Development of Rhodamine 6G Thin Filmn as a Fluorescent Sensor for Explosive Vapor Detection

With increasing public concern for possible future terrorist attacks involving novel explosives, there is a demand for advanced early detection technology. While trained canines are effective at detecting minute quantities of explosive vapors, canines also suffer from false positives, short attention spans, stress, expensive training, and require the assistance of an accompanying handler. Even with these disadvantageous, canines are currently more common in explosive detection due to conventional sensors. The extremely low vapor pressures at room temperature of most explosives limit the number of explosives molecules to be collected in a reasonable detection time pushes limits for most conventional sensors. Most of these sensor devices are big, as they require a vapor collection and pre-concentration system, and require time-consuming procedures. In addition, the concentration of explosive vapors decreases exponentially as a function of distance from the source, and as the function of time the explosive material is present in a location. The detection of trace quantities of explosives in the gas phase is very important in countering terrorist threats. Nanotechnology-enabled sensors could offer significant advantages over conventional sensors, such as better sensitivity and selectivity, lower production costs, reduced power consumption as well as improved stability. The purpose of this research is to provide a fundamental understanding of the materials and mechanisms required and aid in the development of small, inexpensive, effective portable sensor, which is capable for real time detecting explosives vapor at room temperature using only nature vapor pressure. A new array sensing system for explosive gas phase is proposed in this study. This sensor is based on a layered structure of fluorophore deposited onto a few hundred nanometers of a transparent polymer, supported by a glass slide. The fluorophores selected, Rhodamine 6G (Rh6G) and related species, are inexpensive xanthene laser dyes which are widely used as a fluorescence tracers because of their strong absorption properties in the visible light region and a high fluorescence quantum yield. The adsorption of these dyes into solid host systems can alter the photophysical properties of the dye and protect it from thermal decomposition and photobleaching, thus improving the operating and lifetime of the dye. The incorporation of fluorescent dyes into a solid host system becomes of great interest for design of photonic devices with potential applications such as solid tunable lasers and sensors. The type and concentration of aggregates depends on the conditions used for preparation of the hybrid materials. Formation of aggregates can affect the photophysical characteristics by inducing spectral shifts and band splitting. The presence of aggregates also leads to strong fluorescence quenching at high concentrations. Therefore, controlling the state of aggregation of the dye molecules in a solid matrix is a critical condition for efficiency of fluorescence sensor. The solvent system also plays a role in aggregation formation and a variety of solvents with differing polarities will be used to optimize this technology. When this fluorophore layer is applied on an appropriate substrate structure before putting on the glass substrate, such as transparent layer of polymer, a huge emission enhancement will occur. This emission enhancement can be explained by internal reflection: When the light hits on the fluorophore layer, some of the light is reflected at each interface, which allows the light to bounce along the polymer layer, this internal reflection can provide more opportunities for the incident light to be absorbed, as a result, the emission enhancement could be over a factor of 1000. This huge emission enhancement shows the potential to be used as fluorescence-based sensor with improved sensitivity, and it is sensitive enough to detect explosives with low vapor pressures under room temperature. With this new sensing system, selectivity can be improved by using a variety of fluorophores with high quantum yield to create a sensor array. Each array of fluorophores will give a distinctive response to an analyte, resulting in different response pattern for each analyte. A standard pattern for each known analytes can be used on identifying unknowns by this array response pattern. The photophysical properties of the fluorophores structure and polymer layer effect on the fluorophores needs to be further understood in order to develop this array sensor with improved sensitivity and selectivity.

canines are currently more common in explosive detection due to conventional sensors.
The extremely low vapor pressures at room temperature of most explosives limit the number of explosives molecules to be collected in a reasonable detection time pushes limits for most conventional sensors. Most of these sensor devices are big, as they require a vapor collection and pre-concentration system, and require time-consuming procedures.
In addition, the concentration of explosive vapors decreases exponentially as a function of distance from the source, and as the function of time the explosive material is present in a location. The detection of trace quantities of explosives in the gas phase is very important in countering terrorist threats. Nanotechnology-enabled sensors could offer significant advantages over conventional sensors, such as better sensitivity and selectivity, lower production costs, reduced power consumption as well as improved stability. The purpose of this research is to provide a fundamental understanding of the materials and mechanisms required and aid in the development of small, inexpensive, effective portable sensor, which is capable for real time detecting explosives vapor at room temperature using only nature vapor pressure.
A new array sensing system for explosive gas phase is proposed in this study.
This sensor is based on a layered structure of fluorophore deposited onto a few hundred nanometers of a transparent polymer, supported by a glass slide. The fluorophores selected, Rhodamine 6G (Rh6G) and related species, are inexpensive xanthene laser dyes which are widely used as a fluorescence tracers because of their strong absorption properties in the visible light region and a high fluorescence quantum yield.
The adsorption of these dyes into solid host systems can alter the photophysical properties of the dye and protect it from thermal decomposition and photobleaching, thus improving the operating and lifetime of the dye. The incorporation of fluorescent dyes into a solid host system becomes of great interest for design of photonic devices with potential applications such as solid tunable lasers and sensors. The type and concentration of aggregates depends on the conditions used for preparation of the hybrid materials.
