Synergistic enhancement via plasmonic nanoplate-bacteria-nanorod supercrystals for highly efficient SERS sensing of food-borne bacteria

1 Synergistic enhancement via plasmonic nanoplate-bacteria-nanorod supercrystals for highly efficient SERS sensing of food-borne bacteria WeiQiang Wang‡a, Ville Hynninen‡b, Li Qiu‡c, AiWen Zhangc, Tibebe Lemmad, NanNan Zhanga, HongHua Ge*a, J. Jussi Toppari*d, Vesa P. Hytönenb and Jin Wang*a a Institute of Health Sciences and School of Life Science, AnHui University, Hefei, Anhui 230601, P.R. China b BioMediTech, University of Tampere, FI-33520 Tampere, Finland and Fimlab Laboratories, FI-35520, Tampere, Finland c Institute of Intelligent Machines, Chinese Academy of Sciences, Hefei, Anhui 230031, P. R. China dUniversity of Jyväskylä, Department of Physics, Nanoscience Center, P.O. Box 35, FI-40014, University of Jyväskylä, Finland KEYWORDS: Assembly, Asymmetric nanoparticles, Bacteria, SERS, Chemometrics This is the accepted manuscript of the article, which has been published in Sensors and Actuators B: Chemical. 2017, 239, 515-525. http://dx.doi.org/10.1016/j.snb.2016.08.040 2 ABSTRACT Bio-sensing techniques utilizing metallic nanoparticles as a probe have gained more and more attention and play today an important role in the detection of bacteria. To date, although several sensing materials have been tested, there is still a long way to go to achieve a fast, low-cost, ultrahigh sensitive and multifunctional substrate suitable for a universal biosensor for detection of bacterial cells. Here, we report a novel probe design based on anisotropic plasmonic nanoparticles organized to a biocompatible 3D bio-inorganic scaffold, i.e., nanoplate-bacteriananorod supercrystals (NBNS) with extremely high surface-enhanced Raman spectroscopic (SERS) activity as a model of synergistic plasmonic enhancement from nanoparticles and assembly. This unique structure of nanoparticles incorporated into supercrystal assembly allows efficient detection, identification and classification of cells and bacteria. In this design, the NBNS ensures that the target cells take advantage of the superior multifold increase in Raman scattering signals (electromagnetic enhancement from both types of nanoparticles), due to the geometry of the 3D scaffold. The excellent reproducibility and stability of NBNS substrates were confirmed by comparing the SERS activities of different substrates and analytes. Principal component analysis (PCA) applied to the SERS spectra clearly discriminated the homogeneous bacterial samples and their mixtures. Successful detection and identification of bacteria in model samples consisting of two representative bacteria blends in Fanta soft-drink were demonstrated via plasmonic bio-inorganic scaffold combined with PCA analysis. We believe that this work will greatly facilitate the development of ultrasensitive SERS probes for highly advanced biosensor, pioneering the use of SERS for controlling food safety. 3 1. Introduction Food-borne bacteria, e.g.,Escherichia coli,Salmonella etc., have received more attention recently because of the low infectious dose and high health risk associated with these pathogens responsible for many food-borne diseases. Therefore, developing rapid, sensitive and reliable methods for assaying of food-borne pathogens are of an urgent demand. Many modern techniques, e.g., microfluidic device [1,2], colorimetric [3,4], electrical device [5,6], fluorescence [7], microcantilever [8], electrochemical [9], chemiluminescence [10],.have been already used for detection of food-borne pathogens. Assembled nanoparticles have also been extensively studied due to their potential applications in electronic and optical sensors owing to their favorable photo-physics and spectroscopic properties. For instance, Chang et al. employed ultrasonic-assisted self-assembling of monolayer graphene oxide to fabricate a highly sensitive and selective field-effect transistor (FET) sensor for an assay of Escherichia coli bacteria [11]. Ondera et al. assembled gold nanopopcorns on single–walled carbon nanotubes as a hybridized sensor for rapid detection of E. coli bacteria [12]. As compared to other spectroscopic techniques, surface enhanced Raman scattering (SERS) based on plasmonic nanoparticles is a highly attractive method for detection of bacteria due to its high-sensitivity, reliability and reproducibility [12-16]. For example Wu et al. successfully utilized a silver nanorod array to obtain SERS signal of foodborne pathogenic bacteria in mung bean sprout samples by using a portable and handheld Raman system [17]. Particularly promising among possible designs for exploiting the unique optical properties of metallic nanoparticles on the functionality of sensing selective materials is the layer-by-layer (LBL) assembly protocol for fabrication of metamolecules. These strongly coupled plasmonic 4 assemblies as metamaterials have been extensively investigated because of their inherent highlysensitive SERS properties originating from the densely packed “hot-spots” within the nanoassembly [17-22]. In this paper, we utilize plasmonic nanoplates and nanorods supercrystal to develop a novel highly sensitive SERS protocol. It is based on 3D bio-inorganic supercrystal assembly of anisotropic nanoparticles of various types, i.e., nanoplate-bacteria-nanorod supercrystals (NBNS), and used here for label-free assay of different Gram-positive and Gram-negative bacteria. The novel construction of NBNS ensures that the target cells take an advantage of the superior multifold enhanced Raman scattering enabling discrimination of different bacteria species based on their SERS spectra. Furthermore, to demonstrate the applicability of the method especially within the area of the food industry, we recorded SERS spectra from bacteria embedded in soft-drink and obtained a clear separation of them from the surrounding matrix with high correlation by using principal component analysis (PCA). This statistical technique coupled with the high sensitivity of the SERS data opens the door to the applicability of this biosensing platform to food quality control. 