PABT was chemically adsorbed onto the surface of AgNPs by the formation of Ag−S bonds. The free PABT was removed from the AgNP paper by washing with ultrapure water. The same procedure was also used for the modifications of other Raman-active molecules with a thiol group, including 4-mercaptotoluene, 4-mercaptobenzoic acid, and 2-methoxybenzenethiol. Their signal intensities of Raman were adjusted by changing the concentrations of these molecules in the modification solution. Preparation of TNT Crystals. TNT crystals were prepared by the recrystallization of crude TNT powder that was supplied by National Security Department of China. The TNT powder was dissolved to obtain a saturation solution by refluxing in ethanol, and the unsoluble materials were removed by filter. After cooling to room temperature, TNT gradually crystallized from the ethanol and needlelike millimeter-sized crystals were obtained. On the other hand, 1.0 × 10−3 M TNT in ethanol was used to form good reproducible microcrystals by slow evaporation at room temperature. In detail, 10 μL of TNT solution was dropped onto a piece of clean glass or a stainless steel slide. After the ethanol evaporated completely, TNT crystallized into micrometer-sized particles on the surface of the substrates. The sizes of TNT particles were 7−8 μm by the observations of an optical microscope under a dark-field mode. To Detect the Odor of TNT Crystals. The experimental operation was performed after the lab’s door and window were closed. Millimeter-scaled TNT crystals were placed on a piece of stainless steel slide. The energy of a near-infrared 1064 nm laser was adjusted to 5 w through the change of input electric current. When the laser illuminated TNT crystals, the shape, size, and color of TNT crystal did not produce any change. Meanwhile, the AgNP papers were located at different heights above these crystals to collect the vapor molecules emanating from TNT crystals. After a set time of collection, Raman spectra were taken from the AgNP paper. To Detect the Odor of TNT Residues. Clothing, leather, envelopes, and soil were used as the target matrices spiked with TNT. Ten microliters of TNT solution with different concentrations was dropped onto these matrices and dried at room temperature. It should be noted that the diffusion area of TNT solution on clothing, leather, and envelopes was controlled in a circle district with a diameter of ∼1 cm by a very slow dipping. Therefore, the TNT residues in these matrices can be calculated quantitatively by the TNT amount per square centimeter. The content of TNT in soil was calculated by the TNT weight per gram soil. A beam of nearinfrared laser with a 1064 nm wavelength and 1 mm light spot was used to illuminate these objects. Before experiments, the laser energies endurable by these inspected objects were individually tested. The highest energies that did not destroy them were applied. Meanwhile, the AgNP papers were located at the different heights above the objects to collect the odor of TNT. After a set time of collection, Raman spectra were obtained from the AgNPs paper.

collected by swabbing the suspicious belonging and clothing, which is unreliable due to the strong adhesion of ultrasmall particulates in various matrices.5 Moreover, the nearly same molecular mass of some explosives (e.g., TNT, 227.13 vs nitroglycerine, 227.09) may lead to an ambiguous decision.6 In the past decades, a variety of sensory materials and mechanisms have been widely proposed for the supersensitive detections of TNT residues, including semiconductive polymers by fluorescence quenching,7,8 dual-color quantum dots paper by particulate visualization,9 gold nanoparticles by colorimetry,10,11 nanoparticles network by plasmonic resonance,12 Ag sol by surface-enhanced resonance Raman scattering,13 etc. In addition, the gas sensing devices for the detection of TNT vapor have also been demonstrated by the beforehand vapor sampling.7,14−16 However, an olfactory-like method to detect the odor of solid explosives in an in situ, nontouching way has yet to be established. Here, we report that the inkjet-printed silver nanoparticle (AgNP) papers detect the airborne species from crystalline TNT and its residues in open environment by a Raman switch approach and an aid of low-energy laser. The AgNPs papers modified with p-aminobenzenethiol (PABT) can efficiently capture the very few airborne molecules emanating from TNT crystals via a charge-transfer complexing reaction. Meanwhile, the interaction at an ultratrace level lights up the ultrahigh Raman scattering of PABT molecules by multiple resonance Raman enhancements, giving a supersensitive signal readout. The chemosensory concept demonstrates a novel strategy for an olfactory-like sensor as well as the rapid explosive screening.

