Disposable, stable media for reproducible surface-enhanced Raman spectroscopy

Abstract

Large numbers of identical and stable SE(R)RS [surface-enhanced (resonance) Raman]-active media, which are convenient to handle and manipulate but sufficiently inexpensive that they can be used once and then discarded, have been prepared by isolating nanoparticles from Ag and Au sols in hydrophilic polymer gels. The preparation simply involves mixing a suitable polymer with the sol to give a viscous suspension that can be coated onto a substrate and dried to form a hard translucent film. The films remain inactive until they are treated with aqueous analyte solution, which causes the film to swell and brings the analyte into contact with the active metal particles. The swollen films give strong SERS spectra which are effectively identical to those obtained from simple sols. The advantage of this method is that the dried polymers can be stored indefinitely before use and that they give a high degree of spectral reproducibility.

---

Introduction

Surface enhanced Raman and resonance Raman spectroscopy (SE(R)RS) can be an extremely sensitive spectroscopic technique [single molecule SER(R)S has been reported^1,2] and has good discrimination because it yields characteristic vibrational spectra of compounds. These characteristics mean that SE(R)RS has great potential as an analytical technique. However, it has not been widely exploited for routine analytical tasks (as opposed to one-off studies) because of problems with signal reproducibility. Even though a very broad range of methods, including electrochemical or chemical roughening of metal surfaces,^3,4 deposition of the metal onto substrate⁵ and preparation of colloidal suspensions of the metals (sols),^6,7 can be used to generate SE(R)RS-active materials, none of these is ideal for preparing media for routine analytical measurements. For these applications, ideally, the media should be identical, sufficiently inexpensive that they can be discarded after a single measurement (to prevent cross-contamination) and sufficiently stable that they can be stored ready for use for extended periods (months or years). The lowest cost approach is to use small aliquots of colloidal solutions but preparation of colloidal solutions with identical surface-enhancing properties is difficult.⁸ This problem is compounded by the fact that the sols are unstable and give different enhancement factors as they age, which means that even if they are prepared reproducibly they cannot be stored and used only as required.

The objective of the work described here was to find a suitable medium which would protect the nanoparticles of a conventional Ag or Au sol during storage but still allow them to act as enhancing media when required. The general strategy we have adopted is to use a hydrophilic swelling polymer, which is mixed with the sol to form a viscous solution that is then dried. In the dry polymer, it would be expected that aggregation of the nanoparticles would be prevented and that they would be protected from external contamination or chemical attack. However, introduction of an aqueous solution of analyte should force the polymer to swell and free the nanoparticles to interact with the analyte to produce the required surface enhancement (see Fig. 1). This approach differs from previous work^9–11 in which a porous or swelling medium is treated with aqueous silver salts and subsequent reduction generates the active particles in situ because here the sol can be prepared under controlled conditions before isolation of the particles.

Fig. 1 Preparation of activated 96 well plates and polymer film coated plates for SE(R)RS.

The advantage of this method is that, if successful, it should allow production of large numbers of stable and inexpensive SE(R)RS-active substrates which give reproducible signals because they are prepared by dividing a large volume of homogeneous stock solution into numerous identical sub-units.

Materials and methods

Silver sols were prepared by the standard citrate reduction method.⁷ The polymer–sol mixtures were prepared by adding an appropriate amount of the solid hydrophilic swelling polymer to the sol with vigorous stirring. The viscous mixture was then allowed to stand for a period of several minutes, to allow full hydration and swelling of the polymer, and was then stirred a second time before being applied to the chosen substrate. The data reported here are from a commercially available polycarbophil polymer (Noveon AA1, B.F. Goodrich Ltd.) but other hydrophilic polymers can also be used.¹² The viscosity of the polymer–sol suspension could be altered by adjusting the polymer concentration: 3.5 g of polymer per dm³ of sol gave solutions that were just thin enough to pipette. The deposited films were then dried under reduced pressure; during the drying process the polymer–sol suspension shrank to form a mechanically hard, uniform, translucent, durable material.

The flexible nature of the viscous suspension means that it can be used to create a range of physically different enhancing media designed for different purposes. Here we show data from just two different forms: glass slides (for spot tests) and polymer coated multi-well plates. However, further variations are possible; for example, coating the inside of glass capillaries gives tubes that can be used to draw up very small samples of analyte, while milling the dried films gives a powder that can be re-hydrated to give a SE(R)RS-active sol solution.

Results and discussion

The polymer used to support the nanoparticles must be hydrophilic, cause minimal perturbation of the nanoparticles’ microstructure and have a small Raman cross-section, so that it does not interfere with signals from the analyte. Furthermore, it must release the nanoparticles in their SERS-active form when required. We have found that polymers with carboxylic side chains best meet these criteria and, of these, the polycarbophil polymer we have used is the most successful, although a range of other hydrophilic polymers can also be used.¹² The polycarbophil polymers do not appear to alter the microstructure of the nanoparticles when added to sol solutions, presumably because the particles in the Ag sols have a surface layer of citrate ions which are replaced by chemically similar aryl carboxylic groups on polymer addition. The UV–visible absorption spectra of sols, which are strongly influenced by particle size/aggregation,² do not change significantly on polymer addition (Fig. 2) and transmission electron micrographs of the dried films show that even after drying the nanocrystals remain unaggregated. However, even in the presence of the gel-forming polymer, the nanoparticles can be aggregated, if required, by rehydrating the dried films with aqueous salt solution. This aggregation is easily observed through the characteristic broadening in the UV–visible absorption spectra of the films (see Fig. 2).

Fig. 2 UV/vis absorption spectra of (a) starting Ag colloid, (b) the viscous colloid–sol mixture made by adding polymer to the colloid shown in (a) and (c) a film that had been dried then rehydrated using 0.1 mol dm⁻³ aqueous NaCl.

