Flexible 3D Plasmonic Web Enables Remote Surface Enhanced Raman Spectroscopy

Abstract Nanoplasmonic materials concentrate light in specific regions of dramatic electromagnetic enhancement: hot spots. Such regions can be employed to perform single molecule detection via surface‐enhanced Raman spectroscopy. However, this phenomenon is challenging since hot spots are expected to be highly intense/abundant and positioning of molecules within such hot spots is crucial to manage with ultrasensitive SERS. Herein, it is discovered that a 3D plasmonic web embedded within a biohybrid (3D‐POWER) exhibits plasmonic transmission, spontaneously absorbs the analyte, and meets these so much needed criteria in ultrasensitive SERS. 3D‐POWER is built with nanopaper and self‐assembled layers of graphene oxide and gold nanorods. According to in silico experiments, 3D‐POWER captures light in a small region and performs plasmonic field transmission in a surrounding volume, thereby activating a plasmonic web throughout the simulated volume. The study also provides experimental evidence supporting the plasmonic field transport ability of 3D power, which operates as a SERS signal carrier (even beyond the apparatus field of view), and the ultrasensitive behavior of this ecofriendly and flexible material facilitating yoctomolar limit of detection. Besides, 3D‐POWER is proven useful in food and biofluids analysis. It is foreseen that 3D‐POWER can be employed as a valuable platform in (bio)analytical applications.


Figure S5.
In silico experiments demonstrating the electromagnetic field distribution in BC/GO35/AuNRs, frontal plane.The panels display 10 different scenarios, where, considering the respective particle density (see Table S3), the AuNRs were randomly oriented.

Figure S6.
In silico experiments demonstrating the electromagnetic field distribution in BC/AuNRs, horizontal plane.The panels display 10 different scenarios, where, considering the respective particle density (see Table S3), the AuNRs were randomly oriented.

Sebum component
Range Raman signal (cm -1 ) Selection of the optimal SERS substrate.We evaluated the SERS performance of the fabricated substrates by incubating millimetric circles (diameter, 6 mm) of the substrates (BC/AuNRs, BC/GO35/AuNRs, BC/GO80/AuNRs BC/GO125/AuNRs) inside a microtube containing 1 mL of fluorescein (free acid), (FSC, CAS 2321-07-5), as a model analyte, at different concentrations, blanks samples were also incubated in this series of experiments.According with Wang [4] and Batistela [48] , FSC has three typical pKa values.Each pH range displays one or more predominant species (see Figure S8a).Under these conditions, using ethanol as a solvent, the fingerprint of two main species, phenolate and carboxylate of FSC was observed in the SERS spectra, see Figure S8b.
We firstly investigated the performance of BC/GO125/AuNRs as a SERS substrate.In these series of measurements, the intensity of the D and G Raman bands of GO masked the fingerprint of FSC even at different FSC concentrations, Figure S9c.Lower exposure times and number of acquisitions were also explored with similar results.BC/GO35/AuNRs and BC/GO80/AuNRs were also evaluated using different concentrations of FSC, from 10 -4 M to 10 -21 M respectively, see Figure S9a-b.In both cases, we realized that we were able to detect extremely low concentrations of FSC, even at the zeptomolar (zM) range (see Table S6).
However, in general, BC/GO35/AuNRs had the highest Raman intensity, the lowest coefficient of variation (CV, which was about 0.17 ± 0.08), and minimal Raman interference such as fluorescence and shot noise, see Figure S9a-b.

Figure S1 .
Figure S1.Raman spectra recorded on bacterial nanocellulose (BC) as a substrate.______________ Figure S2.SEM micrographs of different configurations of the explored biohybrid: superficial view..

Figure S3 .
Figure S3.SEM micrographs of different configurations of the explored biohybrid: transversal view..

Figure S4 .Figure S5 .Figure S6 .Figure S11 .
Figure S4.In silico experiments demonstrating the electromagnetic field distribution in BC/GO35/AuNRs, horizontal plane.__________________________________________________ Figure S5.In silico experiments demonstrating the electromagnetic field distribution in BC/GO35/AuNRs, frontal plane._____________________________________________________ Figure S6.In silico experiments demonstrating the electromagnetic field distribution in BC/AuNRs, horizontal plane.________________________________________________________________ Figure S7.In silico experiments demonstrating the electromagnetic field distribution in BC/AuNRs, frontal plane.___________________________________________________________________ Figure S8.UV-Vis characterization of the synthesized materials._____________________________ Figure S9.a. Species predominance diagram of FSC. b.SERS spectra facilitated by 3D-POWER.____ Figure S10.SERS analysis of FSC at several concentrations._________________________________ Figure S11.Analysis of repeatability and signal intensity robustness.(a-c) SERS spectra of FSC at different concentrations, obtained with 3D-POWER (BC/GO35/AuNRs) fabricated in different batches._______________________________________________________________________ Figure S12.SERS spectra of FSC at several concentrations analyzed with different SERS substrates.

