The present invention is directed to a technique for the detection of nanoparticles, such as viruses, and more particularly to an optical technique using interferometry which does not require knowledge of the dielectric properties of the nanoparticles.
Particle sizing is used in many areas of science and technology. The food industry, cosmetics, pharmaceuticals, paints and coatings, metals, ceramics, explosives, fireworks and semiconductor industries are just a few places that employ particle size measurements. For example, the reflectivity of road signs depend on the size of the glass beads embedded in the paint, the flavor of coffee depends on the size of the milled grains, proper size distribution of medication granules enhances absorption into the body, and particle size determines strength and performance of ceramic materials. The detection of particles is also important in areas of modern society, such as environmental protection and public health. For example, inhalation of ultra-fine particles originating from emissions of various kinds can lead to a number of adverse health effects, including inheritable genetic changes.
Currently, the advent of nanoscience and nanotechnology has made it increasingly important to reliably assess the size of nanometer scale particles. Nanoparticles find use in many areas, such as diagnostics and treatment of tumors, treatment systems for radioactive and biohazard materials, solar power energy conversion, electronic circuits, sensors, lasers, artificial bone implants and others. See, for example, Loo et al., “Nanoshell-Enabled Photonics-Based Imaging and Therapy of Cancer,” Techol. Cancer Res. T., vol. 3, no. 1, pp. 33-40, 2004.
Because of their small size, nanoparticles are not easy to detect, and it is evident that there is high demand for novel techniques for the reliable detection, characterization, sorting, and tracking of nanoscale particles of various sorts. Furthermore, as the feature size of integrated circuits becomes increasingly smaller, contamination control of ultrafine particles poses a challenge for the semiconductor industry.
A nanoparticle detector is especially important for biowarfare detection. This type of warfare is particularly devastating due to the potential for rapid infection from a small amount of biological agents. One need only look at the disruption to the U.S. federal government caused by the mailing of anthrax spores, or to the economic harm caused in many countries due to the outbreak of severe acute respiratory syndrome (SARS), to realize the magnitude of such a threat. Warfare viruses are especially dangerous because no cures exist against many viruses. An early detection is one of the few defenses against such threats. A broad network of sensors, cheap and robust enough to be placed throughout public spaces with credible threats of attack, can provide a reliable early warning of an attack.
The field of particle sizing science is very broad. A database of American Society for Testing and Materials contains over 140 particle sizing methods which have evolved over the past number of years. These methods can be classified into sieving, image analysis, fluid classification, and interaction between particles and external fields. Sieving has been used for thousands of years and is still widely used in industry to sort particles based solely on their sizes. Particles are analyzed by essentially sifting the sample powder through a stack of sieves. Image analysis methods measure particle dimensions from images acquired with optical and electron microscopes. Fluid classification methods include gravitational and centrifugal sedimentation methods, which are based on the settling behavior of particles in a suspension under gravitational or centripetal force. Finally, there are techniques based on interaction between particles and external fields include interactions with electrostatic fields, electromagnetic (optical) and acoustic waves. Most of the developed particle measurement systems are designed to measure micrometer or above diameter particles.
Some optical methods are, however, capable of detecting sub-micron particles. Optical methods for particle detection rely on light scattering. See, for example, L. Fabiny, “Optical Particle Counters,” Opt. Phot. News, vol. 9, pp. 34-38, 1998. In the simplest version, an optical particle counter (OPC) includes a light source, usually a laser, which illuminates a sample volume containing particles of interest. The particles scatter light, which is collected by an off-axis detector. The angular distribution of the scattered light intensity is a function of a number of parameters such as particle size, shape, optical density and concentration. These parameters can be extracted from the measured data by solving the inverse Mie scattering problem. Beyond the basic design, there are many variations of OPCs, some of which count individual particles and others measure ensemble average. Examples of single particle counters are a Flow Cytometer, a Phase Doppler Anemometer (PDA) and some versions of Condensation Nuclei Counters (CNC). Examples of OPCs which measure ensemble average are Dynamic Light Scattering (DLS) sensors, Nephelometers (or multiangle photometer) and other versions of the CNCs.