Formation of aggregates can affect the photophysical characteristics by inducing spectral shifts and band splitting. The presence of aggregates also leads to strong fluorescence quenching at high concentrations. Therefore, controlling the state of aggregation of the dye molecules in a solid matrix is a critical condition for efficiency of fluorescence sensor. The solvent system also plays a role in aggregation formation and a variety of solvents with differing polarities will be used to optimize this technology.
When this fluorophore layer is applied on an appropriate substrate structure before putting on the glass substrate, such as transparent layer of polymer, a huge emission enhancement will occur. This emission enhancement can be explained by internal reflection: When the light hits on the fluorophore layer, some of the light is reflected at each interface, which allows the light to bounce along the polymer layer, this internal reflection can provide more opportunities for the incident light to be absorbed, as a result, the emission enhancement could be over a factor of 1000. This huge emission enhancement shows the potential to be used as fluorescence-based sensor with improved sensitivity, and it is sensitive enough to detect explosives with low vapor pressures under room temperature.
With this new sensing system, selectivity can be improved by using a variety of fluorophores with high quantum yield to create a sensor array. Each array of fluorophores will give a distinctive response to an analyte, resulting in different response pattern for each analyte. A standard pattern for each known analytes can be used on identifying unknowns by this array response pattern.

INTRODUCTION
Rhodamine 6G (Rh6G) is a cationic xanthene dye widely used as a fluorescence tracer because it has strong absorption in the visible region and a high fluorescence quantum yield. [1][2][3][4] Rh6G tends to form aggregates at higher concentration in solutions, leading to the formation of dimers 5-7 and higher order structures. [6][7][8] The solvent system also plays a role in aggregation because of the hydrophobic alkyl groups on the Rh6G molecule. 9 Polar protic solvents promote aggregation while polar aprotic solvents hinder the aggregation process, and nonpolar solvents can induce deaggregation. [9][10][11][12] Efficient procedures have been developed to incorporate Rh6G into solid matrices such as polymers, [13][14][15][16] sol-gel silicates, 17,18 and clays. 19,20 The adsorption of Rh6G into solid host systems can alter the photophysical properties of the dye and protect it from thermal decomposition and photobleaching, thus improving the operating and lifetime of the dye. 21 The incorporation of Rh6G into a solid host system becomes of great interest for design of photonic devices with potential applications such as solid tunable lasers and sensors. 14,21 The type and concentration of aggregates depends on the conditions used for preparation of the hybrid materials. 17,19,21 Controlling the state of aggregation of the dye molecules in a solid matrix is a critical condition for efficiency of a desired application. 17,[19][20][21] Formation of aggregates can affect the absorption characteristics by inducing spectral shifts and band splitting. 21 The spectral results are typically assigned to two main types of aggregates. A blue shift (relative to the monomer peak) is characteristic of an H-dimer, while a red shift is assigned to a J-dimer. 22 Therefore in solution the absorption spectra contain contributions from monomers and dimers in equilibrium, which is concentration dependent, so that deviations from the Beer-Lambert law occur. 23 The presence of aggregates also leads to strong fluorescence quenching at high concentrations. 24 This quenching process is largely due to the transfer of excitation energy from monomers to aggregates, which then decay non-radiatively. 25 Exciton theory can be applied to understand the relationship between the structure of the aggregates and their optical characteristics. [26][27][28] According to this theory, the dye molecule is regarded as a point dipole and the excited state of the dye aggregate splits into two levels through the interaction of the transition dipoles. [26][27][28] In a J-dimer the monomer transition moments are aligned end-to-end. The lower energy exciton state for the J-dimer is an allowed electric-dipole transition while transition to the upper state is forbidden, so red shifted absorption is observed and the dimer is emissive. 24,26,27 In an H-dimer the monomer transition moments are parallel, and only the higher energy exciton state is allowed. Thus, for an H-dimer a blue shifted absorption band is characteristic and the excited state is a non-emissive species. 26,27 The general dimer is referred to as an oblique dimer, as shown in Scheme 1-1. When the chromophore is adsorbed onto a substrate the dimer is characterized by an arbitrary angle, θ, between the transition moments. 26,27,[29][30][31] In the common cases discussed above, an H-dimer has θ = 90o and a J-dimer has θ = 0o. For any other value of θ, transition to both exciton excited states is allowed with transition moments: M'=√2Mcosθ (for transition to the lower energy state, E') and M''=√2Msinθ (for transition to the higher exciton state, E"), where M is the transition moment of the monomer. In addition, an excimer can also be formed by the association of excited and unexcited molecules. [32][33][34][35][36][37] Excimer formation does not affect the absorption spectrum but owing to a mismatch between the monomer Franck-Condon state and the excimer state, a red-shifted emission band with long fluorescence lifetimes are typically observed. 37,38 In this work we examine the absorption, emission, and excitation spectra for Rh6G deposited onto a glass substrate. The spectra change as a function of thickness, as expected. At submonolayer coverages the primary structures found on the surface are monomers that have the transition moment aligned parallel to the substrate surface. As the thickness increases both excimer and exciton spectroscopic features are observed. Finally, at the thickest coverage the formation of large aggregates occurs, which significantly suppresses emission.