Figure S1 .
Figure S1.Raman spectra recorded on bacterial nanocellulose (BC) as a substrate.The substrate (BC) was incubated overnight in 1 mL to Fluorescein (FSC) or Glyphosate (GLY) concentrated at 1.0X10 -3 M and 1.8X10 -3 M respectively.Each spectrum represents the mean of fifty spectra recorded on 2500 µm 2 of the corresponding substrate.Excitation wavelength, 785 nm; laser power 0.08 mW; size of the spot = 1.2 µm; exposure time, 2 s. number of acquisitions, 10.

Figure S2 .
Figure S2.SEM micrographs of different configurations of the explored biohybrid: superficial view.First column, BC/AuNRs; second column BC/AuNRs fabricated in the presence of 35 µg mL-1 of GO; third column BC/AuNRs fabricated in the presence of 80 µg mL-1 of GO and forth column BC/AuNRs fabricated in the presence of 125 µg mL-1 of GO.

Figure S4 .
Figure S4.In silico experiments demonstrating the electromagnetic field distribution in BC/GO35/AuNRs, horizontal plane.The panels display 10 different scenarios, where, considering the respective particle density (see TableS3), the AuNRs were randomly oriented.

Figure S9. a .
Figure S9.a. Species predominance diagram of FSC. b.SERS spectra facilitated by 3D-POWER.The SERS substrate (BC/GO35/AuNRs) was incubated overnight in 1 mL of FSC concentrated at 10 -9 M. The spectrum represent the mean of fifty spectra recorded on 2500 µm 2 of the corresponding SERS substrate, obtained through: excitation wavelength, 785 nm; laser power 0.08 mW; size of the spot = 1.2 µm; exposure time, 2 s; number of acquisitions, 10.

Figure S10 .
Figure S10.SERS analysis of FSC at several concentrations.Different SERS substrates were employed, (a) BC/GO35/AuNRs, (b) BC/GO80/AuNRs and (c) BC/GO125/AuNRs.Each spectrum represents the mean of ten spectra recorded on random sites of the corresponding SERS substrate.The spectra were obtained with an excitation wavelength of 785 nm; laser power, 0.08 mW; size of the spot, 1.2 µm; exposure time, 2 s; number of acquisitions, 10.The resulting SERS intensity of the peak at the Raman shift of 1616 cm -1 in BC/GO35/AuNRs (d), BC/GO80/AuNRs (e) and BC/GO125/AuNRs (f), respectively.

Figure S11 .
Figure S11.Analysis of repeatability and signal intensity robustness.(a-c) SERS spectra of FSC at different concentrations, obtained with 3D-POWER (BC/GO35/AuNRs) fabricated in different batches.a. batch 1, b. batch 2 and c. batch 3.Each spectrum represents the mean of ten spectra recorded on random sites of the corresponding SERS substrate.The spectra were obtained with an excitation wavelength of 785 nm; laser power, 0.08 mW; size of the spot, 1.2 µm; exposure time, 2 s; number of acquisitions, 10. d-f.The resulting SERS intensity of the peak at the Raman shift of 1616 cm -1 in batch 1 (d.), batch 2 (e.) and batch 3 (f.),respectively.

Figure S12 .
Figure S12.SERS spectra of FSC at several concentrations analyzed with different SERS substrates.a. FSC analysis using BC/AuNRs.c. FSC analysis using BC/GO35/AuNRs.Each spectrum represents the mean of fifty spectra recorded throughout 2500 µm 2 of the corresponding SERS substrate.The spectra were obtained with an excitation wavelength of 785 nm; laser power, 0.08 mW; size of the spot, 1.2 µm; exposure time, 2 s; number of acquisitions, 10.The resulting SERS intensity of the peak at the Raman shift of 1616 cm -1 in BC/AuNRs (b.) and BC/GO35/AuNRs (d.), respectively.

Figure S13 .
Figure S13.Detailed SERS spectra of FSC analyzed at different concentrations using 3D-POWER (BC/GO35/AuNRs).Each spectrum represents the mean of fifty spectra recorded on 2500 µm 2 of the corresponding SERS substrate.The spectra were obtained through excitation wavelength, 785 nm; laser power 0.08 mW; size of the spot = 1.2 µm; exposure time, 2 s and number of acquisitions, 10.

