WO2021123801A1

WO2021123801A1 - Dual wavelength nanoparticle detection

Info

Publication number: WO2021123801A1
Application number: PCT/GB2020/053279
Authority: WO
Inventors: Stefan Leo VAN WORKUM; Vladimir Alexander Turek; Marko Dorrestijn; David R. Klug; Callum Robertson SMITH
Original assignee: Vidya Holdings Ltd
Priority date: 2019-12-19
Filing date: 2020-12-18
Publication date: 2021-06-24
Also published as: GB201918894D0

Abstract

The present invention provides a method for detecting a plasmonic nanoparticle label. The method for detecting a plasmonic nanoparticle label comprising: illuminating a sample with first and second beams of electromagnetic radiation having first and second frequencies, respectively; wherein the sample is located in a first optical medium on a surface defining a boundary between the first optical medium and a second optical medium having a higher refractive index; measuring the intensity of the radiation at the first and second frequencies after the first and second beams have traversed the sample; detecting the presence of a plasmonic nanoparticle label in the circumstances that the intensity differential between the first beam and the first traversed beam is greater than the intensity differential between the second beam and the second traversed beam; wherein the first frequency is a known plasmon resonance frequency of the nanoparticle, and wherein the second frequency is an off-resonance frequency of the nanoparticle; and wherein the step of illuminating the sample comprises directing the first and second beams at the sample through the second optical medium at an angle of incidence to the surface which is above the critical angle for total internal reflection.

Description

DUAL WAVELENGTH NANOPARTICLE DETECTION

Technical Field

The present invention relates to improvements in or relating to methods for detecting nanoparticles and in particular to detecting the presence of plasmonic nanoparticle labels in immunoassays. Background

The employment of scattering and absorption from labels such as nanoparticles in immunoassays can lead to very high signals. Extinction coefficients as high as 10^{10 11} M ¹cm ¹ are common. This is five to six orders of magnitude more light than could be generated in an assay using state of the art molecular dyes, which are more commonly used as labels in assays, and could enable faster, more sensitive detection biomarkers in various formats, in principle enabling earlier stage diagnoses. The challenge that arises from using nanoparticle labels in place of fluorophores comes from the fact that scattering and absorption usually involve the incident and scattered or transmitted light having the same wavelength, whereas fluorescence detection uses different wavelengths for excitation and detection purposes. This is problematic because, aside from the desired particle labels, almost any sufficiently large object will have some degree of scattering response if the incident and detected light is of the same wavelength. Scattering by imperfections in optical components, imperfect beam termination, samples containing scattering objects and the use of a porous matrix to retain the sample and so on can also lead to high levels of background light which is hard to distinguish. This means that, although the signals from immunoassays using nanoparticle labels can be substantially higher than when fluorophores are used, the background is also much higher, limiting the overall sensitivity of detection. Even relatively small scattering backgrounds can be variable due to variation in sample scattering properties and therefore accuracy and precision of the assay can be affected as well as the overall sensitivity. Avoiding, reducing or even simply identifying what is a ‘background’ signal is therefore key to improving detection capabilities of assays with nanoparticle labels. Whilst a reduction in scattering can be achieved through an efficient design of the optical path of the assay, filtering the sample and other steps and actions it is not possible to eliminate background signals completely.

Immunoassays based on light typically work by illuminating a sample and receiving light back, the intensity of which determines how much of a reporter is bound on a surface. The most common type of labelling is using fluorophores, molecules which are excited by and which then re-emit light, usually at a longer wavelength than the incident light. By using a long-pass filter, it is possible to cut out most of the light that is returned from the assay that does not originate from fluorophores, since most of this is due to elastic scattering, which maintains the wavelength of the incident light.

KR 10 0760,315 B1 relates to method and apparatus for detecting single molecules by using multi-wavelength Total Internal Reflection Fluorescence (TIRF) techniques to detect light produced from two or more single molecules in real colour and real time by inducing TIRF with different wavelength lights produced one or more than two light sources.

