Surface Enhanced Raman Spectroscopy of Molecules Using Colloidal Silver Nanoparticles
BSc. (Hons) Year 4 Undergraduate Research Project
School of Engineering and Physical Sciences, Heriot Watt University
Edinburgh, EH14 4AS
Student: Luke Joseph Mulvey
Registration No: H00155119
Email: [email protected]
Submission Date: 03/03/2017
Supervisor: Lynn Paterson
Moderator: Rory Duncan
Table of Contents
BSc. (Hons) Year 4 Undergraduate Research Project 1
School of Engineering and Physical Sciences, Heriot Watt University 1
Edinburgh, EH14 4AS 1
Student: Luke Joseph Mulvey 1
Registration No: H00155119 1
Email: [email protected] 1
Submission Date: 03/03/2017 1
Supervisor: Lynn Paterson 1
Moderator: Rory Duncan 1
1. Introduction 1
2. Background & Theory 2
2.1 Raman Scattering 2
2.2 Raman Spectrometer 4
2.3 SERS (Surface Enhanced Raman Spectroscopy) 5
3. Experimental Methods & Materials 7
3.1 Synthesis of AgNPs (Silver Nanoparticles) 7
3.2 SEM (Scanning Electron Microscope) 7
3.3 Centrifuging AgNPs 8
3.4 Commercial AgNPs 9
3.5 Silver Substrates 9
3.6 Raman Spectroscopy 9
3.7 SERS Using Colloidal AgNPs 10
4. Results & Discussion 11
4.1 Synthesised AgNPs 12
4.1.1 SEM Images 12
4.1.2 Raman Results of R6G and PS with Synthesised AgNPs 14
4.2 NM300 AgNPs 15
4.2.1 Raman Results of PS with NM300 AgNPs 15
4.2.2 Raman Results of R6G with NM300 AgNPs 16
4.3 Premade Silver Substrates 18
5. Conclusions 20
6. Future Work 20
6.1 Short Term 20
6.2 Long Term 20
I would first and foremost like to thank Dr. Lynn Paterson for all of her guidance and support throughout my research project. I would also like to thank Mark Mackenzie for demonstrating the use of the Raman microscope as well as the synthesis of silver nanoparticles, Mark Leonard for operating the SEM to look at my samples, Faisal Alqahtani for providing me with silver nanoparticles and Howard Chi for the use of silver substrates produced by him.
The use of colloidal AgNPs (silver nanoparticles) has been investigated to enhance the Raman signal detected from a sample. Silver nanoparticles were synthesised and characterised using SEM. Nanoparticles had sizes ranging from below 10 nm up to 1 micron. Using commercially produced NM-300 silver nanoparticles with a maximum size of 20 nm and mean diameter of ?15 nm, we have been able to perform SERS (surface enhanced Raman spectroscopy) and obtain enhancements in the range of 103 times in Raman signals obtained from samples of 1mM R6G (Rhodamine-6G), using a 50 mW 785nm, near-infrared laser at 10% power. Using a previously made SERS substrate covered in roughened silver, we observed enhancements in the range of 104 times of R6G’s Raman spectrum using the same laser at 1% power.
Raman spectroscopy is a powerful analytical technique based on collecting the inelastically scattered light from a sample, that can be used to identify the molecular composition of the sample from its unique spectral fingerprint. In recent years, Raman spectroscopy has become widespread in use in a multitude of fields such as forensics, pharmaceutics, microbiology and chemistry, due to improvements and increasing affordability of laser and optical technology1. One application of Raman spectroscopy which is of particular interest in our research is its potential use in conjunction with microfluidics in point-of-care biomedical devices for the rapid detection of disease in patients.
One issue with Raman scattering is that it is a weak phenomenon which has an intensity that is generally several orders of magnitude weaker than that of fluorescence emission in most cases and ?1000 times weaker than Rayleigh scattering. Due to the low intensity of Raman signals, Raman spectroscopy has been limited by the sensitivity limits of detectors and the intensity of the Raman signals from the samples.
