Plasmonic Nanotrough Networks for Scalable Bacterial Raman

Subscriber access provided by UNIV OF DURHAM

Functional Nanostructured Materials (including low-D carbon)

Plasmonic nanotrough networks for scalable bacterial Raman biosensing Ran Zhang, Yan Hong, Bjoern Reinhard, Pinghua Liu, Ren Wang, and Luca Dal Negro ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.8b07640 • Publication Date (Web): 27 Jul 2018 Downloaded from http://pubs.acs.org on July 28, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 33 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

Plasmonic nanotrough networks for scalable bacterial Raman biosensing Ran Zhang2, Yan Hong 4,5, Bjoern. M. Reinhard4, Pinghua Liu5, Ren. Wang1 and Luca. Dal Negro 1,2,3 * 1 Department of Electrical and Computer Engineering & Photonics Center, Boston University, 8 Saint Mary Street, Boston, Massachusetts 02215, United States 2 Division of Materials Science and Engineering, Boston University, 15 Saint Mary’s Street, Brookline, Massachusetts 02446, United States 3 Department of Physics, Boston University, 590 Commonwealth Avenue, Massachusetts 02215, United States 4 State Key Laboratory of Electronic Thin-Film and Integrated Devices, University of Electronic Science and Technology of China, Chengdu, 610054, China 5 Department of Chemistry & Photonics Center, Boston University, Boston, Massachusetts 02215, United States

KEYWORDS Electrospinning, Nanotrough, SERS, Bacterial sensing, Cellulose

ACS Paragon Plus Environment

1

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 33

ABSTRACT

We demonstrate a novel approach for fabricating Surface Enhanced Raman Scattering (SERS) substrates for single bacterial biosensing based on Ag Cylindrical Nanotrough Networks (CNNs). This approach is developed with large scalability by leveraging a cellulose nanofiber template fabrication via facile electrospinning. Specifically, a concave nanotrough structure consisting of interconnected concave Ag nano shells is demonstrated by depositing a thin layer of Ag atop a sacrificial electrospun nanofibers template and then completely removing the cellulose core in water. Our investigations of the scattering properties and SERS performances of single isolated Ag nanotroughs of different diameters reveal that nanotrough-based substrates provide tunable optical responses and enhanced SERS intensities. Further, Ag CNNs are fabricated in highly-interconnected networks that yield reproducible SERS signals for molecular monolayers and whole bacterial cells, enabling rapid spectral discrimination between different bacterial strains. Finally, by performing Principal Component Analysis (PCA) on a large number of measured SERS spectra (40 spectra per bacterium) we demonstrate successful spectral discrimination between two types of Escherichia coli (E. coli) bacteria, i.e., E. coli K12 with its derivative E. coli DH 5α and E. coli BL21(DE3). The demonstrated cost-effective substrates feature several advantages over conventional SERS substrates including environmentally friendly and scalable fabrication compatible with versatile devices and provide an alternative approach to rapid SERS detection and screening of biochemicals.


2

Page 3 of 33 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1. Introduction Since its discovery nearly 40 years ago, Surface-Enhanced Raman Scattering (SERS) has stimulated a great research effort ranging from analytical chemistry to life science and environmental science

1-3

. In SERS spectroscopy the intensity of individual vibrational signals

can be enhanced up to 10 orders of magnitude when molecules are placed on or near a rough metal surface 3. Its superior sensitivity down to the single molecule level and its capability to provide fingerprinting spectra makes SERS a very powerful and non-destructive approach for rapid molecular detection and identification

4

. Excitation of localized surface plasmon

resonances on roughened metal substrates and the accompanying strong electric (E-) field localization was shown to provide the major contribution to the SERS enhancement effect

5,6

.

Therefore, extensive efforts have been made to design and control different nanostructures of noble metals (mainly Au or Ag) ranging from 2D aperiodic particle arrays nanoholes, nanodiscs, nanorings, and nanopetals

12-18

7-11

to nanopillars,

to nanoparticles assembled in 3D

templates, and corrugated bipyramidal microcrystals 19-25. These SERS substrates are commonly fabricated using controllable methods like electron beam lithography, nanosphere lithography, and focused ion beam patterning and produce ensemble-averaged Raman enhancement factors >107 with good reproducibility. However, the high cost, lack of scalability, and poor integration with biomedical devices greatly limit their use to specialized spectroscopic applications

26

.

Therefore, it is compelling to develop plasmonic active functional materials with tunable and enhanced SERS responses using inexpensive fabrication methods that guarantee the integration into versatile devices over a relatively large area in a simple and cost-effective manner.


3


Page 4 of 33

In the past few decades, great efforts have been made by many groups to make SERS substrates suitable for scalable devices by depositing metal structures on PDMS or PET using Ebeam evaporation or sputtering

25,27

. Among various proposed solutions, polymer-based

nanofibers are interesting because of their mechanical strength, flexibility, and bio-compatibility, particularly when combined with cost-effective electrospinning that does not require the use of hazardous chemicals

28-32

. This approach has not only attracted increasing attention for

applications to light emission, confinement, guiding, and fluorescence sensing, but it also holds the potential to be applied to SERS substrates for molecular detection 33-36. For example, He and colleagues developed scalable SERS substrates by incorporating metal nanoparticles in electrospun nanofibers

37

. However, their approach uses the nanofibers as a matrix to support

material nanostructures, which limits their sensing volume to only small molecules and it is potentially affected by the working environment and the swelling ratio and degradation of the polymer. Camposeo and colleagues deposited gold nanorods on the surface of electrospun polymer nanofibers to enhance the intensity of Raman scattering 38. However, the need for highdensity nanorods poses a significant challenge for the development of high-throughput SERS substrates. Metal coated templates based anodic aluminum oxide (AAO) nanochannels or electrospun fibers have also been investigated by researchers