■

EXPERIMENTAL SECTION Chemicals and Materials. Sodium citrate (99.8%), silver nitrate (AgNO3, 99%), toluene, nitrophenol (NP) and nitrobenzene (NB) were purchased from Sinopharm Chemicals Reagent Co., Ltd. 4-Mercaptotoluene, p-aminobenzenethiol (PABT), 4-mercaptobenzoic acid, and 2-methoxybenzenethiol were obtained from Sigma-Aldrich. All of these reagents were used without further purification. 2,4,6-Trinitrotoluene (TNT) and 2,4-dinitrotoluene (DNT) were supplied by the National Security Department of China and recrystallized with ethanol before use. Commercial hydrophilic cellulose paper was supplied from Shanghai Chemicals Company. Ultrapure water (18.2 MΩ cm) was produced by a Millipore water purification system. Preparation of AgNP Ink. Silver nanoparticles (AgNPs) (45 nm) were synthesized by the reduction of AgNO3 with sodium citrate. The size of AgNPs can be controlled by the ratio of AgNO3 to sodium citrate. Typically, 250 mL of aqueous solution containing 90 mg AgNO3 was first heated to boil, and then 10 mL of 1% sodium citrate was quickly injected into the above boiling solution. After refluxing for 1 h, the resultant yellow-green colloid was cooled to room temperature and filtered through a 0.22 μm Millipore membrane. The freshly prepared AgNP colloid was directly used as the “ink” for inkjet printing. Inkjet Printing of AgNPs Colloid. A common ink cartridge of the commercial inkjet printer was washed many times with ultrapure water until the ink was completely cleared away. After the vacant cartridge was dried in an oven, 2 mL of the AgNP colloid, freshly prepared as ink, was injected into the vacant cartridge. The commercial hydrophilic cellulose papers were sticked on a piece of A4 paper. Subsequently, the printing was carried out by an inkjet printer connected with a computer. 3339

dx.doi.org/10.1021/ac403409q | Anal. Chem. 2014, 86, 3338−3345

Analytical Chemistry

Article

Measurements. A near-infrared 1064 nm fiber laser (SPI, SP-20P-HS-B-A-A-A) was used to heat the TNT crystals or their residues in various matrices. The energy of laser was adjusted for different measurements by the change of the input of electric current. Raman measurements were conducted with Thermo Fisher DXR Raman Microscope equipped with a CCD detector in backscattered configuration using a 10× objective, and the Raman spectra were recorded with a 532 nm laser, 2 mW power, 50 μm aperture slit, and 2 s integral time. The AgNPs were characterized by field-emission scanning microscopy (FE-SEM, Sirion-200) and transmission electron microscopy (TEM, JEOL-2010). UV−vis absorption spectra for solution phase were recorded with a Shimadzu UV-2550 spectrometer. The reflection spectra of AgNPs papers were obtained with Shimadzu Solid Spec-3700DUV spectrometer with an integral sphere detector unit.

by-layer manner (Figure S4 of the Supporting Information). After dryness at room temperature, we can clearly see that the color of the printed surface gradually changes to uniform grass green with the increment of layer numbers (Figure 1B). Before using, the AgNP papers were stored in nitrogen-saturated water to prevent the slow oxidation of AgNPs in air. We assume that the surface of substrate membrane is completely flat and the printed AgNPs are homogeneously distributed on the surface. The concentration of AgNPs is about 0.5 nM, which is calculated using Beer’s law and the extinction coefficient.25 In the inkjet printing, 30 layers for 3.0 × 3.0 cm2 consumed about 0.6 mL of the AgNP colloid. It can be calculated that the AgNPs density is 2.0 × 1010 particles/cm2 on the surface. In accordance with the size of one AgNP (∼45 nm diameter), the coverage percent of AgNPs on the surface was only 32.0%, if the aggregation of AgNPs on the paper was not considered. That is to say, although 30 layers were printed, the state of AgNPs on the surface was still a monolayer form with some interspaces. In order to investigate the properties of printed AgNPs paper, the reflection spectra of freshly printed AgNPs paper were measured and transformed into the absorption spectra. The reflection and absorption of AgNPs papers increasingly intensified with the printing layer numbers of AgNPs (Figure S5 of the Supporting Information). As shown in Figure 1C, the AgNPs layers displayed a main absorption at ∼400 nm. In contrast to the intrinsic absorption of original AgNPs in the colloid, the absorption was obviously widened. Meanwhile, a new wide absorption appeared at the range from 460 to 750 nm when the printing layer number was up to 30 on the paper. The differences from the AgNP colloid result from the slight aggregation of AgNPs on the surface of paper. Here, the absorption enhancement implies that SERS effects of AgNP paper can be adjusted by the change of layer numbers. On the other hand, the new absorption band is very important for the spectral resonance Raman enhancement (see the text herein).23 Charge-Transfer Complexing Reaction on AgNP Papers. The Raman signals of TNT were usually weak due to its small Raman cross section13,26 and lacks the specific interaction with AgNPs. Therefore, it is impossible to directly use the bared AgNPs paper to collect the odor of TNT and detect it by the SERS approach. However, TNT is a Lewis acid with an electron-deficient aromatic ring due to three electrondrawing nitro groups. Our previous work9,27,28 has confirmed that TNT is deprotonated at the methyl group by the electronrich amines, leading to the formation of the charge-transfer Meisenheimer complex. Here, we thus modified Raman-active PABT molecules onto AgNP paper through the formation of Ag−S bonds. In fact, the complexing reaction between PABT and TNT can clearly be observed by a color mutation from colorless to purple upon the addition of PABT into TNT solution (Figure S6 of the Supporting Information). Meanwhile, the appearance of a new visible absorption with a wavelength range of 415−800 nm was also monitored by UV− vis absorption spectra (Figure 1D). It is thus expected that PABT molecules modified on AgNPs paper can specifically capture TNT species by the formation of PABT−TNT chromophores. That is to say, the PABT-modified AgNP papers may have the potential to collect the very few airborne TNT molecules in the open environment and can prevent the dynamic desorption of analyte from the paper surfaces. Multiple Spectral Resonances and Raman “On” Mechanism. Importantly, the complexing reaction can