In SERS (Surface Enhanced Raman Spectroscopy), the Raman signal of a sample is greatly enhanced by being placed on or near a roughened metal surface. The enhancements of the signal due to SERS can be as high as one billion times, allowing single molecules to be detected2.
The aim of this project is to demonstrate SERS, using colloidal silver nanoparticles as a SERS substrate to enhance the Raman signals of molecules that are known to give weak Raman signals, such as R6G (Rhodamine-6G). To generate SERS spectra, silver nanoparticles are synthesised and used as colloidal SERS substrates, commercially available silver nanoparticles are also used, and for comparison, in-house made roughened silver surfaces are also used.
2. Background & Theory
2.1 Raman Scattering
When a photon interacts with a molecule, it can be scattered elastically (Rayleigh Scattering) in which case the energy and thus the frequency and wavelength of the photon after the interaction remain the same. Approximately 1 in 10 million scattered photons are inelastically scattered (Raman Scattering) by an excitation of the molecule, in which case the frequency of the photon changes. If the photon loses energy to the molecule in the inelastic scattering, it is known as ‘Stokes Scattering’ and its wavelength will be longer, but if it gains energy from the molecule it is known as ‘Anti-Stokes Scattering’ which is much less common, as a consequence of the Boltzmann distribution, and its wavelength will be shorter3.
The inelastic exchange of energy between the photons and molecules is due to the photon interacting with vibrational, rotational and electronic energies of the molecule, although we are most interested in the vibrational energies.
When a photon interacts with a molecule it can be momentarily absorbed, causing a transition of the vibrational energy from the ground state to a virtual state, and a new photon is then created and scattered by a transition from this virtual state. In Rayleigh scattering, a transition from the virtual state to the ground state will cause the emission of an elastically scattered photon, with the same energy as the incident photon (i.e. the same wavelength as the incoming light), which is what happens in most cases.
Rayleigh Scattering: ?E_i=E?_e= h?_0 (eq. 1)
In this equation, E_i is the energy of the incident photon, E_e is the energy of the emitted photon and ?_0 is the frequency of the incident photon. The energy of the incident and emitted photons are equal, so this is an elastic scattering. In Fig. 1, we see this as a photon being absorbed, the molecule being excited to the virtual state, and a photon with the same energy as the incident photon emitted.
In Raman scattering, a transition from the virtual state to the first excited vibrational state causes the inelastic emission of a photon which will have energy:
Stokes Scattering: E_e= h?(??_0- ?_v) (eq. 2)
?h??_v is the energy required to raise the vibrational energy of the molecule to the 1st excited state, which is the amount of energy transferred from the incident photon to the molecule resulting in an emitted photon with less energy. This case is known as Stokes scattering, which is see in fig 1.
In Raman scattering, if a photon is absorbed by a molecule that is already in the 1st excited energy level, a photon will be emitted with energy:
Anti-Stokes Scattering: E_e= h?(??_0+ ?_v) (eq. 3)
The emitted photon has gained energy h?_v from the molecule, which is the energy for the vibrational energy level to decrease to the ground state from the 1st excited state. This is known as Anti-Stokes scattering.
Fig. 1 Vibrational energy level diagram of a molecule showing the excitation of the vibrational energy level to a virtual state caused by an incident photon and the emission of scattered photons. The arrow thicknesses represent the likelihood of each type of scattering with Rayleigh scattering being most likely and Anti-Stokes being least likely.
2.2 Raman Spectrometer
In Raman spectroscopy, a Raman spectrometer is used to obtain the Raman spectrum of a sample to determine its molecular structure. In a Raman spectrometer, a high-intensity monochromatic light source (i.e. a laser) is shone through an objective lens onto the sample. The elastic scattered light is filtered out using a notch filter, while the inelastically scattered light passes onto a mirror where it is reflected onto a dispersion grating or prism that separates the light into its constituent wavelengths. The separated wavelengths are then detected by a CCD (charge-coupled detector), although the signal at this point will be very weak (106 times less intense than the laser excitation light), so the signal must be amplified before it is sent to a computer where the data is processed to create a Raman spectrum of the sample.