Figure
Figure S14.a. Species predominance diagram of GLY, b.SERS spectra facilitated by 3D-POWER.The SERS substrate (BC/GO35/AuNRs) was incubated overnight in 1 mL of GLY concentrated at 10 -9 M. The spectrum represent the mean of fifty spectra recorded on 2500 µm 2 of the corresponding SERS substrate obtained through: excitation wavelength, 785 nm; laser power 0.08 mW; size of the spot = 1.2 µm; exposure time, 2 s. number of acquisitions, 10.

Figure S15 .
Figure S15.Detailed SERS spectra of GLY analyzed at different concentrations using 3D-POWER (BC/GO35/AuNRs).Each spectrum represents the mean of fifty spectra recorded on 2500 µm 2 of the corresponding SERS substrate.The spectra were obtained through: excitation wavelength, 785 nm; laser power 0.08 mW; size of the spot = 1.2 µm; exposure time, 2 s and number of acquisitions, 10.

Figure S16 .
Figure S16.In silico experiments were performed to investigate the light transport capability of the materials composing 3D-POWER.The simulations were performed by establishing the pumped region at the upper right corner of the corresponding material.a. Simulation of a plane depicting the light transport capabilities of BC. b.Simulation of a plane depicting the light transport capabilities of BC/GO.c. Simulation of a plane depicting the plasmonic transmission of BC/AuNR.d.Simulation of a plane depicting the plasmonic transmission of 3D-POWER (BC/GO35/AuNR).e-f.3D simulations depicting the plasmonic field transport capabilities of 3D-POWER throughout different volumes.

Figure S17 .Figure S18 .
Figure S17.3D simulations (x = 800 nm, y = 800 nm and z = 50 nm) revealed that 3D-POWER was able to capture light in a small region (5 x 10 5 nm 3 ) and perform plasmonic field transmission, thereby generating a plasmonic web throughout the simulated volume.AuNRs were randomly placed.a. 3D simulation highlighting plasmonic transmission at the bottom of the simulated volume (excitation perpendicular to the self-assembled layers of 3D-POWER).b. 3D simulation highlighting plasmonic transmission at the top of the simulated volume (excitation perpendicular to the self-assembled layers of 3D-POWER).c. 3D-In silico experiment demonstrating highly efficient transport of surface plasmon polaritons in 3D-POWER under an excitation (top corner-vertical) that is parallel to the self-assembled layers of 3D-POWER.

Figure S19 .
Figure S19.Remote SERS experiment using a rectangular piece of 3D-POWER.a. 0.5 µL of FSC concentrated 10 -9 M (or 9.5X10 8 molecules) were drop casted onto the corner of the substrate.b.A Raman mapping in the frontier of the substrate that was not reached by the sample was performed.Signal intensities depicted in the Raman mapping correspond to the intensity at the Raman shift of 1500 cm -1 .The spectra were obtained through excitation wavelength, 785 nm; power 0.08 mW; size of the spot = 1.2 µm; exposure time, 2 s; number of acquisitions 10.

Figure
Figure S20.a. 0.1 µL of FSC concentrated at 10 -16 M (c.a. 10 molecules) were drop casted onto the center of a circular piece of 3D-POWER.Raman mapping with a resolution of 500 µm per step was recorded throughout the diameter (6 mm) of the substrate.b.Signal intensities depicted in the Raman mapping correspond to the intensity at the Raman shift of 1500 cm -1 .The spectra were obtained through excitation wavelength, 785 nm; power 0.08 mW; size of the spot = 1.2 µm; exposure time, 2 s and number of acquisitions, 10.

Figure S21 .
Figure S21.Corn analysis.a. SERS spectra of commercially available white cornmeal, incubated at different concentrations.b.SERS spectra of commercially available blue cornmeal, incubated at different concentrations.c.SERS spectra of non-commercially available cornmeal incubated at different concentrations.Each spectrum represent the means of at least one hundred spectra recorded in an area of 2500 µm 2 of the corresponding SERS substrate.The spectra were obtained through an excitation wavelength of 785 nm; laser power 0.08 mW; size of the spot = 1.2 µm; exposure time, 2 s and number of acquisitions, 10.

Figure S22 .
Figure S22.Pesticide detection using 3D-POWER as SERS substrate in commercially available cornmeal (white corn: Sample A; blue corn: Sample B) and non-commercially available white cornmeal (Sample C).The samples were diluted in the order of attograms (10 -18 ) per milliliter.Each spectrum represent the average of at least one hundred spectra recorded in an area of 2500 µm 2 of the corresponding SERS substrate.The spectra were obtained through an excitation wavelength of 785 nm; laser power, 0.08 mW; size of the spot, 1.2 µm; exposure time, 2 s and number of acquisitions, 10.