39, 40

. However, all these studies

have not demonstrated the successful removal of the template material without effecting the sensing structure, which significantly reduces the integration of SERS substrates with costeffective devices. In recent years, Cui’s group has developed a new type of transparent conductive oxide (TCO) surface by using an electrospinning fiber network as a template, and showed excellent conductivity, robustness and transparency when coupled to various


4

Page 5 of 33 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


substrates41. However, plasmonic optical and sensing capabilities have not been reported for these types of novel alternative transparent electrodes. In our paper, by successfully removing the template materials and thus creating plasmonic tunable highly interconnected structures, we successfully demonstrate networked Ag nanotroughs (concave Ag nano shells) using controllable electrospun nanofibers and we develop a novel type of scalable, cost-effective plasmonic functional materials that can be successfully utilized as SERS substrates. In our previous work, high-quality nanofibers were synthesized from solution-based processing cellulose, which is the most abundant organic polymer in the world. Hydroxypropyl cellulose (HPC), a cellulose derivative that is soluble in both water and organic solvents, was utilized to fabricate the HPC nanofiber template via electrospinning

30-32

. We

discovered that after coating a single HPC nanofiber with a thin layer of Ag (~30 nm) via thermal evaporation followed by the dissolution of the polymer core in water, a concave metal nanotrough structure remains firmly attached to the substrate (Figure 1). Based on this principle, instead of a single isolated nanofiber, a non-woven network of cellulose nanofibers was used successfully as a sacrificial template for Ag deposition and fabrication of Ag cylindrical nanotrough networks (CNNs). The high degree of interconnectivity among the nanofibers in the film gives rise to a stable and tunable plasmonic response that can be harnessed effectively for SERS-based applications, and it also provides a suitable environment for the attachment of bacterial cells. In this paper, we characterize the optical scattering spectra and Raman signal intensity enhancement as a function of the nanotrough diameter on a single nanotrough and on randomly networked structures. Finally, by utilizing the Principal Component Analysis (PCA) statistical method on a large number of measured SERS spectra (40 spectra per bacterium) we demonstrate successful spectral discrimination between two types of Escherichia coli (E. coli)


5


Page 6 of 33

bacteria, i.e., E. coli K12 with its derivative E. coli DH 5α and E. coli BL21(DE3). Our results offer an alternative approach for a rapid bacterial diagnostics based on portable devices, which is of great importance for improving the treatment outcomes of serious infections and ensuring the appropriate use of antibiotics 42, 43.

2. Experimental Section 2.1 Fabrication of Ag CNNs. Specifically, HPC (purchased from Sigma, MW: 100,000 g/mol) was dissolved in water to result in a 50 % solution. The solution was then transferred into a 10 mL syringe (diameter 18.94 mm) with a 20-gauge stainless steel needle. Then the syringe was connected to an infusion syringe pump (Braintree Scientific) and its needle to a high power source station (gamma high voltage). The HPC precursor was driven continuously to the needle nozzle by the syringe pump with a stable flow rate of 0.4 mL/h. Due to the high voltage (18 kV) applied between the nozzle and the metal plate collector, the charged HPC precursor was propelled from the nozzle and nanofibers were created during the flight after the evaporation of the solvent. The charged nanofibers gathered at the ground collector (kept 20 cm from the nozzle) and were neutralized and formed a random nanofibers network. Subsequently, the asprepared random nanofibers network was coated with a thin layer of Ag (~30 nm) via thermal evaporation. The cellulose core of the nanofibers was removed by immersion into water, thus leaving only the Ag nanoshell on the substrate, as shown in the SEM image (Figure 1 inset). The morphology of the nanofibers network was preserved even in the case of a 30nm-small thickness of Ag nanoshells, resulting in a highly interconnected network after removal of the cellulose. (Figure 1).


6

Page 7 of 33 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


2.2 Dark-field scattering characterization of Ag cylindrical nanotroughs. Scattering images of the Ag nanotroughs were recorded using an upright microscope (Olympus BX51 WI). A 100 W tungsten halogen lamp with unpolarized white light fitted with an air dark-field condenser in transmission mode was utilized in order to optically characterize the scattering behavior of the sample. Light scattering from the Ag nanotroughs was collected using a 60X objective (NA = 0.65) and dark-field images were acquired by a digital camera with 620×580 pixels active area. The microscope also featured a 150 mm focal length imaging spectrometer (Acton Research, InSpectrum 150) together with a back-illuminated CCD detector (Hamamatsu INS-122B) that allowed us to spectrally analyze the scattered light signal with a 150 lines/mm grating. We corrected the measured scattering spectra by subtracting the background signal originated from an adjacent area of equal size and devoid of any structures. We also corrected the scattering spectra by the excitation profile of the white light illumination source. 2.3 SERS measurements. A Renishaw Raman microscope (model RM-2000) with ~2λ spatial resolution was used to measure the scattering spectrum excited by a 532 nm diode laser. The frequency was calibrated by considering the 520 cm-1 silicon phonon mode. To characterize SERS performance on the Ag cylindrical nanotrough, pMA was used. Samples were immersed in 4 mM pMA ethanol solution for 2h before rinsed with ethanol and dried with a flow of nitrogen gas. A 100X objective was used for collecting the SERS signal. SERS spectra were measured with incident laser powers at 16.1 µw and each spectrum was averaged from three 100 s exposures on the same spot. 2.4 Bacteria growth and sample preparation. Specifically, three bacterial strains (E. coli K12, E. coli DH 5α and E. coli BL21(DE3)) were cultured in the lysogeny broth (LB) overnight for 15h. Then 1 mL of each of solution was taken and washed with saline (0.9% Sodium


7


Page 8 of 33

Chloride aqueous solution) 3 times by centrifuging and then dispersed in 1 mL of saline solution. The Ag CNNs samples were then immersed into the different bacterial suspensions in saline solutions for 1 h and then washed with DI water for three times before SERS measurement.