■

RESULTS AND DISCUSSION Inkjet-Printed AgNP Papers. Surface-enhanced Raman scattering (SERS) has become a powerful tool for supersensitive chemical detection by using metal nanostructures as enhancement substrates.2,17−23 In general, the metal nanoparticles in the colloid and the assembled or microfabricated nanostructures on the solid surface have been widely adopted for the SERS-based detection. Different from these prior methods,20−23 we used a common inkjet printer to print colloidal AgNP ink on a piece of hydrophilic cellulose paper to fabricate SERS-sensitized surfaces.24 Typically, 45 nm AgNPs colloid was synthesized by the reduction of AgNO3 with sodium citrate in water (Figure S3 of the Supporting Information). The aqueous AgNP colloid was injected into a vacant cartridge of commercial inkjet printer and was printed by the control of a computer (Figure 1A), and the printing was repeated until the desired layer number was obtained in a layer-

Figure 1. The AgNPs paper and spectral resonance mechanism. (A) The AgNPs colloid was used as “ink” and printed using a commercial inkjet printer. (B) The photo of AgNP paper by inkjet printing AgNP colloidal ink on hydrophilic cellulose papers in a layer-by-layer manner. (C) The absorption spectra of the original AgNP colloid and 30 layer AgNP paper. (D) The UV−vis absorption spectra of TNT, PABT, PABT−TNT complex (in ethanol), and AgNPs paper. 3340

dx.doi.org/10.1021/ac403409q | Anal. Chem. 2014, 86, 3338−3345

Analytical Chemistry

Article

TNT and giving Raman signal readout. In the sensing scheme, a very dilute PABT (1.0 × 10−7 M) in water was used to modify AgNP paper, which made the Raman signals of PABT not nearly as detectable, although the SERS effect of AgNPs existed. After the paper captures TNT species via a charge-transfer reaction, the multiple spectral resonances lead to the remarkable Raman enhancement of PABT, and as a result, the Raman fingerprints of PABT appear and become increasingly stronger with the TNT amounts. The Raman on mechanism provides a spectroscopic approach to the detection of TNT. Sensitivity, Selectivity, and Reliability to TNT. The Raman sensitivity was first tested with TNT solution. Before the addition of analytes onto the PABT-modified AgNPs paper, very weak Raman fingerprints of PABT were detected on the paper. As expected, no change in the Raman intensity was observed after the addition of solvents such as water or ethanol on the paper. Interestingly, when 5 μL of TNT ethanol was dropped onto the paper with 30-layer printing, Raman signals of PABT clearly appeared and increasingly intensified with TNT concentrations from 1.0 × 10−13 to 1.0 × 10−6 M (Figure 3A). Obviously, the additional enhancement of PABT Raman spectra is attributed to the contribution of TNT by the formation of PABT−TNT chromophores. The assignments of