The configuration of a typical optical particle counter is illustrated in
Most optical particle counting systems are only sensitive to particles above 200 nm. There are only two optical methods capable of measuring nanoparticles below 100 nm in size: the CNC and the DLS sensors. In the CNC method, saturated vapors of water or alcohol are used to grow bubbles around nanoparticles. This way, particles grow in size and become accessible by other optical detection techniques. It is, however, very difficult to grow bubbles in a controlled manner, thus the original particle size information is often unavailable. The DLS method measures the Brownian motion dynamics of particles by monitoring the time fluctuations of a total number of particles within a small volume. Smaller particles enter or leave the monitored volume more often than the larger particles. Therefore, the time autocorrelation of the measured signal contains information about particle size. The DLS method is capable of measuring particle sizes down to 2-3 nanometers in size, is independent of the optical properties of the particles and is very effective in analyzing monodisperse samples. However, the precision of the DLS size measurements decreases with the polydispersity of particle sizes in a sample. Also, since the DLS sensors measure ensemble averages, they require high particle concentrations. For example, state of the art systems can measure concentrations down to 0.1 mg/ml, which correspond to 2×1013 particles/ml for 20 nm polystyrene beads. The cost and complexity of measurements grow quickly as particle size approaches a few tens of nanometers.
There are, however, only a few projects that are aimed at developing OPCs with single particle sensitivity below 100 nm. One such group was able to optimize the standard OPC configuration to detect polystyrene particles down to 90 nm in diameter by minimized light scattered by the media and optical elements which contribute to noise level. See, M. Hercher et al., “Detection and Discrimination of Individual Viruses by Flow Cytometry,” vol. 27, pp. 350-352, 1979. However, the complexity of their setup precludes practical applications of the system in scientific laboratories, as well as in commercial production. Additionally, 74 nm diameter polystyrene spheres have been detected using essentially an inverse dark-field configuration. See, H. Steen, “Flow Cytometer for Measurement of the Light Scattering of Viral and Other Submicroscopic Particles,” vol. 57A, pp. 94-99, 2004. The incident beam in this setup is blocked by placing a field stop in the center of the exit pupil of a collection objective, while light scattered at higher angles was collected.
In both of the above-discussed projects, as well as in other optical detection methods, a very strong dependence of the detected signal on particle size is a main obstacle in detecting nanoparticles. The scattering cross-section for a particle much smaller than the wavelength of excitation source is:
where R is the particle radius, ∈p and ∈m are the dielectric permittivities of the sphere and the surrounding medium, respectively, and λ is the wavelength of the light. Therefore, the signal to noise ratio decreases very rapidly with particle size. In order to lower the detectable particle size in currently available instruments from 200 nm to 20 nm, the noise level would have to be reduced by six orders of magnitude, which is not realistic.
In addition to size of the particles, it is often necessary to know the particle composition. It is relatively easy to analyze bulk materials, whether by looking at their optical (light absorption, fluorescence), physical (stiffness, elasticity) or chemical (solubility, reactivity) properties or atomic composition (percentage amount of carbon, nitrogen or other atoms). Therefore, it is also possible to identify materials in a highly concentrated particle sample. That can be done, for example, by acquiring absorption spectra in a spectrophotometer, or by collecting fluorescence or Raman scattering spectra. These methods, however, are not suitable for single particle identification. Physical properties cannot be accessed on the nanoscale level, chemical reactions cannot be monitored using such small volumes of reagents. Raman scattering and fluorescence cross-sections are very small and do not enable enough information to be collected from single nanoparticles. Only recently have near-field methods been developed that are capable of detecting absorption, luminescence and Raman scattering from a single nanoparticle immobilized on a surface. In principle, X-ray microanalysis can be used to obtain atomic structure of materials, but such method requires expensive equipment, cumbersome sample preparation and lacks high throughput.
Although these methods extend the detection sensitivity to smaller particle sizes, they suffer from other shortcomings which prevent the detection of single nanoparticles in real time. Either they require particle immobilization to ensure sufficiently long acquisition times or they are subject to a background signal originating from Brownian motion or direct detector exposure. Therefore, a new detection scheme is needed for the recognition of viruses and other nanoparticles. Such a scheme needs to provide accurate, simple and affordable ways of detecting small nanoparticles and biological agents. Such detection devices also need to be capable of obtaining chemical signatures and identifying particles with high specificity.
It is therefore an object of the invention to provide a technique for measuring nanoparticles which overcomes the above-noted shortcomings.
To achieve the above and other objects, the present invention is directed to a background-free detection approach which gives unsurpassed real-time detection sensitivity for nanoscale particles. The successful detection and classification of low-index particles has been demonstrated. The detection scheme is well suited for the screening and sorting of various nanoscale particles such as viruses and other bodies and is compatible with current microfluidic technology.