3. Results and discussion 3.1. Fabrication of Ag cylindrical nanotrough structure. Electrospinning technology is a versatile nanofabrication method capable of generating fibers with a diameter ranging from a few tens to hundreds of nanometers using a number of materials

30-32

. In our approach, a special

cellulose derivative was chosen for its solubility in water and other organic solvents. Specifically, 50% HPC aqueous solution was poured into a syringe fitted with a stainless steel needle. Driven by the syringe pump, the HPC precursor was transferred continuously to the needle tip at a constant flow rate. Under the high voltage (18 kV) applied between the nozzle and the metal plate collector, the charged precursor was extracted from the nozzle and thinned into a nanoscale structure during the flight. The charged nanofibers were further got neutralized at the grounded collector, forming either a thin layer of individual nanofibers or random non-woven nanofiber networks, depending on the different deposition time (Figure 1b). Generally, the grounded collector can be any conductive substrate such as aluminum foil, a silicon wafer, or indium tin oxide (ITO) substrate. Afterward, a thin layer of Ag (~30 nm) was deposited onto the nanofibers via thermal evaporation. Because of the deposition directionality of thermal evaporation and the resulting shadowing effects, Ag tended to coat only the top side of the polymer nanofibers. After the cellulose portion was dissolved in water, only the concave Ag nano shell was left on the substrate (i.e., the Ag nanotrough structure). This is demonstrated in


8

Page 9 of 33 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


the scanning electron microscopy (SEM) image in Figures 1c and d, which show an individual HPC nanofiber before deposition and an individual Ag nanotrough after depositing a thin layer of Ag and the removal of the HPC core. In order to directly demonstrate the nanotrough (shell) structure, we have cut our sample across the Ag nanotrough. A hollow concave structure was clearly revealed on the edge of the scratch across the nanotrough, which also proved the complete removal of the cellulose core.

3.2. Influence of the diameter (D) of a single Ag nanotrough on the scattering spectra. In order to characterize the optical response of the single isolated Ag nanotrough, we measured the scattering properties of Ag nanotroughs as shown in Figure S1 with different morphology using dark-field scattering microscopy by using 100 W tungsten halogen lamp as standard setup

30-32

.

Figure 2a shows representative scattering spectra measured on individual Ag cylindrical nanotroughs with varying diameters collected at 10 different randomly chosen positions on the sample. The measured scattering peak resonance wavelengths of nanotroughs were plotted as a function of Ag half-shell diameter (D), and the error bar was calculated by measuring the diameter of nanoshell at different positions along the same nanotrough, as shown in Figure 2b. In particular, figure 2b demonstrates a red shift of the scattering peak from 494 nm to 687 nm as D is increased from 183 nm to 621 nm. This significant red shift of the spectral response of the Ag nanoshell is due to the plasmonic response of the film, which was also confirmed by our 3D finite-difference time-domain (FDTD) simulations

44

. Specifically, we modeled a long (5

microns) Ag nanotrough structure placed atop the ITO substrate with varying external diameter from 100nm to 500nm in increments of 100 nm and fixed thickness of 20 nm. In the idealized model, Ag nanoshells of smaller diameters correspond better to measured data. This could be


9


Page 10 of 33

partly due to single Ag nanotrough’s deformation during it formation, which render the measured diameters larger, or to the unavoidable structural fluctuations at nanotrough level. The results are reported in Figure 2c, where the matching color scheme with Figure 2a has been chosen to facilitate the comparison of the two results, showing both the validity and the limitation of using perfect semi-circular nanoshell as a simplified model. In order to obtain the simulated scattering spectra, in the FDTD method, computed nearfield results are transformed to the far field, where scattered intensities are obtained. In the commercial FDTD software package used, the incident plane wave is automatically filtered out (total field = incident + scattered) during the calculation of scattered intensity, giving scattering cross-sections purely based on the scattered field. We also used a normally-incident and linearly polarized plane wave illumination with an electric field in the transverse plane (perpendicular to the long-axis of the semi-cylindrical nanoshell). In our simulations we also included the effect of the ITO substrate considering its refractive index measured using spectroscopic ellipsometry (n = 1.8). The computed red shift of the scattering spectra shows that the most relevant geometrical parameter contributing to the scattering behavior of a single Ag nanoshell is the external diameter D. Fig. 2d shows the electric field distribution at the peak wavelength near 600 nm for diameter = 200 nm nanoshell, which corresponds well to the measured spectrum of the 273 nm-diameter nanotrough part. As expected, the field profile is predominantly dipolar. The fact that field inside the nanoshell is present even at the peak of the spectral resonance is due to penetration of the external fielding into the structure at this low for a metal thickness of only 20 nm, which resonates inside the nanoshell. However, the peak intensities of the field on both the internal and external sides of the shell are the comparable. In the FDTD environment, staircasing effects are particularly strong when dealing with dispersive materials (silver in our case) and nanostructures with curved


10

Page 11 of 33 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


surfaces. As a result, an average surface field must be computed in order to better visualize its local distribution. We have achieved this goal by performing standard cubic interpolation on a 3000 by 3000 pixel grid, and plotted the square root of the field amplitude in Fig. 2d.