Fig. 2 Diagram of a Raman spectrometer, adapted from: [http://www.chem.umd.edu/wp-content/uploads/2014/01/RamanSpectroscopy.jpg] from the University of Maryland
2.3 SERS (Surface Enhanced Raman Spectroscopy)
2.3.1 Physical Theory of SERS
SERS is a technique that greatly enhances the Raman signals of molecules that are adsorbed on roughened surfaces or particles, usually of noble metals (eg. gold & silver due to their high reflectivity), with enhancements of up to 106 times being possible, allowing the detection of single molecules in some cases. The strength of the effect in the case of colloidal metal particles, is strongly dependant on particle size, shape and uniformity and in the case of metal surfaces is dependant on the surface roughness and in both cases is dependant on the type of metal used and on the laser wavelength selected.
There are two different effects that are said to be responsible for surface enhancement, namely chemical enhancement and, field enhancement, the latter of which is mostly responsible for the surface enhancement, so we will focus our discussion on this effect.
Field Enhancement: The strength of the Raman signal of a molecule is dependant on the strength of the incoming oscillating electromagnetic field (i.e. the intensity of the laser light beam used to excite the sample), so a higher powered laser would create a stronger Raman signal. When laser light of a specific wavelength is incident on rough metal surfaces, a field enhancement occurs, increasing the electromagnetic field experienced by the adsorbed molecules, thus enhancing the Raman signal of these molecules.
If nano-sized metal particles are positioned in an oscillating electromagnetic field (i.e. a laser beam), the electrons on the surface of the metal particles can oscillate in phase with the field, forming surface plasmons. These oscillating electrons generate an electromagnetic field that is in phase with the electrons and thus the the laser beam, thus the beam is amplified and analyte molecules experience a stronger field and thus produces a stronger Raman signal.
The effect works best if the analyte molecules are directly in contact with the metal particles/ surfaces. The field enhancement provided is proportional to r/d^3 , where d is the distance between the particle and the molecule and r is the radius of the metal particle, so the enhancement falls off exponentially with distance by a factor of d3. Larger particles can have produce a stronger enhancement, but the effect only holds for particles that are sufficiently small compared to the laser wavelength and have a size in the region of r<?/10.
Fig. 3 SERS of a molecule using a small metal particle as a SERS substrate to enhance the Raman signal of an analyte molecule
2.3.2 Current Research in SERS
Nebogatikov et al. have used laser ablation to produce micron sized rings of silver nanoparticles and have seen SERS enhancement factors of up 105 times in organic molecules such as R6G.
3. Experimental Methods & Materials
3.1 Synthesis of AgNPs (Silver Nanoparticles)
The following method4 was used to synthesize colloidal silver nanoparticles with all solutions being prepared using distilled water and analytical grade chemicals that needed no purification.
First, the solutions had to be prepared. To prepare 50ml of a 1% solution of trisodium citrate, 0.500g of trisodium citrate powder was added to 50ml of de-ionised water. To prepare 5ml of a 0.001M solution of silver nitrate, 0.5ml of 0.1M silver nitrate was added to 4.5ml of de-ionised water.
After these solutions were prepared, 50 ml of .001 M silver nitrate (AgNO3) solution was brought to boiling temperature in a beaker. 5 ml of 1 % trisodium citrate (C6H5O7Na3) solution was added drop-wise from a pipette to the silver nitrate while vigorously stirring and continuing to heat the mixture until a colour change to a pale yellow colour was evident. The mixture was then removed from the heating element and stirred until it had cooled to room temperature. A second sample was produced by heating it for longer amount of time during its synthesis, in order to obtain a sample with a larger average size of nanoparticles, which was determined by the colour of the sample turning to a brown colour.