2. MATERIALS AND METHODS 2.1 Materials HAuCl4·4H2O (99.9%), NaBH4 (99%), Vitamin C (99.9%), cetyltrimethylammonium bromide (CTAB) (99%), AgNO3 (99.9%), 2-Quinolinethiol (2-QT), poly(sodium styrenesulphonate) (PSS), trisodium citrate and vancomycin hydrochloride were purchased from Sigma-Aldrich, Gram-negative Salmonella enterica (S. enterica) bacteria (ATCC 14028), Gram- 5 negative Escherichia coli (E. coli) (CMCC 44568), Gram-positive Staphylococcus xylosus (S. xylosus) (ZSL-3) were purchased from China General Microbiological Culture Collection Center (CGMCC). The deionized water that was used throughout the experiments was purified by using a Milli-Q system. 2.2 Fabrication of plasmonic 3D nanorod supercrystals All of the silicon chips with (1x1 cm2) were firstly cleaned by ultrasonication in acetone, ethanol and MilliQ water for 1h. Then, the silicon chips were immersed into a piranha solution (H2SO4:H2O2 = 3:1) at 80˚C for 1 h and rinsed with Milli-Q water. The clean silicon chips were dried by nitrogen gas flow at room temperature. Synthesis of nanorods is explained in supporting information (S1.1). Two samples of 1 mL asprepared nanorod solution was centrifuged at the 8000 rpm for 10 min to remove the excess CTAB surfactants and the precipitates were redispersed in 50 μL and 200 μL Milli-Q solution, respectively. Then, two samples of the obtained nanorod colloidal solution with 5 μL were dropped on the cleaned silicon substrate and the chips were kept in Petri dish under cover at humid environment with temperature 20°C for 24 h so as to yield vertical and parallel 3D nanorod supercrystals, respectively. 2.3 Bacterial culturing Lyophilized powder of E. coli bacteria (CMCC44568), Salmonella enterica (S. enterica) (ATCC 14028) and Staphylococcus xylosus (S. xylosus) bacteria (ZSL-3) were resuspended to 5mL LB nutrient and cultivated for 14h in shaking incubator (200 rpm), respectively. Bacterial 6 suspension was plated on an LB agar plate using sterile plastic inoculating loops to carry out plate streaking and the plate was settled in a constant temperature incubator and cultivated for 14h to obtain single colonies. Single colonies were collected via sterile plastic inoculating loops and dissolved into 1 mL aquae sterilisata. 100 μL bacteria solution was coated on LB agar base and incubated in constant temperature incubator at 37 °C. Subsequently, the bacteria was dispersed in Milli-Q water and centrifuged for 10 min at 3000 rpm so as to protect the cell membrane. Finally, the supernatant liquid was discarded and the bacterial cells were redispersed in Milli-Q water. The purified bacteria were diluted to a concentration of 103cfu/mL. The density of bacteria cells was determined by counting the number of colonies grown on the Petri dish after 12 h of cultivation. 2.4 Fabrication of 3D scaffold of plasmonic nanoplate aggregate-bacteria-nanorod supercrystals The plasmonic 3D vertical nanorod supercrystal substrate was immersed into 10-2 M aqueous solution of vancomycin hydrochloride for 1 h and then dried in air to yield a vancomycin-coated supercrystal substrate. 1 mL of the bacterial suspension was placed on the as-prepared supercrystal substrate and then settled in a constant temperature incubator for 1 h at 37°C. After that, the substrate was washed three times by using Milli-Q water. The bacteria-immobilized supercrystal substrate was immersed into freshly prepared Ag nanoplates to obtain the 3D scaffold of plasmonic nanoplate aggregate-bacteria-nanorod supercrystal and then dried in air for characterization and measurement. Synthesis of Ag nanoplates is shown in supporting information (S1.2). 7 2.5 Characterization Scanning electron microscope (SEM) images were acquired with a Quanta200FEG. UV-VisNIR absorption spectra were recorded by using a Solidspec-3700 spectrophotometer. SERS experiment were performed via Renishaw Invia Reflex Raman spectrometer equipped with a 280 mW semiconductor laser emitting at a 785 nm line. Surface enhanced Raman spectra were collected through 100 x objective lens at the condition of 8.5 mW laser power and10 s integration time. 2.6 Chemometrics The Unscrambler® X software (version 10.3, CAMO Software; Oslo, Norway) including Unscrambler Service Pack 2014 was used to analyze SERS spectra. First, the spectra were visually inspected and significantly differing spectra were removed from the analysis. Then, linear baseline correction with baseline offset and standard normal variate (SNV) were used as a pretreatment. 400.78 cm-1 and 1761.03 cm-1 wavenumbers were used as the defining wavelengths in the linear baseline correction. In the baseline offset the lowest point in the spectrum is subtracted from all the variables (wavenumbers) and, thus, the minimum value is set to 0 and the rest of the variables have positive values. Following the baseline correction, the spectra were normalized using the standard normal variate (SNV) technique, which is a commonly used mathematical transformation method to remove multiplicative scatter and particle-size interferences from spectral data [23, 24]. Following the pretreatments, mean centered principal component analysis (PCA) was performed in the 400 cm-1 to 1800 cm-1 range. Initially, the 800 cm-1 to 1800 cm-1 range was also tested but the longer range produced more accurate results. Thus, the 400 cm-1 to 1800 cm-1 range was used throughout the experiments. 8 PCA reduces the dimensionality of the data set while retaining the dominant features that contribute most to its variance. This allows categorization of the data according to spectral features that might otherwise be hardly distinguishable [25, 26]. 