3.3. Influence of the diameter (D) of a single Ag nanotrough on the SERS performance. In the SERS measurements, para-mercaptoaniline (pMA) was used as a test analyte to study quantitatively the influence of geometry of individual Ag cylindrical nanotrough to the SERS signal from pMA molecules adsorbed on the surface. The Ag nanotroughs were immersed in a 4 mM pMA ethanol solution for 2 h to form a uniform monolayer on the Ag surface. Then the prepared sample was rinsed with ethanol and then transferred to a Renishaw Confocal Raman Microscope for the SERS measurement using a 532 nm line for the excitation. The pump power and the acquisition time were kept at 16.1 µW and 30s, respectively, which is suitable for highthroughput detection that does not modify the Ag nanoshell. All subsequently reported SERS measurements were performed with these acquisition parameters unless otherwise stated. In order to build a direct relationship between the SERS performance and the scattering spectra of individual Ag nanotrough, we measured the scattering properties and the SERS performances on nanotroughs fabricated with different Ds. Figure 3 shows all pMA SERS spectra from Ag isolated cylindrical nanotroughs with different values of D ranging from 183 nm to 621 nm, corresponding to the same Ag nanotroughs that were measured in the previous scattering measurement. Two factors contributed to the SERS intensity with varying Ds. First is the amount of pMA molecules attached to the structure in the detection area. Here, we assume that the Ag nanoshell structure has a half cylinder shape that is


11


Page 12 of 33

hollow inside (i.e., a cylindrical nanotrough, as shown in figure 1) so that the total SERS active area will increase by increasing D. Secondly, the Raman signal enhancement generated by the plasmonic nanostructures reaches its maximum value when the excitation wavelength matches the plasmonic resonances of the corresponding substrates

30-32

. The scattering resonant peak

wavelength approaches the laser excitation wavelength (532 nm) when the diameter D of nanotrough is decreased from 631 nm to 183 nm. The increasing overlap between the excitation wavelength and the scattering peak boosts the local field enhancement and the resulting SERS detection signal 42. However, we should keep in mind that the scattering intensity of nanotrough with D =183 nm is much weaker (~10 fold) due to the reduced number density across the sample. In order to quantitatively compare the SERS intensity enhancement factor for the nanotroughs with different Ds, the intensity obtained for the 1077 cm-1 of pMA on the Ag nanoshell was measured at different points along the nanotrough. In the inset of Figure 3, the signal intensities were plotted as a function of D, and the SERS signal intensity was found to reach its maximum for the Ag nanotroughs with Ds of 273 nm and 318 nm. The nanotroughs with D larger than 318 nm showed lower intensity because their plasmon resonance peak is significantly detuned from the excitation wavelength. On the other hand, the SERS signal of nanotrough with D=183 nm decreases due to its lower scattering efficiency and significantly reduced active volume compared to the nanotroughs with larger D.

3.4. Scattering property and SERS performance of Ag CNNs. For applications to single bacterial detection, we fabricate a highly interconnected Ag CNNs that increases the entrapment of bacteria. The morphologies of the HPC random network template and of the Ag CNNs are shown in Figure 4a. We notice that the Ag CNNs structure obtained by coating the HPC


12

Page 13 of 33 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


networks with Ag and then dissolving the core preserves the same morphology of the HPC network (Inset of Figure 4a right). The thickness of the Ag layer is held constant at around 30 nm. The Ag layer needs to be thick enough to sustain this concave structure and to maintain the integrity of the nanotrough network after the cellulose has been dissolved. One the other hand its thickness needs to be small enough in order to support localized plasmon modes on the substrate, which are essential for efficient SERS detection. We found that the humidity conditions during the electrospinning are a significant factor contributing to the average diameter of HPC nanofibers. Different Ag CNNs samples were fabricated by applying controlled electrospinning humidity of 20%, 40%, 55%, and 70% (Figure 4b). The average diameter was measured and calculated from the SEM images and then plotted as a function of humidity. Figure 5e shows that the average diameter decreases from 582 nm to 272 nm when increasing the humidity from 20% to 70%. Investigating the scattering property in dark-field, a broad scattering peak (Figure S2) emerged in the scattering spectra for all the samples due to the large dispersion of sizes of Ag nanotroughs in the random networks (Figure 4c). To better understand the scattering property of the Ag CNNs, we fit the scattering spectrum using a multiple-peak model. Figure 4c shows that scattering spectrum is well reproduced using a two-peak Gaussian model (Peak 1 is centered around 480 nm and Peak 2 around 650 nm), which reflects the excitation of two main plasmon modes associated to the binomial size distribution of the CNN sample. In order to investigate the effect of humidity to the scattering property, we plot the ratio of the areas of the two Gaussian fitting curves (area of Gaussian with peak 1 divided the area of the Gaussian with peak 2) for different humidity values (Figure 4d). We find that the ratio increases by increasing the humidity, resulting in larger scattering intensity around the excitation wavelength for the sample fabricated under the highest humidity conditions. SERS measurements on pMA are performed


13


Page 14 of 33

using the same procedure described before. The SERS spectra are presented in Figure 5a-d where 9 different random spots were used on each Ag CNNs sample to obtain an average value. In Figure 5f we show the average intensity at 1077 cm-1 plotted as a function of humidity. Comparing the average size of the samples with the single nanotrough scenario, the SERS signals on the samples fabricated with 40%, 55%, and 70% humidity are in good agreement with the trend of SERS intensity versus the different D measured for isolated nanotroughs. This is evidence that the Ag CNNs maintain similar levels of SERS performances even when forming an interconnected random network. Moreover, a more compact uniform network structure was achieved by performing electrospinning in high humidity conditions, as supported by the smaller standard deviation in the average diameter of the nanotrough structures (Figure 5e). In addition, this compact network structure results in a larger active volume for the pMA molecule at the excitation spot and it increases the reproducibility of the SERS signal across the samples (Figure 5f). In order to quantify the level of reproducibility of the SERS signal on our sample, we measured SERS spectra of pMA on 9 different randomly selected points on the Ag CNNs and calculated the relative standard deviation (RSD) over the main peak intensity 37. The Raman intensity of pMA at 1077 cm-1 was utilized to calculate the RSD value of 0.196 (Figure S3b), which shows a good reproducibility over a large scale.