These structural conclusions are independent of the coating method used, which include dip-coating, spin-coating, and drop casting. This is in contrast to previous reports, which only assign exciton type dimers 9 and have not identified an excimer.
Rhodamine 6G chloride (Rh6G) was supplied by Acros Organics and was used without further purification. A range of Rh6G solutions were prepared and used for thin film preparation: 1×10 -9 M -1×10 -2 M in 95% ethanol. Borosilicate glass (BSG) microscope slides were prepared by sonication in 95% ethanol and distilled water (15 minutes each) followed by blow-drying with dry nitrogen gas.
A Laurell Technologies spin coater was used for spin-casting films. 250 μl of the Rh6G solution was placed on the BSG slide and the acceleration was set to 1200 rpm and the maximum rotation was set between 400 and 5000 rpm. The thin films were then allowed to dry in air. Dip-coating was done using an MTI Corp.
HWTL-01 desktop dip coater with temperature chamber set at room temperature.
After coating, the thin film on one side of the BSG slide surface was removed using 95% ethanol. Finally, drop-coated films were prepared by placing a known volume (5 to 20 µL) of a known concentration (5×10 -7 M To 9×10 -4 M) of Rh6G on the BSG slide and allowed to air dry. All thin films were prepared at low humidity (<20%).
The absorption spectra were obtained using a Perkin Elmer Lambda 1050 spectrometer. The slit width was set to 2 nm, the wavelength range was set from 700 nm to 350 nm, and the integration time was set to 10 s for samples made with The scan area was 95 µm by 5.9 µm.
Reflection spectra were obtained using a Filmetrics F40 microscope Thin Film Analyzer and thicknesses were fit using the software provided. Optical constants for Rh6G were determined by spin-casting films onto fused silica and the wavelength dependence of n and k are shown in Fig. 1 The absorption spectra of a series of films are shown in Figure 1 However, there are also significant changes in the lineshape as the film thickness changes, indicating structural changes in the Rh6G film. As shown in the normalized spectra in Figure 1-1B, when using the lowest concentrations the film spectrum has a maximum at 525 nm and a small shoulder at ~490 -500 nm. As the concentration of the casting solution increases and the film becomes thicker, the main peak shifts to lower energy and the high energy shoulder becomes more prominent. in Figure 1-2A, λmax shifts from ~520 nm to ~560 nm between film thicknesses of < 1 nm to about 10 nm. When the film is greater than 10 nm thick, λmax and the overall lineshape does not change with increasing film thickness. Similar behavior is seen for the absorbance at λmax, where it appears that three regions of different slopes are observed: 0 -~1 nm, ~ 1 nm -~ 10 nm, and > 10 nm.
The different thickness regimes require different methods for measurement. As demonstrated in Figure 1-2, we used reflection spectroscopy, XPS, AFM, and a simple area and density method to determine the film thickness.
The data can be used empirically to determine the film thickness.
Depending upon the orientation of the Rh6G molecule on the surface, a monolayer should be 0.9 -1.4 nm thick. The break in slopes at ~ 1 nm suggests that the absorbance of films thinner than 1 nm is no longer being controlled by the thickness of the film but by the surface density of the Rh6G molecules.
The absorption spectra were deconvoluted into the component parts. Only four peaks were required to fit the observed spectra over the entire thickness range. The absorption maxima and linewidths are given in Table 1-1 and representative fits for several thicknesses are given in Figure 1-S2. Consequently, further discussion will be based on films spin-cast at 1200 rpm, unless otherwise noted. Since all of the deposition techniques used here are evaporative, the initial solution concentration always increases during the deposition process. This allows some dimerization to occur even for the most dilute solution. However, the slow evaporation process (dip-coating and drop coating) and the fast evaporation process (spin-coating) leading to similar structures suggests that some fraction of the Rh6G is deposited onto the substrate by adsorption from solution before any evaporation starts. This would account for the observation that the rotation rate in the spin-coating process has little effect on the film thickness. Also, the low concentrations of the casting solutions means that there is little viscosity change for the different solutions, which also is consistent with the spin rate having only a small effect on the resulting film thickness.
Polarized absorbance spectra were recorded to provide further understanding of the Rh6G thin film structure. With the available instrumentation, only polarization in the plane of the glass substrate (x-y) was accessible. Thus, if the Rh6G transition moment lies parallel to the surface, absorption would be observed but if the transition moment is perpendicular to the surface then no light would be absorbed. Figure 1-4 shows the polarized spectra for several different film thicknesses. For the thinnest films, Figure 1-4A, the x-y polarized spectra show absorption for the low energy peak and significant noise around the high energy shoulder. As the thickness is increased, the polarized spectra have the same shape as the unpolarized spectrum but with less intensity, figure 1-4B. With further increase in thickness, Figure 1-4C, the polarized spectra almost match the unpolarized spectra. Finally, at the highest thicknesses, Figure 1-4D, the polarized spectra is similar to unpolarized spectra at high energy, but the polarized spectrum is less intense in the low energy region.
The steady state emission spectra are shown in Figure 1-5. Similar to the absorption spectra, both the intensity and the emission wavelength maximum are strongly affected by the film thickness. As shown in Figure 1-5A, the emission maxima shift to longer wavelength as the film becomes thicker. The normalized spectra shown in Figure 1-5B demonstrate the lineshape changes. As the films become thicker the peaks become broader. All of the emission spectra could be deconvoluted into 4 peaks, with the parameters given in Table 1 Table 1-1, as found for the absorption spectra.