3. RESULTS AND DISCUSSION As shown in Fig. 1, the assembling protocol of the NBNS can be described as three steps: (I) fabricating 3D plasmonic nanorod supercrystal on a chip; (II) immobilization of bacteria on the plasmonic supercrystal; (III) deposition of plasmonic nanoplates on the bacteria. This layer-bylayer fabricated hybridized superstructure substrate has several advantages: (I) 3D nanorod supercrystal provides a large-scaled uniform substrate; (II) bacteria immobilized on the vancomycin coated supercrystal provides a robust scaffold for aggregating nanoplates; (III) synergistic plasmonic enhancement from the 3D hybridized nanoplate aggregates and the nanorod supercrystal. 9 Fig. 1. Cartoon graph of the protocol of fabricating the 3D-shaped controllable plasmonic nanoplate aggregate-bacteria-nanorod supercrystal and utilizing it for label-free sensing of bacteria. 3.1. Fabrication of 3D plasmonic nanorod supercrystal. As an initial step for the layer-by-layer (LBL) fabrication of NBNS, Au nanorods with a longitudinal localized surface plasmon resonance (LSPR) at 783 nm (shown in Fig. S1) were produced (see supporting information for details). The excellent overlap of the LSPR band with the utilized excitation laser line at 785 nm yields optimized plasmonic enhancement. Thanks to the stronger LSPR effects of nanosilver components as compared to nanogold, it is necessary to give a Ag shell coating on the Au nanorods. As shown in Fig. S1, the sensitive longitudinal LSPR band can be remarkably blueshifted in contrast to the insensitive transverse LSPR band, around 500 nm, which only slightly blueshifts when Au nanorod is coated with Ag shell. The more silver is deposited the more the LSPR bands are shifted as clearly seen from Fig. S1. Additionally, in the sample with the highest amount of Ag, a broad LSPR band located around 400 nm, indicating formation of Ag components in addition to the Au/Ag core-shell nanorods. Herein, Au nanorods with the thinnest Ag shell, corresponding to the longitudinal LSPR band blueshifted from 783 nm to 765 nm, were selected as a building blocks for the 3D supercrystal, providing maximum plasmonic coupling with the excitation line of 785 nm. As reported from previous study, the plasmonic band of core-shell nanorod supercrystal is dramatically brodened and enranged from Visible to NIR regions as compared to that of core-shell nanorods colloid solution, which can be efficently coupling with different exciation laser lines [27]. 3D supercrystal of the Au/Ag core-shell nanorods described above was formed on a piranhatreated silicon chip by a controllable slow-evaporation method. In contrast to what is observed 16 supercrystal of plasmonic nanorods; F: S. xylosus on a vancomycin-coated 3D vertical supercrystal of plasmonic nanorods The 3D vertical nanorod supercrystal provides a round top-surface with nanometer-scale gaps (“nanogaps”) as binding sites for vancomycin so as to trap vancomycin nanoparticles with a few nanometers in diameter. Moreover, the nanogaps between the adjacent nanorods can act as important hot-spots (plasmonic gap modes) for amplifying exponentially the Raman scattering. As shown in Fig. 5A-C, in the case of the 3D-vertical nanorod supercrystal without vancomycin, the cell wall of bacteria can only protrude onto round top-surface. In contrast, the cell wall of bacteria can penetrate into a vancomycin-coated 3D-vertical nanorod supercrystal due to the vancomycin nanoparticles entrapped within the nanogaps, as shown in Fig. 5D-F. Furthermore, a stretched shadow around the cell wall of bacteria can be clearly observed on the vancomycincoated 3D supercrystal conversely to the bacteria on a 3D supercrystal without vancomycin. When a cell wall of bacteria is partially entrapped into hot-spot of a nanogap via vancomycin immobilization, great enhancement on a Raman signal of bacteria is expected. Indeed, as observed from Fig. 6, the intensity of the Raman signal of bacteria, S. xylosus,S. enterica and E. coli was increased ca. 9 times, 5 times and 3 times, respectively. The capturing ability of the vancomycin-coated 3D nanorod supercrystal is different for Gram+ and Grambacteria due to a different cell wall structure of Gram+ and Grambacteria, and is reflected in both SEM observations and SERS enhancement. Compared to Grambacteria, e.g.,E. coli and S. enterica, the interaction between vancomycin and peptidoglycan in the cell wall of Gram+ bacteria, e.g., S. xylosus, is much stronger. This can be related to the fact that cell walls of Staphylococcus could be destroyed and a release of cytoplasmic components may happen thanks to vancomycin 17 attacking the cell wall of bacteria by aid of hydrogen bonding to N-acetylglucosamine/Nacetylmuramic acid-peptide linker, peptidoglycan and teichoic-acid, reflected from the more obvious rim attached to the substrate (shown in Fig. 5F). Fig. 6. SERS spectra of A: S. xylosus on a plasmonic vertical nanorod supercrystal, B: S. xylosus bacteria on a vancomycin-coated plasmonic vertical nanorod supercrystal, C: S. xylosus embedded in a plasmonic nanoplate aggregate/vertical nanorod supercrystal, D: S. enterica on a plasmonic vertical nanorod supercrystal, E: S. enterica on a vancomycin-coated plasmonic vertical nanorod supercrystal, F: S. enterica embedded in a plasmonic nanoplate 18 aggregate/vertical nanorod supercrystal, G: E. coli on a plasmonic vertical nanorod supercrystal, H: E. coli on a vancomycin-coated plasmonic vertical nanorod supercrystal, I: E. coli embedded in a plasmonic nanoplate aggregate/vertical nanorod supercrystal. The absorption efficiency of bacteria in NBNS is depended on the vancomycin coatednanorod supercrystal. The capturing ability of van-nanorod supercrystal has been found to be determined by the coverage of the vancomycin coating. In order to obtain the dependence of the adsorption efficiency about the coverage of vancomycin on the nanorod supercrystal, the ratio between the number of the staphylococcus bacteria immobilized on the supercrystal substrate with vancocymin and without vancomycin. Fig. S7 illustrates that the ratio increase accompany with increasing vancomycin coating coverage of van-nanorod supercrystal. However, the dependence of the SERS ability of S. xylosus bacteria on the different van coveraged nanorod supercrystal can be reflected from Fig. S7 It should be pointed out that the ratio of Raman enhancement reaches the maximum and subsequent decrease with alteration of vancomycin converage on the NBNS substrate. The optimized coverage indicate that increasing the Raman signal of absorbed bacteria can be determined via increasing the vancomycin coverage, leading to more bacteria receiving SERS effects. However, the distance between bacteria and plasmonic supercrystal can be increased, which causes the loss of direct enhancement from the nanorod supercrystal. 