According to at least one embodiment, the invention is directed to a method for detecting a particle in a location. The method includes emitting electromagnetic radiation, splitting the electromagnetic radiation into a first component and a second component, directing the first component into a reference arm and directing the second component into the location. The method further includes receiving light backscattered from the location, causing the backscattered light to interfere with the first component from the reference arm to produce an interference intensity distribution using at least two wavelengths, detecting the interference intensity distribution with a detector at the at least two wavelengths and detecting the particle in accordance with a difference among detection signals.
In addition, the light may come from a white light source and a plurality of paired photodetectors. The interference pattern may also differentiate different wavelengths of light into different angles using an optical grating. Also, the light may come from multiple lasers and be detected through multiple split detectors. The number of lasers of the multiple lasers and the number of split detectors of the multiple split detectors may be equal. The method may also include oscillating a position of a mirror in the reference arm to modulate the phase of the first component. A frequency dependence of the particle's polarizability may be sampled and a composition of the particle may be predicted.
According to at least another embodiment, a system for detecting a particle in a location includes a source of electromagnetic radiation, a beam splitter for splitting the electromagnetic radiation into a first component and a second component and a reference arm receiving the first component from the beam splitter. The system also includes focusing optics, receiving the second component from the beam splitter, for directing the second component into the location and for receiving light backscattered from the location, thereby causing the backscattered light to interfere with the first component from the reference arm to produce an interference intensity distribution, a detector comprising a plurality of components for detecting the interference intensity distribution at the at least two wavelengths and a data acquisition system for detecting the particle determining the particle's absolute size in accordance with a difference among detection signals from the plurality of components.
A preferred embodiment of the present invention will be set forth in detail with reference to the drawings, in which:
a) shows a typical photo-detector signal,
The present invention will be set forth in detail with reference to the drawings, in which like reference numerals refer to like elements or operational steps throughout.
Particles are moving through the focus via liquid flow. As a particle travels through the focus, the incident light is scattered back. The scattered light is collected by the focusing objective 110 and is recombined with the reference beam using the same beamsplitter. Both the reference beam and the scattered light are then incident on the optical grating 112, which separates light of different wavelengths into different angles. Such angular light arrangement is then collimated by a computer-controlled holographic optical element 114 and is collected by an array of paired detectors 116. For each wavelength, scattered light and light from the reference arm create an interferometric pattern on the corresponding pair of detectors. The time-varying signal from the detectors is collected by a computer using a high-speed data acquisition card. The size and the spectral signature of the particles are extracted by analyzing the acquired signal. The detector signal is about one millisecond long, which is about the time it takes for a single particle to cross the focus of the focusing objective 110.
Parts of the setup can be varied depending on the performance requirements and costs. For example, two or more single-frequency lasers 301 & 302 can be used instead of a white light source. Such an embodiment 300 is illustrated in
The present invention operates under three main principles of operation to achieve high sensitivity and specificity in particle detection. First, by interfering the scattered light and the incident beam on a split photodetector, or on a single pair of detectors within an array of detectors, the scattered light amplitude is measured. The scattered light amplitude is proportional to the third power of the particle size, i.e. R3. Currently marketed particle sensors detect scattered light power, which is proportional to R6. The weaker particle size dependence leads to a higher signal-to-noise ratio for smaller particles, compared to R6 methods.
Second, the pseudoheterodyne detection approach removes noise associated with phase variations. Interferometric measurements are usually very sensitive to phase differences between two interface beams. For example, air currents can change the effective path length in the arms of an interferometer, thus leading to phase changes. Small vibrations in optical elements can also lead to large phase variations. In the present invention, the largest measurement error comes from the phase variations within the light focus. Even in small nanochannels, the phase of the light scattered by the particles varies rapidly depending on a particle's trajectory. Such noise leads to errors in particle size measurements and limits resolving power of the sensor. The use of the oscillating mirror in the reference arm and a smart detection algorithm helps eliminate phase error contributions to the measured signals.
Third, multiple frequency light is used to probe the particle scattering efficiency (or particle polarizability) at different frequencies. This information is unique for each material and can be used to identify particle composition. The inelastic scattering cross-section is much higher compared to common inelastic scattering techniques of fluorescence and Raman scattering. Thus, this leads to a higher signal-to-noise ratio for small particles and the ability to measure the spectral properties of single particles within the millisecond time frame. To measure spectral properties, interferometric data for each wavelength is collected independently. The sensor embodiment in
A nanoparticle placed in a laser field acts as a dipole with the induced dipole moment:
where E0 is the excitation electric field. The electric field scattered by a nanoparticle is proportional to the induced dipole moment and is therefore proportional to the third power of particle size. A typical photodetector cannot be used to directly measure scattered field amplitude, because it measures power or amplitude squared of the incident light. However, if the scattered and excitation light are interfered on a split photodetector, the difference in signal from one half and the opposite half of the detector (A-B) is proportional to the scattered amplitude.