The mechanism of the reaction can be expressed as follows4:
4Ag+ + C6H5O7Na3 + 2H2O ? 4Ag0 + C6H5O7H3 + 3Na+ + H+ + O2? (eq. 7)
Fig. 4: Apparatus for synthesising silver nanoparticles
3.2 SEM (Scanning Electron Microscope)
A SEM is a type of microscope that uses a focused beam of electrons to scan a sample and image it. Optical microscopy has a resolution limit of approximately 2000 Å, but a SEM can achieve resolutions higher than 1 nanometer. A SEM was used to characterise our two samples of synthesised colloidal AgNPs that were deposited onto glass slides. The two samples were imaged, and the first sample was also imaged a second time after being given a gold coating by DC sputtering, which is used in SEM microscopy to make the surface of the sample electrically conductive, which is required to prevent the accumulation of charge on non-conducting surfaces, which creates scanning faults and imaging artefacts.
3.3 Centrifuging AgNPs
In SERS it is important that colloidal metal particles have uniform shape and size, to achieve good surface enhancement5. A centrifuge can be used to separate particles of different sizes in a solution. A centrifuge was used to obtain particles of 50 nm and lower in size.
A 1 ml sample of our first batch of synthesised AgNPs and a 1 ml sample of a stock sample of silver nanoparticles (NM300) were placed into 1.5ml centrifuge tubes on opposite sides from each other in the centrifuge. According to Fig. 8 (which was obtained from the website of Cytodiagnostics inc, a Canadian company that produces nanomaterials for research purposes), to obtain particles of 50nm and below in size, we needed a G force of 1,800 g for 30 mins. Using a nomogram, we approximated the required Rpm to obtain 1,800 g of force, in a centrifuge with a 4 cm radius, to be 5000 Rpm.
Fig. 8: Appropriate G forces and times for the centrifugation of 1 ml samples of AgNPs in
1.5 ml centrifuge tubes in a standard centrifuge to obtain nanoparticles of a specific size, taken from: http://www.cytodiagnostics.com/store/pc/Silver-Nanoparticle-Handling-and-Storage-d13.htm
Particles that were presumed to be larger than 50 nm in size were seen to sediment at the bottom of both samples of AgNPs. The supernatants were removed and sediments re-suspended in purified water. Both supernatants (? 50 nm) and re-suspended sediments (?50 nm) from the synthesised AgNPs as well as the stock ample of AgNPs were used in initial experiments.
3.4 Commercial AgNPs
Commercially produced NM-300 AgNPs were used in our investigation. NM-300 is a dispersion of AgNPs of <20 nm originating from a single batch of commercially manufactured nano-silver. The particles were about 15 nm in size with a narrow size distribution and were known to be stable for up to 12 months. Two different samples of NM300 were used, one was on old sample, produced in 2011 and the other was a more recently produced sample. These samples were sonicated in a sonic bath prior to use, to break up aggregates that had formed in the samples.
3.5 Silver Substrates
Premade silver substrates were used in our investigation. They were made for Howard Chi in May 2016 for his MSc project. Substrates were made using laser photoreduction of a silver salt solution on a fused silica substrate, by using a femtosecond laser to scan over a solution of trisodium citrate and silver nitrate with a ratio of 3:1. Substrates were created by scanning repeatedly over the same region with a scan speed of 20 ?m/s and with laser powers of 10, 15, 20, 30 and 40 mW to create 5 separate substrates. A solution of 1mM R6G was deposited onto these substrates and allowed to dry.
Fig. : 5 Silver substrates, produced using laser powers varying from 10-40 mW with deposited R6G visible
3.6 Raman Spectroscopy
A Renishaw inVia Raman microscope was used to to generate Raman spectra, and Renishaw’s WiRE™ software was used to control the Raman acquisitions and to record and analyse Raman spectra. There were two lasers that could be used with the microscope, a
50 mW, 514 nm, green laser and a 50 mW, 785 nm, near-infrared laser, although the 785 nm laser was used for most of the experiments due to the background fluorescence caused by the 514 nm laser.