3.4. Deposition of plasmonic nanoplates on the bacteria 19 Due to the height (ca. 100 nm~300 nm) of bacterial cells, e.g.,E. coli (Fig. S7), they protrude from the surface of the 3D vertical nanorod supercrystal. Therefore, only the interface between the lower edge of the bacteria and the endcap of the nanorod array can be effectively affected by a near-field plasmonic effects, implying that only Raman signal from the interface can be enhanced. Therefore, it is necessary to further deposit plasmonic nanoparticles into the gaps between the bacterial cells in order to embed the bacteria into the aggregates of plasmonic nanoparticles. In this case, the electron transfer between the plasmonic nanoparticle aggregates and nanorod supercrystal leads to additional strengthening of the plasmonic field. On the other hand, the bacteria can receive electromagnetic enhancement from the plasmonic field of 3D nanorods supercrystal at the bottom of the hybridized scaffold. Simultaneously, the bacteria can also obtain the electromagnetic enhancement from plasmonic field of the 3D nanoparticles aggregates amongst the gaps of hybridized scaffold. Triangular silver nanoplate is a powerful candidate within anisotropic plasmonic nanoparticles for constructing the chemical or biological sensors due to their thinness as compared to nanosphere, leading to a significant extension of the electromagnetic field farther from their surface. Furthermore, the three tips of the nanoplates with larger refractive index sensitivity dramatically contributes to their near-field electromagnetic enhancement localized the tips [3335]. As shown in Fig. S8A, the nanoplates with triangular tips can be clearly observed. By the present seed-mediated protocol (see supporting information for details), continuous tuning of LSPR band corresponding to the dipole resonance of Ag nanoplates, can be obtained ranging from visible region to near-infrared region (shown in Fig. S8B). We selected the Ag nanoplates with LSPR band located around 795 nm for the experiment as they produce maximum overlap 20 with the laser line excitation at 785 nm, which further leads to stronger SERS enhancement. Therefore, the selected Ag nanoplates were used to efficiently fill the gaps between the immobilized bacteria to form rodlike or spherical nanoaggregates. Detailed comparison of averaged SERS spectra (shown in Fig. S9) of S. xylosus bacteria adhered on different combinations of the substrates, suggests that the enhancement from the combined 3D Ag nanoplate aggregate-nanorod supercrystal is clearly higher than those from the 3D nanorod supercrystal or Ag nanoplate aggregates alone, indicating that synergistic enhancement from both Ag nanoplates aggregates and nanorod supercrystal can maximize the Raman signal of bacteria. 3.5. Analysis of SERS spectra of Gram+ and Grambacteria in the NBNS Based on the plasmonic 3D hybridized bio-inorganic scaffold, the inherent Raman vibrational signal of bacteria is dramatically amplified so that Raman signals of different bacteria can be clearly discriminated. Compared to Grambacteria, e.g.,S. enterica and E. coli, SERS spectrum of Gram+ bacteria S. xylosus is simpler due to a relatively simple structure of the cell wall. Detailed assignment of vibrational bands of bacteria in SERS spectra have been given in Tabs S1-S3. As observed from Fig. 6, the vibrational band between 600~800 cm-1 can be described as a fingerprint region of bacterial DNA/RNA due to the dominant SERS signal of adenine and guanine compared to the others. The order of the SERS cross-section is adenine > guanine > cytosine > thymine [36, 37]. Particularly, the bands at ca. 650 cm-1 and 730 cm-1 are clearly 21 different for all the three bacteria studied here. In the case of Gram+ bacteria, e.g.,S. xylosus, the band at 730 cm-1 is very strong in contrast to the very weak band at 666 cm-1, suggesting that activity of the ring stretching of adenine is stronger than that of guanine. However, in the case of Grambacteria, e.g.,S. enterica, two strong vibrational bands at 653 cm-1 and 727 cm-1 indicate that activities of the ring stretching of adenine and guanine are both high. In contrast, in another representative Grambacteria, e.g.,E. coli, the band at 658 cm-1 assigned as guanine vibration is obviously stronger than the band at 730 cm-1 ascribed to adenine vibration. Besides the fingerprint region between 600~800 cm-1, another obviously discriminating region in the SERS spectra of bacteria is observed between 1100~1500 cm-1. However, in contrast to Gram+ bacteria, the vibrational bands located at the region are quite complicated for Grambacteria due to multiple components in different bacteria. As far as S. xylosus is concerned, the two strong vibrational bands (1327 cm-1 and 1449 cm-1), can be assigned as ring stretching of cytosine/uracil and C-H deformation, respectively. Although analogous vibrational bands of Grambacteria could be observed within the spectral region, the profile of the series of bands is different, i.e., intensity ratios of the bands located at ca. 1329 cm-1, 1370 cm-1, 1413 cm-1 and 1454 cm-1, can be acted as another fingerprint spectral region for discriminating bacteria. It should be pointed out that the weak C=O stretch located at ca. 1660 cm-1 originating from Amide I of protein, and C-N-H deformation located at ca. 1280 cm-1 originating from Amide III, can be observed for all three bacteria. However, the similarities of the profile and position of the vibrational band of the different bacteria make it difficult to use them as fingerprint bands. In orde to obtain a limit of Detection (LOD) of Gram-negative bacteria E. coli and Grampositive bacteria S. xylosus via NBNS SERS substrate, we plotted the SERS peak intensity of E. 22 coli at I658 and SERS peak intensity of S. xylosus at I730 against the bacterial concentration (Fig. S10-S11), respectively. As far as each concentration is concerned, mean and standard deviation of the peak intensities can be obtained on the basis of the collected 6 spectra from two different substrates. In the present case, the LOD of the SERS detection can be determined by the lowest concentration at which a obvious bacterial SERS spectrum can be obtained. As shown from Fig. S10B and S11B, the lowest concentration of E.coli and S.xylosus bacteria that can be detected is 103 CFU/mL and 102CFU/mL, respectively. 