3.4. Bacterial strains differentiation based on Principle component analysis (PCA). Their facile and cost-effective fabrication make Ag CNNs particularly interesting for complex biosensing applications, such as cell-identification through spectral fingerprinting. Due to potential clinical applications and heightened biothreat concerns in recent decades, there is an increasing interest in developing SERS substrates for rapid detection and identification of whole


14

Page 15 of 33 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


bacterial cell with species and strain specificity 35. Given the distinctive distance dependence of field enhancement, SERS probes secreted molecules and metabolites as well as the components of the outer layer of bacterial cells, which provide great chemical selectivity

45,46

. Thus, in

principle, SERS can enhance both sensitivity and specificity of bacterial identification, which makes it a promising tool for bacterial diagnostics

47-49

. Substrate surface morphology and

bacterial binding affinity to the substrate are two important factors that impact the frequency intensity of the vibrational bands in SERS fingerprinting 46. In order to demonstrate that Ag cylindrical nanotrough networks are suitable substrates for the differentiation of SERS spectra of bacterial cells, three different bacterial strains E. Coli K12, coli BL21(DE3) and E. coli DH 5α were cultured and then washed and suspended in saline solution. E. coli DH 5α is a derivative of K12 and E. coli BL21(DE3) is a derivative of E. coli strain B which is different from E. coli K12

48

. The Ag CNNs bacterial SERS sensors were

fabricated under 70 % humidity and immersed in the suspensions of three type of different E. coli respectively for 1 h and then washed using DI water afterwards. SERS spectra excited at 532 nm were acquired. Each sample in the SERS measurement was illuminated by a small laser spot (~800 nm) at 532 nm. SERS spectral of bacterial were taken at 40 different random locations for each of the bacterial strains. Considering the bacteria surface coverage and the average sizes of the bacterial cells (see Figure 6a), we estimate that only one or two E. coli bacteria occupy the laser spot. Compared to the area without bacteria (black curve in 7b), we find that quantifiable SERS spectra were unambiguously detected from the bacterial cells adsorbed on the surface of Ag CNNs (red curve in Figure 6b). For control purposes, SERS measurements were also carried out for the Ag CNNs immersed in the broth without any bacteria (blue curve in Figure 6b). The spectra do not contain any characteristic spectral features, which prove that the broth alone does


15


Page 16 of 33

not influence the SERS measurement of the surface-attached bacteria. The SERS spectra of in Figure 6c show many common spectral features. However, when the number of measurements is sufficiently large (40 spectra per bacterium in our case) systematic differences in the intensity and frequency of individual spectral peaks can be rapidly identified and detected using the statistical multivariate analysis method know as Principle component analysis (PCA)50-52. PCA is a common chemometric technique routinely applied for the classification of bacteria on the basis of their spectral fingerprints

49

.

Depending on the sign of the second derivative of the SERS

spectra, the spectra are coded into a series of zeroes and ones (i.e., barcodes) that can be fed to the PCA to obtain sensitive and specific classification of bacterial strains. In our case, the ability of the Ag CNNs materials to differentiate between the three bacteria strains was demonstrated by performing PCA analysis using the Matlab PCA Toolbox (Eigenvector Research, Inc) based on SERS spectra collected from 40 random areas for each bacteria sample. The spectra were normalized to the maximum intensity channel and then smoothed with a SavitzkyGolay filter 53. The PCA reduces the dimensionality of each spectrum from 1400 independent variables to 5 principle components (PCs). As shown in Figure 6d where we plot the PC2 versus the PC1, the SERS signatures of E. coli K12/ E. coli DH 5α and E. coli BL21(DE3) on the Ag CNNs cluster appear very well-separated and localized within nonoverlapping regions shown in Figure 6d and Figure S4. The highlighted ellipses around each cluster region correspond to 95% confidence intervals centered on the mean for each strain cluster. On the contrary, the SERS signature of E. coli K12 and its derivative E. coli DH 5α were found to form overlapping cluster regions which result from the similar SERS spectra in Figure 6c. The clear separation achieved between different strain clusters and overlapping between


16

Page 17 of 33 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


origin type and its derivative prove that the Ag CNNs are suitable to achieve scalable biodetection and rapid spectroscopic identification of different bacterial strains.

4. Conclusion This study demonstrates scalable plasmonic functional materials that can be used as SERS substrates driven by the tunable plasmonic response of Ag CNNs templates. These materials are fabricated by facile electrospinning of environmentally friendly cellulose nanofiber templates followed by metal evaporation and HPC dissolution. The cost-effective electrospinning technique is utilized as a sacrificial template for the engineering of novel SERS substrates. The plasmonics resonances of fabricated materials are systematically controlled by the Ag cylindrical nanotrough diameters. Using the PCA multivariate statistical analysis on a large number of measured bacterial spectra, we demonstrate fingerprinting discrimination among two different types of bacteria (E. Coli K12, E. coli DH 5α and E. coli BL21(DE3)). This approach offers significant advantages in SERS detection, including improved versatility, cost effectiveness, specificity, and scalability that are all key features for bacterial diagnostics compared to alternative SERS substrates technologies.


17


Page 18 of 33

Figure 1. (a) Schematic figure of the procedure for fabricating Ag cylindrical nanotrough structures. (b) Configuration of electrospinning system for the nanofibers template. (c) SEM images of HPC nanofibers and (d) 30 nm thick Ag nanotrough after coating HPC nanofibers with Ag and dissolving HPC, the inset shows the hollow structures of an individual Ag nanotrough structure.