Fig. 10 Renishaw inVia Raman microscope, with microscope and stage controls on
the left hand side and the spectrometer in its casing in the centre
Before the Raman microscope could be used it needed to be calibrated using a piece of silicon that had a known peak of 520 cm-1. This is very important in situations where minute differences in wavenumbers are being investigated, but for our purposes of trying to observe enhancements in the intensity of Raman spectra, it was less important.
3.7 SERS Using Colloidal AgNPs
Samples were prepared initially by pipetting AgNPs and analyte molecules directly onto slides, but later were pre-mixed before being pipetting onto slides which gave better results. To analyse the samples, the slides were inserted into the stage of the Raman microscope and the lens was manually focused onto the sample, usually using the 50x objective lens. The laser powers, acquisition times and range of wavenumbers over which the sample were scanned were selected using the WiRe software. Raman acquisitions were taken by performing 10 second extended acquisitions between 0 cm-1 and 3200 cm-1.
Glass slides had a vinyl sticker attached that had a circular chamber for the liquid sample. The channel could hold 20 µl of liquid and a glass cover slip was fixed on top of the sample. The glass slides were masking the Raman signals of our samples by producing spectral noise so quartz slides and cover slips were later used instead because quartz produces less fluorescence than boro-silicate glass.
SERS was investigated using two different molecules, PS (polystyrene) and R6G. PS is known to give a relatively strong Raman signal, with characteristic peaks at 620.9 cm-1, 795.8 cm-1, 1001.4 cm-1 and 1031.8 cm-1, 1155.3 cm-1,1450.5 cm-1,1583.1 cm-1,1602.3 cm-1,2852.4 cm-1, 2904.5 cm-1 and 3054.3 cm-1. We used 10 µm PS microspheres produced by Polysciences.
Fig. : Raman spectrum of PS from the University of Alberta, Canada
R6G gives a weak Raman signal, but has been shown in the literature to give a strong SERS signal. It has characteristic peaks at 613 cm-1, 774 cm-1, 1186 cm-1, 1311 cm-1, 1363 cm-1 and 1508 cm-1 1575 cm-1 and 1651 cm-1. R6G samples were produced by adding R6G crystals to de-ionised water and diluting to the required concentration.
Fig. : Characteristic Raman signal of R6G from a reduced silver substrate (red)
and silver nanoparticles ( blue)6
4. Results & Discussion
4.1 Synthesised AgNPs
Our initial experiments focused on investigating SERS using AgNPs that we had synthesised, as discussed in section 3.1. Here we will discuss our findings of our experiments involving these synthesised AgNPs.
4.1.1 SEM Images
Two samples of synthesised AgNPs were prepared as discussed in section 3.2. We see here SEM images of our two samples showing particle sizes, shapes and distributions. Sample 2 was heated for longer during its synthesis than sample 1 and we expected to see larger particles in this sample.
Fig. : Sample 1 – a) Distribution of AgNPs in the sample
b) Round AgNPs with diameters ranging from ?125-225nm
c) Diamond shaped AgNPs had widths and heights ranging from ?650nm to ?850nm
d) Gold coated sample showing AgNPs ranging in size from 10s to 100s of nms
Fig. : Sample 2 – a) Distribution of AgnPS in the sample with sizes ranging from 10s to 100s of nms
b) Aggregated AgNPs with sizes up to 2 µm
c) AgNPs ranging in size from ?140-240 nm
d) Large silver aggregates, up to ?45 µm
We can see from figs. 1 and 2 that there was a range of particle shapes with diamond shaped, rounded and irregular shaped AgNPs, there was a wide range of sizes in both samples ranging from the 10s of nanometres, up to larger silver aggregates that were in the range of a few microns in size, and there didn’t seem to be an even distribution of AgNPs throughout the sample, but since the samples had been deposited and dried on glass slides, it was difficult to determine from these images how the particles would be distributed in a colloidal solution. It was predicted that the second sample would have a larger average particle size than the first sample, due to the fact that it was heated for longer. The darker colour of the second colloidal solution compared to the first solution indicated that it should have had a larger average particle size, but it was hard to determine this from looking at the SEM images of these two samples as there was a wide range of particle sizes in both samples, although in Figs. 2b and 2d we can see larger aggregates of silver which would be expected in a solution that was heated for longer. The average particle size of the two samples could have been further determined using UV-Visible spectrometry.