3.6. Analysis of SERS spectra of mixed bacteria solutions. On the basis of SERS analysis of the three bacteria, experiments with mixtures of different bacteria with different ratios has been carried out to test multiplex sensing abilities via NBNS. As far as the SERS spectra of the two Grambacteria, i.e.,E. coli and S. enterica are concerned, the intensity ratio of bands at ca. 650 cm-1 and 730 cm-1 for the mixed bacteria (with ratio 1:1) is different from the pure bacteria as shown Fig. 7. In the case of S. enterica, the band at 650 cm-1 is obviously lower than the band at 730 cm-1; on the contrary, as for E. coli bacteria, the band at 650 cm-1 is dramatically higher than that at 730 cm-1. However, the intensities of the two bands in the mixed bacteria sample with ratio 1:1 are very close, implying that it is possible to discriminate between the mixture of bacteria and pure bacteria. When the ratio of E. coli and S. enterica is changed from 1:1 to 1:2, the intensity ratio of vibrational bands at 650 cm-1 and 730 cm-1 is remarkably decreased, implying that it is also possible to use this method to determine the concentration of the bacteria in a mixture. Furthermore, as the ratio of E. coli and S. enterica is continually increased to 1:5, the SERS spectra of mixed bacteria appears similar to the SERS 23 spectrum of S. enterica, which implies that the percentage of salmonella bacteria in the bacteria mixture is dominant. Additionally, another fingerprint spectral region between 1100 cm-1 and 1500 cm-1, can also be utilized to discriminate the mixed bacteria and pure bacteria. As seen from Fig. 7, the profiles of the bands of the mixed bacteria are quite different from the pure bacteria, i.e.,S. enterica and E. coli. It should be mentioned that the intensities of the bands assigned as C-O, C-C stretch, C-O-H deformation, located at the region of 900 cm-1 ~ 1200 cm-1, in the SERS spectra are obviously increased as compared to the pure bacteria. As far as the mixed Gramand Gram+ bacteria are concerned (see Fig. S12), E. coli and S. xylosus were also mixed with ratios of 1:1, 1:2 and 1:5 to study the effects of kinds of bacteria and concentration on the SERS signal. As shown by Fig. S12, the SERS spectra of mixed E. coli and S. xylosus with ratio 1:1, is different from SERS spectra of pure E. coli and S. xylosus samples. As compared to the SERS spectra of pure S. xylosus, the band at 660 cm-1 can be clearly observed in the mixture, indicating the presence of S. enterica in the sample. On the other hand, the band at 1326 cm-1 in the SERS spectra of the mixture is remarkably stronger than the other vibrational bands located in the region from 1100 cm-1 to 1500 cm-1. Because in real samples the density of E. Coli is usually higher than that of either S. enterica or S. xylosus, we compare the SERS spectrum of E. Coli bacteria and S. xylosus bacteria with the ratio of 10:1 to the 1:1 mixture or monospecies. As shown from Fig. S13, the relative intensity ratio of the bands at 730 cm-1 and 658 cm-1, can be dramatically changed accompany with the ratio changing from 1:1 to 10:1. For example, the intensity of the band at 658 cm-1 in the mixture with ratio 1:1, is much lower than that of the band at 730 cm-1; however, the intensity of the band 24 at 658 cm-1 in the mixture with ratio 10:1 can be comparable to that of the band at 730 cm-1, suggesting that the mixture can be differentiated from the mixture with ratio of 1:1. O the other hand, the relative intensity of the the band at the bands at 730 cm-1 and 658 cm-1 in the mixture with the ratio of 10:1 (shown in Fig. S13) appears somewhat different from the monospecies (E.Coli bacteria), which implies it is possible to discriminate extreme amounts of E-Coli bacteria in the mixture samples fom the monospecies. Fig. 7. A comparison on normalized SERS spectra of the mixed and monospecies bacteria collected from a 3D plasmonic nanoplate aggregate-bacteria-vertical nanorod supercrystal. A: GramE. coli; B: GramS. enterica; C: mixture of E. coli and S. enterica with ratio 1:1; D: mixture of E. coli and S. enterica with ratio 1:2; E: mixture of E. coli and S. enterica with ratio 1:5. 3.7. PCA analysis of monospecies and mixedbacteria SERS spectra. 25 When utilizing principal component analysis (PCA) on monospecies samples and the mixtures (1:1 ratio) of two different bacterial species together, the monospecies samples formed their own distinct clusters, as shown in Fig. 8a. Thus, the ability of PCA to identify monospecies samples based on their SERS spectra appeared possible. The mixture clusters, in turn, were partially overlapping and located in between of the monospecies samples. However, this was expected, because the mixtures contain features from the two different species and, therefore, they inevitably share some common characteristics. Nonetheless, by evaluating enough PCs, also the identification of the mixtures was shown to be possible, as is demonstrated in Fig. 8b where the clusters are plotted according to PC1 and PC5. Cumulatively, the first five PCs explained 72 % of the total data variance. The SERS spectra of the monospecies samples and 1:1 mixed samples are shown in Supplementary Information Fig. S14~S19. Fig. 8．