US 7,420,675 B2 relates to a multi-energy system that generates and/or forms images of targets/structures by applying Mueller matrix imaging principles and/or Stokes polarimetric parameter imaging principles to data obtained by the multi- energy system.

The drawbacks to using fluorescent molecular labels include relatively weak interactions with light meaning difficult to detect signals, and photobleaching. The latter means that there is a limit to how much light can be used before the labels degrade. These drawbacks can be alleviated by using non-fluorescent, scattering labels such as nanoparticles. For example, an 80nm particle can return 10⁶ times more light per second than a single fluorophore, and photobleaching is not observed which means that detection can be continuous under higher laser powers. The limitation to this method, however, is that elastic scattering from surfaces at the sample can be even larger than the signals from the nanoparticles. Even if the optical configuration of the device is optimised, these background signals can still be high enough to limit the sensitivity of detection. For this reason, immunoassays using nanoparticles tend to either further amplify the signals through, for example, reducing metals onto the adsorbed labels, or to rely on colour differences between the adsorbed particles and the surrounding medium in lateral flow assays.

In de-skilled environments such as consumer tests, reduction of metals onto the particles is usually unfeasible, while the colour change requires sophisticated optical detection configurations that are expensive. Therefore, currently the mode of detection of choice for such assays tends to be a colour change visible to the naked eye. This is the case with many commercial lateral flow assays, used for example in pregnancy tests. In these there is a scattering background from the porous wicking material used to carry the assay components. This material approximates to a white scatterer that is unbiased with wavelength, whilst the nanoparticles used in such assays are coloured due to a combination of wavelength dependent absorption of light and scattering. However the results of such assays are not quantitative, but are instead limited to a binary decision based on a subjective and imprecise judgment from the user. As such these types of consumer assay are sufficient, for example, for pregnancy tests, but for assays where a quantitative or non-binary result is obtained, the user cannot be expected to reliably interpret this information by eye. In some circumstances a relatively small change in the intensities can be important and the human eye in an uncontrolled environment cannot judge with sufficient accuracy to make a meaningful interpretation of the result presented.

In order to be able to truly increase the limits of detection, accuracy and precision for commercially available assays, while maintaining economical components, it is clear that methods based on nanoparticle scattering need to be able to efficiently disentangle sources of background scattering from the scattering or absorption from the labels. It is against this background that the present invention has arisen.

Provided herein is a method based on a differential scattering cross-section at different wavelengths between nanoparticles and potential sources of background. Summary of Invention

According to an aspect of the present invention, there is provided a method for detecting a plasmonic nanoparticle label comprising: illuminating a sample with first and second beams of electromagnetic radiation having first and second frequencies, respectively; wherein the sample is located in a first optical medium on a surface defining a boundary between the first optical medium and a second optical medium having a higher refractive index; measuring the intensity of the radiation at the first and second frequencies after the first and second beams have traversed the sample; detecting the presence of a plasmonic nanoparticle label in the circumstances that the intensity differential between the first beam and the first traversed beam is greater than the intensity differential between the second beam and the second traversed beam; wherein the first frequency is a known plasmon resonance frequency of the nanoparticle, and wherein the second frequency is an off-resonance frequency of the nanoparticle; and wherein the step of illuminating the sample comprises directing the first and second beams at the sample through the second optical medium at an angle of incidence to the surface which is above the critical angle for total internal reflection.

According to another aspect of the present invention, there is provided method for detecting a plasmonic nanoparticle label, comprising illuminating a sample with first and second beams of electromagnetic radiation having first and second frequencies, respectively. The method further comprises measuring the intensity of the radiation at the first and second frequencies after the first and second beams have traversed the sample, and detecting the presence of a plasmonic nanoparticle label in the circumstances that the intensity differential between the first beam and the first traversed beam is greater than the intensity differential between the second beam and the second traversed beam, wherein the first frequency is a known plasmon resonance frequency of the nanoparticle, and wherein the second frequency is an off-resonance frequency of the nanoparticle.