18

Page 19 of 33 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 2. (a) Measured dark-field scattering spectra of a single Ag nanotrough with different diameters. (b) The relation between diameter of isolated nanotroughs and peak positions of darkfield scattering spectrum. (c) FDTD simulations of the scattering spectra (normalized to 1 in each case) of isolated semi-circular Ag nanoshells with different external diameters from 100nm to 500nm as indicated on the figure. The thickness is fixed to be 20nm, and length fixed to be 5µm. The colors correspond to closest experimental results in panel (a). (d) Shows the square root of the field amplitude (after cubic interpolation on a 3000 by 3000 pixel grid) for the simplified model of Ag nanothrough with external diameter D = 200nm and thickness of 20 nm on an ITO substrate (refractive index=1.8), illuminated by a linearly polarized plane wave at normal incidence.


19


Page 20 of 33

Figure 3. SERS spectra of pMA adsorbed on the surface of individual Ag nanotroughs with different diameters. In the inset we show the intensity of the peak at 1077 cm-1 for each sample.


20

Page 21 of 33 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure. 4. (a) SEM images of (a) Left top: HPC nanofiber networks, Bottom: Ag CNNs after fabrication, Right top: the hollow half shell structures of Ag CNNs. (b) SEM images of the Ag CNNs fabricated under different humidity levels as written in the panels. (c) Representative scattering spectrum of a Ag CNNs sample fabricated under 55% humidity fitted with a two-peak Gaussian model. Red curve is the fitting curve (d) Ratio of the areas of the two Gaussian fitting curves (area of Gaussian with peak 1 divided the area of the Gaussian with peak 2) for different humidity conditions.


21


Page 22 of 33

Figure 5. The SERS signal of pMA was measured on Ag CNNs fabricated under different electrospinning humidity conditions according to: (a) 20 % (b) 40 % (c) 55 % (d) 70 %; (e) The relation between the humidity and average diameter of Ag CNNs. (f) The peak intensity at 1077 cm-1 in SERS spectra corresponding to (a-d). SERS signals were measured at different locations on each sample in order to estimate the error bars.


22

Page 23 of 33 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 6. (a) The Ag CNNs before (left) and after (right) being immersed in the E. coli saline solution. (b) SERS bacteria spectra measured at the location where bacteria are attached to the CNNs sample (A black) at the location without any bacteria attached (B red). SERS spectra also measured on Ag CNNs immersed only in the broth solution without any bacteria (C blue) for control. (c) The average normalized SERS spectra obtained from the E. Coli K12(red), E. coli DH 5α E(Blue) and E. coli BL21(DE3) (green) (d) PCA analysis for the SERS spectra of E. Coli K12(red), E. coli BL21 (green) and E. coli DH 5α E (Blue) for 95 % confidential ring.

ASSOCIATED CONTENT


23


Page 24 of 33

Supporting Information. Dark field scattering image of the Ag nanotrough and Ag CNNs. Bright field image and reproducibility test. 3D PCA plot for different bacterial strains. (PDF) AUTHOR INFORMATION Corresponding Author *Email: [email protected] Funding Sources MultiScale Multidisciplinary Modeling of Electronic Materials (MSME): W911NF-12-2-0023 National Science Foundation (NSF): CHE-1609778 ACKNOWLEDGMENT L. D. N. would like to acknowledge the support of the Boston University Nanotechnology Innovation Center received through the BUnano pilot grant “Biocompatible Plasmonic Nanostructures for Noninvasive Brain Imaging.” Additionally, L. D. N. would like to acknowledge the partial support of the Army Research Laboratory through the Collaborative Research Alliance (CRA) for MultiScale Multidisciplinary Modeling of Electronic Materials (MSME) under Cooperative Agreement Number W911NF-12-2-0023.

BMR acknowledges

partial support from the National Science Foundation (NSF) through grant CHE-1609778. R. Z would like to thank Can Nathchar in Prof. Pinghua Liu’s group for providing the E. coli bacterial strains. REFERENCES


24

Page 25 of 33 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(1) Jeanmai, D. L.; Van Duyne, R. P. Surface Raman Spectroelectrochemistry: Part I. Heterocyclic, aromatic, and aliphatic amines adsorbed on the anodized silver electrode. J Electroanal Chem, 1977, 84, 1-20. (2) King, F. W.; Van Duyne, R. P.; Schatz G. C. Theory of Raman scattering by molecules adsorbed on electrode surfaces. J. Chem. Phys. 1978, 69, 4472-4481. (3) Allen, C.S.; Van Duyne R. P. Molecular Generality of Surface-Enhanced Raman Spectroscopy (SERS). A Detailed Investigation of the Hexacyanoruthenate Ion Adsorbed on Silver and Copper Electrodes. J. Am. Chem. Soc. 1981, 103, 7497-7501. (4) Smekal, A. Zur Quantentheorie der Dispersion. Naturwissenschaften 1923, 11, 873-875. (5) Hudson, S. D.; Chumanov G. Bioanalytical applications of SERS (Surface-Enhanced Raman spectroscopy). Anal Bioanal Chem, 2009, 394, 679-686. (6) Kerker, M.; Wang, D.; Chew, H. Surface enhanced Raman scattering (SERS) by molecules adsorbed at spherical particles: errata. Appl. Opt. 1980, 19, 4159-4174. (7) Gopinath, A.; Boriskina, S.; Feng, N. N.; Reinhard, B. M.; Dal Negro, L. Photonic-Plasmonic Scattering Resonances in Deterministic Aperiodic Structures. Nano Lett. 2008, 8, 2423-2431. (8) Gopinath, A.; Boriskina, S.; Reinhard, B. M.; Dal Negro, L. Deterministic aperiodic arrays of metal nanoparticles for surface-enhanced Raman scattering(SERS). Opt. Express, 2009, 17, 3741-3753.