We measured the excited state lifetimes for the radiative decay for the films as a function of thickness, as shown in Figure  comprises the largest fraction of decay for films less than 0.7 nm thick (see Table   1-S1). The emission peak at 573 nm and the peak found at ~575 nm in the TRES with a lifetime of 4.0 ns is consistent with the assignment of an excimer. As the films become thicker, the absorption associated with M becomes small and that with A increases. The emission of the thicker films arises primarily from E and D.
Only for the thickest films does the fraction of emission from A become significant, although the absolute emission is quenched.
Thin films of Rh6G formed by spin-casting, dip-coating, and drop-coating methods exhibit the same spectroscopic and structural characteristicsthe coating method does not affect the film structure. At extremely low dye density, when the surface coverage is low, the spectroscopic signature of the monomeric structure is dominant. But even at submonolayer coverages evidence for exciton dimers and excimer formation is found since most of the Rh6G molecules have a nearby neighbor. Based on the exciton geometry, inferred from the intensities of the two peaks associated with the excited state oblique structure, the Rh6G molecules with nearest neighbors are tilted with respect to the surface at ~67°. A further increase in dye density, when the second layer starts to form, results in aggregation or crystallization. This leads to significant quenching of the emission spectrum and suppression of the monomer, excimer, and exciton emission.
Film thickness and structure can be determined by simply measuring the absorption spectrum. For the thinnest films, one monolayer or less (< 1.2 nm), the absorption maximum is found between 520 and 530 nm and the absorbance is less than 0.025. For films of 1 to 10 monolayers the absorption maximum gradually shifts from ~530 nm to ~560 nm as the films become thicker and the absorbance maximum varies between ~0.025 and ~0.14. For films greater than 10 monolayers, where crystallization may be starting, the absorption maximum is   Table 1-1. Deconvoluted peak composition (using a Gaussian lineshape) for absorption and emission spectra as a function of nominal film thickness and Rh6G concentration used. The peak positions (λ) and full width at half-maximum (Γ) are given in pairs. The uncertainties for both λ and Γ are ± 2 nm.          Table 1-S1: Fluorescence lifetimes for a range of different thickness of Rh6G thin films on glass. All decay curves are 3-exponential fits: Figure 1-S1. Refractive index and extinction coefficient for Rh6G on fused silica. These provided the parameters used to determine thickness.

INTRODUCTION
Rhodamine 6G is a well-known dye for fluorescence application due to its strong emission. Rh6G tends to self-aggregate at higher concentrations. 1,2 This could induce dramatic color changes, that is, changes of the extinction coefficient. 3,4 It could also modify the absorption characteristics such as spectral shifts and band splitting. [5][6][7] Moreover, the fluorescence quantum yield and decay time could also be decreased. [8][9][10] The first stage of the dye aggregation should be the formation of a dimer, but further increases in the dye concentration would lead to the formation of highorder aggregates. [11][12][13] In order to form the simplest aggregate, a dimer, the dyedye interaction must be strong enough to overcome any other forces that would favor solvation of a monomer. 14,15 Hydrogen bonding and hydrophobic interactions are the possible binding mechanism between Rh6G aggregates. [16][17][18] Therefore, the formation of aggregates is influenced by the hydrophobicity of the environment and the electrostatic interaction of the dye molecules. 13,[17][18][19][20] The exciton theory suggests when dimerization occurs, the two-excited states split. 19,21,22 The relative orientation of the transition moment vector of the monomeric unit in the aggregate affects the energy gap and the transition probabilities from the ground to excited states. 5,11,19,[22][23][24] A blue shift (relative to the monomer peak) is characteristic of an H-dimer, while a red shift is assigned to a J-dimer. 9,[22][23][24] An H-dimer is non emissive and normally associated with a sandwich type structure, and a J-dimer is emissive and is a head-to-tail linear structure. 9,[22][23][24] An electronically excited monomer tends to form dimers with an adjacent ground state monomer at higher concentration, forming a electronically excited state dimer and is called an excimer. [25][26][27][28] The fluorescent quantum yield decreases at high concentration, but the absorption spectrum will not change since there is no ground state association. [29][30][31][32] Similarly, a complex formed between an excited state monomer and a ground state molecule of a different nature is called an exciplex. [33][34][35] As previously reported, Rh6G thin films were spin cast onto a glass substrate using ethanol (EtOH) as a solvent. 36 The thickness of the Rh6G film was controlled by the concentration of the casting solutions, surface coverage and dye density on the surface. 36 At extremely low dye density, when the surface coverage is low, monomeric structure is dominant, excited state excimer and exciton dimer formation is also found in most of the Rh6G molecules have a near by neighbor.
Further increase in surface coverage and dye density, additional layer of Rh6G forms and aggregated Rh6G grow until eventually the aggregated Rh6G dominates the structure. 30,[36][37][38][39][40][41] Theoretically, the behavior of Rh6G molecules in solution could be similar as in the solid, changing from monomeric unit to aggregates as a function of dye density. [36][37][38][39][40][41] The purpose of this study is to investigate the structure of Rh6G molecules in different solvents at different concentrations. A series of 5 different solvents (acetone, acetonitrile (CH3CN), dimethylformamide (DMF), ethanol (EtOH) and water) were studied. A wide range of concentrations was studied. The transition from primarily monomers to aggregate was abrupt in pure solvents. Thus, a mixed solvent system of EtOH/Water was used to better control particle growth.