The invention also recognises and takes advantage of the ongoing reduction in cost of light sources, be they light emitting diodes, laser diodes or other light sources, means that adding a second frequency to the readout of the assay does not produce a significant increase in cost.

This method takes advantage of the differential scattering cross-section between nanoparticles and potential sources of background at different wavelengths that arises due to plasmonic resonances, and the resultant effect on reflected or transmitted intensity in the radiation. Utilising the distinct plasmon resonance peaks of nanoparticle labels to distinguish them from the spectral profile of the background radiation overcomes the problem of the incident radiation and the radiation which is scattered or absorbed by the nanoparticles being of the same frequency, allowing for effective identification and isolation of background noise. The employment of nanoparticle labels in, for example, immunoassays can lead to very high signals and thus enable faster, more sensitive detection of biomarkers in various formats, in principle enabling earlier stage diagnoses. The term “scatter” as used herein refers to radiation, and in particular light, which is scattered in the conventional sense, such as by Rayleigh or Raman scattering effects, as well as to reflected radiation in general.

In some embodiments, detecting the presence of the plasmonic nanoparticle label further requires the difference between the intensity differential for the first beam and the intensity differential for the second beam to be above a predefined threshold. This is because small differences in the measured intensities could result from effects other than the plasmon resonances of the nanoparticle labels. Setting a predefined threshold difference for the differentials of the resonant and non-resonant frequency beams identifies radiation which has traversed the nanoparticle labels and background radiation with greater certainty.

In some embodiments, the second frequency is higher than the first frequency. The background radiation will generally trend upwards to higher scattering cross sections at the shorter wavelengths, so if an off-resonant frequency is used which is higher than the resonant frequency, a nanoparticle resonance peak will break that trend by having a higher scattering cross section despite having a lower frequency. This general trend changes to scattering between resonant and off-resonance wavelengths is a more reliable metric than, for example, attempting to compare the absolute scale of a decrease or increase.

Furthermore, if the particles on a surface being measured are in close proximity to each other: for example through aggregation or due to high surface density, new resonances may appear in the red-shifted wavelengths which may become stronger that the initially resonant wavelength. This may mean that the directionality of signal change between the two wavelengths would change. Conversely, new blue shifted peaks from the resonance of single spherical particles will not occur under any experimentally feasible conditions. Accordingly, having an off resonance frequency higher than the resonant frequency decreases the risk of a false negative determination where an algorithm identifies a plasmon resonance as background radiation because a peak from the background radiation has a higher intensity due to a non-plasmonic resonance.

In some embodiments, the sample is located in a first optical medium on a surface defining a boundary between the first optical medium and a second optical medium having a higher refractive index; and the step of illuminating the sample comprises directing the first and second beams at the sample through the second optical medium at an angle of incidence to the surface which is above the critical angle for total internal reflection. Such a configuration will ensure total internal reflection of the illuminating beams and an evanescent field from the two beam wavelengths is then scattered in all directions by the sample particles, enabling dark field detection modes. The losses of the total intensity of the reflected beams can also be measured using bright field techniques. As the sample is only illuminated by the evanescent field on the surface, only a small fraction of the solution interacts with light, thereby reducing the background signals.

In other embodiments, the sample is located in a first optical medium on a surface defining a boundary between the first optical medium and a second optical medium having a different refractive index; and the step of illuminating the sample comprises directing the first and second beams at the sample through the first optical medium. In this more general reflection configuration, the substrate on which the assay is carried out is less constrained by its optical properties and could in principle be any surface, as opposed to the total internal reflection configuration which puts certain constraints on the substrate used, as it needs to interact with radiation which is incident from angles above the critical angle appropriately.

In some embodiments, the step of measuring comprises measuring transmitted illumination to determine the amount of illumination that has been absorbed after traversing the sample.

In some embodiments, the step of measuring comprises measuring scattered illumination to determine the amount of illumination that has been scattered after traversing the sample.

In some embodiments, the step of measuring is performed simultaneously by two separate detectors configured for dark field and bright field measurement, respectively. Simultaneous measurement using different techniques enhances the dynamic range of the assay.