25


Page 26 of 33

(9) Dallapiccola, R.; Gopinath, A.; Stellacci, F.; Dal Negro, L. Quasi-periodic distribution of plasmon modes in two-dimensional Fibonacci arrays of metal nanoparticles. Opt. Express. 2008, 16, 8, 5544-5555. (10) Yan, B.; Wang, J.; Yang, L.; Boriskina, S. V.; Yan, B.; Reinhard, B. M. Spectroscopic Ultra-Trace Detection of Nitroaromatic Gas Vapor on Rationally Designed Two-Dimensional Nanoparticle Cluster Arrays. Anal. Chem. 2011, 83, 2243-2249. (11) Yan, B.; Boriskina, S. V.; Reinhard, B. M. Design and Implementation of Noble Metal Nanoparticle Cluster Arrays for Plasmon Enhanced Biosensing. J. Phys. Chem. C, 2011, 115, 24437-24453. (12) Tripp, R. A.; Novel nanostructures for SERS biosensing, Nano Today 2008, 3, 31-37. (13) Caldwell, J. D.; Glembocki, O.; Bezares, F. J.; Bassim, N. D.; Rendell, R. W.; Feygelson, M.; Ukaegbu, M.; Kasica, R.; Shirey, L.; Hosten, C. Plasmonic Nanopillar Arrays for LargeArea, High-Enhancement Surface-Enhanced Raman Scattering Sensors. ACS Nano 2011, 5, 4046-4055. (14) Lee, S. H.; Bantz, K. C.; Lindquist, N. C.; Oh, S. H.; Haynes. C. L. Self-assembled plasmonic nanohole arrays. Langmuir 2009, 25, 13685-13693. (15) Yu, Q.; Guan, P.; Qin, D.; Golden, G.; Wallace, P. M. Inverted Size-Dependence of Surface-Enhanced Raman Scattering on Gold Nanohole and Nanodisk Arrays. Nano Lett. 2008, 8, 1923-1928.


26

Page 27 of 33 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(16) Galvan, D. D.; Yu, Q. Surface-Enhanced Raman Scattering for Rapid Detection and Characterization of Antibiotic-Resistant Bacteria. Adv. Healthcare Mater. 2018, 278, 1701335. (17) Zhang, L.; Xu, J.; Mi, L.; Gong, H.; Jiang, S.; Yu, Q. Multifunctional magnetic-plasmonic nanoparticles for fast concentration and sensitive detection of bacteria using SERS, Biosens. Bioelectron. 2012, 31, 130-136. (18) Xu, J.; Zhang, L.; Gong, H.; Homola, J.; Yu, Q. Tailoring Plasmonic Nanostructures for Optimal SERS Sensing of Small Molecules and Large Microorganisms. Small 2011, 7, 3, 371– 376. (19) Banaee, M. G.; Crozier, K. B. Gold nanorings as substrates for surface-enhanced Raman scattering. Opt. Lett. 2010, 35, 760-762. (20) Rout, C. S.; Kumar, A.; Xiong, G.; Irudayara, J.; Fisher, T. S. Au nanoparticles on graphitic petal arrays for surface-enhanced Raman spectroscopy. Appl. Phys. Lett. 2010, 97, 133108133108. (21) Wei, H.; Kanson, U. H.; Yang, Z.; Xu, H. Individual Nanometer Hole–Particle Pairs for Surface‐ Enhanced Raman Scattering. Small 2008, 4, 1296-1300. (22) Bhuvana, T.; Kulkarni, G. U. Femtoliter silver cups as surface enhanced Raman scattering active containersNanotechnology. 2009, 20, 1-5. (23) Huang, F. M.; Wilding, D.; Speed, J. D.; Russel, A. E.; Bartlett, P. N.; Baumberg, J. J. Dressing Plasmons in Particle-in-Cavity Architectures. Nano Lett. 2011, 11, 1221-1226.


27


Page 28 of 33

(24) Mettela, G.; Boya, R.; Singh, D.; Kumar, G. V. P.; Kulkarni, G. U. Highly tapered pentagonal bipyramidal Au microcrystals with high index faceted corrugation: Synthesis and optical properties. Sci. Rep. 2013, 3, 1-8. (25) Singh, J.P.; Chu, H. Y.; Abell, J.; Tripp, R. A.; Zhao, Y., Flexible and mechanical strain resistant large area SERS active substrates. Nanoscale 2012, 4, 3410-3414. (26) Willets, K. A.; Van Duyne, R. P. Localized Surface Plasmon Resonance Spectroscopy and Sensing, Annu. Rev. Phys. Chem. 2007. 58, 267-297. (27) Bianco, G. V.; Losurdo, M.; Giangregorio, M. M.; Capezzuto, P.; Bruno, G. Direct Fabrication Route to Plastic-Supported Gold Nanoparticles for Flexible NIR-SERS. Plasmonics 2013, 8, 159-165. (28) Di Benedetto, F.; Camposeo, A.; Pagliara, S.; Mele, E.; Persano, L.; Stabile, R.; Cingolani, R.; Pisignano, D. Patterning of light-emitting conjugated polymer nanofibers. Nat. Nanotechnol. 2008, 3, 614-619. (29) Wang, P.; Wang, Y.; Tong, L. Functionalized polymer nanofibers: a versatile platform for manipulating light at the nanoscale. Light Sci. Appl. 2013, 2, 1-9. (30) Sugimoto, H.; Zhang, R.; Reinhard, B. M.; Fujii, M.; Perotto, G.; Marelli, B.; Omenetto, F. G. Enhanced photoluminescence of Si nanocrystals-doped cellulose nanofibers by plasmonic light scattering. L. Dal. Negro, Appl. Phys. Lett. 2015, 107, 041111.