A split photodetector can be formed from a PIN photodiode with a circularly shaped detection area, which is divided into two equal parts by small (10-30 mm) insulator gaps. Each half is independent from the other and has its own output. A quadrant detector, which has four independent parts instead of two, may be used, where the quadrant detector is also called a position sensing detector (PSD). A quadrant detector can be converted into a split photodetector by connecting adjacent quadrants.
Denoting E and Es as the reference and scattered light fields, respectively, on the split photodetector, the two fields create an interference pattern. The intensity distribution of such a pattern is:
I(x,y)=|E+Es|2=|E|2+2Re(E·Es)+|Es|2 (3)
Typically, the scatter field intensity is much smaller than the laser intensity and thus |Es|2 can be neglected compared to the other terms in eq. (3), thus:
I(x,y)=|E|2+2Re(E·Es) (4)
The differential detector signal, S=A-B, is obtained by integrating the intensity distribution over the corresponding halves of the split detector:
where α denotes the integration area, ⊂ and ⊃ denote the two halves of the photodetector surface and ∘ denotes the entire photodetector surface. In the absence of a passing particle, the reference beam and the light backreflected by optical elements are adjusted into the center of the split photodetector such that the differential signal S is zero. The interference between the reference beam and the backreflected light does not affect the detection method because it is stationary and therefore does not generate any differential signal. Thus, S is a background-free signal similar to fluorescence that is commonly used to detect and track single molecules.
Using equations (4) and (5):
Assuming that the reference beam spot is positioned at the center of the photodetector, the intensity distribution is due to |E|2 being symmetric with respect to the insulating gap on the detector and thus the first two terms of eq. (6) cancel each other and:
It can be seen that the differential signal in eq. (7) is proportional to the scattered field and therefore depends on the third power of the particle size. In order to make the signal insensitive to the noise in the laser power, it can be normalized to the total power incident on the detector:
The normalized differential signal is then proportional to the ratio between the scattered field strength and the laser field strength:
The scattered field, Es, is proportional to the particle's dipole moment, p, and therefore to the electric field in the microscope objective focus, E0, therefore:
As a particle passes through the focus, the amount of scattered light varies depending on where the particle is located with respect to the center of the focus, resulting in a non-zero time-dependent photodetector signal. The amplitude of the signal is constant for particles of the same size, given the maximum illumination conditions are the same within the nanochannel.
The signal-to-noise ratio (SNR) found from the use of the present invention should be compared with that of the standard scattering-based detection. Ultimately, the highest SNR can be achieved when the background light acts as a reference beam. The absolute maximum of the interferometric amplitude in this case is achieved when the interference patterns concentrate all of the energy on one half of the split photodetector (s=1). That can only happen if the scattered field amplitude is equal to the amplitude of the background light Eb, i.e. |Es|=|Eb|. For sufficiently strong powers, the SNR becomes {S/N}=(1/θ) Es/Eb, where t is the angular pointing instability of the light source beam. On the other hand, the maximum SNR in standard light scattering can be written as {S/N}=(1/η)Es2/Eb2, where η is the laser power noise.
First, the SNR in the present invention is proportional to Es2/Eb2, versus Es2/Eb2 for scattering-based detection, and therefore proportional to the third power of the particle size, versus the sixth power of the particle size for scattering-based approaches. Second, the SNR in standard light scattering methods depends on laser power noise, which cannot be easily controlled. On the other hand, the present invention does depend on the angular pointing stability of the laser which can be controlled, for example, by reducing the optical path length. Furthermore, the dimensionless pointing instability coefficient θ for the laser is much smaller (by orders of magnitude) than typical noise power.
With respect to pseudoheterodyne detection, if E=E exp(iω0t+φ) and Es=Es exp(iω0t+φs), the differential signal in eq. (7) becomes:
S∝EEs cos(φ−φs) (11)
where ω0 is the optical frequency, and φ and φs are phases of the scattered and reference beams. It is immediately clear that the signal S not only depends on the amount of scattered light Es, but also on the phase φs. Because of the harmonic behavior, the signal S can change with the full dynamic range of −EEs to +EEs. Small variations in phase can result in large measurement errors in the amplitude and therefore in size. Pseudoheterodyne detection allows for the separate measurement of EEs and cos(φ−φs).