In some embodiments, the first and second beams are configured to alternate and illuminate the sample sequentially. Performing the illumination sequentially simplifies the required optics such that only a single detector is required for measurement. This is preferred where either space or cost is at a premium.

In some embodiments, the sample may be illuminated, preferably using a single light source, to provide two different spectra where the ratio of pixel values between the two images can be analysed. Illuminating a sample containing nanoparticles with at least two different spectra allows for the removal of undesired artefacts, including but not limited to dust, contaminants and/or scratches. Thus, this provides a processed image with a higher signal to background ratio for nanoparticles, and thus clearer identification of nanoparticles. Hence, this approach can be used to distinguish between nanoparticle scattering and scattering from other scattering sites. In some embodiments, the first and second beams are derived from two separate light sources. These may be substantially monochromatic light sources of either matched power density or of predetermined different power density. Alternatively, the first and second beams are derived from a single, broad spectrum illumination source. A single, broad spectrum light source results in a reduction in complexity of apparatus. It also ensures that the light is incident from the same direction across first and second wavelengths. Furthermore, the power density profile of a single source will be known and therefore matching of power densities will be less complex than embodiments using two separate sources.

The invention will now be further and more particularly described, by way of example only, and with reference to the accompanying drawings, in which:

Fiqures

Figure 1 shows an example absorbance spectrum of spherical gold nanoparticles;

Figure 2 shows scattering from a 100 nm silica particle; Figure 3 shows a schematic of an example background reduction mechanism according to the present invention;

Figure 4 shows an example illumination configuration according to the present invention using total internal reflectance geometry;

Figure 5 shows an example illumination configuration according to the present invention using reflection geometry;

Figure 6 shows a further example illumination configuration according to the present invention using total internal reflection geometry with a shutter;

Figure 7 shows a further example illumination configuration according to the present invention using total internal reflection geometry with a filter;. Figure 8 shows an example illumination configuration according to the present invention using reflection geometry with a shutter;

Figure 9 shows an example illumination configuration according to the present invention using reflection geometry with a filter;

Figure 10A shows a spectrum using a single light source with two filters; Figure 10B provides an example of a scattering cross section for 70 nm gold nanoparticles;

Figures 11A provides an image of a green spectrum illumination and Figure 11 B provides an image of a blue spectrum illumination;

Figure 12 shows a histogram plot; Figures 13A shows show an original and Figure 13B shows a processed image; and Figure 14 shows an image of a binary mask with pixels above and below a 2.6 ratio.

Detailed Description

In order to further explain various aspects of the present disclosure, specific embodiments of the present disclosure will now be described by way of example only in conjunction with the accompanying drawings. The present invention makes use of the phenomenon of plasmon resonance. The electrons in metals can interact with the electric field of light, thereby causing oscillations. In nanoparticles such as gold and silver especially these oscillations, plasmons, can meet resonance criteria at certain wavelengths, depending on the composition, shape and size of the nanoparticle as well as its surrounding environment.

Referring to FIG. 1, an example absorbance spectrum of spherical gold nanoparticles is shown. Typically, for spherical gold particles in water, resonance criteria are met in the 520-550 nm wavelength region, which is illustrated in FIG.1 as a peak in the UV-visible absorbance spectrum. The spectrum in FIG. 1 is characteristic of spherical nanoparticles in general, while background scattering will exhibit an entirely different spectral response. Referring to FIG. 2, an example scattering spectrum of a 100 nm glass silica particle is shown, the scattering properties of which would be comparable to dust of a similar size. In essence, non-nanoparticle scattering centres generally approximate to white light scatterers, with a bias towards the blue end of the spectrum. Flowever, nanoparticles have a resonance, which manifests itself as a peak, often in the ‘visible’ or near IR region of the spectrum. In the example shown FIG. 1 this corresponds to the ‘resonant wavelength’ region of the spectrum. This region has a higher intensity than both the blue and red-shifted wavelengths, as opposed to the scattering of the 100 nm silica particle which has decreasing scattering cross-sections at higher wavelengths. The majority of background scatterers will show either similar responses to that shown in FIG. 2 or will at least not display any sharp features.