Polymer films that have been dried onto a microscope slide cover slip form a very convenient matrix for SER(R)S spot testing. Fig. 3 shows spectra obtained by spotting a series of different analyte solutions (chosen to demonstrate the fact that the polymers give strong SER(R)S signals over a range of excitation wavelengths and with different sample types) onto such a film. Application of ca. 1 µl of sample causes the film immediately under the applied droplet to swell into the liquid drop. There is negligible creep of solution across the surface and the only indication of swelling is a marked increase in the viscosity of the applied droplet. The laser can be directed onto this viscous spot to obtain the SERS data. A single 18 × 18 mm slide can typically be used for ca. 16 of these spot tests. The reproducibility of these measurements is determined by the homogeneity of the dried polymer layer but we have found that spreading the viscous suspension over a linear array of slides placed between 200 µm high guide rails gives films that are cosmetically uniform over their entire area.

Fig. 3 SE(R)RS spectra obtained by spotting ca. 1 µl of <10⁻⁶ mol dm⁻³ aqueous solutions of (a) Ru(2,2’-bipyridyl)₃²⁺, (b) Crystal Violet and (c) tetra-4-N-methylpyridylporphyrin onto a polymer film coated glass slide. λ_ex = 457.9 nm (a, c), 785 nm (b).

For accurate quantitative work, the viscous polymer–sol suspension can be dispensed into standard 96 well plates (typically 100 µl per well) and allowed to dry. In the analysis sufficient analyte solution is added to swell the polymer–sol suspension back to its original volume. These swollen gels are not sufficiently viscous to stand unsupported, as they do in the low volume spot test, but they are contained within the well. Since wells are prepared by subdividing the original homogenous polymer–sol suspension the dried films within each are identical. The most important consequence of this is that the signal from each of the wells is remarkably consistent. Fig. 4 shows replicate spectra of a sample of the putative anti-cancer drug AQ4N,¹³ which were normalized by addition of an internal standard. Under these conditions, even crude quantitative analysis (ratioing the peak heights of the strongest band in the sample and standard) shows only a very small (3%) standard deviation between spectra obtained from each of the 20 different wells used in the experiment.

Fig. 4 Replicate SE(R)R spectra of 100 µl aliquots of aqueous AQ4N solution (2.75 × 10⁻⁶ mol dm⁻³) added to individual wells of a pre-treated 96 well plate along with a Ru(2,2’-bipyridyl)₃²⁺ internal standard. λ_ex = 633 nm.

The high reproducibility of spectra obtained from the rehydrated polymer films, coupled with the fact that large numbers of films can easily be produced at low cost (1 dm³ of stock polymer–sol solution provides enough material to prepare 100 multi-well plates, i.e., ca. 10⁴ individual tests), means that this approach can be used for routine high volume analysis.¹⁴ Moreover, since the films are mechanically durable and can be stored indefinitely before use, they should also allow analysts to make occasional use of SE(R)RS methods without the need to repeat time-consuming colloid preparation or standardization each time they need to make a measurement.

In summary, these polymer supported colloids have the potential to transform SE(R)RS into a simple and reproducible routine analytical technique.

Acknowledgement

The authors would like to thank Professor L. Patterson for the generous gift of a sample of the putative anti-cancer drug, AQ4N.

Notes and references

1. K. Kneipp, H. Kneipp, I. Itzkan, R. R. Dasari and M. S. Feld, Chem. Rev., 1999, 99, 2957.
2. A. M. Michaels, M. Nirmal and L. E. Brus, J. Am. Chem. Soc., 1999, 121, 9932.
3. D. D. Tuschel, J. E. Pemberton and J. E. Cook, Langmuir, 1986, 2, 380.
4. G. Sauer, U. Nickel and S. Schneider, J. Raman Spectrosc., 2000, 31, 359.
5. R. P. Vanduyne, J. C. Hulteen and D. A. Treichel, J. Chem. Phys., 1993, 99, 2101.
6. N. Shirtcliffe, U. Nickel and S. Schneider, J. Colloid Interface Sci., 1999, 211, 122.
7. P. C. Lee and D. Meisel, J. Phys. Chem., 1982, 86, 339.
8. Z. Q. Tian, Int. J. Vib. Spectrosc., 2000, 4, 2 www.ijvs.com.
9. M. Volkan, D. L. Stokes and T. Vo-Dinh, J. Raman Spectrosc., 1999, 30, 1057.
10. Y. Kurokawa, Y. Imai and Y. Tamai, Analyst, 1997, 122, 941.
11. A. B. R. Mayer and J. E. Mark, Polymer, 2000, 41, 1627.
12. We have also found that less soluble polymers which swell to a smaller extent, such as copolymers of hydroxyethylmethacrylate with methacrylic acid, can also be used as substrates. In this case the polymers are prepared as dry films or sheets and then immersed in the sol where the nanoparticles diffuse into the polymer. The film can then be dried and rehydrated in the same way as polycarbophil films..
13. M. M. Ali, M. C. R. Symons, F. A. Taiwo and L. Patterson, Chem.-Biol. Interact., 1999, 123, 1.
14. Since the production scale of each batch of substrates can be adjusted to provide extremely large numbers of individual substrates we have not attempted to optimize batch-to-batch reproducibility at this stage but we would expect that variation between batches should be sufficiently low (±10%) that it can be accounted for by simple linear scaling. Moreover, because the batch sizes can be made very large, such recalibration should be needed only rarely..

---

Footnote

† Present address: Avalon Instruments Ltd., 10 Malone Road, Belfast, UK BT9 5BN.

---