Rhodamine 6G chloride (Rh6G) was supplied by Acros Organics and was The fluorescence emission spectra and excitation spectra was acquired using a Horiba (Jobin Yvon) Fluorolog spectrometer. The light source used was a Xenon arc lamp. For emission spectra, the excitation wavelength was set at the absorbance maximum, the slit width was set to 2 nm for low concentration set from 700 nm to 350 nm, the cuvette pathlength was 1 mm for concentrations lower than 1 × 10 -5 M, and 0.1 mm pathlength cuvette was used for concentrations higher than 1 × 10 -5 M. For excitation spectra, the detection wavelengths used were 550 nm, 570 nm, 600 nm, and the maximum emission wavelength.
Number weighted size distribution, volume weighted size distribution, and intensity weighted size distributions were obtained by dynamic light scattering using a Malvern Zetasizer Nano instrument. Scattering optics were set at 90 degree with the cuvette temperature set at 20°C, the integration time was automatically set to get a good signal-to-noise ratio, every measurement were performed after 2 minutes waiting time to allow solutions to be at rest, and every measurement was repeated 10 times.   Rh6G molecules tend to form aggregates at increased concentrations. As shown in Figure 2-1B, increasing total dye concentration leads to an increase in the absorption intensity of the high energy shoulder. The shoulder peak intensity is clearer in water, the absorbance maximum peak shift from 530 nm to 500 nm. At the same time, the fluorescence spectra change their shape in the high concentration solution. The emission peak shifted to the lower energy side and emission at long-wavelength region appeared.
At this concentration, aggregate formation is favored in the presence of water.
In order to analyze the emission spectra more quantitatively, emission peak wavelength and intensity are plotted as a function of Rh6G concentration in Water and ethanol were chosen to study the aggregation process in detail as follows.
both Rh6G in ethanol and water, the absorbance spectra consist of a 530 nm main peak and a small shoulder peak around 500 nm as shown in Figure 2-3. The absorbance spectra shape did not change with increasing concentration in ethanol. But in water, the 500 nm peak slowly increases with increasing dye concentration.
As shown in Figure 2 In the volume distribution, the area of the peak is depending on the volume of the particle (volume of a particle sphere is equal to 4/3π(r) 3 ), and is proportional to the 3 rd power of its diameter 42,43 . In intensity distribution, the area of the peak is proportional to the sixth power of its diameter from Rayleigh's approximation. The size distribution was much wider in water compared to in ethanol 42,43 .
In ethanol, the size distribution shows the size of molecule was ranging from 0. The absorption spectra were deconvoluted into the component parts, only four peaks were required to fit the observed spectra. The absorption maxima and the line widths are given in Table 2 In monomeric solution, peak 1 at 508 nm and peak 2 at 531 nm are required to fit the spectra that account for vibronic shoulder and monomeric absorbance peak. The ratio between these two peaks are 509 nm / 531 nm = 0.88.
As formation of dimer starts, the 531 nm peak slowly decreases, with formation of new peaks at 499 nm and 508 nm. The 499 nm peak dominates in the most concentrated Rh6G in water.
A particularly interesting feature is the change of band shape of the fluorescence emission for 1 x 10 -3 M Rh6G in water, and this is due to detector saturation. The emission spectra can no longer be obtained at this concentration.
As shown in Figure 2 Table 2-2. At low concentration, majority of Rh6G are in the solvated monomeric state, have the absorbance peak at 531 nm and small shoulder at 508 nm. In the excitation spectra, there is a major peak at 531 nm, and smaller peak at 508 nm in excitation spectra, the absorption and excitation spectra superimpose as shown in Figure 2 The monomeric emission decreases as the concentration goes up as a result of formation of dimer, the 499 nm absorbance peak is associated with the dimer absorbance, and it is non emissive, which means it is a H-type dimer.
In this work, the solvent and concentration effects in the absorption and fluorescence spectra of the Rh6G were studied. A full Rh6G concentration range in five different solvents, from the solubility limit to highly dilute systems, was studied in detail.
A very small change of the spectrum was observed in all solvents at low concentration.
When the dye concentration increases in water, a significant change is observed in both absorption and emission.
A careful analysis of the spectra revealed the formation of molecular dimers at high concentration. This is indicated by the appearance of a second absorbance spectral band, which is red-shifted with respect to the fully solvated, isolated monomers. The formation of dimer is also studied by the size change from monomeric dominating solution to dimer dominating solution. It shows that the monomeric unit has an average size of 1.0 ± 0.2 nm, and average size of 3.25 ± 0.6 nm for the dimer dominated solution.
The ratio of monomer to dimer structure can be altered either by increasing the concentration in water, or increasing the water solvent ratio in the mixed solvent, and this provides the tunability between 550 nm to 610 nm for emission wavelength. The type of dimer formed in water is H-type non-emissive dimer.