4.1.2 Raman Results of R6G and PS with Synthesised AgNPs
Initial experiments attempting to observe SERS with R6G and our synthesised AgNPs with varying concentrations were unsuccessful. Centrifuging and filtering our synthesised AgNPs did not produce any better results. The poor results could be attributed to the non-uniformity of the particle shapes and sizes in our samples as well as low particle concentrations in the samples. We decided then to investigate SERS using PS microspheres which gives a strong, characteristic Raman signal as below in fig. 11. Initial experiments using our synthesised AgNPs showed very little, unrepeatable enhancement of the Raman signal of PS microspheres.
Fig. 11: Raman spectra comparing PS microspheres with and without the addition of equal parts of our synthesised AgNPS, using a 50 mW, 785nm laser on 10% power. Each spectrum was averaged over 3 acquisitions.
In Fig. 11, we can see SERS peaks at 620 cm-1, 796.4 cm-1, 1000.9 cm-1, 1031.3 cm-1, 1451 cm-1 and 1602.83 cm-1, with the Raman peaks for the PS microspheres by themselves having roughly the same peaks. These peaks are in line with figures shown in Fig. x and can thus be attributed to PS. Any minor difference in wavenumbers can potentially be attributed to peak shifting due to SERS or calibration errors. There is a small increase in Raman signal of PS of ?1.5 times although this is not a particularly strong enhancement and is not a repeatable result. The low enhancement can be attributed to the fact that 10 µm PS microspheres are quite large in relation to the size of the AgNPs, which would decrease the liklieness of SERS occurring, as well as the fact that our AgNPs were not ideal SERS substrates.
4.2 NM300 AgNPs
Due to synthesised AgNPs giving little SERS, commercially produced NM300 AgNPs (as discussed in section 3.4) were used instead, as they were expected to be more uniform in diameter and more concentrated.
4.2.1 Raman Results of PS with NM300 AgNPs
We compared the Raman spectrum of PS microspheres compared to a sample of PS microspheres mixed with a 1% dilution of NM300 AgNPs. The ratio of 1 % AgNP solution to microsphere suspension was 1:4.
Fig. 12: Raman spectra comparing PS microspheres with and without the addition of NM300 AgNPS, using a 50 mW, 514 nm laser on 100 % power. Each spectrum was averaged over 3 acquisitions.
There was a higher signal overall when the AgNPs were present although individual peaks were not enhanced by much and in some cases were less intense after the addition of AgNPs. Some peaks were missing and there were some new peaks after the addition of AgNPs which would be expected. Characteristic peaks were observed in both the sample of PS as well as the sample of PS and AgNPs at 1002 cm-1, 1033 cm-1, 1604 cm-1, 2858 cm-1, 2910 cm-1 and 3054 cm-1. An additional peak is observed in the sample of PS and AgNPs at 1400 cm-1 and this can be attributed either to peak shifting due to SERS or to the Raman signal of the carrier liquid of the AgNPS or PS microspheres.