transform PABT from a nonresonance molecule into an electronic resonant state like dyes by the formation of visible chromophore. In principle, the highest Raman enhancement on the surface of metal nanoparticles is achieved in the double resonances: molecular electronic resonance and surface plasmon resonance in SERS.23,29 The double resonances are not realized for PABT alone because it is a nonresonant molecule. However, the purple PABT−TNT chromophore has a wide spectral overlapping (415−800 nm) with the plasmonic absorption of 30-layer AgNPs paper (460−750 nm), leading to the surface resonance enhancement (Figure 1D). Meanwhile, the visible absorption of PABT−TNT chromophore is centered at 529 nm. When the 532 nm laser is used to excite the Raman spectrum, the chromophore can effectively absorb the exciting frequency to bring about its electronic resonance to greatly enhance the Raman signals of PABT.23 Generally, the visible chromophores of resonance dye may provide at least 2−3 orders of magnitude of additional enhancement relative to the electromagnetic enhancement alone by metal surface.29 Together with SERS effect, these multiple spectral resonances will lead to an ultrahigh Raman scattering of PABT upon binding with TNT. On the basis of the above results, a novel sensing concept by Raman spectroscopic approach using AgNP paper to detect TNT was proposed in Figure 2 (panels A and B). While the AgNPs on paper were used as SERS substrate, the modified PABT molecules played the key roles of capturing airborne

Figure 3. The sensitivity and reliability of AgNPs paper. (A) Significant Raman enhancements of PABT were observed upon the addition of 5 μL of TNT solution with different concentrations onto the PABT-modified AgNPs paper (the printing layer number was 30). (B) 5 μL of 1.0 × 10−6 M TNT ethanol was dropped on the 10 random sites on the AgNPs paper. (C) The corresponding Raman spectra obtained at these sites. All Raman spectra were recorded with an excitation of a 532 nm laser.

Figure 2. The Raman switch concept. (A) The PABT-modified AgNP paper and its SERS response. (B) The working principle for SERS detection of TNT on the AgNPs paper. 3341

dx.doi.org/10.1021/ac403409q | Anal. Chem. 2014, 86, 3338−3345

Analytical Chemistry

Article

Figure 4. The working principle to detect the odor from TNT crystals. (A) The scheme of molecular flows near the TNT particle boundary and molecular diffusion away from the surface. (B) TNT crystals were illuminated with a beam of near-infrared laser (1064 nm) with a low energy of 5.0 w, in which the shape of crystals still remain unchanged. Meanwhile, a piece of AgNP paper was located at a defined height above TNT crystals to collect the vapor of TNT. (C) The photo of TNT needlelike crystals (taken using a digital camera). (D) The PABT Raman spectra obtained from the AgNPs papers that were located at the indicated heights to collect TNT vapor for 4 min.

Therefore, the AgNPs paper with 30 printing layers was chosen for all the following tests. The selectivity of the Raman “on” was examined by the comparison with other TNT-like analytes, including toluene, nitrophenol, nitrobenzene, and dinitrotoluene (Figure S10 of the Supporting Information). At a high concentration of 1.0 × 10−6 M, they did not cause any obvious enhancement of Raman signals of PABT. In contrary, AgNPs paper was modified with other Raman-active molecules, and the addition of TNT did not also cause any Raman enhancement of these molecules (Figure S11 of the Supporting Information). These observations confirm again the crucial roles of complexing chromophore in the Raman enhancement to PABT. On the other hand, the uniformity of sensitivity at every site is very important for a paper sensor to detect the airborne species. Ten random sites in the different districts of AgNPs paper were chosen, and 5 μL of 1.0 × 10−6 M TNT ethanol was individually dropped there (Figure 3B). The monitored Raman intensities of PABT at these 10 different sites were highly uniform (Figure 3C), suggesting the identical capabilities of different sites in the reaction with TNT and in the Raman enhancement to PABT. This should attribute to the layer-bylayer printing of AgNPs to overcome the difference of AgNP distribution on the paper. Therefore, the supersensitivity, excellent selectivity, and high uniformity of the AgNP papers imply an extraordinary potential for the collection and detection of airborne TNT species. Molecular Flows near the Surface of TNT Particles. Now, let us consider the diffusive motion of molecules near the surface of a solid particle. As shown in Figure 4A, TNT molecules occur three fluxes: the desorption from the particle