INTRODUCTION
The surface morphology of polymer films can have a strong influence on the interfacial properties and has been studied for many years in terms of adhesion. 1,2 Of more recent interest is the ability to pattern polymer thin films that then can be used as substrates for a variety of diverse areas, ranging from microelectronics 3-6 to understanding cell growth. 7,8 One common method for patterning a polymer surface is to induce wrinkling, 9,10 which, ideally, induces a periodic series of ridges and valleys. Wrinkles have been proposed to be used for stretchable electronics, 5,6 microlens applications, 11 and measurement of thin film mechanical properties. [12][13][14][15][16][17] Wrinkles on polymer thin films can be formed spontaneously and is done by a variety of methods. 9,10,18-26 Typically these are multistep processes: a template is produced using a mechanically or thermally induced compressive stress (generally poly(dimethylsiloxane), PDMS) and then the template is used to transfer the wrinkle pattern to another substrate. The other common approach is to use block co-polymers, which aggregate into different phases that create the wrinkle. [27][28][29][30][31][32][33][34][35] A compendium of the different patterning approaches can be found in the comprehensive review by Rodríguez-Hernández. 10 It has been reported that spin-casting can be used to create wrinkles in selected polymer systems. 28,35,36 In this research we are able to demonstrate wrinkle patterns in poly(methylmethacrylate) (PMMA) films by simple spincasting without any templating or additional field to induce the periodic deformation. PMMA is one of the polymers used to establish the foundational parameters for spin-casting. [37][38][39][40][41] In developing models to describe spin-casting, it was always assumed that the films had a uniform thickness. However, it was later recognized that solvent-polymer interactions could be used to influence surface roughness, [42][43][44] but no periodicity was reported. The spin-cast PMMA forms wrinkle pattern with wavelengths on the order of tens of microns and with amplitudes of tens of nanometers. These are the correct length scales to create a photonic structure that could lead to interesting properties. [45][46][47][48][49][50][51] This work is aiming at understanding the parameters needed to create wrinkle patterns in polymer thin films. The periodicity of the wrinkles correlates with polymer thickness as controlled by changing the concentration of PMMA in the spin-casting solution. These data will be used to develop a model to predict wrinkle patterns. Finally, the study of wrinkle patterns act as a photonic structure and can influence the light emission of rhodamine 6G (Rh6G), a common laser dye will be investigated in the next paper.
Microscope glass slides were cut into dimensions of 3.75 cm x 1. The edge of the films tend to be thicker than the center of the films at this rotation rate, so all the thickness were measured at the center of the film.
Spin-casting has been used to create polymer films for many years. The virtues of the method are that it is simple, highly reproducible, and gives films of controlled thickness. The thickness of a spin-cast can be estimated by where is the mass fraction of polymer in the initial solution, is the viscosity of the initial solution, ∞ is the mass fraction of pure solvent in the gas phase at equilibrium, = rotational velocity, Dg is the diffusivity of the solvent in the overlying gas, g is the kinematic velocity of the overlying gas, is the solution density, psol is the vapor pressure of the pure solvent, Mpol is the molecular weight of the polymer, R is the ideal gas constant, T is the absolute temperature, and c = constant. Equation [1] demonstrates that controlling the thickness of the polymer film is complicated and can be achieved multiple ways: changing the concentration of the initial solution or changing the spin rate will have the biggest effect but changing the solvent, polymer molecular weight, temperature, or the composition of the overlying gas can all have small effects.
Primary control of the polymer thickness is attained either via spin rate or solution concentration.  3 , where the higher order terms in the brackets in equation [1] have been truncated. Again, the fit is reasonable except for the lowest concentration. This will allow us to determine other parameters in equation [1]. More importantly, equation [1] can serve as a starting point for understanding the details of the formation of the PMMA films.
A closer examination of the surfaces of the PMMA films shows that there is a periodic wrinkle observable in the film.   First, the wrinkles only started to form at the concentration of 1% (w/v), the thickness of this PMMA film is ~160 nm. Second, the wavelength of the wrinkles is longer at the edges than in the center of the film, for sample 3, the wavelength is ~80 m, significantly higher than found in the center of the film (~70 m). Third, the periodicity is reasonable constant over the millimeter length scale. Finally the amplitude of the periodic structure is on the order of nanometers, and the amplitude is increasing as the film thickness increases, the 1%, 2% and 3% PMMA film has the amplitude as 1.0 nm, 9.8 nm, and 21 nm respectfully. Finally, in case of the 3% w/v solution, which about 315 nm thick, the amplitude varies noticeable across the film: center is 13.5 nm, more cross to the right is 20.9 nm, and at the right edge is 37.5 nm. In contrast, the 2% w/v solution, which is about 200 nm thick, the amplitude is much more consistent.