A 10% dilution of NM300 AgNPs in purified water was filtered with a 0.2µm filter in an attempt to remove larger aggregated Ag particles. These were then used in SERS of PS, with a 1:1 ratio of NM300 and PS suspension. A 50 mW, 514 nm laser was used on 100% power and spectra were averaged of 3 acquisitions. We did not see evidence of SERS when using these size filtered AgNPs with polystyrene microspheres, perhaps as the concentrations were too low.
Centrifuged NM300 AgNPs also did not show evidence of SERS in early experiments with Polystyrene Microspheres and R6G using 514nm and 785nm lasers using varying concentrations. Poor results could be attributed to dilution of samples weakening the SERS effect, not premixing the samples and the fact that the NM300 sample was old and the AgNPs had aggregated a lot.
4.2.2 Raman Results of R6G with NM300 AgNPs
After observing some weak evidence of SERS of PS, using NM300 AgNPs, we then investigated SERS of R6G using NM300 AgNPs. Initial experiments did not show evidence of SERS, although when using a newer sample of NM300 and mxing the AgNPs with the R6G solution before pipetting onto the slides, strong SERS signals were observed.
Fig. : 13: Raman spectra of 3 samples, using a 50 mW, 785nm laser at 10% power, each with 5µl of undiluted AgNPs, and with varying concentrations of R6G
We see one clearly defined peak at 1040 cm-1 in the 0.4 mM R6G sample although this is not a characteristic peak of R6G. Other less defined peaks in all 3 samples are not characteristic of R6G. These Raman signals can possibly be attributed to carrier liquids in the sample for AgNPs and PS microspheres (which were used to focus the lens correctly on the sample).
Three samples were prepared using increasing ratios of R6G to NM300 AgNPs, so as the concentration of R6G was increased, the amount of AgNPs in the sample decreased. The ratios were 1:1, 1:2, 1:3 of undiluted AgNPs to 1 µm R6G.
Fig. : 13: Raman spectra of 3 samples of varying concentrations of R6G with NM300 AgNPs, using a 50 mW, 785nm laser at 10% power, compared with the spectra for R6G by itself
Characteristic peaks were not observed for R6G, and the peaks can probably be attributed to carrier liquids in the sample. The ratio of the AgNPs: R6G that provided the highest signal output was a 1:2 ratio.
This sample of NM300 AgNPs was not a good SERS substrate as it it was not much evidence of SERS. The sample was produced in 2011 and the nanoparticles in the sample had aggregated to a high degree over that time and thus was not producing strong results.
A newer sample of NM300 AgNPs was acquired and used to demonstrate SERS near the end of our project as seen in fig Z.
Fig. : 13: Raman spectra of 3 samples of varying ratios of undiluted NM300 AgNPs1 to 1mM R6G with, using a 50 mW, 785nm laser at 10% power, compared with the spectra for R6G by itself
Here we can characteristic peaks of R6G at 610 cm-1, 773 cm-1, 1183 cm-1, 1363 cm-1 and 1507 cm-1. This is clear evidence of SERS, and we see the strongest enhancement in the sample with a 1:1 ratio of AgNPs to R6g. We can attribute the stronger SERS effect in these samples to the fact that the NM300 AgNPs used in this experiment were produced more recently and were thus less aggregated than the previously used batch of NM300. We also mixed the AgNPs with the R6G solution before depositing onto the slides, so this may also have improved adsorption of the R6G molecules to the AgNPs.
4.3 Premade Silver Substrates
As discussed in section 3.5, we used premade silver substrates that had a 1mM R6G solution deposited onto their surfaces to investigate their SERS capabilities to compare to colloidal AgNPs.
Figs. A and B were both taken from the same substrates using different laser powers. There was a general trend of stronger Raman signals being produced by substrates that had been written with stronger laser powers, presumably because there was a higher density of silver in those substrates made with higher laser powers. Characteristic R6G peaks can be seen in both samples at 610 cm-1, 771 cm-1, 1190 cm-1, 1358 cm-1, 1505 cm-1, 1600 cm-1 and 1646 cm-1, much like the peaks seen in our SERS with NM300 AgNPs, although visible at lower laser powers.