By illuminating a sample with two beams, a first beam at a first frequency which is a resonant frequency of a plasmonic nanoparticle label and a second beam at a second frequency that is not a resonant frequency of the nanoparticle label, the sharp features on the nanoparticle spectral profile create a differential in intensity that is present under illumination from the first beam and not present under illumination from the second beam. This allows intensity measurements at the two frequencies to distinguish between a nanoparticle label with a spectral profile similar to that in FIG. 1 and background radiation arising from scatterers having spectral profiles similar to the glass silica particle of FIG. 2.

In some embodiments, the non-resonant frequency is chosen to be at a higher “blue- shifted” frequency with respect to the resonant frequency beam. While the spectrum of the nanoparticle in FIG. 1 suggests that there is a greater distinction between the resonant and the red-shifted wavelengths, this is not the ideal comparison to use to identify background.

This is partly because the background scattering profile shown in FIG. 2 displays a decreased intensity in the red as compared to the resonant wavelength, whereas the comparison between the blue shifted and the resonant wavelengths would have the opposite behaviour than the nanoparticle. The general trend of increased or decreased scattering between resonant and off-resonance wavelengths is a more reliable metric than attempting to compare the absolute scale of a decrease or increase.

Furthermore, FIG. 1 shows a very low red-shifted scattering, which is not always the case for plasmonic nanoparticles. If the sample particles on the surface are in close proximity to each other, for example through aggregation or due to high surface density, new resonance peaks may appear in the red-shifted wavelengths which can become stronger than the plasmonic resonance peak. In this case the directionality of the signal change between the two wavelengths would reverse. On the other hand, new blue shifted peaks from the resonance of single spherical particles may not occur under any experimentally feasible conditions, ensuring a similar trend irrespective of the progress of the assays.

While mapping a full spectrum of each scatterer in the assay could ensure differentiation between background signals and the nanoparticles, measuring the intensity of scattering from a resonant and non-resonant wavelength beam is sufficient and saves time and cost.

Referring to FIG.3, an example schematic of the above-described background identification mechanism is illustrated. If the intensity close to the frequency of the peak 1 shown in FIG.1 , and another at a shorter wavelength 2 is measured, background signals 5 from, for example, dust will either increase or remain similar at shorter wavelengths, while signals coming from nanoparticles 3 on the substrate 4 should decrease.

There are multiple optical configurations in which these effects could be observed. The two geometries described herein are total internal reflection and reflection, outlined in FIGs. 4 and 5. Flowever it should be recognised that other configurations not described herein may also be suitable for achieving the same effect.

Referring to FIG. 4, there is illustrated an example of a total internal reflection configuration with two different light sources 6 and 7 refracted by a prism 8 to illuminate the nanoparticles 9 bound to the surface 16 in the assay. The assay is not required to be directly on a prism 8 but can also be performed on a separate chip by refractive index matching liquids to ensure that angle of incidence of the light between the medium its travelling in and assay’s medium is above the critical angle. The evanescent field from the two wavelengths is scattered in all directions by the particles, enabling dark field detection modes. The two main methods of detection of this would be to detect the signals 10 directly above the nanoparticles and, if the prism has an optically flat and transparent bottom, the signals 11 directly below. As a further addition, the losses of the total intensity of the reflected beams could be measured using bright field techniques to check the intensity reduction in the transmitted radiation 12 and 13 having traversed the sample.

Referring to FIG. 5, an example reflection configuration is illustrated, the two light sources 14 and 15 directed at an angle with respect to the surface 16 on which the nanoparticles are bound. As with total internal reflection, this would induce scattering of the illuminating radiation which can be detected in dark field from signals 10 above and, if the surface 16 is transparent, signals 11 below the assay. Furthermore, the transmitted radiation 17 and 18 can similarly be measured in bright field. The substrate on which the assay is carried out is less crucial in the reflection mode, as fewer optical constraints are necessary, and could in principle be any surface 16.