all peaks were listed in Table S1 of the Supporting Information. As shown in our previous work,23 the Raman signals of TNT at relatively high concentration may simultaneously be detected by the resonance Raman enhancement through the formation of the TNT−PABT chromophore. As compared with TNT, PABT is a highly Raman-active molecule and thus exhibits a much stronger Raman readout. The strongest peak of TNT at 1369 cm−1 became the weak left shoulder of strong PABT Raman peaks at 1392 cm−1 in Figure 3A. In most cases, however, the weak peaks of the Raman spectra of TNT molecules were usually embedded in the background of the extremely strong Raman spectrum of PABT. Moreover, Figure 3A shows that even at 1.0 × 10−13 M TNT, the Raman signals of PABT are still 1.5 fold that of the blank by the comparison of the strongest peak at 1434 cm−1 (I/I0). Thus, a sensitive limit of 1.1 × 10−14 M TNT was calculated by a linear regression equation (Figure S8 of the Supporting Information). It was observed that a droplet of 5 μL of TNT ethanol diffused to a circle area of ∼1 cm diameter on the paper, and thus the limit of sensitive density of the paper was 1.6 × 10−17 g/cm2 TNT. This suggests that the PABT-modified AgNP paper exhibits an ultrahigh Raman sensitivity to TNT. Figure S9 of the Supporting Information has shown the spectral data of the TNT sensitivity with different printing layers. In fact, the sensitivity increased with the layer number of AgNP printing. When the layer number was more than 30, however, we can see that the background Raman peaks from substrate materials were also very strong and caused an interruption in the detection, in particular, for the ultratrace analyte. In the experiments, 30 layers had the best performances with strong analyte readout and weak background signals. 3342

dx.doi.org/10.1021/ac403409q | Anal. Chem. 2014, 86, 3338−3345

Analytical Chemistry

Article

Here, needlelike crystals with a size of 1−5 mm were prepared by the recrystallization in ethanol (Figure 4C). The above operation was carried out in a lab, closing the door and window to reduce air flow to the largest extent. Surprisingly, the strong Raman fingerprints of PABT were monitored from the AgNPs paper after the collecting time of 4 min at the height of 1 cm (Figure 4D). The Raman signals were at least 6 fold those of the contrast paper and became gradually weaker when the height of AgNP paper was raised. At the height of 5 cm, the Raman enhancement relative to the contrast paper could still clearly be observed. On the other hand, the Raman enhancement was related to the collecting time until the PABT at AgNPs paper was completely reacted (Figure S12 of the Supporting Information). The saturation vapor concentration of TNT at the temperature of 45 °C is ∼171 ppb.32 In accordance with eq 3, we can calculate that the equilibrium vapor concentration is ∼10 ppt at the height of 5 cm, about 17000 fold lower than the saturation concentration. The results clearly confirm that the AgNPs paper efficiently collected the ultratrace molecules diffusing from TNT crystal surface and gave out an enhanced Raman response. Similar experiments were also carried out on the micrometer-sized TNT particulates (7−8 μm) using a smaller laser energy of only 2.6 w. After a collecting time for only 30 s, the Raman enhancement of PABT became increasingly obvious when the collecting time was prolonged (Figure S14 of the Supporting Information). To Detect the Odors of Ultratrace TNT Residues in Various Matrices. We have further developed the utility of this concept to detect the odor of ultratrace residues of TNT in various matrices. Clothing, leather, envelopes, and soil were chosen as target matrices that were spiked with the dilute TNT ethanol and subsequent dryness at room temperature. Similarly, the near-infrared laser was used to evaporate or sublimate TNT residues in these matrices by illuminating the inspected objects. The energies of laser were first tested on clothing, leather, envelopes, and soil individually because the high energy may cause the burning and destruction of objects. The sustainable laser energies were 0.67 w for clothing, 0.77 w for leather, 2.6 w for envelope, and 1.0 w for soil. Besides the difference of material components themselves, another reason is that they exhibit different properties in the absorption and reflection to the laser. Thus, the above indicated energy lasers were used to illuminate these spiked targets, respectively. Meanwhile, the freshly printed AgNP papers were located at the height of 1 cm above the targets. Detailed experiments revealed that the collecting time of 2 s was enough to produce a detectable Raman response. In the case of clothing texture with 7.2 ng/ cm2 TNT, the AgNP paper still gave out the PABT Raman signals with ∼1.0 fold stronger than those of the contrast paper (Figure 5A). Moreover, the Raman signals became much stronger with the increment of TNT residues. Similarly, 2.9 ng/ cm2 TNT on leather, 5.7 ng/cm2 TNT on the envelope, and 1.4 ppm TNT in soil also produced the remarkable Raman enhancements by the same operation (Figure 5, panels B−D, respectively). Due to the very little amount of TNT residues in these matrices, it was immediately vaporized upon exposure to the laser. Therefore, prolonging the collecting time did not produce any further enhancements of Raman. Namely, the AgNP paper can instantly probe the odor from ultratrace TNT residues by laser scanning. The first discussion point relates to the SERS paper we innovated by inkjet printing AgNPs on cellulose papers. Theory