We have achieved our aim of performing SERS on R6G and have observed an enhancement in the Raman spectrum of R6G in the range of 103 times, using commercially produced NM300 AgNPs. We have synthesised and characterised colloidal AgNPs as having a wide range of sizes and shapes and have shown that they were unsuitable as SERS substrates, due to their non uniformity and low concentrations. Pre-made silver substrates showed enhancements in the range of 104 times using lower laser powers than in experiments using colloidal AgNPs. The Substrates produced stronger SERS than the AgNPs because they had a higher concentration of silver particles per unit area, and also due to the fact that the colloidal AgNPs did not have optimal particle sizes for use with a 785nm laser.
6. Future Work
6.1 Short Term
SERS has been demonstrated using NM300 AgNPs, however we would like to see SERS being demonstrated using our synthesised AgNPs. Synthesised AgNPs could be optimised for SERS by producing samples of AgNPs with known particle sizes and with narrow size distributions. This could be achieved by repeated size filtering and centrifugation of samples and by characterisation of the particles using UV-Visible spectroscopy and SEM. To produce better AgNPs, very clean glassware and ultrapure water could be used and reaction conditions could be carefully controlled. Different sized AgNPs could be compared to find the optimal size of AgNP for use as a SERS substrate using a 785nm laser. To be able to get high enough enhancements that acquisition times and laser powers could be significantly reduced would be a worthwhile end goal. Alternative methods of producing AgNPs that produce uniform samples would be worth investigating as well.
6.2 Long Term
We would like to see SERS using colloidal AgNPs incorporated into microfluidic devices that could be used to identify pathogens rapidly. Optimised AgNPs could provide high levels of surface enhancement, allowing low laser powers to be used as well as short acquisition times. Low laser powers could reduce the likelihood of biological samples being damaged by heat produced by powerful lasers and short acquisition times would make this system attractive for use in a biomedical point-of-care setting, if it could identify pathogens a lot faster than traditional methods of producing cultures which can take hours.
1 Vandenabeele, P. Raman spectroscopy. Analytical and Bioanalytical Chemistry 397, 2629-2630, doi:10.1007/s00216-010-3872-8 (2010).
2 Haynes, C. L., McFarland, A. D. & Duyne, R. P. V. Surface-Enhanced Raman Spectroscopy. Analytical Chemistry 77, 338 A-346 A, doi:10.1021/ac053456d (2005).
3 Vandenabeele, P. Practical Raman spectroscopy : an introduction. (Wiley,, 2013).
4 Ratyakshi. Vol. 21 (ed R.P. Chauhan) S113-116 (Asian J. Chem. , 2009).
5 Meyer, M. W. & Smith, E. A. Optimization of silver nanoparticles for surface enhanced Raman spectroscopy of structurally diverse analytes using visible and near-infrared excitation. Analyst 136, 3542-3549, doi:10.1039/C0AN00851F (2011).
6 Rahomäki, J., Nuutinen, T., Karvonen, L., Honkanen, S. & Vahimaa, P. Horizontal slot waveguide channel for enhanced Raman scattering. Optics Express 21, 9060-9068, doi:10.1364/OE.21.009060 (2013).
7 Stamplecoskie, K. G., Scaiano, J. C., Tiwari, V. S. & Anis, H. Optimal Size of Silver Nanoparticles for Surface-Enhanced Raman Spectroscopy. The Journal of Physical Chemistry C 115, 1403-1409, doi:10.1021/jp106666t (2011).
The peak plasmon absorption for AgNPs is centred at ?400nm although this varies with particle size. The optimal size of silver nanoparticles for surface-enhanced Raman spectroscopy has ben found to be ?50-60nm for adsorbed R6G molecules for use with a 785nm laser and it is thought that this result would extend to other adsorbates as well7.
Luke Joseph Mulvey
...(download the rest of the essay above)