The main difference between the reflection and total internal reflection geometries is that in total internal reflection the sample is only illuminated by the evanescent field on the surface. This confinement means that only a small fraction of the solution interacts with light, thereby reducing the background signals. The downside of this is the constraints to the optical elements, namely a prism or other surface that allows access to total internal reflection angles.

The angle of incidence and the position of the illuminating radiation, which can be for example a laser, with respect to the sample can be variable, as can the angle and position of the measuring detector. Such factors can also be varied in a controlled manner during an experiment to differentiate more clearly between background levels and the nanoparticles.

In general, there is a need to be able to differentiate between the resonant and non resonant wavelengths in detection, and for the two separate light sources to be of known power density or of substantially equal power density.

Referring to FIGs. 6 to 9, various possible embodiments of the reflection and total internal reflection geometries are illustrated using shutters and filters to improve the measurement of the illuminating radiation.

Referring to FIG. 6, an example total internal reflection configuration using shutters 21a and 21b is illustrated. A laser that is on resonance with the plasmonic scattering peak 19 and another laser that is 50-100 nm blue-shifted off-resonance 20 are directed to a common beam path through a partially reflective mirror or a beam splitter 22. Perfectly overlapping beam paths without using a beam splitter may also be sufficient; however, beam splitter 22 also ensures similar laser intensities are delivered to the sample.

Dual shutters 21a and 21b can control and alternate which of the two wavelengths, resonant and non-resonant, is illuminating the sample at a given time. This is particularly useful for single detector configurations, as only one frequency need be measured at a time.

Alternatively, a single shutter can be used in front of the off-resonance laser 21b. In such a configuration, either at regular intervals or several times during the assay, shutter 21b would be opened to increase the total light that the detector 24 picks up. If the laser power densities and the scattering cross section of the objects in the field of view of the detector are the same, or known and factorizable, then the total intensity of these objects should double. For non-plasmonic objects, this cross section will either increase or be the same and therefore the signal will double, however for the plasmonics labels this will be substantially below double, thereby allowing algorithmic exclusion of backgrounds.

For dark field measurement, the scattered light would be focused on a detector 24 by lens 23. This could be either above 23a and 24a or below 23b and 24b the assay. In bright field measurement, the reflected laser should be collimated enough that a lens would be unnecessary to focus the light onto the detector 24c.

Whilst only one detector would be required in any of the above described configurations, multiple detectors may also be used. Detection in both dark and bright field simultaneously may enhance the dynamic range of the assay.

Referring to FIG. 7, an example total internal reflection configuration using filters 25 is illustrated. The total internal reflection based detection with filters differs from the geometry with the shutter by both lasers being configured to always illuminate the assay simultaneously, and by the radiation having traversed the sample always being detected by two separate detectors.

In dark field detection, the detectors could be on opposite sides of the assay, with a long-pass filter 25a to exclude the shorter wavelength and a short-pass filter 26a to exclude the longer wavelength. These detectors can also be located on the same side of the assay. Flowever, an optic element such as a beam-splitter would be needed to split the light equally to both detectors. Likewise, the bright field configuration would also require two detectors with a beam-splitter 27, a long pass filter 25b and a short pass filter 26b.

Referring to FIG. 8, an example reflection geometry using shutters is illustrated. The configuration operates in a similar manner to the total internal reflection with shutter geometry described in relation to FIG. 6 in that the resonant laser is under constant operation, while the off-resonance laser is switched on whenever a differentiation between the nanoparticle labels and the background scattering is required. The difference being that total internal reflection of the illuminating radiation is not required.