PCA for monospecies and mixed (1:1) bacterial samples. (a) PCA results plotted according to first two principal components, PC1 and PC2. (b) PCA results plotted according to PC1 and PC5. Blue filled square = E. coli, red filled circle= S. xylosus, purple filled triangle = S. enterica, pink open circle = E. coli +S. xylosus mixture (1:1), green open diamond = E. coli +S. 32 infrastructure support. Dr Meng Fu and JieHong Mao from Merieux NutriSciences-Sino Analytica (QingDao and NingBo) company is gratefully acknowledged. ABBREVIATIONS SERS, surface enhanced raman spectroscopy; LSPR, localized surface plasmon resonance; NBNS, nanoplate-bacteria-nanorod supercrystal; REFERENCES [1] C. Bernabini., D. Holmes, H. Morgan, Micro-impedance cytometry for detection and analysis of micron-sized particles and bacteria. Lab. Chip 11 (2011) 407-412. [2] N. Couniot, L. A. Francis., D. Flandre., Resonant dielectrophoresis and electrohydrodynamics for high-sensitivity impedance detection of whole-cell bacteria. Lab. Chip 15 (2015) 3183-3191. [3] S. Wang, A. K. Singh, D. Senapati, A. Neely, H. Yu., P.C. Ray, Rapid Colorimetric Identification and Targeted Photothermal Lysis of Salmonella Bacteria by Using Bioconjugated Oval-Shaped Gold Nanoparticles. Chem. Eur. J. 16 (2010) 5600-5606. [4] M. S. Verma, P. Z. Chen, L. Jones, F. X. Gu, Branching and size of CTAB-coated gold nanostars control the colorimetric detection of bacteria. RSC. Adv. 4 (2014) 10660-10668. [5] M. S. Mannoor, S. Zhang, A. J. Link, M. C. McAlpine, Electrical detection of pathogenic bacteria via immobilized antimicrobial peptides. PNAS. 107 (2010) 19207–19212. 33 [6] Y. Chen, Z. P. Michael, G. P. Kotchey, Y. Zhao, A. Star, Electronic Detection of Bacteria Using Holey Reduced Graphene Oxide. ACS Applied Materials & Interfaces. 6 (2014) 3805– 3810. [7] G. Feng, Y. Yuan, H. Fang, R. Zhang, B. Xing, G. Zhang, D. Zhang, B. Liu, A light-up probe with aggregation-induced emission characteristics (AIE) for selective imaging, naked-eye detection and photodynamic killing of Gram-positive bacteria. Chem. Commun. 51 (2015) 12490-12493. [8] J. Wang, M. J. Morton, C. T. Elliott, N. Karoonuthaisiri, L. Segatori, S. L. Biswal, Rapid detection of pathogenic bacteria and screening of phage-derived peptides using microcantilevers. Anal. Chem.86 (2014) 1671-1678. [9] M. H. Abdalhai, A. Fernandes, M. Bashari, J. Ji, Q. He, X. Sun, Rapid and sensitive detection of foodborne pathogenic bacteria (Staphylococcus aureus) using an electrochemical DNA genomic biosensor and its application in fresh beef. J. Agric. Food. Chem. 62 (2014) 12659-12667. [10] R. Pacheco-Gómez, J. Kraemer, S. Stokoe, H. England, C. Penn, E. Stanley, A. Rodger, J. Ward, M. Hicks, T. Dafforn, Detection of pathogenic bacteria using a homogeneous immunoassay based on shear alignment of virus particles and linear dichroism. Anal. Chem. 84 (2012) 91-97. [11] J. Chang, S. Mao, Y. Zhang, S. Cui, G. Zhou, X. Wu, C. H. Yang, J. Chen, Ultrasonicassisted self-assembly of monolayer graphene oxide for rapid detection of Escherichia coli bacteria. Nanoscale. 9 (2013) 3620-3626. 34 [12] T. J. Ondera, A. T. HammeII, A gold nanopopcorn attached single-walled carbon nanotube hybrid for rapid detection and killing of bacteria. J. Mater. Chem. B. 43 (2014) 75347543. [13] F. Cui, Z. Q. Hu, M. Azlin, M. S. Lin, Rapid detection of foodand waterborne bacteria using surface-enhanced Raman spectroscopy coupled with silver nanosubstrates. Appl. Microbiol. Biotechnol. 92 (2011) 1053-1061. [14] A. Sivanesan, E. Witkowska, W. Adamkiewicz, L. Dziewit, A. Kamińska, J. Waluk, Nanostructured silver–gold bimetallic SERS substrates for selective identification of bacteria in human blood. Analyst. 139 (2014), 1037-1043. [15] D. Yang, H. Zhou, C. Haisch, R. Niessner, Y. Ying, Reproducible E. coli detection based on label-free SERS and mapping. Talanta. 146 (2016) 457-463. [16] H. Zhou, D. Yang, N. P. Ivleva, N. E. Mircescu, R. Niessner, C. Haisch, SERS Detection of Bacteria in Water by in Situ Coating with Ag Nanoparticles. Anal. Chem. 86 (2014) 15251533. [17] X. Wu, C. Xu, R. A. Tripp, Y. W. Huang, Y. Zhao, Detection and differentiation of foodborne pathogenic bacteria in mung bean sprouts using field deployable label-free SERS devices. Analyst. 138 (2013) 3005-3012. [18] Y. Zhao, X. Yang, H. Li, Y. Luo, R. Yu, L. Zhang, Y. Yanga, Q. Song, Au nanoflower– Ag nanoparticle assembled SERS-active substrates for sensitive MC-LR detection. Chemical Communications. 51 (2015) 16908-16911. 35 [19] C. Zhu, G. Meng, Q. Huang, Z. Zhang, Q. Xu, G. Liu, Z. Huang, Z. Chu, Ag nanosheetassembled micro-hemispheres as effective SERS substrates Chemical Communications. 47 (2011) 2709-2711. [20] X. Wang, D. Yang, P. Huang, M. Li, C. Li, D. Chen, D. Cui, Hierarchically assembled Au microspheres and sea urchin-like architectures: formation mechanism and SERS study. Nanoscale. 4 (2012) 7766-7772. [21] S. Gómez-Graña, C. Fernández-López, L. Polavarapu, J. B. Salmon, J. Leng, I. PastorizaSantos, J. Pérez-Juste, Gold Nanooctahedra with Tunable Size and Microfluidic-Induced 3D Assembly for Highly Uniform SERS-Active Supercrystals. Chemistry of Materials. 27 (2015) 8310-8317. [22] P. Alonso-Gonzalez, M. Schnell, P. Sarriugarte, H. Sobhani, C. H. Wu, N. Arju, A. Khanikaev, F. Golmar, P. Albella, L. Arzubiaga, F. Casanova, L. E. Hueso, P. Nordlander, G. Shvets, R. Hillenbrand, Real-Space Mapping of Fano Interference in Plasmonic Metamolecules. Nano. Lett. 11 (2011) 3922-3926. [23] P. J. De Groot, G. J. Postma., W. J. Melssen., L. M. C. Buydens, V. Deckert, R. Zenobi, Application of principal component analysis to detect outliers and spectral deviations in nearfield surface-enhanced Raman spectra. Anal. Chim. Acta. 446 (2001) 71-83. [24] Å. Rinnan, F. Van den Berg, S. Balling Engelsen, Review of the most common preprocessing techniques for near-infrared spectra. TrAC, Trends Anal. Chem. 28 (2009) 12011222. 36 [25] G. Das, F. Gentile, M. L. Coluccio, A. M. Perri, A. Nicastri, F. Mecarini, G. Cojoc, P. Candeloro, C. Liberale, F. D. Angelis, E. D. Fabrizio, Principal component analysis based methodology to distinguish protein SERS spectra. J. Mole. Struct. 