INTRODUCTION
With the increasing public concern for possible future terrorist attacks involving novel explosives, there is a demand for advanced early detection technology. [1][2][3][4] The extremely low vapor pressures at room temperature of most explosives limit the number of explosives molecules to be collected in a reasonable detection time pushes limits for most conventional fluorescence sensors. [1][2][3][4][5][6] The underlying idea behind analyte fluorescent detection is simple. 5,6 Fluorophore emissions are measured before and after analyte sample exposure. 7 The emission intensity can be quenched, enhanced or unchanged. 7 There are three distinct mechanisms underlying these emission intensity changes: collisional quenching, aggregation changes and energy transfers. [8][9][10] Collisional quenching occurs when analyte molecules collide with a fluorophore in the excited state. [11][12][13][14][15] Additional nonradiative pathways become available for the fluorophore to lose its energy, resulting in a decreased amount of emitted light, thus the emission intensity is quenched. [11][12][13][14][15] In the case of aggregation changes, the analyte can cause a structural change in a fluorophore by either inducing aggregation or by disrupting aggregation, which can lead to either enhancement or quenching of the emission. [19][20][21][22][23][24][25][26] Energy transfer upon exposure to the analyte occurs when the energy transfer from higher energy excited state is disrupted causing a quenching of the emission intensity. [27][28][29][30] To improve emission signal output, several methods have been used to amplify the fluorescent signal, such as metal-enhanced fluorescence, which is a common approach but making reproducible metallic surfaces is challenging. 28,29 Other methods include incorporating the fluorophore into a solid matrix such as polymers and clays, but the photophysical properties of the fluorophore might be changed. [30][31][32][33][34] A new sensing system for gas phase explosive is proposed in this study.
This sensor is based on a layered structure of fluorophore deposited onto a few hundred nanometers of a transparent polymer, supported by a glass slide, as shown in Scheme 4-1. The fluorophore selected is Rhodamine 6G (Rh6G), which is an inexpensive xanthene laser dye, widely used as a fluorescence tracer because of its strong absorption properties in the visible light region and high fluorescence quantum yield. [35][36][37] A monolayer thin film deposited onto a substrate could have several structures, such as ground state monomer and aggregates, excited state excimer and dimer. [38][39][40][41][42] As reported previously, Rh6G thin films were spin cast into glass substrate using EtOH as a solvent. The thickness of Rh6G film was controlled by the concentration of the casting solutions, surface coverage and dye density on the surface. [38][39][40][41][42] At extremely low dye density, when the surface coverage is low, monomeric structure is dominating. [38][39][40][41][42] With increasing dye density, the formation of excited state excimer and excitonic dimer can occur since most of the Rh6G molecules have a nearby neighbor. [43][44][45][46][47][48][49][50][51][52] Further increase in surface coverage and dye density, additional layer of Rh6G forms and aggregated Rh6G grow until eventually the aggregated Rh6G dominates the structure. 44,[56][57][58][59][60][61] When this Rh6G layer is applied on PMMA layer before putting on the glass, a huge emission enhancement will occur. 7,59 This emission enhancement can be explained by internal reflection, as shown in Scheme 4-2: When the light hits on the fluorophore layer, some of the light is reflected at each interface, which allows the light to bounce along the polymer layer. [60][61][62][63][64][65][66] At each bounce on the polymer/fluorophore interface, the light that escapes excites more fluorophore, effectively increasing the efficiency of the use of the excitation light. [60][61][62][63][64][65][66] This huge emission enhancement shows the potential to be used as a fluorescencebased sensor with improved sensitivity, and it is sensitive enough to detect explosives with low vapor pressures under room temperature. [67][68][69][70] In this work we examine the absorption and emission spectra for Rh6G deposited onto glass substrate directly and with a PMMA layer in between. The thickness of the fluorophore will be a key feature in determining the sensitivity of the sensor. The role of the transparent PMMA film interface on Rh6G properties is studied in detail, shows the thickness of the PMMA layer affects the internal reflection, and the surface morphology of PMMA film affects the Rh6G structure by reducing the amount of aggregation of Rh6G. Finally, the reaction mechanism between Rh6G and TNT vapor is examined and studied, the relationship between Rh6G structure and types of mechanism leads to a better mechanistic understanding of the sensor platform.
Rhodamine 6G chloride (Rh6G) was supplied by Acros Organics and was used without further purification. A range of Rh6G solutions were prepared and used for thin film preparation: 1 × 10 −9 M − 1 × 10 −2 M in 95% ethanol.
The polymer used is polymethylmethacrylate (MW ∼120,000, PMMA) and was obtained from Sigma Aldrich. PMMA was dissolved in toluene to form a 4% (w/v) solution.
Borosilicate glass (BSG) microscope slides were prepared by sonication in 95% ethanol and distilled water (15 min each) followed by blow-drying with dry nitrogen gas.
A Laurell Technologies spin coater was used for spin-casting films. A 250 µL portion of the Rh6G solution was placed on the BSG slide, and the rotation was set to 1200 rpm. The thin films were then allowed to dry in air. A 250 µL portion of the poly(methylmethacrylate) (PMMA) solution was placed on the BSG slide, and the rotation was set to 1200 rpm. The three layered samples were prepared by spin-coating the polymer onto the glass slide followed by spin coating the Rh6G solutions into the polymer layer. All thin films were prepared at low humidity (<20%).
The absorption spectra were obtained using a PerkinElmer Lambda 1050 spectrometer. The slit width was set to 2 nm, the wavelength range was set from  In order to understand this emission enhancement, Rh6G structure, role of the interface and PMMA thickness are examined in more detail. The thickness of the Rh6G thin film was measured by reflection spectroscopy, and the formulas given in previous report was used to estimate the Rh6G film thickness, both methods agree with each other on the thickness determination.