As such, and as discussed above, the substrate 16 on which the assay is performed is not as constrained by its optical properties. However, in some embodiments, transparency would be required between the laser and the detector. The angle of incidence can be tuned to minimise direct reflections to the detectors for dark field mode measurements. The angle could range from grazing to normal incidence, though the two extremes would require additional appropriate modifications to the optics. For example, for normal incidence, dark field would need to locate the detectors away from the light path, while grazing incidence would likely require focusing of the beam in order to reach sufficient laser power densities on the sample. Referring to FIG. 9, an example reflection geometry using filters is illustrated. The configuration operates in a similar manner to the total internal reflection with filter geometry described in relation to FIG. 7, whereby shutters can be replaced with long and short pass filters and two detectors to enable continuous tracking of the assay with two wavelengths. Aside from the illustrated schematics, there exist multiple alternative methods in order to implement the invention. The main challenge of any optical configuration is to be able to differentiate between the resonant and non-resonant wavelength beams.

Alternative examples of how this could be achieved include: the light source or the shutters being pulsed, with a lock-in amplifier removing ambient light; the detector itself differentiating the wavelengths; alternating the lasers with, for example, rotating and/or switchable mirrors or a spinning fan blade, which could optionally comprise long and short pass filters; powering the lasers on and off individually during operation. Image processing functions

Further to the physical configurations described above for optimising nanoparticle label detection, provided below is a brief description of example processing methods for filtering background noise from radiation that has traversed the sample location. Although specific processing techniques are described, it should be recognised that other methods for processing the traversed radiation may also be used.

In an example method of processing, every location on the assay can be associated with four different intensities, assuming that intensities for two wavelengths are recorded and both bright field and dark field images are taken for each wavelength.

For every detector element, e.g. every pixel, or group of elements, a filter function / can be applied that is composed of one or more ratios of intensities /. For example, / may be a weighted sum of ratios:

where the enumerator of each ratio is selected to be larger than the denominator for nanoparticles, and subscripts indicate blue (b) and green (g) excitation.

Weights are chosen such that in distinguishing dust from nanoparticles, the sensitivity, e.g. true positive rate, and specificity, e.g. 1 minus the false positive rate, are relatively high, i.e. large area under the receiver operating characteristic curve. Alternatively, the filter function may contain more than two intensities per ratio. For example: r _ h,BF Ib,DF Ig,BF Ig,DF where factors are divided over the enumerator and denominator such that, as mentioned above, the area under the receiver operating characteristic curve is relatively high. An image of / may be a threshold such that dust is filtered out. For example, if for dust, 0.9 < / < 1.1, and for nanoparticles, / > 1.5, an optimal threshold may be close to f_t = 1.3; then, for detector elements with / < f_t, intensities may be set to zero to obtain the filtered image (for any of the four intensity images). If multiple detectors are used, e.g. for bright field and dark field, detector elements that correspond to the same location on the assay on different detectors should be combined in the ratios.

Referring to Figure 10, there is shown a spectrum with two filters that are used to select different portions of the LED spectrum. As shown in Figure 10A, the green spectrum is centred on 559 nm with FWHM 34 nm. As shown in Figure 10A, the blue spectrum is centred on 450 nm with FWHM 40 nm. Both are indicated on the LED spectrum. A single illumination source such as Thorlabs MINTL5 LED can be used.

This approach takes advantage of the spectral dependence of scattering. The nanoparticles exhibit a resonance frequency related to their composition and size. Nanoparticles scattering at this resonance experience a local maximum. An example of scattering cross-section for 70 nm gold nanoparticles at a water interface is shown in Figure 10B, demonstrating a resonance at approximately 556 nm.

Undesired artefacts including dust, contaminants and scratches will have a different spectral response from the nanoparticles of interest. Reviewing at a single wavelength, such artefacts can contribute to the signal and reduce the ability to identify nanoparticles of interest.

Illuminating a sample containing nanoparticles with at least two different spectra allows for the removal of undesired artefacts, including dust, contaminants, scratches and many more. The removal of undesired artefacts produces a processed image with a higher signal to background ratio for nanoparticles, and thus clearer identification of nanoparticles.