993 (2011) 500-505. [26] R. Goodacre, E. M. Timmins, R. Burton, N. Kaderbhai, A. M. Woodward, D. B. Kell, P. J. Rooney, Rapid identification of urinary tract infection bacteria using hyperspectral wholeorganism fingerprinting and artificial neural networks. Microbiology. 144 (1998) 1157-1170. [27] S. Gómez-Graña, J. Pérez-Juste, R. A. Alvarez-Puebla, A. Guerrero-Martínez, L. M. LizMarzán, Self-assembly of Au@Ag nanorods mediated by gemini surfactants for highly efficient SERS-active supercrystals. Adv. Optical. Mater. 1 (2013) 477-481. [28] P. J. Yunker, T. Still, M. A. Lohr, A. G. Yodh, Suppression of the coffee-ring effect by shape-dependent capillary interactions. Nature. 476 (2011) 308-311. [29] X. Lv, W. Ge, Q. Li, Y. Wu, H. Jiang, X. Wang, Rapid and Ultrasensitive Electrochemical Detection of Multidrug-Resistant Bacteria Based on Nanostructured Gold Coated ITO Electrode. ACS Applied Materials & Interfaces. 6 (2014) 11025-11031. [30] H. J. Chung, T. Reiner, G. Budin, C. Min, M. Liong, D. Issadore, H. Lee, R. Weissleder, Ubiquitous Detection of Gram-Positive Bacteria with Bioorthogonal Magnetofluorescent Nanoparticles. ACS Nano. 5 (2011) 8834-8841. [31] M. Schafer, T. R. Schneider, G. M. Sheldrick, Crystal Structure of Vancomycin. Structure. 4 (1996) 1509-1515. 37 [32] P. J. Loll, A. Derhovanessian, M. V. Shapovalov, J. Kaplan, L. Yang, P. H. Axelsen, Vancomycin Forms Ligand-Mediated Supramolecular Complexes. J. Mol. Biol. 385 (2009) 200211. [33] J. E. Millstone, S. J. Hurst, G. S. Metraux, J. I. Cutle, C. A. Mirkin,Colloidal Gold and Silver Triangular Nanoprisms. Small. 5 (2009) 646-664. [34] G. K. Joshi, P. J. McClory, S. Dolai, R. Sardar, Improved localized surface plasmon resonance biosensing sensitivity based on chemically-synthesized gold nanoprisms as plasmonic transducers. J. Mater. Chem. 22 (2012) 923-931. [35] S. H. Ciou, Y. W. Cao, H. C. Huang, D. Y. Su, C. L. Huang, SERS Enhancement Factors Studies of Silver Nanoprism and Spherical Nanoparticle Colloids in The Presence of Bromide Ions. J. Phys. Chem. C. 113 (2009) 9520-9525. [36] B. Aoune, D. M. Zhang, T. Felicia, J. H. Naomi, Surface-Enhanced Raman Spectroscopy of DNA. J. Am. Chem. Soc. 16 (2008) 5523-5529. [37] L. Yilin, H. Qing, M. Guowen, L. J. Wu, J. Zhang, Label-free selective SERS detection of PCB-77 based on DNA aptamer modified SiO2@Au core/shell nanoparticles. Analyst. 139 (2014) 3083-3087. 1 Supporting Information Synergistic enhancement via plasmonic nanoplate-bacteria-nanorod supercrystals for highly efficient SERS sensing of food-borne bacteria WeiQiang Wang‡a, Ville Hynninen‡b, Li Qiu‡c, AiWen Zhangc, Tibebe Lemmad, NanNan Zhanga, HongHua Ge*a, J. Jussi Toppari*d, Vesa P. Hytönenb and Jin Wang*a a Institute of Health Sciences and School of Life Science, AnHui University, Hefei, Anhui 230601, P.R. China b BioMediTech, University of Tampere, FI-33520 Tampere, Finland and Fimlab Laboratories, FI-35520, Tampere, Finland c Institute of Intelligent Machines, Chinese Academy of Sciences, Hefei, Anhui 230031, P. R. China dUniversity of Jyväskylä, Department of Physics, Nanoscience Center, P.O. Box 35, FI-40014, University of Jyväskylä, Finland ‡ These authors contributed equally to this work. * Corresponding author. Emails: [email protected]; [email protected], [email protected] 2 Experimental details S1.1. Synthesis of Au@Ag core shell nanorods (Au@Ag NRs) AuNRs could be synthesized by modified seed-mediated approach [1]. Firstly, 10 mL 0.5 mM HAuCl4was mixed with 10 mL 0.2 M CTAB. Subsequently, 600 μL 0.02 M ice-cold NaBH4 was added to the mixed solution. Vigorous stirring of the seed solution was kept going for 2min, and was subsequently settled at room temperature for 1h. Secondly, 50 mL 0.2 M CTAB, 2.5 mL 4 mM AgNO3, 50 mL 1 mM HAuCl4 were mixed as a growth solution, followed by addition of 700 μL of 0.08 M vitamin C. Finally, the AuNRs were prepared by addition of 120 μL seed solution to the growth solution. The as–prepared AuNRs were centrifuged and redispersed in 10 mL Milli-Q water. 2 mL AuNRs solution was diluted into 10 mL 0.1 M CTAB aqueous solution. 0.5 mL 4 mM AgNO3 solution, 600 μL 0.1 M vitamin C, 1.2 mL 0.1 M NaOH solution were added to the solution. During this step, the color of solution was changed from yellow-reddish to green, indicating formation of silver shell. The as–prepared bimetallic Au@Ag core-shell nanorods were centrifuged and redispersed in 1 mL Milli-Q water. S1.2. Synthesis of Ag nanoplate Ag nanoplates were synthesized via seed-mediated method [2]. As an initial step, 5mL of 2.5 mM NaCA3 was mixed with a mixture prepared by combining 0.25 mL 500 mg/L PSS and 0.3mL 10mM NaBH4. Subsequently, 5mL of 0.5 mM AgNO3 was added to the mixture with a rate of 2mL/min under constant stirring to obtain the Ag seed solution. As the second step, 75μL 10 mM ascorbic acid was added to 5 mL H2O and various amounts of seed solution was added (40 μL ~ 400μL), followed by addition of 3 mL 0.5 mM AgNO3 with a rate of 1mL/min to form Ag nanoplates. Fig. S1. Normalized UV/Vis/NIR spectra of (A) Au nanorods with LSPR band at 783 nm; (B) Au@Ag core-shell nanorods with LSPR band at 765 nm; (C) Au@Ag core-shell nanorods with LSPR band at 720 nm; (D) Au@Ag core-shell nanorods with LSPR band at 664 nm. The gray line means laser line exciation at 785 nm. The amounts of the deposited Ag and thus the shell thickness increases from B to D. 3 Fig. S2. SEM images of A: plasmonic vertical nanorod supercrystal; B: plasmonic parallel nanorod supercrystal; C: expansion of plasmonic vertical nanorod supercrystal; D: expansion of plasmonic parallel nanorod supercrystal. Fig. S3. SEM images of bacteria attached on the parallel supercrystal of plasmonic nanorods. A: E. coli; B: S. enterica; C: S. xylosus 4 Fig. S4. Point-to-point collected SERS spectra of 2-quinolinethiol (2-QT) on A: a plasmonic vertical nanorod supercrystal, B: parallel nanorod supercrystal 11 Fig. S14. Linear baseline corrected SERS spectra of E. coli. 400 cm-1~1800 cm-1 range; n = 45. Fig. S15. Linear baseline corrected SERS spectra of S. enterica. 