Figure S23 .
Figure S23.Food analysis.a. lettuce, b. tomato, c. apple, d.Mango and e. Orange.Each spectrum represent the average of at least 60 spectra recorded in an area of 2500 µm 2 of the corresponding SERS substrate.The spectra were obtained through an excitation wavelength of 785 nm, laser power, 0.8 mW; exposure time, 2 s; size of the spot, 1.2 µm and number of acquisitions, 10.

Table of content
Scheme S1.Fabrication of BC/GO substrate.i. BC is washed with ultrapure water.ii.BC is added in a GO suspension.GO is spontaneously self-assembled by layers within BC. iii.BC/GO is washed and dried.

Table S1
. (Nano)metrology and physicochemical features of each component of 3D-POWER _____ Table S2.UV-Vis features of the explored SERS substrates._________________________________ Table S3.Particle density in the studied biohybrids.______________________________________ Table S4 Mean values and ranges of the resulting electromagnetic enhancement factors () in the explored materials.______________________________________________________________ Table S5 Estimation of the number of FSC molecules._____________________________________

Table S6 .
Statistical analysis of the explored SERS substrates.FSC was employed as a model analyte.______________________________________________________________________________

Table S7 .
Assessment of the behavior of BC/GO35/AuNRs in different batches.FSC was employed as a model analyte.__________________________________________________________________

Table S8 .
Vibrational bands observed in the SERS spectra of FSC (diluted in ethanol).____________

Table S10 .
Estimation of the number of GLY molecules.___________________________________

Table S13 .
Material properties considered in the in silico experiments._______________________

Table S14 .
Vibrational bands observed using 3D-POWER, which were rubbed into the surface of vegetables and fruits.____________________________________________________________

Table S17 .
Cost estimation to 3D-POWER fabrication

______________________________________ Selection of the optimal SERS substrate. _______________________________________________ 4 Scheme S1. Fabrication
of BC/GO substrate.i. BC is washed with ultrapure water.ii.BC is added in a GO suspension.GO is spontaneously self-assembled by layers within BC. iii.BC/GO is washed and dried.Scheme S2.Gold nanorods (AuNRs) synthesis (by growth seed method).

Table S1 .
(Nano)metrology and physicochemical features of each component of 3D-POWER

Table S3 .
Particle density in the studied biohybrids.GO, graphene oxide; AuNRs, gold nanorods.The number of AuNRs per µm 2 was determined by analyzing SEM images (superficial and transversal views, respectively).

Table S4
Mean values and ranges of the resulting electromagnetic enhancement factors () in the explored materials.

Table S5
Estimation of the number of FSC molecules.

Table S6 .
Statistical analysis of the explored SERS substrates.FSC was employed as a model analyte.GO, graphene oxide; AuNRs, gold nanorods, FSC, fluorescein; SD, standard deviation; CV, coefficient of variation.Raman intensities were measured at the Raman shift 1616 cm -1 .

Table S7 .
Assessment of the behavior of BC/GO35/AuNRs in different batches.FSC was employed as a model analyte.GO, graphene oxide; AuNRs, gold nanorods, FSC, fluorescein; SD, standard deviation; CV, coefficient of variation.Raman intensities were measured at the Raman shift 1616 cm -1 .

Table S8 .
Vibrational bands observed in the SERS spectra of FSC (diluted in ethanol).

Table S9 .
Analytical enhancement factor observed in FSC via 3D-POWER.

Table S10 .
Estimation of the number of GLY molecules.

Table S11 .
Vibrational bands observed in in the analysis of GLY concentrated at 10 -9 M (diluted in HPLC grade water).

Table S14 .
Vibrational bands observed using 3D-POWER, which were rubbed into the surface of vegetables and fruits.

Table S15 .
Vibrational bands observed using 3D-POWER substrates, which were impregnated with sweat for ten seconds in the temple of 4 volunteers.

Table S17 .
Cost estimation to 3D-POWER fabrication Bacterial Nanocelullose; GO, Graphene Oxide; AgNO3, Silver Nitrate; NaBH4, Sodium borohydride; HAuCl4, Hydrogen tetrachloroaurate; CTAB, Hexadecyl trimetylammonium bromide; AA, Ascorbic Acid.*The cost was estimated by considering the weight of each reagent employed in the fabrication of one batch with 15 pieces of cellulose, each piece has nine circular SERS substrates.