The basic idea behind pseudoheterodyne detection is to modulate the phase of one of the beams with high frequency. This modulation is implemented by oscillating the position of the reference beam mirror, as shown in
E=Eexp[iφ0t+ik0x0 sin(ωmt)+iφs] (12)
where ωm is the modulation frequency, x is the modulation amplitude and k0x is the phase amplitude due to modulation. The measured differential signal is then:
S∝EEs cos(k0x0 sin(ωmt)+φs−φ) (13)
=EEs[cos(k0x0 sin(φmt))cos(φ−φs)−sin(k0x0 sin(ωmt))sin(φ−φs)] (14)
Using the mathematical expression:
the signal S can be expanded into multiple harmonics, oscillating at frequencies nωm. The amplitude of each harmonic can be extracted using a lock-in amplifier. Therefore, the amplitude of the first harmonic is:
S(1ωm)=EEs[J1(k0x0)sin(φ−φs)] (16)
and the second harmonic is:
S(2ωm)=EEs[J2(k0x0)cos(φ−φs)] (17)
Assuming that x0 can be adjusted to satisfy the relationship J1(k0x0)=J2(k0x0), then the first and second harmonics can be squared and added to together, giving:
S
heterodyne=√{square root over (S2(1ωm)+S2(2ωm))}{square root over (S2(1ωm)+S2(2ωm))}=EEsJ1(k0x0) (18)
It can be seen that the last expression does not contain any phase terms. By measuring heterodyne amplitude, the phase dependence is eliminated and errors associated with phase variations within the focus can also be eliminated. The resolving power, or how close can particles be in size to be separately recognized, is therefore improved.
With respect to multi-color detection, when white light or multiple lasers are used to illuminate particles in the focus, the induced dipole moment should be rewritten as:
where ∈p(ω) describe dielectric properties of a particle at different light frequencies Co. The shape of ∈p(ω) uniquely identifies the composition material of the particle. When many excitation wavelengths (colors) are used in the present invention, ∈p(ω) is probed at those wavelengths. The sensor illustrated in
At each wavelength at which the pseudoheterodyne interferometric amplitude is measured, information can be derived about particle size R and dielectric properties ∈p(ω). The more wavelengths that are probed, the more precise a measure of the particle's size and material can be extracted.
The interference between the scattered light and the frequency-shifted reference light gives rise to a split detector signal oscillations with frequency ωm. The amplitude of the oscillations is modulated depending on the amount of light scattered by the particle.
Similar to the other embodiments, the end result is the phase-insensitive signal, where the maximum amplitude is directly proportional to the third power of particle size and to particle's optical properties at the wavelength of the probing light. Similar to the other preferred embodiments, the split photodetectos renders zero (background-free) signal when particle is absent in the focus, i.e. the interferences between the reference beam and the background reflections from the optical elements or fluidic interfaces do not result in oscillations at the output of the split photodetector (when the photodetector is properly aligned). Similar to the other embodiments, the detection bandwidth is shifted to a higher frequency where less noise is present.
When compared with the other embodiments, the embodiment illustrated in
An alternate embodiment is also provided in
Such an approach can be extended to three or more lasers (or to a white light source) used with three or more AOM's operating at different modulation frequencies, and a single split detector. Illustrative of that is
The present invention establishes new strategies in ultrasensitive particle and virus detection, and will provide new tools relevant to nanoscience and nanotechnology. In addition to detection of agents used in biowarfare and terrorism, the present invention also has applications ranging from contamination control of water, ultrasensitive flow cytometry and environmental monitoring of pollutants.
While a preferred embodiment of the invention has been set forth above, those skilled in the art who have reviewed the present disclosure will readily appreciate that other embodiments can be realized within the scope of the invention. For example, numerical values are illustrative rather than limiting, as are specific techniques for attenuation and the like. Therefore, the present invention should be construed as limited only by the appended claims.
The present application claims the benefit of U.S. Provisional Patent Application No. 60/776,953, filed Feb. 28, 2006, whose disclosure is hereby incorporated by reference in its entirety into the present disclosure.
The work leading to the present invention was supported by NSF Grant No. PHS-0441964. The government has certain rights in the invention.
Number | Date | Country | |
---|---|---|---|
60776953 | Feb 2006 | US |