400 cm-1~1800 cm-1 range; n = 45. Fig. S16. Linear baseline corrected SERS spectra of S. xylosus. 400 cm-1~1800 cm-1 range; n = 45. 12 Fig. S17. Linear baseline corrected SERS spectra of 1:1 E. coli –S. enterica mixture. 400 cm-1 ~1800 cm-1 range; n = 45. Fig. S18. Linear baseline corrected SERS spectra of 1:1 E. coli –S. xylosus mixture. 400 cm-1 ~1800 cm-1 range; n = 45. Fig. S19. Linear baseline corrected SERS spectra of 1:1 S. enterica –S. xylosus mixture. 400 cm-1-1800 cm-1 range; n = 45. 13 Fig. S20. Linear baseline corrected SERS spectra of 1:2 E. coli –S. enterica mixture. 400 cm-1 ~1800 cm-1 range; n = 45. Fig. S21. Linear baseline corrected SERS spectra of 1:5 E. coli –S. enterica mixture. 400 cm-1 ~1800 cm-1 range; n = 45. Fig. S22. Linear baseline corrected SERS spectra of 1:2 E. coli –S. xylosus mixture. 400 cm-1 ~1800 cm-1 range; n = 45. 14 Fig. S23. Linear baseline corrected SERS spectra of 1:5 E. coli –S. xylosus mixture. 400 cm-1 ~1800 cm-1 range; n = 45. Fig. S24. Linear baseline corrected SERS spectra of 1:2 S. enterica –S. xylosus mixture. 400 cm-1 ~1800 cm-1 range; n = 45. Fig. S25. Linear baseline corrected SERS spectra of 1:5 S. enterica –S. xylosus mixture. 400 cm-1 ~1800 cm-1 range; n = 45. 15 Fig. S26. A: PCA for E. coli – S. xylosus 10:1 mixtures plotted according to the first two PCs. Blue circle = E. coli, red circle = S. xylosus, black empty circle = 1:1 mixture, black filled circle = 10:1 mixture (n = 45 for each sample type). Fig. S27. Linear baseline corrected SERS spectra of 10:1 E. coli –S. xylosus mixture. 400 cm-1 ~1800 cm-1 range; n = 45. 16 Fig. S28. Linear baseline corrected SERS spectra of E. coli suspended in Fanta. 400 cm-1~1800 cm-1 range; n = 45. Fig. S29. Linear baseline corrected SERS spectra of S. xylosus suspended in Fanta. 400 cm-1~1800 cm-1 range; n = 45. 17 Fig. S30. Linear baseline corrected SERS spectra of E-coli grown in Fanta to LB nutrient (1:10). 400 cm-1~1800 cm-1 range; n = 45. Fig. S31. Linear baseline corrected SERS spectra of S. xylosus grown in Fanta to LB nutrient (1:10). 400 cm-1~1800 cm-1 range; n = 45. Fig. S32. Linear baseline corrected SERS spectra of 1:1 E. coli –S. xylosus mixture suspended in Fanta. 400 cm-1 ~1800 cm-1 range; n = 45. 18 Tab S1. List of dominant band assignments and description of relevant biochemical components of the SERS spectrum of Salmonella enterica vs-very strong, s-strong, m-medium, w-weak, vw-very weak, sym stretch-symmetric stretch, deform-deformation Wavenumber(cm-1) Assignment [3-8] Biochemical components of S. enterica bacteria 653 vs ring breathing guanine base 727 s ring breathing adenine of nucleic acids 749 m ring breathing thymine of nucleic acids 891 m C-O sym stretch, C-O-H sym deform and lipid layer of cell wall and membrane C-O-C sym deform 990~1152 vw P=O sym stretch nucleic aicds, phospholipids 1159~1288 m C-N sym stretch, N-H sym deform protein C=O sym stretch, Amide III 1327 m ring stretch cytosine and uracil of nucleic acids 1411 m C=O sym stretch peptidoglycan and disaccharide-pentapeptide of cell wall lipopolysaccharide of outer membrane 1454 m C-H deform lipid, carbohydrate and protein 1690 w C=O stretch, amide I (α-helical structure ) protein 19 Tab S2. List of dominant band assignments and description of relevant biochemical components of the SERS spectrum of Escherichia coli Wavenumber(cm-1) Assignment [3-8] Biochemical components of E. coli bacteria 658 vs ring breathing guanine base 730 m ring breathing adenine of nucleic acids 751 m ring breathing thymine of nucleic acids 897 m C-O sym stretch, C-O-H sym deform and lipid layer of cell wall and membrane C-O-C sym deform 1155-1298 m C-N sym stretch, N-H sym deform protein C=O sym stretch, Amide III 1331 m ring stretch cytosine and uracil of nucleic acids 1416 m C=O sym stretch peptidoglycan and disaccharide-pentapeptide of cell wall lipopolysaccharide of outer membrane 1444 m C-H deform lipid, carbohydrate and protein 1664 w C=O stretch, amide I (α-helical structure ) protein 20 Tab S3. List of dominant band assignments and description of relevant biochemical components of the SERS spectrum of Staphylococcus xylosus Wavenumber(cm-1) Assignment [3-8] Biochemical components of S. xylosus bacteria 666 vw ring breathing guanine base 730 s ring breathing adenine of nucleic acids 850-940 m C-O sym stretch, C-O-H sym deform and lipid layer of cell wall and membrane C-O-C sym deform 1241-1307 m C-N sym stretch, N-H sym deform protein C=O sym stretch, Amide III 1327 s ring stretch cytosine and uracil of nucleic acids 1449 m C-H deform lipid, carbohydrate and protein 1678 vw C=O stretch, amide I (α-helical structure ) protein References: 1. B. Nikoobakht, M. A. El-Sayed, Preparation and Growth Mechanism of Gold Nanorods (NRs) using Seed-Mediated Growth Method. Chem. Mater. 15 (2003) 1957-1962. 2. D. Aherne, D. M. Ledwith, M. Gara, J. M. Kelly, Optical Properties and Growth Aspects of Silver Nanoprisms Produced by a Highly Reproducible and Rapid Synthesis at Room Temperature. Adv. Funct. Mater. 18 (2008) 2005-2016. 3. A.C .Williams, H. G. M. Edwards, Fourier Transform Raman Spectroscopy of Bacterial Cell Walls. J. Raman. Spectrosc. 25 (1994) 673-677. 5. W. R. Premasiri, D. T. Moir, M. S. Klempner, N. Krieger, G. Jone II, L. D. Ziegler, Characterization of the Surface Enhanced Raman Scattering (SERS) of Bacteria. J. Phys. Chem. B. 109 (2005) 312-320. 6. K. Maquelin, L. P. Choo-Smith, T. Van vreeswijk, H. P. Endz, B. Smith, R. Bennett, H. A. Bruining, G. J. Puppels, Raman Spectroscopic Method for Identification of Clinically Relevant Microorganisms Growing on Solid Culture Medium. Anal. Chem. 72 (2000) 12-19. 7. K. Maquelin, C. Kirschner, L. P. Choo-Smith, N. Van den Braak, H. P. Endtz, D. Naumann, G. J. Puppels, Identification of Medically Relevant Microorganisms by Vibrational Spectroscopy. J. Microbiol Meth. 51 (2002) 255-271. 8. B. Schrader, ed. 1995, Infrared and Raman Spectroscopy: Methods and Applications. VCH Publishers: NY.