surface (Fd), the readsorption to the particle surface (Fr), and the diffusion of molecules away from the particle (Fe), in which these three fluxes have to be balanced. There is the range of λ where the molecular desorption (Fd) from the particle and the readsorption (Fr) to the particle are equal, in most cases, and termed as the mean free path. A molecule may finally escape away from the particle by diffusion motion beyond the mean free path λ. In the case of low vapor pressure of particles, however, the number of molecules escaping away from the λ range is very few, and thus these solid particles are difficult to be directly identified by the detection of their natural odor. After the desorption of molecules from the particle surface, it is assumed that the convection and accumulation of molecules are negligible, according to Gershanik and Zeiri’s procedure.30 The molecular motion is a quasi-stationary steady-state diffusion and transport process. The solution of molecular concentration (C) at the distance H is31 C=

aC1(b − H ) + bC2(H − a) H (b − a)

(1)

where H is an independent variable between the distance a and b and C1 and C2 are the molecular concentrations at the distances of a and b, respectively. When b is an infinity away from the particle surface, C2 is prone to zero. Then, the equation is simplified as a C = C1 (2) H In theory, the vapor concentration is saturated in the range of λ because Fd is equal to Fr, as shown in Figure 4A. Therefore, when a is equal to λ, the C1 is the saturation vapor pressure Csat that is a constant value related to temperature. The eq 2 for TNT at the temperature T can thus be rewritten as CTNT(T ) =

λ TNT Csat (T ) H

(3)

At the room temperature of 25 °C, the mean free path λ of TNT is 3.18 × 10−6 m (supporting the calculation of λ). In accordance with eq 3, even at the height of 1 cm away from the TNT particle, the vapor concentration drastically reduces from 4 ppb (saturation) to 1.3 ppt in theory. Such a low concentration cannot thus be detected only using the PABTmodified AgNPs papers in such a way as an olfactory system. To Detect the Odor from TNT Crystals in Open Environment. Here, a novel technical scheme to detect the odor of TNT crystals using the AgNPs paper has been proposed as drawn in Figure 4B. The needlelike TNT crystals were cast onto a stainless steel slide. A beam of near-infrared 1064 nm laser with energy modulation was used to slightly heat the TNT crystals. Meanwhile, AgNPs papers were located above the crystals to collect the airborne TNT species sublimating from TNT crystals in the open environment. As revealed by detailed experiment, when the laser of 5 w energy continuously illuminated the TNT crystals, the shape, size, and color of TNT crystals did not have any change. We estimated that the temperature of the crystals was about 45 °C by the heat of the low-energy laser because the softening point and melting point of TNT crystal is 50 and 80 °C, respectively. At a laser energy larger than 5 w, the illuminated TNT crystals became soft or even molten. The slight heating increased the saturation vapor pressure and sped up the sublimation of TNT from the crystal surface. 3343

dx.doi.org/10.1021/ac403409q | Anal. Chem. 2014, 86, 3338−3345

Analytical Chemistry

Article

paper will greatly improve the efficiency of capturing gaseous TNT molecules. However, this will lead to the very strong background signals of PABT through the SERS effect, reducing the sensitivity of detection to TNT because the number of TNT molecules emanating from TNT crystals is very few. Due to the reactivity between TNT and PABT, a best strategy is that AgNPs paper should be modified with a proper concentration of PABT solution (1.0 × 10−7 M) to make the PABT signals just monitored in a very weak intensity. In this case, the Raman “on” of PABT via the additional resonance enhancement upon binding TNT analyte can more obviously be detected.