The Rh6G film thicknesses are 0.7 nm, 1.0 nm, 1.7 nm, 2.6 nm and 9.2 nm respectfully. The thickness of PMMA film measured by reflection spectroscopy is 385 nm.
As the Rh6G casting concentration increases, the films became thicker and the spectra became more intense. Based on assignments from previous study, The absorption spectra of Rh6G on glass and PMMA are similar: both have a low energy peak maximum and a higher energy shoulder. The peak maximum is shifted to the lower energy in the PMMA samples with Rh6G thickness smaller than 1.0 nm, as shown in Figure 4-2A and 4-2B, peak maximum on glass for 0.7 nm Rh6G is 522 nm while on PMMA is 537 nm, and peak maximum on glass for 1.0 nm Rh6G is 522 nm, while on PMMA is 541 nm. This peak maximum is shifted to the higher energy in the PMMA sample when Rh6G is thicker than 1.0 nm as shown in Figure 4-2C to Figure 4-2E, peak maximum for 2.6 nm Rh6G on glass is 560 nm while on PMMA is 548 nm, and peak maximum for 9.2 nm Rh6G on glass is 563 nm while on PMMA is 555 nm.
At extremely low dye density, the surface coverage of this film is low. The spectroscopic signature of the monomeric structure is dominant in both glass and PMMA samples. As shown in Figure 4-2A, the absorbance maxima for Rh6G on PMMA is 537 nm, while on glass is 521 nm, which can be attributed to the polarity difference in the two surfaces.    For Rh6G film below one layer, the number of molecules on glass surface is much larger than on PMMA surface, thus the distance between two Rh6G molecules on glass is smaller than on glass, as a result, PMMA surface is less in favor for aggregates at this surface coverage.
Since further increases in casting solution concentration will start to form the second layer, Figure 4-3C is the bilayer Rh6G on PMMA and glass, and   The red dots are the wrinkle amplitude as a function of PMMA thickness.
As the PMMA thickness increases, the wrinkle amplitude increases. As the wrinkle amplitude increases, the Rh6G emission intensity decreases as shown in The amplitude of the wrinkle is 1 nm. This amplitude is lower than 10 nm, allowing quantum confinement effect to occur. As shown in Figure 4-9B, for the Rh6G thin film with thickness less than 0.9 nm: τ1= 4.0 ± 0.6 ns, τ2 = 1.9 ± 0.4 ns. For the Rh6G film thickness between 0.9 nm to 1.0 nm, the two lifetimes became longer (τ1 = 6.4 ± 0.1 ns, and τ2 = 2.7 ± 0.3 ns), and a new lifetime with τ3 = 0.8 ± 0.2 ns appeared. For the Rh6G film thickness above 1.0 nm, second layer starts to form, the τ1 and τ2 became same as in sub-monolayer thin film, but the 3 rd life time remained as 0.8 ± 0.2 ns. Based on the previous paper, the τ1=4.0 ± 0.6 ns is assigned to the excited state excimer, τ2 = 1.9 ± 0.4 ns is assigned to the monomer, and τ3 = 0.8 ± 0.2 ns is assigned to the aggregate. It shows for the Rh6G on the PMMA surface, at extremely low dye density, the ground state monomeric and the excited state excimer are dominating.
Since B1 is significantly larger than B2, suggests that PMMA surface morphology enhances the excited state excimer rather than simple ground monomer. When Rh6G molecules deposited on the PMMA surface, they are not distributing evenly across the surface. Instead, the molecules tend to accumulate together at a proximity close enough to allow excited state excimer to occur, but too far to form aggregates.
With the dye density increases, right around when the surface coverage reaching the saturation stage, the excited state excimer reaching the maxima, and resulting in the phase transfer, allowing maximum excited state excimer emission.
When second layer of Rh6G starts to form on top of the first Rh6G layer, emission from aggregate structure dominates. In Figure 4-11B and 4-11C, is the monolayer with high coverage, exciton dimers and excited state excimer is presenting in the film. The existence of many radiative and non-radiative pathways shielding the emission change caused by TNT collision. At the same time, the amount of monomeric structure has decreased.
In Figure 4-11D, the second layer starts to form, results in aggregation.
TNT causes a structural change in this film, disrupting aggregation, thus the enhancement of the emission can occur. For Rh6G on glass, either the extremely low surface coverage monolayer thin film or bilayer with the second layer just starts to form could be used in the TNT detection. As the three-layered structure enhancing the emission intensity as the amount of Rh6G is decreased. This is an advantage for improving sensitivity.
Since only a small amount of Rh6G is required for a strong signal, only a small amount of analyte would be required to modulate the signal.
Rh6G was put on the 385 nm PMMA layer to amplify the emission intensity, emission change was observed before and after the exposure as shown in Figure 4-12.
All Rh6G films showed an emission decrease upon exposure to TNT except the thinnest sub-monolayer film as shown in Figure 4-12 A. At this surface coverage, the molecular density is too low, the interaction between monomeric     All decay curves are 3-exponential fits: I(t)=B1e -t/τ 1 ＋ B2e -t/τ 2 ＋ B3e -t/τ 3, the preexponential factor as a fraction as function of Rh6G film thickness.