28

Page 29 of 33 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(31) Zhang, R.; Knitter, S.; Liew, S. F.; Omenetto, F. G.; Reinhard, B. M.; Cao, H.; Dal Negro, L. Plasmon-enhanced random lasing in bio-compatible networks of cellulose nanofibers. Appl. Phys. Lett. 2016, 108, 011103. (32) Chen, D.; Zhang, R.; Wang; R.; Dal Negro, L.; Minteer, S. D. Gold Nanofiber-Based Electrodes for Plasmon-Enhanced Electrocatalysis. J. Electrochem. Soc. 2016, 163, 1132-1135. (33) Ogata, S.; Hatae, T.; Shoguchi, K.; Shinohara, H. The dimensionless correlation of mean particle diameter in electrostatic atomization. Int. Chem. Eng. 1978, 18, 488-493. (34) Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. K.; Whitehouse, C. M. Electrospray ionization for mass spectrometry of large biomolecules. Science, 1989, 246, 64–71. (35) Tomita, Y.; Ishibashi, Y.; Yokoyama, T. Fundamental Studies on An Electrostatic Ink Jet Printer: 1st Report, Electrostatic Drop Formation. Bull. JSME 1986, 29, 3737-3743. (36) De la Mora, J. F. The current emitted by highly conducting Taylor cones. J. Fluid Mech. 1994, 260, 155-184. (37) He, D.; Hu, B.; Yao, Q. F.; Wang K.; Yu S. H. Large-Scale Synthesis of Flexible FreeStanding SERS Substrates with High Sensitivity: Electrospun PVA Nanofibers Embedded with Controlled Alignment of Silver Nanoparticles. ACS nano, 2009, 3, 3993–4002. (38) Camposeo, A.; Spadaro, D.; Magrì, D. Surface-enhanced Raman spectroscopy in 3D electrospun nanofiber mats coated with gold nanorods. Anal Bioanal Chem, 2016, 408, 13571367.


29


Page 30 of 33

(39) Liu, T. Y.; Tsai, K. T.; Wang, H. H.; Chen, Y.; Chao, Y. C.; Chang, H. H.; Lin, C. H.; Wang, J. K.; Wang, Y. L. Functionalized arrays of Raman-enhancing nanoparticles for capture and culture-free analysis of bacteria in human blood. Nat. Commun, 2011, 2, 1-8. (40) Szymborski, T.; Witkowska, E.; Adamkiewicz, W.; Waluk, J.; Kamińska, A.; Szymborski, T.; Witkowska, E.; Adamkiewicz, W.; Waluk, J.; Kamińska, A. Electrospun polymer mat as a SERS platform for the immobilization and detection of bacteria from fluids. Analyst. 2014, 139, 5061-5064. (41) Wu, H.; Kong, D.; Ruan, Z.; Hsu, P.; Wang, S.; Yu, Z.; Carney, T. J.; Hu, L.; Fan, H. S; Cui Y. A transparent electrode based on a metal nanotrough network. Nat Nanotechnol, 2013, 8, 421425. (42) Yan, B.; Thubagere, A.; Premasiri, R.; Ziegler, L. D.; Dal Negro, L.; Reinhard, B. M. Engineered SERS Substrates with Multiscale Signal Enhancement: Nanoparticle Cluster Arrays. ACS Nano, 2009, 3, 1190–1202. (43) Yang, L. L.; Bo, Y.; Premasiri, W. R.; Ziegler, L. D.; Dal Negro, L.; Reinhard, B. M. Engineering Nanoparticle Cluster Arrays for Bacterial Biosensing: The Role of the Building Block in Multiscale SERS Substrates. Adv. Funct. Mater 2010, 20, 2619-2628. (44) Lumerical Solutions, Inc. http://www.lumerical.com/tcad-products/fdtd/. (45) Efrima, S.; Bronk, B. V. Vibrational fingerprinting of bacterial pathogens by surface enhanced Raman scattering (SERS). J. Proceed SPIE. 1999, 164, 19-29.


30

Page 31 of 33 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(46) Kahraman, M.; Yazici, M. M.; Sahin, F.; Culha, M. Convective Assembly of Bacteria for Surface-Enhanced Raman Scattering. Langmuir 2008, 24, 894-901. (47) Premasiri, W. R.; Moir, D. T.; Klempner, M. S.; Krieger, N.; Jones, G.; Ziegler, L. D. Characterization of the Surface Enhanced Raman Scattering (SERS) of Bacteria. J. Phys. Chem. B 2005, 109, 312-320. (48) Blattner, F. R.; Plunkett, G.; Perna, N. T.; Burland, V.; Riley, M.; Collado-Vides, J.; Glasner Rode, C. K.; Mayhew, G. F.; Gregor, J.; Davis, N. W.; Kirkpatrick, H. A.; Goeden, M. A.; Rose, D. J.; Mau, B.; Shao, Y. The Complete Genome Sequence of Escherichia coli K-12. Science, 1997, 1453-1462. (49) Patel, I. S.; Premasiri, W. R.; Moir, D. T.; Ziegler, L. D. Barcoding bacterial cells: a SERS‐based methodology for pathogen identification. J. Raman Spectrosc. 2008, 39, 16601672. (50) Stöckel, S.; Kirchhoff, J.; Neugebauer, U.; Röscha, P.; Poppa, J. The application of Raman spectroscopy for the detection and identification of microorganisms. J. Raman Spectrosc. 2016, 47, 89-109. (51) Hamasha, K.; Mohaidat, Q. I.; Putnam, R. A.; Woodman, R. C.; Palchaudhuri, S.; Rehse, S. J. Sensitive and specific discrimination of pathogenic and nonpathogenic Escherichia coli using Raman spectroscopy—a comparison of two multivariate analysis techniques. Biomed Opt Express 2013, 4, 481-489. (52) http://www.eigenvector.com/software/pca_toolbox.htm


31


Page 32 of 33

(53) Jonsson, P.; Eklundh, L.; TIMESAT—a program for analyzing time-series of satellite sensor data. Comput. Geosci. 2004, 30, 833-845.


32

Page 33 of 33 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Table of Contents Graphic


33

Plasmonic Nanotrough Networks for Scalable Bacterial Raman

Recommend Documents