■

CONCLUSIONS In conclusion, we have developed an inkjet-printed silver nanoparticle paper sensor and an analytically powerful spectroscopic method for the detection of odor emanating from the crystalline explosives in the open environment, and their utility has been established for the instant detection of explosive particulate residues in various matrices using an approach similar to the olfactory system. The results reported here exhibits the possibility of overcoming the fundamental limitations of conventional gas sensors that cannot detect the odor from solid compounds with a very low vapor pressure. The achievement to detect the odor of solid explosives in the open environment attributes to the unique combination of three sensing schemes: (1) the silver nanoparticle paper has an unusual ability of capturing the airborne analyte by a specific reaction; (2) the multiple spectral resonance enhancements give out a supersensitive Raman response; and (3) the ancillary heating of low-energy laser produces ultratrace vapor. We envision that the engineering of this concept may become a powerful tool in screening explosives in airport and probing landmines. More expectations in applications are also possible in the future.

Figure 5. To detect the odors of TNT residues in various matrices. The highest laser energies endurable by clothing, leather, envelopes, and soil were used to illuminate these TNT-spiked matrices. The AgNP papers were located at a 1 cm height above the inspected objects. The Raman spectra of PABT were obtained from the AgNP papers after the collecting time of 2 s: (A) clothing (at 0.67 w), (B) leather (at 0.77 w), (C) envelope (at 2.6 w), and (D) soil (at 1.0 w).

of SERS is now largely established,17 in which most of the enhancement is attributed to the intensity of the electromagnetic optical fields near the gold or silver nanoparticles. Great efforts have been made on the fabrication of ultrahigh SERS substrates such as nanoparticle dimers,20 nanogap or nanotube arrays,21,23 and periodic 2D nanostructures.22 However, these sophisticated techniques for the control of nanostructures have to face these problems with a structural reproducibility, uniformity, and low yield. These drawbacks are not suitable for the fabrication of large-area SERS substrates to meet the requirements for collecting airborne species and producing the uniform Raman response. In this current work, the inkjet printing of AgNPs colloid as ink on cellulose paper is highly simple for the fabrication of large-area SERS substrates. The increment of layer-by-layer printing achieves the highly uniform coverage of AgNPs on the paper. The differences on the various districts of paper for analyte collection and SERS effect are completely eliminated. On the other hand, the change of layer numbers can adjust the surface plasmon properties of aggregating nanoparticles to maximize the SERS response. Similarly, various metal nanostructures such as nanocubes, nanowires, and nanotubes in the colloid can also be printed on paper according to the procedure reported here. The paperbased SERS method is very feasible and inexpensive and will greatly expand the applicable fields of SERS sensors. The second discussion point relates to the sensitive limit that involves the analyte capture and the Raman “on” mechanism. PABT molecules modified on AgNPs paper can capture the airborne TNT molecules and lead to the resonance enhancement of PABT Raman spectra. The detection sensitivity depends on the additional enhancement after binding TNT. In general, a high density of PABT molecules modified on AgNPs

■

ASSOCIATED CONTENT

S Supporting Information *

The calculation of mean free path λ and supporting figures. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by National Natural Science Foundation of China (Grants 21335006, 21175137, 61205152, and 21375131).

■

REFERENCES

(1) Röck, F.; Barsan, N.; Weimar, U. Chem. Rev. 2008, 108, 705−725. (2) Piorek, B. D.; Lee, S. J.; Santiago, J. G.; Moskovits, M.; Banerjee, S.; Meinhart, C. D. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 18898− 18901. (3) Lin, E. C.; Fang, J.; Park, S. C.; Johnson, F. W.; Jacobs, H. O. Nat. Commun. 2013, 4, 1636 DOI: 10.1038/ncomms2590. (4) Craven, B. A.; Neuberger, T.; Paterson, E. G.; Webb, A. G.; Josephson, E. M.; Morrison, E. E.; Settles, G. S. The Anatomical Record 2007, 290, 1325−1340. 3344