Referring to Figures 11A and 11 B, there is provided an image of a solution of nanoparticles illuminated with green and blue wavelengths respectively. The solution of nanoparticles was deposited onto a Borofloat surface. Images were obtained for the green and the blue spectral illumination. As shown in Figure 11 A, the nanoparticles are clearer in the green spectra due to proximity of the resonance.

The pixel values from the green image can be divided by the blue image, thus generating a ratio for each pixel. As a result, a histogram of these ratios can be plotted as shown in Figure 12. A threshold ratio can be selected whereby every pixel below the threshold is set to zero in the green illumination image, and everything above the threshold is unchanged in the green illumination image.

Referring to Figures 13A and 13B, there is provided an original image and a processed image, respectively. The technique as disclosed herein can be applied for a ratio of 2.6, resulting in clear removal of scattering sites such as artefacts i.e. background noise that is not associated with the main nanoparticle deposition.

In one example of the present invention, there is provided a binary mask showing the pixels above a 2.6 ratio (1) and those below 2.6 (0) as shown in Figure 14.

Various further aspects and embodiments of the present invention will be apparent to those skilled in the art in view of the present disclosure.

“and/or” where used herein is to be taken as specific disclosure of each of the two specified features or components with or without the other. For example “A and/or B” is to be taken as specific disclosure of each of (i) A, (ii) B and (iii) A and B, just as if each is set out individually herein.

Unless context dictates otherwise, the descriptions and definitions of the features set out above are not limited to any particular aspect or embodiment of the invention and apply equally to all aspects and embodiments which are described.

It will further be appreciated by those skilled in the art that although the invention has been described by way of example with reference to several embodiments, it is not limited to the disclosed embodiments and that alternative embodiments could be constructed without departing from the scope of the invention as defined in the appended claims.

Claims

1. A method for detecting a plasmonic nanoparticle label comprising: illuminating a sample with first and second beams of electromagnetic radiation having first and second frequencies, respectively; wherein the sample is located in a first optical medium on a surface defining a boundary between the first optical medium and a second optical medium having a higher refractive index; measuring the intensity of the radiation at the first and second frequencies after the first and second beams have traversed the sample; detecting the presence of a plasmonic nanoparticle label in the circumstances that the intensity differential between the first beam and the first traversed beam is greater than the intensity differential between the second beam and the second traversed beam; wherein the first frequency is a known plasmon resonance frequency of the nanoparticle, and wherein the second frequency is an off-resonance frequency of the nanoparticle; and wherein the step of illuminating the sample comprises directing the first and second beams at the sample through the second optical medium at an angle of incidence to the surface which is above the critical angle for total internal reflection.

2. The method of claim 1, wherein detecting the presence of the plasmonic nanoparticle label further requires the difference between the intensity differential for the first beam and the intensity differential for the second beam to be above a predefined threshold.

3. The method of any preceding claim, wherein the second frequency is higher than the first frequency.

4. The method of any of claims 1 to 3, wherein the sample is located in a first optical medium on a surface defining a boundary between the first optical medium and a second optical medium having a different refractive index; and wherein the step of illuminating the sample comprises directing the first and second beams at the sample through the first optical medium.

5. The method of any preceding claim, wherein the step of measuring comprises measuring transmitted illumination to determine the amount of illumination that has been absorbed after traversing the sample.

6. The method of any preceding claim, wherein the step of measuring comprises measuring scattered illumination to determine the amount of illumination that has been scattered after traversing the sample.

7. The method of any preceding claim, wherein the step of measuring is performed simultaneously by two separate detectors configured for dark field and bright field measurement, respectively.

8. The method of any preceding claim, wherein the first and second beams are configured to alternate and illuminate the sample sequentially.

9. The method of any preceding claim, wherein the method further comprises, after a predetermined time interval has elapsed since the sample was last measured, repeating the steps of claim 1 to obtain a series of at least two measurements of the sample.

10. The method of any one of claims 1 to 9, wherein the first and second beams are derived from two separate light sources.

11. The method of any one of claims 1 to 9, wherein the first and second beams are derived from a single, broad spectrum illumination source.