dx.doi.org/10.1021/ac403409q | Anal. Chem. 2014, 86, 3338−3345

Analytical Chemistry

Article

(5) Salinas, Y.; Martinez-Manez, R.; Marcos, M. D.; Sancenon, F.; Costero, A. M.; Parra, M.; Gil, S. Chem. Soc. Rev. 2012, 41, 1261− 1296. (6) Claspy, P. C.; Pao, Y. H.; Kwong, S.; Nodov, E. Appl. Opt. 1976, 15, 1506−1509. (7) Rose, A.; Zhu, Z. G.; Madigan, C. F.; Swager, T. M.; Bulovic, V. Nature 2005, 434, 876−879. (8) Zhou, Q.; Swager, T. M. J. Am. Chem. Soc. 2005, 117, 7017− 7018. (9) Zhang, K.; Zhou, H. B.; Mei, Q. S.; Wang, S. H.; Guan, G. J.; Liu, R. Y.; Zhang, J.; Zhang, Z. P. J. Am. Chem. Soc. 2011, 133, 8424−8427. (10) Jiang, Y.; Zhao, H.; Zhu, N. N.; Lin, Y. Q.; Yu, P.; Mao, L. Q. Angew. Chem., Int. Ed. 2008, 47, 8601−8604. (11) Mathew, A.; Sajanlal, P. R.; Pradeep, T. Angew. Chem., Int. Ed. 2012, 51, 9596−9600. (12) Riskin, M.; Tel-Vered, R.; Lioubashevski, O.; Willner, I. J. Am. Chem. Soc. 2009, 131, 7368−7378. (13) Liu, H. L.; Lin, D. Y.; Sun, Y. D.; Yang, L. B.; Liu, J. H. Chem. Eur. J. 2013, 19, 8789 −8796. (14) Pushkarsky, M. B.; Dunayevskiy, I. G.; Prasanna, M.; Tsekoun, A. G.; Go, R.; Patel, C. K. N. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 19630−19634. (15) Engel, Y.; Elnathan, R.; Pevzner, A.; Davidi, G.; Flaxer, E.; Patolsky, F. Angew. Chem., Int. Ed. 2010, 51, 6830−6835. (16) Spitzer, D.; Cottineau, T.; Piazzon, N.; Josset, S.; Schnell, F.; Pronkin, S. N.; Savinova, E. R.; Keller, V. Angew. Chem., Int. Ed. 2012, 51, 5334−5338. (17) Jeanmaire, D. L.; Van Duyne, R. P. J. Electroanal. Chem. 1977, 84, 1−20. (18) Nie, S. M.; Emory, S. R. Science 1997, 275, 1102−1106. (19) Kneipp, K.; Wang, Y.; Kneipp, H.; Perelman, L. T.; Itzkan, I.; Dasari, R. R.; Feld, M. S. Phys. Rev. Lett. 1997, 78, 1667−1670. (20) Lim, D. K.; Jeon, K. S.; Kim, H. M.; Nam, J. M.; Suh, Y. D. Nat. Mater. 2010, 9, 60−67. (21) Sawai, Y.; Takimoto, B.; Nabika, H.; Ajito, K.; Murakoshi, K. J. Am. Chem. Soc. 2007, 129, 1658−1662. (22) Fang, Y.; Seong, N. H.; Dlott, D. D. Science 2008, 321, 388− 392. (23) Zhou, H. B.; Zhang, Z. P.; Jiang, C. L.; Guan, G. J.; Zhang, K.; Mei, Q. S.; Liu, R. Y.; Wang, S. H. Anal. Chem. 2011, 83, 6913−6917. (24) Yu, W. W.; White, I. M. Anal. Chem. 2010, 82, 9626−9630. (25) Kanjanawarut, R.; Su, X. D. Anal. Chem. 2009, 81, 6122−6129. (26) Kneipp, K.; Wang, Y.; Dasari, R. R.; Feld, M. S.; Gilbert, B. D.; Janni, J.; Steinfeld, J. Z. Spectrochim. Acta A 1995, 51, 2171−2175. (27) Gao, D. M.; Zhang, Z. P.; Wu, M. H.; Xie, C. G.; Guan, G. J.; Wang, D. P. J. Am. Chem. Soc. 2007, 129, 7859−7866. (28) Tu, R. Y.; Liu, B. H.; Wang, Z. Y.; Gao, D. M.; Wang, F.; Fang, Q. L.; Zhang, Z. P. Anal. Chem. 2008, 80, 3458−3465. (29) Witlicki, E. H.; Andersen, S. S.; Hansen, S. W.; Jeppesen, J. O.; Wong, E. W.; Jensen, L.; Flood, A. H. J. Am. Chem. Soc. 2010, 132, 6099−6107. (30) Gershanik, A. P.; Zeiri, Y. J. Phys. Chem. A 2010, 114, 12403− 12410. (31) Grank, J. The Mathematics of Diffusion; Clarendon Press: Oxford, 1975; pp 89−90. (32) Ö stmark, H.; Wallin, S.; Ang, H. G. Propellants, Explos., Pyrotech. 2012, 37, 12−23.

3345

dx.doi.org/10.1021/ac403409q | Anal. Chem. 2014, 86, 3338−3345